January, 1984 CHROMATOGRAPHY - MASS SPECTROMETRY 13 Chromatography - Mass Spectrometry in the Environmental and Life Sciences The following are summaries of six of the papers presented at a Joint Meeting of the Scottish Region, Chroniatographj. and Electrophoresis Group and Joint Pharmaceutical Analysis Group held on March 1st and b n d , 1983, at the University of Edinburgh. The suniniaries are preceded liy a brief note of other features of the meeting by Prof. C. J. W. Brooks.14 CHROMATOGRAPHY - MASS SPECTROMETRY Anal. Proc., Vol. 21 Meeting Note The three scientific sessions included a total of nine lectures, together with two periods occupied by concurrent discussion seminars. The group of speakers included a distinguished visitor from overseas, Professor Karl Ballschmiter of the Department of Analytical Chemistry, University of Ulm, Federal Republic of Germany.The seminars proved very successful, the liveliest interest being elicited by environmental topics and by considerations of modern techniques in sanipling, purification, chromatography and mass spectrometry. Professor Ballschmiter dealt with sample preparation for capillary GC - MS, and participants appreci- ated the chance to draw upon his unique expertise in the analysis of almost every class of organic pollutant. Dr. J. D. Baty’s seminar dealt with the applications of stable isotope-labelled analogues of drugs as internal standards for quantitative analysis and in the study of metabolic pathways. Environmental analysis by GC - MS was discussed in the seminars conducted by Dr.B. Brookes and Mr. S. J. W. Grigson, while microbore LC - MS was the topic in Dr. D. E. Games’s session. Dr. S. J. Gaskell gave a lucid exposition of the analytical capabilities of the various combinations of scanning modes available in modern mass spectrometry systems. i n a Plenary Lecture, Professor C. J. W. Brooks (University of Glasgow) surveyed the role of GC - &IS in the analysis of bio-organic samples, with emphasis on the practical differentia- tion of closely similar compounds. Examples included (i) the selective analysis of tetra- chlorodibenzo-$-dioxins having two chlorine substituents in each benzo ring, based on efficient GLC and on the ion of m / z 176 [C,H,Cl,O,]-• produced via chemical ionisation using oxygen as reagent gas; (ii) the difficulty of distinguishing between sterols with certain structurally or stereochemically isomeric side-chains ; and (iii) the value of appropriate derivatives, and of chemical ionisation, for the characterisation and sensitive detection of prostaglandins and of tlironiboxane B,.The theme was further exemplified by studies in Glasgow in collaboration with Professor T. D. 1’. Lawrie and Mr. J. MacLachlan (Department of Medical Cardiology) on the identification of sterol epoxides in blood plasma. Mr. S. J. \IT. Grigson (Institute of Offshore Engineering, Heriot-Watt University) surveyed the practice and usefulness of analyses of hydrocarbons in marine sediments, exemplified by determinations of triterpanes, steranes and alkylated arenes as indicators of the occurrence and source of oil pollution.Fast atom bombardment (FAB) mass spectrometry was discussed by Dr. R. Large (hl-Scan, Bucks) ; the value of the technique in structural investigations on large biological molecules was well illustrated with reference to peptides, leukotrienes, bleomycins and fungal elicitors of yolygalacturonic acid types, these last compounds yielding molecular ions above nzjz 2 000. Studies on De-acetylation - Re-acetylation Reactions Using Stable Isotopes J. D. Baty, R. G. Willis and Yoke Khim Koh Depavtmepit of BLoclLewical JIedicLrie, iYbnewells Hospital alzd ,IIedLcal Scliool, Dundee, DD1 9SY Acetanailide (see Fig. 1) is largely metabolised in man and other species to paracetamol (see I;ig. 2 ) , which in turn is escreted into the urine as conjugates.Recent work has shown that in a rat liver niicrosomal system the de-acetylation of acetanilide to produce aniline was the major metabolic reaction.’ These iiz i~itvo data would appear to correlate with the in viilo metabolism of acetanilide only if, after de-acetylation, the aniline was re-acetylated to acetanilide. This would have to be a rapid reaction relative to aromatic oxidation. possible metabolic scheme is shown in Fig. 1 . If acetanilide labelled in the acetyl group with deuterium were to be de-acetylated and re-acetylated prior to or after its conversion to paracetamol, the product of this re-acetylation would contain an NHCOCH, group owing to re-acetylation by acetyl-CoA. Consequently, from an inspection of the mass spectra of the residual acetanilide and the paracetamol produced in the reaction, the extent of de-acetylation followed by re-acetylation can be observed.U’e have investigated this reaction in the rat and in man.January , 1984 CHROMATOGRAPHY - MASS SPECTROMETRY NHCOCH3 (CD3) NHCOCH3 (CD3) I 1 Oxidation 1 De-acetylation Acetylation NHCOCH3 I +(j OH I OH YHCOCHj Acetylation ,f+ OH NHCOCH3 I 15 OH 1 2 Fig. 1. Metabolism of acetanilide to aniline and The deuteriomethyl derivative paracetamol in the rat. was used as a substrate in these experiments. An experiment was carried out with ‘three substrates, trideuterioacetanilide, trideuterio- phenacetin and trideuterioparacetamol (Fig. 2). NHCOCD3 4 Deuterioacetanilide NHCOCD3 NHCOCD3 OCzH5 OH Deuteriophenacetin Deuterioparacetamol Fig.2. Substrates for in vivo de-acetylation experi- ments in the rat and man. Two male Wistar rats were used for each experiment with a different substrate. They were given an oral dose (100 mg kg-l) of the substrate and urine was collected over a 24-h period. The urine was hydrolysed and extracted with ethyl acetate as previously described.2 The two forms of acetanilide and the two forms of paracetamol were monitored by GC - MS as their TMS ethers2 and the aniline produced was measured by HPLC., A male volunteer also ingested 50 mg of trideuterioacetanilide and on a separate occasion 50 mg of trideuterio- phenacetin. Urine was collected for the next 8 h and treated as described above. Results None of the substrates showed any exchange of the label in the phosphate buffer (pH 5.4).Any removal of the NHCOCD, group would thus be due to metabolism. We were able to show that in the rat extensive de-acetylation occurred followed by re-acetylation. Table I shows the results of the experiment with trideuterioacetanilide, and Table I1 shows the amount of re-acetylation that occurred with the three substrates. With phenacetin we were unable to monitor the parent compound owing to its rapid metabolism. The amount of re-acetyla- tion of trideuterioacetanilide differed from the amount occurring in the product of oxidation , i . e . , paracetamol. This suggests that the aniline that is produced is oxidised to p-amino- phenol, which is then acetylated to paracetamol. This additional source of re-acetylation would explain the difference in the results between the two compounds.16 CHROMATOGRAPHY - MASS SPECTROMETRY TABLE I Anal.PYOC., Vol. 21 I N V I vo METABOLISM OF TRIDEUTERIOACETANILIDE IN THE RAT Amounts are reported as pmoles in a 24-h urine sample. [ZH,]Acetanilide Acetanilide [ZH,]Paracetamol Paracetamol Aniline Rat 1 .. 1.5 0.6 64.1 37.6 12.7 Rat 2 .. I .4 0.7 57.4 43.1 13.6 In man we observed only a small percentage exchange of acetyl groups (approximately 10%). It may be important to monitor this reaction in people who are fast or slow acetylators of drugs, as the accumulation of amines resulting from de-acetylation reactions may be potentially hazardous. TABLE I1 EXTENT OF DE-ACETYLATION/RE-ACETYLATION FOUND IN THREE DEUTERIATED SUBSTRATES IN THE RAT Re-acetylated paracetamol, y-, Substrate re-acetylation, % f A j r A 7 Substrate Rat 1 Rat 2 Rat 1 Rat 2 [2H,]Acetanilide .. . . .. 37.0 42.9 28.6 33.3 [2H3]Phenacetin . . . . .. 33.7 32.8 - - [ H, ] Parace tam01 .. .. 6.4 8.9 - - References 1. 2. 3. Carlson, G. P., Dziezak, J . D., and Johnson, K. M., Res. Commun. Chem. Pathol. Pharmacol., 1979,25, Baty, J . D., and Robinson, P. R., Biomed. Mass Spectrom., 1976, 3, 60. Sternson, L. A., and Dewitte, W. J . , J . Chromatogr., 1977, 137, 305. 