Thermal behaviour of mixed ligand Co(II), Ni(II) and Cu(II) complexes containing terephthalate ligands

Article abstract of DOI:10.1016/j.tca.2003.10.015

Thermal behaviour of a series of ternary Co(II), Ni(II) and Cu(II) complexes with terephthalate ion and some aromatic diamines has been studied by thermogravimetric and differential scanning calorimetric techniques. Dehydration processes of the complexes are very similar - most of them lose water in a single step. Thermal stability of the dehydrated products increases in the order Co

Full text of DOI:10.1016/j.tca.2003.10.015

Thermochimica Acta 413 (2004) 227–234
Thermal behaviour of mixed ligand Co(II), Ni(II) and Cu(II)
complexes containing terephthalate ligands
Jelena Rogan∗, Dejan Poleti
Department of General and Inorganic Chemistry, Faculty of Technology and Metallurgy, Karnegijeva 4, 11000 Belgrade, Yugoslavia
Received 22 June 2003; received in revised form 20 October 2003; accepted 20 October 2003
Abstract
Thermal behaviour of a series of ternary Co(II), Ni(II) and Cu(II) complexes with terephthalate ion and some aromatic diamines has
been studied by thermogravimetric and differential scanning calorimetric techniques. Dehydration processes of the complexes are very
similar—most of them lose water in a single step. Thermal stability of the dehydrated products increases in the order Co < Cu < Ni
for 2,2ꢀ-dipyridylamine and 2,2ꢀ-bipyridine containing complexes, but Cu < Co < Ni for 1,10-phenanthroline containing complexes. The
investigated compounds with 2,2ꢀ-bipyridine are always the most easily decomposed. The molar dehydration enthalpies are calculated and the
possible decomposition mechanisms are assumed. Relations between thermal behaviour of the complexes and their crystal structures, type of
central atom and aromatic diamine, as well as the presence of coordinated and/or lattice water molecules are discussed.
© 2003 Elsevier B.V. All rights reserved.
Keywords: Transition metal complexes; Terephthalate ion; Aromatic diamines; Enthalpy of dehydration; Thermal stability; Thermogravimetry; Calorimetry
1. Introduction
the Irving–Williams series [11], i.e. the stability order of the
complexes in aqueous solution. However, it seems that the
order Cu < Co < Ni predominate [12], but even an inverse
trend of thermal stability [13] has been observed.
The thermal behaviour of ternary complexes is much less
investigated than the behaviour of binary complexes. Here,
we report the thermal behaviour of nine ternary Co(II), Ni(II)
and Cu(II) complexes containing tpht ion and some aromatic
diamines: 2,2ꢀ-dipyridylamine (dipya), 1,10-phenanthroline
(phen) and 2,2ꢀ-bipyridine (bipy).
Four of investigated complexes were obtained as single
crystals and their crystal structures have been solved and
published elsewhere [2,5,6]. This fact could provide more
insight into the factors determining their thermal stability
and process of dehydration. In addition, an attempt to calcu-
late energy of M–OH2 and hydrogen bonds, and to compare
the obtained values is present.
To the best of our knowledge, only thermal properties of
several binary tpht complexes are described so far [14]. How-
ever, thermal properties of analogous ternary Co(II), Ni(II)
and Cu(II) complexes with the same aromatic diamines and
phthalate ion, pht (dianion of 1,2-benzenedicarboxylic acid)
are studied [15,16]. This allows the comparison of similar
complexes containing 1,2- or 1,4-benzenedicarboxylates.
The dianion of terephthalic (1,4-benzenedicarboxylic)
acid, H2tpht, is a potential bis-bidentate and bridging ligand.
Most studies concerning ternary transition metal complexes
with tpht ligands are concentrated on Cu(II) complexes and
their magnetic properties that are influenced by exchange
interaction through the bridging tpht ligands [1]. Although
some mononuclear tpht complexes are known [2], binu-
clear and polymeric complexes with tpht ion acting as a
bis-monodentate or bis-bidentate and bridging ligand are
much more encountered [3–6].
The review of Donia [7] gives the survey of thermal be-
haviour of a large number of transition metal complexes and
relationships between their structural properties and thermal
stability. It follows that factors affecting the thermal stabil-
ity of transition metal complexes in the solid state are nu-
merous and not well understood. There are some examples
in which the following order of the thermal stability was
established: Cu < Ni < Co [8–10]. This is just reverse of
∗
Corresponding author. Tel.: +381-11-3370477;
fax: +381-11-3370387.
E-mail address: rogan@elab.tmf.bg.ac.yu (J. Rogan).
