GB2128019A - Infrared radiation detection device - Google Patents

Abstract

An IR detector comprises a strip of infrared photosensitive material (3) having bias contacts (7,9) and a multiplicity of low-resistance read-out taps (11) distributed along the strip (3) between the bias contacts (7,9). The detector (1) may be biassed using one of two different modes: an equilibrium mode employing low frequency current bias; or, a dynamic mode employing a sequence of fast current pulses of alternating polarity. Each tap (11) may be referred to one of a similar number of differential amplifiers (15) and compared with a common reference. Alternatively, the taps (11) may be connected in consecutive pairs to differential amplifiers (15) to produce first order difference output signals. A second set of differential amplifiers (17) may be added to allow extraction of second order difference output signals. The strip (3) may be mounted upon semi- conductor material and the amplifiers (15,17) integrated in this material. <IMAGE>

Description

SPECIFICATION Infrared detection Technical field This invention concerns detectors, detector systems and operational methods, for the detection of infrared radiation; particularly photoconductive detectors and systems. The detectors embodying the invention, as described below, are intended to have application in so called "staring" array systems; in slow-scanned (transverse or longitudinal) linear array systems; in long integration time ultra-sensitive detector systems; and, in position sensitive point source detection systems.

Background art Starting systems incorporating detector arrays of photo-sensitive elements either junction photodiode sensing elements, or photo-conductive sensing elements, are known. In these systems radiation from a thermal scene is focussed onto a shielded, cooled, element array and photosignals developed by each element are extracted and processed either for image reconstruction or for other information presentation or analysis.

Operation of photoconductive detectors requires current bias. In a conventional photoconductive detector system, the steady bias current produces a standing DC output may times larger than photo-signal, and because of non-uniformities in bias pedestal it is difficult or impractical to back this off accurately, for each array element, so that desired scene dependant signals can be extracted, processed, and used to control display without introducing an unacceptable degree of fixed pattern noise. The intrinsic photoconductive detector elements normally employed exhibit low resistance even in the absence of infrared long wavelength radiation, and only small fractional changes of resistance when exposed to signal radiation.In a typical example a 50pm square photoconductive element used with an F/1 lens and shield receives ~ 0.3 > W of radiation from an ambient temperature scene and an additional 0.4 nWfrom an object 0.1 K above ambient filling its field of view. For an element resistance of 50Q and typical responsivity of 5 x 1 04vow at 3mA bias, it exhibits a standing bias voltage (ie bias pedestal) in the dark of 150 mV; a voltage 4mV lower in the presence of background radiant illumination; and only a further reduction of 8#V in the presence of the object radiation (ie a reduction a factor 5 x 10-5 of the dark signal).Because of the probable large ratios between pedestal variations and the derivedsignals-these ratios will vary from element to element, and will vary with operational temperature; - compensation proves impractical in staring arrays.

Photovoltaic detectors - eg junction photodiode array detectors- do not require bias, and fixed pattern noise, which is of different origin, is not so difficult to suppress. For this reason these detectors are much favoured. They are of high impedance, offer low power dissipation and low heat load, and can be matched readily to integrated or hybrid change transfer devices for array addressing and signal processing in-situ.

However, it has not yet been demonstrated that adequate diode quality can be achieved in routine fabrication either for 8-14 Fm band 770K liquid cooled operation or for 3-5 ijm band 200"Kthermoelectric cooled operation. The principle drawbacks are poor stability during high temperature storage - inevitable in a military user environment; and, poor long-term operational stability of the diode characteristics - ie limited operational reliability and relatively short service lifetime.

The principal attraction of photoconductive detectors for staring array applications, which makes it worthwhile considering solutions to the pedestal problem, is the better established fabrication technology in device material - eg in cadmium mercury telluride (CMT), and better stability during high temperature storage. Device stability is particularly important since the detector characteristics must be well-defined, known, and, stored, in off-focal plane processing electronics to correct for non-uniform response to background radiation.

Disclosure of the invention This invention is intended to provide methods of operation, apparatus and photoconductive detectors that obviate the problem of bias pedestal. In accordance with the invention there is provided a detector comprised of at least one strip of infrared photosensitive material, the strip having bias contacts and between these contacts a multiplicity of thin low resistance read-out contacts close spaced and distributed along the length of the strip.

In accordance with the invention there is also provided a method of biassing the detector aforesaid, wherein there is applied to the bias contacts a current bias signal cyclic in time and of such waveform and frequency as to produce, for each spaced read-out contact, zero net average signal both in the dark and in the presence of infrared radiation of uniform intensity over the area of the detector.

