Formation of carbon supported PtRu alloys: an XRD analysis

Journal of Alloys and Compounds 315 (2001) 118–122 L www.elsevier.com / locate / jallcom Formation of carbon supported PtRu alloys: an XRD analysis E. Antolini a , *, F. Cardellini b a Scuola di Scienza dei Materiali, Via 25 Aprile 22, 16016 Cogoleto, Genova, Italy b ENEA, Via Anguillarese 301, 00060 Santa Maria di Galeria, Rome, Italy Received 26 April 2000; accepted 7 October 2000 Abstract Carbon supported PtRu alloys were prepared by impregnation of Pt and Ru precursors on a porous carbon support, followed by reduction of the metals with Na 2 S 2 O 4 . After reduction, the samples were thermal treated in argon up to 7008C. The samples were characterized by atomic absorption (AAS) and X-ray diffraction (XRD) measurements. Before thermal treatment only carbon reflexions were visible in XRD pattern. The reflexions of face centered cubic (f.c.c.) PtRu alloy were revealed in XRD pattern starting from thermal treatment at 3008C. No hexagonal close packed (h.c.p.) RuPt reflexions were detected. During thermal treatment, part of Ru reacted with sulphur forming RuS 2 . Ru content in the alloy increased with increasing thermal treatment temperature. The results indicated that, first, f.c.c. Pt with few Ru alloyed was formed, then, with increasing thermal treatment temperature, part of Ru atoms present in the sample in an amorphous form entered in the crystal structure of the platinum by a diffusion-controlled mechanism. Also after thermal treatment at high temperatures there was a large part of unalloyed Ru, with only about 49% of the Ru alloyed with the Pt. PtRu particle size was in the range 15–20 nm.  2001 Elsevier Science B.V. All rights reserved. Keywords: Electrode materials; Intermetallics; Nanostructures; Nanofabrication; X-ray spectroscopy 1. Introduction Carbon supported PtRu alloys seem to be the best anode materials for both direct methanol fuel cell (CH 3 OH as the fuel) and polymer electrolyte fuel cell (H 2 / CO as the fuel), because the problem of poisoning of the electrode by carbon monoxide is less severe than in pure Pt catalysts [1–4]. The crystal structure of pure Pt is face centered cubic (f.c.c.), while that of Ru is hexagonal close packed (h.c.p.). For Ru atomic fractions up to about 0.7, Pt and Ru form a solid solution with Ru atoms replacing Pt atoms on the lattice points of the f.c.c. structure. The lattice constant of ˚ at 0.675 atomic fraction Pt decreases from 3.923 to 3.83 A Ru [5]. Above 0.7 at Ru, another solid solution can form with Pt atoms replacing Ru in an h.c.p. structure. The interactions of the metals with the carbon support affect the formation of PtRu alloy. Unsupported PtRu alloys are easily obtained. For example, Chu and Gilman obtained single phase PtRu alloys with different composition up to the nominal Pt:Ru atomic Ratio 1:1 (obviously the composition of the alloy was the same of nominal composition) *Corresponding author. by thermal treatment at 2508C in H 2 /Ar of a aqueous solution of H 2 PtCl 6 and RuCl 3 precursors [6]. For that regarding carbon supported PtRu alloys, instead, it is difficult to obtain single phase PtRu, and it is almost difficult to obtain a Pt:Ru atomic ratio in the supported alloy like nominal Pt:Ru in the sample. Rauhe et al. tried to obtain PtRu / C alloy by codeposition of metal chlorides precursors in the Pt:Ru atomic ratio 1:1 from aqueous solution onto high-surface area graphite, followed by reduction in a hot H 2 / N 2 stream [7]. They obtained pure Pt and a Ru-rich alloy. X-ray absorption studies (XAS) were done by Mukerjee et al. on a carbon supported PtRu catalyst with Pt:Ru atomic ratio 1:3.4 [8]. They found that Pt was alloyed with Ru and that the Ru content was about 25 atomic percent. There was a large excess of unalloyed Ru, with only about 10% of the Ru alloyed with the Pt. He et al. tried to prepare carbon supported PtRu alloy in the atomic ratio 1:1 by dissolving the metal precursors in water, reducing them with Na 2 S 2 O 3 , adding the carbon support, further reducing in H 2 at 2508C, and, finally, heating at 6608C in N 2 [9]. XRD analysis showed that the ˚ indicating Pt[220] peak resulted in a d-spacing of 1.38 A, the formation of a PtRu alloy with Pt:Ru atomic ratio less than 1:1. An XRD analysis on the dependence of the formation of carbon supported PtRu alloy with nominal 0925-8388 / 01 / $ – see front matter  2001 Elsevier Science B.V. All rights reserved. PII: S0925-8388( 00 )01260-3 E. Antolini, F. Cardellini / Journal of Alloys and Compounds 315 (2001) 118 – 122 PtRu atomic ratio 1:1 on the Me /(Me1C) weight ratio, with Me5Pt1Ru, indicated that Ru content in PtRu alloy increased with increasing the metal content of the samples [10]. The aim of this work is to investigate the formation mechanism of carbon supported PtRu alloy, by analyzing the changes of XRD patterns, following thermal treatment of the samples at various temperatures. 