181. Complementary Role of Mass Spectrometry in the Chromatographic Analysis of Environmental Samples K. Ballschmiter Department of Analytical Chemistvy, University of Ulm, Ulm-Donau, Federal Republic of Germany Environmental science deals with multi-matrix and multi-component problems and mostly with concentrations of the looked-for components in the microgram/gram or nanogram/gram range.Outstanding examples of organic trace analysis with environmental samples are the determinations of polychlorobiphenyls (PCBs) or polychlorodibenzodioxins (PCDDs) in a broad spectrum of matrices; 209 single PCB components are possible and about 100 can be found in technical formulations and in the envir0nment.l Seventy-five polychlorodibenzo- dioxins are possible and four are of particular interest owing to their extreme toxicity. Environmental analysis is further complicated by abiotic and biotic degradation in the environment. However specific a determination can be made in mass spectrometry, the basis of any analytical scheme for environmental samples is a set of separation steps.2 The depletion of the matrix alone requires a well designed sequence of separations.The combination of a highly effective mode of separation, especially high-resolution gas chromatography, with a highly specific mode of detection makes the combination of gas chromatography (GC) with mass spectrometry (MS) extremely appealing for environmental analysis. The availability of open-tubular columns of fused silica has removed the psycho- logical barriers often connected with capillary gas chromatography and has made possible theJanuary, 1984 CHROMATOGRAPHY - MASS SPECTROMETRY 17 optimum interfacing of the gas chromatograph with the mass spectrometer. Either the direct inlet into the source or the so called “open coupling,” where only a defined and constant portion (50-80~0) of the GC effluent is brought into the ion source, are standard designs.The “open coupling” has the possibility of cutting off the solvent peak, which can be a major advantage when injection volumes of 2-5 pl are used to obtain an effective “solvent effect” i n j e c t i ~ n . ~ , ~ Considering the mass spectrometer only as a detector clarifies its operational needs- simplicity in operation and servicability, stability in performance and sensitivity in creating the analytical signal-with respect to the ions to be detected. By early 1983 only two major instrument companies had followed this approach. The type of ion source in the mass spectrometer governs the type of analytical signal with which one can operate, Either electron impact or chemical or negative ion chemical ionisa- tion, owing to the variations in fragmentation pattern and ion yields, gives the basis of the analytical signal.In terms of a GC detector, any ionisation process that gives only a few but typical fragments and high yield is to be preferred; it will give both sensitivity and specificity. Specificity of the detection can be sharply increased by using high-resolution mass spectro- metry, but with a consequent trade-off of sensitivity and stability under the operation conditions, coupled with the high cost of the instrumentation. In most instances it is more reasonable to concentrate the analytical efforts on the chromatographic part of a GC - MS combination than on the resolution of the MS part. Whether a full-scan mode (60-600 a.m.u.) or a multiple-ion mode (4-10 ions) is used in GC - MS depends on the analytical problem.Any structure elucidation of unknowns will require the former mode and any detection or search for expected compounds is best carried out by the latter mode. In either instance one should bear in niind that a peak of 2 s width requires a scan time or sum of dwell times of 0.6 s to ensure reasonable integration. Both modes of )IS operation give three basic types of information: (1) retention time; (2) data regarding the molecular structure; and (3) amount. The first type of information is still often overlooked, although a transformation to retention index can standardise the retention time and add further structural aspects to the analytical information. Quantitation by GC - MS should always operate on the basis of keeping all parameters constant, because only in this way can the inherent systematic errors be regarded as similar in the calibration and the analytical runs.One should keep in mind that any fluctuation in the ion yield can strongly distort the quanti- t a t i ~ n . ~ The use of internal standards can improve the quantitation considerably, or even make it possible. Sensitivity in GC - MS, in addition to requiring optimised functioning of the ion source and the ion separation, includes primarily the structural parameters of the molecules to be investig- ated. Strong molecular ions or fragments such as are obtained with many aromatics clearly give optimum sensitivity. Even with an electron-impact source 1-5 pg of an aromatic com- pound can be detected in the multiple-ion mode.In the scan mode it will require 5-10 ng to give a useful mass spectrum. Negative ion chemical ionisation has detection limits well into the femtogram range, if one uses the right molecules for this demonstration. However, its specificity makes it an even more distorting detector than the electron-capture detector, and it strongly overestimates highly chlorinated compounds. As negative ion chemical ionisation includes chemistry in the gas phase, this can be used to create a structurally defined specificity. One drawback of GC - JIS is often overlooked, namely that molecules with undefined frag- mentation patterns and/or low ion yields are nearly invisible in the GC - MS system. A typical example is the polychlorinated terpene Tosaphene, when using an electron-impact source.One should always bear in mind when using a specific detector such as a mass spectro- meter that it only answers the questions being asked. A more general detector such as the flame-ionisation detector (FID) or even the electron-capture detector (ECD) should be run in parallel in environmental science. The ECD is in general about 50 times more sensitive than an electron-impact source and a corresponding dilution or concentration of the samples has to be carried out. Operating a gas chromatograph with either an FID or ECD in parallel with the GC - JIS combination under identical chromatographic conditions can be extremely helpful with environmental samples. Matrix depletion is a multi-functional separation scheme for environmental samples, and consequently one should push the resolution of the chromatography prior to the detection to18 CHROMATOGRAPHY - MASS SPECTROMETRY Anal.Proc., Vol. 21 its highest possible level. In addition to being sound analytical strategy, it helps to optimise the specificity of the detector. However, both the structural information and the information obtained by the gas chromatograph should be used. The MS detector is an extremely helpful positive identification device in environmental science, whereas most other detectors merely detect the analogy between an unknown and a reference standard. References 1. 2. Zell, M., and Ballschmiter, K., Fresrnius 2. Anal. Chem., 1980, 304, 337. Rallschmiter. K., in Nunisto, L., Editor, “Euroanalysis IV, Reviews on Analytical Chemistry,” AkadCmiai Kiad6, Budapest, and Association of Finnish Chemical Societies, Helsinki, 1982, pp.Grob, I<., and Grob, I<., Jr., J . High Resolut. Chromatogr., Chromatogr. Commun., 1978, 1, 57 and 263. Jennings, W. G., Freeman, It. K., and Roney, T. A., J . High Resolut. Chromatogr. Chromatogr. Schafer, W., and Ballschmiter, I<., Frcsrnius 2. Anal. Chrm., 1983, 315, 475. 139-156. 3. 4. 5. Commun., 1978, 1, 215. Techniques for Improved Quantitative Analysis in Environmental Trace Analysis Using Capillary Gas Chromatography - Mass Spectrometry D. E. Wells DA FS, Freshwater Fisheries Laboratory, Pitlochry Over the past 3 years there has been a considerable growth in the number of reports describing the use of capillary columns, particularly fused-silica columns, for the quantitative GC - MS multi-component analysis of contaminants in environmental samples.’ This has made the use of internal standards (IS) mandatory in any serious quantitative determination.Although this is reflected in most branches of medical science (GC - MS Abstracts, 1980-82; in the Clini- cal Chemistry section 73% of the applications used ISs), the same does not appear to be so for environmental science (GC - MS Abstracts, 1980-82; 10% of these applications use 1%). In keeping with medical science some methods have employed isotopic analogues,192 particularly for tetrachlorodibenzo-P- dioxin (TCDD) and polycyclic aromatic hydrocarbons (PAH) determinations, but clearly this cannot be seriously contemplated on financial or logistical grounds for the extensive multi- coniyonent analyses for both pollutants and metabolites in the environmental field.As one of the main sources of error in quantitative capillary GC - MS, apart from tlie initial sampling, occurs during sample introduction, it is often sufficient to add tlie IS just prior to injection. This laboratory uses GC - MS, GC with electron-capture detection (ECD)3 and high-perform- ance liquid chromatography (HPLC) with ultraviolet (UV)4 and fluorescence detection and samples are often cross-checked by a second method. I t was highly desirable to select ISs that were detectable with as many of these systems as possible, as well as fulfilling the usual criteria of purity and stability.The IS was therefore required (i) to have a similar retention time (R,) but be resolved from its analyte(s), (if) to have a limited number of mass spectral fragments, preferably in/z >120 (above the JIS background) and (iii) to be electron capturing or UV absorbing (and/or to fluoresce). The method of sample introduction was also investig- ated and tlie discrimination experienced in the injection of compounds of differing polarity and volatility was considerably improved by replacing the splitless injector with an on-column injector (OCI). The instruinentation used was a Finnigan 9500 GC, fitted with a split/splitless Grob-type injector, interfaced to a Finnigan 320OF 11s and 6100 Data System. The GC column was 25 m ~ 0 . 2 5 inn1 i.d. fused-silica CP Sil5 (Chroinpak UK, London) and installed from the injector through the interface oven directly into the ion source.The split/splitless injector was used in the splitless mode with 20 ml min-l of helium as the purge gas. The split valve was closed, the sample (1 pl) injected at 260 O C , the injector vented after 45 s and the programme coninienced after 1 niin. An OCI-2 on-column injector [SGE (UK), Milton Keynes] was retrofitted to the Finnigan One of the main problems in this field is the choice of IS.January, 1984 CHROMATOGRAPHY - MASS SPECTROMETRY 19 9500 GC oven top, between the end of the interface oven and the GC oven edge using the manufacturer's recommendations. The sample was introduced into the OCI at ambient temperature, using a syringe with a 0.17 mm i.d. fused-silica needle (SGE) through the pneumatic seal.Following insertion and sealing of the syringe needle the carrier gas pressure was permitted to re-establish (10-20 s) before the sample was injected. The injection time was (1 s for 1-p1 samples and 2-3 s for 2-5-pl samples. On-column Injection Following installation of the OCI system, it was tested with a wide variety of mixtures and found to be superior to the splitless system in sensitivity, sample discrimination and repro- ducibility, particularly for relatively less volatile compounds such as some aromatic amines and sulphonamides. Many of these points have been adequately covered by other w o k e r ~ ~ ~ ~ so that this report is confined to additional information only. Grob6 has reported that up to 8 p1 can be injected over a period of 20 s without harmful effects to the column or to the detriment of the chromatography, particularly for chemically bonded phases. Although this holds true for the GC alone, the upper limit on solvent volume for GC - MS, when the column is interfaced directly into the ion source, is about 5 p1 (the more efficient turbomolecular pumps on newer instruments will handle a greater solvent volume).A summary of results obtained for two more difficult analytes is given in Table I. Ter- butryn, a triazine-based aquatic herbicide, and polychlorinated aminodiphenyl ethers (PADs) were both difficult to quantify using the splitless injector, but showed a marked improvement when introduced via the OCI. 9-Bromoanthracene was used as the internal standard for the determination of terbutryn and decachlorobiphenyl (DCBP) for the PADs.TABLE I COMPARISON OF Determinand Terbutryn .. De-ethylterbutryn PADs . . .. DCBP .. . I INJECTION PRECISION USING SPLITLESS AND ON-COLUMN INJECTORS Coefficient of variation (95%) t 7 OCI Split less Internal NO. of , - - * - - I ,-A-$ mlz standard m/z runs RSN* KPAt RSN* RPAt . . 185 9-Bromoanthracene 256 7 2.3 3.9 1.1 16.6 .. 156 256 7 2.5 5.9 2.6 . . 321 DCBP 500 7 1.6 6.1 2.63 20.7 . . 498 DCBP 500 7 - 2.0 - 6.8 * RSN = relative scan number. t RPA = relative peak area. Following Grab's' detailed descriptions of the solvent effects immediately following injection, most workers have chosen to inject their samples in pentane or hexane around ambient temperatures, which is particularly suitable for the analysis of determinands with a wide boil- ing range or for volatile compounds. However, many pesticide residues and industrial pollut- ants are less volatile, eluting from most columns at higher temperatures (>150 "C).For these analyses a successful alternative approach has been to select both the starting tempera- ture and programming rate as determined by the separation requirements of the analytes. A solvent of the appropriate boiling range, within 20 "C of the initial temperature, is selected (Table 11). This approach has been particularly successful when determining less volatile pollutants using multiple-ion monitoring (MIM) , giving fast analysis times and good precision and separation efficiency. There is, however, one particular disadvantage using high-boiling (>lo0 "C) carriers with direct interface capillary GC - MS, where the solvent is not dumped at the interface stage.After a small number of injections, traces of solvent remain on the inner surfaces of the analyser, increasing the background spectra as the samples are injected. This can be troublesome when using decane if a full mass spectrum, extending below 120 a.m.u., is required or if the20 CHROMATOGRAPHY - MASS SPECTROMETRY TABLE I1 Anal. Proc., Vol. 21 SOLVENT SELECTION FOR DIFFERENT INITIAL GC OVEN TEMPERATURES Final sample solvent B.p./"C Pentane . . . . 36 Hexane . . . . . . 69 Isooctane . . . . 99 Octane . . . . . . 126 Decane .. . . 174 GC oven temperature Injection range/"C at injectionlac 30-50 50 65-80 80 95-1 10 100 120-135 130 170-1 85 180 ions monitored in the MIM mode are coincident with the solvent spectrum. On such occasions a lower boiling solvent such as octane was used.This pumped away rapidly and allowed the normal background subtraction techniques to obtain the spectrum of the eluting chrom- atographic peaks. Choice of Internal Standards During the past 2 years a number of ISs have been incorporated into analytical schemes used at this laboratory.* These now cover the analysis of a large number of persistent pesticide residues and non-volatile industrial pollutants. Most of these compounds are readily available, and to date none has been detected in samples taken from Scottish freshwater, coastal and marine environments. The ISs are listed in order of elution (Table 111) obtained on a 25 m x 0.25 mm i.d.CP Sil fused-silica column and cover the temperature range 80-270 "C. The bromohydrocarbons were chosen to cover the lower temperature range as the lower chlorinated aromatics tend to be more widespread at trace levels in the environment. With the exception of mirex, meth- oxychlor and dechlorane, they all have an aromatic nucleus and an intense molecular ion. The halohydrocarbons also give a choice of ions in the molecular cluster at M + 2, M + 4, etc., depending on the number of halogen atoms, which can be particularly useful when it is necessary to avoid a more intense background ion or co-eluting peak. Of the 15 ISs selected, 10 are suitable for GC - ECD analysis, 13 are UV absorbing and 8 fluoresce.TABLE I11 INTERNAL STANDARDS FOR USE WITH GC - MS, GC - ECD AND HPLC ANALYSIS OF PESTICIDE AND INDUSTRIAL POLLUTANT RESIDUES Internal standard [2H,]Naphthalene . . . . o-Dibromobenzene . . . . 2-Bromonaphthalene . . . . 4-Bromodiphenyl ether . . . . [2H,,]Phenanthracene . . . . 2-Bromofluorene . . . . . . 1,2,3,4-Tetrachloronaphthalene [2H,,]Pyrene . . . . . . 9-Bromoanthracene . . . . [2H,,]Chrysene . . . . . . Methoxychlor . . . . . . [2H12]Perylene . . . . . . Decachlorobiphenyl . . . . Dechlorane . . . . . . Mirex . . . . . . . . Masses selected, lnlz 136 236,238 206,208 248,250 188 244,246 264,266 212 256,258 238 2277 272,2747 264 496,498 261,263 Elution order on CP Sil5,* k GC - ECD 3.49 - 3.98 + 10.34 + 18.72 + 21.31 - 26.09 + 28.43 + 32.85 - 33.33 + 44.07 - 45.99 + 47.56 + 55.23 - 57.31 + + - HPLC UV + + + + + + + + + + + + + - - HPLC - + + + + + + + fluorescence - - - - - + - - * Injected a t 100 OC, then programmed at 2 0C min-'.t Base peak is not the molecular ion. Multiple Internal Standards Most analyses reported to date involving ISs have used a single compound, the response of which is related to all eluants in the chromatograph. This has certainly been sufficiently accurate for most isothermal analyses on packed columns and short run times (about 10 min)January, 1984 CHROMATOGRAPHY - MASS SPECTROMETRY 21 using capillary chromatography. However, for multi-component analyses normally associ- ated with capillary columns one IS does not always cover the whole chromatographic range with similar precision.This has been reported by Sauter et aL9 in the analysis of priority pollutants and is further confirmed in this work. A study was completed for a mixture of mothproofing agents currently being determined in this laboratory. In recent years dieldrin has been replaced by Eulan WA New, which has a series of polychlorinated sulphonamidodiphenyl ethers (PCSDs) as active ingredients and a corresponding amine (PAD) as an impurity and as a primary metab~lite.