Photochemical generation of the N-isopropylbenzonitrilium ion and unusually slow trapping by azide ion

Article abstract of DOI:10.1039/a807346e

The imidate ester 4-cyanophenyl N-isopropylbenzimidate undergoes efficient photoheterolysis in water producing the 4-cyanophenoxide ion and the N-isopropylbenzonitrilium ion. The latter is detected as a transient intermediate in flash photolysis experiments and has a relatively long lifetime (1/k_w) in water of 2.5 ms. The rate constants for hydroxide and azide are respectively 5.2 × 10⁶ and 3.9 × 10⁶ dm³ mol^-1 s^-1. This pattern is not seen with sp² hybridized cations, where k_az is normally significantly greater than k_OH. Further indications that azide ion is unexpectedly poor as a trap for this cation is seen in considering the k_az:k_w value in the context of ratios directly measured for sp² hybridized cations. For the latter a k_az:k_w of less than 10⁵ can be taken to mean that the reaction with azide is at (or certainly very close to) the diffusion limit (i.e. 5-10 × 10⁹ dm³ mol^-1 s^-1). The nitrilium ion 3 has k_az:k_w equal to only 10⁴ mol^-1 dm³ and yet its azide reaction is three orders of magnitude slower than diffusion.

Full text of DOI:10.1039/a807346e

View Article Online / Journal Homepage / Table of Contents for this issue
Photochemical generation of the N-isopropylbenzonitrilium ion and
unusually slow trapping by azide ion
Patrick H. Ruane,a Robert A. McClelland,*a A. Frank Hegartyb and Steen Steenkenc
a Department of Chemistry, University of Toronto, Toronto, Ontario M5S 3H6, Canada
b Department of Chemistry, University College Dublin, Belfield, Dublin 4, Ireland
c Max-Planck-Institut fur Strahlenchemie, D-45413 Mulheim, Germany
Received (in Cambridge) 21st September 1998, Accepted 10th November 1998
The imidate ester 4-cyanophenyl N-isopropylbenzimidate undergoes efficient photoheterolysis in water producing
the 4-cyanophenoxide ion and the N-isopropylbenzonitrilium ion. The latter is detected as a transient intermediate
in flash photolysis experiments and has a relatively long lifetime (1/kw) in water of 2.5 ms. The rate constants for
hydroxide and azide are respectively 5.2 × 106 and 3.9 × 106 dm3 molϪ1 sϪ1. This pattern is not seen with sp2
hybridized cations, where kaz is normally significantly greater than kOH. Further indications that azide ion is
unexpectedly poor as a trap for this cation is seen in considering the kaz :kw value in the context of ratios directly
measured for sp2 hybridized cations. For the latter a kaz :kw of less than 105 can be taken to mean that the reaction
with azide is at (or certainly very close to) the diffusion limit (i.e. 5–10 × 109 dm3 molϪ1 sϪ1). The nitrilium ion 3 has
kaz :kw equal to only 104 molϪ1 dm3 and yet its azide reaction is three orders of magnitude slower than diffusion.
Introduction
-
O
CN
O
CN
Nitrilium ions are a well established class of reactive inter-
mediate1–4 and can also be prepared as stable salts with weakly
nucleophilic counter-ions.5,6 Although their major resonance
contributor is 1a, their reaction with nucleophiles normally
occurs at the sp-hybridized carbon center. In other words they
react through form 1b. In this sense they resemble vinyl cations
and acylium ions.
hυ
C
+
4
(1)
CH(CH3)2
N–CH(CH3)2
H2O
+
C
N
2
3
azide ion well below the diffusion limit, so that the use of
‘clock’ methods for estimating rate constants for sp-hybridized
cations may be questionable. In particular our results suggest
that the lifetimes of imidinium ions ArN᎐Cϩ–NR may be con-
+
+
R
C
N
R'
R
C
N
R'
᎐
2
1a
1b
siderably longer than those calculated11 by ‘clock’ methods.
