вход по аккаунту


Relationships between Structures and Mass Spectra of Organic Compounds.

код для вставкиСкачать
Relationships between Structures and Mass Spectra of Organic Compounds [*J
Our knowledge conceri?ing the rcJlationsliips between the structure and mass Jpectra of
organic compounds has widened iirimenscly during the past few years. It ha5 become pmsJible
to alter the course of fragmentation in the mars spectrometer by changing certain substituents using simple chemical reaction i. The fragmentation products formed are then
entirely diferent from those of the starting material. A combination of controlled massspectrometric decomposition with chemical microreactions often permits the solution of
complicated structural problems with a minimum expenditure of time and material. One
example for the alteration of the fragmentation pattern by a simple chemical reaction is
discussed in connection with rhe elucidation qf’ the structure of pleiocarpamine.
I. Introduction
The qualitative analysis of organic compounds by means
of mass spectrometry is a method of very recent origin.
During the initial period between 1950 and 1959, Beynon [I], Stenhagen and Ryhage [ 2 , 3 ] , Meyerson [4] and
particularly McLaferty [ 5 ] investigated the characteristic fragmentation reactions of simple classes of compounds. The general laws of fragmentation determined
in this way then allowed the elucidation of the structures
of unknown compounds, mainly natural products. This
second stage of development was initiated by the
elucidation of the structure of phthiocerol, a wax-like
constituent of the capsule of the tubercle bacillus [6].
The tremendous possibilities opened up by the new
method were demonstrated by Biernann [7] and also by
Djerassi and Rudzikiewicz [8] by solving complicated
structural problems in the fields of amino acids and
11. Mass-Spectrometric Investigations of
Organic Compounds
In the ionization chamber of the mass saectrometer, the
and Prague; cf. also G. Spiteller, Chem. Weekblad 59, 205 (1963);
2. analyt, Chemie 197, 1 (1963); G, Spite//er and M , Spite//er.
Friedmann, Ind. chim. belge 29, 357 (1964).
[**I Present address: Institut fur Organische Chemie der Universitat Gottingen (Germany).
[I] J . H. Bejwov, Nature (London) 174, 735 (1954); Mikrochim.
Acta 1956, 437; Mass Spectrometry and Its Application to
Organic Chemistry. Elsevier, Amsterdam 1960.
[2] Review: R. Ryhage and€. Stenhogen, J . Lipid Res. I , 361 (1960).
[3; Review: R. Rvhagr and E. Stenhagen in F. W . M c L a f e r t y :
Mass Spectrometry o f Organic Ions. Academic Press, New York
1963, pp. 399-452.
[4] Review: H . M . Grubb and S. Meyerson in [3], pp. 453-528.
[ 5 ] F. W. McLafferry, Analytic. Chem. 28, 306 (1956); Appl.
Spectroscopy / I , 148 (1957); in [3], pp. 309-342.
[6] L. Ahlquist, R. Ryhage, E. Stenhagen, and E. v. Sydow, Arkiv
Kemi I4, 21 I (1959); R. Ryhage, S. Stallberg-Stenhagen, a n d E .
Stenhagen, ibid. 14, 247, 259 (1959).
Angew. Chem. intrrnat. Edit. 1 Vol. 4 (1965) / No. 5
quantity Of energy transferred is not
known and
probably depends on whether the impact is central or
glancing. The energy is generally sufficient to eject an
electron from the molecule, and is also sufficient to
split the molecular ion formed into fragments. In this
way an organic compound is cleaved by the reagent
171 a) K . Biemann, Angew. Chem. 74, 102 (1962); Angew. Chem.
internat. Edit. 1, 98 (1962); b) Mass Spectrometry. Organic
Chemical Applications. McGraw Hill, New York 1962; c) in
[3], pp. 529-556.
[S] a) C. Djerassi, Pure appl. Chem. 6, 575 (1963); b) H . Budzikiewicz, C. Djerassi, and D. H . Williams: Interpretation of Mass
Spectra of Organic Compounds. Holden-Day, San Francisco
1964; c) Structure Elucidation of Natural Products by Mass
Spectrometry. Vol. 1, Alkaloids. Holden-Day, San Francisco
“electron” inside the ionization chamber, and is irrecoverably destroyed.
In the course of degradation, positively and negatively
charged ions, neutral molecules, and radicals are
formed. The charged particles can be separated by
physical means, i.e. by electrical and magnetic fields
(according to their mass and charge), and their mass
spectrum can be determined photographically or with
the aid of a recording device (Fig. 1). In general, only the
positively charged ions, whose formation is more probable than that of the negative ions by a factor of 104,
are measured. The spectra of negative ions, which are
more difficult to record and which have mainly been
studied by von Ardenne [9], can supplement the spectra
of positive ions, especially in the case of compounds of
higher molecular weight containing hydroxy groups [ 101.
Fig, I . Section from the mass spectrum of androstane, C19H32, reduced
to half-size, taken with an Atlas Model CH-4 Mass Spectrometer;
temperature of the TO-4 ion source approhimately 60 “C. The mass/
charge ratios m/e were inserted to facilitate interpretation. The molecular-ion peak is at 260. The entire spectrum from mle = 29 to m/e =
262 is 79.5 cm long.
Except for compounds which contain numerous xelectrons, electrons of the energy normally used (70 eV)
form singly charged particles almost exclusively, so
that the peaks of a mass spectrum indicate the mass
numbers (molecular weights) of the decomposition
products. The quantity of ions having the same mass
number is proportional to the height of the peak.
