1: \documentclass{article}
2: \usepackage{epsfig}
3: \usepackage{oljour}
4:
5: \begin{document}
6:
7: \journalname{zp}
8:
9: \title[en]{Theoretical study of structure and energetics
10: of gold clusters with the EAM method }
11:
12: \begin{author}
13: \anumber{1}
14: \firstname{Denitsa}
15: \surname{Alamanova}
16: \institute{Physical and Theoretical Chemistry}
17: \street{University of Saarland}
18: \number{Bld.\ B2 2}
19: \zip{66123}
20: \town{Saarbr\"ucken}
21: \country{Germany}
22: \tel{+49-681-302-3809}
23: \fax{+49-681-302-3857}
24: \email{deni@springborg.pc.uni-sb.de}
25: \end{author}
26:
27: \begin{author}
28: \anumber{2}
29: \firstname{Valeri G.}
30: \surname{Grigoryan}
31: \institute{Physical and Theoretical Chemistry}
32: \street{University of Saarland}
33: \number{Bld.\ B2 2}
34: \zip{66123}
35: \town{Saarbr\"ucken}
36: \country{Germany}
37: \tel{+49-681-302-4420}
38: \fax{+49-681-302-3857}
39: \email{vg.grigoryan@mx.uni-saarland.de}
40: \end{author}
41:
42: \begin{author}
43: \anumber{3}
44: \firstname{Michael}
45: \surname{Springborg}
46: \institute{Physical and Theoretical Chemistry}
47: \street{University of Saarland}
48: \number{Bld.\ B2 2}
49: \zip{66123}
50: \town{Saarbr\"ucken}
51: \country{Germany}
52: \tel{+49-681-302-3856}
53: \fax{+49-681-302-3857}
54: \email{m.springborg@mx.uni-saarland.de}
55: \end{author}
56:
57:
58: \corresponding{vg.grigoryan@mx.uni-saarland.de}
59:
60:
61: \abstract{
62: Using the Embedded Atom Method as developed by Voter and Chen in combination
63: with the {\it variable
64: metric/quasi-Newton} and our own {\it Aufbau/Abbau} methods, we have identified the
65: three most stable isomers of Au$_N$ clusters with $N$ up to 150. For
66: the first time clusters with tetrahedral symmetry are found to form the
67: ground states of Au$_{17}$ and Au$_{34}$. The Au$_{54}$ {\it icosahedron}
68: without a central atom and the Au$_{146}$ {\it decahedron} are found to be
69: particularly stable, whereas
70: the highly symmetric second and third Mackay icosahedra that could have been obtained
71: for $N$ = 55 and
72: 147, respectively, do not correspond to the particularly stable structures.
73: The three lowest-lying isomers of Au$_{55}$ and Au$_{147}$
74: are low-symmetrical structures. Various structural and energetic properties are
75: analysed, such as stability function, occurrence of magic-sized clusters,
76: construction of icosahedral and {\it fcc} shells, and cluster growth.}
77:
78: \keywords{Gold clusters, embedded-atom-method calculations, structure, stability}
79:
80: \dedication{This work is dedicated to Prof.\ Dr.\ Wolf Weyrich, Physical Chemistry,
81: University of Konstanz, Germany, on the occasion of his 65th birthday.}
82:
83: \received{x}
84: \accepted{x}
85: \volume{x}
86: \issue{x}
87: \class{x}
88: \Year{x}
89:
90: \maketitle
91:
92: \section{Introduction}
93:
94: Since the first part of the 20th century it has been recognized that theoretical
95: studies can constitute an important ingredient of science, first through the
96: dedication of chairs to the field of theoretical physics and later through the
97: establishment of chairs in theoretical chemistry. Simultaneously, by defining
98: `theory' as an independent part of science, the interactions between theory and
99: experiment have in many cases been reduced. Persons who unite these two parts are,
100: therefore, of immense importance and, simultaneously, able to contribute to science
101: in a way that is unmatched by most colleagues. A person, occupying a chair of
102: physical chemistry, and being active in both experimental and theoretical studies
103: of the properties of matter, is, accordingly, unique. Such a person is Wolf Weyrich,
104: who has made significant contributions to the understanding of the properties of a
105: large range of systems, going all the way from smaller molecules to extended solids.
106: In the later years his interests have also turned towards the properties of
107: nanoparticles that somehow lie in between the materials of his earlier interests.
108: Metal nanoparticles are thereby of central interest. Therefore, it is with pleasure
109: that we dedicate the present work to Wolf Weyrich on the occasion of his 65th birthday.
110:
111: Clusters form an important link between isolated atoms and
112: molecules at one extreme and bulk solids at the other. Their large
113: surface-to-volume ratio gives them
114: unique physical and chemical properties. On the other hand, the combination of
115: finite with large size makes it difficult to characterize and analyze their properties
116: in detail. From a theory point of view the central difficulty lies in the determination
117: of the structure of the system of interest and only through comparison with experimental
118: information more definite statements about their properties can be made.
119:
120: Gold clusters represent some of the mostly studied clusters (for a detailed
121: discussion, see Ref.\ \cite{pp04,bf05}). They have recently been investigated in
122: connection with the synthesis of nanostructured materials and devices
123: \cite{rw96,as99,hh03,lds03,hb04}. Their structural and energetic properties
124: have been studied with High-Resolution Electron
125: Microscopy (HREM) and various spectroscopic techniques
126: \cite{kjt92,hh94,ap95,mma97,tgs97,cc97,kk98,bp98,leh99,vas00,mv01,nv01,hh01,
127: vt03,rs04,gm04,bkm04,hm04}. Literature concerning small Au$_N$ clusters
128: is enriched with numerous investigations
129: based on density-functional methods
130: \cite{ll97,odh97,gbp99,hh99,ga00,hl00,jms00,hh02,jo02,jw02,jw03,zw03,by03,
131: mj04,sc04,rmo05,avw05,lx04,fr05} that are not yet capable of giving a definite
132: answer to the problem at what cluster size the structural 2D -- 3D transition
133: occurs. Recent studies combining theory and experiment
134: \cite{hy03,sg02,ff02,mn05} show that the gold clusters are planar at least up
135: to $N$ = 7 for Ref.\ \cite{sg02}, or up to $N$ = 12, according to
136: H\"akkinen \cite{hy03} and Furche {\it et al.} \cite{ff02}.
137:
138: However, global
139: structure optimization is difficult when using {\it ab initio} methods already at very
140: small cluster sizes. Nevertheless, some studies in this direction exist.
141: Thus, the authors of Ref.\ \cite{odh97,mj04} performed
142: density functional calculations on clusters containing more than 30 atoms,
143: relaxing selected high-symmetric configurations. Alternatively, the global
144: optimizations of larger clusters are all based on approximate methods like
145: molecular dynamics
146: \cite{fe91,ha98,wdl98,clc98,clc99,yc01,sp01,scl02,fb02,jr04,ul04,yw04} and
147: semiempirical potentials like the EAM
148: \cite{cc97,clc97,rb99}, Sutton-Chen \cite{dw98}, Murrell-Mottram
149: \cite{ntw00,ntw02}, or the many-body
150: Gupta potential \cite{ilg96,ilg98,ilg99,km99,tl00,sd02}. Using these methods,
151: unbiased structure optimizations
152: were performed up to the 80-atom cluster. Medium-sized
153: clusters ( $ 80 \le N \le 150$) have hardly been
154: studied. Besides the first-principles study of H\"aberlen {\it
155: et al.} \cite{odh97} and the EAM calculations by Cleveland {\it et
156: al.} \cite{cc97,clc97,rb99}
157: considering particular structural motifs, there exists essentially no further
158: investigation on the clusters in this size range.
159:
160: In most of the studies, special attention is paid to the so-called
161: `magic-numbered' clusters, that
162: possess closed electronic and/or geometric shells. Various studies on the
163: smallest `magic' cluster Au$_{13}$ have identified the formation of an
164: icosahedron \cite{ll97,odh97,dw98,ntw00,km99,sd02}. Only the authors of
165: Ref.\ \cite{jo02,jw02} found a
166: disordered structure as the lowest-lying isomer for this cluster size. On the other
167: hand, semiempirical potentials \cite{clc97,dw98,ntw00,sd02} and the
168: density functional study by H\"akkinen {\it et al.} \cite{hh99} on the
169: Au$_{38}$ cluster predict the truncated octahedron to be the
170: global minimum for this cluster size. However, on the basis of first-principles
171: and Gupta potential calculations, the
172: authors of Ref.\ \cite{ilg99,tl00} state that a disordered structure is
173: actually lower in energy than the symmetric. Ultimately, it may be suggested that
174: the obtained structure depends sensitively on the type of the
175: potential, since by using another
176: form of the same potential, Darby {\it et al.} \cite{sd02} found a truncated
177: octahedral structure to be the global minimum
178: of Au$_{38}$. The situation is more clear for the Au$_{55}$ and Au$_{75}$
179: clusters, where a disordered
180: structure \cite{dw98,ilg96,ilg98,km99,tl00,sd02} or a Marks decahedron
181: (m-D$_{5h}$) \cite{ilg98,km99} seem to constitute the global minima.
182:
183: Several groups have performed calculations on larger clusters by minimizing the
184: total energy of initially chosen
185: symmetric structures, although it may be feared that the structures may not be those of
186: the global total-energy minima. Moreover, only few cluster sizes
187: were studied --- the octahedral Au$_{79}$ \cite{clc97,dw98} and Au$_{140}$
188: \cite{clc97}, the decahedral
189: Au$_{101}$, Au$_{116}$, and Au$_{146}$ \cite {clc97}, and the icosahedral
190: Au$_{147}$ \cite
191: {odh97,clc97}. The structures and energetics of the clusters between these high
192: symmetrical ones remain scarcely investigated.
