sec 2 - principle and operation of thermal imaging systems

Maj Harkirat Singh, TSOC−28 (Project Study)Maj Harkirat Singh, TSOC−28 (Project Study)

THERMAL IMAGING TECHNOLOGIES

Introduction

1. Any observation or target engagement process involves the following participants :-

(a) The target and its signatures

(b) The intervening medium

(c) The receiver device

(d) The operator

(e) The weapon

2. The detection, recognition and identification of the target and its subsequent engagement is primarily dependent upon visual means and mechanisms. Visual acquisition process occurs in the visible and Infra-red portions of the electromagnetic spectrum. The technologies involved in infrared, or thermal, imaging are deliberated upon in this section.

The Spectrum

3. The human eye responds to visible light in the colour band from red to violet. Like any other electro-magnetic wave, visible light has a frequency and wavelength and it is differences in frequency that cause the different colours. The wavelengths of visible light range from 0.4 µm (violet) to 0.75 µm (Red). The infra-red radiation has longer wavelengths in the range from 0.75 µm to 1000 µm but current thermal imagers normally operate in a selected part of the range from 1 µm to 15 µm. One military application of UV imaging is the detection of launches of large missiles, e.g. ICBMs. The very hot efflux from such missiles emits large amounts of UV radiation.

Maj Harkirat Singh, TSOC−28 (Project Study)

2

Figure 1

4. Thermal imaging (TI) is a technique which converts the small temperature differences existing in natural scenes to a visual picture similar to that produced by television, with the important difference that no source of illumination is required. Thus thermal imaging can be equally effective by day or by night.

5. The possibility of detecting targets by their IR emissions was pursued both in the UK and Germany before the Second World War In 1938 radar and IR technologies were in their infancy and it was by no means obvious which offered the better technical solution. Radar was backed more strongly in the UK whereas in Germany IR was favoured by the Navy. By the end of the war German IR technology was quite well advanced. Rapid progress occurred after the war in the development of detectors suitable for detecting emissions from the exhaust plumes and hot surfaces of jet aircraft. By the late sixties emissions from objects at normal temperatures could be detected, provided the detector was cooled to a low temperature. The production of appropriate low temperatures was solved by development of the Joule-Thomson minicooler which then made thermal imaging a technical possibility.

6. The principle and operation of thermal imaging devices will be covered in the following parts :-

(a) Radiation theory and atmospheric transmission of thermal radiation,

(b) Design aspects of a generic thermal imaging system.


3

RADIATION THEORY AND ATMOSPHERIC PROPAGATION

Emission of Thermal Radiation

7. All objects, when raised to a temperature greater than absolute zero, emit a continuous spectrum of electromagnetic radiation. This is caused by the continual absorption and radiation of energy by atoms of the body. The figure below illustrates the process of absorption and emission, through which all atoms continuously pass, as energy arrives to be absorbed and then re-radiated, reflected,or transmitted on through the body. For 'everyday' hot objects, most of the radiation lies in the infrared (IR) region of the electromagnetic spectrum. At these frequencies wavelength is normally used to specify the emissions, as is illustrated by the spectrum diagram above.

8. As the temperature of an object increases, the wavelengths that it radiates get smaller while the amount of radiation increases. This is depicted in the set of graphs below. Objects with rough, black surfaces emit much more radiation than objects with smooth, shiny surfaces. Since thermal emissions are affected by the surface conditions of an object, these curves apply only to an ‘ideal radiator’ or black body. However, the radiation pattern of most real objects is broadly similar to that of the ideal radiator, although the amount of radiation is generally less and the curve is not so smooth. Real objects have lower emissivity than ideal radiators. An object with an emissivity of 0.5 would emit half as much thermal radiation as an ideal radiator. Two regions are of interest in the graph; the temperature range around 6 000 K – which is the temperature of the surface of the sun, and 300 K (27oC)– which is the temperature of everyday objects and their surroundings.


4

9. Every object whose temperature is not absolute zero emits what is loosely termed as thermal radiation. Thermal radiation at terrestrial temperatures consists primarily of self-emitted radiation from vibrational and rotational quantum energy level transitions in molecules, and secondarily from reflection of radiation from other heated sources. In most imaging applications, the mechanism is unimportant, the only important factor being the existence of apparent temperature differences. The fundamental element in thermal radiation is the Planck blackbody radiation theorem that postulates generation of thermal radiation by linear atomic oscillators in simple harmonic motion which emit radiation in discrete quanta whose energy E is a function of the radiation frequency v given by E=hv, where h is Planck’s constant.

Laws of Thermal Radiation

10. Three laws govern the power and spectral composition of energy radiated by a blackbody.

(a) Stefan-Boltzmann's Law. The Stefan-Boltzmann law states that the total power (energy per unit time) emitted by a blackbody, per unit surface area of the blackbody, varies as the fourth power of the temperature. This is given the symbol W. Thus, W= σT4 (watt/m2), where T is the absolute temperature in Kelvin and σ is the Stefan Boltzmann’s constant = 5.671x10-8

Wm-2K-4. It can be seen that as T is raised to the fourth power, relatively small increases in temperature have a great effect upon the power radiated. A


5

blackbody raised to 300K, which is roughly room temperature, radiates 0.46 kW/m2, whereas the same body raised to 6000K radiates 73.5 MW/m2. (Note: the average human body radiates about 100W, the same as a standard light bulb).

(b) Planck's Law of Spectral Distribution.

(i) This is a function that denotes the spectral composition of any energy radiated by a blackbody and is denoted by :-

λ λπλλ

= −

2

5 / 3

2 1( , )

1ch T

hc wattW T

e cm

where

c (speed of light in vacuum) = 2.9979X1010 [cm/sec]

h (Planck’s constant) = 6.6256x10-34 [watt sec2]

k (Boltzmann’s constant) = 1.338054 x 10-23 [watt sec/K]

σ (Stefan Boltzmann’s constant) = 5.6697 x 10-12 [watts/cm2 K4]

= π 5 4

2 3

215

kc h

(ii) The equation is clarified in a series of graphs illustrated below. The graphs show power radiated per square centimetre against wavelength for different temperatures of the black body. Each graph is drawn for a specific temperature. Thus the power radiated between two given wavelengths can be estimated by calculating the area under the graph between the wavelengths in question. Clearly, the total power radiated at all frequencies for a given temperature is the total area under the graph of that temperature.


6

(c) Wien's Displacement Law. The Wien’s displacement law states that the wavelength at which the blackbody emission spectrum is most intense varies inversely with the blackbody’s temperature. The constant of proportionality is Wien’s constant (2897 K mm):The spread of wavelengths in the graphs is theoretically infinite, however, for all practical purposes there is a limit at each temperature. Of greater interest is the fact that the peak radiating wavelength changes with temperature and the higher the temperature of the curve the shorter is the wavelength maximum. The equation which determines the actual peak is given by ;

λm x T = constant = 2897µmK. Where λ m is the wavelength in µm at which peak occurs and T is the absolute temperature.

(j) If the wavelength at which peak occurs is calculated for a temperature of 6000K, that of the Sun, the result is 0.5µm. It should come as no surprise that the eye is 'tuned' to accept radiation most readily at this wavelength, which is about in the middle of the light spectrum. On the other hand, if the same calculation is carried out at 300K, roughly room temperature, the result is 10µm, a wavelength well above the light spectrum. Thus, when a weapon system is designed to search for obviously hot bodies, such as a jet efflux, it will normally do so at the shorter wavelengths near the light spectrum, on the other hand, when personnel and equipment are involved, the wavelengths of interest are much longer.


7

(d) Lambert (Cosine) Law.

(i) When radiance of an element of a surface is the same in all direction within the hemisphere over this element then the following relationship is fulfilled

Ι(θ)=Ιncos θ

where I(θ) is the radiant intensity in direction of the angle θ �to the direction normal to the surface, Ιn is the radiant intensity in the direction normal to the surface.

(ii) An ideal surface that fulfils the Lambert cosine law is called the Lambertian surface. For the Lambertian surface there exists a following relationship between the radiant exitance M and the radiant radiance L

M= πL. The relationship is sometimes called the Lambert law


8

Thermal Contrast

11. All objects at ambient temperatures on the Earth will emit thermal radiation and might also reflect incident radiation from surrounding objects. The result is a thermal picture where objects are difficult to distinguish from each other. Under such conditions, the thermal contrast is fairly low.

12. When the same scene is viewed with visible light, where there is virtually no emitted radiation (because the temperatures are too low), then we see our surroundings by reflected light only. There is good contrast between different objects because they reflect different amounts (light-dark) of light and different wavelengths (red-blue) of light. Contrast is further improved because the human eye has evolved to be sensitive to small differences in brightness and colour. The human eye sees so well because it contains different types of receptors that are sensitive to red, green and blue light. This allows the eye to use three views of the brightness of a scene at different wavelengths (colours) and to compare them. Thermal cameras normally use a single type of sensor that just detects the amount of thermal radiation. There is no equivalent to the colour view that the human eye can exploit to see more detail in the scene. Thermal cameras see brightness information (hot and cold) only; human eyes see both brightness (light and dark) with additional detail in colour.

