GB2206447A - Lensed photodetector - Google Patents

Abstract

A photodetector has two regions 14, 24 sharing a common interface and of the same basic semiconductor material, one of the regions is configured as a lens and is degenerately doped to render it partially transparent to radiation absorbed at the interface or in the second region. The photodetector may be a CdxHg1-xTe photodiode pixel of a lensed photodetector array with x=0.26. The lens component 14 provides optical immersion of the pn junction 14/24. Degenerate doping of the lens component produces a Burstein-Moss shift in the absorption edge. The lens 14 accordingly transmits radiation to which it would be opaque with lower doping and for which the pn junction is absorptive. The need for photodiode heterostructures or microlenses is avoided and the structure provides a lensed photodiode array 10 of a single semiconductor material compatible with conventional device growth technology. Carrier tunnelling may be inhibited by a weakly doped n-type layer between the lens and a p region. The invention may be configured as a photoconductor of n<+> nu n<+>construction arranged for carrier exclusion from the nu -type region. <IMAGE>

Description

LENSED PHOTODETECTOR This invention relates to a lensed photodetector, ie a photodetector of the kind incorporating one or more lenses to improve sensitivity.

The use of lenses to improve the performance of radiation detectors is known. In Applied Optics, Vol 1, No 5, September 1962, pages 607-613, R Clark Jones discusses the potential gains obtainable by the use of detectors with optical immersion lenses. Three kinds of immersed detectors are considered, the first being an impractical variety in which the detector is at the centre of a spherical shell of high refractive index material. The second and third cases involve the detector being in contact with the plane surface of a hemisphere or hyperhemisphere. Clark Jones demonstrates that there is a gain in detecting performance due to two independent effects, (a) due to the optical immersion principle and (b) due to the use of a lens.In general, the effect of immersion of a detector using a lens of higher refractive index material is to increase the effective radiation sensitive area of the detector. The degree of improvement depends on the refractive indices of the lens and detector, and whether or not there is a cement layer of different refractive index between lens and detector.

In the Second International Conference on Advanced Infra Red Detector Systems, London, 1983, IEE Publication No 228, pages 4548, Carmichael, Wilson and Dean summarise the advantages of optical immersion. If a large detector is replaced by a smaller otherwise equivalent device presenting an equal apparent area by virtue of immersion, then: (1) Voltage responsivity is increased, (2) Noise equivalent power (NEP) is reduced, (3) A background limited (BLIP) detector of large apparent area is obtained, (4) Detectivity (D*) is increased for non-BLIP detectors, and (5) Photoconductive detectors exhibit increased frequency response.

"BLIP" represents background limited performance. It corresponds to a detector which is sufficiently low noise such that the noise on the detector output signal is dominated by photon noise from a thermal scene. This is the theoretical best performance limit, and is obtained by cooling a semiconductor detector to low temperature.

A typical example of a prior art optically immersed infra-red detector assembly is described in British Patent No 2,132,757B. It comprises a cadmium mercury telluride (CMT) detector disposed centrally of three lens support pads of the same material. The detector and pads are formed from the same piece of CNT material by an etching process which produces lens and pad upper surfaces which are accurately coplanar. A lens is attached to the pads by a very thin adhesive layer which spaces the lens from the pads by a distance less than one tenth of the wavelength of light.

The Carmichael et al device demonstrates a number of practical difficulties associated with implementing optical immersion. The lens must be transparent to radiation which the detector absorbs.

The lens and detector must therefore be of different materials.

The separation between the detector and lens must be less than one tenth of a wavelength of the relevant radiation. Preferably the lens and detector are in contact. These criteria are difficult to achieve, particularly so in the case of physically small arrays of detectors of easily contaminated material such as CMT.

Manufacturing complexity of lensed detector arrays is illustrated by Cross et al in Proc SPIE 1983, pages 73 to 77. This article describes a two-dimensional array of platinum silicide Schottky barrier detectors addressed by charge-coupled devices. The array is formed on the reverse face of a silicon chip, the opposite face of the chip being in contact with a radiation collecting faceplate.