In this sense, mass spectrometry differs fundamentally from
other techniques such as infrared or ultraviolet spectroscopy;
in the latter two methods a molecule is excited by electromagnetic waves but is not altered chemically, so that the sample
used can be recovered unchanged. In contrast, a molecule hit
by a n electron in the ionization chamber of the mass spectrometer generally undergoes cleavage. In infrared and ultraviolet spectroscopy, waveIengths are measured, i. e. the quantities of energy absorbed which are characteristic for certain
structural elements. In the mass spectrometer, the quantity of
energy transferred on collision with the electron varies (when
electrons are used whose energy lies above the ionization potential of the compound investigated) and can therefore be
neither measured nor related t o the structure of the investiga[9] M. v. Ardenne, K . Steinfelder, and R . Tiimmler, Angew. Chem.
73, 136 (1961).
[lo] M . v. Ardenne, K . Steinfelder, and R . Tummler, 2. physik.
Chem. 220, 105 (1962); M. Y . Ardenne, R . Tummler, E. K. Weiss,
and T . Reichstein, Helv. chim. Acta 47, 1032 (1964).
ted compound. On the other hand, the quantity and molecular weight of the fragmentation products can be derived from
themassspectrum. Chemicalandmass-spectrornetricdecornposition processes exhibit many parallel characteristics.
A. Prerequisites for the Investigation of Organic
Compounds by Mass Spectrometry
I . Volatiltiy
In contrast to ultraviolet and infrared spectroscopy,
where a substance is generally investigated in solution
or as a solid, mass spectra can be obtained from substances only if these are present as a gas. This does not
mean that gases or easily volatile liquids and solids
alone are appropriate. Rather, the minimum vapor
pressure required (about 10-6 mm Hg) is so small that
even compounds such as amino acids can be investigated
[l 11. Difficultly volatile substances cannot, however, be
evaporated into the gas reservoir of the mass spectrometer in the normal manner, and introduced through a
nozzle into the ion source [7a], but must be carefully
volatilized in the ion source itself [12-141. With the
exception of a few classes of compounds, nearly all
organic substances can be studied by mass spectrometry in this way. Admittedly, strongly polar compounds which contain numerous O H or NH groups
must be converted into more volatile derivatives, e.g. by
acetylation, in order to make them suitable for massspectrometric investigation [7b].
2. Resolving Power of the Instruments
The applicability of mass spectrometry is also limited
by the resolving power of the instrument, i.e. by its
ability to record ions of different masses as two separate
peaks. The maximum resolving power of currently
available single-focusing instruments is between 1000
and 1500. A resolving power of 1000 means that ions
with a mass of 999 and 1000 can just be separated.
Double-focusing instruments achieve a resolving power
which is approximately twentyfold. However, this increase in resolving power to 20000 does not necessarily
mean that it is possible to obtain spectra of compounds
with molecular weights up to 20000, since most substances with molecular weights above 1000 d o not
possess the required vapor pressure.
The advantage of using double-focusing mass spectrometers is mainly that the empirical formula of an ion
can be determined by an accurate measurement of its
[I 11 K . He.yns and H . F. Grllzninclier, Liebigs Ann. Chem. 667,
194 (1963).
[I21 P.A.Finan and R.I.Reed, Nature (London) 184, 1866 (1959).
R. I. Reed, W . K . Reid, and J . M . Wilson in R. M . Elliot: Advances
in Mass Spectrometry. Pergamon Press, London 1963, pp.
41 6-417.
1131 G. Spireller, C. BrunnPe, K. Heyns, and H . F. Grutzmacher,
2. Naturforsch. 17b, 856 (1962); G. Spiteller and M . SpitellerFriedmann, Mh. Chem. 94, 142 (1963).
[I41 J . F. Lynch, J . M . Wilson, H . Budzikiewicz, and C. DjerasJi,
Experientia 19, 21 1 (1963).
Angew. Chem. internat.
Edit./ Vol. 4 (1965) / No. 5
B. Reproducibility of the Mass Spectra
mass [IS]. For example, it is possible to decide whether
an ion with a mass of 44 has the empirical formula
CzH40 (exact mokcular weight 44.0403) or C3H8 (exact
molecular weight 44.0767) simply by comparison with
a standard sample (peak-matching). However, this
method is still very expensive, so that single-focusing
instruments are preferable for routine measurements.
The ratio of the intensities of the fragment peaks to
the intensity of the molecular-ion peak is strongly
influenced by instrumental factors ; when mass scanning is achieved by altering the electric field, the
focusing quality decreases with increasing molecular
weight of the ions, so that the intensity ratio recorded
in the spectrum is displaced in favor of particles having
a low molecular weight (massdiscrimination effect).
3. Amount of Sample
The quantity of sample required to record a mass
spectrum depends on the method of introduction and
on the type of recording device used. With photographic
recording and direct evaporation of the sample inside
the ion source, approximately 10-3 mg is sufficient,
while 1 mg may be required if the compound is introduced into the ion source via the gas reservoir.
In addition, as in chemical processes, the experimental
conditions are of considerable importance; with increasing electron energy the intensity ratios of fragment
to molecular-ion peaks increases appreciably, but reaches
a limiting value at about 30 eV. Between 50 and 100eV the
spectraaremoreor less independent of theelectronenergy.
For this reason, an energy of 70 eV is generally selected.
A decisive factor affecting the relative abundance of
fragments and the molecular ion produced, is the temperature of the substance undergoing cleavage in the
mass spectrometer [161. Molecules which possess more
kinetic energy require less energy for decomposition, as
they are already thermally excited and more easily
cleaved than “cold” ones [17]. In the spectra of compounds which have been evaporated in cold ion sources
at relatively low temperatures, the peaks of the molecular
ions are more intense than those of the fragments. This
is less so when the same compounds are evaporated in
hot ion sources.
4. Purity
Tn general, the interpretation of a mass spectrum is not
seriously affected by small quantities of impurities.
Since a mass spectrum is composed additively of the
contributions of the individual components, it is often
The upper part of Figure 2 shows the mass spectrum of ntriacontane, taken with a CEC instrument (scanned by variation of the electric field) at a n ion-source temperature of
20 99
-2 10-
20 -
ever, it is very difficult to detect a compound which is
stereoisomeric with the main component, as in general
such isomers afford similar mass spectra.
