ELECTRONIC, MAGNETIC AND CATALYTIC PROPERTIES OF COBALT CONTAINING SPINEL AND PEROVSKITE
TYPE MATERIALS
G.W. Herzog, A. Reichmann, K. Reichmann, IVI. Ruplitsch-Lesnik Institut für Chemische Technologie Anorganischer Stoffe
and K. Gatterer Institut für Physikalische und Theoretische Chemie Technische Universität Graz
Keywords: spinel type materials, perovskite type materials, Co Cobalt spinel type materials C03O4, Co Cobalt perovskite type materials LaCoOa. electrical properties, magnetic properties, meclianicai properties, correlations between properties, practical examples
Abstract: Correlations between electric, magnetic and catalytic properties of various Cobalt containing spinel and perovskite type materials (e.g. C03O4, LaCoOa) were investigated witfi respect to application in environmental catalysis. As a fast test reaction tfie CO oxidation with air up to 400°C was chosen and found to be initiated by high-spin Co^"^ ions at the surface. Thus, catalytic activity depends strongly on the condition of materials preparation. From comparison of the activation energies of the oxidation rate and hopping bulk charges it is concluded, that an activated charge transfer between Co^"*" and Co^^ ions at the surface is the rate determining step. However, the bulk conductance itself does not play any role for catalytic activity. A hindrance for broad application of perovskite type catalysts is the fast deactivation by SO2. As it is chemisorbed and oxidized at the active Co^^ sites to give sulfate, SO2 stops the CO oxidation immediately. Only a partial recovery is possible by heating at higher temperatures, therefore the substitution of Pt metal catalysts by such materials fails, when S02 is present in a waste gas.
Električne, magnetne in katalitske karakteristike materialov spinelnega in perovskitnega tipa, ki vsebujejo kobalt
Ključne besede: materiali tipa spinel, materiali tipa perovskite, Co Kobalt materiali tipa spinekC0304, Co Kobalt materiali tipa perovskite LaCoOs, lastnosti električne, lastnosti magnetne, lastnosti katalitične, korelacija med lastnostmi, primeri praktični
Izvleček: Preiskovali smo korelacije med električnimi, magnetnimi in katalitskimi karakteristikami različnih materialov, ki vsebujejo kobaltove ione in imajo spinelno ali perovskitno strukturo, na primer C03O4 in LaCoOs. Za hitro testiranje smo izbrali oksidacijo CO v zraku pri 400°C. Ugotovili smo, da je reakcija katalizirana s Co^"^ ioni z visokim spinom, ki se nahajajo na površini. To pomeni, da je aktivacijska energija katalize močno odvisna od priprave materiala. Iz primerjave aktivacijskih energij hitrosti oksidacije in »hopping« prevajanja v samem materialu smo ugotovili, da je aktiviran prenos naboja med Co^^ in Co^^ ioni na površini tisti, ki določa hitrost reakcije. Prevodnost samega materiala ne vpliva na katalitsko aktivnost.IVložnost široke aplikacije katalizatorjev perovskitnega tipa pa preprečuje hitra deaktivacija teh katalizatorjev z SO2. SO2 se adsorbira na površini in na aktivnih Co^"^ mestih oksidira v sulfat ter tako takoj prepreči oksidacijo CO. Katalitska aktivnost se lahko samo delno obnovi s segrevanjem pri povišani temperaturi. Zato zaenkrat ni možna zamenjava katalizatorjev na osnovi platine s takimi materiali, če je v odpadnih plinih prisoten SO2.
INTRODUCTION
Since more than two decades Co containing spinels (e.g. C03O4) and perovskites (e.g. LaCoOs) are well-known materials of liigh catalytic activity /1, 2/, for instance for gas phase oxidation with air. Further, applications in gas sensors (for CO) were taken into consideration /3,4/. Unfortunately the relevance of these materials to environmental catalysis or other purposes seems to be extremely restricted and most of the existing literature is not very promising /5-7/. When applied to real gas mixtures, there is a high risk of poisoning. On the other hand, these or similar mixed valence materials are catalytically as active as Pt or Pd and a possible substitution of the latter by oxides is still a certain commercial challenge.
