Reactions of 3-phenyloxetane and 7-oxabicyclo[2.2.1]heptane with dinitrogen pentoxide in dichloromethane

Article abstract of DOI:10.1039/a707260k

3-Phenyloxetane reacts initially almost entirely by ortho and para aromatic nitration. This is followed by oxetane ring-opening to give the corresponding 2-arylpropane-1,3-diol dinitrates. The oxetane ring-opening reactions, and the reaction of 7-oxabicyc

Full text of DOI:10.1039/a707260k

View Article Online / Journal Homepage / Table of Contents for this issue
Reactions of 3-phenyloxetane and 7-oxabicyclo[2.2.1]heptane with
dinitrogen pentoxide in dichloromethane
Jonathan C. Dormer, Kevin A. Hylands and Roy B. Moodie*
Department of Chemistry, University of Exeter, Exeter UK EX4 4QD
3-Phenyloxetane reacts initially almost entirely by ortho and para aromatic nitration. This is followed
by oxetane ring-opening to give the corresponding 2-arylpropane-1,3-diol dinitrates. The oxetane ring-
opening reactions, and the reaction of 7-oxabicyclo[2.2.1]heptane to give exclusively trans-cyclohexane-
1,4-diol dinitrate, are approximately second order in dinitrogen pentoxide and have highly negative
entropies of activation.
O
Introduction
O2NO
ONO2
Oxiranes1 and 3-substituted oxetanes2 react with dinitrogen
pentoxide in dichloromethane to give the corresponding
dinitrates in good yield. These and related reactions are of con-
siderable interest as new synthetic routes for energetic
materials.3 We have sought to investigate the mechanism of
these and related reactions.4–7 We report here on the reaction of
3-phenyloxetane, 1 and of a strained five-membered cyclic
ether, 7-oxabicyclo[2.2.1]heptane, 9, with dinitrogen pentoxide.
In our study of the reaction of 1 with nitric acid in dichloro-
methane,7 we found the reaction to be a mixture of aromatic
nitration (to give 3 and 5) followed by oxetane ring-opening,
and oxetane ring-opening (to give 2) followed by aromatic
nitration. Both routes led to the same pair of products, 4 and 6.
Aromatic nitration kinetics exhibited a higher order in nitric
acid than oxetane ring-opening, and the former route was thus
preferred at the higher concentrations of nitric acid. With
N2O5, initial aromatic nitration is overwhelmingly preferred,
and subsequent oxetane ring-opening is relatively slow, as
discussed below.
1
2
O
O2NO
ONO2
NO2
NO2
3
4
O
O2NO
ONO2
The reaction of 7-oxabicyclo[2.2.1]heptane, 9, a strained five-
membered ring cyclic ether, was of interest in extending the
range of cyclic ethers which can be ring-opened in this way, the
more so because unstrained cyclic five- and six-membered cyclic
ethers give only low yields of the corresponding dinitrates.8
There is the added advantage that stereochemical information
can be gained from the possibility of cis and trans isomer for-
mation in the expected cyclohexane-1,4-diol dinitrate product
10. (Only the trans isomer is depicted.) The stereochemical
advantages have been utilised previously in the reaction of 9
with dimethyl boron bromide9 which gives exclusively the trans
product, inversion of configuration indicating an SN2 process.
In contrast in aqueous hydrogen bromide after a much longer
reaction time a mixture of cis and trans products is formed.10
NO2
NO2
5
6
O2NO
ONO2
NO2
O2NO
ONO2
NO2
NO2
7
8
ONO2
O
Results
Aromatic nitration of 1 by dinitrogen pentoxide is too rapid to
follow kinetically. At Ϫ60 ЊC using a solution of 0.5 mol dmϪ3
dinitrogen pentoxide in dichloromethane, nitration was com-
plete within 5 min and the substrate had been converted quanti-
tatively to 3-(2-nitrophenyl)oxetane, 3, (10%) and 3-(4-nitro-
phenyl)oxetane, 5, (90%). In contrast to the reaction of the
same substrate with nitric acid,7 there is no detectable com-
peting oxetane-ring opening to give 2 in this initial reaction.
