вход по аккаунту


Electron-Impact Induced Rearrangement Reactions of Organic Molecules.

код для вставкиСкачать
J U N E 1967
PAGES 4 7 7 - 5 8 0
Electron-Impact Induced Rearrangement Reactions of Organic Molecules I * *I
Hydrogen migrations accompany many mass spectrometric fragmentation processes and
a detailed study of these migrations, largely by means of deuterium labeling, has shed
much light on electron-impact induced bond fissions. More recently, it has been found
that groups other than hydrogen, such as alkyl, aryl, hydroxyl, etc., also can migrate
after electron bombardment and the present article is concerned with a discussion of
such molecular rearrangements among organic compounds. These observations are not
only of intrinsic mechanistic interest, but a detailed knowledge of the occurrence of such
migration reactions is crucial to the proper interpretation of mass spectra and to the
avoidance of misinterpretations in “element mapping”.
1. Introduction
One of the most significant developments in organic
chemistry d u r i n g this decade has been the very rapid
acceptance and extremely wide use of mass spectrom e t r y for the solution of structural problems. Sten-
hagen’s 111 prediction in 1961 “Es wird wahrscheinlich
nicht lange dauern, bis der massenspektrometrischen
Analyse bei Strukturbestimmungen in der organischen
Chemie dieselbe Stellung zukomnit wie jetzt z.B. der
Infrarotspektrometrie und der magnetischen Kernresonanz” has become a statement of fact in less than
five years. The pursuit of natural products chemistry
without mass spectrometry has become unthinkable [21.
Whereas the rapid acceptance of UV, IR, and N M R spectroscopy or optical rotatory dispersion by the organic chemist was triggered to a large extent by the commercial availability of suitable instruments, this was not the case with mass
spectrometry. Adequate mass spectrometers were available
at least ten years prior to their extensive utilization for
structural organic investigations and it seems that one of
[ * ] Dr. Peter Brown and Prof. Carl Djerassi
Department of Chemistry, Stanford University
Stanford, Calif. (U.S.A.)
[**I Part of this material was covered in a lecture by C. Djerassi
at the EUCHEM Conference o n Mass Spectrometry, Sarlat
(France) September 1965.
[I] E. Stenhagen, 2. analyt. Chem. 181, 462 (1961).
[2] H . Budzikiewicz, C. Djerassi, and D . H . Williams: Structure
Elucidation o f Natural Products by Mass Spectrometry. HoldenDay, San Francisco, 1964. Vol. I, Alkaloids. Vol. 11, Steroids,
Terpenoids, Sugars and Miscellaneous Classes.
Angew. Chem. internat. Edit.
VoI. 6 (1967) No. 6
the principal motivating factors has been the recent “mechanistic” studies which led to the conclusion [31 that many
fragmentation reactions of organic molecules induced by
electron impact are interpretable in the every-day language
of the organic chemist. To a certain extent, the word “mechanism” is a euphemism since the products of such fragmentations in the mass spectrometer are not isolable, but are only
detected in terms of their mass and one is thus dependent
o n indirect evidence. For this purpose, the most useful technique is isotopic labeling 141, but recognition of metastable
ions, high resolution mass measurements, and appearance
potential determinations also play important roles.
Many of the bond fissions promoted by electron bombardment in the mass spectrometer are accompanied by transfer
of one or more hydrogen atoms. Such hydrogen migrations
shed considerable light o n the probable fragmentation paths
of organic molecules. Much of the work in our laboratory 151
has concentrated o n detailed studies of such hydrogen
migrations triggered by a variety of functional groups. As
a n example of the relevance of such hydrogen migrations to
mass spectrometric fragmentation “mechanisms” one may
cite the “McLafferty rearrangement” [61 of ketones such as
n-dibutylketone ( I ) , which has been shown [7] by deuterium
[3] See for instance H . Budzikiewicz, C. Djerassi, and D . H .
Williams: Mass Spectrometry of Organic Compounds. HoldenDay, San Francisco 1967.
[4] For reviews see [2] ,Vol. I, chapter 2 as well as K . Biemann:
Mass Spectrometry. Organic Chemical Applications. McGrawHill, New York 1962, Chapter 5.
[ 5 ] 129th paper in the Stanford series “Mass Spectrometry in
Structural and Stereochemical Problems”; A . M. Dufjield, W.
Carpenter, and C . Djerassi, Chem. Commun. 1967,109.
[6] F. W . McLafferty,Analytic. Chem. 31, 82 (1959).
[7] For most recent investigation and leading references see H .
Budzikiewicz, C . Fenselau, and C. Djerassi, Tetrahedron 22, 1391
labeling to involve,transfer of the y-hydrogen and fission of
the P-bond with production of the enol ion radical (2) 181 and
a neutral olefin (3) [*I. The “McLafferty rearrangement” is
thus the mass spectrometric equivalent of the well known
“Type 11” Norrish photochemical decomposition of
ketones r91.
The frequent occurrence of hydrogen rearrangements
raises the question whether other moieties are also
prone to migrate after electron impact. Answers to
this question are not only of intrinsic mechanistic
interest, but they also have a crucial bearing on the
scope and limitation of Biemann’s “element mapping”
computer technique [lo], which either requires that
such rearrangements do not proceed to a significant
extent or that their occurrence be predictable and
hence discountable. The electron impact induced
migration of alkyl, aryl, and other functional groups
has attracted considerable attention during the past
two years and apparently occurs with much greater
frequency than has hitherto been realized. It is with
this subject that the present review is concerned and
for our purpose, we are principally concerned with the
following process:
in which the migrating group C is at no time bound
to the rest of the molecule by more than one a-bond.
2. Simple Alkyl and Aryl Rearrangements
and a “McLafferty rearrangement” of one of its
methyl groups should then lead to an ion ( 5 ) of mass
142, which was not observed [111.
Similarly, a variety of benzene derivatives of type (6)
have been shown by isotopic labeling 1131 to fragment
by hydrogen transfer [(7), R = HI, but no analogous
methyl migration [(7), R =: CH31 was encountered[l21
when the appropriate hydrogen atoms were substituted
by methyl groups.
(6) R
= H, CH3;
X = CH2, 0, S, NH;
Y = CH2, 0 ; Z = CH2, 0
(7) R
= H
CH2, 0 , S, NH
These results would seem to offer a poor prognosis
for the prevalence of methyl migrations. On the other
hand, some of the early mass spectral results with
13C-labeled hydrocarbons showed quite definitely
that methyl migrations do occur, presumably through
the intermediacy of cationated cyclopropanes [141.
Thus, the mass spectrum of t-butylbenzene (8) exhibits a peak at m/e = 91 (benzyl or tropylium cation)
which in the 13C-labeled analogue [see asterisked
carbon atom in (S)] is distributed between m/e = 91
and 92 in the indicated proportions, thus supporting
the suggestion that cyclopropane (9) is an intermediate.
A similar conclusion[14J can be reached from the
observation 1151 that the C2 species in the mass spectrum of neopentane (10) labeled with 13C at the
central carbon atom are equally represented by
labeled and unlabeled fragments.
The first question that may be asked is whether alkyl
rearrangements can take place in lieu of isotopically
documented hydrogen migrations in substances where
all implicated hydrogen atoms have been replaced by
alkyl groups. The answer is decisively negative 111 121.
Thus 2,2,8,8-tetramethylnonan-S-one
(4) is the completely methylated analogue of n-dibutylketone (1)
[Sl The relevant ehidence is summarized by S . Meyerson and J.D.
McCoNum in C. N . ReiZly: Advances in Analytical Chemistry
and Instrumentation. Interscience, New York 1963, Vol. 2,
pp. 184-199.
I*] A single “fishhook” (cf. 121, Vol. 11, pp. 1-2) denotes hoPi
molytic fission (-C-C-
+ C*
+ -C-),
while the conventional
arrow refers to heterolysis (-C-C- --f -C3+ -CO). Several of
the fragmentations are indicated as concerted processes, but this
is only done for the sake of convenience since no experimental
method is available at the present time of differentiating between
concerted or stepwise processes.
[9] J. G. Culvert and J . N . Pitts j r . : Photochemistry. Wiley &
Sons, New York 1966. See also P. J . Wagner and G. S. Hammond,
J. Amer. chem. SOC.87, 4009 (1965).
[lo] K . Biemann, Pure appl. Chem. 9, 95 (1964) and later papers.
1111 R . R. Arndt and C . Djerassi, Chem. Commun. 1965, 578.
1121 M . Fischer and C . Djerassi, Chem. Ber. 99, 750 (1966).
Another type of methyl migration, which cannot
really be subjected to the same type of experimental
scrutiny as the above mentioned examples, is one
[13] For leading references see [12].
[14] P. N. Rylander and S. Meyerson, J. Amer. chem. SOC.78,
5799 (1956).
1151 C. P. Johnson and A . Langer, J. physic. Chem. 61, 1010
Angew. Chem. internat. Edit.
