Intramolecular hydrogen bonding and molecular structure of 2,5-dihydroxyterephthalaldehyde and 4,6-dihydroxyisophthalaldehyde: A gas-phase electron diffraction and ab initio molecular orbital study

Article abstract of DOI:10.1021/jp9620072

The molecular structure of 2,5-dihydroxyterephthalaldehyde has been determined from a joint electron diffraction/ab initio investigation, and the molecular structure of 4,6-dihydroxyisophthalaldehyde has been obtained from ab initio calculations at the MP2/6-31G* level. There is considerable intramolecular hydrogen bonding in these structures manifested by the O...H and O...O distances as well as by the structural changes in the rest of the molecule. These changes are consistent with the notion of resonance-assisted hydrogen bonding. The hydrogen bonding is somewhat stronger in 4,6-dihydroxyisophthalaldehyde than in 2,5-dihydroxyterephthalaldehyde, and this difference may be linked to the difference in the mutual positioning of the interacting formyl and hydroxy groups in these molecules.

Full text of DOI:10.1021/jp9620072

J. Phys. Chem. 1996, 100, 19303-19309
19303
Intramolecular Hydrogen Bonding and Molecular Structure of
2,5-Dihydroxyterephthalaldehyde and 4,6-Dihydroxyisophthalaldehyde: A Gas-Phase
Electron Diffraction and ab Initio Molecular Orbital Study
Konstantin B. Borisenko,1a Ka´roly Zauer,1b and Istva´n Hargittai*,1a,†
Institutes of General and Analytical Chemistry and Organic Chemistry, Budapest Technical UniVersity and
Structural Chemistry Research Group of the Hungarian Academy of Sciences at Eo¨tVo¨s UniVersity, H-1521
Budapest, Hungary
ReceiVed: July 8, 1996; In Final Form: October 3, 1996X
The molecular structure of 2,5-dihydroxyterephthalaldehyde has been determined from a joint electron
diffraction/ab initio investigation, and the molecular structure of 4,6-dihydroxyisophthalaldehyde has been
obtained from ab initio calculations at the MP2/6-31G* level. There is considerable intramolecular hydrogen
bonding in these structures manifested by the O‚‚‚H and O‚‚‚O distances as well as by the structural changes
in the rest of the molecule. These changes are consistent with the notion of resonance-assisted hydrogen
bonding. The hydrogen bonding is somewhat stronger in 4,6-dihydroxyisophthalaldehyde than in 2,5-
dihydroxyterephthalaldehyde, and this difference may be linked to the difference in the mutual positioning of
the interacting formyl and hydroxy groups in these molecules.
Introduction
SCHEME 1
Our recent investigations of the molecular structures of a
series of ortho-substituted benzene derivatives including 2-ni-
troresorcinol,2,3 2-nitrophenol,4 4,6-dinitroresorcinol,5 and sali-
cylaldehyde6 by electron diffraction and ab initio calculations
indicated considerable intramolecular hydrogen bonding in these
molecules assisted by substantial structural changes in the rest
of the molecule as compared with the respective parent
molecules of phenol3,7 and nitrobenzene3,8 and phenol3,7 and
benzaldehyde.6 The barrier to internal rotation of the nitro group
in the o-nitrophenols is about 10 kJ/mol higher than in
nitrobenzene as a consequence of the restraining effect of the
intramolecular hydrogen bond formation.9 The barrier to
internal rotation in salicylaldehyde was found to be also about
10 kJ/mol higher than in benzaldehyde.6 Our recent ab initio
calculations on 4,6-dinitroresorcinol and 2,5-dinitrohydro-
quinon,10 i.e., the nitro analogs of the present title molecules,
revealed some indications of increased intramolecular interac-
tions in 4,6-dinitroresorcinol as compared with 2,5-dinitrohy-
droquinone. As a continuation of our research into the
geometrical consequences of intramolecular hydrogen bond
formation, we report here the results of electron diffraction and
ab initio study of 2,5-dihydroxyterephthalaldehyde and the
results of ab initio calculations on 4,6-dihydroxyisophthalalde-
hyde.
for 20 h. Then 250 mL of water was added to the still hot
mixture, which was further stirred for 90 min. The mixture
was then diluted with a solution of 12 g of (0.21 mol) potassium
hydroxide in 25 mL of water. The product was extracted first
by mixture of 100 mL of ethyl acetate and 50 mL of ether and
then by 50 mL of pure ether. The combined organic phase was
dried by Na2SO4, the solvent was evaporated, and the rest was
fractionated. The fraction boiling in the 170-215 °C interval
(40 g, 80.5%) was collected and fractionated again. Yield of
the product was 36.8 g (73.3%), bp 212-214 °C, mp 55 °C
(lit.11 bp 216.6 °C, mp 56 °C).
1,4-Dimethoxy-2,5-bis(chloromethyl)benzene. The substance
was prepared in the way described in ref 14.
