Cagelike Si 12 clusters with endohedral Cu, Mo, and W metal atom impurities F. Hagelberg, 1 C. Xiao, 1 and William A. Lester, Jr. 2 1 Computational Center for Molecular Structure and Interactions, Department of Physics, ATM Sciences and General Science, Jackson State University, Jackson, Mississippi 39217 2 Kenneth S. Pitzer Center for Theoretical Chemistry, Department of Chemistry, University of California at Berkeley, Berkeley, California 94720-1460 Received 12 March 2002; revised manuscript received 9 September 2002; published 31 January 2003 In a recent series of mass-spectrometric ion trap measurements H. Hiura et al., Phys. Rev. Lett. 86, 1733 2001, the formation of silicon clusters with endohedral transition-metal impurities was observed. Particular stability was assigned to the experimentally detected species WSi 12 + , which has been shown by ab initio geometry optimization to adopt the shape of a regular hexagonal Si 12 prism with the W atom in the center. A similar geometry—namely, a Si 12 double-chair structure surrounding the metal atom impurity—has emerged from our extensive investigations of silicon clusters in combination with a Cu atom (CuSi N ) as the likely ground-state structure of CuSi 12 . These results suggest the systematic importance of Si 12 cages derived from regular structures with D 6h geometry for the architecture of silicon clusters containing metal atom impurities. In the present comparative study, we discuss the salient features of endohedral M Si 12 clusters with M =Cu, Mo, W, as well as several cationic and anionic species of these systems, with regard to their geometric and electronic structure. The interaction between the Si 12 cage and the enclosed metal impurity is characterized as strongly delocalized bonding for M =Mo, W, while Cu tends to form directed bonds with selected atoms of the cage. Linear extension of the M Si 12 (Me=Mo,W) cells along their principal axes leads to units of the form M 2 Si 18 . DOI: 10.1103/PhysRevB.67.035426 PACS numbers: 61.46.+w, 31.15.Dv, 36.40.Cg, 73.22.-f I. INTRODUCTION Numerous research efforts, both experimental and compu- tational, have been devoted to the understanding of silicon clusters (Si N ). 1,2 The interest in these species is motivated partly by the desire to gain fundamental insight into the mechanisms that govern the evolution of Si systems from the molecular to the macroscopic scale. In addition, there is the prospect of technological innovation through the fabrication of novel materials with Si N clusters as building blocks. In both respects, the investigations of Si N have been guided by dramatic developments in the field of carbon clusters C N during the last two decades. However, no fullerene like ar- chitectures have been identified noncontroversially for Si N units to this date. This disparity between Si N and C N is at- tributed to the finding that the bonding in fullerenes is char- acterized by sp 2 hybridization, which is more favorable for C N than for Si N units. 3 However, from the extensive knowledge accumulated on a wide variety of metal-doped fullerene species, such as La@C N ( N =60,74,82), 4 it has been suggested that implan- tation of a metal impurity into a Si N unit could lead to the formation of a cagelike Si N structure of extraordinary stability. 5 This consideration provides strong motivation for the study of mixed metal-Si clusters. Several recent experimental projects have dealt with these systems. In a pioneering mass spectrometric investigation us- ing a laser vaporization technique, 6 Beck demonstrated the existence of small mixed transition-metal TM-Si clusters, observing various species of the form TMSi N with TM =Cr,Mo,W; N =16,17,18. Additionally, he reported the ob- servation of CuSi N clusters with a pronounced abundance maximum at N =10. This latter study was complemented more recently by experiments of Scherer et al. 7 who identi- fied several series of smaller Cu M Si N clusters. Stimulated by Beck’s experimental work, some computational investiga- tions have been performed on selected TMSi N species with N =15,17, for which the systems were subject to hypotheti- cal symmetry constraints. 8 Moreover, a comprehensive study of the geometric and electronic features of CuSi N has been carried out, 9–12 and the systematics of M Si N ( M =Cr,Mo,W; N 7) clusters 13–15 has been explored. A recent computational study 16 identified numerous clus- ter species of the form M @Si N ( M =Fe,Ru,Os; N =14 and M =Ti,Zr,Hf; N =16). It was shown that some of these clus- ters display highest occupied molecular orbital and lowest unoccupied molecular orbital HOMO-LUMO gaps whose magnitudes are indicative of extraordinary stability. The latest experimental achievement related to metal-Si N clusters was reported by Hiura et al. 17 The authors used an ion trap procedure to detect a cluster series of the form M Si N + ( M =Hf,Ta,W,Re,Ir,...; N =9,11,12,13,14). The method makes possible the recording of time-resolved mass spectra and the observation of the growth of a M Si N + cluster species in considerable detail. The products WSi N H x of the reaction of W + ions with SiH 4 gas molecules were moni- tored in a temporal sequence of spectra. The growth process of the species slows down for N 8 and terminates at N =12. At this number of constituents, a saturation point ap- pears to be reached; the capacity of the metal ion to bind additional Si atoms seems to be exhausted. This tendency in conjunction with the observation that the resulting WSi 12 + cation is dehydrogenated suggests that a highly compact cluster is formed in which the W atom occupies an endohe- dral site. This result can be counted as the first experimental evidence for the existence of a cagelike Si N frame encapsu- PHYSICAL REVIEW B 67, 035426 2003 0163-1829/2003/673/0354269/$20.00 ©2003 The American Physical Society 67 035426-1