X-ray photoelectron spectra of perovskite-type cobalt oxides La1−xSrxCoO3−y (x=0.4, 0.6)

Physica C 160 (1989) 252-258 North-Holland, Amsterdam

X-RAY PHOTOELECTRON SPECTRA OF PEROVSKITE-TYPE La1 _&Co03_, (x=0.4, 0.6)

COBALT OXIDES

A.E. BOCQUET ‘, P. CHALKER b, J.F. DOBSON a, P.C. HEALY a, S. MYHRA a,b and J.G. THOMPSON ’ a Division of Science and Technology, Griffith University, Nathan, QLD 4111, Australia ’ Materials Development Division, AERE - Harwell, Oxon, UK ’ Research School of Chemistry, Australian National University, ACT 2601, Australia Received 17 July 1989

X-ray photoelectron spectra have been collected for bulk samples of the ternary perovskite-type cobalt oxides La, _$r,CoO,_, (x=0.4,0.6). Anomalously low binding energy components for the 0 1s and Sr 3d photopeaks, similar to those recorded for the family of high temperature superconducting copper oxides, have been observed. These observations have been interpreted in terms of an initial-state screening mechanism and the implications for the electronic structure of “metallic” oxides are discussed. The Co 2p signature is typical of low spin Co’+, suggesting that doping by strontium leads to the formation of holes in the oxygen bands.

1. Introduction X-ray photoelectron spectroscopy (XPS) is a surface sensitive technique used by many workers to obtain information on the surface chemical states and occupied electronic states of high-temperature superconducting (HTSC) materials. Both the LnBa2Cu307_-x (l-2-3) system [ 1,2] and the Bi-Sr-CaCu-0 (BSCCO) systems [ 3-51 have been studied, and certain general trends have been observed. These include a well resolved low binding energy (BE) peak in the 0 1s region at ca. 528.5 eV [ 4,6] and significant low BE contributions to the alkaline earth corelevel peaks (Ca 2p, Sr 3d and Ba 3d) [ 3,571. Angular resolved studies [EL lo] have shown that the low BE components ‘are characteristic of the bulk electronic structure of the HTSC phases, while contributions found at “normal” oxide binding energies are surface specific. In our previous studies [ 3 1, we have discussed the possible origins of the low BE alkaline earth contributions in the context of final-state shakedown and initial-state screening mechanisms. Resolution of the question as to whether these features are related to the superconducting properties of HTSC oxides, or are merely characteristic of a 0921-4534/89/$03.50 0 Elsevier Science Publishers ( North-Holland Physics Publishing Division )

broader range of metallic oxides, can be approached through a study of oxides which are metallic, but not superconducting. Here, we report the results of XPS studies on bulk samples of the ternary perovskite La, _-xSrxCo03_-y for x=0.4 and 0.6.

2. Experimental procedures Specimens were prepared by the direct reaction of > 99.9% pure La,O,, SrC03 and Co30, in the molar ratios 9:12:10 (x=0.4) and 6:18:10 (x=0.6) at 1400 and 1300’ C, respectively. The final anneal was done in air at 1100” C. Specimens were characterized by X-ray diffraction (XRD) and Scanning Electron Microscopy and Energy Dispersive X-ray Spectroscopy (SEM/EDS) . Resistivity measurements were recorded down to 4.2 K using the four terminal method. XPS measurements were carried out with a VG ESCALAB II instrument using combined XPSI AES/SAM capabilities (AES=Auger Electron Spectroscopy, SAM = Scanning Auger Microscopy). Some spectra were also obtained with a PHI 560 with similar capabilities [ 111. Fresh surfaces were prepared by abrasion in a dry glove box under an argon atB.V.