2. Experimental procedure The samples consisting of Pt and PtRu, with Pt:Ru atomic ratio 1:1 and 1:3, supported on carbon were prepared by a deposition / reduction process of H 2 PtCl 6 and H 2 RuCl 5 ?H 2 O precursors. The H 2 PtCl 6 and H 2 RuCl 5 ?H 2 O (only for PtRu / C samples) precursors were first adsorbed on the carbon support, then they were ‘in situ’ reduced to metallic Pt and Ru. Ketjenblack (KJB) was used as the carbon support. KJB has a high specific surface area of 900 m 2 g 21 and high basicity (presence of a high concentration of basic functional groups). This commercial carbon was purified (elimination of inorganic impurities) by treatment with an aqueous solution of HNO 3 at room temperature for 48 h.. After HNO 3 treatment, the carbon was washed with deionized water. A PtRu supported on Vulcan XC-72R Cabot (V) was also prepared to evaluate the degree of alloying in the absence of thermal treatment. This carbon has a specific surface area of 254 m 2 g 21 . H 2 RuCl 5 ?H 2 O precursor has been obtained by RuCl 3 ?3H 2 O solution in HCl 0.2 M. An aqueous solution of H 2 PtCl 6 and H 2 RuCl 5 ? H 2 O was added to a suspension of carbon in a water / ethanol mixture. The resulting suspension was heated at 608C (impregnation step). Then an aqueous solution of Na 2 S 2 O 4 , as a reducing agent [11,12], was slowly added (reduction step). Finally, the metallized carbon was filtered, washed in H 2 O and dried in air at 1108C. After drying, the catalyst was submitted to a thermal treatment in argon from room temperature to a temperature in the range 200–7008C at heating rate 158C / min and held at the maximum temperature for 10 min. Chemical analysis of the catalyst was accomplished using atomic absorption spectrophotometry (AAS) by a Perkin Elmer Model 305 atomic absorption spectrophotometer. X-ray diffraction (XRD) measurements were carried out with a Italstructures powder diffractometer, using a focused and monochromatized Cu K a source for Pt and PtRu samples supported on KJB, and a Co K a source for PtRu sample supported on V, with a position sensitive detection 1208. 119 was about 10 w / o for all the samples. Pt:Ru atomic ratio was 1:1 for KJB supported metals and 1:3 for V supported metals. XRD is a bulk method, and reveals information on the bulk structure of the catalyst and its support. The first peak, at the low 2u range, is associated with the carbon support. Fig. 1 shows the XRD pattern of KJB supported pure platinum following thermal treatment up to 4008C. The Pt pattern displayed the [111], [200], [220] and [311] reflexions characteristic of f.c.c. crystal structure. Fig. 2 shows the XRD patterns of KJB supported PtRu / C samples after thermal treatment at different temperatures. XRD patterns of the samples thermal treated at temperatures lower than 3008C displayed the reflexions of the carbon only. This indicates the presence of small and poor crystalline metal particles, with a weak internal organization. Starting from thermal treatment at 3008C, the patterns displayed the reflexions characteristic of platinum f.c.c. crystal structure, with some shift in the position of each reflexion peak. The shifts in 2u correspond to decreased lattice constant due to incorporation of Ru atoms. f.c.c. reflexions showed a tail, which can be attributed to a low composition homogeneity. The reflexions of a RuS 2 phase [13,14] were also revealed, and the maximum RuS 2 amount was formed at 5508C. The reflexions characteristic of an either pure Ru or Ru-rich h.c.p. phase, instead, was not detected. PtRu particle size, calculated from the XRD patterns using the Scherrer formula, was in the range 15–20 nm. To evaluate the thermal crystallization, we have considered as a crystallinity degree index of Pt particles the peak height ratio of the Pt [111] crystal face and the C[0 0 15] reflexion of the carbon. Fig. 3 shows the