0040-6031/$ – see front matter © 2003 Elsevier B.V. All rights reserved.
doi:10.1016/j.tca.2003.10.015

228
J. Rogan, D. Poleti / Thermochimica Acta 413 (2004) 227–234
2. Experimental
ands, complex 3 forms zigzag chains. The coordination of
the Cu(II) ions is very distorted octahedral and one lattice
H2O molecule per formula unit is present [5].
2.1. Preparation and characterisation of the complexes
The crystal structure of [Co(tpht)(phen)(H2O)]n, con-
taining the same ligands as 4, has been solved using room
temperature [5] and low temperature data [4]. Apparently,
significantly higher H2O content in our product (Table 1)
eliminates the possible isostructurality. The distinction
could be attributed to the different synthetic routes.
All complexes were prepared in a similar way using
the procedure which was described in the previous paper
[2]. Microcrystalline products were obtained from dilute
H2O/EtOH solutions containing M2+ ions (M = Co, Ni and
Cu), tpht and diamine ligands in a 1:1:1 molar ratio. They
are stable in air and insoluble in all common solvents. The
complexes were previously characterised by elemental anal-
ysis, magnetic measurements, IR and diffuse-reflectance
spectroscopy [2].
On the other hand, 6 is very similar to the above-mentioned
[Co(tpht)(phen)(H2O)]n and its formula should be written
as [Cu(tpht)(phen)(H2O)]n. The crystal structure determi-
nation [6] has shown that 6 consists of zigzag chains with
tpht acting as amphimonodentate bridging ligand, whereas
Cu(II) ions are in a deformed trigonal-bipyramidal envi-
ronment. However, the investigated sample showed slightly
higher H2O content (Table 1). The XRPD pattern of mi-
crocrystalline product confirmed its identity to the single
crystal data, but some broadening of the diffraction maxima
was observed. The discrepancy between formulae of micro-
crystalline and single crystal products can be explained by
taking into account that in the structure there is a small void
(V ≈ 25 Å3) at 0, 0.5, 0 site [6]. The void volume is smaller
than the expected volume of H2O molecule (ca. 40 Å3).
Thus, during the fast precipitation of the very insoluble
microcrystalline material such voids very probably capture
H2O molecules causing a slight distortion of the crystal
structure and lowering its crystallinity; contrary, the void is
empty in the case of slow growing single crystal. According
to these facts, excess of 0.3H2O molecules present in 6 is lo-
cated statistically in the whole structure, so the true formula
should be written as {[Cu(tpht)(phen)(H2O)]·0.3H2O}n.
The lattice H2O molecule form neither classical O–H · · · O
nor N–H · · · O hydrogen bonds. Only four C–H · · · O con-
tacts, with C · · · O distances in the range 3.09–3.17 Å and the
corresponding –H · · · O distances in the range 2.42–2.67 Å
are possible [6].
Thermogravimetry (30–700 ◦C range) was performed on
a Perkin-Elmer model TGS-2 thermobalance. Differential
scanning calorimetry (45–300 ◦C range) was performed on
a Perkin-Elmer model DSC-2 calibrated with indium as a
standard (ꢀfusH◦; the accuracy 1%). The heating rate was
10 ◦C min−1 using less than 10 mg sample mass. The fur-
nace atmosphere consisted of dry nitrogen at a flow rate
of 60 cm3 min−1. The TG, DTG and DSC curves of the
samples were used to determine the total content of wa-
ter of crystallisation, residual metal oxide and dehydration
enthalpy.
Some of the complexes were characterised by X-ray
powder diffraction (XRPD) analysis using Philips PW1710
diffractometer with a curved graphite monochromator and
Cu K␣ radiation (λ = 1.5418 Å).
3. Results and discussion
The general formula of the investigated complexes is
M(tpht)(diamine)·xH2O (M = Co(II), Ni(II) and Cu(II);
diamine = dipya, phen and bipy), with x = 1–5, while indi-
vidual empirical formulae are given in Table 1. In the case
of known crystal structure, the formulae showing coordina-
tion entity are listed, too. In the further text the complexes
will be denoted by numbers from Table 1.
In the average, Cu complexes contain less H2O than the
similar Co and Ni complexes. This was already observed
in the case of pht complexes [16] and can be attributed to
the more regular coordination polyhedra of Co and Ni com-
plexes, with 6 as an usual coordination number. Thus, co-
ordination polyhedra of Co and Ni are often supplementary
filled up with H2O molecules.
The further discussion will start with a short description
of the known structures, which will be later related to the
thermal properties of obtained microcrystalline products.
