2 Physical Methods Part (i) Organic Mass Spectrometry By R. A. W. JOHNSTONE The Robert Robinson Laboratories The University LiverpoolL69 3BX and F. A. MELLON Agricultural Research Council Unit of Invertebrate Chemistry and Physiology University of Sussex Brighton. BNI 9QJ This our final Report continues the practice adopted in the two previous ones of covering selected areas of mass spectrometry of general interest to organic chemists. A good impression of trends in mass spectrometry can be obtained from the Proceedings of the 6th International Mass Spectrometry Conference,' which should be available in the near future as Volume 6 of Advances in Mass Spectrometry. 1 Theoretical Aspects In earlier Reports,' we have emphasized the limitations of the over-simplified quasi-equilibrium theory equation k = v[(E -Eo)/EIN-',when used for any- thing other than simple teaching or illustrative purposes.This emphasis was simply a reiteration of earlier objections to this equation3 which often seem to have been conveniently overlooked without tacit or implicit justification. The equation is useful for illustrating mass spectrometric fragmentation effects in a quali-quantitative sense. There has been a decline in the use of the simple equation in recent literature possibly because all the parameters which can be varied have now been examined. Nevertheless the equation has been used to calculate tables of estimated kinetic shifts although the authors readily admit the drawbacks. Since the rate given by the equation needs to be folded with the energy distribution function for the ions and the latter is an unknown quantity it is doubtful whether the calculated values provide even an approximate guide to the kinetic shift.Because there is no one rate of ion decomposition all mass spectrometric frag- mentations will yield a spread of rates from zero to about lot4(bond vibration ' Held in Edinburgh. September 1973. * R. A. W. Johnstone and F. A. Mellon Ann. Reports (B) 1972,69 7. See for example A. G. Harrison 'Topics in Organic Mass Spectrometry'. ed. A. L. Burlingame Wiley-Interscience New York 1970 Vol. 8 p. 128. S. Benezra and M. M. Bursey Org. Mass Spectrometry 1973 7 241. 7 R. A. W.Johnstoneand F. A. Mellon frequency) and therefore an ‘observed’ kinetic shift depends on the ability of the instrument to measure the corresponding ion currents.Apart from the ion flight time the sensitivity of the instrument is very important. Measurement of a kinetic shift by obtaining the appearance potentials of normal and metastable ions is considered later. With the decline of interest in the simple QET equation there has been an increase in the number of applications of molecular orbital theory to mass spectrometry. The methods used in these calculations vary from the extremely nai‘ve to the more sophisticated CNDO and INDO approaches but as yet the highly sophisticated ub initio SCF IBMOL and Xa-SW methods which have shown great reliability with inorganic ions have not been used. Perhaps someone with access to the programs could be persuaded to carry out a few calculations.At the simplest MO level for cleavage of a substituent from the para position of 4-styrylquinolines the calculated bond order for the frontier electron correlated linearly with log I where I = the total abundance of the substituent and M -substituent ions;’ it would be interesting to know whether similar correla- tions are obtained between log I and say ionization potentials. At a more sophisticated level extended Huckel theory was used to calculate electron densities and bond orders for oestrone as the neutral molecule and ground-state and excited-state molecular ion.6 These calculations showed that in all the con- figurations considered for the molecular ion the electronic charge was delocalized over the whole structure and more so than in the neutral molecule.These findings are in keeping with earlier qualitative theoretical conclusion^,^ and suggest that net charges are unrelated to fragmentation. However bond densities in the molecular ion did allow a prediction of the gross features of the mass spectrum of oestrone and it was concluded that ionization weakened some bonds near positions of high electron density in the outer molecular orbitals of the ion.6 Just such an argument has been used previously in an attempt to de-polarize the ‘charge localization’ and ‘product stability’ extremes.’ Unfortunately the reliability of the calculations must be questioned because the outermost molecular orbital of oestrone is predicted to be associated largely with the carbonyl function in ring D whereas photoelectron spectroscopy suggests that the outer orbital should be 7r-type and associated mostly with the aromatic ring (c$ the ionization potentials of phenol 8.54 eV,9 and cyclopentanone 9.25 eV’’).It may be that like CNDO calculations extended Huckel theory gives a better guide to charge distributions and not to orbital energies although it has been claimed this is not the case for one Huckel treatment.’ I H. Gusten L. Klasinc and D. Stefanovic Org. Mass Spectrometry 1973 7 1. G. Loew M. Chadwick and D. Smith Org. Mass Spectrometry 1973 7 1241. M. J. S. Dewar Tetrahedron Suppl. 