isotermas de adsorção

© All Rights Reserved

Просмотров: 12

isotermas de adsorção

© All Rights Reserved

- tmp1495.tmp
- Homework 6
- Adsorption
- Adsorption Studies of Direct Red 28 Dye Onto Activated Carbon
- Biosorption and Kinetic Studies on Oil Removal From Produced Water Using Banana Peel
- Treatment of Oil Spill by Buffing Dust as an Efficient Adsorbent - InESPO 2014
- Imaga542014IRJPAC13510_1
- Removal from wastewater and recycling of azo textile dyes by alginate-chitosan beads
- tmpF18F.tmp
- Journal of Environmental Science and Engineering,Vol.6,No.10A,2017
- Studies on the Treatment of Brilliant Green Solution by Combination Microwave Induced Oxidation w
- 1-s2.0-S1878535212002754-main.pdf
- Ict
- Saponins Clay Modified Materials a New Approach Against Callosobruchus Subinnotatus in Stored Products
- Adsorption of Reactive Black 5 From an Aqueous Solution - Equilibrium and Kinetic Studies
- molvig courtney adsorptionofaceticacidbyasolid
- Ball Clay Untuk Adsorpsi Violet Dye
- 1
- Equilibrium Modeling and Kinetic Studies on the Adsorption of Basic Dye by A low-Cost Adsorbent: Sugarcane Bagasse
- Fe3O4ÔÇôwheatstraw Preparation Characterization

Вы находитесь на странице: 1из 10

by-products: Comparison of isotherm models

a,*

Diana Ringot , Benoit Lerzy a, Kathy Chaplain a, Jean-Paul Bonhoure a,b

,

Eric Auclair c, Yvan Larondelle d

a

Institut Superieur dAgriculture de Beauvais, Rue Pierre Waguet, BP 30313, Beauvais Cedex 60026, France

b

Centre de Valorisation des Glucides, 33, Avenue Paul Claudel, 80000 Amiens, France

c

Lesare Feed Additives, 1 Rue du Haut Touquet, 59520 Marquette-Lez-Lille, France

d

Universite catholique de Louvain, Unite de biochimie de la nutrition, Croix du Sud, 2/8, Louvain-la-Neuve 1348, Belgium

Received 16 June 2004; received in revised form 24 May 2006; accepted 28 June 2006

Available online 21 August 2006

Abstract

Biosorption of ochratoxin A (OA) onto yeast biomass appears to be a reasonably low cost decontamination method. In vitro adsorp-

tion of OA onto three yeast industry by-products: a vinasse containing yeast cell walls (EX16), a puried yeast beta-glucan (BETA) and a

yeast cell wall fraction (LEC) was examined at 25 C. Seven classical adsorption models were tested to provide the best description of

toxin adsorption. A comparison of these models was performed using the magnitude of the coecient of determination R2 for the linear

models and the value of the sum of normalised errors (SNE) for linear and non-linear models. Based on the R2 and the SNE values, Hill,

Freundlich and BrunauerEmmettTeller equations produced the best models for OA biosorption onto respectively, EX16, BETA and

LEC. For these best models, the values of isotherm constants were consistent when measured using both linear and non-linear calcula-

tions. The SNE calculation procedure presented in this paper in association with the linear equation analysis method is an appropriate

approach for designing a better adsorption isothermal model.

2006 Elsevier Ltd. All rights reserved.

1. Introduction spices, coee and cocoa products, grape juice and beer

(Commission of the European Union, 2002).

Ochratoxin A (OA) is a secondary metabolite of OA is partially absorbed by the gastrointestinal tract

some species of storage fungi. Aspergillus species such as in monogastrics. In ruminants, OA is hydrolysed by the

A. ochraceus produce OA mainly in tropical climates and microbial ora into a much less toxic metabolite, ochra-

Penicillium species such as P. verrucosum produce OA toxin a. OA is hydrolysed to ochratoxin a by mammalian

in temperate climates. carboxypeptidase A and by some micro-organisms within

Chemically speaking, OA is a chlorine-containing di- the gastrointestinal tract. In hepatic cells, it is hydroxylated

hydroisocoumarine linked to L-phenylalanine. OA and its mainly to 4-R-hydroxyochratoxin A (OA-OH) by mixed

hydroxyl derivatives are toxic when consumed by livestock function oxidases (Li et al., 2000). OA is considered to be

(pigs, poultry) and humans. They are found in human milk, a genotoxic agent in vivo and in vitro in some mammalian

blood serum, plasma, liver and kidney. The main route of species (WHO, 2002), but the mechanism of genotoxicity

exposure is the consumption of contaminated cereals, wine, is unclear and there is no evidence of a direct interaction

with DNA (Bakker and Pieters, 2002; Ringot et al.,

*

Corresponding author. Tel.: +33 344 062 517; fax: +33 344 062 526. 2006). OA is reasonably anticipated to be a possible human

E-mail address: diana.ringot@isab.fr (D. Ringot). carcinogen based on sucient evidence of carcinogenicity

0960-8524/$ - see front matter 2006 Elsevier Ltd. All rights reserved.

doi:10.1016/j.biortech.2006.06.015

D. Ringot et al. / Bioresource Technology 98 (2007) 18121821 1813

in experimental animals (group 2B) (IARC, 1993). OA is Yeasts produce a high quantity of biomass. They are

a nephrotoxic, teratogenic and immunotoxic compound used in a large variety of industrial fermentation processes;

(Muller et al., 1999; Prelusky et al., 1994). OA is relatively moreover, they may be regarded as a good source of adsor-

stable and is only partially degraded under normal process bent material, due to the presence in their cell wall of some

conditions. specic macromolecules such as the mannoproteins and

Dierent approaches have been developed to reduce the beta-glucans. The ability of the Saccharomyces cerevisiae

impact of mycotoxins. They can be divided into pre- and cell wall to bind zeralenone has been reported recently

post-harvest strategies and into biological, chemical and (Yiannikouris et al., 2003).

physical methods (Huwing et al., 2001). Physical methods This research was aimed at examining in vitro ad-

involve extraction with solvent, adsorption, inactivation sorption of OA onto three yeast industry by-products at

by heat and irradiation. Among the physical decontamina- a temperature of 25 C. To this end, several theoretical

tion methods, adsorption onto various types of compounds adsorption models frequently used in the literature were

