Method for manufacturing an optoelectronic device comprising an LED and a photodiode
Patent Information
- Application Number
- FR2022012871
- Authority / Receiving Office
- FR · FR
- Patent Type
- Patents
- Current Assignee / Owner
- Filing Date
- 2022-12-07
- Publication Date
- 2025-03-14
- Estimated Expiration
- 2042-12-07
Smart Images
Figure 00000023_0000 
Figure 00000023_0001 
Figure 00000023_0002
Abstract
Description
Title of the invention: Method for manufacturing an optoelectronic device comprising an LED and a photodiode Technical field
[0001] The present description relates generally to the field of optoelectronic devices. It relates more particularly to the production of an optoelectronic device comprising at least one light-emitting diode (LED), and at least one photodiode. It relates in particular to the simultaneous production, by means of a common epitaxy step, of an active LED emission stack and an active photodiode reception stack, intended to operate in the same wavelength range. Prior art
[0002] Devices have already been proposed, for example in the field of optical communication systems, comprising one or more LEDs configured to emit light signals, and one or more photodiodes configured to receive and measure signals emitted by the LEDs.
[0003] It would be desirable to be able to improve at least in part certain aspects of these systems.
[0004] It would be particularly desirable to be able to simultaneously produce, by means of a common epitaxy step, an active LED emission stack and an active photodiode reception stack. Summary of the invention
[0005] For this, one embodiment provides a method for manufacturing an optoelectronic device comprising at least one LED and at least one photodiode, comprising the following successive steps: a) form, by epitaxy, an active semiconductor emission and reception stack common to the LED and the photodiode; b) forming trenches extending vertically through the active stack and laterally delimiting the LED and the photodiode, wherein the trenches are arranged so that the lateral dimensions of the LED are less than the lateral dimensions of the photodiode.
[0006] According to one embodiment, the trenches are arranged so that the lateral dimensions of the LED are at least two times smaller than the lateral dimensions of the photodiode.
[0007] According to one embodiment, the trenches are arranged so that the lateral dimensions of the LED are at least four times smaller than the dimensions lateral sides of the photodiode.
[0008] According to one embodiment, the trenches are arranged so that the lateral dimensions of the LED are less than 4 μm.
[0009] According to one embodiment, the method comprises, between step a) and step b), a step of transferring and fixing the active stack on a face of an integrated control circuit previously formed in and on a semiconductor substrate.
[0010] According to one embodiment, the transfer and fixing step, the active stack is fixed on said face of the integrated control circuit by molecular bonding.
[0011] According to one embodiment, at the end of the transfer and fixing step, the active stack extends continuously over the entire surface of the integrated control circuit.
[0012] According to one embodiment, the active semiconductor stack comprises one or more III-V or II-VI type semiconductor alloys.
[0013] Another embodiment provides an optoelectronic device comprising at least one LED and at least one photodiode each comprising an active semiconductor emission and reception stack of the same nature and composition, in which the lateral dimensions of the LED are smaller than the lateral dimensions of the photodiode.
[0014] According to one embodiment, the device further comprises an integrated control circuit on one face of which the LED and the photodiode are fixed, the integrated control circuit being adapted to drive the LED with a higher current density than that of the photodiode.
[0015] According to one embodiment, the integrated control circuit is adapted to drive the LED with a current density at least ten times higher than that of the photodiode.
[0016] Another embodiment provides a method of manufacturing an optoelectronic device comprising at least one LED and at least one photodiode, comprising the following steps: a) forming a supporting semiconductor stack comprising at least one doped semiconductor layer; b) simultaneously forming, during a common epitaxy step, an active semiconductor stack for emitting the LED and an active semiconductor stack for receiving the photodiode; (c) forming trenches extending vertically through the support stack and laterally delimiting at least one first support pad and at least one second support pad, in which, at the end of steps b) and c), the active semiconductor emitting stack of the LED covers the first support pad and the active stack of reception of the photodiode covers the second support pad, the method further comprising, after step c), a step d) of porosifying said doped semiconductor layer in the first support pad without porosifying said doped semiconductor layer in the second support pad, or a step of porosifying said doped semiconductor layer in the second support pad without porosifying said doped semiconductor layer in the first support pad.
[0017] According to one embodiment, step c) of forming the trenches through the support stack and step d) of porosifying the doped semiconductor layer are implemented before step b) of epitaxy of the active semiconductor emission stack of the LED and the active semiconductor reception stack of the photodiode, and in which, during step d), said doped semiconductor layer is porosified in the second support pad and is not porosified in the first support pad.
[0018] According to one embodiment, step c) of forming the trenches through the support stack is implemented after step b) of epitaxy of the active semiconductor emission stack of the LED and the active semiconductor reception stack of the photodiode, and, during step d), said doped semiconductor layer is porosified in the first support pad and not in the second support pad.
[0019] According to one embodiment, in step d), the sides of the doped semiconductor layer in the second pad are brought into contact with the electrolyte, while the sides of the doped semiconductor layer in the first pad are protected from contact with the electrolyte by a protective layer.
[0020] According to one embodiment, in step d), the sides of the doped semiconductor layer in the first pad are brought into contact with an electrolyte, while the sides of the doped semiconductor layer in the second pad are protected from contact with the electrolyte by a protective layer.
[0021] According to one embodiment, in step d), a bias current is applied through said doped semiconductor layer.
[0022] According to one embodiment, the method comprises, after steps b) and d), a step of transferring and fixing the LED and the photodiode on one face of an integrated control circuit previously formed in and on a semiconductor substrate.
