This article was downloaded by: [Moskow State Univ Bibliote]

On: 30 August 2013, At: 19:58

Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,
37-41 Mortimer Street, London W1T 3JH, UK

Journal of Modern Optics

Publication details, including instructions for authors and subscription information:
http://www.tandfonline.com/loi/tmop20

History highlights and future trends of infrared sensors

a
Carlo Corsi
a
Centro Ricerche Elettro Ottiche (CREO), L'Aquila, Italy
Published online: 09 Apr 2010.

To cite this article: Carlo Corsi (2010) History highlights and future trends of infrared sensors, Journal of Modern Optics,
57:18, 1663-1686, DOI: 10.1080/09500341003693011

To link to this article: http://dx.doi.org/10.1080/09500341003693011

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained
in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the
Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and
are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and
should be independently verified with primary sources of information. Taylor and Francis shall not be liable for
any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever
or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of
the Content.

This article may be used for research, teaching, and private study purposes. Any substantial or systematic
reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any
form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://
www.tandfonline.com/page/terms-and-conditions
 Journal of Modern Optics
Vol. 57, No. 18, 20 October 2010, 1663–1686

TUTORIAL REVIEW
History highlights and future trends of infrared sensors
Carlo Corsi*
Centro Ricerche Elettro Ottiche (CREO), L’Aquila, Italy
(Received 7 December 2009; final version received 8 February 2010)

Infrared (IR) technologies (materials, devices and systems) represent an area of excellence in science and
technology and, even if they have been generally confined to a selected scientific community, they have achieved
technological and scientific highlights constituting ‘innovation drivers’ for neighbouring disciplines, especially in
the sensors field. The development of IR sensors, initially linked to astronomical observations, since World War
II and for many years has been fostered essentially by defence applications, particularly thermo-vision and, later
on, smart vision and detection, for surveillance and warning. Only in the last few decades, the impact of silicon
Downloaded by [Moskow State Univ Bibliote] at 19:58 30 August 2013

technology has changed the development of IR detectors dramatically, with the advent of integrated signal
read-outs and the opening of civilian markets (EO communications, biomedical, environmental, transport and
energy applications). The history of infrared sensors contains examples of real breakthroughs, particularly true in
the case of focal plane arrays that first appeared in the late 1970s, when the superiority of bi-dimensional arrays
for most applications pushed the development of technologies providing the highest number of pixels. An
impressive impulse was given to the development of FPA arrays by integration with charge coupled devices
(CCD), with strong competition from different technologies (high-efficiency photon sensors, Schottky diodes,
multi-quantum wells and, later on, room temperature microbolometers/cantilevers). This breakthrough allowed
the development of high performance IR systems of small size, light weight and low cost – and therefore suitable
for civil applications – thanks to the elimination of the mechanical scanning system and the progressive reduction
of cooling requirements (up to the advent of microbolometers, capable of working at room temperature). In
particular, the elimination of cryogenic cooling allowed the development and commercialisation of IR Smart
Sensors; strategic components for important areas like transport, environment, territory control and security.
Infrared history is showing oscillations and variations in raw materials, technology processes and in device design
and characteristics. Various technologies oscillating between the two main detection techniques (photon and
bolometer effects) have been developed and evaluated as the best ones, depending on the system use as well as
expectable performances. Analysis of the ‘waving change’ in the history of IR sensor technologies is given with
the fundamental theory of the various approaches. Highlights of the main historical IR developments and their
impact and use in civil and military applications is shown and correlated with the leading technology of silicon
microelectronics: scientific and economic comparisons are given and emerging technologies and forecasting of
future developments are outlined.
Keywords: infrared; smart sensors; thermovision; photon sensors; microbolometers

1. Introduction that have been hit by a photon with suffi-

Infrared (IR) detectors are devices transforming the cient energy.
radiant energy of the IR region (from 0.7 to 300 mm) of (b) Photoelectric internal effects, the most rele-
electro-magnetic (e.m.) spectrum incident on the sensor vant phenomena in sensor development, by
in another form of energy which can be easily generating a couple of hole–electron pairs
inside a photoelectric material; generally a
measurable, generally in the form of an electric signal
photoconductor or a photovoltaic structure.
[1–7].
This conversion of energy is normally done by: . Bolometric effects: effects based on variations of
vibration energy of a crystalline reticle, which
. Photoelectric effects:
are generally measured by the variation of
(a) Photoelectric external effects by generating electrical resistance. Recently advanced develop-
free electrons from the surface of the sensor ments of IR detectors have been achieved

*Email: corsi@romaricerche.it

ISSN 0950–0340 print/ISSN 1362–3044 online

ß 2010 Taylor & Francis
DOI: 10.1080/09500341003693011
http://www.informaworld.com
 1664 C. Corsi

(the so-called silicon microbolometers) and also bolometric sensors; the proportionality to Df is also
the pyroelectric detectors, based on the variation valid in a limited frequency variation.
of the dielectric constant and therefore on the Another parameter often used in thermal analysis
variation of the electrical charge in a capacitive is the noise equivalent temperature difference (NETD),
sensor structure. that is, the temperature difference generating an rms
signal equal to the detector noise.
The IR detectors using photoelectric effects are photon
detectors or quantum detectors and are measuring the
photons with a quantum energy higher than the
1.2. Theory of intrinsic photoconductivity
internal conduction energy gap, while the bolometric
or thermal detectors are measuring the average of The intrinsic photoconductive sensor is essentially a
incoming radiative energy, independently from the photoresistance, that is a resistance changing its value
spectral content (assuming a constant absorbing when it is radiated by some form of e.m. energy –
coefficient). normally light – as is shown in Figure 1. When the
energy of a photon of radiation light is higher than the
forbidden energy gap, Eg, in the photoconductive
Downloaded by [Moskow State Univ Bibliote] at 19:58 30 August 2013

semiconductor, a couple of electron–hole pairs are

1.1. Parameters characterising the IR detectors generated, generating therefore a voltage variation that
is measured as a signal output at the load resistance
The main parameters of IR detectors are: spectral
RL, as shown in Figure 1.
response, signal-to-noise ratio per incident unit power, The voltage responsivity is expressed by
response time and working temperature.
  Vb ðb þ 1Þ
. The spectral response R () in V W1 is RV ¼ , ð3Þ
Lwt hc bn0 þ p0
generally represented by a curve showing the
responsivity versus wavelength, where R is the where w, t, L are the dimensions of the photoconductor
rms (root mean square) of the electrical output (Figure 1) and where b ¼ ‘/h; ‘ is the electronic
voltage of the sensor per unit of radiating rms mobility, h is the hole mobility and Vb is the
power at the wavelength . polarisation voltage and the lifetime of carriers is
. The response time, which is the time needed for given by  ¼ Dnt/Fs.
the signal to achieve 70.7% of the equilibrium This means that the basic requirement for achieving
value. a high responsivity in a photoconductor at a certain
. The signal-to-noise per incident unit power in wavelength  requires a high quantum efficiency ()
watts, the so-called detectivity, is given by: and a long lifetime () of the excess carriers, and also a
small distance between the electrodes, a low concen-
S 1 tration of thermally generated carriers n0, p0 at
D¼ , ð1Þ
N ðE AÞ equilibrium, with the highest applicable polarisation
where S, N are, respectively, the rms of the signal voltage Vb.
voltage and of the noise voltage, E (irradiance) is the In reality this model ignores important effects
rms of the incident radiation, A is the sensitive area of deriving from a too short interelectrode distance or
the IR sensor. The detectivity is a most important by other physical–structural characteristics, which can
parameter for characterising IR detectors and gen- originate ‘draining’ effects or surface recombination.
erally is indicated as D(T, f, Df ), where T is the black
body temperature of the radiating source, f is the
frequency of modulation of the signal and Df is the 1.3. Photovoltaic detectors
normalised bandwidth of the signal output amplifier. The photovoltaic effect is present in sensor structures
with an internal voltage difference which is moving the
A normalised detectivity is given by:
photo-generated carriers (electrons and holes) in
S ðDf Þ1=2 opposite directions. The most common example of
D ¼ : ð2Þ the photovoltaic effect is based on an abrupt p–n
N ðEA1=2 Þ
junction in a semiconductor and is known as a
The assumptions made in the above formula are valid ‘photodiode’.
only if noise is proportional to A1/2 and to Df. The first The photons with energy higher than the
hypothesis is valid only for the photon detectors and energy-gap which are incident over the surface of the
for variation of the sensor area only of one to two sensor are creating a ‘electron–hole’ couple on both
orders of magnitude, and is not valid for some sides of the p–n junction that, thanks to the induced
 Journal of Modern Optics 1665
Downloaded by [Moskow State Univ Bibliote] at 19:58 30 August 2013

Figure 1. Photoconductor functional scheme [10].

field, diffuse within the diffusion path of the junction at polarisation voltage equal to zero
reaching the region of spatial carrier where the high  1
@I
electric field separates the electron–hole couples in R0 ¼ : ð6Þ
such a way that the minority carriers are accelerated, @V Vb ¼0
becoming majority carriers on the other side of the A figure of merit normally used is the value of the
electrical junction. R0A product
In this way the generated photocurrent, modifying  1
the current–voltage characteristics, is given by the @J
R0 A ¼ , ð7Þ
inverse negative current Iph, as is shown in Figure 2. @V Vb ¼0
where J ¼ I/A is the current density.
Iph ¼ qAF, ð4Þ
In normal detection the photodiodes are working at
where A is the photodiode area, F is the flux of zero bias, while the inverse polarisation is used for high
incident photons and  is the quantum efficiency. frequency applications for reducing the RC constant of
Normally the gain in current in a photovoltaic the device.
detector with a simple structure (e.g. not of avalanche
or tunnel type) is equal to 1.
1.3.1. Photodiode currents
When the p–n junction is working with an open
circuit, the accumulation of electron–hole couples in Various mechanisms are involved in the current
the junction carriers produces an open circuit voltage phenomena in photodiodes. The most important are:
which, in the case of a load resistance Rl connected to (a) Dark current mainly due to thermally gener-
the diode, generates a current which achieves its ated carriers in the crystal and in the depletion
highest value when the diode is short-circuited layer of the p–n junction, and surface currents
(short-circuit current Ish). due to surface states and surface leakages.
The open circuit voltage can be obtained by just (b) Diffusion current given by:
multiplying the Ish. per the incremental resistance
JD ¼ Js ½expðqV=kT Þ  1 ð8Þ
R ¼ (v/I ) when V ¼ Vb. That is
where Js is given by
Vph ¼ qAFR, ð5Þ "     #1
1=2 2 1=2 1 e 1=2 1 h 1=2
where Vb is the voltage polarisation and I ¼ f(V ) is the Js ¼ ðkT Þ ni q þ , ð9Þ
ppo e nno h
I–V curve. In many cases the photodiode is operating
 1666 C. Corsi
Downloaded by [Moskow State Univ Bibliote] at 19:58 30 August 2013

