Journal of Chromatography A, 1195 (2008) 117126

Contents lists available at ScienceDirect

Journal of Chromatography A
journal homepage: www.elsevier.com/locate/chroma

Enantiomer identication in the avour and fragrance elds by interactive combination of linear retention indices from enantioselective gas chromatography and mass spectrometry
Erica Liberto a , Cecilia Cagliero a , Barbara Sgorbini a , Carlo Bicchi a , Danilo Sciarrone b , Barbara DAcampora Zellner b , Luigi Mondello c , Patrizia Rubiolo a,
a

Dipartimento di Scienza e Tecnologia del Farmaco, Facolt` a di Farmacia, Universit` a degli Studi di Torino, Via Pietro Giuria 9, Turin 10125, Italy Dipartimento Farmaco-Chimico, Facolt` a di Farmacia, Universit` a degli Studi di Messina, Viale Annunziata, Messina 98168, Italy c Campus-Biomedico, Via E. Longoni 47, Rome 00155, Italy
b

a r t i c l e

i n f o

a b s t r a c t
This study describes the development of a gas chromatographymass spectrometry (GCMS) library to identify optically active compounds in the avour and fragrance eld using enantioselective GC with cyclodextrin derivatives (CDs) as chiral selectors in combination with MS. The library operates on the interactive combination of linear retention indices (IT values) in parallel to MS spectra, so as to identify enantiomers reliably and to measure EE and/or ER unequivocally. Since MS is not a selective probe to discriminate optical isomers, mass spectra (or diagnostic ions in SIM mode) are used to locate the enantiomer(s) in the chromatogram, and IT values to identify it(them) safely and reliably in particular in complex mixtures. The library has been built up through the following steps: (a) Selection of CD derivatives able to cover a wide range of racemate separations. Four cyclodextrin derivatives were used: 2,6-di-O-methyl-3-O-pentyl--CD, 2,3-di-O-methyl-6-O-tertbutyldimethylsilyl--CD, 2,3-di-O-ethyl-6-O-tert-butyldimethylsilyl--CD, and 2,3-di-O-acetyl-6-Otert-butyldimethylsilyl--CD. (b) Determination of the analytes IT values and evaluation of their stability and reliability at both intraand inter-laboratory level. (c) Determination of the range within which the IT of an enantiomer has to fall to be correctly identied, i.e. determination of a common retention index allowance (RIA). (d) Construction of the library, at the moment comprising the enantiomers of 134 racemates. A record has been attributed to each enantiomer including IT values determined on the four CD coated columns, mass spectrum, IUPAC chemical name, CAS number, molecular weight, and, when separated, racemate enantiomer resolution on the CD investigated. Some applications of the library are also reported. 2008 Elsevier B.V. All rights reserved.

Article history: Received 19 February 2008 Received in revised form 16 April 2008 Accepted 18 April 2008 Available online 24 April 2008 Keywords: Enantiomer identication Mass spectral library Enantioselective GCMS Cyclodextrin Linear retention indices Mass spectra Flavours Fragrances

1. Introduction The interaction of a compound with a biological system has long been shown to be stereoselective. Enantiomer recognition and enantiomeric excess (EE) and/or ratio (ER) determination are a very important task in avour and fragrance elds, so as (i) to dene the correlation between chemical composition and organoleptic properties; (ii) to implement quality control and detect fraud or adulteration of natural samples; (iii) to determine the biosynthetic pathway when the formation of a compound is studied or

Corresponding author. Fax: +39 011 2367661. E-mail address: patrizia.rubiolo@unito.it (P. Rubiolo). 0021-9673/$ see front matter 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.chroma.2008.04.045

to classify a sample; (iv) to determine the geographic origin of a natural sample. Cyclodextrin derivatives (CDs) have truly represented a milestone in enantioselective gas chromatography (GC); being rst introduced by Sybilska and Koscielski at the University of Warsaw in 1983 for packed columns [1] and applied to capillary columns in the almost contemporary works of Juvancz et al. [2] and Schurig and Nowotny [3]. Moreover, Schurig and co-workers rst proposed diluting CD derivatives in moderately polar polysiloxane (OV-1701) to provide them with good chromatographic properties and a wider range of operative temperatures [4]. Since then, several groups have investigated CD derivatives as chiral selectors for enantioselective GC applications, and several hundreds of articles have been published dealing with the theory of chiral

