FIELD OF THE INVENTION

The present invention relates to an infrared sensor, comprising a focal plane array (FPA) of resistance microbolometer infrared detectors connected in such a manner to produce different pixel formats to meet specific detection requirements. Typically each imaging pixel may be a mosaic comprising a number of sub-pixels connected in parallel (although other configurations are possible), resulting in enhanced performance and ease of manufacture by micro-fabrication methods. The FPA may be integrated with a readout microcircuit on the same substrate so that with appropriate signal processing one is capable of forming an image of the field of view of interest, facilitating target recognition and very low false alarm rate.

The arrangement of such as mosaic pixel focal plane array will hereinafter be referred to as a MP-FPA.

The present invention also further relates to a method of formulating such an infrared sensor utilising not only novel architecture but also novel fabrication techniques and materials.

BACKGROUND OF THE INVENTION

There are many potential applications for high performance infrared (IR) sensors in the field of security and other related areas in the military, industrial and domestic markets. Current sensors are either too expensive for wide-scale use in this market segment, or have inadequate performance. The common passive infrared (PIR) sensor used widely in security systems is incapable of imaging and produces a simple hot-spot detection of a moving target. These sensors cannot detect stationary targets, or discriminate clearly between a human and other target, and are subject to false alarms. To date there are no IR security, sensors known to the inventor that facilitate detection and recognition of targets of interest with minimal false alarms and at an affordable cost acceptable to this market segment. In particular there . are no known affordable IR security sensors that are able to detect stationary targets or targets where temperature changes relatively slowly.

It is an object of the present invention to overcome the aforementioned problems and to provide the public with a useful alternative.

Accordingly the present invention provides for sensors that facilitate detection and recognition of targets that may be animate, such as human beings, or inanimate, such as a heated area in the field of view that may, for example, indicate the early stages of a fire.

It is a further object of the present invention to provide for sensors that improve on the currently ability to detect targets approaching the sensor (which are often undetected with current PIR sensors), and slow temperature rises related to, for example, electrical equipment failure or a developing fire.

SUMMARY OF THE INVENTION

Therefore in one form of the invention though this need not be the only or indeed the broadest form there is proposed an infrared sensor comprising: a plurality of resistance microbolometer infrared detectors electrically interconnected to form a sub-pixel; and a plurality of sub-pixels forming a focal plane array.

In preference at least some of the microbolometer infrared detectors are connected to each other in parallel.

In preference at least some of the microbolometer infrared detectors are connected to each other in series. It also may be that some are connected in series and some in parallel.

In preference said sensor further includes an infrared transmission optical system adapted to focus infrared radiation onto the focal plane array.

In a further from of the invention there is proposed an infrared sensor comprising: a focal plane array (FPA) including a plurality of pixels, whereby each of said pixels includes a mosaic of electrically interconnected sub-pixels, the mosaic of sub-pixels arranged to form a mosaic pixel focal plane array (MP-FPA) and each said sub-pixel being a resistance microbolometer infrared detector; a signal readout means; an infra red transmitting lens;

a signal processing means; and a display means.

In preference the focal plane array is mounted in a gas-filled or evacuated package fitted with an infrared transmitting window.

In preference a plurality of arrays can be fabricated on the same substrate by step-and- repeat pattern generation.

In preference said sensor further includes a storage means to store signals from the array.

In preference said sensor further includes an alarm to provide an alert when the signal from one or more pixels exceeds a predetermined threshold level above a predetermined ambient background signal.

In preference said sensor further includes an image display enabling the detected image to be displayed.

In preference the bolometer includes heat sensitive material selected from but not limited to undoped amorphous silicon and silicon alloy.

In preference the bolometer includes heat sensitive material selected from but not limited to nanocrystalline or microcrystalline silicon or silicon alloy.

In preference the optical system is a single or multi-element Fresnel lens or other plastic lens design.

In preference the optical system is a single-element germanium or chalcogenide glass lens with aspheric or binary surface geometry.

In preference the focal plane array is packaged at atmospheric pressure in air, nitrogen or xenon.

