~ ) Solid State Communications, Vol. 88, No. 5, pp. 349-353, 1993. Printed in Great Britain.

0038-1098/93 $6.00+. O0 Pergamon Press Ltd

INFRARED REFLECTION STUDY OF "FWO.DIMENSIONAL STRUCTURAL PHASE TRANSITION IN STOICHIOMETRIC Pr2NiO4

Ricardo P. S. M. Lobo, Christine Allan(~on, Krystoff Dembinski, Philippe Odier and Fran(~ois Gervais

Centre de Recherches sur la Physique des Hautes Tempdratures, CNRS 45071 Orleans CEDEX 2, France

Received 16 April 1993 by J. Joffrin Revised form 18 August 1993

The structural phase transition at 117 K of stoichiometric Pr2NiO 4 single crystal has been studied under infrared polarised light, ranging from 50 to 5000 cm -1. Our reflectance spectra, obtained between 77 K and 290 K, allow us to observe the first order character of the Bmab-P42/ncm phase transition. Moreover, it is shown that this transition does not have a group-subgroup relation between the phases. Comparing the spectra on different polarisation directions, we have been able to conclude for a two-dimensional character of this phase transition.

1. INTRODUCTION

The different compounds of the Ln2MO4+ ~ family - Ln being a rare-earth (La, Pr, Nd) and M a transition metal (Cu, Ni, Co) have attracted considerable interest in the seven past years. The main reason is the discovery by Bednorz and M011er 1 of high-T c superconductivity in copper-lanthanum compounds. In addition, there are numerous and various phase transitions (structural, magnetic, electronic) that can be observed in these different compounds 2-4 and particularly in Pr2NiO 4 (5=0). 5-7 Although the presence of high-T c superconductivity has highlighted the Ln2CuO4+~-type compounds, the possibility of obtaining large single crystals of nickelates due to their congruent melting makes the study of this branch of the Ln2MO4+ 8 family easier. For example, complete phonon dispersion curves have been obtained in La2NiO 4 prior to copper compounds. 8 This is not the only reason to compare nickelates and cuprates. The non-stoichiometric compounds show conducting properties. However, the understanding of the "normal' conducting state above T c is still very incomplete and highly controversial. The electronic state of the nickel is basically d 8 in those compound whereas copper is in the d 9 state, hence semiconducting properties of nickel compounds with thermally activated charge carriers9,10 are to be compared to "metallic" behavior in the copper family. In conducting compounds, phonons are partially screened by electronic excitations such as the plasmon. It appears interesting, therefore, to study unscreened phonons in the stoichiometric compound prior to investigating the corresponding conducting oxides.

349

From the structural point of view, one expects, for the Ln2NiO 4 compounds, a tetragonal K2NiF 4- type structure (space group 14/mmm). Actually, at room temperature and for the stoichiometric compound, an orthorhombic (Bmab) phase occurs. Lowering the temperature, a structural phase transition to another tetragonal (P42/ncm) phase is attained. A more detailed discussion of this family structure can be found, for instance, in Ref. 5. Nevertheless, the Pccn symmetry has been reported as an alternative to either the Bmab 11,12 or to the P42/ncm 13 structures. The orthorhombic phase is observed in the stoichiometric materials but, depending on the level of the oxygen non stoichiometry, a transition to the tetragonal 14/mmm structure may occur.

The stoichiometric Pr2NiO4+ 8 presents a room temperature orthorhombic phase with lattice

o

parameters a = 5.410 A, b = 5.582 .~ and c = o

12.235 A. At about 120 K there is a phase transition to the low temperature tetragonal phase with lattice

o o

parameters a -- b = 5.492 A and c = 12.157 A. 7 At higher temperatures, a first order structural phase transition from the orthorhombic to the 14/mmm tetragonal phase has been observed at about 690 K. 14 Materials with 5= 0.02 have been found to present a coexistence of the orthorhombic and low temperature tetragonal phases, while at 5 = 0.06 only the latter is observed up to 300 K. 12 Infrared measurements also show a strong dependence of the sample conductivity on the oxygen non stoichiometry. The stoichiometric compound has an insulator behaviour while the non-stoichiometric one shows a conductive character. 15

350 INFRARED REFLECTION STUDY

In this work, the low-temperature structural phase transition of a stoichiometric (insulator) Pr2NiO 4 single crystal is studied by infrared reflectance between 77 K and 290 K.