118

E. Liberto et al. / J. Chromatogr. A 1195 (2008) 117126

GC recognition with CDs, synthesis of new CD derivatives, their enantioselectivity and applications, many of them concerning the avour and fragrance eld [58]. Chiral recognition of marker compounds in complex real-world samples, as those in the avour and fragrance eld often are, generally requires a two-dimensional approach, because enantioselective GC may double the number of peaks of optically active analytes. This makes some parts of the total chromatograms even more complex and, as a consequence, increases the probability of interferences with a correct EE and/or ER determination. Two complementary but distinct approaches can therefore be adopted; the rst and most popular one is to introduce a second dimension in separation. Chiral recognition is here generally carried out by conventional heart-cut GCGC [912], where the rst column, coated with a conventional phase, serves to locate the peak(s) of the optically active racemate(s), and the second column, coated with a CD stationary phase, separates its(their) enantiomers after online transfer through the heart-cut interface. When the number of components to be investigated is very high, comprehensive twodimensional gas chromatography (GC GC) can also be used but, in this case, column geometry must be inverted because of the high efciency required by the columns coated with CDs in order to give reliable separations [13,14], the rst column has to be coated with the enantioselective CD stationary phase while the second one distributes the peaks over the chromatographic plane [15,16]. The second approach involves the use of a second dimension in identication. In this case, the enantiomer is located and identied by MS detection (or very rarely FT-IR). Single- or multiple-ion monitoringMS (SIM-MS) carried out after a careful choice of suitable diagnostic ions of the optically active marker(s) of the sample under investigation can be applied to clean the part of the chromatogram where the enantiomers elute, thus making correct EE and/or ER determination possible. In general, most GCMS software takes insufcient account of the identication potential of GC, because the identication power of mass spectrometry when used as detector for GC is considered to be, and very often is, exhaustive. Retention indices (Is ) are the most reliable and effective tool for analyte identication by GC data. They were rst introduced by Kovats for isothermal analysis [17] and then by Van den Dool and Kratz for temperature programmed analysis [18], the latter being better known as linear retention indices (IT values). Most GCMS software packages do not include IT values as identication criterion, and only some of them report IT values in the library as blind or inactive data appearing in the legend of each proposed identication record, making them only useful for further or additional conrmation. On the contrary, the interactive use of IT values (i.e. their use as an active identication parameter) can be highly effective since it provides a second independent tool to identify a compound, operating actively and simultaneously in parallel to MS spectra. Moreover, IT values are based on a chemical property of an analyte of a completely different nature compared to MS, which can orthogonally and synergically be combined with its MS fragmentation pattern, i.e. its chromatographic interaction with a given chromatographic separation system or better with a given stationary phase. Mass spectrometry is well known to be unable to discriminate between optical isomers, not being a selective chiral probe in this sense, and therefore giving indistinguishable spectra. As a consequence it cannot be used alone to determine which enantiomer is present in a sample, or to establish the predominant one or to measure its EE and/or ER. In enantioselective GCMS, a given optical isomer can only unequivocally be identied through its IT obtained with a column coated with a chiral selector suitable to separate it from its enantiomer. In the chiral recognition of optically active isomers in a complex mixture, the two identication parameters

(i.e. IT values and MS spectrum) must therefore be combined but, unlike conventional GCMS analysis, mass spectra (or diagnostic ion monitoring) are used to locate the two enantiomers in the chromatogram, and IT values for their identication. This study deals with the development of a MS library specic for the identication of optically active compounds in the avour and fragrance eld using interactive IT values in parallel to MS spectra, so as to identify enantiomers reliably and, when necessary, to enable the measurement of EE and/or ER unequivocally. 2. Experimental 2.1. Racemate standards and essential oils Analyses of 134 racemate standards and pure enantiomers solubilised in cyclohexane at a concentration of 100 ppm were carried out. The enantiomeric recognition of the marker compounds characteristic of commercially available essential oils (e.o.) and extracts was also carried out, in particular balm lemon, bergamot, boronia, cornmint, lavender, lemon, peppermint, and rosemary e.o.s.; apple avour and apricot, peach and coconut headspace sampled by solid-phase microextraction (SPME) were analysed. 2.2. Enantioselective GC columns The library has been built on the basis of IT values obtained on four columns (25 m 0.25 mm I.D., df : 0.25 m) coated with four cyclodextrin derivatives as chiral selectors diluted at 30% in PS-086: 2,6-di-O-methyl-3-O-pentyl--CD (2,6DM3PEN--CD) [19,20] 2,3-di-O-methyl-6-O-tert-butyldimethylsilyl--CD (2,3DM6TBDMS--CD) [21] 2,3-di-O-ethyl-6-O-tert-butyldimethylsilyl--CD (2,3DE6TBDMS--CD) [22] 2,3-di-O-acetyl-6-O-tert-butyldimethylsilyl--CD (2,3DA6TBDMS--CD) [21] All columns were from MEGA (Legnano, Italy). Their performance were periodically tested through the Grob test [23,24] and a laboratory-made chiral test [25] consisting of limonene, 2-octanol, camphor, isobornyl acetate, linalyl acetate, 2-methyl(3Z)-hexenyl butyrate, menthol, hydroxycitronellal, -decalactone and -decalactone racemates. 2.3. Enantioselective GCMS conditions A Shimadzu QP2010 GCMS system was used and results were elaborated with the Shimadzu GCMS Solution 2.51 software (Shimadzu, Milan, Italy). GC conditions: injection mode, split; split ratio, 1:20 for standard solutions, 1:50 for essential oils; injection volume, 1 l. Temperatures: injection, 220 C; transfer line, 230 C; ion source: 200 C; temperature programme: from 50 to 220 C at 2 C/min (if not specied otherwise). Carrier gas: He; ow rate: 1.0 ml/min. 2.4. Library setting-up The library was created on the basis of the analysis of the 134 racemate standards and pure enantiomers, on each column, recording their mass spectra, calculated IT values and enantiomer resolution (R). The IT determination was carried out by injecting an homologous series of n-alkanes containing 17 n-hydrocarbons (C9 C25 ) purchased from Supelco (Bellefonte, PA, USA), each at

