Competitive charge-remote and anion-induced fragmentations of the non-8-enoate anion. A charge-remote reaction which co-occurs with hydrogen scrambling

Article abstract of DOI:10.1039/a607437e

The non-8-enoate anion undergoes losses of the elements of C₃H₆, C₄H₈ and C₆H₁₂ on collisional activation. The mechanisms of these processes have been elucidated by a combination of product ion and labelling (²H and ¹³C) studies, together with a neutralisation reionisation mass spectrometric study. These studies allow the following conclusions to be made. (i) The loss of C₃H₆ involves cyclisation of the enolate anion of non-8-enoic acid to yield the cyclopentyl carboxylate anion and propene. (ii) The loss of 'C₄H₈' is a charge-remote process (one which proceeds remote from the charged centre) which yields the pent-4-enoate anion, butadiene and dihydrogen. This process co-occurs and competes with complex H scrambling, (iii) The major loss of 'C₆H₁₂' occurs primarily by a charge-remote process yielding the acrylate anion, hexa-1,5-diene and dihydrogen, but in this case no H scrambling accompanies the process. (iv) It is argued that the major reason why the two charge-remote processes occur in preference to anion-induced losses of but-1-ene and hex-1-ene from the respective 4- and 2-anions is that although these anions are formed, they have alternative and lower energy fragmentation pathways than those involving the losses of but-1-ene and hex-1-ene; viz. the transient 4-anion undergoes facile proton transfer to yield a more stable anion, whereas the 2-(enolate) anion undergoes preferential cyclisation followed by elimination of propene [see (i) above].

Full text of DOI:10.1039/a607437e

View Article Online / Journal Homepage / Table of Contents for this issue
Competitive charge-remote and anion-induced fragmentations of the
non-8-enoate anion. A charge-remote reaction which co-occurs with
hydrogen scrambling
Suresh Dua,a John H. Bowie,*,a Blas A. Cerda,b Chrys Wesdemiotis,b
Mark. J. Raftery,a Julian F. Kelly,a Mark S. Taylor,a Stephen J. Blanksbya and
Mark A. Buntinea
a
Department of Chemistry, The University of Adelaide, South Australia, 5005, Australia
b Department of Chemistry, The University of Akron, Akron, Ohio 44325-3601, USA
The non-8-enoate anion undergoes losses of the elements of C3H 6, C4H 8 and C6H 12 on collisional
activation. The mechanisms of these processes have been elucidated by a combination of product ion and
labelling (2H and 13C) studies, together with a neutralisation reionisation mass spectrometric study. These
studies allow the following conclusions to be made. (i) The loss of C3H 6 involves cyclisation of the enolate
anion of non-8-enoic acid to yield the cyclopentyl carboxylate anion and propene. (ii) The loss of ‘C4H 8’ is
a charge-remote process (one which proceeds remote from the charged centre) which yields the pent-4-
enoate anion, butadiene and dihydrogen. This process co-occurs and competes with complex H scrambling.
(iii) The major loss of ‘C6H 12’ occurs primarily by a charge-remote process yielding the acrylate anion,
hexa-1,5-diene and dihydrogen, but in this case no H scrambling accompanies the process. (iv) It is argued
that the major reason why the two charge-remote processes occur in preference to anion-induced losses of
but-1-ene and hex-1-ene from the respective 4- and 2-anions is that although these anions are formed, they
have alternative and lower energy fragmentation pathways than those involving the losses of but-1-ene
and hex-1-ene; viz. the transient 4-anion undergoes facile proton transfer to yield a more stable anion,
whereas the 2-(enolate) anion undergoes preferential cyclisation followed by elimination of propene [see
(i) above].
that this should result in competition between charge-remote
Introduction
fragmentation of the carboxylate anion, and anion-directed
fragmentation from the allylic centre. We will show that both
charge-remote and anion induced processes do indeed occur in
this system, but not as a consequence of fragmentation of an
allylic anion.
Collision-induced loss of a neutral from an even-electron nega-
tive ion in the gas phase often involves reaction involving the
charged centre.1,2 There are also fragmentations of even-
electron anions which are ‘charge-remote’, viz. reactions which
occur remote from the charged centre. Amongst the latter
are some radical losses which form stabilised radical anions.3,4
Substantiated reports of charge-remote reactions involving the
loss of even-electron neutrals from even-electron anions are
not common. The classical charge-remote mechanistic proposal
for alkanoate anions is shown in reaction (1). This was first
Results and discussion
We have described above why we chose to study an enoate anion
for this study. The first task is to choose which enoate anion is
the most suitable for study. The mass spectra of a number of
such anions are listed in Table 1. The spectra of the lower
homologues are dominated by loss of carbon dioxide and are
not suitable for this study. In contrast, the higher homologues
show the type of losses previously noted in the spectra of
alkanoate anions,5–7,12 for example, losses of the elements of
C3H6, C4H8 and C6H12. In principle, these losses could all be
charge-remote reactions analogous to that shown in reaction
(1), except of course that the neutral products will be a diene
and dihydrogen. We have chosen to probe the mechanisms of
these processes using the non-8-enoate anion as a suitable
example. The investigation uses a combination of product ion,
labelling (D and 13C), neutral reionisation mass spectrometry
(NRMS) and computational studies.
proposed by Adams and Gross,5–8 but the mechanism has been
questioned.9–11 Support for the charge-remote mechanism has
been provided by Cordero and Wesdemiotis12 who have used
neutralisation reionisation mass spectrometry (NRMS) to
identify butene as a neutral product of the reaction shown in
reaction (1). This charge-remote process has a large activation
barrier; estimates range from ca. 200 kJ molϪ1 (experimental5–8
)
to 385 kJ molϪ1 (computational13).
(a) The evidence based on product ion and labelling studies
The collision-induced negative chemical ionisation tandem
mass spectrum (MS/MS) of the (M Ϫ D)Ϫ ion of CH ᎐CH–
This paper reports the fragmentations of the non-8-enoate
anion, i.e. a system similar to the alkanoate shown in reaction
(1), except that it has double bond functionality at the opposite
end of the molecule to the carboxylate moiety. We anticipated
that collision-induced transfer of a proton from the allylic pos-
ition to the carboxylate anion would yield an allylic anion, and
᎐
2
(CH2)6CO2D is recorded in Fig. 1. The first part of this investi-
gation involves the identification of the structures of the prod-
uct ions formed following losses of ‘C3H6’, ‘C4H8’ and ‘C6H12’.
