Mass Spectrometry Volume 1

A Review of the Literature Published between June 1968 and June 1970

By D. H. Williams

The Royal Society of Chemistry

Copyright © 1971 The Chemical Society
All rights reserved.
ISBN: 978-0-85186-258-3

Contents

Chapter 1 Alternative Methods of Ionisation and Analysis By J. M. Wilson,
Chapter 2 Energetics, Kinetics, and Ion Structures By I. Howe,
Chapter 3 Reactions of Specific Functional Groups By J. H. Bowie,
Chapter 4 Natural Products; including Oligopeptides, Oligonucleotides, and Oligosaccharides By R. G. Cooks and G. S. Johnson,
Chapter 5 Organometallic and Co-ordination Compounds By M. I. Bruce,
Chapter 6 Computerised Data Acquisition and Handling By S. D. Ward,
Chapter 7 Gas Chromatography — Mass Spectrometry By C. J. W. Brooks,
Author Index, 309,

CHAPTER 1

Alternative Methods of Ionisation and Analysis

BY J. M. WILSON

1 General Introduction

This chapter is a discussion of four techniques which have been used in the past four years as an alternative to what I should call the ‘conventional’ mass spectroscopic methods. The first three are alternative methods of ion production and are of interest because they may solve some of the problems which at present conventional electron impact methods cannot solve.

The disadvantages of using an electron bombardment–positive ion source fall into four categories:

(i) A considerable number of compounds do not have stable molecular positive ions.

(ii) In some groups of compounds the tendency to undergo rearrangements often makes it very dangerous to attempt an interpretation in terms of the structure of the unionised molecule.

(iii) Mass spectra are often insensitive to stereochemical differences between molecules.

(iv) The energy distribution in ions produced by electron impact is in most cases unknown.

The first problem is of some importance since one of the first pieces of information which is required in a structure determination is the molecular weight and formula. Field ionisation, chemical ionisation, and negative ion spectra all appear to have advantages for this purpose. Field ionisation would appear to be more generally useful, but there is as yet little work reported on the use of chemical ionisation spectra using reactant gases other than methane. The attractive feature of chemical ionisation is that there is a wide range of gases which could possibly be used. Negative ion spectra should also be useful for this purpose, but no-one has yet defined the optimum experimental conditions for the general production of negative ion spectra from organic compounds.

The tendency to rearrangement appears to be considerably reduced both in field ion and negative ion spectra, although the low abundance of fragment ions in the former may be a disadvantage in some cases. The usefulness of chemical ionisation spectra for this purpose has not yet been defined.

The energy difference between a pair of stereoisomers is usually much less than the average energy transferred in electron bombardment. It should be expected that low-energy processes such as field ionisation or chemical ionisation would be more useful for stereochemical studies. The fourth problem, that of the quantitative measurement of energy transferred in the ionisation process, is one which can be solved by chemical ionisation. The thermodynamic quantities involved in ionisation by charge exchange are often well known, and spectra obtained in this manner should be useful in theoretical studies.

The fourth technique discussed, ion cyclotron resonance, is a different method of ion analysis. Its main applications are to the study of ion–molecule reactions, and it has made possible the elucidation of information on reaction pathways which is not generally available from studies using conventional instrumentation. Its relevance to studies of the mass spectra of organic compounds lies in the ability to measure the reactivity of ions. The structure of ions in the mass spectrometer is still an unsolved problem; the conventional mass spectrometer can only measure the mass of ions, and give some information on the thermodynamics of their formation and their unimolecular decomposition. The ability to study their reactivity in bimolecular processes should lead us somewhat closer to an understanding of their structure.

2 Chemical Ionisation

This method involves the production of ions from the sample under inspection by reaction with ions produced in a reactant gas. The usual experimental conditions which are required are a pressure of 1 Torr of reactant gas and 10-4 Torr of sample. At such pressures, it is much more probable that a sample molecule will react with an ion than be ionised by an electron.

Most of the early work was carried out by Field and his collaborators using methane as reactant gas. The principal ionic processes in methane are as follows:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

In the spectrum at 1 Torr, about 90% of total ionisation is accounted for by the ions CH5+ and C2H5+. These do not react further with methane but, in the presence of other molecules introduced as impurities in the methane, further reactions can be observed.