[I51 Review: R . A . Snunders and A. E. Williams in 131, pp. 343
to 398.
A n g e w . C h e m . internuf. Edit.
Vol. 4 (1965) / No. 5
1161 M . Spiteller-Friedmann, S. Eggers, and C . Spitelier, Mh.
Chem. 95, 1740 (1964).
[I71 D . P. Stevenson, J. chem. Physics 17, 101 (1949).
[ * ] The A.P.I. number refers to the “Catalogue of Mass Spectral
Data”, issued by the American Petroleum Institute (A.P.I.) as
Research Project 44.
sity ratios of the ions can be ascribed principally to the difference in temperature. (These and all following diagrams are
line spectra derived from original spectra.)
The reaction conditions and the type of instrument influence
the intensity ratio of the ions, but have no effect on the main
fragmentation reactions. The same fragments are therefore
formed from a particular substance irrespective of the nature
of the instrument and conditions, and a qualitative analysis
of the spectrum becomes possible. Using identical reaction
conditions and the same instrument, the mass spectra are exactly reproducible and perfectly suitable for quantitative analysis (e.g. hydrocarbon mixtures in the petroleum industry).
C . The Course of Decomposition Reactions
in the Mass Spectrometer
While in a chemical reaction the reagent attacks a
particular group or bond, the electronic attack in the
mass spectrometer is directed towards no particular site
of the molecule. Whereas a chemical reaction owing to
the relative sizes of the reaction partners takes place
only on the surface of the molecules, it must be assumed
that the much smaller electron can penetrate a molecule
at any point, transferring part of its energy to the
bonding and non-bonding electrons of the molecular
orbitals penetrated and leading to the excitation oE
different centers. Within about 1 0 - 1 3 seconds [18], an
electron is ejected from the excited molecule leaving
behind a positively charged molecular ion. According
to the “statistical theory of mass spectra” [IS], the
energy in excess of the ionization potential is assumed
to distribute itself uniformly over all bonds, and fragmentation of the molecular ion then occurs. Numerous
examples have shown that, with a few exceptions, this
theory is at least qualitatively valid; in larger molecules,
however, only partial distribution of the transferred
energy takes place before fragmentation occurs [191.
80 I
Fig 3. Mass spectrum of naphthalene (A.P.I. no. 41).
Ordinate: intensity I [%I.
Abscissa: mle.
In a saturated aliphatic hydrocarbon, electrons can be
removed only from o-bonds. Since these are more stable
than x-bonds, appreciably more energy is necessary to
bring about ionization. Stabilization of the charge in
the molecular ion is not possible. The latter must
contain a single-electron bond, so that the difference
between the energy required to form the ion and that
required to decompose it is relatively small. This means
that the fragment peaks have a very high intensity
compared to the molecular-ion peak (Fig. 2).
It was frequently attempted to solve stereochemical
problems with the aid of mass spectrometry. Normally,
however, stereoisomeric compounds have similar mass
spectra, due to the fact that the main fragmentation
reactions do not involve the center of stereoisomerism.
Thus, stereoisomeric differences lead to only slight
variations in peak intensity, particularly when the center
of stereoisomerism is remote from the bonds which are
preferentially cleaved. It is only when bonds participating in the center of stereoisomerism are broken in the
main fragmentation reactions that the mass spectra
differ sufficiently for stereochemical conclusions to be
1. Single Bond Cieavoge
D. Fragment Formation
Fragments are formed preferentially with a minimum
expenditure of energy. It follows that the intensity of a n
ion peak will be the greater the less energy is required
for ion formation and the more is needed for its further
cleavage. The intensity of an ion peak therefore depends
on both the probability of formation and the probability
of decomposition of the ion.
The ionization of an unsubstituted aromatic hydrocarbon requires a relatively small amount of energy
owing to the abundance of x-electrons, while cleavage
of the resultant molecular ion, in which the positive
charge is stabilized by the residual n-electrons, requires
a considerable amount of additional energy. For this
reason, fragment formation is not very probable and
the molecular-ion peaks of such compounds exhibit a
very high intensity (Fig. 3).
[IS] H . M . Rosensrock and M . Krauss in 131, p. 2.
[I91 V. HanuS and Z . Dolej?eek, Collect. czechoslov. chem.
Commun. 28, 653 (1964).
(19al R . A . Friedel,J. L . Shultz, and A . G. Slinrkey, Analytic.
Chem. 28, 926 (1956).
Since fragmentation occurs preferentially with minimum
energy expenditure, weaker bonds are ruptured rather
than stronger ones, and cleavage reactions in which
stable fragments are formed are preferred, because here
the energy required for rupture is at least partially
balanced by the gain in stabilization energy.
In saturated aliphatic hydrocarbons, for example, the
weaker C-C bonds are therefore split more readily
than the stronger C-H bonds, the main fragments being
ions with the general formula CnH,, + , and masses 15,
29, 43, 57, efc.
By cleavage of a C-C bond in a straight-chain saturated
hydrocarbon, primary carbonium ions of approximately
equal stability are formed (subsequent rearrangement
leading to secondary and tertiary carbonium ions may
then occur). All C-C bonds in such a hydrocarbon
therefore have an equal probability of being ruptured.
The continuous decrease in peak intensity with increasing
molecular weight of the fragments (Fig. 2) is due to the
fact, that ions of higher molecular weight undergo
secondary cleavage reactions more readily (see Section
IL.D.2a), which lead to the accumulation of low molecular-weight ions as final degradation products.
Anqew. Chem. intertiat. Edit.
