2 Physical Methods Part (i) Organic Mass Spectrometry ~~~~~ By R. A. W. JOHNSTONE and F.A. MELLON Dept. of Organic Chemistry University of Liverpool Liverpool L69 3BX Many areas of mass spectrometry have received comprehensive coverage in a recent Specialist Periodical Report ;l this publication is an extensive catalogue of recent work2 and some of the sections such as the one on computer methods give a clear factual account of the current position and trends. To avoid serious overlap with the Specialist Periodical Report this section attempts to assess the influence on organic chemistry of new knowledge and techniques in mass spectrometry. For these reasons the coverage here is selective and certainly not exhaustive in references for which the Specialist Periodical Report should be used.A new introductory book on mass spectrometry has appeared which includes most of the more recent advance^,^" and also one on techniques and applications.3b 1 Theoretical Aspects The empirical application of mass spectrometry to the solution of analytical problems is widespread in organic chemistry. Used empirically the technique is very valuable but a clearer understanding of the theory of mass spectrometry could help its empirical application. Attention to the theory has concentrated in the past mostly on qualitative attempts to explain the fragmentation of mole- cules of known structure in terms of the bond cleavage electron shift and func- tional group approaches long familiar to organic chemists. Knowledge or rationalizations accumulated in this way can then be used to guide the inter- pretation of the mass spectra of unknown compounds (see also Section 5 on the use of computers).The two most widely used of these qualitative theories are based on product-stability arguments4 and charge-localization concept^.^ ‘Mass Spectrometry,’ ed. D. H. Williams (Specialist Periodical Reports) The Chemical Society London 197 1 Vol. 1. R. I. Reed Narure 1971 234 112. (a) R. A. W. Johnstone ‘Mass Spectrometry for Organic Chemists,’ Cambridge University Press 1972; (b) G. W. A. Milne ‘Mass Spectrometry Techniques and Applications,’ Wiley New York 1972. K. Biemann ‘Mass Spectrometry Organic Chemical Applications,’ McGraw-Hill New York 1962 p. 76. H. Budzikiewicz C.Djerassi and D. H. Williams ‘Mass Spectrometry of Organic Compounds,’ Holden-Day San Francisco 1967 p. 9 -27. 5 R. A. W.Johnstone and F. A. Mellon Both theories have their adherents but one tends to be more fashionable6 mostly because 'arrows' or 'fish-hooks' can be drawn to represent the movement of electrons sometimes in an apparently logical manner. A recent article has attempted an unbiased critical review of these theories and their ba~kground.~ The mass spectrum of selenomethionine has been interpreted as a success for charge-localization theory over product-stability arguments,8 and has been cited widely' as a significant success. Extensive energy measurements and more careful investigation of the mass spectra of selenomethionine and of methionine strongly suggest that for these compounds at least charge-localization theory is incorrect although it may be useful for post fact0 rationalization of a mass spectrum.In a more quantitative way the use of quasi-equilibrium theory (QET) to interpret mass spectrometric fragmentation has been tried on two levels one extremely simple the other complex. The full QET" is based on several major assumptions and for its use requires some information which is not easy to obtain. For example the energy distributions in ions the correct vibrational frequencies to use and the possibility of decomposition from excited electronic states are typical parameters which are difficult to assess. Because of these factors the use of QET in polyatomic systems has developed relatively slowly.After some early successes in predicting the mass spectra of alkanes," QET has been used only sporadically. Very recently attempts have been made to 'predict' the observed fragmentation behaviour of some aromatic compounds. ' In view of the known' differences between photon-impact and electron-impact mass spectra the reasons for which have been discussed,' it is unfortunate that these attempts to calculate electron-impact mass spectra' * were based on energy distributions for the ions derived from photoelectron spectroscopy. Electron- impact and photon-impact energy distributions are generally not similar.' At a much less sophisticated level QET can be reduced to the very simple equation (1)14 by which the rates of fragmentation (k)of ions are described in terms of the number of oscillators (N)in the ion the excess of energy (E)over and above the energy required for fragmentation (EJ and a frequency factor (v).k = v[l -E,/EIN-' (1) See any issue of e.g. Org. Mass Spectrometry. ' T. W. Bentley and R. A. W. Johnstone in 'Advances in Physical Organic Chemistry,' ed. V. Gold Academic Press 1970 Vol. 8 pp. 151-269. H. J. Svec and G. A. Junk J. Amer. Chem. Soc. 1967 89 790. T. W. Bentley R. A. W. Johnstone and F. A. Mellon J. Chem. SOC.