Integrated photonic device comprising an electro-optical layer and method for manufacturing this device

Abstract

The invention relates to a photonic device comprising a support substrate (1e), a layer of photonic components (2) disposed on and in contact with the support substrate (1e) and comprising at least one waveguide (2a) disposed in a covering material (2), a buried dielectric layer (1b) disposed on the layer of photonic components (2), in contact with the waveguide (2a), and an electro-optical layer on the buried dielectric layer (1b) extending at least partially over the waveguide (2a) in order to develop a hybrid optical mode in the waveguide (2a) and in the electro-optical layer (3). The invention also relates to a method for preparing such a device. Figure 1g

Description

Title of the invention: integrated photonic device comprising an electro-optical layer and method of manufacturing this device FIELD OF THE INVENTION

[0001] The present invention relates to an integrated photonic device comprising an electro-optical layer. The invention also relates to a method of manufacturing such a device. TECHNOLOGICAL BACKGROUND OF THE INVENTION

[0002] Electro-optical materials are materials whose optical properties are modified by the application of an electrical quantity. These may be materials developing the Pockels effect, such as lithium niobate (LNO) or barium titanate (BTO). These materials are promising materials for, in particular, the production of integrated ultra-high-speed, low-loss optical modulators. Reference may be made to document WO2021202853 or to that of F. Eltes et al, "Thin-film BTO-based modulators enabling 200 Gb / s data rates with sub 1 Vpp drive signal," in Optical Fiber Communication Conference (OFC) 2023, Technical Digest Series (Optica Publishing Group, 2023), paper Th4A.2.

[0003] The paper by Shayan Mookherjea, Viphretuo Mere, Forrest Valdez; Thin-film lithium niobate electro-optic modulators: To etch or not to etch. Appl. Phys. Lett. 20 Mardi 2023; 122(12):120501 presents several approaches to form modulators from a thin-film electro-optic material. These approaches involve etching the electro-optic material to form the waveguide in the thin-film itself and / or forming a waveguide on the electro-optic material by deposition and patterning and / or bonding the thin-film electro-optic material to a support substrate already incorporating a waveguide. In all cases, we seek to develop light radiation, i.e. an optical mode, propagating at least partly in the electro-optical material, with a view to modifying a characteristic of this light radiation by means of an electric field applied to the material.Each of the approaches mentioned in the aforementioned document poses technological problems of integration or performance.

[0004] For example, the etching of electro-optical materials generates particles of the etched composites, which makes it difficult to integrate it into an industrial process of silicon-based microelectronics (known as "CMOS technology"). This etching also leads to the formation of the side walls of the non-vertical waveguides, which affects their performance.

[0005] In the bonding approach, the thin layer of electro-optical material is transferred onto a support substrate incorporating waveguides, based on silicon and / or based on silicon nitride, arranged in an encapsulation material, for example silicon dioxide. This approach requires very good uniformity of the thickness of the layer of encapsulation material separating the thin layer of electro-optical material from the waveguide. This uniformity is difficult to obtain due to the surface topology linked to the presence of the photonic components, such as the waveguide and / or the metal tracks, in the support substrate. This topology is all the more important as a substantial stack of elements is formed in the component layer, in particular metal tracks.

[0006] This uniformity issue is documented and quantified in the paper by Peter O. Weigel et al, "Bonded thin film lithium niobate modulator on a Silicon photonics platform exceeding 100 GHz 3-dB electrical modulation bandwidth" Opt. Express 26, 23728-23739 (2018). In this paper, the thickness of the planarized encapsulation material layer, between the silicon-based waveguide and the lithium niobate film, varies with an amplitude of 150 nm over the 150 mm substrate extent. This thickness non-uniformity gives rise to variations in modulator bandwidth (from 50 to 250 GHz), and modulation efficiency (from 6 to 8 V.cm) due to variations in confinement of the optical mode in the hybrid guide. A thinner planarized layer will confine the mode further within the silicon-based waveguide, resulting in reduced modulation efficiency.

[0007] The document by F. Eltes et al., “A BaTiO3-Based Electro-Optic Pockels Modulator Monolithically Integrated on an Advanced Silicon Photonics Platform,” in Journal of Lightwave Technology, vol. 37, no. 5, pp. 1456-1462, 1 Marchl, 2019, doi: 10.1109 / JLT.2019.2893500 proposes a scheme for the monolithic integration of electro-optic modulators based on barium titanate. The thin film of this material is added to the integrated photonic device at the end of the integration process (“back end of the line” according to the Anglo-Saxon terminology used in the field), after the production of the metal interconnection lines (metal levels). This leads to placing the modulator at a relatively large distance (several micrometers) from the other photonic components, which are in turn placed under the metal interconnection lines and distant from them, in order to avoid or at least limit optical losses.However, in the integration scheme proposed by this document, the electro-optical layer is very far from the waveguides so that the optical coupling required between these components and the modulator is of low performance.