~ Subsequently a series of cis,trans-permethrin-based mothproofers were also marketed. As effluents containing these products were discharged to the same catchment it was desirable to quantify most of these residues in a single chromatogram. Dieldrin, PADS and perrnethrin occur in the same clean-up eluate and were analysed together with TCN, mirex and DCBP as ISs.The results of the reproducibility at two series of concentrations are given in Table IV. The solvent used was decane with 1 pl injected at a column temperature of 80 O C , held for 1 min, followed by programming at 2 “C min-l. The peak areas and scan times were measured and ratioed against each of the three internal standards, in turn. Both the relative scans (relative retention time) and the peak areas exhibit the same trends. The precision of the measure- ment declines as the analyte and IS are separated by an increasing time margin. TABLE IV REPRODUCIBILITY OF PEAK SCAN NUMBER AND AREA FOR MOTHPROOFING AGENTS WITH DIFFERENT INTERNAL STANDARDS Coefficient of variation (95%) Mass on col u nin / Compounds in elution order ng inlz Tetrachloronaphthalcne (TCS) .. 0.1 266 Dieldrin . . . . . . . . 2.0 277 Mirex . . .. .. . . 0.5 272 PADS . . . . .. . . 2.0 321 I’erme thrin . . . . . . 0.2 183 Decachlorobiphenyl (DCBP) . . 2.0 496 0.5 10.0 2.5 10 1 .0 10 r - ’lC9 T---A-- 7 n I<SK* IWAt 7 IS IS 7 IS IS 7 2.9 10.5 7 5.6 7 4.0 9.4 7 3.9 7 4.2 17.8 7 6.8 7 2.0 28.6 7 10.1 7 4.0 12.5 7 13.3 Mirex -A-7 I<SN* l<J?At 4.1 9.6 5.4 0.8 14.0 7.6 IS IS IS IS 0.4 22.0 8.9 1.0 37.8 12.3 1.5 12.8 15.3 DCBP +--7 I<SN* RPAt 4.1 14.8 11.4 1.0 13.2 19.5 1.4 12.6 13.6 1.6 13.2 6.1 0.6 32.4 8.8 IS IS IS IS * I<SX =: relative scan number. t HSA : relative peak area. Conclusions Following the retrofit of the on-column injector to the GC - MS system, superior quantitative results were obtained compared with the splitless injector, particularly for less volatile, non- polar herbicides and pesticides.The final sample solvent can be selected on the basis of the desired chromatography, giving shorter analysis times and minimum delay between injections. The precisions in peak retention timesand areas are markedly improved by the use of multiple internal standards tliat elute within a 10-min time window of the analyte. Many of the internal standards selected can be used with GC - MS, GC - ECD and HPLC with C\’ and fluorescence detection. References 1 .. 131-ooks, C. J . I\‘., liditor, Gas C‘hromatogvafihy - .Wuss Spectvometvy Abstracts, 1980-82, Volumes 11-13, P I t M Science arid Technology Agency, London. 2. Sorlvicki, H. G., I)c\.inc, I ( .I:., ant1 liietla, C. .I., i n \-ail Hall, C. I:., Iiditov, “Mcasurcmcnt of Organic I’ollutants in \\’atcr :inti \\-astc \\‘atcr,” .ISl.lI S‘l1’ 686, I\mcrican Society for Testing and Mater- ials, T’hilatlclphia, lYi!), pp. 1.70-1.51.22 CHROMATOGRAPHY - MASS SPECTROMETRY Anal. Proc., Vol. 21 3. 4. 5. 6. 7. 8. 9. Wells, D. E., and Cowan, A. A, Analyst, 1981, 106, 862. Wells, D. E., and Johnstone, S. J., J . Chvomatogr. Sci., 19'81, 137. Watanabe, C., Tomita, H., Sato, K., Masada, Y., and Hashimoto, K., J . High. Resolut. Chromatgr. Chromatgr. Commun., 1982, 5 , 630. Grob, K., J . HigJt Resolut. Chromatgr. Chromatgr. Commain., 1978, 1, 263. Grob, I<., J . High Resolut. CJwomatgr. Chrotnatgr. Commun., 1978, 1, 57. Wells, D. E., Anal. Pvoc., 1980, 17, 116.Sautcr, A. D., Betowski, L. D., Smith, T. R., Strickler, \'. X., Beimer, K. G., Colly, B. N., and Wilkin- son, J . E., J . HigJz Resolttt. Clivonzatgr. Chvotnatgr. Commun., 1981, 4, 366. Sampling and Gas Chromatography - Mass Spectrometry Analysis of Polar Volatile Organic Pollutants B. 1. Brookes Regional Chemist's Departwient, StvatJiclyde Regional Council, 8 Elliot Place, Glasgow. G3 8E J Water-insoluble Compounds Determinations for volatile water-insoluble compounds (benzene, chloroform, decane, etc.) can be achieved routinely a t parts per lo9 concentrations in both air and water. - I n the simplest methods for air analysis, the air is drawn through a tube containing Tenax GC,I-3 Porapak Q1*4 or graphitised carbon black (GCB) ,5 and, subsequently, the adsorbed compounds are transferred on to the gas chromatography (GC) column by thermal desorption.More powerful adsorbents such as activated charcoal6 are also used but require greater manual dexterity as milligram amounts of the adsorbent and microlitre volumes of extraction solvent must be employed if high recovery efficiencies and high concentration factors are to be obtained. The total cryogenic sampler of Penkett et aL7 is the most elegant technique for the very volatile compounds. I t ensures lOOq/, sampling efficiency and has the greatest freedom from artefacts . Adsorption tube techniques have also been applied to water analysis by bubbling a pure, inert gas through the sample and trapping the vapours in a sample tube.8~~ Water-soluble Compounds Impressive results are obtained for water insolubles with comparative ease and chemists have been tempted to use these same techniques for the water-soluble compounds such as methanol, methylamine and formic acid.However, problems ensue: GCB and Tenax GC, in keeping with their low retention of water vapour, show poor retention of water-soluble com- p o u n d ~ . ~ * ~ Porapak Q, a more powerful adsorbent, can be used for acids and neutrals,'JO although the retention volumes are st+H much less than those of the water-insoluble compounds of similar volatility. Activated charcoal is not generally recommended for water solubles, and the total cryogenic sampler, as developed so far, is unsuitable as the water solubles remain in the water film that has condensed on the walls of the sample vessel.Attempts have been made to use the gas-phase stripping technique for water-soluble compounds in water," but the inherent difficulties of this approach have not been effectively overcome. Vt'hereas the water insolubles can be analysed on an)' non-polar or semi-polar capillary column, different columns are necessary for the different types of water-soluble compounds. Also, as any device suitable for sampling wa ter-soluble compounds will also sample some water, the GC column must be chosen to withstand repeated injections of water. It is therefore a mistake to use the procedures for water insolubles as a starting place for the development of methods for the water solubles. Instead, consideration should first be given to the chemical properties of the compounds.-4 good example of where this has been done is the method of Peters12 for the concentration of volatile water-soluble compounds in water. The sample is heated in a flask and the vapours rise up a short fractionating column to a micro-Soxhlet device built inside a condenser. The system is allowed to come to equilibrium, when 800, of the compound will have been transferred into the Soshlet trap and a concentra- tion factor of sel-era1 liundred achieved. As an alternative to the method of Peters, acids can be concentrated in water samples by There are also problems with the chromatography of water-soluble compounds.January, 1984 CHROMATOGRAPHY - MASS SPECTROMETRY 23 rendering 10 ml of the sample 0.01 N alkaline and reducing the volume by a factor of 100 to 0.1 ml.Bases can be similarly concentrated after rendering the sample 0.01 N a~idic.1~ The equivalent technique for air samples ought to be the condensate sampler, but it can suffer from poor trapping efficiency and consequently other, less general techniques are also employed. Table I gives a summary of the air sampling methods and the analytical procedures. Appropriate combinations of these methods permit most volatile compounds to be analysed in almost any atmosphere or water. Very reactive compounds are an exception as they must be converted into stable derivatives by the sampling technique. The less volatile water- soluble