Application of Gas–Liquid Chromatography to the Analysis of Essential Oils Part XVII.† Fingerprinting of Essential Oils by Temperature-programmed Gas–Liquid Chromatography Using Capillary Columns With Non-polar Stationary Phases Analytical Methods Committee‡ The Royal Society of Chemistry, Burlington House, Piccadilly, London, UK W1V 0BN Problems in obtaining reproducible results when ‘fingerprinting’ essential oils by temperature-programmed gas–liquid chromatography have been reported on in Parts VII and VIII of this series.Those reports were concerned with the general problems and the use of packed columns. This report is concerned with the use of capillary columns and non-polar stationary phases. A collaborative study using capillary columns with non-polar stationary phases has resulted in a method which specifies the ‘g-pack value’ of a column and gives reproducible relative retention indices for the test compounds limonene, acetophenone, linalol, naphthalene, linalyl acetate and cinnamyl alcohol. The method has been applied successfully to the examination of oil of rosemary.A recommended method is given for the reproducible temperature-programmed gas–liquid chromatographic fingerprinting of essential oils using capillary columns with non-polar stationary phases. Keywords: Essential oil analysis; gas–liquid chromatography; fingerprinting; non-polar capillary column The Analytical Methods Committee (AMC) has received and has approved for publication the following report from its Essential Oils Sub-Committee.Report The constitution of the Sub-Committee responsible for the preparation of this report was: Mr. M. J. Milchard (Chairman), Mr. A. M. Humphrey (Chairman to October 1993), Mr. N. Boley, Mr. B. Conway, Mr. R. Esdale, Ms. M. Flowerdew, Miss D. M. Michalkiewicz, Mr. D. A. Moyler, Mr. A. Osbiston, Mr. D. Powis, Mr. A. Sherlock, Mr. R. Smith, Mr. S. Smith, Mr. B. Starr and Mr. T. M. Stevens, with Mr.J. J. Wilson as Secretary. The Sub-Committee would like to thank the following manufacturers of capillary chromatography columns for the interest which they have shown in this work and their willingness to participate in the collaborative trial: Alltech Associates Inc., Chrompack UK Ltd., Hewlett-Packard Ltd., J&W Scientific/Jones Chromatography, Quadrex Corporation, Restek/Thames Chromatography and SGE (UK) Ltd. Their assistance is gratefully acknowledged.Introduction The development of gas–liquid chromatography (GLC) in the early 1950s began a new era in the analysis of essential oils. Until that time the qualities, purities and origins of oils were assessed by physical measurements and chemical assays. This new method gave improved results over the older methods, many of which we now know gave precise but inaccurate results. The technique was soon applied to the accurate determination of major and other components of interest.2–11 However, the concept of ‘fingerprinting’ oils had not been addressed successfully.The chemical nature of essential oils makes them particularly suitable for analysis by GLC. If temperature-programmed operation is used, a very high proportion of the total number of components present can be resolved. Many attempts have been made to establish libraries of chromatograms from temperatureprogrammed GLC analysis. This would allow sample and reference tracings to be compared and the authenticity and quality of the sample to be determined. However, it was quickly found that the conditions of the GLC analysis had to be strictly controlled, but even then reproducibility was poor, particularly among different laboratories.It was apparent that the temperature programming and the nature of the column itself caused the greatest variation in the results. This has led to individual libraries being established, which makes comparison difficult. The Sub-Committee has been studying the fingerprinting of essential oils over many years.Initial work on packed columns established that the lack of reproducibility of results was due to the lack of reproducibility of the columns themselves. This in turn was attributed to problems with coating the support and the subsequent ageing of the packing with use. There was, therefore, a requirement for the standardisation of the column efficiency and of its selectivity without using one as a factor of the other.A publication by van den Dool12 described a method for the characterisation of GLC columns using the relative retentionindices (RRIs) of a group of six test compounds. By calculating the RRIs of the compounds in the mixture on a particular column and then applying a series of further calculations to these RRIs, van den Dool obtained a figure representing the polarity factor of that particular column which he called the g- † For Part XVI, see ref. 1. ‡ Correspondence should be addressed to the Secretary, Analytical Methods Committee, Analytical Division, The Royal Society of Chemistry, Burlington House, Piccadilly, London, UK W1V 0BN. Analyst, October 1997, Vol. 122 (1167–1174) 1167pack value. This value will vary with the different types of stationary phase and their condition. Van den Dool’s work was based exclusively on packed columns and concentrated on two particular stationary phases: SE-30 as an example of a non-polar type and Carbowax 20M as a moderately polar type.However, the g-pack concept can be applied to any stationary phase. The mixture of compounds used in the determination, known as the NC (Netherlands Committee) mixture, was chosen to represent a range of compounds with functional groups similar to those found in essential oils. Also, the mixture was chosen so that there are two pairs of compounds in which the components of each pair elute close together on the two different phases, which can be used to give a measure of the resolving power of the column. However, the greatly increased efficiency of capillary columns over packed columns means that this property of the mixture is not so significant except in cases of extreme column degradation.The NC mixture consists of limonene, linalol, linalyl acetate, acetophenone, naphthalene and cinnamyl alcohol. The Sub-Committee has published several papers, based on the work of van den Dool, defining the methodology to be used in obtaining these fingerprints13–15 and standard fingerprint traces of selected oils.1,16,17 During these studies on the application of the g-pack concept it was found that the value for a packed column was unaffected by changes in operating parameters such as carrier gas flow rate, temperature programming rate, initial temperature and final temperature hold.However, the value could be decreased by loss of stationary phase over a period of time, due to column bleeding, and could be increased by modification of the stationary phase due to oxidation.The g-pack value can, therefore, give a good indication of the condition of the column. With the wider availability and advances in capillary column technology the Sub-Committee decided that the technique should be updated to make use of this technology and set about determining the optimum operating parameters. It was appreciated that the application to capillary columns was not likely to be as straightforward as with packed columns.A major difference was in the control of the stationary phase. For packed columns, good agreement among laboratories was only obtained when there was control over the preparation of the stationary phase and packing of the column. This involved developing a method for coating the stationary phase onto the support (the absorption