2 Physical Methods Part (i) Organic Mass Spectrometry By R. A. W. JOHNSTONE The Robert Robinson Laboratories University of Liverpool P.O. Box 147 Liverpool L69 3BX and F. A. MELLON Agricultural Research Council Unit of Invertebrate Chemistry and Physiology The Chemical Laboratory The University of Sussex Brighton Sussex BNI 9QJ In this Report we continue the practice established in the last Report of covering selected areas of organic mass spectrometry. The coverage is far from exhaustive and many references have been omitted for brevity. However the references given are leading ones and for greater coverage the reader is advised to consult the Specialist Periodical Reports or the incredibly condensed yet wide-ranging survey-cum-catalogue of mass spectrometry which has appeared with over 1700 references to current or recent work.2 As a guide to trends in mass spectrometry the reader would find value in consulting the abstracts of the 20th Annual Con- ference of the American Society for Mass Spe~trometry.~ A very useful book on the uses of mass spectrometry in biochemistry has a~peared,~ along with another introductory text on general mass spectrometry,’ and a review.6 1 Theoretical Aspects Most of the recent theoretical investigations on rate processes still use the grossly over-simple quasi-equilibrium equation k = v( 1 -E,/E)N-’.This equation has sufficient parameters to allow it to be ‘fitted’ to many experimental observations without difficulty. Although the equation is very useful for teaching or illustra- tive purposes there are many dangers inherent in its indiscriminate application ‘Mass Spectrometry’ ed.D. H. Williams (Specialist Periodical Reports) The Chemical Society London 1971 Vol. 1; 1973 Vol. 2. A. L. Burlingame and G. A. Johnson Anafyt. Chem. 1972 44 33713. Abstracts of the 20th Annual Conference on Mass Spectrometry and Allied Topics arranged by the American Society for Mass Spectrometry in co-operation with ASTM Committee E-14 Dallas June 1972. ‘Biochemical Applications of Mass Spectrometry’ ed. G. R. Waller J. Wiley Chichester 1972. ‘Principles of Organic Mass Spectrometry’ D. H. Williams and I. Howe McGraw-Hill London 1972. ‘Mass Spectrometry’ ed. A. Maccoll M .T.P. International Review of Science Physical Chemistry Series 1 Vol.5 Butterworth Baltimore 1972. 7 8 R. A. W.Johnstone and F. A. MelIon to experimental data. For example calculation of the pre-exponential factors of the above equation from mass spectrometric data was adduced as evidence for hydrogen bonding in the transition state for loss of C2H,0 from the molecular ions of ortho-substituted acetanilides.’ Further experiment has led to a retrac- tion of this hypothesis.* As in our previous Report,’ we strongly recommend that if quasi-equilibrium calculations are to be carried out at all the full equation be used especially in view of the ready access of most scientific investigators to computers. We have criticized the use of photoelectron spectra as energy distributions in QET calculations on electron-impact-induced decompositions citing our own unpublished workg Further experimental evidence that photoelectron spectros- copy provides too crude an approximation to electron-impact-induced internal energy distributions in ions has now appeared.lo An interesting investigation into the relative accuracies of expressions for determining rate constants in ion-molecule reactions was undertaken with the aid of ion-cyclotron mass spectrometry.’ 2 Ionization and Appearance Potentials The question of the accuracy of ionization and appearance potentials obtained from electron-impact ionization-efficiency curves appears to be reaching a stage where a definitive answer can be given in the near future. There seems little doubt that ionization and particularly appearance potentials determined by semi- empirical methods such as the semi-log plot vanishing current and Warren techniques or variations on these are somewhat ambiguous.Although the reproducibility of these may be very good their accuracy is questionable especially when there is considerable curvature at the foot of the ionization-efficiency curve. Mathematical analysis of ionization-efficiency curves promises to overcome the drawbacks in the semi-empirical methods and there is currently a lot of activity in this area. The various methods of determining appearance potentials have been evaluated with reference to the critical-slope approach. l2 Of the more interest- ing results of this investigation the following are highly relevant (i) differences between the appearance potentials of normal and metastable fragment ions were on average 0.45 eV when determined by the popular semi-log plot method but were only 0.04 eV when determined by the critical-slope method -investigators of kinetic shifts beware! This problem was highlighted in a recent review.130 (ii) The differences between the appearance potentials of fragment ions deter- mined by the semi-log plot and critical-slope methods were on average 0.45 & 0.17 eV. (iii) Four high-energy competitive fragmentations gave appearance ’ S. A. Benezra and M. M. Bursey J. Chem. SOC.