193:
194: The purpose of the present study is to carry out unbiased calculations on small and
195: intermediate gold clusters,
196: and, subsequently, to investigate the occurrence of magic clusters and the growth
197: patterns, especially for larger clusters. To our knowledge, two earlier studies
198: on gold clusters with the same version of EAM have been performed previously, i.e., the
199: work of Rey and
200: coworkers \cite{cr93,cr94} and that of Sebetci and G\"uven\c{c} \cite{as04}.
201: The first study considered only the energetics and stability of small gold
202: clusters ($2 \le N \le 23$). Sebetci {\it et al.} used a
203: basin-hopping Monte Carlo minimization approach to find the global minima of
204: Al$_N$, Au$_N$, and
205: Pt$_N$ clusters with $N \le 80$. The total energies, point groups, and
206: structural assignments were presented.
207:
208: In the present study the structure and energetics of the three most stable
209: isomers of small and
210: medium-sized Au$_N$ clusters with $2 \le N \le 150$ have been determined for
211: each cluster size by
212: using a combination of the embedded-atom method in the version
213: of Voter and Chen \cite{vo87,vo93,vo95}, the {\it variable metric/quasi-Newton}
214: method,
215: and our own {\it Aufbau/Abbau} method. The paper is organized as follows.
216: In Sec.\ 2 we briefly outline the embedded-atom method, and in Sec.\ 3 we
217: present our structural-determination methods. The main results are given in
218: Sec.\ 4, and a brief summary is offered in Sec.\ 5.
219:
220:
221: \section{The Embedded Atom Method}
222:
223: The main idea of the EAM was initially presented by Daw, Baskes, and Foiles (DBF)
224: \cite{da83,da84,fo86} in 1983--1986, and since then the generality of
225: the functions of the EAM
226: of DBF has been successfully tested through numerous applications to different
227: systems of metals and alloys, including defects, surface and
228: interface structures, surface and bulk phonons, etc. In a previous paper
229: \cite{our} we reported results for the global minima of Ni$_N$, Cu$_N$, and
230: Au$_N$ clusters with up
231: to 60 atoms, obtained with two different versions of the EAM. There we discussed the
232: incapability of the so-called DBF version of
233: EAM to describe properly the properties of the smallest gold clusters,
234: which could be related to the parameterization of the potential
235: only to bulk properties. The version developed by Voter and Chen
236: \cite{vo87,vo93,vo95} takes into account also the properties of the dimer,
237: which makes this method more suitable for the description of the
238: smallest clusters.
239: Accordingly, in this study we use this version of the EAM for the calculation of
240: the total energy of a given cluster.
241:
242: The principle of the method is to split the total energy of the
243: system into a sum over atomic energies:
244: \begin{equation}
245: E_{tot}=\sum_{i}^{N}E_i.
246: \label{1}
247: \end{equation}
248: The embedding energy is obtained by considering each atom as an impurity
249: embedded into a host
250: provided by the rest of the atoms. The electron-electron interaction is
251: presented as sums of short-ranged, pair potentials. Accordingly,
252: \begin{equation}
253: E_{tot}=\sum_i F_i(\rho_i^h)+{\frac{1}{2}\sum_{i\ne j}\phi_{ij}(r_{ij})}
254: \label{2}
255: \end{equation}
256: where $\rho_i^h$ is the local electron density at site $i$, $F_i$ is the
257: embedding energy, i.e.,
258: the energy required to embed an atom into this density, and $\phi_{ij}$ is a
259: short-range potential between atoms $i$ and $j$ separated by distance $r_{ij}$.
260: The pair potential, according to the Voter-Chen version, is taken to be a Morse
261: potential,
262: \begin{equation}
263: \phi(r)=D_M[1-e^{-\alpha_M(r-R_M)}]^2-D_M
264: \label{3}
265: \end{equation}
266: where the three parameters, $D_M$, $R_M$, and $\alpha_M$, define depth, position
267: of the minimum,
268: and a measure of the curvature at the minimum, respectively. The local density
269: at site $i$ is assumed being a superposition of atomic electron densities:
270: \begin{equation}
271: \rho_i^h=\sum_{j \,(\ne i)}\rho_j^a(r_{ij})
272: \label{4}
273: \end{equation}
274: where $\rho_j^a(r_{ij})$ is the spherically averaged atomic electron density
275: provided by atom $j$
276: at the distance $r_{ij}$. The density function is taken as the density of a
277: hydrogenic $4s$ orbital:
278: \begin{equation}
279: \rho(r)=r^6[e^{-\beta r}+2^{9}e^{-2\beta r}]
280: \label{5}
281: \end{equation}
282: where $\beta$ is an adjustable parameter. Because $r^6e^{-\beta r}$ turns over
283: at short $r$, the
284: second term has been added to maintain the monotonically decreasing character
285: of $\rho(r)$ at
286: shorter $r$. This $4s$ orbital density, appropriate for Ni and Cu, also works
287: well for gold.
288: To ensure that the interatomic potential and its fist derivatives are
289: continuous, both $\phi(r)$
290: and $\rho(r)$ are cut off at $r$=$r_{\rm cut}$.
291: In the fitting procedure, the five parameters defining $\phi(r)$ and $\rho(r)$
292: ($D_M$, $R_M$,
293: $\alpha_M$, $\beta_M$ and $r_{\rm cut}$) are optimized by minimizing the
294: root-mean-square deviation
295: between the calculated and reference properties of different selected systems like
296: molecules, surfaces, solids, and defects. Because
297: $F(\rho^h)$ is
298: redefined for each choice of the parameters, the potential always gives perfect
299: agreement with
300: experimental values such as $a_0$, $E_{\rm coh}$ and the bulk modulus $B$. The
301: reference properties are
302: the three cubic elastic constants ($C_{11}$, $C_{12}$ and $C_{44}$), the
303: unrelaxed vacancy
304: formation energy ($E_{vac}^f$), and the bond length ($R_e$) and bond strength
305: ($D_e$) of the
306: diatomic molecule. The values of $\rho_i^a$, $\phi_{ij}$ and $F_i(\rho_i)$ that
307: were used by the Voter-Chen version, are
308: available in numerical form for Ni, Pd, Pt, Cu, Ag, Au and Al.
309:
310: Our reason for choosing the EAM was dictated by the good agreement to
311: experiment, as well as to first principles
312: calculations, and last but not least by the high computational efficiency
313: allowing one to investigate clusters with more than 100 atoms without severe
314: constraints on the initial geometry, which is
315: impossible with first principles methods.
316: In a previous work \cite{our} we performed
317: calculations on smaller gold clusters with $2 \le N \le 60$ atoms comparing the
318: EAM of Daw, Baskes, and Foiles (DBF) and the version used in this work, and we
319: found that the DBF overestimated the
320: binding energy of the dimer by 209{\%} and underestimated the bond distance by
321: 37{\%}. For comparison, the EAM of Voter and Chen gives dimer binding energy
322: corresponding to 99.6{\%} of the experimental value, and bond distance that is
323: 92.2{\%} of the experimental value. For this reason we
324: chose to work with the Voter-Chen version that describes correctly the dimer
325: properties.
326:
327: \section{Structure optimization}
328:
329: Using expression (\ref{1}) we can calculate the total energy of
330: any cluster with any structure as a function of structure, i.e., of the atomic coordinates
331: $\{ {\vec R}_i\}$, $E_{\rm tot}({\vec R}_1,{\vec R}_2,\, \dots \, ,{\vec
332: R}_N)$.
333: In order to obtain the closest local total-energy minimum we use the
334: {\it variable metric/quasi-Newton} method \cite{numer92}.
335:
336: For searching the global minima we have developed our own {\it Aufbau/Abbau}
337: method that is described in details in previous works \cite{cpl03,prb04}. It
338: consists of the following steps:
339:
340: 1) We consider two cluster sizes with $N$ and $N+K$ atoms with $K\simeq 5-10$.
341: For each of those we randomly generate and relax a large set of structures,
342: from which those with the lowest total energy are selected.
343:
344: 2) One by one, each of the $N$ atoms is displaced randomly, and the closest
345: local minima is determined. If the new structure has a lower total energy than
346: the original one, this new one is kept, and the old one discarded. This is
347: repeated approximately $1000-2000$ times depending
348: on cluster size.
349:
350: 3) This leaves us with two `source' clusters, Au$_N$ and
351: Au$_{N+K}$ with their lowest total energies. One by one an atom is added
352: at a random position to the structure with $N$ atoms
353: (many hundred times for each size), and the structures are relaxed. In
354: parallel,
355: one by one an atom is removed from the structure with $N+K$ atoms --- for each
356: intermediate cluster with $N'$ atoms we consider all $N'+1$ possible
357: configurations, that one can obtain by removing one atom from the Au$_{N'+1}$
358: cluster.
359: From the two series of structures for $N\le M\le N+K$ those structures of the
360: lowest energies are chosen and these are used as seeds for a new set of
361: calculations. First, when no lower total energies are found in the two
362: sets of calculations, it is assumed
363: that the structures of the global-total-energy minima
364: have been identified.
365:
366: \section{Results and Discussion}
367:
368: \subsection{Small gold clusters}
369:
370: Like all other semiempirical potentials, the one of the embedded-atom method
371: does not include explicitly the electrons and their orbitals.
