The mechanism of solvolysis of 2-adamantyl azoxytosylate: Isotopic labelling, medium effect, and attempted deoxygenation studies

Article abstract of DOI:10.1139/v98-071

Rates of solvolysis of 2-adamantyl azoxytosylate (1) have been measured over a range of temperatures in ethanoic acid, methanoic acid, 50:50 (v/v) trifluoroethanol:water, 80:20 (v/v) trifluoroethanol:water, 97:3 (w/w) trifluororoethanol:water, and 70:30 (v/v) ethanol:water. For comparison, rates of solvolysis of 2-adamantyl tosylate (2) have also been measured in 50:50, 80:20, and 90:10 (v/v) trifluoroethanol:water, and for both compounds, activation parameters have been determined. These and results published earlier allow a correlation of the two reactions and indicate that the m value for 2-adamantyl azoxytosylate solvolysis is only 0.46. This is one of the lowest m values for a reaction that is unambiguously an S(N)1 solvolysis. We have also recorded the ¹⁷O NMR spectrum of the 2-adamantyl tosylate formed from ¹⁷O-labelled 2-adamantyl azoxytosylate in deuteriochloroform, and the millimeter-wave spectrum of the nitrous oxide evolved in the hydrolysis of ¹⁸O-labelled 2-adamantyl azoxytosylate. These labelling studies have provided more detailed knowledge of the S(N)l fragmentation mechanism of 1 and exclude a mechanism of reaction via rearrangement to N- nitroso,N-(2-adamantyl),O-(p-toluenesulfonyl)hydroxylamine (5). Attempted deoxygenation of 1 to give 2-adamantyl diazotosylate (8) and a subsequent fragmentation proved unsuccessful.

Full text of DOI:10.1139/v98-071

862
The mechanism of solvolysis of 2-adamantyl
azoxytosylate: isotopic labelling, medium effect,
and attempted deoxygenation studies
John K. Conner, Johanna Haider, M.N. Stuart Hill, Howard Maskill, and
Monique Pestman
Abstract: Rates of solvolysis of 2-adamantyl azoxytosylate (1) have been measured over a range of temperatures in ethanoic
acid, methanoic acid, 50:50 (v/v) trifluoroethanol:water, 80:20 (v/v) trifluoroethanol:water, 97:3 (w/w) trifluoroethanol:water,
and 70:30 (v/v) ethanol:water. For comparison, rates of solvolysis of 2-adamantyl tosylate (2) have also been measured in
50:50, 80:20, and 90:10 (v/v) trifluoroethanol:water, and for both compounds, activation parameters have been determined.
These and results published earlier allow a correlation of the two reactions and indicate that the m value for 2-adamantyl
azoxytosylate solvolysis is only 0.46. This is one of the lowest m values for a reaction that is unambiguously an SN1
solvolysis. We have also recorded the 17O NMR spectrum of the 2-adamantyl tosylate formed from 17O-labelled 2-adamantyl
azoxytosylate in deuteriochloroform, and the millimeter-wave spectrum of the nitrous oxide evolved in the hydrolysis of
18O-labelled 2-adamantyl azoxytosylate. These labelling studies have provided more detailed knowledge of the SN1
fragmentation mechanism of 1 and exclude a mechanism of reaction via rearrangement to N-nitroso,N-(2-adamantyl),O-
(p-toluenesulfonyl)hydroxylamine (5). Attempted deoxygenation of 1 to give 2-adamantyl diazotosylate (8) and a subsequent
fragmentation proved unsuccessful.
Key words: nitrous oxide, carbenium ion, isotopic labelling, solvolysis, m value.
Résumé : On a mesuré les vitesses de solvolyse de azoxytosylate d’adamant-2-yle (1) à diverses températures, dans l’acide
éthanoïque, l’acide méthanoïque, le trifluoroéthanol:eau (50:50, v/v), le trifluoroéthanol:eau (80:20, v/v), le
trifluoroéthanol:eau (97:3, p/p) et l’éthanol:eau (70:30, v/v). Pour fins de comparaison, on a aussi mesuré les vitesses de
solvolyse du tosylate d’adamant-2-yle (2) dans des mélanges (50:50, 80:20 et 90:10, v/v) de trifluoroéthanol:eau et, dans les
cas des deux composés, on a déterminé les paramètres d’activation. Ces résultats couplés à d’autres publiés antérieurement
permettent d’établir une corrélation entre les deux réactions et ils indiquent que la valeur m pour la solvolyse de
l’azoxytosylate d’adamant-2-yle n’est que 0,46. Ceci correspond à l’une des valeurs de m les plus faibles pour une réaction
qui est sans ambiguïté SN1. On a aussi déterminé le spectre RMN du 17O du tosylate d’adamant-2-yle formé à partir de
l’azoxytosylate d’adamant-2-yle marqué au 18O lors d’une réaction dans le chloroforme deutéré ainsi que le spectre d’onde à
un millimètre de l’oxyde nitreux qui se dégage lors de l’hydrolyse de l’azoxytosylate d’adamant-2-yle marqué au 17O. Ces
études de marquage ont fourni des connaissances plus détaillées du mécanisme de fragmentation SN1 du composé 1 et elles
ont permis d’exclure un mécanisme réactionnel impliquant un réarrangement en N-nitroso-N-(adamant-2-yl)-O-(p-
toluènesulfonyl)hydroxylamine (5). Il a été impossible de provoquer une désoxygénation du composé 1 pour donner du
diazotosylate d’adamant-2-yle (8) et provoquer une fragmentation subséquente.
