Photonic device for providing optical radiation involving optical modes in a waveguide - Patents.com

Abstract

The present invention relates to a photonic device (DP) for providing optical radiation, comprising a waveguide (2), an N-type semiconductor layer (1n) covering the waveguide (2), and an active region (QW) formed by a stack of III-V materials. The photonic device also comprises a plurality of P-type pillars (1p, 1'p) disposed in and contacted with the active region (QW). At least first metal pads (3n, 3'n) are in ohmic contact with the free portion of the N-type layer (1n), and at least second metal pads (3p, 3'p) are in ohmic contact with the P-type pillars (1p, 1'p).

Description

[Technical field]

[0001] The technical field of the invention is that of integrated optical elements using the properties of semiconductor materials capable of emitting light and conventionally used in integrated circuits for implementing logic and / or analog functions. The invention relates more specifically to photonic devices aimed at establishing optical radiation comprising optical modes in a waveguide. It may in particular be a heterogeneous laser on silicon. [Background technology]

[0002] Optical transceivers typically consist of active optical devices (such as lasers, modulators, photodiodes) and passive optical devices (such as waveguides, optical filters) optionally complemented by electronic circuits. These building blocks can be integrated into photonic devices using techniques and materials commonly used to fabricate integrated electronic circuits.

[0003] In such photonic devices, and as is well known per se, an active region formed by a stack of III-V materials forms the optical amplification medium of the laser. This active region may contain at least one quantum well, quantum dot or quantum box, or a number of such wells, dots or boxes. It is made of materials selected from the non-limiting list: InP, AsGa, InGaAlAs, InGaAsP, InAsP, InAs.

[0004] The active region is sandwiched between an N-type semiconductor layer and a P-type semiconductor layer. These layers are typically based on InP or AsGa, allowing current to circulate through the active region and electrically pump the gain medium to generate light. To inject and extract charge into and from the active region and enable this pumping, conductive metal pads are in ohmic contact with the P-type and N-type semiconductor layers, respectively. These pads are also electrically connected to electrical interconnect tracks, which allow charge to circulate within the device.

[0005] The active region is arranged along a section of a waveguide, for example made of silicon, called the "hybrid section". The optical mode generated in such an arrangement is called "hybrid" because it is partially located in the active region and in the waveguide. The waveguide extends on at least one side of the hybrid section to allow the propagation of the generated mode. Non-Patent Document 1 recalls that the more the optical mode overlaps with the current injection zone of the active region, the greater the amplification.

[0006] To allow this laser effect, the photonic device can also include an optical feedback structure to form a resonant cavity in the gain medium, which can be formed by a distributed reflector, for example a Bragg grating, placed in the active region, or preferably in a waveguide.

[0007] In general, we are interested in optical radiation having a wavelength useful for a chosen application. Thus, in the telecommunications field, such wavelengths are typically between 1200 and 1600 nm. Some elements of the photonic devices presented above are configured to emit radiation in a chosen wavelength range.

[0008] Figure 1 shows an embodiment of a photonic device according to the prior art. In a thin layer DL, on a support (not shown), a waveguide 2 made of silicon is formed. The latter extends longitudinally in the plane of the main direction X in the coordinate system (X, Y, Z) of Figure 1.

[0009] The waveguide has a contour at its edges (i.e. its contour in a plane perpendicular to the main direction of propagation of the guided light). It therefore consists of a ribbon 2a with a lateral dimension extending (in the Y direction) and a rib 2b with a smaller lateral dimension, which is placed below this ribbon 2a and is central to this ribbon 2a. The thickness (in the Z direction) of the ribbon 2a and the rib 2b is each of the order of 100 nm, typically between 50 and 500 nm. An N-type semiconductor layer 1n is placed on the thin layer DL, covering the waveguide 2. Between the waveguide 2 and the N-type semiconductor layer 1n, an electrical insulator, for example a silicon oxide layer, may be provided. This layer 1n has a thickness of the order of 100 nm.