The outstanding catalytic activity is supposed to be connected with a special spin or bond situation of the Go ions, which can be two, three or fourfold charged /8-13/. Most of such ideas are based on mixed-valence and crystal field models, but a direct experimental cor-
relation or evidence for a favored high spin action has never been obtained. A similar lack of knowledge exists for a possible correlation with electric properties, which would be vital for an extended use for sensor elements.
In general spinel and perovskite type materials can be tailored in a wide stoichiometric range, behaving either insulating, semiconducting or metallic in the catalytically interesting temperature range /14-16/. So the first aim of this work was to prepare thermally stable and active materials and to investigate, whether magnetic or electric correlations with catalytic activity can be found or not. In following this way, of course, special attention had to be paid to the interpretation of results because of the basic differences of bulk and surface properties.
The second aim of this work was to investigate the applicability of spinel and perovskite type materials for the CO oxidation in waste gases with a high level of SO2, caused by the low cost incineration of S containing fuels. To gain more basic know-how for an eventual
application, catalytical experiments were carried out not only with powders but also with ceramic honey combs under realistic gas conditions, i.e. with reasonable space velocities /17,18/.
EXPERIMENTAL
For our investigations we chose materials, that were proved to be thermally stable in the temperature range up to 600°C. Those were:
1.	Spinel type materials: ZnNiMn04, C03O4.
2.	Perovskite type materials: LaCoOs, LaNiOs, LaNii-xCoxOs, LaAH-xCoxOs, LaTii-xCoxOs, Lao.sSro.sCoOs.
Except C03O4, which was purchased from MERCK (analytical grade), both spinel and perovskite type materials were prepared by coprecipitation of about 0,05m nitrate solutions with 3 m NaOH. After careful washing with deionised water and drying at 110°C, the calcination of hydroxides was carried out in air at different temperatures up to 1150°C for several hours. To characterize the powders the chemical stoichiometry was determined by wet chemical methods, BET surface was measured with a Quantachrome surface analyser and X-ray spectra were taken on a Siemens D5005 diffrac-tometer. For susceptibility measurements the native powders were used of C03O4, LaCoOs and LaTii-xCoxOs at temperatures between 80K and 600K the magnetic susceptibility was measured with a modified Faraday balance. Low-temperature measure-ments between 4,2 and 80 K were performed with a Lake Shore 7121 magnetometer using liquid Helium for cooling. KCr(S04)2.12H20 was used for calibration in both cases.
For electrical measurements the powders were pressed into pellets and sintered taking into accountthe range of stability. The DC bridge consisted of a Wenking Galvanostat LT-78 and a Knick Voltmeter. Catalytical measurements of the CO oxidation were usually made with spherical agglomerates of 0,1 to 1/jm particles, sieved into a fraction of 0,5 to 1,25 mm, in an electrically heated double-tube reactor made of Pyrex glass with a inner diameter of 2 cm. The gas flow was kept constant by means of two Brooks mass flow meters 5858 TR at a space velocity of about 50 000 h""!. Most experiments were carried out with a mixture of air (H2O content <0,5vol%) and CO (0,1-1vol%). In some cases instead of air a water free mixture of 2% O2 and 98% N2 was used. The influence of SO2 was studied by adding 0,1 % SO2 to the normal gas stream. The conversion of CO into CO2 was continuously monitored by means of a CO/CO2 IR detector Ultramat 22P from Siemens, varying the temperature stepwise. The reactor was usually filled with 2-5 g material equivalent to a height of about 1-3 cm.
Additionally small honey combs (2,5x2,5x10 cm with 25 channels) were prepared from powder mixtures of TiOs/WOs (10%wt)/LaCo03 (20%wt) and Si02(90 %wt) /LaCo03(10 %wt) by extrusion, with the know-how and facilities of the catalyst producer FRAUENTAL KERAMIK, Austria. These honey combs were tested with a separate micro reactor, which was in use at the above mentioned company.
