A microfluidic refractometer was designed based on previous optofluidic intracavity spectroscopy (OFIS) chips utilized to distinguish healthy and cancerous cells. The optofluidic cavity is realized by adding high reflectivity dielectric mirrors to the top and bottom of a microfluidic channel. This creates a plane-plane Fabry-Perot optical cavity in which the resonant wavelengths are highly dependent on the optical path length inside the cavity. Refractometry is a useful method to determine the nature of fluids, including the concentration of a solute in a solvent as well as the temperature of the fluid. Advantages of microfluidic systems are the easy integration with lab-on-chip devices and the need for only small volumes of fluid. The unique abilities of the microfluidic refractometer in this thesis include its spatial, temperature, and wavelength dependence. Spatial dependence of the transmission spectrum is inherent through a spatial filtering process implemented with an optical fiber and microscope objective.;A sequence of experimental observations guided the change from using the OFIS chip as a cell discrimination device to a complimentary refractometer. First, it was noted the electrode structure within the microfluidic channel, designed to trap and manipulate biological cells with dielectrophoretic (DEP) forces, caused the resonant wavelengths to blue-shift when the electrodes were energized. This phenomenon is consistent with the negative dn/dT property of water and water-based solutions. Next, it was necessary to develop a method to separate the optical path length into physical path length and refractive index. Air holes were placed near the microfluidic channel to exclusively measure the cavity length with the known refractive index of air. The cavity length was then interpolated across the microfluidic channel, allowing any mechanical changes to be taken into account.;After the separation of physical path length and refractive index, it was of interest to characterize the temperature dependent refractive index relationship, n(T), for phosphate buffered saline. Phosphate buffered saline (PBS) is a water-based solution used with our biological cells because it maintains an ion concentration similar to that found in body fluids. The n(T) characterization was performed using a custom-built isothermal apparatus in which the temperature could be controlled. To check for the accuracy of the PBS refractive index measurements, water was also measured and compared with known values in the literature. The literature source of choice has affiliations to NIST and a formulation of refractive index involving temperature and wavelength dependence, two parameters which are necessary for our specialized infrared wavelength range. From the NIST formula, linear approximations were found to be dn/dT = -1.4x10-4 RIU °C-1 and dn/dlambda = -1.5x10-5 RIU nm-1 for water.;A comparison with the formulated refractive indices of water indicated the measured values were off. This was attributed to the fact that light penetration into the HfO2/SiO2 dielectric mirrors had not been considered. Once accounted for, the refractive indices of water were consistent with the literature, and the values for PBS are believed to be accurate. A further discovery was the refractive index values at the discrete resonant wavelengths were monotonically decreasing, such that the dn/dlambda slope for water was considerably close to the NIST formula. Thus, n(T,lambda) was characterized for both water and PBS.;A refractive index relationship for PBS with spatial, temperature, and wavelength dependence is particularly useful for non-uniform temperature distributions caused by DEP electrodes. First, a maximum temperature can be inferred, which is the desired measurement for cell viability concerns. In addition, a lateral refractive index distribution can be measured to help quantify the gradient index lenses that are formed by the energized electrodes. The non-uniform temperature distribution was also simulated with a finite element analysis software package. This simulated temperature distribution was converted to a refractive index distribution, and focal lengths were calculated for positive and negative gradient index lenses to a smallest possible length of about 10mm.
展开▼