(hydrated sodium calcium aluminosilicate HSCAS, tested for their ability to describe the equilibrium sorption

kaolin, silica binding agent, bentonite, etc.) has been ex- data. For these models, ve error functions and the norma-

tensively studied in recent years (Grant and Philips, 1998; lised error sum (SNE) were examined. This enabled the

Scott, 1998). Dierent binding agents including activated determination of the best parameter set and thus provided

carbons (Galvano et al., 1998), zeolites (Tomasevic-Cano- an accurate equilibrium adsorption model.

vic et al., 2003), diatomaceous earth (Natour and Yousef,

1998), cholestyramine and mixtures of sterilized yeast and 2. Methods

fermentation residua from beer production (Bauer, 1994)

have been reported to remove OA in vitro. In contrast, acti- 2.1. Adsorbents and reagents

vated charcoal and HSCAS have been reported to have a

limited ecacy against OA in vitro (Ramos et al., 1996). Three adsorbent materials, namely EX16, BETA and

Castellari et al. (2001) examined several ning treatments LEC, were obtained from Bio-Springer, Maison-Alfort,

to reduce OA levels in red wine. They found that potassium France. Two of these adsorbents, EX16 (a vinasse contain-

caseinate and activated carbon were the best agents to ing 16% liquid yeast cell walls) and LEC (a dry yeast cell

remove OA. wall fraction) are industrial by-products of the yeast indus-

The phenomenon of biosorption is dened as a metabo- tries. BETA is the dried puried beta-glucans fraction of

lism independent adsorption of pollutants based on the cell walls. The dry matter content of the by-products was

partition process on a microbial biomass. Microbial bio- 56% for EX16, 95% for BETA and 97.6% for LEC.

mass consists of small particles, with low density, poor OA was purchased from Sigma Chemical Company

mechanical strength and little rigidity. This phenomenon (St. Louis, Missouri, USA). Primary methanolic stock

is generally based on a set of chemical and physical mech- solution (200 mg/L) was prepared in methanol. Five OA

anisms (involving physico-chemical interactions such as test solutions with concentrations of 0.5, 1.0, 2.0, 5.0 and

electrostatic interactions, ion exchange, complexation, che- 10.0 mg/L were prepared by the dilution of the methanolic

lation and precipitation) leading to the immobilization of a stock solution with deionised distilled water.

solute component on the microbial cell wall components. It

has been investigated by dierent authors (Aksu, 2003; 2.2. Experimental system

Aksu and Donmez, 2002; Aksu and Yenner, 1998; Tsezos

and Bell, 1989). Samples of 500 mg of adsorbents were placed in screw

The complexity of the microbial structure implies that cap test tubes with aliquots (10 mL) of the OA test solu-

there are many ways for the pollutant to be captured by tions. Two sets of controls were prepared, one by adding

the cells. Biosorption mechanisms are therefore various 10 mL of mycotoxin test solutions without the addition

(physical adsorption, chemical binding of ionic groups, of adsorbent, and another by adding 500 mg adsorbent

ion exchange, etc.) and in some cases they are still not very to 10 mL deionised distilled water. All experiments were

well understood (Veglio and Beolchini, 1997). Cell surface carried out in triplicate.

sorption is a physico-chemical interaction between the The controls and test tubes were placed in a horizontal

toxin and functional groups of the cell surface, based on shaker-incubator at 400 rpm and at a controlled tempera-

physical adsorption, ion exchange and complexation, ture of 25.0 1.0 C. Preliminary investigations showed

which is not dependent on metabolism. Cell walls of micro- that equilibrium uptake was attained rapidly, with prac-

bial biomass, mainly composed of polysaccharides, pro- tically no change observed after a period of 30 min. All

teins and lipids, oer abundant functional groups, such adsorptions were run for 90 min to ensure complete

as carboxyl, hydroxyl, phosphate and amino groups, as uptake. After the incubation period, solid particles (adsor-

well as hydrophobic adsorption sites such as aliphatic bent) were separated from the supernatant by centrifu-

carbon chains and aromatic rings (Ringot et al., 2005). gation at 6000g for 20 min at 25 C in an ALC 4239R

This physico-chemical phenomenon is quick and can be centrifugation system (Fisher Scientic Labosi, Elancourt,

reversible. France).

1814 D. Ringot et al. / Bioresource Technology 98 (2007) 18121821

methanol:deionised distilled water (50:50). The mixture

was vortexed for 5 min. This extract was diluted, ltered 0.16

and analysed by HPLC.

0.12

Qeq, mg/g

2.3. Chromatographic analysis

0.08

The HPLC chromatographic system consisted of a Spec-

traSYSTEM P1500 isocratic pump (Thermo Separation 0.04

EX16

Product, Fremont, California, USA) equipped with a AS BETA

LEC

3000 model injection valve (100 ll) (Thermo Separation 0.00

Product, Fremont, California), a Spectroow 980 uores- 0 2 4 6 8

cence spectrophotometric detector (Applied Biosystems, Ceq, mg/L

Ramsey, New Jersey, USA) equipped with a 150 W xenon

lamp (kexcitation = 333 nm and kemission = 470 nm) and Fig. 1. Experimental OA adsorption on yeast by-products.

