A comparative study of the thermal reactivities of some transition metal oxalates in selected atmospheres

Article abstract of DOI:10.1016/j.tca.2004.08.021

The nonisothermal, solid-state thermal decompositions of the oxalates of six divalent transition metals, were investigated in alternative flowing atmospheres. Derivative thermogravimetry and differential scanning calorimetry response peak maxima, provided a measure of reaction temperatures. It was found that in inert and reducing atmospheres, the decomposition temperature increased with rise in enthalpy of formation of the divalent transition metal oxide. The results show that the trends of reactivity changes with salt compositions, offer a complementary approach to that provided by the study of single salts.

Full text of DOI:10.1016/j.tca.2004.08.021

Thermochimica Acta 429 (2005) 57–72
A comparative study of the thermal reactivities of some transition
metal oxalates in selected atmospheres
Mohamed A. Mohameda, Andrew K. Galweyb,∗, Samih A. Halawya
a
Department of Chemistry, Faculty of Science, South Valley University, Qena 83523, Egypt
b
5, Regents Wood, Malone Road, Belfast BT9 5RW, UK
Received 10 March 2004; received in revised form 5 August 2004; accepted 9 August 2004
Available online 30 December 2004
Abstract
A comparative investigation has been made of the nonisothermal, solid-state thermal decompositions of the oxalates of six divalent transition
metals (cations: manganese, iron, cobalt, nickel, copper and zinc) in alternative flowing atmospheres, inert (N2, CO2), reducing (H2) and
oxidizing (air). Derivative thermogravimetry (DTG) and differential scanning calorimetry (DSC) response peak maxima, providing a measure
of reaction temperatures, have been used to determine salt reactivities and thus to characterize the factors that control the relative stabilities
of this set of chemically related reactants. Two trends were identified. Trend (1): in the inert and reducing atmospheres, the decomposition
temperature (salt stability) increased with rise in enthalpy of formation of the divalent transition metal oxide, MO. It is concluded that the
rupture of the cation-oxygen (oxalate) bond is the parameter that determines the stability of salts within this set. Trend (2): the diminution of
decomposition temperatures from values for reactions in inert/reducing atmosphere to those for reactions in an oxidizing atmosphere increased
with the difference in formation enthalpy between MO and the other participating oxide (MO3/2 or MO1/2). The change of cation valence
tended to promote reaction, most decompositions in O2 occurred at lower temperatures, but the magnitude of the effect varied considerably
within this set of reactants. Observed variations in stoichiometric and kinetic characteristics with reaction conditions are discussed, together
with the mechanisms of thermal decompositions of these solid oxalates.
This approach to the elucidation of crystolysis reaction mechanisms emphasizes the value of comparative investigations within the group
of chemically related reactants. Previous isothermal kinetic studies had been made for each of the reactants selected here. From these, much
has been learned about the form of the (isothermal) solid-state yield–time curves, often interpreted to provide information about the geometry
of interface development for the individual rate processes. However, identification of the controls of reactivity, reaction initiation (nucleation)
and advance (nucleus growth), is much more difficult to characterize and less progress has been made towards elucidation of the interface
chemistry. The trends of reactivity changes with salt compositions, identified here, offer a complementary approach to that provided by the
study of single salts. Much of the recent literature on thermal decompositions of solids has been concerned with individual reactants, but
many results and conclusions are not presented in the widest possible perspective. Comparisons between systematically related reactants are
identified here as providing a chemical context for the elucidation of the chemical steps that participate in interface reactions. The article
advocates the use of a more chemical approach in investigations of crystolysis (solid-state chemical) reactions.
© 2004 Published by Elsevier B.V.
Keywords: Crystolysis reactions; Oxalate decompositions; Reaction kinetics; Reaction mechanisms; Solid-state chemistry; Solid-state reactions
1. Introduction
By focussing interest on the relative thermal reactivities
of a series of transition metal oxalates, this study is primarily
∗
concerned with the chemical properties, including the mech-
anisms, of a set of selected crystolysis reactions [1,2]. Many
Corresponding author.
E-mail address: andrew@galwey.freeserve.co.uk (A.K. Galwey).
0040-6031/$ – see front matter © 2004 Published by Elsevier B.V.
doi:10.1016/j.tca.2004.08.021

58
M.A. Mohamed et al. / Thermochimica Acta 429 (2005) 57–72
recent reports, concerned with solid-state thermal decompo-
sitions, have tended to be preoccupied with mathematical
aspects of kinetic analyses, in particular with the calculation
of activation energies, E, many examples are given in [3].
This trend appears to have been accompanied by a marked
diminution in interest in the chemical characteristics of such
rate processes and, somewhat surprisingly, also in the signif-
icances of those magnitudes of E which have been reported
for numerous reactions. We note that many studies have been
directed towards the development of novel calculation rou-
tines for the kinetic analysis of reaction rate measurements.
A purpose of this paper is to emphasize the importance of
the chemical processes involved in a group of such reactions.
The present work was undertaken to characterize the reactiv-
ities and controls applicable within a set of selected related
rate processes, involving relatively simple reactions of known
stoichiometry [4–6].
Thermal decompositions of solids, crystolysis reactions,
are chemical changes and, in principle, behaviour patterns
should be capable of being reconciled with the character-
istic chemical properties of the reactant constituents. This
aspiration has not, however, always been achieved [2,3,6].
Transition metal oxalates were selected for this compara-
tive study because the reactions are relatively simple, the
reactants are readily prepared and many have already been
studied [2,4,6]. The extension of a recently published study
[5] of the relative abilities of various metal oxides, con-
taining the same cations as the oxalates studied here, to
promote the decomposition of potassium oxalate, provided
one motivation for the present program. A second rea-
son was that alternative controlling processes and chemical
steps have been proposed by different researchers to pro-
vide mechanistic explanations for metal oxalate breakdown
reactions [4,6]. The present report is intended to contribute
towards resolution of some of these outstanding problems,
by understanding the chemistry of these reactions in greater
detail.
of each one of the three dominant bonds in the carboxylate
structure, R CO O M. Clearly, further work is required to
resolve these inconsistencies.
To contribute towards the resolution of these problems,
we report here comparative studies of the thermal decomposi-
tions of the transition metal oxalate series, manganese to zinc,
in different atmospheres: N2 (inert), CO2 (inert/product), H2
(reducing) and air (oxidizing). An early study [7], concerned
with four of these salts (Fe, Co, Ni and Cu), concluded that
oxide was the initial product and this was reduced to the
metal by carbon monoxide. Dollimore and Griffiths [4] (see
also [8]) reported DTA studies of the thermal reactions of
25 metal oxalates in N2 and in O2. For the divalent ox-
alates of Mn, Fe, Co and Ni, Macklen [9] showed that the
decomposition temperatures (DTA) in N2 were in the se-
quence of decreasing cation electronegativities. The temper-
atures of these reactions in air, an oxidizing atmosphere,
diminished in the sequence of ease of cation oxidation,
M2+ → M3+.
Boldyrev et al. [10] have concluded that the rupture of
the C C link within the oxalate anion is the first step in the
thermal breakdown of all oxalates. This bond is regarded as
the weakest, supported by observations in [11]. The same
model has recently [12] been used to introduce a particularly
comprehensive discussion of the thermal decomposition of
silver oxalate.
Acheson and Galwey [13] and later Galwey and Mohamed
[14] identified the energy barrier to the vacuum thermal de-
compositions of oxalates as the strength of the cation to
oxalate–oxygen bond. For the Ni, Ag, Hg, Mn and Pb ox-
alates, L’vov concluded [15] that the first step in reactant
breakdown is a dissociative evaporation step, with simulta-
neous condensation of the products of low volatility. The
primary volatile products from decompositions of the ox-
alates containing the first three cations mentioned were iden-
1
tified as CO + CO2 + O2 (or O) and for Mn and Pb were
2
1
2CO + O2.
2
Although there have been numerous studies of the solid-
state decompositions of diverse metal carboxylates [2,4,6],
littleisknownforcertainaboutthefactorsthatdeterminetheir
relative and absolute stabilities. The controlling bond rupture
step(s) in oxalate anion breakdowns have not been reliably
identified. However, thermodynamic considerations have en-
abled the principal products to be predicted. CO and/or CO2
are the usual gaseous products and the residue may be the
metal, oxide or carbonate, depending on the electronegativ-
ity of the constituent cation. It is also shown in the present
study, for widely used reaction conditions, that mixtures of
solid products may be given, including cation valencies that
differ from those in the original reactant. Different kinetic
studies of the same reactant have lead to the identification
of different mechanistic conclusions and different stability
controls, including alternative possible ‘rate limiting’ pro-
cess, some of which are mentioned below. As explained in
[6], various literature reports have concluded that the initial
step in anion breakdown may be the rupture or rearrangement
2. Experimental
2.1. Materials
All chemicals used throughout the present study were of
Analytical Grade: sodium oxalate and oxalic acid (BDH,
England); CoCl2·6H2O and MnCl2·4H2O (Koch-Light Lab-
oratories, England); CuCO3·Cu(OH)2 (Prolabo, France);
ZnCO3 (Hopkin & Williams, England) and NiCl2·6H2O
(May & Baker, England).
2.2. Preparations of the reactant metal oxalates
2.2.1. Manganese(II) oxalate
9.9 g MnCl2·4H2O (0.05 M) dissolved in 250 ml distilled
water was slowly added with continuous stirring to 6.7 g