coating technique), which led to columns giving good reproducibility and high efficiencies. While it is possible to prepare and coat capillary columns, in reality laboratories buy in their columns from one of the specialist manufacturers.This means that individual laboratories have no control over the preparation of the columns and that different methods of preparation between manufacturers could lead to slightly different performances of nominally the same stationary phase. It was concluded, therefore, that any method for fingerprinting essential oils on capillary columns would have to be based on commercially available columns and be robust enough to cope with a range of operating parameters.If the method is to be widely applicable it is unrealistic to expect laboratories to buy columns from a specific manufacturer or to change carrier gas. Experimental It was decided that the initial examination would be carried out on a specified stationary phase on columns which the members of the committee had available in their laboratories. A non-polar phase was chosen equivalent to SE-30.A protocol was provided specifying the samples to be examined and the initial column temperature, temperature programme rate and final column temperature. The choice of carrier gas was left to individuals with the proviso that the flow rate was optimised for the column configuration. The three samples examined were: a mixture of the NC mix and a series of even numbered carbon aliphatic hydrocarbons (C8–C24) known as the NC–HC mixture; Spanish rosemary oil; and Spanish rosemary oil plus the hydrocarbon mixture.The initial results showed that there were differences between different column manufacturers but very similar results were obtained between laboratories using similar columns from the same manufacturer. It was, therefore, decided to approach the major capillary column manufacturers and invite them to examine the same samples on their columns. All of the manufacturers were very willing to collaborate in this examination with the given protocol.With their inclusion a total of 16 results were obtained. Results The results of the collaborative study using the NC mixture with the proposed procedure are given in Table 1 in ascending order of g-pack values and are summarised in Table 2. It can be seen from Table 2 that the mean RRIs for the test compounds obtained in this examination compare well with results previously obtained on packed columns. Table 3 shows the elution temperatures of the test compounds and n-alkane hydrocarbons.There is a much wider variation in elution temperatures with capillary columns than was found with packed columns. The results from the examination of oil of rosemary are given in Tables 4 and 5. The results obtained were considered satisfactory by the Sub-Committee. Discussion Throughout all of the collaborative exercises, the Sub-Committee was conscious of the fact that many of the parameters examined were chosen with arbitrary limits and that some justification for the choices should be made.In particular, this applies to the nature of the test compounds in the NC mixture and the use of the RRI system. This system is widely used and most gas chromatographers are acquainted with it. It is a simple system and, although it is relative rather than absolute, it allows for a choice of reference compounds which can be made according to other requirements. For the purpose of this and the previous investigation it was felt that the homologous series of n-alkanes was the most suitable as reference compounds.They are readily available in pure form, are extremely stable and are the least likely compounds to exhibit chromatographic anomalies. In addition, they formed the basis of the work by van den Dool. The elution of a variety of homologous series under temperature-programmed conditions is not linear and the departure from linearity increases as the programming rate decreases.18 Also, this departure is greater for homologous series of polar compounds than for the n-alkanes and this was another reason for the choice of the latter.This non-linearity of elution raises the question of the method of application of the system and if an assumption of linearity between successive n-alkanes introduces errors. Van den Dool recommends the use of both even and odd numbered carbon nalkanes to reduce any error, whereas all of our collaborative work has been carried out with the even numbered carbon nalkanes making the assumption of linearity between each pair.Comparisons have been made of the effect on the calculated RRIs of the six NC text compounds when using all of the n- 1168 Analyst, October 1997, Vol. 122alkanes against using only the even numbered carbon ones and also by calculating on the assumption of linearity as well as by a graphical method to obtain more accurate figures. The differences that were found were so small (in some instances zero) that they were considered to be insignificant.Therefore, our recommended procedure uses only the even numbered carbon n-alkanes and assumes linearity between them. The six compounds in the NC mixture were selected by van den Dool on the basis of their similarity to the types of compounds found in essential oil analyses. Their choice could perhaps be criticised if used in conjunction with other sample types. However, the NC mixture is used for the calibration of the column irrespective of its ultimate use.As the mathematical treatment of the results has already been worked out by van den Dool, there seemed to be no advantage in changing the mixture. The choice of six compounds is a reasonable compromise between having an excessive number of interfering peaks and mathematical calculations and a reduced amount of data leading to a less accurate result. With packed columns, more control was exercised over the operating conditions by arranging for the C24 alkane to elute at the upper temperature of the temperature programme run by adjustment of the carrier gas flow rate.This approach is not relevant for capillary columns as, to obtain satisfactory results, it is necessary to optimise the flow rate, which depends on the dimensions of the individual columns and the nature of the carrier gas. The superior resolving power of capillary columns led to the conclusion that measurement of the resolution between the limonene and acetophenone peaks would not be meaningful as there was baseline separation in all cases.The column would be in poor condition or the operating conditions be significantly different from those recommended for these compounds not to be resolved. Developments in gas chromatographic instrumentation have led to discussions on the relative merits of operating the carrier gas system under conditions of constant flow or constant pressure. The NC–HC mixture was chromatographed on a column in an instrument equipped to run under both conditions.As shown below, the RRIs were virtually identical under both conditions. However, the chromatogram of the sample run under constant pressure conditions showed an increase in the retention times of the components compared with those run under constant flow conditions. This is because of a reduction in flow rate during temperature-programmed operation under constant pressure. RRI Compound Constant pressure Constant flow Limonene 1020 1020 Acetophenone 1034 1034 Linalol 1084 1084 Naphthalene 1158 1156 Linalyl acetate 1242 1241 Cinnamyl alcohol 1274 1273 g-pack 0.999 This demonstrates that either method of carrier gas control should be suitable for undertaking fingerprint determinations using this procedure.The RRI and area % composition of the 17 most abundant components in the sample of oil of rosemary were calculated from chromatograms run under the same conditions as for the HC–NC mixture.From the data in Tables 4 and 5, the Sub- Committee considers that the collated results of the collaborative trials show good agreement. The aromatic compounds show greater variability in their RRIs than the non-aromatic compounds. This effect was noted in the previous publications on fingerprinting by the Sub-Committee on packed columns. 