E. Liberto et al. / J. Chromatogr. A 1195 (2008) 117126

119

100 ppm in hexane. The enantiomer stereochemistry was conrmed either through authentic samples or by combining literature data and analysis of essential oils or fruit avour headspaces or extracts. 3. Results and discussion 3.1. Basic approach As already mentioned, the safest approach to identify unequivocally a given optical isomer by enantioselective GCMS, in particular in complex mixtures, is to combine two identication parameters (i.e. IT values and MS spectrum), one of them suitable to distinguish between enantiomers (IT values). Unlike conventional

GCMS analysis, mass spectra (or monitoring by diagnostic ions (SIM)) are used to locate the two enantiomers in the chromatogram, and IT values for their identication. The present study was divided into two main steps: the rst one concerns the building up of the library using the approach described above and the evaluation of its reliability, and the second one involves its application to its everyday use in a routine laboratory. 3.2. Development of the library The library has been created through the following steps (a) choice of a set of chiral selectors (CD derivatives) able to cover a wide range of racemate separations, (b) analyte IT

Fig. 1. Chiral test proles carried out on the four columns investigated. (1) Limonene, (2) 2-octanol, (3) camphor, (4) isobornyl acetate, (5) linalyl acetate, (6) 2-methyl-(3Z)hexenyl butyrate, (7) menthol, (8) hydroxycitronellal, (9) -decalactone, (10) -dodecalactone; (a) (R) enantiomer, (b) (S) enantiomer, x and y: enantiomer conguration not assigned.

120

E. Liberto et al. / J. Chromatogr. A 1195 (2008) 117126

Table 1 IT variation of three marker analytes as a consequence of different injection time intervals of the hydrocarbon standard mixture Hydrocarbon Injection frequency: every week Anal. A1 Anal. A2 Anal. A3 Anal. A4 Anal. A5 2,6DM3PEN--CD (S)-()-Limonene (R)-(+)-Limonene (1R, 2S, 5R)-()-Menthol (1S, 2R, 5S)-(+)-Menthol (R)--Decalactone (S)--Decalactone 2, 3DM6TBDMS--CD (S)-()-Limonene (R)-(+)-Limonene Menthol (R)--Decalactone (S)--Decalactone 2,3DE6TBDMS--CD (S)-()-Limonene (R)-(+)-Limonene (1R, 2S, 5R)-()-Menthol (1S, 2R, 5S)-(+)-Menthol (R)--Decalactone (S)--Decalactone 2,3DA6TBDMS--CD Limonene (1R, 2S, 5R)-()-Menthol (1S, 2R, 5S)-(+)-Menthol (R)--Decalactone (S)--Decalactone 1061 1068 1350 1352 1610 1616 1082 1097 1371 1635 1645 1056 1072 1282 1285 1573 1588 1053 1383 1393 1809 1819 1061 1068 1350 1352 1610 1615 1081 1096 1370 1633 1643 1056 1072 1282 1285 1573 1588 1053 1384 1394 1809 1819 1060 1067 1348 1351 1608 1614 1081 1096 1370 1633 1644 1057 1073 1283 1285 1574 1588 1051 1382 1392 1807 1817 1060 1067 1348 1350 1608 1613 1081 1096 1370 1633 1644 1056 1072 1282 1285 1573 1588 1051 1081 1091 1807 1816 1059 1067 1348 1350 1607 1613 1081 1096 1370 1633 1643 1056 1072 1282 1284 1573 1587 1051 1081 1091 1806 1816 2 1 2 2 3 3 1 1 1 2 2 1 1 0 1 1 1 2 3 3 3 3 IT Hydrocarbon Injection frequency: every 3 days Anal. B1 Anal. B2 Anal. B3 1061 1068 1350 1352 1610 1616 1082 1097 1371 1635 1645 1056 1072 1282 1285 1573 1588 1053 1383 1393 1809 1819 1061 1068 1350 1352 1610 1615 1081 1096 1370 1633 1643 1056 1072 1282 1285 1573 1588 1053 1384 1394 1809 1819 1060 1067 1348 1351 1608 1614 1081 1096 1370 1633 1644 1057 1073 1283 1285 1574 1588 1051 1382 1392 1807 1817 1 1 2 1 2 2 1 1 1 2 2 1 1 0 0 1 0 2 2 2 2 2 IT Hydrocarbon Injection frequency: every ve analyses Anal. C1 Anal. C2 Anal. C3 Anal. C4 Anal. C5 1061 1068 1350 1352 1610 1616 1082 1097 1371 1635 1645 1056 1072 1282 1285 1573 1588 1053 1383 1393 1809 1819 1061 1068 1350 1352 1610 1616 1082 1097 1371 1634 1645 1056 1073 1283 1285 1574 1588 1053 1384 1394 1809 1819 1061 1068 1350 1352 1610 1616 1082 1097 1371 1635 1645 1056 1073 1282 1285 1573 1588 1053 1384 1394 1810 1820 1061 1068 1349 1352 1610 1615 1082 1097 1371 1634 1645 1056 1073 1282 1285 1573 1588 1053 1383 1393 1810 1819 1061 1068 1349 1352 1610 1615 1082 1097 1371 1635 1645 1056 1072 1282 1284 1573 1588 1053 1083 1093 1810 1820 0 0 1 0 0 1 0 0 0 1 0 0 1 1 1 1 0 0 1 1 1 1 IT

