Density functional theory assessment of the thermal degradation of diclofenac and its calcium and iron complexes

Article abstract of DOI:10.1016/j.molstruc.2005.06.021

Thermogravimetric analyses of diclofenac sodium, its Ca²⁺ and Fe³⁺ complexes manifested a decreasing trend of the onset decomposition temperatures at which these compounds dissociated. The drop in the temperature was metal ion dependent; the sodium salt showed thermal stability up to 245°C, whereas the complexes started their degradation processes at temperatures starting from 90°C. While G* for the cleavage of the acetate moiety in the sodium salt was 63.76 kJmol^-1, it was 82.06 and 140.57 kJmol^-1 in the cases of Ca²⁺ and Fe³⁺, respectively. However, their complete fusion took place at 187.65, 150.34 and 98.77°C, respectively, displaying a reversed trend which is probably indicative of some catalytic part on the binding metals. Using the Gaussian 98 W package of programs, ab initio molecular orbital treatments were applied to diclofenac and its Ca²⁺ and Fe³⁺ metal complexes to study their electronic structure at the atomic level. The thermochemistry of diclofenac sodium was followed through the TG fragmentation peak temperatures using the density functional theory calculations at the 6-31G(d) basis set level. The FT-IR data were in good agreement with the theoretically calculated values. Single point calculations at the B3LYP/ 6-311G(d) level of theory, were used to compare the geometric features, energies and dipole moments of these compounds to detect the effect of the binding metal ions on the thermal dissociation of their diclofenac complexes.

Full text of DOI:10.1016/j.molstruc.2005.06.021

Journal of Molecular Structure 754 (2005) 61–70
www.elsevier.com/locate/molstruc
Density functional theory assessment of the thermal degradation
of diclofenac and its calcium and iron complexes
*
Ihsan M. Kenawi
Department of Chemistry, Faculty of Science, Cairo University, Giza, Egypt
Received 4 April 2005; revised 15 June 2005; accepted 17 June 2005
Available online 18 August 2005
Abstract
Thermogravimetric analyses of diclofenac sodium, its Ca2C and Fe3C complexes manifested a decreasing trend of the
onset decomposition temperatures at which these compounds dissociated. The drop in the temperature was metal ion
dependent; the sodium salt showed thermal stability up to 245 8C, whereas the complexes started their degradation processes
at temperatures starting from 90 8C. While G* for the cleavage of the acetate moiety in the sodium salt was 63.76 kJmolK1
,
it was 82.06 and 140.57 kJmolK1 in the cases of Ca2C and Fe3C, respectively. However, their complete fusion took place at
187.65, 150.34 and 98.77 8C, respectively, displaying a reversed trend which is probably indicative of some catalytic part on
the binding metals.
Using the Gaussian 98 W package of programs, ab initio molecular orbital treatments were applied to diclofenac and its Ca2C and Fe3C
metal complexes to study their electronic structure at the atomic level. The thermochemistry of diclofenac sodium was followed through the
TG fragmentation peak temperatures using the density functional theory calculations at the 6-31G(d) basis set level. The FT-IR data were in
good agreement with the theoretically calculated values.
Single point calculations at the B3LYP/ 6-311G(d) level of theory, were used to compare the geometric features, energies and dipole
moments of these compounds to detect the effect of the binding metal ions on the thermal dissociation of their diclofenac complexes.
q 2005 Elsevier B.V. All rights reserved.
Keywords: Thermal degradation; Density functional theory; Energy parameters; Dissociation constants; Diclofenac complexes
1. Introduction
different metals and organic bases such as
CH2COONa
Diclofenac sodium, monosodium {2-[(2,6-dichloroani-
lino) phenyl acetate, (dic), is germane as a mediator of the
inflammatory processes. It possesses high promising value
as a potential for inhibiting the conversion of arachidonic
acid to prostalglandins. It is also a potent analgesic for
painful attacks of non-rheumatic origin [1]. Due to its amino
acid salt structure, dic is easily converted into salts of
NH
Cl
Cl
those prepared by O’Conner et al [2] including 2-amino-,
2-methyl-1,3-propanediol, 2-amino-2-methylpropanol, tert-
butylamine, benzylamine and deanol. Diclofenac formed
covalently linked compounds with acrylic type polymers
(homopolymer of 2-hydroxyethyl methacrylate and its
copolymers with acrylamide) to obtain macromolecule
prodrugs [3]. It also has strong binding abilities forming
coordinate bonds with some d-metal ions forming dinuclear
macromolecules as Kovala-Demertzi et al. [4] reported.
Abbreviations dic, Diclofenac sodium.
*
Corresponding author. Tel.: 20 263 53228; fax: 20 256 85799.
E-mail address: ihsanata77@hotmail.com.
0022-2860/$ - see front matter q 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.molstruc.2005.06.021

62
I.M. Kenawi / Journal of Molecular Structure 754 (2005) 61–70
program [6]. The activation energies, E*, S*, G*, the
O
H2
Cl
Cl
activation enthalpy, H*, as well as the specific reaction rate,
kr, and dissociation equilibrium constant, Ks, for each step,
were calculated. The Arrhenius equation [7] was used to
calculate the reaction rates.
