Date of Award


Degree Type


Degree Name

Doctor of Philosophy (PhD)



First Advisor

Garth J. Simpson

Committee Chair

Garth J. Simpson

Committee Member 1

Lynne S. Taylor

Committee Member 2

Adam Wasserman

Committee Member 3

Mary J. Wirth


The detection and characterization of crystallinity is critical throughout the drug development process. From the initial establishment of an active pharmaceutical ingredient’s (API) crystal structure to stability testing and quality control, the phase of an API affects the solubility, bioavailability, stability, and efficacy of a drug product. Second harmonic generation (SHG) microscopy has recently been developed as a selective and rapid method for imaging crystallinity in drug formulations. While SHG microscopy can enable the high signal-to-noise (SNR) detection of crystallinity, the intrinsic chemical information content within SHG images is relatively low. In cases of trace crystallinity and/ or small crystal volumes, new tools capable of rapid, qualitative crystal characterization are needed to fill this measurement gap. Several strategies for increasing the chemical information content of SHG microscopy were developed. Following combined computational and experimental studies to help determine the body of crystalline API structures amenable to imaging by SHG microscopy, measurements by confocal Raman spectroscopy and synchrotron X-ray diffraction were performed on regions of interest (ROI) identified by SHG. In both cases, spatial restriction of the spectroscopic technique to these regions of interest lowered the detection limits of Raman and synchrotron X-ray diffraction by several orders of magnitude. To further expand the capabilities of SHG microscopy, nonlinear optical Stokes ellipsometric (NOSE) microscopy was developed to assess crystal structure characteristics through the polarization dependence of SHG. Rapid (8 MHz) polarization modulation enabled NOSE microscopies at video rates (up to 15 Hz). Following development and validation, NOSE microscopy was used in conjunction with an iterative, nonlinear least-squares fitting algorithm to discriminate polymorphic crystal forms of the small molecule D-mannitol. Finally, to extend the linear dynamic range of photon counting measurements as described here-in, a novel digital filter derived from linear discriminant analysis (LDA) was developed and validated via theoretical and experimental nonlinear optical (NLO) measurements.