Optical Blocking Performance of CCDs Developed for the X-ray Astronomy Satellite XRISM Hiroyuki Uchida (Kyoto University): [email protected] T. Tanaka, T. G. Tsuru, H. Okon, Y. Amano (Kyoto Univ.), H. Noda, T. Yoneyama, M. Hanaoka, K. Okazaki, K. Asakura, K. Hayashida, A. Ishikura, S. Sakuma, K. Hattori, H. Matsumoto (Osaka Univ.), K. Mori, Y. Kanemaru, J. Sato, T. Takaki (Univ. of Miyazaki), H. Tomida (ISAS/JAXA), J. S. Hiraga, Y. Urabe (Kwansei Gakuin Univ.), K. K. Nobukawa, M. Saito (Nara Women's Univ), M. Nobukawa, T. Sako (Nara Univ. of Education), H. Uchiyama (Shizuoka Univ.), H. Nakajima, A. Kashimura (Kanto Gakuin Univ.), S. B. Kobayashi, K. Hagino (Tokyo Univ. of Science), H. Murakami (Tohoku Gakuin Univ.), and Xtend team “Light Leaks” found in Hitomi CCD data Causes and Measures for XRISM CCD

In-orbit CCD (SXI) image aboard Hitomi (ASTRO-H) satellite CCD design of Hitomi We have been developing P-channel There are two origins that cause light-leak Optical Blocking Layer (OBL) CCDs for the upcoming X-ray events: Astronomy Satellite XRISM, planned One is "pinholes" found on the aluminum CCD Pinholes to be launched in 2022, based on optical blocking layer (OBL) deposited on the Die Bonding Sheet surface of the CCDs. the design of the Hitomi (ASTRO-H) Silicon Base CCDs Soft X-ray Imager; SXI The other is an end-surface leakage that One of the largest problems happens near the physical boundaries of the recognized in the Hitomi data is so- imaging areas. Cosmic Ray Events called light-leak events, which were To suppress the generation of pinholes, CCD design of XRISM

only found in the data taken during which is considered to be a deterioration of Double-layered (100+100 nm) OBL time when the backside of the the aluminum layer, we adopted a double- layered OBL with an 100+100-nm thickness. CCD Aluminum Plane End-surface Leakage spacecraft is toward the day earth. A similar design is applied for other X-ray Die Bonding Sheet

0 4 12 28 61 126 255 511 1029 2052 4091 Sun light from the day earth mission eRosita (Granato 2012). The main light paths that caused the light- Silicon Base leak events are originating from holes opened We also added an extra for other instruments on the back plane of aluminum plane (1-μm Hitomi. thickness) on the backside of Although XRISM is designed to close these the CCDs to block light from a holes to block the outside light, we further transparent die bonding sheet, improved the optical blocking performance of A similar design to block the end- cf. TEM image of the optical which is the main path of the surface leakage shown by Ryu et al. the CCDs as a kind of fail-safe design. blocking filter of the eROSITA (2017): see figure 11 CCD (Granato 2012) end-surface leakage. Optical Blocking Performance Test of the XRISM SXI CCDs

Test Sample CCD LED: Sato parts DB2-N (peak wavelength ~568 nm) Hitomi (ASTRO-H)

6‒10% pinhole pixels

Xtend CCD

LED

Visible LED LED-irradiated CCD images of ASTRO-H ↑ and XRISM ↓ Infrared LED 568 nm 940 nm

XRISM We developed test sample CCDs and irradiated optical/infrared LED light on them to evaluate <0.1% pinhole pixels their optical blocking performance. After confirming that the light leak was effectively reduced, we performed a CCD screening test to select the best-four chips to be mounted on the flight camera of XRISM (see also poster by Yoneyama et al.). In the screening test, the optical blocking performance test was also performed for all the 12 flight-model candidate CCDs.