A.E. Bocquet et al. IPerovskite-type cobalt oxides Lal_,Sr,Co03_,

253

mosphere. The specimens were then transferred in argon with a purpose-built vessel to the UHV environment of the spectrometer. The exciting X-ray radiation was an Al K, source run at 300 W. The base vacuum ranged from 5 x 10m6to 5 x 1Oe8 Pa. Detailed scans were made over the major XPS regions (Co2p, Ols, Sr3d, La3dand Cls), as wellas the major X-ray induced Auger (XAES ) excitations (Co LMM and La MNN), using a constant analyzer pass energy (CAE) of 25 eV. The surface charge shifting was negligible as determined by the C 1s position of graphitic carbon corresponding to a binding energy of 285.0 eV.

3. Results and discussion The overall compositional trends, as determined by EDS and XPS quantification, were in accord with the expected stoichiometries. The refined unit cell parameters from XRD were: &.,&,4CooJ_; Rhombohedral a=7.666( 1) A, cu=90.347(3)“; La&$r0.6Co03_-y: Cubic a= 7.663( 1) 8, (cf. literature values for LaCo03: a= 7.651 A, (~~90.65’ [ 12 ] ). Thus for these compounds only minor structural differences exist (with a slight rhombohedral distortion from cubic symmetry observable for the x= 0.4 specimen vis a vis x=0.6). The phase and grain development were investigated by routine SEM and typical microstructures are shown in fig. 1. The two micrographs show similar morphologies for x=0.4 and 0.6 for the major phases. However, in the case of the x=0.6 sample, there is a higher density of small grains (bright contrast) which may charge up when exposed to the SEM electron beam. Analysis by EDS suggests that these may be a La rich oxide minor phase, but no clear identification could be made from XRD analysis. Measurements of electrical resistance (R) show that at room temperature, both La&&.&oOJ and La0.4C00.603_-ywere good conductors with R values of 7 m&2and 30mQ respectively. Measurement of R as a function of temperature showed that La&!&.&oO3_-y has a metallic temperature coefficient, while LaO.&&oO~_Y showed apparent semiconductive behaviour. The results of these measurements are shown in fig. 2. It has been noted in previous work [ 12,13 ] that as the semi-conductive

Fig. 1. SEM micrographs of L~&x,,~COO~_, (bottom) Lao.4Sr0.6Co03_-y(top). The bars are of 10 pm length.

and

LaCoO, is doped with Sr, the resistivity at room temperature rapidly decreases with a minimum value observed for x=0.5. it has been argued that the introduction of strontium into the LaCoO, lattice results in the formation of “impurity bands” thus decreasing the resistivity. However, the results obtained from SEM and EDS for the La&&&,Coo~_-y compound indicate the presence of minor phase regions. Thus the semiconductive character of the resistivity curves for this compound may be due either to the presence of minor phases regions at the grain boundaries between the major phase grains and thus the creation of semiconducting “weak links”; or to a temperature dependent effect on the carrier mobility

254

A.E. Bocquet et al. / Perovskite-type cobalt oxides La,_,SrXCoOJ_Y

‘:L’ 0

40

80

120

160

200

240

280

T W

Fig. 2. The electrical resistivity as a function of temperature for La,_,.&r0.4CoOg_,(right-hand axis) and La,&&$o03_, (lefthand axis ) . Table I Binding Energies (eV), Kinetic Energies (eV ) and modified Auger parameters ff* for La, _~r,CoO,_,

01s

Co 2Pw2

Co LMM CP

531.4 (58.8%) 528.8 (41.2%)

780.5 772.4 1552.8

533.5 (8.7%) 531.2 (50.0%) 528.5 (41.3%)

780.0 772.2 1552.8

SrN/2

133.8 (28.5%) 131.8 (71.5%)