13,14 It is attributed to the greater variability in RRI of the aromatic compounds with changes in elution temperature when compared with the non-aromatics.This observation is being studied and will be reported on separately. The data in Table 5 are given as an indication of the composition of oil of rosemary. They assume that all of the compounds have the same response to the flame-ionisation detector and that the oil does not contain any non-volatile material. Also, different integrator parameters will affect the percentage composition data. Conclusion The Sub-Committee recommends the procedure given in the Appendix for the reproducible fingerprinting of essential oils by temperature-programmed GLC using capillary columns with non-polar stationary phases.Although the procedures have been developed and investigated for the analyses of essential oils, it is felt that they have a Table 1 Results of collaborative study on the NC mixture Laboratory 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Stationary phase CPSil5 DB1 BP1 AT1 BP1 007-1 HP1 BP1 DB1 BP1 Rtx1 CPSil5 HP1 CPSil5 HP1 DB1 Column length/m 25 30 25 25 25 25 30 25 60 50 30 25 50 25 25 60 Internal diameter/mm 0.32 0.25 0.22 0.32 0.22 0.25 0.25 0.22 0.25 0.22 0.25 0.25 0.20 0.25 0.32 0.32 Film thickness/ mm 0.25 0.25 0.25 0.30 0.25 0.25 0.25 0.25 0.25 0.25 0.25 0.25 0.50 0.25 1.05 0.25 Carrier gas He H2 H2 He He He He N2 He H2 He He He He He He Test compounds, RRI values— Limonene 1014 1015 1018 1019 1021 1020 1020 1021 1024 1025 1027 1028 1028 1029 1028 1029 Acetophenone 1025 1026 1030 1032 1034 1036 1034 1034 1037 1038 1042 1043 1044 1047 1048 1046 Linalol 1076 1079 1083 1084 1085 1086 1084 1085 1085 1087 1088 1089 1090 1090 1094 1091 Naphthalene 1139 1144 1153 1155 1158 1158 1156 1159 1167 1168 1173 1176 1176 1179 1176 1177 Linalyl acetate 1240 1240 1241 1242 1242 1243 1241 1242 1242 1242 1243 1242 1244 1245 1248 1238 Cinnamyl alcohol 1270 1268 1271 1274 1277 1279 1273 1276 1276 1279 1283 1281 1284 1290 1288 1268 g-pack 0.992 0.994 0.997 0.998 0.999 0.999 0.999 1.000 1.001 1.002 1.004 1.005 1.005 1.006 1.007 1.007 Analyst, October 1997, Vol. 122 1169much wider application and should find use in many other fields of GLC analysis. Appendix Recommended Method for the Reproducible Fingerprinting of Essential Oils by Temperature-programmed Gas–Liquid Chromatography Using Non-polar Stationary Phases The column The preparation of the types of capillary column available today requires considerable experience and expertise. Chromatographers, therefore, have to rely on specialist manufacturers as few have the required skills.However, this study has shown that columns obtained ‘off-the-shelf’ from the major manufacturers were all suitable for use with this method. The stationary phases referred to in Table 1 are the manufacturers’ trade names for polysiloxane phases with no modifications. The dimensions of the column are not critical to the successful application of this method but the following are recommended.Column length, 25–30 m; internal diameter, 0.22–0.25 mm; film thickness, 0.25 mm. A film thickness of more than 3 mm should not be used as different effects are observed. Columns are usually delivered ready for use but any manufacturers’ instructions concerning conditioning should be heeded. Gas chromatographic conditions The temperature control in the ovens of modern gas chromatographs is very accurate. However, if older instruments are used or if any doubt exists, oven settings should be checked using a thermometer of known accuracy.The temperature range for this procedure is 50–250 °C. The programme rate is very important and must be checked to ensure that it is linear. In this case a rate of 4 °C min21 is used. Any of the three commonly used carrier gases, helium, hydrogen and nitrogen, may be used. However, it is most important that the linear gas velocity is adjusted so that the column is operating at optimum efficiency. The actual velocity will depend on the carrier gas and the dimensions of the column.Column manufacturers will advise on this. Carrier gas control may be by either constant flow or constant pressure. Preparation of test mixtures Prepare a mixture of equal masses of the even carbon numbered n-alkanes from C8 to C24. Prepare a mixture of 1.00 part of limonene, 1.37 parts of linalol, 1.60 parts of linalyl acetate, 1.40 parts of acetophenone, 1.13 parts of naphthalene and 1.80 parts of cinnamyl alcohol (NC mixture).Prepare a mixture of 55% m/ m of the NC mixture and 45% m/m of the hydrocarbon mixture. Each n-alkane is then 5% and each NC component is then approximately 10% of the total mixture. These mixtures are now available commercially. Test chromatogram Set up the chromatograph, with the prepared column, for temperature-programmed operation between 50 and 250 °C at 4 °C min21. Inject the combined test mixture and start the programme and integrator. Continue heating at 250 °C until a stable baseline is obtained.Repeat the run if necessary, adjusting the attenuation to bring all the peaks on-scale. A typical chromatogram is shown in Fig. 1. The sample size, Table 2 Summary of results on the NC mixture (Table 1) and comparison with those on packed columns previously examined Relative Capillary columns— Standard standard Test compound Mean RRI deviation deviation (%) Limonene 1023 4.87 0.48 Acetophenone 1037 6.98 0.67 Linalol 1086 4.36 0.40 Naphthalene 1163 12.03 1.03 Linalyl acetate 1242 2.19 0.18 Cinnamyl alcohol 1277 6.48 0.51 Relative Packed columns*— Standard standard Test compound Mean RRI deviation deviation (%) Limonene 1027 1.63 0.16 Acetophenone 1041 2.99 0.29 Linalol 1086 1.80 0.17 Naphthalene 1172 3.55 0.30 Linalyl acetate 1241 1.90 0.15 Cinnamyl alcohol 1280 3.80 0.30 * See ref. 14. Table 3 Elution temperatures of the NC–HC mixture (°C) Laboratory 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Limonene 72 76 87 93 96 91 94 96 117 112 124 125 127 112 122 145 Acetophenone 73 77 89 95 98 93 96 98 119 114 126 127 130 114 125 148 Linalol 79 84 96 103 106 100 102 106 127 122 133 134 137 121 132 155 Naphthalene 86 92 106 113 116 111 112 117 139 134 147 147 150 134 144 169 Linalyl acetate 98 105 118 125 128 123 123 129 151 146 158 157 159 144 154 180 Cinnamyl alcohol 102 108 123 130 133 128 127 133 155 151 164 162 165 150 159 185 g-pack 0.992 0.994 0.997 0.998 0.999 0.999 0.999 1.000 1.001 1.002 1.004 1.005 1.005 1.006 1.007 1.007 C8 59 59 64 69 71 66 70 70 86 80 90 89 92 82 88 112 C10 71 74 85 91 93 88 91 93 113 108 119 120 123 107 118 140 C12 93 99 113 120 122 117 118 123 145 139 151 151 153 137 147 173 C14 118 126 140 148 151 145 143 151 173 169 181 178 181 165 174 209 C16 143 150 165 173 176 170 167 177 199 195 208 202 206 190 198 1170 Analyst, October 1997, Vol. 122dilution and split ratio should be such that the capacity of the column is not exceeded.Calculation of results When a satisfactory chromatogram has been obtained with baseline separation of all peaks, calculate