The quantitative reactivity patterns of sp2 hybridized carbo-
cations have been extensively investigated. Mayr and co-
workers have established a general relation for the reaction with
π nucleophiles in solvents such as dichloromethane and aceto-
nitrile.7 In aqueous and alcohol solvents, very stable carbo-
cations have been studied extensively by Ritchie,8 and less stable
examples using the technique of laser flash photolysis (LFP)9
or indirectly with the azide clock method.10 Considerably less is
known about the reactivity of sp-hybridized examples. There is
an expectation that they may be more reactive than sp2-
hybridized examples.11 This appears to be true for acylium ions,
since the p-(dimethylamino)benzoyl acylium ion, despite the
strongly stabilizing substituent, is estimated to have a lifetime in
water of only 0.1–1 ns.12 Vinyl cations, however, studied by LFP
in solvents such as 2,2,2-trifluoroethanol (TFE) and aceto-
nitrile, do not show a substantial difference from sp2-hybridized
analogs.13–18 For example, the lifetime of the 4-methoxyphenyl-
vinyl cation in TFE is very similar to that of the 4-methoxy-
phenethyl cation.18
In this paper we report a flash photolysis study of the N-iso-
propylbenzonitrilium ion 3. This cation was formed in aqueous
solutions upon irradiation of the imidate 2, through photo-
heterolysis of the C–O bond. The 4-cyanophenoxide anion is a
reasonable leaving group in the excited state, as suggested by
someearlyworkofZimmermanandSomasekhara,19 andapplied
in LFP studies of diarylmethyl and triarylmethyl systems.20,21
Despite the sp hybridized center, the nitrilium ion 3 is
relatively long-lived in water. We also find that it reacts with
Results
Ground state hydrolysis
The imidate 2 undergoes slow hydrolysis in aqueous solutions,
forming N-isopropylbenzamide and the 4-cyanophenoxide ion
in base and 4-cyanophenyl benzoate in acid. The isopropyl-
ammonium ion is presumably also formed in the latter case.
O
aq. NaOH
–
Ph
Ph
C
C
+
OAr
OAr
NiPr
NHiPr
Ph
C
(2)
O
aq. HClO4
+
2
+
iPrNH3
OAr
The rate–pH profile (Fig. 1) shows plateau regions in acid
and base with a change-over around pH 4–5.
Eqn. (3) can be fitted to the data, where kacid (6.6 × 10Ϫ4 sϪ1
)
is the plateau rate constant in acid, kbase (1.35 × 10Ϫ4 sϪ1) is the
plateau in base and Ka (2.15 × 10Ϫ5) is the acidity constant of
protonated 2.
kobs = (kacid[Hϩ] ϩ kbaseKa)/([Hϩ] ϩ Ka)
(3)
Experiments were also performed in dilute sodium hydrox-
ide solutions containing small amounts of sodium azide.
J. Chem. Soc., Perkin Trans. 2, 1999, 43–48
43

View Article Online
Table 1 Spectral data (aqueous solutions unless otherwise specified)
ε/dm3 molϪ1 cmϪ1
Compound
λmax/nm
λmax
at 280 nm
Imidate 2
242
247
274
226
244
2.7 × 104
1.9 × 104
2.5 × 104
1.0 × 104
1.2 × 104
9.7 × 102
5.4 × 102
2.3 × 104
1.2 × 102
2.9 × 103
4-Cyanophenol
4-Cyanophenoxide
N-Isopropylbenzamide
N-Ethylbenzonitriliuma
a As BF4Ϫ salt in 85% H2SO4.
Fig. 1 Variation of log kobs with pH for the hydrolysis of the imidate 2
in aqueous solution at 25 ЊC and ionic strength 1 mol dmϪ3 maintained
with sodium perchlorate.
Fig. 3 Absorbance change at 280 nm following 248 nm laser excitation
of the imidate 2 (5 × 10Ϫ4 mol dmϪ3) in 1 × 10Ϫ3 mol dmϪ3 NaOH. Inset
A shows the small first-order decay that is observed at longer times.
Inset B shows the dependence of the first-order rate constants on the
concentration of sodium hydroxide.
benzonitrilium ion, whose spectrum could be recorded in
concentrated sulfuric acid. The 4-cyanophenoxide ion stands
out as being different, with an absorbance centered at 270–285
nm.
Fig. 2 Dependence on sodium azide concentration of the yield of
N-isopropylbenzamide (squares) and tetrazole 5 (crosses) for the
ground state reaction of imidate 2 in 3 × 10Ϫ4 mol dmϪ3 NaOH at 25 ЊC
and ionic strength 1 mol dmϪ3 maintained with sodium perchlorate.
The flash photolysis experiments to be discussed in the next
section could only be carried out with detection at wavelengths
greater than 265 nm, since the strong absorbance of the precur-
sor imidate precluded observing at lower wavelengths. The
majority of the experiments were carried out with detection at
280 nm. As shown in Table 1, the imidate, amide and 4-cyano-
phenol are relatively weakly absorbing at this wavelength. This
is true also of the nitrilium ion, although there is a shoulder
in the spectrum of this species that gives it a slightly higher
extinction coefficient than the other three at 280 nm.
