Photodetector comprising a photonic crystal structure optically coupled to an active layer with improved quantum efficiency

Abstract

The invention relates to a photodetector comprising a support layer (10), a lattice structure (20) and an active layer (30). The lattice structure (20), referred to as photonic crystal structure, comprises a structured layer (22) formed of patterns (22.1) and of a continuous sublayer (22.2) with a non-zero constant thickness eccs, the patterns (22.1) widening in a direction oriented towards the continuous sublayer (22.2). Moreover, the photonic crystal structure (20) has dimensions such that the structured layer (22) forms a photonic crystal that supports at least one guided mode able to be excited by the incident light radiation of interest, the photonic crystal being optically coupled to the active layer (30) such that the guided mode is optically confined in the structured layer (22) and the active layer (30).

Description

[0001]PHOTODETECTOR COMPRISING A PHOTONIC CRYSTAL STRUCTURE OPTICALLY COUPLED TO AN ACTIVE LAYER WITH IMPROVED QUANTUM EFFICIENCY TECHNICAL FIELD The field of the invention is that of photodetectors, for example matrix photodetectors, comprising a grating structure and an active layer adapted to absorb the light radiation to be detected. STATE OF THE PRIOR ART Photodetectors of electromagnetic radiation are adapted to receive and detect the electromagnetic radiation by converting it into a light or electrical output signal, such as, for example, an output signal formed from charge carriers photogenerated in an active layer. Document WO2014 / 056550 describes an example of an electronic radiation photodetector, such as an image intensifier tube or a photomultiplier tube.As illustrated in Figure 1, the photodetector A1 comprises an input window A10 transparent to the radiation to be detected, and a photoemissive active layer A30 arranged on a rear face of the input window A10. The latter is adapted to receive the incident electromagnetic radiation and to emit in response a flux of photoelectrons. An output device (not shown) receives the photoelectrons and generates in response an output signal, here after having multiplied the photoelectrons into a flux of secondary electrons. The quantum efficiency of the active layer A30 corresponds to the ratio between the number of photoelectrons emitted and the number of incident photons received. It depends in particular on the absorption rate of the incident photon, the transport rate of the photoelectron to the emission face (rear face of the active layer), and the emission rate of the photoelectron out of the active layer A30.Since the absorption and transport rates depend on the thickness of the active layer A30 according to opposite trends, namely that the absorption rate increases with the thickness while the transport rate decreases, there is therefore an optimal thickness allowing to maximize the product of these two rates. However, it appears that the absorption rate decreases sharply for long wavelengths, in particular here from about 800nm. To improve the performance of the photodetector A1, a natural solution would then be to increase the thickness of the active layer A30 to increase the absorption rate, but, as we have seen, this would lead to a degradation of the transport rate. Also, the photodetector A1 comprises a grating structure A20 forming a transmission diffraction grating, located between the entrance window A10 and the active layer A30.Also, an incident photon transmitted by this diffraction grating penetrates the active layer A30 with a non-zero diffraction angle with respect to the normal to the front face of the active layer A30. The apparent thickness of the active layer A30, seen by the photon, is therefore higher, which improves the absorption rate without degrading the transmission rate. Thus, the quantum efficiency of the active layer A30 is improved. However, there is a need to improve the performance of such a photodetector, and more broadly of any photodetector formed of a transparent support layer, a grating structure, and an absorbing active layer.STATEMENT OF THE INVENTION For this, the subject of the invention is a photodetector comprising a transparent support layer, a network structure (photonic crystal structure) and an absorbent active layer, this photodetector having improved performance, in particular due to an increase in the quantum efficiency of the active layer.For this, the subject of the invention is a photodetector suitable for detecting light radiation in a predefined detection spectral band, comprising a stack formed of: o a support layer, made of a transparent material, suitable for receiving the light radiation via a front face and transmitting it via a rear face; o a network structure, arranged on the rear face of the support layer, comprising patterns arranged in a main plane parallel to the front face of the support layer; o an active layer, arranged on a rear face of the network structure, made of a semiconductor material suitable for absorbing the light radiation.According to the invention, the grating structure comprises a structured layer, which is formed from said patterns and a continuous sub-layer of constant thickness eccs not being zero, the patterns being oriented towards the support layer and having a profile widening along a direction oriented towards the continuous sub-layer. In addition, the grating structure, called a photonic crystal structure, has dimensions such that the structured layer forms a photonic crystal which supports at least one guided mode capable of being excited by the