As the microelectronics industry pushes microfabrication processes further, the lab-on-a-chip field has continued to piggy-back off the industry's fabrication capabilities with the goal of producing total chemical and biological systems on small chip-size platforms. One important function of such systems is the ability to perform single molecule detection. There are currently many methods being researched for performing single molecule detection, both macro and micro in scale. This dissertation focuses on an optofluidic, lab-on-a-chip platform called the ARROW biosensor, which possesses several advantages over macro-scale single molecule detection platforms. These advantages include an amplification-free detection scheme, cheap parallel fabrication techniques, rapid single molecule detection results, and extremely low volume sample probing, which leads to ultra-sensitive detection. The ARROW biosensor was conceived in the early 2000s; however, since then it has undergone many design changes to improve and add new functionality to the lab-on-a-chip; however, water absorption in the plasma enhanced chemical vapor deposited silicon dioxide has been a problem that has plagued the biosensor platform for some time. Moisture uptake in the oxide layer of the ARROWs leads to loss of waveguiding confinement and drastically decreases the overall sensitivity of the ARROW biosensors. New ARROW designs were investigated to alleviate the negative water absorption effects in the ARROWs. The new waveguide designs were tested for resiliency to water absorption and the buried ARROW (bARROW) design was determined to be the most successful at preventing negative water absorption effects from occurring in the PECVD oxide waveguides. The bARROWs were integrated into the full biosensor platforms and used to demonstrate high sensitivity single molecule detection without any signs of water absorption affecting the bARROWs' waveguiding capabilities. The bARROW biosensors are not only water resistant, they also proved to be the most sensitive biosensors yet fabricated with average signal-to-noise ratios around 80% higher than any previously fabricated ARROW biosensors.