[0023] According to one embodiment, during the transfer and fixing step, the LED and the photodiode are fixed on the face of the integrated control circuit by molecular bonding.
[0024] According to one embodiment, the trenches are arranged so that the lateral dimensions of the LED are smaller than the lateral dimensions of the photodiode.
[0025] According to one embodiment, the active semiconductor emission stack of the LED and the active semiconductor reception stack of the photodiode comprise one or more III-V or II-VI type semiconductor alloys.
[0026] Another embodiment provides an optoelectronic device comprising at least one LED comprising an active emission semiconductor stack and at least one photodiode comprising an active reception semiconductor stack, the device further comprising a doped semiconductor layer facing the LED and the photodiode, in which the doped semiconductor layer is porous facing the LED and non-porous facing the photodiode, or in which the doped semiconductor layer is porous facing the photodiode and non-porous facing the LED.
[0027] According to one embodiment, the device further comprises an integrated control circuit on one face of which the LED and the photodiode are fixed, the integrated control circuit being adapted to drive the LED with a higher current density than that of the photodiode.
[0028] According to one embodiment, the integrated control circuit is adapted to drive the LED with a current density at least ten times higher than that of the photodiode. Brief description of the drawings
[0029] These characteristics and advantages, as well as others, will be explained in detail in the following description of particular embodiments given without limitation in relation to the attached figures among which:
[0030] [Fig.lA], [Fig.lB], [Fig.lC] and [Fig.lD] are sectional views illustrating steps of an example of implementation of a method for manufacturing an optoelectronic device according to a first embodiment;
[0031] [Fig.2A], [Fig.2B], [Fig.2C], [Fig.2D], [Fig.2E] and [Fig.2F] are sectional views illustrating steps of an example of implementation of a method for manufacturing an optoelectronic device according to a second embodiment;
[0032] [Fig.3A], [Fig.3B], [Fig.3C], [Fig.3D] and [Fig.3E] are sectional views illustrating steps of an exemplary implementation of a method for manufacturing an optoelectronic device according to a third embodiment; and
[0033] [Fig.4] is a diagram illustrating the response of an active LED emitting stack and an active photodiode receiving stack made using a common epitaxy step. Description of the embodiments
[0034] The same elements have been designated by the same references in the different figures. In particular, the structural and / or functional elements common to the different embodiments may have the same references and may have identical structural, dimensional and material properties.
[0035] For the sake of clarity, only the steps and elements useful for understanding the embodiments described have been shown and are detailed. In particular, the production of the electrical connections and the control circuits of the LEDs and photodiodes of the devices described has not been detailed, the embodiments described being compatible with the usual embodiments of these elements or the production of these elements being within the reach of the person skilled in the art from the indications of the present description. In addition, the applications likely to benefit from the embodiments described have not been detailed, the embodiments described being able to advantageously be used for any application comprising one or more LEDs and one or more photodiodes intended to operate in the same wavelength range, for example a wavelength range of visible, ultraviolet or near infrared light.
[0036] Unless otherwise specified, when referring to two elements electrically connected to each other, this means directly connected without intermediate elements other than conductors, and when referring to two elements connected (in English "coupled") to each other, this means that these two elements can be connected or be connected by means of one or more other elements.
[0037] In the following description, when reference is made to absolute position qualifiers, such as the terms "front", "back", "top", "bottom", "left", "right", etc., or relative position qualifiers, such as the terms "above", "below", "upper", "lower", etc., or to orientation qualifiers, such as the terms "horizontal", "vertical", etc., reference is made unless otherwise specified to the orientation of the figures.
[0038] Unless otherwise specified, the expressions "about", "approximately", "substantially", and "of the order of" mean to within 10%, preferably to within 5%.
[0039] According to one aspect of the described embodiments, a method of manufacturing an optoelectronic device is provided in which an active emissive stack of LEDs and an active photosensitive stack of photodiodes are produced simultaneously during the same epitaxy step.
[0040] One advantage lies in the cost reduction compared to methods comprising separate specific epitaxy steps to successively produce the active emissive stack of LEDs and the active photosensitive stack of photodiodes.
[0041] The LED and the photodiode can be integrated monolithically in the same optoelectronic chip, or be separated by cutting at the end of the process to be integrated into separate chips, for example intended to be assembled within the same optoelectronic device.
[0042] The active emissive stack of LEDs and the active photosensitive stack of photodiode are for example inorganic semiconductor stacks, for example based on III-V type semiconductor materials, for example based on group III nitrides, for example gallium, aluminum, indium or an alloy based on one or more of these materials. Alternatively, the active emissive LED stack and the active photosensitive photodiode stack are based on II-VI type semiconductor materials, for example ZnCdSe (zinc-cadmium-selenium).
[0043] The same active stack based on gallium nitride can, for example, be used in transmission as an active LED stack, or in reception, as an active photodiode stack. The photodiode then advantageously has a very low dark current and a narrow optical reception bandwidth, which makes it possible to obtain a very good signal-to-noise ratio.
[0044] However, a difficulty then lies in the fact that the optimal emission wavelength of the LED (emission peak) is shifted upwards by a few tens of nanometers, typically of the order of 20 nm for an active stack based on gallium nitride (GaN), for example based on indium-gallium nitride (InGaN), compared to the optimal reception wavelength of the photodiode (absorption peak). This is the Stokes shift, which comes in particular from the binding energy of the electron-hole pairs. This affects the sensitivity of the photodiode in the emission wavelength range of the LED, and consequently the efficiency of the LED-photodiode system.
[0045] This phenomenon is notably illustrated by [Fig.4].
