Determination of the volatiles from tobacco by capillary gas chromatography with atomic emission detection and mass spectrometry

Research Article

Determination of the volatiles from tobaccoby capillary gas chromatography withatomic emission detection and massspectrometry

A new gas chromatograph-atomic emission detector (GC-AED) coupled with Deans

switching technique for analyzing volatiles from tobaccos were developed. The detector

operating parameters (reagent gas pressure and make-up gas flow rate) were optimized.

The detection limits for the elements carbon (193 nm), hydrogen (486 nm) and oxygen

(171 nm) ranged 0.05–0.2, 0.05–0.3 and 1–11 ng, respectively, depending on the

compound. The sensitivity and linearity for the elements carbon (193 nm), hydrogen

(486 nm) and oxygen (171 nm) decreased in the order O4H4C. Calibration curves were

obtained by plotting peak area versus concentration, and the correlation coefficients

relating to linearity were at least 0.9359. Elemental response factors measured on these

channels, relative to the carbon 193-nm channel, were hydrogen, 0.38–0.48 (mean

%RSD 5 5.64), and oxygen, 0.085–0.128 (mean %RSD 5 14.9). The evaluation was also

done for the new technique and for an established GC-MS technique for the same real

samples. The results of GC-AED and GC-MS showed that there was a relatively good

agreement between the two sets of data.

Keywords: Deans switching / Elemental response factors / GC-AEDDOI 10.1002/jssc.201100732

1 Introduction

The volatile constituents of tobacco leaf are of interest since

their structures will provide information about the cigarette

flavor. Numerous studies on volatile components of the

cigarette and its smoke have been carried out, and hundreds

of components have been reported [1–5]. A sampling

method is required, such as liquid–liquid extraction (LLE)

[6–8], simultaneous distillation–extraction [9, 10], solid-

phase extraction [11], a combination of the former

techniques [12] or solid-phase microextraction [13], which

will provide sufficient quantities of material to obtain

information about the trace components, since many of

the important volatile compounds occur only in very low

concentration.

Gas chromatography with mass spectrometric detection

(GC-MS) has become the main analytical tool for the

determination of the structure and elemental composition

of the components of complex mixtures, for example,

traditional Chinese medicine [14] or herbal drugs [15].

Furthermore, it has been suggested that other technical

improvements, such as the Deans switch microfluidic

device, offer more options and potential for problem-solving

and optimization of GC-based analytical systems [16, 17].

The Deans switching device should make easy set-up for

multidimensional gas chromatographic analysis and heart-

cutting properties possible.

Recently, gas chromatography with atomic emission

detection (GC-AED) is a powerful technique that offers

high-resolution separation of components in a complex

matrix and highly selective spectrometric detection [18–21].

GC-AED provides information about the occurrence of

individual elements in the analyzed compounds. Monitor-

ing at different wavelengths is carried out to measure the

atomic emission of the chosen elements, and data are

recorded as element selective chromatograms. Furthermore,

information about possible empirical formulae can be

obtained by calibrating the instrument with standard

compounds of known composition and comparing it with

the response of the unknown impurity. However, the large

amount of the solvent comes with a risk of rapidly

contaminating the fused-silica discharge tube of the AED

Gang Li1,2

Da Wu2

Ye Wang2

Wenyan Xie2

Xiangmin Zhang1

Baizhan Liu2�

1Department of Chemistry, FudanUniversity, Shanghai,P. R. China

2Technology Center, ShanghaiTobacco Group Co., Ltd,Shanghai, P. R. China

Received August 17, 2011Revised October 24, 2011Accepted October 24, 2011

Abbreviations: ERF, elemental response factor; GC-AED, gaschromatography with atomic emission detection; GC-MS,gas chromatography with mass spectrometry; LLE,liquid–liquid extraction; psi, pounds per square inch

�Additional correspondence: Professor Baizhan Liu

E-mail address: [email protected]

Correspondence: Professor Xiangmin Zhang, Department ofChemistry, Fudan University, Shanghai 200433, P. R. ChinaE-mail: [email protected]: 186-21-65643983

& 2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com

J. Sep. Sci. 2012, 35, 334–340334

detector, potentially limiting its usefulness, and requiring

frequent system maintenance.

