354 J. Opt. Soc. Am. B/Vol. 22, No. 2 /February 2005 Nikogosyan et al.

Fiber Bragg grating inscription by high-intensityfemtosecond UV laser light: comparison

with other existing methods of fabrication

Stephen A. Slattery and David N. Nikogosyan

Department of Physics, National University of Ireland, University College Cork, Cork, Ireland

Gilberto Brambilla

Optoelectronics Research Centre, University of Southampton, Southampton SO17 1BJ, UK

Received June 16, 2004; revised manuscript received August 31, 2004; accepted September 8, 2004

By use of high-intensity ('200 GW/cm2) femtosecond 264-nm laser light and a phase mask technique, Bragggrating inscription in a range of different photosensitive and standard telecom fibers (both H2-free andH2-loaded) was studied. The dependences of the induced refractive index modulation versus the incident flu-ence as well as the thermal decay curves were compared with similar dependences for gratings fabricated byother existing methods. It was shown that with high-intensity UV laser irradiation, two-quantum photore-actions occur in the irradiated fiber core, that result in a significant photosensitivity enhancement of the in-vestigated fibers in comparison with conventional low-intensity 248-nm exposure (by 6–128 times, dependingon fiber type and irradiation intensity). © 2005 Optical Society of America

OCIS codes: 140.3610, 190.4180, 220.4610, 230.1480, 060.2290.

1. INTRODUCTIONFor the inscription of fiber Bragg gratings (FBGs) andlong-period fiber gratings (LPFGs) standard KrF* exci-mer lasers at 248 nm, as well as continuous UV sources at'240 nm, are generally used.1–3 The refractive indexchange generated in fibers at these wavelengths is con-nected with the absorption band of defects in germanosili-cate glass and is mainly of a single-photon nature. Re-cently it was shown that LPFG recording with deep-UV(157 nm), F2 excimer laser radiation,4,5 as a result of thehigh energy of the light quanta, considerably reduces (bymore than two orders of magnitude) the necessary lightfluence. However, it is practically impossible to use thistechnique for Bragg grating fabrication because of theproblem of transporting deep-UV radiation through theair, phase mask and fiber cladding.

Another way to deliver a high excitation energy intothe fiber core is to use multiple-photon absorption of high-intensity, Ti:sapphire laser radiation at 800 nm. Such anapproach was used for LPFG fabrication five years ago.6

Very recently, by means of tight focusing through a phasemask, Bragg gratings were written in standard SMF-28telecom fiber and in silica-core fiber.7,8 It was shown thatmutiple-photon IR inscription could be done without fiberhydrogenation, and the recorded FBGs possessed goodtemperature stability. However, according to recentpublications9,10 the implementation of this approach re-quires careful positioning of the phase mask with respectto the irradiated fiber; failure to meet this requirementimmediately leads to a high value of outband (‘‘gray’’)losses near the transmission loss peak in the irradiatedfiber.7,8

A third way to reach a high value of excitation energy

0740-3224/2005/020354-08$15.00 ©

inside the germanosilicate fiber core is to use high-intensity UV laser irradiation with a subsequent two-step(two-quantum absorption through an intermediate elec-tronic state) or two-photon (two-quantum excitationthrough a virtual-state) excitation process (see Fig. 1). Itis well known that with the rise of irradiation intensityfrom the megawatt to the gigawatt range, the probabilityof the occurrence of two-quantum photoprocesses (bothtwo-step and two-photon) in the irradiated mediumgreatly increases.11 This leads to an enhancement in thephotosensitivity (quantum yield of refractive index modu-lation) of UV irradiated fibers (indeed, their cores) be-cause of the involvement of additional excitation routes.In the case of photosensitive fibers possessing a large lin-ear absorption at the irradiation wavelength, these addi-tional excitation pathways are mostly two-step photoreac-tions sharing their first excitation quantum with thelinear one-step (single-quantum) process. In the case ofstandard telecommunication fibers having much lowerlinear absorption at UV irradiation wavelengths (e.g.,'2.8 cm21 at 264 nm),12 the photosensitivity improve-ment is due to the realization of the two-photon excitationchannel (see the discussion below). Although FBG re-cording by a phase mask technique through two-step ex-citation was realized a decade ago (with an excimer ArF*laser at 193 nm and a light intensity of '0.1GW/cm2),13,14 the second possible excitation channel(through two-photon absorption) was experimentallydemonstrated by us recently with high-intensity ('100GW/cm2) UV laser light.15 We have also investigated thephotosensitivity enhancement in photosensitive and stan-dard fibers with the change from relatively low-intensity248-nm irradiation to high-intensity 264-nm

2005 Optical Society of America

Nikogosyan et al. Vol. 22, No. 2 /February 2005 /J. Opt. Soc. Am. B 355

irradiation,15,16 and have shown that in standard telecomfiber, the photosensitivity improvement is more noticeableas a result of realization of a two-photon excitation path-way.

