Swift UVOT MULLARD SPACE SCIENCE LABORATORY UNIVERSITY COLLEGE LONDON Authors: H. Kawakami Life time performance of FM-intensifier in analog mode Document Number: Swift-UVOT/MSSL/TC/00??.01 11-July-00 Distribution: XMM-OM Project Office A Dibbens Orig. Mullard Space Science Laboratory K Mason A Smith T Kennedy B Hancock H Kawakami DEP 11 July 2000 Hajime Kawakami Life time performance of FM-intensifier in analog mode 1. Introduction The previous report (Swift-UVOT/MSSL.TC/0002) showed capability of high time resolution observation with analog mode by applying lower H.V.s to MCPs, when a target star was brighter than 17.4mag. Swift UVOT may have several bright stars in the field of view during chasing time variation of a gamma ray burster. It is dangerous for the intensifier to observe a bright star for a long time in photon counting mode (high gain in MCPs). The analog mode observation will offer the longer observation time safely, since gain in MCPs is less than 1/5 of the photon counting mode. This intensified CCD detector demonstrated far longer life time (>100 times) than typical position sensitive detectors in terms of accumulated electrons onto anode (XMM-OM/MSSL.TC/0059). This difference may be due to the lower gain in MCPs with our detector, i.e. 5x10^5 with ours while 5x10^7 with the position sensitive detectors. If the life time is extremely sensitive to the gain, the reduction of the gain by the factor of 5 may extend the life time far more than x5. In this report, the image intensifier was operated in the analog mode and was exposed to an intense pinhole illumination for 100 hours. Gain depletion of MCPs and sensitivity loss in F-F images were assessed against accumulated photons and electrons. 2. Electric gain The previous report (Swift-UVOT/MSSL/TC/0002) showed best results in analog mode when applying 90% of nominal MCP voltage for photon counting mode with DEP_#5 intensifier. This damage test was carried out with DEP_#8 intensifier, whose nominal MCPs voltage is 2250V. Therefore, 2020V, 90% of the nominal, was applied to the MCPs for the exposure to the intense pinhole illuminations. The mask pattern consists 11x11 pinholes and their brightness changes along columns by the factor of ~2E+4 (Fig. 1). The light source is made of 64 green LEDs coupled with diffuser and 5300-5700A interference filter. The brightness of the LEDs can be controlled by a constant current source by the factor of ~2E+4 (see detail in XMM-OM/MSSL.TC/0057). The brightness of the pinhole array was calibrated by 3 exposures with the 3 LED current levels, 1,3 and 10 in order to overcome small dynamic range of the detector. The lower LED current (L=1) was used for determining brightness ration among bright pinhole columns, while the medium LED current (L=3) was for faint pinhole columns (Table 1). The highest LED current (L=10) was only for the faintest pinhole column (col=1). Photon losses due to coincidence were corrected for precise photometries (Table 2). Finally, the absolute brightness of the pinhole columns at the LED current level of 10 is tabulated in Table 3. Table 1 Raw counts /(hour x spot) 21 June 2000 DEP_#8 ---------------------------------------------------- LED = 1 3 10 ---------------------------------------------------- col=11 264920.0 N/A N/A col=10 254720.0 N/A N/A col=9 32620.0 N/A N/A col=8 24460.0 N/A N/A col=7 2381.2 N/A N/A col=6 2131.9 23104.0 N/A col=5 158.2 1680.4 N/A col=4 84.1 1042.4 N/A col=3 22.5 186.2 N/A col=2 ( 7.0) 169.0 (253160.0) col=1 14.2 164.6 264110.0 Table 2 True counts /(sec x spot) 21 June 2000 DEP_#8 ---------------------------------------------------- LED = 1 3 10 ---------------------------------------------------- col=11 94.58122 col=10 90.11536 col=9 9.74329 col=8 7.26678 col=7 .69737 col=6 .62426 6.85783 col=5 .04627 .49191 col=4 .02459 .30502 col=3 .00658 .05446 col=2 (.00205) .04942 (89.43997) col=1 .00415 .04814 94.22338 ---------------------------------------------------- Table 3 Pinhole brightness at LED