Medium effect on the rotational barrier of carbamates and its sulfur congeners

Article abstract of DOI:10.1021/jo061934u

The solvent effect on rotation about the conjugated C-N bond has been studied for methyl N,N-dimethylcarbamate (1), S-methyl N,N-dimethylthiocarbamate (2), O-methyl N,N-dimethylthiocarbamate (3), and methyl N,N- dimethyldithiocarbamate (4). The present investigation included experimental determination of activation parameters (ΔH?, ΔS?, and ΔG?) combined with theoretical calculations via both quantum and classical approaches. Rotational barriers were measured through dynamic NMR experiments in solvents of varied polarity and proton donor ability. In the less polar solvents, the values were 15.3 ± 0.5 (CS₂), 14.0 ± 1.1 (CS₂), 17.5 ± 0.4 (CCl₄), and 14.6 ± 0.5 kcal/mol (CCl₄) for 1, 2, 3, and 4, respectively. Upon changing to an aqueous solution, the greatest variations occurred for 2 and 4, whereas for 1 and 3, there was no observable effect. Quantum chemical calculations at the HF/6-311+G(2d,p) and B3LYP/6-311+G(2d,p) levels, with the inclusion of solvation effects via the isodensity polarizable continuum model (IPCM), correctly reproduced the experimentally observed trends but failed to account for some of the measured rotational barrier's magnitudes. Hydrogen-bonding effects were included by performing molecular dynamic simulations. For these latter calculations, it was necessary to parametrize the force field against energies of water-solute complexes calculated at B3LYP/6-31+G-(d,p). Through the results of radial distribution functions, solution rotational barriers could be calculated, presenting good agreement with experimental determinations and revealing the role of hydrogen bonding. Interestingly, only for 2, the rotational barrier is predicted to increase as a result of complexation with water. For the remaining compounds, hydrogen bonding causes the barrier to decrease, contrasting with most of the molecular systems studied up to now.

Full text of DOI:10.1021/jo061934u

Medium Effect on the Rotational Barrier of Carbamates and Its
Sulfur Congeners
Rodrigo M. Pontes, Ernani A. Basso,* and Francisco P. dos Santos
Departamento de Qu´ımica, UniVersidade Estadual de Maringa´, AV. Colombo, 5790,
87020-900 Maringa´-PR, Brazil
eabasso@uem.br
ReceiVed September 19, 2006
The solvent effect on rotation about the conjugated C-N bond has been studied for methyl
N,N-dimethylcarbamate (1), S-methyl N,N-dimethylthiocarbamate (2), O-methyl N,N-dimethylthiocar-
bamate (3), and methyl N,N-dimethyldithiocarbamate (4). The present investigation included experimental
determination of activation parameters (∆Hq, ∆Sq, and ∆Gq) combined with theoretical calculations via
both quantum and classical approaches. Rotational barriers were measured through dynamic NMR
experiments in solvents of varied polarity and proton donor ability. In the less polar solvents, the values
were 15.3 ( 0.5 (CS2), 14.0 ( 1.1 (CS2), 17.5 ( 0.4 (CCl4), and 14.6 ( 0.5 kcal/mol (CCl4) for 1, 2,
3, and 4, respectively. Upon changing to an aqueous solution, the greatest variations occurred for 2 and
4, whereas for 1 and 3, there was no observable effect. Quantum chemical calculations at the HF/6-
311+G(2d,p) and B3LYP/6-311+G(2d,p) levels, with the inclusion of solvation effects via the isodensity
polarizable continuum model (IPCM), correctly reproduced the experimentally observed trends but failed
to account for some of the measured rotational barrier’s magnitudes. Hydrogen-bonding effects were
included by performing molecular dynamic simulations. For these latter calculations, it was necessary to
parametrize the force field against energies of water-solute complexes calculated at B3LYP/6-31+G-
(d,p). Through the results of radial distribution functions, solution rotational barriers could be calculated,
presenting good agreement with experimental determinations and revealing the role of hydrogen bonding.
Interestingly, only for 2, the rotational barrier is predicted to increase as a result of complexation with
water. For the remaining compounds, hydrogen bonding causes the barrier to decrease, contrasting with
most of the molecular systems studied up to now.
Introduction
increases its rotational barrier by more than 10 kcal/mol as
compared to ordinary amines, for which roughly 4 kcal/mol is
required to twist a C-N linkage.6 Such an elevated energy
barrier is largely responsible, for instance, for the conformational
stability of proteins and enzymes through peptide bonds.7-9
Currently, a great deal of knowledge has been accumulated
about the origin of the barrier and its response to the medium
When a nitrogen atom is attached to a double bond (as in
>N-CdO), its lone pair can delocalize over the π system
forming an approximately planar three-atom framework (Figure
1).1-5 The double bond character acquired by the C-N bond
(1) Stewart, W. E.; Siddall, T. H., III Chem. ReV. 1970, 70, 517-551.
(2) Wiberg, K. B.; Rablen, P. R.; Rush, D. J.; Keith, T. A. J. Am. Chem.
Soc. 1995, 117, 4261-4270.
(3) Laidig, K. E.; Cameron, L. M. J. Am. Chem. Soc. 1996, 118, 1737-
1742.
(4) Lauvergnat, D.; Hilberty, P. C. J. Am. Chem. Soc. 1997, 119, 9478-
9482.
(5) Lim, K.-T.; Francl, M. M. J. Phys. Chem. 1987, 91, 2716-2721.
10.1021/jo061934u CCC: $37.00 © 2007 American Chemical Society
Published on Web 02/17/2007
J. Org. Chem. 2007, 72, 1901-1911
1901

Pontes et al.
FIGURE 1. Resonance structures used to explain the barrier to C-N
rotation in amides and correlated compounds.
FIGURE 3. Structural formulas of the studied compounds.
solvents. Rablen et al.22,23 performed related studies for N,N-
dimethylaminoacrylonitrile and also for this compound they
verified variations with the solvent.
Carbamates, -O-(CdO)-N<, bear the same amide frame-
work and would be expected to have their rotational barrier
changed by the solvent polarity. However, Cox and Lectka8
demonstrated that the rotational barrier of carbamates is practi-
cally insensitive to the medium. Additionally, Rablen24 calcu-
lated the dipole moments for the ground (GS) and transition
states of methyl N,N-dimethylcarbamate and found an interesting
situation because the differences between the dipoles of GS and
the preferred transition state (TS1) were similar to those of
amides. Even so, the calculations corroborated the experimental
results of a small or vanishing solvent effect. The paradox was
then solved by arguing that the solvent effect should be
proportional to the GS dipole moment. This proposal was
successful in explaining the distinct behavior of amides and
carbamates because the dipole moment of the latter is about
one-half of that of amides; this difference further increases if
one assumes a quadratic behavior like the one of the Onsager
theory.25 Additionally, carbamates are poorer proton acceptors,
which makes them less sensitive than amides to protic solvents.24
Complementary studies by the present authors helped to confirm
that the solvent insensitivity is a characteristic of the functional
group carbamate, also clarifying the role of dipole moments
and dipole moment variations in the rotation process.26
To get a better understanding of the rotational barrier in
amide-like systems, it is necessary to extend the studies to a
broader class of compounds. On the basis of this, we wished to
see if sulfur substitution would maintain the carbamate behavior
of no increase with the solvent polarity or proton donor ability.
This was not the case, as described below. The inclusion of a
third-row element was expected to affect the molecular struc-
tures in all of its aspects that are important to the rotational
barrier behavior: polarizability, proton affinity, conjugation, and
steric effects. Carbamates and thiocarbamates have been studied
on several occasions regarding their rotational barrier, but
unfortunately, the set of data is not enough to definitely establish
the role of sulfur substitution on the solvent effect.9,10,27-39 As
we shall see, this is an issue where the combination of
experimental measurements and theoretical calculations is of
the most importance to achieve a clear understanding.
