Position-specific secondary deuterium isotope effects on basicity of pyridine

Article abstract of DOI:10.1021/ja105331g

Secondary isotope effects (IEs) on basicities of various deuterated pyridine isotopologues and of 2,6-lutidine-2,6-(CD₃)₂ have been accurately measured in aqueous solution by an NMR titration method applicable to a mixture. Deuteration at any position of pyridine increases the basicity, but the IE per deuterium is largest for substitution at the 3-position and smallest for the 2-position, which is closest to the site of N-protonation, smaller even than that for 2-CD₃ substitution. Computations at the B3LYP/cc-pVTZ level overestimate the magnitude of the measured IEs but largely reproduce the variability with isotopic position. Because the calculated IEs are based on changes in vibrational frequencies on N-protonation, the correspondence between calculated and experimental IEs implies that they arise from zero-point energies, rather than from inductive effects.

Full text of DOI:10.1021/ja105331g

Published on Web 08/11/2010
Position-Specific Secondary Deuterium Isotope Effects on
Basicity of Pyridine
Charles L. Perrin* and Phaneendrasai Karri
Department of Chemistry & Biochemistry, UniVersity of CaliforniasSan Diego, La Jolla,
California 92093-0358
Received June 17, 2010; E-mail: cperrin@ucsd.edu
Abstract: Secondary isotope effects (IEs) on basicities of various deuterated pyridine isotopologues and
of 2,6-lutidine-2,6-(CD ) have been accurately measured in aqueous solution by an NMR titration method
3 2
applicable to a mixture. Deuteration at any position of pyridine increases the basicity, but the IE per deuterium
is largest for substitution at the 3-position and smallest for the 2-position, which is closest to the site of
3
N-protonation, smaller even than that for 2-CD substitution. Computations at the B3LYP/cc-pVTZ level
overestimate the magnitude of the measured IEs but largely reproduce the variability with isotopic position.
Because the calculated IEs are based on changes in vibrational frequencies on N-protonation, the
correspondence between calculated and experimental IEs implies that they arise from zero-point energies,
rather than from inductive effects.
Introduction
Recently we have verified that secondary deuterium IEs
increase the basicity of a wide variety of amines. Similarly,
7
Isotope effects (IEs) are informative features of rates and
equilibria. In contrast to primary IEs, where a bond to the
acidities of carboxylic acids and phenols are decreased detect-
ably by deuterium substitution, even at remote positions. For
1
8
isotope is broken, in the case of secondary IEs that bond remains
intact. Kinetic IEs continue to provide valuable insight into
aliphatic acids the magnitudes of the IEs diminish as the
deuterium becomes more distant from the OH, as expected. Yet
the magnitudes of the IEs in both phenol and benzoic acid do
not decrease as the deuterium moves from ortho to meta to para.
The data are supported by ab initio computations that relate
the IEs to changes of vibrational frequencies and zero-point
energies. However, the experimental IEs are considerably
smaller than the calculated ones, a discrepancy that is likely to
be due to solvation. Except for the simple case of formate anion,
it is difficult to account for the IEs in terms of electron
delocalization that lowers vibrational frequencies. To probe the
generality of a nearly constant IE upon deuteration at ortho,
meta, and para positions, we now address secondary deuterium
IEs on the basicity of pyridines.
2
3
mechanisms of organic, organometallic, and enzyme-catalyzed
4
reactions. Equilibrium IEs provide subtle details regarding
5
conformational preferences and intermolecular interactions. The
6
topic of secondary equilibrium IEs has been reviewed recently.
(
1) Melander, L.; Saunders, W. H. Reaction Rates of Isotopic Molecules;
Wiley: New York, 1980. Carroll, F. A. PerspectiVes on Structure and
Mechanism in Organic Chemistry: Brooks/Cole: Pacific Grove, CA,
1
998; pp 349-365. Anslyn, E. V.; Dougherty, D. A. Modern Physical
Organic Chemistry; University Science Books: Sausalito, CA, 2006;
p 421ff.
(
2) Perrin, C. L.; Zhao, C. Org. Biomol. Chem. 2008, 6, 3349. West, J. D.;
Stafford, S. E.; Meyer, M. P. J. Am. Chem. Soc. 2008, 130, 7816.
