1. Electronic Devices in Optical and Infrared Astronomy

2. Basic Principles: Absorption of Light in Silicon Band Gap The electrons of a single isolated atom occupy atomic orbits, which form a discrete set of energy levels. When individual atoms come close together to form a crystal, electrons in the outermost orbits (or upper energy levels, 원자가 ) of nearby atoms interact to bind the atoms together. When a large number of atoms are brought together to form a solid, the number of orbits becomes exceedingly large, and the difference in energy between them becomes very small, so the levels may be considered to form continuous bands of energy. However, some intervals of energy contain no orbits, no matter how many atoms are aggregated, forming band gaps ( 띠간격 ).

3. Si atomic nucleus + 4 e valence electron

4. Figure below shows a simplified picture of the bands in a solid that allows the three major types of materials to be identified: metals, semiconductors and insulators. In metals ( 금속 ), the valence and conduction bands overlap and so any of the many valence electrons are free to roam throughout the solid, and to move in response to the force of an electric field.

5. An insulating material ( 부도체 ) has a highly ordered structure and a very wide forbidden energy gap. The conduction band is totally empty of electrons and so cannot contribute to an electrical current flow.

6. In a semiconductor ( 반도체 ), a few electrons can be elevated from the valence band to the conduction band across the forbidden gap only by absorbing heat energy from the random, microscopic motion of the crystal structure at room temperature. The electrons promoted to the conduction band ( 전도띠 ) can conduct electricity. In other word, they are free to move under the influence of an electric force field.

7. Table below is a section of the periodic table ( 주기율표 ) of the elements. The primary semiconductors like Si and Ge belong to the "4th column" elements. Compounds of elements on either side of the 4th column can be firmed and these alloys ( 합금 ) will also have semiconductor properties (e.g. indium antimonide, InSb and mercury- cadmium-telluride, HgCdTe).

8. When a photon is absorbed in the crystalline structure of silicon, it energy is transferred to "photoelectron", which is displaced from its normal location in the valence band into the conduction band. When the electron reaches the conduction band, it can migrate through the crystal. Migration can be controlled by applying a voltage by means of small metal plates called gates (or electrodes, 전극 ).

9. Absorption of photons in silicon is a function of the photon energy (and hence wavelength). The photon flux at depth z in the material is given by where  is the absorption coefficient. At a temperature of 300K,  ~5  m -1 at 400nm in the blue, but only  ~0.1  m -1 at 800nm in the far red.

10. At 77K,  reduced to 4.0, 0.25, and 0.005  m -1 at 400, 600 and 800 nm respectively. Clearly, red (low-energy) photons pass deeper into the silicon before being absorbed. Eventually, for the reddest light, there is simply not enough energy to elevate a valence electron to the conduction band. CCD 42-80

11. NameT(K)E G (eV) c (  m) Silicon(Si)295 1.12 1.11 Germanium(Ge)2950.671.85 Indium antimonite(InSb)2950.422.95 77 0.23 5.4 Lead sulphide(PbS) Mercury cadmium telluride(HgCdTe) 77 0.5 2.5 In other words, for each semiconductor there is a wavelength of light beyond which the material is insensitive to light because the photons are not energetic enough to overcome the forbidden energy gap (E G below) in the crystal. Band gap energy (E G ) and long-wavelength photo-absorption limits for some common semiconductors ( c ) are shown below.

12. When silicon atoms in the crystal structure are replaced with other atoms, the semiconductor is said to be doped. If the impurity atom has more valence electrons than the semiconductor, then it will donate these negative charges to the conduction band (n-type). Conversely, if the impurity atom has fewer valence electrons than the semiconductor, then the a positively charged hole is left in the valence band (p-type). The p-type semiconductor is ready to accept any available electrons.

13. positive semiconductornegative semiconductor B: Boron As: Arsenic or P: Phosphorus Note that these are electrically neutral.