PC1000 integration software (Thermo Separation Product,

Fremont, California, USA). The analytical column was an

Alltima reversed-phase column C18 (25 cm 4.6 mm i.d., sorbed amounts. In this study, seven adsorption models

5 lm particles) preceded by an Alltima C18 precolumn often reported in the literature were tested to provide the

(7.5 mm 4.6 mm i.d., 5 lm particles) (Alltech, Temple- best description of toxin adsorption, namely Freundlich

mars, France). The columns were left at room temperature. (F), Langmuir (L), BrunauerEmmettTeller (BET), Hill

The mobile phase was a mixture of HPLC grade aceto- (H), RedlichPeterson (RP), RadkePrausnitz (RkP) and

nitrile:water:acetic acid (45:54:1), ltered through a Toth (T). These models are presented below (Eqs. (2)

0.22 lm lter membrane, degassed and used at a ow-rate (17)). It should be noted that for Freundlich, Langmuir,

of 1.0 ml/min. Retention time was 7.5 min. BET and Hill models, the set of isotherm parameters was

The quantity of OA adsorbed was determined by the calculated not only by linearization but also in non-linear

following equation: forms, while the other models were only evaluated in their

non-linear form. For the linear models, the coecient of

C 0 C eq V determination (R2) is widely used in assessment of isotherm

Qeq 1

m accuracies and is commonly determined by the least

squares method (Dagnelie, 1998). For the non-linear

Qeq quantity of OA adsorbed per gram of adsorbent forms, the various constants of these models were calcu-

(mg/g) lated by optimising the sum of the squares of the errors

C0 initial concentration of OA in solution (mg/L) (see Eq. (18)) using the solver add-in from Microsofts

Ceq residual toxin concentration at equilibrium (mg/L) spreadsheet, Excel (Microsoft, 2001). The sum of the

V volume of solution (L) squares of the errors in the non-linear models was used

m mass of adsorbent (g) as an error indicator analogous to the coecient of

determination in the linear models.

The maximum error calculated for the OA concentration For all models, the isotherm set parameter values are

in the supernatant was 0.259 mg/L for EX16 experiments, presented in Tables 13. Their corresponding linear and

0.315 mg/L for BETA and 0.031 mg/L for LEC. non-linear isotherms are illustrated in Figs. 24 and 57

respectively.

3. Results and discussion

3.1.1. Freundlich empirical model

3.1. Equilibrium studies The empirical Freundlich (Freundlich, 1906) equation

based on sorption onto a heterogeneous surface is given

Adsorption equilibrium is established when the quantity by Eq. (2).

of the toxin being adsorbed (Qeq) is equal to the quantity

Qeq K F C 1=nF

2

being desorbed. Then, the equilibrium concentration in eq

solution (Ceq) remains constant. where KF and nF are the Freundlich constants charac-

The plotting of Qeq = f(Ceq) for all adsorption experi- teristic of the system. KF and nF are indicators of adsorp-

ments is presented in Fig. 1. Adsorbent EX16 was able to tion capacity and adsorption intensity, respectively. The

bind 3243% of the initial OA, BETA adsorbed 3751% linearized form of the Freundlich equation is given in

and LEC was a very good adsorbent, able to bind 95 Eq. (3).

100% of the initial OA.

Many equations have been published to describe the 1

ln Qeq ln K F ln C eq 3

equilibrium relationship between adsorbed and unad- nF

D. Ringot et al. / Bioresource Technology 98 (2007) 18121821 1815

where

2.1430E + 00

1.0489E04

6.3800E03

1.6513E01

2.0480E02

3.0620E01

4.5782E01

5.1636E01

6.9748E01

1.0313E+00

1.7238E+00

Ceq residual toxin concentration at equilibrium (mg/L)

2.228

3.275

0.297

Radske-Prausnitz Toth

nF Freundlich adsorption constant

1.1685E + 00

Qeq adsorbed toxin quantity per gram of biomass (mg/g)

3.5948E05

2.8338E03

4.3922E01

5.6592E02

1.2376E02

1.3601E01

1.9499E01

3.1204E01

4.6883E01

1.1587E+00

A plot of ln(Qeq) versus ln(Ceq) should indicate a

0.018

0.387

0.552

straight line of slope 1/nF and intercept ln KF.

6.3231E05

4.2593E03

6.5255E01

9.9544E02

1.6416E02

2.0442E01

2.8970E01

4.1389E01

5.7472E01

1.4204E+00

3.1.2. Langmuir model

The Langmuir (Langmuir, 1916) model is valid for

0.020

0.161

1.000

RP

6.6124E06

8.3289E04

1.6161E01

6.3976E01

1.0410E02

5.2780E03

3.9973E02

7.1746E02

1.3307E01

2.5886E01

model is given by the following Eq. (4), which can be

linearized as in Eq. (5).

0.071

3.950

1.751

Hill

Q K L C eq

Qeq max 4

1 K L C eq

1.0000E + 00

1.0000E + 00

1.0000E + 00

1.0000E + 00

1.0000E + 00

6.3521E04

2.0836E02

3.9663E02

2.2525E+00

2.4714E+00

4. 28.851

where Qeq (mg/g) and Ceq (mg/L) are the amount of adsor-

bent per unit weight of biomass and the unadsorbed toxin

0.015

BET

is the maximum amount of toxin per unit weight of bio-

6.3231E05

4.2583E03

6.5235E01

9.9544E02

1.6416E02

2.0437E01

2.8961E01

4.1388E01

5.7464E01

1.4202E+00

Langmuir

related to the anity of the binding sites.

0.161

Non-linear models

0.23

C eq 1 C eq

5

1.2879E04

8.0123E03

2.0275E01

2.2476E02

3.8454E01

6.0604E01

5.6667E01

7.7939E01

1.3651E+00

1.9262E+00

Freundlich

0.017

1.391

Ceq residual toxin concentration at equilibrium (mg/L)

1.4166E05

5.4398E04

2.9507E02

2.9417E01

2.2301E02

5.7813E03

2.6107E02

1.3100E02

1.1903E01

1.4576E01

Qmax maximum specic uptake corresponding to sites

0.996

0.078

4.054

1.470

saturation (mg/g)

Hill

6.3470E04

1.9625E02

9.9920E01

3.9027E02

9.4188E01

8.5044E01

9.3918E01

9.8396E01

1.9156E+00

2.3211E+00

line of slope 1/Qmax and intercept of 1/KLQmax.

Isotherm parameters and error functions for OA adsorption onto EX16

26.036

0.867

0.015

BET

For the signication of isotherm parameters see Eqs. (2)(17).

1.9406E + 00 1.0413E + 00

8.7231E05

2.6211E03

2.2285E01

8.2689E01

1.3733E01

1.3669E02

1.2580E01

9.8934E02

3.3458E01

3.4463E01

Qeq K H C eq 6

Langmuir

0.982

0.150

0.109

Linear models

3.5272E04

6.4523E03

1.4662E01

7.5132E01

5.5529E01

2.8023E02

3.0967E01

6.5092E02

3.0400E01

7.0654E01

Freundlich

applied model in studies of gassolid equilibrium. This

0.974

0.013

to describe adsorption phenomena when successive mole-

Q0max , Qmax H,

KD, ar, R, At

nH, br, p, t

HYBRID

KF, Qmax,

Kr, A, Kt

ERRSQ

monolayer.

