Академический Документы
Профессиональный Документы
Культура Документы
ANALYTICAL CHEMISTRY
DEVELOPMENTS, APPLICATIONS
AND CHALLENGES IN FOOD ANALYSIS
No part of this digital document may be reproduced, stored in a retrieval system or transmitted in any form or
by any means. The publisher has taken reasonable care in the preparation of this digital document, but makes no
expressed or implied warranty of any kind and assumes no responsibility for any errors or omissions. No
liability is assumed for incidental or consequential damages in connection with or arising out of information
contained herein. This digital document is sold with the clear understanding that the publisher is not engaged in
rendering legal, medical or any other professional services.
ANALYTICAL CHEMISTRY
DEVELOPMENTS, APPLICATIONS
AND CHALLENGES IN FOOD ANALYSIS
MARCELLO LOCATELLI
AND
CHRISTIAN CELIA
EDITORS
All rights reserved. No part of this book may be reproduced, stored in a retrieval system or
transmitted in any form or by any means: electronic, electrostatic, magnetic, tape, mechanical
photocopying, recording or otherwise without the written permission of the Publisher.
We have partnered with Copyright Clearance Center to make it easy for you to obtain permissions
to reuse content from this publication. Simply navigate to this publication’s page on Nova’s website
and locate the “Get Permission” button below the title description. This button is linked directly to
the title’s permission page on copyright.com. Alternatively, you can visit copyright.com and search
by title, ISBN, or ISSN.
For further questions about using the service on copyright.com, please contact:
Copyright Clearance Center
Phone: +1-(978) 750-8400 Fax: +1-(978) 750-4470 E-mail: info@copyright.com.
Independent verification should be sought for any data, advice or recommendations contained in
this book. In addition, no responsibility is assumed by the publisher for any injury and/or damage
to persons or property arising from any methods, products, instructions, ideas or otherwise contained
in this publication.
This publication is designed to provide accurate and authoritative information with regard to the
subject matter covered herein. It is sold with the clear understanding that the Publisher is not
engaged in rendering legal or any other professional services. If legal or any other expert assistance
is required, the services of a competent person should be sought. FROM A DECLARATION OF
PARTICIPANTS JOINTLY ADOPTED BY A COMMITTEE OF THE AMERICAN BAR
ASSOCIATION AND A COMMITTEE OF PUBLISHERS.
Additional color graphics may be available in the e-book version of this book.
Preface ix
Chapter 1 Method Validation and Hyphenated Techniques:
Recent Trends and Future Perspectives 1
Marcello Locatelli, Roberta Cifelli, Sonia Vitalei,
Pierluigi Santini, Elisa De Luca, Giuseppe Bellagamba,
Christian Celia, Simone Carradori, Luisa Di Marzio
and Adriano Mollica
Chapter 2 Sample Preparation in Food Analysis:
Practices, Problems and Future Outlook 23
Abuzar Kabir and Kenneth G. Furton
Chapter 3 Applications of Magnetic Nanoparticles for the Selective
Extraction of Trace Species from a Complex Matrix 55
Halil İbrahim Ulusoy
Chapter 4 A Review of Phenolic Compounds from Medicinal Plants
and Nutraceuticals, and Their Characterization by Different
Antioxidant Assays 77
Gokhan Zengin, Andrei Mocan, Sengul Uysal,
Ramazan Ceylan, Gianina Crișan
and Abdurrahman Aktumsek
Chapter 5 NMR Methodologies in Food Analysis 103
Luisa Mannina, Anatoly Petrovich Sobolev, Violetta Aru,
Alessia Bellomaria, Fabio Bertocchi, Bruno Botta,
Laura Ruth Cagliani, Augusta Caligiani, Francesco Capozzi,
Dorisa Çela, Flaminia Cesare Marincola, Alessandra Ciampa,
Laura Del Coco, Roberto Consonni, Carmelo Corsaro,
Maurizio Delfini, Valeria Di Tullio, Francesco Paolo Fanizzi,
and food supplements analysis and are a proceeding of chapter 5. Chapter 8 describes the
voltammetry and atomic absorption spectroscopy for the analysis of toxic metals in
seafood, and food supplements. Chapter 9 includes the application of thermal techniques
in foods and food supplements. Chapter 10 includes the analysis of proteins in foods, and
innovative approaches to isolate, purify and characterize proteins in foods and food
supplements. Analytical chemists and researchers working in the field of validation
methods, foods and food supplements, and using standard and innovative instruments to
analyze products; and/or innovative extraction procedures, could be a potential audience
of these chapters.
Dr. Marcello Locatelli and Prof. Cristiano Celia, professors at the "G. d'Annunzio"
University of Chieti-Pescara at the Department of Pharmacy, have been curating this
edition with the collaboration of Italian and foreign colleagues, internationally renowned
in the field of analytical chemistry and analysis of food and nutritional supplements.
Experts in the field of extraction techniques for quantitative analysis and new instrumental
configurations that can also be adopted to provide an important contribution in order to
obtain an up-to-date result with regard to the latest discoveries and applications in this field.
Chapter 1
METHOD VALIDATION
AND HYPHENATED TECHNIQUES:
RECENT TRENDS AND FUTURE PERSPECTIVES
ABSTRACT
*
Corresponding Author: marcello.locatelli@unich.it.
quality of results and still represents the “Achilles heel” in bioanalytical chemistry using
the hyphenated techniques.
Actually, several papers report the potential drawbacks of ME, and provide suitable
strategies to suppress this phenomenon. Furthermore, different factors, e.g., the co-
precipitation of biological samples, interfaces, chromatography and spectrometric
conditions, can affect the bioanalytical assays and validation processes at different steps.
This chapter discusses the potential strategies to suppress the ME and highlights the
importance of validation process according to International Guidelines. In this attempt, the
best solution should be the combination of on-line techniques miniaturized systems, new
ionisation sources, chemometrics approach, high performance liquid chromatography
based and hyphenated techniques.
LIST OF ABBREVIATIONS
CV = Coefficient of Variation;
DAD = Diode Array Detector;
ELISA = Enzyme-Linked ImmunoSorbent Assay;
FLD = Fluorescent Detector;
GC = Gas-Chromatography;
HILIC = Hydrophilic Interaction Liquid Chromatography;
HPLC = High Performance Liquid Chromatography;
HPLC-MS = High Performance Liquid Chromatography-Mass Spectrometry;
ICH = International Conference of Harmonization;
LLE = Liquid-Liquid Extraction;
LOQ = Limit of Quantification;
ME = Matrix Effect;
MRM = Multiple Reactions Monitoring;
MS = Mass Spectrometer;
PE = Process Efficiency;
PP = Precipitation of Proteins;
QCs = Quality Control Samples;
RE = Recovery;
RIA = Radioimmunoassay;
SIM = Single Ion Monitoring;
SPE = Solid Phase Extraction;
SRM = Single Reactions Monitoring;
UHPLC = Ultra High Performance Liquid Chromatography.
1. INTRODUCTION
Analytical chemistry allows to develop suitable methods for analysis of real samples.
This throughput plays a main role in drug discovery and development. The analytical data
are used to set up potential applications in the fields of medicine, synthesis, pharmaceutics,
and test stability. The standardization of data set plays a main role to develop analytical
methods and analyze new chemical entities (NCE) successfully. In particular, the
development and validation of method can affect through the quality of data. The validation
of method also represents a complex process depending on the analytical procedures and
instrumentations.
The guidelines of United States Pharmacopoeia (USP), the International Conference
on Harmonisation (ICH) and the Food and Drug Administration (FDA), provide a
framework to apply this process for pharmaceutical analysis. Basically, the analytical
method should include the following parameters of analysis and particularly, specificity,
linearity, accuracy, robustness of drug concentration range, limit of quantification (LOQ)
and limit of detection (LOD). Although there are main parameters that should be evaluated
to validate analytical methods, there is actually a great variability of approaches in the final
reports.
Recently, several papers report suitable procedures to validate analytical methods. In
particular, Kruve and coworkers revised the standard guidelines to edit the validation of
analytical method [1, 2], even if they do not have included all parameters for the analysis.
It is extremely important to collect a set of data for the best performance of analytical
protocols.
The bioanalytical techniques include instrumental analysis, e.g., gas-chromatography
(GC), high performance liquid chromatography coupled to mass detector spectrometry
(HPLC-MS), and biological assayes, e.g., radioimmunoassay (RIA), enzyme-linked
immuno sorbent assay (ELISA). To improve the quality of anlysis, various orthogonal
instrumental methods are combined to detect and quantify unknown compounds dispersed
in biological samples. The combinations of these methods shoot down the ME and its
related drawbacks [3].
The ME was firstly described by Tang and Kebarle [4], which reported the detection
of a standard solution dispersed in biological samples, e.g., plasma, and the modification
of unknown peak under different experimental conditions [5]. Furthermore, the IUPAC
settled the ME as a combination of different effects derived from sample components and
a quantitative measurement of analytes [6].
This chapter reports an overview of current approaches to validate analytical studies
and provides a dossier to apply chromatographic assay in pharmaceutical analysis. The
validation procedures depend on the NCE setting up, characterization and performances.
2.1. Selectivity
Selectivity and specificity are generally used as synonymous but they indicate different
parameters [11-13]. The selectivity allows to separate and quantify unknown compounds
in biological samples with or without additional components. Blank samples such as
plasma, urine, and blood are used to perform and validate the selectivity of analytical
methods. Additional components included in biological samples, which represent potential
sources of interferences, are endogenous factors, e.g., age, sex, race, geographical location,
metabolites, degradation products, drug association, and xenobiotics. The ME is also a
potential source of interferences by applying the hyphenated instrument configurations,
such as high performance liquid chromatography coupled to mass spectrometry detector
(HPLC-MS). The performance of selectivity test is carried out by developing the
chromatographic conditions in presence of blank samples at a drug concentration closed to
the LOQ [14].
According to the ICH guidelines [8], the blank samples should have an area under the
curve (AUC) neither over 20% of LOQ areas at a specific retention time of analytes, nor
an AUC over 5% of internal standard (IS) area at this specific retention time. These
parameters are important to set up the bioanalytical methods.
2.2. Accuracy
2.2.1 Trueness
The trueness indicates the narrow distribution of values, which are obtained during the
development of bioanalytical methods. This parameter depends on systematic errors and is
carried out as bias, or relative bias. Sometimes, the trueness can be evaluated as recovery
[1, 2, 11]. The trueness was carried out by analyzing replicates of samples added with
different amounts of analytes, such as quality control samples (QCs). This parameter can
be calculated at three different concentration and the limit of quantification (LOQ) levels
by using at least five repetitive analysis per concentration. The mean value should be
obtained within ±15% of the nominal value except for LOQ, where this parameter should
not be over ±20% [8].
The trueness can be also measured within intra-batch and intra-day, for single batch
analysis; and between inter-batch, inter-day and over time for the complete run of analysis.
Figure 1. Accuracy shows as the simultaneous determination of precision and trueness according to
FDA guidelines.
2.2.2 Precision
The precision reports the analysis of replicates in homogeneous volume of single
biological samples. The precision is carried out by analyzing replicates of samples added
with different amounts of analytes, such as quality control samples (QCs). This parameter
can be determined at three different concentration and the limit of quantification (LOQ)
levels by using at least five determinations per concentration. The precision is carried out
at different concentration levels, which should not exceed 15% of the coefficient of
variation (CV) except for the LOQ. The LOQ should not exceed 20% of the CV. The
precision can be also measured within intra-batch and intra-day, for single batch analysis;
and between inter-batch, inter-day, and over time for the complete run of analysis [8].
2.3. Recovery
Anf Aadd
Re cov ery (%) x100
A f
The calibration curve represents the correlation between the instrument response and
concentrations of analytes in the samples, which are used to calculate the calibration curve.
The correlation between response and concentration of standards should be continuous and
reproducible within a specific concentration range of anlytes.
Figure 2. Different statistical approaches used to calculate three calibration curves and linearity using
the same set of data.
Basically, the calibration curve was calculated and derived using the biological or
blank samples performed during the analysis. A standard concentration of analytes
wasanlytes is spiked in the blank sample and calibration curve is carried out as a result of
analyte response in the linear range of analysis. However, the calibration curve can be
further calculated using a synthetic sample showing similar performance of real samples.
Figure 2 shows three different calibration curves, which were performed using the
same set of data. The calibration curves are plotted using different statistical approaches,
such as unweighted model to 1/x and to 1/x2 statistical approach. The use of weighted linear
least-squares regression analysis, according to method validation guidelines, demonstrated
that the fitting of standard curve depends on the concentration-response [7]. This procedure
allows maintaining the sensitivity of the method in all steps of analysis and can be used to
perform the pharmacokinetic and/or bioequivalence studies and minimize drawbacks of
diluted samples.
The weighting factor was applied to x-axis (concentration values) and y-axis
(instrumental response). This factor was previously reported by Meyer [18], which used
the Monte Carlo simulation to calculate the weighted linear least-square fit according to
following parameters:
A linear correlation was evaluated through the range of analysis testing a series
concentration of a standard solution and/or weighting synthetic mixture of lead
compounds. The linearity was calculated plotting the Area under the curve (AUC or y-
axis) and the concentration of drug (x-axis). A statistical test is further applied to calculate
the significance of data, which show a linear correlation throughout the analysis. The least
square was used to evaluate the linear regression, the correlation coefficient and the lack-
of-fit test (LoF test). Awesome, for enzymatic assay, the linearity was tested plotting the
response of method and sample concentrations. However, results must be converted using
mathematical algorithms and the resulting data are used to calculate the regression model.
At least 5 standards at different concentrations is recommended to test the linearity of
analytical method.
2.5. Range
The range of analysis took from linearity and depends on the analytical procedure,
which provides a suitable degree of linearity and accuracy (precision and trueness) if
analytes are in the lineer range of range of Method.
The limit of quantification is the minimum amount of analyte, which can be quantified
in samples using a precision and accuracy of ±20%. The limit of quantification is a specific
parameter for a quantitative analysis at low levels of compounds in samples. LOQ is
especially useful in the presence of contaminantes and/or degradation products in unknown
samples during the analysis. There are several approaches to calculate the limit of
quantitation (Table 1).
Figure 3. LOD and LOQ tests using a blank standard deviation [8].
The LOD and LOQ can be also carried out using the blank standard deviation. In
particular, LOD represents a 3-folds blank standard deviation, while LOQ represents 10-
folds blank standard deviation.
The detection limit is a statistical test of hypothesis, which provided two different
parameters α and β. α and β represented the probability to detect the presence of unspecific
and specific compounds in biological samples, respectively. α and β was shown as value
of 0.05 (Figure 3).
2.7. Robustness
Flow rate;
Mobile phase composition;
Organic phase percentage;
Column temperature used to separate compounds;
Other parameters, e.g., dedector response, can also affect the validation procedure.
Sometimes, the ruggedness is used to indicate the robustness but these parameters are
not comparable according to several international guidelines. The United States
Pharmacopea (USP) 23rd Edition discriminates two parameters [19]; conversely, this
difference is not included in the USP 35th Edition [20]. The European Medicine Agency
(EMA) and Food and Drug Administration (FDA) guidelines missed this statement.
Recently, a review article discussed the importance of robustness and different approaches
to evaluate ruggedness and robustness [21].
The stability test provides suitable information about the stability of unknown
compounds in biological samples. The FDA guidelines for method validation [7] and
AAPS/FDA white paper [22] require performing the stability test of compounds at different
stages and different concentration levels. In particular, the following parameters must be
tested to evaluate the stability test of unknown compounds in biological samples:
3 Freezing-thawing cycles;
Bench-top stability under sample preparation condition or room temperature;
Long term stability at -20°C or -70°C for tissue samples;
Stability in the autosampler, at least for the sample-list runtime.
The matrix composition is one of the main parameters limiting step of bioanalytical
chemistry and quantitative analysis [24]. In fact, the biological matrices contain different
components, which can form complexes with analytes. So, samples must be purified and
extracted before the analysis. The composition of biological samples can also affect the
instrumental setting-up and the performance of apparatus.
The ME is described for the first time by Tang and Kebarle [4], which obtained
different results using a mass spectroscopy detector by comparing a standard solution of
compounds and the same compound added to biological samples [5]. Differences in
analytical signals between standard solution and biological samples may depended on a
potential competition between analyte species and components of matrices for stationary
and mobile phases.
The standard definition of ME according to IUPAC is below reported. The ME is “a
series of different effects derived from sample components, which are detected over the
quantitative measure of analytes.” Components, which derived from ME and were not
included in biological samples, represented the interferences [25]. Basically, it is important
to discriminate the ME meaning and interferences; in fact, interferences represented a pre-
determined constant error, which affects the analyte response; while the ME is the response
of analyte and depends on its concentration [6]. For examples, xenobiotics and metabolites
are internal interferences arising from plastic materials, e.g., syringes or containers. The
interferences arising from mass spectrometry and biological samples were suppressed by
changing native compounds in their ionic counterparts, which modify the ionic effect of
analytes and biological samples, and improve the detection through ionic detector coupled
to mass spectrometry [26]. The physicochemical properties of solvents can also affect the
ionization strange and evaporation, thus providing compounds suitable for the analysis.
The sources of ion suppression include organic and inorganic compounds, which have a
structure similar to unknown samples and compounds, which may be included in the
sample during its manufacture, e.g., residue of polymers, detergents, buffers. Differences
for analyses were also carried out using non-homogeneous biological samples, e.g., plants
and biological fluids, or combining different techniques, e.g., direct injection, protein
precipitation (PP), solid phase extraction (SPE), liquid-liquid extraction (LLE) [27].
The ion suppression depends on different mechanisms, which can improve the signal
response of chromatograms collected using mass spectrometry coupled to different
apparatus. In particular the ion suppression can depend on:
provide the ionic suppression [32]. The impact of mobile phase on sample analysis
can also depend on the instrument setting-up, e.g., the flow rate, type of column
and affinity of compounds for the stationary phase. Furthermore, the ionic
derivatives are neutralized interacting with basic compounds, which suppressed its
native signal;
The ionization of samples at the interface for HPLC-MS and HPLC-MS/MS (ESI
and APCI) and its relative reaction depends on the energy intensity and proton
affinity of sprayed gas for molecules and can compete with interferences. The ionic
suppression also depended on changes of transfer efficiency from liquid to gas,
and/or charge transfer in APCI;
The signal suppression depends on mass spectrometry and procedures.
Interferences can generate ionic derivatives or fragments, with m/z ratio similar to
unknown compounds, which are co-eluted with samples, and provide false positive
results. These drawbacks occured frequently using single quadrupole detector in
single ion monitoring (SIM), rather than a triple quadrupole detector in mode-
selected reaction monitoring (SRM), which increased the specificity and
sensitivity of analysis.
3.2. ME Results
There are several effects of ion suppression and instrumental responses in bioanalytical
chemistry:
The detection capability decreases due to the reduction of signals and changes for
the limit of detection (LOD) and signal-to-noise (S/N);
The accuracy (precision and trueness) depends on sample components;
The linearity and limit of quantification (LOQ) depends on the interaction between
biological samples and unknown compounds [33].
Severe ion suppression can lead to a non-disclosure of compounds. This effect is
occurred during the analysis of samples, which do not show suitable criteria for
detection and provide false negative results. Conversely, the overestimation of
compounds can provide false positive results. This effect depended on the ionic
suppression of internal standard [31].
The lack of database access due to modifications of mass spectra parameters. The
analysis of unknown samples during the mass spectrometry analysis can depend
on their correlation with library databank. The lack of database access is provided
mismatching of results and avoided the identification of right spectrum during the
analysis.
3.3. ME Evaluation
The evolution of biological samples was used to provide technical and useful
guidelines for bioanalytical scientists. The main approaches are:
The method of matrix and solvent fortifications represents a complex evolution of the
post extraction procedure, which compares the peak area of unknown compounds,
dispersed in the biological samples before and after extraction with organic solvent. This
procedure allows quantifying the recovery (RE) of analytical method as well as the ME
during the analysis [33]. In particular, samples are listed as A, B and C, where A represents
the standard before the extraction (solvent); B represents the empty biological samples
(white) added to unknown compounds after extraction; and C represents the white matrix
added to samples and then extracted.
The recovery, matrix effects, and process efficiency were evaluated and reported using
the equations below reported equations as in:
C
RE% 100
B
B
ME% 100
A
C RE
PE % 100 ME
A 100
The ME represents the matrix effect as absolute value. Matuszewski et al. [36] defined
the absolute ME as a comparison between the response signal of the standard added to
samples extracted from a single batch and the response of the standard in the cleaning
solution of mobile phase.
The lack of ME shows an absolute value of 100%, which demonstrates that samples
have the same value in the mobile phase and resulting analysis. An absolute ME value
higher than 100% showed that the ionic effect increased during the analysis and vice versa.
The presence and magnitudo of ME can be also evaluated by comparing different
standard guidelines obtained through various ionization techniques. The analysis is carried
out using the same compounds through HPLC-MS with similar standards and
chromatographic conditions [37].
The ME was performed by comparing the slope of standard calibration curves and
fortified samples. The slope and intercept are equal to the standard deviation of
experimental error in two batches for samples without unknown samples in the resulting
biological samples. Samples, containing unknown compounds in the biological samples
but without ME, show the same slope and intercept of the standard. A slope below the
reference calibration shows the lack of signal suppression; while an increased slope is
referred to signal enhancement [38].
Collection of samples;
Flow rate;
Particle size;
Temperature of the apparatus;
Column diameter and length;
Mobile phase and additive concentrations.
A similar effect can be also obtained by using high polar stationary phase and HILIC
or UPLC. An improvement of stationary phase has been recently obtained by using the
fused-core technology, which allows using short columns and high flow rates under
chromatographic conditions. The main advantages are decreasing of mass transfer [46],
distribution of particle in a narrow range, and the column packaging.
The dilution process can also decrease or remove the ME. The injection of less particles
increases the performance of analysis owing to decreasing in amount of components in the
sample. The dilution, as well as the injection of small volumes, also decreases the number
of molecules competitors to the droplet surface [47, 48].
The regulations and quality assurance play a main role in different fields of analysis,
from drug and metabolites to environmental applications. Recently, the increased use of
hyphenated HPLC-MS techniques requires international guidelines and protocols to
develop and validate bioanalytical processes.
The use of validated and approved bioanalytical methods provides suitable data to
process analysis in drug discovery and metabolisms. The validation of bioanalytical
method provides standard procedures for routine and/or laboratory analysis. A classical
approaches for validation of bioanalytical methods include consists of performing analysis
by using reference values according to instrumental setting-up and analytical apparatus. In
particular, the system configuration depends on multiple components.
The validation of bioanalytical method provides suitable parameters for the analysis of
samples under specific chromatographic conditions. Actually, three method validation
guidelines, e.g., US FDA, ANVISA and EMA, are used to perform and validate HPLC-
MS (and MS/MS) analysis of unknown samples in biological matrices as single or
combined agents. These validation procedures depend on the matrix effects if hyphenated
instrument configurations are used. The combination of these procedures allows to obtain
reproducible data and accurate analysis. The development of uniform and validate
international guidelines by combining the three major international procedures should be
performed to obtain a standard analysis without any potential mismatching of data.
Conflict of Interests
ACKNOWLEDGMENTS
This work was supported by publishing databank of the University “G. d’Annunzio”
of Chieti – Pescara, Chieti, Italy.
REFERENCES
Chapter 2
ABSTRACT
Rapid increase in the demand of food products due to the exponential growth of human
population along with the global distribution network of food and allied products have
posed enormous challenge in recent years to ensure food safety, quality, and authenticity.
Although analytical instrument have continued to enhance high resolution, fast analysis,
robustness, miniaturization, and portability; analysis of food using these state-of-the-art
instrument still requires extensive sample preparation. In fact, sample preparation is often
considered as the bottleneck and the most error prone step of the entire analytical workflow.
Realizing the significance of sample preparation and its critical role in obtaining quality
analytical data, substantial efforts have been made during last two decades to overhaul
sample preparation technologies. Current chapter discusses all aspects of food sample
preparation including solid phase extraction, liquid phase microextraction, solid-phase
microextraction, stir bar sorptive extraction, fabric phase sorptive extraction; highlights the
recent developments, discusses their advantages and shortcomings and projects the future
outlook.
*
Phone: 305-348-2396; E-mail: akabir@fiu.edu.
1. INTRODUCTION
Although analytical instrument used in food analysis has continued to enjoy a robust
growth and improvements in recent years, marked with miniaturization, enhanced
resolution, reduced operating costs, and shorter run time, sample preparation is still the
least evolved part of the analytical workflow and is often neglected [1]. The significance
of a simple and efficient sample preparation technique cannot be overstated when the
analytes of interest are dispersed at trace or ultra-trace level of concentrations in a very
diverse and complex matrices like food. Food represents a complicated sample matrix
comprised of peptides, lipids, carbohydrates, amino acids, fatty acids, organic acids,
nucleic acids, phytochemicals, colorants, aromas, preservatives and numerous other
exogenous compounds [2]. The compositional complexity and diversity of food samples
pose enormous challenges to analytical chemists who routinely deal with food sample for
analysis. Due to the inherent complexity, food samples are not generally suitable for direct
injection into the chromatographic/electrophoretic systems. Several factors contribute to
this problem: (a) ingredients of the complex food matrix may adversely affect the
performance of the analytical instrument; (b) co-elution of the unwanted matrix component
with the target analyte may convolute the analytical data; (c) concentration of the analyte(s)
of interest may be below the limit of detection/quantitation of the analytical instrument;
and (d) the analyte(s) of interest needs to be transferred into a suitable solvent compatible
with the analytical instrument prior to the analysis. The selection of an analytical method
for food analysis primarily depends on the objective of analysis. However, regardless of
the objective of the analysis, it is generally accepted that special attention has to be paid in
selecting an appropriate sample preparation technique.
The sample preparation step has a profound impact on the overall quality of the
analytical data. In fact, the success of the entire analytical process largely depends on the
selection and implementation of an appropriate sample preparation process. It is also very
important that the portion of the sample subjected to the analysis truly represents the bulk
food sample. The sample, immediately after receiving, should be properly stored under
appropriate conditions prior to the sample preparation and instrumental analysis.
Due to the enormous complexicity and inherent diversity of food samples, the sample
preparation stategy must be tailored in a way that matches well with the analytical
instrument and capable of delivering the required degree of accuracy and precision [3]. A
balanced sample preparation strategy should consider many factors including the nature of
the sample, objective of the analysis, precision and accuracy of the analysis to be achieved,
resources that can be used, skill of the person involved in the process and many others. A
generalized scheme of the overall analytical process that undergoes through a sample,
leading to qualitative/quantitative analytical report is presented in Figure 1. As can be seen
from the scheme, samle preparation plays a very important role in the overall analytical
process and if appropriate measures are not taken to minimize the error in this process, the
quality of the analytical data may be severely compromised.
Sampling
Sample Pre-treatment
(homogenization/filtration/
centrifugation/protein precipitation)
Extraction
Fractionation/cleanup
Sample Post-treatment
(solvent evaporation, sample
reconstitution/derivatization
Instrumental Analysis
Data handling
Reporting
Figure 1. Schematic representation of food analysis process highlighting the role of sample preparation.
Sample pre-treatment is the process which prepares the sample for analyte
extraction/fractionation/cleanup and often include homogenization, milling, grinding,
filtration, centrifugation etc. Sometime, this step is integrated with the sample preparation
technique e.g., Quick, easy, cheap, effective, rugged and safe (QuEcheRS) [4].
Homogenization is usually achieved by chopping or grinding. Usually the sample are
blended and mixed afterwards. If, due to the physical state of the sample, room temperature
grinding is nor practical, cryogrinding can be used as an alternative to ambient grinding
[5].
(50-200 mL for a 10 g sample), vulnerable to thermally labile analytes [3, 10]. Due to the
use of large volume of organic solvent, a solvent evaporation and sample reconstitution
step is inevitably needed after Soxlet extraction.
In order to reduce the long extraction time used in Soxhlet extraction, a new
modification named microwave-integrated Soxhlet extraction (MIS) has been introduced
recenly [10]. By combining powerful microwave irradiation to classical Soxhlet extraction,
the new technique has demonstrated equivalent extraction performance both qualitatively
and quantitatively in extracting oil from olives in 32 min compared to values obtained by
conventional Soxhlet extraction in 8 h.
In an attempt to eliminate or reduce the usage of toxic and hazardous organic solvent(s)
in sample preparation, researchers have been continuously searching for environmental
friendly green alternatives without compromising the quality of the analytical data.
Pressurized hot water extraction (PHWE) represents one of these emerging green sample
preparation techniques [11-13]. Due to its benignness, simplicity, relative inexpensiveness
and other attractive features, PHWE enjoys rapid increase in applications in food and
botanical samples. A number of easily controllable parameter such as temperature,
pressure, extraction time, flow rates and addition of modifiers/additives can be exploited
to increase the extraction efficiency of PHWE. Although the high polarity of water
(dielectric constant, 80 at 25°C) works against its use in extracting medium polar and low
polar analytes from food samples, the polarity value of water can be substantially reduced
to 27 when applied pressure at 50 bar and temperature at 250°C (comparable to dielectric
constants of 33 and 24 for methanol and ethanol, respectively). At these conditions, the
viscosity and surface tension of water decrese drastically, allowing easy extraction of the
target analytes even from difficult to reach core of solid food and plant samples. The high
extraction efficiency of PHWE compared to ambient water is attributed to (a) the
improvenent in the solubility and mass transfer effects due to the reduced dielectric
constant, increased diffusivity and reduced viscosity; (b) increased disruption of surface
equilibria [14]. In addition to high temperatute and pressure applied to water in PHWE, the
selectivity of water can be futher enhanced by adding a number of modifiers/additives such
as ammonia (0.01% w/v) [15], ethanol (5%) [16].
A large number of recent applications targeting a wide variety analytes such as
sulfonide antibiotic residue in meat [17], banned azo-dyes in chilli and hot chilli [18],
curcumin from turmeric rhizome unequivocally attests the growing popularity and future
potential of this green sample preparation technique.
Figure 2. Schematic diagram of a laboratory-built pressurized hot water extraction unit (reprinted with
permission from Ref. [13]).
the core of the sample. As a consequence, MAE is able to extract the target analyte faster
with high extraction efficiency [21].
The success of MAE as a viable and green alternative to classical solvent extraction
techniques largely depends on the optimum values of a number of independent variables
that need to be studied during the method development and include solvent composition,
solvent volume, extraction time, extraction temperature, physicochemcal characterstics of
the sample matrix etc. Extraction solvent used in MAE should possess high dielectric
constant so that it can efficiently absorb mecrowave energy. Suitable solvents for MAE
include methanol, ethanol, water. Non-polar solvents possessing low dielectric constants
such as hexane, toluene are not suitable solvents for MAE when used independently.
However, their extraction selectivity as well as extraction efficiency can be reguated by
using mixures of solvents. Mixtures of hexane-acetone and toluene-water are widely used
in MAE [21, 22].
Figure 3 demonstrates an on-line assembly of MAE with a HPLC system. Ability to
connect directly to the analytical instrument substantitally simplifies the overall analytical
workflow and minimizes the potential errors in sample preparation as well as the cost
burden.
MAE offers a number of advantages including low energy consumption, high
extraction efficiency, low solvent consumption, ability to complete automation among
others. One major shortcoming of MAE is its lack of selectivity, resulting in a sample with
high volume of co-extracted matrix interferents that potentially complicate the downtream
instrumental analysis. The sample obtained from MAE may need further clean-up using
solid phase microextraction [23] or solid-phase extraction [24].
A number of recent applications of UAE attest the growing interests of this green
sample preparation techniqie in food analysis [32-34]. Z. Pineiro et al. [32] demonstrated
application of UAE for rapid extraction of stilbenes from grape canes. Due to the
ultrasound alongwith temperature at 75° and 60% ethanolic solution in water, the extracted
time was reduced to only 10 min.
C.A. Ledesma-Escober et al. [33] compared the effect of sample pretreatment and
extraction on the determination of flavonoids from lemon and observed that UAE was the
best method for isolating these compounds when compared with microwave-assisted
extraction (MAE), superheated liquid extraction (SHLE), shaking extraction (SE).
D.M. Chen [34] used UAE to extract 120 drugs in animal derived food samples
including muscles and liver of swine, bovine, sheep, and chicken as well as eggs and dairy
milk. The target drugs were extracted in acetonitrile water with the help of UAE followed
by a final sample cleanup with solid phase extraction.
(PCBs) from fish samples. SPLE has reduced solvent consumption from 200 mL to 90 mL
and the extraction clean-up time from 18 h to 20 min. A large number of commercially
available adsorbents with wide range of polarity can be efficiently utilized to obtain a clean,
matrix interferent free sample for instrumental analysis.
Figure 5. Steps involved in selective pressurized liquid extraction (SPLE) (reprinted with permission
from Ref. [35]).
to centrifugation so that the extracting phase forms a distict layer at the bottom of the
solution. An aliquote of the sedimented layer of the extracting phase containg enriched
analyte(s) is then withdrawn by a syringe needle and injected into the chromatographic
system. The dispersing solvents used in DLLME are essentially water soluble and include
acetone, actonitrile and methanol. The extracting solvent, on the other hand, should be
highly soluble in the dispersing solvent but poorly soluble in water. In addition, the density
of the extracting solvent should differ widely from that of water so that phase separation
occurs when the sample solution is subjected to centrigafugation.
Extraction efficiency of DLLME is measured in terms of enrichment factor (EF) which
is defined as the ratio of the concentration of analyte(s) in the extracting phase to the initial
concentration of the analyte(s) in the sample. The EF value depends on the selection of
both the extracting and the dispersing solvents and should be optimized alongwith their
volumes. The EF value can be further enhanced by the addition of salt and by adjusting the
pH value of the sample,specially when the target analyte(s) is polar in nature.
DLLME is primarily utilized in analyzing pesticides in food samples. Some recent
applications of DLLME include analysis of organophosphorous pesticides (OPPs) in
cucumber and watermelon [38], cholesterol in food sample [39] and biogenic amines in
beer [40].
A schematic representation of DLLME is shown in Figure 6. The settled phase or
sedimented phase, after the process, contains the extracted analyte(s) for chromatographic
analysis.
Figure 6. Dispersive liquid-liquid microextraction (reprinted with permission from Ref. [41]).
Figure 7. Single drop microextraction produre and its different modifications (reprinted with permission
from Ref. [43]).
Sample preparation techniques using solid sorbents can be classified into solid phase
extraction (SPE) and solid phase microextraction (SPME). SPE is an exhaustive sample
preparation technique that isolate the target analyte(s) quantitatively from the sample
matrix when the sample passes though a bed of solid adsorbent. On the contrary, SPME is
an equilibrium sample preparation technique with relies upon the partition of the target
analyte between the extracting phase and the sample matrix.
the target analyte(s) and fractionation of organic compounds from a wide variey of sample
matrices including food. As such, soild-phase extraction (SPE) is considered as the gold
standard among all sample preparation techniques available for food sample preparation
[44-47]. Cleanup of the matrix interferents, preconcentration and fractionation of the
analyte(s) of interest are the most time consuming step in the analytical work-flow and in
some account, consume ~60% of the total time spent in the entire analytical process [48].
Solid-phase extraction has been known since the early 1970s, and therefore considered
as a mature sample preparation technique. Figure 8 demonstrates different formats of solid
phase extraction. Regardless of the format used amomg SPE cartridges, disks, pipette-tips,
96-well pate, the aqueous solution containing the target analyte(s) passes through a bed of
solid particulate sorbent during which the target analyte(s) are retained by the solid sorbent
via different intermolecular interactions between the solid sorbent and the target analyte(s).
Solid-phase extraction enjoys the commercial availability of a plethora of solid sorbents
including C18, C8, polystyrene-divinylbenzene (PS-DVB), alumina, magnesium silicate,
graphatised carbon as well as numerous recently developed phases [49]. However, C18
SPE phase is by far the most commonly used SPE sorbent in food sample preparation. One
of the major problem when using C18 SPE sorbent is its poor affinity towards highly polar
analytes. PS-DVB, Oasis™ HLB or other polar SPE sorbents can be used instead [50].
Usually, the particle size of SPE sorbents range between 10 and 60 µm, however, in the
disk format, use of smaller particle size e.g., 5 µm is a common practice. The sorbents used
in SPE can be classified into nonpolar, polar, ionexchange and mixed mode. The selection
of the apropriate sorbent depends on the analyte of interest, nature of the food matrix and
the potential interferents present in the food matrix. For the simultaneous extraction of
analytes possesing different polarity can be accomplished by using two catridges with
different sorption mechnisms.
Althouh, SPE is already an established sample preparation technique, developing a
SPE method is still challeging and requires substantial knowledge in analytical chemistry
as well as in separation science. Systematic SPE method development is the key to success
of this sample preparation technique and factos such as sample pH adjustment, selection of
the appropriate sorbent chemistry, flow rate during the sample loading, elution solvent and
its flow rate should be properly optimized.
Solid-phase extraction is carried out in four steps: conditioning, sample loading,
selective washing and elution [46]. In order to avoid clogging of the SPE bed, additional
care should be taken to ensure that the aqueous sample is absolutely free of particulate
matter.
SPE has gone through appreciable technological advancement in recent decades,
enabling this technique suitable in both off-line and on-line mode. Figure 9 demonstrates
different steps involved in both the off-line and off-lie mode of operation.
Figure 8. Different formats of solid phase extraction, mass of sorbent used and their estimated solvent
consumption (reprinted with permission from Ref. [51]).
Figure 9. Comparison between off-line and on-line solid phase extraction (reprinted with permission
from Ref. [52]).
Figure 10. Schematic presentation of magnetic solid phase extraction (reprinted with permission from
Ref. [54]).
Figure 11. Microextraction by packed sorbent (MEPS) reprinted with permission from Ref. [56]).
MEPS utilizes only 2 mg of the classical SPE sorbents including reverse phases,
normal phases, ion-exchage sorbents as well as mixed mode chemistries. As such, a MEPS
method can be easily adopted from a corresponding SPE method [56].
During the extraction, the sample is drwan/ejected once or multiple times into and out
of the syringe barrel to allow interaction between the solid sorbent and the target analyte.
After the extraction, the solid adsorbent is washed with deionized water to remove matrix
interferents. The extracted analytes are then eluted using a solvent volume of 20-50 µL.
The analyte enriched solution can be introduced into the analytical instrument. Typical
MEPS method development incudes the selection of appropriate solvent for conditioning,
loading, and elution; sample flow rate;wasing solution; and elution solvent type and
volume.
Recent applications of MEPS include analysis od E-resveratrol in wine [57], volatile
and semivolatile compounds in wine [58], metatonin and other antioxidants in foodsuff
[59].
Figure 11 demonstrate a MEPS device and the location of the sorbent. Figure 12
demonstrates the steps involved in MEPS.
Figure 12. Schematic diagram of microextraction by packed syringe and its different poerational steps
(reprinted with permission from Ref. [60]).
6.1.4. Quick, Easy, Cheap, Effective, Rugged, and Safe (Quechers) Method
Quick, easy, cheap, effective, rugged, and safe (QuEChERS) method was introduced
by M. Anastassiades and co-workers as a quick, convenient and environment friendly
alternative to conventional LLE [61]. It simultaneously reduces steps in sample preparation
and organic solvent consumption. The technique was primarily developed for analyzing
residual pesticide in food matrices. In its original form, food samples are homogenized,
minimum amount of acetonitrile is added to the sample, mixing the solvent by hand-
shaking or vortexing, a small quantity of salt mixture (anhydrous magnesium sulfate and
sodium chloride, 2:1 w/w) is added to the mixture followed by vortex mixing and
centrifugation. Subsequently, the supernatant extract undergoes clean-up and water
removal exercises by passing through a bed consisting of primary-secondary amine (PSA)
and anhydrous magnesium sulfate. The clean-up process removes residual water and many
matrix interferents including organic acids, polar pigments, and sugar [21].
Due to the advantages over LLE such as high analyte recovery, high sample
throughput, minimized steps, and low solvent consumption, the application of QuEChERS
has been continued to rise. However, since QuEChERS does not evaporate the solvent after
extraction and matrix cleanup, the ultimate concentration of the target analyte in the final
solution is always low compared to other sample preparation techniques that employs
solvent evaporation and sample reconstitution as the integral part of the sample preparation
work-flow. As such, in some cases, QuEChERS is followed by solvent evaporation and
sample reconstitution in order to achieve desired limit of quantitation [21].
Figure 12. Steps involved in QuEcheRS process (reprinted with permission from Ref. [62]).
the polymerization, the template is removed by Soxhlet extraction or using other processes.
The removal of template results in 3-D nanocavity that is complementary to the size, shape
and functionality of the template.
Figure 13. Illustration of the synthesis of molecularly imprinted polymer using organic synthesis
approach (reprinted with permission from Ref. [64]).
Due to the high specicificity of the MIP, applications of these artificial receptors are
continuously rising. A recent review exhaustively described the application of molecularly
imprinted polymer in food samples [63].
vial; direct immersion SPME, where SPME fiber is immersed directly inside the aquous
sample matrix containing the analyte(s) of interest.
Extraction of analyte(s) in SPME is based on the partitioning of the analyte between
the extracting phase and the bulk sample matrix. Once the extraction equilibrium is
reached, the extracted analyte(s) are thermally desorbed by introducing the fiber into the
hot inlet of a gas chromatography injection port or may be redissolved in an organic solvent
followed by analysis in gas chromatography, liquid chromatography or capillary
electrophoresis [65]. The application of SPME in food samples have been reviewed
extensively in numerous articles [66-70].
The advantages of SPME as a green sample preparation technique can fully
materialized when all the parameters that affect the extraction sensityvity are properly
optimized. These parameters include extraction sorbent chemistry, extraction mode,
method of sample agitation, sample pH, addition of salt, sample temperature, extraction
time and desorption conditions. Among these parameters, the extraction sorbent chemistry
plays the most important role in determining the extraction efficiency. In general, polar
sorbent coated SPME fibers such as polyacrylate (PA), Carbowax-polyethylene glycol
(PEG) favor extracting polar analytes, whereas nonpolar sorbents such as
polydimethylsiloxane (PDMS), C18, Carboxen-polydimethylsiloxane (CAR/PDMS),
polydimethylsiloxane-divinylbenzene (PDMS/DVB), divinylbenzene-carboxen-
polydimethyl-siloxane (DVB/CAR/PDMS) are frequently used for nonpolar/medium polar
analyte(s).
Food is a very complex sample matrix consisting of proteins, fat, acids, bases, salts,
and numerous other endogenous and exogenous chemicals. PDMS coating is generally
favored in food sample analysis due to its smooth surface that prevents from adsorbing
matrix component when direct immersion SPME is used. This adsorption of the matrix
interferents are often irreversible, leading to permanent damage to the sorbent coating.
However, PDMS being a nonpoar polymer is less effective, specially when polar analytes
are targetted.
Figure 14 presents SPME fiber and some other microextraction devives which were
invented based on the same principle.
sorbent coating are immobilized on the outer surface of the glass tubing containing the
encapsulated magnet. Compared to SPME fiber, stir bar possesses 24-126 µL of the
polymeric sorbent which is approximately 50-250 times more than a SPME fiber. Due to
the substantially larger sorbent loading in stir bar, higher mass of analyte is expected to
accumulate on it during extraction, leading to improvent in limit of detection and
quantitation.
Figure 14. SPME and its different modifications (reprinted with permission from Ref. [1]).
a prolonged period of time, partition equilibrium between the extraction sorbent and the
sample matrix is not often reached which consequently leads to poor extraction efficiency.
SBSE can be carried out in direct immersion mode (DI-SBSE) as well as in headspace
extraction mode (HS-SBSE). Regardless of the extraction mode used, after the completion
of the extraction, extracted analytes accumulated onto the stir bar can be desorbed by
immersing the stir bar into a small volume of organic solvent (liquid desorption, LD) or
can be exposed to thermal desorption (TD) using a special device known as thermal
desorption unit (TDU).
SBSE method development involves optimization of a number of factors including
extraction time, temperature, sample pH adjustment, addition of an inert salt to facilitate
salting out effect, stirring rate, and sample volume. When solvent mediated desorption is
used, solvent type, volume and exposure time into the solvent should also be optimized.
For thermal desorption, temperature and time span of the heat-shock are of prime
importance and should be optimized in order to maximize the extraction sensitivity.
Considering the high potential of SBSE technique as a green sample preparation
approach, a number of new coating technology have been introduced recently. Analyte
mass transfer rate in the new coatings are lot faster than conventional PDMS coating and
the new sorbents also possess unique selectivity, high resistant to solvent and thermal
degradation [72-74].
Some recent applications of SBSE in food analysis include extraction of estrogens in
pork and chicken [75], benzimidazoles in milk and honey [76] and sulfonamides in pork
and chicken [77]. A detailed account on the application of SBSE in food analysis can be
found in a recent review article [78].
the extraction equilibrium, amount of the sorbent loading is directly related to extraction
efficiency. On the other hand, kinetics controls the rate of extraction and consequently the
time required to reach the extraction equilibrium. A shorter sample preparation time is
always favored by the practicing scientists as well as by the routine analytical labs. As a
result, there is a pressing demand for developing new microextraction techniques that can
simultaneously satisfy the required sensitivity as well as minimize sample preparation time.
Shortcomings of the sorbent-based sorptive microextraction technique originate from:
(1) the characteristics of the extraction sorbent (pristine polymer or composite material);
(2) coating technology used in immobilizing the sorbent on the substrate surface [85] and,
(3) the geometrical format of the extraction system [83] that defines the primary contact
surface area (PCSA) of the device where the extraction sorbent makes contact with the
sample matrix containing the analytes. Therefore, if a sample preparation technique is to
be highly sensitive as well as fast, the composition of the sorbent, the coating technology
and the primary contact surface area have to be improved.
Some of the shortcomings suffered by the microextraction techniques are due to the
sorbent coatings used in commercially available microextraction technologies such as
bleeding, washing away with organic solvent, long extraction equilibrium time, limited
selectivity, extraction reproducibility, and swelling of the sorbent originate primarily from
the process of immobilizing the organic polymer on the substrate surface. These coatings
are generally created by physical deposition, followed by free-radical cross-linking
reactions [86, 87]. The lack of chemical bonding between the polymer sorbent and the
substrate is believed to be the root cause of these coating-related problems. A number of
alternative coating techniques have also been proposed including physical deposition [88,
89] electrochemical deposition of conducting polymers [90, 91], gluing with adhesives [92]
and sol-gel column technology [85, 1]. Nonetheless, sol-gel column technology pioneered
by Malik and co-workers [85, 1] have been proven to be the most convenient and versatile
[1]. In addition to the convenience of the coating process, sol-gel technology opens up the
possibility of making multi-component materials that can be conveniently used to
customize the surface morphology, selectivity, and affinity of the sorbent. The sorbent
coating created by sol-gel technology is highly porous and is chemically bonded to the
substrate. As an outcome, such coatings demonstrate remarkable thermal, solvent, and
chemical stability. Due to its inherent porosity, a thin film of sol-gel coating can extend
equivalent or higher sensitivity than commercially available, thick SPME coatings. The
high porosity of the sol-gel coating also makes it possible to reach extraction equilibrium
in a fraction of the time that is required by commercial SPME fibers.
Although tremendous efforts have been made to increase the sensitivity of the
microextraction systems by merely focusing on sorbent loading, little work has been done
to increase the primary contact surface area (PCSA) of the extraction device. The increase
in PCSA of the extraction device not only allows higher sorbent loading without increasing
the coating thickness, but also considerably reduces the extraction equilibrium time.
TFME, SBSE, RDSE etc. were developed to increase the PCSA, but the use of
conventional sorbent immobilization/use of membrane did not offer much benefit in
boosting the sensitivity of these systems.
In addition to improving the coating process and enhancing the primary contact surface
area of the microextraction device, some other factors which may potentially raise the
quality of sample preparation to a new level need further consideration: (1) the ability to
preconcentrate target analytes directly from the unmodified samples without any clean-up
exercises; (2) resistance to harsh chemical environments (i.e., highly acidic and basic) so
that matrix pH can be adjusted to wider pH values; (3) the ability to use any organic solvent
to elute the extracted analytes so that the final solution can be injected into an analytical
instrument of choice; (4) equal effectiveness in field sampling and sample preparation to
eliminate the burden of sample collection, transportation, storage, and sample preparation
in the laboratory; (5) the ability to achieve a high preconcentration factor during the
extraction so that solvent evaporation and sample reconstitution are unnecessary (a
frequent practice in LLE and SPE); (6) the ability to reach extraction equilibrium quickly
so that field sampling and sample preparation do not become an inconvenient task; and (7)
the ability to analyze the same sample using both gas chromatography (GC), high
performance liquid chromatography (HPLC), and/or capillary electrophoresis (CE) to
obtain complementary information depending on the analytical need.
Fabric phase sorptive extraction (FPSE), developed by Kabir and Furton in 2014, is
one of the most recent member of the modern sorbent based sorptive microextraction
techniques [93] which include SPE, SPME, SBSE, MEPS and many others. All these
sample preparation techniques can be classified as either (1) exhaustive extraction
technique (such as SPE); or (2) equilibrium extraction technique (such as SPME).
Exhaustive extraction often requires the sample passing through a fixed bed of the sorbent.
In equilibrium extraction, analytes are diffused towards the sorbent, stabilized on a
substrate to establish the equilibrium between the sorbent and the bulk medium (sample
matrix). The analyte mass transfer towards the sorbent continues until a partition
equilibrium is reached. FPSE has very intuitively combined both the extraction mode
(exhaustive and equilibrium extraction) into a single technology platform. The FPSE media
can be directly inserted into sample container holding the sample for extracting the target
analyte. During this equilibrium extraction process, analyte(s) diffuses though the sample
and interact with the sorbent for successful interaction.
Although, FPSE is a new sample preparation technique, it has already demonstrated its
potential by extracting amphenicols [94], residual penicillin antibiotics [95] and
sulfonamides directly from whole, intact milk samples without any protein precipitation,
sample evaporation and sample reconstitution. Figure 15 illustrates different steps involve
in FPSE process. A comparison between FPSE and matrix solid phase dispersion (MSPD)
given in Figure 16.
Figure 15. Schematic representation of fabric phase sorptive extraction of sulfa drug residues from
whole milk (reprinted with permission from Ref. [96]).
Figure 16. Comparison between fabric phase sorptive extraction (FPSE) and matrix solid phase
dispersion (MSPD) (reprinted with permission from Ref. [94]).
CONCLUSION
Both solvent mediated extraction and solid sorbent mediated extraction techniques are
in rapid transition towards solid-phase microextraction and liquid phase microextraction
and their different format, modifications and implementations. As such, the
implementation of the principles of Green Analytical Chemistry (GAC) in sample
preparation is in progress. Recent microextraction techniques, both solvent-based and
sorbent-based, have substantially reduced the solvent consumption. In addition, many of
the new sample preparation techniques are fully automated or in the process of being
automated. It can be stipulated by the current trend that sample preparation will be totally
automated in near future. Many researchers have contributed significant effort in
developing high performance sorbent materials including sol-gel hybrid inorganic-organic
composite materials, graphene oxide, multi-walled carbon nanotube, chitosan, aptamer etc.
Unfortunately, most of the newly developed advanced materials are not commercially
available and thefore are not available to routine testing laboratories, the major consumer
of the sample preparation technologies. It is expected that these advanced material systems
will soon be available to the testing laboratories, so that their unique capability in food
analytical chemistry can be totally exploited to ensure food safety, security and
authenticity.
REFERENCES
[1] A. Kabir, K.G. Furton, A. Malik, TrAC Trends in Analytical Chemistry 45 (2013)
197.
[2] G.P. Danezis, A.S. Tsagkaris, F. Camin, V. Brusic, C.A. Georgiou, Trac-Trends in
Analytical Chemistry 85 (2016) 123.
[3] K. Ridgway, S.P.D. Lalljie, R.M. Smith, Journal of Chromatography A 1153 (2007)
36.
[4] S. Armenta, M. de la Guardia, Trac-Trends in Analytical Chemistry 80 (2016) 517.
[5] S.S. Nielsen, Food analysis, Springer, New York; Dordrecht, 2010.
[6] H.-G. Schmarr, H.B. Gross, T. Shibamoto, Journal of Agricultural and Food
Chemistry 44 (1996) 512.
[7] N. Mulinacci, M. Innocenti, G. La Marca, E. Mercalli, C. Giaccherini, A. Romani, S.
Erica, F.F. Vincieri, Journal of Agricultural and Food Chemistry 53 (2005) 8963.
[8] G. Singh, P. Marimuthu, C.S. de Heluani, C.A.N. Catalan, Journal of Agricultural
and Food Chemistry 54 (2006) 174.
[9] M. Suhaj, Journal of Food Composition and Analysis 19 (2006) 531.
[59] L. Mercolini, R. Mandrioli, M.A. Raggi, Journal of Pineal Research 53 (2012) 21.
[60] L. Yang, R. Said, M. Abdel-Rehim, Journal of Chromatography B-Analytical
Technologies in the Biomedical and Life Sciences 1043 (2017) 33.
[61] S.J. Lehotay, K. Mastovska, Abstracts of Papers of the American Chemical Society
225 (2003) U101.
[62] V. Di Stefano, G. Avellone, D. Bongiorno, V. Cunsolo, V. Muccilli, S. Sforza, A.
Dossena, L. Drahos, K. Vekey, Journal of Chromatography A 1259 (2012) 74.
[63] X.L. Song, S.F. Xu, L.X. Chen, Y.Q. Wei, H. Xiong, Journal of Applied Polymer
Science 131 (2014) 18.
[64] T. Kubo, K. Otsuka, TrAC Trends in Analytical Chemistry 81 (2016) 102.
[65] C.H. Xu, G.S. Chen, Z.H. Xiong, Y.X. Fan, X.C. Wang, Y. Liu, Trac-Trends in
Analytical Chemistry 80 (2016) 12.
[66] L.B. Abdulra'uf, W.A. Hammed, G.H. Tan, Critical Reviews in Analytical Chemistry
42 (2012) 152.
[67] H.H. Jelen, Journal of Chromatographic Science 44 (2006) 399.
[68] W. Wardencki, M. Michulec, J. Curylo, International Journal of Food Science and
Technology 39 (2004) 703.
[69] M.C. Boyce, E.E. Spickett, Food Australia 54 (2002) 350.
[70] J. Beltran, F.J. Lopez, F. Hernandez, Journal of Chromatography A 885 (2000) 389.
[71] E. Baltussen, P. Sandra, F. David, C. Cramers, Journal of Microcolumn Separations
11 (1999) 737.
[72] Z.G. Xu, Y.L. Hu, G.K. Li, Chinese Journal of Analytical Chemistry 39 (2011) 1766.
[73] X.J. Huang, D.X. Yuan, Critical Reviews in Analytical Chemistry 42 (2012) 38.
[74] H.H. Jelen, M. Majcher, M. Dziadas, Analytica Chimica Acta 738 (2012) 13.
[75] C. Hu, M. He, B. Chen, B. Hu, Journal of Agricultural and Food Chemistry 60 (2012)
10494.
[76] X. Huang, L. Chen, D. Yuan, X. Luo, Journal of Separation Science 34 (2011) 3418.
[77] X. Huang, L. Chen, D. Yuan, Analytical and Bioanalytical Chemistry 405 (2013)
6885.
[78] M. Kawaguchi, A. Takatsu, R. Ito, H. Nakazawa, TrAC Trends in Analytical
Chemistry 45 (2013) 280.
[79] J. Pawliszyn, S. Liu, Analytical Chemistry 59 (1987) 1475.
[80] S. Bigham, J. Medlar, A. Kabir, C. Shende, A. Alli, A. Malik, Analytical Chemistry
74 (2002) 752.
[81] R. Eisert, J. Pawliszyn, Analytical Chemistry 69 (1997) 3140.
[82] P. Richter, C. Leiva, C. Choque, A. Giordano, B. Sepulveda, Journal of
Chromatography A 1216 (2009) 8598.
[83] I. Bruheim, X.C. Liu, J. Pawliszyn, Analytical Chemistry 75 (2003) 1002.
[84] R. Lucena, Analytical and Bioanalytical Chemistry 403 (2012) 2213.
[85] S.L. Chong, D.X. Wang, J.D. Hayes, B.W. Wilhite, A. Malik, Analytical Chemistry
69 (1997) 3889.
[86] B.W. Wright, P.A. Peaden, M.L. Lee, T.J. Stark, Journal of Chromatography 248
(1982) 17.
[87] L.G. Blomberg, Journal of Microcolumn Separations 2 (1990) 62.
[88] C.M. Graham, Y.J. Meng, T. Ho, J.L. Anderson, Journal of Separation Science 34
(2011) 340.
[89] F.M. Musteata, M. Walles, J. Pawliszyn, Analytica Chimica Acta 537 (2005) 231.
[90] A. Mohammadi, Y. Yamini, N. Alizadeh, Journal of Chromatography A 1063 (2005)
1.
[91] H. Bagheri, A. Mir, E. Babanezhad, Analytica Chimica Acta 532 (2005) 89.
[92] Y. Liu, Y.F. Shen, M.L. Lee, Analytical Chemistry 69 (1997) 190.
[93] A. Kabir, K.G. Furton, United States Patents and Trademark Office, 2016.
[94] V. Samanidou, L.-D. Galanopoulos, A. Kabir, K.G. Furton, Analytica Chimica Acta
855 (2015) 41.
[95] V. Samanidou, K. Michaelidou, A. Kabir, K.G. Furton, Food Chemistry 224 (2017)
131.
[96] E. Karageorgou, N. Manousi, V. Samanidou, A. Kabir, K.G. Furton, Food Chemistry
196 (2016) 428.
Chapter 3
ABSTRACT
Sample pretreatment procedures are the most important steps prior to analysis.
Accuracy and reliability of results are directly affected by these steps. An ideal
pretreatment procedure should be simple, cost effective, environmentally friendly, and it
should be compatible with detection method. There are two fundamental goals of pre-
treatment procedures; separation of matrix components and pre-concentration of trace
analyte species.
Solid phase extraction (SPE) methods are mainly preferred owing to useful and
practical properties. These methods are mainly applied as column and batch type.
Application of column methods for routine analysis is mostly limited by sample volume,
process time, complex structure of sample, and costs. The most important challenge in
batch methods are final separation of solid phase and selectivity of solid phase for target
species. Use of magnetic solids in SPE methods provides important advantages to
overcome these complications. Their magnetic properties furnish an easy phase separation
using a magnet, after accumulation of target molecules on particles. Moreover, the required
selectivity can be obtained by covering the core with functional groups specific for analyte
species. These are ideal properties expected from an effective pre-concentration and
separation method.
E-mail: hiulusoy@yahoo.com.
Magnetic nanoparticles have been used to pre-concentrate trace analytes and also to
separate them from matrix media. Analyte species are concentrated during this process,
and they can be determined by most of the conventional techniques even if they are at trace
level in the original samples. Chemical synthesis of magnetite particles is well known since
a long time: aqueous mixtures of Fe(II) and Fe(III) salts are mixed at room temperature in
alkali medium. Then, the immediate precipitation of a black magnetic product constituted
of rocklike magnetite particles, which are negatively charged and associated to the counter-
ions of the alkali and organic and inorganic species, was obtained. They can be modified
with silica supports, carbon nanotubes, alumina, and organic polymers in accordance with
analyte species. Magnetic nanoparticles are used for a lot of kind of sample type:
biological, environmental, plants, etc. Synthesis, characterizations, and some applications
of magnetic particles as effective material will be discussed throughout this chapter. The
future perspectives of magnetic solid applications will be also submitted with a few
proposals.
LIST OF ABBREVIATIONS
1. INTRODUCTION
In last two decades, the use of nanoparticles has daily increased owing to their unique
physical and chemical properties. Nanoparticles (NPs) with various shapes such as
spherical, nanotubes, nano horns, nano cages can be produced by using different starting
materials. Many kinds of nano-materials have been produced and used by most of the
researchers in different areas. Nanomaterials have begun to revolutionize the world around
us; they have been used in various scientific fields such as biotechnology, engineering,
biomedical, environmental, catalysis, drug delivery and material science [1].
The most used raw materials for this purpose are carbon, gold, organic dendrimers,
liposomes, semiconductors, and various metal oxides. These materials have already been
used in medicine, physics, chemistry, electronics, and pharmacy. The physical and
chemical properties of particles are fairly different at nanoscale. The main reason of this
effect is known as quantum size effect [2-4].
Today, the most used nanomaterials for analytical purposes are magnetic nanoparticles
(M-NPs). They have useful properties for scientists from various areas such as controllable
size, penetrable structures by an external effect, versatile chemical derivatization. The most
important advantage of magnetic nanoparticles is their ability to bind an analyte at
molecular size by keeping superparamagnetic properties.
The integration of magnetic nanoparticles with analytical methods provides important
contributes to separation, purification, pre-concentration, and trace quantitative analysis.
Participation of M-NPs to new developed and conventional analytical procedures ensured
great facilities in experimental process and improvements on analytical signals for trace
species in the complex mediums.
Magnetic solids are among the most exciting research topics for analytical chemists
who study on separation experiments. Indeed, they can be easily separated by an external
magnetic force from any matrix medium. Furthermore, once this external magnetic force
is removed, the particles do not form agglomeration because they have not magnetism.
Throughout this section, it will be discoursed about synthesis, characterization, and
applications of magnetic nanoparticles.
There are a lot of types of magnetic nanoparticles used in various scientific areas. The
most used ones: Fe3O4 (magnetite), α-Fe2O3 (hematite, weakly ferromagnetic or
antiferromagnetic), γ-Fe2O3 (maghemite, ferromagnetic), FeO (wüstite,
antiferromagnetic), ε-Fe2O3, β-Fe2O3.
Figure 1. Crystal structure and crystallographic data of the hematite, magnetite and maghemite
(the black circle is Fe2+, the green circle is Fe3+ and the red circle is O2−) [5].
Among these particles, magnetite and maghemite are the most popular ones owing to
biocompatibility properties, biomarker properties, and easy operation procedures. In
addition, some nanoparticle types (CoPt, FeCo, FePt, Ni) are used in industrial area owing
to both magnetic and catalytic properties. Fe3O4, MnFe2O4, and CoFe2O4 are preferred
especially in separation and purification science because their synthesis and application at
microscale are easy. Iron oxides and magnetic Nano beads are used in the preparation of
electrochemical biosensors owing to their wide surface area and biocompatible properties
[6]. The main application field for this type particles are pre-concentration, purification,
separation, and identification of some molecules.
When diameters of magnetic nanoparticles are smaller than 30 nm, they are called
super paramagnetic, which means that they have not magnetic memory. In this way, when
the external magnetic force is removed, the magnetic properties disappeared and
agglomeration is not observed [7, 8].
3. SYNTHESIS OF NANOPARTICLES
There are various synthesis methods for iron oxide based nanoparticles. These methods
are: thermal decomposition, microemulsion, co-precipitation, sol–gel synthesis, hydro-
thermal and solvo-thermal syntheses, ultrasound irradiation and biological synthesis. The
main classification of these methods is carried out by considering reaction medium as
aqueous or non-aqueous.
Many types of iron oxides are mainly preferred as nanoparticles, which are widespread
in nature and can be readily synthesized in the laboratory. Magnetic iron oxides have served
humans for centuries, for example regarding the application of small iron oxide
nanoparticles (IONPs) [5].
The magnetic solid phases used in extraction procedures are generally composed of
iron minerals and magnetic iron oxides such as magnetite (Fe3O4) and maghemite
(γ-Fe2O3).
The most used synthesis procedures for magnetic nanoparticles are based on three type
chemical processes:
Co-precipitation;
High temperature thermal decomposition with/without reduction;
Templated synthesis in the interior of micelles.
The preparation of a magnetic particles for analytical purposes usually involves three
main steps, including obtaining the magnetic particle (magnetite or maghemite) by a
synthesis procedure as in below, the coating of the magnetic core by suitable functional
material and the modification of the resultant core–shell structure by considering
application area [9].
In the most known conventional method, the co-precipitation method consists of
mixing ferric and ferrous ions in a 1:2 molar ratio in very basic solutions at room
temperature or at elevated temperature [5].
The reaction mechanism can be simplified as:
2.25 g of FeCl2.4H2O and 8.48 g FeCl3.6H2O were dissolved in 300 mL of pure water
and vigorously stirred at 800 rpm in the presence of inert medium (N2). Then, the solution
was heated up to 80℃, and 20 mL of ammonia was quickly added to solution. The product
was stirred for further 10 min, and the magnetic particles were removed by an external
magnet. The product was washed by water for 4-5 times in order to remove unreacted
reagents and dried at 50°C in vacuum oven for 3 hours. In these applications, Fe(II) and
Fe(III) ions in the solution are co-precipitated in the basic medium by adding ammonia.
Maghemite is formed in the first step and then magnetite is formed by oxidation. The
formed nanoparticles are covered by using suitable material according to usage area.
In another method used by Liand et al. coating of magnetic particles was carried out
by Stöber process. In this approach, spherical nanoparticles can be obtained at the same
size by a controlled process. The basis of this method is related with the production of silica
particles by hydrolysis of silane precursors in the presence of ammonia with ethanol [11].
In alternative, surface of the obtained maghemite nanoparticles by co-precipitation method
is coated by silica and then modified by amino silane groups. Thus, the obtained material
was successively used for immobilization and separation of some proteins [3, 12]. The
above mentioned co-precipitation methods involve fast particle formation rates, and
therefore the controlling of particle size and size distribution are difficult in many
processes. Various alternative strategies can be used in order to avoid the limitations of this
method such as non-aqueous thermal decomposition strategies.
A few of selected applications of magnetic nanoparticles are listed in Table 1 with their
name, composition, functional structure, and application.
The characterization process finds out important data about adsorptive behavior,
morphological properties, and possible impurities of new synthesis materials. Some
modifications during synthesis can improve some properties of magnetic particles such as
particle size, uniformity, specific surface area, functionality, pore size, magnetism power.
The first step of characterization is generally verifying the existence of iron oxides in
solids such as magnetite, by using X-ray diffraction method (XRD). Six characteristic
peaks belonging to Fe3O4 are observed at XRD spectrum. (2Ɵ = 30.1˚, 35.5˚, 43.1˚, 53.4˚,
57˚, and 62.6˚). Thus, the existence of Fe3O4 in the obtained new material is verified by
this method.
The existence and amount of Fe3O4 can be also demonstrated by Inductive Coupled
Plasma Mass Spectroscopy (ICP) or Atomic Absorption Spectroscopic (AAS) methods
[18-20].
Another important step of characterization is Transmission Electron Microscopy
(TEM). It is one of main techniques used for unrevealing of hetero-structured material.
This technique is based on detecting electron density in the scanned zone [21]. TEM
provides opportunities to observe surface morphology of nanoparticles. Similarly,
scanning electron microscopy (SEM) is a powerful technique used for the determination of
characteristic diversities on surface morphology of nanoparticles [22].
Another characterization technique for magnetic particles is the measurement of
magnetism power for the obtained new material. In this approach, magnetization shown
with M and magnetic field shown with H were determined and then calculated by using a
magnetization curve obtained from the experimental data. The magnetic properties and
magnetization levels of newly synthesized particles were determined by magnetic
susceptibility measurements. [23-24].
As an adsorptive material, the surfaces of magnetic particles are also related with their
adsorption capacities and need to be determined. This also provides important data for new
adsorbents. The most common technique for this measurement is Brunauer-Emmett-Teller
(BET) approach. In this method, the surface area is determined, especially for pores
structures, by measuring N2 adsorption [25].
In addition to these techniques, Fourier Transform Infrared Spectroscopy (FTIR) also
provides important information about special functional groups on the surface of particles.
FTIR is a useful and powerful technique to obtain detailed information about chemical
structure of particles. The data obtained with FTIR allow investigating the existence of
desired functional groups [26].
Predictably, the properties of magnetic particles are pretty based on synthesis
conditions. The main parameters affecting properties of final product are reaction
time, temperature, reactant concentration, type and amount of catalyst, and quality of
solvent, etc.
The most used coating material is silica-based derivatives [2]. Sol-gel procedure is the
most common approach used for the preparation of silica-based magnetic nanoparticles.
This procedure is based on condensation and hydrolysis of alkoxy silane. In this method,
tetraethyl orthosilicate (TEOS) is mostly added to a medium including magnetic particles,
and coating is achieved in the presence of acid or base catalysts [27]. The other common
coating materials are native silica, organic polymers, and metal oxides [28, 29].
Silica-based materials are known as an ideal support for newly synthesized magnetic
particles, owing to stable inert structure under even acidic conditions [30, 31]. So, silica-
functionalized magnetic materials have several advantages including stability under
aqueous conditions, easy surface modification, and easy control of inter-particle
interactions [32].
Silica-based materials are mostly preferred as coating material owing to chemically
inert structure, non-dispersive properties, easily structure to bind new groups as a new
shell. Schematic presentation of binding silane agent on NPs surface is illustrated in Figure
2.
Other main benefits of coating can be highlighted as in below.
Biocompatibility;
Shape recognition;
Fluorescent signaling;
Antigen detection;
Physic sorption.
Nanoparticles can form aggregates in order to decrease their energies due to large
surface area. This event blocks their activities.
The particles can lose their magnetic properties by easily oxidizing with
atmosphere components, because iron oxides have high chemical activities.
One of the issues to be considered in the coating process is the thickness of coating. In
fact, there are few studies about changing magnetic properties with thickness of coating
material [34].
The synthetic polymers obtained by using template molecules are useful materials
mostly used for selective separations. In this process, a template molecule is used in
addition to starting reagents (monomer, cross linker, stabilizer, etc.) in a conventional
polymerization reaction.
Molecular imprinted polymers are preferred in solid phase extraction applications
owing to important properties such as: highly selectivity; chemical resistance to acids,
bases, and organic solvents, mechanical stability, robustness at high temperature and
pressure, and low cost.
The first application of molecular imprinted polymers including magnetic
nanoparticles was presented by Ansell and Mosbach [35]. The other selected studies about
molecular imprinted polymers including magnetic particles are listed in Table 2.
Organic polymers, natural or synthetic, are also preferred as coating materials for
magnetic particles. These groups include cellulose, polystyrene, poly acrylic derivatives,
and the other trade polymers; they provide important advantages in order to recognize
various analyte species [42, 43]. The most used functional structures for this purpose are
amine, carboxylic acid, aldehyde, thiol, epoxy, hydroxyl, streptavidin, protein A and G,
albumin and the other antibodies. In this structure, the nanoparticle cores are encapsulated
in an inorganic or an organic coating that renders the whole particle stable and
biocompatible, and may serve as a support for biomolecules.
If the surface modifications of nanoparticles were made with an amino binder or linker
such as amino propyl triethoxysilane (APTS), active amino terminals (-NH2) in this
structure can be used for immobilization of a single strand DNA chain. So, bioactive DNA-
modified magnetic nanoparticles were obtained for biological applications. In this way,
selective determination of certain bioactive molecules can be carried out according to the
sequence structure of DNA chain linked to NPs surface.
A few applications can be seen in Table 4.
5. ANALYTICAL APPLICATIONS
Magnetic nanoparticle based extraction methods are mostly preferred among a lot of
separation procedures owing to useful properties and operation speed.
Among various separation and pre-concentration techniques, solid phase extraction
(SPE) has important advantages over other conventional sample pretreatment techniques
like liquid-liquid extraction, precipitation, co-precipitation, etc., in terms of simpler and
faster procedures, higher enrichment factors with better recoveries, quicker phase
separation, lower cost and reduced consumption of organic solvents as well [57,58].
Magnetic solid phase extraction (MSPE) by using magnetic particles as a useful
adsorbent has aroused great interest by various scientist in recent years. This approach was
firstly reported by Robinson et al. [59] in 1973 for biotechnological purposes. In 1987,
Wikstrom et al. [60] reported faster phase separations compared with conventional
methods such as liquid–liquid extraction and classical solid phase extraction by using
magnetically susceptible additives (ferrofluids or iron oxide particles) and an external
magnetic field.
Magnetic nanoparticles at nanoscale are important sources for labelling studies in bio-
sensing, due to their strong magnetic properties which are not found in biological systems.
Modification of the particle composition according to application area, suitable size and
magnetic properties permit their use in a variety of instruments and formats for bio-sensing
[7, 66, 72].
Fe3O4-based particles are most commonly used in developing biosensors due to their
superparamagnetic properties, biocompatibilities with antibodies and enzymes and ease of
preparation. However, Fe3O4 magnetic dipolar attraction and its large ratio of surface area
to volume may lead to aggregation in clusters when exposed to biological solutions.
Functionalization can overcome this problem and also enhance biocompatibility [53].
Some examples of magnetic particles used as biosensor are shown in Table 7.
Conflict of Interests
REFERENCES
[5] Wu, W., Wu, Z., Yu, T., Jiang, C; Kim, W. Recent progress on magnetic iron oxide
nanoparticles: synthesis, surface functional strategies and biomedical applications.