[0046] [Fig.4] is a diagram representing the evolution, as a function of the wavelength W (on the abscissa), of the quantum efficiency Q in reception (curve 401) and in emission (curve 403) of an active stack of diodes based on gallium nitride (GaN), for example based on indium-gallium nitride (InGaN).
[0047] According to one aspect of a first embodiment, provision is made to form, by epitaxy, an active semiconductor stack common to the LED and the photodiode, then to form trenches extending vertically through the active stack and laterally delimiting the LED and the photodiode. According to the first embodiment, the LED has lateral dimensions smaller than those of the photodiode. This makes it possible to relax the mechanical stresses in the active LED stack more strongly than in the active photodiode stack. This results in a reduction of the internal electric field in the active LED stack relative to the internal electric field in the active photodiode stack. This reduction of the internal electric field in the active LED stack leads to a downward shift (this is referred to as a blue shift) of the emission peak of the LED.This allows to at least partially compensate for the Stokes shift between the peak. emission peak and the absorption peak of the active stack. Thus, the emission peak of the LED approaches the absorption peak of the photodiode, which improves the efficiency of the system.
[0048] Figures 1A to 1D are sectional views schematically illustrating steps of an example of implementation of a method for manufacturing an optoelectronic device according to the first embodiment.
[0049] [Fig.1A] illustrates a structure comprising an active semiconductor emission and reception stack 103 arranged on the upper face of a support substrate 101.
[0050] The active stack 103 comprises, for example, a semiconductor layer 103a doped with a first conductivity type, for example N-type, coating the upper face of the substrate 101, an active layer 103b coating the face of the layer 103a opposite the substrate 101, i.e. its upper face in the orientation of [Fig. 1A], and a semiconductor layer 103c doped with a second conductivity type, for example P-type, coating the face of the layer 103b opposite the layer 103a, i.e. its upper face in the orientation of [Fig. 1A]. By way of example, the layer 103b is in contact, by its lower face, with the upper face of the layer 103a, and, by its upper face, with the lower face of the layer 103c.
[0051] The layers 103a, 103b and 103c of the active stack 103 each extend, for example, continuously and with a substantially uniform thickness over the entire surface of the substrate 101.
[0052] The layers 103a, 103b and 103c are for example formed successively by epitaxy on the upper face of the support substrate 101.
[0053] For example, the support substrate 101 is made of sapphire or silicon. The semiconductor layers 103a and 103c of the active stack 103 are, for example, made of gallium nitride. The active layer 103b comprises, for example, a stack of layers each forming a quantum well, for example based on indium-gallium nitride (InGaN).
[0054] A buffer layer, not shown, may provide an interface between the upper face of the substrate 101 and the lower face of the lower layer 103a.
[0055] [Fig. 1A] further illustrates a step of depositing, on the upper face of the active stack 103, a metal layer 105. In the example shown, the layer 105 extends continuously and with a substantially uniform thickness over the entire upper surface of the active stack 103. By way of example, the layer 105 is in contact, by its lower face, with the upper face of the upper layer 103c of the active stack.
[0056] [Fig. 1B] schematically represents an integrated control circuit 110, previously formed in and on a semiconductor substrate 111, for example a substrate in silicon. In this example, the control circuit 110 comprises, on its upper face side, for each of the LEDs of the device, a metal connection pad 113L intended to be connected to one of the electrodes (anode or cathode) of the LED, so as to be able to control a current flowing in the LED and / or apply a voltage to the terminals of the LED. The control circuit 110 further comprises in this example, on its upper face side, for each of the photodiodes of the device, a metal connection pad 113P intended to be connected to one of the electrodes (anode or cathode) of the photodiode, so as to be able to read an electrical signal representative of the intensity of light radiation received by the photodiode in its sensitivity wavelength range.
[0057] The control circuit comprises, for example, for each LED, connected to the metal pad 113L dedicated to the LED, an elementary control cell comprising one or more transistors, making it possible to control the current flowing in the LED and / or a voltage applied to the terminals of the LED, and, for each photodiode, connected to the metal pad 113P dedicated to the photodiode, an elementary reading cell comprising one or more transistors, making it possible to read an electrical signal representative of the intensity of a light radiation received by the photodiode in its sensitivity wavelength range. The reading circuit comprises, for example, a transimpedance amplifier used to amplify the current of the photodiode.
[0058] The control circuit 110 is for example produced in CMOS technology. The metal pads 113L, 113P may be laterally surrounded by an insulating material 114, for example silicon oxide, so that the control circuit 110 has a substantially flat upper surface comprising an alternation of metal regions 113 and insulating regions 114. The contact on the electrodes of the LEDs or photodiodes not connected to the pads 113L, 113P, may be made collectively, for example in a peripheral region of the control circuit 110, by means of one or more connection pads (not visible in the figure) of the control circuit 110.By way of example, the control circuit 110 comprises, on the side of the upper face of the substrate 111, a stack of insulating and conductive levels forming an interconnection network 112 comprising in particular the connection pads 113L, 113P, the upper face of the interconnection network 112 defining the upper face of the circuit 110.
[0059] [Fig. 1B] further illustrates a step of depositing, on the upper face of the integrated control circuit 110, a metal layer 115. In the example shown, the layer 115 extends continuously and with a substantially uniform thickness over the entire upper surface of the circuit 110. By way of example, the layer 115 is in contact, by its lower face, with the upper face of the interconnection network 112 of the control circuit 110.
