1: \documentstyle[12pt,psfig]{article}
2: \textwidth 6.1in
3: \textheight 8.2in
4: \topmargin =-.25in
5: \renewcommand{\baselinestretch}{1.6}
6: \begin{document}
7: \begin{center}
8: \textbf{{\Large{}\textbf {Effects of temperature and sulphur on the
9: composition profile of Pt-Rh nanocatalysts : A comparative study.}}}
10: \end{center}
11: \begin{center}
12: \textbf{Abir De Sarkar and Badal C. Khanra}\textbf{$^{\rm{}*}$}
13: \end{center}
14: \begin{center}
15: \textit{Condensed Matter Physics Group, Saha Institute of Nuclear Physics,}
16: \end{center}
17: \begin{center}
18: \textit{ 1/AF, Bidhannagar, Calcutta - 700064, India}
19: \end{center}
20:
21: \noindent
22: \textbf{Abstract}
23:
24:
25: \noindent
26: Monte-Carlo simulation technique has been used to investigate the
27: effect of temperature and adsorbed gases on the composition
28: profile of unsupported Pt-Rh nanocatalysts. For a 2406 atom fcc
29: cubo-octahedral Pt$_{\rm{}50}$Rh$_{\rm{}50}$ nanocatalyst the
30: shell-wise composition for all the eight shells has been
31: simulated. For the temperatures 700 K, 1000 K and 1300 K, the
32: top shell of clean Pt-Rh nanocatalysts is found to be
33: mildly Pt-enriched, while the second shell is Pt-depleted. The
34: Pt concentration of the top shell shows a maximum at T = 1000
35: K. In presence of a quarter monolayer of adsorbed oxygen the
36: top shell shows Rh enrichment, while all the other shells show
37: Pt-enrichment. This is true for all the three temperatures for
38: which the composition profiles have been studied.
39:
40:
41:
42: \noindent
43: PACS: 61.46.+W, 61.66Dk,68.10. Jy, 68.35. Md
44:
45:
46: \noindent
47: \textit{Keywords:} Nanocatalysts, segregation, Monte-Carlo
48: simulation, adsorption
49:
50:
51: \noindent
52:
53: $\bullet$ Corresponding Author\\
54: FAX (91) (33) 337 4637\\
55: Email: <badal@cmp.saha.ernet.in>
56:
57: \newpage
58: \noindent
59: \textbf{1. Introduction}
60:
61:
62: \noindent
63: The ceria-supported Pt-Rh bimetallic catalysts have been found to be
64: very effective three-way catalysts for simultaneous elimination of
65: the CO , NO and the hydrocarbons from automobile exhaust gases [1-3].
66: These catalysts are increasingly used globally in the catalytic
67: converters of automobiles. It is important, therefore, to have a
68: thorough understanding of the structural, electronic and catalytic
69: properties of these catalysts. In this work , we present the results
70: of our theoretical investigations on the composition profile of
71: unsupported Pt-Rh bimetallic nanostructures as a function of
72: temperature and adsorbates.
