1. Introduction

Fiber Bragg gratings are key elements in high speed optical networks and fiber sensors. These devices are fabricated by photo-inducing periodic variations in the refractive index of the core of optic fiber via exposure to an UV interferometric pattern. The fibers used for this application are photosensitive over most of the UV spectrum, although they are generally only weakly photosensitive at wavelengths >270nm [1]. Consequently, ArF (193nm) [2] and KrF excimers (248nm) [3], frequency doubled argon ion (244nm) [4] and copper lasers (255nm) [5], and frequency quadrupled Nd:YLF (262nm) [6] and Nd:YAG lasers (266nm) [7] are all effective sources for fabricating fiber Bragg gratings. Unfortunately, the protective polymer jacket of standard fiber is generally highly absorbing at wavelengths <300nm and therefore, must be removed.

There are two main techniques for removing the fiber jacket. The most common of these is mechanical stripping, however, process repeatability can be a problem as the blade in these systems wears with time and commercial devices are not suited to removing the jacket from the centre of a length of fiber. Another commonly used method for jacket removal is chemical stripping using solutions of dichloromethane, however, volatile chemicals such as this are undesirable for clean room environments. Perhaps the predominant problem associated with these methods is that they both depend on a degree of physical contact with the fiber, which ultimately weakens the fiber and reduces its longevity [8]. In the former case the fiber must be clamped in position while the jacket is forcibly removed, while in the latter the fiber must be physically cleaned in a secondary process to remove any chemical residue.

Although novel fiber Bragg grating writing techniques have been demonstrated which obviate the need to remove the outer coating, such as writing at near UV wavelengths [9,10] or in fibers with UV transmissible outer coatings [11], it is still necessary to remove small sections of the coating when packaging these devices into athermal protective cases. This is because mounting via the soft outer polymer jacket causes excessive drift of the central wavelength. As a result, there is significant interest in developing non-contact methods of polymer jacket removal. One example is chemical stripping using hot sulphuric acid, which has been shown to result in a similar tensile strength to pristine fiber [12]. The disadvantage of this method (shared by other “wet” processing methods) is that it is difficult to accurately control the region to be stripped. Another method generating significant interest is polymer jacket removal via laser ablation, which as well as being a “dry” processing technique also permits better control of the region of jacket to be removed. Fiber jacket removal has already been successfully demonstrated using short wavelength ArF and KrF excimer lasers, and long wavelength CO₂ lasers (10.6μm) [13], however, there are a number of problems associated with using these wavelengths for this purpose. Firstly, laser ablation using CO₂ laser radiation is a pyrolytic process thus it is difficult to produce well defined features within the polymer jacket using this source. Furthermore, glass is highly absorbing at 10.6μm [14], therefore, exposure to CO₂ radiation causes heating of the fiber which increases the potential for destructive damage, and accelerates outgassing of the hydrogen load resulting in a loss of fiber photosensitivity. UV laser radiation is better suited to polymer ablation because material removal occurs largely via photolytic mechanisms. In particular, polymers are highly absorbing in the UV and UV photons often have sufficient energy to directly break molecular bonds, resulting in less thermal loading of the substrate and cleaner ablation. The down side in this case is that the standard fibers used in the fabrication of fiber Bragg gratings are very photosensitive to the ArF and KrF excimer wavelengths of 193 and 248nm respectively. Pre-exposure to intense light at these wavelengths can result in bleaching of the fiber core making the fabrication of fiber Bragg gratings more difficult.

Laser ablation of fiber polymer jackets is best achieved using high peak power, UV laser sources with wavelengths between 270-300nm for which these fibers are only weakly photosensitive yet polymer absorption is high. One such source is the frequency doubled copper laser or UV-CVL that produces 255, 271 or 289nm pulsed output. In this paper we present results of a study demonstrating the efficacy of UV-CVLs as sources for both removing the polymer jacket and writing fiber Bragg gratings. Moreover, both steps are achieved without extra handling of the fiber.

2. Experiment

The copper vapour laser used in these studies was nominally a 20W device (tube diameter 25mm, length 1m) operating at a pulse rate of 10kHz. The output beam was reduced in size to ~4mm using a mirror telescope and a 1mm diameter aperture placed at the focus of telescope removed the low beam quality amplified spontaneous emission. The remaining 10W of HBQ (~6W of green and ~4W of yellow) was line focussed (using an f=63mm cylindrical lens) into a non-linear crystal, a technique that has been shown to reduce the chance of crystal damage and minimise the effects of walk-off [15]. The non-linear crystal was beta barium borate (BBO) with dimensions 6mm × 4mm × 8mm and was cut for type 1 second harmonic generation (θ_pm = 47°, ϕ=90°). For these experiments we varied the tuning angle to frequency double either the 511 or 578nm fundamental, producing 255 and 289nm light respectively; note however, that two non-linear crystals could be used to generate 255 and 289nm output simultaneously. The UV output was separated from the fundamental using a Pellin Broca prism. Output powers of ~500mW were available at both UV wavelengths.

Nufern photosensitive fiber (GF1) was used throughout these experiments. The fiber was also hydrogenated to increase its photosensitivity. The fiber was fixed in position on top of a y-z-tilt positioning stage, which in turn was mounted upon a motorised translation stage (Physik Instrumente) allowing computer controlled movement in the x direction (parallel to the fiber axis/perpendicular to the beam path). During polymer jacket removal the UV beam was focussed to a spot ~10um in diameter using a 10x microscope objective. Alignment of the focal spot with the centre of the fiber was facilitated by an on-line imaging system. Fiber Bragg gratings were written using standard phase mask techniques which in our case included a cylindrical lens (f=75mm) to focus the light onto the fiber via the phase mask (aligned parallel to and approximately 0.5mm in front of the fiber). Note that the phase mask used in these experiments was optimised for a frequency doubled argon ion laser (244nm) and had a significant zero order component, effects which contribute to an overall reduction in the peak rejection levels of the gratings written with this system. Both the cylindrical lens/phase mask pair used during grating writing, and the microscope objective used during jacket removal were mounted upon standard kinematic plates, which allowed for easy interchange between the two processes. The fiber Bragg gratings written with this system were monitored using a laboratory built broadband light source and an Anritsu Optical Spectrum Analyser (MS9710c).

3. Results and Discussion

In order to gauge the response of photosensitive fiber to 289nm radiation we first investigated the growth rates of fiber Bragg gratings written with both 255nm and 289nm light, as shown in Figure 1. Note that in this experiment the polymer jackets were removed using mechanically strippers. Other than the differing wavelengths each grating was written under identical conditions, namely, exposure of a 3-4mm length of fiber to 50mW of UV focussed onto the fiber with the cylindrical lens (fluence = 60W/cm²). Figure 2 shows the corresponding shift in the Bragg wavelength with time for these two gratings. The grating written with 255nm light took ~5 minutes to grow to a maximum depth of 11dB after which the grating started to erase [16,17]. The Bragg wavelength shifted by ~2 nm during this time.

 

Figs. 1 and 2. Comparison of the growth dynamics and spectral shift of Bragg gratings written with 255 and 289nm laser light. Note that these fibers were prepared using mechanical stripping.

Download Full Size | PDF

By comparison, even after 20 minutes exposure the grating written with 289nm light still measured <2dB and showed no signs of erasure. Furthermore, the Bragg wavelength shifted by <1nm during this period. These results highlight the low photosensitivity of this fiber to 289nm light. In particular, the average refractive index increased by 2.4 × 10^-3 after 5 minutes exposure to 255nm light, and by only 4.8 × 10^-4 after 20 minutes exposure to 289nm light. In effect the photosensitivity of this fiber to 289nm light is a factor of 20 smaller than that to 255nm light at the same power densities.