[0008] A similar approach, and with the same limitations, is proposed in the paper by Nicholas Boynton et al, “A heterogeneously integrated Silicon photonic / lithium niobate traveling wave electro-optic modulator, » Opt. Express 28, 1868-1884 (2020).

[0009] Generally speaking, it is sought to place the electro-optical material layer of a waveguide as close as possible to each other in order to allow efficient optical coupling between these two elements. This promotes the confinement of the optical mode in the electro-optical layer, which ensures better modulation efficiency. The electrodes making it possible to apply a modulation electric field to the electro-optical material layer must also be placed as close as possible to this layer for the same reasons of modulation efficiency. This need for compactness and in particular the presence of metal electrodes close to the waveguide leads to high optical losses. It is therefore often necessary in known integration schemes to find a compromise between the efficiency of the application of the modulation field and the optical losses. SUBJECT OF THE INVENTION

[0010] An aim of the invention is to remedy, at least in part, the problems which have just been presented. More particularly, an aim of the invention is to propose a scheme for integrating an electro-optical layer into an integrated photonic device making it possible to have a better compromise between optical losses versus modulation efficiency than in known solutions. BRIEF DESCRIPTION OF THE INVENTION

[0011] With a view to achieving one of these aims, the subject of the invention proposes a method for preparing a photonic device comprising the following steps: - providing a starting substrate comprising a base substrate, a buried dielectric layer disposed on and in contact with the base substrate and a surface layer having a refractive index greater than the refractive index of the buried dielectric layer disposed on and in contact with the buried dielectric layer; - treating the starting substrate to form a layer of photonic components, the photonic components being in a covering material and comprising at least one waveguide formed in the surface layer; - transfer the photonic component layer onto a support substrate and remove the base substrate to expose the buried dielectric layer; - transferring an electro-optical layer onto at least part of the buried dielectric layer, the electro-optical layer having a first surface in contact with the buried dielectric layer and a second surface opposite the first, the electro-optical layer extending at least partly above the waveguide in order to be able to develop a hybrid optical mode in the waveguide and in the electro-optical layer; - form at least two electrodes arranged laterally on either side of the waveguide in order to be able to develop a modulation electric field intercepting the hybrid optical mode.

[0012] According to other advantageous and non-limiting characteristics of the invention, taken alone or in any technically feasible combination: • the surface layer is formed of a semiconductor material; • the electrode formation step includes the deposition of a layer encapsulation in contact with the second face of the electro-optical layer and / or in contact with the buried dielectric layer; • the step of forming the electrodes comprises the deposition of metal pads above the second face of the electro-optical layer and / or above the buried dielectric layer; • the electrode formation step comprises the formation of at least two metal vias; • the step of treating the starting substrate comprises the formation of at least two doped pads in the semiconductor surface layer, the two doped pads (6) being arranged laterally on either side of the waveguide; • the step of processing the starting substrate comprises the formation of at least two metal tracks in the layer of photonic components.

[0013] According to another aspect, the invention provides a photonic device comprising: - a support substrate; - a layer of photonic components arranged on and in contact with the support substrate and comprising at least one waveguide arranged in a covering material; - a buried dielectric layer arranged on the layer of photonic components, in contact with the waveguide; - an electro-optical layer on the buried dielectric layer, the electro-optical layer having a first surface in contact with the buried dielectric layer and a second surface, opposite the first, extending at least partly above the waveguide in order to be able to develop a hybrid optical mode in the waveguide and in the electro-optical layer; - at least two electrodes arranged laterally on either side of the waveguide in order to develop a modulation electric field intercepting the hybrid optical mode.

[0014] According to other advantageous and non-limiting characteristics of the invention, taken alone or in any technically feasible combination: • the buried dielectric layer has a thickness of less than 300 nm, preferably less than 100 nm and / or has a thickness uniformity of less than or equal to 50 nm; • the waveguide is formed from a semiconductor material, preferably formed from silicon, or silicon nitride; • the waveguide extends longitudinally on either side of the electro-optical layer, the waveguide having a reduced thickness at the level of the electro-optical layer than in the rest of the waveguide; • the device comprises an encapsulation layer in contact with the second face of the electro-optical layer and / or in contact with the buried dielectric layer; • the at least two electrodes respectively comprise two metal pads arranged above the electro-optical layer and / or above the buried dielectric layer; • the at least two electrodes comprise at least two metal vias; • the at least two electrodes respectively comprise at least two pads formed from a doped semiconductor material and arranged in the layer of photonic components, the two doped pads being arranged laterally on either side of the waveguide; • the at least two electrodes respectively comprise two metal tracks arranged in the layer of photonic components, the metal tracks being located in planes placed between the waveguide (2a) and the support substrate; • the photonic device comprises two waveguides arranged at least partly under the electro-optical layer, the two waveguides forming the two arms of a modulator. Brief description of the drawings

