J. Cent. South Univ. Technol. (2007)04-0504-05

DOI: 10.1007/s11771-007-0098-9 

Determination of volatile components in cut tobacco with gas chromatography–mass spectrometry and chemometric resolution

HUANG Lan-fang(黄兰芳)^{1, 2}, WU Ming-jian(吴名剑)^{1, 3}, SUN Xian-jun(孙贤军)¹, ZHONG Ke-jun(钟科军)¹,

GUO Zi-ming(郭紫明)¹，DAI Yun-hui(戴云辉)^{1, 2}, HUANG Ke-long(黄可龙)², GUO Fang-qiu(郭方遒)²

(1. Technical Center of Changde Cigarette Factory, Changde 415000, China;

2. School of Chemistry and Chemical Engineering, Central South University, Changsha 410083, China;

3.College of Chemistry, Xiangtan University, Xiangtan 411105, China)

_____________________________________________________________________________

Abstract：Chromatography–mass spectrometry (GC–MS) was used to analyze the volatile components of cut tobacco samples with the help of heuristic evolving latent projections (HELP). After extracting with simultaneous distillation and extraction method, the volatile components in cut tobacco were detected by GC–MS. Then the obtained original two-dimensional data were resolved into pure mass spectra and chromatograms. The qualitative analysis was performed by similarity searches in the national institute of standards and technology(NIST) mass database with the obtained pure mass spectrum of each component and the quantitative results were obtained by calculating the volume of total two-way response. The accuracy of qualitative and quantitative results were greatly improved by using the two-dimensional comprehensive information of chromatograms and mass spectra. 107 of 141 separated constituents in the total ion chromatogram of the volatile components were identified and quantified, accounting for about 88.01% of the total content. The result proves that the developed method is powerful for the analysis of complex cut tobacco samples.

Key words: chemometrics; gas chromatography–mass spectrometry; heuristic evolving latent projection; cut tobacco; volatile component

_____________________________________________________________________________

1 Introduction

Cigarette consists of numerous materials, such as cut tobacco, paper material and additives. Obviously cut tobacco is one of the most important materials. The volatile components are the main compositions in cut tobacco, which contribute to the flavor of cigarettes during smoking. Therefore volatile component in cut tobacco is a very important factor to appraise the quality and the commercial values of tobacco. Many methods have been reported for the determination of the volatile components in cut tobacco^[1-5]. Usually chromatography– mass spectrometry (GC-MS) is used for the determination of the concentrations of the volatile components in cut tobacco. However, although extraction and concentration methods are used to the complicated samples such as cut tobacco^[1-3], it is still difficult to achieve complete separation even if rigorous conditions are imposed on the chromatographic separation process. Because the overlapping chromatographic peaks always present a problem for similar search from the MS library, only very few chemical components could be identified. Furthermore, quantitative analysis cannot be performed to the overlapped peaks, since it would be also difficult to determine the area of each component. So the results obtained by the above mentioned would be questionable.

These years with the development of hyphenated instruments, multidimensional data revealing the compositions of samples obtained from GC-MS, HPLC-DAD can be resolved with chemometric methods. Many associated methods, such as evolving factor analysis (EFA)^[6], window factor analysis (WFA)^[7], heuristic evolving latent projections (HELP)^[8], subwindow factor analysis (SFA)^[9], evolving window orthogonal projection (EWOP)^[10] and orthogonal projection resolution (OPR)^[11], have been developed to provide more information for chemical analysis both in chromatographic separation and in spectral identification, which makes it possible to interpret these complex systems^[12-14].

In this paper, a method for the determination of the volatile components in cut tobacco with gas chromatography–mass spectrometry and chemometric resolution was developed. After extracting with simultaneous distillation and extraction method, the volatile components in cut tobacco were detected by GC–MS under appropriate chromatographic conditions. Then the pure chromatogram and mass spectrum of each component can be obtained through unique resolution from the obtained two-dimensional data with the HELP method. Qualitative identification of each component was performed using retention times and mass spectra. Finally, the quantitative analyses were carried out with the overall volume integration method^[14].