Vol. 6 (1967) / No. 6
where only the observed fragmentation mode suggests
its occurrence. For instance, the M - 43 peak (loss
of CH3CO) in 5a-pregnan-20-one (11) is much smaller
than the corresponding one in A16-5a-pregnen-2O-one
(12), which is a surprising result since vinylic fissions
are usually very unfavorable processes in mass spectrometry. Consequently, it has been suggested
that the
expulsion of the acetyl radical in (12) is accompanied
by concerted migration of the angular methyl group
to furnish the highly stabilized carbonium ion (13).
for this reason that migration of the 9-10 bond[*] in
the hypothetical intermediate (17b) was postulated [18J,
the final C6HloNO fragment then being depicted as
(18), which is completely consistent with the deuterium
labeling results.
oH (176)
A similar case is provided by the mass spectrum"71
of camphor (14), whose base peak occurs at nije = 95
( C ~ H1)I (16) owing to the loss of ketene and a methyl
radical. As shown by isotopic labeling, nearly 50 %
rather than the expected 33 % of the methyl loss
originates from C-1. This is best interpreted by
assuming that the fragment ion ( 1 5 ) , resulting from
expulsion of ketene, first suffers methyl migration to
(I5a) (the driving force being the generation of an
"ianized double bond") in which either the labeled or
unlabeled methyl group would be expected to be lost
to an equal extent.
The postulation of such "undetectable" migrations typified
by (IZ), ( I S ) , and (176) is tempting and great caution should
be exercised before resorting to such mechanistic rationalizations. An example of an unnecessary postulation of a methyl
migration is provided by the recently proposed "91 reaction
path (19) + (20) + (21) to account for the appearance of
a n ion of mass 173 in the spectrum of santonin (19).
In fact a more plausible decomposition path can be formulated as (22) + (23), which does not require the intervention of a methyl migration nor does it utilize the unfavorable electron sextet depicted in (20) and (21), simply
by assuming that the fragmentation is initiated by the molecular ion (22).
c H.
Yet another example is provided by the most abundant
fragment ion of mass 112 (C~HIONO)in the mass
spectrum of 5cc-androstan-3-one oxime (17), which
by deuterium labeling (see asterisked positions) was
shown [1*J to arise formally by the bond cleavages
indicated by the wavy line in formula (17). Such a
formulation would require the very unfavorable fission
of two bonds connected to one carbon (C-10) and it is
[I61 L. Tokds and C. Djerassi, Steroids 6, 493 (1965).
1171 D. S. Weinberg and C. Djerassi, J. org. Chemistry 31, 115
(1966); D. R . Dimmel and J . Woiinsky, J. org. Chemistry 32, 410
[IS] D. Goldsmith, D . Becher, S. D . Sample, and C. Djerassi,
Tetrahedron, Suppl. 7 , 145 (1966).
Angew. Chem. internat. Edit. 1 VoI. 6 (1967)
1 No. 6
With the above material as background, it is now
possible to consider recent authenticated cases of alkyl
or aryl migrations. They may be divided into reactions
[*] A very similar rationale has been offered (see H . E. Audier,
S. Bory, G. Defaye, M . Fdtizon, and G . Moreau, Bull. SOC. chim.
France, 1966,3181) for the presence of an intense peak at m/e=117
in the mass spectrum of the diterpenoid hydrocarbon (IBa).
1191 D.G.B. Boocock and E.S. Waight, Chem. Commun. 1966,90.
where the principal driving force appears to be the
generation of favorable carbonium ion or radical intermediates and those where the migration is accompanied by the expulsion of a stable neutral molecule
such as CO or C 0 2 . It is the latter type of rearrangement reaction that is most likely to lead to erroneous
results in “element mapping” [lo], unless its occurrence
is readily recognized.
The first example of an authentic 1,2-alkyl migration
is provided by the mass spectra of trans-A3-lO-methyl2-octalone (24) and related bicyclic ketones, which
exhibit an intense peak at m/e = 69 that was shown
by high resolution mass measurements to correspond
to the composition C 4 H s 0 . Isotopic labeling [201 (see
asterisked positions in (24)) demonstrated that methyl
migration had occurred and that the resulting fragment
is probably best expressed in terms of (25). Similar
conclusions could be reached 1211 from the mass
spectrum of Al-5u-androsten-3-one (26) and its
labeled analogues.
(27) as well as analogous D-homo steroids (ZS), whose
mass spectra displayed the most intense peak at mle =
121 (27a) and which was displaced by the appropriate mass increment i n the various arylidene
derivatives. Again two reaction paths (A and B) can
be entertained [231 which are analogous to those
proposed in the fragmentation of the octalone (24) or
the steroid (26).
\ I
Two plausible pathways to the rearrangement ion (25)
are outlined below, the principal difference being the
nature of the uncharged species which may result from
either a 1,3-(path A) or two consecutive 1,2-(path
B) [221 shifts.
A similar methyl migration process was discovered [231
through appropriate deuterium labeling (see asterisks
in (27)) in 2-arylidene- and 2-furfurylidene-1-decalones
[20] F. Komitsky j ) ., J. E. Gurst, and C. Djerassi, J. Amer. chem.
SOC.87, 1398 (1965).
[21] R. H. Shapiro and C. Djerassi, J. Amer. chem. SOC.86,2825
(1964), as well as unpublished high resolution mass measurements which showed that 70 % of the m/e=69 peak in the mass
spectrum of (26) corresponds to C4H50.
[22] This possibility was first raised by Dr. W. J. Richter at the
EUCHEM Conference on Mass Spectrometry, Sarlat (France)
[23] C . Djerassi, A. M . Duffieid, F. Komitsky jr., and L.Tokes,
J. Amer. chem. SOC. 88, 860 (1966).
A possible distinction between paths A and B can be
made by examining the minimum structural requirements for such methyl migration. The striking observation was made [241 that whereas 4,4-dimethylcyclohex-2-en-1-one (29) and 2-furfurylidene-6,6-dimethylcyclohexanone (30) exhibit practically no
methyl rearrangement peaks in their mass spectra,
introduction of an additional methyl substituent (see
(31) and (32)) at the site of the original ring juncture
restores again the proclivity towards methyl transfer
observed in their bicyclic analogues ((24) and (27)).
Since the presence of this additional methyl function
in (31) and (32) should play no noticeable role in
reaction path A, route B may be considered to be
more plausible.
(29) R = H
(31) R = CH,
(30) R = H
(32) R = CH3
Not only methyl groups, but other alkyl and aryl
groups have also been shown [241 to participate in such
rearrangements. These observations raise the question
about the relative migratory aptitudes of various
substituents in electron impact promoted rearrangements as compared to the migratory aptitudes of the
same substituents in carbonium ion or free radical
[241 R. L. N. Harris, F. Komitsky j r . , and C . Djerassi, J. Amer.
chem. SOC.,in press.
Angew. Chem. internat. Edit. / Vol. 6 (1967)
No. 6
rearrangements 1251. Thus a comparison [241 of the
relative intensities of the mle = 69 and 131 peaks in
the mass spectrum of 4,5-dimethyl-4-phenylcyclohex2-en-1-one (33) indicates a ten-fold preference for
phenyl as compared to methyl migration. Such a
conclusion is only qualitative, since it presupposes an
identical rate (or lack) of further decomposition of the
rearrangement ions of mass 69 and 131.
m / e = 167
(25) m / e = 6 9
(25a) m / e = 1 3 1
m / e = 243
m / e = 57
A similar conclusion about the preference of phenyl
over methyl migration can be reached by comparing
the relative intensity of the hydrocarbon portion
(CsH9) of the m/e = 105 peak in the mass spectrum 1261
of [2-D]-1-phenyl-l,2-epoxypropane(34) with that
of the m/e = 106 peak. The virtue of this comparison
is that these two hydrocarbon species yield effectively
the same ion and any difference in their subsequent
decomposition would be relatively small since it would
be only due to an isotope effect.
m / e = 105
m / e = 43
m / e = 71
2,2-Diphenylchroman (39) exhibits in its mass spectrum 1281
a significant peak at m/e = 195, which by high resolution
mass measurements and isotopic labeling has been shown [291
to involve a phenyl migration resulting in the eventual
expulsion of a benzyl radical, a moiety which is not present
in the original molecule (39). One plausible mechanism,
though by no means the sole oneI291, is depicted below:
m / e = 105
m / e = 106
Epoxides appear t o be a relatively fertile field for
encountering molecular rearrangements. Thus the
presence of m/e = 167 and wile = 243 peaks in the
mass spectra 1261 of stilbene oxide (35) and tetraphenylethylene oxide (36) demonstrates the occurrence of
phenyl migrations, while high resolution and isotope
labeling experiments 1273 in dialkyl epoxides such as
2,3-dimethyl-2,3-epoxybutane(37) and 3,4-epoxyhexane (38) revealed the occurrence of ethyl and
methyl migrations, respectively.
These skeletal rearrangements among epoxides would
complicate o r confuse the interpretation of “element
maps” [lo] were it not for the rather low abundance of
such rearrangement peaks. However, much more
substantial rearrangement peaks, which could affect
conclusions from “element mapping” 1101, have been
encountered in other compounds and three such
examples involving aryl or benzyl migration may be
cited in this connection.
[25] See e.g. E. S . Could: Mechanism and Structure in Organic
Chemistry. Henry Holt and Co., New York 1959, pp. 607and 758.
.Dupin, M . Fitizon, and Y.Hoppilliard, Tet[26] H. Audier, .
rahedron Letters 1966, 2077, and personal communication from
Professor Fitiron.
1271 P. Brown, .
and C. Djerassi, Tetrahedron Suppl.8.
Pt. 1, 241 (1966).