dependence of Pt[111] / C[0015] height ratio vs. thermal treatment temperature. The crystallisation of PtRu in the f.c.c. structure 3. Results AAS analysis indicated the metal content Me /(Me1C) and Pt:Ru atomic ratio of the samples. The metal content Fig. 1. X-ray diffraction pattern of carbon (KJB) supported platinum, thermal treated at 4008C. 120 E. Antolini, F. Cardellini / Journal of Alloys and Compounds 315 (2001) 118 – 122 height slightly decreased owing to the broadening of the reflexion due to Pt-support interactions at high temperatures. The dependence of d-spacing of PtRu[220] and PtRu[311] reflexions on thermal treatment temperature is shown in Fig. 4a,b. The d-spacing of both reflexions decreased with increasing annealing temperature. This reduction in d-spacing corresponds to a reduction of the lattice parameter, resulting from the replacement of the platinum atoms on the lattice points of the f.c.c. crystal structure by the smaller ruthenium atoms. From d-spacing values of [220] and [311] reflexions, we have calculated the lattice parameter of PtRu alloys. Fig. 5 shows the dependence of PtRu lattice parameter on thermal treatment temperature. Lattice constant values were lower than that of pure platinum supported on KJB (upper dashed line), and higher than that of unsupported PtRu alloy with Pt:Ru atomic ratio 1:1(lower dashed line). The lattice parameter decreased with increasing temperature. The lowest value of the lattice parameter was similar to the value of the lattice Fig. 2. X-ray diffraction patterns of carbon (KJB) supported PtRu with nominal Pt:Ru 1:1, thermal treated at different temperatures. took place starting from 3008C, and the intensity of PtRu[111] reflexion increased, owing to the increase of the crystallinity, up to around 4508C. Then the PtRu[111] peak Fig. 3. PtRu[111] / C[0015] peak height ratio vs. thermal treatment temperature. Fig. 4. Dependence of PtRu d-spacing for [220] (a) and [311] (b) ˚ reflexions on thermal treatment temperature (d-spacing error60.001 A). E. Antolini, F. Cardellini / Journal of Alloys and Compounds 315 (2001) 118 – 122 121 From the values of lattice parameters, we have calculated Ru atomic fraction in PtRu alloy. The dependence of PtRu lattice parameter of unsupported alloy on Ru content follows the Vegard’s law and can be obtained from literature data [5,6,16]. For supported alloys, it is difficult to evaluate the degree of alloying, as the lattice constant of pure supported platinum is lower than that of unsupported Pt, owing platinum-carbon interactions [17,18], and the Ru atomic fraction in f.c.c. alloy is less than the nominal Ru content of the sample. Assuming that the dependence of the lattice parameter on Ru content is the same for supported and unsupported Pt, the lattice constant of carbon supported PtRu l PtRu / C results: l PtRu / C 5 l os 2 kx Ru Fig. 5. PtRu lattice constant vs. thermal treatment temperature. Lattice ˚ (d) PtRu / C (KJB) with Pt:Ru atomic ratio parameter error60.0025 A. 1:1; (s) PtRu / C (V) with Pt:Ru 1:3; (upper dashed line) pure Pt on KJB carbon thermal treated at 4008C; (lower dashed line) unsupported PtRu with Pt:Ru 1:1 [6]. parameter of commercial PtRu / C by E-TEK with Me / (Me1C) 30% and nominal Pt:Ru 1:1 [10]. Towards the aim to obtain a value of C-supported PtRu lattice constant in the absence of thermal treatment, we have prepared a sample with nominal Pt:Ru atomic ratio 1:3, deposited on untreated Vulcan. Indeed, the formation of metal crystallites also depends on the characteristics of the substrate [12,15]. Fig. 6 shows the XRD pattern of PtRu supported on Vulcan. In this case we obtained large and well crystallized PtRu particles also without thermal treatment. On the basis of the values of d-spacing, f.c.c. PtRu alloy was formed with low Ru content. No Ru-rich h.c.p. and RuS 2 phases were revealed. Then, a large part of Ru was present in an amorphous form. Also with a high Ru content in the sample, Ru content in the alloy was low. (1) ˚ is the lattice parameter of pure carbon where l os 53.9155 A ˚ is a constant, obtained supported platinum, and k50.124 A from data related to unsupported alloys. Table 1 reports the value of x Ru calculated from Eq. (1). Using these values of x Ru , we have calculated the amount of Ru alloyed (Ru al ) from the relation: Ru al 5 x Ru / [(1 2 x Ru )(Ru / Pt) nom ] (2) where (Ru / Pt) nom is the nominal Ru / Pt atomic