Comparison of the Thermal Tuning Capability of

Different Types of Bragg Grating Filters for

Wavelength Division Multiplexing Applications

M. J. N. Lima, A. L. J. Teixeira, J. R. Ferreira da Rocha

Department of Electronics and Telecommunications and Institute of

Telecommunications, University of Aveiro, 3810-193 Aveiro, Portugal

Tel: +351 234 370383, Fax: +351 234 381128, E-mail: mmcc@ua.pt

O. Frazão

INESC Porto – UOSE, Rua Campo Alegre, 687, 4169-007 Porto, Portugal

P. S. B. André

Department of Physics and Institute of Telecommunications, University of Aveiro,

3810-193 Aveiro, Portugal

Abstract

In this contribution we compare the thermal tuning capabilities of two type I gratings

written in an unloaded and a hydrogen-loaded germanium doped silica fibres, and a

type IIa grating written in unloaded fibre. The hydrogen-loaded grating is annealed

after writing, to remove any unreacted hydrogen. We further study grating response

stability with temperature, an important property when these devices are used as

optical filters in wavelength division multiplexing systems. Experimental results

show that the worst option is the type I grating in unloaded fibre which presents the

highest bandwidth and group delay variations with temperature, so the less suitable as

thermo-tuneable optical filter in wavelength-multiplexed systems. For the other two

options, the bandwidth and group delay variations are much smaller, with the lowest

values obtained for the hydrogen-loaded type I grating. However, the best thermal

tuning efficiency is obtained for the type IIa grating.

Key-words: Optical communications, wavelength division multiplexing, Bragg

gratings, thermal tuning.

1. Introduction

The wavelength tuning capabilities of fibre Bragg gratings (FBG) are related to their

capacity to shift the reflection spectrum central wavelength. To obtain such tuning

effect two main methods are used: by modifying the fibre refraction index or by

changing the grating period. These parameter variations can be induced thermally

[1][2] or by mechanical stress [2][3]. Although the latter approach leads typically to a

broader tuning range and a higher tuning speed [3], the first allows higher

reproducibility and reversibility of the tuned frequencies [4], important properties for

wavelength-division multiplexing (WDM) applications.

However, as the FBG is tuned by varying its central reflection wavelength, it is also

crucial that the shapes of its reflection spectrum and group delay characteristics

remain stable, so that there is negligible performance variation when selecting

different WDM channels.

Bragg gratings are nowadays routinely fabricated in germanosilicate fibre by

exposing the fibre core to an interference pattern of pulsed or continuous-wave

ultraviolet (UV) light. The majority of the studies made so far on thermal stability of

germanosilicate fibre gratings focused on the analysis of the reflectivity peak decay

with temperature [5]-[8].

In this contribution we compare the thermal tuning capabilities of type I gratings,

written in unloaded and hydrogenated germanosilicate fibres, and type IIa gratings

written in unloaded fibres, and analyse their characteristics stability as the

temperature is changed. We observe not only the reflection peak amplitude variation,

but also the thermally-induced changes in other characteristics of the FBG response,

like the bandwidth and group delay, critical parameters for their performance as

WDM filters.

2. Gratings Fabrication

As referred above, for this study we consider type I and type IIa gratings. The latter

are often observed in highly germanium-doped fibres, and only in non-hydrogenated

ones [9]. We considered two type I gratings written in a hydrogen-loaded and an

unloaded fibre, and one type IIa grating.

We have used the same fibre and writing conditions in all cases, so that the observed

behaviour depends exclusively on the type of grating considered. The FBGs were

written in a commercial germanosilicate optical fibre (Fibercore SM1500 4.2/125),

by exposing it to an interference pattern originated from a UV (248 nm) beam

through a phase-mask. We have used a KrF excimer laser, with a fluence of 300

mJ/cm2 per pulse, and a pulse repetition rate of 30 Hz. The phase-mask period was

1067 nm and its length 10 mm.

To hydrogenate the fibre, it was kept under high-pressure hydrogen atmosphere (120

bars) during one week, in order to enhance its photosensitivity. After the UV

exposure, the hydrogenated grating without coating was annealed for 12 h at 160 ºC,

to remove any unreacted hydrogen and increase its stability [7].

In Fig. 1 we present the measured maximum reflectivity (Rmax) and mean refractive

index perturbation (<∆neff>) as a function of the irradiation time, considering the

formation of type I and type IIa gratings, using respectively the hydrogen loaded and

the unloaded germanosilicate fibres. For the latter case, we have considered the FBG

after 100 minutes irradiation (accumulated fluency ∼54 kJ/cm2), denoted by FBG1

(type IIa curves) in the figure, and after only 7.5 minutes (∼4 kJ/cm2), FBG2 (type I).

