GERMANIUM PHOTODIODE WITH OPTIMIZED METAL CONTACTS

Description

TECHNICAL FIELD

[0001] The field of the invention is that of passivated planar photodiodes made from germanium comprising optimized metal contacts. The invention finds application in particular in the field of detecting light radiation in the near infrared. STATE OF THE PRIOR ART

[0002] The optoelectronic photodetection devices may comprise a matrix of passivated planar photodiodes. The photodiodes then extend along the same main plane, between first and second surfaces that are opposite and parallel to each other. They then each comprise a detection portion, formed for example of a first region of n-doped germanium and flush with the first surface, of a second region of p-doped germanium and flush with the second surface, and of an intermediate region of intrinsic or very lightly p-doped germanium, and located between the first and second doped regions. A passivation layer made of a dielectric material may cover the first surface in order to limit the contribution of the dark current to the electric current measured by each photodiode.

[0003] The electrical polarization of the first n-doped region and the second p-doped region may, in the case of passivated planar photodiodes, be carried out on the side of the first surface. Thus, a peripheral semiconductor portion, for example made of p-doped polycrystalline silicon, surrounds the detection portion in a main plane of the photodiode and comes into contact with the second p-doped region. In addition, metal contacts, arranged on the side of the first surface, come into contact with the first n-doped region and the peripheral p-doped semiconductor portion.

[0004] However, it appears that a metallic contact on n-doped germanium is of the rectifier type and not of the ohmic type, due to the blocking of the Fermi level ( Fermi level pinning, in English) at the metal / germanium interface. However, there are different solutions for making the metal / n-doped germanium contact ohmic.

[0005] Thus, the paper by Paramahans et al. entitled Contacts on n-type germanium using variably doped zinc oxide and highly doped indium tin oxide interfacial layers, Appl. Phys. Express 8, 051302 (2015) describes the intercalation of a very thin dielectric layer between the metal and the n-doped germanium, such as ZnO at least 2 nm thick or ITO at least 2.4 nm thick. Furthermore, the paper by Wu et al. entitled Ohmic contact to n-type Ge with compositional Ti nitride, Appl. Surf. Sc. 284, 877-880 (2013) describes another solution which consists of making a TiN x / n-Ge type contact with x at least equal to 0.8 to obtain an ohmic contact. However, these ohmic contacts can then be very resistive, which degrades the performance of the photodiode.

[0006] Furthermore, document EP3657556A1 describes a passive planar photodiode made from germanium, in which an intercalated semiconductor portion of n-doped polycrystalline silicon is formed on and in contact with the detection portion. It forms a reservoir of n-type dopants intended to diffuse into the germanium to form the first n-doped region. A metal contact is made on and in contact with the intercalated semiconductor portion, and makes it possible to electrically polarize the first n-doped region.

[0007] However, there is a need for such a passivated planar photodiode with improved performance.

[0008] Document EP3660930A1 describes a germanium PIN photodiode. STATEMENT OF THE INVENTION

[0009] The invention aims to remedy at least in part the drawbacks of the prior art, and more particularly to propose a passivated planar photodiode which has improved performance.

[0010] For this, the object of the invention is a photodiode comprising: a detection portion, having a first surface and a second surface opposite each other and parallel to a main plane, made of a first crystalline germanium-based semiconductor material, comprising: ∘ a first n-type doped region, and flush with the first surface; ∘ a second p-type doped region, and flush with the second surface; ∘ an intermediate region, located between the first region and the second region and surrounding the first region in the main plane; a peripheral semiconductor portion, made of a second p-type doped semiconductor material, surrounding the detection portion in the main plane and coming into contact with the second region; an interposed semiconductor portion, arranged on and in contact with the first region of the detection portion;metal contacts, arranged on the side of the first surface, and adapted to electrically polarize, on the one hand the first region by means of the intermediate semiconductor portion, and on the other hand the second region by means of the peripheral semiconductor portion.;

[0011] According to the invention, the intercalated semiconductor portion is made of a third crystalline semiconductor material having: a natural lattice parameter equal, to within 1%, to a natural lattice parameter of the first germanium-based semiconductor material; and a band gap energy at least 0.5eV higher than that of the first germanium-based semiconductor material.

[0012] Some preferred but non-limiting aspects of this photodiode are as follows.

[0013] The intercalated semiconductor portion may comprise n-type dopants identical to those present in the first region.

[0014] The intercalated semiconductor portion can be made of a III-V crystalline semiconductor compound, and preferably of AlAs or GaAs.

[0015] The interposed semiconductor portion may be located in a notch of the first surface of the sensing portion, such that the interposed semiconductor portion is surrounded, in the principal plane, by the first region.

[0016] One of the metal contacts, called the central metal contact, may be located on and in contact with the intermediate semiconductor portion.

[0017] The photodiode may comprise an upper semiconductor portion located on and in contact with the interposed semiconductor portion, made of an n-type doped semiconductor material with dopants identical to those of the interposed semiconductor portion and the first region.

[0018] One of the metal contacts, called the central metal contact, can then be located on and in contact with the upper semiconductor portion.

[0019] The upper semiconductor portion may be made of a material identical to that of the peripheral semiconductor portion.

[0020] The upper semiconductor portion and the peripheral semiconductor portion may be made of a silicon-based semiconductor material. The photodiode may then have a silicided upper zone in contact with the metal contacts.

[0021] The invention also relates to a method of manufacturing a photodiode according to any one of the preceding characteristics, comprising the following steps production of a stack comprising a first sub-layer intended to form the second region and a second sub-layer intended to form the intermediate region; production of an upper insulating layer covering the stack; production of the peripheral semiconductor portion through the stack and the upper insulating layer to open onto the first sub-layer; production of the intermediate semiconductor portion by epitaxy from the second sub-layer of the stack, through an opening in the upper insulating layer.

