Materials Chemistry and Physics 77 (2002) 912–917

Magnetic properties of the perovskite compoundsYFe1–xCrxO3 (0.5≤x≤1)

A. Dahmania,b, M. Taibib, M. Noguesc, J. Arideb, E. Loudghiria, A. Belayachia,∗a Laboratoire de Physique des matériaux, Faculté des Sciences Université Mohammed V, B.P. 1014, Rabat, Morocco

b Laboratoire de Physico-chimie des Matériaux, Associé à l’AUF (LAF 502) ENS, B.P. 5118, Takaddoum Rabat, Moroccoc Laboratoire de Magnétisme et d’Optique de l’Université de Versailles (URA 1531), Batiment Fermat,

45 Avenue des Etats Unis, 78035 Versailles Cedex, France

Received 12 October 2001; received in revised form 2 April 2002; accepted 11 April 2002

Abstract

Powder oxides with the formula YFe1–xCrxO3 (0.5 ≤ x ≤ 1) have been studied. The samples have been prepared using the ceramicsolid state reaction in air. The X-ray diffraction spectra indicated that the samples crystallise in the perovskite structure with orthorhombicdeformation. The magnetic properties of these materials are concentration and temperature dependent. Zero-field-cooled and field-cooledmagnetization curves show strong irreversibilities and complex hysteresis loops. The saturation magnetization for the mixed compoundsshows a unusual behaviour. The results are discussed by considering the existence of many weak ferromagnetic components induced bythe random distribution of the transition metal cations in the octahedral site. This leads to frustration interactions and short-range orderthat smooth the phase transition.© 2002 Elsevier Science B.V. All rights reserved.

Keywords: Perovskite; Weak ferromagnetism; Frustration; Random distribution; Substitution; Phase transition

1. Introduction

In the randomly mixed oxides competition of several kindsof magnetic order leads to a great variety of disordered phe-nomena[1–4]. The perovskite compounds with general for-mula LnBO3, where Ln is a rare earth or yttrium and B isa transition metal have attracted increasing interest due totheir relatively simple structure correlated to the large vari-ety of magnetic properties that these compounds can exhibit.These materials have regained much interest in many fieldsof technological applications[5–8].

In the perovskite materials, the magnetic properties aregenerally determined by the B site cations and, thus, alloyingdifferent kinds of cations in this site may induce a varietyof interesting phenomena as has been observed by manyauthors[9].

In contrast to the manganese-based perovskites, LnMnO3that have attracted considerable interest since the discoveryof colossal magnetoresistance in these manganites[10],relatively few investigation has been performed on theorthoferrite–orthochromite mixed perovskites Ln(FeCr)O3.

∗ Corresponding author. Tel.:+212-37-778-973x51;fax: +212-37-778-973.E-mail address: belayach@fsr.ac.ma (A. Belayachi).

Early studies have established that the basic compoundsYFeO3 and YCrO3 are orthorhombically distorted per-ovskites. Below their Nèel temperatures of 648 and 141 K,respectively, the compounds order antiferromagneticallywith G-type structure. Each ion has six antiparallel nearestneighbours. The weak ferromagnetism observed in thesecompounds results from small canting between the sublat-tices[11] which was attributed to the Dzyaloshinsky–Moriyaantisymmetric exchange[12,13].

Wold and Croft[14] have observed solid solubility be-tween YFeO3 and YCrO3 in the YFe1–xCrxO3 system.However, to our knowledge, the only systematic investiga-tion of the magnetic properties of the series YFe1–xCrxO3has been performed by Kadomtseva et al.[15]. The authorsreported that the substitution of Cr3+ ions for Fe3+ onesinduces anomalous in the temperature and concentration de-pendence of the magnetic moment. The results are explainedby two mean weakly ferromagnetic components. The au-thors asserted that the most complicated region for the analy-sis of the magnetic behaviour is concentration nearx = 0.5,where the used impurity–matrix model is completely loss.