In preference the sub-pixel detector is an active microbolometer fabricated by thin film transistor technology

In preference the electronic readout or both readout and FPA is/are fabricated optionally in JFET or p-i-n diode technology

In preference the electronic readout or both readout and FPA is/are fabricated entirely with thin film transistor technology on silicon, glass or other suitable substrate.

In preference the sensor includes a matrix of sub-pixels microbolometer detectors, each pixel in the focal plane array comprising a number N sub-pixels, where N may range from 1 (single-detector pixel) to a number determined by the desired performance.

In preference the sub-pixels each may include a different number of bolometers.

In preference said sensor further includes an integrated CMOS or CCD daylight CCTV camera.

In preference said sensor further includes a light sensor to detect whether it is night or day.

In a further form of the invention there is proposed a method of detecting an infrared signal using an apparatus as defined in any one of the above claims and including the step of defόcusing or averaging the signal output from adjacent pixels.

In a yet further form of the invention there is proposed a resistance microbolometer IR-sensitive focal plane array (FPA) consisting of a plurality of imaging pixels, each imaging pixel in the FPA being comprised of a mosaic of sub-pixels formed by groups of detectors electrically interconnected, each sub-pixel including a thin-film temperature sensitive component, an integral infrared absorption mechanism, and a thermal isolation means.

In preference the FPA is formed on the same substrate including a readout integrated microcircuit (ROIC).

In preference said FPA is mounted in an evacuated or gas-filled package integrated with a infrared lens.

In preference said FPA is formed on a substrate selected from the group including but not limited to oxidised silicon and glass.

In preference said ROIC is formed during the same fabrication process as the FPA.

BRIEF DESCRIPTION OF THE DRAWINGS

The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate several implementations of the invention and, together with the description, serve to explain the advantages and principles of the invention. In the drawings,

Figure 1 is a schematic of an imaging array embodying the present invention and illustrating an imaging MP-FPA with each pixel comprising an array of sub-pixels, each sub-pixel being an individual detector element;

Figure 2 illustrates how groups of sub-pixels may be connected to form imaging pixels;

Figure 3 is a schematic diagram of an individual sub-pixel detector element; and

Figure 4 is a typical example of how a pixel may be electrically connected and readout out from the FPA.

DESCRIPTION OF THE PREFERRED EMBODIMENTS

The following detailed description of the invention refers to the accompanying drawings. Although the description includes exemplary embodiments, other embodiments are possible, and changes may be made to the embodiments described without departing from the spirit and scope of the invention. Wherever possible, the same reference numbers will be used throμghout the drawings and the following description to refer to the same and like parts.

Illustrated in Figure 1 is a MP-FPA 1 in both plan and cross-section comprising an array of 4x4 imaging pixels. Each pixel comprises an array of sub-pixels 2, each sub-pixel 2 being a functional IR detector 3, illustrated in cross-section. In this example there are 4x4 sub-pixels.

The FPA is formed on and integrated with a ROIC 4 and mounted in package 5. The package 5 has an IR transmitting window 6 and is coupled to an IR lens 7. In some instances, the lens 7 may fit directly to the package 5 so that the window 6 is not required. The FPA 1 may be packaged at atmospheric pressure in suitable filling gas such as nitrogen or xenon using simplified packaging. Where higher performance is required, and in preference it may be packaged in vacuum using wafer level packaging. .

Typically a number of FPA's will be fabricated at the same time on a single substrate. Different array formats may also be fabricated on the same substrate. In particular, it is advantageous to fabricate arrays with different numbers of imaging pixels, but the same

number of sub-pixels, the former varying only in the manner in which the sub-pixels are interconnected.

For example, a generic 160x120 format MP-FPA could have 160x120 pixels, each pixel comprising a single detector; alternatively 40x30 pixels, each pixel comprising 4x4 sub- pixel detectors.

In some applications it may be desirable to include two or more arrays of identical or different pixel format in the same sensor, in order to increase the overall field of view or have a higher resolution in some sectors. These multi-arrays may be diced from the same wafer.

