UV transparent, on-fiber elastomeric phase-masks by casting directly on holographic glass phase-mask plates are demonstrated for the first time. These translation insensitive masks have been used to fabricate fiber Bragg gratings by simple exposure to 262 nm UV radiation. The stretchable phase-masks allow the Bragg wavelength of the inscribed grating to be changed easily. Multiple Bragg gratings at widely different wavelengths have been inscribed in one fiber using the same phase-mask. Inexpensive and disposable, the phase-masks allow the inscription of gratings in planar waveguides and fibers without causing contact-damage to either, and show the possibility of creating widely tuneable 1-D band-gap structures. In-situ fine tuning of the desired Bragg wavelength or correcting for errors in the phase-mask during inscription are other possible uses of this technology.
© 2005 Optical Society of America
Fiber Bragg gratings are usually written behind a UV transparent fused silica phase-mask with an interferometer . These phase-masks are usually made by an expensive process such as e-beam writing or holographic exposure. Phase-masks are also difficult to replicate without repeating the entire exposure methods and/or processing with the e-beam. When using it close to an optical fiber or planar waveguide, there is a finite probability of causing damage to the waveguide or contaminating the phase-mask. For behind the phase-mask writing, it is very difficult to tune the Bragg wavelength of the inscribed grating, without using a complex phase-mask design (eg chirped); an alternative is to use a stock of expensive phase-mask, each at a different wavelength.
Recent developments in the writing of Bragg gratings through UV transparent polymer coatings  has allowed the fabrication of Bragg gratings by the use of near UV radiation . Use of rigid polymer phase-masks for writing quality Bragg gratings has demonstrated that optical properties of some polymers yield excellent results [5–7]. This paper demonstrates a technique for fabricating elastomeric phase-masks that are transparent to 262 nm UV radiation, by casting. No damage is caused to the fiber or waveguide when the soft phase-masks are placed in direct contact with each other. We also report elastomeric on-fiber phase-masks (EOFPM) that are an integral part of the coating. These, then are simply exposed to UV radiation to inscribe a Bragg grating in the fiber core. As these phase-masks are highly stretchable, they can be tuned over a wide range of wavelengths, either directly in front of the waveguides or in an interferometer. This novel EOFPM remains in contact with the fiber as a permanent attachment, like its primary coating.
2. Fabrication of EOFPM
The polymer used is a two-part room temperature vulcanizing silicone rubber encapsulant. Transmission spectrum measurements of this polymer show it to be significantly transparent from 240nm upwards. A standard UV-transparent phase-mask made by holographic exposure with a period of 1.09 µm was used as a template for our polymer phase-mask. The polymer is cured for 24 hours at room-temperature so that it is as flat and homogeneous as possible. The curing process is finished by placing the polymer for 10 minutes under a standard infrared lamp, while it is still on the glass phase-mask. During the curing process, it is important for mask quality to keep the polymer, in a clean room so that dust deposits do not distort the surface of the polymer. As the elastomer does not stick to silica, the polymer phase-mask is peeled from the surface, ready for use. Dozens of casts were made from a single phase-mask without damaging or leaving residues on the silica phase-mask. Figure 1 shows an SEM photograph of the phase-mask template in the polymer.
A photosensitive optical fiber [Stocker Yale PS-1550-Y3 non-hydrogenated, NA=0.17, GeO2 conc. of 8 mole%, cut-off wavelength=1360 nm], was used in all experiments with the EOFPM except when specified otherwise. The phase-mask has a period of 1.09 µm. The fiber is placed in contact with the phase-mask ensuring that the lay is perpendicular to the periodic structure before casting. After curing, the polymer is cut around the fiber using a sharp hot blade, as shown in Fig. 2, so that the phase-mask becomes the coating for the fiber.
