Coupled Reduction with Platinum Group Metals

THE PREPARATION OF HIGH PURITY RARE METALS

By U. Berndt, B. Erdriianii and C. Keller Institute for Radiochemistry, Karlsruhe Nuclear Research Centre

presence of noble metals (platinum, palladium, Very pure metals of the actinide, rhodium, iridium) can be applied to prepare lanthanide and alkaline earth series can very pure noble metal alloys with actinide, be prepared from their oxides by lanthanide and alkaline earth elements, coupled reduction by hydrogen in the titanium, zirconium, hafnium, vanadium, presence of platinum group metals. niobium, tantalum, and even with lithium During the process the platinum group (2-6). metals form intermetallic phases with The direct synthesis of alloy phases of the metals from the oxides. These noble metals with base metals involves many phases are decomposed to yield the pure serious problems caused by the relatively metals in gaseous form and the platinum high vapour pressure of the base metals. In group metals in solid form. The pure addition, the choice of the crucible material gaseous metals condense to produce and of the gas atmosphere affects the results. pure solid specimens. Furthermore, ordered phases are obtained only by repeated homogenising and by pro- longed and careful annealing. Also the usual In the course of our investigations dealing methods of preparation of highly radioactive with phase equilibria in uranium oxide-rare metals are somewhat difficult, and in some earth oxide systems, we made the following cases the elements, such as protoactinium unexpected observations (I) : (islPa) and curium (244Cm),are only prepared (I) In fluorite phases with a high content of in milligram quantities and have not yet been rare earth oxides it is not possible to investigated sufficiently. Further methods for reduce higher valency states of uranium the preparation of intermetallic phases are to tetravalent uranium, at least when the reactions of platinum metals with using hydrogen and at temperatures nitrides (7,8), sulphides (9), borides (IO), below IOOC-IIOO~C. or carbides (II,IZ) of the base metals; the (2) At higher temperatures and when reactions of the carbides with platinum using noble-metal crucibles, we ob- metals have been investigated more served weight changes which indicated thoroughly. However, the results of the a reduction of the uranium to valencies reactions of base metal carbides with platinum below four. The metallic appearance metals do not give single-phase products of the reaction product made a chemical but mixtures, e.g. UMe, and U,MeC, reaction between the substance and the (Me=Ru, Os, Rh, Ir, Pt), or UMe, and crucible very possible. carbon corresponding to the initial metal Detailed studies proved this assumption concentration. The application of these and we were able to demonstrate that the methods to the preparation of intermetallic hydrogen reduction of metal oxides in the phases is further hindered, as the stoichio-

Platinum Metals Rev., 1974, 18, (l), 29-34 29 metric constitution of the carbides or nitrides of hydrogen or ammonia. Reaction mech- must be examined by analyses before starting anisms of coupled reactions in the case of the the reaction. The preparation of the pure refractory oxides with practical reference to oxygen-free carbides and nitrides of the platinum thermocouples have been studied by transuranium elements is also complicated by Darling and his colleagues (15-19). practical problems. In the cases of curium Details of this method of obtaining the very and protoactinium no carbides and nitrides pure noble metal alloy-phases are as follows. have yet been prepared. When heating the oxides with finely powdered From metallurgical practice it is known noble metals in an extremely pure stream that oxides difficult to reduce are more easily of hydrogen, a reaction takes place which we reduced in the presence of nobler metals. have called “coupled reduction”, i.e. the re- This is the case for the preparation of ferrous duction of a metal oxide (carbonate)/fluoride alloys and results from the fact that with in the presence of a noble metal, leading to higher temperatures the oxygen partial the formation of a base metal-noble metal pressure of the oxide is raised by the presence alloy. of the more noble metal. Reduction is then The formula is general for the method used possible if both metals form mixed crystals (in the case of an oxide): and this is explained by the affinity forces xPt -1- Me,O + Hz S PtxMe,+HzO occurring during the formation of the mixed crystals (13). Up to the present time we have been able On this basis it was possible for Klemm to show that coupled reduction is possible and Bronger (14) to prepare some noble when using platinum, palladium, iridium and metal alloy phases by reduction of base metal rhodium as the noble metals. Attempts to oxides with platinum metals in the presence prepare alloy phases with ruthenium, osmium,

Platinum Group Metal-Actinide Alloy Phases Prepared by Coupled Platinum Group Metal-Actinide Alloy Phases Prepared by Coupled Reduction during This Work I Reduction during This Work -

Pu Am 3 U NP I n - ~_-Th Pa U - NP __ Pu %.Am - Cf - __ - Rh Rh,Pu Rh,Am Rh Rh,Pu Rh,Am "n