MPSD

EABS

AER

SNE

R2

Q0max C BET C eq

parametersa

Qeq 7

C s C eq 1 C BET 1 C eq =C s

function

Isotherm

Table 1

C eq 1 C BET 1 C eq

Error

8

Qeq C s C eq Q0max C BET Q0max C BET C s

a

1816

Table 2

Isotherm parameters and error functions for OA adsorption onto beta-glucanes

Linear models Non-linear models

Freundlich Langmuir BET Hill Freundlich Langmuir BET Hill RP Radske- Toth

Prausnitz

Isotherm R2 0.961 0.920 0.915 0.940

parametersa KF, Qmax, 0.016 0.134 0.019 0.174 0.015 3.237 0.024 0.766 1.052 0.017 22.644

Q0max ,

Qmax H, Kr,

A, Kt

nF, KL, 1.030 0.169 9.912 8.004 0.964 0.005 9.549 51.324 67.378 1.051 48.190

CBET, KD,

ar, R, At

nH, br, p, t 1.040 1.087 0.000 0.571 0.538

Error ERRSQ 1.0550E04 4.7852E04 1.0284E04 2.6054E04 1.0011E04 1.0593E04 2.6720E05 1.1239E04 1.0186E04 1.9465E04 1.1074E04

function HYBRID 4.5149E03 9.7567E03 5.7889E03 7.5246E03 4.5383E03 4.4048E03 5.2701E03 4.8674E03 4.3496E03 5.5218E03 4.5287E03

MPSD 3.3860E01 9.4996E01 9.0677E01 7.4106E01 2.8610E01 3.2898E01 1.4667E+00 2.8242E01 3.1662E01 4.1335E01 3.7048E01

AER 1.1492E+00 1.4792E+00 1.5350E+00 1.3612E+00 1.0789E+00 1.1559E+00 1.4283E+00 1.0645E+00 1.1410E+00 1.2667E+00 1.1850E+00

EABS 1.8329E02 3.2715E02 2.0782E02 2.6770E02 1.9104E02 1.9760E02 8.5830E03 2.0140E02 1.9451E02 2.5957E02 1.9671E02

2.2046E01 1.0000E + 00 2.1492E01 5.4447E01 2.0920E01 2.2137E01 5.5839E02 2.3488E01 2.1286E01 4.0678E01 2.3142E01

4.6275E01 1.0000E + 00 5.9333E01 7.7122E01 4.6514E01 4.5147E01 5.4015E01 4.9888E01 4.4581E01 5.6594E01 4.6416E01

2.3086E01 6.4770E01 6.1825E01 5.0527E01 1.9507E01 2.2431E01 1.0000E + 00 1.9256E01 2.1588E01 2.8183E01 2.5260E01

7.4863E01 9.6362E01 1.0000E + 00 8.8678E01 7.0282E01 7.5299E01 9.3046E01 6.9347E01 7.4330E01 8.2516E01 7.7194E01

5.6027E01 1.0000E + 00 6.3525E01 8.1829E01 5.8395E01 6.0400E01 2.6236E01 6.1561E01 5.9456E01 7.9344E01 6.0128E01

SNE 2.2230E + 00 4.6113E + 00 3.0617E + 00 3.5260E + 00 2.1562E + 00 2.2541E + 00 2.7888E + 00 2.2354E + 00 2.2124E + 00 2.8732E + 00 2.3214E + 00

a

For the signication of isotherm parameters see Eqs. (2)(17).

Table 3

Isotherm parameters and error functions for OA adsorption onto LEC

Linear models Non-linear models

Freundlich Langmuir BET Hill Freundlich Langmuir BET Hill RP Radske- Toth

Prausnitz

Isotherm R2 0.896 0.981 0.997 0.964

parametersa KF, Qmax, Q0max , 0.202 0.109 0.045 1.892 0.897 26.275 0.040 18.706 4.296 0.593 0.795

Qmax H,

Kr, A, Kt

nF, KL, CBET, 2.048 12.460 22.750 4.854 0.732 0.021 19.924 20.063 6.846 8.719 9.298

KD, ar, R, At

nH, br, p, t 0.762 1.389 0.000 1.146 5.976

Error ERRSQ 5.0766E03 8.7417E03 7.4881E05 2.3796E03 1.5754E03 1.6555E03 7.0877E05 1.5848E03 1.6470E03 1.6436E03 1.6471E03

function HYBRID 3.6971E02 4.9814E02 2.3850E03 1.9182E02 4.3355E02 2.6388E02 2.5069E03 4.4200E02 2.6430E02 2.7106E02 2.6431E02

MPSD 6.1672E01 3.4450E01 9.5714E02 1.9502E01 1.9014E+00 7.9567E01 1.2103E01 1.9471E+00 8.0085E01 8.4852E01 8.0084E01

AER 1.5035E+00 8.1778E01 5.3580E01 7.9404E01 2.6561E+00 1.8442E+00 6.2723E01 2.6815E+00 1.8490E+00 1.8980E+00 1.8490E+00

EABS 1.0070E01 9.9503E02 1.3048E02 7.4544E02 8.1587E02 7.8147E02 1.5304E02 8.1713E02 7.8002E02 7.8521E02 7.8001E02

5.8073E01 1.0000E + 00 8.5660E03 2.7221E01 1.8021E01 1.8938E01 8.1079E03 1.8129E01 1.8841E01 1.8801E01 1.8842E01

7.4217E01 1.0000E + 00 4.7877E02 3.8508E01 8.7033E01 5.2973E01 5.0324E02 8.8730E01 5.3057E01 5.4413E01 5.3059E01

3.1674E01 1.7693E01 4.9158E02 1.0016E01 9.7652E01 4.0865E01 6.2160E02 1.0000E + 00 4.1131E01 4.3579E01 4.1130E01

5.6070E01 3.0497E01 1.9981E01 2.9612E01 9.9054E01 6.8775E01 2.3391E01 1.0000E + 00 6.8953E01 7.0783E01 6.8952E01

1.0000E + 00 9.8812E01 1.2958E01 7.4027E01 8.1020E01 7.7605E01 1.5198E01 8.1146E01 7.7461E01 7.7976E01 7.7460E01

SNE 3.2003E + 00 3.4700E + 00 4.3499E01 1.7938E + 00 3.8278E + 00 2.5916E + 00 5.0648E01 3.8801E + 00 2.5944E + 00 2.6555E + 00 2.5944E + 00

a

For the signication of isotherm parameters see Eqs. (2)(17).