3. Properties of the elastomer
The primary interest in choosing an elastomeric polymer is for its capability of stretching. With an elastomeric phase-mask, the Bragg wavelength of gratings written in a photosensitive fiber can be tuned up to 160% (maximum elongation). Considering normal incidence and interference of the first orders through the phase-mask, the relation between the phase-mask period and the Bragg wavelength is :
For a phase-mask with a period of 1.09 µm and a refractive index of 1.465, we obtain a Bragg wavelength of 1596 nm. For an elongation of 2.5% (approximately 1.2 mm for a 5 cm long phase-mask), the Bragg wavelength will shift by 40 nm. Casting from a silica phase-mask with a period of 1.045 µm, gratings with a Bragg wavelength anywhere in the telecommunication C and L bands can be written.
Another important property of the elastomer is UV-transparency. The transmission and reflection spectra were measured using a Perkin-Elmer 2-beam spectrophotometer from 200 nm to 1600 nm. A restricted wavelength transmission plot can be seen in Fig. 3, for a 3 mm thick sample. Nonetheless, absorption in the UV is greater than in the visible range, but it is still acceptable, as the thickness of the polymer can be less than 0.5 mm. The transmission at 240 nm is around 60%, increasing to 67% at 262 nm. During the writing process, we used an incident UV power of 305 mW and the 0.5 mm thick phase-mask transmitted 268 mW to the fiber, which is more than enough for grating fabrication.
3. Bragg Gratings Written with Elastomeric On-Fiber Phase-Masks
Two configurations were investigated for writing Bragg gratings with the EOFPM. The first configuration was with the fiber embedded in the polymer phase-mask. The phase-mask was placed on a silica slide and precisely aligned horizontally, so that during the scanning process the intensity at the fiber core remains constant. The second one was with the polymer phase-mask simply stuck on the surface with the relief-grating facing the fiber. No alignment was required because the fiber was clamped by the vacuum chuck and the phase-mask stuck on the fiber afterwards.
The exposure was done by directly focusing a laser beam from a quadrupled YLF laser operating at 262 nm, with an average power of 305 mW. We used static exposure and also scanned the beam across the fiber. On exposure to UV light, a blue emission is seen from the core, due to interaction between UV light and the Ge-doped core of the fiber .
The polymer phase-mask may be damaged by the incident beam if the UV scanning speed is too slow, or in the case of a static exposure, if exposure time is too long. If the power of the beam is too high (>350 mW) it may also burn the polymer. This can be easily adjusted to obtain optimal gratings writing conditions without damaging the mask. We were able to write 7 gratings with the same polymer phase-mask, at a scanning speed of 0.2 mm/s, and no damage was observed on the surface of the mask. In the second configuration (fiber inside the phase-mask and stuck on a silica slide), exposure time was even more limited as slight heating of the glass-slide had a detrimental effect on the inscription quality, even before damage was observed on the polymer phase-mask. When damage occurs, the grating rapidly degrades as the fiber is exposed only to the zeroth order of the phase-mask. The grating inscribed in the fiber bleached rapidly after damage was sustained.
4. Experimental results
4.1 Fiber inside the polymer phase-mask
This is the first configuration as explained in section 3. We compared the results of gratings written by static exposure and by scanning beam. Figure 4(a) shows the reflection spectrum of the set of four fiber gratings written at the same wavelength with static exposure to 305 mW of 262 nm radiation. Figure 4(b) shows the reflection spectrum of an 8 mm long grating written using a scanning beam at a speed of 0.10 mm/s and average power of 305 mW for 262 nm radiation.
The reflection spectrum as seen in Fig. 4(a) developed from a single curve with bulk structure, into one with fine interference detail from the multiple Fabry-Perot filters of the four in-line Bragg gratings. We note that the gratings were reasonably well overlapped in their spectra, although they were written with an elastomeric phase-mask. The peak reflection is around 1596 nm, and the peak reflectivity at the Bragg wavelength is ~0.35%. This low reflectivity is a result of the intense zeroth-order transmitted and the low photosensitivity of the fiber. The exposure time was also limited to 1 min. to prevent damage to the phase-mask.