M.A. Mohamed et al. / Thermochimica Acta 429 (2005) 57–72
59
sodium oxalate (0.05 M) dissolved in 250 ml distilled wa-
ter. The reaction mixture was maintained at 60 ◦C for 2 h.
The precipitated product was filtered, washed several times
with distilled water and finally dried at 100 ◦C.
Thermal analysis experiments were performed in a dy-
namic atmosphere (flow rate 40 ml/min) of dry N2, H2,
CO2 or air. Variations of sample mass and particle sizes are
known [17] to have significant effects on the thermal analy-
sis responses (peak shape and maximum peak temperature).
Therefore, constant masses (10–15 mg) of gently crushed
reactant samples were always used. Each thermal analysis
experiment was repeated three times and all mass loss val-
ues reported were the average of three measurements: re-
producibility was better than 1%. Peak temperatures from
DTA and DSC response curves were similarly recorded and
were reproducible within 1 ◦C. Highly sintered ␣-Al2O3
powder (Shimadzu) was used as the reference material in
DTA and DSC experiments. Calibrations of the DTA and
the DSC instruments for temperature and enthalpy mea-
surements at heating rates between 2 and 30 ◦C/min were
carried out using Specpure indium and zinc metals, John-
son Matthey: melting points 156.6 and 419.5 ◦C, ꢀH, 28.24
and 101.50 Jg−1, respectively [18]. All gases used in the
present work (i.e., N2, H2, CO2 and air) were 99.0% pure
products from the Egyptian Company of Industrial Gases
(Nasr City, Egypt).
2.2.2. Ferrous oxalate
The Prolabo product, the dihydrate, was used as received
without further purification.
2.2.3. Cobalt oxalate
5.95 g CoCl2·6H2O (0.025 M) dissolved in 250 ml dis-
tilled water was added with continuous stirring to 3.35 g
of sodium oxalate (0.025 M). The procedures were identical
with those used in the preparation of manganese oxalate.
2.2.4. Nickel oxalate
The preparation of nickel oxalate was similar to that for the
cobalt and manganese salts, through the addition of an aque-
ous solution of 5.95 g NiCl2·6H2O in 250 ml distilled water
to an aqueous solution of 3.35 g sodium oxalate in 250 ml
distilled water.
2.2.5. Copper(II) oxalate
2.4.2. X-ray diffraction
10.5 g basic copper carbonate (approximate composition,
CuCO3·Cu(OH)2) was added in small amounts with contin-
uous stirring to 12.6 g oxalic acid in 250 ml distilled water at
60 ◦C[16]. Theblueproductprecipitatedwasfiltered, washed
several times with distilled water and dried at 100 ◦C.
X-ray powder diffraction analysis of the metal oxalates
and of the decomposition products from reactions in each of
the atmospheres, and at different temperatures, used a model
D5000 Siemens diffractometer (Germany) with Ni-filtered
˚
Cu K␣ radiation (␭ = 1.5406 A). An on-line data acquisition
and handling system permitted an automatic JCPDS library
search [19] and match (Diffrac AT software, Siemens) for
phase identification.
2.2.6. Zinc oxalate
Zinc oxalate was prepared by the slow addition of 6.27 g
(0.05 M) ZnCO3, with continuous stirring, to 6.3 g (0.05 M)
oxalic acid dissolved in 250 ml distilled water maintained at
60 ◦C. Stirring was continued for a further 2 h. The white
precipitate was filtered, washed several times with distilled
water and dried at 100 ◦C.
2.4.3. IRspectroscopy
Infrared spectra of the reactants, samples partially de-
composed at selected temperatures (under different dynamic
atmospheres) and the residual solid products were car-
ried out adopting the KBr disk technique using a Magna-
FTIR 560 (USA) instrument operated by Nicolet Omnic
software.
2.3. Elemental analysis
The carbon and hydrogen contents of all reactants were de-
termined using standard combustion techniques. The metal
content (i.e., Mn, Fe, Co, Ni, Cu and Zn) of each salt was de-
termined using the inductively coupled plasma method (ICP)
with a model OES ARL instrument, Switzerland.
2.4.4. Enthalpy determinations
Enthalpy (ꢀH) measurements for the decomposition pro-
cesses (and dehydrations of the hydrated salts) were obtained
from the DSC response traces at different heating rates and
in the different flowing gases.
2.4. Experimental techniques
2.4.1. Thermal analysis
3. Results and discussion
Thermal analysis measurements (thermogravimetry
(TGA), derivative thermogravimetry (DTG), differential
thermal analysis (DTA) and differential scanning calorimetry
(DSC)) were carried out by a Shimadzu “Stand Alone”
thermal analyzer (TGA-50H, DTA-50 and DSC-50, Japan).
This was equipped with a data acquisition and handling
system (model TA-50WSI, Japan).
3.1. Elemental analysis
Elemental analyses of the reactants showed that the man-
ganese, copper and zinc oxalates were prepared in the an-
hydrous form and the other three salts, ferrous, cobalt and
nickel oxalates, were the dihydrates. Analytical data are