Experimental Section
2,5-Dimethoxyterephthalaldehyde. A mixture of 24.1 g (0.10
mol) of 1,4-dimethoxy-2,5-bis(chloromethyl)benzene and 30 g
(0.21 mol) of hexamethylenetetramine in 200 mL of chloroform
was boiled for 30 min and then left at room temperature
overnight. The solvent was removed under reduced pressure,
and the rest was boiled with stirring in 320 mL of 50% aqueous
acetic acid for 12 h. Then 25 mL of saturated hydrochloric
acid was added to the mixture, and boiling continued further
for 8 h in an oil bath at 120 °C. The precipitated product was
filtered and washed with water and alcohol. Yield of the product
was 2.1 g (10.5%), mp 201 °C (lit.12 mp 207 °C). From the
filtrate an additional 0.8 g of the product could be obtained; its
mp was 200 °C. The substance was used for the next step
without purification.
Synthesis. The sample of 2,5-dihydroxyterephthalaldehyde
was synthesized according to Scheme 1. Steps I, II, and IV
are different from those reported in refs 11, 12, and 13,
respectively.
1,4-Dimethoxybenzene. A mixture of 40 g (0.36 mol) of 1,4-
dihydroxybenzene, 400 mL of acetone, 80 g (0.57 mol) of
powdered dry potassium carbonate, and 80 mL (106 g, 0.84
mol) of dimethyl sulfate was boiled under reflux with stirring
† For the 1996/97 academic year, Distinguished Visiting Professor,
Department of Chemistry, University of North Carolina at Wilmington,
Wilmington, NC 28403.
X Abstract published in AdVance ACS Abstracts, November 15, 1996.
S0022-3654(96)02007-2 CCC: $12.00 © 1996 American Chemical Society

19304 J. Phys. Chem., Vol. 100, No. 50, 1996
Borisenko et al.
Figure 2. Experimental (E) and theoretical (T) radial distributions and
three times their difference (3∆) for 2,5-dihydroxyterephthalaldehyde.
Figure 1. Experimental (E) and theoretical (T) molecular intensities
and three times their differences (3∆) for 2,5-dihydroxyterephthalal-
dehyde. 50 cm and 19 cm refer to the two camera ranges used in the
experiments (cf. Table 1).
TABLE 1: Experimental Conditions for
2,5-Dihydroxyterephthalaldehyde
no. of camera
plates dist., mm temp, °C
nozzle wavelength, data intervals, data steps,
Å
Å-1
Å-1
7
5
501.4
192.4
152
152
0.049 53 2.000-13.875
0.049 53 8.50-35.25
0.125
0.25
2,5-Dihydroxyterephthalaldehyde. To 2.1 g (10.8 mmol) of
2,5-dimethoxyterephthalaldehyde was added 30 mL of 48%
aqueous hydrobromic acid, and the mixture was boiled under
reflux with stirring in an oil bath at 130 °C for 35 h. The
mixture was then evaporated at reduced pressure to ap-
proximately one-fourth of the initial volume, and 50 mL of 48%
aqueous hydrobromic acid was added. Then the mixture was
diluted by adding 50 mL of acetic anhydride in portions with
shaking. It was further boiled for another 25 h. The solvent
was removed under reduced pressure, and the product was
extracted with benzene. After the organic phase was evaporated
and benzene was removed the rest (2,5-dihydroxyterephthal-
aldehyde, 0.8 g, 44.7%) was purified by sublimation in vacuum
(0.5 Torr) in an oil bath at 160-180 °C, yielding 0.5 g (25.7%),
mp 251-252 °C (lit.13 mp 258-260 °C). For the purpose of
the electron diffraction experiment the substance was sublimed
twice, and its purity was checked by thin layer chromatography.
Electron Diffraction Experiment. The electron diffraction
photographs of 2,5-dihydroxyterephthalaldehyde were recorded
in our modified EG-100A apparatus15 with a membrane nozzle
system16 from two nozzle-to-plate distances. Some of the
experimental conditions are summarized in Table 1. The
experimental total intensities are given as Supporting Informa-
tion and may also be obtained from the authors. The atomic
electron scattering factors were taken from available compila-
tions.17 The experimental and theoretical molecular intensities
and radial distributions are shown in Figures 1 and 2. The
numbering of atoms is presented in Figure 3.
Figure 3. The numbering of atoms for 2,5-dihydroxyterephthalalde-
hyde (I) and 4,6-dihydroxyisophthalaldehyde (II).