372: Therefore, such a potential tends to prefer high-symmetry, compact structures,
373: whereas structures of lower symmetry that can be explained through electronic
374: effects are not found. As a result, for the smallest gold clusters,
375: where spin-orbit interactions
376: play an important role, our global-minima structures are compact (see Table
377: \ref{tab01}), and the planar structures (that are believed to be those of the
378: true total-energy minima) only metastable. In the size range $N$
379: = 4 -- 7 first principles studies obtain these 3D configurations as
380: higher-lying isomers, which can serve as an example of how the inclusion of
381: electronic effects can change the energetic ordering of the isomers. On the
382: other hand, the addition of electronic effects in the semiempirical potentials would
383: restrict their use only to small and relatively larger clusters with pre-chosen structures. An
384: appropriate choice in this respect could be the Density Functional Tight-Binding methods (DFTB)
385: that include explicitly the electrons and are computationally more efficient
386: than the common density functionals. Actually, in a recent study \cite{cole} we demonstrated
387: the important role that the electronic effects can play for the
388: binding energy and the stability functions. However, except for $N$ = 4, where
389: a rhombus was the lowest-lying isomer according to the DFTB method, all the
390: ground state structures for the smallest gold clusters Au$_5$ -- Au$_9$ had 3D
391: shapes. On the other hand, even the most recent density-functional studies are
392: still not in agreement at which cluster size the structural transition 2D -- 3D
393: occurs. According to the LDA study of Wang {\it et al.} \cite{jw02}, it is the
394: pentagonal bipyramid that forms the global minimum of Au$_7$. Remacle and
395: Kryachko \cite{fr05} suggested that gold clusters are planar at least up to
396: $N$ = 9, while Walker \cite{avw05} predicted that the transition occurs at
397: Au$_{11}$. Using ion mobility measurements and {\it ab-initio} molecular
398: dynamics Kappes {\it et al.} \cite{sg02,ff02} found that the 3D transitions
399: occur at Au$_{12}^-$ and Au$_8^+$. The same group studied the adsorption of CO
400: on isolated gold cluster cations in the size range $N$ = 1 -- 65. The smallest
401: clusters with $4 \le N \le 6$ as well as Au$_8$ were found to be planar, while for
402: Au$_7$ the global minimum was a 3D structure, but not a bipyramid, in contrast
403: to the results of Wang {\it et al.} \cite{jw02}. In a combined experimental and
404: theoretical study H\"akkinen and coworkers \cite{hy03} confirmed the 2D -- 3D
405: transition at Au$_{12}^-$, however, Xiao and Wang \cite{lx04} suggested that for the
406: neutral clusters this transition occurs first at Au$_{15}$. In most of
407: the cases the planar structures are competing with 3D isomers, and the
408: energetic differences are insignificant, which in turn means that the
409: ordering of the isomers depends strongly on the used functional and the
410: starting conditions. For example, H\"akkinen {\it et al.} \cite{hh02} compared
411: the global minima of relativistic and nonrelativistic Au$_7^-$ clusters and
412: found that for the nonrelativistic gold the lowest-lying isomer was a capped
413: octahedron that corresponds to our second isomer for this cluster size. At
414: larger cluster sizes the potential
415: used in this study yields results in agreement with density-functional and
416: experimental studies. The study of H\"akkinen {\it et al.} \cite{hh99} on the
417: Au$_{38}$ cluster predicted the truncated octahedron to be the global minimum,
418: and a recent experiment \cite{hm04} showed that the Au$_{55}^-$ cluster most
419: probably is not an icosahedron, but a structure with a low symmetry, in
420: agreement with our results.
421:
422: In summary, we can conclude that although our results for the smallest gold
423: clusters correspond to higher-lying isomers within the first principles
424: methods, due to the lack of electronic effects, our method is sufficiently
425: accurate in describing the larger gold clusters with $N>9$, where most probably
426: the planar structures begin to compete with 3D configurations.
427:
428: \subsection{Energetic properties}
429:
430: The high stability of the so-called `magic-numbered' clusters has become a
431: subject of great interest
432: in connection with its relevance in the medicinal and the colloidal
433: chemistry, as well as in the production of catalysts and high-tech nanomaterials.
434:
435: In Fig.\ \ref{bind} we show the binding energy per atom for the
436: global-minima structures, as well as the difference between the
437: total energies of the lowest-lying isomers
438: obtained by us and those found by Sebetci {\it et al.} \cite{as04} using exactly
439: the same potential for the interatomic interactions. One
440: can see that the latter difference increases almost linearly with the number of atoms
441: and has its maximum at $N$ =
442: 79, where we obtained a truncated octahedron in contrast to the structure with
443: D$_{3h}$ symmetry found by Sebetci and G\"uven\c{c}. Except for few cases,
444: the total-energy difference is marginal (about 5 meV/atom, which may be due to
445: numerical differences), giving support for the quality of both theoretical
446: approaches in optimizing the structure.
447: At $N$ = 52 they obtained an uncentered
448: icosahedron-like structure with C$_{2h}$ symmetry, that lies energetically
449: between our second and third lowest isomers.
450:
451: In order to identify the particularly stable clusters we have considered the following
452: criteria. The clusters can be considered as very stable if their binding energy
453: per atom is much
454: larger than that of the two neighboring clusters. This can be quantified
455: through the stability function,
456: $E_{\rm tot}(N+1.1)+E_{\rm tot}(N-1.1)-2E_{\rm tot}(N.1)$, where $E_{\rm
457: tot}(N.k)$ is the
458: total energy of the energetically $k$-lowest isomer of the Au$_N$ cluster. This
459: function, that has maxima for particularly stable clusters, is shown in Fig.\
460: \ref{stability}.
461: Here we can identify a large number of particularly stable clusters, i.e.,
462: so-called magic
463: clusters. These are found for
464: $N$ = 4, 6, 10, 13, 15, 17, 23, 28, 30, 36, 38, 40, 42, 45, 49, 54, 58, 61,
465: 64, 66, 68, 73, 75, 77, 79, 82, 84, 89, 92, 95, 101, 109, 111, 116, 118, 124,
466: 128, 133, 135, 140, 144, and 146. The most pronounced peaks (marked in the figure)
467: occur at $N$ = 13, 30, 40, 54, 75, 79, 82, 124, 133, 140, and 146. In agreement with
468: Sebetci and G\"uven\c{c} \cite{as04},
469: the 54-atom icosahedron without a central atom is found to represent a
470: magic-numbered cluster, whereas the Au$_{55}$ cluster does not.
471: The latter possesses a distorted icosahedral structure with C$_{3v}$
472: symmetry, lying 0.374 eV lower than the perfect icosahedron,
473: 2.9 eV lower than the decahedron, and 3.27 eV lower than
474: the cuboctahedron. In our study, all the three lowest-lying
475: isomers of Au$_{55}$ have lower energy than
476: the symmetric structures, in agreement with previous studies where disordered
477: configurations were
478: found as global minima for Au$_{55}$ \cite{ilg96,ilg98,tl00,sd02}. For Au$_{38}$
479: and Au$_{75}$, a
480: cuboctahedron \cite{hh99,dw98,ntw00,sd02,as04} and a Marks decahedron
481: \cite{dw98,ilg98,km99,as04} were
482: obtained, in agreement with first-principles and semiempirical studies.
483: However, two studies employing
484: the many-body Gupta potential identified amorphous structures as those of the global minima
485: of Au$_{38}$
486: \cite{ilg99,tl00}, which is most probably due to the parameterization of the
487: potential, since Darby {\it et al.} \cite{sd02} found an octahedron as the
488: lowest-lying isomer by using another version of the same potential.
489:
490: Another striking result of our study is that the 146-atom Marks decahedron represents
491: a peak in the stability function, whereas
492: the Au$_{147}$ icosahedron does not. According to our study, the third Mackay
493: icosahedron lies 2.89 eV lower than
494: the cuboctahedron, which in turn is 2.53 eV lower than the decahedron, but 0.37 eV
495: higher than a disordered structure
496: with partly decahedral construction. To our knowledge, this is the first study
497: predicting a disordered global minimum for the Au$_{147}$ cluster.
498:
499: According to our other criterion for a particularly stable cluster, such a
500: cluster occurs if the energy difference between the two energetically lowest
501: isomers $E_{\rm tot}(N.2)-
502: E_{\rm tot}(N.1)$ is large. This energy difference is shown in Fig.\
503: \ref{diff}, and comparing to Fig.\ \ref{stability} we can see that many of the
504: clusters that are particularly stable according to the first criterion are
505: stable also according to the second one.
506:
507: \subsection{Structural properties}
508:
509: In this subsection, instead of discussing in particular the structures of the
510: individual clusters, we shall introduce different quantities that are devised
511: to reduce the available information to some few key numbers. The theoretical
512: background of the descriptors used in this subsection was introduced by us in a
513: previous work \cite{prb04}.
514:
515: The shape analysis, based on the eigenvalues of the matrix with the moments of
516: inertia and whose results are shown in Fig.\ \ref{shape}, separates the clusters into being
517: overall spherical, more cigar-like shaped, or more lens-like shaped.
518: One can see that only few clusters have a spherical shape (these are
519: found for the energetically lowest isomer for $N$ = 4, 6, 13, 17, 34, 38, 54,
520: 79, and 140, and
521: for the next one for $N$ = 42 and 116), all of them corresponding to
522: high-symmetrical isomers
523: (cf.\ Table \ref{tab01}) and, for the lowest-energy isomer, most of them to the
524: class of magic clusters.