T T Section A and B: Is calculated every consecutive days on hydrocarbons injected on the day A1/B1. Section C: Is calculated every consecutive days on hydrocarbons injected each day every ve analysis.

determination and evaluation of their reliability, and (c) denition of a correct procedure to select a suitable retention index allowance (RIA) that included determination of the optimal injection amount and measurement of the average total analyte tailing factor. All results and considerations reported in the present article are based on data resulting from enantioselective GCMS analyses of 134 racemates whose IT values were determined on four columns coated with different CD chiral selectors (see below).
Table 2 Average IT s and column

3.2.1. Selection of the chiral selector set Four cyclodextrin derivatives were chosen as chiral selectors to build up this library (see Section 2.2). The choice of a relatively large number of chiral selectors (columns) was mainly due to the fact that a derivatised CD with a universal enantioselectivity has not yet been found, therefore a number of derivatives suitable to cover most of the usual separation in the avour and fragrance eld had to be used. This lack is due to the intrinsic mechanism of chiral recognition with CD derivatives in gas chromatography that is based on a

IT s of chiral test components carried out over three weeks, on a period of 2 months randomly repeated for three couples of 2 consecutive days for each CD 2,6DM3PEN--CD Week 1 Week 2 Week 3 2,3DM6TBDMS--CD IT Week 1 Week 2 Week 3 0 1 1 0 0 0 0 0 0 1 0 0 0 0 1 1082 1097 1128 1130 1193 1199 1257 1261 1254 1256 1246 1249 1371 1447 1450 1635 1646 1641 1646 1081 1096 1128 1130 1193 1198 1256 1260 1254 1255 1245 1249 1370 1446 1449 1634 1645 1640 1645 1082 1095 1128 1130 1193 1199 1256 1260 1254 1256 1245 1249 1371 1447 1450 1634 1645 1641 1645 1 2 0 0 0 1 1 1 0 0 1 0 1 1 1 1 1 1 1 2,3DE6TBDMS--CD IT Week 1 Week 2 Week 3 1056 1072 1111 1112 1133 1141 1220 1222 1231 1237 1240 1244 1282 1285 1373 1374 1573 1588 1586 1588 1056 1073 1111 1113 1133 1141 1220 1222 1231 1237 1240 1244 1282 1285 1373 1374 1573 1588 1586 1588 1056 1071 1110 1112 1133 1141 1220 1223 1231 1237 1240 1244 1282 1285 1373 1374 1573 1588 1586 1588 0 1 1 1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 2,3DA6TBDMS--CD IT Week 1 Week 2 Week 3 1052 1236 1241 1258 1321 1301 1303 1296 1298 1383 1393 1646 1651 1809 1819 1866 1875 1053 1237 1242 1259 1322 1303 1304 1297 1299 1383 1394 1647 1652 1810 1820 1867 1876 1053 1235 1242 1258 1321 1301 1303 1296 1298 1382 1392 1645 1651 1810 1820 1866 1875 1 2 1 1 1 2 1 1 1 1 2 2 1 1 1 1 1 IT

(S)-()-Limonene (R)-(+)-Limonene (S)-2-Octanol (R)-2-Octanol (S)-()-Camphor (R)-(+)-Camphor (X)-Isobornyl acetate (Y)-Isobornyl acetate (R)-()-Linalyl acetate (S)-(+)-Linalyl acetate (X)-2-Methyl-(3Z)-hexenyl butyrate (Y)-2-Methyl-(3Z)-hexenyl butyrate (1R, 2S, 5R)-()-Menthol (1S, 2R, 5S)-(+)-Menthol (X)-Hydroxycitronellal (Y)-Hydroxycitronellal (R)--Decalactone (S)--Decalactone (S)--Decalactone (R)--Decalactone

1061 1068 1147 1184 1269 1273 1240 1243 1245 1350 1352 1438 1610 1616 1637

1061 1068 1146 1184 1269 1273 1240 1243 1245 1349 1352 1438 1610 1616 1636

1061 1067 1146 1184 1269 1273 1240 1243 1245 1349 1352 1438 1610 1616 1636

Analyses were repeated three times each day for a total of 18 determinations for each column.

E. Liberto et al. / J. Chromatogr. A 1195 (2008) 117126 Table 3 RIA window values expressed as average start-apex and apex-stop IT variation of the groups of racemate analysed with each CD column Class 2,6DM3PEN--CD Start-apex Esters Alcohols Ketones Acids Hydrocarbons Aldehydes Lactones Mean 1.6 1.6 1.6 1.5 1.1 2.0 1.8 1.6 Apex-stop 1.4 1.8 1.7 2.4 1.8 2.0 2.0 1.9 2,6DM6TBDMS--CD Start-apex 1.3 1.3 1.5 1.1 1.5 1.0 1.5 1.3 Apex-stop 1.8 1.8 2.1 2.6 1.5 1.5 2.2 1.9 2,6DE6TBDMS--CD Start-apex 1.3 1.2 1.4 1.1 1.3 1.0 1.6 1.3 Apex-stop 1.7 1.5 1.5 4.3 2.0 1.0 2.5 2.1 2,6DA6TBDMS--CD Start-apex 1.4 1.4 1.5 1.3 1.3 1.0 1.8 1.4 Apex-stop 1.3 1.9 1.6 4.8 1.6 1.5 1.8 2.1 1.4 1.4 1.5 1.3 1.3 1.2 1.7 1.4 1.6 1.8 1.7 3.5 1.7 1.5 2.1 2.0 Average start-apex RIA