1963,19,89; T. W. Bentley and R. A. W. Johnstone Adv. Phys. Org. Chem. 1970 8 154. R. A.W. Johnstone ‘Mass Spectrometry for Organic Chemists’ Cambridge University Press 1972 pp. 3945. R. A. W. Johnstone and F. A. Mellon J.C.S. Faraday II 1973 69 36. lo D. Chadwick D. C. Frost and L. Weiler J. Amer. Chem. SOC., 1971 93 4320. I’ P. A. Cox S. Evans A. F. Orchard N. V. Richardson and P. J. Roberts Faraday Discuss. Chem. SOC. 1972 No. 54 p. 26. Physical Methods-Part (i) Organic Mass Spectrometry 9 INDO and CNDO calculations have been carried out on substituted phenyl- acetates and the calculated bond orders were interpreted as suggesting bond formation between the carbonyl oxygen and an ortho-fluorine and hence a ‘lower frequency factor’ for elimination of C,H,O from the molecular ion. l2 Although it is not clear in the paper,12 the authors appear to have used the CNDO and INDO density matrix Ppv to obtain the bond order.In looking for bond formation this approach can lead to unreliable results because Ppv is derived from the electron densities at the atoms (p v) and does not take into account the degree of positive overlap (Spv)between them. l3 It would be interesting to know if the same result was obtained from a population analysis13 obtained by multiplying and summing the correct off-diagonal elements of the density and overlap integral matrices i.e. ~Pp~.Spv.’~ As has been found in many other applications of INDO and CNDO techniques,15 the predicted ordering of 71-and a-levels seems unreal. From a further application of INDO it was suggested that scrambling of hydrogen atoms in the molecular ion of ethane took place uia a diborane-type bridge structure (1) with hydrogen atoms on opposite sides of the C-C bond rather than when they are on the same side as in structure (2).16 H H H H,H ,/NH\\ H\ / ‘\ I I /’ c-c c-c / ‘ I’ \H /\ H ‘H H H INDO calculations also predict near co-planarity for the aryl rings in two para-arylacetophenones and therefore the enhanced resonance stabilization nay account for the low abundance of C,H,O ions in their mass spectra.” Finally an attempt has been made to evaluate mass spectrometric fragmentation in terms of the total energies of the fragments as calculated by CNDO and hence the lowest energy for fragmentation.Given Hammond’s postulate’ * the approach seems valid within the limitations of CNDO at which point it is perhaps wise to recall Dewar’s warning on the use of molecular orbital methods to predict small energy changes uiz.the accuracy to be expected from estimating a small energy change as the difference between two large inaccurate energies can only be low. l9 12 C. E. Parker J. R. Has M. M. Bursey and L. G. Pederson Org. Mass spectrometry 1973 7 1189. 13 R. S. Mulliken J. Chem. Phys. 1955 23 1841; ibid. 1962,36 3428. I4 J. A. Pople and D. L. Beveridge ‘Approximate Molecular Orbital Theory’ McGraw- Hill New York 1970 p. 43. 15 J. E. Bloor and D. L. Breen J. Phys. Chem. 1968,72,716; see also ref. 11 and R. A. W. Johnstone and F. A. Mellon J.C.S. Furaduy IZ 1973 69 1155. 16 C. E. Parker M. M. Bursey and L.G. Pederson Org. Mass Spectrometry 1973 7 1077. 17 C. E. Twine. C. E. Parker and M. M. Bursey Org. Muss. Spectrometry 1973 7 1179. I8 G. S. Hammond J. Amer. Chem. Soc. 1955,77 334. 19 M. J. S. Dewar in ‘Aromaticity‘ Special Publication No. 21 The Chemical Society London 1967 p. 199. R.A. W.Johnstone and F. A. Mellon We add to this a reiteration of the poor energies generally obtained using standard CNDO methods. l5 Before leaving this discussion of theoretical models of mass spectrometric fragmentation it is worth emphasizing that almost all applications of simple and full quasi-equilibrium theory equations have explicitly or implicitly assumed that the density of states does not include excited ionic species i.e. that a short time after ionization the molecular ion has converted excess of electronic excitation energy into excess of vibrational energy shared amongst all the oscillators and rotamers and therefore that fragmentation occurs from a vibrationally and rotationally excited ground-state ion.The first experimental body-blow to this cosy assumption has been dealt by the observation from photoelectron-mass spectrometric coincidence techniques that ions in the first excited-state of the C,F cation decompose without prior internal conversion to the ground- state ion.” 2 Ionization and Appearance Potentials These potentials are of considerable value in giving thermochemical data and in providing information on the energetics of mass spectral fragmentations. There is still confusion in the literature over both the significance and the use of adiabatic and vertical ionization potentials and this has led to frustration manifesting itself in print.21 Quite simply for most thermochemical arguments the adiabatic potential should be used.The confusion arises over the data available in the literature because it is not certain exactly what the various mass spectrometric methods measure. Earlier measurements of ionization and appearance potentials using a mass spectrometer have frequently given considerably higher values than have similar measurements made in photo-ionization studies and this led to a belief that the electron-impact methods measured some nebulous ‘vertical’ ionization potential. That some of these differences were due simply to instru- mental insensitivity and the poor methods of evaluating ionization efficiency curves was made apparent in a recent paper.” There is therefore reason to believe that given a sufficiently sensitive method of gathering ionization efficiency data and a proper means of evaluating the data ionization potentials measured by electron or photon impact would be identical and adiabatic ones.The literature has been further confused by another use of the term ‘vertical’ ionization potential in photoelectron spectroscopy. This vertical potential is the name given to the position of a maximum on a band in the photoelectron spectrum and is of dubious significance since it measures a maximum arising from many overlapping processes in polyatomic molecules and alters with changes in the resolving power of the instr~ment.’