compounds, such as nicotine, also present problems. They can be adsorbed on fine particles in air1* and a specialised sampling device must be used.TABLE I ANALYTICAL TECHNIQUES AND AIR SAMPLING PROCEDURES Limitations of Compound type GC column Air sampling method sampling method* Water insolubles Non-polar' or Tenax G P 3 V semi-bolar capillary Water-soluble neutrals Polar capillary or Chromosorb loll Acids Bases Capillaryls or Chromosorb 101' Triton X-100 - KOHL Chromosorb 102 - KOH17 or KOH-modified capillary'* Activated charcoals GCB5 Porapak Q4 Total cryogenic7 Porapak Ql Silica geP5 Condensate' Porapak Q1 Condensate Alkali bubbler'5 Condensatel Silica gel'5*'* Acid bu bbler13 9 l7 v, w v, w V LV W MA W * The letters indicate that the sampling method is unsuitable for certain compounds or atmospheres as follows: V = very volatile compounds, e.g., methane; LV = less volatile compounds, e.g., naphthalene; W = very wet atmospheres; MA = atmospheres rich in mineral acids such as sulphur dioxide.It is usually preferable to select an analytical technique that permits full scan monitoring, as it provides the most comprehensive data. However, selective ion monitoring (SIM) has the advantage of greater quantitative accuracy and the ability to distinguish co-eluting compounds. The packed columns in Table I may sometimes produce poor results if water- insoluble compounds are co-eluting with water sohbles. This particular problem can be overcome, without resort to capillary columns or SIM, by using a sampling procedure which distinguishes the two types of compound. Fig. 1 shows two analyses for the minor organic components in landfill gas using a Chromosorb 101 column.The first chromatogram was obtained from a Porapak Q sample and most of the alcohols have their analyses impaired by fluorocarbons and hydrocarbons that are also present. In the second chromatogram the interferences have been eliminated by analysing only the aqueous phase of a condensate sample of the same gas. Whatever type of column is used it is generally advantageous to operate at as high a temperature as possible to avoid condensation of water at the head of the column.' With mass spectrometric detection it may also be necessary to restrict the end of the column, or use make-up gas. Derivatisation should be avoided for the analysis of volatiles in environmental samples. This saves the analyst from having to make assumptions about the type of compound he is looking for ; environmental samples are sufficiently complex without the unnecessary addition of reactive chemicals.The analytical methods in Table I give linear calibration graphs for injections in the range 1-100 ng. Water-soluble compounds can therefore be measured by GC - MS at parts per lo9 levels in both air and water. The techniques are simple, but attention must be given to the choice of sampling or concentration method that will suit both the type of compound to be measured and the analytical procedure. For atmospheric samples it must also suit the type of atmosphere.24 CHROMATOGRAPHY - MASS SPECTROMETRY Anal. Proc., Vol. 21 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. (6) w c- i, a I 50 100 150 Temperature/’C Fig.1. Minor components of a landfill gas analysed on a 1 m x 2 mm i.d. 80-100 mesh, Chromosorb 101 column, monitoring m/z 22-220. (a) Porapak Q sample; and (b) aqueous phase of a condensate sample. Peaks: B, = butan-1-01; B, = butan-2-01; D = diethyl ether; E = ethanol; P, = propan-1-01; Pz = propan-2-01; and W = water. References Brookes, B. I., and Young, P. J., Talanta, in the press. Brown, R. H.. and Purnell, C. J., J. Chromatogr., 1979, 178, 79. Pellizzari, E. D., US Environmental Protection Agency, Report Number EPA-600/2-79-057, National Roe, A. B., J . Inst. Water Eng. Sci., 1982, 34, 118. Ciccioli, P., Bertoni, G., Brancaleoni, E., Fratarcangeli, R., and Bruner, F., J. Chromatogr., 1976, 126, Grob, K., and Grob, G., J. Chrornatogr., 1971, 62, 1. Penkett, S.A., Prosser, N. J. D., Rasmussen, R. A., and Khalil, M. A. K., J. Geophys. Res., 1981, 86, Fed. Reg., 1979, 44, No. 233, 69532. Grob, K., Grob, K., Jr., and Grob, G., J. Chrornatogr., 1975, 106, 299. Brookes, B. I., Jickells, S. M., and Nicolson, R. S., J. Assoc. Public Anal., 1978, 16, 101. Ramstad, T., and Nestrick, T. J., Water Res., 1981, 15, 375. Peters, T. L., Anal. Chem., 1980, 52, 211. Mosier, A. R., Andre, C. E., and Viets, F. G.. Jr., Environ. Sci. Technol., 1973, 7 , 642. Brookes, B. I., and Forbes, G. F., Environ. Health Scotl., 1981, 11, No. 4, 12. “NIOSH Manual of Analytical Methods,” US Government Printing Office, Washington, DC, 1983. Hrivnac, M., Sykora-Cechova, L., and Muller-Aerne, M., J. High Resolut. Chromatogr. Chromatogr. Kuwata, K., Yamazaki, Y., and Uebori, M., Anal.Chem., 1980, 52, 1980. Becher, G., J. Chromatogr., 1981, 211, 103. Technical Information Service, US Dept. of Commerce, Springfield, VA, 1979. 757. 5172. Commun., 1981, 4, 323. Microbore Liquid Chromatography - Mass Spectrometry and its Analytical Potential David E. Games, Nicholas J. Alcock and Mark A. McDowaII Department of Chemistry, University College, P.O. Box 78, Cardiff, CF1 1XL Microbore Liquid Chromatography There is considerable current interest in the use of smaller bore columns for high-performance liquid chromatography (LC).1-3 The types of column in use range from packed microbore columns of 0.25-2 mm i.d.4-12 to o p e n - t u b ~ l a r ~ ~ - ~ ~ and packed capillary column^^^^^^ with inner diameters less than 100 pm.Use of smaller bore columns requires modification of the injection, pumping and detection systems used for conventional LC. In addition, specialised techniques are required for the production of suitable columns. These problems are much less formidable if microbore columns are used and suitable injectors, pumps, columns and detectorsJanuar-y , 1984 CHROMATOGRAPHY - ,MASS SPECTKOMETRY 25 are commercially available. The requirements for columns of this type are that the injection volumes should be 1 pl or less, the pumps should be capable of operating down to 10 p1 min-1 and the cell volume of the detector should be of the order of 1 pl or less. Currently the best columns appear to be made from glass-lined stainless stee16s7 or but PTFE10 and fused-silica11J2 columns have also been used.Packing technology has not yet been fully developed and most columns are packed with 10-pm material. A major advantage in using microbore LC is the considerable reduction in solvent costs that can be obtained, e.g., use of a 250 x 1 mm i.d. column with a flow-rate of 50 p1 min-l results in a 9504 reduction in solvent consumption compared with a 250 x 4.5 mm i.d. column with a flow-rate of 1 ml min-l. This means that the use of unusual, expensive solvents can be contemplated for difficult separations, e g . , optically active or deuteriated20 solvents. Other advantages include the obtaining of high plate numbers by connecting columns in series, which can be advantageous when complex mixtures are being studied. The small peak volumes obtained in microbore LC result in increased mass sensitivity with concentration-sensitive detectors.Thus, for the same amount of sample injected on to both a 4.6 and a 1 mm i.d. column, the detector response for the peak eluting from the microbore column is of the order of 20 times greater than that eluting from the conventional column. Finally, because of the lower flow-rates used with microbore columns, coupling of a liquid chromatograph with a mass spectrometer can be undertaken without the necessity to split off some of the column eluent which is necessary with some types of interface. A problem with microbore LC is that there is a decrease in the volume and amount of sample (of the order o f 1 pl and 10 pg, respectively) which can be injected on to the column without affecting chromatographic performance. The first problem can be overcome by the use of pre-column concentration techniques2 and the latter by use of column-switching techniques.21 Combined Liquid Chromatography - Mass Spectrometry (LC - MS) ,4 variety of methods for combining a liquicl chromatograph with a mass spectrometer have Currently three approaches been developed and the area has been extensively ~-eviewed.