The prediction of1H chemical shifts in amines: A semiempirical and ab initio investigation

Article abstract of DOI:10.1002/mrc.2043

Twenty one conformationally fixed amines and their N,N-dimethyl derivatives were obtained commercially or synthesized. These included cis and trans 4-o-butyl cyclohexylamine, 2-exo and 2-endo norbornylamine, 2-adamantylamine, 4-phenylpiperidine, 1-napthyl

Full text of DOI:10.1002/mrc.2043

MAGNETIC RESONANCE IN CHEMISTRY
Magn. Reson. Chem. 2007; 45: 749–757
Published online 19 July 2007 in Wiley InterScience
(www.interscience.wiley.com) DOI: 10.1002/mrc.2043
The prediction of 1H chemical shifts in amines:
a semiempirical and ab initio investigation†
Ernani A. Basso,1 Gisele F. Gauze1 and Raymond J. Abraham2∗
1
Chemistry Department, State University of Maringa´ , CEP-87020-900, Maringa´ , PR, Brazil
Chemistry Department, The University of Liverpool, Liverpool L69 3BX, UK
2
Received 14 May 2007; Accepted 24 May 2007
Twenty one conformationally fixed amines and theirN,N-dimethyl derivatives were obtained commercially
or synthesized. These included cis and trans 4-t-butyl cyclohexylamine, 2-exo and 2-endo norbornylamine,
2-adamantylamine, 4-phenylpiperidine, 1-napthylamine and tetrahydro-1-napthylamine. The 1H NMR
spectra of these amines were measured in CDCl3 solution, assigned and the 1H chemical shifts given. This
1
data was used to investigate the effect of the amino group on the H chemical shifts in these molecules.
These effects were analyzed using the CHARGE model. This calculates the electric field and steric effects
of the amino group for protons more than three bonds removed, together with functions for the calculation
of two-bond and three-bond effects. The rotational isomerism about the C–N bond of the amino group was
investigated by ab initio calculations of the potential energy surface (PES) about this bond at the HF/3–21G
level. The resulting conformers were then minimized at the B3LYP/6-311++G (d,p) level. These geometries
were then used to calculate the 1H chemical shifts in the above compounds by CHARGE and the ab initio
gauge-invariant atomic orbital (GIAO) method at the B3LYP/6–311++G(d,p) level and the shifts were
1
compared with those observed. The compounds investigated gave 170 H chemical shifts ranging from
0.60 to 8.2 ppm. The rms errors (obs. – calc.) were ca 0.1 ppm (CHARGE) and ca 0.2 ppm (GIAO). Large
deviations of ca 1.0 ppm were observed for the NH protons in the GIAO calculations. The complex spectra
of alkyl and aryl amines can thus be successfully predicted by both ab initio and semiempirical methods
except for the NH protons, for which the ab initio calculations are not sufficiently accurate. Copyright 
2007 John Wiley & Sons, Ltd.
KEYWORDS: NMR; 1H chemical shifts; amines; NMR prediction; GIAO; CHARGE
chemical shifts and noted that it was a function of the ring
size and the nitrogen substitution.
INTRODUCTION
Amines are one of the most important classes of organic
compounds, being constituents of all alkaloids, amino acids,
proteins, etc., and many other natural products. Also, many
drugs are amines (e.g. amphetamines, morphine, nicotine,
etc.). They are also versatile synthetic intermediates in
organic and organometallic chemistry. Despite this common
There are two possible reasons for the absence of any
systematic investigation into amino chemical shifts. One is
the strong basicity of alkyl amines. In a recent investigation,
1
Abraham et al.4 found that the H shifts of the NH protons
of aliphatic amines showed little consistency in either CDCl3
or DMSO solvent and often the signals Cwere too broad
to be observed. They note that the NH2 proton signals
of dimethylamine hydrochloride occur at 9.2 ppm in CDCl3
compared to those of the free base at ca 1 ppm,5 and therefore
traces of acid will produce large shifts for these protons and
also broaden the signal owing to exchange. They suggested
that this was the most probable reason for the lack of
consistency of these shifts in either solvent and concluded
that these protons are not suitable for diagnostic purposes.
Other near protons would be expected to experience similar
but much smaller effects because of protonation. Aryl amines
are much less basic and we shall show that the chemical shifts
of the NH protons in these compounds can be satisfactorily
predicted once all the factors affecting their chemical shifts,
including ꢀ contributions, are evaluated.