The faceplate is configured as an array of contiguous partcylindrical lenses having parallel axes, each lens being arranged to converge radiation on a respective column of the array of detectors. The detectors are 55 pm by 59 pm rectangles with spacings between centres of 80 pm in each column and 160 pm in each row. The cylindrical lenses have a centre to centre spacing of 160cm to match that of the detectors in each row, and the lens radius of curvature is 194 pm in each case. Cross et al suggest manufacturing the faceplate using precision diamond machining processes, but they admit that this would be difficult at least for a faceplate less than 70 pm thick. They mention the possibility of forming the lenses in the silicon chip surface rather than employing a separate faceplate if a suitable indexing technique could be devised.This demonstrates the engineering problems associated with providing lenses for physically small detectors.

Lenses for radiation detectors while desirable are accordingly complex and expensive to produce. Ideally, the lens and detector would be implementable with modern layer growth technologies capable of defining both of these elements with high accuracy.

Machining or other expensive mechanical engineering at extreme accuracy would then be unnecessary. This approach has been adopted in the area of light emitting diodes (LEDs). In the IEEE Journal of Quantum Electronics, Vol QE-17, No 2, February 1981, pages 174178, Wada et al describe providing an array of LEDs with respective lenses. The array is a double heterostructure of InCaAsP/InP, where the InP layer is photochemically etched to provide lenses and the quaternary InGaAsP region provides the LED array. The reason for this complex double heterostructure construction arises from the requirement that the lens material must be transparent to radiation emitted by and therefore of wider bandgap than the LED region.Furthermore, the lens and LED regions must be chemically and crystallographically compatible to avoid unwanted chemical reactions, doping and crystal lattice distortion in material interface regions. These criteria very severely restrict choice of materials, and similar considerations apply to lensed detectors.

Generally speaking, heterostructure technologies are disadvantageous because of conduction and valence band discontinuities, lattice mismatch and chemical incompatibility.

Furthermore, heterostructures are physically more difficult to grow and are more expensive than homostructures, ie devices comprising a single semiconductor material. At first sight, alloy semiconductors such as CdxHgl~xTe would appear to offer a convenient route to forming heterostructures, since only a change in the compositional parameter x is required. However, in practice, a change in doping is also required. Accordingly, simultaneous control over two growth parameters is necessary, which gives rise to considerable practical difficulty in conventional sliding boat liquid phase epitaxy.

It is an object of the present invention to provide a lensed infrared detector wherein the lens and detector are formed from the same semiconductor material.

The present invention provides a photodetector having a first region sharing a common interface with a second region, the regions being of the same semiconductor material, and wherein the first region is configured as a lens arranged to converge radiation on the second region and is degenerately doped to provide a Burstein Moss shift rendering it at least partially transparent to radiation absorbed in the interface or the second region.

The invention provides the advantage that the lens and detector are of one material, the lens being rendered transparent to radiation by degenerate doping. It is accordingly unnecessary to provide a separate lens of a different material with associated constructional complexity and expense. The invention is well suited to production by conventional semiconductor device technology.

In a preferred embodiment, the invention is formed of semiconductor material having a band gap of less than 0.5eV and comprising at least two constituent elements. For detection of wavelengths in the range 3 pm to 5 um, the material may be CdxHgl~xTe with x = 0.26 and an n-type region doping of 1 x 1018 cm 3 to 5x1018cm 3, preferably 2.0 + 1.0 x 1018 cm 3. For detection of wavelengths in the range 8 um to 11 pm, cadmium mercury telluride may be employed once more but with x = 0.19 and an n-type doping level of 1 x iso17 cm 3 to 5 x 1018 cm 3 preferably 4.0 + 2.0 x 1017 cm~3.

The invention may be arranged as a two-dimensional pixel array in which each pixel comprises a respective lensed n-type first region surmounting a second region. The array may be arranged in the focal plane of an objective lens of F-number between 0.5 and 2.0.

