Proton transfer reactions of N-aryl triazolium salts: Unusual ortho-substituent effects

Article abstract of DOI:10.1002/poc.3399

Previous studies of the C(3)-hydrogen/deuterium exchange reactions of the triazolium ion conjugate acids of triazolyl N-heterocyclic carbenes revealed a change of mechanism under acidic conditions with N1-protonation to a dicationic salt. Interestingly, the data suggested an increase in pK_a^N1 in the presence of a N-pentafluorophenyl substituent relative to other N-aryl substituents with hydrogens or methyl substituents rather than fluorines at the ortho-positions. To probe the presence of an apparent donor effect of a N-pentafluorophenyl substituent, which differs from the more common electron withdrawing effect of this group, we have studied the analogous deuterium exchange reactions of four triazolium salts with heteroatoms or heteroatom substituents in the 2-position and/or 6-position of the N-aryl ring. These include triazolium salts with N-2,4,6-tribromophenyl 11, N-2,6-dichlorophenyl 12, N-2-pyridyl 13 and N-2-pyrimidinyl 14 substituents. The log k_ex - pD profiles for 11, 12 and 14 were found to show similar trends at lower pDs as for the previously studied N-pentafluorophenyl triazolium salt, hence supporting the presence an apparent donor effect on pK_a^N1. Surprisingly, the log k_ex - pD profile for N-pyridyl salt 13 uniquely showed acid catalysis at lower pDs. We propose herein that this data is best explained by invoking an intramolecular general base role for the N-(2-pyridyl) substituent in conjunction with N1-protonation on the triazolium ring. Finally, the second order rate constants for deuteroxide ion catalysed C(3)-H/D exchange (k_DO, M^-1 s^-1), which could be obtained from data at pDs 1.5, were used to provide estimates of C(3)-carbon acid pK_a^C3 values for the four triazolium salts 11-14.

Full text of DOI:10.1002/poc.3399

Special Issue Article
Received: 17 September 2014,
Revised: 23 October 2014,
Accepted: 24 October 2014,
Published online in Wiley Online Library: 7 December 2014
(wileyonlinelibrary.com) DOI: 10.1002/poc.3399
Proton transfer reactions of N-aryl triazolium
salts: unusual ortho-substituent effects†
David E. Tuckera, Peter Quinna, Richard S. Masseya, Christopher J. Collettb,
David J. Jasiewicza, Christopher R. Bramleya, Andrew D. Smithb**
and AnnMarie C. O’Donoghuea*
Previous studies of the C(3)-hydrogen/deuterium exchange reactions of the triazolium ion conjugate acids of triazolyl
N-heterocyclic carbenes revealed a change of mechanism under acidic conditions with N1-protonation to a dicationic
N1
salt. Interestingly, the data suggested an increase in pKa in the presence of a N-pentafluorophenyl substituent rela-
tive to other N-aryl substituents with hydrogens or methyl substituents rather than fluorines at the ortho-positions.
To probe the presence of an apparent donor effect of a N-pentafluorophenyl substituent, which differs from the more
common electron withdrawing effect of this group, we have studied the analogous deuterium exchange reactions of four
triazolium salts with heteroatoms or heteroatom substituents in the 2-position and/or 6-position of the N-aryl ring. These
include triazolium salts with N-2,4,6-tribromophenyl 11, N-2,6-dichlorophenyl 12, N-2-pyridyl 13 and N-2-pyrimidinyl 14
substituents. The log kex – pD profiles for 11, 12 and 14 were found to show similar trends at lower pDs as for the
N1
previously studied N-pentafluorophenyl triazolium salt, hence supporting the presence an apparent donor effect on pKa
.
Surprisingly, the log kex – pD profile for N-pyridyl salt 13 uniquely showed acid catalysis at lower pDs. We propose herein
that this data is best explained by invoking an intramolecular general base role for the N-(2-pyridyl) substituent in con-
junction with N1-protonation on the triazolium ring. Finally, the second order rate constants for deuteroxide ion catalysed
C(3)-H/D exchange (kDO, Mꢀ1 sꢀ1), which could be obtained from data at pDs >1.5, were used to provide estimates of
C3
C(3)-carbon acid pKa values for the four triazolium salts 11–14. Copyright © 2014 John Wiley & Sons, Ltd.
Keywords: deuterium exchange; N-heterocyclic carbene; organic catalysis; proton transfer; triazolium
dependence on deuteroxide ion and exchange via triazol-3-ylidenes
7 (path A, Scheme 1). At lower pD values, upward deviations were
INTRODUCTION
observed from the line of unit slope through the log kex – pD data
consistent with a change in mechanism for deuterium exchange.
Of the large series of triazolium salts studied in aqueous solu-
tion, the altered dependence of log kex values on pD under more
acidic conditions was most prevalent and occurred at signifi-
cantly higher pD values, for N-pentafluorophenyl triazolium salt
6a (Ar = C6F5).[21] For the structurally homologous series of
triazolium salts 6, the pD value for onset of the change in slope
Triazol-3-ylidenes 1 form a class of stable N-heterocyclic carbenes
(NHC) that are frequently utilized as efficient, selective organic cat-
alysts in a broad range of transformations.[1–7] Closely related to
other stable NHCs including imidazole-2-ylidenes 2, imidazolin-2-
ylidenes 3, thiazol-2-ylidenes 4 and trihydropyrimidin-2-ylidenes
5, these carbenes have seen application in diverse areas of chem-
istry in addition to organic catalysis.[8–17] Common to most applica-
tions of NHCs of this type is the in situ generation of the carbene
from the conjugate acid azolium salt precursor by use of an appro-
priate base. We and others have reported the kinetic acidities
towards hydroxide ion and estimates of the aqueous pKa values
of the conjugate acid precursors to NHCs 1–5.[18–23]
* Correspondence to: AnnMarie C. O’Donoghue, Department of Chemistry,
Durham University, South Road, Durham, DH1 3LE, UK.
E-mail: annmarie.odonoghue@durham.ac.uk
** Correspondence to: Andrew D. Smith, EaStCHEM, School of Chemistry, University
of St Andrews, North Haugh, St Andrews, Fife KY16 9ST, UK.