1
occurrence, there are few studies on the H NMR chemical
shifts of the amino group, and a definitive analysis of amino
substituent chemical shifts (SCS) in 1H spectra has not been
performed to date. Zurcher’s1 pioneering investigation did
not consider amines, probably because of lack of good data.
Pretsch2 reported literature data for some amines, mainly of
attached methyl groups. More recently, Alkorta and Elguero3
in their theoretical (GIAO) calculations of H shieldings in
amines gave literature data for the amines they investigated
(methyl and ethylamine and some heterocyclic amines). They
1
1
calculated the influence of the nitrogen lone pair on the H
ŁCorrespondence to: Raymond J. Abraham, Chemistry
Department, The University of Liverpool, PO Box 147, Liverpool
L69 3BX, UK. E-mail: abrahamr@liv.ac.uk
The other factor is simply the trivalent nature of the
nitrogen atom. This gives additional degrees of freedom
†This is part 26 of 1H Chemical Shifts in NMR series.
Copyright  2007 John Wiley & Sons, Ltd.

750
E. A. Basso, G. F. Gauze and R. J. Abraham
when compared to monovalent substituents. For example,
the SCS of the amino group is dependent on the position of
the lone pair.3 In solution, the NH2 group is usually not in a
single orientation, but the precise populations of the different
conformers are often unknown. For example, ethylamine has
two distinct conformers with the lone pair anti or gauche to the
C–C bond but the conformer energies in solution have not
to our knowledge been measured accurately. Gas-phase ab
initio calculations by Alkorta and Elguero3 gave ethylamine
as a 50 : 50 mixture but solvent effects were not considered.
In more complex acyclic amines, the number of conformers
is prohibitively large for any quantitative calculation.
The theory distinguishes between short-range substituent
effects over one, two, and three bonds, which are attributed
to the electronic effects of the substituents and long-range
effects due to the electric fields, steric effects and anisotropy
of the substituents.
The CHARGE scheme calculates the effects of neighbor-
ing atoms on the partial atomic charge of the atom under
consideration on the basis of the classical concepts of induc-
tive and resonance contributions. If we consider an atom I in
a four-atom fragment, I-J-K-L, the partial atomic charge on
I (qI) is due to three effects. There is an ˛-effect from atom
J, given by the difference in the electronegativity of atoms I
and J and a ˇ-effect from atom K proportional to both the
electronegativity of atom K and the polarizability of atom I.
There is also a ꢁ-effect from atom L given by the product of
the atomic polarizabilities of atoms I and L for I D H and
L D F, Cl, Br, I. However for chain atoms (C, N, O, S etc.)
the ꢁ effect (i.e. C–C–C–H) is parameterized separately and
is given by A C B cos ꢂ where ꢂ is the C–C–C–H dihedral
angle and A and B empirical parameters.
For this reason we have investigated a variety of cyclic
amines in which the carbon framework is in a rigid
conformation (Fig. 1), and determined the orientation of the
amino group by ab initio calculations. These molecules give
1
rise to complex H NMR spectra, which can, however, be
completely assigned to give an ensemble of 1H chemical
shifts. This ensemble provides a stringent critique for any
calculation of 1H chemical shifts in these molecules and will
be used to parameterize the CHARGE model for the amino
group.
The total charge is given by summing these effects and
the partial atomic charges (qI) converted to shift values using
the equation
In previous parts of this series, the CHARGE program
has been developed to provide a model capable of accurately
predicting the 1H chemical shifts of a variety of organic
compounds in CDCl36–8 and DMSO4 solutions. This program
consists of a neural-network-data-based approach for one-
, two- and three-bond substituent effects plus a classical
calculation of the long-range effects of substituents. This
calculates the electric field, anisotropy and steric effect of
the substituent plus, for conjugated systems, the effect of
the ꢀ-electrons and for aromatics the ring-current shifts. The
υ D 160.84qI ꢀ 6.68
ꢃ1ꢄ
The effects of distant atoms on the proton chemical shifts
are due to steric, anisotropic and electric field contributions.
Ð Ð Ð
H
H steric interactions are shielding in alkanes and
Ð Ð Ð
deshielding in aromatics and X H (X D C, O, N, Cl,
Br, I) interactions deshielding, according to a simple rꢀ6
dependence, where aS is the steric coefficient for any given
atom.