In this paper we thoroughly investigate FBG fabrica-tion in a number of commercial, standard telecom andphotosensitive fibers (both H2-free and hydrogenated) asa result of high-intensity (100–300 GW/cm2) 264-nm fem-tosecond irradiation and provide comparisons with otherexisting methods of FBG inscription.

2. EXPERIMENTAL SETUPThe scheme of the experimental setup is given in Fig. 2.For the inscription of FBGs we used the fourth-harmonicradiation of a commercial, femtosecond, Nd:glass laser(Twinkle, Light Conversion Ltd., Lithuania).17 This la-ser system emits femtosecond 264-nm pulses that areGaussian in time and space and were characterized by usearlier.18 The measured pulse duration was 220 fs(FWHM), the beam diameter 0.3 cm (FWHM), the repeti-tion rate 27 Hz, and the pulse energy up to 300 mJ. Thefemtosecond UV pulses were focused by a fused silica cy-lindrical lens of 21.8-cm focal length through a phasemask and onto the fiber (with acrylate coating removed),which was placed behind the phase mask at a distance of'100 mm. The phase mask was fixed together with thefiber onto a three-dimension positioner that could bealigned relative to the laser beam. In all the experi-ments described below we used the static phase masktechnique (as opposed to the scanning phase mask tech-nique employed in Ref. 15).

A specially designed phase mask (Ibsen Photonics) witha 1.07-mm pitch optimized for 264-nm illumination and ofsmall thickness (1 mm) was employed in the experiments.This allowed us to reduce significantly the losses due totwo-photon absorption in the substrate of the phasemask. In a separate experiment we measured the two-photon absorption coefficient of the phase mask materialand found it to be equal to 2.3 3 10211 W/cm2, which isclose to the best values for the Suprasil fused silicasamples (1.7 3 10211 cm/W) reported in Ref. 18.

The incident laser pulse energy was measured with apyroelectric detector (Ophir Optronics Inc. PE10) con-nected to a Laserstar energy meter system (Ophir Op-

Fig. 1. Simplified scheme of photoexcitation and energy levelsin germanosilicate glass.

tronics Inc.). The longitudinal displacement of the lenswith respect to the fiber allowed us to vary the incidentUV irradiation intensity. In accordance with Ref. 18 theincident pulse peak intensity and total incident irradia-tion fluence were calculated from the following equations,respectively:

I 5ep

S p

ln 2 D 3/2 tw2

4 S F 2 S

F2

l

nF D , (1)

E 5epN

pw2

2 ln 2 S F 2 S

F2

l

nF D , (2)

where ep is the pulse energy incident on the fiber; F is thefocal length of the lens at 264 nm; S is the distance be-tween the focusing lens and the fiber; l is the thickness ofthe phase mask; n is the refractive index of fused silica(phase mask) at 264 nm (n 5 1.5007);19 t is the pulse du-ration at FWHM, equal to 220 6 10 fs;18 w is the beamradius at FWHM, equal to 1.51 6 0.02 mm;18 and N is thetotal number of laser pulses incident on the fiber. Therelative accuracy of our fluence and intensity measure-ments were '1% and 4%, respectively.18

The monitoring of Bragg grating transmission duringinscription was done with an AQ4222 EE LED source andAQ6317C optical spectrum analyzer (both supplied byYokogawa Europe BV). The UV exposure was computercontrolled with LabVIEW software and a peripheral com-ponent interface–general-purpose interface bus card(both from National Instruments).

For comparison purposes we also performed the FBGinscription with conventional 248-nm radiation from aKrF* excimer laser (GSI Lumonics). The beam size,pulse duration and repetition rate were 7 3 10 mm, 20ns, and 20 Hz, respectively. The beam was focused by acylindrical lens of focal length 150 mm onto a 3-mm thickphase mask optimised for 248-nm exposure with a pitchof 1.060 mm. The fiber was positioned within a few mi-crometers of the phase mask. The incident intensity wasfixed at 25 MW/cm2. A small section (3 mm long) of thebeam was used to write the grating. The real-time moni-toring of Bragg grating transmission was done with abroadband LED source and an optical spectrum analyzer(HP 70950A).