14. PN Junction regions are used many times in semiconductor structures. When a PN junction is formed, electrons in the n region tend to diffuse into the p region at the junction, and fill up some of hole. Similarly, the diffusion of holes from the p to the n side leads to an more positive electrical potential. A pn junction in thermal equilibrium with zero bias voltage applied

15. Following transfer, the injected electrons come into contact with holes on the p-side and are eliminated by recombination. Likewise for the injected holes on the N-side. As the result, electrons and holes are gone, leaving behind the charged ions adjacent to the interface (pn junction) in a region with no mobile carriers (depletion region). A pn junction in thermal equilibrium with zero bias voltage applied depletion region

16. When equilibrium is reached, the charge density is approximated by the displayed step function. In fact, the region is completely depleted of majority carriers (leaving a charge density equal to the net doping level), and the edge between the depletion region and the neutral region is quite sharp (see figure for Q). The depletion region has the same charge on both sides of the pn interfaces, thus it extends farther on the less doped side (the n side in figures A and B). Integrating the electric field across the depletion region determines what is called the built-in voltage.  V  Built-in voltage

17. Forward bias (P positive with respect to N) narrows the depletion region and decreases the barrier to carrier injection. The diffusion component of the current greatly increases and the drift component decreases. In this case the current is rightward in the figure of the pn junction. The carrier density is large (it varies exponentially with the applied bias voltage), making the junction conductive and allowing a large forward current. e.g. light-emittion diode (LED) + -

18. Under reverse bias (or photovoltaic mode, P negative with respect to N), the potential drop (i.e., voltage) across the depletion region increases. This enlarges the depletion region, which increases the drift component of current and decreases the diffusion component. In this case the net current is leftward in the figure of the pn junction. The carrier density then is small and only a very small reverse saturation current flows.

19. Photodiodes ( 광다이오드 ) are similar to regular semiconductor diodes except that they may be either exposed or packaged with a window or optical fiber connection to allow light to reach the sensitive part of the device. When a photon of sufficient energy strikes the diode, it excites an electron, thereby creating a mobile electron and a positively charged electron hole. If the absorption occurs in the junction's depletion region, these carriers are swept from the junction by the built-in field of the depletion region. Thus holes move toward the anode, and electrons toward the cathode, and a photocurrent is produced.

20. When used in zero bias (or photoconductive mode), a voltage builds up. The diode becomes forward biased and a current begins to flow across the junction in the direction opposite to the photocurrent. This mode is is the basis for solar cells ( 태양전지 ). In the mode that the diode is reverse biased, dramatically reducing the response time at the expense of increased noise. This increases the width of the depletion layer, which decreases the junction's capacitance resulting in faster response times. The reverse bias induces only a small amount of current (known as saturation or back current) along its direction while the photocurrent remains virtually the same. light

21. MOS When we make an imaging device, which composed of numerous individual pixels, we want to store the photogenerated electrons into the pixels. To do it, the pixels requires a electrostatic field to attract the charged electron. Thus, we need to create a storage region (capacitor, 콘덴서 ) capable of holding many charges. This can be done by applying metal electrodes ( 전극 ) to the semiconductor silicon together with a thin (0.1  m) separation layer made from silicon dioxide (insulator). This structure behaves like a parallel plate capacitor (MOS, metal-oxide-semiconductor). Charge storage, charge-coupling, clocking Semiconductor (Si) Insulator (SiO 2 ) Metal gate +V

22. If the material is p-type, then a positive voltage on the gate will drive away the holes and sweep out a region depleted of charge. This is similar to pn-junction! When a photon is absorbed, it produces an electron-hole pair. The hole is driven out of the depletion region, whereas the electron is attracted towards the electrode. Insulator (SiO 2 ) Metal gate +V    Depletion region

23. The MOS capacitor is the combination of two parallel plate capacitors, the oxide capacitor and the silicon depletion region capacitor. The capacitance is given by where d denotes the separation between two plate capacitors, A the area of the plate,  the dielectric constant of the insulator (SiO 2 : ~4.5) Semiconductor (Si) Insulator (SiO 2 ) Metal gate V d A  0

24. Charge-coupling CCD is a bit like measuring the rainfall over a plantation. Suppose you distribute a many buckets in a rectangular pattern (CCD pixels) over the field. After it has stopped raining, we measure the amount of water in each bucket by shifting the entire array of buckets towards a conveyor belt (serial output register) located at one end of the field. bucket brigade ( 물통 릴레이 ?)

25. rain=photon output empty output empty Example of one-dimensional array (buckets in the rain)

26. General layout of a CCD showing numerous square pixels laid out in a grid

27. Schematic view of a three-phase CCD Semiconductor silicon is covered with a thin electrical insulating layer of silicon oxide on top of which are placed three sets of metal electrode stripes.

