WO2012160343A1

WO2012160343A1 - Optical fibre coupling arrangement

Info

Publication number: WO2012160343A1
Application number: PCT/GB2012/050910
Authority: WO
Inventors: Catherine RUSHWORTH; Claire Vallance; Dean JAMES
Original assignee: Isis Innovation Limited
Priority date: 2011-05-20
Filing date: 2012-04-25
Publication date: 2012-11-29
Also published as: GB201108557D0

Abstract

An optical coupling arrangement for an optical fibre (100) is made by cutting through the optical fibre, removing a wedge (9) from a first cut end (5) of the optical fibre to expose a flat face (10) that is coated with reflective material (13), and then splicing the cut ends (5, 6) of the optical fibre. This forms a notch (11) between the flat face of the first cut end and a face (6) of the second cut end of the optical fibre that is not coated by reflective material. This arrangement is straightforward to manufacture because the reflective material is applied to the cut end, rather than to a single surface of a notch in an optical fibre.

Description

Optical Fibre Coupling Arrangement

The present invention relates to the provision of a low-loss optical coupling arrangement in an optical fibre.

Optical coupling arrangements can be provided in optical fibres to provide coupling of electromagnetic (EM) radiation into or out of the optical fibre. The present invention is concerned with applications demanding a high coupling efficiency (that is the fraction of the EM radiation that is coupled into the optical fibre) and with applications demanding that the optical coupling arrangement minimises losses.

One example of an application that demands a high coupling efficiency is fibre-loop cavity ring-down spectroscopy (FLCRDS) which is a type of cavity ring-down spectroscopy (CRDS). CRDS is an extremely sensitive absorption spectroscopic technique which has been primarily used for measuring absorptions of gases with weak absorption bands or present at low concentrations. Currently, its use is also being developed for real-time detection applications in microfluidics and other analyses involving small liquid volumes. FLCRDS is a cavity ring-down method using a fibre loop as an optical cavity to enhance the optical path length, and therefore the detection sensitivity, by many orders of magnitude over single pass absorption spectroscopy techniques. For certain applications, a cavity consisting of a loop of optical fibre, in place of the two-mirror optical cavity used in gas-phase measurements, has a number of advantages, for example facilitating adaptation to the liquid phase. Achievement of high detection sensitivities requires minimisation of losses around the fibre-loop, and maximisation of EM radiation intensity transmitted to the detector. Therefore the optical coupling arrangement for coupling of EM radiation into and out of the optical fibre desirably has a high coupling efficiency and introduces low coupling losses. Existing techniques for FLCRDS typically use optical coupling arrangements such as a bend coupler for coupling into the optical fibre. This type of coupler has a relatively poor coupling efficiency and introduces losses within the loop each time the EM radiation circulating within the loop passes the coupler.

Many types of optical side-coupling arrangement for an optical fibre have been developed, primarily for use in the telecommunications field.

One such optical coupling arrangement, disclosed for example in US-5,037, 172, is a notch formed in an optical fibre with two flat faces, one but not the other being coated by reflective material. In principle, such an optical coupling arrangement can provide a high coupling efficiency into stable modes of the optical fibre. The coupling efficiency for EM radiation coupled into the optical fibre can be close to 100%, subject to the practical limitation of focussing the EM radiation on the reflective material. The amount of EM radiation coupled out of the optical fibre can be controlled by the size of the notch and incidental losses are low.

The method of making the optical coupling arrangement disclosed in US-5,037, 172 is to machine the notch by milling using a rotating, circular knife and to deposit the coating of reflective material by vacuum metallization using a razor edge mask to shield the uncoated face of the notch. However, this method of manufacture presents challenges in practice. For example, in the vacuum metallization step, it is difficult to restrict the coating to the desired face without collection of the reflective material on the other face. This reduces the coupling efficiency and creates losses for EM radiation passing around the fibre loop.

Accordingly, it would be desirable to provide an optical coupling arrangement in an optical fibre that provides desirably high coupling efficiency and desirably low losses that can be more easily manufactured in practice.

According to one aspect of the present invention, there is provided a method of making an optical side-coupling arrangement for an optical fibre, the method comprising:

cutting through the optical fibre;

removing a wedge from a first cut end of the optical fibre so that a flat face extending at a non-zero angle to the optical axis is exposed;

coating reflective material on the flat face; and

splicing the cut ends of the optical fibre so that a notch is formed between the flat face of the first cut end of the optical fibre and a face of the second cut end of the optical fibre that is not coated by reflective material.

Thus there is provided a method of making an optical coupling arrangement for an optical fibre that comprises a notch in an optical fibre formed between two faces, with only one face, that is the angled face, coated by reflective material. Thus the optical coupling arrangement provides the advantages discussed above of having a high coupling efficiency and low losses for EM radiation passing along the fibre. The optical coupling arrangement can couple EM radiation into the optical fibre and can couple EM radiation out of the optical fibre, which in the latter case allows for its as either an output coupler or as a beamsplitter. The fraction of EM radiation output from the optical coupling arrangement may be tuned by controlling the relative areas of the flat face and the complete fibre cross section.

The advantageous optical coupling arrangement described above is made using a technique that is relatively straightforward to implement and can be performed without specialised equipment, for example as compared to the method of manufacture disclosed in US-5,037, 172. This is achieved by cutting the optical fibre, performing the processing to form the notch and then splicing the cut ends. Thus, both the formation of the notch by removal of a wedge and also the coating of the reflective material are performed on a single cut end, which facilitates both processes. Removal of a wedge to form the notch is straightforward. Coating the flat face is also straightforward, without any risk of reflective material collecting on the other face of the notch. Due to the large open angle between the angled face of the notch and the adjacent cut face of the end of the optical fibre, it is relatively simple to remove any reflective material deposited on the fibre end by polishing following the coating process.

The optical coupling arrangement may be used in a wide range of applications. One application is in FLCRDS or more generally in CDRS (for example linear fibre CRDS), in which the optical coupling arrangement may be used to couple EM radiation into or out of the optical loop, or more significantly both simultaneously. The optical coupling arrangement provides particular advantages over typical techniques, for example using two bend couplers to couple electromagnetic radiation separately into and out of the optical fibre. In particular, the high coupling efficiency allows a high proportion of the EM radiation to be coupled into the optical fibre, thereby reducing the power requirements of the light source, and allowing improved signal levels and/or a smaller, cheaper light source and/or detector. Furthermore, the optical coupling arrangement provides relatively low loss of EM radiation passing around the cavity which can improve the detection sensitivity of the spectroscopy significantly.

According to a further aspect of the present invention, there is provided an optical fibre including an optical coupling arrangement made by a similar method to the first aspect of the present invention.