The observed rate constants are independent of azide con-
centration,† as they are of hydroxide concentration. The
4-cyanophenoxide ion continues to be formed in quantitative
yield, but the tetrazole 5 is also found as a product. This forms
at the expense of the N-isopropylbenzamide, with the relative
amounts of the two depending on azide concentration as shown
in Fig. 2. A feature of the data is that at high azide concen-
tration, the yields level out at around 90% tetrazole and 10%
amide.
Photochemistry
The imidate 2 (approximately 0.0001 mol dmϪ3) was irradiated
with 254 nm light in aqueous solutions of pH 5–13. The only
products that are observed by HPLC are 4-cyanophenol and
N-isopropylbenzamide. These are formed over 2–3 minutes,
under conditions where there is less than 3% of the ground state
hydrolysis. The chemical yield is 100%, and the quantum yield is
~20%.
N
N
Ph
N
N
iPr
5
Absorption spectra
Working in sodium hydroxide solutions, LFP experiments
reveal a prompt formation (Fig. 3) of a large absorbance cen-
tered at 275–280 nm, whose spectrum obtained immediately
after the laser pulse (not shown) closely resembles that of the
4-cyanophenoxide ion. The appearance of this species is not
surprising since the ion is a product of the photolysis. The
absorbance is obviously permanent in basic solutions where the
phenol exists in the anionic form. There is a more weakly
Table 1 summarizes the relevant spectral information. The
imidate 2 is strongly absorbing at 240–250 nm, with a spectrum
that overlaps that of 4-cyanophenol. The N-isopropyl-
benzamide also overlaps with these two, as does the N-ethyl-
† At lower pH (e.g. in phosphate buffer at pH 7), the rate constants do
increase with added azide ion.
44
J. Chem. Soc., Perkin Trans. 2, 1999, 43–48

View Article Online
Fig. 5 The points plot the ∆OD obtained by LFP at the completion of
the fast kinetic decay for the solution of Fig. 4. The line is the spectrum
of N-ethylbenzonitrilium BF4Ϫ in 85% H2SO4, drawn so that it overlaps
with the LFP spectrum at 290–300 nm.
Fig. 4 Absorbance change at 280 nm following 248 nm laser excitation
of imidate 2 (5 × 10Ϫ4 mol dmϪ3) in 0.01 mol dmϪ3 NaH2PO4–0.0002
mol dmϪ3 Na2HPO4 (pH = 5.3). The inset shows the absorbance change
at 280 nm obtained on lamp photolysis of the same solution.
absorbing transient species whose signal overlaps that of the
aryl oxide. This is seen on working at longer times, and applying
an offset so as to enlarge the absorbance change (Inset A to Fig.
3). This reveals a small first-order decay, with an absorbance
change of about 10% of the total signal. As shown in Inset B,
the rate constants for the decay depend on the concentration of
sodium hydroxide, with a second-order rate constant kOH of
5.2 0.2 × 106 dm3 molϪ1 sϪ1 (20 ЊC and ionic strength = 1.0
mol dmϪ3).
LFP experiments were also performed at pH 5–8 in phos-
phate buffer solutions. The prompt initial absorbance observed
in base is again present, but now undergoes a rapid first-order
decay (Fig. 4). The rate constants are strongly dependent on the
concentration of the phosphate buffer, so much so that it is not
possible to accurately extrapolate to zero buffer concentration.
Experiments at different buffer ratios show the relation given in
eqn. (4) with kp = (5.7 0.2) × 108 dm3 molϪ1 sϪ1 and kd = (1.5
Fig. 6 Conductivity change following 248 nm laser irradiation of a
solution of imidate 2 in 1 × 10Ϫ5 mol dmϪ3 HClO4.
2Ϫ
kfast decay = kp[H2PO4Ϫ] ϩ kd[HPO4
]
(4)
Sodium azide (0–400 µmol dmϪ3) accelerated the decay in a
linear fashion, with kaz = (3.9 0.3) × 106 dm3 molϪ1 sϪ1 (20 ЊC
and ionic strength = 1.0 mol dmϪ3).
Fig. 5 shows the spectrum obtained using ∆OD values
obtained by LFP at the end of the fast decay but before any
of the slow decay, along with the spectrum of the authentic
N-ethylbenzonitrilium ion.