1 Vol. 4 (1965) 1 No. 5
bility at one of the C-C bonds adjacent to the heteroatom, since in this way stabilized ions can be formed.
a) I n f l u e n c e of a n Alkyl S u b s t i t u e n t o n t h e
C o u r s e of F r a g m e n t a t i o n
The C-C bonds in a branched saturated hydrocarbon
are likewise approximately equal in strength. By fission
of bonds at branching points, however, secondary and
tertiary carbonium ions are formed. Because of the
inductive electron donor effect of alkyl groups, a
positive charge is better stabilized in such intermediates
than in primary carbonium ions. For this reason, less
energy is required to split branched saturated hydrocarbons than their straight-chain isomers. The molecular-ion peaks of branched hydrocarbons therefore
exhibit a lower intensity than those of their straightchain isomers (Fig. 4). The preferential splitting of
bonds at branching points is demonstrated by the
relatively high intensities ot the peaks of ions formed by
such reactions. However, it must be remembered tnat
secondary decomposition reactions are less favored because of the greater stability of secondary and tertiary
2 -c. + oc-x
-cI c-x
N H i , N H R , NR2,
OH, O R , SH, S R , Halogen
5! 5)
60 -
83 (7)
20 -
- LO
Fig. 5 . Mass spectrum of 3-octanol (41 [19al.
Ordinate: Intensity I 1x1.
Abscissa: mle.
Reprinted with permission of the copyright owner.
20 -
rn le
Fig. 4. Mass spectrum of 9-octyldocosane ( I ) (A.P.I. no. 866).
Ordinate: Intensity I
Abscissa: m/e.
For example, the mass spectrum of 3-octanol ( 4 ) (Fig. 5 ) contains t w o characteristic fragments with mass numbers 59 (5)
and 101 ( 6 ) , formed by cleavage of a C-C bond on either
side of the functional group.
carbonium ions. It is therefore possible, in the absence
of functional groups, to deduce the position of
branching from the mass spectrum (Fig 4). For example, 9-octyldocosane ( I ) is decomposed into the
carbonium ions (2) and (3).
H3C-(CHz)7-CH-(CHz)1,-CH3 (3)' m a s s number 309
H3C - ( CH2)4- C$
(6), m a s s number 101
+C - CH z - c H~
is). m a s s number 5 9
(7), [ m a s s number 83]@
(Z), m a s s numbc- 239
b) T h e I n f l u e n c e of a S u b s t i t u e n t C o n t a i n i n g
a Singly B o n d e d H e t e r o - A t o m with a L o n e
Electron Pair
Functional groups containing a singly bonded heteroatomwith a lone pairof electrons (first-ordersubstituents)
have a much stronger effect on fragmentation than alkyl
groups. Such compounds are cleaved with high probaAngew. Chem. intrrimt. Edit. 1 Vol. 4 (1965) / Nu. 5
The low intensity of the peak of fragment (6) compared to
that of fragment (5) can be attributed at least in part to the
greater tendency of heavier ions to undergo further decomposition reactions. For example, (6) decomposes very readily
to form (7) with the elimination of one molecule of water.
In general, molecular ions of alcohols and amines show
a lower intensity than those of straight-chain saturated
hydrocarbons with the same number ot carbon atoms,
in spite of the fact that ionization occurs more readily
due to the lone pair of electrons at the hetero-atom. The
reason is that the probability of decomposition of the
molecular ions is greater (because of the high stability
of the fragments [ "C-X:]) than their probability of
formation, so that the energy required to form the
fragments is smaller than in the case of hydrocarbons.
An increase in the tendency of the hetero-atom to
supply a negative charge to stabilize a positive charge
on an adjacent carbon atom leads to increased fragmentation probability. The fragments of amines formed
in such reactions therefore produce a much higher
intensity in the mass spectrum than those of alcohols of
comparable constitution. The ability of halogens to
supply a negative charge is so small that such fissions
are hardly ever observed, especially since the lower
bond strength of the carbon-halogen bond (compared
to that of the C-C bond) strongly favors release of the
halogen over other possible decomposition reactions.
The ability to stabilize charges decreases from nitrogen
to the halogens via sulfur and oxygen. It does not run
parallel to the electronegativity [20] (which refers to
the ease of displacement of the electrons of a a-bond)
but is a measure of nucleophilic properties (i.e of the
availability of a lone x- or p-electron pair). Although
bimolecular nucleophilic reactions in solution and monomolecular decomposition reactions in the gas phase
(such as take place in the mass spectrometer) cannot be
directly compared, it is in some cases possible to carry
out comparative basicity measurements by mass spectrometry; for example in methoxyethanol (8) the negative
charge density and hence the basicity of the ether
oxygen is greater than that of the alcoholic hydroxy
group because of the inductive electron-donating effect
of the methyl group. When the compound is decomposed in the mass spectrometer, formation of fragment
(9) is much more likely than that of (10) (Figure 6).
(lo), mass number 31
(9), m a s s number 45
Fig. 6. Mass spectrum of methoxyethanol (8) taken with an Atlas CH-4
instrument; temperature of the TO-4 ion source approximately 50 OC.
Ordinate: Intensity I
Abscissa: mle.
A comparative determination of basicity by mass spectrometry has the advantage over wet chemical processes that interfering effects of the solvent are eliminated. For example,
tertiary amines in aqueous solution exhibit a lower basicity
than secondary ones, although an increase could be expected
from the additional alkyl group and the consequent increase
1201 Cf. [7b], p. 88.
tertiary amines [>C-NRz]
form fragments much more fre0
quently than secondary [>C-NHR]
ones [22].
or primary [>C-NHz]
2. Multiple Bond Cleavage
So far we have considered only the stability of positively
charged fragments. No allowance has been made for
the simultaneously formation of uncharged particles.