(B) 1971 1800. lo See M. Vestal in 'Fundamental Processes in Radiation Chemistry,' ed. P. Ausloos Interscience 1968 pp. 59-1 18 for a comprehensive review of current QET applications. l1 H. M. Rosenstock M. B. Wallenstein A. L. Wahrhaftig and H.Eyring Proc. Nur. Acad. Sci. U.S.A. 1952 38 667. l2 F. W. McLafferty T. Wachs C. Lifshitz G. Innorta and P. Irvine J. Amer. Chem. SOC.,1970 92 6867. R. A. W. Johnstone and F. A. Mellon unpublished data. The multiplicity of processes found in 'mono-energetic' electron impact IE curves also support this contention. l4 H. M. Rosenstock A. L. Wahrhaftig and H. Eyring 'The Mass Spectra of Large Molecules Part 11 The Application of Absolute Rate Theory,' Technical Report No. 2 Institute for the Study of Rate Processes University of Utah 25 June 1952. Physical Methods-Part (i) Organic Mass Spectrometry To use this equation to predict mass spectrometric ion abundances [A] the rates of decomposition (k)must be folded with the energy distribution P(E) in the ions [equation (2);assuming no subsequent decomposition of the fragment [A1 = n,(E) * P(E) (2) ion; nA = number of ions A at an energy El.Equation (1)is so oversimplified that it should not be surprising to find that calculations based upon it have needed to adjust often questionably one or more parameters to obtain even modest correlations with observed fragmentations of organic compounds. It is doubtful whether this simple equation will ever be satisfactory for explaining mass spectra except in a quali-quantitative fashion. Attempts to calculate mass spectra with equations (1) and (2) have used ‘adjusted’ v N E P(E) and combinations of these parameter^.'^ The time is ripe to relegate equation (1) to the place it deserves in history except where it can be used in a semi-quantitative descriptive sense.Some of the quali-quantitative predictions based on equation (1) have been predicted qualitatively without its help. l6 2 Ionization and Appearance Potentials Ionization and appearance potentials are important thermochemical quantities. Experimentally the determination of ionization and appearance potentials using a mass spectrometer leaves much to be desired and there is little doubt that many literature values for these quantities are seriously in error. The appearance potential of an ion measured by mass spectrometry may be inaccurate for a number of reasons but probably the most important of these are the ‘kinetic shift’ and the inadequacy of past experimental methods for dealing with curvature at the foot of an ionization efficiency curve.The kinetic shift caused by the fact that an ion must have some energy in excess of the thermochemical appearance potential for fragmentation to be observed in the mass spectro- meter is likely to be important when a fragmentation requires a large activation energy (excess of energy in the ion must be concentrated along the breaking bond) or when the decomposing ion has a large number of degrees of freedom amongst which its excess of energy can be dissipated. It has been suggested that kinetic shift can be estimatedI7 by comparing the appearance potentials of metastable and normal ions. Theoretically this concept seems sound but as applied experimentally,’ the results can be very misleading7 because of inadequacies in the experimental methods used.It has been found that insensitivity in the mass spectrometric detecting system may be a major cause of many observed ‘kinetic shifts’.l9 Is See for example A. N. H. Yeo and D. H. Williams Org. Mass Spectrometry 1971 5 135 and references therein. l6 F. H. Field and J. L. Franklin ‘Electron Impact Phenomena and the Properties of Gaseous Ions,’ Academic Press New York 1957 pp. 78 79. I. Hertel and Ch. Ottinger 2. Naturforsch. 1967 22a 40. ’*J. H. Beynon J. A. Hopkinson and G. R. Lester Internat. J. Mass Spectrometry Ion Phys. 1969 2 291. l9 R. A. W. Johnstone and F. A. Mellon unpublished data. 8 R. A. W.Johnstone and F. A. Mellon Mathematical analytical methods are available” for dealing with curvature at the foot of an ionization efficiency curve.This curvature is made up from an approximately Maxwellian spread of energies in the electrons from a heated filament together with ‘true’ curvature caused by closely separated or overlapping electronic and vibronic states. The mathematical methods of analysing experi- mental ionization efficiency curves remove most of the effects of the spread in electron energies and leave the ‘true’ curvature; in ideal situations this ‘true’ curvature is found to be compounded of many short straight lengths. Care is still necessary with some of these mathematical methods one in particular (Morrison’s second-derivative technique) has been shown theoretically to give positive errors of up to 0.35 eV in some cases.21 In experiments utilizing a computer to gather and interpret data appearance and ionization potentials seem much more accurately measurable22 than by older methods.An added advantage of this computer-aided method is the tremendous reduction in time needed for the measurements. Using these computer-analytical methods accurate heats of formation and evidence for the existence or otherwise of kinetic shifts can be obtained.’ Other workers have reported similar computer- aided methods for the acquisition of ionization efficiency curves,’ and one 3924 has used appearance potentials found by such a method as a distinguish- ing factor between structural isomers. 3 Ionization Methods The appearance of the mass spectrum of any compound depends on the excess of energy (electronic and vibronic) left in the molecular ion after the ionization step because the excess of energy eventually gives rise to the observed frag- mentation.Thus different methods of ionization can have large effects on the mass spectrum. So far this discussion has centred on electron-impact induced fragmentations of ions as this is still the most widespread method of ionization used in the mass spectrometry of organic compounds. However there have been significant advances in other methods of ionization and these hold con- siderable promise for future general use in organic mass spectrometry. Photo-ionization has been used for a long time and a recent review has discussed its applications.2s A disadvantage of photo-ionization which has held up its general application is the low ion yield due to the relatively low intensity of light which can be directed into the ion-source.The molecular ion in a photo- ionization mass spectrum is usually abundant and the number of fragment ions much reduced compared with electron-impact mass spectra. Mass spectra 2o J. D. Morrison J. Chem. Phys. 1953 21 1767; 1963 39 200; R. E. Winters J. H. Collins and W. L. Courchene ibid. 1966 45 1931. 21 G. G. Meisels and B. G. Giessner Internat. J. Mass Spectrometry Ion Phys. 1971 7,489. 22 R. A. W. Johnstone F. A. Mellon and S. D. Ward Internat. J. Mass Spectrometry Ion Phys. 1970 5 24 1. 23 M. L. Gross and C. L. Wilkins Anafyt. Chem. 1971 43 1624. 24 R. G. Dromey J. D. Morrison and J.C. Traeger Internat. J. Mass Spectrometry Ion Phys. 1971 6 57. 25 N. W. Reid Internat. J. Mass Spectrometry Ion Phys. 1971 6 1. Physical Methods-Part (i) Organic Mass Spectrometry 9 produced by photon-impact may be much simpler to interpret and could be obtained more widely if suitable light sources were available. Field ionization in which a steep electrical potential gradient enables quantum tunnelling of an electron from a molecule to an electrode to occur continues to make steady progress,26 and ion sources based on this technique are available commercially. The excess of energy imparted to a molecular ion during field ionization is very low so that fragmentation is greatly reduced compared with electron-impact ionization at 70 eV and abundant molecular ions are obtained.Sometimes the abundance of fragment ions is embarrassingly small for structural analysis and the commercial ion sources are designed to operate either in the field or electron-impact ionization mode. The spectra of a compound ionized under both conditions can then be compared. Another way in which fragmen- tation of molecular ions produced by field ionization could be induced is by ion-molecule collisions.27 Here the molecular ions are passed through a region of gas at sufficiently high pressure for ion-molecule collisions to occur and the collisional energy appears as vibrational energy leading to fragmentati~n.~~ A disadvantage of field-ionization compared with electron-impact methods has been the low usable ion yield but it is now reported that with ‘super-emitting’ electrodes the ion yields of the two methods are comparable.** A second disadvantage of field ionization applies also to electron-impact sources viz.the sample must be introduced into the source as vapour and therefore thermal energy must be supplied before ionization. Especially for labile aliphatic com- pounds or compounds which are difficult to volatilize this extra thermal energy can lead to excessive fragmentation to such an extent that no molecular ions are observed and usable information from fragment ions is nil. The problem has been overcome in field-ionization sources by the field-desorption tech- nique.” The substance to be investigated is applied as a very thin coating to the field-ionizing electrode.No heat is required to vaporize the sample and ions formed in the condensed phase at the electrode are shot out into the mass spectro- meter. Results obtained in this way with compounds which are difficult to deal with by electron-impact mass spectrometry are impressive. 