To allow better understanding, an embodiment of the present invention will now be described by way of non- limitative example with reference to the accompanying drawings, in which:

Fig. 1 is a flow chart of a method of making an optical coupling arrangement for an optical fibre;

Fig. 2 is a cross-sectional view of an optical fibre along the optical axis;

Fig. 3 is a cross-sectional view of an optical fibre that has been cut through;

Fig. 4 is a cross-sectional view of a cut end of an optical fibre with a wedge removed;

Fig. 5 is a perspective view of the cut end of Fig. 4;

Fig. 6 is cross-sectional view of the cut end of Fig. 4 with a coating applied;

Fig. 7 is a cross-sectional view of the cut ends of optical fibre spliced to form the optical coupling arrangement;

Figs. 8 and 9 are cross-sectional views of the optical coupling arrangement showing alternative optical configurations;

Fig. 10 is a view of the mode structure of light coupled into an optical fibre using bend coupling;

Figs. 11 and 12 are views of the mode structure of light coupled into an optical fibre using the optical coupling arrangement;

Fig. 13 is a cross-sectional view of the optical coupling arrangement showing the location of detectors used in a simulation;

Figs. 14 and 15 are graphs of loss against penetration depth of notch;

Fig. 16 is a cross-sectional view perpendicular to the optical axis of the optical fibre at the notch;

Fig. 17 is a cross-sectional view perpendicular to the optical axis of the optical fibre with a notch having an alternative shape; Fig. 18 is a graph of loss against penetration depth for the notches of Figs. 16 and 17;

Fig. 19 is a cross-sectional view of an optical fibre of alternative construction formed with an optical coupling assembly;

Fig. 20 is a diagram of a FLDRS apparatus employing the optical coupling arrangement; Figs. 21 and 22 are graphs of normalised signal intensity over time for a conventional

FLDRS apparatus and the FLDRS apparatus of Fig. 20, respectively; and

Fig. 23 is a graph of normalised ring-down values against end separation for optical fibres of different diameters.

A method of making an optical coupling arrangement for an optical fibre is shown in Fig. 1. The method is performed on an optical fibre 100 comprising a core 1 and a cladding 2 that surrounds the core 1, as shown in Fig. 2. The core 1 may be made of glass, or any other suitable material, for example silica or a plastic. The cladding 2 has a lower refractive index than the core 1 and may be made of any suitable material for example silica, glass or plastic. The optical fibre 100 in this example is a multimode fibre in which the cladding 2 is thin relative to the diameter of the core 1, but in general the optical fibre 100 may be of any other type, for example a multiply clad fibres, such as an all-glass double-clad fibre.

The optical fibre 100 is encased in a coating 15 which has the purpose of providing desired mechanical properties to the optical fibre 100 and plays no part in the guiding of EM radiation. The coating 15 is made of any suitable material such as a plastic.

In step SI, the optical fibre 100 is cut through, in this example along line X-X that is perpendicular to the optical axis O, although in principle the cut could be made at an acute angle to the optical axis O. The cutting in step SI splits the optical fibre 100 into two portions 3 and 4, as shown in Fig. 3, having respective fibre end faces 5 and 6 that are cut faces at their respective cut ends 7 and 8. To facilitate the cutting, a short section of the coating 15 around the site of the cut is removed beforehand.

The cut is preferably made to make fibre end faces 5 and 6 as flat as possible. Any suitable cutting technique may be used, for example a specially designed fibre cutter tool, or a capillary cutter (for example a Shortix™ capillary cutter) that has for has been found to give the best results in lieu of a specially designed fibre cutter tool where the optical fibre 100 is of relatively large diameter. The first cut end 7 of the first portion 3 is further processed in steps S2 to S7, and the second cut end 8 of the second portion 4 is further processed in step S8.

In step S2, the fibre end face 5 of the first cut end 7 of the first portion 3 is polished to produce a flat face, for the purpose of subsequently splicing to the second cut end 8 of the second portion 4. This polishing, and indeed all the polishing steps described below, may be performed using a polishing wheel consisting of a rotating flat plate made of, or covered by, a polishing material, for example a commercially available 0.1 μιη diamond lapping film. During the polishing the first portion 3 may be clamped, for example in a pin vice, pin chuck or specially-designed Perspex mounts.

In step S3, a wedge 9 of the first cut end 7 of the first portion 3 is removed, as shown in Fig. 4, exposing an angled face 10 on the first cut end 7 that is a flat face extending at an obtuse non-zero angle to the fibre end face 5 and at an acute non-zero angle to the optical axis O. The angle to the optical axis O is desirably 45°, but may vary from that. The removed wedge 9 extends through the cladding 2 into the core 1.

As can be seen from Fig. 5 (in which the coating 15 is omitted for clarity), the angled face 10 in this example is shaped as a segment of the fibre end face 5 of the first cut end 7, as viewed along the optical axis O. The removal of the wedge 9 may be performed simply by polishing the first cut end 7 of the optical fibre 100 to the desired penetration depth. The angle of the polishing is set at the desired angle of the angled face 10, for example by clamping the first portion 3 at a different angle to that in step S2.

In steps S4 to S7, the first cut end 7 is treated chemically to coat reflective material 13 on the angled face 10. The reflective material 13 is silver in this example, but may be any suitable material, for example another metal. In this example, the reflective material 13 is coated on the angled face 10 using a coupling agent that couples the reflective material 13 to the material of the optical fibre 100, in particular the core 1. In this example, the coupling agent is VECTABOND™ but as detailed below a wide variety of coupling agents capable of providing such coupling are available and may be used instead.

In step S4, the coupling agent is coated on the entire first cut end 7, that is simultaneously on the fibre end face 5 and on the angled face 10. As the coupling agent is not restricted to the angled face 10, this is straightforward, for example by painting or dipping.

In step S5, the fibre end face 5 of the first cut end 7 is polished to remove the coupling agent from the fibre end face 5. This leaves the coupling agent solely on the angled face 10. Steps S4 and S5 overall achieve coating of the angled face 10 by the coupling agent using a simple process.

In step S6, the reflective material 13 is applied to the angled face 10. As it is difficult to restrict the reflective material 13 to the angled face 10, this is done by applying the reflective material to the entire first cut end 7, that is simultaneously to the fibre end face 5 of the first cut end 7 and to the angled face 10. The reflective material is applied in this example by precipitation from solution, for example in the case of silver from Tollens' reagent that precipitates silver by reduction of silver nitrite. Such an application of a solution is straightforward, for example by painting or dipping. The coating adheres strongly to the angled face 10 by being coupled by the coupling agent, but weakly to the fibre end face 5.

In step S7, any excess reflective material on the fibre end face 5 is removed, if necessary by polishing.

Steps S4 to S7 overall achieve coating of the angled face 10 by the reflective material using a simple process. Any of the polishing steps, that is steps S2, S5 or S7, may also be controlled to reduce the radial depth of the angled face 10 to a desired penetration depth into the optical fibre 100. This is optional but allows for a finer control of the size of the angled face 10 than can be achieved from step S2 alone.