~ It is by no means certain and is probably highly doubtful that electron-impact and photoelectron ‘vertical’ ionization potentials are equivalent.2o I. G. Simm C. J. Danby and J. H. D. Eland J.C.S. Chem. Comm. 1973,832. * See ref. 27 p. 1205 for example. *’ R. A. W. Johnstone and F. A. Mellon J.C.S. Furaday fI 1972,68 1209. ” D. W. Turner C. Baker A. D. Baker and C. R. Brundle ‘Molecular Photoelectron Spectroscopy’ Wiley-Interscience London and New York 1970 p. 284. Physical Methods-Part (i) Organic Mass Spectrometry 11 Because the difference between photoelectron adiabatic and vertical ionization potentials can vary from zero to about 0.5 eV and different methods of measuring electron-impact ionization potentials can easily have this range also it is not surprising there is confusion in the literature.At present it seems the only values to be reasonably certain of are the photo-ionization and photoelectron adiabatic ionization potentials and the electron-impact potentials obtained with sensitive apparatus and mono-energetic electrons24 or proper mathematical deconvolution of ionization efficiency data.25 As if this confusion is not sufficient there is an added one associated with the term ‘accurate’. Frequently in the literature values for ionization potentials obtained by electron-impact methods are claimed to be ‘accurate’ to say k0.2eV as the result of several determinations. Whilst it can be appreciated that the values obtained are reproducibleto within those limits i.e.the experimental work has been carried out diligently there is no guarantee that the results are accurate. In other words several determinations of an ionization potential could be repro-ducible within close limits but the absolute value obtained could be inaccurate simply because the methods used to acquire and evaluate data are inadequate. The popular ‘semi-log’ method appears to be particularly suspect in this regard,26 and even within a series of compounds relative ionization potentials are unlikely to be accurate (cf:bandshapes for first ionization potentials of benzene and aniline for example). . The usefulness of ionization and appearance potential measurements has again been re~iewed.~’ The differences in appearance and ionization potentials for primary fragmentation of a series of stereoisomers of 1.3-oxathians were found to correlate well with conformational energy differences indicating release of energy due to non-bonded interactions.28 Almost mono-energetic electron beams have been used to measure the ionization potentials of C3H3 and C3H5 C4H,radicals from various sources and hence the heats of formation of the corresponding cations.24 From these values and measurements of the heats of formation of these ions from organic compounds it was concluded that the C3H ion seemed always to have a cyclopropenyl structure and the C3H,ion to have an ally1 structure.27 On the basis of appearance potentials of benzoyl ions from a variety of pre- cursors it has been claimed that excited states of benzoyl ions could be detected.29 The existence of the excited states was deduced from the high values found for the heats of formation of benzoyl ions from different precursor ions.The high results l4 F. P. Lossing Canad. J. Chem. 1972,50 3973. l5 See for example J. D. Morrison J. Chem. Phys. 1953 21 1767; R.E.Winters J. H. Collins and W. L. Courchene ibid. 1966,45 1931 ;J. Vogt and C. Pascual Internat. J. Mass Spectrometry Ion Phys. 1972 9,441 ; G.D.Flesch and H. J. Svec ibid. p. 106; and reference 35. 26 J. L.Occolowitz and B. J. Cerimele in ‘Abstracts of the 20th Annual Conference on Mass Spectrometry and Allied Topics’ ASTM Committee E-14 Dallas June 1972 p. 95. 2’ J. Jalonen and K.Pihlaja Org. Mass Spectrometry 1973 7 1203. ’* J. Jalonen P. Pasanen and K. Pihlaja Org. Mass Spectrometry 1973 7,949. F. Benoit Org. Mass Spectrometry 1973 7 1407. 12 R. A. W.Johnstone and F. A. Mellon were not attributed to experimental error because only a small spread was found in the heats of formation of C6H cations from various sources. It must be pointed out that the heats of formation for C6H ions obtained in this study although consistent are more than 1 eV greater than the currently accepted best values obtained by photoionization methods3' and are 1 eV greater than values obtained using the IE/EDD method.22 Further in measurements made three years ago,31 the appearance potentials by the IE/EDD method for C,H,CO ions from acetophenone and methyl benzoate were 0.94 and 0.53 eV respectively lower than the values reported in the present worktg and lead to the suspicion that the semi-log method of evaluating the ionization efficiency data has failed spectacularly.Efforts have been made to get some idea of the extent of kinetic shift by com- parison of the appearance potentials of normal ions and the longer-lived meta- stable ions.32 The maximum lifetime of a normal ion is very close to the lifetime of a metastable ion in most mass spectrometers so that a sensitive measurement of the appearance potential of a normal ion should not be expected to differ much from that of a metastable. Recent data33 on appearance potentials of some normal and metastable ions using the sensitive IE/EDD method did not provide evidence for any measurable kinetic shift in fragmentations where they had previously been reported.32 A new method has been described for measuring double and triple ionization potentials from the kinetic-energy loss distributions in reactions of the type m+ + N -+ m2+ + N + e- and gives values which compare