2. EXPERIMENTAL

A Pr2NiO4+ ~ single crystal has been obtained via the floating zone method with a CO 2 laser. The ratio [Pr]/[Ni] is very close to 2 but a subsequent reduction process at 470K during 10 hours under an Argon flow is required to set 8 = 0. In order to avoid breaking of the crystal, it has been kept in an Araldite resin for polishing and measurements.

The infrared spectra have been obtained with a Bruker IFS 307 Fourier Spectrometer and the liquid- nitrogen cryostat is equipped with polyethylene or thallium bromoiodine windows (same materials for polarisers substrates) in order to cover the regions up to and above c.a. 300 cm -1, respectively. The temperature control has been attained with accuracy of 0.5 K by means of a CLTS probe. Measurements have been made between 77 K and 290 K within the spectral range 50-5,000 cm -1. Spectra show a constant reflectivity level above 1,000 cm -1 .

3. RESULTS

In order to fit reflectance data, we have used the factored form of the dielectric function 16

1.0

w~--"~" ~ ~- ->--" >- 0 . 0 0 ' 5 - ~ 1 i

" O.C

O:o 0 200

I 400

Vol. 88, No. 5

~'1"£

600 800 1000 FRESI.IENCY (cm-1)

Fig. 1. Temperature dependence of reflectance infrared spectra within the ab plane of stoichiometric Pr2NiO 4 obtained heating the sample from 77 K to 290 K. Solid and dashed lines represent experimental data and best fits respectively.

experimental data and the dashed ones are the best fits using equations (1) and (3). In this figure it is clear that a phase transition occurs between 121 K and 126 K. Note the presence of some bands that exist only in either the low temperature or room

~2 _¢o 2 e(to) = ] - [ jLO +iyjLOt°

,£= a'ji K)~TO - f02 + i"/jTO(0 (1)

temperature phases such as, for instance, the mode at 500 cm -1 (room temperature) and the modes at 250 cm -1 (low temperature).

We present, in Fig. 2, the reflectance spectra obtained (and best fits as in Fig. 1) with light polarised along the c axis of the material at room

where ~'jLO and ~~iT O are the longitudinal and transverse optical modes of the/th oscillator and "/jLO and ~jTO their dampings, respectively. E~ is the dielectric constant at high frequencies. The oscillator strength (Aej), which represents the contribution of each oscillator to high the dielectric constant from e~

to % (the static dielectric constant), is not an adjustable parameter but may be deduced from TO- LO splitting via

temperature and at 77 K. Apart small bands in the low temperature spectrum at 480 and 670 cm -1, both spectra present the same phonons in both orthorhombic and tetragonal phases.

Figure 3 details the temperature dependence of the reflectance due to the oscillators present in the spectral range of 500-750 cm -1, upon cooling (light polarised in the ab plane). It is remarkable the existence of three bands between 122 K and 113 K while outside this region only two modes are

2 f~2 [ - i (nkLO - j~o) e~ k (2)

A£j = ~2 2 .iTO 1-I (n~TO - njrO)

k~j

The reflectance spectrum R(co) near normal

observable in each phase. Figures 4 and 5 present the thermal behaviour

of the adjusted frequencies, damping terms and oscillator strengths for the modes observed between 600 and 700 cm 1. Symbols are fitting parameters and the lines are guides for the eye. For the sake of clarity only the data obtained upon heating the

incidence is related to the dielectric function via

. f~_ ] 2 R - ,j~ + 1 (3)

In Fig. 1, typical infrared spectra within the ab plane obtained upon heating the sample from 77 K to 290 K are shown. Solid lines represent

sample have been shown in these figures. In this region, also, one can see an oscillator with ~iTO ~ 615 cm 1 which exists only in the tetragonal phase while another mode having &'~jmo ~ 640 cm -1 is present only in the orthorhombic phase. From our results we can determine a transition temperature of 123 K upon heating the sample and 117 K on cooling runs, yielding a thermal hysteresis of 6 K for this phase transformation.

