1. Micro-discharge debug with CCD Shuichi Shimma Univ. of Tsukuba Contents 1.Introduction for micro-discharge 2.Setup 3.I-V characteristics 4.Procedure of micro-discharge debugging 5.Results 6.Summary JPS meeting @ Tohoku Gakuin Univ. Mar. 29, 2003 ATLAS Japan Silicon Group (KEK, Tsukuba, Okayama, Hiroshima, Kyouto univ. of educ.) 1/14

2. 1. Introduction for micro-discharge Micro-discharge is caused by high electric field. edge design of electrode, bulk non-uniformity (defects), scratches on the silicon surface, etc. When the electric field becomes high, electron avalanche develops. In this situation, electron temperature becomes higher than lattice temperature. Hot electrons Using a cooled CCD camera, we can make images of infrared radiation from hot electrons. S. Shimma, JPS meeting @ Tohoku Gakuin Univ. : Mar. 29, 2003 2/14

3. 2. Setup Specifications HPK digital CCD camera Effective area: 12.3mm x 12.3 mm Effective # of pixels: 512 (H) x 512 (V) Pixel size : 24 um x 24 um Read out noise : 12 electrons r.m.s Dark current : 3 electrons/pixel/sec @ -50degC A/D converter : 16 bit We used, Spectral response characteristics wide-angle lens, (covers ~½ of the module) microscope lens, (enlarged view up to 300  m x 300  m). S. Shimma, JPS meeting @ Tohoku Gakuin Univ. : Mar. 29, 2003 3/14

4. 3. I-V characteristics Sample modules : 2022017020xxxx, from mass production (xxxx = 0069, 0099, 0100, 0101, 0102, 0106,0202) S. Shimma, JPS meeting @ Tohoku Gakuin Univ. : Mar. 29, 2003 4/14

5. 4. Procedure of micro-discharge debugging [0] Choose a module with micro-discharge, [1] Take a picture in light, [2] Take a noise image, [3] Take a hot spot image applying HV up to 550 V, [4] Subtract noise from hot spot image ([3] – [2]), [5] Superimpose [4] on the [1]. Exposure time 10 mins (wide-angle) 1 to 3 mins (microscope) CCD temperature : -55degC S. Shimma, JPS meeting @ Tohoku Gakuin Univ. : Mar. 29, 2003 5/14

6. 5. Results Modules 0069, 0101, 0102 and 0106. The cause of spots was sensor itself. 0069 (S1)0101 (S1)0106 (S4) Wide-angle view(upper), x125 magnification picture (lower) 0102 (S1) 7.8  A -> 5.6  A11  A -> 7  A6.7  A -> 4.5  A6.2  A -> 3.5  A S. Shimma, JPS meeting @ Tohoku Gakuin Univ. : Mar. 29, 2003 6/14

7. 5. Results (Cont’d) Module 0100 Enlarged Hot spot was observed on the sensor 4 (backside of the module). The cause of hot spot is investigated by the visual inspection. (X 12.5 maginification) 13  A -> 11.3  A S. Shimma, JPS meeting @ Tohoku Gakuin Univ. : Mar. 29, 2003 7/14

8. 5. Results (Cont’d) Visual inspection The cause of hot spot was a shave off a bonding wire. It was in contact with the surface of detector. Because of the contact, decay time of micro-discharge is longer. 0100 0106 S. Shimma, JPS meeting @ Tohoku Gakuin Univ. : Mar. 29, 2003 8/14

9. Module 0099 5. Results (Cont’d) Hot spot was observed on the sensor 1 (frontside of the module). (X 125 maginification)3.8  A -> 2.4  A The initial I-V curve was good. After the long-term test, micro-discharge appeared. Two hot spots exists at the neighboring electrodes. S. Shimma, JPS meeting @ Tohoku Gakuin Univ. : Mar. 29, 2003 9/14

10. Module 0202 (the worst I-V curve) 5. Results (Cont’d) 31.2  A -> 2.3  A (X 125 maginification) All sensors did not show any micro-discharge before module-assembly. Some problems occurred at module-assembly? S. Shimma, JPS meeting @ Tohoku Gakuin Univ. : Mar. 29, 2003 10/14

11. 5. Results (visual inspection) No scratches were found around the hot spot. Instead, several black dots were found near the hot spot. (X 250 magnification)(X 375 magnification) S. Shimma, JPS meeting @ Tohoku Gakuin Univ. : Mar. 29, 2003 11/14

12. 6. Summary We have started to locate micro-discharge points in series modules with the CCD camera. We examined 7 modules, their causes of microdischarge are; 0100 : shave off bonding wire, Sensor itself, 0069 : 0101 : 0102 : 0106 : 0099 : impurities (?), long-term applying HV, 0202 : dusts on the sensor surface. (discharge appeared at sensor test.) S. Shimma, JPS meeting @ Tohoku Gakuin Univ. : Mar. 29, 2003 12/14 These modules can be used as the test modules.

13. To identify the spots, we set the HV power OFF, the power to ASICs from SCTLV OFF and VME power ON. We still found the luminous spots. The enlarged view of the luminous spot is … X62.5 magnification According to chip designer: “The luminous spot is due to a diode for the prevention of ESD. When the difference of potential between Vdd and Reset B is larger than 0.6 V, the diode starts working.” ResetB = 1.5 V @ Vdd off We measured the ResetB with a tester. Misc.

15. To confirm the function of ESD diode, we took pictures with Vdd = 4.0 V. ResetB = 4.5 V @ Vdd on (design value : 4.0 V) Working correctly !! Previous spots disappeared. Other luminous spots were due to heat in the ASICs. Misc.