CHARGE program is available as part of NMRPredict,9
modeling 1H and 13C software package.
a
υsteric D aS/r6
ꢃ2ꢄ
An alternative method of calculating NMR chemical
shifts is by the ab initio gauge-invariant atomic orbital
(GIAO)10,11 method, in which the nuclear shielding tensor
is calculated. This method has been used successfully in
the calculation of heavy-atom chemical shifts. Pulay et al.12
in a discussion of the GIAO method noted that since the
chemical shift range of 1H is the smallest of all atoms, it will
be very sensitive to variation in the methodology such as the
geometry and basis set. An investigation by Lampert et al.13
on phenol and benzaldehyde derivatives led to deviations
of ca 0.5–1.0 ppm depending on the procedure and basis
set used, and this may well represent the limit of accuracy
of such calculations. Alkorta and Elguero3 calculated amine
shieldings at the GIAO/B3LYP/6–311CCGŁŁ level but noted
The effects of the electric field of the C–X bonds (X D H,
F, Cl, Br, I, O, N) on the C–H protons are obtained from
the component of the electric field along the C–H bond. The
electric field for a single bonded atom (e.g.–O) is calculated
as due to the charge on the oxygen atom and an equal and
opposite charge on the attached carbon atom. The vector sum
gives the total electric field at the proton and the component
of this field along the C–H bond is proportional to the proton
chemical shift.
The magnetic anisotropy of a bond with cylindrical
symmetry (e.g. C C) is obtained from the appropriate
McConnell14 equation:
υanis D ꢅꢃ3 cos2 ϕ ꢀ 1ꢄ/3R3
ꢃ3ꢄ
1
that their H calculations were not as well correlated with
the observed shifts as the analogous 13C calculations. We use
here an approach similar to that of Alkorta and Elguero in
that the same density function theory and basis set will be
used for the geometry optimization and GIAO calculations.
The GIAO-derived chemical shifts will then be compared
with the shifts from the CHARGE parameterization.
In Eqn (3) R is the distance from the perturbing group to
the nucleus of interest in A, ϕ is the angle between the vector
R and the symmetry axis and ꢅ the molar anisotropy of the
˚
C
C bond. ꢅ D ꢅparl ꢀ ꢅperp where ꢅparl and ꢅperp are the
susceptibilities parallel and perpendicular to the symmetry
axis, respectively.
For a nonsymmetric group such as the carbonyl group,
Eqn (3) is replaced by the full McConnell equation:
THEORY
As the theory has been given previously,6–8 only a brief sum-
mary of the latest version (CHARGE8c) will be given here.
υanis D [ꢅparlꢃ3 cos2 ꢂ1 ꢀ 1ꢄ C ꢅperpꢃ3 cos2 ꢂ2 ꢀ 1ꢄ]/3R3
ꢃ4ꢄ
Copyright  2007 John Wiley & Sons, Ltd.
Magn. Reson. Chem. 2007; 45: 749–757
DOI: 10.1002/mrc

Prediction of 1H chemical shifts in amines: Part 26
751
Figure 1. Amino compounds used for parameterization of CHARGE 177 ð 203 mm (600 ð 600 DPI).
where ꢂ1 and ꢂ2 are the angles between the radius vector R
and the x- and z-axes, respectively, and ꢅparlꢃꢅz ꢀ ꢅxꢄ and
ꢅperpꢃꢅy ꢀ ꢅxꢄ are the parallel and perpendicular anisotropy
for the C O bond, respectively.
current density for the ring, being 1.0 for substituted ben-
zenes.
υrc D fc ꢆꢃ3 cos2 ꢂ ꢀ 1ꢄ/R3
ꢃ6ꢄ
The effect of the excess ꢀ-electron density at a given
carbon atom on the proton chemical shifts of the neighboring
protons is given by Eqn (5), where q˛ and qˇ are the excess
ꢀ-electron density at the ˛- and ˇ-carbon atoms, respectively.
The ꢀ-electron densities are calculated using Huckel theory
parameterized to reproduce the values obtained from ab
initio calculations.8
The above contributions are added to Eqn (1) to give the
calculated shift of Eqn (7).