Lopez del Amo, J. M.; Langer, U.; Torres, V.; Buntkowsky, G.; Vieth,
H.-M.; P e´ rez-Torralba, M.; Sanz, D.; Claramunt, R. M.; Elguero, J.;
Limbach, H.-H. J. Am. Chem. Soc. 2008, 130, 8620. Baciocchi, E.;
Bietti, M.; Lanzalunga, O.; Lapi, A.; Raponi, D. J. Org. Chem. 2010,
According to UV spectrophotometric measurements in aque-
H
ous buffers, pyridine-d
5
is more basic than pyridine, with K
a
/
D
9
K
a
equal to 1.048 ( 0.028, corresponding to a ∆pK
a
of 0.02.
However, a subsequent report, using the same method, gave a
∆pK of 0.11 ( 0.015 at 293 K. The substantial disagreement
75, 1378. Garver, J. M.; Fang, Y.; Eyet, N.; Villano, S. M.; Bierbaum,
10
V. M.; Westaway, K. C. J. Am. Chem. Soc. 2010, 132, 3808.
3) Deprez, N. R.; Sanford, M. S. J. Am. Chem. Soc. 2009, 131, 11234.
Horino, Y.; Yamamoto, T.; Ueda, K.; Kuroda, S.; Toste, F. D. J. Am.
Chem. Soc. 2009, 131, 2809. Zimmer-De Iuliis, M.; Morris, R. H.
J. Am. Chem. Soc. 2009, 131, 11263. Amiel-Levy, M.; Hoz, S. J. Am.
Chem. Soc. 2009, 131, 8280.
a
(
(
(
can be ascribed to difficulty in making the measurements, owing
to interference from the vibrational structure present on the UV
absorbances, which varies not only between pyridine and
pyridinium ion but also between isotopologues. More recently,
the gas-phase equilibrium constant for proton transfer from
4) Lau, S. T. B.; Tanner, M. E. J. Am. Chem. Soc. 2008, 130, 17593.
Munos, J. W.; Pu, X.; Mansoorabadi, S. O.; Kim, H. J.; Liu, H.-w.
J. Am. Chem. Soc. 2009, 131, 2048. Panay, A. J.; Fitzpatrick, P. F.
J. Am. Chem. Soc. 2010, 132, 5584. McCusker, K. P.; Klinman, J. P.
J. Am. Chem. Soc. 2010, 132, 5114.
pyridinium ion to pyridine-d was found to be 1.24 ( 0.06 at
5
5) Saunders, M.; Wolfsberg, M.; Anet, F. A. L.; Kronja, O. J. Am. Chem.
Soc. 2007, 129, 10276. Hayama, T.; Baldridge, K. K.; Wu, Y.-T.;
Linden, A.; Siegel, J. S. J. Am. Chem. Soc. 2008, 130, 1583. Dunitz,
J. D.; Ibberson, R. M. Angew. Chem., Int. Ed. 2008, 47, 4208.
Mugridge, J. S.; Bergman, R. G.; Raymond, K. N. J. Am. Chem. Soc.
(7) Perrin, C. L.; Ohta, B. K.; Kuperman, J. J. Am. Chem. Soc. 2003,
125, 15008. Perrin, C. L.; Ohta, B. K.; Kuperman, J.; Liberman, J.;
Erd e´ lyi, M. J. Am. Chem. Soc. 2005, 127, 9641. Perrin, C. L.; Dong,
Y. J. Am. Chem. Soc. 2008, 130, 11143.
(8) Perrin, C. L.; Dong, Y. J. Am. Chem. Soc. 2007, 129, 4490.
(9) Batts, B. D.; Spinner, E. J. Chem. Soc. B 1968, 789.
(10) Bellobono, I. R.; Beltrame, P. J. Chem. Soc. B 1969, 620.
2
010, 132, 1182.
(
6) Perrin, C. L. AdV. Phys. Org. Chem. 2010, 44, 123.
10.1021/ja105331g  2010 American Chemical Society
J. AM. CHEM. SOC. 2010, 132, 12145–12149 9 12145

A R T I C L E S
Perrin and Karri
1
1
4.0 M DCl/D
2
O were added and 13C NMR spectra were recorded
3
31 K, corresponding to a ∆pK
a
of 0.09. This agrees well
with the second value in aqueous solution and also with the IE
derived from vibrational frequencies calculated at the B3LYP/
after each addition, until there was no further change in chemical
shifts. The number of aliquots varied from 12 to 20. For each
deuterated pyridine isotopologue at least two carbons undergo
substantial changes in chemical shifts upon N-protonation and were
thus suitable as reporter nuclei. When necessary, signals of the
isotopologues were assigned from the stoichiometry or by adding
authentic undeuterated material. For carbons attached to D the
chemical shift of the central component of the CD triplet or of the
cc-pVTZ level. However, the agreement with the solution ∆pK
a
may be fortuitous.