Mots clés : oxyde nitreux, ion carbénium, marquage isotopique, solvolyse, valeur m.
[Traduit par la rédaction]
Introduction
ions are amongst the earliest to have been investigated, and
early proposals that some of the most deep seated of molecular
rearrangements (e.g., amongst terpenoids) occur via such in-
termediates were some of the boldest (1). These proposals
were accepted only after fierce debate and overwhelming evi-
dence, and much of current organic chemistry reaction mecha-
nisms vernacular originates from that Olympian period earlier
this century. However, even though the principal features of
substitution and elimination reactions that involve carbenium
ions are well known, some details of the natures of these inter-
mediates have been inferred from only indirect evidence or
still remain obscure.
Mechanisms of reactions involving intermediate carbenium
This paper is dedicated to Professor Erwin Buncel in
recognition of his contributions to Canadian chemistry.
Received October 7, 1997.
J.K. Conner. Chemistry Department, University of Stirling,
Stirling FK9 4LA, United Kingdom.
J. Haider, H. Maskill,1 M. Pestman, and M.N.S. Hill.
Chemistry Department, University of Newcastle, Newcastle
upon Tyne NE1 7RU, United Kingdom.
1
According to a frequently encountered model for SN1 reac-
tions (2a), the carbenium ion is generated in an initial rate-lim-
iting step, then is captured by a nucleophile (which could be a
Author to whom correspondence may be addressed.
Telephone: 44-191-222-7088. Fax: 44-191-222-6929. E-mail:
h.maskill@ncl.ac.uk
Can. J. Chem. 76: 862–868 (1998)
© 1998 NRC Canada

863
Conner et al.
Scheme 1.
solvolysis of a standard compound may be quantified using the
Winstein–Grunwald equation (4) (or an extended (5) or modi-
_
N O
2
_
Ad N O OTs
*
*
+
Ad-OTs
fied (6) version). An m value, the gradient of the log(kcompound
)
2
versus log(kstandard) correlation, close to unity means that the
solvent effect for the solvolysis of the compound in question
is the same as for the standard compound, and this is usually
interpreted as evidence for very closely similar SN1 solvolytic
mechanisms. The original standard compound was tert-butyl
chloride but 2-adamantyl tosylate is now regarded as more
appropriate for substrates with arenesulfonate nucleofuges (7,
8). An m value appreciably lower than unity is usually inter-
preted as indicating that the reaction involves a low degree of
charge separation in the transition structure (assuming there
are other reasons for believing that the reaction is still an SN1),
or that the mechanism is not an SN1 at all.
Recently, there has been an expansion in our knowledge of
carbenium ion chemistry partly through developments in fast
reaction methods pioneered by McClelland et al. (9), partly
through new experimental techniques developed, for example,
by Olah et al. (10) and Mayr and Patz (11) for preparing car-
benium ions under nonnucleophilic conditions, partly through
improved theoretical procedures for studying isolated carbe-
nium ions themselves, for example, by Schleyer and co-work-
ers (12), and partly due to better conceptual models for
solvolytic and related reactions offered by Jencks and Richard
(13). It is within this last context that the present project had
its origins. Some years ago (14), we introduced the solvolysis
of alkyl azoxyarenesulfonates as a heterolytic reaction, which
is unambiguously unimolecular and allows the study of the
SN1 mechanism uncomplicated by competitive bimolecular
substitution. The carbenium ions so formed are the same as
when produced by other means, but their initial environment
is different. We proposed a mechanism that is a link between
the solvolysis of alkyl arenesulfonates on the one hand and
nitrous acid induced deamination of primary alkylamines (and
related reactions) on the other (14, 15). We also presented
evidence that the initial fragmentation to give an ion pair and
nitrous oxide is rate-determining (i.e., there is no internal
2
_
N O
2
A
_
*
_
TsO
Ar
_
2 H O
2
O
N
SO
2
+
+
H O
3
O
*
Ad-OH
+
Ad
N
1
*
_
_
B
N O
2
2 H O
2
_
Ar
TsO
SO
_
Ad N O OTs
2
*
+
O
N
Ad-OTs
2
5
*
*
_
N O
2
2
O
Ad
N
*
17
O = O or
18
Ts = p-Me-C H -SO
2
Ar = p-Me-C H
6
O
6
4
4
solvent molecule) in a subsequent product-forming step. In this
simple model, the transition structure in the first step involves
vestigial bonding between the developing trigonal electro-
philic carbon and the nucleofuge. According to the more com-
plete Winstein–Shiner ion-pair model (2b, 3), the rate-limiting
step in the solvolysis of 2-adamantyl tosylate, for example, is
not the initial ionization but the separation of the first-formed
contact (or intimate) ion pair to give the solvent-separated ion
pair. Clearly, this model requires that the contact ion pair un-
dergoes internal return in competition with, but faster than,
progression to the solvent-separated ion pair. In this model, the
transition structure in the rate-limiting step of the solvolytic
process involves a fully developed positive charge in the elec-
trophilic fragment, and the solvent-separated ion pair is the
main product-forming intermediate from 2-adamantyl tosy-
late.
