A. Berkessel, A. C. OꢀDonoghue et al.
When two indicators were employed, there is good agree-
ment within experimental error between the final pKa
values. As an alternative to the more reactive dimsyl anion,
we chose the hydroxide ion as a base to generate the conju-
gate base of HA in solution, since the former could poten-
tially undergo direct reaction with a number of our aryl sub-
strates. As a result, our DMSO solution contained up to
2 vol% of water. For acids 7a and 8, as representative ex-
amples of nitrogen and oxygen Brønsted acids, we repeated
the measurements in anhydrous DMSO using the dimsyl
anion as a base and found that the pKa values are very simi-
lar to, and within experimental error of, those determined
by our experimental method (Table 2). Furthermore, we de-
termined the pKa values for trifluoroacetic acid and saccha-
rin under our experimental conditions, again representative
examples of oxygen and nitrogen acids, and these are in ex-
cellent agreement with values determined by Bordwell[17] for
these compounds in anhydrous DMSO (see the Supporting
Information). These results demonstrate that the small per-
centage of water has an insignificant effect on our measure-
ments. This is unsurprising because all of the pKa values re-
corded in Table 2 are relatively low (ꢅ5.7), and the more
stabilized Brønsted bases are less dependent on hydrogen
bonding from a solvent for stabilization.
Ionic strength effects on our experimental data were
found to be negligible. pKa measurements for diphenylphos-
phoric acid 8 were repeated at constant ionic strength main-
tained using tetrabutylammonium bromide (see the Sup-
porting Information). Absorbance values at a given buffer
ratio of [HA]/[Aꢀ] agree closely (<5%) with and without
the addition of salt. For each pKa measurement, the concen-
tration of all species was maintained at ꢅ10 mm to disfavor
homoconjugation and ion association. Bordwell has reported
that the ion pairing of anions with metal cations is of little
or no importance in dilute millimolar solutions in DMSO,
except for strongly chelating ions such as that formed from
acetylacetone.[17,23,26,27] Decomposition of the chiral Brønsted
acids in Table 2, due to hydrolysis or other competing reac-
tions, does not occur over the timescale of the measure-
ments (see the Supporting Information).
In general, the pKa values for the chiral Brønsted acids in
Table 2 fall within narrow ranges: the BINOL phosphoric
acids 4 and the phosphoric amide 5 can all be found in the
range 2.4–4.2. The sulfonyl imides 6 (BINBAMs) and sulfu-
ryl imides 7 (JINGLEs) are stronger acids and have rather
similar pKa values in the range 1.7–1.9. The fluorinated
TADDOL analogues (TEFDDOLs) 3a and 3b were found
to have pKa values of 5.7 and 2.4, respectively. Their re-
markable dependence of acidity on structure is discussed in
more detail below.
Despite a large change in the steric and electronic nature
of the 3- and 3’-substituents of phosphoric acids 4, there is a
small variation in pKa values.[28] This proves that remote
functionality at the 3- and 3’-positions has only a small
effect on the thermodynamic acidity of the phosphoric acid
despite the observation of widely varying effects of these
substituents on the enantio- and diasteroselectivities of
many organic transformations. As the 3- and 3’-substituents
are not directly conjugated to the reaction centre, the small
variation in pKa within the series will be due to inductive
and steric effects. Overall, the pKa is observed to decrease
as the substituent becomes more inductively electron-with-
drawing (e.g., the bromo substituent in 4b) due to the
better stabilization of the anionic phosphate conjugate base.
The effect of addition of bulky 2-naphthyl substituents (as in
4g) at the 3- and 3’-positions is to increase the pKa by about
0.4 units relative to the parent acid 4a. This small increase
in pKa could be due to the slight disruption of DMSO solva-
tion of the phosphate conjugate base in the presence of
these large aryl substituents.
N-Triflyl phosphoric amide 5 has a pKa value of 0.9 units
lower than analogous phosphoric acid ent-4e. A phosphoric
amide would be expected to have a higher pKa than a corre-
sponding phosphoric acid by analogy with the difference in
pKas between carboxylic acids and amides (the pKas of
acetic acid and acetamide are 12.3 and 25.5 in DMSO, re-
spectively[17]). In this case, there is a compensating decrease
in pKa due to the strongly electron-withdrawing N-triflyl
substituent. Bordwell has also determined a pKa value of
9.7[17] for N-triflyl amide (CF3SO2NH2) in DMSO, which in-
dicates that the effect of the phosphoryl substituent in 5 is
to decrease the pKa by about 6.4 units.
Bis-sulfonyl and sulfuryl imides 6 and 7a have almost
identical pKa values within experimental error thus showing
that the BINOL oxygen has little influence on the pKa,
which is largely governed by the two sulfonyl substituents in
each case. Koppel has measured a pKa of 16.0 for benzene
sulfonamide in DMSO solution.[21a–c] This value may be com-
pared with the pKa value of 30.6 determined by Bordwell
for aniline,[17] and shows that the introduction of one sulfo-
nyl substituent decreases the pKa by 14.6 units. The intro-
duction of two sulfonyl substituents, as in 6, decreases the
pKa by a further 14.2 units, although this bisACTHNUGRTENUNG(sulfonyl)imide
has a naphthyl rather than a phenyl substituent.
As part of a comprehensive investigation of equilibrium
acidities in DMSO, Bordwell and co-workers have reported
the pKa values for aliphatic alcohols and analogous fluori-
nated derivatives. The pKa values are reported for methanol
(29.0),[25] 2,2,2-trifluoroethanol (23.5 (TFE)),[17] and per-
fluoro-tert-butanol (10.7 (PFTB)).[17] Here, we report a pKa
value of 17.2 for 1,1,1,3,3,3-hexafluoro-2-propanol (HFIP)
(see the Supporting Information). This means that the addi-
tion of one CF3 group decreases the pKa value of the ali-
phatic alcohol by about 6 units (methanol>TFE>HFIP>
PFTB). TEFDDOL 3a, a fluoroalcohol with two CF3
groups, which are positioned a to every hydroxyl group,
would be expected to have a pKa value in the range of
HFIP (17.2), but in fact has a pKa value of 5.77. This large
increase in acidity can be explained with help of the X-ray
crystal structure of 3a (Figure 2).
The X-ray structure[29] of 3a reveals an intramolecular hy-
drogen bond between the hydroxyl groups as well as an in-
termolecular hydrogen bond between two TEFDDOL mole-
cules (Figure 2). The intramolecular hydrogen bond enhan-
8526
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 8524 – 8528