73:
74:
75:
76: The structural and catalytic properties of the Pt-Rh
77: alloy surfaces have earlier been exhaustively studied by
78: several groups [4-11]. Van Delft et al. [4] used AES to
79: investigate the Pt$_{\rm{}62 }$Rh$_{\rm{}38 }$ and
80: Pt$_{\rm{}10 }$Rh$_{\rm{}90}$ alloy systems. They found
81: a surface enrichment with Pt for both the systems. When
82: studied as a function of temperature this surface
83: enrichment showed a maximum at about 1000 K for both the
84: systems. This phenomenon was explained in terms of a
85: vibration entropy contribution to the enthalpy. Tsong et
86: al. [5] made a layer-by-layer analysis of a Pt-Rh alloy
87: tip at 1000 K to show that the topmost layer is
88: Pt-enriched and the second layer is Pt-depleted. The
89: distribution of Pt and Rh atoms depend on the annealing
90: temperature. Hirano et al. [6] , on the other hand,
91: showed that the presence of adsorbates like oxygen
92: and/or nitric oxide has a more pronounced effect on the
93: chemical restructuring process of the Pt-Rh surface
94: layers than the annealing temperature. While there have
95: been a large number of theoretical studies on this
96: segregation phenomenon in clean and gas-covered alloy
97: surfaces [4-7], the number of segregation studies for
98: supported particles are comparatively few. Some
99: theoretical investigations for Pt-Rh particles have been
100: made by Strohl and King [8], Schoeb et al. [9] and Yang
101: et al [10]. These works are based either on rigid-lattice
102: Monte-Carlo (MC) simulations or on molecular dynamics /
103: Monte-Carlo-correlated effective medium theory, where the
104: interaction energies were generated from a simplified
105: version of the non-selfconsistent
106: electron-density-functional correlated effective medium
107: (CEM) theory [11]. For the fcc cubo-octahedral particles
108: the atomic distribution at the corner, edge and fcc(100)
109: and (111)-like faces were obtained as a function of the
110: particle size and composition. A shell-by-shell
111: composition profile of the particles were not, however,
112: studied in these works. It is the main purpose of the
113: present work to investigate the shell-wise composition
114: of Pt-Rh nanoparticles as a function of temperature
115: and adsorbate coverage. For the 2406 fcc cubo-octahedral
116: Pt-Rh nanoparticles a Monte-Carlo simulation has been
117: carried out to analyse the atomic distribution in top
118: few layers . We show that the presence of adsorbate like
119: oxygen drastically changes the composition profile, while
120: the effect of temperature from 700K-1300 K on the
121: composition profile is relatively much small. These
122: studies are useful in understanding the structure and
123: catalytic properties of ceria-supported Pt-Rh catalysts
124: used in the catalytic converter . The plan of the paper is
125: as follows. In section 2, the theoretical model
126: developed earlier and used in the present work is
127: briefly outlined. The results for clean and
128: oxygen-adsorbed Pt-Rh nanoparticles are presented in
129: section 3. The conclusions are drawn in section 4.
130:
131:
132:
133:
134: \noindent
135: \textbf{2. Theoretical model}
136:
137:
138: \noindent
139: The theoretical model used in this work is the MC simulation
140: technique employed elsewhere in connection with surface segregation
141: in presence and in absence of adsorbed species[12-16]. One
142: determines the pair-bond energy between two nearest neighbour atoms
143: E$_{ij}$ in the form
144:
145:
146: \begin{equation}
147: E_{\rm{}ij} = w_{\rm{}ij }/Z + E_{\rm{}c}^{\rm{}i} (n)/n + E_{\rm{}c}^{\rm{}j} (m) /m
148: \end{equation}
149:
150: \noindent
151: where the indices i and j are the component metals A or B of the
152: bimetallic system AB ; and w$_{ij}$ is the interchange energy for
153: two dissimilar atoms. w$_{ij}$ = 0 if i = j and this interchange
154: energy is related to the excess heat of mixing. Z is the bulk
155: coordination number. E$_{c}^{i}$ (n) is the partial bond energy of the
156: i-atom with n nearest neighbours etc. Usually, the
157: coordination-dependent partial bond energy , E$_{\rm{}c}$$^{\rm{}i}$
158: (n) ( for n = 2 - 12), is calculated from an empirical formula
159: like
160:
161:
162:
163: \begin{equation}
164: E_{\rm{}c}^{\rm{}i} (n) = a^{\rm{}i} + b^{\rm{}i} n + c^{\rm{}i} n^{\rm{}2}
165: \end{equation}
166:
167:
168: \noindent
169: where a$^{\rm{}i}$, b$^{\rm{}i}$ and c$^{\rm{}i}$ are adjustable
170: constants that can be derived from experimental data of dimer
171: energy , surface energy and the cohesive energy of a metal
172: corresponding to n=2, 8 and 12 respectively. Once the pair
173: binding energy E$_{\rm{}ij}$ is evaluated it is the usual MC
174: procedure to compute the configuration energy for a particular
175: distribution of A and B atoms in the nanoparticle. And from various
176: such distributions an equilibrium distribution may be found giving
177: the shell-wise occupation of sites by A or B atoms. In presence of
178: adsorbates, there are additional energy terms to describe the
179: chemisorptive bond and the concomitant metal-metal pair-bond energy
180: variations. King and Donnelly [12] showed that since in Monte-Carlo
181: simulations only the differences in energies are calculated, it is
182: sufficient to add the difference in chemisorption energy of the
183: adsorbate on the two metals to the configuration energy. This has
184: been done in this work to study the role of adsorbed oxygen on the
185: composition profile.