Comparison of the effect of solvents upon the rate of
solvolysis of a compound with the effect upon the rate of
_
O
+
N
OTs
OTs
N
2
1
O
C
CH
OBs
O
Cl
6
7
_
_
O
δ
OTs
+
N
OTs
N
N
N
+
δ
H
8
9
© 1998 NRC Canada

864
Can. J. Chem. Vol. 76, 1998
Scheme 2.
O
N
N
R'
O
_
C
O
N
N
+
R' CO2
N
R'
R
2
Products
C
O
R
R
Scheme 3. (i) NH3OH+Cl–, H2SO4, EtOH; (ii) NaH3BCN, MeOH, HCl; (iii) NaNO2, H3O+Cl–, H2O (using ordinary water or oxygen-labelled
water as appropriate); (iv) TsCl, pyridine. The asterisk denotes the position of the 17O or 18O isotope when the nitrosation was carried out using
labelled water.
(i)
(ii)
NHOH
O
NOH
(iii)
(iv)
_
_
O
H
O
O
N
+
N
*
+
*
OTs
N
N
3
1
return) and concerted (14, 16). In nonnucleophilic, weakly ion-
izing solvents such as chloroform or toluene, the initially
formed ion pair collapses to give the covalent alkyl arenesul-
fonate whereas in ionizing nucleophilic media such as water,
the carbocation is captured by a solvent molecule to give a
solvolytic substitution product (17). These possibilities are il-
lustrated for 2-adamantyl azoxytosylate (1) in Scheme 1, path
A. An alternative mechanism for the decomposition of 1 is
included in Scheme 1 path B, which is based (in part, see later)
upon the probable mechanism for the isomerization of N-al-
kyl,N-nitrosoamides into alkyl diazocarboxylates prior to frag-
mentation in the nitrosoamide method of deamination and
related reactions, Scheme 2 (18). In this route, there is an initial
rearrangement of the substrate to give N-nitroso,N-(2-adaman-
tyl),O-(p-toluenesulfonyl)hydroxylamine (5), possibly but not
necessarily through a concerted cyclic mechanism, followed
by the fragmentation step.
We now report a comparison of the decompositions in a
range of solvents of 2-adamantyl azoxytosylate (1) and 2-
adamantyl tosylate (2), the latter long having been regarded as
an archetypal SN1 substrate (19); we also compare the
solvolyses of 1 and pinacolyl brosylate (3,3-dimethyl-2-butyl
p-bromobenzenesulfonate, 6) (19, 20), an SN1 substrate that
reacts with rate-limiting ionization, i.e., with no internal return,
and 1-adamantyl chloroformate (7), an SN1 substrate that also
solvolyzes with liberation of a gas molecule (21). Addition-
ally, we present oxygen isotope labelling results, which allow
us to refine our earlier mechanistic proposal for the solvolysis
of 1. Finally, attempts to remove the pendant oxygen from 1,
which would lead to the previously unknown but presumably
more unstable 2-adamantyl diazotosylate (8), are described.
Compound 8, which is analogous to the alkyl diazocarboxylate
in Scheme 2 should, if formed, fragment to give a nitrogen-
separated ion pair and thence either 2-adamantyl tosylate or
solvent-derived products.
Results
The two substrates 1 and 2 were made by literature methods,
see Scheme 3 for 1, and all solvolytic media were prepared in
conventional ways. Rate constants were already known for
2-adamantyl tosylate in many solvents, but to allow compari-
son with all our results for 2-adamantyl azoxytosylate, some
new values were required. Rates of reaction of both substrates
in all solvolytic media were investigated by our normal UV
spectrophotometric methods (see experimental section) and
activation parameters were calculated using the Eyring equa-
tion from rate constants measured over a temperature range of
at least 30°C. Our results are shown in Table 1; Fig. 1 is a
correlation of the rate results for the solvolyses of 1 and 2 in 11
solvents (gradient = 0.46, R = 0.96), and Fig. 2 is a correlation
of the rate results for the solvolyses of 1 and pinacolyl brosy-
late (6) in just six solvents (gradient = 0.53, R = 0.96).