incident light radiation of interest, the photonic crystal being optically coupled to the active layer so that the guided mode is optically confined in the structured layer and the active layer. Some preferred but non-limiting aspects of this photodetector are as follows. The patterns may have a height h, between their top and their base, of between one tenth and three quarters of the thickness e. ccsof the continuous sub-layer of the structured layer. The photonic crystal structure may comprise a so-called filler layer, located between the support layer and the structured layer, made of a dielectric material having a refractive index n L less than a refractive index n H of the structured layer and presenting a gap n H -n L at least equal to 0.2. The refractive index nL of the filling layer may have a deviation in absolute value |n L -n cs | with a refractive index of the support layer at most equal to 0.1. A refractive index nH of the structured layer may have a deviation in absolute value |n H -n ca | with a refractive index of the active layer at most equal to 0.5. The sum of a height h of the patterns, of a thickness e ccsof the continuous sub-layer of the structured layer and a thickness eca of the active layer, can be at least equal to λr / (2×<n’H> ), where λr is a wavelength of the guided mode, and where<n’H> is an average index equal to an average of the refractive indices of the structured layer and the active layer weighted by the respective thicknesses. The patterns may have a parabolic profile. The active layer may have a rear face opposite the photonic crystal structure (20), in contact with a vacuum. The filling layer may be made of an oxide, or even a silicon oxide. The active layer may be a photoemissive layer made of a material based on SbNaKCs, SbNa2KCs, SbNaK, SbKCs, SbRbKCs or SbRbCs, or of a material based on AgOCs. The photodetector may be an image intensifier tube or a photomultiplier tube.The invention also relates to a method for manufacturing a photodetector according to any one of the preceding characteristics, comprising the following steps: o producing a stack formed of a thick layer, a so-called filling layer, and an upper thin layer; o producing through-openings in the upper thin layer and opening onto the filling layer, arranged regularly at the pitch p of the patterns; o partial isotropic wet etching of the filling layer from the through-openings, so as to form periodic etched zones intended to form the patterns; then removing the upper thin layer; o depositing on the filling layer a layer forming the structured layer; o depositing on the structured layer the active layer; o removing the thick layer, and assembling the support layer on the face made free of the filling layer.BRIEF DESCRIPTION OF THE DRAWINGS Other aspects, aims, advantages and characteristics of the invention will appear better on reading the following detailed description of preferred embodiments thereof, given by way of non-limiting example, and made with reference to the appended drawings in which: Figure 1, already described, is a schematic and partial view, in cross-section, of a photodetector according to an example of the prior art; Figure 2A is a schematic and partial view, in cross-section, of a photodetector according to one embodiment; Figure 2B illustrates in more detail the photonic crystal structure of the photodetector of Fig. 2A; Figure 2C is a top view of the structured layer of the photodetector of Fig. 2A; Figure 3A illustrates examples of absorption and reflection spectra of a photodetector similar to that of Fig.2A; Figure 3B illustrates examples of absorption and reflection spectra of a photodetector according to a variant of that of Fig. 2A; Figure 3C illustrates an example of power spectral density of light radiation in night vision; Figures 4A to 4F illustrate steps of a method of manufacturing a photodetector similar to that of Fig. 2A. DETAILED DESCRIPTION OF PARTICULAR EMBODIMENTS In the figures and in the remainder of the description, the same references represent identical or similar elements. In addition, the different elements are not shown to scale so as to favor the clarity of the figures. Furthermore, the different embodiments and variants are not mutually exclusive and can be combined with each other. Unless otherwise indicated, the terms “substantially”, “approximately”, “of the order of” mean to within 10%, and preferably to within 5%.Furthermore, the terms "between ... and ..." and equivalents mean that the limits are included, unless otherwise stated. The invention relates to a photodetector adapted to detect light radiation in a detection spectral band Δλ. detpredefined. It comprises a stack formed of, along a main direction, a transparent support layer, a grating structure (hereinafter called 'photonic crystal structure'), and an active layer adapted to absorb the incident light radiation. The main direction is a direction parallel or inclined with respect to the normal to the front face of the transparent support layer through which the light radiation to be detected enters the photodetector. It also defines the optical axis of the photodetector along which the incident light radiation propagates. A main plane can be defined as a plane parallel to the front face of the transparent support layer. In the remainder of the description, the main direction is parallel to the main plane, but an inclination of a few degrees, preferably at most 10°, remains possible.As detailed below, the performance of the photodetector is improved by the fact that the photonic crystal structure has dimensions such that the structured