The second region may be p-type, the interface being a pn junction at which radiation is absorbed. The second region may alternatively comprise a weakly doped n-type layer adjacent the lens region and a p-type layer, thus providing the second region with an internal pn junction for radiation absorption with carrier tunnelling being inhibited by the n-layer.

The invention may also provide a photoconductive detector arranged for elevated temperature operation by means of carrier exclusion.

In this embodiment the second region comprises a v-type (intrinsic but residual n-type) layer and a p-type layer, the v -type layer being sandwiched between the lensed first region and the p-type layer. Under the action of bias voltage carrier exclusion takes place in the v -type layer. This layer accordingly simulates extrinsic characteristics required for photoconductive purposes, but normally obtainable only by cooling to much lower temperatures.

Carrier exclusion in the v-type layer takes place in a region of the order of an optical absorption length in thickness, and within which photons are absorbed. Photon absorption changes the detector resistance and current, the latter being monitored to provide an indication of light intensity.

In order that the invention might be more fully understood, one embodiment thereof will now be described, by way of example only, in which: Figures 1 and 2 are respectively schematic side and plan (underside) views of a photodetector of the invention comprising a 3 x 2 pixel array; Figure 3 is an energy band diagram illustrating the Burstein-Moss shift in a degenerately doped n-type semiconductor material; Figure 4 shows graphs of semiconductor absorption against wavelength for different doping levels illustrating Burstein-Moss shift; Figure 5 shows a photodetector of the invention arranged in the focal plane of an objective lens; Figure 6 shows in more detail part of a photodetector of the invention; Figures 7 and 8 schematically show embodiments of the invention modified to inhibit tunnelling;; Figure 9 schematically shows a photoconductive detector of the invention arranged for carrier exclusion; and Figure 10 is an energy band diagram illustrating the Burstein-Moss shift in a degenerately doped p-type semiconductor material.

Referring to Figures 1 and,2, in which like parts are likereferenced, there are shown two schematic representations of an infra-red detector 10 of the invention. Figure 1 is a sectional side view and Figure 2 is a plan view of the underside of the Figure 1 device.

The detector 10 is a three by two array of like pixels such as 12.

Each pixel 12 consists of a predominantly square section component or first region 14 on a square base 16 having sides of length 55 um in both cases. The component 14 has an upper surface 18 configured as a spherical cap of radius of curvature 45 pm, and is 61 cm in height. The surface 18 has a centre of curvature 20 which is located 16 um vertically above the base centre 22. The component 14 is of degenerately doped n-type or n+ CdxHgl~xTe (CMT) material.

The dopant is aluminium with a concentration of 2 x 1018 cm 3 and the compositional parameter x is 0.26.

A rectangular block or second region 24 having dimensions 12 llm x 12 urn x 3 um has one 121l m x 12 pm face (not shown) in contact with the base 16. The block 24 has 12 pm sides parallel to those of the base 16, and is arranged symmetrically about a vertical line through the base centre 22. The block 24 is of the same semiconductor material as the component 14, ie Cd Hgl,,Te with x = 0.26. However, the doping of the block 24 is weak p-type; it has an arsenic dopant concentration of 1 x 1015 cm 3.

Each pixel 12 has a planar electrode layer 28 on the block 24. The electrode layer 28 is of gold, which forms an Ohmic contact to CMT.

An electrical connection (shown in Figure 1 only) 32 is provided to the electrode 28. A single large gold electrode 34 and a connection 36 are provided for the detector 10. By virtue of the degenerate doping of the components such as 14, their conductivity is metallic and they require only a single common connection.