The complexes 1 and 2 are isostructural and mononuclear;
they consist of discrete complex units with central atoms in a
slightly distorted octahedral environment. Surprisingly, tpht
is coordinated by only one of its COO groups in a chelate
mode [2].
Crystal structures of other investigated complexes are
not known. However, XRPD patterns of 4, 5, 7 and 8
showed that the corresponding Co–Ni pairs (i.e. 4 and 5,
as well as 7 and 8) are isostructural, just as 1 and 2 [2].
As expected, the complexes with different diamines do not
have the same structures because of different bulkiness and
planarity/non-planarity of used diamines. Nevertheless, the
confirmed isostructurality, together with very similar physi-
cal and spectral properties [2], indicate that the coordination
mode of tpht ion is very probably identical in all Co and
Ni complexes, so they could be presented by the formula
[M(tpht)(diamine)(H2O)2]·2H2O similar to the 1 and 2.
Results of TG and DSC analyses: the total water con-
tent, the initial dehydration temperature (Tdeh,i), the fi-
nal dehydration temperature (Tdeh,f), the temperature of
DSC peak maximum (Tmax), molar dehydration enthalpy
(ꢀdehH◦m) and the initial decomposition temperature (Tdec,i
)
Due to the bridging role of tpht ions, which are asym-
metrically coordinated as bis-bidentate and bis-chelate lig-
are listed in Table 1. The TG and DSC curves are given in
Figs. 1 and 2.

Table 1
Results of TG and DSC analyses
No.
Empirical formula and formula showing
coordination entity
H2O content
(wt.%), [found
(calculated)]
Tdeh,i
(◦C)
Tdeh,f
(◦C)
Tmax
(◦C)
ꢀdehH◦m
(kJ mol−1
Tdec,i
(◦C)
Fragmentation
)
Up to
T (◦C)
Mass loss (wt.%) [found
(calculated)]
1
2
3
4
5
6
7
8
9
Co(tpht)(dipya)·5H2O, [Co(tpht)(dipya)(H2O)2]·3H2O
Ni(tpht)(dipya)·5H2O, [Ni(tpht)(dipya)(H2O)2]·3H2O
Cu(tpht)(dipya)·H2O, {[Cu(tpht)(dipya)]·H2O}n
Co(tpht)(phen)·4H2O
17.90 (18.60)
18.32 (18.61)
4.22 (4.32)
14.86 (15.16)
14.55 (15.17)
5.43 (5.43)
15.65 (15.97)
15.98 (15.98)
7.84 (7.79)
64
68
76
86
94
90, 144
100
80
108
121
105
145
152
127, 174
150
162
265
115
128
110
139
148
124, 178
154
177
128
248
243
49
248
218
263
300
279
316
331
277
254
281
265
392
–
75.6 (75.6) (5H2O + CO)
–
344
381
421
323
314
–
35.9 (34.0) (H2O + C5H5N + tpht)
80.4 (78.9) (4H2O + CO)
41.2 (41.0) (4H2O + CO + phen)
57.8 (56.5) (1.3H2O + tpht)
73.7 (74.3) (4H2O + CO)
–
Ni(tpht)(phen)·4H2O
Cu(tpht)(phen)·1.3H2O, {[Cu(tpht)(phen)(H2O)]·0.3H2O}n
Co(tpht)(bipy)·4H2O
16a, 61b (77)c
252
236
104
Ni(tpht)(bipy)·4H2O
Cu(tpht)(bipy)·1.8H2O
89
–
–
a
b
c
The molar dehydration enthalpy after removal of 0.3 mol H2O.
The molar dehydration enthalpy after removal of 1.0 mol H2O.
The overall molar dehydration enthalpy.

230
J. Rogan, D. Poleti / Thermochimica Acta 413 (2004) 227–234
3.1. Dehydration
The complexes 3, 4, 6, 8 and 9 can be easily obtained as
anhydrous compounds, while in 1, 2, 5 and 7 remains 0.1,
0.2, 0.2 and 0.1 mol of the total H2O content, respectively.
These remaining water molecules are eliminated in the sub-
sequent, almost horizontal step and further decomposition
usually begins before the compounds are completely dehy-
drated. Similar behaviour was already noticed during the
dehydration of some ternary pht complexes [15,16].
The dehydration processes for the most of complexes are
very similar—they lose water in a single step and it is not
possible to distinguish between coordinated and lattice H2O
molecules. The only exception is the complex 6, which de-
hydrates in two well separated steps. Its TG curve (Fig. 1)
clearly indicates presence of loosely bound lattice (0.3 mol)
and more strongly bound coordinated (1 mol) H2O molecule.