In the present paper, we perform an experimental investi-gation of the iron substitution effect on the magnetic prop-erties in the system YFe1–xCrxO3 (0.5 ≤ x ≤ 1). Indeed,it has been expected that strong ferromagnetic exchange

0254-0584/02/$ – see front matter © 2002 Elsevier Science B.V. All rights reserved.PII: S0254-0584(02)00188-8

A. Dahmani et al. / Materials Chemistry and Physics 77 (2002) 912–917 913

interactions can be made, in principal, when cations withhalf-filled eg orbitals interact with cations possessing emptyeg orbitals[16]. This should induce a variety of interestingphenomena.

2. Experimental

Polycrystalline samples YFe1–xCrxO3 (x = 0.5, 0.75 and1) are prepared by solid state reaction in air. Stoichiomet-ric quantities of Y2O3, Fe2O3 and Cr2O3 oxides are mixedand ground in ethanol. The mixture was fired once for 24 hat 900◦C and then reground and heated again for 60 h at1350◦C.

X-ray powder diffraction spectra were recorded at roomtemperature in a D5000 Siemens diffractometer using Cu K�radiation. Static zero-field-cooled (MZFC) and field-cooled(MFC) magnetization measurements were carried out in aSQUID (superconducting quantum interference device), us-ing the following procedures. In ZFC measurements, thesample was cooled from room temperature to 5 K in a zeromagnetic field. A field ofH = 100 Oe was applied at 5 K,and then the isofield magnetization was measured in thewarming cycle. For the FC case, the magnetization mea-surements were performed in the warming cycle after thesample temperature was lowered to 5 K in the presence ofthe same fixed magnetic field as was used for the measure-ments in the ZFC case. The hysteresis loops measurementswere recorded in magnetic fields between –60 and 60 kOe ina SQUID at different temperatures between 4.2 and 300 K.For the compounds withx = 1 and 0.75, the magneticsusceptibility measurements�dc (dc = direct current) wereperformed over the temperature range 300–900 K, using aDMS4 magnetometer under 1 T magnetic applied field whilefor the compound YFe0.5Cr0.5O3 that has complex magneticproperties the susceptibility was performed over the temper-ature range 5–900 K, using a FONER type magnetometer.

3. Results

3.1. X-ray diffraction analysis

The X-ray diffraction analysis has established a solidsolubility between YFeO3 and YCrO3. The compoundscrystallise in the perovskite structure with orthorhombicdeformation, with the most probable space groupD16

2h (Pbnm). The values of the cell parametersa, b and c aresummarised inTable 1.

Table 1Cell parameters determined from X-ray diffraction spectra

Compositionx a (Å) b (Å) c (Å)

0.50 5.253 5.557 7.5460.75 5.244 5.532 7.5431.00 5.236 5.511 7.538

The substitution of Fe3+ for Cr3+ induces an increase ofthe cell parameters in agreement with the difference of ionicradius of Cr3+ and Fe3+ (0.615 and 0.645 Å, respectively).

3.2. ZFC-FC magnetization

The as-obtainedMZFC–MFC curves are as shown inFig. 1.Strong irreversibilities between the two curves are observedfor the three samples studied. Furthermore, the magnetiza-tion versus temperature evolution shows a concentration de-pendence.

• For YCrO3 (x = 1), the FC magnetization follows aBrillouin-like function, whereas, the ZFC curve has a cuspat TN = 144 K. In the rest of the paper, we simply wishto use YCrO3 to demonstrate that in the case of a sim-ple weak ferromagnet the magnetization curves follow aBrillouin function as in the classical ferromagnets.

• For YFe0.25Cr0.75O3, MZFC shows a maximum at about160 K which is identified to the Nèel temperature.MFC hassaturation like value below 30 K and decreases between30 and 150 K above which it drops significantly. The Néeltemperature determined from the jump of the curve isTN = 160 K.

• The ZFC and FC curves do not merge at the Nèel temper-ature, the irreversibilities persist untilTirreversible= 210 Kwell aboveTN. This must indicate that the true paramag-netic behaviour is not reached immediately aboveTN.