It is worthwhile to stress that complementary informa-

tion can be gathered from GC-AED coupled with Deans

switch. The potential merit of this system could be also

applied to the multidimensional GC technique. Separation

may be enhanced through ‘heart-cutting’ peaks from one

GC column onto a second GC column with a stationary

phase of different selectivities using a Deans switch

assembly. For added benefit, the solvent is not introduced in

the second column or in the AED detector, avoiding

contaminating the fused-silica discharge tube, column

degradation, target analyte peak detection/integration and

frequent system maintenance issues.

As far as we know, the application of GC-AED coupled

with Deans switching device as a routine tobacco screening

tool has not yet been reported. In view of this, in this

contribution, we describe an approach using GC-AED

coupled with Deans switching for the analysis of volatile

fractions of tobaccos. Detection limits, regression analyses

and elemental response factors (ERFs) are determined for

carbon, hydrogen and oxygen, although the main aim of this

study is to evaluate the ability of quantitative GC-MS and

GC-AED coupled with Deans switch when analyzing

tobacco samples.

2 Materials and methods

2.1 Samples and chemicals

Typical aged flue-cured tobacco leaves from different

provinces in China were prepared for these studies. A 10-g

aliquot of each sample was collected and mixed thoroughly,

put into an oven at 401C for 8 h to remove moisture, and

then ground to 40–60 mesh powder.

Anhydrous dichloromethane used as solvents from

Merck (Darmstadt, Germany) was of analytical grade

quality. It was used without further purification. Phenethyl

acetate (Sigma, St. Louis, MO, USA) was used as an

internal standard. Stock anhydrous dichloromethane stan-

dards of phenylmethanal, benzenemethanol, b-damascen-

one, solanone, (E)-geranylacetone and b-ionone were all

purchased from Sigma and were all in GC purities. Stock

standard solutions of 500 mg/mL of each compound were

prepared by dissolving 5 mg in 10 mL of dichloromethane

and were used to optimize GC conditions. All were refri-

gerated at 41C during storage. Anhydrous sodium sulfite

(J&K Scientific, Shanghai, China) was prepared before use.

The plasma gas and carrier gas used for GC was helium

(99.9999% purity). A VICI Model I-23572-HP2 (USA)

helium purifier was used inline between the helium tank

and the plasma cavity. The reagent gases for the AED

were oxygen, hydrogen and 10% methane-in-nitrogen.

Nitrogen was used for purging the AED system. All the

gases were supplied by Shanghai No. 5 Steel Works

(Shanghai, P. R. China).

2.2 Isolation of the volatile components

A 10.0-g tobacco sample was placed in a 500-mL flask,

combined with 350 mL water, previously salted with 10%

anhydrous sodium sulfite. The solution was water steam

distilled for 20 min. A distillate of 150 mL was collected in a

separation funnel to perform a LLE, which was extracted by

anhydrous dichloromethane (each 15� 3 mL) under

mechanical shaking for 5 min. The organic phases were

separated from the aqueous phase, and the funnel was

washed with 10 mL of dichloromethane. The resulting

extract layers were combined and then were added to

150 mL of internal standard solution (4.5 mg/mL). The

organic phases were collected in a flask and concentrated

to dryness using a rotary vacuum evaporator at 401C. The

dry extract was dissolved in 1 mL of dichloromethane. No

clean-up was necessary. All glasswares were cleaned and

heated for 2 h at 1501C prior to use.

2.3 Chromatography

2.3.1 GC-AED

The AED used in this study was a G-2350 Model (Joint

Analytical Systems GmbH, Germany) equipped with a

photodiode array (PDA) having a nominal wavelength range

of 160–800 nm and a spectral resolution of 0.1 nm at

400 nm. The concave holographic grating had a flat focal

plane along which the PDA could be moved. Plasma power

was supplied by a microwave magnetron tube. The AED had

been described in detail elsewhere [22–24]. The 1.0-mm

id� 47-mm long fused-silica discharge tube was water

cooled and fitted with a fused-silica window, which was

purged with 35 mL/min of reagent grade helium. Groups of

elements were scanned, without background correction, for

carbon, hydrogen and oxygen at 193, 486 and 171 nm,

respectively.