(B) 1971 1515. S. A. Benezra and M. M. Bursey J.C.S. Perkin II 1972 1537. Annual Reports (B) 1971 68 6. 10 G. Innorta S. Torroni and S. Pignataro Org.Mass. Spectrometry 1972 6 43. I‘ T. McAllister Internat. J. Mass Spectrometry Ion Phys. 1972 8 162. J. L. Occolowitz and B. J. Cerimele ref. 3 p. 95. l3 T. W. Bentley and R. A. W. Johnstone Adv. Phys. Org. Chem. 1970 8 (a) p. 185; (6)p. 164. Physical Methods-Part (i) Organic Mass Spectrometry 9 potentials by the semi-log plot method greater by 0.9-1.0 eV than those obtained by the critical-slope method. Further evidence for the potential of electron-impact data obtained from conventional ion sources has been pr~vided'~ by a fairly comprehensive investi- gation into the use of the electron-energy-distribution-difference te~hnique'~ allied to computer acquisition of ionization data. The accuracy of measuring ionization potentials was comparable with photoionization and photoelectron methods and for appearance potentials it was shown that some cases of reported kinetic shifts were due to the inaccurate methods previously used to measure appearance potentials.It is interesting to note in view of the results obtained with the critical-slope method that the energy-distribution-difference approach is not a proper mathematical folding but it can be shown that it is also a critical- slope method. '' A third-derivative technique17 has been used in attempts to overcome the resolution problems in the second-derivative method. '* Good agreement with photoionization and monoenergetic electron-impact data was found although the method suffered from problems due to random noise the effects of which are multiplied greatly when second or third derivatives are taken from ionization- efficiency curves.To overcome the noise problem ionization-efficiency curves have been scanned many times and the data averaged to give excellent results." However appearance potentials determined by the second-derivative method may be subject to large errors in some cases.2o A simple new technique for measuring ionization and appearance potentials2 ' will undoubtedly suffer from the drawbacks inherent in the other empirical methods. 3 Ionization Methods The use of electron impact to produce mass spectra continues to be the most widely used of all the methods of ionization. Its attraction is partly due to its earlier evolution but is also due to the relative robustness and ease of operation of ion sources constructed on this principle.However two other methods chemical and field ionization are rapidly gaining ground and offer attractive features unavailable by electron impact. Field ionization (FI) and field desorp- tion (FD) mass spectrometry continue to be areas in which there are few but active groups of workers. Both FI and FD methods give abundant molecular ions but relatively little fragmentation and they prove ideal therefore for examining 14 R. A. W. Johnstone and F. A. Mellon J.C.S. Faraday 11 1972 68 1209. 15 R. E. Winters J. H. Collins and W. L. Courchene J. Chem. Phys. 1966 45 1931. 16 R. A. W. Johnstone and B. N. McMaster to be published. 17 C. D. Finney and A. G. Harrison Internat.J. Mass Spectrometry Ion Phys. 1972 9 221. I8 J. D. Morrison J. Chem. Phys. 1953 21 1767. 19 R. G. Dromey J. D. Morrison and J. C. Traeger Internat. J. Mass Spectrometry Ion Phys. 1971 6 57. 20 G. G. Meisels and B. G. Giessner Internat. J. Mass spectrometry Ion Phys. 1971 7 489. 21 G. D. Flesch and H. J. Svec Internat. J. Mass Spectrometry Ion Phys. 1972 9 106. 