[0015] Other characteristics and advantages of the invention will emerge from the detailed description of the invention which follows with reference to the appended figures in which:

[0016] [Fig.la][Fig.lb][Fig.lc][Fig.ld][Fig.le][Fig.lf][Fig.lg][Fig.lh]

[0017] Figures 1a, 1b, 1c, 1d, 1e, 1f, 1g, 1h represent steps of a preparation process in accordance with the invention;

[0018] [Fig.2a] [Fig.2b] [Fig.2c] [Fig. 3a] [Fig.3b]

[0019] Figures 2a, 2b, 2c, 3a, 3b represent different embodiments of an integrated photonic device according to the invention;

[0020] [Fig.4]

[0021] [Fig.4] represents an integrated photonic circuit according to the invention implementing operates a modulator;

[0022] [Fig.5a] [Fig.5b]

[0023] Figures 5a, 5b represent different views of a first mode of implementation of the photonic component of [Fig.4];

[0024] [Fig.6a] [Fig.6b]

[0025] Figures 6a, 6b represent different views of another mode of implementation of the photonic component of [Fig.4];

[0026] [Fig.7a][Fig.7b][Fig.7c]

[0027] Figures 7a, 7b, 7c represent other embodiments of an integrated photonic device according to the invention. DETAILED DESCRIPTION OF THE INVENTION

[0028] Figures 1a to 1h show the different steps of a method according to the invention.

[0029] In a first step shown in [Fig. 1a], a starting substrate 1 is provided. This substrate, for example of the silicon-on-insulator type, is formed from a base substrate 1a, typically made of silicon and having a thickness of several hundred microns. A buried dielectric layer 1b, typically made of silicon oxide, is arranged on and in contact with the base substrate 1a. This layer may have a thickness of between a few tens of nanometers and several thousand nanometers. In the context of the present description, and for reasons which will become apparent in the remainder of this presentation, it is preferable to choose this layer 1b to be relatively thin, for example between 5 nm and 300 nm, and preferably between 5 nm and 200 nm or between 5 nm and 100 nm.According to an important characteristic, this buried dielectric layer 1b has good thickness uniformity, with thickness variations less than 50 nm over the entire extent of the substrate, or even less than 5 nm, or 3 nm or less than 1 nm. Finally, the starting substrate 1 comprises a semiconducting surface layer 1e arranged on and in contact with the buried dielectric layer 1b. This layer 1e can in particular be formed from monocrystalline silicon and have a thickness of between 50 nm and 1000 nm.

[0030] The starting substrates 1 of silicon-on-insulator type are widely available on the market, in a wide variety of dimensions, and for thickness specifications, in particular of the semiconducting surface layer 1c and of the buried dielectric layer 1b, which can be freely chosen. The provision of this starting substrate 1, which constitutes the first step of a method according to the invention, therefore does not involve any particular difficulty. These substrates have in particular a uniformity of thickness of the buried dielectric layer 1b less than the preferred one of 50nm, over the entire extent of the substrate.

[0031] It is noted that the starting substrate may be of a different type than that which has just been presented. In particular, the surface layer 1c may be chosen to be of a different nature than that, semiconducting, of a semiconductor material. It may in particular be silicon nitride.

[0032] More generally therefore, the starting substrate 1 is formed of a base substrate 1a, which may be made of silicon, a buried dielectric layer 1b, which may be made of silicon dioxide, and a surface layer 1c having a refractive index greater than the refractive index of the buried dielectric layer 1b. This surface layer may in particular be made of or comprise silicon nitride.

[0033] During a second step of a method according to the invention, illustrated in [Fig.lb], the starting substrate 1 is treated to form, in and on the surface layer 1c, components, in particular photonic components.

[0034] As is well known in itself, this treatment can comprise any type of conventional technological steps in the world of microelectronics: deposition, etching, photolithography, in order to define patterns making it possible to make these components functional... By way of illustration, these conventional technological steps can be chained together to form, by etching the surface layer 1c, waveguides 2a. Selective deposition steps, for example of germanium, make it possible to form a photodetector 2c. Silicon nitride deposition steps make it possible to form additional waveguides 2b. It is also possible to form metal tracks 7, at different levels, to conduct electrical signals.Multiple deposition steps, which may be followed by a polishing step, of a covering material 2e, for example silicon dioxide, makes it possible to encapsulate the assembly and form a component layer 2 resting on the base substrate 1a, via the buried dielectric layer 1b. The covering material 2e may be opened by etching, during the initial substrate processing step to form metal vias 2d, which may come into contact with the metal tracks 7, making it possible, for example, to make contacts on the active components of the component layer or to produce the modulation field of the electro-optical layer as will be explained later.