O
HN
O
–
+
M2
Some theoretical progress was made to cover the optimal
set and geometry of the above mentioned adducts. By
applying the Gaussian 98 W package of programs [8], the
density functional theory was used to implement and
provide new insights into their chemical behaviour through
a systematic study of the electronic structure and bonding
characteristics, comparing the energy quantities and dipole
moments of dic and its calcium and iron complexes. This
was done using the B3LYP level of theory at the 6-311G(d)
basis set. To give a comprehensive view of the reaction
mechanism resulting by fragmentation through the degra-
dation process, a theoretical follow up of the thermochem-
istry of dic was calculated at the dissociation peak
temperatures using the same theoretical method at the
6-31G(d) basis set level. The resulting frequencies were
compared with the experimental values obtained by FT-IR
studies to assess the applicability of the calculations, and to
detect the presence of imaginary frequencies.
NH
–
O
Cl
O
Cl
O
H2
Scheme 1. Divalent metal complexes.
Kenawi et al. [5] prepared some s and d-metal complexes
with diclofenac having the formulae Ca{(dic)2.2H2O},
Mg{(dic)2.2H2O}, Zn{(dic)2.2H2O} and Fe{dic}3.3H2O,
according to Schemes 1 and 2.
Calcium and iron can cause metabolic disorders, which
may be health hazardous, connected with both deficiencies
and excessive amounts of these ions. Calcium deposits in
bones and teeth; also in soft tissues. Iron normally functions
as a protein hydrolysate chelate, which transports oxygen
through the body. Thus, any instability in the content of
either of these metal ions would result in grave disorders.
The article [5] covering the synthesis of the calcium and
iron complexes of divalent 2 s and trivalent 3d electrons
lacked any information concerning the structural geometries
or kinetic parameters. These data are an essential backup for
the biomedical and pharmaceutical studies of such drugs; as
complexes provide mechanisms for electron storage, for
electron transport and for catalytic behaviour. A study of the
thermal stability of these complexes would help in the
design of new chemical profiles for the diclofenac binding
sites. The TG parameters characterize the stepwise pyrolysis
of these compounds, thus throwing more light on their
structure formed through several-stepped mechanisms
where the temperature range for each step may overlap,
leading to irregular mass-temperature curves that can be
difficult to analyze via practical methods. The change in the
values of these parameters on complex formation is traced
through mathematical calculations using the Coats–Redfern
2. Experimental
Thermogravimetric analyses were run on Shimadzu
TGA-50H in a nitrogen atmosphere with a rate flow of
(qZ20 ml/min) and (10 8C/min) using from 3 to 6 mg test
samples and heating up to 800 8C. Differential analyses
were run on a Shimadzu DTA-50 H in nitrogen atmosphere
at a rate flow 15 ml/min with 10 8C/min up to 1000 8C. The
thermally inert reference material used was a-Al2O3
(Shimadzu Corp.). A HANNA H18417 pH meter was used
to adjust the pH of the drugs with concentrated carbonate-
free NaOH. IR spectra were recorded on a Pye-Unicam SP3-
300 (500–4000 cmK1) and a MATTSON 5000 FT IR
(200–800 cmK1) spectrometers. A Magnetic Substance
Balance, Sherwood Scientific, Cambridge Science Park;
Cambridge, England (sample tube length 1.5 cm) was used
to detect the magnetic susceptibility of the ferric complex.
All chemicals used were analytical grade and water was
twice distilled.
O
Cl
Cl
O
–
Cl
NH
NH
+
M3
2.1. Metal chelates
O
–
.
3H2O
–
Cl
O
The complexes were prepared by mixing 100 ml of (0.
01 M) calcium nitrate and ferric chloride with 200 and
300 ml, respectively, of 0.01 M aqueous solutions of dic and
the pH was adjusted to w6.5, stirring the precipitate for 6 h.
The formed precipitate was filtered, washed and dried under
vacuum, thus, being ready for elemental, thermal analyses
and IR spectroscopy.
O
O
Cl
NH
Cl
Scheme 2. Fe complex.

I.M. Kenawi / Journal of Molecular Structure 754 (2005) 61–70
63
2.2. Methods of calculations
functionals involve both the values of the electron spin
densities and their gradients. Single-point refinements of dic
and its complexes were carried out at the B3LYP/
6-311G(d) level of theory on the fully optimized geome-
tries. The B3LYP is the keyword for the hybrid functional,
[10], which is a linear combination of the gradient
functionals proposed by Becke [11] and Lee, Yang and
Parr [12], together with the Hartree–Fock local exchange
function [13].
2.2.1. Integral method
The energy parameters resulting from the thermal
analyses were calculated using the Coats–Redfern equation
[6]. The energy of activation, (E*), was calculated from the
proposed integral equation:
 Á
Ã
log ln Wf=Wf KWt =T2
Most ferric ion complexes [14] are in the high spin state
t2g; eg , but those subjected to a strong field ligand are in
Â
ÃÀ ÃÁÃ
À
Á
Z log AR=fE 1 K2RT=E
KEÃ=2:303RTO
À
Á
the low spin state t2g . To write the correct molecule
specification (spin multiplicity) of the iron complex,
magnetic susceptibility of the solid complex was studied
and the multiplicity was found to be 2. This agrees with our
previous uv spectroscopic studies [15] that showed the
presence of charge transfer in dic. The frequency vibrations
and activation energies, at STP and at the TG dissociation
peak temperatures, were calculated at the B3LYP/6-31G(d)
level of theory, after being fully optimized.