This is in accordance with the previous discussion and the
formula given in Table 1.
Sample 3 is monohydrate and the present H2O is not co-
ordinated [5]. Its final dehydration temperature is 105 ◦C
and it is evident (Table 1) that this lattice H2O molecule
can be removed up to about 100 ◦C. The loss of lattice H2O
molecules which occurs below 120 ◦C is also found for some
ternary Ni(II) complexes [17]. In other cases the final dehy-
dration temperatures are higher (Table 1), which is in agree-
ment with approved or assumed (see above) coexistence of
coordinated and lattice H2O molecules.
Fig. 1. TG curves obtained at heating rate of 10 ◦C min−1 in flowing N2
(the numbers correspond to the number of complex from Table 1).
For dipya and phen complexes the initial dehydration
temperatures, Tdeh,i, increase in the order Co < Ni < Cu,
and for bipy complexes in the order Ni < Cu < Co. The
observed sequence for phen complexes is analogous to the
previously described pht complexes [16]. It is also interest-
ing to compare Tdeh,i for pairs of isostructural Co and Ni
complexes: with one exception (for 7 and 8) all character-
istic temperatures are higher for Ni complexes.
Although in respect to phen and bipy, dipya is capable
to form an additional hydrogen bond, its complexes always
have the lowest Tdeh,i, Tdeh,f, and Tmax (Table 1). This means
that they easily and quickly lose H2O that can be attributed
to the nature of dipya, which is the only non-planar ligand,
used. According to this, the escape of H2O should be easier
through the channels in the structure when present diamine
ligand is not planar.
The observation that we were not able to distinguish
between coordinated and lattice H2O molecules can be ex-
plained by the fact that a single hydrogen bond is weaker
than coordinative M–OH2 bond, but every lattice H2O
molecule is usually involved in two, sometimes in three hy-
drogen bonds. The energy of two and three hydrogen bonds
could be higher than the energy of one coordinative bond.
In order to obtain more insight into dehydration processes,
DSC analysis of the complexes was performed up to 300 ◦C.
The expected endothermic effect for dehydration is observed
in the DSC curve of all complexes, and◦from this effect
the molar dehydration enthalpies (ꢀdehHm) are calculated
Fig. 2. DSC curves obtained at heating rate of 10 ◦C min−1 in flowing
N2 (the numbers correspond to the number of complex from Table 1).

J. Rogan, D. Poleti / Thermochimica Acta 413 (2004) 227–234
231
and together with DSC peak maxima (Tmax) summarised in
Table 1.
DSC curves for the most complexes compose of a unique
endothermic peak (Fig. 2). Only DSC curve for 6 has two
maxima (at 124 and 178 ◦C), because this complex, as shown
by TGA, is dehydrated in two steps.
to calculate the energy of M–OH2 bonds. Thus, about 68
and 63 kJ mol−1 is obtained for 1 and 2, respectively. Since
in both cases there are two coordinated H2O molecules, the
calculated values should be divided by 2, so the final mean
M–OH2 bond energies for 1 and 2 are 34 and 32 kJ mol−1
respectively.
,
DSC curves for all Co and Ni complexes (Fig. 2) are very
similar and their maxima are shifted to the higher tempera-
tures in◦the following order: dipya < phen < bipy (Table 1).
(217.5–251.9 kJ mol−1), and dipya complexes do not clearly
show the presence of an extra H2O molecule. The values
for Ni complexes are always lower than for Co complexes
with the same ligand. If different number of present H2O
molecules is taken into account, similar ꢀdehH◦m values
are already found for some ternary Co and Ni complexes
[18,19].
The maxima of DSC curves for Cu complexes are shifted
to the higher temperatures (Fig. 2) in the following order:
dipya < phen < bipy, but ꢀdehH◦m values are substantially
smaller than those for Co and Ni complexes (Table 1). This
is in agreement with smaller number of H2O molecules in
their formulae. Published ꢀdehH◦m values for some other
ternary Cu complexes [18] are similar to the values observed
for 3, 6 and 9 in this study.
Similar calculations for complexes 4, 5, 7 and 8 with eight
hydrogen and two M–OH2 bonds assumed resulted in mean
M–OH2 bond energies: 60 (for 4), 45 (for 5), 62 (for 7) and
54 kJ mol−1 (for 8).
ꢀdehHm for Co and Ni complexes are close to each other
Owing to always smaller ꢀdehH◦m values for Ni com-
plexes in comparison to Co complexes with the same
diamine ligand, the Ni–OH2 bonds are weaker than the
Co–OH2 bonds. This is just opposite to the trend of hydra-
tion energies of gaseous Co(II) and Ni(II) ions, lattice ener-
gies of corresponding compounds [20] and to the M–OH2
bond distances, where for isostructural complexes Ni–OH2
distances are shorter than Co–OH2 [2].