In our experiments we used three different single-modefibers: (1) standard telecom fiber SMF-28 (Corning) with

Fig. 2. Scheme of the experimental setup. EE-LED, edge-emitting LED; OSA, optical spectrum analyzer.

356 J. Opt. Soc. Am. B/Vol. 22, No. 2 /February 2005 Nikogosyan et al.

a core diameter of 8.2 mm and numerical aperture of 0.14;(2) photosensitive GF1 fiber from Nufern with a core di-ameter of 9.0 mm and numerical aperture of 0.13; and (3)photosensitive, boron-codoped PS1250/1500 fiber from Fi-bercore with a core diameter of 6.6 mm and numerical ap-erture of 0.13. Sensitization of the fibers was performedin a hydrogen atmosphere at '15 MPa and 70 °C for 2weeks.

The induced refractive index modulation was calcu-lated by

Dnmod 5l

phLtanh21~1 2 T !1/2

5l

2phLlnF1 1 ~1 2 T !1/2

1 2 ~1 2 T !1/2G , (3)

where T is the transmission loss peak of a uniform grat-ing; l is the grating operation wavelength; L is the lengthof the grating; and h is the mode overlap parameter (frac-tion of the fiber mode power contained in the core), whichis given by

h 5p2d2K2

l2 1 p2d2K2, (4)

where d is the fiber core diameter and K is the numericalaperture of the fiber. The latter equation can be easilydeduced from the ratio of the power of the lowest-ordermode in the fiber core to that in the cladding, given in Ref.20 as

Pcore /Pclad 5p2d2K2

l2. (5)

Temperature stability studies of recorded FBGs wereperformed by the isochronal annealing approach.21 Thefiber was kept in a quartz cylinder inside an oven (Carbo-lite MTF 12/25/250). The temperature of the oven wasincreased from room temperature to 50 °C (or 100 °C) andkept at that temperature for 0.5 h. The changes in thegrating strength and the position of the transmission losspeak were monitored by an optical spectrum analyzer(Ando AQ6317C). After that the temperature was raisedin steps of 50 °C (or 100 °C) and kept at each temperaturefor a half hour; each time, the grating signal was re-corded. The normalized refractive index modulation(proportional to the integrated coupling constant) was cal-culated and graphed against the temperature.

3. RESULTS AND DISCUSSIONFigures 3–5 show the increase in the transmission losspeak and the simultaneous redshift of the FBG wave-length with the irradiation fluence for three different fi-bers under high-intensity 264-nm exposure. It is alsoseen from these graphs that the rise in grating strength isaccompanied by an increase in the grating linewidth.Figures 6 and 7 demonstrate the experimental curves ofthe refractive index modulation versus the irradiation flu-ence in FBGs recorded in four different fiber samples withhigh-intensity femtosecond UV irradiation. From these

graphs it follows that the photosensitivity of the in-vestigated fibers increases in the following order:H2-free Nufern , H2-free Fibercore , H2-loaded SMF-28, H2-loaded Nufern. For SMF-28 the refractive indexmodulation is proportional to the irradiation intensity(Fig. 6); this observation confirms the two-quantum char-acter of the photochemical reactions involved. This alsoagrees with our previous findings.15

Using the elementary expressions for single-quantumand two-photon absorption and the fixed-field approxima-tion, and assuming a rectangular shape for the laserpulse, one can easily estimate the fraction of UV light ab-sorbed in the cladding and core of SMF-28 fiber at high-intensity 264-nm excitation. In our previous work12 wemeasured the linear and two-photon absorption coeffi-cients for 3.5-mol.% Ge-doped fused silica preform, whichcorresponds in dopant concentration to the core of

Fig. 3. Growth of grating strength and shift of the transmissionloss peak in a FBG recorded in H2-loaded Nufern GF1 fiber withdifferent light fluences and incident intensity of 231 GW/cm2.The linewidth values of the recorded gratings at the 3 dB levelare 0.14, 0.28, and 0.59 nm at fluences 14, 38, and 120 J/cm2, re-spectively.

Fig. 4. Growth of grating strength and shift of the transmissionloss peak in a FBG recorded in H2-loaded Corning SMF-28 fiberwith different light fluences and incident intensity of 241GW/cm2. The linewidth values of the recorded gratings at the 3dB level are 0.25, 0.40, and 0.62 nm at fluences 81, 115, and 228J/cm2, respectively.