Sci. Technol. Adv. Mater., 2015, 16, 1-43.
[6] Koh, I., Josephson, L. Magnetic Nanoparticle Sensors. Sensors, 2009, 9(10), 8130-
8145.
[7] Beveridge, J. S; Stephens, J. R; Williams, M. E. The Use of Magnetic Nanoparticles
in Analytical Chemistry. Annu. Rev. Anal. Chem., 2011, 4, 251–73.
[8] Aguilar-Arteagaa, K., Rodrigueza, J. A., Barradob, E. Magnetic solids in analytical
chemistry: A review. Anal. Chim. Acta, 2010, 674, 157–165.
[9] Giakisikli, G., Anthemidis, A. N. Magnetic materials as sorbents for metal/metalloid
preconcentration and/or separation. A review. Anal. Chim. Acta, 2013, 789, 1– 16.
[10] Safari, M., Yamini, Y; Tahmasebi, E., Ebrahimpour, B. Magnetic nanoparticle
assisted supramolecular solvent extraction of triazine herbicides prior to their
determination by HPLC with UV detection. Microchim. Acta, 2016, 183(1), 203-210.
[11] Alejandro-Arellano, M., Blanco, A., Ung, T., Mulvaney, P., Liz-Marzán, L. M. Silica
coated metals and semiconductors. Pure Appl. Chem., 2000, 72, 257-265.
[12] Xu, H., Tong, N., Cui, L., Lu, Y., Gu, H. Preparation of hydrophilic magnetic
nanospheres with high saturation magnetization. J. Magn. Magn. Mater., 2007, 311,
125-132.
[13] Mosier-Boss, P. A., Lieberman, A. H. Surface-Enhanced Raman Spectroscopy
Substrate Composed of Chemically Modified Gold Colloid Particles Immobilized on
Magnetic Microparticles. Anal. Chem., 2005, 77, 1031–1037.
[14] Safarikova, M., Kibrikova, I., Ptackova, L., Hubka, T., Komarek, K., Safarik, I.
Magnetic solid phase extraction of non-ionic surfactants from water. J. Magn. Magn.
Mater., 2005, 293, 377-381.
[15] Son, A., Dosev, D., Nichkova, M., Ma, Z., Kennedy, I. M., Scow, K. M., Hristova,
K. R. Quantitative DNA hybridization in solution using magnetic/luminescent core-
shell nanoparticles. Anal. Biochem., 2007, 370, 186–194.
[16] Chiang, C. L., Sung, C. S., Wu, T. F., Chena, C. Y., Hsu, C. Y. Application of
superparamagnetic nanoparticles in purification of plasmid DNA from bacterial cells.
J. Chromatogr. B, 2005, 822, 54–60.
[17] Zhao, G., Song, S., Wang, C., Wu, Q., Wang, Z. Determination of triazine herbicides
in environmental water samples by high-performance liquid chromatography using
graphene-coated magnetic nanoparticles as adsorbent. Anal. Chim. Acta, 2011,
708(1-2), 155-159.
[18] Aguilar-Arteaga, K., Rodriguez, J. A., Miranda, J. M; Medina, J., Barrado, E.
Determination of non-steroidal anti-inflammatory drugs in wastewaters by magnetic
matrix solid phase dispersion–HPLC. Talanta, 2010, 80, 1152–1157.
[19] Dong, J., Xu, Z., Wang, F. Engineering and characterization of mesoporous silica-
coated magnetic particles for mercury removal from industrial effluents. Appl. Surf.
Sci., 2008, 254, 3522–3530.
[20] Latham, A. H., Wilson, M. J., Schiffer, P., Williams, M. E. TEM-induced structural
evolution in amorphous Fe oxide nanoparticles. J. Am. Chem. Soc., 2006, 128,
12632–12633.
[21] Zou, Z., Ibisate, M., Zhou, Y., Aebersold, R., Xia, Y., Zhang, H. Synthesis and
evaluation of superparamagnetic silica particles for extraction of glycopeptides in the
microtiter plate format. Anal. Chem., 2008, 80, 1228–1234.
[22] Brucea, I. J., Taylor, J., Todd, M., Davies, M. J., Borioni, E., Sangregorio, C., Sen,
T. Synthesis, characterisation and application of silica-magnetite nanocomposites. J.
Magn. Magn. Mater., 2004, 284, 145–160.
[23] Chiang, C. L., Chen, C. Y., Chang, L. W. Purification of recombinant enhanced green
fluorescent protein expressed in Escherichia coli with new immobilized metal ion
affinity magnetic absorbents. J. Chromatogr. B, 2008, 864, 116–122.
[24] Wang, Q., Kanel, S. R., Park, H., Ryu, A., Choi H. Controllable synthesis,
characterization, and magnetic properties of nanoscale zerovalent iron with specific
high Brunauer–Emmett–Teller surface area. J. Nanopart. Res., 2009, 11(3), 749-755.
[25] Galhouma, A. A., Mahfouz, M. G., Goma, N. M., Vincent, T., Guibal, E. Chemical
modifications of chitosan nano-based magnetic particles for enhanced uranyl
sorption. Hydrometallurgy, 2017, 168, 127–134.
[26] Wu, W., He, Q., Jiang, C. Magnetic Iron Oxide Nanoparticles: Synthesis and Surface
Functionalization Strategies. Nanoscale Res. Lett., 2008, 3, 397-415.
[27] Wei, J., Zhang, Y., Wang, Y., Tan, F., Liu, J., Cai, Y., Qian, X. Highly efficient
enrichment of phosphopeptides by magnetic nanoparticles coated with zirconium
phosphonate for phosphoproteome analysis. Rapid Commun. Mass Spectrom., 2008,
22, 1069–1080.
[28] Liu, X., Guan, Y., Shen, R., Liu, H. Immobilization of lipase onto micron-size
magnetic beads. J. Chromatogr. B, 2005, 822, 91–97.
[29] Erdem, F., Sayar, H., Karadeniz, Guven, G., Ozsoz, M; Piskin, E. Development of
Streptavidin Carrying Magnetic Nanoparticles and Their Applications in
Electrochemical Nucleic Acid Sensor Systems. Electroanalysis, 2007, 19, 798–804.
[30] Chen, L., Wang, T., Tong, J. Application of derivatized magnetic materials to the
separation and the preconcentration of pollutants in water samples, TrAC,, 2011,
30(7), 1095-1108.
[31] Strelko, V., Malik, D. J., Streat, M. Interpretation of Transition Metal Sorption
Behavior by Oxidized Active Carbons and Other Adsorbents. Sep. Sci. Technol.,
2004, 39, 1885-1905.
[32] Lu, A. H., Salabas, E. L., Schuth, F. Magnetic Nanoparticles: Synthesis, Protection,
Functionalization, and Application. Angew. Chem., Int. Ed. Engl., 2007, 46, 1222-
1244.
[33] Ünak, P. Imaging and Therapy with Radionuclide Labeled Magnetic Nanoparticles,
31 Braz. arch. biol. technol., 2008, 51, 31-37.
[34] Tian, H., Li, J., Shen, Q., Wang, H., Hao, Z., Zou, L., Hu, Q. Using shell-tunable
mesoporous Fe3O4@HMS and magnetic separation to remove DDT from aqueous
media. J. Hazard Mater., 2009, 171, 459-464.
[35] Ansell, R. J., Mosbach, K. Magnetic molecularly imprinted polymer beads for drug
radioligand binding assay. Analyst, 1998, 123, 1611-1616.
[36] Wua, X., Huanga, M., Zhoua, T., Mao, J. Recognizing removal of norfloxacin by
novel magnetic molecular imprinted chitosan/γ-Fe2O3 composites: Selective
adsorption mechanisms, practical application and regeneration. Sep. Purif. Technol.,
2016, 165 (13), 92–100.
[37] Riveros D., Cordova, K., Michiels, C., Verachtert, H., Derdelinck, G. Polydopamine
imprinted magnetic nanoparticles as a method to purify and detect class II
hydrophobins from heterogeneous mixtures. Talanta, 2016, 160 (1), 761–767.
[38] Li, H., Xie, T., Ye, L. Core-shell magnetic molecularly imprinted polymer
nanoparticles for the extraction of triazophos residues from vegetables. Microchim.
Acta, 2017, doi:10.1007/s00604-017-2096-4.
[39] Anirudhan, T. S., Christa, J., Deepa, J. R. Extraction of melamine from milk using a
magnetic molecularly imprinted polymer. Food Chem., 2017, 227, 85–92.
[40] Sheykhaghaei, G., Hossainisadra, M., Khanahmadzadeh, S., Seyedsajadia, M.,
Alipouramjad, A. Magnetic molecularly imprinted polymer nanoparticles for
selective solid phase extraction and pre-concentration of Tizanidine in human urine.
J. Chromatogr. B, 2016, 1011, 1–5.
[41] Penga, H., Luo, M., Xiong, H., Yub, N., Ning, F., Fana, J., Zeng, Z., Lic, J., Chen, L.
Preparation of photonic-magnetic responsive molecularly imprinted microspheres
and their application to fast and selective extraction of 17β-estradiol. J. Chromatogr.
A, 2016, 1442, 1–11.
[42] Uemur, U., Ohb, M., Kitagaw, S. Size and Surface Effects of Prussian Blue
Nanoparticles Protected by Organic Polymers. Inorg. Chem., 2004, 43 (23), 7339–
7345.
[43] Stuart, C., Humphrey, H. P., Dobson, Y. J. Magnetic nanoparticles for gene and drug
delivery. Int. J. Nanomedicine, 2008, 3(2), 169–180.
[44] Liu, Z., Liu, Y., Yang, H., Yang, Y., Shen, G., Yu, R. A phenol biosensor based on
immobilizing tyrosinase to modified core-shell magnetic nanoparticles supported at
a carbon paste electrode. Anal. Chim. Acta, 2005, 533, 3-9.
[45] Sanchez, C., Di, D. H., Connor, V. C., Cler, P. C., Meunier E., Pipy, Payre, B., Tan,
R. P., Gougeon, M., Carrey, J., Gigoux, V., Fourmy, D. Targeting a G-Protein-
[58] Das, D., Gupta, U., Das, A. K. Recent developments in solid phase extraction in
elemental speciation of environmental samples with special reference to aqueous
solutions. TrAC, 2012, 38, 163–171.
[59] Robinson, P. J., Dunnil, P., Lilly, M. D. The properties of magnetic supports in
relation to immobilized enzyme reactors. Biotechnol. Bioeng., 1973, 15, 603–606.
[60] Wikstrom, P., Flygare, S., Grondalen, A., Larsson, P. O. Magnetic aqueous two-
phase separation: A new technique to increase rate of phase-separation, using
dextran-ferrofluid or larger iron oxide particles. Anal. Biochem., 1987, 167, 331–339.
[61] P. Ashtari, K. Wang, X. Yang, S. Huang, Y. Yamini, Novel separation and
preconcentration of trace amounts of copper (II) in water samples based on
neocuproine modified magnetic microparticles. Anal. Chim. Acta, 2005, 550, 18–23.
[62] Lee, P. L., Sun, Y. C., Ling, Y. C. Magnetic nano-adsorbent integrated with lab-on-
valve system for trace analysis of multiple heavy metals. J. Anal. At. Spectrom., 2009,
24, 320–327.
[63] Chen, B., Hu, B., He, M., Huang, Q., Zhang, Y., Zhang, X. Speciation of selenium
in cells by HPLC-ICP-MS after (on-chip) magnetic solid phase extraction. J. Anal.
At. Spectrom., 2013, 28, 334–343.
[64] Chen, B., Heng, S., Peng, H., Hu, B., Yu, X., Zhang, Z., Pang, D., Yue, X., Zhu, Y.
Magnetic solid phase microextraction on a microchip combined with electrothermal
vaporization-inductively coupled plasma mass spectrometry for determination of Cd,
Hg and Pb in cells. J. Anal. At. Spectrom., 2010, 25, 1931–1938.
[65] Faraji, M., Yamini, Y., Rezaee, M. Extraction of trace amounts of mercury with
sodium dodecyle sulphate-coated magnetite nanoparticles and its determination by
flow injection inductively coupled plasma-optical emission spectrometry. Talanta,
2010, 81, 831–836.
[66] Locatelli, M., Sciascia, F., Cifelli, R., Malatesa, L., Bruni, P., Croce, F. Analytical
methods for the endocrine disruptor compounds determination in environmental
water samples. J. Chromatogr. A, 2016, 1434, 1-18.
[67] Suna, L., Chen, L., Sun, X., Du, X., Yu, Y., He, D., Xu, H., Zeng, Z., Wang, H.,
Ding, L. Analysis of sulfonamides in environmental water samples based on
magnetic mixed hemimicelles solid-phase extraction coupled with HPLC–UV
detection. Chemosphere, 2009, 77(10), 1306–1312.
[68] Heidari, H., Razmi, H. Multi-response optimization of magnetic solid phase
extraction based on carbon coated Fe3O4 nanoparticles using desirability function
approach for the determination of the organophosphorus pesticides in aquatic
samples by HPLC–UV. Talanta, 2012, 99(15), 13–21.
[69] Lo, C., Chen, W., Chen, C., Chen, Y. Rapid Enrichment of Phosphopeptides from
Tryptic Digests of Proteins Using Iron Oxide Nanocomposites of Magnetic Particles
Coated with Zirconia as the Concentrating Probes. J. Proteome Res., 2007, 6 (2),
887–893.
Chapter 4
ABSTRACT
Medicinal plants have been a primary source of bioactive compounds that are finally
being developed as commercial drugs. Nowadays, many phytochemicals are associated
with health benefits and they are used as antioxidants, antimicrobials, anticancer and anti-
inflammatory agents. Phenolic compounds are the main secondary metabolites and are
highly emphasized due to their well-documented bioactivities. These compounds may be
classified into different groups as a function of the number of phenol rings that they contain
and on the basis of structural elements that bind these rings to one another. Numerous
studies indicated a positive association between health-promotion and well-being and the
consumption of a polyphenolic rich diet. Generally, the health benefit effects of phenolic
compounds depend on their antioxidant effects. Various methods are used to investigate
the antioxidant properties of plant extracts or phenolic compounds. Especially concerning
radical scavenging (DPPH, ABTS, etc.) and reducing power (CUPRAC, FRAP, etc.),
Corresponding Author: gokhanzengin@selcuk.edu.tr.
phosphomolybdenum assays are amongst the most common antioxidant assays. This
chapter aims to critically review several aspects concerning the occurrence of phenolics in
different medicinal plant matrices and nutraceuticals, and their characterization by HPLC
coupled with different antioxidant assays. Moreover, a special aim of this work would be
to create a glimpse concerning modern approaches used to assess the antioxidant capacity
of phenolics by HPLC online coupling techniques.
1. INTRODUCTION
Several epidemiological studies suggest that high intakes of plants or plant products
(vegetables or fruits) reduce the risk of some types of chronic and degenerative diseases
such as cancer, cardiovascular diseases, diabetes mellitus, and Alzheimer’s disease. These
properties are attributed to a variety of phytochemicals including vitamins, minerals or
different plant secondary metabolites (Bernhoft, 2010). Secondary metabolites are not
essential for growth and they have a wide range of chemical structures and biological
activities. Phenolic compounds are one of the most known classes of plant secondary
metabolites. They posses a large structural variety and exhibit different biological
activities. For example, the antioxidant abilities of phenolic compounds is of then
considered a marker of several other biological activities (Carocho and Ferreira, 2013; Dai
and Mumper, 2010). Phenolic components have one or more hydroxyl groups in their
phenyl rings and they play as hydrogen donors. As a result of this, phenolic compounds
are good antioxidants being involved in scavenging several free radicals. Also, phenolic
compounds can chelate metal ions involved in the production of free radicals (Rice-Evans
et al., 1996). Flavonoids are one of the main classes of phenolic compounds and they
possess a wide spectrum of biological activities including anti-inflamatory, anti-cancer,
anti-microbial or analgesic (Corradini et al., 2011; Grassi et al., 2010; Nijveldt et al., 2001).
With these perspectives, the characterization of phenolic compounds and theirs biological
activites have been investigated in many studies (Carradori et al., 2014; Gidaro et al., 2015;
Gidaro et al., 2016; Zengin et al., 2016a; Zengin et al., 2016b). Consequently, nowadays
there is a high emphasis on phenolic compounds, their occurrence in different plants as
well as in assessing their biological activities.
This chapter reports on overwiev of phenolics, their characterization by antioxidant
assays, and their involvement in different health beneficial effects.
Several thousands of phenolic compounds have been identified in the plant kingdom,
and can be divided into different classes according to the type and number of phenolic rings
(D Archivio et al., 2007). The main classes of phenolic compounds are: phenolic acids,
flavonoids, stilbenes, coumarins, lignans and tannins.
Phenolic acids account for approximately one-third of the dietary phenolics. Phenolic
acids may be present in plants in free and bound forms (Robbins, 2003). They can be
divided in two subgroups, the hydroxybenzoic and hydroxycinnamic acids. The
hydroxybenzoic and hydroxycinnamic acids are derived from non-phenolic molecules of
benzoic and cinnamic acid, respectively (Figure 1).
The hydroxybenzoic acid derivatives are commonly present in the form of glucosides
in medicinal or edible plants. The most common hydroxybenzoic acids are gallic, p-
hydroxybenzoic, protocatechuic, vanilic and syringic acids. The hydroxycinnamic acids
mainly occur in medicinal or edible plants as simple esters with quinic acid or glucose.
Hydroxycinnamic acids include caffeic, ferulic, p-coumaric, and sinapic acids. Unlike
other phenolic compounds, hydroxybenzoic and hydroxycinnamic acids posses an acidic
character due to the presence of one carboxylic group (Annie and Jean-Jacques, 2009).
(Figure 2).
Flavonoids are a predominant class of phenolics, and they constitute about two-third
of the dietary phenolics. Flavonoids consist of fifteen carbons arranged in a C6-C3-C6
configuration. Flavonoids can be consindered derivatives of the 2-phenylchromone parent
compound having a structure which consists of three phenolic rings (A, B, and C) (Figure
3). Nonetheless, flavonoids can be classified as flavones, flavonols, isoflavones,
flavanones, flavanonols, flavanols, and anthocyanidins (Das, 1994; Yao et al., 2004).
Flavones and flavonols can naturally occure as aglycones in foods. About 200
flavonols and 100 varieties of flavones have been identified in plants. Flavones and
flavonols have a double bond between C2 and C3 (Table 1 and Figure 4). Flavonols are
different from flavones due to the absence of a hydroxyl group in the position 3 (Shahidi
and Naczk, 2004) (Table 2 and Figure 5). Flavonones and flavononols are characterized by
the presence of a saturated C2-C3 and carbonyl group in the position 4 (Table 3 and Figure
6). Flavononols are different from flavonones due to possessing a hydroxyl group in the
position 3 (Table 4 and Figure 7). Isoflavonoids consist of a phenyl ring (the A ring), a
heterocyclic C-ring and another phenyl ring (the B ring) at the position C-3 (Han et al.,
2009). Common isoflavones are genistein, daidzein and glycitein (Jaganath and Crozier,
2010). Anthocyanins and catechins are known as flavans because of lacking the carbonyl
group in the position 3. Flavans can be categorised as flavan-3-ols and flavan-3,4-diols
(Shahidi and Naczk, 2004) (Table 5 and Figure 8). Anthocyanins are pigments that are
colored in red, purple, or blue related to pH (Konczak and Zhang, 2004). Anthocyanidins
are the primary structure of the anthocyanins (without sugar moiety) (Table 6 and Figure
9). Pyranoanthocyanins structure arises from the cyclisation between C4 and the hydroxyl
group at C5 of the original flavylium moiety yielding a fourth ring (D). Pyranoanthocyanins
are more stable than anthocyanins (Schwarz et al., 2003).
Substitution
Name 5 6 7 8 3’ 4’
Baicalein OH OH OH H H H
Baicalin OH OH Glucuronide H H H
Wogonine Glucuronide OH H Glucuronide OMe H H
Chrysin OH H OH H H H
Apigenin OH H OH H H OH
Luteolin OH H OH H OH OH
Name Substitution
3 5 7 2’ 3’ 4’ 5’
Kaempferol OH OH OH H H H OH
Quercetin OH OH OH H OH OH H
Morin OH OH OH OH H OH H
Myricetin OH OH OH H OH OH OH
Hyperoside Galactoside OH OH H OH OH H
Quercetin-3- Galactouronide OH OH H OH OH H
galactouronide
Quercetin-3- glucoside Glucoside OH OH H OH OH H
Miquelianin Glucuronide OH OH H OH OH H
Rutoside -Rha-Glc OH OH H OH OH H
Trihydroxyethylrutoside -Rha-Glc OH 7;3’;4’=OC2H5OH
Tetrahydroxyethylrutoside -Rha-Glc 5;7;3’;4’= OC2H5OH
Substitution
Name 5 7 8 2’ 3’ 4’ 5’
Naringenin OH OH H H H OH H
Eriodictyol OH OH H H OH OH H
Lysinonotin OH OH OMe OMe H OMe H
Hesperetin OH OH H H OH OMe H
Dihydrotricin OH OH H H OMe OH OMe
Naringin OH -Rha-Glc H H H OH H
Hesperidin OH -Rha-Glc H H OH OMe H
Substitution
Name 3 5 7 3’ 4’ 5’
Fustin OH H OH OH OH H
Taxifolin OH OH OH OH OH H
Amelopsin OH OH OH OH OH OH
Substitution
Name 3 5
(+)-Catechin OH H
(+)-Gallocatechin OH OH
(+)-Gallocatechin-3-O-gallate Galloyl OH
(-)-Epicatechin OH H
(-)-Epicatechin gallate Galloyl H
(-)-Epigallocatechin OH OH
(-)-Epigallocatechin gallate Galloyl OH
Procyanidin B2 Dimer of (-)-epicatechin
Procyanidin C1 Trimer of (-)-epicatechin
Tetrameric proanthocyanidin Tetramer of (-)-epicatechin
Procyanidin B4 (+)-Catechin-(4α-8)-(-)-epicatechin
Substitution
Name 3 5 7 3’ 4’ 5’
Pelargonidin OH OH OH H OH H
Cyanidin OH OH OH OH OH H
Delphinidin OH OH OH OH OH OH
Tannins constitute the third most important group of phenolic compounds (Porter
et al., 1989). Tannins are classically divided into hydrolysable and condensed
(proanthocynidins), depending on their chemical structures (Figure 10). Proanthocyanidins
are polymeric flavonoids (particularly flavan-3-ols) and oligomers, whereas hydrolysable
tannins are yielding gallic acid when hydrolyzed (Ferreira and Slade, 2002; Khanbabaee
and van Ree, 2001). Phlorotannins are a third group of tannins. These compounds are a less
familiar group and they are not important for the human diet (Bravo, 1998; Porter et al.,
1989).
Phenolic compounds are an important part of the human diet (Table 7) and have been
largely studied due to their biological activities including antioxidant, antimutagenic,
anticancer, antiallergenic, antiinflammatory, antiviral, antiulcer, antidiarrheal,
anthelmintic, antihepatoxic, and antiproliferative (Carocho et al., 2014; Muchuweti et al.,
2007). The average daily intake of dietary phenolic compounds is about 1g per person,
however this values is largely debatable according to different studies (Scalbert and
Williamson, 2000). Epidemiological research reported that the consumption of phenolic
compounds is correlated with a reduced incidence of diseases like coronary heart disease,
age related eye degeneration, and cancer (Hertog and Hollman, 1996; Hertog et al., 1993;
Knekt et al., 1996; McCullough et al., 2012; Steinmetz and Potter, 1991). Generally,
phenolic compounds can protect from diseases related with reactive oxygen species (ROS)
(Armatu et al., 2010), and the protective effects of these compounds are ascribed to their
ability of direct scavenging of free radicals (Heim et al., 2002).
Different authors described that some flavonoids exhibited antilipoperoxidant,
antitumoural, antiplatelet, antiischaemic, antiallergic, and anti-inflammatory activities
(Cao et al., 1997; Deschner et al., 1991; Middleton and Kandaswami, 1992; Terao et al.,
1994; Tzeng et al., 1991). Additionally, hydroxycinnamic acid conjugates and flavonoids
possess a great number of antioxidant facets in vitro. Recent evidence reported that tea
polyphenols might protect against different stages of carcinogenesis(Khan and Mukhtar,
2009). For example, epigallocatechin-3-gallate (EGCG) found in green tea acts as a cancer
chemopreventive agent (lungs, liver, skin and prostate cancer, gastrointestinal tract),
antiobesity, and cardiovascular protective agent (Corradini et al., 2011; Khan and Mukhtar,
2009; Klaus et al., 2005; Luo et al., 2001). Catechins are reported to possess antibacterial,
anticancer, anticataract, antifungal, antihypercholesterolemic, antimutagenic, antiviral,
antiproliferative activities (Akagi et al., 1997; Chaudhuri et al., 1997; Davis et al., 1997;
Geetha et al., 2004; Hirasawa and Takada, 2004; Matsui et al., 2006; Mikaberidze and
Moniava, 1974; Riso et al., 2002). Miyake et al. (2005) reported that a high level of
consumption of soy and its isoflavones might be related to reducing the prevalence of
allergic rhinitis.
Resveratrol, a phenolic compound notably found in grapes, pistachio, peanuts and
berries possesses an important antioxidant capacity. Besides, this compound has protective
effects against several diseases such as cancer, type 2 diabetes, cardiovascular disease, and
neurological conditions (Marques et al., 2009). Resveratrol can also be used in minimizing
or preventing lipid oxidation in pharmaceutical products.
The authors have reported that caffeic acids and its esters may have protectivite activity
against colon carcinogenesis (Olthof et al., 2001; Robbins, 2003). Caffeic and ferulic acids
also showed in vitro antioxidant activities (Piazzon et al., 2012; Villaño et al., 2005) and
they have also antimicrobial effects aganist pathogenic bacteria and fungi (Alves et al.,
2013). Anthocyanins possess several pharmacological and biological properties including
anti-inflammatory and antioxidant (Kong et al., 2003).
Gallic acid has exhibited various bioactivities such as antineoplastic, bacteriostatic,
antimelanogenic and antioxidant properties (Kim, 2007) as well as antimalarial, antifungal,
and antiviral potential (Chaudhuri et al., 1997; Florov and Mishenkova, 1969; Mahadevan
and Koti Reddy, 1968; Thompson et al., 1953).
Many researches have reported that p-hydroxybenzoic acid possesses several
biological activities like antioxidant activity against free radicals, antimicrobial activity
against pathogenic bacteria and fungi, estrogenic, and antimutagenic activity (Heleno et
al., 2013; Pugazhendhi et al., 2005; Rice-Evans et al., 1996). Additionally, protocatehuic
acid has antioxidant, antimicrobial, cytotoxic, chemopreventive, neuroprotective, and
apoptotic properties (Alves et al., 2013; An et al., 2006; Ferreira et al., 2009; Yin et al.,
2009; Yip et al., 2006).
Tannins possess several biological activities because they are metal ion chelators,
protein precipitating agents, and antioxidants. Tannins can inhibit lipid peroxidation,
lipoxygenases in vitro, and scavenge radicals like hydroxyl and superoxide (Gyamfi and
Aniya, 2002). Lignans play important applications in cancer chemotherapy and various
other pharmacological activities (Saleem et al., 2005).
Reactive oxygen species (ROS) such as superoxide anion (O2˙-), hydroxyl radical
(HO˙), peroxyl radical (L˙), and alkoxyl radical (RO˙), and reactive nitrogen species such
as nitric oxide (NO˙) and various oxidants, play a dual role as both toxic and beneficial
compounds, since they can be either harmful or helpful to the human body. These
compounds have an important role in oxidative stress by damaging important
biomolecules, including lipids, nucleic acids, proteins, and carbohydrates (Gülçin, 2012).
In the human body, various antioxidant enzymes as well as antioxidant molecules such as
phenolic compounds are responsible for neutralizing these free radicals. In addition, small
amounts of these compounds are required for normal physiological functions because they
have an important role in various biological processes such as signal transduction. This
part of the chapter provides a brief overview of the various methods used to assess the
antioxidant capacity.
Plant derived natural products such as polyphenols are excellent antioxidants. There
are various ways of measuring antioxidant capacity of isolated compounds or herbal
extracts. Antioxidant activity can be monitored by a variety of assays based on different
mechanisms, including hydrogen atom transfer (HAT), single electron transfer (SET),
reducing power, and metal chelation, among others (Huang et al., 2005). In general,
depending on the involved chemical reactions, these assays fall into two categories: of
hydrogen atom transfer (HAT) reaction-based assays and single electron transfer (SET)
reaction-based assays.
HAT reaction-based assays measure the ability of antioxidants to quench free radicals
by donating hydrogen atoms. The relative reactivity in HAT-based methods is determined
by the bond dissociation energy of the hydrogen donating group in the potential
antioxidant. The majority of HAT reaction assays apply a competitive reaction scheme in
which antioxidant and substrate used in the assay compete for thermally generated peroxyl
radicals through the decomposition of an azo compound. If the antioxidant can trap peroxyl
radicals, signal is not decreased. Therefore, the magnitude of reduction of signal (mostly
fluorescence) inversely correlates with antioxidant capacity of the specimen.
SET reaction-based assays detect the ability of an antioxidant to transfer one electron
to reduce any compounds, including radicals, carbonyls, or metal ions. Relative reactivity
of an antioxidant in SET methods is based mainly on the ionization potential of the reactive
functional group in the molecule. In SET-based assays, when the oxidant is reduced, it
changes color, and the degree of color change correlates with antioxidant capacity. The
basis of SET-based assays is the following reaction:
hydrazyl – assay is originally described by Blois (1958). DPPH radical scavenging assay
is one of the methods that are used most frequently and presents the first approach for
evaluating antioxidant activity. This method is a SET-based with HAT mechanism being
only a marginal reaction pathway in the assay (Prior et al., 2005). DPPH is a stable radical
with a deep purple color. The DPPH scavenging assay is based on reduction of this free
radical by an antioxidant (Figure 14). The reaction is accompanied with color change of
the DPPH measured at 517 nm, and the discoloration shows the antioxidant efficacy.
In ABTS radical scavenging assay (an electron transfer-based assay), the radical cation
has a blue-green color and it exhibits maximal absorption at 734 nm. The radical cation
form of ABTS (ABTS˙+), is generated from ABTS (Figure 15) in the presence of strong
oxidizing agents such as potassium persulfate (Re et al., 1999). Similar to the DPPH assay,
the potential antioxidant efficacy is determined by the discoloration of the preformed cation
radical by the antioxidant molecule donating an electron or hydrogen atom.
The ferric ion reducing antioxidant power (FRAP) assay is based on the SET by an
antioxidant to reduce the ferric ion (Fe3+) ligand complex to the intensely blue colored
ferrous (Fe2+) complex in acidic media. The original FRAP assay uses tripyridyltriazine
(TPTZ) as the iron-binding ligand. When the ferric-tripyridyltriazine (Fe+3-TPTZ) complex
is reduced to the ferrous (Figure 16), the complex have maximum absorbance at 593 nm
with an intense blue color (Antolovich et al., 2002).
The CUPRAC assay is a cupric reduction assay improved as a different version of the
FRAP assay. In this assay, copper is used as oxidant instead of iron in the FRAP assay.
The method measures the reducing power of antioxidants to convert cupric ion (Cu 2+) to
cuprous ion (Cu+) (Figure 17). The chromogenic oxidizing reagent used for the CUPRAC
assay is the neocuproine (Nc; 2,9-dimethyl-1,10-phenanthroline). When an antioxidant
reduces copper (II)-Nc to copper (I)-Nc an orange-yellow color appears with maximum
absorbance at 450 nm (Apak et al., 2004).
Figure 17. CUPRAC reaction by an antioxidant molecule (HA: antioxidant molecule, A+: an oxidized
antioxidant molecule).
The total phenolic content (TPC) is an important parameter in determining the total
antioxidant capacity, and widely used for evaluation of antioxidant extracts, including
extracts from plants and derivatives. The Folin–Ciocâlteu assay is the most well-known
method for determination of TPC. The Folin–Ciocâlteu method was originally designed
for the analysis of proteins taking advantage of phenolic amino acid tyrosine in proteins
(Folin and Ciocalteu, 1927) and expanded by Singleton et al. (1999) to analyse phenolic
components in wine samples. It has become a routine analysis for antioxidant assessment
of food and plant extracts since that time.
The Folin–Ciocâlteu assay is based on the electron transfer from phenolic compounds
to Folin-Ciocâlteu reagent in alkaline medium (carbonate). The Folin-Ciocâlteu reagent
contains phosphomolybdic/phosphotungstic acid complexes. The molybdenum center
from this complex is reduced from Mo(VI) to Mo(V) by an electron donated by a phenolic
antioxidant, and during this process the complex with light yellow color changes to a blue
color with an absorption maximum at 765 nm. Gallic acid is widely used as reference
standard for this assay, and the results are expressed as gallic acid equivalents.
Transition metals such as iron are known to stimulate lipid oxidation via Fenton
reaction. Antioxidative compounds such as flavonoids and phenolic acids are potent metal
chelators. These antioxidants can deactivate prooxidant metal ions and thus prevent metal
ion-induced lipid oxidation (Hippeli and Elstner, 1999).
Metal chelation capacity is detected by measuring the chelating effects of antioxidants
for ferrous ion (Fe +2). Ferrous sulphate and ferrous chloride which are the most widely
used sources of ferrous ion, and ferrozine (Figure 18) can quantitatively form complexes
with Fe+2. The ferrozine-Fe +2 complex produces a red chromophore with absorbance that
can be measured at 562 nm. The results of metal chelation capacity of antioxidants are
mostly expressed as EDTA (ethylenediaminetetraacetic acid) equivalents.
Currently, in the search for new natural antioxidants, often complex mixtures are
frequently encountered and oftentimes a loss of antioxidant activity happens during the
isolation and purification processes which might be due to the loss of the synergism
between various molecules that are present in the crude extract (Geng et al., 2015). At the
same time, bioassay-guided fractionation of plant samples is tedious, time consuming and
often with low efficiency.
The initial selection of herbal compounds possessing the antioxidant effects is carried
out by in vitro methods. Traditionally used spectrophotometric in vitro methods are widely
applied as mentioned above (Llorent-Martínez et al., 2016; Mocan et al., 2016; Zengin et
al., 2017; Zengin et al., 2014). These methods are able to determine the total amount of
antioxidants based on different chemical mechanisms and to evaluate the total antioxidant
activity in the researched complex herbal samples being very useful to compare between
them samples of the same type (Huang et al., 2005).
Nowadays, most frequently these methods use stable free radicals, such as 1,1’-
diphenyl-2-picrylhydrazyl radical (DPPH•) and 2,2’-azinobis- (3-ethylbenzothiazoline-6-
sulfonic acid) radical cation (ABTS•+) which are among the most common ones. After the
reaction with an antioxidant or a mixture of antioxidants (herbal extract) these stable free
radicals convert to colorless compounds and the decrease of absorption is recorded
(Shahidi and Zhong, 2015). In this case, radical scavenging ability is measured according
to alteration in absorption and evaluated by standard estimates such as well-known natural
or synthetic widely used antioxidants (ascorbic acid, Trolox, BHT, BHA, etc.).
The main problem of natural antioxidant mixtures is that they are not pure substances,
and consequently there is a lack of data concerning safety use and efficacy. Moreover, the
fact of being natural does not guarantee any safe effects. Therefore, efficacy and safety
studies on natural antioxidants are mandatory in the perspective of long use as herbal drugs
or dietary supplements.
In this perspective, one can identify the need for the systems that would help to
distinguish, evaluate and, if possible, to identify individual compounds possessing
antioxidant activity in complex herbal samples, and eventually to establish quality markers
of herbals. Once the antioxidant active substance is determined and the scavenging power
of radicals is assessed according to standard estimates, the performance of efficacy and
safety researches inside in vivo systems becomes possible (Geng et al., 2015; Gong et al.,
2011; Nuengchamnong et al., 2009; Raudonis et al., 2010). However, a solution for
evaluating individual antioxidant capacity is fractionation and isolation which is tedious,
produces a lot of waste (i.e., solvents, plant material), and consequently is not eco-friendly.
To study the antioxidants in plants, separation and purification are necessary steps
traditionally although they are time-consuming and labor-intensive processes (Zhang et al.,
2014).
Therefore, more elaborate techniques and approaches are needed (Mocan et al., 2016).
Currently the on-line based methods, which combine HPLC separation and post-column
derivatization, are designed for the search of individual antioxidants (Figure 19). The
greatest advantages of on-line methods are their selectivity, informatory capability and
high sensitivity for precise determination of antioxidant-active compound and evaluation
of its activity in complex mixtures, as well as being able to detect the individual
contribution of a compound to the overall antioxidant activity (Raudonis et al., 2010).
These systems could also be applied to assess the quality of herbal raw materials, food
supplements and other herbal preparations as well as to establish quality markers for raw
materials emphasized for their high concentrations of antioxidants. Currently, several
manufacturers advertise numerous food supplements and herbal preparations as a rich
source of antioxidants; however, there is no reliable methodology for assessing the
antioxidant activity, safety and stability of the antioxidants present in the respective
products (Raudonis et al., 2010).
Online detection methods used for determining individual components antioxidant
capacity evaluation are generally applied to medicinal plants extracts,
phytopharmaceuticals (Bandonien and Murkovic, 2002; Geng et al., 2015; Gong et al.,
2011; Mocan et al., 2016; Raudonis et al., 2010; Zhang et al., 2014), infusions, decoctions
(teas) (Goulas et al., 2014; Riehle et al., 2013), as well as to edible plants and food matrices
(Damašius et al., 2014; Fiol et al., 2013; Pedan et al., 2016; Zietz et al., 2010). The post-
column derivatization reagents mostly used are: ABTS+ (Fiol et al., 2013; Mocan et al.,
2016; Riehle et al., 2013), DPPH (Geng et al., 2015; Goulas et al., 2014; Nuengchamnong
et al., 2009; Pedan et al., 2016), or combinations based on different antioxidant reagents
such as DPPH and FRAP (Zhang et al., 2015).
Figure 19. HPLC-ABTS equipment system for radical scavenging analysis of individual antioxidants.
REFERENCES
Akagi, M., Fukuishi, N., Tomoko, K., Sagesaka, Y. M., Akagi, R., 1997. Anti-allergic
effect of tea-leaf saponin (TLS) from tea leaves (Camellia sinensis var. sinensis).
Biological and Pharmaceutical Bulletin 20, 565-567.
Alves, M. J., Ferreira, I. C., Froufe, H. J., Abreu, R., Martins, A., Pintado, M., 2013.
Antimicrobial activity of phenolic compounds identified in wild mushrooms, SAR
analysis and docking studies. Journal of applied microbiology 115, 346-357.
An, L., Guan, S., Shi, G., Bao, Y., Duan, Y., Jiang, B., 2006. Protocatechuic acid from
Alpinia oxyphylla against MPP+-induced neurotoxicity in PC12 cells. Food and
chemical toxicology 44, 436-443.
Annie, F., Jean-Jacques, M., 2009. Phenolic acids in fruits and vegetables, Flavonoids in
Health and Disease, 2nd ed. CRC Press., New York, USA.
Antolovich, M., Prenzler, P. D., Patsalides, E., McDonald, S., Robards, K., 2002. Methods
for testing antioxidant activity. Analyst 127, 183-198.
Apak, R., Güçlü, K., Özyürek, M., Karademir, S.E., 2004. Novel total antioxidant capacity
index for dietary polyphenols and vitamins C and E, using their cupric ion reducing
capability in the presence of neocuproine: CUPRAC method. Journal of agricultural
and food chemistry 52, 7970-7981.
Armatu, A., Colceru-Mihul, S., Bubueanu, C., Draghici, E., Pirvu, L., 2010. Evaluation of
antioxidant and free scavenging potential of some Lamiaceae species growing in
Romania. Romanian Biotechnological Letters 15, 5274-5280.
Bandonien, D., Murkovic, M., 2002. The detection of radical scavenging compounds in
crude extract of borage (Borago officinalis L.) by using an on-line HPLC-DPPH
method. Journal of biochemical and biophysical methods 53, 45-49.
Bernhoft, A., 2010. A brief review on bioactive compounds in plants. Bioactive compounds
in plants-benefits and risks for man and animals.
Blois, M.S., 1958. Antioxidant determinations by the use of a stable free radical. Nature
181, 1199-1200.
Bravo, L., 1998. Polyphenols: chemistry, dietary sources, metabolism, and nutritional
significance. Nutrition reviews 56, 317-333.
Cao, G., Sofic, E., Prior, R. L., 1997. Antioxidant and prooxidant behavior of flavonoids:
structure-activity relationships. Free Radical Biology and Medicine 22, 749-760.
Carocho, M., Barreiro, M.F., Morales, P., Ferreira, I.C., 2014. Adding molecules to food,
pros and cons: A review on synthetic and natural food additives. Comprehensive
Reviews in Food Science and Food Safety 13, 377-399.
Carocho, M., Ferreira, I. C., 2013. A review on antioxidants, prooxidants and related
controversy: natural and synthetic compounds, screening and analysis methodologies
and future perspectives. Food and Chemical Toxicology 51, 15-25.
Carradori, S., D’Ascenzio, M., Chimenti, P., Secci, D., Bolasco, A., 2014. Selective MAO-
B inhibitors: a lesson from natural products. Molecular diversity 18, 219-243.
Chaudhuri, T., Das, S., Vedasiromoni, J., Ganguly, D., 1997. Phytochemical investigation
of the roots of Camellia sinensis L. (O. Kuntze). Journal of the Indian Chemical
Society 74.
Corradini, E., Foglia, P., Giansanti, P., Gubbiotti, R., Samperi, R., Laganà, A., 2011.
Flavonoids: chemical properties and analytical methodologies of identification and
quantitation in foods and plants. Natural product research 25, 469-495.
D Archivio, M., Filesi, C., Di Benedetto, R., Gargiulo, R., Giovannini, C., Masella, R.,
2007. Polyphenols, dietary sources and bioavailability. Annali-Istituto Superiore di
Sanita 43, 348.
Dai, J., Mumper, R. J., 2010. Plant phenolics: extraction, analysis and their antioxidant and
anticancer properties. Molecules 15, 7313-7352.
Damašius, J., Venskutonis, P., Kaškonienė, V., Maruška, A., 2014. Fast screening of the
main phenolic acids with antioxidant properties in common spices using on-line
HPLC/UV/DPPH radical scavenging assay. Analytical Methods 6, 2774-2779.
Das, D.K., 1994. [40] Naturally occurring flavonoids: Structure, chemistry, and high-
performance liquid chromatography methods for separation and characterization.
Methods in enzymology 234, 410-420.
Davis, A. L., Lewis, J. R., Cai, Y., Powell, C., Davis, A.P., Wilkins, J.P., Pudney, P.,
Clifford, M.N., 1997. A polyphenolic pigment from black tea. Phytochemistry 46,
1397-1402.
Delmas, D., Lançon, A., Colin, D., Jannin, B., Latruffe, N., 2006. Resveratrol as a
chemopreventive agent: a promising molecule for fighting cancer. Current drug
targets 7, 423-442.
Deschner, E.E., Ruperto, J., Wong, G., Newmark, H.L., 1991. Quercetin and rutin as
inhibitors of azoxymethanol-induced colonic neoplasia. Carcinogenesis 12, 1193-
1196.
Ferreira, D., Slade, D., 2002. Oligomeric proanthocyanidins: naturally occurring O-
heterocycles. Natural product reports 19, 517-541.
Ferreira, I. C., Barros, L., Abreu, R., 2009. Antioxidants in wild mushrooms. Current
Medicinal Chemistry 16, 1543-1560.
Fiol, M., Weckmüller, A., Neugart, S., Schreiner, M., Rohn, S., Krumbein, A., Kroh, L.
W., 2013. Thermal-induced changes of kale’s antioxidant activity analyzed by HPLC–
UV/Vis-online-TEAC detection. Food chemistry 138, 857-865.
Florov, A., Mishenkova, E., 1969. Inhibitory effects of some preparations from higher
plants on reproduction of influenza virus in vitro and in vivo. Mikrobiolohichnyi
zhurnal 32, 628-633.
Folin, O., Ciocalteu, V., 1927. On tyrosine and tryptophane determinations in proteins. J.
biol. Chem 73, 627-650.
Fu, L., Xu, B.-T., Xu, X.-R., Gan, R.-Y., Zhang, Y., Xia, E.-Q., Li, H.-B., 2011.
Antioxidant capacities and total phenolic contents of 62 fruits. Food Chemistry 129,
345-350.
Geetha, T., Garg, A., Chopra, K., Kaur, I. P., 2004. Delineation of antimutagenic activity
of catechin, epicatechin and green tea extract. Mutation Research/Fundamental and
Molecular Mechanisms of Mutagenesis 556, 65-74.
Geng, D., Chi, X., Dong, Q., Hu, F., 2015. Antioxidants screening in Limonium aureum by
optimized on-line HPLC–DPPH assay. Industrial Crops and Products 67, 492-497.
Gidaro, M. C., Alcaro, F., Carradori, S., Costa, G., Vullo, D., Supuran, C.T., Alcaro, S.,
2015. Eriocitrin and apigenin as new carbonic anhydrase VA inhibitors from a virtual
screening of calabrian natural products. Planta medica 81, 533-540.
Gidaro, M. C., Astorino, C., Petzer, A. l., Carradori, S., Alcaro, F., Costa, G., Artese, A.,
Rafele, G., Russo, F. M., Petzer, J. P., 2016. Kaempferol as Selective Human MAO-A
Inhibitor: Analytical Detection in Calabrian Red Wines, Biological and Molecular
Modeling Studies. Journal of agricultural and food chemistry 64, 1394-1400.
Golukcu, M., Ozdemir, F., 2010. Changes in phenolic composition of avocado cultivars
during harvesting time. Chemistry of natural compounds 46, 112-115.
Gong, Y., Plander, S., Xu, H., Simandi, B., Gao, Y., 2011. Supercritical CO2 extraction of
oleoresin from marigold (Tagetes erecta L.) flowers and determination of its
antioxidant components with online HPLC‐ ABTS•+ assay. Journal of the Science of
Food and Agriculture 91, 2875-2881.
Goulas, V., Exarchou, V., Kanetis, L., Gerothanassis, I. P., 2014. Evaluation of the
phytochemical content, antioxidant activity and antimicrobial properties of mountain
tea (Sideritis syriaca) decoction. Journal of Functional Foods 6, 248-258.
Grassi, D., Desideri, G., Ferri, C., 2010. Flavonoids: antioxidants against atherosclerosis.
Nutrients 2, 889-902.
Gülçin, I., 2012. Antioxidant activity of food constituents: an overview. Archives of
toxicology 86, 345-391.
Gyamfi, M. A., Aniya, Y., 2002. Antioxidant properties of Thonningianin A, isolated from
the African medicinal herb, Thonningia sanguinea. Biochemical Pharmacology 63,
1725-1737.
Han, R.-M., Tian, Y.-X., Liu, Y., Chen, C.-H., Ai, X.-C., Zhang, J.-P., Skibsted, L.H.,
2009. Comparison of flavonoids and isoflavonoids as antioxidants. Journal of
agricultural and food chemistry 57, 3780-3785.
Heim, K. E., Tagliaferro, A.R., Bobilya, D.J., 2002. Flavonoid antioxidants: chemistry,
metabolism and structure-activity relationships. The Journal of nutritional
biochemistry 13, 572-584.
Heleno, S. A., Ferreira, I.C., Esteves, A. P., Ćirić, A., Glamočlija, J., Martins, A., Soković,
M., Queiroz, M.J.R., 2013. Antimicrobial and demelanizing activity of Ganoderma
lucidum extract, p-hydroxybenzoic and cinnamic acids and their synthetic acetylated
glucuronide methyl esters. Food and Chemical Toxicology 58, 95-100.
Hertog, M., Hollman, P. C., 1996. Potential health effects of the dietary flavonol quercetin.
Hertog, M. G., Feskens, E. J., Kromhout, D., Hollman, P., Katan, M., 1993. Dietary
antioxidant flavonoids and risk of coronary heart disease: the Zutphen Elderly Study.
The Lancet 342, 1007-1011.
Hippeli, S., Elstner, E. F., 1999. Transition metal ion‐ catalyzed oxygen activation during
pathogenic processes. FEBS letters 443, 1-7.
Hirasawa, M., Takada, K., 2004. Multiple effects of green tea catechin on the antifungal
activity of antimycotics against Candida albicans. Journal of Antimicrobial
Chemotherapy 53, 225-229.
Huang, D., Ou, B., Prior, R. L., 2005. The chemistry behind antioxidant capacity assays.
Journal of agricultural and food chemistry 53, 1841-1856.
Iacopini, P., Baldi, M., Storchi, P., Sebastiani, L., 2008. Catechin, epicatechin, quercetin,
rutin and resveratrol in red grape: Content, in vitro antioxidant activity and interactions.
Journal of Food Composition and Analysis 21, 589-598.
Ignat, I., Volf, I., Popa, V. I., 2011. A critical review of methods for characterisation of
polyphenolic compounds in fruits and vegetables. Food chemistry 126, 1821-1835.
Jaganath, I. B., Crozier, A., 2010. Dietary flavonoids and phenolic compounds. Plant
phenolics and human health: biochemistry, nutrition, and pharmacology 1, 1-50.
Kelebek, H., Selli, S., 2011. Evaluation of chemical constituents and antioxidant activity
of sweet cherry (Prunus avium L.) cultivars. International Journal of Food Science &
Technology 46, 2530-2537.
Khan, N., Mukhtar, H., 2009. 7 Green Tea Catechins: Anticancer Effects and Molecular
Targets. Plant Phenolics and Human Health: Biochemistry, Nutrition and
Pharmacology 1, 165.
Khanbabaee, K., van Ree, T., 2001. Tannins: classification and definition. Nat Prod Rep
18, 641-649.
Kim, Y.-J., 2007. Antimelanogenic and antioxidant properties of gallic acid. Biological
and Pharmaceutical Bulletin 30, 1052-1055.
Klaus, S., Pültz, S., Thöne-Reineke, C., Wolfram, S., 2005. Epigallocatechin gallate
attenuates diet-induced obesity in mice by decreasing energy absorption and increasing
fat oxidation. International journal of obesity 29, 615-623.
Knekt, P., Jarvinen, R., Reunanen, A., Maatela, J., 1996. Flavonoid intake and coronary
mortality in Finland: a cohort study. Bmj 312, 478-481.
Konczak, I., Zhang, W., 2004. Anthocyanins—more than nature's colours. BioMed
Research International 2004, 239-240.
Kong, J.-M., Chia, L.-S., Goh, N.-K., Chia, T.-F., Brouillard, R., 2003. Analysis and
biological activities of anthocyanins. Phytochemistry 64, 923-933.
Llorent-Martínez, E., Ortega-Barrales, P., Zengin, G., Uysal, S., Ceylan, R., Guler, G.,
Mocan, A., Aktumsek, A., 2016. Lathyrus aureus and Lathyrus pratensis:
characterization of phytochemical profiles by liquid chromatography-mass
spectrometry, and evaluation of their enzyme inhibitory and antioxidant activities. RSC
Advances 6, 88996-89006.
Luo, F.-j., Hu, Z., Deng, X.-y., Zhao, Y., Zeng, L., Dong, Z.-g., Yi, W., Cao, Y., 2001.
Effect of tea polyphenols and EGCG on nasopharyngeal carcinoma cell proliferation
and the mechanisms involved. Chinese Journal of Cancer Research 13, 235-242.
Mahadevan, A., Koti Reddy, M., 1968. The effect of phenolic compounds on growth,
polygalacturonase production and activity of Fusarium oxysporum f. vasinfectum.
European Journal of Plant Pathology 74, 87-90.
Marques, F. Z., Markus, M. A., Morris, B. J., 2009. Resveratrol: cellular actions of a potent
natural chemical that confers a diversity of health benefits. The international journal
of biochemistry & cell biology 41, 2125-2128.
Matsui, Y., Kumagai, H., Masuda, H., 2006. Antihypercholesterolemic activity of
catechin-free saponin-rich extract from green tea leaves. Food science and technology
research 12, 50-54.
McCullough, M. L., Peterson, J. J., Patel, R., Jacques, P. F., Shah, R., Dwyer, J. T., 2012.
Flavonoid intake and cardiovascular disease mortality in a prospective cohort of US
adults. The American journal of clinical nutrition 95, 454-464.
Middleton, E., Kandaswami, C., 1992. Effects of flavonoids on immune and inflammatory
cell functions. Biochemical pharmacology 43, 1167-1179.
Mikaberidze, K., Moniava, I., 1974. Umbelliferone from tea leaves. Chemistry of Natural
Compounds 10, 81-81.
Miyake, Y., Sasaki, S., Ohya, Y., Miyamoto, S., Matsunaga, I., Yoshida, T., Hirota, Y.,
Oda, H., 2005. Soy, isoflavones, and prevalence of allergic rhinitis in Japanese women:
the Osaka Maternal and Child Health Study. Journal of allergy and clinical
immunology 115, 1176-1183.
Mocan, A., Schafberg, M., Crișan, G., Rohn, S., 2016. Determination of lignans and
phenolic components of Schisandra chinensis (Turcz.) Baill. Using HPLC-ESI-ToF-
MS and HPLC-online TEAC: Contribution of individual components to overall
antioxidant activity and comparison with traditional antioxidant assays. Journal of
Functional Foods 24, 579-594.
Muchuweti, M., Kativu, E., Mupure, C., Chidewe, C., Ndhlala, A., Benhura, M., 2007.
Phenolic composition and antioxidant properties of some spices. American Journal of
Food Technology 2, 414-420.
Nijveldt, R. J., Van Nood, E., Van Hoorn, D. E., Boelens, P. G., Van Norren, K., Van
Leeuwen, P. A., 2001. Flavonoids: a review of probable mechanisms of action and
potential applications. The American journal of clinical nutrition 74, 418-425.
Nuengchamnong, N., Krittasilp, K., Ingkaninan, K., 2009. Rapid screening and
identification of antioxidants in aqueous extracts of Houttuynia cordata using LC–
ESI–MS coupled with DPPH assay. Food Chemistry 117, 750-756.
Obón, J., Díaz-García, M., Castellar, M., 2011. Red fruit juice quality and authenticity
control by HPLC. Journal of Food Composition and Analysis 24, 760-771.
Olthof, M. R., Hollman, P. C., Katan, M. B., 2001. Chlorogenic acid and caffeic acid are
absorbed in humans. The Journal of nutrition 131, 66-71.
Pedan, V., Fischer, N., Rohn, S., 2016. An online NP-HPLC-DPPH method for the
determination of the antioxidant activity of condensed polyphenols in cocoa. Food
Research International 89, 890-900.
Piazzon, A., Vrhovsek, U., Masuero, D., Mattivi, F., Mandoj, F., Nardini, M., 2012.
Antioxidant activity of phenolic acids and their metabolites: synthesis and antioxidant
properties of the sulfate derivatives of ferulic and caffeic acids and of the acyl
glucuronide of ferulic acid. Journal of agricultural and food chemistry 60, 12312-
12323.
Porter, L., Dey, P., Harborne, J., 1989. Methods in plant biochemistry. Methods in plant
biochemistry 1.
Prior, R. L., Wu, X., Schaich, K., 2005. Standardized methods for the determination of
antioxidant capacity and phenolics in foods and dietary supplements. Journal of
agricultural and food chemistry 53, 4290-4302.
Pugazhendhi, D., Pope, G., Darbre, P., 2005. Oestrogenic activity of p‐ hydroxybenzoic
acid (common metabolite of paraben esters) and methylparaben in human breast cancer
cell lines. Journal of Applied Toxicology 25, 301-309.
Raudonis, R., Bumblauskiene, L., Jakstas, V., Pukalskas, A., Janulis, V., 2010.
Optimization and validation of post-column assay for screening of radical scavengers
in herbal raw materials and herbal preparations. Journal of Chromatography A 1217,
7690-7698.
Re, R., Pellegrini, N., Proteggente, A., Pannala, A., Yang, M., Rice-Evans, C., 1999.
Antioxidant activity applying an improved ABTS radical cation decolorization assay.
Free radical biology and medicine 26, 1231-1237.
Rice-Evans, C. A., Miller, N. J., Paganga, G., 1996. Structure-antioxidant activity
relationships of flavonoids and phenolic acids. Free radical biology and medicine 20,
933-956.
Riehle, P., Vollmer, M., Rohn, S., 2013. Phenolic compounds in Cistus incanus herbal
infusions—antioxidant capacity and thermal stability during the brewing process.
Food research international 53, 891-899.
Riso, P., Erba, D., Criscuoli, F., Testolin, G., 2002. Effect of green tea extract on DNA
repair and oxidative damage due to H 2 O 2 in Jurkat T cells. Nutrition research 22,
1143-1150.
Chapter 5
*
Corresponding Author address. Email: anatoly.sobolev@cnr.it.
2
Laboratorio di Risonanza Magnetica “Annalaura Segre”, Istituto di Metodologie
Chimiche, CNR, Monterotondo, Rome, Italy
3
Chemometrics & Analytical Technology, Department of Food Science,
University of Copenhagen, Copenhagen, Denmark
4
Dipartimento di Scienze e Tecnologie Chimiche, Università di Roma
“Tor Vergata”, Rome, Italy
5
Istituto per lo Studio delle Macromolecole, Lab NMR, CNR, Milan, Italy
6
Dipartimento di Scienze degli Alimenti, Università degli Studi di Parma,
Parma, Italy
7
Dipartimento di Scienze e Tecnologie Agro-Alimentari,
Università di Bologna, Cesena, Italy
8
Dipartimento di Scienze Chimiche e Geologiche, Università di Cagliari,
Monserrato (Cagliari), Italy
9
Consiglio per la Ricerca in Agricoltura e l’Analisi dell’Economia Agraria – Centro
di Ricerca per lo Studio delle Relazioni tra Pianta e Suolo (CREA-RPS), Rome, Italy
10
Dipartimento di Scienze e Tecnologie Biologiche ed Ambientali,
Università del Salento, Campus Ecotekne, Lecce, Italy
11
CNR-IPCF, Istituto per i Processi Chimico-Fisici del CNR di Messina,
Messina, Italy
12
Dipartimento di Chimica, Sapienza Università di Roma, Rome, Italy
13
Dipartimento di Ingegneria Civile, Ambientale, del Territorio, Edile e di Chimica
(DICATECh), Politecnico di Bari, Bari, Italy
14
Italian Institute of Technologies, @CNLS Sapienza, Rome, Italy
15
Dipartimento di Chimica, Università degli Studi di Bari “Aldo Moro”,
Bari, Italy
16
Centro di Ricerca di Risonanze Magnetiche CERM, Università di Firenze,
Sesto Fiorentino (Florence), Italy
17
Consorzio interuniversitario per lo sviluppo dei Sistemi a Grande Interfase - CSGI,
Sesto Fiorentino (Florence), Italy
18
Dipartimento di Scienze Chimiche, Università degli Studi di Padova, Padova, Italy
19
Centro Interdipartimentale per la Risonanza Magnetica Nucleare per l’Ambiente,
l’Agro-Alimentare ed i Nuovi Materiali (CERMANU), Università di Napoli Federico
II, Portici (Naples), Italy
20
Istituto Zooprofilattico Sperimentale dell’Abruzzo e del Molise “G. Caporale”,
Teramo, Italy
21
Dipartimento di Scienze Chimiche e Geologiche,
Università di Modena e Reggio Emilia, Modena, Italy
22
Dipartimento di Farmacia, Università degli Studi di Salerno, Salerno, Italy
23
Dipartimento di Farmacia, Università di Napoli Federico II, Naples, Italy
24
Dipartimento di Scienze per la Qualità della Vita, Università di Bologna,
Rimini, Italy
25
Dipartimento di Scienze Biomediche, Odontoiatriche e delle Immagini
Morfologiche e Funzionali, Università degli Studi di Messina, Messina, Italy
26
Dipartimento di Scienza Applicata e Tecnologia, Politecnico di Torino,
Tourin, Italy
27
Dipartimento di Medicina Sperimentale e Clinica,
Università di Firenze, Florence, Italy
28
Dipartimento MIFT, sezione di Fisica, Università di Messina, Messina, Italy
ABSTRACT
1. INTRODUCTION
Taking into account the growing contribution of NMR methodologies in food science
and the increasing interest of NMR community in this field, in 2015, in the frame of Italian
NMR Discussion Group (GIDRM), the Italian Group of Magnetic Resonance in Food
Science was established. This interdisciplinary group includes NMR experts interested in
food science. This chapter, fruit of the collective work of Italian Group of Magnetic
Resonance in Food Science, is aimed to introduce some fundamental aspects of NMR
methodologies suitable for the application in food analysis. NMR methodologies offer a
comprehensive characterization of foodstuff owing to the possibility to study a sample
from different points of view including structural, compositional, functional etc. aspects.
High resolution NMR spectroscopy applied to semi-solid food samples or to extracts in
solution is used to determine the foodstuff chemical composition. In particular, the
comprehensive analysis of food metabolites (food metabolomics) by high-resolution NMR
has found more and more attention. Moreover, nutritional metabolomics focused on
functional aspects of food intake is the object of numerous studies. Finally, low field NMR
relaxometry and imaging applied to intact food samples give precious information
regarding their morphology and texture.
Finally, different online databases related to different aspects of NMR and food science
are available, see section 2.6. In particular, an attempt to create a comprehensive database
including NMR spectra of different foodstuffs and data on their origin, composition,
variety etc is undertaken by the Italian Group of Magnetic Resonance in Food Science in
the frame of an Italian project entitled “e-ALIERB: un OPEN LAB per caratterizzare e
valorizzare i prodotti alimentari ed erboristici del territorio laziale” funded by Regione
Lazio Lr13/2018.
2.1. HR-MAS
High-resolution magic angle spinning (HR-MAS) NMR allows the investigation of not
liquid samples without any pretreatment obtaining spectra with high resolution. In fact, in
not liquid systems, the NMR spectral sensitivity and resolution are limited by strong
anisotropic interactions that provoke line-broadening effects also at high magnetic fields.
These line broadening effects are mainly due to a non-uniform magnetization (the magnetic
susceptibility changes with spin position) and, in addition, to the spins dipolar coupling,
and to the chemical-shift anisotropy (CSA) effect. The HR-MAS technique indeed has been
developed to reduce the two main line-broadening mechanisms that are important in
acquiring spectra of a tissue or cell sample, namely dipolar coupling and heterogeneous
isotropic susceptibility (Chen & Singer, 2007). As mentioned, the local magnetic field
created by dipolar coupling and by heterogeneous susceptibility changes from position to
position and as well as the resonance frequency. This provokes a frequency distribution for
each resonance frequency that depends on the intensity of the two effects and causes the
line-broadening of NMR peaks. To bypass these problems, in HR-MAS technique, the
sample is put in a rotor and spun at the magic angle of about 54.74° by few thousands of
Hertz to reduce these mechanisms. Nowadays, the rotor can spin at frequencies between
about 1 and 100 kHz: at higher frequencies the signal results more enhanced and with
higher resolution. For example, Figure 1 shows the HR-MAS proton spectra of a plant leaf
acquired on a 400 MHz spectrometer with spinning frequencies ranging from 4 to 19 kHz.
It is noteworthy that the signal increases with the spinning frequency as well as the
resolution. In fact, HR-MAS maintains, even for semi-solid samples, the high instrumental
sensitivity and precision of liquid state NMR, permitting to obtain a large number of
information on the whole metabolic profiles of complex matrices by means of a single
experiment (Corsaro et al., 2015; Mannina et al., 2012).
In addition, HR-MAS has the intrinsic advantage that the preparation of the
investigated sample is minimal or absent. Furthermore, micro-quantities of sample
(heterogeneous samples such as tissues and cells) (Corsaro et al., 2015), are sufficient to
acquire high-resolution (HR) spectra with a very good signal/noise ratio (Figure 2).
HR-MAS NMR can be considered a hybrid technique between classical solution state
NMR and solid-state NMR. In fact, this technique keeps the classical solution NMR
experiments including bidimensional pulse sequences, but, similarly to solid-state NMR,
it uses the magic angle spinning approach (Andrew et al., 1959; Lindon et al., 2007). Of
course, as well as classical NMR, the HR-MAS technique allows carrying out quantitative
analyses on the concentrations of the identified metabolites within the sample by means
for example of a reference compound with known concentration (see next paragraphs).
Figure 1. A comparison between NMR proton spectra obtained with several spinning frequencies in
HR-MAS NMR techniques. Figure adapted from the Bruker manual “High Resolution Magic Angle
Spinning Spectroscopy”.
Although HR-MAS spectroscopy is a tool of great value for the analysis of foodstuffs,
relevance and value have not completely expressed yet. Just recently, this new NMR
technique, due to the simultaneous detection in the same experiment of polar and apolar
metabolites in talis qualis samples, has become more popular in food science and in the
biological and biomedical fields (Lindon et al., 2009; Corsaro & Mallamace, 2011; Santos
et al., 2015).
As mentioned before, HR-MAS offers the unique possibility to study the metabolic
changes of intact tissue due to different factors (for example, seasonal variations,
transgenic modification, aging, etc.) and/or the effect of a specific “remedy” on a matrix
(Corsaro et al., 2015; Santos et al., 2015).
Figure 2. A comparison between the static NMR and HR-MAS NMR techniques. Figure adapted from
Fitch et al. (Fitch et al., 1994).
In conclusion, the HR-MAS technique is one of the most suited approaches to study
foodstuffs and their metabolic pathways. HR-MAS quickly provides a broad range of
information directly from the material under investigation and that can be used to solve
problems related to food science. As shown in the next paragraphs, this important
spectroscopic tool has a lot to offer for the analysis of very different food matrices.
The use of NMR spectroscopy for quantification purposes has been exhaustively
reviewed by Bharti and Roy. In the present chapter, some points of the review have been
extracted and updated with new findings (Gallo et al., 2015a).
Basics for NMR quantification rely on the fundamental relationship existing between
the signal intensity (I) in the NMR spectrum and the moles (n) of the nuclei generating that
particular signal. I is directly proportional to n according to equation (1):
I = k•n (1)
where k is the spectrometer constant and depends on acquisition settings such as pulse
excitation, repetition time, broad-band decoupling, etc.
In principle, without any kind of calibration, NMR spectroscopy is intrinsically able to
provide immediately quantitative information on the studied system being also capable to
distinguish between different isomers. The methods that NMR spectroscopy can furnish to
measure the concentration of the different metabolites within a mixture can be divided into
relative and absolute (Bharti & Roy, 2012; Malz & Jancke, 2005).
Relative quantification methods allow the determination of the molar ratio between
two different compounds, providing that the corresponding 1H NMR signals were correctly
assigned. In details, by denoting with Ii and Ni the integrated area and the number of nuclei
(originating that signal) of the i-th metabolite, respectively, the molar ratio between the
metabolites X and Y can be written as (Bharti & Roy, 2012; Malz & Jancke, 2005):
(2)
Note that in optimized 1H NMR experiments (Griffiths & Irving, 1998) the
spectrometer constant (k in equation 1) is the same for all resonances, therefore it does not
appear in the expression of the molar ratio (equation 2). In analogy, the molar fraction of
the compound X in a mixture of Z components can be thus calculated by means of the
following formula:
(3)
It is noteworthy that this procedure is independent from the solvent and from its amount
used to dissolve the mixture, since its resonance frequency can be excluded from the
calculation. The limitations of these methods, with particular emphasis to that described by
equation 2, include the needs of the complete assignment of the spectrum and the possible
occurrence of overlaps among the signals of different compounds. Both these
circumstances usually take place in complex matrices such as those of food products and
in this case different approaches can be applied for the metabolites quantification.
Once the signal from a compound of known concentration (reference compound) is
present in the acquired NMR spectrum, all the areas of the compounds of interest can be
referred to the reference area with the aim to obtain the absolute concentration of that
compound within the studied mixture. Note that, the reference compound can be internal,
that is dissolved within the same mixture, or external, that is for example contained in a
capillary within the same sample holder. Nowadays several reference compounds are
available in the market, and sometimes already in prepared solutions. They are often non-
volatile substances chosen as a function of their solubility with respect to the solvent used.
Moreover, usually their signals fall in regions of the spectrum where the signals of interest
do not appear.
The purity of the component X (PX) can be calculated as (Bharti & Roy, 2012; Malz &
Jancke, 2005; Griffiths & Irving, 1998):
(4)
where I and N have the meaning already seen, m represents the molar mass, W the
gravimetric weight, the subscripts R and T refer to the reference standard compound and
total sample, respectively.
Note that the detection, and thus the quantification, of minor compounds in mixture
are facilitated by the use of intense magnetic fields. In fact, for higher magnetic fields, the
sensitivity is enhanced because the macroscopic magnetization increases linearly with the
amplitude of the field. Furthermore, the signal-to-noise ratio is proportional to the square
root of the cube of the magnetic field. Finally, higher fields provide also higher chemical
shift dispersion (in Hz units) allowing an easier compounds identification.
The quantification procedures are constantly improved thanks to the development of
modern technology. In fact, there is a growing interest towards the improvement of online
platforms and databases, see below.
In the following, the reasoning to determine the theoretical slope for a generic signal
is reported. In a NMR spectrum, the signal (a) has intensity Ia generated by specific protons
belonging to the analyte of interest and the signal (r) has intensity Ir generated by specific
protons in a reference compound.
Applying eq. 1 to Ia and Ir gives:
2.3. Metabolomics
Metabolomics is one of the -omic sciences and its goal is to detect and quantify
metabolites (small molecules <2000 Da) in a biological sample. The ensemble of all
metabolites contained in one cell, tissue, organ or organism (depending on the system
investigated) is called the metabolome (Fiehn, 2002).
Strictly speaking, there exist two similar terms to identify this discipline:
metabonomics, defined as “the quantitative measurement of the dynamic multiparametric
metabolic response of living systems to pathophysiological stimuli or genetic modification”
(Nicholson et al., 1999), and metabolomics, introduced later and defined as the
“comprehensive and quantitative analysis of all metabolites” in a system (Fiehn, 2002).
Even if there is a little philosophical, rather than technical, difference between the two
terms, they are considered to be equivalent and used interchangeably by the scientific
community (Everett, 2015). Though these terms were coined years later, the concept was
born with early attempts to the simultaneous analysis of metabolites present in biological
fluids through 1H NMR spectroscopy in the 1980’s (Nicholson & Wilson, 1989).
The metabolome can be considered the final fulfillment of the genome: for this reason
the metabolic space represents an optimal level at which to analyze changes in biological
systems with high sensitivity (Li et al., 2003). In other words, the metabolome reflects the
current biological state of an organism because it is the endpoint of all interactions and
reactions among the genome, transcriptome, proteome, microbiome, including also the
effects of environment, lifestyle, diet, physical exercise, and pollutants.
Targeted and untargeted approaches are possible in metabolomics, the former focusing
on the analysis of a subset of known compounds, the latter focusing on the whole array of
metabolites detected. Using both approaches hundreds to thousands of metabolites can be
detected (but not necessarily identified or quantified). Obtained data are usually analyzed
following standardized metabolomics pipelines (Salek et al., 2015), and information is
extracted using state-of-the-art statistical tools (Cacciatore et al., 2014).
Mass spectrometry (MS) and 1H NMR spectroscopy represent the analytical techniques
most commonly employed in metabolomics.
NMR is the ideal technique for untargeted screening of biological samples, as it is
characterized by high-reproducibility, high throughput, minimal and cost effective sample
preparation, and it is intrinsically robust and quantitative without requiring dedicated
calibration. Unfortunately, its sensitivity is limited to the micromolar range. In contrast,
mass spectrometry is better suited for targeted analysis, and thanks to its very high
sensitivity the detection limit easily reach the nanomolar range (Emwas, 2015). In a large
scale assessment (Dumas et al., 2006) it was proved that NMR spectroscopy combined
with multivariate pattern recognition is a robust and precise approach for metabolomics
studies, outperforming other “-omic” technologies in terms of reproducibility. The
instrumental reproducibility of NMR spectroscopy has been assessed in a large scale ring-
test (Gallo et al., 2015a) in which the different participants were able to produce NMR
spectra of a given mixture that were statistically equivalent in terms of relative intensities
of the signals. In practice NMR and MS can be considered two highly synergistic
techniques, with NMR mainly used for hypothesis generation (fast untargeted analysis)
and MS for hypothesis verification by means of the fine-tuned analysis of the low
concentrated metabolites presumably involved in the pathways enlightened by NMR.
The NMR spectrum of food matrix can be called “metabolic fingerprint”, because it
constitutes a snapshot of the portion of the whole metabolome detectable by NMR. From
mono-dimensional 1H NMR spectra it is possible to extract useful information using
multivariate statistical analysis, and this feature could confer to metabolomics a key role
in food investigation. The metabolic fingerprint in fact acts as a sort of “complex
biomarker” more sensitive than each of the individual molecules it is composed of. The
fingerprint of the food itself can be obtained to extract information about origin, traceability
and composition of the products.