60
M.A. Mohamed et al. / Thermochimica Acta 429 (2005) 57–72
recorded below with the theoretically calculated values in parenthesis:
MnC2O4
Mn, 37.95 (38.44%) and C, 16.33 (16.79%).
FeC2O4·2H2O
CoC2O4·2H2O
NiC2O4·2H2O
CuC2O4
Fe, 30.71 (31.05%); C, 12.98 (13.34%) and H, 2.41 (2.22%).
Co, 31.51 (32.39%); C, 12.87 (13.19%) and H, 2.01 (2.20%).
Ni, 31.74 (32.13%); C, 13.40 (13.13%) and H, 1.92 (2.19%).
Cu, 41.38 (41.93%) and C, 15.42 (15.84%).
ZnC2O4
Zn, 42.11 (42.62%) and C, 15.41 (15.65%).
X-ray diffraction observations for the reactants provided
confirmation of these elemental analyses. The diffraction
patterns obtained gave good matches with the correspond-
ing standard JCPDS data references [19]: MnC2O4 (32-
0646); FeC2O4·2H2O (22-0635); CoC2O4·2H2O (25-0250);
NiC2O4·2H2O (14-0742); CuC2O4 (21-0297) and ZnC2O4
(37-0718).
3.2.1.3. In hydrogen. On heating at 5 ◦C min−1, a large
DTG peak maximum occurred at 397 ◦C (350–420 ◦C) (at
10 ◦C min−1 values were 409 ◦C (365–438 ◦C)). A small ini-
tial response was detected at 341 (and 350 ◦C). Mass loss
values, 50.9% (and 51.0%) were close to, or slightly greater
than, expectation for product MnO.
3.2.1.4. In air. On heating at 5 ◦C min−1, decomposition
was completed in a single rate process at relatively lower tem-
3.2. Thermal analysis
perature, DTG peak at 290 ◦C (275–325 ◦C). At 10 ◦C min−1
,
the response, 299 ◦C, was completed within a much dimin-
ished temperature interval, 287–307 ◦C. Mass losses, 44.1
and 46.4% on heating at 5 and 10 ◦C min−1, respectively,
were in accordance with expectation for the formation of
product mixtures containing both Mn3O4 and Mn2O3.
Quantitative measurements for each reactant studied are
individually presented in Fig. 1a–f, TGA and Fig. 2a–f, DSC
response traces for decompositions of the six reactants in the
four atmospheres used, N2, CO2, H2 and air. The significant
results obtained are tabulated (Table 1), total mass losses and
DTG peak maxima (Table 2), DSC peak maxima (Table 3),
decomposition enthalpies, ꢀH and (Table 4), XRD identifi-
cation of the residual solid products [19].
3.2.1.5. DTA measurements. The reaction in air (oxygen)
wasexothermic, expectediftheinitialCOgasandtheresidual
MnOproductswerebothsubsequentlyoxidized. Thereaction
in nitrogen was endothermic, consistent with formation of the
lower oxidation states of the same products.
3.2.1. Manganese oxalate
ObservationsandresultsaresummarizedinFigs. 1aand2a
and Tables 1–4.
3.2.1.6. DSC measurements. Manganese oxalate decompo-
sitions in N2 (414 ◦C), CO2 (425 ◦C) and H2 (407 ◦C)
atmospheres were endothermic with DSC peak tempera-
ture maxima (listed in Table 2) close to the corresponding
DTG values. In contrast, the reaction in air (321 ◦C) was
exothermic.
3.2.1.1. In nitrogen. On heating at 5 ◦C min−1, a large broad
(315–415 ◦C) DTG composite peak was detected, including
two maxima at 337 and 389 ◦C, extensive overlap resulted
in no perceptible minimum. The total mass loss, 47.4%, cor-
responded with expectation for a mixture of products, MnO
and Mn2O3. On heating at 10 ◦C, the first peak was almost
eliminated, appearing only as an early distortion in the single
dominant response at 418 ◦C (360–450 ◦C). Here the mass
loss, 50.2%, was in accordance with the formation of MnO.
After completion of reaction there was a small rise in product
mass, attributable to oxidation by trace amounts of oxygen
in the flowing nitrogen.
3.2.1.7. Discussion. It has been shown [20] (see also
[21,22]) that manganese oxalate decomposition in oxygen
proceeds in a lower temperature interval (227–317 ◦C) than
the vacuum reaction (343–363 ◦C). This was ascribed to the
active role of the Mn3+ ion in promoting oxalate anion break-
down. Evidence was provided [20] that, in vacuum, a small
excess of CO was formed in the early stages of reaction, be-
lieved to be accompanied by the formation of an appropriate
amount of Mn3+. In oxygen, cation (and CO) oxidation is
expected.
The above results confirm the significant role of oxygen in
promoting this reaction in air at 290–299 ◦C, attributable to
participation by the oxidized cation, Mn3+, and at the higher
heating rate completion was particularly rapid. These resid-
ual products were (relatively) the most highly oxidized. In
3.2.1.2. In carbon dioxide. On heating at 10 ◦C min−1, a
small DTG response at about 356 ◦C overlapped extensively
with the dominant peak at 425 ◦C (352–453 ◦C). The mass
loss, 50.4% was in accordance with expectation for the for-
mation of MnO. At 5 ◦C min−1 the mass loss was 49.5% and
DTG peaks were at 350 and 408 ◦C.

M.A. Mohamed et al. / Thermochimica Acta 429 (2005) 57–72
61
Table 1
Effects of different gas atmospheres on the total mass loss (ML% on completion of reaction) and Tmax (from the DTG curves) values for decompositions of the
different metal oxalates at a heating rate of 10 ◦C/min
Gas atmosphere (40 ml/min)
N2
Mn(II) oxalate
Fe(II) oxalate
Co(II) oxalate
Ni(II) oxalate
Cu(II) oxalate
Zn(II) oxalate
ML%
Tmax
50.2%
418 ◦C
55.4%
201.6
264
351 ◦C
55.4%
191.5
316.9 ◦C
59.4%
229.1
367.6 ◦C
56.9%
299.8 ◦C
48.0%
411.8 ◦C
CO2
ML%
Tmax
50.4%
425 ◦C
56.6%
204.7
284
384
411.4 ◦C
55.5%
190.4
340.1 ◦C
60.0%
236.7
375
392 ◦C
56.0%
309.6 ◦C
48.0%
421.6 ◦C
H2
ML%
Tmax
51.0%
409 ◦C
69.3%
208.8
258.7
407.3 ◦C
68.0%
190.4
359.5
370.7 ◦C
68.1%
258.7
348.3 ◦C
59.6%
283.5 ◦C
48.1%
405.3 ◦C
Air
ML%
Tmax
46.4%
299 ◦C
55.8%
204.5
231.4 ◦C
56.0%
199.2
303.1 ◦C
60.3%
270.9
340.1 ◦C
54.6%
307.1 ◦C
48.2%
406.9 ◦C
ML%: percentage mass loss on completion of reaction. Tmax: DTG peak maxima (◦C).
Table 2
DSC peak temperature maxima (◦C) for the dehydration and the decomposition of the metal oxalates on heating at 10 ◦C/min in different gas atmospheres,
atmosphere flow rate 40 ml/min
Metal oxalate
Anhyd.
Process
Tmax in N2 atmosphere
Tmax in CO2 atmosphere
Tmax in H2 atmosphere
Tmax in air atmosphere
Dehyd.
–
–
–
–
MnC2O4
Decompn.
414 (n)
425 (n)
407 (n)
321(x)
FeC2O4·2H2O
Dehyd.
Decompn
183 (n)
392 (n), 412 (n)
186 (n)
388 (n), 416 (n)
179 (n)
381 (n), 422 (n)
200 (n)
244 (x)
CoC2O4·2H2O
NiC2O4·2H2O
Dehyd.
Decompn
185 (n)
387 (n)
181 (n)
404 (n)
173 (n)
375 (x)
179 (n)
320 (x)
Dehyd.
Decompn
223 (n)
340 (n), 357 (n)
230 (n)
375 (n)
214 (n)
351 (x)
227 (n)
353 (x)
Anhyd.
CuC2O4
Anhyd.
ZnC2O4
Dehyd.
–
–
–
–
Decompn
Dehyd.
296 (x)
–
299 (x)
–
304 (x)
–
318 (x)
–
Decompn
407 (n)
419 (n)
405 (n)
390 (n)
DSC response maxima: (n) endothermic, (x) exothermic.
Table 3
Enthalpy (ꢀH) changes on dehydration and on decomposition of some metal oxalates during heating at 10 ◦C/min in different gas atmospheres flowing at
40 ml/min
Metal oxalate
Atmosphere (units)
Air (kJ/mol)
H2 (kJ/mol)
CO2 (kJ/mol)
N2 (kJ/mol)
Anhyd. MnC2O4
Dehydn.
Decomp.
Dehydn.
Decomp.
Dehydn.
Decomp.
Dehydn.
Decomp.
Dehydn.
Decomp.
Dehydn.
Decomp.
–
111
61.2
79.2
71.0
78.7
73.1
70.4
–
–
30.4
35
41.6
32.1
10.9
20.0
33.6
–
–
–
78.8
96.2
60.5
84
39.7
82.3
29.3
–
88.0
94.1
70.2
83
40.8
87
32.5
–
11.9
–
FeC2O4·2H2O
CoC2O4·2H2O
NiC2O4·2H2O
Anhyd. CuC2O4
Anhyd. ZnC2O4
43.5
–
16.6
21.1
–
25.6
18.0
–
62.0
72.1
Trends of ꢀHdecomp
: In nitrogen and in carbon dioxide: Mn > Zn > Fe > Co > Ni > Cu; in hydrogen: Fe > Ni > Mn > Zn > Cu > Co; in air:
Mn > Fe > Co > Ni > Cu > Zn.