Figure 2. Photovoltaic functional scheme [10].

where ni is the concentration of intrinsic for linearly graded junctions, is showing the
carriers, po and no are the concentrations of proportional dependence for the g-r current.
the majority carriers and  e and  h are the Moreover, the g-r current is proportional to ni,
lifetime of electrons and holes respectively in while the diffusion current is proportional to
the p and n regions [8–10]. n2i . Therefore, there is a temperature Te where
The diffusion current is changing versus the two currents are comparable, while below
temperature as the square of intrinsic density Te the g-r current is dominant.
of electrons (n2i ) and is generally the current
Other current phenomena are related to tunnelling
dominating at high temperatures.
effects and more importantly, to surface leakages. The
(c) Generation-recombination current:
dark current taking care of all these phenomena is
Such a mechanism can be the dominant one at expressed by
low temperatures and is given by [8]:  
qðV  IRs Þ V  IRs
qni w 2 sinhðqV=2kT Þ I ¼ Is exp 1 þ þ IT , ð12Þ
JGR ¼ f ðbÞ, ð10Þ kT Rsh
ðeo ho Þ1=2 qðVbi  V Þ=kT
where Rs is the series resistance and Rsh is the shunt
where Vbi is the induced voltage and  eo and  ho resistance of the photodiode.
are the lifetimes of the electrical carriers in the If the diffusion current is dominating, the
depletion layer and f(b) is a complex function b coefficient is close to 1, but if the main carrier
which is normally close to 1. The generation- transport is due to g-r current, b is equal to a  2.
recombination (g-r) current can be simplified by
qwni
JGR ¼ , ð11Þ
2o 1.3.2. R0A product
which, taking into account that the width of the In the case of a classic diode, where d  Le
barrier is varying as the square root of the  1=2
applied voltage (w ffi V1/2) in the presence of ðKT Þ1=2 e
ðR0 AÞD ¼ 3=2 2 Na : ð13Þ
abrupt junctions or as the cubic root (w ffi V1/3) q ni e
 Journal of Modern Optics 1667

If the thickness of the crystal layer is smaller than the of the free electronic carriers due the lattice
diffusion length of the minority carriers, the R0A vibrations of the semiconductor crystal causing a
product is increased and in such a case, we obtain current fluctuation at the microscopic level.
The g-r noise, for an intrinsic photoconduc-
Vb 0
ðR0 AÞGR ¼ ð14Þ tor, is given by
qni w
   1=2
2Vb 1 þ b np 1=2 Df
with an increase in the product R0A equal to Le/d. Vgr ¼ : ð17Þ
ðLwtÞ1=2 bn þ p n þ p 1 þ !2  2
It is interesting to underline that the measure-
1.4. Noise mechanisms ments of the g-r noise is allowing one to obtain
All the IR sensors are limited in their detection the value of the lifetime  just by measuring with
performance capabilities by the various forms of a spectrum analyser the knee of the curve at
noise generated either from the sensor itself or from v ¼ 1/.
fluctuations in the radiation environment, either from . The so-called 1/f noise characterised by a spec-
the electronic amplifier for the signal read-out (with the trum where the noise power is inversely propor-
Downloaded by [Moskow State Univ Bibliote] at 19:58 30 August 2013

most recent very low-noise amplifiers this type of noise tional to the frequency f, according to
can be ignored – in comparison to the other two) [8,9].  
KIb Af 1=2
The noise due to the fluctuation of background I1=f ¼ , ð18Þ
radiation is given by f
ð where K is a proportionality factor, Ib is the
2p1=2 Vb 1 þ b 1 ðÞ2 expðh=kTb Þ dv
Vph ¼ current bias, is a constant almost equal to 2
ðlwÞ1=2 t bno þ p 0 c2 ½expðh=kTb Þ  12 and  is a constant almost equal to 1.
tðDf Þ1=2
1=2
, ð15Þ The 1/f noise is normally associated with the
ð1 þ ! 2  2 Þ presence of potential microbarriers at the boundaries
where Tb is the background temperature, and no is the of polycrystalline grains in the semiconductor and the
frequency corresponding to the cut-off wavelength 1/. reduction of 1/f noise is almost an art in which one has
The most important internal noise types are: to take care with the realisation of the electrical
contacts and the preparation of photosensitive
. The thermal noise (Johnson–Nyquist noise) surfaces.
associated with any device with a resistance ‘R’ Normally the IR photodetectors show a 1/f noise at
(pure capacitors and inductors don’t have this low frequencies, while at higher frequencies the
type of noise, although they can have other predominant noise is the g-r noise until the 1/
forms of noise, e.g. capacitive noise due to frequency where the Johnson noise starts to prevail
electronic switching). (Figure 3).
This type of noise is related with the random
thermal fluctuations of the electrical carriers
which are moving within the semiconductor (the
2. Complex devices: IR focal plane arrays (FPA)
total fluctuation of the carriers is generating the
other type of noise, the so-called g-r noise The most important application of IR detectors is in
described later on). thermovision which is the ability to see the thermal
The thermal noise is present in the absence of emission of the scene. Thermovision systems have
external biasing and generates a current fluctu- strongly evolved over time since the first thermovision
ation independently from the method of mea- systems developed during World War II, based on a
surement. simple opto-mechanical scanning focusing scene onto a
The root-mean square (rms) value is given by single detector [10].
Such type of image reconstruction, based on a
Vj ¼ ð4kTRDf Þ1=2 , ð16Þ serial scanning of the image points (pixels), has been
improved, mainly by enhancing the signal-to-noise
where k is the Boltzmann constant, T is the
ratio thanks to the increase in sensor numbers. This
temperature and Df is the frequency bandwidth.
was first achieved by using linear detector arrays, with
This type of noise has a flat distribution and is
a serial read-out integrated in time (time delay
therefore called ‘white noise’: integration [TDI], Figure 4(a)) or by using a parallel
. The generation recombination noise (g-r or structure to read simultaneously more rows of the
flicker noise) is due to the random fluctuations scanned image and then by integrating them in the
 1668 C. Corsi

Figure 3. Electronic noise versus frequency [10].

Downloaded by [Moskow State Univ Bibliote] at 19:58 30 August 2013

Figure 4. (a) Parallel and (b) serial scanning functional scheme [10].

image reconstruction (Figure 4(b)) (both applied espe- positioned in the focal plane of the image avoiding the
cially to satellite remote sensing). use of any opto-mechanical scanning, in a way similar
The importance of the development of FPA to human vision, is self-evident.
detectors, working in the so-called ‘staring’ mode, i.e. Various types of electronic read-out have been
capable of seeing the image by a simultaneous vision developed since the 1970s, starting with a pseudo-
of the scene thanks to a mosaic sensor structure bidimensional read-out based on the sequential
 Journal of Modern Optics 1669
Downloaded by [Moskow State Univ Bibliote] at 19:58 30 August 2013

Figure 5. (a) (x–y) addressing by CMOS switching. (b) Rows–columns (x–y) scanning. (c) Rows–columns (x–y) scanning.
(d ) Integrated (x–y) scanning [10].

read-out of rows and columns by using multiplexers originated by the need to define the differences more
and a shift-register (Figure 5(a)), passing then to the precisely than just the adjective hot/hotter and cold/
so-called X–Y addressing by using integrated read-out colder. Based on the ideas of Aristotle who had
devices (Figure 5(b)) or external addressing capable of defined four qualities: hot, cold, moist and dry. Galen
selecting specific areas of the mosaic sensors (130–200 AD) was the first man to describe heat and
(Figure 5(c)). cold by a number and the word ‘temperature’ was
The read-out was continuously evolving until 1970, originated from ‘temper’, determining the ‘complexion’
when the new charge coupled device (CCD) read-out, of a person by the proportion in which the above four
based on electronic charge transfer, allowed a com- qualities were ‘tempered’ [11,12].
pletely integrated read-out (Figure 5(d )), with the The instrument by which the temperature of bodies
development of FPA IR sensors with more than one is measured is called ‘ thermometer’ (warmth measurer)
million pixels of detectors. and the history of thermometry is a recent part of the
history of science. In fact, although the ancient Greeks
(Philo of Byzantium, second century BC, and Heron of
Alexandria, first century BC), knew of the principle of
3. Historical scenario air thermal expansion and Abû Alı̂ ibn Sı̂nâ (known as
Infrared as radiation theory and advanced technology Avicenna) in the early eleventh century already demon-
is new (only two centuries old), but the concept of heat strated that the expansion and contraction of air moved
and temperature measurements is quite old. In fact the the position of the water–air interface inside a tube with
ancient Egyptians moved their hands across the surface air inside and partially filled with water, it wasn’t until
of the human body to scan and monitor changes in the sixteen/seventeenth centuries that the first air
temperature. The concept of temperature was thermometers were developed.
 1670 C. Corsi
Downloaded by [Moskow State Univ Bibliote] at 19:58 30 August 2013

Figure 6. (a) Galileo thermometer. (b) Floating ampoules. (The colour version of this figure is included in the online version of
the journal.)