current level = 10 -------------------------------------------------------------------------column 1 2 3 4 5 6 7 8 9 10 11 -------------------------------------------------------------------------Brtness 94.22 96.73 106.59 597.0 962.8 13.4k 15.0k 156k 209k 1938k 2034k (c/s) B0 star 16.5 16.5 16.4 14.5 14.0 11.1 11.0 8.4 8.1 5.7 5.6 (mag) ------------------------------------------------------------------------- The ratio of gains between V_mcp=2250V (nominal) and 2020V were determined by both of the brightness of event splash at phosphor screen and anode current. The pulse height distributions of the event splash with the 2 different voltages to MCPs were shown in Fig. 2. A F-F with the count rate of 15,000 c/s (full area) was used as an input light source. It was difficult to determine the ratio accurately, because the pulse height distribution with V_mcp=2020V was squashed to the lower energy end. The brief ratio determined from the peak positions was > 5.5 times. The brighter F-F input was used for the measurement of the anode current to provide sufficient current at V_mcp=2020V. The detected count rate for the input F-F measured in photon counting mode was 86,100 c/s (full area). After the correction of the coincidence loss, the true incoming rate is estimated to be 94,000 c/s (full area). The procedure of the coincidence correction followed XMM-OM/MSSL.TC/0050. Where coincidence area of event splashes was assumed to be 12 (CCD_pixels)^2 from other 2 intensifiers, though there was no specific measurement for DEP_#8. A 99.91k Ohm resister was inserted at the anode cable, whose voltage was 8000V, and the small voltage drop across the resister was measured with a precision multimeter, FLUKE 87 IV, in the readout accuracy of 1uV. The resistance value was also calibrated by the FLUKE 87 IV. The small voltage drops were 1012uV and 151uV for V_mcp=2250V and 2020V. Hence, currents were 10.23nA and 1.53nA. The input impedance of Fluke 87IV is 10M Ohm, therefore anode currents were corrected by the factor of 1.01. The full detector area is (3.37 x256 CCD_pixels)^2, while photocathode area is circle with the diameter of 25mm. Since the anode current was induced from all photocathode area, incoming rate of electrons creating node current were 94,000 c/s * (D=25mm) / (3.37 x256 (CCD_pixels)^2 = 94,000 c/s * 1.2467 = 117,000 c/s. Therefore, the electric gain for low count rate is 5.4x10E+5 with V_mcp=2250V and 8.1x10E+4 with V_mcp=2020V. The ratio of the gain is x6.7 times. The electric gain for high count rate was measured using the pinhole illumination in the LED current range of L=1-10 for V_mcp=2250V and L=3-10 for V_mcp=2020V. Columns=1-9 of the pinhole array was blocked, so that the brightest 2x11 pinholes with nearly same brightness from columns=10-11 only were used for the illumination. Since voltage display of the FLUKE 87 IV was not stable in the last 2 digits (10uV, 1uV), the display was read 10 times and averaged for the lowest 2 illuminations (i.e. LED current levels=3 and 4 with V_mcp=2020V and L=1 and 2 with V_mcp=2250V). The results for the both of V_mcp=2250V and V_mcp=2020V were tabulated in Table 4 and were plotted in Figs. 3 and 4. The higher electric gains with pinhole input than F-F input at the low count rate are due to global gain variation of MCPs. 1) Electric gain of the intensifier is 5.7E+5 at low count rate with V_mcp = 2250V and 8.1E+4 with V_mcp = 2020V. 2) Gain depletion is 1/9.7 at the count rate of 2E+6 c/s with V_mcp = 2250V, while 1/8 with V_mcp = 2020V, assuming pore paralysis is negligible at the count rate of 100 c/s, Electric gain of MCPs at pinhole positions should have changed during the heavy photon dose. The anode current was measured after completing the 100 hours photon dose by illuminating exactly same pinhole positions. This gauges the level of the change before and after the photon dose. Again, columns=1-9 of the pinhole array was blocked, so that the brightest 2x11 pinholes from columns=10-11 only were used