We therefore have undertaken to resolve the disagreement
between the two measurements in aqueous solution. This topic
is of further interest because of the recent discovery of a low-
temperature polymorph of pyridine-d
5
, different from that of
Also, all the previous IE studies addressed only
, and the errors of (0.02 precluded measurement of
3
CD septet was used.
1
2
Instrumentation. All H NMR spectra and all 13C NMR spectra
1
pyridine-h
pyridine-d
5
.
of lutidine were recorded on a JEOL ECA500 spectrometer. All
5
1
3
the small contributions of individual deuteriums. It would be
informative to distinguish the separate contributions from the
various positions around the ring as well as to assess the IE of
C NMR spectra of pyridines were recorded on a Varian VX-500
spectrometer (125 MHz), using an XSense direct-detection cold
probe. A 7.0-µs, 45° excitation pulse, 128-800 acquisitions, 32K
data points zero-filled to 128 K, and standard broad-band 1
H
3
CD groups in 2,6-lutidine (2,6-dimethylpyridine).
decoupling were used, with the temperature regulated at 298 K.
Data Analysis. Chemical shifts at various pyridine positions
during the course of a titration of a pair of isotopologues were
NMR titration can measure relative basicities with great
1
3
precision. When successive small aliquots of acid are added
to a mixture of bases, the one that is more basic is preferentially
protonated. Its chemical shift thus moves ahead of that of the
H
D
analyzed according to eq 1. The resulting values of K
a
/K
a
were
H
D
converted to ∆pK () log10(K /K )) and averaged for each
a a
less basic one. The acidity constants K
a
and chemical shifts δ
isotopologue. Because precision varied with the reporter nucleus,
weighted averages were calculated (weighted inversely to the square
of the standard deviation), and associated errors are either errors
of the weighted average or standard errors of the mean, whichever
is larger.
0
+
can then be related through eq 1, where δ or δ is the shift for
the deprotonated or protonated form, respectively, as measured
at the beginning or end of the titration. Therefore a plot of the
0
+
quantity on the left vs (δ
H
- δ
H
)(δ
D
D
- δ ) should be linear
Computations. Ab initio density-functional calculations on
with a zero intercept and with a slope equal to the ratio of acidity
constants. Any reporter nucleus can be utilized for this purpose,
pyridine, pyridinium ion, and their 2,6-dimethyl derivatives were
16
performed at the B3LYP/cc-pVTZ level, as recommended, using
1
3
but C is most suitable because its chemical shifts are quite
sensitive to the state of pyridine protonation and because there
are several nuclei in the molecule that provide independent
17
Gaussian 03, Revision D.01. Harmonic vibrational frequencies
0 2 2 5 3 2 2 2
for d , 2,6-d , 3,5-d , 4-d, d , (CD ) , and anti-(CH D) isotopologues
were calculated at very tightly optimized geometries on an ultrafine
optimization grid. The contributions from molecular masses and
moments of inertia, or from the product of vibrational frequencies
H
D
a a
measurements of K /K .
1
8
+
H
0
D
H
a
D
a
0
H
+
D
via the Redlich-Teller product rule, are negligible and are
ignored. The double difference, ∆∆Σν, of sums of all the frequen-
cies of the four species was calculated according to eq 2. This
difference was then converted through zero-point energies to a
calculated ∆pK at 298 K. A separate ∆∆Σν for high-frequency
(
δ - δ )(δ - δ ) ) (K /K )(δ - δ )(δ - δ )
H
D
H
D
(
1)
We now report highly accurate position-specific secondary
deuterium IEs on the basicities of pyridine and 2,6-lutidine. We
have also modeled the IEs by ab initio computations. In all cases
deuterium substitution increases basicity, and we attribute the
IEs to changes of vibrational frequencies on N-protonation.
-
1
C-H vs C-D stretching modes (frequencies > 3000 cm or 2100
-1
cm , respectively) could also be obtained. For the pyridine
isotopologues vibrations can be distinguished as in-plane and out-
of-plane, according to the z components of the displacements, and
a ∆∆Σν due to all the in-plane vibrations was extracted.