[0035] At the end of this processing step, and regardless of the components formed in and on the surface layer 1e, a layer of photonic components 2 is available, these components being embedded in a covering material 2e. The photonic components 2a, 2b, 2c, 2d, 7 of the layer 2 comprise at least one waveguide 2a formed in the surface layer 1e. It is noted that this waveguide 2a having been formed in the surface layer 1e, it rests on and is in contact with the buried dielectric layer 1b.

[0036] At the end of this treatment step, it may be provided to polish the exposed surface of the layer of photonic components 2 in order to facilitate the following transfer step.

[0037] In a following step shown in [Fig. 1c], the layer of photonic components 2 is transferred onto a support substrate 1c. This transfer can be carried out by any suitable technique. This transfer generally comprises the assembly of the layer of photonic components 2 carried by the base substrate 1a with this support substrate 1c. This may for example be an assembly by molecular adhesion. Once the assembly has been carried out, the base substrate 1a is removed to expose the buried dielectric layer 1b. This removal can be carried out by grinding assisted by dry or wet etching, the buried dielectric layer 1b forming a particularly effective barrier layer for this etching. [Fig. 1d] shows the layer of photonic components 2 transferred onto the support substrate 1c at the end of this step.It is possible at this stage to thin the buried dielectric layer 1b, for example by chemical etching, to give it a chosen thickness. In this case, care should be taken not to excessively degrade its thickness uniformity.

[0038] In a following step of the preparation method according to the invention, an electro-optical layer 3 is transferred onto the buried dielectric layer 1b. This electro-optical layer 3 may for example be made of lithium niobate (LNO) or barium titanate (BTO). The layer 3 preferably has a monocrystalline quality. It may be transferred by assembly-thinning of a substrate made of bulk electro-optical material or of a substrate on which a layer of electro-optical material has been previously deposited, typically by epitaxy. It may also be a substrate of the “electro-optical material on insulator” type comprising a surface layer of electro-optical material, for example lithium niobate, transferred onto a handling substrate, for example a silicon substrate provided with a layer of silicon oxide.

[0039] In any event, the electro-optical layer 3 which has been placed on the component layer 2 has a first surface in contact with the buried dielectric layer 1e and a second surface opposite the first. The electro-optical layer 3 extends at least partly above the waveguide 2a in order to be able to develop a hybrid optical mode in the waveguide and in the electro-optical layer.

[0040] It is noted in this regard that the electro-optical layer 3 is separated from the waveguide 2a only by the thickness of the buried dielectric layer 1b. It was indicated previously that the latter was preferably chosen to be relatively thin and uniform in thickness, and it is now understood that this choice makes it possible to facilitate the development of the hybrid optical mode when the integrated photonic device is in operation.

[0041] It is not necessary for the electro-optical layer 3 to extend over the entire extent of the buried dielectric layer 1b. It may be preferable to preserve a portion of this layer 1b not covered by the electro-optical layer 3, for example to more easily access certain underlying photonic components or to allow the entry / exit of light radiation into the integrated device as will be shown in one of the particular embodiments of the invention, described in a later section of this description.

[0042] The configuration in which the electro-optical layer does not entirely cover the buried dielectric layer 1c, shown in [Fig.lf], can be obtained in multiple ways. According to a first approach, the step of transferring the electro-optical layer 3 can consist of vignetting, that is to say transferring one or a plurality of blocks formed from the electro-optical material, each block having a reduced dimension, and smaller than that of the exposed surface of the buried dielectric layer 1b, but sufficient to overhang, at least in part, at least one waveguide of the photonic component layer 3. This approach leads to directly obtaining the configuration of [Fig.lf] directly after transferring the electro-optical layer 3. Advantageously, in order to simplify the manufacturing method, it is sought to limit the number of blocks transferred onto the buried dielectric layer 1b.For this purpose, a pad may overhang a plurality of waveguides of the photonic component layer 3. .

[0043] According to another approach, an electro-optical layer 3 is applied, completely covering the buried dielectric layer 1b, and then a portion of the electro-optical layer 3 is removed to expose a portion of the buried dielectric layer 1b. This removal can be obtained by etching.

[0044] It is of course possible to preserve the electro-optical layer 3 covering the buried dielectric layer 1b, as shown in [Fig. 1e], and the two approaches which have just been presented are perfectly optional. In all cases, the electro-optical layer 3 is preserved at least to partially cover the waveguide 2a in order to be able to develop the hybrid optical mode.

