Application of GC in Food Analysis
Application of GC in Food Analysis
Application of GC in Food Analysis
9+10, 2002
Jana Hajšlová
Institute of Chemical Technology; Faculty of Food and Biochemical Technology; Department of Food
Chemistry and Analysis, Technická 3, 166 28 Prague 6, Czech Republic
0165-9936/02/$ - see front matter # 2002 Published by Elsevier Science B.V. All rights reserved.
PII: S0165-9936(02)00805-1
trends in analytical chemistry, vol. 21, nos. 9+10, 2002 687
food and environmental samples. GC has been We shall refer to the types of analyses that
instrumental in helping humans realize that we answer these questions as relating, respectively, to:
must use caution with agricultural and industrial
1. composition;
chemicals to avoid harming our health, the food 2. additives and contaminants; and,
supply, and the ecosystem that we rely upon to
3. transformation products.
sustain ourselves. The scientific discoveries
made with the help of GC in agricultural and These categories are not always clear or even
food sciences have contributed to more plenti- important, but they are helpful for the purpose
ful and healthier food, longer and better lives, of describing the types of applications in food
and an expanding population of 6 billion people. analysis that are the subject of this article.
Other recent articles have reviewed the analy-
tical chemistry of food analysis [1], and parti- 1.2. Composition
cular food applications involving GC, such as
carbohydrates and amino acids [2], lipids and Food is composed almost entirely of water,
accompanying lipophilic compounds [3,4], proteins, lipids, carbohydrates, and vitamins and
aroma and flavors [5–8], and pesticide residues minerals. Water is often a very large component
[9,10]. The purpose of this article is to mention of food, but it is generally removed by drying
the main applications of GC and discuss current before compositional analysis is conducted.
trends in food analysis. We hope to provide Mineral content (as measured by ash after
insight into how state-of-the-art techniques may burning) is generally a very small component of
impact analytical food applications in the future. food, thus a compositional triangle of the remain-
There is no space in this article discuss all ing major components (lipids, proteins, and car-
advances being made in GC of food applications, bohydrates) can be devised as shown in Fig. 1 [11].
and we have chosen to focus on fast-GC/MS, This food-composition triangle can be used to
which we believe is the developing technology describe and categorize foods based on their che-
that will have the most impact in the coming mical content, and the division of the triangle into
decade if it can be applied in routine food nine sections, as shown, can be very helpful to the
applications. chemist in deciding the appropriate analytical
techniques to use in making measurements [9].
1.1. Needs for food analysis Nutritional labeling laws in many countries
require all processed foods to be analyzed and
Most needs for food analysis arise from the reporting of their composition to the con-
nutrition and health concerns, but other reasons sumer. The food processor also has an interest
for food analysis include process-control or (and necessity!) to analyze carefully the compo-
quality-assurance purposes, flavor and palat- sition of its product, thus a great number of
ability issues, checking for food adulteration, food compositional analyses are conducted
identification of origin (pattern recognition), or every day. Although GC is rarely used in bulk
‘‘mining’’ the food for natural products that can compositional assays, it is the primary tool for
be used for a variety of purposes. All analytical analysis of fatty acids, sterols, alcohols, oils,
needs for food analysis originate from three aroma profiles, and off-flavors, and in other
questions: food-composition applications [12]. GC is also
the method of choice for analysis of any volatile
1. What is the natural composition of the food(s)? component in food.
2. What chemicals appear in food as an additive or
byproduct from intentional treatment, unintended 1.3. Additives and contaminants
exposure, or spoilage (and how much is there)?
3. What changes occur in the food from natural or Many agrochemicals are used to grow the
human-induced processes? quantity and quality of food needed to sustain
688 trends in analytical chemistry, vol. 21, nos. 9+10, 2002
Fig. 1. Food-composition triangle divided into nine categories and examples of different foods in each category. Redrawn from
[11] with permission from the author.
the human population. Many of the agrochem- to make it appear better to the consumer or to
icals are pesticides (e.g. herbicides, insecticides, alter its taste or texture.
fungicides, acaricides, fumigants) that may All these types of additives and contaminants
appear as residues in the food. Other types of are regulated by government agencies world-
agrochemicals that may appear as residues in wide. Without doubt, more than a million ana-
animal-derived foods are veterinary drugs (e.g. lyses of food contaminants and additives are
antibiotics, growth promotants, anthelmintics). conducted worldwide per year by industry,
Different types of environmental contaminants government, academic, and contract labora-
(e.g. polyhalogenated hydrocarbons, polycyclic tories. GC is the primary tool for the measure-
aromatic hydrocarbons, organometallics) can ment of many chemical contaminants and
appear in food through their unintended expo- additives.
sure to the food through the air, soil, or water.