525: It is interesting that the average value follows more or less
526: the same curve for all the three isomers, with some deviations at
527: $N$ = 130, 146, and 147. Also the
528: largest differences show a similar behaviour, except for some few cases mainly
529: for $N$ below 40 and between 80 and 85. Therefore, except when the eigenvalues
530: are all very similar (which occurs
531: for $N$ around 50, 70, 100, 116, and 140), the overall shape
532: (i.e., lens- or cigar-like) is the same for all three isomers.
533:
534: The construction of atomic shells can be easily seen from the distribution of
535: radial distances (i.e., the distance for each individual atom to the center of mass) shown
536: in Fig.\ \ref{radial} for the ground state structures as function of the
537: cluster size. Up to $N$ around 50, no trends can be identified, with an exception
538: around $N=13$. But for
539: $N$ just above 50 a clear tendency towards shell construction can be seen for
540: the first isomer. This corresponds to the formation of the Au$_{54}$
541: icosahedral cluster. Also for $N$ close to 110 and around 140 shell
542: constructions for the lowest-lying isomer are observed. In the latter case,
543: this corresponds to the formation of an octahedron. The radial distributions
544: for the second and the third isomers are not shown, as they are quite similar
545: to that for the first isomer. Particular shell constructions are found
546: only for highly symmetrical clusters corresponding to $N$ = 42, 48,
547: 80, 101, 116, and around Au$_{130}$ for the second isomer, and around $N$ = 40,
548: 60, 116, and 130 for the third isomer.
549:
550: In Fig.\ \ref{koord} we show the average and minimal coordination numbers
551: and the average bond lengths of the clusters. We define two atoms as being
552: bonded if their interatomic distance is less than 3.49 \AA,
553: which is the average value between the nearest-neighbour distance (2.89\
554: {\AA})
555: and the next-nearest-neighbour distance (4.08\ {\AA}) in bulk Au. Moreover, we
556: distinguish between inner atoms with a coordination number of 12 or larger and
557: surface atoms with a coordination number less than 12.
558:
559: Fig.\ \ref{koord}(a) presents the average coordination number as a function of
560: $N$. A saturation towards
561: the bulk limit of 12 is seen, although one has to remember that even for the
562: largest cluster
563: of our study 94 out of 150 atoms are characterized as surface atoms. Also, the
564: function increases in general with the size of the system, with
565: oscillations in particular for the clusters with $N$ = 17 and 18, which is
566: due to the formation of a tetrahedron for Au$_{17}$, and a structure with
567: C$_{4v}$ symmetry at $N$ = 18, respectively. The latter has already earlier been obtained
568: with the EAM method (see Ref.\ \cite{as04}), but it is the first time
569: that a tetrahedral configuration is found for the Au$_{17}$ cluster.
570:
571: The minimum atomic coordination for each cluster size is shown in Fig.\
572: \ref{koord}(b). The existence of low-coordinated atoms, i.e with coordination
573: numbers of 3 or 4, could point to the occurrence of a cluster growth, where
574: extra atoms are added to the surface of
575: the cluster, whereas higher coordination numbers could indicate a growth
576: where atoms are inserted inside the
577: cluster, or, alternatively, upon a strong rearrangement of the surface atoms.
578: The latter is the case for the gold clusters, with few exceptions at $N$ = 14, 17,
579: 18, 78, 83, and 134, where lower coordinations are found. The lowest
580: coordination corresponding to Au$_{14}$ is in connection with the
581: formation of an icosahedron plus one additional atom on the surface. At $N$ =
582: 17 and 18, some structural changes take place, as discussed above. Au$_{78}$ and
583: Au$_{83}$ correspond to structures with a decahedral motif capped with one
584: additional atom. This is also the case for Au$_{134}$ where the C$_{2v}$
585: symmetry of the decahedral structure corresponding to $N$ = 133 is lowered by the
586: addition of an atom to the surface.
587:
588: Fig.\ \ref{koord}(c) shows the average bond length as a function of the cluster
589: size. The dashed line corresponds to the bulk value of 2.89\ {\AA}. The
590: average bond length for all the structures is smaller than the bulk value,
591: especially for Au$_{17}$ and Au$_{18}$, where more compact structures are
592: formed. However, this property approaches the bulk value faster than the
593: average coordination number.
594:
595: One important issue in many of the molecular dynamics studies on gold clusters is
596: to identify how the clusters grow and if the cluster with $N$ atoms could be
597: derived
598: from the one with $N-1$ atoms simply by adding one atom. In order to quantify
599: this possibility we use the concept of similarity functions,
600: introduced by us earlier \cite{cpl03,prb04}.
601:
602: The similarity function $S$, shown in Fig.\ \ref{similar}(a), approaches 1 if the
603: Au$_N$ cluster is very similar to the Au$_{N-1}$ cluster plus
604: an extra atom. We see indeed that for $N$ up to around 50,
605: $S$ is significantly different from 1, confirming that in this range
606: the growth is complicated. The most pronounced peaks occur at $6<N<9$,
607: $15<N<20$, 34, 38, 39, 52, 56, 79, 80, 85, 111, 126, 140, 141, and $145< N
608: <147$. Many of these correspond to highly symmetrical clusters, however some of
609: the clusters with larger peaks ($N$ = 39, 56, 62, 85, 111, 126, 141, and 145)
610: have lower symmetry. The octahedral Au$_{38}$ and the low-symmetrical
611: Au$_{39}$ are structurally very different from their $N$-1-atom neighbours.
612: Au$_{56}$ marks the end of the icosahedral shell built between Au$_{52}$ and
613: Au$_{55}$, and the clusters resume their disordered growth. The octahedral
614: Au$_{61}$ is followed by the disordered Au$_{62}$, and the decahedral
615: Au$_{85}$ comes after the disordered Au$_{84}$. Between the decahedral
616: Au$_{110}$ and Au$_{112}$ lies the disordered Au$_{111}$. The addition of one
617: atom to the disordered Au$_{125}$ leads to the formation of an unfinished but
618: regular decahedron at $N$ = 126. The decahedral Au$_{141}$ comes immediately
619: after the octahedron corresponding to $N$ = 140. Although Au$_{144}$ has
620: partly decahedral construction, its $N$+1-atom neighbour is disordered. It
621: seems that for each cluster size there is a rearrangement of the gold atoms,
622: and no particular growth motif can be identified. This, in turn, means that the
623: cluster growth is very complicated and it is difficult to consider it as an
624: one-by-one atom addition.
625:
626: Finally, some selected, high-symmetry clusters are shown in Fig.\ \ref{pict}.
627:
628: \section{Summary and Conclusions}
629:
630: We have determined the three energetically lowest isomers of gold
631: clusters in the range $2 \le N \le 150$ by using a combination
632: of the embedded-atom method in the version of Voter and Chen
633: (for the calculation of the total energy for a given structure), the
634: {\it variable metric/quasi-Newton} method (for the determination of the closest
635: total-energy minimum), and our own {\it Aufbau/Abbau}
636: method (for the determination of the global total-energy minimum). Although the
637: calculations
638: provide a large amount of information for each individual cluster, instead of
639: discussing each cluster separately, we focused on identifying general trends
640: such as total energy per
641: atom, overall symmetry and shape, average bond length and coordination number,
642: and similarity with $N-1$-atom clusters.
643:
644: The version of EAM used in the present
645: calculations is parameterized to bulk, as well as to the dimer properties,
646: which allows it to describe properly the properties of the smaller gold
647: clusters.
648:
649: This study predicts a number of particularly stable clusters, i.e.,
650: `magic-numbered' clusters that in
651: many cases are in agreement with results obtained by first principles and other
652: semiempirical studies when such exist, but the advantage of our study is that
653: the structures were obtained by using a completely
654: unbiased approach. These magic numbers were clearly visible both in the
655: `stability function'
656: and in the total-energy difference between the energetically lowest and
657: higher-lying isomers.
658:
659: We also found that even for our largest clusters the binding energy per atom has
660: still not converged to the bulk limit. Similarly, the average coordination
661: number is far from the bulk value, but higher than for nickel clusters, where
662: several structures with shell constructions and corresponding low coordination
663: numbers were formed \cite{prb04}. The average bond distance for gold has not
664: reached the bulk value, due to the rearrangement of the atoms for each
665: cluster size that leads to the formation of very compact structures.
666:
667: The shape analysis showed that roughly spherical clusters corresponded mainly
668: to the energetically lowest isomer, but in some cases also to the
669: second-lowest one, and that these often belong to particularly stable
670: structures.
671:
672: By analysing the distribution of radial distances as a function of the cluster
673: size we could
674: identify a region with $N$ around 55, where a shell construction was formed.
675: Comparing to previous results for nickel clusters \cite{prb04}, where clear
676: shell
677: constructions were formed at $N$ around 13, 55, and 147, here the atoms
678: rearrange for each global minimum, and therefore particular shell constructions
679: can not be observed. The similarity function also points to the lack of regular
680: growth.
681:
682: \section{Acknowledgments}
683: We gratefully acknowledge {\it Fonds der Chemischen Industrie} for the very
684: generous support.
685: This work was supported by the SFB 277 of the University
686: of Saarland and by the German Research Council (DFG) through project
687: Sp439/14-1.
688: Finally, it is a pleasure to dedicate this work to Wolf Weyrich who, through
689: his work, demonstrated to one of the present authors (MS) the importance of
690: combining theory and experiment.
691:
692:
693: \begin{thebibliography}{9}
694:
695: \bibitem{pp04}
696: P. Pyykk\"o, Angew. Chem. Int. Ed. {\bf 43}, 4412 (2004).