This was done by comparing (i) MS/MS/MS fragmentation
J. Chem. Soc., Perkin Trans. 2, 1997
695

View Article Online
Table 1 MS/MS data from CH2᎐᎐CH(CH2)nCO2Ϫ ions
Loss
Ϫ
CH2᎐᎐CH(CH2)nCO2
n
Formation
CH2CO2
H2O a CO2 C3H6 C4H8 C5H10 C6H12 C8H16
Ϫ
ؒ
Hؒ
H2
1
2
3
4
5
6
8
100
65
100
100
100
54
3
10
34
72
16
72
51
88
100
29
5
7
10
1
2
9
5
10
15
20
5
2
8
17
100
100
5
3
14
48
39
56
10
22
10
5
a The loss of H2O is a standard reaction of carboxylate species and proceeds following proton transfer to form the enolate anion.14 To take a specific
example relevant to the enoic acids: CH2᎐᎐CHCH2CD2CO2Ϫ loses HOD and D2O in the ratio 2:1. The processes are as follows: CH2᎐᎐CHCH2CD2-
CO2Ϫ → CH2᎐᎐CHCH2 CDCO2D → [(CH2᎐᎐CHCH2CD᎐C᎐O) DO ] → CH2᎐᎐CH CHCD᎐C᎐O ϩ HOD and CH2᎐᎐CHCH2C᎐COϪ ϩ D2O.
Ϫ
Ϫ
Ϫ
᎐
᎐
᎐
᎐
᎐
᎐
However, if the loss of ‘C3H6’ involves a ‘charge-remote’ pro-
cess, the ionic product will be CH ᎐CH(CH ) CO Ϫ. The data in
᎐
2
2
3
2
Table 2 identify the product ion as the cyclopentyl carboxylate
anion and therefore the loss of C3H6 must involve a cyclisation
process.
We now wish to study the losses of ‘C3H6’, ‘C4H8’ and ‘C6H12’
in detail, and to assist with this endeavour, we have prepared
carboxylate anions from five D2 compounds [those indicated
by an * in 1] and one 13C compound [that depicted by # in 1].
The MS/MS data derived from all labelled species are recorded
in Table 3.
The cyclisation process—loss of C3H6. It has already been
shown that the product anion of this process is the cyclopentyl
carboxylate anion. The deuterium labelling results indicate that
both the hydrogens at the 7 position and one hydrogen at pos-
ition 5 are lost as part of C3H6. These data may be rationalised
as shown in Scheme 2. Transfer of a proton from position 2 to
Fig. 1 The collision-induced negative chemical ionisation tandem
mass spectrum (MS/MS) of the (M Ϫ D)Ϫ ion of CH2᎐᎐CH(CH2)6-
CO2D. VG ZAB 2HF instrument. For experimental conditions, see
Experimental section.
data of each product anion [using an MS 50 TA instrument
(through the courtesy of Dr R. N. Hayes and Professor M. L.
Gross)] and (ii) MS/MS fragmentation data of source formed
product ions (using the VG ZAB 2HF instrument), with the
MS/MS data of known ions produced by deprotonation of the
appropriate neutral molecules. These data are recorded in Table
2. Results are summarised in Scheme 1. (B2/E) Linked scans of
Scheme 2
the carboxylate anion produces the key enolate intermediate (2),
which undergoes the cyclisation process shown. We have drawn
the cyclisation process [see intermediate 2] in synchronous fash-
ion because of the deuterium labelling results {the alternative
stepwise process in which an intermediate [(cycloC5H9CO2H)
Ϫ
ϪCH CH᎐CH ] decomposes to yield cyclo-C H CO
and
᎐
2
2
5
9
2
CH CH᎐CH , would involve the loss of one H from position 2
᎐
3
2
(the precursor enolate anion is formed following proton trans-
fer from position 2 to the carboxylate anion centre), not, as
observed, from position 5}.
The losses of ‘C4H8’ and ‘C6H12’. The two most plausible mech-
anisms for the loss of ‘C4H8’ are the charge-remote reaction (2)†
Scheme 1
† There are anionic processes like those summarised in formulae (3) and
(4), which would give the same products as those formed by charge-
remote process (2). These processes are less favourable on energetic
grounds than that shown in reaction (3). In addition, we will show later,
that such ionic processes do not occur because the precursor 4-anion
preferentially undergoes facile proton transfer to form carboxylate (and
enolate) anions.
source formed m/z 71, 99 and 113 confirm that each is formed
directly from the parent (M Ϫ H)Ϫ ion.
The product ions shown in Scheme 1 which result from the
losses of ‘C4H8’ and ‘C6H12’ are those that would be formed
from the appropriate charge-remote process [cf. reaction (1)].
696
J. Chem. Soc., Perkin Trans. 2, 1997

View Article Online
Table 2 MS/MS and MS/MS/MS data for some product ions from the non-8-enoate anion
Precursor ion (m/z)
Product ion (m/z) Spectrum type Spectrum [m/z (loss) abundance]
Ϫ
Ϫ
CH2᎐᎐CH(CH2)6CO2
(155)
CH2᎐᎐CH(CH2)3CO2
᎐C3H6
(113)
MS/MS/MSa
MS/MS
MS/MS
111 (H2) 100, 71 (C3H6) 60
ؒ
111 (H2) 100, 71 (C3H6) 65, 44 (C5H9 ) 4
ؒ
112 (H ) 100, 111 (H2) 10, 95 (H2O) 34, 71 (C3H6) 2
ؒ
(113)
69 (CO2) 29, 58 (C4H7 ) 10
Ϫ
ؒ
ؒ
CH3CH2CH᎐CHCH2CO2
MS/MS
MS/MS
112 (H ) 18, 97 (CH4) 8, 84 (C2H5 ) 4, 69 (CO2) 100,
᎐
ؒ
ؒ
(113)
67 (CO2 ϩ H2) 4, 58 (C4H7 ) 2, 53 (CO2 ϩ CH4) 3, 44 (C5H9 ) 3
Ϫ
ؒ
ؒ
CH3(CH2)2CH᎐CHCO2
112 (H ) 68, 111 (H2) 28, 95 (H2O) 14, 84 (C2H5 ) 62, 69 (CO2) 100
᎐
(113)
ؒ
MS/MS
112 (H ) 8, 111 (H2) 2, 95 (H2O) 100, 85 (28) 8,
83 (CH2O) 64, 71 (C3H6) 6, 69 (CO2) 2, 57 (C4H8) 2, 43 (70) 1
ؒ
MS/MS
111 (H2) 100, 71 (C3H6) 72, 44 (C5H9 ) 5
MS/MS/MSa
MS/MS
98 (H ) 70, 81 (H2O) < 10 , 58 (C3H5 ) < 10 , 55 (CO2) 100
Ϫ
b
b
ؒ
ؒ
CH2᎐᎐CH(CH2)6CO2
(155)
CH2᎐᎐CH(CH2)2CO2
᎐C4H8
(99)
Ϫ
ؒ
ؒ
98 (H ) 65, 97 (H2) 5, 81 (H2O) 10, 58 (C3H5 ) 7, 55 (CO2) 100
(99)
CH2᎐᎐CH(CH2)6CO2
(155)
CH2᎐᎐CHCO2
᎐C6H12
(71)
MS/MS/MSa
MS/MS
70 (H ) 100, 27 (CO2) 35
Ϫ
ؒ
Ϫ
ؒ
ؒ
70 (H ) 100, 44 (C2H3 ) 4, 27 (CO2) 29
(71)
a
b
The MS/MS/MS spectra are weak: peaks < 5% are lost in baseline noise. The spectra are weak—the abundances of these peaks are only
approximate.