2 Experimental

2.1 Instruments

A Hewlett-Packard 6890 gas chromatograph interfaced with a Hewlett-Packard mass selective detector 5973N (Agilent Technologies, USA) operated by HP enhanced ChemStation software, G1701DA MSD ChemStation Rev. D.00.00.38.  was employed in this study. Simultaneous distillation and extraction (SDE) apparatus was manufactured by University of Science and Technology of China (Hefei, China) and R2003KE rotary evaporator was manufactured by Shanghai Senco Technology Co. Ltd.(Shanghai, China).

2.2 Materials and regents

Cut tobacco from Fujian Province of China was provided by Technical Center of Changde Cigarette Factory. Dichloromethane and sodium sulfate were of analytical grade.

2.3 Extraction of volatile components

The volatile components in cut tobacco were extracted with simultaneous distillation and extraction method with a micro version apparatus^[15]. Cut tobacco was dried at 35 ℃ for 2 h and ground into powder at first. Then, for each extract, 25.0 g of cut tobacco was weighed and placed in a 1 000 mL flask. 150 g sodium sulfate and 350 mL redistilled water were also placed in this 1 000 mL flask. 40 mL dichloromethane was placed in a 100 mL flask, which was placed in a 60 ℃ water-bath. They were both distilled for 2.0 h at atmospheric pressure. At last, about 40 mL extract was obtained. The extract was condensed to 1 mL with a rotary evaporator.

2.4 Detection of volatile components

GC-MS was used to obtain chromatograms of the extract. The column fitted with a HP-5MS fused silica column (5% phenyl methyl polysiloxane 30 m×0.25 mm i.d., film thickness 0.25 μm) was used. The oven was held at 70 ℃ for 1 min during injection, then temperature was programmed at 3 ℃/min to a final temperature of   210 ℃ and held for 5 min. Inlet temperature was kept at 270 ℃ all the time. 5.0 ?L of essential oil was injected into the GC. Helium carrier gas at a constant flow-rate of 1.0 mL/min and a 5:1 split ratio were used simultaneously. Mass spectrometer was operated in full scan and electron impact modes with an electronic energy of 70 eV. Interface temperature: 270 ℃; MS source temperature: 230 ℃; MS quadrupole temperature: 160 ℃. In the range of m/z from 30 to 500, mass spectra were recorded with velocity of 3.12 s/scan.

2.5 Data analysis

Data analysis was performed on a Pentium based IBM compatible personal computer. All programs of the chemometrical resolution methods were coded in MATLAB 6.3 for windows. The library searches and spectral matching of the resolved pure components were conducted on the National Institute of Standards and Technology (NIST) MS database containing about 107 000 compounds.

3 Results and discussion

3.1 Component resolution

The real total ionic current (TIC) chromatogram of the volatile components of cut tobacco sample from Fujian is shown in Fig.1. As could be seen from it, there are a lot of peaks and their contents vary greatly. Although chromatographic separation was optimized, some of eluted components overlapped with one another and the concentrations of many volatile components were very low. If directly searched in the NIST mass database without further data processing, incorrect identification of compounds may be obtained. First, the similarity indices (SIs) obtained from direct searching with the NIST MS database were quite low for most of these chromatographic peaks and sometime the same component was searched at different chromatographic scan points. Furthermore, the component with low content was very difficult to be identified straight with the NIST mass database, since two-dimensional data obtained by mass spectral measurement unavoidably contained peaks associated with column background and residual gases. However, if the overlapped peaks and the components with low content were resolved into pure spectra and chromatograms with HELP method, the qualitative analysis of components would be improved to a reliable extent.

Fig.1 Total ion chromatogram of volatile components in cut tobacco from Fujian area, China

Brief depiction of HELP method was showed here, because it has been described in detail in the Ref.[8].

According to the Lambert-Beer law, a two-dimensional data A produced by GC–MS can be represented as follows:

        (1)

where  A denotes an absorbance matrix expressing i components of m chromatographic scan points at n atom mass units or wavelength points, C is the pure chromatographic matrix, S is the pure mass spectral matrix and E represents the noise. The unique resolution of a two-dimensional data into chromatograms and spectra of the pure chemical constituents was carried out with HELP method.

Divide the matrix (A) into many submatrix and take the peak cluster B within 15.65-16.05 min as example to demonstrate the whole analysis procedure mentioned above(Fig.2).