697:
698: \bibitem{bf05}
699: F. Baletto and R. Ferrando, Rev. Mod. Phys. {\bf 77}, 371 (2005).
700:
701: \bibitem{rw96}
702: R. L. Whetten, J. T. Khoury, M. M. Alvarez, S. Murthy, I. Vezmar, Z. L. Wang,
703: P. W. Stephens, C. L. Cleveland, W. D. Luedtke, and U. Landman, Adv. Mater.
704: {\bf 8}, 428 (1996).
705:
706: \bibitem{as99}
707: A. Sanchez, S. Abbet, U. Heiz, W. -D. Schneider, H. H\"akkinen, R. N. Barnett,
708: and U. Landman, J. Phys. Chem. A {\bf 103}, 9573 (1999).
709:
710: \bibitem{hh03}
711: H. H\"akkinen, S. Abbet, A. Sanchez, U. Heiz, and U. Landman, Angew. Chem. Int.
712: Ed. {\bf 42}, 1297 (2003).
713:
714: \bibitem{lds03}
715: L. D. Socaciu, J. Hagen, T. M. Bernhardt, L. W\"oste, U. Heiz, H. H\"akkinen,
716: and U. Landman, J. Am. Chem. Soc. {\bf 125}, 10437 (2003).
717:
718: \bibitem{hb04}
719: U. Heiz and E. L. Bullock, J. Mater. Chem. {\bf 14}, 564 (2004).
720:
721:
722:
723: \bibitem{kjt92}
724: K. J. Taylor, C. L. Pettiette-Hall, O. Cheshnovsky, and
725: R. E. Smalley, J. Chem. Phys. {\bf 96}, 4 (1992).
726:
727: \bibitem{hh94}
728: H. Handschuh, G. Gantef\"or, P. S. Bechthold, and W. Eberhardt, J. Chem. Phys.
729: {\bf 100}, 7093 (1994).
730:
731: \bibitem{ap95}
732: A. Pinto, A. R. Pennisi, G. Faraci, G. D'Agostino, S. Mobilio, and F.
733: Boscherini, Phys. Rev. B {\bf 51}, 5315 (1995).
734:
735: \bibitem{mma97}
736: M. M. Alvarez, J. T. Khoury, T. G. Schaaff, M. Shafigullin, I. Vezmar, and R.
737: L. Whetten, Chem. Phys. Lett. {\bf 266}, 91 (1997).
738:
739: \bibitem{tgs97}
740: T. G. Schaaff, M. N. Schafigullin, J. T. Khoury, I. Vezmar, R. L. Whetten, W.
741: G. Cullen, P. N. First, C. Guti\'errez-Wing, J. Ascensio, and M. J.
742: Jose-Yacam\'an, J. Phys. Chem. B {\bf 101}, 7885 (1997).
743:
744: \bibitem{cc97}
745: C. L. Cleveland, U. Landman, M. N. Shafigullin, P. W. Stephens, and R. L.
746: Whetten, Z. Phys. D {\bf 40}, 503 (1997).
747:
748: \bibitem{kk98}
749: K. Koga, H. Takeo, T. Ikeda, and K. Ohshima, Phys. Rev. B {\bf 57}, 4053
750: (1998).
751:
752: \bibitem{bp98}
753: B. Palpant, B. Pr\'evel, J. Lerm\'e, E. Cottancin, M. Pellarin, M. Treilleux,
754: A. Perez, J. L. Vialle, and M. Broyer, Phys. Rev. B {\bf 57}, 1963 (1998).
755:
756: \bibitem{leh99}
757: L. E. Harrell, T. P. Bigioni, W. G. Cullen, R. L. Whetten, and P. N. First, J.
758: Vac. Sci. Technol. B {\bf 17}, 2411 (1999).
759:
760: \bibitem{vas00}
761: V. A. Spasov, Y. Shi, and K. M. Ervin, Chem. Phys. {\bf 262}, 75 (2000).
762:
763: \bibitem{mv01}
764: M. Vogel, K. Hansen, A. Herlert, and L. Schweikhard, Eur. Phys. J. D {\bf 16},
765: 73 (2001).
766:
767: \bibitem{nv01}
768: N. Vandamme, G. Verschoren, A. Depuydt, M. Cannaerts, W. Bouwen, P. Lievens, R.
769: E. Silverans, C. V. Haesendonck, Appl. Phys. A {\bf 72}, 177 (2001).
770:
771: \bibitem{hh01}
772: H. H\"ovel, Appl. Phys. A {\bf 72}, 295 (2001).
773:
774: \bibitem{vt03}
775: V. Torma, O. Vidoni, U. Simon, and G. Schmid, Eur. J. Inorg. Chem. 1121
776: (2003).
777:
778: \bibitem{rs04}
779: R. Salvati, A. Longo, G. Carotenuto, L. Nicolais, S. De Nicola, and G. P. Pepe,
780: Eur. Phys. J. B {\bf 41}, 43 (2004).
781:
782: \bibitem{gm04}
783: G. Mills, B. Wang, W. Ho, and H. Metiu, J. Chem. Phys. {\bf 120}, 7738 (2004).
784:
785:
786: \bibitem{bkm04}
787: B. K. Min, W. T. Wallace, A. K. Santra, and D. W. Goodman, J. Phys. Chem. B
788: {\bf 108}, 16339 (2004).
789:
790: \bibitem{hm04}
791: H. H\"akkinen, M. Moseler, O. Kostko, N. Morgner, M. A. Hoffmann, and B. v.
792: Issendorff, Phys. Rev. Lett. {\bf 93}, 093401 (2004).
793:
794: \bibitem{ll97}
795: L. Lamare and F. Michel-Calendini, Int. J. Quant. Chem. {\bf 61}, 635 (1997).
796:
797: \bibitem{odh97}
798: O. D. H\"aberlen, S. - C. Chung, M. Stener, and N. R\"osch, J. Chem. Phys. {\bf
799: 106}, 5189 (1997).
800:
801: \bibitem{gbp99}
802: G. Bravo-P\'erez, I. L. Garz\'on, and O. Novaro, J. Mol. Str. (Theochem) {\bf
803: 493}, 225 (1999).
804:
805: \bibitem{hh99}
806: H. H\"akkinen, R. N. Barnett, and U. Landman, Phys. Rev. Lett. {\bf 82}, 3264
807: (1999).
808:
809: \bibitem{ga00}
810: H. Gr\"onbeck and W. Andreoni, Chem. Phys. {\bf 262}, 1 (2000).
811:
812: \bibitem{hl00}
813: H. H\"akkinen and U. Landman, Phys. Rev. B {\bf 62}, 2287(R) (2000).
814:
815: \bibitem{jms00}
816: J. M. Soler, M. R. Beltr\'an, K. Michaelian, I. L. Garz\'on, P. Ordej\'on, D.
817: S\'anchez-Portal, and E. Artacho, Phys. Rev. B {\bf 61}, 5771 (2000).
818:
819: \bibitem{hh02}
820: H. H\"akkinen, M. Moseler, and U. Landman, Phys. Rev. Lett. {\bf 89}, 033401
821: (2002).
822:
823: \bibitem{jo02}
824: J. Olviedo and R. E. Palmer, J. Chem. Phys. {\bf 117}, 9548 (2002).
825:
826: \bibitem{jw02}
827: J. Wang, G. Wang, and J. Zhao, Phys. Rev. B {\bf 66}, 035418 (2002).
828:
829: \bibitem{jw03}
830: J. Wang, G. Wang, and J. Zhao, Chem. Phys. Lett. {\bf 380}, 716 (2003).
831:
832: \bibitem{zw03}
833: J. Zhao, J. Wang, and J. G. Hou, Phys. Rev. B {\bf 67}, 085404 (2003).
834:
835: \bibitem{by03}
836: B. Yoon, H. H\"akkinen, and U. Landman, J. Phys. Chem. A {\bf 107}, 4066
837: (2003).
838:
839: \bibitem{mj04}
840: M. P. Johansson, D. Sundholm, and J. Vaara, Angew. Chem. Int. Ed. {\bf 43},
841: 2678 (2004).
842:
843: \bibitem{sc04}
844: S. Chr\'{e}tien, M. S. Gordon, H. Metiu, J. Chem. Phys. {\bf 121}, 3756 (2004).
845:
846: \bibitem{rmo05}
847: R. M. Olson, S. Varganov, M. S. Gordon, H. Metiu, S. Chretien, P. Piecuch, K.
848: Kowalski, S. A. Kucharski, and M. Musia, J. Am. Chem. Soc. {\bf 127}, 1049
849: (2005).
850:
851: \bibitem{lx04}
852: L. Xiao and L. Wang, Chem. Phys. Lett. {\bf 392}, 452 (2004).
853:
854: \bibitem{avw05}
855: A. V. Walker, J. Chem. Phys. {\bf 122}, 094310 (2005).
856:
857: \bibitem{fr05}
858: F. Remacle and E. S. Kryachko, J. Chem. Phys. {\bf 122}, 044304
859: (2005).
860:
861:
862:
863:
864: \bibitem{hy03}
865: H. H\"akkinen, B. Yoon, U. Landman, X. Li, H.-J. Zhai, and L.-S.
866: Wang, J. Phys. Chem. A {\bf 107}, 6168 (2003).
867:
868: \bibitem{sg02}
869: S. Gilb, P. Weis, F. Furche, R. Ahlrichs, and M. M. Kappes, J.
870: Chem. Phys. {\bf 116}, 4094 (2002).