Table 3 MS/MS data for labelled non-8-enoate anions
Precursor ion (m/z)
Spectrum [m/z (loss) abundance]
Ϫ
a
ؒ
ؒ
CH2᎐᎐CH(CH2)5CD2CO2
(157)
156 (H ) 55, 155 (D ) 32, 138/137 (HOD, D2O) 15,
115 (C3H6) 100, 101 (C4H8) 30b, 102 (C4H7D) 25b, 72 (C6H11D) 48, 60 (C7H13 ) 22
ؒ
Ϫ
ؒ
ؒ
CH2᎐᎐CH(CH2)3CD2(CH2)2CO2
156 (H ) 65, 155 (D ) 15, 139 (H2O) 38, 115 (C3H6) 100,
ؒ
(157)
101 (C4H8) 15, 100 (C4H7D) 30, 71 (C6H10D2) 45, 58 (C7H11D2 ) 24
Ϫ
ؒ
CH2᎐᎐CH(CH2)2CD2(CH2)3CO2
(157)
156 (H ) 70, 139 (H2O) 28, 114 (C3H5D) 100, 101 (C4H8) 15, 100
ؒ
(C4H7D) 8, 71 (C6H10D2) 38, 58 (C7H11D2 ) 18
Ϫ
b
ؒ
CH2᎐᎐CHCH2CD2(CH2)4CO2
(157)
156 (H ) 45, 139 (H2O) 22, 115 (C3H6) 100, 101 (C4H8) 8 ,
100 (C4H7D) 15b, 99 (C4H6D2) 25b, 71 (C6H10D2) 52, 58 (C7H11D2 ) 25
ؒ
Ϫ
ؒ
ؒ
CH2᎐᎐CHCD2(CH2)5CO2
(157)
156 (H ) 30, 155 (D ) 30, 139 (H2O) 28, 113 (C3H4D2) 100,
100 (C4H7D) 15b, 99 (C4H6D2) 29b, 71 (C6H10D2) 38, 58 (C7H11D2 ) 18
ؒ
Ϫ
12
CH2᎐᎐CHCH213CH2(CH2)4CO2
(156)
155 (H ) 50, 138 (H2O) 15, 114 ( C3H6) 100,
ؒ
99 (12C313CH8) 28, 71 (12C513CH12) 12, 58 (12C613CH13 ) 5
ؒ
a
Composite peak not resolved. b Peaks not fully resolved.
and the anionic process (3).‡ Deuterium labelling studies
should resolve the problem as to which of processes (2) and (3)
is operative. In practice, this turns out not to be the case, since
the deuterium labelling data [Table 3, and cf. 1] indicate a par-
ticularly complex scenario. These data show that hydrogens at
positions 2, 4, 5, 6 and 7 have partially scrambled prior to or
during the loss of ‘C4H8’. Hydrogen scrambling in anion sys-
tems can, in principle, occur by one of two general processes, viz.
(i) by specific proton transfer processes occurring to particular
anionic sites in the molecule, or (ii) by some type of skeletal
rearrangement of the carbon skeleton, or (iii), some combin-
ation of both (i) and (ii). The latter scenario is reminiscent of
the fragmentations of positively charged long chain carboxylate
esters, where both H and C rearrangements abound.15 We pre-
pared the 6-13C labelled derivative of non-8-enoic acid in order
to distinguish between these possibilities. If carbon skeletal
rearrangement is occurring in this negative ion system, both
‘12C313CH8’ and ‘12C4H8’ will be lost from the 6-13C labelled
‡ An alternative anionic process involves the reaction of the 4-anion
shown in (5).† This would result in the formation of but-1-ene and the
enolate anion of pent-4-enoic acid. We were unable to find a transition
state for this reaction using AM1 calculations and have not considered
the process further.
J. Chem. Soc., Perkin Trans. 2, 1997
697

View Article Online
Fig. 2 Semi-empirical AM1 calculations for charge-remote reaction
(2). Energy (kJ molϪ1) reaction coordinate plot. For computational
procedures, see Experimental section.
Fig. 4 ϪNfRϩ spectra of the CID neutrals from (a) the non-8-enoate
anion, and (b) the nonanoate anion. VG AutoSpec instrument. For
experimental procedures, see Experimental section.
firm the hydrogens involved in the loss of ‘C6H12’. No detectable
hydrogen scrambling competes with this loss.
The evidence to date does not allow us to differentiate
between charge-remote and anion-induced mechanisms for
‘C4H8’ or ‘C6H12’ loss.
Fig. 3 Semi-empirical AM1 calculations for anionic mechanism reac-
tion (3). Energy (kJ molϪ1) reaction coordinate plot.
(b) The evidence based on neutralisation reionisation mass
spectrometry
parent anion. The (M Ϫ H)Ϫ ion from 6-13C-non-8-enoic acid
shows exclusive loss of ‘12C313CH8’: no carbon scrambling pre-
cedes or accompanies this or any other fragmentation (see
Table 3).