[0045] A method according to the invention also comprises a step aimed at forming at least two electrodes arranged on one side and the other of the waveguide 2a (and as close as possible to the electro-optical layer) in order to be able to develop a modulation electric field intercepting the hybrid optical mode. The modulation electric field, imparted by the electrical potential difference between the two electrodes, makes it possible to vary the optical properties of the electro-optical material in order to modulate the light radiation propagating therein. In the context of this description, the term "electrode" means the elements making it possible to form a conductive path making it possible to conduct the modulation electric signal as close as possible to the electro-optical layer 3.

[0046] This step of forming the electrodes may consist of forming two pads 4a by depositing a metallic material, for example gold, on the second face of the electro-optical layer 3, that is to say on its exposed surface. This configuration is shown in [Fig.lg].

[0047] In [Fig. 1h], an encapsulation layer 5 has been formed on the second face of the electro-optical layer 3 to cover it. The two metal pads 4a are arranged on the encapsulation layer 5 and, by etching this layer, two metal vias 4b passing through the encapsulation layer 5 have been formed. In the embodiment of [Fig. 1h], the two metal vias 4b are in contact with, on the one hand, the metal pads 4a and, on the other hand, with the second face of the electro-optical layer 3. The vias can be made of aluminum.

[0048] It should be noted that the preparation method can be separated into two distinct sequences: the first sequence of steps of the method aims to prepare the layer of photonic components 2 transferred onto the support substrate 1c and the second sequence of steps of the method, which follows the first sequence, aims to transfer the electro-optical layer 3 covering the buried dielectric layer 1b and to form at least part of the electrodes. In particular, these two sequences can be carried out in different industrial environments: the first sequence can be carried out in a CMOS technological environment, incompatible with certain steps forming the second sequence, in particular the etching of the electro-optical material and the use of gold for the formation of the metal pads 4a. This second sequence can be implemented in a compatible environment, different from the CMOS technological environment in which the first sequence is implemented.

[0049] In any event, at the end of the method which has just been described, we have an integrated photonic device comprising on a support substrate: - a layer of photonic components 2 arranged on and in contact with the support substrate 1e and comprising at least one waveguide 2a arranged in a covering material (2); - a buried dielectric layer 1b arranged on the layer of photonic components 2, in contact with the waveguide 2a. - an electro-optical layer 3 on the buried dielectric layer 1b, the electro-optical layer 3 having a first surface in contact with the buried dielectric layer 1b and a second surface opposite the first and extending at least partly above the waveguide 2a in order to be able to develop a hybrid optical mode in the waveguide and in the electro-optical layer; - at least two electrodes arranged laterally on either side of the waveguide 2a in order to develop an intercepting modulation electric field hybrid optical mode.

[0050] The buried dielectric layer 1b advantageously has a low thickness, for example less than 300nm, which makes it possible to establish easy and lossless (or low loss) coupling between the waveguide 2a and the electro-optical layer 3. It also has a thickness uniformity less than or equal to 50nm.

[0051] It is noted that the metal tracks 7 are located in this device in planes placed between the waveguide 2a and the support substrate 1c, which clearly shows the fact that the component layer has been transferred from the base substrate 1a onto the support substrate 1c.

[0052] [Fig.2a] shows another embodiment of a photonic device according to the invention. This embodiment makes it possible to move the metal part of the electrodes away from the hybrid optical mode, which further reduces optical losses in comparison with the embodiment of [Fig.1h] or [Fig.1g]. In this configuration the electrodes comprise doped semiconductor pads 6, arranged in the layer of photonic components 2. The waveguide 2a is undoped to limit optical losses.

[0053] [Fig.2a] shows the base substrate 1c on which rests, successively, the layer of photonic components 2, the buried dielectric layer 1b and the electro-optical layer 3. The latter is arranged overhanging a waveguide 2a as in the main implementation mode.

[0054] In the embodiment of [Fig.2a], two pads 6 formed from the doped semiconductor material are also provided, the two doped pads 6 being arranged laterally on either side of the waveguide. These two pads are formed during the step of processing the starting substrate 1 ([Fig.lb]) in the semiconductor surface layer. The doping can be carried out by ion implantation, as is well known per se. The doping makes it possible to make these pads 6 from conductive semiconductor materials and to limit optical losses. This doping can be of the p or n type and have a dopant density of the order of 2.0 10A19 atoms per cmA3. This doping can be uniform over the entire extent of the pad or present a gradient from the end 6b furthest from the waveguide 2a where the dopant density can be of the order of 2.0 10A19 at / cmA3 towards the end 6a closest to the waveguide 2a, where the dopant density can be of the order of 2. 10A18 at / cmA3

[0055] Also shown in [Fig.2a] are the metal pads 4a formed on the encapsulation layer 5 and metal vias 4b passing through the encapsulation layer 5 and the electro-optical layer 3, the metal vias 4b being respectively electrically connected to at least two doped pads 6. The metal pads 4a and the metal vias 4b are produced during the electrode formation step. In this embodiment, the metal vias 4b directly contact the doped pads 6 at their ends 6b furthest from the waveguide 2a and the therefore deviate from the hybrid optical mode. The electrodes are therefore formed from metal pads 4a, metal vias 4b and doped pads 6.