The detector 10 operates as follows. It is designed to detect radiation in the 3 pm to 5 pm infra-red wavelength band. The pn junction interface between the block 24 and component 14 of each pixel is a homojunction, ie it is a junction between like semiconductor materials which differ only in their dopant species and concentrations. Infra-red radiation passing through lensshaped upper surface 18 of each pixel 12 is converged on to the pn junction interface 14/24. The component 18 is at least partly transparent to radiation in the 3-5 pm band by virtue of the Burstein-Moss absorption edge shift arising from degenerate n-type doping to be described later.The p-type component 24 absorbs the radiation to generate free charge carriers which may be swept out by a reverse bias voltage applied across electrodes 28 and 34.

Alternatively, the photovoltaic mode of operation may be employed, in which the pn junction is unbiased. In this case, charge carriers liberated by incident radiation are swept out by the internal field across the pn junction.

The spherically surfaced pixel component 14 transmits infra-red radiation in the 3 pm to 5 pu wavelength interval only because it is degenerately doped. If it were to have appreciably lower doping, it would absorb this radiation, and pre-empt absorption in and detection by the p-type component 24. The reason for this is that, in a heavily doped semiconductor, the energy required to activate a valence electron to the conduction band is increased to a value greater than the energy gap EG. This is because degenerate doping results in filling of electron states at the bottom of the conduction band. Accordingly, the conduction band states available to receive valence band electrons are at least predominantly those well above the conduction band minimum. This is known as the Burstein-Moss effect, and is described inter alia by Dingrong et al, Solid State Communications, Vol 56, No 9, pages 813-816, 1985.

Referring now to Figure 3, there is schematically shown the energy band structure for a heavily doped semiconductor material exhibiting the Burstein-Moss effect. This drawing shows the conduction band 40 with a minimum at 42, together with the light and heavy hole valence bands 44 and 46 with a common maximum at 48.

The semiconductor material is degenerately doped n-type (n+); and conduction band states are filled up to the Fermi level 50 at energy EF above the valence band maximum 48.

By virtue of conduction band filling, electrons excited from the valence band to the conduction band must receive energy at least equal to EF. If the semiconductor material were to have been lightly doped, electrons could have been excited to the conduction band with an energy of EG. The effect of increasing the doping level from light to degenerate is therefore to shift the optical absorption edge from a photon energy of EG to EF. This is known as the Burstein-Moss shift equal to EF-EG. It ignores valence band curvature which increases the energy shift.

Referring now to Figure 4, there are shown two graphs 60 and 62 of optical transmission (arbitrary units) against infra-red wavenumber (cam 1)* Wavenumber increases to the left and the scale is halved to the left of the ordinate axis as compared to that to the right, this being a feature of the output of the spectrometer employed.

Wavelengths of 3 ijm and 5 pm are indicated by arrows 64 and 66 respectively.

Both graphs 60 and 62 were obtained using CdxHgl-xTe material with x = 0.26. Graph 60 corresponds to material with an n+ doping of 1 x 1018 cm 3 of aluminium, and graph 62 corresponds to p-type material. The effect of introducing degenerate or n+ doping is to produce a Burstein-Moss shift of the absorption edge from the 5 pm region to the 3 pm region.

Figure 4 demonstrates that degenerate doping of lensed component 14 in Figures 1 and 2 renders it at least partially transparent to radiation absorbed by the same material with p-type doping. In particular, the component 14 transmits radiation absorbed by the p region 24. The detector 10 accordingly exhibits the advantage that transparency of the lens component required to increase responsivity is obtained without the difficulties associated with heterostructure semiconductor technologies or with attaching miniaturised discrete lenses to a detector array. The detector lû may accordingly be produced as a homostructure of a single semiconductor material with doping change between parts 14 and 24.

The lensed component 14 has a refractive index of 3.4 allowing for Burstein-Moss shift. Providing a detector lens increases the apparent area presented to radiation by the detector. The advantage of reducing the actual detector area while retaining its apparent area with a lens is that thermal noise is reduced.

Thermal noise is proportional to the square root of detector volume. Since detector thickness is largely fixed as that required to absorb the majority of incident photons, the internal noise of two detectors of dissimilar areas will be in the ratio of the square root of their area ratio if they are otherwise equivalent.

