Comparison of the results

We compare the result of the representative device. We believe that the optimization of the operation (clock timing, clock voltage and bias voltage) has been almost completed. There might be, however, some space to improve the performance (e.g. readout noise) because we have not tried all the possible combinations of the parameters of the operation. .

Readout noise and time

The following figure shows the total readout time versus readout noise of several different CCDs. The time includes both transferring charges and sampling signals. We change the integration time of our correlated double sampling circuit. SITe looks pretty slow but we obtained tail after X-ray events when we reduced the width of the clocking more. Further investigation is necessary about SITe.

[Topdrawer file] [Excel file]

Quantum efficiency(QE)

The following figure shows the RELATIVE comparison of four types of CCDs. We set the peak value of the MIT(high-rho) at 1.0. The estimated error of the measurement is +-2%. See the discussion here regarding the error. All the data were measured here at the detector lab except W62C2 which was measured at Lick observatory. Because W6C1 was tested both here and Lick, we compared the results. The shapes of QE agreed very well and our relative QE value is 10% larger than the Lick's QE value. Therefore, a factor of 1.1 was multiplied to the QE of W62C2 measured at Lick when plotted in the figure for reference.

comparison41.gif (10714 バイト)

Non-uniformity of QE

MIT

CCID-20 from the R&D phase (Phase-1) has non-uniform light response especially in blue. It is called 'brick-wall pattern'. The laser annealing process produces the pattern. MIT/LL is exploring the parameters of the process and we expect to have more uniform device from the following phase 2 and 3.

Part of W6C1 image under 350 nm light (1000X1000)

The degree of no-uniformity which is defined by the (max - min)/mean in small area (which covers several black and white region) are the following,

W58C2	45 % at 360 nm	0.9 % at 960 nm
W6C1	31 % at 350 nm	0.6 % at 960 nm

SITe

Under blue light, SI001S shows small dots whose diameter are approximately 100 micron. The dots has 7% less light response. The following image shows how they look like (a part of 350 nm image). They disappear over 640 nm.

EEV

EEV001S also shows non-uniformity at 350 nm probably due to the laser stepping pattern. The (max-min)/mean is about 7 %.

Part of EEV001S image under 350 nm light (1000X1000)

Last update: 98/07/17

Satoshi MIYAZAKI satoshi@anela.mtk.nao.ac.jp