TY - GEN
T1 - Mismatch reduction for dark current suppression
AU - Sander, David
AU - Abshire, Pamela
PY - 2010
Y1 - 2010
N2 - In this paper we present a dark current suppression technique for low-light image sensor arrays fabricated in a standard CMOS process. It has been shown that reducing the reverse bias of a p-n junction minimizes the thermally generated dark current and increases the signal to noise ratio. While this work well for single sensors, arrays of sensors suffer from mismatch, which limits the ability to apply a consistent junction bias. In this work we show simulation results for a floating gate mismatch compensation technique which reduces this biasing mismatch. While simulations of an idealized structure provided a 40X reduction in mismatch compensation, a worst-case simulation provided only a 6X reduction in mismatch. For the idealized simulation, the standard deviation of the input-referred mismatch was reduced from 7.7 mV to 196 μV, while in the worst case simulation the standard deviation of the input-referred mismatch was reduced to 1.4 mV, corresponding to an approximate dark current reduction of 10X and 5X respectively.
AB - In this paper we present a dark current suppression technique for low-light image sensor arrays fabricated in a standard CMOS process. It has been shown that reducing the reverse bias of a p-n junction minimizes the thermally generated dark current and increases the signal to noise ratio. While this work well for single sensors, arrays of sensors suffer from mismatch, which limits the ability to apply a consistent junction bias. In this work we show simulation results for a floating gate mismatch compensation technique which reduces this biasing mismatch. While simulations of an idealized structure provided a 40X reduction in mismatch compensation, a worst-case simulation provided only a 6X reduction in mismatch. For the idealized simulation, the standard deviation of the input-referred mismatch was reduced from 7.7 mV to 196 μV, while in the worst case simulation the standard deviation of the input-referred mismatch was reduced to 1.4 mV, corresponding to an approximate dark current reduction of 10X and 5X respectively.
UR - https://www.scopus.com/pages/publications/79951882974
U2 - 10.1109/ICSENS.2010.5689925
DO - 10.1109/ICSENS.2010.5689925
M3 - Conference contribution
AN - SCOPUS:79951882974
SN - 9781424481682
T3 - Proceedings of IEEE Sensors
SP - 1696
EP - 1700
BT - IEEE Sensors 2010 Conference, SENSORS 2010
T2 - 9th IEEE Sensors Conference 2010, SENSORS 2010
Y2 - 1 November 2010 through 4 November 2010
ER -