Nikogosyan et al. Vol. 22, No. 2 /February 2005 /J. Opt. Soc. Am. B 357

SMF-28 fiber (3 mol.% Ge). The values obtained at 264nm were 2.8 cm21 and 42 3 10211 cm/W. The two-photon absorption coefficient for pure fused silica (fibercladding) at 264 nm is 1.8 3 10211 cm/W.12 Using thisdata and the geometrical parameters of SMF-28 givenabove, we can show that with high-intensity (240GW/cm2) 264-nm irradiation, the fiber cladding absorbsonly 2.5% of the incident light energy (by the two-photonabsorption mechanism), while the fiber core absorbs 8.5%of the light energy entering the core (8.3% by two-photonabsorption and 0.2% by linear absorption), or 8.3% of theincident light energy. Therefore we can ignore linear ab-sorption (and hence the two-step excitation process) in thefiber core of SMF-28 and consider the FBG inscription pro-cess in SMF-28 fiber (Fig. 6) a solely two-photon one.

On the other hand, from these estimates it follows thatthe light energy absorbed in the fiber core exceeds by only3.3 times the amount absorbed in the cladding. Thebandgap energy value for fused silica is 9.3 eV,22 which isbelow the doubled quantum energy value 2hn

Fig. 5. Growth of grating strength and shift of the transmissionloss peak in a FBG recorded in H2-free Fibercore PS1250/1500fiber with different light fluences and incident intensity of 205GW/cm2. The linewidth values of the recorded gratings at the 3dB level are 0.13, 0.28, and 0.49 nm at fluences 298, 775, and1728 J/cm2, respectively.

Fig. 6. Refractive-index modulation versus fluence forH2-loaded Corning SMF-28 fiber irradiated at two different inci-dent intensities.

5 9.39 eV. Therefore, in principle there could be somerefractive index modulation in fiber cladding as well; how-ever, we never observed any significant outband (‘‘gray’’)losses near the transmission loss peak, which are charac-teristic of the multiple-photon IR approach.7,8

At the moment we cannot give similar estimates forphotosensitive fibers irradiated with high-intensity264-nm radiation as the corresponding two-photon ab-sorption coefficient values for the fiber core material areunknown. However, it is clear that with an increase indopant concentration the relative amount of light energyabsorbed in the core will increase in comparison withstandard telecom fiber.

From Fig. 7 it is clear that hydrogenation of NufernGF1 fiber increases its photosensitivity by '40 times.Corresponding to this, we failed to record FBGs in H2-freeSMF-28 under present experimental conditions. Furtherincrease of the irradiation intensity (above 250 GW/cm2)immediately led to a rise in two-photon absorption lossesin the phase mask substrate, which limited the UV inten-sity incident on the fiber.

Fig. 7. Comparison of refractive index modulation curves forthree photosensitive fiber samples: H2-free and H2-loadedNufern GF1 and H2-free Fibercore PS1250/1500.

Fig. 8. Comparison of refractive index modulation curves forH2-free Fibercore PS1250/1500 fiber after femtosecond (264 nm,205 GW/cm2) and nanosecond (248 nm, 25 MW/cm2) irradiation.

358 J. Opt. Soc. Am. B/Vol. 22, No. 2 /February 2005 Nikogosyan et al.

Figures 8 and 9 present the results of a comparativestudy of the refractive index modulation change withlight fluence for H2-free Nufern GF1 and FibercorePS1250/1500 fibers taken with femtosecond and nanosec-ond UV irradiation at wavelengths of 264 and 248 nm, re-spectively. From these graphs it follows that the photo-sensitivity of these fibers is significantly higher at thefemtosecond (264 nm) irradiation than at the nanosecond(248 nm) one. Similar data were obtained for H2-loadedNufern GF1 and Corning SMF-28 fibers (data not shown).

The observed behavior of the wavelength shift of thetransmission loss peak (Figs. 10 and 11) was similar tothat of the refractive index modulation curves (Figs. 6 and7), which is not surprising because according to simpletheoretical considerations for low exposure times the shiftin FBG wavelength is proportional to the refractive indexchange.2

Our results on thermostability of FBGs recorded withhigh-intensity 264-nm femtosecond irradiation are pre-sented in Figs. 12–14. From these graphs it follows thatFBGs recorded in Fibercore PS1250/1500 fiber exhibit the

Fig. 9. Comparison of refractive index modulation curves forH2-free Nufern GF1 fiber after femtosecond (264 nm, 180GW/cm2) and nanosecond (248 nm, 25 MW/cm2) irradiation.