[0060] The layer 115 is for example made of the same material as the layer 105. For example, the layers 105 and 115 each comprise an upper layer called a bonding layer. The bonding layers of the layers 105 and 115 are preferably made of the same material, for example titanium.
[0061] [Fig.lC] illustrates the structure obtained at the end of a step of transferring the active stack of LEDs and photodiodes 103 onto the upper face of the control circuit 110. For this, the structure of [Fig.lA] can be turned over, then transferred onto the structure of [Fig.lB] so as to bring into contact the face of the metal layer 105 opposite the substrate 101 (i.e. its lower face in the orientation of [Fig.lC], corresponding to its upper face in the orientation of [Fig.lA]) with the face of the metal layer 115 opposite the substrate 111 (i.e. its upper face in the orientation of figures 1B and 1C). During this step, the active stack 103 is bonded to the control circuit 110. For example, the bonding of the active stack 103 to the control circuit 110 can be obtained by molecular bonding between the two surfaces brought into contact.Alternatively, the two surfaces can be fixed by thermocompression, eutectic bonding, or any other suitable fixing method.
[0062] Once the bonding has been carried out, the support substrate 101 is removed so as to expose the upper face (in the orientation of [Fig.1C]) of the semiconductor layer 103c of the active stack 103. The substrate 101 is for example removed by grinding and / or etching from its face opposite the active stack 103. As a variant, in the case of a transparent substrate 101, for example a sapphire substrate, the substrate 101 can be detached from the active stack 103 by means of a laser beam projected through the substrate 101 from its face opposite the active stack 103 (laser lift-off type method). More generally, any other method for removing the substrate 101 can be used. After removal of the substrate, an additional etching step may be provided to remove any buffer layers remaining on the upper face side of the semiconductor layer 103c.Furthermore, part of the thickness of the layer 103c can be removed, for example by etching. At the end of this step, the active stack 103 covers substantially the entire surface of the control circuit 110, without discontinuity. For example, the thickness of the active stack 103 at the end of the step of [Fig. 1D] is between 0.5 and 2 μm.
[0063] At the end of this step, the mechanical constraints of the epitaxial active stack 103 are partially transferred into the substrate 111 of the control circuit 110.
[0064] [Fig. 1D] illustrates a step subsequent to the step of [Fig. 1C], during which trenches 120 are formed in the active stack 103, from its upper face, for example by lithography then etching, so as to delimit one or more LEDs L and one or more photodiodes P, each corresponding to a portion of the active stack 103 in the form of an island or mesa. In the example shown, the trenches 120 extend vertically over the entire height of the active stack 103 and open onto the upper face of the metal layer 105. The trenches 120 can be aligned with marks previously formed on the control circuit 110. In the example shown, each LED L is located, in vertical projection, opposite a single metal pad 113L of the control circuit 110, and each photodiode P is located, in vertical projection, opposite a single metal pad 113P of the control circuit 110. By way of example, each LED L and each photodiode P has, in top view, a substantially square or rectangular shape. The trenches 120 form, for example, in top view, a grid or a grid laterally separating the LEDs L and the photodiodes P of the device from each other.
[0065] The trenches can then be extended through the metal layers 105 and 115 to individualize the electrical connections on the lower semiconductor layer 103c of the active stack 103 of each LED L and each photodiode P. Subsequent steps can then be implemented to resume individual or common electrical contact on the upper semiconductor layer 103a of the active stack 103 of each LED L and each photodiode P. These steps have not been detailed and are within the scope of the person skilled in the art from the indications of the present description. By way of example, these steps are similar to what was described in the patent application WO2017194845 or in the patent application WO2019092357 previously filed by the applicant.
[0066] During the step of [Fig. 1D] of etching the active stack 103, an additional relaxation of the mechanical stresses present in the epitaxially grown active stack 103 occurs via the edges of the etched islands or mesas. This relaxation depends on the size of the islands or mesas. In particular, small islands or mesas exhibit significant stress relaxation, while larger islands or mesas retain relatively high mechanical stress. The relaxation may further depend on the nature of the substrate, which may for example comprise a stack of a gallium nitride layer on a silicon layer, or a stack of a gallium nitride layer on a sapphire layer, or a stack of a porous gallium nitride layer on a silicon layer.
[0067] According to one aspect of the first embodiment, it is planned to define: - LEDs L of relatively small lateral dimensions so as to obtain a significant relaxation of the mechanical stresses in the active stack 103 and consequently a relatively significant downward shift of the emission peak, and - P photodiodes with relatively large lateral dimensions so as to obtain less significant relaxation of mechanical stresses in the active stack 103 and consequently a relatively small downward shift of the absorption peak.
[0068] This makes it possible to at least partially compensate for the Stokes shift naturally present between the emission peak and the absorption peak of the active stack 103.
[0069] By way of example, the islands or mesa forming the LEDs L have lateral dimensions less than or equal to 5 pm, for example less than or equal to 4 pm, for example less than or equal to 2 pm. This makes it possible to obtain almost total relaxation of the active stack during the etching of the LED. The islands or mesa forming the photodiodes P have lateral dimensions greater than those of the LEDs, for example at least twice greater than those of the LEDs, for example at least four times greater than those of the LEDs, so as to maintain relatively high mechanical stresses in the active stack 103 of the photodiodes P.
[0070] As a non-limiting example, for an active stack based on GaN and for square L LEDs with sides of approximately 1 pm and P photodiodes with sides of 8-10 pm, an alignment of the emission peak of the L LEDs with the absorption peak of the P photodiodes is observed.