In step S8, the second cut end 8 of the second portion 4 is polished to produce a flat face, for the purpose of subsequently splicing to the first cut end 7 of the first portion 3. Step S8 may be performed at any time relative to steps S2 to S7.

Finally, in step S9, the first and second cut ends 7 and 8 are spliced by the fibre end faces 5 and 6, as shown in Fig. 7. Any suitable splicing technique may be applied, for example mechanical splicing. Suitable techniques may include: clamping the first and second cut ends 7 and 8 together, in which case optionally an index matched material may be applied between the cut faces 5 and 6 to assist the splicing; or bonding the first and second cut ends 7 and 8 with an adhesive that is preferably index-matched to optical fibre 100. The splicing reconstitutes the optical fibre 100 with the fibre end faces 5 and 6 providing contacting surfaces, but due to the processing in steps S3 to S8 a notch 11 is formed between the first cut end 7 of first portion 3 and the second cut end 8 of the second portion 4. In particular the notch 11 has (a) a face formed by the angled face 10 that is coated by the reflective material 13 and (b) a further face 14 formed by the fibre end face 6 of the second cut end 8 that is not coated by the reflective material 13, being a portion of the fibre end face 6 that extends coplanarly from the contacting surface of the fibre end face 6. The notch 11 extends into core 1 through the cladding 2. Thus fabricated, the notch 11 acts as an optical coupling arrangement 12.

In overview, the optical coupling arrangement 12 is made in the optical fibre 100 by a method that is simple and inexpensive, requiring no highly specialised equipment.

Variations are possible as follows.

As an alternative to the further face 14 of the notch 11 that is not coated by the reflective material 13 being a portion of the fibre end face 6, the second cut end 8 of the second portion 4 could have a wedge removed similarly to the first cut end 7 of the first portion 3 to expose a further flat face forming a face of the notch 11. However, this alternative requires additional processing and introduces a requirement to align the cut ends 7 and 8 more closely.

In this example, the angled face 10 of the notch 11 that is coated by the reflective material

13 extends at an angle of 45° to the optical axis O which is desirable for coupling EM radiation into and out of the core 1 perpendicularly. However, this angle may be changed by adjusting the angle of the angled face 10 during removal of the wedge 9. The angle is selected so that the input EM radiation after reflection from the angled face is incident on the boundary between the core 1 and the cladding 2 at an angle of at least the critical angle. This means there are a range of possible angles, between limits derived from the boundaries between the core 1 and the cladding 2 on the opposite sides of the optical fibre 100. This range can be calculated using ray tracing, but is dependent on the incidence angle of input EM radiation relative to the optical axis O, and on whether the first or second optical configuration described below is used. For example, if the incidence angle of input EM radiation relative to the optical axis O is 90° and the first optical configuration is used, then the angle of the angled face 10 relative to the optical axis O can be in the range from (C/2) to (90-C/2), where C is the critical angle of the boundary between the core 1 and the cladding 2, because there is no refraction at the face 14. If the incidence angle of input EM radiation relative to the optical axis O is 90° and the first optical configuration is used, then the angle will be in a range that differs slightly, taking into account the refraction at the face 14. Furthermore, other angles are permitted by changing the incidence angle of input EM radiation relative to the optical axis O, in theory permitting any angle for the angled face 10, although practical considerations such as alignment limit the deviance of the incidence angle of input EM radiation relative to the optical axis O.

Another practical consideration is that making the angle of the angled face 10 relative to the optical axis O shallower increases the target area for the input EM radiation, which provides a benefit with the practical constraint that very shallow angles make it difficult to position the source.

Changing the angle of the angled face 10 may also affect the efficiency of coupling of EM radiation into and out of the core 1 by changing the coupling into modes of the optical fibre 100,. Change in this angle may also require the position and/or orientation of optical components, such as a source or a detector, that couple with the optical coupling arrangement 12.

In this example, the further face 14 of the notch 11 that is not coated by the reflective material 13 extends perpendicular to the optical axis O which is desirable for minimising parasitic reflections at that face. However, this angle may be changed by changing the angle of the fibre end faces 5 and 6, either during the cutting in step S2 or the subsequent polishing in steps S2, S5, S7 or S8, and/or by the alternative mentioned above of removing a wedge from the second cut end 8 of the second portion 4. Change in this angle may affect the efficiency of coupling of EM radiation into and out of the optical fibre 100 by changing the amount of EM radiation reflected at that face and/or the refraction at that surface, but in general any angle may be selected that provides adequate optical coupling. The further face 14 may be inclined towards the notch 11 (i.e. towards the angled face 10), although this may affect the coupling efficiency by reducing the space for a beam incident on, or reflected from, the angled face 10. The further face 14 may be inclined away from the notch 11 (i.e. away from the angled face 10), provided that, in the case of EM radiation being coupled out of the core 1 by the angled face 10 passing through the further face 14 (in the first optical configuration of Fig. 8), the further face is not so shallow that the angle of incidence is greater than the critical angle (the angle of incidence being the angle between the normal of the further face 14 relative to the optical axis O).

Step S3 of removing a wedge 9 of the first cut end 7 of the first portion 3 may alternatively be performed by any suitable technique other than polishing, for example laser machining or focused ion beam milling (FIB).

The optical coupling arrangement 12 may be used to optically couple EM radiation, including but not restricted to visible light, into and/or out of the core 1 as follows, i.e. as an input coupler, an output coupler, or both. Two alternative optical configurations are shown in Figs. 8 and 9 wherein the arrows indicate the passage of EM radiation.

EM radiation is coupled into the optical fibre 100 by reflection from the reflective material 13 on the angled face 10. This may be achieved by directing a beam of EM radiation onto the angled face 10.

In the first optical configuration of Fig. 8, the input EM radiation is incident on the internal side of the angled face 10, by being directed laterally through the optical fibre 100 from the opposite side of the optical fibre 100 from the notch 11. The EM radiation is reflected from the angled face 10 inside the core 1 away from the notch 11 along the optical fibre 100.

In the second optical configuration of Fig. 9, the input EM radiation is incident on the external side of the angled face 10, by being directed into the notch 11 from the same side of the optical fibre 100 as the notch 11. The EM radiation is reflected inside the notch 11 and enters the optical fibre 100 through the further face 14 of the notch 11 that is not coated with reflective material 13.

In both optical configurations, the EM radiation reflected from the angled face 10 couples into stable optical modes of the core 1 with a high coupling efficiency. By way of comparison, the coupling efficiency is significantly better than with bend coupling, which is particularly problematic for multi-mode optical fibre of relatively large diameter. Bend coupling is an arrangement in which EM radiation is coupled into a optical fibre by illuminating a bend in the optical fibre at an appropriate angle. At a bend, an optical fibre suffers so-called "macrobending losses", in which some of the EM radiation propagating in the optical fibre no longer strikes the boundary between the core and the cladding at an angle greater than the critical angle, and is instead refracted out through the cladding. Bend coupling uses the same process in reverse to couple light into the optical fibre. Using bend coupling, only a few percent of the incident EM radiation is at best coupled into the core, and the EM radiation is coupled preferentially into high order core modes and cladding modes.