The lens 7 is typically a single or multi-element plastic Fresnel lens or other plastic lens design or a single lens of germanium, silicon or a chalcogenide glass, with aspheric/binary surface geometry. The lens may be plano-convex, so designed with a short back focal length to enable mounting directly onto the FPA package.

The actual pixel format of an FPA depends on the specific application. The size and number of imaging pixels will be determined, respectively, by the target size and range, and the desired overall field of view.

For example, if it were desired to detect a target of dimensions Im x Im at a distance of 100m, then for a lens of focal length 20mm the required pixel size is 200μm x 200μm. For the array of Figure 1, the sub-pixel size would be 50μm x 50μm. The overall field of view would be 4m x 4m at 100 metre range, or an angular field of about 2.3 degrees. In practice there would be many more pixels in the array in order to achieve a suitably large field of view.

Figure 2 illustrates in plan view how sub-pixels may be interconnected, although other layouts are possible. Continuing with the 4x4 array example of Figure 1, thin-film electrical interconnect rails 8, 9, and 10 are biased at positive potential, and rails 11 and 12 at negative (ground) potential. At each of the sixteen sub-pixel detectors, contact electrodes 13 connect the detector body 15 to the interconnect rails 8, 9 and 10, and contact electrodes 14 connect the detector body to the interconnect rails 11 and 12; thus electrically the detectors are wired in parallel. The entire structure is supported above the ROIC by electrically conducting columns 16. It is possible with this specific design format to achieve a high optical fill factor, thus maximising the IR absorption area.

It is important to note that if there is a fault in one of the sub-pixels, the pixel will remain functional. This redundancy in operation is but one important feature of the invention, and has a significant impact on process yield, hence production cost.

A further advantage of using a parallel network is that the pixel. electrical resistance is much less than the sub-pixel resistance, which allows pixel resistance to be best matched to the ROIC input whilst using a higher resistivity detector material, and thus a higher Temperature Coefficient of Resistance (TCR). See also the discussion following equation (2) below.

The resistance network is, however, not limited to parallel connection. Optionally, it may be desirable to employ a series network, for example to increase resistance when employing a high conductivity detector material; and the best solution in specific FPA designs may be a parallel-series network.

Figure 3 illustrates a cross-section schematic of a sub-pixel detector. The detector body 3 is supported on conducting pillars 16, above the associated ROIC 4. The body 3 is comprised of a thin metal reflector 17, on which is deposited an insulating layer 18. Two metal contact electrodes 13 and 14 make electrical contact to a thin-film semiconductor heat- sensitive layer 19. The structure further includes a second insulator 20 and a thin top metal film 21. The conducting pillars 16 make contact with the ROIC 4 through metal pads 22. The .'detector stack', comprising reflector-insulator-semiconductor-insulator-top metal, forms an integral optical cavity tuned for maximum IR absorption at lOμm wavelength.

The detector body is formed by micro-engineering (MEMS) techniques on a ^• polyimide sacrificial layer (not shown), which is subsequently removed by plasma ashing.

In preference, the semiconductor layer is an amorphous silicon alloy or doped hydrogenated amorphous silicon, the two insulator layers are silicon nitride, and the metal contacts (which also form the thermal loss path from detector body to heat sink) are a 60:40 Ni-Cr alloy (nichrome) or titanium. The reflector is aluminium. The metal pillars are aluminium or a dual Ti:Pt metallisation. The top metal is nichrome, titanium or platinum. These layers will be deposited by plasma-enhanced chemical vapour deposition (PECVD) or physical vapour deposition (PVD). The layers are patterned by conventional microlithography and dry etching.

Amorphous silicon and its alloys can be modified to favourably enhance performance for security sensor applications, for example by varying electrical conductivity. Amorphous silicon-hydrogen alloy (a-Si:H) can be doped to increase conductivity allowing the desired electrical resistance to be obtained. There is a limit to the conductivity which can be achieved in this manner, but a higher conductivity may be obtained by forming an alloy of silicon with phosphorous or boron, where these elements are alloyed rather than act as dopants. Other alloys with silicon include tin, antimony, indium, germanium and gallium.