XPS data from literature show, that the Co content at the surface strongly depends on calcination temperature /19-21/. Therefore we varied additionally the calcination temperature of LaCoOs between 620 and 1150°C. Some calcined products were treated with water, because this significantly enhanced the catalytic activity. For comparison we included a commercial support coated with Pd (supplied by Kanzler Verfahrenstechnik, Austria) into our investigations.
RESULTS
Catalytic Properties
In the following figures all conversion/temperature curves refer to the gas inlet temperature, because the heat evolution during oxidation and thereby caused temperature distribution makes the kinetics a dynamical one. So self heating, formally acting as an autocata-lytic reaction was the reason why we usually did not analyze the kinetics. But from several measurements approaching equilibrium temperatures, achieved with a smaller amount of catalyst and CO (reduced self-heat-ing), we can roughly state, that up to a high degree the conversion curves correspond to an Arrhenius' law of activation, excluding a transport limitation in the gas phase up to about 300°C. It is remarkable that activation energies of such a variety of materials like LaCoOs, LaNiOs, C0SO4, prepared at different calcination temperatures are almost equal and amount to 0,6± 10%eV (fig. 1).
300	250
T CO]
200	150
£ c
S 10-
\
. V
■	.3C0O, icalciisiion öSO'C 5ri 3m--
•	{8WC 5h.
A	LaCoO, (550'C. 5h ! Im^.-g)
T	LaCoO	;fi 12m-.g>
♦	Co O (lOrr.-'g,
2.2	2 4
1000/T [K ■)
Fig. 1: Arrhenius ' evaluation of conversion curves.
As can be noticed from fig. 2, one of the most CO active substances except Pd covered supports is C03O4. CO/CO2 conversion starts below 100°C and reaches values of 100% at about 140°C, whereas LaCoOs needs at least 50° higher inlet temperature. The conversion curves of the Ni and AI substituted LaCoOs materials display one common feature: the higher the Co content in the bulk and hence at the surface, the more active is the material. Because of the much higher Co content, C0SO4 fits well into this simple model.
A similar conclusion can be drawn from the conversion behavior of the honey combs and the reground honey combs, showing no essential difference. Such honey
connbs work like a mixture of powders. Furthermore, it does not matter, whether the ions substituted for Co in LaCoOs are active 3d-ions (Ni3+) or not (A|3+). But at this stage we cannot decide, which kind of Co ions with respect to valence and spin state at the surface is responsible for the catalytic oxidation. The support Ti02/W03, which is the commercial basis for ceramic catalysts for NOx-reduction, is inactive for CO-oxidation and the activity depends only on the amount of LaCoOs.
40 60 80 ICO 120 140 1 60 160 200 220 240 260 280 300 320
TpC]
Fig. 2: Conversion cun/es of various materials (reactor filling 5 g).
Two preparation phenomena, which lead to a veritable increase of the catalytical activity are documented in figs. 3a and 3b. The first one concerns the calcination temperature and the second one the treatment of already calcined products with water. Calcination at lower temperatures leads to higher, surface-normalized reaction rates and treatment with water shifts the conversion curve to lower temperatures.
By normalization with respect to surface area for the conversion curves in fig.3a an activity ratio of 17:7:4 is found. X-ray analysis of these materials proved perovskite type structure and no additional impurity phases were found. For materials calcined at lower temperatures we assume, that these are more active because of unoxidized Co2+ species. On the other hand, the treatment with water (fig.3b) might have cleaning or dissolution effects, or the Co3+ ions are reduced to more active Co2+ ions approaching the case of C03O4.
70 60
O 30 30
— LaCoO, powdetKH^O —L3C0O powde* —•— honoycomb+H^O —O— hon By comb
/ // /
? / i f
n !
ß
y
a
B.»

