Crystal structures and hydrogenation properties of palladium-rich compounds with elements from groups 12–16

2.1 Binary palladium-rich compounds with elements of group 12

The compounds PdCd, PdHg (both AuCu type) and Pd_0.75Zn_0.25 (solid solution of Cu type) were synthesized from the elements. The latter was formed in an attempt to prepare an ordered phase Pd₃Zn via mineralization by iodine within 3 months. Crystal structures refined on the basis of XRPD data (Table 1) are in good agreement with the literature [14–16]. No reaction with hydrogen is observed for any of the three compounds under given terms (Table 1).

2.2 Binary palladium-rich compounds with elements of group 14

The compounds Pd₂Sn (Co₂Si type), Pd5Pb3 (Ni5Ge3 type), and Pd₁₃Pb₉ (Pd₁₃Pb₉ type) do not show any reaction with hydrogen according to DSC experiments and by comparing XRPD data before and after the hydrogenation experiment. Refined crystal structures (Table 1) agree well with those from the literature [17–19]. The structures of Pd₃Sn and Pd₃Pb are also in accordance with literature [20]. However, an increase of the unit cell volume by up to 0.4 and 0.6 % upon hydrogenation, respectively, indicates a possible hydrogen uptake. [Pd₆] octahedral sites, which are preferred by hydrogen [5], are present in their structures (AuCu₃ type). Several hydrides MPd3Hx with occupation of these interstices, resulting in a cubic anti-perovskite type structure, are known already (M = Mg [3], In [5], Tl [6], Y [7], Mn [8], Ce [9]). However, no significant thermal signal during the in situ DSC experiment could be observed. This may be assigned to low hydrogen content and small bonding energy of hydrogen in these compounds.

2.3 Binary palladium-rich compounds with elements of group 15

From the wide range of Pd–As compounds known [11, 21–33] only for PdAs₂ and Pd₅As high quality crystal structure data are available [33, 11]. For Pd₃As [26] only a structure type was assigned, but no crystal structure was refined. This prompted us to reinvestigate the crystal structure of Pd₃As and perform a Rietveld refinement based on XRPD data. The results (Table 2, Fig. 1) confirm the assignment of the Fe₃P structure type (D0_e) and provide the first refined structure data for Pd₃As.

Fig. 1:

Rietveld refinement of the crystal structure of Pd₃As based on X-ray powder diffraction data (CuK_α radiation); refined crystal structure data and residual values in Table 2.

Further details of the crystal structure investigation may be obtained from Fachinformationszentrum Karlsruhe, 76344 Eggenstein-Leopoldshafen, Germany (fax: +49-7247-808-666; e-mail: crysdata@fiz-karlsruhe.de, http://www.fizkarlsruhe.de/request_for_deposited_data.html) on quoting the deposition number CSD-430924.

XRPD data confirm the NiAs structure type for PdSb and the refined lattice parameters (Table 1) are in good agreement with literature [34, 35]. The correspondence of refined lattice parameters for Pd₅Sb₂, Pd₈Sb₃, Pd₂₀Sb₇, and Pd₅Bi₂ with those from literature [36–38] is also reasonable. According to the absence of thermal effects in DSC experiments and the absence of significant unit cell volume changes (no XRD for Pd₂₀Sb₇ and Pd₈Sb₃ after hydrogenation), none of the Pd–As, Pd–Sb and Pd–Bi phases investigated shows any signs for hydrogen uptake.

2.4 Binary palladium-rich compounds with elements of group 16

Pd₄Se and Pd₁₇Se₁₅ were synthesized and their structure are in good accord with literature data [39, 40]. The unit cell volumes of both compounds decrease slightly during the hydrogenation experiment, which might be connected to the reaction of the side product in case of Pd₄Se (Table 1). DSC and XRPD experiments suggest the absence of significant hydrogen incorporation.