[0056] This embodiment is advantageous in that the modulation electric field develops from the end 6a of the doped pads 6 located closest to the waveguide 2a, which ensures high efficiency of application of this modulation electric field on the electro-optical layer 3, at the level of the optical mode which develops there. Being formed of a semiconductor and non-metallic material, the pads 6 have little effect on the optical mode which develops and the optical losses remain limited.

[0057] [Fig.2b] represents a variant of the embodiment of [Fig.2a] which allows to exploit metal tracks 7 formed in the layer of photonic components 2 in order to have better electrical contacts with the doped pads 6. Indeed, the metal tracks 7 formed during the manufacturing step of the layer of components 2 benefit from a surface layer of silicide, known to form very good electrical contacts. To do this, during the step of processing the starting substrate 1, two electrically conductive metal tracks 7 were provided in the layer of photonic components 2, encapsulated in the covering material and arranged under the doped pads 6. These two metal tracks 7 are respectively in electrical contact with the two doped pads 6. In this embodiment, the metal vias 4b of the electrodes are respectively in contact with the two metal tracks 7, by their “landing” ends.The electrodes are therefore respectively formed of metal pads 4a, metal vias 4b, metal tracks 7 and doped pads 6.

[0058] In the variant of [Fig.2c], the two electrodes are sufficiently far apart laterally from each other so that the metal vias 4b do not pass through the electro-optical layer 3. The two electrodes are formed of metal pads 4a arranged on the encapsulation layer 5 (or, when it is absent, on the buried dielectric layer 1e) and the metal vias 4b are respectively electrically connected at least to the two doped pads 6, without passing through the electro-optical layer 3.

[0059] Figures 3a and 3b show another embodiment of a photonic device according to the invention also making it possible to move the metal portion of the electrodes away from the hybrid optical mode. In this embodiment, the electric field is applied to the electro-optical layer 3 by the metal tracks 7 arranged in the photonic component layer 2. These metal tracks 7 are electrically connected to the metal pads 4a by means of the metal vias 4b with which they are in contact. To facilitate the routing of the electrical signals, these metal tracks 7 may be in contact with other metal tracks 7a arranged at distinct levels in the photonic component layer 2, such as this is shown in the figures, these tracks 7,7a being connected to each other by intermediate vias. As already noted previously, the metal tracks 7,7a are located in planes placed between the waveguide 2a and the support substrate 1e.

[0060] In the variant of [Fig.3a], the two metal pads 4a are formed on the encapsulation layer 5 and the vias 4a pass through the encapsulation layer 5 and the electro-optical layer 3. In the variant of [Fig.3b], the two metal pads 4a are sufficiently laterally distant from each other so that the vias 4b do not pass through the electro-optical layer 3. The two metal pads 4a are formed on the encapsulation layer 5 and the metal vias 4b are respectively electrically connected to at least two doped pads 6, passing through the encapsulation layer 5 but not the electro-optical layer 3. As already noted, the encapsulation layer is optional and, in its absence, the metal pads 4a are then arranged in contact with the buried dielectric layer 1b.

[0061] [Fig.4] represents an embodiment implementing a Mach-Zehnder type modulator. Two waveguides 2a are arranged under the electro-optical layer 3 and three electrodes 4G, 4S are arranged to apply an electric field dedicated to the two hybrid optical modes which develop respectively in the waveguides 2a and the electro-optical layer. As is well known per se, the central electrode 4S is generally intended to carry the modulation signal, and the two lateral electrodes 4G an electrical ground. It is noted that the electrodes 4S, 4G can extend over the entire length of the modulator MOD to increase the modulation efficiency.

[0062] A two-branch MOD modulation block is thus formed. A plurality of such modulation blocks could naturally be provided in an integrated photonic device according to the invention.

[0063] [Fig.5a] represents a top view of a very advantageous example of implementation of the device of [Fig.4]. This figure clearly shows the electro-optical layer 3 (in the form of a thumbnail) overhanging two waveguides 2a, and metal tracks 7. These metal tracks 7 are arranged in planes lower than the waveguides 2a in the layer of photonic components 2 (as in the configuration of [Fig.3a]). The waveguides 2a and the metal tracks 7 extend longitudinally beyond the electro-optical layer 3, on either side of this layer 3. The end portions of the waveguides 2a and the metal tracks 7 are therefore not overhung by this layer 3, which in particular makes it possible to place the metal pads 4a at the end portions of the metal tracks 7 and thus prevent the metal vias 4b from passing through this layer 3.

