Platinum Metals Rev., 1993, 37, (2), 86
The Modification of Superconductors
Platinum Group Metals Confer Notable Properties
Since the discovery of the first high temperature superconductor, YBaCuO, many attempts have been made to improve its properties by the addition of modifiers; indeed almost all of the elements in the Periodic Table have already been tried in this way. The present survey, covering both our own work and the published literature, examines the effects of adding the platinum group metals to the superconductor, resulting in the formation of several new compounds, many of which are semiconducting and some of which may have potential value in engineering and chemical technology. The complex oxides formed are characterised and data on the superconducting properties of the modified YBaCuO oxide phases are presented.
Immediately after the discovery of high temperature superconductors (1) a series of papers appeared discussing their modification by other metals; now almost all the elements of the Periodic Table have been used as modifiers. The aims of this research were as follows: to produce a modification-induced increase in the critical temperature (Tc), to verify the hypothesis on the mechanism of superconductivity by substituting atoms of one type for those of another, to find the effect of additions to the initial mixture on the size of the particles formed, and to increase the rate of formation of the main phase, for example, YBa2Cu3O7-δ.
The high reactivity of the components in the superconductor cause it to react with the container material, forming new phases which affect the properties of the superconductor. Such phenomena, observed for the YBa2Cu3O7-δ + platinum metals system, were discussed recently (2). In this review we present data from the literature and from our own work on the interaction of YBa2Cu3O7-δ with all die platinum group metals; and new phases containing platinum metals are characterised.
The YBa2Cu3O7-δ + Ruthenium System (3)
Compounds of the platinum group metals are known to be catalytically active for redox reactions, with ruthenium exhibiting the highest activity towards inorganic ions (4). In the absence of ruthenium, the synthesis of YBa2Cu3O7-δ usually takes place at 920-950°C and requires about 60 hours (5). However, the Y-Ba-Cu-Ru-O system is characterised by two peculiarities: a shortening of the synthesis time of YBa2Cu3O7-δ, by a factor of approximately six, and a decrease in the reaction temperature to 880°C, which can be explained by assuming that ruthenium acts as a catalyst for the synthesis. It should be noted that an increase in the duration of the synthesis causes degradation of the superconducting properties; thus, sample No. 15, whose composition is presented in Table I, required 10 hours to sinter, whereas sample No. 16 took 26 hours.
From Table I it may be seen that certain specific ratios of barium, ruthenium and copper are required to produce materials having a high superconducting transition temperature, Tc, (see sample Nos. 2,4, 6–9 and 11–15). The appearance of a superconducting phase with a lower Tc is characteristic of samples Nos. 1, 3, 5, 10 and 16. Samples containing more than 0.5 moles of ruthenium per mole of yttrium loose their superconductivity.
Thermal cycling of samples with high Tc results in hysteresis for practically all of these sintered samples after the first cycle (293–77 K and 77–180 K), the Tc being 94 K on cooling and 107 K on heating. No hysteresis arises during further cycling; Tc first increases, then gradually decreases (see Figure 1.) before becoming stable at 94 K after the 17th cycle. Storing the samples in liquid nitrogen (6), as well as thermal cycling (7), is known to increase Tc, which may be caused by some micro- or macrostructural phase transformation up to the quasi-equilibrium state.
Samples with Tc higher than the liquid nitrogen temperature were studied by X-ray phase analysis. Samples containing 0.1-0.7 moles of ruthenium (per mole of yttrium) appeared to be two phase (Figure 2a), the YBa2Cu3O7-δ phase was superconducting, while the other ruthenium-containing phase exhibited semiconducting behaviour. In the sample with a ruthenium content of 0.8 moles only the semiconducting phase was found to be present. Unit cell parameters are listed in Table II. When the ruthenium content was above 0.9 moles (per mole of yttrium) the sample was multiphase. However, the samples are superconducting, see Table III when the ratio of the intensities of the main peaks on the X-ray diffraction pattern of the sample under study of the superconducting, I sc:I nsc non-superconducting (semiconducting) phases, is above 0.485. This provides evidence for the existence of a “percolation limit” in the superconducting phase contents, below which superconductivity is not observed. This may be interpreted in terms of the infinite three-dimensional cluster (8) of Josephson junctions (9).
