Документ взят из кэша поисковой машины. Адрес оригинального документа : http://xmm.vilspa.esa.es/calibration/documentation/epic_cal_meetings/200904/chen-anomalous-states.pdf Дата изменения: Wed May 14 18:30:29 2014 Дата индексирования: Wed Apr 13 13:40:51 2016 Кодировка: Поисковые слова: dark energy

Low energy noise on the MOS CCDs
C M Huber t Chen, CEA Saclay 31- 3-2009

- A.k.a. the MOS2CCD5 effect (Stuhlinger 2008) or `anomalous states' (Kuntz and Snowden 2008). - Appears as plateaux in spectra below 1 keV. - Present in multiple CCDs in both MOS 1 and 2. - No switch on/off detected within observation. - No correlation with radiation monitors. - Excess in Patterns 0, 2 and 4.

Low energy noise on the MOS CCDs ­ p. 1/1


Symptoms of the noise in images

MOS1 CCD2

MOS2 CCD5

Low energy noise on the MOS CCDs ­ p. 2/1


Event patterns 1 to 4
-

Charge sharing between two adjacent pixels produces patterns 1­4. Spectral line at E0 appears as diagonal line E 2 = -E 1 + E0 in the ODF. Identity of Pixel 2 depends on orientation: ie, pattern value:

-

-

Low energy noise on the MOS CCDs ­ p. 3/1


Excess of patterns 2, 4 is abnormal
Kuntz and Snowden (2008):

Lines: no noise symbols: noisy

Pixels with patterns 2 or 4 are in different potential wells, so charge-sharing rates should be lower, but they are not.

Low energy noise on the MOS CCDs ­ p. 4/1


Noise has correlation in ODF E1,E2 !

Pattern 2: Slope = -3/4 Pattern 4: Slope = -4/3 Key: Pixels w/ patterns 2 and 4 are read out consecutively, but in opposite order: Pixel 1 precedes Pixel 2 Pixel 1 follows Pixel 2

Low energy noise on the MOS CCDs ­ p. 5/1


A unified view in (E 1, E 2)-space
Patterns 2 + 4 (E 1 E 2):
-

Energies E 1 and E 2 in ODF are correlated for noise events. Slope = -3/4 instead of -1. Noise energy is actually discrete, and in multiples of 1.53; plateau formed by smearing due to -3/4 slope. Patterns 0, 2, 4 only: problem is in read-out.

-

Pattern-0 spectrum:
-

Low energy noise on the MOS CCDs ­ p. 6/1


Correlation is different in different CCDs

MOS 1 CCD 4

MOS 1 CCD 5

MOS 2 CCD 2

MOS 2 CCD 5

Low energy noise on the MOS CCDs ­ p. 7/1


Application: Are the central CCDs noisy?
MOS 1 CCD 1 MOS 2 CCD 1 MOS 2 CCD 1 Pattern-0 spectrum

(E1,E2)-space shows little sign of noise, but some spectra from MOS 2 CCD 1 appear to have (small) plateaux.

Low energy noise on the MOS CCDs ­ p. 8/1


Correlates w/ RAWX, iff RAWX < 4
Unique (E1,E2)-track in each of the columns RAWX=0,1,2.

Pattern-0 data hints at more noise tracks below threshold.

Low energy noise on the MOS CCDs ­ p. 9/1


No correlation w/ RAWY periodicity
Noise appears as dark/bright rows (same RAWY), but all noise tracks brighten/darken together.

Low energy noise on the MOS CCDs ­ p. 10/1


Charge injection is not the cause

Charge injection test shows no track in (E1,E2)-space.

Low energy noise on the MOS CCDs ­ p. 11/1


Summary
-

Energies E 1 and E 2 in ODF are correlated for lowenergy noise in MOS. Problem lies in read-out amplifier circuit, not in pixels. Interference w/ column/rowselect lines possible. True cause and mechanism still not understood.

-

Low energy noise on the MOS CCDs ­ p. 12/1