Angew. Chem. internat. Edit. 1 Vol. 6 (1967) 1 No. 6
m / e = 195
A second case is illustrated by Biemann’s observation [301
that the tribenzylamine (40) mass spectrum contains an
appreciable peak at m!e = 181 of composition C14H13. Since
the parent amine (40) does not contain a n uninterrupted
sequence of fourteen carbon atoms, a benzyl rearrangement
must have occurred in the genesis of the m/e 181 fragment.
A third example is the production of a CgHIONZS ion (m/e =
178) from 3-(P-hydroxy-~-phenylethyl)-2-iminothiazolidine
(41) (R = O H ; R = H) in a one-step process (as shown by
the presence of the appropriate metastable ion), which may
be depicted 1311 as a six-center phenyl migration, although
the structural requirements for such a migration must be
fairly specific since no such rearrangement is encountered 1321
[28] C. S . Barnes, M . I. Strong, and .
L. Occolowitz, Tetra-
hedron 19, 839 (1963).
1291 S. D . Sample and C. Djerassi, unpublished.
[30]K . Biernann, Abstracts of 13th Annual Conference on Mass
Spectrometry, ASTM Committee E-14, St. Louis, Mo., 1965,
p. 427.
[31] 3. R. Webster, Chem. Commun. 1966, 124, and personal
[32] Unpublished experiments by Dr. J . K . MacLeod.
48 1
in the related pyrrolidone (42) nor in the ketone (41) (R= 0 ;
R = H) or various N-acetyl derivatives (41) (R’ = COCH3;
R = 0 or OH)[311.
It must be borne in mind that there are many skeletal
rearrangements, which are mechanistically important
and readily detectable by isotopic labeling, but where
the occurrence of such a rearrangement, irrespective
of its abundance, has no bearing on the interpretation
of “element maps”. Such isotopically documented
examples include the loss of a propyl radicalL331
encompassing carbon atoms 2, 3, and 4 of long chain
fatty acid methyl esters (43), the loss of ethylene1341
from positions 2 and 3 (22 %) or 3 and 4 (78 %) in
isohexyl cyanide (44), and finally the apparent ethyl
migration 1351 involved in the expulsion of a methyl
radical in N-propylsuccinimide (451.
c H~4 ~
1- --S
m/e = 72
Alkyl or aryl groups are not the only molecular species
that can suffer migration. When the migrating group
contains a heteroatom, then it follows automatically
that the composition of the resulting peak will afford
misleading information in “element mapping”. Whether such information will be fatal depends on the
number of such rearrangement peaks in one such
molecule, their relative abundance, and whether their
occurrence can be pinpointed as a result of internal
inconsistency of the computer-analysed data. The net
result of such rearrangements, namely a combination
in the fragment ion of heteroatoms and carbon atoms
which is not present in the molecular ion, can also
arise from the expulsion of neutral molecules such as
carbon monoxide, formaldehyde, nitriles, etc. Before
discussing them, it is worth while discussing some
recent examples of oxygen rearrangements.
One of the most abundant ions in the mass spectrum[37aJ of
4-hydroxycyclohexanone (47) corresponds to C2H402 ( m / e =
60). Similarly, the base peak in the spectrum of 4-methoxycyclohexanone (48) is associated with a n ion of mass 74
(C3H.502). The presence of two oxygen atoms in a twocarbon fragment from (47) and in a three-carbon moiety
from (48) requires the intervention of a n oxygen rearrangement in each instance. By labeling all of the positions with
deuterium, it was found [37aI that one of the hydrogen atoms
at C-3 or C-5 becomes equivalent to the C-4 hydrogen atom,
and a reaction path consistent with these data is depicted
(48) R = CH3
R = H,
H~C’ d H
m / e = 60
R = CH3, m / e = 74
These results raise the question of how general such
transannular rearrangements might be and how they
are able to compete with other fragmentation reactions.
In this regard, it is relevant to point out that the
second most intense peak in the mass spectrum of
2-ethylcyclohexanone (49) is due to a standard
McLafferty rearrangement (see (1)
(2)) with loss
of ethylene to give (SO), while in 4-methoxy-2-ethylcyclohexanone ( 5 1 ) this fragmentation is greatly
The expulsion of an ethyl radical[36] in 2-propylisothiocyanate (46) would be misleading in “element
mapping”, but fortunately it is of low abundance.
[331 E. Stenhagen, 2. analyt. Chem. 205, 109 (1964); N. DinhNguyen, R . Ryhage, S. Stallberg-Stenhagen, and E. Stenhagen,
Ark. Kemi 18, 393 (1961).
[34] R. Beugelmans, D . H . Williams, H . Budzikiewicz, and C .
Djerassi, J. Amer. chem. SOC. 86, 1386 (1964).
[35] A. M. Duffeld, H. Budzikiewicz, and C . Djerassi, J. Amer.
chem. SOC.87, 2913 (1965).
[36] E. Bach, A . Kjaer, R . H. Shapiro, and C. Djerassi, Acta
chern. scand. 19,2438 (1965).
(51) ‘CH3
m/e = 74
Aiigew. Chem. internat. Edit. j Vol. 6 (1967) ] No. 6
reduced [37a3 at the expense of the transannuiar
methoxyl rearrangement with eventual production of
the m/e = 74 ion.
together as a contiguous chain in the original molecule, and may prove misleading unless the scope of
these reactions is vigorously explored and delineated.
In the related cyclohexane diol system, rearrangement
ions of composition CH3O2 (m/e = 47) and abundance
c35 = 1.1 % (70 eV) I*] have recently been reported [37b]
for the 1,3- and 1,4-isomers, but no evidence for
hydroxyl migration was found inthel,2-compound[37bl.
Kinetic methods [42,431 have so far proved inapplicable to the interpretation of the mass spectra of all
but the simplest organic compounds; the role of
thermodynamic considerations has consequently been
enhanced. Thus one finds that major fragmentation
pathways can almost always be rationalized (if not yet
predicted) by stabilizing the positive charge (and
radical, if present) at all intermediate stages, and by
finally losing a stable neutral (or radical) species [441.
Methoxyl rearrangements have also been encountered
in mass spectrometric investigations [38al of permethylated sugars, a n instructive example being the genesis
of the m/e = 105 peak in the mass spectrum of methyl
(52), as proved by appropriate deuterium labeling. An analogous
acetoxy migration has been proposed [38bl in the
fragmentation of pyranose penta- and hexa-acetates.
It is not very surprising then, to discover that a large
proportion of electron impact induced skeletal rearrangements are accompanied by expulsion of such
neutral entities as carbon dioxide, carbon monoxide,
sulfur dioxide, formaldehyde, etc., all of which have
favorably negative heats of formation [451 (Table 1).
Table 1. Heats of formation (gas) at 298 OK.
m / e = 105
Rearrangements involving hydroxyl migration have
been observed in benzoic acid[39,40],and in oximes of
benzophenone 1181 and acetophenone 1411; these examples are further discussed below.
3. Rearrangements Accompanied by Ejection of a
Stable Neutral Molecule
Many of the skeletal reorganization processes encountered in mass spectrometry have in common the
loss of a stable neutral moiety from a non-terminal
position in the original molecule. The consequence of
such rearrangements, especially if they occur to a
significant extent, on the element map[lol of a particular compound is obvious. Entries will be recorded
for ions comprising atoms that were not in fact bonded
[37a] M . M . Green, D . 5’.Weinberg, and C . Djerassi, 3. Amer.
chem. SOC.88, 3883 (1966).
[*I C35 signifies the intensity of a peak relative to the sum of
intensities of all peaks from m/e = 35 to the molecular ion peak.
[37b] H. F. Griitzmacher, J . Winkler, and K . Heyns, Tetrahedron
Letters 1966, 6051.
[38a] K . Heyns and D. Miiller, Tetrahedron 21, 55 (1965); N . K .
Kochetkov and 0. S. Chizhov, ibid. 21, 2029 (1965).
[38b] K . Heyns and D. Miiller, Tetrahedron Letters 1966, 6061.
[39] S. Meyerson and J. L. Corbin, J. Amer. chem. SOC.87, 3045
[40] J . H . Beynon, B. E. Job, and A . E. Williams, Z . Naturforsch.
ZOa, 883 (1965).
1411 K . G . Das and P. S. Kulkarni, 14th Annual Conference on
Mass Spectrometry, A.S.T.M. Committee E-14, Dallas. Texas,
Angew. Chem. internat. Edit. ! Vol. 6 (1967)
1 No. 6
a) Loss of Carbon Monoxide
In an important paper, Beynon [461 called attention to
the ejection of carbon monoxide from p-benzoquinones, e.g. anthraquinone (53), and from benzot r o p m e (54) and phenol (55). In each instance, major
peaks result from such rearrangement.
(53) m / e = 208
m / e = 180
m/e = 152
m / e = 66
[42] H. M . Rosenstock, M . B. Wallenstein, A. L. Wahrhaftig, and
H. Eyring, Proc. nat. Acad. Sci. U.S.A. 38, 667 (1952).
1431 See, however, M . M . Bursey and F. W . McLafferty, J. Amer.
chem. SOC.88, 529 (1966); 88, 4484 (1966); F. W . McLafferty,
M . M . Bursey, and S . M . Kimball, ibid. 88. 5022 (1966); M . M .
Bursey and F. W . McLafferty, ibid. 88, 5023 (1966); 89, 1 (1967).
