New crystalline modification of terephthalic acid

Article abstract of DOI:10.1016/S0022-2860(01)00493-8

The new crystalline modification of terephthalic acid can be obtained either from the commercially available triclinic modification by heating at 250 deg C or directly by a thermal hydrolysis of p-dicyanobenzene. It crystallises in a monoclinic system, space group C2/m with a = 8.940(2), b = 10.442(2), c = 3.790(1) Angstroem, V = 353.7(2) Angstroem³, β = 91.21(3) deg and Z = 2. Both carboxyl groups of terephthalic acid are coplanar with the phenyl ring. The molecules in the crystal are linked through the hydrogen bonds in the carboxyl groups into infinite chains. The orientational disorder is observed in the carboxyl groups. The structure of the monoclinic form is compared with the structure of both the triclinic modification.

Full text of DOI:10.1016/S0022-2860(01)00493-8

Journal of Molecular Structure 595 ꢀ2001) 77±82
www.elsevier.com/locate/molstruc
New crystalline modi®cation of terephthalic acid
Â
M. Sled z , J. Janczak, R. Kubiak
*
W. Trzebiatowski Institute of Low Temperature and Structure Research, Polish Academy of Sciences, Ok o lna 2 str.,
P.O. Box 1410, 50-950 Wrocøaw, Poland
Received 21 December 2000; accepted 15 January 2001
Abstract
The new crystalline modi®cation of terephthalic acid can be obtained either from the commercially available triclinic
modi®cation by heating at 2508C or directly by a thermal hydrolysis of p-dicyanobenzene. It crystallises in a monoclinic
3
ꢀ
ꢀ
system, space group C2=m with a ˆ 8:940ꢀ2†; b ˆ 10:442ꢀ2†; c ˆ 3:790ꢀ1† A; V ˆ 353:7ꢀ2† A ; b ˆ 91:21ꢀ3†8 and Z ˆ 2:
Both carboxyl groups of terephthalic acid are coplanar with the phenyl ring. The molecules in the crystal are linked through the
hydrogen bonds in the carboxyl groups into in®nite chains. The orientational disorder is observed in the carboxyl groups. The
structure of the monoclinic form is compared with the structure of both the triclinic modi®cations. q 2001 Elsevier Science
B.V. All rights reserved.
Keywords: Terephthalic acid; Powder diffraction; Crystal structure; Polymorphism; Phase transition
1
. Introduction
and Brown [3] obtained both forms of TPA by recrys-
tallization of pure acid from a water solution in a glass
tube at about 1508C. They stated that form I is more
stable at room temperature than form II as it has a
higher density. Additionally they suggested that the
form II is a high-temperature modi®cation. The phase
II crystallises with a different habit ꢀparallelepipeds)
in comparison to form I ꢀplates and needles). The
drastic change in the crystal habit of TPA at tempera-
tures above 1508C observed by Gaines and Myerson
[4] was attributed to the phase transition of the high-
temperature phase II. Gerasimov et al., [5] Colombo
et al. [6] and Fjaer and Samuelsen [7] recognise both
the triclinic forms of TPA in their Raman spectro-
scopic studies. However, Saska and Myerson [1,8]
using the powder X-ray diffraction technique based
on a wide range of TPA samples obtained under
different conditions have not obtained any evidence
for the presence of form II. In turn Davey et al. [9]
using the scanning electron microscope ꢀSEM)
Terephthalic acid C H ꢀCOOH) ꢀabbreviated as
6
4
2
TPA) is the main starting material for the synthesis
of polyethylene terephthalate ꢀPET) which is used in
the plastics manufacturing industry [1]. Commercially
it is produced by the catalytic air oxidation of para-
xylene [1,2]. According to Bailey and Brown [3] TPA
exists in two crystalline modi®cations ꢀforms I and II).
Both modi®cations crystallise in the centrosymmetric
space group of the triclinic system, with one molecule
per unit cell. In both forms the molecules are linked
together by hydrogen-bonded carboxyl dimers into
in®nite chains, forming two-dimensional layers. The
main difference between these triclinic forms can be
found in the mutual arrangement of the chains. Bailey
*
Corresponding author. Tel.: 148-71-343-5021;
fax: 148-71-441-029.
E-mail address: sledz@int.pan.wroc.pl ꢀM. Sled z ).
Â
0022-2860/01/$ - see front matter q 2001 Elsevier Science B.V. All rights reserved.
PII: S0022-2860ꢀ01)00493-8