10 R.A. W. Johnstone and F. A. Mellon mixtures since the mass spectrum of a mixture is composed almost entirely of molecular ions.22 High-resolution FI mass spectrometry gives the molecular formulae of the components of a mixture.23 Although FI and FD methods give some fragmentation the reproducibility of the mass spectral fragmentation pat- terns is apparently not good and spectra are frequently accumulated on photo- plates rather than by a more convenient electrical recording device.The FD method gives particularly good results with labile or difficultly volatile materials that are available only in small quantities because the mass spectra are obtained at room temperature. For example the mass spectra of 0.1-1.0nmole of amino-acids (including arginine and cystine) can be obtained without the prepara- tion of derivatives and the mass spectrum of a pentapeptide containing arginine has been obtained similarly.24 The low temperatures employed in FD work constitute probably the biggest advantage of this method over chemical ioniza- tion. There are now many groups of workers with chemical ionization (CI)sources and several important advances have been made in practical techniques.CI mass spectroscopy suffers like FI work in producing relatively little fragmenta- tion although molecular ions are abundant. The absence of fragmentation patterns is a serious drawback because a great deal of structural information is then not available. To combine the advantages of abundant molecular ions with extensive fragmentation CI mass spectroscopy has been carried out with mix- tures of water and argon as reagent gases.25 In the technique described argon is used as the carrier gas in gas chromatography and water is injected into the effluent before it passes into the ion source where two sorts of reaction (1) and (2) occur. Reaction (1) is protonation of a molecule (M) by H30+ to produce M + H30+ -+ MH+ + H,O M + Ar" -+ M+' + Ar a quasi-molecular ion (MH) having little excess of internal energy and therefore there is little fragmentation of the MH ion i.e.it is abundant in the mass spectrum. Reaction (2) is a charge-exchange reaction to produce a molecular ion with a considerable excess of internal energy so that fragmentation is rapid and exten- sive i.e.abundant fragment ions are produced. Thus reactions (1)and (2) together produce a mass spectrum with abundant molecular ions and fragment ions. This method also illustrates another advantage of CI with g.c.-m.s. namely that no molecular separator is needed between the end of the g.c. column and the ion source.26 22 H. D. Beckey H. Knoppel G. Metzinger and P. Schultze 'Advances in Mass Spectro- metry' Vol.3 ed. W. H. Mead institute of Petroleum London 1966 p. 35; H. D. R. Schuddemage and D. 0. Hummel ibid. Vol. 4 ed. E. Kendrick 1968 p. 857. 23 J. B. Forehand and W. F. Khun Analyf. Chem. 1970,42 1839; H. R. Schulten H. D. Beckey H. L. C. Menzelaar and A. J. H. Boerboom Analyr. Chem. 1973 45 191. 24 H. W. Winkler and H. D. Beckey Org. Mass Spectrometry 1972 6 655. 25 D. F. Hunt and J. F. Ryan Analyr. Chem. 1972 44 1306. 26 G. P. Arsenault J. J. Dolhun and K. Biemann Chem. Comm. 1970 1542. Physical Methods-Part (i) Organic Mass Spectrometry 11 Other reagent gases have been used in CI sources. Ammonia has been pro- posed as a selective reagent for certain functional groups in molecules e.g. conjugated ketones,27a and nitric oxide has been used to ‘identify’ functional groups by the nature of the quasi-molecular ion pr~duced.~” These investiga- tions promise significant advances in structure elucidation by mass spectrometry of very small quantities of material.Other uses of CI are included in Section 6. An intriguing development is the production of CI spectra at low ion-source pressures. Normally in C1 work the source is operated at a relatively high gas pressure and this can cause difficulties through high-voltage arcing. By the use of ion-storage ion-molecule reactions can be obtained at much lower pressures than in the normal CI source. The device is a three-dimensional quadrupole ion storage trap and hence its name quistor.28 4 Chromatographic-Mass Spectrometric Methods The most widely used combination of two analytical methods is that employing gas chromatography and mass spectrometry (g.c.