The detector 10 may be produced by known CMT layer growth techniques. A continuous p-layer may be laid on an n+ substrate, the p-layer being selectively etched to define blocks 24. The substrate is then ion-etched to define lens surfaces 18. To achieve this, a photoresist layer is deposited on the substrate and configured with varying thickness resembling the required lens array. The photoresist layer is then ion-etched away and its surface becomes reproduced by that of the substrate.

The detector 10 of Figures 1 and 2 is shown as 3 x 2 array for reasons of illustrational clarity. In practice, a small picture of a scene on a television monitor would require a 64 x 64 pixel array. Such an array would be employed with an objective lens.

Referring now to Figure 5, there is shown a 64 x 64 pixel array 70 arranged to receive radiation from a scene via an objective lens 72. The lens 72 is a concavo-convex meniscus of focal length locum and diameter 7cm, which provides an F-number of about 1.4.

Generally speaking, an F-number betweem 0.5 and 2 is desirable.

The pixel array 70 is shown drawn approximately to scale with respect to the lens 72, and is located in the lens focal plane 74.

Each pixel of the array 70 has the same construction and dimensions as pixel 12 in Figures 1 and 2.

The arrangement illustrated in Figure 5 produces a factor of about 4 or about 300% improvement in D* when the array 70 is operated at temperatures for which thermal noise in the detector material is the dominant source of noise. This substantial improvement is as compared to an otherwise equivalent pixel array incorporating 55 lem x 55 Pm radiation sensitive pixels without individual pixel lenses such as 14 in Figure 1. Use of individual pixel lenses accordingly allows shrinkage in radiation sensitive pixel area with consequent noise reduction without corresponding reduction in radiation collection efficiency.

The foregoing improvment in D* may be further enhanced by the use of an objective lens 72 of larger F-number. Moreover, additional improvement is obtainable by increasing the curvature of individual pixel lens surfaces 18 in Figure 1, ie reducing their radius of curvature.

The invention is most relevant to detectors formed from semiconductor materials exhibiting an appreciable Burstein-Moss shift. The shift in terms of optical absorption edge change measured as a wavelength interval is more pronounced for narrow gap semiconductors having a band gap EG less than 0.5eV.Such materials comprise the binary, ternary or higher order compounds of the group Ill-V, II-VI, IV-IV, and IV-VI varieties, of which the more important are set out in Table 1.

4 GROUP BINARY TERNARY QUARTERNARY Ill-V InSb InAsSb InGaAsSb InAs InSbBi II-VI HgCdTe PhSnSeTe HgZnTe HgCdZnTe HgMnTe PbSnTe IV- IV SnGe SnSi IV-VI PbS PbSe PbTe

TABLE 1 In the foregoing specific embodiment intended for detection of radiation in the 3 pm to 5 pm wavelength interval, the detector material was CdxHgl~xTe with x = 0.26 and an n-region doping level of 2 x 1018 cm 3. A doping level within + 10% of this value is preferred. More generally, for this material and wavelength interval the doping level may be in the range 1 to 5 x 1018 cm 3.

CMT may also be employed to detect radiation in the 8-11 pm wavelength interval. For this interval the compositional parameter x is equal to 0.19 (2000K operating temperature), and the n-region doping level is in the range 1 x 1017 cm 3 to 5 x 1018 cm 3, preferably 4.0 + 2.0 x 1017 cm 3.

Referring now to Figure 6, which is not to scale, there is shown a detailed practical embodiment of part of a photodetector of the invention. Parts equivalent to those described with reference to Figures 1 and 2 are like referenced with the prefix 100. An n first region 114 having a lensed upper surface (not shown) is contiguous with a p region 124. A sputtered ZnS insulating layer 126 coats undersurface areas of the n+ region 114 and the p region 124. It also defines a contact window 140, and is partly coated with a thin Ni/Cr layer 142. A gold contact layer 144 is deposited over the Ni/Cr layer 142, and bears an indium solder bump 146. A contact pad 148 of a readout circuit indicated by 150 contacts the solder bump 146. The readout circuit 150 may be an array of MOSFET switches, or charge coupled device readout means of known kind.