This is illustrated for example in Fig. 10 which shows the mode structure within the core following bend coupling into the fibre. This is, for example, a far from ideal spatial distribution of EM radiation for spectroscopic interrogation of samples. Using the optical coupling assembly, virtually 100% of input EM radiation may be coupled into the core 1 , as shown by a Zemax™ optical modelling simulation. Furthermore, EM radiation is introduced into the full range of cavity modes after propagating some distance along the core 1. This is illustrated for example in Figs. 11 and 12 which show two examples of the mode structures achieved by the optical coupling assembly. The mode structure may be tuned by selection of the radial penetration depth of the notch 11 into the core 1 , or by the alignment of the notch relative to the flat fibre end. In principle all of the incident EM radiation is reflected from the reflective material 13, although there may be relatively small losses on entry into the optical fibre core 1 in the first optical configuration, or at the further face 14 of the notch 11 that is not coated with reflective material 13 in the second optical configuration.

EM radiation passing along the optical fibre 100 is coupled out of the optical fibre 100 by reflection from the reflective material 13 on the angled face 10.

In the first optical configuration of Fig. 8, the output EM radiation passes into the notch 11 through the further face 14 of the notch 11 that is not coated with reflective material 13 and is incident on the external side of the angled face 10. The EM radiation is reflected inside the notch 11 and exits laterally the optical fibre 100 on the same side as the notch 11.

In the second optical configuration of Fig. 9, the output EM radiation is incident on the internal side of the angled face 10. The EM radiation is reflected from the angled face 10 inside the core 1 and after passing through the optical fibre 100 exits laterally from the optical fibre 100 on the opposite side from the notch 11.

In both optical configurations, whilst some of the EM radiation is output, the remainder of the EM radiation passes along the optical fibre 100. The fraction of EM radiation output from the optical fibre 100 may be tuned by controlling the cross-sectional area of angled face 10 relative to the cross-sectional area of the core 1. The optical coupling assembly may therefore be used as an output coupler or a beam splitter.

The configuration of the optical coupling assembly allows for the output coupling to occur whilst minimising parasitic losses in the light passing along the optical fibre 100 past the optical coupling assembly. By way of comparison, such losses are significantly better than with bend coupling. Using 'bend coupling', the losses introduced by the bend are typically as high as 10-15%. Using the optical fibre coupling arrangement, the losses may be made lower and in some cases reduced to as little as 1%.

Simulations of the optical coupling assembly used as a beam splitter have been carried out using the Zemax™ optical modelling software suite, Zemax™ being an optical ray-tracing program which is commonly used for optical design and tolerance analyses. The simulations assumed multimode propagation of EM radiation. In the simulations, rays are launched from a fibre-coupled divergent source, whose angle of divergence (9 degrees) matches a typical maximum acceptance angle of the optical fibre. This provides a fairly realistic mode picture of the optical fibre core 1 and cladding 2, so that the range of propagation angles of EM radiation escaping from the notch is modelled correctly.

Fig. 13 is a cross-sectional view of the core 1 and cladding 2 showing the location of detectors 21, 22 and 23 used in the software for the quantitative determination of output coupling efficiency. Detector 21 measures the amount of EM radiation lost at the notch 11. Detector 22 measures the amount of EM radiation in the core 1 after the notch 11. Detector 23 measures the small amount of EM radiation which travels in the cladding 2. The percentage of propagating EM radiation L that is coupled out of the fibre is determined from the signals D_\, D₂ and Z¼ recorded at the three detectors 21, 22 and 23:

A

D_{1 +} D_{2 +} D ^Xl0° ⁽D

This percentage loss L is plotted as the solid points in Figs. 14 and 15 as a function of penetration depth d of the notch 11 into the core 1 , as shown in Fig. 16 which in this example has a penetration depth d of lOum in an core 1 of diameter 365μιη with cladding 2 of total diameter 400μιη.

Predictions may be made for a perfect notch 11 , wherein the output coupling efficiency is simply given by the fractional area of the optical fibre 1 taken up by the notch 11. Straightforward trigonometry shows this to be:

100

L ( 2 cos^_1x - sin(2 cos^_1x)) (2)

r - d

where x = , where r is the radius of the optical fibre 1 and d is the penetration distance.

These predictions are plotted by the continuous line in Figs. 114 and 15 which show that the predictions are in virtually perfect agreement with the results of the Zemax™ simulations.

Desirably, the penetration depth d of the notch 11 into the core 1 is selected to lie in the low loss region of interest shown in Fig. 15, where the penetration depth d is less than 10% of the radius of the core 1 , which in this example provides a loss per pass of less than 6%.

The angled face 10 of the notch 11 described above is shaped as a segment of the fibre end face 5, as viewed along the optical axis O, but the notch 11 may be arranged to have other shapes.

One alternative possibility is for the angled face 10 to be shaped to protrude into the optical fibre 100 from the edge of the optical fibre 100, as viewed along the optical axis O. In principle, as such a notch 10 does not need to extend as a segment across the full optical fibre 100, a lower loss could be achieved. As an example of this, Fig. 17 illustrates a notch 11 formed by an angled face 10 shaped as a rectangle, as viewed along the optical axis O. For such a rectangular notch, 11 the percentage loss L can be approximated as:

'notch cross-sectional area^ (Ιι Ιτλ

L = 100 -. : « 100 (3)

v core cross-sectional area J pr J '

where _/ and are the width and height of the notch cross section in the core, as shown in

Fig. 17.

Fabrication of such notches is more difficult than when the angled face 10 is shaped as a segment. However, the notch 11 could be made using FIB. This technique yields extremely high resolution structures, although it is slow and expensive, as compared to polishing. Nonetheless, this type of notch 11 would give lower total losses, as the cross-sectional area of the notch 11 is lowered. As long as the area of the core 1 is greater than 8um by 8um, then EM radiation from a single-mode fibre can still be totally coupled into the core 1. For example, a notch 11 that is shaped as a 1 Ομιη by ΙΟμιη square in a core 1 of diameter 365μιη should entail a loss of only 0.1%. This is in contrast to notch shaped as a segment of penetration depth d equal to ΙΟμιη in a core 1 of diameter 365μιη, for which a loss of 1.5% is calculated. The calculated losses for the notches 11 shown in Figs. 16 and 17 are demonstrated in Fig. 18, plotted for the rectangular notch 11 of Fig. 17 having a constant width (/_/) of ΙΟμιη and a height (l₂) equal to the notch penetration depth d of the segment notch 11 of Fig 16.

