Detectors for T N G

CCD CHARACTERIZATION REPORT

This document is available in postscript format.

INDEX

Vertical Clocks Waveform

It is shown the waveform of the parallel clocks used to drive the CCD. To have a good transfer efficiency a total time of 70 microseconds was settled to shift each row. It is important to note that the Multi Pinned Phase (MPP) operational mode has been selected. In fact, at the end of the waveform all the three clocks are at the low level.

Horizontal Clocks Output 0 Horizontal Clocks Fast Skip Output 0

Here are shown the waveform of the serial clocks used to read the pixel charge or to skip the unwanted one. The charge is coupled to the right output (0) on-chip amplifier. A time of 38 microseconds has been settled to read a pixel. About 160 seconds are required to read an entire image from one output.

Horizontal Clocks Output 1 Horizontal Clocks Fast Skip Output 1

Here are shown the waveform of the serial clocks used to read the pixel charge or skip the unwanted one from the left output (1) on-chip amplifier.

Conversion Factor (e-/DN) Out 0 Conversion Factor (e-/DN) Out 1

By setting the bias voltage levels and the other parameters to those reported on these pages we obtained a gain of 2.0 e-/DN and a readout noise of 6.8 e- r.m.s. in the out 0 case, and a gain of 2.45 e-/DN and a readout noise of 6.9 e- r.m.s in the out 1 case. The method used to evaluate the gain will be described in details in another document (in preparation). It is essentially based on the "photon transfer" method that consists in taking pairs of images at various signal levels. The statistics for gain evaluation of the two outputs is reported in these pages.

Output 00.5 s FF image (500 e-) 10 s FF image (9900 e-) 100 s FF image (96000 e-)

These pages show a flat field image at different signal levels (500 e-, 9600 e-, 96000 e-, i.e. from low levels to the saturation) with the charge is read from the output 0. At the bottom of each page is shown the mean value and the standard deviation calculated in the marked areas. It is visible the major problem of this CCD: 4 bad columns near the middle of the device.

Output 10.5 s FF image (500 e-) 10 s FF image (9900 e-) 100 s FF image (96000 e-)

 These pages show a flat field image at different signal levels (500 e-, 9900 e-, 96000 e-), with the charge is read from the output 1. The image read from this output shows a slightly bad column at low signal level.

30 min Dark image

The dark current rate is evaluated by taking three 30-minutes dark images and averaging them to remove the cosmic rays. This page shows the result of this operation and the statistics in the lower panel indicates a relatively low dark current rate.

Hot pixels

Hot spots with a dark signal greater than 30 e- in the 30 minutes exposure are mapped. The complete list is shown. Only two isolated pixel and a small cluster show a dark greater than 30 e-; all the other listed hot pixels are located in the 4 bad columns above mentioned.

Linearity

On this page the linearity of the response over all the dynamic range is plotted. The saturation, or the full-well allowed by this device, is about 110000 e-. The gain is set to 2 e-/DN and this means that the 16-bit A/D converter used in the TNG CCD controller is able to cover the whole dynamic range (65535 x 2 = 130000). Furthermore, the readout noise is well sampled (3-4 DN r.m.s.). Full-wells of 110000 electrons are typical for this kind of CCD operated in Multi Pinned Phase (MPP) mode. The linearity plot shows that the low level signal is obtained with a short exposure time, and this is the explanation of the 1.5 % deviation from linearity at low level. At shorter exposure times the accuracy of the shutter has to be taken into account. Further measurements with fainter sources and higher exposures will overcome the problem.

Quantum Efficiency in the 1400- 10500 Å

On this page the Quantum Efficiency (QE) in the 140 - 1050 nm range is plotted. This curve is obtained by using the optical apparatus described in the web page of our laboratory. Details on the measurements will be found in another document (in preparation). The special technique of backside charging, developed at the Steward Observatory, called "Chemisorption" and devoted to enhance the blue-UV response, applied to this device improve the QE response in all the visible and UV spectral range. In fact, our measurements show a 95% peak in the 550 - 650 nm range and 50 - 60% in the blue region. Furthermore a 40% of QE at 900 nm is observed. Information regarding the set-up conditions are given also.

QE vs AR coating thickness in the 2000- 10500 Å

On this page the Quantum Efficiency (QE) in the 200 - 1050 nm range obtained for this chip is compared with the QE obtained for the W1-1 chip. The difference between the two CCD is the HfO₂ thickness. As can be seen the 600 Å of HfO₂ on the CCD W1-1 (0,0) produces a red shift of QE peak.