the RRIs of the NC components assuming a linear span between adjacent hydrocarbon peaks. Using the values obtained, tabulate the results as in Table 6 and calculate the g-pack value for the column following the given worked example. This calculation can easily be performed by using a computer spreadsheet. Calculation of relative retention indices (RRIs) RRI = � - + - + [ ( )] ( ) 200 2 100 Rtc Rtn Rt n Rtn n where n = carbon number of 1st hydrocarbon; Rtc = retention time of the compound; Rtn = retention time of 1st hydrocarbon; and Rt(n + 2) = retention time of 2nd hydrocarbon.Table 4 Results of collaborative study on rosemary oil: relative retention indices Laboratory Compound 1 2 3 4 5 6 7 8 9 10 11 12 13 Mean s s (%) a-Pinene 906 909 916 918 922 918 920 928 928 930 937 936 935 923 9.61 1.04 Camphene 918 921 929 930 935 932 934 942 942 946 951 949 950 937 10.41 1.11 Sabinene —* —* 956 957 959 959 959 —* 963 966 966 964 970 962 4.35 0.45 b-Pinene 949 952 960 960 964 963 964 971 971 975 978 976 977 966 9.07 0.94 Myrcene 972 973 977 977 979 979 979 981 981 982 983 982 983 979 3.41 0.35 p-Cymene 1005 1006 1009 1009 1011 1010 1011 1014 1014 1017 1018 1016 1017 1012 4.10 0.41 Limonene/ cineol 1012 1014 1017 1018 1022 1019 1019 1024 1024 1027 1030 1029 1033 1022 6.14 0.60 Linalol 1074 1076 1082 1081 1084 1083 1083 1084 1086 1088 1088 1087 1092 1084 4.70 0.43 Camphor 1100 1106 1114 1117 1121 1118 1118 1127 1127 1132 1136 1135 1140 1122 11.43 1.02 Borneol 1130 1135 1142 1143 1147 1146 1147 1153 1155 1160 1161 1159 1163 1149 9.89 0.86 Terpinen-4-ol 1143 1148 1157 1156 1159 1159 1160 11 1166 1171 1171 1169 1174 1161 8.84 0.76 a-Terpineol 1157 1161 1161 1167 1171 1170 1171 1175 1177 1181 1181 1179 1184 1172 7.85 0.67 Verbenone 1162 1166 1173 1175 1178 1178 1178 1184 1186 1193 1192 1190 1200 1181 10.57 0.90 Bornyl acetate 1255 1259 1265 1265 1268 1269 1269 1276 1275 1280 1280 1278 1283 1271 8.25 0.65 b-Caryophyllene 1392 1433 1409 1410 1414 1414 1416 1430 1427 1434 1437 1434 1433 1422 13.07 0.92 a-Humulene 1425 1467 1444 1443 1447 1447 1450 1464 1461 1468 1471 1467 1467 1455 13.27 0.91 g-pack 0.992 0.994 0.997 0.998 0.999 0.999 1.000 1.001 1.002 1.004 1.005 1.005 1.007 * Compound not resolved.Table 5 Results of collaborative study on rosemary oil: area % of selected components Laboratory Compound 1 2 3 4 5 6 7 8 9 Mean s s (%) a-Pinene 22.5 18.6 20.2 24.3 19.8 20.4 21.4 20.0 20.3 20.8 1.59 7.6 Camphene 9.2 9.4 8.4 9.8 8.2 8.3 9.1 8.6 8.6 8.8 0.52 5.9 Sabinene 1.3 1.4 1.3 1.4 1.1 1.3 1.2 1.2 1.3 0.10 7.7 b-Pinene 4.5 3.2 2.9 3.2 2.9 2.8 2.9 2.9 2.9 3.1 0.50 16.1 Myrcene 4.2 4.2 4.2 4.3 4.1 3.9 4.1 4.0 3.9 4.1 0.13 3.2 p-Cymene 2.3 2.5 1.9 2.2 1.8 1.8 2.3 1.8 2.1 2.1 0.25 11.9 Limonene/cineol 26.5 26.9 25.2 26.6 25.4 25.6 26.0 26.0 26.0 26.0 0.53 2.0 Linalol 0.8 1.2 1.1 0.9 1.1 1.0 1.1 1.5 1.1 1.1 0.19 17.3 Camphor 17.5 21.2 18.7 16.9 18.9 18.8 18.3 19.4 20.1 18.9 1.22 6.5 Borneol 2.8 3.7 3.2 2.7 3.2 2.8 3.0 2.9 3.2 3.1 0.29 9.4 Terpinen-4-ol 0.7 0.9 0.8 0.7 0.8 0.7 0.8 0.7 0.8 0.8 0.07 8.8 a-Terpineol 1.2 1.5 1.4 1.1 1.4 1.3 1.3 1.4 1.5 1.3 0.13 10.0 Verbenone 1.5 1.9 1.7 1.4 1.6 1.5 1.6 1.6 1.7 1.6 0.14 8.8 Bornyl acetate 0.8 1.2 0.9 0.7 0.9 0.9 0.9 1.0 1.0 0.9 0.13 14.4 b-Caryophyllene 1.8 2.3 2.2 1.5 2.2 2.1 1.9 2.2 2.2 2.0 0.25 12.5 a-Humulene 0.6 0.7 0.7 0.5 0.7 0.6 0.6 0.7 0.7 0.6 0.07 11.7 g-pack 0.994 0.997 0.999 0.999 1.000 1.002 1.004 1.005 1.007 Analyst, October 1997, Vol. 122 1171For example, if the retention times of the C10 hydrocarbon, limonene and C12 hydrocarbon are 17.7, 18.6 and 25.2 min, respectively, then n = 10 and RRI for limonene = [200 (18.6 - 17.7)] 25.2 - 17.7 � + = 1000 1024 The operating conditions and chromatographic system may be considered satisfactory if the results lie within the following specification: g-pack 1.0 ± 0.005 RRI values: Limonene 1018–1028 Acetophenone 1030–1044 Linalol 1082–1090 Naphthalene 1151–1175 Linalyl acetate 1240–1244 Cinnamyl alcohol 1271–1283 Standardised chromatograms of essential oils When the performance of a column within a chromatographic system has been satisfactorily established according to the procedure given above, it can be used under similar conditions for analysing essential oils.The performance of the system should be checked regularly. A sample of the essential oil should be injected and run under the conditions established above (first chromatogram). On completion of the run, a sample of a mixture of the essential oil and the n-alkane hydrocarbon mixture should be run under Fig. 1 Typical chromatogram of NC–hydrocarbon mixture. Table 6 Calculation of g-pack value Test compound RRI Y factor X factor Z factor Limonene 1020 (RRI30.14) + 2 = 1.0629 1.05843Y = 1.1250 1.058423Y = 1.1907 136.23 Acetophenone 1034 (RRI30.14) + 2 = 1.2216 1.33503Y = 1.6308 1.335023Y = 2.1772 120.14 Linalol 1084 (RRI30.14) + 2 = 0.9969 1.02183Y = 1.0186 1.021823Y = 1.0408 154.24 Naphthalene 1156 (RRI30.14) + 2 = 1.2784 1.33613Y = 1.7081 1.336123Y = 2.2822 128.16 Linalyl acetate 1241 (RRI30.14) + 2 = 0.8954 0.87973Y = 0.7877 0.879723Y = 0.6929 196.28 Cinnamyl alcohol 1273 (RRI30.14) + 2 = 1.3432 1.49933Y = 2.0139 1.499323Y = 3.0194 134.17 Sum SY = 6.7984 SX = 8.2841 SZ = 10.4032 3factor f f1SY = 6.79843 f2SX = 8.28413 f3SZ = 10.40323 1.07977 = 7.3407 2.88734 = 23.9190 1.49758 = 15.5796 g-pack value = f2SX2f1SY2f3SZ = 23.919027.3407215.5796 = 0.9987 1172 Analyst, October 1997, Vol. 122identical conditions (second chromatogram). Comparison of the two chromatograms should show identical retention times and comparisons of the peak heights should be consistent with any dilution due to the n-alkane mixture. These comparisons allow for a check on the reproducibility of the system and also enable the position of the n-alkanes to be transferred from the second chromatogram (Fig. 3) to the first (Fig. 2) in such a manner that visual interference is avoided. In instances where a component of the essential oil overlaps or obscures an n-alkane peak, its position can be determined by comparison with the chromatogram obtained for the g-pack calculation (Fig. 1). Thus, it is possible to determine RRIs for any of the peaks of interest in the essential oil chromatogram.References 1 Analytical Methods Committee, Analyst, 1993, 118, 1089. 2 Analytical Methods Committee, Analyst, 1971, 96, 887. 3 Analytical Methods Committee, Analyst, 1973, 98, 616. 4 Analytical Methods Committee, Analyst, 1973, 98, 823. 5 Analytical Methods Committee, Analyst, 1975, 100, 593. 6 Analytical Methods Committee, Analyst, 1977, 102, 607. Fig. 2 Typical chromatogram of oil of rosemary Spanish. Fig. 3 Typical chromatogram of oil of rosemary–hydrocarbon mixture.Analyst, October 1997, Vol. 122 11737 Analytical Methods Committee, Analyst, 1978, 103, 375. 8 Analytical Methods Committee, Analyst, 1981, 106, 456. 9 Analytical Methods Committee, Analyst, 1987, 112, 1315. 10 Analytical Methods Committee, Analyst, 1988, 113, 657. 11 Analytical Methods Committee, Analyst, 1990, 115, 459. 12 van den Dool, H., Standardisation of G.C. Analysis of Essential Oils, Proefschrift, Rijksuniversiteit te Groningen, Rotterdam, 1974. 13 Analytical Methods Committee, Analyst, 1980, 105, 262. 14 Analytical Methods Committee, Analyst, 1981, 106, 448. 15 Analytical Methods Committee, Analyst, 1984, 109, 1339. 16 Analytical Methods Committee, Analyst, 1984, 109, 1343. 17 Analytical Methods Committee, Analyst, 1988, 113, 1125. 18 Grant, D. W., and Hollis, M. G., J. Chromatogr., 1978, 158, 3. Paper 7/04651K Accepted July 2, 1997 1174 Analyst, October 1997, Vol. 122 Application of Gas–Liquid Chromatography to the Analysis of Essential Oils Part XVII.