that image is interesting - it looks to me like all the red and blue pixels are reading back as the decimal value 4094, while the green pixels have real signal. it does seem that the sensor is damaged, but not necessarily optically as i would not expect all those pixels to be stuck at 4094. that value is somewhat suspicious since if the sensor is a 12-bit design then 4095 is the maximum value you'd ever read back for a saturated pixel. but i would not expect the red and blue channels to be completely overexposed while the green channel seems reasonably OK, so it seems like something wrong in the sensor electronics rather than some optical/mechanical problem with scraping the bayer matrix...
rob