186:
187:
188:
189:
190:
191: \noindent
192: \textbf{3. Results and Discussions}
193:
194:
195: \noindent
196: \textbf{ }We are interested in the composition profile of the outer
197: shells of the Pt-Rh nanoparticles since it is only the outer shells
198: which are important in catalysis and can be influenced by adsorbed
199: atoms. For convenience, we avoid the usual description of shell
200: structure where the shells are constructed by atoms equidistant from
201: the centre. As may be found from the work of Yang and DePristo [10]
202: that to describe a 200-atom fcc cubo-octahedral particle one needs
203: 11 shells. This concept may be very useful for describing the
204: stability and morphological structure of a nanoparticle. However, for
205: a large fcc cubo-octahedron particle of 2406 atoms as considered in
206: this work this description would require a large number of shells
207: which is very difficult to handle. In our treatment , the outermost
208: shell (we also call it `top shell') comprises all the atoms at the
209: corner sites, bridge sites, fcc(100) and fcc(111)-type faces. The
210: total number of such atoms is 752. If we peel off these atoms from
211: the outermost shell, the second shell would be exposed to the adsorbed
212: gases. The number of atoms in the second shell is 582. In this way, we
213: may have just eight shells with total number of atoms in the shells (
214: counting from the outermost shell inward) as 752, 582, 432, 302,
215: 192, 102, 38 and 6 respectively. The number of Pt and Rh atoms
216: are counted in each shell and the fraction of these atoms in the
217: shells are calculated. In this work we have calculated the fraction
218: of Pt and Rh atoms for a bulk Pt$_{\rm{}50}$Rh$_{\rm{}50}$ particle.
219:
220:
221:
222: \noindent
223: The values of the parameters used for the present MC simulation are
224: the following:
225:
226:
227:
228: Pt: a$^{\rm{} }$ = -0.4287398, b$^{\rm{} }$ = -0.0474969,
229: c$^{\rm{} }$ = 0.0035483 , E$_{\rm{}ad }$(O) = 85 kcals/mole
230:
231:
232: Rh: a$^{\rm{} }$ = -0.4744544, b$^{\rm{} }$ = - .0341727,
233: c$^{\rm{} }$ = 0.0028138 , E$_{\rm{}ad }$(O) = 102 kcals/mole
234:
235:
236:
237: \noindent
238: and w$_{\rm{}ij}$ /Z = -0.00691 eV. E$_{\rm{}ad}$(O) is the
239: chemisorption energy of oxygen atom on the metal.
240:
241:
242: \noindent
243: We have calculated the composition profile for T = 700 K, 1000 K and
244: 1300 K. The studies at T = 700 K is of particular importance
245: since the exhaust temperature at the three-way catalyst is of this
246: order. To study the effect of adsorbed gases we have considered
247: one coverage ( = 0.25 ) of adsorbed oxygen. This is because it is
248: the presence or absence of oxygen that controls the CO oxidation and
249: NO reduction in three-way catalysts. The calculation , however, can
250: easily be repeated for other coverages and other adsorbed gases . The
251: present work attempts to study the qualitative change of composition
252: profile as a function of temperature and adsorbate coverage.
