Nielsen2010 FoodAnalysis 4thEdanuragaja.staff.ipb.ac.id/files/2017/05/3_GC.pdf · 2017-05-04 · umn chromatography on silica gel), simple solvent extraction, or some combination

29chapter

Gas Chromatography

Michael C. Qian∗Department of Food Science and Technology, Oregon State University,

Corvallis, OR 97331-6602, [email protected]

and

Devin G. Peterson and Gary A. ReinecciusDepartment of Food Science and Nutrition, University of Minnesota,

St. Paul, MN 55108-6099, [email protected]

[email protected]

29.1 Introduction 51529.2 Sample Preparation for Gas

Chromatography 51529.2.1 Introduction 515

29.2.2 Isolation of Solutes from Foods 51529.2.2.1 Introduction 51529.2.2.2 Headspace Methods 51629.2.2.3 Distillation Methods 516

S.S. Nielsen, Food Analysis, Food Science Texts Series, DOI 10.1007/978-1-4419-1478-1_29,c© Springer Science+Business Media, LLC 2010

513

514 Part V • Chromatography

29.2.2.4 Solvent Extraction 51729.2.2.5 Solid-Phase Extraction 51729.2.2.6 Direct Injection 519

29.2.3 Sample Derivatization 51929.3 Gas Chromatographic Hardware and

Columns 52029.3.1 Gas Supply System 52029.3.2 Injection Port 520

29.3.2.1 Hardware 52029.3.2.2 Sample Injection

Techniques 52129.3.2.2.1 Split Injection 52129.3.2.2.2 Splitless

Injection 52129.3.2.2.3 Temperature

ProgrammedInjection 521

29.3.2.2.4 On-ColumnInjections 522

29.3.2.2.5 Thermal DesorptionInjection 522

29.3.3 Oven 52229.3.4 Column and Stationary Phases 522

29.3.4.1 Packed Columns 52229.3.4.2 Capillary Columns 52329.3.4.3 Gas–Solid (PLOT)

Chromatography 52429.3.5 Detectors 524

29.3.5.1 Thermal ConductivityDetector 52429.3.5.1.1 Operating

Principles 52429.3.5.1.2 Applications 525

29.3.5.2 Flame Ionization Detector 52529.3.5.2.1 Operating


29.3.5.3 Electron Capture Detector 52629.3.5.3.1 Operating


29.3.5.4 Flame Photometric Detectorand Pulsed Flame PhotometricDetector 526

29.3.5.4.1 OperatingPrinciples 526

29.3.5.4.2 Applications 52729.3.5.5 Photoionization Detector 527

29.3.5.5.1 OperatingPrinciples 527

29.3.5.5.2 Applications 52829.3.5.6 Electrolytic Conductivity

Detector 52829.3.5.6.1 Operating


29.3.5.7 Thermionic Detector 52829.3.5.7.1 Operating


29.3.5.8 Hyphenated GasChromatographicTechniques 528

29.3.5.9 Multidimensional GasChromatography 52829.3.5.9.1 Conventional

Two-DimensionalGC 529

29.3.5.9.2 ComprehensiveTwo-DimensionalGC 529

29.4 Chromatographic Theory 53029.4.1 Introduction 53029.4.2 Separation Efficiency 530

29.4.2.1 Carrier Gas Flow Rates andColumn Parameters 530

29.4.2.2 Carrier Gas Type 53129.4.2.3 Summary of Separation

Efficiency 53229.5 Applications of GC 532

29.5.1 Residual Volatiles in PackagingMaterials 533

29.5.2 Separation of Stereoisomers 53329.5.3 Headspace Analysis of Ethylene Oxide in

Spices 53329.5.4 Aroma Analysis of Heated Butter 534

29.6 Summary 53529.7 Study Questions 53529.8 References 536

Chapter 29 • Gas Chromatography 515

29.1 INTRODUCTION

The first publication on gas chromatography (GC) wasin 1952 (1), while the first commercial instrumentswere manufactured in 1956. James and Martin (1) sep-arated fatty acids by GC, collected the column effluent,and titrated the individual fatty acids for quantita-tion. GC has advanced greatly since that early workand is now considered to be a mature field that isapproaching theoretical limitations.

The types of analysis that can be done by GCare very broad. GC has been used for the determi-nation of fatty acids, triglycerides, cholesterol andother sterols, gases, solvent analysis, water, alcohols,and simple sugars, as well as oligosaccharides, aminoacids and peptides, vitamins, pesticides, herbicides,food additives, antioxidants, nitrosamines, polychlo-rinated biphenyls (PCBs), drugs, flavor compounds,and many more. The fact that GC has been used forthese various applications does not necessarily meanthat it is the best method – often better choices exist.GC is ideally suited to the analysis of thermally stablevolatile substances. Substances that do not meet theserequirements (e.g., sugars, oligosaccharides, aminoacids, peptides, and vitamins) are more suited to anal-ysis by a technique such as high-performance liquidchromatography (HPLC) or supercritical fluid chro-matography (SFC). Yet gas chromatographic methodsappear in the literature for these substances.

This chapter will discuss sample preparation forGC, gas chromatographic hardware, columns, andchromatographic theory as it uniquely applies to GC.Texts devoted to GC in general (2–4) and food applica-tions in particular (5, 6) should be consulted for moredetail.

29.2 SAMPLE PREPARATION FOR GASCHROMATOGRAPHY

29.2.1 Introduction

One cannot generally directly inject a food productinto a GC without some sample preparation. Thehigh temperatures of the injection port will result inthe degradation of nonvolatile constituents and cre-ate a number of false GC peaks corresponding tothe volatile degradation products formed. In addi-tion, very often the constituent of interest must beisolated from the food matrix simply to permit con-centration such that it is at detectable limits for the GCor to isolate it from the bulk of the food. Thus, onemust generally do some type of sample preparation,component isolation, and concentration prior to GCanalysis.

Sample preparation often involves grinding,homogenization, or otherwise reducing particle size.There is substantial documentation in the literatureshowing that foods may undergo changes duringsample storage and preparation. Many foods containactive enzyme systems that will alter the compositionof the food product. This is very evident in the area offlavor work (7–9). Inactivation of enzyme systems viahigh-temperature-short-time thermal processing, sam-ple storage under frozen conditions, drying the sam-ple, or homogenization with alcohol may be necessary(see Chap. 5).

Microbial growth or chemical reactions also mayoccur in the food during sample preparation. Chem-ical reactions often will result in the formation ofvolatiles that will again give false peaks on the GC.Thus, the sample must be maintained under condi-tions such that degradation does not occur. Microor-ganisms often are inhibited by chemical means (e.g.,sodium fluoride), thermal processing, drying, orfrozen storage.

29.2.2 Isolation of Solutes from Foods

29.2.2.1 Introduction

The isolation procedure may be quite complicateddepending upon the constituent to be analyzed. Forexample, if one were to analyze the triglyceride boundfatty acids in a food, one would first have to extractthe lipids (free fatty acids; mono-, di-, and triglyc-erides; sterols; fat-soluble vitamins, etc.) from the food(e.g., by solvent extraction) and then isolate only thetriglyceride fraction (e.g., by adsorption chromatogra-phy on silica). The isolated triglycerides then wouldhave to be treated to first hydrolyze the fatty acidsfrom the triglycerides and subsequently to form estersto improve gas chromatographic properties. The twolatter steps might be accomplished in one reaction bytransesterification (e.g., borontrifluoride in methanol)as described in Chap. 8, Sect. 8.3.1.6, and Chap. 14,Sect. 14.6.2. Thus many steps involving several typesof chromatography may be used in sample prepara-tion for GC analysis.

The analysis of volatiles in foods (e.g., packagingor environmental contaminants, alcohols, and flavorsor off-flavors) may be a simpler task. These mate-rials for GC analysis may be isolated by headspaceanalysis (static or dynamic), distillation, prepara-tive chromatography (e.g., solid-phase extraction, col-umn chromatography on silica gel), simple solventextraction, or some combination of these basic meth-ods. The procedure used will depend on the foodmatrix as well as the compounds to be analyzed.The primary considerations are to isolate the com-pounds of interest from nonvolatile food constituents


(e.g., carbohydrates, proteins, vitamins) or those thatwould interfere with GC (e.g., lipids). Some of thechromatographic methods that might be applied tothis task have been discussed in the basic chromatog-raphy chapter (Chap. 27). Methods for the isolationof volatile substances will be covered briefly as theypertain to the isolation of components for gas chro-matographic analysis.

It should be emphasized that the isolation pro-cedure used is critical in determining the resultsobtained. An improper choice of method or poor tech-nique at this step negates the best gas chromatographicanalysis of the isolated solutes. The influence of iso-lation technique on gas chromatographic analysis ofaroma compounds has been demonstrated (10). Thesebiases are discussed in the sections that follow andin more detail in books edited by Marsili (11) andMussinan and Morello (12). While these books relateto the analysis of aroma compounds in foods, the tech-niques for the isolation of these volatiles are the sameas used in the analysis of other volatiles in foods.

29.2.2.2 Headspace Methods

One of the simplest methods of isolating volatilecompounds from foods is by direct injection of theheadspace vapors above a food product. There aretwo types of headspace sampling: direct (or static)headspace sampling and dynamic headspace sam-pling.

Direct headspace sampling has been used exten-sively when rapid analysis is necessary and majorcomponent analysis is satisfactory. At equilibrium,the headspace of the sample is taken using a gas-tight syringe and then injected directly into theGC. Examples of method applications include mea-surement of hexanal as an indicator of oxidation(13, 14) and 2-methylpropanal, 2-methylbutanal,and 3-methylbutanal as indicators of nonenzymaticbrowning (15). The determination of residual solventsin packaging materials also may be approached by thismethod. Unfortunately, this method does not providethe sensitivity needed for trace analysis. Instrumen-tal constraints typically limit headspace injectionvolumes to 5 ml or less. Therefore, only volatilespresent in the headspace at concentrations greaterthan 10−7 g/l headspace would be at detectable levels[using a flame ionization detector (FID)].