Fig.2 TIC chromatograms of peak cluster B

1—Original chromatogram; 2—Chromatogram after background removing

Curve 1 in Fig.2 is an original chromatogram from 15.65 to 16.05 min (peak cluster B in Fig.1). It can be seen that the frontal peak is not baseline separated from the latter peak. By direct search from mass library, matching values are quite low. However, if this overlapped peak is resolved into pure chromatographic profiles and mass spectra, matching values would be improved to an acceptable extend.

At first, background was subtracted with principal component projection analysis(PCA)^[9-10]. Curve 2 in Fig.2 is the chromatogram after background removing. Obviously, baseline is down to the bottom. Then the rank map was obtained with fixed size moving window eloving factor analysis(FSMWEFA)^[16-17] and is shown in Fig.3.

Fig.3 Evolving eigenvalues(e) of peak B obtained using FSMWEAF with window size of 8

From the rank map, the chromatographic eluting order can be determined and so can the number of components in the system, the selective regions and zero-concentration regions of all the constituents. With all the information determined, the two-dimensional data matrix was uniquely resolved into pure chromatographic profiles and mass spectra of all components. The corresponding obtained pure chromatographic profiles are shown in Fig.4. Resolved mass spectrum of each component is shown in Figs.5 and 6, respectively. Other peaks in the studied sample were also performed in the same way as described as above.

Fig.4 Resolved chromatographic profiles of peak cluster B

1—Resolved chromatogram of component 1; 2—Resolved chromatogram of component 2

Fig.5 Resolved mass spectrum of component 1 by SFA(a) and standard mass spectrum of 9-methyl-5-methylene-8- decen-2-one (C₁₂H₂₀O) (b)

Fig.6 Resolved mass spectrum of component 2 by SFA(a) and standard mass spectrum of [1α, 1β, 3β, 6α]-1, 2-cyclohexanediol(C₁₀H₂₀O₂)(b)

3.2 Qualitative analysis

The qualitative analysis was performed by similarity searches in the NIST mass database with the obtained pure mass spectrum of each component. The standard spectrum of each searched component in peak cluster B from the NIST MS library is also shown in Figs.5 and 6, respectively. The result shows that these two components in peak cluster B are 9-methyl-5-methylene-8- decen-2-one (C₁₂H₂₀O) and [1α, 1β, 3β, 6α]-1, 2-cyclohexanediol (C₁₀H₂₀O₂) with the match values 0.923 and 0.916.

Identification of components in other peaks can be performed in the same way. 141 peaks were separated, but only 107 were identified here. Two reasons resulted in this. One is that ratios of signal to noise of some components are too low, and the other is that some of the components are not included in the NIST MS database. If some of these unidentified constituents were of interest, further study seemed to be necessary.

3.3 Quantitative analysis

The quantitative results were obtained by calculating the volume of total two-way response, namely C_iS_i^T. Similar to the general chromatographic quantitative method with peak area or height, the concentration of each component is proportional to the overall volume of its two-way response (C_iS_i^T). The advantage of this quantitative method over general peak-area integration is that all mass spectral absorbing points are taken into consideration. The identified components amounted quantitatively to 88.01% of the total content. The identified components, whose relative concentrations were above 0.5%, are listed in Table 1, where β is relative content of each component.

Table 1 Identification and quantification of volatile components in cut tobacco from five different sources

4 Conclusions

1) A gas chromatography–mass spectrometry (GC–MS) is developed for analysis of constituents in essential oils of cut tobacco with the help of combined chemometric method, such as PCA, FSMWEFA and HELP. 107 of 141 separated volatile constituents in cut tobacco are identified and quantified, accounting for about 88.01% of the total content.

2) This study shows the developed method can reveal chemical constituents of cut tobacco and can be used for the quality control of cut tobacco.

References

[1] CLARKSON P, MICHAEL C. The identification of an unusual volatile component in processed tobacco by gas chromatography with mass spectrometry and atomic emission detection[J]. Analytica Chimica Acta, 1996, 335(3): 253-259.

[2] PENG Fu-min, SHENG Liang-quan, LIU Bai-zhan, et al. Comparison of different extraction methods: Steam distillation, simultaneous distillation and extraction and headspace co-distillation, used for the analysis of the volatile components in aged flue-cured tobacco leaves[J]. Journal of Chromatography A, 2004, 1040(1): 1-17.