The charge-remote and anionic processes described above
can be differentiated based on the neutral molecules they
release. For example, the charge-remote reaction (2) produces
butadiene plus dihydrogen, whereas the anionic mechanism
reaction (3) yields but-1-ene. Which neutrals are cleaved can be
found by a NRMS-type experiment, i.e. by post-ionising the
neutral CID losses and directly detecting them in the corre-
sponding neutral fragment-reionisation, ϪNfRϩ, spectrum.12,18
That of the non-8-enoate anion is shown in Fig. 4(a).
For the interpretation of this spectrum, it is important to
realise that all neutrals eliminated from the collisionally-
activated non-8-enoate precursor anion (see Fig. 1) are post-
ionised simultaneously. Each of the neutrals produces several
ions upon collision-induced dissociative ionisation (CIDI),19
the resulting ϪNfRϩ spectrum arising by superimposition of the
individual CIDI spectra.18 Because of this convolution, spectral
interpretation and structural assignments are substantially
facilitated if the ϪNfRϩ spectrum of the unknown [Fig. 4(a)] is
compared against the collisional ionisation spectra of relevant
reference molecules. An alkene and an alkadiene [Figs. 5(a) and
5(b)] as well as the neutral losses from the saturated nonanoate
anion [see Fig. 4(b)] were chosen as the pertinent references in
this case and will be discussed first.
Since we cannot distinguish between reactions (2) and (3) by
deuterium labelling, perhaps computational studies may assist.
We are not in a position to do high level ab initio calculations
with systems of such complexity, but estimates of the reaction
coordinate profiles for reactions (2) and (3) have been made
using semi-empirical calculations (at AM1 level) using GAUS-
SIAN 9416 and GAMESS-US17 (see Experimental section for
full details of the procedures used), and these are summarised
in Figs. 2 and 3. The data shown in Figs. 2 and 3 should only be
used qualitatively, but they do indicate (i), that both processes
are energetically unfavourable with high activation barriers [an
earlier AM1 calculation of the activation energy of a charge-
remote reaction of an alkanoate anion gave a value of 385 kJ
molϪ1 (cf. the value in Fig. 2 of 372 kJ molϪ1), even though the
value was nominally reduced using an experimental correc-
tion 13], and (ii) the anionic process has the lower barrier to the
transition state, it is less endothermic overall, but it is the more
complex in terms of the number of steps involved. It would be
unwise using these data, to propose which of reactions (2) or
(3) is the more kinetically favoured.
Turning now to the loss of ‘C6H12’: this process could occur
by either the charge-remote reaction (4) or from the enolate
anion as shown in reaction (5). Since the same hydrogens are
eliminated in both processes, deuterium labelling will not dis-
tinguish between the two possibilities. The data in Table 3 con-
Collisional ionisation spectra of hex-1-ene and hexa-1,5-diene.
Depending on the decomposition mechanisms of the non-8-
enoate anion (see above), the major neutral CID fragments are
either alkenes or alkadienes. The types of ions formed upon
collisional ionisation of these neutrals can be found by
698
J. Chem. Soc., Perkin Trans. 2, 1997

View Article Online
CnH2nϪ1ϩ (m/z 27, 41, 55, 69 and 83), but much less (if any) alkyl
ions CnH2nϩ1ϩ (m/z 29, 43, 57, 71 and 85). These spectral charac-
teristics reveal that the formal CnH2nϩ2 losses cannot be alkanes
ؒ
or alkyl radicals plus H , because such species would have pro-
ϩ
duced considerable CnH2nϩ1 after reionisation.12 As discussed
in the foregoing section for hex-1-ene, alkenyl ions are diagnostic
of alkenes.12,21 Thus the ϪNfRϩ data indicate that the major
neutrals eliminated from the nonanoate anion are CnH2n ϩ H2,
in agreement with the charge-remote fragmentation depicted in
reaction (1). Parallel conclusions were reached from the ϪNfRϩ
data of other saturated fatty acids.12
The relative intensities in the reference ϩNRϩ spectrum of hex-
1-ene and the ϪNfRϩ spectrum of nonanoate [Figs. 5(a) and
4(b)] do not match. This is because the latter also contains con-
tributions from other smaller alkenes as well as from other CID
neutrals (see below). Also, alkenes formed by charge-remote
fragmentations are energised because of the large reverse acti-
vation energy associated with such processes (cf. Fig. 2). This
ϩ
Fig. 5 ϩNRϩ spectra of the molecular cations from (a) hex-1-ene and
(b) hexa-1,5-diene. VG AutoSpec instrument.
explains why C2H<2nϪ1 fragments, which lie higher in energy
ϩ 23
than C2H2nϪ1
are more dominant in the ϪNfRϩ spectrum
[Fig. 4(b)] than in the reference ϩNRϩ spectrum [Fig. 5(a)]. For
the same reason, alkene molecular ions (CnH2n ϩ) have a lower
᝽
relative abundance in the ϪNfRϩ spectrum than in reference
ϩNRϩ spectra. For example, compare m/z 84 (C6H12 ϩ) in Figs.
᝽
ϩ
ؒ
4(b) and 5(a) (the peak height of C6H12 is artifically enhanced
in the reference ϩNRϩ spectrum owing to the small peak width
of precursor ions, which are not subject to kinetic energy
release. In contrast, the C6H12 ϩ peak in the ϪNfRϩ spectrum is
᝽
much wider because it originates from the larger nonanoate ion,
and through a process of large reverse activation energy that
leads to substantial kinetic energy release).
Fig. 6 The CID mass spectrum of the nonanoate anion. VG AutoSpec
instrument.
Three nominal radical losses are observed, namely the losses
ؒ
ؒ
ؒ
ؒ
of C5H11 , C7H11 and C8H17 (see Fig. 6). Reionised C5H11
yields a detectable C5H11ϩ ion (m/z 71), whereas only a trace of
C7H15ϩ (m/z 99) is formed, and there is no C8H17ϩ (m/z 113) above
noise level. It is possible that the larger radicals decompose
more extensively upon post-ionisation. More likely, the CID
ions of m/z 58 and 44 are not formed in one step by elimination
neutralisation-reionisation(ϩNRϩ)of theirmolecularions.ϩNRϩ
experiments were conducted with hexa-1,5-diene and hex-1-ene,
the larger neutral fragments generated by reactions (4) and (5).
Owing to the high kinetic energy, and hence superior trans-
mission and reionisation efficiencies of large neutrals, such
moieties generally provide dominant contributions to NfR
spectra.18
ؒ
ؒ
of C7H15 or C8H17 respectively, but originate by consecutive
processes, e.g. from m/z 86 by losses of C2H4 or C3H6.
