1. Introduction

The previous report (Swift-UVOT/MSSL.TC/0002) showed capability of analog mode for high time resolution photometry, when a target star was brighter than 17.4mag. Swift UVOT may have several bright stars in the field of view during monitoring 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 life time, since the electric gain of MCPs is less than 1/6 of the photon counting mode.

Quantifying very bright star as a stand alone system is one of the safety requirements for Swift UVOT. Analog mode is one of the candidate to fulfil this requirement. It is, however, essential to know, how long does the detector withstand against the intense star illumination during the brightness assessment.

Our intensified CCD detector demonstrated far longer life time (>100 times) than typical position sensitive detectors in terms of accumulated anode charge (XMM-OM/MSSL/TC/ 0059). This difference may be due to the lower gain of MCPs with our detector, i.e. ~5x10^5 with ours while ~10^7 with the position sensitive detectors. If the life time depends on the gain strongly, the reduction of the gain by the factor of 6 may extend the life time far longer than x6.

In this report, the image intensifier was operated in the analog mode and was exposed to intense pinhole illuminations for 100 hours. Gain depletion of MCPs and sensitivity loss in F-F images were investigated against accumulated charge.
 
 

2. Electric gain

The best result in analog mode was obtained with 90% of nominal photon counting MCP voltage in the test with DEP_#5 intensifier. The experiments in this report was carried out with DEP_#8 intensifier, whose nominal MCPs voltage is 2250V. The 2020V, 90% of the nominal, was applied to the MCPs during the photon dose. The illumination pattern was a 11x11 pinhole array and the brightness has gradient along the column in the dynamic range of ~2E+4 (Fig. 1). The light source is made of 64 green LEDs covered with a diffuser and a 5300-5700A interference filter. The brightness of the LEDs is controlled by a constant current source in the dynamic range of ~2E+4 (see detail in XMM-OM/MSSL.TC/0057).
 
 

Table 1. Raw counts /(hour x spot) 21 June 2000 DEP_#8

Because of the large gain loss in the low input rate while no gain loss in the high input rate after the 100 hours photon dose, the gradient of the gain curve against input rate becomes flat. This suggests a very hard scrubbing may lighten the effect of pore paralysis, hence may extend dynamic range of MCPs.
 

Table 4. Electric gain of XMM-OM tube in high count rate

3. Gain loss of MCPs

A reference pulse height distributions for individual pinhole columns, col=4-11 (600-2E+6 c/s), were measured with V_mcp=2250V (photon counting mode) before starting the photon dose. The photon doses were carried out 3 times for 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 reference pulse height distribution and the one after the 100 hours dose by the 2E+6 c/s pinholes.

The gain reduced to 1/2.5 of the beginning. The gain loss was quantified from peak positions of the pulse height distributions. The gains after each dose were tabulated in Table 6, and were plotted against accumulated charge in Fig 6. Fig. 7 is the extract from XMM-OM/MSSL.TC/ 0059, in which the intensifier was operated in photon counting mode during the photon dose. The plots in analog mode coincides with those in photon counting mode. This implies that the gain loss can be described by the single parameter, accumulated charge, in any condition (i.e. different gain, input count rate, exposure time etc.).

Table 6. Gain depletion of MCPs

Ref-3 Files used in this section

4. Sensitivity loss in photon counting image

A reference F-F image with the blue (460nm) LED illumination was integrated for 15 hours in photon counting mode before starting the photon dose. F-F image integrations were followed after each intense illumination to see the sensitivity losses in the different level of dose. The integrations started after the sufficient period since the end of the intense illumination in order to avoid fluorescence, i.e. 80 hours in the 1st day, 38 hours the 2nd day and 27 hours the last day.

Fig. 8 shows 4 F-F images, one was taken prior to the photon dose for reference and the other 3 were after 15, 30 and 100 hours dose. The F-F after 100 hours dose clearly showed 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 the pinhole positions in the accuracy of 8um along H_intensifier direction (bias direction of the 1st MCP), while systematically shifted by 15um along V_intensifier direction. The day by day growth of the black spots is shown in Fig. 9. The image contains all factors, i.e. fluorescence, gain depletion and photocathode sensitivity loss. Fig. 10 shows profiles of the averaged black spots in the last day. Y-width of the slice is 3 twixel (= 58um). The depth of black spots reached 25% at the brightest pinhole after 100 hours dose. The black spots were noticeable for the illumination intensities of > 13kc/s after 100 hours dose, but not for the lower intensity illuminations. Since the F-F integrations were started after the sufficiently long period since the end of the photon dose, the peak depths were only little affected by fluorescence (less than 1.4%).

The 6-10 dark frames were acquired in photon counting mode before and after the F-F integrations in order to correct the effect of fluorescence for further precision. Standard fluorescence profiles are shown in the lower panel of Fig. 9. The size of the fluorescence is far larger than both of the black spots and projected pinhole, i.e. from 220um to 270um. The fluorescence spots have offset from the pinhole positions by 200-240um along H_intensifier direction (bias direction of the 1st MCP). The fluorescence spot got doughnut shape (i.e. black spot in central part) with the increase of the dose level. The details of these characteristics were identical to those described in the former report, XMM-OM/MSSL/TC/0057.

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 (1.4 in maximum) was subtracted. The results were tabulated in Table 7 and were plotted against accumulated charge in Fig. 11. The sensitivity did not decrease till 1E-5 couloms/spot. It started to decrease steeply from 1E-4 couloms/spot.
 
 

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 (threshold=15ADU). The results were tabulated in Table 8 and were shown in Fig. 12. The sensitivity loss of photocathode is not noticeable up to 3E-5 couloms/spot.

The photocathode sensitivity loss for the same intensifier but dosed in photon counting mode was extracted from XMM-OM/MSSL/TC/0057 (Fig. 13). The sensitivity loss dosed in analog mode shows slightly faster decrease above 2E-4 couloms/spot. This result did not imply that ion barrier characteristics of the 1st MCP improved with the lower voltage.

A Gaussian profile was fitted to the black spots to quantify the spatial extent. The results are shown in Fig. 14. There are clear correlation with accumulated charge. The width increased from 80um(FWHM) to 110um after acquiring 1E-5 to 2E-3 couloms/spot.

Table 7. Sensitivity change in blue F-F image at peak position

Table 8. Photocathode sensitivity change at peak position

5. Summary

i) Gain loss dosed in analog mode was same as that in photon counting mode in terms of accumulated anode charge.

ii) Photocathode sensitivity loss dosed in analog mode was slightly faster (by ~30%) than that in photon counting mode in terms of accumulated anode charge.

iii) Gain loss measured in the low input rate was 1/2 - 1/3 after 100 hours photon dose, while that in the high input rate almost nothing. In the consequence, pore paralysis curve was flatten after the dose. This suggests a possibility of extending dynamic range of MCPs by a hard scrubbing.
 
 
 
 

Appendix. Experiment procedure for DEP_#8 intensifier in analog mode