To vent the solvent before entering the plasma and

prevent carbonaceous deposition on the wall of the

discharge tube, the analyses were carried out on the GC-

AED coupled with Deans switching device combination. As

illustrated in Fig. 1, the GC was equipped with a split/

splitless inlet and a capillary flow technology based Deans

switching system. The primary and the second columns

were both 30 m� 0.25 mm id� 0.25 mm HP-5 (Agilent

Technologies) installed in the GC oven. A restrictor, R1

(3 m� 180 mm id deactivated fused silica, Agilent Technol-

ogies), was connected between the second output of the

Deans switch and a monitoring flame ionization detector

(FID). In our experiment, the resolution of all examined

peaks obtained with the system was sufficient, thus using

the columns with a stationary phase of different selectivities

coupled with Deans switch assembly was not necessary.

The Deans switching operation was demonstrated

previously [25, 26]. In this configuration, a solenoid valve,

located outside the oven, was in the off position. The

J. Sep. Sci. 2012, 35, 334–340 Other Techniques 335


effluent from the first column was ‘pushed’ with an addi-

tional flow towards the monitor FID detector. When the

solenoid valve was actuated (from the software), the flow

now entered the other arm to the Deans switching. And the

flow from the first column was redirected into the second

column. In both valve positions, pressures through restric-

tor R1 and column 2 were similar.

Sample volumes of 1 mL were injected in splitless mode

at an injection port temperature of 2801C, applying a pres-

sure pulse of 26 pounds per square inch (psi). The initial

column temperature was set at 451C for 1 min and then

increased to 2501C (2.51C/min, 5 min hold). The pressure at

the outlet of the first column and the inlet of the second

column was 18.4 psi, resulting in a flow of 2 mL/min

through the second column (to AED). The final part of the

GC column was used as a transfer line to the detector. The

transfer line and the detector cavity were maintained at

2801C. The scavenger gases were hydrogen, oxygen and 10%

methane-in-nitrogen at 11.5, 21 and 30 psi, respectively. The

spectrometer was purged with nitrogen at 400 mL/min.

Filter and back amount adjustments were set according to

the default specifications.

2.3.2 GC-MS

Samples were analyzed with a Hewlett-Packard Model 7890

GC interfaced to a Model 5975A mass-selective detector. A

30-m� 0.32-mm capillary column coated with 0.25-mm HP-

5 was used for all analyses. The inlet and the detector were

set at 2801C, respectively. The MS ion source was heated to

2301C, and the quadrupole was kept at 1501C. Their

collision energy for MS fragmentation was at 70 eV,

scanning from m/z 40 to 330 in one scan. The mass

spectral identification of the compounds were carried out by

comparing with the NIST98 (National Institute of Standards

and Technology, Gaithersburg, MD, USA) mass spectral

library as well as to the Wiley 7.0 (Wiley, New York, USA)

mass spectral library. Qualitative analysis (mass spectral

data) was verified by comparing the retention indices and

mass spectra of identified compounds with those of

authentic reference substances. Quantitative data were

obtained by the internal standard method using phenethyl

acetate as the internal standard or standards as reference

substances, respectively, without considering calibration

factors (i.e. F 5 1.00 for all compounds). The GC column

and chromatographic conditions and settings were the same

as those used for GC-AED runs.

3 Results and discussion

3.1 Optimization of AED parameters

With GC-AED, the elements were in groups on the

condition that close emission line wavelengths and the

same scavenger gases were used [27]. For this reason, three

sequential chromatographic runs were required to obtain

chromatograms for C, H and O. The chromatogram for

carbon (193 nm) was obtained using oxygen and hydrogen

as scavenger gases. While the emission line for hydrogen lay

in 486 nm, this element required oxygen as the sole

scavenger gas. The chromatogram for oxygen (171 nm)

was obtained using hydrogen and 10% methane-in-nitrogen

as the scavenger gas. Their presence, combined with the

retention time, would usually be to identify a given sample.

