Thermal behaviour of diclofenac, diclofenac sodium and sodium bicarbonate compositions

Article abstract of DOI:10.1007/s10973-006-8182-1

This work aims to investigate the thermal behaviour of diclofenac, diclofenac sodium, and NaHCO₃ both as single components and binary mixtures. In particular, the melting point and latent heat of fusion binary diagrams of the diclofenac sodium/diclofenac mixtures at different mole fraction compositions were investigated in order to gain information about the thermal behaviour of their solid mixtures. A good agreement between liquidus curves, calculated by the Schroeder-Van Laar equation from fusion enthalpies and temperatures, and the experimental results was found. For all binary compositions, an endothermic effect at 153°C, probably due to the eutectic fusion, is present.

Full text of DOI:10.1007/s10973-006-8182-1

Journal of Thermal Analysis and Calorimetry, Vol. 90 (2007) 3, 903–907
THERMAL BEHAVIOUR OF DICLOFENAC, DICLOFENAC SODIUM
AND SODIUM BICARBONATE COMPOSITIONS
*
Irene Pasquali, R. Bettini and F. Giordano
Department of Pharmacy, University of Parma, via Usberti 37/A, 43100, Parma, Italy
3
This work aims to investigate the thermal behaviour of diclofenac, diclofenac sodium, and NaHCO both as single components and
binary mixtures. In particular, the melting point and latent heat of fusion binary diagrams of the diclofenac sodium/diclofenac mix-
tures at different mole fraction compositions were investigated in order to gain information about the thermal behaviour of their
solid mixtures. A good agreement between liquidus curves, calculated by the Schroeder–Van Laar equation from fusion enthalpies
and temperatures, and the experimental results was found. For all binary compositions, an endothermic effect at 153°C, probably
due to the eutectic fusion, is present.
Keywords: binary melting point diagram, diclofenac, diclofenac sodium, eutectics
Introduction
different proportions of unreacted DCLNa, DCLH
and NaHCO
3
.
Diclofenac, 2-[(2,6-dichlorophenyl)amino]benzeneace-
tic acid (DCLH), is a non-steroidal anti-inflammatory
To achieve a better interpretation of the thermal
phenomena shown by these multicomponent systems,
DSC analyses of the single components and of possible
binary mixtures were carried out and the results are re-
ported in this paper, along with the corresponding bi-
nary melting point and the heat of fusion diagrams [5].
a
drug with analgesic properties. DCLH (pK 3.80 at
–
5
2
2
5°C) shows very low aqueous solubility (6·10 M at
5°C) [1]. Salts from a variety of inorganic and organic
bases [2, 3] (in particular the sodium salt, DCLNa) are
commonly used to improve solubility and, conse-
quently, the dissolution rate of the drug.
In a previous work [4], DCLNa was chosen as
model drug to study the solubility in supercritical car-
Experimental
Materials
2
bon dioxide (SC–CO ) of the tetrahydrate form in
DCLNa (Lisapharma, Como, Italy) and sodium bicar-
bonate (Carlo Erba, Milan, Italy) were used as pur-
chased.
comparison with the anhydrous form. Solid-state
modifications due to the interaction of DCLNa with
SC–CO
ning calorimetry (DSC), thermogravimetric analysis
TG), hot stage microscopy (HSM), Fourier transform
2
were also investigated by differential scan-
(
Preparation of DCLH
infrared spectroscopy (FTIR) and Karl Fischer
titrimetry. The DSC trace of untreated DCLNa
–1
A DCLNa solution in water (10 g L ) was acidified
with 0.1 N HCl until a pH value of 3–4 was reached to
achieve the complete precipitation of DCLH. The pre-
cipitate was filtered, thoroughly washed with distilled
water, air dried and stored in a desiccator. The forma-
tion of DCLH was confirmed by Fourier transform
infrared spectroscopy.
(
Fig. 1a) showed an endothermic peak in the 40–60°C
temperature range (assigned to water removal, TG),
followed, above 260°C, by complex endo–exo phe-
nomena mainly due to decomposition; on the other
hand, the DSC patterns of the solid phase recovered
2
after SC–CO treatment evidenced a series of thermal
events between 100 and 170°C (Fig. 1b). This pecu-
liar behaviour was attributed to possible reaction of
DCLNa with the supercritical fluid, mediated by crys-
Preparation of binary mixtures
Binary physical mixtures of DCLH/DCLNa,
DCLNa/NaHCO , DCLH/NaHCO in different mole
tal water (in the case of DCLNa·4H
moisture with anhydrous DCLNa. DCLNa, in the
presence of SC–CO and water, generated solid resi-
2
O) or surfacial
3
3
fraction compositions (ranging from 0.1 to 0.9) were
prepared by simple gentle mixing in a mortar to
achieve a randomly even distribution of components
prior to scanning.
2
dues whose thermal behaviour was difficult to inter-
pret, owing to the formation of mixtures containing
*
Author for correspondence: giordano@unipr.it
1
388–6150/$20.00
Akadémiai Kiadó, Budapest, Hungary
Springer, Dordrecht, The Netherlands
©
2007 Akadémiai Kiadó, Budapest

