One of the most important characteristics of CCD detectors is their extended sensitivity. CCDs have spectral response to photon's with energies extending over several orders of magnitude – from near IR to X-ray spectral regions.

Back-illuminated CCD detectors are actually able to efficiently detect radiation in a very wide spectral region, which ranges from 1100 nm. to 0.1 nm. -- four energy decades (1 eV-10 keV). In this technical note, the performances of a Scientific Imaging Technologies, Inc. (SITe) back-illuminated CCD are presented in the 0.3 to 1100 nm spectral region, which ranges from (4.5 keV-1 eV) and comprises near IR, Visible, near and far UV, EUV and soft X-ray spectral regions.

Moreover, the native oxide surface is very susceptible to charging effects. For example, signal charge will get trapped at the surface negating the positive oxide charge causing the backside well to shrink in size. This phenomena leads to an increase in quantum efficiency. However, over time, the trapped charge will recombine lowering the quantum efficiency to its original level. Quantum efficiency hysterisis (QEH) is a term used to describe such CCD behavior. Many groups use a high intensity UV lamp to stimulate this effect and exploit the enhanced responsivity. However, the temporal instability limits this solution for most practical implementations.

To achieve stable high and stable quantum efficiency, the backside of the CCD must be negatively charged to remove the backside depletion region and to drive the signal electrons toward the front surface where they can be collected, transferred, and detected.

X-ray detection occurs in a CCD when a high-energy primary x-ray dissipates its energy in the silicon of the back-illuminated CCD. Every 3.6 eV of energy lost by the x-ray photon generates approximately one electron-hole pair. In contrast to visible light, which is absorbed near the depletion region of the back-illuminated CCD, x-ray photons are absorbed within several tens of Angstroms from the CCD's back surface . Figure 2 shows the generation function of electro-hole pairs as a function of incident energy. (Figure 4 also shows the absorption depth of various energy photons) The left side of the graph represents the back surface of the CCD. As shown in the Figure, for electrons with energy levels below 2 keV generate electron hole pairs within tens of angstroms from the back surface. Thus, any surface states that are present due to improper passivation will cause the electrons to recombine and not reach the CCD’s depletion region.

Diffusion in the silicon separates the electron-hole pairs. The substrate connection collects the holes. The potential wells formed by the applied gate voltages collect the electrons, which constitute the amplified signal charge. To the first order, the signal is proportional to the energy of the photons prior to their impinging on the CCD back surface.

For electron energies below 2 keV, it can be assumed that all of the electron generation occurs within several tens of Angstroms from the back surface (see Figure 2) -- away from the depletion region. The probability of the electrons generated by the process being collected in the depletion region and not re-combining at the back-surface is shown in Figure 3 as a function of a number of CCD characteristic parameters, including Eo, the electric field created by the back surface p-type implant, and So the surface recombination velocity. Both of the variables are a function of the efficacy of the back surface processing and environmental effects on the back surface.

Shown in Figure 4 is the x-ray spectral response of a SITe back-illuminated CCD measured by three separate groups. Also shown in the Figure is the absorption length of x-ray photons in silicon as a function of their energies. Photons absorbed closer to the CCD back surface as subject to more variability to processing and environmental effects. In light of the discussions above, the variability of the measurements of the three groups can be aniticipated.

The tested detector exhibits high values of quantum efficiency and good uniformity of response in a very wide spectral region, representing a very useful, almost optimal detector of radiation in the extended optical domain.

As is anticipated by the discussion above, one group found variability during the quantum efficiency measurement in the EUV spectral range. They measured a slow decreasing of the quantum efficiency in time. After 2 hours from the CCD cooling at -30° C and with a pressure in the calibration chamber of 5.10-⁷ mbar, in the worst case at 93 nm., a 10 percent decrease in QE was measured. The effect is however critically dependent on the pressure inside the vacuum chamber: at wavelength of 121.6 nm., the decrease was up to 35% after 1 hour at pressure of 2.10-⁵ mbar, up to 20% after 1 hour at pressure of 7.10-⁶ mbar and up to 8% after 1 hour at pressure of 5.10-⁷ mbar.

A possible explanation for this decreasing can be found in the presence of some residual contaminants inside the vacuum chamber which deposit and condense on the cooled CCD surface, in particular water vapor.

It was noted that at -30° C the quantum efficiency is slightly higher than at -50° C and the decreasing in time is less evident. This effect can also be explained by the presence of water vapor which deposits on the CCD surface; at higher temperatures, the condensation decreases and, consequently, the quantum efficiency remains higher.

These observations point out the importance of stabilizing the environmental influences on the back-illuminated CCD.

The tested detector exhibits high values of quantum efficiency and good uniformity of response in a very wide spectral region, representing a very useful, almost optimal detector of radiation in the extended optical domain.

Results are highly dependent on CCD processing, and in the case of "unpinned" CCD back-surface treatement, the results are dependent on environmental conditions.