Page 266 - Academic Press Encyclopedia of Physical Science and Technology 3rd InOrganic Chemistry
P. 266

P1: GNH Final Pages
 Encyclopedia of Physical Science and Technology  EN009M-428  July 18, 2001  1:6







              Metal Particles and Cluster Compounds                                                       523

              infrared spectroscopy. The CO stretching frequencies on  Clusters on surfaces, or supported clusters, lend them-
              clusters of one to four nickel (or copper) atoms rapidly  selves to a variety of X-ray and electron spectro-
              approach the values found for CO chemisorbed on the  scopies, such as extended X-ray absorption fine structure
              polycrystalline bulk metal. Thus, the CO to metal bond-  (EXAFS), X-ray and ultraviolet photoelectron spec-
              ing appears to be local as expected.              troscopy (XPS and UPS), and transmission electron mi-
                In mixed deposits of metal, methane, and argon only  croscopy (TEM) to name a few. Most studies of model
              A1 atoms of some 18 metals tried appears to react at 10 K  systems have used sodium chloride or carbon substrates
              without photolysis. The reactivity of A1 appears to be  and ultrahigh-vacuum techniques.
                                                2
              unique and probably due to its radical like P state. The-  Several authors have modeled the kinetics of cluster
              oretical studies lend support to the idea that atoms with  growth. Nucleation occurs at specific sites, often associ-
              partially filled p orbitals would be most reactive in C H  ated with lattice defects on the substrate. Cluster growth
              insertion processes.                              commences as atoms impinge near each other; in a se-
                The question as to whether clusters or atoms are more  ries of experiments it was found that gold atoms will be
              reactive in carbon-halogen bond breaking processes was  captured if they land within 6.5 A of a growing cluster.
              recently considered. It was found that in an argon matrix  As cluster growth continues the energetics regarding
              at 10 K Mg 2 and Mg 3 reacted with CH 3 Br (presumably to  two- or three-dimensional growth must be considered.
              form CH 3 Mg 2 Br and CH 3 Mg 3 Br), but Mg atoms did not.  Calculations indicate that if the heat of vaporization of
              This higher reactivity was attributed to the lower ioniza-  the metal 	 0 is greater than three times the heat of des-
              tion potential of these small clusters and to the fact that  orption of the metal atom from the surface, E des , three-
              in the free state Mg 2 and Mg 3 are very weakly bound, but  dimensional clusters should form. However, if the heat
              in the product should be strongly bound. Similar results  of metal vaporization is less than three times the energy
              have been found for Ca, Ca 2 , and Ca 3 . Ab initio theoret-  of desorption minus the energy of diffusion E diff ,two-
              ical studies convincingly support these experimental re-  dimensional cluster growth is favored. Depending on the
              sults. However, even after numerous attempts, these clus-  surface, E des and E diff will vary:
              ter compounds have eluded isolation, and are apparently
              quite unstable.                                                      	 0   3E des
                Kinetic studies of metal atom aggregation in cold matri-  (three-dimensional growth favored)
              ces has received some attention. A statistical frozen matrix
              approach (calculating probability that M and M are neigh-          	 0   3E des − E diff
              bors and react to give M 2 ), and a highly mobile metal    (two-dimensional growth favored)
              atom approach (diffusion being rapid in quasiliquid layer)
              have been used. It was found that the diffusion mecha-  These theoretical considerations predict that metal clus-
              nism was supported best by the experimental results. The  ters growing on most clean metal surfaces and semicon-
              eventual M 2 concentration was found to be proportional  ductors would grow initially in two dimensions, and this
              to the square of the M/substrate ratio. Concentrations of  is found experimentally.
              higher metal aggregates vary as some higher power of the  Small clusters on surfaces have some unusual proper-
              M/substrate ratio.                                ties. Their geometrical shapes usually do not resemble the
                Since the statistical approach underestimates the for-  bulk element. For example, using Moire interference pat-
              mation of clusters in the matrix, further analysis has been  terns and TEM the smallest colloids of gold are shown
              done. Both dilute and high-concentration matrices have  to have pentagonal symmetry. The presence of multiple
              been dealt with experimentally and mathematically. The  twinned tetrahedra such as icosahedra, again suggests that
              best model was found to be one which simulates freeze-  icosahedral structures are preferred over the bulk structure
              out by assuming the reactions to stop abruptly after a cer-  during the initial growth phase.
              tain time. It should also be pointed out that the trend in  If perfect crystallites are not formed in the initial stages
              product distribution as a function of metal concentration  of growth, there will be some strain energy in the cluster.
              is adequately described by this model and greatly aids  As metal thickness increases, the desorption energy of a
              spectroscopic assignment of bands to metal atoms and  metal atom on the surface changes. For Na on a tungsten
              clusters.                                         surface, E des is initially 2.5 ± 0.2 eV, but after four layers
                A film produced by deposition of atoms or particles  are deposited becomes 1.06 eV, which corresponds closely
              on a surface forms in several stages: (1) nucleation, (2)  to the bulk heat of vaporization of Na. Since the Na Na
              cluster growth, (3) coalescence, (4) further thickening, and  bonding energy is so high initially, lattice relaxation via
              (5) recrystallization (perhaps). We will be concerned with  two-dimensional dislocations is difficult so more strain
              steps (1)–(3).                                    energy is created in the growing Na cluster.
   261   262   263   264   265   266   267   268   269   270   271