1817

1818 D. Ringot et al. / Bioresource Technology 98 (2007) 18121821

0.09 0.08

0.08 0.07

0.07

0.06

0.06

0.05

0.05 Experimental

Qeq, mg/g

Qeq, mg/g

0.04 Freundlich

0.04

Experimental Langmuir

0.03 0.03 BET

Freundlich

0.02 Langmuir Hill

0.02

BET Redlich-Peterson

0.01

Hill 0.01 Toth

0 RadKe-Prausnitz

0 2 4 6 8 0

Ceq, mg/L 0 2 4 6 8

Ceq, mg/L

Fig. 2. Linear isotherms for OA adsorption onto EX16.

Fig. 5. Non-linear isotherms for OA adsorption onto EX16.

0.10

0.09 0.09

0.08 0.08

0.07 0.07

0.06

0.06

Qeq, mg/g

Experimental

Qeq, mg/g

0.05

0.05 Freundlich

0.04 Experimental

0.04 Langmuir

0.03 Freundlich

BET

Langmuir 0.03

0.02 Hill

BET

0.01 Hill

0.02 Redlich-Peterson

0.00 0.01 Toth

0 1 2 3 4 5 6 RadKe-Prausnitz

0

Ceq, mg/L

0 1 2 3 4 5 6

Fig. 3. Linear isotherms for OA adsorption onto BETA. Ceq, mg/L

0.20

0.18

0.20

0.16

0.18

0.14

0.16

0.12

0.14

Qeq, mg/g

0.10

0.12

Qeq, mg/g

Experimental

0.08

Experimental 0.10 Freundlich

0.06 Freundlich 0.08 Langmuir

0.04 Langmuir BET

0.06 Hill

BET

0.02 0.04 Redlich-Peterson

Hill

0.00 0.02 Toth

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 RadKe-Prausnitz

0.00

Ceq, mg/L

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35

Ceq, mg/L

Fig. 4. Linear isotherms for OA adsorption onto LEC.

Fig. 7. Linear isotherms for OA adsorption onto LEC.

CBET BET adsorption constant relating to the energy of

interaction with the surface (L/mg)

Ceq residual toxin concentration at equilibrium (mg/L) Cs saturation concentration of the solute correspond-

Qeq adsorbed toxin quantity per gram of biomass (mg/ ing to monolayer saturation (mg/L)

g)

Q0max maximum specic uptake corresponding to mono- A plot of Ceq/[Qeq*(Cs Ceq)] versus Ceq/Cs allows

layer saturation (mg/g) calculation of CBET and Q0max .

D. Ringot et al. / Bioresource Technology 98 (2007) 18121821 1819

3.1.4. Non-ideal competitive adsorption (NICA) model This model has three isotherm parameters, Kr, ar and br

At the origin of the NICA (Koopal et al., 1994, 2001) (0 < br < 1) which characterize the isotherm.

model, it was postulated that an equation like the Hill When br = 1, then the equation becomes similar to the

equation (Hill, 1910) could be used to describe the binding Langmuir model (Eq. (13)).

of dierent species onto a homogeneous substrate. This

K r C eq

model assumes that the adsorption is a cooperative phe- Qeq 13

nomenon due to the ability of ligand binding at one site 1 ar C eq

on a macromolecule to inuence ligand binding at a dif-

When br = 0, then it becomes similar to Henrys law (Eq.

ferent site on the same macromolecule.

(14)).

The Hill equation is:

K r C eq

Qmax H C nH

eq Qeq 14

Qeq 9 1 ar

K D C nH

eq

This equation can be linearized through logarithmic terms 3.1.6. The RadkePrausnitz model

(Eq. (10)) The RadkePrausnitz (Radke and Prausnitz, 1972)

! isotherm model is depicted in Eq. (15).

Qeq

ln nH ln C eq ln K D 10 A R C peq

Qmax H Qeq

Qeq 15

A R C p1

eq

where

This equation can be converted to a linear form (Eq. (16))

K D K nH

d 11 !

1 1 1

Ceq residual toxin concentration at equilibrium (mg/L) ln ln R p ln C eq 16

Qeq A C eq

Qeq adsorbed toxin quantity per gram of biomass (mg/

g)

KD Hill constant Ceq residual toxin concentration at equilibrium (mg/L)

Kd dissociation constant per site (mg/L), Kd is equal Qeq adsorbed toxin quantity per gram of biomass (mg/

to residual toxin concentration at half saturation, g)

K d C 50 A constant of RadkePrausnitz isotherm (L/g)

eq ; also, Kd = 1/Ka, where Ka is the associ-

ation constant R constant of RadkePrausnitz isotherm, (mg/

Qmax H maximum specic uptake corresponding to sites g) * (mg/L)p

saturation (mg/g) p constant of RadkePrausnitz isotherm

nH Hill cooperativity coecient of the binding inter-

action

3.1.7. Toth model

Thus, three possibilities can occur The Toth isotherm (Toth, 1971) is derived from the

potential theory and is applicable for heterogeneous

nH > 1, positive cooperativity in binding, adsorption. It assumes a quasi-Gaussian energy distribu-

nH = 1, non-cooperative or hyperbolic binding, tion. Most sites have an adsorption energy lower than

nH < 1, negative cooperativity in binding. the maximum adsorption energy (Eq. (17))

K t C eq

Qeq 17

3.1.5. RedlichPeterson model At C teq 1=t

This model (Redlich and Peterson, 1958) approaches

the Freundlich model at high adsorbate concentration (it

describes equilibrium on heterogeneous surfaces and hence Ceq residual toxin concentration at equilibrium (mg/L)

does not assume monolayer capacity) and Henrys law at Qeq adsorbed toxin quantity per gram of biomass (mg/

g)

low concentration (Eq. (12)).