[3] WU Ming, ZHAO Ming-yue, ZHAO Xiao-dong, et al. Comparative analysis of several cut tobaccos from inside and outside China[J]. Chinese Tobacco Acta, 2002, 8(4): 1-9.(in Chinese)

[4] GORDON B M, UHRIG M S, BORGERDING M F, et al. Analysis of flue-cured tobacco essential oil by hyphenated analytical techniques[J]. J Chromatogr Sci, 1988, 26(4): 174-180.

[5] CHIDA M, SONE Y, TAMURA H. Aroma characteristics of stored tobacco cut leaves analyzed by a high vacuum distillation and canister system[J]. J Agric Food Chem, 2004, 52(26): 7918-7924.

[6] MAEDER M, ZILIAN A. Evolving factor analysis: A new multivariate techniques in chromatography[J]. Chemom Intell Lab Syst, 1988, 3(3): 205-213.

[7] DEN Wei, EDMUND R. Malinowski. Investigation of copper (II)-ethylenediaminetetraacetate complexation by window factor analysis of ultraviolet spectra[J]. Journal of Chemometrics, 1993, 7(2):89-98.

[8] LIANG Y Z, KVALHEIM O M, KELLER H R, et al. Heuristic evolving latent projections: Resolving two-way multicomponent data. Part 2: Detection and resolution of minor constituents[J]. Anal Chem, 1992, 64(8): 946-953.

[9] MANNE R, SHEN H L, LIANG Y Z. Subwindow factor analysis[J]. Chemom Intell Lab Syst, 1999, 45(1/2): 171-176.

[10] XU Chen-jian, JIANG Jian-hui, LIANG Yi-zeng. Evolving window orthogonal projections method for two-way data resolution[J]. Analyst, 1999,124 (10): 1471-1476.

[11] SANCHEZ F C, RUTAN S C, GIL GARCIA N D. Resolution of multicomponent overlapped peaks by the orthogonal projection approach, evolving factor analysis and window factor analysis[J]. Chemom Intell Lab Syst, 1997, 36(2): 153-164.

[12] GUO Fang-qiu, LIANG Yi-zeng, XU Cheng-jian, et al. Determination of the volatile chemical constituents of Notoptergium Incium by gas chromatography-mass spectrometry and iterative or non-iterative chemometrics resolution methods[J]. Journal of Chromatography A, 2003, 1016(1): 99-110.

[13] HUANG Lan-fang, LI Bo-yan, LIANG Yi-zeng, et al. Application of combined approach to analyze the constituents of essential oil from angelica sinensis[J]. Analytical and Bioanalytical Chemistry, 2004, 378(2): 510-517.

[14] WU Ming-jian, SUN Xian-jun, DAI Yuan-hui, et al. Determination of constituents of essential oil from angelica sinensis by gas chromatography–mass spectrometry and chemometric resolution[J]. Journal of Central South University of Technology, 2005, 12(4): 250-254.

[15] CAI J B, LIU B Z, LING P, et al. Analysis of free and bound volatiles by gas chromatography and gas chromatography–mass spectrometry in uncased and cased tobaccos[J]. Journal of Chromatography A, 2002, 947(2): 267-275.

[16] KELLER H R, MASSART D L. Peak purity control in liquid chromatography with photodiode-array detection by a fixed size moving window evolving factor analysis[J]. Anal Chim Acta, 1991, 246(4): 379-390.

[17] LIANG Yi-Zeng, KVALHEIM O M, RAHMANI A. Resolution of strongly overlapping two-way multicomponent data by means of heuristic evolving latent projections[J]. Journal of Chemometrics, 1993, 7(1):15-43.

_____________________________

Foundation item: Project supported by the Postdoctoral Foundation of Changde Cigarette Factory; Project(20060400887) supported by China Postdoctoral Science Foundation

Received date: 2006-12-24; Accepted date: 2007-01-27

Corresponding author: HUANG Lan-fang, PhD, Professor; Tel: +86-731-8836376; E-mail: lf18huang@yahoo.com.cn

(Edited by YANG Hua)