2.5 Ternary palladium-rich compounds with Pd₅TlAs type and related structures

The structure type of Pd₅TlAs may be described as an intergrowth structure with AuCu₃ like and CsCl like slabs [41]. It may also be derived from a cubic close packing by doubling of one lattice parameter, i.e. a cubic-tetragonal transition, and ordering the eight atomic positions in a 5:1:1:1 fashion with five palladium, one thallium, one arsenic atom and one vacancy. It thus represents one of many ordered intermetallic compounds related to the cubic close packing by crystallographic group–subgroup relationships [42–44]. This structure type exhibits many octahedral sites, which seem to be attractive for incorporation of hydrogen. Compounds with this type of structure thus seem to be good candidates for hydrogenation. The refined structures of Pd₅InAs, Pd₅HgSe and Pd₅InSe are in good agreement with literature [43, 45–47]. For all other compounds listed in Tables 1 and 3 these are the first refined crystal structure data. The only free positional parameter, z(Pd2) is higher for the compounds with arsenic than for those with selenium resulting in shorter Pd–As as compared to Pd–Se distances as expected from atomic sizes (Table 3). The c/a ratio of the Pd₅TlAs type structure is always considerably smaller than two. That means that the smaller selenium or arsenic atoms are packed closer along the crystallographic c direction than the larger cadmium, mercury, indium or thallium atoms.

From all compounds listed above only Pd₅InSe exhibits reactivity towards hydrogen (Table 1). The distorted [Pd₄M₂] (M = Cd, Hg, In) and [Pd₅X] (X = As, Se) octahedral sites in Pd₅MX are two possible hydrogen positons in this structure. Occupation by hydrogen would lead to distances between 1.96 and 2.06 Å, which are comparable with palladium-hydrogen distances in Pd₃InH_0.89 (2.01 Å [5]). However, unreasonably short distances As–H, Se–H, Cd–H, Hg–H, In–H, Tl–H would result, which probably prevents hydrogen from entering the structures.

Pd₅InSe and Pd₈In₂Se [46] form Pd₃InH_x during the hydrogenation experiment. Differential thermal analysis (DTA) of Pd₅InSe in helium atmosphere was executed and the results have shown two reversible thermal signals with a hysteresis (Fig. 2, top). X-ray powder diffraction after the DTA confirmed that the reaction is fully reversible. To investigate the intermediates, a temperature-resolved XRPD experiment was carried out. The false color plot (Fig. 2, bottom) shows the formation of Pd₈In₂Se at about 525 °C. Upon further heating Pd₈In₂Se decomposed at about 625 °C and Pd₃In (ZrAl₃ type) was formed. On cooling (not shown here) Pd₅InSe returned.

Fig. 2:

Differential thermal analysis (top) and temperature-resolved XRPD (MoK_α1 radiation; bottom) of Pd₅InSe.

We suggest the following reactions to take place:

2 Pd5InSe  ⇄351 ∘C576 ∘C  Pd8In2Se + (2 Pd + Se)(l)⇄579 ∘C655 ∘C  2 Pd3In (ZrAl3 type) + 2 (2 Pd + Se)(l)

For the proposed liquid phase we have indirect evidence from annealing experiments. Pd₅InSe samples were annealed at 600 and 800 °C, i.e. after the first and second thermal signal in the DTA, respectively, and the silica glass ampoules were quenched in water. The main phases were Pd₈In₂Se at 600 °C and Pd₃In at 800 °C with secondary phases Pd₃₄Se₁₁, Pd₇Se₄ and Pd₁₇Se₁₅. The latter forms from the liquid (2 Pd + Se) by quenching according to the phase diagram [48] of the system Pd–Se. Thus, hydrogen plays a role in the reactions (Table 1) only insofar that it reacts with Pd₃In formed by thermal decomposition of Pd₅InSe.

2.6 The half-antiperovskite Pd₃Bi₂Se₂

The half-antiperovskite Pd₃Bi₂Se₂ (Ni₃Bi₂S₂ type) [49] was investigated for its hydrogenation properties. [Pd₆] octahedral sites in its crystal structure, albeit strongly distorted, suggested this structure family to be a good hydrogenation candidate from geometric reasons. However, neither in situ DSC nor comparison of unit cell volumes before and after the hydrogenation experiment hint towards any hydrogen uptake.