1441 F. W. McLafferty in F. W. McLafferty: Mass Spectrometry
of Organic Ions. Academic Press, New York 1963, p. 318.
[45] Selected Values of Chemical Thermodynamic Properties,
Circular of the National Bureau of Standards No. 500, U.S.
Government Printing Office, Washington, D.C., 1952.
[46] J . H. Beynon, G. R. Lester, and A . E. Williams, J. physic.
Chem. 63, 1861 (1959).
However, the elimination of a neutral molecule from a ring
falls outside our definition of “group migration” (see Introduction). The term “migration” implies movement of a
group from one site t o another, accompanied by bond
breakage and bond formation in the transition state. In
many cases of expulsion of a stable, neutral species from a
cyclic ion, recyclization is not a mechanistic necessity 1471,
although it is almost universally depicted as such in fragmentation schemes.
The perils inherent in assuming that loss of carbon
monoxide from cyclic carbonyl compounds will
generate a simple recyclized ion are apparent from a
study [48aJ of 2-pyrone(56)and its deuterated analogues
(57) and (58). If a symmetrical species such as (59)
were formed, then the mass spectra of [3-D]-2-pyrone
(57) and [6-D]-2-pyrone (58) would be identical. In
fact they were not, gross differences being noted in the
isotope distribution in ions formed by further decomposition of the M-CO species. However, the
(56) R’ =
R’ =
RZ = H
H, R2 = D
(58)R’ = D , R 2 = H
carbon monoxide. Skeletal rearrangements leading to
M-CO ions were not encountered to any significant
extent in dialkyl sulfones and sulfoxides [52,531.
Carbon monoxide loss has also been discerned in the
spectra of acetylacetone [46,54,551 (65), methyl and
ethyl acetoacetate [54,561 (66) and (67), isopropenyl
acetate L571 (68), distyryl ketone 1541 (69), phenyi
styryl ketone[46,541 (70), benzoic acid [39,401 (71),
benzophenone [49,501 f 72), and trichlorotrifluoroacetone[58al (73) (Table 2), and a variety of other carbonyl compounds containing at least one additional site
of unsaturation 158bl.
Table 2. Rearrangement peaks involving loss of carbon monoxide
interpretation of these labeling results by Pirkle 148al
have recently been critically questioned 14%) and the
intermediacy of a cyclic M - CO ion (59) in the
fragmentation of 2-pyrone remains an open question.
Elimination of carbon monoxide occurs somewhat
unexpectedly in the mass spectra of diphenyl
ether [46,49,501 (60) and various aryl substituted sulfones L51.521 and sulfoxides [521. For example, methyl
phenyl sulfoxide 1521 (61), diphenyl sulfoxide [521 (62),
and dibenzothiophene sulfone [51.521 (63) all display
high intensity peaks, corresponding to operation of an
M-CO process. It has been suggested[5*>521that
isomerization of the molecular ion of (63) to that of
the sulfinate ester (64) occurs prior to the ejection of
1’1 ‘ 8 0 labeling showed 1541 that principally the ketone carbonyl is lost
as carbon monoxide.
The operation of a 1,4-phenyl migration associated
with C O expulsion has been uncovered [591 by isotope
labeling studies with 2-phenoxy-4,5-benzotropone(74).
Separate scrutiny of both the 180-labeled carbonyl
compound and the 180-labeled ether analogue disclosed that in the formation of the two most important high mass peaks (M-OH, 26 %; M-CO, 60 %)
both oxygen atoms are involved. In the carbon mon-
m/e = 142
m / e = 141
220 ( 2 0 % )
m / e = 222 (80%)
[47] For evidence concerning acyclic ions in the mass spectra of
aromatic hydrocarbons, see P. Natalis and J . L . Franklin, J.
physic. Chem. 69, 2935 (1965), and references cited therein.
[48a] W . H . Pirkle, J. Amer. chem. SOC.87, 3022 (1965).
[48b] P.Brownand M.M.Green, J. org. Chemistry, 32, 1681 (1967).
[49] P . Natalis and J . L . Franklin, J. physic. Chern. 69, 2943
1501 J. H . D . Eland and C . J. Danby, J. chem. SOC.(London)
1965, 5935.
[51] E. K . Fields and S . Meyerson, J. Amer. chem. SOC.88, 2836
[52] J. H . Bowie, D. H. Williams, S:O. Lawesson, J. 0 . Madsen,
C. Nolde, and G. Schroll, Tetrahedron 22, 3315 (1966).
f 75)
(61) R = en3
(62) R = CsH5
of M-CO
peak ( %)
[531 R. G. Gillis and J . L . Occolowitz, Tetrahedron Letters 1966,
1997; I.D.Entwistle, R.A. W . Johnstone, and B. J.Millard, J. chem.
SOC.(London) C , 1967, 302.
[54] J. H. Bowie, R . Grigg, D . H . Williams, S.-0. Lawesson, and
G . Schroll, Chem. Commun. 1965, 403.
[55] J. H. Bowie, D . H. Williams, S . - 0 . Lawesson, and G . Schroll,
J. org. Chemistry 31, 1384 (1966).
[56] J. H . Bowie, S . - 0 . Lawesson, G . Schroll, and D . H . Williams,
J. Amer. chern. SOC.87, 5742 (1965).
[57] A . S. Newton and P. 0.Strom, J. physic. Chem. 62,24(1958).
[58a] F. W . McLafferty, Appl. Spectroscopy 11, 148 (1957).
[58b] J. H . Bowie, R. G. Cooks, S.-0. Lawesson, P. Jakobsen, and
C. Schroil, Chem. Commun. 1966, 539.
[591 0. L. Chapman, T. H . Kinstle, and M.T. Sung, J. Amer.
chem. SOC. 88,2618 (1966).
Angew. Chem. internat. Edit. I VoI. 6 (1967)
No. 6
oxide loss (M-CO), 8 0 % of t h e oxygen involved
derives from the ether position, which can be readily
accounted for by consideration of the relative stabilities
of the t w o postulated intermediates (76) and (77).
In the loss of OH (M -17) process, the oxygen
originates to approximately equal extent from each
position 1591, implicating again 1,4-phenyl migration
[(74) % (75)l. No thermal phenyl migration was
observed in this system but other instances of photochemically induced phenyl shifts (1,s i n cis-dibenzoylethylene) to oxygen are on record L6*,611.
(82) on electron impact. In both compounds, a-cleavage
affords an ion of mass 75, which then expels carbon monoxide
to give the rearrangement species at m / e = 47. The relative
abundance of this peak is 8 % for (81) and 47 % for (82).
Also falling in this category are the recent reports[64,651 of
carbon monoxide loss from certain n-bonded organometallic
compounds. For example, with the ferrocene derivatives
(83) R
(83), migration of the group R to the metal atom apparently
occurs. This type of rearrangement is common to a wide
range of suitably substituted x-complexes of Fe, Cr, and
Mn 1651.
231 ( M - 1 7 )
A closely related methyl migration has been reported [621 in
the electron impact induced fragmentation of 2-methoxythioanisole (78). An appropriate metastable ion verifies the
expulsion of carbon monosulfide from the M - 15 species
(SO), furnishing an ion of mass 95. In the S-trideuteriomethyl
analogue, approximately 65 % of mje = 95 increases to
m/e = 98, and again a corresponding metastable ion is
observed, indicating that methyl migration between sulfur
(79) and oxygen (80) precedes loss of carbon monosulfide.
O g C H3
A further example of carbon monoxide elimination attended
by skeletal rearrangement is apparent in the fragmentation
of dimethyl acetalc63461 (81) and trimethyl orthoformate 1631
[601 G. W. G r f j n and E. J . O'Connel, J. Amer. chern. SOC.84,
4148 (1962).
[61] H . E. Zimmerman, H . G . C . Durr, R . G . Lewis, and S . Braun,
J. Amer. chem. SOC.84, 4149 (1962).
[62] J . H. Bowie, S . - 0 . Lawesson, J. 0. Madsen, C . Nolde, and
D. H . Williams, J. chem. SOC.(London) B 1966, 951.
Angew. Chem. internat. Edit.
I Vol. 6 (1967) 1 No. 6
(84) R = CH3, CsHs
m / e = 298
Another reorganization process i n t h e mass spectra
of some binuclear transition metal x-complexes has
been observed by ReedC661. Thus t h e chromium
complex (84) decomposes t o an ion of mass 298,
[63a] G. Schroll, H. J . Jakobsen, S . - 0 . Lawesson, P . Brown, and
c, Djerassi, Ark, Kemi 26, 279 (,966),
[63b] M . J . Rix, A . J . C. Wakefield, and B. R . Webster, Chem.
[64] A. Mandelbaum and M . Cais, Tetrahedron Letters 1964,
[65] N. Maor, A. Mandelbaum, and M . Cais, Tetrahedron
Letters 196s, 2087.
1661 F. J . Preston and R . I . Reed, Chem. Commun. 1966, 51.
which in turn loses the elements of CrS2. This behavior
was also noted for the corresponding Fe and Ni
Skeletal rearrangement ions have also been observed [671
spectra of a variety of other transition
in the
metal complexes [(SS), (86), (87)].