[0071] The embodiments described are not limited to the example arrangement of the LEDs L and the photodiodes P shown in [Fig. 1D]. By way of example, the device may comprise, on a first part of the surface of the integrated control circuit 110, a plurality of LEDs L, for example identical (apart from manufacturing dispersions), for example arranged in a matrix according to rows and columns, for example with a constant inter-LED pitch. The device may further comprise, on a second part of the surface of the integrated control circuit 110, a plurality of photodiodes P, for example identical (apart from manufacturing dispersions), for example arranged in a matrix according to rows and columns, for example with a constant inter-photodiode pitch. The inter-LED pitch in the first region is for example identical to the inter-photodiode pitch in the second region.The lateral dimensions of the LEDs in the first region are, on the other hand, smaller than the lateral dimensions of the photodiodes in the second region.
[0072] In addition to the differentiated size of the LEDs L compared to the photodiodes P, another parameter making it possible to reduce the wavelength shift between the emission peak of the LEDs and the absorption peak of the photodiodes is the density of charge carriers in the active stack and in particular in the quantum wells of the active layer 103b. More particularly, a high carrier density will lead to screening the electric field present in the active stack, and consequently to shifting downwards the optimal operating wavelength of the active stack.
[0073] Thus, advantageously, the control circuit 110 is configured to drive the LEDs L at higher voltages than the photodiodes P. This makes it possible to obtain a higher carrier density in L LEDs than in P photodiodes, and consequently to reduce the shift between the emission peak of the L LEDs and the absorption peak of the P photodiodes. For example, the driving voltages are chosen so as to have a carrier density in the L LEDs at least twice as high, for example at least five times as high, for example of the order of ten times as high, than in the P photodiodes.
[0074] The value of the wavelength shift linked to the increase in current density in the LED depends on the structure of the active stack, and in particular on the width of the quantum wells of the active layer 103b. In particular, the wider the wells, the greater the screening of the electric field linked to the increase in carrier density, and consequently the greater the downward shift of the optimal emission wavelength of the LED linked to the increase in carrier density. Increasing the width of the wells, on the other hand, leads to a longer radiative recombination time, which can be detrimental for communication applications requiring a short recombination time. The person skilled in the art will be able to choose the appropriate compromise according to the needs of the application.As an illustrative, non-limiting example, for an LED comprising 4 nm thick InGaN quantum wells with an indium content of 14.3%, driving the LED with a current density of the order of 100 A / cm2 leads to a blue shift of the emission peak of approximately 6 nm compared to driving the same LED at a current density of the order of 10 A / cm2.
[0075] In order to fully compensate for the Stokes shift, it may be possible, for example, to combine the mechanical relaxation effect described above by providing smaller LEDs than the photodiodes, and the field screening effect by the carriers by providing a higher current density in the LEDs than in the photodiodes. As an illustrative, non-limiting example, for gallium nitride-based LEDs comprising InGaN quantum wells, a wavelength shift of the emission peak is observed between a 4 pm wide LED driven at a current density of the order of 200 A / cm2 and an LED of the same type with a width of 25 pm driven at a current density of the order of 10 A / cm2, of the order of 30 nm. Of this 30 nm shift, about 20 nm is related to the size difference, the rest (about 10 nm) being related to the current density difference.This shift is typically of the same order as the Stokes shift between transmission and reception in the active stack.
[0076] It will be noted that compensation by differentiation of the carrier densities between the LEDs and the photodiodes can also be obtained in a device having L LEDs with the same lateral dimensions as the P photodiodes, or even lateral dimensions greater than those of the P photodiodes.
[0077] According to one aspect of a second embodiment, provision is made, before the common epitaxy step during which the active emission and reception stacks are formed simultaneously, to locally porosify, opposite the photodiodes of the device, a support layer made of a semiconductor material, on which the active stack is intended to be epitaxially grown. This results in a relaxation of the mechanical stresses in the active stack of photodiodes during epitaxy, in particular during the formation of the active layer 103b of the stack. This relaxation leads to a difference in proportions of the species of the semiconductor alloy forming the active layer 103b between the photodiodes and the LEDs. In particular, in the case where the active layer comprises InGaN quantum wells, this results in a greater incorporation of indium in the quantum wells of the photodiodes than in the quantum wells of the LEDs.This leads to a red shift, i.e. upwards, of the absorption peak of the photodiodes, and thus to at least partially compensating the Stokes shift between the emission peak and the absorption peak of the active stack.
[0078] Figures 2A to 2F are sectional views schematically illustrating steps of an example of implementation of a method for manufacturing an optoelectronic device according to the second embodiment.
[0079] [Fig.2A] illustrates a structure comprising a support semiconductor stack 210 on one face of a support substrate 101. The support substrate 101 is for example identical or similar to what has been described previously. The support semiconductor stack 210 is for example made of a III-V type semiconductor material, for example gallium nitride. The support semiconductor stack 210 comprises at least one doped semiconductor layer 210b, with a doping level chosen to make the layer 210b porous during a subsequent electrolytic porosification step. For example, the layer 210b is N-type doped. For example, the layer 210b is made of N-type doped gallium nitride with a doping level of between 1019 and 1.5*1019 atoms / cm3.
[0080] In the example shown, the support stack 210 further comprises, on the side of the lower face of the layer 210b, for example in contact with the lower face of the layer 210b, a semiconductor layer 210a. The layer 210a is for example made of the same material as the layer 210b but with a doping level lower than that of the layer 210b, for example with a doping level at least ten times lower than that of the layer 210b. As a variant, the layer 210a is made of a material different from the material of the layer 210b.