Fig. 6 shows the results of an experiment involving conduct-
ivity detection, for a solution of the imidate in a very dilute
perchloric acid solution. The laser flash causes a conductivity
decrease, followed by an increase that returns the conductivity
to its original value. The decrease does not follow good first-
order behaviour. The increase on the other hand does, with
k = (3.9 0.1) × 102 sϪ1, the same rate constant obtained for the
slow absorbance change at zero ionic strength in neutral
solutions.
0.4) × 108 dm3 molϪ1 sϪ1 (20 ЊC and ionic strength = 0.01 mol
dmϪ3).
The pKa of 4-cyanophenol is 7.95, so that, depending on the
pH, there are small amounts of 4-cyanophenoxide present as a
final stable product. This ion obviously gives rise to a perman-
ent positive ∆OD at 280 nm. However, even at pH 5–5.5 where
the relative amount of 4-cyanophenoxide is small, the rapid
decay proceeds to a final ∆OD that is about 10–15% of the
initial one (see Fig. 4). This absorbance, moreover, is not pres-
ent in a spectrum recorded on a diode array spectrometer 10
seconds after laser irradiation. This led us to suspect that the
transient species seen in base was still present, now considerably
longer-lived since no decay could be observed even at the
longest time scale of the LFP apparatus (~100 µs). Indeed,
working with a lamp flash photolysis apparatus, a small first-
order decay could be observed (see Inset to Fig. 4). The rate
constants (kslow decay) were independent of pH in the region pH
5–8,‡ with a rate constant kw = (3.0 0.2) × 102 sϪ1 at ionic
strength 1.0 mol dmϪ3 and (4.0 0.3) × 102 sϪ1 at zero ionic
strength.
Discussion
Ground state hydrolysis
On the surface 2 is exhibiting the normal behaviour of imidate
esters,22–25 with the N-protonated form reacting with water or
hydroxide to form a tetrahedral intermediate 6 that preferen-
tially loses aryl oxide in base and isopropylamine in acid (see
top part of Scheme 1). In terms of the rate–pH profile the
‡ The rate constants were also independent of the phosphate buffer
concentration, provided the latter was kept below 0.01 mol dmϪ3
.
J. Chem. Soc., Perkin Trans. 2, 1999, 43–48
45

View Article Online
Table 2 Fraction of imidate reacting by SN1 ionization, and amide:
k1(H2O)
k2[ -OH]
OAr
C
O
Ar
O
Ar
tetrazole competition for reaction of nitrilium ion in aqueous base
Ph
C
Ph
2H+
C
Ph
NHiPr
+
(kw ϩ kOH[OHϪ]):kaz/mol dmϪ3
Ka
2
NiPr
6
OH
NiPr
H
ACID
-iPrNH2
[NaOH]
fion
Solvolysisa
Flash photolysisb
BASE
-ArOH
kion
0.0003
0.001
0.92
0.91
4.6 × 10Ϫ4
1.4 × 10Ϫ3
4.8 × 10Ϫ4
kw
-ArO-
OH
kOH[OH-]
1.42 × 10Ϫ3
O
O
Ph
C
Ph
C
Ph
C
a Determined from ground state solvolysis by fit of eqn. (6) to experi-
mental data. b Calculated from hydroxide ion concentration and
absolute values of kw, kOH and kaz measured by flash photolysis.
7
+
NiPr
NiPr
N3
NHiPr
OAr
Ph
C
3
N
N
N
Ph
C
Ph
-
kaz[N3 ]
to its conjugate acid. The term fion = kion/(kion ϩ k2Kw/Ka) is
defined as the fraction that proceeds via the former. Amide
is the only product of the tetrahedral intermediate pathway
(at high pH), and this is indicated by the second term in the
equation. The ionization pathway has a second competition,
involving nucleophilic trapping of the nitrilium ion by water
and hydroxide ion to produce amide or by azide ion to ultim-
ately give tetrazole. This competition is reflected in the term in
brackets containing the ratio (kw ϩ kOH[OHϪ]):kaz.
N
8
5
NiPr
iPr
Scheme 1
plateau region in acid represents the water addition at a pH
where the imidate is fully protonated. As the pH is increased
and the equilibrium shifts to neutral imidate, the rate becomes
first-order in Hϩ. The flat region at even higher pH occurs
where the hydroxide addition becomes important, with the
hydroxide dependency cancelled by the requirement for the
protonated imidate.
