Hydrogen isotope fractionation factors for N,N-dimethylbenzyl-ammonium ion and some related species: An unusually strong preference for deuterium over protium

Article abstract of DOI:10.1039/a604674f

Deuterium fractionation factors were determined by the ¹H and ¹³C NMR methods in aqueous solution for PhCH₂NLMe₂⁺ (φ = 1.47 ± 0.05), PhCH₂OL (φ = 1.04 ± 0.06), PhCO₂L (φ = 1.04 ± 0.08), and CH₃CO₂L (φ = 0.99 ± 0.02). The medium effect for transferring PhCH₂NMe₂ from H₂O to D₂O, Φ = 1.025 ± 0.003, was also determined by partitioning this substance between water and immiscible organic solvents, and a UV spectroscopic method was used to measure the solvent isotope effect on the acid ionization of PhCH₂NLMe₂⁺, (Q_a)_H/(Q_a)_D = 4.88 ± 0.16. This solvent isotope effect agrees well with the value predicted using the relevant fractionation factors, (Q_a)_H/(Q_a)_D = 4.38 ± 0.28. The unusually large value of φ for PhCH₂NLMe₂⁺ is attributed to stiffened bending vibrations of its N-L bond imposed by the tetrahedral structure of the ion and the bulk of its methyl groups.

Full text of DOI:10.1039/a604674f

View Article Online / Journal Homepage / Table of Contents for this issue
Hydrogen isotope fractionation factors for N,N-dimethylbenzyl-
ammonium ion and some related species: an unusually strong
preference for deuterium over protium
Hong-Xun Guo and A. Jerry Kresge*
a
Department of Chemistry, University of Toronto, Toronto, Ontario M5S 3H6, Canada
1
13
D euterium fractionation factors were determined by the H and C N M R methods in aqueous solution
+
for PhCH N LM e2 (φ = 1.47 ± 0.05), PhCH OL (φ = 1.04 ± 0.06), PhCO L (φ = 1.04 ± 0.08), and CH CO L
2
2
2
3
2
(
φ = 0.99 ± 0.02). The medium effect for transferring PhCH N M e from H O to D O, Φ = 1.025 ± 0.003, was
also determined by partitioning this substance between water and immiscible organic solvents, and a U V
spectroscopicmethod was used to measurethesolvent isotopeeffect on theacid ionization of PhCH N LM e ,
2
2
2
2
+
2
2
(
Q ) /(Q ) = 4.88 ± 0.16. This solvent isotope effect agrees well with the value predicted using the relevant
a
H
a D
+
fractionation factors, (Q ) /(Q ) = 4.38 ± 0.28. The unusually large value of φ for PhCH N LM e is
attributed to stiffened bending vibrations of its N ᎐L bond imposed by the tetrahedral structure of the ion
a
H
a
D
2
2
and the bulk of its methyl groups.
Deuterium fractionation factors have proved to be especially
tion frequencies, but this effect was offset by compensatory
useful in analysing and predicting solvent isotope effects in light
increases in bending vibration frequencies. These increases were
attributed to the tetrahedral structure of ammonium ions
which restricts N᎐L bending motion in these substances. The
hydronium ion, on the other hand, is trigonal with a very flat
pyramidal structure that allows considerably less restricted
‘out-of-plane’ bending motion. The lowered stretching fre-
quencies of the hydronium ion are consequently not offset by
increased bending frequencies, and its fractionation factor is
therefore less than that of water.
1
(
H O) and heavy (D O) water. Such fractionation factors are
2
2
partial isotope effects attributable to individual hydrogens.