† Fingerprinting of Essential Oils by Temperature-programmed Gas–Liquid Chromatography Using Capillary Columns With Non-polar Stationary Phases Analytical Methods Committee‡ The Royal Society of Chemistry, Burlington House, Piccadilly, London, UK W1V 0BN Problems in obtaining reproducible results when ‘fingerprinting’ essential oils by temperature-programmed gas–liquid chromatography have been reported on in Parts VII and VIII of this series.Those reports were concerned with the general problems and the use of packed columns. This report is concerned with the use of capillary columns and non-polar stationary phases. A collaborative study using capillary columns with non-polar stationary phases has resulted in a method which specifies the ‘g-pack value’ of a column and gives reproducible relative retention indices for the test compounds limonene, acetophenone, linalol, naphthalene, linalyl acetate and cinnamyl alcohol.The method has been applied successfully to the examination of oil of rosemary. A recommended method is given for the reproducible temperature-programmed gas–liquid chromatographic fingerprinting of essential oils using capillary columns with non-polar stationary phases. Keywords: Essential oil analysis; gas–liquid chromatography; fingerprinting; non-polar capillary column The Analytical Methods Committee (AMC) has received and has approved for publication the following report from its Essential Oils Sub-Committee.Report The constitution of the Sub-Committee responsible for the preparation of this report was: Mr. M. J. Milchard (Chairman), Mr. A. M. Humphrey (Chairman to October 1993), Mr. N. Boley, Mr. B. Conway, Mr. R. Esdale, Ms. M. Flowerdew, Miss D. M. Michalkiewicz, Mr. D. A. Moyler, Mr. A. Osbiston, Mr. D. Powis, Mr. A. Sherlock, Mr. R. Smith, Mr. S. Smith, Mr. B. Starr and Mr. T. M.Stevens, with Mr. J. J. Wilson as Secretary. The Sub-Committee would like to thank the following manufacturers of capillary chromatography columns for the interest which they ha shown in this work and their willingness to participate in the collaborative trial: Alltech Associates Inc., Chrompack UK Ltd., Hewlett-Packard Ltd., J&W Scientific/Jones Chromatography, Quadrex Corporation, Restek/Thames Chromatography and SGE (UK) Ltd. Their assistance is gratefully acknowledged.Introduction The development of gas–liquid chromatography (GLC) in the early 1950s began a new era in the analysis of essential oils. Until that time the qualities, purities and origins of oils were assessed by physical measurements and chemical assays. This new method gave improved results over the older methods, many of which we now know gave precise but inaccurate results. The technique was soon applied to the accurate determination of major and other components of interest.2–11 However, the concept of ‘fingerprinting’ oils had not been addressed successfully.The chemical nature of essential oils makes them particularly suitable for analysis by GLC. If temperature-programmed operation is used, a very high proportion of the total number of components present can be resolved. Many attempts have been made to establish libraries of chromatograms from temperatureprogrammed GLC analysis. This would allow sample and reference tracings to be compared and the authenticity and quality of the sample to be determined.However, it was quickly found that the conditions of the GLC analysis had to be strictly controlled, but even then reproducibility was poor, particularly among different laboratories. It was apparent that the temperature programming and the nature of the column itself caused the greatest variation in the results. This has led to individual libraries being established, which makes comparison difficult. The Sub-Committee has been studying the fingerprinting of essential oils over many years.Initial work on packed columns established that the lack of reproducibility of results was due to the lack of reproducibility of the columns themselves. This in turn was attributed to problems with coating the support and the subsequent ageing of the packing with use. There was, therefore, a requirement for the standardisation of the column efficiency and of its selectivity without using one as a factor of the other.A publication by van den Dool12 described a method for the characterisation of GLC columns using the relative retentionindices (RRIs) of a group of six test compounds. By calculating the RRIs of the compounds in the mixture on a particular column and then applying a series of further calculations to these RRIs, van den Dool obtained a figure representing the polarity factor of that particular column which he called the g- † For Part XVI, see ref. 1. ‡ Correspondence should be addressed to the Secretary, Analytical Methods Committee, Analytical Division, The Royal Society of Chemistry, Burlington House, Piccadilly, London, UK W1V 0BN.Analyst, October 1997, Vol. 122 (1167–1174) 1167pack value. This value will vary with the different types of stationary phase and their condition. Van den Dool’s work was based exclusively on packed columns and concentrated on two particular stationary phases: SE-30 as an example of a non-polar type and Carbowax 20M as a moderately polar type. However, the g-pack concept can be applied to any stationary phase.The mixture of compounds used in the determination, known as the NC (Netherlands Committee) mixture, was chosen to represent a range of compounds with functional groups similar to those found in essential oils. Also, the mixture was chosen so that there are two pairs of compounds in which the components of each pair elute close together on the two different phases, which can be used to give a measure of the resolving power of the column.However, the greatly increased efficiency of capillary columns over packed columns means that this property of the mixture is not so significant except in cases of extreme column degradation. The NC mixture consists of limonene, linalol, linalyl acetate, acetophenone, naphthalene and cinnamyl alcohol. The Sub-Committee has published several papers, based on the work of van den Dool, defining the methodology to be used in obtaining these fingerprints13–15 and standard fingerprint traces of selected oils.1,16,17 During these studies on the application of the g-pack concept it was found that the value for a packed column was unaffected by changes in operating parameters such as carrier gas flow rate, temperature programming rate, initial temperature and final temperature hold.However, the value could be decreased by loss of stationary phase over a period of time, due to column bleeding, and could be increased by modification of the stationary phase due to oxidation. The g-pack value can, therefore, give a good indication of the condition of the column.With the wider availability and advances in capillary column technology the Sub-Committee decided that the technique should be updated to make use of this technology and set about determining the optimum operating parameters. It was appreciated that the application to capillary columns was not likely to be as straightforward as with packed columns.A major difference was in the control of the stationary phase. For packed columns, good agreement among laboratories was only obtained when there was control over the preparation of the stationary phase and packing of the column. This involved developing a method for coating the stationary phase onto the support (the absorption