100 120 140 ',60 ISO 200 220 2'lG
Trc]
Fig. 3b: Shift of conversion curves by water treatment.
Some additional experiments were carried out with Sr2+- and Ti'*+-substituted materials (fig. 4). Both the Sr2+ substitution of La3+ ions and the substitution of Co3+ ionsshiftthe conversion to lowertemperatures.
Here and in all other cases of Co and Ni containing catalysts the influence of SO2 is extremely negative: The CO oxidation is completely stopped within a few minutes, indicating a lethal chemical adsorption of SO2 or SO3. As shown by Lao.sSro.sCoOa in fig. 4 a certain but not total recovery is only achievable by heating up to temperatures above 600°C. Pd coated supports do not show this behavior and oxidize SO2 into SO3 in parallel to CO, even in a 2% O2 containing gas.
t c'c]
Fig. 4: Conversion influenced by substitution and SO2 poisoning.
/
J f /
/ y	-
62a'C/2b teom^ 650'C.'5h {S5m<") A 850-C;Sh (ISm^
250	300
tfc)
Fig. 3a: Conversion curves of 5g LaCoOs calcined at different temperatures.
Electrical Properties
Referring to room temperature, the bulk conductivity of sintered pellets of the materials under investigation range from metallic (e.g. LaNiOs) to semiconducting (e.g. LaCoOa, ZnNiMn04) and low conducting materials (heavily AI and Ti doped LaCoOs) (fig. 5).
In the catalytically interesting temperature range of 100 to 400°C, LaCoOs behaves semiconducting with an activation energy of about 0,55eV /22-24/. Above 400°C the behavior changes to a metallic type. Doping with Ti decreases the conductivity and increases the activation
energy to about 0,6-0,7eV /25-28/. C03O4 also behaves semiconducting, but with an activation energy changing from low to high temperatures (0,3-0,7eV) /29, 30/. On the other hand, in a wide range LaNiOs is a Pauli type metallic conductor /31, 32/.
Trc]
400
200
100
103-
102 ,
10°-10-' -
fe 10-J
o «
C
cn
10-3-

10-5-
10-'^



-•-LaCoOj(l200'C) - + - COjO^ (850-C) -A-ZnNiMnO^{1120*C) -T- LaAI^,COj^jOj(1250-C)
1,5
2.0
2,5
3.0
3.5
1000/T [K-1]
Fig. 5: Conductivity data of some active materials.
Magnetic Properties
C03O4 is a mixed-valence compound with normal spinel structure. Co2+ ions are located at tetrahedrally and Co3+ ions at octahedrally oxygen coordinated lattice positions /33, 34/. It was first shown by Cossee /35/, that the observed magnetic susceptibility is consistent with a material, where Co2+ ions are high-spin (total spin S=3/2) with a spin-only moment of /jso=3.87/jb (equ.1) and 00^+ ions are low-spin (S=0) because of the strong octahedral crystal field.
^so
= ^48(3+1) ■ ,ub = 3,87lib
(1)
iUso: /is:
spin-only moment Bohr magneton
The temperature dependence of the susceptibility (fig. 6) shows Curie-Weiss behavior (equ. 2). The evaluation including a diamagnetic correction with Pascal constants /36/ leads to an effective moment of^eff=4,57/iB (equ. 3). The deviation from the spin-only value 3,87 can be attributed satisfactorily to spin-orbit coupling /35/.
X-0: C:
l~ieff
C
^ T-0 Susceptibility
Weiss-Temperature (-109K) Curie Constant (2,607 cm^Kmol"^)
(2)
F^eff
= >/8C-Ub =4,57j.ii;
(3)
effective moment
Most of the sintered Co and Ni containing materials conduct via activated hopping of electronic charges, i.e. small polarons. So one is tempted to assume, that the Arrhenius' like conversion curves (fig. 1), which are proportional to the temperature dependence of the oxidation rate, are kinetically determined by the balance of hopping charges. Activation energies for hopping conduction and conversion rates are of the same order of magnitude (0,6 eV).
