Radiation Damage from HRMA Focussed X-rays

The ACIS CCDs are sensitive to exposure to X-rays as well as to protons and particles with greater charges. Radiation tests with tungsten X-rays show that 200 krads is sufficient to change the operating point of a CCD by 2 Volts, which renders the CCD inoperative. In the case of AXAF with a focused beam of X-rays, a small voltage shift can create a pocket that can change the charge transfer efficiency in a given column of a CCD. The physical mechanism that creates this condition is the ionization of silicon dioxide atoms and silicon nitride atoms in the layer separating the gate structure from the buried channel in the CCD (see cross section below). This charge is trapped in the dielectric and can not be discharged, hence the buildup is cumulative. If sufficient charge accumulates over time, the charge transfer efficiency of the CCD is affected. For this reason, some precautions must be used in the exposure of ACIS to bright X-ray sources on orbit. The charging of the dielectric layer is a function of x-ray energy, being strongest at low energy in the front illuminated (FI) CCDs. The back illuminated (BI) CCDs are less sensitive to charging because the layer of silicon that is about 40 microns thick, overlays the insulation layer and provides shielding for this layer. A gate thickness of only about 0.6 $\mu$m shields the FI CCDs from the charging effect.

To compute the radiation damage from cosmic sources it is necessary to assume a spectrum for the source - or use a measured spectrum. To define the problem take the photon number spectrum to be given by N(E)dE, the absorption coefficient for silicon as $\mu_{SiO_2}(E)$, the absorption coefficient for SiO₂ as $\mu_{Si_3N_4}(E)$, and the absorption coefficient for Si₃N₄ as $\mu$_Si3N4, the densities of three components as , and the pixel area to be A_pixel. The cross section of the CCD containing two pixels is shown in Figure 6.46. X-rays from the focusing optics, which are located at the top of the page, can be imagined as directing a stream of X-rays down onto the pixel from above. The gate structure absorbs a fraction of the X-rays depending upon the energy, and the rest pass through with a small fraction absorbed in the Si₃N₄ layer and another fraction absorbed in the SiO₂. For a back illuminated CCD the X-rays are flowing upward on the page into the pixel where a significant fraction is absorbed in the Depleted Silicon region before it reaches the oxide layer separating the gates from the depleted silicon.
 
 

Figure 6.46:  Cross section of two CCD pixels showing
the different layers of absorbing material. The overlapping
portion of the gate structure is exaggerated.

 

In order to compute the amount of ionizing radiation deposited in the insulating layer between the depleted silicon and the gate structure, it is necessary to integrate over the incoming spectrum. The following integral serves to illustrate the dose of ionizing radiation received by the SiO₂.

D = Abs / Area (6.15)

where

The first term in 6.16 is the area of the telescope, the second term the fraction of the flux into a single pixel for the on-axis position (Fig. 6.47), the third term is the input photon number flux after the absorption by the interstellar medium has been included, the fourth term is the absorption in the aluminum/polyimide optical blocking filter, the fifth term the absorption in the SiO₂ overcoating, the sixth term the silicon in the gate structure, the seventh term the silicon nitride insulation layer and finally, the absorption in the silicon dioxide insulation layer. The actual gate structure is used in the calculation, but the fourth term with a gates subscript is a placeholder to represent this term. The divisor converts the absorbed energy into the absorbed energy in units of 100 ergs per gram of material in the pixel which is the dose in rads, D.
 
 

Figure 6.47:  Fraction of incident X-ray flux falling within central pixel (averaged
over all sub-pixel positions)

The following figures (6.48-6.51) present the results of computations for a range of input spectral parameters. The input spectrum has been normalized to one photon cm^-2-s^-1, as seen at the top of the Earth's atmosphere.
 
 

Figure 6.48:  Expected radiation damage from a 1 photon/cm²-s blackbody source
incident at the HRMA aperture.  Top: FI chip; Bottom: BI chip.

Figure 6.49:  Expected radiation damage from a 1 photon/cm²-s power law source
incident at the HRMA aperture.  Top: FI chip; Bottom: BI chip.

Figure 6.50:  Expected radiation damage from a 1 photon/cm²-s thermal plasma source incident at the HRMA aperture.  Top: FI chip; Bottom: BI chip.

Figure 6.51:  Expected radiation damage from a 1 photon/cm²-s thermal bremsstrahlung source incident at the HRMA aperture.  Top: FI chip; Bottom: BI chip.