[0081] In the example shown, the support stack 210 further comprises, on the side of the upper face of the layer 210b, for example in contact with the upper face of the layer 210b, a semiconductor layer 210c. The layer 210c is for example made of the same material as the layer 210b but with a doping level lower than that of layer 210b, for example with a doping level at least ten times lower, preferably at least 100 times lower, than that of layer 210b. Alternatively, layer 210c is made of a material different from the material of layer 210b.
[0082] The layers 210a, 210b and 210c of the support stack 210 each extend, for example, continuously and with a substantially uniform thickness over the entire surface of the substrate 101.
[0083] The layers 210a, 210b and 210c are for example formed successively by epitaxy on the upper face of the support substrate 101.
[0084] For example, the support substrate 101 is made of sapphire or silicon. A buffer layer, not shown, may optionally provide an interface between the upper face of the substrate 101 and the lower face of the lower layer 210a of the support stack 210.
[0085] [Fig.2B] illustrates a step of forming trenches 220 in the support stack 210, from its upper face, for example by lithography then etching, so as to define in the stack 210 a plurality of support pads SL and SP in the form of islands or mesas. Each support pad SL is intended to receive, on its upper face, an LED L of the device, and each support pad SP is intended to receive, on its upper face, a photodiode P of the device.
[0086] In the example shown, the trenches 220 extend vertically from the upper face of the stack, pass entirely through the layers 210c and 210b, and open into the layer 210a without passing entirely through it. As a variant, the trenches 220 pass entirely through the layer 210.
[0087] The trenches 220 form, for example, in top view, a grid or a grid laterally separating from each other the support pads SL and SP intended to receive the LEDs L and the photodiodes P of the device.
[0088] The support pads SP and SL have, for example, all the same lateral dimensions, for example between 1 μm and 25 μm, for example between 2 and 8 μm. For example, the support pads SP and SL have, in top view, a square or rectangular shape.
[0089] At this stage, in each support pad SL and SP, the sides of the doped semiconductor layer 210b of the support stack are exposed.
[0090] [Fig.2C] illustrates the structure obtained at the end of a step of selective porosification of the layer 210b, located only in the support pads SP of the photodiodes P of the device. During this step, the layer 210b of the support pads SP is made porous by electrolytic etching or electroporosification. The layer 210b of the support pads SL, on the other hand, remains non-porous.
[0091] For this, the sides of the support pads can be previously coated with a protective layer (not visible in the figure), for example made of an insulating material, for example, silicon oxide or nitride. The protective layer is, for example, initially deposited over the entire upper face, then removed locally, for example by photolithography and etching, so as to expose the sides of the SP support pads without exposing the sides of the SL support pads.
[0092] The structure can then be immersed in an electrolytic bath (not visible in the figure), for example a solution based on oxalic acid, for example an aqueous solution of oxalic acid.
[0093] A bias voltage is then applied so as to cause a current to flow in the doped semiconductor layer 210b. For example, the voltage is applied between a first electrode (not visible in the figure) connected to the layer 210a and the electrolyte (not visible in the figure) connected by the edge to the layer 210c.
[0094] Under the effect of the bias current, the portions of the layer 210b in contact, by their sides, with the electrolyte, that is to say the portions of the layer 210b included in the support pads SP of the photodiodes P of the device, become porous. The portions of the layer 210b protected from contact with the electrolyte, that is to say the portions of the layer 210b included in the support pads SL of the LEDs L of the device, remain intact (non-porous).
[0095] It will be noted that, in this example, the doping levels of layers 210a, 210b and 210c of the support stack are chosen so that only layer 210b is made porous during the electroporosification step.
[0096] At the end of this step, the protective layer covering the sides of the SL support pads can be removed.
[0097] [Fig.2D] illustrates the structure obtained at the end of a common epitaxy step during which an active semiconductor stack 103 is formed on each support pad SL and on each support pad SP. The epitaxy is for example located in openings previously etched in a dielectric layer, not shown.
[0098] On each support pad SL and SP, the active stack 103 covers for example the entire upper surface of the pad. The portion of the active stack 103 covering each pad SL defines an LED of the device. The portion of the active stack 103 covering each pad SP defines a photodiode of the device.
[0099] On each support pad SP and SL, the active stack 103 comprises, in order from the upper surface of the pad, a semiconductor layer 103a, a semiconductor layer 103b, and a semiconductor layer 103c, for example identical or similar to what has been described previously in relation to FIGS. 1A to 1D. The layers 103a, 103b and 103c are for example formed successively by epitaxy from the upper face of the pads SP and SL. For example, in each pad SP and SL, the lower semiconductor layer 103a of the active stack 103 is in contact, by its lower face, with the upper face of layer 210c.
[0100] Due to the presence of the porous layer 210b in the support pads SP, a greater mechanical relaxation is obtained in the active stack of the P photodiodes than in the active stack of the L LEDs. This results, during epitaxy, in an incorporation of the different species in the active layer 103b of the active stack of the L LEDs and in the active layer 103b of the active stack of the P photodiodes. In particular, in the case of an active layer 103b based on InGaN, this results in a greater incorporation of indium in the active layer 103b of the P photodiodes than in the active layer 103b of the L LEDs. Thus, the presence of the porous layer 210b in the support pads SP of the P photodiodes leads to shifting the absorption peak of the P photodiodes upwards in wavelength (towards the red), and thus to bringing it closer to the peak emission of LEDs L.
[0101] [Fig.2E] illustrates the structure obtained at the end of a step of forming, on each LED L, a contact metallization 232L on and in contact with the upper face of the upper semiconductor layer 103c of the active stack 103 of the LED, and, on each photodiode P, a contact metallization 232P on and in contact with the upper face of the upper semiconductor layer 103c of the active stack 103 of the photodiode.