[0022] The method may comprise, before the step of producing the intercalated semiconductor portion, a step of producing a notch in the second sub-layer of the stack, through the opening, followed by the epitaxy of the intercalated semiconductor portion in the notch.

[0023] When the intercalated semiconductor portion is unintentionally doped during its epitaxy, the process can then include the following steps: production of the upper n-type doped semiconductor portion; annealing adapted to cause the diffusion of the dopants contained in the upper semiconductor portion through the intercalated semiconductor portion to form the first region in the detection portion.

[0024] The upper semiconductor portion and the peripheral semiconductor portion may be made of the same silicon-based material. The method may then include a step of making the metal contacts simultaneously, one in contact with the upper semiconductor portion, and the other in contact with the peripheral semiconductor portion.

[0025] The intercalated semiconductor portion can be n-type doped during its epitaxy. The process can then include the following step: making metal contacts, one in contact with the intercalated semiconductor portion, and the other in contact with the peripheral semiconductor portion. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] 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: there Figure 1A is a schematic and partial cross-sectional view of a passivated planar photodiode according to one embodiment, comprising a metal contact in contact with the interposed semiconductor portion, which is in contact with the first region of the detection portion; Figure 1Bis a schematic and partial view, in cross-section, of a passivated planar photodiode according to an alternative embodiment, comprising an upper semiconductor portion, arranged between and in contact with the metal contact and the intermediate semiconductor portion; Figures 2A to 2L illustrate different steps of a method of manufacturing a photodiode according to the embodiment illustrated in the fig.1B ; THE Figures 3A to 3C illustrate different steps of a method of manufacturing a photodiode according to the embodiment illustrated in the fig.1A . DETAILED DESCRIPTION OF SPECIFIC EMBODIMENTS

[0027] 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 enhance the clarity of the figures. Furthermore, the different embodiments and variants are not mutually exclusive and may 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 equivalent mean that the limits are included, unless otherwise indicated.

[0028] The invention relates generally to a passivated planar photodiode, and preferably to a photodiode array, as well as to the manufacturing method. Each photodiode comprises a detection portion made from germanium, and is adapted to detect light radiation in the near infrared (SWIR, for Short Wavelength IR, in English) corresponding to the spectral range from approximately 0.8µm to 1.7µm, or even approximately 2.5µm.

[0029] The detection portion of the photodiodes has a first surface and a second surface opposite each other and parallel to a main plane of the photodiodes. The two surfaces extend along identical planes for each of the photodiodes, and vertically delimit (along the thickness axis) the detection portion. The photodiodes do not have a mesa structure insofar as they are optically isolated from each other by peripheral trenches filled with a doped semiconductor material. They thus have a particularly high fill factor. Furthermore, they are said to be passivated insofar as the first surface is partly covered by a dielectric passivation layer. This contributes to reducing the surface component of the dark current.

[0030] Generally speaking, the photodiode includes: ∘ a detection portion, made of a first crystalline semiconductor material based on germanium, and comprising: a first n-type doped region, a second p-type doped region, and an intermediate region located between the two doped regions and surrounding the first doped region in a main plane of the photodiode; ∘ a peripheral semiconductor portion, made of a second p-doped semiconductor material, surrounding the detection portion in the main plane and coming into contact with the second doped region; ∘ an intercalary semiconductor portion, arranged on and in contact with the first n-doped region, and made of a crystalline, n-doped semiconductor material, and coming into contact with the first n-doped region;∘ metal contacts, arranged on the side of the first surface, and adapted to electrically polarize, on the one hand the first region by means of the intermediate semiconductor portion, and on the other hand, the second region by means of the peripheral semiconductor portion.;

[0031] The crystalline semiconductor material of the intercalated semiconductor portion has a natural lattice parameter equal to that of the first germanium-based semiconductor material to within 1%, i.e., it is equal to plus or minus 1%, and preferably to within 0.5%. By natural lattice parameter is meant the lattice parameter of the unstressed, i.e., relaxed, material. The lattice parameter in question here is oriented in the principal plane of the photodiode. In addition, the crystalline semiconductor material of the intercalated semiconductor portion has a bandgap energy at least 0.5eV higher than that of the first germanium-based semiconductor material. The bandgap energy corresponds to the energy difference between the minimum energy of the conduction band and the maximum energy of the valence band.The crystalline semiconductor material of the intercalated semiconductor portion is preferably a III-V compound, for example a binary III-V compound such as AlAs and GaAs, or even a ternary III-V compound such as GaAlAs or GaInP.

[0032] There Figure 1A is a partial and schematic view, in cross-section, of a passivated planar photodiode 1 according to one embodiment, belonging to a matrix of photodiodes. The photodiodes 1 are made from germanium, and are reverse-polarized from the first surface 10a, while being isolated from each other by trenches filled with a p+ doped semiconductor material. They each comprise an interposed semiconductor portion 27 located here between and in contact with a metal electrical polarization contact 32.1 and the first n+ doped region 11.

[0033] Here and for the remainder of the description, a direct three-dimensional reference frame XYZ is defined, where the X and Y axes form a plane parallel to the main plane of the photodiode 1, and where the Z axis is oriented along the thickness of the detection portion 1 of the photodiode, from the second surface 10b in the direction of the first surface 10a.