Generally speaking, the metabolome is influenced by foods, drinks, drugs, and other
environmental products. Because different nutritional habits and host/guest interactions
with the gut microbiome lead to different metabolic signatures, metabolomics can reveal
dietary intake patterns (Lenz et al., 2004; Lloyd et al., 2011; Solanky et al., 2005; Zuppi et
al., 1998) and represents a powerful tool for the understanding of the outcomes of dietary
intervention (Andersen et al., 2014; O’Sullivan et al., 2011).
In summary, NMR-based metabolomics can be considered a sort of “universal” robust
and quantitative analytical technique (Goodacre et al., 2007; Simmler et al., 2014) and the
method of choice for studies involving health, nutrition and food composition (Goodacre
et al., 2007; Simmler et al., 2014).
2.4. Nutri-Metabolomics
Modern nutrition has the goal of understanding how diet and food intake regulate the
metabolic state of humans (Bordoni & Capozzi, 2014). Mammals are extremely
biologically complex systems, thus it is fundamental to intertwine different technologies
applied in systems biology to pursue current nutrition objectives.
Metabolomics, as already described, is a very powerful holistic approach capable of
identifying and measuring simultaneously many low molecular weight compounds present
in bio-samples, thus characterizing a whole biological system in a fast and thorough way.
The biochemical profiles obtained through metabolomics techniques are influenced by
many inherent factors such as genotype or environmental factors, like the gut microbiota
or a specific diet.
In order to validate the emerging nutrigenomic findings and assess the down-stream
effects of biological variations due to differences in structural genomic, it is clear the
importance of an approach capable of actually shedding a light onto the metabolic end-
points. (Özdemir & Kolker, 2016)
Nutritional metabolomics, or nutri-metabolomics, is thus increasingly employed for
the investigation of the molecular interaction between diet and the biological host. Being
an unbiased approach, it is possible to analyze and assess many physiological
measurements, allowing also a rapid identification of metabolic diseases and the influence
of nutrients and foods on them (LeMieux et al., 2014).
The first studies in nutri-metabolomics concerned the assessment of the molecular
relations between biochemical processes and nutrition, identifying food biomarkers in
order to later employ them as a control in dietary interventions and to understand the
complex trans-genomic interactions between the microbiota and its host. (Claus & Swann,
2013).
The applications of nutri-metabolomics are rapidly increasing, in the light of the search
of a tailor-made nutrition capable of guaranteeing the best health state possible (Bordoni
& Capozzi, 2015).
The main fields of interest and application of nutri-metabolomics can now be
summarized in three main points (Llorach et al., 2012):
For what concerns the first point, it is in fact necessary to evaluate the accuracy of the
reported dietary questionnaires usually employed by nutritional studies and nutri-
metabolomics can be an efficient tool in this sense (Capozzi and Trimigno, 2015). In effect,
the reported food intakes can be validated by the presence in the metabolome (metabolic
profile of a bio-fluid) of specific food or dietary biomarkers (Trimigno et al., 2015).
Garcia-Aloy et al., for example, employed HPLC-q-TOF-MS and multivariate
statistical tools to identify the biomarkers of bread consumption in a free-living population.
It was found how a consistent bread consumption affected the urinary metabolic profile
with a specific subset of compounds and how, therefore, a specific biomarker pattern was
able to identify bread-consumers (Garcia-Aloy et al., 2015). A similar approach was
employed by Vázquez-Fresno et al., but with NMR tools, to identify wine biomarkers after
an intervention, validating their metabolite pattern through the employment of the
calculated model on a free living population with wine-consumers. In this way, the wine-
consumer biomarkers were found and can help in both future dietary interventions or
investigation and on the assessment of the potential benefits of moderate wine
consumptions through the study of the mechanism of action of these biomarkers (Vázquez-
Fresno et al., 2015a).
The importance of specific food components and nutrients in the health status has been
highlighted more and more. It is therefore necessary to investigate the real accessibility of
these bioactive nutrients and their effect on the metabolome and this is how nutri-
metabolomics comes to hand again. This type of studies can be done both in vitro,
employed simulated gastric models, and in vivo, with more expensive and lengthy
interventions. Pan et al. investigated the down-stream products of ham digestion through
an in vitro model by NMR spectroscopy, showing how the consumption of this protein-
rich product can effectively make important nutrients available for human beings (Pan et
al., 2014).
Zhu et al. employed an acute intervention study to define the biomarkers of whole grain
wheat intake through both non-targeted and targeted nutri-metabolomics approached (Zhu
et al., 2016). In this way, the specific biomarker pattern for this food product was found
and can be of great aid in understanding the metabolic effect of its consumption and its
actual impact on health, since whole grain is believed to help in the prevention of many
chronic diseases such as cancer or cardiovascular pathologies.
The importance of the assessment of the real molecular and metabolic mechanisms of
bioactive nutrients is one of the main drives in nutri-metabolomics studies, as evident.
Also Bondia-Pons et al. employed this method, with NMR techniques, to investigate
the effects of lycopene ingestion, comparing the consumption of two tomato sauces with
different concentrations of this metabolite and thus the global effect on the human
metabolome of this interesting carotenoid (Bondia-Pons et al., 2013).
Moreover, these nutri-metabolomics approaches can be employed, as stated, for the
assessment of the effects of specific diets, considered potentially beneficial, such as the
Mediterranean diet. Váquez-Fresno et al. found how the Mediterranean diet impacted on
the urinary 1H NMR metabolome, affecting both the human and the gut-microbiota
metabolism. This study highlights the great potential of this methodology for the evaluation
of metabolic changes caused by dietary interventions (Váquez-Fresno et al., 2015b).
It is thus clear how the nutri-metabolomics field is of great potential in the discovery
of the underlying relations between nutrients, diets and human health, starting from the
assessment of food biomarkers and ending up in the explanation of the biochemical
changes occurring in human metabolism after specific food consumptions or dietary
interventions and their impact on the global health status and on the prevention of specific
diseases.
Furthermore, the application of NMR spectroscopy has been of great help in this field
of research, thanks to its reproducibility and its fast analysis and its capability to survey a
good portion of the metabolic profile of a specific biological specimen. The universality of
NMR has in fact classified it as a preferred procedure in nutri-metabolomics assessments.
2.5. Chemometrics
In the previous sections the great potential of High Resolution (HR) 1H NMR
spectroscopy for unraveling the significant metabolites, biomarkers responsible of many
sought or unwanted characteristics in food matrices and living systems, has been
introduced.
The strength of the 1H NMR analytical approach for foodstuff is its ability to look at
most of the metabolic components of a mixture in a simultaneous fashion, which enables
an extensive metabolite quantitative analysis, thus generating a large amount of data,
highly collinear and covariant. It is therefore necessary to extract and identify the relevant
information out of the data and this requires a quite challenging data mining strategy.
Chemometrics represents the perfect solution to such a challenge, since it can reduce
the dimension of the NMR spectral data and, at the same time, points out the existence of
possible patterns among the analyzed sample (Giuliani et al., 1991; Giuliani et al., 1993;
Giuliani et al., 2004; Khakimov et al., 2015).
But how can we define chemometrics? Scientific quantitative studies have historically
been based on cause-effect one-to-one relationships, thoroughly exploiting and developing
univariate data analysis methods coupled with randomized or structured design of
experiments. This approach allowed understanding those phenomena that could be
ascribed to uni-or low- variated causes and was very well suited for the relative low amount
of information that univariate instruments provided, but represents nowadays a limitation
for investigating the overwhelming amount of data that modern analytical platforms can
collect in a short time and that can dig into the essence of more complex systems, such as
those within food metabolomics, which are characterized by latent many-to-many
relationships. For unraveling these hidden relationships, multivariate methods, able to
perform unsupervised data exploration, are required. Chemometric methods rely on the
extraction of common latent factors (i.e., principal components) from underlying common
or latent structures in the analyzed original data. Chemometric methods for data exploration
and mining are designed to handle large multivariate data sets, exploiting the collinearity
between variables and projecting the original multivariate data into a few dimensions
(latent variables) that are much easier to be described by simple graphical representations.
As it will be presented in the following sections, chemometric unsupervised and
supervised approaches to foodstuff analysis have been successfully applied, giving relevant
results, for the authentication of the geographical origin, the differentiation of varieties or
breeding systems and the evaluation of several food quality aspects that are crucial for
foodstuff protection, quality assessment and safety.
2.5.1. Preprocessing
Still, it must be stressed that the data collected from the NMR spectrometer are most
likely affected by several biases that are both user, instrumental and experimental setup
dependent. If FID and spectral processing (carried out after the signal acquisition) help
correcting for most of the instrumental biases and the compliance with well-designed
Standard Operative Protocols (SOPs) drastically reduces the risk of user dependent sources
of unwanted variance, the spurious variance introduced by unavoidable food matrix
processing and sample collection and/or preparation, often makes the raw data, directly
collected from the instrument, non-suitable for a reliable data mining. It becomes therefore
necessary to take a series of corrective actions that are commonly known as data
preprocessing. In the recent years, many algorithms have been proposed for coping for
different aspects of data bias. It is the intention of this section to concisely present the sort
of problems that data preprocessing can cope with and list the most used algorithmic
strategies that are taking place in the Metabolomic and Foodomic fields. They can be
grouped into three main groups which correspond to data horizontal alignment, data
normalization and data scaling.
As a rule of thumb, for a successful application of chemometrics to spectral data, these
should be low-rank bilinear. Bilinearity entails that the spectra signal intensities must be
solely proportional to metabolite concentration and that signal intensities are additive. In
addition, a given chemical compound is expected to give a unique spectral signature which
is composed by a series of signals (singlets and/or multiplets) that are located at precise
and well known values of chemical shift (ppm); this compound will therefore have a
common spectral fingerprint across all samples and only vary in intensity from sample to
sample.
One of the most valuable features of 1H NMR spectroscopy is that the observed signals’
chemical shifts are extremely sensitive to the local chemical environment each type of
proton is subjected to. However, this sensitivity also means that resonance frequencies are
prone to fluctuations in the samples’ pH, temperature, and external magnetic field (Wishart,
2008), causing the collected data to lose the required bi-linearity since the metabolites’
signal alignment across spectra of different samples results compromised. Thus, it is crucial
before multivariate data analysis that, unless they are part of the sought source of variance,
shift errors are removed to prevent such variation misleading the following pattern
recognition step.
For coping with such a problem, several different peak alignment methods have been
developed. While overall spectrum-to-spectrum variations due to small variations in
spectrometer frequency can be solved by a simple horizontal translation of the entire
spectra, either by using a pattern recognition method or by using an internal reference peak,
it is more difficult to handle local peak-to-peak chemical shift variations due to variations
in, for example, samples’ pH or ionic strength. Traditionally, data smoothing and/or
binning (bucketing) are a practical solution to the problem (Spraul et al., 1994), but they
imply loss of spectral resolution and uncontrolled data reduction which may lead to the
loss of some relevant latent data structure. Therefore, in recent times other more general
and refined methods have been proposed that do not implicitly require data reduction and
that, within the same spectrum, can cope with local chemical shift variations different in
intensity and direction. These include, for instance, Correlation Optimized Warping
(COW) (Nielsen et al., 1998), Recursive Segment-wise Peak Alignment (Veselkov et al.,
2009) and interval Correlation Optimized Shifting (icoshift) (Savorani et al., 2010).
In contrast to signal alignment, which represents a horizontal change to the NMR data
matrix, normalization is a vertical, row-wise change to the matrix of data, meaning that the
whole spectral data of a single row are increased or decreased in intensity by a common
coefficient that can be obtained using different approaches according to the kind of
intensity bias to be solved. Each row will therefore have its own coefficient that might or
might not be the same as that of other rows (samples). Acquired NMR metabolite profiles,
obtained from food matrices, often present systematic variations in intensity throughout
the full measured spectroscopic range which can hinder the relevant differences between
samples or make two identical samples to result different. The source of this unwanted
variation is normally related to intrinsic inhomogeneity of samples, minor differences in
sample preparation and minor fluctuations of the instrumental conditions which are
difficult to estimate and correct. Data normalization methods are typically applied to
remove unwanted systematic bias in signal intensities while preserving the interesting
metabolic information. The normalization of metabolite profiles is therefore a process of
data standardization, intended to remove this spurious between-sample variation from the
data.
Another kind of vertical correction of the intensities, this time to correct for intensity
bias among variables, is data scaling. This represents a column-wise vertical correction of
the data matrix that differently affects the different regions of the spectra and it has the
purpose of highlighting the importance of those variables that, using different criteria, are
considered to carry a relevant part of the sough information in the data. A different
coefficient is therefore calculated and applied based on the characteristics of the involved
column of variables. Column-wise scaling (a.k.a. Scaling) is considered a pre-treatment
method before multivariate date analysis because it equalizes the importance of the
different variables, in a way affecting the results that can be obtained during the data mining
and, as well as any other data preprocessing, must be performed with care and its effects
validated before acceptance. Furthermore, any normalization and scaling method carries
the risk of artificially introducing new spurious correlations to the data if not properly
chosen (Capozzi et al., 2011). It is therefore obvious that the selected method might
drastically change the results that can be obtained during the successive application of
multivariate data mining methods.
A great number of established methods for normalizing metabolite profiles has been
made available in the literature in the recent years and every method is based on certain
assumptions to the nature of the intensity problem the data can be affected from. Among
the most utilized normalization methods it is worth mentioning: (i) constant sample
intensity sum (1-norm); (ii) unit sample vector length (2-norm or Euclidian norm); (iii)
probabilistic quotient normalization (Dieterle et al., 2006) and (iv) the use of a reference
metabolite present in all samples at a constant level and whose signals (at least one) are
intense and well separated from the other signals. To this extent, also the chemical
standards TSP and DSS, often used for chemical-shift referencing for NMR spectra, may
provide quantitative reference feature if added in an accurate and precise way to all
samples, but only if no chemical reaction occurs between them and the analyzed food
matrix.
Since any normalization or scaling method may be destructive to the quantitative
information buried in the metabolomics data, it is always recommended to use an external
validation for the chosen method. This can be optimally achieved if external quantitative
knowledge from reference chemical analysis is available to the sample set. Then, a simple
multivariate regression (vide infra) to a given metabolite can evaluate whether the
normalization deteriorates or improves the correlation between its signal intensities across
spectra and its reference values.
X=TPT+E (6)
where T is the matrix which collects the coordinates of the samples onto the new set of
components (scores), which is used for visualizing the distribution of the samples, while P
is the matrix containing the coefficients relating the principal components to the
experimental variables (loadings), which provides the trait d’union for interpreting the
samples’ pattern in the scores space in terms of experimentally measured variables. Since
PCA entails an approximation of the data matrix, the extent of such approximation is
summarized by the matrix E, i.e., the residuals, which contain the difference between the
true values of the data matrix X and its PCA estimate TPT. Here, it must be stressed that,
while in general PCA is the most appropriate tool for EDA, as it projects the data onto
directions resulting in the best fit of the points in the multivariate space (i.e., which account
for the maximum explained variance), for some specific problems other projection
approaches could result more useful for the particular task. One of this is multivariate curve
resolution (MCR), which, rather than aiming at extracting abstract directions, try to
calculate components that can have a chemical interpretation, through the imposition of
tailored constraints. Another possibility, in the case where samples can be a priori
differentiated according to some criterion, i.e., classified (vide infra), is to identify
directions into the multivariate space where the separation among these different groups of
samples are maximum (as it occurs with canonical variate analysis, CVA). Furthermore, it
can happen, especially in problems related to metabolomic studies, where samples are
collected according to an underlying experimental design, where one or more factors (e.g.,
time and/or treatment) are varied in a controlled way. In such cases, the direct use of
exploratory techniques as PCA on the whole data matrix could lead to a poor interpretation
of the effects of the factors under investigation, which may appear as mixed in the bilinear
model. To cope with this issue, various approaches have recently been proposed in the
literature which allow to analyze multivariate data collected from designed experiments,
obtaining information on the significance of the effect of the individual factors and their
mutual interactions and, at the same time, a multivariate interpretation of the observed
effect through a component model. The most famous of this family of methods is ANOVA-
simultaneous component analysis (ASCA) (Smilde et al., 2005), which uses the main
concept of traditional analysis of variance to partition the experimental matrix into sub-
matrices associated to the main effect of factors and their interactions and then analyzes
each effect submatrix by PCA to achieve chemical interpretation of the impact of the factor
(or interaction) on the measured variables.
In some occasions, e.g., when the number of analyzed samples is limited, or when
reference measurements of some additional properties or characteristics of the samples are
not available to be related with the experimental data from NMR, exploratory analysis
represents not only the first but also the only data processing step that can be carried out.
However, in any other situation, the spectral profiles recorded by NMR constitute the basis
to build predictive models, i.e., models which allow not only the description of the main
features of the samples under investigation, but also to accurately predict one or more
characteristics of new samples. In particular, depending on the nature of the properties to
be predicted, two families of methods can be distinguished: calibration methods, which are
used to predict quantitative properties (e.g., peroxide number or iodine number for a lipid
moiety, protein content in cereals and so on), and classification methods, which provide
qualitative answers (as, for instance, in traceability studies, when one could be interested
in verifying if the origin of the good is compliant with what declared in the label). Since
the final aim of these methods is to be able to accurately predict the desired properties on
new (unknown) samples, in order to be applicable to a particular problem, it is mandatory
that (at least) the following two requirements are met: a sufficient number of training
samples, for which the value of the property to be predicted is known beforehand and that
span as much as possible the range of variability to be expected for future observations,
and a careful validation step, to test the reliability of the approach providing also a
quantitative estimate of the error. Validation (Westad & Marini, 2015) is indeed an
unavoidable step for any predictive model, as it guarantees the applicability of the methods
to any future sample: in its most essential form, it consists in applying the model to a new
set of samples, which are treated as unknown by the model itself but for which the “true”
values of the responses to be predicted are available to the experimenter. Accordingly, by
comparing these reference values with the ones predicted by the model it is possible to
have an estimate of how reliable is the proposed approach.
As anticipated above, one speaks of calibration problems when a predictive relation is
sought between a matrix of experimentally measured variables (in our case, the NMR
spectra or some features extracted from them) X and one or more quantitative variables to
be predicted Y (Marini, 2013). In most cases, this relation is expected to be linear (or at
least to be approximable by a linear function), so that one can express the calibration
problem as:
Y = XB + E (7)
where B is the matrix collecting the coefficients of the model and E represents the residuals.
Equation 7 is a general form and there are various approaches to calibration which differ
among one another in the way the coefficients B are calculated. The simplest linear
regression approach (multiple linear regression) is an extension of ordinary univariate least
squares but requires the data matrix X to have more observation (samples) than predictors
(variables), and that the variables are as uncorrelated as possible (characteristics that are
rarely met in the case of NMR data for which the variables are highly collinear). Indeed,
when spectral data are concerned, the most commonly used approach is the so-called
Partial Least Squares Regression (PLS-R). In PLS-R, data are first projected to a relevant
lower dimensional subspace (in a way which is analogous to PCA, but that, instead of using
maximum fit/explained variance to orient the components, identifies the new axes as the
ones along which there is maximum covariance between X and Y), and then uses these
scores as predictors for the responses.
When the answer to be predicted is of qualitative nature (i.e., discrete), then the
corresponding approaches for predictive modeling are referred to as classification methods
The increasing interest in the analysis techniques with Nuclear Magnetic Resonance
Spectroscopy (NMR) is producing a huge amount of raw data. The real value of the data
is when it can be accessed, analyzed and correlated to produce answers.
The need to make available and access this huge amount of data, is the driver of a new,
world-wide, connected network of Databases.
New skills NMR spectroscopy to identify and quantify small molecules in solution for
biomarker studies and metabolic flux 1D 1H, 13C and 2D 1H-13C spectra are emerging for
more than a thousand compounds and data. At this point the bases are providing vital
information for the NMR-based metabolomics and Chemometric quantities. This is the
reason why emerging public and private databases containing NMR data of metabolites in
standard conditions are useful to compare the chemical changes: these information
provided concern J coupling, peak multiplicity and peak intensities.
Typically, data comes from samples collected in standard conditions of temperature,
pH, and concentration of buffers such as the Human Metabolome Database (HMDB)
(Wishart et al., 2007) which were collected in H2O, 298 K and at a pH of 7.0. Also there
are additional contents, beyond the standards, as in the case of BMRB, HMDB, MQMCD
and Platform for RIKEN Metabolomics (PRIME) (Sakurai et al., 2013) that provide
connections for each metabolite pathway to external databases, such as the Kyoto
Encyclopedia of genes and genomes (KEGG) (Ogata et al., 1999; Kanehisa et al., 2006) to
track relevant metabolic pathways, or the connections HMDB metabolites to a
MetaboCard, which contains, among other things, information from the literature and from
other databases such as the average concentrations and disorders concerning. In addition
to the experimental data derived from pure standard care, the BMRB, HMDB, and
MQMCD also provide chemical shift data derived from empirical calculation. In MQMCD
we can find a Java applet that relates Jmol to its corresponding NMR spectrum (1H 1D and
only 13C) so we can select a peak in the NMR spectrum and see the corresponding atoms
shown in Jmol structure and vice versa in an interactive way. For the metabolites within a
database just do text-based searches, based on the name, chemical formula or ID numbers
as in the case of PubChem ID. The real power of these databases is the ability to identify
and quantify the metabolites automatically using spikes from experimental data. The recent
MetaboLights database (http://www.ebi.ac.uk/ metabolights/) (Drefahl, 2011) is a
comprehensive resource, cross-species and cross-technique, even working as a repository
for NMR experimental data or MS, which users can post to provide detailed information
about metabolites and concentrations, also including descriptive metadata of experiments
(Ellinger et al., 2013). Here follows a summary, taken from the Metabolomics Society web
site, (http://metabolomicssociety. org/resources/metabolomics-databases), of some of the
available databases, related to Comprehensive Metabolomic, Metabolic Pathway,
Compound or Compound-Specific, Drug, Spectral, Disease & Physiology Databases.
PRODORIC is also maintained. Several tools for the analysis of stored data and for the
visualization of the corresponding results are provided, enabling a quick understanding of
metabolic pathways, genomic arrangements or promoter structures of interest.
MetaboLights database for metabolomics experiments and derived information. The
database is cross-species, cross-technique and covers metabolite structures and their
reference spectra as well as their biological roles, locations, concentrations and
experimental data from metabolic experiments. MetaboLights offer user-submission tools
and have strong reporting capabilities. We will utilize and further develop de-facto standard
formats where various components are encapsulated, such as the encoded spectral and
chromatographic data, and associated information about the chemical structure, as well as
metadata describing assays and the study as a whole.
food products through, and are made available in FooDB in zip format, to download and
view (Wishart et al., 2007; Wishart et al., 2009).
Users can view chemical structures and molecular weights by mean of a specific
chemical structure search by FoodView functions, listing the foods or ChemView, listing
the chemicals. This is a great benefit for food companies to provide better food labels for
the nutraceutical industry, for physicians and public health policies that can better observe
CSI effects of food and address the food choices.
An attempt to create a comprehensive database including NMR spectra of different
foodstuffs and data on their origin, composition, variety etc is undertaken by the Italian
Group of Magnetic Resonance in Food Science (e-ALIERB OpenLab, 2016).
collected from existing biochemical databases plus more than 400,000 computationally
generated human phase-I and phase-II metabolites of these known compounds. The IIMDB
database features a user-friendly web interface and a programmer-friendly RESTful web
service.
both searchable and downloadable and have been carefully collected under defined
conditions on several types of GC/MS instruments (quadrupole and TOF).
Metlin, Metabolite Database is a repository for mass spectral metabolite data. All
metabolites are neutral or free acids. It is a collaborative effort between the Siuzdak and
Abagyan groups and Centre for Mass Spectrometry at The Scripps Research Institute.
METLIN is searchable by compound name, mass, formula or structure. It contains 15,000
structures, including more than 8000 di- and tripeptides. METLIN contains MS/MS,
LC/MS and FTMS data that can be searched by peak lists, mass range, biological source
or disease.
Fiehn GC-MS Database, this library contains data on 713 compounds (name, structure,
CAS ID, other links) for which GC/MS data (spectra and retention indices) have been
collected by the Fiehn laboratory. A locally maintain program called BinBase/Bellerophon
filters input GC/MS spectra and uses the spectral library to identify compounds. The actual
GC/MS library is available from several different GC/MS vendors.
BML-NMR, Birmingham Metabolite Library Nuclear Magnetic Resonance database is
a freely available resource containing 3,328 NMR spectra of 208 common metabolite
standards. This database includes both 2D 1H J-resolved spectra and 1D 1H spectra,
recorded at 500 MHz using various water suppression methods and acquisition parameters,
for solutions at pH values of 6.6, 7.0 and 7.4. The raw and processed data, including
associated metadata, are housed in a purpose-built MySQL database that is compliant with
the Metabolomics Standards Initiative (MSI) endorsed reporting requirements, with some
necessary amendments. Library data can be accessed freely and searched through a custom
written web interface. FIDs, NMR spectra and associated metadata can be downloaded
according to a newly implemented MSI-compatible XML schema.
MetaboLights, database for metabolomics experiments and derived information. The
database is cross-species, cross-technique and covers metabolite structures and their
reference spectra as well as their biological roles, locations, concentrations and
experimental data from metabolic experiments. MetaboLights offer user-submission tools
and have strong reporting capabilities. We will utilize and further develop de-facto standard
formats where various components are encapsulated, such as the encoded spectral and
chromatographic data, and associated information about the chemical structure, as well as
metadata describing assays and the study as a whole.
mzCloud features a searchable collection of high resolution/accurate mass spectral
trees using a new third generation spectra correlation algorithm. mzCloud is free and
available for public use online. mzCloud also represents an open consortium of dedicated
research and scientific groups aiming to establish a comprehensive library of high quality
spectral trees to improve the structure elucidation of unknowns. mzCloud tries to address
identification bottleneck by considering all mass spectrometric relevant aspects, looking at
number of experimental and computational details and in some cases allowing
identification of unknowns even if they are not present in library.
Low field NMR is a widely used method in food science for its ability to provide
information about the mobility and compartmentalization of water and about molecular
interactions between water and food components.
NMR systems operating at 20 MHz for 1H are the most popular and they have
dominated the low field market for the last decades. The typical 20MHz low field NMR
configuration allowed to obtain just by pressing a button a few and reliable measurements
such as total signal intensity, relaxation time distributions or diffusion measurements for
droplet size distributions. For this reason these instruments have found wide applications
in industrial environment.
Indeed the development of the first commercial dedicated pulsed Low field NMR
benchtop instrument started in 1970 within industrial collaboration between a food
company and an instrument manufacturer with the aim to set up a new method for
determination of solid content in fats (SFC). Nowadays the NMR SFC method is a well
accepted method for determination of solid content in food.
Although initially the benchtop NMR were developed for routine quality and process
control in industrial environment, the academic scientific community rapidly recognized
the potentiality of this technique for research purpose. This interest promoted the
development in electronic equipment, in hardware and software, allowing more
sophisticated low field NMR applications to a wide range of foods.
Figure 3. Schematics of standards laboratory permanent magnet configurations: a) parallel plate magnet
and b) cylindrical Halbach magnet (Mitchell et al., 2014).
Figure 4. Schematic representation of the used prototype: (a) the magnets have a size of 30x 30 mm and
enclose a sample space of the same size. The housing (grey) has a hinge at the lower left side and has a
buckle at the opposite side. (b) Scheme showing the opened device. (Windt et al., 2011).
The iron yoke is used to minimize the stray field. The amount and type of magnetic
material used in the pole pieces, plus the pole gap, determines the field strength B0 between
the poles. B0 field strength between 0.12 and 1.4T can be obtained corresponding to proton
Larmor frequencies in the range of 5-60 MHz. These field strengths gain sufficient
sensitivity for observing abundant molecular species in foods such as water and oil.
Commonly air gaps for most magnets are between 25 an 100mm allowing use for tube
diameters ranging from 10 up to 100 mm (Fukushima, 2009). Larger air gap can be
employed; one however has to compromise on a lower Larmor frequency to obtain more
easily stabilization and homogeneity. The largest air gaps are wide enough to pass an intact
fruit or conveyor belt (Chayaprasert & Stroshine, 2005). Typically the alloy used for the
pole pieces is formed from neodymium, iron, and boron (NdFeB). An alternative
permanent magnetic material used in some designs is samarium–cobalt (SmCo) alloy
which has a lower temperature coefficient compared to NdFeB. Therefore, magnet designs
incorporating both SmCo and NdFeB magnetic material can provide a magnetic field that
is less sensitive to temperature. Improvement in the B0 homogeneity may be obtained by
the inclusion of shim coils; gradient coils can also be added to enable diffusion and imaging
experiments. Solenoid rf coils are simpler to design and tune, and provide improved B1
homogeneity compared to the alternatives.
An alternative magnet design that has become popular for benchtop NMR is the
cylindrical Halbach permanent magnet (Halbach, 1980) (Figure 3b). In Halbach magnets
many small permanent magnets are arranged in the indicated way (Figure 3b) in order to
create a homogeneous magnetic field inside the cylinder. In contrast to the other designs,
the stray-field outside the magnet is rather small. The magnetic field direction is transverse
to the cylinder axis so that a simple axial solenoid can be used as rf coil. A Halbach magnet
can also be built to be very compact (Raich & Blümich, 2004). Homogeneity of the
magnetic field depends on the number bar magnets used in the Halbach magnet design.
Each magnet should have the same magnetization and size to produce homogeneous
magnetic field for NMR devices. A best compromise between homogeneity and field
strength is obtained with a high number of magnets (Dogan et al., 2009).
Open access Halbach magnet designs have also proposed for on line inspection of
intact food products (Mitchell et al., 2014; Hills et al., 2005; Blümich et al., 2014). This
open-access Halbach concept is embodied in the NMR-CUFF, used to monitor water
transport in plant stems (Windt et al., 2011). A schematic representation of NMR CUFF is
shown in Figure 4a and 4b.
Halbach magnet systems of the NMR-CUFF design are ideally suited for samples
which as a whole are too large to fit into a normal NMR magnet, but in which the section
of interest is small or slender. Examples of such objects are non-metallic tubes in
experimental or industrial setups, extremities and necks in animals or humans, or stems
and branches in plants. Halbach magnet systems have not reached commercial
implementation yet.
The primary advantages of Halbach magnets are the reduction of the magnet volume
and weight by three orders of magnitude and a large bore/magnet size ratio (Danieli et al.,
2010). Thus, Halbach magnets can be very compact, light, highly homogeneous, and safe
and thus make a great option for industrial environments. Halbach magnets with small and
large bores have been built for medium- and low-resolution applications. In 2010 Danieli
et al., built a 0.5-kg SmCo Halbach magnet with approximately 0.7 T (30 MHz for 1H) for
medium-resolution NMR (∼0.2 ppm of the resonance frequency) (Danieli et al., 2010). The
magnet is a cylinder with 80 mm length, 35 mm diameter, and 15 mm bore which allows
the use of standard 5-mm NMR tubes (Osán et al., 2011). Figure 5 presents a wide-bore
(10 cm) low-resolution NMR Halbach 0.23-T (∼8.7 MHz for 1H) magnet for applications
in oil and food industry (Colnago et al., 2014).
Figure 5. Halbach magnet for flow analysis with pre-polarizer magnet (Colnago et al., 2014).
to consider the relaxation times T1 and T2 of the nuclei under study. In particular, the PFG
STE pulse sequence should be used when T1 >> T2 for the investigated system. The
motivation to apply this method in food science arises from the fact that the observed signal
attenuation is proportional to the self-diffusion constant (D) and three adjustable NMR
parameters: the amplitude (G), the duration () and the time interval between the de- and
refocusing pulses () of the magnetic field gradients. The self-diffusion constant D can be
derived from the exponential attenuation of signal in a field gradient as a function of the
expression G22 ( – /3).
Figure 6. Comparison of T1–T2 distribution functions (left) and D–T2 distribution functions (right)
measured on four different dairy products: skim milk, heavy cream. The dashed lines in the T 1–T2
distribution functions indicate T1=T2, whereas in the D–T2 distribution functions, they indicate the
diffusion coefficient of water. Contour lines are shown at 10, 30, 50, 70, and 90% of maximum values
in each panel. For the samples of heavy cream and Brie, we show in addition the 5% line.
The advent of a fast algorithm for two dimensional Laplace inversion has opened the
way to a revolution in multidimensional relaxometry and diffusometry (Hurlimann et al.,
2006; Song, 2006).
Correlation of diffusion and T2 has been achieved by combining common
diffusometric approaches, such as PFGSE or PFGSTE, with a CPMG sequence.
Subsequently, D–T2 correlations can be obtained using different data analysis approaches.
Discrete data analysis approaches (e.g., NNLS, SPLMOD) have been useful to distinguish
the diffusion coefficient values of water fractions associated with different compartments,
(Van Dusschoten et al., 1995) as well as surface/volume ratios of cells (Raffo et al., 2005).
Recently the inverse Laplace method was also applied, obtaining faster processing. As an
example in Figure 6 are shown T1–T2 and D–T2 distribution functions measured for milk
and heavy cream (Hurlimann et al., 2006).
The skim milk sample exhibits a single narrow signal in both T 1-T2 and D-T2
distribution functions, indicating that the only detected signal is that of water; the protein
signal decays too quickly to appear in the 2D map. The measured diffusion coefficient is
close to the molecular diffusion coefficient of water, shown as dashed line in the D–T2 plot,
implying that water diffusion is much less affected by the minor components, in this case
protein. On the other hand, both distributions of the cream sample show two components.
The upper component is still relatively sharp and its diffusion coefficient is close to that of
water. The lower component has a T2 distribution that is about 1.5 decades wide and
appears at a much lower diffusion coefficient. This component is associated with liquid fat.
Signals from solid fat and proteins have relaxation times less than 100 ms and could not be
detected with the present experiments.
Such results indicate both T1-T2 and D-T2 are reliable suited methods to study
heterogeneous systems including emulsion, biological and food samples.
Figure 8. NMR signal vs depth of the two probeheads. The maximum intensity of the signal is obtained
at about 0.1 mm for the surface probehead (left), at about 3 mm for the 3 mm probehead (right).
where γ is the gyromagnetic ratio. In this case the echo amplitude does not decay in a
simple exponential manner. Because of the τ3 dependence, the effect of diffusion is
particularly pronounced for large values of τ that are used to measure long T 2 relaxation
times. Thus, it is preferable to use CPMG pulse sequence with a small τ to reduce the
diffusion attenuation of the spin echo. In the case of most conventional NMR magnets, the
field inhomogenety is sufficiently small that at short τ values the second term of the
equation becomes negligible even for fast diffusing molecules such as water. Instead, in
the case of single-sided NMR the magnetic field gradient is usually so strong that the
second term of equation (8) cannot be neglected. As a consequence, T2 decay measured in
a magnetic field that usually has a large field gradient result always much more short than
those measured in a homogeneous one.
The magnet described above has a strong gradient along the depth direction which can
be used to obtain spatial localization into the object simply by changing the excitation
frequency. Besides, this magnet offers a typical resolution of about a millimeter one could
hardly talk about depth profiling but only about measurements at some depths. A depth
profiling can be obtained using a purposely built single-sided NMR sensor by RWTH
Aachen University, Aachen, Germany (Blumich et al., 2008). This sensor generates a
magnetic field with an extremely uniform gradient to resolve near surface structure of
arbitrary large samples. For depth profiling the sample remains on top of the sensor and
the profile is acquired by exciting the whole region of interest in a single experiment
(Figure 9). The magnet geometry consists of four permanent magnet blocks positioned on
an iron yoke (Figure 9). Two magnets are polarized along y and two along -y. Magnets
with the same polarization are separated by a small gap ds while magnets with opposite
polarization are separated by a gap dB. The sensor is placed on a lift which moves the
sensitive volume constituted of a thin slice at a well defined position through the object by
varying the distance between the sensor and the object. The sensitive slice is 25x25 mm2
and the thickness of the sensitive volume can be adjusted from a few to 150 µm and
depends on the setting of the measurement parameters. The amplitude of the signal from
the sensitive slice is proportional to the spin population in that slice.
The sensitive volume is a thin slice above the device, at a distance of 1 cm, parallel to
its surface.
The amplitude values as a function of the sensor displacement produce a mono-
dimensional image of spin population as a function of the depth into the object. However,
depending on the way one wants to define the contrast, not only the signal amplitude but
also other parameters such as for example relaxation times can be reported against the
depth scanned. Presently, penetration depths up to 2.5 cm may be achieved with these
sensors (Perlo et al., 2005). With respect to the previous portable NMR probes, in the
single-sided NMR the presence of an extremely uniform gradient (20 T/m) allows the
measurements of auto-diffusion coefficients. The pulse sequence used is based on the
generation of Hahn (SE) (Hahn, 1950; Stejskal & Tanner, 1965) and stimulated echoes
(STE) (Tanner, 1970), both operating in the presence of a steady gradient. To improve the
sensitivity of these experiments, a CPMG sequence is always applied after the diffusion-
editing period to generate an echo train. By the sensitivity improvement achieved by adding
the echo train, complete diffusion curve can be obtained in short times. Beside, large
gradients simplify measurements of the diffusion coefficient in heterogeneous materials
such as porous material and biological systems, since it reduces the contributions of the
local magnetic fields due to susceptibility variation within the sample. The use of a strong
static gradient allows measurement of root-mean square molecular displacement as small
as 20 nm and self-diffusions until 10-16 m2/s (Casanova et al., 2011).
Magnetic Resonance Imaging (MRI) represents a rapid, reliable, robust and non-
invasive NMR technique enabling the acquisition of detailed images of whole tissues and
revealing different and complementary information on their molecular composition and
dynamics. Up to date, such an advanced technique has been prevalently exploited to
investigate on mammalians in the context of diagnostic medicine and pharmaceutical
applications. Only in the last two decades, MRI applications have been successfully
extended to other research fields, including Food Science. As proved by a large body of
literature, MRI technique has the potential to assess food quality, characterize agrofood
products, optimize food process design and understand the mechanisms of both heat and
mass transport in food (Hills, 1998; Martinez et al., 2003; Butz et al., 2005). This technique
is very useful not only because it permits to visualize and evaluate the static structure and
the texture of food products but also because it allows to follow, non-invasively and in real
time, the dynamic changes occurring in food as well as to ascertain key factors such as
distribution and diffusivity of water and fats in food compartments (Kirtil et al., 2016).
The principles on which MRI technique is based have been extensively discussed
earlier (Hills, 1998; Clark et al., 1997; Patel et al., 2015). Briefly, the nuclear magnetic
resonance phenomenon occurs when nuclei of magnetically active atoms (spin number ≠
0) are immersed in a static magnetic field and exposed to a second oscillating magnetic
field. Then, following the pulse interaction between the applied electromagnetic radiation
and the dipolar moments of the nuclei subjected to the static field, the resonance energy is
relaxed thus producing a Free Induction Decay (FID) that contains detailed information
about the structure, dynamics and chemical environment of the molecular material under
study (Levitt, 2008). In MRI, differently from other NMR techniques, the excited spins are
subjected to a spatial encoding (phase and frequency encoding) that is substantially
conducted by applying a linear variation of the local magnetic field through a set of three
orthogonal gradient coils. This step is critical since it enables the modulation of spins
precession frequencies as a function of their position in the sample. The resulting matrix
forms a time-domain representation of an image in the so-called “k-space” in which the
horizontal (x) and vertical (y) axes are commonly associated to frequency- and phase-
encoding, respectively, and whose reconstruction, through 2D Fourier Transform, leads to
the MRI image (Hills, 1998; Canela et al., 2016).
Depending on the chosen pulse sequence, it is possible to carry out different MRI
experiments which enable the acquisition of a number of complementary information and
are potentially helpful in circumventing problems such as the improvement of images
quality or the shortening of the experiment duration in order to minimize the time-
dependent food perishability. The most common MRI pulse sequences, which are
substantially based on spin-echo or gradient-echo families, have been clearly summarized
by Bitar et al. (Bitar et al., 2006).
Each 2D MRI image results from the analysis of a specific sample area, referred to as
slice, which is selectively excited by applying directional gradients. The set of slices to be
acquired represents the experiment geometry which is conventionally defined as axial
(transverse plane), coronal (frontal and longitudinal plane) or sagittal (lateral and
longitudinal plane) as a function of slices orientation in respect to the sample placement in
the magnet. According to the selected geometry, it is possible to focus on a restricted area
of food sample and accurately evaluate its anatomical peculiarities and chemical-physical
properties. It is remarkable that the MRI technique also offers the opportunity to produce
3D images by applying a three-dimensional reconstruction of a set of 2D slices acquired
along a third axis orthogonal to the plane of slices. Moreover, 3D MRI enables the
advantage to precisely quantify the volume of internal compartments in whole food
samples (Van Dusschoten et al., 2016).
Most of MRI applications on food focus on the detection of the 1H nucleus (the
“proton”), although additional probe inserts and protocols for other magnetically active
nuclei such as 13C, 31P, 19F, 2H and 23Na are also available. The water proton signal is
commonly preferred because of the benefits rising from its great naturally occurring
abundance in agrofood products, such as fresh fruits, cereals and legumes. However, it is
also possible to optimize the MRI experiments in order to focus on other 1H frequencies
typical of food substrates such as carbohydrates or fats.
A most peculiar advantage of MRI is the fact that several MRI pulse sequences can be
applied to modulate contrast factors (weighting) as a function of intrinsic molecular
properties, such as nuclear relaxation times (spin-lattice T1, spin-spin T2 and T2*) and
diffusivity. In particular, NMR relaxation is the process by which the total energy of a
nuclear system, acquired after its excitation by a radio frequency pulse, is progressively
lost and the equilibrium state is recovered. Given the absence of interfering factors, the
relaxation is related to structural variations, the overall molecular mobility, or the
interactions undergone by the nucleus. Conversely, diffusion coefficients may evaluate the
apparent translational self-diffusion of molecules (i.e., the water) through food
compartments, thus allowing to estimate the extent of free diffusion distance, rigidity of
food products or time-dependent variations in the internal texture, and permeability. On
this basis, the appropriate set up of both relaxation- and diffusion-based MRI experiments
may usefully provide weighted images that lack suppressed areas due to fast relaxation or
high diffusion coefficient, thus improving the visualization of hidden or overlapped sites.
In Figure 10 is shown a representative example of the application of a T2-contrast to isolate
specific tissues in a kiwi fruit, such as the fruit pulp, and enhance the detection of site-
specific details.
Figure 10. Spin density (left-hand) and T2-weighted (right-hand) axial MR images of a kiwifruit
(Actinidia chinensis). The MR images refer to the fruit region included in the orange rectangle of the
central reference image. The application of a spin-spin modulation (total spin-echo delay of 150 ms)
permitted to suppress the columella tissue and highlight the fruit pulp (Unpublished data).
It is also important to consider that such MRI pulse sequences not only produce
contrasted images, but also enable the reliable site-specific quantification of relaxation
times and diffusion coefficients, thus providing additional molecular information. For
instance, in Figure 11 are shown two axial MR images of both stem and primary root of a
young maize plant revealing, through a parametric image, the spatial variability of
diffusivity as a function of plant tissue.
Figure 11. Spin density (A) and diffusion-based parametric (B) axial MR images of both stem and
primary root of a young maize plant (Zea mays). The colors associated to the latter image are
modulated as a function of diffusivity and permit to identify its spatial variability within maize plant
tissues (Unpublished data).
A further method that enhances image quality and improves MRI response in food
products consists in the selective suppression of either images areas (through saturated
slices) or very intense signals due to water or lipids. For example, the comparison of images
acquired with or without lipid suppression may indirectly allow the quantification of fatty
components, while the suppression of water signals from samples intrinsically rich of water
may improve the image quality due to the increased visibility of signals of lower intensity.
Finally, several MRI experiments have been conceived to acquire imaging-related
spectra that distinguish aromatic, carboxyl, hydroxyalkyl and alkyl groups in specific sites
of samples. The most used experiments are the Localized Spectroscopy (LS) and the
Chemical Shift Imaging (CSI) (Melkus et al., 2009). The former produces a single
spectrum resulting from the NMR measurement of a sample whole voxel, while the latter
provides a set of spectra (whose number depends on the matrix dimension) that indicate
distribution of molecular components in different locations of the image. Such techniques
enable the spatial identification of both selected functional groups and, in some cases, the
specific markers of a transformation process. However, the required spectral resolution for
these experiments may be strongly hampered by typical problems related to solid-state
NMR (strong dipolar interactions and chemical shift anisotropy), as well as by insufficient
correction of local dishomogeneities due to extremely complex texture of samples.
The density and distribution of mobile protons, as well as their mobility extent,
contribute to determine the quality and number of details in 1H MR images. Therefore,
MRI experiments based on water peak frequency are potentially very advantageous when
applied to fruit and vegetables, since these agro-food products are generally characterized
by a relatively large abundance of mobile and homogenously distributed water. In addition,
the resulting high signal/noise ratio permits to reduce the analysis time and concomitantly
offers the opportunity to reveal many anatomical and morphological details. In the context
of plant science, it has been proved that MR images of fresh vegetables may lead to a
defined detection of skin, vascular tissue, cortex, seed-bed and seeds, which can be even
improved by applying contrast factors. With this aim, one of the most exploited MRI pulse
sequence consists in Multi-Slice-Multi-Echo (MSME) which produces detailed T2-
weighted MR images without any gradient-related artifact. In particular, the site-specific
signal intensity in T2-weighted MR images is proportional to the water content and
modulated as a function of the transverse relaxation time T 2. Since the latter is related to
the correlation time of water (and thus to its mobility), the relatively dark and light areas
of T2-weighted MR images generally refer to tissues containing water molecules
characterized by low and fast mobility, respectively. Such a site-specific difference is most
likely due to the interaction of water with cellular tissues and their components, in terms
of bond strength (hydrogen bonding and solvation), as well as to the contribution of other
factors, such as viscosity or naturally occurring paramagnetic metals.
ACKNOWLEDGMENTS
This work has been carried out within the project of Regione Lazio Lr 13/2008 entitled
“e-ALIERB: un OPEN LAB per caratterizzare e valorizzare i prodotti alimentari ed
erboristici del territorio laziale”.
REFERENCES
Andersen, M.-B. S., Rinnan, Å., Manach, C., Poulsen, S. K., Pujos-Guillot, E., Larsen, T.
M., Astrup, A., & Dragsted, L. O. (2014). Untargeted metabolomics as a screening tool
for estimating compliance to a dietary pattern. Journal of Proteome Research, 13,
1405-1418.
Andrew, E. R., Bradbury, A., & Eades, R. G. (1959). Removal of dipolar broadening of
nuclear magnetic resonance spectra of solids by specimen rotation. Nature, 183, 1802-
1803.
Azeredo, R. B. D., Colnago, L. A., & Engelsberg, M. (2000). Quantitative analysis using
steady-state free precession nuclear magnetic resonance. Analytical Chemistry, 72(11),
2401-2405.
Azeredo, R. B. V., Colnago, L. A., Souza, A. A., & Engelsberg, M. (2003). Continuous
wave free precession: practical analytical tool for low-resolution nuclear magnetic
resonance measurements. Analytica Chimica Acta, 478(2), 313-320.
Baroni, S. (2009). Magnetic Resonance in Food Science. Challenges in a Changing World.
Cambridge, UK: Royal Society of Chemistry.
Bharti, S. K., & Roy, R. (2012). Quantitative 1H-NMR spectroscopy. Trends in Analytical
Chemistry, 35, 5-26.
Bitar, R., Leung, G., Perng, R., Tadros, S., Moody, A. R., Sarrazin, J., McGregor, C.,
Christakis, M., Symons, S., Nelson, A., & Roberts, T. P. (2006). MR pulse sequences:
what every radiologist wants to know but is afraid to ask. Radiographics, 26(2), 513-
537.
Blümich, B., Blümler, P., Eidmann, G., Guthausen, A., Haken, R., Schmitz, U., Saito, K.,
& Zimmer, G. (1998). The NMR-mouse: construction, excitation, and applications.
Magnetic Resonance Imaging, 16, 479-484.
Blümich, B., Anferova, S., Sharma, S., Segre, A. L., & Federici C. (2003a). Degradation
of historical paper: nondestructive analysis by the NMR-MOUSE. Journal of Magnetic
Resonance, 161, 204-209.
Blümich, B., Anferova, S., Kremer, K., Sharma, S., Herrmann, V., & Segre, A. (2003b).
Unilateral NMR for quality control: the NMR-MOUSE. Spectroscopy, 18, 18-34.
Blümich, B., Perlo, J., & Casanova, F. (2008). Mobile single-sided NMR. Progress in
Nuclear Magnetic Resonance Spectroscopy, 52, 197-269.
Blümich, B., Casanova, F., & Appelt, S. (2014). NMR at low magnetic fields. Progress in
Nuclear Magnetic Resonance Spectroscopy, 76, 1-60.
Bondia-Pons, I., Cañellas, N., Abete, I., Rodríguez, M. Á., Perez-Cornago, A., Navas-
Carretero, S., & Martínez, J. A. (2013). Nutri-metabolomics: subtle serum metabolic
differences in healthy subjects by NMR-based metabolomics after a short-term
nutritional intervention with two tomato sauces. Omics: A Journal of Integrative
Biology, 17(12), 611-618.
Bordoni, A., & Capozzi, F. (2014). Foodomics for healthy nutrition. Current Opinion in
Clinical Nutrition & Metabolic Care, 17(5), 418-424.
Bordoni, A., & Capozzi F. (2015). The foodomics approach for discovering biomarkers of
food consumption in nutrition studies. Current Opinion in Food Science, 4, 124-128.
Cacciatore, S., Luchinat, C., & Tenori, L. (2014). Knowledge discovery by accuracy
maximization. Proceedings of the National Academy of Science, 111, 5117-5122.
Butz, P., Hofmann, C., & Tauscher, B. (2005). Recent developments in noninvasive
techniques for fresh fruit and vegetable internal quality analysis. Journal of Food
Science, 70(9), R131-141.
Canela, N., Rodrıguez, M. A., Baiges, I., Nadal, P., Arola, L. (2016). Foodomics imaging
by mass spectrometry and magnetic resonance. Electrophoresis, 37, 1748-1767.
Capozzi, F., Ciampa, A., Picone, G., Placucci, G., & Savorani, F. (2011). Normalization is
a necessary step in NMR data processing: finding the right scaling factors. In J.-P.
Renou, P. S. Belton, & G. A. Webb (Eds.), Magnetic Resonance in Food Science: An
Exciting Future (First edition, pp. 147-60). Cambridge (UK): The Royal Society of
Chemistry.
Capozzi, F., & Trimigno, A. (2015). Using metabolomics to describe food in detail. In J.-
L. Sébédio, & L. Brennan (Eds.), Metabolomics as a tool in nutrition research (First
edition, pp. 203-229). Cambridge, (UK): Elsevier.
Casanova, F., Perlo, J., & Blümich, B. (2011). Single-Sided NMR (First edition). Verlag
Berlin Heidelberg, Germany: Springer.
Chayaprasert, W., & Stroshine, R. (2005). Rapid sensing of internal browning in whole
apples using a low-cost, low-field proton magnetic resonance sensor. Postharvest
biology and technology, 36(3), 291-301.
Chen, J. H., & Singer, S. (2007). High-resolution magic angle spinning NMR spectroscopy.
In J. C. Lindon, J. K. Nicholson, & E. Holmes (Eds.), The handbook of metabonomics
and metabolomics (First edition, pp. 113-147). New York, USA: Elsevier.
Clark, C. J., Hockings, P. D., Joyce, D. C., & Mazucco, R. A. (1997). Application of
magnetic resonance imaging to pre- and post-harvest studies of fruits and vegetables.
Postharvest Biology and Technology, 11, l-21.
Claus, S. P., & Swann, J. R. (2013). Nutrimetabonomics: applications for nutritional
sciences, with specific reference to gut microbial interactions. Annual review of food
science and technology, 4, 381-399.
Colnago, L. A., Azeredo, R. B. V., Marchi Netto, A., Andrade, F. D., & Venâncio, T.
(2011). Rapid analyses of oil and fat content in agri-food products using continuous
wave free precession time domain NMR. Magnetic Resonance in Chemistry, 49(SI),
S113-S120.
Colnago, L. A., Andrade, F. D., Souza, A. A., Azeredo, R. B. V., Lima, A. A., Cerioni, L.
M., Osán, T. M., & Pusiol, D. J. (2014). Why is Inline NMR rarely used as industrial
sensor? Challenges and opportunities. Chemical Engineering & Technology, 37, 191-
203.
Corsaro, C., & Mallamace, D. (2011). A Nuclear Magnetic Resonance study of the
reversible denaturation of hydrated lysozyme. Physica A, 390, 2904-2908.
Corsaro, C., Mallamace, D., Vasi, S., Ferrantelli, V., Dugo, G., & Cicero, N. (2015). 1H
HR-MAS NMR spectroscopy and the metabolite determination of typical foods in
mediterranean diet. Journal of Analytical Methods in Chemistry, 2015, 175696 (14
pp.).
Danieli, E., Perlo, J., Blümich, B., & Casanova, F. (2010). Angewandte Chemie
International Edition, 49 (24), 4133-4135.
Dieterle, F., Ross, A., Schlotterbeck, G., & Senn, H. (2006). Probabilistic quotient normal-
ization as robust method to account for dilution of complex biological mixtures.
Application in 1H-NMR metabonomics. Analytical Chemistry, 78(13), 4281-4290.
Doğan, N., Topkaya, R., Subaşı, H., Yerli, Y., & Rameev B. (2009). Development of
Halbach magnet for portable NMR device. Journal of Physics: Conference Series, 153,
012047 (3 pp.).
Drefahl, A. (2011). CurlySMILES: a chemical language to customize and annotate
encodings of molecular and nanodevice structures. Journal of Cheminformatics, 3,
1-7.
Dumas, M. E., Maibaum, E. C., Teague, C., Ueshima, H., Zhou, B., Lindon, J. C.,
Nicholson, J. K., Stamler, J., Elliott, P., Chan, Q., & Holmes, E. (2006). Assessment
of analytical reproducibility of 1H NMR spectroscopy based metabonomics for large-
scale epidemiological research: the INTERMAP study. Analytical Chemistry, 78,
2199-2208.
Dunn, K. J., Bergman, D.J., & Latorraca, G.A. (2002). Handbook of geophysical
exploration seismic exploration: Nuclear Magnetic Resonance petrophysical and
logging application (1°edition). Oxford, UK: Elsevier Science.
e-ALIERB OpenLab. (2016). “e-ALIERB OpenLab”. Last modified 2016. http://www.e-
alierb.it/nmr
Eidmann, G., Savelsberg, R., Blümer, P., & Blümich, B. (1996). The NMR-MOUSE, a
mobile universal surface explorer. Journal of Magnetic Resonance, Series A, 122, 104-
109.
Ellinger, J. J., Chylla, R. A., Ulrich, E. L., Markley, J. L. (2013). Databases and Software
for NMR-Based Metabolomics. Current Metabolomics, 1, 22 pp.
Ellis, L. B., Speedie, S. M., & McLeish, R. (1998). Representing metabolic pathway
information: an object-oriented approach. Bioinformatics, 14, 803-806.
Emwas, A.-H. M. (2015). The strengths and weaknesses of NMR spectroscopy and mass
spectrometry with particular focus on metabolomics research. Methods in Molecular
Biology, 1277, 161-193.
English, A. E., Whittall, K. P., Joy M. L. G., & Henkelman, R. M. (1991). Quantitative
two-dimensional time correlation relaxometry. Magnetic Resonance in Medicine, 22,
425-434.
Everett, J. R. (2015). A new paradigm for known metabolite identification in
metabonomics/metabolomics: metabolite identification efficiency. Computational and
Structural Biotechnology Journal, 13, 131-144.
Fiehn, O. (2002). Metabolomics - the link between genotypes and phenotypes. Plant
Molecular Biology, 48, 155-171.
Fiehn, O., & Kind, T. (2007). Metabolite profiling in blood plasma. Methods in Molecular
Biology, 358, 3-17.
Fitch, W. L., Detre, G., & Holmes C. P. (1994). High-resolution 1H-NMR in solid-phase
organic synthesis. Journal of Organic Chemistry, 59, 7955-7956.
Fan, T. W.-M. (1996). Metabolite profiling by one- and two-dimensional NMR analysis of
complex mixtures. Progress in Nuclear Magnetic Resonance Spectroscopy, 28, 161-
219.
Fukushima, E. (2009). Magnetic resonance microscopy: spatially resolved NMR
techniques and applications, (First edition). Weinheim, FRG: Wiley.
Garcia-Aloy, M., Llorach, R., Urpi-Sarda, M., Tulipani, S., Salas-Salvadó, J., Martínez-
González, M. A., Corella, D., Fitó, M., Estruch, R., Serra-Majem, L., & Andres-
Lacueva, C. (2015). Nutrimetabolomics fingerprinting to identify biomarkers of bread
exposure in a free-living population from the PREDIMED study cohort.
Metabolomics, 11(1), 155-165.
Gallo, V., Intini, N., Mastrorilli, P., Latronico, M., Scapicchio, P., Triggiani, M.,
Bevilacqua, V., Fanizzi, P., Acquotti, D., Airoldi, C., Arnesano, F., Assfalg, M.,
Benevelli, F., Bertelli, D., Cagliani, L. R., Casadei, L., Cesare Marincola, F.,
Colafemmina, G., Consonni, R., Cosentino, C., Davalli, S., De Pascali, S. A., D’Aiuto,
V., Faccini, A., Gobetto, R., Lamanna, R., Liguori, F., Longobardi, F., Mallamace, D.,
Mazzei, P., Menegazzo, I., Milone, S., Mucci, A., Napoli, C., Pertinhez, T., Rizzuti,
A., Rocchigiani, L., Schievano, E., Sciubba, F., Sobolev, A., Tenori, L., & Valerio, M.
(2015a). Performance assessment in fingerprinting and multi component quantitative
NMR analyses. Analytical Chemistry, 87, 6709-6717.
Gallo, V., Intini, N., Mastrorilli, P., Latronico, M., Scapicchio, P., Cremonini, M. A.,
Benevelli, F., Milella, A., Rizzuti, A. & Cafagna, I. (2015b). Validation of a 1D 1H-
NOESY experiment for fingerprinting of wheat and flour: NMR Inter-Laboratory
Comparisons Series, Vol. 1 (NeP Edizioni Srls). Rome, IT.
Gallo, V., Intini, N., Mastrorilli, P., Latronico, M., Scapicchio, P., Milella, A., Rizzuti, A.,
Todisco, S., Pascazio, A., Triggiani, M., Bevilacqua, V., Cobas, C., & Mari, S. (2016).
Validation of NMR fingerprinting methods: effects of processing on measure
reproducibility and laboratory performance assessment: NMR Inter-Laboratory
Comparisons Series, 2 (NeP Edizioni Srls). Rome, IT.
Giuliani, A., Capuani, G., Miccheli, A., Aureli, T., Ramacc, M. T., Conti, F. (1991).
Multivariate data analysis in biochemistry: a new integrative approach to metabolic
control in brain aging. Cellular and Molecular Biology, 37(6), 631-638.
Giuliani, A., Capuani, G., Aureli, T., Miccheli, A., Manetti, C., Conti, F. (1993).
Multivariate data analysis as applied to NMR results: a window on biological
complexity. Journal of Magnetic Resonance in Medicine and Biology, Vol. 1 (CIC
Edizioni Internazionali, pp. 5-12). Rome, IT.
Giuliani, A., Zbilut, J. P., Conti, F., Manetti, C., Miccheli, A. (2004). Invariant features of
metabolic networks: a data analysis application on scaling properties of biochemical
pathways. Physica A, 337, 157-170.
Gonzalez-Beltran, A., Maguire, E., Sansone, S. A., & Rocca-Serra, P. (2014). LinkedISA:
semantic representation of ISA-Tab experimental metadata. Bmc Bioinformatics, 15,
1-15.
Goodacre, R., Broadhurst, D., Smilde, A. K., Kristal, B. S., Baker, J. D., Beger, R., Bessant,
C., Connor, S., Capuani, G., Craig, A., Ebbels, T., Kell, D. B., Manetti, C., Newton, J.,
Paternostro, G., Somorjai, R., Sjostrom, M., Trygg, J., Wulfert. F. (2007). Proposed
minimum reporting standards for data analysis in metabolomics. Metabolomics, 3,
231-241.
Griffiths, L., & Irving, A. M. (1998). Assay by nuclear magnetic resonance spectroscopy:
quantification limits. Analyst, 123, 1061-1068.
Hahn, E. L. (1950). Spin Echoes. Physical Review, 80(4), 580-594.
Halbach, K. (1980). Design of permanent multipole magnets with oriented rare earth cobalt
material, Nuclear Instruments and Methods, 169, 1-10.
Hills, B., (1998). Magnetic resonance imaging in food science (First edition). New York,
USA: John Wiley & Sons.
Hills, B. P., Wright, K. M., & Gillies, D.G. (2005). A low-field, low-cost Halbach magnet
array for open-access NMR. Journal of Magnetic Resonance, 175, 336-339.
Hurlimann, M. D., Burcaw L., & Song, Y. Q. (2006). Quantitative characterization of food
products by two-dimensional D-T2 and T1-T2 distribution functions in a static
gradient. Journal of Colloid and Interface Science, 297, 303-311.
IUPAC Norm Version 2150, Solid content determination in fats by NMR – Low resolution
nuclear magnetic resonance, Blackwell Scientific Publication, Oxford 1987.
Jackson, J. A., Burnett, L., & Harmon, J. F. (1969). Remote (inside-out) NMR. III.
Detection of nuclear magnetic resonance in a remotely produced region of
homogeneous magnetic field. Journal of Magnetic Resonance, 41(3), 411-421.
Kanehisa, M., Goto, S., Hattori, M., Aoki-Kinoshita, K. F., Itoh, M., Kawashima, S.,
Katayama, T., Araki, M., & Hirakawa, M. (2006). From genomics to chemical
genomics: new developments in KEGG. Nucleic Acids Research, 34, D354-D357.
Khakimov, B., Gurdeniz, G., & Engelsen, S. B (2015). Trends in the application of
chemometrics to foodomics studies. Acta Alimentaria, 44(1), 4-31.
LeMieux, M. J., Aljawadi, A., & Moustaid-Moussa, N. (2014). Nutrimetabolomics.
Advances in Nutrition: An International Review Journal, 5(6),792-794.
Lenz, E. M., Bright, J., Wilson, I. D., Hughes, A., Morrisson, J., Lindberg, H., & Lockton,
A. (2004) Metabonomics, dietary influences and cultural differences: a 1H NMR-based
study of urine samples obtained from healthy British and Swedish subjects. Journal of
Pharmaceutical and Biomedical Analysis, 36, 841-849.
Levitt, M. (2008). Spin Dynamics: Basics of Nuclear Magnetic Resonance (Second
edition). Chichester, UK: Wiley.
Li, J., Brazhnik, O., Kamal, A., Guo, D., Lee, C., Hoops, S., & Mendes, P. (2003).
Databases and visualization for metabolomics. In G. H., & R. G. (Eds.), Metabolic
profiling - its role in biomarker discovery and gene function analysis (edition 1, pp.
293–309). New York, USA: Springer.
Lindon, J. C., Holmes, E., & Nicholson, J. K. (2007). Metabonomics in pharmaceutical R
& D. FEBS Journal, 274, 1140-1151.
Lindon, J. C., Beckonert, O. P., Holmes, E., & Nicholson, J. K. (2009). High-resolution
magic angle spinning NMR spectroscopy: application to biomedical studies. Progress
in Nuclear Magnetic Resonance Spectroscopy, 55, 79-100.
Llorach, R., Garcia-Aloy, M., Tulipani, S., Vazquez-Fresno, R., & Andres-Lacueva, C.
(2012). Nutrimetabolomic strategies to develop new biomarkers of intake and health
effects. Journal of Agricultural and Food Chemistry, 60(36), 8797-8808.
Lloyd, A. J., Beckmann, M., Favé, G., Mathers, J. C., & Draper, J. (2011). Proline betaine
and its biotransformation products in fasting urine samples are potential biomarkers of
habitual citrus fruit consumption. British Journal of Nutrition, 106, 812-824.
Malz, F., & Jancke, H. (2005). Validation of quantitative NMR. Journal of Pharmaceutical
and Biomedical Analysis, 38, 813-823.
Mannina, L., Sobolev, A. P. & Viel, S. (2012). Liquid state 1H high field NMR in food
analysis. Progress in Nuclear Magnetic Resonance Spectroscopy, 66, 1-39.
Marini F. (2013), Chemometrics in Food Chemistry (1°edition). Oxford, UK: Elsevier.
Martinez, I., Aursand, M., Erikson, U., Singstad, T. E., Veliyulin, E., & van der Zwaag, C.
(2003). Destructive and non-destructive analytical techniques for authentication and
composition analyses of foodstuffs. Trends in Food Science & Technology, 14, 489-
498.
McKenzie, J. S., Donarski, J. A., Wilson, J. C., & Charlton, A. J. (2011). Analysis of
complex mixtures using high-resolution nuclear magnetic resonance spectroscopy and
chemometrics. Progress in Nuclear Magnetic Resonance Spectroscopy, 59, 336-359.
Melkus, G., Rolletschek, H., Radchuk, R., Fuchs, J., Rutten, T., Wobus, U., Altmann, T.,
Jakob, P., & Borisjuk, L. (2009). The metabolic role of legume endosperm: a non-
invasive imaging study. Plant Physiology, 151, 1139-1154.
Mètis, A., & Mariette, F. (2003). Determination of water self-diffusion coefficient in
complex food products by low field 1 H PFG-NMR: comparison between the standard
spin-echo sequence and the T 1-weighted spin-echo sequence. Journal of Magnetic
Resonance, 165, 265-275.
Mitchell, J., Gladden, L. F., Chandrasekera, T. C., & Fordham, E. J. (2014). Low-field
permanent magnets for industrial process and quality control. Progress in Nuclear
Magnetic Resonance Spectroscopy, 76, 1-60.
Nicholson, J. K., & Wilson, I. D. (1989). High-resolution proton magnetic resonance
spectroscopy of biological fluids. Progress in Nuclear Magnetic Resonance
Spectroscopy, 21, 449-451.
Nicholson, J. K., Lindon, J. C., & Holmes, E. (1999). Metabonomics: understanding the
metabolic responses of living systems to pathophysiological stimuli via multivariate
statistical analysis of biological NMR spectroscopic data. Xenobiotica, 29, 1181-1189.
Nielsen, N. P. V., Carstensen, J. M., & Smedsgaard, J. (1998). Aligning of single and
multiple wavelength chromatographic profiles for chemometric data analysis using
correlation optimised warping. Journal of Chromatography A, 805, 17-35.
Novoa-Carballal, R., Fernandez-Megia, E., Jimenez, C., & Riguera, R. (2011). NMR
methods for unravelling the spectra of complex misture. Natural Product Reports, 28,
78-98.
Osán, T. M., Olle, J. M., Carpinella, M., Cerioni, L. M. C., Pusiol, D. J., Appel, M.,
Freeman, J., & Espejo, I. (2011). Journal of Magnetic Resonance, 209 (2), 116-122.
Ogata, H., Goto, S., Sato, K., Fujibuchi, W., Bono, H., & Kanehisa, M. (1999). KEGG:
Kyoto Encyclopedia of Genes and Genomes. Nucleic Acids Research, 27, 29-34.
O’Sullivan, A., Gibney, M. J., & Brennan, L. (2011). Dietary intake patterns are reflected
in metabolomic profiles: potential role in dietary assessment studies. The American
Journal of Clinical Nutrition, 93, 314-321.
Özdemir, V., & Kolker, E. (2016). Precision nutrition 4.0: a big data and ethics foresight
analysis - convergence of agrigenomics, nutrigenomics, nutriproteomics, and
nutrimetabolomics. Omics: A Journal of Integrative Biology, 20(2), 69-75.
Pan, X., Smith, F., Cliff, M. T., Capozzi, F., & Mills, E. C. (2014). The application of
nutrimetabolomics to investigating the bioaccessibility of nutrients in ham using a
batch in vitro digestion model. Food and Nutrition Sciences, 5(1), 17.
Patel, K. K., Khan, M. A., & Kar, A. (2015). Recent developments in applications of MRI
techniquesfor foods and agricultural produce-an overview. Journal of Food Science
and Technology, 52(1), 1-26.
Pauli, G. F., Jaki, B. U., & Lankin, D. C. (2005). Quantitative 1H-NMR: Development and
Potential of a Method for Natural Products Analysis. Journal of Natural Products, 68,
133-149.
Peemoeller, H., & Pintar, M. M. (1980a). Two-dimensional time-evolution approach for
resolving a composite free-induction decay. Journal of Magnetic Resonance, 41, 358-
360.
Peemoeller, H., Shenoy, R. K, & Pintar, M. M. (1980b). Two-dimensional NMR time
evolution correlation spectroscopy in wet lysozyme. Journal of Magnetic Resonance,
45, 193-204.
Peemoeller, H. (1989). NMR spin grouping. Bulletin of Magnetic Resonance, 11, 19-30.
Perlo, J., Casanova, F., & Blümich, B. (2005). Profiles with microscopic resolution by
single-sided NMR. Journal of Magnetic Resonance, Series B, 176, 64-70.
Raffo, A., Gianferri, R., Barbieri R., & Brosio, E. (2005). Ripening of banana fruit
monitored by water relaxation and diffusion 1H-NMR measurements. Food Chemistry,
89, 149-158.
Raich, H., & Blümler, P. (2004). Design and construction of a dipolar Halbach array with
a homogeneous field from identical bar magnets: NMR Mandhalas. Concepts
Magnetic Resonance B, 23, 16-25.
Rutledge, D. N. (1992). Low resolution pulse nuclear magnetic resonance in the agro-food
industry. Journal de chimie physique et de physico-chimie biologique, 89, 273-285.
Salek, R. M., Neumann, S., Schober, D., Hummel, J., Billiau, K., Kopka, J., Correa, E.,
Reijmers, T., Rosato, A., Tenori, L., Turano, P., Marin, S., Deborde, C., Jacob, D.,
Rolin, D., Dartigues, B., Conesa, P., Haug, K., Rocca-Serra, P., O’Hagan, S., Hao, J.,
van Vliet, M., Sysi-Aho, M., Ludwig, C., Bouwman, J., Cascante, M., Ebbels, T.,
Griffin, J. L., Moing, A., Nikolski, M., Oresic, M., Sansone, S. A., Viant, M. R.,
Goodacre, R., Günther, U. L., Hankemeier, T., Luchinat, C., Walther, D., & Steinbeck,
C. (2015). COordination of Standards in MetabOlomicS (COSMOS): facilitating
integrated metabolomics data access. Metabolomics, 11, 1587-1597.
Sakurai, T., Yamada, Y., Sawada, Y., Matsuda, F., Akiyama, K., Shinozaki, K., Hirai, M.
Y., & Saito, K. (2013). PRIMe update: innovative content for plant metabolomics and
Tang, H. R., Belton, P. S., Ng, A., Waldron K. W., & Ryden, P. (1999a). Solid state 1H-
NMR studies of cell wall materials of potatoes. Spectrochimica Acta Part A: Molecular
and Biomolecular Spectroscopy, 55, 883-894.
Tang H. R. & Belton, P. S. (1999b). Advances in magnetic resonance in food science.
Cambridge, UK: Royal Society of Chemistry.
Tanner, J. E. (1970). Use of stimulated echo in NMR diffusion studies. The Journal of
Chemical Physics, 52(5), 2523-2526.
Trimigno, A., Picone, G., Capozzi, F. (2015) A 1H NMR-based metabolomics approach on
dietary biomarker research in human urine. In F. Capozzi, L. Laghi, & P. S. Belton
(Eds.), Magnetic Resonance in Food Science: Defining Food by Magnetic Resonance
(1° edition, 143-153). Cambridge, (UK): Royal Society of Chemistry.
van Dusschoten, D., de Jager, A., & Van As, H. (1995). Extracting diffusion constants from
echo-time-dependent PFG NMR data using relaxation-time information. Journal of
Magnetic Resonance, Series A, 116 (1), 22-28.
Van Dusschoten, D., Metzner, R., Kochs, J., Postma, J. A., Pflugfelder, D., Bühler, J.,
Schurr, U., & Jahnke, S. (2016). Quantitative 3D analysis of plant roots growing in soil
using magnetic resonance imaging. Plant Physiology, 170, 1176-1188.
Vázquez-Fresno, R., Llorach, R., Urpi-Sarda, M., Khymenets, O., Bulló, M., Corella, D.,
Fitó, M., Martínez-González, M. A., Estruch, R., & Andres-Lacueva, C. (2015a). An
NMR metabolomics approach reveals a combined-biomarkers model in a wine
interventional trial with validation in free-living individuals of the PREDIMED study.