The methane chemical ionisation spectra of saturated hydrocarbons are generally dominated by an (M — 1)+ peak. Although the detailed mechanism of the reactions which produce these ions is not well understood, the simple rationalisation that CH5+ reacts as a Bronsted acid and C2H5+ reacts as a Lewis acid appears to be satisfactory. In the mass spectrum of n-octadecane (Figure 1), both ions are probably precursors of the abundant C18H37+ ion.

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (1)

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (2)

It is possible that reaction (1) is a two-step process, protonation being followed by decomposition of the unstable C18H39+ intermediate. The other ions in the C.I. spectrum of n-octadecane are all CnH2n+1+ ions, and these can be explained by an electrophilic substitution process, e.g.

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

Field has further rationalised his observations of the C.I. mass spectra of hydrocarbons in terms of a random attack of reactant ions followed by specific decomposition at the point of attack.

The spectra of branched alkanes exhibit slightly less abundant (M – 1)+ ions than do those of the straight-chain hydrocarbons. This is in contrast to the E.I. mass spectra, in which there is drastic reduction in the relative abundance of the M+ ion in going from straight-chain to branched hydrocarbons. In the mass spectrum of 7-n-propyltridecane (Figure 2), it can be seen that (in comparison with Figure 1) there is very little difference in the intensity of the (M – 1)+ peaks. It can also be seen that C — C bond fission is enhanced at the branch, and this has been attributed to β-fission. Such a process will be endothermic for a straight-chain secondary alkyl ion, e.g.

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

but exothermic for a branched ion.

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

It is generally found in studies of ion-molecule reactions that only exothermic or thermoneutral processes are observed; the collision frequency at pressures of the order of 1 Torr is too low for frequent thermal activation. The abundance of the (M – 1)+ ion varies with the degree of branching. The relationship

Ii = 0.32(Ni/Nn)

has been found to approximately describe the intensity Ii of the (M — 1)+ peak from an isoalkane, where Ni is the number of available hydrogen atoms in the isoalkane and Nn is the number of available hydrogen atoms in the n-alkane with the same number of carbon atoms. Hydrogens which are ‘not available’ for abstraction are those in methyl groups and β to branches. In some multiply-branched hydrocarbons, e.g. (1) in Scheme 1, an ion which could lead to β-fission could also be stabilised by a 1,2-hydride shift. Such a possibility limits the accuracy of such calculations, but the results will still give some indication of the degree of branching.

Skeletal rearrangements can, of course, be expected in carbonium ion chemistry. The presence of an [M – C3H7]+ ion in the spectrum of (2) can be explained by a series of rearrangements, as in Scheme 2. Compounds such as (3) which cannot undergo such rearrangements show no [M – C3H7]+ ion.

The cycloalkanes are the only saturated hydrocarbons which exhibit a protonated molecular ion. The (M + 1)+ ions are formed presumably by protonation followed by C — C bond cleavage within the ring. For acyclic saturated hydrocarbons larger than ethane, the protonated molecular ions are very unstable and decompose either by loss of H2 or by C — C bond fission. Generally the intensity of the (M – 1)+ ion is about ten times that of the (M + 1)+ ion. The (M – 1)+ intensity can be calculated by a method similar to that used for branched acyclic hydrocarbons. Although in methane C.I. mass spectra we find mostly even-electron species (the ionising species CH5+ and C2H5+ are both even-electron), there are some useful exceptions. The only even-mass ions in the cycloalkane spectra are formed by fission of ring–substituent bonds, e.g. as in Scheme 3. Examination of cyclo-C6D12 shows that the formation of C6H13+ and C6H11+ from cyclohexane takes place with no exchange of hydrogen atoms between the ion and the molecule other than the transfer of a single particle. This result effectively eliminates the possibility of reaction (4) since the isotopic species C6D12H+ would not be likely to decompose by elimination of HD only. A more probable explanation is that C6H11+ is formed by hydride abstraction only.

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (3)

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (4)

In the C.I. mass spectra of alkylbenzenes there are usually very abundant (M + 1)+ ions. In all cases where there is an intense (M + 1)+ peak, it is accompanied by (M + 29)+ and (M + 41)+ peaks. The suggested mechanism for formation of the former is as follows:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

another molecular collision being necessary for stabilization. The protonated or alkylated molecular ions can decompose in two ways, as shown in Scheme 4.