Fig. 2 (a)
|d,nm||I, per cent||d,nm||I, per cent|
Using X-ray microanalysis the composition of the semiconducting rumemum-containing phase, for all samples with ruthenium contents less than 0.9 moles was found to be Y2Ba5CuRu2Ox. But in the sample with ruthenium content of 0.8 moles, only this phase was found to be present. The compound is brown in colour and its melting point was found by thermal gravimetric analysis to be 1090°C. The valence states of ruthenium and the other elements were examined by X-ray photoelectron spectroscopy (XPS). The peaks of the elements contained in the Y-Ba-Cu-Ru-Ox system are shown in the wide-scan spectrum, see Figure 3, where data for all the systems under investigation are summarised.
The oxidation state of the ruthenium on the surface of Y-Ba-Cu-Ru-O samples can be evaluated from the dependence of Eb(Ru3d 5/2), where Eb is the binding energy, on n (10), where n is the coefficient in RuOn. However, before the localisation of the RU3d 5/2, and the Ru3d line intensities, it is necessary to distinguish between the contribution from the Ru3d 5/2 and Cls lines. The carbon may result from surface contamination, residual gases in the vacuum system, non-reacting part of the starting powders, and so on, which overlap in the XPS spectrum for Y-Ba-Cu-Ru-O samples, Figure 4. For this purpose the intensity of a non-overlapping line, Ru3p, and the ratio Ru3p :Ru3d have been determined for RuO2 powder. The Ru3p :Ru3d ratio measured coincided with the value calculated to an accuracy of 3 per cent (11). In the case of RuO2, it was not difficult to estimate the separate Ru3d and C1s contributions, since the most intensive peak, Ru3d 5/2, did not overlap with the Cls spectrum, Figure 4.
The Ru3d 5/2 peak position (284.7 eV) for two Y-Ba-Cu-Ru-O samples provides evidence for the high oxidation state of the ruthenium atoms. The value for Eb(Ru3d 5/2) is 283.2–283.6 eV for RuO2, depending on the calibration method (10). Such high values of Eb(Ru3d 5/2) were obtained in the Y-Ba-Cu-Ru-O samples for which the values of Eb(O1s ) and Eb(Cu2p 3/2) were usual, that is, coincident with those for YBa2Cu3O7-δ (12–14). Consequently, the oxidation state of ruthenium in die compositions under consideration is high, Ru8+ being most probable. Therefore, the final composition of the ruthenium phase maybe represented as Y2Ba5CuRu2O17 (3, 15), a compound of sufficient stability under high vacuum.
In Table IV the bulk and surface compositions of the samples are given, the surface composition being determined from the following formula:
where [Cx] is the concentration (in atomic per cent) of element x, I is the integral intensity of the analytical lines and s is the photoionisation cross-section. The summation was performed along the analytical lines for the metals. A comparison of the bulk and surface compositions showed that the surface of all the samples in the Y-Ba-Cu-M-O system, where M is a platinum group metal, was barium enriched, as in the case of YBa2Cu3O7-δ. (The Ba3d 5/2 line of relatively low kinetic energy, about 470 eV, was taken to be an analytical one). The coefficient, γ, where γx = [Cx]sur/[Cx]bulk, is presented in Table IV for the platinum group metals which, in our opinion, are the most interesting. (Values of [Cx]sur were obtained by XPS, while values of [Cx]bulk were obtained by chemical analyses).
The wide-scan X-ray photoelectron spectrum of the sample is shown on Figure 3
The elements of the platinum group may be divided into three sub-groups: ruthenium and iridium, with γ much greater than 1, constitute the first sub-group; the second sub-group contains platinum, rhodium and osmium, while palladium is the only representative in the third sub-group with γ = 0, even at very high palladium concentrations.
Considerable enrichment of the surface by a modifying agent may take place when it is not incorporated into the lattice of the main crystal phase. When YBa2Cu3O7-δ is the main phase, the presence of a platinum group metal, M, does not decrease the Tc value, as shown by the measurements of the temperature dependence of the magnetic susceptibility, and as observed for the ruthenium containing samples.
The YBa2Cu3O7-δ + Rhodium System
Like ruthenium, the other platinum group metals are catalysts for the synthesis of YBa2Cu3O7-δ, which is shown by a faster reaction, the absence of a “green” phase, (Y2BaCuO5) and a decrease in the reaction temperature.