Metabolomics, 11(4), 797-806.
Vázquez-Fresno, R., Llorach, R., Urpi-Sarda, M., Lupianez-Barbero, A., Estruch, R.,
Corella, D., Fitó, M., Arós, F., Ruiz-Canela, M., Salas-Salvadó, J., & Andres-Lacueva,
C. (2015b). Metabolomic pattern analysis after mediterranean diet intervention in a
nondiabetic population: a 1- and 3-year follow-up in the PREDIMED study. Journal
of Proteome Research, 14(1), 531-540.
Veselkov, K. A., Lindon, J. C., Ebbels, T. M. D., Crockford, D., Volynkin, V., Holmes, E.,
Davies, D. B., & Nicholson, J. K. (2009). Recursive Segment-Wise Peak Alignment
of Biological 1H- NMR spectra for improved metabolic biomarker recovery. Analytical
Chemistry, 81, 56-66.
Zhu, Y., Wang, P., Sha, W., & Sang, S. (2016). Urinary biomarkers of whole grain wheat
intake identified by non-targeted and targeted metabolomics approaches. Scientific
Reports, 6, 36278 (16 pp.).
Zuppi, C., Messana, I., Forni, F., Ferrari, F., Rossi, C., & Giardina, B. (1998). Influence of
feeding on metabolite excretion evidenced by urine 1H NMR spectral profiles: a
comparison between subjects living in Rome and subjects living at arctic latitudes
(Svaldbard). Clinica Chimica Acta; International Journal of Clinical Chemistry, 278,
75-79.
Westad, F., & Marini, F. (2015). Validation of chemometric models - a tutorial. Analytica
Chimica Acta, 893, 14-24.
Windt, C. W., Soltner, H., van Dusschoten, D., & Blümler. P. (2011). A portable Halbach
magnet that can be opened and closed without force: The NMR-CUFF. Journal of
Magnetic Resonance, 208, 27–33.
Wishart, D. S., Tzur, D., Knox, C., Eisner, R., Guo, A. C., Young, N., Cheng, D., Jewell,
K., Arndt, D., Sawhney, S., Fung, C., Nikolai, L., Lewis, M., Coutouly, M. A.,
Forsythe, I., Tang, P., Shrivastava, S., Jeroncic, K., Stothard, P., Amegbey, G., Block,
D., Hau, D. D., Wagner, J., Miniaci, J., Clements, M., Gebremedhin, M., Guo, N.,
Zhang, Y., Duggan, G. E., MacInnis, G. D., Weljie, A. M., Dowlatabadi, R., Bamforth,
F., Clive, D., Greiner, R., Li, L., Marrie, T., Sykes, B. D., Vogel, H. J., & Querengesser,
L. (2007). HMDB: the human metabolome database. Nucleic Acids Research, 35,
D521-D526.
Wishart, D. S. (2008). Quantitative metabolomics using NMR. TRAC-trends in Analytical
Chemistry, 27, 228-237.
Wishart, D. S., Knox, C., Guo, A. C., Eisner, R., Young, N., Gautam, B., Hau, D. D.,
Psychogios, N., Dong, E., Bouatra, S., Mandal, R., Sinelnikov, I., Xia, J. G., Jia, L.,
Cruz, J. A., Lim, E., Sobsey, C. A., Shrivastava, S., Huang, P., Liu, P., Fang, L., Peng,
J., Fradette, R., Cheng, D., Tzur, D., Clements, M., Lewis, A., De Souza, A., Zuniga,
A., Dawe, M., Xiong, Y. P., Clive, D., Greiner, R., Nazyrova, A., Shaykhutdinov, R.,
Li, L., Vogel, H. J., & Forsythe, I. (2009). HMDB: a knowledgebase for the human
metabolome. Nucleic Acids Research, 37, D603-D610.
Online References
http://metabolomicssociety.org/resources/metabolomics-databases
https://www.drugbank.ca/w/databases
http://www.ebi.ac.uk/metabolights/
http://foodb.ca/
Chapter 6
1
Laboratorio di Risonanza Magnetica “Annalaura Segre,” Istituto di Metodologie
Chimiche, CNR, Monterotondo (Rome), Italy
2
Dipartimento di Chimica e Tecnologie del Farmaco,
Sapienza Università di Roma, Rome, Italy
3
Chemometrics and Analytical Technology, Department of Food Science,
University of Copenhagen, Copenhagen, Denmark
4
Dipartimento di Scienze e Tecnologie Chimiche,
Università di Roma “Tor Vergata,” Rome, Italy
5
Istituto per lo Studio delle Macromolecole, Lab NMR, CNR, Milan, Italy
6
Dipartimento di Scienze degli Alimenti, Università degli Studi di Parma,
Parma, Italy
7
Dipartimento di Scienze e Tecnologie Agro-Alimentari,
Università di Bologna, Cesena, Italy
8
Dipartimento di Scienze Chimiche e Geologiche, Università di Cagliari,
Monserrato (Cagliari), Italy
9
Consiglio per la ricerca in agricoltura e l’analisi dell’economia agraria –
Centro di Ricerca per lo Studio delle Relazioni tra
Pianta e Suolo (CREA-RPS), Rome, Italy
10
Dipartimento di Scienze e Tecnologie Biologiche ed Ambientali,
Università del Salento, Campus Ecotekne, Lecce, Italy
11
CNR-IPCF, Istituto per i Processi Chimico-Fisici del CNR di Messina,
Messina, Italy
12
Dipartimento di Chimica, Sapienza Università di Roma, Rome, Italy
13
Dipartimento di Ingegneria Civile, Ambientale, del Territorio,
Edile e di Chimica (DICATECh), Politecnico di Bari, Bari, Italy
14
Italian Institute of Technologies, @CNLS Sapienza, Rome, Italy
15
Dipartimento di Chimica, Università degli Studi di Bari “Aldo Moro,” Bari, Italy
16
Centro di Ricerca di Risonanze Magnetiche CERM, Università di Firenze,
Sesto Fiorentino (Florence), Italy
17
Consorzio interuniversitario per lo sviluppo dei Sistemi a Grande Interfase - CSGI,
Sesto Fiorentino (Florence), Italy
18
Dipartimento di Scienze Chimiche, Università degli Studi di Padova, Padova, Italy
19
Centro Interdipartimentale per la Risonanza Magnetica Nucleare per l’Ambiente,
l’Agro-Alimentare ed i Nuovi Materiali (CERMANU),
Università di Napoli Federico II, Portici (Naples), Italy
20
Istituto Zooprofilattico Sperimentale dell'Abruzzo e del Molise
“G. Caporale,” Teramo, Italy
21
Dipartimento di Scienze Chimiche e Geologiche,
Università di Modena e Reggio Emilia, Modena, Italy
22
Dipartimento di Farmacia, Università degli Studi di Salerno, Salerno, Italy
23
Dipartimento di Farmacia, Università di Napoli Federico II, Naples, Italy
24
Dipartimento di Scienze per la Qualità della Vita,
Università di Bologna, Rimini, Italy
25
Dipartimento di Scienze Biomediche, Odontoiatriche e delle Immagini
Morfologiche e Funzionali, Università degli Studi di Messina, Messina, Italy
26
Dipartimento di Scienza Applicata e Tecnologia,
Politecnico di Torino, Tourin, Italy
27
Dipartimento di Medicina Sperimentale e Clinica,
Università di Firenze, Florence, Italy
28
Dipartimento MIFT, sezione di Fisica, Università di Messina, Messina, Italy
ABSTRACT
Keywords: liquid state NMR, HR-MAS NMR, food science, food composition,
chemometrics
1. INTRODUCTION
In this chapter, some fundamental aspects of NMR methodologies in food science are
described. Both Parts A and B are dedicated to the most relevant state-of-the-art practical
applications of NMR in food science. The examples reported are chosen by members of
the Italian Group of Magnetic Resonance in Food Science actively involved in the
development of new NMR methodologies to study food matrices using various NMR
approaches. In this chapter, the applications of high-resolution NMR in liquid-state and in
semi-solid matrices (HR-MAS in the analysis of food components and entire food samples
are demonstrated using examples of different food matrices (vegetal and animal origin) and
different problems related to food safety, traceability, geographical and botanical origin,
farming methods, food processing, maturation and ageing, etc. are discussed.
In this chapter the applications of high resolution NMR in liquid state and for semi-
solid samples commonly used in food science are described.
2.1. Honey
considered in their interpretation: plant morphology, contamination of the hive with pollen
unrelated to the nectar source, contamination during uncapping and processing,
ultrafiltration of honey (Corvucci et al., 2015). Furthermore, pollen analysis is labor
intensive and requires highly trained and experienced technicians. Despite these
shortcomings, it is the only technique used to date to discriminate between polyfloral and
monofloral honey and to identify its geographical origin. The European Union Commission
is encouraging the development of new analytical methods for this purpose.
The differentiation of honey types can in principle be achieved through the
quantification of several analytes. Flavonoids, proteins, phenolic compounds, volatiles,
carbohydrates, amino acids, and trace elements have all been proposed (Kaškoniene &
Vneskutonis, 2010). If more than one class of compounds is considered, the classification
that can be achieved is more accurate, although multiple techniques and/or protocols might
be required, complicating the analysis. Using spectroscopic techniques, coupled to
statistical analysis, multiple components can be detected simultaneously; this approach
seems the most promising to achieve geographical and botanical honey classification.
High-resolution NMR spectroscopy is particularly suited for this problem because of the
possibility to quantify a wide range of chemical species in a single experiment.
This work contains a survey of NMR-based metabolomic studies performed to
determine honey provenance in terms of geographical area and botanical origin. Paragraph
2.1.1. describes the results of a few studies performed on honey samples dissolved in water.
The aqueous spectra are dominated by strong sugar signals, but amino acids and organic
acids could also be identified. These spectra were used as fingerprint for geographical
honey classification, for the detection of honey adulteration with sugar syrup
(Schwarzinger et al., 2015), and, in a few cases, some botanical origins were identified.
Paragraph 2.1.2. describes the study of minor components extracted with chloroform.
Sugars are absent in the chloroform extracts, whereas many other classes of organic
compounds are present, both volatile and less volatile ones, some of which are secreted by
the bees and many of which are transferred from the flower to the honey. These molecular
profiles are therefore diagnostic for entomological (Schievano et al., 2013a) but also
botanical and geographical classification. In this paragraph, a new approach for a botanical
classification is presented. The combination of NMR data from water solution and organic
extracts gives an overall picture of almost all types of organic compounds present in the
honey in a precise and accurate way, with simple sample preparation and rapidity of
analysis. This seems to be, to date, the most promising technique in honey authenticity
assessment and in this direction future research could be developed.
and acacia honey samples from different EU and non EU countries. The PCA model
showed a clear differentiation between polyfloral and acacia honey samples with the latter
characterized by a higher amount of sucrose and fructose with respect to polyfloral
samples. PCA analysis performed with Hungarian and Italian acacia honeys also led to a
clear separation (Figure 1).
Figure 1. Hierarchical PLS-DA score plot performed by considering 13 polyfloral honey samples of
certain origins, constituting the training set. Filled symbols represent honey samples from Hungary
(diamond), Italy (circle), and Argentina (triangle). PC1 =38.7% and PC2 = 22.7%. R2X= 91.9%, R2Y
=79.7%, and Q2 = 72.7% (Reprinted with permission from Consonni R. & Cagliani L. R., 2008a).
deviations of some carbons of both glucose and fructose isoforms with respect to other
acacia honeys analyzed. In another work, Consonni et al. (Consonni et al., 2012a) used the
saccharide content to characterize chestnut, polyfloral, acacia, rhododendron, and high
mountain Italian honeys; nineteen saccharides (mono, di, tri, and tetra saccharides) were
identified. An OPLS-DA model allowed a discrimination of highly valued honeys, such as
high mountain polyfloral from rhododendron and from polyfloral honeys. In another work,
(Consonni et al., 2013), the same authors used the content of the 19 saccharides identified
to distinguish the origin of Chinese, Hungarian, Italian, and South American polyfloral
honey using an OPLS-DA model (Figure 3).
Figure 2. Classification list for all polyfloral honey test set samples reprojected onto the hierarchical PLS-
DA model performed by considering 13 polyfloral honey samples as a training set a (Reprinted with
permission from Consonni R. & Cagliani L. R., 2008a. Copyright (2008) American Chemical Society).
i.e., Piedmont, Trentino-Alto Adige, Lombardy, and Aosta Valley by performing a two
steps OPLS-DA classification.
Figure 3. Score plot of OPLS-DA performed by considering 22 polyfloral honey samples: filled boxes,
dots, triangles, and diamonds represents Chinese, Hungarian, Italian and South American polyfloral
honey samples respectively. R2X = 91.5%, R2Y = 85.5% and Q2 = 73.2% (Consonni et al., 2013).
Figure 4. Score plot of OPLS-DA performed by considering 17 rhododendron honey samples: filled
triangles, dots, and diamonds represents honey samples from Lombardy, Piedmont and Aosta Valley
regions of Italy. R2X = 77.4%, R2Y = 83.7% and Q2 = 66.1% (Consonni et al., 2013).
Figure 5. Example of a 1H chloroform extract spectrum of a honey with the assignment of some
compounds.
The present study focused on chloroform extracts where many different chemical
classes of compounds are present: essential oils, flavonoids, terpenes, phenols, fatty acid
derivatives, aldehydes and others (Figure 5). Some of these derive on the physiology of the
bee, some arise in processing after harvest and others derive from the original sources of
nectar.
Each plant has a distinct profile of these molecules, so some of them may be used as
specific markers for the botanical origin of honey. The extraction procedure used is very
simple, relatively quick, and reproducible. It consists in a water-chloroform extraction, and
recovery of the organic phase (Schievano et al., 2010). The study was performed on 1155
samples of 18 different monofloral honeys reported in Table 2.1 (Schievano et al., 2016).
The spectrum of a chloroform extract (Figure 5) appears very crowded in the entire
spectral range. However, closer inspection of specific regions reveals some distinct
differences in the chemical composition of metabolites among the various honey types
(Figure 6). These significant differences make the NMR spectra of chloroform extracts
excellent fingerprints for a botanical classification.
Figure 6. Portion of 1H NMR spectra of chloroform extracts of different monofloral honeys. Black dots
highlight the diagnostic resonances for each honey type.
Figure 7. Portion of 1H NMR spectrum of citrus honey chloroform extracts (a). Portion of 1H NMR
spectra of the fractions with isolated markers of (E)-2,6-dimethylocta-2,7-diene-1,6-diol (b) and caffeine
(c).
Figure 8. Up: chemical structure of the dehydrovomifoliol, the eucalyptus marker. Down: selected
portions of 600 MHz two-dimensional NMR spectra and peak assignments.
Figure 9. S-plot for the OPLS-DA model built to characterize the class of thistle (top). Representative 1H
NMR spectrum of thistle honey (bottom). Symbols indicate the position of the principal marker
resonances in the S-plot, in the NMR spectrum, and the corresponding protons in the molecular structures.
Figure 10 shows a few examples of the classification obtained. The first sample has a
high value (0.85) of the predictive component for the ailanthus class and a low value (0.24)
for linden. Using the above criterion, it is classified as monofloral ailanthus honey with a
contribution of linden. In a similar way, the second sample is assigned to the acacia class
with a contribution of wildflower. In the third case, non-negligible yk values are obtained
for 3 classes; this is a polyfloral honey with the contribution of the floral sources reported
in the last column. The novelty of this metabolomic approach, with respect to all the other
ones present in the literature, is that for an unknown sample, it gives not only the
compliance to the declared class but it also reveals the principal and the secondary floral
origins present in that honey.
The class assignment and the floral contribution identification obtained in this way on
120 additional samples was compared with that obtained by traditional analysis (Table 2.1)
and there was complete agreement for the 90% of the samples. For the remaining 10%, the
botanical species contributing to the honey found with the two methods were the same and
the different final assignments were clearly a consequence of the different approaches of
the two methods. In some instances, the two methods do not reach the same assignment
because NMR provides the real composition of the honey while the traditional method
evaluates the compliance to a specific profile. Specifically, the NMR data reflect the
botanical species present while the traditional method assesses the suitability to a
commercial profile. For example, disagreement is common for honeys in which the main
botanical species is underrepresented in the pollen composition or does not produce pollen.
One such case is asphodel, a so-called light-honey: the honey has light color and flavor,
and the flower produces very low amounts of pollen. Pollen analysis is unable to identify
it, and sensory analysis is necessary, but a small amount of a strong-flavored and/or colored
honey is sufficient to cover its origin. In this case, NMR identifies it as asphodel, as it is
the main botanical origin, but the complete profile does not meet the commercial one for
monofloral asphodel honey. Finally, the assignment is different for the two methods when
a species is present, which has not yet been studied by NMR. A future perspective will be
to enlarge the database by introducing other minor types.
Figure 10. Up: NMR definition of monofloral and polyfloral honey. In down table: Examples of the NMR
classification. The yk scores are reported for each sample and for each OPLS-DA model. Values on a
dark grey background are above the set threshold. The species that were found to contribute significantly
are reported in light grey.
The NMR profile provides indications also on the production area. That is, beside the
signals of the main botanical species, other typical floral species of a given production
contribute to the spectrum, allowing a geographical classification. In this regard, the first
results have already been achieved (Schievano et al., in preparation).
In conclusion, we can say that the proposed new method may give a comprehensive
solution to the problem of fraudulent practices in honey labeling, and allow rapid, less
expensive and objective quality control. It may prove to be a valid alternative to the
classical methods that often require several physical-chemical measurements and sensory
evaluation to support melissopalynological analysis, which makes the assessment very
demanding and complex.
Table 2.1. Number of honey samples used to build the statistical model
(second column) and number of honeys used as control samples
(third column) of different botanical origin (first column)
Milk provides essential nutrients and is an important source of dietary energy, high-
quality proteins and fats. Human breast milk is the gold standard in neonate nutrition, able
to satisfy all the needs of the newborns. However, animal milk can play an important role
in the diets of children when breast milk is not available and in populations with very low
fat intakes and limited access to other animal source foods (FAO). Animal milk for human
consumption and/or for dairy processing is mainly from: cow (Bos taurus), buffalo
(Bubalus bubalis), camel (Camelus bactrianus), small ruminants such as sheep (Ovis aries)
and goat (Capra hircus), yak (Bos grunniensis) and equine (horse and donkey). The Codex
Alimentarius defines a milk product as a “product obtained by any processing of milk,
which may contain food additives and other ingredients functionally necessary for the
processing.” The range of milk products varies significantly from region to region and
among countries, depending on dietary habits, the milk processing technologies, market
demand, and social and cultural circumstances (FAO). Milk undergoes different treatments
that yield different typologies such as pasteurized milk, skimmed milk, reconstituted milk,
ultra-high-temperature (UHT) milk and fortified milk, more recently lactose-free milk is
available, while raw milk is less and less consumed. Fermented milk products include
yoghurt, kefir and other. Cheeses are produced through the coagulation of milk protein
(casein), which is separated from the milk’s whey. Hundreds of varieties of cheese are
produced, many of them being characteristic to a particular region and protected by
different designations. Cheese can be soft, hard, semi-hard, hard ripened or unripened.
Cheese’s diverse characteristics derive from differences in the compositions and types of
milk, processes applied and microorganisms used. Butter and cream are fatty milk
products. Researches on milk and dairy products cover different areas, such as chemical
characterization, quality assessment, traceability, herd management, industrial processes,
shelf-life, and fraud detection. In this contest, NMR is a quite active research tool in the
study of different aspects of milk and dairy products (Maher & Rochfort, 2014; Sundekilde
et al., 2013a). HR-NMR studies on milk and cheese are usually carried out on the lipophilic
and hydrophilic extracts, seldom milk has been studied as it is (Monakhova et al., 2012).
Different separation and extraction procedures can be applied, such as the Folch (Scano et
al., 2011; Piras et al., 2013), or the Bligh-Dyer (Tsiafoulis et al., 2014; Papaemmanouil et
al., 2015; Cesare Marincola et al., 2012) methods, moreover, the lipid fraction of cheese
can be simply extracted by deuterated-CHCl3 without further preparation of sample
(Schievano et al., 2008). Sonication of samples is sometimes recommended for the
presence of the original fat globules in raw milk or post-processing assembly of
amphipathic materials.
al. the combination of NMR profiling and multilinear regression was reported as a powerful
method for the evaluation of the unknown composition of milk mixtures from different
animal species (Lamanna et al., 2011).
Some studies were focused on the change in metabolic profile of cheese samples during
ripening and in the presence of different microorganisms. The metabolic profiles of
potential probiotic or synbiotic cheeses were studied by Rodrigues et al. (Rodrigues et al.,
2011). The combination of HR-1H NMR spectral data and multivariate statistical analysis
allowed aqueous profiles of cheese samples to be distinguished by maturation time, added
probiotic bacteria and added prebiotic. Principal Component Analysis (PCA), of 1H NMR
spectral data was able to characterize Gouda cheeses at different ripening time, adjunct
cultures and brining conditions. Ruyssen et al. showed a separation in the PCA score plot
according to the ripening time and particular adjunct cultures (Ruyssen et al., 2013). Fiore
Sardo Protected Designation of Origin (PDO) cheese is produced from raw ewe’s milk
without the use of exogenous starter cultures or with the addition of exclusively
autochthonous lactic acid bacteria (LAB). To study the effects of different LAB cultures
on cheese characteristics, (Piras et al., 2013), by a PLS approach, successfully linked
variation of the 1H-NMR metabolite profile to the starter culture used and to the ripening
stage. Parmigiano Reggiano and Grana Padano are important PDO cheeses, both produced
in northern Italy. The influence of the ripening time on the metabolic profile of Grana
Padano cheese was studied by De Angelis Curtis et al. (De Angelis Curtis et al., 2000),
where the increase of free amino acid content was associated to proteolytic processes. The
modification of chemical profile during ripening, and the possibility to discriminate
Parmigiano Reggiano cheese samples from the analogue “Grana type” cheeses produced
in east Europe countries, were also the subject of the research by Consonni et al. (Consonni
& Cagliani, 2008b). Mozzarella di Bufala Campana cheese is an unripened soft cheese with
a PDO trademark produced in Southern Italy obtained only from buffalo milk. The
product’s characterization was performed with an integrated approach based on low- and
HR-NMR techniques (Gianferri et al., 2007).
Preliminary results from one of the authors (Tenori et al., in preparation) suggest that
NMR metabolomics can be successfully used for milk traceability. These data show that,
through the NMR metabolic profile of milk, it is possible to accurately distinguish different
milk samples originating from different farms (cross-validated recognition accuracy > 95%
using PLS discriminant analysis) even when they are located in the same restricted
geographical area. Furthermore, nutritional patterns characteristic of different seasons and
of different farms are also well reflected in the milk profiles and characterized by NMR.
The most important difference was between using or not silage feedings. Metabolite
analysis showed a 3-fold increase of lecithin in the milk of cows fed with silages. This
finding is very relevant to assess the quality of dairy products. A prominent example is
Parmigiano Reggiano, the preparation protocol of which excludes milk from silage-fed
cows. Finally, correlations between spectral and nutritional data were analysed to
understand relations between feeding and metabolites, and to evaluate whether and at
which extent the production of some metabolites can be influenced by cow nutrition.
Moreover, to study the possibility to discriminate the geographical origin of buffalo
milk and buffalo mozzarella cheese, Brescia et al. studied the products coming from two
areas of Southern Italy with different analytical techniques. The authors successfully
obtained discrimination of mozzarella samples with a combination of NMR data and
isotopic ratios (Brescia et al., 2005). An interesting application of HR-1H NMR for the
quantitative determination of histamine in different cheese samples was reported by
Schievano et al. (Schievano et al., 2009). Health status of the animals is of primary
importance to have high quality milk and dairy. Ketosis in cow was studied by Klein (Klein
et al., 2011), it has been assessed that healthy animals had higher levels of milk GPC
(glycerophosphocholine) and lower levels of milk PC (phosphocholine) compared to
animals suffering ketosis. Furthermore, Klein et al. found that animals with high milk
GPC/PC ratios showed a lower milk fat levels and lower plasma PtC (phosphatidylcholine)
levels (Klein et al., 2011), so linking changes in milk metabolites to those in plasma due to
pathological conditions. In healthy dairy cow, Maher et al. explored the metabolic
relationships between blood and milk. The study confirmed that milk is a distinct metabolic
compartment with a metabolite composition largely not influenced by plasma composition
under normal circumstances (Maher et al., 2013). One of the parameter used to assess milk
quality, onset of mastitis, aptitude of milk in cheese making is the count of somatic cells
(SCC) in milk. Sundekilde et al. (Sundekilde et al., 2013b) correlated milk metabolite
levels to medium and high SCC in cow milk. These researchers endorsed the 1H NMR
metabolic profiling as a useful tool for diagnosis of milk animals diseases to ameliorate
herd management. Moreover, Sundekilde et al. investigated the relationship between the
metabolite profile of bovine milk and important technological properties by using 1H and
13
C NMR metabolomics approach. The authors reported a good relationship between the
metabolite profile and the coagulation properties of the milk (Sundekilde et al., 2011).
spectra of milk fat by Papaemmanouil et al. (Papaemmanouil et al., 2015), and in 1H and
13
C NMR spectra by Scano et al. (Scano et al., 2011) in pecorino sardo cheese, the latter
particularly rich on this FA. Tsiafoulis (Tsiafoulis et al., 2014) achieved to characterize the
1
H NMR signals of four CLA isomers in the lipid fraction of lyophilized milk. Positional
isomerism of FA in triacylglycerols, as studied by Andreotti et al. (Andreotti et al., 2002),
is of particular importance for their availability during digestion and for the development
of dairy aromas, diacylglycerol and monoacylglicerols, minor constituents of the milk and
cheese lipid fraction, are extremely informative on the lipolytic process in cheese,
organoleptic properties and health implication of dairy. The presence of di- and mono-
acylglycerols was detected in 1H spectra of milk fat (Papaemmanouil et al., 2015), and in
cheese fat also free fatty acids, aldehydes, and hydroperoxides were observed (Scano et al.,
2011). Phospholipids are valuable minor components of milk and dairy (0.3−1.9% in
cheese fat). Phospholipids have a high nutritional value and useful technological properties.
Phospholipids in cheese were detected and quantified by 1D 31P and 2D 31P-1H NMR
(Kaffarnik et al., 2013). In milk of different species, namely human, cow, camel, mare,
Garcia et al. (Garcia et al., 2012), by 31P NMR, identified a large number of different
phospholipids, among which the bioactive plasmalogens. Moreover Andreotti et al.
identified by 31P-NMR spectroscopy different phosphorylated compound from buffalo and
cow milk (Andreotti et al., 2006).
In the field of fraud detection, the results from four separate analytical methods: 1H
NMR, 13C NMR, stable isotope data and the α-linolenic acid content, were combined by
multivariate statistical analysis improving authentication of organic vs. conventional milk.
The investigated set of samples comprised various milk categories, from organic and
conventional farms and retail milk (Erich et al., 2015). A very interesting research
highlighted the presence of a peculiar FA (cyclopropyl-FA) in rumen, milk and cheese of
cows exclusively feed with maize silage. Cyclopropyl-FA was identified by 1H NMR and
content determined by GC-MS. This novel FA, of microbial origin, is now proposed as
biomarker of silage feeding so to detect fraud, mislabeling dairy products such as
Parmigiano Reggiano DOP that doesn’t allow silage in the cow’s diet (Marseglia et al.,
2013). In the context of food surveillance to validate the labeling of milk-based products,
the 1H NMR study of the lipid fraction allowed to distinguish cheese and ice cream
produced by vegetable fats, fraudulently used as cheap substitutes of milk fat. PLS
calibration on 1H NMR spectral data with GC-MS data were set up to be used as a first
screening to detect the adjoin of vegetable fats (Monakhova et al., 2013). An interesting
research showed that PDO Asiago cheese produced in alpine farms can be differentiate
from that produced in lowland and mountain industrialized factories by applying
chemometric analysis to both 1H and 13C NMR data set (Schievano et al., 2008).
2.2.3. HR-MAS
Analogously to liquid state HR studies, some topics were investigated by HR-MAS. In
particular, commercial and certified mozzarella cheese from Campania buffalo milk
(MBC) were compared in order to assess quality and traceability (Mazzei & Piccolo, 2012).
Regarding the former, it was possible to discriminate between commercial and certified
samples on the basis of galactose, lactose, acetic acid and glycerol content, and it was
observed that 2 days old MBC samples showed increasing signals for isobutyric alcohol,
lactic acid, and acetic acid. Emmental, another PDO diary product, was also examined
(Shintu & Caldarelli, 2006) with the aim to discriminate among cheese wheels ailing from
neighboring European countries. The ripening of Parmigiano Reggiano (Shintu &
Caldarelli, 2005) was examined at 4, 8, 12, 18 and 24 months, and the results confirm the
relevant role played by amino acids and other low molecular weight compounds, for the
successfully evaluation of the process. Moreover, HR-MAS has proven to be able to
characterize the Carganico Cacioricotta cheese (Apulia) (Sacco et al., 2011), starting from
the raw materials (goat milk) and arriving to commercially ripe cheese wheels studied at
different times.
Extra virgin olive oil (EVOO) (International Olive Council) is the fifth most important
oil crop in the world behind wheat, rice and coarse grains. Presently, due to the beneficial
nutritional and sensory properties and high economic value, the production and
consumption of extra virgin olive oil (EVOO) involves not only Mediterranean area but
also other areas. The unique physical–chemical and nutraceutical properties, as well as
sensory characteristics, of the EVOOs are due to the high content of unsaturated acids
(oleic, linolenic and linoleic acids) and antioxidant compounds (polyphenols, carotenoids,
tocopherols, sterols, squalene etc) and depend on many factors including the geographical
origin and cultivar used for olive oil production. As reported in many papers (Mannina &
Sobolev, 2011), 1H NMR methodology enables the detection of major and minor
components present in olive oils and, together with suitable chemometric methods,
represents a powerful tool to obtain the information on olive oil quality, authenticity,
geographical origin and variety.
Here, we report some relevant aspects regarding the detection of minor and major
compounds in olive oils and some results of olive oil characterization in terms of
authenticity, geographical origin and variety.
Figure 11. 1H-NMR general profile of the Tonda Iblea 2014 EVOO, the expansions show specific regions
populated by characteristic resonances coming from minor components.
The single pulse 1H experiment provided information on the main components of the
olive oil samples (saponifiable fraction: triglycerides) and minor compounds (sterols,
terpenes, aldehydes etc) (Sacchi et al., 1998), (Figure 11). It is known that minor
constituents could play an important role for authenticity assessment purposes being more
difficult to be adulterated, but unfortunately their 1H NMR signals can be hidden by the
major components in a standard single pulse NMR proton spectrum especially if the 1H
experiment is carried out at low or medium magnetic field. In order to detect better the
signal of minor constituents of olive oil (unsaponifiable fraction) it is possible to use
selective NMR experiments that allow the suppression of signals due to the main
components by multipulse sequence (Longobardi et al., 2012). In Figure 12, the
comparison between a typical NOESYGPPS and ZG1H spectrum is reported illustrating
the main effect of the multiple suppression: the intensities of the triglyceride signals are
drastically reduced whereas the unsaturated signals remained substantially unaffected
getting information on the unsaponifiable fraction and allowing, if necessary, the
quantification of its constituents. Moreover, since the dominating lipid signals are no
longer present in the NOESYGPPS spectra, the receiver gain can be increased improving
the signal-to-noise ratio (about 10 times higher) and therefore the NMR sensitivity. Thus,
the dynamic range of concentrations covered by the two experiments was of the order of
100,000 allowing for a more comprehensive NMR assessment of the samples.
Figure 12. Comparison of 1H spectra obtained with NOESYGPPS (bottom trace) and ZG1H (top trace)
sequences.
The use of selective pulses coupled to gradient spin-echo refocusing called DPFGSE
(Rastrelli et al., 2003), is another very smart NMR strategy suggested to overcome the
problem of the dynamic range, paving the way to the easy detection of some specific minor
components, especially those presenting resonances away from other interfering
resonances just as aldehydes. Rotondo et al. have successfully used this strategy to analyze
the aldehydic profile of several Sicilian EVOOs of the 2009 campaign (Rotondo et al.,
2011). Many aldehydic signals were detected as reported previously (Mannina et al., 2001;
Mannina et al., 2003) but Sicilian olive oils from typical cultivar show new signals at 9.18
and 9.58 ppm at different magnetic fields (300 MHz and 500 MHz). These new signals
disappeared after thermal stress.
In order to assign these new signals, the extended 2D version of the DPFGSE
technique, proposed as useful strategy in food analysis (Koda et al., 2011), was used. It
corresponds to the F2-selection of the specific aldehydic region by the final “fid-reading.”
This sequence uses the fine optimization of the shaped-pulses and pulse sequence designed
by Bruker (Falländen, Switzerland, Figure 13). In this way it was possible to determine
(Dugo et al., 2015) the structure of four aldehydic compounds (Figure 14) which contribute
to nutraceutical properties and olive oil flavor (Perez-Trujillo et al., 2010; Karkoula et al.,
2012). This study suggests that the DPFGSE analysis is the best tool to detect the real (not-
modified) aldehyde profile (Rotondo et al., 2011; Dugo et al., 2015) since other
Figure 13. Off-set stack plot of 1H-NMR traces for “Uovo di Piccione 2011” EVOO: (a) Standard proton
spectrum; (b) Bruker DPFGSE with optimized Gaussian shaped pulse; (c) Default DPFGSE Agilent
“seduce” pulse sequence. The middle trace (b) evidencing the best sensitivity and selectivity, was used
for the 2D extensions. On the right the pulse sequence of the DPFGSE followed by the modified 2D
TOCSY and NOESY versions.
quantified (Salvo et al. 2016), (Figure 15). It was possible to observe that: 1,3DG/1,2DG
is: a) not immediately related to the EVOO milling date, b) cultivar dependent as the
isomerization rate depends on the presence of macromolecules (lipases) and/or metal ions,
c) not really related to the acidity which is pretty constant along the one year shelf-life.
Figure 14. Four different aldehyde types chemically identified by the 2D extensions (TOCSY, NOESY
and HSQC) of the DPFGSE sequence on EVOO just as it is. Below: several aldehyde compounds known
to be components of EVOOs are also represented in the scheme.
Figure 15. Total assignment of the glyceridic 1H-NMR region of the EVOO; integration regions indicated
are those used to infer the 1,3-DGs quantification, on the right the chemical structure of the assigned
species.
Figure 16. Supervised analysis and relative OPLS-DA score plots for Coratina, Cima di Mola, Ogliarola
and Peranzana monovarietal EVOO samples. Due to partial overlap of 2013/2014 and 2014/2015
Coratina samples, the prediction ability for the year classification related to Coratina (Q2 = 0.82) resulted
about 10% lower with respect to Ogliarola (Q2 = 0.95), Cima di Mola (Q2 = 0.93) and Peranzana (Q2 =
0.91) cultivars.
1
H NMR together with chemometric analysis were used to create statistical models to
define and confirm the declared olive oil geographical origin choosing olive oil samples
from Liguria, an Italian region as a model class (Mannina et al., 2010). Mediterranean olive
oils (totally 896 samples) from three consecutive harvesting years have been analyzed.
Good classification PLS-DA and SIMCA models have been created and olive oils from
Liguria have been well defined with respect to the other olive oils. Ligurian olive oils
turned out to be always characterized by a high amount of terpenes (an important sensory
marker) and a low amount of saturated fatty acids, an important “quality marker.”
Recently, a new approach based on the combination of two types of 1H-NMR
fingerprinting experiments with multivariate statistical analysis has been used to classify
Italian and Greek monocultivar extra virgin olive oils (EVOOs) according to their
geographical origin (Longobardi et al, 2014). In particular, 104 authentic monocultivar
EVOO samples coming from three different Apulian regions in Italy (Dauno, Terra di Bari,
Terra d’Otranto) and four different regions in Greece (islands of Kefalonia, Kerkira,
Lefkada, Zakinthos) have been analyzed by using two rapid NMR sequences carried out
on each sample: a standard single pulse NMR experiment (ZG1H) and multi-signal
suppression NMR sequence (NOESYGPPS).
For data treatment, in the article, different data matrices were taken into account:
matrix 1 (obtained from ZG1H spectra), matrix 2 (from NOESYGPPS spectra), and matrix
3 obtained by combining the two previous ones.
Subsequently, on each matrix, a Monte-Carlo embedded cross-validation was used to
demonstrate that a combination of principal component analysis, canonical analysis, and
classification via nearest class mean can be used to predict the origin of olive oil samples.
For matrices 1, 2, and 3 the obtained average correct prediction probabilities were 74%,
65% and 78%, respectively. Moreover, in order to test the reliability and the validity of the
models, artificial randomization of origin-assignment of the matrix 3 was carried out and
a “false random data set” was obtained. The random correct prediction rates ranged from
4 to 19% demonstrating that a remarkable part of the variance in the final model was
effectively due to the different geographical origins, whilst only a small contribution of
noise was modeled. Considering all results reported above and the rather limited number
of samples tested, the proposed method can be considered a promising tool for
geographical origin prediction of olive oils.
analyst evokes, measures, analyzes and interprets human responses to stimuli as perceived
through the senses. Human sensory tests are regularly employed in the food and beverages
industries and they are sometimes integrated by a number of techniques, including the
electronic nose and the electronic tongue (Deisingh & Stone, 2004). Unfortunately, the
usual artificial tongue /nose is used to determine only very specific components of the
analyzed food. Furthermore, not all instrumental techniques are able to analyze directly the
genuine mixture interacting with our sense without any extraction/concentration
procedures. For example, mass spectrometry and gas chromatography require
volatilization of the analyzed compounds that very often is obtained with a chemical
derivatization. In this frame, the potentiality of NMR spectroscopy as predictive tool to
measure sensory descriptors have been tested, without performing any complementary
chemical analyses. In particularl, an NMR metabolomic approach, in combination with
multivariate analysis, has been used to find correlations between traditional sensory
features and chemical components of canned tomato (Malmendal et a., 2011) and extra-
virgin olives (Lauri et al., 2013). In the case of olive oils, this approach revealed that, in
certain conditions, sweet, grassy, artichoke, leaf, tomato, bitter, pungent and rosemary
tastes can be predicted. Interestigly, it turned out that bitter, pungent and artichoke tastes
are highly correlated and display a very similar chemical profile. They all show a strong
anticorrelation to sweet taste, which displays the inverse chemical profile. Tomato and
rosemary tastes also show a strong anticorrelation, while grassy and tomato taste display
similar profiles without being significantly correlated. Unfortunately, not all NMR signals
related to a given sensory descriptor could be unambiguously assigned. Nevertheless, it
was possible to obsverve that the lack of hexenal seems to increase sweet, tomato, grassy
and fruity tastes, whereas the increment of its concentration increases the perception of leaf
and rosemary tastes. Secoiridoids, trans-alkenals and 4-hydroxy-trans-alk-2-enal are in a
way related to the sweet/bitter relationship of the EVOO. Finally, trans-2-hexenal seems
to be correlated to fruity, grassy and tomato tastes. The obtained results are very promising
and pave the way to the use of NMR as powerful analytical method in food industry.
Vinegar is undoubtedly one of the oldest solutions of acetic acid produced by any
fermentable sugary substrate (cider, wine, sorghum etc) after various fermentation
processes and methods. Usually three stages of fermentation are included. The first stage
is starch saccharification, followed by the alcoholic fermentation, which converts
fermentable sugars into ethanol mainly by yeast action. The third stage is the acetic acid
fermentation, which is the oxidation of ethanol into acetic acid by acetic acid bacteria. The
final product shows unique characteristics, due to many kinds of chemical compounds
formed during the fermentation process. The quality of this product is subjected to many
factors, basically the raw materials and the microbial diversity involved in vinegar
fermentation. Several vinegars are worldwide present, each of those endowed with specific
characteristics, reflecting the national typicality, and according to the raw material they
have been classified into three categories: vinegar obtained from vegetables (rice, onion,
tomato etc), vinegar obtained from fruits (cider, mango, pineapple), and vinegar obtained
from animals (honey, whey, etc). Different types of vinegar have different applications:
some are used as preservatives or condiments, some are used as drinks, and others are used
to treat diseases in traditional Chinese medicine.
After the oxidation process, additional ageing processes could take place, performed
into barrel made of different types of wood, thus inducing various aromatic tastes
(balsamic) which make the typicality. Italy has a different culture about vinegar. The
characteristic products, largely appreciated worldwide, consist of Balsamic and Traditional
Balsamic vinegars of Modena (BVM and TBVM respectively). BVM obtained the PGI
trademark in 2009 (Reg. CE n° 583/2009 of 3 July 2009) and it is made from wine vinegar
with the addition of caramel and a small quantity of aged wine vinegar (max 10%), as
described in the set rules (D.M., December 3, 1965) for the preparation procedures. BVM
aging can take up to three years (BVM refined) or up to more than three years (BVM aged)
for a higher quality product. Local market could also develop other low quality products
obtained with a mixture of wine vinegar and aromas or natural dyes, typically, the so called
“industrially produced balsamic vinegars.”
TBVM is conversely, a very different product. It is a typical Italian food with unique
quality features and characteristics that come from the combination of several human and
climatic factors. It is obtained simply from cooked must, according to the precise standards
of the product specification. It is then left to age in batteries made of wooden barrels of
decreasing sizes observing a process of decanting and topping up ending to the smallest
barrel placed at the end of the battery. The different types of wood (cherry, chestnut, oak,
etc) are able to confer aromas to the product during the ageing process. The ageing process
is needed at least for 12 years in order to be called “traditional balsamic vinegar” (TBVM
refined) but can also reach 25 years of age (TBVM extra old) and even more. TBVM
obtained the PDO trademark in 2000 (Reg. CE n° 812/2000, GUCE L. 100 of 20 April
2000). During this ageing process the cooked must experiences several chemical
modifications such as sugar degradation, acetylated derivatives formation, enrichment of
aroma from barrels etc. Currently the quality of both BVM and TBVM are assessed by
means of sensorial analysis and by very simple chemical–physical property determinations,
like total acidity, density and dry residual. In the last years different studies appeared in
literature proposing the NMR spectroscopy in combination with multivariate statistical
analysis protocols as a powerful analytical and objective approach able to characterize the
BVM and TBVM and to assess its quality and aging.
The first NMR study appeared in 2004 (Consonni & Gatti, 2004) showing that the
quantification of five selected metabolites, being ethanol, acetic acid, malic acid, glucose,
and HMF, could lead to a BVM and TVBM aging evaluation, by using PCA. Samples of
vinegars resulted distributed in the PCA score plot in accordance to their aging process.
Successively, Consonni and Cagliani (Consonni & Cagliani, 2007) approached relaxation
analysis of BVM samples by measuring the spin-lattice relaxation time (T1) of acetic acid
and β-glucose. A PLS model based on these measurements combined with quantitative
determination of the five selected metabolites previously discussed resulted in a very good
BVM aging process determination. A further study followed for the characterization of 72
BVM and TBVM samples with different aging processes (Consonni et al., 2008c) (Figure
17). A hierarchical PLS-DA model resulted in a high-predictable capability in terms of the
aging process whose validation was checked on both training and test sets and further
confirmed by accurate prediction of 41 unknown samples. In this model, the loading plot
suggested acetate, ethanol, and 3-hydroxy-2-butanone as the variables characteristic for
samples aged for less than 12 years while sample aged for more than 12 years presented a
higher content in sugars and HMF. This approach could be usefully employed as an
analytical test for the balsamic vinegar aging process, based also on the high reproducibility
of the NMR measurements. Chemical components of BVM were also presented in
Caligiani et al. (Caligiani et al., 2007a).
The same authors investigated the evolution of two chiral minor components of TBVM
upon ageing into barrels, namely 2-hydroxybutanone (acetoin) and 2,3-butanediols
(Caligiani et al., 2007b). The stereoisomer composition of 2,3-butanediol appears related
to indigenous microflora of the specific battery and that of 2-hydroxybutanone a good
marker for ageing process, giving possible contributions in determining the age of TBVM.
A further improvement in the knowledge about BVM and TVBM was perform by
Consonni (Consonni et al., 2008d) by applying the 13C NMR spectroscopy in TBVM
authenticity determination. As a matter of fact, TBVM has a very high economic value
because of the ageing process that can reach more than 25 years, leading it to be subjected
to different frauds practices. In this work different isoforms of glucose and fructose were
investigated. TBVM samples dissolved into water showed the “natural ratio” of two
isoforms for glucose (pyranosidic) and three isoforms for fructose (furanosidic
and pyranosidic). When organic solvent, such as DMSO, was used, a differentiation in
the isoform ratios with respect to the natural one, for both glucose and fructose, was
observed. In particular, the fructose isoform ratio showed the degradation of the most
abundant pyranosidic isoform most likely due to saccharides degradation that took place
during the must cooking process. By analyzing the different glucose and fructose isoforms
ratio in BVM and TBVM sample three different evidences that can suggested the presence
of a fraud were identified (Figure 18): the presence of the fructose in pyranosidic
isoform that is indicative for initial discrimination of TBVM samples (that isoform is
absent in the BVM samples), the intensity ratio for C2FP/C2FF signals, values higher
or equal to one suggest a potential TBVM defraud, and finally the measurement of the
C2FF and C3FF chemical shift deviations, whose values of 1 and 0.5, respectively, were
chosen for authentication alert level. Values lower indicate a potential TBVM defraud. This
procedure has been patented (Consonni, 2007 MI2007A001489) and was successfully
verified with some fraud samples.
Figure 17. 1H NMR spectra recorded at 600MHz of a Traditional Balsamic Vinegar of Modena; resonance
assignment shared among all the NMR studies performed by different authors. From the bottom to the
top traces: expansions of the aromatic/aldehydes, anomeric and aliphatic regions respectively.
In 2009 Cirlini et al. (Cirlini et al., 2009) monitored the formation of glucose and
fructose acetates during the vinegar maturation process. Further improvements derived
from relaxometric studies of vinegars, by the use of field cycling, allowing information
about the interaction of water molecules with paramagnetic and large macromolecular
systems (Baroni et al., 2009). This NMR technique allowed switching the magnetic field
from about 10-4 T up to 0.4 T in the milliseconds time scale by using a relaxometer. The
longitudinal relaxation time (T1) of water protons is affected by the occurrence of
interaction among macromolecules and paramagnetic metal ions. The comparison of
relaxation profiles for authentic and suspected TBVM samples allowed identification of
counterfeit specimens. In particular, counterfeit samples showed much lower values for
longitudinal relaxation times respect to authentic samples, due to the lack of
macromolecular systems, which are naturally formed during the ageing process. In
addition, a relationship correlating the age of TBVM and longitudinal relaxation time could
be found.
Figure 18. Anomeric sugar region of the 13C spectrum relative to three TBVM samples (A), two BVM
samples (B) and three unknown declared TBVM samples differently defrauded (C). Dotted lines in (C)
indicate the shift of fructose isoforms due to different degrees of adulteration (Reprinted with permission
from Consonni R. et al., 2008d. Copyright (2008) Elsevier).
Additionally, studies have been performed by other groups in the recent years (Papotti
et al., 2015), adopting different discriminant analysis protocols to discriminate BVM
against TBVM, but they essentially confirmed the previous findings of other groups.
Additionally, some other information like phenolic content and antioxidant activity
(Bertelli et al., 2015), obtained from other analytical procedures than NMR, could be
included to NMR dataset, also because themselves they did not result sufficiently
discriminant. A different analytical approach for balsamic vinegar authentication was
recently proposed by Werner et al. (Werner & Rossmann, 2015) and Perini et al. (Perini et
al., 2014) groups, based on the evaluation of stable isotope parameters ( 13C and 18O) and
stable isotope ratios (D/H, 13C/12C and 18O/16O) respectively. These approaches are
following the already introduced official methods based on isotope ratio mass spectrometry
and 2H-NMR (Thomas & Jamin, 2009). The first method was proposed to detect possible
addition of components from both C4 and C3 plant sources as mixtures, known as
‘intelligent sugar mixture’, measuring 13C of acetic acid and 18O of water content, following
a stepwise procedure. The second approach provided the experimental evidence that wine
databank could also be used for gathered-products, as BVM represents respect to wine
vinegar.
2.5.1. Fish
In the last decade, several scientific studies on fish and seafood, including wild and
farmed fish and shellfish, both of marine and freshwater origin, have shed light on the
importance of these products in the human nutrition and in the global food supply as
healthy products. The main quality attribute is related to the superior nutritional profile and
benefits especially due to the presence of essential amino acids (arginine, histidine,
isoleucine, leucine, lysine, methionine, phenylalanine, threonine, tryptophan and valine),
highly digestible protein, vitamins (A, D and B complex), minerals and a high content of
polyunsaturated fatty acids, PUFA (Tacon & Metian, 2013; Thilsted et al., 2016). The latter
represent the only source of omega-3 fatty acids, whose importance in prevention of human
diseases has been largely discussed and abundantly presented in literature. As NMR
technology can identify simultaneously many molecular compounds in biological sample,
it has been largely adopted to assess fish quality and freshness. Its application, which
encompasses the study of both aqueous and lipid fraction, covers a wide range of species
and different topics like fish physiology and development, pollutant effects on fish, fish
condition and disease, and fish as foodstuffs. In this last case, other perspective application
areas are sea food safety and quality, fraud, authentication, quality loss during food storage
and processing and nutrition value. For example, in the field of edible fish oils and fats,
NMR provides useful information on the composition of fats and of components as n-3
polyunsaturated fatty acids. 13C and 1H-NMR are complementary techniques for studying
fish lipid composition, structures and chemical alterations: 13C-NMR offers the advantage
of higher resolution but requires longer time; proton NMR, on the other hand, is very rapid
but with reduced resolution, thus lipid proton signals are present in a narrow spectral range.
Therefore, different applications have been carried out considering advantages and
disadvantages of the two spectral frames (Sacchi et al., 2008). The application of high-
resolution 13C-NMR spectroscopy had focused, in preliminary studies, on studying the
composition and structure of fish fatty acids (Aursand & Grasdalen, 1992; Sacchi et al.,
1993). In fact, the 13C-NMR approach provides, in one shot observations, information
about the lipid classes, the fatty acid profile, phospholipids, the positional distribution of
fatty acids in both triacylglycerides and phospholipids and cholesterol/cholesteryl content.
Gradually, the application of 13C-NMR, combined with pattern recognition (multivariate
statistical analysis), has been adopted to classify fish species according to their wild,
farmed and geographical origin in order to move toward the European Commission’s
requests about fish (food) traceability and authenticity in the EU. Aursand et al., in the aim
of traceability and authenticity, investigate on the application of 13C-NMR to discriminate
between farmed and wild Atlantic salmon (Salmo salar L.), between different geographical
origins, and to verify the origin of market samples (Aursand et al., 2009). Muscle lipids
from 195 Atlantic salmon of known origin in addition to market samples were analyzed by
13
C-NMR spectroscopy and multivariate analysis. In the analysis of market samples, five
fish labeled and purchased as wild salmon were classified as farmed salmon (indicating
mislabeling), and there were also some discrepancies between the classification and the
product declaration with regard to geographical origin (Aursand et al., 2009).
Proton NMR provides also fundamental information on the molecular profile, called
metabolic fingerprinting, of fish sample and at the same time allows to investigate
simultaneously a wide range of metabolites which can be relevant when freshness and
nutritional value are verified. K-index and TMAO content are two indices of quality-
freshness and spoilage and both can be determined by using 1H-NMR. For example 1H
NMR data was used to calculate K-index during the storage of fresh aqua-cultured sea bass
(Dicentrarchus labrax) packaged in modified atmosphere or γ-irradiated (Reale et al.,
2008). 1H NMR together with microbiological, sensory and chemical analyses indicated
that the gas mixture composed by 60:35:5 CO2/N2/O2 was the best inhibitory treatment to
extend the shelf life of sea bass.
The degree of freshness of fish muscle can be also monitored by the changes in
muscle’s amino acid composition as free amino acids are key molecules in both autolysis
and biological spoilage reactions. Ciampa et al. propose, as an alternative method to the K-
index, 1H-NMR molecular profiling to follow Bogue (Boops boops) quality loss during
storage, based on the amino acids degradation (Ciampa et al., 2012). The research work
highlighted how the concentration of some free amino acids increases in the fish flesh
because of enzymatic reactions, during the first days of storage, and due to bacterial
development afterwards; in fact, the storage temperature seemed to mainly affect bacterial
development rate, modulating the amino acid concentrations starting from day 4 of storage.
The amino acids profile was shown to be sensitive to the phenomena leading to the
compositional changes occurring during fish storage (Ciampa et al., 2012). Other research
papers confirm the 1H-NMR as suitable technique for freshness and quality monitoring:
combined with multivariate statistical analysis, it forms the basis for an advanced
diagnostic route to check the quality of fish. Picone et al. found that free amino acids and
other low molecular weight molecules, quantified by 1H-NMR, allowed classification of
farmed Gilthead Seabream (Sparus aurata) according to the aquaculture system employed
(Picone et al., 2011). The obtained results of the investigation show that the fish
metabolome accurately reflects the rearing conditions. Many metabolites covariate indeed
with the rearing conditions and a few metabolites including glycogen (stress indicator),
histidine, alanine and glycine display significant changes dependent on the aquaculture
system and on the storage times.
Different rearing conditions and dietary of wild and cultured see bass resulted in clear
separation of their metabolite profiles monitored by 1H and 13C NMR spectroscopy
(Mannina et al., 2008). The identified metabolites included water-soluble ones belonging
to different classes such as sugars, amino acids, dipeptides and organic acids as well as
liposoluble metabolites such as lipids, sterols and fatty acids. The content of diunsaturated
fatty acids in cultured samples was about six times as high as that in wild fish, whereas
cultured fish resulted to be poor in polyunsaturated fatty acids, especially in
docosahexaenoic and eicosapentaenoic acids. Among the other metabolites, higher content
of cholesterol, phosphatidylethanolamine and TMAO and lower level of some water-
soluble metabolites such as choline, glutamine, fumaric and malic acids were observed in
wild fish with respect to cultured see bass.
Recently, the development of in vitro digestion protocols gives the opportunity to
evaluate also metabolites’ release and protein hydrolysis in cooked fish, as demonstrated
by Vidal et al. who applied 1H NMR to study micro-wave cooked European sea bass
(Dicentrarchus labrax) fillets, along different points of their in vitro digestion: after oral,
gastric and subsequent duodenal digestion, pointing to a different digestibility between
farmed and wild animals (Vidal et al., 2016).
In summary, NMR could become a useful tool for inspective authorities in terms of
authenticating food labeling of fish products (e.g., production method, catchment’s
area/fish farm of origin, nutritional value, storage and processing methods). The benefit of
NMR to aquaculture industries could be to assess the health status of the fish or to
investigate the effects of breeding conditions (for example feed, temperature, medication,
pollutants) and nutritional quality (lipids profile for instance) of the final product
depending on storage and processing conditions.
2.5.3. Algae
High-resolution 1H and 13C NMR has been used to investigate the metabolome of
Aphanizomenonflos-aquae (AFA), (Righi et al., 2016) a unicellular microalga that
spontaneously grows in Upper Klamath Lake and is widely employed as dietary
supplement. AFA can be found on the market as tablets, capsules and dry powder. Samples
suitable for NMR analysis can be obtained from ground tablets or from dry powder after
suspension in D2O (Figure 19). In this last case, better spectral resolution is obtained
(Figure 19a) and spectral interference from added ingredients (e.g., pea fiber) is avoided.
Probes for liquids were employed for AFA study, because no appreciable improvements
were observed on spectra acquired with HR-MAS probes.
Figure 19. Water-presaturated 1H NMR spectra of AFA samples obtained from: a) commercial dry
powder (Farmalabor s.r.l., Canosa di Puglia, Italy) suspended in D 2O; b) ground tablet (Eos s.r.l. Musestre
di Roncade, Treviso, pea fiber addition is declared).
1
H and 13C NMR spectra of AFA show a plethora of signals that reflects the very
complex metabolome of these cyanobacteria. Both broad and narrow components are
observed in the water-presaturated 1H NMR spectrum (Figure 19). The broader ones are
due to species slowly tumbling, usually high molecular weight molecules or adducts,
whereas the narrow signals derive from small species free to move in solution. NMR
proved to be very helpful in the identification of some interesting secondary metabolites
(Figure 20) such as mycosporine-like amino acids, porphyra-334 and shinorine, and low
molecular weight glycosides, glyceryl β-D-galactopyranoside and a glyceryl 6-amino-6-
deoxy-α-D-glucopyranoside. Furthermore, a new species, cis-3,4-dehydrolysine, was
detected for the first time. These identifications were not based on data present in NMR
databases, but were rather accomplished using the correlations found in homonuclear and
heteronuclear 2D spectra (COSY, TOCSY, NOESY, ROESY, HSQC, HMBC and HSQC-
TOCSY) to reconstruct the molecular skeletons. The majority of these assignments were
confirmed by mass-spectrometry.
The signals of the above molecules overlap those of other 50 metabolites or residues:
amino acids, oligopeptides, osmolytes, mono- and polysaccharides, glycosides, organic
acids, nucleobases. The chance of success in the unambiguous identification of a species
in such crowded spectra is strongly related to the presence of at least one not overlapped
signal, giving clear correlations with other nuclei in 2D spectra. In this respect, TOCSY
experiment (Figure 21) proved to be very effective in highlighting intriguing spin systems
and 1H-13C HSQC-TOCSY in extracting proton spin systems starting from HSQC proton-
carbon correlations. However, AFA metabolome is so complex that this work can only be
considered the starting point for further studies, in which NMR can be flanked by mass-
spectrometry.
The same approach can in principle be applied also to other algae used as dietary
supplement, as Spirulina (Arthrospira Platensis), even though in this case the stability of
the powder suspension in the NMR tube and the signal line shape were less satisfactory
(Figure 22) and this prevented us in undertaking such a complex study. The hydration
process of dry spirulina alga seems to compete with alteration processes.
Figure 22. Water-presaturated 1H NMR spectra of AFA samples (left) and Spirulina samples (right)
obtained on the day of preparation (above) and after one week (below).
2.5.4. Bottarga
The eviscerated roes of striped mullet (Mugil cephalus) are manufactured in several
countries, and the salted and dried product can be found worldwide under different names
and typologies. In the Mediterranean Sea, Sardinia Island (Italy) has a long tradition in
manufacturing mullet roes to obtain a product called “bottarga.” Bottarga has become so
increasingly popular in international markets as a delicacy for appetizers or as a pasta
seasoning that mullets of the Mediterranean Sea are not enough to satisfy the request of
this product. Consequently, the Sardinian producers must turn their attention to other
fishing areas located in different regions of the globe for roe supplies among which FAO
31 (central-western), 34 (central-eastern), and 41 (southwestern). Even if the raw material
is not necessarily original to the island, Sardinian bottarga has its own peculiar rheological
and organoleptic profile due to the skills of the local producers, inheritors of an ancient
tradition in processing this delicacy. For this reason, Sardinian manufacturers of bottarga
are requesting a Protected Geographical Indication (PGI) designation for this product.
The lipid profile of mullet raw roes has been extensively investigated by NMR
spectroscopy, providing useful information on the compositional changes occurring upon
processing and storage (Scano et al., 2008, Scano et al., 2009; Rosa et al., 2009, Rosa et
al., 2012). Besides, three metabolomics studies were focused on the water-soluble
metabolites. The first one was designed to investigate the metabolome of bottarga in
relation to the geographical origin of mullet. To this aim, the metabolic profile of 25
2.6. Meat
In recent years, interest in meat quality and authenticity has increased. Many
consumers are concerned about the meat they eat and accurate labeling is important to
inform consumer choice. So there is the need for new, rapid and reliable analytical
methodologies and easily quantifiable markers to be used for meat authentication. Fields
of particular interest are meat origin (meat cuts, breed, feed intake, slaughter age, wild vs.
farmed meat, organic vs. conventional meat, and geographical origin), meat substitution
(meat species), meat processing treatment (irradiation., fresh vs. thawed meat and meat
preparation) and non-meat ingredient addition (Ballin, 2010). Analytical methods used in
meat authentication include a diverse range of equipment and techniques. Regarding NMR,
the most widely explored area in meat science is traditionally 1H relaxometry, due to its
potential in characterizing water and structure in heterogeneous systems like muscle/meat
(for example water-holding capacity), together with dynamic studies of physical changes
during conversion of muscle to meat. NMR imaging studies have also been frequently
applied in the elucidation of structural changes and brine dynamic during meat processing,
as freezing, curing and high pressure treatment. 31P NMR spectroscopy has been applied
to investigate the effect of genetic, feeding, and slaughter factors on meat quality through
their action on phosphorous metabolism post mortem. All these aspects were reviewed by
Bertram et al. (Betram & Andersen, 2004). In the present paragraph, only the applications
related to the use of high resolution NMR based metabolomics in meat authentication,
quality control and processing monitoring will be highlighted.
the geographical origin of beef samples, showing significant separation among beef
originating from four countries: Australia, Korea, New Zealand, and the United States. The
major metabolites responsible for differentiation were succinate and various amino acids
including isoleucine, leucine, methionine, tyrosine, and valine, which can be evaluated as
possible biomarkers of origin (Youngae et al., 2010). HR MAS 1H NMR was also applied
for the selection of potential molecular markers of specific quality or geographical origin.
Dried beef samples of certified origin were tested, concluding that fat content as well as
specific metabolites, probably linked to feeding system, are shown to be good candidates
for markers of origin (Shintu et al., 2007).
The same approach was applied on lamb meat from three areas located in Apulia
(Southern Italy) and 1H HR-MAS NMR spectra were registered directly on the minced
solid sample, acquiring information on a large number of metabolites (amino acids, fatty
acids, sugars, etc.). The application of multivariate statistical analysis to a data set
containing 1H HR-MAS NMR spectral data together with isotope ratios provided a
satisfactory origin differentiation of lamb meat samples (Sacco et al., 2005).
2.6.4. Breeds
Some NMR-based metabolomics studies were conducted to identify different
metabolite profiles in animal breeds, both in pigs and cows. For example, meat extracts
from five different pig crossbreeds including Duroc/Landrace/Yorkshire, Iberian/Duroc,
Iberian/Duroc/Landrace, Mangalitza/Duroc, and Mangalitza/Landrace/Yorkshire were
analyzed by NMR based metabolomics. Amino acids including alanine, carnosine,
isoleucine, methionine, phenylalanine, and valine, as well as lactate, inosine
The effect of gamma-ray irradiation on the fatty acid profile of beef meat was examined
at doses of 2.5, 5.0, 7.5, 10.0 and 15.0 kGy by means of 1H NMR spectroscopy. NMR
results revealed a clear trend toward a decrease in the amount of polyunsaturated fatty acids
with increasing the irradiation dose (Stefanova et al., 2011).
A metabolomic approach combining 1H NMR lipid profiling with chemometric
procedures, as stepwise linear discriminant analysis (sLDA) and artificial neural networks
(ANNs), provided a successful discrimination between the groups of irradiated and non
irradiated beef meat. sLDA allowed the classification of 100% of the samples into
irradiated or non-irradiated beef groups; the same result was obtained by ANNs using the
1 kGy irradiation dose as discriminant value suggested by the network. Furthermore, sLDA
allowed the classification of 81.9% of the beef samples according to the irradiation dose
(2.5, 4.5 and 8 kGy) (Zanardi et al., 2013).
Beside lipid fraction, also 1H NMR analysis of polar metabolites was found useful to
discriminate beef meat according to irradiation (0, 2.5, 4.5 and 8 kGy). Classification trees
revealed that three metabolites (glycerol, lactic acid esters and tyramine or a p-substituted
phenolic compound) are important biomarkers for classification of the irradiated and non-
irradiated beef samples. The use of the NMR-based approach simplifies sample preparation
and decrease the time required for analysis compared to available official analysis
procedures (Zanardi et al., 2015).
2.7. Fruits
Fruits are an essential part of human alimentation, since they not only represent a
source of energy in the form of simple and complex carbohydrates, but also because they
provide a wealth of nutrients such as vitamins, fibers and polyphenols. They can be
consumed as it is, as in the case of fresh fruits eaten at the end of a meal, or they can be the
raw material for more elaborate preparations such as confectionery or even, as it is for
shelled fruits, roasted and employed as finger food. The great availability and valuable
contribution to human health make fruit one of the most diffused crop in the world as is
proven by the large amount of varieties present on the market, such as apples (Malus),
bananas (Musa), kiwi fruits (Actinidia), peaches (Prunus persica), grapes (Vitis vinifera),
pears (Pyrus), hazelnuts (Corylus avellana), pistachios (Pistacia vera), almonds (Prunus
dulcis) and so on. The molecular basis of their benefits lies in their chemical composition,
which is influenced by several factors such as the genetic one since, for example, the color
of the peel of apples, and thus their phenol content, varies according to the examined specie
or cultivar. Other important factors are the climatic and geographic conditions on which
the aliment is grown, since soil composition, average precipitations, temperature and
humidity can heavily affect the production of amino acids, carbohydrates and secondary
metabolites. Among the other factors influencing fruit chemical composition, the ripening
stage is one of paramount importance since metabolites with a protective activity against
external pathogens decrease their concentration as the fruit matures. This aspect is
important since many molecules with such activities are potent antioxidants, such as
ascorbic acid, caffeic acid and its esters, and thus can be useful to human health. Several
spectroscopic techniques can be employed to evaluate the composition of foods, yet
Nuclear Magnetic Resonance (NMR) is the leading one because it can provide both
qualitative (i.e., can identify the species in a sample) and quantitative information with the
employment of bidimensional experiments such as COSY, TOCSY, HSQC and HMBC.
therefore, the harvest time could be adjusted taking into account the potential effect on
consumer health and the presence of molecules, such as pectines and polysaccharides,
which could have industrial relevance. Ripening time is also a function of the fruit cultivar,
as observed for pistachios (Sciubba et al., 2016). In this study the ripening rate of two
pistachio cultivars, namely “Bianca” and “Gloria” was assessed by HR-NMR and it was
observed that molecules such as shikimic acid, gallic acid and their esters sharply decrease
a month before commercial maturity. Since these molecules possess many beneficial
properties, this information is of great importance for food processing strategies.
2.7.5. Cocoa
Cocoa beans represent the seeds of the fruits of cocoa tree (Theobroma cacao L., order
Sterculiacae) growing in areas about 20°latitude north and south of the Equator. Three main
variety of cocoa exist: Forastero (bulk cocoa, 85% of the world cocoa production), Criollo
(fine aromatic cocoa, 5% of the production) and a cross hybrid of the previous two,
Trinitario. Cocoa beans are transformed in cocoa powder or chocolate by a multi-step
process performed in part in the origin countries (beans fermentation and drying) and in
part in the developed countries (roasting, dutching, conching, tempering) (Caligiani et al.,
2016). Due to the complexity of the whole process, the variability of the raw material and
of fermentation process, NMR approach can face different issues in cocoa bean
characterization:
The definition of botanical origin is important due to the different commercial value of
fine and bulk cocoa. Principal component analysis applied to 1H NMR dataset of cocoa
polar extracts showed that Arriba (fine cocoa, Forastero sub-specie native from Ecuador)
and Criollo beans, form groups separated from those of cocoa beans of Forastero varieties
(Caligiani et al., 2014b).
Also the assessment of cocoa geographical origin is growing in importance due to the
increasing market of mono origin chocolate. A first NMR method aimed at the
classification of cocoa beans according to the geographic origin was developed and
patented in 2005, by applying a multivariate statistical analysis of heterocorrelated
bidimensional NMR experiments (HMBC) of polyphenolic extracts (Soulet et al., 2005).
Further studies showed that 1H NMR and HR-MAS 1H NMR metabolomics approach
combined with chemometrics were effective in discriminating cocoa coming from different
continents (America, Africa, Asia/Oceania), but not single countries (Marseglia et al.,
2016; Caligiani et al., 2010).
The difficult to define the cocoa origin arise also from the effect of fermentation level,
which was proven as a factor influencing cocoa metabolome more than botanical or
geographical factors. In fact a study applying PCA to cocoa bean 1H NMR dataset showed
that well fermented brown beans form a group clearly separated from unfermented, slaty,
and underfermented, violet, beans, independently of the variety or geographical origin; as
a consequence, 1H NMR was a good method to evaluate cocoa fermentation level, which
is currently determined by visual inspection of bean color (Caligiani et al., 2014b).
The metabolomics-quantitative approach permitted also the simultaneous detection
and quantification of all the main classes of cocoa metabolites as amino acids,
polyalcohols, organic acids, sugars, methylxanthines, catechins, and phenols (Caligiani et
al., 2010).
NMR was also used in cocoa characterization for the structural identification of new
cocoa metabolites as C-glycosylated flavan-3-ols (Stark & Hoffman, 2006), N-
phenylpropenoyl-L-amino acids (Stark & Hoffman, 2005).
Finally, a widespread use of pulsed NMR can be found in cocoa butter characterization,
in particular to detect the amount of crystallized solids present in cocoa butter (Haeupler et
al., 2014) or in cocoa butter replacers (Jahurul et al., 2014).
homo- and heteronuclear NMR experiments (CPMG, LED, COSY, TOCSY and HSQC)
and more than fifty metabolites were found.
1
H NMR spectra (Figure 23) clearly show differences between lemon and citrus
flavedo: the former is richer in terpenes, the latter in carbohydrates and waxes. A very
similar composition is instead found for the oil gland content (spectra not shown), where
nine constituents (limonene, γ-terpinene, β-pinene, α-pinene, geranial, neral, citronellal,
myrcene, sabinene, α-thujene, nerol and geraniol esters) were identified through 1D and
2D NMR experiments and quantified through 1H NMR. In this case, HR-MAS probe was
used as a nano-probe for the acquisition of spectra directly on 40 μL of oil-water emulsion,
obtained cutting the most superficial layer of flavedo and pressing gently on the surface.
Figure 23. 1H HR-MAS NMR water-presaturated spectra of (a-d) lemon and (e-h) citron: (a, e) flavedo,
(b, f) albedo, (c, g) pulp and (d, h) seeds.
1
H NMR spectra of albedo from both fruits are dominated by carbohydrate signals,
mainly glucose, fructose and sucrose, whereas low signals from terpenes and free amino
acids (asparagine, alanine, threonine, valine, glutamine and proline) are detected. No
signals from flavonoids were observed in 1D and 2D NMR spectra.
The main components of the pulp of both fruits are citric acid and sugars, but also
amino acids (asparagine, alanine, proline, glutamine and γ-aminobutyric acid), polyols
(myo-inositol and scyllo-inositol), stachydrine, malic acid and small amounts of
trigonelline (especially in citron pulp) are present.
2.7.7. Kiwifruits
Kiwifruit has become a valuable horticultural crop in international food markets.
Fragrance, flavor, and healthful properties are attributes which depend on the fruit chemical
composition. Specifically, the flavor of the fruit flesh depends on the balance between
soluble sugars and non-volatile organic acids. Different sugars are responsible of different
levels of sweetness while the perception of acidity depends on the level of organic acids.
Consequently, the analysis of kiwifruit metabolites and the knowledge of the nutritional
profile may help the marketing. Besides, the knowledge of the metabolite profile at
different stages of development is important for determining the most suitable time of
harvesting and for industries interested in extracting specific compounds from the fruit to
obtain foodstuff additives or substances of pharmaceutical interest. In fact, the
concentration of these compounds depends on the stage of development. The metabolite
profile of Hayward, Zespri, and CI.GI. kiwifruits aqueous extracts have been investigated
by high field NMR (Capitani et al., 2010; Capitani et al., 2013a), the profile was monitored
over the season (June-December). Metabolites belonging to different classes such as
organic acids (malic, citric, quinic, ascorbic, shikimic, and lactic acid), amino acids
(alanine, glutamine, threonine, arginine, glutamate, asparagine, aspartate, valine, leucine,
isoleucine, lysine, triptophane, -aminobutyrate phenylalanine, and histidine), and
carbohydrates (glucose, galactose, xylose, myo-inositol, sucrose, raffinose, fructofuranose,
-D-fructopyranose, fructose-6-phosphate and glucose-6-phosphate) were identified. Also
the presence of other metabolites such as orthophosphate, adenosine, choline, uridine, 3-
O--D-glucopyranosyl-trans and cis-caffeic acids, 3-O-α-L-rhamnopyranosyl quercetin,
neochlorogenic acid, and catechin, was detected. The metabolite profile of the three
kiwifruits cultivar displayed common features, such as high levels of quinic, citric, and
ascorbic acid, but also differences during the growth and the development of fruits.
For example, in Hayward and CI.GI. the highest concentration of some amino acids such
as arginine and alanine was found in the early stages of development and then decreased.