[0010] On top of the N-type semiconductor layer 1n is an active region QW, on which a P-type semiconductor layer 1p rests. This assembly is disposed along a hybrid portion 2h of a waveguide 2, in particular along the ribs 2b of the waveguide 2, in contact with the N-type layer 1n. The hybrid portion of the waveguide is configured for the definition of a Bragg grating that defines the optical feedback structure of the device, for example the lateral relief shape of the waveguide 2 shown in FIG.

[0011] The assembly formed by the active region QW and the P-type semiconductor layer 1p has the shape of a slab with lateral dimensions smaller than those of the N-type semiconductor layer 1n, so that the free parts of this layer 1n located on either side of the active region QW are neither covered by the active region QW nor by the P-type semiconductor layer 1p. The active region QW has a relatively small thickness, on the order of 200 nm, and extends along the hybrid portion 2h of the waveguide, which may have a longitudinal distance on the order of 500 microns.

[0012] In the free part of the N-type semiconductor layer 1n, metal pads are arranged in ohmic contact with this layer, here two vias 3n, 3'n are arranged on either side of the active area QW. These metal vias 3n, 3'n provide an electrical connection between the N-type semiconductor layer 1n and electrical interconnection tracks IC protruding above the assembly. Similarly, metal pads 3p in the form of other vias are in ohmic contact with the P-type semiconductor layer 1p to ensure an electrical connection between this layer and another interconnection track IC (these tracks are not shown in the top view of FIG. 1 in order to maintain the readability of this view).

[0013] The optical mode M that is generated when a current is injected into the active region QW via the P-type semiconductor layer 1p is shown by a dotted line in the cross section of FIG. 1. It is observed that this mode has a very large extent in the direction perpendicular to the plane (X, Y) in which the waveguide is defined. In order to prevent this mode from disturbing the metal elements of the device, in particular the second metal pad 3p, the thickness of the P-type layer 1p is particularly large, greater than one micron and reaching up to several microns. In this way, the metal pads 3n, 3'n, 3p are sufficiently spaced from the active region QW and absorption of the optical radiation in the metal that constitutes the pads is avoided.

[0014] However, this configuration is not advantageous. It requires, first of all, a significant thickness of material forming the P-type semiconductor layer 1p, which is undesirable during the fabrication of the photonic device. Indeed, etching this significant thickness is long and produces a high surface topology, which is problematic for the remaining fabrication operations, in particular the formation of the metal pads. Moreover, the P-type semiconductor layer 1p, which is typically made of InP, has a conductivity of, for example, about 2E18 at / cm 3 doped with zinc, has a high absorption coefficient (especially in the range of wavelengths useful in the field of optical communication), in particular of the order of -40 to -70 dB / cm. This coefficient is, for example, 2E18 / cm 3 This should be compared with N-type semiconductor layers based on InP doped with sulfur at about -15 dB / cm, and silicon at about -2 dB / cm.

[0015] Thus, the architecture of the photonic device shown in FIG. 1 is not optimal, on the one hand because the generated radiation is highly absorbed in the P-type semiconductor layer 1p, and on the other hand because its manufacture is difficult if the aim is to provide a flat surface.

[0016] To solve this, the aforementioned non-patent document proposes providing an active region QW and a P-type semiconductor layer 1p that spans the lateral dimension along the Y direction, allowing to spread the optical mode generated in this direction and also to flatten the optical mode in the Z direction perpendicular to the plane defined by the waveguide 2. In order to allow the current injection and the electrical pumping to overlap with the optical mode, this solution requires the presence of a lateral blocking layer arranged between the active region QW and the P-type semiconductor layer 1p. The current injection is therefore in the center of the structure, since it is blocked laterally by the blocking layer. In this way, the optical emission has a mode that is not spread in the metal forming the metal pad. The disadvantage of this method is the need for epitaxy to form the P-type semiconductor layer that covers the lateral blocking layer. This step is carried out at very high temperatures, which generates stresses in the active region QW that cause its degradation.

[0017] The present invention aims to provide an alternative solution to the prior art to ameliorate the problems caused by the presence of a mandatory thick P-type layer. [Prior art documents] [Patent documents]

[0018] [Patent Document 1] U.S. Patent No. 8,110,823 [Non-patent literature]

[0019] [Non-Patent Document 1] Seifried et al., “Monolithically Integrated CMOS-Compatible III-V on Silicon Lasers,” in IEEE Journal of Selected Topics in Quantum Electronics, vol. 24, no. 6, pp. 1-9, Nov.-Dec. 2018, Art no. 8200709 Summary of the Invention

[0020] Aiming to achieve one of these objectives, the subject of the present invention is a photonic device for providing optical radiation comprising an optical mode in a waveguide, comprising: a waveguide extending in a plane along a primary direction of propagation of the optical mode; an N-type semiconductor layer disposed to cover the waveguide; an active region formed of deposited layers of III-V materials, the active region being disposed along a portion of the waveguide and in contact with a portion of the N-type layer, the other portion of the N-type layer being referred to as a free portion and not in contact with the active region; a plurality of P-type semiconductor pillars disposed on and contacting the active region, the P-type semiconductor material having a first optical index, the P-type pillars being separated from one another by an encapsulant material having a second optical index lower than the first; at least one first metal pad in ohmic contact with the free portion of the N-type layer and at least one second metal pad in ohmic contact with the P-type pillar; Equipped with.

[0021] By creating a P-type layer in the form of several pillars laterally separated from each other by an encapsulation material with a low optical index, it is possible to shape the optical mode generated by the photonic device and to prevent (or limit) the spreading of the optical mode in the Z direction perpendicular to the plane in which the waveguide lies, overlapping a strongly absorbing metal pad. At the same time, it is possible to limit the thickness of the P-type pillars compared to the thickness of the P-type layers of the prior art, without the need for a blocking layer. This increases the efficiency of the photonic device.

[0022] According to other advantageous, non-limiting features of the present invention, taken alone or in any technically feasible combination, the waveguide is configured to form an optical feedback structure; The optical feedback structure is a Bragg grating having a lateral or vertical roughness. a photonic device further comprising an assembly layer of a dielectric material disposed between the N-type layer and the waveguide; a photonic device including at least three P-type pillars; At least one P-type pillar does not make ohmic contact with the second metal pad; and the encapsulation material comprises air, silicon dioxide, silicon nitride, or aluminum oxide; the P-pillars having different widths or being laterally separated from one another by different distances; the P-type pillar is disposed along a lateral central zone of the active region; the P-pillar having a height of less than 1 micron; the active area having a width between 0.5 and 5 microns; a photonic device further comprising two first metal pads disposed on either side of the active area; The active area is implemented by including a damaged zone disposed between two P-type pillars. [Brief description of the drawings]

[0023] Further features and advantages of the present invention will become apparent from the following detailed description of the invention, which proceeds with reference to the accompanying drawings. [Figure 1] FIG. 1 illustrates a prior art photonic device. [Diagram 2] FIG. 1 shows a first embodiment of a photonic device DP according to the invention. [Figure 3a] 1 shows three simulated architectures of a photonic device that highlight the advantages of the photonic device according to the present invention; [Figure 3b] 1 shows three simulated architectures of a photonic device that highlight the advantages of the photonic device according to the present invention; [Figure 3c] 1 shows three simulated architectures of a photonic device that highlight the advantages of the photonic device according to the present invention; [Figure 4] FIG. 13 illustrates the absorption losses at the metal pads of three simulated architectures as a function of the thickness of the P-type semiconductor material. [Diagram 5] FIG. 1 shows a summary of the results of a simulation carried out within the framework of the preparation of the present invention. [Figure 6] FIG. 2 illustrates another embodiment of a photonic device according to the present invention. [Figure 7] FIG. 2 illustrates another embodiment of a photonic device according to the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0024] To simplify the following description, the same reference numbers are used for identical elements or elements that perform the same function in the prior art or in different embodiments of the described photonic devices.

[0025] FIG. 2 shows a first embodiment of a photonic device DP according to the present invention.

[0026] In the substrate S, which may be of silicon-on-insulator type, there is a silicon waveguide 2 arranged in a thin layer DL of this substrate S. The waveguide 2 extends longitudinally along a main direction (X direction in FIG. 2) to transmit light to other components of the photonic device 1, for example a modulator. In the example of FIG. 2, the waveguide 2 is composed of a ribbon 2a. Its maximum width (along the Y direction, transverse to the main direction) can be between 2 and 500 microns. The illustrated waveguide is also composed of a rib 2b in the transverse center of the ribbon 2a, which extends in the main direction from one end of this ribbon 2a to the other. The rib 2b can have a width between 0.05 and 20 microns. The waveguide has a thickness (along the Z direction) of between 5 nm and 5 microns.

[0027] In this example, the combination of ribbons 2a and ribs 2b constitutes a waveguide 2 with contoured edges, a structure known per se, forming an optical feedback structure. As can be seen from the top view in FIG. 2, the waveguide 2 here is constructed in the form of a Bragg grating with transverse concave-convex shape, but the grating may also have a longitudinal concave-convex shape, and the optical feedback structure is integrated into the optical amplification medium of the photonic device DP.

[0028] The waveguide 2 is embedded in a dielectric material, typically silicon dioxide, and the optical confinement within the waveguide 2 is obtained by the difference in optical index between the silicon of the waveguide 2 and the dielectric material, which has a lower optical index than silicon. The assembly consisting of the dielectric and the waveguide 2 forms a thin layer DL of the substrate S.

[0029] Naturally, the invention is not limited to the shape of the waveguide 2 shown in FIG. 2, which may have a contour different from that exemplified here. It is also not necessary that the substrate S is of the silicon-on-insulator type, although the use of such a substrate greatly facilitates the manufacture of the waveguide 2. The waveguide 2 also does not have to be made of silicon. For example, the waveguide 2 can be formed from silicon nitride deposited on a silicon substrate having a surface oxide layer, the nitride waveguide being considered as being encapsulated by the deposited silicon oxide layer. It is therefore not necessary to completely encapsulate the waveguide by a dielectric material, and the thin layer DL can have a surface topology that leads to the formation of cavities in the structure when the photonic device DP is replenished, as for example in US Pat. No. 5,399,363.

[0030] Continuing with the embodiment of FIG. 2, the waveguide 2 is covered by an emitting structure consisting of an active region QW sandwiched between layers of N-type semiconductor material 1n and P-type semiconductor material 1p.

[0031] More specifically, the photonic device of Fig. 2 comprises an N-type semiconductor layer 1n arranged to overlap the waveguide 2, i.e. to extend laterally on either side of this waveguide over at least a portion of its length. It typically has a width (in the Y direction) of 20 to 200 microns. The N-type semiconductor layer may be made of InP doped with sulphur and may generally have a thickness between 50 and 500 nm if the waveguide 2 is made of silicon. This N-type semiconductor layer 1n may be in direct contact with the waveguide 2, more generally with a thin layer DL, or, as in Fig. 2, an assembly layer BL may be provided between the thin layer DL and the N-type semiconductor layer 1n. This assembly layer is preferably thin, with a thickness between a few nanometers and 150 nm.

[0032] For simplicity, in the remainder of this specification, an N-type semiconductor layer will be referred to by the expression "N-type layer," with it being understood that this layer is made of a semiconductor material.

[0033] On the N-type layer 1n there is an active region QW formed from a deposition of III-V materials. The active region QW is arranged in contact with only a portion of the N-type layer 1n along a longitudinal portion 2h of the waveguide 2, called the hybrid portion. Thus, the so-called "free" portions of the N-type layer 1n are not in contact with the active layer QW, which are now arranged laterally on either side of the active region QW and can be used to form ohmic contacts.

[0034] The active region QW is generally of a thickness of between 10 and 500 nm, typically of the order of 50 nm, and extends along the hybrid portion 2h of the waveguide 2 for a length typically of between 100 and 2,000 microns. It has a width between 0.5 and 30 microns, for example equal to 5 microns.

[0035] The photonic device DP according to the present specification also comprises at least one first metal pad in contact with the free part of the N-type layer 1n. In the illustrated example, two vias 3n, 3'n filled with a metal material, for example tungsten, are arranged on either side of the active layer QW. These metal vias ensure the electrical connection between the N-type layer 1n and the interconnection tracks IC of the photonic device DP (not shown in the top view of FIG. 2 to maintain the readability of this view).

[0036] Finally, the photonic device DP of Fig. 2 comprises a number of pillars of P-type semiconductor material 1p arranged on and in contact with the active region QW. These pillars will be denoted in the remainder of the specification by the expression "P-type pillars", and will be understood to be made of semiconductor material in any case. The figure thus shows a first P-type pillar 1p made of P-doped InP arranged along a first side of the active region QW, and a second P-type pillar 1'p arranged along the other side of this region. The lower surface of these pillars 1p, 1'p is in contact with the active region QW, and the upper surface is in ohmic contact with second metal pads 3p, 3'p, respectively. These P-type pillars 1p, 1'p are separated from each other by an encapsulation material with a lower optical index than the semiconductor material of which the pillars are formed. It may therefore be a silicon oxide, silicon nitride, a polymer dielectric based on benzocyclobutene, or even an air-based dielectric material. In any case, the semiconductor material forming the P-type pillars has a first optical index and the encapsulation material separating the P-type pillars has a second index lower than the first index.

[0037] Such an arrangement of P-shaped pillars makes it possible to shape the optical mode generated by the photonic device and to prevent (or limit) the spreading of the optical mode in the Z direction perpendicular to the plane in which the waveguide 2 lies.

[0038] The P-type pillars 1p and 1'p extend over the entire length of the active region QW, or at least over a large part of this length. The width Wb of each pillar and the spacing e between each pillar 1p, 1'p naturally depend on the width of the active region QW and the number of pillars. The width of the pillars or the spacing between two pillars do not all have to be the same. For example, the width Wb of a pillar may be 0.5-3 microns and the spacing e between two pillars may be 0.1-2 microns depending on the width of the active region QW located above it.

[0039] As mentioned above, the N-type layer 1n, the active region QW, and the P-type pillars 1p are embedded with an encapsulation material, typically silicon oxide. First and second metal pads are disposed within this material for electrically connecting the various elements to electrical interconnection tracks IC projected into the assembly.

[0040] The optical mode M, which is generated in a hybrid manner in the active region QW and the waveguide 2 by the photonic device DP shown in Fig. 2, is shown by a dotted line on the cross section of this figure. The P-pillars 1p, 1'p shape this optical mode M so that it spreads preferentially laterally in the Y direction of Fig. 2 and prevents it from spreading in the Z direction perpendicular to the plane in which the active region QW lies. This reduces the risk of this optical mode covering the metal zones, in particular the metal pads 3p, 3'p, which are in ohmic contact with the P-pillars 1p, 1'p.

[0041] In general, the number of P-type pillars, the width Wb of these pillars and the spacing e between two P-type pillars are adapted to the width of the waveguide 2. They are selected to create a geometry in which the hybrid optical mode is preferentially propagated laterally.

[0042] The height of these pillars 1p, 1'p can therefore be reduced compared to the thickness of 1 or 2 microns or more encountered in prior art devices. This feature is particularly advantageous since the P-type semiconductor material from which these pillars are formed has a particularly high optical absorption coefficient. The absence of any blocking layer in this structure makes it easy to form these pillars, does not require an epitaxy recovery step, and provides photonic devices with a flat surface. As is evident from the results presented in the next section of this specification, P-type pillars 1p, 1'p can be formed with heights of less than 1 micron, even 500 nm, or 300 nm, while keeping the absorption losses of the metal pads to less than 0.1 dB.

[0043] In order to fully demonstrate the advantages of replacing the continuous P-type layer of the prior art with multiple P-type pillars 1p, 1'p separated by an encapsulating material with a low optical index, the Applicant has proceeded to simulate several structures DP1, DP2, DP3 of the architecture shown in Figures 3a to 3c.

[0044] The first structure DP1 in Fig. 3a corresponds to a prior art photonic device made on the basis of InP. The active layer QW spans a width of 3 microns. A continuous layer of P-type InP 1p has a width of 2 microns, centered on the active layer and aligned with the rib of the waveguide 2. This layer is covered with a 500 nm metal pad 3p, 1 micron thick. An assembly layer BL is placed between the N-type layer 1n and the waveguide 2.

[0045] The second structure DP2 of Fig. 3b is according to the invention. The active region QW has a width of 2 microns and is protruded by the rib of the waveguide 2. This second structure DP2 also has two P-pillars 1p of P-doped InP with a width Wb of 0.5 microns and spaced apart by a distance e of 0.5 microns. The P-pillars are each located 0.25 microns from the edge of the active layer QW. The two P-pillars 1p of this second structure have a height of 500 nm.

[0046] A third structure DP3 of Fig. 3c is also according to the invention and is similar to the second structure DP2, but this time it is provided with four P-type pillars made of P-doped InP, each having a width of 0.25 microns. They are spaced apart from each other by a distance e of 0.25 microns. The four P-type pillars of this third structure DP3 have a height of 250 nm.

[0047] For each of these structures, the overlap of the generated optical mode M with the various elements that compose them (waveguide 2, active region QW, P-type layers or pillars Ip) was measured by simulation. The following table summarizes the results obtained.

[0048] [Table 1]

[0049] It is observed that by introducing pillars separated by an encapsulation material into the architecture of the device, it is possible to limit the increased portion of mode M in the waveguide 2 and in the active region QW. At the same time, the range of this radiation in the P-type semiconductor material forming the continuous layer of the first structure DP1 or in the P-type pillars of the second and third structures PD2, DP3 is limited. Recall that P-doped InP has an absorption coefficient of optical radiation of the order of 50 dB / cm, if the active region QW is made from InN, it has an absorption coefficient of the order of 15 dB / cm, and silicon waveguides of the order of 2 dB / cm. It is therefore very advantageous to limit the range of radiation generated in the P-type semiconductor material, as observed when this material is arranged in the form of pillars.

[0050] In a second series of simulations, the height of the P-doped InP pillar 1p was varied for each of the three structures DP1, DP2, DP3. The absorption losses in the metal pad 3p protruding into the P-doped semiconductor material were evaluated as a function of this thickness. The graph in Figure 4 shows these results. The thickness, in nm, of the P-type semiconductor material forming the pillar 1p of the first structure DP1 or the P-type pillar 1p of the second and third structures DP2, DP3 is placed along the x-axis of this graph. The y-axis shows the absorption losses (in dB) in the metal pad 3p of these structures. It is observed in this graph that in order to reduce these losses to less than 0.1 dB in a conventional structure such as the first DP1 structure, it is necessary to provide a thickness of more than 1 micron for the P-type layer 1p, as reported in the introduction of the prior art. In the case of structures DP2 and DP3 according to the present invention, the thickness of this semiconductor material can be reduced to less than 650 nm when two P-type pillars 1p are provided as in the second structure DP2, and to less than 300 nm when four P-type pillars 1p are provided as in the third structure DP3.

[0051] Finally, FIG. 5 shows a summary of the results of simulations of the absorption losses carried out and compared for each of the three structures DP1, DP2, DP3 shown in FIGS. 3a-3b. As can be seen in this FIG. 5, each structure DP1, DP2, DP3 has a P-type semiconductor material thick enough to limit the absorption losses in the metal pad 3p, which are limited for each of these structures and are well below 1 dB / cm. In the case of the first structure DP1, where the generated optical mode is widely spread in the relatively thick P-type material, the absorption losses in this material reach 10 dB / cm. This is not the case in the second and third structures DP2, DP3, where the diffusion of the optical mode and the reduction in the thickness of the P-type semiconductor material in the pillar 1p limit the absorption losses in this material to less than 5 dB / cm.

[0052] In looking at the total losses shown in this FIG. 5, the full advantage of the present invention can be appreciated.

[0053] Many modifications can be made to the embodiment shown in FIG. 2 without departing from the scope of this specification. Thus, as already mentioned in connection with the presentation of simulation results, it may be advantageous to provide more than one P-type pillar 1p, 1′p. Any number of P-type pillars 1p, even or odd, for example 3, 4, 5 or more, may thus be provided. It may be advantageous to place the P-type pillar substantially in the center, along the Y direction, of the active region QW. This facilitates the evacuation of heat by thermal conduction within the pillar, which heat is generated predominantly in the central zone of the active region QW. This can be achieved, for example, by choosing an odd number of P-type pillars 1p, one of which is placed along this central zone. This configuration is the one shown in the example of FIG. 7, but of course other configurations are possible.

[0054] As previously mentioned, the pillars can have a variable width Wb and spacing e between them. If at least three P-type pillars 1p are provided, only a number of these pillars can be in electrical contact with at least one second metal pad, with the remaining P-type pillars 1p not being electrically connected to the interconnect track IC. By providing non-electrically connected pillars of semiconductor material embedded in an encapsulation material, the optical mode M can be confined and shaped to extend laterally.

[0055] 6 thus shows a cross-sectional view of a photonic device DP' having four P-type pillars 1p. The outer P-type pillars are not connected to the second metal pads 3p and therefore to their interconnect tracks IC. The distance e separating the central P-type pillars is greater than the distance e' separating the outer pillars from the central pillar.

[0056] In general, conventional simulation means can be used to determine the geometric parameters of the multiple P-shaped pillars, their number, and their relative positions to give an appropriate shape to the optical mode.

[0057] It is also possible to invert the waveguide 2 with respect to the arrangement of Fig. 2, as in the case of the photonic device DP' shown in Fig. 6. In such a configuration, ribbon 2a is disposed between N-type layer 1n and ribbon 2b. The assembly layer BL of the photonic device DP' of Fig. 6 is optional.

[0058] To help confine the optical mode and shape it to diffuse primarily laterally, some of the active region QWs can be damaged and made optically inactive, particularly the part of this region QW that is not protruded by the P-pillar 1p. This damage to the crystal structure of the active region QWs can be accomplished by implanting a heavy species such as silicon.

[0059] The methods of manufacturing the photonic devices DP, DP' according to the various embodiments described above are very similar to the prior art methods.

[0060] The waveguide 2 is first formed in a substrate in a completely conventional manner, by etching, deposition, oxidation, etc. This step may involve transferring the waveguide 2 from the starting substrate to the substrate S on which the photonic device is formed. Alternatively, the waveguide 2 may be formed directly in this substrate S.

[0061] A label covering the waveguide 2 is transferred onto the substrate S on which this waveguide 2 is present, the label consisting of a stack formed by an N-type semiconductor layer, an active layer and a P-type semiconductor layer. This label is intended to form the emission structure of a photonic device. The P-type semiconductor layer therefore has a thickness substantially equal to the thickness of the P-type pillar of the photonic device in question, for example less than 1 micron, or less than 500 nm. Typically, the label has a small thickness of less than 2 microns.

[0062] The label forms a slab of these stacked layers and is mounted on the substrate S overlapping the waveguide 2, with the N-type semiconductor layer being mounted on the side of the waveguide 2. There may be an assembly layer BL mounted between the waveguide 2 and the N-type semiconductor layer.

[0063] In a subsequent local etching step, this label is treated to precisely define the emission structure. For this purpose, a part of the P-type semiconductor layer is removed to form the P-type pillar 1p and to expose the active layer. A lateral part of this layer is removed to define the active region QW and to expose at least one free part of the N-type semiconductor layer. This free part can also be partially removed to form the N-type layer in its final dimensions.

[0064] The assembly is then encapsulated by depositing an encapsulating material and can be planarized, for example using a mechanical-chemical polishing step. It is noted that since the label has a relatively small thickness initially, the local etching steps are performed quickly, the surface topology after these steps is smaller, the thickness of the encapsulating material required to cover the emission structures is reduced, and the polishing steps aimed at planarizing the final surface are easier.

[0065] In a complementary step of the method for the manufacture of a photonic device, recesses are formed by etching in the encapsulation material filled with a metal material to form first and second metal pads in ohmic contact with the free portion of the N-type layer 1n and at least some of the P-type pillars 1p.

[0066] Naturally, the invention is not limited to the described embodiments, but variations are possible without departing from the scope of the invention as defined by the claims.

[0067] Thus, although P-type semiconductor pillars are shown here completely separated from each other by an encapsulation material having a lower optical index, this need not always be the case. Thus, at least some of the P-type pillars can be fixed at their feet to adjacent pillars over only a portion of their height. In either case, the P-type semiconductor layer of the label is constructed on the waveguide as detailed herein in such a way that the shape of the generated optical mode is preferentially extended laterally.

Claims

1. A photonic device (DP, PD') for providing optical radiation comprising an optical mode in a waveguide, comprising: a waveguide (2) extending in a plane along a main direction of propagation of said optical mode; An N-type semiconductor layer (1n) arranged to cover the waveguide (2); an active region (QW) formed of a deposited layer of III-V material, said active region (QW) being disposed along a portion (2h) of said waveguide and in contact with a portion of said N-type semiconductor layer (1n), another portion of said N-type semiconductor layer (1n) being called a free portion and not in contact with said active region (QW); a plurality of P-type semiconductor pillars (1p, 1'p) disposed in and contacting the active region (QW), the P-type semiconductor material having a first optical index, the P-type semiconductor pillars (1p, 1'p) being separated from one another by an encapsulation material having a second optical index lower than the first; at least one first metal pad (3n, 3'n) in ohmic contact with the free portion of the N-type semiconductor layer (1n) and at least one second metal pad (3p, 3'p) in ohmic contact with the P-type semiconductor pillar (1p, 1'p); A photonic device comprising:

2. The photonic device (DP, PD') of claim 1, wherein the waveguide (2) is configured to form an optical feedback structure.

3. The photonic device (DP, PD') according to claim 2, wherein the optical feedback structure is a laterally rough or vertically rough Bragg grating.

4. The photonic device (DP, PD') according to any one of claims 1 to 3, further comprising an assembly layer (BL) made of a dielectric material, arranged between the N-type semiconductor layer (1n) and the waveguide (2).

5. A photonic device (DP, PD') according to any one of claims 1 to 4, comprising at least three P-type semiconductor pillars (1p, 1'p).

6. The photonic device (DP, PD') according to claim 5, wherein at least one of said P-type semiconductor pillars (1p, 1'p) is not in ohmic contact with a second metal pad (3p, 3'p).

7. The photonic device (DP, PD') according to any one of claims 1 to 6, wherein the encapsulation material comprises air, silicon dioxide, silicon nitride, or aluminium oxide.

8. Photonic device (DP, PD') according to any one of claims 1 to 7, wherein said P-type pillars have different widths or are laterally separated from each other by different distances.

9. A photonic device (DP, PD') according to any one of claims 1 to 8, wherein a P-type semiconductor pillar (1p, 1'p) is arranged along the lateral centre of the active region (QW).

10. The photonic device (DP, PD') according to any one of claims 1 to 9, wherein said P-type semiconductor pillar (1p, 1'p) has a height of less than 1 micron.

11. A photonic device (DP, PD') according to any one of claims 1 to 10, wherein said active region (QW) has a width between 0.5 and 5 microns.

12. The photonic device (DP, PD') according to any one of claims 1 to 11, wherein the active region (QW) comprises a damaged zone located between two P-type semiconductor pillars (1p, 1'p).

Images