FIGURE 2. Transition states (TS1 and TS2) generated by rotation
over the central C-N bond of the planar ground state (GS).
for amides.2-4,10-20 The comprehensive study of Wiberg et al.2
demonstrated that the rotational barrier in amides increases with
the solvent polarity and proton donor ability. For N,N-dimethyl-
formamide and N,N-dimethylacetamide, variations of 4 and 2
kcal/mol, respectively, were observed when passing from the
gas phase to a water solution. Such increases could be
rationalized on the basis of the transition states for the process
as follows. Rotation over the C-N bond breaks the conjugation
and generates two transition states, TS1 and TS2 (Figure 2).
Because the transition and ground states are expected to possess
different dipole moments, the rotational barrier of amide-like
systems should, in principle, vary with the solvent polarity, and
this is indeed observed for amides.2 Similar results were obtained
for the corresponding sulfur derivatives, N,N-dimethylthio-
formamide and N,N-dimethylthioacetamide.21 The rotational
barriers are larger for the thioamides but still increase in polar
(6) Benson, S. B. Thermochemical Kinetics: Methods for Estimation of
Thermochemical Data and Rate Parameters, 2nd ed.; John Wiley & Sons:
New York, 1976.
(7) Cox, C.; Lectka, T. Acc. Chem. Res. 2000, 33, 849-858.
(8) Cox, C.; Lectka, T. J. Org. Chem. 1998, 63, 2426-2427.
(9) Souza, W. F.; Kambe, N.; Sonoda, N. J. Phys. Org. Chem. 1996, 9,
179-186.
(10) Deetz, M. J.; Forbes, C. C.; Jonas, M.; Malerich, J. P.; Smith, B.
D.; Wiest, O. J. Org. Chem. 2002, 67, 3949-3952.
(11) Duffy, E. M.; Severance, D. L.; Jorgensen, W. L. J. Am. Chem.
Soc. 1992, 114, 7535-7542.
(12) Feigel, M.; Strassner, T. J. Mol. Struct. (Theochem.) 1993, 283,
33-48.
(13) Jackman, L. M.; Cotton, F. A. Dynamic Nuclear Magnetic Reso-
nance Spectroscopy; Academic Press: New York, 1975.
(14) Vassilev, N. G.; Dimitrov, V. J. Mol. Struct. 1999, 484, 39-47.
(15) Wiberg, K. B.; Breneman, C. M. J. Am. Chem. Soc. 1992, 114,
831-840.
In this way, we selected four compounds, namely, methyl
N,N-dimethylcarbamate (1), S-methyl N,N-dimethylthiocarbam-
ate (2), O-methyl N,N-dimethylthiocarbamate (3), and methyl
N,N-dimethyldithiocarbamate (4) (Figure 3). Each of them was
prepared and subjected to experimental measurements by
(16) Wiberg, K. B.; Hadad, C. M.; Rablen, P. R.; Cioslowski, J. J. Am.
Chem. Soc. 1992, 114, 8644-8654.
(17) Wiberg, K. B.; Laidig, K. E. J. Am. Chem. Soc. 1997, 119, 5935.
(18) Wiberg, K. B.; Laidig, K. E. J. Am. Chem. Soc. 1987, 109, 5935-
5943.
(19) Wiberg, K. B.; Rablen, P. R. J. Am. Chem. Soc. 1995, 117, 2201-
2209.
(20) Wiberg, K. B.; Rablen, P. R. J. Am. Chem. Soc. 1993, 115, 9234-
9242.
(21) Wiberg, K. B.; Rush, D. J. J. Am. Chem. Soc. 2001, 123, 2038-
2046.
(22) Rablen, P. R.; Miller, D. A.; Bullock, V. R.; Hutchinson, P. H.;
Gorman, J. A. J. Am. Chem. Soc. 1999, 121, 218-226.
(23) Rablen, P. R.; Pearlman, S. A.; Miller, D. A. J. Am. Chem. Soc.
1999, 121, 227-237.
(24) Rablen, P. R. J. Org. Chem. 2000, 65, 7930-7937.
(25) Onsager, L. J. Am. Chem. Soc. 1936, 58, 1486-1493.
(26) Basso, E. A.; Pontes, R. M. J. Mol. Struct. (Theochem.) 2002, 594,
199-206.
1902 J. Org. Chem., Vol. 72, No. 6, 2007

Rotational Barrier In Carbamate Congeners
TABLE 1. Activation Parameters for Compounds 1-4 Determined through DNMR Total Line Shape Analysis Measurementsa
1
2
∆Hq
∆Sq
∆Gq
∆Hq
∆Sq
∆Gq
CCl4
Bz-d6
CS2
CD2Cl2
CD3OD
CD3CN
20% D2O/CD3OD
DMSO-d6
D2O
-
-
-
-
-
-
-
-
-
-
-
-
10.8 ( 0.3
10.4 ( 0.6
14.2 ( 0.6
12.9 ( 1.4
10.1 ( 0.8
-
-15.0 ( 1.2
-16.6 ( 2.1
-3.1 ( 2.1
-7.0 ( 5.4
-17.1 ( 2.9
-
15.3 ( 0.5
15.3 ( 0.9
15.2 ( 0.9
15.0 ( 2.1
15.2 ( 1.2
-
12.5 ( 0.7
14.7 ( 1.0
12.6 ( 1.1
13.5 ( 0.7
9.7 ( 0.3
-
-5.4 ( 2.8
1.7 ( 4.1
-6.7 ( 4.2
-2.8 ( 2.7
-17.7 ( 1.2
-
14.0 ( 1.1
14.2 ( 1.6
14.6 ( 1.6
14.3 ( 1.1
14.9 ( 0.5
-
11.6 ( 1.0
-10.6 ( 3.4
14.8 ( 1.4
11.6 ( 0.5
-10.8 ( 1.7
14.8 ( 0.7
3
4
∆Hq
∆Sq
∆Gq
∆Hq
∆Sq
∆Gq
CCl4
Bz-d6
CS2
CD2Cl2
CD3OD
CD3CN
20% D2O/CD3OD
DMSO-d6
D2O
17.4 ( 0.3
-0.2 ( 0.9
17.5 ( 0.4
13.8 ( 0.4
15.1 ( 0.2
-
-2.7 ( 1.2
0.3 ( 0.5
-
14.6 ( 0.5
15.0 ( 0.2
-
19.9 ( 0.8
8.0 ( 2.5
17.6 ( 1.1
-
-
-
-
-
-
-
-
-
-
-
-0.2 ( 0.9
-
-
14.2 ( 0.1
15.7 ( 0.1
14.8 ( 0.2
16.0 ( 0.2
12.7 ( 0.2
-3.1 ( 0.4
0.4 ( 0.3
-2.5 ( 0.6
0.0 ( 0.5
-10.8 ( 0.6
15.2 ( 0.2
15.6 ( 0.1
15.5 ( 0.3
16.0 ( 0.2
15.9 ( 0.3
17.4 ( 0.3
17.5 ( 0.4
-
-
18.7 ( 0.3
-
2.7 ( 0.7
-
17.9 ( 0.3
-
a ∆Hq and ∆Gq in kcal/mol and ∆Sq in cal/K mol.
dynamic nuclear magnetic resonance (DNMR) in solvents of
varied polarity and proton-donor ability. The theoretical study
was composed of electronic structure routines (ab initio and
DFT) together with molecular dynamics simulations.
entropies are still negative in the gas phase where no solvation
structure can affect its values.
The probable reason for the negative entropy is the loss of
one vibrational mode when going from GS to TS, once the C-N
torsion becomes an imaginary frequency and no longer con-
tributes to the thermodynamic properties of the system.2 The
few cases for which the entropy is positive can be understood
in terms of complexation with solvent molecules. For instance,
it is known40 that aromatic solvents complex to the planar GS
of amides due to the effect associated with resonance structure
b (Figure 1), an effect that disappears in the TSs. Thus, for the
cases where the activation entropy is positive, the ground states
seem to be more organized with the solvent molecules overcom-
ing the negative gas-phase entropy.
