Astro890L5 - Detectors Wide, High, Deep, and Sensitive...

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: Detectors Wide, High, Deep, and Sensitive April 23, 2008 Astro 890 The Plan Wax poetic (and silly) about the landscape Get real and focus on stuff I actually know something about Photon Detection Ways of organizing large numbers of Pixels Processing signal ADCs System Examples of real detectors April 23, 2008 Astro 890 Paul said talk about Detectors Radar Detector Radon Detector Smoke Detector Lie Detector Leak Detector Snake Detector April 23, 2008 Astro 890 Even if I stop being silly... Neutrino Detector? Cosmic Ray Detector? Photon Detector? Yeah, Photon, that's the one April 23, 2008 Astro 890 Still a Tall Order DC to Daylight doesn't cut it (daylight is only 1015 Hz) ~108Hz to 2x1028Hz 4 x10-7to 1014 eV (~100 ergs) April 23, 2008 Astro 890 Stick to what you know? We "do" ground based optical/NIR instruments Ground based means to atmospheric cutoff at ~300 nm (3000 ) caused by O3 in upper atmosphere. NIR, for us, means to 2.5 or J H and K bands. At any longer wavelength the background fills the wells so fast additional hardware is required for real time co-adding. April 23, 2008 Astro 890 First Astronomical CCD Picture Fairchild 100x100 CCD, 1974 April 23, 2008 Astro 890 Photon Detection (0.3 to 2.5) Electron into vacuum (photocathode). Silver halide to silver or other photo chemistry (Call Kodak) Create hole electron pair(s) April 23, 2008 Astro 890 Silicon Works Great up to ~1m 1=1.24 eV Si band gap =1.1 eV Beyond that ... InSb out to ~5 HgCdTe which can be tuned to 2.5 , 5 or even 12 April 23, 2008 Astro 890 Avoid Solid State Physics at all Cost April 23, 2008 Astro 890 Care for (half of) the Pair Don't let it spend any time in an area with a lot of holes (electrons) Move a packet of them around without loss April 23, 2008 Astro 890 QE (DQE) Get Photon into detector Fresnel losses, Si has index of 3.4! R~30% Anti-reflection coatings can help Semi-transparent electrodes. Losses ~50% This is the main reason for thinning. Recombines near surface P channel has field in the right direction Not letting it get out of detector Si becomes transparent much beyond 1 April 23, 2008 Astro 890 Absorption length in Si April 23, 2008 Astro 890 Moving charge Two ways, really the same Make a moving potential well (CCD) Make an array of switches (Most IR detectors) April 23, 2008 Astro 890 CCD wells Buried "channel stops" to define columns Array of insulated electrodes to define pixels Additional array of electrodes to shift a line at a time April 23, 2008 Astro 890 CCDs 3 Phase Serial Register 1 2 3 1 2 3 1 2 3 1 2 3 Output April 23, 2008 Astro 890 Three 2D Archtectures CCD with Parallel and Serial shifts CCD with only Parallels and an output per column X-Y array of switches April 23, 2008 Astro 890 CCD Analogy RAIN (PHOTONS) VERTICAL CONVEYOR BELTS (CCD COLUMNS) BUCKETS (PIXELS) HORIZONTAL CONVEYOR BELT (SERIAL REGISTER) April 23, 2008 Astro 890 MEASURING CYLINDER (OUTPUT AMPLIFIER) Exposure finished, buckets now contain samples of rain. April 23, 2008 Astro 890 Conveyor belt starts turning and transfers buckets. Rain collected on the vertical conveyor is tipped into buckets on the horizontal conveyor. April 23, 2008 Astro 890 Vertical conveyor stops. Horizontal conveyor starts up and tips each bucket in turn into the measuring cylinder . April 23, 2008 Astro 890 After each bucket has been measured, the measuring cylinder is emptied , ready for the next bucket load. ` April 23, 2008 Astro 890 April 23, 2008 Astro 890 April 23, 2008 Astro 890 April 23, 2008 Astro 890 April 23, 2008 Astro 890 April 23, 2008 Astro 890 April 23, 2008 Astro 890 A new set of empty buckets is set up on the horizontal conveyor and the process is repeated. April 23, 2008 Astro 890 April 23, 2008 Astro 890 April 23, 2008 Astro 890 April 23, 2008 Astro 890 April 23, 2008 Astro 890 April 23, 2008 Astro 890 April 23, 2008 Astro 890 April 23, 2008 Astro 890 April 23, 2008 Astro 890 April 23, 2008 Astro 890 April 23, 2008 Astro 890 April 23, 2008 Astro 890 April 23, 2008 Astro 890 April 23, 2008 Astro 890 April 23, 2008 Astro 890 April 23, 2008 Astro 890 April 23, 2008 Astro 890 Eventually all the buckets have been measured, the CCD has been read out. April 23, 2008 Astro 890 Q=CV V=Q/C High responsivity means very small C 1cm x 1cm x 1mm ~ 1pF (10^-12 F) Good CCDs now have node capacity of ~10 fF (10^-14F) 10 V/electron April 23, 2008 Astro 890 Turning Charge Packet into a Voltage Vref Reset Switch Buffer Vout "node" capacitance, C Electron Packet April 23, 2008 Astro 890 IR array Architecture Ref. Voltage Row Switches Column Switches Output Amp. April 23, 2008 Astro 890 Signal Processing 1 "kTC" noise Any time you reset a capacitor the resulting voltage is uncertain by kTC. Good CCD will have C~10-14F, T~150K, noise ~30 electrons RMS Must use "dual slope integrator" or "double correlated sampling" (DCS) to reject kTC noise. DCS also rejects "noise" at any frequency less than 1/ (integration time) April 23, 2008 Astro 890 Dual slope Integrator Output Voltage Vref Vref-nq/C Sampling Function +1 0 -1 Integrator Output } Vo time April 23, 2008 Astro 890 Signal Processing 2 Optimal S/N for White Noise is Simple differential averager (aka dual slope integrator) As with all white noise problems S/N grows as t where t is the integration time Noise from amplifiers is always 1/f like If 1/f knee is below 1/t then noise is essentially White If 1/f knee is well above 1/t and noise is really 1/f then S/N is independent of t. April 23, 2008 Astro 890 Dark Current (silicon) Room Temperature ~1/40 eV Band gap ~1.2 eV Leads to a dark current of ~20pA/cm2 Or about 300 electrons/second for 15 x 15 pixel. Must cool, typically 190K for 1 electron/hour, although surface effects can cause problems. 2.5 HcCdTe requires 77K or lower April 23, 2008 Astro 890 The next step Systems have typically had 16 bit ADC Not enough resolution for low noise CCDs with a full well of 200,000 electrons 18 bit ADCs now work (and are a lot cheaper than old style 16 bit units) April 23, 2008 Astro 890 Telescopes are harsh Fiber optic coupling between telescope focus and computers helps with noise problems Old=120 Mbits/sec, New=2 Gbits/sec April 23, 2008 Astro 890 System (Warm Part) Computer "Sequencer" Generates clock patterns Receives data from the cold Fiber interface April 23, 2008 Astro 890 System (Cold Part) Head Electronics Drive electronics for detector Signal processing and analog to digital conversion Fiber interface April 23, 2008 Astro 890 Real World Detectors 800 x 800 Texas Instruments, Original HST detector 3 x 8 k e2v, Worlds Largest IR array in Barbie Coffin April 23, 2008 Astro 890 ...
View Full Document

This note was uploaded on 07/17/2008 for the course ASTRO 890 taught by Professor Martini during the Spring '08 term at Ohio State.

Ask a homework question - tutors are online