• For YFe0.50Cr0.50O3, the behaviour of the ZFC–FC mag-netization is more complex. On one hand,MZFC shows abroad peak at about 80 K. This type of peaks, taken aloneis generally seen as a signature of spin-glass behaviour.On the other hand, below the temperatureTI = 160 K,the FC curve lies above the ZFC one. Whereas, unusu-ally, aboveTI , theMFC curve passes below theMZFC oneand achieves a minimum value at about 210 K.

3.3. Magnetization measurements

A typical hysteresis loops are as shown inFig. 2 for thethree samples studied. The loops are well closed and sym-metric around zero. TheseM–H curves reveal a complexmagnetic behaviour. Above a particular fieldHF, which de-pends on temperature and composition, the magnetizationincreases linearly and continuously with increasing field atall the temperatures. This suggests that the as obtainedM–Hcurves are the sum of two contributions. An hysteresis loopwhich saturates atHF and an antiferromagnetic componentmanifested by the linear increase of the magnetization. So,the high–field part of theM(H) evolution can be representedasM(H) = χAFH + σs, whereχAFH is the antiferromag-netic contribution andσ s is the saturation magnetization ofthe weak ferromagnetism[17]. Theσ s is obtained by the ex-trapolation of the linear part ofM–H curve to zero. The evo-lution ofσ s as a function of temperature is as shown inFig. 3

914 A. Dahmani et al. / Materials Chemistry and Physics 77 (2002) 912–917

Fig. 1. ZFC–FC magnetizations for YFe1−xCrxO3: (a) x = 0.5; (b)x = 0.75; (c) x = 1.

in the case of the mixed compounds (x = 0.5 and 0.75). Thisparameter has a different behaviour for the two compounds.

• For YFe0.5Cr0.5O3, σ s decreases with increasing tempera-ture until reaching a minimum value at about 250 K above

Fig. 2. As-obtained hysteresis loops for YFe1−xCrxO3: x = 0.5; (b)x = 0.75; (c) x = 1 at 5 K.

which it increases slightly. In addition,σ s does not followa Brilloiun function.

• For YFe0.25Cr0.75O3, the saturation magnetization firstlydecreases and goes to zero at about Tcompensate= 50 K.This temperature (50 K) is attributed to a compensation-likeeffect. With further increasing temperature,σ s increasesto achieve a maximum at about 130 K and decreasesagain drastically but does not vanish until 210 K.

A. Dahmani et al. / Materials Chemistry and Physics 77 (2002) 912–917 915

Fig. 3. Saturation magnetization for YFe1−xCrxO3 (x = 0.5, and 0.75)determined from hysteresis loops.

3.4. Susceptibility

For YFe0.25Cr0.75O3 and YCrO3, the magnetic suscepti-bility measurementsχdc were performed in the temperaturerange 300–900 K. A quasi-linear variation of the inverse ofthe susceptibility is observed. This suggests that only a para-magnetic regime occurs over the studied temperature rangein these samples. The data are fitted by the Curie–Weiss lawand the determined Curie parameters (Curie constantC andasymptotic Curie temperatureθp) are reported inTable 2.

For the compound YFe0.5Cr0.5O3 that exhibits a morecomplex magnetic behaviour, the susceptibility measure-ments were performed over the extended temperature range5–900 K in order to obtain maximum of information. Theinverse of the susceptibility as function of temperature is asshown inFig. 4. At very high temperature (T > 300 K),χ−1 behaves quasi-linearly indicating a spin non-correlatedbehaviour. In this temperature range, the data obey to theCurie–Weiss law. The determined Curie parameters are re-ported inTable 2. With decreasing temperature a local min-imum is observed at aboutTmin = 250 K below whichχ−1

increases slightly and reaches a local maximum at aboutTmax = 200 K. When the temperature is further decreased,χ−1 falls in way similar to that observed in weak ferromag-

Table 2Experimental and theoretical Curie constants, the paramagnetic Curietemperatures and the ordering temperatures determined by magnetic mea-surements

Compositionx Cexperimental

(Kemu mol−1)Ctheoretical

(Kemu mol−1)–θp(K) TN (K)

0.50 3.13 3.12 476 210–2500.75 2.69 2.49 322 1601.00 1.91 1.87 292 144

Fig. 4. Inverse of dc magnetic susceptibility vs. temperature forYFe0.5Cr0.5O3.

nets. However, the decrease in this case takes place smoothlyand quasi-linearly in an extended temperature range.