Consider now the activation Gibbs energies that are what we
effectively call rotational barriers. The values for 1 in Table 1
agree with previous studies8,24,26 for carbamates both in the
barrier magnitude and in its response to the medium, i.e., the
absence of a representative variation upon changing the solvent
polarity or hydrogen-bonding ability. (The rotational barrier in
water is actually 0.5 kcal/mol smaller than that in CS2 suggesting
a decrease in the rotational barrier; however, the variation is
smaller than the experimental error and this hypothesis cannot
be confirmed by these data only.) Compound 2 presents a
rotational barrier about 1.0 kcal/mol smaller than that for 1 in
the apolar solvent CS2, but the values become essentially equal
for both in water. Thus, experiments suggest an increase in the
rotational barrier of 2. There is a limitation in this conclusion
imposed by the experimental errors. To overcome this problem,
we also used ∆Gq values calculated through the coalescence
temperature procedure.1 The method consists of measuring the
limiting separation of the absorption frequencies of the exchang-
ing methyl groups together with the temperature at which the
two signals coalesce. Limitations due to the solvent freezing or
boiling, however, prevented us from obtaining suitable data for
Results and Discussion
NMR Measurements. The results of total line shape analysis
(TLSA) for compounds 1-4 are presented in Table 1. Our main
interest is in activation Gibbs energies (∆Gq), but before
commenting on them, let us make a brief analysis of entropies
and enthalpies of activation. There are some cases in which the
rotational barrier can be thought of almost completely as an
enthalpic phenomenon,8 e.g., compound 4 in DMSO or com-
pound 3 in CCl4, whereas for most of them, it is indispensable
to include the entropic contribution. With a few exceptions, the
activation entropies (∆Sq) are negative, and we are tempted to
imagine a more organized solvation shell for the transition states
(TS1 and TS2). However, as already reported for amides2 and
also verified in the calculations to be presented (Table 4), the
(27) Valega, T. M. J. Org. Chem. 1966, 31, 1150-1153.
(28) Martin, M. L.; Marbon, F.; Trierweller, M. J. Phys. Chem. 1981,
85, 76-78.
(29) Smith, B. D.; Goodenough-Lashua, D. M.; D’Souza, C. J. E.; Norton,
K. J.; Schmidt, L. M.; Tung, J. C. Tetrahedron Lett. 2004, 45, 2747-2749.
(30) Yamagami, C.; Takao, N.; Takeuchi, Y. Aust. J. Chem. 1986, 39,
457-463.
(31) Lemire, A. E.; Thompson, J. C. Can. J. Chem. 1970, 48, 824-829.
(32) Kost, D.; Kornberg, N. Tetrahedron Lett. 1978, 35, 3275-3276.
(33) Gayathri Devi, K. R.; Sathyanarayana, D. N.; Manogaran, S.
Spectrochim. Acta 1981, 37A, 31-36.
(34) Julia`, S.; Ginebreda, A.; Sala, P.; Sancho, M.; Annunziata, R. C. F.
Org. Magn. Reson. 1983, 21, 573-575.
(35) Kornberg, N.; Kost, D. J. Chem. Soc., Perkin Trans. 2 1979, 1661-
1664.
(36) Schlottmann, B. U. Tetrahedron Lett. 1971, 1221-1224.
(37) Lustig, E.; Benson, W. R.; Duy, N. J. Org. Chem. 1967, 32, 851-
852.
(38) Kleinpeter, E.; Kretschemer, M.; Borsdorf, R.; Widera, R.; Mu¨hl-
sta¨dt, M. J. Prakt. Chem. 1980, 322, 793-797.
(39) Kleinpeter, E.; Widera, R.; Mu¨hlsta¨dt, M. J. Prakt. Chem. 1977,
319, 133-139.
(40) Wayland, B. B.; Drago, R. S.; Henneike, H. F. J. Am. Chem. Soc.
1966, 88, 2455.
J. Org. Chem, Vol. 72, No. 6, 2007 1903

Pontes et al.
TABLE 2. Rotational Barriers (∆Gq in kcal/mol) for Compounds 1, 2, and 4 Determined through the Coalescence Temperature Methoda
1
2
4
b
b
b
Tc
∆νc
∆Gq
Tc
∆νc
∆Gq
Tc
∆νc
∆G‡
CCl4
Bz-d6
CS2
CD2Cl2
CD3OD
CD3CN
20% D2O/CD3OD
DMSO-d6
D2O
-
-
-
-
-
-
-
31.5
54.7
-
35.1
149.0
-
15.2d
15.5
-
-
-
-
-5.8
-14.6
-1.0
-7.8
13.7
-
-
-
21.6
5.9
10.9
3.6
4.1
1.8
6.0
-
15.4
15.2
15.4
15.1
15.2d
-
6.6
2.1
4.0
2.3
8.2
-
14.2
14.3
14.7
14.6
15.1
-
-
-
-
11.6
-1.0
11.8
-
43.5
43.6
49.1
54.7
61.1
42.4
43.5
40.5
34.1
38.3
15.7
15.7
16.0
16.5
16.7
10.8
3.0
15.5d
17.7
9.3
15.3d
a
Experimental errors for ∆Gq are within (0.2 kcal/mol. b Coalescence temperature in °C. c Difference in the absorption frequencies of the nonequivalent
d
methyl protons, in Hz. For these cases, it was not possible to attain a suitable ∆ν, so these values are less reliable as compared to the remaining ones.
compound 3. For the remaining compounds, it was possible to
measure the rotational barriers with this method (Table 2). The
advantage of using the coalescence temperature comes from the
fact that the experimental errors are in the vicinity of (0.2 kcal/
mol, which is better, for example, than the TLSA values for 2.
The rotational barrier of compound 1, according to Table 2,
agrees with those obtained through TLSA both in magnitude
and in the absence of solvent effect. The entries in Table 2 also
show an increase of ∼1.0 kcal/mol for 2, and now this value
lies well above the experimental error, confirming that the
rotational barrier for compound 2 does experience an increase
with the solvent polarity and maybe with hydrogen bonding
also.
FIGURE 4. Possible conformations for the studied compounds.
On passing from CCl4 to water, the rotational barrier in
compound 4 undergoes an increase of 1.3 kcal/mol according
to TLSA (Table 1), which is a representative increase above
the experimental errors. The same behavior can be noticed in
Table 2, and thus it is possible to assert that the rotational barrier
of compound 4 suffers a representative solvent effect. Wiberg
and Rush21 verified that the barriers in thioamides are greater
than that of amides and so are the solvent effects. A similar
behavior is observed here for thiocarbamates; i.e., the rotational
barrier variations in carbamates, if any, could not be detected
by the experiments so far performed, but sulfur substitution puts
the variations in a measurable plateau, except for compound 3.
For this particular one, we could not acquire data in protic
solvents due to an apparent sample decomposition. The com-
pound initially formed a homogeneous solution but soon
separated phases accompanied by the evolution of a white
smoke. This indicates a very strong water-solute interaction
compromising even the molecular stability; we did not go deeper
into the reactivity of this compound as it is not the focus of the
present study. In aprotic media, there are no representative
variations for compound 3.
optimization and to frequency calculations (HF and B3LYP)
to establish their stationary point nature. Both 1 and 2 exist in
only one form from B3LYP as well as from HF, namely, GS1.
Contrastingly, we found two conformations for 3 and 4, and in
these cases, the theoretical methods provide different results.
Conformations GS2 and GS3 are stable at HF for 3, with
the two being almost isoenergetic (0.02 kcal/mol in favor of
GS2), but for B3LYP, only GS2 was obtained. A possible
reason for this divergence is the differences in the structural
parameters provided by each method. The C-H bond eclipsing
the (CdS)-N bond in the anti-methyl group of GS3 (Figure
4) is 2.197 Å from the oxygen atom (-O-) at HF, whereas at
B3LYP, this distance is slightly smaller (2.188 Å) increasing
the steric repulsion. Moreover, it is known that electron
correlation methods tend to put more electron density on the
molecular periphery relative to HF.41 Together, both factors
make GS3 unstable at B3LYP.