Kt constant of Toth isotherm (mg/g)

K r C eq At constant of Toth isotherm (L/mg)

Qeq 12

1 ar C beqr t constant of Toth isotherm

Qeq adsorbed toxin quantity per gram of biomass (mg/ 3.2. Error analysis

g)

Kr constant of RedlichPeterson isotherm (L/g) The classical method to determine isotherm parameters

ar constant of RedlichPeterson isotherm (L/mg)br is to use linear regression with transformed variables.

br constant of RedlichPeterson isotherm Linearization of the isotherm equation alters the error

1820 D. Ringot et al. / Bioresource Technology 98 (2007) 18121821

structure and normality assumption of standard least For EX16, the data presented in Table 1 show, based on

squares. This could explain earlier observations that Fre- the sum of normalised errors, that the Hill model pro-

undlich isotherms produced better results at low concen- duced the best goodness of t. Furthermore, for the linear

tration and Langmuir isotherms tend to t data better equations, the highest coecients of determination were

for high concentration (Richter et al., 1989). Recently, also obtained using the Hill model. Since the value of

the utilisation of sum of normalised errors (SNE) calcula- the Hill coecient nH is higher than 1, the binding of

tion procedure in adsorption studies has been presented OA by EX16 is a positive cooperative interaction. For this

by several authors (Allen et al., 2003; Ho et al., 2002; Por- model, the values of isotherm constants are comparable

ter et al., 1999). In these publications, a non-linear com- when they are calculated by linear and non-linear

plex optimisation procedure was performed for the equations.

estimation of the SNE values. In the present study, ve For BETA, the SNE values presented in Table 2 show

error functions were examined to determine and evaluate that the non-linear Freundlich model produced the best

the t of isotherm models to the experimental data t across the range of non-linear models. Moreover, in lin-

(Eqs. (18)(22)) and only the sum of the squares of the ear models, the best value of linear R2 was also determined

errors (ERRQS) was optimised as a relevant error indica- for the Freundlich model. Based on the R2 and on the SNE

tor comparable to the coecient of determination in the values, the Freundlich model in its linear and non-linear

linear models. forms produced a reasonable model for OA biosorption

onto beta-glucans. The magnitude of R2 is an indication

1. The sum of the squares of errors (ERRSQ) of the relative quality of t of the linear isotherm. The fact

X

p that the nF value (in the Freundlich model) is very close to

2

Qe;meas Qe;calc i 18 1 means that this model becomes Henrys law. Then, OA

i1 adsorption onto BETA is a simple linear solid/liquid parti-

2. The hybrid fractional error function (HYBRID) tion which follows Henrys law.

" # For LEC, the data presented in Table 3 show that, based

Xp 2

Qe;meas Qe;calc on the SNE values, the BET model is the most appropriate

19 model for the biosorption of OA on this yeast cell product.

Qe;meas

i1 i Furthermore, the highest value of R2 was also calculated

3. A derivative of Marquardts percent standard deviation for the BET model. Based on the R2 values and on the

(MPSD) SNE value, the BET model appears to be a very good

!2 model for OA biosorption onto LEC. For this model,

Xp

Qe;meas Qe;calc the values of isotherm constants calculated with the linear

20

i1

Qe;meas equation are very close to those obtained using non-linear

i

calculation.

4. The average relative error (AER) The various proles of toxin adsorption onto the three

Xp adsorbents studied suggest dierent mechanisms of ad-

Qe;meas Qe;calc

21 sorption according to their dierent compositions. In a

Q

i1 e;meas i

previous study (Ringot et al., 2005), a thermodynamic

5. The sum of absolute errors (EABS) approach to characterization of the binding of OA onto

X

p

EX16, BETA and LEC was presented, enabling suggestion

Q Qe;calc i 22 of some hypotheses regarding the mechanism of OA yeast

e;meas

i1 cell wall biosorption. In order to investigate these hypoth-

eses in future research, it is proposed to study the inuence

The process of minimising the respective error functions of adsorption conditions (pH and adsorbent mass) on the

across the range of experimental concentrations permitted biosorption phenomenon.

calculation of the isotherm constants. The values of the

errors obtained for each error function for each set of lin- 4. Conclusions

ear and non-linear isotherm constants were normalised by

dividing by the highest value for that error function across The present work allowed identication, among the

a range of dierent isotherms. These normalised errors most commonly described models in the literature, of the

were added for each parameter set to get the SNE. A com- isothermal equation which best tted the adsorption of

parison of the various SNE could thus be undertaken and OA onto each of the three yeast by-products studied. Lin-

allowed the identication of the best set of isotherm con- ear and non-linear optimisation techniques were applied to

stants to describe the measured data. The parameter set determine isotherm parameters. For the best models iden-

thus providing the smallest normalised error sum was con- tied, the non-linear calculation of isotherm parameters

sidered to be optimal. presented in this paper produced comparable data to those

The values of all these minimum error functions and of obtained using the linear method based on the least squares

the optimum isotherm constants are listed in Tables 13. calculation.

D. Ringot et al. / Bioresource Technology 98 (2007) 18121821 1821

Based on the error analysis, the Hill, Freundlich and IARC, 1993IARC Monographs on the Evaluation of Carcinogenic Risks

BET equations, in both their linear and non-linear forms, to Humans: Some Naturally Occurring Substances, Food Items and

Constituents, Heterocyclic Aromatic Amines and Mycotoxins, vol. 56.

appear to be appropriate models for OA biosorption onto International Agency for Research on Cancer, Lyon, France, pp. 489

EX16, BETA and LEC, respectively. For these best mod- 521.

els, the values of isotherm constants were very close when Koopal, L.K., Van Riemsdijk, W.H., de Wit, J.C.M., Benedetti, M.F.,

calculated using linear and non-linear equations. 1994. Analytical isotherm equation for multicomponent adsorption to

These results indicate that the linear equation analysis heterogeneous surfaces. J. Colloid Interface Sci. 166, 5160.

