Vacuum-ultraviolet laser-induced refractive-index ... - IEEE Xplore

JOURNAL OF LIGHTWAVE TECHNOLOGY, VOL. 21, NO. 9, SEPTEMBER 2003

1969

Vacuum-Ultraviolet Laser-Induced Refractive-Index Change and Birefringence in Standard Optical Fibers Kevin P. Chen, Peter R. Herman, Fellow, OSA, Rod Taylor, and Cyril Hnatovsky

Abstract—Strong photosensitivity responses with the use of above-bandgap 157-nm F2 -laser radiation are reported for standard germanosilicate fiber. A large 1.3 10 3 effective index change at 1.55 m was inferred by trimming strong and weak Bragg gratings in hydrogen-free fiber. The F2 -laser fluence-pro50 mJ cm2 is much lower than with cessing window of traditional ultraviolet (UV) lasers. For hydrogen-soaked fiber, highly asymmetric refractive-index profiles were noted by atomic force microscopy and microreflection microscopy, yielding a peak index change of 0 01 across a small 1- m penetration depth at the fiber core. The index asymmetry appears to underlie the large 5 10 5 value of laser-induced birefringence. Index Terms—Atomic force microscope (AFM), birefringence, Bragg grating, F2 excimer laser, laser, microreflectivity, optical fiber, refractive-index profile.

I. INTRODUCTION

T

HE discovery of photosensitivity in germanosilicate glass fibers [1] has enabled the manufacture of high quality, low-cost, and ultralow loss in-fiber photonic devices, supporting today’s fiber optics telecommunication industry. Fiber Bragg grating (FBG) devices serve essential functions such as wavelength multiplexers [2], dispersion compensators [3], gain equalizers [4], and mode converters [5]. Because of an intrinsically weak photosensitivity response, enhancement procedures, such as hydrogen loading [6], are typically applied to speed grating writing times and increase the available refractive-index change in standard telecommunication fiber. can be With hydrogen enhancement, index changes of 5 induced with modest doses of 248-nm KrF-laser radiation [6]. However, hydrogen enhancement brings several drawbacks, such as short shelf life, 1390-nm OH-band transmission losses [7], and aging-related relaxation [8]. Alternative enhancement mechanisms are also available in specialty fibers doped with high concentrations of germanium, boron, and phosphor [9]. However, modal mismatch with standard optical fiber may increase insertion losses. Another approach to improving the fiber photosensitivity response is extending light sources to shorter wavelength. Significant enhancements are available by moving systematically from 488-nm Ar-ion laser light to light from the 248-nm KrF laser, the Manuscript received April 29, 2003. This work was supported by the Natural Science and Engineering Research Council of Canada and the Canadian Institute for Photonics Innovation. K. P. Chen and P. R. Herman are with the Department of Electrical and Computer Engineering, University of Toronto, Toronto, ON M5S-3G4, Canada (e-mail: [email protected]). R. Taylor and C. Hnatovsky are with the Institute for Microstructural Sciences, National Research Council of Canada, Ottawa, ON K1A 0R6, Canada. Digital Object Identifier 10.1109/JLT.2003.815647

193-nm ArF laser [10], and most recently, the 157-nm F laser was reported [11]–[14]. An refractive-index change of [10] for standard telecommunication fiber without any enhancement when applying 193-nm radiation. A two-photon process was inferred [15] to excite direct bandgap transitions, in contrast to the bleaching of inter-bandgap defects for the 248-nm 30 kJ cm at laser case. Because large 193-nm doses of single-pulse fluence of 650 to 1300 mJ cm per pulse were applied, thermal stresses can be induced that lead to fiber damage, fiber brittleness, and excess optical loss. Strong photosensitivity responses were also noted for record short-wavelength 157-nm F -laser radiation [11], [12] by the rapid formation of strong long period gratings in standard fibers. This paper presents a comprehensive extension of this F -laser work, reporting on the 157-nm photosensitivity responses of standard telecommunication fiber with and without the use of hydrogen enhancement. Unlike traditional ultraviolet lasers, the 7.9-eV photon from the F laser directly excites electrons across the 7.1-eV bandgap of low concentration GeO -doped (5%) silica glass [16] to initiate strong single-photon photosensitivity processes. Effective were readily induced in refractive-index changes of untreated fiber as noted by trimming FBG. The strong 157-nm absorption in the core is further responsible for generating a highly asymmetric refractive-index profile, as observed by atomic force microscopy (AFM) and microreflection microscopy, that sharply contrasts the uniform refractive-index profiles noted for KrF- and ArF-laser irradiation, even under hydrogen soaking conditions [17], [18]. A peak refractive-index is reported. Such asymmetry appears change of . to underlie an unusually strong birefringence of 5 The F -laser therefore presents a unique opportunity for fabricating birefringence mode converters [5], rocking filters [19], and other polarization-controlling devices in fibers. The characterization of 157-nm photosensitivity responses, the refractive-index profiles, and the laser-induced birefringence are described in Sections II–IV, respectively, followed by a general discussion of the results in Section V. II. 157-nm PHOTOSENSITIVITY IN STANDARD FIBERS Photosensitivity responses in standard telecommunication fiber (Corning SMF-28) were determined by trimming FBGs with uniform 157-nm light. Fibers were hydrogen-soaked at 1900 psi for 10 d at room temperature, and then inscribed with – ) and strong both weak (index contrast – ) FBGs by using a (saturated index change 248-nm KrF laser and a phase mask. The FBG samples were