62
M.A. Mohamed et al. / Thermochimica Acta 429 (2005) 57–72
heating rate, 3 ◦C min−1. (ii) In O2, the occurrence of smaller
DTG peaks at lower temperatures than the dominant reaction
in N2, CO2 and H2 may be associated with facilitation of the
precursor reaction step required to generate the Mn3+ that
is an essential participant in manganese oxalate decomposi-
tion. This is confirmed in Fig. 3, where the decomposition
temperature of the MnC2O4/Mn3O4 mixture was diminished
by about 55 ◦C, compared with that of the pure salt. (iii) There
is strong evidence that a small amount of impurity gas in the
reaction atmosphere can exert a perceptible influence on the
kinetic characteristics of the overall reaction, accounting for
the low temperature peaks during reactions in N2. This pos-
sibility must be remembered in designing rate studies where
the small amount of promoter (O2 yielding Mn3+) can have a
large kinetic consequence. (iv) The mass gain after comple-
tion of decomposition in air is ascribed to oxidation of the
lower oxides initially formed.
hydrogen the peak maxima appeared at significantly higher
temperatures, 397–409 ◦C and the residual oxide was MnO.
These conclusions are entirely consistent with the previous
work [20]. The formation of MnO from reactions in N2 and
H2 is confirmed by X-ray diffraction (Table 4). During the
reaction in N2 at 10 ◦C min−1 the initial response is small
(about 10% reaction) whereas on heating at 5 ◦C min−1 this
peak represents about 50% reaction. Product mass increases
after reactions in N2 and CO2 are consistent with some MnO
oxidation by traces of O2 in these flowing gases.
The following conclusions are reached. (i) The pattern of
behaviour is entirely consistent with the previous proposal
[20] that Mn3+ ions participate in controlling the rate of ox-
alate ion breakdown. This accounts for the relatively much
more rapid reaction in air (oxygen), which diminished the
reaction temperature by 110 10 ◦C. This is somewhat less
than the difference found by Macklen [9], 150 ◦C, at the lower
Fig. 1. Thermogravimetric plots of mass losses (%) against temperature (◦C) for thermal decompositions, in a flowing, 40 ml/min, atmosphere of N2, CO2, H2
or air, on heating at 10 ◦C min−1 of: (a) manganese oxalate, (b) ferrous oxalate dihydrate, (c) cobalt oxalate dihydrate, (d) nickel oxalate dihydrate, (e) copper
oxalate and (f) zinc oxalate. Atmospheres: N2, continuous line; H2, short dashes; CO2, long dashes; air, alternate long and short dashes.

M.A. Mohamed et al. / Thermochimica Acta 429 (2005) 57–72
63
Fig. 2. Differential scanning calorimetric plots of enthalpy output/input (mW) against temperature (◦C) for thermal decompositions, in a flowing, 40 ml/min,
atmosphere of N2, CO2, H2 or air, on heating at 10 ◦C min−1 of: (a) manganese oxalate, (b) ferrous oxalate dihydrate, (c) cobalt oxalate dihydrate, (d) nickel
oxalate dihydrate, (e) copper oxalate and (f) zinc oxalate. Atmospheres represented as in Fig. 1.
3.2.2. Ferrous oxalate dihydrate
Observations and results are summarized in Figs. 1b
and 2b and Tables 1–4.
DTG response peaks, maxima at 264 and 351 ◦C with an in-
distinct minimum at 315 ◦C. At the lower rate of heating,
5 ◦C min−1, a closely similar broad response included max-
ima at 262 and 334 ◦C. The total mass loss (including dehy-
dration) was 55.4%, close to expectation for the formation of
product Fe2O3, 55.6%.
3.2.2.1. Dehydration. The first large TG response was as-
cribed to the precursor dehydration reaction: maxima were
between 197 and 207 ◦C for samples heated at 10 ◦C min−1
(peaks were between 170 and 230 ◦C) and close to 185 ◦C
for 5 ◦C min−1 (160–210 ◦C). Mass losses were 19.9 0.2%
which agrees well with expectation (20.0%) for the loss of
2H2O. It is concluded that dehydration was (largely) com-
pleted before the onset of anion breakdown, for reactions in
all gases except air, where the contribution from dehydration
overlapped with the onset of decomposition and its comple-
tion could not be distinguished on the DTG response traces.
3.2.2.3. In carbon dioxide. Again, on heating at
10 ◦C min−1, two broad overlapping DTG peaks were
found between 240 and 445 ◦C, but the first, 284 ◦C,
was relatively smaller (less than half) the dominant peak,
maximum 384 ◦C. The mass loss, 56.6% was slightly greater
than the reaction in N2 and is consistent with the formation
of a residual mixture containing approximately equal masses
of Fe2O3 and Fe3O4.
3.2.2.2. In nitrogen. On heating in N2 at 10 ◦C min−1, this
3.2.2.4. In hydrogen. On heating at 5 ◦C min−1, two DTG
reactant showed two broad and overlapping (235–415 ◦C)
peaks were distinguished, maxima close to 255 and 401 ◦C

64
M.A. Mohamed et al. / Thermochimica Acta 429 (2005) 57–72
Table 4
XRD results for identifications of the final solid decomposition products on heating metal oxalates in different gas atmospheres (reaction temperatures shown)
Metal oxalate
MnC2O4
In N2 atmosphere
In H2 atmosphere
In air atmosphere
At 460 ◦C
At 450 ◦C
At 320 ◦C
MnO (maj), card 7-0230
Mn2O3 (min), card 24-0508
MnO (maj)
MnO2 (min), card 12-0141
Mn3O4 (maj)
Mn5O8 (min)
MnO2 (min)
FeC2O4·2H2O
At 440 ◦C
Fe2O3 (maj), card 15-0615
Fe3O4 (min), card 11-0614
At 480 ◦C
Fe (maj), card 6-0696
Fe3O4 (min)
At 260 ◦C
Fe2O3 (maj)
Fe3O4 (min)
Fe (min)
Fe2O3 (min)
CoC2O4·2H2O
NiC2O4·2H2O
CuC2O4
At 380 ◦C
Co3O4 (maj), card 43-1003
CoO (min), card 9-0402
At 400 ◦C
Co (maj), card 15-0806
Co3O4 (min)
At 320 ◦C
Co3O4 (maj)
Co2O3 (min)
At 400 ◦C
Ni (maj), card 4-0850
NiO (maj), card 22-1189
At 360 ◦C
Ni (only)
At 360 ◦C
NiO (only)
At 350 ◦C
Cu (maj), card 4-0836
CuO (min), card 41-0254
Cu2O (min), card 5-0667
At 350 ◦C
Cu (only)
At 350 ◦C
Cu (maj)
CuO (maj)
Cu2O (maj)
ZnC2O4
At 460 ◦C
At 460 ◦C
At 460 ◦C
ZnO (only), card 36-1451
ZnO (only)
ZnO (only)
maj = major component of residue; min = minor. ‘Cards’ mentioned refer to [19].
(response ratios about 1:4) for both heating rates, the mini-
mum was close to 300 ◦C. The first peak represented 10.7%
mass loss and the total on completion of decomposition
on heating at 10 ◦C min−1 was 69.3%, close to expectation
(68.9%) for the formation of residual Fe metal.
maximum, 220–233 ◦C, followed by rapid completion of re-
action. Thetotalmasschange, 55.8%wasclosetoexpectation
for Fe2O3 product, 55.6%.
3.2.2.6. Discussion. The first response peaks in Table 1,
maxima 202–209 ◦C, are identified as dehydration (−2H2O).
The predominant mass losses due to decomposition, Fig. 1b,
are associated with the peak maxima at 411, 407 and 351 ◦C
for reactions in CO2, H2 and N2, respectively. These occur
3.2.2.5. In air. Decomposition overlapped significantly with
the precursor dehydration step on heating at both 5 and
10 ◦C min−1. After onset of the former there was a sharp
Fig. 3. The decomposition of a manganese oxalate/Mn3O4 mixture on heating in N2 at 10 ◦C min−1, a plot of mass loss (%) against temperature (◦C). The
presence of Mn3O4 diminishes reaction temperature by about 55 ◦C.

M.A. Mohamed et al. / Thermochimica Acta 429 (2005) 57–72
65
Fig. 4. Plot of mass loss (%) against temperature (◦C) for the thermal decomposition of cobalt oxalate dihydrate on heating in N2 at various rates 2–30 ◦C min−1
Both stoichiometry and reaction rate change with conditions, for discussion see text.
.
(as expected) at a temperature somewhat above that of the
isothermal reaction. The lower temperature of onset of de-
composition, Fig. 1b, the peaks at 384, 259 and 264 ◦C in
CO2, H2 and N2, is not consistent with the earlier study [14].
However, the most probable explanation is that a proportion
of the Fe2+ becomes oxidized during the last stages of de-
hydration either by disproportionation in water vapour or by
O2 traces in the atmosphere, except for the reactions in H2.
Promotion of oxalate breakdown by Fe3+ accounts for the
lower temperature reaction in N2. The decomposition of the
ferric salt, response maximum at 231 ◦C, Table 1, is rapid at
these reaction temperatures [23,24].
A previous study [14] of the isothermal decomposition of
ferrous oxalate in vacuum identified the reaction as occur-
ring between 323 and 365 ◦C. The residual products were the
mixture (Fe + Fe3O4) resulting from the disproportionation
of unstable FeO. Decomposition of ferric oxalate occurred
within a lower temperature interval, 137–177 ◦C [23,24].
From isothermal, vacuum decomposition studies it was esti-
mated that the reaction of iron(II) oxalate was completed in
about 10 min at 370–380 ◦C [14] and for iron(III) oxalate at
180–185 ◦C [23]. This difference, about 193 ◦C agrees well
with the value reported by Macklen [9] for the change of reac-
tion temperatures (192 ◦C), peak maxima, on heating iron(II)
Fig. 5. Plot of mass loss (%) against temperature (◦C) for the thermal decomposition of cobalt oxalate dihydrate on heating in N2 at 10 ◦C min−1. The reaction
rate is increased by the presence of Co3O4, but diminished by CoO, for discussion: see text.