Dynamic headspace sampling or purge and traphas found wide usage in recent years. This con-centration method may involve simply passing largevolumes of headspace vapors through a cryogenictrap or, alternatively, a more complicated extractionand/or adsorption trap. A simple cryogenic trapoffers some advantages and disadvantages. A cryo

trap (if properly designed and operated) will collectheadspace vapors irrespective of compound polarityand boiling point. However, water is typically themost abundant volatile in a food product, and, there-fore, one collects an aqueous distillate of the prod-uct aroma. This distillate must be extracted with anorganic solvent, dried, and then concentrated for anal-ysis. These additional steps add analysis time and pro-vide opportunity for sample contamination. A morecommonly used technique is adsorbent traps.

Adsorbent traps offer the advantages of provid-ing a water-free volatile isolate (trap material typicallyhas little affinity for water) and are readily auto-mated. The adsorbent initially used for headspacetrapping was charcoal. The charcoal was either solventextracted (CS) or thermally desorbed with backflush-ing (inert gas) to recover the adsorbed volatiles. Theuse of synthetic porous polymers as headspace trapmaterial now dominates. Initially, Tenax (a porouspolymer very similar to the skeleton of ion-exchangeresins) was most commonly used; however, combi-nations of Tenax and other polymers are now seeinggreater application. These polymers exhibit good ther-mal stability and reasonable capacity. Adsorbent trapsare generally placed in a closed system and loaded,desorbed, and so on via the use of automated multi-port valving systems. The automated closed systemapproach provides reproducible GC retention timesand quantitative precision necessary for some stud-ies. The primary disadvantage of adsorbent traps istheir differential adsorption affinity and limited capac-ity. Buckholz et al. (16) have shown that the mostvolatile peanut aroma constituents will break throughtwo Tenax traps in series after purging at 40 ml/minfor only 15 min. Therefore, the GC profile may onlypoorly represent the actual food composition due tobiases introduced by the purging and trapping steps.

29.2.2.3 Distillation Methods

Distillation processes are quite effective at isolatingvolatile compounds from foods for GC analysis. Prod-uct moisture or outside steam is used to heat andcodistill the volatiles from a food product. This meansthat a very dilute aqueous solution of volatiles results,and a solvent extraction must be performed on the dis-tillate to permit concentration for analysis. The distilla-tion method most commonly used today is some mod-ification of the original Nickerson–Likens distillationhead. In this apparatus, a sample is boiled in one sideflask and an extracting solvent in another. The productsteam and solvent vapors are intermixed and con-densed; the solvent extracts the organic volatiles fromthe condensed steam. The solvent and extracted dis-tillate return to their respective flasks and are distilled


to again extract the volatiles from the food. While thismethod is convenient and efficient, artifacts from sol-vents used in extraction, antifoam agents, steam sup-ply (contaminated water), thermally induced chemicalchanges, and leakage of contaminated laboratory airinto the system may contaminate the volatile isolate.

29.2.2.4 Solvent Extraction

Solvent extraction is often the preferred method for therecovery of volatiles from foods. Recovery of volatileswill depend upon solvent choice and the solubility ofthe solutes being extracted. Solvent extraction typi-cally involves the use of an organic solvent (unlesssugars, amino acids, or some other water-soluble com-ponents are of interest). Extraction with organic sol-vents limits the method to the isolation of volatilesfrom fat-free foods (e.g., wines, some breads, fruit andberries, some vegetables, and alcoholic beverages), oran additional procedure must be employed to sepa-rate the extracted fat from the isolated volatiles (e.g., achromatographic method). Fat will otherwise interferewith subsequent concentration and GC analysis.

Solvent extractions may be carried out in quiteelaborate equipment, such as supercritical CO2 extrac-tors, or can be as simple as a batch process in aseparatory funnel. Batch extractions can be quite effi-cient if multiple extractions and extensive shaking areused (17). The continuous extractors (liquid–liquid)are more efficient but require more costly and elabo-rate equipment.

29.2.2.5 Solid-Phase Extraction

The extractions discussed above involved the use oftwo immiscible phases (water and an organic solvent).However, a newer and a very rapidly growing alter-native to such extractions is solid-phase extraction(18, 19). In one version of this technique, a liquidsample (most often aqueous based) is passed througha column (2–10 ml vol) filled with chromatographicpacking or a TeflonR filter disk (25–90 mm in diameter)that has the chromatographic packing embedded in it.The chromatographic packing may be any of a num-ber of different materials (e.g., ion-exchange resins or ahost of different reversed- or normal-phase HPLC col-umn packings). When a sample is passed through thecartridge or filter, solutes that have an affinity for thechromatographic phase will be retained on the phasewhile those with little or no affinity will pass through.The phase is next rinsed with water, perhaps a weaksolvent (e.g., pentane), and then a stronger solvent(e.g., dichloromethane). The strong eluent is chosensuch that it will remove the solutes of interest.

29-1f igure

Schematic of a solid-phase microextraction(SPME) device (21). (Courtesy of Dr. JanuszPawliszyn, Department of Chemistry, Univer-sity of Waterloo, Waterloo, Ontario, Canada.)

Overall, solid-phase extraction has numerousadvantages over traditional liquid–liquid extractionsincluding: (1) less solvent is required; (2) speed;(3) less glassware is needed (less cost and potentialfor contamination); (4) better precision and accuracy;(5) minimal solvent evaporation for further analysis(e.g., GC); and (6) it is readily automated. Solid-phaseextraction has limitations, but new variations of thetechnique seek to overcome some of these.

The most recent version of solid-phase extractionis called solid-phase microextraction (SPME). Thismethod was developed originally for environmentalwork (20, 21). In this adaptation, the phase is boundonto a fine fused silica filament (ca. the size of a 10-µlsyringe needle, Fig. 29-1). The filament is immersed ina sample or in the headspace above a sample. Afterthe desired extraction time, the filament is pulled intoa protective metal sheath, removed from the sam-ple, and forced through the septum of a gas chro-matograph where the adsorbed volatiles are thermallydesorbed from the filament (Fig. 29-2).

SPME is an equilibrium technique and, there-fore, the volatile profile (i.e., volatile recovery) thatone obtains is strongly dependent upon sample com-position and careful control of all sampling param-eters. This includes the specific phase coating andthickness on the filament/fiber used. Several differ-ent phases of fiber are commercially available, andcompounds with a wide range of polarity or volatil-ity can be analyzed. PDMS (polydimethylsiloxane) isa nonpolar phase coating and can be used to extract


29-2f igure

Schematic showing the steps involved in the use of a solid-phase microextraction (SPME) device. (Reprinted withpermission of Supelco, Bellefonte, PA 16823, USA.)

nonpolar compounds. Polar analytes can be extractedwith polar phases (e.g., polyacrylate and Carbowaxcoatings). Porous fibers such as Carboxen or divinyl-benzene (DVB) coating are good for highly volatilecompounds. The coating has various film thicknesses.Thicker film fibers (100 µm) are better for volatiles,whereas thinner film fibers (7 µm and 30 µm) are betterfor larger molecules. Multiphase fibers (such as Car-boxen/PDMS, Carboxen/DVB/PDMS) are also avail-able to extract both polar and nonpolar compounds.

Due to its simplicity, SPME is very popular forvolatile aroma analysis of food and beverages. WhileHarmon (22) notes that the method can give excel-lent results, Coleman (23) cautions that the fibershave a definite linear range and competition betweenvolatiles for binding sites can introduce errors. Otherconcerns are for sensitivity limitations, precision, andlife of the filament. If the filament must be replaced(breakage), there is the issue of reproducibility of thenew vs. the old filament.

Solid-phase dynamic extraction (SPDE) isanother technique for volatile extraction. SPDE is sim-ilar to SPME, except the polymer is coated inside aspecial needle. A gas-tight syringe is used for SPDEto draw the headspace of food, and the volatiles areabsorbed by the phase. The process can be repeatedmany times by moving the plunger up and down toachieve maximum absorption. The needle can then beinjected into the GC for analysis. Different phases areavailable and the volume of the phase is about 4.5 µlcompared with only 0.6 µl for SPME, so the SPDE hasless issue with analyte saturation and competition.

Stir bar sorptive extraction (SBSE) is a new tech-nique for volatile extraction. In SBSE (Fig. 29-3), a

29-3f igure

Diagram of stir bar sorptive extraction(SBSE) device. (Courtesy of Gerstel, Inc.,Linthicum, MD.)

magnet stir bar is jacketed with glass, and the glass iscoated with a layer of absorbent (PDMS). The bar spinsin the sample solution and absorbs the analytes fromthe sample solution. The stir bar can also just hangon the headspace for volatile extraction the same wayas the SPME. After the extraction, the volatiles are thenthermally desorbed and introduced into a GC. Stir barhas almost 50 times more volume of absorbent thanSPME. For SPME, the PDMS volume is about 0.5 µl;with SBSE, it is 24–126 µl. Due to the increased volumeof absorbent phase, SBSE has much higher sensitivitythan SPME and has minimum competition and satu-ration effects (24, 25). The high sensitivity (ppt to ppg)


and flexibility of SBSE for nonpolar and medium polarcompounds makes it an effective and time savingmethod for extracting trace volatile compounds fromcomplex matrices (25). The PDMS phase is robust;does not absorb water, alcohol, or pigment; and isvery good for flavor extraction in alcoholic beverages.However, the PDMS phase is not selective for shorter-chain acids and polar compounds. Other phases ofSBSE need to be used to analyze polar compounds.

29.2.2.6 Direct Injection

It is theoretically possible to analyze some foods bydirect injection of the food into a gas chromatograph.Assuming one can inject a 2- to 3-µl sample intoa GC and the GC has a detection limit of 0.1 ng(0.1 ng/2 µl), one could detect volatiles in the sam-ple at concentrations greater than 50 ppb. Problemswith direct injection arise due to thermal degradationof any nonvolatile food constituents, damage to theGC column, decreased separation efficiency due towater in the food sample, contamination of the col-umn and injection port by nonvolatile materials, andreduced column efficiency due to slow vaporizationof volatiles from the food (injection port temperaturesare reduced to minimize thermal degradation of thenonvolatile food constituents). Despite these concerns,direct injection is commonly used to determine oxida-tion in vegetable oils (26,27). A relatively large volume

of oil (50–100 µl) can be directly injected into an injec-tion port of a GC that has been packed with glasswool. Since vegetable oils are reasonably thermallystable and free of water, this method is particularlywell suited to oil analysis.