TABLE 2: 2,5-Dihydroxyterephthalaldehyde: Computed
Geometriesa
parameter
HF/6-31G*
MP2(FC)/6-31G*
C1-C2
1.4009
1.4191
1.3931
1.4020
1.0885
120.66
118.47
120.87
120.50
1.4645
1.2388
1.1057
124.19
116.19
120.67
1.3543
0.9872
107.11
123.12
1.8151
2.6866
145.37
99.54
C2-C3
1.3823
1.3897
1.0751
120.57
118.03
121.40
120.23
1.4775
1.1969
1.0908
124.34
115.55
121.18
1.3380
0.9538
110.29
123.80
1.9130
2.7146
140.10
100.30
-605.882 410 1
C3-C4
C3-H
C6-C1-C2
C1-C2-C3
C2-C3-C4
C1-C6-H
C1-C7
CdO
C7-H
C1-C7dO
C1-C7-H
C2-C1-C7
C-O
O-H
C-O-H
C1-C2-O
O8‚‚‚H11
O8‚‚‚O10
O-H‚‚‚O8
CdO‚‚‚H11
energy (au)
-607.599 388 2
Ab Initio Calculations. The ab initio molecular orbital
calculations were carried out using the GAUSSIAN 94 series
of programs.18 The polarized standard 6-31G* basis set19 was
used throughout. Initially, RHF/6-31G* optimizations, as well
as frequency analyses, were performed for both 2,5-dihydroxy-
terephthalaldehyde (I) and 4,6-dihydroxyisophthalaldehyde (II)
to confirm that the computed planar structures are indeed stable
states. The calculations for these molecules were then followed
by second-order Møller-Plesset20 optimizations with only
valence orbitals active (MP2(FC)/6-31G*) to include the effect
a re equilibrium bond lengths (Å) and angles (deg).
of electron correlation into the calculations. The results of the
computations are presented in Tables 2 and 3.
Potential energy curves corresponding to the rotation of the
formyl groups were calculated for both molecules at the HF/
6-31G* level. In these single-point total energy calculations
the molecular geometries were assumed, while one of the formyl
groups was rotated stepwise in the interval 0-180° with 10°

Molecular Structure of 2,5-Dihydroxyterephthalaldehyde
J. Phys. Chem., Vol. 100, No. 50, 1996 19305
group in 4,6-dihydroxyisophthalaldehyde is calculated to be 85.0
kJ/mol at 95°, and that in 2,5-dihydroxyterephthalaldehyde is
63.7 kJ/mol at 100°.
Structure Analysis
The electron diffraction structure analysis of 2,5-dihydroxy-
terephthalaldehyde was carried out by the least-squares fitting
of calculated molecular scattering intensities to the experimental
data, applying conventional single-start21 and multistart Monte
Carlo global optimization techniques.22
Initially the molecule was assumed to have C2h symmetry,
and its geometry was described by the following independent
parameters.
The benzene ring geometry was described by the mean value
of the ring C-C bonds, two ∆CC bond length differences
(between C1-C2 and C3-C4, ∆CC1, and between C1-C2 and
C2-C3, ∆CC2), and angles C1-C2-C3 and C6-C1-C2. All
C-H bonds of the benzene ring were assumed to have equal
length and directed along the bisectors of the respective C-C-C
angles. Following some initial attempts to refine the two ∆CC
differences independently, it was decided to assume these
differences at the computed values. The choice of these
assumed differences had no appreciable influence on the other
parameters.
Figure 4. Potential energy curve of internal rotation of the formyl
groups in 2,5-dihydroxyterephthalaldehyde and 4,6-dihydroxy isoph-
thalaldehyde calculated at the HF/6-31G* level.
TABLE 3: 4,6-Dihydroxyisophthalaldehyde: Computed
Geometriesa
parameter
HF/6-31G*
MP2(FC)/6-31G*
C6-C1
1.4197
1.4236
1.3940
1.3934
1.0854
1.0916
120.29
118.84
121.59
120.16
119.21
119.92
1.4568
1.2400
1.1062
124.36
116.11
121.20
1.3456
0.9908
107.22
121.79
1.7908
2.6745
146.66
98.77
C1-C2
1.3840
1.3841
1.0722
1.0777
120.48
118.15
122.56
120.18
118.72
119.91
1.4585
1.2017
1.0921
124.79
115.51
121.62
1.3186
0.9577
110.51
121.69
1.8676
2.6893
142.21
99.19
C5-C6
C5-H
C2-H
The geometry of the formyl group was described by the bond
length differences between C1-C7 and (C-C)mean, ∆CC3;
between CdO and (C-C)mean, ∆CO; and between C7-H and
(C-H)mean of the benzene ring, ∆CH; by the angle C1-C7dO;
and by the deviation of the direction of the C7-H bond from
the bisector of the C1-C7dO angle, ∆CCH. Varying the
differences involving hydrogen atoms, i.e., ∆CH and ∆CCH,
has led to suspiciously large values. Thus these parameters were
also assumed at the calculated ab initio values in the refinements
throughout.
C5-C6-C1
C6-C1-C2
C1-C2-C3
C4-C5-C6
C3-C2-H
C6-C5-H
C3-C7
C7dO
C7-H
C3-C7dO
C3-C7-H
C4-C3-C7
C-O
The hydroxy group was described with a ∆CO bond length
difference between C1-C2 and C-O and with a ∆OH difference
between (C-H)mean and O-H. In addition, the C-O-H angle
was refined as an independent parameter. The angle of torsion
around the C-O bond could not be determined in the least-
squares refinements, and it was assumed to be zero (allowing
for the closest contact between the hydroxy group hydrogen
and the formyl group oxygen), as suggested by the ab initio
calculations.