The YBa2Cu3O7-δ + rhodium system is very similar to that containing ruthenium, that is, its Tc is dependent on the ratio of the components. The presence of only one rhodium-containing phase, Y3Ba10CuRh3Ox, is characteristic of the system (X-ray microanalysis data). The results of the XPS study of this compound are presented below, its wide-scan spectrum being shown in Figure 3.
The Rh3d5/2 peak, which is the most frequently used analytically for rhodium compounds, is easy to distinguish in the XPS spectrum of the Y-Ba-Cu-Rh-O composition. Its position at 309.4 eV is evidence for rhodium being in the 3+ oxidation state in the oxide matrix (16). It should be noted that the “typical” value of Eb(Rh3d 5/2) for the Rh3+ complexes is 310.7 eV. However, the value of 308.9 eV has been reported elsewhere for the Rh2O3 oxide (17).
Thus, the formal oxidation state of the rhodium modifying agent in the Y-Ba-Cu-Rh-O samples is 3+, however, the positive charge on the rhodium atom in the sample studied is noticeably higher than for Rh2O3. These data allowed the refined formula of the rhodium phase to be determined as Y3Ba10CuRh3O20.
The YBa2Cu3O7-δ + Palladium System
The modification of the YBa2Cu3O7-δ system by palladium of more than 0.001 atomic per cent results in the complete loss of superconductivity (18, 19). In subsequent literature palladium has been shown to substitute for copper in the superconductor lattice, forming YBa2Cu2.5Pd0.5O7 (20). The absence of palladium on the surface of the Y-Ba-Cu-Pd-O samples with high palladium contents is evidence of this (Table IV, XPS data), the copper content being somewhat increased (when compared with the other samples containing the same amount of modifying agent M).
Of the samples in the Y-Ba-Cu-Pd-O system studied, only No. 8 had sufficient surface concentration of palladium to differentiate the Pd3d 5/2 peak from the XPS background noise (Table IV). The Eb(Pd3d 5/2) value of 336.6 eV is evidence that the palladium oxidation state in the under-surface layer is 2+. For PdO the value of Eb(Pd3d 5/2) is 336.5 eV (21), which is almost the same as obtained in our measurements.
Nevertheless, it seems unreasonable to expect the oxidation state of the promoting element to be the same as in sample No. 8.
By means of X-ray microanalysis three new phases were detected in the YBa2Cu3Pd2O8.5 sample annealed at 900°C.
The YBa2Cu3O7-δ + Osmium System
This system was studied as thoroughly as that of ruthenium. A series of YBa2Cu3OsxOy, samples (with x = 0.1–1.0) was synthesised according to the procedure described in (3).
All the samples, are multiphase compounds. In samples with x = 0.1–0.8 two superconducting phases are distinguishable, one having Tc = 93 K, while the Tc values of the other phase may vary from 62 K to 83 K (Figure 5). Samples with x = 0.9 and x = 1.0 are not superconducting. The XPS method was used to determine the oxidation state of the osmium (see Figure 3 for the wide-scan spectrum). Since there is no information in the literature on the dependence of the most intensive osmium peak, 4f 7/2, on the osmium oxidation state, our data in Figure 6 may be of interest. The Eb(Os 4f 7/2) values were found to be 53.5 and 54.3 eV for samples with x = 0.7 and 1.0, respectively, which correspond to Os4+ (x = 0.7) and Os5+ (x = 1.0). The osmium 4f line shape (a larger half-width) is evidence for the presence of states other than the main states of the osmium atoms in the sample under study (Figure 7).
X-ray phase analysis showed that samples with x = 0.1-0.8 consisted of the main phase: YBa2Cu3O7-δ (Pmmm, a = 0.3882, b = 0.3891, c = 1.1688 nm for the sample with x = 0.7), and an admixture phase: BaOsO3 (In3m, a = 0.935 nm) (see Figure 8). Copper oxide and unidentified compounds with diffraction peaks at d = 0.4833 and 0.4187 nm were also detected among the admixtures.
