Gas phase and condensed phase SNi reactions. The competitive six and seven centre cyclisations of the 5,6-epoxyhexoxide anion. A joint experimental and ab initio study. A comparison with SNi reactions of homologous epoxyalkoxide anions

Article abstract of DOI:10.1039/a808853e

A. Ab initio calculations [at the MP2 Fc/6-31+G(d) level of theory] indicate that the barriers to the transition states for the competitive six and seven centre S_Ni cyclisation processes of the 5,6-epoxyhexoxide anion are 35.0 and 39.7 kJ mol^-1 respectively. Experimental studies show that (i) in solution, the 5,6-epoxyhexoxide anion cyclises (and at the same time opens the ethylene oxide ring) to give tetrahydropyran-2-methanol as the predominant product on workup, and (ii) collisional activation of the 5,6-epoxyhexoxide anion in the gas phase gives the 2-tetrahydropyranmethoxide anion as the exclusive anionic product. It is proposed that frequency factors (Arrhenius A factors) control the courses of these kinetically controlled gas phase reactions. A comparison of the calculated harmonic vibrational partition functions for the two possible transition states confirms a higher value of Q_Vib for the reaction proceeding through the six-membered transition state. B. A comparison is made of the reported competitive S_Ni reactions for 2,3-epoxypropoxide, 3,4-epoxybutoxide, 4,5-epoxypentoxide and 5,6-epoxyhexoxide anions. For all but the 3,4-epoxybutoxide system, the exclusive or major product is that which contains the smaller of the two ring systems for both gas phase and condensed phase reactions. In the case of the 3,4-epoxybutoxide system: (i) in the gas phase, both four and five membered ring S_Ni products are formed in comparable yield, and (ii) in the condensed phase, the major product is that with the larger ring.

Full text of DOI:10.1039/a808853e

View Article Online / Journal Homepage / Table of Contents for this issue
Gas phase and condensed phase S i reactions. The competitive
N
six and seven centre cyclisations of the 5,6-epoxyhexoxide anion.
A joint experimental and ab initio study. A comparison with SNi
reactions of homologous epoxyalkoxide anions†
John M. Hevko, Suresh Dua, John H. Bowie and Mark S. Taylor
Department of Chemistry, The University of Adelaide, South Australia, 5005, Australia
Received (in Cambridge) 12th November 1998, Accepted 22nd December 1998
A. Ab initio calculations [at the MP2 Fc/6-31ϩG(d) level of theory] indicate that the barriers to the transition states
for the competitive six and seven centre S i cyclisation processes of the 5,6-epoxyhexoxide anion are 35.0 and 39.7
N
Ϫ1
kJ mol respectively. Experimental studies show that (i) in solution, the 5,6-epoxyhexoxide anion cyclises (and at
the same time opens the ethylene oxide ring) to give tetrahydropyran-2-methanol as the predominant product on
workup, and (ii) collisional activation of the 5,6-epoxyhexoxide anion in the gas phase gives the 2-tetrahydropyran-
methoxide anion as the exclusive anionic product. It is proposed that frequency factors (Arrhenius A factors) control
the courses of these kinetically controlled gas phase reactions. A comparison of the calculated harmonic vibrational
partition functions for the two possible transition states confirms a higher value of QVib for the reaction proceeding
through the six-membered transition state.
B. A comparison is made of the reported competitive S i reactions for 2,3-epoxypropoxide, 3,4-epoxybutoxide,
N
4
,5-epoxypentoxide and 5,6-epoxyhexoxide anions. For all but the 3,4-epoxybutoxide system, the exclusive or major
product is that which contains the smaller of the two ring systems for both gas phase and condensed phase reactions.
In the case of the 3,4-epoxybutoxide system: (i) in the gas phase, both four and five membered ring S i products are
N
formed in comparable yield, and (ii) in the condensed phase, the major product is that with the larger ring.
Introduction
determine the relative rates of the two possible cyclisation pro-
cesses, viz. the barriers to the transition state and the Arrhenius
We have reported the competitive gas phase S i reactions
of systems of the type shown in Scheme 1 where n = 1–3.
N
(
frequency) factor A. Both are considered separately below.
1–4
a. The results of ab initio calculations for competing cyclis-
ations of the 5,6-epoxyhexoxide anion (1). The results of an ab
initio computational study using Gaussian 94 for the system
8
–
O
O
(
) n
where n = 4 are summarised in Fig. 1. The geometries of the
local minima and the transition states were optimised at the
RHF/6-31ϩG(d) level of theory, with energies determined at
the MP2-Fc/6-31ϩG(d) level of theory. Energies of 1–3, A and
B together with the geometries of the two transition states are
listed in Table 1. Full geometric data for 1–3 are listed in the
Supplementary Data.
O
O –
(
) n
–
O
O ( ) n
The data presented in Fig. 1 and Table 1 show that the
barriers to the two transition states A and B are 35.0 and
Scheme 1
Ϫ1
3
9.7 kJ mol at the level of theory indicated. If these modest
5
–7
Baldwin’s rules (which refer to stabilised carbanion cyclis-
ations) pertain when n = 1 and 3, i.e. the predominant cyclis-
ation is that which forms the smaller of the two possible ring
systems. The kinetics of these reactions are influenced by the
barrier to the transition state and/or the Arrhenius factors A.