The mean Cu–OH2 bond energy in 6 and 9 can be cal-
culated, too. If 0.3 mol of lattice H2O molecule and the
well-separated thermal effect of its elimination in 6 are
neglected, this compound has explicitly one coordinated
H2O, which builds two additional hydrogen bonds. Based
on the above-calculated mean energy of hydrogen bond
(≈16 kJ mol−1) Cu–OH2 bond energy is ca. 29 kJ mol−1
.
The ratio of ꢀdehH◦m values for 3, 6 and 9 (49:77:104 ≈
1:1.6:2.1) only approximates the ratio (1:1.3:1.8) of H2O
molecules found (Table 1). From ꢀdehH◦m the bond ener-
gies of lattice H2O are 49 for 3 and 53 (=16/0.3) kJ mol−1
for 6, while the bond energy of coordinated H2O for 6
is 61 kJ mol−1. According to these values, it seems that 9
contains 1 mol of coordinated and 0.8 mol of lattice H2O.
Because of the polymeric nature of 3 [5] and 6 [6], the
probable structural formula for 9 is {[Cu(tpht)(bipy)(H2O)]·
0.8H2O}n.
With assumption that in 9 there are 0.8 lattice and 1 coor-
dinated H2O molecule and that both of them participate in
two hydrogen bonds, the calculated energy of Cu–OH2 bond
is about 46 kJ mol−1. Thus, within series of Co, Ni and Cu
complexes containing the same diamine ligands, Cu–OH2
bonds are always the weakest.
From the previous considerations and calculated values,
it can be concluded that the energy of two hydrogen bonds
is approximately equal to or slightly lower than the en-
ergy of one coordinative bond. This observation explains
the impossibility to distinguish coordinated and lattice water
molecules in described complexes.
3.2. Energies of coordinative M–OH2
and hydrogen bonds
Some limitations of the above approach must be em-
phasised. First, bond energies depend on bonds geometry,
which are different within one complex and from complex
to complex, too; these fine details are not known for all
samples. This could be a reason for dispersion (from 29 to
62 kJ mol−1) of calculated M–OH2 bond energies, which
also present mean values only. Second, three hydrogen bonds
in 3 are in the range 2.743–2.920 Å [5], i.e. they could be de-
scribed as “normal” to long. It is, therefore, possible that the
Determined enthalpies of dehydration and known crystal
structures of some investigated complexes enable to calcu-
late the mean energy (enthalpy) of hydrogen and M–OH2
bonds. Apparently, the first three complexes (Table 1) can be
used as a suitable starting set for this calculation. In 1 and 2
there are 11 hydrogen bonds per molecule [2], which means
that, besides the H2O molecule which is also bonded to the
amine N from dipya, every H2O molecule participates in 2
hydrogen bonds; in addition two M–OH2 bonds exist. The
complex 3 also contains dipya and three hydrogen bonds,
but H2O molecule is not coordinated [5].
calculated mean energy of hydrogen bonds (16 kJ mol−1
)
is slightly underestimated causing that the mean energy
of M–OH2 bonds is overestimated. Finally, we expect that
there is something like “resistance of crystal lattice” to the
water escaping. This is previously shown by disordering and
partial or total amorphisation of investigated samples after
dehydration [16]. During this process some bonds are bro-
ken, some other bonds could appear, but, generally speaking,
the energy of such system must be higher than the energy
of well-ordered crystal lattice. Thus, the calculated bond
ꢀdehH◦m value of 3 is 49 kJ mol−1. Since only three hy-
drogen bonds exist therein, the mean energy of hydrogen
bond (per 1 mol of H2O) is about 16 kJ. Because in 1 and 2,
as stated above, there are 11 hydrogen bonds, their overall
energy is approximately 180 kJ mol−1. By subtracting this
result from the experimental ꢀdehH◦m values it is possible

232
J. Rogan, D. Poleti / Thermochimica Acta 413 (2004) 227–234
energies should be described as “effective values” and they
take into account this and probably some other phenomena.
mentation very probably is not so simple or that decompo-
sition steps are strongly overlapped.