However, a simple inspection of the conductivity and activity data of LaNiOs reveals, that there is no direct correlation between electrical bulk and catalytical surface properties. This is also supported by the results obtained with Sr- and Ti-substituted LaCoOs. As Sr-sub-stitution increases the electronic conductivity and Ti-substitution decreases it, the enhanced catalytic activity in both cases shows that electronic conductivity does not have any influence on the conversion rate. So one is led to state, that bulk and surface regions are not coupled. If there is still an Arrhenius' like temperature dependence, it must either be caused by a quite different activated step or by a charge transfer between Co2+ and Co3+ ions (Ni2+ and Ni3+ ions respectively) because of a restructured LaCoOs surface. In other words, we conclude, that in all cases the catalytic activity depends on a properly mixed valence state at the surface, as it is given in the case of the spinel C03O4.
850
aco
650 600 550 500 450 4C0 350 3C0 250 2C0 150 100
300 350
T(K1
Fig. 6: Inverse molar susceptibilities.
The low-temperature susceptibility (fig. 7 curve a) goes through a maximum at about 40 K (Neel-temperature), which is consistent with an antiferromagnetic ordering of Co2+ spins /34/. A C03O4 sample exposed to 0,1% SO2 under conversion conditions (45 min, 310°C) has the same Neel-temperature, but shows an increase of the magnetic susceptibility towards lower temperatures (fig. 7 curve b). The absence of a field dependence excludes the presence of ferromagnetic impurities in the exposed sample. The additional component is paramagnetic and most probably surface C0SO4. The
relative amount of C0SO4 in the sample can be roughly estimated to about 0,2%, when the observed Curie constant of the paramagnetic component is compared with the one of C0SO4.
Table 1: Curie-Weiss analysis
0 £
Fig. 7: Low-temperature molar susceptibility of C03O4 (a) and C03O4 exposed to SO2 (b)
Sample	C, [cnr Kniol ' l	0[K]	[Hb]
LaCoO, (1 l50°C/5h)	1.234	-158	3.14
LaCoO, (1050°C/5h)	1,240	-160	3.15
LaCoO, (950°C/5h)	1,229	-158	3,14
LaCoO, (850°C/5h)	1,211	-155	3,1 1
LaCoO, (7S0°C/5h)	1.204	-153	3,10
LaCoO, (650°C/5h)	1,194	-155	3,09
LaTi,,_,Coo.,0,(900°C/2h)	0.888	-54	2,67
LaTi„ ,Co;,,0, (900°C/2h)	0,8,37	-52	2,59
LaTio,Co,,iO,(900'C/2h)	0,762	-51	2,47
Co,0,	2,607	-109	4,57
On the other hand, Co3+ ions in LaCoOs occupy octahedral positions and their spin state changes in a wide temperature range /8, 9, 37/. The magnetic properties are governed by a subtle balance between the crystal field strength and the intra-atomic exchange energy, so that small temperature changes may cause conversion from low-spin to high-spin state and vice versa. The 1/% values of LaCoOs samples calcined at different temperatures are almost identical in the temperature range 150 to 600 K (fig 6). Above 400 K the inverse susceptibility reaches a temperature independent plateau, indicating a metal like phase /9, 38/. From 150 to 400 K the Curie-Weiss law applies, below 100 K severe discrepancies of the differently calcined materials occur. Samples calcined at temperatures above 900°C show a minimum in the 1/x curve, while those calcined below 900°C do not show a minimum. The 1/% curve decreases rapidly towards lowertemperatures and a notable field dependence is observed indicating an additional ferromagnetic phase /9, 38, 39/.
The presence of this ferromagnetism, which dominates the low-temperature susceptibility excludes a direct observation of the paramagnetic surface C0SO4 species in SO2 exposed LaCoOs samples. Curie Constants, C, Weiss temperatures, 0 and effective moments, fjeff, of a Curie-Weiss analysis with diamagnetic correction are given in table 1.