υtotal D υcharge C υsteric C υanis C υel C υꢀ C υrc
ꢃ7ꢄ
APPLICATION TO AMINES
In the CHARGE routine, single bonds have been considered
to be isotropic and the same assumption will be made
here for the C–N and N–H bonds. The other terms in the
CHARGE formulation (electric field, steric effects, ꢀ-effects
and ring currents) all need to be considered. The electric
field shift is given directly from the calculated partial atomic
charges and these can be checked by comparison of the
observed and calculated dipole moments. For methylamine,
υꢀ D 10.0 q˛ C 2.0 qˇ
ꢃ5ꢄ
For aromatic molecules, the effect of the aromatic ring
current has to be included, and this is given by the equiva-
lent dipole approximation (Eqn 6). In Eqn 6, R is the distance
of the proton from the benzene ring center, ꢂ the angle of
the R vector with the ring symmetry axis, ꢆ the equiv-
alent dipole of the aromatic ring and fc the ꢀ-electron
Copyright  2007 John Wiley & Sons, Ltd.
Magn. Reson. Chem. 2007; 45: 749–757
DOI: 10.1002/mrc

752
E. A. Basso, G. F. Gauze and R. J. Abraham
dimethylamine, and trimethylamine, the observed versus
calculated dipole moments are 1.33, 1.01, and 0.63 D15
versus 1.32, 1.10, and 0.86 D, respectively. The agreement
is such that the electric field term may be used directly. The
nitrogen steric effect is given by Eqn 2, where the value of
the steric coefficient aS needs to be found. The aromatic ring
current in substituted benzenes has been found in previous
investigations to be the same as in benzene, and this will
be assumed here for the amino benzenes. The effect of the
ꢀ-electron density on the ring protons is given by Eqn 5,
but the effect of the ꢀ-density on the nitrogen atom on the
chemical shifts of the attached proton will require a different
coefficient. There is also the influence of the nitrogen lone
pair on the shielding of neighboring protons. This was not
included explicitly for the analogous case of the oxygen atom
in alcohols and ethers6, being covered by the steric effect of
the oxygen atom plus the ˛-, ˇ- and ꢁ-effects of the oxygen
atom for the near protons. The same procedure will be used
for the nitrogen atom of the amines considered.
Computational methods
The geometries of all compounds used for parameteriza-
tion (Fig. 1) were minimized using ab initio calculations
with the Gaussian 03W program.21 For all compounds the
potential energy surfaces (PES) were constructed at the
HF/3-21G level in order to determine the preferred amino
group orientation. The stable conformers were then mini-
mized at the B3LYP/6–311CCG(d,p) level, and the GIAO
calculations were performed using the recommended22
B3LYP/6–31G(d,p) level. The chemical shifts were refer-
enced to methane (minimized and calculated in the same
manner) and converted to TMS using the experimental value
of 0.23 ppm for methane.23
RESULTS AND DISCUSSION
Conformational analysis
The hydrocarbon fragments in the molecules studied were
chosen as rigid structures; therefore only the rotational
isomerism about the C–N bond needs to be determined.
In order to determine the most stable rotamers for the
amino group, we performed PES calculations by varying
Thus we need to evaluate the steric coefficient aS (Eqn 2),
ꢀ-electron coefficient (Eqn 5) and the ˛-, ˇ- and ꢁ-effects of
the nitrogen atom on the chemical shifts of the near protons.
These parameters were obtained from an iterative calculation
on the observed shifts.
°
the C2–C1–N–H(C) dihedral angle with increments of 10
to all compounds where applicable. Although the nitrogen
lone pair is not defined in either the ab initio or the
CHARGE calculations, it is more convenient to describe
the rotational isomerism about the H–C–N–R2 (R D H, Me)
bond in terms of the H–C–N-lone-pair dihedral angle (ꢂ).
For compounds 1, 1M, and 2, two populated rotamers were
found, and the PES for each compound is shown in Fig. 2.
Note that the energy is given relative to the more stable
conformer in each case. In trans-4-t-butylcyclohexylamine
(1) the gauche conformer (ꢂ D 600) is calculated to be
slightly less stable (0.3 kcal molꢀ1) than the trans (ꢂ D 1800)
conformer. The statistical weight of two for the gauche
conformer results in almost equal populations of the two
forms (anti : gauche, 45 : 55). In the dimethyl derivative (1M),
the symmetric conformer is much more stable than the
gauche by 1.7 kcal molꢀ1. In cis 4-t-butylcyclohexylamine (2),
EXPERIMENTAL
Experimental details
All the amino (R NH2) compounds and solvents used were
obtained commercially, as well were the dimethylamino
compounds 8M, 9M and 10M (Fig. 1). The remaining
dimethylamino derivatives were synthesized according to
literature procedures, as follows.