E-mail: ads10@st-andrews.ac.uk
†
This article is published in Journal of Physical Organic Chemistry as a special issue
on the International Symposium on Reactive Intermediates and Unusual Molecules
2014 on Physical Organic Chemistry by Robert Moss (Rutgers University USA) and
Anna Gudmundsdottir (University of Cinncinnati, USA).
In particular, we showed that the pD rate profiles for the
deuterium exchange reactions of triazolium salt precursors 6 to
triazol-3-ylidenes 7 reveal distinct differences from analogous data
for NHCs 2–5.[21] The presence of the additional ring nitrogen in
triazolium ions 6 allows for alternative deuterium exchange mech-
anisms under more acidic conditions. Deuteronation at N(1) can
occur to give dicationic triazolium ions 8, which are precursors to
monocationic NHCs 9. At higher pD values, log kex values were
observed to increase linearly with pD consistent with a first order
a D. E. Tucker, P. Quinn, R. S. Massey, D. J. Jasiewicz, C. R. Bramley, A. C.
O'Donoghue
Department of Chemistry, Durham University, South Road, Durham
DH1 3LE, UK
b C. J. Collett, A. D. Smith
EaStCHEM, School of Chemistry, University of St Andrews, North Haugh, St Andrews,
Fife KY16 9ST, UK
J. Phys. Org. Chem. 2015, 28 108–115
Copyright © 2014 John Wiley & Sons, Ltd.

PROTON TRANSFER REACTIONS OF N-ARYL TRIAZOLIUM SALTS
Scheme 2. Triazolium catalysis of the benzoin condensation
in the presence of ortho-heteroatom substituents on the N-aryl
ring of catalyst.[24] The common first step of both of these reactions
involves the reaction of an aryl aldehyde and triazolium salt in the
presence of base to give a deuteroxy aryl intermediate 10 (Scheme 2,
shown for benzoin condensation). We have observed that N-aryl
ortho-X-heteroatom substituents significantly increase rate and
equilibrium constants for formation of adducts 10 relative to para-
substituted and ortho-alkyl analogues. One possible explanation of
these observations is the presence of an O-D…X interaction in the
adduct similar to the N+-H…ortho-F interaction proposed earlier.
In order to further probe these ortho-substituent effects, we
have studied the deuterium exchange reactions of a number of
additional triazolium salts 11–14 with heteroatoms or hetero-
atom substituents in the 2-position and/or 6-position of the
N-aryl ring. The study of alternative triazolium salts with differ-
ent ortho-halo substituents on the N-aryl ring was a logical next
step. Ortho-Bromo and chloro-substituted substrates 11 and 12
were chosen for initial study because of the straightforward ac-
cess to the relevant precursor N-aryl hydrazines required for
their synthesis. N-2-Pyridyl and 2,6-pyrimidinyl salts 13 and
14 allow ortho-nitrogen atom substituent effects to be probed.
The data for these substrates also provides further evidence
that path C, and not path B, is accountable for the altered depen-
dence of log kex values on pD under more acidic conditions. In
addition, we observed that the N-pyridyl triazolium salt 13
displays a different N-aryl substituent effect to all other triazolium
substrates. Formal acid catalysis of deuterium exchange was
uniquely observed for N-pyridyltriazolium salt 13 under more
acidic conditions, providing evidence for a possible intramolecular
deprotonation reaction involving the pyridyl substituent.
Scheme 1. Potential mechanistic pathways for deuterium exchange at
C(3)-H of triazolium ions 6
decreased in the order 6a > 6b > 6c > 6d ~ 6e ~ 6f (Ar = C6F5 (a);
4-CN-Ph (b); 4-F-Ph (c); Ph (d) [Correction added on 12 January
2015, after first online publication: H (d) was corrected to Ph
(d)]; 2,4,6-(Me3)-Ph (e); 4-MeO-Ph (f)). Only in the case of
triazolium salt 6a was the initial decreased dependence on pD
further followed by a downward break at the lowest pDs studied.
Data for 6a could be fit by an equation describing either of two
kinetically equivalent mechanistic options (path
B or C,
Scheme 1), which both require protonation at N1 of the
triazolium ring. Path B involves a solvent promoted deuterium
exchange reaction of the cationic triazolium salt 6a with removal
of active species via N-protonation to a dicationic triazolium ion
8a. Path C entails no solvent reaction of the monocationic salt 6a
but instead the initial N-protonation at N(1) of salt 6a followed
by deuteroxide-catalysed exchange on the dicationic
triazolium salt 8a. Both of these mechanisms allow for the
observed continued decrease in kex, the first order rate con-
stants for deuterium exchange, with pD. On the basis of an anal-
ysis of the results of fitting through assumption of either
pathway, we suggested that path C was the more likely option.
In order to explain the differences in the observed kinetic data
for 6a relative to 6b–f, we postulated that the presence of an
ortho-heteroatom, for example, fluorine, could favour proton-
ation (or deuteronation) at N(1) resulting in an increased
prevalence of dicationic triazolium salts 8 in the normal pD range
and a higher pKa (N1).[21] Of the substrates studied, only the
N-pentafluorophenyltriazolium salt 6a had ortho-heteroatoms
rather than hydrogen or methyl groups on the N-aryl ring. A
higher pKa (N1) for salt 6a seems counter-intuitive on the basis of the
electron withdrawing, through-bond, inductive substituent effect of
fluorine, which would be expected to decrease basicity at N(1). How-
ever, protonation at N(1) may be favoured in order to suppress
unfavourable electrostatic interactions between this nitrogen and
spacially proximal N-aryl ortho-fluorines, or, as a result of stabilizing
through-space N+-H…ortho-F interactions in the N-protonated salt.
In organocatalytic applications, varying the triazolium N-aryl
substituent can have dramatic effects on yields and selectivities
of typical triazolylidene-mediated transformations. In related
mechanistic studies of triazolium catalysis of benzoin and Stetter
reactions, we have also observed unexpected substituent effects
EXPERIMENTAL
The syntheses of triazolium salts 11–14, the preparation of solutions,
the determination of pD and NMR methods are described in the
Supporting Information.