Vol. 88, No. 5 INFRARED REFLECTION STUDY 351

1.0

0.5 F- i,i

0.0

0.0 0 I I

200 400

EI/'£

600

77 K

800 1000 FREQUENCY (cm-1)

Fig. 2. Experimental data (solid lines) and best fits (dashed lines) for infrared reflectance along the c axis of stoichiometric Pr2NiO 4 at liquid nitrogen and room temperature.

4. DISCUSSION

The jumps observed for the frequencies shown in Fig. 4, specially in the highest frequency mode, together with the divergence present in the temperature behaviour of the damping terms (Fig. 5 (a)), the "anomalies" of oscillator strengths (Fig. 5 (b)), and the thermal hysteresis of 6 K, indicate the first order character of this transition. Moreover,

observing Figs. 1 and 4, we can see that some bands are present only in the low temperature region while others appear only in the room temperature phase. For example, it is to be emphasised that Fig. 4 clearly shows the discontinuous character of the phase transition with a restricted temperature range in which modes of both phases coexist but do not merge ! These effects allow us to conclude that this transition does not have a group-subgroup relation between the phases. As the Pccn structure is a subgroup of the both Bmab and P42/ncm, it can be attributed neither to the room temperature phase nor to the low temperature one, at least in stoichiometric Pr2NiO 4 single crystal. On the other hand, the absence of a group-subgroup relation between Bmab and P4~/ncm is compatible with the discontinuous transition shown by the infrared spectra. Thus, we are led to conclude for a Bmab structure in the orthorhombic room temperature phase and a P4Jncm in the low temperature tetragonal phase, discarding the possibility of a Pccn

7OO

680

62O

(cm-'@

Fig. 3. Thermal evolution of bands relative to the NiO 2 vibrations within the plane ab, upon cooling the sample from 290 K down to 77 K.

TOO

• -t, (11) ~

E(I) o.....,,o.._......o.~+., e.

Fig. 4. Thermal behaviour of longitudinal and transverse optical modes fitted frequencies for NiO 2 vibrations in the ab plane. (I), (11) and (111) represent each pair of TO (circles) and LO (squares) modes. Lines are guide to the eye.

352

4O

2O

10 SO

0.4

0.3

02

0.1

ol SO

INFRARED REFLECTION STUDY

! i i I

(a) l~c. l m I Ln I ".

100 '50 ' 200 2SO 300 TEHPERATURE (K)

! i i i

(b) +

",..

~-op_ t ""~-

I I I I

100 150 200 250 100 TEHPERATURE (K)

Fig. 5. (a) Damping terms fitted for NiO 2 vibrations (I, II and III in Fig. 4) on the plane ab and (b) oscillator strengths for their transverse optical modes. In both panels, the same symbols have been adopted for each mode. Only the data concerning the heating of the sample is shown.

symmetry in our sample. Although no group- subgroup relation exists between P4Jncm and Bmab, both structures are subgroups of the parent structure 14/mmm.

Due to the very near a and b lattice parameters, our sample is polydomain due to 90 °- twinning and spectra correspond to a mixed (a,b) spectrum. Measurements at 300 K with light polarised along some directions within the ab plane have allowed us to estimate a difference of about 5 cm °1 between the frequencies of these modes. No further attempt to work in polarised light have been tried afterwards.

Four E u modes are expected in the parent phase 14/mmm. De Andres et aL expect from group theory 8 B2u + 5 B3u (13 altogether) in the (a,b) plane of the room temperature Bmab phase. They also

Vol . 88 , N o . 5

predict 13 E u modes in the low temperature P42/ncm. Looking at Figs. 3 and 4, one observes two modes in the 500-750 cm -1 region. The same number of modes is observed below the phase transition but at a different frequency. This is consistent with group theory expectations and our phase assignments. As proposed by Fernandez et al., 17 the splitting of the bands in the low- temperature phase with respect to the high- temperature parent phase 14/mmm is due to the loss of equivalence of the four oxygen atoms in the NiO 2 equatorial plane, caused by a change in the symmetry axis direction of the Ni 2+ point group (C2h).