Experimental Section
∆
∆ΣV ) (ΣVHPy
+
- ΣVHPydn+) - (ΣVPy - ΣVPydn
)
(2)
Preparation of Deuteropyridines. Except for commercially
5
available pyridine-d , each deuteropyridine or dideuteropyridine was
prepared from the corresponding bromopyridine or dibromopyridine
Results
1
4
with two portions of excess Zn dust in D
Lutidine-2,6-(CD was prepared by two successive exchanges in
O containing K
2 4 2
SO /D O at 90 °C. 2,6-
)
2
The intrinsic isotope shift due to deuteration is always to
lower frequency (upfield), except for C3,5 of lutidine-2,6-(CD ) .
3
1
5
D
2
2 3
CO at 180 °C, but in a Biotage Initiator
microwave oven at 15 bar. Incorporation of deuterium to >95%
3
2
13
C NMR chemical shifts of pyridines and pyridine isotopo-
1
was verified by H NMR.
logues at the beginning and end of titration are listed in Table
S1. Experimental data from NMR titrations of deuterated
pyridine isotopologues are listed in Table 1. For every titration
the correlation coefficient R for the linear fit to eq 1 was greater
than 0.99995, except for C4 of 2,6-lutidine, where R drops to
Preparation and Titration of NMR Samples. Pyridine samples
(
∼1:1 H/D) were prepared as 1.0 M solutions in D
dimethoxybenzene as the internal standard (δ 114.5) and with a
small aliquot of NaOD/D O to guarantee that the pyridine was fully
deprotonated. The 2,6-lutidine sample was prepared similarly, but
in 20% DMSO-d in D O, owing to a lower solubility. The samples
were then degassed under vacuum. Successive 5-µL aliquots of
2
O with 1,4-
2
0
.9998 because the intrinsic isotope shift of the distant deute-
6
2
riums is insufficient to resolve the isotopologues near the
beginning and end of the titration. Figure 1 shows not a typical
plot, but the second worst such plot, for C2,6 of pyridine-d
for which the slope is 1.0828 ( 0.0030 but for which R is only
.99995.
5
,
(
(
11) Mu n˜ oz-Caro, C.; Ni n˜ o, A.; D a´ valos, J. Z.; Quintanilla, E.; Abboud,
J. L. J. Phys. Chem. A 2003, 107, 6160.
12) Crawford, S.; Kirchner, M. T.; Bl a¨ ser, D.; Boese, R.; David, W. I. F.;
Dawson, A.; Gehrke, A.; Ibberson, R. M.; Marshall, W. G.; Parsons,
S.; Yamamuro, O. Angew. Chem., Int. Ed. 2009, 48, 755.
0
(
13) Perrin, C. L.; Fabian, M. A.; Armstrong, K. B. J. Org. Chem. 1994,
(16) Szafran, M.; Koput, J. J. Mol. Struct. 2001, 565-566, 439–448.
(17) Frisch, M. J.; et. al. Gaussian 03, revision D.01; Gaussian, Inc.:
Wallingford, CT, 2004.
5
9, 5246. Perrin, C. L.; Fabian, M. A. Anal. Chem. 1996, 68, 2127.
(
(
14) Huber, S.; Grassi, G.; Bauder, A. Mol. Phys. 2005, 103, 1395.
15) Kebede, N.; Pavlik, J. W. J. Heterocycl. Chem. 1997, 34, 685.
(18) Bigeleisen, J.; Mayer, M. G. J. Chem. Phys. 1947, 15, 261.