Both the initial products, 3 and 5, subsequently undergo
oxetane-ring opening to give 4 and 6 respectively. The reaction
3→4 is too quick for accurate determination of the rate con-
stant, but it appears to be 25–50 times quicker at 0 ЊC than
O2NO
10
9
5→6. Rate constants for the reaction 5→6, and approximate
rate constants for 3→4, deduced by computer fitting11 of data
as in Fig. 1 are in Table 1. The much faster reaction of 3→4
compared with 5→6 is as found previously in our investigation
of the reaction of 1 with nitric acid.7 (As previously,7 3 has not
been isolated. Its structure is inferred from the NMR spectrum
and the fact that further reaction gives rise to 4.)
The reaction 5→6 is approximately second order in N2O5.
J. Chem. Soc., Perkin Trans. 2, 1998
243

View Article Online
Table 1 Rate constants k for the reactions 3→4 and 5→6 deduced
from the yields of 3 ϩ 5, 4 and 6 in samples quenched at time intervals
following reaction of 1 (0.04 mol dmϪ3) with dinitrogen pentoxide in
dichloromethane
Table 2 Rate constants k for the reaction 9→10 deduced from the loss
of 9 in samples quenched at time intervals following reaction of 9
(7 × 10Ϫ3 mol dmϪ3) with dinitrogen pentoxide in dichloromethane
T/ЊC
[N2O5]/mol dmϪ3
k/10Ϫ5 sϪ1
k/10Ϫ4 sϪ1
32
32
25
25
25
25
25
25
25
25
17
17
17
17
10
10
10
0.127
0.098
0.264
0.186
0.154
0.133
0.126
0.116
0.087
0.063
0.317
0.141
0.120
0.118
0.210
0.100
0.084
33
14
71
60
24
13
24
10
5.5
3.3
65
16
15
10
41
6.5
5.1
T/ЊC
Ϫ60
Ϫ60
Ϫ21
[N2O5]/mol dmϪ3
3→4a
5→6b
0.50
1.0
1.0
0.50
0.75
1.0
4
6
12
4
14
35
d
0.0054c
d
0.134
0.157
0.351
0.76
0
0
0
0
1.8
1.66
a
b
50%.
5% unless otherwise indicated. c Based on two points only,
50%. d Not determined.
R
R
•
NO3
•
H
ONO2
•
NO2
R
R
-HNO3
NO2
NO2
Fig. 1 Yields (% of initial 1) of products of reaction of 1 with dinitro-
gen pentoxide (1 mol dmϪ3) in dichloromethane at 0 ЊC, as a function of
time (t/ks) Triangles (observed) and full line (calculated); 3 ϩ 5.
Squares (observed) and dot-dash line (calculated); 4. Circles (observed)
and dashed line (calculated); 6. Calculated11 curves are based on the
rate constants in Table 1.
H
H
ONO2
Scheme 1
The product of nitration of 9 was exclusively the trans-
cyclohexane-1,4-diol dinitrate 10. This was identified by its H
1
Third order rate constants were calculated by dividing the first
order rate constants in Table 1 by the square of the concen-
tration of N2O5. From these third order rate constants over the
range Ϫ60 to 0 ЊC, activation parameters for the reaction 5→6
were calculated: ∆H‡ = 26 ( 4) kJ molϪ1, ∆S‡ = Ϫ232 ( 20)
NMR spectrum at Ϫ100 ЊC, which showed two signals in the
methine region of unequal intensities corresponding to the
axial–axial and equatorial–equatorial conformations. The cis
isomer would have given two signals of equal intensity. (Com-
parison was made with a mixture of the cis and trans isomers
prepared by nitration of a mixture of cis and trans-cyclo-
hexane-1,4-diols.) A study of the kinetics of ring-inversion of
10 studied by variable temperature NMR spectroscopy is
reported in the appendix.
