Thermal decomposition of freeze-dried μ-oxo-carboxylates of manganese and iron

Article abstract of DOI:10.1016/s0040-6031(98)00611-x

The decomposition of freeze-dried mixed carboxylates of manganese and iron was investigated by means of DTA, TG, mass spectroscopy and X-ray powder diffractometry. The three main steps of decomposition are characterized as release of (a) H₂O, (b) carboxylic acid and CO₂/CO, and (c) the corresponding carbonyl compound and CO₂. In particular, the course of process (b) strongly depends on the stability of the metal-carboxylate link in the three investigated carboxylates. Well-crystallized single-phase manganese ferrites can be obtained on decomposition of formates of appropriate composition and thermal treatment of decomposition products at 600°C while maintaining a p(O₂) within the coexistence field of manganese ferrite.

Full text of DOI:10.1016/s0040-6031(98)00611-x

Thermochimica Acta 327 (1999) 173±180
Thermal decomposition of freeze-dried m-oxo-carboxylates
of manganese and iron
*
H. Langbein , S. Christen, G. Bonsdorf
Institute of Inorganic Chemistry, Dresden University of Technology, Mommsenstr. 13, D-01062, Dresden, Germany
Received 2 October 1998; received in revised form 16 November 1998; accepted 18 November 1998
Abstract
The decomposition of freeze-dried mixed carboxylates of manganese and iron was investigated by means of DTA, TG,
mass spectroscopy and X-ray powder diffractometry. The three main steps of decomposition are characterized as release of (a)
2 2 2
H O, (b) carboxylic acid and CO /CO, and (c) the corresponding carbonyl compound and CO . In particular, the course of
process (b) strongly depends on the stability of the metal±carboxylate link in the three investigated carboxylates. Well-
crystallized single-phase manganese ferrites can be obtained on decomposition of formates of appropriate composition and
2
thermal treatment of decomposition products at 6008C while maintaining a p(O ) within the coexistence ®eld of manganese
ferrite. # 1999 Elsevier Science B.V. All rights reserved.
Keywords: Manganese ferrite; Freeze-dried precursor; Thermal decomposition; Thermal analysis; Mass spectroscopy
1
. Introduction
composition of the solid products strongly depends on
the nature of precursor compounds and the thermal
treatment. The main reason for the different results
seems to be the dif®cult synthesis of compounds with
an exact Fe : M-stoichiometry. The conditions for a
reproducible synthesis are known only for a few
suf®cient less soluble m-oxo-acetates [5±7]. Because
of these dif®culties, we have developed a freeze-
drying method for synthesis of reactive amorphous
acetates with any M : Fe-stoichiometry. The freeze-
dried precursors show a decomposition pathway ana-
logous to the crystalline m-oxo-acetates of appropriate
composition [8,9].
Trinuclear m-oxo-carboxylates Fe (III)M(II)O-
2
(
RCOO)6L3 (R ˆ alkyl, H; L ˆ H O, RCOOH,
2
M ˆ Zn, Co, Fe, Ni, Mn) are characterized by the
metal-ion stoichiometry of spinels on the molecular
level. The thermal decomposition of these compounds
proceeds in the temperature interval from 408 to
4
carboxylates should be suitable precursors for the soft
008C [1]. From this point of view, the trinuclear
synthesis of ferrites, MFe O , with spinel structure.
4
2
Some investigations in this ®eld have shown that
thermal decomposition of appropriate m-oxo-carbox-
ylates leads to spinel ferrites but, in most cases, more
or less single oxides are formed [2±4]. The phase
This paper deals with the effect of the nature of the
carboxylate anion (formate, acetate, propionate) on
the thermal decomposition of freeze-dried Mn±Fe-
carboxylates. Common features and differences in
the decomposition behaviour are discussed. The most
*
Corresponding author. Tel.: +49-0351-463-4366; fax: +49-
0
0
351-463-7287; e-mail: Hubert.Langbein@chemie.tu.dresden.de
040-6031/99/$ ± see front matter # 1999 Elsevier Science B.V. All rights reserved.
P I I : S 0 0 4 0 - 6 0 3 1 ( 9 8 ) 0 0 6 1 1 - X