Finally, a very interesting result may be deduced from the comparison of Figs. 1 and 2. In Fig. 1 the structural phase transition of this system is observed via the modifications of the vibration modes observed in the ab plane. On the other hand, in Fig. 2, apart the two small modes at 480 and 670 cm -1 in the low temperature spectrum (which may be regarded as contributions from other polarisation), the spectra at 77 K and 300 K present basically the same phonons. This giant difference in the thermal behaviour between the spectra obtained with light polarised perpendicular and along the tetragonal axis allow us to conclude that the atomic relative displacements leading to this phase transition occur only in the ab plane, i.e., this phase transition has a strict two dimensional character.

5. CONCLUSIONS

The analysis of a single crystal of Pr2NiO 4 under infrared polarised light between 77 and 290 K has allowed us to verify the first order character of the low temperature structural phase transition without a group-subgroup relation between the phases. Results support the P42/ncm symmetry for the lowest temperature phase and Bmab to the room temperature one. Jumps in the frequencies of the vibration modes and a small thermal hysteresis of 6 K have been observed around an average transition temperature of 120 K. In our spectra, not only is a characteristic splitting of vibrations within the NiO 2 plane in the low temperature phase observed, but also we have been able to separate vibrations due to the orthorhombic difference between the lattice parameters a and b. Comparison of the spectra obtained along the c axis and within the ab plane allowed us to conclude for a two-dimensional character of this phase transition.

Acknowledgements- We are grateful to Dr. P. Simon for many helpful discussions and pertinent comments. One of us (RPSML) acknowledges the Brazilian agency CNPq for fellowship support towards this work.

Vol. 88, No. 5 INFRARED REFLECTION STUDY

REFERENCES

353

1. J.G. Bednorz & K. A. MQIler, Z. Phys. B 64, 189 (1986).

2. P. Gopalan, M. W. McEIfresh, Z. Kakol, J. Spalec & J. M. Honig, Phys. Rev. B 45, 249 (1992).

3. J. B. Torrance, P. Lacorre, A. I. Nazzal, E. J. Ansaldo & Ch. Niedermeyer, Phys. Rev. B 45, 8209 (1992).

4. See, for example, J. B. Torrance, Ph. Lucorre, Ch. Asavavoengchai & R. Metzger, Physica C 182, 351 (1991).

5. J. L. Martinez, M. T. Fernandez-Diaz, J. Rodriguez-Carvajal & P. Odier, Phys. Rev. B 43, 13766 (1991 ).

6. A. de Andr6s, J. L. Martinez & P. Odier, Phys. Rev. B 45, 12821 (1992).

7. M.T. Fernandez-Diaz, J. Rodriguez-Carvajal, J. L. Martinez, G. Fillion, F. Fernandez & R. Saez- Puche, Z Phys. B 82, 275 (1991).

8. L. Pintschovius, J. M. Bassat, P. Odier, F. Gervais, G. Chevrier, W. Reichart & F. Gompf, Phys. Rev. B 40, 2229 (1989).

9. J. M. Bassat, P. Odier & F. Gervais, Phys. Rev. B 35, 7126 (1987).

10. F. Gervais, P. Echegut, J. M. Bassat & P. Odier, Phys. Rev. B 37, 9364 (1988).

11. J. D. Sullivan, D. J. Buttrey, D. E. Cox & J. Hriljac, J. Solid State Chem. 94, 337 (1991).

12. D. J. Buttrey, J. D. Sullivan, G. Shirane & K. Yamada, Phys. Rev B 42, 3944 (1990).

13. J. Rodriguez-Carvajal, M. T. Fernandez-Diaz & J. L. Martinez, J. Phys. • Condens. Matter 3, 3215 (1991 ).

14. M. T. Fernandez-Diaz, J. Rodriguez-Carvajal & J. L. Martinez, Proc. Congress React. Solids, Madrid (1992).

15. F. Gervais, R. P. S. M. Lobo, C. AIlan~:on, J. M. Bassat, J. P. Loup & P. Odier, Solid St. Commun. (in press).

16. F. Gervais, in Infrared and Millimeter Waves, Vol. 8, Ed. K. J. Button, Academic Press, pp. 279-339 (1983).

17. F. Fernandez, R. Saez-Puche, M. T. Fernandez- Diaz, J. Rodriguez-Carvajal, J. L. Martinez, I. L. Botto & E. J. Baran, Eur. J. Inorg. Solid State Chem. T 28, 510 (1991).