We also prepared samples of 2-adamantyl azoxytosylate
specifically enriched with 17O and with 18O in the oxygen
marked with an asterisk in Scheme 3 by reaction of tosyl chlo-
ride (22) with a specifically labelled sample of N-(2-adaman-
tyl)hydroxydiazenium oxide (3), which in turn had been
prepared by nitrosation of N-(2-adamantyl)hydroxylamine by
a method based upon earlier work by Dahn and Ung-Truong
(23). The 17O NMR spectrum of the 2-adamantyl tosylate from
the reaction of the labelled 2-adamantyl azoxytosylate in deu-
teriochloroform after more than 10 half-lives at 65°C unambi-
guously showed enhancement of the signal at δO 158.4 ppm
© 1998 NRC Canada

865
Conner et al.
Table 1. Rate constants at 25°C and enthalpies and entropies of activation for the solvolysis of 2-adamantyl
azoxytosylate (1) and 2-adamantyl tosylate (2).a
106 k25°C (s–1)
∆H‡ (kJ mol–1)
∆S‡ (J K–1 mol–1)
b
Compound
Solvent
YOTs
1
50TFE
80TFE
97TFEd
HCO2H
70E
2.14
1.95c
1.83
3.04
0.47
–0.61
2.14
1.95c
1.9c
22.2
18.3
13.3
19.4
3.56
0.589
2.98e
1.83
1.63
98
98
–5
–6
101
105
107
110
101
96
–1
16
10
AcOH
50TFE
80TFE
90TFE
3
2
–12
–32
–41
94
a Rate constants at 25°C are by extrapolation from results at other temperatures, which in turn are means of several
determinations and estimated to be reliable to 5%. The uncertainties in ∆H‡ and ∆S‡ are estimated to be 5 kJ mol–1 and
8 J K–1 mol–1
.
b Reference 8.
c Estimated from ref. 8.
d Less reliable results of ∆H‡ = 89 kJ mol–1 and ∆S‡ = –35 J K–1 mol–1 were reported earlier in ref. 14b.
e A result of 3.35 × 10–6 s–1 is reported in ref. 8.
Fig. 1. Graph of log k/s–1 for solvolysis of 2-adamantyl
azoxytosylate 1 plotted against log k for solvolysis of 2-adamantyl
tosylate 2.
Fig. 2. Graph of log k/s–1 for solvolysis of 2-adamantyl
azoxytosylate 1 plotted against log k for solvolysis of pinacolyl
brosylate 6.
-4.5
HFIP
97HFIP
50E
HCO2H
-4
97TFE
-5
50TFE
50E
80TFE
50E
HCO2H
TFE
97TFE
-5
-6
-5.5
80E
70E
80E
AcOH
-8
-6
AcOH
-6
-6.5
-6
log(k25AdOTs)
-4
-4
-2
log(k25Pin-OBs)
due to the sulfonyl oxygens; this compares with the signal at
δO 157.5 ppm due to the sulfonyl bonded oxygens in the start-
ing material. The millimeter-wave spectrum of the gas pro-
duced from the hydrolysis of the 18O-labelled substrate was
also recorded (24), which identified it unambiguously as ni-
trous oxide by the J 4←3 band at 96 512 MHz due to N216O;
there was no trace of the J 3←2 band at 71 155 MHz pre-
viously identified in the spectrum of N2O containing the 18O
isotope (25).
Using a variety of reaction conditions and established de-
oxygenating agents (acetyl bromide (26), triethyl phosphite
(27, 28), thionyl chloride (26), phosphorus trichloride (26, 29),
triphenylphosphine (28), tri(n-butyl)phosphine (28b), and lith-
ium metal (30)), we found no unambiguous evidence for the
formation of 2-adamantyl tosylate from 2-adamantyl azoxyto-
sylate under nonnucleophilic conditions.
Scheme 4 has been identified as the first step in the mechanism
of reaction of 3; it involves a proton transfer either inter-
molecularly by protonation/deprotonation under the acidic
conditions, or intramolecularly along a trajectory beginning in
a strong hydrogen bond within 3 (25). In the next acid-cata-
lysed step of this mechanism, the fragmentation is not con-
certed but stepwise via the oxodiazonium ion. In view of this
preferred reaction of 3 via the N-nitroso tautomer, an intra-
molecular rearrangement of 2-adamantyl azoxytosylate (1)
into the isomeric N-nitroso,N-(2-adamantyl),O-(p-toluenesul-
fonyl)hydroxylamine (5) via a cyclic transition structure is in-
cluded as a preliminary to a credible alternative solvolytic
mechanism for 1 in Scheme 1, path B. To distinguish between
paths A and B in Scheme 1, we prepared 2-adamantyl azoxy-
tosylate specifically labelled separately with 17O and 18O in the
tosylated oxygen marked with an asterisk in Scheme 1.
Clearly, if the mechanism follows path A, the labelled oxygen
will end up in the 2-adamantyl tosylate in a nonnucleophilic
medium whereas if the mechanism follows path B it will be
lost as nitrous oxide. We have already shown that nitrous oxide
Discussion
Isomerization of 3, the product of nitrosation of N-(2-adaman-
tyl)hydroxylamine, into 4 prior to the fragmentation in
© 1998 NRC Canada

866
Can. J. Chem. Vol. 76, 1998
Scheme 4.