66
M.A. Mohamed et al. / Thermochimica Acta 429 (2005) 57–72
oxalate in N2 and in O2 at 3 ◦C min−1. The present DTG ob-
servations measured the first peaks in N2 and in H2 at 264
and258 ◦C, respectively, and219–232 ◦CinO2, allsomewhat
above values expected for decomposition of the iron(III) salt.
The second peaks on heating in N2, 351 ◦C, was just below
the decomposition of iron(II) oxalate and the higher temper-
ature of reaction in H2, 407 ◦C, is ascribed to reduction of
residual oxide to iron metal product. From all the present re-
sults, we estimate the peak temperature maximum difference
between decompositions in N2 and O2 to be 150 20 ◦C.
of reaction. The mass loss, 55.5%, was the same as that in
N2.
3.2.3.4. In hydrogen. The DTG response (335–365 ◦C) for
decomposition on heating at 5 ◦C min−1 was an overlapping
doublet, maxima 344 and 354 ◦C, separated by a small mini-
mum. At 10 ◦C min−1, the components peaks were no longer
separated, the single maximum 359 ◦C was for the first rate
process. The total mass loss, 68.0%, identified cobalt metal
as the residual product.
3.2.3. Cobalt oxalate dihydrate
ObservationsandresultsaresummarizedinFigs. 1cand2c
and Tables 1–4.
3.2.3.5. In air. Reaction was very rapidly completed at both
heating rates, 5 and 10 ◦C min−1, and DTG response maxima
were detected at 294 and 303 ◦C. The mass loss, 56.0% is
close to expectation (56.1%) for the formation of product
Co3O4.
3.2.3.1. Dehydration. Water evolutions were completed be-
low 220 ◦C, in a single DTG response that did not overlap
with the subsequent decomposition. The measured mass loss
was 18.6%, at low heating rates but relatively increased with
heating rate, reaching expectation for the dihydrate, 19.7%
at 30 ◦C min−1. It is considered that this small inconsistency
does not affect the reactions of current interest.
3.2.3.6. Discussion. The rate and the products of decom-
position in N2 vary with heating rate (Fig. 4). At heating
rates below about 10 ◦C min−1, reaction was completed by
about 345 ◦C and the mass losses identify the products as
being predominantly Co3O4 and Co2O3. In contrast, at the
higher heating rates an initially relatively slow reaction rate
subsequently increased to yield residual products containing
CoO and cobalt metal. This pattern of behaviour is consis-
tent with enhanced reduction of initial product oxides, at the
higher heating rates, during temporarily high concentrations
of CO within the reactant mass. Separate experiments, Fig. 5,
showed that Co3O4 significantly promoted cobalt oxalate de-
composition whereas CoO significantly decreased the rate.
This pattern of conditions-sensitive behaviour is consistent
with previous work by Broadbent et al. [25] who found that
the kinetics of cobalt oxalate breakdown, 298–320 ◦C, were
influenced by the temperature of the precursor dehydration
step.
3.2.3.2. In nitrogen. DTG response peaks were broad and
slightly asymmetrical, the onset slope being steeper than the
subsequent return to the baseline. Maxima at both heating
rates were very close to 317 ◦C for responses at 10 ◦C min−1
(307–372 ◦C) and 5 ◦C min−1 (295–345 ◦C). Mass losses,
55.4% were slightly greater than expectation for Co2O3
formation (54.7%), evidence for the presence of Co3O4
(56.12%) in the residue.
Further comparative studies were made of the decompo-
sition in N2 at various heating rates (Fig. 4). Reactions at 2,
5 and 10 ◦C min−1 were rapidly completed between 290 and
345 ◦C, mass losses were close to 55.4% (Co2O3 with some
Co3O4). Heating at 20, 25 and 30 ◦C min−1 gave responses
close to a single line, which included a relatively slow re-
action 340–400 ◦C resulting in about 17% mass loss. This
was followed by a more rapid decomposition completed at
about 430 ◦C. The total mass losses of 63.0–63.7%, identify
the products as a mixture of comparable amounts of CoO and
cobalt metal. Intermediate behaviour was found for heating
At the lower rates of heating, decomposition occurred be-
tween 305 and 325 ◦C, yielding Co2O3 and Co3O4, whereas
on rapid heating reaction yielded metal in the product from
a reaction at up to 410 ◦C. From the above data, it is esti-
mated that oxygen reduces the reaction temperature by about
40 10 ◦C, which is appreciably less than the value 110 ◦C,
in [9].
at 15 ◦C min−1
.
Further experiments in N2 gas, Fig. 5, showed that the
addition of Co3O4 slightly decreased (−20 ◦C) the temper-
ature of onset of cobalt oxalate decomposition and reaction
was completed at about 330 ◦C, comparable with reactions
at the lower heating rates. The addition of CoO increased the
decomposition temperature, reaction proceeded more slowly
in a higher temperature interval and was only completed at
about 470 ◦C.
3.2.4. Nickel oxalate dihydrate
Observations and results are summarized in Figs. 1d
and 2d and Tables 1–4.
3.2.4.1. Dehydration. The first DTG response peak
(180–270 ◦C), maxima 230–260 ◦C, was identified as
dehydration and was always completed prior to the onset
of decomposition. Mass losses measured, 20.4–20.6%,
were slightly greater than expectation for the loss of 2H2O
(19.7%), presumably due to a small water excess retained
on drying.
3.2.3.3. In carbon dioxide. The DTG peak, maximum
340 ◦C at 10 ◦C min−1, was unsymmetrical (312–385 ◦C),
again onset was more rapid than the subsequent completion