O-H
C-O-H
C3-C4-O
O8‚‚‚H11
O8‚‚‚O10
O-H‚‚‚O8
CdO‚‚‚H11
energy (au)
-605.898 234 2
-607.606 350 1
a re equilibrium bond lengths (Å) and angles (deg).
Preliminary calculations, in which only one conformer of 2,5-
dihydroxyterephthalaldehyde was considered, lead to a model
with the C1-C7 bond, ra, 1.478(3), being the same as in
benzaldehyde, ra, 1.477(2).6 Previous experience with similar
intramolecular interactions in o-nitrophenols and the ab initio
calculations both suggested this bond to be considerably shorter
in 2,5-dihydroxyterephthalaldehyde than in benzaldehyde. Thus
we performed Monte Carlo optimizations searching for a
minimum in which certain conditions concerning the structural
variations in 2,5-dihydroxyterephthalaldehyde as compared to
benzaldehyde and phenol would be satisfied. The intramolecular
motion of the formyl groups was described in terms of their
rotational potential function in these calculations.
steps. The other formyl group was kept in the plane of the
benzene ring. This approach permitted estimation of the barrier
heights and some of the other parameters of the rotational
potential energy functions. The interaction terms are canceled
in this case. The results of these calculations along with the
fitted rotational potential energy function (in the form of Fourier
series),
V(æ1,æ2) ) 1/2V1(2 - cos æ1 - cos æ2) +
1/2V2(2 - cos 2æ1 - cos 2æ2) + 1/2V3(2 - cos 3æ1 -
cos 3æ2) + 1/2V4(2 - cos 4æ1 - cos 4æ2)
According to the estimations of parameters of the potential
function describing the internal rotation of the formyl groups
in 2,5-dihydroxyterephthalaldehyde and 4,6-dihydroxyisoph-
thalaldehyde from ab initio calculations, the contribution of the
first two terms in the Fourier series representation is essential.
The rotational potential energy function for 2,5-dihydroxy-
are shown in Figure 4. We obtained the following Fourier
coefficients for 2,5-dihydroxyterephthalaldehyde, V1 ) 35.5, V2
) 42.9, V3 ) 5.4, V4 ) -0.5, and for 4,6-dihydroxyisophthal-
aldehyde, V1 ) 36.8, V2 ) 62.9, V3 ) 7.0, V4 ) -1.0 (all values
in kJ/mol). The barrier height to internal rotation of the formyl

19306 J. Phys. Chem., Vol. 100, No. 50, 1996
Borisenko et al.
Figure 5. Results of Monte Carlo optimization with different assumed values of the parameters of the potential energy function of internal rotation
of the formyl groups in 2,5-dihydroxyterephthalaldehyde (V1 and V2) and initial geometries. The results are represented in the form of an R-factor
map. The computed values (HF/6-31G*, see text) are indicated by the arrow.
TABLE 4: Results of Electron Diffraction Least-Squares
Refinementsa of 2,5-Dihydroxyterephthalaldehyde
terephthalaldehyde was thus taken as
V(φ1,φ2) ) 1/2V1(2 - cos φ1 - cos φ2) +
parameter
model Ab
parameter set Bc
1/2V2(2 - cos 2φ1 - cos 2φ2)
(C-C)mean
C1-C2
1.405(2)
1.420(4)
1.394(6)
1.403(3)
1.111(14)
121.2(5)
119.3(2)
119.5(7)
1.466(3)
1.229(1)
1.128(16)
124.6(4)
116.0(10)
120.5(2)
1.369(4)
0.976(6)
121.8(3)
102.2(10)
1.751(9)
2.662(4)
2.63
1.404(3)
1.418(3)
1.392(3)
1.401(3)
1.111(16)
121.5(5)
118.7(3)
119.8(3)
1.469(8)
1.229(2)
1.128(16)
124.8(8)
115.9(4)
120.2(5)
1.370(4)
0.974(5)
122.8(5)
101.0(12)
1.765(8)
2.677(8)
C2-C3
in the electron diffraction structure analysis, ignoring higher
order and interaction terms. Such an assumption may be
reasonable because the higher order terms appear considerably
smaller than these 1-fold and 2-fold terms in the results of the
ab initio calculations. Interaction between well-separated formyl
groups is supposed to be also small. The molecular model
describing internal rotation of the formyl groups in 2,5-
dihydroxyterephthalaldehyde included 91 rigid pseudoconform-
ers with differen φ1 and φ2 angles of formyl group torsion in
the 0-180° interval with 15° steps. The other geometrical
parameters were assumed to be the same in all conformers. The
contribution of each pseudoconformer was calculated from a
Boltzmann distribution with double statistical weights for
conformers in which φ1 * φ2. The calculations were performed
assuming different values of the parameters V1 and V2 in the
0-100 kJ/mol intervals using 200 different initial parameter
sets in our Monte Carlo optimization. All other structure
parameters and vibrational amplitudes were varied in these
refinements. Results of the calculations in the form of the
R-factor surface map as a function of V1 and V2 values are
presented in Figure 5. The results indicate that there can be
several models with different V1 and V2 values equally well
approximating the experimental data, with one of these models
corresponding to the computed ab initio parameters of the
rotational potential function. Thus we assumed the V1 and V2
parameters at the calculated ab initio values in subsequent
calculations.