1
74
H. Langbein et al. / Thermochimica Acta 327 (1999) 173±180
suitable formates were also investigated in regard to
the phase formation in the solid products. Because of
the strong in¯uence of the reaction gas atmosphere on
the oxidation states of iron and manganese, the p(O2)
was measured continuously with the help of solid
electrolyte cells.
2
. Experimental
The complex carboxylate solutions for the freeze-
2
Fig. 1. Experimental arrangement for continuous p(O ) measure-
drying process were obtained as follows: Mn(II)- and
Fe(II)-carboxylates were prepared by dissolving the
metal powders in the appropriate carboxylic acid in an
inert atmosphere. After that, the Fe(II)-carboxylate
ment during decomposition. C1, C2, represent solid electrolyte
cells, working temperature 6008C. 1, flow meter; 2, furnace; and
N gas flow 10 l/h.
2
�
2
� 1
was oxidized in a 10 ±10 M solution of the appro-
priate carboxylic acid with a twofold excess of H O .
cells, C1 and C2, were calibrated with gas mixtures
of known p(O2).
2
2
A deep brown Fe(III)carboxylate solution was formed.
The maximum attainable concentrations are 0.2 M
(
formate), 0.6 M (propionate) and 1 M (acetate). An
3. Results and discussion
adequate quantity of the appropriate Mn(II)carboxy-
late solution was added to 500 ml of the Fe(III)car-
boxylate solution. The mixture was quickly frozen in
liquid nitrogen. The drying process was carried out in
a vacuum chamber of a freeze-drying apparatus
Fig. 2 shows the TG curves of the freeze-dried Fe±
Mn-formate, Fe±Mn-acetate and Fe±Mn-propionate.
In an inert atmosphere, all the decomposition pro-
cesses are endothermic. Decomposition takes place in
three separate main steps. The mass spectroscopic
analysis of gaseous products allows the conclusion
that these steps represent ®ve somewhat superimposed
decomposition processes. The temperature ranges for
these processes and the appropriate main primary
gaseous decomposition products are listed in Table 1.
Fig. 3(a and b) illustrate the mass spectroscopic ana-
lysis for the freeze-dried formate in a more detailed
form. Instead of the also observed molecule peaks of
�
3
(
Christ) from � 40 to ‡208C at 10 mbar. The com-
position of the ®ne-grained, soft agglomerated pow-
ders were determined by complexometric titration and
elemental analysis as follows:
MnFe2OꢀHCOO† ꢀH2O†3
6
MnFe OꢀCH COO† OHꢀH O†
2
3
5
2
MnFe OꢀC H COO† ꢀH O†
2
2
5
6
2
3
Depending on the conditions during the freeze-drying
process, the water and carboxylic acid content can be
somewhat different. The analysis of the IR spectra
allows the conclusion that in the freeze-dried amor-
phous products at the very least partially complex
trinuclear m-oxo-carboxylate species are present.
The thermal decomposition was investigated by
means of a Netzsch thermal analyzer STA 409,
coupled with a mass spectrometer QMS 125 (Balzers).
X-ray diffraction was performed using a powder
diffractometer D5000 (Siemens). The experimental
arrangement shown in Fig. 1 was used for the con-
HCOOH (m ˆ 46) and H CO (m ˆ 30), the more
2
intensive fragment peaks of COOH (m ˆ 45) and
HCO (m ˆ 29) are shown. For all the investigated
carboxylates the decomposition is almost complete at
ca. 3508C. The solid decomposition products of acet-
ates and propionates contain, respectively, 1±3% car-
bon formed by pyrolytic side reactions. In an oxygen
containing atmosphere, this carbon is oxidized at ca.
6008C.
The decomposition starts with the elimination of
coordinated H O (process 1). In case of acetate and
2
propionate, the H O elimination is superimposed by
2
tinuous measurement of p(O ) throughout the thermal
2
the elimination of carboxylic acid. This process, 2(a),
can be understood as a thermal hydrolysis of a metal±
decomposition. The ZrO -based solid electrolyte
2