The layers of ZnS 126, Ni/Cr 142 and gold 144 are well known in the art of CWT photodetectors and will not be described further. The Ni/Cr layer 142 in particular is a known pretreatment for seeding growth of a gold layer.

Referring now to Figures 7 and 8, which are not to scale, there are schematically shown alternative embodiments 160 and 170 of the invention. These embodiments are equivalent to a pixel 12 of Figures 1 and 2 with an n layer interposed between n+ region 14 and p region 24. Pixel 160 has an n layer 162 between n+ region 154 and p region 166, and pixel 170 and n layer 172 between n+ region 174 and p region 176. The pixels 160 and 170 differ only in that n layer 162 extends the full width of n+ region 164, whereas the extent of n layer 172 is that of p region 176. The purpose of n layers 162 and 172 is to inhibit tunnelling of electrons across the band gap under reverse voltage bias in the depletion region of a n+p junction. This is a known effect in pn junctions where the n region is heavily doped, and is inhibited by interposing a weakly doped n layer.The n layer 162 or 172 should have n-type doping in the range 1014 - 1015 cm 3 and should be 2 + 0.5 Pm in thickness.

Carrier tunnelling of this kind is discussed inter alia by Anderson in Infrared Physics, Vol 20, pp 353-361 (1980).

Embodiments of the invention previously described have been photodiodes, ie devices each including a pn junction.

Photoconductive detectors in accordance with the invention may also be provided. Referring now to Figure 9, which is no. to scale, a photoconductive detector 180 of the invention is shown schematically. The detector 180 comprises an n+ lensed region 182 equivalent to part 14 of Figure 1. This surmounts a weakly doped n-type or v region 184, which is 10Pm in thickness and has a residual doping level < 1015 cm 3. This compares with an intrinsic contribution to conductivity or carrier concentration of 1016 cm 3 at 200 K, so intrinsic conduction predominates.

The v region 184 in turn surmounts a further n+ region 186 having a gold contact 188. The detector 180 accordingly has an n+vn+ structure, and is formed from CdxHg1#xTe with x = 0.19. The n regions have doping levels in the range 1017 - 5 x 1018 cm 3, as previously described for part 14. The detector 180 is designed for operation in the 8 to 10pm wavelength interval at 200 K. If the compositional parameter x is changed to 0.26, the operating wavelength becomes 3-5#m at ambient temperature, but n+ doping levels of 1 - 5 x 1018 cm 3 are required.The operating temperatures correspond to greatly reduced cooling requirements, since CMT photoconductive detectors normally require cooling to 800K and 2000K for operation in the 8-1Opm and 3-511m wavelength intervals respectively.

The detector 180 operates as follows. The v region 184 has predominantly intrinsic characteristics in the absence of bias current. It is envisaged that the photodetector 180 is part of an array of like pixels as illustrated in Figures 1 and 2 with a common electrical connection (not shown) to the lensed region 182 equivalent to connection 34/36. In operation, the common connection and lensed region 182 are biassed positive. This creates a dynamic situation at the n+ interface 182/184 wherein electrons (majority carriers) may flow freely from 184 to 182 but only a very small hole or minority carrier current may flow from 182 to 182. The n+ interface is therefore an excluding contact to region 184, a known form of contact having properties discussed inter alia in published European Patent Application No. 0167305.

The properties of the excluding n+ contact under bias denude the v region 184 of holes (minority carriers). Space charge considerations require a corresponding fall in the intrinsic contribution to the electron or majority carrier population, this contribution being required to remain equal to the minority carrier population. The result is that biasing n+ region 182 positive with respect to n+ region 186 results in a substantial reduction in the intrinsic contribution to conduction in v region 184, which accordingly simulates extrinsic behaviour by virtue of carrier exclusion. Whereas intrinsic characteristics predominate in the absence of bias, carrier exclusion reduces the intrinsic contribution below the extrinsic equivalent arising from doping.