Kinetic measurements
The kinetic procedures for the measurement of rate constants for
deuterium exchange for triazolium salts 11–14 were identical to those
previously reported for the study of analogous salts 6a–f.[21] Because of
the lability of the triazolium salts towards C(3)-H/D exchange in unbuf-
fered D2O solvent, reactions were initiated by addition of a solution,
containing internal standard (tetramethylammonium deuteriosulfate)
and buffer or DCl, directly to the rigorously dried triazolium salt. The final
substrate and internal standard concentrations in the D2O reaction solu-
tions were 5 and 1 mM, respectively. Reaction solutions in NMR tubes
were incubated at 25 °C in a thermostated water bath. pD values were
J. Phys. Org. Chem. 2015, 28 108–115
Copyright © 2014 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/poc

D. E. TUCKER ET AL.
recorded at the beginning and end of each reaction and were found to
be constant within error ( 0.05). The progress of the C(3)-H/D deuterium
exchange reaction was followed by 1H-NMR spectroscopy during the
disappearance of 75–90% of the C(3)-H signal of each substrate. There
was no change in the integrated areas of signals of all other protons of
triazolium salts 11–13 during this period, and no appearance of new
signals, consistent with the absence of any parallel decomposition or
hydrolysis reactions under the reaction conditions. For N-pyrimidinyl
salt 14, a competing reaction was observed to occur to a small extent
accounting for ≤6% of total products. On the basis of the NMR signals
Representative NMR spectral overlays of the deuterium exchange reac-
tions, first order kinetic plots, tabulated kex data and log kex – pD rate pro-
files are included in the Supporting Information for each triazolium salt
11–14 (Figs. S1–S16, Tables S1–S4).
RESULTS AND DISCUSSION
The C(3)-H/D deuterium exchange reactions of triazolium salts
11–14 were performed in aqueous acetic acid buffer or DCl solu-
tions at a range of pD values and at constant ionic strength,
I = 1.0 (KCl). For these substrates, deuterium exchange was too
fast to monitor above pD 4 at 25 °C. Buffer catalysis of deuterium
exchange was found to be insignificant in all previous studies of
azolium ion conjugate acids of NHCs including representative
triazolium salts.[18,20,21] Hence, it was assumed that buffer cataly-
sis of exchange was not significant, and the observed pseudo
first order rate constants for exchange, kex (sꢀ1), for reactions
performed in acetic acid buffers were used directly on pD rate
profiles of deuterium exchange.
of the products formed, this reaction was presumed to be
a
nucleophilic aromatic substitution reaction of DOꢀ or D2O at C(2) of
the pyrimidinyl ring.
The observed pseudo first order rate constants for exchange of the
C(3)-proton for deuterium, kex (sꢀ1), were obtained from nonlinear least
square fitting of reaction progress against time to a first order exponential
decay function. Reaction progress was defined by values of f (s), the fraction
of remaining unexchanged substrate, which were calculated from Eqn (1),
where AC3H and Astd are the integrated areas of the singlet of the C(3)-H
of the triazolium salt and the broad triplet at 3.3 ppm because of the methyl
hydrogens of internal standard, tetramethylammonium deuterosulfate. In
the case of triazolium salt 14, small corrections were made for the parallel
SNAr reaction of substrate such that the corrected f (s) values represented
reaction because of deuterium exchange only.
As in the previous study of a large series of triazolium salts,
values of log kex for 11 and 12 (Fig. 1,
and ) increase with
pD in the region from pD = 0–4.5. Figure 1 shows the pD rate
profiles for these salts in comparison with those for two previ-
ously studied N-pentafluorophenyl and N-phenyl triazolium salts
6a and 6d, respectively. Figure 2 shows the pD rate profiles for
N-pyridyl and N-pyrimidinyl salts 13 and 14 also in comparison
with previous data for 6a and 6d, respectively. The data for
N-pyridyl triazolium salt 13 (Fig. 2, ) is distinctly different from
all other triazolium salts studied including N-pentafluorophenyl
substrate 6a. Values of log kex for 13 decrease with pD in the region
ðAC H=Astd
Þ
3
ðAC H=AstdÞt
fðsÞ ¼
(1)
3
0
Figure 1. pD rate profiles for the deuterium exchange reactions of the
C(3)-proton of triazolium salts 11 ( ) and 12 ( ) in D2O at 25 °C and
I = 1.0 (KCl). Also plotted for comparison are log kex – pD data taken
from R. S. Massey et al.[21] for the deuterium exchange reactions of
triazolium salts 6a ( ) and 6d ( ). The solid lines show the fits of
the data at ionic strength, I = 1.0, to Eqn (3) except for 6d[21] for which
data are fit to an equation allowing for only a first order dependence on
Figure 2. pD rate profiles for the deuterium exchange reactions of the
C(3)-proton of triazolium salts 13 ( ) and 14 (}) in D2O at 25 °C and
I = 1.0 (KCl). Also plotted for comparison are log kex – pD data taken
from R. S. Massey et al.[21] for the deuterium exchange reactions of
triazolium salts 6a ( ) and 6d ( ). The solid lines show the fits of the data
to Eqn (3) for 6a and 14 and Eqn (4) for 13. For 6d,[21] the data are fit to an
equation allowing for only a first order dependence on DOꢀ
DOꢀ. For triazolium salt 11 and 12, extra data points (
and ) were
obtained in 2 M DCl (shown as open symbols on the same profiles, as
for 6a additional data points[21] in 1.2 and 2 M DCl)
wileyonlinelibrary.com/journal/poc
Copyright © 2014 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2015, 28 108–115

PROTON TRANSFER REACTIONS OF N-ARYL TRIAZOLIUM SALTS
from pD = 0–1.3 and increase with pD in the region 1.3–4.5. By con-
trast, the log kex – pD profile for N-pyrimidinyl salt 14 (Fig. 2, }) is
comparable with those for 11 and 12 displaying an increase of rate
constants for exchange with pD in the whole region studied.