They are expressed as D :H ratios and are defined in terms of
isotope exchange equilibria involving the solvent: the fraction-
ation factor at some particular hydrogenic site, φ , is equal to the
i
D :H ratio at that site, (D/H) , compared to the D :H ratio of
i
the solvent with which it is equilibrated, (D/H) , eqn. (1).
s
φ = (D/H) /(D/H)
(1)
It follows from this hypothesis that replacement of the
hydrogens of an ammonium ion with larger groups should
restrict bending motion even more and lead to still greater
fractionation factors. We have therefore examined the N,N-
dimethylbenzylammonium ion, 3, in order to see whether or not
this is so. We determined the fractionation factor for this sub-
i
i
s
Fractionation factors reflect the tightness of bonding to the
hydrogens they represent: φ > 1 denotes tighter, and φ < 1 loos-
er, bonding than that in the reference oxygen–hydrogen bonds
of water. For example, the less than unit value of the fraction-
ation factor for the hydronium ion in aqueous solution,
1
2
stance by the traditional H NMR method and also by the
1
,2
13
5
13
φ = l† = 0.69, indicates that the bonding of the hydrogens in
this substance is looser than that of the hydrogens in water,
which is consistent with the lower vibrational frequencies of the
more recently devised C NMR method. We made the
C
1
3
NMR measurements using substrate enriched in C at the ben-
1
3
zylic position, which we prepared from [α- C]benzoic acid; we
therefore also determined the fractionation factor for benzoic
acid, its reduction product benzyl alcohol and acetic acid as well.
3
hydronium ion than those of water. This looser bonding is
believed to be a consequence of the positive charge of the
hydronium ion which draws electron density out of the bonds
of this substance, making them weaker.
It is generally believed that this difference in fractionation
factor between hydronium ion and water extends to all posi-
tively charged and neutral oxygen–hydrogen bonds. A similar
difference, however, appears not to apply to positively charged
and neutral nitrogen–hydrogen bonds: we have recently shown
that the fractionation factor for the positively charged N᎐L ‡
bonds of benzylammonium ion, 1, φ = 1.08, is actually greater
Experimental
Materials
13
[
α- C]N,N-Dimethylbenzylamine was prepared by lithium alu-
13
minium hydride reduction of [α- C]N,N-dimethylbenzamide,
itself obtained from [α- C]benzoic acid (Isotec, Inc. 99 atom%
C at the labelled position) via the acid chloride. [α- C]Benzyl
alcohol was prepared by lithium aluminium hydride reduction
of the same benzoic acid. [α- C]Methyltrimethylammonium
13
13
13
1
3
+
+
4
PhCH NLMe
2
PhCH NL3
2
PhCH NL2
2
2
iodide was prepared previously. The identities of these sub-
stances were confirmed by their H NMR spectra. All other
1
1
2
3
materials were best available commercial grades.
than that for the neutral N᎐L bonds of benzylamine, 2,
4
+
φ = 0.96. Examination of vibrational frequencies of NL and
NMR
4
+
NL as surrogates for PhCH NL and PhCH NL showed that
3
2
3
2
2
NMR measurements were made with Varian Gemini 200 or
Gemini 300 instruments operating at probe temperatures of
the expected bond weakening accompanying positive charge
introduction was indeed reflected in lowered stretching vibra-
2
5 ЊC. Coaxial 5 mm sample tubes were used, with the sample
solution containing substrate plus the tetramethylammonium
iodide (0.001 ) reference placed in the outer compartment and
†
The special symbol ‘l’ is generally used for the fractionation factor of
the hydronium ion.
The symbol ‘L’ denotes either protium or deuterium, i.e. L = H or D.
D O (to provide a frequency lock) placed in the inner com-
2
‡
partment. Because the reference and substrate were present
J. Chem. Soc., Perkin Trans. 2, 1997
295

View Article Online
a
together in the same solution, no correction for differences in
magnetic susceptibility of the H O, HDO and D O solvents
were necessary.
Table 1 Summary of fractionation factors
2
2
φ
Weighted
average
1
3
b
1
b
pQ determination
The acid dissociation constant, pQ ,§ of N,N-dimethylbenzyl-
ammonium ion was determined spectrophotometrically using
the change in absorbance at λ 224 nm that accompanies
this ionization reaction. Measurements were made with a Cary
a
Species
PhCH NLMe
C NMR
H NMR
a
+
1.482 ± 0.058
1.102 ± 0.115
1.063 ± 0.129
0.990 ± 0.130
1.461 ± 0.076
1.024 ± 0.065
1.031 ± 0.107
0.981 ± 0.067
0.985 ± 0.026
1.474 ± 0.046
1.043 ± 0.056
1.044 ± 0.082
0.986 ± 0.022
2
2
PhCH OL
2
PhCO L
2
CH CO L
3
2
1
2
18 spectrometer, operating with the cell compartment at
5.0 ± 0.05 ЊC, on solutions of fixed stoichiometric N,N-
0
.990 ± 0.067
Ϫ5
a
b
dimethylbenzylammonium ion concentrations (ca. 5.5 × 10
In aqueous solution, 25 ЊC, ionic strength = 0.10 . Uncertainties are
standard deviations obtained by propagating errors in constituent
NMR parameters.