-m.s.).Gas chromatography has gone through a long development period and has now reached a stage where separation of complex mixtures is routine. Similarly mass spectrometry has been developed as a powerful analytical tool which by the use of empirical rules gives a great deal of information on the structures of molecules. The two methods in tandem yield an analytical instrument far more powerful than either of the individual methods alone. There has been one major difficulty in operating gas chromatographs in tandem with mass spectrometers and this is caused by the need to have relatively high gas flows and pressures in the former instrument but relatively very low gas pressures in the latter.The following section describes the efforts made to circumvent the difficulty. Commonly open tubular (capillary) gas-chromatographic columns use gas flows of less than about 2mlmin-’ and these gas flows can be accepted by modestly sized pumps in the mass spectrometer. It is possible therefore to couple a capillary gas-chromatographic column directly into the ion source of the mass spectrometer so that all the g.c. efffuent is available for a mass spectrum. If the gas flow through the capillary is a little too high a splitter can be used to vent a proportion of the effluent to the atmosphere although of course that proportion is then lost and is not available for mass spectrometric analysis.Recently the use of very high-capacity pumps has been described so that one can allow the total effluent of a g.c. column (up to 20 ml min- ’) to flow directly into the mass ~pectrometer.~’ Under these circumstances even packed g.c. columns can be used. As mentioned briefly earlier CI sources work at relatively high pressures and use high-capacity pumps so that g.c. effluents can be passed directly into the 27 (a) I. Dzidic and J. A. McCloskey Org. Mass Spectrometry 1972 6 939; (b) D. F. Hunt and J. F. Ryan J.C.S. Chem. Comm. 1972 620; see also F. H. Field Accounts Chem. Res. 1968 1 42 and M. S. Wilson I. Dzidic and J. A. McCloskey Biochem. Biophys. Acta 1971 240 623. J. F. J. Todd and G. Lawson Chem. in Brit. 1972 8 373. 29 R.F. Bonner G. Lawson and J. F. J. Todd J.C.S. Chem. Comm. 1972 1179. 30 W. Henderson and G. Steel Analyt. Chem. 1972 44 2302. 12 R.A. W. Johnstone and F. A. Mellon ion source and the g.c. carrier gas becomes the reactant gas of the CI process. Thus with these sources methane or argon is frequently used as the g.c. carrier gas. When electron-impact sources are used with g.c. apparatus it is not only essen- tial to maintain a low pressure in the ion source but the residual carrier gas passed into the source must have an ionization potential high enough for it not to be ionized along with the organic material under investigation. For this reason helium is the carrier gas of choice when electron-impact sources are coupled to g.c. instruments and the electron energy is maintained below 20 V.Therefore although the carrier gas does not pose much of a problem the need for low gas pressures in the ion source does. A device is needed to separate as much carrier gas as possible from the organic material eluted from the g.c. column (enrichment) and to pass as much organic material as possible to the mass spectrometer (transfer efficiency). The ideal device (molecular separator) would provide infinite enrichment with 100% transfer but most practical models are far from this standard and much of the organic material eluted from the g.c. column is lost resulting in a reduction in effective sensitivity of the combined g.c.-m.s. technique compared with either technique alone. The simplest device for passing organic effluent from a g.c.column into a mass spectrometer is simply a short length of cooled capillary tube at the end of the g.c. column in which the organic material is condensed. The capillary tube and its contents are then carried to the mass spectrometer and a mass spectrum is obtained in the usual way. This method is somewhat clumsy and inefficient and not recommended for regular use but where say only one or two components eluted from a g.c. column need to be looked at occasionally it provides a simple cheap method. Generally molecular separators are on-line devices that are used continuously. A brief list of the currently used molecular separators is given here for the benefit of the reader but for a more detailed description a number of recent reviews should be c~nsulted.~~ The Ryhage (Jet) separator32 depends on the relatively much greater rate of diffusion of helium compared with organic com- pounds of higher molecular weight from a stream of them passing through a low- pressure area.This low-pressure area is simply a small evacuated gap between a narrow jet and a slightly larger orifice. Despite its all-metal construction thermal decomposition is usually low because the sample traverses the separator at speeds in excess of the velocity of sound. Nevertheless some compounds particularly polar ones are lost in the separator especially when present in low concentrations and there is a need for an all-glass separator. The transfer effi- ciency is variable but in favourable cases can exceed 50 %.The Llewellyn (mem- brane) separator33 allows organic material to dissolve in and diffuse through a very thin non-porous membrane into the mass spectrometer but carrier gas ” A. N. Freedman Analyt. Chim. Acta 1972 59 19; R. Ryhage and S. Wikstrom in ‘Mass Spectrometry. Techniques and Applications’ ed. G. W. A. Milne Wiley London 1971; G. A. Junk Internat. J. Mass Spectromelry Ion Phys. 1972 8 1. 