134.2 (23.6%) 131.8 (76.4%)

b W/2

836.8 833.4

837.7 833.4

such as ionized impurity scattering. It is therefore likely that, for the present specimens, the major phase regions are “metallic” for both x=0.4 and 0.6. Within experimental uncertainties, only slight differences can be observed between the XPS signatures of La,&&.&oO~_, and La0$r0.&o03_, Binding energies (BE) for the XPS peaks, kinetic energies (KE) for the XAES peaks and the modified Auger parameter, cy*= [BE(XPS)+KE(XAES)] [ 141, are presented in table I. All BE values are referenced to the graphitic C 1s peak. Results for a selection of HTSC compounds are presented in table II, as an aid to discussion. Peak signatures of the 0 1s and Sr 3d levels for the perovskite-type cobalt oxides

are found to be similar to those found for the HTSC compounds. In both systems, several distinct contributions can be observed in these regions. These envelopes were “stripped” using subtraction of Gaussian peaks by a standard software package, to resolve them into high BE and low BE components. For HTSC compounds, the 0 1s envelope can usually be resolved into two components, a low BE peak at ca. 528 eV and a high BE peak at ca. 53 1.0 eV. For sintered samples, the relative contribution of the latter peak is related to differences in fabrication route and surface preparation, and is found to diminish upon scraping in UHV [ 6,151. For single crystal specimens [ 7 1, very little emission is found for this peak with the dominant feature invariably being the low BE peak. Furthermore, angular resolved studies have shown that the high BE contribution is surface specific. These results have led us and many other investigators to attribute the high BE component of the 0 1s envelope to extrinsic effects (such as contamination, presence of minor phases and/or environmental degradation) and to conclude that the low BE contribution is characteristic of oxide anions within the bulk material. Significant low BE contributions have also been observed for the alkaline earth species of HTSC compounds. Contributions to the peak envelope with binding energies well below those found for the binary alkaline earth oxides, and even below those of the alkaline earth metals themselves, have been observed for the Ba 3d and Ba 4d levels of l-2-3 compounds [ 1,2,8] and more recently for the Sr 3d and Ca 2p levels of the BSCCO phases [ 3,7]. Again, angular resolved studies indicate that these contributions are characteristic of the bulk HTSC phases. The Sr 3d and 0 1s envelopes for the La, _Sr,Co03_, cobalt oxides are shown in figs. 3a and 3b, respectively. Our results, along with those reported in the literature [ 16- 18 1, show similar low BE contributions in these regions. Low BE contributions have also been observed in the 0 1s region of La, _xCexCo03_,, [ 193. Peak “stripping” of the 0 1s region reveals a low BE component at ca. 528.5 eV contributing ca. 41% to the total envelope, and a high BE contribution at ca. 531.3 eV (fig. 4). By comparison, these data differ significantly from the known cobalt oxides, Co0 and Co,O,, which show low and high BE 0 1s components at 529.4 and

A.E. Bocquetet al. /Perovskite-typecobaltoxides Lal_xSrxCo03_Y

255

Table II Binding Energies (eV) and Kinetic Energies (eV) for selected HTSC compounds and relevant metals and metal oxides Compound

0 1s

YBa#h,O,_, low BE

531.2 528.7

BI&~CUO, low BE

531.4 529.4 (21%)

133.5 131.7 (50%)

Bi2Sr2CaCu08 low BE

531.2 529.4 (28%)

133.8 132.2 (54%)

TlzBa2CaCuOs low BE

531.3 528.8 (18%)

Sr metal SlQ

530.3

Ca metal CaO

529.9

Ba metal BaO

528.4

Sr 3d,/z

Ca 2~~2

Ba 3d,/z

Ref.

719.6 777.8

1121

[31

131

346.6 345.2 142%) = 345.0

779.7 777.4 (5%)

134.2 135.1

53 1.5 eV, respectively [ 20,2 11, or the lanthanum and strontium oxides, Laz03 and SrO, which have 0 1s peaks at 529.0 and 531.5eV [22,23]. The Sr3d envelope can also resolve into two spin-orbit split doublets, with a low BE 3d5,* component at 131.8 eV (73%) and a high BE 3ds12 component at 134.0 eV (fig. 5 ) . As in the case of the HTSC compounds, these low BE components occur at binding energies 2-3 eV below those observed for other related compounds, such as the ATiOs (A = Ca, Sr, Ba) perovskites [ 241, in which the alkaline earths are in known divalent configurations. Fig. 3c shows the Co 2p region of La, _$r$oO,_, The observed features of the envelope are typical of those found for spectra of a wide range of diamagnetic cobalt (III) complexes [25,26] and there is little evidence for the presence of paramagnetic Co2+ in these samples. The 2p,,, binding energies are similar to those observed previously [ 16- 18 ] and are somewhat higher than the values found for Co304 (779.6 eV [ 20,21]), but within the range found for octahedral cobalt (III) compounds where the observed binding energies are dependent on the ligand field strength; higher BE correlating with the more ionic ligands (for example: [Co(S2CN(C,H,),),], 779.0eV [26]; [Co(en),]C13, 780.2eV [25];