121

Average apex-stop RIA

hostguest interaction of each enantiomer of a racemate with the CD selector, depending the separation on the rather low difference in the energy of interaction of each enantiomer with the CD chiral selector [13,14]. On the basis of the authors experience, laboratories involved with enantioselective GC in the avour and fragrance eld should be provided with at least two columns coated with different CD derivatives, which enables to separate at least 80% of the most common racemates in this eld. The selected columns were periodically tested through the Grob test [23,24] to evaluate their chromatographic performance over time and a homemade chiral test [25] containing racemates with different volatility, structure and polarity to control their enantioselectivity. Its composition is reported in Section 2.2. Fig. 1 reports the chiral test proles carried out on the four columns investigated. 3.2.2. Analyte IT determination and reliability As mentioned above, IT values are the identifying parameter and can actively and successfully be used only if they are highly stable over time; the repeatability of IT values over time is fundamental for a reliable identication. The development of this library required some fundamental investigations, not necessary in routine work, to evaluate the reliability of the system. The frequency of injection of the hydrocarbon reference mixture to obtain the highest IT repeatability was rst studied. The intra-laboratory IT stability versus the frequency of injection of the hydrocarbons was determined by a series of periodical experiments carried out on some of the components of the chiral test on all columns investigated. Table 1 reports the IT variation of the marker analytes by injecting hydrocarbon
Table 4 Inter-laboratory average IT s and

standard mixture at different intervals. These results clearly show that the highest IT stability is obtained when the hydrocarbons are injected every ve analyte injections. This frequency was then adopted for the creation of the library. Less frequent hydrocarbon injections are required for routine analysis. Control analyses were carried out over a period of 2 months, they were randomly repeated for three couples of 2 consecutive days for each CD column. Analyses were repeated three times each day for a total of 18 determinations for each column. Table 2 reports the average IT values and IT of each enantiomer calculated over the whole period investigated. These results show that IT values are highly repeatable and their IT values never exceed two units. 3.2.3. Determination of the retention index allowance, RIA The RIA window determination, i.e. the range within which the IT of an analyte has to fall to be correctly identied, is a key point for an univocal identication. A friendly operating library should be based on a unique RIA window to be automatically applied to all enantiomers when analysed on all columns investigated. The ideal RIA should be narrow enough to include only one of the two enantiomers; RIAs choice is therefore strongly conditioned by the enantiomer resolution obtained with a given chiral selector and it is critical, in particular, for those enantiomers whose resolution on a given column is below 1.5, i.e. where partial peak overlapping occurs. On the other hand, RIA cannot be too narrow to avoid that the IT of a given analyte falls outside the range because of retention variation. A reliable RIA therefore requires not only highly stable IT values, as mentioned in Sections 3.2.2 and 3.2.4, but also highly inert columns to avoid peak tailing that can affect enantiomer IT

IT s of the chiral test components 2,6DM3PEN--CD Laboratory 1 Laboratory 2 I 1 1 0 1 1 1 1 1 0 0 0 0 1 0 0

T

2, 3DM6TBDMS--CD Laboratory 1 Laboratory 2 1082 1096 1128 1130 1193 1199 1257 1260 1254 1256 1245 1249 1371 1447 1450 1634 1645 1641 1645 1082 1097 1129 1131 1193 1199 1256 1261 1254 1256 1246 1249 1371 1447 1450 1634 1645 1641 1645 0 1 1 1 0 0 1 1 0 0 1 0 0 0 0 0 0 0 0 I
T

2, 3DE6TBDMS--CD Laboratory 1 Laboratory 2 1056 1072 1111 1112 1133 1141 1220 1222 1231 1237 1240 1244 1282 1285 1373 1374 1573 1588 1586 1588 1056 1072 1110 1112 1134 1142 1221 1224 1233 1239 1241 1245 1283 1285 1374 1375 1573 1587 1587 1589 0 0 1 0 1 1 1 2 2 2 1 1 1 0 1 1 0 1 1 1 I
T

2, 3DA6TBDMS--CD Laboratory 1 Laboratory 2 1053 1237 1242 1258 1321 1302 1303 1296 1298 1383 1393 1646 1652 1810 1820 1866 1876 1053 1237 1244 1260 1322 1302 1304 1297 1298 1385 1394 1648 1654 1813 1823 1869 1878 0 0 2 2 1 0 1 1 0 2 1 2 2 3 3 3 2 IT

(S)-()-Limonene (R)-(+)-Limonene (S)-2-Octanol (R)-2-Octanol (S)-()-Camphor (R)-(+)-Camphor (X)-Isobornyl acetate (Y)-Isobornyl acetate (R)-()-Linalyl acetate (S)-(+)-Linalyl acetate (X)-2-Methyl-(3Z)-hexenyl butyrate (Y)-2-Methyl-(3Z)-hexenyl butyrate (1R, 2S, 5R)-()-Menthol (1S, 2R, 5S)-(+)-Menthol (X)-Hydroxycitronellal (Y)-Hydroxycitronellal (R)--Decalactone (S)--Decalactone (S)--Decalactone (R)--Decalactone