1
2146 J. AM. CHEM. SOC. 9 VOL. 132, NO. 34, 2010

Deuterium Isotope Effects on Basicity of Pyridine
A R T I C L E S
Table 1. Deuterium IEs on Basicity As Measured at Each Reporter Nucleus, Average IEs, and Average IE per D
13
H
a
D
a
Py-d
n
C
K
/K
average ∆pK
∆pK per D
2
2
3
3
3
3
3
4
4
,6-D
2
C2,6
C3,5
C2,6
C4
C2,6
C3,5
C4
C3,5
C4
C2,6
C3,5
C4
1.0139 ( 0.0016
1.0141 ( 0.0005
1.0258 ( 0.0010
1.0229 ( 0.0002
1.0523 ( 0.0010
1.0493 ( 0.0006
1.0487 ( 0.0002
1.0174 ( 0.0004
1.0172 ( 0.0002
1.0828 ( 0.0030
1.0825 ( 0.0023
1.0809 ( 0.0025
1.0703 ( 0.0005
1.0699 ( 0.0007
1.0660 ( 0.0063
1.0692 ( 0.0015
0.0060 ( 0.0002
0.0030 ( 0.0001
0.0104 ( 0.0006
0.0106 ( 0.0002
,6-D
-D
2
0.0104 ( 0.0006
0.0212 ( 0.0004
-D
,5-D
,5-D
,5-D
-D
2
2
2
0.0074 ( 0.0001
0.0343 ( 0.0006
0.0074 ( 0.0001
-D
D
D
D
5
5
5
s
2,6-(CD
2,6-(CD
2,6-(CD
2,6-(CD
3
3
3
3
)
2
)
2
)
2
)
2
C2,6
C3,5
C4
0.0294 ( 0.0010
0.0049 ( 0.0002
CH
3
Table 2. Calculated Secondary Isotope Effects on Basicities of Deuterated Pyridines
a
n
∆∆Σν, cm-1
∆∆Σν, cm-1
b
∆∆Σν, cm-1
c
Py-d
n
ΣνPydn
ΣνHPydn+
∆pK
Py
2
-
2
3
38 900
35 998
36 043
37 467
31 704
62 719
54 367
59 926
45 168
42 243
42 261
43 717
37 881
68 951
60 562
66 137
0
0
0
0
,6-d
2
22.8
49.4
18.1
90.3
0
0.024
0.052
0.019
0.095
0
28.4
13.9
5.1
47.3
0
26.2
31.5
12.6
70.5
s
3
4
d
2
2
2
,5-d
-d
2
4
5
2, 3, 4
s
3
3
,6-(CH
,6-(CD
,6-(CH
3
3
2
)
)
D)
2
2
37.2
21.0
0.039
0.022
17.0
19.0
s
s
d
2
a
b
c
d
Number of bonds between D and N. C-H/D stretches. In-plane vibrations. Anti D.
d
2
, -3,5-d
2
, and -4-d is 0.0347 ( 0.0005, which agrees very
. Both of these
well with 0.0343 ( 0.0006 of pyridine-d
5
comparisons show that the IEs at the various positions are
additive.
Computational results on deuterated pyridine isotopologues
are listed in Table 2. The values for the total zero-point energies
5
for pyridine, pyridinium ion, and their d isotopologues differ
1
1
-1
from the ones calculated previously by no more than 4 cm .
A separate ∆∆Σν for high-frequency C-H vs C-D stretching
modes is also included, as well as ∆∆Σν from in-plane
vibrations for pyridine isotopologues.
Discussion
The data show that deuteration at any position increases the
basicity of pyridines. This is consistent with the general
phenomenon seen in many previous studies of secondary
deuterium IEs on acidity and basicity.
According to Table 1, the ∆pK
0.0343 ( 0.0006. This result settles the disagreement between
the two previously measured values of 0.02 and 0.11.
a 5
value for pyridine-d is
Figure 1. Linearized plot (eq 1) for C2,6 signals in NMR titration of
pyridine plus pyridine-d .
5
9
,10
Both
The data in Table S1 verify that the chemical shifts of C3
and C4 move to higher frequency (downfield) on N-protonation,
whereas those of C2 and 2-CH move to lower frequency,
3
are wrong, but the former one is closer to the correct value,
probably fortuitously, because both measurements were subject
to severe interference from vibrational fine structure in the UV
spectra. Certainly our result is much more reliable, inasmuch
as even the worst entries in Table 1 are more precise than the
despite the depletion of electron density. This reversal does not
spoil the applicability of eq 1. However, the sensitivity of
chemical shift to the state of protonation is greatest for C4 and
least for C3,5. Consequently the precision of the IE generally
increases from C3,5 to C2,6 to C4, except when the reporter
nucleus was a deuterated carbon, for which the signal was weak.
The data are internally consistent. The IE measured for each
isotopologue is the same, within a small but variable experi-
mental error, regardless of reporter nucleus. The IE of pyridine-
a
(0.02 ∆pK of those previous studies.