[0034] The photodiode 1 comprises a detection portion 10 extending along the Z axis between a first and a second reference surface 10a and 10b, parallel to each other and opposite to each other. The first surfaces 10a of the photodiodes are parallel to each other, and the second surfaces 10b are also parallel to each other. The first surface 10a is defined by a part of the detection portion 10 where the first n+ doped region 11 and the intermediate region 13 are flush. The second surface 10b is opposite the first surface 10a along the Z axis.

[0035] The maximum thickness of the detection portion 10, defined along the Z axis between the first and second surfaces 10a, 10b, is here substantially constant from one photodiode to another, for example is between a few hundred nanometers and a few microns, for example between approximately 1 µm and 5 µm. The thickness is chosen so as to obtain good absorption in the wavelength range of the light radiation to be detected. The detection portion 10 has a transverse dimension in the XY plane which can be between a few hundred nanometers and a few tens of microns, for example between approximately 1 µm and 20 µm.

[0036] The detection portion 10 is made of a crystalline, preferably monocrystalline, germanium-based semiconductor material. By germanium-based, it is meant that the crystalline semiconductor material corresponds to germanium or is a compound formed of at least germanium. Thus, the photodiodes can be made, for example, of germanium Ge, silicon germanium SiGe, germanium tin GeSn, or even silicon germanium tin SiGeSn. In this example, the detection portion 10 is made of at least one layer made of germanium. It can thus be a layer or a substrate made of the same semiconductor material and have regions of different types of conductivity (homojunction) so as to form a PN or PIN junction. It can alternatively be a stack of sub-layers of different semiconductor materials (heterojunction), which are then formed based on germanium.

[0037] The detection portion 10 is thus formed of a first n-type doped (n or n+) region 11, which is flush with the first surface 10a and forms an n-type doped well, and a second p+ doped region 12, which is flush with the second surface 10b. By flush, we mean “reaching the level of”, or “extending from”. An intermediate region 13, not intentionally doped (in the case of a PIN junction), or p-doped (in the case of a PN junction), is located between and in contact with the two doped regions 11 and 12, and surrounds the first n+ doped region 11 in the main plane. In this example, the semiconductor junction is of the PIN type, the first region 11 being n+ doped, the second region 12 p+ doped and the intermediate region 13 is intrinsic (not intentionally doped).

[0038] The first n-type doped region 11 extends here from the first surface 10a and is surrounded by the intrinsic region 13 in the main plane. It is distant from the lateral edge 10c of the detection portion 10 in the XY plane, the lateral edge 10c being defined by the internal face of a peripheral p+ doped semiconductor portion 25. It thus forms an n+ doped well which is flush with the first surface 10a and is spaced by a non-zero distance from the lateral edge 10c as well as from the second surface 10b. The first n+ doped region 11 thus participates in delimiting the first surface 10a. It may have a doping which may be between 5.10 18< and 10 21< at / cm 3< approximately.

[0039] The second p+ doped region 12 extends in the XY plane, flush with the second surface 10b, here from the lateral edge 10c. It extends along the Z axis from the second surface 10b. It may have a substantially uniform thickness along the Z axis and thus be flush with only a lower zone of the lateral edge 10c. Alternatively, as illustrated in the fig.1A and 1B , the second p+ doped region 12 may have a p+ doped lateral region 14 which is continuously flush with the lateral edge 10c along the Z axis and extends over the entire periphery of the detection portion 10. The second p+ doped region 12 may have a doping which may be between 10 18< and 10 19< at / cm 3< approximately.

[0040] The intermediate region 13 is located between the two n+ and p+ doped regions 11, 12, and surrounds the first n+ doped region 11 in the XY plane. It is here made of an intrinsic semiconductor material so as to form a PIN junction but can be weakly p-doped, to form a PN junction (cf. Fig. 1B ).

[0041] The photodiode 1 here comprises a lower insulating layer 21, made of a dielectric material, covering the second surface 10b of the detection portion 10, as well as, as described later, the lower face of the p+-doped peripheral semiconductor portion 25. The lower insulating layer 21 may further be adapted to form an anti-reflection function with respect to the incident light radiation. It in fact forms the receiving face of the light radiation intended to be detected.

[0042] The detection portion 10 of the photodiode 1 is here delimited laterally, in the XY plane, by a trench, preferably continuous, filled with a p+ doped semiconductor material, and forming a peripheral semiconductor portion 25, here p+ doped. The peripheral through portion 25 participates in electrically polarizing the photodiode 1, here from the side of the first surface 10a, and in pixellizing the photodiode matrix (optical isolation). It extends here over the entire thickness of the detection portion 10 to open onto the lower insulating layer 21, but alternatively, it may not open onto the lower insulating layer 21 and may end in the second p+ doped region 12. The internal face of this p+ doped peripheral semiconductor portion 25 then defines the lateral edge 10c of the detection portion 10.The semiconductor material is preferably made from silicon, for example amorphous silicon, polycrystalline silicon, silicon germanium, or can even be made from amorphous germanium.

[0043] An upper insulating layer 23 covers the first surface 10a of the photodiode 1, and makes it possible to electrically insulate the metal contacts 32.1 and 32.2. It is thus in contact with the first n+ doped region 11 as well as the intermediate region 13. It is made of a dielectric material, such as a silicon oxide, a silicon nitride, or a silicon oxynitride. Other dielectric materials may be used, such as a hafnium or aluminum oxide, or even an aluminum nitride, among others. It has a thickness for example between 50nm and 500nm.