There are numerous other approaches for the iso-lation of volatiles from foods. Some are simple varia-tions of these methods, while others are unique. Sev-eral review articles are available that provide a morecomplete view of methodology (11, 12, 28).

29.2.3 Sample Derivatization

The compounds one wishes to determine by GCmust be thermally stable under the GC conditionsemployed. Thus, for some compounds (e.g., pesti-cides, aroma compounds, PCBs, and volatile contami-nants), the analyst can simply isolate the componentsof interest from a food as discussed above and directlyinject them into the GC. For compounds that are ther-mally unstable, are too low in volatility (e.g., sugarsand amino acids), or yield poor chromatographic sepa-ration due to polarity (e.g., phenols or acids), a deriva-tization step must be included prior to GC analysis(see also Chaps. 10 and 14). A listing of some of thereagents used in preparing volatile derivatives for GCis given in Table 29-1. The conditions of use for thesereagents are often specified by the supplier or can befound in the literature.

29-1table Reagents Used for Making Volatile Derivatives of Food Components for GC Analysis

Reagent Chemical Group Food Constituent

Silyl reagents Hydroxy, amino carboxylic acids Sugars, sterols, amino acidsTrimethylchlorosilane/

hexamethyldisilazaneBSA [N, O-bis(trimethylsilyl) acetamideBSTFA [N, O-bis (trimethylsilyl)

trifluoroacetamide]t-BuDMCS

(t-butyldimethylchlorosilyl/imidazole)

Esterifying reagentsMethanolic HCIMethanolic sodium methoxideN, N-Dimethylformamide dimethyl acetalBoron trifluoride (or trichloride)/methanol

Carboxylic acids Fatty acids, amines, amino acids,triglycerides, wax esters,phospholipids, cholesteryl esters

MiscellaneousAcetic anhydride/pyridine Alcoholic and phenolic Phenols, aromatic hydroxyl groups,

alcoholsN-trifluoroacetylimidizole/

N-heptafluorobutyrlimidizoleHydroxy and amines Same as above

Alkylboronic acids Polar groups on neighboring atomsO-alkylhydroxylamine Compounds containing both hydroxyl

and carbonyl groupsKetosteroids, prostaglandins


29-4f igure

Diagram of a gas chromatographic system. (Courtesy of Hewlett Packard Co., Analytical Customer Training,Atlanta, GA.)

29-2table

Gas Chromatographic Hardware and

Operating Conditions to be Recorded for

All GC Separations

Parameter Description

Sample Name and injection volumeInjection Type of injection [e.g., split vs. splitless

and conditions (injection port flowrates)]

Column Length, diameter (material-packedcolumns), and manufacturer

Packing/phase Packed columns – solid support; sizemesh; coating; loading (%)

Capillary columns – phase material andthickness

Temperatures Injector; detector; oven and anyprogramming information

Carrier gas Flow rate (velocity) and typeDetector TypeData output Attenuation and chart speed

29.3 GAS CHROMATOGRAPHIC HARDWAREAND COLUMNS

The major parts of a GC are the gas supply system,injection port, oven, column, detector, electronics,and recorder/data handling system (Fig. 29-4). Thehardware as well as operating parameters used inany GC analysis must be accurately and completelyrecorded. The information that must be included ispresented in Table 29-2.

29.3.1 Gas Supply System

The gas chromatograph will require at least a supply ofcarrier gas, and, most likely, gases for the detector (e.g.,hydrogen and air for a FID). The gases used must beof high purity and all regulators, gas lines, and fittingsmust be of good quality. High-quality pressure regula-tors must be used to provide a stable and continuousgas supply. The regulators should have stainless steelrather than polymer diaphragms since polymers willgive off volatiles that may contribute peaks to the ana-lytical run. All gas lines must be clean and contain noresidual drawing oil. Nitrogen, helium, and hydrogengases are typically used as the carrier gas to transportthe analytes in the GC column. The carrier gas lineshould have traps (moisture trap, oxygen trap, andhydrocarbon trap) in line to remove any moisture andcontaminants from the incoming gas. These traps mustbe periodically replaced to maintain effectiveness.

29.3.2 Injection Port

29.3.2.1 Hardware

The injection port serves the purpose of providing aplace for sample introduction, its vaporization, andpossibly some dilution and splitting. Liquid samplesmake up the bulk of materials analyzed by GC, andthey are always done by syringe injection (manual orautomated). The injection port contains a soft septumthat provides a gas-tight seal but can be penetrated bya syringe needle for sample introduction.


29-5f igure

Schematic of a GC injection port. (Courtesy of Hewlett-Packard Co., Analytical Customer Training, Atlanta, GA.)

Samples may be introduced into the injection portusing a manual syringe technique or an automatedsampling system. Manual sample injection is gener-ally the largest single source of poor precision in GCanalysis. Ten-microliter syringes are usually chosensince they are more durable than the microsyringes,and sample injection volumes typically range from 1to 3 µl. These syringes will hold about 0.6 µl in the nee-dle and barrel (this is in addition to that measured onthe barrel). Thus the amount of sample that is injectedinto the GC depends upon the proportion of this 0.6 µlthat is included in the injection and the ability of theanalyst to accurately read the desired sample volumeon the syringe barrel. This can be quite variable forthe same analyst and be grossly different between ana-lysts. This variability between injections and the smallsample volumes injected is the reason why internal(vs. external) standards are common for GC.

29.3.2.2 Sample Injection Techniques

The sample must be vaporized in the injection portin order to pass through the column for separation.This vaporization can occur quickly by flash evapora-tion (standard injection ports) or slowly in a gentlermanner (temperature-programmed injection port oron-column injection). The choice depends upon thethermal stability of the analytes. Due to the varioussample as well as instrumental requirements, there areseveral different designs of injection ports available.

29.3.2.2.1 Split Injection Capillarycolumnshavelim-ited capacity, and the injection volume may have to bereduced to permit efficient chromatography. The injec-tion port may serve the additional function of split-ting the injection so that only a portion of the ana-lyte goes on the column (i.e., split injection) (Fig. 29-5).The injection port is operated about 20◦C warmer thanthe maximum column oven temperature. The samplemay be diluted with carrier gas to accomplish a split(1:50 to 1:100 preferred), whereby only a small por-tion (1 part) of the analyte (more exactly, 1 part ofgas flow) goes on the column, and the majority (44–99 parts) of the analytes are vented to the split vent.High split ratio typically gives a sharp, narrow peak.

29.3.2.2.2 Splitless Injection To increase the sensi-tivity, a splitless injection mode can be used. In split-less injection, the split vent valve is closed and all ofthe analyte goes on the column (Fig. 29-5). Similar tothe split injection, the temperature of the injector isoperated at 20◦C higher than the maximum columnoven temperature. Splitless injection requires to setup the initial column temperature at least 20◦C lowerthan the boiling point of the sample solvent, so thesolvent can recondense in the column for acceptablechromatography of early eluting compounds.

29.3.2.2.3 Temperature Programmed Injection Fortemperature-programmed injection ports, the sample


is introduced into an ambient temperature port andthen it is temperature programmed to some desiredtemperature. Since the sample is not introduced to thehot injector, the technique is desired for temperature-sensitive analytes. In addition, this technique is veryuseful to inject a large amount sample when it isused together with a split/splitless injection mode toincrease the sensitivity. For example, 10 µl of liquidsample can be injected at low temperature using a highsplit ratio to let the solvent vent out and then the injec-tion mode can be changed to “splitless” as the injectoris heated up to transfer all analytes onto the column.

29.3.2.2.4 On-Column Injections On-column injec-tion is a technique whereby the sample is directlyintroduced into the column whose temperature is atthat of the GC oven or that of the room. The sample isthen slowly volatized as the oven heats up. The initialoven temperature needs to be below the boiling pointof the solvent. This technique is good for thermallylabile analytes.

29.3.2.2.5 Thermal Desorption Injection The vola-tiles can be introduced onto the head of GC columnfor chromatographic separation directly from foodsamples through thermal desorption. The sample isheated in a thermal desorption unit, and the volatilesare carried through a purge gas to a split/splitlessinjector. Cryofocusing with liquid nitrogen either inthe injector or column is needed to attain sharp peaks.Alternatively, the volatiles can be retained using a trap

such as TenaxTM

during the purge stage and then ther-mally desorbed onto the column. The samples can beextracted with SPME or SBSE described previously(Sect. 29.2.2.5) and then thermally desorbed into thecolumn for analysis. This technique has gained pop-ularity to analyze volatile aroma compounds in foodsincluding fruits (29, 30) and wine (31).

29.3.3 Oven

The oven controls the temperature of the column. InGC, one takes advantage of both an interaction ofthe analyte with the stationary phase and the boil-ing point for separation of compounds. Thus, theinjection is often made at a lower oven tempera-ture and is then temperature programmed to someelevated temperature. While analyses may be doneisothermally, compound elution time and resolu-tion are extremely dependent upon temperature, sotemperature-programmed runs are most common. Itshould be obvious that higher temperatures will causethe sample to elute faster and, therefore, be at a cost ofresolution. Oven temperature program rates can range

from as little as 0.1◦C/min to the maximum tempera-ture heating rate that the GC can provide. A rate of2–10◦C/min is most common.

The capillary column also can be directly heatedwith an insulated heating wire based on low ther-mal mass (LTM) technology. A temperature sensoris mounted on the column. The column, the heat-ing wire, and the sensor are all coiled together andwrapped with aluminum foil. The column can be uni-formly heated very rapidly to improve the separationand efficiency. Since the system does not have muchvoid volume and other insulation materials, it coolsvery quickly. The total heating and cooling cycle ismuch shorter than the traditional standard GC oven,which makes it ideal for fast GC analysis. The mod-ule is available with almost any standard capillary GCcolumn.