30 For example the electron-impact mass spectrum of glucose has no molecular ion and few fragment ions of diagnostic value ; field-ionization yields a molecular ion and many fragment ions ; field-desorption yields an abundant molecular ion with a few fragment ions in low abundance.29 Further developments of this technique are awaited with interest particularly by those research workers who need to analyse such mass spectrometically difficult compounds as sugars carbohydrates glycosides peptides and nucleotides.26 See H. D. Beckley Angew. Chem. Internat. Edn. 1969 8 623 for a comprehensive review. K. R. Jennings Internat. J. Mass Spectrometry Ion Phys. 1968 1 227. 28 H. D. Beckley H. Hilt A. Maas M. D. Mighaed and E. Ochterbeck Innternat. J. Mass Spectrometry Ion Phys. 1969,3 161 ;M. D. Mighaed and H. D. Beckley ibid 1971,7 1. 29 H. D. Beckey Innternat. J. Mass Spectrometry Ion Phys. 1969 2 500. ’”H. Krone and H. D. Beckey Org. Mass Spectrometry 1971 5 983. 10 R. A. W.Johnstone and F. A. Mellon Another method of ionization making rapid progress and which promises to be of great value for labile compounds is chemical i~nization.~' In this technique a reactant gas ion (commonly CH,+ from methane) ionizes a molecule (M) by proton exchange to yield a 'quasi-molecular' ion (MH') as shown in equation (3).Various other reactant gas ions have been used e.g. argon ethane CH5' + M -+ CH + MH' (3) propane and i~obutane.~~ Reactant gas ions like CHsf are thought to be essentially thermally equilibrated by collision with CH molecules before they collide with and ionize the molecule (M) under investigation. Fairly closely defined amounts of energy are transferred from the reactant gas ion to the quasi-molecular ion during its formation. The quasi-molecular ions are even- electron species containing little excess of vibrational energy and fragmentation of these MH' ions is greatly reduced compared with the Mt ions formed by electron impact. Generally there is enough excess of energy in the quasi- molecular ions to give moderately abundant fragment ions for structural analysis ; these chemical-ionization mass spectra are notable for their abundant quasi- molecular ions.Again the sample must be vaporized into the ion-source for chemical ionization which can be detrimental to labile compounds. Nevertheless very encouraging results are being obtained as illustrated by the reported mass spectra of oligopeptide~.~~ When the reactant gas ion is monatomic (e.g.argon) a precisely defined amount of energy is transferred during ionization [equation (4)]. The molecular ion produced in this case is an odd-electron species not a quasi-molecular ion and Af+M-+Mf+A (4) the excess of energy it contains is the difference between the ionization potentials of argon and the molecule (M).Commercial chemical-ionization sources are available. Ion-molecule reactions utilizing hydrated species (H904' and D904') have been used recently as an alternative ionizing technique to obtain the amino-acid sequence of a ~eptide.~~ The method was observed to give a large protonated 'molecular ion' and intense sequence ions (compare remarks on increasing fragmentation through ion-molecule reactions in the section of field ionization above). F. H. Field Accounts Chem. Res. 1968 1 42. 32 F. H. Field P. Hamlett and W. F. Libby J. Amer. Chem. SOC.,1967 89 6035; J. Michnowicz and B. Munson Org. Mass Spectrometry 1970 4 1481; F. H. Field J. Amer. Chem. Soc. 1969 91 2827. 33 A.A. Kiryushkin H. M. Fales T. Axenrod E. J. Gilbert and G. W. A. Milne Org. Mass Spectrometry 1971 5 19. 34 R. J. Beuhler L. J. Green and L. Friedman J. Amer. Chem. SOC.,1971 93 4307. Physical Methods-Part (i) Organic Mass Spectrometry 4 Gas Chromatographic-Mass Spectrometric (GC-M.S.) Methods Gas chromatography coupled with mass spectrometry is a valuable analytical technique for dealing with small quantities of mixtures. Usually the mixture is separated completely or partially by standard gas-chromatographic techniques and the separated components are identified from their mass spectra and retention times. The separated components can be condensed out of the gas-chromato- graphic column effluent and introduced into the mass spectrometer by the usual inlet systems.A more efficient use of g.c.