The mixture of cis and trans 4-dimethylamino tert-
butylcyclohexane isomers (1M) and (2M) and also N-
methyl-4-phenyl piperidine (3M) were prepared through
the method of Leuckart reductive alkylation,16–18 using
formaldehyde in the presence of formic acid. The 2-endo-
dimethylaminonorbornane (4M) was obtained from selective
reduction19 of norcamphor, with sodium cyanohydridob-
orate (NaBH3CN). A variation of this procedure20 using
just formaldehyde and the reductive agent (NaBH3CN) was
used to obtain the dimethylamino compounds (5M), (7M)
and (11M), as the Leuckart reaction did not work for these
compounds.
°
the symmetric form (ꢂ D 180 ) is more stable than the gauche
conformer, with one hydrogen atom pointing into the ring
(E D 1.0 kcal molꢀ1), and is again the more populated form
(77%). In the dimethyl derivative (2M), the methyl groups
are too bulky to point into the ring and only the symmetric
form is present. The conformations of piperidine derivatives
analogous to (3) and (3M) have been determined previously.
In piperidine and N-methyl piperidine, the N–H equatorial
conformer is favored over the axial NH conformer by 0.4 and
3.0 kcal molꢀ1, respectively.24 The remaining compounds
were found to have one major conformer. In both endo
aminonorbornane (4) and the dimethyl derivative (4M), the
stable conformer has the lone pair anti to the CH proton but in
the corresponding exo compounds (5) and (5M) the lone pair
is anti to the C2 –C3 bond. Both 2-adamantanamine (7) and its
dimethyl derivative (7M) exist in a symmetric conformation
Spectroscopy experiments
1H and 13C NMR spectra were obtained on a Bruker
Avance spectrometer operating at 400.13 MHz for proton
and 100.63 MHz for carbon. Compounds 1M, 2M, 3M, 4M,
1
and 10 gave complex, overlapping H spectra at 400 MHz,
and the 1H spectra were obtained on a Bruker Avance
spectrometer operating at 700.13 MHz. COSY, HMQC, and
HMBC experiments were also performed as needed. The
spectra were recorded in 20 mg cmꢀ3 solutions in CDCl3,
with a probe temperature of ca 300 K and TMS as reference.
To avoid the presence of residual HCl, all solutions were
filtrated through columns of basic alumina before the spectra
acquisition.
°
with the lone pair anti to the CH bond (ꢂ D 180 ).
The ab initio calculations iterate to a planar nitrogen atom
for aniline (8), N,N-dimethylaniline (8M) and ortho-toluidine
(9), but in (9M) the nitrogen atom is pyramidal with one
Copyright  2007 John Wiley & Sons, Ltd.
Magn. Reson. Chem. 2007; 45: 749–757
DOI: 10.1002/mrc

Prediction of 1H chemical shifts in amines: Part 26
753
proton is given by Eqn (5), in which the coefficient for the
attached proton was determined as 10 ppm/ꢀ electron.23
In some current investigations on phenols,25 it was found
that the coefficient for the effect of ꢀ-electron on the
oxygen atom on the OH proton was much larger (ca
40 ppm), and it is of interest to determine the analogous
value for nitrogen. This can be obtained by comparing the
NH proton shifts of some para-substituted anilines with
the calculated ꢀ-densities at the nitrogen atom. The para
substituent is sufficiently distant from the nitrogen atom so
that the only factor affecting the NH shift is the ꢀ-electron
density. This data is shown in Table 1 for a variety of para-
substituted anilines. There is an excellent linear correlation
between the shifts and the ꢀ-densities to give a coefficient
of 42.8 ppm/ꢀ electron with an rms error of 0.142. The
observed shifts and shifts calculated by CHARGE with
the above coefficient inserted are shown in Table 1. The
agreement clearly demonstrates the ꢀ-dependence of the
shifts, but it is important to note that the coefficient depends
on the ꢀ-calculation. The modified Huckel program detailed
previously6 is used here. This coefficient was included in
the calculations of the NH proton chemical shifts of the
aromatic amines in Table 3. It has no effect on the NH
shifts of the aliphatic amines detailed here, as none of them
are conjugated. The remaining parameters necessary for
the calculation of the proton chemical shifts were obtained
from an iterative calculation using the nonlinear least mean-
square program CHAP8.26 The coefficients A and B of the
ꢁ-effects (H–C–C–N and H–C–N–C) were determined
Figure 2. Potential energy surface of compounds 1, 1M, and 2
about the C2–C1–N–H(C) dihedral angle. 101 ð 76 mm
(600 ð 600 DPI).