According to a recent reinvestigation of LaCoOs /9/ the following conclusions can be drawn. The calculated effective moment for a 1:1 mixture of low- and high-spin Co3+ ions is3,46/jb, however, smaller values are found experimentally. The reason for this difference could either be due to the presence of less than 50% high-spin Co3+ (^so=4,9 fJB) or to a partial replacement of the high-spin species by Co2+ (^so=3,87/jb) or to both at the same time. The fact that the samples calcined at lower temperatures give the smaller effective moments is consistent with a larger amount of surface which also explains the observed field dependence.
For the catalytic activity of LaCoOs materials the existence of Co2+ at the surface is obviously of great importance. When the materials are calcined at lower temperatures a larger amount of unoxidized Co2+ species is retained causing higher activity (see fig. 3a). At the same time, treating the materials with water causes reduction of Co3+ to thus, again enhancing the catalytic activity (see fig. 3b).
In the attempt to create 00^+ ions in LaCoOa we substituted Ti'^+ for 00^+ ions. The inverse susceptibilities of three samples LaTixCoi^xOs (x = 0,2, 0,4, 0,5) are also shown in fig. 6. is diamagnetic, hence a reduction of the susceptibility with increasing substitution is expected and is observed (table 1). The 1/x behavior is consistent with a metallic LaCoOs phase coexisting with paramagnetic 00^+ ions. The effective moment can be estimated according to
etf = ^X ■ ,u|o(Co^'-) + (1 ~ 2x) ■ uI,(LaCoOg) ■ Ug
(4)
Insertion of 3,87 for and 3,10 for LaCoOs leads to the values 2,95 for x = 0,2, 2,81 for x = 0,4 and 2,73 for x = 0,5. From chemical analysis we know, that the com-
Table 2: Chemical analysis of Co^oxidation state in LaTh-xCoxOs materials
Pciovski;c	A	\-cra,uc Co oxidalion stale		■ m-: r sui-fiKc oi^ 1 (.aicüiod 1 nnileritiLs [niv'g]
	nomiruU	obsL'l'V'yd		
		calciiicd WO=C	sintered r2.šO-C	
LoCoO, ((i20°r,'21i)	3	2,«6		12
LaCoO, |900°C/2M	3	2,93		
La 11, .Co,/',	2,75	2,7	2.65	111
l-a ri, ,(:o, ,0,	2.57	2,43 i		1 1
LiiTi„,Cu„.0,		2,.58 1 2.22 '		1.5
Lari„,Co„;Ü,		! 2,62 i 1	2,06	17
position given by x is not really achieved by a heat treatment at 900°C (table 2). So we understand why the applied estimation gives too high moments, but the trend is well reproduced. The strong increase and the field dependence of x below 100 K again indicate the presence of at the surface.
DISCUSSION
Although XPS spectra show a Co deficit, when calcination is carried out at low temperatures, XPS cannot differentiate between Co2+ and 00^+ ions, So one needs additional arguments for the decision, which of the Co ions or possibly both together create catalytical activity atthe surface. Because of the stoichiometry and spinel type structure of C03O4 both ions are present to a high extent in the bulk and at the surface, whereas in LaCoOs only Co3+ ions exist in the bulk. In a well crystallized perovskite lattice Co2+/Co'^+ ion pairs can be created by temperature dependent disproportion of Co3+, but to a much smaller amount.
Under the real conditions of preparation some unre-acted Co2+ portions may remain, for example, when nitrates decompose, but do not react completely. Another possibility is given by a reducing treatment with aqueous solutions or by decomposition at too high temperatures leaving an oxygen deficiency. It was recently proved experimentally, that Co2+ ions are responsible for catalytical activity in C03O4, which was prepared with different Co2+ content /40/. As the spinel is the much better CO catalyst than the perovskite, we suppose that both ions need to be present for a fast reaction, operating via charge transfer after a first O2 adsorption step.