253:
254:
255:
256: \noindent
257: Figure 1 shows the \% of Pt atoms in top four shells. The
258: top shell ( shell number 1) corresponds to the surface layer
259: and shows Pt-enrichment for all the three temperatures;
260: while the second and third shell shows Pt-depletion for the
261: three temperatures. The no segregation line is represented
262: by the dots. The behaviour from shell number 4 to higher (
263: moving towards the core of the particle ) is
264: divergent. However, for surface and catalytic properties
265: the top 2-3 shells are of real interest since they
266: constitute the surface layers . Besides, the top shell for a
267: 2406-atom particle represents 31\% of total atoms, sometimes
268: known as dispersion; while the top 2 , top 3 and top 4
269: shells represent 55\% , 73\% and 86\% of total atoms for
270: this particle respectively. If we compute the \% of Pt
271: atoms averaged over top 2, top 3 and top 4 shells, the
272: results would be as shown in Figure 2. The results are
273: computed for all the three temperatures. The results averaged
274: over all the eight shells would ,of course, be the bulk Pt
275: concentration which is 50\%. This is why all the curves meet
276: at shell number 8. It may be noted that for T = 700 K the
277: top shell is Pt-enriched; but the averages over top 2, top 3
278: and top 4 shells show Pt-depletion. For T = 1000 K and 1300
279: K, however, the averages over top 2, top 3 and top 4 shells
280: show Pt-enrichment. This signifies that the properties,
281: that depend on the composition of the top shell only, would
282: not change with temperature, since the shell is
283: Pt-enriched. However, the properties that depend on the
284: average composition of top 2 , top 4 or top 4 shells will
285: change with temperature. In Figure 3 we plot the same
286: shell-averaged Pt concentration as a function of
287: temperature. In addition to the features described above
288: for Figure 2, an important feature becomes obvious in
289: figure 3. It is that the Pt-enrichment shows a maximum at T
290: = 1000 K. This result agrees very well with the
291: experimental findings of Van delft et al. [4] for the
292: polycrystalline alloy surface. The authors [4] explained this
293: behaviour with a surface segregation model containing both
294: enthalpy and vibrational entropy contribution. It is likely
295: that for the present system of nano-sized Pt-Rh particles
296: also the entropy contribution may be responsible for the
297: temperature dependence of surface segregation. In Figure 4
298: we plot the \% of Pt atoms in presence of 0.25 monolayer of
299: adsorbed oxygen and compare them with the results for clean
300: system, shown already in Figure 1. It may be noticed that the
301: presence of 0.25 monolayer of oxygen has a very strong effect
302: on the composition of the top shell as well as on other
303: shells. Because of higher binding energy of oxygen on Rh
304: compared to that on Pt the Rh atoms are pulled to the top
305: shell at all the three temperatures. This segregation is so
306: strong that the rest of the shells of the particle become
307: Pt-enriched. On comparing the effect of temperature and
308: adsorbates on the shell composition it may be noted from
309: Figure 4 that the effect of temperature on the composition
310: profile is mild. This is obvious from the spreadth of the
311: composition for three temperatures in both the clean and
312: adsorbed cases. However, the large separation of the two
313: sets of curves ( each set consisting of the results for three
314: temperature), one for the clean case and the other for the
315: oxygen-adsorbed case, signifies that the effect of
316: adsorbates is much stronger on the composition. We believe,
317: these results can be exploited in the three-way catalyst
318: where Rh acts as the main component in reducing NO.
319: Furthermore, it may be mentioned that the present study
320: considers one bulk composition and one coverage of one
321: adsorbed species. This may be repeated for other
322: compositions, adsorbates and coverages as the situation
323: demands.
324:
325:
326:
327: \noindent
328: \textbf{4. Conclusions}
329:
330:
331: \noindent
332: A Monte-Carlo procedure has been utilised to compute the atomic
333: composition of a Pt-Rh nanoparticle in the form of shells
334: gradually peeled off from the outer layer. The composition profile
335: correctly predicts Pt segregation in the top layer ; but the
336: variation of the composition of the outer shells with temperature
337: is small. In contrast, the variation of the composition of the
338: shells with presence of adsorbed oxygen is very large.
339:
340:
341:
342:
343: \newpage
344: \noindent
345: \textbf{References}
346:
347: \noindent
348: [1] K. C. Taylor, \textit{In} Automotive Catalytic Converters,
349: Springer, Berlin(1984).
350:
351:
352: \noindent
353: [2] K. Taschner, \textit{In} Catalysis and Automotive Pollution
354: Control II, Ed. A. Crucq, Elsevier, Amsterdam,(1991),p-17.
355:
356:
357: \noindent
358: [3] C. Howitt, V. Pitchon, F. Garin and G. Maire, \textit{In}
359: Catalysis and Automotive Pollution Control III, Eds. A. Frennet, and
360: J. M. Bastin, Elsevier, Amsterdam (1995), p-149.