Since however the t o t a1 energy balance determines the
course of a decomposition reaction, this consideration
is valid only when fissions are compared in which particles of roughly equal stability are formed, e.g. alkyl
radicals. Very often, neutral molecules may be split off
during decomposition reactions. They are much more
stable than radicals and therefore can no longer be
Uncharged particles, i. e. neutral molecules and radicals,
cannot be observed directly in the mass spectrum, but
their mass can be obtained from the mass difference
between peaks of positively charged ions. For example,
a mass difference of 28 could be due to the loss of a
molecule of carbon monoxide or one of ethylene.
A decision whether a neutral molecule or a radical has been
eliminated can readily be made in many cases; compounds
which consist only of C , H and 0 must have an even-numbered molecular weight. If a radical is eliminated from such a
compound by single bond cleavage, a n ion of odd mass number is formed, and the difference in mass between the molecular ion and the fragment must likewise be an odd number.
In contrast, release of a neutral molecule from such a compound can be recognized by an even difference in mass. The
formation of a neutral molecule from nitrogen-containing
compounds, which d o not necessarily possess an even-numbered molecular weight, cannot be so easily recognized from
their mass spectra, since nitrogen-containing neutral molecules of uneven mass as well as nitrogen-free neutral molecules of even mass or with an even number of nitrogen atoms
can be eliminated.
a) M u 1ti s t a g e D e c o m p o s i t i o n React i o n s
l.5 19)
l Or
of electron density on the nitrogen [21]. The decrease actually
observed is due entirely to the water, since with chloroform as
solvent the basicity rises on passing from secondary t o tertiary amines. Similarly, mass-spectrometric studies show that
Uncharged particles, e.g. H20, CH3, and CO, are
generally eliminated in stages. Normally, almost any
number of stable neutral molecules but only a single
radical can be eliminated. Multiple homolytic bond
cleavage would lead to radical-ions of low stability, so
that such processes do not occur for energetic reasons
if other fissions can take place. Although this rule is not
always valid, its application may frequently help to
exclude erroneous interpretations of fission mechanisms.
For example, a mass difference of 30 might be interpreted as the loss of two methyl groups, but much more
likely represents the release of one molecule of CHzO
or NO.
[21] H . A. Staah: Einfuhrung in die theoretische oiganische
Chemie. 2nd Edit., Verlag Chernie, Weinheim, 1960, p. 631,
where further references are given.
1221 L. DolejS, V. HanuS, V. ('ern?. and F. Sorm, Collect. czechoslov. chem. Commun. 28, 1586 (1963); N . Nertner-Jehle, H .
Nesvadba, and G . SpiteNer, Mh. Chem. 95, 687 (1964);Cf. [8b],
p. 77; [8c], p. 167.
Angew. Cliern. internnt. Edit.
Val. 4 (1965) I N g . 5
An exception t o this rule is, for example, the decomposition
of an ion which is formed by the homolytic fission of a saturated, branched hydrocarbon at the branching point.
- R-e~-C,-H‘
This ion decomposes with the elimination of a hydrogen radical from one of the C H groups adjacent t o the positive charge
center, as the radical cation formed is more stable than the
primary ion (see Figure 4).
b) R e a r r a n g e m e n t R e a c t i o n s
McLuffirty [23] was the first to show that neutral molecules are frequently split off during rearrangement reactions, via a six-centered intermediate. For example,
the elimination of an olefin is characteristic for compounds containing a double bond and a y-CH group.
If X = 0, S, or NR, the free p-electrons may contribute
considerably to the stabilization of the rearranged fragment.
X = CH,, CHR. 0, S , NR
Y = H, R, OR, OH, NH2
The positive charge may reside in the olefin or in the
rearranged fragment, depending on the degree of stabilization. In the mass spectra of aliphatic carboxylic
acids (X = 0,Y = OH), for example n-butyric acid [23a],
the rearranged fragment has a mass number of 60. The
intensity of the peak of the rearranged fragment is particularly marked in the case of low-molecular compounds
for this reason, is designated as a “specific” rearrangement [23]). It is frequently observed that labelled atoms
are more or less uniformly distributed over all the fission
fragments. Such “random rearrangements”, which occur predominantly in hydrocarbons, generally lead to
fragments of low peak intensities and are often only of
secondary importance in the interpretation of a spectrum.
While the origin of the hydrogen atom in the McLafferty
rearrangement has been clarified, there is still no agreement as to whether the hydrogen migrates in the form
of a radical [8b,8c] or as a proton [26]. A proton
migration is supported by the fact that an increase in
electron density at the atom X facilitates the reaction.
The rearrangement probability decreases in the order
X = NR>O>CHz.
The main argument for a radical migration comes from
measurements of the ionization potentials of simple
molecules, which indicate that the positive charge of
the molecule-ion is located at the atom X. This argument, however, does not appear to be conclusive, since
no allowance has been made for the fact that in recording mass spectra of organic compounds the electron
energy is not maintained in the region of the ionizing
potential but at 70 eV, so that initially any electron can
be eliminated from the molecule. In addition, it has
been demonstrated that the course of the rearrangement
is independent of the site of the positive charge [26].
The “McLafferty rearragement”, which in many cases strongly influences the decomposition of compounds with secondorder substituents, finds parallels in certain chemical decomposition reactions, for example in the preparation of olefins
by Tschugnef’s method or in the thermal decomposition of
/ \
The exclusive participation of y-hydrogen atoms in the rearrangement reaction was demonstrated by using a carboxylic
acid in which the CH2 groups had been successively replaced
by CD2 groups. While in the mass spectrum of the a-labelled
compound the rearranged fragment had the mass number
62, a fragment of mass number 60 occurred in the case of the
p-labelled compound, and the spectrum of the y-labelled compound showed the rearrangement peak at mass number 61. It
was concluded that the rearranged fragment contains the x CH2 group and one y-H atom [24], in agreement with
McLqfferty’s postulate [23].
The labelling of carbon atoms in a molecule by hydrogen-deuterium exchange [25] is frequently used in mass
spectrometry to explore decomposition mechanisms.