29.3.4 Column and Stationary Phases

The GC column may be either packed or capillary.Early chromatography was done on packed columns,but the advantages of capillary chromatography sogreatly outweigh those of packed column chromatog-raphy that few packed column instruments are soldany longer (Fig. 29-6). While some use high resolutiongas chromatography (HRGC) to designate capillaryGC, GC today means capillary chromatography tomost individuals.

29.3.4.1 Packed Columns

The packed column is most commonly made of stain-less steel or glass and may range from 1.6 to 12.7 mmin outer diameter and be 0.5–5.0 m long (generally 2–3 m). It is packed with a granular material consistingof a “liquid” coated on an allegedly inert solid sup-port. The solid support is most often diatomaceousearth (skeletons of algae) that has been purified, possi-bly chemically modified (e.g., silane treated), and thensieved to provide a definite mesh size (60/80, 80/100,or 100/120).

The liquid loading is usually applied to the solidsupport at 1–10% by weight of the solid support. Whilethe liquid coating can be any one of the approximately200 available, the most common are silicone-basedphases (methyl, phenyl, or cyano substituted) andCarbowax (ester based).

The liquid phase and the percent loading aredetermined by the analysis desired. The choice of liq-uid is typically such that it is of similar polarity tothe analytes to be separated. Loading influences timeof analysis (retention time is proportional to loading),resolution (generally improved by increasing phase


29-6f igure

Comparison of gas chromatographic separa-tion of perfume base using packed (top)and capillary columns (bottom). (Courtesy ofHewlett-Packard Co., Analytical CustomerTraining, Atlanta, GA.)

loading, within limits), and bleed. The liquid coat-ings are somewhat volatile and will be lost from thecolumn at high temperatures (this is dependent uponthe phase itself). This results in an increasing baseline(column bleeding) during temperature programming.

As many as 200 different liquid phases have beendeveloped for GC. As GC has changed from packed tocapillary columns, fewer stationary phases are now inuse since column efficiency has substituted for phaseselectivity (i.e., high efficiency has resulted in bet-ter separations even though the stationary phase isless suited for the separation). Now we find fewerthan a dozen phases in common use (Table 29-3). Themost durable and efficient phases are those based onpolysiloxane (−Si−O−Si−).

Stationary phase selection involves some intu-ition, knowledge of chemistry, and help from the col-umn manufacturer and the literature. There are gen-eral rules, such as choosing polar phases to separatepolar compounds and the converse or phenyl-basedcolumn phase to separate aromatic compounds. How-ever, the high efficiency of capillary columns oftenresults in separation even though the phase is not opti-mal. For example, a 5% phenyl-substituted methyl

silicone phase applied to a capillary column will sep-arate polar as well as nonpolar compounds and is acommonly used phase coating.

29.3.4.2 Capillary Columns

The capillary column is a hollow fused silica glass(<100 ppm impurities) tube ranging in length from5 to 100 m. The walls are so thin, ca. 25 µm, thatthey are flexible. The column outer walls are coatedwith a polyamide material to enhance strength andreduce breakage. Column inner diameters are typi-cally 0.1 mm (microbore), 0.2–0.32 mm (normal capil-lary), or 0.53 mm (megabore).

Megabore columns (0.53 mm i.d.) were initiallydesigned to replace packed columns without modifi-cation of instrumentation hardware. The most com-monly used capillary columns are now 0.32 mm and0.25 mm i.d. columns. Smaller diameter columns(0.10 mm and 0.18 mm i.d.) are used for fast GC anal-ysis. The most common lengths of the GC column are15, 30, and 60 m, although special columns can be over100 m. Longer columns require longer analysis time.Although a longer column gives improved resolution,this benefit of better separation is not particularly obvi-ous due to already high resolution power of capillaryGC column.

Liquid coating is chemically bonded to the glasswalls of capillary columns and internally crosslinkedto give phase thicknesses ranging from 0.1 to 5 µm.Film thickness directly affects separation. Thickerfilms retain compounds longer in the stationary phase,thus the analytes will have longer interaction withthe stationary phase to achieve separation. Generally,thick filmed column should be used to separate veryvolatile compounds. For example, a FFAP columnwith 1-µm film thickness can effectively hold and sep-arate dimethyl sulfide (H2S) and other highly volatilesulfur compounds (32). However, a thick film alsowill give a higher baseline due to bleeding. A thinfilm (0.25 µm) column is usually used to separate highmolecular weight compounds; the analytes will stay inthe stationary phase less time. Thin film columns alsohave less breeding at high temperature, and they arefrequently used for GC-MS.

Most compounds can be separated using nonpolar5% phenyl 95% dimethylpolysiloxane-based columns(e.g., DB-5, HP-5, RTX-5). This type of column hasa very wide temperature range (−60◦C to 325◦C)and is very stable. However, to separate very polarcompounds, such as alcohols and free fatty acids, apolar column is needed such as XX-WAX (polyethy-lene glycol) or XX-FFAP (polyethylene glycol treatedwith nitroterephthalic acid). A wax-type column hassuperior separation power; however, it has a narrow


29-3table Common Stationary Phases

Composition Polarity Applicationsa

Phases with SimilarMcReynoldsConstantsb Temperature Limitsc

100% Dimethyl polysiloxane(gum)

Nonpolar Phenols, hydrocarbons,amines, sulfur compounds,pesticides, PCBs

OV-1, SE-30 −60◦C to 325◦C

100% Dimethyl polysiloxane(fluid)

Nonpolar Amino acid derivatives,essential oils

OV-101, SP-2100 0–280◦C

5% Phenyl 95% dimethylpolysiloxane

Nonpolar Fatty acids, methyl estersalkaloids, drugs,halogenated compounds

SE-52, OV-23, SE-54 −60◦C to 325◦C

14% Cyanopropylphenylmethyl polysiloxane

Intermediate Drugs, steroids, pesticides OV-1701 −200◦C to 280◦C

50% Phenyl, 50% methylmethyl polysiloxane

Intermediate Drugs, steroids, pesticides,glycols

OV-17 60–240◦C

50% Cyanopropylmethyl, 50%phenyl methyl polysiloxane

Intermediate Fatty acids, methyl esters,alditol acetates

OV-225 60–240◦C

50% Trifluoropropylpolysiloxane

Intermediate Halogenated compounds,aromatics

OV-210 45–240◦C

Polyethylene glycol-TPAmodified

Polar Acids, alcohols, aldehydes,acrylates, nitrites, ketones

OV-351, SP-1000 60–240◦C

Polyethylene glycol Polar Free acids, alcohols, esters,essential oils, glycols,solvents

Carbowax 20M 60–220◦C

aSpecific application notes from column suppliers provide information for choosing a specific column.bMcReynolds constants are used to group stationary phases together on the basis of separation properties.cStationary phases have both upper and lower temperature limits. Lower temperature limit is often due to a phase change (liquidto solid) and upper temperature limit to a volatilization of phase.

usable temperature range (40–240◦C). It bleeds highlyat high temperature and becomes solid (lost separa-tion power) at low temperature. It is also sensitiveto residue oxygen in the carrier gas, and it deterio-rates quickly if oxygen is not removed in the carriergas. Other specialty phase columns have been devel-oped to improve specific resolution. Cyanopropyl-based columns (SP-2560, CP-Sil 88) are good for transfatty acid esters. A cyclodex-based column is useful toseparate stereoisomers of many flavor compounds.

29.3.4.3 Gas–Solid (PLOT) Chromatography

Gas–solid chromatography is a very specialized areaof chromatography accomplished without using a liq-uid phase – the analyte interaction is with a porousmaterial. This material has been applied both topacked and capillary columns. For the capillary col-umn, the porous material is chemically or physi-cally (by deposition) coated on the inner wall ofthe capillary and the column is called porous-layeropen-tabular (PLOT) column. The most popularporous materials are alumina oxide, carbon, molec-ular sieve, and synthetic polymers such as Poropakor Chromosorb (trade names of polymers based on

vinyl benzene). Separations usually involve water orother very volatile compounds such as headspace gascomposition (N2, O2, CO2, CO) in packaged food andethylene during fruit ripening and storage.

29.3.5 Detectors

There are numerous detectors available for GC, eachoffering certain advantages in either sensitivity (e.g.,electron capture) or selectivity (e.g., atomic emissiondetector). The most common detectors are the FID,thermal conductivity (TCD), electron capture (ECD),flame photometric (FPD), pulsed flame photomet-ric (PFPD), and photoionization (PID) detectors. Theoperating principles and food applications of thesedetectors are discussed below. The characteristics ofthese detectors are summarized in Table 29-4.

29.3.5.1 Thermal Conductivity Detector

29.3.5.1.1 Operating Principles As the carrier gaspasses over a hot filament (tungsten), it cools thefilament at a certain rate depending on carrier gasvelocity and composition. The temperature of the fil-ament determines its resistance to electrical current.


29-4table Characteristics of Most Common Detectors for Gas Chromatography

Thermal Conductivity Flame Ionization Electron Capture Flame Photometric PhotoionizationCharacteristic Detector Detector Detector Detector Detector

Specificity Very little; detectsalmost anything,including H2O;called the“universaldetector”

Most organics Halogenatedcompounds andthose with nitro orconjugated doublebonds

Organic compoundswith S or P(determined bywhich filter isused)

Depends onionization energyof lamp relative tobond energy ofsolutes

Sensitivitylimits

ca. 400 pg; relativelypoor; varies withthermal propertiesof compound

10–100 pg for mostorganics; verygood

0.05–1 pg; excellent 2 pg for S and 0.9 pgfor P compounds;excellent

1–10 pg dependingon compound andlamp energy;excellent

Linear range 104 – poor;response easilybecomesnonlinear

106–107 – excellent 104 – poor 104 for P; 103 for S 107 – excellent

29-7f igure

Schematic of the thermal conductivity detector.(Courtesy of Hewlett-Packard Co., AnalyticalCustomer Training, Atlanta, GA.)