In this paper we (i) report the behaviour in both the con-
densed phase and gas phase of the next homologue, i.e. where
n = 4, and (ii) consider all systems so far reported in terms of
the parameters influencing the competing cyclisations.
barriers are controlling the relative rates of the two processes,
both 2 and 3 should be formed following collision activation of
1 in the gas phase, with 2 being marginally favoured. Factors
which may influence the height of the barriers are as follows. (i)
The OCO angles for A are 154.42Њ (dihedral angle 172.93Њ) and
for B, 156.38Њ (dihedral angle 172.25Њ). These values are very
similar: the higher value of 156.38Њ may, in principle, favour
reaction through B (cf. ref. 7), however any such effect will be
marginal. (ii) The larger electrophilicity of the substituted car-
bon of the ethylene oxide ring (according to Mulliken charge
8
1
2
calculations, C = 0.168 and C = 0.419) favours nucleophilic
attack at the more substituted carbon; i.e. formation of 2
through transition state A. (iii) The changes in ring strain from
Results and discussion
Part A. The 5,6-epoxyhexoxide system
1
to transition states A and B should be marginal in determin-
1
. A theoretical approach. There are two major factors which
ing relative kinetics: in contrast, the release in ring strain from 1
to products, makes 2 more thermodynamically stable than 3 [2
is some 11 kJ mol more negative in energy than product 3
†
5
Supplementary material is available for this paper (SUPPL. NO.
7479, 9 pp.). For details of the Supplementary Publications Scheme
see ‘Instructions for Authors’, J. Chem. Soc., Perkin Trans. 2, available
via the RSC web page (http://www.rsc.org/authors). The supplementary
data are also available on the RSC’s web server, http://www.rsc.org/
suppdata/p2/1999/457/.
Ϫ1
(
strain energy of ethylene oxide, tetrahydropyran and oxepan
Ϫ1 9
are 112.5, 6.3 and 26.7 kJ mol respectively )].
b. The calculation of Arrhenius factors. Frequency factors for
the two competitive S
i processes shown in Fig. 1 may be influ-
457
N
J. Chem. Soc., Perkin Trans. 2, 1999, 457–464

View Article Online
Table 1 Energies of species shown in Fig. 1, and geometries of transi-
δ _
O8
a
δ _
tion states A and B
8
2
3
1
O
3
4
1
4
7
2
O
B
A
_
δ
Energies/hartrees
δ _
O
5
6
Ϫ1
5
Species
(kJ mol
)
Geometries/Å or Њ
6
1
Ϫ384.4306105
nominally 0)
Ϫ384.4678042
Ϫ88.0)
Ϫ384.4636616
Ϫ76.7)
Ϫ384.4172435
ϩ35.0)
b
( +39.7 )
(
(
+35.0 )
2
b
b
(
3
(
1
2
3
4
5
6
7
A
C C
C C
C C
C C
C C
C O
O C
C O
C C C
C C C
C C C
C C C
C C O
C O C
C C O
C O C
O C C
O C C O
C C
C C
C C
C C
C C
C O
O C
C O
O C
C C C
C C C
C C C
C C C
C C O
1.4450
1.5068
1.5333
1.5282
1.5356
1.3767
2.0844
1.7837
121.58
113.68
114.10
111.37
111.62
107.07
154.42
52.09
2
(
3
(O)
4
5
O
6
7
2
8
2
_
1
2
3
4
5
6
7
O
2
3
1
3
4
4
5
5
6
6
7
2
8
1
2
1
( - 76.7 )
7
2
2
8
O –
8
1
( - 88.0 )
51.02
7
2
8
3
172.93
1.4432
1.5199
1.5326
1.5380
1.5445
1.3741
2.0751
1.7467
1.4283
121.05
116.54
115.34
117.22
115.63
104.93
156.38
52.93
O
1
2
3
4
5
6
7
O –
B
Ϫ384.2565604
ϩ39.7)
2
(
3
2
O
4
5
6
Fig. 1 Ab initio calculations at MP2-Fc/6-31G(d) level (Gaussian 94)
7
1
8
2
Ϫ1
for the reactions 1→2 and 1→3. Energies in kJ mol . See Table 1 for
1
the geometries of transition states A and B. Structures 1, 2 and 3 have a
number of stable conformers. Those shown are the most stable (see
Supplementary Data for geometries and energies of the most stable
conformers).
8
1
2
3
4
5
6
7
2
3
3
4
4
5
5
6
Since we are dealing with anions, it seems reasonable to
assume that QElec = 1. We are dealing with unimolecular
rearrangements, so QTrans of the reactant and transition state
are identical, and we assume (to a first approximation) that the
same is true for QRot. Thus approximation of the A factor for
each process simplifies to evaluating QVib for the reactant and
for the competing transition states. However it must be noted
that the anions in a mass spectrometer, especially following
collisional activation, will not follow a Boltzman (thermalised)
distribution of internal energies. Thus the following calcu-
lations must only be considered in a qualitative sense.
We have calculated, at the HF/6-31ϩG(d) level of theory,
the harmonic vibrational frequencies of each structure using
Gaussian 94 [cf. ref. 11]. These values (scaled by a factor of
.9131) are listed as Supplementary Data to the paper. A dif-
ficulty which often occurs with such an evaluation arises due to
6
7
1
8
C O C
O C O
C O C
O C C
O C C
O C C O
7
1
1
8
2
1
3
2
8
2
52.15
8
2
118.18
172.25
7
1
8
a
Calculations at MP2(fc)/6-31ϩG(d) level. All energies include zero
b
point energy scaling (factor 0.9676). See Supplementary Data.
enced by (i) the initial nucleophilic attack to form the transition
state (i.e. the ability of the nucleophile to access the appropriate
channel, and the depth of that channel), and/or (ii) the nature
of the transition state [i.e. whichever is ‘looser’ (more dis-
ordered) will give the higher rate]. To determine the relative
abilities of the nucleophile to access the two channels requires
an intimate knowledge of the potential surface maps for those
processes. We do not have these data. However, the entropic
natures of the two transition states may be determined by calcu-
lating the relative A factors.
0
12
the problem of hindered rotors.