layer forms a photonic crystal that supports at least one guided mode capable of being excited by the incident light radiation of interest, which extends in the plane of the photonic crystal. In addition, the photonic crystal is optically coupled to the active layer, so that the guided mode is optically confined in the photonic crystal and the active layer. For this, the guided mode is excited at normal or inclined incidence (along the optical axis) by a so-called guided-mode resonance (GMR) effect, at a predetermined large wavelength of the detection spectral band Δλ. det , called the resonance wavelength λ r, where the absorption rate of the active layer is usually relatively low in the absence of such a photonic crystal structure. The absorption rate is therefore particularly high at this resonance wavelength λ r, while the transport rate remains unchanged. The photonic crystal is formed by the patterns of a structured layer made of a dielectric with a high refractive index nH. These patterns are preferably arranged periodically, but a quasi-periodic or even non-periodic arrangement, such as correlated disorder, is also possible. The photonic crystal gives rise to band gaps for the propagation of light in certain directions and at certain wavelengths, similar to the electronic band gaps in semiconductors. The photonic crystal is preferably two-dimensional in the sense that the patterns are arranged along two axes distinct from the principal plane in which the structured layer extends. Furthermore, as described later, it is possible to slow down the light in the photonic crystal, and it is confined laterally in the plane of the structured layer.As detailed below, the performance of the photodetector is improved by the fact that the reflection rate, defined at the interface between the support layer and the photonic crystal structure, is reduced. This is achieved by the fact that the patterns of the photonic crystal, which extend vertically from a continuous sub-layer of constant thickness towards the support layer, have a profile that widens along the main direction, so that the fill factor associated with the patterns increases along the main direction from a zero value to at least 80%, or even at least 90% or more. By this profile of the patterns, which is preferably parabolic, the refractive index, associated with the patterns (and therefore their fill factor) (explained below), has a value that increases along the main direction, going from a low value n. Lsubstantially equal to the refractive index of a filler layer and the support layer up to a high value substantially equal, at least 10%, to the high index nH of the material of the structured layer of the grating structure. It appears that this reduction in the reflection rate is broadband, which improves the performance of the photodetector. Figure 2A is a schematic and partial cross-sectional view of a photodetector 1 according to one embodiment. Figure 2B illustrates a portion of the photonic crystal structure 20, and Figure 2C is a top view of the structured layer 22 of the photonic crystal structure 20. In this example, the photodetector 1 is an image intensifier tube (e.g., EB-CCD or EBCMOS type) or a photomultiplier tube, and the active layer 30 is a semi-transparent photocathode. The light radiation to be detected has a spectral range that covers the detection spectral band Δλ det, which here ranges from 0.5 µm to approximately 1 µm. According to the invention, the photodetector 1 comprises at least a stack formed, along the main direction, of a transparent support layer 10, a photonic crystal structure 20, and an active layer 30. An output device (not shown) is located downstream of the active layer 30, and is adapted to receive the charge carriers photogenerated in the active layer 30 to generate in response an optical or electrical output signal. In the case here where the photodetector 1 is an image intensifier tube or a photomultiplier tube, the output device may then comprise a photoelectron multiplier device (e.g. a microchannel plate) followed by a phosphor screen and / or, where appropriate, a CMOS or CCD matrix sensor. However, the invention is not limited to this type of photodetectors or to photocathodes.Here and for the remainder of the description, a direct three-dimensional orthogonal XYZ reference frame is defined, where the X and Y axes form a main plane along which the layers extend, and where the Z axis is oriented along the thickness of the layers. The +Z direction is oriented from the support layer 10 towards the active layer 30. It forms the main direction along which the light radiation to be detected enters the photodetector (the main direction may, however, be inclined with respect to the normal to the main XY plane, preferably by at most approximately 10°). Furthermore, the terms “front” and “rear” are understood as relating to an increasing positioning along the main +Z direction. The support layer 10 forms an entry window for the light radiation to be detected in the photodetector 1.It has a front face 10av by which the light radiation is received, and a rear face 10ar, opposite the front face 10av, by which the received light radiation is transmitted towards the photonic crystal structure 20. The front faces 10av and rear faces 10ar are here substantially parallel to the main plane XY. The support layer 10 is made of a material transparent in the detection spectral band Δλ. det , that is to say that its transmittance (transmission rate) is very high, for example close to 100%. It is made of a material chosen according to the spectral range of the light radiation