J molϪ1 KϪ1
.
A preliminary investigation was also made of the reaction
of 1 with N2O5 at 30 ЊC and it was found that side reactions
intruded. A significant amount (about 10%) of an additional
product, possibly the m-nitrophenyl analogue, 7, was detected
by NMR spectroscopy. It was the major product initially but
its concentration did not increase with time. This is remin-
iscent of some results of ozone-mediated nitrations with
nitrogen dioxide, (Kyodai nitration12) and attributed to oxid-
ation by the nitrate radical to give the aromatic radical cation
and nitrate ion. In the early stages of the reaction these
can combine and be followed by combination with NO2 and
nitric acid elimination to give the meta nitro product, as in
Scheme 1. Later, the acid produced is thought to capture the
nitrate ion and prevents further reaction by this pathway.12
In the present work, the NO3 and NO2 radicals would be
formed by homolytic decomposition of dinitrogen pentoxide.
Traces of another reaction product, possibly the dinitro com-
pound 8, which also was formed initially but did not increase
with time, were also detected. These reactions were not investi-
gated further. In a separate experiment, it was demonstrated
that there was no significant reaction of 1 with dinitrogen
tetroxide. Compound 1 was recovered in 95% yield after 1
hour at 22 ЊC after treatment with N2O4 (1.54 mol dmϪ3 in
dichloromethane).
The kinetics of the reaction of 9 with N2O5 to give 10 were
investigated by GC analysis of quenched samples (see Experi-
mental section). With N2O5 in excess, the reaction was quanti-
tative and followed a pseudo-first order course; rate constants
are in Table 2. These results indicate a second-order dependence
on the concentration of N2O5. Activation parameters calcu-
lated from the derived third-order rate constants are: ∆H‡ = 29
( 6) kJ molϪ1; ∆S‡ = Ϫ217 ( 22) J molϪ1 KϪ1
.
Reactions with higher concentrations of 9, approximately
equimolar with N2O5, were less successful. The yield of 10 was
substantially reduced and there was an unidentified product.
This was not investigated further.
There was no reaction of 9 with anhydrous nitric acid under
the conditions of these experiments. However reactions with
N2O5 in the presence of nitric acid showed rate enhancement
over those where nitric acid was absent. The order in N2O5
remained approximately two. Details are in Table 3. When 15
N
enriched N2O5 was used, the 15N NMR analysis during the
course of the nitration at 25 ЊC revealed no chemically induced
dynamic nuclear polarization (CIDNP) signals. There was thus
no evidence for the intrusion of a radical reaction.
244
J. Chem. Soc., Perkin Trans. 2, 1998

View Article Online
Table 3 Rate constants k for the reaction 9→10 deduced from the loss
of 9 in samples quenched at time intervals following reaction of 9
(7 × 10Ϫ4 mol dmϪ3) with dinitrogen pentoxide in dichloromethane, in
the presence of HNO3 at 25 ЊC
Table 4 Rate constants (k/sϪ1) for ring inversion of 10
Minor to major
conformer
Major to minor
conformer
T/K
[N2O5]/mol dmϪ3
[HNO3]/mol dmϪ3
k/10Ϫ4 sϪ1
193
198
203
208
216
223
90
165
300
520
1000
2750
68
124
226
392
754
0.101
0.152
0.163
0.174
0.260
0.037
0.056
0.057
0.066
0.098
2.0
4.0
5.6
5.0
15
2075
Discussion
Aromatic nitration of 1 is too rapid for kinetic study but the
isomer proportions are similar to those observed in nitric acid
nitration,7 and it seems likely that a similar, normal nitronium
ion mechanism is operative. With nitric acid, the reacting sub-
strate was identified as the oxetane hydrogen-bonded to nitric
acid. With N2O5 it must be the free oxetane which is reacting.7
Even so the preference for para nitration is again marked.