The calculated vibrational
partition functions may underestimate the actual hindered
rotor partition functions. Since we are concerned with the ratio
of the A factors of two competing cyclisation channels which
To assess the relative Arrhenius A factors for the two compet-
have the same reactant, we have not addressed the hindered
10
‡
ing cyclisation channels, we have used transition state theory,
viz. eqn. (1), where kB is Boltzman’s constant, h is Planck’s
rotor problem. The value of Q
for the transition state lead-
Vib
ing to the the six-membered ring product (2) is 63.1, whereas
that for the seven-membered product (3) is 26.3, so the transi-
tion state leading to the formation of 2 is ‘looser’ than that
giving 3 (see Supplementary Data for details). The A factor for
the formation of 2 is thus larger than that for the formation of
, within the approximations we have employed.
In summary, (i) ab initio calculations indicate that the
barriers to the formation of both 2 and 3 from 1 are modest,
that both 2 and 3 should be formed under conditions of
collision activation, with 2 being the major product. (ii) The
relative Arrhenius factors indicate that A is the looser transition
‡
k(T) = (k T/h) (Q /Q )exp(E /k T)
(1)
B
R
o
B
constant, E is the energy difference between reactant and tran-
sition state at 0 K, and Q and Q are the molecular partition
functions of transition state and reactants respectively. The
partition function can be factorised into partition functions
for translation, rotation, vibration and electronic state, viz.
o
‡
R
3
10
eqn. (2).
Q = QTrans QRot QVib QElec
(2)
4
58
J. Chem. Soc., Perkin Trans. 2, 1999, 457–464

View Article Online
Table 2 Reactions of 4,5-epoxyhexanol (C), tetrahydropyran-2-meth-
anol (D) and 3-oxepanol (E) with 10% aqueous sodium hydroxide
Product ratioa
Reactant
T/ЊC
t/h
0.16
3
60
[(C:D:E)%]
C
20
70:24:6
0:72:28
0:70:30
0:61:39
0:60:40
0:63:37
0:100:0
0:0:100
2
2
00
00
00
0
0
1
1
1
0.16
3
60
60
60
D
E
100
100
a
Ratios determined by measuring the areas under each GC peak. Each
measurement was carried out twice—error ±2%.
Fig. 3 MS/MS Data for 2. Peak widths at half height, 97 (27.7), 85
33.1) and 67 (30.7).
(
Fig. 4 MS/MS Data for 3. Width at half height of m/z 97 = 33.4 volts.
Fig. 2 MS/MS Data for 1. ZAB 2HF spectrometer. For experimental
procedures see Experimental section. Peak widths at half height, m/z
(
volts ± 0.2): 97 (27.4), 85 (33.1) and 67 (30.7).
peaks in these spectra are recorded in the legend to each figure.
Data pertaining to product ion studies are listed in Table 3, and
1
8
state by a significant factor. Taken together, these calculations
suggest that 2 should be the major cyclisation product in gas
phase conditions.
the mass spectra of two O labelled derivatives are recorded in
Table 4.
The spectra of 1 and 2 (Figs. 2 and 3) are the same within
experimental error, including identical peak widths at half
height of the major peaks m/z 97, 85 and 67. These spectra are
quite different from that (Fig. 4) of 3. This indicates that 1 and
2 are fragmenting through a common intermediate (perhaps
following cyclisation of 1 to 2), but that cyclisation of 1 to 3 is
not occurring under the reaction conditions. We now need to
confirm that (i) the fragment ions from 1 and 2 have the same
structures, (ii) there are differences between these structures
and those of the fragment anions of 3, and (iii) fragment-
ations observed in the spectra of 1 occur following conversion
of 1 to 2.
2
. The experimental approach. a. The competitive cyclisation
reactions in the condensed phase. The reactions of 5,6-epoxy-
hexanol with 10% aqueous sodium hydroxide are summarised
in Table 2. Tetrahydropyran-2-methanol is the major product
formed at 20 ЊC, with the product ratio of tetrahydropyran-2-
methanol to 3-oxepanol being 65:35 after one hour. The analo-
gous ratio obtained at 100 ЊC is 60:40. There is no significant
change in these product ratios (at constant temperature) with
increasing time. Tetrahydropyran-2-methanol and 3-oxepanol
are not interconvertible under these conditions, and neither
reverts to 5,6-epoxyhexanol. Thus the two competing SNi
reactions of deprotonated 5,6-epoxyhexanol are kinetically
controlled under these reaction conditions.
All data concerning product ion studies are collected in Table
3. Whenever possible, both collisional activation and charge
13
reversal (positive ion) mass spectra of source formed product
anions are compared with those of authentic anions formed by
independent syntheses. In some cases, the collisional activation
spectra of the fragment ions are either not characteristic (e.g.
m/z 67), or are too weak to make meaningful comparison (e.g.
m/z 59). In these cases, identification is based on the charge
reversal spectra. The results of these comparisons are consist-
ent with the data summarised in Scheme 2. The product anions
in the spectra of 1 and 2 are the same, but different from those
in the spectrum of 3.
b. Gas phase cyclisation of the 5,6-epoxyhexoxide anion.
Alkoxide ions 1, 2 and 3 (shown in Fig. 1) were formed in the
gas phase by nucleophilic displacement of the appropriate
acetate in the ion source of the VG ZAB 2HF mass spec-
trometer. Forming the alkoxide ions indirectly is necessary,
Ϫ
since gas phase deprotonation of the alcohols by HO will
occur both at OH and elsewhere on these molecules. The
collision induced mass spectra of the three alkoxides 1–3 are
recorded in Figs. 2–4. Peak widths at half height for the major
J. Chem. Soc., Perkin Trans. 2, 1999, 457–464
459

View Article Online
a
Table 3 Product ion studies using collisional activation (CA) and charge reversal (CR) spectra
Spectrum
[
CA: m/z(loss or formation)abundance]
Parent (m/z)
Product (m/z)
Mode
[CR: m/z(abundance)]
ؒ
ؒ
1
2
3
(115)
(115)
(115)
ϪH O (97)
CA
CR
96(H )100, 95(H )26, 82(CH )11, 79(H O)59, 67(CH O)68.