1061 1068 1147 1184 1269 1273 1240 1243 1245 1350 1352 1438 1610 1616 1637

1062 1069 1147 1185 1270 1274 1241 1244 1245 1350 1352 1438 1611 1616 1637

122 Table 5 List of compounds included in the library Hydrocarbons -Phellandrene -Pinene -Citronellene -Phellandrene -Pinene Camphene Caryophyllene Limonene Sabinene Heterocyles Ambroxide Menthofuran Rose oxide Esters -Terpinyl acetate Bornyl acetate Bornyl isovalerate Butyl butyrolactate cis-2-Methyl-3-hexenylbutyrate cis-Carvyl acetate Dihydrocarvyl acetate Dimethyl methylsuccinate Ethyl 2-methylbutyrate Ethyl 2-phenylbutyrate Ethyl 3-hydroxybutyrate Ethyl 3-hydroxyhexanoate Ethyl 3-methyl-3-phenylglicidate Isobornyl acetate Isobornyl isobutyrate Lavandulyl acetate Linalyl acetate Linalyl cinnamate Linalyl propionate Menthyl acetate Methyl 3-hydroxyhexanoate Methyl dihydrofarnesoate Neomenthyl acetate Nopyl acetate Propylene glycolbutyrate Styrallyl acetate Lactones Aerangis lactone 3-Methyl--decalactone -Decalactone -Dodecalactone -Heptalactone -Hexalactone -Nonalactone -Octalactone -Undecalactone -Decalactone -Dodecalactone -Decalactone -Dodecalactone -Heptalactone -Hexalactone -Nonalactone -Octalactone -Pentadecalactone -Pentalactone -Tetradecalactone -Undecalactone Massoia decalactone Massoia dodecalactone Whyskey lactone Ketones 1,8-Epoxy p-menthan-3-one 3,6-Dimethylocta 2-en-6-one 3-Methylcyclohexanone 3-Oxocineole -Damascone -Ionone -Irone

E. Liberto et al. / J. Chromatogr. A 1195 (2008) 117126 Table 5 (Continued ) Camphor Camphorquinone Carvone Fenchone Isomenthone Menthone Methylcyclopentenolone Nootkatone Piperitone Pulegone Verbenone Aldehydes Citronellal Cyclamen aldehyde Hydroxycitronellal Myrtenal Perillyl aldehyde Alcohol -Bisabolol 1-Octen-3-ol 1-Phenyl ethanol 1-Phenyl-1-propanol 1-Phenyl-2-pentanol 2-Butanol 2-Heptanol 2-Hexanol 2-Methylbutanol 2-Octanol 2-Pentanol 2-Phenyl-1-propanol 3-Hexanol 3-Octanol 4-Methyl-1-phenylpentanol 6-Methyl-5-hepten-2-ol -Terpineol Borneol cis-Myrtanol Citronellol Fenchyl alcohol Geosmin Isoborneol Isomenthol Isopinocampheol Isopulegol Lavandulol Linalool Linalool oxide Menthol Neoisomenthol Neomenthol Nerolidol Octan-1,3-diol Patchouli alcohol Perillyl alcohol Terpinen-4-ol Tetrahydrolinalool trans-Myrtanol Viridiorol Acids Citronellic acid 2-Methylbutyric acid 2-Phenylpropionic acid Chrysanthemic acid

E. Liberto et al. / J. Chromatogr. A 1195 (2008) 117126

123

Fig. 2. Extract ion proles (77, 99, 114 m/z) of (): Cocus nucifera L. fruit head space sampled by SPME, and (- - -) -octalactone standard solution analysed on a 2,3DA6TBDMS-CD column.

values and/or resolution, in particular with polar analytes. RIA is therefore directly related to the tailing factor of a peak and is determined by measuring the IT values at the start, apex and stop points of the peak; the window is dened through the following expressions:
T T T T T T T (Iap Istart ) < Ix < Ix + (Istop Iap ) Ix T is the IT of the analyte considered, I T is the IT at the apex where Ix ap T T is the IT at the peak starting point and Istop is the IT of the peak, Istart at the peak nal point. This approach gives two sub-window values (the rst one before the peak apex, the second one after the peak apex) limiting identication mistakes with asymmetrical peaks. A correct RIA is also conditioned by the injected analyte concentration that should be such to avoid column overloading in particular with CD stationary phases because of the peak defocusing phenomenon [26]. The unique RIA for the present library was obtained from the average RIA of each class of compounds, in its turn determined from the individual RIA of each of the enantiomers of 134 racemates with different structures, polarities and volatilities analysed three times on the four enantioselective columns adopted. Table 3 reports the average RIA window values of the different groups of compounds analysed with each column. The good RIA homogeneity between the different classes of compounds allowed us to adopt a unique average RIA value of 1 and +2 for all analytes analysed on the investigated columns.