O
N
N
O
O
+ H+
O
O
N
H
N
N
+
O
H
_
H+
Ad
N
+
H
-
Ad
Ad
H
3
4
_
_
H2O
H2O
+
Ad N N O
+ 2 H2O
Ad +
N2O
Ad OH
H3O+
+
+
enriched with 18O can be readily identified by observation of
the J 3←2 rotational band at 71 155 MHz (25); its absence in
the sample isolated in the present case established that the label
is not in the nitrous oxide. To corroborate this negative evi-
dence and now using 17O enriched starting material 1 (25) we
recorded an enhanced 17O NMR signal at δO 158.4 ppm in the
2-adamantyl tosylate product (2). Clearly, the oxygen marked
with an asterisk in 2-adamantyl azoxytosylate (1) in Scheme 1
ends up in the 2-adamantyl tosylate product (2), and not in the
nitrous oxide, so path B via rearrangement to N-nitroso,N-(2-
adamantyl),O-(p-toluenesulfonyl)hydroxylamine (5) in Scheme 1
may now be ruled out.
tosylate (2) in these solvents are faster than anticipated on the
basis of their reactivity in aqueous alcohols. The most credible
explanation on the basis of existing mechanistic understanding
is that the bidentate carboxylic acid molecules are more effec-
tive than the aqueous alcohols at preventing internal return in
the solvolysis of 2. Internal return is also absent in the
solvolysis of pinacolyl brosylate (6) but for a reason different
to that for 1. In the solvolysis of 6, rearrangement is either
concerted with the ionization or follows on so rapidly that
internal return cannot compete (20). Consequently, in the cor-
relation of pinacolyl brosylate with 2-adamantyl tosylate,
points for carboxylic acids are also not on the line defined by
reactions in aqueous ethanol (19). Low extents of internal re-
turn in the acetolysis and formolysis of 2-propyl brosylate
(compared with trifluoroethanolysis) have been reported ear-
lier (20). A fair correlation is obtained between the log k values
for solvolyses of 1 and 6, neither of which undergoes internal
return, as shown in Fig. 2; the gradient (0.53) corresponds to
an appreciably higher solvent effect for 6 than for 1. This in-
dicates a transition structure for the ionization of 6 with a
higher degree of dipolarity than that in the ionization of 1.
Another model for the reaction of 1 is 1-adamantyl chloro-
formate (7), although it is tertiary rather than secondary. A low
m value determined over a wide range of solvents and appre-
ciably positive ∆S‡ values were reported by Kevill and col-
leagues (21); their solvolytic mechanism for 7 involves
ionization to give the 1-adamantyl cation, carbon dixide, and
chloride. Clearly, fragmentation in this system also cannot be
followed by internal return to give covalent 7, although the
ions of the carbon dioxide separated ion pair can collapse to
give 1-adamantyl chloride.
The logarithmic correlation of solvolytic rate constants at
25°C for 2-adamantyl azoxytosylate (1) and 2-adamantyl to-
sylate (2) is shown in Fig. 1. The nine points for aqueous al-
coholic media constitute an excellent linear plot (R > 0.99),
and the points for the two carboxylic acid solvents are notice-
ably off; omitting these two points has a negligible effect upon
the gradient. Since 2-adamantyl tosylate is the standard com-
pound for defining the solvent parameter Y(OTs), the gradient
of this plot is the m value for 2-adamantyl azoxytosylate (1);
the actual plot for 1 of log k/s–1 against Y(OTs) with the two
carboxylic acid solvents omitted gives m = 0.46 (R > 0.99).
Confidence in our earlier proposal that solvolysis of 1 involves
rate-limiting fragmentation with formation of an ion pair sepa-
rated by a molecule of nitrous oxide (14) is strengthened by
our now having excluded the alternative mechanism for reac-
tion of 1 via rearrangement through 5 (see above). Conse-
quently, this very low m value for a reaction that is
unambiguously an SN1 indicates that the transition structure
for this fragmentation has only a modest degree of dipolarity
spread over four atoms, structure 9. This is in accord with our
conclusion from a double Hammett investigation of solvolyses
of substituted benzyl azoxyarenesulfonates, i.e., that the bonds
from the carbon to one nitrogen and from the other nitrogen to
oxygen are about half broken in the transition structure (16).
In contrast, the rate-limiting step in the solvolysis of 2 involves
conversion of one type of ion pair into another (17, 31), i.e.,
ionization occurs in a prior (reversible) step, and the higher
dipolarity in this transition structure arises from fully devel-
oped charges even though they are at only a small interionic
separation. Our m value for 1 establishes that such a low result
alone cannot be taken as a reliable indicator that a reaction of
a simple secondary alkyl system is not an SN1.
The newly determined activation parameters for the
solvolysis of 1 shown in Table 1 are unexceptional except that
they provide further supportive evidence that the reactions are
SN1 solvolyses, i.e., the enthalpies of activation are relatively
high, and the entropies of activation are small (close to zero).