13. In detection systems the most important factor of interest is the contrast between target and background. This is readily seen in black and white television and the viewing of other objects within the light spectrum. Thermal contrast is no exception to this and in order to detect a thermal target it will be necessary to distinguish it from its background.

14. Definition. The thermal contrast, C, between a target and its background is defined as, C = (Wt-Wb)/(Wt+Wb). Where Wt is the power radiated by the target and Wb is the power radiated by the background. When the temperatures of both target and background are approximately the same it can be shown that C=2×∆T/T, where T is the background temperature at the target and ∆T is the difference in temperature between the two. The proviso is that both target and background behave as blackbodies. For a situation where background temperature is 300K, ie, T=300, and the temperature difference is 1K, ie�� ∆t=1, then C=2/300, which is approximately 0.00666, that is 0.67%. For visible wavelengths scene contrasts are much higher, typically about 30% to 40%. Thermal imaging therefore suffers by comparison with daylight television from inherently low contrast scenes.

Radiation From Real Bodies

15. Thermal radiators are characterised by their radiation emission efficiencies into three categories; blackbodies, greybodies and selective radiators. This radiation efficiency property is described by the quantity spectral emissivity, ε(λ), which may


9

be considered as radiation emission efficiency at a given wavelength. Any object (solid, liquid or gas) in thermal equilibrium with its surroundings, which means maintained at a constant temperature, simultaneously radiates and absorbs energy at the same rate, in the form of a continuous spectrum of infrared radiation.

16. A theoretically perfect emitter and absorber of radiation is called a blackbody. It is considered to absorb all radiation independently of wavelength, shape, material or surface characteristics. The radiation emitted by such a body is known as blackbody radiation and depends only upon the absolute temperature of the body. It is also, by definition, the most efficient radiator possible. All other bodies will be less efficient radiators/absorbers depending upon their structure and surface characteristics. The blackbody is an idealisation; it emits and absorbs the maximum theoretically available amount of thermal radiation at a given temperature. A blackbody has ε=1 within a given wavelength range of interest, and a greybody has ε=constant<1 within a specified band. A selective radiator has 0≤ε(λ)≤1. Actual radiations are determined by a number of factors such as the radiating efficiency of a body, called its emissivity, the ability to absorb or reflect energy, and the temperature to which the body is raised.

17. The radiation emitted by a blackbody is wholly dependent on its absolute temperature and it follows that the thermal contrast between two blackbodies will depend upon their temperature difference only. With real bodies the situation is more complicated because both temperature and emissivity of the surface must be taken into account.

18. Temperature. Surface temperature is affected by a number of factors including:-

(a) Absorptivity of the surface.

(b) Spectral composition of incident radiation.

(c) Thermal capacity and thermal conductivity of the material.

(d) Prevailing meteorological conditions. (day or night, clear or cloudy, wet or dry, calm or windy etc.)

(e) Emissivity of the surface.

19. Emissivity.

(a) The blackbody laws can be applied to real bodies with the inclusion of an 'efficiency' factor which will account for the difference in behaviour between a blackbody and the real one. The efficiency factor for a body is known as its emissivity, E, and is defined as:-

E = rb

bb

Power radiated by a real body at temperature T (W [T])

Power radiated by a black body at temperature T (W [ ])T


10

(b) This has the effect of modifying the original blackbody equation for radiation from Wbb = σ⋅Τ4, to Wrb = E× σ⋅Τ4 for a real body. Since a blackbody is by definition the most efficient radiator possible, Ebb=1. and Erb<1. Where Ebb is the emissivity of a black body and Erb is the emissivity of a real body. It can be seen that no real body can acquire a radiating efficiency greater than that of a blackbody.

(c) If the values of E for a range of common materials were listed it would show that they depend very much upon three factors, the type of material (metal, semiconductor or insulator), surface finish and temperature. In general, for highly polished metal surfaces E is very low, about 0.5, but for the rough surfaces of semiconducting and insulating materials E may be close to 1, implying 50% and 100% of the efficiency of a blackbody radiator respectively.

(d) It is also possible for the emissivity of a body to vary with the wavelength being emitted. In practice two types of variation are found to be important, one called a greybody radiator, for which the emissivity remains constant over all the wavelengths transmitted, such as with the tailpipe of a jet aircraft, vehicle exhaust or human being. The other is called a selective radiator for which the emissivity may vary greatly particularly at certain wavelength, such as with hot water vapour and carbon dioxide. Both are illustrated graphically at Fig.4.


11

(e) The emissivity of a body does denote its radiating ability as the temperature must also be taken into account as it also affects the power and wavelengths radiated. In order to obtain an accurate picture of the radiations from a real body, its emissivity and the blackbody curves, must be taken together. This is the case in the figure above, where the selective radiator curve of Fig.4, and 1000K blackbody curve of Fig.3, have been added together and plotted. The same exercise has also been carried out for a greybody where it can be clearly seen that such a radiator has similar features to a blackbody but with a lower level of emissions, that level depending upon the body characteristics.

20. Emissivity and Thermal Contrast.

(a) The emissivity of a surface is influenced by the nature and state of the surface, particularly if it is covered with a film of water or moisture. With real bodies, thermal contrast is not only a matter of the difference between radiation emitted by a target and its background. The fact that the emissivity of a real body is less than 100% when compared to a blackbody means that the surface possesses a finite reflectivity and the radiation leaving the target may contain two components:-


12

(i) A direct component controlled by the temperature and emissivity of the surface.

(ii) A reflected component of the background and any neighbouring targets.

(b) Sometimes the reflected component dominates the direct component and vice-versa. Radiation leaving the background is also affected by the presence of a target or targets and as a result certain simplifying assumptions must be made. In general target signature radiated will depend upon the emissivities of, as well as the temperature difference between, target and background

21. Absorptivity.

(a) Energy arriving at a surface can be absorbed, transmitted or reflected. Transmission implies that the energy passes through the body, and in the vast majority of real targets this factor is zero. In other words the energy is either absorbed or reflected. Since the sum of reflected and absorbed energy must be equal to the incident energy, a good absorber is a poor reflector and vice-versa. It can be shown that the ability of a body to absorb is equal to its emissivity, which is consistent with the postulates on blackbody radiators.

(b) The emissivity of most surfaces falls abruptly for viewing angles (measured perpendicular to the surface) greater than 60o. The radiating character of the surface can change from being a good emitter at low viewing angles to a good reflector at large viewing angles. This is a somewhat similar effect to glint in the radar spectrum of a target. The figure below, illustrates the effect. It should be noted that surfaces which do not reflect at light wavelengths, such as roughened metal surfaces, could be good reflectors at infrared wavelengths, where the level of absorption may drop significantly.

.


13

Atmospheric Propagation

22. On a clear day, visible light passes through the Earth’s atmosphere with ease and we can see for many kilometres. When the air is not clear then visibility is poor - perhaps only a few hundred metres. The principal reason for poor visibility is not that the light is absorbed or blocked by the atmosphere - it is caused by the scattering of the light as it passes through the air. When light encounters particles in the air that are of a similar size to its wavelength then some of the light is scattered in a random direction by the particle. The visible effect when there are only a few particles in the air, is that of a grey haze as light from the bright parts of the scene is scattered into what should be the darker parts of the scene. When there are many particles in the air, as in mist, fog or smoke, then practically all the light that we see has been scattered and everything is a uniform grey.

23. Once the thermal signature of a target and its background has been generated, the signal they comprise must pass through the atmosphere and other media between the target and the detector. At the detector it is necessary to discriminate between target, background and noise, in order to present a usable signal to the display system. When infra-red radiation passes through the same particles then, because the infra-red radiation has a longer wavelength, there is much less scattering. Consequently, a thermal camera is much less affected by mist and battlefield smoke. If the particles were larger then they would scatter infra-red also, but such particles would fall to the ground quite quickly. Heavy rain, which consists of large particles, will scatter a significant amount of infra-red radiation and make it difficult to use a thermal camera over any great distance. The large quantity of (falling) water in the air also causes significant absorption of both visible and infra-red radiation.

24. Thermal radiation in its passage through the atmosphere suffers attenuation due to both absorption and scattering by the gas molecules, molecular clusters (aerosols), rain, snow, and suspensions such as smoke, fog, haze, and smog. The following molecules (in order of importance) absorb radiation in the thermal infrared in broad bands centred on the wavelengths indicated : water (4.8, 9.6, 6.3 µm), carbon dioxide (2.7, 4.3, 15 µm), ozone (4.8, 9.6, 14.2 µm ), nitrous oxide (4.7, 7.8 µm ), CO (4.8 µm ) and methane (3.2, 7.8 µm ). Of the two processes molecular absorption proves to be more dominant at infrared wavelengths, unlike scattering prevailing at visible wavelengths. Water vapour, ozone and carbon dioxide are the most significant absorbers. The 6.3 µm water band and the 2.7 and 15 µm carbon dioxide bands effectively limit atmospheric transmission in the 2 to 20 µm thermal spectral range to just two atmospheric windows – the 3.5 to 5 µm and 8 to 14 µm bands.

