Advances in microchannel plate detectors for UV/visible Astronomy Dr. O.H.W. Siegmund Space Sciences Laboratory, U.C. Berkeley Advances in:- Photocathodes (GaN, Diamond, GaAs) Microchannel plates (Silicon MCP s) Readouts (Cross strip) Are changing the tools and performance of photon counting imaging detectors available for future UV missions achieving better QE, lower background, higher resolution, better uniformity, linearity and better lifetimes. Work funded by NASA grants, NAG5-8667, NAG5-11547, NAG-9149 1
EUV Photocathodes,, 100Å - 2000Å Alkali Halide UV Photocathodes have improved substantially as a result of better fabrication techniques and geometrical optimization 0.7 1 QDE 0.6 0.5 0.4 0.3 0.2 CsI 1985 30 CsI 1985 20 CsI #3 2/99 20 CsI #3 2/99 30 CsI #2 1/99 20 CsI #2 1/99 30 Quantum Efficiency 0.1 CsI KI KBr 0.1 0 0 500 1000 1500 2000 Wavelength (Å) Opaque Alkali Halide Photocathodes - MCP substrate 0.01 0 500 1000 1500 2000 Wavelength (A) Opaque Alkali Halide Photocathodes - metal substrate 2
EUV Photocathodes,, 100Å - 2000Å Diamond Photocathodes on Silicon and Si MCP s Polycrystalline boron doped diamond, band gap - 5.47 ev (227 nm) - Solar blind. Hydrogenated and cesiated diamond exhibit NEA. Cathodes are air stable (<10% drop in 18 hours)and are very mechanically robust. Can be ultrasonic cleaned in water and alcohol. Re-hydrogenation restores QE. 1 Diamond coated Silicon MCP 0.1 Diamond photocathode UV efficiency QDE #2 0.01 #1 #5 #8 21201 Si MCP 20801 20501 Pre-hydrogenated values 0.001 0 500 1000 1500 2000 Wavelength (Å) Diamond Photocathodes on Silicon and Si MCP s 3
EUV Photocathodes,, 100Å - 2000Å Diamond cathodes have been grown on Si MCP s,, appropriately for EUV sensors Activation of diamond indicates that very high EUV DQE s are possible sible Diamond Photocathode on Silicon MCP (Nanosciences) At smaller incidence angles than 90 the QE improves significantly for planar layers 4
GaN Photocathodes,, 1000-4000Å Potentially high UV QE and cutoffs from 350nm to 450nm Basic strategy is to achieve NEA using Cs activation. Material, cleaning & activation processes all under study. Working with Northwestern U. (Ulmer, Wessels). 1 T samples processed by Nanosciences 2003 Quantum Detection Efficiency 0.1 0.01 Early GaN photocathode comparisons [Ulmer 2001] 0.001 100 150 200 250 300 350 400 Wavelength (nm) Opaque GaN photocathodes on Sapphire UCB supplied samples, Nanosciences processing 5
GaN UV Photocathodes,, 1000-4000Å 1 JG238/T/TT samples 2003 Ulmer et al 2001 GaN Opaque Photocathodes First tests showed poor QE with Cs activation - except at short wavelength. Quantum Detection Efficiency 0.1 0.01 Early sealed tube, MgF 2 window JG238 early attempt Initial sealed photodiode tubes were slightly better, and showed no degradation over > 6 months. Improvements in processing have improved the QE substantially. Cutoff is quite sharp at 370-390nm. 390nm. As NEA improves the longer wavelength QE and enhances the cutoff characteristics. 0.001 150 200 250 300 350 400 450 Wavelength (nm) Development of opaque GaN photocathodes on sapphire substrates, with Cs activation ~40% QE for < 250nm is significantly better than CsI or CsTe,, and shows promise for further >250nm improvements. 6
GaAs Visible and NIR Photocathodes GaAs has high visible/nir QE (50%), and noise is 10 events/sec at -20 C. Can adjust bandpass blue end response. Photon counting with time response ~1ns. No cosmic ray background effects. Can combine with high spatial resolution large area imaging MCP readouts. Astronomy - interferometry, time resolved spectroscopy and imaging. GaAs photocathode efficiency (courtesy ITT) GaAs photocathode blue variants (courtesy ITT) 7