) but variable, and known, hydrogen ion concentrations; ionic
strength was maintained at 0.10 . Equilibrium constants, Qa,
were evaluated by non-linear least-squares fitting of eqn. (2), in
+
A Q + A [L ]
B
a
BH
+
(2)
Aobs
=
Q + [L ]
a
which A and A are absorbances of the substrate completely
B
BH
+
in its basic and acidic forms respectively and [L ] is hydrogen
ion concentration.
Partition coefficients
Partition coefficients for the distribution of N,N-dimethyl-
benzylamine between organic solvents and water were deter-
mined spectrophotometrically using the absorbance of the
amine at λ 260 nm. Measurements were made with a Cary 118
spectrometer whose cell compartment was thermostatted at
2
5.0 ± 0.05 ЊC. The absorbance of aqueous solutions of the
Ϫ3
amine (ca. 1 × 10 ) was first measured; 5.00 ml aliquots of
these solutions were then shaken vigorously with 5.00 ml vol-
umes of organic solvent, the resulting mixtures were kept in a
Fig. 1 Dependence of the chemical shift of the α-carbon atom of
N,N-dimethylbenzylammonium ion on the concentration of this ion in
H O solution
2
2
5.0 ± 0.05 ЊC bath for 4–5 h to achieve phase separation, and
the absorbance of the organic layer was then measured again.
Partition coefficients, K, were calculated using eqn. (3) in which
conditions: in H O solution (δ ), in D O solution (δ ) and in
2
H
2
D
an H O–D O mixture of deuterium atom fraction x (δ ). The
2
2
x
deuterium to protium ratio at the exchanging site is then equal
A0
to (δ Ϫ δ )/(δ Ϫ δ ), and the fractionation factor at that site is
H
x
x
D
(3)
K =
equal to this ratio divided by the deuterium to protium ratio of
the solvent, as shown in eqn. (4).
A Ϫ A
0
eq
A and A are absorbances of the aqueous solutions before and
after equilibration with the organic solvent.
0
eq
δ Ϫ δ
x
x
H
(4)
φ = ͩ ͪ
δ Ϫ δ
x
΋1 Ϫ x
H
Results
13
Observed C NMR chemical shifts of the enriched α-carbon
atom of N,N-dimethylbenzylammonium ion proved to be lin-
early dependent on the concentration of this substance in all
Fractionation factors
Amines and ammonium ions dissolved in water exchange the
hydrogens of their N᎐L bonds with the solvent rapidly, and this
makes it difficult to remove these substances from solution in an
H O–D O mixture for conventional isotopic analysis without
three solvents employed, H
O, D O and a 50:50 H O–D O
2
2
2
2
mixture. Measurements were consequently made over a range
of concentrations, at a fixed ionic strength (0.10 , NaClO ),
2
2
4
and the data were extrapolated to zero concentration by linear
least-squares analysis. An example is shown in Fig. 1. The zero-
concentration chemical shifts were then used to calculate the
fractionation factors listed in Table 1. Similar measurements
were made on benzyl alcohol, benzoic acid and acetic acid and
the fractionation factors obtained for these substances are also
listed in Table 1.
disturbing their deuterium content. Fortunately, methods have
been developed for conducting the isotopic assay directly on
1
3
the dissolved, rapidly exchanging substances. The C NMR
method is based upon the fact that substitution of deuterium
for protium in a bond to hydrogen produces an isotope effect
on the chemical shift of nearby carbon atoms. In a rapidly
exchanging system, the change in chemical shift caused by this
isotope effect will depend upon the deuterium to protium ratio
in the isotopically substituted bond, and the phenomena can
consequently be used to report the average deuterium content
of that bond.