coating technique), which led to columns giving good reproducibility and high efficiencies. While it is possible to prepare and coat capillary columns, in reality laboratories buy in their columns from one of the specialist manufacturers. This means that individual laboratories have no control over the preparation of the columns and that different methods of preparation between manufacturers could lead to slightly different performances of nominally the same stationary phase.It was concluded, therefore, that any method for fingerprinting essential oils on capillary columns would have to be based on commercially available columns and be robust enough to cope with a range of operating parameters. If the method is to be widely applicable it is unrealistic to expect laboratories to buy columns from a specific manufacturer or to change carrier gas.Experimental It was decided that the initial examination would be carried out on a specified stationary phase on columns which the members of the committee had available in their laboratories. A non-polar phase was chosen equivalent to SE-30.A protocol was provided specifying the samples to be examined and the initial column temperature, temperature programme rate and final column temperature. The choice of carrier gas was left to individuals with the proviso that the flow rate was optimised for the column configuration. The three samples examined were: a mixture of the NC mix and a series of even numbered carbon aliphatic hydrocarbons (C8–C24) known as the NC–HC mixture; Spanish rosemary oil; and Spanish rosemary oil plus the hydrocarbon mixture.The initial results showed that there were differences between different column manufacturers but very similar results were obtained between laboratories using similar columns from the same manufacturer. It was, therefore, decided to approach the major capillary column manufacturers and invite them to examine the same samples on their columns. All of the manufacturers were very willing to collaborate in this examination with the given protocol.With their inclusion a total of 16 results were obtained. Results The results of the collaborative study using the NC mixture with the proposed procedure are given in Table 1 in ascending order of g-pack values and are summarised in Table 2. It can be seen from Table 2 that the mean RRIs for the test compounds obtained in this examination compare well with results previously obtained on packed columns. Table 3 shows the elution temperatures of the test compounds and n-alkane hydrocarbons.There is a much wider variation in elution temperatures with capillary columns than was found with packed columns. The results from the examination of oil of rosemary are given in Tables 4 and 5. The results obtained were considered satisfactory by the Sub-Committee. Discussion Throughout all of the collaborative exercises, the Sub-Committee was conscious of the fact that many of the parameters examined were chosen with arbitrary limits and that some justification for the choices should be made.In particular, this applies to the nature of the test compounds in the NC mixture and the use of the RRI system. This system is widely used and most gas chromatographers are acquainted with it. It is a simple system and, although it is relative rather than absolute, it allows for a choice of reference compounds which can be made according to other requirements. For the purpose of this and the previous investigation it was felt that the homologous series of n-alkanes was the most suitable as reference compounds.They are readily available in pure form, are extremely stable and are the least likely compounds to exhibit chromatographic anomalies. In addition, they formed the basis of the work by van den Dool. The elution of a variety of homologous series under temperature-programmed conditions is not linear and the departure from linearity increases as the programming rate decreases.18 Also, this departure is greater for homologous series of polar compounds than for the n-alkanes and this was another reason for the choice of the latter.This non-linearity of elution raises the question of the method of application of the system and if an assumption of linearity between successive n-alkanes introduces errors. Van den Dool recommends the use of both even and odd numbered carbon nalkanes to reduce any error, whereas all of our collaborative work has been carried out with the even numbered carbon nalkanes making the assumption of linearity between each pair.Comparisons have been made of the effect on the calculated RRIs of the six NC text compounds when using all of the n- 1168 Analyst, October 1997, Vol. 122alkanes against using only the even numbered carbon ones and also by calculating on the assumption of linearity as well as by a graphical method to obtain more accurate figures. The differences that were found were so small (in some instances zero) that they were considered to be insignificant.Therefore, our recommended procedure uses only the even numbered carbon n-alkanes and assumes linearity between them. The six compounds in the NC mixture were selected by van den Dool on the basis of their similarity to the types of compounds found in essential oil analyses. Their choice could perhaps be criticised if used in conjunction with other sample types. However, the NC mixture is used for the calibration of the column irrespective of its ultimate use.As the mathematical treatment of the results has already been worked out by van den Dool, there seemed to be no advantage in changing the mixture. The choice of six compounds is a reasonable compromise between having an excessive number of interfering peaks and mathematical calculations and a reduced amount of data leading to a less accurate result. With packed columns, more control was exercised over the operating conditions by arranging for the C24 alkane to elute at the upper temperature of the temperature programme run by adjustment of the carrier gas flow rate.This approach is not relevant for capillary columns as, to obtain satisfactory results, it is necessary to optimise the flow rate, which depends on the dimensions of the individual columns and the nature of the carrier gas. The superior resolving power of capillary columns led to the conclusion that measurement of the resolution between the limonene and acetophenone peaks would not be meaningful as there was baseline separation in all cases. The column would be in poor condition or the operating conditions be significantly different from those recommended for these compounds not to be resolved. Developments in gas chromatographic instrumentation have led to discussions on the relative merits of operating the carrier gas system under conditions of constant flow or constant pressure.The NC–HC mixture was chromatographed on a column in an instrument equipped to run under both conditions.As shown below, the RRIs were virtually identical under both conditions. However, the chromatogram of the sample run under constant pressure conditions showed an increase in the retention times of the components compared with those run under constant flow conditions. This is because of a reduction in flow rate during temperature-programmed operation under constant pressure.RRI Compound Constant pressure Constant flow Limonene 1020 1020 Acetophenone 1034 1034 Linalol 1084 1084 Naphthalene 1158 1156 Linalyl acetate 1242 1241 Cinnamyl alcohol 1274 1273 g-pack 0.999 This demonstrates that either method of carrier gas control should be suitable for undertaking fingerprint determinations using this procedure. The RRI and area % composition of the 17 most abundant components in the sample of oil of rosemary were calculated from chromatograms run under the same conditions as for the HC–NC mixture.From the data in Tables 4 and 5, the Sub- Committee considers that the collated results of the collaborative trials show good agreement. The aromatic compounds show greater variability in their RRIs than the non-aromatic compounds. This effect was noted in the previous publications on fingerprinting by the Sub-Committee on packed columns. 