Food may also be contaminated by toxins from 1.4. Transformation products
various micro-organisms, such as bacteria or
fungi (e.g. mycotoxins), or natural toxins already Transformation products are those chemicals
present in the food or that arise from spoilage. that may occur in food due to unintended
Packaging components (e.g. styrenes, phtha- chemical reactions (e.g. Maillard reactions, auto-
lates) can also leach into foods unintentionally. oxidation), industrial processes (e.g. drying,
In addition, chemical preservatives and syn- smoking, thermal processing, irradiation), and/
thetic antioxidants may be added after harvest or other processes (e.g. cooking and spoilage).
or during processing of the food to extend The types of chemicals that are categorized as
storage time or shelf-life of food products. transformational products (or endogenous con-
Other chemical additives (such as dyes, emulsi- taminants arising from transformational pro-
fiers, sweeteners, synthetic flavor compounds, cesses) are polycyclic aromatic hydrocarbons,
and taste enhancers) may be added to the food heterocyclic amines, urethane, nitrosamines,
trends in analytical chemistry, vol. 21, nos. 9+10, 2002 689
Fig. 2. Comparison of GC and HPLC in major food applications over three time periods (11 years each) of scientific literature
abstracted in PubMed [13]. In addition to year of publication, all searches were limited by ‘‘GC OR gas chromatography’’ or
‘‘HPLC or high performance liquid chromatography’’ AND ‘‘food.’’ Specific terms were used in the searches of each category as
follows: 1) pesticides=‘‘pesticide OR herbicide OR insecticide OR fumigant OR fungicide’’; 2) environmental con-
taminants=‘‘dioxin OR PAH OR PCB OR organometallic’’; 3) drugs=‘‘pharmaceutical OR drug OR antibiotic OR hormone’’;
4) toxins=‘‘toxin OR mycotoxin OR alkaloid’’; 5) additives=‘‘additive OR preservative OR sweetener OR emulsifier’’; 6,7,8) terms
as listed were used for nitrosamines, packaging, and irradiation; 9) amino acids=‘‘amino acid OR protein’’; 10) lipids=‘‘fat OR
lipid OR oil OR fatty acid OR sterol OR cholesterol’’; 11) carbohydrates=‘‘carbohydrate OR sugar OR fiber OR fibre’’;
12) vitamins=‘‘vitamin OR nutrient OR mineral’’; and, 13) aroma/flavor=‘‘aroma OR flavor’’.
GC applications, such as separations of lipids, Does this mean that only those techniques
HPLC has begun to rival GC in terms of that meet the analytical quality objectives (lower
publications. detection limits with greater selectivity) will sur-
vive (at least until an even better approach
2.1. Analytical trends comes along)? Can a faster, cheaper, easier
method with a smaller instrument that gives
The future of analytical food applications is lower quality results or lacks automation
impossible to predict with certainty, but it is become widespread in useful applications?
helpful in trying to predict the future by study- A test case to answer these questions is solid-
ing the past. The major goals in routine appli- phase microextraction (SPME) [14–17]. In
cations of analytical chemistry have always been combination with GC, SPME is able to extract
the same: to achieve better accuracy, lower and to detect volatiles in food in an easy, and
detection limits, and higher selectivity with relatively fast and cheap approach. In the
faster, easier, and cheaper methods using more decade since its introduction, SPME has been
robust, highly versatile, and smaller instruments. the subject of nearly 1,000 publications, but
The goals of lower detection limits and greater because of complications in quantitation, strong
selectivity with smaller instruments have devel- dependence on matrix, and certain practical
oped into actual trends, and, overall, many matters, some quality in the results is sacrificed
techniques today provide greater sample for speed and ease. The strengths of SPME
throughput with more ease (as a result of make it helpful in monitoring transformational
automation), but they are rarely cheaper! changes or obtaining qualitative information,
trends in analytical chemistry, vol. 21, nos. 9+10, 2002 691
but as Fig. 2 indicates, such transformational method. Thus, greater selectivity (in sample
monitoring is a niche market. It will be inter- preparations, analytical separations, and detec-
esting to see the status of SPME in 10 years. tion techniques) is always another welcome
feature that helps to provide better results at
2.2. Predictions from the 1980s lower detection limits. The continuing ability to
achieve lower detection limits with selective
In 1982, Tanner [18] attempted to extrapolate GC/MS(-MS) analysis, for example, has been a
the trends in food analysis for the 1980s. The major advancement [19]. In industrialized
major trend in GC at that time was that capillary nations, in addition to providing confirmatory
columns were replacing packed columns, and it results, GC/MS has become a primary GC tool
was an easy prediction to make that this trend for some food-analysis laboratories because of
would continue. In retrospect, another easy its ability to quantify many analytes at
prediction was that the use of computers for sufficiently low concentrations.