Since the whole logarithmic term is practically constant
for the temperature range over which all the reactions
generally occur, therefore
 Á
Ã
À
ÃÁ
log ln Wf=Wf KWt =T2 Zlog AR=fE KEÃ=2:303RTO
Where AZArrhenius constant; WfZmass loss at com-
pletion of the reaction; WtZmass loss up to temperature T
(K); WfKWtZmass remaining at T (K); TZabsolute
temperature K; RZgas constant; E*Zactivation energy in
J moleK1; fZheating rate, deg. minK1
.
A plot of the left hand side against 1/T would give a
straight line from whose slope E* may be calculated; and
form whose intercept the Arrhenius constant is calculated.
The activation entropy, S*, the activation enthalpy, H*, the
free energy of activation, G*, and the specific rate, kr, and
stability, Ks, constants were calculated concurrently, from
the following equations:
3. Results and discussion
Elemental analysis showed the Ca2C and Fe3C
complexes of dic to be 1:2 and 1:3 M:L, respectively
3.1. Thermal processes
SÃ Z 2:303 log Ah=kTR
HÃ Z EÃ KRT
The thermograms of dic and its calcium and iron
complexes, Fig. 1, traced the weight loss steps into four
distinct changes, always losing water molecules in their first
mass loss step, thus agreeing with Fini et al. [16]. The results
in Table 1 are reported in terms of mass loss percent with the
coinciding temperature ranges of the TG peak decompo-
sition temperatures and DTA peak temperatures. The
cleaved groups and the remaining moieties are calculated
GÃ Z HÃ KTSSÃ
kr Z A expKEÃ=RT
S
DG ZKRT ln KS
2.2.2. Computational methods used
The computer used in the theoretical studies was a CPU:
AMD ATHLON XP 1.8 GHz, RAM: 256 DDR/333 MHz;
CACHE: 512 K; GRAPHIC PROC.: RADEON ATI7200 64
DDR mb.
The computations of single-point refinements, geometric
parameters, vibrational frequencies and activation energies
were carried out at the Density Functional level of theory
(DFT) using the Gaussian 98 W package of programs [8].
The DFT methods compute electron correlation via general
functionals of the electron density. These partition the
electronic energy into several components, which are
computed separately; the kinetic energy, the electron-
nuclear interaction, the coulomb repulsion and an
exchange-correlation term accounting for the remainder of
the electron-electron interaction [9]. The gradient corrected
Fig. 1. TG(–) and DTA (–-) thermograms of dic and its complexes.

64
I.M. Kenawi / Journal of Molecular Structure 754 (2005) 61–70
Table 1
Thermodynamic functions of dic and its calcium and iron complexes
Sample
Dic
Temp. range oC
% mass loss
Lost groups
DTA peaks K
Peak
20–95
19.9
15.1
4H2O
68.44
282
Endothermic
Exothermic
Exothermic
Endothermic
Exothermic
248–395
CH2COO
300
397–570
570–647
19.4
29.6
15.2
15.9
7.7
Phenyl
548
Phen.NHCl
NaCl
617
Residue
Dic2Ca
21–132
2H2O,CO2,O
2CH2,CO
Phen.NH
3Phen,NH,2Cl
CaCl2
54
Endothermic
Exothermic
Exothermic
Endothermic
136–299
300–420
420–664
269
325
747
13.1
48.4
14.8
Residue
Dic3Fe
Stable to 92 3H2O
92–153
5.6
13.8
4.0
2CO
135
174
355
488
Exothermic
Endothermic
Exothermic
Exothermic
160–323
3CH2,CO2,O2
3NH
325–400
410–670
61.7
16.4
6Phenyl, 3Cl
FeCl3
Residue
and therein depicted. Agreement with the experimental
values was within the range of 0.25 to 1.8%. An important
observation was drawn from these results; which is that the
thermal dissociation of these compounds took place
according to a, more or less, regular pattern. The acetate
moiety was the first to go, yet more readily so from its
attachment in the cases of the complexes, where the
enthalpy change commenced at a lower onset temperature;
almost 100 8C less than in the case of the sodium salt.
Meanwhile, the acetate groups present in the Ca2C and
Fe3C complexes were not completely dissociated until the
second mass loss step. On the other hand, the rupture of the
phenyl ring from the remaining dichloro–diphenylamine
moieties occurred at very close onset temperatures in the
two complexes, but at w100 8C less than the salt. The final
fusion resulted in all three cases in a metallic chloride
residue. The onset temperatures were almost equal in
the complexes but 150 8C less in the sodium salt. The results
that were confirmed by DTA, showed that coordination
caused steric constraints that weakened the bonds. A fact
that agrees well with their consecutive melting points
(283–5 Na, 250–1 Ca and 150 8C for Fe) that are previously
reported [5]. It would then be appropriate to assume that
thermal stability is affected by the metal ion type involved in
binding, to the order of NaOCaOFe.
These reactions may be considered bimolecular nucleo-
philic addition reaction, as seen from this data [17].
The activation energy, E*, necessary to cleave the acetate
group from the parent compounds was a positive quantity
that increased remarkably (as shown in Scheme 4) in
the case of the complexes; being 56.58, 84.76 and
149.11 kJmolK1 for NaC, Ca2C and Fe3C, respectively.
The acetate detachment, in the complexes, was com-
pleted in the second mass loss step; having values of E*
64.19 and 38.21 kJmolK1 for calcium and iron, respectively.