Table 1. Second order rate constants for deuteroxide-catalysed
hydrogen–deuterium exchange at C(3) (kDO, Mꢀ1 sꢀ1), carbon
acid pKaC3 and pKaN values in aqueous solution at 25 °C and ionic
strength, I=1.0 (KCl)
C3c
N
Salt
kDO (Mꢀ1 sꢀ1 a
)
pKa
Ka (M)d
pKNa
Deuterium exchange reactions of N-2,4,6-tribromophenyl 11 and
N-2,6-dichlorophenyl 12 triazolium tetrafluoroborates
11
12
13
14
6a
6d
4.29 ( 0.13) × 108a
2.71 ( 0.12) × 108a
1.05 ( 0.05) × 108a
1.07 ( 0.03) × 108a
6.82 ( 0.25) × 108b
6.82 ( 0.13) × 107b
16.7
16.9
17.3
17.3
16.5b
17.5b
1.1 ( 0.3)
1.3 ( 1.3)
0.9 ( 0.4)
2.3 ( 1.6)
ꢀ0.04
ꢀ0.11
0.05
ꢀ0.36
Deuterium exchange kinetic data for N-2,4,6-tribromophenyl-
and N-2,6-dichlorophenyl triazolium salts 11 and 12 show the
same dependencies on pD as observed for N-pentafluorophenyl
salt 6a. In particular, there is a marked change in the depen-
dence of log kex values on pD under more acidic conditions closely
similar to data for 6a and more significant than for 6b–f. At
pDs >1.5, values of log kex for 11 and 12 increase linearly
with pD, and the data may be fit by a line of unit slope
indicating a first order dependence on deuteroxide ion con-
centration in this region. This is consistent with a mechanism
involving deuteroxide-catalysed C3-H/D exchange of the
monocationic triazolium ion substrate (path A, Scheme 1).
At pDs <1.5, the dependencies of log kex on pD decrease,
and data points in this region deviate upwards from the line
of unit slope that fits the remaining data at higher pDs. There is
also the beginning of a further downward break at the lowest pD
values as observed previously for N-pentafluorophenyltriazolium
salt 6a. By comparison, the profile for unsubstituted N-phenyl salt
6d (Fig. 2, ) is essentially linear with slope unity for all data points
except at pDs <0.2.
1.5 ( 0.4)b ꢀ0.18b
e
e
aValues of kDO (Mꢀ1 sꢀ1) obtained by fitting log kex – pD data
to Eqn (2) or (3).
bTaken from R. S. Massey et al.[21]
C3
cpKa values obtained by application of Eqn (5).
dValues of KaN (M) obtained by fitting log kex – pD data to Eqn
(2) or (3).
eValue could not be determined because of limited data
points showing an altered dependence on pD as described
in the main text.
Table 2. Predicted rate constants for kD2O, kDO′ or k′ based
on kinetic fitting
As in our previous study for 6a, the data for salts 11 and 12 fit
well to a kinetic scheme allowing for the occurrence of either
path B or C at lower pDs in conjunction with path A at higher
pDs. The log kex – pD data fits well to Eqn (2) or (3) that allow
for paths A and B or paths A and C, respectively. In these
equations, kDO (Mꢀ1 sꢀ1) is the second order rate constant for
deprotonation of monocationic triazolium ion (cf. 6, Scheme 1)
by deuteroxide, Kw = 10ꢀ14.87 is the ion product of D2O at 25 °C,
Salt
kD2O (sꢀ1 a
)
k )
DO’ (Mꢀ1 sꢀ1 d
11
12
14
6a
13
2.0 × 10ꢀ5 ( 3.8 × 10ꢀ6
8.0 × 10ꢀ6 ( 6.6 × 10ꢀ6
4.7 × 10ꢀ6 ( 2.6 × 10ꢀ6
)
)
)
)
)
1.2 × 1010 ( 2.2 × 109)
5.7 × 109 ( 4.7 × 109)
5.9 × 109 ( 3.3 × 109)
3.3 × 1010 ( 2.0 × 109)b
6.1 × 10ꢀ5 ( 3.6 × 10ꢀ7 b
1.4 × 10ꢀ4 ( 4.3 × 10ꢀ5 c
aValues of kD2O (sꢀ1) obtained by fitting log kex – pD data to
Eqn (2).
γDO = 0.73 is the activity coefficient for deuteroxide ion under
our experimental conditions, KNa is the acidity constant for
ionization at N, kD2O (sꢀ1) is the first order rate constant for
deprotonation of monocationic triazolium ion at C(3) by sol-
vent D2O and kDO′ (Mꢀ1 sꢀ1) is the second order rate constant
for deprotonation of dicationic triazolium ion (cf. 8, Scheme 1)
by deuteroxide ion. Fitting to either equation yields identical
values for kDO and KaN (Table 1), whereas values for kD2O or
bTaken from R. S. Massey et al.[21]
cValue of k′ (sꢀ1) obtained by fitting log kex – pD data to Eqn (4).
dValues of kDO′ (Mꢀ1 sꢀ1) obtained by fitting log kex – pD data
to Eqn (3).
difference across this series is only 10-fold (Table 1). Similarly,
small N-aryl substituent effects were observed for the 20
triazolium salts previously studied with kDO values only varying
by a maximum of 37-fold across this large series. The order of
reactivity of triazolium salts corresponds to an increase in rate
constant for deprotonation at C(3) with more electron withdraw-
ing N-aryl substituents. The kDO value for 2,4,6-tribromophenyl
triazolium salt 11 is 2.6-fold higher than for 2,6-dichlorophenyl
analogue 12. Although a chloro substituent is more electron
withdrawing than bromo, the greater number of halogen atoms
in 11 results in a higher observed kinetic acidity towards
deuteroxide ion.
k
DO′ are obtained by fitting to Eqn (2) or (3), respectively
(Table 2). For comparison, previously reported data[21] for
N-pentafluorophenyl and phenyl salts 6a and 6d are also included
in Tables 1 and 2.