The determined Curie constants for the three compoundsare in agreement with the theoretical ones calculated for thespin only state. The asymptotic Curie temperaturesθp arenegative, suggesting the predominance of antiferromagneticinteractions in the studied materials.

4. Discussions

The X-ray diffraction analysis revealed a complete sol-ubility between YFeO3 and YCrO3 and the substitution ofchromium ions by iron ones results in a decrease of thecell parameters while the structure is preserved. However,the magnetic investigation evidenced strong temperature andconcentration dependence of the magnetic parameters.

In the ZFC–FC magnetization measurements, we haveobserved strong irreversibilities for all samples studied, buttheir temperature variations are very sensitive to the chem-ical composition. In the case of the non-substituted sampleYCrO3, MFC that is proportional to the magnetic moment,follows a Brillouin function, but for the mixed compoundsthe substitution shifts the magnetization away from theabove simple behaviour. This effect is more pronounced inthe case of YFe0.5Cr0.5O3 as this compound is known toexhibit the more anomalous properties[15]. The ZFC–FCevolution present in this case has some features of a spinglass-like behaviour with the presence of peak in theMZFCcurve. However, the hypothesis of spin glass is discardedwith the following arguments:

1. The system shows a branching ofMZFC and MFC at atemperature that is well above the temperature maximumof the peaks.

2. The peak inMZFC is very broad, indicating the pro-gressive nature of the involved process in contrast to

916 A. Dahmani et al. / Materials Chemistry and Physics 77 (2002) 912–917

spin glass-like model[18] characterised in particular bya true thermodynamic phase transition occurring at awell-defined temperature.

We conclude, thus, that the magnetic properties of themixed compound are frustrated. The frustration, in thiscase, is not strong enough to lead to a true spin glass statealthough it is enough to perturb the arrangement of spinsin the material. On the other hand, the unusual behaviourin this compound is manifested by the relative variations ofMZFC and MFC. In fact, aboveTI , MFC lies belowMZFC.Such event suggests the occurrence of different magneticregimes that have different temperature variations, as hasbeen reported for other materials[19].

The complex shape of the obtained hysteresis loopsreveals the existence of antiferromagnetic structures super-imposed to weak ferromagnetism. In addition, the negativevalues of the asymptotic temperatures indicate the pre-dominance of antiferromagnetic interactions. Indeed, in themother compounds YFeO3 and YCrO3 the small cantingbetween the sublattices is responsible of the weak fer-romagnetic that is induced by the Dzyaloshinski–Moriyaantisymetric exchange and according to Moriya[13] thecanting angle, in this case, is constant with temperature.In the mixed compounds, the hypothesis of a simple weakferromagnetism with a constant canting angle although canexplain the occurrence of hysteresis but it cannot justify thenon-Brillouin-like temperature dependence of the saturationmagnetization. To explain these variations we consider theexistence of many weak ferromagnetic components anti-ferromagnetically coupled, having different canting anglesand distributed in the matrix. For simplification we con-sider only two weak ferromagnetic components and use theNéel’s model for ferrimagnetism[20]. So each of the abovetwo weak ferromagnetic components has a magnetizationMi and concentrationNi . As the temperature increases, thetwo spontaneous magnetizations vary differently. The netspontaneous magnetization that results from the algebraicsum ofM1 andM2 might have different behaviours depend-ing on the concentration. The case of YFe0.25Cr0.75O3 isof particular interest, because in this materialσ s(T) goesto zero at a temperature lower than the ordering one. Thisbehaviour is typical of a Néel’s N-type ferrimagnet. At thispoint (called compensation temperature) the moments ofthe two weak ferromagnetic components cancel each otherand the net moment will pass through zero.

