Berkeley Lab-UCO/Lick high-resistivity CCD's
Figures, mainly prepared for talks

1. Absorption of light in silicon

A picture goes here Absorption coefficient (intensity, not amplitude) of light in pure silicon. Under most conditions, an absorbed photon results in an electron-hole pair, one of which is detected. Note that the absorption length changes by nearly four orders of magnitude over the interesting optical range.

This is for room temperature, where the bandgap is at 1103 nm (1.125 eV). At -110 C the bandgap is at 1075 nm (1.154 eV).

Full size postscript version, (48616 bytes)

2. Transverse diffusion (related to MTF)

 picture goes here "Fakestar:" CCD is illuminated with blue light through a contact mask with array of 2 um, 4 um, and 8 um square holes, at two different back surface bias potentials. In the top case the substrate is not fully depleted. See Holland's Bruges paper for details.

Full size postscript version, (1.25 Mbytes)

2. LBLN/Lick CCD design concepts 

A picture goes here Conventional front-illuminated and thinned scientific CCDs. 

Full size postscript version, (48616 bytes)


A picture goes here The Berkeley Lab proposal: Use a high-resistivity (~10 kohm-cm) n-type substrate, and operate fully depleted. A conductive rear window (ITO with a possible SiO_2 coating) serves as rear contact and an antireflective window. 

The gate structure is conventional, as are the buried channel and oxide layer. An epitaxial layer is not necessary. 

Since n-type material is used, there is no harmful effect from electron accumulation at the rear surface. However, it is necessary to leave about 10 nm of doped polysilicon on the rear surface to maintain low leakage. 

Full size postscript version, (48 616 bytes)

2. Quantum efficiency

A picture goes here Quantum efficiency (QE) of the LBL/Lick CCd as compared with standard front-illuminated, thinned back-illuminated, and thinned back-illuminated antireflective coated CCDs. (The latter are from M. Blouke and M. Nelson, SPIE 1900, 228-240 (1993). 

The solid curve is the LBL/Lick QE as measured on an equivalent photodiode using a monochromater at LBL. The dots are nominally the same measurement, taken with a prototype CCD using narrow bandpass filters at Lick. The difference is not understood. 

The thickness of the single AR coating was chosen for maximal transmission at 400 nm. Future windows, with an additional quartz coating, should provide substantially better QE in the 1000 nm region. 

Full size postscript version, (81 846 bytes)

Many more figures will be added when I can find time!

Don Groom