IV DETECTORS Lit.: C.R.Kitchin: Astrophysical Techniques, 2009 C.D.Mckay: CCD s in Astronomy, Ann.Rev. A.&A. 24, 1986 G.H.Rieke: Infrared Detector Arrays for Astronomy, Ann.Rev. A&A 45, 2007 up to 1837: HUMAN EYE + : relatively high Q.E. (~10%), very high dynamic range (10 10!), - : no integration capability (! ~ 0.1 s), limited spectral bandwidth (400-700 nm), rather low spatial resolution not suitable for quantitative measurements 1837 Daguerre: invention of photography first sky object photographs (daguerrotypes): 1840 Moon J.W. Draper 1843 Solar Spectrum 1845 Sun Foucault +Fizeau 1870 invention of dry gelatine emulsions this enabled astronomical applications! photography dominated astronomical detection for more than a century: ~1870-1980 pro s and con s of photographic plate: + : ability to integrate (up to many hours) large detector area (up to 50x50 cm) enormous storage capacity (10 6-10 8 pix /cm 2 ) high resolution (grains ~ 1-10 µm) wavelength coverage " ~ 300-1000 nm - : low Q.E. (~ 0.1 3%, with hypersensit.: <10%) non-linearity (Q.E. depends on intensity) reciprocity failure : non-linearity in exposure time limited dynamic range (< 10 4 ) emulsion properties depend on batch, processing Daguerrotype of the Moon, John W. Draper March 26, 1840 New York
IV.1 PHOTOELECTRIC DETECTORS photoelectric effect discovery: Hertz 1887 explanation: Einstein 1905 first photocells : ~1910 principle: solid absorbs photon with h# > h# 0 =E lim and emits electron with E el = h# - h# 0 #(photoelectrons) $ #(infalling photons) electron current through resistor % voltage device to measure photon flux electronically: photocell problem: currents are very small (e.g. 10-17 A)! S/N dominated by thermal resistance noise Johnson noise : RMS noise voltage V n $ (ktr) 1/2 astronomical application pioneered by Stebbins & Whitford 1910-1930! beginning of electronic revolution in astronomical detection schematic of a photomultiplier tube solution: 1940-45 development of photomultiplier tube! current amplification without resistance gain factor 4-5 per dynode stage! 10 6-10 7 amplification for 10 dynodes! photon counting possible QE 40% 20% photomultipliers were the prime detectors for astronomical precision (spectro)photometry during the period 1950-1990 10% 1% + : good Q.E. (up to 30%) wide "&range (~100-1000 nm) linear device high dynamic range (~ 10 6 ) suitable for photon counting photon-noise limited (when cooled) - : not usable for imaging! a wide variety of photocathode "-response profiles was produced by many different cathode materials: e.g.: Cs-Sb, Ag-O-Cs, Ag-Bi-O-Cs, Na-K-Cs-Sb, K-Cs-Sb, etc.
IMAGING PHOTO-ELECTRIC DETECTORS although photomultipliers as single pixel detectors are unsuitable for imaging, imaging devices based on photoelectric detection have been developed: 1 st generation image intensifiers: multiplication of photons (but: low gain) "-conversion via phosphor properties before ~1980: final detection by photographic plate Micro-Channel Plates (MCP s), or Multi-Anode Microchannel plate Arrays (MAMA s) Q.E. up to 20%, ~10 6 electrons/photon high spatial resolution (10 µm channels) many pixels (up to ~10 8 ) well-suited for EUV + soft X-rays! successfully used in Einstein, Rosat, Chandra IV.2 SEMICONDUCTOR DETECTOR ARRAYS creation of photo-electrons: photo-ionisation of a solid photo-conduction: photo-excitation in a solid 'E gap impurity band electron energy levels of atoms in a solid split and merge into energy bands with many sub-levels atoms are kept together by valence electrons in the outer shells; combined energy levels of those shells form the valence band excitation can move valence band electrons to higher energy levels in the same band, or (if 'E> 'E gap ) into higher conduction bands conduction of electrons can occur in valence or higher bands only if free E-levels are available case a) : E-levels in the valence band partly filled! electrons can move easily: solid is a conductor case b) : E-levels in valence band filled, 'E gap is large! electrons cannot jump to free levels: solid is insulator case c) : as in case b), but with small 'E gap ; excitation (thermal or photonic) can now move electrons into the conduction band: solid is a semiconductor impurities