M.A. Mohamed et al. / Thermochimica Acta 429 (2005) 57–72
67
3.2.4.2. In nitrogen. A single DTG response (330–420 ◦C),
broad and slightly asymmetric, gave maxima between 365
and 370 ◦C at both heating rates, 5 and 10 ◦C min−1. Overall
mass losses, were consistent with formation of a product mix-
ture containing both Ni metal and NiO. At low heating rates,
2, 5 and 10 ◦C min−1, the mass losses, 59.4%, agreed with
expectation for product NiO (59.1%) and at higher heating
rates, 15, 20, 25 and 30 ◦C min−1, the product contained as
much as 50% Ni metal. The peak maximum increased with
heating rate between 330 and 390 ◦C.
product was Ni or NiO. From the present observations, it is
concluded that the promoting influence of O2 (presumably
through cation oxidation, as with the other cations discussed
above) is relatively small, diminishing the reaction tempera-
ture by only10–20 ◦C.
3.2.5. Copper(II) oxalate
ObservationsandresultsaresummarizedinFigs. 1eand2e
and Tables 1–4.
3.2.5.1. In nitrogen. On heating at 5 ◦C min−1 a large, rel-
atively sharp (290–305 ◦C) peak was detected, maximum
296 ◦C. The mass loss, 55.5%, indicated the formation of
some copper metal (expectation 58.1%), but containing sig-
nificant amounts of the lower oxide (losses to form Cu2O
and CuO were 52.8 and 47.5%, respectively). The response
for heating at 10 ◦C min−1 was closely similar (maximum
300 ◦C) and the residue (56.9%) contained a higher propor-
tion of metal. Under both conditions, a slow mass increase
after completion of decomposition is ascribed to oxidation
by impurity O2 in the flowing atmosphere.
3.2.4.3. In carbon dioxide. The DTG peak maximum tem-
perature, 392 ◦C, was slightly higher than that in N2 and the
mass loss (60.0%) indicated NiO product.
3.2.4.4. In hydrogen. Decomposition occurred with peak
maxima at 327–350 ◦C and within a comparatively narrow
temperature interval, 20–25 ◦C. The mass loss indicated that
Ni metal was the only product.
3.2.4.5. In air. At both heating rates, 5 and 10 ◦C min−1, re-
action was completed within a relatively narrow temperature
interval (20–25 ◦C) and DTG peak maxima were similar to
those measured in H2, 330–340 ◦C. Residual masses were in
accordance with the formation of product NiO only.
3.2.5.2. In carbon dioxide. On heating at 5 ◦C min−1, a sin-
gle sharp peak (300–318.5 ◦C) was detected with maximum
at 310 ◦C and mass loss 56.0%. This reaction was very similar
to that in N2.
3.2.4.6. Discussion. There have been many previous studies
of the thermal decomposition of nickel oxalate [2,6,26,27]
under a variety of conditions and the residual products have
been alternatively identified as Ni metal, NiO and their mix-
tures, as reported above. The present results show that re-
actions under oxidizing conditions (in air, 330–340 ◦C) and
under reducing conditions (H2, 327–350 ◦C) proceed within
similar, and relatively narrow, temperature intervals. Decom-
positionsininertatmospheresoccurredatslightlyhighertem-
peratures, 365–370 ◦C peak maxima, and reaction in CO2
was even greater, 392 ◦C. These slight diminutions in reac-
tivity may be ascribed to some contribution, or participation,
by the volatile products in the controlling breakdown step
within the active reaction interface. CO2 and/or CO formed
and temporarily retained in the vicinity of the reaction contact
(nickel oxalate)/(Ni and/or NiO) is regarded as apparently in-
fluencing interfacial activity and thereby reducing slightly the
overall rate of salt decomposition.
The significant conclusion is that the rates of nickel ox-
alate decomposition are closely similar, whether the metal
and/or its oxide is the residual product. This is entirely con-
sistent with previous comparative studies [28] of the thermal
reactions of some nickel carboxylates (these did not include
the oxalate) in vacuum and in O2. It was concluded that re-
actions proceeded as interface processes and isothermal ki-
netic characteristics were similar, even identical, for each salt,
during reactions in vacuum and in O2. Although Arrhenius
parameters showed some variations with the presence or ab-
sence of oxygen, the absolute reactivity of each reactant was
influenced only slightly by O2 and/or whether the residual
3.2.5.3. In hydrogen. Reactions at 5 and 10 ◦C min−1 gave
relatively broad response peaks between, approximately, 225
and 300 ◦C, maxima at 270 and 283.5 ◦C, respectively. Mass
losses confirmed expectation that Cu metal was the only
residual product.
3.2.5.4. In air. The maxima of the very sharp response peaks
on heating at 5 and 10 ◦C min−1 were at 296.5 and 307 ◦C, re-
spectively. The mass loss at the higher rate of heating, 54.6%,
approached that expected for Cu2O formation whereas at the
lower rate indicated the almost exclusive formation of metal.
3.2.5.5. Discussion. In a previous study [16], it was demon-
strated that copper(II) oxalate decomposed through step-
wise cation reduction, Cu2+ → Cu+ → Cu◦. This mechanism
has subsequently been identified in the thermal breakdowns
of copper(II) formate and other copper(II) salts of organic
acids, see [16,29] and references therein. In copper(II) ox-
alate decomposition, 242–277 ◦C, the two consecutive reac-
tions overlapped to a greater extent than in the other related
reactants, to give a sigmoid-shaped isothermal yield–time
curve. Evidence for this two-step process [16] was obtained
from the kinetic analysis, which did not fit the usual solid-
state reaction rate models [2] and from analytical measure-
ments. Similarly, the present observations do not distinguish
individual contributions from two-rate processes.
Decomposition in vacuum yielded copper metal as the
sole residual product [16]. Under the present reaction con-
ditions, except in H2 where oxidation was not possible, the

68
M.A. Mohamed et al. / Thermochimica Acta 429 (2005) 57–72
mass losses gave evidence that appreciable proportions of the
oxides were retained in the products, attributable to O2 im-
purity in the atmospheres. The influence of O2 on the rate of
this reaction was small, this reactant had not been considered
in [9].
Few reports of solid-state decompositions place conclusions
in a wider, systematic and chemical, context.
The alternative approach used here, through comparative
investigations of reactivity data obtained for a set of related
reactants, recalls and develops an earlier method of eluci-
dating mechanisms of solid-state reactions [7,9,10]. Former
searches for systematic chemical trends and properties from
behaviour variations within sets of comparable reactions ap-
pears to have become replaced by a preoccupation with the
methods of calculating values of E, but without much ap-
parent interest in any chemical significance of the results so
obtained [3]. The present comparative study extends the for-
mer study [9] to measure variations of reactivity for this set of
six reactants that contain divalent transition metal cations oc-
cupying consecutive positions in the periodic table. The con-
sequences of selected variations of reaction conditions and
the presence of reactive gases were determined. The work
was also intended to elucidate chemical aspects of the con-
trols and mechanisms of the thermal reactions for this series
of oxalates. Systematic variations of behaviour, elucidating
the chemical properties of reactant cations, are expected to
contribute to the coherent and ordered advance of the subject.
The thermoanalytical (nonisothermal) responses observed
on heating each of the six reactants, in the different atmo-
spheres, have been compared with results from the already
reported (isothermal) studies, to characterize the chemical
changes and rate processes to which each of the peaks relate.
Some of the peak shapes observed differed from expecta-
tion, based on previous observations for the same reaction,
perhaps due to different experimental conditions. Such varia-
tions are of potentially considerable significance in quantita-
tive kinetic studies: peak shape determines the rate equation
from which the reaction geometry is deduced [2,6]. For ki-
netic studies is essential to ensure that the rate data measured
refer to rate processes controlled by the slowest reaction, the
chemical controlling step. During rate measurements under-
taken to elucidate reaction mechanisms, it is important to
confirm that reaction rates are not modified by secondary
controls, as appears sometimes to have been inadequately
recognized [3] in kinetic analyses of thermal data for crystol-
ysis reactions. Variations of apparent behaviour with reaction
conditions have important implications in the interpretation
of kinetic data and caution is always essential when inferring
chemical conclusions from a minimum of rate measurements.
However, such apparent variations are interpreted differently
here because comparisons of these effects between the differ-
ent reactions are of potential value in identifying the factors
thatcontributetowardsdeterminingabsolutereactivityand/or
reaction mechanisms and controls.
3.2.6. Zinc oxalate
Observations and results are summarized in Figs. 1f and 2f
and Tables 1–4.
3.2.6.1. In nitrogen, hydrogen and air. The TG response
peak temperatures for decomposition on heating at both 5
and 10 ◦C min−1 in these three atmospheres were identical,
within experimental error: a single peak 385–425 ◦C, max-
ima near to 397 and 407 ◦C, respectively, were detected. Mass
losses were close to 48.0%, which is about 1% greater than
expectation for the formation of ZnO (47.0%), evidence of a
small amount of dissociative sublimation on heating.
3.2.6.2. In carbon dioxide. On heating at 10 ◦C min−1, the
single peak was observed at about 10 ◦C higher. Presumably
CO2, adsorbed onto the product ZnO, slightly opposes anion
breakdown, which is unexpected because ZnCO3 is known to
decompose at a significantly lower temperature, CO2 is lost
at 300 ◦C [30].
3.2.6.3. Discussion. Apart from the small influence of CO2,
the decomposition of zinc oxalate was independent of the
presence of these gases. Reaction temperature is consistent
with previous studies [31,32] in which reaction yielded ZnO
without the intervention of the carbonate. The present ob-
servations detected no perceptible influence of O2 on zinc
oxalate decomposition, a reaction not studied in [9].
4. Conclusions
4.1. Experimental data and kinetic interpretation
The present comparative investigation of relative reactiv-
ities, within this set of chemically related and comparable
reactants, is intended to complement the more conventional
kinetic methods of elucidating thermal behaviour [2,6]. Stud-
ies of this type for the present reactants, undertaken to iden-
tify the rate equation and to calculate the Arrhenius param-
eters, to obtain insights into the reaction controls and mech-
anisms, have already been completed: representative litera-
ture citations are given above. The emphasis throughout has
been placed upon the, often older, isothermal kinetic mea-
surements because of perceived uncertainties [3] about some
methods and theories widely used in recent interpretations of
nonisothermal rate data. Many isothermal kinetic and mecha-
nistic research programs have been directed towards the prop-
erties and behaviour of a single, individual salt and usually
do not explicitly explore the possibility of trends, similarities
or differences across sets of chemically related substances.
4.2. Mechanisms of metal oxalate decompositions
The pattern of reactivities revealed by this study confirms
and extends the trend reported by Macklen [9]. It further pro-
vides support for the previous conclusion [13,14] that oxalate
stability is determined by the strength of the divalent cation