The v region accordingly simulates the behaviour of an equivalently doped like material cooled to much lower temperature to reduce intrinsic conduction.

Carrier exclusion does not take place throughout the whole of v region 184. It occurs in a volume indicated at 190 extending from the 182/184 interface to a chain line 192 distant 3pm from the interface. Carrier exclusion is greatly beneficial in photoconductive detectors, since it permits detector operation at higher temperatures with reauceci cooling requirementse In view of its lower doping concentration and carrier exclusion effects, the resistance of the v region 184 is the predominant contribution to the resistance of the detector 180. Infrared radiation converged on the region 184 by the lensed region 182 reduces this resistance measurably, and the degree of reduction indicates infrared intensity.The thickness of the exclusion region 190 is in the order of an optical absorption length, so it absorbs most of the radiation transmitted by the lensed region 182 in the appropriate wavelength iatesvalo Referring now to Figure 10, there is shown an alternative energy band structure indicated generally by 200 which may be employed in devices of the invention. This drawing shows a conduction band 202 with a minimum at 204, together with a single valence band 206 with a maximum at 208. The structure has a Fermi level 210 below the valence band maximum 208, which corresponds to a degenerately doped p-type semiconductor material. A shaded region 212 denotes wholly empty electron states in the valence band. The structure 200 is typical of semiconductors such as PbSnTe.

Since electron states in the region 212 are unpopulated, electrons may only be excited to the conduction band 204 from states such as 214 in the valence band 206 below the Fermi level 210. Comparison of Figures 3 and 5 demonstrates that an absorption edge shift may be produced by degenerate p-type doping as well as degenerate ntype doping. Accordingly, the conductivity types of the semiconductor parts in Figures 1, 2 and 6 to 9 may be reversed. In particular, in Figure 1, lens component 14 may be degenerate p-type or p+ and block 24 may be n-type. For the purposes of this specification, the expression "Burstein-Moss shift" shall be construed as embracing absorption edge shift by degenerate doping by either n-type or p-type dopants.

Claims (12)

1. A photodetector having a first region sharing a common interface with a second region, the regions being of the same semiconductor material, and wherein the first region is configured as a lens arranged to converge radiation on the second region and is degenerately doped to provide a Burstein Moss shift rendering it at least partially transparent to radiation absorbed in the interface or the second region.

2. A photodetector according to Claim 1 wherein the semiconductor material has at least two constituent elements and has a band gap of less than 0.5eV.

3. A photodetector according to Claim 1 or 2 arranged as one pixel of an array of like pixels.

4. A photodetector according to Claim 1, 2 or 3 of CdxHgl~xTe wherein x = 0.26 and the first region is n-type with a doping level in the range 1 x i018 cm 3 to 5 x 1018 cm 3.

5. A detector according to Claim 4 wherein the doping level is in the range 2.0 + 1.0 x 1018 cm 3.

6. A photodetector according to Claim 1, 2 or 3 of CdxHgl~xTe wherein x = 0.19 and the first region is n-type with a doping level in the range 1 x 1017 cm 3 to 5 x 1018 cm 3.

7. A photodetector according to Claim 6 wherein the doping level is 4.0 + 2.0 x 1017 cm3,

8. A photodetector according to Claim 1 wherein the second region consists at least partially of p-type material.

9. A photodetector according to Claim 8 wherein the second region includes an n-type region of lower doping than the first region between the first region and the p-type material and arranged to inhibit tunnelling.

10. A photodetector according to Claim 1 wherein the second region comprises v-type and n±type layers, the w-type region being arranged for carrier exclusion.

11. A photodetector according to any preceding claim arranged in the focal plane of an objective lens having an F-number in the range 0.5 to 2.

12. A photodetector substantially as herein described with reference to the accompanying drawings.