13,14 It is attributed to the greater variability in RRI of the aromatic compounds with changes in elution temperature when compared with the non-aromatics.This observation is being studied and will be reported on separately. The data in Table 5 are given as an indication of the composition of oil of rosemary. They assume that all of the compounds have the same response to the flame-ionisation detector and that the oil does not contain any non-volatile material. Also, different integrator parameters will affect the percentage composition data. Conclusion The Sub-Committee recommends the procedure given in the Appendix for the reproducible fingerprinting of essential oils by temperature-programmed GLC using capillary columns with non-polar stationary phases.Although the procedures have been developed and investigated for the analyses of essential oils, it is felt that they have a Table 1 Results of collaborative study on the NC mixture Laboratory 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Stationary phase CPSil5 DB1 BP1 AT1 BP1 007-1 HP1 BP1 DB1 BP1 Rtx1 CPSil5 HP1 CPSil5 HP1 DB1 Column length/m 25 30 25 25 25 25 30 25 60 50 30 25 50 25 25 60 Internal diameter/mm 0.32 0.25 0.22 0.32 0.22 0.25 0.25 0.22 0.25 0.22 0.25 0.25 0.20 0.25 0.32 0.32 Film thickness/ mm 0.25 0.25 0.25 0.30 0.25 0.25 0.25 0.25 0.25 0.25 0.25 0.25 0.50 0.25 1.05 0.25 Carrier gas He H2 H2 He He He He N2 He H2 He He He He He He Test compounds, RRI values— Limonene 1014 1015 1018 1019 1021 1020 1020 1021 1024 1025 1027 1028 1028 1029 1028 1029 Acetophenone 1025 1026 1030 1032 1034 1036 1034 1034 1037 1038 1042 1043 1044 1047 1048 1046 Linalol 1076 1079 1083 1084 1085 1086 1084 1085 1085 1087 1088 1089 1090 1090 1094 1091 Naphthalene 1139 1144 1153 1155 1158 1158 1156 1159 1167 1168 1173 1176 1176 1179 1176 1177 Linalyl acetate 1240 1240 1241 1242 1242 1243 1241 1242 1242 1242 1243 1242 1244 1245 1248 1238 Cinnamyl alcohol 1270 1268 1271 1274 1277 1279 1273 1276 1276 1279 1283 1281 1284 1290 1288 1268 g-pack 0.992 0.994 0.997 0.998 0.999 0.999 0.999 1.000 1.001 1.002 1.004 1.005 1.005 1.006 1.007 1.007 Analyst, October 1997, Vol. 122 1169much wider application and should find use in many other fields of GLC analysis. Appendix Recommended Method for the Reproducible Fingerprinting of Essential Oils by Temperature-programmed Gas–Liquid Chromatography Using Non-polar Stationary Phases The column The preparation of the types of capillary column available today requires considerable experience and expertise. Chromatographers, therefore, have to rely on specialist manufacturers as few have the required skills.However, this study has shown that columns obtained ‘off-the-shelf’ from the major manufacturers were all suitable for use with this method. The stationary phases referred to in Table 1 are the manufacturers’ trade names for polysiloxane phases with no modifications. The dimensions of the column are not critical to the successful application of this method but the following are recommended.Column length, 25–30 m; internal diameter, 0.22–0.25 mm; film thickness, 0.25 mm. A film thickness of more than 3 mm should not be used as different effects are observed. Columns are usually delivered ready for use but any manufacturers’ instructions concerning conditioning should be heeded. Gas chromatographic conditions The temperature control in the ovens of modern gas chromatographs is very accurate. However, if older instruments are used or if any doubt exists, oven settings should be checked using a thermometer of known accuracy.The temperature range for this procedure is 50–250 °C. The programme rate is very important and must be checked to ensure that it is linear. In this case a rate of 4 °C min21 is used. Any of the three commonly used carrier gases, helium, hydrogen and nitrogen, may be used. However, it is most important that the linear gas velocity is adjusted so that the column is operating at optimum efficiency.The actual velocity will depend on the carrier gas and the dimensions of the column. Column manufacturers will advise on this. Carrier gas control may be by either constant flow or constant pressure. Preparation of test mixtures Prepare a mixture of equal masses of the even carbon numbered n-alkanes from C8 to C24. Prepare a mixture of 1.00 part of limonene, 1.37 parts of linalol, 1.60 parts of linalyl acetate, 1.40 parts of acetophenone, 1.13 parts of naphthalene and 1.80 parts of cinnamyl alcohol (NC mixture).Prepare a mixture of 55% m/ m of the NC mixture and 45% m/m of the hydrocarbon mixture. Each n-alkane is then 5% and each NC component is then approximately 10% of the total mixture. These mixtures are now available commercially. Test chromatogram Set up the chromatograph, with the prepared column, for temperature-programmed operation between 50 and 250 °C at 4 °C min21. Inject the combined test mixture and start the programme and integrator.Continue heating at 250 °C until a stable baseline is obtained. Repeat the run if necessary, adjusting the attenuation to bring all the peaks on-scale. A typical chromatogram is shown in Fig. 1. The sample size, Table 2 Summary of results on the NC mixture (Table 1) and comparison with those on packed columns previously examined Relative Capillary columns— Standard standard Test compound Mean RRI deviation deviation (%) Limonene 1023 4.87 0.48 Acetophenone 1037 6.98 0.67 Linalol 1086 4.36 0.40 Naphthalene 1163 12.03 1.03 Linalyl acetate 1242 2.19 0.18 Cinnamyl alcohol 1277 6.48 0.51 Relative Packed columns*— Standard standard Test compound Mean RRI deviation deviation (%) Limonene 1027 1.63 0.16 Acetophenone 1041 2.99 0.29 Linalol 1086 1.80 0.17 Naphthalene 1172 3.55 0.30 Linalyl acetate 1241 1.90 0.15 Cinnamyl alcohol 1280 3.80 0.30 * See ref. 14. Table 3 Elution temperatures of the NC–HC mixture (°C) Laboratory 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Limonene 72 76 87 93 96 91 94 96 117 112 124 125 127 112 122 145 Acetophenone 73 77 89 95 98 93 96 98 119 114 126 127 130 114 125 148 Linalol 79 84 96 103 106 100 102 106 127 122 133 134 137 121 132 155 Naphthalene 86 92 106 113 116 111 112 117 139 134 147 147 150 134 144 169 Linalyl acetate 98 105 118 125 128 123 123 129 151 146 158 157 159 144 154 180 Cinnamyl alcohol 102 108 123 130 133 128 127 133 155 151 164 162 165 150 159 185 g-pack 0.992 0.994 0.997 0.998 0.999 0.999 0.999 1.000 1.001 1.002 1.004 1.005 1.005 1.006 1.007 1.007 C8 59 59 64 69 71 66 70 70 86 80 90 89 92 82 88 112 C10 71 74 85 91 93 88 91 93 113 108 119 120 123 107 118 140 C12 93 99 113 120 122 117 118 123 145 139 151 151 153 137 147 173 C14 118 126 140 148 151 145 143 151 173 169 181 178 181 165 174 209 C16 143 150 165 173 176 170 167 177 199 195 208 202 206 190 198 1170 Analyst, October 1997, Vol. 122dilution and split ratio should be such that the capacity of the column is not exceeded.Calculation of results When a satisfactory chromatogram has been obtained with baseline separation of all peaks, calculate