[0064] In this top view, the waveguides 2a and the metal tracks 7 extend to longitudinally define a succession of zones: an input zone zl, a transition zone z2, a core zone in which the modulation takes place, a new transition zone z2 and an output zone z4.

[0065] [Fig.5b] represents a section of the device shown in [Fig.4], respectively at the level of the input zone zl and output z2 of the modulator MOD (section AA in [Fig.4]), at the level of the transition zone z2 (section BB) and at the level of the core zone z3 (section CC).

[0066] In this [Fig.5b] we observe the evolution of the thickness of the waveguide 2a according to the different zones of the modulator. We observe in particular that this thickness Ep2 is less in the core zone z3 than the thickness Epi in the transition zones z2 and input-output zl. A relatively thick thickness in the transition zone z2 makes it possible to avoid reflections linked to the presence of the electro-optical layer 3. A relatively thin thickness in the core zone Z3 makes it possible to obtain increased confinement in the material of the electro-optical layer 3, and therefore greater modulation efficiency. In this advantageous configuration, the waveguide 2a, which extends longitudinally on either side of the electro-optical layer 3, therefore has a smaller thickness at the level of the electro-optical layer 3 than in the rest of the waveguide 2a.

[0067] It is also preferable to use a wider waveguide 2a in the input-output zones z1, z4 and in the transition zone z2 than in the core zone z3 to limit these reflections. The reduced width waveguide in the core zone z2 of the modulator makes it possible to deconfine the optical mode of the waveguide 2a and confine it further in the material of the electro-optical layer 3.

[0068] The configuration shown in Figures 5a, 5b therefore makes it possible to control the MH optical mode (shown in [Fig.5b]) so that it propagates efficiently and develops, at the level of the core zone Z3, in the electro-optical layer 3, in order to modulate this MH optical mode efficiently.

[0069] It will be noted that in the method presented with reference to Figures 1a to 1h, it is particularly easy to shape and structure the semiconducting surface layer 1c to give it the desired width and thickness.

[0070] [Fig.6a] represents a top view and [Fig.6b] represents a section of the device of [Fig.4], at the level of the core zone z3 (section CC of [Fig.6a]), when the electrodes are implemented by doped pads 6 as has been explained in relation to the description of figures 2, 2a and 2b. In this configuration, the metal tracks 7 make it possible to electrically contact the doped pads 6.

[0071] Of course, the invention is not limited to the embodiments described and variant embodiments can be made without departing from the scope of the invention as defined by the claims.

[0072] It is noted that the encapsulation layer 5, in all the implementation modes exposed, is perfectly optional. An illustration without encapsulation layer 5 is shown in [Fig.7c]. In this case, the metal pads 4a can rest directly on the buried dielectric layer 1b.

[0073] Furthermore, it is possible to integrate other photonic components into the device than those which have already been presented explicitly. By way of illustration, the support substrate 1c can be provided with a cavity 8 filled with air. This air cavity 8, arranged under and in line with the waveguide 2a, can be used to improve the adaptation between the optical mode and the radiofrequency mode, thus improving the modulation bandwidth. This cavity 8 is illustrated in the embodiment of [Fig. 4], but is applicable to all the embodiments set out in the present description.

[0074] When the photonic device aims to form a modulator, the latter is not necessarily implemented by a Mach Zehnder interferometer as has been illustrated in the present description. The modulator(s) can thus be implemented, for example by resonant ring modulators or by Michelson type interferometers.

[0075] Furthermore, the photonic device may comprise other components exploiting the properties of the electro-optical layer in addition to or as a replacement for the modulator(s) taken as an example.

[0076] In Figures 7a, 7b, a photonic device is shown comprising two modulation blocks MOD and in which the electro-optical layer 3 does not entirely cover the buried dielectric layer 1b. This configuration is used to place, at the level of the buried dielectric layer 1b exposed and not covered by the electro-optical layer, surface couplers GC, GC' making it possible to inject / produce external light radiation into the photonic device or from this device. In [Fig.5a], the surface coupler GC is formed in the semiconductor layer 1c of the starting substrate, during the step of forming the layer of photonic components 2. A metal mirror M has been placed, embedded in the covering material of the covering layer, making it possible to guide the light radiation towards the waveguide 2a. In the case of [Fig.5b], the surface coupler GC' is also embedded in the covering material.It may be composed of silicon nitride. Again, the surface couplers illustrated in these figures 5a, 5b may be incorporated into all the embodiments set forth in the present description.

Claims

Claims

1. A method for preparing a photonic device comprising the following steps: - providing a starting substrate (1) comprising a base substrate (1a), a buried dielectric layer (1b) arranged on and in contact with the base substrate (1a) and a surface layer (1e) having a refractive index greater than the refractive index of the buried dielectric layer (1e) arranged on and in contact with the buried dielectric layer (1b); - treating the starting substrate (1) to form a layer of photonic components (2), the photonic components being in a covering material (2e) and comprising at least one waveguide (2a) formed in the surface layer (1e); - transferring the layer of photonic components (2) onto a support substrate (1e) and removing the base substrate (1a) to expose the buried dielectric layer (1b);- transferring an electro-optical layer (3) onto at least part of the buried dielectric layer (1b), the electro-optical layer (3) having a first surface in contact with the buried dielectric layer (1b) and a second surface opposite the first, the electro-optical layer (3) extending at least partly above the waveguide (2a) in order to be able to develop a hybrid optical mode in the waveguide (2a) and in the electro-optical layer (3); - forming at least two electrodes arranged laterally on either side of the waveguide (2a) in order to be able to develop a modulation electric field intercepting the hybrid optical mode.;

2. Preparation method according to the preceding claim in which the surface layer (Ic) is formed from a semiconductor material.

3. Preparation method according to one of the preceding claims in which the step of forming the electrodes comprises the deposition of an encapsulation layer (5) in contact with the second face of the layer electro-optical (3) and / or in contact with the buried dielectric layer (1b).

4. Preparation method according to one of the preceding claims in which the step of forming the electrodes comprises the deposition of metal pads (4a) above the second face of the electro-optical layer (3) and / or above the buried dielectric layer (1b).

5. Preparation method according to one of the preceding claims in which the step of forming the electrodes comprises the formation of at least two metal vias (4b).

6. Preparation method according to one of the preceding claims in which the step of treating the starting substrate (1) comprises the formation of at least two doped pads (6) in the semiconductor surface layer (1e), the two doped pads (6) being arranged laterally on either side of the waveguide (2a).

7. Preparation method according to one of claims 5 to 7 in which the step of treating the starting substrate (1) comprises the formation of at least two metal tracks (7) in the layer of photonic components (2).

8. Photonic device comprising: - a support substrate (1e); - a layer of photonic components (2) arranged on and in contact with the support substrate (1e) and comprising at least one waveguide (2a) arranged in a covering material (2); - a buried dielectric layer (1b) arranged on the layer of photonic components (2), in contact with the waveguide (2a); - an electro-optical layer on the buried dielectric layer (1b), the electro-optical layer (3) having a first surface in contact with the buried dielectric layer (1b) and a second surface, opposite the first, extending at least partly above the waveguide (2a) in order to be able to develop a hybrid optical mode in the waveguide (2a) and in the electro-optical layer (3);- at least two electrodes arranged laterally on either side of the waveguide (2a) in order to develop a modulation electric field intercepting the hybrid optical mode.;

9. Photonic device according to the preceding claim in which the buried dielectric layer (1b) has a thickness of less than 300 nm, preferably less than 100 nm and / or has a thickness uniformity of less than or equal to 50 nm.

10. Photonic device according to one of claims 8 and 9, in which the waveguide (2a) is formed of a semiconductor material, preferably formed of silicon, or silicon nitride.

11. Photonic device according to one of claims 8 to 10 in which the waveguide (2a) extends longitudinally on either side of the electro-optical layer (3), the waveguide (2a) having a smaller thickness at the level of the electro-optical layer (3) than in the rest of the waveguide (2a).

12. Photonic device according to one of claims 8 to 11 comprising an encapsulation layer (5) in contact with the second face of the electro-optical layer (3) and / or in contact with the buried dielectric layer (1b).

13. Photonic device according to one of claims 8 to 12 in which the at least two electrodes respectively comprise two metal pads (4a) arranged above the electro-optical layer (3) and / or above the buried dielectric layer (1b).

14. Photonic device according to one of claims 8 to 13 in which the at least two electrodes comprise at least two metal vias (4b).

15. Photonic device according to one of claims 8 to 14 in which the at least two electrodes respectively comprise at least two pads (6) formed from a doped semiconductor material and arranged in the layer of photonic components (2), the two doped pads (6) being arranged laterally on either side of the waveguide (2a).

16. Photonic device according to one of claims 8 to 15 in which the at least two electrodes respectively comprise two metal tracks (7) arranged in the layer of photonic components (2), the metal tracks being located in planes placed between the waveguide (2a) and the support substrate (le).

17. Photonic device according to one of claims 8 to 16 comprising two waveguides (2a) arranged at least partly under the layer electro-optical (3), the two waveguides forming the two arms of a modulator (MOD).

Images