cause extra E-levels in-between the bands of pure material: impurity bands! electrons can be excited more easily into the conduction band simple picture of energy levels in solids this explains a number of facts: photons can excite electrons in conductors, but these get lost in the sea of thermally excited electrons! no good for detectors 'E gap in semiconductors can be changed by controlled doping with impurities pure Si (=semicond.) has 'E gap = 1.10 ev this corresponds to the observed "-cutoff of 1.11 µm in CCD s for Gallium-doped Si 'E gap drops to 0.07eV! "-cutoff = 17 µm smaller 'E gap requires deeper cooling to suppress thermal excitation
CHARGE-COUPLED DEVICES - CCD s 1969 invented by W.Boyle + G.Smith (Bell Labs) for application in computer memory ~1975 first applications as astronomical detectors summary of main properties: absorption of hv (>1.10 ev) in Si! excited electron + hole electrodes on the chip create + potential wells that collect the electrons, holes diffuse away into the Si lattice high QE of this process: for " = 500-800 nm ~80%! pixels are created by combination of: grid of pos. electrode strips on the Si and ( grid of channel stops inside the Si (thin barriers to stop e - transfer along electrodes) schematic structure of a 3-phase CCD the + electrodes are split into multiple electrode sets (n=3 for 3-phase CCD ); each set has its own voltage by clocking phased voltage differences after exposure, the accumulated pixel charges are transferred along the columns between channel stops into output registers similar charge transfer feeds the charges from the output register into the output amplifier charge transfer efficiency: 99.9999 %! chargetransfer in a 3-phase CCD more facts about CCD s cooling is required to reduce thermal dark current but: T) * E gap + (Si: 1.10 %1.15 ev for T 300 %100K) this leads to reduced red response I dark $ e -A/kT! modest cooling is OK (typically ~200K) monochromatic light (especially red) will cause interference in the Si layer (Fabry-Pérot fringing) this can be reduced by AR-coating on Si surface broad-band imaging: fringing not a serious problem (exception: bright sky lines) spectroscopy: fringing is major flat-fielding problem! NB: FP-fringing occurs in all 2D detector arrays! " saturated pixels (25 µm pix: ~0.5x10 6 e - /pix) cause charge leakage to nearby pixels ( blooming ) and non-linear response the lower sensitivity level is set by CCD readout noise (~ few e - /pix/readout)! together: dynamic range ~ 10 5 FIELD (IFU image slices) detector fringing in JWST-MIRI spectrometer cosmic rays produce high spikes; this limits max. t exp! long t exp requires shorter sub-exposures the CCD readout chain contains several analog amplification stages with gain factors,1 and an ADC! ADC output number (ADU), (# detected photons) this ADU conversion factor needs to be calibrated electrodes cover significant part of CCD front side! back-illuminated CCD s have higher QE blooming of saturated star image largest CCD s: 16.8 Mp (e.g. Kodak-16801: 4096 2 9 µm pix)
IV.3 INFRARED DETECTORS most modern IR detector arrays are based on photoconductivity in semiconductors many different materials are used from semiconductor bandgap properties we can understand: Near-IR % pure materials can be used Mid/Far-IR % doped materials are needed the "-cutoffs in the table are reflected in the detectors of the Spitzer instruments: BIB (blocked impurity band) detectors use heavily doped Si which would normally lead to high dark current; to suppress this there is an extra blocking layer of pure Si present maximum sizes of IR arrays: NIR: 2048 2 ; short-wave MIR : 1024 2 ; long-wave MIR : 256 2 unlike CCD s, IR-arrays don t use charge transfer, but are read pixel-by-pixel by means of complex readout circuits and MOSFET amplifiers in an underlying multiplexer ( MUX )! allows non-destructive/multiple readouts bolometers are sensitive thermometers based on change in electrical resistance after conversion of absorbed radiation into heat in the NIR/MIR they have been replaced by photoconductor arrays, but semiconductor bolometers are still used in the far IR (" > 100 µm) overview of photoconductor detector arrays for the NIR/MIR (G.H.Rieke 2007): - - IBC = BIB