favourably with ocher literature values.34 A new method has been proposed for mathematical analysis of ionization efficiency curves obtained with a conventional mass spectrometric ion source. The method takes account of the effective electron energy distribution in the ion source at the time of the experiment and does not need any arbitrary normalization procedures because both co-ordinates have the same units of energy.35 3 Ionization Methods There have been two developments in ionization methods which could become very important in the future.The first of these is an atmospheric-pressure ionization source in which ionization is induced in a stream of gas at atmospheric pressure and the ions produced are allowed into the mass spectrometer through '13 H. M. Rosenstock J. T. Larkins and A. J. Walker Internat. J. Mass Spectrometry Ion Phys. 1973 11 309; Yu. Sergeev M. E. Akopyan and G. I. Vilesov Optika i Spek-' troskopiya 1972 32 230. F. A. Mellon Ph.D. thesis Liverpool University 197 1. 32 J. H. Beynon J. A. Hopkinson and G. R. Lester Internat. J. Mass Spectrometry Ion Phys. 1969 2 291; P. Brown Org. Mass Spectrometry 1970 3 639. 33 T. W. Bentley R. A. W. Johnstone and B.N. McMaster J.C.S. Chem. Comm. 1973 510. 34 R. G. Cooks,T. Ast and J. H. Beynon Internat. J. Mass Spectrometry Ion Phys. 1973 11 490. 35 R. A. W. Johnstone and B. N. McMaster J.C.S. Chem. Comm. 1973 730. Physical Methods-Part (i) Organic Mass Spectrometry 13 a small apert~re.~~.~’ The primary source of electrons is a radioactive 63Ni foil and the positive ions are formed by a complex series of ion-molecule reactions; the source appears to be very similar in its principle of operation to the radioactive sources once very widely used in gas chr~matography.~~ Samples can be intro-duced directly into the gas stream in suitable solvents and therefore the source can be used for biological fluids without prior work-up or the preparation of derivatives.Using a computer to process the data and with multiple single- or many-ion scanning corresponding sensitivities of 5-10 picogram or 25 picogram could be attained. For example nicotine was detected in smokers non-smokers and room air.37 The second ion source uses an electrohydrodynamic ionization technique.39 The interaction of a small conducting liquid meniscus with a strong electrostatic field leads to a high potential gradient and ionization similar to that involved in field desorption. Although the technique has as yet been applied only to liquid metals the authors indicate their intention of using it for labile organic compounds. With the relatively little-used photo-ionization method differences in the fragmentation behaviour of stereoisomers have been readily dete~ted.~’ Field desorption (FD) continues to build up an impressive range of mass spectra of organic compounds which are thermally labile or of low volatility.Recently indirect infrared heating of the field anode has been used to yield more abundant molecular ions relative to fragment ions?’ Pyrolysis of sub-microgram quantities of DNA on the high-temperature activated tungsten emitter of an FD source gave a mass spectrum showing ions for all five bases in the molecule as well as ions corresponding to nucleosides nucleotides and dinu~leotides.~~ High-resolution mass measurements and computer analysis of fragmentation patterns were used to assign ion composition and the possibility of using this technique for sequencing is under investigation.Underivatized nucleosides and nucleotides have also been examined by conventional FD and high-resolution measure- ment~.~~ The need for optimal adjustment of the ion source to obtain good high-resolution data and the importance of a suitable solvent for applying the compound to the emitter were dem~nstrated.~~ Abundances of quasi-molecular and fragment iGns in FD mass spectra of glycosides allowed a differentiation between z-and P-isomer~.~~ The remarkably successful handling by FD of 36 E. C. Horning M. G. Horning D. I. Carroll I. Dzidic and R. N. Stillwell Analyr. Chem. 1973,45 936. 37 E. C. Horning M. G. Horning D. I. Carroll I. Dzidic and R. N. Stillwell Life Sci. 1973 13 1331 and references therein.38 For a description see for example H. P. Burchfield and E. E. Storrs ‘Biochemical Applications of Gas Chromatography’ Academic Press New York and London 1962 p. 58. 39 B. N. Colby and C. A. Evans Anafyt. Chem. 1973,45 1884. 40 Z. M. Akhtar C. E. Brion and L. D. Hall Org. Mass Spectrometry 1973 7 647. 4’ H. U. Winkler and H. D. Beckey Org. Mass Spacrrometry 1973 7 1007. 42 H.-R. Schulten H. D. Beckey A. J. M. Boerboom and H. L. C. Menzelaar Analyr. Chem. 1973,452358. 43 H.-R. Schulten and H. D. Beckey Org. Mass Spectromerry 1973 7 861. 44 W. D. Lehmann H.-R. Schulten and H. D. Beckey Org. Mass Spectrometry 1973,7 1103. 14 R. A. W.Johnstone and F. A. Mellon molecules which prove intractable by conventional ionization methods is illustrated by the success in obtaining mass spectra of and high-resolution data on disodium deoxyfluoro-D-glucose 6-pho~phates.~~ Field ionization (FI) mass spectroscopy together with high-resolution mass measurements and electron-impact mass spectroscopy has been used to analyse a crude mixture of alkaloids from Erythrina seeds;several new alkaloids were tentatively identified.46 The detection of multiply-charged ions in FI mass spectra has been used to derive kinetic data for ionic decomposition^.