Microcantilever arrays have been shown to be promising label-free nanomechanical sensing devices with high sensitivity. Two factors that affect the usefulness of microcantilevers in sensing scenarios are the sensitivity of the transduction method for measuring changes in microcantilever properties and the ability to create large compact arrays of microcantilevers. In this dissertation, we demonstrate that microcantilevers with an in-plane photonic transduction method are attractive because they maintain the sensitivity of the traditional laser beam reflection method while being scalable to simultaneous readout of large microcantilever arrays. First I demonstrate the integration of a compact waveguide splitter network with in-plane photonic microcantilevers which have amorphous silicon strip loading differential splitter and simultaneous microcantilever readout with an InGaAs line scan camera. A 16-microcantilever array is fabricated and measured. Use of a scaled differential signal yields reasonable correspondence of the signals from 7 surviving released microcantilevers in the array. The average sensitivity is 0.23 µm-1. To improve the sensitivity and consistency, and reduce fabrication difficulties, a new differential splitter design with 4 µm long double-step multimode rib waveguide is introduced. Furthermore, a modified fabrication process is employed to enhance the performance of the device. A new 16-microcanitiler array is designed and fabricated. The sensitivity of a measured 16-microcantilever array is improved to approximately 1 µm-1, which is comparable to the best reported for the laser reflection read out method. Moreover, most of the microcantilevers show excellent uniformity. To demonstrate large scale microcantilever arrays with simultaneous readout using the in-plane photonic transduction method, a 64-microcantilver array is designed, fabricated and measured. Measurement results show that excellent signal uniformiy is obtained for the scaled differential signal of 56 measured microcantilevers in a 64-array. The average sensitivity of the microcantilevers is 0.7 µm-1, and matches simulation results very well.

We have invented an in-plane all-photonic transduction method for photonic microcantilever arrays that is scalable to large arrays for sensing applications in both bio- and nanotechnology. Our photonic transduction method utilizes a microcantilever forming a single mode rib waveguide and a differential splitter consisting of an asymmetric multimode waveguide and a Y-branch waveguide splitter. The differential splitter's outputs are used to form a differential signal that has a monotonic response to microcantilever deflection. A differential splitter using an amorphous silicon strip-loaded multimode rib waveguide is designed and fabricated to demonstrate the feasibility of the in-plane photonic transduction method. Our initial implementation shows that the sensitivity of the device is 0.135×10^-3 nm^-1 which is comparable to that of other readout methods currently employed for static-deflection based sensors. Through further analysis of the optical characteristics of the differential splitter, a new asymmetric double-step multimode rib waveguide has been devised for the differential splitter. The new differential splitter not only improves sensitivity and reduces size, but also eliminates several fabrication issues. Furthermore, photonic microcantilever arrays are integrated with the differential splitters and a waveguide splitter network in order to demonstrate scalability. We have achieved a measured sensitivity of 0.32×10^-3 nm^-1, which is 2.4 times greater than our initial result while the waveguide length is 6 times shorter. Analytical examination of the relationship between sensitivity and structure of the asymmetric double-step rib waveguide shows a way to further improve performance of the photonic microcantilever sensor. We have demonstrated experimentally that greater sensitivity is achieved when increasing the step height of the double-step rib waveguide. Moreover, the improved sensitivity of the photonic microcantilever system, 0.77×10^-3 nm^-1, is close to the best reported sensitivities of other transduction methods (~10^-3 nm^-1).

Microcantilevers have been investigated as high sensitivity, label free biosensors for approximately 15 years. In nearly all cases, a thin gold film deposited on the microcantilevers is used as an intermediate attachment layer because of the convenience of thiol-gold chemistry. Unfortunately, this attachment chemistry can be unstable when used with complex sample media such as blood plasma. The Nordin group at BYU has recently developed an all-silicon in-plane photonic microcantilever (PMCL) technology to serve as a platform for label-free biosensing. It has the advantage of being readily scalable to simultaneous readout of many PMCLs in array format, and allows integration with polymer microfluidics to facilitate the introduction of biological samples and reagents. An essential processing step for the transformation of the PMCL into a practical biosensor is the ability to effectively immobilize active biological receptors directly on silicon PMCL surfaces such that ligand binding generates sufficient surface stress to cause measureable PMCL deflection. This dissertation presents the development of a method to functionalize the sensor surface of all-silicon in-plane photonic microcantilever (PMCL) arrays. This method employs a materials inkjet printer for non-contact jetting and a fluid that is custom designed for ink-jetting and biological applications with approximately 1 pL droplet size. The method facilitates the application of different receptors on select PMCLs with drop placement accuracy in the +/- 7.5 μm range. The functionalization fluid facilitates further processing using humidity control to achieve full coverage of only the PMCL's top surface and removal of dissolved salts to improve uniformity of receptor coverage and to prevent fouling of the sensor surface. Once a functionalization method was successfully developed, a series of experiments were performed to investigate the amount of surface stress that can be generated when receptors are immobilized directly to the silicon surface. In one series of experiments, a 4.8 μM streptavidin solution was used with biotin immobilized on multiple PMCLs to demonstrate adsorption-induced surface stress and concomitant deflection of the PMCL. The group observed ~ 15 nm PMCL deflection on average, with a corresponding surface stress of approximately 4 mN/m. These experiments yield the sensor response in real-time and employ a combination of multiple PMCLs functionalized as either sensors or unfunctionalized to serve as references. Investigation of various attachment chemistries is included, as well as a comparison with and without passivation of non-sensor surfaces. Investigated passivation strategies prevented ligand binding from generating a differential surface stress. Failure modes and physical mechanisms for adsorption-induced surface stress are discussed. Immobilization and passivation strategies for antibody-based biosensing are demonstrated with fluorescence microscopy and a corresponding PMCL sensing experiment using rabbit anti-goat F(ab') fragments as the receptors and Alex Fluor 488 labeled goat anti-rabbit IgGs as the ligand. While the results of these experiments remain inconclusive, suggestions for future research involving the PMCL sensor array are recommended.