H. Langbein et al. / Thermochimica Acta 327 (1999) 173±180
175
�
1
Fig. 2. Thermal analysis of the freeze-dried Mn±Fe-carboxylates. Heating rate 5 K min ; atmosphere, argon; 1, formate; 2, acetate; and 3,
propionate.
O(CO)R-link without alteration of the oxidation num-
ber of metal ions. On the contrary, the elimination of
formic acid begins only after the end of the delivery of
the decomposition products represent only the com-
position. It must be supposed that after the dehydration
reaction all processes are cross-linkages of molecular
species forming M±O±M bridges. As in any thermal
decomposition involving reactions between primary
gaseous products and solids or gas-phase reactions,
the actual yield in each reaction product is very
sensitive to experimental conditions (heating rate,
gas ¯ow, sample mass).
H O (process 2(b)). Formic acid originates in an
2
elimination of two HCOO-moieties connected with
the reduction of Fe(III) to Fe(II) and an H-transfer
process from one moiety to the other (2 HCOO!
HCOOH‡CO ). In case of acetates and propionates,
2
an analogous reductive elimination process also
begins at ca. 2008C. The simultaneous formation of
According to Dollimore and Tonge [10], there are
two main routes for the decomposition of anhydrous
metal formates to oxides:
carboxylic acid and CO needs the additional forma-
2
tion of CH and C H . An indirect proof of the reactive
2
2 4
CH species is the mass spectroscopic detection of
2
MꢀOOCH† ! MO ‡ 2CO ‡ H2O
(1)
(2)
propanoic acid during the decomposition of the
freeze-dried acetate. Propanoic acid can arise by a
characteristic insertion reaction of CH in a C±H-link
2
MꢀOOCH† ! MO ‡ CO ‡ CO2 ‡ H2
2
2
of acetic acid.
The authors refer to the formation of different organic
byproducts (for instance H CO) depending on reac-
Following the reduction of Fe(III) to Fe(II), car-
boxylic acid is released together with CO instead of
CO . In case of acetates and propionates this process
2
tion conditions. In the case of manganous formate, the
gas evolution analysis suggests that at the beginning of
decomposition the reaction (1) is favoured kinetically,
but gradually assumes less importance in comparison
to reaction (2). This result seems to be contradictory to
our results, but considering the different reaction
conditions the difference can be understood. The main
results in [10] were obtained in time-dependent
experiments at #>4008C without variation in the
oxidation number of metal ions. The lower tempera-
ture in our experiments allows the detection of pri-
2
2
(c) is completely superimposed with process 2(b) and
partially superimposed with process 3. On the other
hand, in case of formate the decomposition processes
2
(b) and 2(c) can be observed separately but the
processes 2(c) and 3 are completely superimposed.
The results of the decomposition of the formate
sample are summarized in Fig. 4. The mass losses
observed by TG analysis correlate with the process
steps in the reaction scheme. The given formulas for

Table 1
2
Primary gaseous decomposition products of freeze dried MnFe -carboxylates
Formate
process
Acetate
Propionate
temperature
#max/8C
products
temperaturerange/
#
max/8C
products
temperature range/
#
max/8C
products
range/8C
8C
8C
1
70±200
112
H
2
O
70±200
110±210
210±300
210±300
250±330
140
150
265
265
300
H
2
O
70±200
100±180
180±280
180±280
260±330
110
140
251
251
293
2
H O
2a
2b
2c
3
CH
CH
CH
CH
3
3
3
3
COOH
C
2
C
2
C
2
C
2
H
H
H
H
5
COOH
200±290
260±350
260±350
270
324
324
HCOOH, CO
HCOOH, CO
2
COOH, CO
COOH, CO
2
5
5
5
COOH, CO
COOH, CO
2
HCOH, CO
2
COCH
3
, CO
2
2 5 2
COC H , CO

H. Langbein et al. / Thermochimica Acta 327 (1999) 173±180
177
Fig. 3. Observation of mass numbers during the decomposition of the formate precursor. (a) m ˆ 18, H
2
O; m ˆ 44, CO
2
; m ˆ 45, HCOO
fragment; and (b) m ˆ 28, CO; m ˆ 29, HCO fragment; and m ˆ 44, CO
2
.
mary gaseous products HCOOH and H CO as main
2
Independent of kinetic in¯uences, the exothermic
reaction (3) should gradually assume less importance
with rising temperature, unlike the endothermic reac-
tion (4). The absence of CO below 3008C can also be
explained by the consecutive reaction (5).
components besides CO, CO , H O and H which are
2
2
2
also observed. Because of the reductive decomposi-
tion step 2(b), CO instead of CO is formed at ®rst in
2
addition to HCOOH. Decomposition of HCOOH takes
place already below 3008C in accordance with
Eqs. (3) and (4).
CO ‡ HCOOH ! CO ‡ H CO
(5)
2
2
The shoulder in the intensity graph of the fragment
pr ee aa ck t iwo ni t .h m ˆ 29 at ca. 2808C (Fig. 3(b)) refers to this
HCOOH ! H2 ‡ CO2
(3)
(4)
HCOOH ! H O ‡ CO
2