Silicon MCP Developments Hexagonal pore Si MCP with ~7µm pores, >75% open area Silicon MCP s Silicon MCP s are made by photo-lithographic methods Photolithographic etch process - very uniform pore pattern No multifiber boundaries & array distortions of glass MCP s Large substrate sizes (100mm) OK, with small pores (5µm) High temperature tolerance - CVD and hot processes OK UHV compatible, low background (No radioactivity) Development in collaboration with Nanosciences. Typical Silicon microchannel plates in test program 25mm diameter (75mm currently feasible) 40:1 to 60:1 L/D (>100:1 possible) 7µm pore size, hexagonal and square pore ~2 bias and 8 bias, resistances ~GΩ, to <100MΩ possible Working on processing techniques to improve uniformity Techniques for gain & QE enhancement under investigation 8cm Si MCP on 100mm substrate 8
Silicon MCP Performance Characteristics Many Si MCP s of 25mm diameter with ~7µm pores have been tested The performance is improving as production is being refined. Gain is similar to glass MCP s. Open area ratio is up to >75% for hexagonal pores Gain decrease during scrub is smaller and faster than glass MCP s Gain 1.00E+05 1.00E+04 1.00E+03 1.00E+02 1.00E+01 Si 208 CVD 105-1 CVD107-5 CVD106-1 CVD120-2 CVD114-1-c CVD122-2 CVD122-4 CVD122-4-BEO CVD134-1 CVD134-4 CVD139-2 CVD139-4 ITT Glass MCP CVD169-3 Glass MCP 1.00E+00 0 500 1000 1500 2000 Voltage (V) Gain evolution of single Si MCP s MCP gain as a function of extracted charge, for one Si MCP. 9
Silicon MCP Performance Characteristics Gain and pulse height very similar to glass MCP s, stacks of Si MCP s (4) with gain up to 10 6 Quantum detection efficiency is similar to good bare glass MCP s (COS, EUVE, 12/10/6µm) The background rate is lower (0.02 events cm -2 sec -1 ) than normal or low background glass Gain and response uniformity are reasonably good. No hex modulation! 0.2 0.15 12/10µm COS Si MCP Bare glass Si Hex MCP 6µm pore MCP 0.1 0.05 0 200 400 600 800 1000 1200 1400 Wavelength (Å) QDE for Si & bare glass MCP s vs Wavelength Contrast enhanced image of the fixed pattern response to a Hg vapor lamp with a stack of 4 Si MCP s. ~14mm area, 10 7 counts, ~50µm resolution XDL. 10
Cross strip anode readout Cross strip is a multi-layer layer cross finger layout. Fingers have ~0.5mm period on ceramic. Charge spread over 3-53 5 strips per axis, Event position is derived from charge centroid. Can encode multiple simultaneous events. Fast event propagation (few ns). Bottom fingers 32mm x 32mm XS anode, 0.5mm period Anodes up to 32 x 32mm have been made Signals are routed to anode backside by hermetic vias Packaging can be compact with amp on anode backside Overall processing speed should support >> MHz rates Compact and robust (900 C). 11
Cross Strip Anode Electronics Chain Basic encoding sequence X Fingers Preamp Small, low power ASIC encoding with sparsification reduces data throughput requirements 16 16 16 16 Shaper 50 Ohm Driver Discriminator ADC 16 DIGITAL INTERFACE Cross strip anode position encoding electronics test-bed system. All signals amplified and digitized. Can choose up to 12 bits per signal. Anode backside showing connectors for the external board where preamplifier chips are mounted. Currently amplifiers have ~600e- rms noise. 12
Cross Strip Anode Readout Outstanding Spatial Resolution/Linearity ~7µm pores are resolved, <3 µm electronic resolution with 10 bit encoding electronics Image linearity is ~1µm level and shows pore misalignments and multi-fiber boundaries Gain required is <4 x 10 5, allows higher local event rates than normal readouts Lower gain means longer overall MCP lifetime due to reduced scrubbing Similar image of 7µm pore MCP pair at 2 x 10 6 Gain. Flood image of 12µm pore MCP pair at 4 x 10 6 Gain, 1mm square area. 13
Resolution of Cross Strip MCP Sensors Gain 1.3 x 10 6 4 x 10 5 1.1 x 10 6 1.8 x 10 5 7µm pore MCP pair 10µm pore MCP pair 14
Advanced MCP Sensors for Astrophysics Developing Detector Prospects High QE cathodes (Diamond, GaN, GaAs) ) with ~50%QE covering 20nm - 850nm Si MCP s with low fixed pattern noise, <5µm pores, background 0.02 events cm -2 sec -1 Cross-strip strip readouts with <5µm resolution, >50mm formats, >10k x 10k resels resels Packaging is shrinking while spatial resolution & QE are increasing ing dramatically. GALEX 65mmsealed tube XDL detector COS 2 x 85mm XDL detector 15