~~-~~ appear to have particular merit, as follows.Direct Liquid Introduction With this approach a portion of the effluent from the liquid chromatograph is fed into the mass spectrometer ion source where the solvent acts as a chemical ionisation (CI) reagent gas for ionising the solute molecules. Early systems suffered from problems due to blocking of the capillary inlet when low-volatility compounds were being examined.These problems have been 0verconie~8-~~ by use of cooled probes with a membrane having a hole of 1-5 pm diameter, which enables efficient nebulisation of the LC eluent to occur. A second important feature is the incorporation of a desolvation chamber and finally cryogenic pumping is of considerable assistance in handling some polar mobile phases. A potential ability to handle mass spectrometrically very difficult molecules has been shown with the recent obtaining of a negative CI spectrum of vitamin B12.32 It should be noted, however, that the obtaining of on-line LC - )IS data is not as easy as obtaining spectra of samples just introduced in solu- tion. There are many impressive examples of the use of this type of interface in the litera- t ~ r e .~ ~ The main drawbacks are that only a portion of the LC effluent can be handled if conventional LC columns are used, solvents have to be carefully purified and often solvent systems developed for a particular separation cannot be used because they fail to give good CI spectra (this can be a particular problem when gradient elution is used), finally, only CI mass spectral information is obtained. The latter is not necessarily a problem for many types of stud)., C.R. , quantitative studies and additional structural information can be obtained by use of collision induced dissociation with double or triple quadrupole instrument^.^**^^ On the positi\re sick is the lack of thermal decomposition of thermally labile compounds at low levels, wliich has been demonstrated with commercial interfaces of this type.Moving Belt Systems \\'it11 interfaces of this type the effluent is fed on to a continuously moving belt made of Kapton. Solvent is remo\.ed by a heater and in two vacuum locks and the solute is flash vaporised into the mass spectrometer ion source where either electron impact (EI) or CI26 CHROMATOGRAPHY - MASS SPECTROMETRY Anal. Proc., Vol. 22 spectra can be obtained. Three commercial systems are available, two of which enter directly into the mass spectrometer ion ~ o u r c e , ~ ~ ~ ~ ~ and the other is interfaced to the ion source block.38 Two of the system^^^.^^ exhibit very similar behaviour in terms of their ability to handle com- pounds and are suitable for handling compounds that are amenable to direct insertion probe mass spectrometric study.However, problems are encountered owing to thermal decom- position of thermally labile compounds at low levels. The third system3’ is superior in its ability to handle compounds and enables the range to be extended to those compounds which can only be handled by desorption chemical i o n i ~ a t i o n . ~ ~ Systems of this type exhibit good retention of chromatographic integrity and their ability to provide both E I and CI data has proved to be very effective in the identification of known and unknown compounds. They have been extensively used for the solution of a wide range of problems and their utility and problems encountered in their use have recently been reviewed.40 Initial problems in handling aqueous reversed mobile phase systems have now been largely overcome and with recent improvements in detection levels using microbore LC41 this type of interface is ideal for qualita- tive studies of compounds which fall within the ionisation range mentioned earlier.Thermospray Ionisation and Liquid Ion Evaporation As a result of studies of LC - MS using molecular beam techniques a new method for the ionisation of low-volatility compounds which can be introduced into the mass spectrometer in ionic form has been developed.42 The liquid chromatographic effluent passes through a 0.015 mm i.d. stainless-steel tube which is embedded in an electrically heated copper block, a supersonic jet of vapour is produced which traverses the ion source of a mass spectrometer and enters a 1 cm diameter pumping line which is connected to a 300 ml min-1 mechanical vacuum pump.A conical ion exit aperture is connected to the ion source and a high-capacity source heater is embedded in the ion source cavity. The system produces ions without the ion source filament being on but can also be used with it on to produce solvent induced CI spectra. The system has been shown to be capable of providing mass spectral data from a wide range of difficult molecules, c.g., underivatised peptides, nucleotides and vitamin B,,, and impressive sensitivities have been demonstrated with on-line LC - MS of nucleosides. Unlike the other systems described, this type of interface is capable of handling 2 ml min-1 of water.A system based on this approach has recently become commercially available from Finnigzin-MAT and it appears to be the method of choice for handling low-volatility com- pounds that can be ionised in solution. Liquid ion evaporation is a similar technique to thermospray ionisation in that it is an extremely effective technique for handling compounds that are introduced into the mass spectrometer in ionic form.43v44 This technique has been used for LC - MS with an atmos- pheric pressure ion (.\PI) source. Ions that are present in liquid solution are emitted into the .API source bjr use of a strong electric field. In general (SfH)+ or (JI-H)- ions are formed with little fragmentation ; lwwwr, use of collision-induced dissociation with a triple quad- rupole instrument enables further structural information to be obtained.This technique has not yet been as fully developed as thermospray ionisation for LC - 31s but shows consider- able promise for handling low-volatility ionic compounds. Microbore LC - MS Although this approach appears to have little advantage for the third of the approaches to LC - SIS, it has considerable advantages for interfaces of the first two types. The advan- tages with interfaces of the direct liquid introduction type are immediately apparent in that if flow-rates of the order of 20 pl niin-l are used all of the eluent from the LC can be fed into the mass spectrometer ion source, resulting in considerable improvements in detection levels measured in terms of sample injected on-column. One of the problems with this type of interface is that some loss of chromatographic performance occurs unless the column is directly in the interface and special interfaces have to be designed.45 Jlicrobore LC has been used exte~isively~l+-~~ with interfaces of this type and with interfaces of the jet5” and vacuum nebulisingj8-61 types.Although interfaces of the moving belt type are capable of handling higher solvent flow-rates than those of the direct liquid introduction type, problems can be encountered through thermal decomposition or steam distillation of sample at the high temperature required in the infraredJanuary, 1984 CHROMATOGRAPHY - MASS SPECTROMETRY 27 heater to handle high percentage aqueous mobile phases and there are additional problems due to solvent beading on the belt if systems containing over 50% of water are used.Use of a nebulising system to spray sample on to the belt is one solution to the p r ~ b l e m . ~ ~ , ~ ~ Micro- bore LC provides an equally effective ~ o l u t i o n . ~ ~ ~ ~ ~ Our initial studies21 in this area used a JASCO microbore LC system which uses 0.5 mm i.d. PTFE tubing for columns. We were able to develop efficient methods for packing columns of this type, but their performance did match those of conventional columns. However, we were able to show that use of this approach enables better detection limits to be achieved in terms of sample.injected on-column, as with aqueous mobile phases all the column's effluent could be fed on to the interface without the necessity to split off some of the solvent as occurs with conventionalcolumns. In addition, by feeding ethanol on to the belt behind the eluent from the microbore column, high percentage aqueous mobile phases could be readily handled. Recently we41 have explored the use of commercially available glass-lined stainless-steel and stainless-steel columns of 1 mm i.d.and packed with 10-pm material. These columns give considerably improved chromatography equivalent to that obtained from conventional columns packed with similar material. The approach has been used in studies of pesticides and natural products and has also been used with gradient elution64 to study extracts from test well samples taken from landfill sites. In addition to the benefits mentioned earlier by increasing the flow-rate, microbore LC enables fast analysis and flow programming to be per- formed with interfaces of the moving belt type.This type of study would necessitate splitting of some of the