The electron impact mass spectrum of toluene has provided some interesting examples of rearrangement processes. One of these is the essentially random loss of H· or D· from the molecular ion of any of the deuteriated toluenes. Field has found that the principal processes in the methane C.I. spectrum of toluene are formation of (M + 1)+ and (M – 1)+ ions. In the hydride abstraction process of C6H5CD3, the ratio of H loss to D loss is 0.14. This would suggest that there is almost no rearrangement involved, only mainly a simple and specific abstraction process. Similarly, [7-2H]cycloheptatriene undergoes hydride abstraction fairly specifically, although formation of C7H7+ after electron impact is random.

Olefins give series of CnH+2n-1 and CnH+2n+1 ions, the former predominating at carbon numbers up to the molecular weight. There is considerably more fragmentation of the carbon chain than is found with saturated compounds. Protonation with CH5+ is about 57 kcal mol-1 exothermic, so it is possible for the protonated molecule both to rearrange and fragment. It would appear that like E.I. spectra, methane C.I. spectra are not very sensitive to the position of the double bond in acyclic olefins.

The introduction of an oxygen atom into the molecule leads to spectra with very specific cleavages. This is understandable in terms of the current theory of rates of ion–molecule reactions. The probability of reaction is dependent upon an ion–dipole interaction. In a molecule with, a polar group, one would expect the ion to attack specifically at the polar centre. In the methane C.I. spectra of esters, nearly all of the ions can be explained in terms of proton transfer to oxygen or ion attachment to oxygen as the first step. By far the most abundant ion in the spectrum of methyl propionate is (M + 1)+. [M + C2H5]+ and [M + C3H5]+ ions are also observed, but the only abundant fragment ion is C2H5CO+. With larger alkyl groups, the spectra become more complex, e.g. that of n-pentyl propionate (Scheme 5). The abundance of the various ions produced varies considerably with the structure of the alkyl group. Where the alkyl group can form a stable carbonium ion, the principal reaction scheme is as follows:

[FORMULA NOT REPRODUCIBLE IN ASCII]

This process has been studied in greater detail using isobutane as reactant gas. At a pressure of 1 Torr the principal ionic reactions in isobutane are hydride abstractions from the neutral molecule, e.g.

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

The resulting spectrum consists almost entirely of C4H9+, the t-butyl cation, which behaves as a very mild Bronsted acid. This can be seen from a comparison of the proton affinities of CH4 (118 kcal mol-1) and i-C4H84 (187 kcal mol-1). The chemical ionisation mass spectrum of benzyl acetate using methane as a reactant gas does not show much change with temperature, as can be seen in Table 1, but there is a pronounced effect on the isobutane C.I. mass spectrum.

The reaction

[FORMULA NOT REPRODUCIBLE IN ASCII]

is sufficiently exothermic that the change in thermal energy between 91 °C and 193 °C has little effect on the further decomposition of the protonated molecular ion. At low temperatures there are also observed association reactions of the type:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

It was therefore possible to measure the rate constants for thermal decompositions and equilibrium constants for the association reactions. Field found that the rates of decomposition of benzyl acetate were about one tenth those of t-amyl acetate. From variable-temperature studies, he was able to show that although the activation energies were the same, the pre-exponential factors in the Arrhenius equation differed by a factor of 10. This can be explained if one assumes that there is free rotation about the C — phenyl bond in protonated benzyl acetate, but in the activated complex for dissociation this rotation must be stopped to allow the CH2 group to become coplanar with the benzene ring.

In a further study of substituted benzyl acetates, Field has shown that substantially the same processes occur, except that there is no evidence for the formation of the p-nitrobenzyl ion from p-nitrobenzyl acetate. The other compounds gave values for rate constants and activation energies which gave reasonable fits on Hammett plots. One explanation for the exceptional behaviour of nitro-compounds is that the attack of C4H9+ may be specifically at the nitro-group.

Weeks and Field have found that the corresponding process in methoxymethyl acetate will not take place with isobutane, but can be observed with methane.

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

The change in rate of production of CH3O[??]H2 fits an Arrhenius plot and has a low pre-exponential factor which may be due to the necessity for a cyclic transition state involving the ether protonated species (4).

(Continues…)Excerpted from Mass Spectrometry Volume 1 by D. H. Williams. Copyright © 1971 The Chemical Society. Excerpted by permission of The Royal Society of Chemistry.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.