Thus, the order of activation energy required would be
EFeOECaOENa. This relates to the ionization potential
values of these metal ions, namely, 5.14, 11.87 and 30.64 eV
for NaC, Ca2C and Fe3C, respectively [14]. However, E*
required for the completion of the third and final mass loss
steps, in which the amine completely dissociates, mani-
fested a reversed trend where ENaOECaOEFe having the
values 127.07and188.92 for NaC, 62.96 and 148.87 for
Ca2C and 18.84 and 91.63 kJmolK1 for Fe3C. This is in
agreement with the melting points of these compounds. The
above mentioned results point to the weakening of the
London forces [18] due to complex formation.
Another point of importance is the very large drop in EFe
starting from the second loss step where the metal ion was
broken from the complex. This value may be ascribed to
some catalytic behaviour on part of the iron.
Thus, thermal dissociation might probably follow a
somewhat similar trend as in Scheme 3, after the loss of
water.
Cl
Cl
Cl
In an attempt to further, understand such a dissociation
mechanism, an investigation of the kinetic pyrolysis
parameters was carried out, utilizing the mathematical
algorithms introduced by Coats–Redfern [6] and Arrhe-
nius [7]. The activation energies and the stability and
specific rate constants are all depicted in Table 2 The
correlation coefficient, R2, that evaluates the accuracy of
the calculations, is also reported.
_
_
397 570
248 395
N
N
H
H
NH
Cl
Cl
Cl
CH2COONa
_
570 647
NaCl
Scheme 3.

I.M. Kenawi / Journal of Molecular Structure 754 (2005) 61–70
65
Table 2
Energy parameters of the thermal degradation of dic and its Ca2C and Fe3C complexes
Adduct
Dic
Step
R2
Ts K
E*
62.54
S*
K8.59
H*
59.78
G*
62.63
kr mK1
Ks
1
2
3
4
1
2
3
4
1
2
3
4
0.999
0.997
0.993
0.993
0.999
0.994
0.994
0.997
0.991
0.996
1.000
0.993
332
586
803
901
333
540
602
807
415
500
637
751
1.91E-01
1.34E-01
1.83E-01
2.11E-01
3.00E-01
2.22E-01
1.60E-01
1.55E-01
3.22E-01
0.63E-01
0.63E-01
1.20E-01
1.42E-10
2.11E-06
3.09E-09
1.33E-11
1.42E-13
2.14E-07
9.18E-07
1.87E-10
2.08E-18
1.64E-05
6.82E-03
1.35E-07
56.58
127.07
188.92
84.76
K20.56
K13.14
K6.91
51.70
120.39
181.42
81.98
63.76
130.94
187.65
82.06
Dic2Ca
Dic3Fe
K0.22
64.19
K17.29
K19.44
K10.15
12.26
59.69
69.03
62.96
57.94
69.65
148.87
149.11
38.21
142.15
145.66
34.05
150.34
140.57
45.83
K23.56
K20.23
K17.84
18.84
13.54
26.42
91.63
85.39
98.77
E*,H*,G* in kJmolK1; S* in JKmolK1.
.
The values of the entropy of activation, S*, of all three
compounds are negative numbers that are average in value,
thus the reaction rate would be also of moderate value;
neither faster nor slower than predicted by the collision
theory [19]. The presence of the methyl group, in the ortho
position, exerts both inductive and steric effects. These
cause an increase of the negative charge on the nitrogen
atom, which result in the lowering of entropy values [17].
Thus S* should increase once the acetate group is
dissociated. The only positive S* value of the activation
entropy is that of the first mass loss step in the iron complex,
which is 12.26 JKK1molK1, indicating an irreversible
change where the two CO lost have different ranges of
rotational and vibrational levels open to them. There also
exists a high probability of an increase in the rate of reaction
due to a catalytic behaviour of the ferric ions.
The activation enthalpies, H*, or the heats of atomiza-
tion, of the thermal dissociation of these compounds were
calculated to be positive values, thus indicating endothermic
reactions that need energy to break the existing molecular
bonds. In the sodium salt, the loss of water required
59.78 kJmolK1; this was decreased to 51.70 kJmolK1 for
the acetate fission. This is again much less in the case of the
break away of this group in the complexes, where it took
place along two loss steps with H* being equal to 81.98 and
59.69 kJmolK1for calcium; and 145.66 and 34.05 kJmolK1
for iron. The doubled and tripled amounts of enthalpy are
probably due to the presence of more bonds to break due to
the existence of two and three acetate groups attached to the
metal ions in the complexes. The methyl group, in the
o-position to nitrogen, exerts both inductive and steric
effects. The inductive effect increases the effective negative
charge on the nitrogen atom, in other words nucleophility
increases [17]. Therefore, H*, is higher after the acetate
group cleavage. The loss of one phenyl group in dic needed
double that amount (120.39 kJmolK1), while the final fusion
of the salt with NaCl remaining as residue, used up
181.42 kJmolK1. This is much less than the bond enthalpy
of a chlorine molecule (C242 kJmolK1), and DH8 of NaCl
(K411.15 kJmolK1) [19].
S* for the complete release of the acetate group
has close values in all three, being K20.56,K17.29
andK23.56 JKK1molK1, for NaC,Ca2C and Fe3C, respect-
ively. Yet, the release of the CO, CO2 and O in the first
mass loss steps in calcium and iron, possess a much larger
value of entropy, K0.22 and 12.26 JKK1molK1, probably
due to some catalytic behaviour of these metal ions which
increase the entropy by increasing the rate of reaction. In
the sodium salt, contrary to the expected, there was a large
drop of S* (K8.59 toK20.56) as the temperature was
raised. This is due to the loss of water of hydration which is
easily vaporized and the remaining molecules possessing
smaller entropy values, as they are solids and thus do not
move as freely. On increasing the temperature, the entropy
follows the regular trend in increasing gradually until it
reaches its peak value (K6.91 JKK1molK1) at the final
step of losing all the phenyl rings.