Fig. 10. Wavelength change of transmission loss peak versuslight fluence for H2-loaded Corning SMF-28 and Nufern GF1 fi-bers.

lowest thermal stability. The refractive index modula-tion value decreases by five times at a temperature of400 °C (Fig. 12). The refractive index modulation valuemeasured in annealed Nufern GF1 fiber decreases by fivetimes at a temperature of '700 °C. It should be empha-sized that H2-loaded and H2-free Nufern GF1 fibers havedifferent decay curves. In the temperature range 100–600 °C the H2-free Nufern fiber is more temperature re-sistant than the H2-loaded one, whereas at temperaturesabove 600 °C the H2-loaded Nufern GF1 fiber is more re-sistant (Fig. 13). The thermal decay curves for FBGs cre-ated in hydrogenated Nufern fiber at high-intensity264-nm and conventional 248-nm irradiations are similar(data not shown). In contrast, the thermal decay curvesfor FBGs recorded in H2-loaded SMF-28 by 264-nm and248-nm laser sources differ significantly (Fig. 14). Theseresults show that FBGs fabricated in hydrogenated stan-dard telecom fiber SMF-28 possess the highest thermalstability of all fibers investigated; its refractive indexmodulation value decreases by five times at a tempera-ture of more than 800 °C.

Fig. 11. Wavelength change of transmission loss peak versuslight fluence for H2-free Fibercore PS1250/1500 and Nufern GF1fibers.

Fig. 12. Thermal stability of FBG recorded in a H2-free Fiber-core PS1250/1500 fiber with 264-nm radiation at incident inten-sity of 138 GW/cm2 and incident fluence of 0.38 kJ/cm2; the initialgrating strength was 12.7 dB.

Nikogosyan et al. Vol. 22, No. 2 /February 2005 /J. Opt. Soc. Am. B 359

During the annealing process the grating wavelengthexperiences a redshift due to thermal expansion of the fi-ber. The typical experimental dependence of a FBGtransmission loss peak wavelength versus the tempera-ture recorded in H2-free Nufern GF1 fiber is given in Fig.15. The measured thermal shift of the Bragg wavelengthfor all fibers used was in the range (0.9–1.1)3 1022 nm/°C, which agrees well with the estimate of(1 –2) 3 1022 nm/°C given in Ref. 2.

Let us compare the photosensitivity results obtained byus with a high-intensity 264-nm femtosecond sourceagainst those of a conventional 248-nm KrF* excimer la-ser. The results are presented in Table 1.

We discuss first the non-loaded fibers (Figs. 8 and 9).To record a 9.95-dB grating in a Fibercore PS1250/1500fiber at 248-nm irradiation, an incident laser fluence of4.4 kJ/cm2 was needed. For the inscription of a similargrating at 264-nm irradiation with incident intensity of

Fig. 13. Thermal stability of FBGs inscribed in H2-free andH2-loaded Nufern GF1 fibers with 264-nm radiation at incidentintensities of 164 and 231 GW/cm2 and incident fluences of 7.18and 0.26 kJ/cm2, respectively; the initial grating strengths were8.4 and 31.7 dB, respectively.

Fig. 14. Thermal stability of two FBGs recorded in H2-loadedCorning SMF-28 fiber at 248- and 264-nm irradiation. The in-cident intensities were 0.025 and 166 GW/cm2 and incident flu-ences were 16.5 and 0.24 kJ/cm2, respectively. The initial grat-ing strengths were 10.4 and 11.6 dB, respectively.

205 GW/cm2, 5.7 times less fluence was necessary. Tofabricate a FBG of 6.92 dB in H2-free, Nufern GF1 fiberwith 248-nm radiation, 11.9 kJ/cm2 incident fluence wasapplied. This is 7.6 times more than the amount thatwas needed in the case of 264-nm irradiation with inci-dent intensity of 180 GW/cm2.