^^ Like FD and FI methods chemical ionization (CI)continues to find increasingly wide use both conventionally and for ‘difficult’ organic compounds.CI mass spectra of underivatized oligopeptides were obtained by introduction of the sample on the end of a probe directly into the reactant gas plasma.These spectra showed quasi-molecular ion peaks sometimes weak and good ‘sequence’ peaks.48 It was suggested that the 100ng of sample required was a considerable gain in sensitivity over electron-impact methods requiring extensive prior derivatization of the peptide with concomitant loss of sample. For sequencing work the CI method was superior to FD in that it gave more abundant sequence ions. Temperature effects in the CI mass spectra of amino-acids and peptides have been investigated4’ and show the expected increase in fragmentation with increase in ion source temperature. As with photo-ionization FD and FI several stereochemical and conformational effects have been observed in CI spectra. The quasi-molecular ions of steroidal 1,2- and 1,3-amino-alcohols eliminated H20 but only when the N-0 distance was too great for hydrogen-bonding.” Considerable differences were observed between the CI mass spectra of epimeric pairs.Other investigations have shown that CI fragmentations can occur at sites remote from that of protonation.” CI mass spectra have been obtained for biogenic amines which do not give good results under electron-impact i~nization.~~ Finally mention should be made of the increased interest in negative ion mass spectroscopy. Following an earlier passing comment of ours2 regarding the then current small interest in and general usefulness of negative ion mass spectroscopy there has been a vigorous reaction from one group of workerss3 to the extent of formulating axioms to describe negative ion fragmentation.similar to those used by proponents of charge-localization (see earlier comments on p. 8). As the 45 H.-R. Schulten H. D. Beckey E. M. Bessell A. B. Foster M. Jarman and J. H. Westoal J.C.S. Chem. Comm. 1973,416. “ D. E. Games A. H. Jackson and D. S. Millington Tetrahedron Letters 1973 3063. ‘’ H. D. Beckey M. D. Mighaed and F. W. Rollgen Internat. J. Mass Spectrometry Ion Phys. 1973 10 471. 48 M. A. Baldwin and F. W. McLafferty Org. Mass Spectrometry 1973 7 1353. 49 M. Meot-Ner and F. H. Field J. Amer. Chem. SOC.,1973 95 7207. ’’ P. Longevialle G. W. A. Milne and H. M. Fales J. Amer. Chem. SOC.,1973 95 6666. ’’ R.L. Foltz A. F. Fentiman C. A. Mitscher and H. D. Showalter J.C.S. Chem.Comm. 1973,872. 52 G. W. A. Milne H. M. Fales and R. W. Colburn Analyt. Chem. 1973 45 1952. 53 R. G. Alexander D. B. Bigley and J. F. J. Todd Org. Mass Spectrometry 1973 7 643. Physical Methods-Part (i) Organic Mass Spectrometry 15 authors themselves suggest there is not yet a wide range of results with which to test the axioms adequately and it is therefore unfitting to comment on their value. If the axioms do yield an empirical rationalization for negative ion mass spectra they will be useful. 4 Chromatographic-Mass SpectrometricMethods With the increasing use of high-speed liquid chromatography have come attempts to couple this apparatus to a mass spectrometer and possible interfacing systems are being reported. In one procedure a stopped-flow technique enabled the solvent from small liquid samples from the liquid chromatographic apparatus to be evaporated at the tip of a probe which was then driven by a motor into the ion source and heated to vaporize the organic residue.54 Satisfactory mass spectra were obtained in this way but a cycle time is required of 3-5 minutes during which the liquid chromatograph is stopped.As a preliminary investigation into the direct coupling of a liquid chromatograph with a mass spectrometer 10 microlitres of solutions were introduced into a CI source through a 2mm capillary The ideal solvent requires a combination of volatility good reagent gas activity in the CI source and few interfering peaks in the mass spectrum (presumably it should also be a good eluent for the liquid chromato- graph !)and various solvents were tried.CI spectra of high specificity were obtained but increased source pumping speeds will be necessary before direct coupling can be made. A good deal of interesting work has been carried out with coupled gas chromatographic-mass spectrometric (g.c.-ms.) systems which are now in widespread routine use following their ready commercial accessibility. Much of this work is covered in Sections 5 and 6 and only instrumental topics are covered here. Of general interest is the demonstration that higher resolution and sensitivity can be achieved using open tubular capillary columns rather than packed columns.56 This point is also mentioned in the course of an excellent review of integrated g.c.-m.s.technol~gy.~~ Readers might like to be reminded of the usefulness of packed capillary columns which have better resolutions than ordinary packed columns and allow a higher sample loading than open capillary columns.58 A system of flow-programming has been described which by use of an auxiliary helium supply and vent achieves maximum separator yield for differing flow rates from the gas chromatographic apparatus.5g 54 R. E. Lovins S. R. Ellis G. D. Tolbert and C. R. McKinney Analyt. Chem. 1973,45 1553. ” M. A. Baldwin and F. W. McLafferty Org. Mass Spectrometry 1973 7 11 1 I. ” B. F. Maume and J. A. Luyten J. Chromatog. Sci. 1973 11 607. ’’ R. Ryhage Quart. Rev. Biophys. 1973,6 3 11. For leading references see C. A. Cramers J.Rijks and P. Bocek J. Chromatog. 1972 65 29. ’’ M. A. Grayson R. L. Levy and C. J. Wolf Analyr. Chem. 1973,45,806. 16 R. A. W.Johnstoneand F. A. Mellon Defects in resolution on coupling a gas chromatographic apparatus to a mass spectrometer were shown to arise in the connecting lines and not in the separator itself.60 The defects were not simply due to inadequate heating of the lines. 