0733-8724/03$17.00 © 2003 IEEE

1970

trimmed by the 157-nm F laser immediately after the FBG inscription for the hydrogen-soaked case, or baked at 150 C for 5 d prior to the 157-nm irradiation to provide hydrogen-free samples. A commercial F laser (Lambda Physik LPG200I) provided unpolarized 157-nm radiation at up to 100-Hz repetition rate. A CaF cylindrical lens adjusted the on-fiber fluence from 4 to 150 mJ cm . All optical beam paths were flushed with nitrogen or argon gas to provide a transparent beam path for the vacuum ultraviolet (VUV) radiation. The transmission spectra of the FBGs were monitored in-situ during laser irradiation by a broadband light source (Thorlab ASE7001) and an optical spectrum analyzer (Ando 6317B). ac and dc refractive-index changes were inferred from the changes in FBG transmittance and spectral peak wavelength, respectively. Spectral recordings were delayed by 20 min following the F laser irradiation to eliminate laser-heating effects. Fig. 1 shows the dc [Fig. 1(a) and (c)] and ac [Fig. 1 (b) and (d)] refractive-index changes induced in hydrogen-free fiber, for the weak [Fig. 1(a) and (b)] and strong [Fig. 1(c) and (d)] FBG cases. Three single-pulse fluence values of 5.3, 61, and 136 mJ cm were applied for the weak grating case [Fig. 1(a)], yielding nearly , saturated effective refractive-index changes of 1.3 , and 0.45 , respectively, following a total 0.75 radiation dose of 25 kJ cm . All reported index changes are large for a hydrogen-free telecommunication fiber. The photosensitivity is weaker at high single-pulse fluence (61 and 136 mJ cm ), possibly owing to an accumulation of laser damage inside the fiber and opacity generation in the cladding. Although a weak 157-nm cladding absorption of 77.5 cm was reported by Dyer et al., [20] fiber damage becomes visible under an optical microscope after prolonged irradiation at a fluence mJ cm per pulse, which is twofold smaller than the of 300-mJ cm ablation threshold of germanosilicate glass [21]. Such laser damage narrows the F -laser processing window to fluence values well below that typically applied (0.5 – 1 J cm ) with longer-wavelength lasers. A compaction phenomenon is also anticipated in the 157-nm irradiated fiber cladding, predicting a fractional volume com200 ppm for a 25–kJ cm dose [22]. paction of The associated refractive-index increase presumably lowers the index contrast between the germanosilicate core and the fused silica cladding, reducing the observable change in effective index. The relative roles of cladding compaction, core photosensitivity, and possible stress relief in the 157-nm treated fibers could not be delineated with the present instrumentation. The 248-nm FBG formation step was found to pre-sensitize the fiber core, increasing the post 157-nm photosensitivity response as seen in Fig. 2. The grating reflectance increased sixfold, from 2.15 to 13 dB, following a 3.5-kJ cm dose of F -laser light at 5.3-mJ cm fluence per pulse. Corresponding to this increase is a fourfold change in ac refractive index from base created by the KrF laser exposure to the 3.8 as seen in Fig. 1(b). Fig. 1(b) also shows that the amplification effect is twofold weaker for the case of 136-mJ cm single-pulse fluence. Laser amplification effects have been previously reported but are generally much weaker with longer wavelength radiation. [23] The evolution of the ac index con-


Fig. 1. F -laser photosensitivity responses in hydrogen-free SMF-28 fiber inferred from trimming FBGs. dc index changes are shown in (a) and (c) as a function of accumulated fluence for weak and strong FBG cases, respectively. The ac index modulation is shown in (b) and (d) for weak and strong FBG cases, respectively. Horizontal dashed lines indicate the ac index modulation induced by the KrF laser prior to the F -laser dose. Single-pulse fluence values are indicated for each curve.