The experiments with sodium azide indicate however that the
situation at higher pH is more complicated. In particular, the
observation of the product 5 incorporating azide with no
increase in the rate constant suggests the occurrence of an SN1
mechanism in which the nitrilium ion 3 is formed in a rate-
determining ionization of 2, prior to product-determining
steps. This ion subsequently reacts with azide ion to form the
adduct 8, which would cyclize to 5.26 Reaction with water or
hydroxide ion gives the imidic acid 7, which tautomerizes to the
amide product.§
Eqn. (6) was fitted to the experimental data to provide the two
parameters fion and the ratio (kw ϩ kOH[OHϪ]):kaz. The results
for two hydroxide concentrations are given in Table 2. It can be
seen that the SN1 ionization accounts for slightly over 90% of
the reaction. Thus the pH independent rate constant at high pH
can be broken down into contributions from kion of 1.26 × 10Ϫ4
sϪ1 and from k2Kw/Ka of 9 × 10Ϫ6 sϪ1 (which means that k2 is
1.9 × 104 dm3 molϪ1 Ϫ1). That the tetrahedral intermediate
s
pathway does contribute is indicated by the yield of amide
levelling off at high contributions of sodium azide (Fig. 2). Thus
there must be some pathway that continues to produce amide
even under conditions where trapping of the nitrilium ion by
azide ion dominates over water and hydroxide trapping.
To the best of our knowledge, this is the first demonstration
that an imidate ester can hydrolyze via an SN1 reaction. Such an
ionization is however not surprising, since imidoyl chlorides are
well established to rapidly form nitrilium ions by loss of chlor-
ide ion.3,4 The 4-cyanophenoxide is obviously not as good a
leaving group. We have however previously observed that tri-
arylmethyl 4-cyanophenoxides also undergo a slow SN1 reac-
tion.21 Thus, the combination of the polar protic solvent water
and a very stable cation, the nitrilium ion in the present system,
can result in a situation where SN1 ionization of the 4-
cyanophenoxide can occur.
Thus, under basic conditions there are two kinetically equiva-
lent reactions leading to the same products, N-isopropylbenz-
amide and 4-cyanophenoxide. The kinetic expression for the
entire pH region is given by eqn. (5).
kobs = [k1(H2O)[Hϩ] ϩ (k2Kw/Ka ϩ kion)Ka]/([Hϩ] ϩ Ka) (5)
The constants k1(H2O) and k2 refer respectively to water and
hydroxide addition to the protonated imidate, with Ka the
acidity constant for this species. The constant kion is the rate
constant for the SN1 ionization, a process that occurs with the
neutral imidate. This equation has the same form as eqn. (3),
with kacid = k1(H2O), and kbase = k2Kw/Ka ϩ kion. Because of the
kinetic ambiguity, the latter cannot be separated into the con-
tributions from the two pathways, at least on the basis of the
rate–pH profile.
Mechanism of photochemical reaction
While the products of irradiation are obviously consistent with
a relatively efficient photoheterolysis (Scheme 2), other path-
Such a breakdown can however be carried out based on the
products in the presence of sodium azide (Fig. 2). Eqn. (6) is the
expression for the yield of the amide product.
OAr
+
hυ
Ph
C
Ph
kw
C
NiPr
+
ArO-
< 20 ns
NiPr
-
2-
kOH[OH-]
kp[H2PO4 ]
kd[HPO4
]
Percent amide =
-
{(kw ϩ kOH[OHϪ])/kaz}
{(kw ϩ kOH[OHϪ])/kaz} ϩ [N3
kaz[N3 ]
100fion
ͩ
ͪ
ϩ
Ϫ
]
ArOH
Scheme 2
100(1 Ϫ fion
)
(6)
ways could also be considered, such as one where irradiation
In base there is an initial competition involving ionization of
the neutral imidate or the pH equivalent hydroxide addition
results in a rapid C᎐N hydration.27
᎐
Consistent with the ionization mechanism however is the
observation that the 4-cyanophenoxide ion is formed very rap-
idly, within the 20 ns laser pulse in both basic and weakly acidic
solutions. In base, this ion is obviously stable and there is a large
permanent positive ∆OD at 280 nm. In acid and neutral solu-
tions, the aryloxide is protonated so that its absorbance decays.
A detailed study of this process was carried out in phosphate
buffers. The rate constants [eqn. (4)] measured in such solutions
§ A reviewer suggested that an SN1 reaction involving the O-protonated
intermediate seems more likely. This however is not consistent with the
kinetics which show unambiguously that the neutral imidate is the
species reacting by this mechanism. It can also be noted that at pH 6,
there is evidence that the product 5 forms by nucleophilic attack on the
protonated imidate. This is characterized by the formation of this
product concomitant with an increase in the rate constant.