In contrast to Kadomtseva et al.[15] who assumed thatthe contribution of the impurity or the matrix predominatesbelow or above the compensation temperature, respectively,we believe that the magnetically predominant componentwill be oriented parallel to the applied field, whereas, thatof the second component will be aligned antiparallel.

In classical materials (ferromagnetic or antiferromag-netic) the transition from the ordered state towards theparamagnetic one is manifested by an abrupt variation inthe magnitude parameters. This is not so in our mixed

compounds YFe0.25Cr0.75O3 and YFe0.5Cr0.5O3. The de-crease of the magnetization (or susceptibility) takes placeprogressively in a large temperature interval. This givesmore evidence that the magnetic properties of these materi-als are not usual but rather are frustrated. This fact is morepronounced in the case ofx = 0.5 compound. In effect, inthis sample a precise determination of the Néel temperatureis not available. The inverse susceptibility versus temper-ature shows a local minimum aroundTmin followed by alocal maximum. A similar behaviour has been reported inthe case of LiCoF4 [21] where the neutron diffraction anal-ysis has given proof that the Nèel temperature correspondsto Tmin. We believe that it is so in our compound. In addi-tion, Tmin corresponds to the minimum observed inσ s(T).However, the occurrence of hysteresis loops up to 290 K isunlikely with this argument. To clarify this problem we haveperformed a Mössbauer spectrum at 300 K. The obtainedresults suggest the paramagnetic behaviour, indicating thatthe Nèel temperature is situated below room temperature.

Furthermore, in the two mixed compoundsMZFC andMFCcurves merge only at a temperature well aboveTN and thesaturation magnetization does not vanish completely atTN.This must indicate that the true paramagnetic regime is notyet reached and that some short range order still exist aboveTN [22]. The occurrence of hysteresis loops at this tempera-tures reveals that the short range order gives rise to magneticclusters[23].

All the above events suggest that the substitution processinduces frustration in the magnetic properties of the mixedperovskites. The presence of frustrated interactions is notsurprising because the occurrence of collinear long-rangemagnetic coupling is not appropriate in structurally disor-dered systems[24]. In this case, the disorder is caused bytopological cationic distribution of the iron and chromiumions (Fe3+, t32ge

2g and Cr3+,t32g), having antagonist character,

in the octahedral B site that is expected to be random, lead-ing to a distribution of the canting angles causing frustrationthat is responsible for the observed non-collinear behaviour.In fact, in solid solutions the substitution of atoms is alwaysin random positions[18]. The hypothesis of random distri-bution is only justified if the solid solution is between twooxides with identical structures and similar lattice parame-ters, containing cations of the same valence and ionic radius[25]. Such conditions are in fact realised in the mixed com-pounds studied. The unusual behaviour induced cationicdistribution in perovskite materials are well established andhave been reported by several authors[26–29].

Frustration can also be induced by competition be-tween the antiferromagnetic interactions Fe3+–O–Fe3+

,

Cr3+–O–Cr3+ and the mixed one Fe3+–O–Cr3+. This laterinteraction, if at all present, is expected to have the maxi-mum probability for YFe0.50Cr0.50O3 as this compound hasan equal number of Fe3+ and Cr3+ ions. In fact, followingGoodenough rules[16] the mixed superexchange interac-tion Fe3+–O–Cr3+ must be ferromagnetic, if the superex-change pathway is perfectly linear. However, Moskovin

A. Dahmani et al. / Materials Chemistry and Physics 77 (2002) 912–917 917

et al. [30] have reported that the superexchange angle is146◦ and that this interaction is antiferromagnetic in thecase of YFe1–xCrxO3. Other investigations are needed toclarify this question.

5. Conclusions

The investigation of the magnetic properties in the per-ovskite oxides YFe1−xCrxO3 (0.5 ≤ x ≤ 1) suggestedstrong temperature and concentration dependence andanomalous behaviour. The unusual magnetic behaviourobserved in the mixed compounds YFe0.5Cr0.5O3 andYFe0.25Cr0.75O3 are attributed to the substitution process.These results are explained by considering that the transi-tion metal ions Fe3+ and Cr3+, having different electronicconfiguration, are randomly distributed in the octahedralsite. This leads to a distribution of the weak ferromagneticcomponents inside the material. Mössbauer effect measure-ments are in progress and they will provide more helpfulinformations about the local cationic distribution in thesecompounds.