As a compound elutes with the carrier gas, the cool-ing effect on the filament is typically less, resulting ina temperature increase in the filament and an increasein resistance that is monitored by the GC electronics.Older style TCDs used two detectors and two match-ing columns; one system served as a reference and theother as the analytical system. Newer designs use onlyone detector (and column), which employs a carriergas switching value to pass alternately carrier gas orcolumn effluent through the detector (Fig. 29-7). Thesignal is then a change in cooling of the detector asa function of which gas is passing through the detec-tor from the analytical column or carrier gas supply(reference gas flow).

The choice of carrier gas is important since differ-ences between its thermal properties and the analytesdetermine response. While hydrogen is the best choice,helium is most commonly used since hydrogen isflammable.

29.3.5.1.2 Applications The most valuable proper-ties of this detector are that it is universal in responseand nondestructive to the sample. Thus, it is used infood applications for which there is no other detec-tor that will adequately respond to the analytes (e.g.,water, permanent gases, CO) or when the analystwishes to recover the separated compounds for furtheranalysis (e.g., trap the column effluent for infrared,nuclear magnetic resonance (NMR), or sensory anal-ysis). It does not find broad use because it is relativelyinsensitive, and often the analyst desires specificity indetector response to remove interfering compoundsfrom the chromatogram.

29.3.5.2 Flame Ionization Detector

29.3.5.2.1 Operating Principles As compounds elutefrom the analytical column, they are burned in ahydrogen flame (Fig. 29-8). A potential (often 300 V)is applied across the flame. The flame will carry acurrent across the potential which is proportional tothe organic ions present in the flame from the burn-ing of an organic compound. The current flowingacross the flame is amplified and recorded. The FIDresponds to organics on a weight basis. It gives virtu-ally no response to H2O, NO2, CO2, H2S and limitedresponse to many other compounds. Response is bestwith compounds containing C−C or C−H bonds.


29-8f igure

Schematic of the flame ionization detectordesigned for use with capillary columns.(Courtesy of Hewlett-Packard Co., AnalyticalCustomer Training, Atlanta, GA.)

29.3.5.2.2 Applications The food analyst is mostoften working with organic compounds, to which thisdetector responds well. Its very good sensitivity, widelinear range in response (necessary in quantitation),and dependability make this detector the choice formost food work. Thus, this detector is used for vir-tually all food analyses for which a specific detector isnot desired or sample destruction is acceptable (col-umn eluant is burned in flame). This includes, forexample, flavor studies, fatty acid analysis, carbo-hydrate analysis, sterols, contaminants in foods, andantioxidants.

29.3.5.3 Electron Capture Detector

29.3.5.3.1 Operating Principles The ECD contains aradioactive foil coating that emits electrons as it under-goes decay (Fig. 29-9). The electrons are collected onan anode, and the standing current is monitored byinstrument electronics. As an analyte elutes from theGC column, it passes between the radioactive foiland the anode. Compounds that capture electronsreduce the standing current and thereby give a mea-surable response. Halogenated compounds or thosewith conjugated double bonds give the greatest detec-tor response. Unfortunately this detector becomes sat-urated quite easily and thus has a very limited linearresponse range.

29-9f igure

Schematic of the electron capture detector.(Courtesy of Hewlett-Packard Co., AnalyticalCustomer Training, Atlanta, GA.)

29-10f igure

Schematic of the flame photometric detector.(Courtesy of Hewlett-Packard Co., AnalyticalCustomer Training, Atlanta, GA.)

29.3.5.3.2 Applications In food applications, theECD has found its greatest use in determining PCBsand pesticide residues (see Chap. 18). The specificityand sensitivity of this detector make it ideal for thisapplication.

29.3.5.4 Flame Photometric Detector andPulsed Flame Photometric Detector

29.3.5.4.1 Operating Principles The FPD detectorworks by burning all analytes eluting from the analyt-ical column and then measuring specific wavelengthsof light that are emitted from the flame using a fil-ter and photometer (Fig. 29-10). The wavelengths


29-11f igure

Comparison of flame photometric detector (a) and pulsed flame photometric detector (b). (Courtesy of VarianInc., Palo Alto, CA.)

of light that are suitable in terms of intensity anduniqueness are characteristic of sulfur (S) and phos-phorus (P). Thus this detector gives a greatly enhancedsignal for these two elements (several thousandfoldfor S- or P-containing organic molecules vs. non-S orP-containing organic molecules). Detector response toS-containing molecules is nonlinear and thus quantifi-cation must be done with care.

The PFPD is very similar to FPD. Unlike tradi-tional flame photometric detection (FPD), which usesa continuous flame, the PFPD ignites, propagates, andself-terminates 2–4 times per second (Fig. 29-11). Spe-cific elements have their own emission profile: hydro-carbons will complete emission early while sulfuremissions begin at a relatively later time after combus-tion. Therefore, a timed “gate delay” can selectivelyallow for only emissions due to sulfur to be integrated,producing a clean chromatogram. This timed “gatedelay” greatly improves the sensitivity. The PFPDcan detect sulfur-containing compounds at a muchlower detection limit than nearly all other methods ofdetection (33).

29.3.5.4.2 Applications Both the FPD and the PFPDhave found major food applications in the determi-nation of organophosphorus pesticides and volatilesulfur compounds in general. The determination ofsulfur compounds has typically been in relation toflavor studies.

29-12f igure

Schematic of the photoionization detector.(Courtesy of Hewlett-Packard Co., AnalyticalCustomer Training, Atlanta, GA.)

29.3.5.5 Photoionization Detector

29.3.5.5.1 Operating Principles The photoioniza-tion detector (PID) uses ultraviolet (UV) irradiation(usually 10.2 eV) to ionize analytes eluting from theanalytical column (Fig. 29-12). The ions are acceleratedby a polarizing electrode to a collecting electrode. Thesmall current formed is magnified by the electrometerof the GC to provide a measurable signal.


This detector offers the advantages of being quitesensitive and nondestructive and may be operatedin a selective response mode. The selectivity comesfrom being able to control the energy of ionization,which will determine the classes of compounds thatare ionized and thus detected.

29.3.5.5.2 Applications The PID finds primary usein analyses for which excellent sensitivity is requiredfrom a nondestructive detector. This is most oftena flavor application in which the analyst wishes tosmell the GC effluent to determine the sensory char-acter of the individual GC peaks. While this detectormight find broader use, the widespread availabilityof the FID (which is suitable for most of the sameapplications) meets most of these needs.

29.3.5.6 Electrolytic Conductivity Detector

29.3.5.6.1 Operating Principles Compounds enter-ing the electrolytic conductivity detector (ELCD) aremixed with a reagent gas (oxidizing or reducingdepending on the analysis) in a nickel reaction tubeproducing ionic species. These products are mixedwith a deionized solvent, interfering ions are scrubbedfrom the effluent, and the ionic analyte-transformationproduct is detected within the electrolyte conductiv-ity cell. This detector can be used for the specificdetection of sulfur-, nitrogen-, or halogen-containingmolecules. For example, when operated in the nitro-gen mode, analyte is mixed with H2 gas and hydro-genated over a nickel catalyst at 850◦C. Acidic hydro-genation products are removed from the effluent bypassage through a Sr(OH)2 trap and the NH3 fromthe analyte passes to the conductivity cell where it ismeasured (34).

29.3.5.6.2 Applications This detector can be usedin many applications for which element specificityis desired. Examples would be pesticide, herbicide,nitrosamine, or flavor analysis. The ELCD is veryselective and quite sensitive having detection limitsof 0.1–1 pg of chlorinated compounds, 2 pg for sulfur,and 4 pg for nitrogen.

29.3.5.7 Thermionic Detector

29.3.5.7.1 Operating Principles The thermionic det-ector (also called the nitrogen phosphorus detector,NPD) is a modified FID in which a nonvolatile ceramicbead is used to suppress the ionization of hydrocar-bons as they pass through a low-temperature fuel-poor hydrogen plasma. The ceramic bead is typicallycomposed of rubidium which is heated to 600–800◦C.Most commonly this detector is used for the selective

detection of nitrogen- or phosphorus-containing com-pounds. It does not detect inorganic nitrogen orammonia.

29.3.5.7.2 Applications This detector is primarilyused for the measurement of specific classes of flavorcompounds, nitrosamines, amines, and pesticides.

29.3.5.8 Hyphenated Gas ChromatographicTechniques

Hyphenated gas chromatographic techniques arethose that combine GC with another major technique.Examples are GC-AED (atomic emission detector),GC-FTIR (Fourier transform infrared), and GC-MS(mass spectrometry). While all of the techniques areestablished methods of analysis in themselves, theybecome powerful tools when combined with a tech-nique such as GC. GC provides the separation and thehyphenated technique the detector. GC-MS has longbeen known to be a most valuable tool for the identifi-cation of volatile compounds (see Chap. 26). The MS,however, may perform the task of serving as a specificdetector for the GC by selectively focusing on ion frag-ments unique to the analytes of interest. The analystcan detect and quantify components without their gaschromatographic resolution in this manner. The samestatements can be made about GC-FTIR (see Chap. 24).The FTIR can readily serve as a GC detector.

A relatively new combination is GC-AED. Inthis technique, the GC column effluent enters amicrowave-generated helium plasma that excites theatoms present in the analytes. The atoms emit light attheir characteristic wavelengths, and this emission ismonitored using a diode ray detector similar to thatused in HPLC. This results in a very sensitive andspecific elemental detector.

29.3.5.9 Multidimensional GasChromatography

Multidimensional gas chromatography (MDGC)greatly increases the separation ability of gas chro-matography (35). By simply coupling two GCcolumns, each of opposite polarity, an overallimprovement in separation can be accomplished (36).However, this tandem operation of GC columns doesnot actually represent multidimensionality, but ratherresembles the use of a mixed-stationary phase column(35). True MDGC involves a process known as orthog-onal separation in which a sample is first dispersedby one column, and the simplified subsamples arethen applied onto another column for further sepa-ration. MDGC techniques can be generally divided


into two classes: (1) conventional, or “heart-cut,”MDGC and (2) comprehensive two-dimensional gaschromatography (GC×GC).