32 R. Ryhage Analyt. Chem. 1964,36 759; Arkiv. Kemi 1967 26 26. 33 P. Llewellyn and D. Littlejohn 16th Annual Conference on Mass Spectrometry and Allied Tapics ASTM Committee E14 Pittsburgh 1968. Physical Methods-Part (i) Organic Mass Spectrometry 13 being less soluble does not dissolve in the membrane material and is vented straight to atmosphere.The membrane is usually made of silicone rubber but because the solution-diffusion process can take a relatively long time (0.1 s) some loss of the g.c. resolution may be noticeable particularly with capillary columns through mixing of components in the membrane. Although it is a good separator with high transfer and enrichment factors it cannot be used for extended periods at temperatures much above 200 “C. An all-glass membrane separator is now marketed. In the Watson-Biemann separator,34 enrichment is obtained by the different rates of diffusion of carrier gas and organic material on passage through a short length of tubular glass frit. The efficiency varies between 10 and 50% and being of all-glass construction this type is relatively trouble- free although the large surface area leads to some thermal decomposition prob- lems.The above three separators are the ones most commonly used and all are available commercially either as separate items or built into g.c.-m.s. units. Other separators which have been described include the silver membrane,35 the variable,36 and the reaction types,37 each of which has advantages and dis- advantages but has not been widely used. G.c.-m.s. methods are utilized extensively in chemical biochemical and bio- medical research and are mentioned again later. Other couplings of analytical techniques with mass spectrometry include protein ~equenator,~~ centrichr~matography,~~ and liquid-liquid chromatog-raph~.~~ 5 Computers The use of computers for acquisition of data from mass spectrometers is well established and the main area of interest has shifted to their use as aids to the interpretation of data.Storage of mass spectra on ‘files’ or in a ‘library’ which can be searched by computer has attracted considerable attention. To reduce the amount of information about each mass spectrum that needs to be stored abbreviated spectrum files are used in which only the more important charac- teristics of a mass spectrum are retained.40 For example it has been shown recently4* that by noting only two peaks in every 14 mass units along with some information on abundances a library of spectra can be set up and searched by computer in a matter of seconds. With the mass spectrum of a compound of unknown structure the library user can request the computer to search its files.34 J. T. Watson and K. Biemann Analyt. Chem. 1964 36 1135; ibid. 1965 37 844. R. Cree Conference on Analytical Chemistry and Applied Spectroscopy Pittsburgh March 1967. 36 C. Brunee H. J. Buttemann and G. Kappus Abstracts of the 17th Annual Conference on Mass Spectrometry and Allied Topics ASTM Committee E14 Dallas 1969 p. 121. ” P. Simmonds G. R. Shoemaker and J. E. Lovelock Analyt. Chem. 1970 42 881. 38 F. W. Karesch and P. W. Rasmussen Analyt. Chem. 1972 44 1488. 39 R. E. Lovins J. Craig and T. Fairwell ref. 3 p. 163; R. E. Lovins J. Craig F. Thomas and C. McKinney Analyt. Biochem. 1972 47 539. 40 H. S. Hertz R. A. Hites and K. Biemann Analyt. Chem. 1971 43 681. 4‘ S.L. Grotch Analyr. Chem. 1973 45 2. 14 R.A. W.Johnstone and F. A. Mellon If the computer has any spectra to match the unknown they are printed out. This sort of retrieval system has been developed to a considerable level of sophistication by incorporating a ‘conversational’ mode of operation between the user and the computer.42 In the conversational mode the user can ask a series of previously specified questions which the computer answers by search- ing the stored information from about 9O00 spectra. In this way the user can obtain a short list of possible identifications of an unknown very rapidly by feeding a minimum of information to the computer. Other workers have described a similar interrogation system.43 After examining the short list the user can add more information if that is necessary to obtain a closer identifica- tion.In a similar system library search facilities have been extended to include data from nuclear magnetic resonance and i.r. as well as from mass spectro- metr~.