In the above examples, the removed wedge 9 and hence the notch 11 extends through the cladding 2 into the core 1. However this is not essential and in some cases the removed wedge 9 and hence the notch 11 may extend only into the cladding 2. An example where this is the case is shown in Fig. 19, wherein the optical fibre 100 has an alternative construction in which the optical fibre 100 is a single mode fibre in which the core 1 is of smaller diameter relative to the cladding 2, than in the example shown in Figs. 2 to 7. In this case, the optical coupling arrangement 12 is formed in the same manner as described above except that the removed wedge 9 and hence the notch 11 extend only into the cladding 2 and not into the core 1. The coupling agent is selected to couple the reflective material 13 to the material of the cladding 2. This optical coupling arrangement 12 operates in the same manner as described above and has similar advantages, except that the coupling efficiency may be lower because the EM radiation couples with high efficiency only into the modes of the cladding 2, not into the modes of the core 1 (although during subsequent propagation will couple into the core 1).

The optical coupling arrangement 12 may have a range of uses, as will now be described. The optical coupling arrangement 12 is particularly suitable for coupling EM radiation into and out of an optical cavity in CRDS, including in particular FLCRDS.

Fig. 20 shows an example of an apparatus 30 for FLCRDS using the optical coupling arrangement 12. The apparatus 30 includes the optical fibre 100 arranged in a loop which therefore constitutes a cavity.

The apparatus 30 has a sample holder 31 arranged to hold a sample and disposed within the loop of the optical fibre 100 to absorb EM radiation passing around the cavity. The sample holder 31 may have two alternative configurations shown in the insets of Fig. 20.

In the first alternative configuration, the sample holder 31 is a container 32 for the sample placed between the two ends of the optical fibre 100 which couple through the container 32 to form the loop. Thus the EM radiation is transmitted directly through the sample in the container 32. The first alternative configuration is particularly suitable for a sample that is a liquid, although the sample could alternatively be a gas

In the second alternative configuration, the sample holder 31 is a container 33 arranged to hold a sample that is typically a gas around a thinned portion 34 of the optical fibre 100 shaped to overlap the evanescent field of the EM radiation with the sample. Other sample configurations are possible.

The apparatus 30 includes the optical coupling assembly 12 within the optical fibre 100 used as both an input coupler and an output coupler, in either the first or second optical configuration of Fig. 8 or Fig. 9. The apparatus includes a source 35 of EM radiation and a detector for the EM radiation both coupled into the optical fibre 100 by the optical coupling arrangement 12. The source 35 may be a laser as is conventional for CRDS. However, the high coupling efficiency reduces the power of the source 35 needed to achieve a desired signal power, thereby allowing the possibility of higher signal powers and/or cheaper types of source, possibly in some applications allowing the use of lower intensity sources such as a light emitting diode (LED). The detector 35 may be of any suitable type, for example a photomultiplier tube as is conventional for CRDS, but the optical coupling arrangement 12 also opens up the possibility in some cases of a lower cost detector such as a photodiode.

Although the apparatus 30 is configured for FLCRDS by arranging the optical fibre 100 in a loop to form the cavity, the apparatus 30 could be adapted for CRDS using a linear cavity by changing the configuration of the optical fibre 100 to include mirrors at each end, for example by coating the ends of the optical fibre 100 with reflective material.

FLCRDS is an extremely sensitive spectroscopic technique being developed for real-time detection applications in microfluidics and other analyses involving small liquid volumes.

Micro fluidic, or 'lab on a chip', systems show great promise for reaction prototyping and many different types of chemical analysis, but the extremely small sample volumes provide a challenge for any detection technique. Only a few on-chip spectroscopic detection schemes have been

demonstrated, and most of these are based on fluorescence. This is only useful if the analyte is a fluorescent species or can be fluorescently tagged. As a result, there is considerable interest in developing ultra-sensitive absorption spectroscopy techniques, since virtually every chemical species absorbs somewhere in the range of easily accessible wavelengths.

In any absorption spectroscopy, the absorption signal depends on the absorption coefficient and concentration of the substance under study, and on the optical path length through the sample. Cavity ring-down methods such as CRDS use an optical cavity to enhance the optical path length, and therefore the detection sensitivity, by many orders of magnitude. Cavity-based spectroscopies have become the ultra-sensitive detection method of choice for gas-phase measurements, and have the potential to be adapted to the liquid phase. FLCRDS replaces the two-mirror optical cavity used in gas-phase measurements with a loop of the optical fibre 100.

Achievement of high detection sensitivities requires minimisation of losses around the fibre loop, and maximisation of EM radiation intensity transmitted to the detector. Use of the optical coupling arrangement 12 satisfies both of these requirements. In general, in a conventional pulsed CRDS experiment, for example using a bend coupler as the input coupler, only a tiny fraction of the incident laser pulse (at best a few percent) is coupled into the cavity. Use of the optical coupling arrangement 12 allows virtually all of the laser pulse to be coupled into the cavity, leading to greatly improved signal levels without compromising the performance of the cavity. It is estimated that the optical coupling arrangement 12 will improve the detection sensitivity of FLCRDS over conventional techniques by one to two orders of magnitude.

One other example achieving high degrees of input coupling into the cavity of a CRDS apparatus is disclosed in Fabian et al., "Optical fibre cavity for ring-down experiments with low coupling losses", Measurement science & technology 21, 2010, 094034 relating to a linear fibre cavity in which the input coupler is formed by fabricating a small central hole in a reflective end face of the fibre using FIB which is effective, but time consuming and expensive technique that cannot be applied to a cavity in the form of a loop.

Thus, use of the optical coupling assembly 12 allows FLCRDS to be used for the study of fluids with the advantages discussed above, for example with lower intensity sources, possibly an LED, rather than a laser, greatly reducing associated costs and making the technique more attuned to the 'cheap as chips' philosophy of microfluidics.

In overview, the optical coupling arrangement 12 may open up CRDS and related techniques to a much wider range of EM radiation sources, thereby reducing costs considerably. As well as a reduction in costs associated with the source, the higher EM radiation intensity circulating in the cavity and incident on the detector will allow the relatively expensive photomultiplier generally required at present to be replaced with much lower cost photodiodes.

Use of the optical coupling arrangement 12 has been tested in a CRDS apparatus as shown in

Fig. 20 with the first configuration of the sample holder 31. Representative results for a conventional apparatus employing bend coupling and for the optical coupling arrangement 12 are shown in Figs. 21 and 22 respectively being graphs of the normalised signal intensity over time.

Fig. 21 shows the ring-down signal for bend-coupled multimode optical fibre with a diameter of 105μιη and cladding of diameter 125μιη. For larger diameter optical fibres the high losses associated with bend coupling render this approach impracticable. In this example, the loop length was 10.30 m and two separate bends with 9 mm bend radii were required to act as input and output couplers. The loss per pass was 16%, the ring-down time (exponential decay constant) was 235 ns, and the peak signal intensity recorded by a photomultiplier connected to a digital oscilloscope was 50 mV.

