Quantum efficiency model for p+-doped back-illuminated CCD imager

An analytical model has been developed for predicting the spectral response of thinned, p+-doped back-illuminated charge-coupled device (CCD) imagers. The device is divided into two regions: a thin, uniformly doped p+ layer used to passivate the illuminated back surface from external electrical effects, and a p- region that extends from the p+ region across the approximately 10-micrometers thickness of the device to the potential well in the buried channel. The one-dimensional steady-state continuity equation for low-injection conditions has been solved analytically for the surface p+ region, which is characterized by electron diffusion length and coefficients appropriate for the doping level and a surface recombination velocity Sn that represents the loss of photoelectrons at the surface. All photoelectrons generated in the p- region are assumed to be collected in the buried channel because of the long diffusion length and the presence of a field sweeping the carriers into the CCD channel. The effect of multiple internal reflections on photoabsorption at long wavelengths is included. The quantum efficiency of this device is calculated as a function of the depth and recombination velocity of the p+ surface layer, using Sn as the only independent fitting parameter, and matches experimental results well over the wavelength range from 360 to 1100 nm.