Another method of determining fractionation factors for
1
rapidly exchanging hydrogens uses
H NMR. This method is
1
based upon the fact that the H NMR signal for hydrogens
exchanging rapidly with the solvent is a composite line whose
position depends upon the relative amounts of protium in the
solvent and at the exchanging site. Introduction of deuterium
into the system will change these relative amounts (unless the
fractionation factor is unity), and that will change the position
of the NMR signal. This change in position, plus knowledge of
the stoichiometry of the system, may then be used to calculate
the fractionation factor.
1
3
Use of this method requires measurement of the C chem-
ical shift of the reporting carbon atom under three different
§
The acid dissociation constants determined here are concentration
quotients applicable at the ionic strength of the measurements (0.10 ).
The symbol ‘Q ’ is used for such concentration quotients and ‘K ’ is
a
a
reserved for thermodynamic acidity constants referred to zero ionic
strength.
The relationship governing this phenomenon for a single sol-
2
96
J. Chem. Soc., Perkin Trans. 2, 1997

View Article Online
Table 2 Partition coefficients for distribution of N,N-dimethylbenzyl-
a
amine between water and organic solvents
K
Organic solvent
Hexane
H O
D O
Φ
2
2
0.1144
0.1163
0.1169
1.016
1.029
0
.1136
Cyclohexane
Heptane
0.1161
.1169
0.1195
0.1193
1.029
1.020
0
0.08758
.08850
0.09041
0.09041
1.032
1.022
0
Average: Φ = 1.025 ± 0.003
At 25 ЊC; ionic strength of aqueous phase = 0.10 .
a
Fig. 2 Relationship between N,N-dimethylbenzylammonium ion con-
centration and the composite chemical shift of this ion and water in
H O solution; chemical shifts are referenced to the methyl group signal
2
of the ion
1
b,2,6
ute is given in eqn. (5),
where δ is the chemical shift of the
0
δs[S]
(5)
δobs = δ +
0
1
Ϫ x + xφ
solvent and δ is that of the solute. The observed chemical shift,
s
δobs, is thus expected to be a linear function of solute concentr-
ation. This was found to be so for all of the substrates examined
here; Fig. 2 shows a typical example. The slope of the linear
relationship in H O, where x = 0, provides δ , and that, in com-
2
s
bination with the slope determined in an H O–D O mixture of
2
2
atom fraction x, then affords φ.
This method was used to determine fractionation factors for
Fig. 3 Titration curves for the ionization of N,N-dimethylbenzyl-
ammonium ion in H O (O) and D O (∆)
1
3
each of the four substrates to which the C NMR method had
been applied; the results obtained are also listed in Table 1. It
may be seen that the fractionation factors determined by the
two methods agree with each other very well: the differences
between the results obtained by the two methods are all well
within the combined uncertainties of the individual determin-
ations. This is significant, because fractionation factors deter-
2
2
+
gens in L O , each with the fractionation factor l. The conju-
3
gate base PhCH NMe has no exchangeable hydrogens, but it
2
2
may contribute a medium effect and the term Φ
therefore included. (Following Albery, we use lower case φ for
individual site fractionation factors and upper case Φ for
is
PhCH NMe
2 2
7
1
mined by the H NMR method include medium effects whereas
1
3
4
medium effects.) The fractionation factor for L O is unity, and
2
those determined by the C NMR method do not; medium
effects in the present systems are thus evidently too small to be
detected by the methods employed.
the term for this substance therefore drops out of eqn. (7).
Medium effects can be evaluated as solvent isotope effects on
partition coefficients for distribution of substrates between
water and an immiscible organic solvent. Such isotope effects
were determined for N,N-dimethylbenzylamine. The results
are summarized in Table 2; they lead to the medium effect
Φ = 1.025 ± 0.003. This medium effect is quite small, which is
consistent with our inability to determine medium effects as
Medium effects are produced by water molecules in the solv-
ation shells of the substances to which they refer. These water
molecules are different from bulk water, and their fractionation
factors are consequently different from the unit value that
applies to bulk water. The difference is usually quite small, but
because the number of solvating water molecules is appreciable,
their combined influence could be significant.