M.A. Mohamed et al. / Thermochimica Acta 429 (2005) 57–72
69
Fig. 6. Trend (1) (see text). Plot of mean reaction temperature, from the present study, for decompositions of the divalent oxalates in inert (N2, CO2) or reducing
(H2) atmospheres against the enthalpy of formation of the divalent oxide, MO. This trend supports the previous conclusion [13,14] that solid-salt stability is
determined by the metal to oxalate–oxygen bond strength, M OCO · COO · · ·.
bond with the carboxylate anion, · · · OCO M OCO · · · as
showninFig. 6. Anapproximatelylinearrelationshipisfound
between the mean DTG peak maximum temperatures for (di-
valent) salt decompositions in an inert atmosphere and the
enthalpy of divalent metal oxide formation, MO. The latter
parameter is regarded as providing a measure of the cation-
to-anion bond strength in the metal oxalate. The trend in both
parameters is:
correlate. This trend is regarded as less satisfactory whereas
the increase in reaction temperature with MO stability is con-
sistent with decomposition control by M O bond strength.
4.2.1. Isothermal activation energies
Magnitudes of E reported from isothermal studies of these
decompositions do not readily or obviously correlate with the
sequence in Trend (1). Representative values of E (kJ mol−1
)
from reports of decomposition studies of divalent oxalates
include the following: Mn: 141 [2], 180 [22]; Zn: 210 [31],
197 [32]; Fe: 157 [14]; Ni: 210 [26], 160 [27]; Co: (small),
168 [25] and Cu: (about) 200 [34], 180 [16]. No trend is
discerned, indeed the scatter of variations between individual
valuesreported by different workers, or evenwithin particular
studies of a single reactant [25,34], is comparable with the
apparent differences between the different salts. We conclude
that magnitudes of E do not provide a measure of absolute
reactivity [3], it appears that values are condition-dependent.
Other observations, from the present comparative stud-
ies, that identify potential difficulties in the interpreta-
tion of nonisothermal TGA kinetic data include the fol-
lowing. The stoichiometries and rate characteristics (the
yield–time–temperature curve shape) of the decompositions
of cobalt (Fig. 4) and nickel oxalates varied significantly with
reactionconditions. Thiswasidentifiedhereasaconsequence
of oxide reduction during those reactions in which a sufficient
CO concentration was maintained within the reactant sam-
ple. The possible participation of secondary reactions must
be considered in thermoanalytical kinetic studies where data
Mn > Zn > Fe > Ni > Co > Cu
(1)
The temperatures of completion of oxalate decompositions
vary significantly in the sequence: Mn(445 ◦C) > Zn(439 ◦C)
> Fe(409 ◦C) > Ni(387 ◦C) > Co(370 ◦C) > Cu(320 ◦C) (the
molten and monovalent potassium oxalate decomposes
at a significantly higher temperature, 600 ◦C [5]). This
systematic variability is regarded as evidence that rupture of
the ‘central’ C C anion bond is not the controlling step in
the oxalate decompositions [10–12].
This conclusion also differs from that of Macklen [9] who
finds that reaction temperatures decrease with decreasing
electronegativity of the metal and concludes that decomposi-
tion occurs by rupture of the C O bond to give the metal
oxide. This sequence was not adequately confirmed here.
The relatively small changes of electronegativities [33] for
Ni (1.91), Cu (1.90) and Co (1.88) do not follow the same
sequence as the significant differences of reaction tempera-
tures, given above. For the other salts the electronegativity
values show a wider variation, Fe (1.83), Zn (1.65) and Mn
(1.55) but again the reaction temperature differences do not

70
M.A. Mohamed et al. / Thermochimica Acta 429 (2005) 57–72
interpretation is sensitive to reaction–response curve shape.
It is further noted, for the decompositions of manganese and
ferrous oxalates, that the presence of a small quantity of O2
impurity in the (inert) flowing atmosphere introduces into
the reactant a sufficient amount of higher valent cations to
cause considerable distortions of the early response peaks.
Clearly, the provision of an alternative, preferred reaction
path through the formation of an active, but unintended and
unrecognized, intermediate (a higher valent cation) may in-
validate kinetic and mechanistic conclusions. Both of the ef-
fectsdescribedherearesignificantandarepotentiallycapable
of influencing apparent kinetic behaviour.
Oxygen (reactions in air) considerably promoted the de-
compositions of Mn and Fe oxalates (Table 1), reaction tem-
peratures were diminished by more than 100 ◦C. The effect is
significantly less for cobalt oxalate, about 40 ◦C, even lower
(15 ◦C) for nickel oxalate and not detected for copper(II) and
zinc oxalates, consistent with the observations reported by
Macklin [9]. Results are summarized in Fig. 7, showing Trend
(2):
Fe > Mn > Co > Ni > Cu
(2)
In Fig. 7, a measure of the promotion of decomposition by
O2 (the temperature difference between reaction in an in-
ert atmosphere and that in air) is plotted against the differ-
ence of the enthalpies of formation of the divalent oxide and
the other oxide participating. For copper, the Cu2O value
was taken and zinc forms only ZnO. More detailed investi-
gations of the mechanisms of these individual reactions are
required to identify the rate-controlling parameters. It is pos-
sible that the reactions with oxygen may proceed at a hetero-
geneous reactant–product interface, through the participation
of a residual oxide phase at which superficial oxidation oc-
curs. However, hitherto, little information has been obtained
concerning the chemical characteristics of such reactions.
From parallels with kinetics and mechanisms discussed
for the decompositions of nickel formate [35] and of copper
formate [36], there is the possibility that the reactions stud-
ied here proceed by heterogeneous catalytic breakdown of
the oxalate anion on the product phase, metal and/or oxide.
All these reactions have been studied within approximately
comparable temperature intervals. Because the factors con-
trolling such surface processes have not yet been reliably
or comprehensively characterized, it is premature to attempt
to formulate detailed mechanisms of the present reactions.
Chemical behaviour, however, appears to be more compli-
cated than has been recognized.
4.2.2. Chemistries and enthalpies of thermal reactions
The pattern of reaction enthalpies reported in Table 3 is
consistent with the variations of chemical characteristics be-
tween these salts. In both inert atmospheres, N2 and CO2,
ꢀH values for decompositions were similar and Trend (1) is
found again. A difference is that the (approximately similar)
values for the Ni and Co oxalates are here reversed which
may be due to different compositions of the product oxides.
We also note that the largest enthalpies, and highest reaction
temperatures, areassociatedwiththeoxide-formingreactions
(MnO, ZnO, FeO), lower for Ni and Co, where reaction con-
ditions determine whether oxide or metal is given (Fig. 4) and
lowest for the copper salt which yields the metal. Residual
products usually contain a proportion of oxide, which may
be evidence that MO is the primary decomposition product,
subsequently possibly (depending on conditions) reduced by
the CO concurrently formed. Enthalpies of reactions in H2
were generally less, attributable to the formation of CO and
usually the lower oxides. The exothermic reactions in air re-
sult from the oxidation of CO to CO2 and the formation of the
(sometimes) higher oxides of the residual metals. Enthalpies
correlate more satisfactorily with the specific chemical prop-
erties of the cations in these salts than with the pattern of
isothermal kinetic behaviour.
4.2.4. Comment: isothermal and nonisothermal kinetic
studies of the decompositions of divalent transition metal
oxalates
4.2.3. The role of oxygen in promoting decompositions
of transition metal oxalates
Published kinetic and mechanistic studies of the salts stud-
ied here [2,6] have characterized reaction geometries and
other aspects of the thermal rate processes for these indi-
vidual compounds. The forms of the yield–time curves have
been correlated with appropriate kinetic models representing
solid-state advancing interface reactions. Much less progress
has been made towards elucidating interface chemistry and
the overall controls of reactivity. Measured values of E have
not always been identified with a specific rate controlling
step, for each individual salt or within any collective pattern,
and chemically significant trends have not been recognized.
The present study is intended to complement this approach
through these comparative studies of absolute reactivities
of chemically related salts, together with the influences of
gaseous reactants. This is intended to introduce additional
perspectives from which the detailed chemistry of these re-
Decomposition temperatures of the Mn, Fe, Co and Ni
salts (those containing oxidizable cations) were significantly
diminished by the presence of oxygen. This trend followed
the pattern described earlier by Macklen [9], the largest
diminutions being found for the most easily oxidized cations.
He suggested that the first step in salt breakdown is cation oxi-
dation, M2+ → M3+. It is alternatively possible that the chem-
ical changes proceed at an active (solid-state) reaction inter-
face [2], here a metal oxalate/oxide contact, where oxygen at
the oxide surface interacts with the cations and anions of the
salt. This is represented as a heterogeneous catalytic-type re-
action in which adsorbed oxygen promotes anion breakdown
resulting in the formation of CO2 and metal oxide (whereas
retained adsorbed CO would be expected to inhibit activity
and possibly reduce the oxide).