29.3.5.9.1 Conventional Two-Dimensional GC Con-ventional two-dimensional GC is achieved by usingcoupled capillary columns for which a small portion,or heart-cut, of the effluent from the first (“presepa-ration”) column is transferred to the second (“analyt-ical”) column. The concept of conventional MDGC isalmost identical to that of preparative GC operations,for which one column is used to obtain a partially sep-arated fraction of a complex aroma mixture, whichis then reinjected onto another GC column, usuallywith an opposite stationary phase, for further separa-tion. The only difference is that with MDGC there areno requirements for manual collection of the effluentobtained from the preseparation column since the twocolumns are directly connected.

Because the second column in the MDGC sys-tem is only injected with a small portion of the totalsample at one time, a large quantity of the sample can

be injected onto the first column without the worryof chromatographic band smearing during analyti-cal separations (37). Therefore, trace compounds canbe easily enriched for more successful detection andidentification.

The MDGC technique is particularly useful tostudy enantiomers of flavor compounds. The inter-ested compound can be “heart-cut” and transferredto an analytical column with an enantioselective sta-tionary phase for good separation of targeted chiralcompounds.

29.3.5.9.2 Comprehensive Two-Dimensional GCComprehensive two-dimensional MDGC is amongthe most powerful two-dimensional gas chromato-graphic techniques that have been developed today(Fig. 29-13). Unlike conventional MDGC in whichonly particular segments are transferred from thepreseparation column onto the analytical column,comprehensive MDGC, or GC × GC, involves thetransfer of the entire effluent from the first col-umn onto a second column by way of a modulation

29-13f igure

Total ion chromatograms and their respective two-dimensional contour plots for an Arabica coffee extract sep-arated by GC × GC using two different column sets: polar × nonpolar (a and c) along with nonpolar × polar(b and d). [Reprinted from reference (45), used with permission.]


interface so that complete two-dimensional data canbe obtained for the entire run of the first column. Theoperation of the modulator involves the generation ofnarrow injection bands from the first column, whichare continuously, but individually, sent to the sec-ondary column for final separation. GC × GC requiresthat the second column can operate quickly enough togenerate a complete set of data during the time thata single peak elutes from the first GC column, gener-ally within 5 s (35, 38). The data from both time axesare combined to create a set of coordinates for eachpeak so that the resultant chromatogram is actually atwo-dimensional (2D) plane rather than a straight line.Peak area information can be obtained by summingthe integration over both dimensions.

In comprehensive GC × GC, the two columnsperform independently of each other, therefore theoverall peak capacity becomes the product of thecapacities for each column. Because analytes elutefrom the second column so quickly, data acquisitionmust be adequately fast enough for proper detec-tion. Time-of-flight mass spectrometry (TOF-MS) andrapid-scanning quadrupole mass spectrometry (qMS)have both been used as effective detection methods forGC × GC to obtain mass spectral information (39, 40).

Although the instrumentation can be quite expen-sive, the use of comprehensive GC × GC for volatilearoma analysis has exponentially increased over thepast few years as methodologies have become moreestablished and systems have become commerciallyavailable. Overall, the application of MDGC, bothconventional and comprehensive, has allowed foradvanced separations of complex aromas to occur byusing state-of-the-art instrumentation.

29.4 CHROMATOGRAPHIC THEORY

29.4.1 Introduction

GC may depend on several types (or principles)of chromatography for separation. The principles ofchromatographic separation are discussed in Chap. 27,Sect. 27.4. For example, size-exclusion chromatogra-phy is used in the separation of permanent gases suchas N2, O2, and H2. A variation of size exclusion is usedto separate chiral compounds on cyclodextrin-basedcolumns; one enantiomorphic form will fit better intothe cavity of the cyclodextrin than will the other form,resulting in separation. Adsorption chromatography isused to separate very volatile polar compounds (e.g.,alcohols, water, and aldehydes) on porous polymercolumns (e.g., TenaxR phase). Partition chromatogra-phy is the workhorse for gas chromatographic sep-arations. There are over 200 different liquid phasesthat have been developed for gas chromatographic use

over time. Fortunately, the vast majority of separationscan be accomplished with only a few of these phases,and the other phases have fallen into disuse. GCdepends not only upon adsorption, partition, and/orsize exclusion for separation, but also upon solute boil-ing point for additional resolving powers. Thus, theseparations accomplished are based on several proper-ties of the solutes. This gives GC virtually unequaledresolution powers as compared with most other typesof chromatography (e.g., HPLC, paper, or thin-layerchromatography).

A brief discussion of chromatographic theory willfollow. The purpose of this additional discussion is toapply this theory to GC to optimize separation effi-ciency so that analyses can be done faster, less expen-sively, or with greater precision and accuracy. If oneunderstands the factors influencing resolution in GC,one can optimize the process and gain in efficiency ofoperation.

29.4.2 Separation Efficiency

A good separation has narrow-based peaks and ide-ally, but not essential to quality of data, baseline sep-aration of compounds. This is not always achieved.Peaks broaden as they pass through the column – themore they broaden, the poorer is the separation andefficiency. As discussed in Chap. 27, Sect. 27.5.2.2.2,a measure of this broadening is height equivalent toa theoretical plate (HETP). This term is derived fromN, the number of plates in the column, and L, thelength of the column. A good packed column mighthave N = 5000, while a good capillary column shouldhave about 3000–4000 plates per meter for a total of100,000–500,000 plates depending on column length.HETP will range from about 0.1 to 1 mm for goodcolumns.

29.4.2.1 Carrier Gas Flow Rates and ColumnParameters

Several factors influence column efficiency (peakbroadening). As presented in Chap. 27, these arerelated by the Van Deemter equation [1]: (HETPvalues should be small.)

HETP = A + B/u + Cu [1]

where:

HETP = height equivalent to a theoretical plate

A = eddy diffusion

B = band broadening due to diffusion

u = velocity of the mobile phase

C = resistance to mass transfer


29-14f igure

Illustration of flow properties that lead tolarge eddy diffusion (term A).

29-15f igure

The influence of column diameter on col-umn efficiency (plates/meter). (Courtesy ofHewlett-Packard, Analytical Customer Train-ing, Atlanta, GA.)

A is eddy diffusion; this is a spreading of theanalytes in the column due to the carrier gas hav-ing various pathways or nonuniform flow (Fig. 29-14).In packed column chromatography, poor uniformityin solid support size or poor packing results in chan-neling and multiple pathways for carrier flow, whichresults in spreading of the analyte in the column. Thus,improved efficiency is obtained by using the highperformance solid supports and commercially packedcolumns.

In capillary chromatography, the A term is rela-tively very small. However, as the diameter of thecapillary column increases, the flow properties dete-riorate, and band spreading occurs. The most effi-cient capillary columns have small diameters (0.1 mm),and efficiency decreases rapidly as one goes to mega-bore columns (Fig. 29-15). Megabore columns areonly slightly more efficient than packed columns.While column efficiency increases as we go to smallercolumns, column capacity decreases rapidly. Micro-bore columns are easily overloaded (capacity may be1–5 ng per analyte), resulting again in poor chromatog-raphy. Thus, column diameter is generally chosen as0.2–0.32 mm to compromise efficiency with capacity.

B is band broadening due to diffusion; solutes willgo from a high to a low concentration. The term u isvelocity of the mobile phase. Thus, very slow flowrates result in large amounts of diffusion band broad-ening, and faster flow rates minimize this term. Theterm u is influenced by the carrier gas choice. Larger-molecular-weight carrier gases (e.g., nitrogen) are

more viscous than the lighter-molecular-weight gases(e.g., helium or hydrogen) and thus peak spreadingis less for nitrogen than for helium or hydrogen car-rier gases. This results in nitrogen having the lowestHETP of the carrier gases and theoretically being thebest choice for a carrier gas. However, other consid-erations that will be discussed in Sect. 29.4.2.2 makenitrogen a very poor choice for a carrier gas.

C is resistance to mass transfer. If the flow (u) istoo fast, the equilibrium between the phases is notestablished, and poor efficiency results. This can bevisualized in the following way: If one molecule ofsolute is dissolved in the stationary phase and anotheris not, the undissolved molecule continues to movethrough the column while the other is retained. Thisresults in band spreading within the column. Otherfactors that influence this term are thickness of the sta-tionary phase and uniformity of coating on the phasesupport. Thick films give greater capacity (ability tohandle larger amounts of a solute) but at a cost in termsof band spreading (efficiency of separation) since thickfilms provide more variation in diffusion propertiesin and out of the stationary phase. Thus phase thick-ness is a compromise between maximizing separationefficiency and sample capacity (too much sample –overloading a column – destroys separation ability).Phase thicknesses of 0.25–1 µm are commonly used formost applications.

If the Van Deemter equation is plotted, giving thefigure discussed in Fig. 27-13, we see an optimum inflow rate due to the opposing effects of the B and theC terms. It should be noted that the GC may not beoperated at a carrier flow velocity yielding maximumefficiency (lowest HETP). Analysis time is directly pro-portional to carrier gas flow velocity. If the analysistime can be significantly shortened by operating abovethe optimum flow velocity and adequate resolutionis still obtained, velocities well in excess of optimumshould be used.

29.4.2.2 Carrier Gas Type

The relationship between HETP and carrier gas flowvelocity is strongly influenced by carrier gas choice(Fig. 29-16). Nitrogen is the most efficient (lowestHETP) carrier gas, as discussed in Sect. 29.4.2.1, butits minimum HETP occurs at a very low flow velocity.This low mobile phase velocity results in unnecessar-ily long analysis times. Considering the data plottedin Fig. 29-16, nitrogen has an HETP of about 0.25 atan optimum flow velocity of 10 cm/s. The HETP ofhelium is only about 0.35 at 40 cm/s flow velocity. Thisis a small loss in resolution to reduce the analysis timefourfold (10 cm/s for nitrogen vs. 40 cm/s for helium).