Fig. 22 shows the ring-down signal the optical coupling arrangement 12 provided in an optical fibre 100 that is a multimode fibre with core 1 of diameter 365μιη and cladding 2 of diameter 400μιη. The loop length was 18.69m. The loss per pass attributed to coupling losses was 1.5% (there is an additional loss from fibre absorptions at the laser wavelength, in this case the manufacturer's specification for this fibre absorption at 532nm is 0.00345m^"1, and therefore the total losses from the fibre absorption with an 18.69m loop is 6.5%>. This loss is reduced simply by reducing the loop length; for example, we have investigated loops only 3m in length, in this case losses from fibre absorptions are only 1%, i.e. similar to the loss in the coupler themselves), the ring-down time was 1.166μ8, and the peak signal was 500mV. The true peak signal level is considerably higher, as in these measurements a gated photomultipher was employed in order to delay signal acquisition until the EM radiation inside the cavity had decayed to levels that would not saturate the detection. We have so far achieved a maximum signal of 4V in our experiments.

In addition, Fig. 23 is a graph of normalised ring-down values against end separation for optical fibres 100 of different diameters in the FLCRDS apparatus of Fig. 20. This shows that increasing size of the core 1 reduces the losses across a gap introduced into the loop, thereby greatly simplifying the design of sample regions. Use of an core 1 of diameter larger than ΙΟΟμιη is made possible by the optical coupling arrangement 12.

As can seen, use of the optical coupling arrangement 12 improves the detection sensitivity by reducing the loop losses and increasing the amplitude of the signal. Introducing the optical coupling arrangement 12 reduces the loop losses dramatically, leading to a much longer exponential ring- down decay, and the improved coupling efficiency into the loop leads to greatly improved signal amplitude. Use of the optical coupling arrangement 12 also allows use of an optical fibre lOOhaving a core 1 of larger diameter which is advantageous in FLCRDS as it greatly simplifies the design of the sample holder 31 of the first configuration due to reduced losses across the container 32.

Besides CRDS, the optical coupling arrangement 12 has broader application in the general area of optical fibre couplers/splitters. There are many other applications in which side-coupling of EM radiation into and/or out of an optical fibre 100 is desirable and in which the optical coupling arrangement 12 may be employed. Fibre couplers and splitters are widely available in

coupling/splitting ratios ranging from 50:50 to 99:1, but are generally optimised for a specific wavelength. The optical coupling arrangement 12 has the advantage that the coupling/splitting ratio is determined almost purely by geometrical considerations, thereby allowing fabrication of couplers/sp litters with any desired coupling/splitting ratio and which function over a broad range of wavelengths dependent only on the reflectivity of the metal deposited on the notch. There may also be applications in the field of fibre lasers for introducing a pump beam or beams to the lasing medium.

The nature of the coupling agent will now be discussed. In general, the coupling agent may be any material that is capable of coupling the reflective material to the material of the optical fibre 100 or to the core 1. The reflective material should be secured to the fibre end in a way that is both mechanically and chemically-stable. A wide variety of such coupling agents are available and may be used.

In the case that the reflective material is a metal and the core 1 is made of glass (or silica), the coupling agent may be a linker compound of the type commonly used to deposit metal on glass. Many such linker compounds are known, as such deposition is important in a wide range of scientific applications, for example electrode fabrication. In order to increase the adhesion of a metal such as gold or silver to the glass surface, the surface of the glass is typically modified using an organic coupling agent, most commonly an organo- functional silane. Such coupling agents may have one end that can bind to the glass surface, and an opposite end that is attractive to metal binding. These compounds may comprise a three-part structure illustrated below, where the group designated Nu (indicating an electron-rich, nucleophilic group) binds metals, and the Si(OR)₃ group bonds to glass. The linker chain, L, is typically a hydrocarbon chain and can be any number of carbon atoms in length.

Thus, typically, the coupling agent is a compound of formula (I)

R

N u— L < Si /°-^R2

\

R³ (I)

wherein:

Nu is a nucleophilic group which is capable of binding to a metal;

L is a linker chain of formula -(Y)„-, wherein n is a positive integer and the or each Y, which may be the same or different where n is greater than 1 , is selected from -alk-, -arylene-, -X- alk-, -alk-X-, -X-alk-X-, -X-arylene-, -arylene-X-, -X-arylene-X-, wherein X is O, C(O), C(0)0, OC(O), N(R"), C(0)N(R"), N(R")C(0), and wherein alk is an unsubstituted or substituted Ci_-20 alkylene group, which C₁.₂₀ alkylene group is optionally interrupted by O, C(O), C(0)0, OC(O), N(R"), C(0)N(R"), N(R")C(0) or arylene, and wherein R" is H, Ci_₆ alkyl or aryl; and

R¹, R² and R³, which are the same or different, are independently selected from H, unsubstituted or substituted CM_O alkyl and aryl, wherein at least one of R¹, R² and R³ is an unsubstituted or substituted CM_O alkyl group or an aryl group.

Nu may be any suitable nucleophilic group which is capable of binding to a metal; many such nucleophilic groups are known to the skilled person. Such groups include amino (NH₂) and thiol (SH) groups, and aryl- and alkyl- derivatives of amino and thiol groups. Thus, Nu may be selected from NH₂, SH, NHR⁴, NR⁴R⁵ and SR⁴, wherein R⁴ and R⁵ are selected from Ci_-6 alkyl and aryl.

The positive integer n may be an integer of from 1 to 200. Thus, L may be a polymeric or oligomeric linker. However, more typically, n is an integer of from 1 to 50, or for instance from 1 to 10. The integer n may for instance be from 1 to 5.

The or each Y is typically an unsubsituted or substituted CM_O alkylene group, preferably Ci_ ₁₀ alkylene.

The integer n is typically 1. Thus, the linker chain L may be an unsubsituted CM_O alkylene group, or for instance an unsubsituted Ci_₆ alkylene group such as methylene, ethylene, propylene, butylene, pentylene or hexylene. Typically, R¹, R² and R³, which are the same or different, are selected from unsubstituted or substituted Cuo alkyl and aryl. R¹, R² and R³ may for instance be unsubstituted CMO alkyl groups, or for instance unsubstituted CM alkyl groups such as methyl, ethyl, propyl or butyl.

Examples of suitable commercially-available coupling agents include 3- aminopropyltriethoxysilane (APTES) or 3-mercaptopropyltrimethoxysilane (APMS) Another example of a suitable coupling agent is VECTABOND™, which provides a simple fabrication route to glass binding, because the coupling agent may be applied simply by immersing the optical fibre 100 in an acetone solution of VECTABOND™ for a suitable period, typically around 5 minutes. VECTABOND™ is a silane-based reagent that has a similar chemical structure to APTES and uses an amine group to bind the metal.