to be detected. In the detection spectral band Δλ detranging from approximately 0.5 to 1 µm, the material may be, for example, quartz or borosilicate glass. The support layer 10 may have a thickness ranging from a few hundred microns to a few millimeters. For example, the support layer 10 is made of quartz, and has a thickness of approximately 2 mm. It has a refractive index noted ncs. The photonic crystal structure 20 is arranged on the rear face 10ar of the support layer 10. It comprises patterns 22.1 arranged in the main plane XY, here periodically. It is sized to form a photonic crystal making it possible to improve the absorption in the active layer 30, and the patterns 22.1 have a profile which widens from the top towards the base so as to reduce the reflection rate at the interface between the support layer 10 and the photonic crystal structure 20.The photonic crystal structure 20 comprises a first layer 21, called the filling layer, made of a dielectric material, transparent in the detection spectral band Δλ. det , and having a low refractive index n L . It extends here in the main plane XY in contact with the support layer 10, but an interlayer (not shown), also made of a transparent material with a low refractive index, can be located between and in contact with the support layer 10 and the filling layer 21. The refractive index nL has a deviation in absolute value |nL-ncs| from the refractive index n cs of the support layer 10 preferably less than or equal to 0.1, defined at the resonance wavelength λ rof the guided mode supported by the photonic crystal, included in the detection spectral band Δλdet, so as to limit the reflection of the light radiation at the interface between the support layer 10 and the filling layer 21. Preferably, the refractive index n L is substantially equal to that of the support layer 10. For example, the filling layer is made of a silicon oxide, for example SiO2, whose refractive index nL varies from 1.4623 to 1.4504 in the spectral band Δλdet = [0.5; 1µm]. The filling layer 21 here comprises a continuous sub-layer 21.1 of constant thickness e ccr, here not zero but which can be zero, from which extend along the main direction +Z portions 21.2 which fill the space located between the patterns 22.1 of the photonic crystal structure 20. The thickness eccr is defined here along the Z axis between the front face 20av of the photonic crystal structure 20 and a plane passing through the top of the patterns 22.1. For example, in the case of Δλ det = [0.5; 1µm], the thickness e ccr can be of the order of approximately 0.5 µm. As described later, this value can be optimized so as to improve the absorption rate in the active layer 30 by resonance effect of the guided modes, and so as to reduce the reflection rate at the interface between the support layer 10 and the photonic crystal structure 20. The photonic crystal structure 20 also comprises a second layer 22, called the structured layer, made of a dielectric material, transparent in the detection spectral band Δλdet , and having a high refractive index n H greater than n L It extends in the main XY plane, and is formed of the patterns 22.1 and a continuous sub-layer 22.2 of constant thickness e ccs non-zero, from which the patterns 22.1 extend towards the support layer 10. The refractive index nH has a deviation nH-nL with respect to the index nL of the filling layer 21 preferably greater than or equal to 0.2 at the resonance wavelength λ r , so as to ensure optical confinement of the guided mode in the photonic crystal. For example, the structured layer 22 is made of a titanium oxide, for example TiO2, whose refractive index nH varies from 2.7114 to 2.4856 in the spectral band Δλdet = [0.5; 1µm], so that the gap n H – n Lvaries between 1.2491 and 1.0352. This deviation is 1.0794 at the resonance wavelength λr of 0.75µm. In addition, the index nH has a deviation in absolute value |nH-nca| with respect to the refractive index nca of the active layer 30 less than or equal to 0.5, and preferably is substantially equal to the latter, so that the guided mode supported in the structured layer 22 (the photonic crystal) sufficiently covers the active layer 30 (this is then referred to as a supermode). Preferably, the sum of the thicknesses of the high-index media, namely the height h of the patterns 22.1, the thickness e ccs of the continuous underlay 22.2 and the thickness e caof the active layer 30 is such that: h + eccs + eca is at least equal to λr / 2<n’H> , and preferably at least equal to λr / <n’H> , or even at least equal to 3λr / 2<n’H> . Thus, the conditions are improved so that the structured layer 22 and the active layer 30 support at least one supermode. The refractive index<n’H> is here an average of the nH index of the structured layer 22 and the nca index of the active layer 30, weighted by the respective thicknesses (h+nccs for the structured layer 22; e ca for the active layer 30), or, more precisely, an average of the permittivities of layers 22 and 30 weighted by the respective volume fractions (according to the Bruggeman model of effective media). The continuous sub-layer 22.2 has a thickness e ccsdefined along the Z axis between a plane passing through the base of the patterns 22.1 and the rear face 20ar of the photonic crystal structure 20. For example, in the case of Δλdet = [0.5; 1µm], the thickness eccs can be of the order of approximately 0.35µm. As described later, this value can be optimized to minimize the reflection rate in the detection spectral band Δλ detand maximize the absorption rate at the resonance wavelength λr. The patterns 22.1 extend