Significant rearrangement peaks have been noted in
the mass spectra of all the compounds given in the
table. For all the compounds listed, except the cyanoacetates (94) -(97) and the diethyl malonates (9a), an
unsaturated system is in conjugation with the ester
carbonyl group [691. Effective localization of the charge
on the adjacent group (102) would then provide an
electron deficient site for alkyl migration [102)
(103)], attended by simultaneous loss of carbon dioxide. Only the simpler alkyl groups (e.g. methyl,
ethyl) give significant rearrangement peaks, probably
due to the larger number of alternative reactions
possible in the more complex groups.
In the spectra of tetramethylcyclobutane-l,3-dione
(88) and its photoisomer (89), peaks [681 are seen which
correspond to the sequential loss of two carbon monoxide molecules. Although the structures of the ions
(88)- 9%
(89) 4%
depicted are no less tentative than those proposed in
many mass spectral mechanistic schemes, nevertheless,
carbon-carbon bond formation must accompany or
precede the final ejection of carbon monoxide in order
to furnish the ion of mass 84. The observation of a
metastable peak for the decomposition of the molecular ion to mje = 84 betrays the simultaneous operation of aconcerted M-2C0 process.Certaincorrelations
between the electron impact induced and the photochemical behavior of ketene dimers such as (88) have
been made
b) Loss of Carbon Dioxide
Skeletal rearrangement accompanied by expulsion of
carbon dioxide in the mass spectrometer is a well documented process, occurring particularly in certain
types of esters, and closely related compounds
(Table 3).
1671 J . M . Wilson and M . Haas, Symposium on “Newer Physical
Methods in Structural Chemistry”, Oxford 1966.
[681 N . J.Turro, D. C . Neckers, P. A . Leermakers, D. Seldner,
and P. D’Angelo, J. Amer. chem. SOC.87, 4097 (1965).
However, although the migration of the alcohol R group in
a-p unsaturated esters seems to be a fairly general process [731,
only one case of the converse situation (i.e. a vinyl ester of a
saturated acid) expelling carbon dioxide with concomitant
alkyl shift has been reported [(IGG), Table 31 1121. Migration
of the t-butyl group in phenyl pivalate is observed, but no
corresponding methyl migration in phenyl acetate has been
The substituted cyanoacetates (96) and (97) are
believed [7OJ to suffer synchronous y-cleavage (to the
carbonyl group), alkyl migration, and loss of carbon
m / e = 96 (39%)
m / e = 68 (100%)
[691 J. H. Bowie, D. H. Williams, P. Madsen, G. Schroll, and
S.-0. Lawesson, Tetrahedron 23, 305 (1967).
[lo] J. H. Bowie, R. Grigg, S:O. Lnwesson, P. Madsen, G. Schroll,
and D. H. Williams, J. Amer. chern. SOC.88, 1699 (1966).
[71J J . H. Bowie, D. H. Williams, S . - 0 . Lawesson, and G. Schroll,
J. org. Chemistry 31, 1792 (1966).
[72] F. Weiss, A. Isard, and G. Bonnard, Bull. SOC.chim. France
I965, 2332.
[?3] It must be conceded that although rearrangement occurs
in t-butyl benzoate [12], no significant analogous methyl
migration is apparent with methyl benzoate [12].
Angew. Chem. internat. Edit.
1 Vol. 6 (1967) / No. 6
Table 3. Loss of carbon dioxide from unsaturated carboxylic esters.
abundance %
Rearrangement peak
M - C02
M - HCOz
M - C02
M - HCO2
< 1
M - CO?
M - HCO2
M - C02
M - HCOz
M - C02
- HCO2
M - COz
M - CO2
M - C02
M - CO2
M - HC02
I ::
M - C02
M - HCOl
M - HCO2
- COz
M - HC02
+ C02)
M - HCOz
R 1 = CH3
R2 = CH3
R1 = CH3
R2 = C ~ H S
+ C02)
+ C02)
R1 = H
R2 = n-C3H7
+ C02)
+ C02)
M - CO2
M - CO2
M - HCOz
M - COz
dioxide, as outlined for ethyl s-butylcyanoacetate
(104). Further elimination of ethylene from the rearrangement ion (mass 96) provides one route to the
species (m/e = 68) responsible for the base peak in tha
Both alkyl and aryl group migrations have been
described 1741 in the electron impact induced decomposition of trimethylsilyl esters. Rearrangement peaks
[74] R . M. Teeter, Abstracts Tenth Annual Conference on Mass
Spectrometry, ASTM Committee E-14, New Orleans, La., 1962,
p. 51.
, Vol. 6 (1967) 1 No. 6
R' = H
R2 = i-C,H,
Angew. Chem. internnt. Edit.
[M - (15 + 44)] were apparent in the spectra of the
trimethylsilyl derivatives of octanoic, decanoic, and
octadecanoic acids ( I O S ) , and also alanine. In the case
of aromatic acid derivatives, support for the scheme
(106) -+ M - 15
M - (15 + 44) is provided by
the presence of appropriate metastable peaks, and
also by an increase in the intensity of the M - (15
+ 44) peak relative to M - 15 as R is varied from
an electron withdrawing to an electron donating
group. In the mesitoic acid derivative, the rearrangement is still operative.
+ 44)
Similar rearrangements have been noted 175,761 with
N-substituted ethyl carbamates. Where a-cleavage to
nitrogen is possible [e.g., ethyl N-ethylcarbamate
(107)], it apparently precedes ethyl migration and
carbon dioxide loss. In instances where such cleavage
is prohibited [e.g. ethyl N-phenylcarbamate (IOS)],
the sequence of events appears to be reversed. Deuterium labeling studies [see asterisks in (107) and
( l O S ) ] were undertaken in support of the proposed
m/e = 102 (M-15)
= 1 2 1 (M-44)
R;L-(15 + 44)
+ 15)
Both hydrogen and methyl shifts accompanied by
carbon dioxide elision are found 1771 in the fragmentation of certain cyclic carbonates. Analogous rearrangements with loss of sulfur dioxide are also encountered [771 in the corresponding cyclic sulfites,
further discussed below. As an example of the former,
the two largest peaks in the mass spectrum of butane2,3-diol cyclic carbonate (109) are at n?/e 43 (100 %)
and m/e = 29 (38 %). Each peak is a doublet by exact
mass measurements, indicating that both rearrangement species (110) and ( I l l ) are formed in approximately 1 :3 ratio.
carbonates (114). The rearrangement peak is of high
intensity for the methyl cwnpounds (112) (R = CH3,
53.2 % &) and (113) (R = CH3, 11.8 % &), and
likewise for the diphenyl compound (114) (Ar = C&,
8.3 % c40), but rapidly falls off with the larger R groups.
Evidence was adduced1781 for a two stage decomposition of the dialkyl carbonates (112), with alkyl
migration and loss of carbon dioxide following initial
a-cleavage (see (IIS), diisopropyl carbonate). With the
alkyl phenyl carbonates (113) (except for R = CH3),
both this decomposition path and the converse (initial
loss of carbon dioxide, then cc-cleavage) appear to
operate, while the diaryl compounds (114) expel carbon dioxide 1791, followed by carbon monoxide from
the diaryl ether [461 rearrangement ion.
m / e = 131 (hf-15)
m / e = 87
M-(15 + 44)
A more detailed study has since been undertaken [80al in
a n attempt to clarify further the mechanism of methyl
migration in methyl phenyl carbonate [(I13), R =
CH31. Previous resolts 1781 already indicated that
migration was taking place to aryl oxygen, rather than
to the ortho ring position, and two separate 1 8 0 labeling experiments [see asterisks in (116) and (117)]
This type of rearrangement is also prevalent 1781 in the
mass spectra of symmetrical dialkyl carbonates (112),
alkyl phenyl carbonates (113), and aryl phenyl
C. P. Lewis, Analyt. Chern. 36, 176 (1964).
C. P. Lewis, Analyt. Chem. 36, 1582 (1964).
P. Brown and C. Djerassi, unpublished.
P. Brown and C. Djerassi, J . Arner. chern. SOC.88, 2469
1791 The pyrolytic analogue of this rearrangement has been
reported by P. D. Ritchie, J. chern. SOC.(London) 1935, 1054.
[SOa] P. Brown and C. Djerassi, J. Amer. chern. SOC. 89, 2711
Angew. Chem. internat. Edit.1 VoI. 6 (1967) 1 No. 6
verified [80al that the aryl oxygen was indeed retained
exclusively in the M - 44 rearrangement ion.
By variation of the mrta- and para-substituents [X in
(118)] in the aromatic ring, the propensity for rearrangement was expressed [8OaI in kinetic terms [431,
and a Hammett pa plot of positive slope obtained. A
reasonable interpretation of these data is that methyl
migration takes place in this system preferentially to
the more positively charged site.
In this connection, it is worth recalling that alkyl migration
in dialkyl carbonates [781(112) and ethyl N-ethylcarbamate [751
(107) occurs only after initial a-cleavage. This cleavage
removes the inductively stabilizing effect of a n alkyl group
on the ion radical, and simultaneously generates a full
positive charge o n the “receptor site”. Alkyl migration takes
place in the molecular ion of methyl phenyl carbonate[781
[(I13), R = C H 3 1 and ethyl N-phenyl carbamate[751 (108),
where the phenyl substituent can inductively enhance the
positive character of the “receptor site”. The data for trimethylsilyl derivatives of substituted benzoic acids [741 (106)
also support the hypothesis that the migrating group will be a
nucleophilic species and the “receptor site” electrophilic in
A substituent effect study has also been performed [SOa] on
the electron impact induced decarboxylation of aryl phenyl
carbonates. In this reaction, phenyl migration [(l18a) +
(118b)l was distinguished from aryl migration [(118a) +
(118c)I by 180-labeling. Thus relative migratory aptitudes
of the aryl groups could be assigned (Table 4).