ꢀꢀ
ꢁ
ꢁ
2
4
3
5
ꢂ
ꢃ
kDOKw
γDO
KaN
10pD þ KNkD
2
a
O
log kex ¼ log
log kex ¼ log
(2)
(3)
KNa þ 10ꢀpD
ꢀꢀ
ꢁ
ꢁ
ꢀ
ꢁ
2
3
k′DOKw
γDO
kDOKw
The acidity constants for protonation at nitrogen, KaN, increase
in the order 11 < 12 < 6a and correspond to pKa values of
ꢀ0.04, ꢀ0.11 and ꢀ0.18, respectively (Table 1). The relatively
large fitting errors associated with these KaN values are expected
as only ~50% N-protonation has occurred at pD = 0. The values
for pKa are similar within error, and the data shows that there
is a small increase in the degree of N-protonation within the
normal pH range in the order 6a < 12 < 11. Importantly, as
KaN
10pD
þ
γDO
N
4
5
ꢂ
ꢃ
KNa þ 10ꢀpD
N
Values of the second order rate constants, kDO, for C(3)-
deprotonation by deuteroxide ion of triazolium salts 6a, 6d, 11
and 12 decrease in the order 6a > 11 > 12 > 6d, although the
J. Phys. Org. Chem. 2015, 28 108–115
Copyright © 2014 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/poc

D. E. TUCKER ET AL.
observed previously for 6a, the pKaN values for 11 and 12 must
be substantially higher than for N-phenyl salt 6d (Fig. 2) and
other N-aryl salts 6b–f. The pD profiles for 6b–f are linear
through most of the pD region studied, with only one to
three data points at the lowest pDs showing upward devia-
N
tion; thus, pKa values for 6b–f could not be obtained. It
N
can be presumed that the pKa values are significantly less
than zero for 6b–f with no substantial protonation at N1 in
the normal pD range.
The new data in Table 2 for triazolium salts 11 and 12
supports these conclusions. The kD2O values calculated for 11
and 12 using Eqn (2) are 85-fold to 213-fold larger than ob-
served for the true pD-independent solvent reactions of the
N-cyanomethyl thiazolium salt 15 (R1 = CN, R2 = H), whereas
there is a much smaller ring effect on kDO values (<10-fold),
again suggesting that path B does not occur for the former
triazolium ions. Estimates for kDO′ calculated for C(3)-deprotonation
of the dicationic conjugate acids of 11 and 12 (cf. 8 in
Scheme 2), for reaction via path C, are all at the diffusional
limit (Table 2). This is logical given that kDO values for the
less reactive monocationic salts are already as high as
~108 Mꢀ1 sꢀ1. Further, as will be seen in the succeeding texts,
the observed acid catalysis in the case of N-pyridyl salt 13 is
most logically accounted for in tandem with path C for the
other triazolium salts at lower pDs.
Notably, the profiles for N-4-cyanophenyl 6b and 2,6-
dichlorophenyl 12 substrates are almost superimposable at pDs
>1.5 yielding similar kDO values (Fig. S17); however, the altered
dependence on pD is greater for 12 under more acidic condi-
tions. As the electronic substituent effect on kDO is similar in both
cases, this supports the existence of an additional ortho-chloro
N
substituent effect to explain the observed differences in pKa
.
This N-aryl net donor effect on pKaN for 6a, 11 and 12 is opposite
to the normal inductive through-bond electron accepting
substituent effect of these substituents, which would favour an
increase in acidity. Instead, we observe an increase in the basicity
of N(1) for 6a, 11 and 12 relative to other N-aryl triazolium ions
6b–f. The only possible explanation of these substituent effects
N
on pKa lies in through-space interactions as proposed earlier,
as otherwise a normal inductive substituent effect would be
N
expected as observed on acidity at C(3). Although the pKa
values for 6a, 11 and 12 are similar within error, the highest
observed value is for triazolium salt 11 with the largest
ortho-bromo substituents, and it is possible to speculate that
this facilitates a donor interaction with N(1).
Deuterium exchange reactions of N-pyridyltriazolium 13 and
N-2,6-pyrimidinyl 14 triazolium tetrafluoroborates
As for triazolium salts 11–12, values of log kex for N-pyridyl salt
13 increase linearly at pDs >1.5 (Fig. 2, ), and the data may
be fit by a line of unit slope consistent with a first order depen-
dence on deuteroxide ion concentration in this region and
deuterium exchange via path A (Scheme 1). In contrast with data
for the other triazolium salts in Fig. 1, values of log kex for 13
decrease with pD in the region from pD = 0–1.3. This observed
acid catalysis of deuterium exchange requires protonation of
substrate and a subsequent pD-independent C(3)-deprotonation
of the resulting dicationic substrate.
The log kex – pD data for N-pyridyl salt 13 fits well to Eqn (4),
which allows for both deuteroxide-catalysed exchange on the
monocationic triazolium salt (path A, Scheme 1) and, addition-
ally, a pD-independent C(3)-deprotonation reaction on the
dicationic substrate. In Eqn (4), kDO, Kw and γDO are as defined
earlier in the text. The first order rate constant k′ (sꢀ1) refers to
pD-independent deprotonation of dicationic substrate, and
mechanistic options for this process are discussed in the
succeeding texts. In this case, either the pyridyl nitrogen or N1
of the triazolium ring could potentially be protonated to give a
dicationic species. The pKaN = 0.05 calculated for N-pyridyl
substrate 13 (Table 1) is similar to pKNa values for triazolium salts
6a, 11 and 12 for which additional protonation can only occur at
N(1) of the triazolium ring. However, pKaN = 0.05 may also be
consistent with the acidity constant for the pyridinium nitrogen.
Numerous solution studies establish the pKa of the N-protonated
pyridinium ion at ~5,[26] and this would be expected to sub-
stantially decrease in the presence of a monocationic triazolium
substituent.