1
78
H. Langbein et al. / Thermochimica Acta 327 (1999) 173±180
H CO, according to Eq. (6)
2
H2CO ! H2 ‡ CO
(6)
results in a reaction corresponding to Eq. (2).
Traces of oxygen in the gas ¯ow cause a minor
importance of the reductive elimination of formic acid
and, in agreement with this fact, the formation of
Mn(II)±Fe(III)-oxide. Decomposition in air atmo-
sphere results in the formation of a-(Fe,Mn) O .
2
3
In an atmosphere poor in oxygen, the p(O) above
2
the sample in the course of decomposition is deter-
mined by the released gaseous products. Fig. 5 repre-
sents the results of a continuous measurement of p(O2)
during the decomposition of a freeze-dried formate in
an arrangement shown in Fig. 1. The dotted area
designates the oxygen coexistence pressure for sin-
gle-phase spinels MnFe O at 6008C. The upper and
2
4Æꢀ
lower p(O ) were calculated according to the hypo-
Fig. 4. Decomposition scheme for Mn±Fe-formate.
2
thetic coexistences (7) and (8) using thermodynamic
data of Barin [11].
After ®nishing the reductive elimination, the for-
mation of HCOOH is connected with the release of
CO (process 2(c)). Taking into account the consecu-
tive reaction (4), the whole process fully corresponds
to reaction (1).
2Mn3O4 ‡ 6Fe2O3@6MnFe2O4 ‡ O2
MnFe2O4@6MnO ‡ 4Fe3O4 ‡ O2
We have shown that this approximation is suf®-
ciently exact for the determination of preparation
conditions for manganese ferrite [12]. Actually, the
(7)
(8)
6
In a completely overlapped reaction 3 (Table 1 and
Fig. 4) CO and H CO are formed. Decomposition of
2
2
Fig. 5. Analysis of p(O
phase spinels MnFe
2
) during the thermal treatment of the Mn±Fe-formate precursor. Dotted area, oxygen coexistence pressures for single-
2
O
2 2
4Æd at 6008C; cell 1, p(O ) at C1; cell 2, p(O ) at C2 (see Fig. 1).

H. Langbein et al. / Thermochimica Acta 327 (1999) 173±180
179
reduction of Fe(III) in MnFe O at the lower phase
2
oxidation of the solid decomposition product. This
is con®rmed by the X-ray analysis of samples taken at
B, C and D (Fig. 6). The portion of spinel phase
MnFe O in the sample B is noticeably higher than
this portion at A. After a reaction time corresponding
to point C, a single-phase manganese ferrite spinel is
obtained. In view of the small phase width Áꢀ in
4
boundary, ®rst of all, results in the formation of an Fe-
2
‡
2‡
3‡
rich spinel, (Mn1
Fex Fe2 O ), besides an Mn-
� x
rich manganow uÈ stite phase (Mn1 Fe O). At the
4
� y
y
2 4
upper phase boundary the oxidation of Mn(II) ®rstly
leads to the formation of an Mn-rich spinel
2
Mn Mn2
‡
3‡
3‡
(
Mn) O .
Fex O ) and an Fe-rich a-(Fe,-
� x
4
MnFe O
2
after point C, the p(O ) measured at C2
2
2
3
4Æꢀ
The oxygen content in the N gas stream before the
2
quickly increases to a value higher than the coexist-
ence pressure at the upper phase boundary of manga-
nese ferrite. As a consequence, sample D already
contains some a-Fe O or an Fe-rich a-(Fe,Mn) O
sample is held constantly and measured with the solid
�
4
electrolyte cell C1 as p(O ) ˆ 10 bar. Below a
2
sample temperature of 2008C, the same oxygen partial
pressure in the gas stream is measured at cell C2 after
2
3
2 3
phase. Because of the very low oxygen content in the
gas stream, the complete formation of equilibrium
phases containing only Mn(III) needs a very long
reaction time.
passing the sample. Above 2008C, the p(O ) measured
2
at C2 decreases because of the reaction of gaseous
decomposition products with traces of oxygen in the
gas stream. At point A in Fig. 5, the decomposition is
®
nished and the released gases are outside the reaction
zone. In correspondence with the above-mentioned
assumption, the X-ray powder diffractogram of the
solid A demonstrates the presence of a w uÈ stite phase
4. Conclusions
Thermal decomposition of freeze-dried complex
carboxylates of Mn and Fe is a suitable pathway to
synthesize manganese ferrites with spinel structure.
Well-crystallized single-phase ferrites can be obtained
already on decomposition of formates of appropriate
composition and thermal treatment of decomposition
beside a spinel phase (Fig. 6). When the p(O ) is
2
measured at C2 over a long period, it continuously
increases and reaches the original value measured at
C1 after a duration of more than 24 h. The local
minimum at ca. 380 min/5508C is probably originated
by the oxidation of traces of pyrolysis carbon. The
further consumption of oxygen is due to the slow
products at 6008C, maintaining a p(O ) within the
2
coexistence ®eld of manganese ferrite.
Fig. 6. X-ray powder diffractograms of solid decomposition products (formate precursor). Sampling after a thermal treatment corresponding
to points A, B, C and D (see Fig. 5).