Furthermore, the data in Table 1 show that we can separate
the contributions from each of the positions around the pyridine
ring. For the isotopologues of pyridine the IE per deuterium is
2 2
largest for pyridine-3,5-d and smallest for pyridine-2,6-d , even
though in the latter the deuterium substitution is closest to the
site of N-protonation. The IE per deuterium in 2,6-lutidine is
also small, even though the methyl groups are again quite close
to the site of N-protonation. It is larger though than the IE in
3
,5-d
2
is double that of pyridine-3-d, or equivalently, the IE
per D is the same. The sum of the ∆pK values of pyridine-2,6-
J. AM. CHEM. SOC. 9 VOL. 132, NO. 34, 2010 12147

A R T I C L E S
Perrin and Karri
pyridine-2,6-d
2
. These are unexpected results, and the observa-
A higher gas-phase IE, corresponding to a ∆pK
observed and also calculated for pyridine-d ,
5
a
of 0.09, as
further supports
a substantial role for solvation in reducing the IE. That IE
1
1
tion that the largest IE is from deuteration at C3,5 is quite
surprising.
1
0
The calculations in Table 2 corroborate the experimental
observations that deuteration at any position increases the
basicity of pyridines. Because the calculated IEs are based on
changes in vibrational frequencies on N-protonation, the cor-
respondence between calculated and experimental implies that
the IEs arise from zero-point energies. This excludes the
possibility that the major contributor to the IEs is from inductive
effects arising from anharmonicity, which makes a C-H bond
coincides with the 0.11 obtained earlier in aqueous solution,
but the agreement must be fortuitous because that aqueous value
is wrong.
It is difficult to assign the IEs to specific isotope-sensitive
vibrations. Because of variations in coupling of the individual
bond vibrations into normal modes, a frequency in pyridine
n
cannot be correlated with one in pyridine-d or in a pyridinium
ion. Prime candidates though are the C-H stretching modes,
whose vibrational frequencies and zero-point energies increase
on N-protonation. According to the calculations in Table 2,
stretching modes are indeed largely responsible for the small
1
9
longer than a C-D bond and changes the dipole moment.
Such an origin was suggested for pyridine and some substituted
9
pyridines, and it cannot be rejected as a minor contributor, but
it is not the dominant contributor. This is the same conclusion
that was reached for secondary deuterium IEs on the basicity
of amines, where the IE could be shown to be of stereoelectronic
origin, reaching a maximum when the deuterium is antiperipla-
IE in pyridine-2,6-d
2
and for the IE of the anti D in lutidine-
2,6-(CH D) , but not for the other, larger IEs.
2
2
Another separation of vibrational modes in pyridine is
between in-plane and out-of-plane. According to the results in
Table 2, the in-plane modes contribute nearly 80% of the IE in
7
nar to the nitrogen lone pair.
The values in Table 2 separate the contributions from each
of the positions around the pyridine ring. The calculated IEs,
pyridine-d
consistent with the previous conclusion regarding stretching
modes of pyridine-2,6-d , which are necessarily in-plane. Less
than 20% of the overall IE in pyridine-d comes from the out-
of-plane bending modes, which represent 30% of the total
modes.
5 2
and all of the small IE in pyridine-2,6-d . This is
per deuterium, decrease in the order pyridine-3,5-d
2
> pyridine-
2
4
-d > lutidine-2,6-(CD > pyridine-2,6-d , which is the same
3
)
2
2
5
as the order obtained experimentally, except for a reversal of
the two lowest. Therefore we can be confident that the IE does
not fall off with distance from the site of N-protonation, but
that the smallest IEs are seen at the positions nearest the
nitrogen.
5
Alternatively, the IE in pyridine-d has been associated with
a calculated shortening of the C-H distance and a strengthening
1
1
of the C-H bond on N-protonation. However, the calculated
shortening decreases in the order C2-H > C3-H > C4-H,
whereas the IEs do not vary monotonically with distance. Instead
In both 2,6-lutidine and 2,6-lutidinium ion the lowest-energy
conformation of each methyl is calculated to have one hydrogen
antiperiplanar to the nitrogen and two hydrogens syn (1a,1b).
The calculated ∆pK due to deuteriums in those two anti
positions is nearly half of that due to six deuteriums. Therefore
the anti deuteriums contribute a larger IE, per deuterium, than
do syn deuteriums. Of course, this feature of the calculation
could not be verified experimentally.