1
differences between fractionation factors obtained by the H
1
3
and C NMR methods as discussed above.
We determined the acid ionization constant of N,N-
dimethylbenzylammonium ion in H O and in D O by making
absorbance measurements in perchloric acid and sodium
hydroxide solutions as well as in biphosphate ion, boric acid
and hydrogencarbonate ion buffers. Ionic strength was main-
pQa
2
2
The solvent isotope effect on the acid ionization of N,N-
dimethylbenzylammonium ion, eqn. (6), can be used to provide
Q
a
+
+
PhCH NLMe2 + L O
PhCH NMe + L O
(6)
2
2
2
2
3
tained at 0.10  by adding NaClO as required. Hydrogen ion
4
concentrations of the buffers were obtained by calculation,
a check on the reliability of the presently determined fraction-
ation factor for this substance. This isotope effect may be
expressed in terms of fractionation factors for all of the sub-
stances taking part in this reaction, as shown in eqn. (7). In this
using pK (H O) = 7.201 and K (H O)/K (D O) = 3.43 for
a
2
a
2
a
2
8
biphosphate ion, pK (H O) = 9.223 and K(H O)/K(D O) =
a
2
2
2
9
3
5
.28 for boric acid and pK = 10.329 and K(H O)/K(D O) =
10
a
2
2
.07 for hydrogencarbonate ion, plus activity coefficients
expression, φPhCH 2NLMe2
+
and l are individual site fraction-
11
recommended by Bates; the isotope effect on the autoproto-
lysis of water was taken to be K(H O)/K(D O) = 7.28.
As Fig. 3 illustrates, the results gave well-defined titration
+
+
ation factors for single hydrogens in PhCH NLMe2 and L O ,
12
2
3
2
2
and l is raised to the third power because there are three hydro-
curves. Two sets of measurements were made in H O and two
2
(
Qa)
φPhCH 2NLMe2
+
H
in D O. Least-squares fitting of eqn. (2) gave Q = (1.172 ±
=
(7)
2
a
3
Ϫ9
Ϫ9
(
Qa)D ΦPhCH 2NMe2
l
0.026) × 10
and (1.112 ± 0.028) × 10
 for H O and
2
J. Chem. Soc., Perkin Trans. 2, 1997
297

View Article Online
Ϫ10
Ϫ10
Q = (2.316 ± 0.086) × 10
and (2.387 ± 0.101) × 10
 for
the general belief that fractionation factors for all neutral O᎐L
bonds are similar and close to the unity value of bulk water.
a
D O. The weighted averages of these values are Q (H O) =
1.144 ± 0.019) × 10  and Q (D O) = (2.346 ± 0.066) × 10
, and the ratio of these averages gives the solvent isotope effect
2
a
2
Ϫ9
Ϫ10
(

It is interesting that the fractionation factors for PhCH OL
a
2
2
and PhCO L are similar despite the great difference in acidity
2
Q (H O)/Q (D O) = 4.88 ± 0.16. The value for H O may be
of these two substances. It was once believed that solvent iso-
tope effects on the ionization of acids increase with decreasing
strength of the acid, an effect that should be reflected in a ten-
dency for fractionation factors of the acids to increase with
decreasing acid strength. With the accumulation of more evi-
a
2
a
2
2
converted into a thermodynamic acidity constant by applica-
1
1
tion of appropriate activity coefficients; the result gives pK =
a
8
.926 ± 0.007, in good agreement with pK = 8.91 reported in a
a
13
previous study.
1
6
Evaluation of eqn. (7) using the values of φ
+
and
dence, however, this idea was found to be incorrect.
2 2
PhCH NLMe
2
ΦPhCH 2NMe2 determined here plus l = 0.69 ± 0.01 gives the iso-
tope effect 4.38 ± 0.28. This agrees well with the directly meas-
ured result 4.88 ± 0.16, and that adds confidence to the reliabil-
ity of both the fractionation factor and acid ionization constant
determinations.
Acknowledgements
We are grateful to the US National Institutes of Health for
financial support of this work under Grant No. GM 47539.