2
2
3
2
2
97(70), 96(100), 95(78), 81(56), 79(54), 77(34), 69(22), 67(78), 66(44), 65(50), 63(16), 57(4),
5(14), 53(33), 52(34), 51(38), 50(33), 43(8), 41(45), 39(53), 31(6), 29(14), 27(21), 26(16).
5
ؒ
ؒ
ϪH O (97)
CA
CR
96(H )100, 95(H )22, 82(CH )13, 79(H O)66, 67(CH O)70.
2
2 3 2 2
97(67), 96(100), 95(75), 81(49), 79(52), 77(38), 69(24), 67(84), 66(47), 65(54), 63(19), 57(4),
5(12), 53(29), 52(30), 51(29), 50(30), 43(6), 41(35), 39(46), 31(3), 29(12), 27(20), 26(15).
5
Ϫ
ϪH O (97)
CA
CR
95(H )24, 79(H O)100, 67(CH O)1, 41(C H )2.
2
2 2 2 3 5
97(21), 96(76), 95(100), 94(8), 81(52), 79(77), 77(45), 69(18), 67(45), 66(32), 65(28), 63(12),
7(4), 55(8), 53(12), 51(14), 41(21), 39(34), 31(4), 29(8), 27(15), 26(4).
5
ؒ
ؒ
CA
CR
96(H )100, 95(H )28, 82(CH )5, 79(H O)65, 67(CH O)52.
2 3 2 2
O–
O–
97(68), 96(100), 95(72), 81(52), 79(58), 77(39), 69(21), 67(83), 66(50), 65(55), 63(21), 57(4),
5(12), 53(31), 52(39), 51(32), 50(30), 43(6), 41(37), 39(46), 31(6), 29(12), 27(19), 26(12).
5
(
97)b
Ϫ
CA
CR
95(H )24, 79(H O)100, 67(CH O)1, 41(C H )1.
2
2
2
3
5
97(19), 96(70), 95(100), 94(8), 81(53), 79(66), 77(46), 69(18), 67(39), 66(29), 65(25), 63(11),
57(4), 55(8), 53(13), 51(15), 41(22), 39(37), 31(4), 29(6), 27(11), 26(4).
97)b
(
c
1
2
(115)
ϪCH O (85)
CA
CR
83(H )<20, 67(H O)60, 55(CH O)100.
2
2
2
2
85(4), 84(22), 83(12), 82(5), 81(2), 68(9), 67(60), 66(35), 65(36), 56(28), 55(85), 53(32),
1(26), 50(21), 43(6), 41(42), 39(100), 29(62), 28(16), 27(45), 26(20).
5
c
(115)
ϪCH O (85)
CA
CR
83(H )<20, 67(H O)68, 55(CH O)100.
2
2 2 2
85(6), 84(21), 83(13), 82(8), 81(4), 68(14), 67(62), 66(34), 65(38), 56(32), 55(87), 53(33),
1(28), 50(18), 43(8), 41(45), 39(100), 29(65), 28(18), 27(51), 26(20).
83(H )6, 67(H O)38, 55(CH O)100.
85(3), 84(9), 83(12), 82(6), 81(4), 70(5), 69(5), 67(7), 66(6), 65(5), 56(36), 55(100), 54(48),
53(46), 51(22), 50(22), 43(3), 41(24), 39(77), 29(52), 28(28), 27(42), 26(16).
5
OH
OH
CA
CR
2
2
2
–
85)d
(
CA
CR
67(H O)100, 55(CH O)1.
85(3), 84(83), 83(97), 82(5), 81(2), 70(1), 69(2), 68(15), 67(17), 66(10), 65(10), 57(2), 56(8),
55(29), 53(12), 51(9), 50(10), 43(6), 41(43), 39(100), 29(76), 28(26), 27(64), 26(18).
2
2
–
–
H
(
85)e
1
2
(115)
(115)
Ϫ(H O ϩ CH O) (67)
CR
CR
CR
67(100), 66(48), 65(34), 64(3), 63(6), 62(3), 61(2), 53(1), 52(4), 51(6), 50(5), 49(2), 41(7),
2
2
4
0(3), 39(10), 38(2), 37(1), 27(2), 26(1).
67(100), 66(50), 65(35), 64(2), 63(7), 62(3), 61(2), 53(1), 52(4), 51(6), 50(5), 49(2), 41(7),
0(2), 39(8), 38(2), 37(1), 27(2), 26(1).
Ϫ(H O ϩ CH O) (67)
2
2
4
67(100), 66(54), 65(36), 64(2), 63(7), 62(3), 61(2), 53(1), 52(4), 51(4), 50(3), 49(1), 41(6),
40(2), 39(7), 38(2), 37(1), 27(2), 26(1).
–
(
67)f
CR
67(100), 66(51), 65(39), 64(2), 63(7), 62(2), 61(1), 53(1), 52(5), 51(8), 50(7), 49(2), 41(22),
4
0(10), 39(39), 38(7), 37(4), 27(8), 26(3).
–
67)
g
(
3
(115)
ϪC H (59)
CR
CR
59(8), 58(29), 57(6), 56(8), 45(2), 44(4), 42(13), 41(15), 31(34), 30(48), 29(100), 28(32).
59(10), 58(28), 57(4), 56(12), 45(2), 44(6), 42(18), 41(19), 31(29), 30(43), 29(100), 28(34).
4
8
Ϫ
h
OCH CHO (59)
2
a
The abundances in both collisional activation and charge reversal spectra are dependent on both source conditions and the collision gas pressure.
b
As a general guide when comparing spectra, the abundance of an individual peak should be correct to ±10%. Formed by deprotonation of the
Ϫ
c
appropriate alcohol with HO . Weak spectrum: difficult to measure the abundance of the peak formed by loss of H because of baseline noise.
2
d
Ϫ
e
Ϫ
f
Formed by decarboxylation of HOCH CH CH᎐CHCH CO
. Formed by deprotonation of MeCH CH᎐CHCH OH with HO . Formed by
2 2
᎐
Ϫ h
2
2
᎐
2
2
Ϫ
g
deprotonation of penta-1,3-diene with HO . Formed by deprotonation of cyclopentene with HO . Reported by spectrum, see ref. 3.