46
J. Chem. Soc., Perkin Trans. 2, 1999, 43–48

View Article Online
represent an approach to equilibrium, so that there is a term
kp[H2PO4Ϫ] for protonation of the aryloxide and a term
kd[HPO4Ϫ2] for the reverse process, the deprotonation of the
4-cyanophenol. The ratio kp :kd should be equal to the ratio of
acidity constants Ka(H2PO4Ϫ):Ka(ArOH), and there is good
agreement — 3.8 for the former and 4.2 for the latter. It can also
be seen that the rate constants are typical of a normal proton
transfer that is nearly thermoneutral.28
Identification of nitrilium ion transient in flash photolysis
experiments
The N-isopropylbenzonitrilium ion is obviously the other
product of heterolysis. As seen with an authentic sample of the
N-ethyl derivative, N-alkylbenzonitrilium ions have spectra that
overlap that of the imidate 2. Thus, attempts to observe such
intermediates by flash photolysis must necessarily be carried
out at wavelengths where they are less strongly absorbing. A
small decay is observed in both acid and phosphate buffers at
280 nm, and there is good evidence that this represents the
reactions of the nitrilium ion. (i) The absorbance change in this
decay is around 10% of the signal due to the 4-cyanophenoxide.
This is approximately the ratio of extinction coefficients at 280
nm (Table 1). (ii) The transient spectrum observed at pH 5–5.5
after the decay of the phenoxide should be that of the nitrilium
ion. As shown in Fig. 5, this is true above 285 nm. The deviation
below 285 nm occurs because the precursor imidate is begin-
ning to absorb in these solutions so that the ∆OD measured by
LFP is not just that of the nitrilium ion, but rather is the differ-
ence between the OD of this species and the imidate. (iii) The
transient is quenched by the nucleophiles azide and hydroxide,
as is typical of cationic intermediates. (iv) Most importantly,
the trapping ratios (kw ϩ kOH[OHϪ]):kaz calculated with the
three absolute rate constants measured by flash photolysis are
in excellent agreement with values obtained in the ground state
solvolysis from product analysis (Table 2).
Fig. 7 Log kaz versus log (kaz :kw) for triarylmethyl cations (closed
squares, ref. 29), diarylmethyl cations (open squares, ref. 29), the xan-
thylium ion (X, ref. 30) and biarylnitrenium ions (closed circles, ref. 31).
ion reacts at the diffusion limit.10 A rate constant of kaz = 5 ×
109 dm3 molϪ1 sϪ1 that is typical of encounter control is taken
for kaz. The absolute kw is then calculated from the selectivity.
The ratio kaz :kw for the nitrilium ion in this work is 1.3 × 104
dm3 molϪ1, so that the azide-clock approach predicts a kw of
4 × 105 sϪ1. This is three orders of magnitude greater than the
actual rate constant directly measured by flash photolysis. This
obviously occurs because the reaction with azide ion is well
below the diffusion limit.
The unusually low reactivity of azide ion with the nitrilium
ion 3 is seen in the fact that it is actually slightly less reactive
than hydroxide ion. This is not typical of sp2 hybridized cations,
where kaz is normally significantly greater than kOH. The data in
Fig. 7 are also relevant. This plots log kaz versus log (kaz :kw) for
various types of sp2 hybridized carbocations — triarylmethyl
cations,29 diarylmethyl cations,29 the xanthylium ion30 and
biarylnitrenium ions31 (which react with azide ion as cyclo-
hexadienyl cations32). Within each of these series, the more
reactive ions all react with azide ion at a nearly independent
rate constant. The limiting kaz are typical of diffusion control,
falling in the range 5–10 × 109 dm3 molϪ1 sϪ1 slightly dependent
on cation structure. Selectivities kaz :kw obviously depend on
structure, but changes reflect only variations in kw. Ultimately,
for very stable cations, kaz becomes smaller than the diffusion
limit. It can been seen however that any cation that has
kaz :kw < 105 (or kw > 105 sϪ1) has kaz that is close to or at the
limit. On this basis, it was concluded that an experimentally
measured kaz :kw lower than 105 dm3 molϪ1 would mean that
the azide-clock approach would provide a reasonably good
estimate for kw.9 The point for the nitrilium ion 3 in Fig. 7
shows that this is clearly not true for this system. Despite
only a 104 azide:water ratio this cation reacts well below the
diffusion limit with azide ion. We therefore caution against
the application of the azide-clock method (or other clock
approaches) to estimating the reactivities of sp-hybridized carbo-
cations, until more is known about the absolute reactivities of
such ions.