from the continuous sub-layer 22.2 towards the support layer 10. They have a height h along the Z axis, defined between a plane passing through their top and a plane passing through their base (at the interface with the sub-layer 22.2). They are arranged regularly in the XY plane with a pitch p (constant if periodic arrangement). The arrangement of the patterns 22.1 is preferably two-dimensional, in the sense that it extends along two distinct axes in the XY plane, as illustrated in fig.2C. It makes it possible to make the photodetector 1 insensitive to the polarization of the incident light radiation. In addition, a two-dimensional photonic crystal is capable of supporting several guided modes at different resonance wavelengths (several resonances in the spectral band Δλdet).In such a case, we essentially consider the resonance wavelength λr which most effectively improves the absorption at the band gap edge of the active layer 30. The height h is preferably between eccs / 10 and 3eccs / 4 so that the resonance has a significant effect on the absorption rate. In addition, it is defined taking into account the thicknesses e. ccs summer ca , and refractive indices n H and n ca so that the supermode is supported and spreads correctly in the active layer 30. For example, the height h can be equal to approximately 0.35µm, and is a value to be optimized to minimize the detection spectral band reflection rate Δλ det and maximize the absorption rate at the resonance wavelength λ r. Furthermore, the patterns 22.1 have a profile whose width lH increases along the main direction +Z. The profile is defined as the envelope of a cross-section of the patterns 22.1 along a plane passing through the top of the pattern and parallel to the Z axis. In other words, the width lH of the patterns 22.1 increases from the top to the base of the patterns 22.1. The patterns 22.1 may have a cone shape, with a truncated top or not, and a circular, oval, polygonal base, etc. The profile of the patterns 22.1 may be triangular or curved, preferably parabolic. Also, this makes it possible to reduce the reflection rate at the interface between the support layer 10 and the photonic crystal structure 20, over the entire detection spectral band Δλ det . Thus, when we consider an elementary volume V elemof the photonic crystal structure 20 in an XY plane, which passes through the center of the patterns 22.1 as illustrated in Fig. 2B and 2C, we can define an average refractive index as being the average of the refractive index n L and the refractive index n H using the fill factor ff of the 22.1 patterns present in the elementary volume Velem. In other words:<n(z)> = nH×ff(z) + nL×(1-ff(z)). The filling factor ff(z) varies along the main direction +Z as the ratio of the surface s H (z) patterns in the XY plane at the ordinate z on the total surface S elem,tot of the elementary volume V elem The value of the average refractive index increases along the +Z direction, going from the low value nL at the top of the 22.1 patterns to a high value close to n H at the base of the grounds 22.1, or preferably at least equal to 90% of n H. Preferably, the profile of the patterns 22.1 is parabolic so that the average refractive index<n(z)> increases linearly. Indeed, it appears that a parabolic profile ensures an optimal reduction of the reflection rate in the detection spectral band Δλ det . Furthermore, as indicated previously, the arrangement pitch p and the height h are chosen, as well as the thickness eccs of the continuous sub-layer 22.2, and preferably also the thickness e ccrof the continuous sub-layer 21.1, such that the structured layer 22 forms a photonic crystal that supports at least one guided mode capable of being excited by the incident light radiation of interest. In addition, the photonic crystal is optically coupled to the active layer 30, such that the guided mode is optically confined in the photonic crystal and the active layer 30. More specifically, the light radiation incident on the photonic crystal structure 20 excites one or more modes supported by the photonic crystal, by a guided mode resonance (GMR) effect. A description of this guided mode resonance effect can be found in particular in the article by Magnusson & Ko entitled Guided-mode resonance nanophotonics: Fundamentals and applications, SPIE 9927, Nanoengineering: Fabrication, Properties, Optics, and Devices XIII, 992702 (15 September 2016).Also, a guided mode, of predefined wavelength λr, is present in the waveguide formed by the structured layer 22 and is oriented in the XY plane. The resonance wavelength λ. r is obviously included in the detection spectral band Δλ det , and is located in the long wavelengths, for example around 850nm or 900nm approximately, where the absorption rate of the active layer 30 normally drops, i.e. in the absence of the photonic crystal structure 20 according to the invention. For this, the thickness e ccs of the continuous sub-layer 22.2 is chosen so that the resonant guided mode spreads transversely over the active layer 30 (supermode), so that the absorption rate of the active layer 30 is then enhanced at the resonance wavelength λ r, without the transport rate being degraded (due to an increase in the thickness of the active layer 30). In addition, the arrangement pitch p is chosen so that the propagation constant of the guided mode of the nth diffraction order, defined as β n = n×2π / p, is between the refractive index n L and the minimum value among the refractive index n Hand that of the active layer 30. In addition, the height h, as well as the arrangement pitch p, are chosen so as