In general terms, these and earlier results (14) show that the ∆S‡
values are somewhat less negative than the corresponding re-
sults for 2. We attribute this to a positive contribution to the
overall entropy of activation arising from the nascent N2O
molecule in the reaction of 1 not found in the reaction of 2.
However, even this does not compensate in all cases for the
substantial negative contribution to ∆S‡ from the enhanced
solvation that occurs due to the increased dipolarity in the
transition state from 1.
The deviations for the two carboxylic acids in the correla-
tion of Fig. 1 indicate that the reactions of 2-adamantyl
The original isolation of alkyl azoxyarenesulfonates as
© 1998 NRC Canada

867
Conner et al.
stable compounds was initially surprising in view of the insta-
bility of analogous compounds prepared by White et al. (18,
32), which fragment at low temperatures into cation, stable
molecule, and anion. We reasoned that abstraction of the pen-
dant oxygen atom of azoxytosylates such as 1 would give the
corresponding diazotosylate (8) and that fragmentation of 8
would now liberate nitrogen (rather than nitrous oxide), and
consequently be an easier reaction. We anticipated that 8, if
formed, would be too unstable to isolate under the deoxygena-
tion reaction conditions and sought its expected decomposition
product under non nucleophilic conditions, i.e., 2-adamantyl
tosylate (17). Using a range of deoxygenating reagents, we
found no firm evidence by TLC or NMR for the formation of
2 and conclude that 1 does not undergo an easy deoxygenation.
were then made by our usual Tipson (14b, 22) method from
3-17O and 3-18O; both had melting points and 1H NMR spectra
identical with the unlabelled compound.
Decomposition of 17O-labelled 2-adamantyl azoxytosylate
A sample (85 mg) of 17O-labelled 2-adamantyl azoxytosylate
(δO 157.5 (STO)) was dissolved in a 1:4 mixture of chloro-
form:deuteriochloroform (4 cm3) in an NMR tube and main-
1
tained at ca. 50°C in an oil bath for 10 h. The H NMR
spectrum indicated only about 10% conversion, so the tem-
perature of the oil bath was raised to about 60°C, and the
reaction was continued for a further 15 h by which time the
1
conversion was about 40% by H NMR, and the 17O NMR
spectrum showed a single strong but broad band centred at
158 ppm. After 4 days at this temperature, the 1H NMR spec-
trum showed complete conversion to 2-adamantyl tosylate;
there was no trace of the starting material, and the product
showed a strong 17O NMR signal at δO 158.4 ppm (STO).
Experimental
Ethanoic acid was heated under reflux for 12 h with ethanoic
anhydride, then fractionally distilled; trifluoroethanol was
either stirred over molecular sieves for several days, then frac-
tionally distilled, or heated under reflux over calcium hydride
for 2 h, then fractionally distilled; spectroscopic grade abso-
lute ethanol was used as supplied; methanoic acid was purified
by standing over boric anhydride (B2O3) for several days fol-
lowed by successive distillation from fresh B2O3 under re-
duced pressure until the separation of the OH and the CH
signals in the 1H NMR spectrum of the neat distillate became
constant at ca. 3.158 ppm. Aqueous solvolytic media were pre-
pared by mixing appropriate volumes of glass distilled water
with either ethanol or trifluoroethanol except 97TFE, which
was prepared by weight; the ethanoic acid was buffered with
0.15 mol dm–3 anhydrous sodium ethanoate, but all other sol-
vents were unbuffered. Isotopically enriched water was pur-
chased from Goss (10% enriched with 17O) and from Aldrich
(10% enriched with 18O).
Decomposition of 18O-labelled 2-adamantyl azoxytosylate
A gentle flow of nitrogen was passed through a stirred solution
made up from aqueous perchloric acid (70%, 6.1 cm3), water
(24 cm3), and18O-labelled 2-adamantyl azoxytosylate (85 mg,
0.25 mmol) in a flask fitted with a reflux condenser. The reac-
tion was maintained at 69°C for 6 h, then for another 30 min as
the reaction mixture cooled down; the gas flow passed through
a KOH drying tower and a gas trap immersed in liquid nitro-
gen. The contents of the gas trap were then transferred at room
temperature to a millimeter-wave spectrometer using standard
vacuum line procedures, and predetermined regions of the
spectrum were scanned. An intense absorption was recorded at
96 512 MHz due to the J 4←3 transition of N216O; no absorp-
tion was detected in the region of 71 155 MHz previously
identified as due to the J 3←2 transition of N218O (25).
Kinetics
2-Adamantyl azoxytosylate (1) and 2-adamantyl tosylate
(2)
Rates of solvolytic reactions were measured using either a
Pye–Unicam SP8-300 spectrophotometer connected to an Ap-
ple microcomputer with in-house software for data collection
(34) or a Cecil CE 5502 spectrophotometer connected to an
Elonex PC using commercial data collection software. In the
former case, rate constants were calculated also using in-house
software (34); in the latter case, GraFit version 3.0 (Erithacus
Software Ltd.) was used. In both cases, cell holders for up to
four cells were thermostatted by circulating water, the cell-
holder block temperature being measured with a fitted plati-
num resistance thermometer, and the system was programmed
to record absorbance data automatically at predetermined time
intervals. Rate constants were calculated from absor-
bance–time data collected over about three or four half-lives
using standard computational techniques.