M.A. Mohamed et al. / Thermochimica Acta 429 (2005) 57–72
71
Fig. 7. Trend (2) (see text). Plot of temperature differences between decompositions of divalent oxalates in inert or reducing atmospheres (N2, CO2, H2) and
those in oxidizing atmospheres against difference of enthalpy of formation of the divalent oxide (MO) and the other metal oxide, MnO3/2, FeO3/2, NiO3/2
CuO1/2, and CoO4/3. No value for CoO3/2 could be found, possibly explaining its relatively greatest distance from the line, no higher oxide of zinc is known.
,
actions may be elucidated in greater detail. Interpretations in-
corporating complementary types of evidence are more likely
to be reliable than conclusions based on limited data.
The observations reported above reveal a pattern of ab-
solute reactivity in which salt decomposition temperature is
identified as increasing with the strength of the M O bond,
Trend (1), Fig. 6. It is suggested, therefore, that control is
through electron transfer, involving rupture/dissociation of
the bond between oxalate–oxygen and the divalent cation.
Furthermore, from Trend (2) it is concluded that the availabil-
ity of more highly oxidized cations, M3+, (during reactions in
oxygen) facilitates this electron transfer, thereby diminishing
reaction temperature (and again recognizing the importance
of electron transfer). This model is represented mechanisti-
cally as follows.
followed by breakdown of the link with the anion. Subse-
quent steps involve rupture of the · · · C C · · · link in the ox-
alate radical ion (the weakest bond [12]) forming CO2 and
CO, the latter may result in reduction of initial MO prod-
uct to metal. It is suggested that this sequence of chemical
changes takes place at an active interface advancing progres-
sively into each reactant crystallite. No step can be regarded
here as a dominant (exclusive) controlling process, the acti-
vation energies cited in the text do not correlate with reac-
tion temperatures. The sequence of anion breakdown steps
may be similar in all the oxalates and, within the interface,
exert some influence (control) on the step identified above
as ‘dissociation’.
From the above pattern of behaviour we conclude that
chemical trends have been discerned within this set of ther-
mal decompositions involving related reactants. Informa-
tion not directly obtained from studies of single reactants is
thereby provided and is, therefore, to be welcomed as an al-
ternative approach to the elucidation of interface chemistry.
This method of comparative studies has been available for
decades, e.g. [9], but has become neglected in recent years.
It is recommended that the value of such considerations, ex-
ploiting the methods of systematic chemistry, should now be
reappraised in the search for future advancement in a field
where many reports appear as isolated and individual con-
tributions to the literature [3]. Scientific advance implies the
development and expansion of coherent and systematic theo-
retical foundations, trends and aspirations that are not always
obvious in many recent reports of solid-state decompositions,
crystolysis reactions [1,2].
The first step is dissociation through electron transfer (fa-
cilitated by a cation of higher valency):
O M2+ O2− CO · CO · O · M · O · · ·
→ · · · O M O · CO · CO · O M O · · ·
This may then revert to the original structure or rearrange
through breakdown of the oxalate radical, including rupture
at the weakest [12] bond (· · · C C · · ·):
· · · O M CO2 ↑ CO ↑
Reactant
Gases
↓ O M O · · ·(on or to residual solid product)
Residual Product
The temperature of the initial dissociation step is specific
to each cation, varies with its oxidation state and may be

72
M.A. Mohamed et al. / Thermochimica Acta 429 (2005) 57–72
[19] JCPDS-ICDD, PDF2 Data Base 1996.
References
[20] M.E. Brown, D. Dollimore, A.K. Galwey, J. Chem. Soc., Faraday
Trans. I 70 (1974) 1316.
[1] A.K. Galwey, M.A. Mohamed, J. Chem. Soc., Faraday Trans. I 81
(1985) 2503.
[2] A.K. Galwey, M.E. Brown, Thermal Decomposition of Ionic Solids,
Elsevier, Amsterdam, 1999.
[3] A.K. Galwey, Thermochim. Acta 397 (2003) 249;
A.K. Galwey, Thermochim. Acta 399 (2003) 1;
[21] D. Dollimore, J. Dollimore, J. Little, J. Chem. Soc. A (1969) 2946.
[22] P.E. Yankwich, P.D. Zavitsanos, Pure Appl. Chem. 8 (1964) 287.
[23] A.K. Galwey, M.A. Mohamed, Thermochim. Acta 213 (1993) 279.
[24] D. Broadbent, D. Dollimore, J. Dollimore, J. Chem. Soc. A (1967)
451.
A.K. Galwey, Thermochim. Acta 407 (2003) 93;
A.K. Galwey, Thermochim. Acta 413 (2004) 139.
[4] D. Dollimore, D.L. Griffiths, J. Thermal Anal. 2 (1970) 229.
[5] M.A. Mohamed, A.K. Galwey, S.A. Halawy, Thermochim. Acta 387
(2002) 63.
[6] M.E. Brown, D. Dollimore, A.K. Galwey, Comprehensive Chemical
Kinetics, 22, Elsevier, Amsterdam, 1980, p. 210.
[7] V.P. Kornienko, Ukrain. Khim. Zhur. 23 (1957) 159.
[8] D. Dollimore, Thermochim. Acta 117 (1987) 331.
[9] E.D. Macklen, J. Inorg. Nucl. Chem. 30 (1968) 2689.
[10] V.V. Boldyrev, I.S. Nev’yantsev, Yu.I. Mikhailov, E.F. Khairetdinov,
Kinet. Katal. 11 (1970) 367.
[25] D. Broadbent, D. Dollimore, J. Dollimore, J. Chem. Soc. A (1966)
1491.
[26] P.W.M. Jacobs, A.R. Tariq Kureishy, Trans. Faraday Soc. 58 (1962)
51.
[27] D.A. Dominey, H. Morley, D.A. Young, Trans. Faraday Soc. 61
(1965) 1246.
[28] B.R. Wheeler, A.K. Galwey, J. Chem. Soc., Faraday Trans. I 70
(1974) 661.
[29] M.A. Mohamed, A.K. Galwey, S.A. Halawy, Thermochim. Acta 411
(2004) 13.
[30] N.A. Lange, Handbook of Chemistry, 10th ed., McGraw-Hill, New
York, 1967, p. 345.
[11] K.O. Hartman, I.K. Hisatsune, J. Phys. Chem. 71 (1967) 392.
[12] V.V. Boldyrev, Thermochim. Acta 388 (2002) 63.
[13] R.J. Acheson, A.K. Galwey, J. Chem. Soc. A (1967) 1167.
[14] M.A. Mohamed, A.K. Galwey, Thermochim. Acta 213 (1993) 269.
[15] B.V. L’vov, Thermochim. Acta 364 (2000) 99.
[16] M.A. Mohamed, A.K. Galwey, Thermochim. Acta 217 (1993)
263.
[17] M.A. Mohamed, S.A. Halawy, J. Thermal Anal. 41 (1994) 147.
[18] D.R. Lide (Ed.), Handbook of Chemistry and Physics, 78th ed., CRC
Press, New York, 1997–1998, p. 12–191.
[31] P.E. Yankwich, P.D. Zavitsanos, J. Phys. Chem. 68 (1964) 457.
[32] J.D. Danforth, J. Dix, J. Am. Chem. Soc. 93 (1971) 6843.
[33] D.R. Lide, (Ed.), Handbook of Chemistry and Physics, 78th ed.,
CRC Press, Boca Raton, 1997–1998, p. 9–74.
[34] D. Dollimore, T.A. Evans, Y.F. Lee, Thermochim. Acta 194 (1992)
215.
[35] J. Fahrenfort, L.L. van Reyen, W.H.M. Sachtler, in: J.H. de Boer
(Ed.), The Mechanism of Heterogeneous Catalysis, Elsevier, Ams-
terdam, 1960, p. 23.
[36] D.M. Jamieson, A.K. Galwey, J. Catal. 34 (1974) 156.