GB2213954A - Optical waveguide connecting component having tapered core - Google Patents

Abstract

An optical connecting component 1 comprises an optical waveguide having a tapered optical field spot size 2 in use, the field spot size being determined by the waveguide's dopant diffusion profile. The optical connecting component 1 may comprise a length of optical fibre of conventional construction which has been subjected to a localised non-uniform heating step to produce a dopant profile which varies along the length of the fibre. Preferably, the optical fibre comprises silica with fluorine doping. The connecting components 1 may be used with a single ferrule 18 having a partial cut 19 in which is mounted an element 16 e.g. an optical filter. <IMAGE>

Description

OPTICAL CONNECTING COMPONENT The present invention relates to optical connecting components. It finds particular application in coupling optical radiation between waveguiding elements for use in optical communications.

An optical waveguiding element comprises in general a length of dielectric material having a cross sectional refractive index profile which is peaked such that optical radiation travelling along the length of material is guided to remain within the material of the waveguiding element. If none of the energy of the radiation leaks from the material of the waveguiding element, it can be described as adiabatic.

In the case of an optical fibre, the refractive index profile can show for instance a curved peak (graded index fibre) or a rectangular peak (step index fibre). The material within the peak region is generally known as the core, and the material outside the peak region is generally known as the cladding. For most applications, and particularly long distance applications such as international optical communications cables, a step index fibre is now used which has a particularly narrow core.

The diameter of the core of this type of fibre, known as monomode fibre because the optical confinement dictates a single transverse mode of propagation, is commonly of the order of 8zm. The mode spot size, the cross sectional width of an optical field propagating in the fibre as measured between the 1/e intensity levels, is then generally of the order of 9m. (Mode spot size is known to be a function of the wavelength :; the optical field and the value of about 9 m applies = wavelengths of the order of 1.3 or 1.5 m. Mode spot size can alternatively be expressed in other words, including "beam spot size", "optical field size", or "optical field spot size".Half the mode spot size may be referred to as the "mode field radius", or "field spot radius".) The refractive index profile in an optical fibre, particularly in monomode fibres mad: of silica, is generally created by the use of a droant. This may be Ge for instance, which increases the refractive index.

In the case of an optical device such as a buried heterostructure laser, the waveguiding refractive index profile is generally achieved diffewently, for instance by using an epitaxially layered struct-e. A typical beam spot size for this type of device mr be of the order of only 4 or 5 m or even only 1 or 2 m.

In optical communications, it m- be necessary to transmit optical radiation between e ements which have extremely small optical field dimens ons. For instance, it may be necessary to couple optical radiation between monomode fibres, as described above. or between a monomode fibre and an optical device such as s buried heterostructure laser. In the abserce of any supplementary equipment or arranges nt, such coupling can call for extremely accurate positio- ng of the elements concerned, with regard to their distance apart as well as alignment. Otherwise too great an @@tical loss is introduced.

It is known to spread the optical field-at the end of an optical fibre so as to reduce the effect of transverse offset in coupling to other elements. For instance, it is known to heat and stretch the end o= a fibre. This reduces the transverse dimensions oi the fibre but has the effect of spreading the optical field. However the resultant component is extremely delicate and difficult to handle.

It is also known to increase the transverse dimensions, that is to "fatten" the end of a fibre. This means that tight optical confinement is lost but again, the optical field is spread. However, such fattening is difficult to do. Further, the change in outer dimension of the fibre is difficult to accommodate accurately in fibre to fibr or fibre to device connectors which often align fibres by reference to their outermost diameters.

The object of the present invention is to provide an optical connecting component which can be used to facilitate the optical coupling of elements having relatively small optical field dimensions.

According to a first aspect of the present invention, there is provided an optical connecting component comprising an optical waveguide having a tapered optical field size but a constant external dimension.

Preferably, the connecting component is at least substanially adiabatic in operation.

The optical waveguide may comprise an optical fibre and the tapered optical field size may be achieved by means of a dopant profile which varies with length along the waveguide. In particular, the combination of dopant fluorine with silica fibres has been found suitable because in practice the diffusion constant of fluorine in silica fibres allows an optical connecting component to be manufactured in a relatively quick and simple manner.

According to a second aspect of the present invention, there is provided a method of making an optical connecting component having a tapered optical field size, which method comprises applying a non-uniform temperature profile to a sample optical waveguide whose refractive index profile is determined by dopant distribution, for a period sufficient to cause a change in distribution of the dopant or dopants in the optical waveguide.

It has not been realised in the past that such redistribution can be achieved in a practical manner, for use in coupling commonly used optical elements. For instance, Ge is widely used as a dopant in monomode silica optical fibres because although it introduces a degree of optical loss to the fibre, it is relatively cheap.

However, the conditions required to cause redistribution of Ge in a fibre are such that it would not generally be practical to do so in regular production processes. In particular, in the Ge-doped monomode fibre commonly used in long distance communications, it only occurs at very high temperatures or over long periods.

It has now been discovered, in making the present invention, that there are material and dopant systems in which the dopant distribution can be modified in a manner suitable for use in a production process, to achieve non-uniform redistribution of the dopant profile.

The non-uniform temperature profile can be applied to change the distribution of a dopant or dopants already in the optical waveguide, or alternatively a dopant could be caused to diffuse into the optical waveguide from outside.

It has further been discovered that in spite of the different material systems, an optical connecting component according to an embodiment of the present invention can be used effectively to provide Ge-doped fibres or other optical waveguides with a tapered dopant profile.

According to a third aspect of the present invention, there is provided an optical connecting component, for coupling a first optical waveguide to a second optical waveguide, comprising a third optical waveguide having a tapered optical field spot size determined by the dopant diffusion profile of the third optical waveguide.

Preferably the optical field spot size is tapered such that the connecting component is at least substantially adiabatic in use.

By using an independent optical connecting component, the optical field spot size of a waveguide can effectively be converted, without introducing unacceptable optical losses, to a different field spot size. This becomes possible even when the waveguide itself comprises a material and dopant system in which it would be impractical to change the field spot size.

A suitable material and dopant system for use in optical connecting components according to embodiments of the present invention comprises a silica fibre doped with fluorine. It has been found that significant changes in optical field spot size can be achieved by applying a non-uniform temperature profile to such a fibre for a time period of the order of an hour, or even 30 minutes or less. Not only can such changes be achieved by using maximum temperatures already commonly applied to optical fibres in other contexts, but a dopant profile can be selected which avoids the introduction of significant losses when the optical connecting component is in use.

The combination of time period and temperature profile can of course be varied and it may be found preferable to use lower temperatures but longer time periods, for instance of up to three hours, in order to achieve a more adiabatic connecting component.

Connecting components according to embodiments of the present invention are significantly less susceptible to losses introduced by longitudinal as well as transverse offset. This benefit can be exploited in particular by employing the connecting components in devices based on the insertion of an element such as a filter into a gap in a waveguide or between a pair of waveguides. Such devices will be referred to hereinafter as "gap devices". Using known arrangements, the losses introduced by increasing longitudinal offset tend to limit the size of gap which is practicable, and therefore the size of an element to be inserted. Using connecting components according to embodiments of the present invention, a gap device can be designed in which significantly larger elements can be inserted without introducing an unacceptable degree of loss.

Optical connecting components and methods of making them, according to embodiments of the present invention, will now be described by way of example only, with reference to the accompanying drawings in which: Figure la shows schematically an optical connecting component in combination with its associated optical field spot size; Figure ib shows the refractive index profile associated with the cross section indicated by the line A-A on Figure la; Figure 2 shows a temperature profile used in a method of making the connecting component of Figure la; Figure 3 shows changes in the refractive index profile of the connecting component during its manufacture, measured in two orthogonal directions across the cross section indicated by the line B-B of Figure la;; Figure 4 shows the mode spot radius of the connecting component of Figure la, measured at different points along its length; Figure 5 shows the effect of transverse offset on optical loss in coupling a pair of connecting components, compared with coupling a pair of monomode fibres; Figure 6 shows in cross section an experimental arrangement for achieving the results of Figure 5; Figure 7 shows the effect of longitudinal offset on optical loss in coupling a pair of connecting components, compared with coupling a pair of monomode fibres; Figure 8 shows in side elevation a connecting component being used effectively to provide a Ge-doped monomode fibre with a tapered dopant profile; Figure 9 shows in three quarter elevation a gap device incorporating two connecting components according to Figure la; and Figure 10 shows in cross section an alternative type of gap device to that of Figure 9.

It should be noted that, except for the graphs in Figures 3, 4, 5 and 7, the above Figures are not drawn to scale.

Referring to Figure la, an optical connecting component 1 comprises a length of optical fibre whose optical field spot profile is tapered. At a first (untreated) end 3 of the component, the optical field spot profile 2 is that of a monomode fibre while at the second (treated) end 4, the profile 2 is approximately double in diameter.

The component is based on a pure silica core, monomode, fluorine - doped fibre. Referring to Figure lb, such a fibre has a rectangular cross-sectional refractive index distribution. At 633nm, the central core region 5 has a refractive index of about 1.456 while the cladding region 6 has refractive index about 0.3 per cent lower than that of the core, at about 1.452. At 1300nm, these figures are 1.447 and 1.443 respectively. The core radius is 4.Swm and the cladding radius is 62.5m.

Such a fibre 7as an inverse dopant distribution with respect to the refractive index profile because the presence of fluc-ne decreases the refractive index of the silica fibre rather than increasing it. Hence the core of the fibre is dopLlt free, which leads to the term used above, a pure s'ica core" fibre.

In a method =: making the optical connecting component of Figure la, a portion of a length of fibre is subjected to local heating. then cleaved at the centre of the portion. The cleave gives the second, treated end 4 of the component. ,e first untreated end 3 can also be produced by a cleave, away from the heated portion. The overall length o the component is about 1 m but this is a matter of convenfence and relatively unimportant in determining the optical coupling characteristics of interest in the component which are controlled by the heated portion.

In more detail, the fibre portion is held between two clamps and heater in a silica boat about 35mm long, using a flame about Strit wide at the point where it impinges on the silica boat. sing an infra red pyrometer, the maximum temperat--e of the fibre under the action of the flame was estima ssd to be 14500C plus or minus 500C.

A heating time o- 30 mins was used. (The silica boat acts to protect the f:=re and to spread the flame).

Sufficient t Zsion is applied to the fibre by means of the clamps, during heating, to prevent the fibre deforming but not enough t= cause any tapering.

Referring to -igure 2, the temperature profile of the fibre, in the region where the flame impinges on it, has a central maximum - which is about 5mm wide, corresponding to the flame width. The temperature profile then shows lateral shoulders 3 corresponding to the presence of the silica boat and which are therefore about 35mm apart.

Beyond the shoulders 8, the temperature profile drops back to room temperature of 2O0C.

After the heating step, when the fibre has cooled, the connecting component is produced by cleaving the fibre at the centre of the heated portion corresponding to the central temperature maximum 7.

Referring to Figure 3, the heating has the effect of heavily modifying the rectangular refractive index peak corresponding to the fibre core 5. Using a refractive index profiler to assess the refractive index profile before and after heating, shown in Figures 3a and 3b respectively, it can be seen that the peak at the second end 4 of the component is smoothed. It loses its rectangular shape, and broadens out.As shown .n Figure 3 the same effect occurs in each of two orthogonal axes across the fibre. (The profiler in each case was operated at 633 nm.) Measuring the width of the refractive index peak corresponding to the fibre core 5 after heating, between the lie points of the curve, it is estimated that the fibre core diameter has increased under heating from 9zm to 14 rm. That is, by about 550/o.

The size of the fibre core is not however a direct indication of the optical field spot size. In order to assess the latter, a sample component, made in the manner described above, was subjected to transverse offset versus loss measurements. (A method of making such measurements is described below with reference to Figure 6.) It was found that the optical field spot size had approximately doubled, from about 4.75m to about 9zm in radius.

In order to assess the spot profile along the length of the ted region, the sample was glued into a glass ferrule and then polished before each of a series of measurements so as to expose, each time, an end corresponding to a different point along the sample with respect to the centre of the temperature profile.

In this series of transverse offset versus loss measurements, the same receive component was used for each measurement. This means of course that the mode spot sizes of the two components became progressively more different. However, this was taken into account in plotting the profile of the optical field spot radius, using a formula described by D. Marcuse in "Loss analysis of single-mode fibre splices", Bell Systems Technical Journal No. 56 p 703 (1977).

Referring to Figure 4, the profile of the optical field spot radius 9 was found to be enlarged from a radius of about 435 rm, equivalent to that of the untreated fibre, to a maximum radius of about 8 Rm at the centre of the heated region. The tapered portion 9, extending between those two values, was about 1 cm long.

(It should be noted that the maximum radius of about 8 ssm might be a low rather than a high estimate. This is because the first measurement was taken after the sample was cleaved at the centre of the temperature profile, but then polished away from the centre slightly.) It was noted that the optical field spot radius profile corresponded well to the temperature profile used, as shown in Figure 2.

It will be realised that variations can be made in the way that a particular optical field spot radius profile is achieved. For instance, instead of using a fixed, matching temperature profile, a narrower heat source could be used and the temperature profile achieved by moving the sample relative to the heat source in a preselected manner.

Regarding the optical coupling characteristics of the connecting component described above with reference to Figure 1, the effect of offset, both transverse and longitudinal, on optical loss was investigated.

Referring to Figure 6, a pair of connecting components 1 were aligned, firstly with reference to their outer diameters, then more accurately by maximising optical radiation transmitted from one to the other via the cores 5. One of the components 1 was then moved in a transverse direction with respect to the other and the optical loss introduced at the coupling between the components 1 was measured at a series of different transverse offset positions. The resolution of the transverse offset position measurements was 0.1m.

A 1300nm stabilised LED source was used where necessary to launch light into the components.

The experiment was then repeated using a pair of untreated monomode fibres.

Referring to Figure 5, the relative optical loss curves for transverse offset between a pair of connecting components 10, and between a pair of untreated fibres 11, were plotted. It could be seen that transverse offset had a significantly smaller effect on coupling between the connecting components.

The drop 12 in optical power at the limit of the mode field radius, 4.3dB, is included on Figure 5 for comparison.

Referring to Figure 7, the experiments described above were repeated but with regard to longitudinal instead of transverse offset. In this case the connecting component curve 13, showing the relative losses between connecting components separated by varying distances, indicated an approximately twofold improvement with respect to the untreated fibre curve 14. Hence at a longitudinal offset of 400 zm, the relative loss between connecting components was about 4.4 dB but between untreated, monomode fibres it was over 10.5 dB. These values at 250cam were 2.5dB and 7.5dB respectively.

The loss versus offset measurements shown in Figures 5 and 7 are normalised. That is, they show losses with respect to the fully aligned position of a pair of connecting components or fibres. The additional absolute loss introduced by coupling two connecting components as described above, compared with coupling two untreated fibres, was found to be about 0.5 dB. This indicates that the components were not perfectly adiabatic in operation.

To achieve a more adiabatic component, it would in general be necessary to control the heating process in a different way. For instance an improvement in adiabaticity can be achieved by heating the fibre so as to produce a smoother, more symmetric temperature profile during dopant diffusion.

Referring to Figure 8, a connecting component 1 according to an embodiment of the present invention can be used effectively to provide a Ge-doped fibre 18 with a tapered dopant profile. Using a fusion splice, a connecting component 1 about 1m in length is connected to the end of the Ge-doped fibre 18, as a tail.It has been found that the losses introduced by such a splice show a mean of only 0.08dB with a standard deviation of 0.07dB over a sample of twenty splices, in spite of the different dopant/material systems. (The cladding outer diameter and the core diameter of both the connecting component 1 and the Ge-doped fibre 18 in each case were about 125,m and 8 or 9m respectively, giving roughly equal mode field radii.) Referring to Figure 9, a gap device incorporating connecting components 1 according to embodiments of the present invention, comprises two such components 1 mounted in ferrules 17 which are aligned by a V groove in a substrate 15. An element 16, such as an optical filter, is mounted between the aligned ends of the connecting components 1. The distance between the aligned ends is 200;m.

Referring to Figure 10, it is not necessary to use a pair of separate ferrules 17. Instead of aligning a pair of ferrules in a Y groove, a single ferrule 18 may be used, the element 16 being mounted in a partial cut 19 in the ferrule 18. This may reduce alignment and assembly problems.

As shown earlier, the loss between connecting components at a 250zm separation is only 2.5dB. In a gap device, the presence of an element 16 affects the optical path length between the components and the loss is reduced to 2.1dB, taking Fresnel loss and normalisation into account. The equivalent loss using untreated monomode fibre in a gap device would be 5.2dB. Hence the use of connecting components according to embodiments of the present invention in gap devices can lead to a 3.1dB reduction in qp-Ncal loss.

Fluorine has been found to have diffusion characteristics in silica which make it suitable for use in embodiments of the present invention. Heating fluorine-doped fibre will cause dopant ions in the cladding to diffuse into the core. This will modify the refractive index profile, and hence the fundamental mode field width, under reasonable conditions of temperature and time. The following is a theoretical proposition for analysing the diffusion process for fluorine in a silica fibre.

The diffusion process depends on the product, r, of the fluorine diffusion constant and the time of heating.

=Kt (1) where: K is the diffusion constant (m~2s 1); t is the time(s).

Specifying the required field broadening as 5O0/o(say) implies a constraint on r; for physically reasonable values of t ( < 1 hr, say), this implies a constraint on K. As K depends on temperature T, this provides a limit on T.

It is possible to derive an expression for the fluorine dopant concentration c, appropriate to zero flow of fluorine dopant through the fibre cladding-air boundary, using the cylindrical diffusion equation:

where: K is the diffusion constant; r is the radius; t is the time.

The result is:

where: a is a numerical coefficient; a J0 is the Bessel function of order 0; r is the radius; b is the cladding radius.

The refractive index profile of fluorine-doped silica is given by: 2 2 n=cn&alpha;+(1-c )nco (3) (3) where: nco is the core refractive index; ncl is the cladding refractive index.

Using a Gaussian approximation to the fundamental mods and requiring a 500/0 increase in the field width, the exponent r in equation (2) is required to satisfy the following equations:

where v is the normalised frequency.

Equation (4) was solved numerically for the required &gamma;, using twenty terms in the summation over a.

Assuming the diffusion constant varies with temperature as: log(K)=log (K0)-g (6) T where Ko and g are determined by fitting to experimental data, the required heating temperature T is given by:

where r is determined by equation (4) and T is in Kelvin.

The refractive index profiles of two separate 1m lengths of fluorine-doped fibre were assessed, again at 633nm, and found to be almost identical. The data gave a step-profile defined by:

Parameter value a b 65.0 m nc2o 2.0953 2.0839 where a is the core radius and other quantities have been defined previously.

Data on the diffusion of fluorine in silica is scarce. Two models based on a temperature dependence described by equation (5) were found with parameters given below.

The calculation of heating temperature T from equations (4)-(7) was performed for each diffusion model separately and for both 25 /o and SO 0/o field broadening. The following values of r were obtained:

% broadening &gamma; (m-2) 25 3.1 x 10-12 50 6.0 x 10-12 For a heating time of t = 1hr, the values of T, converted to Celsius for convenience, are:

Diffusion Ref. g(deg K) log (K0) % broadening Tmax (deg C) 4 1.8 x 104 -4.9 25 1535 4 1.8 x 104 -4.9 50 1585 5 9.7 x 103 -7.8 25 1065 5 9.7x103 -7.8 50 1120 On the basis of the above theoretical analysis and available fluorine diffusion data, heating at temperatures of order 1300 degrees C for times of order 1 hour would be sufficient to cause substantial (of the order of 500/0) broadening of the fundamental mode field.However, it was noted that in practice, as described above, heating at a temperature of about 1450C for only 30 minutes can produce 1O00/o broadening of the mode field. This discrepancy could be explained at least in part by the scarcity of data on the diffusion constant of fluorine in silica.

Although connecting components according to the present invention have been described as being based on waveguides comprising fluorine doped silica fibres, other material - dopant systems could be used, provided the diffusion constant of the dopant in the waveguide material was sufficiently high to allow changes in dopant distribution under reasonable conditions of temperature and time.

Also other refractive index or field spot profiles might be used but the effect of the steepness of the spot profile taper on optical losses would have to be borne in mind, as mentioned above.

Claims (17)

1. An optical connecting component comprising an optical waveguide having a tapered optical field spot size in use and a constant external diameter, with respect to the waveguiding direction of the component.

2. A component according to Claim 1 which is at least substantially adiabatic in operation.

3. A component according to either one of Claims 1 and 2 wherein the optical waveguide comprises an optical fibre.

4. A component according to Claim 3 wherein the optical fibre has a tapered dopant profile which provides the tapered optical field size.

5. A component according to Claim 4 wherein the optical fibre comprises silica.

6. A component according to either one of Claims 4 or 5 wherein the tapered dopant profile is provided by the dopant fluorine.

7. An optical fibre for use in optical communications wherein the fibre is adapted for connecting to another optical component by means of a tail connected to an end thereof, the tail comprising an optical connecting component according to any one of the preceding Claims.

8. A fibre according to Claim 7 wherein the fibre comprises a Ge-doped silica fibre.

9. A gap device comprising a component according to any one of Claims 1 to 6.

10. An optical connecting element, for coupling first and second optical waveguides, comprising a third optical waveguide having a tapered optical field spot size in use, the field spot size being determined by the dopant diffusion profile of the third optical waveguide.

11. An optical connecting element according to Claim 10 wherein the third optical waveguide comprises a length of optical fibre, the tapered optical field spot size occurring adjacent an end of the fibre such that the field spot size increases to a maximum value at or substantially at the end face of the fibre.

12. An optical connecting element according to Claim 11 wherein the length of optical fibre comprises fluorine doped fibre.

13. A method of making an optical connecting component having a tapered optical field spot size, which method comprises applying a non-uniform temperature profile to a sample optical waveguide whose refractive index profile is determined by dopant distribution, for a period sufficient to cause a change in distribution of the dopant or dopants in the optical waveguide.

14. A method according to Claim 13 wherein the period is not more than three hours.

15. A method according to Claim 14 wherein the period is not more than one hour.

16. A method according to Claim 14 wherein the period is not more than thirty minutes.

17. A method according to any one of Claims 14, 15 or 16, wherein the maximum temperature of the temperature profile is not more than 1600 C.