-m.s. methods is obtained by passing the already volatilized separated components directly from the g.c. column into the mass spectrometer. By using a chemical-ionization source and methane as the carrier gas in the g.c. apparatus the carrier gas becomes the reactant gas in the ion source.35 Chemical-ionization sources are designed to operate at higher pressures than electron-impact sources and for the latter it is necessary to remove as much g.c. carrier gas as possible before it reaches the source. The carrier gas may be separated from entrained organic components by using a molecular separator of which there are now several types known and some of them commercially available. A recent article has reviewed the desirable and undesirable features of these separat01-s.~~ A major advance which occurred recently has been the successful coupling of a gas chromatographic apparatus to a field-ionization source.37 The combined g.c.-m.s.method is finding ever wider use as the evidence of a rapidly increasing number of research publications shows.38 Articles and reviews of the subject have appeared3’ and there is no call to cover the same ground here. It is pertinent to note that complex mixtures analysed in this way would take many times longer if tackled by other methods. Particularly attractive are the advances which can now be made in investigations of drug metabolism on a micro-~cale.~~ An instrumental advance of importance for g.c.-m.s. appli- cations has been the development of a ‘multiple ion detector’ system capable of recording three different masses simultaneously.This has also been converted into a peak-matching device capable of a ‘dynamic resolution’ of 70000 for a static resolution of 5-00 in a single-mass spe~trometer.~~ 5 Computers The use of computers to acquire accurate mass data is now well established so that with one sweep of the mass spectrum it is possible to obtain not only the molecular formula of a compound (elemental composition of molecular ion) 35 G. P. Arsenault J. J. Dolhun and K. Biemann Chem. Comm. 1970 1542. 36 G. A. Junk Internat. J. Mass Spectrometry Ion Phys. 1972 8 1. 37 P. G. Burkhalter Analyt. Chem. 1971 43 17. 38 See for example numerous publications in J.Chromatography and J. Sci.Instr. 39 C. J. W. Brooks pp. 288-307 in ref. 1 ; see also ref. 36. 40 C. J. W. Brooks A. R. Thawley P. Rocher B. S. Middleditch G. M. Anthony and W. G. Stilwell in ‘Advances in Chromatography,’ ed. A. Zlatkis Chromatography Symposium University of Houston Houston 1970 p. 262. 41 C.-G. Hammar and R. Hessling Analyr. Chem. 1971 45 298 307. 12 R. A. W.Johnstone and I;. A. Mellon but also the elemental compositions of all fragment ions. This kind of information is of considerable value in elucidating the fragmentation of complex molecules. The mass spectrum of a compound with even a modest molecular weight may contain many fragmented ions for each of which the computer data-acquisition system produces one or more elemental compositions.Such a list of elemental compositions is not easy to sift for information of use in structural analysis. For this reason it was proposed that element maps should be con~tructed.~~ These element maps contain the elemental compositions arranged in columns of increasing complexity and simplify the task of information searching. Little use has been made of element maps and it is difficult to evaluate their general usefulness. Now that mass spectrometric data-acquisition systems are more widely distributed (several universities in the U.K. possess them) more people may make use of element maps and make a proper evaluation possible. In a slightly less sophisticated manner routine mass spectra may be stored by a com- puter and compared with spectra already held in the computer memory bank.With a large number of stored spectra it is then possible to identify a compound from its mass spectrum if the mass spectrum has been obtained previously and stored.43 A recent theoretical investigation using the principles of information theory has implied that a mass spectrum can be encoded in 48 bits of information or two words of core store in a computer with a word length of 24 bits.44 This means that a relatively small computer could be used to conduct library searches. At a more advanced level computers are being made not simply to compare spectra but to recognize patterns. The computer is programmed to abstract patterns when presented with the low-resolution mass spectra of a group of similar known compounds.Information acquired by this technique is then applied to unknown compounds which the computer will hopefully classify according to patterns ‘learnt’ during the ‘training’ procedure. This is the learning machine approach so-called because information can be continuously updated and the probability of correct assignments correspondingly increased the larger the number of compounds examined.45 A recent improvement in the classifica- tion of low-resolution mass spectral data has resulted in a predictive ability of up to 98%.46 As a further example the fragmentation behaviour of ketones can be rationalized in terms of the functional keto-group and other attached groups and the information stored in the computer. Examination of the mass spectrum of an unknown ketone by the computer and comparison with the stored information from known ketones enables it to make deductions of likely structure.In a similar fashion the computer can be programmed to distinguish between for example ketones ethers and amines. Initial results seem 42 K. Biemann P. Bommer and D. M. Desiderio Tetrahedron Letters 1964 1725. 