Journal of Vacuum Science and Technology A: Vacuum, Surfaces and Films

Atomic-scale control and manipulation of the microstructure of polycrystalline thin films during kinetically limited low-temperature deposition, crucial for a broad range of industrial applications, has been a leading goal of materials science during the past decades. Here, we review the present understanding of film growth processes—nucleation, coalescence, competitive grain growth, and recrystallization—and their role in microstructural evolution as a function of deposition variables including temperature, the presence of reactive species, and the use of low-energy ion irradiation during growth.

Structural analysis of the surface reconstructions investigated by ultrahigh vacuum (UHV) transmission electron microscopy (TEM) and diffraction (TED) is shown. By TED intensity analysis a new structural model of Si(111)-7×7 is derived. The model basically consists of 12 adatoms arranged locally in the 2×2 structure, nine dimers on the sides of the triangular subunits of the 7×7 unit cell and a stacking fault layer. UHV–HREM of Si (111)-7×7 surface is commented.

The recent development in the field of superhard materials with Vickers hardness of ⩾40 GPa is reviewed. Two basic approaches are outlined including the intrinsic superhard materials, such as diamond, cubic boron nitride, C3N4, carbonitrides, etc. and extrinsic, nanostructured materials for which superhardness is achieved by an appropriate design of their microstructure. The theoretically predicted high hardness of C3N4 has not been experimentally documented so far. Ceramics made of cubic boron nitride prepared at high pressure and temperature find many applications whereas thin films prepared by activated deposition from the gas phase are still in the stage of fundamental development. The greatest progress has been achieved in the field of nanostructured materials including superlattices and nanocomposites where superhardness of ⩾50 GPa was reported for several systems. More recently, nc-TiN/SiNx nanocomposites with hardness of 105 GPa were prepared, reaching the hardness of diamond. The principles of design for these materials are summarized and some unresolved questions outlined.

Sculptured thin films with three dimensional microstructure controlled on the 10 nm scale were fabricated with an evaporation technique. Glancing angle deposition (GLAD) and substrate motion were employed to “sculpt” columnar thin film microstructure into desired forms ranging from zigzag shaped to helical to four-sided “square” helical. Computer control of substrate motion was used to accurately position the substrate and to achieve the desired film structures. The growth mechanics of this novel thin film deposition technique are investigated with density measurements, scanning electron microscopy analysis, and measurements of effective refractive index. Adatom diffusion and atomic shadowing are the dominant growth mechanisms with glancing angle deposition conditions creating extreme shadowing. With controlled rotation of the substrate about two axes during deposition, a dense capping layer can be produced on top of the porous sculptured films. The success of the capping process was found to be strongly dependent on the technique used, with an exponential decrease (θ∝[1−A⋅eB⋅t]) with time of incident flux angle found to be the best to reduce filling of the porous film and fracturing of the capping film. The GLAD technique was found to have potentially promising application in optical, biological, and chemical devices and materials.

Thin films prepared under conditions of low adatom mobility are characterized by a highly anisotropic physical structure with a wide range of systematically varying column and void sizes. The structure zone models, previously developed to classify the larger sized physical structures, are revised to account for the evolutionary growth stages of structure development as well as the separate effects of thermal- and bombardment-induced mobility. The zone T introduced by Thornton is shown to be a subzone within zone 1.

We have studied the local atomic structure of silicon suboxide (SiOx, x<2) thin films using infrared (IR) spectroscopy. The films were prepared by plasma enhanced chemical vapor deposition (PECVD) of silane (SiH4) and nitrous oxide (N2O) mixtures, which were then diluted with He. The IR spectra were found to vary significantly with the degree of He dilution. Films grown with no He showed SiN, NH, and SiH bonding groups in addition to the three characteristic vibrations of the Si–O–Si linkage. The addition of He reduced the strength of the SiN, NH, and SiH absorption bands, and resulted in systematic increases in the frequency of the Si–O–Si asymmetric stretching vibration. The frequency of this Si–O–Si stretching vibration scales linearly with the oxygen concentration from approximately 940 cm−1 in oxygen doped amorphous silicon to 1075 cm−1 in stoichiometric noncrystalline SiO2. A deposition temperature of 350 °C and a He dilution of 50% gave a film composition close to SiO1.9. We propose a model for the deposition process that emphasizes the role of the He dilution.

It has become common practice to employ, as a binding energy reference for x-ray photoelectron spectroscopy studies on nonconductive materials, the C(1s) spectra of the ubiquitous (adventitious) carbon that seems to exhibit an instantaneous presence on all air exposed materials. Despite this commonality, surface scientists, including many practitioners, have expressed substantial concerns about the validity of this approach. A detailed discussion of the method is presented including consideration of the types of materials and the electronic energy states involved, e.g., Fermi edges, vacuum levels, etc., and the couplings that must exist for the referencing method to be correctly applied. A number of other surface environments for which the carbon referencing method may be fallacious are also presented. This leads to a consideration of the electron spectroscopy for chemical analysis results for different types of adventitious species and how the presence of some of these may confuse the use of the method. In this regard, we will also discuss the use of other methods to establish binding energy scales, such as Fermi edge coupling and select doping (e.g., the Au dot approach).