In 1603 Galileo Galilei built a thermoscope made of by the botanist Linnaeus: increasing heat was indicated
a glass cylinder filled with a liquid whose density by higher temperatures. In 1948 the name of the
sensibly expands with temperature (Figure 6). Inside ‘Celsius’ scale was given to the centigrade scale.
the glass cylinder there are small glass ampoules filled Another branch of thermometry is gas thermo-
with coloured liquids: such ampoules have different metry and thermodynamic thermometry. Before the
densities and they are labelled with a certain temper- Fahrenheit thermometer appeared, in 1660 Robert
ature value. When the device is in thermal equilibrium Boyle reported on his study of air trapped in a U tube
with the surrounding ambient, one can read the that he found that volume at constant pressure
temperature on the lowest floating ampoule [11–13]. was a function of temperature. While Fahrenheit
Robert Fludd (1574–1651) was also regarded as an studied his liquid-in-glass thermometer, the French
independent inventor of the thermometer, modifying scientist Amontons developed the constant volume gas
Philo’s apparatus that demonstrated the expansion of thermometer. He used air as the thermometric
air by heat. One limit of the air thermometer was that it medium, and concluded that the lowest temperature
also responded to the change in atmospheric pressure, which could exist would correspond to a zero gas
even if originally there was no concept of atmospheric pressure. This must have been the first step on the way
pressure. The sealed liquid-in-glass thermometer to understanding the concept of temperature.
invented by Ferdinand II, Grand Duke of Tuscany, According to Amontons we could define the temper-
in 1654 was to overcome two other problems, the lack ature as being simply proportional to the pressure of a
of portability and the evaporation of water, just by gas, and thus we would need only one fixed point to
using the thermal dilation of a liquid instead of air to define a scale [14]. But the new temperature scale didn’t
sense temperature changes (the first liquid in glass was appear at that time, maybe because of the cumbersome
spirits of Chianti wine). In 1663 the Members of the operation of the gas thermometer. Jacques-Alexandre
Royal Society of London agreed to use one of several Charles studied the phenomena again in 1787.
thermometers made by Robert Hooke as the standard Jaspeh L. Gay-Lussac extended Charles’ work, all his
for the calibration of other thermometers. gases – air, oxygen, nitrogen, hydrogen, and carbon
In 1714 Fahrenheit improved the precision of the dioxide – expanded the same amount when heated
thermometer using a cylinder rather than a globe as the from the ice point to the boiling point. The result
bulb of the thermometer and substituting alcohol showed that the mean volumetric coefficient of thermal
spirits with mercury, with a more linear thermal expansion, at constant pressure, is equal to 1/267
expansion, and fixing the lower temperature point by (degree)–1. In 1847, Victor Regnault obtained a better
using salt with ice water and the higher point to boiling value of 1/273. Later experiments revealed that all
water at 212 degrees. Celsius, in 1742, created a gases had very slightly different thermal coefficients,
decimal scale, using ‘zero’ as the boiling point of water however, all were found to approach a common
and 100 C as the freezing point. His scale was reversed value 1/273.15 C as the pressure approached zero.
 Journal of Modern Optics 1671

The gases that obey the temperature–volume relation- Convention du Metre, the founding treaty for the SI,
ship exactly at constant pressure (and the pressure– and was revised with the current version agreed to in
volume relationship exactly at constant volume) were 1990, and known as ITS90. The ITS are empirical
defined the perfect (‘ideal’ or ‘simple’) gas. In fact, all temperature scales giving a close approximation to the
gases at extremely low pressure approach perfect gas known thermodynamic scale, but are more precise and
behaviour: the volume of the perfect gas in the gas easier to use [65].
thermometer will approach zero as t ¼ 273.15. We The history of IR radiation sensors starts in 1800
can establish a new zero of temperature at the ice point when the astronomer, William Herschel, discovered the
273.15, namely the absolute zero degree, and the existence of infrared radiation, a form of energy in the
perfect gas temperature scale T ¼ t þ 273.16 was light beyond the ‘red’ (from the Latin ‘infra’ – below),
defined in 1954 [15]. by trying to measure the heat of the separate colours of
The precise meaning of temperature came after the rainbow spectrum cast on a table in a darkened
Kelvin’s thermodynamic temperature appeared. In room. After noticing that the temperature of the
1824, Carnot had laid down his principle of reversibil- colours increased from the violet to the red part of the
ity with regard to ‘quantity of heat’ in an ideal ‘body’ spectrum, Herschel decided to measure the tempera-
that ‘undergoes changes’ and in 1845 (after 30 years of ture just beyond the red portion of the spectrum in a
Downloaded by [Moskow State Univ Bibliote] at 19:58 30 August 2013

experiments) Joule demonstrated mechanical equiva- region apparently devoid of sunlight. To his surprise,
lence. Heat now became an engineering measurable he found that this region had the highest temperature
‘quantity’ and the higher the temperature the greater of all. In April 1800 he reported it to the Royal Society
the quantity of heat in Carnot’s ‘body’. A real relation as Dark Heat and, making further experiments on
of temperature to heat was achieved by the first and what he called the ‘calorific rays’ that existed beyond
second laws of thermodynamics: it is a standing proof the red part of the spectrum, he found that they were
that for a reversible heat engine operating over a reflected, refracted, absorbed and transmitted just like
Carnot cycle between two temperatures and the ratio visible light [16,17].
of the heat taken in at the higher temperature to that The basic laws of IR radiation (Kirchhoff’s law,
given out at the lower temperature is proportional Stefan–Boltzmann’s law, Planck’s law and Wien’s
simply to the ratio of the same function of each of displacement law) were developed many years after
the two temperatures. William Thomson (later Lord the discovery of IR radiation. In 1859 Gustave
Kelvin) realised, in 1848, that this relation could be Kirchhoff found that a material that is a good
used to define the ratio of any two temperatures. absorber of radiation is also a good radiator.
The values of the temperatures would depend upon Kirchhoff’s law states that the ratio of radiated
the functional form, but by taking the simplest power and the absorption coefficient: (1) is the same
possible form of the function he defined a temperature, for all radiators at that temperature, (2) is dependent
which he called thermodynamic temperature, T. on wavelength and temperature, and (3) is independent
Thermodynamic temperature thus has the property of the shape or material of the radiator. If a body
that ratios of T are defined in terms of the properties of absorbs all radiation falling upon it, it is said to be
reversible heat engines and are independent of the ‘black’. For a black body the radiated power is equal to
working substance. The definition of the quantity the absorbed power and the emissivity (ratio of emitted
thermodynamic temperature then has to be completed power to absorbed power) equals one. In 1884, L.E.
by assigning a particular numerical value to an Boltzmann, starting from the physical principles of
arbitrary fixed point of temperature, triple point of thermodynamics, derived the theoretical formula for
water. So the real meaning of the temperature is a the T4 black body radiation law, stated empirically
physical quantity which is fundamental for the field of in 1879 by J. Stefan, by developing the Stefan–
thermodynamics and is directly related to the basic Boltzmann’s law (19):
laws of thermodynamics [14,15]. In principle, any
W ¼ T 4, ð19Þ
suitable thermodynamic interaction may be used as the
basis for a thermometer, such as the acoustic ther- where W ¼ radiation power, T ¼ absolute temperature,
mometer, the thermal noise thermometer, the gas and ¼ Stefan–Boltzmann’s constant.
thermometer, the radiation thermometer, etc. In 1901, Nobel Prize laureate Max Karl Ernst
However, thermodynamic thermometers cannot Ludwig Planck developed Planck’s law which stated
achieve the desired precision, and are complex and that the radiation from a black body at a specific
time consuming to use. To overcome these difficulties, wavelength can be calculated from
the above empirical temperature scale was defined
as the International Temperature Scale (ITS) by the 2h 1
IðÞ  ¼  ð20Þ
Comité des Poids et Measures (CIPM) under the c2 expðh=kT Þ  1
 1672 C. Corsi

(where I() is the radiation power emitted per unit of (PbSe), lead telluride (PbTe), and indium antimonide
surface and solid angle unit, in the frequency interval (InSb) cooled detectors extended the spectral range
(   þ ); T ¼ absolute temperature; c ¼ speed of beyond that of PbS, providing sensitivity in the 3–5 mm
light; h ¼ Plank’s constant). medium wavelength (MWIR) atmospheric window
Soon after Wilhelm Wien (Nobel prize 1911) (extrinsic photoconductive germanium detectors were
established Wien’s displacement law taking the deriv- allowing one to reach the long wavelength spectral
ative of Plank’s law equation to find the wavelength for region, needing very low temperature with the use of
maximum spectral radiance at any given temperature: liquid helium). But we had to wait until the 1960s to see
the first advanced developments coming out thanks to
M T ¼ 2897:8 ðmmKÞ: ð21Þ direct gap photon materials based on ternary semicon-
IR detector developments, even after the discovery of ductor compounds (HgCdTe and PbSnTe) [18]. This
infrared radiation by Sir H. Herschel in 1798, were was a real breakthrough, because in the meantime
mainly based on the use of thermometers which microelectronics was offering new advanced manufac-
dominated IR applications until World War I, turing technologies like photomasking and integrated
although in 1821 J.T. Seebeck had discovered the microsoldering and assembly. So, thanks to these
advanced technologies, the first linear arrays of tens
Downloaded by [Moskow State Univ Bibliote] at 19:58 30 August 2013

thermoelectric effect and in 1829 L. Nobili had

of elements were developed at the end of the 1960s with
fabricated the first thermocouple, allowing in 1833
strong competition between HgCdTe and PbSnTe (this
the multi-element thermopile development by
compound could offer more stability and reliability in
Macedonio Melloni with the first detection of a
performance, moreover at the beginning of the 1970s
human being at 10 m. Early thermal detectors,
Lincoln Labs, MIT researchers were developing the
mainly thermocouples and bolometers, were sensitive
first solid state IR 10 mm lasers applied to environ-
to all infrared wavelengths and operating at room
mental control) [19]. On the other hand, PbSnTe was
temperature normally, and until a few years ago, they
showing higher dielectric constant, limiting high fre-
were of relatively low sensitivity and slow
quency performances, and a high thermal expansion
response time.
coefficient, with a strong limitation for integration into
The first photon detectors (based on the photo-
silicon microelectronics. (The speed limit could be
conductive effect discovered by Smith in 1873 in overcome if forecasting the FPA array developments
selenium and, later on, by Bose in photovoltaic lead which could allow the use of slower detectors.)
sulphide, but not applied for many years) were The choice, supported by US industries, of
developed by Case in 1917. In 1933 Kutzscher concentrating efforts and resources in a specific tech-
developed IR PbS detectors (using natural galena nology produced the ‘first generation of linear detector
found in Sardinia); these sensors were widely used arrays’, which allowed one to obtain BLIP detectors at
during War World II. These detectors have been liquid nitrogen temperature (this first generation of
extensively developed since the 1940s. Lead sulfide CMT linear arrays was the basis for the ‘common
(PbS) was the first practical IR detector, sensitive to modules’ LWIR FLIR systems with a number of pixels
infrared wavelengths up to 3 mm. In the meantime from 60 up to 180, each detector connected with
Cashman developed TaS, PbSe and PbTe IR detectors feedthroughs to the room temperature read-out
with high performances supporting developments in electronics).
the UK and US. The invention of charge coupled devices (CCDs)
High level results were achieved in the 1940s, in 1969 [20] made it possible to start developing the
especially in lead salts (good PbSe and PbTe stable ‘second generation’ FPA detector arrays coupled with
cells were developed by the Office of Scientific Research on-focal-plane electronic analogue signal readouts
and Development [OSRD] in co-operation with MIT, which could multiplex the signal from a very large
Harvard and the British Telecommunication Research array of detectors. In the mid-1970s, while the first
Establishment [TRE], Great Malvern). The history of common module IR arrays were produced, the first
IR detector developments therefore has been almost CCD IR bidimensional arrays [21,22] were appearing
coincident with optoelectronics for military applica- in the USA and, the first Smart Sensors based on LTT
tions for many decades, strongly conditioning the RF sputtered thin films, using X-addressing read-out,
cultural behaviour of the IR industry and in some way were developed in Italy [23].
of R&D labs. In 1975 the first CCD TV camera was realised and
Thanks to the discovery of transistors in 1948, this allowed the forecast of the ‘second generation
infrared technology during the 1950s was enjoying a FPAs’ capable of staring vision, although the necessity
great growth, especially in the development of solid for very high spatial resolution and high reliability
state IR sensors. Almost contemporarily, lead selenide even in complex structures, with an extremely high
 Journal of Modern Optics 1673