The bias signal aforesaid, may be an alternating signal - eg a signal of square or sinusoidal waveform. For this mode of operation the frequency of the signal is chosen so that photocarriers produced in the presence of radiation can follow changes in bias signal and establish equilibrium for each half-cycle of the bias current. In this mode the equilibrium distribution of photocarriers is changed and differs for the up-field and down-field directions. The output signal follows these distribution changes and thus differs in each half-cycle, not only in sign but also in amplitude. The response is non-linear in the presence of non-uniform illumination.

Alternatively, the bias signal may be of pulse waveform - a train or burst of pulses, the consecutive pulses being of opposite polarity. For this mode of operation the pulse lengths chosen are short compared with carrier lifetime. The pulse spacings, however, are long enough to allow equilibrium to re-establish between each pulse. in this mode the photocarriers are shifted either towards or away from the output contacts, on consecutive pulses. As previously, the response is non-linear in the presence of non-uniform illumination.

For both modes, common mode voltage - ie the sum of bias pedestal and background signal is readily capable of rejection - eg by following the outputs with time integration or harmonic extraction to extract useful signal.

There is also provided in accordance with this invention apparatus for performing the methods aforesaid, comprising in combination the detector and a bias source connected to the bias contacts of the detector, this source providing the current bias of either alternating or pulse waveform.

The output signals from the detector may be extracted via one or more banks of differential input integrating amplifiers: For image reconstruction signals may be taken from the outputs of a single bank of amplifiers. Each read-out contact is connected to one of the differential inputs of a corresponding one of the amplifiers and a common reference signal from one of the bias contacts or from one of the read-out contacts is supplied to all other differential inputs.

Alternatively, for different image construction, the output contacts may be connected pairwise each to the two differential inputs of corresponding amplifiers. Such differential image serve to enhance the boundary edges of objects in the thermal scene. A second bank of differential amplifiers may be used to derive second-difference signals.

Brief introduction of the drawings In the drawings accompanying this specification: Figure 1 is an illustrative plan view showing a detector adapted in accord with this invention; Figures 2, 3 and 4 are schematic diagrams of apparatus, incorporating the detector shown in the preceding Figure, adapted for the derivation of first difference, second difference and zero difference output signals, respectively; Figure 5 is a graph illustrating the assymetric equilibrium distributions of photocarriers in a detector corresponding to line source excitation for positive bias (bold outline), and, for negative bias (broken outline);; Figures 6, 7 and 8 are graph representations illustrating the detector signal response, spatial derivative, and spatial integral response, respectively, calculated as a function of the distance of an infinitesimally close pair of contacts from a line source image focussed upon the detector, for three different values of alternating bias current amplitude; Figure 9 is a graph illustrating the symmetric equilibrium and non-equilibrium distributions of photocarriers in a detector due to line source excitation, for zero, negative fast pulse and positive fast pulse bias; Figure 10 is a graph comparing the time variance of (a) photocarrier concentration in the proximity of a pair of read-out contacts, and (b) the corresponding read-out signal level, for fast pulse bias.

Figures 11 and 12 are graph representations illustrating the detector signal response and spatial integral response respectively, calculated as a function of the distance of an infinitesimally close pair of contacts from a line source image focussed upon the detector, for three different values of fast pulse bias current amplitude.

Description of the preferred embodiments A detector 1 is shown in Figure 1. It is comprised of a strip 3 of photosensitive material mounted upon a supporting substrate 5 of insulating material. Bias contacts 7 and 9 are provided at each end of this strip and between these there is distributed a multiplicity of low-resistance read-out contacts 11. These contacts 11 are close-spaced and evenly distributed (ie equispaced) along the length of the strip 3. Even spacing is chosen so that the resolution is constant for all read-outs 11.

In this example the strip 3 is a filament of cadmium mercury telluride (CMT) photosensitive material, and this has been formed by cutting and thinning down material from a well characterised good quality bulk crystal. The dimensions of this filament are as follows: width : 50 pm length: 500 > m depth : 1 0 > m The width is that of a conventional photodetector element, though of course widths between 25 > m and 100 > m would be practical. The length chosen affords a reasonable number of resolution elements, though any length between say a factor 3 to 10 times the width again would be quite practical.

The depth has been chosen as a compromise value. It is neither too thin, for then the detector response would be dominated by undesirable surface effects, nor is it too thick, for then power dissipation would be unnecessarily high. The dimensions quoted above are those for 3-5 Fm band infrared sensitive CMT material (Cdy Hg1-y Te: y = 0.28). For thin material, the minority carriers have at an operating temperature T of 200"K a typical lifetime of 10secs, a characteristic diffusion length X of between 20 and 50 iim, and a typical mobility of 200 cmV-1S-1.

