Ex situ variable angle spectroscopic ellipsometry studies on chemical vapor deposited boron-doped diamond films: layered structure and modeling aspects
Metadata[+] Show full item record
[ACCESS RESTRICTED TO THE UNIVERSITY OF MISSOURI AT AUTHOR'S REQUEST.] Diamond possesses several outstanding physical properties (mechanical hardness, thermal conductivity, chemical inertness and high index of refraction) that may qualify it to become an engineering material. However, diamond is a wide bandgap semiconductor with E[subscript g] = 5.5 eV, but when doped with boron, the material becomes p-type semiconductor enabling various electronic and electrochemical applications. The motivation to perform this research is manifold that owes to the electrical conductivity increase with doping, wide optical transparency (from IR to near UV) and relatively large dielectric constant of the diamond materials which is utilized in a wide range of technological applications including electrochemical micro-electrodes, electro-optic devices (large coupling coefficients, i.e. [kappa]31 and [kappa]33) and bio-photonics, for instance. Essentially, all of the above mentioned applications require a detailed structural and physical property characterization with an objective to establish (micro) structure-property correlation. In this thesis, we present our recent optical and electrical property measurements on the boron-doped diamond films which were synthesized using microwave plasma-assisted chemical vapor deposition technique on Si (100) substrate using methane in high hydrogen dilution and trimethyl boron as a boron precursor. These films were analyzed using rotating analyzer variable angle spectroscopic ellipsometry (SE) from the near IR to UV range (830-193 nm). By applying the conventional Bruggeman effective medium theory and linear regression analyses to the raw data that is [psi[[lambda iota], [delta lambda iota] and pseudo-dielectric ( [less than] [epsilon]r [lambda iota] [less than] , [less than] [epsilon iota] [less than] ) function, we determined the most appropriate model fit. The SE modeling was performed using three different approaches including normal and point by point fit methods combined with the coupled and uncoupled bulk and surface layers approaches providing the details about the films' microstructure in terms of: a) the multilayer (main component layer and the surface layer) thicknesses; b) the volume fraction of the constituents (f[subscript sp3C] and f[subscript sp2C]) and of voids (f[subscript v]) in the layers; c) the inhomogenity of the surface along the growth axis; and d) optical (n, [kappa]) and corresponding complex dielectric ([epsilon]r, [epsilon iota]) constant. An estimator i.e. mean squared error (MSE, [chi][superscript 2]) is used to assess the accuracy of the model fit. A simplified three-layer model that simulates the SE data reasonably well is with the coupled point by point method. The results obtained through ellipsometry modeling, such as surface roughness and overall film thickness, were compared with those from atomic force microscopy to validate the model employed. The results (f[subscript v] and f[subscript sp2C) were discussed with respect to process parameters, especially boron concentration in gas phase, and complemented with other analytical tools including scanning electron microscopy, Raman spectroscopy, and electrical I-V property measurements. Typically, surface roughness values around 6 nm were found for films grown under different boron concentration which is almost five to six times smaller than those determined using atomic force microscopy. In this context, we determined an approximate linear relationship between these two variables. Lastly, we have also used Lorentz oscillator approach to determine the oscillator strengths and broadening and some of the preliminary results are presented with varying boron concentration.
Access is limited to the campus of the University of Missouri--Columbia.