to obtain this guided mode resonance effect. When this guided mode resonance effect is obtained, the resonant guided mode satisfies a Bloch condition, so that its group velocity ^^^^ / ^^^^ is substantially zero (ω is the pulsation and k is the wave vector). Also, the resonant guided mode forms a standing wave, and can be referred to as a “slow Bloch mode”. Also, the resonant guided mode remains located, in the XY plane, at the location where it was excited by the incident light radiation. It is then absorbed in the active layer 30 in this same area in the XY plane, which improves the contrast of the photodetector 1, or even makes it possible to limit the effects of crosstalk or halo in the case of a matrix photodetector 1.For example, the pitch p can be equal to approximately 0.55µm, and is also a value to be optimized to minimize the reflection rate in the detection spectral band Δλdet and maximize the absorption rate at the resonance wavelength λ. r. The active layer 30 rests on the rear face 20ar of the photonic crystal structure 20. It is adapted to absorb the light radiation of interest in the detection spectral band Δλdet. It is made of a semiconductor material adapted to absorb the light radiation of interest, including the resonant guided mode oriented in the XY plane and optically confined in the structured layer 22 (photonic crystal) and in the active layer 30. It has a thickness eca that is substantially constant in the XY plane. In the case of a photocathode, it may be, for example, a crystalline material based on antimony and alkali metal, or based on silver oxide, for example a material chosen from SbNaKCs, SbNa2KCs, SbNaK, SbKCs, SbRbKCs or SbRbCs. The photocathode material may also be formed based on AgOCs. As previously stated, the semiconductor material has a refractive index n casubstantially equal to that of the material of the structured layer, or having a deviation in absolute value |n ca -n H| at most equal to 0.5. It has a thickness preferably less than or equal to 300nm, for example equal to approximately 140nm. Thus, the photodetector 1 has improved performance thanks to the photonic crystal structure 20 which ensures, on the one hand, a broadband reduction of the reflection at the interface between the support layer 10 and the photonic crystal structure 20, and on the other hand, a resonant excitation of at least one guided mode supported in the structured layer 22 which forms a photonic crystal, this resonant guided mode also being optically confined in the active layer. The photodetector 1 may be a matrix, and may comprise a CMOS matrix sensor (not shown), located downstream of the active layer, and more precisely downstream of a phosphor screen.Thus, the backside of the active layer may be in contact with vacuum, and a multiplier device may be located between the active layer, to multiply the photoelectrons into a high flux of secondary electrons oriented towards the phosphor screen. The photons generated by the phosphor screen are then detected by the CMOS matrix sensor. Preferably, each detection pixel of the CMOS sensor has a dimension, in the XY plane, equal to an integer number of periods p. Note here that an approach to determine the values ​​of the pitch p, the height h, and the thicknesses eccr and eccs, so as to form the photonic crystal and the guided mode resonance, may consist of solving Maxwell's equations by a rigorous coupled-wave analysis (RCWA) method, by doing a parametric study. Thus, each of the parameters p, h, e. ccr summer ccsvaries over a predefined range, and the spectral response of the photodetector is calculated in absorption A(λ) and in reflection R(λ). The presence of an absorption peak at the resonance wavelength λr can thus be verified. Then, to obtain the optimal values, two figures of merit can be calculated, namely a FOMA figure of merit associated with the absorption spectrum, which is defined as the integral over the detection spectral band of a correlation of the absorption spectrum A(λ) with a predefined intensity spectrum NV(λ) (power spectral density of a light radiation in night vision). Similarly, a FOM figure of merit is defined R associated with the reflection spectrum, as being the integral of the correlation of the reflection spectrum R(λ) with the intensity spectrum NV(λ). The two figures of merit are therefore written: m The optimal values ​​that maximize the absorption rate A at the resonance wavelength while minimizing the reflection rate R over the detection spectral band can thus be chosen on the basis of one and / or the other of the figures of merit. In this respect, Figures 3A and 3B illustrate examples of the absorption spectra A(λ) and reflection R(λ) of the photodetector, for two examples of sizing of the photonic crystal structure. These spectra are obtained by numerical resolution of Maxwell's equations, here by an RCWA technique. The detection spectral band here extends from 0.5 to 1 µm. Furthermore, Figure 3C illustrates an example of a power spectral density of light radiation in night vision.In these examples, the photodetector 1 is made of a transparent support layer 10 made of a silicon oxide, a photonic crystal structure 20 formed of a first filling layer 21 made of SiO2 and a second structured layer 22 made of TiO2, then a thin layer of Al2O3, and finally an active layer 30 such as a photocathode made from antimony and at least one alkali metal (of type S25). The filling layer 21 comprises a continuous sub-layer 21.1 of constant thickness eccr, and the structured layer 22 comprises a continuous sub-layer 22.2 of constant thickness e. ccsand 22.1 patterns