The implementation of Co2+ ions in LaCoOa by substitution of Ti4+ ions atthe calcination-temperature 900°C enhances both the Co2+ content and the activity (table 2 and fig. 6). Thus we believe, that in all LaCoOs materials there is always a kind of mixed-valence surface phase present. The lower the calcination temperature the more mixed-valence states are formed. Moreover, the range of activation energy for electronic hopping process (0,4-0,7eV) fits well to the activation energy of the catalytical reaction (0,5-0,7eV).
CONCLUSIONS
From the experimental results of this work the following conclusions can be drawn. For any Co and Ni containing spinel or perovskite type material to be a good CO burning catalyst, it must contain mixed valences of 3d-ions at the surface, to facilitate a fast charge transfer when O2/CO adsorption takes place.
It can be stated, that the catalytic activity is mainly influenced by the ratio of Co3+ and Co2+ at the surface of the material. A surface treatment of LaCoOs can be done by a partial reduction of already calcined products with aqueous solutions or calcination at low enough temperatures to get also large surface area. The limiting case of any improvement would be similar to the C03O4 surface itself.
Although the activation energy of the CO oxidation corresponds to the activation energy of hopping
charges in the bulk, no direct correlation with the DC bulk conductivity is observed. On the other hand, the magnetic susceptibility of native LaCoOs powders, calcined at low temperatures, indicate directly the presence of high-spin Co2+ (mixed valence states).
As anticipated before, SO2 is a very strong inhibitor and blocks the exceptionally fast CO reaction instantly. Most probably it is chemisorbed at oxygen covered Co2+ sites and leads finally to sulfate formation. Decomposition of pure NiS04 and C0SO4 takes place around 800°C, so that the problem of recovery at lower temperatures is evident and all catalysts of this type fail in low temperature applications of SO2 containing waste gases. Quite the same SO2 problem appears with honey combs and impregnated supports, otherwise being as fast as Pd coated supports. The Pd catalyst withstands SO2, but during CO/CO2 reaction it is oxidized to SO3.
REFERENCES
/1/ LG. Tejuca. U.LE.FIerro, J.M.D.Tascon, Adv. Cat. 36 (1989) 237.
/2/ T. Seiyama, Catal. Rev. Sei. Eng. 34 (1992) 281-425.
/3/ C.M. Chiu, Y.H. Chang (Journal Paper) Thin Solid Films, 342 1-2 (1999) 15-19.
/4/ R. Sorita, T. Kawano, Sensors & Actuators B-Chemical, B36 1-3 (1996) 274-7.
/5/ R.J.H. Voorhoeve, D.W. Johnson Jr., J.P. Remeika, P.K. Gallagher, Science 195 (1977) 827.
/6/ Y.F.Yu Yao, J. Catal. 36 (1975) 266.
/7/ D.Y.Rao, D.K. Chakrabarty, Ind. J. Chem. 23A (1984) 375.
/8/ P.M. Raccah, J.B. Goodenough, Phys. Rev. 155 (1967) 932.
/9/ M.A. Senaris-Rodriguez. J.B. Goodenough, J. Solid State Chem. 116 (1995) 224.
/10/ R.J.v. Verhoeve. Advanced Materials in Catalysis, ed. by J.J. Burton, R.L. Garten, Academic Press, N.Y. 1977.
/11/ T. Shimizu, Chem. Lett. (1980) 1.
/12/ T. Arakawa, A. Yoshida, J. Shiokawa, Mat. Res. Bull. 15 (1980) 347.
/13/ O. Parkash et al., Mat. Res. Bull. 9 (1974) 1173.
/14/ A. Macher, K. Reichmann, O. Fruhwirth. K. Gatterer, G.W. Herzog, Informacije MIDEM 26 (1996) 79.
/15/ N. Katsarakls, O. Fruhwirth, W. Sitte, Fourth Euro Ceramics, ed. by G. Gusmano, E. Traversa, Faenza, Vol.5 (1995) 89.