29-16f igure

Influence of carrier gas type and flow rateon column efficiency. (Courtesy of Hewlett-Packard Co., Analytical Customer Training,Atlanta, GA.)

One can potentially even push the flow velocity upto 60 or 70 cm/s and accomplish separation in evenshorter times.

The plots in Fig. 29-16 suggest that hydrogen isan even better choice for a carrier gas than helium(i.e., has a flatter relationship between carrier gas flowvelocity and HETP). However, there are some con-cerns about hydrogen being flammable and reports inthe literature that some compounds may be hydro-genated in the GC system. Additionally, some detec-tors cannot use hydrogen as a carrier gas (e.g., a massspectrometer) and, thus, one may be limited to heliumas a good compromise.

29.4.2.3 Summary of Separation Efficiency

In summary, an important goal of analysis is to achievethe necessary separation in the minimum amount oftime. The following factors should be considered:

1. In general, small diameter columns (packed orcapillary) should be used since separation effi-ciency is strongly dependent on column diam-eter. While small diameter columns will limitcolumn capacity, limited capacity often can becompensated for by increasing phase thickness.Increased phase thickness will also decreasecolumn efficiency but to a lesser extent thanincreasing column diameter.

2. Lower column operating temperatures shouldbe used – if elevated column temperatures arerequired for the compounds of interest to elute,use a shorter column if resolution is adequate.

29-17f igure

Relationships among column capacity, effi-ciency, resolution, and analysis speed.

3. One should keep columns as short as possible(analysis time is directly proportional to col-umn length – resolution is proportional to thesquare root of length).

4. Use hydrogen as the carrier gas if the detec-tor permits. Some detectors have specific carriergas requirements.

5. Operate the GC at the maximum carrier gasvelocity that provides resolution.

The pyramid shown in Fig. 29-17 summarizesthe compromises that must be made in choosing theanalytical column and gas chromatographic operatingconditions. One cannot optimize any given operat-ing conditions and column choices to get one of theseproperties without compromising another property.For example, optimizing chromatographic resolution(small bore capillary diameter, thin phase coating, longcolumn lengths, and slow or optimum carrier gas flowrate) will be at the cost of capacity (large bore columnsand thick phase coating) and speed (thin film coating,high carrier gas flow velocities, and short columns).Capacity will be at a cost of resolution and speed, etc.The choice of column and operating parameters mustconsider the needs of the analyst and the compromisesinvolved in these choices.

29.5 APPLICATIONS OF GC

While some detail on the application of GC to foodanalyses has been presented in Chaps. 10, 14, and18, a few additional examples will be presentedbelow to illustrate separations and chromatographicconditions.


29-18f igure

Typical capillary gas chromatographic separation of residual volatiles in a food packaging film. [From (41), usedwith permission.]

29.5.1 Residual Volatiles in PackagingMaterials

Residual volatiles in packaging materials can be aproblem both from health (if they are toxic) and qual-ity standpoints (produce off flavors in the food). As theindustry has turned from glass to polymeric materi-als, there have been more problems in this respect.GC is most commonly used to determine the residualvolatiles in these materials.

The chromatograms presented in Fig. 29-18 wereproduced by steam distilling a food packaging film,extracting the volatiles from the distillate in an organicsolvent, concentrating the solvent extract, and thenchromatographing it on a capillary column (top chro-matogram in Fig. 29-18) (41). The extreme complexityof the chromatogram required that the concentratebe further fractionated on silica gel and each frac-tion rechromatographed. The chromatograms labeled“cuts 1–5” are the chromatograms resulting from elut-ing the silica gel with: (1) hexane removing satu-rated hydrocarbons from the gel bed (cut 1); (2) 10%CH2Cl2/hexane removing the unsaturated and aro-matic hydrocarbons (cut 2); (3) CH2Cl2 removing theketones and aldehydes (cut 3); (4) methyl-t-butyletherremoving the acids, unsaturated ketones, and aldehy-des (cut 4); and (5) alcohol removing the remainingpolar volatiles (cut 5). One can see that the prefrac-tionation of the extracted packaging volatiles greatlysimplified the chromatography and permitted theresearcher to focus on the volatiles responsible for theoff odor in the packaging material.

29.5.2 Separation of Stereoisomers

GC has found extensive application in the separationof chiral volatile compounds in foods (e.g., D andL-carvone). Chiral separations are most commonlyaccomplished using cyclodextrin-based gas chromato-graphic columns. Cyclodextrins are molecules (6-, 7-,or 8-membered rings of glucose) that have an internalcavity of suitable dimensions to permit the inclusion ofmany small organic molecules. While optical isomersof molecules have virtually identical physical prop-erties and thus they are difficult to separate by mostchromatographic methods, they differ in spatial con-figuration. Stereoisomers of a given compound willbe included in the cyclodextrin cavity of the gas chro-matographic column to a lesser or greater extent asthey flow through a cyclodextrin capillary column andbecome separated.

The chromatogram presented in Fig. 29-19 showsthe separation of six stereoisomers of α and β-irone(42). This separation was accomplished using anoctakis (6-O-methyl-2,3-di-O-pentyl)-γ-cyclodextrin/OV-17 capillary column.

29.5.3 Headspace Analysis of Ethylene Oxidein Spices

Ethylene oxide (ETO) is a highly volatile compoundthat has found use in the food industry as a fumi-gant for spices (43). It has been classified as a suspecthuman carcinogen and thus its residual concentration


29-19f igure

Enantiomeric analysis of irone stereoisomersin orris oil. [Reprinted from (42), p. 180, bycourtesy of Marcel Dekker, Inc., New York.]

in spices is of concern. Because of its volatility, ETO iswell suited to determination by GC.

Woodrow et al. (43) chose to use a headspacemethod for ETO determination. This is reasonablesince ETO is very volatile, sensitivity is adequate,and headspace techniques are simple to perform. Themethod involved adding 1 g of ground spice to a 22-mlheadspace vial (a vial that has a Teflon septum closurefor sampling), adding internal standard (1-octanol),incubating the vial at 60◦C for 20 min, and then remov-ing and injecting ca. 1 ml of the headspace into thegas chromatograph. ETO was separated from othervolatiles in the sample using a porous polymer cap-illary column (divinylbenzene homopolymer). Typicalchromatograms of pure ETO, spice, and spice spikedwith ETO are shown in Fig. 29-20.

29.5.4 Aroma Analysis of Heated Butter

Volatile aroma compounds are important contributorsto the quality of foods. The composition and the con-centration of volatile aroma compounds impact theflavor perceived. GC has been widely applied to definea volatile chemical fingerprint to characterize the fla-vor quality of food products. A chromatogram of the

29-20f igure

Headspace gas chromatographic analysis ofethylene oxide in spices. (a) 3 µg pure ETO, (b)1 µg pure ETO, (c) 3 µg ETO in spice, (d) 1 µg inspice, (e) the pure spice. [From (43), used withpermission.]

volatile compounds in heated butter isolated by astatic headspace-Tenax absorbent technique and sub-sequently analyzed by GC on a wax capillary columnis illustrated in Fig. 29-21. Seven select aroma com-pounds reported to contribute to the flavor of heatedbutter are displayed. Changes in the concentrationsof the volatile flavor compounds can be related tochanges in the flavor properties of foods and provideinsights into the role of processing, storage, ingredi-ents, packaging, etc. on food flavor. Volatile flavorcompounds that originally contributed desirable fla-vor properties can also become undesirable at elevatedlevels and result in off-flavor development. For exam-ple, an off-flavor defect in butter developed duringstorage has been related to an increase in lactone con-centration, such as δ-decalactone (44). The predictionof product shelf-life based on off-flavor developmenttypically involves very volatile compounds, such ashexanal (a common indicator of lipid oxidation) (45).


29-21f igure

Heated butter static headspace GC chroma-togram with select aroma compounddisplayed. [Adapted from (46), used withpermission.]

29.6 SUMMARY

GC has found broad application in both the foodindustry and academia. It is exceptionally well suitedto the analysis of volatile thermally stable compounds.This is due to the outstanding resolving properties ofthe method and the wide variety of detectors that canprovide either sensitivity or selectivity in analysis.

Sample preparation generally involves theisolation of solutes from foods, which may be acc-omplished by headspace analysis, distillation, pre-parative chromatography (including solid-phaseextraction), or extraction (liquid–liquid). Some solutescan then be directly analyzed, while others must bederivatized prior to analysis.

The gas chromatograph consists of a gas supplyand regulators (pressure and flow control), injectionport, column and column oven, detector, electron-ics, and a data recording and processing system. Theanalyst must be knowledgeable about each of theseGC components: carrier and detector gases; injectionport temperatures and operation in split, splitless,temperature-programmed, or on-column modes; col-umn choices and optimization (gas flows and temper-ature profile during separation); and detectors (TCD,FID, NPD, ECD, FPD, PFPD, and PID). The character-istics of these GC components and an understandingof basic chromatographic theory are essential to bal-ancing the properties of resolution, capacity, speed,and sensitivity.

Unlike most of the other chromatographic tech-niques, traditional GC has reached the theoretical lim-its in terms of both resolution and sensitivity. Thus,this method will not change significantly in the futureother than for minor innovations in hardware or asso-ciated computer software. However, two-dimensionalGCs, both heart-cut GC–GC and comprehensive GC ×

GC, are still developing quickly in both instrumenta-tion and applications, especially in the field of flavoranalysis.

GC as a separation technique has been combinedwith AED, FTIR, and MS as detection techniques tomake GC an even more powerful tool. Such hyphen-ated techniques are likely to continue to be developedand refined.

29.7 STUDY QUESTIONS

1. For each of the following methods to isolate solutes fromfood prior to GC analysis, describe the procedure, theapplications, and the cautions in use of the method:

(a) Headspace methods(b) Distillation methods(c) Solvent extraction

2. What is solid-phase extraction and why is it advanta-geous over traditional liquid–liquid extractions?

3. Why must sugars and fatty acids be derivatized beforeGC analysis, while pesticides and aroma compoundsneed not be derivatized?