arranged regularly along two distinct axes contained in the XY plane. The 22.1 patterns have a height h, a pitch p and a maximum filling factor ffmax defined at the base of the 22.1 patterns. The 22.1 patterns have a parabolic shape with a circular base, so that the filling factor ff, and therefore the average refractive index, varies linearly along the +Z direction. The parametric study is carried out by varying the parameters p, h, eccr and eccs over the following ranges: for p, from 0.5 to 0.63µm, for h from 0.25 to 0.55µm, for e ccr from 0.23 to 0.55µm and for e ccs from 0.23 to 0.45µm. Fig. 3A illustrates the absorption spectra A(λ) and reflection R(λ), in the case where the photonic crystal structure 20 has been sized to maximize the FOMA figure of merit. In this case, the pitch p is equal to 0.6µm, the height h is equal to 0.39µm, the thickness e b1of the upstream region is equal to 0.49µm and the thickness e b2 of the downstream region is equal to 0.29µm. It appears that the figure of merit FOM A is equal to 8.9966 over the 0.5-1µm detection spectral band. This high value is due to the fact that the absorption spectrum has an absorption peak around 875nm, representative of the presence of a guided mode resonance in the photonic crystal and therefore in the active layer. In addition, the reflection is particularly low over the entire detection spectral band, with a FOM figure of merit R equal to 1.8846. Fig. 3B illustrates the absorption spectra A(λ) and reflection R(λ), in the case where the photonic crystal structure 20 has been sized to maximize the figure of merit FOMB. In this case, the pitch p is equal to 0.53µm, the height h is equal to 0.33µm, the thickness e b1 of the upstream region is equal to 0.55µm and the thickness e b2of the downstream region is equal to 0.35µm. It appears that the figure of merit FOM A remains high, with a value of 7.4456, due to the presence of an absorption peak around 850nm linked to a guided mode resonance in the photonic crystal and therefore in the active layer. In addition, the reflection is further reduced, with a figure of merit FOM Requal to 1.2261. Also, the photodetector 1 has a particularly high quantum efficiency, because the photonic crystal structure 20 makes it possible both to optimize the absorption rate in the active layer, in particular at the resonance wavelength, and to reduce the reflection rate on the detection spectral band, without degrading the transport rate. Figures 4A to 4F illustrate steps of a method for manufacturing a photodetector 1 identical or similar to that of Fig. 2A. With reference to Fig. 4A, a stack is first provided formed, from bottom to top, of a thick layer 40 of silicon (for example a plate, or wafer, of silicon with a thickness of several hundred microns), of a layer 21 of buried oxide with a thickness of approximately 2 µm, of a thin layer 41 of amorphous silicon with a thickness of 100 nm, and of a thin layer 42 of silicon oxide with a thickness of approximately 100 nm.The buried oxide is preferably made by oxidation of the silicon of the thick layer. With reference to Fig. 4B, periodic openings 43 are formed through the thin oxide layer 42 and the thin silicon layer 41, which open onto the upper face of the buried oxide layer 21. The openings 43 are made at the desired pitch p for the structured layer. For this, a hard mask is deposited on the thin oxide layer 42, which is structured to form through openings. Then the openings 43 are made through the thin oxide layer 42 by RIE type dry etching. Then the openings are made through the thin silicon layer 41 by RIE type dry etching. With reference to Fig. 4C, the thin oxide layer 42 is removed, for example by RIE dry etching. The upper face of the structured thin silicon layer 41 is then made free. With reference to Fig.4D, a partial etching of the buried oxide layer 21 is carried out, from the periodic openings present in the thin silicon layer 41, so as to form etched areas 44 having a desired periodic profile. The etching here is an isotropic wet etching, for example of the BOE etching type (buffered oxide etch). The etching depth corresponds to the desired height h of the patterns of the photonic crystal structure. The ratio between the vertical etching speed and the horizontal etching speed makes it possible to control the profile of the etched areas 44. A continuous lower part of constant thickness eccr can be defined in the buried oxide layer 21. With reference to fig. 4E, the structured thin silicon layer 41 is then removed, to give access to the upper face of the buried oxide layer 21.A layer 22 of a high refractive index material, for example here TiO2, is then deposited so as to fill the etched areas 44, which forms the patterns, and to form a continuous sub-layer of constant thickness e. ccs, and the upper face of the high-index layer 22 is planned, for example by mechanical-chemical planarization. An intermediate thin layer 45 (optional) can be deposited, for example made of Al2O3. Then, the active layer 30 is deposited so as to cover the high-index layer 22. With reference to FIG. 4F, the stack obtained is turned over, then the thick silicon layer 40 is removed, and a support layer 10 made of quartz or borosilicate glass is fixed on the face made free of the buried oxide layer 21. Thus, a photodetector 1 identical or similar to that illustrated in FIG. 2A is obtained, which comprises the transparent support layer 10, the photonic crystal structure 20, and the active layer 30. Particular embodiments have just been described. Different variants and modifications will appear to those skilled in the art.