Meanwhile, the sudden decrease in the value of H*,
in the second and third loss steps of the iron complex
After the first step, the complexes show a similar trend of
decreasing entropy, which may be attributed to the existence
of a steric factor as a result of the large size of the existing
groups causing hindrance to the course of the formation of
complexions and thus decrease in S*. Once these groups
break away, the entropy follows its general course of
increasing with temperature.
Scheme 4.

66
I.M. Kenawi / Journal of Molecular Structure 754 (2005) 61–70
(34.05 and 13.54 kJmolK1), points out its catalytic
behaviour.In the case of Ca2C, a slight decrease in H* is
also noticed at the third mass loss step (57.94 kJmolK1).
This may be explained as a degree of instability caused by
steric constraints. However, the final fusion requires three
folds the amount (142.15 kJmolK1); probably caused by the
formation of the ionic CaCl2 residue; a value incomparable
to DH8 of both Cl2 and CaCl2 (K795.8 kJmolK1) [19]. On
the other hand, the final fusion of the iron complex exceeds,
by six fold, the preceding enthalpy value (85.39 kJmolK1).
The positive values of the free energies of activation, G*,
calculated from the thermograms, indicate that these
dissociation processes are non-spontaneous reactions that
do not do work and are not thermodynamically favoured.
They are endergonic processes obtaining their energy
content from the gradually raised temperature. Though
these values showed, in the case of dic, a somewhat regular
increasing trend accompanying temperature increase, the
pattern changed in the cases of the complexes. The loss of
the four water molecules of hydration of dic contained
62.63 kJmolK1, which is far more than the reported
(K220.6) standard free energy of formation of H2O(g)
[20]. The loss of the acetate group had a G* value of
63.76 kJmolK1, which is almost equal to that of the loss
of water, and that is due to the corresponding very low value
of S* (K20.56J KK1molK1), even though the temperature
each compound; except in the case of iron, where the
complete loss of the acetate and the three NH groups are 10
fold faster than all other loss steps being 0.632!10K1 and
0.627!10K1 minK1, respectively. It then decreases back to
1.2!10K1 minK1; due to the catalytic behaviour of the
Fe3C ions. Generally, these dissociation processes are of
moderate rate, a fact that agrees with the not very large
negative number of the entropy values of these reactions
[19]. The cation type slowed down the rate slightly from
sodium to calcium, but the catalytic abilities of iron
somewhat counteracted this effect, thus; kFeOkNaOkCa.
The dissociation equilibrium constants, Ks, of the TG mass
loss steps of dic and its complexes, ranged in the order of
10K3 up to 10K18. According to Le Chatelier’s principal,
the increase of temperature favours the products of an
endothermic reaction. The resulting Ks values increased as
the thermal dissociation proceeded, but increased once more
as the final fusion processed. By comparing Ks for each step
in all three compounds, it appears that the calcium complex
reached equilibrium at a slower rate than the sodium salt;
and was less stable as a result of the binding with NH which
probably caused steric constraints. The equilibrium constant
of Ca2C was 2.14!10K7 and that of dic was 2.11!10K6
.
The Fe3C complex, on the other hand, exerted a catalytic
effect and reached equilibrium quickly with a much higher
Ks value of 1.65!10K5. Naturally, a catalyst will increase
both the forward and reverse rate processes so that a reaction
will reach equilibrium sooner. This is quite apparent in the
third loss step where Ks is 6.82!10K3 for Fe3C, while it is
3.09!10K9 and 9.18!10K7 for NaC and Ca2C, respect-
ively. Stability of the complexes is reached quicker and the
products are more stable due to the higher values of electro-
negativity of these metal ions (Na 0.9, Ca 1.0, Fe 1.8) [18].
The formed CaCl2 and FeCl3 consequently form stronger
ionic bonds than NaCl; Ks being 1.87!10K10, 1.35!10K7
and 1.33!10K11, respectively.
was 586 K. The value of G* was doubled (130.94 kJmolK1
)
on the loss of one phenyl group and went further up to
187.65 kJmolK1 for the final fusion of the compound
yielding NaCl. This is so much higher than the reported
value (K384.14 kJmolK1) for DG8 of the formation of solid
NaCl [19]; thus confirming the endergonic character of the
process. Meanwhile, the decrease of the values of G* in the
complexes was noticed. The first mass loss step (82.06 and
140.57 kJmolK1 for Ca2C and Fe3C, respectively) suddenly
dropped in the case of iron to 45.83 kJmolK1. This is a result
of its catalytic character, which lowered the saddle point of
its energies of activation. The second and third mass loss
steps in the calcium complex were almost equal (69.03 and
69.65 kJmolK1) and almost half that of the corresponding
loss step in dic, probably due to the presence of two
diclofenac groups in the complex. The third loss step in the
iron complex had a G* value of 26.42 kJmolK1, half the
amount of the previous step that completed the acetate
group cleavage. This, again, may be ascribed to the catalytic
character of Fe3C, being activated by its breaking loose. In
the final step in both complexes, G* increased again as
dissociation was completed.
The amount of reactants at the dissociation peak
temperatures were 32.05, 39.99 and 51.86% for dic, calcium
and iron complexes. This means that the iron complex was
the one that was least thermally stable, followed by the
calcium complex and the sodium salt was the most stable of
all three. Thus, it may be concluded that the replacement of
the NaC ion by these metals caused steric hindrance due to
coordination that affected the thermal stability of the drug,
which is confirmed by the lowering of the respective
melting points.