Pinhole Events End-surface Leakage ASTRO-H FM Spare CCD XRISM Test Sample CCD The amount of the end- Threshold of the definition of “light leak” 3 103 10 = 40 ch surface leakage is also 2 102 10 significantly reduced (see 10 10

1 1 the projections shown in Number of Pixels that on-axis region exceed the Light Leak 1 10 102 103 1 10 102 103 (Field of view of the other Light Leak: LED - dark (ch) Light Leak: LED - dark (ch) the summary). instrument SXS on board XRISM) We compared the number of the “pinhole” pixels of the All the pixels in the 45 pinhole (light leak) pixels out of 10000 pixels (=0.45%) XRISM CCDs with that of the ASTRO-H spare by irradiating imaging area except for the cf. Hitomi CCDs: 5–7% the optical LED light on the detectors. outermost edges are not As a result, the light leak was effectively reduced affected by the end-surface compared with that of Hitomi's CCDs. The pinhole pixel leakage, which satisfies the rate of XRISM CCD is below 1/10 of that of Hitomi; much requirements of the XRISM less than the requirements of the XRISM mission. mission.

S/N Numbering by PI Number of pinholes Number of pinholes x10 luminosity Summary FM01-13 (ASTRO-H Spare) Selected as the flight model ↓ 27 = 0.03% (AB) 265 = 0.2% (AB) EQ103-1-07-CP01A-4 FM02-01 FM01-15 (ASTRO-H Spare) 15 = 0.02% (CD) 232 = 0.1% (CD) 11 = 0.01% (AB) 129 = 0.1% (AB) ✔ EQ103-2-10-CP01A-6 FM02-02 We have been developing P-channel Charge-Coupled Devices (CCDs) 12 = 0.01% (CD) 132 = 0.1% (CD) 23 = 0.02% (AB) 164 = 0.1% (AB) EQ103-2-12-CP01A-2 FM02-03 11 = 0.01% (CD) 139 = 0.1% (CD) for the upcoming X-ray Astronomy Satellite XRISM, planned to be 1503 = 0.9% (AB) 4914 = 3.1% (AB) EQ104-1-07-CP01A-6 FM02-04 1447 = 0.9% (CD) 5036 = 3.1% (CD) 29 = 0.03% (AB) 2111 = 1.3% (AB) launched in 2022. EQ104-3-19-CP01A-2 FM02-05 Hitomi in-orbit 30 = 0.03% (CD) 1965 = 1.2% (CD) FM02-01 83 = 0.08% (AB) 900 = 0.6% (AB) EQ104-4-24-CP01A-2 FM02-06 FM02-02 ✔ To reduce the light-leak events found from the in-orbit data of the 133 = 0.1% (CD) 1438 = 0.9% (CD) 79 = 0.08% (AB) 1015 = 0.6% (AB) FM02-03 EQ104-4-22-CP01A-2 FM02-07 60 = 0.06% (CD) 961 = 0.6% (CD) FM02-04 Hitomi (ASTRO-H) CCD, we changed the design of the XRISM CCDs and 73 = 0.07% (AB) 1144 = 0.7% (AB) FM02-05 EQ104-4-24-CP01A-6 FM02-08 139 = 0.1% (CD) 1236 = 0.8% (CD) FM02-06 42 = 0.04% (AB) 179 = 0.1% (AB) improved their optical blocking performance. ✔ EQ105-2-12-CP01A-6 FM02-09 FM02-07 15 = 0.02% (CD) 105 = 0.1% (CD) FM02-08 47 = 0.05% (AB) 182 = 0.1% (AB) ✔ EQ105-2-14-CP01A-4 FM02-10 FM02-09 ✔ As a result, the light leak was effectively reduced compared with 24 = 0.02% (CD) 143 = 0.1% (CD) FM02-10 ✔ 49 = 0.05% (AB) 65 = 0.0% (AB) EQ105-2-14-CP01A-6 FM02-11 XRISM CCDs that of Hitomi's CCDs. We thus conclude that the light leak found in 35 = 0.04% (CD) 82 = 0.1% (CD) FM02-11 14 = 0.01% (AB) 228 = 0.1% (AB) FM02-12 EQ105-3-21-CP01A-6 FM02-12 18 = 0.02% (CD) 162 = 0.1% (CD) FM02-13 ✔ 15 = 0.02% (AB) 92 = 0.1% (AB) Hitomi will be addressed by these two new designs. EQ103-2-11-CP01A-6 FM02-13 ✔ 18 = 0.02% (CD) 121 = 0.1% (CD)