To create a 10.2-dB grating in photosensitive,H2-loaded Nufern GF1 fiber, we needed to apply 18 timesless fluence with 264-nm irradiation than with 248 nm.Similarly, for H2-loaded, Corning SMF-28 fiber at 264-nmirradiation, we needed to use 85 times less fluence (if ir-radiating with 167 GW/cm2 intensity), or even up to 128times less fluence value (if irradiating with 241 GW/cm2

intensity), compared with 248-nm irradiation.Further comparison could be done with literature data

on FBG inscription by a low-intensity (0.6 MW/cm2)pulsed UV copper vapor laser emitting at 255 nm.23 Toinduce a refractive index change of 0.6 3 1023 in H2-freeFibercore PS1250/1500, it was necessary to apply a totalincident fluence of 5 kJ/cm2. In our case, the same re-fractive index change was produced with a total incidentfluence of 1.5 kJ/cm2 (Fig. 7), which is '3 times smaller.When irradiating H2-loaded SMF-28 fiber with 255-nmpulses, a refractive index modulation of 13 1023 was produced with a total incident fluence of 12.5kJ/cm2. In our case, irradiating at 264 nm with an inten-sity of 241 GW/cm2 (Fig. 6), we reached the same refrac-tive index modulation with a total incident fluence of 0.32kJ/cm2, which is '40 times smaller.

The same copper vapor laser can operate at 271 nmwith a pulse intensity of 1.2 MW/cm2.23 It is interestingto compare the photosensitivity of H2-loaded SMF-28 fi-ber at this wavelength with our results at 264 nm. Toproduce a refractive index modulation of 1023 one shouldapply 271-nm light with a fluence of 158 kJ/cm2. For264-nm irradiation with an intensity of 241 GW/cm2, weneed a fluence value of only 0.32 kJ/cm2 (Fig. 6), which isdifferent by 494 times. This comparison supports oursuggestion made above regarding the two-photon nature

Fig. 15. The shift of the FBG wavelength versus temperaturerecorded in H2-free Nufern GF1 fiber with an incident intensityof 164 GW/cm2 and an incident fluence of 7.18 kJ/cm2; the initialgrating strength was 8.4 dB.

360 J. Opt. Soc. Am. B/Vol. 22, No. 2 /February 2005 Nikogosyan et al.

Table 1. Comparison between High-Intensity 264-nm Approach of FBG Inscription and the ConventionalOne

Fiber

IrradiationWavelength

[nm]

IncidentIntensity[GW/cm2]

TransmissionPeak Loss

[dB]

Incident LaserFluence[kJ/cm2]

H2-free Fibercore PS1250/1500 248 0.025 9.95 4.4264 205 10.65 0.775

H2-free Nufern GF1 248 0.025 6.92 11.9264 180 7.02 1.573

H2-loaded Nufern GF1 248 0.025 10.17 0.7264 231 10.19 0.038

H2-loaded Corning SMF-28 248 0.025 10.31 14.7264 167 10.34 0.172264 241 9.77 0.115

of the photochemistry involved in the case of high-intensity 264-nm irradiation of H2-loaded SMF-28 fiber.

It is also of interest to compare our results with themultiple-photon infrared approach to the fabrication ofFBGs,7,8,24,25 which utilizes 120-fs, 800-nm pulses gener-ated by a Ti:sapphire laser. The authors of Refs. 7, 8, 24,and 25 give contradictory data on UV irradiance: InRefs. 7 and 24 the incident intensity of 12,000 GW/cm2 forFBG inscription in H2-free SMF-28 is mentioned,whereas in Refs. 8 and 25 the cited infrared peak inten-sity is only 2900 GW/cm2 ('4 times less). In Ref. 25 theauthors state that for FBG inscription in H2-loadedSMF-28 fiber they needed of an order less infrared inten-sity. If we combine all the data from Refs. 7, 8, 24, and25 it follows that to induce a refractive index change of1.85 3 1023 in H2-loaded SMF-28 by high-intensity(3000–12,000 GW/cm2) 800-nm femtosecond radiation, anincident fluence of 0.8–3.2 kJ/cm2 is needed. In our casewith 264-nm irradiation at a 241-GW/cm2 incident inten-sity, we required a fluence value of 0.65 kJ/cm2 (see Fig.6), which is in all cases smaller than that under the IRmultiple-photon approach.

We tried to fabricate Bragg gratings with the highestpossible values for the transmission loss peak with high-intensity 264-nm irradiation. We have shown that thevalues of 30 dB for photosensitive fibers and of 45 dB forhydrogenated standard telecommunication fibers are eas-ily reachable by our technique. The experimental trans-mission spectra with maximal grating strength recordedin H2-free Fibercore and H2-loaded SMF-28 fibers are pre-sented in Figs. 16 and 17, respectively. The further in-crease is limited by the sensitivity of our spectral analy-ser and/or by the absence of a splicer for making loss-freeconnections in the case of photosensitive fibers. It shouldbe emphasized that we did not observe any saturation ofthe refractive index modulation, especially in the case oftwo-photon excitation (SMF-28 fiber).

