Abstract

Use of CCD detector arrays as visible imagers in space telescopes has been problematic. Charge-coupled devices rapidly deteriorate due to damage from the high radiation environment of space. CMOS-based imagers, which do not transfer charge, offer an alternative technology that is more tolerant of a high-radiation environment.

This dissertation evaluates the performance of four "pathfinder" 1K by 1K hybridized silicon P-I-N detector arrays made by Raytheon under subcontract to RIT as candidates for use in a space telescope application. Silicon P-I-N arrays have photon capture properties similar to back-thinned CCD's and should be far more robust than CCD's in the high-radiation environment of space. The first two devices, 180 µm thick prototypes, demonstrate crisp imaging with lateral diffusion of 5 microns at 35 Kelvin. The nodal capacitance is estimated to be 41 fF and the quantum efficiency is remarkably good (typically > 0.75) over a spectral range from 410 to 940 nm. A second pair of devices, fabricated with detectors thinned to 40 µm, exhibits similar performance but with blue-enhanced spectral response from an improved anti-reflective coating.

Operating, testing, and evaluating imaging devices similar to the ones tested here is also problematic. Precise, low-noise, flexible control systems are required to operate the devices, and interpretation of the data is not always straightforward. In the process of evaluating these pathfinder devices, this dissertation surveys and advances systems engineering and analysis (i.e. the application of linear and stochastic system theory) generally useful for operating and evaluating similar hybridized "staring" focal plane arrays. Most significantly, a previously unaccounted for effect causing significant errors in the measurement of quantum efficiency - inter-pixel capacitive coupling - is discovered, described, measured, and compensated for in the P-I-N devices. This coupling is also shown to be measurably present in hybridized indium antimonide arrays. Simulations of interpixel coupling are also performed and predict the coupling actually observed in the P-I-N devices.

Additional analysis tools for characterizing these devices are developed. An optimal estimator of signal on a multiply-sampled integrating detector in the presence of both photon and read noise is derived, modeling a pixel as a simple linear system, and is shown to agree with known limiting cases. Theories of charge diffusion in detectors are surveyed and a system model based on the steady state diffusion equation, infinite lifetime, and contiguous pixels is derived and compared to other models. Simulations validate this theory and show the effect of finite mean free path, finite lifetime, and non-contiguous pixels upon it. A simple method for modeling and evaluating MTF from edge spread is developed and used. A model that separately measures system and device noise in multichannel systems is developed, and shown to agree with measurements taken with the same device in both a quiet and a somewhat noisy system. Hardware and software systems that operate these devices are also surveyed, and 'agile' technologies and development methodologies apprate for detector research are employed to build a simple and flexible array control system, primarily from open-source components. The system is used to collect much of the experimental data.

Library of Congress Subject Headings

Metal oxide semiconductors, Complementary--Testing; Optical detectors--Testing; Imaging systems--Testing; Silicon diodes--Testing; Telescopes--Equipment and supplies--Testing

Publication Date

2005

Document Type

Dissertation

Student Type

Graduate

Degree Name

Imaging Science (Ph.D.)

Department, Program, or Center

Chester F. Carlson Center for Imaging Science (COS)

Advisor

Zoran Ninkov

Advisor/Committee Member

Ian Gatley

Advisor/Committee Member

William J. Forrest

Comments

Physical copy available from RIT's Wallace Library at TK7871.99.M44 M66 2005

Campus

RIT – Main Campus

Download

Share

COinS