3.2. Theoretical simulations
These pyrolysis results show that the energies of dic are
cation dependant on the s and d orbital metals studied in this
work. Complex formation resulted in a change in some
physical properties, as well as steric constraints that must be
operative to give energy-preferred orientation of the system
resulting in stronger binding free energy.
3.2.1. Thermochemistry of dic fragments
The thermochemistry of dic and its fragments that
resulted during the thermal dissociation were studied
theoretically using the density functional theory, applying
the B3LYP/6-31G(d) level of theory. By default, analysis is
carried out at 298.15 K and 1 atmosphere of pressure using
the principle isotope for each element type [21]. However,
The specific reaction rates, kr, were all of the order 10K1
showing slight differences between the mass loss steps for
,

I.M. Kenawi / Journal of Molecular Structure 754 (2005) 61–70
67
Table 3
Thermal energies predicted at the B3LYP / 6-31G(d) level of theory
Loss step fragments
SCF,au (STP)
DE,au (TS K)
DE,kJ
DH, au (Ts K)
DH,kJ
TsZ586 K
Dic
1579.631496
1268.745554
344.876663
0.198513
0.169251
0.049348
522.93
445.85
129.99
0.199105
521.96
445.90
131.55
Cl2FNHF
CH3COONa
DE*Z52.91 kJmolK1
TsZ803 K
Cl2FNHF
Cl2FNH2
0.170084
0.050181
DH*Z55.47 kJmolK1
1268.745554
1064.874363
204.936088
0.169998
0.108631
0.108856
447.82
286.16
286.75
0.170111
448.11
286.05
284.24
0.109111
F
0.108421
DH*Z122.18 kJmolK1
DE*Z125.07 kJmolK1
by specifying the READISOTOPE option to the FREQ
keyword in the route section, it was possible to study the
thermochemistry of these reactions at the specified peak
temperatures, Ts.
dic and its calcium complex were simulated through single
point energy calculations at the B3LYP level of theory with
a 6-311G (d) basis set after being fully optimized. This high
basis set has been chosen to detect the dipole moments at a
very accurate level. The first optimization step of the
complexes was carried out on the Hyperchem 6 package
using ZINDO 1 semi-empirical methods. To obtain a higher
level of optimization these structures were run on the
Gaussian98 W package. The sodium salt was optimized
using the B3LYP level with a basis set of 6-31G (d). The
calcium complex was a difficult optimization case, thus
the approximate matrix was improved through use of the
computed first derivatives obtained from the STO-3G size
basis set. The iron complex would not be optimized as
convergence would not be met, an error message repeatedly
appearing at the L502 link. Thus, a comparison of its dipole
moment with the other compounds was restricted to the
ZINDO1 level of theory.
The difference, DE, for the two intermediate dissociation
reactions whose energies were computed at STP and the
consecutive Ts temperatures, was taken as an equivocal
value of the amount of activation energy, E*, that would be
required to complete such reactions. A slight deviation from
the experimental values might be because of the choice of
the medium sized basis set 6-31G (d). The choice of such a
size of basis set was because higher sets may lose the
negative frequency values that indicate a saddle point of
energy. This imaginary frequency is due to the presence of
an intermediate complexion that takes part in the reaction.
However, one may reliably compare the computed values to
the experimental ones since these results, Table 3, are much
closer than expected. The change in the energies, as
dissociation processes, followed a similar trend to the
experimentally calculated parameters. That is, cleavage of
the acetate group was computed to have DE*Z53 kJmolK1
and DH*Z55 kJmolK1, whereas the experimental values
were 56 kJmolK1 and 52 kJmolK1, respectively. The loss of
one phenyl ring from the dichloro diphenylamine was
The increasing electronegativity [18] of the substituted
metals will cause the carbon in the COOKgroup to supply
more p orbital character which results in distortion, of the
geometric parameters, from the ideal values. The output
standard orientation of the optimized structures of all three
computed at DE*Z125 kJmolK1 and DH*Z122 kJmolK1
whereas the experimental values were 127 kJmolK1 and
120 kJmolK1
,
Table 4
FT-IR and theoretically calculated frequencies, cmK1, at 6-31G(d) level,
for dic and its complexes
.
As the calculation of such complicated systems becomes
more complex, the theories must still be tested. Thus, a
comparison of the experimental ir frequencies of dic with
the computed values is important. The results, Table 4, were
in good agreement. In the simulated frequencies, there
appeared four negative values, thus confirming the presence
of an intermediate complexion involved in the procession of
the thermal dissociation of dic.
Assignments
Ring stretch
Experimental
Theoretical
417
418
459
460
513
516
548
545
683
687
718
720
gCH–Cl
gCH
dCH
760
759
869
867
1161
1292
1409
1454
1472
1625
2968
3422
1162
1290
1405
1454
1471
1622
2963
3429
nC–N
dsCOOK
nM
dCH2COOK
dNH
nCH2
3.2.2. Electronic structural modeling.
The substitution in dic of the sodium ion by either
calcium or ferric ions, or in other words the formation of
coordinated adducts, resulted in steric constraints that
initiated a change in geometry to give an energy preferred
orientation. Thus, the molecular energies and structures of
nNH

68
I.M. Kenawi / Journal of Molecular Structure 754 (2005) 61–70
Table 5
Na
Bond angles, order and lengths of dic and its complexes computed at the
B3LYP/6-311G(d) level
O
Atom no.