In the case of 4, conformations GS2 and GS4 are stable at
B3LYP and only GS4 exists with HF. A similar argument may
be invoked here. The distance between the carbon in the syn-
methyl group and the double-bonded sulfur atom is 2.973 Å at
HF and 2.994 Å at B3LYP in GS2. Therefore, steric repulsion
should be greater at HF making conformation GS2 converge
to GS4.
To understand how the rotational barriers of these four
compounds respond to the solvent medium, we performed
theoretical calculations using electronic structure methods as
well as liquid simulations by molecular dynamics, as described
hereafter.
For the transition states TS1 and TS2, we used structures
like those depicted in Figure 2, both of them having only one
imaginary frequency corresponding to the (CdZ)-N torsion
which characterizes these structures as the correct first-order
saddle points.
Rotational Barriers Determined through a Continuum
Solvation Model. Amides and thioamides commonly adopt the
conformation labeled as GS1 in Figure 4.14 That was also our
starting point for the present compounds, but the outcome was
somewhat different for the sulfur derivatives. Thus, it was
necessary to conduct a conformational search prior to advancing
with the theoretical studies. There are in principle four
candidates for the ground-state conformation, labeled as GS1-
GS4 in Figure 4, and these were submitted to full geometry
(41) Hadad, C. M.; Rablen, P. R.; Wiberg, K. B. J. Org. Chem. 1998,
63, 8668-8681.
(42) Lide, D. R. CRC Handbook of Chemistry and Physics, 77th ed.;
CRC Press: Boca Raton, 1996.
1904 J. Org. Chem., Vol. 72, No. 6, 2007

Rotational Barrier In Carbamate Congeners
TABLE 3. Gas-Phase Activation Parameters for Compounds 1-4
Calculated at the HF/6-311+G(2d,p) and B3LYP/6-311+G(2d,p)
Levels of Theorya
TABLE 4. Rotational Barriers (∆Gq, in kcal/mol) for Compounds
1-4 Calculated at the HF/6-311+G(2d,p) and B3LYP/
6-311+G(2d,p) Levels of Theory with Solvation Effects Included
through the IPCM Methoda
HF
B3LYP
TS1
B3LYP
TS1
B3LYP
eff.b
B3LYP
∆Hq
∆Sq
∆Gq
∆Hq
∆Sq
∆Gq
HF
HF
HF
1
TS1
TS2
eff.b
2
TS1
TS2
eff.b
3c
TS1
TS2
eff.b
4d
TS1
TS2
eff.b
12.86
13.89
-7.08
-6.71
14.97
15.89
14.86
13.86
14.77
-8.28
-7.90
16.33
17.12
16.19
1
ꢀ
ꢀ
ꢀ
ꢀ
1.00
2.23
2.61
32.61
35.69
46.83
78.36
14.97
15.41
15.47
15.84
15.84
15.85
15.86
16.33
16.81
16.88
17.29
17.30
17.30
17.32
15.89
16.03
16.04
16.03
16.03
16.03
16.02
17.12
17.41
17.44
17.59
17.59
17.59
17.60
14.9
15.2
15.3
15.5
15.5
15.5
15.5
16.2
16.6
16.7
17.0
17.0
17.0
17.0
9.71
13.51
-6.53
-5.90
11.66
15.29
11.66
11.66
14.82
-5.52
-5.80
13.40
16.55
13.40
15.16
16.14
-9.00
-8.46
17.84
18.66
17.71
14.52
15.97
-9.21
-8.86
17.27
18.61
17.21
2
1.00
2.23
2.61
32.61
35.69
46.83
78.36
11.66
12.56
12.70
13.70
13.70
13.73
13.76
13.40
14.60
14.75
15.72
15.73
15.75
15.78
15.29
15.45
15.47
15.51
15.51
15.51
15.51
16.55
17.84
17.88
18.08
18.08
18.09
18.09
11.7
12.6
12.7
13.7
13.7
13.7
13.7
13.4
14.6
14.8
15.7
15.7
15.7
15.8
11.12
15.38
-11.56
-10.59
14.56
18.53
14.56
11.45
15.54
-7.78
-6.90
13.77
17.60
13.77
a
b
∆Hq and ∆Gq in kcal/mol and ∆Sq in cal/K mol. Effective barrier,
incorporating the contribution of both transition states. c Conformation GS2
3c
d
was used. Conformation GS4 was used.
1.00
2.23
2.61
32.61
35.69
46.83
78.36
17.84
19.17
19.37
20.76
20.77
20.80
20.84
17.27
18.03
18.17
19.04
19.04
19.06
19.09
18.66
19.06
19.14
19.59
19.59
19.60
19.61
18.61
18.70
18.73
18.90
18.90
19.91
18.91
17.7
18.7
18.8
19.5
19.5
19.5
19.5
17.2
17.9
18.0
18.6
18.6
18.6
18.6
The stable conformations were then used to obtain vacuum
rotational barriers (Table 3). As mentioned before, the gas-phase
activation entropies are negative, a behavior that is maintained
on passing to solutions for most cases as demonstrated by the
experimental results. Nevertheless, gas-phase barriers must not
be directly compared to experiments, as it is first necessary to
include condensed-phase effects.
The solvation energy may be dissected into two components,
namely, the bulk solvent polarity and specific solute-solvent
interactions such as hydrogen bonding. The former can be
directly handled by quantum methods, whereas specific interac-
tions require the inclusion of statistical aspects of liquids, a task
accomplished by classical-based protocols (Monte Carlo or
molecular dynamics). Let us first consider the solvent polarity.
For this purpose, we employed the isodensity polarizable
continuum model (IPCM) as described in the Experimental
Section.
The solution rotational barriers were obtained by simply
adding to ∆Gq (from Table 3) the corresponding solvation
energies from single-point calculations of the gas-phase struc-
tures (Table 4). The rotational barrier of 1 increases slightly
when one passes from the gas phase to water (0.6 kcal/mol at
HF and 0.8 kcal/mol at B3LYP). If we consider only those
solvents for which measurements have been done, i.e., CS2 and
D2O, the variation is only 0.2 kcal/mol with HF and 0.3 kcal/
mol with B3LYP. These small variations agree with experi-
ments. Although both methods agree with each other in the
response to the solvent, the B3LYP values are about 2 kcal/
mol above the experimental results (∼15 kcal/mol, Table 1).
Regarding compound 2, the predicted solvent effect is 1.0
kcal/mol for both HF and B3LYP, which agrees with the
experimental results. However, the calculated barriers are, on
average, 1.0 kcal/mol below the experimental values at HF and
about 1.0 kcal/mol above the experimental values at B3LYP.
This result gives evidence that a further correction needs to be
added to the calculated barriers.
4d
1.00
2.23
2.61
32.61
35.69
46.83
78.36
14.56
15.41
15.54
16.55
16.56
16.58
16.62
13.77
14.31
14.40
15.06
15.06
15.08
15.10
18.53
19.65
19.80
20.73
20.74
20.76
20.78
17.60
18.38
18.50
19.11
19.11
19.13
19.14
14.6
15.4
15.5
16.6
16.6
16.6
16.6
13.8
14.3
14.4
15.1
15.1
15.1
15.1
a
Dielectric constants, corrected to 25 °C, correspond to a vacuum,
carbon tetrachloride, carbon disulfide, methanol, acetonitrile, dimethyl
sulfoxide, and water.42 Effective barrier, incorporating the contribution
b
c
d
of both transition states. The conformation GS2 was used.
The
conformation GS4 was used.
to ensure the observation of such an effect. What is most
probable is that weak solute-solvent complexation effects mask
the variation. This interpretation is supported by the measured
entropy values. To see this, let us remember that the gas-phase
theoretical activation entropies are negative (Table 3), whereas
the experimental ones are positive or very close to zero for 3.
These results suggest that some kind of ordered state is broken
on passing from the ground to the transition state, like a solute-
solvent complex. Concerning the rotational barrier magnitude,
neither HF nor B3LYP furnishes satisfactory results using only
IPCM.