°
methyl group at ca 90 to the phenyl ring. Interestingly the
ab initio calculations give a nonplanar nitrogen atom for
both the naphthalene derivatives (10) and (10M) and the
tetrahydronapthalenes (11) and (11M).
Where the contribution of each rotamer was significant
as in compounds 1, 1M, 2, and 3, the chemical shifts
were calculated for each conformer and the final chemical
shifts obtained from the weighted average. This procedure
is necessary because for some protons the chemical shifts
change by up to 0.5 ppm in the different conformers.
with the nitrogen steric coefficient (aS, Eqn 2), which equals
6
˚
66.6 ppm/A .
Proton chemical shift calculations
CHARGE model versus GIAO calculations
Effect of the ꢀ-density on the nitrogen atom on NH
chemical shifts
The experimental and calculated proton chemical shifts of the
aliphatic amines are presented in Table 2. The experimental
chemical shifts range from ca 0.60 to 3.30υ and were predicted
with the CHARGE model with an rms error of 0.108 ppm.
The majority of the shifts show deviations of less than
0.1 ppm, showing excellent agreement between observed
and calculated chemical shifts.
The effect of the excess ꢀ-electron density at an aromatic
carbon atom on the proton chemical shifts of the attached
Table 1. NH2 1H chemical shifts versus nitrogen ꢀ-electron
excess for para-substituted anilines
One of largest deviations of the calculated shifts is
associated with the gamma substituent effect (N–C–C–H).
This has large dihedral angle dependence. The norbornane
derivatives and the 4-t-butylcyclohexane derivatives have
very different dihedral angles (N–C–C–H) see Fig. 3 and
this may be the reason for such deviations. The ꢁ-effect has
been parameterized using a simple (cos ꢇ) term and this
may not be adequate to fully account for this dihedral angle
dependence.
The observed versus calculated proton chemical shifts for
the aromatic amines are given in Table 3. The majority of
the shifts were predicted to within 0.05 ppm, showing even
better agreement than the aliphatic amines.
It is of some interest to compare the CHARGE model with
the GIAO method. The data are also included in Tables 2 and
3. The GIAO calculations are, in general, less accurate than
CHARGE calculations. In particular they gave all the amino
proton chemical shifts significantly more shielded than the
experimental value. The discrepancies between theory and
experiment for these protons were so large that they were
NH2 Chemical Shift (ppm)
ꢀ-Electron
excess
(me)c
Substituent
Obs.a
Cal.b
NMe2
OEt
F
Me
Cl
Br
I
H
135
140
141
145
149
148
149
149
152
153
158
3.28
3.40
3.51
3.48
3.63
3.64
3.57
3.56
3.90
4.19
4.30
3.146
3.371
3.481
3.585
3.739
3.732
3.720
3.775
3.920
3.986
4.202
CF3
CO2Me
CN
a Data from Ref. 5.
b CHARGE program.
c Millielectrons.
Copyright  2007 John Wiley & Sons, Ltd.
Magn. Reson. Chem. 2007; 45: 749–757
DOI: 10.1002/mrc

754
E. A. Basso, G. F. Gauze and R. J. Abraham
Copyright  2007 John Wiley & Sons, Ltd.
Magn. Reson. Chem. 2007; 45: 749–757
DOI: 10.1002/mrc

Prediction of 1H chemical shifts in amines: Part 26
755
Copyright  2007 John Wiley & Sons, Ltd.