PASQUALI et al.
ciated to the broad double peak, showed a mass loss
of 25% in the 90–200°C interval.
Thermal behaviour of single components
Diclofenac
TG and DSC traces of DCLH are reported in Fig. 2. The
DSC pattern of DCLH shows a sharp endothermic peak
at 181°C, corresponding to fusion [6]. The latent heat of
–1
fusion, DH= –133 J g was calculated as the area under
the peak by numerical integration. DCLH mass loss un-
der dynamic flow of dry nitrogen was below 10% before
fusion, while massive degradation of the compound af-
ter fusion was evidenced by TG.
Fig. 1 DSC traces of DCLNa a – before and b – after super-
critical treatment [4]
Methods
Differential scanning calorimetry was performed with
a Mettler DSC 821e (Mettler Toledo, USA) driven by
STARe software (Mettler Toledo) on 4–5 mg samples
(
Microbalance MT5 Mettler Toledo) placed in sealed
and pierced 40 μL aluminium pans. Scans of each
component and the binary mixtures were carried out
–
under a flux of dry nitrogen (100 mL min ) between
1
–
1
0–290°C at 10 K min . Calibration of temperature
3
and enthalpy values was performed with indium.
Thermogravimetric analyses (TG50, Mettler To-
ledo, USA) were carried out on samples placed in 70 μL
alumina pans with a pierced cover. The samples were
Fig. 2 a – TG and b – DSC traces of DCLH
–1
heated under a flux of nitrogen (200 mL min ) at
10 K min in the 30–290°C temperature range.
Diclofenac sodium
–1
The DSC trace of DCLNa (Fig. 3) shows an
endotherm at 280°C, followed by an exotherm. This
result is indicative of melting followed by decomposi-
tion. The melting point of the drug was 287°C with
Each sample was analysed at least in triplicate.
All data are expressed as mean value.
Only first heating DSC and TG scans were con-
sidered owing to massive decomposition of DCLH,
–
1
an enthalpy variation (DH) of –85.9 J g (within the
65–295°C range).
3
DCLNa and NaHCO on heating.
2
NaHCO
The DSC trace of NaHCO
thermic events at 96°C (DH= –27.3 J g ) and at
3
Results and discussion
3
(Fig. 4c) shows two endo-
–
1
As mentioned above, the treatment of DCLNa with
SC–CO , in presence of water, both in the case of
–
1
2
150°C (DH= –717.2 J g ). TG (Fig. 4a) evidences the
thermal decomposition of NaHCO , that occurs gen-
crystal water and surfacial moisture, generated inside
the vessel of the supercritical fluids apparatus, mix-
3
erating H O and CO and sodium carbonate, which
2
2
tures of unreacted DCLNa, DCLH and NaHCO
different proportions depending on the experimental
conditions, according to the following reaction:
3
in
starts approximately at 50°C ending at 170°C with a
measured mass loss of 36.8% (calculated 36.9%). The
st
agreement between the DSC (Fig. 4c) and 1 deriva-
tive of the TG (Fig. 4b) traces is a further confirma-
tion of the decomposition of the salt.
DCLNa+H
2
O+CO
2
« DCLH+NaHCO
3
(1)
The broad double peak recorded between 158
and 175°C (Fig. 1b), therefore, includes all thermal
events due to the complex mixture of chemicals pres-
ent inside the vessel [4]. In particular, the in situ reac-
tions are mainly responsible for the thermal behav-
iour. The TG reported in our previous work [4], asso-
Thermal behaviour of binary systems
DCLNa/DCLH mixtures
The thermal traces of the DCLNa/DCLH mixtures at
different mole fraction compositions are reported in
9
04
J. Therm. Anal. Cal., 90, 2007