Rh,Am 33 Rh,U Rh,Np Rh,Pu wa == Rh,Th Rh,Pa Rh,U Rh,Np Rh,Pu Rh,Am 'in Ir Ir,Npa Ir,Pua Ir2Arna Ir Ir,Npa Ir,Pua Ir,Cma Ir,U ... Ir,Pa Ir,U. .- - - - Pd Pd,Np Pd,Pu Pd,Am Pd Pd3NP Pd,Pu Pd,Am Pd,Cm Pd,Th Pd,U Pd,Th I Pd,U Pd,U Pd,U Pt Pt,Pu Pt,Arn Pt Pt,Pu Pt,Am Pt,Cm Pt,Thb Pt,Pa Pt,U Pt,Np Pt,Pu Pt,Thb Pt,Pa Pt,U Pt,Np Pt,Pu Pt,Th Pt,Pa Pt,U Pt,Np Pt,Pu Pt,Arn Pt,Th Pt,Pa Pt,U Pt,Np Pt,Pu Pt,Arn Pt,Cm (Pt,Cf)c a No Ir,A tcompounds exist a No Ir,A compounds exist b Not obtained in pure form b Not obtained in pure form c Only indirect proof c Only indirect proof Compounds in bold type were prepared for the first time Compounds in bold type were prepared for the first time

Platinum Metals Rev., 1974, 18, (1) 30 nickel, or cobalt by means of coupled re- duction have been started. The temperature needed to achieve complete reduction in- creases when going from platinum to iridium and decreases with increasing atomic number of the actinide element. Furthermore, alloys with a high noble metal : actinide ratio are more easily prepared than compounds having a lower ratio. This means, for example, that Pt,Cm may be prepared at about 12oooC, whereas the preparation of the iridium compound requires temperatures of 1550°C. The table shows the platinum-, palladium-, iridium-, and Rh-actinide intermetallic phases prepared by coupled reductions, among them the twenty com- pounds prepared for the first time. Intermetallic Phase Structure The structure of PtjA phases (A=Th, Np-Cm) was elucidated by comparison with compounds of the Pt,RE type which were prepared by coupled reduction and were investigated by Bronger (20). The hitherto unknown structure of Pt,Th and Pt,Pu proved to be isostructural with Pt,Sm; Pt,Pa and PtjU crystallise in the cubic Ni,U-structure type. The structure of the other new alloy phases could be established by comparison with known types of structures. The 2:1 compounds crystallise in the cubic Laves- phase type, whilst most 3:r phases have the ordered Cu,Au-structure type. Deviation from this structure type could be observed only for the platinum alloy phases of the light actinide elements. Four different types of structures were found for the 3:r com- thanide or the actinide series has the same pounds of the platinum system. Pt,Th is the stoichiometric formula but different lattice only compound whose type of structure is structures. The structures of these different unknown. The Pt,Pa and Pt,U alloy phases lattice types, however, are very similar, only crystallise in the hexagonal Cd,Mg-structure the stacking order being different. The type, Pt,Np has the hexagonal Ni,Ti- lattice constants of the isostructural series lattice, while the Pt3Pu phase has the ex- follow a trend which is generally known pected Cu,Au-structure type. This fact is when comparing the radii of the actinide remarkable because no case has been reported elements. In the case of the Ir,A (A=Th, as yet in which a simple compound of three U-Cm) and Rh3A (A=Th-Cm) com- neighbouring elements in either the lan- pounds this may be seen in Fig. I. A mini-