However, a particular hydrogen atom is not always
shifted to a precisely defined site in the molecule, as in
the above-mentioned McLafferty rearrangement (which,
[23] F. W. McLaflerty, Analytic. Chem. 31, 82 (1959).
[23a] G. P . Happ and D . W. Stewart, J. Amer. chem. SOC.74,4404
( I 952).
[23] N g . D i d - N g u y e n , R . Ryhage, S. Stallber~-Stenhogen,and
E. Stenhagen, Arkiv Kemi 18, 393 (1961).
(251 For a review, see [k],
Chapter 2.
Aiigew. Chcin. internat. Edil. 1 Vol. 4 (1965)
No. 5
Other decomposition processes occurring in the mass
spectrometer cannot be compared with chemical reactions and cannot be interpreted by a simple displacement of pairs of electrons. One example is the
decomposition of o-anisidine ( I I ) . Here the mass
spectrum shows that the molecular ion releases a
methyl group and then decomposes to form a pyridinium ion (12) with the elimination of one molecule of
CO [27]. These fission products have a particularly low
energy content, so that alternative fragmentation routes
[26] G. Spiteller and M . Spiteller-Friedmann. Mh. Chem. 95, 2 5 1
( I 964).
[27] G. Spiteller and M . Sprteller-Friedmonn, M h . Chem. Y3, 1395
are not available in this and similar decomposition
reactions [28] because of the stability of the original
(IS), m a s s n u m b e r 1 0 9
c) D e c o m p o s i t i o n R e a c t i o n s which P r o c e e d
via Radical Ions
The interpretation of the cleavage of alicyclic conipounds is appreciably more difficult. Except when
fragments are formed by hydrogen abstraction, at least
two carbon-carbon bonds must be broken, and naturally more energy is required for this purpose than for
the ruptuie of a single bond. Therefore, the molecularion peaks of unsubstituted alicyclic compounds are
much more intense than those of open-chain saturated
compounds with the same number of carbon atoms
(Fig. 7).
/ J
i 151,
mass number 81
mass number 82
The decomposition of the primary fission product (17) by
elimination of a butyl radical in the course of rearrangement
may explain the formation of (16) in a different manner. Probably, both reactions occur simultaneously. The loss of one
molecule of cyclopropane from (17) leads to an ion (19) from
which, once again, a hydrogen atom can be eliminated. However, fragment (20) can also be formed in a single step from
(1 7) by rearrangement.
(f7), m a s s number 138
Fig. 7. Mass spectrum of cis-decalin ( 1 3 ) (A.P.I. no. 992).
Ordinate: Intensity 1 [%I.
Abscissa: m/e.
As the mass spectrum of cis-decalin (13) shows, alicyclic hydrocarbons decompose to give numerous fragments which
are, however, not particularly characteristic. Their formation
can proceed along various routes, so that relationships between structure and mass spectrum are not readily apparent.
The primary cleavage occurs preferentially at a ring junction;
in this way relatively stable carbonium ions can be formed.
The primary decomposition product is a radical ion (17) in
which thecenter of positive charge is separated from the radical
site by several saturated carbon atoms; stabilization of the decomposition product is therefore insignificant. Formation of
such radical ions makes the decomposition of alicyclic hydrocarbons fundamentally different from that of aliphatic hydrocarbons, which form simple ions.
For the decomposition of such radical ions, the radical character of the particles appears to be of decisive significance. The
elimination of one molecule of ethylene from (17) I291 leads to
the formation of the radical ion (14) with mass 110, which
can decompose further with the elimination of a hydrogen
atom or a molecule of ethylene. The fragment (15) formed by
the twofold elimination of ethylene can be stabilized by the
loss of a hydrogen radical to form an ally1 cation (16) with
mass 8 1 .
1281 J . H . Beynon, G . R . Lester, and A. E. Williums, J . physic.
Chem. 31, 1861 (1959).
[29] Following a suggestion of Budzikiewic-, Djerussi, and Williums [8b], the displacement of a pair of electrons is symbolized
by an arrow and that of a single electron by a small hook.
However, unlike these authors, we are indicating each electron
to avoid errors.
i 141,
m a s s number 110
( 171.
m a s s number 138
1/91, m a s s number 96
(ZO), mass number 95
I n analogy with observations by N . Neuner-Jehle with quinolizidines, it seems that C3H6 can also be eliminated as propylene in the course of a twofold hydrogen shift:
Although the formation of only a small proportion of
the fragments appearing in the spectrum of cis-decalin
has thus been interpreted, and only a few ways of
formation of the individual ions have been discussed,
this example does show the many possible decomposition reactions of cyclic compounds and hence the
difficulties encountered in the interpretation of their
spectra. If functional groups are present, however, they
may exert a specific influence on the mode of fragmentation.
Angew. Chein. interrmt. Edit.
Vol. 4 (1965)
!No. 5
111. General Rules for the Interpretation of
Mass Spectra
A. The Formation of Key Fragments
The rules derived so far for the decomposition of organic
molecules in the mass spectrometer allow us to formulate
some principles governing the interpretation of mass
A relationship between the structure and the mass
spectrum of a compound will be apparent if some decomposition reactions require much less energy than all
others, so that spectra are formed which contain few,
but intense peaks. Such characteristic peaks frequently
represent the key to the elucidation of the structure of
an unknown compound, and may be designated as
peaks of "key fragments".
Key fragments differ from others which occur in nearly
every spectrum by their particularly high intensity and/
or their unusual mass number. For example, an ion of
mass number 41, in general an ally1 cation, cannot be
designated as a key fragment since, owing to its great
stability, it arises from many organic compounds with
different structures. The same is true for other hydrocarbon fragments unless their peaks are distinguished
by abnormally high intensities (see Fig. 4).
C H3
simultaneous presence of peaks at mass numbers 74 and M
minus 31 therefore indicates a methyl ester rather than an amethylcarboxylic acid.