DICLOFENAC, DICLOFENAC SODIUM AND SODIUM BICARBONATE COMPOSITIONS
Fig. 3 DSC trace of anhydrous DCLNa
Fig. 6 Phase diagram of the DCLH/DCLNa binary system
(
empty squares: experimental data obtained from DSC
curves; see Fig. 7 — and - - - calculated temperature
composition curve)
calculated using the Schroeder–Van Laar equation in
its simplified form [7]:
A
DH æ 1 1 ö
f
lnx=
ç
ç
è
-
Tf Tf
÷
÷
ø
(2)
A
R
where x is the mole fraction of the more abundant
component of a mixture whose melting terminates at
A
–1
A
T
f
(K); DHf (cal mol ) and Tf (K) are the enthalpy
of fusion and melting point of the pure component, re-
spectively, and is the gas constant,
.9869 cal mol K . For mixtures that form a true
R
Fig. 4 a – TG, b – DTG and c – DSC of NaHCO
3
–
1
–1
1
eutectic (i.e., complete miscibility of the liquid phases
and immiscibility of the solid phases) the melting
point of each component is lowered by the presence
of the other component and, therefore, the melting
point decreases from both sides of the diagram. A sat-
isfactory agreement between theoretical and experi-
mental liquidus curves could be found only regarding
binaries with xDCLH>0.6 (150–181°C temperature
range). For mixtures with xDCLH<0.6, it was practi-
cally impossible to evaluate accurately the thermal
parameters due to extensive decomposition of both
components, thus leading to a clearly unsatisfactory
agreement between calculated (dashed line) and ex-
perimental values (above 180°C). All binary compo-
sitions show at 153°C an endothermic effect, proba-
bly due to the eutectic fusion.
A second endothermic effect at constant temper-
ature was also observed experimentally at higher tem-
perature (250°C). Unfortunately, the extensive de-
composition undergone by DCLH above its melting
point makes quite difficult to give a convincing
explanation of this phenomenon.
Aiming to demonstrate the presence of a true
eutectic between DCLH and DCLNa at 153°C, data
obtained from DSC analyses were used to draw the la-
tent heat of fusion diagram (Fig. 7). The theoretical
Fig. 5 DSC traces of the mixtures DCLH/DCLNa at different
mole fractions
Fig. 5. Samples from 0.1 to 0.7 DCLNa mole fraction
exhibited two endothermic peaks around 153 and
2
(
0
50°C. A further endothermic event around 170°C
Fig. 1b) was present in the DSC trace of the 0.8 and
.9 mole fractions. All samples exhibited several
endo-exothermic events above 250°C. Thermal data
measured from the DSC traces of samples having dif-
ferent compositions were used to draw the binary
melting point diagram (Fig. 6). In the same diagram
the experimental data are reported along with the data
J. Therm. Anal. Cal., 90, 2007
905

PASQUALI et al.
Fig. 8 DSC traces of the DCLH/NaHCO
mole fractions
3
mixtures at different
Fig. 7 Latent heat of fusion diagram at 153°C of the binary
mixture DCLH/DCLNa (empty squares: calculated
data, empty diamonds: experimental data; bars repre-
sent the standard deviation, n=3)
It may be useful to recall the following reactions
which can take place when DCLH and NaHCO
heated together,
3
are
DCLH+NaHCO
NaHCO ® Na
2 2
DCLH+Na CO «2 DCLNa+H O+CO
3
«DCLNa+H
2
O+CO
2
(4a)
(4b)
(4c)
heat of fusion at 153°C for each DCLNa/DCLH mix-
ture was calculated with the following equation:
2
3
2
CO +H O+CO
3
2
2
2
2
3
DHfus (mixt) = SDHfusx
X
x
(3)
where, DHfus (mixt) (J g ) is the heat of fusion at 153°C
for the mixture, DHfusx and X denote the specific heat
It should be stressed that the reaction (4a) in the ex-
perimental conditions is not quantitative, i.e., although
–
1
x
3
an excess NaHCO or DCLH is present, mixtures of the
of fusion and the mole fraction of each component x
in the mixture. As shown in the relevant diagram, the
latent heat of fusion at 153°C of the mixture increases
with increasing the amount of DCLH. Discrepancies
between calculated and measured latent heats of fu-
sion were found, due to intrinsic difficulties of peak
integration (mixtures 0.8 and 0.9, Fig. 5) or dishomo-
geneity of the sample (mixture 0.1). For all other mole
fraction compositions there is a good agreement be-
tween experimental and calculated data.
solid products and reactants are formed. The overall
composition, therefore, progressively changes on heat-
ing due to the formation and decomposition phenomena.
To demonstrate this assumption, mixtures of DCLH and
3
NaHCO were treated in aqueous medium in order to
achieve a complete reaction. After water removal, the
DSC scan of the solid residue showed the pattern re-
ported in Fig. 9 (trace b) in comparison with the trace
obtained before the treatment with water (trace a). As
can be seen, the same DSC trace of pure anhydrous
DCLNa (Fig. 3) was obtained.
DCLH/NaHCO
3
mixtures
The DSC traces of 0.1–0.5 DCLH mole fraction
compositions show (Fig. 8) the NaHCO decomposi-
3
tion with two peaks in the 90 and 150°C range, the
melting of DCLH around 180°C and the DCLNa de-
The DSC scans of DCLH/NaHCO mixtures are re-
ported in Fig. 8. For all samples, an endothermic
event between 80 and 100°C, reasonably due to initial
3
thermal decomposition of NaHCO
dition, several endothermic effects, between 140 and
80°C, depending on the composition of the mixture,
are present. In particular, all mixtures within the
.1–0.5 DCLH mole fraction interval show an analo-
gous trend, characterised by peaks at 153, 180 and
85°C. The peak at higher temperature (285°C) can
be attributed to the decomposition of DCLNa formed
by reaction of DCLH with NaHCO or its decomposi-
tion product, Na CO . This is further confirmed by
the similar patterns shown by DCLH/DCLNa and
DCLH/NaHCO mixtures with compositions
0.6 DCLH mole fraction (Figs 5 and 8).
3
, is present. In ad-
2
0
2
3
2
3
3
³
3
Fig. 9 DSC trace of a – DCLH/NaHCO 0.5 mole fraction;
b – the same after complete reaction (see text)
9
06
J. Therm. Anal. Cal., 90, 2007