43 S. A. Abrahamson Sci. Tools 1967 1429; H. S. Hertz R. A. Hites and K. Biemann Analyt. Chem. 1971 43 681. 44 L. E. Wangen W. S. Woodward and T. L. Isenhour Analyt. Chem. 1971,43 1605. 45 T. L. Isenhour and P. C. Jurs Analyt. Chem. 1971 43 20A reviews many aspects of learning machines applied to chemistry (including mass spectrometry). 46 P. C. Jurs Analyt. Chem. 1971,43 22. PhysicaI Methods-Part (i) Organic Mass Spectrometry 13 promising47 but there is still a very long way to go before automatic positive structural analysis is possible by artificial intelligence.Another area in which full use can be made of the exceptional data storage ability of computers lies in g.c.-m.s. work which generates a great deal of infor- mation. After a substance has emerged from a g.c. column its mass spectrum is obtained and then stored on a magnetic disc or tape ;the spectrum may then be recalled from the store at will. In this way a complete gas-chromatographic analysis can be completed and all the mass spectra stored for printing out or display later at leisure.48 It is also possible to programme the computer to search for the mass spectrum of a specified compound amongst all those held in its store without having all the mass spectra printed out.The use of computers for data acquisition in connection with ionization and appearance potential measurements has already been discussed above. Also it should be mentioned that the mathematical analysis techniques described earlier can be used to abstract high-resolution data from low-resolution ~pectra.~’ 6 Metastable Ions In any fragmentation process m -+ rn; the parent ion (rn,f) can decompose to the daughter ion (m:) anywhere in the mass spectrometer. With conventional mass spectrometers of the type generally used by organic chemists this frag- mentation is almost complete for many m ions before they leave the ion source and only rn; ions are observed.Some m ions have insufficient excess of energy to decompose and appear as m,f ions. These m mi ions can be termed ‘normal’ ions. However some m,f ions have just sufficient excess of energy to decompose to mi between the ion source and collector. Because the momentum of m is shared between its fragmentation products these mi have less translational energy than ‘normal’ m; ions and appear in a different place in the mass spectrum. The my ions with less than the full translational energy are the inaptly named metastable ions and they may be observed as broad peaks of low abundance. In a routine mass spectrum the apparent mass of a metastable ion (m*)is given by the expression m* = mz2/rn1 where m, m2 are the masses of the ions m rnz.These metastable ions are very useful for deciphering fragmentation pathways. Metastable abundance ratios for production of C3H8N+ ions together with 13C and deuterium labelling have been used to classify ion structures from fifteen different isomers.50 It was claimed to be possible to distinguish five structurally distinct ions. Metastable ions have also been used to observed deuterium isotope effects in mass spectra. 51 Many more metastable transitions can be observed in double-focusing mass spectrometers by the technique of 47 A. M. Duffield A. V. Robertson C. Djerassi B. G. Buchanan G. L. Sutherland E. A. Feigenbaum and T. Lederberg J. Amer. Chem. SOC.,1969 91 2977. ‘* R. Binks R. L. Cleaver J. S. Littler and J. Macmillan Chem. in Britain 1971 7 8 reviews this and other aspects of low-resolution mass spectra data acquisition.49 R. G. Dromey and J. D. Morrison Internat. J. Mass Spectrometry Zon Phys. 1971,6,253. N. Uccella I. Howe and D. H. Williams J. Chem. SOC.(B),1971 1933. ” I. Howe and D. H. Williams Chem. Comm. 1971 1195. 14 R. A. W.Johnstone and F. A. MeIlon ‘defocusing’ the normal ion beam.52 In defocusing metastable ions (m;)formed after rn; leaves the ion source but before it enters the field of the electric sector can be detected by altering the accelerating potential at the ion source but keeping the potential across the sector plates constant. Normal ions are then no longer focused by the electric sector but metastable ions are. In a variation on this technique the accelerating voltage at the ion source is held constant and the electric sector voltage is varied.By putting a second electron multiplier in the path of the ion beam the metastable ions can be detected ; this method has been termed ion kinetic energy (i.k.e.) spectro~copy.’