The data from XPS and X-ray phase analysis may lead to the conclusion that the compound with four-valent osmium, BaOsO3, is formed in samples with x = 0.1–0.8, see Table V.
|d, nm||l, %||d, nm||l, %|
The non-superconducting samples, with x = 0.9 and 1.0, consist of two unknown osmium phases (as the main components) and a small amount of YBa2Cu3O7-δ, Y2BaCuO5, BaCuO2 and Y2Cu2O5. The unknown phases were identified as Y2Ba3Cu2OsOx and YBa3Cu4Os0.1Ox by X-ray microanalysis. Since the oxidation state of osmium in samples with x = 0.9 and 1.0 is 5+ (XPS data), the phases may be presented as Y2Ba3Cu2OsO10.5 and YBa3Cu4Os0.1O8-δ. Attempts to synthesise the phase Y2Ba3OsO10.5 in its pure form were a failure and gave a mixture of compounds, such as Y2BaCuO5 (Pbnm; a = 0.7131, b = 1.2164, c = 0.5657 nm), BaCuO2 (Im3m, a = 1.8294 nm), and Y2Cu2O5. The other diffraction peaks in the XRD pattern of this sample were attributed to the Y2Ba3Cu2Os(V)O10.5 phase (Table VI). The synthesis of the other osmium phase, YBa3Cu4Os0.1O8-δ, was not a success either. X-ray phase analysis showed the presence of YBa2Cu3O7-δ, BaCuO2 and Y2BaCuO5. The XRD data on the fourth phase are given in Table VI. They were thought to belong to YBa3Cu4Os0.1(V)O8+δ.
The YBa2Cu3O7-δ + Iridium System
This system is a close analog to the ruthenium and rhodium systems. The dependence of Tc on the ratio of the initial elements was also similar to that of the ruthenium system (see Table I). X-ray phase analysis and X-ray microanalysis detected only one iridium compound in the system, which appeared to be Y3Ba6CuIr6Ox, with a face centred cubic F-cell (a = 0.8341 nm).
According to the XPS data the Eb(Ir4f 7/2) value for the Y-Ba-Cu-Ir-O sample is 63.2 eV. It should be noted that the Eb(Ir4f 7/2) values obtained for K3IrCl6 and K2lrCl6 complexes are 62.6 and 63.7 eV, respectively, (16) and the Eb(Ir4f 7/2) value for the main state of the Ir/Al2O3 catalyst stored in air is 62.5 eV (25), and it was thought that the main iridium state on the catalyst surface was Ir4+ (25). Therefore it may be expected that the Ir4+ oxidation state is highly probable in the Y-Ba-Cu-Ir-O system, at least in the undersurface layer. It is also noteworthy that the Ir4f line of the sample is rather narrow; consequendy iridium is included in the composition of only one compound on the surface, which agrees with the X-ray microanalysis data. A high coefficient of surface enrichment, γ, with iridium may lead to the supposition that the size of the iridium containing particles is smaller than that of the main phase present, namely YBa2Cu3O7-δ.
Thus, from the above data the iridium compound can be represented as Y3Ba6CuIr6O23.5.
The YBa2Cu3O7-δ + Platinum System
This system has been studied in the greatest detail, and was considered in (2), and results of the XPS study are described here.
The main state of platinum in samples of the Y-Ba-Cu-Pt-O system was found to be Pt4+, since measurements gave Eb(Pt4f 7/2) = 74.4 eV, which is closer to 74.5 eV for PtO2 than to 73.7 eV for PtO (26). However, considerable shoulders were seen on the Pt4f spectrum in the region of higher binding energies (Figure 9). At present it is difficult to find an unequivocal explanation of this phenomenon, although we considered two hypotheses: first, the influence of the Kα3,4 satellite of the Ba4d line and second, the presence of platinum ions in an oxidation state higher than 4+.
A careful examination of the spectral site of samples not containing platinum shows that the Kα3,4 satellite of the Ba4d line does overlap with the Pt4f spectrum. The line shape of the Pt4f spectrum is very dependent on the Pt:Ba ratio, so that at low values of the ratio the height of the Pt4f 5/2 peak may be greater than that of the Pt4f 7/2 peak. It should also be noted that the shape of the Kα3,4 satellite of the Ba4d line is rather smooth, and it seems impossible to explain the shoulder in the spectrum Pt4/ (marked * in Figure 9) only by means of the contribution of this satellite. We now attribute this effect to the presence of Pt6+. However, we have not succeeded in observing this effect on the spectrum of the known Y2Ba3Cu2PtO10 (27), which is evidence for the presence of as yet undetected phases with PtPt6+ in the YBa2Cu3O7-δ + platinum system.
