Ultraviolet sensitivity of a teledyne-e2v EMCCD

Electron multiplying charged coupled devices (EMCCD’s) can provide significantly greater signal to noise ratios in low light conditions and/or for higher speed readout than traditional CCDs. Due to the electron multiplication before readout, the effective readout noise can be at the sub-electron level, enabling single photon counting. Traditional far UV (150 – 200 nm) imaging detectors have utilized micro-channel plates to detect usually scarce UV photons at low efficiency, amplify them into electron showers which strike a phosphor, allowing a silicon detector array to perform the final detection of the resulting visible light pulse. The typical efficiencies of UV photo detection with MCP systems ranges from a low of a few percent to as high as 25%. Given that the theoretical probability of absorption of UV photons in silicon is at least 30% in this wavelength range, then it should be possible to make use of a photon counting EMCCD to directly detect UV photons that is competitive with MCP performance. We approached Teledyne-e2v and they confirmed that a backside thinned EMCCD with their ‘astro no-coat’ process should provide reasonable quantum efficiency (ie. > 30%) in this range. The primary application in which we are interested is UV imaging of the aurora from space-based platforms. In this application there are system level advantages to replacing an MCP based detector with an EMCCD which is directly sensitive to UV illumination, namely the elimination of a high voltage power supply and higher spatial resolution. An MCP produces an electron shower which degrades image quality and also requires a relatively thick detector window which has to be accommodated in the imager optical design. We acquired five CCD201 engineering model EMCCDs with e2v’s ‘astro no-coat’ process, and incorporated one of these into a standard flexible liquid nitrogen cooled EMCCD camera produced by Nuvu Camēras. Once installed the EMCCD operation was confirmed with standard Nuvu Camēras test procedures. The camera was then mounted in a test vacuum chamber along with a McPherson UV monochromator so that the UV performance could be assessed. A NIST traceable photodiode was used for the absolute calibration. The resulting intrinsic QE was found to be 34% at 180 nm rising to 44% at 150 nm. The quantum yield was found to be quite low, only a few percent at 180 nm rising to only 1.13-1.18 at 150 nm. This is considerably lower than comparable results from CCDs where delta-doping has been used to improve the responsive quantum efficiency and also lower than a Teledyne-e2v CMOS sensor with the same surface treatment.