CHEN et al.: VUV LASER-INDUCED REFRACTIVE-INDEX CHANGE AND BIREFRINGENCE IN STANDARD OPTICAL FIBERS

Fig. 2. Transmission spectrum of a weak 2.4-dB FBG formed by a 248-nm laser before (inset) and following (main) a uniform 3.5-kJ=cm postexposure dose by the 157-nm F laser. The grating reflection was amplified to 14.3 dB.

trast in Fig. 1(b) can be well represented (solid curve) as a function of 157-nm fluence (in units of J cm ) by the expression

(1) for the case of 5.3-mJ cm single-pulse fluence. The two fluence-dependent factors possibly relate to a saturation of the ac photosensitization phenomenon and washing out of the ac modulation by the uniform 157-nm irradiation. The sensitization effect of 248-nm preirradiation is no longer evident for the case of strong FBGs as shown in Fig. 1(d). The gratings become weaker with F -laser exposure, and the ac index modulation follows a simple exponential decay of

1971

Fig. 1(c) (strong FBG) with Fig. 1(a) (weak FBG). The 248-nm irradiation of hydrogen-soaked fiber possibly contributes new precursor states that remain photosensitive to 157-nm light even after thermal annealing. This 248-nm photosensitization is further exemplified by the amplification of the ac index change in Fig. 1(b). The 157-nm photosensitivity responses for hydrogen-loaded fiber are summarized in Fig. 3. For the weak FBG case [Fig. 3(a)], a large and unsaturated index change of 2 was induced with a modest 1-kJ cm dose of 157-nm laser light when applied at 5-mJ cm fluence. This index value is twofold larger than observed in the hydrogen-free fiber [Fig. 1(a)], and developed more than ten times faster with accumulated fluence. The strong hydrogen enhancement is even more pronounced at lower accumulated fluence, becoming fiftyfold faster, for index change. The 157-nm photoexample, to reach a 1 sensitivity response is approximately one-third weaker for the higher single-pulse fluence of 88-mJ cm in Fig. 3(a), showing that high-fluence laser damage also arises in hydrogen-soaked fiber. However, the 248-nm photosensitization effect observed in the hydrogen-free fibers [Fig. 1(b) and (c)] was not evident here. The 157-nm photosensitivity response dropped by 50% for the strong FBG case of Fig. 3(c) in contrast to the twofold enhancement observed in the hydrogen-free fiber in Fig. 1(c) relative to Fig. 1(a). The 248-nm photosensitivity enhancement was also not evident in the ac index changes shown in Fig. 3(b) and (d). One additional contrast is the similarity of refractive-index changes in Fig. 1(c) for the dissimilar single-pulse fluence values of 5 and 58 mJ cm . The combination of large 248-nm exposure and hydrogen soaking may create a high density of photosensitive precursor states that sharply reduce the 157-nm penetration depth into the fiber core, creating only a narrow band index modification that moderates the observed change in effective refractive index. An examination of laser-induced refractive-index profiles in the next section helps elucidate the cause of these contradicting responses. III. INDEX CHANGE PROFILES

for

mJ cm

(2a)

for

mJ cm

(2b)

The ac refractive-index fall-off is related to the saturation and washing out of a strong grating index modulation formed by high 248-nm exposure. Nevertheless, the F -laser radiation proin the effective index as vided a strong change of 1.2 noted in Fig. 1(c) for single pulse fluence of 5.0 mJ cm . This saturated dc index change is identical with the value obtained in the weak FBG case of Fig. 1(a) for the similar condition of 5.3-mJ cm exposure, even though the 248-nm FBG formation in step consumed a large saturated index change of 6 – 8 the strong FBG case. The strong 248-nm pre-irradiation therefore does not appear to consume the photosensitivity capacity of the 157-nm process, suggesting a different photosensitivity channel for the above bandgap radiation. Moreover, a twofold acceleration of the 157-nm photosensitivity response is noted in the strong grating case as seen by comparing the index change in

Underlying the above observations in effective refractive index are more subtle changes to the guiding properties such a waveguide birefringence or polarization-mode dispersion that depend on the index profiles modified by the deep UV laser radiation. In this section, a combination of AFM [17], [18] and microreflectivity methods are employed to study the index profiles of 157-nm irradiated fibers. A. Laser Modified Index Profiles: AFM and Chemical Etching Hydrogen-soaked and hydrogen-free SMF-28 fibers were uniformly irradiated with 157-nm light at up to 2200-J cm total dose at low single-pulse fluence of mJ cm . A reference H -loaded fiber was also radiated with 248-nm KrF-laser light. AFM cross-sectional scans were first performed on the clean-cleaved end facets of the irradiated fiber sections to examine for the presence of VUV-laser produced divots or complementary hill-like topography due to laser-induced brittleness. The topography was found to be flat over the entire examined cores and cladding regions for all of the radiated

1972

Fig. 3. F -laser photosensitivity responses in hydrogen-soaked SMF-28 fiber inferred from trimming FBGs. dc index changes are shown in (a) and (c) as a function of accumulated fluence for weak and strong FBG cases, respectively. The ac index modulation is shown in (b) and (d) for weak and strong FBG cases, respectively. Single-pulse fluence values are indicated for each curve.

samples. The end facets of these fibers were then etched in 1% (by volume) HF solution and subsequently examined by AFM.


The differential etching rates of the Ge-doped core and fused silica cladding were used to resolve the index of refraction profile. The known index contrast of 0.005 between the core and the cladding of SMF-28 fibers was used to calibrate the linear etch rate and subsequently relate the etch rate to the refractive-index change. The index change owing to VUV exposure was then determined from the accelerated etch rate in the core by assuming that the index change from the VUV exposure follows the same linear relationship between the etch rate and the index change as the un-exposed SMF-28. This assumption is validated in Sections III-B. The sensitivity of this etching technique permits laser-induced index of refraction differences of to be observed. This sensitivity could not be improved with longer etching times because deeper etches ( 100 nm) could not be accurately tracked by the AFM probe, especially in regions of greatest etch gradient near the core-cladding interface. Fig. 4 compares AFM topographies and cross-sectional scans of 248-nm and 157-nm irradiated fibers. The AFM results from a H -loaded fiber irradiated by 20 kJ cm of 300 mJ cm are KrF radiation at a single-pulse fluence of shown in Fig. 4(a). The 12-min etch in 1% HF solution reveals a uniform etch profile across the core region. A reliable value for the laser-induced index change was not obtainable since the (from Section II) barely anticipated index change of exceeds the 0.0005 sensitivity limit for this method. However, a uniform index profile is expected since few and weak cladding modes were observed in the FBG reflection spectra produced in Section II under similar illumination conditions. Further, other groups [17], [18] have noted that hydrogen loading does not significantly distort the symmetric index profile of standard telecommunication fibers. A uniform etch profile is also noted in Fig. 4(b) for a hydrogen-free fiber irradiated with 6 kJ cm of 157-nm light mJ cm . According to the at single-pulse fluence of data in Fig. 1(a), this exposure should provide a 7 change in effective refractive index, which again barely exceeds the 0.0005 sensitivity limit of the present acid-etching AFM technique. At this level of index change, any asymmetry in the index profile that may arise from strong above-bandgap absorption in the germanosilicate core should become observable by concentrating the refractive-index change to higher value on the radiated side of the fiber. The flat profile in Fig. 4(b) suggests a uniform index profile that contrasts with the results of hydrogen-free planar waveguide photosensitivity studies in which a penetration depth for the index change of 5 m was deduced using the same 157-nm laser dosage [14]. Co-dopants in the case of planar waveguides may reduce the 157-nm penetration depth. The uniform 157-nm fiber-core response [Fig. 4(b)] is dramatically altered with hydrogen loading as seen by the AFM image shown in Fig. 4(c). A total 157-nm dosage of 1.3 kJ cm with single-pulse fluence of 5 mJ cm generated a sharp preferential etching toward one side of the core after 12 mins in a 1.5 times above that HF bath. This accelerated etch rate is for the core plateau, implying a peak index change of . The fall-off of the VUV induced trench is 900 nm, indicating a very shallow VUV penetration depth in the hydrogen-loaded fiber core.


1973

Fig. 4. AFM surface topography (left) and cross-sectional analysis (right) of the core region of SMF-28 fiber uniformly irradiated under the following conditions: (a) 20-kJ=cm total fluence of 248-nm KrF-laser light at 300-mJ=cm single-pulse fluence on hydrogen-soaked fiber; (b) 6-kJ=cm total fluence of 157-nm F -laser light at 5-mJ=cm single-pulse fluence on hydrogen-free fiber; and (c) 2200-J=cm total fluence of 157-nm F -laser light at 5-mJ=cm single-pulse fluence on hydrogen-loaded fiber.

A systematic study of the peak index change and VUV penetration with increasing laser dose provided the data in Fig. 5. was generated within a A peak index change of narrow 0.7- m crescent with a 2.2-kJ cm dose. This is probably the largest index change induced in a standard fiber ever , can be well represented reported. The peak index change,

by a power law relationship of , where NF (in J cm ) is the accumulated fluence dose. This relationship supports a one-photon photosensitivity process, consistent with the strong direct-bandgap absorption anticipated in the gerpenetration depth reduces gradmanosilicate core [24]. The ually from 1.4 m to 0.7 m with increasing fluence dose up

1974


Fig. 5. Evolution of the peak index change (solid circle) and the 1=e penetration depth (open circle) of 157-nm radiated hydrogen-loaded SMF-28 fiber as inferred from AFM profiles of acid-etched fiber cores. The peak index F (solid curve). data is well represented by n