Ortho nitration leads to oxetane 3 which ring-opens to 4
much more rapidly than 5 to 6, (indicative of intramolecular
assistance of ring-opening by the ortho-nitro group as has been
discussed before7) and it was possible to find conditions (0 ЊC, 5
min) where the products were a mixture of 4 (7%) and 5 (93%)
which could readily be separated chromatographically. Thus 5
can be prepared in high yield by this method.
The ring-openings 5→6 and 9→10 proceed at similar rates.
Both reactions are approximately second order in N2O5, and
both show large negative entropies of activation. It seems likely
that the two reactions proceed by a similar mechanism. We
could find no evidence that the reactions are radical in char-
acter, and it seems more likely that one molecule of N2O5 pro-
vides an incipient nitronium ion to the ether oxygen whilst the
other concertedly provides an incipient nitrate ion for backside
nucleophilic attack. It is conceivable in view of the entropies
of activation that in this relatively non-polar medium full
charge development is avoided in a 10-membered cyclic transi-
tion state as in 11. (Entropies of activation for third-order reac-
Fig. 2 Observed (l.h.s.) and computed (ref. 14, r.h.s.) 1H NMR
spectra in the methine region of 10
and aryloxirane→arylethane-1,2-diol dinitrate on the other are
substantially different.
Appendix
O
The kinetics of ring-inversion of trans-cyclohexane-1,4-diol
dinitrate, 10, were investigated by dynamic NMR. Spectra at
various temperatures, and the corresponding computer fits,14
are shown in Fig. 2. The spectrum at Ϫ100 ЊC (not shown) has
two unequal sharp peaks in the methine region, with the major
equatorial–equatorial conformation being favoured over the
minor axial–axial one in the ratio 3:2. At higher temperatures
these peaks broaden and coalesce. Rate constants for inter-
conversion are in Table 4. Activation parameters deduced from
these results are: minor to major conformer: ∆H‡ = 39.4 ( 0.6)
kJ molϪ1; ∆S‡ = Ϫ0.3 ( 2.9) J molϪ1 KϪ1; major to minor con-
former: ∆H‡ = 39.4 ( 0.6) kJ molϪ1; ∆S‡ = Ϫ2.6 ( 2.9) J molϪ1
O
N
O
O
N
O
O
O
O
N
O
O
N
O
11
tions cannot be compared directly with those for second-order
reactions because the comparison is dependent on the concen-
tration unit used. However on the molar scale as here, pre-
association of any two components would be expected to have
KϪ1
.
Experimental
an associated entropy change of about Ϫ50 J KϪ1 mol
.
Ϫ1 13 This
accounts for only a part of the observed large negative entro-
pies of activation.)
Apparatus
All NMR spectra were recorded on either a 250 MHz Bruker,
a 300 MHz Bruker or a 400 MHz Bruker spectrometer. IR
spectra were recorded by a Perkin-Elmer 881 Infra-red
spectrophotometer.
These reactions share some characteristics (second order in
N2O5, large negative activation entropies) with the oxirane
ring-opening reactions of aryl oxiranes. However 4-nitro-
phenyloxirane is 300–400 times more reactive than 3-(4-nitro-
phenyl)oxetane, 5, and aryloxirane ring-opening is accom-
panied by racemisation, not inversion.4 The oxirane reaction
shows substituent effects indicative of aryl-ring stabilised
carbocation character in the transition state. No such possibil-
ity exists with the ring-opening reactions of 5 and 9. It therefore
seems likely that despite the similarities mentioned the transi-
tion states for the reactions 5→6 and 9→10 on the one hand
Materials
Dichloromethane was dried by distillation in silanised vessels
from calcium hydride. Precautions were taken to exclude
moisture throughout the experiments described. 7-Oxabicyclo-
[2.2.1]heptane, 9, was fractionally distilled (bp 118 ЊC). Dini-
trogen pentoxide,5 anhydrous nitric acid,15 3-phenyloxetane,
1,15 2-(o-nitrophenyl)propane-1,3-diol dinitrate, 4,7 3-(p-nitro-
J. Chem. Soc., Perkin Trans. 2, 1998
245

View Article Online
phenyl)oxetane, 515 and 2-(p-nitrophenyl)propane-1,3-diol di-
nitrate, 6,7 were prepared as described.