4. Why is the injection port of a GC at a higher temperaturethan the oven temperature?

5. Differentiate packed columns from capillary columns(microbore and megabore) with regard to physical char-acteristics and column efficiency.

6. You are doing GC with a packed column and notice thatthe baseline rises from the beginning to the end of eachrun. Explain a likely cause for this increase.

7. The most common detectors for GC are TCD, FID, ECD,FPD, and PID. Differentiate each of these with regardto the operating principles. Also, indicate below whichdetector(s) fits the description given.

(a) Least sensitive(b) Most sensitive(c) Least specific(d) Greatest linear range(e) Nondestructive to sample(f) Commonly used for pesticides(g) Commonly used for volatile sulfur compounds

8. What types of chromatography does GC rely upon forseparation of compounds?

9. In GC, explain why a balance has to be maintainedbetween efficiency and capacity. Also, give an exam-ple situation in which you would sacrifice capacity forefficiency.

10. You plan to use GC to achieve good chromatographicseparation of Compounds A, B, and C in your food


sample. You plan to use an internal standard to quanti-tate each compound. By answering the following ques-tions, describe how using an internal standard works forthis purpose (see also Chap. 27, Sect. 27.5.3).

(a) How do you choose the internal standard for yourapplication?

(b) What do you do with the internal standard, relativeto the standard solutions for Compounds A, B, and Cand relative to the food sample? Be specific in youranswer.

(c) What do you measure?(d) If you were to prepare a standard curve, what would

you plot?(e) Why are internal standards commonly used for GC?

11. A fellow lab worker is familiar with HPLC for food anal-ysis but not with GC. As you consider each componentof a typical chromatographic system (and specificallythe components and conditions for GC and HPLC sys-tems), explain GC to the fellow worker by comparingand contrasting it to HPLC. Following that, state ingeneral terms the differences among the types of sam-ples appropriate for analysis by GC vs. HPLC and giveseveral examples of food constituents appropriate foranalysis by each (see also Chap. 28).

29.8 REFERENCES

1. James AT, Martin AJP (1952) Gas–liquid chromatogra-phy: the separation and microestimation of volatile fattyacids from formic acid to dodecanoic acid. BiochemJ 50:679

2. Niessen WMA (2001) Current practice of gas chromatog-raphy – mass spectrometry. Marcel Dekker, New York

3. Rood D (1999) A practical guide to the care, mainte-nance, and troubleshooting of capillary gas chromato-graphic systems, 3rd edn. Weinheim, New York

4. Schomburg G (1990) Gas chromatography: a practicalcourse. Weinheim, New York

5. Gordon MH (1990) Principles and applications ofgas chromatography in food analysis. E. Horwood,New York

6. O’Keeffe M (2000) Residue analysis in food: principlesand applications. Harwood Academic, Amsterdam

7. Drawert F, Heimann W, Enberger R, Tressl R (1965)Enzymatische Verandrung des naturlichen Apfelaro-mass bei der Aurfarbeitung. Naturwissenschaften52:304

8. Fleming HP, Fore SP, Goldblatt LA (1968) The formationof carbonyl compounds in cucumbers. J Food Sci 33:572

9. Kazeniak SJ, Hall RM (1970) Flavor chemistry of tomatovolatiles. J Food Sci 35:519

10. Leahy MM, Reineccius GA (1984) Comparison of meth-ods for the analysis of volatile compounds from aqueousmodel systems. In: Schreier P (ed) Analysis of volatiles:new methods and their application. DeGruyter, Berlin

11. Mresili R (1997) Techniques for analyzing food aroma.Marcel Dekker, New York

12. Mussinan CJ, Morello MJ (1998) Flavor analysis. Ameri-can Chemical Society, Washington, DC

13. Sapers GM, Panasiuk O, and Talley FB (1973) Flavorquality and stability of potato flakes: effects of rawmaterial and processing. J Food Sci 38:586

14. Seo EW, Joel DL (1980) Pentane production as an indexof rancidity in freeze-dried pork. J Food Sci 45:26

15. Buttery RG, Teranishi R (1963) Measurement of fatoxidation and browning aldehydes in food vapors bydirect injection gas–liquid chromatography. J Agric FoodChem 11:504

16. Buckholz LL, Withycombe DA, Daun H (1980) Applica-tion and characteristics of polymer adsorption methodused to analyze flavor volatiles from peanuts. J AgricFood Chem 28:760

17. Reineccius GA, Keeney PA, Weiseberger W (1972) Fac-tors affecting the concentration of pyrazines in cocoabeans. J Agric Food Chem 20:202

18. Majors RE (1986) Sample preparation for HPLC and GCusing solid-phase extraction. LC-GC 4:972

19. Markel C, Hagen DF, Bunnelle VA (1991) New technolo-gies in solid-phase extraction. LC–GC 9:332

20. Pawliszyn J (1997) Solid phase microextraction: theoryand practice. VCH Publishers, New York

21. Zhang Z, Yang ML, Pawliszyn J (1994) Solid phase-microextraction: a solvent-free alternative for samplepreparation. Anal Chem 66:844A–857A

22. Harmon AD (1997) Solid phase microextraction for theanalysis of flavors. In: Marsili R (ed) Techniques foranalyzing food aroma. Marcel Dekker, New York

23. Coleman WMI (1996) A study of the behavior of Mail-lard reaction products analyzed by solid-phase microex-traction gas chromatography-mass selective detection.J Chromatogr Sci 34:213–218

24. Pfanncoch E, Whitecavage J (2002) Stir bar sorptiveextraction capacity and competition effects. GerstelGlobal, Baltimore, MD, pp 1–8

25. David F, Tienpont B, Sandra P (2003) Stir-bar sorptiveextraction of trace organic compounds from aqueousmatrices. LC-GC Europe 16:410

26. Dupuy HP, Fore SP, Goldbatt LA (1971) Elution and anal-ysis of volatiles in vegetable oils by gas chromatography.J Am Oil Chem Soc 48:876

27. Legendre MG, Fisher GS, Fuller WH, Dupuy HP,Rayner ET (1979) Novel technique for the analysis ofvolatiles in aqueous and nonaqueous systems. J Am OilChem Soc 56:552

28. Widmer HM (1990) Recent developments in instrumen-tal analysis. In: Bessiere Y, Thomas AF (eds) Flavorscience and technology. Wiley, Chichester, p 181

29. Malowicki SMM, Martin R, Qian MC (2008) Volatilecomposition in raspberry cultivars grown in the PacificNorthwest determined by stir bar sorptive extraction-gas chromatography-mass spectrometry. J Agric FoodChem 56:4128–4133

30. Du X, Qian M (2008) Quantification of 2,5-dimethyl-4-hydroxy-3(2H)-furanone using solid-phase extractionand direct microvial insert thermal desorption gaschromatography-mass spectrometry. J Chromatogr A1208:197–201


31. Fang Y, Qian MC (2006) Quantification of selectedaroma-active compounds in Pinot noir wines from dif-ferent grape maturities. J Agric Food Chem 54:8567–8573

32. Fang Y, Qian MC (2005) Sensitive quantification of sulfurcompounds in wine by headspace solid-phase microex-traction technique. J Chromatogr A 1080:177–185

33. Amirav A, Jing H (1995) Pulsed flame photometer detec-tor for gas chromatography. Anal Chem 67:3305–3318

34. Buffington R, Wilson MK (1987) Detectors for gas chro-matography. Hewlett-Packard Corp., Avondale, PA

35. Shellie R, Marriott P (2003) Opportunities for ultra-highresolution analysis of essential oils using comprehen-sive two-dimensional gas chromatography: a review.Flavour Fragrance J 18:179–191

36. Merritt C (1971) Application in flavor research. In:Zlatkis A, Pretorius V (Eds) Preparative gas chromatog-raphy. Wiley-Interscience, New York, pp 235–276

37. Kempfert KD (1989) Evaluation of apparent sensitivityenhancement in GC/FTIR using multidimensional GCtechniques. J Chromatogr Sci 27:63–70

38. Phillips JB, Xu J (1995) Comprehensive multi-dimensional gas chromatography (Review). J Chro-matogr A 703:327–334

39. Shellie R, Mondello L, Marriott P, Dugo G (2002) Char-acterization of lavender essential oils by using gaschromatography-mass spectrometry with correlation oflinear retention indices and comparison with compre-hensive two-dimensional gas chromatography. J Chro-matogr 970(1/2):225–234

40. Adahchour M, Brandt M, Baier H-U, Vreuls RJJ,Batenburg AM, Brinkman UAT (2005) Comprehensivetwo-dimensional gas chromatography coupled to arapid-scanning quadrupole mass spectrometer: princi-ples and applications. J Chromatogr A 1067:245–254

41. Hodges K (1991) Sensory-directed analytical concentra-tion techniques for aroma-flavor characterization andquantification. In: Risch SJ, Hotchkiss JH (eds) Foodpackaging interactions II. American Chemical Society,Washington, DC, p 174

42. Bernreuther A, Epperlein U, Koppenhoefer B (1997) In:Marsili R (ed) Techniques for analyzing food aroma.Marcel Dekker, New York, p 143

43. Woodrow JE, McChesney MM, Seiber JN (1995) Deter-mination of ethylene oxide in spices using headspace gaschromatography. J Agric Food Chem 43:2126

44. Keeney PG, Patton S (1956) The coconut-like flavordefect of milk fat. I. Isolation of the flavor com-pounds from butter oil and its identification as delta-decalactone. J Dairy Sci 39:1104–1113

45. Ryan D, Shellie R, Tranchida P, Casilli A, MondelloL, Marriott P (2004) Analysis of roasted coffee beanvolatiles by using comprehensive two-dimensionalgas chromatography-time-of-flight mass spectrometry.J Chromatogr A 1054:57–65

46. Peterson DG, Reineccius GA (2002) Determination of thearoma impact compounds in heated sweet cream butter.Flavour Fragrance J 18:320–324

Documents

Nielsen2010 FoodAnalysis 4thEdanuragaja.staff.ipb.ac.id/files/2017/05/3_GC.pdf · 2017-05-04 · umn chromatography on silica gel), simple solvent extraction, or some combination