1 = 4 2 10

to 2200 J cm . This shrinking penetration is consistent with our 157-nm photosensitivity studies of planar waveguides [14], and is likely related to laser-generated precursors and laser-induced damage. Similar factors were cited in Section II to reduce the photosensitivity response when large single-pulse fluence was applied to the fiber. Because of the finite penetration depth in hydrogen-soaked fiber, the peak index values in Fig. 5 are six1.55 m as fold larger than the effective index change at measured by the FBG trimming method [Fig. 3(a)]. This combination of VUV irradiation and hydrogen loading provide strong anisotropy in profiling refractive-index changes in germanosilicate waveguides. To best of our knowledge, such a unique extreme anisotropic index profile induced by laser radiation has only been observed at this record-short wavelength of 157 nm [14]. B. Laser-Modified Index Profiles: Microreflectivity Microscopy To validate the AFM plus chemical etching technique as a means of measuring laser induced refractive-index profiles, microreflectivity measurements were also conducted on the sectioned fiber samples. Microreflectivity data obtained using a microscope 633 nm (He–Ne) laser together with a objective were collected from cleaved surfaces (i.e., without any etching ). The output end of the fibers was terminated in index matching epoxy to prevent any back reflected light from contributing to the front surface reflected signal. The optical resolution was estimated to be 500 nm. The microreflectivity data was then converted to refractive-index changes using the Fresnel reflection formula and using a cladding index of refraction of 1.46 as a reference. Fig. 6 shows the comparison of the index profile from both the AFM plus the chemical etching method and the microreflectivity measurement on a hydrogen-loaded SMF-28 fiber irradiated with 2.2 kJ cm of 157-nm accumulated dosage. The index profiles are very similar when one takes into account the poorer resolution of the microreflectivity measurements,

Fig. 6. Refraction index profiles at the core of F -laser irradiated hydrogen-loaded SMF-28 fiber obtained from a combination of HF chemical etching and AFM (solid circle), and from microreflectivity microscopy (open circle). The peak component measured with the AFM technique (9-m position) was normalized to index difference between the fiber core and ). The total F dosage was 2200 J=cm at 5-mJ=cm cladding (4.5 single-pulse fluence.

210

demonstrating the excellent agreement between the two techniques. The 157-nm laser induced index change obtained by of 0.005 normalizing the AFM data with a core–cladding , agreeing with the (SMF-28) from the plateau region is predicted value from the microreflectivity data. As expected, the peak index contrast of 0.005 – 0.017 at the 9- m position in Fig. 6 exceeds the corresponding change in effective index , as extrapolated from the data in Fig. 3(a) to of 2 – 3 the same accumulated fluence of 2.2 kJ cm . One discrepancy is noted for microreflectivity data in Fig. 6, which exhibited a raised index region instead of a plateau on the nonirradiated side of the core. Both techniques lacked the sensitivity required to reveal the presence of anisotropic index profiles in the hydrogen-loaded fibers when the 157-nm accumulated fluences J cm . were less than IV. BIREFRINGENCE IN BRAGG GRATINGS The highly anisotropic index profile generated by F -laser radiation introduces large birefringence into germanosilicate optical fibers, offering new opportunities for in-fiber photonics component fabrication. In this section, we choose the FBG as an example to study the modification of polarization characteristics under 157-nm laser radiation. Fig. 7 shows transmission spectra of a weak FBG along the fast and slow polarization axes following a uniform F -laser exposure of 10-kJ cm dose, at single-pulse fluence of 5 mJ cm in a hydrogen-free fiber. Since birefringence was not observed in the FBG with the KrF-laser exposure, a large spectral res0.05 nm is therefore fully attributable onance shift of to the F -laser radiation, yielding a waveguide birefringence of , where is the FBG resonance wavelength and is the effective index. This birefringence is index change, and is at least ten about 5% of the total 1.0 times larger than the birefringence values reported in standard fibers using a 240-nm laser [25]. The post-F -laser irradiation


1975

Fig. 7. Transmission spectra of a FBG along slow and fast axes in hydrogen-loaded SMF-28 fiber showing a 5 10 birefringence induced by a 10-kJ=cm postlaser dose of 157-nm radiation.