CID neutrals from the non-8-enoate anion. Comparison of the
ϪNfRϩ spectra of the non-8-enoate and nonanoate anions [Figs.
4(a) and 4(b)] shows that the former contains more unsaturated
product ions. Fig. 4(a) includes dominant peaks at m/z 54, 67
and 81 but barely any heavier ions within the corresponding
peak groups. These products (several of which belong to the
CnH2nϪ3ϩ series) also appear in the ϩNRϩ spectrum of hexa-1,5-
diene [Fig. 5(b)] and are characteristic for alkadienes. The small
relative abundances of m/z 55, 69 and 83 [see insets in Fig. 4(a)],
which are indicative of alkenes [see previous section and Fig.
5(a)], further attests that these alkenes are not an important
component in the neutral loss mixture released from the non-
8-enoate anion.
The ϩNRϩ spectrum of ionised hex-1-ene [Fig. 5(a)] contains
a prominent alkenyl ion series, viz. CnH2nϪ1ϩ (m/z 27, 41, 55, 69).
These ions, which also appear in the EI spectrum of hex-1-ene
[inset in Fig. 5(a)],20 are diagnostic for alkenes.21 The given
alkenyl ions are less important or even absent in the ϩNRϩ spec-
trum of ionised hexa-1,5-diene [Fig. 5(b)]. Instead, this spec-
trum includes a sizeable CnH2nϪ3ϩ series (m/z 39, 53, 67, 81) and
other highly unsaturated ions, consistent with the higher
degree of unsaturation in the diene.21 Highly unsaturated frag-
ments also dominate the EI spectrum of hexa-1,5-diene20 [see
inset in Fig. 5(b)]. It can be seen from Fig. 5, that alkenes and
alkadienes can be distinguished from a consideration of their
collisional ionisation products. It should be noted that the
relative fragment ion abundances in ϩNRϩ and EI spectra differ:
this is a consequence of the different internal energy distri-
butions deposited in parent ions by neutralisation-reionisation
compared with electron ionisation.22 Thus collisional ionisation
reference spectra are most suitable for comparisons with NfR
data.
The sizeable m/z 54 signal (C4H6 ϩ) in Fig. 4(a) is due to the
᝽
elimination of butadiene, consistent with the operation of
charge-remote reaction (2). Similarly, the negligible relative
ϩ
intensities of m/z 55 and 56 (C4H7 and C4H8 ϩ) indicate the
᝽
loss of but-1-ene (C4H8) from the non-8-enoate anion to be an
unfavourable process.
The abundant peak corresponding to m/z 81 (C6H9ϩ) in Fig.
CID neutrals from the nonanoate anion. The majority of the
fragment ions generated upon CID of the (M Ϫ H)Ϫ ion from
the saturated nonanoic acid correspond to nominal CnH2nϩ2
losses (marked by an * in Fig. 6). The only important fragments
which do not belong to this series are those at m/z 139 (H2O
4(a), shows a small shoulder corresponding to m/z 82 (C6H10 ϩ).
᝽
The presence of these two products confirms that C6H10 is elim-
inated from the non-8-enoate anion, supporting the operation
of charge-remote process 5. That the neutral is hexa-1,5-diene
is confirmed by the reference collisional ionisation spectrum
shown in Fig. 5(b). All ions present in the reference spectrum
are seen in the ϪNfRϩ spectrum of non-8-enoate [Fig. 4(a)]. The
abundance ratio of m/z 54:67 is larger in the ϪNfRϩ than the
ϩNRϩ spectrum, because the former includes a contribution
from C4H6 loss [reaction (2)]. As explained earlier, the large
ؒ
ؒ
ؒ
loss), 86 (C5H11 loss), 58 (C7H15 loss) and 44 (C8H17 loss).
Overall, the CnH2nϩ2 losses make up the largest fraction of the
neutral CID fragments and should therefore be the principle
contributor of the ions observed in the ϪNfRϩ spectrum [Fig.
4(b)]. The latter contains a relatively abundant alkenyl series
J. Chem. Soc., Perkin Trans. 2, 1997
699

View Article Online
reverse activation energy of charge-remote reactions supplies
excess energy to the neutral products, promoting extensive
fragmentation upon reionisation and reducing peak resolution
[cf. Figs. 4(a) and 5(b)].
reaction (3), since the enolate anion precursor [in reaction (5)] is
some 130 kJ molϪ1 more negative in energy than that shown for
the precursor 4-anion [in reaction (3)] (see Fig. 3). This leads
to the final question: why then does not anionic reaction (5)
occur to the exclusion of charge-remote reaction (4)? The
answer must be that the enolate anion preferentially undergoes
the cyclisation process already described (see Scheme 2) rather
than reaction (5).
The alkenyl ions C6H11ϩ and C6H12 ϩ (m/z 83 and 84) are of
᝽
very small abundance in the ϪNfRϩ spectrum of the non-8-
enoate anion [Fig. 4(a)]. The nonanoate anion loses hexene,
and its ϪNfRϩ spectrum [Fig. 4(b)] shows a small peak corre-
spondng to C6H11ϩ. Based on the relative ratios of m/z 83:81 in
Figs. 4(a) and 4(b), we conclude that the loss of hex-1-ene from
the non-8-enoate anion is, at best, a minor process compared
with that resulting in loss of hexa-1,5-diene.
Experimental
Mass spectrometric methods
Thus ϪNfRϩ data are consistent with (i) charge-remote process
(2) operating to the exclusion of anionic process (3), and (ii) the
operation of charge-remote process (4) with the possibility of a
minor contribution from anionic process (5).
Collisional activation (CID) mass spectra (MS/MS) were
determined with a VG ZAB 2HF mass spectrometer.25 Full
operating details have been reported.26 Specific details are as
follows: the chemical ionisation slit was used in the chemical
ionisation source, the ionising energy was 70 eV, the ion source
temperature was 100 ЊC, and the accelerating voltage was 7 kV.
Non-8-enoic acid was introduced into the source via the direct
probe with no heating [measured pressure of sample 1 × 10Ϫ6
Torr (1 Torr = 133.322 Pa)]. Deprotonation was effected using
HOϪ (from H2O) or DOϪ (from D2O) for deuterium-labelled
derivatives. The measured pressure of H2O or D2O was 1 × 10Ϫ5
Torr. The estimated source pressure was 10Ϫ1 Torr. Argon was
used in the second collision cell (measured pressure, outside the
cell, 2 × 10Ϫ7 Torr), giving a 10% reduction in the main beam,
equivalent to single collision conditions.
