Effect of MGy dose level γ-radiation on the parameters of fbgs written in a Ge-doped silic

European Workshop on Optical Fiber Sensors, Scotland, Peebles, 8-10 July 1998

Effect of MGy dose level γ-radiation on the parameters of FBGs written in a Ge-doped

silica fiber

A.I.Gusarov*, A.Fernandez Fernandez ?, F.Berghmans ?

, O.Deparis*, Y.Defosse*, P.Mégret*,M.Décreton ?, M.Blondel*

* Faculté Polytechnique de Mons (FPMs), Service d’Electromagnétisme et de

Télécommunications, 31 bd. Dolez, B-7000 Mons, fax: (+32 65) 37.41.99? Belgian Nuclear Research Center (SCK ?CEN), 200 Boeretang, B-2400 Mol

1. Introduction

Optical fiber sensors (OFS) are developed to replace conventional electro-mechanical sensing systems, which are well established, have proven their reliability and can be manufactured at low cost. Advantages of OFS are resistance to electromagnetic interference, intrinsic safety,mechanical simplicity and small size, the possibly of high sensitivity and multiplexing capabilities. These features make OFS an interesting alternative for application in the nuclear industry. However, the feasibility of using OFS in a radiation environment still needs due assessment. For example, Berghmans et al. [1] showed that γ-radiation at doses below 100 kGy can result in the failure of commercial fluorescence and Fabry-Perrot temperature sensors,whereas semiconductor absorption temperature sensors did not show any degradation for doses in excess of 450 kGy.

A fiber Bragg grating (FBG) is an optical filter constituted by a single-mode fiber in which the core refractive index is periodically modulated along its axis [2,3]. It can be produced by exposing a photosensitive optical fiber to an ultraviolet laser beam, of which the intensity varies in space. The result is a perturbation of the effective refractive index. The mechanisms of the FBGs formation are still not well established, but it is generally accepted that the underlying physic is related to the generation of defects by UV-light.

The reflection spectrum of such a filter is peaked around the Bragg wavelength λB defined by the Bragg condition: λB = 2n eff Λ, where n eff is the refractive index of the guided mode and Λ is the grating pitch. The use of a FBG as a temperature sensor is based on the dependence of λB on the temperature T. For a restricted temperature interval this dependence can be well approximated by a linear expression:()()()000T T T T B B ?+=αλλ,

where α0 is the FBG temperature sensitivity at T = T 0. If the values of λB (T 0) and α0 are given,the temperature of the FBG can be easily calculated.

The most prominent effect of ionizing radiation on an optical fiber is an increase of the fibre’s attenuation. An advantage of a FBG-sensor is that the information about the measured parameter (temperature) is wavelength-encoded and should therefore be insensitive to radiation-induced losses. Radiation effects in a FBG are related to the generation of radiation defects. Such defects change the glass parameters resulting in a change of the Bragg wavelength λB (T 0) → f(D,T 0).The application of FBGs for temperature sensing in a radiation environment therefore depends on to what extent λB (T 0) and α0 are influenced by radiation.

To the best of our knowledge there are only a few publications on this subject [4-6]. For a Ge-doped silica fiber the shift of the resonance can be as high as 0.1 nm at a dose of 0.5 MGy [4,5],corresponding to the error in temperature estimation (about 10 °C) not acceptable for most applications. For a N-doped fiber, which was shown to possess a high radiation hardness as far as transmission in the visible region of the spectrum is concerned, the shift can be even higher

[6]. The value of α0 however is expected not to change very much under radiation.

2. Experimental procedure

The FBGs with λB around 1546 nm were written at the FPMs in Mons in a 10 mol.% Ge-doped silica fiber using a phase mask. Their main characteristics are listed in Tab. 1.

Tab. 1 FBGs parameters. L is the FBG length, ?λ? is the FBG bandwidth, R is the FBG reflectivity. The total fluence of pulsed excimer laser light during the writing process is also given.

FBG

#1#2#3#4#5#6#7L, cm

0.60 1.70 1.70 1.70 1.70 1.700.75?λ?, nm

0.1820.1400.1200.0970.1690.1130.176R, dB

1.80 3.03

2.85 1.057.40 2.31

3.16Total fluence, kJ/cm 241.53

4.927.77.956.051.92

5.8

The basic optical setup used to characterize the FBGs in situ under radiation is shown in Fig. 1.