~ It is beyond the scope of this article to discuss i.k.e. spectroscopy at length but many metastable ions can be observed by its use and of particular interest to the organic chemist it appears that small structural differences between isomers are dete~table,’~ which is often not the case from an examination of routine mass spectra. Mention may be made of a new device for enhancing metastable ions which makes use of the different translational energies of these and normal ions.” 7Mass Spectra of Organic Compounds Data continue to accumulate on the mass spectra of a wide variety of compounds’ and wider-ranging attempts are being made to rationalize and correlate frag- mentation behaviour.56 Some classes of compounds are receiving particularly marked attention because of their biological importance. Thus the identification of prostaglandins can be made by g.c.-m.s. method^,'^ and mass spectra of nucleotides have been ~btained.’~ There is heavy emphasis on the mass spectro- metry of peptides because it is possible to obtain the sequence of amino-acids in a peptide from its mass spectrum. This specialized topic was of sufficient importance for a European summer school to be ~rganized.’~ The ability to extract an amino-acid sequence from one mass spectrum is attractive but there have been and still are difficulties preventing the completely successful application of mass spectrometry to this problem.Not least of these difficulties is one of scale. Peptide chemists and biochemists are accustomed to handle nanomole amounts of peptides and are unlikely to be much impressed with results obtained with millimole quantities. For example 10 millimoles of a peptide of molecular weight 1000 is 10mg which is ‘bucket-scale’ in peptide work and very routine even for mass spectrometry. 52 M. Barber and R. M. Elliot ASTM Committee E14 Conference on Mass Spectrometry Montreal June 1964; K. R. Jennings J. Chem. Phys. 1965 43 4176. 53 J. H. Beynon J. W. Amy and W. E. Baitinger Chem. Comm.1969 723. 54 E. M. Chait and W. B. Askew Org. Muss Spectrometry 1971 5 147. s5 A. McCormick N. R. Daly R. E. Powell and R. Hayes 5th Mass Spectroscopy Group Meeting University of Bristol July 197 1. 56 T. W. Bentley and R. A. W. Johnstone J. Chem. Sac. (B),1971 1804. ’’ E. Granstrom and B. Samuelsson J. Amer. Chem. SOC. 1969 91 3398. 58 A. M. Lawson R. N. Stillwell M. M. Tacker K. Tsuboyama and J. A. McCloskey J. Amer. Chem. SOC.,1971 93 1015. 59 European Molecular Biology Organisation Meeting Paris September 1971 ; (a) reported by R. A. W. Johnstone. Physical Methods-Part (i) Organic Mass Spectrometry 15 Another difficulty in the mass spectrometry of peptides is that of volatilizing them into the ion-source and from this point of view field-desorption experiments may prove particularly valuable.Generally peptides with approximately 8-12 amino-acid residues will give good mass spectra provided the peptide has been suitably modified. The zwitterionic character of a peptide is removed by acylation (acety16' or ethoxycarbony16 I) of the N-terminal and methylation of the C-terminal. It is then often necessary to N-methylate the amide bonds to reduce intermolecular hydrogen bonding and increase volatility.6 Several methods of N-methylation have been described and one recommended for general use for all but arginine residues.63 Despite all these researches only one peptide structure has genuinely been deduced from just the mass spectrum.64 A second example is of a somewhat dubious nature because extensive pyrolysis of the peptide occurred on its introduction into the mass spe~trometer.~~ Cyclic peptides frequently are much more volatile than linear ones and can be volatilized without the preparation of derivatives.However because they are cyclic and may open indiscriminately on fragmentation it is often difficult to decipher the amino-acid sequence. Another method for sequencing long peptides is to break them down chemi- cally or enzymically into shorter peptides. These shorter usually di- or tri- peptides are separated and identified by g.c.-ms. analysis of their trifluoroacetyl derivatives.66 From the total information obtained it is possible to build up the structure of the original peptide as was done with fungal antitoxin^.^^ Other experiments along these lines but using acetylated derivatives have been des- cri bed.9a 8 Miscellaneous Some recent mechanistic studies of note have included attempts to evaluate factors influencing rates of fragmentation in ortho-disubstituted benzenes using qualitative QET arguments.68 Steric effects were documented in the same fashion as the difference between appearance and ionization potentials has been used to describe electronic effects. Steric effects have also been included in a system of 'interaction values' used to explain the observed variations of ion intensities in the mass spectra of tri- methylsilylated pyran~ses.