A different trend was observed in the case of Zespri, in fact arginine increased over the
season. The concentration of arginine, glutamine, threonine, asparagines, and alanine was
always higher in Zespri than in Hayward and CI.GI. As expected, in the three cultivars, the
content of sugars such as sucrose, glucose and fructose increases during fruit growth,
reaching the highest value in December. With regard to secondary metabolites, a high level
of 3-O--D-glucopyranosyl-trans caffeic acid was found in Hayward and CI.GI. kiwifruits
harvested in July and August, whereas in Zespri Gold the level of this metabolite was low
at any stage of development. A rather high level of epicatechin was found in Hayward and
CI.GI. kiwifruits harvested in August while in Zespri Gold the level was always low. A
high level of quercetin-3-rhamnoside was detected in Hayward and GI.CI. kiwifruits
harvested in June, July, and August, whereas in Zespri this metabolite was not present at
all or was under the limit of detection. A low level of neo-chlorogenic acid was detected
only in Zespri Gold over the season. PCA analysis was applied to the intensity of selected
NMR signals. Figure 24a shows the mean and standard deviation PCA scores at each stage
of development relative to the first two principal components obtained from the whole data
matrix. PC1 is related to the developmental stage of the fruit, whereas PC2 is related to the
cultivar. By observing the evolution of the scores along PC1, it can be seen that Zespri
Gold metabolite profile always precedes the Hayward and CI.GI. counterparts, whereas
Hayward and CI.GI. profiles are substantially coupled along both PC1 and PC2. This
behavior is in agreement with the common origin of Hayward and CI.GI. which are both
varieties of Actinidia deliciosa. To underline differences between the two cultivars, PLS2
analysis was carried out on Hayward and CI.GI. samples using the same variables used for
PCA analysis. Figure 24b shows the means and standard deviations of the scores of the
first two out of our significant components of the PLS2 regression model between Hayward
and CI.GI. kiwifruits. The score plot obtained shows that, with the exception of June and
July, data relative to CI.GI. and Hayward are not separated along the stage of development
(PC1), whereas they are well separated along PC2. The most relevant metabolites
responsible for the separation, namely arginine, choline, myo-inositol, quinic acid, sucrose,
uridine and histidine, were indicated by their regression coefficients of the PLS2 model
with the cultivar. Lower levels of arginine, myo-inositol, uridine, and histidine were found
in CI.GI. than in Hayward, whereas higher levels of quinic acid, choline, and sucrose were
found in Hayward kiwifruits.
Figure 24. a) Score plot of means and standard deviations obtained from the PCA analysis applied to the
intensity of selected 1H resonances monitored as a function of the developmental stage in Zespri Gold,
Hayward and CI.GI kiwifruits. b) Score plot of means and standard deviations obtained from the PLS2
analysis applied to the intensity of selected 1H resonances in Hayward and CI.GI. kiwifruits.
2.7.8. Peach
The metabolite profile of aqueous extracts of two peach varieties, Percoca Romagnola
7 and Flaminia, with different susceptibility to Ceratitis capitata attack has been studied
by high field NMR spectroscopy (Capitani et al., 2013b; Sobolev et al., 2015). The study
was carried out for investigating possible differences in their profiles which could account
for the different susceptibility to insect attack exhibited by the two varieties. Water soluble
metabolites belonging to different classes such as carbohydrates (fucose, fructose, fructose-
6-phosphate, glucose, glucose-6-phosphate, rhamnose, sucrose, xylose, and myo-inositol),
organic acids (citric, fumaric, malic, quinic, shikimic, and succinic acid), and aminoacids
(alanine, asparagine, isoleucine, threonine, and valine) were identified. Besides, also other
metabolites such as choline, trigonelline, catechin, chlorogenic and neo-chlorogenic acid,
orthophosphate, and α-L- glycerophosphorylcholine, were detected. PCA analysis applied
to the intensity of 29 selected 1H resonances for comparing the profiles of the two varieties,
well grouped Percoca Romagnola 7 with respect to Flaminia. The first two PCs accounted
for 52.9% of variability within the data, PC1 accounting for 35.5% and PC2 for 17.4%.
The metabolites mostly responsible for the grouping were found to be valine, isoleucine,
glucose, xylose, myo-inositol, choline, alanine, chlorogenic and neo-chlorogenic acids,
quinic and fumaric acids, sucrose, and -fucose. The pulp of Percoca Romagnola 7, the
more resistant variety, exhibited greater amounts of alanine, quinic acid, chlorogenic and
neo-chlorogenic acids, than Flaminia. These compounds are reported to be related to the
defense against fungal and insects attack in other members of Rosaceae family, indicating
that the phenypropanoid pathway is at least partially involved in the repulsion of Ceratitis
capitata. Moreover, whereas valine and isoleucine derived volatiles are mainly involved in
fruit aroma; alanine is also a precursor of compounds involved in plant defense. According
to this observation, Percoca Romagnola 7 variety, the more resistant and less appealing
variety, showed a lower amount of valine and isoleucine along with a significantly higher
level of alanine compared to Flaminia. On the basis of these literature data it is possible to
underline the importance of the principal components. PC1, with its emphasis on defense
metabolites, describes the capacity to repel insects’ attacks, whereas PC2, because of its
higher levels of flavor precursors and free carbohydrates, is an index of good sensorial
properties.
2.7.9. Blueberry
High field NMR spectroscopy was applied for a detailed blueberry characterization.
Specifically, untargeted metabolite profile of blueberry aqueous and organic extracts as
well as targeted analysis focused on anthocyanins and other phenols, have been reported
(Capitani et al., 2014). Water soluble metabolites belonging to different classes such as
carbohydrates (glucose, sucrose, fructofuranose, fructopyranose, and myo-inositol), amino
acids (alanine, arginine, asparagine, -aminobutyrate, glutamine, glutamate, valine, and
leucine), organic acids (malic, citric, quinic and shikimic acid), and other metabolites such
as chlorogenic acid and choline, were identified. In the organic extracts triglycerides, fatty
acids and -sitosterol were detected. Five anthocyanins, namely malvidin-3-glucoside,
malvidin-3-galactoside, delphinidin-3-glucoside, delphinidin-3-galactoside and petunidin-
3-glucoside, together with 3-O-α-L-rhamnopyranosyl quercetin and chlorogenic acid, were
identified in solid phase extracts. This combination of untargeted metabolite profile and
targeted NMR analysis focused on phenols was proposed to monitor comprehensively the
nutritional-nutraceuticals properties of different blueberry varieties.
2.7.10. Tomato
Tomato (Solanum lycopersicum) is one of the most important crop whose production
is increasing linearly over the years and has doubled since 1993 to 2013 reaching 1.6 x 108
tons, see FAO stat website. There are many varieties of tomato and it is consumed in many
different ways: fresh vegetable, dehydrated tomato, tomato sauce, tomato paste, tomato
soup, tomato powder, etc. In particular, fresh tomatoes must have an acceptable flavor
given by the combination of its main constituents (sugars, organic acids, free amino acids
and salts) to satisfy the consumers demand. Indeed, tomato is important both from an
economical point of view and for its relevant nutritional value. In fact, tomato is low in
calories, is a good source of antioxidants and possesses antitumoral, and antidepressive
properties (Basu & Imrhan, 2007).
Antioxidants properties are mainly guaranteed by the presence of carotenoids
(lycopene and β-carotene), polyphenols (flavonoids and hydroxycinnamic acids) and
vitamins (ascorbic acid, vitamin E). For these reasons, tomato is always present in healthy
diets.
The NMR technique has been widely applied to study many aspects of tomato, starting
from the characterization of the plant, to the metabolic profiling of the fruit, to the quality
assessment of greenhouse-grown tomato, to the geographical identification of protected
species, etc (Ishida et al., 1989; Le Gall et al., 2003; Sobolev et al., 2003; Deborde et al.,
2009; Consonni et al.; 2009, Ferreres et al., 2010; Pérez et al., 2010; Mallamace et al.,
2014).
The first NMR attempt on tomato fruits was performed in 1989 by means of Magnetic
Resonance Imaging. Ishida et al. measured the distribution and the relaxation times of
water in tomato fruits and they were able to distinguish the physiological variations among
different types of tissues and the physiological changes during maturation of tomato fruit
(Ishida el al., 1989). In Figure 25, the differences in the water distribution among various
tissues in tomato fruits from green immature to red mature conditions obtained by Ishida
et al. are reported. Note that the amount and the mobility of water are strictly related with
the biological reactions in enzymatic and metabolic processes in tissue cells.
Figure 25. Images of tomato fruits acquired by means of the saturation recovery method with a short
repetition time of 0.5 s. Figure adapted from Ishida et al., 1989.
Lately, Sobolev et al. in 2003 performed the first study at high field (600 MHz) on
tomato juice and pulp. By means of bidimensional pulse sequences, they were able to
assign each spin system and then they calculated the relative molar ratios of the main
components by peaks integration.
In Figure 26 two expansions of the bidimensional pulse sequences used to assign the
spectral components for the juice of a particular tomato variety known as Red Setter is
reported. In particular, on the right side it is represented the expansion of the high-field
region of the TOCSY and, on the left side the expansion of carbohydrates region of 1H–
13
CHSQC spectrum (Sobolev et al., 2003). For more details and the complete assignment
of the NMR spectrum of tomato juice, see the work by Sobolev et al., 2003.
More recently, the new challenges include the discrimination between different tomato
products (both fresh and paste for examples) for the geographical identification and further
for the determination of some biomarkers that can represent the fingerprint of such a
product.
For example, High-Resolution techniques such as Magic Angle Spinning, and
chemometric analyses such as the PCA, were applied in order to determine the metabolic
profile of some protected varieties as that of the cherry Tomato of Pachino (Mallamace et
al., 2014; Corsaro et al., 2015; Corsaro et al., 2016). The Sicilian cherry tomato of Pachino
was the first tomato accredited by the European PGI certificate, guaranteeing its
geographical origin and quality. In the mentioned study, the concentration of some
metabolites was found to be discriminant for the characterization of the variety. In fact,
from the score and loading plots of the PCA, some metabolites were identified to have a
significant concentration that was able to characterize the studied variety with respect to
others. In details, the concentration of glucose, fructose, GABA (aminobutyric acid),
glutamic acid, trigonelline, tryptophan and tyrosine was higher in Pachino cherry tomatoes
whereas that of alanine, guanosine, and methanol was higher in non-Pachino ones
(Mallamace et al., 2014; Corsaro et al., 2015). Indeed, the combination of these values
conferred the particular taste and high nutritional quality for this precious variety of tomato.
Figure 26. Expansions of two different parts of the spectrum of the Red Setter tomato juice. Left side:
carbohydrates region of 1H–13C HSQC spectrum where the indicated signals refer to 1: Methanol; 2:
Choline; 3: Asparagine; 4: Aspartate; 5: Glutamine and glutamate; 6: Threonine; 7, 15, 22, 23, 30: -D-
glucose; 8, 15, 18, 19, 20, 29: -D-glucose; 9, 11, 21, 24, 26: -D-fructofuranose; 10, 12, 13, 14, 17: -
D-fructopyranose; 16: Malate; 25, 27, 28: -D-fructofuranose. Right side: the high-field region of the
TOCSY experiment that reports some of the assigned compounds. Figure adapted from Sobolev et al.,
2003.
Figure 27. PCA score plot performed considering a total of 47 triple concentrated tomato paste samples
(harvest 2007): diamonds and dots represent Chinese and Italian samples respectively. PC1 = 38.0%, PC2
= 21.9%. R2X =79.6%, and Q2 = 56.0%.
Figure 28. PCA score plot performed by considering 92 double and triple concentrated tomato paste
samples of certain origin and production year: open and filled diamonds represent Chinese samples
produced in 2007 and 2008 respectively while open and filled dots represent Italian samples produced in
2007 and 2008 respectively. PC1= 42.2%; PC2=21.4%; PC3=10.3%. These first three PCs accounted for
R2X = 74% and Q2 = 67.4%. (A) PC1 versus PC2 and (B) PC1 versus PC3.
The scores of PCA were used to select training and test sets with D-Optimal Onion
Design protocol and an OPLS-DA model was successively performed on the training set
samples obtaining a correct classification for all samples analyzed. The corresponding S-
plot highlighted citrate as the most discriminant variable characterizing Chinese samples,
while sugars (glucose and fructose) resulted to be characteristic for the Italian ones.
Because of the possible use of citrate for pH tomato paste correction and bacteria growth
inhibition (even if not allowed by law), the discrimination of samples based on this variable
resulted biased. By excluding the citrate contribution, a new OPLS-DA was performed
obtaining again a good classification for all samples according to the provenience. Italian
samples showed again a higher sugar content while the Chinese samples resulted
characterized by higher levels in aspartic acid and glutamine. The same group successively
investigated the influence of the production year on the geographical discrimination of
Italian and Chinese triple and double concentrated tomato paste samples (Consonni et al.,
2010). A total of 119 samples produced in 2007 and 2008 were analyzed by 1H NMR and
data handled by PCA and bidirectional Orthogonal Projection to Latent Structures
Discriminant Analysis (O2PLS-DA). This latter model is well-suited for noisy reduction
and uncorrelated variables removal, allowing to obtain robust classification models and a
clear interpretation of the systematic variation useful to characterize each class. Among
119 samples, 92 were of known origin (candidate set) while the remaining 27 (validation
set) were directly bought on the market. The PCA performed on the 1H NMR spectra
bucketed (excluding the citrate contribution) of the 92 lyophilized samples, led to obtained
a clear discrimination of samples according to the geographical origin with the first two
PCs (Figure 28) while the third component routed a samples discrimination according to
the production year (Figure 28). This funding suggested that even though some variables
contributed for the discrimination of samples according to the year of production, the
information related to geographical origin resulted stronger.
A confirmation was obtained by a deeper analysis of the orthogonal space and residuals
of O2PLS-DA model performed on the selected training set from the 92 samples obtaining
a clear sample separation according to the production year irrespectively of the provenance
performing a PCA analysis while the predictive part of the O2PLS-DA model was not able
to distinguish samples according to the production year. Moreover, the O2PLS-DA model
performed considering only samples of 2007 and the O2PLS-DA performed by considering
samples of know origin of 2007 and 2008 indicated the same discriminant variables
suggesting that the two models were comparable from a qualitative point of view.
Therefore, the prediction of origin of both candidate and validation samples sets was
evaluated considering the O2PLS-DA model performed only on samples produced in 2007
obtaining a correct classification for more than 95%.
2.8. Coffee
Coffee constitutes the second most traded commodity in the world and the largest
consumed beverage after water. It is mainly produced in the so-called “coffee bean belt,”
a well-defined geographic area, comprising Central and South America, Africa and
Middle-Southeast Asia between the “Tropic of Cancer” and the “Tropic of Capricorn.” The
pedoclimatic characteristics, such as high altitudes, moist tropical climate, rich soil, and
perfect temperatures strongly affect the organoleptic peculiarities leading to high quality
coffee. The economic value is correspondingly affected by variety and provenance; in this
view, objective analytical tools aimed to investigate and to determine the origin would
largely benefit the quality determination assessment. Nowadays, the geographic origin of
coffee is evaluated by the “in cup taste testers,” whose response could not be shared
overwhelmingly. In these last years, a large number of different analytical approaches
appeared, focused mainly on coffee quality assessment, and specifically on the detection
of adulteration practices that could involve the declared geographic origin, the blend
composition in terms of coffee variety, or the addition of cheaper components.
Roasted coffee, like other food matrices, represents a complex mixture of several
compounds, and is an ideal subject for NMR investigations. The first NMR work on roasted
coffee dealing with “metabolite profiling” was published in 2011 (Wei et al., 2011); it used
13
C-1H two-dimensional heteronuclear spectra and provided a large resonance assignment
database. Further studies from the same group were performed monitoring 30 components
upon roasting process (Wei et al., 2012). Their results indicated caffeine and myo-inositol
as relatively thermally stable, and multivariate data analysis indicated that some
components such as sucrose, chlorogenic acids, quinic acids, and polysaccharides could
serve as chemical markers during coffee bean roasting. Their results suggest useful
chemical markers to control and characterize the coffee-roasting process.
Following the “metabolomic” approach performed to study several foods, Consonni’s
group performed some studies on quality assessment of roasted coffee by NMR and
multivariate statistical analysis (Consonni et al., 2012b). Specifically, Coffea arabica
samples from the main production areas of the world, i.e., Africa, Asia and America, were
analyzed on the basis of their metabolic profile of aqueous extracts. The 1H NMR spectra
of water extracts of coffee represent a complex pool of metabolites, contain several classes
of chemical compounds, such as esters, lactones, aromatics, amides etc, and are dominated
by large amounts of chlorogenic acids (present in different isoforms of caffeoyl/feruloyl-
quinic acids), caffeine, trigonelline and few organic acids. Roasting conditions
(temperatures and periods) would strongly affect isomerization and degradation processes
of these chemical moieties. Notwithstanding different roasting conditions, 40 samples of
coffee from America, Asia, or Africa were successfully characterized according to their
origins by using a three classes OPLS-DA model (Figure 29).
For a deeper understanding of the role played by the metabolites as responsible for the
geographical differentiations, series of two classes OPLS-DA models were further
investigated, and these latter models were validated by using training and test sets selected
on the basis of D-optimal onion design (Olsson et al., 2004). The corresponding S-plots
indicated fatty acids for American, acetate and trigonelline for Asian and finally
chlorogenic acids and lactate for African as the responsible metabolites for sample
differentiation.
NMR spectroscopy showed its potentiality in quantitative determinations as well. In
the case of coffee, more than 100 different species are known, but from the commercial
point of view, Coffea arabica L. and Coffea canephora var. robusta, (commonly known as
arabica and robusta, respectively) represent the two most relevant and widely cultivated
species, accounting for 56% and 44% of the world’s production in 2011, respectively,
reaching 134 millions of bags (FAOSTAT source).
Figure 29. (A) and (B) Score and loading plots of OPLS-DA performed by considering all roasted coffee
samples: filled triangles, dots and diamonds represents American, Asian and African roasted coffee
samples respectively. R2Y = 81.5% and Q2 = 69.7%. (Reprinted with permission from Consonni et al.,
2012b. Copyright (2012) Elsevier).
The price gap between the two species is significantly different in favor of arabica and
as a consequence of this a growing financial incentive to unlawfully replace of high quality
arabica coffee with the cheaper robusta is increased in these last years. The requirement of
a reliable analytical method able to evaluate the ratio between the two species in coffee
blends is therefore economically very important. Several attempts have been performed
approaching arabica vs robusta differentiation with different analytical techniques. Some
of them have as a drawback the sample preparation/pretreatment or the extraction
procedure that could impair the reproducibility and the speed of the determination.
Conversely, NMR spectroscopy is highly reproducible and rapid, being characterized by
null sample preparation avoiding derivatization or purification procedures. The success of
NMR determination in investigating arabica and robusta roasted coffee blends has been
shown in a recent paper (Cagliani et al., 2013) where the compositional evaluation of the
blends was determined by applying 1H–NMR spectroscopy to water extracts of blends,
enabling the detection of different classes of chemical compounds simultaneously, with a
single experiment (Figure 30).
Figure 30. NMR spectra of water extracts of C. arabica (bottom trace) and C. canephora var. robusta
(top trace) with the assignment of main chemical compounds. Spectra are referenced and scaled against
external standard. Letters stands for: A: acetate; CGA: chlorogenic acids; C: caffeine, Q: quinic acids; T:
trigonelline; F: 2-furyl methanol; NP: N-methyl pyridine and FO: formiate. (Reprinted with permission
from Cagliani et al., 2013. Copyright (2013) Elsevier).
roasted coffee blends, in spite of different geographic origin and roasting conditions,
making this approach suitable for authentication of unknown arabica and robusta coffee
blends (Figure 31).
Figure 31. Predicted (YPred) versus reference (YVar) percentage composition inarabica obtained for
training set samples. (Reprinted with permission from Cagliani L. R. et al., 2013. Copyright (2013)
Elsevier).
This method has also been patented (Consonni & Cagliani, 2012 MI2012A001281).
Recent examples reported still the use of NMR in authentication process, such as in
determining tocopherol content as a marker for coffee adulteration with maize (Tavares et
al., 2016) or by detection of 16-O-methylcafestol (16-OMC) from robusta in coffee blends
by hydrophobic extracts (Schievano et al., 2014) and to differentiate origins (Arana et al.,
2015). A different coffee derived product, coffee oil, was investigated by another group; at
first, an extensive characterization of the matrix was conducted employing different
extraction methods and different conditions for the acquisition of the NMR spectra
(D’Amelio et al., 2013). The main components of green bean coffee oil were identified and
quantified from 13C NMR spectra. Specifically, the content of free fatty acids,
triacylglycerols, diterpene esters, and 1,3-diacylglycerols were determined in samples from
several different origins, both arabica and robusta, using appropriate 13C signals. The fatty
acid composition was also determined. Particular attention was given to the assignment
and quantification of diterpenes, which were the topic of subsequent publications.
Comparison of the NMR results with standard chromatographic methods underlined the
potential for the development of methods that could replace time-consuming analyses in a
routine setting, such as the detection of caffeine in green coffee oil or the quantitative
Figure 32. Expanded region of the 1H NMR spectrum showing the diagnostic resonances of protons 17
and 21: (a) 100% Robusta sample, (b) 100% Arabica sample, and (c) 16-OMC standard. In the insets on
the top of the figure, the signals of 17 and 21 protons of 16-OMC ester are enlarged. They are clearly
visible in the Robusta sample (top), but they are absent from the Arabica sample. The arrow shows the
weak methyl signals in position 21 of 16-OMC in its free form.
Compared to the DIN method, the NMR one has a lower Limit of Detection and the
proposed extraction procedure is more efficient. The quantification of 16-OMC could in
principle allow the percent composition of arabica/robusta blends to be determined.
Unfortunately, the amount of this compound is very variable in robusta samples (Speer et
al., 2006). Recently, Finotello et al. investigated the variability of 16-OMC by NMR both
on green and roasted beans (Finotello et al., in preparation) and concluded that Asian
samples are much less differentiated than African ones. Although the precise determination
Figure 33. Zoom of the aliphatic region of the spectra of saliva extracts at different times after caffeine
intake.
Finally, the expertise of the Padova group in both honey and coffee was the perfect
combination to study a still rare product, such as coffee honey (Schievano et al., 2015b).
Using a water extraction, several compounds were quantified by NMR. Specifically, the
simultaneous presence of the three alkaloids, caffeine, theobromine, and trigonelline, was
suggested to be the molecular signature of coffee honey. The Coffea pollen content
determined in the three samples studied correlated well with the amount of these three
compounds, supporting this hypothesis.
The use of bench-top NMR instruments is gaining interest in the field of food science,
largely because of their low cost. The group in Padova contributed to the study of coffee
with the determination of water distribution in green beans using time-domain NMR
(Venditti et al., 2011). Venditti compared the T2 behavior of beans hydrated in controlled
conditions and found that up to ~10% water protons exhibit restricted mobility. Part of this
effect is a consequence of chemical exchange of the water protons with those of the solid
matrix. They interpreted their results in terms of the antiplasticizer and plasticizer effect of
water. Finally, an interesting pilot study was published by Wei’s group (Wei et al., 2014)
where they proposed NMR as a sort of "magnetic tongue" for the characterization and
prediction of the tastes of foods, since it provides a wealth of information in a
nondestructive and nontargeted manner. Different sensations of roasted coffee taste (on
both arabica and robusta extracts) were identified by the combination of NMR-based
metabolomics and human sensory test by the use of multivariate projection method of
orthogonal projection to latent structures (OPLS).
2.9. Saffron
Saffron, the most expensive spice in the world, is obtained from the pistils of Crocus
sativus L. flowers after drying process according to the trade standard ISO 3632 (ISO,
2011). This process is carried out by the producers with traditional procedures with some
differences in the practices according to geographical origin (Ordoudi & Tsimidou, 2004).
The main production is located in Iran, followed in Asia by India while in Africa and
Europe Morocco and, Greece, Spain and Italy constituted the main producers. Saffron is
traditionally used for coloring and flavoring food but it is also endowed with a range of
health promoting benefits (Melnyk et al., 2010; Winterhalter & Straubinger, 2000;
Tarantilis & Polissiou, 2004).
The typical color, taste, aroma and flavor of saffron are due to three main secondary
metabolites: crocins, responsible for the strong coloring capacity, picrocrocin, conferring
the bitter taste, and safranal, giving rise to the characteristics saffron odor and aroma
(Tarantilis et al., 1994; Kanakis et al., 2004).
The high market value of this spice is mainly related to manual cultivation cost which
is not mechanized yet. Because of its high price and limited production, saffron has been
subjected to various types of adulteration throughout last years (Hagh-Nazari & Keifi,
2007). Common fraudulent practices include artificial colorants addition as well as inferior
plant material with similar appearance, mainly when the spice is in powder form (Hagh-
Nazari & Keifi, 2007; Torelli, et al., 2014).
The quality of saffron and its commercial evaluation are determined by specifications
of ISO/TS-3632 standard (ISO, 2010; ISO, 2011) that established spectrophotometric
quantification of crocins, picrocrocin and safranal in aqueous saffron extracts by
absorbance measurements at 440, 257 and 330 nm respectively. Unfortunately this method
presents some disadvantages because first of all safranal is only scarcely water soluble and
in the range of 320-340 nm, cis-crocin isomers gave adsorption as well (Kanakis et al.
2004; Tarantilis et al., 1995; Zougagh et al., 2006).
Figure 34. NMR spectra of DMSO extract of Italian PDO saffron from Aquila. (A) One dimensional
proton spectrum with primary assignments reported. (B) Expansion of the aromatic region of TOCSY
spectrum reporting the resonances assignment of kaempferol. (C) Anomeric region of HSQC spectrum
reporting the assignment for bound and unbound saccharides (Cagliani et al., 2015).
The first NMR data about a group of constituents of saffron was reported by Wittwer
(Pfander & Wittwer, 1975) in late 1975 on isolated glycosyl esters of crocetin. The
evidence of a geometrical isomer of crocin was successively reported by Speranza
(Speranza & Dadà, 1984), highlighting the spectroscopic characterization of 13-cis crocins
together with most abundant all trans-crocins in a Greek saffron sample. Others studies
followed focusing the structure elucidation of saffron compounds combining NMR
spectroscopy with other analytical techniques such as HPLC, MS, LC-MS, FTIR, FT-
Raman, UV (Assimiadis et al., 1997; Pfister et al. 1996; Straubinger et al., 1997;
Straubinger et al., 1998; Van Calsteren et al., 1997). Only few years ago, in 2010, appeared
the first NMR works based on the metabolic fingerprinting of saffron extracts with the aim
to distinguish among authentic Iranian saffron and commercial samples obtained from
retail stores in Denmark, Sweden and Turkey (Yilmaz et al., 2010; Yilmaz et al., 2011).
The reported data suggested the possibility to obtain discrimination by PCA and Parallel
Factor analysis by using mono and two dimensional NMR data respectively but the authors
inexplicably did not report detailed metabolite assignment. Further NMR investigations
revealed the presence of food additive (E1518) in one group of sample, while in others the
presence of bio-adulterants, like C. sativus stigmata, Curcuma longa and Carthamus
tinctorius flowers were detected. Successively Cagliani et al. performed a deeper metabolic
NMR analysis of Italian PDO saffron (Figure 34) (Cagliani et al., 2015).
Figure 35. OPLS-DA score plot performed by considering all saffron samples analyzed: R 2X = 98.4%,
R2Y= 98.8% and Q2=84.6%. Filled dots and diamonds represent Italian PDO and commercial saffron
samples respectively (Cagliani et al., 2015).
Figure 36. PCA score plot performed considering all (n=98) saffron samples. PC1=68.8%; PC2=17.1%.
R2X=92.8%; Q2=90.4%. The samples are colored according to the origin: black triangles, light grey
boxes, black diamonds and light grey circles represent Greek, Iranian, Italian and Spanish samples,
respectively; the storage period for each sample is reported from Ordoudi et al., 2015.
The obtained model was validated by training and test set and used for the quality
evaluation of some commercial samples of unknown storage history by Consonni et al.
(Consonni et al., 2016). In this study NMR data were compared with those obtained by
FTIR on the same samples. The outcomes were very encouraging as it was possible to
establish a limit value to establish substandard quality of samples. The combination of
metabolomic techniques offered a useful tool to combat saffron mislabeling and
counterfeiting with low-quality saffron material irrespective of provenience. Other studies
were focused on the adulteration of saffron. In 2015 Petrakis et al. performed a preliminary
study for the detection of plant adulterated saffron and the identification of the adulterant
used by means of 1H NMR and chemometrics (Petrakis et al., 2015). Authentic Greek
saffron samples and four typical plant-derived materials utilized as bulking agents in
saffron such as Crocus sativus stamens, safflower, turmeric, and gardenia were
investigated. A two-step approach, relied on the application of both OPLS-DA and O2PLS-
DA protocols to the 1H NMR data, was adopted to perform the authentication and the
prediction of authentic and adulterated saffron respectively. The first model led to
discriminate authentic from artificial counterfeit mixtures of saffron containing 20% (w/w)
of each plant adulterant while the second one allowed to discriminate the type of plant
adulteration (Figure 37). Both models presented a good predictive capability, being
validated by using a test set, strongly supporting the validity of the approach proposed.
Figure 37. O2PLS-DA score plot (PC1 versus PC3) performed by considering only adulterated saffron
divided into four classes according to the type of plant adulterant: saffron adulterated with 20%
concentration (w/w) with Gardenia jasminoides fruit extract, safflower, C. sativus stamens and turmeric
are presented with light gray (1), black (2), black (3), and grey (4) circles, respectively. R2X = 95.2%,
R2Y = 97.6% and Q2 = 96% (Petrakis et al., 2015).
Finally, the identification and the quantification of saffron adulteration with Sudan
dyes I–IV (Figure 38) were investigated (Petrakis et al., 2017). Sudan dyes constitute a
family of lipophilic azo-compounds, widely used as coloring agents in industrial products
such as textiles, plastics, waxes, and polishes while their use in food is forbidden in most
countries, including the European Union (Commission Decision, 2005) being the dyes
Sudan I–IV classified as Group 3 carcinogens by the International Agency for Research on
Cancer (IARC, 1975). Notwithstanding the illegality of these dyes, their presence in a
range of foodstuffs including saffron, have been reported to EU rapid alert system
(RASFF).
Figure 38. Structure and labeling scheme for Sudan I–IV dyes analyzed (Petrakis et al., 2017).
By HR-NMR spectroscopy, a complete 1H and 13C NMR assignment for all Sudan dyes
was achieved allowing to identify, in the aromatic region of 1H NMR spectra, as reported
in Figure 39, specific resonances that allowed the univocal identification of each Sudan
dye in adulterated saffron. Moreover, by applying the qHNMR the quantification of Sudan
III, as a paradigm, in adulterated saffron at levels that may practically impact the visual
color (0.14–7.1 g/kg), was performed considering the signal occurring at 8.064 ppm. The
high linearity, accuracy and rapidity obtained encourage the investigation by 1H NMR
spectroscopy as a powerful technique for the evaluation of saffron adulteration with Sudan
dyes.
Figure 39. Aromatic region of 1H NMR spectra of Sudan I–IV dyes and the pure Greek saffron analyzed.
From the bottom to the top pure saffron, Sudan I, Sudan II, Sudan III, and Sudan IV are represented.
Specific signals for the identification of each Sudan dye in adulterated saffron are highlighted with the
red grid (Petrakis et al., 2017).
Quality is a very important parameter of the industrial and artisanal products and
services. Many definition of quality can be briefly summarized as perception of the degree
to which the product or service meets the customer's expectations.
In other words, quality can be defined as the characteristics of a product or service that
bear on its ability to satisfy stated or implied needs and customer satisfaction, or, simply,
a product or service free of deficiencies. Obviously there is a statistical implication because
the control cannot be limited to only one sample but to a population of products and
services and the results must appear uniform around a target value.
In foods, quality is well documented by the requirements in terms of healthy,
nutritional value, correctness of the processing and storage, place of production, absence
of contamination(s) and the particular value of the product in the market.
Many techniques have been developed to check the quality and among them magnetic
resonance has developed a particular important role in testing those aspects that constitute
quality. Several examples, often successfully applied in the food industry, have been
proposed with particular good results that allowed to form a robust set of tests that together
with other techniques more sensitive to low concentration (HPLC and HPLC-MS) to be
used in a large variety of food products.
In the field of nutraceutics proteins constitute a class of products that are considered
very healthy due to their composition and effects on human. “Nutraceutical” is a fusion of
the words “nutrition” and “pharmaceutical” thus including nutrients, dietary supplements
and herbal products. Briefly they can be grouped in dietary supplements as nutrients
derived from food concentrated products (vitamins, minerals, herbs, amino acids, proteins,
enzymes), functional foods or foods enriched which are foods added with the components
that are important for human health. In these classes of compounds many chemical entities
are to be measured and determined. One different problem is often constituted by proteins
present in this preparation. In fact the determination of proteins concentration is a very
simple procedure whilst the determination of their quality should represent a new
measurable parameter. In fact the proteins are a class of compounds where the quality of
proteins does depend on their biological activity. Obviously biological activity of every
protein should be often measured by a specific test. This cannot be easily measured in the
case of nutraceuticals due to the number and variety of proteins present in these
preparations. On the other hand the quality of a protein can be easily considered as the level
of its biological integrity. The degree of folding (or oppositely the level of the unfolding
process as a denaturing process) or, alternatively the degree of damages induced to proteins
by the industrial process and storage, can be considered a parameter that helps in the
estimate of the level of quality or, in this case of “bioquality.” We can assume that the
higher the structural characteristics higher can be considered the quality of the proteins in
a product. This is particularly relevant in nutraceutics where proteins are added to the
formulation to give an optimal final product.
Proteins are an important source of amino acids that have to be introduced constantly.
Obviously this has a direct connection with the quantity but more specifically with the
quality of the protein sources (Ramani et al., 2010). There are several protein nutraceutical
formulations in the market using as protein sources plants (including here fungi or and
algae) and animals. Among these sources proteins derived from milk proteins have a
central role. In fact, several baby milk formula and or nutraceuticals with protein based
formulation usually are derived from soluble fractions of proteins from cow milk. The
soluble protein fraction (20% of milk protein fraction) contains proteins that are commonly
named whey proteins, whereas insoluble proteins are primarily the caseins (Jenness, 1979).
The principal whey proteins components from milk are β-lactoglobulin, α-lactalbumin,
serum albumin, immunoglobulins, lactoferrin and other minor fractions composition, being
very heterogeneous both from the point of view of structure and molecular weight (MW).
All of them present a relevant number of branched chain amino acids whereas their
counterparts in the insoluble fraction of milk, caseins, are principally composed by
aromatic and sulfur containing amino acids (Tang et al., 2009). If one refers to the Protein
Digestibility Corrected Amino Acid Score (PDCAAS) (Haug et al., 2007; Schaafsma,
2000; Schaafsma, 2012) considering human requirements for amino acids, digestibility,
and bioavailability milk proteins are considered the best protein source.
The protein fraction of milk is not only important from the point of view of nutrients
but it exerts also direct effects on human health. In fact, proteins such as Lysozyme,
Lactoferrin, and Lactoperoxidase are well known for their action as important
antimicrobial agents (Jenssen & Hancock, 2009; Min et al., 2005). Futhermore, proteins
such as Lactoferrin, β-lactoglobulin and α-lactalbumin present a suppressing action in
tumor development (Parodi, 2007). It is worth note that Lactoferrin is very important also
in iron absorption and exerts antioxidant and anticarcinogenic effects while β-lactoglobulin
is a retinol carrier with antioxidant proprieties (Gonzalez-Chavezet, 2009). Moreover,
peptides obtained from enzymatic cleavage of milk proteins present their own biological
activity. These peptides have potential functions in different biological pathways as
modulators and thus may exert beneficial physiological effects (Meisel & FitzGerald,
2003).
by densitometry analysis (down panels) revealing a clear difference both from the point of
view of MW and protein concentration. Probably this variability must be due to the source
protein raw material used from each manufacturer. In fact, there are several proceedings
used to obtain milk whey proteins among raw material producers, which may have an
impact on the protein profile of milk whey proteins used for these preparations.
Figure 40. SDS-PAGE analysis of protein fraction of three different nutraceutical products (A, B and C)
taken from the market (upper panel) and their densitometry analysis (down panels) revealing clearly
different compositions of proteins in terms of MW and concentrations.
spread of the resonances observed in NMR spectra of native proteins is, oppositely, the
indicator of a well structured secondary and tertiary structure. In fact, the local magnetic
fields determined by structural elements are able to shift the resonances by a large different
spread. This can be measured by NMR thus obtaining parameter related to the degree of
the unfolding process in these proteins. The use of autocorrelation function derived from
1
H NMR spectra where chemical shifts multiplicity is governed by the details of the 3D
solution structures of proteins can be used as an indicator of protein folding (Figure 41). In
this method one can use the autocorrelation function directly from the 1H NMR spectra
correlated with the topological complexity. By using sets of folded native proteins or their
mixtures one may evaluate the protein folding by taking in consideration the value of the
autocorrelation function at C(0.5) (Hoffman et al., 2005). The evaluation, takes into
account that natively folded proteins display C (0.5) values > 0.5, while partially folded or
unfolded proteins have values of < 0.4. In general, in terms of folding we can observe from
Table 2.2 referring to the same three precedent samples the that the Highest Score (HS)
was achieved by sample C, followed by sample A with an Intermediate High Score (IHS)
while B was classified with a Very Low Score (VLS).
Molecular Weight assuming a globular shape for such proteins. In Figure 42 we can
observe the enlargement three DOSY experiments. The Diffusion coefficient (D) is here
reported as logD and is in first approximation for globular and quite spherical proteins
related to the radius of the molecular species by the Stokes-Einsten Equation (Einstein,
1956).
Figure 42. DOSY spectra of protein of the three nutraceutical protein based formulations (A, B and C)
present in the market. From the spectra one can easily evaluate the mean molecular weight of the proteins
present in the sample.
Moreover, the radius of the species in the solution can be related to their MW, the
measure of D may give information on the mean MW of the protein fraction. Without
knowing previously, the mean molecular weight of the samples we can classify them in
base of their Diffusion coefficient as described in Table 2.2.
CONCLUSION
protocols that can be validated and possibly recognized as official methodologies for the
control of food quality, authenticity and provenance.
ACKNOWLEDGMENTS
This work has been carried out within the Italian project of Regione Lazio Lr 13/2008
entitled “e-ALIERB: un OPEN LAB per caratterizzare e valorizzare i prodotti alimentari
ed erboristici del territorio laziale.”
REFERENCES
Al-Jowder, O., Casuscelli, F., Defernez, M., Kemsley, E. K., Wilson, R. H., & Colquhoun,
I. J. (2001). High resolution NMR studies of meat composition and authenticity. In G.
A. Webb, P. S. Belton, A. M. Gil, & I. Delgadillo (Eds.), Magnetic Resonance in Food
Science: A View to the Future (First edition, pp. 232-238). Cambridge, UK: Royal
Society of Chemistry.
Andreotti, G., Trivellone, E., Lamanna, R., Di Luccia, A., & Motta, A. (2000). Milk
identification of different species: 13C-NMR spectroscopy of triacylglycerols from
cows and buffaloes’milks. Journal of Dairy Science, 83, 2432-2437.
Andreotti, G., Lamanna, R., Trivellone, E., & Motta, A. (2002). 13C NMR spectra of tag:
an easy way to distinguish milks from different animal species. Journal of the
American Oil Chemists' Society, 79, 123-127.
Andreotti, G., Trivellone, E., & Motta, A. (2006). Characterization of buffalo milk by 31P-
nuclear magnetic resonance spectroscopy. Journal of Food Composition and Analysis,
19, 843-849.
Arana, V. A., Medina, J., Alarcon, R., Moreno, E., Heintz, L., Schafer, H., & Wistl, J.
(2015). Coffee’s country origin determined by NMR: The Colombian case. Food
Chemistry, 175, 500-506.
Aru, V., Pisano, M. B., Scano, P., Cosentino, S., & Cesare Marincola, F. (2015). Changes
in the 1H NMR metabolic profile of mussels (Mytilus galloprovincialis) with storage
at 0°C. In F. Capozzi, L. Laghi, & P. S. Belton (Eds.), Magnetic Resonance in Food
Science: Defining Food by Magnetic Resonance. (First edition, pp. 181-198).
Cambridge, UK: Royal Society of Chemistry.
Aru, V., Pisano, M. B., Savorani, F., Engelsen, S. B., Cosentino, S., & Cesare Marincola,
F. (2016). Metabolomics analysis of shucked mussels’ freshness. Food Chemistry, 205,
58-65.
Bogdanov, S. (1997). Nature and origin of the antibacterial substances in honey. LWT-
Food Science and Technology, 30, 748-753.
Brescia, M. A., Mazzilli, V., Sgaramella, A., Ghelli, S., Fanizzi, F. P., & Sacco, A. (2004).
1
H NMR characterization of milk lipids: a comparison between cow and buffalo milk.
Journal of Food Composition and Analysis, 81, 431-436.
Brescia, M. A., Monfreda, M., Buccolieri, A., & Carrino, C. (2005). Characterisation of
the geographical origin of buffalo milk and mozzarella cheese by means of analytical
and spectroscopic determinations. Food Chemistry, 89, 139-147.
Cagliani, L. R, Pellegrino, G., Giugno, G., & Consonni, R. (2013). Quantification of
coffeaarabica and coffeacanephora var. robusta in roasted and ground coffee blends.
Talanta, 106, 169-173.
Cagliani, L. R., Culeddu, N., Chessa, M. & Consonni, R. (2015). NMR investigations for
a quality assessment of Italian PDO saffron (Crocus sativus L.). Food Control, 50, 342-
348.
Caligiani, A., Acquotti, D., Palla, G., & Bocchi, V. (2007a). Identification and
quantification of the main organic components of vinegars by high resolution 1H NMR
spectroscopy. Analytica Chimica Acta, 585, 110-119.
Caligiani, A., Silva, G., & Palla, G. (2007b). Determination of 2,3 butanediol and
2hydroxybutanone stereoisomers in batteries of Traditional Balsamic vinegar. Journal
of Agricultural and Food Chemistry, 55, 7810-7815.
Caligiani, A., Acquotti, D., Cirlini, M., & Palla, G. (2010). 1H NMR study of fermented
cocoa (Theobroma Cacao L.) beans. Journal of Agricultural and Food Chemistry, 58,
12105-12111.
Caligiani, A., Coisson, J. D., Travaglia, F., Acquotti, D., Palla, G., Palla, L., & Arlorio, M.
(2014a). Application of 1H NMR for the characterisation and authentication of “Tonda
Gentile Trilobata” hazelnuts from Piedmont (Italy). Food Chemistry, 148, 77-85.
Caligiani, A., Palla, L., Acquotti, D., Marseglia, A., & Palla, G. (2014b). Application of 1H
NMR for the characterisation of cocoa beans of different geographical origins and
fermentation levels. Food Chemistry, 157, 94-99.
Caligiani, A., Marseglia, A., & Palla G. (2016). Cocoa: production, chemistry, and use.
Reference Module in Food Science, from Encyclopedia of Food and Health, 185-190.
Capitani, D., Mannina, L., Proietti, N., Sobolev, A. P., Tomassini, A., Miccheli, A., Di
Cocco, M. E., Capuani, G., De Salvador, R., & Delfini, M. (2010). Monitoring of
metabolic profiling and water status of Hayward kiwifruits by Nuclear Magnetic
Resonance. Talanta, 82(5), 1826-1838.
Capitani, D., Mannina, L., Proietti, N., Sobolev, A. P., Tomassini, A., Miccheli, A., Di
Cocco, M. E., Capuani, G., De Salvador, F. R., & Delfini, M. (2013a). Metabolic
Profiling and Outer Pericarp Water State in Zespri, CI.GI., and Hayward kiwifruits.
Journal of Agricultural and Food Chemistry, 61(8), 1727-1740.
Capitani, D., Sobolev, A. P., Tomassini, A., Sciubba, F., De Salvador, R., Mannina, L., &
Delfini, M. (2013b). Peach fruit: a metabolic comparative analysis of two varieties with
different resistance to insect attacks by NMR spectroscopy. Journal of Agricultural
and Food Chemistry, 61(8), 1718-1726.
Capitani, D., Sobolev, A. P., Delfini, M., Vista, S., Antiochia, R., Proietti, N., Bubici, S.,
Ferrante, G., Carradori, S., De Salvador, F. R., & Mannina, L. (2014). NMR
methodologies in the analysis of blueberries. Electrophoresis, 35, 1615-1626.
Castejon, D., Garcia-Segura, J. M., Escudero, R., Herrera, A., & Cambero, M. I. (2015).
Metabolomics of meat exudate: Its potential to evaluate beef meat conservation and
aging. Analytica Chimica Acta, 901, 1-11.
Cesare Marincola, F., Noto, A., Caboni, P., Reali, A., Barberini, L., Lussu, M., & Fanos,
V. (2012). A metabolomic study of preterm human and formula milk by high resolution
NMR and GC/MS analysis: preliminary results. The Journal of Maternal-Fetal &
Neonatal Medicine, 25(sup5), 62-67.
Ciampa, A., Picone, G., Laghi, L., Nikzad, H., & Capozzi, F. (2012). Changes in the amino
acid composition of bogue (Boops boops) fish during storage at different temperatures
by 1H-NMR spectroscopy. Nutrients, 4, 542-553.
Cicero, N., Corsaro, C., Salvo, A., Vasi, S., Giofré, S. V., Ferrantelli, V., Di Stefano, V.,
Mallamace, D., & Dugo, G. (2015). The metabolic profile of lemon juice by proton
HR-MAS NMR: the case of the PGI Interdonato lemon of Messina. Natural Product
Research, 29, 1894-1902.
Cirlini, M., Caligiani, A., & Palla, G. (2009). Formation of glucose and fructose acetates
during maturation and ageing of balsamic vinegar. Food Chemistry, 112, 51-56.
Codex Alimentarius Commission, (2002). Codex standard 12, Revised Codex Standard for
Honey. Standard Methods, 11, 1-8.
Codex Stan 12-1981. (2001). “Codex standard for honey.” Last modified 2001.
www.fao.org/input/download/standards/310/cxs_012e.pdf.
Cohen, Y., Avram, L., & Frish, L. (2005). Diffusion NMR spectroscopy in supramolecular
and combinatorial chemistry: an old parameter-new insights. Angewandte Chemie
International Edition, 44, 520-554.
Commission Decision (2005). Commission decision 2005/402/EC of 23 May 2005 on
emergency measures regarding chilli, chilli products, curcuma and palm oil. Official
Journal of the European Union, L135, 34-36.
Consonni, R., & Gatti, A. (2004). 1H NMR studies on Italian Balsamic and Traditional
Balsamic vinegars. Journal of Agricultural and Food Chemistry, 52, 3446-3450.
Consonni, R. (2007). Patent no. MI2007A001489, Italy.
Consonni, R., & Cagliani, L. R. (2007). NMR relaxation data for quality characterization
of Balsamic vinegar of Modena. Talanta, 73, 332-339.
Einstein, A. (1956). Investigation on the Theory of the Brownian Movement (First edition).
New York, NY: Dover Publicatios.
Ellis, R. P., Spicer, J. I., Byrne, J. J., Sommer, U., Viant, M. R., White, D. A., &
Widdicombe, S. (2014). 1H NMR metabolomics reveals contrasting response by male
and female mussels exposed to reduced seawater pH, increased temperature, and a
pathogen. Environmental Science and Technology, 48, 7044-7052.
Erich, S., Schill, S., Annweiler, E., Waiblinger, H.-U., Kuballa, T., Lachenmeier, D. W.,
& Monakhova, Y. B. (2015). Combined chemometric analysis of 1H NMR, 13C NMR
and stable isotope data to differentiate organic and conventional milk. Food Chemistry,
188, 1-7.
Fanizzi, F. P., Del Coco, L., Girelli, C. R., & De Pascali, S. A. (2015). Harvest year effects
on Apulian EVOOs evaluated by 1H NMR based metabolomics. In IV International
Conference on Foodomics, Cesena (FC). Book of abstract pp. 31-32. Edited by
University of Bologna.
Ferreres, F. Taveira, M., Pereira, D. M., Valentão, P., & Andrade, P. B. (2010). Tomato
(Lycopersiconesculentum) Seeds: New Flavonols and Cytotoxic Effect. Journal of
Agricultural and Food Chemistry, 58, 2854-2861.
Finotello, C., Forzato, C., Gasparini, A., Mammi, S., Navarini, L., & Schievano, E. NMR
quantification of 16-O-Methylcafestol and Kahweol in Coffeacanephora var. robusta
beans from different geographical origins Food Control – submitted.
Fotakis, C., & Zervou, M. (2016). NMR metabolic fingerprinting and chemometrics driven
authentication of Greek grape marc spirits. Food Chemistry, 196, 760-768.
Garcia, C., Lutz, N. W., Confort-Gouny, S., Cozzone, P. J., Armand, M., & Bernard, M.
(2012). Phospholipid fingerprints of milk from different mammalians determined by
31
P NMR: Towards specific interest in human health. Food Chemistry, 135, 1777-1783.
Gazzetta Ufficiale number 57, March 9, (2006). Passata di pomodoro. Origine del
pomodoro fresco. Ministero delle Politiche Agricole e Forestali, Feb 17, 2006.
Gianferri, R., Maioli, M., Delfini, M., & Brosio, E. (2007). A low-resolution and high-
resolution nuclear magnetic resonance integrated approach to investigate the physical
structure and metabolic profile of Mozzarella di Bufala Campana cheese. International
Dairy Journal, 17, 167-176.
Girelli, C. R., Del Coco, & L. Fanizzi, F. P. (2016a). 1H NMR spectroscopy and
multivariate analysis as possible tool to assess cultivars, from specific geographical
areas, in EVOOs. European Journal of Lipid Science and Technology, 118, 1380-1388.
Girelli, C. R., Del Coco, L., Papadia, P., De Pascali, S. A., & Fanizzi, F. P. (2016b). Harvest
year effects on Apulian EVOOs evaluated by 1H NMR based metabolomics. PeerJ, in
press.
Gonzalez-Chavez, S. A., Arevalo-Gallegos, S., & Rascon-Cruz, Q. (2009). Lactoferrin:
structure, function and applications. International Journal of Antimicrobial Agents, 33,
301.e (pp. 8).
Gosling, E. (2015). Phylogeny and evolution of bivalve molluscs. In J. Wiley & Sons
(Eds.), Marine Bivalve Molluscs (Second edition pp. 1-11). Chichester, UK: Wiley-
Blackwell.
Graham, S. F., Kennedy, T., Chevallier, O., Gordon, A., Farmer, L., Elliott, C., & Moss,
B. (2010). The application of NMR to study changes in polar metabolite concentrations
in beef longissimus dorsi stored for different periods post mortem. Metabolomics, 6(3),
395-404.
Graham, S. F., Farrell, D., Kennedy, T., Gordon, A., Farmer, L., Elliott, C., & Moss, B.
(2012). Comparing GC-MS, HPLC and 1H NMR analysis of beef longissimus dorsi
tissue extracts to determine the effect of suspension technique and ageing. Food
Chemistry, 134(3), 1633-1639.
Haeupler, M., Peyronel, F., Neeson, I., Weiss, J., & Marangoni, A. G. (2014). In situ
ultrasonic characterization of cocoa butter using a chirp. Food and Bioprocess
Technology, 7(11), 3186-3196.
Hagh-Nazari, S., & Keifi, N. (2007). Saffron and various fraud manners in its production
and trades. Acta Horticulturae, 739, 411-416.
Haszprunar, G., Schander, C. & Halanych, K. M. (2008). Relationships of higher
molluscan taxa. In W. Ponder, & D. R. Lindberg Editor (Eds.), Phylogeny and
Evolution of the Mollusca (First edition pp. 19-32). Berkeley, USA: University of
California Press.
Haug, A., Høstmark, A. T., & Harstad, O. M. (2007). Bovine milk in human nutrition - a
review. Lipids in Health and Disease, 6(25), 16 pp.
Hoffmann, B., Eichmüller, C., Steinhauser, O., & Konrat, R. (2005). Rapid assessment of
protein structural stability and fold validation via NMR. Methods in Enzymology,
394,142-175.
IARC (1975). Monographs on the evaluation of the carcinogenic risk of chemicals to man:
some aromatic azo compounds (Vol. 8), (pp. 224–231). Lyon, France: International
Agency for Research on Cancer.
Ishida, N., Kobayashi, T., Koizumi, M., & Kano H. (1989). 1HNMR imaging of tomato
fruits. Agricultural and Biological Chemistry, 53, 2363-2367.
ISO, (2010). International Standard ISO 3632-2: Saffron (Crocus sativus L.) Test Methods.
Geneva: The International Organization for Standardization.
ISO, (2011). International Standard ISO 3632-1: Saffron (Crocus sativus L.) Specification.
Geneva: The International Organization for Standardization.
Jahurul, M. H. A., Zaidul, I. S. M., Norulaini, N. A. N., Sahena, F., Abedin, M. Z., Ghafoor,
K., & Mohd, O. A. K. (2014). Characterization of crystallization and melting profiles
of blends of mango seed fat and palm oil mid-fraction as cocoa butter replacers using
differential scanning calorimetry and pulse nuclear magnetic resonance. Food
Research International, 55, 103-109.
Jakes, W., Gerdova, A., Defernez, M., Watson, A. D., McCallum, C., Limer, E.,
Colquhoun, I. J., Williamson, D. C., & Kemsley, E. K. (2015). Authentication of beef
versus horse meat using 60 MHz 1H NMR spectroscopy. Food Chemistry, 175, 1-9.
Jenness R. (1979). Comparative aspects of milk proteins. Journal of Dairy Research, 46,
197-210.
Jenssen, H., & Hancock, R. E. W. (2009). Antimicrobial properties of lactoferrin.
Biochimie, 91, 19-29.
Kaffarnik, S., Ehlers, I., Gröbner, G., Schleucher, J., & Vetter, W. (2013). Two-
dimensional 31P,1H NMR spectroscopic profiling of phospholipids in cheese and fish.
Journal of Agricultural Food Chemistry, 61, 7061-7069.
Kanakis, C. D., Daferera, D. J., Tarantilis, P. A., & Polissiou, M. G. (2004). Qualitative
determination of volatile compounds and quantitative evaluation of safranal and 4-
hydroxy-2,6,6-trimethyl-1-cyclohexene-1-carboxaldehyde (HTCC) in Greek saffron.
Journal of Agricultural and Food Chemistry, 52, 4515-4521.
Karkoula, E., Skantzari, A., Melliou, E., & Magiatis, P. (2012). Direct measurement of
oleocanthal and oleacein levels in olive oil by quantitative 1H NMR. Establishment of
a new index for the characterization of extra virgin olive oils. Journal of Agricultural
Food Chemistry, 60, 11696-11703.
Kaškoniene, V., & Venskutonis, P. R. (2010). Comprehensive Reviews in Food Science
and Food Safety, 9(6), 620-634.
Klein, M. S., Buttchereit, N., Miemczyk, S. P., Immervoll, A. K., Louis, C., Wiedemann,
S., & Gronwald, W. (2011). NMR metabolomic analysis of dairy cows reveals milk
glycerophosphocholine to phosphocholine ratio as prognostic biomarker for risk of
ketosis. Journal of Proteome Research, 11(2), 1373-1381.
Koda, M., Furihata, K., Wei, F., Miyakawa T., & Tanokur M. (2011). F2-selective two-
dimensional NMR spectroscopy for the analysis of minor components in foods.
Magnetic Resonance in Chemistry, 49, 710-716.
Lamanna, R., Braca, A., Di Paolo, E., & Imparato, G. (2011). Identification of milk
mixtures by 1H NMR profiling. Magnetic Resonance in Chemistry, 49, S22-S26.
Lauri, I., Pagano, B., Malmendal, A., Sacchi, R., Novellino, E., & Randazzo, A. (2013).
Application of “magnetic tongue” to the sensory evaluation of extra virgin olive oil.
Food Chemistry, 140(4), 692-699.
Le Gall, G., Colquhoun, I. J., Davis, A. L., Collins, G. J., & Verhoeyen, M. E. (2003).
Metabolite profiling of tomato (Lycopersiconesculentum) using 1H NMR spectroscopy
as a tool to detect potential unintended effects following a genetic modification.
Journal of Agricultural and Food Chemistry, 51, 2447-2456.
Lee, Y. G., Cho, J. Y., Kim, C. M, Lee, S. H., Kim, W. S., Jeon, T. I., Park, K. H., & Moon,
J. H. (2013). Coumaroylquinic acid derivatives and flavonoids from immature pear
(Pyruspyrifolianakai) fruit. Food Science and Biotechnology, 22, 803-810.
Locci, E., Piras, C., Mereu, S., Cesare Marincola, F., & Scano, P. (2011). 1H NMR
metabolite fingerprint and pattern recognition of mullet (Mugilcephalus) bottarga.
Journal of Agricultural and Food Chemistry, 59, 9497-9505.
Longobardi, F., Ventrella, A., Napoli, C., Humpfer, E., Schütz, B., Schäfer, H.,
Kontominas, M. G., & Sacco, A. (2012). Classification of olive oils according to
geographical origin by using 1H NMR fingerprinting combined with multivariate
analysis. Food Chemistry, 130, 177-183.
Luykx, D. M. A. M., & van Ruth, S. M. (2008). An overview of analytical methods for
determining the geographical origin of food products. Food Chemistry, 107, 897-911.
Maher, A. D., Hayes, B., Cocks, B., Marett, L., Wales, W. J., & Rochfort, S. J. (2013).
Latent biochemical relationships in the blood-milk metabolic axis of dairy cows
revealed by statistical integration of 1H NMR spectroscopic data. Journal of Proteome
Research, 12(3), 1428-1435.
Maher, A. D., & Rochfort, S. J. (2014). Applications of NMR in Dairy Research.
Metabolites, 4, 131-141.
Mallamace, D., Corsaro, C., Salvo, A., Cicero, N., Macaluso, A., Giangrosso, G.,
Ferrantelli, V., & Dugo, G. (2014). A multivariate statistical analysis coming from the
NMR metabolic profile of cherry tomatoes (The Sicilian Pachino case). Physica A,
401, 112-117.
Malmendal, A., Amoresano, C., Trotta, R., Lauri, I., De Tito, S., Novellino, E., Randazzo
A. (2011). NMR spectrometers as "magnetic tongues": prediction of sensory
descriptors in canned tomatoes. Journal of Agricultural and Food Chemistry, 59(20),
10831-10838.
Mannina, L., Fontanazza, G., Patumi, M., Ansanelli, G., & Segre, A. L. (2001). Italian and
Argentine olive oils: A NMR and gas-chromatographic study. Grasas Aceites, 52, 380-
388.Mannina, L., Sobolev, A. P., & Segre, A. L. (2003). Olive oil as seen by NMR and
chemometrics. Spectroscopy Europe, 15(3), 6-14.
Mannina, L., Sobolev, A. P., Capitani, D., Iaffaldano, N., Rosato, M. P., Ragni, P., Reale,
A., Sorrentino, E., D’Amico, I., & Coppola, R. (2008). NMR metabolic profiling of
organic and aqueous sea bass extracts: implications in the discrimination of wild and
cultured sea bass. Talanta, 77, 433-444.
Mannina, L., Marini, F., Gobbino, M., Sobolev, A. P., & Capitani, D. (2010). NMR and
chemometrics in tracing European olive oils: the case study of Ligurian samples.
Talanta, 80, 2141-2148.
Mannina, L., & Sobolev, A. P. (2011). High resolution NMR characterization of olive oils
in terms of quality, authenticity, and geographical origin. Magnetic Resonance in
Chemistry, 49, S3-S11.
Mannina, L., Sobolev, A. P., Di Lorenzo, A., Vista, A., Tenore, G. C., & Daglia, M. (2015).
Chemical composition of different botanical origin honeys produced by Sicilian black
honeybees (Apis mellifera ssp. sicula). Journal of Agricultural and Food Chemistry,
63, 5864-5874.
Marseglia, A., Caligiani, A., Comino, L., Righi, F., Quarantelli, A., & Palla, G. (2013).
Cyclopropyl and ω-cyclohexyl fatty acids as quality markers of cow milk and cheese.
Food Chemistry, 140, 711-716.
Marseglia, A., Acquotti, D., Consonni, R., Cagliani, L. R., Palla, G., & Caligiani, A.
(2016). HR MAS 1H NMR and chemiometrics as useful tool to assess the geographical
origin of cocoa beans – comparison with solution 1H NMR. Food Research
International, 85, 273-281.Mazzei, P., & Piccolo, A. (2012). 1H HRMAS-NMR
metabolomic to assess quality and traceability of mozzarella cheese from Campania
buffalo milk. Food Chemistry, 132, 1620-1627.
Meisel, H., & FitzGerald, R. J. (2003). Biofunctional peptides from milk proteins: mineral
binding and cytomodulatory effects. Current Pharmaceutical Design, 9(16), 1289-95.
Melnyk, J. P., Wang, S., & Marcone, M. F. (2010). Chemical and biological properties of
the world’s most expensive spice: Saffron. Food Research International, 43, 1981-
1989.
Mendes, E., Brojo Proenca, E., Ferreira, I. M. P. L. V. O., & Ferreira, M. A. (1998). Quality
evaluation of Portuguese honey. Carbohydrate Polymers, 37, 219-223.
Min, S., Harris, L. J., & Krochta, J. M., (2005). Antimicrobial effects of lactoferrin,
lysozyme, and the lactoperoxidase system and edible whey protein films incorporating
the lactoperoxidase system against Salmonella enterica and Escherichia coli O157:H7.
Journal of Food Science, 70, 332-338.
Monakhova, Y. B., Kuballa, T., Leitz, J., Andlauer, C., & Lachenmeier, D. W. (2012).
NMR spectroscopy as a screening tool to validate nutrition labeling of milk, lactose-
free milk, and milk substitutes based on soy and grains. Dairy science & technology,
92(2), 109-120.
Monakhova, Y. B., Godelmann, R., Andlauer, C., Kuballa, T., & Lachenmeier, D. W.
(2013). Identification of imitation cheese and imitation ice cream based on vegetable
fat using NMR spectroscopy and chemometrics. International Journal of Food
Science, 2013, Article ID 367841(pp. 9).
Morris, K. F., & Johnson, C. S., Jr. (1992). Diffusion-ordered two-dimensional nuclear
magnetic resonance spectroscopy. Journal of the American Chemical Society, 114,
3139-3141.
Mucci, A., Parenti, F., Righi, V., & Schenetti, L. (2013). Citron and lemon under the lens
of HR-MAS NMR spectroscopy. Food Chemistry, 141, 3167-3176.
Olsson, I. M., Gottfries, J., & Wold, S. (2004). Chemometrics and Intelligent Laboratories,
73, 37-46.
Ordoudi S. A., & Tsimidou M. Z. (2004). Saffron quality: effect of agricultural practices,
processing and storage. In R. Dris, & S. M Jain (Eds.), Production Practices and
Ramani, J., Wijesinha-Bettoni, R., & Burlingame, B., (2010). Protein quality evaluation
twenty years after the introduction of the protein digestibility corrected amino acid
score method. British Journal of Nutrition, 108, S183-S211.
RASFF (Rapid Alert System for Food and Feed). “RASFF - Food and Feed Safety Alerts.”
Accessed 07 December 2016. http://ec.europa.eu/food/safety/rasff.
Rastrelli, F., Schievano, E., Bagno, A., & Mammi, S. (2003). Magnetic Resonance in
Chemistry, 15(3), 6-14.
Reale, A., Sorrentino, E., Iaffaldano, N., Rosato, M. P., Ragni, P., Coppola, R., Capitani,
D., Sobolev, A. P., Tremonte, P., Succi, M., & Mannina, L. (2008). Effects of ionizing
radiation and modified atmosphere packaging on the shelf life of aqua-cultured sea
bass (Dicentrarchus labrax). World Journal of Microbiology and Biotechnology, 24,
2757-2765.
Renou, J.-P., Bielicki, G., Deponge, C., Gachon, P., Micol, D., & Ritz, P. (2004).
Characterization of animal products according to geographic origin and feeding diet
using nuclear magnetic resonance and isotope ratio mass spectrometry. Part II: Beef
meat. Food Chemistry, 86(2), 251-256.
Ribó, J. M., Crusats, J., El-Hachemi, Z., Feliz, M., Sanchez-Bel, P., & Romojaro, F. (2014).
High-resolution NMR of irradiated almonds. Journal of the American Oil Chemists'
Society, 81(11), 1029-1033.
Righi, V., Parenti, F., Schenetti, L., & Mucci, A. (2016). Mycosporine-like amino acids
and other phytochemicals directly detected by high-resolution NMR on Klamath
(Aphanizomenonflosaquae) blue-green algae. Journal of Agricultural and Food
Chemistry, 64, 6708-6715.
Ritota, M., Casciani, L., Failla, S., & Valentini, M. (2012). HRMAS-NMR spectroscopy
and multivariate analysis meat characterization. Meat Science, 92(4), 754-761.
Rodrigues, D., Santos, C. H., Rocha-Santos, T. A. P., Gomes, A. M., Goodfellow, B. J., &
Freitas, A. C. (2011). Metabolic profiling of potential probiotic or synbiotic cheeses
by nuclear magnetic resonance (NMR) spectroscopy. Journal of Agricultural Food
Chemistry, 59, 4955-4961.
Rosa, A., Scano, P., Melis, M. P., Deiana, M., Atzeri, A., Dessì, M. A. (2009). Oxidative
stability of lipid components of mullet (Mugil cephalus) roe and its product “bottarga.”
Food Chemistry, 115, 891-896.
Rosa, A., Scano, P., Atzeri, A., Deiana, M., Mereu, S., Dessì, M. A. (2012). Effect of
storage conditions on lipid components and color of Mugil cephalus processed roes.
Journal of Food Science, 77 (1), 107-114.Rotondo, A., Salvo, A., Giuffrida, D., Dugo,
G., & Rotondo, E. (2011). NMR analysis of aldehydes in sicilian extra-virgin olive oils
by DPFGSE techniques. Atti della Accademia Peloritana dei Pericolanti-Classe di
Scienze Fisiche, Matematiche e Naturali, 89, C1A8901002 (pp. 8).
Ruyssen, T., Janssens, M., Van Gasse, B., Van Laere, D., Van der Eecken, N., De Meerleer,
M., Vermeiren, L., Van Hoorde, K., Martins, J. C., Uyttendaele, M., & De Vuyst, L.
Scano, P., Anedda, R., Melis, M. P., Lai, A., & Roggio, T. (2011). 1H-and 13C-NMR
characterization of the molecular components of the lipid fraction of Pecorino Sardo
cheese. Journal of the American Oil Chemists' Society, 88(9), 1305-1316.
Scano, P., Rosa, A., Locci, E., Manzo, G., & Dessì, M. A. (2012). Modifications of the 1H
NMR metabolite profile of processed mullet (Mugilcephalus) roes under different
storage conditions. Magnetic Resonance in Chemistry, 50, 436-442.
Schaafsma, G. (2000). The protein digestibility-corrected amino acid score. Journal of
Nutrition, 130(7), 1865S-1867S.
Schaafsma, G. (2012). Advantages and limitations of the protein digestibility-corrected
amino acid score (PDCAAS) as a method for evaluating protein quality in human diets.
British Journal of Nutrition, 108 Suppl 2, S333-S336.
Schievano, E., Pasini, G., Cozzi, G., & Mammi, S. (2008). Identification of the production
chain of asiago d’allevo cheese by Nuclear Magnetic Resonance spectroscopy and
principal component analysis. Journal of Agricultural Food Chemistry, 56(16), 7208-
7214.
Schievano, E., Guardini, K., & Mammi, S. (2009). Fast determination of histamine in
cheese by Nuclear Magnetic Resonance (NMR). Journal of Agricultural Food
Chemistry, 57, 2647-2652.
Schievano, E., Peggion, E., & Mammi, S. (2010). 1H Nuclear Magnetic Resonance spectra
of chloroform extracts of honey for chemometric determination of its botanical origin.
Journal of Agricultural Food Chemistry, 58(1), 57-65.
Schievano, E., Mammi, S., & Menegazzo, I. (2013a). Nuclear Magnetic Resonance as a
Method to Predict the Geographical and Entomological Origin of Pot-Honey. In P. Vit,
S. R. M. Pedro, & D. Roubik (Eds.), Pot-honey a leg. stingless bees, Springer science
& business media (1stedition, pp. 429-445) New York, NY: Springer.
Schievano, E., Morelato, E., Facchin, C., & Mammi, S. (2013b). Characterization of
markers of botanical origin and other compounds extracted from unifloral honeys.
Journal of Agricultural Food Chemistry, 61(8), 1747-1755.
Schievano, E., Finotello, C., De Angelis, E., Mammi, S., & Navarini, L. (2014). Rapid
authentication of coffee blends and quantification of 16-O-Methylcafestol in roasted
coffee beans by Nuclear Magnetic Resonance. Journal of Agricultural and Food
Chemistry, 62, 12309-12314.
Schievano, E., Finotello, C., Navarini, L., & Mammi, S. (2015a). Quantification of caffeine
in human saliva by Nuclear Magnetic Resonance as an alternative method for
cytochrome CYP1A2 phenotyping. Talanta, 140, 36-41.
Schievano, E., Finotello, C., Mammi, S., Illy, Belci, A., Colomban, S., & Navarini, L.
(2015b). Preliminary characterization of monofloral Coffea spp. honey: correlation
between potential biomarkers and pollen content. Journal of Agricultural and Food
Chemistry, 63, 5858-5863.
Schievano, E., Finotello, C., Uddin, J., Mammi, S., & Piana, L. (2016). Objective definition
of monofloral and polyfloral honeys based on NMR metabolomic profiling. Journal of
Agricultural Food Chemistry, 64, 3645-3652.
Schwarzinger, S., Brauer, F., Rösch, P., & Kämpf, B. (2015). Food fraud: Testing honey
with NMR-profiling. New Food Magazine, 6.Sciubba, F., Capuani, G., Di Cocco, M.
E., Avanzato, D., & Delfini, M. (2014a). Nuclear magnetic resonance analysis of water
soluble metabolites allows the geographic discrimination of pistachios (Pistaciavera).
Food Research International, 62, 66-73.Sciubba, F., Di Cocco, M.E., Gianferri, R.,
Impellizzeri, D., Mannina, L., De Salvador, F. R., Venditti, A., & Delfini, M. (2014b).
Metabolic profile of different Italian cultivars of hazelnut (Corylusavellana) by nuclear
magnetic resonance spectroscopy. Natural Product Research, 28, (14), 1075-1081.
Sciubba, F., Di Cocco, M. E., Gianferri, R., Capuani, G., De Salvador, F. R., Fontanari,
M., Gorietti, D., & Delfini, M. (2015). Nuclear Magnetic Resonance-based metabolic
comparative analysis of two apple varieties with different resistances to apple scab
attacks. Journal of Agricultural and Food Chemistry, 63, 8339-8347.
Sciubba, F., Avanzato, D., Vaccaro, A., Capuani, G., Spagnoli, M., Di Cocco, M. E.,
Tzareva, I. N., & Delfini, M. (2016). Monitoring of pistachio (Pistacia Vera) ripening
by high field nuclear magnetic resonance spectroscopy. Natural Product Research, 1-
8.
Shintu, L., & Caldarelli, S. (2005). High-resolution MAS NMR and chemometrics:
Characterization of the ripening of parmigiano reggiano cheese. Journal of
Agricultural Food Chemistry, 53, 4026-4031.
Shintu, L., & Caldarelli, S. (2006). Toward the determination of the geographical origin of
emmental(er) cheese via high resolution MAS MNR: A preliminary investigation.
Journal of Agricultural Food Chemistry, 54, 4148-4154.
Shintu, L., Caldarelli, S., & Franke, B. M. (2007). Pre-selection of potential molecular
markers for the geographic origin of dried beef by HR-MAS NMR spectroscopy. Meat
Science, 76(4), 700-707.
Sinanoglou, V. J., Kokkotou, K., Fotakis, C., Strati, I., Proestos, C., & Zoumpoulakis, P.
(2014). Monitoring the quality of γ-irradiated macadamia nuts based on lipid profile
analysis and Chemometrics. Traceability models of irradiated samples. Food Research
International, 60, 38-47.
Sinanoglou, V. J., Strati, I. F., Kokkotou, K., Lantzouraki, D., Makris, C., & Zoumpoulakis,
P. (2015). GC-FID and NMR spectroscopic studies on gamma irradiated walnut lipids.
Journal of Spectroscopy, 2015, ID 532762(pp. 10).
Smilowitz, J. T., O’Sullivan, A., Barile, D., German, J. B., Lönnerdal, B., & Slupsky, C.
M. (2013). The human milk metabolome reveals diverse oligosaccharide profiles. The
Journal of Nutrition, 143(11), 1709-1718.