131

[261

[261

345.7 347.1 779.1 778.9

[261

780.5 eV [26] and [Co (sepulchrate ) ] Cls, [CO(NH,)~]C~,, 781.5eV [24]). The BE of these present doped compounds is also slightly higher than that observed for the undoped end member, LaCo03 (779.6eV [27]). The La 3d region for the La1_$r,CoO,_, cobalt oxides is shown in fig. 3d. The spectra are consistent with those previously reported [ 161. The prominent satellites in the La 3d region have been attributed to an 0 2p to La 4f shake-up transition, and are identical to those observed in a wide range of LaM03 (M = transition metal) perovskites [ 28,291 and also for La$uO, [ 301. In summary, the 0 1s and Sr 3d photo-peaks in these two complexes show anomalously low BE contributions whereas the La 3d and Co2p peaks are quite typical of trivalent ions in a range of compounds, although the Co 2p BE is significantly higher than in the case of either non-complex oxides or the undoped LaCoO,. In La, _,.Sr,CoO,_, doping by substitution of S?+ into L..a3+sites will result in charge imbalances which can be compensated for either by the oxidation of co3+ to co4+ or by the formation of oxygen holes. Analogous odservations have been made for the HTSC l-2-3 compounds. In this latter system, the

256

5 s

A.E. Bocquet et al. / Perovskite-type cobalt oxides Lal_,SrxCo03_y

k

5 L

1

I

140

135

sr

130

BlndinQ

EnerQy

3d

535

125

Binding

(et’)

530 Enar~y

525 (eV)

I

790

780 Binding

Energy

770

760

(ev)

570

550

550 Binding

Energy

540

830

&?a

(ev)

Fig. 3. XPS envelopes for: (a) Sr 3d, (b) 0 Is, (c) Co 2p and (d) La 3d for La&r0.4Cou3_-y (bottom) and La,,,4Sr0.6CoOg_,(top).

observation of a large shake-up satellite structure in the Cu 2p region [ 1,2] suggests that the predominant valence state is Cu*+ ( d9). Doping of these compounds gives rise to either a mixed 2 + / 3 + valence state for copper or holes in the oxygen valence bands. The Co4+ (d’) ion, however, is not readily stabilized and would be expected to be paramagnetic, giving rise to a shake-up satellite. Since none is observed for the present compounds it is a plau-

sible conclusion that Co4+ is not formed and that Sr doping results in the formation of holes associated with the oxygen sub-lattice. This may also account for the small increase in the binding energies for the Co 2p levels. Thus we conclude that the metallic conduction of the La-Sr-Co oxides is due to p-type conduction in the hybridized Co (3d)-O( 2p) bands, by analogy with the Cu( 3d)-O(2p) hybridization for the l-2-3 compounds.

A.,?. Bocquetet al. /Perovskite-type cobaltoxidesLa,_,Sr,Co4_,

Fig. 4. The results of stripping the 0 Is envelope of ~.6Sr,JkO~ showing high BE and low BE components.

Fig. 5. The results of stripping the Sr 3d envelope of La,&&,CoOs showing high BE and low BE components.