dichloromethane at 35 ЊC (ca. 20 cm3) was added. The dichloro-
methane was added at this temperature to overcome the tem-
perature loss due to the cold glass. More dichloromethane at
25 ЊC (ca. 20 cm3) was then added followed by a solution of 9 (7
mmol dmϪ3, 5 cm3). This solution also contained the GC refer-
ence standard. The solution was quickly made up to 50 cm3
with dichloromethane and then placed in a thermostatted com-
partment. Four aliquots (10 cm3) were taken at time intervals
and syringed into sealed flasks containing water, and shaken.
Flasks were removed from the glove box and the contents
extracted with dichloromethane (3 × 10 cm3). The organic frac-
tions were combined, dried (MgSO4) and the bulk of the solv-
ent removed by distillation using a small fractionating column,
until only a small volume remained. This was analysed by GC.
(Column Dextril 400 at 100 ЊC, N2 flow rate 40 cm3 min1,
response factor relative to reference standard 3-fluoronitro-
benzene 0.852. Retention times of 9 and ref. standard 3.0 and
14.06 min, respectively.)
Preparation of 15N-labelled N2O5
Anhydrous 15N-labelled HNO3 was added dropwise with stir-
ring to P2O5 in a slow stream of ozonised oxygen, then passed
through a tube containing P2O5 before collection in an NMR
tube fitted with a B 10 joint and cooled in a dry-ice–acetone
bath. After collection the NMR tube was subasealed, removed
to the glove-box, and reweighed. A known volume of dichloro-
methane was added by syringe, and the solution was used
immediately.
Preparation of a mixture of cis- and trans-cyclohexane-1,4-diol
dinitrate
To anhydrous nitric acid (8 g) sulfuric acid (98%, 19.3 g) was
cautiously added in an ice-bath. The mixed acid was cooled to
Ϫ35 ЊC and dry dichloromethane (20 cm3) added. To this cold
solution a mixture of cis- and trans-cyclohexane-1,4-diol (3 g,
0.026 mol) was added slowly, with stirring, over 15 min.
Towards the end of the addition the solution became very
viscous, and the temperature was allowed to rise to Ϫ25 ЊC to
facilitate stirring which was continued for 1 hour. The reaction
mixture was quenched with ice and the organic layer was
washed with sodium hydrogen carbonate solution, separated
and dried over MgSO4. Removal of solvent gave a pale yellow
crystals (2.9 g, 0.14 mol, 54% yield); δH(300 MHz, CDCl3) 5.05
(2H, br s, CH) 1.95 (8H, m, CH2 ring); δC(75.5 MHz, CDCl3)
79.04, 78.65 (2 × CH) 25.74, 25.31 (2 × CH2 ring); νmax/cmϪ1
1625 (asym. NO2 str.) 1274 (sym. NO2 str.).
Acknowledgements
We thank Dr R. P. Claridge, Dr R. W. Millar and Dr J. P. B.
Sandall for helpful discussions, and Dr V. Sik for assistance
with the dynamic NMR studies described in the appendix. This
work has been carried out with the support of the Defence and
Evaluation Research Agency.
References
1 P. Golding, R. W. Millar, N. C. Paul and D. H. Richards,
Tetrahedron, 1993, 49, 7037.
2 P. Golding, R. W. Millar, N. C. Paul and D. H. Richards,
Tetrahedron, 1993, 49, 7051.