Koopal, L.K., Van Riemsdijk, W.H., Kinniburgh, D.G., 2001. Humic

using the R2 calculation associated with the SNE calcula- matter and contaminants. General aspects and modelling ion binding.

tion procedure presented in this work is an appropriate Pure Appl. Chem. 73, 20052016.

method to use for the study of OA adsorption onto yeast Langmuir, I., 1916. The adsorption of gases on plane surface of glass,

by-products. mica and platinum. J. Am. Chem. Soc. 40, 13611403.

Li, S., Marquardt, R.R., Frohlich, A.A., 2000. Identication of ochra-

toxins and some of their metabolites in bile and urine of rats. Food

Acknowledgements

Chem. Toxicol. 38, 141152.

Microsoft Corporation, 2001. Users guide, Microsoft Excel Ver. 2002.

The authors thank Mr. Renaud Trouve from Serendi Muller, G., Kielstein, P., Rosner, H., Berndt, A., Heller, M., Kohler, H.,

Ltd., Mr. Eric Oriol from Bio-Springer Ltd. and Mrs. Pau- 1999. Studies on the inuence of ochratoxin A on immune and defense

line Anton-Gay from ISAB for their scientic interest in reactions in weaners. Mycosis 42, 495505.

Natour, R.M., Yousef, S.M., 1998. Adsorption eciency of diato-

this study.

maceous earth for mycotoxins. Arab Gulf J. Sci. Res. 16, 113E

127E.

References Porter, J.F., McKay, G., Choy, K.H., 1999. The prediction of sorption

binary mixture of acidic dyes using single- and mixed-isotherm

Aksu, Z., 2003. Reactive dye bioaccumulation by Saccharomyces cerevi- variants of the ideal adsorbed solute theory. Chem. Eng. Sci. 54,

siae. Process Biochem. 38, 14371444. 58635885.

Aksu, Z., Donmez, G., 2002. A comparative study on the biosorption Prelusky, D.B., Rotter, B.A., Rotter, R.G., 1994. Toxicology of myco-

characteristics of some yeast for Remazol Blue reactive dye. Chemo- toxins. In: Miller, Treholm (Eds.), Mycotoxins in Grain, Compounds

sphere 50, 10751083. other than Aatoxins. Eagan Press, St. Paul, MN USA, pp. 359

Aksu, Z., Yenner, J., 1998. Investigation of the biosorption of phenol and 403.

monochlorinated phenols on the dried activated sludges. Process Radke, C.J., Prausnitz, J.M., 1972. Adsorption of organic solutes from

Biochem. 33, 483491. dilute aqueous solutions on activated carbon. Ind. Eng. Chem.

Allen, S., Gan, Q., Matthews, R., Johnson, P.A., 2003. Comparison of Fundam. 11, 445450.

optimised models for basic dye adsorption by kudzu. Bioresource Ramos, A.J., Fink-Gremmels, J., Hernandez, E., 1996. Prevention of toxic

Technol. 88, 143152. eects of mycotoxins by means of non-nutritive adsorbent compounds.

Bakker, M., Pieters, M.N., 2002. Risk assessment of ochratoxin A in J. Food. Prot. 59, 631646.

Netherlands. RIVM Report 388802025. Inspectorate for Health Redlich, O., Peterson, C., 1958. Useful adsorption isotherm. J. Phys.

Protection and Veterinary Public Health. Bilthoven, Netherlands, 24p. Chem. 63, 1024.

Bauer, J., 1994. Moglichkeiten zur Entgiftung mycotoxin-haltiger, Fut- Richter, E., Schultz, W., Myers, A.L., 1989. Eect of adsorption equation

termittel. Monatssh. Veterinarmed. 49, 175181. on prediction of multicomponent equilibria by the ideal adsorption

Bruanuer, S., Emmett, P.H., Teller, E., 1938. Adsorption of gases in multi- solution theory. Chem. Eng. Sci. 44, 16091616.

molecular layers. J. Am. Chem. Soc. 60, 309316. Ringot, D., Lerzy, B., Bonhoure, J.P., Auclair, E., Oriol, E., Larondelle,

Castellari, M., Versari, A., Fabiani, A., Parpinello, G.P., Galassi, S., 2001. Y., 2005. Eect of temperature on in vitro ochratoxin A biosorption

Removal of ochratoxin A in red wines by means of adsorption onto yeast cell derivatives. Process Biochem. 40, 30083016.

treatments with commercial ning agents. J. Agric. Food Chem. 49, Ringot, D., Chango, A., Schneider, Y.J., Larondelle, Y., 2006. Toxico-

39173921. kinetics and toxicodynamics of ochratoxin A: an update. Chem. Biol.

Commission of the European Union, 2002. Report of experts participating Interact. 159, 1846.

in task 3.2.7. Assessment of dietary intake of ochratoxin A by the Scott, P.M., 1998. Industrial and farm detoxication process for myco-

population of EU Member States. 153p. toxins. Rev. Med. Vet. 149, 543548.

Dagnelie, P., 1998. Statistique theorique et appliquee. vol. 1, De Boecket Tomasevic-Canovic, M., Dakovic, A., Rottinghaus, G., Matijasevic, S.,

Larcier, Paris-Bruxelles, Belgium pp. 130148. Duricic, M., 2003. Surfactant modied zeolites new ecient

Freundlich, H.M.F., 1906. Uber die Adsorption in Losungen. Z. Phys. adsorbent for mycotoxins. Micropor. Mesopor. Mater. 61, 173

Chem. 57, 385470. 180.

Galvano, F., Pietri, A., Bertuzzi, T., Piva, A., Chies, L., Galvano, M., Toth, J., 1971. State equations of the solid gas interface layer. Acta Chem.

1998. Activated carbons: In vitro anity for ochratoxin A and Acad. Hung. 69, 311317.

deoxynivalenol and relation of adsorption ability to physicochemical Tsezos, M., Bell, J.P., 1989. Comparison of the biosorption and

parameters. J. Food Prot. 61, 469475. desorption of hazardous organic pollutants by live and dead biomass.

Grant, P.G., Philips, T.D., 1998. Isothermal adsorption of aatoxin B1 on Water Res. 23, 561568.