361:
362:
363: \noindent
364: [4] F. C. M. J. M. Van Delft, Surf. Sci. 189/190 (1987) 1129.
365:
366: \noindent
367: [5] T. T. Tsong, D. M. Ren and M. Ahmad, Phys. Rev. B38 (1988)7428.
368:
369: \noindent
370: [6] H. Hirano, T. Yamada, K. Tanaka, J. Sierra and B. E. Nieuwenhuys,
371: Vacuum 41(1990) 134; H Hirano, T. Yamada, K. Tanaka, J. Sierra and
372: E. Nieuwenhuys, Surf. Sci. 222(1989)L804; H Hirano, T. Yamada,
373: K. Tanaka, J. Sierra and B. E. Nieuwenhuys, Surf. Sci. 226(1990)1.
374:
375:
376: \noindent
377: [7] F. C. M. J. M. Van Delft, B. E. Nieuwenhuys, J. Sierra and
378: R. M. Wolf, ISIJ International, 29(1989)550.
379:
380:
381: \noindent
382: [8] J. K. Strohl and T. S. King, J. Catal. 116 (1989)540,
383: \textit{ibid}. 118(1989)53.
384:
385:
386: \noindent
387: [9] A. M. Schoeb, T. H. Raeker, L. Yang, X. Wu, T. S. King and
388: A. E. DePristo, Surf. Sci. Lett 278(1992) L125.
389:
390:
391: \noindent
392: [10] L. Yang and A. E. DePristo, J. Catal. 148 (1994)575.
393:
394: \noindent
395: [11] M. S. Stave, D. E. Sanders, , T. J. Raeker and A. E. DePristo
396: J. Phys. Chem. 93(1989) 1556.
397:
398:
399: \noindent
400: [12] T. S. King and R. G. Donnely, Surf. Sci.141(1984)417.
401:
402: \noindent
403: [13] B. C. Khanra and M. Menon, Chem. Phys. Lett. 305(1999)89.
404:
405: \noindent
406: [14] B. C. Khanra and M. Menon, Physica B 291(2000)368.
407:
408: \noindent
409: [15] A. De Sarkar, M. Menon and B. C. Khanra, Appl. Surf. Sci. (In press)
410:
411: \newpage
412:
413: \noindent
414: \textbf{{\large{}\textbf{Figure Caption}}}
415: \vskip 1.2cm
416: \noindent
417: \textbf{{\large{}Fig. 1}}{\large{}. MC-simulated \% of Pt atoms in
418: different shells of a 2406-atom
419: Pt}$_{\rm{}{\large{}50}}${\large{}Rh}$_{\rm{}{\large{}50}}${\large{}
420: nanoparticle for different temperatures .}
421:
422:
423:
424: \noindent
425: \textbf{{\large{}Fig. 2}}{\large{}. MC-simulated average \% of Pt
426: atoms in top, top 2, top 3 and top 4 shells of a 2406-atom
427: Pt}$_{\rm{}{\large{}50}}${\large{}Rh}$_{\rm{}{\large{}50}}${\large{}
428: nanoparticle for different temperatures.}
429:
430:
431:
432: \noindent
433: \textbf{{\large{}Fig.3}}{\large{}. MC-simulated average \% of Pt
434: atoms in top, top 2, top 3 and top 4 shells of a 2406-atom
435: Pt}$_{\rm{}{\large{}50}}${\large{}Rh}$_{\rm{}{\large{}50}}${\large{}
436: nanoparticle plotted as a function of temperature to demonstrate the
437: existence of a maximum at T = 1000 K.}
438:
439:
440:
441: \noindent
442: \textbf{{\large{}Fig. 4.}}{\large{} MC-simulated \% of Pt atoms in
443: different shells of a clean and oxygen-adsorbed 2406-atom
444: Pt}$_{\rm{}{\large{}50}}${\large{}Rh}$_{\rm{}{\large{}50}}${\large{}
445: nanoparticle . }
446:
447: \end{document}
448:
449:
450:
451:
452:
453: