DEVICE FOR COMBINING SEVERAL LIGHT BEAMS

Abstract

A device for combining several light beams, the device including several hollow input waveguides, at least one per light beam, as well as a hollow output waveguide which is the same for the different light beams, each input waveguide having an input opening to let the corresponding light beam enter, and, at the opposite, an output opening through which it emerges in the output waveguide, the output waveguide, as well as each input waveguide being laterally delimited by one or more metallic reflecting surfaces, and wherein at least a section of the output waveguide is divergent and widens in the direction of an output opening of the output waveguide.

Claims

1. A device for combining a plurality of light beams, the device comprising a plurality of hollow input waveguides, at least one for each light beam, as well as a hollow output waveguide which is the same for the plurality of light beams, each hollow input waveguide having an input opening to let a corresponding light beam of the plurality of light beams enter, and, at an opposite, an output opening, the hollow input waveguide leading to the hollow output waveguide through the output opening of the hollow input waveguide, the hollow output waveguide, and each hollow input waveguide being laterally delimited by one or more metallic reflecting surfaces, wherein at least a section of the hollow output waveguide is divergent and widens when moving towards an output opening of the hollow output waveguide.

2. The device according to claim 1, further comprising a convergent mirror on which said plurality of light beams reflect, after emerging from the hollow output waveguide.

3. The device according to claim 1, wherein said divergent section has an average opening angle (α) greater than 20 degrees.

4. The device according to claim 1, wherein the output opening of the output hollow waveguide has an area that is at least greater than twice an area of an input section of said divergent section.

5. The device according to claim 4, wherein the output opening of the output hollow waveguide has an area that is at least greater than three times an area of an input section of said divergent section.

6. The device according to claim 1, wherein the plurality of hollow input waveguides and the hollow output waveguide each extend parallel to a same plane.

7. The device according to claim 6, further comprising a deflecting mirror arranged to reflect the plurality of light beams emerging from the hollow output waveguide in an averaged direction tilted with respect to said plane.

8. The device according to claim 6, comprising a base with an upper surface, and a cover with a lower surface that is in contact with the upper surface of the base, the base comprising, on its upper surface several grooves, each hollow input waveguide being formed, at least partially, by one of said grooves.

9. The device according to claim 1, comprising a plurality of light sources for emitting said plurality of light beams, each light source being placed so that the light beam emitted by said light source enters one of the hollow input waveguides of the device, and wherein said plurality of light sources are rigidly bound to the hollow input waveguides.

10. The device according to claim 9, wherein the plurality of hollow input waveguides and the hollow output waveguide each extend parallel to a same plane, and wherein the device further comprises at least one input mirror, to reflect the light beam emitted by one of said plurality of light sources towards the input opening of the corresponding hollow input waveguide, a direction perpendicular to said at least one input mirror being tilted with respect to said plane (P).

11. The device according to claim 1, wherein at least some of the hollow input waveguides have a section whose width or diameter is between 0.1 and 1.5 mm.

12. The device according to claim 11, wherein the width or diameter is between 0.1 and 0.7 mm.

Description

BRIEF DESCRIPTION OF THE FIGURES

[0051] The figures are presented for the purposes of information and are in no way limiting.

[0052] FIG. 1 schematically shows a first embodiment of a device for combining beams implementing the instant technology, viewed in perspective.

[0053] FIG. 2 schematically shows a beam combiner of the device of FIG. 1, viewed in perspective and by transparency.

[0054] FIG. 3 is identical to FIG. 2, but shows some dimensions of the combiner.

[0055] FIG. 4 shows in more detail an output mirror of the device of FIG. 1, in perspective.

[0056] FIG. 5 schematically shows, in perspective, light rays propagating in the device of FIG. 1, these rays being obtained by digital simulation.

[0057] FIGS. 6, 7 and 8, obtained by digital simulation, show the irradiance obtained in a given zone to be illuminated, at the output of the device of FIG. 1, when a first, second and respectively third source is in operation, the other sources being switched off.

[0058] FIG. 9 is similar to FIG. 5, but in the case of an output waveguide without a divergent section.

[0059] FIGS. 10, 11 and 12 are similar respectively to FIGS. 6, 7 and 8 but, here again, in the case of an output waveguide without a divergent section

[0060] FIG. 13 schematically shows a second embodiment of a device for combining beams implementing the instant technology, viewed in perspective.

[0061] FIG. 14 schematically shows a third embodiment of a device for combining beams implementing the instant technology, viewed in perspective.

DETAILED DESCRIPTION

[0062] As already mentioned, the instant technology relates to a device 1; 2; 3 for combining together several input light beams, F1, F2 and F3, in such a way as to obtain at the output the same global light beam, for example semi-collimated (see FIGS. 1, 13 and 14). This combination is carried out using a set of hollow metal waveguides.

[0063] This set comprises input waveguides 21, 22, 23 (FIG. 2), one per input light beam F1, F2, F3. These input waveguides each emerge in the same output waveguide 40, which is the same for the different light beams F1, F2, F3, making it possible to combine them.

[0064] Remarkably, at least one section of the output waveguide 40 is divergent and widens in the direction of an output opening 45 of this guide. As explained in the part entitled “summary”, the fact that this waveguide widens as such, in the direction of propagation of the light beams, makes it possible to reduce the divergence of the global light beam F0.sub.1; F0.sub.2; F0.sub.3 that emerges therefrom.

[0065] Three embodiments of this device, which respectively bear the reference number 1, 2 and 3, are shown respectively in FIGS. 1, 13 and 14.

[0066] In the first embodiment, the device 1 comprises an output mirror 9 on which the global light beam F0.sub.1 is reflected, after emerging from the hollow waveguide system (FIGS. 1, 4 and 5). This output mirror 9 is convergent, in order to further reduce the divergence of the beam F0.sub.1. It is moreover tilted, in such a way as to deviate the global light beam F0.sub.1 towards a zone to be illuminated Zs located outside the axis.

[0067] In the second embodiment, the device 2 is devoid of such an output mirror (FIG. 13).

[0068] In the third embodiment, the device 3 is devoid of such an output mirror but on the other hand comprises one input mirror 51, 52, 53 for each input light beam F1, F2, F3 (FIG. 14). These input mirrors are tilted, in such a way as to direct these light beams towards the input waveguides 21, 22, 23, the light beams F1, F2, F3 being initially emitted with emission directions tilted with respect to the plane P wherein the waveguides extend (FIG. 14).

[0069] These three embodiments however have many points in common (in particular relating to the arrangement of the waveguides). Thus, identical or corresponding elements will as much as possible be marked with the same reference signs, from one embodiment to another, and they will not necessarily be described each time. These three embodiments are now described in more detail, one after the other.

First Embodiment

[0070] Different aspects of the device 1 according to the first embodiment can be seen in FIGS. 1 to 5. As can be seen in these figures, the device 1 comprises: [0071] several, here three separate light sources, 11, 12 and 13, [0072] a beam combiner 6, which comprises the hollow waveguides 21, 22, 23, 40 mentioned hereinabove, and which has here the form of a globally parallelepipedic and flat block, and [0073] the output mirror 9.

[0074] The light sources 11, 12, 13 are here laser sources of the QCL type (Quantum Cascade Laser) that each emit a substantially monochromatic radiation (i.e.: of a very narrow spectrum), with an average emission wavelength located between 1 and 15 microns, between 2 and 12 microns or even between 5 and 11 microns. Note moreover that the fact that the beams F1 to F3 are called “light beams” cannot be interpreted as meaning that these beams are visible beams. These three sources have respective average emission wavelengths I.sub.1, I.sub.2 and I.sub.3 that are different from one another, here. Alternatively, the different sources of the device could however have the same emission spectrum.

[0075] The light beams F1, F2, F3 that emerge from sources 11, 12, 13 are highly divergent, here. Each one of these beams has for example an opening angle higher than 20 degrees (even higher than 40 degrees). This opening angle can correspond, as here, to the full width (angular) at mid-height of the irradiance profile of the beam considered, in a first section plane comprising the axis of propagation of the beam. In a second section plane containing the axis of the beam, and perpendicular to the first section plane, each one of these beams has here an opening angle higher than 40 degrees. For the digital simulations presented hereinafter, the beams emitted, of a Gaussian profile, more precisely have an opening angle of 30 degrees and of 60 degrees, respectively in this first and this second cutting plane (these are opening angles corresponding to the type of QCLs used here). Alternatively, the sources 11, 12, 13 could however each be provided with a collimation device, such as a microlens, reducing the divergence of the light beam emitted.

[0076] Again alternatively, other type of laser sources could be used, for example sources of the ICL type (Interband Cascade Laser), other types of laser diodes (with an external cavity mounting, or not), other types of external or internal cavity lasers or tunable lasers. However, among the various laser sources that can be considered, semiconductor sources will desirably be chosen, for their compactness (one of the objectives being to obtain a compact device).

[0077] Incoherent light sources, for example incandescent sources such as silicon carbide bar sources, could also be used instead of the laser sources mentioned hereinabove.

[0078] The sources 11, 12, 13 are arranged one after the other, in a line, along an axis x. The light beams F1, F2, F3 are each emitted in a direction that is parallel to the same axis z, axis which, here, is perpendicular to the axis x.

[0079] The device 1 comprises a heat management module to remove the heat released by the sources 11, 12, 13, or even to adjust the temperature thereof. This heat management module here comprises a block 5 made from a thermally conductive material, for example metal, on which the sources are mounted (FIG. 1). This block plays both the role of a thermostat (due to its thermal inertia) and a heat dissipator. More generally, the heat management module comprises a thermostat and a heat dissipator (cooling fins, Peltier effect module, etc.). The heat management module is here the same for the different sources (which is well suited when the illumination is done successively for the different sources, or if they have the same emission spectrum). But the device could also include thermal management modules that are independent of one another, with one module for each source.

[0080] In this embodiment, the sources 11, 12, 13 are rigidly bound to the waveguides 21, 22, 23, 40, i.e. fixed, without displacement possible with respect to the latter. Indeed, the sources are permanently fixed on the block 5, which itself is rigidly bound to the beam combiner 6 (either because it is fixed to the beam combiner, for example gluing or by screwing, or because the block 5 is formed by a section of a monolithic part that is part of the combiner, this section protruding at the rear of the combiner).

[0081] As can be seen in FIG. 2, the waveguides 21, 22, 23, 40 here extend parallel to the same plane P, and are even coplanar. In other terms, each waveguide extends along an average line, on which it is centred, and these different average lines are located in this same plane P. The plane P in question is parallel to the plane containing the axes x and z mentioned hereinabove (while the axis y shown in the figures is perpendicular to the plane P).

[0082] The beam combiner 6 can comprise as here a base 7 with a flat upper surface 71 (and parallel to the plane P in question). On its upper surface, this base comprises several grooves. Each waveguide 21, 22, 23, 40 is formed, at least partially, by one of these grooves. These grooves can have any transverse profile, for example triangular, semi-circular, or, as here, rectangular (which is convenient in terms of manufacturing). The beam combiner 6 can also comprise a cover 8 with a lower surface 82, planar, that comes into contact with the upper surface 71 of the base (FIG. 4). It is this lower surface 82, metallic and reflecting, that delimits the upper portion of the waveguides (by covering the grooves in question, or, in other terms, by closing off the upper opening of each groove). This arrangement allows for a convenient manufacturing of the set of waveguides, whether via etching of the upper face of a substrate, machining by removing material, of the milling type, or by additive manufacturing or by moulding. The base 7 is a monolithic part in a single piece (i.e.:

[0083] with a continuity of material over the entire part), as well as the cover 8, which contributes to the stability and compactness of the device.

[0084] The beam combiner 6, globally parallelepipedic, is of small dimensions. Its width and its length are for example less than 20 or even 10 mm, while its thickness is for example less than 5 mm or even 3 mm.

[0085] Whether it is formed by this base and cover, or differently, the beam combiner 6 here comprises a substrate wherein the waveguides 21, 22, 23, 40 are formed. This substrate can be formed from a semiconductor material, such as silicon for example, or from glass, the surfaces that laterally delimit the guides then being covered with a metal layer, after having possibly been polished. This metal layer has a high reflectivity over the entire spectral range of use, for example greater than 95% or even 98%. This metal can for example be aluminium or gold, which each have a high reflectivity in the mid-infrared range. The substrate in question can also be made of metal, which makes it possible to overcome a step of metallisation of the surfaces in question. It can be noted that, in the mid-infrared, a surface roughness of about a few hundred nanometres is largely sufficient to obtain a quality specular reflection, and such a roughness is compatible, in a standard way, with the manufacturing techniques mentioned hereinabove.

[0086] Note that, in the case of a semiconductor substrate, all the hollow guides could be carried out by standard microelectronics methods (etching, bonding, metallisation), with this substrate also being used as a support for the light sources (the sources then being semiconductor-based sources). In this way, the alignment (and the packaging) of the different elements would be facilitated.

[0087] Regardless of the nature of the substrate wherein the waveguides 21, 22, 23, 40 are a part, each waveguide is entirely delimited laterally by one or more metallic reflecting surfaces, here. Thus, for a waveguide of circular section, for example, the guide is delimited laterally by an entirely metal cylindrical surface. For waveguides that have a straight section that is rectangular, such as those shown in FIG. 2, each guide, or, at the very least, each segment of the guide is delimited laterally by four metallic reflecting surfaces.

[0088] As already indicated, the waveguides 21, 22, 23, 40 are hollow. The interior volume of these different guides can be filled with air. It can also be filled with air devoid of water vapour, or pure nitrogen, or put into a vacuum, in order to overcome the marked absorption caused by the water vapour and carbon dioxide at some mid-infrared wavelengths.

[0089] The geometrical structure of all the waveguides 21, 22, 23, 40 is now described in more detail.

[0090] Each input waveguide 21, 22, 23 extends from its input opening 24, 26, 28 to an output opening 25, 27, 29 through which it leads into the output waveguide 40 (FIG. 2).

[0091] The input waveguide 22 is straight, and parallel to the axis z of emission of the light beams F1, F2, F3. The two input waveguides 21 and 23 located on either side of the latter are straight piecewise, with, for each one, a short input segment, parallel to the axis z, and a main segment, straight, titled with respect to the axis z in such a way as to progressively bring this guide closer to the other input waveguides. Alternatively, instead of being straight piecewise, the guides could however be curved (i.e. extend along curved average lines). The input waveguides 21, 22, 23 each have a section, here rectangular, that remains the same all along this guide, without widening or narrowing.

[0092] In this first embodiment, each input opening 24, 26, 28 is located in front of, i.e. facing one of the sources 11, 12, 13 of the device 1. The distance between the source considered, 11, 12, 13 and the corresponding input opening, 24, 26 or 28, is reduced, in such a way as to inject most of the emitted beam F1, F2, F3 into the guide despite the strong divergence of this beam. This distance is for example greater than 20 microns (for ease of fabrication), but less than half the width w or the diameter of the input opening 24, 26, 28.

[0093] In terms of lateral dimensions, the input waveguides 21, 22, 23 have: [0094] a height h (extension according to a direction perpendicular to the plane P) which is the same for these different guides, and which is moreover equal to the height h of the output waveguide 40 (FIG. 3), and [0095] the same width w.

[0096] The height h and the width w are greater than 0.1 mm. This renders the alignment of the sources and of the input waveguides relatively easy, and allows for an injection with little loss. Using waveguides that are not too narrow also makes it possible to limit the number metal reflections on the edges of the guide, which makes it possible to reduce the losses by absorption on the metal surfaces that delimit the guide. Moreover, the height h and the width w are less than 1.5 mm, and even less than 0.7 mm, here, in such a way as to limit the total size of the device 1.

[0097] Regarding the output waveguide 40, it extends from a first end 43 to its output opening 45.

[0098] The input waveguides 21, 22, 23 are connected to the output waveguide 40 at its first end 43. At its first end 43, the output waveguide has moreover a width close to three times the width w of any one of the input guides.

[0099] The output waveguide 40 is centred on an axis zo, here parallel to the axis z. The input waveguide 22 is aligned with the axis zo of the output waveguide 40. The two input waveguides 21 and 23 are connected to the output waveguide 40 by forming an angle γ with the axis zo of this guide. The junction angle y is comprised in an embodiment between 10 and 50 degrees (including when the number of input guides is different from that used here).

[0100] As can be seen in FIG. 2, the output waveguide 40 comprises a junction section, 41, followed by a divergent section 42.

[0101] The junction section 41 extends from the first end 43 of the output waveguide, to an input section 44 of the divergent section 42 (this input section is the section of the guide 40 from which it widens). The junction section 41 has a section that remains the same all along the junction section 41 (section which, here, is rectangular).

[0102] The divergent section 42 extends from its input section 44 to the output opening 45 of the output waveguide. It here has a rectangular section (rectangular profile, or, in other words, rectangular contour), that widens all along this section 42 of the output guide. This divergent section is thus delimited by four surfaces planes 46, 47, 48, 49 that together form a horn that widens when moving towards the output of the guide.

[0103] Here, the upper and lower surfaces 48, 49 of this horn are parallel with each other, and parallel to the plane P. Its two lateral surfaces 46 and 47 are on the other hand tilted with respect to one another. They form between them an opening angle α. This opening angle is here higher than 20 degrees, and even 50 degrees, and lower than 100 degrees. For the connecting angles of the input waveguides 21, 22, 23 on the output waveguide 40 comprised typically between 10 and 50 degrees, such opening angle values are well suited for obtaining as the output a global light beam F0.sub.1 wherein the averaged directions of propagation of the initial individual light beams F1, F2, F3 will have been brought closer to one another.

[0104] Here, the divergent section 42 is therefore divergent only in one plane. This plane, wherein the output waveguide 40 is angularly open, is the same as the plane containing the input waveguides, angularly separated from one another (this is the plane P mentioned hereinabove), precisely so that the angular opening of the output guide makes it possible to bring the directions of propagation of the beams closer together, injected into this guide by the input guides.

[0105] In terms of profile, the output opening 45 of the output guide has an area that is at least higher than twice, or even higher than three times the area of the input section 44 of the divergent section 42 of the output guide. This increase in surface makes it possible to substantially reduce the divergence of the global light beam F0.sub.1 that emerges from the beam combiner 6.

[0106] Several alternatives can be considered, for the output waveguide 40 that has just been presented.

[0107] Thus, the junction section 41, non-divergent, could for example be omitted. Other digital simulation results show indeed that a satisfactory combination can be obtained without this junction section (to the point of modifying the dimensions or the opening angle of the output guide), in particular in terms of divergence of the global output light beam and of homogeneity of the light power in the latter.

[0108] The junction section 41 could also be slightly divergent (but less than the divergent portion 42), to favor a propagation of the beams in the direction of the output of the device.

[0109] Instead of comprising a non-divergent section, followed by a divergent section that has a constant opening angle, the output waveguide could have an opening angle that varies progressively, continuously all along this guide (this opening angle increasing for example progressively along this guide). In this case, the output waveguide would then have an average opening angle (average along the divergent section, from its input section to the output opening) higher than 20 degrees, or even 40 degrees.

[0110] On the other hand, instead of being divergent according to a single one of the two transversal directions perpendicular to the axis zo of the guide (here according to the direction x), the divergent section 42 could be divergent according to these two transversal directions (x and y). The divergent section could then have the shape of a cone, or the shape of a horn with four faces such as presented hereinabove but then with upper and lower surfaces 48 and 49 also tilted with respect to one another (instead of being parallel).

[0111] Moreover, the device could comprise a number of input waveguides different from what was presented hereinabove (for example four or five input guides, instead of three).

[0112] Now concerning the output mirror 9, as already indicated, it is tilted in such a way as to deviate the global light beam F0.sub.1 that emerges from the output opening 45, towards a zone to be illuminated Zs located outside the axis.

[0113] Here, the output mirror 9 is tilted in such a way that the global light beam F0.sub.1 has an averaged direction of propagation z.sub.R, after reflection on the output mirror 9, that is perpendicular, or almost perpendicular to the plane P. This makes it possible to illuminate the zone Zs, located outside the axis (shifted apart).

[0114] Being able to illuminate such a zone is interesting in practice for the device 1, which is miniaturised and portable. Indeed, the beam combiner 6, globally planar and that remains one of the most sizeable elements of the device, can then be placed parallel to the surface of an element to be characterised, such as the skin of an individual or a block of material to be analysed, the output mirror then deflecting the global light beam towards this surface to be analysed (with an illumination that is globally in normal incidence).

[0115] The zone to be illuminated Zs corresponds here to a disc, the diameter of which is comprised between 1 and 10 mm, parallel to the plane P.

[0116] As already indicated, the output mirror 9 is convergent, in order to reduce the divergence of the global light beam F0.sub.1. The convergent nature of this mirror makes it possible in particular to increase the total light power received in the zone to be illuminated Zs.

[0117] By way of example, the output mirror 9 can be a parabolic mirror used outside the axis (the reflecting surface of which being formed by a portion of a paraboloid).

[0118] The output mirror could also be a mirror of the parabolic type, but with parabolic profiles and focal points that are different in a section plane parallel to the plane y,z, and in a section plane parallel to the plane x,z. The output mirror could also be a mirror of the toroidal type (circular section, but with different radii of curvature in the plane of the sources, x,z, and in the plane y,z). The output mirror could also have a concave reflecting surface with an arbitrary shape (“freeform” mirror), optimised to illuminate essentially the zone Zs, homogeneously.