number of pixels (up to one million pixels), were impressive growth since the first developments, allow-
pushing towards alternative solutions, with materials ing the real expectation for production of low cost,
less difficult than CMT, in the manufacturing process high performance detector arrays which finally should
(e.g. extrinsic silicon detectors). The high quantum follow the rules of a real global market, opening a real
yield of CMT and the top performances required market for civil applications following the winning
by the military allowed one to improve the perfor- rules of silicon microelectronics. For these reasons the
mances of sensor linear arrays by integrating time emerging room temperature detectors in the 1970s by
delay and integration inside the detector structure itself the use of pyroelectric materials [28], which shows the
(SPRITE detector [24]). From the late 1970s to the limitations of not being fully monolithic, but the more
1980s, CMT technology efforts focused almost exclu- innovative room temperature silicon microbolometers
sively on PV device development because of the need [29] appearing on the IR scene in the 1990s, appear to
for low power and high impedance for interfacing to be a real breakthrough for future IR sensors.
readout input circuits in large arrays (photoconductive Table 1 reports a schematic synthesis of the most
CMT was not suitable due to its low impedance). This important developments in thermal sensing and
effort was concretised in the 1990s with the birth of Table 2 reports the highlights in thermal sensor and
‘second generation IR detectors’ which provided large IR sensor developments since Herschel’s discovery.
Downloaded by [Moskow State Univ Bibliote] at 19:58 30 August 2013

2D arrays with the number of pixels up to many

hundreds of thousands thanks to hybrid integration
(indium bumps or loophole soldering) of CMT 4. State of the art
bidimensional arrays in silicon substrate with CCD Electro-optics (EO) technology is continuously grow-
and more recently CMOS read-out. At the same time, ing in complexity and difficulty in science and in
other significant detector technology developments technology development: in particular, IR sensors
were taking place. Silicon technology generated novel might be considered at the frontier of solid state
platinum silicide (PtSi) detector devices which have technology [17–26].
become standard commercial products for a variety of Electro-optics recently has shown an impressive
MWIR high resolution applications. Monolithic growth in performance thanks to the reliability and
extrinsic silicon detectors were demonstrated first in cost achievable by the integration with advanced
the mid-1970s [25,26]. Thanks to PtSi Schottky barrier silicon microcircuit technology.
IR properties, great attention was dedicated to FPA IR detectors can be structured into two main
arrays based on integrated silicon Schottky sensors classes: cooled (photon detectors) and uncooled
which were showing reliable monolithic silicon CMOS detectors (thermal or bolometer detectors) and each
integrated technology and high uniformity in detectiv- divide further into two classes, hybrid and monolithic.
ity, but were operating in the short wavelength region Infrared history is showing oscillations and variations
and with the limitation of low working temperatures. in raw materials, technology processes and in device
Similar considerations can be made for the long design and characteristics. In the IR sensor technolo-
wavelength GaAs/GaAlAs multiquantum well IR gies we can perceive a general oscillation between the
FPA arrays [27], which, although if with lower two main detection techniques (photon and bolometer
quantum efficiency, are close to CMT performances effects) evaluated as the best ones depending, initially,
even showing higher homogeneity and stability in on the single sensor performances and later, since the
sensitivity thanks to a more reliable manufacturing 1980s, on system use, especially after the development
process, but with the strong limitation of working at of the CCD/CMOS bidimensional FPA (Figure 7).
lower temperatures (577 K). This requires the use of
cryogenic structures with the associated high cost of
purchasing and maintenance; therefore, improving the
4.1. Photon cooled detectors
restriction of the main use to military applications,
limiting the market size and, as a consequence, product 4.1.1. Hybrid CMT detectors
growth. In the form of photoconductive (PC) detectors, the
In all the latest developments the real driving key hybrid photon detectors have been the most fully
technology has been the integration of IR technology developed detectors which utilise the photon-induced
with silicon microelectronics and more and more electronic change in bulk conductance as a detection
emergence of the importance freeing IR from the con- mechanism. PC detectors must be biased with an
straints of the cooling requirements due to its high cost external voltage. This technology has been used
(almost a 1/3 of the total cost) and low reliability and primarily in first generation scanned sensors. It was
heavy need for maintenance. For the above reasons, not a viable option for second generation focal plane
work on uncooled infrared detectors has shown arrays because of their difficulties to be integrated in a
 1674 C. Corsi

Table 1. History of thermal sensors. Table 2. History of IR detectors.

3000 BC Egyptian medical wizards 1800 IR radiation Sir W. Herschel

400 BC Hippocrates 1821 Thermoelectric effect Seebeck
340 BC Aristotle 1829 Thermocouple G. Nobili
II Century BC Galeno 1833 Thermopile Macedonia Melloni
II–I Century BC Philo of Byzantium – Heron of 1836 Optical pyrometer Becquerel
Alexandria 1873 Photodetection Smith
1605 Galileo (selenium)
1654 Ferdinand II Duke of Tuscany 1884 IR radiation law Boltzmann
1724 Fahrenheit 1902 Photoconductivity effect Bose
1742 Celsius/Linneus (0–100) 1917 Lead sulphide Case
1868 C. Wunderlich (thermometry in 1933 Lead sulphide (galena) Kutzsher
medicine) 1940 TI2S Cashman
1877 Lehman (liquid crystal thermometer) 1942 Golay cell Golay – Queen Mary
1660 Boyle/Amonton (gas constant volume) College
1787 J.L. Gay-Lussac 1/267 (degree)1 1948 transistor Bardeen–Brattain–
1847 Regnault 1/273 (degree)1 Shockley
1824 Carnot (ciclo) 1950s PbS, PbSe, PbTe T. Moss RRSE
Downloaded by [Moskow State Univ Bibliote] at 19:58 30 August 2013

1845 J.P. Joule (heat/work) 1959 HgCdTe W. Lawson, J. Putley

1860 W. Thomson (Lord Kelvin) 1960s Ge:X, InSb
273.15 abs e scale 1969 CCD Boyle-Smith (Bell Labs)
1990 ITS (International Temperature Scale) 1970s PbSnTe/HgCdTe,Si:X Lincoln Labs, SBRC
by CIPM (Comité des Poids et Hughes, Honeywell,
Measures) Rockwell, Mullard
1973 Common modules Night vision Lab
1975 IR smart sensors C. Corsi, Elettronica SpA
1978 Si:X/CCD/PtSi/CCD RCA Princeton Lab W.F.
HgCdTe/CCD Kosonocky,
complex structure; moreover, the power dissipation
F. Shepherd D.
due to the continuous bias voltages produces consid- Barbee–F. Milton–J.
erable cooling power requirements. Photovoltaic (PV) Steckel
detectors do not require a bias voltage and can hence 1980s HgCdTe SPRITE T. Elliott RSE
be operated in the low power dissipation mode InGaAs QWIP F. Capasso, L. Esaki,
B.F. Levine,
required for large second generation infrared arrays. M. Razeghi,
PV technology is actually the primary detector tech- L.J. Kozlowski
nology being developed for large array applications; 1990s Pyroelectric FPAs RRSE-BAE R.A. Wood
although PV detectors have a theoretical photo-current Bolometer FPAs (Honeywell) J.L. Tissot,
which is twice that of PC devices, because both holes Multi-colour FPAs P.R. Norton,
Advanced FPAs A. Rogalski, H. Zogg
and electrons contribute due to the charge separation S.D. Gunapala,
caused by the internal junction field. Unfortunately, D.Z. Ting
photovoltaic technology is intrinsically more non- (Jet Propulsion Labs)
uniform than PC detectors because of the variations 2000 MEMS FPAs – cantile- B. Coole, S.R. Hunter,
ver IR nanotubes/ X. Zhang J.M. Xu,
due to intrinsic junction formation and consequent nanowires S. Huang, Y. Zhao,
large variation in diode impedance. PV detectors are J. Xu, Maurer,
normally operated at a slightly negative bias voltage in G. Jiang, D.J. Zook
scanned systems which are not affected by low
frequency 1/f noise; staring PV detector arrays are
normally operated at zero bias in order to minimise the have been developed by integrating CMT on a silicon
1/f noise which is present in the low frequency read-out substrate by MBE GaAs buffer layers for matching the
noise bandwidth. HgCdTe is the most attractive option thermal expansion coefficient. The major disadvantage
as a photon material because of the availability of high of HgCdTe is its non-uniformity. Each of the
quantum efficiency PV detector arrays with the above-described materials is normally associated with
appropriate cut-off wavelength; moreover, HgCdTe a particular multiplexer interconnection. HgCdTe
diodes have quantum efficiencies in the 60%–70% focal plane arrays are hybridly integrated with
range and array non-uniformities in the 10%–30% silicon-based readout electronics. The detector readout
range. Current PV arrays are fabricated using liquid connection can utilise either indium solder bumps, in a
phase epitaxy. MOCVD growth technology is expected flip-chip configuration, or over the edge connections, if
to improve the yield and non-uniformity of these the HgCdTe strips are bonded directly to the silicon
arrays. FPA HgCdTe detectors with a million pixels readout chip.
 Journal of Modern Optics 1675
Downloaded by [Moskow State Univ Bibliote] at 19:58 30 August 2013

Figure 7. Historical waving in IR sensor technologies [66].