~~ 'O A. Prox and K. K. Sun Z. Naturforsch. 1966 tlb 1028. 61 J. D. Kamerling W. Heerma and J. F. G. Vliegenhart Org.Mass Spectrometry 1968 1 351. 62 B. C. Das S. D. Gero and E. Lederer Biochem. Biophys. Res. Comm. 1967 29 211. 63 G. Marino L. Valente R. A. W. Johnstone F. Mohammedi-Tabrizi and G. C. Sodini Chem. Comm. 1971 64 K. L. Agarwal G. W. Kenner and R. C. Sheppard J. Amer. Chem. SOC. 1969,91,3096. " D. M. Desiderio G. Ungar and P. A. White Chem. Comm. 1971 432. '' F. Weygand A. Prox E. C. Jorgensen R. Axen and P. Kirchner Z. Naturforsch. 1963 18b 93. 67 A. Prox J. Schmid and H. Ottenheym Annalen 1969,722 179. " S. A. Berezra and M. M. Bursey J. Chem. SOC.(B) 1971 1515. 69 S. C. Havlicek M. R. Brennan and P. J. Scheur Org. Mass Spectrometry 1971,5 1273. R. A. W.Johnstone and F. A. Mellon An overwhelming majority of publications in organic mass spectrometry is concerned with the mass spectrometric fragmentation of positive ions and little work has appeared concerning negative-ion spectroscopy.Early experiments in negative-ion mass spectrometry were unpromising because negative ions are usually not easy to produce by electron impact. Even when negative ions can be produced there is less fragmentation and what elimination reactions there are tend to be more complex than those of positive ions. More recent work has not changed this impression which is exemplified by the elimination of HNO and HF from p -ni trophen y1trifluoroace tamides. Ion cyclotron resonance (i.c.r.) spectroscopy in which the ions have much longer lifetimes than those produced in conventional mass spectrometers has been re~iewed.~’ The longer ion lifetimes mean there are greater chances of ion-molecule collisions occurring and i.c.r.is proving of considerable value for studying ion-molecule reactions. Mechanistic studies into ion structures have been pursued by utilizing these ion-molecule reactions and particular emphasis has been placed on studies of the structures of ions resulting from the six-centre rearrangement of ketones. These particular rearrangement ions at m/e 58 [equation (5)]could possibly have a number of structures but the enol and enol mje 58 keto keto forms are most favoured” and of the two the available evidence of i.c.r. spectroscopy supports the enol structure. The evidence is based on the ion- molecule reactions undergone by the ion at m/e 58 compared with those of the molecular ion of acetone and with those of the fragment ion at m/e 58 from methylcyclobutanol [equation (6)].Similar experiments have been carried out with other compounds.73 m/e 58 Other i.c.r. investigations include the perennial problem of the mechanism of ‘tropylium ion’ formation in the mass spectrum of benzyl compounds. The reactivity of C7H,f ions in the mass spectrum of toluene was shown to differ 70 J. H. Bowie Austral. J. Chem. 1971 24 989. G. C. Goode R. M. O’Malley A. J. Ferrer-Correia and K. R. Jennings Chem. in Britain 1971 7 12. 72 J. Diekman J. K. MacLeod C. Djerassi and J. D. Baldschwieler J. Amer. Chem. Soc. 1969,91,2069. 73 G. Eadon J. Diekman and C. Djerassi J. Amer. Chem. SOC.,1969 91 3986. Physical Methods-Part (i) Organic Mass Spectrometry from that of norbornadiene or cycloheptatriene in its ability to abstract NO from alkyl nitrates.74 Toluene ions underwent the ion-molecule reaction C7H8++ RONO -+ C,H,NO2+ + RO-whereas the other isomers would not give this reaction.Similarly structural isomers of C,H ions from a variety of sources have been distinguished by their ability to react with amm~nia.’~ to carry Initial attempts have been rep~rted’~ out stereochemical analysis by examining the differing ion-molecule reactions of stereochemical isomers using i.c.r. It was possible to distinguish exo- and endo-borneol from the ion-molecule reactions between their molecular ions and biacetyl. Purely analytical applications of mass spectrometry have reached an advanced level in the work reported on the identification of constituents of human blood and tissues on a very small scale.77 Quantitative high-resolution mass spectro- metry was used to simultaneously determine the amounts (pg) of up to five constituents in about 5mg of desiccated tissue.This sort of work opens up research intro drug metabolism and suggests a possible extension to biosynthetic studies in plants where the simultaneous monitoring of several constituents needs to be carried out over a period of time. ” M. K. Hoffmann and M. M. Bursey Tetrahedron Letters 1971,27,2539;M. M. Bursey M. K. Hoffmann and S. A. Benezra Chem. Comm. 1971 1417. l5 F. W. McLafferty and M. L. Gross J. Amer. Chem. Sac. 1971,93 1267.l6 M. M. Bursey and M. K. Hoffmann Canad. J. Chem. 1971,49 3395. l7 W. Snedden and R. B. Parker Analyt. Chem. 1971,43 1651.