Claims

CLAIMS 1. Photodetector (1) adapted to detect light radiation in a predefined detection spectral band, comprising a stack formed of: o a support layer (10), made of a transparent material, adapted to receive the light radiation via a front face (10av) and to transmit it via a rear face (10ar); o a grating structure (20), arranged on the rear face (10ar) of the support layer (10), comprising patterns (22.1) arranged in a main plane parallel to the front face (10av) of the support layer (10); o an active layer (30), arranged on a rear face (20ar) of the grating structure (20), made of a semiconductor material adapted to absorb the light radiation; characterized in that: ^ the grating structure (20) comprises a structured layer (22), which is formed of said patterns (22.1) and a continuous sub-layer (22.2) of constant thickness e ccsnon-zero, the patterns (22.1) being oriented towards the support layer (10) and having a profile widening along a direction oriented towards the continuous sub-layer (22.2); ^ the grating structure (20), called photonic crystal structure, has dimensions such that the structured layer (22) forms a photonic crystal which supports at least one guided mode capable of being excited by the incident light radiation of interest, the photonic crystal being optically coupled to the active layer (30) so that the guided mode is optically confined in the structured layer (22) and the active layer (30).

2. Photodetector (1) according to claim 1, in which the patterns (22.1) have a height h, between their top and their base, of between one tenth and three quarters of the thickness eccs of the continuous sub-layer (22.2) of the structured layer (22). 3.Photodetector (1) according to claim 1 or 2, in which the photonic crystal structure (20) comprises a so-called filling layer (21), located between the support layer (10) and the structured layer (22), made of a dielectric material. having a refractive index n L less than a refractive index n H of the structured layer (22) and having an nH-nL gap at least equal to 0.

2.

4. Photodetector (1) according to claim 3, in which the refractive index nL of the filling layer (21) has a gap in absolute value |n L -n cs | with a refractive index of the support layer (10) at most equal to 0.

1.

5. Photodetector (1) according to any one of claims 1 to 4, in which a refractive index n H of the structured layer (22) has a deviation in absolute value |n H -n ca| with a refractive index of the active layer (30) at most equal to 0.

5.

6. Photodetector (1) according to any one of claims 1 to 5, in which the sum of a height h of the patterns (22.1), of a thickness e ccs of the continuous sub-layer (22.2) of the structured layer (22) and of a thickness e caof the active layer (30) is at least equal to λr / (2×<n’H> ), where λr is a wavelength of the guided mode, and where<n’H> is an average index equal to an average of the refractive indices of the structured layer (22) and of the active layer (30) weighted by the respective thicknesses.

7. Photodetector (1) according to any one of claims 1 to 6, wherein the patterns (22.1) have a parabolic profile.

8. Photodetector (1) according to any one of claims 1 to 7, wherein the active layer (30) has a rear face opposite the photonic crystal structure (20), in contact with a vacuum.

9. Photodetector (1) according to any one of claims 1 to 8 and claim 3, wherein the filling layer (21) is made of an oxide.

10. Photodetector (1) according to claim 9, wherein the filling layer (21) is made of a silicon oxide. 11.Photodetector (1) according to any one of claims 1 to 10, in which the active layer (30) is a photoemissive layer made of a material based on SbNaKCs, SbNa2KCs, SbNaK, SbKCs, SbRbKCs or SbRbCs, or of a material based on AgOCs.

12. Photodetector (1) according to any one of claims 1 to 11, the latter being an image intensifier tube or a photomultiplier tube.

13. Method for manufacturing a photodetector (1) according to any one of the preceding claims, comprising the following steps: o producing a stack formed of a thick layer (40), a so-called filling layer (21), and an upper thin layer (41); o producing through-openings (43) in the upper thin layer (41) and opening onto the filling layer (21), arranged regularly at the pitch p of the patterns (22.1); o partial isotropic wet etching of the filling layer (21) from the through-openings (43), so as to form periodic etched zones (44) intended to form the patterns (22.1); then removal of the upper thin layer (41); o deposition on the filling layer (21) of a layer (22) forming the structured layer; o deposition on the structured layer (22) of the active layer (30); o removal of the thick layer (40), and assembly of the support layer (10) on the face made free of the filling layer (21).

Images