C3-C4
(C-H)mean
C6-C1-C2
C1-C2-C3
C2-C3-C4
C1-C7
CdO
C7-H9
C1-C7dO
C1-C7-H
C2-C1-C7
C-O
O-H
C1-C2-O
C-O-H
d
O8‚‚‚H11
d
O8‚‚‚O10
R-factor, %
a ra distances (Å) and angles (deg) with least-squares standard
deviations parenthesized in units of the last digit. b The least minimum
from 200 refinements with different initial geometries. c Average
structure using 200 parameter sets from the results of Monte Carlo
optimization. The parenthesized standard deviations are calculated from
these results and indicate the variations of the results as a consequence
of choosing different initial parameters (bond lengths, bond angles, and
amplitudes of vibrations), in reasonable intervals, in the refinements.
d All-planar hydrogen-bonded conformer.
except the parameters of the potential energy function. The
parameter set B in the second column of Table 4 represents an
average of the results from the Monte Carlo optimization. The
parenthesized standard deviations are calculated from these
results and indicate the variations of the results as a consequence
of choosing different initial parameters (bond lengths, bond
angles, and amplitudes of vibrations), in reasonable intervals,
in the refinements. The larger standard deviations of the
parameter set B, in some cases, as compared with those of model
A, indicate the sensitivity of some of the bond lengths and bond
angles to the choice of the initial values of the other parameters.
Further Monte Carlo optimization with 200 initial geometries
assuming the computed parameters of the rotational potential
function has yielded a structure of 2,5-dihydroxyterephthalal-
dehyde that was consistent with the expected structural variations
as compared to benzaldehyde and phenol on the basis of the ab
initio calculations. The results are presented in Table 4. The
standard deviations of model A were calculated in a special
refinement in which all parameters were treated as variables,

Molecular Structure of 2,5-Dihydroxyterephthalaldehyde
J. Phys. Chem., Vol. 100, No. 50, 1996 19307
hydrogen bonding24 has been a very useful concept in interpret-
ing the structural variations in these compound series. It should
be noted that the difference between the longest and the shortest
C-C bond in the benzene ring is slightly larger in 4,6-
dihydroxyisophthalaldehyde, 0.031 Å, than in 2,5-dihydroxy-
terephthalaldehyde, 0.026 Å. This may suggest a greater
electron redistribution in the benzene ring of 4,6-dihydroxy-
isophthalaldehyde than in 2,5-dihydroxyterephthalaldehyde.
TABLE 5: Bond Lengths (rg, Å) and Bond Angles (deg) of
2,5-Dihydroxyterephthalaldehydea with Estimated Total
Errors from the Electron Diffraction Analysis Incorporating
Constraints from ab Initio MO Calculations
(C-C)mean
C1-C2
C2-C3
C3-C4
(C-H)mean
C1-C7
CdO
1.407 ( 0.004
1.422 ( 0.006
1.396 ( 0.009
1.405 ( 0.005
1.12 ( 0.02
1.469 ( 0.005
1.231 ( 0.003
1.13 ( 0.02
1.371 ( 0.006
0.983 ( 0.009
C6-C1-C2
C1-C2-C3
C2-C3-C4
C2-C1-C7
C1-C7)O
C1-C7-H
C1-C2-O
C-O-H
121.2 ( 0.7
119.3 ( 0.3
119.5 ( 1.0
120.5 ( 0.3
124.6 ( 0.6
116.0 ( 1.0
121.8 ( 0.4
102.2 ( 1.0
Formyl Group Torsion. The ab initio calculations indicated
the existence of another stable conformer of 2,5-dihydroxy-
terephthalaldehyde and 4,6-dihydroxyisophthalaldehyde in which
one of the formyl groups is being rotated to about 180° (Figure
4). However, their contribution at the experimental temperature
should be small, as they have about 40 kJ/mol higher energy
than the hydrogen-bonded conformers. Usually single-point ab
initio calculations of the potential energy overestimate the
absolute energy, while the difference between two barrier
heights, calculated under the same conditions, may be free of
such an error. The barrier height to internal rotation of the
formyl groups in 2,5-dihydroxyterephthalaldehyde is 21.3 kJ/
mol lower than that in 4,6-dihydroxyisophthalaldehyde. The
present electron diffraction results concerning the internal
rotation of the formyl groups in 2,5-dihydroxyterephthalaldehyde
are in agreement with those from the ab initio calculations. The
electron diffraction results are rather approximate in this respect
(Figure 5).