References
Discussion
1
See, e.g. (a) D. M. Quinn and L. D. Sutton, in Enzyme Mechanisms
from Isotope Effects, ed. P. F. Cook, CRC Press, Boca Raton, FL,
1991, p. 73; R. L. Schowen, in Mechanistic Principles of Enzyme
Activity, eds. J. F. Liebman and A. Greenberg, VCH, New York,
1988, p. 119; (b) A. J. Kresge, R. A. More O’Ferrall and
M. F. Powell, in Isotopes in Organic Chemistry, eds. E. Buncel and
C. C. Lee, Elsevier, New York, 1987, vol. 7, p. 177.
A. J. Kresge and A. L. Allred, J. Am. Chem. Soc., 1963, 85, 1541;
V. Gold, Proc. Chem. Soc., 1963, 141.
3 R. A. More O’Ferrall, G. W. Koeppl and A. J. Kresge, J. Am. Chem.
Soc., 1971, 93, 1, 9.
The fractionation factor determined here for N,N-dimethyl-
benzylammonium ion, φPhCH 2NLMe2 = 1.47 ± 0.05, is con-
+
siderably greater than that for benzylammonium ion,
φPhCH 2NL3 = 1.08 ± 0.02. This difference supports the hypoth-
esis that the tetrahedral structure of ammonium ions restricts
the bending motion of their N᎐L bonds and the consequent
stiffening of these bonds raises their fractionation factors.
Because methyl groups are larger than hydrogen, this bond
stiffening should be greater for PhCH NLMe than for
PhCH NL3 and the fractionation factor for the former should
be greater than that for the latter, as observed.
The presently determined fractionation factor for PhCH N-
4
2
+
2
2
+
4 C. H. Arrowsmith, H.-X. Guo and A. J. Kresge, J. Am. Chem. Soc.,
2
1
994, 116, 8890.
5
R. M. Jarret and M. Saunders, J. Am. Chem. Soc. 1985, 107, 2648.
2
6 A. J. Kresge and Y. C. Tang, J. Phys. Chem., 1979, 83, 2156.
7 W. J. Albery, in Proton Transfer Reactions, eds. E. F. Caldin and
V. Gold, Chapman and Hall, London, 1975, ch. 9.
+
LMe2 is in fact unusually large. Values of the magnitude of the
present result have been reported recently for some of the back-
1
4
8 R. Gary, R. G. Bates and R. A. Robinson, J. Phys. Chem., 1964, 68,
bone amide hydrogens of staphylococcal nuclease, but frac-
tionation factors much greater than unity are rare. The only
value for an ammonium ion derived from a tertiary amine of
which we are aware is φ = 1.23 for N,N-dimethyl-p-nitro-
anilinium ion; though not as great as our result, this fraction-
ation factor is large, and that offers further support for the
idea that bending vibrations play an important role in deter-
mining the magnitude of fractionation factors of ammonium
ions.
3
806.
9
V. Gold and B. M. Lowe, J. Chem. Soc. A, 1968, 1923.
1
1
0 M. Paabo and R. G. Bates, J. Phys. Chem., 1969, 73, 3014.
1 R. G. Bates, Determination of pH Theory and Practise, Wiley, New
York, 1973, p. 49.
1
5
12 V. Gold and B. M. Lowe, J. Chem. Soc. A, 1967, 936.
1
1
1
1
3 H. Goldschmidt and R. M. Salcher, Z. Physik. Chem., 1899, 29, 89.
4 S. N. Loh and J. L. Markley, Biochemistry, 1994, 33, 1029.
5 V. Gold and C. Tomlinson, J. Chem. Soc. (B), 1971, 1707.
6 R. P. Bell, The Proton in Chemistry, 2nd edn., Cornell University
Press, Ithaca, New York, 1973, p. 234.
None of the fractionation factors determined here for the
oxygen–hydrogen bonds of PhCH OL, PhCO L and CH CO L
2
2
3
2
is significantly different from unity (see Table 1), and that for
Paper 6/04674F
Received 4th July 1996
Accepted 7th October 1996
CH CO L, φ = 0.99 ± 0.02, agrees well with φ = 0.96 ± 0.02
3
2
9
reported for this substance before. This provides support for
2
98
J. Chem. Soc., Perkin Trans. 2, 1997