The nature of the silicon oxide transition region in the vicinity of the Si/SiO2 interface is probed by infrared and x-ray photoelectron spectroscopies. The layer-by-layer composition of the interface is evaluated by uniformly thinning thermal oxide films from 31 Å down to 6 Å. We find that the thickness dependence of the frequencies of the transverse optical and longitudinal optical phonons of the oxide film cannot be reconciled by consideration of simple homogeneous processes such as image charge effects or stress near the interface. Rather, by applying the Bruggeman effective medium approximation, we show that film inhomogeneity in the form of substoichiometric silicon oxide species accounts for the observed spectral changes as the interface is approached. The presence of such substoichiometric oxide species is supported by the thickness dependence of the integrated Si suboxide signal in companion x-ray photoelectron spectra.

We have investigated the fundamental mechanism underlying the hydrogen-induced exfoliation of silicon, using a combination of spectroscopic and microscopic techniques. We have studied the evolution of the internal defect structure as a function of implanted hydrogen concentration and annealing temperature and found that the mechanism consists of a number of essential components in which hydrogen plays a key role. Specifically, we show that the chemical action of hydrogen leads to the formation of (100) and (111) internal surfaces above 400 °C via agglomeration of the initial defect structure. In addition, molecular hydrogen is evolved between 200 and 400 °C and subsequently traps in the microvoids bounded by the internal surfaces, resulting in the build-up of internal pressure. This, in turn, leads to the observed “blistering” of unconstrained silicon samples, or complete layer transfer for silicon wafers joined to a supporting (handle) wafer which acts as a mechanical “stiffener.”

Surface infrared absorption spectroscopy and density functional cluster calculations are used to definitively demonstrate that the Si-Si dimer bond is the target for the initial insertion of oxygen into the Si(100) -͑231͒ surface, following H 2 O exposure and annealing. This reaction, in turn, facilitates the subsequent incorporation of O into the Si backbonds, thereby promoting (local) oxidation.[S0031-9007(97)04218-X]

No abstract

Infrared absorption spectroscopy and density functional cluster calculations are used to identify the intermediate oxide structures formed by high temperature annealing of the water-exposed Si͑100͒-͑2 3 1͒ surface. We find that initially there is a strong tendency for oxygen to agglomerate on single dimer units at T ϳ 800 K. Upon dehydrogenation, a remarkable structural transition is observed, wherein the dangling bonds recombine to form silicon epoxides (three-membered Si-O-Si rings). We demonstrate that these epoxides are the thermodynamically favored product in such constrained systems and, consequently, should be preferentially formed at silica interfaces. [S0031-9007(98)07445-6] PACS numbers: 81.65. Mq, 31.15.Ar, 68.35.Ja, 78.55.Ap Understanding the formation and evolution of SiO (defect) structures at silica surfaces and Si͞SiO 2 interfaces is of prime importance due both to the utilization of ultrathin oxide ͑ϳ10 Å͒ films as the gate dielectric layer in state-of-the-art semiconductor devices [1] and to the production of ultralow loss silica-based fiber for long distance optical communications. The detailed structural understanding of such interfaces poses a formidable scientific challenge due to the lack of long-range order and critical dependence on a wide array of production parameters. Therefore, a comprehensive knowledge of the fundamental physical and chemical properties of such systems must commence with the characterization of related systems that are more reproducible and amenable to control. In accordance with this prescribed approach, a wide variety of experimental [2] and theoretical [3] studies of such model systems have been undertaken over the past decade, but have singularly failed to provide a definitive mechanistic picture. For example, previous theoretical studies have focused primarily on a few simple Si-O structures that may result from the insertion of a single oxygen into the Si͑100͒-͑2 3 1͒ surface, with little exploration of the wider group of multiply oxidized structures that are subsequently formed during oxidation.Similarly, experimental studies have typically employed techniques [e.g., ultraviolet photoemission spectroscopy (UPS), x-ray photoemission spectroscopy (XPS)] and conditions that do not permit identification of the relevant discrete (sub)oxide structures formed. Recently, a combined infrared and theoretical study of the initial water-induced oxidation of Si͑100͒-͑2 3 1͒ [4] has shown that oxygen is first inserted into the dimer bonds and then into the Si backbonds, although the formation of the "real" multiply oxidized Si structures (that are the constituent subunits of extended layer growth) could not be delineated, as the spectral signatures of these oxide species were below the accessible range ͑,1000 cm 21 ͒.In this work, we have implemented a novel spectroscopic configuration that allows access to a broad spectral range ͑550 4000 cm 21 ͒ with submonolayer sen-sitivity to all vibrational components. We have primarily focused on the H 2 O:Si͑100͒-͑2 3 1͒ system a...