A light emitting diode (LED) is used as a broad-band source. Two optical switches (OS) and a 3 d

B coupler allow to characterize all FBGs both in transmission and reflection. The spectra

were recorded by an optical

spectrum analyzer (OSA ANDO AQ-6215A) with a sampling interval of 2.5 pm. The signal

level was -55 dBm and the noise

level was below -75 dBm.

For each spectrum, λB was

determined by fitting the part of the spectrum near λB (41 points)

with a gaussian function. Such a procedure allows the estimation of λB with a pm-level accuracy.The 3 dB width of the Bragg

resonance (?λ?) was derived

from the reflectivity spectra and the amplitude or the maximal reflectivity (R) was estimated based on the transmission data.

The overall performance of a FBG-based sensor is defined by both the sensing head (the actual FBG) and the light transmitting fiber. In [6] radiation-induced transmission losses in the Ge-doped

fiber have made it impossible to measure the FBGs characteristics for radiation levels above 100 kGy. In general, a degradation of the signal to noise ratio means a lower accuracy of the central wavelength (i.e., temperature) estimation and, if the signal drops to the noise floor, no measurement is possible at all. To avoid this problem we spliced short pieces of the “radiation sensitive” Ge-doped fiber, used for FBGs writing, with a radiation resistant Ge-doped fiber with a very low P impurity content. This allowed to run the experiment for doses above the MGy level. To distinguish between the radiation influence on the FBGs and on the transmitting optical fiber, a reference fiber without FBG was used.

The reference fiber and 7 FBGs were inserted in a specially designed temperature controlled oven made as an aluminum cylinder of 100 mm length and 80 mm diameter with 9 axial channels. A heating wire, closely coiled around the cylinder, was connected to a PID-type controller, which allows temperature control with the accuracy of 0.1 °C. The feedback signal to the controller was generated by a thermocouple inserted into a 2 mm diameter axial channel with the channel axis being parallel to the cylinder axis. The fibers were pulled through other FBG Fig. 1 Basic optical set-up

channels located on the same distance from the cylinder axis as the channel with the thermocouple and the FBGs were located in the middle of the aluminum piece. Due to the symmetrical location of the channels and the good heat conductivity of aluminum, the thermocouple returns the same temperature as that one seen by the FBGs.

The measurement campaign included three steps: pre-radiation, radiation and post-radiation tests. For the irradiation, the container with the FBGs was installed into 60Co γ-irradiation underwater facility RITA at SCK ?CEN with a dose rate of 3 kGy/h [7]. The measurements were performed in situ during 15 days with a total accumulated dose in excess of 1 MGy. During the first two days the temperature of the FBGs was maintained at 35 °C. The value of 35oC is assumed high enough to avoid the effect from gamma heating.

After two days the periodic temperature modulation was applied to allow for calculation of the temperature sensitivity coefficient α0. One period of that temperature modulation consists of the following steps. After a measurement set the temperature was increased up to 45 °C and kept fixed until the end of the next measurement set. Then the temperature was decreased down to 40 °C and, further down to 35 °C. Cooling of the system results from natural heat transfer from the aluminum cylinder to the environment.

3. Results and discussion

The results of the radiation test show that there exists an effect related probably with the radiation environment. Fig. 2 compares the response of FGB #2 with the oven temperature. After the initial period of two days, the temperature modulation induced periodic peaks. The

020406002

4

6

8

1012 - T T , °C t, h 0.000.020.040.060.080.10

0.12

- ?λB ?λB , n m Fig. 2 Initial part of radiation test :variation of λB with

the time for the FBG #2 and variation of

temperature.from the temperature modulation data during the irradiation test .