28. 1 2 Schematic view of a three-phase CCD

29. One of the three strips is set to a more positive voltage than the other two, and so it is under this one that the depletion region forms, and the electrons accumulate. We have two walls of the well, which obstructs any movement of charge along the length of the electrode (channel stops). The unique feature of the CCD is the method in which the photo- generated charge is extracted from the storage (or detection) site. Is is called "charge-coupling".

30. To transfer charge from under one electrode to the area below an adjacent electrode, we raise the voltage on the adjacent electrode to the same value as the first one (2). When the voltage on the original electrode is reduced to 0V, the transfer is complete because the collapse of the storage well pushes any remaining charges across to the new electrode (3). In the case of three-phase CCD, Since it takes three electrodes to define one pixel, three of the above transfers are required to move two-dimensional charge pattern by one pixel step. The process of raising and lowering the voltage can be repeated over and over (clocking). These clock pulses can be described in a diagram (timing waveform). (1) (2) (3)

31. 6.3 CCD Constructions 6.3.1 Interline and frame-transfer CCDs

32. Interline CCD have a photosensitive (green) and a masked (red) storage array, but they are interlaced, so that each storage row is adjacent to its photosensitive counterpart. Photosensitive and storage rows are alternated. This means that one shift is required in order to store the electrons from the photosensitive into the storage light-shielded array, instead of a number of shifts equal to the number of the rows. This architecture is useful when we read the electrons at very high speed (e.g. TV frame rates). Interline is widely applied for video cameras but less commonly for the astronomical use due to the narrow sensitive areas.

33. Frame-transfer: The array in frame-transfer CCD is duplicated. One part (photosensitive or image array) collect incoming photons, while the second part acts just as a temporary storage area (storage array). The storage array is shielded from light, so that no electrons are generated by incoming photons. The timing is the following. At the end of the exposure, all the electrons in the image array are transferred to the storage array. Only when all this shifting is over the reading phase begins. This architecture is applied in X-ray satellite and optical imaging camera on spacecrafts.

34. Full Frame CCD is made of an array of photosensitive elements where electrons are created by incoming photons from the time we press the shutter release button throughout the exposure time. Then a shifting phase occurs, shifting these electrons one row at a time to a sensing circuit producing a voltage proportional to the number of electrons. Blurring of the optical scene during readout is not a problem when we install a mechanical shutter. This architecture is widely applied for the astronomical observations because of the large photosensitive area.

35. The transfer direction is terminated by a special row of CCD pixels (output register). The electrodes are arranged at right angles to the main imaging area of the CCD so that it can transfer charge horizontally rather than vertically. Since the output register is a single row, it is called the serial register, while the main area of the CCD the parallel register. At the end of the output register is a single output amplifier. 6.3.2 CCD outputs

36. In general, the pixels in the output register are larger than those in the imaging area to ensure that they have more storage capacity and are much less likely to saturation.

37. The output amplifier of the CCD is shown schematically below. A packet if electrons with charge Q is allowed through the output gate onto an effective storage capacitance C which causes an instantaneous charge V=Q/C in the voltage of the input line of the on-chip transistor which in turn yields a voltage change at the output line. CCD output circuit

38. CCD output circuit and the clocking Signal output V RD

39. So far we described the front-illuminated CCDs in which the illumination flux was compelled to pass through the electrode. These gates are made of very thin polysilicon, which is transparent at long wavelengths, but becomes opaque at wavelengths shorter than 400 nm. Thus, at short wavelengths, gate structure attenuates incoming light. It is possible, using acid-etching techniques, to uniformly thin a CCD to a thickness of approximately 10 μm and focus an image on the backside of the CCD register where there is no gate structure (back-illuminated CCD). Backside-illuminated CCDs

40. One disadvantage of the thinned CCDs is that they are more mechanically fragile. Some thinned CCDs are mounted to a supporting substrate. Interference fringing can occur due to multiple reflections internal to the CCD substrate or between the silicon and supporting substrate.