Again with the in-phase-mask fiber, Fig. 4(b) shows the quality of the back-reflection spectra. This grating written by a focussed beam scanned over 8 mm is better resolved, and indicates a uniform exposure along the fiber. The Bragg wavelength is at 1581.2 nm and the bandwidth is ~0.2 nm (FWHM). The wavelength of the grating in this fiber is at a much shorter wavelength as a result of the in-phase-mask fiber fabrication process. The fiber was pre-stretched before casting the phase-mask and its release resulted in a shorter Bragg wavelength. To obtain a Bragg wavelength shift of 15 nm, we only need a compression of 0.9% or a compression of 400 µm over the phase-mask length. This shows the versatility of this technique. In this configuration, the fiber strength limits the wavelength shift one can achieve, but this is not the case in the other configuration we will discuss next.
4.2 Polymer phase mask stuck on the fiber
With the phase-mask stuck on the fiber (sticky phase-mask), the vacuum chucks hold the fiber horizontally. To have perfect alignment we only have to make sure the grating is perpendicular to the fiber. We chose to expose the fiber by scanning the beam along it, because it eliminates the probability of damaging the mask. We also obtain a better resolution and reflection quality than with static exposure. In Fig. 5, we show the reflection and transmission spectra of a hydrogenated SMF28 fiber using a compressed sticky phase-mask. As we had limited access to hydrogenation facilities, the SMF28 fiber was loaded elsewhere at 200bar for 2 weeks at room temperature. However, the fiber had unfortunately been left in a warm laboratory for two days before using with our phase-masks: hence the low photosensitivity of the fiber. We did, however, want to show that SMF28 fiber can also be used as the writing times are much longer compared to other types of fibers.
If we compare with the fiber inside the phase-mask configuration, we see that the quality of the spectra in Fig. 5 is poorer than in Fig. 4(b). We observe this because it is harder to make the mask stick equally on the fiber whilst keeping the tension or compression uniformly along the length of the fiber. Also, the grating must be as close to the fiber as possible to get a high quality interference pattern; but since the polymer is very soft when the sticky phase mask is stuck on the fiber, it is slightly deformed.
Figure 6 shows three reflection peaks written using a sticky phase-mask. The peaks at 1578, 1592 and 1619 nm were respectively written in a slightly compressed, relaxed and elongated sticky phase-mask. The full-width half-maximum intensity bandwidths of the peaks are 1.1, 2.3 and 3.3 nm respectively. The reflected intensity is very weak mainly due to the limited Ge-doped fiber photosensitivity and the high zero-order that saturates the inscription and limits the visibility of the fringes.
If we compare these results with the results obtained for the beam scanned across the in-phase-mask fiber, we notice the intensity of reflection is weaker. This is because in the latter case, the light passes through the polymer before being diffracted by the grating. Since the polymer was cured without a back-plate, it is not completely flat and this induces distortions in the laser beam, as well as de-focussing of the beam, reducing the overlap of the first orders.
4.3 Quality of the gratings
The quality of the Bragg gratings depend on several factors such as uniformity of the mix of the polymer, presence of bubbles, surface quality and thickness. The first two issues can be addressed by making a larger quantity of polymer than required to ensure an accurate mix and uniformity of the two components and by the proper removal of bubbles by standard evacuation process. The back surface can be made perfect by using an optical flat on the rear face during the casting process. The thickness of the polymer influences the amount of power absorbed and can also be minimized. We therefore believe that very high quality gratings are possible using the above steps with an appropriate zero-order suppressed phase-mask, corrected for the refractive index of the polymer.
The measured first order diffraction efficiency of the master silica phase-mask was I1/Iin=32.6% with a zero-order of 12.6% (as the phase-mask had been optimized for 244nm). On the other hand the polymer phase-mask had a first-order diffraction efficiency of only 15.6% and a zero-order of 39.4%. The transfer of power from the first to the zero-order is mainly due to the difference of refractive index between the materials. We are awaiting a phase-mask with the appropriate parameters to reduce the zero-order to match the silica phase-mask, the results of which will be reported in the future. However, we believe that the quality of Bragg gratings written with the polymer phase mask can be of comparable quality to those written using a silica phase mask.