A further option is to deposit a doped or undoped a-Si:H layer in such a manner to . grow the crystal size, resulting in higher conductivity. In this form, the material is commonly called nanocrystalline or microcrystalline silicon.

There are two main reasons for seeking a higher conductivity semiconductor material: . to reduce the electrical resistance of bolometers to better match the readout circuit, and to reduce electrical noise, hence to achieve enhanced detective performance.

Figure 4 illustrates electronically how detector sub-pixels may be connected to form an imaging pixel. In the preferred embodiment, detectors will be connected in parallel. This results in reduced electronic noise whilst retaining a large absorbing area. Continuing the same example of Figure 1, the 16 sub-pixel detectors, Rbol, are shown in parallel connection, and are wired in series with a bias resistor, Rload. The bias voltage is applied when the switch Sl is closed. The parallel resistance network is connected via a DC blocking capacitance C to a low noise amplifier and signal conditioning circuit, labelled Amp, and the signal is read out by switch S2 and digitised. The bias may be direct current, alternating current or pulse. It will be understood that Figure 4 describes a generic electronic configuration, where a number of variations are possible.

The FPA is employed in conjunction with a Digital Signal Processor (DSP) to produce the desired form of output from the sensor. In a preferred embodiment, the signals from the ambient background will be integrated and stored, and a detection alarm will be recorded when the signal from one or more pixels exceeds a preset threshold above ambient. This may be accomplished by synchronous detection, but other methods will be evident to specialists in the field. At the same time, the output may be displayed as an image, enabling visual target recognition. In preference, a signal processing algorithm will be incorporated in the DSP, whereby the size, shape, speed and direction of movement are analysed for the

purpose of target recognition; and in preference the number of targets in the field of view will be counted and recorded.

Observers of IR sensors may not wish to see pixelated images. The quality of the image may be improved by removing high spatial frequencies, such as slightly de-focussing the optics, or averaging the output from adjacent pixels.

In a practical sensor installation, the IR sensor may be employed in conjunction with a visible wavelength CMOS or CCD CCTV camera for day and night operation. During night operation, the IR sensor may trigger an illumination source for higher resolution examination with the daylight camera. The IR sensor may be used in a fixed installation to monitor a specific target field such as a wall, doorway, path, gate or shop front; but in applications were a wide field of view is required, the sensor may be installed in a pan-tilt-zoom mechanism.

The IR sensor may also be coupled to a data storage device. This would enable a user to examine the data, such as an image, from the device after the event has occurred. For example, in the case of the sensor detecting an intrusion into an area, the operator may be able to review the data to determined what the intrusion was, for example a person or an animal.

In order to further understand the operation and advantages of this invention, it may be noted that for the 8 to 13μm spectral waveband the commonly used figure of merit for resistance microbolometers, the Noise Equivalent Temperature Difference (NETD), is given by:

where Vn is total electronic noise, Tno is the optical system T number (T/no), i#is the radiant responsivity, and Ad is the radiation receiving area of the pixel. The NETD should be as small as possible, so it is clear that the noise and Tno should also be as small as possible, whilst the responsivity and receiving area should be as large as possible. The Tno is usually chosen to be ~ 1.0 for optical reasons, and the area is determined by the requirement to resolve the target at the desired detection range.

For short range security sensors, one may compromise between angular resolution and detective performance, by making the area as large as possible whilst retaining the desired resolution. The area is typically in the range 0.01 to 1.0mm ² .

The low frequency responsivity, expressed in volts per watt, is given by:

K = ^≡- (2)

G where B is a circuit bridge factor, V is the bias voltage, a is the TCR of the temperature sensitive detector material, ε is the detector emissivity (equal to absorptance), and G is the thermal conductance defining heat loss between detectors and the nearby background.

One typically designs the sub-pixel detectors to optimise the values of B, V and G. The value of ε is determined by the design of the detector optical cavity, and is usually of the order 0.8. It is noted that Joule heating usually limits the value of V, but since the electrical resistance of individual sub-pixels may be quite high, the potential exists to use a higher bias voltage compared to other array technologies. Further, as one sees above, the use of a mosaic of sub-pixels provides a means to employ a detector material with a higher value of a. Thus enhancements in responsivity are possible with the sensor design described in this specification.