The data considered to date show that 1 and 2 have iden-
tical fragmentations but do not indicate whether 1 cyclises to
We have not carried out labelling studies on the 3-oxepanol
anion 3: our primary interest in 3 was to determine whether its
fragmentations are different from those of 1 and 2. Even so,
the formation of m/z 59 (Scheme 2) is straightforward, while
the structure of m/z 97 is consistent with the fragmentation
pathway shown in Scheme 4.
2
before fragmentation occurs. This problem is resolved from
18
a consideration of the spectra (Table 4) of O labelled deriv-
atives (of 1 and 2). These spectra demonstrate the equilibration
of the two oxygens of 1 (and 2) prior to fragmentation: in the
case of 1, such equilibration requires rearrangement of 1 to 2
prior to fragmentation. The fragmentations of 2 have been
considered previously: it has been proposed that the oxygen
equilibration occurs by the proton transfer/ring opening equi-
Conclusions to Part A. 1. Ab initio and Arrhenius factor cal-
culations indicate that 1 should cyclise to both 2 and 3. The
barriers to transition states A and B are modest (computed as
14
Ϫ1
libria summarised in Scheme 3. Fragmentation of the ring
opened structures, particularly 4/5 and 6/7, account for both
the oxygen equilibration and the structures of the product
35.0 and 39.7 kJ mol respectively), and the Arrhenius factor
is larger for the formation of 2 than for 3. Thus 2 is predicted
to be the favoured product.
ions (Scheme 2) formed by the respective losses of H O and
2. Experimentally, 1 cyclises to give only 2 in the gas phase:
product 3 is not detected. It is proposed that the larger
2
CH O.
2
4
60
J. Chem. Soc., Perkin Trans. 2, 1999, 457–464

View Article Online
1
8
Table 4 CA MS/MS Data for O labelled 1 and 2
Spectrum
Parent
[m/z(loss or formation)relative abundance]
a
ؒ
18
18
18
Ϫ
1
8
116/115 (H , H )20, 99(H O)100, 97(H O)85, 87(CH O)8, 85(CH O)10, 79(H O ϩ H O)5, 67(C H )47.
O
O –
2
2
2
2
2
2
2
5
7
m/z 117b
1
O –
8
ؒ
18
18
18
Ϫ
1
16(H )24, 115(H )5, 99(H O)64, 97(H O)70, 87(CH O)17, 85(CH O)14, 79(H O ϩ H O)8, 67(C H7 )100.
2 2 2 2 2 2 2 5
O
m/z 117b,c
a
b
18
18
Unresolved peaks. The O label is in the side chain of 1: cyclisation of 1 will give 2 with the O label in the ring. Thus the relative ratios of m/z
c
9
9:97 and 87:85 will be reversed in the two spectra, as observed. The relative abundances of the major fragment peaks in the spectra of 1 and 2 are
critically dependent on collision gas pressure. This explains the major abundance differences in the two spectra above (cf. e.g. m/z 67) (the two spectra
were measured six years apart, see ref. 14): this does not affect the 99:97 and 87:85 ratios.
_
O –
–
H2O
O –
OH
OH
m/z 97
m/z 85
OH
OH
O
O
O
O
_
O –
– CH O
2
_
1
2
+
major
minor
_
O –
– H] –
_
O –
[
OH
OH
O
OH
OH
OH
–
(H2O + CH2O)
_
5
4
or
_
m/z 67
m/z 97
OH
O –
O
_
OH
–
H2O
6
7
O –
O –
Scheme 3‡
O
– C4H8
O –
OH
_
OH
3
–
O
CHO
m/z 59
O
_
O
(
O
Scheme 2
Arrhenius factor for reaction 1 to 2 is a major contributing
factor to the exclusivity of this reaction (i.e. entropically, transi-
tion state A is ‘looser’ than transition state B).
[
O
) HO –
]
O –
+ H2O
Scheme 4‡
3
. Treatment of 1 in 10% aqueous sodium hydroxide at reflux
yields both 2 and 3 in kinetically controlled reactions, with 2
being the predominant product.
1
2
C and C (the two carbons on the ethylene oxide ring) of the
reactant, and (iv) the Arrhenius factor (QVib) for each reaction
(
determined as outlined in Part A). The bond lengths of the
Part B. Comparison of the competing reactions of Scheme 1
where n ؍ 1–4.
breaking and forming bonds in the various transition states are
not listed because they are comparable for all systems (except
for symmetrical 1.1 where the O–C bonds are both 1.89 Å ),
1
We now summarise the results for the four homologous systems
for which competing S i processes have been studied in the
gas phase. A number of parameters which influence the
i.e. the forming bonds are longer (in the range 2.08–2.10 Å)
than the breaking bonds (within the range 1.74–1.89 Å). The
N
1–4
2–4
formation of the eight transition states of the four systems
major (and minor) products obtained both in the gas and con-
densed phases from each pair of transition states are also listed
in Table 5.
Three of the four systems show significant correspondence
between the experimental gas phase (and condensed phase)
data and the theoretical data (n = 1, 3 and 4), while one (n = 2)
shows some deviation. When n = 1, the Payne rearrangement is
the major process in the gas phase: a feature in accord with ab
(
n = 1 to 4; cf. Scheme 1) are listed in Table 5. These kinetic
parameters are calculated for gas phase reactions, i.e. for
reactions occurring in the absence of solvent. The transition
states are numbered as follows: 1.1 (n = 1)—the Payne
rearrangement, S i through three centre state, 1.2 (n = 1)—S i
N
N
through the competing four membered transition state; 2.1
n = 2)—S i through four centre state, 2.2 (n = 2)—S i through
(
N
N
the competing five centred state, etc. The parameters listed are
i) the O–C–O angle for the S i process at the transition state,
(
N
‡
The second structure may be transient or alternatively the formation
together with the corresponding dihedral angle, (ii) the com-
of the third structure from the first may be concerted. The second struc-
ture is included in the Scheme in order that the mechanism for ring
opening is clear.