2

also enhances the index contrast of the FBG [i.e., Fig. 1(b)] to yield different reflectivities of 33% and 35% in Fig. 7 for the fast and slow axes, respectively. Polarization-mode dispersion (PMD) is an important factor to be minimized in telecommunication optical components. Fig. 8(a) shows PMD of 1– 3 ps in the resonant wings of the 1-cm long and 30-dB FBG written by the KrF laser in a hydrogen-loaded SMF-28 fiber. Fig. 8(b) shows a PMD spectrum of a similar FBG (1 cm, 30 dB) formed after the hy-kJ cm drogen-soaked SMF-28 fiber was pretreated with a dose (3.5 mJ cm single-pulse fluence) of F -laser radiation. The PMD of 40–60 ps is now twentyfold larger. However, such F -laser induced PMD can be reduced by two-order of magnitudes by radiating the fiber from opposite sides. V. DISCUSSION AND CONCLUSION A comprehensive study of photosensitivity responses with 157-nm F -laser radiation was undertaken for both hydrogen-free and hydrogen-loaded standard fibers by means of trimming FBGs. The above-bandgap 157-nm light interacted strongly with the germanosilicate fiber core, inducing large in both fiber types. changes in effective index of mJ cm could be Only weak single-pulse fluence of applied to the fiber to avoid damage and increased fiber absorption that prematurely saturated the index changes [Figs. 1(a) less than the and (b) and 3(a) and (b)]. This exposure is fluences typically applied with UV excimer lasers, and offered the additional benefit of only modest laser heating of the fiber to less than 20 C above ambient (at 100 Hz). Thermal damage was not evident, and fibers were not obviously weakened or overly brittle under such low pulse fluence exposure. Since strong 157-nm photosensitivity was demonstrated (Fig. 1) without the need for hydrogen enhancement, index changes are much more stable than that in hydrogen-loaded fibers. This has been confirmed by aging test in long-period grating written by 157-nm laser in hydrogen-free standard fiber (SMF-28). [12] Table I compares the 157-nm F -laser photosensitivity responses (effective index changes) of hydrogen-free and hydrogen-soaked standard fibers to that provided by the more

Fig. 8. Group velocity dispersion (GVD) difference measured between slow and fast axes of FBGs written by a 248-nm KrF laser in hydrogen-loaded SMF-28 fiber with (a) and without (b) a 1-kJ=cm predose of F -laser light at a single-pulse fluence of 2.7 mJ=cm .

common 248-nm KrF and 193-nm ArF lasers. In hydrogen-free fiber, the 193-nm and 157-nm photosensitivity response rates are approximately two orders of magnitude stronger than with the 248-nm KrF laser. A two-photon process was inferred [15] for the 193-nm photosensitivity response, which is approximately fourfold weaker than the single-photon response underlying the 157-nm photosensitivity. Table I also demonstrates the well know enhancement effect of hydrogen loading [6] that dramatically increases the 248-nm photosensitivity in both the size of the index change and the exposure time (accumulated fluence). Using identical hydrogen loading condition (105 atmosphere, 10 d, 20 C), a less is noted in the 157-nm significant enhancement factor of photosensitivity response over the hydrogen-free 157-nm case. The low-concentration germanosilicate fiber core interacts strongly with 157-nm radiation, presumably leaving less capacity for improving the response with hydrogen soaking. Nevertheless, the hydrogen-enhanced 157-nm responses are more than thirtyfold faster than the comparable 248-nm hydrogen-enhanced response. Further, hydrogen-soaked fibers concentrate stronger refractive-index changes into thin crescent-shaped cylinders [Fig. 4(c)] that have less influence on the effective index as seen by the large-area guiding mode. For similar reasons, photosensitivity enhancement effects are

1976


TABLE I COMPARISON OF PHOTOSENSITIVITY RESPONSES IN STANDARD TELECOMMUNICATION FIBER (3% GeO )

also less pronounced when 157-nm radiation is applied to high-concentration germanosilicate fibers. [11] A particularly unique feature of 157-nm photosensitivity is the strongly asymmetric index profile generated in both planar waveguides [14] and hydrogen-soaked fibers [Fig. 4(c)]. Such asymmetry can generate a large polarization-dependent Goos–Hanchen shift that may underlie the strong birefringence noted here. Such strong birefringence has not been observed in standard fibers, even with hydrogen loading, when radiated with longer wavelength radiation [17], [18]. However, a large UV-induced birefringence of was generated in germanium-rich optical fiber, possibly owing to preferential alignment of dipole moments of UV-induced defects along the polarization axis of the UV laser radiation [25]. For the present 157-nm case, the transverse component of the unpolarized incident laser radiation may also generate strong anisotropic index changes that underlie the strong birefringence and is a subject of further investigation. Large waveguide birefringence is strictly unacceptable in certain telecommunication applications such as FBG DWDM filters, but has strong merits for several component applications such as polarization mode converters, polarization trimming in fiber lasers, and birefringence trimming in planar lightwave circuits. The flexible control of waveguide birefringence and refractive-index profile is normally very challenging with traditional UV lasers and requires specialty fibers with high germanium concentration and the presence of other dopants to improve the UV-interaction with the fiber core. The above-bandgap F -laser radiation is therefore a natural and powerful tool for rapidly generating highly anisotropic refractive-index changes inside the core of standard optical fibers. The birefringence depicted in the FBG of Fig. 7 demonstrates that 157-nm radiation can convert standard symmetric fiber into polarization-maintaining fiber with beat length smaller than 1.5 cm—a value comparable with commercially available polarization-maintaining (PM) fiber. The F laser light enables new approaches in birefringence trimming and refractive-index profiling at key locations in optical modules and fiber optical networks without introducing significant insertion loss or without the need of specialty PM fibers. In conclusion, this paper demonstrated the potential application of the F laser in fiber components fabrication. A modest F -laser exposure of 1 – 10 kJ cm provided a large index change in standard fibers, while twofold larger index changes and fifteenfold faster responses were found with hydrogen soaking. The 157-nm fluence processing window of