[0044] Furthermore, the detection portion 10 advantageously comprises a p+ doped lateral region 14 located at the lateral edge 10c. This lateral region 14 has a doping level higher than that of the intermediate region 13 when it is doped. The p+ doped lateral region 14 is flush with the lateral edge 10c and is in contact with the p+ doped peripheral semiconductor portion 25. Thus, the polarization of the second p+ doped region 12 is improved to the extent that the contact surface with the p+ doped peripheral semiconductor portion 25 is increased. In addition, this p+ doped lateral region 14 makes it possible to prevent the space charge zone of the photodiode 1 from extending as far as the lateral edge 10c. Thus, the contribution of this zone (potentially not free from defects linked to the production of the trenches) to the dark current is limited. This improves the performance of photodiode 1.

[0045] Furthermore, the detection portion 10 is made from germanium, for example is made from germanium, and the p+-doped peripheral semiconductor portion 25 is made from silicon, for example doped polycrystalline silicon. The detection portion 10 then advantageously comprises a lateral zone 15 made from silicon germanium. The lateral zone 15 is flush with the lateral edge 10c and is in contact with the p+-doped peripheral semiconductor portion 25. Thus, the lateral zone 15 has a bandgap energy greater than that of the detection portion 10 made from germanium. This lateral “gap opening” makes it possible to reduce the sensitivity of the photodiode 1 to defects present near the trenches. This also improves the performance of the photodiode 1.

[0046] Furthermore, the photodiode 1 comprises an intercalated semiconductor portion 27, arranged on and in contact with the first n+ doped region 11. It is spaced by a non-zero distance from the peripheral semiconductor portions 25 in the XY plane and is here surrounded by the upper insulating layer 23. It is made of a crystalline semiconductor material doped here of the n+ type, which has a natural lattice parameter equal to within 1%, and preferably to within 0.5%, of that of the crystalline semiconductor material of the detection portion 10, here germanium, and has a band gap energy greater by at least 0.5eV than that of the crystalline semiconductor material of the detection portion 10 (germanium). The intercalated semiconductor portion 27 can thus be made of a III-V binary or ternary compound, such as GaAs, AlAs, GaAlAs, GaInP, among others.

[0047] In this example, the intercalated semiconductor portion 27 is advantageously located in a notch formed in the first surface 10a of the detection portion 10, which makes it possible to reduce or even avoid the presence of structural defects (dislocations) in the intercalated portion 27, as described later. Alternatively, it can rest on the first surface 10a which is then continuously planar (no notch). The size in the XY plane of the intercalated semiconductor portion 27 depends on the size of the photodiode 1: for a photodiode with a pitch of 5 µm, it can be between approximately 0.5 and 4 µm (and preferably less than 2 µm), and for a pitch of 10 µm, it can be between approximately 1 and 9 µm.

[0048] It is noted here that the intercalated semiconductor portion 27 comprises n-type dopants identical to those present in the first n+-doped region 11, insofar as, as explained below, the formation of the first region 11 is carried out by diffusion of the dopants through the intercalated semiconductor portion 27 (cf. fig.2A to 2L ) or initially contained in the intercalated semiconductor portion 27 ( fig.3A to 3C ).

[0049] The photodiode 1 further comprises metal contacts 32.1, 32.2 allowing it to be reverse-biased from the side of the first surface 10a. Thus, a metal contact 32.1 is here arranged on and in contact with the intermediate semiconductor portion 27, and allows the first n+ doped region 11 to be electrically polarized. A metal contact 32.2 is here arranged on and in contact with the peripheral semiconductor portion 25, and allows the second p+ doped region 12 to be electrically polarized. The metal contacts 32.1, 32.2 are here electrically isolated from each other in the XY plane by the upper insulating layer 23 and by a dielectric passivation layer 29. The photodiode 1 is intended to be reverse biased, for example by applying a negative electrical potential to the p+ doped peripheral semiconductor portion 25 and by bringing the first n+ doped region 11 to ground.

[0050] Generally speaking, for illustration purposes, the photodiode 1 may have dimensions in the XY plane of between approximately 1 µm and 100 µm. The thickness of the second p+ doped region 12 may be between approximately 20 nm and 500 nm. The thickness of the intrinsic region 13 may be between approximately 0.7 µm and 2.5 µm when the photodiode 1 is intended to detect light radiation in the SWIR or near infrared (NIR) range. The first n+ doped region 11 may have a thickness of between approximately 10 nm and 600 nm. The dielectric layers 23 and 29 may together have a thickness enabling the upper face of the photodiode 1 to be completely covered, for example between approximately 10 nm and 600 nm, and the thickness of the lower insulating layer 21 may be between approximately 50 nm and 1 µm.

[0051] Thus, the photodiode 1 has improved performance compared to the examples of the prior art mentioned previously. Indeed, the photodiode 1, by the presence of the intermediate semiconductor portion 27 located between the metal contact 32.1 and the n-doped germanium 11, does not have a metal contact on the n-doped germanium

[0052] (first region) which would then form a rectifier contact, or even a highly resistive ohmic contact. This limits the series resistances, which in particular makes it possible to limit low-frequency noise. In addition, the intercalated semiconductor portion 27, which is in contact with the first n+ doped region 11, is made of a semiconductor material with a band gap energy at least 0.5eV higher than that of germanium, which makes it possible to reduce the dark current of the photodiode 1 on the one hand, and to reduce the sensitivity of the photodiode 1 to defects possibly present near the notch in which the intercalated semiconductor portion 27 is located. In addition, the absence of mismatch of the lattice parameter between the crystalline material of the intercalated semiconductor portion 27 and that of the detection portion 10 makes it possible to avoid structural defects (dislocations, etc.) which would degrade the performance of the photodiode 1.

[0053] There Figure 1B is a schematic and partial view, in cross-section, of a photodiode 1 according to an alternative embodiment. In this example, the photodiode 1 differs from that illustrated in the fig.1A in particular in that the intermediate region 13 is weakly p-doped (PN junction), and in that an upper semiconductor portion 28 is located between and in contact with the central metal contact 32.1 and the intermediate semiconductor portion 27.