One of the objectives of this invention is to minimise the electrical noise, Vn in equation (1), by replacing a single-detector pixel with a mosaic of sub-pixels. The ultimate performance of bolometer detectors is set by thermal fluctuation noise; however, the dominant noise sources are usually Johnson noise and current (1/f) noise. Other minor noise sources such as shot noise may be present, but for the purpose of this discussion the total noise can be considered to be given by:

where Vj is Johnson noise and V 1/f is current noise, both expressed in units of V/VHZ, and Δf is the noise bandwidth (Hz).

If the electrical resistance^ R, of a single-detector pixel is, for example, the same as each of N sub-pixels, then the resistance of the composite pixel, when the sub-pixels are wired in parallel, is R/N, and the Johnson noise is reduced by 1Λ/N. Thus in Figures 1 and 2,

the resistance of the composite pixel will be 1/16 of a single-detector pixel, and the Johnson noise will be reduced by a factor 4.

In the case of current noise, according to the well-known Hooge relationship the noise is dependent on the electrical resistivity and volume of the temperature sensitive component, reducing with the square root of resistivity and the inverse square root of volume. For a given material, the volume should thus be as large as possible. It can be shown that the 1/f noise of the sub-pixel mosaic is 1/γN that of an individual sub-pixel and the same as a single-detector pixel having the same semiconductor thickness and surface area.

Thus for a given pixel area, the use of an MP-FPA design results in a substantial reduction in electronic noise, compared to that of a single-detector pixel of the same area, or that of a sub-pixel detector. In addition, electronic noise can be further reduced by using a semiconductor material of lower electrical resistivity (higher conductivity), as described above. This results in a lower resistance, hence reduced Johnson noise, and lower 1/f noise as a consequence of the Hooge relationship.

The noise bandwidth will be determined by the sampling (integration time) per pixel, which in turn depends on the number of pixels and the frame time. Thus a further advantage of an IR security sensor, where a given optical field of view is imaged by fewer and larger pixels than higher resolution imagers, is a reduced noise bandwidth, hence lower total noise.

The enhanced performance provided by an increase in responsivity, reduced noise and larger pixel area is a feature of this invention, and in some applications allows the specified performance to be achieved without evacuation, thus offering the potential for use of low cost gas-filled packaging.

The thermal speed of response of a bolometer detector is the ratio of thermal capacitance to thermal conductance. It can be readily shown that the response time of the parallel-connected sub-pixel mosaic is similar to that of a single-element detector of the same pixel size.

The bolometer element shown in Figure 3 is a conventional resistance microbolometer. In an alternative embodiment the sub-pixel mosaic may be formed with active microbolometers, which function as a thin film field effect transistor (TFT), where the temperature sensitive resistive component is the channel resistance of the transistor. Sub- pixels and readout components may advantageously be fabricated with the same TFT

technology on plain silicon, glass or other suitable substrate, giving significant economies in fabrication cost. The detector element and readout components may be optionally fabricated as thin film JFET and/or p-i-n diodes.

In summary, the present invention offers a number of advantages over existing technology including but not necessarily limited to: a. accommodating target size, shape, speed and direction of movement b. provision of an imaging capability c. target recognition d. very low false alarm rate e. detection of stationary and approaching targets f. detection of slowly heating or cooling targets g. array redundancy and low failure rate h. novel low cost fabrication

Further advantages and improvements may very well be made to the present invention without deviating from its scope. Although the invention has been shown and described in what is conceived to be the most practical and preferred embodiment, it is recognized that departures may be made therefrom within the scope and spirit of the invention, which is not to be limited to the details disclosed herein but is to be accorded the full scope of the claims so as to embrace any and all equivalent devices and apparatus.

In any claims that follow and in the summary of the invention, except where the context requires otherwise due to express language or necessary implication, the word "comprising" is used in the sense of "including", i.e. the features specified may be associated with further features in various embodiments of the invention.