The detector operating parameters (reagent gas pressure

and make-up gas flow rate) were optimized to obtain the

highest degree of sensitivity. For those emission lines that

required two scavenger gases simultaneously, independent

optimization of both the oxygen and hydrogen pressures was

carried out. First, the hydrogen pressure was held constant at

5 psi, the default value provided by the software, while oxygen

was varied from 5 to 30 psi to determine the optimum oxygen

pressure. For the C (193 nm) emission line, an increase in

oxygen pressure led to a decrease in peak area up to 15 psi,

which remained almost constant thereafter. A pressure of

20 psi was adopted to avoid accumulation of elemental

carbon in the AED discharge tube. The oxygen reagent gas

pressure was then held constant at its optimum value,

whereas the hydrogen pressure was varied from 5 to 25 psi.

The optimum value of the hydrogen pressure was 11.5 psi.

Using the same procedure, 30 psi of the 10% methane-in-

nitrogen was selected as the optimum value when analyzing

the element of the oxygen (171 nm).

The supplementary helium added to the column flow

prior to detection (make-up gas) affects the sensitivity and

peak shape of the components. To determine the optimum

helium (make-up gas) flow rate, which provided the maxi-

mum sensitivity, the internal standard solution was injected

and the flow rate was varied from 20 to 100 mL/min. A flow

rate of 70 mL/min provided the maximum sensitivity for all

the components at the C (193 nm) emission line. Monitor-

ing the same emission line showed very wide peaks and

Figure 1. Configuration of GC-AED coupled with Deans switch-ing device.

J. Sep. Sci. 2012, 35, 334–340336 G. Li et al.


overlapping at flow rates slower than the optimized value.

Under the optimized conditions, standard solutions deter-

mined eluted in the following order: phenylmethanal,

benzenemethanol, phenethyl acetate, b-damascenone, sola-

none, (E)-geranylacetone and b-ionone.

3.2 Optimization of the extraction procedure

Good results were obtained by LLE. Preliminary experi-

ments were carried out using several organic solvents

(hexane, isooctane, diethyl ether, dichloromethane and ethyl

acetate) and mixtures. The extraction procedure using

dichloromethane provided excellent results. In an attempt

to improve the extraction efficiency by increasing the ionic

strength of the samples, sodium sulfite was added to the

samples, which could improve the extraction of these

components with a high degree of water solubility [28].

The extraction percentages were optimal for a 10:1 water/

dichloromethane ratio. Higher sodium sulfite concentra-

tions did not improve extraction percentages [29, 30].

Different shaking times were assayed and 5 min was

sufficient to obtain the highest extraction percentage using

the proposed GC-AED and GC-MS method. After extraction,

the organic phase was evaporated to dryness and the dry

extract was dissolved with 1 mL of dichloromethane.

3.3 Detection limits and regression analysis

In this experiment, working standard solutions were

prepared at different concentrations. The C (193 nm), H

(486 nm) and O (171 nm) emission lines were monitored.

Each concentration was injected into the GC five times to

improve the quality of the information.

Detection limits were calculated using a signal-to-noise

ratio of 3 for all investigated compounds and values are

given in Table 1. The analyte detection limits for the

nonmetals carbon, hydrogen and oxygen were 0.05–0.2,

0.05–0.3, and 1–11 ng, respectively. This indicated that the

order of decreasing sensitivity to molecular structure was

O4H4C. These detection limit ranges also suggested that

errors using molecular-independent quantification (the use

of a single standard to quantify a series of compounds

having similar elemental distribution but different mole-

cular structures) would be larger for oxygen than for carbon

or hydrogen.