Â
M. Sled z et al. / Journal of Molecular Structure 595 ꢀ2001) 77±82
78
observed both the triclinic forms of TPA. Neverthe-
less, in contrast to the suggestions given by Bailey
and Brown [3], they stated that form II is more
stable at room temperature and normal pressure
than form I. Additionally they suggested that form
I at room temperature is the metastable modi®ca-
tion, which is stabilised by twinning. Domenicano
et al. [10] obtained the triclinic modi®cation of TPA
and they stated that it is ªapparently different from
either of the two forms described by Bailey and
Brownº [3]. However, the powder diffraction pattern
calculated for this modi®cation is quite same as that
calculated for phase I obtained by Bailey and Brown
[3].
Since the available literature data concerning the
different phases and the conditions of the phase
transition in the TPA are inconsistent, we therefore
attempted to determine the conditions in which the
appropriate phases of TPA exists. As a result of our
investigations a new modi®cation of TPA in a crystal-
line form ꢀmonoclinic) has been obtained. In this
contribution we describe the formation conditions
and the crystal structure of the new monoclinic form
of TPA. Additionally, the correlation between the
crystal structures of the monoclinic and triclinic
forms is discussed.
2. Experimental
2.1. Powder diffraction
The commercial TPA ꢀtriclinic form I) was used in
the powder diffraction experiment. The several
samples of TPA were heated at different temperatures
in the range of 100±2508C in sealed glass ampoules
for 5 h and then cooled to the room temperature. Next,
their X-ray powder diffraction diagrams were taken
on a STOE transmission powder diffractometer
equipped with a position sensitive detector ꢀPSD)
and monochromatic CuKa1 radiation ꢀl ˆ
ꢀ
1
both the triclinic forms of TPA were calculated on
:540562 A†: The powder diffraction diagrams for
Fig. 1. X-ray powder diffraction diagrams of the terephthalic acid:
ꢀ
a, b) the calculated diagrams for the triclinic forms I and II, respec-
tively; ꢀc) experimental diagram for the sample heated at 2508; and
d) the calculated diagram for the monoclinic form.
ꢀ