3 R. W. Millar, M. E. Colclough, P. Golding, P. J. Honey, N. C. Paul,
A. J. Sanderson and M. J. Stewart, Philos. Trans. R. Soc. London, A,
1992, 339, 305.
4 J. C. Dormer and R. B. Moodie, J. Chem. Soc., Perkin Trans. 2,
1994, 1195.
5 R. J. Lewis and R. B. Moodie, J. Chem. Soc., Perkin Trans. 2, 1996,
1315.
6 R. J. Lewis and R. B. Moodie, J. Chem. Soc., Perkin Trans. 2, 1997,
563.
7 K. A. Hylands and R. B. Moodie, J. Chem. Soc., Perkin Trans. 2,
1997, 709.
8 P. Golding, R. W. Millar, N. C. Paul and D. H. Richards,
Tetrahedron Lett., 1988, 29, 2731 and 1989, 30, 6431.
9 Y. Guindon, M. Therien, Y. Girard and C. Yoakim, J. Org. Chem.,
1987, 52, 1680.
10 D. S. Noyce, B. N. Bastain, P. T. S. Lau, R. S. Monson and
B. Weinstein, J. Org. Chem., 1969, 34, 1247.
Preparation of trans-cyclohexane-1,4-diol dinitrate, 10
To a solution of dinitrogen pentoxide (5.6 mmol) in dichloro-
methane (20 cm3) at 25 ЊC 7-oxabicyclo[2.2.1]heptane (1.06
mmol) was added, and the mixture was kept at 25 ЊC for 4
hours. After this time it was quenched in water and the organic
layer was separated and dried with MgSO4. The solvent was
removed to leave a pale yellow solid. This was dissolved in
ethanol and refluxed with activated carbon, then filtered hot
and recrystallised from ethanol to give white crystals (mp
124.4–125.1 ЊC, 99% yield); δH(300 MHz, CDCl3) 5.08 (2H, s,
CH) 2.12 (4H, m, CH2 ring) 1.79 (4H, m, CH2 ring). (In the 15
N
proton decoupled spectrum there was a single sharp signal at
Ϫ42.60 relative to external nitromethane. In the proton coupled
spectrum this signal was a doublet with a coupling constant,
3JNH, of 2.6 Hz.); νmax/cmϪ1 1632 (asym. NO2 str.) 1275 (sym.
NO2 str.).
11 ‘Scientist’ package, MicroMath Scientific Software.
12 H. Susuki, T. Murashima and T. Mori, J. Chem. Soc., Chem.
Commun., 1994, 1443.
13 E. A. Moelwyn-Hughes, The chemical statics and kinetics of
solutions, Academic Press, London, 1971.
14 D. A. Kleier and G. Binsch, DNMR3 Computer programme 165
Quantum Chemistry Programme exchange, Indiana University,
1970; V. Sik, Ph.D. Thesis, University of Exeter, 1979.
15 K. A. Hylands and R. B. Moodie, J. Chem. Soc., Perkin Trans. 2,
1996, 2073.
Products of reaction of 1 with N2O5
To each of several aliquots (1 cm3) of a solution of N2O5 in
dichloromethane in separate flasks were added aliquots (0.25
cm3, 0.2 mol dmϪ3) of a solution of 1 in dichloromethane. Each
was quenched in an excess of cold aqueous sodium hydrogen
carbonate solution after an appropriate time interval at
the required temperature. Repeated extractions with dichloro-
methane, followed by drying (MgSO4) and removal of solvent
gave a mixture of products which was dissolved in CDCl3 and
analysed by 1H NMR spectroscopy.
Paper 7/07260K
Received 7th October 1997
Accepted 4th November 1997
Typical kinetic experiment with 7-oxabicyclo[2.2.1]heptane, 9
To N2O5 (7.7 mmol) in a dry volumetric flask in a dry glove box,
246
J. Chem. Soc., Perkin Trans. 2, 1998