25. Atmospheric Constituents.


14

(a) The atmosphere contains, in addition to nitrogen, oxygen, carbon dioxide, water vapour and methane etc, clusters of molecules known as aerosols, a general term denoting suspended liquid and solid particles. Aerosols can be divided into two classes, firstly haze and dust particles with radii between 0.1 and 2.0µm, and secondly clouds of liquid and solid water with radii between 1.0 and 100µm.

(i) Haze and Dust Particles. These particles are ejected directly into the atmosphere or formed in the atmosphere from trace gases. They originate from soil and rock debri, forest fires, sea-salt nuclei, gaseous emissions forming various sulphates or ammonium salts and hydrocarbons. Their effect on transmission occurs mainly above altitudes of a few kilometres.

(ii) Clouds of Liquid and Solid Water. About 50% of the particles described in the previous paragraph are water soluble and act as nuclei for condensing water vapour in relative humidities above 35%. The drop size increases with relative humidity. Fig.1, shows a range of particle sizes.

(b) The presence of suspended matter has a vital but variable effect on atmospheric transmissions, depending upon meteorological factors as well as geographic location. Three models have been designed to help predict their influence. They are:-

(i) Extremely Clear. Aerosol concentration very low.

(ii) Maritime. Aerosol concentration largely dependent on salt nuclei.

(iii) Continental. Aerosol concentration largely dependent on industrial pollutants such as sulphuric acids and ammonium salts.

26. Effects of Haze, Fog, Cloud and Rain. If conditions are hazy or smoke is present, then in most circumstances transmission in the middle and far IR windows is superior to that at visible wavelengths. This is because the mean particle size of about 0.5µm is a relatively small fraction of infrared wavelengths, say 10µm, and therefore scattering by haze or smoke particles will be negligible. Seeing through fogs and clouds with a thermal imager is difficult when operating with either the middle or far IR window because scattering is intense. This is because the droplet sizes present range from 5 to 15 µm. Rain affects the thermal imager in two ways:-

(a) It tends to reduce the thermal contrast of the scene.

(b) It affects atmospheric transmission in both windows (almost equally) particularly in heavy rain. For example in heavy rain attenuation may amount to as much as 20% per km.


15

27. Processes Of Attenuation. Attenuation means the progressive reduction of the energy of a beam of radiation as it passes through a medium. Two processes are responsible for the attenuation, they are absorption and scattering, both due to the molecules and aerosols within the medium. For the atmosphere these consist of the constituents outlined in paragraph 3.

(a) Atmospheric Absorption. Photons of electromagnetic radiation interact with the atoms and molecules of the medium so that the photon energy is reduced or completely changed into another form, normally heat. Absorption in the atmosphere, for visible wavelengths, is quite negligible compared to scattering but is significant in the ultra-violet, infrared and radar regions of the spectrum. Aerosol absorption can be important in humid conditions. Some wavelengths of infra-red radiation are absorbed as they pass through the air. Thermal imagers avoid these wavelengths as they would not be very effective. Figure Three shows the variation in transmission of the Earth’s atmosphere at different wavelengths. Wavelengths between 6 µm and 8 µm suffer very high absorption but, fortunately for thermal cameras, the wavelengths between 9 µm and 13 µm (emitted by human bodies and warm vehicles) are transmitted quite well.


16

(b) Scattering. Radiating photons interact with the particles of the medium and have their directions changed without loss of energy. The amount of scattering depends on particle concentration and the ratio of particle radius, r, to wavelength of radiation, λ. Fig.2, illustrates how scattering is dependent upon wavelength for a variety of aerosols. When the radiating wavelength is lower than the radius of the aerosol, scattering is at its maximum and if the radiating wavelength is greater than the radius of the aerosol, the amount of scattering falls rapidly. Optimum scattering occurs when the diameter of the aerosol particle approaches the wavelength of the emitted energy.

28. Fig.3, shows a plot of atmospheric transmittance against wavelength, measured over a horizontal path length of about 1.9km at sea level. Relative humidity was about 58% at 20°C. The plot illustrate s the combined effects of molecular and aerosol scattering together with absorption, however, selective absorption by water vapour, carbon dioxide and ozone molecules is the dominant influence.


17

29. The curve of fig.2, exhibits several regions of high transmittance, known as atmospheric windows, separated by regions of high absorption. This leads to the infrared spectrum being broken up into a number of operational sections as follows:-

(a) Near IR - 0.8 - 3.0 µm.

(b) Middle IR - 3.0 - 6.0 µm.

(c) Far IR 6.0 - 15.0 µm.

(d) Extreme IR - 15.0 - 1000 µm.

30. Most thermal imagers operate in the middle or far IR regions as these contain the principal wavelengths that are emitted by objects at, or near, ambient temperatures. If a heat source is available to illuminate the scene then the near IR region can be used.

31. Increasing humidity reduces the transmission within each window, though the 8 - 14µm is more susceptible. The cause of this type of broadband absorption is considered to be two-fold, firstly absorption by pairs of water molecules loosely bound together and secondly the cumulative effect of water vapour absorption curves. These peak at wavelengths beyond 20µm but extend down into the two windows.

32. Atmospheric humidities are generally greater in summer than in winter. Thus atmospheric transmission within the windows will be less in the summer, particularly for the 8 - 14µm window.

33. Conclusions. The following general conclusions may be drawn:-

(a) Transmission in the 8 - 14 µm window is dominated by molecular absorption, especially, in summer.

(b) Transmission in the 3 - 5 µm window is influenced more by aerosol scattering in summer and winter.

(c) For short ranges, the 8 – 14 µm window appears to have the better atmospheric transmission for most purposes.

(d) Only in conditions of high humidity and extreme clarity in visual terms does the 3 - 5 µm window have a superior transmission, particularly for ranges in excess of 10km.


18

THERMAL IMAGER SYSTEMS

34. Processes Involved in Thermal Imaging.

(a) To be detected, and subsequently recognised and identified, an object must produce an object-to-background apparent temperature difference of sufficient magnitude to distinguish it from other variations in the background.

(b) The intervening atmosphere must not excessively blur or attenuate this signal.

(c) The operator must use an effective search procedure, know what to look for, and be able to point his sensor in the proper direction.

(d) The sensor must collect the radiant signal with an optic, and convert it to an electrical signal with good signal-to-noise ratio in a detector operating in an appropriate spectral band.

(e) This electrical signal must then be reconverted to an optical signal on a video display.

(f) Finally, the operator must be able to optimize his visual information extraction capability by using video gain and brightness controls.

(g) This entire process of conversion of an infrared scene to an analog visible scene must be performed so that contours, orientations, contrasts, and details are preserved or enhanced, and without introducing artifacts and excessive noise.

35. The basic elements of a thermal imaging system are depicted schematically in the block diagram below in terms of sequential processes in the imaging process. Particular systems may combine some functions and eliminate others, and there are many different specific implementations of these functions.


19

Collecting Optics and Filters

36. The optical system collects, spectrally filters, spatially filters, and focusses the radiation pattern from the scene onto a focal plane containing detector elements. An opto-mechanical scanner is interposed between the optical system and the detector. The ray bundle reaching the detector from the object, with the help of the scanner, traces out a TV-like raster. This process of detecting the scene sequentially is called scene dissection.

37. Materials fulfil a variety of roles in military infrared systems designed for surveillance, target acquisition, tracking and weapon aiming. The main groupings are as:-

(a) Windows through which radiations must pass.

(b) Missile Domes.

(c) Refracting optical systems such as telescopes for thermal imaging systems.

38. Radiation at the 8-13 micron wavelength is barely transmitted by optical glass and, therefore, alternative materials must be employed. Ordinary glass is optically opaque above 2.7µm because of water molecule absorption in the glass. The most widely used is germanium, a 10 mm thick slab of which has a transmission value of between 45% and 35% in the 8-13 micron waveband. Being a metal, germanium demands sophisticated manufacturing techniques to produce lenses with the


20

required high standards of form and 'optical' quality. Of course, germanium is opaque as far as the human eye is concerned, so combined optical and TI sights using common optical components cannot be constructed using germanium.

39. Thermal Imager sight heads which are open to the atmosphere must be capable of being kept clean, especially when used in the ground role where mud and dust is a constant problem. Thus, much development has been carried out to produce coatings for these infra-red glasses. These coatings must be able to resist the abrasion of wash/wipe systems, while maintaining the transmission properties of the glass.