Ϫ1
puted barrier from reactant to the transition state (in kJ mol ),
(
iii) the Mulliken charge ratio (the relative electrophilicities) of
J. Chem. Soc., Perkin Trans. 2, 1999, 457–464
461

View Article Online
Table 5 Parameters influencing the kinetics of competitive S i processes. Major products
N
Product(s)
Gas phase
O–C–O
(dihedral)/Њ
Mulliken
C :C
Arrhenius
Qvib
Ϫ1 a
1
2 b
c
d
Transition state
Barrier/kJ mol
Condensed phase
δ _
O
e
f
1
.1
4
5
160.1 (159.7)
116.5 (153.0)
163.4 (163.1)
0.189:0.349
3.4
2.5
9.8
Major
Minor
1
2
O
δ _
δ _
O
e
1.2
122
70
O
δ _
δ _
O
g
2
.1
0.158:0.336
Major
Minor
δ _
O
δ _
O
g
2
.2
69
48
140.4 (163.1)
156.1 (176.1)
4.7
Major
Major
Major
O
δ _
δ _
O
3
.1
0.181:0.329
20.5
Sole product
δ _
O
δ _
O
3.2
48
152.8 (175.3)
10.2
—
Minor
O
δ _
δ _
O
_O
4.1
3
5
154.4 (172.9)
156.4 (172.3)
0.168:0.419
63.1
26.3
Sole product
Major
Minor
δ
δ _
O
4.2
40
—
O
δ _
a
b
1
2
c
At MP2-Fc/6-31ϩG(d) level. The Mulliken charge distribution is that of C and C in the reactant, see ref. 8. Calculations as described earlier in
d
e
f
this paper. Carried out in 10% aqueous sodium hydroxide at reflux (for full details see refs. 2–4). Calculations at G2 level in these two cases. Not
performed. However alkyl substituted 2,3-epoxypropoxide anions exclusively undergo the Payne rearrangement in solution, cf. ref. 1. Gas phase
g
experiments only indicate that both are formed in approximately comparable yield, see ref. 3.
initio data. The Payne rearrangement (which proceeds through
transition state 1.1) is favoured by the lower barrier of 45 kJ
mol (cf. 122 kJ mol for 1.2). The major contributor to the
larger barrier to transition state 1.2 is the unfavourable geom-
etry of the transition state, i.e. angle O–C–O is computed to be
3.1 and 4.1 respectively. The Arrhenius factor must control the
gas phase reaction when n = 3. The same scenario probably also
pertains when n = 4. Although the barriers are different (35 and
Ϫ1
Ϫ1
Ϫ1
40 kJ mol ), they are very modest, and if they are a major
influence in determining the product ratio, then products
1
16.5Њ (cf. 160.1Њ for 1.1: the closer the O–C–O angle is to the
should be observed from both processes. We propose that fre-
ideal S 2 angle of 180Њ, the more favoured will be the reaction if
quency factors are controlling the gas phase S i reactions where
N
N
other parameters are equal). Although the larger Arrhenius fac-
tor also favours the formation of transition state 1.1, the large
barrier difference is a major factor influencing the competing
SNi reactions when n = 1.
n = 3 and 4. The condensed phase reactions (in 10% aqueous
sodium hydroxide at reflux) show the expected trend. The
smaller of the ring systems is formed predominantly in each
case.
The gas phase results for systems where n = 3 and 4 are
perhaps surprising but they are consistent and qualitatively
similar to the predictions of the ab initio studies. The computed
barriers for each of the competing reactions when n = 3 and 4
The situation with regard to the system when n = 2 is not
as clear as those of the three systems described above. The
3
potential profile summarising the ab initio calculations is
recorded in Fig. 5 to assist the discussion. This system shows
similarities to that where n = 3, i.e. the barriers to the two
processes are comparable (see Fig. 5 and cf. Table 5) as are
the Mulliken and the Arrhenius QVib ratios (see Table 5).
Ϫ1
are comparable in each system (48 kJ mol for 3.1 and 3.2; 35
Ϫ1
and 40 kJ mol for 4.1 and 4.2), yet in both cases there is only
one product formed in the gas phase, viz. those formed through
4
62
J. Chem. Soc., Perkin Trans. 2, 1999, 457–464

View Article Online
5–7
Baldwin’s rules. When n = 2, both products are formed in
comparable yield.
(
ii) The corresponding condensed phase reactions preferen-
tially give the smaller ring system in higher yield when n = 1, 3
and 4. When n = 2, the larger ring system is formed in the higher
yield.
Experimental
Mass spectrometric methods
Collisional activation (CA) mass spectra (MS/MS) were deter-
15
mined with a VG ZAB 2HF mass spectrometer. Full operat-
16
ing details have been reported. Specific details were as follows:
the chemical ionisation slit was used in the chemical ionisation
source, the ionising energy was 70 eV, the ion source temper-
ature was 100 ЊC, and the accelerating voltage was 7 kV. The
liquid samples were introduced through the septum inlet with
Ϫ6
no heating [measured pressure of sample 1 × 10 Torr (1
Torr = 133.322 Pa)]. Alkoxide anions 1–3 were formed by S 2
N
Ϫ
displacement from the appropriate acetate using HO (from
Ϫ5
H O: measured pressure 1 × 10 Torr). The estimated source
2
Ϫ1
pressure was 10 Torr. Argon was used in the second collision
cell (measured pressure, outside the cell, 2 × 10 Torr), giving a
Ϫ7
1
0% reduction in the main beam, equivalent to single collision
Fig. 5 Ab initio calculations [geometries RHF/6-31ϩG(d), energies
conditions. CID MS/MS measurements involved using the
magnet to choose the ion under study, collision activating it (see
above), and scanning the electric sector to analyse the resultant
product anions. Charge reversal (CR) (positive ion) MS/MS
data for negative ions were obtained as for CA MS/MS data,
except that the electric sector potential was reversed to allow the
transmission of positively charged product ions (for full details
see ref. 13). The recorded peak widths at half height are an
average of ten individual measurements and are correct to ±0.2
volts.