[0102] [Fig.2E] further illustrates a step of filling the trenches 220 and the space between the LEDs L and the photodiodes P with an electrically insulating material 234, for example silicon oxide.
[0103] After filling, a planarization step can be implemented, for example by chemical-mechanical polishing (CMP), so that the contact metallizations 232L, 232P are flush with the upper face of the filling material 234.
[0104] [Fig.2F] illustrates a step of transferring and fixing the structure of [Fig.2E] onto a control integrated circuit 110, for example similar to that of [Fig. IB].
[0105] During this step, the contact metallizations 232L, 232P of the structure of [Fig.2E] are brought into contact, by their face opposite the support substrate 101, with the face of the contact metallizations 113L, 113P of the control circuit 110 opposite the substrate 111.
[0106] For example, the structure of [Fig.2E] is fixed and electrically connected to the control integrated circuit 110 by molecular bonding, for example by hybrid metal-metal / oxide-oxide bonding.
[0107] After assembling the two structures, the support substrate 101 of the structure of [Fig.2E] can be removed. In addition, all or part of the support semiconductor stack 210 can be removed, for example by grinding or etching.
[0108] In the example shown, the layer 210a of the support stack 210 is in completely removed, and layers 210b and 210c are retained. The described embodiments are not limited to this example, however.
[0109] Subsequent steps can then be implemented to resume individual or common electrical contact on the upper semiconductor layer 103a of the active stack 103 of each LED L and of each photodiode P. For example, a layer of a transparent electrically conductive material, for example a transparent conductive oxide, for example indium-tin oxide (ITO) is deposited on and in contact with the upper face of the structure of [Fig.2F]. These steps have not been detailed and are within the scope of the person skilled in the art from the indications of the present description.
[0110] Similar to what has been described previously, the control circuit can optionally be configured to drive the LEDs and the photodiodes with carrier densities adapted to reduce the offset between the emission peak of the LEDs and the absorption peak of the photodiodes.
[0111] According to one aspect of a third second embodiment, provision is made to form support pads SP and SL in a manner similar to that described above in relation to FIGS. 2A to 2F, but to selectively porosify the layer 210b of the support pads only after the common epitaxy step during which the active stacks 103 of the LEDs L and the photodiodes P are formed simultaneously. In this third embodiment, the layer 210b is made porous in line with the LEDs L and is kept intact (non-porous) in line with the photodiodes P. This leads to at least partially relaxing the mechanical stresses in the active stack of LEDs L without applying this relaxation in the photodiodes P. This results in a reduction of the internal electric field in the active stack of LEDs compared to the internal electric field in the active stack of photodiodes.This decrease in the internal electric field in the active LED stack leads to a downward shift of the LED emission peak. Again, this at least partially compensates for the Stokes shift between the emission peak and the absorption peak of the active stack. Thus, the LED emission peak moves closer to the photodiode absorption peak, improving the system efficiency.
[0112] Figures 3A to 3E are sectional views schematically illustrating steps of an example of implementation of a method for manufacturing an optoelectronic device according to the third embodiment.
[0113] [Fig.3A] illustrates a structure comprising a semiconductor support stack 210 on one face of a support substrate 101. The support stack 210 and the support substrate 101 are for example identical or similar to what has been described previously in relation to [Fig.2A].
[0114] [Fig.3A] further illustrates a step of forming an active stack 103 of LED and photodiode on the upper face of the support semiconductor stack 210. The active stack 103 is for example identical or similar to what has been described previously, in particular in relation to [Fig.lA]. The layers 103a, 103b and 103c are for example formed successively by epitaxy from the upper face of the support stack 210. By way of example, the lower semiconductor layer 103a of the active stack 103 is in contact, by its lower face, with the upper face of the layer 210c.
[0115] At this stage, the layers of the support stack 210 and the layers of the active stack 103 each extend continuously and with a uniform thickness over the entire surface of the support substrate 101.
[0116] [Fig.3B] illustrates a step of forming trenches 320 in the active stack 103 and in the support stack 210, from the upper face of the active stack 103, for example by lithography then etching, so as to define in the stack 210 a plurality of support pads SL and SP in the form of islands or mesas, each support pad SL being coated, on its upper face, by a portion of the active stack 103 defining an LED L of the device, and each support pad SP being coated, on its upper face, by a portion of the active stack 103 defining a photodiode P of the device.
[0117] In the example shown, the trenches 320 extend vertically from the upper face of the active stack 103, pass entirely through the layers 103c, 103b, 103a, 210c and 210b, and open into the layer 210a without passing entirely through it. As a variant, the trenches 220 pass entirely through the layer 210.
[0118] The trenches 320 form, for example, in top view, a grid or a grid laterally separating the LEDs L and the photodiodes P and the support pads SL and SP.
[0119] The LEDs L and the photodiodes P, and the underlying support pads SP and SL, for example all have the same lateral dimensions, for example between 1 μm and 25 μm, for example between 2 and 8 μm. For example, the LEDs L and the photodiodes P and the support pads SP and SL have, in top view, a square or rectangular shape. More generally, the LEDs L and the photodiodes P can have any shape, for example round or hexagonal.
[0120] At this stage, in each support pad SL and SP, the sides of the doped semiconductor layer 210b of the support stack are exposed.