Previously,[21] we suggested that path C rather than B was the
most likely mechanism under more acidic conditions to explain
the altered dependence of log kex values on pD. This was mainly
based on a comparison of rate constants for deuterium
exchange for triazolium ion 6a with analogous literature data
for thiazolium ions 15.[23,25] There is no additional site for pro-
tonation within the thiazolium ring of 15 unlike for the
triazolium analogue 6a. Washabaugh and Jencks did observe
the onset of a true pD-independent solvent reaction for four
thiazolium salts 15 in concentrated DCl solutions at pDs < < 0
yielding rate constants, kD2O, that range from 1.6 × 10ꢀ9 sꢀ1
up to 9.4 × 10ꢀ8 sꢀ1. In a comparison of data for triazolium
and thiazolium salts, we previously noted that the difference
in kDO values is significantly smaller than for kD2O values. As
an example, kDO = 4.67 × 107 Mꢀ1 sꢀ1 for the N-cyanomethyl
thiazolium salt 15 (R1 = CN, R2 = H), which is 14-fold lower than
kDO = 6.82 × 108 Mꢀ1 sꢀ1 for N-pentafluorophenyltriazolium salt
6a (Table 1), whereas there is a much greater 650-fold differ-
ence for the corresponding kD2O values [kD2O = 9.4 × 10ꢀ8 sꢀ1
for 15 (R1 = CN, R2 = H) versus kD2O = 6.1 × 10ꢀ5 sꢀ1 for 6a
(Table 2)]. It is difficult to explain a substantially larger ring
effect on the solvent compared with the deuteroxide-catalysed
deuterium exchange reactions, and we suggested that this
provides evidence that path B is not the correct mechanism
to account for our data at low pDs. Further, the N-protonated dica-
tionic substrate would be expected to be more acidic at C(3) than
the monocationic analogue, and it seems logical (because of the
requirement for N-protonation to explain the observed log
kex – pD data for 6a) that a deuteroxide reaction of the
dication would be more likely than a water reaction on the
monocation given the greater reactivity of both the substrate
and the base in the former case, albeit at very low concentra-
tions of DOꢀ.
ꢀꢀ
ꢁ
ꢁ
2
3
5
ꢂ
ꢃ
kDOKw
γDO
KaN
10pD þ k′ 10ꢀpD
4
log kex ¼ log
(4)
KNa þ 10ꢀpD
wileyonlinelibrary.com/journal/poc
Copyright © 2014 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2015, 28 108–115

PROTON TRANSFER REACTIONS OF N-ARYL TRIAZOLIUM SALTS
Possible mechanisms formally consistent with the observed
acid catalysis of deuterium exchange for N-pyridyl salt 13 at
lower pDs are shown in Scheme 3. In analysing these options,
a key consideration is whether these mechanisms can explain
why a pD-independent reaction of the N-protonated salt is
possible for 13 but has not been observed for any other
triazolium ion. Option D1 (Scheme 3) involves initial protonation
on N1 of the triazolium ring followed by pD-independent depro-
tonation by D2O without direct involvement of the pyridyl ring.
This mechanism may be discounted as acid catalysis of deute-
rium exchange has not been observed for any other triazolium
ion studied to date, and a remote pyridyl substituent would
not be expected to drastically increase the rate of deprotonation
of dicationic substrate by solvent, especially as kDO values for
exchange via path A (Scheme 1) are similar for 13 and other
triazolium salts studied.
Option D2 (Scheme 3) involves initial protonation on the
pyridyl nitrogen, rather than N1 of the triazolium ring, followed
by pD-independent deprotonation by water. Remote proton-
ation on the adjacent pyridyl ring would not significantly
increase the rate of deprotonation at C(3) by solvent to enable
competition with path C (Scheme 1). The latter would be
expected to be faster because of N1-protonation on the more
proximal triazole ring with subsequent C(3) deprotonation by
more basic deuteroxide ion. N-Protonation of the pyridyl substit-
uent could, in theory, increase the rate of deprotonation by wa-
ter at C(3) to compete with a water reaction of the monocationic
triazolium salt (path B, Scheme 1). However, the data for all other
triazolium salts requires protonation at N1, as the pD rate profiles
do not show pD-independence and log kex values continue to
decrease; hence, this latter option is self-contradictory. Proton-
ation of the more proximal triazolium ring should then also result
in a competing water reaction such as D1, and acid catalysis
would also be expected for the other triazolium salts. On this
basis, option D2 cannot account for the unique observation of
acid catalysis for 13 in competition with either path B or C for
the other triazolium salts.
the N-pyridyl salt only and also supports path C as the main
mechanism for deuterium exchange at lower pDs for the other
triazolium salts 6a, 11 and 12 rather than path B. In this mecha-
nism, observed acid catalysis can be explained by N-protonation
on the triazolium ring accompanied by intramolecular deproton-
ation at C(3) by the pyridyl nitrogen. This intramolecular depro-
tonation could be direct or may involve one or more solvent
molecules. As this option is not available to the other substrates
6a, 11 and 12, it could provide an explanation for the singular
occurrence of acid catalysis in the case of N-pyridyl substrate
13 only.
The log kex – pD profile for N-pyrimidinyl salt 14 (Fig. 2, }) is
comparable with those for 11 and 12 displaying an increase of
rate constants for exchange with pD in the whole region studied,
and no acid catalysis of deuterium exchange is observed at lower
pDs. This suggests that intramolecular catalysis, through C(3)
deprotonation by the adjacent pyrimidine ring, is not significant
in this case. Given the much decreased basicity of a simple
monocyclic pyrimidine relative to a pyridine nitrogen (pKas of
5.1 and 1.1 for N-protonated pyridinium and pyrimidinium ions,
respectively),[26] this is likely because intramolecular deproton-
ation by the less basic pyrimidine nitrogen cannot compete
with intermolecular deprotonation at C(3) by deuteroxide ion
under these conditions. The data for N-pyrimidinyl salt 14 fits
N
well to Eqn (2) or (3), and the resulting values for kDO, Ka and
kD2O or kDO′ are shown in Tables 1 and 2. Significantly, the ob-
served pKaN = ꢀ0.36 for the N-pyrimidinyl substrate 14 is very
similar to that observed for N-pyridyl substrate 13 (pKaN = 0.05)
and provides further evidence that protonation occurs on
the triazolium N(1) rather than the pyrimidine ring, which
would be expected to yield more different pKas. This lends
further support to reaction via path C as the dominant
mechanism for all salts, except N-pyridyl triazolium ion 13,
at lower pDs.