4.1.2. InSb/LTT detectors format up 1024 1024 have been fabricated. The
The importance of integrating infrared detectors within cryogenically cooled InSb and HgCdTe arrays have
a read-out electronics was the main reason for pushing comparable array size and pixel yield at the MWIR
the development of lead salts (PbS, PbTe, PbSnTe and spectral band. However, the shorter response time and
PbSnSe) in the form of layers deposited on silicon high quantum efficiency have made HgCdTe the
substrate by the use of buffer layers of CaF2 and BaF2 preferred material, although if crystal growth and
or by a special thin film technology. It has to be process fabrication need higher complexity, it shows
underlined that lead salts are excellent IR detectors lower reliability.
either in the form of polycrystalline layers as well as
single crystals and are particularly suitable for complex
pattern shapes. Moreover, well acquainted technolo- 4.2. Monolithic photon detectors
gies such as MBE are particularly suitable for lead salt 4.2.1. Silicon Schottky diodes
growth. A well-established material technology is The main developments are based on Schottky barrier
based on using InSb with a CID readout structure. silicon (PtSi) detectors, with a rapid growth in recent
At 77 K working temperature, InSb cut-off wavelength years in the USA and Japan which allowed achieve-
is near 5 mm. Array non-uniformities in the range of ment of the highest number of pixels in FPA (4106
10% have been achieved; this technology has the pixels). PtSi has a relatively low cut-off coupled to the
limitation of short cut-off wavelength. InSb material is need for high cooling; the only operational monolithic
more mature than HgCdTe and good quality few inch development working at 77 K is the PtSi arrays in the
diameter bulk substrates are commercially available: 3–5 mm range, but it has better uniformity than other
InSb devices are usually made with both p-n junction detector technologies and can be integrated with silicon
and MIS capacitors. Fabrication techniques for InSb CCDs in monolithic structures. The negative charge of
photodiodes use gaseous diffusion, and a subsequent silicide is transferred to a CCD by the direct charge
etch results in a p-type mesa on an n-type substrate. To injection method. The effective quantum efficiency in
maximise the available resolution and response of the 3–5 mm atmospheric window is very low, of the
photodiodes, the bulk material is thinned to about order of a few per cent, but useful sensitivity is
10 mm. Backside illuminated staring arrays with a obtained by means of near full frame integration in
 1676 C. Corsi

area arrays. Moreover, for relatively long integration to an underlying stage of a silicon microcircuit. The
time (410 ms), PtSi responsivity can be adequate for performance is nowadays of acceptable level depending
detection and hence PtSi appears to be one of the more on the quality of the thermal insulation and therefore is
promising technologies. Schottky photoemission is normally operating in a vacuum encapsulation and
independent of such factors as semiconductor doping, thermoelectric thermal regulation. The imaging sys-
minority carrier lifetime, and alloy composition, and as tems, based on pyroelectric arrays, usually need a
a result of this, has spatial uniformity characteristics reference obtained by optical modulators which chop
that are far superior to those of other detector the incoming radiation. Barium strontium titanate
technologies. Uniformity is only limited by the geo- (BST) ceramic is a reliable material with a very high
metric definition of the detectors. At present FPAs up permittivity and usually has improved the performance
to millions of pixels have been developed, but slow of pyroelectric FPAs using a bias voltage applied to
progress in the future is expected for this technology. optimise the pyroelectric effect near the phase transi-
tion at the Curie temperature.
Although many applications for this hybrid array
4.2.2. Multi-quantum well detectors technology have been identified, and the imagers
A new class of photon detectors has been developed by employing these arrays have been in mass production
Downloaded by [Moskow State Univ Bibliote] at 19:58 30 August 2013

creating multilayer structures which allow a quantum for some years, no hybrid technology advances are
well superlattice [27]. These structures are based on two foreseen. The reason is that up to now the thermal
mono-atomic layers of different semiconductor materi- conductance of the bump bonds is limiting the NETD
als (GaAs and GaAlAs) with the creation of modulated ( f/1 optics) to about 50 mK. Pyroelectric array tech-
electronic bands (normally referred to as NIPI struc- nology therefore is moving toward monolithic silicon
tures). Detectors with D41010 cm Hz1/2 W1 have microstructure technology with a less complex and
been obtained at working temperatures 70 K. The more reliable process with the problem that most
possibility of this technology to integrate GaAs elec- ferroelectrics lose their properties as the thickness is
tronic devices can foreseeably allow important future reduced, even if lead titanate (PbTiO3) and related
developments although limitations in operative uses are materials hold their properties well, even in
expected due to the low working temperature needed for thin film form. Various techniques for the deposition
limiting the band intermodulation noise. Among the of thin ferroelectric films have been investigated,
different types of quantum well infrared photodetectors including radio frequency magnetron sputtering, dual
(QWIPs), the GaAs/AIGaAs technology is the most ion beam sputtering, sol-gel processing, and laser
mature with a number of military and commercial ablation.
applications. QWIP cannot compete with the HgCdTe
photodiode as a single device, especially at higher
4.3.2. Monolithic: silicon microbolometers
temperature operation (477 K) due to fundamental
limitations associated with inter-sub-band transitions. This completely new approach based on microsystems
Even though QWIP is a photoconductor, several of technology (MST) is showing excellent performances
its properties, such as high impedance, fast response and seems to offer very promising future developments
time, long integration time, and low power consump- [29]. Monolithic silicon FPAs have been fabricated by
tion, comply well with the requirements of large realising silicon-based bolometers in microminiature
FPA fabrication. At low temperature, QWIP has form by etching a thin bridge bolometer layer con-
potential advantages over HgCdTe for VLWIR FPA nected by two thin legs to the underlying silicon
applications in terms of array size, uniformity, yield substrate [30]. The whole bi-dimensional array with a
and cost. high number of pixels (4105), with a 50 mm pixel, is
maintained under vacuum and thermally regulated by
a single thermo-element cooler in order to obtain
outstanding performances (NETD550 mK). This
4.3. Room temperature thermal detectors
technology is moreover showing the advance, being
4.3.1. Hybrid: ferroelectric/pyroelectric detectors integratable with silicon technology, of having the
The main technology has been based on ferroelectric detector read-out circuits enabling a maximum in the
materials which, thanks to the dielectric constant filling factor. The most popular thermistor material
dependence versus temperature, can generate a voltage used in fabrication of the micromachined silicon
variation with capacitive structure detection and bolometers is vanadium dioxide, VO2. From the
therefore a quick response, although measurable as a point of view of an IR imaging application, probably
thermal variation [28]. These detectors are mainly the most important property of VO2 is its high negative
coupled in a hybrid model (normally by InSn bumps) temperature coefficient of resistance (TCR) at ambient
 Journal of Modern Optics 1677

temperature, which exceeds 4% per degree for a single sensors which could avoid cryogenic needs. After a
element bolometer and about 2% for a FPA, with serious, but partially successful effort using the pyro-
monolithic readout circuits integrated into underlying electric effect, limited by low sensitivity and chopper
silicon. At present, several research programmes are requirement, the new microbolometer technology,
focused towards larger array size and enhancement of because of the microsize of the thin film bolometers,
performances with D4109 cm Hz1/2 W1 and some of completely integratable with silicon technology and
them are based on an all-silicon version of the therefore often named silicon microbolometers, have
microbolometer. It is anticipated that some new emerged in the last few years with a highly promising
materials will form the basis of the next generation of future for IR sensor market growth.
semiconductor film bolometers, although the most So, for the first time, thanks to the elimination of
promising material appears to be amorphous silicon cryogenic cooling, wide use of IR Smart Sensors are
which is completely integratable with IC manufactur- emerging on the international market, becoming stra-
ing processes. This new original solution showing tegic components for the most important areas, like
acceptable performances coupled to a low cost transport (especially cars, aircrafts and helicopters),
approach (full silicon technology integrability and environment and territory control, biomedicine and
room temperature operation), can create a real break- helping towards ‘human beings having a better life’
Downloaded by [Moskow State Univ Bibliote] at 19:58 30 August 2013

through in infrared sensor technology. (intelligent buildings, energetic control, thermo-

mechanical structuring, auxiliaries for disabled people,
etc.).
The main efforts nowadays in IR detector devel-
5. Future infrared detectors
opments are oriented towards the focal plane array
Electro-optics (EO) is continuously growing in com- (FPA) structure with the highest number of pixels
plexity and difficulty in science and technology (more than 106 element), with integrated electronics for
developments: in particular, IR sensors might be con- signal read-out and elaboration, with working tem-
sidered at the frontier of solid state technology [30–39]. peratures close to room temperature and with high
Electro-optics has shown winning performances uniformity. Bi-dimensional FPAs coupled to X–Y
thanks to the reliability and cost achievable by readout integrated microelectronics (CCD and
integration with advanced silicon microcircuits. The CMOS devices) are allowing great improvements in
exploding market of optical telecommunications has, performances, but also in reliability and cost, improved
in some way, concentrated the main interest on the by the fact that, thanks to the increased integration
near infrared bands, pushing the infrared technology time possible in staring arrays, NETD values close to
(especially the long wavelengths) towards thermovision the BLIP limit can be more easily achieved. This, with
and surveillance applications. This gave an impressive the main task of achieving high optoelectronic FPA
impulse to the development of FPA arrays with ever performances with smaller and lighter structures, gives
growing performances and tasks for these advanced possibilities for applications in civil areas thanks to
high-tech components and systems. The main effort cost reduction by eliminating opto-mechanical scan-
nowadays in IR detector developments is oriented ning and cryogenic cooling. The possibility of an
towards the highest number of pixels (of the order of advanced IR FPA (with 4106 pixels in 8–12 mm)
106) with a FPA structure, with integrated electronics working at room temperature, allows one to forecast
for signal read-out and elaboration, and with the an incredible growth of uses and applications besides
highest possible working temperature. Key parameters the evident growth of advanced applications of IR
of a single pixel sensor, such as ultimate sensitivity science based on thermal measurements.
(measured by NETD), response time and working There is a large amount research activity directed
temperature, are integrated more and more by key towards 2D staring array detectors consisting of more
parameters of FPAs as the number of pixels, unifor- than 106 elements. IR FPAs have a similar growth rate
mity, reliability and cost. All these requirements will be as dynamic random access memory (RAM) integrated
strongly conditioned by the complete integrability with circuits (ICs) [40,41] (it is a consequence of Moore’s
silicon microelectronics technologies. Law [42], which predicts the ability to double transistor
The competition among various technologies and integration on each IC about every 18 months), but
‘technical schools’ was strong with unforeseeable new with a lag of about 10 years and with some breakpoints
actors emerging in the last years (overall, room (linear arrays and FPAs) and with a saturation trend
temperature microbolometers for future and extremely due to the physical constraints of the diffraction law, as
valid applications in the civil field). Operational explained soon after. Figure 8 illustrates the trend in
requirements (mainly of maintenance and reliability) array sizes for various technologies over the past 40
were pushing IR science to look for new advanced years and some forecasting for the future. (Actually the
 1678 C. Corsi