Magn. Reson. Chem. 2007; 45: 749–757
DOI: 10.1002/mrc

756
E. A. Basso, G. F. Gauze and R. J. Abraham
Table 3. Experimental and calculated 1H chemical shifts of aromatic amines
Compound
Method
H2/6 H3/5
H4
NR2
r.m.s
Aniline (8)
CDCl3
CHARGE
GIAO
CDCl3
CHARGE
GIAO
6.635
6.661
6.159
6.730
6.651
6.402
H3
7.127
7.134
7.028
7.231
7.124
7.172
H4
6.733 3.500
6.678 3.538
6.474 2.837
6.713 2.926
6.669 2.726
6.662 2.923
–
0.035
0.318
–
0.122
0.169
N,N-Dimethylaniline (8M)
H5
H6
CH3
NR2
o-Toluidine (9)
CDCl3
CHARGE
GIAO
CDCl3
CHARGE
GIAO
7.021
6.929
6.918
7.147
7.090
7.189
H2
6.683
6.612
6.471
6.934
6.990
6.959
H3
7.010 6.625 2.128 3.488
6.946 6.585 2.376 3.933
6.960 6.168 2.924 1.860
7.142 7.023 2.322 2.689
7.076 6.980 2.325 2.715
7.121 6.927 2.192 2.507
–
0.060
0.260
0.047
0.102
o-N,N-Dimethyltoluidine (9M)
H4
H5
H6
H7
H8
NR2
1-Aminonaphthalene (10)
CDCl3
CHARGE
GIAO
CDCl3
CHARGE
GIAO
CDCl3
CHARGE
GIAO
CDCl3
CHARGE
GIAO
6.676
6.530
6.503
7.023
7.047
7.024
6.449
6.463
6.217
6.889
6.898
6.836
7.272
7.212
7.257
7.344
7.384
7.365
6.901
6.931
6.913
7.074
7.080
7.083
7.308 7.800 7.441 7.433 7.788 3.966
6.937 7.718 7.465 7.333 8.107 3.973 0.201
7.125 7.647 7.425 7.400 7.732 3.236 0.114
7.474 7.775 7.413 7.443 8.223 2.853
7.433 7.807 7.493 7.431 8.422 2.848 0.080
7.398 7.702 7.466 7.509 8.369 2.729 0.083
6.512 2.704 1.723 1.826 2.398 3.475
6.467 2.679 1.816 1.808 2.550 3.995 0.072
6.451 2.626 1.618 1.713 2.229 2.540 0.129
6.798 2.794 1.760 1.780 2.735 2.669
6.89 2.694 1.808 1.800 2.828 2.679 0.061
6.801 2.737 1.661 1.646 2.706 2.464 0.098
–
1-N,N-Dimethylaminonaphthalene (10M)
5-amino-1,2,3,4-Tetrahydronaphthalene (11)
–
–
5-N,N-Dimethylamino-
1,2,3,4-tetrahydronaphthalene (11M)
–
disregarded in the rms error calculations. The theoretical
predictions for protons attached to the carbon chains do
not show a regular behavior. Some of the predicted shifts
were shielded and others were deshielded. Nevertheless the
rms error values are only 0.096 for CHARGE and 0.187 for
GIAO.
In conclusion the GIAO calculations yield reasonable
results and could be improved by using alternative basis sets
or theories, but this is beyond the scope of this study. Overall,
the semiempirical calculations produce more reliable results
and provide a rapid and useful tool for routine use in
chemical shifts prediction.
We also analyzed the model that produces the greatest
number of hits closest to the observed chemical shifts.
CHARGE produces the largest number of best hits (64%),
while GIAO is better only in 32% of the hits.
Acknowledgements
We thank Dr Mehdi Mobli and GlaxoWellcome plc. for the 700 MHz
spectra. The authorsacknowledge Coordenac¸a˜o de Aperfeic¸oamento
´
de Nıvel Superior (CAPES) and Conselho Nacional de Desenvolvi-
´
mento Cientıfico e Tecnolo´gico (CNPq) for scholarships to G.F.G
and E.A.B and a Leverhulme Emeritus Fellowship to R.J.A.
REFERENCES
1. Zurcher RF. Prog. Nucl. Magn. Reson. Spectrosc. 1967; 2: 205.
2. Pretsch E, Simon W, Seibl J, Clerc T. In Spectral Data for Structure
Determination of Organic Compounds, 2nd edn., Springer-Verlag:
London, 1989.
3. Alkorta I, Elguero J. Magn. Reson. Chem. 2004; 42: 955.
4. Abraham RJ, Byrne JJ, Griffiths L, Perez M. Magn. Reson. Chem.
2006; 44: 491.
5. Pouchert CJ, Behnke J. Aldrich Library of 13C and 1H FT NMR
Spectra. Aldrich Chemical Company: Milwaukee, 1993.