[0054] More specifically, the intercalated semiconductor portion 27, as indicated previously, rests on and in contact with the detection portion 10. It is thus in contact with the first n+ doped region 11. It is made of a crystalline semiconductor material having the properties mentioned above in terms of band gap energy and absence of lattice mismatch with the germanium of the detection portion 10. As indicated below, it was produced by epitaxy from germanium and in an unintentionally doped manner.

[0055] The upper semiconductor portion 28 therefore rests on and in contact with the intercalated semiconductor portion 27. It is spaced from the peripheral semiconductor portion 25 and the metal contacts 32.2 in the XY plane by a non-zero distance. This lateral space is filled here by the passivation dielectric layer 29. The upper semiconductor portion 28 is made of an n+ doped semiconductor material, with dopants capable of n+ doping the germanium to form the first n+ doped region 11, here with phosphorus or arsenic. Thus, the upper semiconductor portion 28 is a reservoir of dopants intended to diffuse, during diffusion annealing, through the intercalated semiconductor portion 27 to join the germanium and form the first n+ doped region 11. It is preferably made from polycrystalline silicon, but other polycrystalline materials can of course be used.

[0056] Finally, the metal contact 32.1 is here arranged on and in contact with the upper semiconductor portion 28. It makes it possible to electrically polarize the first n+ doped region 11 via the upper semiconductor portion 28 then the intermediate semiconductor portion 27.

[0057] THE Figures 2A to 2L are schematic and partial views, in cross-section, of different stages of a method of manufacturing a photodiode according to the example of the fig.1B . This method has the particular advantage of producing the metal contacts 32.1 and 32.2 simultaneously. It also has the advantage of not doping the intercalated semiconductor portion 27 during growth (controlling the diffusion of the dopants can be tricky) or by ion implantation (which can create defects in the intercalated semiconductor portion 27).

[0058] In this example, the photodiodes 1 are made of germanium and have a PIN junction, and are adapted to detect infrared radiation in the SWIR range. The photodiodes 1 are planar and passivated, and are reverse biased from the first surface 10a, and here by means of a control chip 40 hybridized to the photodiode matrix 1.

[0059] In reference to the fig.2A , a first semiconductor sub-layer 22.1 of monocrystalline germanium is produced. The first semiconductor sub-layer 22.1 is secured to a support layer 20, here made of silicon, by means of a lower insulating layer 21, here made of silicon oxide. This stack takes the form of a GeOI substrate (for Germanium On Insulator,in English). This stack is preferably produced using the method described in the publication by Reboud et al. entitled Structural and optical properties of 200mm germanium-on-insulator (GeOI) substrates for silicon photonics applications, Proc. SPIE 9367, Silicon Photonics X, 936714 (February 27, 2015). Such a method has the advantage of producing a semiconductor sub-layer of germanium 22.1 having a total absence or a low rate of structural defects such as dislocations. The germanium may be unintentionally doped or be doped, for example p-type. The semiconductor sub-layer 22.1 may have a thickness of between approximately 100nm and 500nm, for example approximately 300nm, and may be covered with a protective layer (not shown) made of a silicon oxide. The lower insulating layer 21 (BOX, for Buried Oxide,in English) can have a thickness between 50nm and 1µm, for example between 100nm and 500nm, and advantageously provides an anti-reflective function.

[0060] The first sub-layer 22.1 is then doped with p+-doped germanium by ion implantation of a dopant such as boron or gallium, when the first sub-layer 22.1 was initially made of intrinsic germanium. The protective layer, if any, has been previously removed by surface cleaning, and the first sub-layer 22.1 of germanium can be coated with a pre-implantation oxide layer (not shown) with a thickness of a few tens of nanometers, for example equal to 20 nm. The sub-layer 22.1 of germanium then has a doping level of between approximately 10 18 and 10 20 at / cm 3. Diffusion annealing of the dopant can then be carried out under nitrogen for a few minutes to a few hours, for example 1 hour, at a temperature which can be between 600°C and 800°C, for example equal to 800°C. This annealing may not be performed when the 22.1 sublayer was doped at growth.Another way to make this p+ layer is by epitaxy of a layer of boron-doped germanium in situ between 10 18< and 10 19< at / cm 3< approximately on a sub-layer of intrinsic germanium. This epitaxy can be done between 400 and 800°C but preferably at 400°C.

[0061] In reference to the fig.2B , a second semiconductor sub-layer 22.2 of germanium is produced by epitaxy from the first sub-layer 22.1. The two sub-layers 22.1, 22.2 are intended to form the coplanar detection portions 10 in germanium of the photodiode matrix 1. The second sub-layer 22.2 is formed by epitaxy, for example by chemical vapor deposition (CVD, for Chemical Vapor Deposition, in English) and reduced pressure precedence (RPCVD, for Reduced Pressure Chemical Vapor Deposition) or by any other epitaxy technique. Annealing can be carried out to reduce the dislocation rate in the sub-layer 22.2. The pre-implementation oxide layer, if any, has been previously removed by surface cleaning. The second sub-layer 22.2 of germanium is here intrinsic, i.e. not intentionally doped since it is desired to produce a PIN junction. It is intended to form the light absorption zone of the photodiodes 1. Its thickness depends on the wavelength range of the light radiation to be detected in the case of a photodiode 1. In the context of SWIR photodiodes, the sub-layer 22.2 of intrinsic germanium has a thickness for example between 0.5 µm and 3 µm, preferably equal to 1.5 µm.