HSCAS clay. J. Agric. Food Chem. 46, 599605. Veglio, F., Beolchini, F., 1997. Removal of metals by biosorption: a

Hill, A.V., 1910. The possible eects of the aggregation of the molecules of review. Hydrometallurgy 44, 301316.

haemoglobin on its dissociation curves. J. Physiol. (London) 40, ivvii. World Health Organisation, 2002. Evaluation of certain mycotoxins in

Ho, Y.S., Porter, J.F., McKay, G., 2002. Equilibrium isotherm studies for food. WHO Technical Report Series, vol. 906. WHO, Geneva,

the sorption of divalent metal ions onto peat: copper, nickel and lead Switzerland, p. 62.

single component systems. Water Air Soil Pollut. 141, 133. Yiannikouris, A., Poughon, L., Cameleyre, X., Dussap, C.J., Francois, J.,

Huwing, A., Freimund, S., Kappeli, O., Dutler, H., 2001. Mycotoxin Bertin, G., Jouany, J.P., 2003. A novel technique to evaluate

detoxication of animal feed by dierent adsorbent. Toxicol. Lett. 122, interactions between Saccharomyces cerevisiae cell wall and myco-

179188. toxins: application to zearalenone. Biotechnol. Lett. 25, 783789.

- tmp1495.tmpЗагружено:Frontiers
- Homework 6Загружено:tarhuni
- AdsorptionЗагружено:toxictaz
- Adsorption Studies of Direct Red 28 Dye Onto Activated CarbonЗагружено:Alexander Decker
- Biosorption and Kinetic Studies on Oil Removal From Produced Water Using Banana PeelЗагружено:Alexander Decker
- Imaga542014IRJPAC13510_1Загружено:Nórida Pájaro Gómez
- Treatment of Oil Spill by Buffing Dust as an Efficient Adsorbent - InESPO 2014Загружено:Gale Ramadan
- Removal from wastewater and recycling of azo textile dyes by alginate-chitosan beadsЗагружено:IJEAB Journal
- tmpF18F.tmpЗагружено:Frontiers
- Journal of Environmental Science and Engineering,Vol.6,No.10A,2017Загружено:Οδυσσεας Κοψιδας
- Studies on the Treatment of Brilliant Green Solution by Combination Microwave Induced Oxidation wЗагружено:Alin Druc
- 1-s2.0-S1878535212002754-main.pdfЗагружено:NataliakusumaDewi
- IctЗагружено:div
- Saponins Clay Modified Materials a New Approach Against Callosobruchus Subinnotatus in Stored ProductsЗагружено:IJSTR Research Publication
- Adsorption of Reactive Black 5 From an Aqueous Solution - Equilibrium and Kinetic StudiesЗагружено:Dr. Arif Hidayat
- molvig courtney adsorptionofaceticacidbyasolidЗагружено:api-405393737
- Ball Clay Untuk Adsorpsi Violet DyeЗагружено:edywiyono2013
- 1Загружено:Mandeep Singh
- Equilibrium Modeling and Kinetic Studies on the Adsorption of Basic Dye by A low-Cost Adsorbent: Sugarcane BagasseЗагружено:Amin Mojiri
- Fe3O4ÔÇôwheatstraw Preparation CharacterizationЗагружено:Cristina Ene
- 23. 01-15 Research Journal of Pharmaceutical, Biological and Chemical Sciences 6(1) Page No. 730.pdfЗагружено:Narendrakumar Gopakumaran
- Information 06 00014Загружено:1982fernan
- carbon activado acacia.pdfЗагружено:Katherin Ricardo
- CT=58(1646-1653)OD12Загружено:Shalini Sriram
- Mittal2008. MYЗагружено:Fatma Maharani
- 1-Batch Studies for Methylene Blue Removal and Recovery by Untreated Coffee ResiduesЗагружено:Οδυσσεας Κοψιδας
- Kinetics PaperЗагружено:Shahul Hameed Kp
- Comparative Study on Biosorption of Pb(..) and Cr(..) ByЗагружено:Yogi Pernanda
- Klp 20Загружено:Febryan Chuang
- Purificacion de Agua Con LevaduraЗагружено:luisbetox

- CFX FSI 14.5 L9 Immersed_Solids 26Загружено:Moh Sen
- Olimpiada Po Angliiskomu Yazyku2Загружено:Regina Iskhakova
- Madrasas and Muslim MilitancyЗагружено:golamhossainju
- ubd-in-a-nutshellЗагружено:api-304509886
- jurnal kustaЗагружено:Robertus Elvantora
- Lab-on-a-chip_06-2013_150dpiЗагружено:Kunal Bose
- Drum-KitЗагружено:Jenna Roubos
- Mongodb on Red hatЗагружено:Luis Contreras
- resume 1Загружено:api-439136131
- choice boardЗагружено:api-315700132
- Md 254 Process ImprovementЗагружено:santoshtandav
- CompanyBrochureЗагружено:Rahul
- Nigeria Charismatic Pentecostal HistoryЗагружено:Richard Oluseyi Asaolu
- Chemical and Reagent HazardsЗагружено:sarcali
- The Hunter Farmer FishermanЗагружено:Vaibhav Misra
- Ignou Mba Solved Assignments for January 2018Загружено:Dharmendra Singh Sikarwar
- untitleddocumentЗагружено:api-335814742
- IC4_L1_WQ_U7to8Загружено:Paul Sebastian
- Civil Compromise Agreement of Brian Cole and VictimЗагружено:Christopher Christie
- Advanced Processor Superscalarclass PptЗагружено:Kanaga Varatharajan
- Icecat CSV Documentation Ver 1.9Загружено:higgs1
- Petri NetЗагружено:Munawar Zaman
- Itil Dumps Question and AnswersЗагружено:Prasad Hassan ಪ್ರಸಾದ್ ಹಾಸನ್
- SQL JoinsЗагружено:Akira Kei
- Understand Arabic in 12 Coloured TablesЗагружено:speed2kx
- Euromedlab 2013 abstracts 211 and 219Загружено:Jagannadha Rao Peela
- Electromagnetic PropulsionЗагружено:pras_scribd
- 1a GM, Activos IntangibleЗагружено:Juan Sebastian Quintero Botero
- 310 Discussion on Pile StiffnessЗагружено:magdyamdb
- Drawing on Burlesque Excessive Display and Fat Desire in the Work of Cristina Vela_Jamie Ratcliff.pdfЗагружено:jen-lee