The comparison between our FBGs and those with sup-pressed cladding modes recently fabricated by the IRmultiple-photon approach9,10 shows that in both cases therecorded gratings possess similar values of off-band lossesand linewidth. On the other hand FBGs inscribed by

high-intensity 264-nm femtosecond laser pulses are ther-mally erased at lower temperatures (at least partly),which in the case of photosensitive fibers could be con-nected with the involvement of single-quantum processes.

Fig. 16. Maximal grating strength achieved in H2-free Fiber-core PS1250/1500 fiber irradiated at incident intensity of 205GW/cm2 and incident fluence of 2.68 kJ/cm2.

Fig. 17. Maximal grating strength achieved in H2-loaded Corn-ing SMF-28 fiber irradiated at incident intensity of 241 GW/cm2

and incident fluence of 1.22 kJ/cm2.

Nikogosyan et al. Vol. 22, No. 2 /February 2005 /J. Opt. Soc. Am. B 361

4. CONCLUSIONTwo-quantum high-intensity 264-nm inscription of Bragggratings in standard telecom (Corning SMF-28) and pho-tosensitive (Nufern GF1, Fibercore PS1250/1500) fiberswas investigated. The refractive index change depen-dences on incident fluence at 264 nm irradiation obtained,as well as thermostability results, were compared withsimilar graphs obtained for low-intensity irradiation at248 nm, 255 nm, and 271 nm, as well as with the resultsobtained with high-intensity, IR multiple-photon excita-tion at 800 nm. It was shown that for the fiber samplesinvestigated, the efficiency of grating inscription is maxi-mal with high-intensity femtosecond 264-nm irradiation,whereas the thermal stability of our gratings is compa-rable with that obtained by other techniques but less thanthat observed under 800-nm exposure.

ACKNOWLEDGMENTSWe acknowledge the contribution of Adrian Dragomir(currently with the European Patent Office), who partici-pated in the early stages of this study, and the fruitful co-operation with Kristian Buchwald (Ibsen Photonics).The authors are also grateful to the Science FoundationIreland for their generous financial support (grant 0.1/IN.1/IO58).

Corresponding author D. N. Nikogosyan’s e-mail ad-dress is niko@physics.ucc.ie.

REFERENCES1. P. St. J. Russell and J. L. Archambault, ‘‘Fiber gratings,’’ in

Optical Fiber Sensors, Vol. 3, Components and Subsystems(Artech House, Boston, Mass., 1996), pp. 9–67.

2. R. Kashyap, Fiber Bragg Gratings (Academic, San Diego,Calif., 1999).

3. A. Othonos and K. Kalli, Fiber Bragg Gratings: Funda-mentals and Applications in Telecommunications and Sens-ing (Artech House, Boston, Mass., 1999).

4. K. P. Chen, P. R. Herman, R. Tam, and J. Zhang, ‘‘Rapidlong-period grating formation in hydrogen-loaded fibre with157-nm F2-laser radiation,’’ Electron. Lett. 36, 2000–2001(2000).

5. K. P. Chen, P. R. Herman, J. Zhang, and R. Tam, ‘‘Fabrica-tion of strong long-period gratings in hydrogen-free fiberswith 157-nm F2-laser radiation,’’ Opt. Lett. 26, 771–773(2001).

6. Y. Kondo, K. Nouchi, T. Mitsuyu, M. Watanabe, P. G. Ka-zansky, and K. Hirao, ‘‘Fabrication of long-period fiber grat-ings by focused irradiation of infrared femtosecond laserpulses,’’ Opt. Lett. 24, 646–648 (1999).

7. S. J. Mihailov, C. W. Smelser, P. Lu, R. B. Walker, D. Grob-nic, H. Ding, G. Henderson, and J. Unruh, ‘‘Fiber Bragggratings made with a phase mask and 800-nm femtosecondradiation,’’ Opt. Lett. 28, 995–997 (2003).

8. S. J. Mihailov, C. W. Smelser, D. Grobnic, R. B. Walker, P.Lu, H. Ding, and J. Unruh, ‘‘Bragg gratings written in all-SiO2 and Ge-doped core fibers with 800-nm femtosecond ra-diation and a phase mask,’’ J. Lightwave Technol. 22, 94–100 (2004).

9. C. W. Smelser, D. Grobnic, and S. J. Mihailov, ‘‘Generationof pure two-beam interference grating structures in an op-

tical fiber with a femtosecond infrared source and a phasemask,’’ Opt. Lett. 29, 1730–1732 (2004).