[0121] [Fig.3C] illustrates the structure obtained at the end of a step of selective porosification of the layer 210b, located only in the support pads SL of the LEDs L of the device. This step is similar to what was previously described in relation to [Fig.2C], with the difference that, in the example of [Fig.3C], the layer 210b of the support pads SL is made porous, while the layer 210b of the support pads SP support remains intact (non-porous).
[0122] For this, during the electroporosification step, the sides of the support pads SP can be protected from contact with the electrolyte by a protective layer (not visible in the figure), the sides of the support pads SL being on the other hand in contact with the electrolyte.
[0123] In the example of [Fig.3C], the bias voltage allowing a current to flow in the layer 210b is for example applied between a first electrode (not visible in the figure) connected to the layer 210a and the electrolyte (not visible in the figure) connected by the edge to the layer 103c.
[0124] Due to the porosification of the layer 210b in the support pads SL, a greater mechanical relaxation is obtained in the active stack of LEDs L than in the active stack of photodiodes P. This leads to shifting the emission peak of the LEDs downwards, and thus bringing it closer to the absorption peak of the photodiodes P.
[0125] [Fig.3D] illustrates the structure obtained at the end of a step similar to what was described previously in relation to [Fig.2E] of forming, on each LED L, a contact metallization 232L on and in contact with the upper face of the upper semiconductor layer 103c of the active stack 103 of the LED, and, on each photodiode P, a contact metallization 232P on and in contact with the upper face of the upper semiconductor layer 103c of the active stack 103 of the photodiode.
[0126] [Fig.2E] further illustrates a step of filling the trenches 320 and the space between the LEDs L and the photodiodes P with an electrically insulating material 234, for example silicon oxide.
[0127] After filling, a planarization step can be implemented, for example by chemical mechanical polishing (CMP), so that the contact metallizations 232L, 232P are flush with the upper face of the filling material 234.
[0128] [Fig.3E] illustrates a step similar to what was described above in relation to [Fig.2F], of transferring and fixing the structure of [Fig.3D] on a control integrated circuit 110, and of removing the support substrate 101, and, possibly, all or part of the support semiconductor stack 210.
[0129] In a similar manner to what has been described previously, subsequent steps can then be implemented to resume individual or common electrical contact on the upper semiconductor layer 103a of the active stack 103 of each LED L and of each photodiode P.
[0130] Similar to what has been described previously, the control circuit can optionally be configured to drive the LEDs and photodiodes with densities carrier adapted to reduce the shift between the emission peak of the LEDs and the absorption peak of the photodiodes.
[0131] Various embodiments and variations have been described. Those skilled in the art will understand that certain features of these various embodiments and variations could be combined, and other variations will occur to those skilled in the art. In particular, the described embodiments are not limited to the examples of materials and dimensions mentioned in the description.
[0132] Furthermore, although exemplary embodiments have been described above in which the active LED and photodiode stacks 103 are attached to the control integrated circuit by direct full-plate metal-to-metal bonding or by direct metal-to-metal / dielectric-to-dielectric hybrid bonding, the described embodiments are not limited to these particular examples. More generally, the active LED and photodiode stacks 103 may be attached to the control integrated circuit by any other means, for example by direct full-plate oxide-to-oxide bonding.
[0133] Furthermore, it will be noted that the first and third embodiments can be combined.
[0134] Finally, the practical implementation of the embodiments and variants described is within the reach of those skilled in the art from the functional indications given above.
Claims
Claims
1. Method for manufacturing an optoelectronic device comprising at least one LED (L) and at least one photodiode (P), comprising the following successive steps: a) forming, by epitaxy, an active semiconductor stack (103) for emission and reception common to the LED and the photodiode; b) forming trenches (120) extending vertically through the active stack (103) and laterally delimiting the LED (L) and the photodiode (P), in which the trenches are arranged so that the lateral dimensions of the LED (L) are smaller than the lateral dimensions of the photodiode (P).
2. The method of claim 1, wherein the trenches (120) are arranged such that the lateral dimensions of the LED (L) are at least twice smaller than the lateral dimensions of the photodiode (P).
3. A method according to claim 1 or 2, wherein the trenches (120) are arranged so that the lateral dimensions of the LED (L) are at least four times smaller than the lateral dimensions of the photodiode (P).
4. A method according to any one of claims 1 to 3, wherein the trenches (120) are arranged such that the lateral dimensions of the LED (L) are less than 4 pm.
5. Method according to any one of claims 1 to 4, comprising, between step a) and step b), a step of transferring and fixing the active stack (103) on a face of an integrated control circuit (110) previously formed in and on a semiconductor substrate (101).
6. Method according to claim 5, wherein during said transfer and fixing step, the active stack (103) is fixed on said face of the integrated control circuit (110) by molecular bonding.
7. Method according to claim 5 or 6, wherein, at the end of said transfer and fixing step, the active stack (103) extends continuously over the entire surface of the integrated control circuit (110).
8. A method according to any one of claims 1 to 7, wherein the active semiconductor stack (103) comprises one or more IILV or ILVI type semiconductor alloys.
9. Optoelectronic device comprising at least one LED (L) and at least one at least one photodiode (P) each comprising an active semiconductor stack (103) for emission and reception of the same nature and composition, in which the lateral dimensions of the LED (L) are smaller than the lateral dimensions of the photodiode (P).
10. Device according to claim 9, further comprising a control integrated circuit (110) on one face of which the LED (L) and the photodiode (P) are fixed, the control integrated circuit (110) being adapted to drive the LED with a higher current density than that of the photodiode (P).
11. Device according to claim 10, in which the integrated control circuit (110) is adapted to drive the LED (L) with a current density at least ten times higher than that of the photodiode (P).