In the case of compound 4, when one goes from CCl4 to
H2O, the rotational barrier varies by 1.2 kcal/mol at HF and by
0.8 kcal/mol at B3LYP, in agreement with experiments. The
HF barriers are somewhat above the experimental ones, and
those calculated by B3LYP lie, on average, below the DNMR
measurements.
To summarize, although the IPCM model is able to reproduce
the solvent effect to some extent, it is still necessary to include
explicit solute-solvent interactions. Among these interactions
are those of very strong impacts on the solute properties, i.e.,
hydrogen bonds, and those that are more common among highly
polarizable solvents, such as CS2 or DMSO, which correspond
to somewhat weaker, less geometrically restricted interactions.
When one passes from CCl4 to DMSO, the rotational barrier
for compound 3 varies by 0.8 kcal/mol at HF and by 0.7 kcal/
mol at B3LYP. The experimental variation is 0.4 kcal/mol and
stays within the experimental errors, so that it is not possible
J. Org. Chem, Vol. 72, No. 6, 2007 1905

Pontes et al.
FIGURE 5. Water complexes used in the force-field parametrization of compounds 1-4.
For the present study, we will treat only hydrogen bonding
through molecular dynamics simulations. Whether HF or
B3LYP provides better rotational barriers can be judged only
after including these interactions.
built complexes of the solute with a water molecule placed in
strategic positions and then calculated the complex energy at
the B3LYP/6-31+G(d,p) level (Figure 5). Through this calcula-
tion, the solute was kept at the B3LYP/6-311+G(2d,p) geometry
while the water molecule was fixed with its TIP4P parameters
(r(O-H) ) 0.9752 Å, (H-O-H) ) 104.52°);43 only the
distance between the two molecules was optimized. The partially
optimized geometries were then used to perform a counterpoise
correction for the basis set superposition error (BSSE) and to
obtain interaction energies, ∆Eint. There are three atomic
quantities composing Coulomb and Lennard-Jones potentials,
namely, the electron charge, q, and the two Lennard-Jones
Force-Field Parametrization. Liquid simulations require the
specification of solute-solvent potential interaction functions.
These are defined by Coulomb (electron charges) and Lennard-
Jones (van der Waals) parameters adjusted for every atom or
atom groups.11 Unfortunately, there were no such parameters
for the present compounds; those reported for amides did not
furnish satisfactory results when extended to carbamates and
thiocarbamates in our tests. Therefore, it was necessary to
calibrate the potential interaction functions before running the
liquid simulations. To accomplish this task, we followed a
procedure similar to that used by Jorgensen et al.11 We first
(43) Jorgensen, W. L.; Chandrasekhar, J.; Madura, J. D. J. Chem. Phys.
1983, 79, 926-935.
1906 J. Org. Chem., Vol. 72, No. 6, 2007

Rotational Barrier In Carbamate Congeners
TABLE 5. Solute-Water Complexation Energies (in kcal/mol)
Calculated at the B3LYP/6-31+G(d,p) Level of Theory (DFT) and
through the Fitted Force Field (FF)a
total of 12 simulations, as described in the Experimental Section.
The coordinates stored during MD runs were used to construct
radial distribution functions (rdf) for each pair of solute-solvent
atoms. These functions are constructed such that in structured
portions of the liquid, like those close to the solute, its value
differs from the unit with a maximum indicating a likely region
for finding solvent molecules. In the bulk liquid, far from the
solute, rdf’s approach unit values and no region is more likely
to contain solvent molecules than others. Hydrogen bonding
can be identified by a sharp peak at about 2 Å (O‚‚‚H rdf). The
rdf’s for compounds 1-4 are presented in Figure 6.
1
2
3
4
DFT
FF
DFT
FF
DFT
FF
DFT
FF
GS
CP1 -6.21 -6.32 -4.87 -4.43 -2.79 -2.93 -2.77 -2.68
CP2 -5.91 -5.98 -5.51 -5.89 -2.46 -2.60 -2.42 -2.57
CP3 -1.20 -1.32 -1.07 -0.80 -1.23 -1.42 -0.68 -0.56
CP4 -2.24 -2.37 -2.01 -1.74 -0.51 -1.22 -2.30 -2.30
CP5 -0.79 -0.65 -0.24 -0.84 -0.85 -0.57 -0.28 -0.42
CP6 -0.83 -0.80 -0.97 -1.22 -1.02 -0.63 -1.15 -1.11
rmsb
0.11
0.39
0.37
0.11
Solvation peaks are clearly observed in the GS of 1 for both
CdO‚‚‚O(water) and CdO‚‚‚H(water) rdf’s. The distance
between the maxima of the two rdf’s is 0.95 Å, very close to
the O-H bond length in the TIP4P model, showing that there
is an efficient complexation between the solute and water
molecules at the carbonyl oxygen. On the other hand, there is
no indication of hydrogen bonding at the nitrogen of GS, as
expected, following from the sp2 hybridization of that atom.
The situation is a little different in the TSs. Here, beyond the
bonding at the carbonyl oxygen, water molecules also complex
at the nitrogen which is now sp3 and has a lone pair available
to donate electron density. These results suggest, even before
we quantify the effects, that the rotational barrier of 1 decreases
as a consequence of hydrogen bonding because the TSs have
two binding sites compared to only one in the GS.
TS1
CP1 -5.38 -5.28 -4.64 -4.39 -2.26 -2.16 -2.20 -2.11
CP2 -4.96 -4.90 -4.49 -4.66 -1.58 -1.67 -1.54 -1.77
CP3 -5.91 -5.93 -4.47 -4.61 -5.13 -5.10 -3.99 -3.95
CP4 -1.62 -1.56 -1.35 -1.30 -1.05 -0.50 -1.03 -0.91
CP5 -1.32 -1.54 -0.57 -1.05 -1.21 -1.22 -0.72 -0.86
CP6 -0.54 -0.98 -0.30 -1.01 -0.36 -0.50 -0.21 -0.73
rmsb
0.21
0.38
0.24
0.25
TS2
CP1 -5.36 -5.53 -4.46 -4.34 -2.32 -2.32 -2.29 -2.17
CP2 -4.44 -4.16 -3.92 -4.05 -1.28 -1.29 -1.25 -1.32
CP3 -5.88 -5.55 -5.45 -5.32 -5.68 -5.66 -5.49 -5.07
CP4 -1.04 -0.93 -1.73 -1.28 -0.42 -0.46 -1.36 -1.26
CP5 -1.51 -1.24 -0.74 -1.16 -1.52 -1.66 -0.69 -1.14
CP6 -0.55 -1.11 -0.56 -1.30 -0.39 -0.54 -0.48 -1.02
rmsb
0.32
0.40
0.09
0.34
a
b
Complexes CP1-CP6 are defined in Figure 5. Root mean square
The rdf’s of compound 2 resemble those for 1 in their
qualitative aspects. However, the N‚‚‚H and N‚‚‚O peaks in the
TSs are, relative to 1, less intense. As we are going to see, this
causes the effect of hydrogen bonding in 2 to be the opposite
of that for 1; i.e., the rotational barrier will increase from
complexation to protic solvents.
deviations (in kcal/mol).
parameters accounting for van der Waals forces, ꢀ and σ. These
parameters determine the interaction energy according to the
following equation:
Another aspect that can be explained by rdf’s is the negative
activation entropies in the protic solvents of Table 1. The
measured ∆Sq values in D2O are about twice the calculated gas-
phase values for 1 and 2, which suggest that in the corresponding
transition states water molecules might be more organized
around the solute compared to the ground state. Looking at the
rdf’s, the reason becomes clear, for the GSs of 1 and 2 have
only one protonation site compared to two in the TS, confirming
the more ordered state of the latter. A similar effect could be
invoked to explain the negative entropy in D2O/CD3OD. The
effect in this case, however, is more intense, but without detailed
molecular dynamic data in methanol, we cannot unequivocally
describe this subtle behavior.
