Simulated vibrational sum frequency generation from a multilayer thin film system with two active interfaces

In the field of surface-specific vibrational sum frequency generation spectroscopy (VSFG) on organic thin films, optical interferences combined with the two-interface problem presents a challenge in terms of qualitative assessment of the data and quantitative modeling. The difficulty is amplified when considering systems comprised of more than a single material thin film layer. Recently, in our lab we have developed a generalized model that describes thin film interference in interface-specific nonlinear optical spectroscopies from arbitrary multilayer systems. Here, we apply the model to simulate VSFG spectra from the simplest multilayer: a system of two thin films, one of which is an organic small molecule and the other is a dielectric layer on a semiconductor substrate system where we idealize that the organic interfaces are equally VSFG active. Specifically, we consider the molecule N,N′-dioctyl-3,4,9,10-perylenedicarboximide (PTCDI-C₈) deposited on a silicon wafer with a thermally grown oxide dielectric. We present results for the four polarization experiments that sample the nonzero nonlinear susceptibility elements of macroscopically centrosymmetric materials (ssp, sps, pss, and ppp) and in two mIR frequency windows (the imide carbonyl stretches around 1680cm^-1 and the alkyl stretches around 2900 cm^-1) as a function of both thin film thicknesses with fixed input beam angles. We use frequency dependent refractive indices for all materials. The goal is to illustrate some of the intricacies contained in the intensity data of such systems. Of particular interest is the effect of the relative polar orientation of modes at the interfaces and the possibility of designing a system where the collected signal is exclusively attributable to a single interface. Our calculations indicate that in order to unambiguously identify the relative polar orientation one must experimentally vary an additional system parameter such as thin film thickness or input beam angle and for quantitative modeling one cannot ignore either interfacial contribution. The results show that proper modeling of thin film interference effects is essential for accurate data analysis and should include the frequency dependent refractive indices, especially for modes with larger mIR absorption cross sections, even when absorptive losses are small.