3.3. Thermal decomposition of anhydrous compounds
3.4. Thermal stability
Thermal properties of the complexes were observed from
room temperature to 700 ◦C and only complexes 2, 7 and
8 are totally decomposed in this range (Fig. 1). However,
while 2 and 7 are already decomposed up to 399 and 481 ◦C,
respectively, the residual mass of 8 slowly decreases above
451 ◦C and reaches the theoretical value about 700 ◦C. The
oxides of MO type were confirmed as residues (2: found
15.5%, calculated 15.4%; 7: found 17.8%, calculated 16.6%;
8: found 16.8%, calculated 16.6%). In other cases, the
residue does not reach constant mass up to 700 ◦C perhaps
forming mixtures of different oxides and/or carbides, which
are also found in the study of some ternary pht complexes
[15]. Under N2 atmosphere, the constant mass in some pht
complexes is not always reached even up to 850 ◦C [16].
After dehydration, all Co complexes (1, 4 and 7) decom-
pose via several steps (Fig. 1 and Table 1), also confirmed
by the presence of several DTG maxima. In all cases decom-
position process starts by decarboxylation and lost of CO.
Similar fragmentation is found during the thermal decom-
position of mixed Cu(II) complexes with pht ligands [15]
and zinc phthalate [21]. Further steps are more or less irre-
solvable. Still, up to 479 ◦C, removal of 4H2O + tpht could
be assumed for complex 4, but agreement between found
(53.8%) and calculated (50.3%) value is not so good.
The decomposition of Ni and Cu complexes is practically
a single step process, but only 2 is fully decomposed in this
step, while the other samples show a predominant step and
slow weight loss after that (Fig. 1).
The initial decomposition temperatures (Tdec,i) for dipya
and bipy complexes increase as Co < Cu < Ni, and for
phen complexes as Cu < Co < Ni (Table 1, Fig. 3). This
means that anhydrous Ni complexes are the most stable,
no matter what diamine ligand is present. It should also
be emphasised that Ni complexes are always more stable
It could be assumed that the residue mass in decom-
position of complex 3 corresponds to the removal of
H2O + C5H5N + tpht, while the residue mass for com-
plex 6 corresponds to the removal of 1.3H2O + tpht. Both
possibilities were found previously during the study of the
analogous mixed Cu complexes with dicarboxylate ions and
heterocyclic polyamines [15], as well as in the investigation
of thermal stability of some ternary pht complexes [16].
The differences were explained by the different ␲-acceptor
abilities of the aromatic polyamine ligands [15]. In the com-
plexes containing strong ␲-acceptors (such as phen), the
Cu–N bond should be stronger than the Cu–O bond and the
dicarboxylato ligand is removed first. If diamine ligands are
weak ␲-acceptors (like dipya), the order of elimination is
reversed. According to this, it can be assumed that 3 prob-
ably decomposes in two running steps that were not sepa-
rable. The first step corresponds to the removal of C5H5N
segment of dipya and subsequent to the removal of tpht.
However, complex 5 does not fit in this scheme (Table 1).
For complexes 8 and 9 removal of total H2O content +
C5H4N +tpht in the predominant step could be expected by
analogy. However, the calculated residual masses are 3–5%
higher than the observed values. This indicates that the frag-
Fig. 3. Relationships between the initial decomposition temperature and
(a) diamine ligand, (b) transition metal.

J. Rogan, D. Poleti / Thermochimica Acta 413 (2004) 227–234
233
(Fig. 3) than analogous and isostructural Co complexes. The
same order of the thermal stabilities is found for isostructural
nitrato Co, Ni and Cu complexes with dipya and bipy [7].
On the other hand, thermal stability for Cu complexes
increases in the order bipy < phen < dipya, while for Co
and Ni complexes the order is bipy < dipya < phen. In
the view of these results, bipy complexes are most unstable
(Table 1, Fig. 3), no matter what transition metal is present.
The inverse trend in thermal stability (phen < bipy) has
been found for some Ni(II) complexes [22].
The comparison of obtained results with data for similar
ternary complexes does not reveal analogous behaviour. For
example, thermal behaviour of isostructural [23] pht com-
plexes [M(pht)(phen)(H2O)3]·H2O, M = Co, Ni showed
that Co complex is more stable than Ni complex [16]. Here
described tpht complexes (4 and 5) with identical empiri-
cal formulae are generally more stable than pht complexes,
but 4 has lower Tdec,i than 5. Also, dipya containing pht
complexes: [M(pht)(dipya)] (M = Co or Cu) were found to
be significantly more stable than complexes with phen [16],
which is not the case for here investigated complexes. How-
ever, previously described dipya complexes do not contain
water of crystallisation and are very probably of polymeric
nature.
tical and very small (5.2◦) [24]. Consequently, the whole
molecule is close to the coplanar conformation, which
should be the most stable, at least in gas phase [25]. On the
other hand, these angles in tpht and similar aromatic poly-
carboxylate complexes vary in very wide range [1–6,23].