(118~)( M - 4 4 )
(1186) ( M - 4 6 )
It was found that charge distribution in the molecular ioo
and the relative nucleophilicities of the migrating species are
major factors controlling the overall migration tendencies.
Very generally, electron withdrawing substituents on the aryl
group favored phenyl migration, and electron releasing
substituents produced more aryl migration. Correlation of
these migratory aptitudes (1,3-0 to 0 shift) with those
obtained in solution chemistry for the pinacol rearrangement
(1,2-C to C shift) [sob] is only qualitative (see Table 4).
As a logical extension of our previous workL781, we
have studied [811 the relative ease of elimination of
various possible neutral molecules from the series of
all possible methyl phenyl (119) and diphenyl (120)
mono-, di-, and trithiocarbonates (Table 5). In addition
to this rearrangement, a further process involving
Table 5. Rearrangement peaks in the mass spectra of thiocarbonates
(all numerical values refer to C a ) .
1 M - co 1 M - cs I M - coZJ
M - C O S ~ M - cs2
I 11.8 I
I 0.7
I 0.7
isomerization of the molecular ion of compounds
containing the part structure C6H50-CS was disclosed
[see (121)
(122)j. This latter type of rearrangement
has good analogy in the corresponding thermal process.
Table 4. Relative migratory aptitudes of aryl groups in the electron
impact induced decarboxylation of aryl phenyl carbonates.
f coz
< 1
< 1
< 1
< 1
Angew. Cheni. internat. Edit.
< 0.02
< 0.01
< 0.01
< 0.01
-c 0.01
I .9
Theobserved eliminations (Table 5 ) can be rationalized
by locating the charge at a favorable site, and assuming
a relative migratory order of CH3 > O C ~ H S> SCsHs
> C6H5. I n this way, neutral species with negative
heats of formation are preferentially produced (see
Table l),but this factor alone does not exert the sole
VoI. 6 (1967) 1 No. 6
[Sob] W. E . Bachmann and J . W. Ferguson, J. Amer. chem. SOC.
56, 2081 (1934).
[Sl] J. B.Thomson, P. Brown, and C . Djerassi, J. Amer. chem.
S O C . 88, 4094 (1966).
Controlling influence. Thus in (123), carbon dioxide is
lost to a greater extent than carbonyl sulfide, whereas
in (124) carbon monoxide loss prevails over that of
carbonyl sufide or carbon dioxide, and in (125) only
elision of carbonyl sulfide is apparent. In the diphenyl
series, (126) suffers loss of carbon monoxide rather
than carbon dioxide or carbonyl sulfide, and (127)
eliminates carbon monoxide rather than carbonyl
A r-S-A r'
ments in these compounds are more pronounced than
in the dialkyl analoguesL531. In addition t o the unexpected carbon monoxide expulsion discussed in one
of the preceding sections, loss of sulfur dioxide from
sulfones and sulfur monoxide from sulfoxides seems
to require at least one aromatic ring (site of unsaturation) in the molecule. Analogous eliminations of sulfur
dioxide from sulfones have been observed under
radiolytic 1851 and pyrolytic
(600-700 "C) conditions.
Table 6. Some rearrangement peaks in the mass spectra of sulfoxide
and sulfones.
In the course of this work [8*1, it was noted 1771 that
the spectrum of phenyl chloroformate (128) contains
two rearrangement peaks, due to loss of carbon monoxide (4 % relative abundance), and of carbon dioxide (14%).
Two recent reports call attention to the expulsion of
carbon dioxide from certain aromatic imides, e.g. Nmethylphthalimide (128a), R = C H 3 [ 8 2 a J and the Nphenyl analogue (128a), R = C6H5[82b1. (128a),
R = C ~ H ~ - P - C ~also
H ~ ,gives rise to carbon dioxide
elimination, but an o-phenyl substituent (128a), R =
C ~ H ~ - O - C ~ results
H ~ ,
in different fragmentation
pathways I82bI.
M - SO1
M - HS02
c ) Loss
of Sulfur Dioxide and Sulfur Monoxide
The mass spectra of diaryl sulfones [52,83,841 (129) and
sulfoxides [52,841 (130) have recently been published,
and it is apparent (Table 6) that skeletal rearrange[82aJ R. A. W. Johnstone, B. J. Millard, and D . S. Millington,
Chem. Commun. 1966, 600.
[82bl J. L. Cotter and R . A . Dine-Hart, Chem. Commun. 1966,
3,3'-dithiophene sulfone
M - HSOz
M - H2S02
M - HSOz
M - HzS02
- SO2
M - HS02
M - HzSOz
1 M - SO2
1'1 These are only representative examples of a large number of methylsubstituted diphenyl sulfones listed in ref. [831.
A further rearrangement of the unsymmetrjcal diaryl sulfone
(129) molecular ion to sulfinate esters (131) and (132) is
envisagedt831 to account for the presence of aryloxy ions in
the mass spectra. Significantly, the more highly methyl
[83] S. Meyerson, H. Drews, and E. K. Fields, Analyt. Chem. 36,
1294 (1964).
[84]J. 0. Madsen, C . Nolde, S O . Lawesson, G . Schroll, J. H .
Bowie, and D . H . Williams, Tetrahedron Letters 1965, 4377.
185) L. Kevan, P. L . Hall, and E.T. Kaiser, J. physic. Chem. 70,
[86] E. C. Leonard, J. org. Chemistry 30, 3258 (1965).
Angew. Chem. internat. Edit.
Vol. 6 {1967) / No. 6
substituted aromatic nucleus was found to migrate preferentially in every case.
Loss of sulfur dioxide from sulfonamides on electron
impact has also been reported [873 (Table 7). Since the
intensity of the M - SO2 peak increases as the basicity
of Z increases (see (133) and Table 7 ) , and taking into
account steric considerations, a general scheme -can
be proposed (133) -+ (134) + (135)][871.In support
15 8 5
4 3 57
pears to rearrange to the 3,3-dimethyl-2-butanone molecular ion (138), in what may be described as an
electron impact induced pinacol rearrangement. The
peaks at rnle=43 and 57 have the compositions expected, and together comprise 52 0
4 &.
d) Loss of Formaldehyde
(135) M-SOz
of this rationalization, it should be noted that 4aminobenzosulfonamide (Table 6, R = H) exhibits no
M - SO2 peak, whereas with 4-nitrobenzosulfonamide, this process is responsible for the base peak.
More recently, an analogous expulsion of sulfur dioxide from N-n-butyl-N‘-p-toluenesulfonylurea
has been detected [881. By a combination of exact mass
measurements and deuterium labeling techniques, two
fragmentation pathways werk established, the main
process being identical to that described above for
other sulfonamides[(l33) + (134) -+ (135)l. At low
ionizing voltage (19 eV), the relative importance of
the rearrangement was enhanced.
Table 7. Rearrangement peaks in the mass spectra of sulfonamides.
~ - R - C ~ H ~ - S O Z - N H - R [87].
M - SO2
M - SO2
M - SO2
M - SO2
M - SO2
M - SO2
M - SO2
[I .3]-Diazin-2-yl
M - SO2
M - SO2
M - SO2
M - SO2
M - SO2
(140) M - 3 0
R = CzH5, i-C3H7, t - C 4 H g
(14O), that surprisingly involve loss of the u-carbon of
the alcohol. Acetals of secondary and tertiary alcohols,
on the other hand, first suffer wcleavage and then
rearrangement [141)
(142) + (143)], exactly as in
the dialkyl carbonates (112). Alkyl aryl acetals (144)
eliminate formaldehyde directly from the molecular
ion, again paralleling the behavior of the corresponding carbonates (113), in which alkyl migration takes
place preferentially to the most positive site. It should
be noted that the analogous rearrangement process
apparently does not operate in dithioacetals 1901.
( %)
(142) M-15
3 RodQ
As previously mentioned,certain cyclic sulfites undergo
alkyl migration accompanied by sulfur dioxide loss 1771,
analogous to the corresponding cyclic carbonates
losing carbon dioxide [(109)
( I I O ) ] . For example,
2,3-dimethylbutane-2,3-diolcyclic sulfite (137) ap--f
[87] G. Spiteller and R . Kaschnitr, Mh. Chern. 94, 964 (1963).
[88] H . F. Grostic, R. J. Wnuk, and F. K . MacKeilar, 3. Amer.
chern. SOC.88, 4664 (1966).
Angew. Chern. internat. Edit. 1 Vol. 6 (1967) / No. 6
One of the few well documented cases of skeletal
rearrangement with formaldehyde expulsion taking
place in the mass spectrometer has been described [891
for formaldehyde acetals. The primary alcohol
derivatives (139) display low intensity M - CH2O peaks
During the course of deuterium labeling studies on
butyl acetate [911(146),a small but characteristic M - 30
peak was detected. Loss of formaldehyde from butyl
acetate (146) and of acetaldehyde from I-methyl-propyl
acetate (147) was confirmed by precise mass measurements. Similar rearrangements occur with neopentyl
esters of acetic, propionic, benzoic, and some substituted benzoic acids 1921. I n an independent investiga[89] P . Brown, C . Djerassi, G . Schroll, H. J . Jakobsen, and S.-0.