While individual key fragments normally indicate only
the group to which a compound belongs, they may
provide conclusive evidence of the presence of certain
structural elements when taken together. Hence they
can make a substantial contribution to the elucidation
of the structure of an unknown compound.
Key fragments are formed above all from compounds
containing basic nitrogen atoms or aromatic ring systems, since stabilized ions can be derived from these
with particular ease. It is therefore not surprising to
find, for example, that in alkaloid chemistry mass
spectrometry has been widely employed to determine
B. Spectrum-Comparison Technique
The use of mass spectrometry in the elucidation of
alkaloid structures depends to a significant extent on
the ability of many plants to produce several alkaloids
which differ only by the presence or absence of functional groups but contain the same carbon skeleton. If
these functional groups do not substantially influence
the course of fragmentation, then the same main
cleavage reactions ensue on decomposition of all these
mass number 74
m a s s number 7 4
It is self-evident that a single key fragment is not
sufficient to postulate the existence of a certain group
in the compound studied, since ions of the same empirical formula may have different structures. For
example, the spectra of a-methylcarboxylic acids (21)
and of unbranched methyl esters (22) show a key fragment with mass number 74, which in both cases is
formed in the course of a McLafferty rearrangement.
However, i n simple methyl esters (23), release of a n -0CH3
radical is also favored, since a stable ion can be formed in this
manner, too. Consequently, the spectra of simple methyl
esters exhibit a fragment with mass number M minus 31 (M =
molecular ion) corresponding to the loss of a methoxyl radical, in addition to the rearrangement fragment with mass number 74. On the other hand, carboxylic acids cannot yield fragments of 31 mass units ( M . U . ) , provided that substituents
-e; -t)CH,
alkaloids in the mass spectrometer. According to the
masses and positions of the substituents, the positions of
the key-fragment peaks differ merely by a certain
number of mass units, so that the carbon skeleton can
be identified simply by comparison of the spectra
(Figure 8). For example, the positions of the key-fragment
peaks of isoreserpinine (24a) and reserpiline (246) differ
by 30 mass units (1 CH3O minus 1 H).
This spectrum-comparison technique was initially used
by Bieinann [30]. It permits a structure determination
only when spectra of compounds with knownstructure
s ~ i c has CHzOH are absent, since to achieve this both a n OH
group and a CH2 group (in the form of a carbene) must be
eliminated, which is impossible for energetic reasons. The
Angrw. Chenr. iwtrrnat. Edit. 1 Vol. 4 (1965) ,INo. 5
(a): R
H; (b): R = OCH3
39 1
are available for comparison, or when the key fragments
give some indication of the groups present so that
reference compounds can be selected.
l Or
m /e
of an N-CH3 group [32]. Its ultraviolet spectrum indicated
the presence of a methoxyindole chromophore and the infrared spectrum theabsence of N H and O H groups as well as the
presence (possibly) of a five-membered cyclic ketone [32]. The
N M R spectrum showed an ethylidene side-chain, and methylation revealed that the second nitrogen atom also was tertiary
The small amount of substance available made the elucidation of structure by chemical means very difficult, and resort
was made to mass spectrometry.
The mass spectrum of pleiocarpamine (28) (Fig. 9) contained
the molecular-ion peak at mass number 322. This led to a
correction of the empirical formula from C~iH2402Nzto
C20H2202NZ [3 I]. A key fragment at mass number 263 corresponded to the loss of a particle with mass 59, presumably a
methoxycarbonyl group.
m le
Fig. 8. Mass spectra of isoreserpinine ( 2 4 4 (top) and reserpiline 1246)
(bottom), both taken with the Atlas CH-4 mass spectrometer; temperature of the ion source approximately 120°C. Only key fragments
are shown.
Ordinate: lntensity I
Abscissa: mie.
IV. Combination of Chemical and MassSpectrometric Decomposition Reactions
M- 59
< 110-0 ~
JL rl_ 1.
u . & l
Fig. 9. Mass spectrum of pleiocarpamine (28) 1311.
Ordinate: Intensity I
Abscissa: mle.
Even when a structure cannot be derived directly, e.g.
for lack of reference material, the mass spectrum almost
invariably gives some indication of the structural
elements present. These indications can be confirmed or
disproved by chemical degradation reactions. Samples
of 2-5 mg are sufficient when the reaction products are
also studied by mass spectrometry. Their purification is
generally unnecessary, since impurities do not adversely
affect the recognition of the usually intense key-fragment peaks, although they do increase the number of lowintensitypeaks. In addition, volatile constituents originating from impurities can be removed within the ion
source by gradual heating.
Frequently, the course of a reaction can be deduced merely
from the change in molecular weight. Simple chemical reactions frequently alter the basicity of key atoms and thus influence the course of fragment formation in such a way that
the spectra exhibit characteristic cleavage products first for
one and then for the other part of the molecule.
A. Elucidation of the Structure of Pleiocarpamine
The structure of the alkaloid pleiocarpamine (28) was determined by combining small-scale chemical reactions with mass
spectrometry and N M R measurements 13I]. Elementary analysis gave the empirical formula CziH2402N2 and showed the
presence of OCH3 and C-CH3 groups as well as the absence
[30] K . Biemann, Tetrahedron Letters 1960, No. 15, 9.
f311 M. Hesse, W . V .Philipsborn, D. Schumann, C. Spiteller, M .
Spitefler-Friedmafin, W . I . Tayfor, H . Sehmid, and P. Karrer, Helv.
chim. Acta 47, 878 (1964).
This finding was in contradiction to previous results I321 but
could be confirmed by a reinterpretation of the infrared and
ultraviolet spectra. I n order to fully clarify this point, pleiocarpamine was reduced with LiAIHI, giving a product with
a molecular weight of 294 (determined by mass spectrometry).