ratio. Fig. 7 shows the dependence of the percent of Ru alloyed on thermal treatment temperature. The amount of Ru alloyed increased with temperature, but it is less than 50% after thermal treatment at 7008C. This result confirms the difficulty to obtain carbon supported PtRu alloy in the nominal Pt:Ru atomic ratio. Also, after thermal treatment at high temperatures, only part of Ru entered into Pt lattice. Part of unalloyed Ru formed RuS 2 , while another part was in an amorphous form. Then, we can suppose the presence of a diffusion process of Ru atoms into f.c.c. Pt structure. The diffusion kinetics can be expressed as: Ru al 5 kt n (3) where k5k 0 exp(2Ea /RT ) is the rate constant; Ea is the activation energy of the diffusion process; t is the thermal treatment time; and the exponent n is related to the reaction mechanism. Plotting log Ru al vs. 1 /T at constant time, we can determine the activation energy of the Table 1 Ru atomic fraction in PtRu alloys, calculated from Eq. (1)a Fig. 6. X-ray diffraction pattern of carbon (V) supported PtRu with Pt:Ru atomic ratio 1:3, thermally untreated. Sample (Nominal composition) Thermal treatment Temperature (8C) Ru atomic fraction in PtRu [from Eq. (1)] PtRu PtRu PtRu PtRu PtRu PtRu PtRu No 300 350 450 550 600 700 0.10 0.04 0.07 0.14 0.25 0.33 0.33 a 1:3 1:1 1:1 1:1 1:1 1:1 1:1 on V on KJB on KJB on KJB on KJB on KJB on KJB Ru atomic fraction error60.02. 122 E. Antolini, F. Cardellini / Journal of Alloys and Compounds 315 (2001) 118 – 122 4. Conclusions The interaction of the Ru with the support hinders the formation of PtRu / C alloy. In the absence of thermal treatment an amorphous structure or f.c.c. Pt with few Ru alloyed (about 4%), depending on the carbon substrate characteristics, was obtained. Thermal treatment in argon supported the formation of the alloy. Ru content in the alloy increased with temperature, but only part of the Ru (about 49%) alloyed with the Pt. The incorporation of Ru atoms into Pt lattice was diffusion-controlled and the value of activation energy of the process was 34 kJ mol 21 . During thermal treatment, part of the Ru reacted with sulphur to form RuS 2 . Fig. 7. Dependence of Ru alloyed on thermal treatment temperature. process. Fig. 8 shows the dependence of ln Ru al on the reciprocal of absolute temperature. The linearity of the plot attests the presence of a diffusion process. The value of Ea was 34 kJ mol – 1 . On the basis of the plot shown in Fig. 8 the percent of Ru alloyed by diffusion process in the absence of thermal treatment should be about zero. The value of alloyed Ru of untreated PtRu supported on V with Pt:Ru 1:3, instead, was 4%. In the presence of an excess of Ru or / and of a different carbon support (different metalsupport interactions), perhaps Ru atoms substitute Pt atoms into f.c.c. structure during the formation of the crystallites. Fig. 8. Arrhenius plot of Ru alloyed against 1 /T. References [1] H.A. Gasteiger, N. Markoviv, P.N. Ross, E.J. Cairns, J. Phys. Chem. 98 (1994) 617. [2] H.A. Gasteiger, N. Markovic, P.N. Ross, J. Phys. Chem. 99 (1995) 8290. [3] F. Richarz, B. Wohlmann, U. Vogel, H. Hoffschelz, K. Wandelt, Surf. Sci. 335 (1995) 361. [4] A. Kabbabi, R. Faure, R. Durand, B. Beden, F. Hahn, J.M. Leger, C. Lamy, J. Electroanal. Chem. 444 (1998) 41. [5] N.V. Ageev, V.G. Kuznetsov, Izvest. Akad. Nauk. S.S.S.R. (Khim) (1937) 753. [6] D. Chu, S. Gilman, J. Electrochem. Soc. 143 (1996) 1686. [7] B.R. Rahue, F.R. McLarnon, E.J. Cairns, J. Electrochem. Soc. 142 (1995) 1073. [8] J. McBreen, S. Mukerjee, J. Electrochem. Soc. 142 (1995) 3399. [9] C. He, H.R. Kunz, J.M. Fenton, J. Electrochem. Soc. 144 (1997) 970. [10] E. Antolini, F. Cardellini, L. Giorgi, E. Passalacqua, J. Solid State Electrochem. (in press). [11] K. Tsurumi, T. Nakamura, A. Sato, US Patent 4,956,331 (1990). ` V. Alderucci, V. Antonucci, N. Giordano, Appl. [12] T. Torre, A.S. Arico, Catal. A 114 (1994) 257. [13] R. Juza, W. Meyer, Z. Anorg. Chem. 213 (1933) 273. [14] O. Knop, Can. J. Chem. 41 (1963) 1832. [15] H.E. van Dam, H. van Bekkum, J. Catal. 131 (1991) 335. [16] H.A. Gasteiger, P.N. Ross, E.J. Cairns, Surf. Sci. 293 (1993) 67. [17] A.K. Shukla, M.K. Ravikumar, A. Roy, S.R. Barman, D.D. Sarma, ` V. Antonucci, L. Pino, N. Giordano, J. Electrochem. Soc. A. Arico, 141 (1994) 1517. ` K.M. El-Kathib, H. Kim, P.L. Antonucci, V. [18] A.K. Shukla, A. Arico, Antonucci, Appl. Surf. Sci. 137 (1999) 20.