EP0000529A1 - Coupling device for an optical fibre - Google Patents

Abstract

The invention relates to a coupling device for optical fiber. Mode coupling means (14,16,12) by alternating curvatures of the fiber (18) are associated with an index adapter (3) surrounding the fiber and traversed by a set of light rays all making the same angle with the axis (2) thereof, and this adapter is provided with a conical optical surface (5) for matching a beam parallel to this set of rays. Application to telecommunications.

Description

The invention relates to a coupling device for optical fiber.

It is known that an optical fiber consists of a core with an optical index N surrounded by a sheath with an index a smaller than N and makes it possible to guide a light in the core which can, for example, be modulated for telecommunication purposes. . There is then an advantage in having a coupling device allowing, without interrupting the optical fiber, either to introduce light into the fiber, or to derive a fraction of the light which it guides in order to be able, for example, to disperse, in the vicinity of the point of diversion, information that the fiber carries further. Despite this obvious need, no truly effective device of this kind has been proposed. Such a device must in particular apply to multimode fibers. These fibers are those in the heart of which light can propagate in several distinct modes. They are the only ones whose industrial use is practically envisaged because of their relatively large diameter (0.15 mm for example with the sheath). The diameter d of the core verifies the relation dF root of (N ^{2 -} n ² ) greater than 0.7656.c, f being the frequency of the light used and c the speed of propagation of light in a vacuum.

A known coupling device is described in US Pat. No. 3,931,518 (Miller). According to this patent, for the purpose of extracting light, a coupling is performed beforehand between modes with a high propagation constant propagating in the core and modes with a low propagation constant which can only propagate in the sheath.

The change in propagation constant is obtained by inducing alternating curvatures in the fiber by clamping against a parallel network engraved in a transparent block. We know indeed (D. Marcuse: Coupled Mode Theory for Round Optical Fibers Bell. Syst. Tech. J. 52, p 817 - 1973) that when the position of the axis of the fiber, its curvature, the index or the diameter of the core fluctuate along the axis of propagation there are energy exchange between different modes corresponding to different values of the propagation constant K.

More precisely, if the defect thus introduced is sinusoldal with a spatial frequency P in radians per unit of length, the energy exchange takes place between two modes of propagation constant K1 and K2 such that, K1 - K2 = P, this exchange that can be done in both directions (WJ Stewart: "Mode Conversion due to periodic distortions of the fiber axis" Optical Fiber Communications Conference, September 16-18, 1975).

c étant la vitesse de la lumière dans le vide et f la fréquence de la lumière utilisée. La dérivation d'une fraction significative de l'énergie transportée résulte d'une part du fait que l'amplitude de la déformation induite dans la fibre est suffisamment importante et d'autre part du fait que la succession des courbures alternées induit, dans la lumière se propageant dans le coeur de la fibre, une succession de transpositions de mode en nombre suffisant pour obtenir une valeur finale de constante de propagation au-dessous de la limite inférieure de propagation dans le coeur. La lumière dont la constante de propagation a été ainsi diminuée se propage désormais dans la gaine. Elle peut être aisément extraite de celle-ci grâce à un "adaptateur d'indice", constitué par un milieu transparent d'indice optique au moins sensiblement égal à celui de la gaine et en contact optique avec la surface extérieure de celui-ci. L'indice de l'adaptateur ne doit pas descendre au-dessous de 0,8 fois celui de la gaine. Cet adaptateur est constitué, dans le brevet Miller précédemment cité, par le disque de couplage (coupling disk) 18. La lumière ayant pénétré dans l'adaptateur est transmise par celui-ci à un dispositif d'utilisation, tel qu'une diode photodétectrice.The upper KM and lower Km limits of the propagation constant K of the modes likely to subsist in the core are given by the relations:

c being the speed of light in a vacuum and f the frequency of light used. The derivation of a significant fraction of the energy transported results on the one hand from the fact that the amplitude of the deformation induced in the fiber is sufficiently large and on the other hand from the fact that the succession of alternating curvatures induces, in the light propagating in the core of the fiber, a succession of mode transpositions in sufficient number to obtain a final value of propagation constant below the lower limit of propagation in the core. The light, the propagation constant of which has thus been reduced, now propagates in the sheath. It can be easily extracted therefrom thanks to an "index adapter", constituted by a transparent medium with an optical index at least substantially equal to that of the sheath and in optical contact with the external surface of the latter. The index of the adapter must not drop below 0.8 times that of the sheath. This adapter is constituted, in the Miller patent cited above, by the coupling disk (coupling disk) 18. The light having penetrated into the adapter is transmitted by the latter to a device for use, such as a photodetector diode .

Another device for extracting light from a fiber is described in a communication by C. and WJ Stewart "Directional coupler for single multimode optical fiber" at the "second European colloquium on optical fiber transmissions", Paris, September 27-30, 1976 and published by the "Committee of the international colloquium on optical fiber transmissions, 11 rue Hamelin 75783 PARIS Cedex 16 (Cables and connections, Part 2, p 267-258).

In this device, the index adapter consists of a thin plate of transparent material, one edge of which carries an undulation constituting a network ("grating") and is applied to the fiber so as to ensure periodic deformation of that -this. The optical contact between this edge and the sheath of the fiber allows light to pass through this thin plate, in which it is directed towards a curved edge which reflects it and forms it on a detector placed in contact with another edge.

These various light extraction devices have the disadvantage of only sending a small fraction of the light circulating in the fiber to the use member (detector). This fraction is that which leaves in a half-plane starting from the axis of the fiber, or more exactly in the small angle diedre formed by two half-planes starting from the axis of the fiber and very close to each other . In the case of the device described in the Miller patent, the light exit plane is a plane passing through the axis of the fiber and perpendicular to the surface of the coupling disc 18. In the case of the device described in the article from Stewtrt, the light exit plane is that of the thin plate. There are by. in addition, interest in making a coupling device making it possible to introduce light into the fiber with good efficiency. This problem arises in particular when it is desired to introduce into the fiber a fraction, as large as possible of the light produced by a light-emitting diode (LED). It is known for this to cut the fiber and to concentrate the light coming from the diode using a lens on the cut end of the fiber. The performance of such a light introducing device is very poor because the die emits light from a non-small emitting surface, and in a solid angle which is large. If we use a lens which receives all the light emitted by the diode, and which concentrates it all on the surface of the section of the core of the fiber, most of this light makes with the axis of the fiber an angle too much great for this light to propagate through the fiber. No conventional optical system makes it possible to avoid this drawback, due to a known law of optics, sometimes called "theorem for conserving the geometric extent of a beam". This law says that no optical system can, without loss of light, reduce the product of the solid angle of divergence of a beam by the area of the section of this beam. This law is expressed in particular in the book "Principle of Optics" by M. Born and E. Wolf (3rd edition, Pergamon Press, p 120, equation 54). Given the characteristics of known light-emitting diodes, it follows that the efficiency of introducing their light into a fiber is always poor. The efficiency of introducing light into the fiber can be considerably increased if the light from a laser is used, which has a very small divergence. But other drawbacks then appear such as the price of the laser, its size, etc.

These known devices for introducing light also have the disadvantage of forcing the fiber to be cut.

The object of the present invention is to produce a coupling device for optical fiber making it possible to obtain good coupling efficiency, in particular for coupling with a light emitting diode emitting light in a large solid angle, without requiring cutting the optical fiber.

It relates to a coupling device for optical fiber characterized in that it comprises - coupling means between the interior and exterior of the fiber to ensure coupling between on the one hand the light propagating in the fiber and on the other hand a set of light rays propagating outside the fiber by surrounding it, the rays of this set passing in the vicinity of the axis of the fiber over the entire length of a segment of this axis by making all the same predetermined angle. with this axis; - and an optical surface external to the fiber and having a shape in a point of revolution around this axis, to ensure coupling between this set of -ayons and a light beam parallel to this axis or converging at a point aligned on this axis,

With the aid of the diagrammatic figures 1 to 7 attached, a description will be given below, without implied limitation, of how the invention can be implemented. The elements which correspond in several of these figures are designated therein by the same reference signs. The paths of the light rays are represented in these figures by dashed lines, with arrows showing the direction of propagation of the light. Figure 1 shows a device according to the invention seen in section through a plane passing through its axis, the coupling means between the inside and the outside of the fiber not being shown.

2 shows a view of the device of Figure 1 in section through a plane perpendicular to its axis.

FIG. 3 represents a sectional view of a device for injecting light into an optical fiber, the section plane passing through the axis of this device.

Figure 4 shows a view of the device of Figure 3 in section through a plane perpendicular to its axis.

FIG. 5 represents, in strong lines, a view of a coupling device according to the invention in section through a plane passing through its axis. In this figure fine lines represent. various lines and a hyperbola that can be traced in the section plane to allow a better understanding of the shape of the optical surface of the device.

FIG. 6 represents a view of a light extraction device according to the invention, with the use of a reflecting optical surface, this device being cut by a plane passing through its axis.

FIG. 7 represents a view of a device for injecting and extracting light according to the invention, in section through a plane passing through its axis.

In these various figures, this device has an axis 2, shown in phantom, this axis also being that of a portion of straight optical fiber and / or the middle axis of a portion of wavy fiber;

It has been indicated above that the invention uses coupling means between the inside and the outside of the fiber. In the case where the invention applies to an optical fiber consisting of a core surrounded by a sheath with a smaller optical index, these coupling means can advantageously include means for coupling modes capable of creating alternating curvatures in a portion of the fiber and thus give this portion an undulating shape oscillating around a mean axis. This makes it possible to couple modes with high propagation constants propagating in the core with modes with low propagation constants propagating in the sheath. These mode coupling means are associated with an index adapter 3 (Fig. 1 and 2) consisting of a transparent medium whose index is not significantly lower than that of the sheath. This adapter makes it possible to couple the modes propagating in the sheath with light propagating in this transparent medium and constituting said set of rays.

Said optical surface is then constituted by a surface 5 of this index adapter. In FIGS. 1 and 2, the optical fiber is represented by a single line along the axis 2. The index adapter has the shape of a solid cylinder, of revolution around the axis 2 and therefore wraps around the fiber . This solid cylinder ends at one end with a cone of revolution around the same axis and forming a point towards the outside. It is the surface of this cone which constitutes said optical surface.

The mode coupling means which have just been indicated are not shown in FIGS. 1 and 2. They are analogous to the arrangements described in the previously mentioned Miller patent. However, it has been found, according to the present invention, that it is advantageous to use a property, little exploited previously, of the light which propagates in the index adapter either by leaving the sheath, or, conversely, by being able to enter it to form the previously mentioned modes. This property is that these light rays all form substantially the same angle with the axis of the fiber. They therefore constitute a set of spokes of a particular type, comprising for example the spokes 4,6 8 and 10 of FIGS. 1 and 2. This set is different from a parallel beam because the rays are located in various planes passing through the axis. This set is on the other hand different from a beam converging on the axis because it comprises rays passing through various points of the axis.

In fact, the rays of this set do not pass exactly through the axis, but only in the vicinity of it. There is therefore a difference between the real rays and those of a set of rays exactly passing through the axis. This difference is less than the radius of the fiber and does not matter for the use of light when considering what is happening at a distance from the fiber significantly greater than its diameter. This is why, the previously mentioned optical surface preferably extends to a distance from the fiber greater than ten times its diameter, so that the rays arriving through this surface are, at least in majority, deflected by this surface. practically in the same way as if their extension exactly met axis 2.

It should also be noted that the useful property mentioned above exists even when the fiber has a wavy shape oscillating around a rectilinear mean axis. In this case, it is this mean axis which constitutes axis 2.

Finally, it should be noted that the angle (a) (fig. 1) that the rays of this set make with the axis 2 is not perfectly predetermined. It may for example undergo a variation of 1.5 'in. Plus or minus. This variation is somewhat troublesome for the implementation of the invention. However, the lower the index of the material constituting the adapter, the higher it is. This is why, this index is preferably chosen not only at least equal to that of the fiber cladding, which is roughly necessary to allow coupling. between the light as propagating in the adapter and that of the modes propagating in the sheath, but still at least equal to the index of the heart of the fiber, the latter index always being significantly higher than that of the sheath. The choice of a high index for the adapter also has the advantage that the angle (a) is increased, which avoids excessive length of the adapter. The value of this angle in radians is approximately given by the formula

(n) étant l'indice de l'adaptateur 3 et (n₁) étant l'indice du coeur de la fibre.

(n) being the index of the adapter 3 and (n ₁ ) being the index of the fiber core.

The surface shown in FIG. 1 is a diopter, that is to say that it separates two media of different indices, that of the adapter (n) and that of the air (1), and that it is crossed by light. To ensure the coupling between all the rays propagating in the adapter forming the angle (a) with the axis 2, and an external beam parallel to the axis 2, it is then necessary to give this surface the shape of a rib having a half angle at the top (b), such that

The mode coupling means previously mentioned are shown in FIGS. 3 and following, in which optical fiber is represented with a strongly gross dianeter and with exaggerated deformations, of naniere to facilitate understanding of the drawing. These means are notably studied in a conference by L. Jeunhomme and JP Poeholle "T Coupler for multimode optical fiber", (North Atlantio Treaty Organization, Advisory Group for Aerospace Research and Development, AGARD, 7 rue Ancelle 92200 NEUILLY SUR SEINE, France) . The report of this conference can be obtained from ONERA 29 avenue de la Division Leclerc, 92 CHATILLON SOUS BAGNEUX, France and from the National Technical Information Service (NTIS) 5285 Port Royal Road, SpringField, VIRGINI 22151, USA.

Par exemple, pour une fibre typique pour laquelle D² = 6 x 10^-3 et r = 0,0425 mm, P doit être voisin de 2mm. En fait, il est apparu que P pouvait être comprise entre 1,4mm et 4mm.The optical action of these means can be defined by two quantities: the pitch P of the regular suction of alternating curvatures, and the amplitude of the deformation. This amplitude is typically between 10 and 100 microns. The step of the succession must be defined more precisely. If we call (r) the radius of the fiber core, n ₁ the index of this core and n ₂ that of the cladding, we can write that the optimal value of P varies as r / 2D for a fiber with an index step, and as r / 1.41D for a fiber with a parabolic index gradient, with

For example, for a typical fiber for which D ² = 6 x 10 ^-3 and r = 0.0425 mm, P must be close to 2 mm. In fact, it appeared that P could be between 1.4mm and 4mm.

When it comes to making a device for extracting light from the fiber, the length of the corrugated portion of the latter can be between 10 and 50 mm approximately, and must be followed by a long portion d '' at least 10mm in optical contact with the adapter. When it comes to injecting light into the fiber, this wavy portion must be in optical contact with the index adapter, and its length must in principle be sufficient for any ray coming from said optical surface arrives on this wavy portion, which must extend downstream beyond the illuminated portion.

Although the mode coupling means indicated above seem to be most advantageous, other means could be used, such as, for example, a regular succession of thinning and thickening of the fiber core.

The exemplary embodiments which will now be described use an index adapter having the general shape of a full cylinder of revolution and consisting of a glass of index n = 1.4859 for a light wavelength of 6.328 angstroms . However, this adapter could equally well be made of a hard, moldable transparent material with a high index. The optical fiber has a core diameter of 85 microns, a core index n ₁ = 1.4645 and a cladding index n ₂ = 1.4565. The angle (a) is a = 12 ° 6. In cases where the optical surface is a diopter, the half angle at the top of the cone is b = 54 ⁰ 29 '. The adapter of index 3 consists of two parts 14 and 16 each having in section the shape of a semicircle so as to constitute the complete circle by bringing these two parts together with the help of a pressure means such than a screw 12 (Fig. 3). The flat faces opposite these two parts each have a regular succession of projections and hollows with a pitch P = 2mm, arranged so that the hollows of one succession are opposite the projections of the other. The portion of the fiber 18 clamped between these two parts is thus given a wavy shape. The combination of the two successions of projections and hollows is often called a "network". The optical contact between the fiber and the adapter is improved by the use of an appropriate transparent oil, of intermediate index between that of the sheath and that of the adapter. This oil is retained by capillarity. This may for example be the silicone grease sold under the name Polyectrene 128 by the firm PRODELEC. The light injection device shown in Figures 3 and 4 has an adapter having a diameter of 26mm, and a length on the axis of 75mm, including the conical part. This length is entirely occupied by the network. A converging step lens 20 (Fresnel lens) is arranged coaxially with the adapter on the side of the cone. It has a diameter of 25.4mm and a focal length of 10mm. At the focal point of this lens beyond this is disposed the emitting surface of a light-emitting diode 22 of the usual type whose radiation diagram is close to Lambert's law. Most of the light emitted by this diode is received by the lens 20, transformed into a beam parallel to the axis 2, then transformed by the conical optical surface of the adapter into a set of rays making the suitable angle with this axis and propagating towards this axis inside the adapter. It therefore meets the sheath of the fiber into which it penetrates and propagates by forming the previously mentioned modes, which are rapidly transformed into modes propagating in the core thanks to the mode coupling means constituted by the network. The length of the latter is chosen to be sufficient for this latter transformation to be almost complete. A high injection yield is thus obtained. It is not necessary to interrupt the fiber 18, but only to bend it outside the adapter, to make room for the lens 20 which must be as close as possible to the adapter.

• . appeler (e) la distance entre ce sommet S et la surface émettrice de la diode 22 et (n) l'indice optique de l'adaptateur 3.
• . couper la surface optique par un plan passant par l'axe 2, par exemple le plan de la figure 5,
• . tracer dans ce plan un axe Oy passant par le centre 0 de cette surface émettrice, cet axe Oy étant parallèle à un rayon dudit ensemble de rayons dans l'adaptateur, c'est-à-dire faisant avec l'axe 2 l'angle (a) précédemment défini, et un axe Ox perpendiculaire à Oy,
• . considérer, une moitié seulement de la section de la surface optique par ce plan, par exemple celle qui est située au-dessous de l'axe 2 sur la figure 5,
• . On peut alors préciser que cette demi section de la surface optique par ce plan est un arc d'une hyperbole définie par l'équation :
  

The use of the lens 20 can be avoided by giving the optical surface 5 a rounded point shape in retaining the value of the half angle at the top. Such an arrangement is represented in FIG. 5 which represents in strong lines elements having a real existence. The shape of the pointed optical surface is then no longer conical and can be defined as follows 1 - It is of revolution around the axis 2. 2 - It retains the same vertex S as above, which constitutes the end of The point. 3 - It retains the same value of the half angle at the top, that is to say that at the top it is tangent to the cone previously described. 4 - It is generated by the rotation of a hyperbola arc passing through the vertex S and rotating around the axis 2. 5 - To clarify the definition, we can:

• . call (e) the distance between this vertex S and the emitting surface of the diode 22 and (n) the optical index of the adapter 3.
• . cut the optical surface by a plane passing through the axis 2, for example the plane of FIG. 5,
• . trace in this plane an axis Oy passing through the center 0 of this emitting surface, this axis Oy being parallel to a radius of said set of rays in the adapter, that is to say making with the axis 2 the angle (a) previously defined, and an axis Ox perpendicular to Oy,
• . consider, only half of the section of the optical surface by this plane, for example that which is located below the axis 2 in FIG. 5,
• . We can then specify that this half section of the optical surface by this plane is an arc of a hyperbola defined by the equation:
  

We have drawn this hyperbola in FIG. 5 in fine lines as well as its axes Ox and Oy and its asymptotes. It should however be clearly understood that these elements drawn in fine lines cannot be seen on the injection device itself. They only serve to help understand _Q ow we can define the optical surface.

The light extraction device shown in FIG. 6 includes an index adapter 30 similar to the previous one. Its diameter is 30mm. It comprises a network extending from its rear face over a length of 44mm, and extending forwards by a long zone this 16mm in which the fiber 18 does not undergo deformation, the optical contact being preserved. This adapter ends forwards by a conical convex optical surface 32 with a half angle at the top equal to

This optical surface is metallized so that the light which reaches it from the fiber is reflected in the adapter by forming a beam parallel to the axis 2 which reaches the front face 34. This front face constitutes a diopter converging of a type well known to opticians, and which makes it possible to converge the beam leaving the adapter towards a receiving diode 36. This converging diopter is eccentric relative to the axis 2, so as to allow the diode to be placed 36 outside the axis 2. It is thus possible not only not to cut the fiber 18, but also not to bend it outside the adapter.

We have just described, with the aid of FIGS. 3 and 4, a light injection device using a dioptric optical surface, and with the aid of FIG. 6, a light extraction device using a reflective optical surface. It is however obvious that one could use a dioptric optical surface in a light extraction device, and a reflective optical surface in a light injection device.

It is also possible, as shown in FIG. 7, to provide injection and extraction of light with a single index adapter, comprising a single network, and, provided at each of its front ends and rear of a dioptric optical surface, an optical input surface 42 and an optical output surface 44. These optical surfaces can be conical. They must then each be associated with a converging lens, respectively 46 and 48, if it is desired to couple with elements of small dimensions, respectively a light-emitting diode 50 and a receiving diode 52.

Claims (14)

1. Coupling device for optical fiber characterized in that it comprises - coupling means (14, 16, 12) between the interior and the exterior of the fiber (18) to ensure coupling between on the one hand the light propagating in the fiber and on the other hand an assembly of light rays propagating outside the fiber surrounding it, the rays of this assembly passing in the vicinity of the axis (2) of the fiber over the entire length of a segment of this axis making all the same predetermined angle with this axis, - And an optical surface (5) external to the fiber (18) and having a shape in point of revolution around this axis, to ensure a coupling between this set of rays and a light beam parallel to this axis or converging at a point aligned with this axis. 2. Device according to claim 1, applicable to an optical fiber (18) consisting of a core surrounded by a sheath of smaller optical index, characterized in that said coupling means between the inside and the outside of fiber (18) comprise - mryens of coupling of modes (14,16,12) capable of creating alternating curvatures in a portion of the fiber and thus giving this portion a wavy shape oscillating around an average axis (2), so as to couple modes with high propagation constants propagating in the core with modes with low propagation constants propagating in the sheath, - And an index adapter (3) consisting of a transparent medium whose index n / is not significantly lower than that of the sheath, in cotic contact with the sheath, so as to couple said modes propagating in the sheath with light propagating in this transparent medium and constituting said set of rays, - a surface (5) of this index adapter constituting said optical surface. 3. Device according to claim 2 characterized in that said optical surface (5) extends to a distance from the fiber (18) greater than _. ten. times the diameter of this fiber, and said index adapter (3) has an optical index at least equal to that of the core of the fiber. 4. Device according to claim 3 characterized in that said optical surface (5) is a diopter permitting the refraction of light. 5. Device according to claim 4 characterized in that said diopter (5) has the shape of a cone of revolution around said axis (2) so as to ensure a reciprocal transformation between said set of rays and a parallel beam of light to this axis. 6. Device according to claim 5 characterized in that it further comprises a converging system (20) external to said index adapter (3) and ensuring the focusing of this parallel beam at a point (22) external to this adapter . 7. Device according to claim 3 characterized in that said diopter (5) has the shape of a surface of revolution obtained by rotating an arc of a hyperbola around said axis (2) of the fiber (18), this hyperbola being traced in a plane passing through this axis and having an axis of symmetry forming with said axis of the fiber the same angle as the rays of said set of rays, so as to ensure a reciprocal transformation between this set of rays and a beam of light converging at a point aligned on this axis of the fiber and on said axis of symmetry of the hyperbola. 8. Device according to claim 5, characterized in that the said optical surface is a reflecting surface (32) for reflecting the light towards the interior of the index adapter (30), and has the shape of a cone of revolution around said axis (2) so as to ensure reciprocal transformation between said set of rays and a beam of light parallel to this axis and inside this adapter. 9. Device according to claim 8 characterized in that the adapter further has, on the path of said inner parallel beam, a converging diopter (34) ensuring the reciprocal transformation between this inner parallel beam and a converging outer beam. 10. Device according to claim 9 characterized in that said converging diopter (34) ensures the transformation between the internal parallel beam and an external beam converging at a point (36) at a distance from the axis (2) of the fiber ( 18). 11. Device according to claim 3 characterized in that said means for creating alternating curvatures comprise two networks formed by two regular successions of hollows and projections formed respectively on two surfaces facing said index adapter (14,16), the projections of one network being opposite the hollows of the other, - and pressure means (12) for pressing the projections against two opposite sides of the optical fiber, - Means being provided to ensure optical contact between the adapter and the wavy region of the fibers. 12. Device according to claim 11, applicable to the injection of light into the optical fiber (18) and characterized in that said grating extends over the length of the area receiving light from said optical surface (42) and extends downstream beyond this zone. 13. Device according to claim 11 applicable to the extraction of light from the optical fiber (18) and characterized in that said index adapter (40) has a zone additional this outlet at least a centimeter long, and allowing optical contact with a straight portion of the fiber consecutive to said wavy portion bearing against said network, the axis (2) of this straight portion forming the extension of the mean axis (2) of this wavy portion, said optic surface (44) of revolution around this common axis being arranged so as to directly receive the light exiting from this rectilinear portion and from at least part of this wavy portion adjacent to this straight portion. 14. Device according to claim 13 characterized in that said index adapter (40) has an optical input surface (42) and an optical output surface (44) on either side of said network, so to allow the use of the mode coupling means constituted by this network both for injecting light into the fiber (18) through the optical entrance surface and for extracting light from this fiber to through the optical exit surface.

Images