Â
M. Sled z et al. / Journal of Molecular Structure 595 ꢀ2001) 77±82
79
pressed into a pellet. The pellet ꢀ,0.5 g) was inserted
into an evacuated glass ampoule and heated at 2508C
for 5 h and the ampoule was then cooled in the
temperature gradient. In the cooler part of the
ampoule the crystals of monoclinic form of
terephthalic acid as a result of sublimation are grown.
2.3. X-ray data collection
Scheme 1.
A colourless single crystal of C H ꢀCOOH) with
6
4
2
3
approximate dimensions 0:38 £ 0:30 £ 0:20 mm was
used for data collection on a four-circle KUMA KM-4
diffractometer equipped with a graphite monochro-
the basis of the X-ray single-crystal data [3]. The
powder diffraction diagram for the monoclinic form
was calculated on the basis of the present single
crystal structure. All the diagrams are shown in Fig. 1.
ꢀ
mator and MoKa radiation ꢀl ˆ 0:71073 A†: The
lattice parameters were re®ned by least-squares ®t of
2
6 re¯ections measured in the 2q range of 16±268: A
total of 928 re¯ections was measured in the range of
# 2q # 588 using the v ±2q scan technique. Three
2
.2. Preparation
4
standard re¯ections were monitored at every 50 and
they showed no intensity variations. Intensities and
their standard deviations were corrected for Lorenz
and polarisation effects using a Kuma Km-4 software
The crystals of the monoclinic form of terephthalic
acid were prepared by a thermal hydrolysis of 1,4-
dicyanobenzene according to the equation outlined
in Scheme 1.
[11]. Face indexed analytical absorption was calcu-
The 1,4-dicyanobenzene and H SeO used in a
2
3
lated using the Shelxtl program [12]. Four hundred
and sixty-seven independent re¯ections were used for
crystal structure solution and re®nement.
molar proportion of 1:4 were mixed together and
Table 1
Crystallographic data and ®nal re®nement parameters for mono-
clinic form of terephthalic acid
2.4. Structure solution and re®nement
Chemical formula
Molecular weight
Temperature ꢀK)
Space group
C
6
H
4
ꢀCOOH)
2
The structure was solved using the Shelx97
program [12]. The direct methods gave the positions
of carbon and oxygen atoms. The hydrogen atoms in
the ring were located from the difference Fourier
synthesis. The positions of the hydrogen atoms in
the phenyl ring were re®ned, but their isotropic
168.14
295
C2/m
Unit cell dimensions
Ê
a ꢀA)
Ê
b ꢀA)
8.940ꢀ2)
10.442ꢀ2)
3.790ꢀ1)
353.72ꢀ14)
2
Ê
c ꢀA)
thermal parameters were ®xed as 1,2U of the carbon
eq
3
Vol. V ꢀA )
Ê
atom linked directly to the H-atom. As the C±O bond
lengths in the carboxyl groups are symmetrically
equivalent ꢀspace group C2=m† the hydrogen atoms
Z
a
D
D
obs.
1.57
calc.
1.579
Ê
Radiation, MoKa ꢀA)
0.71073
0.128
of the COOH groups should be disordered. In the
2
Abs. Coef®cient, m ꢀmm 1)
Face-indexed anal. Abs.
correction, Tmin., Tmax
2
nal re®nement ꢀbased on F ), the hydrogen atom of
®
0.953, 0.975
carboxyl group was included with the occupation
factor of 0.5. The ®nal unweighted and weighted
2
2
b
R [F . 2sꢀF )]
b
0.0254
0.0318
1.029
2
2
wRꢀF )
agreement factors converged to R‰F Š ˆ 0:0254 and
S
wR ˆ 0:0318; a goodness-of-®t calculation resulted in
a value of 1.029 at the ®nal convergence. The ®nal
difference Fourier synthesis showed no peaks of
chemical signi®cance, the largest was 10.091 and
2
3
Ê
Drmax./Drmin. ꢀeA
)
1 0.091, 20.092
a
Measured by ¯oatation.
P
P
P
P
b
2
2
c
2
4 1=2
R ˆ uuF u 2 uF uu= F ; wR ˆ { ‰wꢀF 2 F † Š= wF }
:
o
c
o
2
o
o
2
1
2
2
2
o
2
23
w
ˆ s ꢀF † 1 ꢀ0:0001P† 1 0:0334P; where P ˆ ꢀF 1 2F †=3:
Ê
o
c
20.092 eA . Scattering factors for neutral atoms