column effluent with interfaces of the direct liquid introduction type. Conclusions Microbore LC provides an effective means of reducing solvent costs, improving detection limits and separating complex mixtures. It provides improved detection limits with LC - MS interfaces of the direct liquid and moving belt types and in the latter instance enables aqueous mobile phases to be more effectively handled. Combined LC - MS is currently in a fluid state, interfaces of the moving belt and direct liquid introduction types are effective for handling a wide range of compounds but systems of the thermospray and liquid ion evaporation types appear to be the methods of choice if low- volatility ionic compounds are being investigated.However, use of surface ionisation tech- niques with moving belt systems may also be effective in this area.37 There appear to be five clearly defined areas for LC - MS : (1) compounds that are amenable to gas chromatographic study but are analysed by LC because it is more efficient ; (2) compounds that are thermally labile and decompose on gas chromatography (GC) or do not have sufficient volatility for GC but are amenable to direct insertion probe mass spectrometry ; (3) compounds amenable only to mass spectral study by desorption CI; (4) compounds amenable only to mass spectral study by field desorption, fast atom bom- (5) compounds not currently amenable to mass spectral study.Compounds of classes 1-3 are amenable to study by systems of the direct liquid introduction, moving belt and thermospray types. However, it should be noted that not all systems of the first two types fall into this category. The choice of system depends on the type of study being undertaken. If it is of a qualitative nature then the belt systems appear to have advantages in that they provide data that can be readily compared with those in the literature or computer library to establish identification. However, if compounds are being studied where there is considerable variation in relative amounts problems could occur owing to thermal decomposition of the compounds present in small amounts and the thermospray or direct liquid introduction approach would appear to be preferable.For class 4 and ionic com- pounds thermospray ionisation appears to be the method of choice, although liquid ion evaporation may be equally effective or better but remains to be fully proven. Use of belt systems with surface ionisation may have an application in this area and there are some indications that direct liquid introduction may also be used. Class 5 awaits further develop- bardment or californium-252 plasma desorption ; and28 CHROMATOGRAPHY - MASS SPECTROMETRY Anal. Proc., VoZ. 21 ment in mass spectral instrumentation-the limitations here are not necessarily in the ionisa- tion methods but in the capabilities of mass spectrometers in providing data from molecules of high relative molecular mass.N. J.A. and M.A.M. thank the Analytical Trust Fund of the Royal Society of Chemistry and the SERC and Beecham Pharmaceuticals, respectively, for financial support. We thank the Royal Society and SERC for financial assistance in the purchase of liquid chromato- graphic and mass spectrometric equipment. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. References Novotny, M., Anal. Chem., 1981, 53, 129A. Knox, J. H., J. Chromatogr. Sci., 1980, 18, 453. Guiochon, G., Anal. Chem., 1981, 53, 1318. Scott, R. P. W., and Kucera, P., J. Chromatogr., 1976, 25, 251; 1979, 169, 51; 1979, 185, 27. Scott, R. P. W., J.Chromatogr. Sci., 1980, 18, 49. Ryall, R. R., and Kessler, Jr., H. D., Int. Lab., 1982, June, 68. Tsuji, K., and Binns, R. B., J. Chromatogr., 1982, 253, 227. Vadukul, N. K., and Loscombe, C. R., J. High Resolut. Chromatogr. Chromatogr. Commun., 1982, 5, Hermansson, J., Chromatographia, 1981, 4, 43. Ishii, D., Asai, K., Hibi, K., Jonkuchi, J., and Nagaya, M., J. Chromatogr., 1977, 144, 157. Yang, F. J., J. Chromatogr., 1982, 236, 265. Ishii, D., and Takeuchi, T., J. Chromatogr., 1983, 255, 349. Novotny, M., J. Chromatogr. Sci., 1980, 18, 473. Tsuda, T., Hibi, K., Nakanishi, T., Takeuchi, T., and Ishii, D., J. Chromatogr., 1978, 158, 277. Hibi, K., and Ishii, D., J. Chromatogr., 1980, 189, 179. Ishii, D., and Takeuchi, T., J. Chromatogr. Sci., 1980, 18, 462. Tijssen, R., Bleumer, J.P. A., Smit, A. L. C., and Van Kreveld, M. E., J. Chromatogr., 1981, 218, 137. Yang, F. J., J. Chromatogr. Sci., 1982, 20, 241. McGuffin, V. L., and Novotny, M., J. Chromatogr., 1983, 255, 381. Jinno, K., J. High Resolut. Chromatogr. Chromatogr. Commun., 1982, 5, 364. Games, D. E., Lant, M. S.. Westwood, S. A., Cocksedge, M. J., Evans, N., Williamson, J., and Wood- Arpino, P. J., and Guichon, G., Anal. Chem., 1979, 51, 682A. McFadden, W. H., J. Chromatogr. Sci., 1979, 17, 2; 1980, 18, 97. McFadden, W. H., Anal. Proc., 1982, 19, 258. Games, D. E., Anal. Proc., 1980, 17, 110 and 322. Games, D. E., Biomed. Mass Spectrom., 1981, 8, 454. Games, D. E., in Morris, H. R., Editor, “Soft Ionization Biological Mass Spectrometry,” Heyden, Melera, A., Adv.Mass Spectrom., 1980, 8B, 159. Arpino, P. J., and Guiochon, G., J. Chromatogr., 1979, 185, 529. Arpino, P. J., Krien, P., Vajta, S., and Devant, G., J. Chromatogr., 1981, 203, 117. Arpino, P. J., and Guiochon, G., J. Chromatogr., 1982, 251, 153. Dedieu, M., Juin, C., Arpino, P. J . , and Guiochon, G., Anal. Chew., 1982, 54, 2375. Edmonds, C. G., McCloskey, J. A., and Edmonds, V. A., Biomed. Mass Spectrom., 1983, 10, 237. Henion, J. D., Thomson, B. A., and Dawson, P. H., Anal. Chem., 1982, 54, 451. Voyksner, R. D., Hass, J. R., and Bursey, M. M., Anal. Lett., 1982, 15, 1. Millington, D. S., Yorke, D. A., and Burns, P., Adv. Mass Spectrom., 1980, 8B, 1819. Dobberstein, P., Korte, E., Meyerhoff, G., and Pesch, R., Int. J. Mass Spectrom. Ion Phys., 1983, 46, McFadden, W.H., Schwartz, H. L., and Evans, S., J. Chromatogr., 1976, 122, 389. Games, D. E., McDowall, M. A., and Levsen, K., unpublished work. Alcock, N. J., Eckers, C., Games, D. E., Games, M. P. L., Lant, M. S., McDowall, M. A., Rossiter, M., Alcock, N. J., Corbelli, L., Games, D. E., Lant, M. S., and Westwood, S. A., Biomed. Mass Spectrom., Vestal, M. L., Int. J. Mass Spectrom. Ion Phys., 1983, 46, 193. Thomson, B. A., Iribarne, J. V., and Dziedzic, P. J., Anal. Chem., 1982, 54, 2219. Shushan, B., Fulford, J . E., Thornson, B. A., Davidson, W. R., Danylewych, L. M., Ngo, A., Nac- Brophy, J. J., Nelson, D., and Withers, M. K., Int. J. Mass Spectrom. Ion. Phys., 1980, 36, 205. Henion, J. D., and Maylin, G. A., Biomed. Mass Spectrom., 1980, 7 , 155. Henion, J. D., J.Chromatogr. Sci., 1981, 19, 57. Henion, J. D., and Wachs, T., Anal. Chem., 1981, 53, 1963. Eckers, C., Skrabalak, D. S., and Henion, J.. Clin. Chem., 1982, 28, 1882. Eckers, C., Henion, J . D., Maylin, G. A., Skrabalak, D. S., Vessman, J., Tivert, A. M., and Greenfield. J . C., Int. J . Mass Spectrom. Ion Phys., 1983, 46, 205. 360. hall, B. J., Biomed. Mass Spectrom., 1982, 9, 215. London, 1981, p. 54. 185. Smith, R. W., Westwood, S. A., and Wong, H.-Y., J. Chromatogr., 1982, 251, 165. 1982, 9, 499. son, S., and Tanner, S. D., J. Mass Spectrom. Ion Phys., 1983, 46, 225.January, 1984 CHROMATOGRAPHY - MASS SPECTROMETRY 29 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. Schafer, K. H., and Levsen, K., J . Chromatogr., 1981, 206, 245. Levsen, K., and Schafer, K.H., Int. J . Mass Spectrom. Ion Phys., 1983, 46, 209. Yanianchi, E., Mizuno, T., and Azuma, K., Shibak Shitsuryo Bunseki, 1980, 28, 227. Krien, P., Devant, G . , and Hardy, M., J . Chromatogr., 1982, 251, 129. Okada, S., and Iida, Y., Shibak Shitsuryo Bunseki, 1981, 29, 287. Bruins, A. P., and Drenth, B. F. H., Int. J . Mass Spectrom. Zon Phys., 1983, 46, 213. Takeuchi, T.. Hirata, Y., and Okumara, Y., Anal. Chem., 1978, 50, 659. Tsuge, S., Hirata, Y., and Takeuchi, T., Anal. Chem., 1979, 51, 166. Yoshida, Y., Yosida, H., Tsuge, S., Takeuchi, T., and Mochizuki, K . , J . High Resolut. Chromatogr. Chromatogr. Commun., 1980, 3, 16. Tsuge, S., Yoshida, Y., Takeuchi, T., Mochizuki, K., Kokubun, N., and Hibi, I<., Chem. Biomed. Environ. Instrum., 1980, 10, 405.Yoshida, H., Matsumoto, K., Itoh, K . , lsuge, S., Hirata, Y., Mochizuki, I<., Kokubun, N., and Yoshida, Y., Fresenius 2. Anal. Chem., 1982,311, 674. Smith, R. D., and Johnson, A. L., Anal. Chem., 1981, 53, 739. Lankmayr, E. P., Hayes, M. J., Karger, 13. L., Vouros, P.,and McGuire, J . M., Znt. J . Mass Spectrom. Foster, M . G . , Meresz, O., Games, D. E., Idant, M. S.,and Westwood, S. A., Biomed. Mass Spectrom., Ion Phys., 1983, 46, 177. 