instrument control and data processing would
lead to time-saving and automated operation 2.3. View from 1990s
that would greatly increase sample throughput.
The computer revolution has been essential in If one was to predict the future in 1990, it
all aspects of science, and nearly all modern may have been easy to make erroneous assess-
analytical instruments and many chemists could ments of the impact of state-of-the-art tech-
not function without computers. niques at the time. For example, the atomic
However, Tanner also believed that, in food emission detector (AED) was introduced [20]
applications, the trend of lowering detection with a great deal of marketing and genuine sci-
limits would not be as important in the 1980s. entific interest in 1990. The advantages related
The more important factor was the accuracy of to the highly selective detection of several ele-
the determinations at the trace levels already ments and simultaneously made the instrument
being found. This is sometimes true in food- potentially very powerful in many GC applica-
composition applications, and one could make tions [21,22]. The reality was that the detection
the same argument today that food applications limits for important elements were not low
do not require lower limits of quantitation enough in comparison to other element-selec-
(LOQ). tive detectors, and matrix interferences in other
During the last 20 years, the trend to lower elemental channels limited the usefulness of
LOQ has continued, and, even though lower these channels. The AED could provide key
detection limits may not be needed in some information to help in the identification of ana-
applications per se, lower LOQ enable the lytes [23], but MS by itself can provide struc-
injection of more dilute samples, which is tural elucidation and analyte identification. The
always a welcome feature, especially in GC (to cost of AED was much higher than the worth
reduce coinjection of non-volatiles). Instru- of the questionable benefits it could provide in
ments that give lower detection limits can also most food applications. In 2001, the only com-
reduce the need for clean-up and solvent-eva- mercial manufacturer of the AED announced
poration steps. Indeed, the last 20 years have the termination of the product.
brought the analytical community away from The 1990s saw the rise and decline of other
multi-step, labor-intensive, large-volume, wet- ‘‘advantageous’’ techniques with severe limi-
chemical methods and into simpler, miniatur- tations in most food-analysis applications. A
ized approaches, in part because of the lower partial list includes supercritical fluid extraction,
LOQ possible with modern instruments. supercritical fluid chromatography, microwave
However, lower instrumental detection limits assisted extraction, capillary electrophoresis,
have no impact when matrix interferences are automated trace enrichment and dialysis,
the limiting factor in detection limits for the enzyme-linked immunosorbent assays, molecular
692 trends in analytical chemistry, vol. 21, nos. 9+10, 2002
imprinted polymers, and matrix solid-phase In the case of MS, its combination of qualita-
dispersion. Of course, some of these techniques tive and quantitative features gives it the
are continuing in certain analytical and/or non- advantages needed to become the biggest ‘‘king
analytical applications, but they are not used of the hill’’, and some day, selective GC detec-
widely in food applications for which they were tors will possibly be relegated to niche applica-
marketed. tions. The detectors with greater selectivity and/or
sensitivity that complement MS, such as PFPD
2.4. Current and future trends and XSD are likely to remain, and there is
always a need for lower cost and reliable detec-
Any new approach has to compete in an tors that meet the needs of simpler analyses
uphill struggle with the ‘‘kings of the hill’’ in [24]. But the future of GC (and LC) detection
analytical chemistry. GC, HPLC, traditional and applications is tied with MS. The key ques-
selective detectors, MS, solid-phase extraction tion for MS will continue to be: how much extra
(SPE), and liquid-liquid extraction (LLE) are the capital expense will the laboratory pay to gain
current leading approaches in analytical food the benefits of MS?