Â
M. Sled z et al. / Journal of Molecular Structure 595 ꢀ2001) 77±82
80
Table 2
Fractional atomic coordinates and equivalent isotropic displace-
P P
i
2
Ê
ment parameters ꢀA ) U ˆ 1=3
eq
U a a a a †
ij i j i j
j
Atom
x
y
z
U
ij
O1
C1
C2
C3
0.3595ꢀ1)
0.1429ꢀ1)
0.0718ꢀ1)
0.2974ꢀ1)
0.39529ꢀ7)
0.5
0.3652ꢀ2)
0.1435ꢀ3)
0.0719ꢀ2)
0.3013ꢀ3)
0.0896ꢀ3)
0.0448ꢀ3)
0.0556ꢀ2)
0.0538ꢀ3)
0.38653ꢀ8)
0.5
Fig. 2. Molecular structure of terephthalic acid with numbering of
the atoms. The H atoms of the carboxyl groups are disordered with
the occupation factor of 0.5.
and the correction for anomalous dispersion were as
given in the Shelx97 program [13]. Final agreement
factors, details of data collection and crystallographic
data are listed in Table 1. Final atomic co-ordinates
and equivalent isotropic displacement parameters are
given in Table 2. Bond lengths and angles are listed in
Table 3.
tions applied, we therefore suppose that the additional
lines observed at the powder diffraction pattern
belong to the new, unknown up to now, modi®cation
of TPA. Unfortunately we have not been able to
obtain the pure phase of the new TPA modi®cation
by the method mentioned above ꢀeven at higher
temperatures and/or longer heating times). Thus,
both phases ꢀtriclinic form I and a new phase) co-
exist at the conditions applied above ꢀat a temperature
of 2508C). In order to obtain the pure phase of a new
modi®cation of TPA we decided to carry out the
synthesis of TPA by a thermal hydrolysis of 1,4-
dicyanobenzene at a temperature at which this modi-
3
. Results and discussion
In the course of the heating process of several
samples of TPA in an evacuated glass ampoule we
have stated, on the basis of the X-ray powder diffrac-
tion technique, that in all the samples heated below
2
508C only the triclinic form I is present. Thus the
triclinic form I is stable in the said conditions. Our
observations are consistent with those described by
Saska and Myerson [1,7]. The powder diffraction
diagram taken for the sample heated at 2508C is
more complicated ꢀFig. 1), because besides the
powder diffraction lines, which belong to the triclinic
form I, additional lines are also observed. As can be
seen from Fig. 1 these new lines do not correspond to
the triclinic phase II. Since we have not observed any
evidence on the decomposition of TPA at the condi-
®
1
cation has been observed. The thermal hydrolysis of
,4-dicyanobenzene leads to the formation of good
quality single crystals of TPA, which in fact appeared
to be the new phase obtained by the former procedure.
The prominent lines observed on the powder diffrac-
tion diagram calculated on the basis of the X-ray
single crystal data described below ꢀFig. 1d) are
consistent with those observed at the powder diffrac-
tion diagram taken for the samples heated at 2508C
ꢀFig. 1c).
The crystal of monoclinic form of TPA is built up
Table 3
Ê
Bond lengths ꢀA) and angles ꢀ8)
of hydrogen-bonded chains. This structural motif is
typical for the crystal structures of dicarboxylic
acids. The phenyl ring of TPA molecule is closely
planar, but not perfectly hexagonal. This distortion
can be attributed to the substitution effect of the
carboxyl groups. Both carboxyl groups of the TPA
molecule in the monoclinic modi®cation are coplanar
with the phenyl ring ꢀFig. 2). Similar orientation of
both carboxyl groups in relation to the phenyl ring can
be found in the triclinic form II, but different in the
O1±C3
C1±C3
C2±C2
1.248ꢀ1)
1.493ꢀ2)
1.384ꢀ2)
122.4ꢀ1)
120.1ꢀ1)
119.9ꢀ1)
C1±C2
C1±C2
1.369ꢀ1)
1.369ꢀ1)
a
b
a
O1±C3±O1
O1±C3±C1
C2±C1±C3
a
C2 ±C1±C3
118.8ꢀ1)
120.1ꢀ1)
120.1ꢀ1)
b
a
C1±C2±C2
C2±C1±C2
a
Symmetry code: x, 1 2 y; z
Symmetry code: 2x, y, 2z
b