the RRIs of the NC components assuming a linear span between adjacent hydrocarbon peaks. Using the values obtained, tabulate the results as in Table 6 and calculate the g-pack value for the column following the given worked example. This calculation can easily be performed by using a computer spreadsheet. Calculation of relative retention indices (RRIs) RRI = � - + - + [ ( )] ( ) 200 2 100 Rtc Rtn Rt n Rtn n where n = carbon number of 1st hydrocarbon; Rtc = retention time of the compound; Rtn = retention time of 1st hydrocarbon; and Rt(n + 2) = retention time of 2nd hydrocarbon.Table 4 Results of collaborative study on rosemary oil: relative retention indices Laboratory Compound 1 2 3 4 5 6 7 8 9 10 11 12 13 Mean s s (%) a-Pinene 906 909 916 918 922 918 920 928 928 930 937 936 935 923 9.61 1.04 Camphene 918 921 929 930 935 932 934 942 942 946 951 949 950 937 10.41 1.11 Sabinene —* —* 956 957 959 959 959 —* 963 966 966 964 970 962 4.35 0.45 b-Pinene 949 952 960 960 964 963 964 971 971 975 978 976 977 966 9.07 0.94 Myrcene 972 973 977 977 979 979 979 981 981 982 983 982 983 979 3.41 0.35 p-Cymene 1005 1006 1009 1009 1011 1010 1011 1014 1014 1017 1018 1016 1017 1012 4.10 0.41 Limonene/ cineol 1012 1014 1017 1018 1022 1019 1019 1024 1024 1027 1030 1029 1033 1022 6.14 0.60 Linalol 1074 1076 1082 1081 1084 1083 1083 1084 1086 1088 1088 1087 1092 1084 4.70 0.43 Camphor 1100 1106 1114 1117 1121 1118 1118 1127 1127 1132 1136 1135 1140 1122 11.43 1.02 Borneol 1130 1135 1142 1143 1147 1146 1147 1153 1155 1160 1161 1159 1163 1149 9.89 0.86 Terpinen-4-ol 1143 1148 1157 1156 1159 1159 1160 1165 1166 1171 1171 1169 1174 1161 8.84 0.76 a-Terpineol 1157 1161 1161 1167 1171 1170 1171 1175 1177 1181 1181 1179 1184 1172 7.85 0.67 Verbenone 1162 1166 1173 1175 1178 1178 1178 1184 1186 1193 1192 1190 1200 1181 10.57 0.90 Bornyl acetate 1255 1259 1265 1265 1268 1269 1269 1276 1275 1280 1280 1278 1283 1271 8.25 0.65 b-Caryophyllene 1392 1433 1409 1410 1414 1414 1416 1430 1427 1434 1437 1434 1433 1422 13.07 0.92 a-Humulene 1425 1467 1444 1443 1447 1447 1450 1464 1461 1468 1471 1467 1467 1455 13.27 0.91 g-pack 0.992 0.994 0.997 0.998 0.999 0.999 1.000 1.001 1.002 1.004 1.005 1.005 1.007 * Compound not resolved. Table 5 Results of collaborative study on rosemary oil: area % of selected components Laboratory Compound 1 2 3 4 5 6 7 8 9 Mean s s (%) a-Pinene 22.5 18.6 20.2 24.3 19.8 20.4 21.4 20.0 20.3 20.8 1.59 7.6 Camphene 9.2 9.4 8.4 9.8 8.2 8.3 9.1 8.6 8.6 8.8 0.52 5.9 Sabinene 1.3 1.4 1.3 1.4 1.1 1.3 1.2 1.2 1.3 0.10 7.7 b-Pinene 4.5 3.2 2.9 3.2 2.9 2.8 2.9 2.9 2.9 3.1 0.50 16.1 Myrcene 4.2 4.2 4.2 4.3 4.1 3.9 4.1 4.0 3.9 4.1 0.13 3.2 p-Cymene 2.3 2.5 1.9 2.2 1.8 1.8 2.3 1.8 2.1 2.1 0.25 11.9 Limonene/cineol 26.5 26.9 25.2 26.6 25.4 25.6 26.0 26.0 26.0 26.0 0.53 2.0 Linalol 0.8 1.2 1.1 0.9 1.1 1.0 1.1 1.5 1.1 1.1 0.19 17.3 Camphor 17.5 21.2 18.7 16.9 18.9 18.8 18.3 19.4 20.1 18.9 1.22 6.5 Borneol 2.8 3.7 3.2 2.7 3.2 2.8 3.0 2.9 3.2 3.1 0.29 9.4 Terpinen-4-ol 0.7 0.9 0.8 0.7 0.8 0.7 0.8 0.7 0.8 0.8 0.07 8.8 a-Terpineol 1.2 1.5 1.4 1.1 1.4 1.3 1.3 1.4 1.5 1.3 0.13 10.0 Verbenone 1.5 1.9 1.7 1.4 1.6 1.5 1.6 1.6 1.7 1.6 0.14 8.8 Bornyl acetate 0.8 1.2 0.9 0.7 0.9 0.9 0.9 1.0 1.0 0.9 0.13 14.4 b-Caryophyllene 1.8 2.3 2.2 1.5 2.2 2.1 1.9 2.2 2.2 2.0 0.25 12.5 a-Humulene 0.6 0.7 0.7 0.5 0.7 0.6 0.6 0.7 0.7 0.6 0.07 11.7 g-pack 0.994 0.997 0.999 0.999 1.000 1.002 1.004 1.005 1.007 Analyst, October 1997, Vol. 122 1171For example, if the retention times of the C10 hydrocarbon, limonene and C12 hydrocarbon are 17.7, 18.6 and 25.2 min, respectively, then n = 10 and RRI for limonene = [200 (18.6 - 17.7)] 25.2 - 17.7 � + = 1000 1024 The operating conditions and chromatographic system may be considered satisfactory if the results lie witthe following specification: g-pack 1.0 ± 0.005 RRI values: Limonene 1018–1028 Acetophenone 1030–1044 Linalol 1082–1090 Naphthalene 1151–1175 Linalyl acetate 1240–1244 Cinnamyl alcohol 1271–1283 Standardised chromatograms of essential oils When the performance of a column within a chromatographic system has been satisfactorily established according to the procedure given above, it can be used under similar conditions for analysing essential oils.The performance of the system should be checked regularly.A sample of the essential oil should be injected and run under the conditions established above (first chromatogram). On completion of the run, a sample of a mixture of the essential oil and the n-alkane hydrocarbon mixture should be run under Fig. 1 Typical chromatogram of NC–hydrocarbon mixture. Table 6 Calculation of g-pack value Test compound RRI Y factor X factor Z factor Limonene 1020 (RRI30.14) + 2 = 1.0629 1.05843Y = 1.1250 1.058423Y = 1.1907 136.23 Acetophenone 1034 (RRI30.14) + 2 = 1.2216 1.33503Y = 1.6308 1.335023Y = 2.1772 120.14 Linalol 1084 (RRI30.14) + 2 = 0.9969 1.02183Y = 1.0186 1.021823Y = 1.0408 154.24 Naphthalene 1156 (RRI30.14) + 2 = 1.2784 1.33613Y = 1.7081 1.336123Y = 2.2822 128.16 Linalyl acetate 1241 (RRI30.14) + 2 = 0.8954 0.87973Y = 0.7877 0.879723Y = 0.6929 196.28 Cinnamyl alcohol 1273 (RRI30.14) + 2 = 1.3432 1.49933Y = 2.0139 1.499323Y = 3.0194 134.17 Sum SY = 6.7984 SX = 8.2841 SZ = 10.4032 3factor f f1SY = 6.79843 f2SX = 8.28413 f3SZ = 10.40323 1.07977 = 7.3407 2.88734 = 23.9190 1.49758 = 15.5796 g-pack value = f2SX2f1SY2f3SZ = 23.919027.3407215.5796 = 0.9987 1172 Analyst, October 1997, Vol. 122identical conditions (second chromatogram). Comparison of the two chromatograms should show identical retention times and comparisons of the peak heights should be consistent with any dilution due to the n-alkane mixture. These comparisons allow for a check on the reproducibility of the system and also enable the position of the n-alkanes to be transferred from the second chromatogram (Fig. 3) to the first (Fig. 2) in such a manner that visual interference is avoided. In instances where a component of the essential oil overlaps or obscures an n-alkane peak, its position can be determined by comparison with the chromatogram obtained for the g-pack calculation (Fig. 1). Thus, it is possible to determine RRIs for any of the peaks of interest in the essential oil chromatogram. References 1 Analytical Methods Committee, Analyst, 1993, 118, 1089. 2 Analytical Methods Committee, Analyst, 1971, 96, 887. 3 Analytical Methods Committee, Analyst, 1973, 98, 616. 4 Analytical Methods Committee, Analyst, 1973, 98, 823. 5 Analytical Methods Committee, Analyst, 1975, 100, 593. 6 Analytical Methods Committee, Analyst, 1977, 102, 607. Fig. 2 Typical chromatogram of oil of rosemary Spanish. Fig. 3 Typical chromatogram of oil of rosemary–hydrocarbon mixture. Analyst, October 1997, Vol. 122 11737 Analytical Methods Committee, Analyst, 1978, 103, 375. 8 Analytical Methods Committee, Analyst, 1981, 106, 456. 9 Analytical Methods Committee, Analyst, 1987, 112, 1315. 10 Analytical Methods Committee, Analyst, 1988, 113, 657. 11 Analytical Methods Committee, Analyst, 1990, 115, 459. 12 van den Dool, H., Standardisation of G.C. Analysis of Essential Oils, Proefschrift, Rijksuniversiteit te Groningen, Rotterdam, 1974. 13 Analytical Methods Committee, Analyst, 1980, 105, 262. 14 Analytical Methods Committee, Analyst, 1981, 106, 448. 15 Analytical Methods Committee, Analyst, 1984, 109, 1339. 16 Analytical Methods Committee, Analyst, 1984, 109, 1343. 17 Analytical Methods Committee, Analyst, 1988, 113, 1125. 18 Grant, D. W., and Hollis, M. G., J. Chromatogr., 1978, 158, 3. Paper 7/04651K Accepted July 2, 1997 1174 Analyst, October 1997, Vol. 122