검출기 눈, 사진, Photoelectric device, Photomultipliers, Image intensifiers, Charged Coupled Device,

Charged Coupled Device 장점 : 1. 높은 양자효율 (60-80%) (사진 :2-10%, PMT: 20-40%) 2. 선형 반응 (사진=특성곡선) 3. digital 정보 (사진 =아날로그)  콤퓨터 영상처리 4. 이차원 : 측광 및 측성 (PMT =측광 only) 5. 넓은 다이나마익 범위 > 10등급 이상 (사진=5등급) 6. 넓은 파장 영역에 반응 (사진 =특수처리 spectral sensitization) 단점 : 1. 좁은 시야 (망원경시야 >> CCD에 의해 제한된 시야) 2. 엄청난 자료 (512*512 CCD = 262,144 pixels, 4K*4K = 4096*4096 = 16,777,216 pixels) 3. 각 pixel의 다른 반응 :  flat 보정 필수 4. readout 잡음  관측된 자료를 읽어 올 때 발생 5. RN 잡음을 포함한 잡음을 줄이기 위한 냉각 (-100도 C ) 필수

검출기들의 양자효율

CCD 측정 CCD counts : bias + dark current + signal related to the number of electrons liberated by photons in each pixel Only fraction of photons generate electrons ; QE Photons = number of electrons (bias & dark removed) /QE = gain * DN / QE CCD = integrating device is controlled by a mechanical shutter or electrically( changing voltages in CCD) READ NOISE : on-chip amps average noise per pixel = read noise : 5 -20 electrons per pixel per read out (same whether exp 0.1 or 3 hours) Bias frame :simply read out zero image = bias signal changing with changing CCD Temp, also with read noise Dark frame : Without any light, signal caused by thermal excitation of electrons in the CCD : very sensitive to Temp cooling

CCD 측정 –cont Flat frame : uniformly illuminating the CCD will NOT generate an equal signal in each pixel ; small scale (pixel to pixel) sensitivity variation : caused by slight differences in pixel sizes Large scale (over large fraction of chip) sensitivity variation : by small variations in the silicon thickness across the chip These can be up to 10% variations over the chip Data frames : Exp time for best S/N Signal =photons  “root N” noise ; no way to get rid of this noise : poisson noise Dark noise, read out noise, cosmic ray

CCD 관측 1. a number of bias frames - median combine to a single low noise bias frame 2. a number of dark frames – no light, finite integration time equal to the data frame integration) , combine dark frames into a single low-noise frame (after subtracting bias frames) 3. a number of flat frames in each filter – twilight sky , dome flat (bias, dark subtracted)  normalize the flat so that the average signal in each pixel is 1.00 ( simply by dividing by the average signal) 4. subtract low-noise bias frame and low-noise dark frame from object frame. Then, divide this by the normalized frame Reduction frame =[(row-object frame)-(low-noise bias frame)-(low-noise dark frame)]/(normalized flat frame)] Amateur : CCD 20-40 C below the ambient T (0 to -40 C)  dark : changes in the dark – source of problems  take darks before and after object, these averaged and subtracted from the data frames (here dark really “dark+bias”)  inefficient at using telescope time – much used to dark frames

Night sky emission lines produce fringes on a CCD image; also additive noise ; usually in near IR I, Z bands : observe the sparse field with the same exp of the targets