References

[1] T. Bruckel, W. Praud, P. Convert, J. Phys. C, Solid State Phys. 2565(1988) 20.

[2] M. Nogues, D. Fiorani, J. Tejeda, J.L. Dormann, S. Sayouri, A.M.Testa, E. Agostinelli, J. Magn. Magn. Mater. 1641 (1992) 104.

[3] P.D. Battle, J.R. Frost, S.H. Kim, J. Mater. Chem. 1003 (1995) 5.[4] K. Tezuka, K. Henmi, Y. Hinatsu, NM. Masaki, J. Solid State Chem.

591 (2000) 154.

[5] L.W. Tai, M.M. Nasrallah, H.U. Anderson, J. Solid State Chem. 117(1995) 118.

[6] A.S. Bhalla, R.Y. Guo, R. Roy, Mater. Res. Innov. 3 (2000) 4.[7] P. Fumagali, C. Speath, G. Güntherodt, IEEE Trans. Magn. 3277

(1995) 31.[8] W. Zhong, W. Chen, W.P. Ding, N. Zhong, A. Hu, Y.W. Du, Q.J.

Yang, J. Magn. Magn. Mater. 112 (1999) 195.[9] P.D. Battle, T.C. Gibb, A.J. Herod, S.H. Kim, P.H. Munns, J. Mater.

Chem. 865 (1995) 5.[10] B. Raveau, C. Martin, A. Maignan, J. Alloys Comp. 461 (1998) 277.[11] B.C. Tofield, B.E.F. Fender, J. Phys. Chem. Solids 2741 (1970) 31.[12] J. Dzyaloshinsky, J. Phys. Chem. Solids 241 (1967) 4.[13] T. Moriya, Phys. Rev. 91 (1960) 120.[14] A. Wold, W. Croft, J. Phys. Chem. 447 (1959) 63.[15] A.M. Kadomtseva, A.S. Mskovin, I.N. Bostrem, B.M. Wanklyn, N.A.

Khafizova, Sov. Phys. JETP 1202 (1977) 45.[16] J.B. Goodenough, Phys. Rev. 564 (1955) 100.[17] Y. Doi, Y. Hinatsu, J. Phys, Condens. Matter 4813 (1999) 11.[18] J.L. Dormann, M. Nogues, J. Phys, Condens. Matter 1223 (1990)

63.[19] A.V. Mahajan, D.C. Johnston, D.R. Torgeson, F. Borsa, Phys. Rev.

B 10966 (1992) 46.[20] L. Néel, Ann. Phys. 137 (1948) 3.[21] P. Lacorre, Ph.D. Thesis, Lemans, 1988.[22] I. Nowik, I. Felner, V.P.S. Awana, J. Magn. Magn. Mater. 67 (1999)

192.[23] S.H. Kim, P.D. Battle, J. Solid State Chem. 174 (1995) 144.[24] K.S. Roh, M.G. Kim, K.S. Ryu, C.H. Yo, Solid State Comm. 565

(1996) 100.[25] J.M.D. Coy, Phys. Rev. B 3240 (1972) 6.[26] A. Belayachi, M. Nogues, J.L. Dormann, M. Taibi, Eur. J. Inorg.

Chem. 1039 (1996) 33.[27] C. Shinzer, J. Alloys Comp. 65 (1999) 288.[28] E. Loudghiri, A. Belayachi, M. Nogues, M. Taibi, A. Dahmani, M.

Elyamani, J. Aride, Morrocan J. Condens. Matter 98 (2000) 3.[29] P.D. Battle, J.G. Gore, J.R. Frost, Phys. Rev. B 51 (1995) 8624.[30] A.S. Moskovin, N.S. Ovanesyan, V.A. Trukhtanov, Hyp. Interact.

265 (1975) 1.