mJ cm that avoided damage in standard germanosilicate fiber was much lower than with traditional UV-laser photosensitivity because of the strong above-bandgap interactions of the 157-nm light. The F -laser modified index profiles were studied by a combination of selective chemical etching and AFM in conjunction with microreflectivity, and proved the generation of a highly anisotropic change. A peak index change of 0.01 confined within a 1- m penetration depth was noted in hydrogen-soaked fiber. Such asymmetry is associated with strong waveguide and presents new opportunities birefringence of up to 5 for fabricating in-fiber devices and controlling birefringence. ACKNOWLEDGMENT The authors would like to thank the Natural Sciences and Engineering Research Council (Canada) and the Canadian Institute for Photonics Innovation for their support and JDS Uniphase, Ottawa, ON, Canada, particularly D. Grobnric, for the polarization-mode dispersion measurement. K. P. Chen would like to thank Materials Manufacturing Ontario for the A. W. Miller Scholarship. REFERENCES [1] K. O. Hill, Y. Fujii, D. C. Johnson, and B. S. Kawasaki, “Photosensitivity in optical fiber waveguides: Application to reflection filter fabrication,” Appl. Phys. Lett., vol. 32, pp. 647–648, 1978. [2] L. R. Chen, D. J. F. Cooper, and P. W. E. Smith, “Transmission filters with multiple flattened passbands based on chirped Moiré gratings,” IEEE Photon. Technol. Lett., vol. 10, pp. 1283–1285, Sept. 1998. [3] J. A. R. William, I. Bennion, K. Sugden, and N. J. Doran, “Fiber dispersion compensation using a chirped in-fiber Bragg grating,” Electron. Lett., vol. 30, pp. 985–986, 1994. [4] S. Y. Ko, M. W. Kim, D. H. Kim, S. H. Kim, J. C. Jo, and J. H. Park, “Gain control in erbium-doped fiber amplifiers by tuning centre wavelength of a fiber Bragg grating constituting resonant cavity,” Electron. Lett., vol. 34, pp. 900–901, 1998. [5] D. W. Morey, G. Meltz, J. D. Love, and S. J. Hewlett, “Mode-coupling characteristics of UV-written Bragg gratings in depressed-cladding fiber,” Electron. Lett., vol. 30, pp. 730–732, 1994. [6] P. J. Lemaire, R. M. Atkins, V. Mizrahi, and W. A. Reed, “High pressure H loading as a technique for achieving ultra-high UV photosensitivity and thermal sensitivity in GeO doped optical fibers,” Electron. Lett., vol. 29, pp. 1191–1193, 1993. [7] P. J. Lemaire, “Reliability of optical fibers exposed to hydrogen: Prediction of long-term loss increases,” Opt. Eng., vol. 30, pp. 780–789, 1991. [8] T. Erdogan, V. Mizrahi, P. J. Lemaire, and D. Monroe, “Decay of ultraviolet-induced fiber Bragg gratings,” J. Appl. Phys., vol. 76, pp. 73–80, 1994. [9] M. Douay, W. X. Xie, T. Taunay, P. Bernage, P. Niay, P. Cordier, B. Poumellec, L. Dong, J. F. Bayon, H. Poignant, and E. Delevaque, “Densification involved in the UV based photosensitivity of silica glasses and optical fibers,” J. Lightwave Technol., vol. 15, pp. 1329–1342, Aug. 1997.