Other details of suitable coupling agents may be found in Goss, Charych and Majda, Anal. Chem., 1991 , 63, pp 85-88 which is incorporated herein by reference.

In the case that the reflective material is a metal and the core 1 is made of plastic, a suitable coupling agent may be employed which comprises a group Nu which is capable of binding to a metal, linked to a group Z which is capable of binding to a plastic. Such coupling agents include compounds that have the same structure as formula (I) above except that a group, Z, which is capable of binding to a plastic, is employed in place of the Si(OR¹)(OR²)(OR³) group. Thus, in the case that the reflective material is a metal and the core 1 is made of plastic, the coupling agent may be a compound of formula (II)

Nu— L Z (Π)

wherein Nu and L are as defined above for the compound of formula (I) and Z is a group which is capable of binding to a plastic. Many suitable groups Z which are capable of binding to plastics are known to the skilled person. For example, in the case of the plastic being PMMA, the end functional group of the PMMA is a methyl ester, which is easy to functionalise with a linker compound. For example, the linker compound may be a compound such as N-lithioethylenediamine, which has an amine group on either end, would be suitable, for forming a strong amide bond with the PMMA at one end, and then leaving a free amine group to bind a metal, for example as disclosed in Alyssa et al., "Surface Modification of Poly(methyl methacrylate) Used in the Fabrication of Microanalytical Devices", Analytical Chemistry, 2000 72 (21), pp. 5331 -5337, which is incorporated herein by reference.

The following substituent definitions apply with respect to the compounds of formulae (I) and (II).

A CMO alkyl group is an unsubstituted or substituted, straight or branched chain saturated hydrocarbon radical. Typically it is Ci_6 alkyl, for example methyl, ethyl, propyl, butyl, pentyl, or hexyl, or CM alkyl, for example methyl, ethyl, i-propyl, n-propyl, t-butyl, s-butyl or n-butyl. When an alkyl group is substituted it typically bears one or more substituents selected from Ci_6 alkyl which is unsubstituted, aryl (as defined herein), cyano, amino, CMO alkylamino, di(Ci_io)alkylamino, arylamino, diarylamino, arylalkylamino, amido, hydroxy, halo, carboxy, ester, keto, Ci_6 alkoxy, aryloxy, haloalkyl, sulfonic acid, sulfhydryl (i.e. thiol, -SH), Ci.io alkylthio, arylthio and sulfonyl. Examples of substituted alkyl groups include haloalkyl, hydroxyalkyl, aminoalkyl, alkoxyalkyl and alkaryl groups. The term alkaryl, as used herein, pertains to a Cuo alkyl group in which at least one hydrogen atom (e.g., 1 , 2, 3) has been replaced with an aryl group. Examples of such groups include, but are not limited to, benzyl (phenylmethyl, PhCH₂-), benzhydryl (Ph₂CH-), trityl

(triphenylmethyl, Ph₃C-), phenethyl (phenylethyl, Ph-CH₂CH₂-), styryl (Ph-CH=CH-), cinnamyl (Ph-CH=CH-CH₂-).

An aryl group is a substituted or unsubstituted, monocyclic or bicyclic (typically

monocyclic) aromatic group which typically contains from 6 to 14 carbon atoms, preferably from 6 to 10 carbon atoms in the ring portion. Examples include phenyl, naphthyl, indenyl and indanyl groups. An aryl group is unsubstituted or substituted. Typically, an aryl group is a substituted or unsubstituted phenyl group. When an aryl group as defined above is substituted it typically bears one or more substituents selected from Ci-Ce alkyl which is unsubstituted (to form an aralkyl group), aryl which is unsubstituted, cyano, amino, Cuo alkylamino, di(Cuo)alkylamino, arylamino, diarylamino, arylalkylamino, amido, hydroxy, halo, carboxy, ester, keto, Ci_6 alkoxy, aryloxy, haloalkyl, sulfhydryl (i.e. thiol, -SH), Cuo alkylthio, arylthio, sulfonic acid and sulfonyl. Typically it carries 0, 1 , 2 or 3 substituents. The term aralkyl as used herein, pertains to an aryl group in which at least one hydrogen atom (e.g., 1 , 2, 3) has been substituted with a Ci_6 alkyl group. Examples of such groups include, but are not limited to, tolyl (from toluene), xylyl (from xylene), mesityl (from mesitylene), and cumenyl (or cumyl, from cumene), and duryl (from durene).

A Ci-20 alkylene group is an unsubstituted or substituted bidentate moiety obtained by removing two hydrogen atoms, either both from the same carbon atom, or one from each of two different carbon atoms, of a hydrocarbon compound having from 1 to 20 carbon atoms (unless otherwise specified), which may be aliphatic or alicyclic, and which may be saturated, partially unsaturated, or fully unsaturated. Thus, the term "alkylene" includes the sub-classes alkenylene, alkynylene, cycloalkylene, etc., discussed below. Typically it is Cuo alkylene, for instance Ci_6 alkylene. Preferably it is C alkylene, for example methylene, ethylene, i-propylene, n-propylene, t- butylene, s-butylene or n-butylene. It may also be pentylene, hexylene, heptylene, octylene and the various branched chain isomers thereof. An alkylene group may be unsubstituted or substituted as specified above for alkyl.

An arylene group is an unsubstituted or substituted bidentate moiety obtained by removing two hydrogen atoms, one from each of two different aromatic ring atoms of an aromatic compound, which moiety has from 5 to 14 ring atoms (unless otherwise specified). Typically, each ring has from 5 to 7 or from 5 to 6 ring atoms. An arylene group may be unsubstituted or substituted, for instance, as specified above for aryl. Typically, the arylene group is phenylene. Ci-₂₀ alkylene groups as defined herein are either uninterrupted or interrupted by one or more heteroatoms or heterogroups, such as O, C(O), C(0)0, OC(O), N(R"), C(0)N(R"), N(R")C(0) wherein R" is H, Ci_6 alkyl or aryl (typically phenyl), or by one or more arylene (typically phenylene) groups. The phrase "optionally interrupted" as used herein thus refers to a C₁.₂₀ alkylene group, as defined above, which is uninterrupted or which is interrupted between adjacent carbon atoms by a heteroatom such as oxygen, by a heterogroup such as C(O), C(0)0, OC(O), N(R"), C(0)N(R"), N(R")C(0) wherein R" is H, aryl or C_rC₆ alkyl, or by an arylene group.

For instance, an alkylene group such as n-butylene may be interrupted by the heterogroup N(R") as follows: -CH₂N(R")CH₂CH₂CH₂-_; -CH₂CH₂N(R")CH₂CH₂-, or -CH₂CH₂CH₂N(R")CH₂-.

Claims

1. A method of making an optical side-coupling arrangement for an optical fibre, the method comprising:

cutting through the optical fibre;

coating reflective material on the flat face; and

2. A method according to claim 1, wherein said step of coating reflective material on the flat face comprises:

applying the reflective material simultaneously to the cut face of the first cut end of the optical fibre and to said flat face; and

polishing the first cut end of the optical fibre so as to remove reflective material from the cut face.