Summary and conclusions
The above study shows that competitive charge-remote and
anion-induced processes occur following collisional activation
of the non-8-enoate anion.
In summary:
(i) The loss of C3H6 is a reaction of the enolate anion (formed
following proton transfer to the carboxylate centre), which
effects cyclisation to yield the cyclopentyl carboxylate anion
and propene. Hydrogen scrambling does not accompany this
process.
(ii) The loss of ‘C4H8’ is a charge-remote process resulting
in the formation of the pent-4-enoate anion together with
butadiene and dihydrogen. This reaction co-occurs with exten-
sive proton transfer processes occurring between various
carbanions on the carbon skeleton.
MS/MS/MS tandem mass spectra were measured (through
the courtesy of Professor M. L. Gross and Dr R. N. Hayes)
with an MS TA 50 mass spectrometer as described previously.27
The pressure of He collision gas in both collision cells was
1 × 10Ϫ6 Torr, producing a reduction in the main beam of
30%.
The NRMS-type experiments were conducted with an E1BE2
tandem mass spectrometer (VG AutoSpec at Akron) that
has previously been described.28 This instrument houses two
collision cells (C-1 and C-2) and an intermediate deflector in the
interface region between E1B and E2. The carboxylate anions
(M Ϫ H)Ϫ from nonanoic acid and non-8-enoic acid were
formed by FAB ionisation, using a 20 keV Csϩ ion gun and
triethanolamine as the matrix. The (M Ϫ H)Ϫ precursor anions
were accelerated to 8 keV, selected by E1B and dissociated
with He in C-1. CID coproduces ionic and neutral fragments.
After exiting C-1, the ionic fragments and any undissociated
(M Ϫ H)Ϫ ions were deflected from the beam path, and the
remaining neutral losses were post-ionised in C-2 by collision-
induced dissociative ionisation with O2.19 The newly formed
cations were mass-analysed by E2 and recorded in the neutral
fragment-reionisation, ϪNfRϩ, spectrum.12,28 The superscripts
besides N and R in the NfR notation designate the charge of
the precursor ion (from which the neutrals are eliminated)
and the charges of the ultimate product ions (to which the
neutrals are reionised), respectively.18 Neutralisation-
reionisation (ϩNRϩ) spectra of the molecular cations of hex-1-
ene and hexa-1,5-diene were measured similarly by replacing
He in C-1 (which causes CID) with trimethylamine (which
causes charge exchange neutralisation).18 In these experiments,
molecular cations were generated by electron impact, and were
accelerated to the kinetic energy with which hex-1-ene and
hexa-1,5-diene would be eliminated from 8 keV nonanoate and
8 keV non-8-enoate, respectively (4.2 to 4.3 keV).
(iii) The major process producing overall loss of ‘C6H12’ is a
charge-remote reaction which yields the acrylate anion together
with hexa-1,5-diene and dihydrogen. A minor accompanying
anionic process may involve loss of hex-1-ene. No detectable
hydrogen scrambling accompanies this (these) process(es).
In conclusion:
(a) the observation of a charge-remote fragmentation
[Ϫ(C4H6 ϩ H2)] co-occurring with anionic proton scrambling
processes is unique in negative ion chemistry. The fact that
charge-remote reaction (2) occurs in preference to the anionic
process may be rationalised as follows. The 4-carbanion, whose
formation is inferred from deuterium labelling data, rearranges
by facile proton transfer (to form either or both of the stable
enolate or carboxylate anions) in preference to the reaction
shown in reaction (3) (there is ample precedent for unstable
carbanions undergoing proton transfer to form more stable
anions24). We have tested whether such a process is feasible by
the use of AM 1 semiempirical calculations. When the 4-
carbanion (see Fig. 3) achieves the conformation shown in (6),
proton transfer to form the non-8-enoate anion is immediate:
no stable hydrogen bonded intermediate is formed during this
process.
(b) The situation concerning loss of ‘C6H12’ is arguably even
more complex. NRMS data are consistent with the charge-
remote reaction (4) being predominant, and the lack of com-
petitive hydrogen scrambling in this instance implies that the
process must have a lower activation barrier than that of the
corresponding reaction (2). Further, anionic reaction (5) must
have a lower activation barrier than that of the corresponding
Syntheses of unlabelled and labelled precursor molecules
Hex-1-ene, hexa-1,5-diene, but-3-enoic acid, nonanoic acid and
ε-caprolactam were commercial products. The following com-
pounds were made by reported procedures: pent-4-enoic acid,29
hex-5-enoic acid,30 hept-6-enoic acid,29 oct-7-enoic acid,29 non-
8-enoic acid 31 and cyclopentane carboxylic acid.32
2,2-[2H2]pent-4-enoic acid. This was prepared as for the
unlabelled analogue,29 except that hydrolysis of the precursor
700
J. Chem. Soc., Perkin Trans. 2, 1997

View Article Online
ester was carried out with NaOD and the final acidification by
D2SO4 in D2O. Overall yield 75%, D2 = 90%.
obtained from the frequency calculations. Calculated zero-
point energies tend to overestimate actual energies by up to ca.
15% and are therefore scaled by 0.89.46 The reaction barriers
were determined by comparing zero point-corrected energies of
the appropriate local minima and transition state. Intrinsic
reaction coordinate (IRC) calculations from the transition state
were undertaken to confirm that the calculated transition state
geometry does indeed connect the relevant local minima on the
overall reaction potential energy surface.
The following computational protocol was employed in this
study. Optimised geometries for reactants and products were
determined using GAUSSIAN 94. The molecular geometry of
the transition state linking local minima on the reaction poten-
tial energy surface (Fig. 3) were then determined using the
Linear Synchronous Transit (LST) approach.47 The structure
resulting from an LST calculation was subsequently used
as input for a saddle point geometry optimisation using
GAMESS. The saddle point geometry determined from the
GAMESS calculation was then input, along with the reactant
and product geometric specifications, into a QST3 transition
state optimisation using GAUSSIAN 94. QST3 Transition state
optimisations employ the Synchronous Transit-Guided Quasi-
Newton (STQN) method developed by Schlegel and co-
workers.48 Finally, IRC calculations were used to confirm that
computed transition state geometries do connect the reactants
and products of interest.
