Relative Basicities of Some Endo and Exo Norbornylamines
J . Org. Chem., Vol. 62, No. 21, 1997 7207
hydrogen bond basicity factor alone, the pKa values of
the 2-aminonorbornanes should be lower in DMF solvent.
This is true in the cases of all of the amines studied
except that of 2-morpholinonorbornane and N-methyl-
morpholine.
as the negative of the gas phase heat of reaction of
B(g) + H+(g) f BH+(g)
that is,
The reason for this “morpholine effect” can be ascer-
tained by observing the dipole moments of the amine
salts (see Table 2). The calculated dipole moments of
these amines correspond very closely to the experimental
values in four test cases including those of the morpho-
lines. So it is assumed that all of the ab initio (HF/6-
31G*) and semiempirical (AM1) calculated dipole mo-
ments are close to the actual values. The large dipole
moments of the salts of pyrrolidino-, hexamethylenimino-,
and piperidinonorbornane show that they possess very
localized charges. So they require much greater solvation
than the 2-morpholinonorbornane salts which have much
smaller dipole moments and hence a much more diffuse
charge. This more diffuse charge distribution is due to
the presence of the ring oxygen atom which causes the
2-morpholinium norbornane salt to be more polarizable
than the other salts. This is shown by its much smaller
dipole moment than would be anticipated if the positive
charge were more localized. Solvents with high polar-
izability are good solvators for highly polarizable cat-
ions.10 The polarizability of DMF and AN can be deter-
mined through the Lorenz-Lorentz equation from
refractive indexes. Solvents with a large refractive index
have strong dispersion forces. The refractive index at
20 °C for DMF and AN are 1.4305 and 1.3441, respec-
tively.10 These values correspond to polarizabilities (R0)
of 1.06 × 10-23 and 4.39 × 10-24, respectively.10 This
difference in solvents is shown also by the fact that the
solvent polarity-polarizability scale of DMF is 0.954
whereas that of AN is 0.895.11 The 2-morpholinonorbor-
nane aminium cations are more solvated in more polariz-
able solvents like DMF than in a less polarizable solvent
like AN. Hence this cation is more stable in DMF than
in AN, and the acid/base equilibrium for 2-morpholinon-
orbornane is shifted more toward the conjugate acid side
in DMF, resulting in its being a stronger base in this
solvent. A similar effect has been reported in comparing
the relative basicities of both morpholine and DABCO
(1,4-diazabicyclo[2.2.2]octane) with those of triethy-
lamine, pyrrolidine, and piperidine in different solvents.12
These are both 1,4-bifunctional molecules. The solvents
used were water and DMSO. The former is a polar protic
solvent, and the latter is an aprotic, dipolar protophilic
solvent5,9 like DMF.
PA ) -[∆Hf(BH+) - ∆Hf(B) - ∆Hf(H+)]
The experimental value for the heat of formation for
the proton, ∆Hf(H+) ) 367.2 kcal/mol,13 was used in the
semiempirical calculations.
The semiempirical methods used were AM114,15 and
PM3.16 These methods were used to calculate the heats
of formation of the free amines and their aminium salts
along with their dipole moments. Full geometric opti-
mizations were carried out for these molecules including
the use of the lowest energy conformation about the
bicyclic-C to N bond. The results of these calculations
are shown in Table 3. All of the PA’s tend to be smaller
with PM3 calculations than with AM1 calculations, and
the two experimental PA’s that have been reported (endo-
and exo-2-aminonorbornane1) have values closer to those
of AM1. This is consistent with the comparison of these
two methods for calculating PA’s made by Burk and co-
workers.17
Using ab initio methods, the proton affinity is deter-
el
mined by with ∆E0 the electronic energy difference of
el
PA ) -∆E0 - ∆ZPE + 5/2RT
E(BH+) - E(B), ∆ZPE the zero point energy difference,
and 5/2RT the H+ contribution to CP.
The ab initio calculations were made on HF level using
STO-3G, 3-21G*, and 6-31G* basis sets. Two calculations
were made at the BLYP level which includes some
estimate of correlation energy. These methods were used
to calculate the total electronic energies as well as dipole
moments. Again full geometric optimizations were car-
ried out for these molecules. The lowest energy confor-
mation about the bicyclic-C to N bond was used in these
calculations. The change in zero point energy between
the free amines and the corresponding salts (from STO-
3G calculations on exo- and endo-2-aminonorbornane)
was determined to be 9.5 kcal/mol, so this number was
used throughout the calculations. The results of these
calculations are shown in Table 3. The ab initio calcu-
lated gas phase PA’s show the same trends as those
observed in the experimental solution pKa’s; namely, the
exo isomers have larger PA’s than the corresponding endo
isomers. However, the semiempirical calculations have
these reversed. Both methods show the morpholine
bicyclic amines have smaller PA’s than the other three
tertiary bicyclic amines. The agreement between the
measured proton affinity and the ab initio calculations
improved dramatically as we went to larger basis sets.
The calculated hybridization for the lone electron pair
on the nitrogen atoms show greater p character for the
exo isomers than for the corresponding endo isomers (87
to 99% p character). This also leads to the prediction
This “morpholine effect’ was further demonstrated by
using another aprotic dipolar protophilic solvent for some
pKa determinations of exo-2-morpholinonorbornane (3d )
and endo-2-pyrrolidinonorbornane (4a ). This additional
solvent was 1-methyl-2-pyrrolidone with a refractive
index10 at 20 °C of 1.4700, and a polarizability of 1.06 ×
10-23. The pKaof 3d in this solvent is 6.03 ( 0.08, a pKa
increase of 0.28 units over AN solvent. The pKa of 4a in
this solvent is 7.02 ( 0.04, a pKa decrease of 1.22 units
over AN.
Theoretical proton affinities (PA) for each of the bicyclic
amines were determined using both semiempirical and
ab initio methods. Proton affinity of a base, B, is defined
(13) Lias, S. G.; Bartness, J . E.; Liebman, J . F.; Holmes, J . L.; Levin,
R. D.; Mallard, W. G. J . Phys. Chem. Ref. Data 1988, 17, Suppl. No. 1.
(14) Dewar, M. J . S.; Zoebisch, E. G.; Healy, E. F.; Stewart, J . J . P.
J . Am.Chem. Soc. 1985, 107, 3902.
(10) Solvents and Solvent Effects in Organic Chemistry; Reichardt,
C., Ed.; VCH Press: Weinheim, 1988; p 13.
(11) Catalan, J .; Lopez, V.; Perez, P.; Martin-Villamil, R.; Rodriguez,
J . G. Liebigs Ann. Chem. 1995, 241.
(15) Dewar, M. J . S.; Dieter, K. M. J . Am. Chem. Soc. 1986, 108,
8075.
(16) Stewart, J . J . P. J . Comput. Chem. 1989, 10, 209.
(17) Burk, P.; Herodes, K.; Koppel, Iv.; Koppel, Il. Int. J . Quantum
Chem., Quantum Chem. Symp. 1993, 27, 633.
(12) Crampton, M. R.; Robotham, I. A. J . Chem. Res. (S) 1997, 22.
Cook, A. Gilbert