for the light source. The gains at the brightest pinholes for various input rate were tabulated in Table 5 and were plotted in Figs. 3 and 4 overlaying original gains. In spite of the large gain depletion at the low input rate, the gain in the saturated count rate does not change before and after the 100 hours dose. This is particularly true for the illumination above 1E+5 c/(sec x spot) with V_mcp = 2250V and 1E+6 c/(sec x spot) with V_mcp = 2020V. From these results we can assume anode currents at columns=10 and 11 were constant throughout the dose, hence we can determine total accumulated charge precisely. There is no measurement on the change of gain at other places, i.e. columns=1-9. Since the total accumulated charges themselves are smaller, the anode currents were hopefully same before and after the photon dose. Because of the large gain depletion at the low count rate region while no gain depletion at the high count rate after the 100 hours photon dose, the gradient of the gain curve against input rate becomes flat. This suggests very hard scrubbing may lighten pore paralysis effect. Table 4. Electric gain of XMM-OM tube in high count rate ------------------------------------------------------------------------LED Intensity Anode current (pA) Electric Gain (c/s pinhole) from 22 pinholes 2020V 2250V 2020V 2250V ------------------------------------------------------------------------ F-F 94000 1530 10230 8.1 E+4 5.4 E+5 L=1 92.35 (6.7) 184 (2. E+4) 5.7 E+5 L=2 352 132 585 10. E+4 4.7 E+5 L=3 1014 295 1370 8.3 E+4 3.8 E+5 L=4 3426 800 2880 6.6 E+4 2.4 E+5 L=5 16500 2330 9260 4.0 E+4 1.6 E+5 L=6 51000 5250 20200 2.9 E+4 1.1 E+5 L=7 139000 11100 43300 2.3 E+4 0.88E+5 L=8 410000 29700 114000 2.1 E+4 0.79E+5 L=9 984000 57800 225000 1.7 E+4 0.65E+5 L=10 1986000 102000 412000 1.5 E+4 0.59E+5 ------------------------------------------------------------------------ Table 5. Electric gain after 100 hours dose ------------------------------------------------------------------------LED Intensity Anode current (pA) Electric Gain (c/s pinhole) from 22 pinholes 2020V 2250V 2020V 2250V ------------------------------------------------------------------------ L=1 92.35 --- 64 --- 2.0 E+5 L=2 352 --- 219 --- 1.8 E+5 L=3 1014 103 664 3. E+4 1.9 E+5 L=4 3426 239 1810 2. E+4 1.5 E+5 L=5 16500 1041 6550 1.8 E+4 1.1 E+5 L=6 51000 2790 16900 1.6 E+4 0.94E+5 L=7 139000 7700 38400 1.6 E+4 0.78E+5 L=8 410000 24000 108000 1.6 E+4 0.75E+5 L=9 984000 50200 225000 1.4 E+4 0.65E+5 L=10 1986000 94100 419000 1.3 E+4 0.60E+5 ------------------------------------------------------------------------ Ref-2 Files used in this section /swift/ZPHD010.dat ZPIN011.dat,ZPIN012.dat,ZPIN013.dat,ZPIN014.dat 3. Gain depletion Pulse height distributions (hereafter, PHD) for individual pinhole columns, 4-11 (600-2E+6 c/s), in the pinhole array were measured with V_mcp=2250V before starting the photon dose as reference. The photon doses were followed 3 times, 15 hours, 15 hours and 70 hours with V_mcp=2020V. The pulse height distributions were measured after the each photon dose. Fig. 5 shows the original PHD before the dose and the one after the 100 hours dose by the 2E+6 c/s pinholes. The gain reduced to 1/2.5 of the original. The gain depletion was quantified from peak positions of the PHD. Day by day change of the gain is tabulated in Table 6. The results were plotted against accumulated charge as shown in Fig 6. Fig. 7 is the extract from XMM-OM/MSSL.TC/0059, in which the intensifer was operated in photon counting mode. The plots of anolog mode coincides with that of photon counting mode. This implies that the gain depletion can be described by the single parameter, accumulated charge, for any condition (i.e. different gain, input count rate, exposure time etc.). Ref-3 Files used in this section /swift/ZPHD016.dat,ZPHD028.dat,ZPHD047.dat,ZPHD064.dat 4. Photocathode sensitivity loss The plot of photocathode sensitivity loss against accumulated photons showed split branches in high dose end