MP2-Fc/6-31ϩG(d)] for competitive S i cyclisations of the 3,4-epoxy-
N
butoxide anion (from ref. 4).
There is however one feature which distinguishes this system
from the other three, viz., the two product alkoxides have
very different thermodynamic stabilities. The five membered
Ϫ1
ring product is 82 kJ mol more negative in energy than the
four membered counterpart (see Fig. 5) [cf. Ϫ27 (n = 1, i.e.
1
oxetane product more stable than ethylene oxide product),
4
Ϫ1
Ϫ26 (n = 3) and ϩ 11 kJ mol (n = 4, see Fig. 1 this paper)].
This large difference of 82 kJ mol is due principally to the
Ab initio calculations
Ϫ1
8
difference in ring strain between the oxetane and tetrahydro-
The Gaussian 94 suite of programs was used for all calcu-
Ϫ1
9
furan systems (107.5 and 25.1 kJ mol respectively ).
lations, which were carried out on Silicon Graphics Power Chal-
lenge. The geometries of the local minima and the transition
states were optimised at the RHF/6-31ϩG(d) and MP2 (Fc)/6-
31ϩG(d,p) levels of theory. Harmonic frequency analyses were
performed on each stationary point in order to characterise
them as either a local minimum or transition state. A local
minimum is characterised by possessing all real vibrational
frequencies and its hessian matrix possessing all positive eigen-
values. A transition state is characterised by possessing one
(and only one) imaginary frequency and its hessian matrix
possessing one (and only one) negative eigenvalue. Intrinsic
reaction coordinate (IRC) calculations were performed (begin-
ning from both transition structures) to verify that each transi-
tion structure connected particular local minima. Final energies
are quoted at the MP2 Fc/6-31ϩG(d) level of theory, and
include a scaled (0.8929) zero point vibrational energy correc-
tion [which is based on the RHF/6-31ϩG(d) optimised
geometry].
The system depicted in Fig. 5 is the most complex we have
studied from an experimental viewpoint. Consideration of the
spectra of the three alkoxide anions shown in Fig. 5 together
2
18
with those of H and O labelled derivatives, show that in the
gas phase, (i) both cyclisation products are formed in compar-
able yield, (ii) the three and four membered systems are inter-
convertible under conditions of collisional activation, and (iii)
the tetrahydrofuranyl methoxide product does not interconvert
to the reactant. It is possible that thermodynamic control
explains the formation of at least some of the tetrahydrofuranyl
methoxide product: the exothermicity of this process and the
consequent high reverse energy barrier are in keeping with such
a proposal.
This system has also been studied in two different solvent
systems, i.e. 10% aqueous sodium hydroxide and sodium
hydride/tetrahydrofuran, both at reflux. The products were
identified using gas chromatography/mass spectrometry with a
3
Finnigan GCQ ion trap mass spectrometer. The experimental
results (which are reproducible) show that tetrahydrofuran-
methanol is the major product in both solvent systems. We
anticipated that this result should be explicable in terms of
the reactions being thermodynamically controlled in solution.
However this appears not to be the case: tetrahydrofuran-
methanol is the kinetic product as evidenced by the product
ratios being the same (within experimental error) at reaction
Materials
Tetrahydropyran-2-methanol, penta-1,3-diene and pent-4-en-
1
-ol were commercial samples. Tetrahydropyran-2-methanol
17 18 19
acetate, 5,6-epoxyhexanol, 5,6-epoxyhexyl acetate, 3-oxe-
panol, hexa-3,5-diene-1-ol, and hexa-2,5-dien-1-ol were
made by reported procedures.
19
20
21
3
times within the range 10 min to 60 h.
3
-Acetoxyoxepan. 3-Oxepanol (0.1 g) was acetylated by a
17
standard procedure to yield 3-acetoxyoxepane (bp 121–
Conclusions to Part B. For the scenario outlined in Scheme 1:
123 ЊC/760 mmHg; 0.11 g, 81% yield). Found, C, 60.6; H, 9.2%:
ϩ
(
i) The gas phase S i reactions for the systems (n = 1, 3 and 4)
C H O requires C, 60.7; H, 8.9%. MH = 159 (18%), base
N
8
14
3
1
give the smaller ring system either exclusively or in the larger
yield. This is in accord with the theoretical studies and with
peak m/z 43. H NMR (300 MHz, CDCl ) δ 1.35–1.52 (4H, m),
3
1.70–1.82 (2H, q), 2.01 (3H, s), 3.4 (2H, d), 3.69 (2H, t), 4.1
J. Chem. Soc., Perkin Trans. 2, 1999, 457–464
463

View Article Online
1
3
(
(
1H, m). C NMR (CDCl ) δ 21.1 (CH ), 21.39 (CH ), 27.5
J. M. H., M. S. T and S. D. thank the ARC for a PhD scholar-
ship, a research associate position and a research assistant
position respectively.
3
3
2
CH ), 31.82 (CH ), 64.1 (CH), 70.4 (CH ), 170.5 (C᎐O).
2
2
2
18
[
1- O]5,6-Epoxyhexan-1-ol. Hex-5-en-1-ol (2 g) was oxid-
22
ised to hex-5-enoic acid (1.9 g, 83%), which was stirred at
References
3
2
0 ЊC with oxalyl chloride (1.5 cm ) and N,N-dimethyl-
3
formamide (1 drop) in anhydrous diethyl ether (30 cm ) for 3 h.