1983, 10, 338. Analysis of Endogenous and Exogenous Compounds in Tumour Tissue by Gas Chromatography- High-resolution Mass Spectrometry Simon J. Gaskell, Heather M. Leith and Brian G. Brownsey Tenovus Instztute for Cancer Research. Welsh Natzonal School of Medicine, Heath, Cardig, CF4 4XX The mechanism of action of natural oestrogens, in common with other steroid hormones, is now considered to involve binding to a specific cytosol receptor protein, transformation of the steroid - receptor complex and translocation to the nucleus.Interaction with the chroma- tin stimulates RNA and DNA synthesis and ultimately promotes protein synthesis and normal cell function and growth. Synthetic anti-oestrogens, now widely used in the treatment of breast cancer, are thought to exert their beneficial effect, at least in part, by sequestering the receptor but failing to elicit a full oestrogenic response. Study of natural oestrogens and synthetic drugs requires a knowledge of circulating concentrations and of the levels achieved in target tissues. The concentrations of oestrogens in blood plasma are low (generally in the pg ml-1 range) and, of this total, only a small proportion is non-protein bound and hence physiologically active.A direct measure of this fraction is obtained by analyses of saliva where the concentrations of oestradiol-17/3, for example, rarely exceed 20 pg ml-I. According- ly, the analyses of natural and synthetic compounds which bind to the oestrogen receptor are particularly demanding in terms of both sensitivity and selectivity. While analyses based on gas chromatography - mass spectrometry (GC - MS) are clearly of great potential for such studies, considerable attention must be given to achieving specificity during both sample work-up and mass spectrometric detection. Sample Work-up Procedures \.\'ork-up procedures where fractionations are based on criteria different from those which apply during GC - 11s characterisation are particularly beneficial.Anion-exchange chromatography, for example, may be used to exploit the weak acidity of the natural oestro- gens (and synthetic oestrogens such as diethylstilboestrol) attributable to their phenolic structure. The use of triet hylammoniohydroxypropyl-Lipidex 5000 was developed by Axelson and Sjovalll and has been applied to the analysis of oestrogens in urine, saliva2 and tissue3 extracts. Fractionation on a single micro-column in sequential reversed-phase, straight-phase and ion exchange modes provides a rapid separation and high degree of sample purification. Cation-exchange chromatography, using sulphoethyl-Sephadex LH-20,4 has been applied to the isolation of the basic anti-oestrogenic drug tamoxifen from extracts of plasma5 and tissue.6 An alternative procedure for the isolation of natural oestrogens is provided by the use of solid-phase coupled antisera.Immunoadsorption, a technique borrowed and adapted from the immunoassaJ-ist, permits the direct extraction of selected analytes fiom a biological fluid or fractionation of tissue extracts. Table I gives the extraction efficiencies, from blood plasma,30 CHROMATOGRAPHY - MASS SPECTROMETRY Anal. Proc., Vol. 21 of various steroids using an oest radiol-l7/3 antiserum coupled to microcellulose and compares the cross-reactions determined conventionally according to the Abraham criteria.’ A typical immunoadsorption procedurea incorporates addition of a suspension of the solid-phase coupled antiserum to the biological fluid, brief incubation, followed by centrifugation and washing of the solid phase with buffer and water.The analyte and coextractants are recovered by methanol stripping, after which the coupled antiserum is available for re-use. The extremely rapid technique is valuable primarily because the stereochemical discrimination implicit in its use complements the subsequent GC - MS characterisation. TABLE I EXTRACTION EFFICIENCIES OF STEROIDS FROM PLASMA USING A CELLULOSE-COUPLED ANTI-OESTRADIOL-17fl SERUM, AND COMPARISON WITH THE CROSS-REACTIVITIES OF THE LIQUID ANTISERUM Data from Teference 8. Steroid Extraction efficiency,* :(, Cross-reaction,? yo Oestradiol-17fi . . .. .. .. .. 81 100 Oestrone .. .. .. .. .. .. 65 24 Oestriol . . .. . . .. .. .. 53 9.5 Testosterone . . .. .. .. .. 3 (0.1 Dehydroepiandrosteroiie . . .. .. . . 3 (0.1 Cortisol . . .. .. . . .. . . <1 - * Extraction of 3H-labelled steroid from 0.5 nil of plasma using 0.5 nil of a suspension of solid-phase-coupled t Determined according to the --\braham criteria,’ with an antiserum dilution of 1 : 150 000. antiscruni a t a dilution of 1 : 100. GC - MS Characterisation and Quantification The sensitivities required for analysis of many steroid hormones and drugs in body fluids and tissues dictate the use of selected ion monitoring (SIM) techniques during GC - MS. The implicit neglect of most of the features of the mass spectrum of the analyte may lead to un- certainty of compound identification and various approaches may be taken to enhance the selectivity of detection.High sensitivi- ties and selectivities have been reported in favourable cases using negative chemical ionisation (CI),s although the exploitation of the technique in the biomedical area is at an early stage. The use of halogenated derivatives estends the range of application of negative CI in the electron-capture mode. Thus, for example, oestradiol-17p bisheptafluorobutyrate may be detected at the low picograni level during GC - negative CIlIS/SIlI of (M-HF)- ions. The use of high 11s resolution during SIJIlO represents a technique of general applicability to increase selectivity of detection. Signal intensities are necessarily reduced but the useful sensitivity, defined as the signal to noise ratio observed during the analysis of biological samples, niay actually be increased owing to the reduction in “chemical noise.’’ The superior selectivity has been demonstrated in analyses of, itttcv d i n , oestradiol-17/3 in blood plasma8 and testosterone in prostatic tissue.ll GC - high resolution W3/SIJl has been used in analyses of the anti-oestrogen taniosifen in blood plasma5 and tumour tissue6 of treated breast cancer patients. The quantitative data were shown to be consistent with the hypothesis of the mechanism of action of the drug as a coinpetitor to oestradiol-17/3 for binding to the oestrogen receptor. More recent studies of the concentrations of the three “classical” oestrogens (oestrone, oestradiol-17/3 and oestriol) in human breast tuniours have further illustrated the value of the high resolution technique. Concentrations are generally low, with levels of oestriol, for example in the range 0-500 pg g-1 in tissue. Netastable peak monitoringllJ2 provides an experimentally simple, although less widely applicable, alternati1.e to high-resolution SIN. ,in appropriate choice of derivative has an important bearing on the sensitivity achie\?ed. -4nalyses of steroids in both tissue and plasma samples have been reported. Further enhancement of selectivity is dependent on an im- proved resolution of the parent ion beam. This may be achieved using one of the recently described configurations of the multiple analyser mass spectronieter ; the value of such instruments in biochemical trace analysis, however, has yet to be demonstrated. The most attractive of these is selective ionisation.January, 1984 EQUIPMENT NEWS The support of the Tenovus Organisation is gratefully acknowledged. 31 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 1 1 . 12. References Axelson, M., and Sjovall, J., J . Steroid Biochem., 1977, 8, 683. Gaskell, S. J., Finlay, E. M. H., and Pike, A. W., Biomed. Mass Spectrom., 1980, 7, 500. Axelson, M., Clark, J. H., Eriksson, H., and Sjovall, J., J . Steroid Biochem., 1981, 14, 1253. Setchell, K. D. R., AlmC, B., Axelson, M., and Sjovall, J., J . Steroid Biochem., 1976, 7, 615. Daniel, C. P., Gaskell, S. J., Bishop, H., and Nicholson, R. I., J . Endocrinol., 1979, 83, 401. Daniel, C. P., Gaskell, S. J., Bishop, H., Campbell, C., and Nicholson, R. I., Eur. J . Cancer CEin. OncoE., Abraham, G. E., J . Clin. EndocrinoZ. Metab., 1969, 29, 866. Gaskell, S. J., and Brownsey, B. G., Clin. Cham., 1983, 29, 677. Markey, S. P., Lewy, A. J., and Colburn, R. W., in De Leenheer, A. P., Roncucci, R. R., and Van Peteghem, C., Editors, “Quantitative Mass Spectrometry in Life Sciences, 11,” Elsevier, Amster- dam, 1978, p. 17. 1981, 17, 1183. Millington, D. S., J . Steroid Biochem., 1975, 6, 239. Gaskell, S. J., Finney, R. W., and Harper, M. E., Biomed. Mass Spectrom., 1979, 6, 113. Gaskell, S. J., and Millington, D. S., Biomed. Mass Spectrom., 1978, 5, 557.