40. For all three roles the chosen material should have:-

(a) Good transmission properties in the 3 - 5 or 8 – 14 µm bands as required. For optical windows, this implies both the visible and 8 - 14 µm bands over a wide temperature range. It being necessary to offset the effects of aerodynamic heating whilst the optical window is being carried in high speed airborne vehicles. For missile domes it means the 3 - 5 µm band, since a missile is programmed to seek and attack high temperature targets appearing within that band.

(b) Good mechanical strength, resistance to thermal shock for some applications and rain erosion resistance.

(c) Additionally, lens materials must be of high quality to avoid aberrations in both focus and colour. Suitable coatings should be available to reduce the losses due to reflection. Where germanium is used reflection losses are liable to be rather high and with several surfaces to pass through, the overall transmission losses could be severe. A family of coatings have been developed for use on both external and internal surfaces. The one used on


21

external surfaces has the diamond-like structure of carbon, which easily meets the requirements for high surface transmission and high erosion resistance.

41. Many classes of materials inherently have both military and commercial applications. While commercial applications are generally at lower performance levels than those of the military, this is not always the case. This includes those materials that provide specific military advantage and covers the physical properties, mechanical properties, behavior, and/or processing required to achieve that advantage. To illustrate the same, a tabulated summary of the US Defense Acquistions Office material performance requirements is listed below :-


22

Detecting Thermal Radiation

42. In order to be used to generate a picture for display the infra-red radiation from the scene must be converted into an electrical signal. This is accomplished by using the energy of the incoming radiation to cause an electric current to flow in a sensor (or detector) made from a suitable material. The brighter parts of the scene cause more energy to fall on the sensor and a greater current is produced. An ordinary television camera operates in a similar manner. However, an ordinary television camera is not very effective at a wavelength of around 10 µm for three simple reasons:

(a) The glass lens absorbs strongly at 10 µm so very little IR will enter the camera.

(b) The sensor, or detector, that converts visible light (0.6 µm) into an electrical signal does not operate at 10 µm. This sensor is often made from a silicon chip.

(c) The inside of the camera is at a temperature of around 300 K and thermal emissions from the inside of the camera would impair the picture.

43. The solution to the above problems require each point to be addressed and a change of material or operating environment made to compensate, as follows:

(a) The lenses of thermal cameras are made from a material that transmits IR radiation. Germanium is commonly used - but is quite soft and needs to be coated with a harder material to resist abrasion, etc. It also does not transmit visible light thus denying any visual transmission of imagery through the lens of a thermal camera.

(b) The material used to make the detector must be carefully chosen so that it responds to the wavelengths of the IR radiation. One common material is Cadmium Mercury Telluride (CMT). A list of those materials that are commonly used to detect IR radiation is in Table One, at the end of this handout. (The cut-off wavelength is the longest wavelength to which each responds.)

(c) To enable the sensor to respond to external thermal radiation from objects at a temperature of around 300 K, the sensor must be cooled to much lower temperatures. A sensor at 300 K would be blinded by its own thermal radiation. The normal operating temperature of these sensors is around 90K (or –180°C).


23

44. The incident electromagnetic field from the scene produces a disturbance within the detectors which is usually proportional to the incident energy. The type of disturbance depends on the type of detector, and may be an induced charge separation, an induced current, or a resistance change. Usually the incident radiation produces a voltage, and the two-dimensional object radiance pattern is converted into a one-dimensional analog voltage train by a scanning process.

45. Implementation. In the early days, only single element detectors were available and the systems produced were necessarily simple, being essentially thermal pointers. Using these, it was possible to identify a 'hot spot', but with only one detector element it was not feasible to scan the whole scene and produce a full picture. Thus the 'hot spot' would have to be identified using either an optical or an II sight. However, Thermal Imaging technology developed very rapidly and thermal pointers were never deployed in operational equipment. With the advance of technology, it has become possible to produce detector arrays containing multiple detector elements. Using opto-mechanical means, the field of view can now be scanned by the detector array to build up an image of the scene. The electrical outputs from the detectors are then processed and fed to a viewer, where a picture is built up using the television raster scan technique.

Infrared Detectors

46. Infrared detectors fall into two major categories depending on how the absorbed radiant energy interacts with the atoms of the detector material, namely thermal detectors and photon detectors.

47. Thermal Detectors.

(a) In this type of detector the radiation is converted to heat in the detector material and produces an increase in temperature. The materials used possess some temperature dependent property, such as resistance, which changes with the arrival of the radiation, producing an electrical output, eg, a change in current, proportional to the level of radiation. Examples of thermal detectors are the:-

(i) Thermocouple or Thermopile. A thermoelectric device (thermocouple or thermopile) is based on the presence of one or several junctions between two materials. The junctions properly arranged and connected develop a thermo-emk that changes with temperature, the so called Seebeck effect. In order for the sensitivity to be high the Seebeck coefficientshouldbeas high as possible. Certain alloys containingantimony and bismuth have very high Seebeck coefficients of 150 uV/K. The CMOS compatible combination aluminum/polycrystalline silicon gives about 65 uV/K. Thermopiles are the oldest type of infrared detectors and utilize thermo-electromotive force generated between two different types of conductors. Thermopiles are made from both metals and semiconductors.


24

(ii) Thermistor. Thermistors change their resistance with changes in temperature in a rather exaggerated way. They are primarily of two types :-

(aa) Positive Temperature Coefficient (ptc) thermistors whose resistance increases with increasing temperature (as it does for a pure metal), however, the response is usually extremely nonlinear

(ab) Negative Temperature Coefficient (ntc).thermistors, whose resistance decreases with increasing temperature.

(iii) Bolometer. The microbolometer detector is a thermal detector rather than a photon counter. It actually heats up as a result of being exposed to IR energy, which changes its electrical resistance proportionately. The most benefit from a microbolometer is that cryogenic cooling devices are not necessary. They also operate in the long-wave length region making them useful in outdoor and low-temperature applications. Thermal detectors usually have much lower sensitivity than photon detectors. As a result, they will not replace photon detectors in critical, low signal-to-noise applications

(iv) Pyro-Electric Detector. With pyroelectric detectors, thermal changes alter the electrical polarization which appears as a voltage difference. When infrared radiation is incident, temperature changes are generated in the crystal. An electric charge is then generated on the surface of the crystal in accordance with the amount of temperature variation. Pyroelectric detectors are current sources with an output proportional to the rate of change of its temperature. Capable of extremely rapid response and insensitive to DC effects, they are frequently used in industrial radiometric systems and in the analysis of infrared lasers.

(v) Pneumatic Detectors. There are two detectors: Golay cells and capacitor microphones. In Golay cells, the sealed xenon gas expands


25

when it is warmed by incident infrared radiation. The resultant variation of pressure shifts a mirror located between a light source and a photocell, varying the amount of light entering the photocell and thus changing the output of the photocell. In capacitor microphones, the varying expansion of the gas affects the capacitor film, which in turn produces the variation in the electrostatic capacity

(b) The response of thermal detectors is independent of wavelength if the detector is perfectly grey (ie, the response per watt is constant). It is also rather slow because of the inherent thermal capacity of the detector. This limits their ability to deal with rapid changes, eg, moving targets. With the exception of the bolometer and the pyro-electric detector, thermal detectors are unsuitable for thermal imagers.

48. Photon (or Quantum) Detectors.

(a) Quantum type detectors feature high detectivity and fast response speed. Responsivity is wavelength dependent and, except for detectors in the near infrared range, cooling is normally used with these detectors. Quantum type detectors are further classified into intrinsic types and extrinsic types.

(i) Intrinsic Type Detectors. Intrinsic type detectors have detection wavelength limits determined by their inherent energy gap and responsivity drops drastically when the wavelength limit is exceeded. Among them. the photoconductive detectors, which change their conductivity when infrared radiation is incident, have high responsivity and allow simple signal processing. The photovoltaic detectors generate an electric current when infrared radiation is incident and have high responsivity and a fast response speed. HgCdTe or PbSnTe detectors are also included in the intrinsic type detectors. The wavelength of peak responsivity of these detectors can be changed by controlling the composition of the ternary mixture. In particular, the HgCdTe detectors are useful since they respond to wavelengths in the 3 to 5 µm and 7 to 13 µm ranges.

(ii) Extrinsic Type Detectors. These are photoconductive detectors whose wavelength limits are determined by the level of impurities doped in high concentrations to the Ge or Si semiconductors. The biggest difference between intrinsic type detectors and extrinsic type detectors is the operating temperatures. Extrinsic type detectors must be cooled down to the temperature of liquid helium.

(b) These devices are also referred to as photon detectors because the absorption process involves direct interaction between arriving photons and the electrons of the detector material. Two types of interaction are possible:-


26

(i) Photoemissive Detectors. By which electrons are released from a photocathode into vacuum. This is unsuitable for detecting middle and far infrared radiation since the photon energies are too low.