NaC
Ca2C
Fe3C
H
Bond angles, in degree
O2 C3 O4
O2 C3 C5
C5 C6 C7
C6 C7 N8
C7 N8 C9
N8 C9 C11
C9 C10 Cl12
Bond order
M1–O2
114.998
121.059
127.340
128.458
112.630
117.134
120.428
120.947
117.108
123.702
130.795
116.013
110.280
116.663
115.690
122.606
127.391
125.203
124.799
128.260
116.675
C
N
Cl
dic
0.100
0.917
0.954
1.381
1.155
1.101
1.102
0.863
0.831
–
0.616
0.992
0.983
1.617
1.212
0.947
1.032
0.862
0.862
0.179
0.186
1.018
0.967
1.688
0.950
0.985
0.991
0.852
0.863
0.339
C6–C5
C5–C3
C3–O4
C3–O2
C7–N8
Ca
N8–C9
C11–Cl13
C10–Cl12
N8–M1
Bond length, in Angstrom
M1–O2
C6–C5
3.211
1.514
1.492
1.308
1.363
1.405
1.371
1.741
1.743
2.541
1.514
1.514
1.451
1.411
1.456
1.446
1.777
1.779
2.445
2.451
1.485
1.483
1.294
1.383
1.427
1.462
1.735
1.745
2.264
dic2Ca
C5–C3
C3–O4
C3–O2
C7–N8
N8–C9
C11–Cl13
C10–Cl12
N8–M1
Fig. 2. Optimized structures of dic and its calcium complex.
compounds, Fig. 2, shows all the atoms to be occupying the
x, y and z planes with the two phenyl rings of the dichloro–
diphenyl amine at a dihedral angle of w1088. Thus, the
cations did not affect the torsion angles of these two rings.
However, some bond angles (Table 5) varied in their values
due to steric constraints resulting on complex formation and
consequently, a slightly different hybridization. The angles
O2-C3-O4, C5-C3-O2 and C5-C6-C7 (Scheme 5) were
increased by w1 to 38 indicating that the CZO and
CH3COOK groups moved away from the diphenylamine
moiety. At the same time, the NH between these two phenyl
rings showed a decrease in its angles indicating their
simultaneous approach towards each other. This agrees with
the fact that as the electronegativity of the substituted metal
ion increases, the extent of decrease in these angles
increases [18]. At the same time, the C9-C10-Cl12 angle
increased in the case of Ca2C and decreased in the
Fe3Ccomplex.
–
electronegativity is higher than those of the other two. The
presence of the d5 electrons in the ferric ion strains the
oxygen slightly away. As for the nitrogen attachment
between the phenyl rings, the bonds were longer in both
complexes, showing that the drug becomes less stable at this
site.
The bond order of these compounds, Table 5, assessed
their fragmentation patterns that are depicted in Table 1 The
M–O bond for each has the smallest value (0.100 NaC and
0.186 Fe3C) except in Ca2C(0.616), where the Ca–N order
(0.179) is smaller. The Fe-N order is 0.339.These values are
4
O
2
O
3
C
The strain in the bond lengths, Table 5, accounting for
the change in the nuclear-nuclear repulsions and electron–
nuclear attractions was accentuated at the M–O and M–N
bonds. Because of the electronegativity values of the
cations, the M–O bond lengths are shorter in the complexes
than in dic. The coordinate M–N bonds are O2.4A8, and
thus unstable. The C6–C5 and C5–C3 bonds in Ca2C, of the
acetate group, did not differ much from dic; but in the iron
complex were shorter. The C–O bonds were longer in
Ca2Cthan in dic, but shorter in Fe3C (Fig. 3), though its
1
M
5
CH2
13
Cl
8
HN
11
6
7
9
Cl
12
10
Scheme 5.

I.M. Kenawi / Journal of Molecular Structure 754 (2005) 61–70
69
central metal ion; and since Fe3C contains d5 electrons in its
outermost shell, the mCaOmFe is reasonable.
Since mNa is the largest value, 11.4 D, (Ca being 4.3 D) it
is apparent that all the data coincide. The dipole moment
values agree with the fact that 100% ionic character
possesses the largest m, decreasing as the character becomes
partial. The lone pair on nitrogen gives a very asymmetric
distribution of the electronic charge cloud around this atom,
producing a large atomic dipole (lone-pair dipole) that
opposes the molecule bond dipoles, thus reducing the
resultant moment [19].
Fe
dic3Fe
Fig. 3. Optimized structure of Fe3C complex.
4. Conclusion
in accordance with their very long bond lengths, showing
that as in the TG dissociation process, the metals break loose
at the very start. In dic, the C6–C5 bond order (0.917) is
smaller than that of C5–C3 (0.954); but in the complexes it
has a larger value. This explains the dissociation of the
acetate moiety as a whole in the sodium salt, while in the
complexes the COOK breaks away first. Though the C-Cl
bond order in all three is w0.860, the delocalized p-electron
cloud of the phenyl ring exerts a mesomeric effect that will
result in delayed activity, namely, the stabilization of these
bonds [22]. On the other hand, the nitrogen atom, which has
bond orders in the range 0.947 to 1.102, is between two
phenyl rings. Therefore, it will break its attachments in the
same order as the TG fragmentation process.