5 Computers Computers are increasingly used in mass spectrometry to acquire and process data to control mass spectrometers to compare incoming data with established 'libraries' and to make structural assignments. That the computer has not yet shrugged off the mass spectroscopist is evident from the highly successful inter- active systems in which an operator 'converses with' or alters the operating mode of a computer according to changing requirements.Data acquisition by computer has been extended to the growing field of multiple ion detection (MID). The utility of MID for monitoring g.c. effluents is well-established. In MID only selected ions from the known mass spectrum of a compound are looked for in the ionized effluent from the g.c. column. This technique of detecting selected ions is more rapid than whole scans of all the ions in a spectrum and has an inherently greater sensitivity since only the more abundant ions need be considered. For example if the mass spectrum of a compound has four abundant characteristic ions the mass Spectrometer can be set to cycle round and continuously monitor the abundances of just those ions ; when the compound emerges from the g.c.column and is ionized those four characteristic ions tire sensed even at very low levels. Because quantitative information is required only for known compounds it is not important that most of the information contained in the total spectrum is not used. In this way accurate identification and estimation can be made of small quantities of for example pesticides in soil or drugs in metabolic fluids. In conventional MID methods the selected peaks are examined by altering the ion-accelerating voltage in controlled jumps and recording the ion abundance data on a pen or galvano- meter recorder. It is this process which has been simplified through the use of computers. The computer is used to control the changes in ion-accelerating voltage and to monitor the resulting MID The information can be displayed instantly on a cathode ray tube6' and exact masses can be 'typed' into the computer which then sets the correct accelerating voltages in the ion source.It proved possible to monitor up to eight selected fragment ions in one application and assays of picogram quantities of myoinositol deuteriated alanine and glucose were made.62 On a less sophisticated level a computer was used to monitor selected ion abundances without controlling the ion-accelerating voltage.63 Such a use of the computer was used also to identify individual ion current profiles 'O F. Bruner P. Ciccioli and S. Zelli Analyr. Chem. 1973 45 1002. '* K. Elkin L. Pierrou U. G.Ahlbor B. Holmstadt and J.-E. Lindgren J. Chromafog. 1973 81,47. 62 W. F. Holmes W. H. Holland B. L. Shore D. M. Bier and W. R. Sherman Analyr. Chem. 1973,45,2063. 63 P. D. Klein J. R. Haufman and W. J. Eider Analyr. Chem. 1973 45 308. Physical Methods-Part (i) Organic Mass Spectrometry 17 which would have overlapped on a conventional data record of composite profiles.64 This same system was also 'interactive' in that the operator could modify the instructions to the computer according to changing requirements. As an alternative to MID of selected ions in a mass spectrum magnetic scanning of a limited range of the total mass spectrum has been used with computerized data acquisition and processing to assay The relative merits of single-ion monitoring and repetitive magnetic scanning have been assessed using cholestene as the test substance.Which technique is to be used must ultimately depend on the amount of information already obtained or the amount required. Single-ion monitoring which is approximately a thousand times more sensitive than repetitive scanning was recommended for detection of trace quantities and repetitive scanning for the location of drug metabolites.66 In this respect an electronic resetting device operating during a scan has led to considerable signal enhancement compared with a normal scan and suggests that the above figures for relative sensitivities may need re~ision.~' Apart from simple data processing a computer can be used to recognize a compound by comparing its mass spectrum with those held in a library or file and finding the best match(es).There are two main requirements to make this library searching fast and efficient namely that a minimum of information be stored and that the search and computation time should be short. To some extent these criteria are interdependent. By coding only the most intense peak in every section of 14 mass units throughout a mass spectrum and using only a minimum of ion abundance information the 7000 mass spectra in one collection could be searched in only 10seconds ;'* this latter paper concludes with a timely warning on the use of library search methods. Information theory has led to the adoption of a statistical search and match technique which gave a high percentage of correct identifications even when matching was confined to only 8peaksin the spectrum.69 The variability of mass spectra was highlighted by these authors who noted that the differences between mass spectra of the same compound run on different instruments were sometimes greater than the variations in spectra of geometrical isomers run on the same instrument.Further developments of a search and retrieval system using two out of every 14 ion peaks and available by telephone have been This system will accept instructions to search for molecular weights and complete or partial molecular formulae besides the usual library searching to match spectra. Attempts have been made to classify the pharmacological activity of drugs according to their mass spectra and preliminary findings were that some drugs " J.T. Watson D. R. Pelster B. J. Sweetman J. C. Frohlich and J. A. Oates Anulyt. Chem. 1973,45 2071. 