ICMCTF2014 Session C2: Novel Aspects in Thin Film Characterization and Data Modeling

Information on material characteristics such as chemical composition, order (crystal phase, grain size, or disorder), and sample structure (thin film thicknesses, interface formation, spatial non-uniformity, general morphology) are extracted from spectroscopic ellipsometry measurements of thin films in photovoltaic device configurations. Optical properties in the form of the complex dielectric function (ε = ε₁ + iε₂) are extracted for semiconductor layers, transparent conducting oxides (TCOs), and metal contacts. Variations in e are compared with structure for amorphous and nanocrystalline hydrogenated silicon (Si:H), grain size and strain for cadmium telluride (CdTe), composition of copper indium gallium diselenide (CIGS), and electronic transport properties as reflected using Drude oscillator models for tin oxide and zinc oxide TCOs and molybdenum and silver metal contacts. All of these properties influence the behavior of the thin film in the device as well as the device’s ultimate performance. In situ real time spectroscopic ellipsometry (RTSE) applied during thin film growth has been to track structural evolution in Si:H, compositional changes including phase segregation of copper selenide and CIGS, and interface filling on textured substrates used for amorphous and nanocrystalline Si:H and CdTe solar cells. In particular, growth evolution of complete Si:H devices (including the doped and undoped semiconductor layers, transparent conducting oxides, and metal back contacts) will be compared on specular and textured substrates. Ex situ mapping spectroscopic ellipsometry measurements are conducted for the same samples studied by RTSE to demonstrate that structural models developed by RTSE can be applied to assess spatial non-uniformity over large areas—in essence connecting fundamental material studies to manufacturing issues. The influence of layer structural characteristics (thicknesses) and intrinsic properties of each material are connected to device performance through mapping spectroscopic ellipsometry measurements collected over arrays of small area photovoltaic devices. Comparison of these mapping results enable a large amount of devices with slightly different properties, arising from spatial non-uniformity of each component layer, to be evaluated with only a few complete device fabrication processes. In this manner, principles guiding optimization of photovoltaic devices can be developed.

Localized plasmon resonances (LPR) can be used to efficiently concentrate EM energy by several orders of magnitude. They can be tailored to enable single-molecule chemical detection arising from Surface Enhanced Raman Spectroscopy (SERS), alternatively they can be used to develop hybrid photonic/plasmonic nanostructures such as NW-based plasmonic lasers with small cross section area mode capable to bridge the size gap between electronic and optoelectronic devices. We will first present an overview of the current efforts to develop hybrid photonic/plasmonic devices for energy and sensing applications with special emphasis their optical modeling.

The optical characterization of such plasmonic devices is thus a current issue of scientific and technological interest. Spectroscopic ellipsometry (SE) is a mature thin film characterization technique that can provide quantitative information about thickness, surface roughness, and dielectric functions of thin films using regression analysis and a transfer matrix formalism using Fresnel coefficients. For periodic samples, such as 1D gratings, the use of numerical simulation techniques, such as rigorous coupled wave analysis, SE has been capable to provide exquisite subwavelength structural characterization which is widely used for critical dimension metrology in the semiconductor industry. However, and for a number of reasons that will be discussed in the presentation, it is desirable to develop additional numerical capabilities to assist in data interpretation of complex samples. We have recently demonstrated that Finite-Difference Time Domain (FDTD) can provide quantitative SE data analysis of thin films and non-planar samples with sensitivity approaching 1/2 monolayer in the case of an ideal dielectric film on c-Si substrate. In the second part of this presentation we will summarize our recent breakthroughs in the use of FDTD as a powerful tool capable to provide quantitative modeling of SE data for plasmonic structures.

Structural phase transformation in thin films has a significant influence on properties, including electrical conductivity, hardness, and optical appearance. This presentation will address the underlying driving forces that contribute structural transformations in these materials. Using an in-situ laser interferometer curvature measurement technique, the intrinsic growth stress evolution was monitored for a series of Ti/Nb multilayered films. Each multilayer had a different bilayer spacing but each individual layer thickness within the bilayer was equivalent. Upon the HCP Ti to BCC Ti transformation, the overall compressive stress, for a given thickness, increased. This resulted in a measurable change in the interfacial stress, with the Nb layer exhibiting a more compressive response as compared to the Ti layer. Electron microscopy and atom probe tomography characterized the films. Molecular dynamic simulations were used in conjunction with the experimental results to explain the phase transformation in relationship to interfacial stress and chemical intermixing.

Different approaches have been developed concerning growth description of the compact nitride layers, especially those produced by ammonia. Nitriding by plasma uses a glow discharge technology to introduce nitrogen to the surface which in turn diffuses itself into the material. During this process, the ion bombardment causes sputtering of the specimen surface.

This paper presents a study of the kinetics of compound layer formation during plasma nitriding of pure iron. The study considers the erosion effect at the plasma-nitriding interface due to sputtering. Experimental data obtained by the authors as well as other sources are discussed. A mathematical model which simulates the process is introduced. The erosion effect is computer simulated and adjusted in order to consider its contribution to the study of layer growth kinetics. Finally, experimental and numerical results are compared in order to provide a better understanding of the process.