(ii) Photoconductive Detectors. These absorb photons to elevate an electron from the valence band to the conduction band of the material, and so increase the conductivity of the detector. The additional current output can be processed either as a voltage or a current signal. To detect far IR (8-12µm) radiation they must be cooled to eliminate the noise generated by thermally generated carriers. Examples of substances used in photoconductive detectors include indium antimonide(InSb), quantum well infrared photodetector (QWIP), mercury cadmium telluride (mercad, MCT), lead sulfide (PbS) and lead selenide (PbSe).

(iii) Photovoltaic Detectors. They absorb photons to create an electron-hole pair across a PN junction to produce a small electric current or potential difference. Substances used include indium antimonide (InSb), mercury cadmium telluride (MCT) and platinum silicide (PtSi) -silicon Schottky barrier.

(c) The table above lists semiconductor materials suitable for photo-detectors. The cut-off wavelength, λc, for each is determined by the energy gap for the material. At short wavelengths the photon energy, E = h×v (where h=Planck's constant and v=frequency) is high and easily elevates electrons to the conduction band. At wavelengths above λc, the photon energy is not high


27

enough to elevate any charge carriers to the conduction band and response ceases sharply. Thus λc is inversely proportional to the energy gap for the material.

(d) The figure below shows a typical photon detector response. Output increases with wavelength because, for a given power radiated, there will be more photons at longer wavelengths than at the shorter ones. Hence, more charge carriers are released per watt of arriving energy.

(e) Early infrared homing missiles were equipped with detectors which could only sense the hottest parts of the target. (eg, Lead Sulphide with λc at 3.1 µm) To 'see' more of the thermal signature for the target these detectors were gradually replaced by Indium Antimonide, InSb, which can respond out to 6µm. Thermal imaging of ordinary scenes at 300K, radiating strongly between 8 and 12 µm, requires a different detector material. The discovery of Cadmium Mercury Telluride, CMT, made thermal imaging possible, as against the sensing and tracking of heat which was carried out in those earlier missile systems.

(f) A popular detector material is Platinum Silicide (PtSi) which is sensitive in the 1-5.5 µm region. PtSi has low quantum efficiency (less than 1%). Most IR detectors are photon counters; that is they count IR photons over very short periods of time. Quantum efficiency refers to the relative efficiency at which IR photons are collected and converted into electrical charges. PtSi operates in the short wavelength region (1-5 µm), has good sensitivity (as low as 0.05oC), and has excellent stability. Two other detector materials (both need cooling) commonly used are :-

(i) Indium Antimonide (InSb) (responsivity 2.2 - 4.6 µm, high quantum efficiency of around 80-90%) and ;


28

(ii) Mercury Cadmium Telluride (HgCdTe) - (responsivity 8 - 12 µm). The detector material, CMT, is composed of Cadmium Telluride, CdTe, a semiconductor with a fairly large energy gap, and Mercury Telluride, HgTe, another semiconductor with a very small energy gap. When the two are combined in appropriate proportions, it becomes a very effective detector in the 8 - 12 µm band. By adjusting the proportions, CMT can be 'tailor made' for operation in the 3 - 5 µm window. The penalty of operating with a material that has the small energy gap required to be effective at longer wavelengths, is that a significant number of electrons are elevated to the conduction band by thermal excitation, thus swamping the photo-generated carriers. By cooling the detector to 77K, the temperature of liquid air, thermal excitation is virtually suppressed and 8 - 12 µm operation is then possible. By varying the Hg, Cd, Te mixture, it can be tailored to the MWIR or LWIR. The most popular is the LWIR.

(g) Both can work with cooling in the 3 µm band, but MCT requires cryogenic temperatures from liquid Nitrogen or Stirling cycle coolers to work at 8 to 12 µm.

(h) Photon detectors are characterised by a photon-electron interaction, which means that the absorbed energy is not transferred to the material as heat, because the electron-hole pairs are conducted out of it by an operating potential difference placed across the external terminals. The response increases linearly with wavelength up to cut-off and operation is extremely fast by comparison with thermal detectors because the thermal capacity is not involved.

49. Uncooled Detectors. "Uncooled" thermal imaging detectors refers to detector arrays that operate at or above room temperature. The term "uncooled" is used to distinguish this technology from the historical norm, which is to use detectors that only operate at cryogenic temperatures, e.g. the temperature of liquid nitrogen (77oK) or lower. Two basic uncooled detector types have emerged today, ferroelectric detectors and microbolometers. Ferroelectrics have been developed by Texas Instruments and GEC Marconi; microbolometer technology has been developed by Honeywell.

(a) Ferroelectric detector technology takes advantage of a ferroelectric phase transition in certain dielectric materials. At and near this phase transition, the electric polarization of the dielectric is a strong function of temperature; small fluctuations of temperature in the material cause large changes in polarization. Then, if the sensor is maintained at a temperature near the ferroelectric phase transition and if the optical signal is modulated (with a synchronous chopper), then, an infrared image can be readout that reflects the scene temperatures. The ferroelectric BST array is a ceramic material consisting of barium, strontium and titanium salts. The nominal desired composition is Ba0.66Sr0.34TiO3. Because ferrolectrics retain their


29

electric polarization after application and removal of an electric field, their polarization depends on temperature.

(b) Microbolometer arrays, on the other hand, consist of detectors made from materials whose electrical resistivity changes with temperature. Each detector is part of a readout circuit that measures the resistance of the element as a signal.

(c) The advantage of uncooled systems is system lifetime and cost (cooled sensor systems need to be chilled to temperatures as low as 77K which requires the use of expensive and intricate cryogenic systems).

Detector Cooling

50. Current high performance TI systems, operating in the 8-13 µm waveband, have their detectors cooled to about -200oC. At this temperature, the IR-induced signal is clearly distinguishable from the naturally occurring 'noise', enabling target temperature differences of less than 1 oC to be detected. The two most commonly used methods of cooling detectors are the Reverse Stirling Cycle engine and the Joule-Thompson minicooler.

51. Reverse Stirling Cycle Cooler.


30

(a) Principle. The Reverse Stirling Cycle engine is in effect a closed-cycle refrigerator, in which a gas (usually Helium) is compressed in one cylinder and then allowed to expand through a nozzle in the other the resulting drop in pressure producing a drop in temperature. It requires a supply of electricity to power the electric motor that turns the pump, but the gas is sealed in the system and does not require topping-up. Helium is used because it remains gaseous to very low temperatures (ordinary air begins to liquefy at around 80K or –190°C) and because Helium has low viscosity and flows easily through small pipes.,. The cold gas is then recycled through the compressor until the required temperature is achieved. For rapid cooling, when first switched on, the flow of gas can be increased then, once the operating temperature is reached, the flow can be reduced to maintain the temperature. The compressed gas can come from a re-chargeable gas bottle or from a small compressor. As the air heats up during compression, it must be allowed to cool to ambient temperature before use.

(b) Advantages. The Reverse Stirling Cycle cooler offers a number of advantages, being reasonably small and light and having a small input power requirement. As a closed cycle system, it does not require an external gas supply from a compressor or bottles and this makes it simple from the operator's point of view. The small size of the Stirling Cycle cooler and the fact that it is a self contained unit are also attractive features for system designers, since units can be built into imaging systems.

(c) Disadvantages. On the other hand, the system does have two major drawbacks.

(i) First, the compressor motor can produce audible noise and generate vibrations at the detector, leading to a degradation of the thermal picture. However, these problems are being minimised by the use of linear electric motors (with no rotating parts) and by having the cold finger separated from the compressor by a narrow bore pipe, which isolates the detector from the vibration source.

(ii) The second and most important problem is the long cool-down time of the Stirling Cycle cooler, which can be of the order of ten minutes. This can be an important factor if a device needs to come into action at very short notice.

52. Joule-Thompson Minicooler.

(a) Principle. The Joule-Thompson (JT) minicooler also uses the rapid expansion of a high pressure gas to produce a temperature reduction, but in this case the gas is externally supplied, either from bottles or a compressor (or both). The incoming high pressure gas is fed through a heat exchanger to a nozzle, where it expands and cools. The gas is then exhausted to atmosphere, passing back over the heat exchanger as it does so and thus


31

cooling both the detector and the incoming high pressure gas. Once a continuous flow has been established, the temperature falls rapidly to a point where the gas emanating from the nozzle becomes a liquid. At this point, the change of state can be detected and the flow of gas automatically reduced, to give a stable temperature with minimal gas consumption.

(b) Advantages. The JT minicooler itself is a simple device with few moving parts and is, therefore, fairly reliable and silent in operation. From the operational point of view, its ability to reach operating temperature rapidly—cool down times of less than one minute being achievable—is a decisive factor in its favour.