According to Kaduk and Golab [25], the energy barrier for
rotation around Caromatic − Ccarboxyl bond does not exceed
50 kJ mol−1, and for the rotation (torsion) angle of 20◦
it has the value of only about 9 kJ mol−1 for two twisted
groups. The observed deviations from planarity could be
ascribed to the significant improvement of crystal packing
[25]. However, our calculations show that the formation of
only one hydrogen bond should be more than enough to
stabilise high rotation angles of COO groups.
Finally, it is worth noticing that in all cases dehydration
enthalpies per mole of H2O are similar to or slightly higher
than the sum of enthalpy of fusion and enthalpy of evap-
oration for H2O (46.7 kJ mol−1). Thus, energy required to
remove water of crystallisation is close to the energy of its
sublimation. In other words, in the investigated complexes,
state of energy of coordinated and/or lattice H2O molecules
is similar to the energy of the pure H2O in solid state.
Acknowledgements
4. Conclusions
Support of this work by the Ministry of Science, Tech-
nologies and Development of the Republic of Serbia is grate-
fully acknowledged.
Some factors influencing the thermal stability of transition
metal complexes in the solid state and mechanism of their
decomposition are summarised in the review article of Donia
[7]. Here described results and comparison with some pre-
vious studies on similar compounds [9,10,16] showed that
there are some other factors, which should be added to the list
as probably very important. They are: (a) a decrease of the in-
termolecular forces in the solid state with increasing strength
of the intramolecular metal–ligand bonds, (b) ␲-bonding
ability of ligands (and possibly ␲-stacking interactions
between them), (c) bulkiness and planarity/non-planarity
of ligands, and (d) discrete or polymeric nature of the
complexes.
However, the factors are numerous and apparently com-
petitive, so they can be used as a predictive tool solely
within a series of similar complexes. Strongly speaking,
only the thermal behaviour of isostructural complexes could
be compared and the knowledge of crystal structure is very
desirable. If examined complexes contain water of crystalli-
sation, redistribution of bonds and changes of their geom-
etry are expected after dehydration. Since it is very hard to
characterise such, at least partially disordered intermediary
products, changes induced by dehydration are more or less
unpredictable. This introduces an additional uncertainty in
the prediction and explanation of the thermal stability of
complexes.
References
[1] M. Verdaguer, J. Gouteron, S. Jeannin, Y. Jeannin, O. Kahn, Inorg.
Chem. 23 (1984) 4291–4296;
E.G. Bakalbassis, J. Mrozinski, C.A. Tsipis, Inorg. Chem. 24 (1985)
3548–3553;
E.G. Bakalbassis, J. Mrozinski, C.A. Tsipis, Inorg. Chem. 25 (1986)
3684–3690;
Ph. Chaudhuri, K. Oder, K. Wieghardt, S. Gehring, W. Haase, B.
Nuber, J. Weiss, J. Am. Chem. Soc. 110 (1988) 3657–3658;
E.G. Bakalbassis, A.P. Bozopoulos, J. Mrozinski, P.J. Rentzeperis,
C.A. Tsipis, Inorg. Chem. 27 (1988) 529–532;
E. Bakalbassis, P. Bergerat, O. Kahn, S. Jeannin, Y. Jeannin, Y.
Dromzee, M. Gullot, Inorg. Chem. 31 (1992) 625–631;
E. Bakalbassis, C. Tsipis, A. Bozopoulos, W. Dreissig, H. Hartl, J.
Mrozinski, Inorg. Chim. Acta 186 (1991) 113–118;
C.E. Xanthopoulos, M.P. Sigalas, G.A. Katsoulos, C.A. Tsipis, A.
Terzis, M. Mentzafos, A. Hountas, Inorg. Chem. 32 (1993) 5433–
5436.
[2] J. Rogan, D. Poleti, Lj. Karanovic´, G. Bogdanovic´, A. Spasojevic´-de
Biré, D.M. Petrovi
c´, Polyhedron 19 (2000) 1415–1421.
[3] Zh.L. Deng, J. Shi, Z.H. Jiang, D.Zh. Liao, Sh.P. Yan, G.L. Wang,
Polyhedron 11 (1992) 885–887;
L. Li, D. Liao, Z. Jiang, Sh. Yan, Polyhedron 20 (2001) 681–684;
J.Ch. Kim, J. Cho, A.J. Lough, Inorg. Chim. Acta 317 (2001) 252–
258.