The correlation coefficients derived from linear regres-

sions and the standard errors of the estimates are also

included in Table 1. Linear calibration curves were obtained

by plotting peak areas versus concentrations. The results

Table 1. Calibration parameters at different emission lines

Analyte GC-AED

Carbon 193 nm Hydrogen 486 nm Oxygen 171 nm

Slope r2 LOD (ng) Slope r2 LOD (ng) Slope r2 LOD (ng)

Phenylmethanal 111.3 0.9985 0.15 42.1 0.9818 0.3 29.6 0.9507 1.5

Benzenemethanol 79.75 0.9986 0.15 38.6 0.9857 0.2 34.8 0.9423 2

Phenethyl acetate 85.53 0.9996 0.2 46.4 0.9941 0.2 31.5 0.9494 1

b-Damascenone 95.62 0.9993 0.05 58.3 0.9954 0.15 27.8 0.9451 8

Solanone 125.1 0.9997 0.05 67.8 0.9963 0.05 14.4 0.9007 6

(E)-Geranylacetone 82.33 0.9995 0.05 63.5 0.9966 0.075 10.9 0.9303 8

b-Ionone 99.81 0.9997 0.05 71.1 0.9958 0.09 10.6 0.9328 11

Table 2. Elemental response factors (ERF)a)

Analyte H (486 nm) (%RSD) O (171 nm) (%RSD)

Furfural 0.455 (10.6) 0.128 (8.2)

2-Furanmethanol 0.42 (9.7) 0.123 (10.6)

Protoanemonine 0.41 (9.8) 0.119 (7.5)

Phenylmethanal 0.425 (11.6) 0.103 (15.4)

6-Methyl-5-hepten-2-one 0.385 (6.85) /

Benzenemethanol 0.4 (6.23) 0.123 (16.2)

Benzeneacetaldehyde 0.39 (7.21) 0.123 (17.7)

2-Acetylpyrrole 0.405 (8.6) 0.118 (15.5)

Phenylethyl alcohol 0.6 (5.79) 0.121 (6.8)

Indole 0.465 (5.88) /

4-Vinylguaiacol 0.38 (6.15) 0.118 (6.2)

Solanone 0.42 (5.03) 0.121 (24.1)

b-Damascenone 0.395 (6.89) /

(E)-b-Bamascone 0.425 (6.57) /

(E)-Geranylacetone 0.435 (4.38) 0.106 (20.5)

b-Ionone 0.385 (7.2) /

Dihydroactinidiolide 0.455 (2.97) 0.085 (13.7)

Megastigmatrienone 1 0.415 (2.76) /

Megastigmatrienone 2 0.39 (3.32) 0.118 (19.5)

Megastigmatrienone 3 0.405 (2.05) /

Megastigmatrienone 4 0.44 (2.15) 0.12 (18.3)

3-Oxo-a-ionol 0.43 (5.28) /

Neophytadiene 0.44 (1.15) /

Farnesyl acetone 0.42 (2.43) 0.102 (17.8)

Methyl palmitate 0.46 (3.77) 0.117 (20.3)

Palmitic acid 0.48 (2.33) /

Mean ERF (mean %RSD) 0.428 (5.64) 0.115 (14.9)

a) (Elemental area counts/carbon 193 area counts)� (no. of

carbon atoms/no. of element atoms).



obtained showed the following order of sensitivity:

C4H4O, as measured by the coefficient of determination

(r2). The poorer precision for oxygen could result from

chromatographic tailing resulting from the greater polarity

of these compounds relative to the neutral analytes.

3.4 Elemental response factor

An ERF was calculated from the average of ERF for each

concentration level used for regression analysis using the

expression [31]. ERF 5 (elemental area counts/carbon 193-

nm area counts)� (no. of carbon atoms/no. of element

atoms). Table 2 shows the ERF data for the hydrogen 486-

nm and oxygen 171-nm channels. The order of increasing

%RSD (mean value) was H (5.64)oO (14.9), and the order

of increasing ERF was O (0.115)oH (0.428).

In principle, the very high temperatures in the MIP

plasma should result in the complete atomization of analyte

molecules entering from the GC column so that AED

response should be independent of molecular structure. In

fact, compound-independent calibration had not proved

Figure 2. Total ion chromatogram of volatilecomponents obtained from flue-cured tobac-co from Henan in China. For identified peaknumbers, see Table 3.