4.1 Synthesis of intermetallic compounds

Intermetallic compounds were synthesized from stoichiometric mixtures of the elements in evacuated silica tubes. Temperature treatment is given in the following as final temperature, holding time and heating rate, e.g. 875 °C 24 h 1 °C min^–1 describes a temperature treatment of heating with 1 °C min^–1 to 875 °C, holding this temperature for 24 h and cooling with the natural cooling rate of the furnace. In some cases a few mg of iodine was added (chemical vapor transport reaction), denoted by “I₂”.

• Pd_0.75Zn_0.25: 1000 °C 4 d

• PdCd: 770 °C 3 h, 280 °C 14 d

• PdHg: I₂, 400 °C 9 d 1 °C min^–1

• Pd₂Sn: I₂, 720 °C 5 d 1 °C min^–1

• Pd₃Sn: 720 °C 72 h 1 °C min^–1

• Pd₅Pb₃: 1200 °C 2 h 1.5 °C min^–1 quenched at 900 °C, ground in mortar, 370 °C 11 d 1 °C min^–1

• Pd₁₃Pb₉: 1200 °C 2 h 1.5 °C min^–1 quenched at 900 °C, ground in mortar, 520 °C 9 d 1 °C min^–1

• Pd₃Pb: I₂, 875 °C 24 h 1 °C min^–1, 430 °C 7 d

• Pd₃As: 650 °C 2 h 3.5 °C min^–1, 1000 °C 72 h 2 °C min^–1

• Pd₂₀Sb₇: 900 °C 168 h 2.1 °C min^–1, 25 °C 1.2 °C min^–1

• Pd₈Sb₃: 1000 °C 120 h 1.8 °C min^–1

• Pd₅Sb₂: 1000 °C 168 h 1.3 °C min^–1

• PdSb: 850 °C 6 h 2 °C min^–1, 750 °C 48 h 1.7 °C min^–1

• Pd₅Bi₂: I₂, 450 °C 4 d 0.5 °C min^–1

• Pd₁₇Se₁₅: 430 °C 7 d 1 °C min^–1

• Pd₄Se: 700 °C 2 h 1 °C min^–1, 375 °C 7 d

• Pd₅TlAs: 650 °C 2 h 3.4 °C min^–1, 1000 °C 30 h 1.9 °C min^–1

• Pd₅CdSe, Pd₅CdAs: 750 °C 6 d 1 °C min^–1

• Pd₅HgSe I₂, 400 °C 10 d

• Pd₅InSe, Pd₅InAs: I₂, 950 °C 4 h; 700 °C 7 d

• Pd₈In₂Se 950 °C 4 h; 750 °C 6 d

• Pd₃Bi₂Se₂: 1200 °C 1 h 3 °C min^–1; 500 °C 500 h

• Products were obtained as gray powders, some of them with a silvery luster.

4.2 X-ray powder diffraction (XRPD)

XRPD data were collected using flat reflection samples on a Panalytical X’Pert at T = 23(1) °C with CuK_α radiation or using flat transmission samples on a Huber Guinier G670 camera with an image plate system using either CuK_α1 or MoK_α1 radiation. Rietveld refinements were carried out with the program FullProf [13] and pseudo-Voigt as profile function. Absorption effects were modelled with a fixed overall thermal displacement parameter of –1.8 Ų in the refinement of the crystal structure of Pd₃As.

4.3 Thermal analysis (in situ DSC and ex situ DTA)

Differential scanning calorimetry was performed in situ under hydrogen pressures on a Q1000 DSC (TA Instruments) equipped with a gas pressure chamber. Twenty to thirty milligrams of the powdered intermetallics was put in aluminum crucibles, which were closed with an aluminum lid. These were placed inside the pressure chamber, which was then purged several times with hydrogen gas before filling it to the final hydrogen gas pressure of 5.0 MPa. Samples were heated to 430 °C with 10 °C min^–1, held at that temperature for a minimum of 1 h, and cooled to 27 °C with 10 °C min^–1. Usually, two or three such runs were performed, before the hydrogen pressure was released, the sample taken out and structural characterization undertaken. After releasing the gas pressure, the products were characterized ex situ by XRPD. The difference thermal analysis (DTA) of Pd₅InSe in helium atmosphere was carried out on a Netzsch F1 Jupiter device using sintered alumina crucibles and a heating rate of 10 K min^–1.