The bias contacts 7 and 9 and the read-out contacts 11 have been provided by metal deposition and defined photolithographically. The read-out contacts 11 are as thin as is practical, typically 5 iim wide or less, in order to optimise the area of the detector strip 3 that can be exposed to radiation.

In this example the read-out contacts 11 are spaced a distance 20 Fm apart (i.e. 25 m from centre to centre). This is a spacing that is large compared with the widths of each contact; the proportion of material that is uncovered by metal is thus high. At the same time the spacing is at most comparable if not significantly less than the characteristic diffusion length. For a line source image, the equilibrium distribution of the photocarriers will then extend across several of the read-out contacts 11. Horizontal resolution is, at least in part, determined by this spacing. Vertical resolution is determined by the width of the strip 3. When incorporated as part of a detecting system, the detector 1 is matched up to an F/2 lens and cold shield and is mounted upon a thermo-electric cooler and encapsulated in a dewar.

The principles of operation will now be considered, and reference will be made to Figures 5 to 8 of the drawings. These principles are most easily demonstrated for an alternating current bias of square waveform applied to the bias contacts 7 and 9 of the detector 1 and for line source illumination of the detector 1. This bias produces a driving field of half-amplitude E each half cycle. The period of the alternating bias is such that diffusion and recombination mechanisms have time to establish equilibrium and in each half cycle the carriers attain a steady rate of distribution. This equilibrium condition imposes a limitation on bias frequency f: f liT Thus for the example described above, lifetime T of 10 Fsecs, a suitable frequency f would be 10 kHz.

The two steady state distributions, one for positive bias (+E) and one for negative bias (-E) are illustrated in Figure 5. As can be seen from Figure 5 the photocarrier distribution is in each case assymetric. This arises because of drift effects in the presence of bias.

The distributions are solutions of the continuity equation: D dp - E dP = p/# = G#(x); -l-G5( dx where p is the carrier density; D the diffusion constant (200 K); G the line source image density; and x the distance from image line centre.

These solutions are as follows:

where Z = FET and '= DT 2# The carrier distribution curve each side of the line image (x = o) is similar to that of the zero bias distribution, but each corresponds to a modified distribution length (Au, Ad) :- #u = #/[Z + @(1 + Z2) ] & Xd = - Z + #(l + Z2) ] .

Consider now an infinitesimally close pair of read-out contacts a distance x' from the line image. At the point x' the concentration of carriers switches from high to low and back again as the field alternately drifts the carriers towards and away from the contact pair. The scale of the concentrations is proportional to the source intensity G, and the photoconductive signal is therefore proportional to intensity, and also to the product of carrier concentration and field. Clearly when x' is positive, the drift gives a larger positive photosignal than a negative one when the bias field is reversed.

The average (ie time integral) output from the contacts is non-zero. If the point of generation is at the point x' (x' = 0) symmetry demands no net signal. If the point of generation is a long way from x' (ie several diffusion lengths) carriers recombine before reaching the read-out zone, also giving no net signal.

The signal < pE > per unit width of operation is given by:

This function is illustrated by Figure 6 for different values of bias; the bias parameter Z = 1, 2 and 3.

The response to a line source is therefore derivative-like, with no pedestal contribution. A corollary is that there is no response to a uniform scene, or to background illumination.

A practical apparatus for extracting signal information at points along the length of the detector strip 3 is shown in Figure 2. In this apparatus circuit a constant current alternating signal source 13 (ie a high impedance source) is connected across the two bias contacts 7 and 9. Signals from the read-out contacts 11 are fed to a first bank of differential input integrating amplifiers 15. These may be external to the mounted detector.

It is however inconvenient to have a large number of leads leaving the detector 1. The detector is dewar encapsulated and these leads would only increase heat conduction to the detector and load the cooler. To avoid this, it is convenient to mount the strip 3 on a substrate of insulated semiconductor material, having integrated these amplifiers 15, as also multiplex components, in the semiconductor.

As shown in Figure 2, the read-out contacts 11 are connected in pairs to the differential inputs of the amplifiers 15.

In the analysis above, it was assumed that the read-out contacts were spaced infinitesimally close together. This of course is not the case for the practical detector 1 described, for they are about a diffusion length or less apart. The actual output will thus be broadened by this amount. (It is in effect a spatial integral over a series of infinitely closely spaced pairs; a convolution of the function illustrated by Figure 6).