1
80
H. Langbein et al. / Thermochimica Acta 327 (1999) 173±180
The thermal stability of precursors decreases in the
References
sequence formate±acetate±propionate. In the same
sequence, the content of pyrolytically formed carbon
in the decomposition products increases. In case of
acetates and propionates a thermal hydrolysis process
forming the corresponding carboxylic acids already
begins below the boiling point of the acid. This results
in a partial separation of components in the former
homogeneous precursor. The worse homogeneity
involves a higher temperature for the synthesis of
single-phase complex oxides. Moreover, the inter-
mediate occurrence of ¯uid suspensions leads to hard
agglomerated solids. In case of formate decomposi-
tion, the release of formic acid begins only above the
boiling point and loose, soft agglomerated powders
are formed.
[1] Kh.M. Yakubov, V.A. Logvinenko, T.A. Zhemchuzhnikova,
Sh.Kh. Abdullaev, G.V. Gavrilova, A.N. Mikheev, J. Thermal
Anal. 30 (1985) 1095.
[
[
[
[
[
2] L. Patron, A. Contescu, G. Munteanu, M. Brezeanu, E. Segal,
Rev. Roum. Chim. 31 (1986) 811.
3] I. Ilie, M. Brezeanu, L. Patron, E. Segal, G. Filoti, Rev. Roum.
Chim. 32 (1987) 1109.
4] I. Ilie, M. Brezeanu, L. Patron, E. Segal, Rev. Roum. Chim.
33 (1988) 803.
5] A.B. Blake, A. Yavari, W.E. Hatfield, C.N. Sethulehshmi, J.
Chem. Soc., Dalton Trans. 12 (1985) 2509.
6] Ch.T. Dziobkowski, J.T. Wrobleski, Inorg. Chem. 20 (1981)
679.
[
[
7] H. Langbein, P. Eichhorn, Z. Chem. 30 (1990) 4.
8] H. Langbein, C. Michalk, K. Knese, P. Eichhorn, J. Eur.
Ceram. Soc. 8 (1991) 171.
[
9] S. Christen, H. Langbein, K. Jaenicke-R oÈ bler, Thermochim.
Acta 209 (1992) 253.
[
[
10] D. Dollimore, K.H. Tonge, J. Inorg. Nucl. Chem. 29 (1967)
21.
Acknowledgements
6
11] I. Barin, Thermodynamic Data of Pure Substances, VCH
Verlagsgesellschaft mbH, Weinheim, 1989.
The authors gratefully acknowledge ®nancial sup-
port by the Deutsche Forschungsgemeinschaft and the
Fonds der Chemischen Industrie.
[12] G. Bonsdorf, K. Sch aÈ fer, K. Teske, H. Langbein, H. Ullmann,
Sol. Stat. Ionics 110 (1998) 73.