Fabrication and Characterization of Microcrystalline Silicon Near Infrared Photodiode Detector Pixel on Glass Substrate for Large Area Electronics
This thesis is focused on development of near infrared (NIR) photodetectors on glass substrate at low temperatures for large area electronics applications. In the first part of this thesis we study the optical properties of plasma enhanced chemical vapor deposition (PECVD) prepared hydrogenated microcrystalline silicon (mc-Si:H) material. We demonstrate that uc-Si:H film has an absorption coefficient of 10^3 cm^-1 at wavelength of 850 nm which is more than three orders of magnitude higher than the absorption coefficient in conventionally used hydrogenated amorphous silicon (a-Si:H) material in large area electronics. However, in spite of its high absorption coefficient in NIR region, we demonstrate that metal-semiconductor-metal (MSM) detectors based on uc-Si:H have a weak dynamic range (DR) of operation of about 50 at wavelength of 850 nm per 1 mW/cm^2 of incident optical power density. Furthermore we demonstrate that NIR DR for uc-Si:H MSM detectors is very close to the one for a-Si:H MSM device and hence uc-Si:H detector is not efficient for NIR light detection. As a result we decided to focus on photodiodes instead of MSM detectors as an alternate device structure with high DR capability. For this purpose we developed and characterized doped uc-Si:H layers and fabricated an n type-intrinsic-p type (nip) uc-Si:H photodiode with intrinsic layer thickness of 343 nm. This device shows an external quantum efficiency (EQE) of 2 % at 850 nm with a DR of almost 500 for 1 mW/cm^2 of incident optical intensity which is an order of magnitude higher than the one for uc-Si:H MSM detector. By increasing the intrinsic layer thickness to about 2 um we observed that the EQE was increased to 8% without a notable change in DR due to proportional increase in dark current level as compared to the photocurrent level. By introduction of a 1 um thick textured AZO back scatterer, however, we were able to reach an EQE of 19.2 % with DR of more than 1000 at wavelength of 850 nm per mW/cm^2 of incident optical density which is 20 times higher than the one for uc-Si:H MSM device. Furthermore, we developed an HSpice circuit model parameter extraction method for our high DR photodiode demonstrating a non-ideality factor of 1.54, reverse saturation current of 4.94 10^-11 A, shunt resistance of 1.35 G.ohm, series resistance of 191:3 k.ohm, and parallel capacitance of 40 pF in a photodiode with area of 500 by 500 um^2. In the second part of this thesis, we focused on development of a-Si:H thin film transistor (TFT) in order to fabricate a pixel circuit based on our high DR photodiode and test its feasibility for implementation of 2D imaging arrays based on uc-Si:H photodiodes. During the design of our TFT fabrication process, integration of our uc-Si:H photodiode had been taken into consideration and hence a bottom gate structure was adopted compared to the top gate architecture. In order to design a hybrid TFT/photodiode pixel circuit we needed to come up with an accurate HSpice model representation for our TFT. As a result we adopted the HSpice Level 61 transistor model and presented an step by step parameter extraction procedure for our TFT obtaining the 29 TFT parameters in this model. Our HSpice simulation results proved to accurately model the behaviour of the TFT in both above threshold and subthreshold regimes in comparison to the experimental TFT data. The fabricated TFT showed a very low threshold voltage of 3.6 V with an on/off ratio of 10^6 which is suitable for uc-Si:H photodiode pixel circuit design. In the third part of this thesis we focused on integration of the developed photodiode and TFT for realization of a hybrid photodiode/TFT pixel circuit for imaging arrays where we presented three different pixel design and fabrication processes based on the uc-Si:H photodiode and a-Si:H TFT. We discussed the design, simulation, analysis, fabrication, and experimental measurements of conventional pixel with one TFT and one photodiode. We demonstrated that the conventional pixel suffers from saturation problem and signal drift due to high dark current flow of the uc-Si:H photodiode (compared to a-Si:H photodiode) during pixel wait time. In order to solve the saturation problem we presented a design with integrated capacitance underneath the photodiode to enhance pixel capacitance. However, the enhanced pixel capacitance comes at the cost of slower response and the pixel still suffers from signal drift during wait time. As a result we proposed a new pixel design with two TFTs, one capacitor, and one photodiode which proved to reduce the signal drift of the pixel by an order of magnitude during the wait time. As a result this new proposed pixel shows promising characteristics for a large area NIR photodetector pixel on glass substrate which can be used in smart screens and wearable sensor applications.
200 University Avenue West
Waterloo, ON N2L 3G1