~~ Computer-aided classification and interpretation of mass spectra has been reviewed4’ as has the use of computers with g.c.-m.s. Another active area of research has been the application of computers as ‘learning’ or ‘artificial intelligence’ machines and an important advance has been the introduction of a pattern classifier of mass spectral information from non-linear chemical system^.^' In a more specific application of artificial intel- ligence a computer has been used to identify oestrogenic steroids through interpretation of high-resolution mass spectral data.48 6 Chemical Biochemical and Biomedical Uses of Mass Spectrometry The high sensitivity of mass spectrometry is being exploited increasingly for investigations of small samples of tissue or blood plasma in forensic science drug detection and metabolic studies.By the use of specialized techniques pico- gram quantities of known compounds can be identified and measured accurately. With such sensitivity only tiny amounts of tissue or plasma are needed. The specialized techniques needed for these studies are straightforward with the right equipment and several manufacturers supply suitable instruments. To identify small amounts of known compounds the whole range of the mass spectrometer is not scanned but 14 of the most abundant or characteristic peaks in the mass spectrum are scanned rapidly in succession.This method is termed mass fragmentography and has been reviewed with emphasis on its use for detecting biogenic amines psycho-active compounds pesticides pollutants 42 S. R. Heller H. M. Fales and G. W. A. Milne Org. Mass Spectrometry 1973 7 107; S. R. Heller Analyt. Chem. 1972 44 1951. 43 D. H. Smith and G. Eglinton Nature 1972 235 325. 44 T. Emi and J. T. Clerc Helc. Chim. Acta 1972 55 489. 45 D. H. Smith Analyt. Chem. 1972 44 536. 46 F. W. Karasek Analyt. Chem. 1972 44 32A. 47 J. B. Justice D. N. Anderson T. L. Isenhour and J. C. Marshall Analyt. Chem. 1972 44 2087. 48 D. H. Smith B. G. Buchanan R. S. Engelmore A. M. Duffield A.Yeo E. A. Feigen-baum J. Lederberg and C. Djerassi J. Amer. Chem. SOC. 1972 94 5962. Physical Methods-Part (i) Organic Mass Spectrometry 15 steroids purines prostaglandins and glucose.49 If we suppose a known com- pound has two characteristic peaks in its mass spectrum at say m/e 100 and 200 then in mass fragmentography the tissue sample or extract to be examined is heated in the ion source whilst the mass spectrometer ‘looks’ only for peaks at m/e 100 and 200 alternately. If the peaks are present then the compound can be assumed to be present and its amount estimated. As an example the minimum quantity of the drug imipramine which could be ‘identified’ in this way was 50 pg.” The accurate estimation of the amount of a substance present in these small quantities requires the integration of the intensity of mass fragmentographic peaks with time.Thus in the hypothetical example given above the intensities of the peaks at m/e 100 and 200 would be recbrded at definite time intervals from when they first appeared until they last disappeared. Intensity of peak plotted against time then gives a curve under which the area represents a measure of the amount of compound present. To obtain the most accurate measure- ments the intensity of a peak in the mass spectrum of a standard compound vaporized into the ion source must be measured simultaneously. This method as first introduced,’ has been adversely criticized because considerable fluctua- tions in sensitivity and scattering of quantitative data were found because it was not possible to maintain constant or reproducible conditions in the ion source.” However a small modification was made so that the standard and the compound being investigated were flash-evaporated together into the ion source.The modification led to a great improvement in quantitative estimations as exempli- fied with dansyl derivatives of amines.’2 Integrated mass fragmentography has been used for example in estimations of barbiturates and metabolites in plasma samples of new-born children,’ lidocaine and its metabolite^,^^ heroin,’ ’ prostaglandin^,^^ and biogenic amine~.