and agricultural applications. These techniques
have usurped previous major analytical tools,
such as thin-layer chromatography, Soxhlet 3. Faster GC/MS
extractions, tedious wet chemical methods, and
non-selective GC detectors. The features and Increasing the speed of analysis has always
performance of the current leading technologies been an important goal for GC separations. The
are established parameters, and any new tech- time of GC separations can be decreased in a
nique will have to match or better them for a number of ways: 1) shorten the column; 2)
reasonable price. Are there any new technolo- increase carrier-gas flow; 3) reduce column-film
gies that can join, or even usurp, any of these thickness; 4) reduce carrier-gas viscosity; 5)
‘‘kings of the hill?’’ increase column diameter; and/or, 6) heat the
Advantageous approaches that were intro- column more quickly. The trade-off for
duced for bench-top operations in the last 15 increased speed however is reduced sample
years with strong applicability to food analysis capacity, higher detection limits, and/or worse
include the major advances in HPLC/MS (and separation efficiency. How much of these fac-
MSn) and GC/MSn>, and other instrumental tors is the analyst willing to sacrifice for speed?
devices, such as programmable temperature Not much, apparently, because separation times
vaporization (PTV), pulsed flame photometric in typical routine applications have been much
detection (PFPD), halogen specific detection the same for decades (20–50 min). Perhaps as
(XSD), and pressurized liquid extraction (PLE), more laboratories begin to use instruments with
which is also known as accelerated solvent higher inlet-pressure limits, faster oven-tem-
extraction (ASE). Each of these techniques has perature program rates, electronic pressure
been on the market for at least six years, and control, and faster electronics for detection,
they provide benefits in breadth of scope, fast-GC with micro-bore columns will become
selectivity and/or detectability that are likely to more widely used, but the inherent trade-off will
make them useful for years to come. remain.
Other potentially useful fairly new commercial In practice, the GC conditions should be
devices for GC analysis of foods include large- designed to give the shortest analysis time while
volume injection (LVI), direct sample introduc- still providing the necessary selectivity (i.e.
tion (DSI) (commercially known as the separation of both analyte-analyte and matrix-
ChromatoProbe), and resistively heated capil- analyte). The use of element-selective detectors
laries. These techniques are not yet established may improve matrix-analyte selectivity, but, in
and it is not clear what their fate will be. that case, analyte-analyte selectivity must be
trends in analytical chemistry, vol. 21, nos. 9+10, 2002 693
the different approaches are focused this appli- columns quickly, and, since sample capacity is
cation. The reader is directed to the literature for reduced by a cubed factor in relation to column
descriptions of other food applications [37–39]. diameter [41], increased LOQ and decreased
ruggedness result, so such narrow columns can
3.1. GC/TOF-MS rarely be used in real-life applications.
TOF-MS can also give wide spectral mass
An advantage of the micro-bore GC/TOF- range and/or exceptional mass resolution (typi-
MS method versus the other approaches is that cally at the expense of speed, however). More-
separation efficiency need not be compromised over, GC/TOF-MS techniques do not
for speed of analysis. Modern quadrupole necessarily need to use short, micro-bore
instruments are capable of sufficiently fast scan columns to achieve short analysis times. Short,
rates for fast-GCMS [40], but quadrupole wider columns, ballistic or resistive heating of
instruments cannot match the potential of TOF columns, comprehensive 2-dimensional GC,
for this purpose. Rapid deconvolution of spec- and/or low pressure may become more suitable
tra (‘‘scanrate’’) with TOF-MS makes it the only approaches to meet food-application needs in
MS approach to achieve several data points GC/TOF-MS in the future.
across a narrow peak in full scan operation.
Fig. 3 gives an example of rapid GC/TOF-MS 3.2. LP-GC/MS
for the analysis of pesticides in a solution.
However, the injection of complex extracts LP-GC/MS, commercially known as Rapid-
deteriorates the performance of micro-bore MS. is an interesting approach to speed the
Fig. 4. Chromatogram of pesticides in toluene solution in conventional GC-MS and LP-GC/MS (5 ng injected).