DICLOFENAC, DICLOFENAC SODIUM AND SODIUM BICARBONATE COMPOSITIONS
commonly be utilized to obtain information concern-
ing such interactions [8]. To overcome the intrinsic
difficulties when investigating pharmaceutical prod-
ucts, the simplification of the multi-component
systems as binaries is sometimes necessary.
In this work, a good agreement between liquidus
curves calculated by the Schroeder–Van Laar equa-
tion from fusion enthalpies and temperatures, and the
experimental results obtained by first heating runs
was found. For all binary compositions, an endother-
mic effect at 153°C, probably due to the eutectic fu-
sion, is present. The presence of a true eutectic be-
tween DCLH and DCLNa at 153°C, was further dem-
onstrated by drawing the latent heat of fusion diagram
at the same temperature.
Fig. 10 DSC traces of the mixture DCLH/NaHCO
a – 0.33 and b – 0.5 mole fraction composition
3
composition at 280°C. It was not possible to put into
evidence the eutectic fusion for two reasons: i) the
temperature very close to the decomposition tempera-
ture of NaHCO
3
, ii) the small amount of the eutectic
References
mixture. Within the 0.6–0.9 DCLH mole fraction
range, the two peaks at 150 and 180°C tend to over-
lap, and a broad peak between 170–180°C is evident.
A further final demonstration of the non-quanti-
1
2
3
4
A. Chiarini, A. Tartarici and A. Fini, Arch. Pharm.,
17 (1984) 268.
A. Fini, G. Fazio and I. Rapaport, Drug Exp. Clin. Res.,
9 (1993) 81.
K. M. O’Connor and O. I. Corrigan, Int. J. Pharm.,
26 (2001) 163.
3
1
tative reaction between DCLH and NaHCO
achieved by comparing the DSC trace of
DCLH/NaHCO mixture at 0.33 mole fraction com-
position with the DSC trace of the DCLNa/NaHCO
3
was
2
3
R. Bettini, G. Bertolini, E. Frigo, A. Rossi, I. Casini,
I. Pasquali and F. Giordano, J. Therm. Anal. Cal.,
77 (2004) 625.
3
at 0.5 mixture composition (Figs 10a and b, respec-
tively). A quantitative reaction between 1 mole of
5 A. K. Galvey, J. Therm. Anal. Cal., 82 (2005) 423.
6
F. Giordano, A. Rossi, I. Pasquali, R. Bettini, E. Frigo,
A Gazzaniga, M. E. Sangalli, V. Mileo and S. Catinella,
J. Therm. Anal. Cal., 73 (2003) 509.
DCLH and 2 mole of NaHCO
lead to the 1:1 DCLNa/NaHCO
Fig. 10b, where, however, it is still clearly visible the
melting peak of unreacted DCLH (around 180°C).
3
would theoretically
mixture reported in
3
7
8
J. T. Carstensen, Advanced Pharmaceutical Solids, Marcel
Dekker, New York 2001, p. 171.
K. Khimeche and A. Dahmani, J. Therm. Anal. Cal.,
84 (2006) 47.
Conclusions
Received: September 12, 2006
Accepted: November 17, 2006
OnlineFirst: February 26, 2007
Pharmaceutical formulations are multi-component
systems which are influenced by possible
physico-chemical interactions. The prediction and
comprehension of such interactions are therefore of
particular interest. Thermal analysis methods can
DOI: 10.1007/s10973-006-8182-1
J. Therm. Anal. Cal., 90, 2007
907