Grob test to evaluate their performance. IT s of the chiral test components were then determined by injecting both the hydrocarbon standard mixture and three times the chiral test. Table 4 reports IT values of the chiral test components and IT values between the two laboratories. These results show that the IT reproducibility is very high and that the system proposed can be used at an interlaboratory level provided that rigorously standardised conditions are applied. 3.2.5. Creation of the library The library was then created by attributing a record to each enantiomer. Each record includes retention indices on four columns, mass spectrum, IUPAC chemical name, CAS number, molecular weight, and, when separated, racemate enantiomer resolution on the chiral selector investigated and, in the included data base, relative retention and tailing factor together with the original source of the each enantiomer (see below). Table 5 reports the list of the compounds at present included in the library. The library with interactive IT values described here is dedicated to GC/MS systems adopted here, and operates together with the related data collecting software (see Section 2.3) [27]. This software is provided with an automated retention index calculation option. IT calculation rst requires the creation of the reference index table obtained by injecting an homologous series of standards (in this case the C9 C25 n-alkane series). IT values of enantiomers or of real-world sample components are then determined relatively to the reference index table. Moreover, up to ve distinct retention index values (i.e. determined on ve stationary phases) for each compound whose mass spectrum is appended to the mass spectral library can be introduced in each record, and a column selection tool is available in order to display in the similarity search result window only the IT values determined on the investigated stationary phase.

3.2.4. Inter-laboratory IT s reliability The IT reliability was also tested by an inter-laboratory round robin test carried out in both the authors laboratories. The chiral test was simultaneously analysed with two sets of four CD columns under rigorously controlled conditions (see experimental). Each CD column was rst submitted to a reconditioning cycle (2 h) and to a

Fig. 3. Enantioselective GCMS prole of Lavandula angustifolia P. Mill. essential oil analysed on a 2,6DM3PEN--CD column.

124

E. Liberto et al. / J. Chromatogr. A 1195 (2008) 117126

Fig. 4. Enantioselective GCMS prole of linalool in Lavandula angustifolia essential oil with a 2,6DM3PEN--CD column. Library search with inactive (1) and active (2) RIA.

The correct elution order of the enantiomers of a racemate was determined by analysing pure standards either commercially available or supplied by other laboratories, and natural sources containing identied enantiomer conrming the results with lit-

erature data. The determination of the correct elution order of -octalactone enantiomers on the four columns is here reported as an example of the latter approach. Coconut fresh fruits were analysed by HS-SPME-enantioselective GCMS on the four CD columns

Fig. 5. Identication of linalyl acetate in Lavandula angustifolia essential oil.

E. Liberto et al. / J. Chromatogr. A 1195 (2008) 117126

125

and -octalactone located in the four chromatograms through its MS spectrum. (Fig. 2). The absolute R conguration was attributed to the identied peak on the basis of literature data reporting the elution order on the same stationary phases [28]. 3.3. Application of the library to real-world samples The reliability of the library was then evaluated through the chiral recognition of some components of Lavandula angustifolia P. Mill (lavender) e.o. These examples are also reported to show how the long and careful work required for the library development resulted in a very easy and quick operation in the everyday work. Since the MS spectra of optical isomers are not distinguishable, their EE and/or ER can directly be determined from the TIC integration, provided that: (a) when complex matrices are analysed, their peaks are perfectly separated and/or diagnostic ions suitable to discriminate them in extract ion mode from other coeluting peaks are present in their mass spectra or that a suitable spectral deconvolution programme is available, and (b) a dramatically high difference of enantiomer abundances do not inuence the mass spectrum relative ion abundances. Moreover, the fact that each record also includes the enantiomer resolution of a racemate on the four columns enables the operator to nd the column(s) (within the four investigated) able to separate from an unresolved racemate, provided that it has been identied through its mass spectrum. Three examples involving the identication and ER determination of well resolved racemates, unresolved racemates, and poorly resolved compounds are discussed here. 3.3.1. Well resolved racemates Two different strategies for library searches can be applied depending if RIA option is operated or not. Enantiomers can be distinguished through interactive IT values only when the RIA is operative. This parameter indicates the IT window where to run the search; only the spectra of analytes with IT values falling in the selected window are considered, thus making possible their

Fig. 6. Enantioselective GCMS proles of linalyl acetate in Lavandula angustifolia essential oil ( ) and of its racemate standard (- - -) analysed on a 2,3DE6TBDMS-CD column.

unequivocal identication. If RIA is not operative, IT values are displayed but they do not actively operate for identication: the enantiomers are therefore located by MS but not distinguished. These considerations are clearly illustrated by the enantioselective GCMS analysis of linalool in lavender e.o with a 2,6DM3PEN--CD column (Figs. 3 and 4) where R enantiomer was found to elute as rst and to be the most abundant (ER >95%) [29]. 3.3.2. Unresolved racemates Fig. 5 shows the reported results after library search when linalyl acetate is analysed with the identical column used above for the same e.o. IT values and RIA are active but specic enantiomers are not identied because they are not separated. However, library search displays the resolution obtained with the four chiral columns enabling us to nd the one separating them. Fig. 6

Fig. 7. Enantioselective GCMS prole of -pinene in Lavandula angustifolia essential oil analysed on a 2,3DE6TBDMS--CD column. Library search with narrow (1) and wide (2) RIA.