largest HgCdTe FPA is a short wavelength IR (SWIR) applications even hyperspectral resolving detection are
hybrid 2048 2048 with a unit cell size of required) [66]. This can allow one to foresee either high
18 mm 18 mm for astronomy and low background sensitivity tunable bandgap photon detectors or wide-
applications [43].) band microbolometers with integrated coating multi-
The pixel size is conditioning the achievable spectral filters. The cost of an FPA depends on the type
number of pixels achievable in a feasible FPA chip and maturity of the technology with different reduc-
size conditioned by the size and the cost of the optics, tion factors in the case of mass quantity production
and overall by the fundamental limit on the pixel size (again production based on silicon technology is more
determined by the diffraction law [34,35]. In fact the competitive) [37,67]. The cost of an IR sensors array
size of the diffraction-limited optical spot, or Airy disk, can be shared in the following main blocks: (a) IR FPA
is given by: physical detectors, including read-out electronics, (b)
assembly and testing, (c) optics with mechanical
d ¼ 2:44F, ð22Þ
structure, and (d) cryogenics in the case of cooled
where d is the diameter of the spot,  is the wavelength, detectors. Up to a few years ago there was the thumb’s
and F is the f number of the focusing lens (for high evaluation of almost 25% for each part, although for
Downloaded by [Moskow State Univ Bibliote] at 19:58 30 August 2013

luminosity F/1 optics at 10 mm wavelength, the diffrac- high performance photon-cooled hybrid detectors,
tion limited spot size is 25 mm). more than 50% was due to the sensor chip and a
Therefore, in the future a reduction for pixel size significant cost improvement in the cryogenic cooler
mainly thanks to oversampling (up to a factor of 4) was to be taken into account in the case of working
could be achieved for applications requiring high temperatures, lower than 77 K. This explains why PtSi
spatial resolution: the LWIR pixels could reach a and QWIP detectors requiring, respectively, wider
limit of 5 mm and SWIR pixel sizes could shrink to luminosity optics and lower operating temperatures,
correspondingly smaller dimensions. Anyway the pixel have a cost comparable to other photon detectors, even
size, unlike the size of a DRAM memory cell, cannot though the raw material (Si or GaAs) and fabrication
be significantly reduced and therefore to increase the technologies are much cheaper than for HgCdTe.
number of pixels in IR FPAs, the chip size should grow IR imagers mainly for military applications, actually
with a high cost and limited handling (this a further using cryogenic or thermoelectric coolers, complex IR
reason for which room temperature IR sensor tech- optics, and expensive sensor materials have typical costs
nologies allowing monolithic integration with silicon of some tens of thousands of Euros, while IR cameras
microelectronics are favourite for future developments) using emerging micromachined silicon bolometer
[66]. New requirements in most advanced applications arrays with NEDT close to 10 mK with costs of a few
in military surveillance, astronomy, medical and envi- thousands of Euros are expected in the civil market in
ronment fields are pushing the request, as a strategic the near future. This cost in the case of mass production,
parameter, of the multi-spectral operability (for some as is expected for collision avoidance and guidance

Figure 8. Number of pixels in infrared detector arrays versus time (Moore’s law for comparison) [66].
 Journal of Modern Optics 1679

assistance in automobile production, could be reduced 6. Smart sensors

down to less than one thousand Euros. Associated with the push towards the highest number
The actual explosions of microsystem technologies of pixels (4106) and the highest working temperatures
(sensors, control and actuators), especially in automo- (close to room temperature), the general trends of
tive applications, is, moreover, increasing this positive future detectors will show more and more an increase
growth of the IR market with a real possibility of high of the ‘intelligence’ of the sensors which will integrate
level products at contained cost. New markets (auto- the sensing function with the signal extraction, pro-
motive, law enforcement, intelligent building, environ- cessing/‘understanding’ (Smart Sensors) [44–47].
mental control) and old markets (biomedical and The term ‘Smart Sensors’ originated to indicate
medical, industrial process monitoring, energy control, sensing structures capable of gathering in an ‘intelli-
surveillance and warning) will get strong benefits from gent’ way and of pre-processing the acquired signal to
the new technology feeding also new military market give aimed and selected information. The Smart Sensor
applications (especially portable equipment, soldier of technology, based on the use of a smarter sensor
2000). In Figure 9 is shown the single pixel detector architecture, allows one to integrate technical design
cost in the last 50 years (at actual value of the year and development (from optics, detector materials,
2000) with a reduction of more than 106, that is
Downloaded by [Moskow State Univ Bibliote] at 19:58 30 August 2013

electronics, and algorithms) into the sensor’s functions

comparable only to the cost reduction for 1 Mb rather than trying to get the required performance by
memory chip for computers obtained by silicon just relying on one aspect of the technology, for
microelectronics, which is the most advanced technol- instance the number of sensor pixels.
ogy developed up to now (this is another factor One of the most advantageous application areas for
demonstrating the strong liaison between IR FPAs Smart Sensors is the infrared field where the informa-
and silicon microelectronics). tion to be extracted is generally based on very small
In conclusion, the cost of IR sensors strongly signals buried in highly intensive and diffused back-
reduced for a single pixel cost is however still high ground noise and often high intensity ‘unwanted
because of the improved performances of the new high signals’. In general, background clutter consists of
quality thermal camera (similar analysis can be done for extended objects that are more slowly varying spatially
PC where the cost of a memory bit has been reduced by than the target, therefore temporal filtering as well as
a factor of 108, but the cost of the top computer is spatial filtering, further complemented by multispectral
reduced of a factor of 10). A strong reduction of the filtering, are required for target signal detection and
cost, even for high number pixel IR FPAs is expected extraction. This implies that infrared imaging devices
thanks to silicon technology, which in the case of the require some processing of detector output signals to
silicon microbolometer is avoiding the high cost of correct non-uniformity and remove the background
cryogenic components. For future consistent cost effect and to avoid this, without on-focal-plane
reduction, particular efforts should be dedicated to processing, most of the data would be useless clutter
low cost, high luminosity IR optics based on organic or unwanted data. Because of the whole acquired
materials highly transparent in the long wavelength IR pattern only a few pixels contain the target information
region. of the selected targets.

Figure 9. Single sensor pixel cost evolution [66].

 1680 C. Corsi
Downloaded by [Moskow State Univ Bibliote] at 19:58 30 August 2013

Figure 10. Insect fly eye [40,45].

Therefore, conventional approaches need to pro- like is performed by an insect eye thanks to a spatial–
cess these confused data through all the chain of the temporal correlation [44,45]. IR Smart Sensors have
readout electronics, the analogue to digital converters been developed by integrating a columnar
and the digital signal processor before, finally, three-electrodes sub-structure designed to capture a
separating and rejecting the clutter. In contrast, the fast moving point target by a read out processing for
Smart Sensor rejects this clutter before it is read off the clutter rejection performed in an analogue way within
focal plane sensors so that most of the useless data is the multielectrode sensors on the focal plane. A scheme
not processed. The Smart Sensor design concept is for the three-electrode electronics read-out with a
based on the processing capabilities, at least at some reticule structure or implementing this analogue pro-
stage of the threshold, inside the sensor structure itself. cessing within the sensor on the focal plane is shown
This means that the Smart Sensor in some way in Figure 12. In Figure 13 the first IR smart FPAs with
emulates a living eye in the simplest stage, at least at a 1024 sensors, each one structured in 21 sub-pixels
primordial level, such as an ‘insect eye’ (Figures 10 and [44,45] is shown.
11), and, in future perspective, could reach perfor- The basic objectives of new IR smart sensors are
mances close to the ‘human eye’, thanks to neuronal much more demanding because significant improve-
network development, which can allow pattern recog- ments in the performance of VLSI processors and
nition and object discrimination [45,48,68]. infrared mosaic detector arrays are being achieved,
One of the simplest feature extractions and at the especially in the pre-threshold stage. Target signals are
same time most appealing for the numerous applica- expected to be deeply buried in background clutter
tions, is the discrimination of point sources from noise which can be much higher than the target
extended background emissions and/or of fast intensity. Therefore, imaginative pattern recognition
events (moving targets or changeable emissions) for processing techniques using all spatial, temporal and/
static or slow moving scenario just as is operating in or spectral information of both targets and back-
the fly eye. ground clutter should be developed for suppressing
In this case a reticule structured detector, which is background clutter and unwanted signal, but main-
electronically modulated to obtain a spatial–temporal taining or even enhancing the target signal. Finally, it
correlation of the focused spot target, normally buried is important to underline that the important recent
within the diffused background emission, can allow the developments of neural networks for advanced com-
detection of a point source or a well-defined shaped puting allow foreseeably an impressive growth of the
target improving the signal-to-clutter ratio: this corre- Smart Sensor concept especially for those detector
lation associated with appropriate temporal signatures, technologies which will take advantage of the possi-
can allow one to discriminate and identify the targets, bility of integrating processing devices.
 Journal of Modern Optics 1681

Figure 11. Smart sensors – foveated sensor [40,45].

Downloaded by [Moskow State Univ Bibliote] at 19:58 30 August 2013

Figure 12. Reticule structured smart sensor [40,45].