[10] B. Malo, J. Albert, K. O. Hill, F. Bilodeau, D. C. Johnson, and S. Thériault, “Enhanced photosensitivity in lightly doped standard telecommunication fiber exposed to high fluence ArF excimer laser light,” Electron. Lett., vol. 31, pp. 879–880, 1995. [11] K. P. Chen, P. R. Herman, R. Tam, and J. Zhang, “Rapid long-period grating formation in hydrogen-loaded fiber with 157-nm F -laser irradiation,” Electron. Lett., vol. 36, pp. 2000–2001, 2000. [12] K. P. Chen, P. R. Herman, J. Zhang, and R. Tam, “Fabrication of strong long-period gratings in hydrogen-free fibers with 157-nm F -laser radiation,” Opt. Lett., vol. 26, pp. 771–773, 2001. [13] K. P. Chen, P. R. Herman, and R. Tam, “Strong fiber Bragg grating fabrication by hybrid 157- and 248-nm laser exposure,” IEEE Photon. Technol. Lett., vol. 14, pp. 170–172, Feb. 2002. [14] K. P. Chen, P. R. Herman, and J. Zhang, “Large photosensitivity in germanosilicate planar waveguides induced by 157-nm F -laser radiation,” in Optical Fiber Communication Conf., Tech. Dig. OSA Trends in Optics and Photonics (TOPS), vol. 54, Washington, DC, 2001, Paper WDD81. [15] J. Albert, B. Malo, K. O. Hill, F. Bilodeau, D. C. Johnson, and S. Theriault, “Comparison of one-photon and two-photon effects in the photosensitivity of germanium-doped silica optical fibers exposed to intense ArF excimer laser pulses,” Appl. Phys. Lett., vol. 67, pp. 3529–3531, 1995. [16] J. Nishii, N. Kitamura, H. Yamanaka, H. Hosono, and H. Kawazoe, “Ultraviolet-radiation-induced chemical reactions through one and twophoton absorption processes in GeO -SiO glasses,” Opt. Lett., vol. 20, pp. 1184–1186, 1995. [17] D. Inniss, Q. Zhong, A. M. Vengsarkar, W. A. Reed, S. G. Kosinski, and P. J. Lemaire, “Atomic force microscopy study of UV-induced anisotropy in hydrogen-loaded germanosilicate fibers,” Appl. Phys. Lett., vol. 65, pp. 1528–1530, 1994. [18] Q. Zhong and D. Inniss, “Characterization of the lightguiding structure of optical fibers by atomic force microscopy,” J. Lightwave Technol., vol. 12, pp. 1517–1523, Sept. 1994. [19] D. C. Psaila, C. M. Sterke, and F. Ouellette, “Double pass technique for fabricating compact optical fiber rocking filters,” Electron. Lett., vol. 31, pp. 1093–1094, 1995. [20] P. E. Dyer, A. -M. Johnson, H. V. Snelling, and C. D. Walton, “Measurement of 157 nm F laser heating of silica fiber using an in situ fiber Bragg grating,” J. Phys. D, vol. 34, pp. L109–L112, 2001. [21] P. R. Herman, K. Beckley, B. Jackson, K. Kurosawa, D. Moore, T. Yamanishi, and J. Yang, “Processing applications with the 157 nm fluorine excimer laser,” in Proc. SPIE, vol. 2992, 1997, pp. 86–95. [22] J. Zhang, P. R. Herman, C. Lauer, K. P. Chen, and M. Wei, “157-nm laser-induced modification of fused-silica glasses,” in Laser Appl. Microelectron. Optoelectron. Manuf., SPIE Proc., vol. 4274, 2001, pp. 125–132. [23] P. E. Dyer, R. J. Farley, R. Giedl, and K. C. Byron, “Amplification of fiber Bragg grating reflectivity by post-writing exposure with a 193 nm ArF laser,” Electron. Lett., vol. 14, pp. 1133–1134, 1994. [24] N. F. Borrelli, C. M. Smith, and D. C. Allan, “Exicmer-laser-induced densification in binary silica glasses,” Opt. Lett., vol. 24, pp. 1401–1403, 2000.

1977

[25] T. Erdogan and V. Mizrahi, “Characterization of UV-induced birefringence in photosensitive Ge-doped silica optical fibers,” J. Opt. Soc. Amer. B, vol. 11, pp. 2100–2105, 1994.

Kevin P. Chen, photograph and biography not avialable at the time of publication.

Peter R. Herman received the B.Eng. degree from McMaster University, Hamilton, ON, Canada, in 1980 and the M.A.Sc. and Ph.D. degrees in lasers and diatomic spectroscopy from the Physics Department, the University of Toronto, Toronto, ON, Canada, in 1982 and 1986, respectively. Following a postdoctoral study of X-ray lasers at the Institute of Laser Engineering, Osaka University, Osaka, Japan, he returned to the University of Toronto to take up a faculty position in the Department of Electrical and Computer Engineering. He is currently a Full-Time Professor, supervising a large research group working on two frontiers of laser technology—extreme short wavelength and ultrafast—directed towards novel materials processing applications. These include laser processes defining photonic components, writing three-dimensional optical circuits, packaging, nanofabrication of photonic crystals, and integration of photonic functions for cell and gene biophotonic chips. His research group interacts with several academic and industrial partners and has published more than 100 scientific journal and conference papers. More information can be found at ww.engineering.utoronto.ca/laserphotonics Prof. Herman is a Fellow of the Optical Society of America (OSA) for pioneering research on F2 laser processing applications and is an active Member of the SPIE and the Laser Institute of America.

Rod Taylor, photograph and biography not avialable at the time of publication.

Cyril Hnatovsky, photograph and biography not avialable at the time of publication.