2
the largest IE is in pyridine-3,5-d . Likewise, the redistribution
of electron density does not account for the IEs, inasmuch as
H2,6 are the hydrogens calculated to acquire the greatest positive
charge on N-protonation, and they show the smallest IE.
It is tempting to try to assign the IEs to n-σ* delocalization,
which weakens C-H bonds and lowers zero-point energies, as
7
,8
can be recognized in amines and the formate anion. Delo-
calization of the nitrogen lone pair into σ*CH, as in 2a and 2c,
would lower the stretching frequency and zero-point energy of
the C2-H and C4-H bonds in the neutral pyridine. This
delocalization would lead to a greater IE for deuteration at C2
and C4, where the observed IEs are smaller. In contrast, for the
3
C3-H and anti-2-CH bonds the lone pair is delocalized instead
Comparison of Tables 1 and 2 shows that the calculated IEs
are larger than the experimental ones, by an average factor of
into σCH, as in 2b and 2d. This delocalization would increase
the stretching frequency and zero-point energy of these bonds
in the neutral pyridine and would lead to an inverse IE, which
is not seen. This may account though for why the IE of the
2
.5. This overestimate arises because hydrogen bonding by
solvent to the nitrogen lone pair makes the pyridine resemble
the pyridinium ion, thereby reducing the magnitude of the IE.
A similar discrepancy was obtained for acidities of carboxylic
2
-CH
3
is so small. It does not account for the surprising
shows the largest IE, little of
8
observation that pyridine-3,5-d
2
acids and phenols.
which arises from stretching modes that would be strengthened
by delocalization as in 2b. Therefore it seems futile to associate
the IEs with specific features of bonding, structure, or vibrational
modes.
Computations that take solvation into account would be more
realistic. Rather than attempt them on pyridines, we carried out
B3LYP/6-31 calculations on formic acid, both without solvation
8
(
as previously) and with the polarized continuum model of
water. According to the vibrational frequencies in Table S8,
the double difference, ∆∆Σν, of sums of all the frequencies
(
eq 2) and thus the calculated secondary deuterium IE on the
acidity of formic acid are reduced by 50% in this solvent. This
is an imperfect model, in that continuum solvation ignores the
specific hydrogen-bonding feature of water, but it does support
a strong role for solvation to reduce the IE.
Conclusions
We have used an NMR titration method to accurately
measure secondary IEs on the basicities in aqueous solution
(
19) Klein, H. S.; Streitwieser, A. Chem. Ind. 1961, 180. Halevi, E. A.
Prog. Phys. Org. Chem. 1963, 1, 109.
1
2148 J. AM. CHEM. SOC. 9 VOL. 132, NO. 34, 2010

Deuterium Isotope Effects on Basicity of Pyridine
A R T I C L E S
of various deuterated pyridine isotopologues and of 2,6-
lutidine-2,6-(CD ) . The data show that deuteration at any
3 2
position increases the basicity. The IE per deuterium is larg-
est for substitution at the 3-position and smallest for the
Acknowledgment. This research was supported by NSF Grant
CHE07-42801. Purchase of the NMR spectrometers was made
possible by NSF Grant CHE-0741968.
Supporting Information Available: Complete ref 17. Table
2
-position, which is closest to the site of N-protonation,
13
n
S1 ( C chemical shifts of pyridines and pyridines-d at
smaller even than that for 2-CD substitution. Computations
3
beginning and end of titration). Tables S2-S5 (calculated
energies, optimized geometries, atomic charges, and interatomic
distances of all structures). Tables S6-S7 (calculated vibrational
frequencies for pyridine, pyridinium ion, 2,6-lutidine, and 2,6-
lutidinium ion and their deuterated isotopologues). Table S8
can reproduce the experimental trends with isotopic position,
although they overestimate the magnitude of the measured
IEs, owing to the neglect of solvation. According to the
calculations, the IEs originate from vibrations whose frequen-
cies and zero-point energies are increased on N-protonation,
but it is not possible to associate those increases with
individual vibrational modes or with n-σ* delocalization. It
is also not possible to account for the surprising observations,
confirmed by calculation, that the IE is smallest for deutera-
tion at the nearby C2 or 2-methyl.
(calculated vibrational frequencies for formic acid, formate, and
their deuterated isotopologues, without and with a polarized
continuum model of solvation). This material is available free
of charge via the Internet at http://pubs.acs.org.
JA105331G
J. AM. CHEM. SOC. 9 VOL. 132, NO. 34, 2010 12149