Improving the experimental setup for ultrasound-optical tomography imaging

Sammanfattning: According to Bröstcancer förbundet, mammography is not efficient at detecting tumors in dense breast tissue or diagnosing breast cancer at its early stages. Ultrasound-optical tomography (UOT) is an imaging technique in development and has the potential for deep-tissue imaging. If ultrasound-optical tomography were implemented, it would be easier to differentiate between malignant, benign, and healthy tissue from any type of breast tissue. UOT is an imaging technique that takes advantage of high penetration depth and high spatial resolution of ultrasound imaging and optical imaging. In UOT, a laser light and an ultrasound pulse propagate through the tissue simultaneously at a frequency f$_L$ and f$_{US}$, respectively. The light will scatter while it propagates through the tissue and some of this scattered light will become frequency shifted by ultrasound pulse due to the acousto-optic effect. The tagged light will have the frequency $f_T = f_L + f_{US}$. The tagged (frequency shifted) light can be separated from the untagged light (unshifted light) using a thulium-doped lithium niobate, Tm$^{3+}$:$~$LiNbO$_3$, crystal as a filter. The crystal is kept at a temperature close to zero kelvin because then it exhibits unique characteristics, e.g. it has a narrow linewidth and long-lived hyperfine levels at this temperature. The filter is created by a method known as spectral hole burning (SHB). A laser beam is used to transfer electrons from the ground state to the excited state to create a hole at a specific wavelength. The spectral hole is created at the frequency of the tagged light and hence a narrow bandpass filter is constructed inside the crystal. The tagged light is fully transmitted through the filter while it highly attenuates untagged light. The tagged light is detected with a photodiode and processed in MATLAB after it has been transferred to an oscilloscope. This thesis aims to model and design a phantom probe that minimizes vibration and other unwanted movements or disturbances during measurements. The automated phantom holder will be used for the recording of 3D images. Another task of the thesis was to obtain the absorption spectrum of a 0.005$\%$ Tm$^{3+}$:$~$LiNbO$_3$ crystal when it is cooled down to 3$~$K to ensure that the crystal has the same absorption characteristics as predicted in literature. The absorption line at $\sim$ 800$~$nm is of interest since oxyhemoglobin and deoxyhemoglobin have similar absorption coefficients at $\sim$ 800$~$nm. Optical absorption and scattering information will help determine if the sample contains a cancerous region. The phantom probe was modeled in Solid Works and manufactured through 3D printing. In this setup, the sample holder was chosen to be translated while the ultrasound transducer was stationary to generate less blurry images. The design of the probe has to accommodate two detection schemes, reflection and transmission mode. The phantom probe was automated using a linear servo actuator since it was controlled with pulse-width modulation (PWM). It used a square signal as an input that could be generated with an Arbitrary signal generator (AWG). Using a device that operates with a signal was important because it would make it easier to integrate it into the experimental setup. The whole phantom probe was constructed in a cost-efficient way and in a way that it could be easily incorporated into the experimental setup. The absorption spectrum showed that the crystal has an absorption line at $\sim$ 794.3$~$nm. This absorption spectrum was compared to an absorption spectrum taken at 8$~$K on the same crystal and captured with a different method. Both absorption spectra had the same absorption peaks at almost the same wavelengths but they also showed few discrepancies that may depend on the temperature difference and the recording method. In this thesis, the absorption spectrum data taken was captured by sweeping the wavelength. The signal was captured with a photodiode, transferred to an oscilloscope, and then processed in MATLAB. The absorption spectrum data at 8$~$K was obtained using a Fourier transform spectrometer, resulting in data with little noise and well resolved peaks. To conclude, a functional and robust phantom probe was designed and manufactured that could withstand vibration and other undesired movements. An absorption spectrum of Tm$^{3+}$:$~$LiNbO$_3$ crystal was obtained at 3$~$K and compared to absorption taken at 8$~$K and compared to literature and previous measurements under similar conditions.