2,2-[2H2]Non-8-enoic acid. Non-8-enoic acid 33 (0.5 g) was
heated under reflux in MeOD–MeONa [Na (0.2 g) in MeOD (5
cm3)] for 15 h. After being cooled to 25 ЊC, the solvent was
removed in vacuo, the residue dissolved in water (5 cm3), cooled
(0 ЊC), acidified to pH 1 with aqueous hydrogen chloride (10%),
and extracted with diethyl ether (5 cm3). Removal of the sol-
vent gave the desired product (95% yield; D1 = 10, D2 = 90%).
4,4-[2H2]Non-8-enoic acid. Diethyl succinate was reduced 34
with lithium aluminium deuteride in refluxing tetrahydrofuran
to yield 1,1,4,4-[2H4]butane-1,4-diol (yield 85%), which was
tosylated 35 with tosyl chloride§ (1 equiv.) in anhydrous pyridine
to yield
a mixture of mono- and di-tosylated 1,1,4,4-
[2H4]butane-1,4-diol. The mono-tosylated product was ob-
tained in 60% yield following column chromatography [on
silica eluting with diethyl ether–hexane (1.5:8.5)]. The mono-
tosylate was coupled 36 with pent-4-ene magnesium bromide to
yield 4,4-[2H2]non-8-enol (yield 72%) which was oxidised 37 with
chromium trioxide to give the desired 4,4-[2H2]non-8-enoic acid
(overall yield 55%; D2 = 99%).
5,5-[2H2]Non-8-enoic acid. This was synthesised by a similar
methodology as that used above for 4,4-[2H2]non-8-enoic acid.
Glutaric anhydride was reduced with lithium aluminium
deuteride to obtain 1,1,5,5-[2H4]pentane-1,5-diol. Overall yield
of 5,5-[2H2]non-8-enoic acid, 45%; D2 = 99%.
6,6-[2H2]Non-8-enoic acid. This was prepared by first coup-
ling38 4,4-[2H2]-4-bromobut-1-ene39 with 5-magnesium bromide
pentanol tetrahydropyranyl ether 38 to obtain 6,6-[2H2]non-8-
enol tetrahydropyranyl ether in 67% yield. 6,6-[2H2]Non-8-enol
was formed following deprotection of the tetrahydropyranyl
ether with acid,40 and then oxidised 37 to give 6,6-[2H2]non-8-
enoic acid in 50% yield (D2 = 99%).
Acknowledgements
We thank the Australian Research Council (J. H. B.) and both
the National Institutes of Health of the U.S.A., and the Ohio
Board of Regents (B. A. C. and C. W.), for the financial support
of this project. Two of us (S. D. and M. J. R.) thank the ARC
for research associate positions. Our thanks also to Dr R. N.
Hayes and Professor M. L. Gross for MS/MS/MS data.
7,7-[2H2]Non-8-enoic acid. This was synthesised using a simi-
lar methodology as that used for 4,4-[2H2]non-8-enoic acid.
1,1,7,7-[2H4]Heptane-1,7-diol was prepared by reduction of
dimethyl pimelate with lithium aluminium deuteride.34 Overall
yield of 7,7-[2H2]non-8-enoic acid, 55% (D2 = 99%).
References
6-13C-Non-8-enoic acid. The reaction 41 between allyl bromide
(1.2 g) and Cu13CN (1.1 g, 13C = 99%)42 gave allyl (13C cyanide)
in 70% yield. The allyl (13C cyanide) was hydrolysed 43 to the
labelled vinyl acetic acid (65% yield), which was reduced 34 with
lithium aluminium hydride in refluxing tetrahydrofuran to form
1-13C-but-3-en-1-ol in 78% yield. The alcohol was tosylated 44
with tosyl chloride. The tosylate was purified by column chro-
matography over silica in diethyl ether–hexane (1.5:8.5),
coupled 36 with 5-magnesium bromide pentanol tetrahydro-
pyranyl ether 38 to yield 6-13C-non-8-enol tetrahydropyranyl
ether, which was then deprotected to give the alcohol39 and then
oxidised 37 to 6-13C-non-8-enoic acid (overall yield from 1-13C-
but-3-en-1-ol, 40%: 13C = 99%).
1 J. H. Bowie, Mass Spectrom. Rev., 1990, 9, 349.
2 J. H. Bowie, in Experimental Mass Spectrometry, ed. D. H. Russell,
Plenum Press, New York and London, 1994, pp. 1–38.
3 P. C. H. Eichinger and J. H. Bowie, Int. J. Mass Spectrom. Ion
Processes, 1991, 110, 123.
4 S. Dua, J. H. Bowie and J. C. Sheldon, J. Chem. Soc., Perkin Trans. 2,
1994, 543.
5 J. Adams and M. L. Gross, J. Am. Chem. Soc., 1986, 108, 6915.
6 J. Adams and M. L. Gross, J. Am. Chem. Soc., 1989, 111, 435, and
references cited therein.
7 J. Adams, Mass Spectrom. Rev., 1990, 9, 141.
8 M. L. Gross, Int. J. Mass Spectrom. Ion Processes, 1992, 118/119,
137.
9 V. H. Wysocki, M. H. Bier and R. G. Cooks, Org. Mass Spectrom.,
1988, 23, 627.
10 V. H. Wysocki, M. M. Ross, S. R. Horning and R.G. Cooks, Rapid
Commun. Mass Spectrom., 1988, 2, 214.
Computational methods
11 V. H. Wysocki and M. M. Ross, Int. J. Mass Spectrom. Ion
Processes, 1991, 104, 179.
Semi-empirical molecular orbital calculations using the AM1
model Hamiltonian 45 were performed with the GAUSSIAN
9416 and GAMESS-US17 systems of programs. Calculations
were performed on a variety of computational platforms,
including DECStation 5000/25 and Silicon Graphics Indigo 2
xZ workstations, as well as Thinking Machines CM-5 and
Silicon Grapics Power Challenge supercomputers.
Optimised molecular geometries were characterised as local
minima or transitions states by subsequent vibrational fre-
quency calculations. Molecular geometries representing local
minima possess all positive vibrational frequencies; transition
states are identified as possessing one (only) imaginary fre-
quency as well as one (only) negative eigenvalue of the Hessian
matrix. The zero-point energy for each structure was also
12 M. M. Cordero and C. Wesdemiotis, Anal. Chem., 1994, 66, 861.
13 M. M. Siegel and N. B. Colthup, Appl. Spectrosc., 1988, 42, 1214.
14 K. M. Downard, J. C. Sheldon, J. H. Bowie, D. E. Lewis and
R. N. Hayes, J. Am. Chem. Soc., 1989, 111, 8112, and references
cited therein.