C7-H9
C-O
O-H
a Intramolecular motion of the formyl groups was treated in terms
of their rotational potential function; see text.
Results and Discussion
Bond lengths and bond angles from the joint electron
diffraction/ab initio investigation of the molecular structure of
2,5-dihydroxyterephthalaldehyde with estimated total errors23
are presented in Table 5. The results are compared with those
of phenol and benzaldehyde in Tables 6 and 7, referring to the
calculations and to electron diffraction, respectively. The
parameters of salicylaldehyde are also listed for completeness.
Benzene Ring Geometry. The rg(C-C)mean bond length of
the benzene ring in 2,5-dihydroxyterephthalaldehyde, 1.407 (
0.004 Å, shows a slight increase as compared with salicylal-
dehyde,6 1.404 ( 0.003 Å. This increase is within experimental
error but may be better determined than it appears because both
experiments may have the same systematic error which is the
single largest component of the estimated total experimental
error. The slight increase of the (C-C)mean bond length also
appears in the results of the ab initio MP2(FC)/6-31G*
calculations on 2,5-dihydroxyterephthalaldehyde, 1.405 Å, as
compared with salicylaldehyde,6 1.400 Å. An even larger
difference is observed if compared with rg(C-C)mean of some
monosubstituted benzenes, such as benzaldehyde, 1.397 ( 0.003
Å from electron diffraction and 1.397 Å from ab initio MP2-
(FC)/6-31G* calculations.6 According to the ab initio MP2-
(FC)/6-31G* calculations, the C-C bond between the substit-
uents is considerably longer than the other C-C bonds in the
benzene ring in both 2,5-dihydroxyterephthalaldehyde and 4,6-
dihydroxyisophthalaldehyde (Tables 2 and 3). This is in
agreement with the consequences of the intramolecular hydrogen
bond formation observed for o-nitrophenols and contribution
of o-quinonoid resonance forms of the molecules, presented in
Scheme 2. Generally speaking, the notion of resonance-assisted
Intramolecular Hydrogen Bonding. Some important struc-
ture parameters from the computations are compiled in Table
6. There is relatively strong intramolecular hydrogen bonding
assisted by resonance in both molecules. This is inferred from
the short O‚‚‚H and O‚‚‚O distances and from structural changes
in the rest of the molecule, as compared with phenol and
benzaldehyde. The shortening of the C1-CHO and C-O bonds
and the lengthening of the CdO and O-H bonds as compared
to the corresponding bonds in benzaldehyde and phenol are
suggested by their resonance structures (Scheme 2). Note that
changes in the C1-CHO, C-O, and O-H are more pronounced
in 4,6-dihydroxyisophthalaldehyde than in 2,5-dihydroxytereph-
thalaldehyde.
An increase of the hydrogen bond strength in the series 2,5-
dihydroxyterephthalaldehyde to salicylaldehyde to 4,6-dihy-
droxyisophthalaldehyde can be discerned in the calculated
O‚‚‚H/O‚‚‚O distances: 1.815/2.687 Å, 1.803/2.681 Å, and
1.791/2.675 Å, respectively. The trend can also be seen in the
changes of the other parameters; for example, the C-O, C1-
CHO bonds shorten, the C(C)-C(O)-O, C-C(C)-C angles
TABLE 6: Comparison of Some Important Geometrical Parametersa of Phenol, Benzaldehyde,
2,5-Dihydroxyterephthalaldehyde (I), Salicylaldehyde, and 4,6-Dihydroxyisophthalaldehyde (II) from ab Initio MP2(FC)/6-31G*
Calculations
difference
difference
difference
difference
parameter
phenolb benzaldehydec
I
salicylaldehydec
II
I - phenol I - benzaldehyde II - phenol II - benzaldehyde
(C-C)mean
1.396
1.397
120.3
1.397
1.401
1.405
1.419
118.5
120.7
1.465
1.239
120.7
124.2
1.354
0.987
123.1
107.1
1.815
2.687
1.400
1.414
119.3
119.7
1.460
1.240
120.8
124.5
1.353
0.989
122.7
107.0
1.803
2.681
1.404
1.424
120.3
118.8
1.457
1.240
121.2
124.4
1.346
0.991
121.8
107.2
1.791
2.675
+0.009
+0.022
-1.7
+0.008
+0.018
+0.008
+0.027
0.0
+0.007
+0.023
C1-C2
C-C(O)-C
C-C(C)-C
C1-C(dO)
CdO
120.3
1.480
1.227
119.9
124.2
+0.4
-1.5
-0.015
+0.012
-0.8
-0.023
+0.013
-1.3
C-C(C)-C(dO)
C-CdO
C-O
0.0
+0.2
1.375
0.973
122.8
108.3
-0.021
+0.015
+0.3
-0.029
+0.018
-1.0
O-H
C(C)-C(O)-O
C-O-H
O‚‚‚H
-1.2
-1.1
O‚‚‚O
a re distances (Å) and angles (degrees). b Reference 3. c Reference 6.