∆Eint
)
{q q e2/r + 4ꢀ [(σ /r )12 - (σij/rij)6]} (1)
i j ij ij ij ij
∑ ∑
i
j
in which
ꢀ ) ꢀ ꢀ , σ ) σ σ
x
x
ij
i
j
ij
i
j
Throughout the force-field calculations, we used a model with
each methyl group being described as a single atom, giving a
seven-point representation of the solute molecular structure. A
further simplification was to describe the two CH3’s attached
to N by the same atom type. Lennard-Jones parameters (ꢀ, σ)
were taken from similar compounds,11,44-47 and only the atomic
charges were varied to reproduce DFT interaction energies.
Table 5 presents the results of the adjusted force field together
with the DFT values. The complete set of fitted electron charges
along with the Lennard-Jones parameters used are included
as Supporting Information. The agreement between DFT and
force-field values was satisfactory in most cases, so they can
be considered as appropriate to be used in the liquid simulations.
Radial Distribution Functions. After obtaining the force-
field parameters, we performed molecular dynamics calculations
on water solutions of the four compounds. Because each
compound has three species (GS, TS1, and TS2), there is a
Although for 1 and 2 we observe hydrogen bonding at the
carbonyl oxygen of GS, compounds 3 and 4, contrastingly,
present no signal of complexation at the corresponding double-
bonded sulfur atoms. On the other hand, both transition states
engage in hydrogen bonding to water, although in a smaller
fashion in the case of TS1. Because only the transition states
are stabilized by complexation, we expect the rotational barriers
of 3 and 4 to experience a decrease in protic solvents.
Rotational Barriers with the Inclusion of Hydrogen
Bonding. The following step is to quantify the effect of
hydrogen bonding on the rotational barriers. The solvation peaks
previously identified can be integrated out to the first minimum,
furnishing, in this way, the average number of water molecules
engaging in a water-solute complex. Table 6 presents these
results. Complexation energies were obtained by simply mul-
tiplying the average number of water molecules in the solvation
shell by the corresponding energies in Table 5. For instance,
(44) Jorgensen, W. L. J. Phys. Chem. 1986, 90, 1276-1284.
(45) Jorgensen, W. L. J. Phys. Chem. 1986, 90, 6379-6388.
(46) Jorgensen, W. L.; Madura, J. D.; Swenson, C. J. J. Am. Chem. Soc.
1984, 106, 6638-6646.
(47) Jorgensen, W. L.; Swenson, C. J. J. Am. Chem. Soc. 1985, 107,
569-578.
J. Org. Chem, Vol. 72, No. 6, 2007 1907

Pontes et al.
FIGURE 6. Radial distribution functions for compounds 1-4.
when one water molecule binds the GS of 1 at the carbonyl,
the system gains 6.06 kcal/mol of stability (average between
CP1 and CP2, Table 5); however, 1.7 water molecules, on
average, surround GS according to the rdf’s (Table 6), and thus
the overall stabilization becomes -10.21 kcal/mol. The last
column of Table 6 lists the contribution of hydrogen bonding
to the rotational barriers.
The smallest effect occurs for 1 corresponding to a decrease
in the barrier. Decreases are observed also for 3 and 4 with
higher intensity than those in 1. Only for 2, hydrogen bonding
1908 J. Org. Chem., Vol. 72, No. 6, 2007

Rotational Barrier In Carbamate Congeners
TABLE 6. Analysis of Radial Distribution Functions for
Compounds 1-4
occasion, however, we did not use the thermodynamic quantities
but rather total energies corrected to zero-point vibrations.
CdZ
N
total
∆Eb coord.c
Methyl N,N-Dimethylcarbamate (1)
Let us now analyze the reasons for the above-mentioned
behaviors. As we said while commenting on the rdf’s for
compound 1, the transition states engage in hydrogen bonding
in two places, at the carbonyl and at the nitrogen, and the ground
state accepts bonding only at the carbonyl. Integration of the
corresponding rdf’s (Table 6) shows that these two interactions
in the TSs are responsible for the rotational barrier diminution.
This behavior is readily explained by the resonance model
because the nitrogen passes from sp2 to sp3 hybridization on
going from GS to TS and becomes a good proton acceptor
(Figure 2).
species
coord.a
∆Eb
coord.a
∆Ed
∆∆Ee
f
f
GS
1.69 (2.53) -10.21
TS1 1.08 (2.43) -5.60 0.89 (2.63) -5.28 1.97 -10.88 -0.67
TS2 1.12 (2.38) -5.50 0.92 (2.63) -5.41 2.04 -10.91 -0.70
-
-
1.69 -10.21
S-Methyl N,N-Dimethylthiocarbamate (2)
1.56 (2.48) -8.09
TS1 0.93 (2.38) -4.24 0.44 (2.68) -1.97 1.37
TS2 0.71 (2.18) -2.95 0.84 (2.68) -4.58 1.55
f
f
GS
-
-
1.56
-8.09
-6.21 +1.88
-7.53 +0.56
O-Methyl N,N-Dimethylthiocarbamate (3)
f
f
f
f
f
f
f
f
GS
TS1
TS2
-
-
-
-
-
-
-
-
0.92 (2.73) -4.72 0.92
1.01 (2.68) -5.71 1.01
-4.72 -4.72
-5.71 -5.71
A similar pattern is observed for compound 2, but in this
case, the hydrogen-bonding strength produces a different
outcome. The nitrogen here is not as good of a proton acceptor
as it is for 1 in the TS, so despite the TSs having two protonation
sites, the GS experiences the greater stabilization. It is seen from
Table 6 that the preferred transition state for 2, TS1, forms about
one-half of the hydrogen bonds of the corresponding TS of 1,
0.44 vs 0.89. Indeed, Mulliken analysis gives -0.326 e for the
nitrogen in the TS1 of 1, compared to -0.058 e for the nitrogen
in the TS1 of 2 (at HF/6-311+G(2d,p)). Moreover, the large
sulfur atom of 2 makes the approach of water molecules at TS1
difficult and also contributes to lowering its proton-acceptor
ability.
Methyl N,N-Dimethyldithiocarbamate (4)
f
f
f
f
f
f
f
f
GS
TS1
TS2
-
-
-
-
-
-
-
-
0.36 (2.73) -1.44 0.36
0.93 (2.68) -5.10 0.93
-1.44 -1.44
-5.10 -5.10
a
Number of water molecules coordinated to solute obtained by
integration of the (Cd)Z‚‚‚H(water) rdf; cutoff radii are presented in
parentheses and in Å. b Solute-water complexation energies (in kcal/mol)
calculated on the basis of the data of Table 5 and the above coordination
c
numbers. Total number of coordinated water molecules obtained by
summing up the contributions of (Cd)Z and N. d Total complexation energy
(in kcal/mol) obtained by summing up the contributions of (Cd)Z and N.
e
Total contribution of hydrogen bonding (in kcal/mol) to the rotational
barriers. No solvation shell observed for these rdf’s.
f
Consider now compounds 3 and 4. The rotational barrier
decreases for both as a consequence of hydrogen bonding, and
the effect here is substantially larger than in 1. For these
compounds, as mentioned before, there is no hydrogen bonding
with the double-bonded sulfur atom in either GS or TS, so that
the complexation at the nitrogen of TS causes the barrier to
decrease. Similarly to what we mentioned above for the TS1
of 2, the additional sulfur atom of 4 repels the approach of water
molecules to the nitrogen region, making hydrogen bonding in
the TS1 of 4 less effective. When these effects are added to the
IPCM results, the calculated rotational barrier of 4 is found to
be in satisfactory agreement with the experimental value.
Unfortunately, there is no experimental data for compound 3
in water. Given the good performance of IPCM + MD for the
other compounds, we should expect a rotational barrier for 3 in
water in the vicinity of 14 kcal/mol, about 4 kcal/mol lower
than the value in DMSO (Table 4).
TABLE 7. Rotational Barriers in Water (in kcal/mol) for
Compounds 1-4 Calculated Using the Combination IPCM + MD
TS1
B3LYP
TS2
B3LYP
eff.a
HF
HF
B3LYP
HF
exptl
1
2
3
4
17.00
17.99
14.39
13.66
15.19
15.64
16.12
15.18
17.25
19.08
13.24
14.00
15.32
16.07
13.90
15.68
16.7
17.9
13.2
13.4
14.8 14.8 ( 1.4
15.4 14.8 ( 0.7
13.9
-
15.0 15.9 ( 0.3
a
Effective barriers, including the contribution of both transition states.
contributes to an increase in the rotational barrier. These results
contrast with those of amides which have their barriers raised
in protic solvents,19 and as far as we are aware, there is no
previous report of a negative solvent effect on the rotational
barrier of an amide-like system.