126

E. Liberto et al. / J. Chromatogr. A 1195 (2008) 117126

reports the enantioselective GC prole of the same e.o. analysed on a 2,3DE6TBDMS--CD where linalyl acetate is base-line separated, R-isomer eluted as rst and was prevailing (ER > 99%) [29]. 3.3.3. Poorly separated enantiomers The choice of a correct RIA is fundamental for those racemates whose resolution on a given column is below 1.5, i.e. when they partially overlap. In this case, too wide a RIA may include both enantiomers making their identication problematic, even if their IT values are different; on the other hand, if RIA is too narrow, the risk that IT falls outside the window is concrete, in particular if the chromatographic system is not perfectly stable. The rst possibility is clearly illustrated in Fig. 7 for -pinene enantiomer identication in the same lavender essential oil when it is analysed with a 2,3DE6TBDMS--CD column and a RIA between 3 and +3 is applied. On the other hand Fig. 7, shows that the general RIA adopted for this library, from 1 to + 2, provides a correct enantiomer identication showing that the R isomer elutes as rst with an ER above 75%. 4. Conclusions The results reported here demonstrated the reliability of the adopted approach for an unequivocal enantiomer identication by enantioselective GCMS in the avour and fragrance eld and the fundamental importance of combining IT values and mass spectra actively for the correct identication of an enantiomer. IT values have also been demonstrated to be highly stable, thus affording the adoption of a RIA value common to all class of racemates investigated between +1 and 2 IT units, provided that the reference hydrocarbon standard mixture is injected every ve analyses when the library is created. The library has also been shown to operate effectively in the analysis of real-world samples and at present it includes the enantiomers of 134 racemates analysed on columns coated with four different CD derivatives. The number of records is being constantly increased although its implementation takes time because of the difculty to nd new racemates and the related pure enantiomers.

Acknowledgements This research was carried out within the project entitled: Sviluppo di metodologie innovative per lanalisi di prodotti agroalimentari (FIRB Cod.: RBIP06SXMR 002) of the Ministero ` e della Ricerca (MIUR) (Italy). dellIstruzione, dellUniversita References
[1] D. Sybilska, T. Koscielski, J. Chromatogr. 261 (1983) 357. [2] Z. Juvancz, G. Alexander, J. Szejtli, J. High Resolut. Chromatogr. Chromatogr. Commun. 10 (1987) 105. [3] V. Schurig, H.P. Nowotny, J. Chromatogr. 441 (1988) 155. [4] H.P. Nowotny, D. Schmalzing, D. Wistuba, V. Schurig, J. High Resolut. Chromatogr. 12 (1989) 383. [5] C. Bicchi, V. Manzin, A. DAmato, P. Rubiolo, Flavour Fragr. J. 10 (1995) 127. [6] C. Bicchi, A. DAmato, V. Manzin, in: K. Swift (Ed.), Flavours and Fragrances, Royal Society of Chemistry, Cambridge, UK, 1997, p. 57. [7] C. Bicchi, A. DAmato, P. Rubiolo, J. Chromatogr. A 843 (1999) 99. [8] W.A. Koenig, D.H. Hochmuth, J. Chromatogr. Sci. 42 (2004) 402. [9] G. Schomburg, H. Husmann, E. Hubinger, W.A. Koenig, J. High Resolut. Chromatogr. Chromatogr. Commun. 7 (1984) 404. [10] A. Bernreuther, N. Christoph, P. Schreier, J. Chromatogr. 481 (1989) 363. [11] A. Mosandl, U. Hener, U. Hagenauer-Hener, A. Kustermann, J. High Resolut. Chromatogr. 12 (1989) 532. [12] L. Mondello, M. Catalfamo, G. Dugo, P. Dugo, J. Chromatogr. Sci. 36 (1998) 201. [13] M. Jung, D. Schamalzing, V. Schurig, J. Chromatogr. 552 (1991) 43. [14] V. Schurig, H.P. Nowotny, Angew. Chem., Int. Ed. Engl. 29 (1990) 939. [15] R. Shellie, P. Marriott, C. Cornwell, Anal. Chem. 74 (2002) 5426. [16] R. Shellie, L. Mondello, G. Dugo, P. Marriot, Flavour Fragr. J. 19 (2004) 582. Helv. Chim. Acta 41 (1958) 1915. [17] E. Kovats, [18] H. Van den Dool, P.D. Kratz, J. Chromatogr. 11 (1963) 463. [19] C. Bicchi, G. Artuffo, A. DAmato, V. Manzin, J. High Resolut. Chromatogr. 15 (1992) 710. [20] W.A. Koenig, B. Gehrcke, D. Ichelin, P. Evers, J. Donnecke, W. Wang, J. High Resolut. Chromatogr. 15 (1992) 367. [21] B. Maas, A. Dietrich, D. Bartschat, A. Mosandl, J. Chromatogr. Sci. 33 (1995) 223. [22] C. Bicchi, A. DAmato, V. Manzin, A. Galli, M. Galli, J. Chromatogr. A 742 (1996) 161. [23] K. Grob, G. Grob, K. Grob, J. Chromatogr. 156 (1978) 1. [24] K. Grob, G. Grob, K. Grob, J. Chromatogr. 219 (1981) 13. [25] C. Bicchi, G. Artuffo, A. DAmato, V. Manzin, A. Galli, M. Galli, J. High Resolut. Chromatogr. 16 (1991) 209. [26] H.G. Schmarr, B. Maas, A. Mosandl, S. Bihler, H.P. Neukom, K. Grob, J. High Resolut. Chromatogr. 14 (1991) 317. [27] R. Costa, M.R. De Fina, M.R. Valentino, P. Dugo, L. Mondello, Nat. Product Commun. 2 (2007) 413. [28] D. Lehmann, B. Maas, A. Mosandl, Z. Lebensm, Unters. Forsch. 201 (1995) 55. [29] P. Kreiss, A. Mosandl, Flavour Fragr. J. 7 (1992) 187.