The main limitations in the use of elastomer phase-masks are based on practical considerations rather than fundamental difficulties. Keeping the surfaces of the elastomer very clean is very important. We believe that the fabrication of the phase-mask can be much improved by using a mould optimized for zero-order suppression and a hydrogenated photosensitive optical fiber would certainly improve the strength of the grating. We believe that the quality of the gratings can be comparable to the ones produced by silica phase-masks.
We have demonstrated a simple technique to fabricate elastomeric phase-masks that are period-tuneable and UV transparent for the inscription of Bragg gratings. We fabricated Bragg gratings by directly exposing the phase-mask to 262 nm wavelength radiation. These phase-masks are robust and easily cloned using the first cast polymer impression as a template.
With the fiber in-phase-mask configuration, we obtained better quality and resolved peaks because the grating was very close to the fiber core and light did not pass through the polymer phase-mask before being diffracted. The difficulty of tuning the grating, which is of interest in this technique, led us to investigate the simplicity of the sticky phase-mask, which also gave good results. We note that it is very easy to write blazed gratings using on-fiber phase-masks by simple in plane-rotation of the mask.
We believe that this technique will open new possibilities in Bragg gratings fabrication of chirped and other novel filters by local deformation of the phase-mask. Once a phase-mask has been incorporated as a coating on the optical fibre, it can be protected in a short tube housing as a “dark” or unwritten grating. At a later date, the tube may be removed and the fibre exposed to any UV radiation with sufficient power,, without the need for an external interferometer for the grating to be written. This concept of “dark gratings” will allow gratings to be fabricated without stripping the fiber at some time in the future.
This research was partially supported by the Canadian Institute for Photonics Innovation and by the Canada Research Chairs Program of Natural Science and Engineering Research of Research Council of Canada. The authors gratefully acknowledge Paul Lefebvre and Claude Beaulieu of LxSix for the providing the hydrogenated SMF28 fiber.
References and links
1. R. Kashyap, Fiber Bragg Gratings (Academic Press, 1999), pp. 57–74.
2. R P Espinola, R M Atkins, N P Wang, D A Simoff, M A Paczkowski, R S Windeler, D L Brownlow, D S Shenk, P A Glodis, T A Strasser, J J DeMarco, and P J Chandonnet, “40dB fiber Bragg grating written through the fiber coating at 257nm,” In Bragg Gratings, Photosensitivity, and Poling in Glass Fibers and Waveguides: Applications and Fundamentals, Vol. 17 of OSA Technical Digest Series (Optical Society of America, Washington DC, 1997), Postdeadline paper PD2.
3. D S Starodubov, V Grubsky, J Feinberg, D Dianov, S L Semjonov, A N Guryanov, and N N Vechkanov, “Fiber Bragg gratings with reflectivity >97% fabricated through polymer jacket using near-UV radiation,” In Bragg Gratings, Photosensitivity, and Poling in Glass Fibers and Waveguides: Applications and Fundamentals, Vol. 17 of OSA Technical Digest Series (Optical Society of America, Washington DC, 1997), Postdeadline paper PD1.
4. Y Duval, R Kashyap, S Fleming, and F Ouellette, “Correlation between uv-induced refractive index and photoluminescence in Ge-doped fibre,” Appl. Phys. Lett.61 (25), 2955–2957, 21 December 1992. [CrossRef]
5. D. S. Starodubov and J. Feinberg, US patents 5,745,617; 5,881,188.
6. D. S. Starodubov, V. Grubsky, and J. Feinberg, “Ultrastrong Fiber Gratings and Their Applications,” In SPIE Conference Photonics East: Optical Fiber Reliability and Testing, Vol. 3848-26, 20 September 1999.
7. D. S. Starodubov, V. Grubsky, and J. Feinberg, “Optical Fiber Gratings: New Fabrication Techniques and Components,” In Nanostructures and Quantum Dots, Digest of the LEOS Summer Topical Meetings, (San Diego, 1999).