19308 J. Phys. Chem., Vol. 100, No. 50, 1996
Borisenko et al.
TABLE 7: Comparison of Some Important Geometrical Parametersa of Phenol, Benzaldehyde,
2,5-Dihydroxyterephthalaldehyde (I), and Salicylaldehyde from Electron Diffraction
difference
I - phenol
difference
I - benzaldehyde
parameter
phenolb
benzaldehydec
I
salicylaldehydec
(C-C)mean
1.399(3)
1.399(3)
121.6(2)
1.397(3)
1.400(7)
1.407(4)
1.422(6)
119.3(3)
121.2(7)
1.469(5)
1.231(3)
120.5(3)
124.6(6)
1.371(6)
0.983(9)
121.8(4)
102.2(10)
1.76(1)
1.404(3)
1.418(14)
120.9(9)
118.2(18)
1.462(11)
1.225(4)
121.4(8)
123.8(12)
1.362(10)
0.985(14)
120.9(11)
104.8(26)
1.74(2)
+0.008
+0.023
-2.3
+0.010
+0.022
C1-C2
C-C(O)-C
C-C(C)-C
C1-C(dO)
CdO
119.9(10)
1.479(4)
1.212(3)
120.9(6)
123.6(4)
+1.3
-0.10
+0.019
-0.4
C-C(C)-C(dO)
C-CdO
C-O
+1.0
1.381(4)
0.958(3)
121.2(12)
106.4(17)
-0.010
+0.025
+0.6
O-H
C(C)-C(O)-O
C-O-H
O‚‚‚H
-4.2
O‚‚‚O
2.667(8)
2.65(1)
a rg distances (Å) and ra angles (degrees) with estimated total errors in the units of the last digit in parenthesis. b Reference 7. c Reference 6.
SCHEME 2
SCHEME 3
mental errors are too large to draw any conclusion from this
structural change.
Conclusions
close, the C-C(C)-C angle opens, and the O-H bond
lengthens in this series (Table 6). Increase of the barrier height
to internal rotation in 4,6-dihydroxyisophthalaldehyde as com-
pared to 2,5-dihydroxyterephthalaldehyde may also be an
indication of stronger intramolecular interactions in 4,6-dihy-
droxyisophthalaldehyde.
1. There is resonance-assisted intramolecular hydrogen
bonding in 2,5-dihydroxyterephthalaldehyde and 4,6-dihydroxy-
isophthalaldehyde. This implies not only short O‚‚‚H and O‚‚‚O
distances but also considerable structural variations in both
molecules comprising a shortening of the C-CHO and C-O
bonds and a lengthening of the CdO, O-H bonds, and the C-C
bond between the substituents in the benzene ring, as compared
with the respective bonds in phenol and benzaldehyde.
2. Somewhat weaker intramolecular interactions can be
discerned in 2,5-dihydroxyterephthalaldehyde than in 4,6-
dihydroxyisophthalaldehyde as compared with parent molecules
from the ab initio calculations. This may be attributed to the
difference in the mutual positioning of the interacting pairs of
the formyl and hydroxy groups.
The strength of the intramolecular hydrogen bonding seems
affected by the mutual positioning of the interacting hydroxy
and formyl groups in 2,5-dihydroxyterephthalaldehyde (I) and
4,6-dihydroxyisophthalaldehyde (II), just like it was recently
observed in the computed structures of 2,5-dinitrohydroquinone
and 4,6-dinitroresorcinol.10 We suggest that the hydroxy
group-formyl group interaction shifts the electron density from
the hydroxy group toward the formyl group, and this shift is in
direct relation with the intramolecular hydrogen bond formation.
The mutual positioning of the two pairs of the interacting groups
may increase this shift of electron density in 4,6-dihydroxy-
isophthalaldehyde because the directions of the shifts caused
by each interacting pair coincide. This leads to greater electron
redistribution and, consequently, stronger hydrogen bonding.
In 2,5-dihydroxyterephthalaldehyde the shifts of the electron
density, caused by each interacting pair of formyl and hydroxy
group, would counteract each other, thus resulting in smaller
overall electron density redistribution and, consequently, weaker
intramolecular hydrogen bonding (Scheme 3).
Acknowledgment. We thank Mrs. Ma´ria Kolonits for
experimental work. Konstantin B. Borisenko thanks the J.
Varga Foundation of the Budapest Technical University for a
Postdoctoral Fellowship. This research has been supported by
the Hungarian National Scientific Research Foundation (OTKA,
No. T 014945).
Supporting Information Available: Two tables showing
total experimental electron diffraction intensities and background
data (4 pages). Ordering information is given on any current
masthead page.