Table 7 lists the final rotational barriers calculated by the
IPCM + MD combination. These values were obtained by
adding the effect of hydrogen bonding just mentioned to the
data of Table 4. The B3LYP values now diverge greatly from
experimental measurements, whereas HF presents a very
satisfactory agreement. Before including both sets of effectss
solvent polarity and specific interactionssit was not possible
to correctly judge the performance of each method. It can be
seen now that the inclusion of electron correlation through DFT
is not advantageous for the calculation of rotational barriers.
Actually, this was noted before by Wiberg et al.2 for gas-phase
calculations. Additionally, Jasien et al.48 found that electron
correlation is not essential for calculating rotational barriers in
amides; their analyses were based on MP2 calculations. We have
conducted an earlier study26 on the rotational barrier of
carbamates using DFT-B3LYP in which good agreement was
obtained between calculated and experimental values. In that
Dipole Moments. As a final step, let us consider the
connection between dipole moments and the response to the
solvent polarity. Because HF and B3LYP agree with each other
about the variations in the IPCM results, we think it is advisable
to analyze dipole moments calculated through both methods,
and these are presented in Table 8. Ground-state dipole moments
for 1 and 2 are very similar, as also are the variations when
passing to TS1. Therefore, similar solvation effects should be
expected for these species, but this does not happen because
the data in Table 4 clearly show a far more pronounced effect
in 2, about 2.0 kcal/mol compared to only 0.4 kcal/mol in 1
(vacuum to water). Neither GS dipole moments nor variations
in dipole moments can alone account for the differences between
these compounds.
Now, let us look at the rotation GS f TS1 for compounds
3 and 4. In this case, both µ and ∆µ are greater for 3, the
compound for which the solvent effect is also larger (+3.00
kcal/mol compared to +2.06 kcal/mol on going from ꢀ ) 1.00
to ꢀ ) 78.36, Table 4). We see, therefore, that considering the
(48) Jasien, P. G.; Stevens, W. J.; Krauss, M. J. Mol. Struct. (Theochem.)
1986, 139, 197-206.
J. Org. Chem, Vol. 72, No. 6, 2007 1909

Pontes et al.
TABLE 8. Calculated Dipole Moments and Dipole Moment
TABLE 9. Partition of Energy Terms Contributing to the Solvent
Effect According to the Spherical Cavity Approximation (ao in Å
and the Remaining Values in kcal/mol)
Differences (in D) for Compounds 1-4a,b
µ (GS)
B3LYP
∆m (TS1)c
∆m (TS2)c
ao
µ
total
total
HF
HF B3LYP
HF
B3LYP
(TS1) calcd in -2φ(ꢀ,ao)µGS∆µ -φ(ꢀ,ao)(∆µ)2 dipole IPCM
1
1
vacuum
water
2.56
3.61
2.48
3.67
-1.65
-2.21
-1.62
-2.30
+0.45
+0.44
+0.18
+0.23
4.09
4.11
4.36
4.33
vacuum
water
0.87
1.65
-0.28
-0.50
+0.59 +0.89
+1.15
2
2
3
4
vacuum
water
2.25
3.63
2.31
3.80
-1.66
-2.28
-1.59
-2.24
-0.21
-0.16
-0.19
-0.29
vacuum
water
0.76
1.68
-0.28
-0.53
+0.48 +2.10
+1.15
3
vacuum
water
4.28
6.04
3.68
5.32
-2.52
-3.82
-2.29
-3.73
-0.72
-1.29
-0.71
-1.32
vacuum
water
1.84
3.93
-0.54
-1.24
+1.30 +3.00
+2.69
4
vacuum
water
3.74
5.34
3.31
4.75
-2.55
-3.19
-2.09
-2.54
-1.12
-1.77
-0.89
-1.32
vacuum
water
1.66
2.97
-0.57
-0.89
+1.10 +2.06
+2.08
a
b
c
GS2 for 3 and GS4 for 4. Basis set 6-311+G(2d,p). Variation on
going from GS to TS [∆µ ) µ(TS) - µ(GS)].
but it fails for 2 probably due to the importance of higher
multipole moments or cavity approximations. The correct trends
in the solvent effect are reproduced with dipole moments from
the gas-phase structures, but only after using the dipoles from
a previous solvated structure (IPCM) can we obtain a reasonable
estimate of the effect on the rotational barrier. The ∆µ values
are all negative for the rotation GS f TS1 implying that the
first term of eq 5 will raise the barrier for compounds 1-4.
The largest effect occurs for 3, which has the largest µ and ∆µ.
The smallest effect is observed for 1, for which both µ and ∆µ
assume their lowest values.
quantities µ or ∆µ, as we did above, may or may not lead to
the correct result for the solvent effect trend. To put things in
a more systematic scheme, we can use the approximation
introduced by Onsager,25 writing the solvation energy of a given
molecule in terms of its dipole moment, µ, and of a spherical
cavity of radius, ao:
ꢀ - 1 µ2
∆Gsolv ) -
(2)
3
2ꢀ + 1
ao
If the dipoles of GS and TS differ by ∆µ, the TS dipole
moment can be written as µTS ) µGS + ∆µ. The solvation
energy for TS then becomes
Conclusions
Rotational barriers in amide-like systems are generally
affected by solvation. Nevertheless, the effect can sometimes
be too small to detect by experimental procedures, as is the
case with carbamates. For carbamates, O-alkyl thiocarbamates,
and dithiocarbamates, protic solvents decrease the barrier
through hydrogen bonding. S-Alkyl thiocarbamates, however,
have their barrier raised in protic media. Carbamates experience
only a little decrease in their rotational barrier due to hydrogen
bonding because the two protonation sites at the transition states,
the carbonyl and the nitrogen, are roughly balanced by hydrogen
bonding at the carbonyl of the ground state. The inability of
the sulfur atom to form hydrogen bonds is responsible for the
large decrease observed for compounds 3 and 4 in water
solutions because in these cases only the transition states will
be stabilized by hydrogen bonding.
∆GTS ) -φ(ꢀ,ao)µGS2 - 2φ(ꢀ,ao)µGS∆µ - φ(ꢀ,ao)(∆µ)2
solv
(3)
where
ꢀ - 1 1
φ(ꢀ,ao) )
3
2ꢀ + 1
ao
The solvent effect for the process GS f TS is simply defined
by
TS
solv
GS
solv
∆∆Gsolv ) ∆G - ∆G
(4)
Combining the continuum solvation method IPCM with
molecular dynamics proved to be a suitable protocol for
analyzing solvation effects in solvents that engage in hydrogen
bonding with the solute. Rotational barriers for carbamate
congeners are better calculated through Hartree-Fock than with
the B3LYP method. Even so, on the basis of the final calculated
barriers, B3LYP furnished satisfactory results when used to
obtain solute-solvent complexation energies. Solute-solvent
complexation energies calculated at the B3LYP/6-311+G(2d,p)
level, statistically corrected by radial distribution functions, can
be used to quantify hydrogen-bonding effects on the rotational
barriers. The final calculated results seem to slightly overesti-
mate the experimental effects but are still reliable in reproducing
the observed barriers and their response to the medium.
Replacing oxygen with a sulfur atom can cause two effects,
depending on where the replacement takes place. In particular,
the sulfur atom in 2 hinders the approach of a water molecule
Under the assumption that the molecular radius does not vary
too much from GS to TS, we finally arrive at
∆∆Gsolv ) -2φ(ꢀ,ao)µGS∆µ - φ(ꢀ,ao)(∆µ)2
(5)
The first thing to be noted when analyzing solvation effects
in terms of dipole moments (µ) or dipole moment variations
(∆µ) is that both quantities must be considered simultaneously,
and eq 5 gives us an idea of how µ and ∆µ relate to the solvent
effect on rotational barriers. If we have two compounds with a
similar µ, they will behave as distinctly as the differences in
their ∆µ. If, instead, the ∆µ are similar, then the relative
behavior will be dictated by µGS. The contributions of each
member of eq 5 are listed in Table 9.