For the hydrogen loaded case, due to the hydrogen loading there is no type IIa

behaviour [9] and another type I grating of higher reflectivity than FBG2 can be

achieved just after irradiating for 7 minutes (∼3.8 kJ/cm2), FBG3 in the figure.

3. Experimental Results

In this section we analyse the behaviour of the three FBGs described above, as we

vary the temperature. We measure the thermal tuning efficiency, that is the

coefficient of central wavelength shift with temperature (pm/ºC), and analyse other

thermally induced alterations on the gratings response, namely on the maximum

reflectivity, the grating bandwidth and the group delay.

The measuring set-up for characterizing the FBGs is based on the phase-delay

technique, but using a lightwave component analyser instead of a vector-voltmeter

[10]. The light from a tuneable laser is modulated using a Mach-Zehnder

interferometer and launched into the input port of the FBG. The amplitude and phase

of the reflected signal are compared with the corresponding parameters of the input

modulated signal, in a lightwave component analyser. As the wavelength of the laser

is tuned with steps of 0.01 nm, the reflectivity and group delay are averaged over 128

samples. The chosen modulation frequency was 1 GHz, thus one degree of phase

change is equivalent to a group delay of 2.77 ps, and the measurements precision is

∼22 fs.

To control the gratings temperature we involved them with thermal conductive glue,

over a thermoelectric Peltier element, so that the temperature could be easily set. The

temperature was then varied from 30 ºC to 75 ºC (with a 15 ºC step) and the grating

characteristics measured for each case.

In Fig. 2 we present the reflection spectra and group delay curves as a function of

wavelength deviation from the peak reflection wavelength, for the three gratings

referred above. Results are presented for the minimum and maximum temperatures

considered in this study (30 ºC and 75 ºC). Fig. 3 shows the peak reflection

wavelength variation (∆λcent) with temperature and Fig. 4 presents the -3 dB grating

bandwidths (BW-3dB) for four temperatures.

4. Results Discussion

Observing Fig. 2 we notice that for the considered temperature range, the peak

reflectivity variations with temperature are negligible for the three FBGs. This

behaviour was expected at these temperatures (<80 ºC), if we consider the thermal

decay studies presented in the literature, for gratings written in hydrogen-loaded and

unloaded germanosilicate fibres [6]-[8].

With respect to the thermal tuning efficiency they present similar values, 13.8 pm/ºC,

13.3 pm/ºC and 13.1 pm/ºC, respectively for FBG1, FBG2 and FBG3 (see Fig. 3).

Nevertheless, the highest value is obtained for the type IIa grating, FBG1, allowing

the highest tuning range for the same temperature variation.

Returning to Fig. 2 we observe that the variations in the reflection spectrum and

group delay curve of FBG2 are much more pronounced than for the other two FBGs.

The measured FBGs bandwidth variations for the temperature range 30 to 75 °C,

shown in Fig. 4, confirm this result. Indeed, the -3 dB bandwidth variations for FBG3

are within ±0.64% of the filter mean bandwidth (0.313 nm), for FBG1 are within

±2.35% of mean 0.42 nm and for the FBG2, the worst result, within ±25% of its

mean bandwidth (0.19 nm), for the considered temperature ranges. The bandwidth

variations obtained for FBG1 and FBG3 are quite reduced, nevertheless they are

smaller for the properly stabilized type I grating in hydrogen-loaded fibre, FBG3.

From these results, it is clear that one should avoid the use of type I gratings like

FBG2, as thermo-tuneable optical filters. On the other hand, type I gratings in

hydrogen-loaded fibres, with proper stabilization after writing, like FBG3, and type

IIa gratings, under similar conditions to FBG1, are suitable options. For a specific

tuning filter application in a WDM system we will choose one of these two options,

depending on the critical issue for the specific case: the tuning range or the

bandwidth variation.

5. Conclusions

We have studied the thermal behaviour of three different grating types, written in

germanosilicate fibres, and analysed the most relevant characteristics for applications

as optical filters in WDM systems. The experimental results show that type I gratings,

obtained from ending the writing process in a non-hydrogenated fibre before a

potential type IIa grating is obtained, represent the worst option, due to their

considerable bandwidth and group delay variations with temperature. These

variations are almost negligible for the type IIa grating and type I grating in

hydrogenated fibre, the latter with proper stabilization. For a particular tuning-filter

application, we will choose one of the options, the type IIa or the type I in

hydrogenated fibre, depending on the most critical aspect for the specific application.