Because of the very long downstream diffusion lengths Ad at high fields E, a penalty for using high fields (to get large signals) is broad spatial response to point sources. The response slope near the origin (Figure 6) gets steeper so a second derivative would give an increasingly peaked output as a function of position (see Figure 7). From this point of view a second derivative detector would serve as a good point source detector, giving position information. The second derivative is found practically by finding the difference in outputs between adjacent pairs of contacts. Practical apparatus is shown in Figure 3. This apparatus is a modification of that shown in Figure 2 and includes a second bank of differential input amplifiers 17.The differential inputs of these amplifiers 17 are connected to pairs of the first bank amplifier 15 outputs and therefore provide second-difference signals at their output.

Note that the full width of the response (Figure 7) is less than A/2. Applying this technique to point source detection in the 3-5 m band with cooling to 2000K it would provide spatial resolution of 25 calm, and assuming the bias parameter Z = 2 would have a power dissipation of 4 #W/resolution element, allowing arrays of many thousands of resolution elements on a thermoelectric cooler. [In the 8-14 pLm band at 77 K a CMT (y = 2.0) material detector would allow resolution down to about 12 ,u m with a power dissipation/resolution element of 5 ,uwW ] .

If true image information is required, the output signals from the read-out contacts 11 can be spatially integrated to restore some resemblance to the original scene. This allows broad images to be reconstructed.

The integral is just the sum of the outputs of all the read-out pairs, and this is the same as the signal from one read-out contact referred to ground at the end of the detector. Apparatus for doing this is shown in Figure 4, a modified configuration of the apparatus shown in Figure 2. In this case each read-out contact is connected to one of the differential inputs, the (+) input, of a corresponding one of the differential input integrating amplifiers 15. The other differential inputs, the (-) inputs, are connected in common to circuit earth. The integral function for a line source image is shown in Figure 8.

Square wave alternating bias is not essential, indeed sinusoidal wave alternating bias is easier to use, and requires lower bandwidth, or only tuned amplifiers to amplify the signal. Nor is it essential to derive output information by time average integration. The read-out signals are non-linear and thus contains harmonics of the bias frequency. Useful signal may then be extracted using tuned filters, filters tuned to the harmonic frequency. This is most suitable for sine wave bias drive, or simple combination of sine waves, perhaps with superimposed DC bias. A particularly important combination of sine waves is represented by an amplitude modulated AC drive, in which case the detected signal is the demodulated waveform, suitably chosen to be at a frequency which avoids low frequency noise.

In the examples given above, the resolution is limited by the long drift/diffusion Xu distance in the direction of the electric fields. An improvement may be gained by applying the bias in short pulses, pulses of alternating polarity. For this mode of operation, in the absence of bias, the equilibrium carrier distribution near the point of application of a line source is a simple exponential with a relaxation length equal to the diffusion length (Z = 0). If a high field is suddenly applied to the existing carrier distribution begins to drift in the direction of the field, initially retaining its shape, but starting to deform to the skew distribution described above.Provided the field is removed in time short compared to a lifetime T, then the deformation does not develop significantly and it can be assumed to a good approximation that the whole distribution is drifted bodily along the direction of the field (Figure 9). After removal of the field there is a period of recovery where the distribution reverts back to the initial zero-bias form. A pulse of opposite polarity is then applied and the read-out output integrated as above to give a net signal. Representative time dependant carrier concentration in the region of the read-out zone, and the corresponding signal output are shown in Figure 10.

The response and integral response functions are shown in Figures 11 and 12 respectively. From Figure 11 it can be seen that the response diminishes rapidly at distances remote from the line image (x = 0), particularly for low field (Z = 1). The integral distribution is of order two diffusion lengths 2A wide for low fields. This mode is therefore preferred for image reconstruction applications.

CLAIMS (Filed on 15.9.83.) 1. An infrared radiation detector comprising at least one strip of infrared photosensitive material, the strip having bias contacts and between these contacts a multiplicity of thin low-resistance read-out contacts close-spaced and distributed along the length of the strip.

2. A detector, as claimed in Claim 1, and in combination therewith a first bank of matched differential amplifiers, each contact and the next consecutive contact being connected to the differential inputs of a corresponding one of the amplifiers.

3. A detector, as claimed in Claim 2, and in combination therewith a second bank of matched differential amplifiers, each first bank amplifier and the next consecutive amplifier being connected to the differential inputs of a corresponding one of the second bank amplifiers.

4. A detector, as claimed in Claim 1, and in combination therewith a first bank of matched differential amplifiers, each contact being connected to a corresponding one of the amplifiers, each amplifier being connected also to a common reference.