Lawesson, J. Amer. chern. SOC.87, 4559 (1965).
[90] 7’. M. Shuttleworth, Appl. Spectroscopy 18, 78 (1964).
[91] D. R . Black, W. H. McFadden, and J. W. Corse, J. physic.
Chern. 68, 1237 (1964).
(921 W . H . McFadden, K . L. Stevens, S . Meyerson, G . J . Karabatsos, and C. E. Orzech, J. physic. Chem. 69, 1742 (1965).
49 1
ment peak appears at nije = 106 [M -(CH3 + HCN),
% relative abundance], and in N1-methyl-iV1,Nzdiphenylformamidine (156) it is shifted to mje -- 168
[M-(CH3 + HCN)]. By using the 15N-labeled analogue
[(157), see asterisks], it was shown that both methyl
and phenyl migrations are taking place, since some
of the mje = 106 peak moved to m/e == 107.
(146) R'
(147) R'
(148) R'
= CH3,
R2 = H, R3 = C2H5
R 2 = R3
= CH3
= CzH5, R 2 =
H, R3 = C2&
tion 1931 using deuterium labeled esters, expulsion of
formaldehyde from butyl propionate (148) was noted.
The rearrangement ion was written as the ketone
molecular ion (149), since or-cleavage in (149) could
then account simply for an otherwise anomalous
peak (C~H-JO)at mje = 71.
It is noteworthy in this connection that thioesters [943
do not lose thioformaldehyde, but rather thiiranes
(1.53). The rearrangement peak corresponds to 3.8 %
2 4 0 for (150), and 1.2% &o for (1.51).
(155) R = CH,
(156) R = c&
m / e = 106
(150) R' = CH3, R2 = C2&
(151) R' = R2 = C&
In a series of experiments designed to explore the feasibility
of elimination of various neutral species (particularly formaldehyde) from trimethylsilyl derivatives of alcohols, a large
rearrangement peak (62 % relative abundance) was detected [95Jin the mass spectrum of the benzyl alcohol derivative
(154). This process is obviously closely related to the loss of
m / e = 107
The occurrence of two rearrangement peaks at mie =
94 and m /e = 103 in the mass spectrum of benzophenone oxime (1.58) has been attributed [**I to a
hydroxyl migration, producing benzonitrile as one of
the species. Operation of this scheme is supported by
both precise mass measurements and observation of
a metastable peak for the transition.
carbon dioxide by trimethylsilyl derivatives of benzoic
acidr741 (106). A similar study of the effect of substituents
in the aromatic ring on the migratory propensity of the aryl
group in this system is in progress [951.
e) Loss of Hydrogen Cyanide and Nitriles
m / e = 94
( 158)
m / e = 103
Similar rearrangements have been claimed [411 in the
spectra of oximes (1.59), semicarbazones (160), and
thiosemicarbazones (161) of p-substituted acetophenones.
Skeletal rearrangement accompanied by elimination of
hydrogen cyanide as a neutral species has been observed 1961 in the spectra of certain amidines. In N1,NIdimethyl- N2-phenylformamidine (155), the rearrange-
[93] C. Djerassi and C. Fensefau, J. Amer. chern. SOC.87, 5756
1941 W. H . McFadden, R . M . Seifert, and J . Wasserman, Analyt.
Chem. 37, 560 (1965).
I951 .
B. Thomson, J. Diekman, and C . Djerassi, unpublished.
[96] A . K . Bose, I. Kugajevsky, P. 7'. Funke, and K . G. Das, Tetrahedron Letters 1965, 3065.
R = H. CH3. OCH3, Br, NO2
A recent studyC971 of the electron impact induced
fragmentation of alkyl and aryl sulfonylhydrazones
has uncovered some important rearrangement processes which fall in this category. For example,
1971 A . Bhati, R . A. W . Johnstone, and B. J. Millard, J. chem.
SOC.C 1966, 358.
Angew. Chem. internat. Edit. / Vol. 6 (1967) I No. 6
m/e = 92
R = H, CH3, OCOCH,, CsH5
benzenesulfonylhydrazones and toluene-p-sulfonylhydrazones of aromatic aldehydes and ketones (162)
generate characteristic peaks at m/e = 92. With the
cinnammaldehyde derivatives (163) two rearrangement peaks are seen at mje = 118 and mje = 91,
produced by successive eliminations of hydrogen
cyanide. A further reorganization process has been
alkyl substituted analogues has been attributed 162,981 to
the lack of competing reactions possible with the
simplest alkyl group.
Table 8. Skeletal rearrangement peaks in the mass spectra of sulfides
and disulfides.
m / e = 118
m / e = 91
% Relative
Rearrangement peak
M - HS
M - HS
recognized in the acetyl derivative (162a). After conventional cleavage of ketene, a subsequent loss of diimide apparently occurs, furnishing a rearrangement
peak at rnle = 246.
f) Loss of Sulfur
Systematic studies on sulfides [53,62,841 and disulfides [84,98,99a] have clearly shown that many skeletal
rearrangements with attendant sulfur loss are possible
(Table 8). However, as with rearrangements involving
carbon dioxide expulsion[69], it has beenproposed[62,9*J
that an unsaturated site in the molecule is a necessary
prerequisite for significant rearrangement. The absence of skeletal rearrangement peaks in the mass
spectra of higher dialkyl sulfidesE99bl may be cited in
favor of this suggestion. Also, the much greater
proclivity of methyl compounds (e.g. dimethyl sulfide [531, dimethyl disulfide 1981, and thioanisole [621) to
undergo skeletal rearrangement compared with other
M - H2S
_M_- _C S_ _
= n-C3H7
M -4 S
= i-C3H7
M -4 S
Angew. Chem. internat. Edit. J Vol. 6 (1967) j No. 6
Rz = -(CH2)5-
M -2 S
1981 J. H . Bowie, S.-0. Lawesson, J . 0. Madsen, C . Nolde,
G . Schroll, and D. H . Williams, J. chern. SOC.(London) B 1966,
[99a] J . 0. Madsen, S.-0. Lawesson, A . M. Duffild, and C. Djerassi, I. org. Chemistry, in press.
[99b] S . D . Sample and C. Djerassi, J. Amer. chem. SOC. 88,
1937 (1966).
In phenyl substituted sulfides and disulfides, the
C7H73 ion of mass 91 is virtually ubiquitous, occurring
for example to the extent (Table 8) of 25 % with thioanisole (621, and 1 2 % with phenyl ally1 sdfide. It is
also worth noting that although the M - CO process
in diphenyl ether [46,49,501 (60) is a relatively favorable one (15-25 % relative abundance [46,491), with
diphenyl sulfide the corresponding M - CS peak is
much less important (3 %, Table 8).
Thiuramdisulfides (994 such as ( 1 6 3 ~ )have small
peaks corresponding to the loss of four sulfur atoms
(Table 8), In addition, a more important (47 % relative
abundance) rearrangement peak appears at m/e = 109,
which can be most simply visualized as arising by
phenyl migration to sulfur, followed by elimination of
the elemencs of methyl isothiocyanate.
A significant peak (10 % relative abundance) at m/e =
93 (M - 30) in the mass spectrum of nitrobenzene
(166) has been rationalized by the assumption 11021
that prior isomerization o f the molecular ion to that
of phenyl nitrite (167) occurs. No complementary
rearrangement is apparent with aliphatic nitro-compounds [1031.
In the spectrum of the triphosphatriboracyclohexane [lo41 (168), the base peak appears at m/e = 41,
which has been attributed to the formation of dimethylboron ion (169). Interestingly, the appearance
potential for this species is greater than 20 eV.
m / e = 109
A small peak (0.2 % &o) at m / e = 149 detected[losl in the
spectrum of N,N-diphenyl phenylacetamide (170) has been
interpreted by invoking a phenyl migration, and substituent
and deuterium labeling results (but no exact mass measurements) are cited as corroborative evidence.
Miscellaneous Rearrangements
A rearrangement peak at m/e = 91, amounting to 20 %
relative abundance, has been reported [loo] in the mass
spectra of a series of alkyl aryl phosphonates (164).
(164) R
H, CH3. CzH,
Many relatively intense peaks in the spectra of triaryl
phosphates 11011 [e.g. triphenyl phosphate (165)] must
involve complex skeletal rearrangements. For example,
peaks at mle = 228 (12 % relative abundance), mie =
170 (40 %), and m/e = 152 (8 %) have been attributed[lOl] to the manifestation of such events. The
trialkyl phosphates, on the other hand, fragment more
predictably, with hydrogen transfer processes predominating.
m/e = 152
[loo] ff.Budzikiewicz and Z . Pelah, Mh. Chem. 96, 1739 (1965).
In the same paper [1057, an electron impact induced “Beckmann rearrangement” was claimed to operate in the fragmentation of benzophenone oxime (158),due to the presence
m / e = 228
m / e = 170
m / e = 194
A . Quay/e, Advances Mass Spectrometry I, 365 (1959).
of a n otherwise anomalous peak at mle = 105 ( 4 % &a),
which would be expected to result from cleavage in the Beckmann product benzanilide (171). However, repetition of this
[lo21 J. H . Beynon, R. A . Saunders, and A . E. WiNiams, Ind.
chim. belg. 1964, 311.