The structural variations in 2,5-dihydroxyterephthalaldehyde
from electron diffraction are consistent with those computed at
the MP2(FC)/6-31G* level (Table 7). There may also be a
slight increase of the O‚‚‚O distance in 2,5-dihydroxytereph-
thalaldehyde as compared with salicylaldehyde, but the experi-
References and Notes
(1) (a) Institute of General and Analytical Chemistry. (b) Institute of
Organic Chemistry.

Molecular Structure of 2,5-Dihydroxyterephthalaldehyde
J. Phys. Chem., Vol. 100, No. 50, 1996 19309
(2) Borisenko, K. B.; Hargittai, I. J. Phys. Chem. 1993, 97, 4080.
(3) Bock, C. W.; Hargittai, I. Struct. Chem. 1994, 5, 307.
(4) Borisenko, K. B.; Bock, C. W.; Hargittai, I. J. Phys. Chem. 1994,
98, 1442.
(5) Borisenko, K. B.; Hargittai, I. J. Phys. Chem. 1995, 99, 13808.
(6) Borisenko, K. B.; Bock, C. W.; Hargittai, I. J. Phys. Chem. 1996,
100, 7426.
(7) Portalone, G.; Schultz, G.; Domenicano, A.; Hargittai, I. Chem.
Phys. Lett. 1992, 197, 482.
(8) Domenicano, A.; Schultz, G.; Hargittai, I.; Colapietro, M.; Porta-
lone, G.; George, P.; Bock, C. W. Struct. Chem. 1990, 1, 107.
(9) Borisenko, K. B.; Hargittai, I. J. Mol. Struct., in press.
(10) Borisenko, K. B.; Bock, C. W.; Hargittai, I. J. Mol. Struct.
(THEOCHEM), in press.
(11) Hlasiwetz, H.; Habermann, J. Justus Liebigs Ann. Chem. 1875, 177,
334.
(12) Angyal, S. J.; Morris, P. J.; Tetaz, J. R.; Wilson, J. G. J. Chem.
Soc. 1950, 2141.
(13) Marx, J. N.; Song, P.-S.; Chui, P. K. J. Heterocycl. Chem. 1975,
417.
(14) Wood, J. H.; Gibson, R. E. J. Am. Chem. Soc. 1949, 71, 393.
(15) Hargittai, I.; Herna´di, J.; Kolonits, M. Prib. Tekh. Eksp. 1972, 239.
Hargittai, I.; Tremmel, J.; Kolonits, M. Hung. Sci. Instrum. 1980, 50, 31.
(16) Hargittai, I.; Herna´di, J.; Kolonits, M.; Schultz, G. ReV. Sci. Instrum.
1971, 42, 546.
(17) (a) Coherent: Bonham, R. A.; Scha¨fer, L. In International Tables
for X-ray Crystallography; Kynoch Press: Birmingham, 1974; Vol IV,
Chapter 2.5. (b) Incoherent: Tavard, C.; Nicolas, D.; Rouault, M. J. Chim.
Phys. 1967, 64, 540.
(18) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Gill, P. M. W.;
Johnson, B. G.; Robb, M. A.; Cheeseman, J. R.; Keith, T.; Petersson, G.
A.; Montgomery, J. A.; Raghavachari, K.; Al-Laham, M. A.; Zakrzewski,
V. G.; Ortiz, J. V.; Foresman, J. B.; Cioslowski, J.; Stefanov, B. B.;
Nanayakkara, A.; Challacombe, M.; Peng, C. Y.; Ayala, P. Y.; Chen, W.;
Wong, M. W.; Andres, J. L.; Replogle, E. S.; Gomperts, R.; Martin, R. L.;
Fox, D. J.; Binkley, J. S.; Defrees, D. J.; Baker, J.; Stewart, J. P.; Head-
Gordon, M.; Gonzalez, C.; Pople, J. A. Gaussian 94, Revision B.2; Gaussian
Inc.: Pittsburgh, PA, 1995.
(19) Hehre, W. J.; Ditchfield, R.; Pople, J. A. J. Chem. Phys. 1972, 56,
2257. Hariharan, P. C.; Pople, J. A. Theor. Chim. Acta 1973, 28, 213.
(20) Møller, C.; Plesset, M. S. Phys. ReV. 1934, 46, 618.
(21) Andersen, B.; Seip, H. M.; Strand, T. G.; Stølevik, R. Acta Chem.
Scand. 1969, 23, 3224.
(22) Borisenko, K. B.; Hargittai, I. J. Mol. Struct. 1996, 376, 195.
(23) Hargittai, M.; Hargittai, I. J. Chem. Phys. 1973, 59, 2513.
(24) See, for example: Gilli, P.; Ferretti, V.; Bertolasi, V.; Gilli, G. In
AdV. in Molecular Structure Research; Hargittai, M., Hargittai, I., Eds.;
1996; Vol. 2, pp 67-102.
JP9620072