If the tuning range is the most important aspect, type IIa FBGs represent the best

solution. For systems very sensitive to bandwidth variations, type I FBGs in

hydrogenated fibre must be chosen.

Acknowledgments

This work was financed by the Portuguese scientific program PRAXIS XXI. We are

thankful to INESC Porto – USOE, for allowing the use of their laboratories facilities

to write the gratings used in this study.

References:

[1] T. Eftimov, M. C. Farries, S. Huang, N. Duricic, D. Grobnic, B. Keyworth, J. S. Obhi, “8-channel tunable drop device with thermal tuning for 100 GHz channel spacing”, in Proceedings of 24th European Conference on Optical Communication, 127-128 (1998).

[2] H. Kumazaki, Y. Yamada, H. Nakamura, S. Inaba, K. Hane, “Tunable wavelength filter using a Bragg grating fiber thinned by plasma etching”, IEEE Photonics Technology Letters 13(11), 1206-1208 (2001).

[3] A. Iocco, H. G. Limberger, R. P. Salathé, L. A. Everall, K. E. Chisholm, J. A. R. Williams, I. Bennion, “Bragg grating fast tunable filter for wavelength division multiplexing”, IEEE/OSA Journal of Lightwave Technology 17(7), 1217-1221 (1999).

[4] P. S. André, J. L. Pinto, I. Abe, H. J. Kalinowski, O. Frazão, F. M. Araújo, “Fibre Bragg grating for telecommunications applications: tuneable thermally stress enhanced OADM”, Journal of Microwaves and Optoelectronics 2(3), 32-45 (2001).

[5] H. Patrick, S. L. Gilbert, A. Lidgard, M. D. Gallagher, “Annealing of Bragg gratings in hydrogen-loaded optical fiber”, Journal of Applied Physics 78(5), 2940-2945 (1995).

[6] L. Dong, W. F. Liu, L. Reekie, “Negative index gratings formed by a 193 nm excimer laser”, Optics Letters 21, 2032-2034 (1996).

[7] B. O. Guan, H. Y. Tam, X. M. Tao, X. Y. Dong, “Highly stable fiber Bragg gratings written in hydrogen-loaded fiber”, IEEE Photonics Technology Letters 12(10), 1349-1351 (2000).

[8] G. Brambilla, H. Rutt, “Fiber Bragg gratings with ultra-high temperature-stability”, Optical Fiber Communication Conference 2002, 660-662 (2002).

[9] A. Othonos, K, Kalli, “Photosensitivity in optical fibers”, Chap. 2 in Fiber Bragg Gratings - Fundamentals and Applications in Telecommunications and Sensing, pp. 9-94, Artech House (1999).

[10] R. Kashyap, “Measurement and characterization of gratings”, Chap. 9 in Fiber Bragg Gratings, pp. 409-441, Academic Press , San Diego (1999).

Biographies:

- M. J. N. Lima was born in Lourosa, Portugal, on April 21, 1972. He

received the Licenciatura degree in Electronics Engineering and

Telecommunications, in July of 1994, and the M.Sc. degree in Telecommunications

Systems, in June of 1998, from the University of Aveiro, Portugal. He is currently

pursuing the Ph.D. degree in Electrical Engineering at the University of Aveiro.

His research interests include optical filtering and add/drop multiplexing in DWDM

optical networks.

He is member of the Portuguese Engineering Association (OE), of The Institute of

Electrical and Electronics Engineers (IEEE) and of the Optical Society of America

(OSA).

- A. L. J. Teixeira was born in S. Pedro do Sul, Portugal, on November

17, 1970. He received the Licenciatura degree in Electronics Engineering and

Telecommunications, in July of 1994, and the Ph.D. degree in Electrical Engineering,

in May of 1999, both at the University of Aveiro. Part of the Ph.D. work was

performed at the Institute of Optics, Rochester NY, under the orientation of Prof.

Govind Agrawal.

He is currently an Assistant Professor at the University of Aveiro and researcher at

the Institute of Telecommunications – Aveiro.

His research interests include optical filtering in DWDM systems, wide-band optical

amplifiers, OCDMA and radio over fibre technologies.

- J. R. Ferreira da Rocha received the M.Sc. degree in Telecommunication Systems

and the Ph.D. in Electrical Engineering from the University of Essex, England, in

1980 and 1983, respectively.