Conductivity detection
Compelling evidence for these interpretations is provided by the
conductivity experiment shown in Fig. 6. The photoheterolysis
produces two organic ions, so that initially there would be a
small conductivity increase. This is not observed since the aryl-
oxide is rapidly protonated by the hydronium ion present, and
the removal of this highly conducting ion results in a conduc-
tivity decrease. The failure of this process to obey first-order
kinetics can be attributed to the fact that a very dilute concen-
tration of Hϩ was employed in order to avoid protonation of
the imidate. Thus the concentration of aryloxide produced by
the laser pulse is likely to be similar to that of the Hϩ.
A slow conductivity increase is then observed. This repre-
sents the release of Hϩ as the nitrilium ion reacts with water
(3 ϩ H2O→7 ϩ Hϩ). This occurs with a rate constant kw, and
there is excellent agreement between the value obtained with
conductivity detection, and the one following the absorbance
decay at 280 nm. The conductivity returns to its original level,
since the net result of the photolysis in this solution is the
production only of neutral species (2 ϩ water→amide ϩ 4-
cyanophenol).
Reactivity of N-isopropylbenzonitrilium ion
Experimental
This cation has a lifetime (1/k) in water alone of 2.5 ms, ex-
tending to 3.3 ms at ionic strength 1 mol dmϪ3. By carbocation
standards this is a relatively long-lived intermediate, despite the
ready access of nucleophiles to the sp-hybridized carbocation
center. This kinetic stability can undoubtedly be attributed
to the stabilization of the cation in its nitrilium resonance
contributor 1a.
N-Ethylbenzonitrilium
tetrafluoroborate,5
1-isopropyl-5-
phenyltetrazole,26 N-isopropylbenzamide33 and 4-cyanophenyl
benzoate34 were prepared by literature methods.
4-Cyanophenyl N-isopropylbenzimidate was prepared from
the imidoyl chloride (prepared from the amide by the method
of Ugi et al.)35 by reacting with 1.1 equivalents of the sodium
salt of 4-cyanophenol in acetonitrile for 16 hours. After
removal of the solvent, water and dichloromethane were added
and separated, the dichloromethane layer was dried with
The azide-clock method converts selectivity ratios (kaz :kw)
into absolute rate constants based on the assumption that azide
J. Chem. Soc., Perkin Trans. 2, 1999, 43–48
47

View Article Online
7 H. Mayr and M. Patz, Angew. Chem., Int. Ed. Engl., 1994, 33, 958.
MgSO4 and, after filtration, the solvent removed to give a white
solid that was recrystallized from petroleum ether. This com-
pound had mp 270 ЊC, δH (200 MHz, CDCl3, Me4Si) 1.21 (6H,
d), 4.02 (1H, sept), 6.97 (2H, d), 7.30–7.45 (3H, mult), 7.57 (2H,
d), 7.76 (2H, mult); m/z 264 (Mϩ, 5%), 146 (60), 104 (100);
calcd. for C17H16NO2, 264.1262; found, 264.1251.
8 C. D. Ritchie, Acc. Chem. Res., 1972, 5, 348; 1986, 64, 2239.
9 R. A. McClelland, Tetrahedron, 1996, 52, 6823.
10 J. Richard, Tetrahedron, 1995, 51, 1535.
11 K. N. Dalby and W. P. Jencks, J. Am. Chem. Soc., 1997, 119, 7271.
12 D. B. Song and W. P. Jencks, J. Am. Chem. Soc., 1987, 109, 3160;
1989, 111, 8470.
13 W. Schnabel, I. Naito, T. Kitamura, S. Kobayashi and H. Taniguchi,
Tetrahedron, 1980, 36, 3229.
14 S. Kobayashi, T. Kitamura, H. Taniguchi and W. Schnabel, Chem.
Lett., 1983, 1117; 1984, 2101.
15 F. I. M. Van Ginkel, R. J. Visser, C. A. G. O. Varma and G. Lodder,
J. Photochem., 1985, 30, 453.
16 S. Kobayashi, Q. Q. Zhu and W. Schnabel, Z. Naturforsch., Teil B,
1988, 43, 825.
17 Y. Chiang, R. Eliason, J. Jones, A. J. Kresge, K. L. Evans and R. D.
Gandour, Can. J. Chem., 1993, 71, 1964.