41. Nowadays, digital still cameras use either a CCD image sensor or a CMOS sensor. These imaging sensors accomplish the same task of capturing light and converting the photon energy into electrical signals. A CMOS chip is a type of active pixel sensor made using the CMOS semiconductor process. Unlike a CCD, an extra circuitry next to each photo sensor converts the light energy to a voltage. Additional circuitry on the chip may be included to convert the voltage to digital data. Potentially, CMOS use less power and/or provide faster readout than CCDs. Usually, CMOS can be developed with low cost. The CMOS architecture is now widely applied for cameras on cell-phones. CMOS image sensor

42. A example of CMOS image sensor

43. Example of IR detector Teledyne (formally Rockwell) 's Hawaii HgCdTe FPA offers IR sensor 1024x1024 format with four outputs, <10 e- read noise, up to 1 MHz data rate per output, <0.1 e - /sec dark current at a convenient 78K, and a simple user interface. The HAWAII 1024 x 1024 readout is structured in four independent quadrants having four outputs, as shown in the right figure. I  IV  II  III (NIC)

45. Each quadrant contains two digital shift registers for addressing pixels in the array; a horizontal register and a vertical register. Each register requires two clocks; one dual edge triggered clock and one level triggered clock. To obtain a raster scan output, the horizontal register is usually clocked in the fast direction with the vertical register being clocked in the slow direction.

48. example timing pattern LSYNC should occur sometime before the first pixel clock edge. LINE cause vertical register to increment at both edges; odd rows on positive edge and even rows on negative edge. RESETB clock is an active low asynchronous reset which resets the entire currently selected row. READ is an active high clock. When active, pixel outputs are allowed to pass to the column bus for subsequent off chip. Line Clock must be low when fsync is pulsed low. Pixel Clock must be low when lsync is pulsed low. Horizontan / Fast Clock (shifted by Pixel/Lsync) Vertical / Slow Clock (shifted by Line/Fsync)

49. Horizontal Register Pixel and Lsync are the two required clocks for the horizontal register. The Pixel input clock is a dual edge triggered clock which will increment the selected column on both edges (odd columns selected on positive edges and even columns selected on negative edges). The Lsync clock is an active low input which will set a '0' in the first latch and a '1' in the remaining latches of the shift register, thereby initializing the shift register to select the first column in the quadrant. Vertical Register Line and Fsync are the two required clocks for the vertical register. The Line input clock is a dual edge triggered clock which will increment the row selected on both edges (odd rows selected on positive edges and even rows on negative edges). The Fsync clock is an active low input clock which will set a '0' in the first latch and a '1' in the remaining latches of the shift register, thereby initializing the shift register to select the first row in the quadrant.

50. Integration time voltage time V1V1 V2V2 X Single sampling:V 2 O Correlated double sampling:V 1 -V 2 O Multiple sampling: V 1 xN-V 2 xN - Ramp sampling reset Readout sequence for HAWAII array

51. 6.4 Astronomical Requirements For astronomy applications, CCDs cannot be used like normal TV cameras taking frames every 1/30th of a second. Instead, the CCD must be used as if it were a photographic emulsion in a camera. We need to take long exposure to build up a charge image from a faint source. When the charge image is removed during the readout process, we cannot do that rapidly either because the charge transfer efficiency will be impaired and the electronic noise is greater at higher readout frequencies. Astronomical CCDs are usually read out very slowly and hence this mode is called slow scan. (e.g. 50kHz for 1Kx1K  21sec) 6.4.1 Slow-scanning, cooling and optimization

52. EEV CCD42-80, provided by the manufacture

53. With such long exposures, however, the second problem arises- dark current. To permit long exposures, astronomical CCDs must be cooled to temperatures well below the freezing point of water, which implies using a vacuum chamber to avoid frosting. The availability of MPP CCDs has helped considerably in reducing the cooling requirements and many applications can be met with simple thermoelectric coolers (e.g. Peltier device), which is great for amateur astronomers! For the most strict applications in astronomical spectrographs, more cooling is required and most CCD cameras at professional observations use modified liquid nitrogen cooling systems. Finally, the performance of a CCD depends on how it is operated, and there is a certain amount of optimization that is required in terms of finding the very best clock voltages, bias levels and currents.

54. Summary Light can be absorbed and converted to electrical charge in a semiconductor using the photoelectric effect. The basic structure of a CCD is a two-dimensional grid of metal-oxide-semiconductor (MOS) pixels controlled by overlaying metallic electrodes arranged in strips. Charges accumulate where the light falls on the CCD and builds up am image. The pattern of charge can be readout by clocking the electrode strips to cause the charge pattern to couple from one pixel to the next and move along the columns of the CCD to an output register. Most modern CCDs are of the buried-channel construction. CCDs can be front-illuminated, or thinned backside-illuminated. For astronomy applications, CCDs must be cooled and operated in slow-scan readout mode.