65 L. Baczynskyj D. J. Duchamp J. F. Zieserl and U. Axen Analyt. Chem. 1973,45,479. 66 D. M. Desiderio and B. S. Middleditch Analyt. Chem. 1973 45 806. " R. P. Page A. V. Nowak and R. Wertzler Anulyt. Chem. 1973,45 994. 68 S. L. Grotch Analyt. Chem. 1973,45 2. 69 S. Farbman R. I. Reed D. H. Robertson and M. E. F. Silva Internat. J. Muss Spectro-metry Ion Phys. 1973 12 123. 70 S. R. Heller H. M. Fales. and G. W. A. Milne Org. Muss Spectrometry 1973.7 107. 18 R. A. W.Johnstone and F. A. Mellon could be classified as sedatives or tranquilizers when the mass spectra were fed into a suitably ‘trained’ computer pr~gram.~ A mass spectral computer-processing technique known as the self-training interpretive and retrieval system (STIRS) has been developed to use nine data classes for library searching and matching.72 These classes include information on ion series primary and secondary losses of neutral particles characteristic ions and fingerprint ions.The system compared well with other matching techniques and appeared to be surprisingly independent of errors in the spectro- scopic data in some cases. 6 Chemical Biochemical and Biomedical Uses There has been informal discussion as to the reality of the word ‘biomedical’ since medicine is always associated with biological systems and as to whether it is any different from biomedical. As used here we imply the interaction of medicine and biology in a gross sense i.4.large-scale effects and associated natural or in- duced metabolic functioning as opposed to the molecular (biochemical) investiga- tion of living systems. For example the estimation of ethanol in a toper would be biomedical :the chemistry of the metabolism of the ethanol would be biomedical. Mass spectrometric methods of peptide and protein sequencing are now becoming more routine and especially when the quantity of material available is not a problem these methods can compete favourably with the Edman degra- dation. Each of the methods mass spectrometric and Edman has its advantages and disadvantages which can be classified mainly under the headings of time involved sensitivity cost ‘difficult’ amino-acids and branching or looping of the peptide chain.Most probably the two methods will gradually come to complement each other as demonstrated by the sequence determination on somatostatin a hypothalamic polypeptide which inhibits the secretion of somatotropin.’ Initially Edman degradation of the complete polypeptide became difficult at the sixth cycle and therefore somatostatin was cleaved with trypsin into three peptides two of which were sequenced by the Edman method and the third by mass spectrometry. A discussion of permethylation methods preferred for use for the ‘difficult’ amino-acid residues methionyl histidyl and arginyl contains recommended procedures for dealing with peptides containing them.74 The methods described are now standard and widely used but for the methionine- and histidine-containing peptides it is recommended that very short permethylation times be used rather than exact equivalents of methyl iodide7’ K.-L.H. Ting R. C. T. Lee G. W. A. Milne M. Shapiro and A. M. Guarino Science 1973,180,417. 72 K.-S. Kwok,R. Venkataraghavan and F. W. McLafferty J. Amer. Chem. SOC.,1973 95 4185. 73 N. Ling R. Burgus J. Rivier W. Vale and P. Brazeau Biochem. Biophys. Res. Comm. 1973,50 127; R. Burgus N. Ling M. Butcher and R. Guillemin Proc. Nat. Acad. Sci. U.S.A. 1973,70 684. 74 H. R. Morris R. J. Dickinson and D. H. Williams Biochem. Biophys. Res. Comm. 1973 51 247. 75 M. L. Polan W. J. McMurray S. R. Lipsky and S. Lande Biochem. Biophys. Res. Comm. 1970 38 1 127 J.Amer. Chem. SOC.,1972,94 2847. Physical Methods-Part (i) Organic Mass Spectrometry 19 to prevent sulphonium and ammonium salt formation. In a criticism of the latter method the authors included a further publication7’ which they mis- rep~rted’~.~~ as implying the use of stoicheiometric quantities of methyl iodide. The earlier p~blication~~ in fact recommended short reaction times and not the use of stoicheiometric quantities of methyl iodide. Since two groups of workers have now found that short reaction times for permethylation do not give salt formation with methionine and histidine it can probably be accepted as a routine procedural detail in the presence of these residues. The application of low- and high-resolution mass spectrometry chemical ionization metastable peak measurement and fractional evaporation mass spectral methods to the analysis of mixtures of peptides has been des~ribed.’~ It was concluded that high-resolution mass data were necessary for the unequivocal assignment of some peptide sequences although statements to the contrary have appeared earlier.’’ Dipeptides isolated from urine and identified by g.c.-m.s. have indicated a break-down in collagen in a patient suffering from undiagnosed purpura.80 The C-terminal amino-acids in peptides have been identified by mass spectro- metry after conversion to thiohydantoins and trimethylsilylation.8 The application of mass spectrometry to the analysis of saccharides and associated substances continues to increase.The elucidation of sequences of monosaccharide units in N-arylglycosylamine acetates and in acetates of 2-phenyI-1,2,3-triazole derivatives has been done in this way.” The analysis of glucuronides by mass spectrometry of the trimethylsilyl derivatives allowed identification of the sugar and agly~one.’