The simple approximations introduced above work well for
compounds 1, 3, and 4, as can be seen by comparing the values
of the dipole moment model to those calculated with IPCM,
1910 J. Org. Chem., Vol. 72, No. 6, 2007

Rotational Barrier In Carbamate Congeners
Geometries were optimized using the restricted Hartree-Fock
(RHF) method as well as the hybrid B3LYP density functional,
both with the 6-311+G(2d,p) basis set. A subsequent frequency
calculation characterized the stationary points (one imaginary
frequency for transition states and zero for ground states), from
which we also obtained thermodynamics quantities.56 The effect
of solvent polarity was included through the isodensity polarizable
continuum model (IPCM)57 at the HF/6-311+G(2d,p) and B3LYP/
6-311+G(2d,p) levels by single-point calculations over the corre-
sponding optimized structures. The IPCM solvation model had been
used in related studies and proved to be suitable for representing
the energy variations due to bulk solvent polarity in rotational
barriers.2,23,24 In this model, the solute is placed in a cavity defined
by an isosurface of the total electron density (typically 0.0004
e/bohr2). The model not only treats dipole moments but also is
equivalent to going to an infinite order in a multipole expansion.57
Molecular dynamics simulations of water solutions were carried
out with the TINKER58 package of programs. The TIP4P model43
was employed to describe the water molecules, whereas for solutes,
it was necessary to conduct a force-field parametrization (as detailed
in Results and Discussion). Rigid models were used. Simulations
were performed in the NVT ensemble with 296 water molecules
plus the corresponding solute in a pre-equilibrated box with
dimensions 20.80 × 20.80 × 20.80 Å. After equilibrating the box
at 298 K for 20 ps, production periods lasted 300 ps with a
temperature couple parameter of 0.3 ps and a time step of 0.001
ps.
to the nitrogen of TS1 and causes this transition state to lose
some of the stabilization energy it could gain from hydrogen
bonding. This contributes to increasing the barrier corresponding
to TS1, which is the preferred transition state of 2.
Three of the four compounds studied behave in a way
explainable by dipole moments and dipole moment variations.
Compound 2 is the exception and is a good example for not
extrapolating conclusions before a careful evaluation of a wide
range of structural systems.
Experimental Section
Syntheses. Compounds were obtained following the procedures
described by Yoder et al.4949 Methyl N,N-dimethylcarbamate (1)
was prepared by the reaction of sodium methoxide with N,N-
dimethylcarbamoyl chloride in tetrahydrofuran (THF) (bp 128 °C/
1
∼760 Torr, lit.49 130-132 °C/∼760 Torr). H NMR (300 MHz,
CDCl3) δ: 3.69 (3H, s); 2.91 (6H, s).49,50
S-Methyl N,N-dimethylthiocarbamate (2) was prepared by the
reaction of sodium thiomethoxide with N,N-dimethylthiocarbamoyl
chloride in tetrahydrofuran. Sodium thiomethoxide was obtained
from the isothiouronium salt procedure51 by bubbling methyl
mercaptan in a mixture of THF and Nao (bp 54 °C/>6 Torr, lit.49
184 °C/∼760 Torr). 1H NMR (300 MHz, CDCl3) δ: 2.96 (6H, s);
2.29 (3H, s).
O-Methyl N,N-dimethylthiocarbamate (3) was prepared by the
reaction of sodium methoxide with N,N-dimethylthiocarbamoyl
1
chloride in THF (bp 76 °C/>7 Torr, lit.49 87-92 °C/12 Torr). H
Acknowledgment. The authors are thankful to Centro
Nacional de Processamento de Alto Desempenho (CENAPAD-
SP) for the computer facilities and to Fundac¸a˜o de Amparo a`
Pesquisa do Estado de Sa˜o Paulo (FAPESP) for financial
support. We also give thanks to Coordenac¸a˜o de Aperfeic¸oa-
mento de Pessoal de N´ıvel Superior (CAPES) for a scholarship
to R. M. Pontes and to Conselho Nacional de Pesquisa (CNPq)
for a fellowship to E. A. Basso.
NMR (300 MHz, CDCl3) δ: 4.02 (3H, s); 3.74 (3H, s); 3.12 (3H,
s).
Methyl N,N-dimethyldithiocarbamate (4) was prepared by the
reaction of N,N-dimethylamine with methyl iodide and carbon
1
disulfide (mp 43-44 °C, lit.49 45-47 °C). H NMR (300 MHz,
CDCl3) δ: 3.57 (3H, s); 3.38 (3H, s); 2.65 (3H, s).
NMR Measurements. A 300 MHz spectrometer was used to
1
acquire H spectra. Samples were prepared by placing 15 µL (for
liquid) or 20 mg (for solid) of the compound in 0.7 mL of the
appropriate solvent in 5-mm o.d. NMR tubes. In the case of carbon
disulfide, acetone-d6 was used as an external reference. Typical
conditions were a sweep width of 2500 Hz, a pulse length of 6.7
µs, 16 scans, and 1 s for the delay time. 32 K of data points were
used for acquisition with further zero filling to 64 K. Line
broadening was not applied. Deuterated solvents were obtained
commercially and used as received, and CS2 was distilled and stored
under molecular sieves prior to use. The variable-temperature probe
was calibrated against vacuum-sealed methanol (-70 to 15 °C) and
ethylene glycol (20-80 °C) standards.52,53 Total line shape analyses
(TLSA) were accomplished using the WINDNMR54 software.
Computational Details. Electronic structure calculations were
conducted using the GAUSSIAN 9855 package of programs.
Supporting Information Available: 1H NMR spectra and
Z-matrices for the optimized structures and force-field parameters
for MD simulations. This material is available free of charge via
the Internet at http://pubs.acs.org.
JO061934U
(55) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb,
M. A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A., Jr.;
Stratmann, R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A.
D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi,
M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.;
Ochterski, J.; Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Malick,
D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.;
Ortiz, J. V.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi,
I.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.;
Peng, C. Y.; Nanayakkara, A.; Gonzalez, C.; Challacombe, M.; Gill, P. M.
W.; Johnson, B. G.; Chen, W.; Wong, M. W.; Andres, J. L.; Head-Gordon,
M.; Replogle, E. S.; Pople, J. A. Gaussian 98, revision A.11.2; Gaussian,
Inc.: Pittsburgh, PA, 1998.
(49) Yoder, H. C.; Komoriya, A.; Kochanowski, J. E.; Suydam, F. H. J.
Am. Chem. Soc. 1971, 93, 6515-6518.
(50) Basso, E. A.; Oliveira, P. R. O.; Caetano, J.; Schuquel, I. T. A. J.
Braz. Chem. Soc. 2001, 12, 215-222.
(51) Frank, R. L.; Smith, P. V. J. Am. Chem. Soc. 1946, 68, 2103-
2104.
(56) Foresman, J. B.; Frisch, Æ. Exploring Chemistry with Electronic
Structure Methods, 2nd ed.; Gaussian, Inc.: Pittsburgh, 1996.
(52) Greet, A. L. v. Anal. Chem. 1970, 42, 679.
(53) Raidford, D. S.; Fisk, C. L.; Becker, E. D. Anal. Chem. 1979, 51,
2050.
(57) Foresman, J. B.; Keith, T. A.; Wiberg, K. B.; Snoonian, J.; Frish,
M. J. J. Phys. Chem. 1996, 100, 16098-16104.
(58) Ponder, J. W. TINKER: Software tools for molecular design; version
4.1; Washington University School of Medicine: St. Louis, 2003.
(54) Reich, H. J. J. Chem. Educ. Software 3D2.
J. Org. Chem, Vol. 72, No. 6, 2007 1911