The results of first principles calculations on H-silsesquioxanes (i.e., (HSiO 3/2 ) n with n ) 4, 6, 8, 10, 12, 14, and 16) are reported here. Double numeric basis sets and local and nonlocal density approximations to density functional theory are employed for calculations. It is shown that use of the nonlocal density approximation is required for the reliable prediction of the most stable isomer for silsesquioxanes. Furthermore, a progression of the preferred building unit with the increase in size of the T cage is revealed. The smaller T cages prefer four-and five-member rings while the larger cages are found to prefer four-and six-member rings. Analysis of the energy of the hydrolysis reaction, binding energy, and fragmentation paths finds the relative stability of the silsesquioxane cages containing four-, five-, and six-member rings in agreement with experimental observations. For the (HSiO 3/2 ) 16 cage, the calculated results predict the stability of the D 2d -6 4 5 0 4 6 configuration over the D 4d -6 0 5 8 4 2 configuration in contradiction to suggestions based on 29 Si NMR measurements. We find a consistent picture for the highest occupied molecular orbitals (HOMOs) of all silsesquioxanes considered showing them to be composed of (lone-pair) oxygen p-type atomic orbitals. On the other hand, the lowest unoccupied molecular orbitals (LUMOs) show size dependence in their composition which appears to cause the presence of a state in the HOMO-LUMO gap for higher silsesquioxane cages. Density of states plots and analysis of molecular orbitals reveal this state to be due to the terminal hydrogens bonded to silicon atoms.

Calculations based on density functional theory (DFT) were performed on various structural isomers of methyl silsesquioxanes, [MeSiO 3/2 ] n where n ) 4, 6, 8, 10, 12, 14, and 16, to study their structural and electronic properties. The calculated results find the stability of methyl silsesquioxanes, except [MeSiO 3/2 ] 4 , against fragmentation and hydrolysis, and of one isomer of [MeSiO 3/2 ] 14 against hydrolysis. The deformation density plots show that chemical bonding in methyl silsesquioxanes is mainly determined by the building block unit, (MeSiO 3/2 ) as also seen in hydridosilsesquioxanes (HSQ). However, unlike HSQ, the large cages of methyl silsesquioxanes do not develop a localized electronic state in the HOMO-LUMO gap.

No abstract

Single and multicomponent mixed layers of silsesquioxane clusters on freshly evaporated gold surfaces have been investigated by X-ray photoelectron spectroscopy and reflection-absorption infrared spectroscopy. Approximately 5-10% of the cluster layers (e.g., H8Si8O12 and H10Si10O15) on gold desorb upon evacuation of the adsorbate from the reaction chamber. These open adsorption sites are an avenue for cluster displacement reactions that yield mixed monolayers (e.g., H8Si8O12/D8Si8O12 and H8Si8O12/C6H13-H7Si8O12) of several compositions on gold. This dynamic behavior is not observed for the C6H13-H7Si8O12 cluster layer on gold. Rather, this molecule acts as a poison to these reported displacement processes at the gold surface.

The chemisorption of the hydridosilsesquioxane clusters H8Si8O12 and H10Si10O15 onto a freshly evaporated gold surface has been observed by X-ray photoemission and reflection-absorption infrared spectroscopies. On the basis of these analytical techniques and supporting nonlocal density functional calculations, a single Si-Au bond at a cluster vertex is created through a surprising gold-mediated Si-H bond activation. The chemisorbed clusters are stable in a vacuum and when exposed to oxygen or water. Coverage-dependent peak frequency and infrared intensity shifts, characteristic of a chemically dynamic interface, are detected as a function of cluster overpressure. An equilibrium between the clusters and the gold surface is proposed, evidenced by a fairly constant rate of adsorption to gold for several coverage regimes. IntroductionProgress in the areas of microelectronic device packing density, speed, and functionality generally requires smaller device dimensions. 1 The ensuing problems associated with smaller line dimensions such as propagation delay, crosstalk noise, and power dissipation must be addressed. As a solution to these problems, new, low dielectric constant (k < 3) materials as well as alternative architectures have been proposed to replace the current Al(Cu) and SiO 2 interconnect technology. However, the replacement of SiO 2 as the insulator of choice is frequently complicated by the occurrence of undesired chemical reactions at the interface between the new low-k material and the metal lines. As dimensions approach the atomic scale, characteristics of microelectronic devices can become dominated by the chemical structure and dynamics of the interfaces.Hydridosilsesquioxane (HSQ) resin (k ) 2.9) is a siloxane-based polymer consisting primarily of O 3 SiH entities with an overall stoichiometry of (HSiO 1.5 ) n . High thermal stability, excellent gap fill capability, low electrical leakage, and relatively low outgassing properties are some of the desirable attributes of this polymer. 2 However, device failure results when the polymer is used in direct contact with metal lines. To avoid this problem, the first step used during device fabrication involves the formation of a very thin layer of SiO 2 directly on the metal lines. The HSQ resin is then spun cast onto the SiO 2 followed by curing and testing. 2 This procedure is problematic because it adds additional steps to the process and adds nonnegligible amounts of SiO 2 (k ∼ 4) to the dielectric layer. Ideally, a low-k dielectric could be applied directly without the formation of the SiO 2 barrier layer. In this paper, the reactions that HSQ resin may undergo when in direct contact with metal surfaces are explored.The interface of metal-insulator metal (MIM) devices is also of substantial interest. MIM devices show interest-

We report the frequency dependence of linear and nonlinear optical susceptibilities of H-silsesquioxanes of various cage sizes and conformations using the INDO/CI method coupled with the SOS method. The average dynamic refractive index of silsesquioxanes is found to decrease as the cage size increases, and its variation is very small for the different conformations of the same cage size at the same incident wavelength. The calculated second-order susceptibilities show that small cages have a larger magnitude which is comparable to that of crystalline R-quartz. It is suggested here that H-silsesquioxanes of a smaller cage size can be a good candidate materials for nonlinear optical applications having low absorption, wide transparency and adequate susceptibility in ultraviolet or vacuum ultraviolet region of the spectrum.

No abstract

Scanning tunneling microscopy is used to determine the bonding geometry of the spherosiloxane cluster, H(8)Si(8)O(12) , on Si(100)-2 x 1. The images obtained are consistent with monovertex bonding to the Si(100)-2 x 1 surface via activation of a single Si-H bond. Filled and empty state images show good agreement with calculations of the electron density distribution of the cluster as well as the Psi(2) highest occupied molecular orbital and lowest unoccupied molecular orbital surface plots of the cluster.