Figure 3. Typical GC-AEDchromatograms of volatilecomponents obtained fromflue-cured tobacco from Henanin China by element wave-length: (A) C 193 nm and (B) O171 nm. For identified peaknumbers, see Table 3.

J. Sep. Sci. 2012, 35, 334–340338 G. Li et al.


reliable [32, 33] for electron-rich aromatic structures with

large dissociation energies because of incomplete decom-

position and/or the formation of small molecular species by

recombination reactions [34]. Similar results were observed

here: the ERF oxygen and hydrogen for different

compounds extracted from tobacco determined under

the same analytical conditions differed slightly (Table 2).

The poorer precision for O might have been based on the

detector performance at the wavelength or self-absorption

between the emission source and the spectrometer.

3.5 Analysis of flue-cured tobacco samples

Complementary measurements for the proposed LLE

method were also performed on a GC-MS system with

exactly the same extracts as used for the GC-AED, with the

aim of confirming the identity of the compounds detected

by GC-AED. The selectivity of the AED was studied by

comparing GC-MS and GC-AED chromatograms of real

samples. The total ion chromatogram and the typical GC-

AED chromatograms of the volatiles from Henan province

are shown in Figs. 2 and 3, respectively. The selectivity of

the AED oxygen channel was evident: hardly any interfering

peaks showed up.

keton

es

aldeh

ydes

alcoh

ols

phen

ols

aliph

atic a

cods

neop

hytad

iene

the ot

her c

ompo

nents

MSDAED

0

20

40

60

80

Figure 4. Comparing the two techniques (GC-MS and GC-AED),for percentages of the volatile components obtained for ketones,neophytadiene, aliphatic acids, alcohols, aldehydes, phenolsand the other components’ fortification level in the flue-curedtobaccos from Henan province in China.

Table 3. The contents of the volatile components of flue-cured tobaccos in Henan, Guizhou, Yunnan and Liaoning province in China

detection with GC-MS and GC-AED

No. Analyte Henan (mg/g) Guizhou (mg/g) Yunnan (mg/g) Liaoning (mg/g) Identified by MS