871:
872: \bibitem{ff02}
873: F. Furche, R. Ahlrichs, P. Weis, C. Jacob, S. Gilb, T. Bierweiler, and M. M.
874: Kappes, J. Chem. Phys. {\bf 117}, 6982 (2002).
875:
876: \bibitem{mn05}
877: M. Neumaier, F. Weigend, and O. Hampe, J. Chem. Phys. {\bf 122},
878: 104702 (2005).
879:
880:
881:
882:
883:
884: \bibitem{fe91}
885: F. Ercolessi, W. Andreoni, and E. Tosatti, Phys. Rev. Lett. {\bf 66}, 911
886: (1991).
887:
888: \bibitem{ha98}
889: H. Arslan and M. H. G\"uven, Springer-Verlag, ARI (1998) {\bf 51}, 145
890:
891: \bibitem{wdl98}
892: W. D. Luedtke and U. Landman, J. Phys. Chem. B {\bf 102}, 6566 (1998).
893:
894: \bibitem{clc98}
895: C. L. Cleveland, W. D. Luedtke, and U. Landman, Phys. Rev. Lett. {\bf 81}, 2036
896: (1998).
897:
898: \bibitem{clc99}
899: C. L. Cleveland, W. D. Luedtke, and U. Landman, Phys. Rev. B {\bf 60}, 5065
900: (1999).
901:
902: \bibitem{yc01}
903: Y. Chushak and L. S. Bartell, Eur. Phys. J. D {\bf 16}, 43 (2001).
904:
905: \bibitem{sp01}
906: S. M. Paik, S. M. Yoo, M. Namkung, and R. A. Wincheski, NASA/CR-2001-211014,
907: ICASE Report No. 2001-17
908:
909: \bibitem{scl02}
910: S. -C. Lee, K. -R. Lee, J. -G. Lee, and N. M. Hwang, Nanotech Vol. 2 (2002).
911:
912: \bibitem{fb02}
913: F. Baletto, R. Ferrando, A. Fortunelli, F. Montalenti, C. Mottet, J. Chem.
914: Phys. {\bf 116}, 3856 (2002).
915:
916: \bibitem{jr04}
917: J. Rogan, R. Ram\'irez, A. H. Romero, and M. Kiwi, Eur. Phys. J. D {\bf 28},
918: 219 (2004).
919:
920: \bibitem{ul04}
921: U. Landman and W. D. Luedtke, Faraday Discuss. {\bf 125}, 1 (2004).
922:
923: \bibitem{yw04}
924: Y. Wang, S. Teitel, and C. Dellago, J. Chem. Phys. {\bf 122}, 214722 (2005).
925:
926:
927:
928:
929: \bibitem{clc97}
930: C. L. Cleveland, U. Landman, T. G. Schaaff, M. N. Shafigullin, P. W. Stephens,
931: and R. L. Whetten, Phys. Rev. Lett. {\bf 79}, 1873 (1997).
932:
933: \bibitem{rb99}
934: R. N. Barnett, C. L. Cleveland, H. H\"akkinen, W. D. Luedtke, C. Yannouleas,
935: and U. Landman, Eur. Phys. J. D {\bf 9}, 95 (1999).
936:
937: \bibitem{dw98}
938: J. P. K. Doye and D. J. Wales, New J. Chem. {\bf 22}, 733 (1998).
939:
940: \bibitem{ntw00}
941: N. T. Wilson and R. L. Johnston, Eur. Phys. J. D {\bf 12}, 161 (2000).
942:
943: \bibitem{ntw02}
944: N. T. Wilson and R. L. Johnston, Phys. Chem. Chem. Phys. {\bf 4}, 4168 (2002).
945:
946: \bibitem{ilg96}
947: I. L. Garz\'on and A. Posada-Amarillas, Phys. Rev. B {\bf 54}, 11796 (1996).
948:
949: \bibitem{ilg98}
950: I. L. Garz\'on, K. Michaelian, M. R. Beltr\'an, A. Posada-Amarillas, P.
951: Ordej\'on, E. Artacho, D. S\'anchez-Portal, and J. M. Soler, Phys. Rev. Lett.
952: {\bf 81}, 1600 (1998).
953:
954: \bibitem{ilg99}
955: I. L. Garz\'on, K. Michaelian, M. R. Beltr\'an, A. Posada-Amarillas, P.
956: Ordej\'on, E. Artacho, D. S\'anchez-Portal, and J. M. Soler, Eur. Phys. J. D
957: {\bf 9}, 211 (1999).
958:
959: \bibitem{km99}
960: K. Michaelian, N. Rend\'on, and I. L. Garz\'on, Phys. Rev. B {\bf 60}, 2000
961: (1999).
962:
963: \bibitem{tl00}
964: T. X. Li, S. Y. Yin, Y. L. Ji, B. L. Wang, G. H. Wang, and J. J. Zhao, Phys.
965: Lett. A {\bf 267}, 403 (2000).
966:
967: \bibitem{sd02}
968: S. Darby, T. V. Mortimer-Jones, R. L. Johnston, and C. Roberts, J. Chem. Phys.
969: {\bf 116}, 1536 (2002).
970:
971:
972:
973:
974:
975:
976: \bibitem{cr93}
977: C. Rey, L. J. Gallego, J. Garc\'ia - Rodeja, J. A. Alonso, and M. P. I\~niguez,
978: Phys. Rev. B {\bf 48}, 8253 (1993).
979:
980: \bibitem{cr94}
981: J. Garc\'ia - Rodeja, C. Rey, L. J. Gallego, and J. A. Alonso, Phys. Rev. B
982: {\bf 49}, 8495 (1994).
983:
984: \bibitem{as04}
985: A. Sebetci and Z. B. G\"uven\c{c}, Modelling Simul. Mater. Sci. Eng. {\bf 13},
986: 683 (2005).
987:
988:
989:
990:
991:
992: \bibitem{da83}
993: M. S. Daw and M. I. Baskes, Phys. Rev. Lett. {\bf 50}, 1285 (1983).
994:
995:
996: \bibitem{da84}
997: M. S. Daw and M. I. Baskes, Phys. Rev. B {\bf 29}, 6443 (1984).
998:
999: \bibitem{fo86}
1000: S. M. Foiles, M. I. Baskes, and M. S. Daw, Phys. Rev. B {\bf 33}, 7983
1001: (1986).
1002:
1003: \bibitem{vo87}
1004: A. F. Voter and S. P. Chen, in {\it Characterization of Defects in Materials},
1005: edited by R. W. Siegal, J. R. Weertman, and R. Sinclair, MRS Symposia
1006: Proceedings No. 82 (Materials Research Society, Pittsburgh, 1987), p. 175.
1007:
1008: \bibitem{vo93}
1009: A. Voter, Los Alamos Unclassified Technical Report No LA-UR 93-3901 (1993).
1010:
1011: \bibitem{vo95}
1012: A. F. Voter, in {\it Intermetallic Compounds}, edited by J. H. Westbrook and
1013: R. L. Fleischer (John Wiley and Sons, Ltd, 1995), Vol. 1, p. 77.
1014:
1015: \bibitem{our}
1016: V. G. Grigoryan, D. Alamanova, and M. Springborg, Eur. Phys. J. D {\bf 34}, 187
1017: (2005).
1018:
1019: \bibitem{cole}
1020: D. Alamanova, Y. Dong, H. ur-Rehman, M. Springborg, and V. G. Grigoryan,
1021: Computing Letters (in press)
1022:
1023: \bibitem{numer92}
1024: W. H. Press, S. A. Teukolsky, W. T. Vetterling, B. P. Flannery, in
1025: {\it Numerical Recipes in FORTRAN: the Art of Scientific Computing},
1026: (Cambridge University Press, Cambridge, 1992), p. 387.
1027:
1028: \bibitem{cpl03}
1029: V. G. Grigoryan and M. Springborg, Chem. Phys. Lett.
1030: {\bf 375}, 219 (2003).
1031:
1032: \bibitem{prb04}
1033: V. G. Grigoryan and M. Springborg, Phys. Rev. B {\bf 70}, 205415 (2004).