~ Aldosylaldonates have been per- methylated and analysed by g.c.-ms. whereby 1 --+ 3 1 -+4 and 1 -* 6 linked residues could be distinguished even though there was no molecular ion (electron- impact i~nization).’~ After acetylation and methylation amino-sugar-containing glycosphingo-lipids were analysed. 85 In a new variation of g.c.-m.s. methods for distinguishing the stereoisomers of chiral alcohols and amines drimanoyl and chrysanthemoyl derivatives were used to give readily separable diastereoisomers.86 H.R. Morris in ‘Mass Spectrometry’ ed. D. H. Williams (Specialist Periodical Reports) The Chemical Society London 1973 Vol. 2 p. 162. 77 G. Marino L. Valente R. A. W. Johnstone F. M. Tabrizi and G. C. Sodini J.C.S. Chem. Comm. 1972 357. H.-K. Wipf P. Irving M. McCannish R. Venkataragharen and F. W. McLafferty J. Amer. Chem. Soc. 1973,% 3369. 79 H. R. Morris D. H. Williams and R. P. Ambler Biochem. J. 1971 125 189. R. A. W. Johnstone T. J. Povall J. D. Baty J.-L. Pousset C. Charpentier and A. Lemonnier Clin. Chim. Acta 1974 52 137. M. Rangarajan R. E. Ardrey and A. Darbre J. Chromafog.,1973,87 499. 82 0.S. Chizhov N. N. Malysheva and N. K. Kochetkov Izvest. Akad. Nauk S.S.S.R. Ser. khim.1973 1030; Carbohydrnre Res. 1973,28 21. 83 S. Billets P. S. Lietman and C. Fenselau J. Medicin. Chem. 1973 16 30. J. N. C. Whyte. Canad. J. Chem. 1973,51 3197. W. Stoffel and P. Hanfland Z. physiol. Chem. 1973 354,21. 86 C. J. W. Brooks M. T. Gilbert and J. D. Gilbert Anafyf. Chem. 1973,45 896. 20 R. A. W.Johnstone and F. A. Mellon Oestrogens have been analysed in normal human placental tissue at term and in 1-50 microlitres of urine during pregnan~y.~’ Similarly urinary acid meta- bolites have been identified and estimated by g.c.-m.s.88 Triglyceride mixtures from a variety of sources from cow’s milk to coconut oil were successfully ~eparated.~’ 7 Metastables Research on metastable ions continues apace. If all ions which arise through fragmentation in regions of the mass spectrometer other than the ion source are designated as metastables then this section should include collisional activation” and -E mass spectra.” Collisional activation spectra are formed by increasing the pressure in a field- free drift region of the mass spectrometer until ion-neutral collisions result with addition of internal energy to the ions.The extra energy may be sufficient to cause fragmentation and this can be studied by conventional methods of meta- stable peak measurement. Such collision-induced metastables9’ have been studied in an instrument having a reversed Nier-Johnson geometry by direct analysis of daughter ions (DADI). The ion decomposition pathways resulting from collisional activation resembled conventional mass spectra and provided extra information on ions in normal spectra.The relative abundances of most product ions in collisionally activated spectra appear to be dependent on structure rather than internal energy;90 the structures of ions from different precursors were compared. The so-called -E mass spectra are formed by generating negative ions from positive ions in the electrostatic analyser of a double-focusing instrument. If the polarities of the analyser are reversed negative ions thought to arise from the process m+ + N -+m- + N2+are transmitted. Two variations on methods of investigating metastable ions have been proposed. One technique uses the electrostatic analyser region of a double-focusing mass spectrometer to obtain DADI spectra,93 and the other uses a similarly modified instrument to present consecutive metastable transitions on a single chart record.94 8 Conclusion The increasing use of mass spectrometry in biochemical and biomedical fields has given added impetus to the development of computer processing and of new ” J.Jakowski H.-S. Ervast and H. Adlercreutz J. Steroid Biochem. 1973 4 181; H. Adlercreuz and D. H. Hunneman ibid. p. 233. T. A. Witten S. P. Levine M. T. Killian P. J. R. Boyle and S. P. Markey Clinical Chem. 1973 19 963. ’’ T. Murata and S. Takahashi Analyt. Chem. 1973,45 1816. 90 F. W. McLafferty P. F. Bente R. Kornfeld S.-C. Tsai and I. Howe J. Amer. Chem. SOC.,1973 95 2120; F. W. McLafferty R. Kornfeld W. F. Haddon K. Levsen I.Sakai P. F. Bente S.-C. Tsai and H. D. R. Schuddemage ibid. p. 3886. 91 T. Keough J. H. Beynon and R.G. Cooks J. Amer. Chem. SOC. 1973 95 1695. 92 K. R. Jennings Internat. J. Mass Spectrometry Ion Phys. 1968 1 227. 93 J. H. Miller J. Ross,J. Rustenburg and G. L. Wilson Analyt. Chem. 1973 45 627. 94 J. H. Miller and G. L. Wilson Internat. J. Mass Spectrometry Ion Phys. 1973 12 225. Physical Methods-Part (i ) Organic Mass Spectrometry ionization techniques. With advances in these fields the need for prior preparation or derivatization of biological samples can be greatly reduced. The establishment of a firm base for the understanding of mass spectrometric fragmentation mechanisms is a slower process but is being tackled through a variety of sophisticated theoretical and instrumental methods.Finally in a light-hearted vein we have been intrigued by the steady prolifera- tion of jargon in the form of abbreviations applied to or resulting from mass spectrometry. Without thinking too hard we append a list of some of the abbreviations we have come across this year. The prize for the first correct solution is an Honorary Fellowship in the Faculty of Sciences of Llareggub University.” PI PE EI CI FD FI GC MS IE EDD IP AP STIRS HSLC QUISTOR TLC MID LRP HRP API QET INDO CNDO MO EHT CRT DAD1 -E BNI 9QJ L69 3BX ADIOS ’’ Entries to The Dean 0. Dr. Williams c/o Dr. H. van’t Klooster Dept. of Analytical Chemistry Utrecht University The Netherlands.