[0062] In reference to the fig.2C, an upper insulating layer 23 is deposited so as to continuously cover the upper face of the second sub-layer 22.2, that is to say so as to cover the detection portions 10 of the photodiodes 1. The upper insulating layer 23 is made of a dielectric material, for example a silicon oxide, nitride or oxynitride. Cleaning of the upper face of the second sub-layer 22.2 may have been carried out beforehand. The upper insulating layer 23 may have a thickness of between 10nm and 600nm.

[0063] In reference to the fig.2D, the trenches 24 intended to pixelate the photodiodes 1 and to participate in their reverse electrical polarization are produced by photolithography and etching (by means of the peripheral semiconductor portions 25 which will then be produced). A localized etching of the upper insulating layer 23, of the sub-layer 22.2 of intrinsic germanium, and of the sub-layer 22.1 of p+ doped germanium is thus carried out, until it opens here onto the upper face of the lower insulating layer 21 (but the trenches 24 can open onto the sub-layer 22.1 without crossing it). Each trench 24 thus preferably extends continuously around a photodiode 1. A plurality of detection portions 10 are thus obtained, separated from each other by a continuous trench 24. They are preferably obtained by an anisotropic etching technique, so as to obtain a lateral border 10c which is substantially vertical along the Z axis.The trenches 24 have a transverse dimension (width) in the XY plane which can be between 300nm and 2µm, for example equal to 1µm. The detection portions 10 can thus have a shape in the XY plane, for example circular, oval, polygonal, for example square, or any other shape.

[0064] In reference to the fig.2E, the peripheral semiconductor portions 25 are then produced. For this, a doped semiconductor material is deposited so as to fill the trenches 24. The semiconductor material is preferably a silicon-based material, for example amorphous silicon, polycrystalline silicon, silicon germanium, or even amorphous germanium. The semiconductor material is doped p+ with boron or gallium, with a dopant concentration of the order of approximately 10 19< to 10 20< at / cm 3<. Thus, the doped semiconductor material comes into contact with the lateral edge 10c via the trench 24. A chemical mechanical polishing (CMP) step is then carried out, stopping on the upper face of the upper insulating layer 23, to remove the excess semiconductor material and planarize the upper face formed by the upper insulating layer 23 and the semiconductor material of the peripheral semiconductor portion 25.We thus obtain a peripheral semiconductor portion 25 doped p+ in each trench 24.

[0065] In reference to the fig.2F, a dielectric layer is deposited so as to cover the upper surface of the peripheral semiconductor portions 25 (this increases the thickness of the upper insulating layer 23 which thus covers the portions 25). Then, by photolithography and etching, a through opening 26.1 is produced within the upper insulating layer 23 to open onto the detection portion 10, opposite a central zone located at a distance from the peripheral semiconductor portions 25 in the XY plane. The width or diameter of the through opening 26.1 depends in particular on the desired width of the first n+ doped region. It may be between 0.3 µm and 5 µm for example. Preferably, the through opening 26.1 also forms a notch within the detection portion 10, for example with a depth of around a hundred nanometers.

[0066] In reference to the fig.2G, an intercalated semiconductor portion 27 is produced, located opposite the central zone of the detection portion 10 and in contact with it. The intercalated semiconductor portion 27 is produced by epitaxy from the germanium of the detection portion, for example by molecular beam epitaxy (MBE, Molecular Beam Epitaxy, in English). It is made of a crystalline semiconductor material which has, on the one hand, a lattice parameter almost identical to that of germanium, that is to say that its natural lattice parameter is equal to that of germanium to within at most 1.0%, and preferably to at most 0.5%; and on the other hand, a band gap energy Eg higher (that is to say strictly higher) than that Eg Ge of the germanium of the detection portion.

[0067] For example, the intercalated semiconductor portion 27 is made of AlAs or GaAs (stoichiometric compounds, i.e. the proportion of the two elements is identical, whose natural lattice parameter is 5.6605Å for AlAs and 5.653Â for GaAs, which corresponds respectively to a deviation of 0.044% and 0.088% with the natural lattice parameter of 5.658Â of germanium. In addition, the band gap energy is 2.12eV for AlAs and 1.424eV for GaAs, which is much higher than the 0.67eV of germanium. Other semiconductor compounds, for example ternary, may also be suitable, such as for example GaAlAs and GaInP depending on the proportions of the different chemical elements.

[0068] In this example, the intercalated semiconductor portion 27 is intrinsic, i.e. not intentionally doped. The formation of the first n+ doped region 11 will be obtained by diffusion of germanium dopants (e.g. phosphorus or arsenic) from a dopant reservoir (the upper semiconductor portion 28) through the intercalated semiconductor portion 27).

[0069] In reference to the fig.2H , an upper semiconductor portion 28 is produced, forming the reservoir of n-type dopants intended to diffuse into the detection portion through the intermediate semiconductor portion 27 to obtain in finethe first n+ doped region 11. It is made of an n+ doped semiconductor material, for example polysilicon or polycrystalline silicon germanium. However, it is advantageously made of a material identical to that of the peripheral semiconductor portion, for example here polysilicon, so as to simplify the production of the metal contacts 32.1 and 32.2.

[0070] For example, the upper semiconductor portion 28 is made of polysilicon doped n+ with phosphorus or arsenic, produced by low pressure chemical vapor deposition (LPCVD) between 400°C and 650°C for example. It has a size in the XY plane preferably greater than that of the intercalary semiconductor portion 27, and a thickness of for example between 50nm and 200nm.