3. A method according to claim 1, wherein said step of coating reflective material on the flat face comprises:

coating a coupling agent, capable of coupling the reflective material to the material of the optical fibre, on the flat face; and

applying the reflective material simultaneously to the cut face of the first cut end of the optical fibre and to said flat face so that the coupling agent couples the reflective material to the flat face.

4. A method according to claim 3, wherein the coupling agent is an organo-functional silane.

5. A method according to claim 3 or 4, wherein said step of coating a coupling agent on the flat face comprises:

coating the coupling agent simultaneously on the cut face of the first cut end of the optical fibre and on said flat face; and

polishing the cut face of the first cut end of the optical fibre, so as to remove the coupling agent from the cut face.

6. A method according to any one of claims 3 to 5, further comprising a further step of polishing the first cut end of the optical fibre after said step of applying the reflective material simultaneously to the cut face of the first cut end of the optical fibre and to said flat face and before said step of splicing the cut ends of the optical fibre.

7. A method according to any one of claims 2, 5 or 6, further comprising a further step of polishing the first cut end of the optical fibre before said step of applying reflective material to the flat face.

8. A method according to any one of claims 2, 5, 6 or 7, wherein said step, or one of said steps, of polishing the first cut end of the optical fibre is performed after said step of removing a wedge so as to reduce the radial depth of penetration of the flat face.

9. A method according to any one of the preceding claims, further comprising a step of polishing the second one of the cut ends of the optical fibre prior to said step of splicing the cut ends of the optical fibre.

10. A method according to any one of the preceding claims, wherein the reflective material is a metal.

11. A method according to claim 10, wherein the metal is silver.

12. A method according to any one of the preceding claims, wherein the optical fibre comprises a core and a cladding surrounding the core, the cladding having a lower refractive index than the core.

13. A method according to claim 12, wherein the removed wedge extends through the cladding and into the core.

14. A method according to claim 12, wherein the removed wedge extends through the cladding but not into the core.

15. A method according to any one of claims 11 to 14, wherein the core is made of glass.

16. A method according to any one of the preceding claims, wherein the optical fibre is encased by a coating, the notch extending through the coating into optical fibre.

17. A method according to any one of the preceding claims, wherein the flat face is shaped as a segment of the cut face of the first cut end of the optical fibre, as viewed along the optical axis.

18. A method according to claim 13, wherein said step of removing a wedge from a first cut end of the optical fibre is performed by polishing first cut end of the optical fibre at said acute angle to the optical axis.

19. A method according to any one of claims 1 to 16, wherein the flat face is shaped to protrude into the optical fibre from the edge, as viewed along the optical axis.

20. A method according to any one of the preceding claims, wherein said face of the second cut end of the optical fibre that forms the notch is a portion of the cut face of the second one of the cut ends of the optical fibre that is spliced to the first cut ends of the optical fibre.

21. A method according to any one of the preceding claims, wherein the cut faces of the cut ends of the optical fibre that are spliced extend perpendicular to optical axis.

22. An optical fibre including an optical coupling arrangement made by a method according to any one of the preceding claims.

23. An optical fibre including an optical coupling arrangement, wherein

two ends of respective portions of the optical fibre are spliced by contacting surfaces of said ends, and

the optical coupling arrangement comprises a notch formed between (a) a flat face exposed by removal of a wedge from the end of a first one of the portions of optical fibre, the flat face extending at a non-zero angle to the contacting surfaces and at a non-zero angle to the optical axis, and (b) a face of the end of the second one of the portions of the optical fibre, the flat face of the end of the first portion of optical fibre having reflective material coated thereon and the face of the end of the second portion of the optical fibre not having reflective material coated thereon.

24. An optical fibre according to claim 23, wherein the contacting surface of the first portion of the optical fibre is polished.

25. An optical fibre according to claim 23 or 24, wherein said the reflective material is coupled to the flat face by a coupling agent.

26. An optical fibre according to claim 25, wherein the coupling agent is an organo- functional silane.

27. An optical fibre according to any one of claims 23 to 26, wherein the contacting surface of the second portion of the optical fibre is polished.

28. An optical fibre according to any one of claims 23 to 27, wherein the reflective material is a metal.

29. An optical fibre according to claim 28, wherein the metal is silver.

30. An optical fibre according to any one of claims 23 to 29, wherein the optical fibre comprises a core and a cladding surrounding the core, the cladding having a lower refractive index than the core.

31. An optical fibre according to claim 30, wherein the removed wedge extends through the cladding and into the core.

32. An optical fibre according to claim 30, wherein the removed wedge extends through the cladding but not into the core.

33. An optical fibre according to any one of claims 30 to 32, wherein the core is made of glass.

34. An optical fibre according to any one of claims 23 to 33, wherein the optical fibre is encased by a coating, the notch extending through the coating into optical fibre.

35. An optical fibre according to any one of claims 23 to 34, wherein the flat face is shaped as a segment of the cut face of the first cut end of the optical fibre, as viewed along the optical axis.

36. An optical fibre according to claim 35, wherein the flat face is polished.

37. An optical fibre according to any one of claims 23 to 34, wherein the flat face is shaped to protrude into the first portion of the optical fibre from the edge, as viewed along the optical axis.

38. An optical fibre according to any one of claims 23 to 37, wherein said face of the end of the second one of the portions of the optical fibre that forms the notch is a coplanarly extended portion of the contacting surface that is spliced to the first portion of the optical fibre.

39. An optical fibre according to any one of claims 23 to 38, wherein the contacting surfaces of the ends of the portions of the optical fibre that are spliced extend perpendicular to optical axis.

40. An apparatus for performing cavity ring-down spectroscopy, the apparatus comprising:

an optical fibre including an optical coupling arrangement according to any one of claims 22 to 39, the optical fibre being arranged in a cavity and having a sample holder arranged to hold a sample so as to absorb EM radiation passing around the cavity;

a source of electromagnetic radiation coupled into the optical fibre by said optical coupling arrangement;

a detector for said electromagnetic radiation coupled into the optical fibre by said optical coupling arrangement.

41. An apparatus according to claim 40, wherein the optical fibre is arranged in a cavity by being arranged in a loop.

42. A method of performing cavity ring-down spectroscopy comprising:

arranging an optical fibre including an optical coupling arrangement according to any one of claims 22 to 39, in a cavity;

arranging a sample to absorb EM radiation passing around the cavity;

coupling electromagnetic radiation into the optical fibre with said optical coupling arrangement;

coupling electromagnetic radiation out of the optical fibre with said optical coupling arrangement; and

detecting the electromagnetic radiation coupled out of the optical fibre.

43. A method according to claim 42, wherein the step of arranging the optical fibre in a cavity comprises arranging the optical fibre in a loop.