according to the illumination intensities (Fig. 35 of Report-B). Again, this seemed to be due to the pore paralysis of MCPs. Photocathode sensitivity loss was plotted against anode current in Fig.51, using Table 28 and assuming constant electric gain throughout the photon dose. The split branches seen in Fig. 35 (v.s. accumulated photons) merged together in Fig. 51. This result implies that ion feed back is proportional to the electron cloud at the 2nd MCP. 4. Sensitivity loss in photon counting image A F-F image with the blue LED (460nm) was integrated for 15 hours in photon counting mode after each intense illumination to see the impact on science image. The integration started at the elapsed time of 80 hours for the 1st day, 38 hours the 2nd day and 27 hours the last day, since the end of the intense illuminations to avoid fluorescence. Fig. 8 shows 2 raw F-F images, one taken prior to the photon dose for reference and the other after the 100 hours dose. The F-F after the dose clearly shows an array of black spots corresponding to the pinhole positions. A F-F image in each day of photon dose was divided by the reference F-F to remove detector artefacts and illumination non-uniformity. Then, the 11x11 array of black spots were averaged along the columns to improve S/N. Central positions of the black spots coincided with pinhole positions in the accuracy of 10um. The day by day growth of the black spots is shown in Fig. 9. These images contain all factors, i.e. fluorescence, gain depletion and photocathode sensitivity loss. White spots appeared at 2E+6 c/s pinhole positions in the 1st day, as the fluorescence dominated photocathode sensitivity loss and MCPs gain depletion. The black spots are seen for the illumination intensities of > 19kc/s after the 100 hours dose but not obvious for the illumination intensities of < 2.1kc/s. This is big improvement from the DEP-QM intensifier, in which black spots were clearly seen for the illumination intensities of 0.8kc/s after 21 hours dose (ref. XMM-OM/MSSL/TC/0044). Fig. 32 shows profiles of the averaged black spots from the 5th to the 9th days. Y-width of the slice is 3 twixel (= 58um). Since the integrations were started after the decay of fluorescence for these 5 F-Fs, the peak depths were not affected by fluorescence more than 0.8%. The depth of black spots reached 30% for the brightest illumination after 100 hours dose. The sensitivity loss at the peak position was quantified from the average of 3x3 twixels square centred on the black spots. The normalization level was determined from 37x37 twixels (=717um) square excluding central D=21 twixels circular area. Then, the effect of fluorescence (3.8% in maximum) was subtracted. The results were tabulated in Table 19 and were plotted against accumulated dose events in Fig. 33. The sensitivity did not decrease up to 1E+8 dose events. It started to decrease steeply from 1E+10 dose events. The sensitivity decreased more slowly for the brighter pinholes. This is again the effect of pore paralysis. The sensitivity loss for DEP's QM-intensifier was plotted in the same frame as shown in Fig. 34. The QM-intensifier already lost sensitivity by 3% at 3E+7 dose events, while the DEP_#8 intensifier did not up to 3E+9 dose events. The ruggedness of DEP_#8 intensifier is clear at the lower dose events. The sensitivity loss in F-F image was averaged over central D=210um (=11 twixels) circular area to characterize spatial extent of damage as well as the depth. The results are tabulated in Table 20 and plotted in Fig. 35 after the correction of fluorescence. A Gaussian profile was fitted to deep black spots to investigate the spatial extent directly. The results are tabulated in Table 21 and shown in Fig. 36. The width of the black spots increased with accumulated dose events. It started from 80um(FWHM) and reached 120um after acquiring 1E+12 dose events. The sensitivity loss seen in F-F image is the combination of gain depletion and photocathode sensitivity loss. The