The following generalised conclusions have been made from the data considered here:
Superconductivity in YBa2Cu3O7-δ + M, where M is a platinum group metal, depends on the amount of M in the system, the lowest values being characteristic of systems with palladium, which substitutes for copper in the superconductor structure. The highest amounts of M have been found for the osmium system, barium osmate being formed in this case. All platinum metals form new non-superconducting (semi-conducting) phases, listed in Table VII.
As seen from Table VII, ruthenium exhibits the highest oxidation state of 8+ in this complicated system, the existence of a Pt6+ compound is assumed, the oxidation state of osmium is 5+, and rhodium, iridium and palladium are all in their standard valence states. In addition, ruthenium, rhodium, and iridium form only one compound with the Y-Ba-Cu-O system, while the platinum and palladium systems are the most diverse. The ability to form perfect crystals (particularly iridium, platinum, osmium), their semiconducting properties, ease of synthesis and stability in air and vacuum make these compounds quite possibly suitable for engineering and chemical technology (as catalysts for oxistability in air and vacuum make these compounds quite possibly suitable for engineering and chemical technology (as catalysts for oxidation and burn-off of waste gases, in fuel cells and for semiconducting techniques).
Readers who wish to receive a copy of XRD data on Y3Ba10CuRh3O20, YBaCu5Pd3O10.5, YBa8Cu8Pd3O20, Y6Ba5Cu3Pd3O20 and Y3Ba6CuIr6O23.5 established during these investigations should send their request to Dr. Yu. M. Shul'ga, Institute of Chemical Physics, Chernogolovka, 142432, Moscow Region, Russia.
- 1 J. D. Bednorz and K. A. Muller, Z. Phys. B, Condens. Mater., 1986, 64, 189
- 2 E. F. Maher, Platinum Metals Rev., 1991, 35, ( 1 ), 2
- 3 E. N. Izakovich,, A. T. Mailybaev,, G. D. Sokolovskaya,, Yu. M. Shul’ga,, V. I. Rubtsov,, Yu. M. Korolev,, V. N. Spector and M. L. Khidekel, Pkys. Status Solidi (a), 1991, 124, 525
- 4 B. R. James, “ Homogeneous Hydrogenation ”, Wiley-Intersci. Publ., New York, 1973
- 5 I. E. Graboi,, A. P. Kaul and Yu. G. Metlin, “ Itogi Nauki i Tekhniki. Ser. Khimiya Tverdogo Tela ” (Russ.), 1988, 6, 37
- 6 D. H. Matthew,, A. Bailey,, R. A. Vaile,, G. J. Russell and K. N. Taylor, Nature, 1987, 328, 786
- 7 H. D. Tostarndt,, M. Calffy,, A. Freimuta and D. Wohllenben, Solid State Commun., 1989, 69, 911
- 8 M. Nunez Requeiro,, P. Esquenazi,, M. A. Izbizky,, C. Duran,, D. Castello,, T. Luzuriaga and G. Nieva, Physica, 1988, C1S3, 1016
- 9 S. R. Shenoy, Physica, 1988, B1S2, 72
- 10 C. D. Wagner,, W. H. Riggs,, L. E. Davis,, J. F. Moulder and H. Mullernerg, eds., “ Handbook of X-Ray Photoelectron Spectroscopy ”, Perkin-Elmer Corp., Minnesota, 1979, p. 190
- 11 J. H. Scofield, Livermore Lab. Rep., UCRL51326, 1973
- 12 N. Nuckel,, J. Fink,, B. Renkel,, D. Ewert,, C. Politis,, P. J. W. Weijs and J. C. Fuggle, Z. Phys., 1987, B67, 1