[0071] In reference to the fig.2I, a diffusion and activation annealing of the dopants is then carried out, for example between 500°C and 800°C and for approximately 10 to 60 min. Thus, the dopants (phosphorus or arsenic) diffuse through the intercalated semiconductor portion 27 then into the detection portion 10. A first n+ doped region 11 is thus obtained which extends in contact with the intercalated semiconductor portion 27, both along the Z axis and in the XY plane.

[0072] Note that the fact that the intercalated semiconductor portion 27 was epitaxially grown in a notch 26.2 of the detection portion 10 makes it possible to reduce the number of structural defects (dislocations) in the intercalated semiconductor portion 27 or even to avoid them. This results in a crystalline material of better crystalline quality, which improves the performance of the photodiode.

[0073] In reference to the fig.2J, a passivation dielectric layer 29 is deposited so as to cover the upper semiconductor portion 28. The dielectric layer can be deposited at 400°C and be made of a dielectric material such as a silicon oxide, nitride or oxynitride, an aluminum oxide or nitride, a hafnium oxide, among others. It can have a thickness of, for example, between 200nm and 1000nm. A CMP-type planarization step is then carried out.

[0074] Through openings 30 are then made through the dielectric layers 23, 29 so as to open onto an upper surface of the peripheral semiconductor portion 25 (to then make the metal contact 32.2) and onto an upper surface of the upper semiconductor portion 28 (to then make the metal contact 32.1). These through openings 30 may have a size in the XY plane of between 0.1 µm and 1 µm, preferably between 0.3 µm and 1 µm. They may be made by plasma etching with an etching stop on the polysilicon surfaces of the upper semiconductor portion 28 and the peripheral semiconductor portion 25. Thus, the fact that the upper portion 28 and the peripheral portion 25 are made of the same material, here silicon, makes it possible to make the metal contacts 32.1 and 32.2 simultaneously, thus simplifying the manufacturing process.

[0075] In reference to the fig.2K , a silicided zone 31 is preferably produced on the upper surface of the upper semiconductor portion 28, and a silicided zone 31 on the upper surface of the peripheral semiconductor portion 25. For this, a thin layer of the type formed by a Ni / TiN stack with thicknesses of 9nm and 10nm is deposited at the bottom of the through openings 30, 30 by physical vapor deposition (PVD), followed by a siliciding annealing between approximately 300°C and 350°C for approximately 10 to 30s. The unreacted Ni and TiN is then removed, and then a second siliciding annealing is carried out between approximately 400°C and 450°C for approximately 10 to 30s.

[0076] The central 32.1 and lateral 32.2 metal contacts are then produced. Here, a thin bonding layer 33 formed from a Ti / TiN / Cu type stack is deposited by chemical deposition in the value phase in the through openings 30.1, 30.2, and the empty space is filled with copper 34 deposited by electrolysis. A CMP type planarization step is then carried out with a stop on the oxide of the dielectric passivation layer 29. The dielectric passivation layer 29 and the metal contacts 32.1 and 32.2 together have the same planar upper face.

[0077] In reference to the fig.2L, the hybridization of the optoelectronic stack thus obtained is carried out on a control chip 40. The connection face of the control chip 40 can thus be coated with an insulating layer 41, made of a dielectric material, crossed by metal contacts 42. the photodiode matrix 1 and the control chip 40 are thus assembled by hybrid molecular adhesion, by contact of the faces formed by the metal contacts and the insulating layers. A bonding annealing can be carried out so as to increase the surface bonding energy between the two faces in contact. The support layer 20 is then removed, for example by abrasion ( grinding, in English), so as to expose the lower insulating layer 21. This thus forms the receiving face of the light radiation to be detected, and advantageously provides an anti-reflection function.

[0078] The manufacturing method thus makes it possible to obtain one or more passivated photodiodes 1, each comprising an interposed semiconductor portion 27 in contact with the detection portion 10, the properties of which in terms of mesh parameters and band gap energy make it possible to avoid having to make a metal contact directly on the n-doped germanium, such a contact being either of the rectifier type or of the ohmic but highly resistive type. One or more photodiodes 1 are thus produced having improved performance, in particular in terms of low-frequency noise (as indicated previously).

[0079] THE Figures 3A to 3C are schematic and partial views, in cross-section, of different stages of a method of manufacturing a photodiode 1 according to the example of the fig.1A .

[0080] Photodiode 1 is different from the one illustrated in the fig.1Bessentially in that it does not include an upper semiconductor portion 28. Also, the metal contact 32.1 is on and in contact with the intermediate semiconductor portion 27.

[0081] Also, the first n+ doped region 11 is produced by diffusion of dopants from the intermediate semiconductor portion 27 which itself forms the dopant reservoir (and not from the upper semiconductor portion 28, as in the Figures 2A to 2L ).

[0082] In reference to the fig.3A , the intermediate semiconductor portion 27 is produced by epitaxial growth directly from the detection portion 10. This step is similar to that of the fig.2Gand is distinguished therefrom in that the intercalated semiconductor portion 27 is n+ doped with dopants suitable for n-doping the germanium of the detection portion 10, for example here with phosphorus, arsenic, or even silicon or zinc. The intercalated semiconductor portion 27 is doped during growth and not by ion implantation, to avoid any degradation of the crystalline quality of the material. Furthermore, the dopants do not modify the properties of the material of the intercalated semiconductor portion 27 in terms of lattice parameter. The bandgap energy remains at least 0.5eV higher than that of the germanium of the detection portion 10. It should also be noted that the epitaxy temperature is controlled to limit any possible diffusion of the dopants in the detection portion 10.

[0083] In reference to the fig.3B, the first n+ doped region 11 is produced by performing a diffusion and activation annealing of the dopants. The temperature and duration of the annealing can be similar to those mentioned previously, in connection with the fig.2I . Thus, the dopants diffuse from the intercalated semiconductor portion 27 into the detection portion 10, which forms the first n+ doped region 11.