Estimation of carbon acid pKa values
The carbon acid pKa values for deprotonation at C(3) for
monocationic triazolium salts 11–14 may be determined using
Eqn (5), which is derived for Scheme 4.[18,20–22,27–31] In this
equation, kHO (Mꢀ1 sꢀ1) is the second order rate constant for
deprotonation at C(3) by hydroxide ion, which may be calculated
Another possibility is that a shared hydrogen bond forms
between N1 on the triazole and the pyridyl nitrogen thereby
accelerating the rate of deprotonation at C(3) by solvent (option
D3, Scheme 3). However, the geometry of this hydrogen bond is
not linear, and it is difficult to envisage why this would occur
only for a solvent deprotonation reaction and not for deproton-
ation by deuteroxide ion. A final option (D4, Scheme 3) is
compatible with the occurrence of acid catalysis in the case of
from the corresponding kDO value using a value of kHO
/
kDO = 2.4[32] for the secondary solvent isotope effect on the
basicity of HOꢀ in H2O versus DOꢀ in D2O. As discussed previ-
ously,[18,20,21] the absence of significant general base catalysis
of exchange provides evidence that the reverse protonation of
the triazol-3-ylidene 7 by water is equal or close to the limiting
rate constant for the physical process of dielectric relaxation of
solvent (kHOH ≤ kreorg = 1011 ꢀ1[33,34]).
s
kHOH
kHO
pKa ¼ pKw þ log
(5)
Scheme 3. Potential mechanistic options for pD-independent depro-
tonation of a dicationic N-deuteronated N-pyridyl triazolium salt
Scheme 4. Equilibrium for deprotonation by hydroxide of triazolium
ions 6 at C(3)
J. Phys. Org. Chem. 2015, 28 108–115
Copyright © 2014 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/poc

D. E. TUCKER ET AL.
The C(3)-H pKa values for triazolium salts 11–14 are similar and
range from 16.7 to 17.3. The values are comparable with those
estimated for related thiazolium ions (cf. 15), however, are
substantially lower than our previously published values for the
conjugate acids of imidazole-2-ylidenes 2, imidazolin-2-ylidenes
3 and trihydropyrimidin-2-ylidenes. This is due to the pres-
ence of the additional electron withdrawing ring nitrogen,
which increases the stability of the formally neutral NHC 7
relatively to the cationic conjugate acid 6. This large in-
plane electron withdrawing effect of nitrogen is relatively
common in heterocyclic systems. As mentioned earlier, the
pKa of N-protonated pyrimidine is four units lower than for the
pyridinium ion because of the additional ring nitrogen atom in
the former system.
result of stabilizing through-space N+-H…ortho-X interactions
in the N-protonated salt.
The present study also reveals unique substituent effects for
the N-(2-pyridyl)-triazolium system 13, which shows distinct acid
catalysis of deuterium exchange at lower pD values. We propose
that this data is best explained by invoking an intramolecular gen-
eral base role for the N-(2-pyridyl) substituent in conjunction with
N1-protonation on the triazolium ring. Future work will investigate
the effect of additional substitution by both donor and acceptor
groups on the 2-pyridyl N-aryl ring, as a means of further
exploring the proposed intramolecular proton transfer reaction.
In addition, the potential of ortho-amino groups of N-2-anilino
substituents will be probed as alternative intramolecular general
base catalysts.
The estimated kDO′ values for C(3)-deprotonation of N-protonated
6a, 11, 12 and 14 by deuteroxide ion (Table 2) are at the diffusional
limit. The new kDO′ data for 11, 12 and 14 are all safely within error
of typical bimolecular values for diffusion of small molecules in so-
lution (kd ~5× 109 Mꢀ1 sꢀ1). This provides confidence in a stepwise
rather than concerted mechanism for deuterium exchange at C(3)
of the dicationic salt, via a distinct monocationic NHC intermediate,
as shown in Scheme 1 (path C). To our knowledge, a monocationic
NHC 9, or a N1-alkylated analogue, has not been directly isolated
to date, and these additional results for 11, 12 and 14 provide
evidence for the transient formation of these species in aqueous
solution under acidic conditions.
Overall, these results highlight the varying roles of ortho-
heteroatoms in influencing the chemical behaviour of widely
used triazolium salts in organic catalysis.
Acknowledgements
We thank the EPSRC (D. E. T., R. S. and C. J. C.) and Durham
University (P. Q., D. J. J. and C. R. B.) for funding of this project.
REFERENCES
[1] A. T. Biju, N. Kuhl, F. Glorius, Acc. Chem. Res. 2011, 44, 1182.
[2] X. Bugaut, F. Glorius, Chem. Soc. Rev. 2012, 41, 3511.
[3] D. Enders, O. Niemeier, A. Henseler, Chem. Rev. 2007, 107, 5606.
[4] J. Mahatthananchai, J. W. Bode, Chem. Sci. 2012, 3, 192.
[5] N. Marion, S. Díez-González, S. P. Nolan, Angew. Chem. Int. Ed. 2007,
46, 2988.
[6] C. A. Rose, S. Gundala, C.-L. Fagan, J. F. Franz, S. J. Connon, K. Zeitler,
Chem. Sci. 2012, 3, 735.
[7] S. J. Ryan, L. Candish, D. W. Lupton, Chem. Soc. Rev. 2013, 42, 4906.
[8] R. W. Alder, M. E. Blake, L. Chaker, J. N. Harvey, F. Paolini, J. Schutz,
Ang. Chem. Int. Ed. 2004, 43, 5896.
[9] R. H. Crabtree, Coord. Chem. Rev. 2013, 257, 755.
[10] S. J. Hock, L.-A. Schaper, W. A. Herrmann, F. E. Kuehn, Chem. Soc. Rev.
2013, 42, 5073.
[11] W. Liu, R. Gust, Chem. Soc. Rev. 2013, 42, 755.
[12] O. R. Luca, R. H. Crabtree, Chem. Soc. Rev. 2013, 42, 1440.