(c) Disadvantages. The main problem associated with this device is the need for a continuous supply of pure high pressure gas, a requirement which imposes significant design, maintenance and logistic penalties. The gas must be kept clean of impurities if it is not to become an explosive mixture and this process can require filters and adsorbant chemicals. With operating pressures of up to 4000 psi, the pipework and valves must be carefully designed and maintained if they are to operate reliably and safely. Finally, the supply of fresh filters, chemical adsorbants and replenishment gas bottles imposes a considerable logistic load. Nevertheless, the JT minicooler is still the preferred system where speed into action is a priority requirement.


32

53. Other Cooling Methods. There are two further methods of cryogenic cooling which have been examined for thermal imagers. :

(a) Liquid nitrogen was considered in the late 1960s but the anticipated difficulties of handling this material under field conditions led to its not being pursued.

(b) Thermo Electric Coolers.

(i) A thermoelectric cooler (TEC), sometimes called a thermoelectric module or Peltier, is a semiconductor-based electronic component that functions as a small heat pump. By applying a low voltage DC power source to a TEC, heat will be moved through the cooler from one side to the other. One cooler face, therefore, will be cooled while the opposite face simultaneously is heated.


33

Consequently, a thermoelectric cooler may be used for both heating and cooling by reversing the polarity (changing the direction of the applied current). This ability makes TECs highly suitable for precise temperature control applications as well as where space limitations and reliability are paramount and CFCs are not desired.

(ii) A typical single stage cooler (Figure 1) consists of two ceramic plates with p- and n-type semiconductor material (bismuth telluride) between the plates. The elements of semiconductor material are connected electrically in series and thermally in parallel. When a positive DC voltage is applied to the n-type thermo-element, electrons pass from the p- to the n-type thermo-element and the cold side temperature will decrease as heat is absorbed. The heat absorption (cooling) is proportional to the current and the number of thermoelectric couples. This heat is transferred to the hot side of the cooler, where it is dissipated into the heat sink and surrounding environment.

(iii) The theories behind the operation of thermoelectric cooling can be traced back to the early 1800s. Jean Peltier discovered there is a heating or cooling effect when electric current passes through two conductors. Thomas Seebeck found two dissimilar conductors at different temperatures would create an electromotive force or voltage. William Thomson (Lord Kelvin) showed that over a temperature gradient, a single conductor with current flow, will have reversible heating and cooling. With these principles in mind and the introduction of semiconductor materials in the late 1950's, thermoelectric cooling has become a viable technology for small cooling applications.

(iv) Coolers based on the Peltier Effect do not require a gas or a liquid. In these, two dissimilar conductive materials are connected to


34

form two junctions. When an electric current is passed through the circuit, the temperature of one junction rises while the other falls. By connecting devices in series, with the hot junction of one adjacent to the cold face of its neighbour, a cumulative effect can be produced. Unfortunately, these systems have a low heat load capacity and can only achieve temperatures of around -loC, which is insufficient for high performance thermal imaging.

Detector and Thermal Imager Performance Parameters

54. The parameters by which infra-red detectors usually specified, are responsivity, cut-off wavelength noise, detectivity, and time constant. These are elaborated upon as given below :-

(a) Responsivity Ry. It is the ratio of electrical output to radiant input power. For a photoconductor it is usually expressed in volts per watt; For a photodiode, which is often used to give a current output, it may be stated in amperes per watt. As the output electrical signal, due to a change of incident radiation, is a very small fraction (about 10"6) of the quiescent voltage or current, responsivity is measured by exposing the detector to chopped radiation from a calibrated source and measuring the alternating voltage component at the topping frequency.

(b) Cut-off wavelength, λΟ. It is the longer of the two wavelengths at which the responsivity (the detector is down to half its maximum. The wavelength of maximum responsivity is designated λpk.

(c) Noise. Noise, together with responsivity. determines a detector's ability to detect small input signals. It is a function of frequency; there is a low-frequency region of flicker or 1/f noise, a flat mid-region, and a high-frequency roll-off determined by the detector time constant. Noise may be specified at one or a number of frequencies, or as a noise spectrum. It is expressed in V/Hz1/2.

(d) Noise Equivalent Power. The performance of all detectors is limited by noise. This is a random signal that appears at the output of a detector, even when there is no input signal. It is generated by the electronic circuitry, the detector and the incident photon stream. Each has the effect of producing a tiny voltage change within the components associated with detection. The noise equivalent power, NEP, of a detector is the minimum power that can be detected. It is defined as the incident radiant power on the detector that produces an output signal equal to the noise output signal. Clearly above this level the target can be extracted from the noise and below it the target is hidden by it, or is effectively part of it. Typical values for a photon detector would be approximately 2×10-8W at 300K and 10-10W at 77K. For a pyroelectric thermal detector it would be around 10-7W. The NEP of a


35

particular detector depends upon its overall area and the operating bandwidth in Hz. For a high definition thermal imager producing a TV-compatible Charge Coupled IR, CCIR, output, the bandwidth is about 4MHz. The photon detector is obviously much the more sensitive, particularly when cooled to 77K.The various noise sources are known as:-

(i) Temperature Noise. Also known as Johnson or Nyquist Noise, is caused by random motion of electric charges in resistors and other components.

(ii) 1/f Noise. Also known as 'Flicker Noise' is caused by fluctuations in the numbers of charge carriers at the contacts of the detector. This is designated as 1/f, where f=frequency, because the associated noise power is inversely proportional to frequency.

(iii) Thermal Generation-Recombination Noise. It is caused by fluctuations in the generation and recombination rates of electrons and holes produced by thermal excitation. It causes fluctuations in the electron and hole populations about mean values. This source of noise is reduced by cooling the detector.

(iv) Background Generation-Recombination Noise It is also called 'Photon Noise'. Fluctuations in the arrival rate of photons emitted from a target back ground also contribute to fluctuations in the electron and hole populations. This is the dominating source of noise when the detector is operated at a sufficiently low temperature to eliminate or reduce other sources. It produces a condition known as Background Limited Perfomance, BLIP.

(v) The first three noise sources mentioned above are internal and can be controlled to a certain extent, whereas photon noise is external and a feature of the target background. It represents the ultimate limit of performance that can be obtained with a thermal imager, as is sometimes the case with an image intensifier. Radio and radar receivers, on the other hand, are limited by temperature noise. The figure below, illustrates the effect when the voltage output signal of a detector is viewed on a Cathode Ray Oscilloscope.


36

(e) Detectivity, D*.

(i) It is a figure of merit (or quantum detectors that takes account of responsivity. noise, and detector element size; it is expressed in cm Hz1/2W-1. An advantage of D* as a figure of merit for quantum detectors, is that it enables one to calculate a theorectical maximum detectivity. It would apply if performance were limited only by noise due to fluctuation of the background radiation. That maximum depends on the cutoff wavelength of the detector and its field of view, and is designated BLIP (tor Background Limited Infra-red Photodetection).

(ii) Detector responsivity is defined as the output voltage level per watt of radiant power absorbed. This can be improved by:-

(aa) Increasing the purity of the detector material.

(ab) Coating the detector surface to minimise reflection and maximise absorption.

(ac) Operating at low temperature.

(iii) Peak values of about 105 Volt/Watt are obtainable at 77K. The reader should not be misled by this. Detected power levels may be of the order nano to pico Watt, which means that actual detector output may be of the order microvolt.

(f) Time constant, τ. IT is determined by measuring the time between incident radiation being cut off and the output of the detector falling by 63%.


37

Thermal Imager Performance

55. The performance of a thermal imager is usually specified in terms of temperature resolution, angular resolution, and field of view.

(a) Temperature resolution is a measure of the smallest temperature difference in the scene that the imager can resolve. Temperature resolution depends on the effciency of the optical system, the responsivity and noise of the detector, and the signal-to-noise ratio of the signal-processing circuitry. It can be expressed in two ways:

(i) Noise Equivalent Temperature Difference (NETD). It is the temperature difference tor which the signal-to-noise ratio at the input to the display is unity.

(ii) Minimum Resolvable Temperature Difference (MRTD). It is the smallest temperature difference that is discernible on the display: it is typically less than 0.3 K and can be less than 0.1 K.

(b) Angular resolution is, in principle, the effective width of the detector element divided by the focal length of the infra-red optics. However. it may be degraded by the optical transfer (unction of the lens or the frequency re-sponse of the signal-processing circuitry or the display. Angular resolution is typically 1 milliradian and can be as small as 0.1 milliradian Angular resolution is sometimes called instantaneous field of view (IFOV).

(c) System field of view, sometimes called total field of view (TFOV), is the plane angle subtended by the scene imaged. When the subtended angles parallel and perpendicular to the scanning direction differ, the angle sub-tended by the diagonal may be given. Together, the required system field of view and angular resolution determine the number of pixels obtained.

(d) The term, field of view, is also used to describe the convergence angle of the infra-red optics to which the detector is matched. Thus, the terms, instantaneous field of view, system field of view, and detector field of view, refer to three distinct quantities and must not be confused.