In the context of the HTSC compounds, the low BE 0 1s feature has been interpreted as arising from an initial-state screening effect, due to the metallic properties arising from hybridization of the Cu ( 3d)0( 2p) bands [ 31. Valence band spectra of the ternary cobalt oxides [ 3 1 ] also indicate the presence of hybridized Co( 3d)-O( 2p) bands, and a similar screening mechanism may be operative for these compounds. For the alkaline earth species, however, any initial-state screening mechanism requires “extra” electronic charge to reside on or near the alkaline earth

257

ion in order to account for the binding energies observed in the XPS spectra. This is contrary to most simple band structure calculations for the HTSC compounds which predict “normal” divalent states for the alkaline earth cations [ 32 1. Despite these theoretical predictions, however, there now exists a body of experimental evidence [ 3 ] to favour an initial-state screening mechanism rather than the alternative explanation based on a shake-down mechanism. For example, recent high-resolution X-ray diffraction data [ 331 suggest anomalously large integrated electron densities near the Ba ions in the I2-3 compounds. Furthermore, low BE anomalies, similar to those found in the core-level XPS signatures, have been observed in soft X-ray emission from HTSC compounds [ 341. If an initial state screening mechanism were, in fact, considered to be responsible for the low BE contributions, then it follows that the alkaline earth layers in these systems are likely to be relatively rich in itinerant and/or weakly bound electrons. In this context, it should be noted that the XPS spectra of the Aurivillius phases, which are related structurally to the HTSC compounds but which are neither superconducting nor metallic, do not exhibit any of the anomalous binding energy features discussed above [351. It is not possible at this stage to determine the detailed cause and effect relationships between these anomalous XPS features and the mechanisms which account for metallic conduction and superconductivity. However, the fact that they are observed for the present non-superconducting metallic oxides as well as the HTSC compounds does suggest that they are likely to be general phenomena associated with the metallic rather than the superconducting properties of these oxides.

Acknowledgements

This work was supported in part by the Harwell Underlying Research Programme. One of us (S.M. ) wishes to acknowledge support and hospitality extended during a period of attachment to the Harwell Laboratory. Some of the XPS work was carried out at the Brisbane Surface Analysis Facility with the assistance of Mr. B. Wood. The assistance of the CSIRO

258

A.E. Bocquet et al. /Perovskite-type cobalt oxides La,_,Sr,Co03_,

Division of Applied Physics in characterizing the compounds is gratefully acknowledged.

References [ 11P.C. Healy, S. Myhra, J.C. Riviere, A.M. Stewart and J.G. Thompson, Philos. Mag. Lett. 58 (1988) 139.

[ 21 S. Myhra, J.C. Riviere, A.M. Stewart and P.C. Healy, Z. Phys. B 72 (1988) 413.

[ 3 ] A.E. Bocquet, J.F. Dobson, P.C. Healy, S. Myhra and J.G. Thompson, Phys. Stat. Solidi B 152 ( 1989) in press; J.F. Dobson, A.E. Bocquet, P.C. Healy, S. Myhra, A.M. Stewart and J.G. Thompson, Aust. J. Phys., in press. [ 41 P. Steiner, S. Hiifner, A. Jungmann, S. Junk, V. Kinsinger, I. Sander, W.R. Thiele, N. Backes and C. Politis, Physica C 156 (1988) 213. [ 51 H.M. Meyer III, D.M. Hill, J.H. Weaver, D.L. Nelson and CF. Gallo, Phys. Rev. B 38 (1988) 7144. [6] Z.-X. Shen, P.A.P. Lindberg, I. Lindau, W.E. Spicer, C.B. Eom and T.H. Geballe, Phys. Rev. B 38 (1988) 7 152. [ 71 S. Kohiki, T. Wada, S. Kawashima, H. Takagi, S. Uchida and S. Tanaka, Phys. Rev. B 38 (1988) 7051; S. Kohiki, T. Wada, S Kawashima, H. Takagi, S. Uchida and S. Tanaka, Phys. Rev. B 38 ( 1988) 8868. [S] J. Halbritter, P. Walk, H.-J. Mathes, B. Haeuser and H. Rogalla, Z. Phys. B 73 (1988) 277. [ 9 ] X.D. Wu, A. Inam, M.S. Hedge, T. Venkatesan, C.C. Chang, E.W. Chase, B. Wilkens and J.M. Tarascon, Phys. Rev. B 38 (1988) 9307. [lo] F. Stucki, P. Briiesch and T. Baumann, Physica C 156 (1988) 461. [ 111 PC. Healy, S. Myhra and A.M. Stewart, Philos. Mag. B 58 (1988) 257. [ 121 H. Ohbayashi, T. Kudo and T. Gejo, Jap. J. Appl. Phys. 13 (1974) c-17. [ 13 ] P.M. Raccah and J.B. Goodenough, J. Appl. Phys. 39 ( 1968) 1209. [ 14 ] D. Briggs and M.P. Seah, eds., Practical Surface Analysis by Auger and X-ray Photoelectron Spectroscopy (Wiley, New York, 1983) and references therein. [ 15 ] E.G. Michel, J. Alvarez, M.C. Asenio, R. Miranda, J. Ibaiiez, G. Peral, J.L. Vicent, F. Garcia, E. Moran and M.A. AlarioFrance, Phys. Rev. B 38 (1988) 5146.