Removal of the solvent in vacuo followed by distillation gave
hex-5-enoyl chloride (1.4 g, 73%), which was allowed to stir
with H2 O (0.26 g, 96% O) and tetrahydrofuran (10 cm ). The
solution was dried (MgSO ) and concentrated under reduced
1 S. Dua, M. S. Taylor, M. A. Buntine and J. H. Bowie, J. Chem. Soc.,
Perkin Trans. 2, 1997, 1991.
2
S. Dua, M. S. Taylor, M. A. Buntine and J. H. Bowie, Int. J. Mass
Spectrom. Ion Processes, 1997, 165/166, 139.
18
18
3
3 J. M. Hevko, S. Dua, J. H. Bowie and M. S. Taylor, J. Chem. Soc.,
Perkin Trans. 2, 1998, 1629.
4
pressure to afford a clear oil which was added dropwise to a
4 J. M. Hevko, S. Dua, M. S. Taylor and J. H. Bowie, Int. J. Mass
Spectrom., 1998, in the press.
5 L. Tenud, S. Farook, J. Seibl and A. Eshenmoser, Helv. Chim. Acta,
1970, 53, 2059.
suspension of lithium aluminium hydride (0.5 g) in diethyl ether
3
(
20 cm ) and heated at reflux for 2 h. The mixture was cooled
3
to 0 ЊC, aqueous hydrogen chloride (30%, 2 cm ) was added,
6
7
G. Stork and J. F. Cohen, J. Am. Chem. Soc., 1974, 96, 5270.
the organic layer separated, the aqueous layer extracted with
J. E. Baldwin, J. Chem. Soc., Chem. Commun., 1976, 734; J. E.
3
diethyl ether (2 × 5 cm ), the combined organic extracts dried
Baldwin, Further Perspectives in Organic Chemistry,
A Ciba
(
MgSO ), concentrated in vacuo, and the residue distilled to give
Foundation Symposium, Elsevier, Amsterdam, 1978, 85.
4
18
18
[
1- O]hex-5-en-1-ol (0.75 g), which was epoxidised and dis-
8 GAUSSIAN 94, Revision C3, M. J. Frisch, G. W. Trucks, H. B.
Schlegal, P. M. W. Gill, B. G. Johnson, M. A. Robb, J. R.
Cheeseman, T. Keith, G. A. Petersson, J. A. Montgomery, 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. Gonzales and J. A. Pople, Gaussian Inc., Pittsburgh, PA, 1995.
18
tilled (60–70 ЊC/20 mmHg) to afford the O labelled alcohol
18
(
0.5 g, 58% yield; O = 48%).
The corresponding acetate was made by a standard
19
technique.
Condensed phase reactions
9
S. W. Benson, Thermochemical Kinetics, Wiley, New York, 1967.
Products of the base catalysed reactions of 5,6-epoxyhexanol,
tetrahydropyran-2-methanol and 3-oxepanol were analysed
using a Finnigan GCQ mass spectrometer. Conditions: Column
phase RTX-SMS (length 30 cm, ID 0.25, GC fused silica
capillary), He carrier gas. Initial column temperature, held
at 50 ЊC for 2 min, then the temperature increases at 15 ЊC
per min. Retention times: 5,6-epoxyhexanol (7.00 min),
tetrahydropyran-2-methanol (6.02 min) and 3-oxepanol (6.32
min).
10 R. G. Gilbert and S. C. Smith, Theory of Unimolecular and
Recombination Reactions, Blackwell Scientific, Cambridge, 1990.
1
1
1 A. P. Scott and L. Radom, J. Phys. Chem., 1996, 100, 16502.
2 J. P. A. Heuts, R. G. Gilbert and L. Radom, Macromolecules, 1995,
2
8, 8771.
1
3 J. H. Bowie and T. Blumenthal, J. Am. Chem. Soc., 1975, 97, 2959;
I. Howe, J. H. Bowie, J. E. Szulejko and J. H. Beynon, Int. J. Mass
Spectrom. Ion Processes, 1980, 34, 99.
14 S. Dua, R. A. J. O’Hair, J. H. Bowie and R. N. Hayes, J. Chem. Soc.,
Perkin Trans. 2, 1992, 1151.
1
5 VG ZAB 2HF, VG Analytical, Manchester, UK.
16 M. B. Stringer, J. L. Holmes and J. H. Bowie, J. Am. Chem. Soc.,
986, 108, 3888.
7 T. Shono, Y. Matsumua, O. Onomura and Y. Yamada, Synthesis,
1987, 12, 1099.
A mixture of 5,6-epoxyhexanol (0.5 g) and aqueous sodium
3
hydroxide (10%, 5 cm ) was allowed to stir for 60 h, (i) at
1
2
0 ЊC, and (ii) at 100 ЊC (see Table 2). The reaction mixtures
1
were sampled at various times. Each sample was acidified with
aqueous hydrogen chloride (10%) until the pH was 6, extracted
18 K. Nishitani, Y. Harada, Y. Nakamura, K. Yokoo and Y.
3
Yamakawa, Tetrahedron Lett., 1994, 35, 7809.
with dichloromethane (5 cm ), the organic extract dried
1
9 M. Mihailovic and D. Marinkovic, Croatica Chem. Acta, 1986, 59,
(
MgSO ) and concentrated. The product composition was
4
1
09.
analysed by GC/MS (see Table 2).
2
2
0 S. Martin, C. Tu and T. Chou, J. Am. Chem. Soc., 1980, 102, 5274.
Tetrahydropyran-2-methanol and 3-oxepanol were each
treated with aqueous sodium hydroxide as detailed above. No
reaction was observed in either case.
1 J. Appo Rao and M. P. Cava, J. Org. Chem., 1989, 54, 2751.
22 E. J. Eisenbraun, Org. Syn. Coll. Vol. 5, 1973, 5274.
Acknowledgements
This research was financed by the Australian Research Council.
Paper 8/08853E
4
64
J. Chem. Soc., Perkin Trans. 2, 1999, 457–464