[0119] In any case, the more or less convergent nature of the output mirror 9 (its focal distance, for example), and the position of this mirror are determined in such a way as to optimise the power received in the zone Zs and/or the homogeneity with which this power is distributed in the zone Zs (and this possibly beam by beam, when the sources have different emission spectra) and/or the collimated nature of the global light beam F0.sub.1.

[0120] The characteristics of the output mirror are in general chosen according to those of the divergent output waveguide 40, since the characteristics of the divergent section of this guide have a substantial influence on the properties of the global light beam that emerges from the guide. By way of example, for a highly divergent guide, the focal point of the output mirror can be chosen closer to the input section 44 than the output opening 45 of the waveguide, and inversely for a hardly divergent guide.

[0121] Values that are well suited for the curvature of the output mirror 9, for its position, as well as for the geometrical characteristics of the waveguides (opening angle, length of each section, etc . . . ) can be determined, by digital simulation, for example by plotting rays, in such a way as to optimise one or more of the criteria mentioned hereinabove.

[0122] An example of a result of such a simulation is presented in FIGS. 5 to 8.

[0123] In this example, the fixed parameters were: the height h of the waveguides (0.2 mm), the width w of the input waveguides (0.18 mm), the spacing between the sources (0.86 mm), and the total length Lo of the beam combiner (about 6 mm). The zone Zs to be illuminated has a diameter of 1.6 mm and is located 6 mm above the plane of the (according to the Y axis).

[0124] The (adjustable) free parameters comprised: the length I3 and the width Lout of the divergent section of the output waveguide, the length I2 of the junction section 41, the length according to the axis z of the input guides, I1, and the position and focal point of the output mirror 9.

[0125] The free parameters were then adjusted by carrying out digital simulations (using the optical simulation software Zemax OpticStudio), in such a way as to optimise the total power received in the zone Zs as well as the homogeneity with which this power is distributed in the zone Zs and this, beam by beam (i.e. with only the beam F1 present as input, then with only the beam F2 present as input, then with only the beam F3 present as input).

[0126] The results shown in FIGS. 5 to 8 are those obtained at the end of this optimisation procedure. For this example, this optimisation procedure led to the following values: connecting angle γ of about 17 degrees, opening angle a of about 60 degrees (the width Lout of the output opening 45 being equal to about 4.4 times the width Lin of the input section 44), I2/Lo ratio of about ⅓, radius of curvature of the output mirror of about 3 mm, the mirror having a diameter of 4 mm and being tilted 45 with respect to the plane P.

[0127] FIG. 5, which shows a portion of the light beams plotted during this simulation, shows that most of the rays emitted effectively contribute to the illumination of the zone Zs, few rays passing outside this zone. More precisely, for the three light sources 11, 12 and 13, the total light power received in the zone Zs is respectively 45% (for the source 11), 57% and 42% of the light power initially emitted by the considered source. The slight difference in energy efficiency between the channel of the source 11 and that of the source 13 (although these two channels are symmetrical with one another) can be explained by a slight stronger absorption of the metal used (here gold) at the wavelength I3 (of about 9 microns), with respect to the wavelength I1 (of about 6 microns).

[0128] FIG. 6 shows, in grayscale, the irradiance obtained at different points of the zone Zs when only the source 11 is in operation (the sources 12 and 13 being switched off). The coordinates x and z that mark the point considered inside this zone, are expressed in mm, in this figure. FIG. 7 and FIG. 8 are identical to FIG. 6, but correspond respectively to the case where only the source 12 is in operation, and to the case where only the source 13 is in operation. As can be seen in these figures, the light power is effectively distributed homogeneously inside the zone Zs to be illuminated, and this for each source. As can be seen in FIGS. 10 to 12, things would be very different without the divergent section of the output waveguide.

[0129] FIGS. 9 to 12 show the results of a digital simulation carried out in the same conditions as for FIGS. 5 to 8, but without a divergent section at the output of the output waveguide. The beam combiner 6′ then comprises the three input waveguides mentioned hereinabove, and an output waveguide with a constant section.

[0130] As hereinabove, the characteristics of the output mirror 9′ (as well as a portion of the geometrical characteristics of the guides) were adjusted in such a way as to optimise the total power received in the zone Zs as well as the homogeneity with which this power is distributed in the zone Zs, and this, beam by beam.