This corresponds to the loss in mass accompanying the reduction of a COOCH3 group t o a C H 2 0 H group (28 M. U.).
At the same time, the unchanged position of the key fragment
at mass number 263 demonstrated that a CHzOH group had
been eliminated in place of a COOCH, group. In this way the
presence not only of a methoxycarbonyl group in pleiocarpamine was demonstrated, but also that of an unsubstituted
indole chromophore, since the molecule could not contain
more than two oxygen atoms.
The structure of the important key fragment with mass number 180 in the spectrum of pleiocarpamine was elucidated in
the following manner. Since this ion is formed not only from
pleiocarpamine but also from the product of reduction with
LiAIH4, it could not contain the functional group, since i n
this case replacement of COOCH3 by C H 2 0 H would have
caused a displacement of the peak by 28 mass units. Removal
of parts of the stable indole system is impossible for energetic reasons. A fragment which is formed by loss of both the
functional group (59 mass units) and the indole part (1 15 mass
units) from pleiocarpamine (molecular weight 322) will have
a maximum molecular weight of 322 - (59 + 115) = 148.
Since the key fragment had a molecular weight of 180, it must
still contain the indole system, and so the only possible empirical formula for the fragment with molecular weight 180 is
Ci3HloN. If in addition it is considered that the indole nitrogen is substituted, then this ion must be a quinolinium ion
(rearrangements of 3-substituted indoles t o quinolinium ions
are frequently observed in mass-spectrometric decomposition
1321 W. G. K u w p and H.Srhirricl, Helv. chim. Acta44,1503 (1961).
Angew. Cliem. internat. Edit. ] Val. 4 (1965)
No. 5
addition to the presence of a piperidine ring. Besides, the de
composition reaction indicated that the methoxycarbonyl
group must be located in ring E.
(29). m a s s n u m b e r 180
These findings suggested a six-membered ring [structure (30)1
which links the indole nitrogen with the neighboring cdrbonatom.
These results together with evidence obtained from spectra and
chemical decomposition reactions, led Schmid et al. to postulate a provisional structural formula (28~1)for pleiocarpamine [3 I].
This formula has been confirmed by a detailed analysis of the
100 Mc IH-NMR spectrum and by a combined chemical and
mass-spectrometric degradation in which many of the reaction products were not characterized by elementary analysis
but by their mass spectra.
The mass spectrum of pleiocarpamine (Fig. 9) contains no
fragments indicating the presence of a piperidine ring. Evidently such fragments are absent because stabilization of a
positive charge is far better in the indole system than in the
piperidine part of the molecule. Reduction of the indole to
the indoline system greatly reduces stabilization of the positive charge in the aromatic part of the molecule, so that the
mass spectrum of dihydropleiocarparnine ( 3 / ) (Fig. 10) is
entirely different.
Isolation in pure form is not always necessary for the
mass-spectrometric elucidation of the structure of a n
unknown compound. By fractional evaporation of the
sample into the reservoir of the mass spectrometer [7a]
or better by continuous gradual volatilization of the
substance in the ion source, a mixture of substances can
be partly or even completely separated. From the
alterations in the intensities of ion peaks it is then
possible to recognize which fragments are formed from
the components of the mixture.
In mixtures of very similar compounds, it is possible
under certain conditions to derive the structures of the
individual components (without isolation of pure substances) by chemical reactions on a microscale followed
by a mass-spectrometric study of the mixture of reaction
products. One example is the structural elucidation
of eight alkaloids from only 30 mg of material [33].
VI. Summary
V. Elucidation of the Structures of Compounds
in Mixtures without Isolation of the Components
2L 0
Fig. LO. Mass spectrum of 2,7.dihydropleiocarpaniine 131) 131 J
Ordinate: Intensity I
Abscissa: mie.
During decomposition of the dihydro derivative ( 3 / ) , an
ally1 cation (32) seems to be formed initially, which then undergoes rearrangement with displacement of a hydrogen
radical to form the key fragment (33). The elimination of one
molecule of ethylene from (33) finally leads to formation of
the fragment (34).
I n this way, the existence of a tryptamine bridge linking the
piperidine part with the indole ring was made probable in
T, d H
1331, m a s s n u m b e r 1 3 5
Angew. Chrm. iiitenmr. Edit. 1 Vol. 4 (1965) No. 5
The principal rules governing the fragmentation of
organic compounds in the mass spectrometer have been
discussed; whether a mass spectrum can be successfully
interpreted depends to a large extent on the magnitude
of the energy difference between the possible decomposition paths of the compound studied. In general,
spectra characterized by the presence of few key fragments can be expected only from compounds which
decompose preferentially in a few directions. This is the
case when stable fragments can be formed, as in the
decomposition of aromatic or heterocyclic derivatives
or of compounds with substituents releasing negatively
charged portions, since in all these substances a positive
charge may become stabilized.
Simple chemical reactions alter the ability of various
substituents to stabilize charges. Thus the course of
mass-spectrometric degradations can be varied to some
extent, with consequent formation of key fragments
from either one or the other part of a molecule.
A particular advantage is that all reactions can be
carried out on a microscale, since impurities usually
do not interfere with the subsequent mass-spectrometric investigation of the reaction products.
We should like to thank Prof. F. Wessely and our collaborators R. Kaschnitz, N . Neuner-Jehle, and Dr. S . Eggers
for numerous stimulating discussions and the Osterreichischer Forschungsrat for the provision of' a mass
Received: July 30th, 1964
[ A 4191216 IE]
German version: Angew. Cheni. 77, 393 (1965)
Translated by Express Transiation London
[33] G . S p i t e l k and M . Spite((er.-Friedt,ionri, M h . Chem. Y4, 779
(1963); Ind. chim. beige 29, 357 (1964).
Без категории
Размер файла
1 101 Кб
mass, structure, compounds, organiz, relationships, spectral
Пожаловаться на содержимое документа