- 13 H. Hiara,, M. Hirabayashi,, N. Terada,, Y. Kimura,, K. Sensaky,, A. Akimoto,, K. Bushida,, F. Kawashima and R. Uzuha, Jpn. J. Appl. Pkys., 1987, 26, L460
- 14 P. Steiner,, V. Kinsinger,, I. Sander,, B. Siegwart,, B. Humer,, C. Politis,, R. Hoppe and H. P. Muller, Z. Phys., 1987, 67, 487
- 15 Yu. M. Shul’ga,, V. I. Rubtsov,, E. N. Izakovich and Yu. G. Borod’ko, Surface (Russ.), 1991, 6, 121
- 16 V. I. Nefedov, “ X-Ray Photoelectron Spectroscopy of Chemical Compounds ” (Russ.), Khimiya, Moscow, 1984, p. 284
- 17 M. N. Firsov,, V. I. Nefedov and E. P. Domashevskaya, J. Struct. Chem. (Russ.), 1979, 20, 49
- 18 Y. Nishii,, S. Moriya and S. Tokunaga, Phys. Lett., A, 1987, 126, 55
- 19 T. J. Wagener,, Y. Gao,, I. M. Vitomirov,, C. M. Aldao,, J. J. Joyce,, C. Capasso and J. H. Weaver, Pkys. Rev . B, 1988, 38, 232
- 20 G. Ferrey,, A. LeBail,, Y. Laligant,, M. Merkieu,, B. Raveau,, A. Suplice and R. Tournier, J. Solid State Chem., 1988, 13, 610
- 21 K. S. Kim,, A. G. Goosmann and N. Winograd, Anal. Chem., 1974, 46, 197
- 22 V. I. Nefedov,, N. M. Spitsyn and Ya. V. Salyn, Koord. Khimiya (Russ.), 1975, 1, 1618
- 23 G. J. Leigh and W. Bremser, J. Chem. Soc., Dalton Trans., 1973, 1216
- 24 A. M. T. Bell, Supercond. Sci. Technol., 1990, 3, 55
- 25 A. V. Bulatov,, T. P. Geidei,, S. N. Ustritskii,, I. Ya. Turaev,, M. L. Khidekel,, Yu. M. Shul’ga and N. M. Yaroshenko, Kinet. Katal. (Russ.), 1988, 29, 238
- 26 K. S. Kim,, N. Winograd and R. E. Davis, J. Am. Chem. Soc., 1971, 93, 6296
- 27 E. A. Genkina,, O. K. Melnikov,, A. B. Bykov and B. A. Maksimov, KristaOografiya, 1989, 34, 1426
- 28 G. Calestani,, C. Rizzoli and G. D. Andreotti, Solid State Commun., 1988, 66, 223
- 29 U. Geiser,, L C. Porter,, H. H. Wang,, J. A. Allen and J. M. Williams, J. Solid State Chem., 1988, 73, 243
- 30 J. S. Swinnea and H. Steinfink, Acta Cryst., 1987, C43, 2436
- 31 N. Toyota,, P. Koorevaar,, J. van der Berg,, P. H. Kes,, T. Shisido,, Y. Saito,, N. Kuroda,, K. Ukei and J. Sasaki, J. Phys. Condens. Matter, 1989, 1, 3721
- 32 Y. Laligant,, G. Ferrey,, M. Hervieu and B. Raveau, Eur. J. Solid State Inorg. Chem., 1988, 25, 111
- 33 P. Gouzerh,, F. Robert,, J. Gouzerh and H. Makram, J. Cryst. Growth, 1990, 102, 1059
- 34 M. Hiorth, Acta Chem. Scand., 1988, 42A, 727
- 35 U. Amador,, E. Moran,, M. Alario-Franco,, A. Vegas,, M. Martinez-Ripoli,, J. Ibanez and J. B. Torrance, Physica C, 1989, C162, 8731
- 36 I. Shishsido,, T. Fukuda,, N. Toyoto,, K. Ukei and T. Sasaki, J. Cryst. Growth, 1987, 85, 599
- 37 Z. Chen,, Y. Qian,, Y. Wu,, D. Sun,, R. Lin,, T. Cheng,, L. Niu,, G. Zhou,, Z. He,, J. Xia,, Y. Zao and Q. Zhang, J. Cryst. Growth, 1989, 94, C277
- 38 H. K. Muller-Buschbaum and A. Teichert, J. Less-CommonMet., 1989, 155, 9
- 39 Y. Qian,, Z. Chen,, T. Chen,, G. Pan,, Q. Zhang,, J. Xia,, Y. Zhao,, Z. He,, G. Zhow,, T. Cheng and L. Niu, Mater. Res. Butt., 1988, 23, 119
- 40 W. Gutau and H. Muller-Buschbaum, Z. Anorg. Attg. Chem., 1990, 584, 7
- 41 W. Gutau and H. Muller-Buschbaum, J. Less-CommonMet., 1990, 159, 223