2-Adamantyl azoxytosylate (1) (mp 111–112°C, lit. (14)
111–113°C) and 2-adamantyl tosylate (2) (mp 82–83°C, lit.
(33) 82.7–83.7°C; δO 159.3 ppm (STO, natural abundance))
were prepared as described previously (14b).
Isotopically labelled N-(2-adamantyl)hydroxydiazenium
oxide (3) and the corresponding azoxytosylates
An ice-cold solution of sodium nitrite (0.20 g) in water (10%
enriched with 18O, 0.5 cm3) was added dropwise over about
10 min to an ice-cold stirred solution of conc. hydrochloric
acid (37%, 0.20 cm3), N-(2-adamantyl)hydroxylammonium
chloride (170 mg), and water (10% enriched with 18O,
0.5 cm3) in ethanol (3 cm3). After about 20 min, the white
crystalline product was filtered at the pump, washed with cold
water, and dried under vacuum (143 mg; mp 134–135°C, lit.
(14b) (unlabelled material) 137–139°C; δH 1.60–2.10 (m,
14H), 4.29 (s, 1H), 12.1 (br, 1H)). The 17O-labelled analogue
(148 mg, mp 135–136°C) was made in the same way from
N-(2-adamantyl)hydroxylammonium chloride (167 mg) using
Attempted deoxygenations
(i) Four drops of triethyl phosphite (27, 28) were added to a
solution of compound 1 (10 mg) in toluene (0.5 cm3), and the
reaction was monitored by TLC for 24 h. The temperature was
then raised to 40°C but TLC showed only starting material
after several days.
1
10% 17O-enriched water and had an identical H NMR spec-
trum.
(ii) A solution of 1 (26 mg) in phosphorus trichloride (26,
29) (2 cm3) was stirred at room temperature under nitrogen and
protected from atmospheric moisture (SiO2 tube). No change
Specifically labelled 2-adamantyl azoxytosylate (1)-17O (δO
157.5 (STO), 291.6 (N-O-S), 452.4 ppm (N-O)) and (1)-18O
© 1998 NRC Canada

868
Can. J. Chem. Vol. 76, 1998
P. Buzek, P. von R. Schleyer, W. Koch, and J.W. de M. Carneiro.
J. Am. Chem. Soc. 115, 259 (1993).
was detected by TLC after 24 h whereupon the reaction was
worked up, and the isolated crystalline product was shown to
be starting material by H NMR and TLC. Further reactions
with acetyl bromide (26) thionyl chloride (26), triphenyl-
phosphine (28), tri(n-butyl)phosphine (28), and lithium wire
(30), were also attempted but gave negative or ambiguous re-
sults.
1
13. W.P Jencks. Chem. Soc. Rev. 10, 345 (1981); Chem. Rev. 85,
511 (1985); R.D. Guthrie and W.P. Jencks. Acc. Chem. Res. 22,
343 (1989); T.L. Amyes, I.W. Stevens, and J.P. Richard. J. Org.
Chem. 58, 6057 (1993); V. Jagannadham, T.L. Amyes, and J.P.
Richard. J. Chem. Soc. Perkin Trans. 2, 171 (1993).
14. (a) H. Maskill, P. Murray-Rust, J.T. Thompson, and A.A. Wil-
son. J. Chem. Soc. Chem. Commun. 788 (1980); (b) H. Maskill,
J.T. Thompson, and A.A. Wilson. J. Chem. Soc. Perkin Trans. 2,
1693 (1984).
Acknowledgement
We thank Dr. J.G. Smith for help with the millimeter-wave
spectroscopy, the ERASMUS Bureau for grants to J.H. and
M.P., the EPSRC for a grant to J.K.C., and Drs. J.S. Roberts
(Stirling University) and C.I.F Watt (Manchester University)
for helpful discussions.
15. J.K. Conner and H Maskill. Bull. Soc. Chim. Fr. 342 (1988).
16. I.M. Gordon and H Maskill. J. Chem. Soc. Chem. Commun.
1358 (1989).
17. H. Maskill, J.T. Thompson, and A.A. Wilson. J. Chem. Soc.
Chem. Commun. 1239 (1981).
18. E.H. White and D.J. Woodcock. In The chemistry of the amino
group. Edited by S. Patai. Interscience, New York. 1968.
pp. 407–497.
References
19. T.W. Bentley and P. von R. Schleyer. J. Am. Chem. Soc. 98,
7658 (1976).
20. V.J. Shiner, R.D. Fisher. and W Dowd. J. Am. Chem. Soc. 91,
7748 (1969).
21. D.N. Kevill and F.L. Weitl. Tetrahedron Lett. 707 (1971); D.N.
Kevill, J.B. Kyong, and F.L. Weitl. J. Org. Chem. 55, 4304
(1990).