1) methamidophos, 2) dichlorvos, 3) acephate, 4) dimethoate, 5) lindane, 6) carbaryl, 7) heptachlor, 8) pirimiphos-methyl,
9) methiocarb, 10) chlorpyrifos, 11) captan, 12) thiabendazole, 13) procymidone, 14) endosulfan I, 15) endosulfan II,
16) endosulfan sulfate, 17) propargite, 18) phosalone, 19) cis-permethrin, 20) trans-permethrin, 21) deltamethrin. Used from
[32] with permission of the publisher.
trends in analytical chemistry, vol. 21, nos. 9+10, 2002 695
analysis by which a relatively short (10 m) mega- and, 7) thermal degradation of thermally-labile
bore (0.53 mm i.d.) column is used as the analytes is reduced.
analytical column. The vacuum from the MS Fig. 4 shows how a three-fold gain in speed
extends into the column, which leads to higher was made in the analysis of 21 representative
flow rate and unique separation properties. A pesticides using LP-GC/MS versus traditional
restriction capillary (0.1–0.25 mm i.d. of appro- GC/MS. Larger injection volume could be
priate length) is placed at the inlet end to pro- made in LP-GC/MS because of better focusing
vide positive inlet pressure and to allow normal of the gaseous solvent at the higher head pres-
GC injection methods. Advantages of LP-GC/ sure and larger column capacity, so overall gains
MS include: 1) fast separations are achieved; 2) in sensitivity were achieved. However, reduced
no alterations to current instruments are nee- separation efficiency occurs with LP-GC/MS
ded; 3) sample capacities and injection volumes and ruggedness of the approach with repeated
are increased with mega-bore columns; 4) peak injections was no better than traditional methods
widths are similar to conventional separations to with a narrow-bore analytical column.
permit normal detection methods; 5) peak
heights are increased and LOQ can be lower 3.3. GC/SMB-MS
(depending on matrix interferences); 6) peak
shapes of relatively polar analytes are improved; GC/MS with current commercial instruments
have a practical 2 mL/min flow limitation
because of MS-instrument designs. GC/SMB-
MS is a very promising technique and instru-
ment that overcomes this flow rate limitation
because SMB-MS requires a high gas-flow rate
at the SMB interface. However, only a single
prototype GC/SMB-MS instrument exists at
this time, and the approach is not commercially
available.
The advantages of GC/SMB-MS include: 1)
the selectivity of the MS detection in electron-
impact ionization is increased because an
enhanced molecular ion occurs for most mole-
cules at the low temperatures of SMB, so losses
of selectivity in the GC separation can be made
up by increased selectivity in the MS detection;
2) the use of very high gas-flow rates enables
GC analysis of both thermally labile and non-
volatile chemicals, thereby extending the scope
of the GC/SMB-MS approach to many analytes
currently done by HPLC; 3) the SMB-MS
approach is compatible with any column
dimension and injection technique; 4) reduced
column bleed and matrix interference occurs
Fig. 5. Fast-GC/SMB-MS analysis of the indicated 13 pes- because of the lower temperatures and
ticides in methanol (3–7 ng injected). Trace B is a zoom of enhanced molecular ions; and, 5) better peak
the upper trace A in order to demonstrate the symmetric shapes occur because tailing effects in MS are
tailing-free peak shapes. A 6 m capillary column with 0.2
mm i.d., 0.33 mm DB-5ms film was used with 10 mL/min
eliminated. Fig. 5 gives an example in the
He flow rate. Used from [34] with permission of the separation of diverse pesticides using GC/SMB-
publisher. MS.
696 trends in analytical chemistry, vol. 21, nos. 9+10, 2002
[33] P.A. Leclercq, C.A. Cramers, Mass Spectrom. Rev. 17 [37] I. Arnault, N. Mondy, F. Cadoux, J. Auger, J. Chromatogr.
(1998) 37. A 896 (2000) 117.
[34] M. Kochman, A. Gordin, P. Goldshlag, S.J. Lehotay, A. [38] L. Mondello, G. Zappia, G. Errante, P. Dugo, LC.GC
Amirav, J. Chromatogr. A (in press). Europe 13 (2000) 496.
[35] A. Amirav, S. Dagan, T. Shahar, N. Tzanani, S.B. Wainhaus, [39] C.C. Chang, G.R. Her, J. Chromatogr. A 893 (2000) 169.
in: E.J. Karjalainen (Editor), Advances in Mass Spectro- [40] A. Amirav, N. Tzanani, S. Wainhaus, S. Dagan, Eur. Mass.
metry, Vol. 14, Elsevier, Amsterdam, 1998, pp. 529–562. Spectrom. 4 (1998) 7.
[36] A. Dagan, A. Amirav, Int. J. Mass Spectrom. Ion Process. [41] R.T. Ghijsen, H. Poppe, J.C. Kraak, P.P.E. Duysters,
133 (1994) 187. Chromatographia 27 (1989) 60.