[13] C. D. Martin, M. Soleilhavoup, G. Bertrand, Chem. Sci. 2013, 4, 3020.
[14] D. Martin, M. Melaimi, M. Soleilhavoup, G. Bertrand, Organometallics
2011, 30, 5304.
The C(3)-carbon acid pKa values for N(1)-protonated dicationic
triazolium salts (cf. 8, Scheme 1) would be predicted to be sub-
stantially lower than for monocationic triazolium ions 6. By
analogy with the first and second pKaNs for simple diprotonated
and monoprotonated nitrogen-containing heterocycles, the
decrease in pKa would be expected to be at least five units. It
is predicted that the reverse protonation of the monocationic
NHCs (cf. 9) will fall below the upper limiting rate constant for
protonation by solvent (kreorg = 1011
s
ꢀ1). Using an average
value of kDO′ = 5 × 109 Mꢀ1 sꢀ1 together with kHOH ≤ 1011 sꢀ1
in Eqn (5) yields an upper limit estimate of pKa ≤ 15.3 for
the C(3)-carbon acidity of a dicationic triazolium ions. These
pKas establish the conjugate acids
8 of monocationic
carbenes 9 as the most acidic of all NHC families studied to
date in aqueous solution.
[15] M. Melaimi, M. Soleilhavoup, G. Bertrand, Ang. Chem. Int. Ed. 2010,
49, 8810.
[16] B. M. Neilson, A. G. Tennyson, C. W. Bielawski, J. Phys. Org. Chem.
2012, 25, 531.
[17] V. P. Reddy, A. Vadapalli, E. Sinn, N. Hosmane, J. Organomet. Chem.
2013, 747, 43.
CONCLUSIONS
Our studies of the deuterium exchange reactions of ortho-
disubstituted triazolium salts 11–12 provide additional evidence
for formal 2-heteroatom donor effects on pKa values at N(1). In
particular, there is a marked change in the dependence of log
kex values on pD under more acidic conditions closely similar
to data for 6a and more significant than for 6b–f where the lat-
ter have only hydrogens or methyl groups rather than halogens
in the 2-position. Although the pD profiles for all the triazolium
ions 6a–f and 11–12 are very similar above pD 1.5 yielding kDO
values within ~10-fold of each other, the altered dependence on
pD under more acidic conditions is more dominant for 2-
halo-substituted ions 6a, 11 and 12 than for 6b–f. The calcu-
[18] T. L. Amyes, S. T. Diver, J. P. Richard, F. M. Rivas, K. Toth, J. Am. Chem.
Soc. 2004, 126, 4366.
[19] J. P. Guthrie, R. A. M. O’Ferrall, A. C. O’Donoghue, W. E. Waghorne, I.
Zrinski, J. Phys. Org. Chem. 2003, 16, 582.
[20] E. M. Higgins, J. A. Sherwood, A. G. Lindsay, J. Armstrong, R. S. Massey,
R. W. Alder, A. C. O’Donoghue, Chem. Commun. 2011, 47, 1559.
[21] R. S. Massey, C. J. Collett, A. G. Lindsay, A. D. Smith, A. C. O’Donoghue,
J. Am. Chem. Soc. 2012, 134, 20421.
[22] A. C. O’Donoghue, R. S. Massey, in: Contemporary Carbene Chemistry,
Vol. 7 (Eds: R. A. Moss, M. P. Doyle), Wiley, Hoboken, New Jersey,
2014, 75.
[23] M. W. Washabaugh, W. P. Jencks, Biochemistry-Us 1988, 27, 5044.
[24] C. J. Collett, R. S. Massey, O. R. Maguire, A. S. Batsanov, A. C.
O’Donoghue, A. D. Smith, Chem. Sci. 2013, 4, 1514.
[25] M. W. Washabaugh, W. P. Jencks, J. Am. Chem. Soc. 1989, 111, 683.
[26] W. P. Jencks, J. Regenstein, in: Handbook of Biochemistry and Molec-
ular Biology (Physical and Chemical Data) 3, Vol. 1 (Ed: G. D. Fasman),
CRC Press, Cleveland, OH, 1976, 305.
N
lated pKa values decrease in the order Br > Cl > F, although
the difference is small across the series. The donor effect on
pKaN, which favours protonation at N(1), may be to suppress
unfavourable electrostatic interactions between N(1) and
spacially proximal N-aryl ortho-heteroatoms or could be a
[27] T. L. Amyes, J. P. Richard, J. Am. Chem. Soc. 1992, 114, 10297.
[28] T. L. Amyes, J. P. Richard, J. Am. Chem. Soc. 1996, 118, 3129.
wileyonlinelibrary.com/journal/poc
Copyright © 2014 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2015, 28 108–115

PROTON TRANSFER REACTIONS OF N-ARYL TRIAZOLIUM SALTS
[29] R. J. Delley, S. Bandyopadhyay, M. A. Fox, C. Schliehe, D. R. W.
Hodgson, F. Hollfelder, A. J. Kirby, A. C. O’Donoghue, Org. Biomol.
Chem. 2012, 10, 590.
[34] U. Kaatze, R. Pottel, A. Schumacher, J. Phys. Chem. 1992, 96, 6017.
[30] J. P. Richard, G. Williams, J. L. Gao, J. Am. Chem. Soc. 1999, 121, 715.
[31] J. P. Richard, G. Williams, A. C. O’Donoghue, T. L. Amyes, J. Am. Chem.
Soc. 2002, 124, 2957.
[32] V. Gold, S. Grist, J. Chem. Soc. Perkin Trans. 2 1972, 89.
[33] U. Kaatze, J. Chem. Eng. Data 1989, 34, 371.
SUPPORTING INFORMATION
Additional supporting information may be found in the online
version of this article at the publisher’s website.
J. Phys. Org. Chem. 2015, 28 108–115
Copyright © 2014 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/poc

Copyright of Journal of Physical Organic Chemistry is the property of John Wiley & Sons,
Inc. and its content may not be copied or emailed to multiple sites or posted to a listserv
without the copyright holder's express written permission. However, users may print,
download, or email articles for individual use.