5. A detector combination as claimed in any one of the preceding Claims 2 to 4, the strip being mounted upon insulated semiconductor material, the amplifiers being embodied as integrated components in this semiconductor material.

6. A method of biassing the detector, claimed in Claim 1, wherein there is applied to the bias contacts an alternating current bias signal, a signal of frequency sufficiently low as to permit photocarriers generated within the detector to follow changes in the bias signal and to establish equilibrium for each half-cycle of the bias current.

7. Apparatus, for performing the method claimed in Claim 6, comprising the detector claimed in Claim 1 together with a connected source of alternating current bias, a bias of frequency sufficiently low as aforesaid.

8. A method of biassing the detector, claimed in Claim 1, wherein there is applied to the bias contacts a period sequence of current pulses alternating in polarity, the duration of each pulse being short in comparison with the characteristic photocarrier lifetime, and the temporal separation of consecutive pulses being sufficiently long as to allow photocarriers to re-establish an equilibrium distribution.

9. Apparatus, for performing the method claimed in Claim 8, comprising the detector claimed in Claim 1 together with a connected current pulse source; the source being capable of producing pulses of alternate polarity, and of the duration and the separation aforesaid.

10. An infrared detector, detector combination, or apparatus, constructed, arranged, and adapted to perform, substantially as described hereinbefore with reference to and as shown in the accompanying drawings.

**WARNING** end of DESC field may overlap start of CLMS **.

Claims (10)

**WARNING** start of CLMS field may overlap end of DESC **. the direction of the field, initially retaining its shape, but starting to deform to the skew distribution described above. Provided the field is removed in time short compared to a lifetime T, then the deformation does not develop significantly and it can be assumed to a good approximation that the whole distribution is drifted bodily along the direction of the field (Figure 9). After removal of the field there is a period of recovery where the distribution reverts back to the initial zero-bias form. A pulse of opposite polarity is then applied and the read-out output integrated as above to give a net signal. Representative time dependant carrier concentration in the region of the read-out zone, and the corresponding signal output are shown in Figure 10. The response and integral response functions are shown in Figures 11 and 12 respectively. From Figure 11 it can be seen that the response diminishes rapidly at distances remote from the line image (x = 0), particularly for low field (Z = 1). The integral distribution is of order two diffusion lengths 2A wide for low fields. This mode is therefore preferred for image reconstruction applications. CLAIMS (Filed on 15.9.83.)

1. An infrared radiation detector comprising at least one strip of infrared photosensitive material, the strip having bias contacts and between these contacts a multiplicity of thin low-resistance read-out contacts close-spaced and distributed along the length of the strip.

2. A detector, as claimed in Claim 1, and in combination therewith a first bank of matched differential amplifiers, each contact and the next consecutive contact being connected to the differential inputs of a corresponding one of the amplifiers.

3. A detector, as claimed in Claim 2, and in combination therewith a second bank of matched differential amplifiers, each first bank amplifier and the next consecutive amplifier being connected to the differential inputs of a corresponding one of the second bank amplifiers.

4. A detector, as claimed in Claim 1, and in combination therewith a first bank of matched differential amplifiers, each contact being connected to a corresponding one of the amplifiers, each amplifier being connected also to a common reference.

5. A detector combination as claimed in any one of the preceding Claims 2 to 4, the strip being mounted upon insulated semiconductor material, the amplifiers being embodied as integrated components in this semiconductor material.

6. A method of biassing the detector, claimed in Claim 1, wherein there is applied to the bias contacts an alternating current bias signal, a signal of frequency sufficiently low as to permit photocarriers generated within the detector to follow changes in the bias signal and to establish equilibrium for each half-cycle of the bias current.

7. Apparatus, for performing the method claimed in Claim 6, comprising the detector claimed in Claim 1 together with a connected source of alternating current bias, a bias of frequency sufficiently low as aforesaid.

8. A method of biassing the detector, claimed in Claim 1, wherein there is applied to the bias contacts a period sequence of current pulses alternating in polarity, the duration of each pulse being short in comparison with the characteristic photocarrier lifetime, and the temporal separation of consecutive pulses being sufficiently long as to allow photocarriers to re-establish an equilibrium distribution.

9. Apparatus, for performing the method claimed in Claim 8, comprising the detector claimed in Claim 1 together with a connected current pulse source; the source being capable of producing pulses of alternate polarity, and of the duration and the separation aforesaid.

10. An infrared detector, detector combination, or apparatus, constructed, arranged, and adapted to perform, substantially as described hereinbefore with reference to and as shown in the accompanying drawings.