To conclude, the present study that dealt with the
mathematical depiction of the energy pyrolysis parameters,
as well as, the theoretical assessment of these results, lead to
the following points:
(1) The activation energies, E*, G*, the activation
entropies, S* and the heats of atomization, H*, led to
the fact that by replacing the NaC by Ca2C or Fe3C the
drug became less thermally stable. The results showed
that iron manifested catalytic character that increased
its kr and Ks values of its latter loss steps.
(2) The thermochemistry of these compounds, that was
studied via ab initio calculations at the B3LYP/6-
31G(d) level of theory, assessed the experimental
results to a quite satisfactory extent.
The dipole moments and energies simulated at the
B3LYP/ 6-311G(d) level of theory are depicted in Table 6.
The SCF energies of dic, Ca2C and Fe3C decreased as the
cation was changed (NaOCaOFe). This means that the
cation type affects the stability of the drug. This is confirmed
by the energy gap values DEZLUMOKHOMO, that
increase in the order EFeOECaOENa. The nuclear-nuclear
(3) Experimental FT-IR data was used to check the validity
of the theoretical calculations. The data were in good
agreement.
(4) Single point calculations at the B3LYP/6-311G(d) level
of theory to compare the geometric parameters,
energies and dipole moments of the compounds,
assessed the fragmentation loss steps to satisfaction.
repulsion energy showed
a
much higher value
(13163.2473au) within the iron complex molecule. These
results match the great difference in melting points of the
studied compounds.
Acknowledgements
The magnetic moments, m, of dic and its complexes
clearly assess the above results. The m of iron was found,
using the ZINDO 1 level of theory, to be w9 D smaller than
that of the calcium complex resulting at the same level of
calculation. The direction of m was always towards the
I would like to extend my deepest gratitude to Prof. Rifaat
Hilalforallowingmetohavethepleasureofenjoyingthework
on his Gaussian 98 W package of programs. I would also like
to thank Mr. Walid M. Ibrahim for his patient guidance during
my learning of how to feed in the input.
Table 6
Energies, au, and dipole moments, D, of dic and its complexes computed at
the 6-311G(d) level
References
Cation SCF
LUMOKHOMO
K1827.51 5.31E-03
K4007.95 5.89E-03
N–N rep. energy
m
NaC
Ca2C
Fe3C
1935.37
5729.25
11.353
4.275
[1] P. Moser, A. Sallmann, I. Wiesenberg, J.Med.Chem. 33 (1990)
2358–2368.
[2] K.M. O’Connor, O.I.Corrigan, Inter. J.Pharm. 226 (1-2) (2001)
K6204.86 7.30 (a-orbitals)
7.32 (b-orbitals)
13163.25
ZINDO
1w8.95a
163–179.
[3] H. Namazi, M. Babazadeh, A. Sarabi, A. Entezami, J.Poly. Mater. 18
(3) (2001) 301–312.
a
ZINDO 1 gave mCaZ17.63 D.

70
I.M. Kenawi / Journal of Molecular Structure 754 (2005) 61–70
[4] D. Kovala-Demertzi, D. Mentzafos, A. Terzis, Polyhedron 12 (11)
(1993) 1361–1370.
[15] I.M. Kenawi, B.N. Barsoum, M.A. Youssef, J. Parm. Biomed Anal. 37
(40), (2005).
[5] I.M. Kenawi, B.N. Barsoum, M.A. Youssef, Under publication.
[6] A.W. Coats, J.P. Redfern, Nature(London) 68 (1964) 201.
[7] S. Arrhenius, Z.Phys.Chem. 4 (1889) 226; see: Trans.Farad.Soc., 33
(1937) 72.
[16] A. Fini, M. Garuti, G. Fazio, J. Alvarez-Fuentes, M.A. Holgado, J.
Pharm.Sciences 90 (12) (2001) 2049–2057.
[17] L.P. Hammett, Physical Organic Chemistry, McGraw-Hill, Koga-
kusha, Ltd, Tokyo, 1970.
[8] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb,
J.R. Cheeseman, et al., Gaussian 98, Revision A.6, Gaussian, Inc.,
Pittsburgh PA, 1998.
[18] L. Pauling, General Chemistry, W.H.Freeman and Co, USA,
1970.
[19] R. Chang, Physical Chemistry for the Chemical and Biological
Sciences, University Science Books, USA, 2000.
[20] P. Matthews, Advanced Chemistry, Cambridge Univ. Press, Great
Britain, 1997.
[9] P. Hohenberg, W. Kohn, Phys. Rev. B 136 (1964).
[10] W. Kohn, L.J. Sham, Phys. Rev. A 140 (1965) 1133.
[11] A.D. Becke, Phys. Rev. A 38 (1988) 3098.
[12] C. Lee, W. Yang, R.G. Parr, Phys. Rev.B 37 (1988).
[13] R.L. Flurry Jr, Molecular Orbital Theories of Bonding in Organic
Molecules, Marcel Dekker, Inc., New York, 1968.
[14] F.A. Cotton, G. Wilkinson, C.A. Murillo, M. Bochmann,, sixth ed.
Advanced Inorganic Chemistry, John Wiley and Sons, Inc., Pte Ltd,
Asia, 2003.
[21] J.B. Foresman, A. Frisch,, 2nd ed Exploring Chemistry with
Electronic Structural Methods, Gaussian, Inc., Pittsburgh, USA,
1996.
[22] J. March,, fourth ed. Advanced Organic Chemistry: Reactions,
Mechanisms, and Structure, John Wiley and Sons, Inc.(Asia), Pte
Ltd, Asia, 2000.