15 W. Sonneveld, D. Van der Steen and H. J. J. Pabon, Recl. Trav.
Chim., Pays-Bas, 1968, 87, 1110; Ng. Dinh-Nguhen, Arkiv. Kemi,
1968, 28, 289.
16 Gaussian 94, Revision C.3, M. J. Frisch, G. W. Trucks,
H. B. Schlegel, P. M. W. Gill, B. G. Johnson, M. A. Robb,
J. R. Cheesman, T. Keith, G. A. Petersson, J. A. Montgomery,
K. Raghavachari, M. A. Al-Latham, V. G. Zakrzewski, J. V. Ortiz,
J. B. Foresman, J. Cioslowski, B. B. Stefanov, A. Nanayakkara,
M. Challacombe, C. Y. Peng, P. V. Ayala, W. Chen, M. W. Wong,
J. L. Andres, E. S. Replogle, R. Gomperts, R. L. Martin, D. J. Fox,
J. S. Binkley, D. J. Defrees, J. Baker, J. P. Stewart, M. Head-Gordon,
C. Gonzalez and J. A. Pople, Gaussian Inc., Pittsburgh, PA, 1995.
§ Tosyl chloride = 4-methylphenylsulfonyl chloride.
J. Chem. Soc., Perkin Trans. 2, 1997
701

View Article Online
17 M. W. Schmidt, K. K. Baldridge, J. A. Boatz, S. T. Elbert,
M. S. Gordon, J. H. Jensen, S. Koseki, N. Matsunaga, K. A.
Nguyen, S. Su, T. L. Windus, M. Depuis and J. A. Montgomery,
J. Comput. Chem., 1993, 14, 1347.
18 M. J. Polce, S. Beranova, M. J. Nold and C. Wesdemiotis, J. Mass
Spectrom., 1996, 31, 1073.
19 P. C. Burgers, J. L. Holmes, A. A. Mommers, J. E. Szulejko and
J. K. Terlouw, Org. Mass Spectrom., 1984, 19, 422.
20 F. W. McLafferty and D. B. Stauffer, Wiley/NBS Registry of Mass
Spectral Data, 1989, Wiley, New York.
21 F. W. McLafferty and F. Turecek, Interpretation of Mass Spectra,
University Science Books, Mill Valley, CA, 4th edn., 1993.
22 S. Beranova and C. Wesdemiotis, J. Am. Soc. Mass Spectrom., 1994,
5, 1093.
23 S. G. Lias, J. E. Bartmess, J. F. Liebman, J. L. Holmes, R. D. Levin
and W. G. Mallard, Gas Phase Ion and Neutral Thermochemistry,
J. Phys. Chem. Ref. Data 17, 1988, Suppl. 1. The computer-based
version was used.
31 H. Stetter and W. Dierichs, Chem. Ber., 1952, 85, 1061.
32 D. W. Gohen and W. R. Vaughan, Org. Syn. Coll. Vol. 4, 1963, 594.
33 T. H. Chan and D. Stossel, J. Org. Chem., 1986, 51, 2423.
34 A. Burger, L. Turnbull and J. G. Dinwiddie, J. Am. Chem. Soc.,
1950, 72, 5512.
35 A. I. Vogel, Textbook of Practical Organic Chemistry, Longman and
Green Pty Ltd., 4th edn., 1978, p. 1103.
36 G. Fouquet and M. Schlosser, Angew Chem., Int. Ed. Engl., 1974,
13, 82.
37 E. J. Eisenbraun, Org. Syn. Coll. Vol. 5, 1973, 310.
38 J. H. Babler and B. J. Invergo, J. Org. Chem., 1979, 44, 3723.
39 A. Padwa, W. F. Rieker and R. J. Rosenthal, J. Org. Chem., 1984, 49,
1353.
40 E. J. Corey, H. Niwa and J. Knolle, J. Am. Chem. Soc., 1978, 100,
1942.
41 J. V. Supniewski and P. L. Salzberg, Org. Syn. Coll. Vol. 3, 1955,
851.
42 H. J. Barber, J. Chem. Soc., 1943, 79.
24 S. T. Graul and R. R. Squires, J. Am. Chem. Soc., 1990, 112, 2506
and references cited therein; K. M. Downard, R. N. Hayes and
J. H. Bowie, J. Chem. Soc., Perkin Trans. 2, 1992, 1815.
25 J. K. Terlouw, P. C. Burgers and H. Hommes, Org. Mass Spectrom.,
1979, 14, 307.
43 E. Reitz, Org. Syn. Coll. Vol. 3, 1955, 851.
44 L. Brandsma and H. D. Verkrnijsse, Synthesis of Ethylenes, Allenes
and Cumulenes, Elsevier, New York, 1981, p. 223.
45 M. J. S. Dewar, E. G. Zoebisch and E. F. Healy, J. Am. Chem. Soc.,
1985, 107, 3902.
26 G. W. Adams, J. H. Bowie and R. N. Hayes, J. Chem. Soc., Perkin
Trans. 2, 1991, 689.
27 D. J. Burinsky, R. G. Cooks, E. K. Chess and M. L. Gross, Anal.
Chem., 1982, 54, 295; M. L. Gross, E. K. Chess, P. A. Lyon,
F. W. Crow, S. Evans and H. Tudge, Int. J. Mass Spectrom. Ion
Phys., 1982, 42, 574.
46 W. Hehre, L. Radom, P. v. R. Schleyer and J. A. Pople, Ab Initio
Molecular Orbital Theory, Wiley, New York, 1986.
47 T. A. Halgren and W. N. Lipscomb, Chem. Phys. Lett., 1977, 49,
225.
48 C. Peng and H. B. Schlegel, Israel J. Chem., 1993, 33, 449.
28 M. J. Polce, M. M. Cordero, C. Wesdemiotis and P. A. Bott, Int.
J. Mass Spectrom. Ion Processes, 1992, 113, 35.
29 R. P. Linstead and H. N. Rydon, J. Chem. Soc., 1933, 580.
30 R. P. Linstead and H. N. Rydon, J. Chem. Soc., 1934, 1995.
Paper 6/07437E
Received 1st November 1996
Accepted 3rd December 1996
702
J. Chem. Soc., Perkin Trans. 2, 1997