He is at present a Full Professor at University of Aveiro, Portugal. Within the

Telecommunications Institute, he is a member of the Management Committee

(Aveiro branch) and national coordinator for the Optical Communications Area. He

has coordinated the University of Aveiro and Telecommunications Institute

participation in various projects included in the following European Union (EU)

Telecommunications R&D Programs: RACE, RACE II; ACTS and IST. He has

published about 170 papers, mainly in international journals and conferences and his

present research interests include modulation formats and receiver design for very

high capacity optical communication systems based on linear and non-linear

transmission and WDM optical networks.

In the past few years he has acted as a technical auditor, evaluator and independent

observer for the evaluation of projects to various EU R&D Programs. He has also

participated in various project evaluation boards set up by the EPSRC (Engineering

and Physical Sciences Research Council), United Kingdom. By invitation of the

ACTS Management Committee, he participated in the ‘Expert Groups on Visionary

Research in Communications’, aiming to create a bridge between the activities

carried out in the 4th and the 5th Framework Programs on EU activities in the field of

research, technological development and demonstration.

- O. Frazão was born in Torres Novas, Portugal. He received his Graduated in

Physics Engineering at the University of Aveiro. From 1997 to 1998 he worked at the

Institute of Telecommunications – Aveiro Plant. In 1999 he joined the INESC Porto –

Optoelectronics and Electronics Systems Unit.

His research interests include applications of fibre Bragg gratings in optical sensors

and optical communications.

- P. S. B. André was born in Luanda, Angola, on April 1971. He

received the Physics Engineering degree in 1996 and the PhD in Physics in 2002,

both from the Universidade de Aveiro, Portugal. In the same year he joined the

Instituto de Telecomunicações – Aveiro as senior researcher.

His current research interests include the study and simulation of optoelectronics

components, fiber Bragg gratings, transparent performance monitoring, multi-

wavelength optical communications systems and networks.

He has author or co-authored of about one hundred research papers and presentations

at scientific conferences.

He is member of the Portuguese Physics Society (SPF) and of The Institute of

Electrical and Electronics Engineers (IEEE).

Figure Captions:

Fig. 1 Measured parameters during grating formation: maximum normalised

reflectivity, Rmax (heavy lines), and mean refractive index perturbation, <∆neff> (hair

lines).

Fig. 2 Measured reflection spectra (heavy lines) and group delay variation (hair lines)

of the studied gratings, FBG1 (a), FBG2 (b) and FBG3 (c), for the minimum (solid

lines) and maximum (dashed lines) temperatures analysed, respectively 30ºC and 75

ºC.

Fig. 3 Measured peak wavelength variation with temperature for the three gratings,

FBG1 (circles), FBG2 (squares) and FBG3 (triangles).

Fig. 4 Measured -3 dB bandwidths of the three gratings, FBG1 (circles), FBG2

(squares) and FBG3 (triangles), for the different temperatures, from 30 ºC to 75 ºC

with step 15 ºC.

Figure 1

0

0.2

0.4

0.6

0.8

1

0 20 40 60 80 100

Time (min.)

Rm

ax

0

0.25

0.5

0.75

1

1.25

(<∆n

eff>

)x10

-3

type IIa type I

FBG1

FBG2

FBG3

Figure 2

-20

-16

-12

-8

-4

0

-0.3 -0.2 -0.1 0 0.1 0.2 0.3

Wavelength deviation (nm)

Ref

lect

ion

spec

trum

(dB

)

40

100

160

220

280

340

Gro

up d

elay

(ps

)

(a)

-20

-16

-12

-8

-4

0

-0.3 -0.2 -0.1 0 0.1 0.2 0.3

Wavelength deviation (nm)

Ref

lect

ion

spec

trum

(dB

)

120

180

240

300

360

420

Gro

up d

elay

(ps

)

(b)

-20

-16

-12

-8

-4

0

-0.3 -0.2 -0.1 0 0.1 0.2 0.3

Wavelength deviation (nm)

Ref

lect

ion

spec

trum

(dB

)

160

220

280

340

400

460

Gro

up d

elay

(ps

)

(c)

Figure 3

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

30 45 60 75

Temperature (ºC)

∆λce

nt (

nm)

FBG1 (13.8 pm/ºC)

FBG2 (13.3 pm/ºC)

FBG3 (13.1 pm/ºC)

Figure 4

0

0.06

0.12

0.18

0.24

0.3

0.36

0.42

0.48

30 45 60 75

Temperature (ºC)

BW

-3dB

(nm

)

FBG1 (0.42 nm±2.35%)

FBG3 (0.31 nm±0.64%)

FBG2 (0.19 nm±25%)