6. Abraham RJ. Prog. Nucl. Magn. Reson. Spectrosc. 1999; 35: 85.
7. Abraham RJ, Griffiths L, Warne MA. J. Chem. Soc., Perkin Trans.
2 1998; 1751.
8. Abraham RJ, Bardsley B, Mobli M, Smith RJ. 2005; 43: 3.
9. NMRPredict. http://www.modgrah.co.uk [accessed 2006],
Modgraph Consultants Ltd. 1, Oakland View, Oaklands
Welwyn, Herts, AL6, ORJ, UK.
Figure 3. Structure and dihedral angle of amines derivatives
(R D H, CH3) of norbornane and 4-t-butylcyclohexane.
127 ð 76 mm (600 ð 600 DPI).
Copyright  2007 John Wiley & Sons, Ltd.
Magn. Reson. Chem. 2007; 45: 749–757
DOI: 10.1002/mrc

Prediction of 1H chemical shifts in amines: Part 26
757
10. Dodds JL, McWeeny R, Sadlej AJ. Mol. Phys. 1980; 41: 1419.
11. Wolinski K, Hilton JF, Pulay P. J. Am. Chem. Soc. 1990; 112: 8251.
12. Pulay P, Hinton JF. Encyclopedia of Nuclear Magnetic Resonance.
Wiley: New York, 1995.
13. Lampert H, Mikenda W, Karpfen A, Ka¨hlig H. J. Phys. Chem. A
1997; 101: 9610.
14. McConnell HM. J. Chem. Phys. 1957; 27: 226.
15. McClellan AL. Table of Experimental Dipole Moments Volume 3.
Rahara Enterprises: El Cerrito, CA, 1989.
16. Pine SH, Sanchez BL. J. Org. Chem. 1971; 36: 829.
17. Meiners AF, Bolze C, Schorer AL, Morris FV. J. Org. Chem. 1958;
23: 1122.
Hratchian HP, Cross JB, Adamo C, Jaramillo J, Gomperts R,
Stratmann RE, Yazyev O, Austin AJ, Cammi R, Pomelli C,
Ochterski JW, Ayala PY, Morokuma K, Voth GA, Salvador P,
Dannenberg JJ, Zakrzewski VG, Dapprich S, Daniels AD,
Strain MC, Farkas O, Malick DK, Rabuck AD, Raghavachari K,
Foresman JB, Ortiz JV, Cui Q, Baboul AG, Clifford S,
Cioslowski J, Stefanov BB, Piskorz IP, Komaromi I, Martin RL,
Fox DJ, Keith T, Al-Laham MA, Peng CY, Nanayakkara A,
Challacombe M, Gill PMW, Johnson B, Chen W, Wong MW,
Gonzalez C, Pople JA. GAUSSIAN 03W (Revision 8.02). Gaussian:
Pittsburgh, 2003.
22. Foresman JB, Frish AE. In Exploring Chemistry with Electronic
Structure Methods, ed. Gaussian Inc: Pitsburg, 2nd edn., 1996.
23. Gunther H. NMR Spectroscopy-An Introduction. John Wiley &
Sons: New York, 1980; 68.
18. Moore ML. Org. React. 1948; 5: 301.
19. Borch RF, Berstein MD, Durst HD. J. Am. Chem. Soc. 1971; 93:
2897.
20. Borch RF, Hassid AI. J. Org. Chem. 1972; 37: 1673.
21. Frisch MJ, Trucks GW, Schlegel HB, Scuseria GE, Robb MA,
Cheeseman JR, Vreven T, Kudin KN, Burant JC, Millam JM,
Iyengar SS, Tomasi J, Barone V, Mennucci B, Cossi M,
Scalmani G, Rega N, Petersson GA, Nakatsuji H, Hada M,
Ehara M, Toyota K, Fukuda R, Hasegawa J, Ishida M,
Nakajima T, Honda Y, Kitao O, Nakai H, Klene M, Li X, Knox JE,
24. Eliel EL, Wilen SH. Stereochemistry of Organic Compounds. Wiley:
New York, 1994; 743.
25. Abraham RJ. Mobli M. unpublished work.
26. Kuo SS. Computer Applications of Numerical Methods. cd. 8.
Addison Wesley: London, 1972; in.
Copyright  2007 John Wiley & Sons, Ltd.
Magn. Reson. Chem. 2007; 45: 749–757
DOI: 10.1002/mrc