Platinum Metals Rev., 1974, 18, (1) 31 mum in the lattice constants is to be Seen measured with an “Hygrometer PA 1000’’ between uranium and neptunium, both made by Panametrics, Waltham, U.S.A. of which metals have been shown to possess The same method was also used to purify the highest valency in their metallic state. helium and argon, which sometimes were used Such a dependence of molar volume with as inert gas when cooling reaction products, the atomic number could not be observed in Coupled reductions for the preparation of the series of platinum compounds because most of the platinum alloy phases and all of of the different types of structures, having the palladium, iridium and rhodium alloy more or less dense packing of the atoms. phases were unsuccessful when performed The preparation of the Pt,RE phases with two commercial hydrogen diffusion (RE =La-Tm, Y) by coupled reduction was apparatuses (3). first performed by Bronger (zo), who also Some comment should be made concerning prepared some Pt,RE phases (RE =Ho-Lu, the analytical purity of the alloys. Because SC). the noble meta1:base metal ratio does not Apart from the Pt,RE (RE=La-Tm, Y) change during the coupled reduction the and Pt,RE (RE=Tb-Lu, Sc, Y) compounds ratio of the elements in the final products is we have been able to prepare other types of the same as in the initial mixtures; this has alloy phases of the lanthanides, such as been demonstrated by chemical analyses on Pt,RE (RE=La-Gd) and Pd,RE (RE- several characteristic compounds. Exceptions La-Lu, Sc, Y) and also Rh,Sc. All these to this rule will be discussed later. Determina- compounds have been prepared earlier by tions of the oxygen contents were made on conventional metallurgical techniques except most of the alloy phases; some phases were PdsTm, which could not have been prepared also analysed for nitrogen and hydrogen. (21) in pure form owing to the relatively low All determinations were performed by vacuum AH, of thulium metal, Further information hot extraction. The oxygen contents of most on Pd,Tm is available elsewhere (3,4). of these alloy phases do not exceed 500 p.p.m.; The hydrogen used was the purest cylinder the hydrogen and nitrogen contents are hydrogen available with oxygen and water always less than IOO p.p.m. This means that vapour contents of about 10 p.p.m. each. these phases are not stabilised by the in- In order to prepare high purity hydrogen corporation of oxygen or nitrogen (3). from it several stages of purification by An increase in stability, however, is to be physical and chemical methods were necessary observed when carbon is incorporated in the (3, 22). After passing through heated lattice of the compounds of Cu,Au-type titanium cuttings, heated platinum/asbestos structure. The reaction, for example, of and molecular sieves, the hydrogen was ThRh, with carbon yields ThRh,C, with a further purified by bubbling through a series maximum value of x=0.5 (25, 26). The of three bottles containing liquid potassium- increase in stability may be obtained from sodium alloy (approximate composition K :Na thermodynamic data. Using a high tempera- =2.5:1 (23)) and finally by cooling to liquid ture galvanic cell of the type nitrogen temperature to remove organic (-)TalU,UF,( CaF,[UF,,UPtJ?tl Ta(+) contaminants. The hydrogen thus purified had an oxygen a preliminary AGIOOOKvalue of -49.1 content of less than IO-~~torr and a partial kcal/mole for UP[, was obtained (27). water vapour pressure of less than IO-’ torr Using similar galvanic cells the following (22, 24). The oxygen content of the purified data were obtained: gas was determined by using a Tho, solid galvanic cell with Fe/FeO as the reference URu, AG,,,,, =-45.5 kcal/mole electrode. The water vapour content was URhB AG,,,,, =--62.4 kcal/rnole

Platinum Metals Rev., 1974, 18, (1) 32 The increase in stability of the carbon- doped compounds may be seen by comparing the AG values of URu, and URu,C,,, the compound with the highest carbon content (12). By using a galvanic cell of the type

( -)PtJ U,UF,I CaF,I UF,,URu,C,,Ru,C/ Pt( -) a AGloooKvalue for URu,Co7 of -60.1 kcal/mole was obtained. The increase in stability increases with the carbon content of the intermetallic phases. The influence of carbon in the octahedral positions on the stability of the metal lattice can be inter- preted by an increase of the effective valency of the platinum metals. Preparation of Metals When preparing americium-platinum alloys This means that metals like Am, Cf, Cm, at temperatures in excess of 13m°C we Ba-Ca, and Li can be prepared by direct observed mainly a loss of americium in the hydrogen reaction of the corresponding reaction product, Detailed studies showed oxides, the noble metals only serving as a that the americium-platinum alloys de- catalyst. This is shown by the scheme composed at temperatures of about 1300°C or in Fig. 2. higher, in vacuum or inert gas atmospheres, The significance of this type of reduction is to yield solid noble metal and americium. that using alloy phases as intermediates it is These could be condensed at cooled parts of possible to prepare volatile metals by hydrogen the reaction apparatus. reduction of the oxide. Preliminary data We also have been able to demonstrate even indicate that it is possible to separate that not only americium but also other two elements, for example, californium from volatile metals (e.g. Cf, Ca-Ba, Li) could be americium or from curium, by utilising the prepared by this method (3, 5, 6). difference in volatility of the actinide metals. Not yet investigated by us is the preparation With tracer amounts of 252Cfwe have been of the rare earth metals by this method. able to show that californium is volatilised However, remembering the above-mentioned by more than go per cent at x25o0C in fact, that the conventional metallurgical extremely pure hydrogen or in vacuum from technique for the preparation of Pd,Tm was californium oxide-platinum mixtures, at not successful (21), it should obviously be which temperature curium does not volatilise. possible to prepare rare earth metals, at Detailed studies of these processes and their least those with a relatively low AH, (e.g. thermodynamic relationships are now under thulium), investigation. Note that we were able to prepare calcium also from the mineral fluorite practically References in one step by coupled reduction to the I H. G. Diehl and C.Keller, J. Solid State Chem., 3, 621 compounds Pt,Ca or Pt,Ca followed by 1971, 2 B. Erdmann and C. Keller, Znorg. Nucl. Chem. decomposition in vacuum at higher temp- Left., 1971, 7, 675 eratures. Detailed studies of these processes 3 B. Erdmann, Kernforschungszentrum Karlsruhe Kept. KFK-1444,1971 and their thermodynamic relationships are 4 B. Erdmann and C. Keller, J. Solid Stare now under oa investigation. Chem., 1973, ?>40