Sobolev, A. P., Segre, A. L., & Lamanna R. (2003). Proton high-field NMR study of tomato
juice. Magnetic Resonance Chemictry, 41, 237-245.
Sobolev, A. P., Mannina, L., Proietti, N., Carradori, S., Daglia, M., Giusti, A. M.,
Antiochia, R., & Capitani, D. (2015). Untargeted NMR-based methodology in the
study of fruits metabolites. Molecules, 20(3), 4088-4108.Soulet, S., Vercauteren, J.,
Troplin, P., & Raybaud, G. (2005). NMR determinationof the geographic origin of
cocoa beans and their products. Fr. Demande FR 2857450 A1 20050114.
Speer, K., & Kölling-Speer, I. (2006). The lipid fraction of the coffee bean. Brazilian
Journal of Plant Physiology, 18(1), 201-216.
Speranza, G., & Dadà, G. (1984). 13 cis-Crocin: a new crocinoid of saffron. Gazzetta
Chimica Italiana, 114, 189-192.
Spevacek, A. R., Smilowitz, J. T., Chin, E. L., Underwood, M. A., German, J. B., &
Slupsky, C. M. (2015). Infant maturity at birth reveals minor differences in the
maternal milk metabolome in the first month of lactation. The Journal of Nutrition,
145(8), 1698-1708.
Stark, T., & Hofmann, T. (2005). Isolation, structure determination, synthesis, and sensory
activity of n-phenylpropenoyl-l-amino acids from cocoa (Theobroma cacao). Journal
of Agricultural and Food Chemistry, 53(13), 5419-5428.
Stark, T., & Hofmann, T. (2006). Application of a Molecular Sensory Science Approach
to Alkalized Cocoa (Theobroma cacao): Structure determination and sensory activity
of nonenzymatically c-glycosylated flavan-3-ols. Journal of Agricultural and Food
Chemistry, 54(25), 9510-9521.
Stefanova, R., Toshkov, S., Vasilev, N. V., Vassilev, N. G., & Marekov, I. N. (2011). Effect
of gamma-ray irradiation on the fatty acid profile of irradiated beef meat. Food
Chemistry, 127(2), 461-466.
Stone, H. & Sidel, J. L. (1998). Quantitative descriptive analysis: Developments,
applications, and the future. Food Technology, 52(8), 48-52.
Straadt, I. K., Aaslyng, M. D., & Bertram, H. C. (2011). Assessment of meat quality by
NMR-an investigation of pork products originating from different breeds. Magnetic
Resonance in Chemistry, 49(S1), S71-S78.
Straadt, I. K., Aaslyng, M. D., & Bertram, H. C. (2014). An NMR-based metabolomics
study of pork from different crossbreeds and relation to sensory perception. Meat
Science, 96(2PA), 719-728.
Straubinger, M., Bau, B., Eckstein, S., Fink, M., & Winterhalder, P. (1998). Identification
of novel glycosidic aroma precursors in saffron (Crocus sativus L.). Journal of
Agricultural and Food Chemistry, 46, 3238-3243.
Straubinger, M., Jezussek, M., Waibel, R., & Winterhalder, P. (1997). Novel glycosidic
constituents from saffron. Journal of Agricultural and Food Chemistry, 45, 1678-1681.
Suárez-Ulloa, V., Fernández-Tajes, J., Manfrin, C., Gerdol, M., Venier, P., & Eirín-López,
J. M. (2013). Bivalve omics: State of the art and potential applications for the
biomonitoring of harmful marine compounds. Marine Drugs, 11, 4370-4389.
Sun, L. L., Gao, W., Zhang, M. M., Li, C., Wang, A. G., Su, Y. L., & Ji, T. F. (2014).
Composition and antioxidant activity of the anthocyanins of the fruit of
BerberisheteropodaSchrenk. Molecules, 19, 19078-19096.
Sundekilde, U. K., Frederiksen, P. D., Clausen, M. R., Larsen, L. B., & Bertram, H. C.
(2011). Relationship between the metabolite profile and technological properties of
bovine milk from two dairy breeds elucidated by NMR-based metabolomics. Journal
of Agricultural Food Chemistry, 59(13), 7360-7367.
Sundekilde, U. K., Larsen, L. B., & Bertram, H. C. (2013a). NMR-Based milk
metabolomics. Metabolites, 3, 204-222.
Sundekilde, U. K., Poulsen, N. A., Larsen, L. B., & Bertram H. C. (2013b). Nuclear
magnetic resonance metabonomics reveals strong association between milk
metabolites and somatic cell count in bovine milk. Journal of Dairy Science, 96, 290-
299.
Tacon, A. G. J., & Metian, M. (2013). Fish Matters: importance of aquatic foods in human
nutrition and global food supply. Reviews in Fisheries Science, 21, 22-38.
Tang, J. E., Moore, D. R., Kujbida, G. W., Tarnopolsky, M. A., & Phillips, S. M., (2009).
Ingestion of whey hydrolysate, casein, or soy protein isolate: effects on mixed muscle
protein synthesis at rest and following resistance exercise in young men. Journal of
Applied Physiology, 107, 987-992.Tarantilis, P. A., Polissiou, M., & Manfait, M.
(1994). Separation of picrocrocin, cis-trans-crocins and safranal of saffron using high-
performance liquid chromatography with photodiode-array detection. Journal of
Chromatography A, 664, 55-61.
Tarantilis, P. A., Tsoupras, G., & Polissiou, M. (1995). Determination of saffron (Crocus
sativus L.) components in crude plant extract using high-performance liquid
chromatography-UV-visible photodiode-array detection-mass spectrometry. Journal
of Chromatography A, 699, 107-118.
Tarantilis, P. A., & Polissiou, M. G. (2004). Chemical analysis and antitumor activity of
natural and semi-natural carotenoids of saffron. Acta Horticulturae, 650, 447-461.
Tavares, K. M., Lima, A. R., Nunes C. A., Silva V. A. Mendes E., Casal S., & Pereira R.
G. F. A. (2016). Free tocopherols as chemical markers for Arabica coffee adulteration
with maize and coffee by-products. Food Control, 70, 318-324.
Thilsted, S. H., Thorne-Lyman, A., Webb, P., Bogard, J. R., Subasinghe, R., Phillips, M.
J., & Allison, E. H. (2016). Sustaining healthy diets: The role of capture fisheries and
aquaculture for improving nutrition in the post-2015 era. Food Policy, 61, 126-131.
Thomas, F., & Jamin, E. (2009). H-2 NMR and C-13-IRMS analyses of acetic acid from
vinegar, O-18-IRMS analysis of water in vinegar: International collaborative study
report. Analytica Chimica Acta, 649, 98-105.
Tikunov, A. P., Johnson, C. B., Lee, H., Stoskopf, M. K., & Macdonald, J. M. (2010).
Metabolomic investigations of American oysters using 1H-NMR spectroscopy. Marine
Drugs, 8, 2578-2596.
Torelli, A., Marieschi, M., & Bruni, R. (2014). Authentication of saffron (Crocus sativus
L.) in different processed, retail products by means of SCAR markers. Food Control,
36, 126-131.
Tsiafoulis, C. G., Skarlas, T., Tzamaloukas, O., Miltiadou, D., & Gerothanassis, I. P.
(2014). Direct nuclear magnetic resonance identification and quantification of
geometric isomers of conjugated linoleic acid in milk lipid fraction without
derivatization steps: Overcoming sensitivity and resolution barriers. Analytica Chimica
Acta, 821, 62-71.
Van Calsteren, M. R., Bissonnette, M. C., Cormier, F., Dufresne, C., Ichi, T., LeBlanc, J.
C. Y., Perreault, D., & Roewer, U. (1997). Spectroscopic characterization of crocetin
derivatives from Crocus sativus and Gardenia jasminoides. Journal of Agricultural and
Food Chemistry, 45, 1055-1061.
Venditti, G., Schievano, E., Navarini, L., & Mammi, S. (2011). Water mobility and
distribution in green Coffee probed by Time-Domain Nuclear Magnetic Resonance.
Food Biophysics, 6, 321-326.
Vidal, N. P., Picone, G., Goicoechea, E., Laghi, L., Manzanos M. J., Danesi F., Bordoni,
A., Capozzi, F., & Guillén, M. D. (2016). Metabolite release and protein hydrolysis
during the in vitro digestion of cooked sea bass fillets. A study by 1H NMR. Food
Research International, 88, 293-301.
Wang, X., Dai, Q., Fang, C., He, J., & Fang, R. (2016). Comparison of the meat metabolite
composition of Linwu and Pekin ducks using 600 MHz 1H nuclear magnetic resonance
spectroscopy. Poultry science, 8 pp.
Wei, F. F., Furihata, K., Hu, F. Y., Miyakawa, T., & Tanokura, M. (2011). Two-
dimensional H-1-C-13 Nuclear Magnetic Resonance (NMR)-based comprehensive
analysis of roasted coffee bean extract. Journal of Agricultural and Food Chemistry,
59, 9065-9073.
Wei, F. F., Furihata, K., Koda, M., Hu, F. Y., Miyakawa, T., & Tanokura, M. (2012).
Roasting process of coffee beans as studied by Nuclear Magnetic Resonance: time
course of changes in composition. Journal of Agricultural and Food Chemistry, 60,
1005-1012.
Wei, F. F., Furihata, K., Miyakawa, T., & Tanokura, M. (2014). A pilot study of NMR-
based sensory prediction of roasted coffee bean extracts. Food Chemistry, 152, 363-
369.
Werner, R. A., & Rossmann, A. (2015). Multi element (C, H, O) stable isotope analysis for
the authentication of balsamic vinegars. Isotopes in Enviromental and Health Studies,
51, 58-67.
Winterhalter, P., & Straubinger, M. (2000). Saffron - Renewed interest in an ancient spice.
Food Reviews International, 16, 39-59.
Wyzgoski, F. J., Paudel, L., Rinaldi, P. L., Reese, R. N., Ozgen, M., Tulio, A. Z., Miller,
A. R., Scheerens, J. C., & Hardy, J. K. (2010). Modeling relationships among active
Online References
Chapter 7
1
Corresponding Author Email: donatella.capitani@cnr.it.
2
Chemometrics & Analytical Technology, Department of Food Science,
University of Copenhagen, Copenhagen, Denmark
3
Dipartimento di Scienze e Tecnologie Chimiche,
Università di Roma “Tor Vergata,” Rome, Italy
4
Dipartimento di Chimica e Tecnologie del Farmaco,
Sapienza Università di Roma, Rome, Italy
5
Istituto per lo Studio delle Macromolecole, Lab NMR, CNR, Milan, Italy
6
Dipartimento di Scienze degli Alimenti,
Università degli Studi di Parma, Parma, Italy
7
Dipartimento di Scienze e Tecnologie Agro-Alimentari,
Università di Bologna, Cesena, Italy
8
Dipartimento di Scienze Chimiche e Geologiche,
Università di Cagliari, Monserrato (Cagliari), Italy
9
Consiglio per la Ricerca in Agricoltura e l’Analisi
dell’Economia Agraria – Centro di Ricerca per lo Studio
delle Relazioni tra Pianta e Suolo (CREA-RPS), Rome, Italy
10
Dipartimento di Scienze e Tecnologie Biologiche ed Ambientali,
Università del Salento, Campus Ecotekne, Lecce, Italy
11
CNR-IPCF, Istituto per i Processi Chimico-Fisici del
CNR di Messina, Messina, Italy
12
Dipartimento di Chimica, Sapienza Università di Roma, Rome, Italy
13
Dipartimento di Ingegneria Civile, Ambientale, del Territorio,
Edile e di Chimica (DICATECh), Politecnico di Bari, Bari, Italy
14
Italian Institute of Technologies, @CNLS Sapienza, Rome, Italy
15
Dipartimento di Chimica, Università degli Studi di Bari
“Aldo Moro,” Bari, Italy
16
Centro di Ricerca di Risonanze Magnetiche CERM,
Università di Firenze, Sesto Fiorentino (Florence), Italy
17
Consorzio Interuniversitario per lo Sviluppo dei Sistemi a
Grande Interfase - CSGI, Sesto Fiorentino (Florence), Italy
18
Dipartimento di Scienze Chimiche, Università degli
Studi di Padova, Padova, Italy
19
Centro Interdipartimentale per la Risonanza Magnetica
Nucleare per l’Ambiente, l’Agro-Alimentare ed i Nuovi Materiali (CERMANU),
Università di Napoli Federico II, Portici (Naples), Italy
20
Istituto Zooprofilattico Sperimentale dell'Abruzzo e del
Molise “G. Caporale,” Teramo, Italy
21
Dipartimento di Scienze Chimiche e Geologiche,
Università di Modena e Reggio Emilia, Modena, Italy
22
Dipartimento di Farmacia, Università degli Studi di Salerno, Salerno, Italy
23
Dipartimento di Farmacia, Università di Napoli Federico II, Naples, Italy
24
Dipartimento di Scienze per la Qualità della Vita,
Università di Bologna, Rimini, Italy
25
Dipartimento di Scienze Biomediche, Odontoiatriche e delle Immagini
Morfologiche e Funzionali, Università degli Studi di Messina, Messina, Italy
26
Dipartimento di Scienza Applicata e Tecnologia,
Politecnico di Torino, Tourin, Italy
27
Dipartimento di Medicina Sperimentale e Clinica,
Università di Firenze, Florence, Italy
28
Dipartimento MIFT, sezione di Fisica, Università di Messina, Messina, Italy
ABSTRACT
1. INTRODUCTION
Low field 1H NMR relaxometry is a suitable tool to study the most abundant
components of intact foodstuffs from measurements of relaxation parameters and
amplitudes of the NMR signals. Information on food microstructure such as water
compartments and diffusion can be obtained by detecting proton signals dominated by H2O
contained in foodstuffs. This technique does not require sample pretreatment and once
developed can be easily used to quality control applications.
Information at the microns level can be obtained by NMR, based on longitudinal
relaxation time (T1), water self-diffusion (Dw) or T2. T2-weighted signals are undoubtedly
those most often employed (Colnago et al., 2015), analyzed by fitting procedures of the so
obtained curve, in order to describe separately the physical food compartments that
contributed to them. The following two examples allow appreciating the power of these
protocols. Even if muscle fibers are characterized by a diameter as thin as a few tens of
microns, it is nowadays routinely possible to observe and study separately intra and extra
myofibrillar water (Petracci et al., 2014). Again, even if the pericarp cell of a kiwifruit has
a diameter of about 100 microns only, the signal due to vacuole, cytoplasm and intercellular
space can be observed separately (Tylewicz et al., 2011).
Historically, the main obstacle in the so described application of NMR relaxation has
been the ascription of the T2 weighted signals to specific compartments of the samples
under investigation. As the decay of the NMR signal due to T2 is exponential, the most
intuitive way to do so consists in its fitting towards the sum of a handful of exponential
curves. This has been done successfully in many cases, among which those just mentioned
for meat (Petracci et al., 2014) and kiwifruit (Tylewicz et al., 2011), or the one of albumen,
in which the fluid and thick portions can be separately observed (Laghi et al., 2005). There
are several reasons to consider this procedure to extract useful information from T2
weighted signals as sub-optimal. One of them is that describing the signal of water in a
certain compartment with a single exponential curve implies that the state of water is
perfectly homogeneous across the entire sample consider, which can be a too gross
approximation.
In order to overall the limits intrinsic to the fitting towards a discrete number of
exponential curves, several researchers attempted to present T 2 decays as T2 quasi-
continuous distribution data (Gao et al., 2016), similarly to the high-resolution spectra
counterparts. In order to do so, Laplace inversion is used. Unfortunately, in the presence of
finite and/or noisy data, such inversion is ill posed, i.e., it does not give rise to a unique
solution. In order to overcome the impasse, several regularization methods have been
described, with the aim of finding a single solution, which could be considered as the
“best,” according to some criteria (Borgia et al., 1998). Prange and Song tried to find an
out-of-the-box solution to the task of finding the best regularization method to Laplace
inversion, by actually giving up looking for the “best” T 2 distribution (Prange & Song,
2009). At its place, they setup a Monte Carlo algorithm generating probable solutions, from
which the statistical properties of the solutions themselves could be analyzed. Interestingly,
they found that the mean solution spectrum was smooth and close to the regularized
solution, even when individual solutions could be spiky. Zou et al. faced from an again
different perspective the problem of the Laplace inversion of T2 weighted signals, because
they focused on the uncertainty introduced in the T2 distribution generated, by applying a
frequentist method (Zou et al., 2016). Echo and T2 distribution amplitudes on one side, T2
distribution slopes and curvatures on the other served as objective basis for the purpose.
Signal-to-noise ratio was found to be the characteristic of T 2 decays mainly affecting the
uncertainty of the T2 distribution, while norm-smoothing method performed better than
curvature smoothing.
In order to extract useful information from T 2 weighted signals by avoiding the
limitations of discrete fitting and quasi-continuous T2 distribution calculation, researchers
have tried to rely on multivariate analysis. This choice, not based on a priori outlined
model, has been attempted in particular in cases of NMR applications described as
dynamic, where NMR observations are performed recursively in order to follow the
progressive changes a food undergoes along processing or storage. Curiously, even if the
extraction of each piece of relevant information by multivariate analysis seems a very
powerful protocol for the elaboration of relaxation time signals, the idea has been exploited
only rarely in the past. Three cases are worth mentioning, in this context. Engelsen et al.
looked for the consequences of baking on breadcrumb, by simulating baking inside a TD-
NMR instrument (Engelsen et al., 2001). In order to highlight the main changes in water
state related to core temperature increase, they analyzed the T2 weighted signals by means
of principal component analysis (PCA). They found two abrupt changes in the water state
of breadcrumb, ascribed to the onset and to the end of gelatinization, respectively. In
parallel, the researchers managed to relate water state, observed by TD-NMR, to bread
texture changes upon staling. In order to do so, they applied successfully a further
chemometric analysis, partial least square regression (PLSR). Micklander et al. looked for
the main transitions of state meat water undergoes upon cooking (Micklander et al., 2002).
In order to do so, they simulated cooking process directly inside a TD-NMR instrument,
operating at 23.2 MHz, and performed data reduction of the so obtained T2-weighted curves
by beans of PCA. While the first component of this multivariate model mainly incorporated
the effect of temperature increase, the second and third showed abrupt changes that were
ascribed to the denaturation of sarcoplasmic proteins and to the shrinkage of the myofibrils.
Laghi et al. simulated in vitro the digestion of Parmigiano-Reggiano cheese, which selling
strategy often revolves around the high digestibility of its protein fraction. In order to
characterize the amount of protein fraction liberated, together with its overall profile, they
performed the digestion on two samples at different ripening levels and highlighted the
differences among the digestates from the two by PCA, applied to centered unscaled T2
weighted signals, obtained at 20 MHz on a Bruker “The Minispec” benchtop instrument
(Laghi et al., 2013).
In very recent years the group held by Luiz Alberto Colnago, operating at the Brazilian
Agricultural Research Corporation, has operated an authentic revamping of the idea, by
publishing a series of works that seem to have the proper characteristics to become
milestones if the field. In one of them (Pereira et al., 2013), the total soluble solid content
of plums was measured by means of a refractometer. In parallel T 2-weighted signals have
been registered at 0.23T on intact fruits. Soft independent modeling of class analogy
(SIMCA), performed on NMR data reduced by PCA, was able to assign the plums to two
classes according to their low (9 to 12%) or high (13 to 22%) soluble solids content, with
a 96 accuracy. In a conceptually similar paper (Pereira & Colnago, 2012), water content
was registered by gravimetric method on samples of beef and then accurately predicted by
T2-weighted NMR signals, again at 0.23T, by means of multivariate analysis. In the
specific case, partial least squares (PLS) and principal component regression (PCR)
showed comparable prediction ability.
T1 and Dw have been far less employed for food microstructure elucidation than T 2,
with a few exceptions (Tylewicz et al., 2016; Santagapita et al., 2013) because the former
requires longer acquisition times and the latter often needs poorly parsimonious models for
interpretation. This state of art can be foreseen to change substantially due to Colnago work
(Colnago et al., 2015), which described a fast protocol to acquire T 1 and T2 relaxation
signals simultaneously, with the TD-NMR instruments available at present.
Horticultural products are challenging food with respect to maintaining quality during
the food chain between field and consumer. Consumer expects its fruit and vegetables to
be of consistent high quality, at optimum ripeness, juiciness and texture, and free from
internal and external quality defects. Whether, harvesting and storage conditions affect the
quality of fruit and vegetables. Off-season fruits coming out of long term cold storage show
particularly high degree of quality variation due to a progressive development of fungal
infection, over ripeness and bruises. For this reason, most fruits and vegetables need to be
examined and selected before to be transported, processed or sold. Therefore, a strong need
exists for rapid and cost-effective quality control tools. Low field proton NMR relaxometry
has become an important tool in quality control analysis of agri-food products. In particular
relaxation times (T1 and T2) and molecular self-diffusion coefficient (D) NMR
measurements have been employed for studying the dynamic of water molecules as well
as the subcellular water distribution within plant tissues and water molecule transport
properties in sub-cellular compartments (Callaghan et al., 1979; Hills & Belton, 1989;
Snaar & Van As, 1992). The role and state of water in food is generally agreed to be of
importance in a range of properties such as texture, microbial growth rates, storage,
deterioration, etc (Duckworth, 1975; Belton, 1984; Hills et al., 1990). Changes in water
NMR dynamics during ripening, processing and storage operations on fruits and vegetables
can, in principle, reveal sub cellular modifications and contribute to a microscopic
understanding of these processes (Gil et al., 1996).
Low-field NMR relaxometry and diffusometry have been applied in the quality
inspection of fruits, and to monitor the fruit maturity and the effects of processing on
morphology. Apple has been one of the most studied fruits. The parenchyma tissue of apple
is representative of multicompartment cellular tissue that has been much studied by one
dimensional relaxation techniques (Hills & Clark, 2003). Four relaxation peaks can be seen
and assigned to the different cell compartments (Snaar & Van As, 1992) (Figure 1).
Figure 1. Distribution of transverse water proton relaxation times in fresh ‘Red Delicious’ apple tissue
measured at spectrometer frequency of 23.4MHz with a CPMG 90–180° pulse spacing of 200 μs.
Numbers refers to relaxation peaks of the water proton in the different cell compartments: 1, the vacuole;
2 and 3, cytoplasm and extra-cellular compartment and 4, cell wall. With the permission of reference
Marigheto et al., 2008.
The peak with shortest T2 (peak 4) probably arises from more rigid components of the
cell walls in the apple matrix. Peaks 3 and 2 can be associated with the water in the
cytoplasm and extracellular compartments. The peak with the longest T 2 (peak 1) is
associated with the water in the vacuole. Diffusion of water between the various subcellular
(vacuolar, cytoplasmic) and extra-cellular water compartments which averages the
magnetization to an extent that depends on cell morphology and membrane permeability.
This aspect complicates the peak assignment.
Keener et al. have investigated the relationship between low-field (0.13 T; 5.4 MHz)
apparent water proton diffusion coefficients (Dw), the dominant CPMG T 2 measured with
a pulse spacing of 2 ms, and degree of ripeness (soluble solids content) as measured by
refractometry (Keener et al., 1999).
The following trends were reported:
a) In healthy “Golden Delicious” and “Granny Smith” apples, the T 2 decreased with
increasing soluble solid content (Brix) in healthy apples.
b) The water diffusivity decreased with increasing soluble solid content in all
samples, both bruised and healthy. This is no doubt a result of the increasing
viscosity of more concentrated sugar solutions.
c) Internal defects caused by bruising, watercore, and internal browning resulted in a
decrease in T2 in all varieties at this low field strength.
distribution between the cell compartment and it could be exploited in the development of
online NMR sensors of fruit quality.
Figure 2. Examples of mealy apple (first and second line, the first one also showing internal breakdown)
and intermediate and fresh fruits (third and fourth line respectively). With the permission of reference
Barreiro et al., 2000.
Figure 3. Experimental T1–T2 cross correlation spectra of apple tissue measured at 23.4MHz with a
CPMG 90–180° pulse spacing of 200 s. (a) Tissue from fresh apple (b) tissue from mealy apple. 1,
vacuole; 2 and 3, cytoplasm and extracellular compartment; 4, cell wall; 5, starch; 6, starch or protein or
vascular tissue; 7, starch; 8, pectin; 9, baseline artifact and 10, exchange peak. The diagonal line denotes
T1 = T2. With the permission of reference Marigheto et al., 2008.
Several 1D and 2D relaxation and diffusion protocols were also tested to find the most
sensitive technique for the detection of mealiness and ripening in apples (Barreiro et al.,
1999; Barreiro et al., 2002). The loss of membrane integrity upon internal browning in
apple and pear tissues was manifested by changes in relaxation time distributions in T 2–T1
and T2–D correlation measurements (Hernandez-Sanchez et al., 2007).
Low field NMR relaxometry and diffusometry have been applied to assess internal
browning and watercore in apples. Internal browning disorder manifests itself as light and
dark brown patches throughout the cortex and core. The disorder is induced by high CO2
concentrations, especially during storage in modified atmospheres, but canal so appear in
unpicked fruit on the tree. Watercore is a physiological disorder affecting apples in which
intercellular air spaces in the flesh adjacent to the vasculature become filled with fluid
having elevated sorbitol and sucrose concentrations (Simons, 1968; Bowen & Watkins,
1997; Kumpoun et al., 2003). Although watercore may disappear during storage in some
apple varieties, it can develop into a severe form of internal browning in others. Currently,
apples with internal browning cannot easily be sorted from good apples when the defect
does not affect their external appearance. In general, lots with excessive levels of internal
browning cannot be packed for the fresh market. In apple varieties where watercore causes
problems during storage, non destructive detection of watercore would allow defective
apples to be sorted out and marketed before the unaffected apples are placed in storage.
Although internal browning or watercore had no effect on the self-diffusion coefficient
of water, differences in T2 values between healthy and defective apple tissue were
observed. Jung et al. evaluated the use of a low-field (0.1256 T) 1H NMR sensor for
detection of two types of internal disorder found in apples (Jung et al., 1998). They
conducted Carr–Purcell–Meiboom–Gill (CPMG) T2 tests on whole apples that were
healthy or that had either internal browning or watercore. The T 2 value of the apples with
internal browning decreased as the severity of internal browning increased. By contrast,
the T2 values of the apples with watercore increased as the severity of watercore increased.
The authors concluded that T2 measurement for detection of apples with severe levels of
internal browning should be possible. Recently Chayaprasert and Stroshine developed a
rapid sensing based on low-cost low-field proton magnetic resonance sensor was to assess
internal browning in whole apples (Chayaprasert & Stroshine, 2005). The MR sensing
system consisted of a permanent magnet equipped with a conveyor belt. Although the T 2
relaxation curves were affected by the apple’s motion, differences between apparently
healthy apples and those with internal browning could be distinguished using apparent T2
values. A similar sensor was used by Cho et al. for on-line detection of internal browning
and watercore in apples (Cho et al., 2008). Changes in T 2 components associated with
different water compartments in affected apples could be associated with a movement of
water from vacuoles into cytoplasm and extracellular spaces (Cho et al., 2008).
Avocado fruits ripen only after picking and, therefore should be harvested when
mature. Unfortunately, there are no visible external changes in the fruit to indicate maturity
but it has been shown that increasing oil content and decreasing water content correlate
closely with sensory measures of maturity and that the oil content correlates well with the
increasing dry weight during maturation. The relationship between NMR parameters and
oil content and /or dry weight in avocado has been investigated by MR techniques (Chen,
1993). Working at a high-filed of 2T (85 MHz) Chen and co-workers showed that T1 and
T2 of e water decrease linearly with increasing dry weight with satisfactory correlation
coefficients.
Besides standard T2 CPMG and T1 methods, more sophisticated relaxation and
diffusion NMR techniques were also tested to assess feasibility of fast online assessment
of quality factors such as maturity, oil content and presence of hard lumps in avocado
(Marigheto et al., 2005). Two-dimensional T2–T1 correlation measurements proved to be
adequate for determination of oil content in avocado tissue, but the long acquisition time
excluded it from online implementation. They are therefore the preferred method for
noninvasive off-line quality control.
Effects of banana ripening during storage were studied by relaxometry (CPMG),
diffusometry (PFGSE) and a combination of (PFGMSE-CPMG) investigated by Raffo and
coworkers (Raffo et al., 2005). Following the methodology of earlier work on apple and
potato tissue the CPMG echo decay for banana were deconvoluted into three components
corresponding to cellular, cytoplasmic and vacuolar water (Hills & Le Floch, 1994; Hills
& Remigerau, 1997). It was then found that the cytoplasmic and vacuolar water T2 showed
significant increases with storage (up to 7 days). The changes were explained as the
progressive enzymatic hydrolysis of starch granules during ripening. Moreover, the
observed water self-diffusion coefficient decrease is related to sugar accumulations as
starch hydrolysis proceeds.
Internal browning in pears has been studied T1-T2 correlation spectroscopy at a proton
frequency of 300.15 MHz. Affected tissue has a shorter transverse relaxation rate compared
to healthy tissue especially at higher magnetic field strength. Tissue disintegration as well
as water evaporation appeared to be the main reason for this difference. Like apple the T2-
D spectra of fresh pears and pears with internal browning both showed two peaks
corresponding to water in the vacuole and from cytoplasm. In all cases examined vacuolar
T2 and its diffusion coefficient of the internally browned pear are longer than those
measured in fresh pear (Hills, 2006).
Kiwifruit is another fruit that has been much studied. Many study performed by MRI
have been focused on the changes in relaxation time and diffusion maps in kiwifruits as
they ripened (Clark et al., 1998a). Measurements were made at 2T (85MHz) and T1 and T2
maps were reported over a 30-days ripening period. It was shown that all relaxation times
show significant increase during ripening, especially in the flesh and locule regions. TD-
NMR has also found widespread use in mechanistic studies on microstructural effects of
processing routes used to minimize food degradation such as air-drying, freeze-drying and
osmotic drying (Van Duynhoven et al., 2009). Osmotic dehydration (OD) is a widely used
method to partially remove water from fruits by immersion of cellular tissue in hypertonic
aqueous sugar solutions. This process is particularly common as a pre-treatment before air-
drying or to obtain minimally processed fruit and vegetables products. In the case of
kiwifruit, the OD is employed to increase the product shelf life, but since by itself is not
enough adequate, generally it is combined with other stronger stabilizing methods such as
freezing or air-, freeze-, vacuum-drying (Rahman & Conrad, 2007). The combined
application of T2 and water self-diffusion coefficient measurements allowed having an
insight of the changes in the cellular compartments of kiwifruit’s outer pericarp promoted
by OD. T2 measurements enabled the quantification of the protons located in the three main
cell compartments within the mobile water is located: vacuole, cytoplasm plus extracellular
space, and cell wall. Water self-diffusion coefficient evaluated through a single component
model was found to be a fast technique to follow OD treatment consequences along time
and to highlight the different response to the treatment from fruits of different ripeness.
Low Field NMR results showed that the OD process influenced water mobility
characteristics within the vacuole and cytoplasm plus extracellular space compartments.
The possibility to highlight how the response to OD treatment was modulated by fruit
ripeness suggests that low field NMR can represent a powerful and versatile tool to
investigate the behavior of vegetable tissues during minimal processing (Santagapita et al.,
2013).
Dry matter content is an important quality parameter for many fruits and vegetables
and it has been used to determine in a non-invasive manner by multi-variate modeling of
CPMG decays of potatoes (Thybo et al., 2003). Multi-variate models based on CPMG
decays obtained from raw potatoes successfully predicted the sensory texture (Thybo et al.,
2000) and other quality parameters of cooked potatoes (Povlsen, 2003).
TD-NMR relaxometry was used to study air-drying (Marques et al., 1991a; Marques
et al., 1991b) and freeze-drying of carrots (Hills & Nott, 1999).
Relaxometry could detect sublimation of the frozen core and removal of non frozen
water during freeze-drying. Also anomalous freeze-drying of potato could be detected.
These studies were followed by more sophisticated approaches where the concepts of non-
freezing water and three-compartment relaxation/diffusion models were deployed. CPMG
relaxation measurements localized non-freezing water in cellular tissue of potato in cell
walls and starch granules. T2 relaxation measurements were used to study changes in sub-
cellular water compartmentation and cell membrane integrity after applying air-drying,
freeze-drying, freeze–thaw processing and rehydration in carrot parenchyma tissue (Hills
& Nott, 1999), which, unlike potato (Hills & Lefloch, 1994) and apples (Hills &
Remigereau, 1997) does not contain large numbers of starch granules or intercellular air
gaps. Carrot tissue distinguishes itself due to the high levels of dissolved sugar in the
vacuole and this strongly affects relaxation behavior during drying and freezing. The
observed decrease in relaxation times during drying was a consequence of vacuolar
shrinkage and the progressive concentration of vacuolar sugars.
Time Domain (TD) NMR is being used throughout all food science and technology
areas since it is a powerful analytical technique employable to investigate the physical and
chemical properties and texture of many substances of agricultural relevance. TD NMR is
implemented on cost-effective and easy-to-use bench top equipment and does not require
pre-treatment or substantial changes (procedures of extraction, solubilization, etc.) of the
samples.
A wide range of TD NMR applications based on FID analysis, relaxometry and self-
diffusion coefficients determinations have been developed in the research area to cover all
food supply chain and also applied to food industry, in research and development but also
quality and control process). Notwithstanding a relatively few published academic works,
TD NMR represents an excellent alternative to some traditional methods of food analysis
(Hills, 2006).
TD NMR is particularly useful as a tool to study dairy products. In particular, it has
been used to characterize the fat and water in cheese, by NMR signal analysis, the fat and
moisture content, by FID-spin-echo application, as well as the diffusion domains and their
distribution, by relaxometry determinations. (Mariette & Lucas, 2005; Brosio & Gianferri,
2011) In these ways, it is also able to investigate the cheese texture attributes by water state
and distribution study (by relaxometry and diffusion measurements), and furthermore their
change as a function of the ripening, shelf-life or production processes, providing a rapid
method for assessment of cheeses quality. (Brosio & Gianferri, 2011)
Since cheese matrix is extremely complex, several studies have been carried out
(Gianferri et al., 2007a; Gianferri et al., 2007b; Castell-Palou et al., 2011) in order to
developed models to evaluate macroscopic physical parameters by T1 and T2 relaxation
times measurements, while auto-diffusion coefficients do not require any interpretive
model (Brosio et al., 2008; Brosio & Gianferri, 2011).
Relaxation measurements (by Carr-Purcell-Meiboom-Gill pulse sequence) represent
most widespread TD NMR application, but they require adequate data-analysis to obtained
structural and compositional information. In fact, in heterogeneous food systems different
NMR relaxation times can be measured due to presence more abundant components, i.e.,
water and lipids, as well as due to presence of pores or fat globules, that are different
structural elements. Furthermore, water relaxation is affected by the cheese texture, which
can be related to interactions between water and macromolecules. Therefore, not only so
translation and rotation of water molecules surely affect relaxation, but also diffusive and
chemical exchange processes between water molecules and biopolymers strongly
contribute to T2 and T1 values.
In cheese, the change of dynamic NMR parameters (T1, T2 and self-diffusion
coefficient), as well as the FID amplitude and intensity, can also be correlated with the
system evolution due to a ripening process, as in the case of Grana Padano cheese (De
Angelis Curtis et al., 2000), or an aging process in the shelf-life, as is the case of Mozzarella
di Bufala Campana cheese (Gianferri et al., 2007b).
Moreover, T2 and T1 value changes can be associated with system modification due to
temperature processing or water sorption as during the storage (Castell-Palou et al., 2011).
shrinkage and lets out water molecules (entrapped water) that, in turn, becomes
“interstitial” water (junction zones water) (De Angelis Curtis et al., 2000).
Bordini et al. have explored two different kinds of Parmigiano Reggiano (aged 15 and
30 months) at different stages of protein hydrolysis process by TD- and HR-NMR (Bordini
et al., 2011). TD-NMR has pointed out that cheeses with different aging times, although
starting from distinct initial compositions, conclude digestion in a similar way, in terms of
free amino acids and small organic compounds, but evolve with different kinetics of
hydrolysis and peptide formation, discriminating the young from the old cheese.
In both studies, the TD NMR spectroscopy has been employed to provide information
about water distribution in hard cheeses and ripening processes. Even if it has to be
emphasized that the ripening process implies many modifications of high complexity, these
approaches provide a way to verify the effect of the ripening process on foodstuffs.
Analysis and quality control of food is an important application area for low field time
domain NMR (McCarthy et al., 2006).
This is primarily due to their relatively low cost, ease of operation, ability to provide
information on product composition within short turnaround times. However, as samples
need to be withdrawn, packaged food, intact plants, and process control are largely
excluded from such an analysis unless strategies related to imaging are employed. Progress
in this direction has been made by development of unilateral NMR sensors, devices that
are able to perform non invasive analysis of foods without needing to cut the sample into
pieces that must fit in the NMR tube.
Interesting applications of portable single sided NMR sensors to food matrices have
been reported in literature (Mitchell et al., 2014; Blümich et al., 2008).
One of the first applications of single sided NMR devices in food was the quantitative
study of the oil and water content in food.
Pedersen et al. investigated the performance of conventional benchtop NMR and
single-sided NMR applied to the study of oil-in-water emulsions (Pedersen et al., 2003).
The study demonstrated how CPMG-like pulse sequence can be used on the NMR MOUSE
to obtain quantitative measurements on model food systems. The results were compared to
those obtained using a conventional low field NMR instrument. In a homogeneous field
the decay rate of pure water obtained after applying a CPMG sequence is much longer than
that of oil, which means that the relaxation behavior of the two components is very
different. In a single side NMR sensor, the presence of a strong magnetic field gradient
heavily modifies the decay of magnetization; in fact, the decay rate of water is much faster
than that of oil because water is much more affected by diffusion than oil. This diffusion
weight makes the single side sensor a suitable instrument for quantifying the oil/water
ratios. The trend in relaxation behavior with increasing oil content measured by single-
sided NMR was found to be the reverse of that obtained by homogeneous benchtop NMR.
The reverse of the trend in the decay rates with decreasing the oil content is also foreseeable
as a result of the strong magnetic field gradient which affects the apparent transverse
relaxation time of water due to its rapid diffusion rate. Authors demonstrated that NMR
decays obtained by both homogenous benchtop NMR and single-sided NMR can be readily
deconvoluted in two components in the case of oil-in-water emulsions with oil content
ranging from 10 to 67%.
Furthermore, single-sided NMR was used to obtain compositional information in a
through-package manner, so the product could be analyzed in sealed conditions. For
example, Guthausen et al. investigated the applicability of a single-sided NMR sensor to
measure the fat content in packaged food products (Guthausen et al., 2004). Fat content is
one of the important parameters of quality control in many food products. In this work two
different low-field NMR methods, namely, a ratio method and a relaxation time method,
were applied and discussed. The processed NMR signal was linearly correlated with the
fat content obtained by reference methods. The linear correlation allowed the application
of single-sided NMR for fat measurements. At this aim two methods were applied
(Guthausen et al., 2006). One is based on differences in relaxation times which depend on
molecular mobility and, consequently on the molecular structure (relaxation time method),
whereas the other one exploits the fact that the diffusion coefficients of water and fat differ
by more than an order of magnitude (diffusion weighted method). A pulse sequence was
purposely developed for measuring fat and water content in packaged coffee creams and
packaged mayonnaises and margarines. The processed NMR signal was linearly correlated
with the fat content obtained by reference methods.
In this case the ratio of the final echoes and the first few echoes were the relevant NMR
parameter, named NMR ratio. The NMR ratio reported as a function of the reference fat
content exhibits a linear dependence fit by a linear regression with a correlation coefficient
of 0.996 indicating that a reasonable determination of the fat content is possible.
The relaxation method was used for calibrating the fat content in packaged
mayonnaises and margarines. In fact, the relaxation time differences of fat and water
protons may be exploited to generate a contrast allowing the relative content determination.
Again, a linear correlation was found between the fat percentage measured by NMR and
the reference fat content. In this case data were analyzed by a chemometric approach
obtaining a correlation coefficient of 0.991. The linear correlation found with both methods
between the processed NMR signal and the fat content obtained by reference methods may
allow the use of single-sided NMR for fat measurements even in packaged food. The
chemometric approach was applied to the NMR signal which can be analyzed using
supervised or unsupervised pattern recognition methods (Vandenginste et al., 1998). These
methods applied to food analysis may be qualitative or quantitative. The knowledge of
reference values allows for the calibration of the NMR response (Guthausen, 2016).
In another paper Martini et al. compared the determination of solid fat content (SFC)
obtained using conventional off-line NMR instrument with determination obtained by
single-sided NMR sensor (Martini et al., 2005). Authors explored the use of single sided
NMR to measure variation of SFC on-line during the crystallization of a fat product.
Interesterified hydrogenated palm oil was added to canola oil in different proportions to
obtain blends with different SFC values. Two different experiments were carried out to
study whether the motion during the crystallization affected the measurements. In one
experiment, agitation was stopped during the measurement and then restarted to let the
crystallization continue until the next measurement. In the other experiment measurements
were carried out under agitation. To eliminate the temperature effect on measurements
carried out by single-sided sensor a proper correction was applied to the collected data. In
the former case the corrected data were close to those collected by a conventional off-line
NMR instrument particularly at intermediate SFC values, whereas in the latter case data
were found to be significantly different from those determined by the off-line instrument.
Haiduc et al. reported a feasibility study on the use of single-sided NMR for the
assessment of the microstructural quality of food material (Haiduc et al., 2007). Authors
investigated model systems consisting of oil-in-water emulsion gels stabilized by proteins.
These systems form complex structures made of fat droplets and a protein aggregate
network where water is dispersed in pores with different size. An important quality
parameter of such systems is the water exudation (WE). In a classical approach WE is
determined by measuring the amount of water lost from the system when subject to
gravitational forces. To establish a relation between signals obtained by single-sided NMR
and WE multivariate calibration techniques were applied. Specifically, to obtain a
calibration model allowing for a physical interpretation, an approach based on multilinear
regression (MLR) was applied to decays collected with a benchtop instrument and with a
single-sided sensor as well. Decays were transformed from the continuous domain to the
discrete domain of T2 distributions and amplitudes using the Nonnegative Least Squares
(NNLS) algorithm. The decays were averaged and the obtained data fitted without any
initial guess on the number of components or T2 values. To summarize, the MLR model
was built using the NNLS amplitudes as responses, and the functional parameters as
predictors. The quality of MLR model applied to both set of data is comparable indicating
that single-sided NMR may also allow for the assessment of WE in a through package
mode.
Single side NMR was also employed in quality control of sealed liquid foods. Stork et
al. applied single-sided and semisingle-sided NMR sensors with a reduced magnetic field
gradient for determining the oxygen content in unopened bottles with superoxygenated
table water, and compared the results obtained with results obtained by conventional NMR
(Stork et al., 2006). A good comparison between the spin-lattice relaxation rate measured
with a single-sided NMR sensor and that measured by conventional NMR was found. The
semisingle-sided sensor was also used for determining the oxygen concentration in a
commercial bottle before and after opening the bottle by measuring the water relaxation
time. The sensor was able of monitoring the oxygen concentration that was constant before
opening, and progressively decreased after opening the bottle, indicating the possible use
of this sensor for on-line application after a suitable calibration procedure.
Another paper reports a method based on the use the profile NMR MOUSE to detect
adulteration of virgin olive oil through sealed bottles (Xu et al., 2014). With this method
the transverse relaxation time and the self-diffusion coefficient of virgin olive oils
adulterated with different percentages of sunflower oil and red palm oil, were investigated.
In this paper authors developed a 2-dimensional Laplace inversion according to the
algorithm by Venkataramanan to reconstruct the 2-dimensional probability density
distribution of the transverse relaxation time and self-diffusion coefficient
(Venkataramanan et al., 2002). Figure 4 shows the D-T2 distribution map of different olive
oils which are well separated in the self-diffusion coefficient (D) direction whereas overlap
in the transverse relaxation time (T2) direction.
Figure 4. D-T2 distribution maps of oils. From the top to the bottom a) pure extra virgin olive oil, extra
virgin olive oil adulterated with 10 and 20% of sunflower oil, pure sunflower oil. b) Pure red palm oil,
extra virgin olive oil adulterated with 10 and 20% of red palm oil, and pure extra virgin olive oil
(reproduced with permission from Xu et al., 2014).
From top to bottom peaks correspond to pure extra-virgin olive oil, extra virgin olive
oil adulterated with 10 and 20% of sunflower oil, and pure sunflower oil. Figure 4b shows
the D-T2 distribution of red palm and extra virgin olive oil mixtures. From the top to the
bottom peaks correspond to pure red palm oil, extra virgin olive oil adulterated with 10 and
20% of red palm oil, and extra virgin olive oil. Oils are well separated in D direction and,
at a lesser extent, in T2 direction. These data indicate that the adulteration of extra virgin
olive oils with sunflower oil may be readily detected from the self-diffusion coefficient
behavior, whereas the adulteration with red palm oil can be detected from both diffusion
and transverse relaxation behavior.
The tomato paste processing industry is very interested in developing methods to detect
tomato spoilage in 1,000 L non-ferrous, metal-lined containers without violating the seal.
Early spoilage detection would eliminate shipping costs and disposal costs when spoiled
tomato paste arrives at destination. Pinter et al. explored the relaxation properties of sterile
and unsterile tomato paste (Pinter et al., 2014). Authors found that spoilage in tomato paste
test samples leads to longer longitudinal relaxation times using a conventional benchtop
NMR system. Specifically, the steady state spin lattice relaxation time T1 obtained by the
regression of T1 values measured as a function of time, was the parameter chosen to
differentiate between spoiled and unspoiled tomato paste. This result prompted them to
extend the study to 1,000 L non-ferrous, metal-lined totes using single-sided NMR. In order
to perform measurements directly on the metal container its effect on the circuit tuning and
impedance matching properties was properly compensated. A modified saturation recovery
sequence was used to measure the longitudinal relaxation with single-sided NMR. The
NMR signal and T1 values obtained from the large format container with the single-sided
sensor suggested that this device can be used to study tomato paste spoilage in factory
process environments.
Unilateral NMR sensor was also able to asses in vivo the fat content in fish,
demonstrating the feasibility for online quality control. Veliyulin et al. developed and
tested a new method for the rapid measurement of the fat content in live (or slaughtered)
Atlantic salmon, based on a mobile low-field NMR analyzer, has been developed and tested
(Veliyulin et al., 2005). The instrument, calibrated against a set of reference samples (fish
oil in agarose), was used for non-destructive fat determination of the Norwegian quality
cut of anaesthetized fish. The distribution of transverse relaxation times when measured
with a conventional benchtop NMR instrument shows three peaks. The shortest component
is usually ascribed to water closely associated with macromolecules, the intermediate one
to intracellular water or water within the myofibrillar structure, whereas the longest one
accounts for both lipids and water. The interference of the fat signal with the signal from
extra- myofibrillar water both contributing to the longest component makes a direct
quantification of the fat or water content from T2 measurements very difficult. When
measurements are performed by a single-sided NMR sensor, the strong field gradient
generated by the sensor makes the self-diffusion coefficient contribute strongly to the
effective transverse relaxation time. Because the average self-diffusion coefficient of water
in fish muscle is considerably faster than that of fat, this difference may be exploited to
make accurate quantification of the fat component from the transverse relaxation response.
Figure 5 compares the transverse relaxation curves of pure fish oil and Atlantic salmon
white muscle obtained applying a CPMG sequence optimized to ensure the best separation
between water and fat components. A significant correlation was found between the fat
content measured by single-sided NMR and that measured by chemical extraction data
obtained after slaughtering the same fish. Furthermore, the mobile single-sided sensor was
used to map the fat distribution over the whole fish surface. The rapid fat determination
with easy calibration routines showed that single side NMR has potential for
implementation in connection with on-line quality control for in vivo assessment of fat
content in salmon.
Figure 5. Transverse relaxation curves of pure fish oil and Atlantic salmon white muscle measured at 4°C
using a single-sided NMR sensor (reproduced with permission from Veliyulin, 2005).
Relaxometry using a single side sensor was employed to investigate the water status
and ripening of fruits. Capitani monitored the water status of kiwifruits as a function of
season using a single-sided NMR sensor (Capitani et al., 2010; Capitani et al., 2013). Using
this sensor it was possible to measure the entire fruit at a depth of about 0.5 cm from the
peel surface without cutting it (Figure 6a). The T2 distribution of a mature kiwifruit
measured by single-sided NMR shows three peaks (Figure 6b). According to the literature
the longest T2 component was assigned to protons in vacuole, the intermediate one to
proton in cytoplasm and extra-cellular space, and the shortest one to proton in cell walls. It
must be noted that the presence of the strong magnetic field gradient of the single-sided
sensor, heavily shortens T2 values. In fact, in the case of ripened kiwifruit measured in
homogeneous fields literature data report average T2 values of 800-1000 ms for the longest
component, 200-400 ms for the intermediate component, and 20-80 ms for the shortest one.
With single-sided NMR the longest component is as short as 20 ms. Nevertheless,
single-sided NMR was suitable to monitor the growth of kiwifruit. Because the shortest
component was very poorly affected by the season, only the intermediate and longest
components were taken into account. Figure 7 reports the average values of the
T T
intermediate component ( 2a ) and the longest component ( 2b ) measured on nine
kiwifruits of three cultivar, namely Hayward (a, b), CI.GI. (c, d), and Zespri (e, f) at
different stages of development. In all cultivar T2 values were found to be rather constant
until October, thereafter they increased. The tendency toward longer T 2 relaxation times
later in the season is consistent with a change in the fruit texture occurring during fruit
development. However, whereas in Hayward and CI.GI. a gradual lengthening of T 2 was
observed, in Zespri, a net and sharp lengthening of both components occurred between
October and November, indicating the earlier maturation of Zespri.
Figure 6. a) Measurement of intact kiwifruit with a single-sided NMR sensor. b) Transverse relaxation
times distribution measured on a ripened kiwifruit.
T T
Figure 7. Average T2 values, namely 2a and 2b measured on nine kiwifruits versus the developmental
stage of Hayward (a, b), CI.GI. (c, d), and Zespri (g, h) kiwifruits (Reproduced with permission from
Capitani et al. 2013).
The ripening of kiwifruits was also monitored in field on fruits attached to the plant
during three campaigns of measurement carried out in October, November and December.
Figure 8 shows the transverse relaxation decays measured in field. In all cultivar a
lengthening of the decays with the season was observed, however in Zespri (Figure 8c) the
process was complete already in November, in fact decays measured in November and
December perfectly overlap. In agreement with data collected in laboratory on intact
kiwifruits, these data confirmed the earlier maturation of Zespri.
Figure 8. Transverse relaxation curves measured in field on Hayward (a), CI.GI (b), and Zespri (c) with
the corresponding relaxation times distributions.
The profile NMR MOUSE was used to get preliminary results on three fresh intact
blueberries and the same blueberries let to wither outside the fridge for three and six days
(Capitani et al., 2014) (Figure 9).
With this sensor NMR depth profiles were collected, these profiles encode the
amplitude of the 1H NMR signal as a function of the depth scanned. The amplitude of
profiles measured on withered blueberries was found to be lower than that measured on
fresh blueberries, indicating a loss of water that was quantified by integrating the profiles.
After three days of withering a loss of water of 13% (a), 11% (b), and 16% (c) was
measured, whereas after six days the loss of water was found to be 30% (a), 20% (b), and
34% (c). Therefore, the integral of the profile is a suitable index of water loss. The same
index may be used to monitor changes in foodstuff texture due to maturation, ripening,
water and osmotic stress, for monitoring the effect of different types of processing on food
matrices, and the effect of storage.
Figure 9. Depth profiles of three fresh blueberries (a, b, and c), and profiles of the same blueberries let to
wither for three and six days (Reproduced with permission from Capitani et al., 2014).
Adiletta et al. used the profile NMR MOUSE to investigate the drying process on pears
(Adiletta et al., 2015). Drying is an important process for conservation and marketing of
fruits due to their high water activity which makes them perishable. Therefore, the
knowledge and optimization of drying process are very important to minimize thermal
damage and quality loss. Information on drying kinetics was obtained by measuring the
intensity of 1H NMR signal as a function of the thickness of the sample. Figure 10a shows
the comparison among profiles of fresh pear and pears dried for 3, 6, 15, 20, 29, and 48
hours. The amplitude of profiles measured in dried samples progressively lowered and also
the thickness of the profiles progressively reduced indicating a loss of water with increasing
the drying time and the consequent shrinkage. Figure 10b shows the good agreement
between values (Mt/M0), obtained by gravimetric method and values (It/I0) obtained by
integrating NMR depth profiles, with a regression coefficient of 0.978. An evaluation of
the loss of water in the outer, intermediate, and central regions of pears samples was
obtained simply by integrating the profiles in three corresponding regions (without cutting
the samples); the results obtained are reported in Figure 10c. As expected the water content
of the outer region decreases more quickly than that of the intermediate and central ones,
such differences are much reduced at long drying times.
Figure 10. a) NMR depth profiles of fresh pear and pear samples dried at 50 °C for 3, 6, 15, 20, 29, and
48 h. b) Relationship between the water loss measured by gravimetric method (M t/M0) and single-sided
NMR (It/I0). c) Water loss in exterior, intermediate, and central part of pear samples measured by single-
sided NMR. (Reproduced with permission from Adiletta et al., 2015).
We hereby reported a short survey of the most relevant works related to the application
of Magnetic Resonance Imaging (MRI) in food chemistry. De facto, a relatively large body
of literature has highlighted the determinant role played by MRI, through different MRI
experiments and applications, as a rapid, quantitative, and non-invasive NMR technique to
evaluate the quality of fresh, stored and processed food products. Although the image
resolution is restricted to few micrometers and the costs of MRI analysis are still rather
expensive, this technique offers the unique advantage to investigate the inner morphology
of food products, thereby progressively enlarging our knowledge on the number of factors
that affect their quality, authenticity, shelf-life, perishability and safety for human health.
Numerous works have so far demonstrated how MRI high-resolution images may serve
as a reliable tool to appreciate critical morphological changes in edible fruits and
vegetables. Sequi et al. demonstrated that MRI spectroscopy is capable to differentiate
cherry tomatoes grown in Protected Geographical Indication (PGI) area from those grown
in non-PGI area on the basis of morphological and physical parameters (Sequi et al., 2007).
In particular, the proposed approach consisted of a set of four empirical equations taking
into account the pericarp thickness, the width of the inner and outer spherical crowns
composing the pericarp itself and their transverse relaxation times T 2. Later, the same
authors revealed changes in both morphological structure and relaxation times of PGI
Pachino cherry tomatoes depending on seasonal growing conditions (Ciampa et al., 2010).
Remarkably, T1-weighted images revealed a white film covering the placental cavities
exclusively for tomatoes harvested in winter and spring seasons (Figure 11). Moreover, the
T1 values associated to the site interested by the film resulted to be relatively low thus
indicating intense interactions with the cellular tissues (Ciampa et al., 2010).
Figure 11. T1-weighted image of Pachino cherry tomatoes (Lycopersicon esculentum cv. Shiren)
harvested during the winter (A) and summer (B) seasons. Reprinted from Journal of Food Chemistry with
permission by Elsevier (Ciampa et al., 2010).
MRI also allowed to quantify the microstructure of apple cortex that is involved in
mechanical and transport properties of fruit tissues (Musse et al., 2010). In order to confirm
the reliability of these results, the authors acquired images on two magnets at different
strength and performed a comparison with X-ray microtomography technique. Spin echo
MRI images enabled to asses the variation in water distribution in pea seeds, harvested at
different seed stages, and indicated a gradual dislocation of water from the inner to the
outer part of cotyledons, likely orientated in the same direction as starch accumulation
(Garnczarska et al., 2008). Thybo et al. demonstrated that, even though MRI was able to
efficiently differentiate several potato varieties, the histogram image analysis did not show
a robust correlation with dry matter content of tubers (Thybo et al., 2003).
The MRI technique may also single out important details concerning the vascular
system of plants. A valid example is represented by a work of Salerno et al. in which the
reported clear images of internal morphology of radishes showed the radial distribution of
xylematic and phloematic vessels (Salerno et al., 2005). Interestingly, Marconi et al. have
observed several changes occurring in the internal structure of radish tuber grown in two
different types of soils (sandy and clay-loam) and irrigated with water contaminated by
different concentrations of Arsenic (V) (Marconi et al., 2010). It was observed that the As-
uptake induced the formation of large cavities, with the detachment of the external cortex,
and that the effect was even enhanced in case of radishes grown in clay soils (Marconi et
al., 2010). The importance of this issue relies on the fact that the uptake of As (V) in edible
vegetables, even in small amounts, constitutes a primary risk for food safety and human
health.
The use of phytoregulators to increase the yields in agriculture represents another issue
of food safety concern. In fact, since these agrochemicals are prohibited in both biological
and integrated agriculture, it is necessary to apply analytical tools capable to identify and
discourage their possible use. MRI has been applied to trace the phytoregulators addition
in kiwi fruit cultivation. Since hormones-related metabolic residues are no longer present
at harvest time in fruits characterized by long maturation period, such as kiwi, their
identification must be obtained during the growth stage. This was achieved on hormone-
treated kiwis by examining the internal morphology and, in particular, the qualitative
changes occurring in the epicarp and mesocarp (Valentini et al., 2009).
The MR structural images have been also used for non-invasive investigations on other
common problems of agrofood products, such as the malformations growth, the effects of
parasites, the early detection of mechanical damages and physiological diseases. For
example, the watercore is a physiological disorder in apples which occurs mostly late in
the season in over-mature fruits when they are still pending on the tree. Watercore-affected
apples are not easily recognized because only their internal tissues are interested while the
outer part of the fruit appears intact. Wang et al. and Clark et al. have applied MRI to study
the watercore occurrence in Red Delicious and Fuji apples apple varieties, respectively
(Wang et al., 1988; Clark et al., 1998b). Both works showed that the disorder implies a
So far, a discrete body of literature has proved that MRI not only represents an
excellent technique to supply a detailed (up to few tens of micrometers) visualization of
internal structure of intact food products, but can also provide indisputable elements to
both assess the quality of food and reveal the influence exerted by important factors such
as the storage procedures, the food industrial processing or the temperature treatments. In
this context, the most relevant applications of MRI will be cited in the following paragraphs
and grouped as a function of applied procedure or treatment.
The application of uncontrolled and undesired physical pressures may occur during
transport and storage of fruit and vegetables with negative influence on the quality. Otero
(Otero & Prestamo, 2009) investigated the effects exerted by the application of controlled
pressures on strawberries and evaluated the extent of physical tissue damage on the basis
of T1-, T2- and diffusion-weighted MRI images.
Several works have used MRI also to study processes occurring during food packaging
and the effectiveness of treatments aimed to preserve food quality. For example, MRI
permitted to assess the capacity of a chitosan-based coating product to slow down the
maturity and decay in two varieties of citrus fruits and, at the same time, their preservation
by inhibiting the fungal growth (Galed et al., 2004). The authors showed that the fungal
formation may be revealed by T2-weighted images as an accumulation of mobile water in
the floral meristem. This is accompanied by a dehydrated area in the epicarp that
corresponds to the tissue portion where the fungi are growing (Galed et al., 2004). MRI
also allowed to monitor the process of moisture redistribution occurring in commercial
sandwiches containing sausages or ham and stored in sealed plastic packages under N2
atmosphere (Ramos-Cabrer et al., 2006). Interestingly, these authors have successfully
used a variant of the Single Point Imaging (SPI) pulse sequence (Kennedy et al., 1998)
with the aim to circumvent the limitation of extremely restricted mobility of water fraction
in sandwiches.
3.3. Freezing-Thawing
A valid strategy to preserve, as much as possible, the quality of food products consists
in the application of freezing, followed by a low-temperature storage (generally around -
20°C, for a maximum period that depends on food properties), and finally by thawing prior
to product consumption. However, extreme or inappropriate freeze-thawing conditions
may significantly influence texture, firmness, water distribution and organoleptic
properties of food and the MRI technique appears especially suitable to investigate these
parameters. In fact, it has been proved that T2–weighting is one of the best MRI tools to
identify freezing injuries in food, as shown by several studies which examined frozen-
thawed products such as courgettes (Duce et al., 1992), blueberries (Gamble, 1994),
kiwifruits (Kerr et al., 1997) and oranges (Hernandez-Sanches et al., 2004). The T2
increase, which commonly results from the freezing procedure, may be ascribed to both
the cell lysis due to ice crystals formation as well as to protein denaturation that affects the
overall food structure (Erikson et al., 2012). Moreover, differently from T 1 values, the T2
values are very sensitive in identifying freeze-damaged cucumbers by showing higher
water mobility localized in damaged tissues (Kotwaliwale et al., 2012). An interesting
study reported by Koizumi et al. showed that spin-echo-based MRI images of a specific
MRI system, enabled to follow the changes during the thawing process of frozen vegetables
such as green soybeans, broad beans, okra, asparagus and taro (Koizumi et al., 2006).
Moreover, T1 values and magnetization transfer rate of water were shown to differentiate
fresh meat samples belonging to several species (lamb, beef and pork) from those deriving
from a freezing–thawing process (Evans et al., 1998). The application of MRI was explored
by Nott (Nott et al., 1999) as a potential tool for the authentication of fresh rainbow trout
and its distinction from the same fish after a freezing–thawing process, while Foucat
(Foucat et al., 2001) monitored the rainbow trout quality by MRI as a function of the frozen
storage period. Measurements by MRI of water content and distribution, as well as
relaxation times, contributed to confirm that ultrasonic treatment is a promising way to
decrease the freezing time of frozen radish samples and better preserve their quality (Xu et
al., 2015). Differences in MRI diffusion coefficients in mozzarella cheeses ("pasta-filata"
and "pasta non-filata") were significantly related to the days elapsed between product
manufacture and freezing, to the length of frozen storage and to the temper period at 5 °C
(Kuo et al., 2003).
Another strategy to store and preserve food products such as fish flesh and meat
consists in the salting procedure. In this context, a noteworthy example is represented by
the work of Aursand (Aursand et al., 2010), in which, through an SPI MRI pulse sequence,
the effects of the salting of Atlantic salmon have been investigated by means of 23Na MRI.
The authors demonstrated that the combination of 1H and 23Na MR images enabled the
quantification of the uptaken sodium and revealed the tight correlation between salmon fat
distribution and salt diffusion/distribution. Recently, MRI has been combined with data
mining to study salt-diffusion during the post-salting stage of Iberian hams and to predict
salt content (Caballero et al., 2016).
Temperature plays a determinant role not only for food preservation, but also in many
processing techniques aiming to improve products quality and safety. Temperature
treatments may be in fact involved in both phases of production and food transformation.
Since these treatments are also applied to dairy products, several interesting MRI studies
have been developed to investigate their effects (Anedda, 2015; Mulas et al., 2013). In
particular, Mulas described a new protocol to discriminate Sardinian sheep milk cheese
originated from either heat-treated or raw milk. The proposed MRI procedure enabled
samples differentiation through proton T2 distribution. A significantly larger abundance of
water protons population strongly interacting with proteins accompanied by a smaller long
T2 population, has been detected in heat-treated milk cheeses as compared with raw milk
counterparts (Mulas et al., 2013). The MRI technique was also employed to examine cereal
products subjected to different thermal processes, such as steeping, at low and high
temperatures, and drying processes. For example, Ruan (Ruan et al., 1992) studied the
pericarp quality of corn kernels as a function of water absorption upon temperature steeping
and drying processes. The authors stated that changes in MRI signal intensity were due to
a lower proton density resulting from a moisture loss. Interestingly, their results proved
how MRI may potentially be exploited for process and quality control, by monitoring
moisture transfer and related structural changes during drying processes.
The cooking process represents a further food transformation of primary importance
when one wants to control sensory, nutritional and technological qualities of cooking end-
products. Bouhrara (Bouhrara et al., 2011) acquired MR images of meat samples during
cooking and revealed that the structural changes occurring in meat were mostly caused by
modifications of both cell and connective-tissue proteins. It has been also demonstrated
that both T1 and T2 values progressively decreased with increased heating time in cooked
meat (Shaarani et al., 2006). They showed a loss of water associated with a decreased
rotational mobility, and spatial differences between the core and outer layers of samples.
The T1 and T2 relaxation times are expected to decrease when the denaturation of structural
proteins begins (myosin), thus inducing the release of water from cells (myofibrils) into
the extracellular space. The expulsion of water is a consequence of the contraction of
connective tissue which expels water first into the inter bundle space and then out of meat
(Bouhrara et al., 2011). Remarkably, MRI helped in identifying the heating pattern
developed in a sauce that contained meat pieces and was processed by either microwave or
conduction-limited conventional heating (Bows et al., 2001). In this work, MRI enabled
the mapping of the spatial distribution of temperature and demonstrated that a different
heating pattern occurred as a function of the heating procedure (Bows et al., 2001).
CONCLUSION
ACKNOWLEDGMENTS
This work has been carried out within the Italian project of Regione Lazio Lr 13/2008
entitled “e-ALIERB: un OPEN LAB per caratterizzare e valorizzare i prodotti alimentari
ed erboristici del territorio laziale.”
REFERENCES
Adiletta, G., Russo, P., Proietti, N., Capitani, D., Mannina, L., Crescitelli, A., & Di Matteo,
M. (2015). Characterization of pears during drying by conventional technique and
portable non invasive NMR. Chemical Engineering Transactions, 44, 151-156.
Anedda, R. (2015). Magnetic Resonance analysis of dairy processing: suitable tools for the
dairy industry. In F. Capozzi, L. Laghi, & P. S. Belton (Eds.), Magnetic Resonance in
Food Science: Defining Food by Magnetic Resonance. (First edition pp. 51-64).
Cambridge, UK: Royal Society of Chemistry.
Aursand, I. G., Erikson, U., & Veliyulin, E. (2010). Water properties and salt uptake in
Atlantic salmon fillets as affected by ante-mortem stress, rigor mortis, and brine
salting: A low-field 1H NMR and 1H/23Na MRI study. Food Chemistry, 120, 482-489.
Barreiro, P., Ruiz-Cabello, J., Fernandez-Valle, M. E., Ortiz, C., & Ruiz-Altisent, M.
(1999). Mealiness assessment in apples using MRI techniques. Magnetic Resonance
Imaging, 17, 275-281.
Barreiro, P., Ortiz, C., Ruiz-Altisent, M., Ruiz-Cabello, J., Fernandez-Valle M. E.,
Recasens, I., Asensio M. (2000). Mealiness assessment in apples and peaches using
MRI techniques. Magnetic Resonance Imaging, 18(9), 1175-1181.
Barreiro, P., Moya, A., Correa, E., Ruiz-Altisent, M., Fernandez-Valle, M., Peirs, A.,
Wright K. M., & Hills, B. P. (2002). Prospects for the rapid detection of mealiness in
apples by nondestructive NMR relaxometry. Applied Magnetic Resonance, 22, 387-
400.
Belton, P. S. (1984). Nuclear magnetic resonance and photoacoustic spectroscopy. In H.
W.-S. Chan, (Ed.), Biophysical Methods in Food Research (pp. 103-37). London, UK:
Blackwell Scientific.
Blümich, B., Perlo, J., & Casanova, F. (2008). Mobile Single-side NMR. Progress in
Nuclear Magnetic Resonance Spectroscopy, 52, 197-269.
Bordini, A., Picone, G., Babini, E., Vignali, M., Danesi, F., Valli, V., Di Nunzio, M., Laghi,
L., & Capozzi F. (2011). NMR comparison of in vitro digestion of Parmigiano
Reggiano cheese aged 15 and 30 months. Magnetic Resonance in Chemistry, 49, S61-
S70.
Borgia, G. C., Brown, R. J. S., & Fantazzini, P. (1998). Uniform-penalty inversion of
multiexponential decay data. Journal of Magnetic Resonance, 132(1), 65-77.
Bouhrara, M., Clerjon, S., Damez J. L., Chevarin, C., Portanguen, S., Kondjoyan, A., &
Bonny, J. M. (2011). Dynamic MRI and thermal simulation to interpret deformation
and water transfer in meat during heating. Journal of Agricultural and food Chemistry,
59, 1229-1235.
Bowen, J. H., & Watkins, C. B. (1997). Fruit maturity, carbohydrate and mineral
contentrelationships with watercore in ‘Fuji’ apples. Postharvest Biology and
Technology, 11, 31-38.
Bows, J. R., Patrck, M. L., Nott, K. P., & Hall, L. D. (2001). Three-dimensional MRI
mapping of minimum temperature achieved in microwave and conventional food
processing. International Journal of Food Science & Technology, 36, 243-252.
Brosio, E., Belotti, M., & Gianferri, R. (2008). Diffusive and chemical exchange model:
practical hits to investigate water dynamics in foods. In L. V. Greco, & M. N. Bruno
(Eds.), Food Science and Technology (First Edition pp. 323-340). Hauppauge (NY):
Nova Science Publishers.
Brosio, E., & Gianferri, R. (2011). NMR spectroscopy in dairy products characterization.
In R. D. Foster (Ed.), Cheese: Types, Nutrition and Consumption (First edition pp. 183-
206). Hauppauge (NY): Nova Science Publishers.
Caballero, D., Caro, A., Rodríguez, P. G., Duran, M. L., Del Mar Avila, D., Palacios, R.,
Antequera, T., & Perez-Palacios, T. (2016). Modeling salt diffusion in Iberian ham by
applying MRI and data Mining. Journal of Food Engineering, 189, 115-122.
Callaghan, P. T., Jolley, K. W., & Lelievr, J. (1979). Diffusion of water in the endosperm
tissue of wheat grains as studied by pulsed field gradient nuclear magnetic resonance.
Biophysical Journal, 28, 133-141.
Capitani, D., Mannina, L., Proietti, N., Sobolev, A. P., Tomassini, A., Miccheli, A., Di
Cocco, M. E., Capuani, G., De Salvador, R., & Delfini, M. (2010). Monitoring of
Colnago, L. A., Moraes, T. B., Monaretto, T., & Andrade, F. D. (2015). Rapid
determination of food quality using steady state free precession sequences in TD-MNR
Spectroscopy. In Magnetic Resonance in Food Science (pp. 1-16).
De Angelis Curtis, S., Curini R., Delfini, M., Brosio, E., D’Ascenzo, F., & Bocca, B.
(2000). Amino acid profile in the ripening of Grana Padano cheese: A NMR study.
Food Chemistry, 71, 495-502.
Defraeye, T., Lehmann, V., Gross, D., Holat, C., Herremans, H., Verboven, P., Verlinden,
B. E., & Nicolai, B. M. (2013). Application of MRI for tissue characterisation of
‘Braeburn’ apple. Postharvest Biology and Technology, 75, 96-105.
De Rossi, A., De Pilli, T., Severini, C., & McCarthy, M. J. (2008). Mass transfer during
osmotic dehydration of apples. Journal of Food Engineering, 86, 519-528.
Duce, S. L., Carpenter, T. A., Hall, L. D. (1992). Nuclear Magnetic Resonance Imaging of
fresh and frozen courgettes. Journal of Food Engineering, 16, 165-172.
Duckworth, R. B. (1975). Water Relations in Foods. London, UK: Academic Press.
Elgar, H. J., Burmeister, D. M., & Watkins, C. B. (1998). Storage and handling effects on
a CO2- related internal browning disorder of ‘Braeburn’ apples. Horticultural Science,
33, 719-722.
Engelsen, S. B., Jensen, M. K., Pedersen, H. T., Nørgaard, L., & Munck, L. (2001). NMR-
baking and multivariate prediction of instrumental texture parameters in bread. Journal
of Cereal Science, 33(1), 59-69.
Erikson, U., Standal, I. B., Aursand, I. G., Veliyulin, E., & Aursand, M. (2012). Use of
NMR in fish processing optimization: a review of recent progress. Magnetic
Resonance in Chemistry, 50, 471-480.
Evans, S. D., Nott, K. P., Kshirsagar, A. A., & Hall, L. D. (1998). The effect of freezing
and thawing on the magnetic resonance imaging parameters of water in beef, lamb and
pork meat. International Journal of Food Science & Technology, 33, 317-328.
Foucat, L., Taylor, R. G., Labas, R., & Renou, J. P. (2001). Characterization of frozen fish
by NMR imaging and histology. American laborator, 22, 38-43.
Galed, G., Fernandez-Valle, M. E., Martınez, A., & Heras, A. (2004). Application of MRI
to monitor the process of ripening and decay in citrus treated with chitosan solutions.
Magnetic Resonance Imaging, 22, 127-137.
Gamble, G. R. (1994). Non-invasive determination of freezing effects in blueberry fruit
tissue by magnetic-resonance-imaging. Journal of Food Science, 59, 571-573.
Gao, Y., Xiao, L., Zhang, Y., & Xie, Q. (2016). The generalized Phillips-Twomey method
for NMR relaxation time inversion. Journal of Magnetic Resonance, 271, 1-6.