/16/ O. Fruhwirth, A. Macher, K. Reichmann, H.G. Schuster, Third Euro Ceramics, ed. by P.Duran, J.F.Fernandez, Faenza, Vol.2 (1993) 395.
/17/ I. Blnder-Begsteiger, G.W. Herzog, E. Megla, M. Tomann-Rosos, Chem. Ing. Tech. 62 (1990) 60.
/18/ M.Lesnik, Thesis, Technische Universität Graz 1994.
/19/ K.Tabata, I.Matsumoto, S.Kohlki, J. Mat. Science 22 (1987) 3037.
/20/ T.Nitadorl, T.lchikl, M.MIsono, Bull. Chem. Soc. Jpn 61 (1988) 621.
/21/ E.A.Lombardo, K.Tanaka, I.ToyoshIma, J. Catalysis 80 (1983) 340.
/22/ V.G.BhIde, D.S.Rajorla, G.R.Rao, C.N.R.Rao, Phys. Rev. B 6 (1972) 1021.
/23/ C.N.R.Rao, O.M.Parkash, P.Ganguly, J. Solid State Chem. 15 (1975) 186.
/24/ G.Thornton, B.C.Totield, D.E.Williams, Solid State Comm. 44 (1982) 1213.
/25/ D.Bahadur, O.M.Parkash, J.Solid State Chem. 46 (1983) 197.
/26/ N.Ramadass, J.Gopalakrishnan, M.V.Sastri, J. Less Common
Metals 65 (1979) 129. /27/ A.Macher, Thesis, Technische Universität Graz, 1994. /28/ K.Koumoto, H.Yaganida, J. Am. Ghem. Sog. 64 (1981) 156. /29/ J.A.Tareen, A.Malecki et al., Mater. Res. Bull. 19 (1984) 989. /30/ J.B.Goodenough, J. Appl. Phys. 37 (1966) 1415. /31/ K.Sreedhar, J.M.Honig et al.. Physical Review B, 46 (1992) 6382.
/32/ N.Katsarakis, Thesis, Technische Universität Graz, 1996.
/33/ P.Gossee, J. Inorg, Nucl. Ghem. 8 (1958) 483.
/34/ W.L.Roth, J. Phys. Ghem. Solids 25 (1964) 1.
/35/ P. Gossee, A.E. van Arkel, J. Phys. Ghem. Solids 15 (1960) 1.
/36/ E.A. Boudreaux, L.N, Mulay, Theory and Applications of
Molecular Paramagnetism, J, Wiley, N.Y. 1976. /37/ G.H.Jonker, J. Appl. Phys. 37 (1966) 1424. /38/ K.Asai, et al.. Physical Review B 50 (1994) 3025. /39/ N. Menyuk, K.Dwight, P.M. Raccah, J. Phys. Ghem. Solids 28 (1967) 549,
/40/ G. Oliva, L. Forni, L. Formaro, Appl. Spectroscopy 50 (1996) 1395.
Gerhard W. Herzog, Angelika Reichmann, Klaus Reichmann, Martina Ruplitsch-Lesnik, Institut für Chemische Technologie anorganischer Stoffe, Technische Universität Graz Stremayrgasse 161 III, A-8010 Graz Tel.: ++43 316 873 8294 Fax.: ++43 316 873 8292 E-mail: f537mach@mbox.tu-graz.ac.at
Karl Gatterer
Institut für Physikalische und Theoretische Chemie Technische Universität Graz Technikerstrasse 4, A-8010 Graz Tel.: ++43 316 873 8239 Fax: ++43 316 873 8225 E-mail: gatterer@ptc.tu-graz.ac.at
ACKNOWLEDGEMENT
The authors are indebted to FRAUENTAL KERAMIK, Frauenthal Austria, for Stimulation and financial support of this work. The low-temperature magnetic equipment was provided by the Austrian Science Foundation (project number: P10713-CHE).
Prispelo (Arrived): 23.01.00
Sprejeto (Accepted): 27.03.00