or standards

MS AED MS AED MS AED MS AED

1 Furfural 15.45 14.85 24.51 22.38 9.631 8.711 23.41 22.26 S

2 2-Furanmethanol 8.408 7.93 6.749 6.027 1.814 1.533 15.29 14.77 S

3 5-Methylfuranone 0.5352 0.348 0.3634 tr 0.112 tr 0.4382 tr MS

4 Protoanemonine 5.182 5.416 5.16 5.923 0.89 0.663 9.388 10.12 MS

5 2-Acetylfuran 0.6238 tr 0.6866 tr 0.3111 tr 1.77 tr MS

6 Phenylmethanal 8.395 7.321 7.762 6.819 1.281 1.107 9.965 9.462 S

7 6-Methyl-5-hepten-2-one 1.402 1.634 1.112 1.342 1.533 1.476 1.1807 0.938 MS

8 Benzenemethanol 11.21 13.03 38.23 39.18 27.57 28.43 35.32 35.93 S

9 Benzeneacetaldehyde 13.81 12.45 1.205 1.088 1.145 1.051 3.083 2.841 MS

10 2-Acetylpyrrole 10.15 8.96 9.364 8.957 5.69 5.137 7.048 7.11 MS

11 Phenylethyl alcohol 14.19 14.03 22.24 22.37 17.2 17.44 21.09 21.84 MS

12 Indole 2.472 tr 3.395 3.021 1.146 tr 1.764 tr MS

13 4-Vinylguaiacol 6.752 5.88 10.13 9.725 3.632 3.193 7.111 6.876 MS

14 Solanone 21.76 22.41 35.62 35.84 41.1 41.72 57.67 59.43 S

15 b-Damascenone 10.23 10.67 13.3 13.53 8.85 9.06 9.457 9.832 S

16 (E)-b-Bamascone 2.087 2.356 2.651 2.411 2.774 2.167 2.242 2.567 MS

17 (E)-Geranylacetone 16.87 14.91 4.939 5.283 5.894 6.112 7.492 7.734 S

18 b-Ionone 2.952 3.045 1.532 1.779 1.754 2.032 0.2094 tr S

19 Dihydroactinidiolide 17.08 17.33 12 13.07 9.757 9.771 10.72 11.13 S

20 Megastigmatrienone 1 8.897 8.626 9.053 8.686 6.226 5.325 5.005 4.792 MS




24 3-Oxo-a-ionol 5.65 5.97 7.613 7.775 tr tr 4.517 4.882 MS

25 Neophytadiene 1080 1101 1463 1501 988.5 1010 871.7 901.7 MS

26 Farnesyl acetone 23.97 24.51 tr tr tr tr 4.488 4.923 MS

27 Methyl palmitate 29.19 29.31 56.35 57.33 20.59 23.11 21.91 23.44 MS

28 Palmitic acid 88.35 84.67 108.5 107.64 47.44 45.78 140.5 138.8 MS

tr, trace amount.



The results of GC-AED and GC-MS analysis of the tobacco

are given in Table 3, which showed that there was a relatively

good agreement between the two sets of data. As it could also

be calculated from Table 3, the proportion of the ketones,

neophytadiene, aliphatic acids, alcohols, aldehydes, phenols

and the other components relative to the total amount of the

volatiles in tobacco was not very different for the tobacco from

different provinces in China [35, 36]. As illustrated in Fig. 4, the

percentages of ketones, neophytadiene, aliphatic acids, alcohols,

aldehydes, phenols and the other components were obtained in

Henan province. The results also showed good agreement

between GC-MS and GC-AED analysis. This indicated that the

AED coupled with Deans switch was a good alternative for

chromatography when analyzing tobacco samples.

4 Concluding remarks

The sensitivity and precision of GC-AED coupled with Deans

switching device for tobacco analysis were reported. Nano-

gram analyte detection limits for carbon, hydrogen and

oxygen were obtained. Precision worsens as measured by

regression coefficient of determination or ERF in the order

C4H4O. Detection limits decreased in the order O4H4C,

which approximated the order of decreasing first ionization

energies. Compared with GC-MS technique, the selectivity of

the GC-AED was evident. The results showed that GC-AED

coupled with Deans switching technique was a useful tool for

analyzing the volatiles from tobaccos.

The authors are grateful to Science and TechnologyCommission of Shanghai (Grant No. 11R21422000) forfinancial support.

The authors have declared no conflict of interest.

5 References

[1] Masahiro, C., Yukio, S., Hirotoshi, T., J. Agric. FoodChem. 2004, 52, 7918–7924.

[2] Yong, L., Tao, P., Ziming, G., Yanli, L., Xiaolin, W.,Jianhua, D., Kejun, Z., Xin, L., Guowang, X., Talanta2010, 81, 650–656.

[3] Yong, L., Tao, P., Yanli, L., Xiaolin, W., Qinghua, L., Xin,L., Guowang, X., J. Sep. Sci. 2011, 34, 1447–1454.

[4] Elizabeth, S., Alan, S., Najat, A. S., Food Chem. Toxicol.2008, 46, 1582–1590.

[5] Serban, C. M., Melissa, K., J. Chromatogr. A 2007, 1141,90–97.

[6] Francois, M., Martial, S., J. Chromatogr. A 2010, 1217,7528–7538.

[7] Xiujuan, L., Jia, H., Changjiang, H., Huili, W., Xuedong,W., J. Sep. Sci. 2009, 32, 4213–4219.

[8] Huang, L. F., Zhong, K. J., Sun, X. J., Wu, M. J., Huang,K. L., Liang, Y. Z., Guo, F. Q., Li, Y. W., Anal. Chim. Acta2006, 575, 236–245.

[9] Gan, H. L., Cheman, Y. B., Tan, C. P., NorAini, I., Nazi-mah, S. A. H., Food Chem. 2005, 89, 507–518.