Infrared Image Capture Mechanisms and Detector Geometry

56. There are two basic types of infrared imaging systems: mechanical scanning systems and systems based on detector arrays without scanner. It should be mentioned that detector arrays as well are used for scanning systems, but the number of detector elements (picture elements - pixels) generally is smaller in this case.


38

(a) A mechanical scanner utilizes one or more moving mirrors to sample the object plane sequentially in a row-wise manner and project these onto the detector. The advantage is that only one single detector is needed. The drawbacks are that high precision and thus expensive opto-mechanical parts are needed, and the detector response time has to be short. As mentioned above, detector arrays are also used for this application. For example, a long linear detector array can be used to simultaneously sample one column of the object plane. By using a single moving mirror the whole focal plane can be sampled. In contrast, when a single detector is used, two mirrors moving in two orthogonal directions must be used, one of them moving at high speed, the other one at lower speed.

(b) Detector arrays operated as focal plane arrays (FPA) (or staring arrays) are located in the focal plane of a camera system, and are thus replacing the film of a conventional camera for visible light. The advantage is that no moving mechanical parts are needed and that the detector sensitivity can be low and the detector slow. The latter is a result of that the integration time can belong. The drawback is that the detector array is more complicated to fabricate. However, with the ascent of rational methods for semiconductor fabrication, economy will be advantageous, provided that production volumes are large. The general trend is that infrared camera systems will be based on FPAs, except for special applications.

57. The spatial resolution of the image is determined by the number of pixels of the detector array. Common formats for commercial infrared detectors are 320x240 pixels (320 columns, 240 rows), and 640x480. The latter format (or something close to it), which is nearly the resolution obtained by standard TV, will probably become commercially available in the next few years. Today, for example indium antimonide and platinum silicide detectors are commercially available in the 320x240 pixels format. Typical pitches between pixels are in the range 20-50 um.

58. Detector arrays are more complicated to fabricate, since besides the detector elements with the function of responding to radiation, electronic circuitry is needed to multiplex all the detector signals to one or a few output leads in a serial manner. The output from the array is either in analogue or digital form. In the former case analogue to digital conversion is usually done external to the detector array. The electronic chip used to multiplex or read out the signals from the detector elements are usually called simply readout integrated circuit (ROIC) or (analogue) multiplexer. The ROIC is usually made using silicon CCD (charge coupled device) or CMOS technology. However, the detector elements must often be fabricated from more exotic materials as discussed above. The exceptions are e. g. platinum silicide or micro-bolometers which can be based on silicon technology. In the former case a hybrid approach is most common, in which case all the detector pixels are fabricated from a separate detector chip. This detector chip is then flip-chip bonded to the ROIC chip. Flip-chip bonding involves the processing of metal bumps onto contact holes, one per pixel, of both the detector chip and the ROIC. Using special equipment, the two chips are first aligned to each other. Then the chips are put in


39

contact, while applying heat and/or mechanical force. During this process the two chips become electrically connected to each other via the metal bumps. Usually indium is used for the bumps due to its excellent low temperature properties.

59. Uniformity of the detector elements across the array is a key issue for obtaining high performance. In fact, individual pixel response characteristics differ considerably across an array in most cases. Therefore so called pixel correction has to be done prior to the presentation of the final image. This amounts to calibrating each individual pixel, by exposing the array to calibrated surfaces of known temperature.

60. Scanning.

(a) The designer of a thermal imager is faced with two alternatives, either he can choose a relatively simple detector array and a complex scanner, or he can choose a simple scanner at the expense of more complicated electronics. The various options are:-

(i) Single element Scan.

(ii) Multi-element Scan.

(iii) Focal plane processing.

(b) Image Formation. The alternative to electronic scanning is mechanical scanning. When mechanical scanning is used then a small number of sensors (perhaps just one) is used and a number of moving mirrors or prisms are placed between the lens and the sensor. As the mirrors move then light from each part of the scene is reflected onto the sensor in turn, usually following a similar raster pattern to that used in ordinary television. In most mechanically scanned systems two moving mirrors are used, one to scan horizontally (for the lines of the picture) and a second to scan vertically (for the fields). To enable the picture to be transmitted, recorded and displayed using commercially available systems, the output signal from the imaging module must be compatible with standard television to reduce costs. Image formation involves the following :-

(i) Field Scanning. The field mirror has to rock back and forth at a rate of 50 Hz and this is well within the capabilities of electromagnetic actuators. The mirror must be kept small, with low mass, so that it can be moved quickly. The movement follows a ‘sawtooth’ pattern with a steady movement to scan down the picture followed by a rapid ‘flip’ back up again, ready to scan the next field.

(ii) Line Scanning. The lines occur at a rate of 15,625 Hz which is far too fast for an oscillating system. To provide the high frequency line scan, two simple techniques are applied. Firstly, the imager scans eight lines at a time - this reduces the frequency by a factor of eight.


40

The group of eight picture lines is called a swathe. Secondly, the imager uses a rotating hexagonal mirror instead of a rocking mirror, and this allows it to scan six groups of lines for each rotation, 48 lines in total. Thus, the hexagonal mirror must rotate by 15,625 ÷ 48 or about 326 revolutions per second, 19,530 rpm, an achievable rotational speed. The system must be kept dust-free as the high rotational speed of the mirror will cause dust to embed itself in the surface of the mirror during any impact. This will damage the mirror. As each facet of the octagonal mirror passes through the light path then each of the sensors is presented with light from one line of the picture. Since a single television line scan normally takes 64 µs and this system scans eight lines at a time then it takes 8×64 µs or 512 µs to complete this. As the line scan occurs, the field mirror is steadily moving through part of its vertical scan of the scene. By the time that the next facet of the hexagonal mirror approaches then the field mirror has moved down by the equivalent of eight picture lines. Thus, the next set of eight lines to be scanned will be immediately underneath the previous eight. This means that the picture has a larger tilt, 8/287.5 or 1.6° down at the right-hand side, than an ordina ry television image which has a tilt of 1/287.5. The tilt is insignificant as most television monitors produce pictures with some distortion.

(iii) One Field. After 39 swathes have been scanned, the field is complete. The number of lines scanned is 8 × 39 = 312. There is no half-line, as there would be in an ordinary television field and a blank half-line is added electronically to each field so that there are 625 lines in each complete picture (two interlaced fields).

(iv) Synchronising. The motor that turns the hexagonal mirror and the actuator that tilts the field mirror must be synchronised to move at the correct times. The circuits that provide the timing signals for this movement also provide the timing signals that are incorporated in the outgoing television signal for line and field synchronisation.

61. Single Element Scan.

(a) The focussed image of the scene is scanned line by line over a single detector element as shown at Fig.3. Scanning is achieved by combining the motion of two moving mirrors, one oscillating rapidly about a vertical axis, to generate a horizontal scan, and the other oscillating more slowly about a horizontal axis, to generate the vertical scan. The varying output of the detector is amplified and applied to a display, which is scanned in synchronism with the detector, to generate a visual picture of the scene. Fig.4, shows the signal from the detector element over the period of one line scan. It is similar to that produced by a vidicon in TV work and merely


41

requires the appropriate line and frame synchronisation signals to be added in order to form the display video signal.

(b) This arrangement is incapable of coping with the data rates typical of TV systems because the frequency of horizontal oscillation is then excessive and the detector is unable to follow the incoming radiation level changes. It has been used extensively in simple low performance imagers where scanning rates are much lower.

62. Multi-Element Scan.

(a) Multi-element detector arrays overcome some of the short-comings of single element detector systems. The elements can be arranged so that a simpler, slower speed, scanning system can be used (parallel scan) or that a cumulative output can be obtained by summing the outputs of the individual elements (serial scan). In practice, combinations of these are much more likely (Serial-parallel scan). In all cases of serial scan the signal-to-noise ratio is improved by √n, where n is the number of elements.

(b) Parallel Scan. In its simplest form the image of the whole scene can be scanned horizontally across a number of detector elements mounted vertically as shown at Fig.5. Fig.6, illustrates how banded parallel scan can be achieved with a multi-faceted rotating mirror (similar to that used in the Milan Infra-Red Adaptor, MIRA). Parallel scanning allows much lower scanning speeds for a given number of picture points but the performance of the detector elements, amplifiers and LED's, must be extremely uniform otherwise the picture will be streaky. An element failure will remove one or in


42

the case of banded scan several lines from the picture. Parallel scanning possesses the virtue of slow scanning speeds at the expense of picture uniformity and defective elements. With serial scanning, since each element samples every pixel of the scene and element-to-element non-uniformities are smoothed out, overall picture uniformity is good. However, scanning speeds are prohibitively high for a single line of detectors providing a TV compatible output.

(c) Serial Scan. This type of scan is essentially a development of the

single element scan described above.