~ Apart from these advances in the estimation of small quantities of known compounds mass spectrometry has been widely used often with gas chromatog- raphy to identify substances of biogenic interest.Thus sulphonamides which usually decompose in g.c.-m.s. systems5 were successfully studied after N-methylati~n.~’Riboflavin and flavoprotein-derived compounds were identified ” A. E. Gordon and A. Frigero J. Chromatog. 1972 73 401. 5o A. Frigero G. Belvedere F. De Nadai R. Fanelli C. Pantarotto E. Riva and P. L. Morselli J. Chromatog. 1973 74 201. 51 J. R. Majer and R. Perry J. Chem. SOC. (A) 1970 822. 52 N. Seiler and B. Knodgen Org. Mass. Spectrometry 1973 7 97; see also 0. Borga L. Palmer A. Linnarsson and B. Holmstedt Analyf. Letrers 1971,4,837; L. Bertilsson A. J. Atkinson J. R. Althaus A. Harfast J. E. Lindgren and B. Holmstedt Analyt. Chem. 1972,44 1434. 53 G. H. Draffan R. A. Clare and F.M. Williams J. Chromatog. 1973 75 45. 54 J. M. Strong and A. J. Atkinson Analyt. Chem.. 1972 44 2287. 55 G. -R. Nakamura T. T. Noguchi D. Jackson and D. Banks Analyt. Chem..,1972,44 408. 56 J. T. Watson D. Pelster B. J. Sweetman and J. C. Frohlich ref. 3 p. 85. 57 A. A. Boulton D. A. Durden and S. Philips ref. 3 p. 102. 58 D. R. Dill and R. L. McKinley J. Gas Chromatog. 1968 6 68; L. Fishbein J. Chromatag. 1967 30 596. 59 B. Blessington Org. Mass Spectrometry 1972 6 347. 16 R. A. W. Johnstone and F. A. Mellon at the 100nmole and an anthelmintic and its metabolites were studied.'l CI methods are increasingly used in this sort of research as the following range of compound types illustrates macrolide antibiotics,62 sugars63 (where ammonia is recommended as the reagent gas) ceramides and gangli~sides,~~ and insecti- cide~.~~ Carbohydrates have been investigated by electron-impact mass spectrometry either after conversion into derivatives (permethylated peracetylated pertri- methylsilylated) or directly by FI and FD methods.From their mass spectra alone it appears possible to identify fructoses,66 to distinguish between cyclic and acyclic forms,' between furanoses and pyranoses,68 and between aldoses and ketde~,~' and to gather information on the nature of glycosidic linkages7' In the last Annual Report we stressed the possible utility of mass spectrometry in research on peptides obtainable in only very small quantities. Very elegant examples of the use of this method have now appeared with the publication of the structures of several releasing hormones from the hypothalamus.Hormones which inhibit the release of melanocyte-stimulating hormone have been identi- fied as Pro-Leu-Gly-NH and Pro-His-Phe-Arg-GlyeNH ,72 and those which release luteinizing (and follicle-stimulating) growth and thyrotropin-stimulating hormones as respectively (pyro)Glu-His-Trp-Ser-Tyr-Gly-Leu-Arg-Pro-Gly-NH ,7 Val-His-Leu-Ser-Ala-Glu-Glu-Lys-Glu-Ala~OH,74 and (pyro)Glu-His- ProaNH .75 The structure of a mutant peptide from an abnormal haemoglobin has been ~onfirrned.~~ G.c.-m.s. has been used to elucidate the structure of a peptide antibi~tic.~~ Pyrolysis of peptides to form diketopiperazines which are identified by g.c. has been proposed as a method of sequencing peptides7* Initial results on a " P.Brown C. L. Hornbeck and J. R. Cronin Org. Mass Spectrometry 1972 6 1383. 61 W. J. A. VandenHeuvel R. P. Buhs J. R. Carlin T. A. Jacob F. R. Koninszy J. L. Smith N. R. Trenner R. W. Walker D. E. Wolf and F. J. Wolf Analyt. Chem. 1972 44 14. 62 R. L. Foltz ref. 3 p. 258. 63 A. M. Hogg and T. L. Nagebhushan ref. 3 p. 31 1. 64 S. P. Markey R. C. Murphy and D. A. Wenger ref. 3 p. 318. 65 F. J. Biros R. C. Dougherty and J. Dalton Org. Mass Spectrometry 1972 6 1161. 66 K. D. Das and B. Thaynmanavan Org. Mass Spectrometry 1972 6 1063. 67 N. K. Kochetov and 0. S. Chizhov Adv. Carbohydrate Chem. 1966 21 39. H. Ch. Curtins M. Miiller and J. A. Vollman J. Chromatog. 1968 37 216. 69 S. Karady and S.H. Pines Tetrahedron 1970 26,4527. '' J. Karkkainen Carbohydrate Res. 1971 17 1; J. P. Kamerling J. F. G. Vliegenthart J. Vink and J. J. De Ridder Tetrahedron 1971 27 4275. " R. M. G. Nair A. J. Kastin and A. V. Schally Biochem. Biophys. Res. Comm. 1971 43 1376. 'Iz R. M. G. Nair A. J. Kastin and A. V. Schally Biochem. Biophys. Res. Comm. 1972 47 1420. '' A. V. Schally R. M. G. Nair T. W. Redding and A. Arimura J. Biol. Chem. 1971 246,7230. '4 A. V. Schally Y. Baba R. M. G. Nair and C. D. Bennet J. Biol. Chem. 1971,276,6647. '' K. Folkers F. Enzmann J. Boler A. V. Schally and C. Y. Bowers J. Medicin. Chem. 1971 14 469; C. Bogentoft J. Chang H. Sievertsson B. Currie and K. Folkers Org. Mass Spectrometry 1972 6 735. 