[0131] But as can be seen in FIG. 10, FIG. 11 and FIG. 12, even with this optimisation procedure, a hardly homogeneous illumination of the zone Zs is then obtained, in particular for the light coming from the source 11, or 13.

[0132] It is observed moreover in FIG. 9 that a substantial proportion of the light rays emitted then pass outside the zone Zs (despite the output mirror 9′, convergent) and do not participate in the illumination thereof. Moreover, for the three light sources 11, 12 and 13, the total light power received in the zone Zs is then respectively 18% (for the source 11), 23% and 16% of the light power initially emitted by the considered source.

[0133] These results illustrate the interest, for such a device for combining beams, of using an output waveguide that widens in the direction of propagation of the beams (instead of narrowing or retaining a constant section), when it is sought to produce at the output a global pseudo-collimated light beam and/or with a high and homogeneous light power density.

Second Embodiment

[0134] The device 2 of the second embodiment, shown in FIG. 13, is identical to the device 1 of the first embodiment, but it is devoid of the output mirror presented hereinabove.

[0135] Its beam combiner 62 has the same structure as the combiner 6 of the first embodiment. Some dimensional characteristics of the device 2, such as the length and the opening angle of the divergent section 42 of the output waveguide 40, can however have different values, with respect to the first embodiment, in such a way as to obtain a global light beam F0.sub.2 as homogeneous and collimated as possible despite the absence of a convergent mirror at the output.

[0136] Moreover, as hereinabove, some dimensional characteristics of the device 2 can be adjusted by digital simulation, in such a way as to optimise the total power received in a given zone to be illuminated, and/or the homogeneity with which this power is distributed in this zone, for this device without an output mirror.

Third Embodiment

[0137] The device 3 of the third embodiment is identical to the device of the first embodiment, but it is devoid of an output mirror (whether a convergent mirror, or simply a tilted flat mirror). On the other hand, it comprises one input mirror 51, 52, 53 for each input light beam F1, F2, F3 (FIG. 14), so as to direct these light beams towards the input waveguides 21, 22, 23. Indeed, in this embodiment, the sources 11, 12, 13 have emission directions that are tilted with respect to the plane P wherein the waveguides extend, and even perpendicular to this plane.

[0138] These tilted input mirrors, 51, 52, 53, make it possible to thus dispose the sources 11, 12, 13, with their emission directions out of plane. This provides additional flexibility in the overall configuration of the device 3, and can in particular facilitate the installation of the cooling and/or thermalisation system of the sources.

[0139] In the present configuration, where the sources 11, 12, 13 are rather separated from the input openings of the input waveguides, it can be interesting to use sources each provided with a collimation system (collimation lens, microlens, etc.), in order to reduce the divergence of the light beams F1, F2, F3 that emerge therefrom. This then makes it possible to retain rather small dimensions for the input openings of the guides (less than 1.5 mm, for example, to retain a compact device 3), while still injecting most of each beam F1, F2, F3 into the corresponding input guide. Alternatively or as a supplement, input mirrors 51, 52, 53 could moreover be used that are both tilted and convergent, instead of providing the sources with collimation systems.

[0140] Again alternatively, instead of comprising several separate input mirrors, one per source, the device could comprise only one input mirror, in a single piece, which is the same for the different sources.

[0141] In any case, the beam combiner 63 of this device 3 has the same structure as the combiner 6 of the first embodiment. But here too, some dimensional characteristics of the device 3, such as the length and the opening angle of the divergent section 42 of the output waveguide 40, can have different values, with respect to the first embodiment, in such a way as to obtain a global light beam F03 as homogeneous and collimated as possible, despite the absence of a convergent mirror at the output, and in light of the divergence that is possibly different of the individual light beams, F1, F2, F3, that enter the device 3. As hereinabove, some dimensional characteristics of the device 3 can be adjusted by digital simulation, in such a way as to optimise the total power received in a given zone to be illuminated, and/or the homogeneity with which this power is distributed in this zone.

[0142] Various alternatives can be made to the embodiments that have just been presented, in addition to those already mentioned. By way of example, the device could comprise both an output mirror, such as described hereinabove, and one or more input mirrors. Moreover, the sources could emit in other wavelength ranges than the one mentioned hereinabove, for example in the visible range.