photocathode sensitivity losses were calculated by removing the effect of gain depletion. The results were tabulated in Table 22 and were shown in Fig. 37. The sensitivity loss of photocathode is not obvious up to 1E+9 dose events. The plot has large scatter at larger dose events, since the calculation becomes less accurate when the gain depletion is large. For instance, 20% of photo-events were lost due to the gain depletion at 2E+6 c/s pinhole position after 100 hours dose. Table 19. Sensitivity change in blue F-F image at peak position --------------------------------------------------------------------------------------------- Total dose Pinhole intensity (counts/sec) (hour) 120c/s 210 1.1k 2.1k 19k 21k 220k 320k 2070k 2070k --------------------------------------------------------------------------------------------- 0.3 .996 1.004 .989 1.000 .996 .991 .994 .984 .979 .981 1.0 .994 .980 .990 .983 .986 .992 .999 .988 .980 .983 3.0 .997 1.003 .996 1.000 .983 .986 .981 .967 .968 .971 7.0 .997 1.002 .998 1.012 .971 .981 .963 .961 .911 .928 15.0 .998 .995 .984 1.000 .970 .975 .934 .937 .870 .894 30.0 .999 1.002 .988 1.001 .954 .960 .905 .896 .785 .821 50.0 .994 .992 .984 .990 .942 .950 .883 .878 .747 .779 70.0 1.009 .996 .979 .990 .929 .917 .844 .836 .667 .694 100.0 1D .996 1.000 .974 .979 .932 .925 .824 .819 .623 .665 100.0 2D 1.003 .980 .976 .981 .942 .938 .881 .848 .652 .698 100.0 5D 1.001 .990 .979 .981 .948 .944 .879 .852 .663 .713 --------------------------------------------------------------------------------------------- Appendix. Experiment procedure for DEP_#8 intensifier in analog mode 20 June - 6 July 2000 ---------------------------------------------------------------------- File Name Pinhole PHD Dark F-F Time(start) ---------------------------------------------------------------------- Before Damage for reference 2000/06/20 PHD010 300FR 17H 51M 05S Pin011 L=3 54000S 19H 01M 34S 2000/06/21 Pin012 L=1 3600S 10H 17M 16S Pin013 L=10 1800S 15H 11M 25S Pin014 L=10 1800S 16H 21M 36S DEP015 54000S 18H 13M 44S 2000/06/22 PHD016 70000FRs 11H 27M 18S Ana017 10000FRs 12H 37M 00S Ana018 10000FRs 16H 50M 13S \/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/ 15 hour Day-1 18:24 - 09:24 2000/06/22 \/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/ 2000/06/23 Drk019 7200S 10H 36M 54S Drk020 7200S 12H 37M 18S Drk021 7200S 14H 37M 42S DEP022 Th=15 54000S 17H 48M 35S 2000/06/24 Drk023 7200S 08H 48M 59S Drk024 7200S 10H 49M 22S Drk025 7200S 12H 49M 45S Drk026 7200S 14H 50M 08S 2000/06/26 Ana027 20000FRs 12H 28M 17S PHD028 70000FRs 14H 34M 34S Drk029 7200S 15H 37M 46S DEP030 Th=15 54000S 17H 38M 10S 2000/06/27 Drk031 7200S 08H 38M 34S Drk032 7200S 10H 38M 58S \/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/ 15 hour Day-2 13:15 - 04:16 2000/06/27 \/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/ 2000/06/28 Drk033 7200S 12H 41M 57S Drk034 7200S 19H 37M 23S Drk035 7200S 21H 37M 47S Drk036 7200S 23H 38M 11S 2000/06/29 Drk037 7200S 01H 38M 35S Drk038 7200S 03H 38M 59S Drk039 7200S 05H 39M 23S PHD040 50000FRs 10H 41M 54S Ana041 30000FRs 11H 23M 58S Drk042 7200S 16H 52M 29S DEP043 54000S 18H 52M 53S 2000/06/30 Drk044 7200S 09H 53M 17S Drk045 7200S 11H 53M 40S Drk046 7200S 13H 54M 03S PHD047 50000FRs 17H 24M 09S \/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/ 70 hour Day-3 18:05 - 16:05 2000/06/30 \/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/ 2000/07/03 Drk048 7200S 17H 20M 36S Drk049 7200S 19H 21M 01S Drk050 7200S 21H 21M 25S Drk051 7200S 23H 21M 49S 2000/07/04 Drk052 7200S 01H 22M 13S Drk053 7200S 03H 22M 37S Drk054 7200S 05H 23M 01S Drk055 7200S 07H 23M 25S Drk056 7200S 09H 23M 49S Drk057 7200S 11H 24M 13S Drk058 7200S 13H 24M 37S Drk059 7200S 15H 25M 01S DEP060 54000S 19H 26M 24S 2000/07/05 Drk061 7200S 10H 26M 48S Drk062 7200S 12H 27M 12S Ana063 30000FRs 15H 47M 40S 2000/07/06 PHD064 50000FRs 09H 39M 36S ----------------------------------------------------------------------------------------------------------