7. Applications highlights did Guglielmo Marconi propose the idea of a

7.1. Surveillance, detection and warning RadioTelemeter for localising metallic objects at dis-
tance and therefore the remote sensing was mainly just
The main application for IR technology was in the past
optical). In fact in 1910, Bellingham had presented a
and will be in the future surveillance and warning and
method to detect the presence of icebergs and steam-
more specifically for military applications, that condi-
ships by using a mirror and the original thermopile. He
tioned for a long time, the development of IR devices later patented this device in 1913. His infrared
and systems. The highlights of this application field are radiometer’s primary advantage over the
quite well known and have allowed one to develop an disappearing-filament optical pyrometer was that it
impressive know-how in system performances. The was able to detect temperatures substantially lower
first historical lesson, missed for lack of knowledge of than ambient. If this device had been installed on the
the users, was the underestimation of the strategic Titanic ship, avoiding that grave tragedy, efforts in
value of IR surveillance systems (at that time RADAR developing IR surveillance systems would probably
was not yet operating although in 1904 Christian have been much greater. During World War II great
Hülsmeyer had used radio waves for detecting ‘the efforts were dedicated to the development of IR
presence of metallic distant objects’, and only in 1922 surveillance systems especially in the Army with both
 1682 C. Corsi
Downloaded by [Moskow State Univ Bibliote] at 19:58 30 August 2013

Figure 13. 1024 21 IR smart sensors (1978 Elt – Italy) [45].

parties capable of IR detection of enemy tanks and diseased zones therefore indicating pathology of the
support to night moving. underlying organs. He wrote, ‘In whatever part of the
After the 1970s R&D developments of IR surveil- body excess of heat or cold is felt, disease is there to be
lance systems, especially for navy applications, were discovered’ [12].
carried out in the US, UK, France and Italy especially Real IR thermography was first used to investigate
(where the first modular FPA staring omni-directional breast cancer in 1957 by Raymond Law. Since then,
surveillance system prototype was designed and there have been thousands of peer-reviewed studies in
realised) [44]. the medical literature, studying hundreds of thousands
In the 1980s the SDI Programme for Ballistic of women [50–56]. After approval by the FDA in 1982
Missile Defence in the US originated a highly advanced of thermography as a breast screening procedure, there
EO surveillance system with performances close to have been an impressive number of studies supporting
BLIP limits. Nowadays, the main efforts are dedicated its effectiveness, as well as the development of strict,
to multispectral detection capability. standardised interpretation protocols, strongly sup-
Lastly, wide application of IR warning is expected ported by great technological advancements in imaging
in automotive applications for smart collision avoid- hardware and computer processing with the introduc-
ance systems in poor visibility conditions: room- tion of digital image processing, later on better known
temperature smart IR receivers could be installed as digital infrared imaging (DII) [55]. DII emerged in
with high reliability and a simple, immediate man– the 1980s, thanks to the use of highly sensitive cameras
machine interface in any type of motor vehicle [49]. to accurately measure heat from the surface of the
breast, detecting temperature differences of close to
0.1 C and allowing automatic computation of the
statistical thermal distribution [57].
7.2. Medical applications DII, a completely non-invasive [69], non-ionising,
Medical applications were among the first and the passive diagnostic technique, demonstrated the capa-
most important civil applications, although they have bility of detecting a pre-cancerous state and/or the
not been successfully recognised for their applicability early signs of breast cancer impossible to detect by
value because of some specific, marginal limits which, physical examination or mammography, and even
at the beginning caused misunderstanding in their benign tumours, simple cysts, fibro cysts, infections,
interpretation, especially among medical users who and other benign conditions could be detected. DII can
were not expert enough in IR knowledge. also be the first indicator that a cancerous situation is
The first use of thermography in medicine was by developing [51–54], even in the early stages of tumour
the Greek physician, Hippocrates, in about 400 BC. He formation, and can supply invaluable information on
was able to gain information about diseases by prognostic evolution. Finally, each patient has a
measuring the temperature distribution on immersing unique infrared map of their breasts (defined IR
the body in wet mud: the areas that dried more quickly pattern signature) which should remain constant in
indicated a warmer region and were considered to be its statistical distribution unless there is a change in the
 Journal of Modern Optics 1683

underlying vascular conditions. The DII has an science, especially when applied to material character-
invaluable diagnostic performance that unfortunately isation [58].
has not yet been recognised worldwide. The main Thermography is a very powerful tool, because with
reason for this disappointment is due to its high a calibrated device it is possible to measure radiance,
sensitivity not coupled to a specific cause: that is IR temporal and surface properties at the same time.
imaging is highly sensitive to thermal variations (down Temperature is not always simple to recover, but in
to better than 0.01 C), but it is not able to identify the favourable cases, accuracy in the order of millikelvins
specific causes of these variations (which moreover are is possible today. Because the lowest degradation step
originated only by vascular links close to the skin) and of any physical phenomena is heat, uses of tempera-
therefore can be easily misinterpreted in the absence of ture maps, so accurate and fast as 1 Mpixel in a few
a deep knowledge in thermal physiology. Thus, being a milliseconds, are far from being fully exploited.
functional and not a morphological diagnostics tool,
while the high sensitivity supplies invaluable informa-
tion for early breast cancer detection especially for
young women, deep necrotic breast cancer might not 7.3.2. Thermal non-destructive evaluation of
materials
be detected in the case of a weak knowledge in
Downloaded by [Moskow State Univ Bibliote] at 19:58 30 August 2013

thermography. New positive perspectives have A particularly important and well-documented area is
emerged recently thanks to new medical applications building inspection, conducting infrared analysis of
in highly specialised fields, where the knowledge of buildings, analysing thermal patterns/thermovision
specific diseases are linked to functional behaviour by with an in-depth understanding of physical underlying
IR thermography which is showing the whole diag- phenomena. Dealing with historical buildings and
nostic potentiality, especially when applied to specific, works of art, different purposes could be achieved,
well pre-defined diseases related with surface vascular for instance, the non-destructive evaluation of materi-
problems (detection of malignant moles, vascular als, the discovery of the hidden history of a monument,
problems, etc.) and/or rheumatism and/or control in for anti-seismic purposes or very recently, the moni-
high blood flood and dermatology surgery. Lastly DII, toring of environment conditions [59,60].
thanks to the expected cost reduction for high sensi- A new approach in monitoring the indoor envi-
tivity room-temperature thermal cameras, is allowing ronment of a building has been recently developed
one to forecast a wide market use in sport-related using a thermodynamic basis. It uses temperature as
trauma and veterinary diagnostics. the driving parameter and is especially suited for
comfort analysis or moisture evaluation. IR thermo-
graphy is the method more suited to cope with such a
7.3. Civil industrial applications: energy control and task. It allows measuring, at the same time, of surface
design temperature and all the basic moist air parameters,
With the help of new IR systems the spectrum of civil such as, air temperature, relative humidity and air
uses can be applied everywhere where processes inside speed. In such a way a very detailed heat and mass
of an object lead to a change of temperature of its balance could be accomplished close to the wall or in
surface. Bearing in mind the current discussion on any other part of the room.
climatic changes and environment control, applied Even the energy saving issue could be deeply
thermography is gaining importance, as for example, investigated using an optical measurement of temper-
it can quickly discover heat losses from building ature and heat flux. Basically, the non-contact nature
envelopes. of such a measurement is much less prone to error than
any contact measurement. In fact, the time constant
needed for the thermal equilibrium, which is a great
7.3.1. Quantitative thermography source of errors in traditional measurement techniques,
An overview of the current R&D works about the use is practically equal to zero. Last, but not least, the
of IR thermography for non-destructive testing is a temperature recorded by any contact probe is a
large task due to the continuously increasing popular- mixture of surface and air temperature. This is not
ity of such a method. It is a matter of fact that ASNT the case using radiometric measurements.
included thermography in the main standardised Furthermore, the imaging of a large surface is the
method [58]. only way to achieve a correct mean value and
The quantitative use of IR thermography covers a measure local anomalies. The literature provides
large number of different disciplines. Actually, some standard practices for thermographic inspections of
applications are on the boundary with photo-thermal buildings [61].
 1684 C. Corsi

7.3.3. Thermal diagnostics in fluid dynamics room-temperature microbolometers for future and
An important application of high resolution, high extremely valid applications in the civil field).
frame rate thermography, especially for the high Operational requirements (mainly of maintenance
scientific knowledge involved, is to visualise and and reliability) are pushing IR science to look for
analyse flow phenomena. Infrared technology has new advanced sensors which could avoid the need for
been employed in wind tunnels and in flight at cryogenics. New microbolometer technologies, because
subsonic to hypersonic conditions with both local of the micro size of thin film bolometers, completely
(that is, when the camera and subject surface are integratable with silicon technology and therefore
mounted on aircraft) and remote camera installations often named silicon microbolometers, have been
[61–64]. With state-of-the-art equipment and advanced emerging in the last few years with a very high promise
methods, greater success and improved resolution of for future IR sensor market growth. For extremely
details of various flow phenomena have become high sensitivity sensors, especially for military and
possible. Infrared thermography has many benefits space applications, the technologies with actual major
over other methods. It is global in nature and possibilities for future development are mainly based
non-intrusive. The basic principle behind IR thermo- on photon sensors (intrinsic and quantum well).
Downloaded by [Moskow State Univ Bibliote] at 19:58 30 August 2013

graphy is the measurement of surface emissions in the Multispectral and hyperspectral capabilities with spa-
IR radiation band, which are directly related to surface tial–temporal filtering capabilities are also emerging,
temperature [62–64]. Surface shear stress and, thereby, especially in military and space applications for target
convective heat transfer with the free stream varies identification. The key emerging factor for future IR
with the boundary layer state where the boundary FPA technologies is room-temperature working for
layer state changes, such as at transition. With uncooled imaging systems and the complete integr-
state-of-the-art equipment the temperature difference
ability with silicon microcircuit technology, especially
changes can be visualised. Shock waves and possibly
because integrated signal processing (smart sensors)
flow can be measured to an accuracy of nearly 0.1 C
will play a fundamental role in future applications
over a small area [63].
where mass production could allow consistent cost
These characteristics make IR thermography a very
reduction (huge markets are expected especially for
powerful tool to visualise certain flow phenomena. In
addition to transition, any flow phenomena that creates automotive applications for the unit cost of a few
a measurable temperature change can be visualised. hundred Euros).
The deep scientific knowledge required in this So, for the first time, thanks also to the elimination
application guarantees the reliability of the IR of cryogenic cooling, wide use of IR Smart Sensors are
measurements, although as a counterpart, it is limiting emerging on the international market, becoming stra-
the market size to a few R&D centres. tegic components for major important areas, such as
transport (especially cars, aircrafts and helicopters),
security, environment and territory control, biomedi-
cine and in helping ‘human beings to a better life’
8. Conclusions
(intelligent building, energetic control structuring,
Future trends for infrared detectors are linked to auxiliaries for disabled people, etc.).
various and complex parameters like the emerging
technologies of sensor fabrication for mass production
especially for civil applications, with a strong two-way
synergy between civil and military applications. New References
markets (automotive, intelligent building, environmen-
tal control) and consolidated markets (biomedical and [1] Smith, R.A.; Jones, F.E.; Chasmar, R.P. Detection and
medical, energy control, surveillance and warning) will Measurement of Infrared Radiation; Oxford University
get strong benefits from microsystem technologies Press: Oxford, 1958.
(sensors, control and actuators) especially in automo- [2] Kruse, P.W.; McGlauchLin, L.D.; McQuistan, R.B.
Elements of Infrared Technology; Wiley: New York, 1962.
tive applications, with a real possibility of high level
[3] Wolfe, W.L.; Zissis, G.J. Infrared Handbook;
products at contained cost feeding a double-wave of
Environmental Research Institute of Michigan:
technology transfer between civil and military applica- Michigan, 1985.
tions (in the future more and more towards new [4] Hudson, R.H. Infrared System Engineering; Wiley-
military market applications, especially portable equip- Interscience: New York, 1969.
ment). The competition among various technologies [5] Elliott, C.T. Infrared Detectors. In Handbook on
and ‘technical schools’ has been strong with unforesee- Semiconductors; Hilsum, C., Ed.; North Holland:
able emerging new actors in the last few years (overall, Amsterdam, 1982; Vol. 4, pp 727–798.
 Journal of Modern Optics 1685