10. D. Grobnic, C. W. Smelser, and S. J. Mihailov, ‘‘Fiber Bragggratings with suppressed cladding modes made in SMF-28with a femtosecond IR laser and a phase mask,’’ IEEE Pho-ton. Technol. Lett. 16, 1864–1866 (2004).

11. D. N. Nikogosyan, ‘‘Two-quantum UV photochemistry ofnucleic acids: comparison with conventional low-intensityUV photochemistry and radiation chemistry,’’ Int. J. Radiat.Biol. 57, 233–299 (1990).

12. A. Dragomir, J. G. McInerney, D. N. Nikogosyan, and P. G.Kazansky, ‘‘Two-photon absorption properties of commer-cial fused silica and germanosilicate glass at 264 nm,’’ Appl.Phys. Lett. 80, 1114–1116 (2002).

13. B. Malo, J. Albert, K. O. Hill, F. Bilodeau, D. C. Johnson,and S. Theriault, ‘‘Enhanced photosensitivity in lightlydoped standard telecommunication fibre exposed to highfluence ArF excimer laser light,’’ Electron. Lett. 31, 879–880 (1995).

14. 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 sensitivity of germanium-doped silicaoptical fibers exposed to intense ArF excimer laser pulses,’’Appl. Phys. Lett. 67, 3529–3531 (1995).

15. A. Dragomir, D. N. Nikogosyan, K. A. Zagorulko, P. G.Kryukov, and E. M. Dianov, ‘‘Inscription of fiber Bragg grat-ings by ultraviolet femtosecond radiation,’’ Opt. Lett. 28,2171–2173 (2003).

16. A. Dragomir, D. N. Nikogosyan, and G. Brambilla, ‘‘In-creased photosensitivity of Ge-doped and Ge, Sn-doped fi-bres under high-intensity 264 nm laser light,’’ Electron.Lett. 39, 1437–1438 (2003).

17. Twinkle, highly integrated, pico/femtosecond, Nd:glass la-ser system with chirped-pulse amplifications; http://www.lightcon.com/lc/scientific/laser.htm.

18. A. Dragomir, J. G. McInerney, and D. N. Nikogosyan, ‘‘Fem-tosecond measurements of two-photon absorption coeffi-cients at l 5 264 nm in glasses, crystals, and liquids,’’ Appl.Opt. 41, 4365–4376 (2002).

19. D. N. Nikogosyan, Properties of Optical and Laser-RelatedMaterials. A Handbook (Wiley, Chichester, UK, 1997), p.177.

20. M. J. Adams, An Introduction to Optical Waveguides (Wiley,Chichester, UK, 1981), Chap. 7.

21. T. Taunay, P. Niay, P. Bernage, M. Douay, W. X. Xie, D.Pureur, P. Cordier, J. F. Bayon, H. Poignant, E. Delevaque,and B. Poumellec, ‘‘Bragg grating inscriptions within strainmonomode NA germania-doped fibres: Part I. Experi-mentation,’’ J. Phys. D 30, 40–52 (1997).

22. J. Nishii, N. Kitamura, H. Yamanaka, H. Hosono, and H.Kawazoe, ‘‘Ultraviolet-radiation-induced chemical reactionsthrough one- and two-photon absorption processes inGeO2-SiO2 glasses,’’ Opt. Lett. 20, 1184–1186 (1995).

23. E. K. Illy and H. J. Booth, ‘‘Comparison of fibre Bragg grat-ing writing at multiple UV wavelengths,’’ in Proceedings of28th European Conference on Optical Communications(European Conference on Optical Communications, Copen-hagen, Denmark, 2002), paper M602.

24. S. J. Mihailov, C. W. Smelser, P. Lu, R. B. Walker, D. Grob-nic, H. Ding, and J. Unruh, ‘‘Fiber Bragg gratings (FBG)made with a phase mask and 800-nm, femtosecond radia-tion,’’ in Optical Fiber Communications Conference, Vol. 86of OSA Trends in Optics and Photonics Series (Optical So-ciety of America, Washington, D.C., 2003), postdeadline pa-per PD30.

25. C. W. Smelser, S. J. Mihailov, D. Grobnic, R. B. Walker, P.Lu, and H. Ding, ‘‘Impact of hydrogen loading on the fabri-cation of fiber Bragg gratings with ultrafast 800-nm laserradiation,’’ in Bragg Gratings, Photosensitivity, and Polingin Glass Waveguides Conference, Vol. 93 of OSA Trends inOptics and Photonics Series (Optical Society of America,Washington, D.C., 2003), paper PD3.