Â
M. Sled z et al. / Journal of Molecular Structure 595 ꢀ2001) 77±82
81
Fig. 3. Molecular arrangement of terephthalic acid in the crystal shows the hydrogen-bonded chains: ꢀ001) projection ꢀa); and ꢀ010)
projection ꢀb).
triclinic form I wherein the carboxyl groups are
inclined by ^5.258 [3]. In the monoclinic form, akin
to the crystals of the triclinic forms, a orientational
disorder in carboxyl groups is observed. The positions
of the acidic hydrogen atoms as well as the single and
double carbon±oxygen bonds are found to be aver-
of one chain are adjacent to the benzene rings of the
neighbouring chain. This is in contrast to the triclinic
form I where the chains in one layer are shifted in such
a way that the benzene rings are adjacent to the
carboxyl dimers of the neighbouring chain. The
neighbouring layers in monoclinic modi®cation of
TPA are separated by c/2 while in both the triclinic
forms the layers are separated by c. Additionally in the
monoclinic modi®cation the chains of the neigh-
bouring layers are shifted by a/2 and b/2 whereas in
both the triclinic forms the chains of one layer lie
above the chains of the neighbouring layer.
Ê
aged ꢀboth C±O bonds have a distance of 1.248ꢀ1) A).
According to the neutron diffraction [14], neutron
scattering [15] and NMR [16] studies on the
terephthalic acid it can be concluded that the disorder
is of a dynamic origin caused by the double proton
exchange associated with an asymmetric double
minimum potential. The disordered hydrogen atom
of the carboxyl group forms two symmetrically
equivalent hydrogen bonds with O´´´O distance of
Ê
4. Supplementary material
2.691ꢀ2) A. They are slightly weaker than those
observed in the triclinic form I ꢀwith O´´´O distance
Additional materials comprising atomic coordi-
nates, anisotropic thermal parameters, full bond
length and angles for terephthalic acid have been
deposited with the Cambridge Crystallographic Data
Centre as supplementary publications no. CCDC
154875.
Ê
of 2.608 A) [3].
The hydrogen-bonded chains in the monoclinic
form of TPA are parallel to [101] direction and form
ꢁ
layers parallel to ꢀ101† plane as shown in Fig. 3. In
one layer, as in the triclinic form II, the benzene rings

Â
M. Sled z et al. / Journal of Molecular Structure 595 ꢀ2001) 77±82
82
References
Cottier, P. Dempsey, J.E. Rout, D.R. Stanley, A. Taylor,
Nature 366 ꢀ1993) 248±250.
[
10] A. Domenicano,G. Schultz,I.Hargittai,M.Colapietro, G. Porta-
lone, P. George, C.W. Bock, Struct. Chem. 1 ꢀ1990) 107±122.
11] Kuma Km-4, User's Guide.
[
1] M. Saska, A.S. Myerson, Cryst. Res. & Technol. 20 ꢀ1985)
201±208.
[
[
[
[
[
[
2] P.M. Brown, A.S. Myerson, AIChE J. 35 ꢀ1989) 1749.
3] B. Bailey, C.J. Brown, Acta Crystallogr. 22 ꢀ1967) 387.
4] S. Gaines, A.S. Myerson, AIChE Symp. Ser. 78 ꢀ1982) 42.
5] V.P. Gerasimov, N.K. Zharikov, V.S. Korobkov, I.V. Ovchin-
nikov, L.V. Rud, Zh. Prikl. Spectrosk. 34 ꢀ1981) 308±311.
6] L. Colombo, V. Volovsek, K. Furi c , J.R. Durig, J. Raman
Spectrosc. 21 ꢀ1990) 169±175.
12] G.M. Sheldrick, Shelxtl program system, Siemens Analy-
tical X-Ray Instruments Inc., Madison, WI, 1990.
13] G.M. Sheldrick, Shelxl97. Program for the solution and
re®nement of crystal structures, University of G oÈ ttingen,
Germany, 1997.
[
[
[
[
[
[
14] P. Fischer, P. Zolliker, B.H. Meier, R.R. Ernst, A.W. Hewat,
J.D. Jorgensen, F.J. Rotella, J. Solid State Chem. 61 ꢀ1986) 109.
15] B.H. Meier, R. Meyer, R.R. Ernst, P. Zolliker, A. Furrer, W.
H aÈ lg, Chem. Phys. Lett. 103 ꢀ1983) 169.
7] H. Fjaer, E.J. Samuelsen, J. Phys. C: Solid State Phys. 19
ꢀ1986) 5945±5956.
[
[
8] M. Saska, A.S. Myerson, AIChE J. 33 ꢀ1987) 848±852.
9] R.J. Davey, S.J. Maginn, S.J. Andrews, A.M. Buckley, D.
16] B.H. Meier, R.R. Ernst, J. Solid State Chem. 61 ꢀ1986) 126.