[10] Teixeira, S., Mendes, A., Alves, A., Santos, L., Anal.Chim. Acta 2007, 584, 439–446.

[11] Ferhat, M. A., Tigrine-Kordjani, N., Chemat, S., Meklati,B. Y., Chemat, F., Chromatographia 2007, 65, 217–222.

[12] Sudhir, K. P., Ki-Hyun, K., Trends Anal. Chem. 2010, 29,804–819.

[13] Careen, M., Fritz, P., Astrid, R., Beat, A., Werner, B.,Frank, S., J. Chromatogr. A 2006, 1116, 10–19.

[14] Gu, M., Yang, F., Su, Z. G., J. Chromatogr. A 2004, 1022,139–144.

[15] Tong, L., Wang, Y. Z., Xiong, J. F., Cui, Y., Zhou, Y. G.,Yi, L. C., Talanta 2008, 76, 80–84.

[16] Frank, D., Karine, J., Pat, S., Andrew, B., Matthew, S. K.,Anal. Bioanal. Chem. 2010, 396, 1291–1300.

[17] Cunha, S. C., Fernandes, J. O., J. Sep. Sci. 2010, 33,600–609.

[18] Rolfes, J., Andersson, J. T., Anal. Chem. 2001, 73,3073–3082.

[19] Carin, V. M., Weeraya, K., Claudia, A. Z., Elina, B. C.,Philip, J. M., J. Sep. Sci. 2006, 29, 1909–1921.

[20] Elci, L., Kolbe, N., Elci, S. G., Anderson, J. T., Talanta2011, 85, 551–555.

[21] Guor-Jien, W., Chi-Tang, H., An, S. H., Food Chem.2009, 116, 774–778.

[22] Meija, J., Montes-Bayon, M., Le Duc, D. L., Terry, N.,Caruso, J. A., Anal. Chem. 2002, 74, 5837–5844.

[23] Soini, H. A., Bruce, K. E., Klouckova, I., Brereton, R. G.,Penn, D. J., Novotny, M. V., Anal. Chem. 2006, 78,7161–7168.

[24] Campillo, N., Penalver, R., Hernandez-Cordoba, M., J.Chromatogr. A 2008, 121, 222–228.

[25] Maikhunthod, B., Morrison, P. D., Small, D. M., Marriott,P. J., J. Chromatogr. A 2010, 1217, 1522–1529.

[26] Stephanie, J. M., Rebecka, C., Francis, M. U., Gwendo-lyn, A. M., J. Chromatogr. B 2007, 858, 59–64.

[27] Bartle, K. D., Hall, S. R., Holden, K., Mitchell, S. C., Ross,A. B., Fuel 2009, 88, 348–353.

[28] Lopez, F. J., Beltran, J., Forcada, M., Hernandez, F., J.Chromatogr. A 1998, 823, 25–33.

[29] Pizarro, C., Perez-del-Notario, N., Gonzalez-Saiz, J. M.,J. Sep. Sci. 2009, 32, 3746–3754.

[30] Juillet, Y., Gibert, E., Beos, A., Bellier, B., Anal. Bioanal.Chem. 2005, 383, 848–856.

[31] Krzysztof, L., Thomas, W., Gunnar, H., J. Chromatogr. A2003, 985, 275–282.

[32] Jelink, J. T., Venema, A., J. High Resolut. Chromatogr.1990, 13, 447–450.

[33] Schafer, W., J. High Resolut. Chromatogr. 1993, 16,674–676.

[34] Janak, K., Colmsjo, A., Ostman, C., J. Chromatogr. Sci.1995, 33, 611–621.

[35] Cai, J. B., Liu, B. Z., Ling, P., Su, Q. D., J. Chromatogr. A2002, 947, 267–275.

[36] Zhu, X. L., Gao, Y., Chen, Z. Y., Su, Q. D., Chromato-graphia 2009, 69, 735–742.

J. Sep. Sci. 2012, 35, 334–340340 G. Li et al.


Documents

Determination of the volatiles from tobacco by capillary gas chromatography with atomic emission detection and mass spectrometry