[lo31 R.T. Aplin, M. Fischer, D . Becher, ff. Budzikiewicz, and
C . Djerassi, J. Amer. chem. SOC.87, 4888 (1965).
[lo41 R . E. Florin, L. A . Wall, F. L. Mohler, and E. Quinn, J.
Amer. chem. SOC. 76, 3344 (1954).
[lo51 P.T. Funke, K. G. Das, and A . K . Bose, J. Amer. chem.
SOC.86, 2527 (1964).
Angew. Chem. internat. Edit. I Vol. 6 (1967)
1No. 6
work1181 suggests that the m / e = 105 peak probably originates
from benzophenone (172) present as impurity in the oxime
sample used.
The more mechanistically feasible claim is also made “051
that pinacol (173) (and related 1,2-diols) may suffer an
electron impact induced pinacol rearrangement. Thus peaks
at M - 18, m / e = 85 (1.6 % &,),
m / e = 57 (2.5 %), and
m / e = 43 (0.8%) are observed in the spectrum of pinacol
( 1 73), but apparently no precise mass determinations have
been carried out.
51 43
Migration of fluorine accompanied by skeletal rearrangement gives rise to major peaks in the mass
spectra of perfluoroxylenes [1061, trifluoromethyl substituted benzenes [lo71 [e.g., (174)], and aroyl fluorides 11071 [e.g., (175)]. The enormous difference in
CH3 11.1
C2H5 4.5
n - C 3 H ~ 3.6
n-C4HB 1.5
m / e = 59
The occurrence of benzyl migrations in the mass
spectra of benzyloxycarbonyl derivatives of certain
amino-acid esters has been reported [1121. Important
peaks (4-52 % relative abundance) were recorded for
a series of mainly dipeptide methyl (X = C H 3 ) , ethyl
(X = CzH5), and phenylthio (X = C~HSS)esters,
corresponding to rearrangement ions (179). It seems
very likely that the rearrangement step (loss of carbon
dioxide) is closely related mechanistically to that
occurring in ethyl carbamate derivatives [75,761.
ease of elimination of C F 2 compared with C H 2 from
organic compounds under electron impact has already
been noted 11081. This trend is further illustrated by the
rearrangement with C F 2 elimination which is suffered [lo91 by simple perfluorohydrocarbons.
A large peak (25 %relative abundance) in the spectrum
of benzylidene malononitrile (176) at mle = 103 has
been attributed [1101, on the basis of precise mass
measurements, to the formal loss of the elements of
cyanoacetylene from the molecular ion, accompanied
by migration of a nitrile group.
R = R‘ = Phenyl
R = Phenyl, R’ = f3-Naphthyl
R = f3-Naphthyl, R’ = Phenyl
R = 2,4,6-Trimethylphenyl. R’= Phenyl
11061 J . R. Majer, J. appl. Chem. 11, 141 (1961).
11071 J. R. Majer, Symposium on “Newer Physical Methods in
Structural Chemistry”, Oxford 1966.
[I081 Cf. [44], p. 330.
(1091 J. H. Beynon: Mass Spectrometry and Its Applications to
Organic Chemistry. Elsevier Publishing Co., Amsterdam 1960,
p. 416.
[110] D. H . Williams, Symposium on “Newer Physical Methods
in Structural Chemistry”, Oxford 1966.
[ I l l ] M. Vandewalle, N . Schamp, and M . Francque, Bull. SOC.
chim. Belge, in press.
Angew. Chem. internat. Edit. 1 VoI. 6 (1967) / No. 6
Extensive rearrangement in the molecular ion of a
series of aromatic azoxy compounds (180) is indicated I1131 by their proclivity to eliminate carbon monoxide followed by loss of nitrogen on electron impact.
The rearranged ion in the case of (180) R = R’ =
phenyl seems to be different from the molecular ion
of o-hydroxyazobenzene, since this species expels
nitrogen followed by carbon monoxide.
R - N=N- R’
The base peak for a series of acyclic 1,3-diketone enol
ethers (177) occurs[1111 at mle = 59, and is due to an
ion of composition C 2 H 3 0 2 @ . Such a species requires
formal migration of the methoxy function, and the
presence of appropriate metastable peaks betray its
origin as M - R fragments. The contribution of
C z H 3 O 2 @ to the total ion current decreases as the
size of R increases.
An important rearrangement process, often generating
the base peak of the spectrum, in glycidic esters and
amides (181) has been noted[1141, involving loss of
C2H02 from the molecular ion. As might be expected from the proposed mechanism, this rearrangement was not encountered 11151 in 1,2-epoxyketones
[e.g., (181) X-R = alkyl).
(I8l) X
0,NH2, NHR
[112] R. T. Aplin, J. H. Jones, and B. Liberek, Chem. Commun.
1966, 794.
[113] J. H. Bowie, R. G. Cooks, and G . E. Lewis, Chem. Commun. 1967, 284.
[114] J. Baldas and Q . N . Porter, Chem. Commun. 1966, 571.
11151 W. Reusch and C . Djerassi, Tetrahedron, in press.
4. Summary
In view of the relatively high incidence of electron
impact induced skeletal rearrangements recognized
during the past year, it seems highly probable that
many additional examples await discovery in the near
future. All such instances will be of intrinsic mechanistic interest, irrespective of the probability of the
process, but only the more abundant peaks will be of
serious consequence to the interpretation of “element
Elimination of stable, neutral species (e.g. CO, COz,
SOz, HCN, CH20, etc.) from a non-terminal position
in aromatic or other unsaturated molecules presently
provides the largest collection of reorganization processes. In many cases, the migrating group may be
visualized as a nucleophilic species, moving preferentially to the most positive site available. Operation of the same rearrangement in the molecular ion
of the corresponding saturated analogues is usually
markedly reduced, or even absent. Prior cleavage, in
order to develop a full positive charge, is often necessary to induce a subsequent group migration. The
most effective migrating groups appear to be those
which have little tendency to suffer fragmentation,
such as methyl and phenyl.
From a consideration of trends already apparent, it
seems likely that thermal, photochemical, and acid
catalysed rearrangements will act as important guides
in the design of electron impact induced parallels. To
date, it is a fairly safe assertion that good fortune has
been involved in the detection of most, if not all,
recorded mass spectral skeletal rearrangements.
Critical comparison of such processes with free-radical
and carbonium ion chemistry is required, and one
avenue which may prove to be profitable in this
connection, presently under active investigation,
involves the study of relative migratory aptitudes.
Received: August Sth, 1966
[A 583 IE]
German version: Angew. Chem. 79, 481 (1967)
Reactions with Molecular Oxygen
Molecular oxygen is an important oxidizing agent both in industrial and in biological
processes. In many of these processes, the 0 2 molecule reacts preferentially with free
radicals, which are frequently paramagnetic metal ions. Homogeneous systems of this
type that are described in the literature are discussed, and an attempt is made to establish
the factors that determine the course and mechanism of the reaction.
1. Introduction
According to its position in the electrochemical series,
molecular oxygen is a powerful oxidizing agent. The
calculated standard potential for the redox equilibrium
(1) i s +1.23 V[**J. However, reactions with molecular
02+4 H + + 4e-
2 H20
(1 )
oxygen at room temperature generally proceed very
slowly in the gas phase as well as in homogeneous
solution. This is even more surprising in view of the
fact that, in the ground state, the oxygen molecule
possesses two unpaired electrons.
Prof. Dr. S. Fallab
lnstitut fur Anorganische Chemie der Universitat Basel
Spitalstr. 51
Basel (Switzerland)
[**I The standard potentials of the following redox systems are
o f the same order of magnitude: PbZ+/PbOz:+1.46 V; CI-/Cl, :
+1.36 V; Cr3+/Cr2072-: +1.33 V; Iz/IO3-: +1.20V; Br-/Br2:
+1.07 V; V02+/V02+: 1.00 V.
The stepwise course of the reduction (1) accounts in
part for the slowness of reaction. The full redox potential is realized only when the oxidation can be carried
out in a more or less synchronous four-electron step
of the type that is assumed to play a part in enzymatic
oxidations [I]. Such processes require a substrate with
a specifically fixed steric position, and this is presumabIy possible only in a macromolecular environment.
If these conditions are not satisfied, the reduction of
the oxygen molecule must take place in two-electron
steps (2) and (3), or even in one-electron steps (4) to (7).
Since the standard potential of reaction (2) is only
+0.68 V[21, the 0 2 molecule is a much weaker oxidizing agent when the structure of the substrate, or the
medium, compels the oxidation to proceed in twoelectron steps. The standard potentials assigned to the
four one-electron steps (4) to (7) are -0.32, 11.68,
[l] J. H. Wung and W. S. Brinigur, Proc. nat. Acad. Sci. 46, 958
(1960); T. Nakumura, Biochim. Biophysica Acta 42, 499 (1960).
[2] P. George: Oxidases and Related Redox Systems. Wiley, New
York 1965, p. 3.
Angew. Chem. internat. Edit. 1 Vol. 6 (1967)I No. 6
Без категории
Размер файла
1 700 Кб
rearrangements, reaction, induced, organiz, molecules, electro, impact
Пожаловаться на содержимое документа