Optofluidic intracavity spectroscopic biosensing system for single cell identification

Abstract

Low cost and label-free identification of single biological cells is of broad interest in a variety of fields, including clinical diagnostics, drug discovery, food safety, environmental monitoring, biology, and homeland security. The goal of this dissertation is to develop a low-cost biosensing system for label-free identification of single biological cells by combining optical and microfluidic techniques. Optical refraction of cells is closely related to cell size, shape, type, and biological states, providing sensitive probes for cell identification. Biological cells inside an optofluidic cavity modify the optical cavity modes, providing a probe for optical properties of the cells that can be used for cell differentiation. In this work, the perturbation effects on an optofluidic cavity transmission spectra induced by cells due to their larger refractive indices than the surrounding media were investigated. Passive optofluidic plane-plane Fabry-Pérot (FP) cavities fabricated on Pyrex glass substrates with a low temperature thermocompressive gold-to-gold diffusion bonding technique were used in the spectroscopic experiments. Complete fabrication processes for gold-patterned optofluidic cavities with HfO2/SiO2 dielectric mirrors using low cost transparency photomasks were developed and characterized. Critical fabrication steps that affect the optofluidic cavity's performance were examined by scanning electron microscopy, atomic force microscopy, and optical spectroscopy. Theoretical modeling of cavity finesse reduction due to mirror roughness and tilt shows good agreement with the measured value. An optofluidic cavity with a cavity finesse of above 40 and a mirror tilt <1° was demonstrated. A customized microscopic system was designed to measure the transmission spectra of the optofluidic cavity. Transmission spectra taken from various microspheres and biological cells excited by an LED at 890 nm are quantitatively different in terms of the number, relative position, and relative wavelength offsets of transverse modes that can be use for cell recognition. For example, intracavity transmission spectra of canine lymphoma cells are distinct from noncancerous lymphocytes, indicating the potential of this technique for cancer detection. The amount of transverse mode shift with respect to the bare cavity mode provides useful information on the refractive indices of cells that are important for cell differentiation. High order transverse modes induced by the transverse optical confinement of spheres and cells were found to be more effectively excited with tilted illuminations. Experiments were complemented by theoretical modeling of the transmission spectra of a cell-loaded FP cavity. Both the longitudinal optical path length change and the lateral optical confinements were thoroughly studied. Paraxial Gaussian beam resonator analysis provides insightful understanding of the effects of cavity length on resonator stability and transverse mode spacing. The range of cavity lengths for stable operation was found and should be considered in sensor design. Analysis of transverse mode spacing due to Guoy phase shift again indicated that a short cavity length would minimize transverse mode spacing variation with the longitudinal positions of the cell. A simplified double homogenous sphere model for cancerous cells was developed to study the spectral changes induced by the nuclei, which are important factors in pre-cancerous detection. The effective index method allows extraction of refractive index and nuclear size relationships from wavelength shifts and shows the impact of transverse mode confinement. Theoretical extraction from the experimental spectra indicates that the nuclei of cancerous cells are larger than those of normal ones. However, full numerical modeling based on the finite-difference time-domain (FDTD) method using commercial software did not produce consistent results. Two different fluid flow control mechanisms were explored to stabilize cells and microspheres within the microfluidic cavity. Pressure driven flow allows a simple microfluidic system to be used while dielectrophoretic (DEP) trapping with electrical fields allows cell immobilization in a more controllable way. Square well shaped dielectrophoretic traps were integrated onto the microfluidic chip by patterning ~10pm wide gold electrodes using a transparency mask. DEP trapping of various types of single microspheres and cells within optofluidic cavities was demonstrated with a 16 MHz, 10 Vpp AC voltage. Effects of DEP trapping on the bare cavity spectra were investigated by varying the AC voltages applied between the microelectrodes. The experimental results indicated that thermal effect was the major factor that caused a red shift of the bare cavity modes with an increase in trapping voltage. Future work is suggested including burying gold microelectrodes beneath the dielectric coatings to reduce the bio-contamination of the electrodes and investigating spectroscopic properties of DEP trapped biological cells. In summary, the biosensing system developed in this work may provide a low-cost, label-free optical technique for recognizing cells in a microfluidic environment.