A slow increase of λ

B is observed at the start. The change of λB is about 25 pm and saturates after the first day of irradiation (at the dose about 100 kGy). Subsequently, λB remains unchanged within the accuracy of our measurements. It is necessary to note that there is no detectable difference in behavior of the seven FBGs written under different conditions. Data in Fig. 2 agree with the expected trend. However, the change of 25 pm seems rather high. As far as such in-situ test is difficult due to high changes in the ambient temperature around the measuring set-up we can not rule out the possibility that the observed shift is in part due to some other effects, not radiation induced. This will be studied in the following experiments. However, we can state with no doubt that the shift in λB due to radiation is definitely less than 25 pm.

The variation of the FBGs temperature (Fig. 2) allowed to calculate the temperature sensitivity

coefficient α0 (Fig. 3). For each data of Fig. 3, sensitivity is calculated by α0 = (Σαi)/3 with αi =?λB,i/?T i and ?T i= +10, -5 and -5 °C (three steps of the temperature modulation). The values of the temperature sensitivity coefficient (averaged on all data and on all FBGs) are: before irradiation α0 = 10.63 ± 0.21 pm/°C; during the irradiation test: α0 = 10.67 ± 0.30 pm/°C; after irradiation: α0 = 10.54 ± 0.20 pm/°C. Fig. 3 shows also that there is no systematical trend in the variation of the sensitivity with dose accumulation.

Moreover, the results show that radiation does not influence the shape of the Bragg resonance:?λ? and R are constant within 5% limit during the radiation test (not shown here). This agrees with the idea that UV- and γ-radiation generate the same radiation defects via an ionization-related process. The spatial distribution of γ-radiation induced defects is uniform and these defects affect only the mean value of the effective refractive index (i.e. λB), but not the modulation amplitude.

4. Conclusions

For the first time, parameters of fibre Bragg gratings exposed to γ-radiation were measured in situ up to very high doses (> 1 MGy).The experimental results allow to draw several

conclusions.

1.The temperature sensitivity coefficient α0is not affected by radiation with the accuracy of

3%.

2.The amplitude and the width of the Bragg resonance are also unchanged under the γ-radiation.

This means, that the radiation does not influence the shape of the reflection and transmission spectra.

3.The change of the Bragg wavelength (λB) as a result of irradiation is not higher than 25 pm

and saturates for the dose of 0.1 MGy.

Consequently, FBG-based temperature sensors are probably capable to maintain the required performance even in a MGy dose level radiation environment.

A.I.Gusarov is supported by SCK?CEN (collaboration contract with FPMs #KNT 909337901). O.Deparis and P.

Mégret are supported by the Inter-university Attraction Pole (IAP IV/07) of the Belgian Government (SSTC). References

1.Berghmans, F., V. Vos, and M. Decréton. Evaluation of three different optical fibre temperature sensor

types for application in gamma-radiation environment. in RADECS'97. 1997. Cannes, France.

2.Hill, K.O., y. Fujii, D.C. Johnson, et al., Photosensitivity in optical fiber waveguides: application to

reflection filter fabrication. APL, 1978. 32(10): p. 647-9.

3.Meltz, G., W.W. Morey, and W.H. Glenn, Formation of Bragg gratings in optical fibers by a transverse

holographic method. OL, 1989. 14(15): p. 823-25.

4.Ferdinand, P., S. Magne, V. Marty, et al. Optical fiber Bragg gratings sensors for structure monitoring

within nuclear power plants. in Optical Fibre Sensing and Systems in Nuclear Environment. 1994. Mol, Belgium: SPIE, V.2425, pp.11-21

5.Niay, P., P. Bernage, M. Douay, et al., Behaviour of Bragg gratings, written in germanosilicate fibers,

against g-ray exposure at low dose rate. IEEE PTL, 1994. 6(11): p. 1350-2.

6.Vasiliev, S.A., E.M. Dianov, K.M. Golant, et al. Performance of Bragg and long-period gratings written in

N- and Ge-doped silica fibers under g-radiation. in RADECS'97. 1997. Cannes, France.

7. S. Coenen, J. Vermunt, L. Van den Durpel, M. Decréton and A. Rahn, "Gamma Irradiation Facilities for

Assessment of Advanced Instrumentation - New Reactor Design and Plant Life Extension Increase their Need", Research Facilities for the Future of Nuclear Energy, H.A. Abderrahim, Proceedings of an ENS Class 1 Topical Meeting, World Scientific Publishing Co., pp. 382-391, 1996.