76 H. R.Morris and D. H. Williams J.C.S.Chem. Comm. 1972 141. " W. A. Konig H. Hagenmaier and E. Bayer 2.analyt. Chem. 1972 259 21 1. 78 A. B. Mauger Chem. Comm. 1971 39. Physical Methods-Part (i) Organic Mass Spectrometry 17 similar approach using g.c.-m.s. for identification of the diketopiperazines have been described.79 7 Metastable-ion Studies Investigations of metastable ions as probes into the nature of transition states for slow decomposition reactions in the mass spectrometer continue. The elimination of keten from p-chloroacetanilide was found to give deuterium isotope effects on competing metastable transitions from which it was concluded that elimination occurred uia a four-membered transition state involving transfer of hydrogen to nitrogen rather than via a six-membered transition state with transfer of hydrogen to the aromatic ring.80 However a cautionary note" has been sounded regarding the characterization of ion structures" by using relative abundances of metastable ions due to two or more competing reactions.Follow- ing experiments on benzylic fluoro-compounds it was pointed out" that ions of identical structure will give a range of intensity ratios because a broad range of internal energies can give rise to metastable decompositions. These workers suggested that changes in the ratio of two or three times could not be considered structurally significant and that much larger changes should be sought before drawing structural inferences. It seems to us that a corollary of this is that two or more different structures could give similar intensity ratios for metastable decompositions especially when the structures are not too dissimilar.Ion structures which are similar could well decompose through a common transi- tion state. In any case there seems no overwhelming reason to suppose that the metastable decompositions bear much resemblance to the faster higher-energy processes occurring in the ion source. 36 A metastable decomposition may occur with release of an excess of internal energy into the reaction co-ordinate. This release of energy causes the peak to broaden. A study of the ratios of this energy release measured by metastable peak widths for loss of hydrogen or deuterium from a number of molecular ions was undertaken and cases in which the activation energy for the reverse reaction was important were identified.83 Ion kinetic energy spectroscopy has been applied to investigations of charge- localization in doubly charged ions of pyrazine pyrimidine and pyrida~ine.~~ The results suggested that the doubly charged ions were linear structures with the charges localized at nitrogen.'' R. A. W. Johnstone and T.J. Povall Meeting of the Protein Group of The Chemical Society Manchester January 1973. N. Uccella 1. Howe and D. H. Williams Org. Mass Spectrometry 1972 6 229. *' K. R. Jennings and A. Whiting Org. Mass Spectrometry 1972 6 921. *' Examples cited in ref. 81 include T. W. Shannon and F. W. McLafferty J. Amer. Chem. SOC.,1966 88 5021; N. A. Uccella I. Howe and D. H. Williams J. Chem.SOC.(B), 1971 1933. 83 M. Bertrand J. €4. Beynon and R. G. Cooks Internat. J. Mass Spectrometry Ion Phys. 1972 9 346. 84 J. H. Beynon R. M. Caprioli and T. Ast Org. Mass Spectrometry 1972 6 273. R. A. W.Johnstone and I;. A. Mellon 8 Conclusion Mass spectrometry in organic chemistry appears to be following the almost classical case-histories of other spectroscopic methods. After years of little interest except to a few physical and petroleum chemists the technique was used by one or two eminent chemists who realized the potential help it could give in structure elucidation. There then followed feverish activity by an increasingly large number of researchers whilst thousands of known compounds were examined to provide empirical guide-lines for understanding mass spectrometric decomposition processes.Lagging slightly behind came increasing improvement and sophistication in instruments and practice. The heyday of the empirical searcher for truth seems to have reached or passed its maximum (if only because there are no more readily available compounds to examine!) leaving mostly the dedicated to worry over mechanisms and theory. However mass spectrometry has been and continues to be taken up enthusiastically as a working tool not only by the organic chemist but also by biochemists and their ilk. Perhaps in a few short years mass spectrometry will be as commonplace and routine as i.r. spectroscopy is to us now but it will not have been without its moments of glory.