Detailed characterisation of polymer jacket ablation at 289nm was undertaken to identify the best set of operating parameters for coating removal. Briefly, it was found that damage to the fiber occurred when output powers >200mW were used while slow translation speeds resulted in additional heat deposition within the polymer jacket and poor edge definition. The best results were achieved using output powers from 100 to 150mW, and several translations at the fastest speed possible from our motorised stages (ie. 12.5mm/s) rather than a single pass at a slow translation speed [18]. Figure 3 shows a scanning electron micrograph of a 20μm wide transverse scribe machined at an output power of 150mW, and 9 translations at a speed of 12.5mm/s. The edge definition of the scribe wall is high pointing to the photolytic nature of the material removal under these conditions. Indeed, upon close inspection the junction between the inner and outer acrylate coatings used in this type of fiber can be seen. Most importantly, the surface of the fiber cladding is fully exposed and free of surface contaminants and tensile strength tests of fibers machined under these conditions showed them to have comparable strength to pristine fiber. The final point to make is that although the fiber core is exposed to very high UV power densities (eg ~200kW/cm²) during the stripping process the total amount of time any particular point is irradiated by UV light is only 1/100 of a second. This equates to a total energy per unit area of 2kJ/cm², well below the value of 75kJ/cm² the fiber experienced during 20 minutes of exposure to 289nm light in the photosensitivity experiment discussed above.

 

Fig. 3. A 20 μm wide transverse scribe machined into the polymer jacket of an optical fiber using 289nm laser radiation.

Download Full Size | PDF

 

Fig. 4. A longitudinal “window” machined into the polymer jacket prior to writing a fiber Bragg grating.

Download Full Size | PDF

Figure 4 shows a “window” machined through the polymer jacket and parallel to the fiber axis using 150mW of 289nm laser light, and 10 translations at a speed of 12.5mm/s. The ripple seen in the walls of this scribe is the result of chatter in the motorised positioning stages. The transmission spectrum for a 20dB fiber Bragg grating written in 3 minutes through one of these windows in the polymer jacket is shown in Figure 5. Although it is possible to engineer a system that rotates the laser beam around the fiber and completely removes the polymer jacket [19,20] for simplicity’s sake we chose to only remove small longitudinal sections (60μm wide and 25mm long) of the polymer jacket via laser ablation. This did however compound the difficulties associated with aligning the line focus with the fiber core and placed additional constraints on the operating parameters for writing fiber Bragg gratings. For example, tight focussing geometries had to be used to fit the UV beam within the boundaries set by the “window” machined in the side of the fiber jacket, and the output power from the UV-CVL had to be reduced to 50mW to avoid ablating the remaining acrylate coating and damaging the phase mask. Despite these constraints, the gratings written in this fashion were similar to those written using the same 255nm source and mechanically stripped fiber in conjunction with conventional focussing geometries and standard UV output powers [5]. In particular, we saw no evidence that the fiber photosensitivity had been lessened when removing the polymer jacket with 289nm laser light. Indeed, the grating shown in Figure 5 grew at a faster rate and to a greater depth that those gratings written with this system after mechanical or chemical stripping. However, further studies are required to determine if the faster growth rates in this case are due to increased photosensitivity, a cleaner fiber surface, a reduction in the stresses induced in the fiber during jacket removal, or some other cause.

 

Fig. 5. Fibre Bragg grating written with 255nm laser radiation after stripping of the polymer jacket using 289nm light.

Download Full Size | PDF

4. Summary

We have demonstrated that the 289nm and 255nm outputs from a frequency doubled copper laser can be used to both strip the outer polymer jacket and write fiber Bragg gratings respectively. Tensile strength tests showed that fibers stripped in this fashion have comparable strength to pristine fiber. The fibers used in study are only weakly photosensitive to 289nm light, thus, strong gratings could still be fabricated after the stripping process.