1034:
1035: \end{thebibliography}
1036:
1037:
1038: \begin{table}
1039: \begin{center}
1040: \caption{Point groups of the optimized gold clusters. $N.k$ marks the energetically $k$-lowest isomer
1041: of the Au$_N$ cluster.}
1042: \begin{tabular}{cccc|cccc|cccc}
1043: \hline\hline
1044: N & N.1 & N.2 & N.3 & N & N.1 & N.2 & N.3 & N & N.1 & N.2 & N.3 \\
1045: \hline
1046: & & & &51 & C$_{\rm 1}$ & C$_{\rm 1}$ & C$_{\rm 1}$ & 101 & C$_{\rm 2}$ & D$_{\rm 5h}$ & C$_{\rm 1}$ \\
1047: 2 & D$_{\rm \infty h}$ & & &52 & D$_{\rm 5d}$ & C$_{\rm 2v}$ & C$_{\rm 2v}$ & 102 & C$_{\rm 1}$ & C$_{\rm 1}$ & C$_{\rm 1}$ \\
1048: 3 & D$_{\rm 3h}$ & & &53 & C$_{\rm 5v}$ & C$_{\rm 3v}$ & C$_{\rm 1}$ & 103 & C$_{\rm 2}$ & C$_{\rm 2}$ & C$_{\rm 1}$ \\
1049: 4 & T$_{\rm d}$ & & &54 & I$_{\rm h}$ & C$_{\rm 1}$ & C$_{\rm s}$ & 104 & C$_{\rm 1}$ & C$_{\rm 2}$ & C$_{\rm 1}$\\
1050: 5 & D$_{\rm 3h}$ & & &55 & C$_{\rm 3v}$ & C$_{\rm s}$ & C$_{\rm 1}$ & 105 & C$_{\rm 2}$ & C$_{\rm s}$ & C$_{\rm 2}$ \\
1051: 6 & O$_{\rm h}$ & C$_{\rm 2v}$ & &56 & C$_{\rm s}$ & C$_{\rm 2v}$ & C$_{\rm 2}$ & 106 & C$_{\rm 3v}$ & C$_{\rm 1}$ & C$_{\rm s}$\\
1052: 7 & D$_{\rm 5h}$ & C$_{\rm 3v}$ & C$_{\rm 2}$ &57 & C$_{\rm 1}$ & C$_{\rm 1}$ & C$_{\rm 1}$ & 107 & C$_{\rm s}$ & C$_{\rm 1}$ & C$_{\rm 1}$ \\
1053: 8 & D$_{\rm 2d}$ & C$_{\rm s}$ & D$_{\rm 3d}$ &58 & C$_{\rm 1}$ & C$_{\rm s}$ & C$_{\rm 1}$ & 108 & C$_{\rm s}$ & C$_{\rm s}$ & C$_{\rm 1}$ \\
1054: 9 & C$_{\rm 2v}$ & D$_{\rm 3h}$ & C$_{\rm s}$ &59 & C$_{\rm s}$ & C$_{\rm 1}$ & C$_{\rm 1}$ & 109 & C$_{\rm 3v}$ & C$_{\rm 1}$ & C$_{\rm 1}$\\
1055: 10 & C$_{\rm 3v}$ & D$_{\rm 4d}$ & D$_{\rm 3h}$ &60 & C$_{\rm s}$ & C$_{\rm s}$ & C$_{\rm 1}$ & 110 & C$_{\rm 2v}$ & C$_{\rm 1}$ & C$_{\rm 1}$ \\
1056: 11 & C$_{\rm 2v}$ & C$_{\rm 2}$ & C$_{\rm 2}$ &61 & C$_{\rm 3v}$ & C$_{\rm s}$ & C$_{\rm 2v}$ & 111 & C$_{\rm 1}$ & C$_{\rm 1}$ & C$_{\rm 1}$ \\
1057: 12 & C$_{\rm 5v}$ & C$_{\rm 2}$ & D$_{\rm 3h}$ &62 & C$_{\rm s}$ & C$_{\rm 1}$ & C$_{\rm 1}$ & 112 & C$_{\rm 2v}$ & C$_{\rm 1}$ & C$_{\rm 1}$ \\
1058: 13 & I$_{\rm h}$ & C$_{\rm 1}$ & C$_{\rm s}$ &63 & C$_{\rm 2v}$ & C$_{\rm s}$ & C$_{\rm s}$ & 113 & T & C$_{\rm 2}$ & D$_{\rm 2}$ \\
1059: 14 & C$_{\rm 3v}$ & C$_{\rm 2v}$ & C$_{\rm s}$ &64 & C$_{\rm 2v}$ & C$_{\rm 1}$ & C$_{\rm s}$ & 114 & C$_{\rm 3}$ & C$_{\rm 2v}$ & C$_{\rm 2}$ \\
1060: 15 & D$_{\rm 6d}$ & C$_{\rm 2v}$ & C$_{\rm s}$ &65 & C$_{\rm 1}$ & C$_{\rm 1}$ & C$_{\rm 1}$ & 115 & C$_{\rm s}$ & C$_{\rm 1}$ & C$_{\rm s}$ \\
1061: 16 & C$_{\rm 2v}$ & D$_{\rm 3h}$ & C$_{\rm 2v}$ &66 & C$_{\rm s}$ & C$_{\rm 2}$ & C$_{\rm 1}$ & 116 & C$_{\rm s}$ & O$_{\rm h}$ & C$_{\rm 3v}$ \\
1062: 17 & T$_{\rm d}$ & D$_{\rm 4d}$ & C$_{\rm s}$ &67 & C$_{\rm s}$ & C$_{\rm 1}$ & C$_{\rm 1}$ & 117 & C$_{\rm s}$ & C$_{\rm s}$ & C$_{\rm 1}$\\
1063: 18 & C$_{\rm 4v}$ & C$_{\rm 2v}$ & C$_{\rm s}$ &68 & C$_{\rm s}$ & C$_{\rm 1}$ & C$_{\rm 1}$ & 118 & C$_{\rm 2v}$ & C$_{\rm 1}$ & C$_{\rm 1}$\\
1064: 19 & D$_{\rm 5h}$ & D$_{\rm 4d}$ & C$_{\rm 2v}$ &69 & C$_{\rm 2}$ & C$_{\rm 1}$ & C$_{\rm 1}$ & 119 & C$_{\rm 1}$ & C$_{\rm 1}$ & C$_{\rm 1}$\\
1065: 20 & D$_{\rm 3d}$ & D$_{\rm 2}$ & D$_{\rm 2h}$ &70 & C$_{\rm 1}$ & C$_{\rm 1}$ & C$_{\rm 1}$ & 120 & C$_{\rm 1}$ & C$_{\rm 1}$ & C$_{\rm 1}$\\
1066: 21 & C$_{\rm s}$ & C$_{\rm 1}$ & C$_{\rm s}$ &71 & C$_{\rm 2v}$ & C$_{\rm 2v}$ & C$_{\rm 1}$ & 121 & C$_{\rm 1}$ & C$_{\rm 1}$ & C$_{\rm 1}$\\
1067: 22 & C$_{\rm 1}$ & C$_{\rm s}$ & D$_{\rm 6h}$ &72 & C$_{\rm s}$ & C$_{\rm s}$ & C$_{\rm s}$ & 122 & C$_{\rm 1}$ & C$_{\rm 1}$ & C$_{\rm 1}$\\
1068: 23 & C$_{\rm 2v}$ & C$_{\rm s}$ & C$_{\rm s}$ &73 & C$_{\rm 2v}$ & C$_{\rm 1}$ & C$_{\rm s}$ & 123 &C$_{\rm s}$ &C$_{\rm 1}$ & C$_{\rm 1}$\\
1069: 24 & C$_{\rm 2}$ & C$_{\rm s}$ & C$_{\rm 3v}$ &74 & C$_{\rm s}$ & C$_{\rm 5v}$ & C$_{\rm s}$ & 124 & C$_{\rm s}$ & C$_{\rm 1}$ & C$_{\rm s}$\\
1070: 25 & C$_{\rm 2}$ & C$_{\rm 1}$ & C$_{\rm 2v}$ &75 & D$_{\rm 5h}$ & C$_{\rm 2v}$ & C$_{\rm s}$ & 125 & C$_{\rm 1}$ & C$_{\rm 1}$ & C$_{\rm 1}$\\
1071: 26 & C$_{\rm 1}$ & C$_{\rm s}$ & C$_{\rm s}$ &76 & C$_{\rm s}$ & C$_{\rm s}$ & C$_{\rm s}$ & 126 & C$_{\rm s}$ & C$_{\rm 1}$ & C$_{\rm 1}$\\
1072: 27 & C$_{\rm s}$ & C$_{\rm 2}$ & C$_{\rm s}$ & 77 & C$_{\rm 2v}$ & C$_{\rm s}$ & C$_{\rm 2v}$&127 & C$_{\rm 1}$ & C$_{\rm 1}$ & C$_{\rm 1}$\\
1073: 28 & C$_{\rm s}$ & C$_{\rm 2}$ & C$_{\rm 2v}$ & 78 & C$_{\rm 2v}$ & C$_{\rm s}$ & C$_{\rm 1}$ &128 & C$_{\rm s}$ & C$_{\rm 1}$ & C$_{\rm s}$\\
1074: 29 & C$_{\rm 2}$ & C$_{\rm s}$ & C$_{\rm 2}$ & 79 & O$_{\rm h}$ & C$_{\rm s}$ & C$_{\rm 1}$ &129 & C$_{\rm 1}$ & C$_{\rm s}$ & C$_{\rm s}$\\
1075: 30 & C$_{\rm 3v}$ & C$_{\rm 1}$ & C$_{\rm 2}$ & 80 & C$_{\rm s}$ & O & C$_{\rm s}$ &130 & C$_{\rm s}$ & C$_{\rm s}$& C$_{\rm s}$\\
1076: 31 & C$_{\rm 3}$ & C$_{\rm 1}$ & C$_{\rm 3v}$ & 81 & C$_{\rm 2v}$ & C$_{\rm s}$ & C$_{\rm 1}$ &131 & C$_{\rm 2v}$ & C$_{\rm 1}$ & C$_{\rm 1}$\\
1077: 32 & D$_{\rm 2d}$ & C$_{\rm 3}$ & C$_{\rm 2v}$ & 82 & C$_{\rm s}$ & C$_{\rm 3}$ & C$_{\rm 2v}$ &132 & C$_{\rm s}$ & C$_{\rm 1}$ & C$_{\rm 1}$\\
1078: 33 & C$_{\rm 2}$ & C$_{\rm 1}$ & C$_{\rm 1}$ & 83 & C$_{\rm s}$ & C$_{\rm s}$ & C$_{\rm s}$ &133 & C$_{\rm 2v}$ & C$_{\rm 1}$ & C$_{\rm 1}$\\
1079: 34 & T$_{\rm d}$ & C$_{\rm 3}$ & C$_{\rm s}$ & 84 & C$_{\rm s}$ & C$_{\rm s}$ & C$_{\rm s}$ &134 & C$_{\rm s}$ & C$_{\rm s}$ & C$_{\rm s}$ \\
1080: 35 & C$_{\rm 2v}$ & D$_{\rm 3}$ & C$_{\rm 2v}$ & 85 & C$_{\rm s}$ & C$_{\rm s}$ & C$_{\rm s}$ &135 & C$_{\rm s}$ & C$_{\rm 1}$ & C$_{\rm 1}$\\
1081: 36 & C$_{\rm 2v}$ & C$_{\rm s}$ & D$_{\rm 2}$ & 86 & C$_{\rm s}$ & C$_{\rm s}$ & C$_{\rm s}$ &136 & C$_{\rm s}$ & C$_{\rm 1}$ & C$_{\rm 1}$\\
1082: 37 & C$_{\rm 2v}$ & C$_{\rm s}$ & C$_{\rm 2}$ & 87 & C$_{\rm 2}$ & C$_{\rm 1}$ & C$_{\rm 1}$ &137 & C$_{\rm 2v}$ & C$_{\rm 1}$ & C$_{\rm 1}$\\