Garnczarska, M., Zalewski, T., & Wojtyla, L. (2008). A comparative study of water
distribution and dehydrin protein localization in maturing pea seeds. Journal of Plant
Physiology, 165, 1940-1946.
Gianferri, R., Maioli, M., Delfini, M., & Brosio, E. (2007a). A low-resolution and high-
resolution nuclear magnetic resonance integrated approach to investigate the physical
Hills, B. P., & Clark, J. (2003). Quality assessment of horticultural products by NMR.
Annual Reports on NMR Spectroscopy, 50, 75-120.
Hills, B. P. (2006). Application of low field NMR to food science. Annual Reports on NMR
Spectroscopy, 58, 177-230.
Hills, B. P. (2008). NMR relaxation and diffusion studies of horticultural products. In G.
A. Webb (Ed.), Modern Magnetic Resonance: Part 1 (First edition, pp. 1721-1727).
Dordrecht, Netherlands: Springer.
Hinrichs R., Gotz J., Noll M., Wolfschoon A., Eibel H., & Weisser H. (2004).
Characterisation of the water-holding capacity of fresh cheese samples by means of
low resolution nuclear magnetic resonance. Food Research International, 37, 667-676.
Jung, K. H., Stroshine, R., Cornillon, P., & Hirst, P. M. (1998). Low field proton magnetic
resonance sensing of water core and internal browning in whole apples. Paper 98-6020.
ASAE, St. Joseph, MI.
Keener, K. M., Stroshine, R. L., & Nyenhuis, J. A. (1999). Evaluation of low field (5.40
MHz) proton magnetic resonance measurements of Dw and T2 as methods of
nondestructive quality evaluation of apples. Journal of the American Society for
Horticultural Science, 124, 289-295.
Kennedy, C. B., Balcom, B. J., & Mastikhin, I. V. (1998). Three-dimensional magnetic
resonance imaging of rigid polymeric materials using single-point ramped imaging
with T-1 enhancement (SPRITE). Canadian Journal of Chemistry, 76(11), 1753-1765.
Kerr, W. L., Clark, C. J., McCarthy, M. J., & de Ropp, J. S. (1997). Freezing effects in fruit
tissue of kiwifruit observed by magnetic resonance imaging. Scientia Horticulturae,
69, 169-179.
Koizumi, M., Naito, S., Haishi, T., Utsuzawa, S., Ishida, N., & Kano, H. (2006). Thawing
of frozen vegetables observed by a small dedicated MRI forfood research. Magnetic
Resonance Imaging, 24, 1111-1119.
Kotwaliwale, N., Curtis, E., Othman, S., Naganathan, G. K., Subbiah, J. (2012). Magnetic
resonance imaging and relaxometry to visualize internal freeze damage to pickling
cucumber. Postharvest Biology and Technology, 68, 22-31.
Kumpoun, W., Motomura, Y., & Harada, Y. (2003). Inhibition of aspergillus rot by sorbitol
in apple fruit with watercore symptoms. Postharvest Biology and Technology, 29, 121-
127.
Kuo, M. I., Anderson, M. E., & Gunasekaran, S. (2003). Determining effects of freezing
on pasta filata and non-pasta filata mozzarella cheeses by Nuclear Magnetic Resonance
Imaging. Journal of Dairy Science, 86, 2525-2536.
Laghi, L., Cremonini, M. A., Placucci, G., Sykora, S., Wright, K., & Hills, B. (2005). A
proton NMR relaxation study of hen egg quality. Magnetic resonance imaging, 23(3),
501-510.
Laghi, L., Babini, E., Bordoni, A., Ciampa, A., Danesi, F., Di Nunzio, M., Picone, G.,
Valli, V., & Capozzi, F. (2013). Time domain measurements and high resolution
Raffo, A., Gianferri, R., Barbieri, R., & Brosio, E. (2005). Ripening of banana fruit
monitors by water relaxation and diffusion 1H-NMR measurements. Food Chemistry,
89, 149-158.
Rahman, M. S., & Conrad, O. P. (2007). Drying and food preservation. In M. S. Rahaman
(Ed.), Handbook of food preservation (Second edition, 403-432). Boca Raton, Florida:
CRC Press.
Ramos-Cabrer, P., Van Duynhoven, J. P. M., Timmer, H., & Nicolay, K. (2006).
Monitoring of moisture redistribution in multicomponent food systems by use of
magnetic resonance imaging. Journal of Agricultural and Food Chemistry, 54, 672-
677.
Ruan, R., Litchfield, J. B., & Eckoff, R. (1992) Simultaneous and nondestructive
measurement of transient moisture profiles and structural changes in corn kernels
during steeping using microscopic nuclear magnetice resonance imaging. Cereal
Chemistry, 69(6), 600-606.
Salerno, A., Pierandrei, F., Rea, E., & Sequi, P. (2005). Definition of internal morphology
and structural changes due to dehydration of radish (raphanussatinus l. cv. Suprella)
using Magnetic Resonance Imaging spectroscopy. Journal of Food Quality, 28, 428-
438.
Santagapita, P., Laghi, L., Panarese, V., Tylewicz, U., Rocculi, P., & Dalla, R. (2013).
Modification of transverse NMR relaxation times and water diffusion coefficients of
kiwifruit pericarp tissue subjected to osmotic dehydration. Food and Bioprocess
Technology, 6, 1434-1443.
Sekiyama, Y., Horigane, A. K., Ono, H., Irie, K., Maeda, T., & Yoshida, M. (2012). T2
distribution of boiled dry spaghetti measured by MRI and its internal structure
observed by fluorescence microscopy. Food Research International, 48, 374-379.
Sequi, P., Dell’Abate, M. T., & Valentini, M. (2007). Identification of cherry tomatoes
growthorigin by means of magnetic resonance imaging. Journal of the Science of Food
and Agriculture, 87, 127-132.
Shaarani, S. M., Nott, K. P., & Hall, L. D. (2006). Combination of NMR and MRI
quantitation of structure and structure changes for convection cooking of fresh chicken
meat. Meat Science, 72, 398-403.
Shaarani, S. M. D., Cardenas-Blanco, A., Gao Amin, M. H., Soon, N. G., & Hall L. D.
(2010). Monitoring development and ripeness of oil palm fruit (Elaeis guneensis) by
MRI and bulk NMR. International Journal of Agricultural and Biology, 12, 101-105.
Simmons, A. L., & Vodovotz, Y. (2012). The effects of soy on freezable bread dough: A
magnetic resonance study. Food Chemistry, 135, 659-664.
Simons, R. K. (1968). The morphological and anatomical characteristics ofwatercorein
apples. Proceedings of the American Society for Horticultural Science, 93, 762-774.
Snaar, J. E. M., & Van As, H. (1992). Probing water compartments and membrane
permeability in plant cells by 1H NMR relaxation measurements. Biophysical Journal,
63, 1654-1658.
Stork, H., Gädke, A., & Nestle, N. (2006). Single-sided and semisingle-sided NMR sensors
for highly diffusive samples: application to bottled beverages. Journal of Agricultural
and Food Chemistry, 54, 5247-5252.
Taglienti, A., Massantini, R., Botondi, R., Mencarelli, F., & Valentini, M. (2009).
Postharvest structural changes of Hayward kiwifruit by means of magnetic resonance
imaging spectroscopy. Food Chemistry, 114, 1583-1589.
Thybo, A. K., Bechmann, I. E., Martens, M., & Engelsen, S. B. (2000). Prediction of
sensory texture of cooked potatoes using uniaxial compression, near infrared
spectroscopy and low field 1H NMR Spectroscopy. LWT-Food Science and
Technology, 33, 103-111.
Thybo, A. K., Andersen, H. J., Karlsson, A. H., Donstrup, S., & Stodkilde-Jorgensen, H.
(2003). Low-field NMR relaxation and NMR-imaging as tools in differentiation
between potato sample and determination of dry matter content in potatoes. LWT-Food
Science and Technology, 36, 315-322.
Tylewicz, U., Panarese, V., Laghi, L., Rocculi, P., Nowacka, M., Placucci, G., & Dalla
Rosa, M. (2011). NMR and DSC water study during osmotic dehydration of Actinidia
deliciosa and Actinidia chinensis kiwifruit. Food Biophysics, 6(2), 327-333.
Tylewicz, U., Aganovic, K., Vannini, M., Toepfl, S., Bortolotti, V., Dalla Rosa, M., Oey,
I., & Heinz, V. (2016). Effect of pulsed electric field treatment on water distribution of
freeze-dried apple tissue evaluated with DSC and TD-NMR techniques. Innovative
Food Science & Emerging Technologies, 37, 352-358.
Valentini, M., Sequi, P., Ciampa, A., Ritota, M., Taglienti, A., Cozzolino, S., Conte, L., &
Terlizzi, M. (2009). Qualità tramite Risonanza Magnetica per Immagini: valutazione
della conservazione e dei trattamenti agronomici. Italus Hortus, 16(5), 324-328.
Van Duynhoven, J., Voda, A., Witek, M., & Van As, H. (2009). Time-domain NMR
applied to food products. Annual Reports on NMR Spectroscopy, 69, 145-197.
Vandeginste, R. G. M., Massart, D. L., Buydens, L. M. C., De Jong, S., Lewi, P. J., &
Smeyers-Verbeke, J. (1998). Handbook of Chemometrics and Qualimetrics: part B
(First edition). Amsterdam, Netherlands: Elsevier Science Publisher.
Veliyulin, E., van der Zwaag, C., Burk, W., Erikson, U. (2005). In vivo determination of
fat content in Atlantic salmon (Salmo salar) with a mobile NMR spectrometer. Journal
of the Science of Food and Agriculture, 85, 1299-1304.
Venkataramanan, L., Song, Y. Q., & Hurlimann, M. D. (2002). Solving Fredholm integrals
of the first kind with tensor product structure in 2 and 2.5 dimensions. IEEE
Transactions on Signal Processing, 50, 1017-1026.
Wang, N., & Brennan, J. G. (1995). Changes in structure, density and porosity of potato
during dehydration. Journal of Food Engineering, 24, 61-67.
Wang, S. Y., Wang, P. C., & Faust, M. (1988). The non-destructive detection of watercore
in apple with Nuclear Magnetic Resonance Imaging. Scientia Horticulturae, 35, 227-
234.
Xin, Y., Zhang, M., & Adhikari, B. (2013). Effect of trehalose and ultrasound-assisted
osmotic dehydration on thestate of water and glass transition temperature of broccoli
(Brassicaoleracea L. var. botrytis L.). Journal of Food Engineering, 119, 640-647.
Xu, B., Zhang, M., Bhandari, B., Cheng, X. F., & Sun, J. (2015). Effect of ultrasound
immersion freezing on the qualityattributes and water distributions of wrapped red
radish. Food and Bioprocess Technology, 8, 1366-1376.
Xu, Z., Morris, R. H., Bencsik, M., & Newton, M. I. (2014). Detection of virgin olive oil
adulteration using low field unilateral NMR. Sensors, 14, 2028-2035.
Zion, B., Chen, P., & McCarthy, M. J. (1995). Detection of bruises in magnetic resonance
images of apples. Computers and Electronics in Agriculture, 13, 289-299.
Zou, Y., Xie, R., Ding, Y., & Arad, A. (2016). Determining Uncertainty in NMR T2
distribution using frequentist method. Applied Magnetic Resonance, 47(10), 1081-
1094.
Chapter 8
ABSTRACT
*
Corresponding Author address. Email: clinio.locatelli@unibo.it.
health in consequence of their consumption. In particular, mussels, clams and oysters are
filtering organisms, and then they call particular attention and require inspections before
being sold on the market.
Other food matrices particularly influenced by pollution from toxic metals are also the
food supplements, e.g., wine and tea. These beverages may be rich in toxic metals, as a
result of pesticides and fertilisers employed in agricultural treatments prior to the harvest
of raw materials.
In any case, the metal determination in food matrices evidently must be accurate,
reproducible and especially it must show very low limits of detection.
The present work reports and discusses the different analytical methodologies for the
voltammetric and spectroscopic determination of some potentially toxic elements (copper,
zinc, cadmium, mercury, tin, lead and antimony) in mussels, clams, oysters, fishes, wine
and tea. This work focuses also on the critical comparison between voltammetric and
spectroscopic instrumental techniques.
1. INTRODUCTION
The present work reports and discusses the different analytical procedure for the
determination of toxic elements [Cu(II), Zn(II), Cd(II), Hg(II), Sn(II), Pb(II) and Sb(III)]
in sea food (mussels, clams, oysters and fishes) and in food supplements (wine and tea).
The proposed methods are partly totally new, and in part are modifications of
procedures already proposed by the same authors; the novelty in this work is the aim to
improve the various fundamental analytical parameters, such as the detection limit,
accuracy (trueness and precision) and, in the case of voltammetry, also the problem of
eventual signal interference between the various metals characterized by very close peak
potential values.
Finally, this work is also proposed as an opportunity to critically highlight and discuss
a comparison between voltammetry and spectroscopy, which undoubtedly are the most
employed instrumental methodologies dedicated to metal determination in environmental
matrices. The reference spectroscopic methodology here chosen is the atomic absorption
spectroscopy (AAS) with electrothermal atomization (ET-AAS) or by cold vapour atomic
atomic absorption spectrometry (CV-AAS).
2. EXPERIMENTAL
Voltammetry
Keeping the temperature at 20.0 ± 0.5°C, the solutions were deaerated by water-
saturated pure nitrogen for 5 min prior to measurements, while a nitrogen blanket was
maintained above the solution during the analysis. The solutions were deaerated after each
standard addition for 1 min. In the electrolysis step the solutions were stirred using a
magnetic stirrer.
Spectroscopy
accurately weighed and transferred into a Pyrex digestion tube together with 1.2 g K2Cr2O7
and 20 mL H2O. A condenser was connected to the digestion tube and 20 mL H2SO4 were
slowly added. The digestion tube was transferred to the hot block preheated to 180°C and
digestion was allowed to proceed for 60 min to completion. After cooling to room
temperature, the condenser was removed, rinsed with three 5-mL portions of H2O and the
washings were added to the digested matter. The open digestion tube, without the
condenser, was returned to the hot-block and boiled for additional 30 min. Finally, after
cooling, the digested solution was diluted to 50 mL.
All acids and chemicals were suprapure grade (Merck, Germany). Acidic stock
solutions of all the elements (1000 mg L-1, Sigma-Aldrich, Germany) were employed in
the preparation of reference solutions at varying concentrations, using for diluting water
demineralised through a Milli-Q system.
NIST-SRM (from National Institute of Standards and Technology, Gaithersburg, MD,
USA) and BCR-CRM (from Institute for Reference Materials and Measurements,
European Commission, Joint Research Centre, Geel, Belgium) were employed as Standard
Reference Materials (SRMs) to optimize and set up the analytical procedures, i.e., Oyster
Tissue NIST-SRM 1566a and Mussel Tissue BCR-CRM 278 (for mussels, clams and
oysters), Cod Muscle BCR-CRM 422 (for fishes), Spinach Leaves NIST-SRM 1570a,
Tomato Leaves NIST-SRM 1573a and Apple Leaves NIST-SRM 1515 (for tea).
4. LIMITS OF DETECTION
In the aqueous reference solution, in the solutions obtained by digestion of the standard
reference material and of all real samples, the limits of detection (LODs) for both
techniques (voltammetry and spectroscopy) were obtained by the equation LOD=K sy/x/b
[10], where sy/x and b are the estimated standard deviation and the slope of the analytical
calibration function of each element, respectively, with a 99.7% (K = 3) confidence level
[11].
In the case of voltammetric technique, since the analytical calibration functions were
determined by standard addition method, it was possible to obtain the LODs directly also
in the real matrix.
5. METALS OF INTEREST
The present work will concern and discuss the most significant elements that are
present in sea foods, wine and tea mainly for environmental and anthropic causes, and the
different analytical methodologies for their voltammetric and spectroscopic determination.
Table 2 reports the metals investigated together with the supporting electrolytes
employed for the voltammetric determination.
The same solutions were also used for the spectroscopic determinations.
Moreover, for each metal or group of metals, a brief introduction is reported to
emphasize the most recent, but also the most significant and interesting works relevant to
the various analytical procedures and techniques for determining the same metals in each
matrix.
Certainly Cu(II), Zn(II), Cd(II) and Pb(II), but also Sn(II) and Sb(III), are the most
studied elements, and many works in the literature concern analytical procedures for the
voltammetric and spectroscopic analytical determination of such metals in mussels, clams,
oysters and fishes.
However, among the six metals in subject, certainly Cd(II), Sn(II), Pb(II) and Sb(III)
deserve particular attention owing to their extreme toxicity and hazard to human health.
With regard to the principal anthropogenic emission sources of Cd(II) in the
environment, they are mainly the industries for production of paints, but a lot of
responsibility for the presence of cadmium into the atmosphere is ascribable to the
combustion of fossil fuels, the wear of the tires and, of course, also the solid waste
incinerators.
Sn(II) is mainly employed in the production of canned goods for the preservation of
food.
Pb(II) is strictly linked to vehicular traffic, since for long time it has been present in
gasoline as lead-tetraethyl with anti-knock function. Other sources of emission
are also industrial activities for the production of paints and incinerators of
municipal solid waste.
Sb(III) is used in the paint and glazes, ceramics, rubbers and in the military industries.
With regard to Cu(II) and Zn(II), they are, as also Sn(II), semi-essential metals, since
absolutely necessary for man, but toxic at high concentrations. Their possible and main
anthropogenic pollution sources are the industries for the production of paints, fungicides
and cosmetics.
In any case, if poisonings by Pb (II), Sn(II), Sb(III) and Cd (II) are possible, although
not very frequent, generically for Cu (II) and Zn (II) these are extremely rare, since these
elements are largely present in all environmental matrices and especially in food, so that
their possible excess in humans occurs when their regulatory mechanisms of absorption
are malfunctioning.
Without doubt, for the Cu(II), Pb(II), Cd(II) and Zn(II) determinations, the most widely
used technique is the spectroscopy in all its versions [12-20]. Other techniques as Neutron
Activation Analysis (NAA) [21-23] or Voltammetry [24-30] are seldom employed.
Moreover, it is important to highlight two papers proposing the use of Energy Dispersive
X-Ray Fluorescence (EDXRF) [31, 32], of slotted tube trap-atomic absorption
spectrometry (STAT-AAS), inductively coupled plasma atomic emission spectrometry
(ICP-AES) [33] and of Laser Ablation Inductively Coupled Plasma Mass Spectrometry
(LA-ICP-MS) [34] for the determination of the four elements in mussels, clams and
oysters.
Finally, very interesting is the work by Wu and coworkers [35] that proposes an
“artificial mussel” for monitoring heavy metals in marine environments.
Works on the Sn (II) and Sb (III) determination in mussels, clams, oysters and fishes
are not numerous, and very often the investigation concerns the organic tin compounds.
They propose spectroscopy [36-40], NAA [41], voltammetry [42] and AFS [43], while
very interesting is the paper by Ferrarello and coworkers [44] that employs size-exclusion
Mussels, clams, oysters and fishes were sampled during summer 2015 in the Goro Bay
(Province of Ferrara, Italy; salinity: 2.7%; average temperature from September to October:
23-25°C). This is a very important area devoted to the fishing and breeding of mussels and
clams for food, but with considerable pollution problems due to its location in proximity
of the Po River mouth, which is the carrier towards the sea of a large amount of industrial
and domestic waste water.
Moreover, the waterways are also strongly influenced by atmospheric pollution load,
and then they result to be also carriers of air pollutants, especially including those related
to vehicular traffic as in particular the platinum group metals (PGMs).
In the Goro Bay flow three branches of the Po river delta mouth, i.e., Po of Volano
(site A), Canal Bianco (site B) and Po of Goro (site C).
Samplings of mussels, clams, oysters were carried out in front of the points where such
branches flow into the bay and in the central area (site D) of the bay itself, while an
additional sampling (site F), for eventual comparisons, was chosen at open sea.
With regard to the samplings of the fishes, they were only two, considering their
extreme mobility in water, one inside the bay as mean sample (site E), and the second at
open sea (site F).
All samplings were performed using a rubber boat without motor to avoid any kind of
pollution.
Again, the fish samples were accurately rinsed with water and the edible parts were
carefully managed by a plastic appliance.
All samples were frozen and then lyophilised for 30 h. After that treatment, the samples
were homogenized thoroughly in an agate mortar.
For all kinds of real samples (mussels, clams, oysters and fishes), the same sample
mineralization procedure was carried out.
The sample preparation of Mussel Tissue BCR-CRM 278, Oyster Tissue NIST-SRM
1566a, Cod Muscle BCR-CRM 422, and of real samples of mussels, clams, oysters and
fishes was the following: approximately 0.5-1.0 g, accurately weighed, was placed in a
platinum crucible and dissolved in 5 mL 69%w/w HNO3 + 3 mL 37%w/w HCl + 7 mL 98%w/w
H2SO4 at 130-150°C. The mixture was evaporated to dryness and, after cooling, soluble
salts were dissolved in 25 mL of the supporting electrolyte.
0.49 mol L-1 ammonium citrate buffer pH 5.5 + 7.1 · 10-3 mol L-1 EDTA-Na2 + 0.15
mol L-1 KCl was employed as supporting electrolyte, which was the aqueous reference
solution.
As regards the composition of the supporting electrolyte, it should be emphasized that
the presence of EDTA-Na2, as already proposed by the authors [45, 46], is necessary to
obtain better resolution for the Sn(II) and Sb(III) peaks in presence of Pb(II).
Indeed, in the marine ecosystems Pb(II) is always and inevitably present.
Unfortunately it shows a peak potential very close to those of Sn(II) and Sb(III).
In the supporting electrolyte without the presence of EDTA-Na2 the peak potentials Ep
(V/AgAgClCl-satd.) are: Ep Pb(II) = -0.542 ± 0.015, Ep Sn(II) = -0.649 ± 0.010 and Ep Sb(III) =
-0.755 ± 0.010.
The problem is decidedly important, considering that the same problem is also present
in the standard reference material and in the real samples collected in the sampling sites.
Also in this case, our methodological procedure proposes the possibility to shift the
interferent peaks towards more cathodic potential values by adding EDTA-Na2. Indeed, the
presence of EDTA-Na2 shifts towards more cathodic potential the values of all the three
elements: Pb(II): -0.563 ± 0.010; Sn(II): -0.719 ± 0.015 and Sb(III): -0.875 ± 0.015 V vs.
AgAgClCl-satd.. Evidently, the new peak position of the elements allows their better
resolution and then also their quantitative determination.
10-mL sample aliquots of supporting electrolyte or of solutions obtained in the
mineralization step of the standard reference material and of real samples were pipetted
into the voltammetric cell and deaerated for 5 min by bubbling water-saturated pure
nitrogen. The simultaneous determination of Cu(II), Zn(II), Cd(II), Sn(II), Pb(II) and
Sb(III) was carried out by Square Wave Anodic Stripping Voltammetry (SWASV) using
HMDE as working electrode.
The voltammetric experimental conditions are reported in Table 3.
The experimental peak potentials Ep (V/AgAgClCl-satd.) of Cu(II), Zn(II), Cd(II),
Sn(II), Pb(II) and Sb(III) are reported in Table 4.
The spectroscopic experimental conditions are reported in Table 1.
Ed -1.300
Ei -1.300
Ef +0.050
td 360
tr 10
dE/dt 100
E 50
0.010
0.100
10
u 600
[a]
Ed: deposition potential (V/Ag, AgCl, Cl-satd.); Ei: initial potential (V/Ag, AgCl, Cl-satd.); Ef: final potential (V/Ag, AgCl,
Cl-satd.); td: electrodeposition time (s); tr: delay time before the potential sweep (s); dE/dt: potential scan rate (mV/s);
ΔE: superposed potential amplitude (mV); τ: sampling time (s); υ: wave period (s); : wave increment (mV); u: stirring
rate (r.p.m.).
Table 5. LODs [a] of Cu(II), Pb(II), Sn(II), Sb(III), Cd(II) and Zn(II) determined in
the aqueous reference solution (µg L-1), in the solutions obtained by digestion of
Mussel Tissue BCR-CRM 278, Oyster Tissue NIST-SRM 1566a SRM and Cod
Muscle BCR-CRM 422, and in the solutions obtained by digestion of real samples
(calculated in µg L-1 and expressed in µg kg-1). The determined values are the mean
of 5 independent determinations. Confidence level: 95%
Table 6. Accuracy of the analytical procedure. The determined values are the mean
of 5 independent determinations. Confidence level: 95%
Voltammetry Spectroscopy
Elements Certified Determined e sr Determined e sr
Values Values (%) (%) Values (%) (%)
Cu(II) 9.600.16 9.050.63 -5.7 5.8 10.250.77 +6.8 6.1
Pb(II) 1.910.04 1.790.15 -6,3 5.7 1.770.17 -7.3 5.9
n(II) 7.73 * 7.300.47 -5.6 5.9 7.270.51 -6.0 6.2
Sb(III) 7.73 * 8.170.50 +5.7 5.8 7.230.55 -6.5 5.8
Cd(II) 0.340.02 0.320.03 -5.9 5.5 0.310.04 -8.8 5.7
Zn(II) 762 71.05.5 -6.6 5.9 81.76.1 +7.5 6.3
Voltammetry Spectroscopy
Elements Certified Determined e sr Determined e sr
Values Values (%) (%) Values (%) (%)
Cu(II) 66.3 70.24.5 +5.9 5.7 70.54.7 +6.3 6.0
Pb(II) 0.371 0.3490.027 -5.9 5.5 0.3950.035 +6.5 5.7
Sn(II) 3 ** 2.820.23 -6.0 5.9 2.800.29 -6.7 6.2
Sb(III) 0.010 ** 0.0090.002 -10.0 5.8 0.0110.002 +10.0 6.1
Cd(II) 4.15 4.380.31 +5.5 5.6 3.890.40 -6.3 5.9
Zn(II) 830 78157 -5.9 5.9 77369 -6.9 6.3
Voltammetry Spectroscopy
Elements Certified Determined e (%) sr Determined e (%) sr (%)
Values Values (%) Values
Cu(II) 1.050.07 0.990.08 -5.7 5.9 0.980.09 -6.7 6.0
Pb(II) 0.0850.015 0.0800.007 -5.9 5.7 0.0910.013 +7.1 5.9
Sn(II) 7.73 * 8.190.52 +6.0 5.9 7.220.55 -6.6 6.3
Sb(III) 7.73 * 7.280.50 -5.8 5.5 8.250.57 +6.7 6.1
Cd(II) 0.0170.002 0.0180.002 +5.9 5.4 0.0160.003 -5.9 5.8
Zn(II) 19.60.5 20.91.7 +6.6 5.8 21.01.9 +7.1 6.2
* Spiked sample concentration.
** Values not certified and given for information only.
Spiked-sample concentrations in the case of Sn(II) and Sb(III) in Mussel Tissue BCR-
CRM 278 and Cod Muscle BCR-CRM 422 were in all cases 7.73 µg g-1 for each metal.
The above reported additions were needed because the metals in subject were not present
in Mussel Tissue BCR-CRM 278 and Cod Muscle BCR-CRM 422 standard reference
materials. This procedure may seem anomalous, but, in our opinion, it was the only way:
certified concentrations of the metals were not available in the same reference materials.
Evidently, under these conditions, the trueness and precision data may be doubtful and
prudentially considered and/or discussed.
Once the procedure for the determination of Cu(II), Zn(II), Cd(II), Sn(II), Pb(II) and
Sb(III) was set up, the method was applied to mussels, clams, oysters and fishes sampled
in the Goro Bay (see section 6.1, “Sea Food Sampling Area and Sampling Sites”).
The experimental results of both techniques are reported in Table 7.
Table 7. Mean values of Cu(II), Pb(II), Sn(II), Sb(III), Cd(II) and Zn(II) (µg g-1) for
oysters, mussels and clams sampled in the Goro Bay. The determined values are the
mean of 5 independent determinations. Confidence level: 95% Voltammetry
Spectroscopy
A large amount of mercury is released every year into the environmental ecosystems
as a consequence of the human activities. The danger posed to human health by this form
of contamination comes mainly from food, i.e., from the ability of this element to enter the
natural alimentary chains, to accumulate in progressively larger quantities, and to reach
high toxic concentrations in the tissues of organisms that play a role in the human diet, a
fortiori for mussels, clams, oysters, marine filtering organisms, and fishes.
For the mercury determination spectroscopic techniques [15, 18, 20, 27, 29, 30, 34, 47-
58] are prevalently employed, rarely voltammetric techniques [15, 26, 59-66].
Although not recent, interesting are the works that quantify the inorganic mercury in
mussel samples by atomic fluorescence spectrometry (AFS) [67] and by NAA [68, 69].
For the mercury determination in sea food, very interesting are the works that propose
innovative approach, as for example thermal decomposition amalgamation atomic
absorption spectrometry (TDA-AAS) [70, 71], cation exchange chromatography
hyphenated with inductively coupled plasma mass spectrometry [72, 73] and flow injection
catalytic cold vapour atomic absorption spectrometry [74].
0.45 mol L-1 HClO4 + 0.25 mol L-1 NaCl + 6.9·10-3 mol L-1 EDTA-Na2 was employed
as supporting electrolyte (aqueous reference solution).
10-mL sample aliquots of supporting electrolyte or of solutions obtained in the
mineralization step of the standard reference material and of real samples were pipetted
into the voltammetric cell and deaerated for 5 min by bubbling water-saturated pure
nitrogen. The determination of Hg(II) was carried out by SWASV using GE as working
electrode.
The voltammetric experimental conditions are reported in Table 8.
Supporting electrolyte: 0.45 mol L-1 HClO4 + 0.25 mol L-1 NaCl + 6.9·10-3 mol L-1 EDTA-Na2.
Ed -0.050
Ei +0.250
Edes -0.800
Ef +0.950
td 270
tded 30
tr 10
dE/dt 100
E 50
0.010
0.100
10
u 600
[a]
Ed: deposition potential (V/Ag, AgCl, Cl-satd.); Ei: initial potential (V/Ag, AgCl, Cl-satd.); Edes: desorption potential (V/Ag,
AgCl, Cl-satd.); Ef: final potential (V/Ag, AgCl, Cl-satd.); td: electrodeposition time (s); tdes: desorption time (s); tr: delay
time before the potential sweep (s); dE/dt: potential scan rate (mV/s); ΔE: superposed potential amplitude (mV); τ:
sampling time (s); υ: wave period (s); : wave increment (mV); u: stirring rate (r.p.m.).
Table 10. LODs[a] of Hg(II) determined in the aqueous reference solution (µg L-1),
in the solutions obtained by digestion of Mussel Tissue BCR-CRM 278,
Oyster Tissue NIST-SRM 1566a and Cod Muscle BCR-CRM 422 standard
reference materials, and of real samples
(calculated in µg L-1 and expressed in µg kg-1).
The determined values are the mean of 5 independent determinations.
Confidence level: 95%
Once the procedure for the Hg(II) determination was set up, the method was applied
to mussels, clams, oysters and fishes sampled in the Goro Bay (see section 6.1, “Sea Food
Sampling Area and Sampling Sites”).
The experimental results relevant to both techniques are reported in Table 12.
Voltammetry Spectroscopy
Certified Determined e (%) sr (%) Determined e (%) sr (%)
Values Values Values
1887 199.613.5 +6.2 5.6 174.914.9 -7.0 5.7
Voltammetry Spectroscopy
Certified Determined e (%) sr (%) Determined e (%) sr (%)
Values Values Values
64.2 60.54.3 -5.8 5.8 60.14.7 -6.4 5.9
Voltammetry Spectroscopy
Certified Determined e (%) sr (%) Determine e (%) sr (%)
Values Values d Values
55916 52739 -5.7 5.9 59645 +6.6 6.1
Table 12. Mean values of Hg(II) (µg g-1) relevant to mussels, clams, oysters and
fishes sampled in the Goro Bay. The determined values are the mean of 5
independent determinations. Confidence level: 95%
As regards inorganic elements, the wine shows to have high concentrations for Na, K,
Ca and Mg, but also low concentrations of metals considered toxic even at trace level
concentrations (Pb and Cd), or potentially toxic, identified also as semi-essential metals,
i.e., essential only at low concentrations, but toxic at high concentrations (Cu and Zn).
These trace elements are generally present in herbicides and pesticides for vineyard
treatments, but the accumulation in grapes and wines may come from different sources
among which mainly the environmental pollution, but also the equipment used and the
methods employed in winery during the wine production cycle from vineyard to bottling.
Considering all the various sources of pollution, all steps of the vineyard-bottling cycle
is strictly regulated, and the analytical methodology employed must follow a precise
protocol [75].
For metals the European Community Legislation [76] and the International
Organisation “Organisation of Vine and Wine” (OIV) [77] have fixed the maximum
concentrations allowed in wine for toxic metals 0.150 mg L-1 [Pb(II)] and 0.010 mg L-1
[Cd(II)], while for semi-essential metals 1.0 mg L-1 [Cu(II)] and 5.0 mg L-1 [Zn(II)].
The Italian legislation provides a rather old and inadequate law, that does not consider
the toxic metals, with the exception of lead in subsequent amendments, but concerning
only copper, zinc, iron and manganese [78].
Concerning the determination of Cu(II), Zn(II), Cd(II) and Pb(II), spectroscopy in all
its versions is the more widely used technique [79-84] together with voltammetry [85-95].
Very interesting is the work proposed by Shen et al. [96] concerning the differentiation
of Chinese rice wines from different wineries based on mineral elemental fingerprinting.
Other two works are of interest: the former proposes the direct determination of Cd(II),
Cu(II) and Pb(II) in wines and grape juices by thermospray flame furnace AAS [97], the
latter concerning the trace element determination in wines using ETAAS and ultrasonic
nebulization coupled to inductively coupled plasma optical emission spectrometry (USN-
ICP-OES) [98].
15-mL red wine and 5-mL 69%w/w HNO3 were placed in a pyrex tube calibrated to 100
mL. The tube was inserted into a home-made block digester, gradually increasing the
temperature up to 130°C, keeping it for 30 min.
After this first step of 30 min, 1 mL H 2O2 35%w/w has been added, increasing the
temperature to 150°C, keeping it for 10 min.
Still keeping the temperature at 150°C, the procedure of additions of H2O2 every 10
min was repeated 10 times.
At the end of the digestive process the sample, completely clarified, was left to cool at
room temperature, and finally transferred into a volumetric flask.
0.1 mol L-1 acetic buffer pH 4.5 was employed as supporting electrolyte (aqueous
reference solution).
18-mL sample aliquots of supporting electrolyte + 2 mL of solutions obtained in the
mineralization step of wine were pipetted into the voltammetric cell and deaerated for 5
min by bubbling water-saturated pure nitrogen. The determination of Cu(II), Zn(II), Cd(II)
and Pb(II) was carried out by Differential Pulse Anodic Stripping Voltammetry (DPASV)
using HMDE as working electrode.
The voltammetric experimental conditions are reported in Table 13.
The experimental peak potentials Ep (V/AgAgClCl-satd.) of Cu(II), Zn(II), Cd(II) and
Pb(II) in the aqueous reference solution and in solutions obtained by digestion of wine
were practically the same, i.e., -0.020 [Cu(II)], -0.900 [Zn(II)], -0.580 [Cd(II)] and -0.430
[Pb(II)].
Table 13. Instrumental parameters for the determination of Cu(II), Zn(II), Cd(II)
and Pb(II) by DPASV[a]. Supporting electrolyte: 0.1 mol L-1 acetic buffer pH 4.5
Ed -1.100
Ei -1.100
Ef +0150
td 180
tr 10
dE/dt 20
E 50
0.065
0.250
r 600
[a]
Ed: deposition potential (V/Ag, AgCl, Cl-satd.); Ei: initial potential (V/Ag, AgCl, Cl-satd.); Ef: final potential (V/Ag, AgCl,
Cl-satd.); td: electrodeposition time (s); tr: delay time before the potential sweep (s); dE/dt: potential scan rate (mV/s);
ΔE: superposed potential amplitude (mV); τ: superposed pulse duration (s); υ: superposed pulse repetition (s); r: stirring
rate (r.p.m.).
In the case of determination of Cu (II), Zn (II), Cd (II) and Pb (II), the only interferent
could be Sn(II), since this metal, in 0.1 mol L-1 acetic buffer pH 4.5 supporting electrolyte,
shows to have a quasi-reversible/irreversible peak around 0.700/0.720 V vs. AgAgClCl-
satd.
Table 14. LODs[a] (µg L-1) determined in the aqueous reference solution and
in the solutions obtained by digestion of real samples. The determined values
are the mean of 5 independent determinations; confidence level: 95%
Table 15. Mean values of Cu(II), Zn(II), Cd(II) and Pb(II) (µg L-1) relevant to three
Italian commercial red wines Sangiovese. The determined values are the mean of 5
independent determinations. Confidence level: 95%
At the experimental conditions employed, the linear range for all the elements in the
aqueous reference solutions is:
Once the procedure for the determination of Cu(II), Zn(II), Cd(II) and Pb(II) was set
up, the method was applied to three Italian commercial red wines Sangiovese: two
produced in Emilia Romagna area (Italy) (sample S12 and S17) and the third in Marche
area (Italy) (sample S13). The analytical results are reported in Table 15.
Tea certainly has a millenary history, especially throughout the Orient, even if its use
in the western world is widespread and decidedly more and more increasing.
Persistent toxic substances tend to accumulate in all the environmental matrices owing
to anthropic activities, causing increasingly detrimental to food safety. Among these
substances, toxic metals, in particular, Cu(II), Zn(II), Cd(II), Pb(II) and Hg(II), are
probably the most dangerous for human health and wildlife [99-104], since they are
0.010 0.010
0.100 0.100
10 10
r 600 600
In the commonly used supporting electrolytes, reduction peak potentials of each metal
are often quite close, thus hindering simultaneous voltammetric determination of
neighbouring elements.
In the specific case of voltammetric determination of Hg(II) at the GE electrode, the
most important interfering species, qualitatively investigated by Hatle [123], appear to be
Cu(II). We partially disagree with the data reported by Hatle [123], who affirms that Cu(II)
does not interfere at concentrations lower than 10 mg L-1. In fact, interference is strictly
linked to the Hg(II)-Cu(II) concentration ratios. Indeed, only qualitative investigations
show that Cu(II) concentrations higher than 400-450 times the Hg(II) concentration, with
Cu(II)-Hg(II) peak potential differences ΔEp lower than 0.21-0.23 V, actually cause strong
voltammetric signal interference, evidently hindering correct determination of Hg(II).
Moreover, given that Hg(II) concentrations are generally very low in real matrices, it is
certainly not correct to fix a concentration limit since, depending on the Hg(II):Cu(II)
concentration ratios, Hg(II)-Cu(II) interference may be present even at Cu(II)
concentrations decidedly lower than 10 mg L-1.
Table 18. Limits of detection (LOD), calculated in g L-1 in the Aqueous Reference
Solutions, and expressed in µg kg-1 in the Standard Reference Materials. The
determined values are the mean of 5 independent determinations; confidence level:
95%.Working electrodes: HMDE (a), GE (b)
Table 19. Accuracy of the analytical procedure. The determined values are the mean
of 5 independent determinations. Confidence level: 95%. Concentrations: μg g-1
Voltammetric measurements.
Supporting electrolyte: 0.4 mol L-1 HCl + 0.2 mol L-1 NaCl + 9.6x10-3 mol L-1 EDTA-Na2.
Working electrodes: GE (a); HMDE (b).
Element Certified Determined e sr
value value (%) (%)
Hg(II) 0.030 0.0320.002 +6.7 5.3
Cu(II) a 12.2 12.90.8 +4.9 5.5
Spinach Leaves Cu(II) b 12.2 11.40.9 -6.6 5.3
NIST-SRM 1570a [a]
Pb(II) (0.20) 0.180.02 -10.0 4.9
Cd (II) 2.89 2.720.20 -5.9 5.2
Zn (II) 82 876 +6.1 5.6
Tomato Leaves Hg(II) 0.034 0.0360.003 +5.9 5.7
NIST-SRM 1573a Cu(II) a 4.70 4.950.29 +5.3 5.3
Cu(II) b 4.70 4.990.31 +6.2 5.5
Pb(II) 10.8 11.50.9 +6.5 5.4
Cd (II) 1.52 1.600.10 +5.3 5.0
Zn (II) 30.9 32.62.3 +5.5 5.8
Apple Leaves Hg(II) 0.044 0.0410.004 -6.8 5.3
NIST-SRM 1515 Cu(II) a 5.64 5.310.38 -5.9 5.7
Cu(II) b 5.64 5.950.39 +5.5 5.5
Pb(II) 0.47 0.440.05 -6.4 5.6
Cd (II) (0.013)[a] 0.0140.002 +7.7 5.0
Zn(II) 12.5 11.71.5 -6.4 5.8
Spectroscopic measurements.
Element Certified Determined e sr
value value (%) (%)
Hg(II) 0.030 0.0330.004 -10.0 5.5
Cu(II) 12.2 13.11.0 +7.4 5.9
Spinach Leaves
Pb(II) (0.20)[a] 0.220.03 +10 5.8
NIST-SRM 1570a
Cd (II) 2.89 3.080.21 +6.6 6.0
Zn (II) 82 767 -7.3 6.0
Hg(II) 0.034 0.0310.003 -8.8 5.9
Cu(II) 4.70 5.000.35 +6.4 5.7
Tomato Leaves
Pb(II) 10.8 11.60.9 +7.4 5.6
NIST-SRM 1573a
Cd (II) 1.52 1.420.13 -6.6 5.8
Zn (II) 30.9 29.12.1 -5.8 6.0
Hg(II) 0.044 0.0480.005 +9.9 5.5
Cu(II) 5.64 5.990.41 +6.2 5.8
Apple Leaves
Pb(II) 0.47 0.510.06 +8.5 5.9
NIST-SRM 1515
Cd (II) (0.013)[a] 0.0120.002 -7.7 5.7
Zn(II) 12.5 13.41.1 +7.2 6.1
[a]
Values in parenthesis are not certified and are given for information only.
Once the procedures for determination of Cu(II), Zn(II) Cd(II), Hg(II), Pb(II) in
standard reference materials were set up, the methods were transferred to three kinds of
green tea from India (A), China (B) and Japan (C), and to three kinds (D, E, F) of
commercial green tea in filters purchased at the local markets in Ravenna (Italy).
To prepare real samples for voltammetric and spectroscopic analyses, all kinds of tea
sample were lyophilised, powdered, homogenised, dried at 80°C for 24h and solubilised
for the analyses as described above (see section 9.1: “Sample Preparation before the
Instrumental Determination”).
Experimental results relevant to tea samples are reported in Table 20.
Table 20. Mean values of the Hg(II), Cu(II), Pb(II), Cd(II), Zn(II) relevant to three
kinds of green tea from India (A), China (B) and Japan (C), and to three kinds (D, E
and F) of commercial green tea in filters purchased at the local markets in Ravenna
(Italy). Concentration: g g-1, dry weight. Working electrodes: Hanging Mercury
Drop Electrode (a), Gold Electrode (b)
Element Technique A B C D E F
Hg SWASV 0.590.02 0.770.05 0.520.03 0.360.03 0.190.03 0.510.03
CV-AAS 0.600.03 0.820.06 0.500.04 0.380.04 0.210.02 0.490.04
Cu SWASVa 20.51.1 29.71.6 25.11.4 14.30.8 10.50.8 15.20.9
SWASVb 21.30.9 30.61.7 24.2.1.7 13.80.9 10.90.6 16.00.7
ET-AAS 18.91.3 30.21.4 26.31.5 14.50.7 11.20.7 15.50.8
Pb SWASV 2.960.19 2.070.09 2.560.15 0.490.03 0.620.05 0.390.04
ET-AAS 3.230.23 2.150.14 2.630.14 0.510.04 0.660.04 0.370.03
Cd SWASV 0.670.04 0.870.06 0.110.02 0.070.02 0.200.02 0.270.02
ET-AAS 0.690.05 0.920.07 0.130.03 0.050.02 0.230.02 0.290.03
Zn SWASV 77.13.9 89.63.7 56.33.7 43.72.7 35.71.7 47.32.7
ET-AAS 79.34.3 87.14.9 59.03.4 45.92.9 37.21.9 49.62.5
Evidently, basing on these results, it is not possible to discriminate between the metal
content fractions ascribed to practices prior to harvest, and the metal content due to
environmental pollution. In fact, specific aspects relevant to this issue would require further
surveys, which are beyond the scope of the present work.
allow simultaneous metal determinations, it does not require enrichment steps, like solvent
extraction and/or particular sample treatments. Another important advantage is that
multiple standard addition method can be carried out by direct sequential additions in the
voltammetric cell. This makes the method particularly rapid and practical, with low
reagent-consumption [128].
However, it is important to underline the fact that both ICP and ICP/MS also permit
multi-element determination. The great advantage in using voltammetry is certainly the
equipment-related costs: very low in the case of voltammetry, extremely high in the case
of ICP and ICP/MS, in the latter case as much as 25-30 times higher.
ACKNOWLEDGMENTS
This investigation was supported by the University of Bologna (Funds for Selected
Research Topics).
REFERENCES
[1] Locatelli, C. Metals. In: Comprehensive Analytical Chemistry, Y. Picò, Editor, 2008,
Vol. 51, Ch. 16, pp. 571-598, Elsevier Publ., Amsterdam.
[2] Jovic, M; Stankovic, S. Human exposure to trace metals and possible public health
risks via consumption of mussels Mytilus galloprovincialis from the Adriatic coastal
area. Food and Chemical Toxicology, 2014, 70, 241-251.
[3] Olmedo, P; Pla, A; Hernandez, AF; Barbier, F; Ayouni, L; Gil, F. Determination of
toxic elements (mercury, cadmium, lead, tin and arsenic) in fish and shellfish
samples. Risk assessment for the consumers. Environmental International, 2013, 59,
63-72.
[4] Bonfil, Y; Brand, M; Kirowa-Eisner, E. Trace determination of mercury by anodic
stripping voltammetry at the rotating electrode. Analytica Chimica Acta, 2000, 424,
65-76.
[5] Shi, Z; Lipkowski, J. Chloride adsorption at the Au(111) electrode surface. Journal
of Electroanalytical Chemistry, 1996, 403, 225-239.
[6] Salaun, P; van den Berg, CMG. Voltammetric detection of mercury and copper in
seawater using a gold microwire electrode. Analytical Chemistry, 2006, 78, 5052-
5060.
[7] Shi, Z; Wu, S; Lipkowski, J. Investigations of Cl- adsorption at the Au(111) electrode
in the presence of underpotentially deposited copper atoms. Journal of
Electroanalytical Chemistry, 1995, 384, 171-177.
[8] Welz, B; Sperling, M. Atomic Absorption Spectrometry, 3rd Edition, 1999, Wiley
VCH, Weinheim.
[9] Landi, S; Fagioli, F; Locatelli, C. Determination of total mercury in sea food and
other protein-rich products. Journal of AOAC International, 1992, 75, 1023-1028.
[10] Miller, JN; Miller, JC. Statistics and Chemometrics for Analytical Chemistry, 2010,
6th Edition, Pearson Education Ltd. Publ., Ashford Colour Press Ltd., Gosport, U.K.
[11] International Union of Pure and Applied Chemistry - Analytical Chemistry Division
1978. Nomenclature, symbols, units and their usage in spectrochemical analysis —
II. Data interpretation. Spectrochim. Acta, 1978, 33B, 241-245.
[12] Psoma, AK; Pasias, IN; Rousis, IN; Barkonikos, KA; Thomaidis, NS. 2014.
Development, validation and accreditation of a method for the determination of Pb,
Cd, Cu and As in sea food and fish feed. Food Chemistry, 2014, 151, 72-78.
[13] Alfonso, JA; Handt, H; Mora, A; Vasquez, Y; Azocar, J; Marcano, E. Temporal
distribution of heavy metal concentrations in oysters Crassostrea rhizophorae from
the centralVenezuelan coast. Marine Pollution Bulletin, 2013, 73, 394-398.
[14] Zmozinski, AV; Passos, LD; Damin, ICF; Espirito Santo, MAB; Vale, MGR; Silva,
MM. Determination of cadmium and lead in fresh fish samples by direct sampling
electrothermal atomic absorption spectrometry. Analytical Methods, 2013, 5, 6416-
6424.
[15] Locatelli, C; Melucci, D. Sequential voltammetric determination of mercuri(II) and
toxic metals in environmental bio-monitors: application to mussels and clams.
International Journal of Environmental Analytical Chemistry, 2010, 90(1), 49-63.
[16] Turkmen, M; Ciminli, C. Determination of metals in fish and mussel species by
inductively coupled plasma-atomic emission spectrometry. Food Chemistry, 2007,
103, 670-675.
[17] Manutsewee, N; Aeungmaitrepirom, W; Varanusupakul, P; Imyim, A.
Determination of Cd, Cu and Zn in fish and mussel by AAS after ultrasound-assisted
acid leaching extraction. Food Chemistry, 2007, 101, 817-824.
[18] Locatelli, C. Heavy metal determinations in algae, mussels and clams. Their possible
employment for assessingthe sea water quality criteria. Journal de Physique, 2003,
JP 107(II), 785-788.
[19] Farkas, A; Salanki, J; Varanka, I. Heavy metal concentrations in fish of Lake
Balaton. Lakes & Reservoirs: Research and Management, 2000, 5, 271-279.
[20] Locatelli, C; Fagioli, F. Determination at trace level of heavy metals in mussels and
clams, Annali di Chimica (Rome), 1997, 87(7-8), 535-538.
[21] Pantelica, A; Ene, A; Georgescu, II. Instrumental neutron activation analysis of some
fish species from Danube River in Romania. Microchemical Journal, 2012, 103,
142-147.
[22] Essumang, DK. Analysis and human health risk assessment of arsenic, cadmium and
mercury in Manta birostris (Manta ray) caught along the Ghanaian coastline. Human
and Ecological Risk Assessment, 2009, 15(5), 985-998.
[23] Fukushima, M; Tamate, H; Nakano, Y. Trace element determination in soft tissues
of marine bivalves by activation analysis. Journal of Radioanalytical and Nuclear
Chemistry, 2003, 255(2), 231-234.
[24] Sobhanardakani, S; Tayebi, L; Farmany, A; Cheraghi, M. Analysis of trace elements
(Cu, Cd and Zn) in the muscle, gill, and liver tissues of some fish species using
anodic stripping voltammetry. Environmental Monitoring and Assessment, 2012,
184, 6607-6611.
[25] Anzano, J; Lasheras, R-J; Bonilla, B; Bonilla, A; Lanaja, J; Peribanez, MA; Gracia-
Salinas, M-J; Anwar, J; Shafique, U. Determination of trace metals by
voltamperometry in zebra mussel (Dreissena polymorpha) employed as
environmental bio-indicator. Green Chemistry Letters and Reviews, 2011, 4(3), 261-
267.
[26] Meucci, V; Laschi, S; Minunni, M; Pretti, C; Intorre, L; Soldani, G; Mascini, M. An
optimized digestion method couplet to electrochemical sensor for the determination
of Cd, Cu, Pb and Hg in fish by square wave anodic stripping voltammetry. Talanta,
2009, 77, 1143-1148.
[27] Locatelli, C; Fabbri, D; Torsi, G. Heavy metal determinations in algae and clams and
their possibile employment for assessing the sea water qualità criteria. Annali di
Chimica (Rome), 2001, 91, 425-434.
[28] Locatelli, C; Torsi, G. Heavy metal determination in aquatic species for food
purposes. Annali di Chimica (Rome), 2001, 91, 65-72.
[29] Locatelli, C. Proposal of new analytical procedures for heavy metal determinations
in mussels, clams and fishes. Food Additives and Contaminants, 2000, 17(9), 769-
774.
[30] Fagioli, F; Locatelli, C; Landi, S. Heavy metals in the Goro Bay: sea water,
sediments and mussels, Annali di Chimica (Rome), 1994, 84, 129-140.
[31] Bilo, F; Moscoso, S; Borgese, L; Delbarba, MV; Zacco, A; Bosio, A; Federici, S;
Guarienti, M; Presta, M; Bontempi, E; Depero, LE. Total reflection X-Ray
fluorescence spectroscopy to study Pb and Zn accumulation in Zebrafish embryos.
X-Ray Spectrometry, 2015, 44, 124-128.
[32] Mages, M; Bandow, N; Kuster, E; Brack, W; von Tumpling, W. Zinc and cadmium
accumulation in single zebrafish (Danio rerio) embryos ― A total reflection X-ray
fluorescence spectrometry application. Spectrochimica Acta Part B, 2008, 63, 1443-
1449.
[33] Yaman, M; Yaman, IH. Determination of trace metals levels of Cd, Pb, Cr, Ni and
Cu in two fish species by STAT-AAS and ICP-AES. Atomic Spectroscopy, 2013,
34(5), 191-198.
[34] Santos, MC; Wagner, M; Wu, B; Scheider, J; Oehlmann, J; Cadore, S; Becker, JS.
Biomonitoring of metal contamination in a marine prosobranch snail (Nassarius
reticulates) by imaging laser ablation inductively coupled plasma mass spectrometry
(LA-ICP-MS). Talanta, 2009, 80, 428-433.
[35] R.S.S. Wu, RSS; Lau, TC; Fung, WKM; Ko, PH; Leung, KMY. An “artificial
mussel” for monitoring heavy metals in marine environments. Environmental
Pollution, 2007, 145, 104-110.
[36] Fernandez, N; Fernandez-Boan, M; Verisimo, P; Freire, J. Assessing the spatial
variability, level and source of organic chemical contaminants in bivalve fishing
grounds on the Galician coast (NW Spain). Marine Pollution Bulletin, 2013, 74(1),
291-301.
[37] Bujan-Villar, MD; Barciela-Alonso, MC; Aboal-Somoza, M; Bermejo-Barrera, B.
Total antimony determination in Galician mussel samples (Mytilus
galloprovincialis) by slurry sampling ― graphite furnace atomic absorption
spectrometry. Preliminary results. Atomic Spectrometry, 2011, 32(1), 34-40.
[38] Morita, Y; Kobayashi, T; Kuroiwa, T; Narukawa, T. Study on simultaneous
speciation of arsenic and antimony by HPLC-ICP-MS. Talanta, 2007, 73, 81-86.
[39] Tariq, J; Jaffar, M; Ashraf, M; Moazzam, M. Heavy metal concentrations in fish,
shrimp, seaweed, sediment, and water from Arabian Sea, Pakistan. Marine Pollution
Bulletin, 1993, 26(11), 644-647.
[40] Mckie, JC. Determination of total tin and tributyltin in marine biological materials
by electrothermal atomic absorption spectrometry. Analytica Chimica Acta, 1987,
197(C), 303-308.
[41] Amer, HA; Shawky, S. Radiochemical neutron activation analysis of total trace tin
and determination of its differentiated compounds in aquatic biota. Radiochimica
Acta, 2002, 90(6), 351-355.
[42] Abollino, O; Aceto, M; La Gioia, C; Sarzanini, C; Mentasti, E. Voltammetric
determination and speciation of inorganic and organometallic tin. Electoanalysis,
2002, 14(15-16), 1090-1097.
[43] Sanchez-Rodas, D; Corns, WT; Chen, B; Stockwell, PB. Atomic fluorescence
spectrometry: a suitable detection technique in speciation studies for arsenic,
selenium, antimony and mercury. Journal of Analytical Atomic Spectrometry, 2010,
25(7), 933-946.
[44] Ferrarello, CN; Fernandez de la Campa, MDR; Muniz, CS; Sanz-Medel, A. Metal
distribution patterns in the mussel Mytilus edulis cytosols using size-exclusion
chromatography and double focusing ICP-MS detection. Analyst, 2000, 125(12),
2223-2229.
[57] Liaw, M-J; Jiang, S-J; Li, Y-C. Determination of mercury in fish samples by slurry
sampling electrothermal vaporization inductively coupled plasma mass
spectrometry. Spectrochimica Acta – Part B Atomic Spectroscopy, 1997, 52 (6 PART
B), 779-785.
[58] Luis, ES; Balagot, AH; Villaflor, AC; Sanchez, FC; Develles, EN. Pb, Cd and Hg
contents of bivalves collected during the different seasons of the year. Food
Chemistry, 1989, 32(4), 239-255.
[59] Bagheri, H; Afkhami, A; Khoshsafar, H; Rezaei, M; Shirzadmehr, A. Simultaneous
electrochemical determination of heavy metals using a triphenylphosphine/
MWCNTs composite carbon ionic liquid electrode. Sensors and Actuators B:
Chemical, 2013, 186, 451-460.
[60] Ensafi, AA; Allafchian, AR; Saraji, M; Farajmand, B. In-situ differential pulse
anodic stripping voltammetry combined with hollow fiber-based liquid-three phase
micro extraction for determination of mercury using Au-nanoparticles sol-gel
modified Pt-wire. Talanta, 2012, 99, 335-341.
[61] Mert, R; Alas, A; Bulut, S; Ozcan, MM. Determination of heavy metal contents in
some freshwater fishes. Environmental Monitoring and Assessment, 2014, 186,
8017-8022.
[62] Zrncic, S; Oraic, D; Celeta, M; Mihaljevic, Z; Zanella, D; Bilandzic, N.
Biomonitoring of heavy metals in fish from the Danube River. Environmental
Monitoring and Assessment, 2013, 185, 1189-1198.
[63] Ly, SY. Voltammetric assay of mercury ion in fish kidneys. Official Journal of
Korean Society of Toxicology, 2008, 24(1), 23-28.
[64] Radulescu, M-C; Danet, AF. Mercury determination in fish samples by
chronopotentiometric stripping analysis using gold electrodes prepared from
recordable CDs. Sensors, 2008, 8(11), 7157-7171.
[65] Augelli, MA; Munoz, RAA; Richter, EM; Cantagallo, MI; Angnes, L. Analytical
procedure for total mercuri determination in fishes and shrimps by
chronopotentiometric stripping analysis at gold film electrodes after microwave
digestion. Food Chemistry, 2007, 101, 579-584.
[66] Silva, NAF; Lopes, MI; Leitao, RAE; Silva, HFA; Matos, MJ. Simultaneous
monitoring of toxic metals on White Poplar (Populus) by SWASV. Journal of the
Brazilian Chemical Society, 2005, 16(6A), 1275-1282.
[67] Liang, N-L; Shi, J-B; He, B; Jiang, G-B; Yuan, C-G. Investigation of methylmercury
and total mercury contamination in mollusc samples collected from coastal sites
along the Chinese Bohai Sea. Journal of Agricultural and Food Chemistry, 2003, 51,
7373-7378.
[68] Yusof, AM; Rahman, NA; Wood, AKH. The accumulation and distribution of trace
metals in some localized marine species. Biological Trace Element Research, 1994,
43-45, 239-249.
[69] Park, KS; Kim, NB; Woo, HJ; Yoon, YY; Lee, KY. An environmental research on
trace elements in freshwater fish by neutron activation analysis. Biological Trace
Element Research, 1990, 26-27(1), 347-354.
[70] Frentiu, T; Butaciu, S; Ponta, M; Senila, M; Darvasi, E; Frentiu, M; Petreus, D.
Determination of total mercuri in fish tissue using a low-cost cold vapor capacitively
couplet plasma microtorch optical emission microspectrometer: comparison with
direct mercury determination by thermal decomposition atomic absorption
spectrometry. Food Analytical Methods, 2015, 8, 643-648.
[71] Torres, DP; Martins-Teixeira, MB; Silva, EF; Queiroz, HM. Method development
for the control determination of mercuri in sea food by solid-sampling thermal
decomposition amalgamation atomic absorption spectrometry (TDA AAS). Food
Additives and Contaminants – Part A Chemistry, Analysis, Control, Exposure and
Risk Assessment, 2012, 29(4), 625-632.
[72] Chen, X; Han, C; Cheng, H; Wang, Y; Liu, J; Xu, Z; Hu, L. Rapid speciation analysis
of mercury in seawater and marine fish by cation exchange chromatography
hyphenated with inductively coupled plasma mass spectrometry. Journal of
Chromatography A, 2013, 1314, 86-93.
[73] Chen, X; Han, C; Cheng, H; Liu, J; Xu, Z; Yin, X. Determination of mercurial
species in fish by inductively coupled plasma mass spectrometry with anion
exchange chromatographic separation. Analytica Chimica Acta, 2013, 796, 7-13.
[74] Zhang, Y; Adeloju, SB. Speciation of mercury in fish samples by flow injection
catalytic cold vapouratomic absorption spectrometry. Analytica Chimica Acta, 2012,
721, 22-27.
[75] CEE Regulation No. 2676/90, the European Community’s Official Journal, L. 272
dated 03/10/1990.
[76] CEE Directive 466/2001/CE and subsequent amendments.
[77] Organization International de la Vigne et du Vin (O.I.V.), decision ENO 13/2006.
[78] D.M. 29/09/1976 Caratteristiche e limiti di alcune sostanze contenute nei vini e negli
aceti. Gazzetta Ufficiale della Repubblica Italiana, No. 262, October 1, 1976.
[79] Chen, A-H; Chen, S-L; Wu, Y-H; Shao, Y; Zhang, C-L. Direct and sequential
determination of six metal elements in cooking wine by HR-CS GFAAS. Advanced
Materials Research, 2014, 1033-1034, 658-662.
[80] Santos, S; Lapa, N; Alves, A; Morais, J; Mendes, B. Analytical methods and
validation for determining trace elements in red wines. Journal of Environmental
Science and Health, Part B, 2013, 48, 364-375.
[81] Tariba, B. Metals in wine ― Impact on wine quality and health outcomes. Biological
Trace Element Research, 2011, 144, 143-156.
[82] Grindlay, G; Mora, J; Gras, L; de Loos-Vollebregt, MTC. Atomic spectrometry
methods for wine analysis: A critical evaluation and discussion of recent
applications. Analytica Chimica Acta, 2011, 691, 18-32.
[96] Shen, F; Wu, J; Ying, Y; Li, B; Jiang, T. Differentiation of Chinese rice wines from
different wineries based on mineral elemental fingerprinting. Food Chemistry, 2013,
141, 4026-4030.
[97] Schiavo, D; Neira JY; Nòbrega, JA. Direct determination of Cd, Cu and Pb in wines
and grape juices by thermospray flame fornace atomic absorption spectrometry.
Talanta, 2008, 76, 1113-1118.
[98] Lara, R; Cerutti, S; Salonia, JA; Olsina, RA; Martinez, LD. Trace element
determination of Argentine wines using ETAAS and USN-ICP-OES. Food and
Chemical Toxicology, 2005, 43, 293-297.
[99] Reilly, C., 2002. Metal Contamination of Food – Its Significance for Food Quality
and Human Health. Oxford, U.K.: Blackwell Science Ltd.
[100] Ebdon, L., Pitts, L., Cornelis, R. Eds., 2002. Trace Element Speciation for
Environment, Food and Health. Cambridge, U.K.: Royal Society of Chemistry.
[101] Merian, E., Anke, M., Ihnat, M., Stoeppler, M., 2004. Elements and their
Compounds in the Environment – Occurrence, Analysis and Biological Relevance.
Weinheim (Germany): Wiley-VCH Verlag GmbH & Co. KGaA.
[102] Intawongse, M., Dean, J.R., 2006. Uptake of heavy metals by vegetable plants grown
on contaminated soil and their bioavailability in the human gastrointestinal tract.
Food Addit. Contam. A 23, 36-48.
[103] Caroli S., 2007. The Determination of Chemical Elements in Food: Applications for
Atomic and Mass Spectrometry. Hoboken, New Jersey, USA: John Wiley & Sons
Inc.
[104] Ferré-Huguet, N., Martì-Cid, R., Schuhmacher, M., Domingo, J.L., 2008. Risk
assessment of metals from consuming vegetables, fruits and rice grown on soils
irrigated with waters of the Ebro river in Catalonia, Spain. Biol. Trace Elem. Res.
123, 66-79.
[105] Locatelli, C., Melucci, D., Locatelli M., 2014. Toxic metals in herbal medicines. A
review., Current Bioactive Compounds, 10, 181 – 188.
[106] Karak, T., Bhagat, R.M., 2010. Trace elements in tea leaves, made tea and tea
infusion: a review. Food Res. Int. 43, 2234-2252.
[107] CEE Directive 1881/2006/CE.
[108] Ning, P., Gong, C., Zhang, Y., Guo, K., Bai, J., 2011. Lead, cadmium, arsenic,
mercury and copper levels in Chinese Yunnan Pu’er tea. Food Addit. Contam. B 4,
28-33.
[109] Salahinejad, M., Aflaki, F., 2010. Toxic and essential elements content of black tea
leaves and their tea infusions consumed in Iran. Biol. Trace Elem. Res. 134, 109-
117.
[110] Gebretsadik, D.W., Chandravanshi, B.S., 2010. Levels of metals in commercially
available Ethiopian black teas and their infusions. Bull. Chem. Soc. Ethiop. 24(3),
339-349.
[111] Bobrowska-Grzesik, E., Jakobik-Kolon, A., 2008. Leaching of cadmium and lead
from dried fruits and fruit teas to infusions and decoctions. J. Food Comp. Anal. 21,
326-331.
[112] Yemane M., Chandravanshi, B.S., Wondimu, T., 2008. Levels of essential and non-
essential metals in leaves of the tea plant (Camellia sinensis L.) and soil of
Wushwush farms, Ethiopia. Food Chem. 107, 1236-1243.
[113] Soylak, M., Tuzen, M., Souza, A.S., das Gracias, M., Korn, A., Costa Ferreira, S.L.,
2007. Optimization of microwave assisted digestion procedure for the determination
of zinc, copper and nickel in tea samples employing flame atomic absorption
spectrometry. J. Hazard. Mater. 129, 264-268.
[114] Mokgalaka, N.S., McCrindle, R.I., Botha, B.M., 2004. Multielement analysis of tea
leaves by inductively coupled plasma optical emission spectroscopy using slurry
nebulisation. J. Anal. At. Spectrom. 19, 1375-1378.
[115] Arain, G.M., Khuhawar, M.Y., 2008. Liquid chromatographic analysis of
mercury(II) and cadmium(II) using dimethylglyoxal bis-(4-phenyl-3-
thiosemicarbazone) as derivating reagent. Acta Chrom. 20(1), 25-41.
[116] Fung, Y.S., Lau, K.M., 2006. Separation and determination of cations in beverage
products by capillary zone electrophoresis. J. Chrom. A 1118(1), 144-150.
[117] Feng, H., Wang, T., Li, S.F.Y., 2003. Sensitive determination of trace-metal
elements in tea with capillary electrophoresis by using chelating agent 4-(2-
pyridylazo)resorcinol (PAR). Food Chem. 81, 607-611.
[118] Shi, Z., Zhang, H., Fu, C., 2000. Separation and determination of trace C(II), Zn(II)
and Pb(II) as their tetraphenylporphirine chelates by reversed-phase high
performance liquid chromatography. J. Liq. Chromatogr. R. T. 23(2), 209-222.
[119] Ichinoki, S., Hongo, N., Yamazaki, M., 1987. Simultaneous determination of heavy
metals in Chlorella and tea leaves by high-performance liquid chromatography. J.
Agric. Food Chem. 35, 448-451.
[120] Zhang, H.-Q., Ni, B.-F., Tian, W.-Z., Zhang, G.-Y., Huang D.-H., Liu, C.-X., Xiao,
C.-J., Sun, H.-C., Zhao, C.-J., 2011. Study on essential and toxic elements intake
from drinking of Chinese tea. J. Radioanal. Nucl. Chem. 287, 887-892.
[121] Banks, C.E., Compton, R.G., 2003. Ultrasonically enhanced voltammetric analysis
and applications: an overview. Electroanalysis 15(5-6), 329-346.
[122] Zhang, Z.-Q., Liu, H., Zhang, H., Li, Y.-F., 1996. Simultaneous cathodic
voltammetric determination of mercury, cobalt, nickel and palladium by mixed
binder carbon paste electrode containing dimethylglyoxime. Anal. Chim. Acta 333,
119-124.
[123] Hatle, M., 1987. Determination of mercury by differential pulse anodic stripping
voltammetry with various working electrodes. Application to the analysis of natural
water sediments. Talanta 34, 1001-1007.
[124] Narin, I., Colak, H., Torkoglu, O., Soylak, M., Dogan, M., 2004. Heavy metals in
black tea samples produced in Turkey. Bull. Environ. Contam. Toxicol. 72, 844-849.
[125] Ferreira, S.L.C., Korn, M.G.A., Ferreira, H.S., da Silva, E.G.P., Araujo, R.G.O.,
Souza, A.S., Macedo, S.M., Lima, D.C., de Jesus, R.M., Amorim, F.A.C., Bousque-
Sendra, J.M., 2007 Application of multivariate techniques in optimization of
spectroanalytical methods. Appl. Spectr. Rev. 42, 475-491.
[126] Santelli, R.E., De Almeida Bezerra, M., De Santana, O.D., Cassella, R.J., Ferreira,
S.L.C., 2006. Multivariate technique for optimization of digestion procedure by
focussed microwave system for the determination of Mn, Zn and Fe in food samples
using FAAS. Talanta 68, 1083-1088.
[127] Almeida, A.A., Lima, J.L.F.C., 2001. Optimized conditions and analytical
performance for the determination of Cu in serum and urine samples using a single
GFAAS procedure. Atom. Spectr. 22, 324-330.
[128] Melucci, D., Locatelli, C., 2012. Multivariate calibration in differential pulse
stripping voltammetry using a home-made carbon-nanotubes paste electrode. J.
Electroanal. Chem. 675, 25-31.
Chapter 9
ABSTRACT
Undoubtedly, the interactions between matter and heat have changed both the world
and humanity’s evolution. From the discovery of fire to our current reality, passing through
all of scientific history, heat control has allowed to protect ourselves, process food,
manipulate metals, change and sterilize materials, accelerate synthetic pathways or even
extract desired information about certain materials, such as thermal stability, melting points
and purity, among others. In this context, this chapter aims to present a concise but
complete briefing on the early history and fundamental concepts concerning thermal
analysis, beginning at thermogravimetric measurements (TGA), differential thermal
analysis (DTA), high-resolution thermogravimetric analysis (Hi-Res TGA®) and evolved
INTRODUCTION
This chapter aims to present briefly but sufficiently complete and clear the main
aspects involved from the preparation and planning to the interpretation of thermal
experiments with analytical purposes; it is neither a reference nor a lengthy review, but a
toolbox for those who intend to take part in this world of analytical chemistry and materials
science. For this reason, the authors will constantly cite reference books and recent
reference works, taking into account that readers are, sometimes, being exposed to this
information for the first time, and thus, in many points, physical parameters, “do´s and do
not´s” and tips will be given.
Among all well-stablished thermal techniques, one of the most applied and popular is
thermogavimetry (TGA), in which mass changes are continuously monitored as a function
of a previously programed thermal path. In other words, this technique allows for the
assessment of the thermal behavior of diverse samples according to modifications caused
by heating (Willard & Dean, 1974; Charsley & Warrington, 1991; Haines, 1995; Ionashiro
& Giolito, 2004; Azevedo & Mothé, 2009).
Although this technique may seem like an analytical application of calcination
processes, it goes even further since, by applying TGA, a controlled heating rate can be
imposed on the samples, while at the same time offering a controlled chemical
environment. Thus, this technique allows for a suitable resolution for analytical purposes,
avoiding overlapping of successive thermal events. In this way, this technique is one of the
most versatile thermal techniques, permitting studies on thermal stability, physical and
chemical processes, cure of polymers, dehydration, crosslinking processes and oxidation-
reduction reactions, among others. In short, every mass change (loss or even gain) related
to heat transfer can be evaluated by TGA under certain conditions.
The main questions are: “How to adapt and synchronize mass and temperature
measurements? How to optimize experimental conditions in order to reach the best possible
resolution and accuracy? And, finally, how to interpret the data?”
For didactic purposes, let us now imagine a 100 g pure chocolate bar: If a person holds
this bar for a while, it is possible to conclude that, depending on the temperature of the
person’s hand, the chocolate will certainly melt but, for obvious reasons, its mass will not
change, only its physical state. On the other hand, if this bar is exposed to a 500oC furnace,
it will certainly decompose, but what happens during the heat transfer to this bar? Does
every particle burn at the same time? What if the “bar” consisted of just 10 mg of finely
pulverized chocolate?
Clearly not only the mass but also the particle size as well as the thermal conductivity
interfere on the heat flow and, thus, sample changes.
According to the heating rate (β, °C min-1) one can imagine that, at a specific moment,
external parts of the bar will melt while its bulk is still solid. So, in case of fast complete
melt is desirable, some variables can be manipulated, such as sample mass/weight (m/wt.),
aggregation degree (powder, crystals, fragments) and particle size (homogeneous,
heterogeneous), as well as heating rate (β).
The same idea can be extrapolated to a “more analytical” example. For instance, let us
consider the calcination of pure hydrated calcium oxalate, often use to calibrate
thermoanalytical systems for routine analyses due to its clear and well-established multi-
step thermal path:
Such thermal path can be didactically exploited in order to extend our talk and dig a
deeper understanding; let us this way consider all these three decomposition steps in a
conventional thermogravimetric experiment. Profiles will present sufficiently resolved
peaks according to the temperature.
In order to obtain good analytical results, which allows for better conclusions, the
combination of many different aspects regarding not only the sample but also the
experimental conditions becomes a priority.
Thermogravimetric experiments are often carried out in thermobalances or
thermogravimetric analyzers that combine the capability of null-type analytical balances
and highly controlled furnaces. In this way, it becomes possible to observe mass changes
under a certain temperature and chemical environment by comparing the sample to a
reference (Figure 1).
Figure 1. The thermogravimetric experiment – sample and reference under the same heating program
and chemical environment, showing weight loss while null-type mechanisms monitor weight changes.
Figure 3. Thermogravimetric data and its respective derivative curve (often used to plan high-resolution
experiments) for the thermal paths of calcium oxalate monohydrate.
Modern equipments are characterized by many different layouts, but all comprise
essentially the same elements: an analytical balance, a furnace, two thermocouples,
reference and sample containers and a controlling computer (Figure 2).
This way, looking back to our previously cited example, one can assess calcium oxalate
thermal paths by using a simple thermal analyzer, obtaining (Figure 3).
It is certainly very important to define the purposes of the experiment to better plan
and define the optimum conditions: what is the sample? Is it a powder? A pure solid? Ionic
or molecular material? A composite material? A solid mixture? Is it widely available? And,
finally, what kind of information am I looking for? Wide is the myriad of potential variables
to be considered!
In our first example, sample consisted of a pure substance, since the purpose is to
calibrate the system in both temperature and mass changes detection.
Weight changes are, in general, related to mass losses, but in some rare cases, it is also
possible to evidence mass gain (for example, when heating copper oxalate under oxygen-
containing atmosphere). Such changes can be classified into chemical and physical
changes. The first group comprises thermal decomposition, oxidative or reductive
processes, dehydration, cure of polymers and crosslinking reactions, among others. The
physical phenomena, on the other hand, include sorption and desorption,
evaporation/condensation, sublimation processes, and so on.
Considering these changes, let us now discuss the main variables involved. In general,
TGA experiments use small solid samples, of masses from 1 to 100 mg (frequently ~ 5
mg), depending on the equipment and the purpose of the analysis. Such small amounts
allow a better heat flow, turning steps and peaks more defined and accurate, our “chocolate
bar” will melt at the same time in all points.
As mentioned, smaller samples often generate well-defined thermogravimetric steps.
However, it is possible to lose information regarding minor components, limited to the
balance capabilities. Higher masses can offer many additional thermal steps and allow for
the identification of intermediate species, but can also take longer to achieve thermal
equilibrium, requiring particular precautions.
The sample weight must be carefully assessed according to both equipment limitations
and the sample characteristics. Still with regard to sample sizes, it is highly suggested the
use 60 to 80% of the mass capability of the system (specially the sample holder),
considering that some materials can expand or even decrepitate during heating processes.
Particle sizes and aggregation degrees are also very important variables. In general,
smaller particles offer high resolution, since they quickly reach thermal equilibrium and
allow easier vapor and gas releases. Larger particles, on the other hand, tend to hold
evolved gases, requiring higher energy levels to allow such a release. Although intuitive,
all these variables must obey physical properties, such as the thermal conductivity of the
material.
Figure 5. Schemes presenting the positions and kinds of furnace (A) regarding thermocouples (B) and
balance arms, note in (C) the reference and sample holders.
Some pans, such as those made of aluminum, are usually disposable. However,
alumina, platinum and many other materials are expensive and can easily be clean and
reused. In this case, the thermal treatment is sometimes enough, but when it is not, the use
of diluted acidic solutions or even solvents may be applied, depending on the assessed
sample and its residue.
Since sample is suitably placed into the pan, the heating program must be chosen. A
furnace, which can be designed in three different positions (above, below and around the
side arm of the balance, as depicted in Figure 5), carries out the heating process. In general,
horizontal furnaces are easier to clean, while above-furnaces tend to more accurate since
minimize convective flows inside the system.
Furnaces are responsible for the thermal program according to the computer, this way,
to reach the best possible heat change between furnace and sample, avoiding heat “losses,”
such part is filled by static or dynamic (renewable) atmosphere, being such environment
chemically inert or reactive. This gaseous environment must allow suitable synchronism
heat program, furnace and sample capabilities.
Dynamic atmospheres are in general used in experiments that require environmental
renewal, or even when one expects to observe solid-gas reactions between sample and
atmosphere components, sometimes the use of synthetic air can simulate specific
situations, thermal oxidative processes can be even observed when heating the sample
under helium-oxygen flow. Among various inert gases used, most frequently used
examples are nitrogen, argon, helium etc.
Sometimes, in order to reach a better understanding, it is of prior importance the
assessment of different purge gases composition and flow (mL min-1). Considering our
example, it becomes easy to note that, according com Le Chatelier Principles, in case the
purge gas contains carbon monoxide and carbon dioxide, second and third steps will be
note that it is also possible to use more than one simultaneously for a wide calibration
(Willard & Dean, 1974; Charsley & Warrington, 1991).
The inherent sensitivity of DTA can indicate any physical or chemical event that causes
enthalpy variations. All DTA interpretations are related to exothermic or endothermic
process. The peak numbers, position, forms, intensity and regions are characteristic of
certain temperature-induced phenomena, enhancing qualitative and quantitative
conclusions regarding the resulting thermogram.
Clays, minerals and oxides were the first materials investigated by thermal analysis.
With the introduction of modern materials science, polymer materials have assumed a key-
role and become the most studied material by thermal techniques. Their multi-applicability
is a huge part of the development of new polymer materials. Undoubtedly, specific
characterizations of these novel products are necessary, including the investigation of the
thermal resistances of the different materials and their respective thermal events
(degradation, glass transition, decomposition), that may compromise their properties
(Skoog & Holler, 2002).
In general, a DTA equipment is composed of a furnace, responsible for the temperature
variations; crucibles, two identical containers that hold the sample and the reference,
respectively; a thermocouple, responsible for monitoring of the temperature difference (ΔT
= Tref – Tsam) was processed and plotted versus temperature or time. Figure 6 displays a
general scheme of a Differential Thermal Analyzer and its respective components.
An electric DTA furnace can function in temperatures ranging from 80 to 3070 K,
using either electromagnetic resistance or radiation as a heat source. The heating geometry
should be symmetric, in a way that both the sample and the reference materials receive the
same amount of heat. Table 4 exemplifies some types of thermal sources (Willard & Dean,
1974; Haines, 1995).
Figure 6. General scheme representative of a Differential Thermal Analyzer (DTA) and its respective
components.
Thermocouples are used as temperature monitoring devices with regard to the electric
potential difference between the metals that comprise a thermocouple. Table 4 also displays
combinations of metals that constitute different thermocouples and their respective
maximum temperatures of application in DTA apparatus (Willard & Dean, 1974; Haines,
1995).
The same caution taken while planning TGA experiments must be followed during
DTA analysis. The choice of atmosphere of a DTA furnace is an important condition,
considering that great part of the reactions that occur in the chamber are classified as
decomposition reactions, producing gases as the product. The atmosphere can frequently
be dynamic (N2, synthetic air) or self-generated (a product of sample decomposition such
as CO and CO2).
Figure 7. Hypothetical thermoanalytical profile obtained from a Differential Thermal Analyzer (DTA).
A hypothetical highly resolved DTA profile is depicted in Figure 7, where some events
can be highlighted. When heated at a certain rate, the sample reaches a temperature in
which event a is observed. Event a is a slight displacement of the baseline, that often points
to a change in the specific heat of the material, typical of glass transition. Event b is an
endothermic peak, intense and large, indicative of a dehydration process. Event c, also
endothermic, is characterized by physical changes, such as melting, solid-state transitions
and crystalline arrangement. Peak c is an exothermic event, of longer duration and higher
intensity. Such findings are indicative of chemical processes, including chemical reactions,
polymerization, decomposition and other compositional changes. The final part of the
process comprises a narrow exothermic peak that indicates a crystallization process.
Each signal generated in a DTA curve can be interpreted, by following the Equation 1.
The areas (A) of the peaks point to a proportionality of the evaluated sample (wt.sam), with
a certain enthalpy involved in the thermic event (∆H), a constant thermal conductivity
related to the sample (G) and a constant (k) depending on the calibration (Skoog & Holler,
2002).
𝐴 = −𝑘𝐺𝑊𝑡𝑠𝑎𝑚 ∆𝐻 Eq. 1
The DTA technique is commonly used as a noble resource to further characterize TGA
data, which indicates mass variations as a function of temperature changes. Taking into
account that all mass gain or loss processes are followed by absorbance or loss of energy,
the DTA technique is an important complementary tool for TGA analyses. However, not
all energy changes caused by temperature result in mass modifications.
For example, Chowdhury et al. (2016) conducted a study wherein the TGA and the
DTA techniques were used together to characterize silver-doped TiO2 nanoparticle
synthesis. The material was monitored in temperatures ranging from 273 to 1273 K, under
increasing temperature rates of 10 K min-1 in an air atmosphere. The authors observed the
thermal events of dehydration of absorbed water (405 K) and hydroxyl groups (558 K),
and also indicate that three thermogram peaks, two exothermic (608 e 737 K) and one
endothermic (688 K), are indicative of surface active agent decomposition progress and
release of inorganic radicals, both TiO2 nanoparticle synthesis precursors. The conclusions
thus extracted from thermoanalytical curves are important arguments regarding the TiO 2
nanoparticle characterization (Chowdhury et al. 2016).
DTA experiments are also capable of provide fingerprinting of certain materials and
lead to better planning of high-resolution experiments.
3. HI-RES TGA®:
A DEEPER LOOK INTO THERMOCHEMICAL PATHWAYS
Another approach available is the Constant Reaction Rate. The heating rate is
controlled to keep a preselected weight loss rate (% of weight change). This means that
whenever the % of weight change exceeds the limit, the heating rate will be reduced, even
cooling if necessary, to maintain the threshold. Another implication of that configuration
is that whenever the weight change is lower than the set limit, the heating rate will be
increased. This algorithm is used to control reaction rate. Another case is when the sample
presents a constant background weight change. The threshold can be configured to keep
the maximum heating rate paired with the constant background, and will be reduced when
other events happen.
The Stepwise Isothermal program uses a constant heating rate until an event begins, or
an amount of weight is lost, and then changes the rate to maintain the temperature constant
until the event finishes. The sequence of constant heating and isothermal control is repeated
for every event involving weight change in the experiment. This algorithm is very useful
for very close or even overlapping transitions. The biggest disadvantage of this approach
is the need for several experiments in order to, properly, optimize the settings, and the
relative slow heating rate needed to avoid overshooting any transition will cause an
increase in experimental time.
In general, one must use DTG curves (as exemplified by Figure 3) obtained from a
preliminary TGA analysis at 20ºC min-1. From such data it must be extracted the maximum
point of weight loss, in order to define the entrance and exit threshold values (entrance
from 0.1 to 0.2 of maximum loss, and exit 1/5 of entrance threshold) (TA - Instruments
TN-40).
Figure 8. TGA (solid line) and DTG (dashed line) of copper sulfate pentahydrate; on the left is find the
conventional TGA profile while, on the right see a high resolution TGA under dynamic rate. The high-
res TGA® was able solve the overlapping events to show each hydration water loss (adapted from TA-
Instruments TA-023 with suitable permission).
The last available algorithm for Hi-Res TGA® is the conventional TGA program,
where a constant heating rate is set to go throughout the entire run. This approach is
suggested to follow the same care of regular TGA parameters for a valid gain in resolution.
Other that the four algorithms of Hi-Res TGA® an ever higher gain on resolution can
be achieved with the use of modulated heating rates. The modulated approach uses a
sinusoidal modulation on the heating rate to improve the obtained results.
Some examples of Hi-Res TGA® applications are kinetic studies and activation energy
calculation, quantifications, degree of crystallinity and thermal decomposition studies.
Here are some examples of these applications found in the literature.
Using dynamic and constant heating rate programs, Brožek et al. (2007) studied the
thermal behavior and composition of block copolymers polyamide-6-polybutadiene,
comparing Hi-Res TGA® with conventional TGA procedures. When using conventional
TGA a reduced pressure condition was required, in order to allow the determination of the
percentage of each component of the block. Without this condition, the events were not
well resolved. On the other hand, by using Hi-Res TGA®, authors reached very similar
results in a reduced experimental time and without the need to control the pressure. Other
than the quantification of each component, the thermal behavior of each component of the
block was studied, and polyamide showed lower thermal stability than polybutadiene. The
use of Hi-Res TGA® presented a simpler and faster experimental procedure to achieve
similar results of TGA under reduced pressure.
Exploiting the enhanced resolution provided by Hi-Res TGA®, Arthanareeswari et al.
(2013) studied the hydration waters and the degree of crystallinity of pantoprazole sodium
sesquihydrate, comparing the results with an analysis by Karl Fisher titration and powder
X-ray diffraction. Karl Fisher titration was used to measure the water content of the sample
while X-ray diffraction to determinate the degree of crystallinity.
The water content found by Karl Fisher was 6.65% (w/w). Using x-ray diffraction it
was possible to calculate the degree of crystallinity of the sample, which was 82.4%. Using
Hi-Res TGA®, it was possible to determinate the water content of the sample, with the
ability to discern the water bounded to the crystal from the total water content, and the
degree of crystallinity in a single experiment. Appling the dynamic rate algorithm and N2
inert atmosphere, the total water content found was 6.534% (w/w), a result that agrees with
the Karl Fisher method, with 5.288% (w/w) being bound water.
The degree of crystallinity found by Hi-Res TGA® was 84.7%, which is comparable
with the x-ray diffraction obtained value. The degree of crystallinity is an important
characteristic from a pharmaceutical point of view, due to its influence on the drug
performance. The authors showed a fast and simple way to study the properties of
crystalline drugs using a single experiment.
Kubo & Kadla (2008) compared conventional TGA and two different procedures using
Hi-Res TGA® (with dynamic heating rate and Hi-Res modulated TGA® (MTGA) in the
study of thermal decomposition and pyrolysis kinetics of lignin, a wood component.
Conventional TGA results have shown a single constant weight loss profile, between 250
and 600ºC. The DTG curve shown overlapping events based on the type of wood
originating the lignin, but without the ability to resolve the overlapping events.
The dynamic rate Hi-Res TGA® approach enhanced the resolution for all studied
samples, but was unable to separate the events for one of the samples. This showed that the
Hi-Res TGA was able to improve the resolution, but was not powerful enough for all
studied samples. Using Hi-Res MTGA an even greater gain on resolution was achieved
and all events were separated on the DTG curves.
Other than the thermal profile of lignin, the activation energy (Ea) was also calculated
with a single experimental run. The obtained values of Ea in MTGA were higher than
previously determined with conventional TGA. This difference was caused by the variable
heating rate applied. To solve this issue the MTGA was performed using constant heating
rates. Using this approach was possible to obtain Ea value accordingly to previous studies.
A thermochemical and kinetic study of pectin was made using Thermogravimetric
Analysis–Differential Scanning Calorimetry/Fourier Transform Infrared spectroscopy
(TGA-DSC/FTIR) and Hi-Res modulated TGA® by Aburto et al. (2015). Hi-Res MTGA®
was used as a fast and precise tool for obtaining more information about the thermal
degradation kinetics of pectin.
Based on the results, was possible to determinate the Ea value and pre-exponential
factor (Z) value as a function of temperature and from the DTGA curves the Ea as a
function of the transformation degree. The results are capable of revealing the complexity
of the reaction, that involves multiple and parallel processes during the pectin degradation.
The study of the complex decomposition mechanism and kinetics of different kinds of the
sample was possible with use of Hi-Res TGA® alone or in combination with other
techniques.
Reddy et al. (2006) proposed a quantification method of pantoprazole using Hi-Res
TGA®. Others have compared two different procedures involving Hi-Res TGA® (with
dynamic rate and Hi-Res modulated TGA®). The thermal decomposition of both the
monohydrate and seshydrate compounds were studied via Hi-Res TGA® and modulated
Hi-Res TGA®, obtaining similar results, although the second approach has shown a higher
resolution and sharper peaks on the DTG curves.
Based on these results, Hi-Res MTGA® was used for the quantifications. The accuracy
and reproducibility were assessed showing recoveries between 96 and 106%. The results
have shown that the accuracy, reproducibility and sensitivity of the Hi-Res modulated
TGA® method were comparable with the well-established FT-IR method, reaching
recoveries between 95 and 105%. The limit of detection in the Hi-Res TGA method was
1.0% and for FR-IR was 2.0%. Looking at these results it is possible to conclude that
applying Hi-Res TGA® for the quantification of hydrate polymorphic forms is not only
valid, but a simpler and better technique.
The detection or analysis of the gases evolved during a chemical reaction, as a function
of temperature, comprises the thermal analysis techniques of evolved gas detection (EGD)
and evolved gas analysis (EGA), respectively. These techniques, which paralleled the
development of modern thermal analysis instrumentation, are currently used to solve many
types of problems in thermal analysis. In recent years, coupled, or hyphenated, techniques
have become popular, and have been successfully applied to the solution of many analytical
problems, since different properties may be measured at the same time.
TGA gives characteristic information about the composition of the measured sample,
in particular the amounts of the various components and their thermal behavior. The
identification of gases released directly from the sample or during thermal treatment cannot
be performed only by thermal analysis, but coupling a spectroscopic method, such as
Fourier Transform Infrared (FTIR) spectroscopy, provides a simple and accurate technique
for identifying the evolved gases from a TGA experiment. The spectra can be used to
characterize the substance or class of substances through spectral interpretation and
comparison with database spectra (Gabbott, 2008).
In addition, this provides information on sample characteristics, such as thermal
stability or decomposition pathways, and can thereby elucidate and solve specific
analytical problems.
IR spectroscopy is a classical technique, which depends on the interaction of infrared
radiation with the vibrating dipole moments of molecules. In the TGA-FTIR technique, the
absorption bands of each spectrum are usually simultaneously integrated over the entire
spectral range or over characteristic spectral regions. The intensity is presented as a
function of time, as so-called Gram–Schmidt curves or chemigrams.
gases. In order to minimize the risk of condensation, the equipment coupling interface is
heated by applying a constant voltage. Optionally, a temperature control system is available
(recommended for condensable gases). The maximum temperature of the entire gas path is
set to 200ºC.
These hyphenated techniques are used in research and development, in quality control,
and to investigate material failure. Typical applications are as follows:
investigated using FTIR coupled with TG. The thermal degradation studies demonstrated
that CO2 and amine are the major degradation products evolved during decomposition.
Figure 9. (a) Simultaneous TG–DTA curves of Mg(Mandelate)2 compounds. (b) IR spectra of the
gaseous products evolved during the thermal decomposition of a magnesium compound.
Worzakowskaa & Scigalski (2012) studied the synthesis, thermal behavior and
characterization of the decomposition products of some linear geranyl diesters. The thermal
behavior of prepared compounds was studied by the TG/DSC/FTIR coupled method. The
analysis of the gases evolved during heating of diesters in an inert atmosphere indicated
asymmetrical disruption of their bonds, resulting in production of a mixture of geraniol
(acyclic and alicyclic monoterpene hydrocarbons) derivatives, like myrcene, ocimene, or
limonene, as main decomposition products.
Worzakowskaa & Scigalski (2014) studied the thermal behavior of cinnamyl diesters
in an inert atmosphere by the TG/FTIR/QMS coupled method. The results confirmed that
diesters decomposed through the main two-step process. The first (and main)
decomposition step was characterized by an asymmetric peak, not well-separated, which
indicates multi-step processes during the pyrolytic cracking of these compounds, observed
between ca. 200 and 460–525°C, with significant mass losses of 85.7% to 90.9%. The
TG/FTIR/QMS and additional DSC analysis indicated that, during the first decomposition
step of ciselimination reactions, partial decarboxylation of formed dicarboxylic acids,
condensation processes of two carboxyl groups and polymerization processes of allene
(benzene-1,2-propadienyl) in a gaseous phase were expected. Consequently, the
production of CO2, H2O and CO as main gases and small amounts of organic compounds,
such as carboxylic acids, cyclic ketones, aldehydes, allene, styrene, ethylbenzene, toluene
and benzene, are observed.
The second decomposition step occurred at higher temperatures (above 460–525°C)
with mass losses from 1.6% to 7.8%. The release of, mainly, CO2 and H2O as gaseous
products was described. This was probably the result of the carbonization process of
polymeric residues formed after the first decomposition step.
Marinho et al. (2015), while studying semi-hard manufactured cheeses, assessed their
lipid fractions by coupled TG-DSC and FTIR techniques, among others. The samples were
divided into three portions: (a) cheeses 24 h after manufacture; (b) cheeses after 60 days of
ripening; (c) cheeses covered with lard and rosemary after 60 days of ripening. The fat
from samples treated with rosemary leaves showed higher decomposition temperatures
(TG), higher resistance to oxidation processes (Tp = 310.2°C) and enthalpy (1228.5 J g−1).
FTIR spectroscopy of the decomposition gaseous products obtained simultaneously with
TG–DSC curves revealed that the main bands could be attributed to –C-H, –COOH
stretching with the formation of CO, CO2 and tetradecanoicacid-7-oxo-methylester
compounds.
In recent years, concern about environmental pollution has increased. Thus, the use of
coupled techniques is of great importance for characterizing the gases released in the
degradation of compounds such as drugs and biological matrices. Some works are briefly
listed below.
Yang et al., (2015) investigated the combustion characteristics of three kinds of
antibiotic residue (AR) materials, investigated by TG–FTIR and others techniques. The TG
results indicated that AR combustion involved three stages. The FTIR spectra identified
evolving gaseous products as CO2, CH4, HCNO, NH3, HCN and NO. A possible pathway
for AR combustion was also tentatively presented.
The management of drug disposal (expired or not) is important in order to minimize
the possibility of environmental pollution. Colman et al. (2016) analyzed aceclofenac
(purity = 99.77% and m.p. = 151.63°C) analyzed by TG-DSC-FTIR in a nitrogen
atmosphere. The TG-DSC curves indicated two main stages of decomposition, with loss of
organic matter, generating carbonaceous derivatives as the final residue. The use of
coupled FTIR suggested the loss of CO, CO2 and/or 2-chloro-propanoic acid/2-chloro-
butanoic acid.
Antibiotic mycelial fermentation residues (AMFRs), which are emerging as solid
pollutants, have been recognized as hazardous waste in China since 2008. Nitrogen (N), an
element present in the environment, is largely retained in AMFR samples derived from
fermentation substrates. Zhu et al. (2016) demonstrated the conversion of N during the
pyrolysis of AMFRs, by using online TG-FTIR-MS (Thermogravimetry-Fourier transform
infrared-Mass spectrometry) technology. In the AMFR sample, organic amine-N, pyrrolic-
N, protein-N, pyridinic-N, were the main N-containing species. The TG-FTIR-MS results
indicated that NH3 and HCN were the main gaseous species, and that their contents were
closely related to amine-N and protein-N, and pyrrolic-N and pyridinic-N content of
AMFRs, respectively.
In their work, Peng et al. (2016) demonstrated the kinetic behavior and evolution
characteristics of the gaseous products of microalgae (MA) and textile dyeing sludge
(TDS) blends during co-pyrolysis by TG–FTIR. The TDS was blended with MA in the
range of 10–90 wt.%, and then heated from 105°C to 900°C at 10, 20 and 40°C min-1 under
an N2 atmosphere, at a flow rate of 80 mL min-1. The initial decomposition temperature of
TDS was lower than for MA, but the pyrolysis intensity of MA was higher than in TDS.
The co-pyrolysis of MA and TDS could improve their pyrolysis performance. The
synergistic interaction between MA and TDS was mainly due to the flowing and sticky
bio-oil and alkali metals catalytic effect. CO2, H2O, CH4, CO, HNCO, NH3, HCN, SO2,
CH, CO and CO groups were the main gaseous products observed from the FTIR spectrums
during co-pyrolysis. The origin of these functional groups was analyzed combined with the
decomposition order of the MA and TDS components.
The evolution of the gaseous products was consistent with the weight loss of the blends
during co-pyrolysis. The yields of CO2, CH, CO, CO, CO, HNCO, NH3, HCN and SO2 did
not monotonically change with the increasing blending ratio of MA, attributed to the
synergistic interaction between MA and TDS.
Another practical application is the assessment of thermal stability of a popular
medicine: acetylsalicylic acid. Such medicine easily loses its acetic acid and becomes
salicylic acid, which then becomes phenol when heated. In this context, it is clear that the
combination of complementary TG-FTIR techniques offer a versatile and complete
analytical system for the detection and identification of gases evolved during thermal
decomposition processes. Decomposition mechanisms can also be elucidated. Finally, TG-
FTIR techniques can provide precise time- and/or temperature-dependent gas evolution
information. Figure 10 displays the thermal decomposition mechanism of acetyl salicylic
acid:
Another interesting field of application is the food industry. Most foodstuff are
subjected to variations in their temperature during production, transport, storage,
preparation and consumption. Processes such as pasteurization, sterilization, liofilization,
and others, must be carefully assessed in order to allow a better choice of temperature and
chemical programs.
Undesirable temperature changes can cause alterations in the physical and chemical
properties of food components, which influence the overall properties of the final product,
e.g., taste, appearance, texture and stability.
In the investigation of foods by thermal analysis and calorimetry techniques, many
physicochemical effects can be observed in the temperature range between –50 and 300°C.
These thermal phenomena may be either endothermic, such as melting, gelatinization,
denaturation, evaporation or exothermic, such as crystallization, oxidation, fermentation.
Glass transitions are observed as a shift in the baseline; this information, associated with
water content and water activity determinations, is of particular interest in relation to
storage of food powders, but also for gas retention in powders foreseen to foam when
dissolved (Raemy, 2003).
Chemical reactions such as hydrolysis, oxidation or reduction may be promoted. A
better understanding of the influence of temperature on the properties of foods enables food
manufacturers to optimize processing conditions and improve product quality (Nollet,
2015).
The use of easier tools to analyze the authenticity of the certificated foods to determine
the quality of final commercial products is a daily challenge.
It is, therefore, important for food scientists to have analytical techniques to monitor
the changes that occur in foods when their temperature varies. Thermal analysis is
recognized as an instrumental method of food analysis able to to give unique information
regarding the nature of the sample or the modifications induced by industrial processing.
The use of thermal analytical techniques and the information obtained is useful in
controlling food quality changes in processing and in food storage (Materazzi, 2005).
These techniques are often grouped under the general heading of thermal analysis. In
principle, most analytical techniques can be used, or easily adapted, to monitor the
temperature-dependent properties of foods, e.g., spectroscopic (NMR, UV-visible, IR
spectroscopy, fluorescence), scattering (light, X-rays, neutrons), physical (mass, density,
rheology, heat capacity) etc. Nevertheless, at present the term thermal analysis is usually
reserved for a narrow range of techniques that measure changes in the physical properties
of foods with temperature, e.g., mass, density, rheology, heat capacity.
Thermal analysis application for food technology is relatively new. The processes of
conversion in complex foods and even in pure food components (e.g., the denaturation of
proteins) often exhibit only very low energy changes.
In the meantime, these performance standards have been attained; with the results that
differential scanning calorimetry (DSC) is now being used more and more in the food
industry for routine process analysis and quality control and not just for research and
development.
The use of DSC to observe protein denaturation or gelatinization of starch has been
well established and thermal analytical techniques are extensively used to study the melting
and crystallization behavior of lipids, which exhibit complex polymorphic forms and
recrystallization phenomena. Carbohydrates and proteins in food systems are generally
water-soluble and thus show both first-order phase transitions (e.g., melting,
crystallization) and state transitions (e.g., denaturation, glass transition) (Roos, 2003).
In many food ingredients the mass loss during oven roasting processes is not only due
to moisture, but many gaseous components, such as CO2 emissions and volatile flavor and
aroma compounds. In order to study such compound losses by thermal analysis, it is
necessary use TG analysis combined with another appropriate methods of thermal analysis,
such as EGA (Ludger, 2007).
Besides, thermal analysis combined with evolved gas analysis techniques (EGA)
allows the optimization of heating treatments as well as the recognition of alterations,
caused by the heat during evaporation of water (Materazzi, 2001).
Milk contains several protein types; among which the casein attains about 94% and it
is increasingly studied due to its numerous using in the cosmetics, drug and food industries
as well as to its importance as an investigation material for elucidating essential questions
regarding the protein chemistry.
Thermal behavior of casein was studied by TG-FTIR. This study brings major
contributions regarding both the casein thermal behavior and thermal degradation
mechanism. The analysis of casein by TG/DTG, DTA in air and under N2 atmosphere is
indicative of a complex and specific degradation mechanism. It was observed the thermal
stability of casein not to depend on the nature of the degradation atmosphere. The
identification of gaseous species evolved indicated that the same species are evolved within
the endothermic domain of the degradation in air and under nitrogen atmosphere, which
would be indicative of the same degradation mechanism (Mocanu, 2012).
FINAL REMARKS
From conventional TGA to Hi-Res® TGA and EGA, analysts can discover thermal
paths of many different thermal processes, assessing very important information regarding
many different fields of science. Thermoanalytical techniques can offer in a fast, reliable
and accurate way data from small amounts of in natura solid samples, from simple thermal
degradation, to the curing of polymers, stability of materials, among others.
Such a toolbox may lead to great conclusions and even accelerate events to predict
stability, shelf-life, etc., as it is very helpful for those in the industry and academia. In
conclusion, by coupling all of this information obtained from the different measured
properties (Figure 11), from different techniques, analysts may shorten the way to a better
analytical understanding of different systems and materials, just by solving the puzzle.
Figure 11. Thermal “puzzle” by which one can define and propose many different thermal paths and
reach a wide range of materials properties.
ACKNOWLEDGMENTS
REFERENCES
Aburto, J., Moran, M., Galano, A. & Torres-García, E. (2015). Non-isothermal pyrolysis
of pectin: A thermochemical and kinetic approach. Journal of Analytical and Applied
Pyrolysis, 112, 94–104.
Arthanareeswari, M., Ravikiran, A., Kamaraj, P. & Praveen, C. (2013). Determination of
Degree of Crystallinity of Pantoprazole Sodium Sesquihydrate by Thermogravimetric
Analysis. Asian Journal of Chemistry, 25, 7337–7340.
Azevedo, A. L. M., Oliveira, R. S., Ponzio, E. A., Semaan, F. S. (2015). Sensor
development exploiting graphite-epoxy composite as electrode material. IOP
Conference Series: Materials Science and Engineering, 97, 012008.
Brown, M. E. (1988) Introduction to Thermal Analysis: Techniques and Applications.
London: Chapman and Hall.
Brožek, J., Budín, J. & Roda, J. (2007). Thermal properties of block copolymers
polyamide-6-polybutadiene. Journal of Thermal Analysis and Calorimetry, 89, 211–
215.
Charsley, E. L. & Warrington, S. B. (1991). Thermal Analysis: Techniques & Applications.
Leeds: Royal Society of Chemistry.
Chowdhury, I. H., Ghosh, S. & Naskar, M. K. (2016). Aqueous-based synthesis of
mesoporous TiO2 and Ag–TiO2 nanopowders for efficient photodegradation of
methylene blue. Ceramics International, 42, 2, 2488-2496.
Colman, M. D., Lazzarotto, S. R. S., Hansel, F. A., Colman, T. A. D. & Schnitzler, E.
(2016). Evolved gas analysis (TG-DSC-FTIR) and (Pyr-GC–MS) in the disposal of
medicines (aceclofenac). Journal of Analytical and Applied Pyrolysis, 119, 157-161.
Fernández-Berridi, M. J., González, N., Mugica, A. & Bernicot, C. (2006). Pyrolysis-FTIR
and TGA techniques as tools in the characterization of blends of natural rubber and
SBR. Thermochimica Acta, 444, 65–70.
Figura L., Teixeira A. A. (2007). Food Physics: Physical propertiess – measurement and
applications. Ed. 1. Springer-Verlag Berlin Heidelberg.
Gabbott P. (2008). Principles and Applications of Thermal Analysis. Blackwell Publishing.
Haines, P. J. (1995). Thermal Methods of Analysis: Principles, Applications and Problems.
Chapman and Hall.
Ionashiro, M. & Giolito, I. (2004). Fundamentos da Termogravimetria, Análise Térmica
Diferencial e Calorimetria Exploratória Diferencial. São Paulo: Giz Editorial.
Kubo, S. & Kadla, J. (2008). Thermal Decomposition Study of Isolated Lignin Using
Temperature Modulated TGA. Journal of Wood Chemistry and Technology, 28, 106–
121.
Leo M. L. Nollet, Fidel Toldra. (2015). Handbook of Food Analysis, 3th edition, v. 1. Taylor
& Francis.
Marinho, M. T., Bersol, L. S., Nogueira, A., Colman, T. A. D. & Schnitzler, E. (2015).
Antioxidant effect of dehydrated rosemary leaves in ripened semi-hard cheese: A study
using coupled TG–DSC–FTIR (EGA). LWT – Food Science and Technology, 63, 2,
1023-1028.
Materazzi S. and Curini R. (2001). Applied Spectroscopy Reviews, 36, 1, 1–9.
Materazzi S., Curtis A. S., Sagone F, Quaglia G. B., Bucarelli F. M., Aquili S.,
Paolesse R., Vecchio S., Curini R., Ascenzo G. D. (2005). The thermal analysis and food
quality: the possibility to qualify the pasta processing. Journal of Thermal analysis and
calorimetry, 80, 465–467.
Materazzi, S. (1997). Thermogravimetry–Infrared Spectroscopy (TG-FTIR) Coupled
Analysis. Applied Spectroscopy Reviews, 32, 4, 385-404.
Mocanu A. M., Moldoveanu C., Odochian L., Paius C. M., Apostolescu N., Neculau R.
(2012). Study on the thermal behavior of casein under nitrogen and air atmosphere by
means of the TG-FTIR technique. Thermochimica Acta, 546, 120–126.
Mothe, C. G. & Azevedo, A. D. (2009). Análise Térmica de Materiais. São Paulo: ArtLiber
Editora.
Nunes, R. S., & Cavalheiro, E. T. G. (2007). Thermal behaviour of glutamic acid its
sodium, lithium and ammonium salts. Journal of Thermal Analysis and Calorimetry,
87, 627-630.
Peng, X., Ma, X., Lin, Y., Guo, Z., Hu, S. Ning, X., Cao, Y. & Zhang, Y. (2015). Co-
pyrolysis between microalgae and textile dyeing sludge by TG–FTIR: Kinetics and
products. Energy Conversion and Management, 100, 391-402.
Raemy A. (2003). Behavior of foods studied by thermal analysis. Journal of Thermal
Analysis and Calorimetry, 71, 273–278.
Ranga V., Reddy, M., Anantha Rajmohan, R., Laxmi Shilpa, D. M., Raut, K.,
Naveenkumar, M. V., Suryanarayana, V. T. & Mathad, A. (2007). Novel quantification
method of pantaprazole sodium monohydrate in sesquihydrate by thermogravimetric
analyzer. Journal of Pharmaceutical and Biomedical Analysis, 43, 1836–1841.
Roos Y. H. (2003). Thermal analysis, state transitions and food quality. Journal of
Thermal Analysis and Calorimetry, 71, 193–199.
Santos, T. A. D., Barreto, L. N., Ritta, A. G. S. L., Meneses, W. S., Nunes, R. S., Semaan,
F. S. (2013). Cost-effective composite electrode for the fast voltammetric screening
and determination of riboflavin (B2) and pyridoxine (B6) in pharmaceuticals. RVq, 5,
548-562.
Silva, A. L., Corrêa, M. M., Oliveira, G. C., Florez-Rodriguez, P. P., Costa, C. A. R.,
Semaan, F. S., Ponzio, E. A. (2017). Development of graphite/silicon composites for
use as flexible electrode materials. Journal of Alloys and Compounds, 691, 220-229.
Silva, R. C., Caires, F. J., Gomes, D. J. C., Colman, T. A. D. & Ionashiro, M. (2014).
Synthesis, characterization, and thermal study of solid-state alkaline earth metal
Chapter 10
ISOLATION, PURIFICATION
AND CHARACTERIZATION OF PROTEINS
ABSTRACT
Protein purification is the most challenging step of the proteins’ exploration. The
protocol of purification varies according to the source of the protein, bacterial,
plant, mammalian, intracellular or extracellular. In order to isolate a protein from its
source it is demanded to disrupt the cell membrane, following processes such
as homogenation, sonication, osmotic shock, repeated freezing and thawing, grinding using
mortar and pestle, dialysis in organic or inorganic acids and enzymolysis or combinations
of them. The aim is to obtain a protein, efficiently, economically and in sufficient purity
and quantity.
Protein purification is a necessary process in many protein applications, such as
identification, immunology, polyclonal and monoclonal antibodies, protein microarrays,
enzymology, biophysical analysis, three-dimensional structure and pharmaceutical.
Proteins can be purified on the basis of characteristics such as solubility, size,
charge, specific binding affinity, hydrophobicity and isoelectric point.
Most common approaches are described in this chapter.
*
Corresponding author: Victoria Samanidou, Tel: +302310997698, FAX: +302310997719, e-mail:
samanidu@chem.auth.gr.
1. INTRODUCTION
Protein purification has been performed for more than 200 years. In 1789, Antoine
Fourcroy isolated substances from plants that had similar properties to egg white (mainly
albumin). Until the middle of the 20th century, the purification methods that were mainly
used were filtration, precipitation, and crystallization. In 1789, Fourcroy was the first who
achieved the isolation of some substances from plants having similar properties to “egg
albumin,” or egg white. In 1840, the first crystals of hemoglobin were prepared by Felix
Hoppe-Seyler. After 50 years, in 1889, Hofmeister attempted to purify the ovalbumin by
repeated crystallization. In the middle of 20th century, Cohn fractionation of plasma was
developed for the purification of albumin and other plasma proteins, as a result of
the intense requirement for blood proteins during the World War II. The method
includes multiple precipitation steps where basic parameters, such as pH, ethanol
concentration, temperature, and protein concentration are different for each step.
Furthermore, Cohn fractionation and precipitation with ammonium sulfate, are still used
today. Another innovative technique for protein’s precipitation is the aqueous two phase
extraction. In 1903, the Russian botanist Mikhail Tswett developed chromatography, as he
produced a colorful separation of plant pigments through a column of calcium carbonate,
a precious tool for the separation of proteins. In 1924, Theodor Svedberg separated proteins
by centrifugation. Since then, many important protein separation methods were introduced,
such as, Affinity Chromatography (1930s), Electrophoresis (1940s), Hydrophobic
interaction chromatography (1940s), Ion exchange chromatography (1960s), Gel filtration
(1960s) and Chromatofocusing (1970s).
Scientists face the challenge of isolating a particular protein from a complex mixture
of proteins or nonproteinaceous materials such as DNA, RNA, polysaccharides, and lipids.
It is known that thousands other proteins (5,000-8,000) with different properties except
from the target-protein exist in any cell type. The aim is to purify the target-protein from
its source efficiently, economically and in sufficient purity and yield.
Protein purification is a necessary process in many protein applications, such as
identification, immunology, polyclonal and monoclonal antibodies, protein microarrays,
enzymology, biophysical analysis, three-dimensional structure and pharmaceutical. Each
field requires different quantity of purified protein, taking into consideration the aims of
each application. In general terms, an efficient and successful protein purification can be
achieved by selecting the most suitable techniques, optimize their performance to suit the
requirements and combine them in order to increase yield using as few as possible steps
(Bonnerjea et al., 1986).
Each purification process depends on the starting material. At first, the protein source
can be bacterial. Generally, microorganisms have to be grown in controlled conditions on
a fairly large scale and be collected during different growth phases. In addition, plants can
be a source of proteins, but they present a large range of difficulties. Plant cells
consist of cell wall, chloroplasts, starch granules and other organelles, so the plant extracts
may be very low in protein. Mammalians, especially animals and humans, are the most
common source of proteins. For example, species like rats, rabbits and meat animals can
be used for studies on proteins, which come from their livers, skeletal muscles, hearts,
brains, kidneys or thymus. Furthermore, tissues from humans, such as blood and placenta
are the most easily obtained. The desired protein can be found intracellularly or
extracellularly, minimizing or maximizing, respectively, the ability of the upcoming
purification.
There are raw materials that have to be used while they are fresh, before the beginning
of natural degradation. On the other hand, frozen storage is demanded in order to maintain
the integrity of a raw material. The common storage temperature ranges between -15℃ and
-25℃ and, in rare cases, below -25℃ down to -80℃. The step after collection and storage
of the source sample includes the homogenization in order to obtain the desired protein in
a solution (Scopes, 1994).
3. CELL LYSIS
Cell disruption is the first step for releasing desired biomolecules from within the cell.
Cells can be disrupted by physical, chemical, mechanical or enzymatic ways in order to
obtain intracellular materials in a crude extract (Lee and Tai, 1999). Physical and
mechanical means include sonication, repeated freezing and thawing, high pressure
homogenization, mechanical tissue homogenizers or grinding with abrasive (e.g., using
mortar and pestle). Also, osmotic shock and solubilization in organic and inorganic acids
are chemical means for cell lysis. In many cases, a combination of the above methods can
be applied for a more efficient cell disruption for particular purposes (Goldberg, 2008).
Table 1 gives a list of techniques that can be used for releasing a protein from its cell.
3.1.1. Ultrasonication
The application of sound waves in solutions can disrupt cells. The main principle is
that the microscale high-pressure sound waves cause disruption by shear forces and
cavitation (Scopes, 1994). Cavitation is a result of the continuous movement of the probe
inside the liquid, so the pressure in the system drops below the vapor pressure and boiling
occurs as a direct effect. The cells are disrupted as the bubbles of the liquid disappear
(Goldberg, 2008).
of harsh physical conditions such as high shear stresses. Also, minimum damage to the
product is ensured (Harrison, 1991). Most bacteriolytic enzymes are not active on viable
cells, hence the biomass requires prior sensitization by heat-inactivation, chemical
pretreatment, freezing or lyophilisation.
All microorganisms used as a raw material for production of lytic enzymes are safe
and non-pathogenic. Lysozyme is one of the most commercially used bacteriolytic
enzymes for large scale application and acts attacking the β-(1,4) linkages of the
polysaccharide chains of peptidoglycan, which form the cell wall. The enzymatic cell
extraction takes place in the present of buffer Tris-HCl [Tris(hydroxymethyl)
aminomethane] and EDTA (ethylendiaminetraacetic acid). Tris-HCl contributes to the
enzyme’s release and EDTA that chelates metal ions in the outer bacterial membrane
(Andrews and Asenjo, 1987). Despite the advantages of this method, the high cost and the
unpleasant odors released during the procedure make it less popular among the cell
disruption techniques.
4. PURIFICATION METHODS
After the preparation of crude extract, various methods for the purification of one or
more proteins can be applied. Usually, the extract is subjected to preparations that separate
the proteins into different fractions based on a particular property such as solubility, size,
charge, specific binding affinity, hydrophobicity and isoelectric point. That process is
known as fractionation. All of these properties can be exploited to separate them from one
another so that they can be studied individually (Scopes, 1994). The most essential methods
for the protein purification and the protein properties that they are based on are listed on
Table 2.
proteins from one aqueous phase to another depending on the molecule’s properties, such
as size, charge, hydrophobicity, pH, temperature and concentrations of biomolecules. Most
of these properties are modified when a change occurs in the aqueous two-phase system.
Thus, ATPE is an attractive technique for recovery and purification of intracellural and
extracellural proteins. It is widely used for laboratory and large-scale protein purification
because of its simplicity and low cost (Palomares, 2004; Diamond and Hsu, 1989).
Affinity Chromatography (AF) is one of the most powerful and widely applicable
chromatographic methods for protein purification. Generally, it can be used for purification
of a specific molecule from complex mixtures. It is based on the use of an immobilized
natural ligand creating a stationary phase, which specifically interacts with the desired
protein. The protein-target is in a mobile phase as part of a complex mixture (Urh et al.,
2009). The mobile phase with desired protein is passed into the column, and the specific
interaction holds back the protein of interest, while others are washed through the column,
as it is represented in Figure 2. After this, the bound protein is eluted by a solution (Scopes,
1994).
The effectiveness of affinity purification depends on various factors, including the
available amount of the ligand, the accessibility of the ligand on the resin, the strength of
the interaction, and the integrity of protein to be immobilized. The optimization of the
whole process is required in order to maximize the interaction between a protein-target and
an immobilized ligand during the binding and wash step. Then, the interaction becomes
less weak allowing the release of the protein (Cuatrecasas, 1970).
suitable for the stationary phase without disrupting it. Size-Exclusion Chromatography is
also called Gel Filtration (GF) when the mobile phase is aqueous (Hong et al., 2012).
Gel Filtration is a powerful method for separating proteins with extremes in size.
Because the supports of this column are non-interactive, they have no trace-enrichment
potential and cannot be used for the purpose of concentrating proteins. Furthermore, it is
not usually used as a first purification step, but it is useful for removing impurities,
oligomers and aggregates of the target protein (Hagel, 2001).
primarily by an increase in the overall entropy (compared with the condition when no
interaction is occurring between the protein and the adsorbent). During separation, the
concentration of salts decreases in a reverse gradient elution. As a result, the strength of
hydrophobic interaction between the protein and the ligand decreases, and the protein will
be desorbed and eluted from the column (Synder et al., 2010).
Hydrophobic interaction chromatography is a gentle technique, with proteins eluting
in their native conformation without losing their biological activity. Also, it is widely used
for the preparative isolation of proteins in laboratory scale-up to process-scale applications.
This method is used to remove various impurities that may be present in the solution,
including undesirable product-related impurities. In particular, HIC is often used to remove
product aggregate species, which possess different hydrophobic properties than the target
monomer species (McCue, 2009).
Adsorbents in HIC are silica and polymer based and they consist of porous beads with
diameters between 5 to 200 μm, a satisfying area for ligand attachment and protein binding.
The strength of hydrophobic binding of the ligand increases with the length of the organic
chain. Moreover, the hydrophobic interaction strength can be affected by ligand density on
the support. The strength becomes higher with higher ligand densities (McCue, 2009;
Ochoa, 1978).
Reversed Phase Chromatography (RPC) is one of the most important and powerful
techniques for protein purification. Several features of this technique are responsible for its
great appeal among the purification techniques. Especially, RPC columns are packed with
small particles redounding to increased resolution, versatility, stability, reproducibility and
high peak capacities for the separation of complex mixtures. Furthermore, it’s
compatibility with mass spectrometry makes it part and parcel of protein studies.
As in Hydrophobic Interaction Chromatography, the separation in RPC is based on the
hydrophobicity of the protein molecule. RPC columns are mainly packed with silica gel or
polymers. Especially, for separation of proteins large pore silicas are required (pore
diameter≥30) for high efficiency. Also, monolithic columns are an alternative choice
avoiding bulk materials. These columns are made from silica or polymer and have a rigid
cylinder shape (Guiochon, 2007; Neville, 2010). The most popular stationary phases are
based on straight-chain alkyl groups (C4, C8, C18).
In RPC the mobile phase consists of an aqueous buffer, often a dilute acid such as
phosphoric, trifluoroacetic, formic or acetic acid, and an organic solvent, such as methanol,
acetonitrile or isopropanol. It is important to maintain pH of aqueous buffer in a stable
range from 2 to 3.5 avoiding the denaturation of proteins. The presence of acid renders the
proteins positively charged and minimizes unpleasant interaction with the stationary phase
(Josic and Kovac, 2010).
4.8. Chromatofocusing
In the middle of 1970s, Sluyterman and his colleagues have introduced a new protein-
separation technique, chromatofocusing. Chromatofocusing is another technique that
separates proteins according to a particular property. In this case, proteins are separated
because of their different isoelectric point. This technique is a specialized form of ion-
exchange chromatography in which proteins are eluted from the column with a pH
gradient. The pH gradient is formed by mixing two different buffer mixtures within the
column. Chromatofocusing is successful with proteins that are stable and soluble at their
isoelectric point, as they are eluted in a buffer close to their pI (Sluyterman et al., 1978).
Firstly, a weak anion exchanger is used for the equilibration of the column defining the
upper pH for high retention of the sample. Then, a second low-pH buffer mixture is applied
as a mobile phase to elute bound proteins at or near their isoelectric point, which is a pH
value at which there is no longer bind to the exchanger. Proteins migrate because of
descending pI values. The protein with the highest pI is eluted first and the protein with the
lowest pI is eluted last. The lower limit of the gradient is defined by the pH of the elution
buffer. Chromatofocusing is therefore a powerful analytical technique with high capacity,
so it is appropriate for preparative separations (Li and Hutchens, 1992).
A short coming of this technique is the fact that the solubility of proteins is decreased
during the elution. It is recommended to add zwitterions, such as taurine and glycine,
nonionic and zwitterionic surfactants in concentrations up to 2 M, which help solubilize
proteins with no change of ionic strength of the elution buffer, which have to be low (Ming
and William, 1992).
4.9. Electrophoresis
The velocity of migration (ν) of a protein in an electric field depends on the electric
field strength (E), the net charge on the protein (z) and the frictional coefficient (f)
according to the Equation 1.
𝐸𝑧
𝑣= (1)
𝑓
𝑓 = 6𝜋𝜂𝑟 (2)
There are two types of electrophoresis, one and two dimensional depending on the
scale of separation. One dimensional electrophoresis is applied for most routine protein
purification.
Separations by electrophoresis are carried out in thin and vertical gels made up of the
cross-linked polymer polyacrylamide because they are chemically inert and are readily
formed. The direction of flow is from top to bottom. Molecules that are small compared
with the pores in the gel readily move through the gel, whereas molecules much larger than
the pores stay at the top. Intermediate-size molecules move through the gel with various
degrees of facility.
The most common applied electrophoretic method for the estimation of purity and
molecular weight is based on the detergent sodium dodecyl sulfate (SDS) (Figure 4). SDS
is an anionic detergent, meaning that when dissolved its molecules have a net negative
charge within a wide pH range. A polypeptide chain binds amounts of SDS in proportion
to its relative molecular mass. SDS-electrophoresis separates proteins on the basis of their
molecular weight. Thus, the progress of the purification process can be monitored as the
number of protein bands visible on the gel decreases after each new fractionation step. If
the position of a protein with known molecular weight in the gel is known, it is possible to
measure the molecular weight of an unidentified protein by monitoring its position. If the
protein consists of two or more different subunits, the subunits will generally be separated
by the SDS detergent and a separate band will appear for each (Stellwagen, 2009).
generated across the gel. Each protein migrates throughout the gel until it meets its pI
(Table 3).
Protein pI
Pepsin <1.0
Egg albumin 4.6
Serum albumin 4.9
Urease 5.0
β-Lactoglobulin 5.2
Phycoerythrin 5.74
Hemoglobin 6.8
Myoglobin 7.0
Chymotrypsinogen 9.5
Cytochrome c 10.7
Lysozyme 11.0
There are some complex mixtures of proteins that can not be separated by one
dimensional gel electrophoresis. In these cases, two-dimensional electrophoresis is applied,
which combines isoelectric focusing and SDS electrophoresis sequentially in a process
achieving the separation of more than one hundred proteins in a polyacrylamide gel. Two-
dimensional polyacrylamide gel electrophoresis (2D-PAGE) is usually performed with
isoelectric focusing in the first dimension and SDS-PAGE in the second dimension.
Especially, this method separates proteins of identical molecular weight in the vertical
direction that differ in pI, or proteins with similar pI values but different molecular weights
in the horizontal direction (Rabilloud and Lelong, 2011).
5.1. Spectroscopy
To quantify the amount and concentration of purified protein, the simplest and most
common method is the use of a spectrophotometer, since proteins absorb light at a specific
wavelength. Most proteins have an absorption maximum at about 280 nm, which is due to
the aromatic amino acids. The type of amino acids that are present in a protein affect the
absorption. Especially, proteins that have similar molecular weight but different amount of
tyrosine and tryptophan will have different absorbance values. Furthermore, there are some
other parameters that may affect the absorbance of a specific protein, such as temperature,
ionic strength, pH and the presence of detergents. The integrity of amino acids depends on
the above parameters, so the ability of aromatic residues to absorb at 280 nm will be
affected, changing the value of the protein’s extinction coefficient. This method of
detection can be used independently in the stand-alone mode using a manual syringe or
pump for sample injection or it can be used on-line with an HPLC system. After a
purification step it is recommended to assess the purity by a spectrophotometer obtaining
quick results without destructing the protein sample.
Apart from UV-visible spectroscopy, fluorescence spectroscopy is used for the
quantitation and characterization of purified proteins. Especially, the fluorescence
emission signal is measured at 280 nm and 340 nm, corresponding respectively to light
scattering and intrinsic protein fluorescence, after excitation at 280 nm. The ratio of the
intensities at 280 nm and 340 nm (I280/I340) is called aggregation index (AI), because it is
related to the degree of aggregation of the sample, but not to the concentration. This index
should have values close to zero for aggregate-free protein preparations and can obtain
values higher than 1 unit, when significant aggregation takes place (Raynal et al., 2014).
purification process. Besides protein purification, this technique can be used for
determining the molecular weight of proteins. The molecular mass of a given protein may
be determined by comparing its ratio of Ve/Vo (Ve: elution volume, Vo: void volume) with
those of protein standards with known molecular mass, which are commercially available.
The column void volume corresponds to the elution volume of a very large molecule.
Plotting the logarithms of the known molecular masses of protein standards against their
respective ratios of elution volume to column void volume, Ve/Vo values, produces a linear
calibration curve (Duong-Ly and Gabelli, 2014).
Nowadays, mass spectrometer detectors has become the most popular detector systems
for bioanalytical methods, such as purification processes. MS detectors are used in two
conformations, the mass selective detector (MSD) and the multiple reaction monitoring
(MRM). The first type of detection is based on the measurement of the protonated
molecular ion (M + H) for each analyte. Whereas, the MRM process is based on the
isolation of the primary ionic species (parent ions), the fragmentation of them into
additional ions (product ions) and the monitoring of them. When mass spectrometry is
combined with HPLC systems, where proteins have been separated, the systems are called
LC-MS or LC-MS/MS, respectively (McMaster, 2005; Gika et al., 2013).
The MS detector interface disperses the liquid sample of proteins into the gas phase,
whereas the pressure must be extremely lower than the atmospheric (760 Torr), in a range
of 10-5 to 10-6 Torr. The most popular ionization methods are electrospray ionization (ESI),
atmospheric pressure chemical ionization (APCI) and matrix-assisted laser desorption-
ionization (MALDI).
One of the most widely used MS detector is the time-of-flight (TOF). In this mass
analyzer, ions are derived from the ion source and are accelerated with a specific energy
and directed through a high vacuum flight tube to the detector. The small ions are faster
than heavy, so they reach detector first, since they demand less energy. The time of flight
of ions is related to the mass (m/z) of the ion.
Now, MALDI-TOF-MS is routinely used in many laboratories for the rapid and
sensitive identification of proteins by peptide mass fingerprinting (PMF). Firstly, proteins
are separated by one- or two-dimensional gel electrophoresis or liquid chromatography.
Then, proteins are de-stained and cleaved with sequence-specific endoproteases, such as
an in-gel trypsin digestion. Proteins can be identified at femtomole levels by MALDI-TOF-
MS and database searching of the PMF data (Thiede et al., 2005; Webster et al., 2012).
CONCLUSION
Nowadays scientists hold many useful tools in their hands to purify and isolate proteins
of interest, so their conformation, properties and specific activities can be studied. The final
application of protein defines the required level of its purity. The assessment of isolation
process can be achieved by a wide range of techniques, which can determine also the
concentration of the purified protein. The most challenging part for scientists is to combine
in the most effective way the techniques that have been described above with the minimum
loss of the desired protein in the fewest steps. In general terms, isolation and purification
of proteins is a field that will not stop evolving as the study of cells and organisms is
limitless.
REFERENCES
Alberts, Bruce., Dennis, Bray., Karen, Hopkin., Alexander, Johnson., Julian, Lewis.,
Martin, Raff., Keith, Roberts. & Peter, Walter. (2009). Cytoskeleton. Essential Cell
Biology. doi:citeulike-article-id:4505949.
Andrews, B. A. & Asenjo, J. A. (1987). “Enzymatic Lysis and Disruption of Microbial
Cells.” Trends in Biotechnology, 5 (10), 273–77. doi:10.1016/0167-7799(87)90058-8.
Bonnerjea, J., Oh, S., Hoare, M. & Dunnill, P. (1986). “Protein Purification: The Right
Step at The Right Time.” Bio/technology, 4 (July), 954–58. doi:10.1038/nbt1186-954.
Boyle, John. (2005). “Lehninger Principles of Biochemistry (4th Ed.): Nelson, D., and Cox,
M.” Biochemistry and Molecular Biology Education, 33, 74–75. doi:10.1002/
bmb.2005.494033010419.
Chaiyarit, Sakdithep. & Visith, Thongboonkerd. (2009). “Comparative Analyses of Cell
Disruption Methods for Mitochondrial Isolation in High-Throughput Proteomics
Study.” Analytical Biochemistry, 394 (2). Elsevier Inc.: 249–58. doi:10.1016/
j.ab.2009.07.026.
Cuatrecasas, P. (1970). “Protein Purification by Affinity Chromatography.” J. Biol. Chem.,
245 (12), 3059–65.
Diamond, D. & Hsu, J. T. (1992). “Aqueous Two-Phase Systems for Biomolecule
Separation.” Advances in Biochemical Engineering/biotechnology, 47, 89–135.
Duong-Ly, Krisna C. & Sandra, B. Gabelli. (2014). “Gel Filtration Chromatography (Size
Exclusion Chromatography) of Proteins.” In Methods in Enzymology. 1st ed. 541, 105-
114. Elsevier Inc. doi:10.1016/B978-0-12-420119-4.00009-4.
Duong-Ly, Krisna C. & Sandra, B. Gabelli. (2014). “Salting out of Proteins Using
Ammonium Sulfate Precipitation.” In Methods in Enzymology. 1st ed. 541, 85-94.
Elsevier Inc. doi:10.1016/B978-0-12-420119-4.00007-0.
England, Sasha. & Seifter, Sam. (1990). “Precipitation Techniques.” Methods in
Enzymology., 182, 285–300.
Gika, Helen G., Georgios, A. Theodoridis., Robert, S. Plumb. & Ian, D. Wilson. (2014).
“Current Practice of Liquid Chromatography-Mass Spectrometry in Metabolomics and
Metabonomics.” Journal of Pharmaceutical and Biomedical Analysis, 87, 12–25.
doi:10.1016/j.jpba.2013.06.032.
Goldberg, Stanley. (2008). “Mechanical/physical Methods of Cell Disruption and Tissue
Homogenization.” Methods in Molecular Biology (Clifton, N.J.), 424, 3–22.
doi:10.1007/978-1-60327-064-9_1.
Guiochon, Georges. (2007). “Monolithic Columns in High-Performance Liquid
Chromatography.” Journal of Chromatography A, 1168, 101–68. doi:10.1016/
j.chroma.2007.05.090.
Greenfield, Norma J. (2006). “Using circular dichroism spectra to estimate protein
secondary structure.” Nature Protocols, 1, 2876-2890.
Harrison, Susan T. L. (1991). “Bacterial Cell Disruption: A Key Unit Operation in the
Recovery of Intracellular Products.” Biotechnology Advances, 9, 217–40. doi:10.
1016/0734-9750(91)90005-G.
Hidayat, U. Khan. (2012). “The Role of Ion Exchange Chromatography in Purification and
Characterization of Molecules.” In Ion Exchange Technologies, edited by Ayben
Kilislioğlu, InTech, DOI: 10.5772/52537.
Hong, Paula., Koza, Stephan. & Bouvier, Edouard S. P. (2012). “Size-Exclusion
Chromatography for the Analysis of Protein Biotherapeutics and their Aggregates.”
Journal of Liquid Chromatography and Related Technologies, 35, 2923-2950.
Sharon, M. Kelly., Thomas, J. Jess. & Nicholas, C. Price. (2005). “How to Study Proteins
by Circular Dichroism.” Biochimica et Biophysica Acta - Proteins and Proteomics,
1751, 119–39.
Shouqin, Zhang., Zhu, Junjie. & Wang, Changzhen. (2004). “Novel High Pressure
Extraction Technology.” International Journal of Pharmaceutics, 278, 471–74.
doi:10.1016/j.ijpharm.2004.02.029.
Snyder, Lloyd R., Joseph, J. Kirkland. & John, W. Dolan. (2010). Introduction to Modern
Liquid Chromatography. http:// 197.14.51.10:81/pmb/BIOLOGIE/ 0470167548_
Chromatography.pdf.
Stellwagen, Nancy C. (2009). “Electrophoresis of DNA in agarose gels, polyacrylamide
gels and in free solution”. Electrophoresis., 30, 188-195.
Thiede, Bernd., Wolfgang, Höhenwarter., Alexander, Krah., Jens, Mattow., Monika,
Schmid., Frank, Schmidt. & Peter, R. Jungblut. (2005). “Peptide Mass Fingerprinting.”
Methods, 35, 237–47.
Urh, Marjeta., Dan, Simpson. & Kate, Zhao. (2009). “Affinity Chromatography. General
Methods.”Methods in Enzymology, 463, 417-438. Elsevier Inc.
Vickery, Hubert Bradford. (1950). “The Origin of the Word Protein.” The Yale Journal of
Biology and Medicine, 22 (5), 387–93. doi:10.1093/nq/s6-II.47.415-e.
Webster, J. & Oxley, D. (2012). “Protein identification by MALDI-TOF mass
spectrometry.” Methods in Molecular Biology, 800, 227-240.
The main research fields of Dr. Christian Celia are focused on the design, synthesis
and physicochemical characterization of vesicular and supramolecular nanocarriers for
systemic and topical delivery of bioactive compounds. Dr. Christian Celia has specific
skills on the physicochemical characterization of vesicular and supramolecular
nanocarriers; analytical evaluation of bioactive compounds by using chromathographic and
spectrophotometer apparatus; evaluation of pharmacokinetic protocols and analysis of
biological samples; bio-distribution of bioactive compounds by using in vitro and in vivo
experimental protocols and models. Dr. Christian Celia carried out his research activity in
Department of Pharmacy, University of Chieti - Pescara “G. d'Annunzio”, and Department
of Nanomedicine, Houston Methodist Research Institute. Dr. Christian Celia is co-author
of 58 publications in peer-reviewed international journals, 1 proceeding paper, 100 meeting
abstracts and communications at national and international conferences and meetingd, 2
Guest Editorials in international peer-reviewed journal, 1 book chapter, and 1 Italian patent.
A C
accelerated solvent extraction (ASE)., 31 calibration, 4, 6, 7, 8, 9, 16, 21, 110, 112, 113, 114,
accuracy, 3, 4, 5, 8, 9, 14, 20, 24, 55, 112, 113, 117, 115, 123, 124, 174, 180, 271, 272, 274, 301, 338,
147, 172, 216, 225, 260, 299, 309, 315, 320, 325, 347, 348, 350, 355, 382
326, 341, 346, 347, 351, 354 carry over, 11
aggregation index, 381 cell disintegration, 369, 371
amino, 56, 62 cell disruption, 369, 370, 371, 384
ammonium sulfate, 368, 372, 384 cell lysis, 283, 369, 370, 371
analytical applications, 56, 66, 69 chemicals, 20, 41, 56, 68
anthocyanidins, 80, 81, 83 chemometrics, 2, 105, 119, 120, 151, 152, 159, 161,
anticancer effects, 98 196, 200, 203, 211, 216, 224, 234, 237, 239, 242,
antioxidant, 77, 78, 86, 87, 88, 89, 90, 91, 92, 93, 94, 243, 249, 257, 260, 285, 293, 295, 329
95, 96, 97, 98, 99, 100, 101, 102, 160, 175, 188, chromatofocusing, 368, 372, 378, 385
228, 234, 244, 251, 365 chromatographic assay, 2, 3
antioxidant capacity, 78, 87, 88, 89, 91, 93, 94, 95, circular dichroism, 383, 384, 386
98, 100, 102, 244 CSA, 46, 47
ANVISA, 1, 18 CV (coefficient of variation), 2, 6, 11, 114, 299, 300,
aqueous two-phase extraction, 372, 385 326, 327
atmospheric pressure chemical ionization, 382
D
B
DAD (diode array detector), 2, 56, 244
BET (Brunauer-Emmett-Teller), 56, 61 differential centrifugation, 370, 371
bioactive compounds, 6, 77, 95, 336 differential thermal analysis, 346, 347
bioanalytical method, 2, 4, 5, 11, 18, 19, 20, 21, 22, dispersion, 47, 48, 70, 75
382 dispersive liquid-liquid microextraction (DLLME),
biological effects, 101 32, 33
E G
EDCs (endocrine disrupting chemicals), 20, 41, 56, GC (gas-chromatography), 2, 3, 20, 21, 47, 56, 126,
68 130, 131, 174, 198, 236, 240, 247, 249, 364, 366
electron, 56, 61, 71 gel filtration, 368, 372, 376, 381, 384
electron microscopy, 56, 61, 71 green analytical chemistry (GAC), 49
electrophoresis, 43, 47, 147, 232, 236, 288, 332, 337,
368, 378, 379, 380, 382, 385, 386
H
electrospray ionization, 20, 21, 102, 382
ELISA (enzyme-linked immunosorbent assay), 2, 3
health, 77, 78, 99, 102, 201
EMA, 1, 11, 18
health benefits, 77, 78, 99, 102, 201
endocrine, 20, 41, 56, 68
high-pressure homogenization, 370
enzyme digestion, 371
HILIC (hydrophilic interaction liquid
evolved gas analysis, 340, 346, 355, 357, 362, 364,
chromatography), 2, 17
366
HPLC (high performance liquid chromatography), 1,
extraction, 13, 17, 23, 26, 31, 35, 36, 37, 38, 39, 43,
2, 3, 4, 11, 13, 14, 15, 16, 17, 18, 21, 29, 47, 56,
56, 64, 66, 67, 70, 72, 74, 332, 335
70, 74, 75, 78, 94, 95, 96, 97, 99, 100, 102, 117,
165, 198, 219, 222, 226, 240, 244, 247, 331, 332,
F 337, 381, 382
HPLC-MS (high performance liquid
fabric phase sorptive extraction (FPSE), 23, 45, 47, chromatography-mass spectrometry), 2, 3, 4, 14,
48 15, 16, 17, 18, 21, 56, 226
FDA, 1, 3, 5, 11, 18, 19 HPLC-online DPPH, 78
flavones, 80, 81, 86 HPLC-online TEAC, 78, 99
flavonoids, 31, 78, 79, 80, 84, 85, 86, 87, 92, 95, 96, HR-MAS NMR, 108, 109, 148, 159, 197, 203, 204,
97, 98, 99, 100, 101, 102, 161, 165, 201, 204, 205, 236, 237, 243, 249
208, 241 HS-SDME, 34
flavonols, 80, 81, 239 hydrophobic interaction chromatography, 368, 376,
FLD (fluorescent detector), 2, 56 377, 385
food, 1, ii, iii, 52, 95, 98, 99, 104, 105, 107, 108, hydroxybenzoic acids, 79, 86
110, 128, 132, 135, 136, 142, 146, 147, 148, 150,
152, 153, 155, 158, 159, 160, 191, 219, 232, 233,
I
234, 235, 238, 241, 243, 244, 245, 246, 247, 256,
257, 267, 286, 287, 288, 289, 290, 291, 292, 293,
ICH (international conference of harmonization), 2,
295, 365
3, 4, 5, 19
food composition, 49, 98, 100, 105, 116, 127, 159,
instrument configurations, 2, 4, 18
233, 235, 257
international guidelines, 2, 11, 18
food science, 1, ii, iii, 52, 95, 98, 99, 104, 105, 107,
international guidelines, 2
108, 110, 128, 132, 135, 136, 142, 146, 147, 148,
ion-exchange chromatography, 373, 378
150, 152, 153, 155, 158, 159, 160, 191, 219, 232,
IONPs (iron oxide based nano particles), 56, 59
233, 234, 235, 238, 241, 243, 244, 245, 246, 247,
isoelectric focusing, 379, 380, 385
256, 257, 267, 286, 287, 288, 289, 290, 291, 292,
isoelectric point, 367, 371, 372, 373, 378, 379, 380
293, 295, 365
isoflavones, 80, 86, 99, 154
fractionation, 26, 36, 93, 368, 371, 379
isolation, 42, 93, 101, 245, 250, 367, 368, 377, 382,
FTIR (fourier transform infrared spectroscopy), 56,
383, 384
61, 222, 224, 354, 355, 356, 357, 358, 359, 360,
362, 364, 365, 366
NPs (nanoparticles), ix, 56, 57, 58, 59, 60, 61, 62,
L 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75,
333
lignans, 79, 85, 86, 88, 99, 101
nutraceuticals, 77, 78, 101, 208, 227, 228, 232
liquid state NMR, 107, 159
LLE (liquid-liquid extraction), 2, 13, 17, 40, 41, 43,
47, 66, 372 O
LOQ (limit of quantification), 2, 3, 4, 5, 6, 8, 9, 10,
11, 14, 20 online methods, 94
low field NMR relaxometry, 257, 260, 285 osmotic shock, 367, 369, 371
lysozyme, 148, 153, 227, 243, 371, 380
P
M
parallelism test, 4, 11
magnetic nanoparticles, ix, 55, 56, 57, 58, 59, 60, 62, PE (process efficiency), 2, 13, 16, 17, 35, 43
64, 66, 67, 68, 69, 70, 71, 72, 73, 75 peak post-extraction method, 15
magnetic solid-phase extraction (MSPE), 38, 66 phenolic acids, 79, 86, 92, 95, 96, 100, 101
MAS, 108, 109, 148, 159, 197, 203, 204, 205, 236, phenolic components, 78, 86, 92, 99
237, 243, 249 phenolics, 78, 79, 80, 84, 96, 98, 100, 101, 102
mass selective detector, 382 phytochemicals, 24, 77, 78, 154, 246
mass spectrometry, 4, 12, 13, 14, 18, 19, 20, 21, 22, post-column infusion method, 15
74, 99, 102, 115, 147, 149, 183, 188, 246, 251, PP (precipitation of proteins), 2, 13, 17, 335
311, 331, 332, 333, 334, 360, 377, 382, 386 precipitation, 2, 13, 17, 47, 56, 58, 59, 60, 66, 368,
matrix, 47, 48, 70, 75 372, 384
matrix solid phase dispersion (MSPD), 47, 48, 70, 75 pressurized hot water extraction (PHWE), 27, 28
matrix-assisted laser desorption-ionization, 382 pressurized liquid extraction (PLE), 31, 32
ME (matrix effect), ix, 1, 2, 3, 4, 12, 13, 14, 15, 16, primary contact surface area (PCSA), 46, 47
17, 18, 19, 20, 21, 22 process efficiency, 16
medicinal plants, 77, 94 proteins, x, 2, 43, 60, 64, 68, 74, 88, 91, 96, 112,
metal chelating, 89 126, 129, 138, 160, 161, 170, 200, 226, 227, 228,
method validation, 7, 11, 18, 19, 20, 114 229, 230, 231, 232, 234, 241, 243, 244, 259, 269,
microextraction by packed sorbent (MEPS), 22, 39, 272, 284, 285, 362, 367, 368, 369, 370, 371, 372,
40, 45, 47 373, 375, 376, 377, 378, 379, 380, 381, 382, 383,
microwave-assisted extraction (MAE), 28, 29, 31 384, 385, 386
M-NPs (magnetic nanoparticles), ix, 55, 56, 57, 58, purification, 50, 57, 58, 70, 71, 93, 166, 215, 367,
59, 60, 62, 64, 66, 67, 68, 69, 70, 71, 72, 73, 75 368, 369, 370, 371, 372, 373, 374, 375, 376, 377,
molecularly imprinted polymers, 41 378, 379, 381, 382, 383, 384, 385
MRM (multiple reactions monitoring), 2, 382
MS (mass spectrometer), 2, 13, 14, 15, 17, 18, 19,
Q
21, 22, 74, 99, 100, 115, 116, 117, 126, 127, 130,
131, 166, 174, 198, 222, 236, 240, 244, 303, 304,
QCs (quality control samples), 2, 5, 6
328, 331, 332, 360, 364, 366, 382, 385
Quick, easy, cheap, effective, rugged, and safe
multiple reaction monitoring, 382
(QuEChERS), 40, 41
N R
natural products, 78, 88, 96, 97, 129
RE (recovery), 2, 4, 5, 6, 15, 28, 41, 135, 136, 155,
new chemical entities, 3
165, 209, 262, 273, 292, 320, 373, 384, 385
NMR-imaging, 257, 285, 295