[ 161 K. Tabata, I. Matsumoto and S. Kohiki, J. Mat. Sci. 22 (1987) 1882. [ 171 K. Tabata and S. Kohiki, J. Mat. Sci. Letts. 6 (1987) 1030. [ 181 S. Kohiki, S. Ozaki, T. Hamada and K. Tabata, Appl. Surf. Sci. 28 (1987) 85. [ 191 K. Tabata, I. Matsumoto and S. Kohiki, J. Mat. Sci. 22 (1987) 3037. [20] T.J. Chuang, C.R. Brundle and D.W. Rice, Surf. Sci. 59 (1976) 413. [21] B.A. Sexton, A.E. Hughes and T.W. Tumey, J. Catal. 97 (1986) 390. [22] V.I. Nefedov (N.G. Shartse trans.), X-ray Photoelectron Spectroscopy of Solid Surfaces, (VSP BV, Utrecht, 1988) and references therein. [23] H. Van Doveren and J.A.T. Verhoeven, J. Elec. Spec. Rel. Phenom. 21 (1980) 265. [24] S. Myhra, H.E. Bishop and J.C. Riviere, Surf. Tech. 19 (1983) 161; C.N. SayersandN.R. Armstrong, Surf. Sci. 77 (1978) 301. [ 25 ] Y. Okamoto, H. Nakano, T. Imanaka and S. Teranishi, Bull. Chem. Sot. Jpn. 48 (1975) 1163. [26] A.A. Archilleos, L.R. Gahan, T.W. Hambley, P.C. Healy and D.W. Weedon, Inorg. Chem. Acta 157 (1989) 209. (271 B.W. Veal and D.J. Lam, J. Appl. Phys. 49 (1978) 1461. [ 281 P. Burroughs, A. Hamnett, A.F. Orchard and G. Thornton, J. Chem. Sot. Dalton Trans. (1976) 1686. [29]D.J.Lam,B.W.VealandD.E.Ellis,Phys.Rev.B22(1980) 5730. [ 30) B. Viswanathan, S. Madhavan and C.S. Swamy, Phys. Stat. SolidiB 133 (1986) 629. [ 311 K. Bolwin, W. Schnumberger and G. Schiller, Z. Phys. B 72 (1988) 203. [32] For example, J. Yu, S. Massida, A.J. Freeman and D.D. Koelling, Phys. Lett. A 122 (1987) 198; J. Yu, A.J. Freeman and J.-H. Xu, Phys. Rev. Lett. 58 (1987) 1035; B. Szpunar and V.H. Smith, Phys. Rev. B 37 (1988) 7525; M.S. Hybertson and L.F. Mattheiss, Phys. Rev. Lett. 60 (1988) 1661; H. Krakauer and W. Pickett, Phys. Rev. Lett. 60 (1988) 1665. [33] E.N. Maslen, unpublished data (1989). [ 341 R.S. Crisp, unpublished data (1989). [35] A.E. Bocquet, J.F. Dobson, P.C. Healy, S. Myhra and J.G. Thompson, Int. J. Mod. Phys. B 3 (1989) 455.