Excitation and readout of a thermally driven time-domain optical coherence tomography system

Abstract

This system presented in this thesis report will be put into practice in a sensory system for dental drilling. The aim of this system is to provide real-time feedback to the practitioner about the anatomical structure in close vicinity of the drill bit. Therefore, the system has to deal with in vivo, in situ bone tissue, bone tissue within a living organism that cannot be removed from the human body for analysis. Hence, non-destructive biopsy is desired. Optical Coherence Tomography (OCT) imaging is able to provide optical biopsy at a high resolution and in real time. With the use of Optical Coherence Tomography (OCT), e.g. blood vessels, nerves and sinuses close to the drill can be detected before the drill bit does damage to these vulnerable anatomical structures. Silicon process technology offers a promising platform for a relatively cheap Time-Domain Optical Coherence Tomography (TD-OCT) system. A first step into the direction of a single-chip TD-OCT system is the development of a thermally excited Optical Delay Line (ODL) or Thermo-Optical Delay Line (TODL). This thesis project is dedicated to the characterisation of the TODL in order to find the voltage excitation waveform that results in a linear variation in effective optical path length. By making use of the models, non-linearites can be compensated and the limited bandwidth can be extended by pre-emphasis of the voltage excitation waveform. Two models are presented, based on the general heat equation for conduction. One model utilizes Fourier analysis to solve the heat equation and is a linear model by definition. The second model uses numerical methods and is able to incorporate non-linearities, such as relations for thermal conductivity, diffusivity and the variation of the refractive index of silicon as a function of temperature. Measurements are conducted to verify the modeling. The linear model has been shown to perform poorly. The non-linear model was able to compensate for the non-linearities mentioned. However, first results showed that the TODL was behaving slightly slower than expected. The non-linear model was corrected for this behaviour by scalling down the relation for thermal diffusivity. The corrected non-linear model achieved a linear variation of the effective optical path lenght over a scanning range of 200 ?m at a line scanning rate of 10 kHz.