Fitting Methods: BI Device Model Parameters.

The spatially averaged quantum efficiency measurements described in section 4.7.2, together with the best fit absolute quantum efficiency model for detector S2 (w182c4) were used to infer absolute quantum efficiency as a function of energy for each BI device at 525 eV, 677 eV, 1740 eV, 2015 eV, 4509 eV, 5894 eV and 8040 eV. These inferred detection efficiencies exceed unity at three energies (525 eV, 677 eV and 2015 eV) in one or both BI detectors, by amounts ranging from 4% to 7%. Obviously, this result demonstrates the presence of significant systematic errors in our analysis of these data. As discussed above, we believe that one major error in our analysis arises because we do not now correct for the effects of source continuum; since the spectral redistribution functions for the BI and FI chips are quite different, the continuum contributes a larger proportion to the BI counting rate than to the FI counting rate. A preliminary discussion of this effect is presented in section 4.7.3. We have not yet completed a BI device response function model adequate for proper analysis of this data. For purposes of this report we have arbitrarily elected to retain the BI CCD deadlayer model published with the previous of of the ACIS Calibration report (October, 1997). Since continuum effects are expected to be smaller at higher energies (where the resemblance of BI and FI redistribution functions is much closer than at low energies), we have fit a model to the MIT CSR relative quantum efficiency data to determine the photosenstive thickness of each BI device, given this fixed deadlayer model. Both the fixed deadlayer model and best-fit depletion thicknesses for the BI devices are shown in Table 4.63. The quoted confidence interval for the depletion depth (15%) is based on an assumed relative quantum efficiency error of 5%.
 
 

Table 4.63: Adopted ACIS Flight Detector Quantum Efficiency Model Parameter Values

 

Location/Device	Model Parameters ($\mu m$)							Reference Detector
	Device-Specific Fit				Generic (from sibling devices)
	Si	SiO2	Si3N4	Depletion Depth	CS Si$^{\dag }$	CS SiO$_{2}^{\ddag }$	CS Width$^{\star}$
I0/w203c4r	0.297	0.350	0.022	65	0.35	0.45	4.1	w190c3
I1/w193c2	0.292	0.339	0.024	65	0.35	0.45	4.1	w103c4
I2/w158c4r	0.272	0.319	0.030	65	0.35	0.45	4.1	w103c4
I3/w215c2r	0.291	0.344	0.020	65	0.35	0.45	4.1	w190c3
S0/w168c4r	0.264	0.318	0.024	64	0.35	0.45	4.1	w190c3
S1/w140c4r	0.0001$^{\ast}$	0.023$^{\ast}$	N/A	30	N/A	N/A	N/A	w134c4r, w103c4
S2/w182c4r	0.270	0.231	0.030	76	0.35	0.45	4.1	w103c4
S3/w134c4r	0.0001$^{\ast}$	0.023$^{\ast}$	N/A	40	N/A	N/A	N/A	w203c2, w103c4
S4/w457c4	0.244	0.314	0.042	72	0.35	0.45	4.1	w190c3
S5/w203c4r	0.289	0.339	0.021	72	0.35	0.45	4.1	w190c3
1-parameter 90% conf. limits for
FI Devices:	$\pm 0.024$	$\pm 0.008$	$\pm 0.010$	$\pm 1.7$	fixed	fixed	fixed
BI Devices:	fixed	fixed	N/A	$\pm 15\%$	fixed	fixed	fixed
$\dag$: typical one-parameter 90 % confidence range from mesh experiments is $^{+0.06}_{-0.03}\, \mu m$.
$\ddag$: typical one-parameter 90 % confidence range from mesh experiments is $^{+0.17}_{-0.11}\, \mu m$.
$\star$: typical one-parameter 90 % confidence range from mesh experiments is $^{+0.3}_{-0.4}\, \mu m$.
$\ast$: Deadlayer model parameters for BI devices fixed at approximate values; see text.