(i) Fig.7, illustrates an array of eight elements, mounted horizontally, over which the image of the scene is scanned. Any pixel of the scene is sampled in turn by each detector element and the output from each element is delayed by τ seconds relative to its neighbour in the scan direction. ( τ is the time required for the scan to move from the centre of one element to the same point on its neighbour, in the direction of scan.)


43

(ii) Fig.8, Shows the start of the sequence of events as the picture scene is scanned over the serial array of detectors. The beam advances across the detectors until detector 1 is viewing picture element A. Its output is a voltage proportional to the thermal content of that element. This signal is now fed through a delay equal to the time required for the scan to move and cover detector 2.

(iii) When the scan arrives at the second detector, as shown by Fig.9, its output rises to the level previously held by detector 1 and the output of detector 1 now changes to a level dependent upon the thermal scene of picture element B. The output of detector 2 is added to the output of detector 1 arriving via the delay. The two outputs are identical and their sum is delivered into the second delay. At the same time, the output of detector 1, now containing the content of picture element B, is delivered to the delay prior to detector 2.


44

(iv) Fig.10, shows the position as the scan moves to detector 3, the output of which rises to the level previously held by detector 2. Detector 2 changes to the level previously held by detector 1 and detector 1 rises to the level appropriate for picture element C. The delayed sum signal from detector 2 is added to the output from detector 3 and passed onto the next delay. It is now three times the value for picture element A. At the same time the delayed signal from detector 1 is added to the output from detector 2 and passed on to the next delay. It is now twice the value for picture element B. Also the output from detector 1 is delivered to the following delay. This sequence continues through the scan until each picture element has been looked at eight times, at which point the scan ceases, the next polygon mirror is brought into play, the frame mirror has moved to its new start position and the next swathe begins.


45

(v) Fig.11, shows the electronics involved. One great advantage is that any noise present, which by its very nature is random, tends to be reduced against the signal, which is constant. The noise sums as the square root of the number of elements and at the same time the signal continues to sum. The signal-to-noise ratio is increased by √n, or √8=2.83 in this case, over that for a single array. The detector outputs could be delivered to a CCD type of array where they would be injected at the appropriate points. The CCD would then act as the delay element. The output of this type of detector is in discrete steps rather than the continuous output of the single detector.

(d) Serial-Parallel Scan. This involves scanning an image of the scene over a matrix of elements, as shown at Fig.12. It is designed to combine the separate virtues of serial and parallel scan, ie, improvement of signals-to-noise ratio and slow speed of scan. The earliest version of TICM2 (Thermal Imager Common Module) used a 6×8 array of CMT elements approximately 50µm square.


46

63. Focal Plane Processing. The drawback of serial scan is the relative complexity of the electronics, and with serial-parallel scan the added disadvantage of the number of electrical connection required. Since the detector elements are cooled to 77K, these connections serve as conduction paths for the flow of heat from outside the mini-cooler. This increases the evaporation rate of liquid air and limits the useful life of a bottle of compressed air. Two solutions to this problem are being developed, they are the SPRITE detector and Focal Plane arrays respectively.

(e) The SPRITE Detector.

(i) A SPRITE (Signal PRocessing In The Element) detector, also known as the TED (Tom Elliott Detector) overcomes the problem of connections to a considerable extent. Instead of a row of discrete detectors scanned serially, the SPRITE consists of a strip of CMT, about 700µm long and 62.5µµµµm wide, with only three connections and one pre-amplifier as shown at Fig.13.

(ii) The long axis of the strip is in the scan direction and the electric field, in the direction of the long axis, is adjusted so that the drift velocity of photo-generated charge carriers precisely matches the scan


47

velocity. As the image of each pixel of the scene, in turn, travels the length of the strip, the carriers it excites travel with it and continue to accumulate. Thus the strip of CMT material not only serves as the detector but also as a delay element and summing circuit, integrating the signal over the length of the strip. The output is picked-off at the read-out zone, a small region about 50000µm long at the end of the strip, as a negative voltage level proportional to the incident radiation at that point in the scene. It is the sum of the contributions obtained by many 'looks' at any one pixel of the scene. All the points together produce one line.


48

(iii) Fig.14, shows the scanning arrangement and detector configuration for the TICM2 showing eight sprite in parallel. This is equivalent to an array of 64 discrete elements but requires only 24 connections instead of 65. The actual TICM2 arrangement is a little more complex than is shown at Fig.14. The imager output is required to be 625 line TV compatible and the interlaced scanning arrangements reflect this. One picture field is formed by using 39 swathes each comprising 8 lines. Thus a single field comprises 39×8=312 lines. The frame is formed by two fields, the first covering the odd lines and the second covering the even lines, hence, there are 624 in all. However, only 512 of the lines are used for picture formation, 112 of them being used as the period during which the two field flybacks are achieved. One line is effectively added to form the TV compatible 625 lines. One field is scanned in 1/50 sec. and two fields comprise the frame, formed in 1/25 sec. as with TV. In order to reduce the angular velocity of the polygon it has six faces. This means that six swathes, or 6×8=48 lines, are formed during one rotation, which must occur in 48/624 of 1/25 sec. or 48/(624×25) sec. The inverse of this is the rate of rotation for the polygon in revolutions/second, ie, 325. Shaft rotation is normally specified in RPM which in this case is 19500 RPM.

(a) Focal Plane Arrays.

(i) The next major advance in thermal imager design will come through the application of CCD technology to provide arrays with considerably larger numbers of elements, eg, 105 to 106 and possibly a non-mechanically scanned or 'staring' array. There are a several distinct technologies of thermal imaging available today. Most of the newer camera designs are based on a focal plane array (FPA) device that is a two-dimensional array of infrared detectors used to create an image. Earlier systems used either a single element detector or a small array of detectors and scanned the scene across the detectors with rotating mirrors. Other popular FPA technologies include both uncooled (microbolometer, proelectric vidicon) and cooled (platinum silicide, indium antimonide) FPA systems. A focal-plane array (FPA) detector is any detector that has more than one row of detectors. A typical infrared FPA system has 256*256 detectors (256 columns and 256 rows). Detector arrays of 256 * 256 are common with MRTD down to 0.03 0C. There are two types of infrared FPAs: monolithic and hybrid. Monolithic FPAs have lower performance compared to their hybrid counterparts because of a lower fill factor. A higher fill factor results in a much higher sensitivity. When a CCD detector is used in a measurement infrared FPA camera, the errors must be compensated for. Optimum battery life is achieved by using a CMOS multiplexer detector readout and high-efficiency rotary Stirling cooler. The


49

problems involved in FPA design and integration are severe because of::-

(aa) The difficulty of extracting the leads from an array having a high density of elements.

(ab) The need to maintain a uniform response over all the elements of the array due to the fact that thermal contrasts, particularly in the 8 - 14 µm window, are so small.

(ii) Image Formation.

(aa) In the television camera, light from the scene is collected by a lens and focused onto a light-sensitive material. In modern camcorders, the sensor is called a CCD and it has no moving parts. The CCD sensor consists of an array of light detectors on a silicon chip, one for each pixel that will appear on the screen. An ideal television screen would display 575 lines and each line would have 575 × 4 ÷ 3 = 767 pixels across it, to give a total of 575 × 767 or about 440 000 pixels. Each little light sensor converts the light that it receives into an electrical signal and outputs its signal when the scan reaches its part of the picture. The scanning is performed electronically by components that are also built into the same silicon chip.

(ab) The lens used has to provide a high quality image and is usually made up from at least five separate lenses. If a zoom lens is used then there might be as many as twelve separate pieces of glass or germanium (for a thermal camera) in the lens The basic property of the lenses used is that the light from the scene is brought to a focus. However, the light from one part of the scene will focus on a different part of the light sensor from light from other parts of the scene.

(ac) The image sensor is located on the focal plane of the system and this is called a focal plane array or a staring array. The system requires one (tiny) light sensor for each pixel and there are no moving parts, except for those required to move the lens back and forth for focusing and to open and close the iris to allow for different light levels. The difficulty with these systems is that a separate sensor is used for each pixel - it is difficult to make 440,000 identical sensors for a complete television picture. The naturally high contrast in a typical scene that is illuminated with visible light helps to mask small differences between the individual pixel-sensors of the detector. Thermal images have low contrast and small variations in the


50

sensitivity of individual pixel detectors make the picture look ‘grainy’.

(ad) Some thermal imaging systems are now being developed with cooled focal plane arrays and these are not much bigger than modern camcorders. However, the lower energy levels of infra-red radiation often lead to the requirement for large aperture lenses (to collect more radiation) and these thermal cameras are often heavier and larger than ordinary camcorders as a result. Additional computer processing might be needed in these systems to compensate for variations in the sensitivity of individual pixel-detectors.

Current and Future State of the Art in Thermal Imaging Technology

64. TO BE ADDED


sec 2 - principle and operation of thermal imaging systems

Documents