Platinum Metals Rev., 1974, 18, (1) 33 5 U. Berndt, B. Erdmann and C. Keller, 14 W. Bronger and W. Klemm, Z. Anorg. Angew. Chem., 1972, 84, 537; Angew. Chem. Allgem. Chem., 1962, 319, 58 Internat. Ed. En., 1972, 11, 515 15 A. S. Darling, G. L. Selman and R. Rushforth, 6 U. Berndt, B. Erdmann and C. Keller, US. Platinum Metals Rea., 1970, 14, 54 Patent 16 Ibid., 1970, 14, 95 7 H. Holleck, KernforschungscentrumKarlsruhe 17 Zbid., 1970, 14, 124 Rept.KF K-1011, 1969 18 Zbid., 1971, 15, 13 8 H. Holleck, Proc. 3rd Internat. Conf. on 19 A. S. Darling, Platinum Metals Rev., 1967, Solid Compounds of Transition Elements, 11, 138 Oslo, 1969 20 W. Bronger,J. Less-commonhfetals, 1967,12,63 9 E. D. Eastman, L. Brewer, L. A. Brornley, 21 R. D. Hutchens, V. U. S. Rao, J. E. Greedan, P. W. Gilles and N. L. Lofgren, J. Am. V. E. Wallace and R. S. Craig,?. Appl. Phys., Ceram. SOC.,1951, 34, 128 1971,423 1293 10 R. A. Mercuri and J. M. Criscione, Abstr. 22 €3. Erdmann, GIT, 1972, 16, 1283 Papers, 158th Mtg., Am. Chem. SOC.,1969, 23 M. Hansen and K. Anderko, Constitution of INOR 33 Binary Alloys, McGraw-Hill, New York, 11 H. Holleck,J. Nucl. Muter., 1968, 28, 339 1958, P- 876 24 Messen u. Steuern, 1972, H29,27 12 H. Holleck and H. Kleykamp,J. Nucl. Mazer., B. Erdmann, 197% 35, 158 25 H. Holleck,J. Nucl. Matcr., I972,&, 278 13 G. Grube and M. Flad, Z. Eleknochem., 26 H. Holleck, Monatsh. Chem., 1971, IOZ, 1699 19423 48, 377 27 N. Schmidt and C. Keller, unpublished data

Science Awards to Platinum Metals Research Workers PROFESSOR GEOFFREY WILKINSON AND DR VLADIMIR HAENSEL The achievements of two research workers the platinum metals. His name is closely prominent in the platinum metals field were associated with the family of catalysts based recognised in October 1973, when awards to on tristriphenylphosphinechlororhodium, Professor Geoffrey Wilkinson and Dr (Ph,P),RhCl, from which the carbonyl Vladimir Haensel were announced. Both complexes RhCl(CO)(Ph,P), and RhH(C0) recipients have contributed to past issues of (Ph,P), used in hydroformylation reactions Platinum Metals Review. In the January 1959 have been developed. issue Dr Haensel described “The Penex Pro- The National Medal of Science, the U.S. cess for Pentane Isomerisation”. In January Federal Government’s highest award for 1964 he contributed “Duofunctional Cat- distinguished achievement in science, mathe- alysts in the Petroleum Industry”, while matics and engineering, was presented by Professor Wilkinson reviewed “Organome- President Nixon at the White House on tallic Compounds of the Platinum Metals”. October 10th to ten scientists, among whom Wilkinson then reviewed “Tertiary Phos- was Dr Haensel, Vice-president for Science phine Complexes of the Platinum Metals” and Technology of the Universal Oil Pro- in the April 1968 issue. ducts Company, Des Plaines, Illinois, for Professor Wilkinson, now of Imperial “his outstanding research in the catalytic College, London, has been awarded the Nobel reforming of hydrocarbons which has greatly Prize for chemistry jointly with Ernst enhanced the economic value of our petroleum Fischer. The award was made for determin- natural resources”. Dr Haensel’s work has, ing the structure of “sandwich” compounds, of course, included responsibility for the initially of ferrocene, and subsequently development and improvement of the series synthesising many other such compounds. of U.O.P. Platforming catalysts since their Ruthenocene is one such platinum metal inception. Platinum-based catalysts for the compound. The original work was carried reforming of hydrocarbons were first de- out in 1952 when Wilkinson was teaching at veloped at U.O.P. and Haensel’s team has Harvard. Since returning to London in 1956 constantly upgraded their performance, more his interests have gradually shifted to other recently by the introduction of bimetallic areas, among which the development of some catalysts in which a second metal synergistic- of the most active homogeneous catalysts is ally improves the reforming performance of of particular importance to the chemistry of platinum. F. J. S.

Platinum Metals Rev., 1974, 18, (1) 34