Reflectivity at different positions along CCD

Figure 4. Variation of total reflectivity across the wafer.

Wavelength (nm)

Figure 4. Variation of total reflectivity across the wafer.

Reflectivity at blue optimisation (minus centre) was measured with 10 nm bandwidth and 5^1 mm spot which suppresses fringe amplitude in the IR compared to reflectivity measured at red optimisation (plus centre), which was measured with a 1 nm bandwidth and 1 mm spot. With these differing measurement conditions absolute magnitude of fringing cannot be directly interpreted between different measured positions.

Reflectivity was also determined by dividing with a running average to give a normalised value; this emphasises changes in reflectivity and fringing effects. Figure 5 shows plots at the red (C+25 mm) and blue (C-20 mm) ends of the device.

The red end data (C+25) shows that the fringe amplitude is minimised at ~870 nm although the total reflection minimum is at ~800 nm. This is because fringes are minimum at the AR coating optimised wavelength (870 nm), which gives minimum backsurface reflectivity. However total reflectivity minimum (and QE maximum) can differ slightly due to the multi-layer structure of the device (Some reflection occurs from the frontside of the chip after double pass through the device at long wavelengths).

The blue end of the device shows fringe amplitudes that increase progressively with wavelength, as expected, since there is no reflectivity suppression for red wavelengths.

normalised reflectivity amplitude

wavelength (nm)

* centre +25mm — centre -20mm normalised reflectivity amplitude

wavelength (nm)

Figure 5. Normalised reflectivity, measured at blue and red ends of the CCD.

Figure 5. Normalised reflectivity, measured at blue and red ends of the CCD.

5. RESULTS FROM THE ISIS SPECTROGRAPH

Sample 44-82 devices on standard (16 ^m) and deep depleted (40 ^m) silicon were measured on the ISIS spectrograph at ING. The sensors were built into cameras at ING and tested on the red arm of ISIS on the William Herschel Telescope.

The first sample chip was a standard, thin CCD44-82 (non-deep depletion 2K*4K, 15^m pixels) with a graded Anti-Reflection (AR) coating. The coating varied along the length of the chip and was designed to match the low dispersion R158 ISIS red grating; dispersion was 0.12 nm/pixel. With the centre wavelength of this grating set to 600 nm, the spectrum incident on the chip was then exactly matched to the AR coat at every pixel. This chip was tested to show the importance of the AR coat for minimisation of fringing since this is crucial to allow good data extraction. The second sample was a deep-depletion CCD44-82 also with a similar graded AR coat.

Firstly, arc spectra were taken for alignment and wavelength calibration. For subsequent fringe measurements a tungsten spectrum was taken. The spectra were block averaged in the spatial (short-axis) direction (x4) to reduce the noise. A single image column was then extracted. The fringe amplitude was measured after a bias subtraction followed by flat-

field division. This flat field was obtained by block averaging the tungsten spectra by a factor sufficiently large to smooth out any fringes.

Two spectra were taken with the graded-AR chip. The first was with the grating angle set to give a perfect match between illumination wavelength and AR coat optimisation, across the whole chip area. The second was taken with the grating tilted so as to project a negative-order reversed-spectrum for comparison. The difference in the fringing between these two spectra showed the importance of a good AR coat (see Fig. 6). The optimally illuminated spectra showed very low fringing even though this was a thinned non-deep-depleted CCD.

Graded ARCptimiiTiIlliTrinaticn

1.10

Graded ARCptimiiTiIlliTrinaticn

1.10

650 700 750 800 850 900 950 Larda (nm)

Figure 6. Fringes in graded-AR standard silicon CCD (a) optimum, (b) anti-optimum. Note that the abscissa scale varies between the two plots.

6. CONCLUSIONS

A variable thickness AR coating can be designed and deposited across large-area CCDs with good accuracy. The measured reflectivity matches expected values, and the spectral response peaks correspond to reflectivity minima. At red wavelengths fringes are drastically reduced if reflectivity is a minimum at the projected wavelength. A graded-AR coating is ideal for use on fixed format spectrographs and can allow maximum QE and minimum fringing. The coating gradient can be tuned by design to match the application.

7. REFERENCE

[1] Andersen, M.J., 2000, Optimising CCDs for Spectrographs, Optical Detectors for Astronomy II, Amico, P. and Beletic, J.W., Editors, ASSL Vol. 252 p. 279.

Section IV: CMOS-Based Sensors

Was this article helpful?

0 0
Telescopes Mastery

Telescopes Mastery

Through this ebook, you are going to learn what you will need to know all about the telescopes that can provide a fun and rewarding hobby for you and your family!

Get My Free Ebook


Post a comment