Files in this item



application/pdf1_Duggirala_Praveen.pdf (87MB)
(no description provided)PDF


Title:Molecular vibrational diagnostics - from quantum state stability to real time histopathology
Author(s):Duggirala, Praveen C.
Director of Research:Gruebele, Martin
Doctoral Committee Chair(s):Martin Gruebele
Doctoral Committee Member(s):Boppart, Stephen A.; Lyding, Joseph W.; McDonald, J. Douglas
Department / Program:Chemistry
Degree Granting Institution:University of Illinois at Urbana-Champaign
Subject(s):vibrational energy flow
IVR threshold
state space
vibrational adiabaticity
fractal dimensions
nonlinear microscopy
spectral interferometry
cancer diagnosis
Abstract:Polyatomic molecules at dissociative energies are generally believed to behave more statistically with increasing molecular size. We show that the size-invariance of the mean vibrational frequency and the molecular dissociation energy conspire to create a stable set of non-statistical quantum states that grows with molecular size. We derive a scaling relation for the number of non-statistical states as a function of molecular size. In accordance with the scaling model, we see highly regular vibrational progressions persisting beyond the dissociation limits in SCCl2. Nearly all of the SCCl2 transitions observed by stimulated emission pumping are assigned and fitted by a simple effective Hamiltonian without resonance terms, up to a total vibrational excitation of 36 quanta. The character of the highly excited vibrational wavefunctions gradually morphs out from the normal modes as energy increases. The number of sharp vibrational features (stable states) observed matches the scaling model predictions. The model and experimental studies are augmented by a numerical survey of ~3.5 million states in the anharmonic vibrational state space of SCCl2 up to the dissociation energy. We analyze the role of specific resonances and regions of state space in shaping the transition from restricted to free energy flow with increasing energy. Using different quantitative measures we find the transition threshold between 250-300 THz. Different resonances act at different energies in different parts of state space, thus softening the threshold. We also verify how the fraction of undiluted spectral features scales with energy. About 1 in 103 feature states remains undiluted at the dissociation limit of SCCl2. This fraction matches the experimental observations when symmetry and Franck-Condon factors are considered and is in agreement with the scaling model. New sensitive assays for rapid quantitative analysis of histological sections, resected tissue specimens, or in situ tissue are highly desired for early disease diagnosis. Stained histopathology is the gold standard, but remains a subjective practice on processed tissue taking from hours to days. We present proof of principle results showcasing the potential of nonlinear interferometric vibrational imaging (NIVI) for cancer diagnosis. NIVI combines chirped-CARS with spectral interferometry to capture background-free vibrational images over a spectral range of 200 cm-1. We show that NIVI acquires Raman-like spectra with the speed of CARS and that the phase retrieval is accurate enough even in diffusely scattering samples like biological tissue. We demonstrate its diagnostic potential by imaging and classifying normal and cancerous mammary tissue in a pre-clinical model of breast cancer. Tissue images acquired in < 5 minutes lead to clear pathological differentiation to beyond 99% confidence intervals. We also present the use of spectrally reconstructed NIVI (SR-NIVI) for real-time molecular histopathology. SR-NIVI combines NIVI with multivariate statistical methods for spectral reconstruction of diagnostic tissue maps. We use singular vale decomposition and logistic regression to reduce the diagnostic information in the NIVI spectra to a simple color code, to construct SR-NIVI images for visualization and decision-making. The sensitivity of SR-NIVI can improve further with diffraction-limited resolution and algorithm development for morphometric analyses.
Issue Date:2010-01-06
Rights Information:Copyright 2009 Praveen C. Duggirala
Date Available in IDEALS:2010-01-06
Date Deposited:2009-12

This item appears in the following Collection(s)

Item Statistics