1083: 38 & O$_{\rm h}$ & D$_{\rm 4h}$ & C$_{\rm s}$ & 88 & C$_{\rm 1}$ & C$_{\rm 1}$ & C$_{\rm 1}$ &138 & C$_{\rm s}$ & C$_{\rm 1}$ & C$_{\rm 1}$\\
1084: 39 & D$_{\rm 3}$ & C$_{\rm s}$ & C$_{\rm 4v}$ & 89 & C$_{\rm 2}$ & C$_{\rm s}$ & C$_{\rm s}$ &139 & C$_{\rm 2v}$ & C$_{\rm 1}$ & C$_{\rm 1}$\\
1085: 40 & D$_{\rm 3}$ & C$_{\rm 1}$ & C$_{\rm 2}$ & 90 & C$_{\rm 1}$ & C$_{\rm 1}$ & C$_{\rm 1}$ &140 & O$_{\rm h}$ & C$_{\rm s}$ & C$_{\rm s}$\\
1086: 41 & C$_{\rm 1}$ & C$_{\rm 1}$ & C$_{\rm 1}$ & 91 & C$_{\rm 1}$ & C$_{\rm 1}$ & C$_{\rm 1}$ &141 & C$_{\rm 2v}$ & C$_{\rm 2v}$ & C$_{\rm s}$\\
1087: 42 & D$_{\rm 4}$ & T$_{\rm d}$ & C$_{\rm 1}$ & 92 & C$_{\rm 1}$ & C$_{\rm 2}$ & C$_{\rm 1}$ &142 & C$_{\rm s}$ & C$_{\rm s}$ & C$_{\rm s}$ \\
1088: 43 & D$_{\rm 2}$ & C$_{\rm 1}$ & C$_{\rm 1}$ & 93 & C$_{\rm 1}$ & C$_{\rm 1}$ & C$_{\rm 1}$ &143 & C$_{\rm s}$ & C$_{\rm 2v}$ & C$_{\rm s}$\\
1089: 44 & C$_{\rm s}$ & C$_{\rm 2}$ & C$_{\rm 1}$ & 94 & C$_{\rm 1}$ & C$_{\rm 1}$ & C$_{\rm 1}$ &144 & C$_{\rm s}$ & C$_{\rm s}$ & C$_{\rm s}$\\
1090: 45 & C$_{\rm s}$ & C$_{\rm 1}$ & C$_{\rm 1}$ & 95 & C$_{\rm 1}$ & C$_{\rm 1}$ & C$_{\rm 1}$ &145 & C$_{\rm 1}$ & C$_{\rm s}$ & C$_{\rm 1}$\\
1091: 46 & C$_{\rm 3}$ & C$_{\rm s}$ & C$_{\rm 1}$ & 96 & C$_{\rm 1}$ & C$_{\rm 1}$ & C$_{\rm 1}$ &146 & D$_{\rm 5h}$ & C$_{\rm s}$ & C$_{\rm 1}$\\
1092: 47 & C$_{\rm 1}$ & C$_{\rm 2}$ & C$_{\rm 1}$ & 97 & C$_{\rm 2}$ & C$_{\rm 1}$ & C$_{\rm 1}$ &147 & C$_{\rm 1}$ & C$_{\rm 1}$ & C$_{\rm 1}$\\
1093: 48 & C$_{\rm 1}$ & D$_{\rm 2d}$ & C$_{\rm 1}$ & 98 & C$_{\rm 1}$ & C$_{\rm 1}$ & C$_{\rm 1}$ &148 & C$_{\rm 1}$ & C$_{\rm 1}$ & C$_{\rm 1}$\\
1094: 49 & C$_{\rm 1}$ & C$_{\rm 1}$ & C$_{\rm s}$ & 99 & C$_{\rm 2}$ & C$_{\rm 1}$ & C$_{\rm 1}$ &149 & C$_{\rm 1}$ & C$_{\rm 1}$ & C$_{\rm 1}$\\
1095: 50 & C$_{\rm 1}$ & C$_{\rm 1}$ & C$_{\rm 1}$ & 100 & C$_{\rm 1}$ & C$_{\rm 1}$ & C$_{\rm 1}$ &150 & C$_{\rm 1}$ & C$_{\rm 1}$ & C$_{\rm 1}$\\
1096:
1097: \hline\hline
1098: \end{tabular}
1099: \label{tab01}
1100: \end{center}
1101: \end{table}
1102:
1103:
1104:
1105: \unitlength1cm
1106: \begin{figure}
1107: \begin{picture}(14,08)
1108: \put(0,0){\psfig{file=Au_bind_sebetci.EPS,width=11cm}}
1109: \end{picture}
1110: \caption{On the left we show the binding energy per atom as a function of size for
1111: the energetically
1112: lowest isomers of $N$ up to 150 with the dashed line giving the {\it bulk} value; on
1113: the right is displayed the difference between the
1114: total energies of the lowest-lying isomers with up to 80 atoms obtained by
1115: Sebetci {\it et al.} and those found here, using different structure-optimization
1116: methods but the same embedded-atom approach.}
1117: \label{bind}
1118: \end{figure}
1119:
1120:
1121: \unitlength1cm
1122: \begin{figure}
1123: \begin{picture}(10,10)
1124: \put(0,0){\psfig{file=Au_stability.EPS,width=11cm}}
1125: \end{picture}
1126: \caption{The stability function as a function of cluster size.}
1127: \label{stability}
1128: \end{figure}
1129:
1130:
1131:
1132: \unitlength1cm
1133: \begin{figure}
1134: \begin{picture}(10,10)
1135: \put(0,0){\psfig{file=diff21.EPS,width=11cm}}
1136: \end{picture}
1137: \caption{The total-energy difference between the two energetically lowest
1138: neighbouring isomers as a function of cluster size.}
1139: \label{diff}
1140: \end{figure}
1141:
1142:
1143: \unitlength1cm
1144: \begin{figure}
1145: \begin{picture}(10,12)
1146: \put(0,0){\psfig{file=I.ps,width=10cm}}
1147: \end{picture}
1148: \caption{Different properties related to the eigenvalues $I_{\alpha\alpha}$ of the
1149: matrix with the moments of inertia. In the upper panel
1150: we show the average value together with points indicating whether clusters with
1151: overall spherical shape (lowest set of rows), overall cigar shape (middle set of rows),
1152: or overall lens shape (upper set of
1153: rows) are found for a certain size. Moreover, in each set of rows, the lowest
1154: row corresponds
1155: to the energetically lowest isomer, the second one to the energetically
1156: second-lowest
1157: isomer, etc. In the lower panel we show the maximum difference of the
1158: eigenvalues for the three different isomers.}
1159: \label{shape}
1160: \end{figure}
1161:
1162:
1163: \unitlength1cm
1164: \begin{figure}
1165: \begin{picture}(10,8)
1166: \put(0,0){\psfig{file=radial_updated.ps,width=10cm}}
1167: \end{picture}
1168: \caption{The distribution of radial distances (in \AA) for the lowest-lying
1169: isomer as a function of cluster size. Each
1170: small line represents (at least) one atom with that radial distance.}
1171: \label{radial}
1172: \end{figure}
1173:
1174: \unitlength1cm
1175: \begin{figure}
1176: \begin{picture}(10,15)
1177: \put(0,0){\psfig{file=Aukoord.EPS,width=10cm}}
1178: \end{picture}
1179: \caption{(a) the average coordination number, (b)
1180: the minimum coordination number, and (c) the average bond length as
1181: functions of cluster size. The dashed lines in (a) and (c) show the
1182: corresponding bulk values for gold.}
1183: \label{koord}
1184: \end{figure}
1185:
1186:
1187: \unitlength1cm
1188: \begin{figure}
1189: \begin{picture}(10,10)
1190: \put(0,0){\psfig{file=similarity.EPS,width=10cm}}
1191: \end{picture}
1192: \caption{The similarity function as function of cluster size. It describes
1193: whether the cluster with $N$ atoms is similar to that of $N-1$ atoms plus an
1194: extra atom.}
1195: \label{similar}
1196: \end{figure}
1197:
1198: \unitlength1cm
1199: \begin{figure}
1200: \begin{picture}(12,10)
1201: \put(0,0){\psfig{file=pict.ps,width=12cm}}
1202: \end{picture}
1203: \caption{Some Au$_N$ clusters with high or peculiar symmetry.}
1204: \label{pict}
1205: \end{figure}
1206:
1207: \end{document}
1208:
1209: