EP3516437A1

EP3516437A1 - Device for coupling a first waveguide to a second waveguide

Info

Publication number: EP3516437A1
Application number: EP17772048.9A
Authority: EP
Inventors: Salim Boutami; Karim HASSAN; Bayram Karakus
Original assignee: Commissariat a lEnergie Atomique CEA; Commissariat a lEnergie Atomique et aux Energies Alternatives CEA
Current assignee: Commissariat a lEnergie Atomique et aux Energies Alternatives CEA
Priority date: 2016-09-23
Filing date: 2017-09-22
Publication date: 2019-07-31
Also published as: FR3056769B1; US20190384004A1; WO2018055139A1; FR3056769A1; US10698160B2

Abstract

The invention relates to a device for coupling a first waveguide (100) to a second waveguide (200), characterized in that the core (210) of the second guide (200) comprises: an end section (211) having at least: a planar end wall (2111) turned to face, and preferably in contact with, the core (110) of the first guide (100); a flared portion (2112) of convex shape extending the end wall (2111) away from the first guide (100), the flared portion (2112) having a cross section that increases with distance from the first guide (100); a constriction; and a main section (212) extending the end section (211) away from the first guide (100) and having a substantially constant cross section.

Description

Device for coupling a first waveguide

with a second waveguide »

TECHNICAL FIELD OF THE INVENTION

The invention generally relates to the coupling of waveguides applied to micro and nanotechnologies.

The invention finds advantageous but not limitative application the field of photodiodes with the coupling between a first guide through which photons penetrate and a second guide comprising an absorption portion for generating electric charges from the absorbed photons.

STATE OF THE ART

In the field of micro and nanotechnologies, it is often necessary to couple two waveguides. The coupling between two waveguides, each comprising a core and a sheath, must promote a maximum of transmission in the heart of the waveguide crossed second.

Several solutions have been developed to ensure the coupling between two waveguides in the field of micro and nanotechnologies, that is to say for waveguides whose length is typically less than a few hundred or even a few tens of micrometers (μηι, 10 ^"6 meters).

A first solution is that the heart of the first guide, typically silicon, tapers gradually over a hundred micrometers until reaching the second guide whose input surface is greater than that of the narrowed end of the first guide.

A major disadvantage of this solution is the size imposed by the length of the narrowed part, usually referred to as "typing" in English. Another solution consists in placing a convergent lens at the outlet of the first guide in order to converge the wave in the second guide.

Such a solution is for example described in the publication of K. V. Acoleyen et al., Published in 201 1 in the journal IEE under the reference

10.1 109 / GROUP4.201 1 .6053748. In this solution the second guide has one end narrowing linearly away from the interface, thus forming a tap. The purpose of this publication is to make a fashion adapter that is shorter in size than conventional tapers.

With this solution, the end narrowed linearly and forming a taper has a length between 10 and 20μη "ΐ.

This dimension is still relatively large and leads to bulky couplings between the two guides.

Another solution is used for the realization of photodiodes with double heterojunction. This type of photodiode is illustrated in FIGS. 1 to 4.

In a double heterojunction photodiode, a first waveguide 100 is formed of a silicon core (Si) inside which the wave is intended to propagate and a sheath 120 of silica (SiO). ₂ ) encapsulating the core 1 10. The second waveguide 200 is formed of a core 210 of intrinsic germanium (Ge-i) and a sheath 220 having two zones 221, 222 of silicon doped respectively n and p and located on either side of the heart 210. It is in the heart 210 of the second guide that takes the absorption of photons and the generation of electric charges. To improve the absorptivity and therefore the performance of the photodiode it is necessary to have the best possible coupling between the first 100 and the second 200 guides. As for the solutions mentioned above, this coupling should preferably allow a mode adaptation, for the shortest distance possible, between the first 100 and the second 200 guides.

At the interfaces between the core 210 of Ge-i and the doped Si zones 221, 222 of the second guide 200, the differences in refractive indices are greater than in conventional photodiodes in which both the core and the doped zones are based on germanium. This increased contrast of refractive indices contributes to confining 310 in the heart 210 of the second guide 200 the light waves 300 from the heart 1 10 of the first guide 100.

This solution thus improves the absorption of light within the heart 210, that is to say that it improves the responsivity of photodiodes compared to standard photodiodes with germanium sheath.

However, as shown in FIG. 1 illustrating a simulation and in the schematic representation of FIG. 4, a portion 320 of the light waves that penetrate into the second guide 200 is diffused in the sheath 220 within the portions 221, 222. These The scattered light waves 320 are therefore not absorbed by the germanium core 210 and thus limit the performance of the photodiode in terms of responsivity.

Thus, there is a need to provide a solution to improve the performance of a coupling between two waveguides while having a small footprint.

This is the purpose of the present invention.

Other objects, features and advantages of the present invention will become apparent from the following description and accompanying drawings. It is understood that other benefits may be incorporated. SUMMARY OF THE INVENTION

To achieve this objective, according to one embodiment, the present invention provides a device for coupling at least two waveguides. The device comprises a first waveguide and a second waveguide, each guide comprising a heart and a sheath enveloping at least a portion of the heart, characterized in that the heart of the second guide comprises:

an end portion having at least: ^■ an end wall facing opposite, and preferably in contact with the heart of the first guide,

^■ a convexly shaped flared portion extending the end wall away from the first guide, the flared portion having a section that increases away from the first guide, the section being taken along a transverse plane (xy) perpendicular to a main direction (z) of extension of the core of the second waveguide,

^■ a narrowing whose section, parallel to said transverse plane (xy), decreases away from the first guide,

o a main portion, extending the end portion away from the first guide and having a section, taken parallel to said transverse plane (xy), substantially constant.

The coupling device according to the invention confers significant advantages in particular in terms of increasing the confinement of light waves within the core of the second waveguide while maintaining a limited space requirement.

In the case where the device forms a photodiode, this improved confinement leads to a reduction of the diffusivity outside the core formed of a photon absorption material. The invention thus makes it possible to increase the responsivity of the photodiode.

Even more unexpectedly, it turns out that for a given width of the absorption portion (the heart of the second guide), the invention makes it possible to very significantly increase the responsivity of the photodiode without unduly reducing the width of the bandwidth.

In other words, the invention makes it possible to significantly increase the responsivity product times ( ^* ) bandwidth.

This represents a considerable advantage over known solutions. Indeed in a conventional coupling device, such as that of a double heterojunction photodiode, the refractive indices of the second guide (n _S i = 3.45 and n _Ge = 4.27) are much higher than the refractive indices of the first guide (n _S i02 = 1 -45 and n _S i = 3.45). A significant part of the light then diffuses out of the photodiode despite the presence of the confinement.

To increase the absorptivity of the intrinsic Ge portion, the skilled person would have at best considered expanding the core section of the second guide. This widening would have the negative effect of reducing the bandwidth, because the charge carriers would then need more time to be extracted. For an acceptable bandwidth, with the solutions of the state of the art, the usual width of the intrinsic germanium Ge-i is Ο.θμηη (10 ^"6 meters) However, for this width, even for a double photodiode heterojunction, the absorptivity remains limited.

The invention as for it, in particular thanks to the shape of the end portion, does not have these drawbacks or at least strongly limits them.

Another object of the present invention relates to a microelectronic device comprising at least and preferably a plurality of coupling devices according to the invention.

By microelectronic device is meant any type of device made with microelectronics means. These devices include, in addition to purely electronic devices, micromechanical or electromechanical devices (MEMS, NEMS ...) as well as optical or optoelectronic devices (MOEMS ...).

BRIEF DESCRIPTION OF THE FIGURES

The objects, objects, as well as the features and advantages of the invention will become more apparent from the detailed description of an embodiment thereof which is illustrated by the following accompanying drawings in which:

FIG. 1 is a schematic sectional view of a Si-p / Ge-i / Si-n double heterojunction photodiode according to the prior art, in which a simulation of the propagation of light waves is represented schematically.

Figures 2 and 3 are views of a photodiode close to that of Figure 1, according to cross sections taken respectively at the second guide and the first guide.

Figure 4 is a schematic view of the photodiode of Figures 2 and 3 in a longitudinal section, that is to say parallel to the main direction of propagation of light waves.

Figure 5 is a longitudinal sectional view of an example of the coupling device according to the invention, this device forming, in this non-limiting example, a photodiode.

Figures 6a to 6c are longitudinal sectional views schematically illustrating and in a modal approach the coupling between the first and the second guide. FIG. 6a relates to a conventional double heterojunction photodiode such as that of FIG. 4; FIG. 6b illustrates a coupling device forming in this non-limiting embodiment for example a double heterojunction photodiode; Figure 6c is an enlarged view of Figure 6b centered on the interface between the first and second guides.

FIG. 7 illustrates the evolution of the bandwidth of a photodiode according to the state of the art and of a photodiode according to the invention as a function of the width of the intrinsic zone formed by the core of the second guide.

Figure 8a illustrates time-domain finite-difference (FDTD) simulations of responsivity of photodiodes versus parameters of the end portion of the core of the second waveguide.

Figure 8b is a table showing the absorptivity, the bandwidth is the product of absorptivity ^* bandwidth for the different photodiodes of Figure 8a.

FIGS. 9a and 9b are simulations according to the FDTD method of light scattering in the core of the second guide, for respectively a coupling device according to the prior art and coupling device according to the invention.

Figures 10a to 10d illustrate alternate embodiments to that of Figure 4.

FIG. 11, comprising FIGS. 11a to 11g, illustrates the main steps of an exemplary method of producing a coupling device according to the invention.

The drawings are given by way of examples and are not limiting of the invention. They constitute schematic representations of principle intended to facilitate the understanding of the invention and are not necessarily at the scale of practical applications. In particular, the relative thicknesses of the different layers as well as the relative dimensions of the different portions of the waveguides are not representative of reality.

DETAILED DESCRIPTION OF THE INVENTION

Before beginning a detailed review of embodiments of the invention, are set forth below optional features that may optionally be used in combination or alternatively.

According to one embodiment, the flared portion is flared in a non-linear profile. According to one embodiment, the flared portion is flared in a sinusoidal profile, preferably in a profile forming a quarter period of a sinusoidal function.

According to one embodiment, the wall delimiting the flared portion forms a curve of which all tangents have a directing coefficient which increases strictly away from the first guide. According to one embodiment, the section of the main portion, taken parallel to said transverse plane, is smaller than that of the end portion. Preferably, the main portion has a section, parallel to said transverse plane, smaller than that of the narrowing wall. The main direction is parallel to the main direction of propagation of light waves.

According to one embodiment, the narrowing extends the flared portion away from the first guide. Preferably the narrowing extends to reach the main portion.

According to an alternative embodiment, the end portion has a portion of constant section, disposed between the flared portion and the narrowing.

- According to one embodiment, the narrowing has a narrowing wall extending between the flared portion and the main portion, said narrowing wall being substantially linear and defines with a plane perpendicular to said main direction an angle of between 0 and 45 degrees, preferably between 0 and 30 degrees and preferably between 0 and 15 degrees. Advantageously, the presence of an angle α having these values allows a particularly good improvement of the coupling while facilitating a simple and reproducible embodiment of the coupling device.

According to one embodiment, the difference between the refractive indices of the core and the sheath of the first waveguide being greater than the difference between the refractive indices of the core and the sheath of the second waveguide. According to one embodiment, the first waveguide has a contrast of refractive indices greater than the refractive index contrast of the second waveguide, the index contrast of a guide being the difference between the refractive index of its core and the refractive index of its sheath.

According to one embodiment, the ratio of the length of the end portion to the total length of the core of the second waveguide is less than or equal to 0.7 and preferably less than or equal to 0.5 and preferably less than or equal to at 0.3, these lengths being taken along said main direction. The length is taken according to the main direction of extension of the second guide. These values make it possible to improve the absorptivity ^* bandwidth product when the coupling device is a photodiode.

According to one embodiment, the cores of the first and second guides are in direct contact. That is, there is no element, void or area between them.

According to one embodiment, the first and second cores form a coupling interface, said interface being plane. According to one embodiment, the interface is contained in a plane parallel to said transverse plane. In all points, the section of the main portion is smaller than that of the end portion.

According to one embodiment, the device forms a double heterojunction photodiode. According to one embodiment, the device forms an avalanche photodiode.

According to one embodiment, at least one and preferably both of the heart of the first guide and the heart of the second guide has, in section along a plane parallel to said transverse plane, an area less than Ι ΟΟμηη ² , preferably less than Ι Ομηη ² , and preferably less than Ι μηη ² . Typically the diameter, if the section of the heart is circular, or the largest side if the section of the heart is a polygon, is less than 50 μηι, preferably less than 10 μηη and preferably less than 1 μηη.

Also optionally the invention may present at least one of the following claims that can be taken in association or alternatively.

According to one embodiment, the flared portion begins at the end of the end portion. Advantageously, the flared portion is delimited by a curve corresponding to a quarter period of the sinus function, centered on 0; naturally this curve has a flat portion centered on 0.

According to one embodiment, the ratio of the maximum width W1max of the end portion to the width W1 of the end portion closest to the first waveguide, that is to say, to the interface with the first waveguide, is between 1.1 and 3, these width being taken in a plane parallel to said transverse plane. This ratio makes it possible to improve the performance of the coupling while keeping in limited space.

According to one embodiment, the maximum width W1max of the end portion is less than or equal to 50 μηη (10 ^-6 meters) and preferably less than or equal to 10 μηη and preferably less than or equal to 1 μηη. ratio improves the performance of the coupling while maintaining limited space.

According to one embodiment, the ratio of the maximum width Wlmax of the end portion to the maximum width W0 of the main portion of the second waveguide is between 1.1 and 6, these widths being taken in a perpendicular plane. to said main direction. This ratio makes it possible to improve the performance of the coupling while keeping in limited space.

According to one embodiment, the length Lr of the narrowing is less than the length L1 of the flared portion, preferably Lr <- and Lr <- and Lr <^, Lr and Le being taken along the main direction (z) of extension of the core of the second waveguide. This makes it possible to improve the mode adaptation between the first waveguide and the second waveguide while maintaining limited space requirements.

According to one embodiment, the heart of the first guide and the heart of the second guide are formed in different materials.

According to one embodiment, the core of the first guide is based on silicon, for example silicon, and the core of the second guide is based on germanium, for example germanium.

According to one embodiment, the heart of one or both guides is formed of a single layer. According to another embodiment, the heart of one or both guides is formed of a stack of layers.

According to one embodiment, the core of the second waveguide is formed of an intrinsic semiconductor material and the sheath of the second waveguide comprises at least two zones of a p-doped semiconductor material and p respectively. not.

According to one embodiment, the core of the second waveguide is in Ge intrinsic (Ge-i) and said zones are zones of silicon (Si) doped p and n.

According to one embodiment, the narrowing has a narrowing wall substantially perpendicular to said main direction.

According to one embodiment, at least one of the sheath of the first guide and the second guide comprises air, a gas or vacuum. According to one embodiment, the sheath of the first waveguide is air, a gas or vacuum.

It is specified that in the context of the present invention, the term "over", "overcomes", "covers" or "underlying" or their equivalents do not necessarily mean "in contact with". For example, the deposition of a first layer on a second layer does not necessarily mean that the two layers are in direct contact with one another, but that means that the first layer at least partially covers the second layer. being either directly in contact with it or separated from it by at least one other layer or at least one other element.

A layer or a core, based on a material A, a layer or a core comprising this material A only or this material A and possibly other materials, for example doping elements, is understood to mean. Conventionally, a doping noted p means that it is a doping by positive charges and a doping n means that it is a doping by negative charges. An example of a coupling device according to the invention will now be described with reference to FIG. 5, then its advantages will be explained with reference to FIGS. 6 to 9.

The coupling device according to the invention comprises at least a first 100 and a second 200 waveguide, preferably used so that the light enters the first guide 100, which then serves as an injection guide, to then reach the second guide 200. Each waveguide comprises a core 1 10, 210 and a sheath 120, 220. Advantageously, for each waveguide, the refractive index of the core is greater than the index refraction of its sheath to confine the light beam in the cores 1 10, 210 avoiding diffusion within the sheaths 120, 220.

Note that the skilled person knows how to calculate the refractive index of a sheath if the latter is heterogeneous on the surface of the heart.

The core 210 of the second waveguide 200 may be surrounded by a heterogeneous sheath 220, as in the nonlimiting example illustrated in FIG. 5. Nevertheless, the flared portion 212 combined with the main portion 212 does not have or that very little collimation effect in the y direction. Its effect is mainly contained along the x axis. Thus, the sheath portion to be taken into account in terms of refractive indices is the lateral sheath, formed in this nonlimiting example by zones 221, 222. In this example, it may be n-doped silicon zone and whose refractive indices are equal to that of silicon.

The invention is not limited to particular materials for the heart and the sheath. It will thus be possible for the core 10, 210 silicon (Si) or intrinsic germanium (Ge-i) or silicon nitride (SiN), and for the sheaths 120, 220 of silicon dioxide (SiO ₂ ), silicon or a gas such as air for example or vacuum.

In the example illustrated, and without being limiting, the core 1 10 of the first guide 100 is in Si, its sheath 120 is in Si0 ₂ , the core 210 of the second guide 200 is in Ge-i and its sheath 220 is formed of a zone 221 of doped Si n (Si-n) and a zone 222 of p-doped Si (Si-p). In this example, the device formed by the union of the two guides 100, 200 thus constitutes a Si-n / Ge-i / Si-p double heterojunction photodiode.

As illustrated in FIG. 5, the core 1 of the first waveguide 100 has a section taken in a xy-section which remains substantially constant along the z-axis.

The norm mark x, y, z is indicated in FIG. 5. The axis z corresponds to a main direction 400 of extension of the core 1 10, 210 of the guides 100, 200. It is parallel to this main direction 400 that the most of the luminous flux propagates in the guides 100, 200.

The heart 210 of the second waveguide 200 has a section, taken in the xy plane, which varies along the z axis. Thus, the core 210 comprises at least one end portion 21 1 having a flared portion 21 12 and a main portion 212, the latter having a section smaller than the section of the end portion 21 1. The end portion 21 21 1 preferably extends from the end 21 1 1 of the heart 210 of the second waveguide 200. According to one embodiment, the heart 210 of the second waveguide 200 is in contact with the heart 1 10 of the first waveguide 100.

Thus, according to this non-limiting embodiment, there is no intermediate element between the cores 1 10, 210 of the two guides 100, 200. The coupling interface is then formed by the end portion 21 1 the heart 210 of the second guide 200 and the end portion of the heart 1 10 of the first guide 100. Preferably, the interface between the cores 1 10, 210 of the two guides 100, 200 is flat. This facilitates the manufacture of the coupling between the two cores 1 10, 210 and promote a good transmission of light waves from one guide to another.

Advantageously, the flared portion 21 of the core 210 is convex. It therefore forms a belly that extends into the sheath in a xy plane.

Thus, preferably, the flared portion 21 12 is not concave. Preferably, it does not have a linear profile in the xy plane. In other words, the wall defining the flared portion 21 12 forms a curve of which all the tangents have a directing coefficient which increases strictly by traversing the z axis. According to a non-limiting embodiment, the wall of the flared portion 21 12 follows a sinus-type function (ignoring the contact wall 21 1 1 which is flat).

The flared portion 21 12 is prolonged by a narrowing. Preferably, the narrowing is defined by a wall 21 14 which extends, along the z axis, from the flared portion 21 12 to the main portion 212. In the nonlimiting example illustrated in FIG. 5, the shrink wall 21 is linear and forms a right angle with the z axis.

According to another embodiment, the wall 21 14 may form with a plane perpendicular to the z axis, a non-zero angle a. This embodiment will be described in detail later with reference to FIG. 10a.

This structure of the core 210 of the second guide 200 significantly improves the coupling of the guide 200 with the first guide 100. This could be explained by a modal approach to the operation of the coupling device. The flared shape of the core 210 of the second guide 200 allows the mode to be adapted between the mode of the first guide 100 and the mode of the second guide 200. This mode adaptation is effected over a very short distance thanks to the end portion 212 and its flared part 21 12.

On the contrary, the conventional steps known in the state of the art are generally very long including those to expand the size of a mode.

A typing of this type is described in document [4] G. Roelkens, P. Dumon, W. Bogaerts, D. Van Thourhout, and R. Baets, "Efficient Silicon-on-Insulator Fiber Coupler Fabricated Using 248-nm- Deep UV Lithography ", IEEE Photonics Technology Letters 17, 2613 (2005). If one had to place a taper having a silicon tip upstream (depending on the direction of propagation of light rays) of the first guide 100, this tip would require several tens of microns. The size of the device would be increased in a very detrimental way.

Similarly, planar lenses in guided optics can adapt mode sizes. They have the disadvantage of being also very long, typically of the order of several tens, or even hundreds of microns. This type of lens is for example described in the publication mentioned in the introduction by K. V. Acoleyen et al., Published in 201 1 in the journal IEE under the reference 10.1 109 / GROUP4.201 1.6053748.

In the present invention, the end portion of the core 210 of the second guide 200 is on the contrary ultra-compact. By the way, it's not a lens. In particular, its base is preferably flat. Its internal walls play rather the role of mirrors to collimate the beams towards the main portion 212 of the second guide 200.

FIG. 6a illustrates a conventional Si-n / Ge / Si-p double heterojunction photodiode. The modes of the first 100 and second 200 guides are clearly different and the structure at the interface of the two guides does not allow adaptation of these modes. We have a divergence of modes. Figure 6b, enlarged in Figure 6c at the interface between the two guides 100, 200, illustrates the mode adaptation through the flared portion 211 of the device according to the invention. The following paragraphs highlight how the sizing of the core 210 of the second guide 200 influences the performance of the coupling device, particularly in terms of reducing the diffusion in the sheath 220 (and therefore in terms of absorptivity / responsivity when it is a photodiode) and in terms of bandwidth.

In the following nonlimiting example, the wall of the flared portion 2112 evolves according to a sinusoidal law. The change in width of the flared portion 2112 as a function of the axis z substantially follows the following equation:

(1) w (z) = wi + Wg.sin-

In which

- Wi is the width of the heart 110 of the first guide 100. Preferably, Wi is the width of the end wall 2111 forming the contact with the first guide.

- W0 is the width of the main portion 212 of the second guide 200, that is to say the portion of substantially constant section;

- The is the length of the end portion 211 and more precisely of the flared portion 2111;

Wg corresponds to the maximum widening of the end portion 212 (that is to say its maximum overshooting of the width with respect to the width Wi). Wlmax is the maximum width of the end portion 211. As illustrated in FIG. 5, Wlmax is equal to the width W with the addition of the widening Wg.

The widths are taken along the x axis, the lengths are taken along the z axis, the sections are taken in planes parallel to the xy plane. In the case where the sections of the cores 110, 210 are circular, the widths then correspond to diameters.

Important design parameters of the flared end portion 211 are therefore the maximum widening W _g , and the length Le of the end portion 211 flared.

In the case of a photodiode, the harmful impact related to the presence of this flared portion 2112 is a decrease in the bandwidth compared to a coupling device without flaring portion 2112. Indeed, a certain portion of the light is absorbed in the flared portion 21 12 of greater thickness than the guide extending it. In this flared portion 21 12 the extraction of the charges is slower.

The overall bandwidth of the flared portion photodiode 21 can be calculated by integrating the local bandwidth along the propagation of the light, and taking into account the attenuation of the absorption density according to the Beer law. Lambert:

(2) BP = ( _r ^ _3F ) £ _Q e - "*. bp w (z). d

In which :

bp (w (z)) is the bandwidth corresponding to an elongated photodiode of width w (z),

- a is the absorption coefficient of the photodiode (m ^" ') (second guide

200)

- I_t is the length of the second waveguide 200.

The changes in the bandwidth of a conventional Si / Ge-i / Si type photodiode (that is to say without flared part as proposed by the invention) (curve referenced 71) and a photodiode according to the invention (curve referenced 72), as a function of their width W0 are shown in Figure 7. As shown in this figure, for these two photodiodes is observed first of all an advantageous increase in the bandwidth when the width decreases. . This is due to a decrease in the transit time of the carriers. Then, when the photodiode width decreases too much (typically when the width is less than 0.3μη-ι), the bandwidth then decreases abruptly, because even if the carrier transit time decreases favorably, the RC constant delay increases. him very strongly.

In the context of the development of the present invention, it was expected that if the responsivity could be increased by the flared portion 12 12, then the bandwidth would necessarily have been degraded too detrimentally.

Unexpectedly, it has been found that the reduction in bandwidth is much less than expected. Indeed during the development of According to the invention, it has been observed that the flared portion 21 does not need to be too wide, nor too long to be optically effective.

This is clear from Figure 7 in which the photodiode according to the invention (curve 72) has a bandwidth very close to that of a conventional photodiode (curve 71) on the domain considered WO widths. Indeed, for this nonlimiting example of a photodiode according to the invention, the latter has a bandwidth very close to that of a photodiode without flared portion for WO widths between Ο.δδμηι. The simulations below of Figures 8a and 8b show that the invention improves the responsiveness product ^* bandwidth. So,

either with responsivity equal to an elongate photodiode, the invention makes it possible to improve the bandwidth,

- Or equal bandwidth, the invention improves responsivity. In these figures, a double heterojunction photodiode, as shown in FIG. 5, is simulated using the FDTD (finite difference time domain) method at the wavelength λ = 1 .55μηΊ (α = 10μηΊ). The parameters that are varied from one simulation to another are those associated with the flared portion 21 12, Wg and Le. Wi is set to Ο.θμηη and WO is also fixed.

With these conditions, and by varying Wg and Le, we obtain a map.

By then varying WO from Ο.θμηη to 0.1 μηη, the mappings illustrated in FIG. 8a are obtained and which mentions the nominal efficiency η ₀ of a corresponding right-hand (flare-off) photodiode. Each time the absorption is recorded in the photodiode (responsivity).

The table in FIG. 8b mentions the absorptivity and the bandwidth calculated according to equation (2) above and the curve of FIG. 7.

The first thing we observe from these simulations is that the invention does provide an improvement in responsiveness compared to a conventional photodiode (without flared portion) of width W0, and nominal efficiency η ₀ .

In this simulation, the maximum improvement is obtained for a small enlargement (Wg = 0.1 μηη typically), and a small flared homeland length 21 1 1 This means that the coupling device according to the invention functions well as a typing the size of which would be very short. In addition, this also indicates that with these small end portion dimensions 21 1, the bandwidth will be only slightly affected by the presence of the flared portion 21 12, contrary to what would have been the case with a typing long dimension. The calculation of the bandwidth according to the formula (2) effectively shows a only low decrease of the bandwidth, to WO fixed.

Thus, the invention allows an improvement of the product absorptivity ^* bandwidth.

We can then apply the invention to derive the most relevant advantage for the application that we want:

- Or keep the same WO as a standard photodiode, to greatly increase the responsivity, while degrading only slightly the bandwidth;

or reduce the WO (of a value close to Wg). The same responsivity is maintained as the initial standard photodiode of WO, but the bandwidth is then greatly increased.

FIGS. 9a and 9b illustrate a FDTD simulation respectively of a standard photodiode with W0 = 0 μm, and of a photodiode according to the invention with The WO is therefore identical for the two photodiodes

This simulation clearly shows:

in FIG. 9a, that the standard photodiode diffuses (see the zones 320, 330) a part of the luminous flux outside the heart of Ge-i;

- In Figure 9b that the photodiode according to the invention has no or very little diffusion outside the heart of Ge-i.

The coupling device described above has particularly improved performance. Nevertheless, the invention covers other structures of coupling devices. Some of these variants are described with reference to Figures 10a to 10d.

Figure 10a schematically illustrates the different slopes or curves that can follow the wall shrinkage 21 14.

In the example of Figure 5, this wall 21 14 forms with a transverse plane, perpendicular to the z axis, a zero angle. Thus, this narrowing wall 21 is perpendicular to the z-axis as illustrated in the example of FIG. 5 or as illustrated by the reference 14a of FIG. 10a. The narrowing is then very abrupt, even immediate.

The angle α may nevertheless be non-zero as illustrated by the reference 21 14b of FIG. 10a. Preferably, this angle a will be positive, that is to say that the value along the z axis of the shrink wall 21 14 will be greater at the level of the main portion 212 than at the flared portion 21 12 Thus the narrowing of the flared part is gradual although fast.

Embodiments with an angle α> 0 have the advantage of facilitating the manufacture of the coupling device. Preferably 0 <oc <45 ° and preferably 0 <oc <30 ° and preferably 0 <κ <15 °.

Preferably the length L _r of the narrowing, taken along the z axis is less than Le. Preferably, the length L _r of the narrowing is less than and preferably less than and preferably less than

In the example of Figure 5, L _r = 0.

In Figures 5 and 10, the narrowing is linear. Nevertheless, it is preferable to have a nonlinear shrinkage, with curved shapes, especially at the junctions with the main part 212 and with the flared shape 21 12.

Figures 10b and 10c illustrate this type of variant with rounded edges.

In practice, because of the inaccuracies of the lithography processes, the hops are generally rounded.

Advantageously, the total length L _t of the second guide 200 is such that: L _t ≥ 2 (Le + 1 _r ).

Figure 10c illustrates various possible shapes for the wall defining the flared portion 21 12. All possible shapes are the curves shown in regular dashed lines which are below the straight line shown in irregular dashed line / dot.

FIG. 10d illustrates a variant in which the narrowing 21 14 does not directly extend the flared portion 21 12. In this variant, an intermediate portion 21 15, preferably cylindrical, extends the flared portion 21 12 to the narrowing wall 21 14.

Advantageously but not limitatively, we will then be entitled to L _t ≥ 2 (Le + li + l _r ) with l _t = length along z of the intermediate part 21 15.

The invention has been described above through a nonlimiting example of application to photodiodes with double heterojunction.

It applies to other types of photodiodes than those illustrated and for example to avalanche photodiodes. By way of nonlimiting example, the following dimensions may be adapted by the coupling devices according to the invention, in particular those illustrated in FIGS. 5, 10a to 10d, but also for the other variants covered by the claims: for a length of wave λ = 1.55μηΊ (α = 10μηΊ '):

- Wg is less than 2μη "ΐ Preferably Wg is between Ο.ΟδμηΊ and 2 μηι, preferably between Ο.ΟδμηΊ and 0.7 μηι, preferably between 0.1 μηη and 0.5 μηι, preferably of the order of 0.1 μηι ;

- The is between 0.1 μηη and 1.5 μηη, preferably between 0.2μηι and 1 .2 μηη, preferably between 0.4μηι and 1 μηι, preferably of the order of Ο.δμηι;

- Wi is equal to 0.2μηι and 1 .0 μηη, preferably between 0.4μηι and 0.8 μηι, preferably between 0.5μηι and 0.7 μηη;

W0 is comprised of 0.15 μm and 1.0 μm, preferably between 0.15 μm and 0.8 μm, preferably between 0.2 μm and 0.6 μm.

When the coupling device forms a photodiode, in particular that illustrated in FIG. 5 and described in the corresponding passages, the above dimensions make it possible to considerably improve the absorbency product ^* bandwidth. These dimensions and advantages also apply to the variants described with reference to FIGS. 10a to 10d. The invention also applies to any other type of coupling device between different materials not necessarily forming a photodiode.

Indeed, if we note nc1 and ng1 respectively, the indices of heart and sheath of the guide of entry of the light; and nc2 and ng2 respectively the core and sheath indices of the light output guide, we can then apply the invention, in particular the shape of the flared portion 21 12, whenever the difference of heart / sheath indices the output guide is smaller than the input one, that is, each time that: nc1 -ng1> nc2-ng2. This index difference between the refractive index of the core of a guide and the refractive index of the sheath of this same guide is called index contrast.

As indicated above, in the nonlimiting example illustrated in FIGS. 5, 6b, 6c and 10a to 10d, the portion of sheath to be taken into account in terms of refractive index is the lateral portion of sheath 220c. that is, extending along the x axis. Thus, in the formula above, ng2 corresponds to the refractive index of the zones 221 and 222. If these zones are Si-n and Si-p, ng2 will therefore be equal to the refractive index of the silicon, Si -n and Sip having the same index values (that of Si). The principle of the present invention works in reverse. If the previous relationship is not verified, it is sufficient to apply symmetry, that is to say that it is sufficient to return the flared portion 21 12 and place it at the end of the injection guide. An exemplary method of manufacturing a coupling device illustrated in Figure 5 or 10a to 10d will now be described in Figures 11a to 11g. Each of Figures 11 to 11 g comprises a sectional view of the second waveguide 200 (seen in a xy section) and a view from above (view parallel to the plane zx.

The first step (FIG. 11a) consists in providing a stack of layers comprising in particular:

a substrate 220, for example a dielectric material such as Si0 ₂ ;

a layer 22 surmounting the substrate 220 and intended to partially form at least the sheath for the core 210. This layer 22 is preferably, but not exclusively, a semiconductor material such as silicon.

According to one embodiment, this layer 22 has at least two doped zones located on either side of the y axis. Thus, an area located for example on the left of the y-axis is doped n and another area to the right of the y-axis is p-doped.

a layer 223, for example made of a dielectric material such as Si0 ₂ , is intended to form a hard mask _.

The steps illustrated in FIG. 11b to 11d aim at engraving the hard mask 223 to give it the shape of the core 210 of the second waveguide 200. This shape appears in FIGS. 11c and 11d.

For example, a layer of resin 224 (FIG. 11b) is opened (FIG. 11c) by any suitable means of lithography (photolithography, electron beam lithography, nano-printing, etc.) to give the aperture the desired shape for the heart 210 of the waveguide 200.

As illustrated in FIG. 11, the hard mask 223 is then etched through the mask formed by the open resin layer 224. The resin layer 224 (FIG. 11) is then removed. The layer 22 is then etched to form part of the sheath through the hard mask 223.

The cavity 540 thus formed is then filled with the material forming the core 210 of the second waveguide 200 (FIG. 11g). The heart 210 then has the desired shape, with in particular the flared end portion 21 1 and the main portion 212.

According to one embodiment, to fill the core 210, epitaxial growth is carried out from a residual layer 560 of the layer 22 left in place at the end of etching.

Thus according to this embodiment the etching of the layer 22 is only partial and leaves a thickness 560 in the bottom of the opening 540.

For example, if the material of the layer 22 is crystalline or polycrystalline silicon, the cavity can be filled by growing germanium by epitaxy.

Preferably, a flattening step is then carried out, for example by chemical mechanical polishing (CMP) so that the upper faces of the layer 22 and the core 210 are substantially included in the same plane.

Then it is possible to cover the upper faces of the layer 22 and the core 210 by a protective layer, preferably a dielectric, for example a layer of Si0 ₂ .

It is then possible to make electrical contacts in the areas doped by all techniques widely known to those skilled in the art, for example by making vias and filling these holes with an electrically conductive material.

This process has the advantage of being simple and perfectly reproducible. It thus makes it possible to manufacture the coupling device in a simple manner and at a limited cost.

The invention is not limited to the previously described embodiments and extends to all the embodiments covered by the claims.

Claims

1. Device for coupling at least two waveguides (100, 200), the device comprising a first waveguide (100) and a second waveguide (200), each waveguide (100, 200) comprising a waveguide core (1 10, 210) and a sheath (120, 220), the difference between the refractive indices of the core (1 10) and the sheath (120) of the first waveguide (100) being greater than the difference between the refractive indices of the core (210) and the sheath (220) of the second waveguide (200), characterized in that the core (210) of the second guide (200) comprises:

an end portion (21 1) having at least:

^■ an end wall (21 1 1) plane in contact with the heart (1 10) of the first guide (100),

^■ a flared portion (21 12) extending convexly form the end wall (21 1 1) away from the first guide (100), the flared portion (21 12) having a section which increases away from the first guide (100), the section being taken along a transverse plane (xy) perpendicular to a main direction (z) extending the core (210) of the second waveguide (200),

^■ a narrowing whose section, parallel to said transverse plane, decreases away from the first guide (100),

o a main portion (212), extending the end portion (21 1) away from the first guide (100) and having a section, taken parallel to said transverse plane, substantially constant.

2. Device according to the preceding claim wherein the flared portion (21 12) is flared in a non-linear profile.

3. Device according to any one of the preceding claims wherein the flared portion (21 12) is flared in a sinusoidal profile, preferably in a profile forming a quarter period of a sinusoidal function.

4. Device according to any one of the preceding claims wherein said narrowing extends the flared portion (21 12) away from the first guide (100) and to reach the main portion (21 1).

5. Device according to any one of the preceding claims wherein the narrowing has a wall (21 14) narrowing extending between the flared portion (212) and the main portion (21 1), said wall (21 14) of narrowing being substantially linear and defines with a plane perpendicular to said main direction (z) an angle α between 0 and 45 degrees, preferably between 0 and 30 degrees and preferably between 0 and 15 degrees.

6. Device according to any one of the preceding claims wherein the length Lr of the narrowing is less than the length Le of the flared portion (21 12), preferably Lr <y and Lr <y and Lr <^, Lr and Le being taken along the main direction (z) of extension of the core (210) of the second waveguide (200).

7. Device according to any one of the preceding claims wherein the ratio of the length (Le + Li + Lr) of the end portion (21 1) over the total length Lt of the heart (210) of the second guide of wave (200) is less than or equal to 0.7 and preferably less than or equal to 0.5 and preferably less than or equal to 0.3, these lengths being taken along said main direction (z).

8. Device according to any one of the preceding claims wherein the ratio of the maximum width Wlmax of the end portion (21 1) to the width of the end portion (21 1) located closest to the first guide of waves (100) is between 1 .1 and 3, these widths being taken in a plane parallel to said transverse plane, the maximum width Wlmax of the end portion (21 1) being less than or equal to 50 μηη (10 ^"6 meters) and preferably less than or equal to 10 μηη and preferably less than or equal to 1 μηη.

9. Device according to any one of the preceding claims wherein the ratio of the maximum width Wlmax of the end portion (21 1) to the maximum width W 0 of the main portion (212) of the second waveguide (200). ) is between 1.1 and 6, these widths being taken in a plane perpendicular to said main direction (z).

10. Device according to any one of the preceding claims wherein the heart (1 10) of the first guide (100) and the heart (210) of the second guide (200) are formed in different materials.

1 1. Apparatus according to any one of the preceding claims wherein the heart (1 10) of the first guide (100) is silicon-based and the core (210) of the second guide (200) is based on germanium.

12. Device according to any one of the preceding claims wherein the core (210) of the second guide (200) is formed of an intrinsic semiconductor material and the sheath (220) of the second guide (200) comprises at least two zones (221, 222) of a doped semiconductor material respectively p and n.

13. Device according to the preceding claim wherein the core (210) of the second waveguide (200) is in Ge intrinsic (Ge-i) and said zones (221, 222) are areas of silicon (Si) p-doped and N.

14. Device according to any one of the preceding claims forming a double heterojunction photodiode.

15. Device according to any one of the preceding claims forming an avalanche photodiode.

16. Apparatus according to any one of the preceding claims wherein at least one of the sheath (120, 220) of the first guide (100) and the second guide (200) comprises air, a gas or vacuum .

17. Microelectronic device comprising a plurality of coupling devices according to any one of the preceding claims.