22. R.S Tipson. J. Org. Chem. 9, 235 (1944).
23. H. Dahn and M-N. Ung-Truong. Helv. Chim. Acta, 71, 241
(1988).
24. J.H. Carpenter, J.D. Cooper, J.B. Simpson, J.G. Smith, and D.H.
Whiffen. J. Phys. E, 7, 678 (1974).
25. J. Haider, M.N.S. Hill, I.D. Menneer, H. Maskill, and J.G. Smith.
1. C.D. Nenitzescu. In Carbonium ions. Vol. 1. Edited by G.A. Olah
and P. von R. Schleyer. Wiley–Interscience, New York. 1968.
pp. 1–75.
2. (a) T.H. Lowry and K.S. Richardson. In Mechanism and theory
in organic chemistry. 3rd ed. Harper Collins, New York. 1987.
p. 329; (b) In Mechanism and theory in organic chemistry. 3rd
ed. Harper Collins, New York. 1987. pp. 335–349.
3. A.F. Diaz, I. Lazdins, and S. Winstein. J. Am. Chem. Soc. 90,
1904 (1968); S. Winstein, B. Appel, R. Baker, and A. Diaz. In
Organic reaction mechanisms. Special Publication No. 19. The
Chemical Society, London. 1965. pp. 109–130; V.J. Shiner and
W. Dowd. J. Am. Chem. Soc. 91, 6528 (1969); V.J. Shiner, D.A.
Nollen, and K. Humski. J. Org. Chem. 44, 2108 (1979); A.D.
Allen, T.T. Tidwell, and O.S. Tee. J. Am. Chem. Soc. 115,
10 091 (1993).
4. E. Grunwald and S. Winstein. J. Am. Chem. Soc. 70, 846 (1948).
5. S. Winstein, E. Grunwald, and H.W. Jones. J. Am. Chem. Soc.
73, 2700 (1951).
6. D.N. Kevill. In Advances in Quantitative structure-property re-
lationships. Vol. 1. Edited by M. Charton. Jai Press, Greenwich,
Conn., U.S.A. 1996. pp. 81–115; D.N. Kevill and M.J. D’Souza.
J. Chem. Soc. Perkin Trans. 2, 257 (1997).
J. Chem. Soc. Chem. Commun. 1571 (1997).
26. J.F. Vozza. J. Org. Chem. 34, 3219 (1969).
27. J.I.G. Cadogan, M. Cameron-Wood, R.K. Mackie, and R.J.G.
Searle. J. Chem. Soc. 4831 (1965); T. Mukaiyama, H. Nambu,
and M. Okamoto. J. Org. Chem. 27, 3651 (1962).
28. (a) J.I.G. Cadogan. In Organophosphorus reagents in organic
synthesis. Edited by J.I.G. Cadogan Academic Press, London
and New York. 1979. pp. 269–294; (b) A.G Rowley. In Organo-
phosphorus reagents in organic synthesis. Edited by J.I.G.
Cadogan Academic Press, London and New York. 1979.
pp. 295–350.
7. T.W. Bentley, F.L. Schadt, and P. von R. Schleyer. J. Am. Chem.
Soc. 94, 992 (1972); F.L. Schadt and P. von R. Schleyer. J. Am.
Chem. Soc. 95, 7860 (1973).
8. F.L. Schadt, T.W. Bentley, and P. von R. Schleyer. J. Am. Chem.
Soc. 98, 7667 (1976).
29. T.R. Emerson and C.W. Rees. J. Chem. Soc. 2319 (1964).
30. V. Kalyanaraman and M.V. George. J. Org. Chem. 38, 507
(1973).
31. V.J. Shiner. In Isotope effects in chemical reactions. Edited by
C.J. Collins and N.S. Bowman. A.C.S. Monograph 167. Van
Nostrand-Reinhold, New York. 1970. pp. 90–159; V.J. Shiner
and R.D. Fisher. J. Am. Chem. Soc. 93, 2553 (1971).
32. E.H. White, H.P. Tiwari, and M.J. Todd. J. Am. Chem. Soc. 90,
4734 (1968).
33. P. von R. Schleyer and R.D. Nicholas. J. Am. Chem. Soc. 83, 182
(1961).
34. H. Maskill and J.T. Thompson. Lab. Microcomput. 1, 11 (1982).
9. R.A. McClelland, V.M. Kanagasabapathy, N.S. Banait, and S.
Steenken. J. Am. Chem. Soc. 111, 3966 (1989); J. Am. Chem.
Soc. 114, 1816 (1992).
10. G.A. Olah, M.D. Heagy, and G.K.S. Prakash. J. Org. Chem. 58,
4851 (1993).
11. H. Mayr and M. Patz. Angew. Chem. Int. Ed. Engl. 33, 938
(1994).
12. S. Sieber, P. von R. Schleyer, H. Vancik, M. Mesic, and D.E.
Sunko. Angew. Chem. Int. Ed. Engl. 32, 1604 (1993); S. Sieber,
© 1998 NRC Canada