Modeling and Parameter Characterization of a Betavoltaic Cell

Darrell S Cheu, Purdue University

Abstract

Betavoltaic cells are a type of nuclear battery where kinetic energy from beta particles are converted into electricity. The goal of this research is to evaluate betavoltaic cell electrical performance and predict its response to temperature changes for potential implementation. To achieve this goal, three tasks were performed: betavoltaic cells were electrically characterized under temperature, critical betavoltaic semiconductor parameters were experimentally determined, and the Shockley-Diode model was used to predict electrical performance and compared to experimental results. Betavoltaic cells were evaluated from -30°C to 70°C. I-V curves were gathered at each temperature step in order to determine open circuit voltage and short circuit current. Open-circuit voltage was observed to decrease with temperature due to the increase in dark current from thermal excitation while short-circuit current increased with temperature due to the increase in mobility in electrons and holes. Open-circuit voltage was 0.75 V and short-circuit current was 70 nA at room temperature. Critical parameters, such as parasitic resistance and doping density were determined. Parasitic resistance was found by evaluating the slopes of I-V curves when I = 0 and V = 0 for shunt and series resistance, respectively, and were determined to be 2.3 × 108 Ω and 1 × 106 Ω, respectively. Doping density was found by determining the capacitance of the cell under AC voltage bias and was determined to be 1 × 1017 cm−3 . Absorption depths were determined in MCNP6 where a monoenergetic point source emitted beta electrons onto a GaAs substrate. Absorption depth was determined at the depth where 99% of energy was deposited into the GaAs substrate for all energies. Backscattering coefficients were also determined by the number of electrons passing through the top layer of the GaAs substrate. The number of particles emitted through the bottom face of the source film was determined in MCNP6 with the F1 tally. Critical parameters were used to model the NanoTritium TM cells with the Shockley-Diode model. The model was solved numerically using MATLAB's fzero function and was also solved explicitly using the lambert-W function. For I-V curves, the lambert-W function was inaccurate, producing curves that shifted 0.1 V, while solving the model numerically was accurate to experimental results. For determining both open-circuit voltage and short-circuit current, the numerical method was accurate while the lambert-W function could not determine results outside of certain temperature ranges.

Degree

M.S.

Advisors

Revankar, Purdue University.

Subject Area

Nuclear engineering|Energy

COinS