In the structure of H2tpht the dihedral angles between
plane of benzene ring and planes of COOH groups are iden-
[4] D. Sun, R. Cao, Y. Liang, Q. Shi, W. So, M. Hong, J. Chem. Soc.,
Dalton Trans. (2001) 2335–2340 and references cited therein.

234
J. Rogan, D. Poleti / Thermochimica Acta 413 (2004) 227–234
[5] Lj. Karanovic´, D. Poleti, J. Rogan, G.A. Bogdanovic´, A. Spaso-
S. Materazzi, G. D’Ascenzo, S. Aquili, K.M. Kadish, J.L. Bear,
Thermochim. Acta 397 (2003) 129–134;
S.T. Breviglieri, É.T.G. Cavalheiro, G.O. Chierice, Thermochim. Acta
356 (2000) 79–84.
jevic´-de Biré, Acta Crystallogr. C58 (2002) m275–m279.
[6] J. Rogan, D. Poleti, Lj. Karanovic´, Proceedings of the X Conference
of the Serbian Crystallographic Society, Abstracts, Serbian Crystal-
lographic Society, Belgrade, 2002, p. 23.
[7] A.M. Donia, Thermochim. Acta 320 (1998) 187–199.
[8] E. Cardarelli, G. D’Ascenzo, A.D. Magri, A. Pupella, Termochim.
Acta 33 (1979) 267–273;
[13] P.V. Khadikar, S.M. Ali, M.A. Farooqui, B.D. Heda, Thermochim.
Acta 91 (1985) 159–171.
[14] C.L. Sharma, T.K. De, A.K. Singh, Indian J. Chem. A 18 (1979)
183–184;
P.S. Bassi, B.S. Randhawa, H.S. Jamwal, Thermochim. Acta 86
(1985) 141–147;
E. Cardarelli, G. D’Ascenzo, A.D. Magri, A. Pupella, Termochim.
Acta 33 (1979) 267–273.
J.R. Allan, G.M. Baillie, J.G. Bonner, D.L. Gerrard, S. Hoey, Ther-
mochim. Acta 143 (1989) 283–288;
[15] R.P. Bonomo, G. Bruno, E. Rizzarelli, A. Seminara, G. Siracusa,
Thermochim. Acta 18 (1977) 207–215.
–
S. Materazzi, K. Kurdziel, U. Tentolini, A. Bacaloni, S. Aquili, Ther-
mochim. Acta 395 (2003) 133–137;
É.T.G. Cavalheiro, F.C.D. Lemos, J.Z. Schpector, E.R. Dockal, Ther-
mochim. Acta 370 (2001) 129–133.
[16] D. Poleti, D. Stojakovi
c´, Thermochim. Acta 205 (1992) 225–233.
[17] A.N. Nelwamondo, D.J. Eve, G.M. Watkins, M.E. Brown, Ther-
mochim. Acta 318 (1998) 165–175.
[18] J.R. Allan, G.H.W. Milburn, F. Richmond, A.S. Wilson, D.L. Gerrard,
J. Birnie, Thermochim. Acta 170 (1990) 147–154.
[19] J.R. Allan, B.R. Carson, P. Jaouen, Thermochim. Acta 170 (1990)
139–145.
[20] F.A. Cotton, G. Wilkinson, Advanced Inorganic Chemistry, 3rd ed.,
Interscience, New York, 1972, pp. 594–596.
[21] J. Sun, L. Yuan, K. Zhang, D. Wang, Thermochim. Acta 343 (2000)
105–109.
[9] R. Currini, Thermochim. Acta 80 (1984) 91–101, and references
cited therein.
[10] G. D’Ascenzo, U.B. Ceipidor, G. De Angelis, Thermochim. Acta 58
(1972) 175–182.
[11] J.E. Huheey, Inorganic Chemistry, 3rd ed., Harper, Cambridge, 1983,
pp. 312–318.
[12] J.A. Real, M.C. Muñoz, J. Borrás, Thermochim. Acta 101 (1986)
83–87;
[22] S. Mitra, Thermochim. Acta 205 (1992) 225–233.
–
J.R. Allan, B.R. Carson, D.L. Gerrard, S. Hoey, Thermochim. Acta
154 (1989) 315–322;
[23] D. Poleti, D.R. Stojakovi
c´, B.V. Prelesnik, Lj. Manojlovic´-Muir, Acta
Crystallogr. C 46 (1990) 399–402.
S. Materazzi, E. Vasca, Thermochim. Acta 373 (2001) 7–11;
[24] M. Bailey, C.J. Brown, Acta Crystallogr. 22 (1967) 387–391.
[25] J.A. Kaduk, J.T. Golab, Acta Crystallogr. B 55 (1999) 85–94, and
references cited therein.