[6] Rogalski, A. Infrared Detectors; Electrocomponent [36] Norton, P.; Campbell III, J.; Horn, S.; Reago, D. Proc.
Science Monograph; Gordon Breach Science Publisher: SPIE 2000, 4130, 226–236.
New York, 2000; Vol. 10. [37] Rogalski, A. Proc. SPIE 2001, Vl 4413, 307–322.
[7] Henini, M.; Razeghi, M., Eds.; Handbook of Infra-red [38] Rogalski, A. Prog. Quantum Electron. 2003, 27,
Detection Technologies; Elsevier: Amsterdam, 2009. 59–210.
[8] Sze, S.M. Physics of Semiconductor Devices; Wiley: [39] Bertrand, F.; Tissot, J.T.; Destefanis, G. 2nd Generation
New York, 1981. Cooled IR Detectors. State of the Art and Prospects.
[9] Van der Ziel, A. Fluctuation Phenomena in In Physics of Semiconductor Devices; Kumar, V.,
Semiconductors; Butterworths: London, 1959. Agarwal, S.K., Eds.; Narosa Publ. House: New Delhi,
[10] Corsi, C. Rivelatori. In Tecniche e strumenti per 1998; Vol. II, pp 713–720.
il telerilevamento ambientale; Gilardini, A., Galati, [40] Corsi, C. Atti Fond. Giorgio Ronchi 1998, LIII (1–3),
G., Eds.; CNR: Rome, 2002; pp 485–542. 11–20.
[11] Infrared Tutorial. www.ZyTemp.com/Radiant [41] Norton, P.R. Proc. SPIE 1999, 3698, 652–665.
Innovation Inc. Version 006, 2000–2003 (accessed [42] Ross, P.E. Inte. Roadmap for Semiconductors; ITRS.
May 10, 2005). http://public.itrs.net (accessed May 10, 2001).
[12] Adams, F. The Genuine Works of Hippocrates; Williams [43] Hodapp, K.W.; Kuhn, J.; Thornton, R.; Irwin, E.;
and Wilkins: Baltimore, 1939. Yamada, H.; Waterson, M.; Kozlowski, L.; Montroy,
Downloaded by [Moskow State Univ Bibliote] at 19:58 30 August 2013

[13] Sherwood, T.F. Ann. Sci. 1942, 5, 129–156. J.T.; Haas, A.; Vural, K.; Cabell, C. Proc. SPIE 1999,
[14] Quinn, T.J. Temperature; Academic Press: New York, 3698, 24–35.
1990. [44] Corsi C. International NATO Electronics Warfare
[15] Zemansky, M.W. Heat and Thermodynamics; Conference, Washington, DC, 1978.
McGraw Hill: New York, 1968. [45] Corsi, C.; di Nola, G.; Marangoni, G.; Salcito, G.
[16] Barr, E.S. Am. J. Phys. 1960, 28, 42–54. Elements Ottico Integrato per filtraggio spaziale inorpor-
[17] Herschel, W. Phil. Trans. Roy. Soc. Lond. 1800, 90, ante un mododulatore di ampiezza di segnali di frequenza
284–292. spaziale desiderata; Technical Report PT-79, Elettronica
[18] Lawson, W.D.; Nielson, S.; Putley, E.H.; Young, A.S. SpA: Rome, 1979.
J. Phys. Chem. Solids 1959, 9, 325–329. [46] Tao, T.F. Proc. SPIE 1979, 178, 2–12.
[19] Melngailis, I.; Harman, T.C. Semiconductors and [47] Corsi, C. Microsystem Technol. 1995, 1, 149–154.
Semimetals; Academic Press: New York, 1970; Vol. 5, [48] Corsi, C. Infrared Phys. 2007, 49, 192–197.
pp 11–174. [49] Corsi, C. Systems for Avoiding Collision of Vehicles in
[20] Boyle, W.S.; Smith, G.E. Bell System Tech. J. 1970, 49, Low Visibility Conditions. European Patent WO/2000/
587–593. 022596, 2000.
[21] Barbe, D.F. Proc. IEEE 1975, 63, 38–67. [50] Early Detection of Breast Cancer. Redding
[22] Steckl, A.J.; Nelson, R.D.; French, B.T; Gudmundsen,
Thermography, Infrared Imaging Center, Redding, CA.
R.A.; Schechter, D. Proc. IEEE 1975, 63, 67–74.
http://www.reddingthermography.com/ (accessed May,
[23] Corsi, C. Proc. IEEE 1975, 62 (14), 14–26.
10, 2007).
[24] Elliot, C.T.; Day, D.; Wilson, D. Infrared Phys. 1982, 22,
[51] Bronzino, J.D., Ed.; Medical Devices and Systems; CRC
31–42.
Press: Abingdon, 2007.
[25] Shepherd, F. Proc. SPIE 1983, 443, 42–49.
[52] Ring, E.F.J. In The History of Thermal Imaging in the
[26] Kosonocky, W. Proc. SPIE 1983, 443, 167–173.
Thermal Image in Medicine and Biology; Amner, K.,
[27] Gunapala, S.D.; Bandara, S.V. Thin Films 1995, 21,
Ring, E.F.J., Eds.; Elsevier: Vienna, 1995; pp 13–20.
113–237.
[53] Louis, K.; Walter, J.; Gautherie, M. Biomedical
[28] Watton, R. Ferroelectrics 1989, 91, 87–108.
Thermology; Alan R. Liss: New York, 1982; pp 279–301.
[29] Wood, R.A.; Han, C.J.; Kruse, P.W. In IEEE, Solid
[54] Gauthrie, M. Biomedical Thermology; Alan R. Liss:
State Sensors & Actuators Workshop; IEEE: USA, June
New York, 1982; pp 897–905.
1992.
[55] Corsi, C. IR Computerized Thermography. Proceedings
[30] Norton, P.R. Proc. SPIE 1998, 3379, 102–114.
[31] Elliott, C.T. Future Infrared Detectors Technologies. of the Biomedical International Conference, S. Elpidio,
Proceedings of Advanced Detectors and Systems, June; Italy, 1981.
IEEE Press: Piscataway, NJ, 1990; pp 61–68. [56] Amalric, R.; Giraud, D.; Spitalier, H. Acta
[32] Corsi, C. Atti Fond. Giorgio Ronchi 1991, XLVI (5), Thermographica 1984, 8, 21–24.
801–810. [57] Corsi, C. Acta Thermographica 1983, 7, 33–39.
[33] Corsi, C. Future Trends and Advanced Development. [58] Maldague, X. Technical Ed., ASNT Handbook, 3rd
In I.R. Detectors: Proceedings of the 2nd Joint Edition, Volume 3, Infrared and Thermal Testing;
Conference IRIS-NATO, London, June 25–28, 1996. Moores, P.O., Ed.; American Society for Nondestructive
[34] Razeghi, M. Opto-Electronics 1998, 6, 155–194. Testing: Columbus, OH, 2001.
[35] Kozlowski, L.J.; Kosonocky, W.F. Infrared Detector [59] Bison, P.; Cernuschi, F.; Grinzato, E. Int. J.
Arrays. In Handbook of Optics; Bass, M., Williams, Thermophys. 2008, 29, 2149–2161.
D.R., Wolfe, W.L., Eds.; McGraw-Hill: New York, [60] Vavilov, V.; Kauppinen, T.; Grinzato, E. Res.
1995; Chapter 23. Nondestruct. Eval. 1997, 9, 181–200.
 1686 C. Corsi

[61] Grinzato, E.; Bison, P.G.; Marinetti, S. J. Cul. Heritage [66] Corsi, C. Giorgio Ronchi Foundation 2002, LVII, 363–369.
2002, 3, 21–29. [67] Corsi, C. IR Technologies: History Lessons and New
[62] Carlomagno, G.; Cardone, G.; Astarita, T. Perspectives. In Advanced Infrared Technology and
Thermography Rev. Gen. Therm. 1998, 37, 644–652. Applications International Workshop 2007, AITA 9;
[63] Banks, D.W.; Quinto, P.F.; Paulson Jr. J.W. Technical Strojnic, M., Ed.; Leon, Mexico, Oct. 8–12, 2007;
Report 2001-210848; NASA Dryden Flight Res. Center: pp 4–19.
Edwards, CA, 2001. [68] Moini, A. Vision Chips or Seeing Silicon; Technical
[64] Astarita, T.; Cardone, G.; Carlomagno, G.M.; Report; University of Adelaide: Australia, 1998.
Meola, C. Opt. Laser Technol. 2000, 32, 593–610. [69] Spitalier, J.; Amalric, D.; Giraud, D. Proceeedings of the
[65] McGee, T.D. Principles and Methods of Temperature International Conference; MTP Press: Boston, 1983;
Measurement; Wiley-InterScience: New York, 1988. pp 173–179.
Downloaded by [Moskow State Univ Bibliote] at 19:58 30 August 2013