18 F. Cozens, R. A. McClelland and S. Steenken, Tetrahedron Lett.,
1992, 33, 173.
Laser flash photolysis experiments involved ca. 20 ns pulses
at 248 nm (60–120 mJ per pulse) from a Lumonics excimer laser
(KrF) emission. A pulsed Xenon lamp providing monitoring
light. After passing through a monochromator, the signal from
the photomultiplier tube was digitized and sent to a computer
for analysis.
Conventional flash photolysis experiments were performed
using an apparatus previously described,36 with the sample
being irradiated with a broad band flash lamp of ca. 100 µs
duration.
Product analyses were performed with a Waters HPLC sys-
tem using a C18 column with 2 cm3 minϪ1 of 50:50 acetonitrile–
water as the eluting solvent and 235 nm detection. Products
were identified by comparison of retention times with those of
authentic samples. The identity of the products, especially the
tetrazole, was verifed by recording the NMR spectrum of a
scaled-up solvolysis. Quantitative analyses were performed by
determining the response factor of the standard samples at 235
nm, and correcting the peak areas of the unknown solutions.
19 H. E. Zimmerman and S. Somasekhara, J. Am. Chem. Soc., 1963,
85, 922.
20 R. A. McClelland, V. M. Kanagasabapathy and S. Steenken, J. Am.
Chem. Soc., 1988, 110, 6913.
21 R. A. McClelland, V. M. Kanagasabapathy, N. Banait and
S. Steenken, J. Am. Chem. Soc., 1989, 111, 3966.
22 G. L. Schmir and B. A. Cunningham, J. Am. Chem. Soc., 1965, 87,
5692.
23 R. K. Chaturvedi and G. L. Schmir, J. Am. Chem. Soc., 1968, 90,
4413.
24 T. C. Pletcher, S. Koehler and E. H. Cordes, J. Am. Chem. Soc.,
1968, 90, 7072.
25 T. Okuyama, T. C. Pletcher, D. J. Sahn and G. L. Schmir, J. Am.
Chem. Soc., 1973, 95, 1253.
26 A. F. Hegarty, unpublished work.
27 F. Haley and K. Yates. J. Org. Chem., 1987, 52, 1817.
28 Y. Chiang, A. J. Kresge and J. F. Holzwarth, J. Chem. Soc., Chem.
Commun., 1982, 1203.
Acknowledgements
R. A. M. acknowledges the continued financial support of the
Natural Sciences and Engineering Research Council of
Canada.
29 R. A. McClelland, V. M. Kanagasabapathy, N. Banait and S.
Steenken, J. Am. Chem. Soc., 1991, 113, 1009.
30 R. A. McClelland, N. Banait and S. Steenken, J. Am. Chem. Soc.,
1989, 111, 2929.
31 D. Ren and R. A. McClelland, Can. J. Chem., 1998, 76, 78.
32 M. Novak, M. J. Kahley, E. Eiger, J. S. Helmick and H. E. Peters,
J. Am. Chem. Soc., 1993, 115, 9453.
33 R. F. C. Brown and I. D. Rae, Aus. J. Chem., 1964, 17, 447.
34 A. A. Humffray and J. J. Ryan, J. Chem. Soc., B, 1967, 468.
35 I. Ugi, F. Beck and U. Fetzer, Chem. Ber., 1962, 95, 126.
36 A. A. Allen, A. J. Kresge, N. P. Schepp and T. T. Tidwell, Can. J.
Chem., 1987, 65, 1719.
References
1 H. Ulrich, The Chemistry of Imidoyl Chlorides, Plenum, New York,
1968.
2 J. Morath and G. W. Stacey, in The Chemistry of the Carbon-
Nitrogen Double Bond, ed. S. Patai, Wiley (Interscience), New York,
1970, ch. 8, p. 327.
3 A. F. Hegarty, J. D. Cronin and F. L. Scott, J. Chem. Soc., Perkin
Trans. 2, 1975, 429.
4 R. Ta-Shma and Z. Rappoport, J. Am. Chem. Soc., 1976, 98, 8460;
1877, 99, 1845.
5 H. Meerwein, P. Laasch, R. Mersch and J. Spille, Chem. Ber., 1956,
89, 209.
6 S. Kabuss, Angew. Chem., Int. Ed. Engl., 1966, 5, 675.
Paper 8/07346E
48
J. Chem. Soc., Perkin Trans. 2, 1999, 43–48