[0084] In reference to the fig.3C, finally, the metal contacts 32.1 and 32.2 are produced. First, a through opening is produced in the insulating layers 23 and 29 which opens onto the peripheral semiconductor portion 25, then the silicided zone 31 is produced. Then, a through opening is produced in the insulating layer 29 which opens onto the intermediate semiconductor portion 27. Finally, a thin bonding layer 33 is deposited in the openings and then the free space is filled with copper 34. This gives the metal contacts 32.1 and 32.2.

[0085] Particular embodiments have just been described. Different variants and modifications will appear to those skilled in the art.

Claims

1. Photodiode (1), comprising: ∘ a detection portion (10), having a first surface (10a) and a second surface (10b) opposite each other and parallel to a main plane, made of a first germanium-based crystalline semiconductor material, comprising: • a first n-type doped region (11), and flush with the first surface (10a); • a second p-type doped region (12), and flush with the second surface (10b); • an intermediate region (13), located between the first region (11) and the second region (12) and surrounding the first region (11) in the main plane; o a peripheral semiconductor portion (25), made of a second p-type doped semiconductor material, surrounding the detection portion (10) in the main plane and coming into contact with the second region (12); o an interposed semiconductor portion (27), disposed on and in contact with the first region (11) of the detection portion (10); o metal contacts (32.1, 32.2), disposed on the side of the first surface (10a), and adapted to electrically bias, on the one hand, the first region (11) by means of the interposed semiconductor portion (27) and, on the other hand, the second region (12) by means of the peripheral semiconductor portion (25); o characterized in that the interposed semiconductor portion (27) is made of a third crystalline semiconductor material having: • a natural lattice parameter equal, to within 1%, to a natural lattice parameter of the first germanium-based semiconductor material; • a bandgap energy at least 0.5 eV higher than that of the first germanium-based semiconductor material.

2. Photodiode (1) according to Claim 1, wherein the interposed semiconductor portion (27) comprises n-type dopants identical to those present in the first region (11).

3. Photodiode (1) according to Claim 1 or 2, wherein the interposed semiconductor portion (27) is made of a III-V crystalline semiconductor compound.

4. Photodiode (1) according to Claim 3, wherein the interposed semiconductor portion (27) is made of AlAs or GaAs.

5. Photodiode (1) according to any one of Claims 1 to 4, wherein the interposed semiconductor portion (27) is located in a notch of the first surface (10a) of the detection portion (10) so that the interposed semiconductor portion (27) is surrounded, in the main plane, by the first region (11).

6. Photodiode (1) according to any one of Claims 1 to 5, wherein one of the metal contacts (32.1), called central metal contact, is located on and in contact with the intermediate semiconductor portion (27).

7. Photodiode (1) according to any one of Claims 1 to 5, comprising an upper semiconductor portion (28) located on and in contact with the interposed semiconductor portion (27), made of an n-type doped semiconductor material with dopants identical to those of the interposed semiconductor portion (27) and of the first region (11).

8. Photodiode (1) according to Claim 7, wherein one of the metal contacts (32.1), called central metal contact, is located on and in contact with the upper semiconductor portion (28).

9. Photodiode (1) according to Claim 7 or 8, wherein the upper semiconductor portion (28) is made of a material identical to that of the peripheral semiconductor portion (25).

10. Photodiode (1) according to any one of Claims 7 to 9, wherein the upper semiconductor portion (28) and the peripheral semiconductor portion (25) are made of a silicon-based semiconductor material and comprising a silicided upper zone (31) in contact with the metal contacts (32.1, 32.2).

11. Method for manufacturing a photodiode according to any one of the preceding claims, comprising the following steps: ∘ producing a stack comprising a first sub-layer (22.1) intended to form the second region (12) and a second sub-layer (22.2) intended to form the intermediate region (13); ∘producing an upper insulating layer (23) covering the stack; ∘ producing the peripheral semiconductor portion (25) through the stack and the upper insulating layer (23) in order to emerge onto the first sub-layer (22.1); ∘ producing the interposed semiconductor portion (27) by epitaxy from the second sub-layer (22.2) of the stack, through an opening (26.1) of the upper insulating layer (23).

12. Manufacturing method according to Claim 11, comprising, before the step of producing the interposed semiconductor portion (27), a step of producing a notch (26.2) in the second sub-layer (22.2) of the stack, through the opening (26.1), followed by epitaxy of the interposed semiconductor portion (27) in the notch (26.2).

13. Manufacturing method according to Claim 11 or 12, wherein the interposed semiconductor portion (27) is unintentionally doped during the epitaxy thereof, the method comprising the following steps: ∘ producing the n-type doped upper semiconductor portion (28); ∘ annealing adapted to cause diffusion of the dopants contained in the upper semiconductor portion (28) through the interposed semiconductor portion (27) in order to form the first region (11) in the detection portion (10).

14. Manufacturing method according to Claim 13, wherein the upper semiconductor portion (28) and the peripheral semiconductor portion (25) are made of the same silicon-based material, the method further comprising the following step: ∘ simultaneously producing metal contacts (32.1, 32.2), one (32.1) in contact with the upper semiconductor portion (28), and the other (32.2) in contact with the peripheral semiconductor portion (25).

15. Manufacturing method according to Claim 11 or 12, wherein the interposed semiconductor portion (27) is n-type doped during the epitaxy thereof, the method comprising the following step: ∘ producing metal contacts (32.1, 32.2), one (32.1) in contact with the interposed semiconductor portion (27), and the other (32.2) in contact with the peripheral semiconductor portion (25).