Dispersion and refractive index measurement for Ge, B-Ge doped and photonic crystal fibre following irradiation at MGy levels

Transcription

1 INSTITUTE OFPHYSICS PUBLISHING Meas. Sci. Technol. 15 (2004) Dispersion and refractive index measurement for Ge, B-Ge doped and photonic crystal fibre following irradiation at MGy levels MEASUREMENTSCIENCE AND TECHNOLOGY PII: S (04) W N MacPherson 1,RRJMaier 1,JSBarton 1,JDCJones 1, A Fernandez Fernandez 2, B Brichard 2, F Berghmans 2, J C Knight 3, P StJ Russell 3 and L Farr 4 1 School of Engineering and Physical Sciences, Heriot-Watt University, Edinburgh EH14 4AS, UK 2 SCK CEN, Belgian Nuclear Research Centre, Boeretang 200, B-2400 Mol, Belgium 3 Optoelectronics Group, Department of Physics, University of Bath, BA2 7AY, UK 4 BlazePhotonics Ltd, University of Bath Campus, Claverton Down, Bath, BA2 7AY, UK Received 24 December 2003 Published 19 July 2004 Online at stacks.iop.org/mst/15/1659 doi: / /15/8/039 Abstract We have investigated the effects of ionizing radiation-induced changes on dispersion and refractive index in photonic crystal fibre, boron co-doped photosensitive fibre and standard SMF-28 fibre up to a level of 7 MGy of 60 Co γ -radiation. We have been unable to observe any changes within an experimental error of 1% for the dispersion value and < for the refractive index. Keywords: dispersion, optical fibre, refractive index, radiation (Some figures in this article are in colour only in the electronic version) 1. Introduction Fibre optic instrumentation for data communications and sensor technology is under investigation for use in nuclear environments such as nuclear waste storage [1], research reactors [2], high energy physics research instrumentation [3, 4] and fusion research [5], with significant benefits over conventional electronic instrumentation operating in environments incompatible with electronic or semiconductor technology [6]. Therefore there is a requirement for optical fibres that can be used in high radiation environments. It is important that the fibre attenuation is not significantly increased as a result of irradiation as this would directly influence the operational lifetime of data communication links and sensor systems. Radiation hard optical fibres, based for example on sol gel derived materials, have been shown to be able to support data links up to MGy levels [6]. Such fibre is typically drawn as multimode fibre. However, fibre optic sensors based on interferometric techniques or, to a lesser extent, fibre gratings, require single mode fibre links to maintain phase information. Radiation induced attenuation measurements in a variety of fibres have been conducted previously, with significant losses at the radiation levels used in our study described below, but no attempt was made to study the effects on dispersion and refractive index. Using fibres as interferometric sensors will require a more detailed understanding of the underlying changes in dispersion and refractive index. The purpose of the measurements reported in this paper is to investigate changes of refractive index and dispersion due to irradiation for a number of fibre samples. In particular we wish to consider SMF-28 (as a benchmark), photosensitized fibre (for fibre Bragg grating applications) and photonic crystal fibres for novel interferometric sensing applications. The photonic crystal fibre (PCF) used in this experiment is made from pure fused silica, the core being surrounded by an array of air channels which form a region of depressed refractive index relative to the core area [7]; a micrograph of /04/ $ IOP Publishing Ltd Printed in the UK 1659

2 W N MacPherson et al Scanning Mirror Sig BBS HeNe BS Lens Test Fibre HeNe Reference Figure 1. Cross section of PCF fibre (left: illustrates end face of the fibre with air-hole structure surrounded by solid silica protective layer, OD 125 µm; right: SEM image illustrates air-hole cladding and solid fused silica core.) the PCF used in this experiment is shown in figure 1. The dimensional design, i.e. the diameter, location and density of the air holes, governs the waveguiding characteristics of the PCF. It is possible to design PCFs with a wide wavelength range of single-mode operation, restricted only by the material properties of bulk fused silica [8], and to manufacture PCF with dispersion that is significantly different to that of conventional fibre [9]. The PCF used in this experiment had an air hole spacing of 3.1 µm (hole diameter 1.5 µm) and for mechanical strength the PCF structure was surrounded by a solid fused silica cladding with an outer diameter of 125 µm. It had a polymer coating ( 250 µm diameter), forming a fibre that was rugged and as straightforward to handle as conventional SMF28. The detailed photo-chemistry of radiation-induced effects in fused silica fibre material is often complex to describe and strongly depends on the glass dopant content. However, we can assert that the best radiation resistance is generally achieved in un-doped pure SiO 2. It is expected that the absence of dopants in PCF will result in optical parameters that are more stable when irradiated than in standard doped fibres. In addition to this, the cladding optical characteristics depend partly upon the properties of air which will be unaffected by irradiation. 2. Experimental details We investigated the radiation-induced changes in dispersion and refractive index in PCF and for comparison in standard Ge-doped telecom single mode fibre (Corning SMF-28). We co-located a set of boron co-doped photosensitive fibres (Fibercore PS1250/1500 and Nortel) with the PCF and SMF-28 samples. Previous work on UV-written fibre Bragg gratings in several fibre types [6, 10] indicates that some gratings show saturation in the radiation-induced shift of the Bragg wavelength. It is not clear if the reported observed drift is an underlying radiation-induced change in the fundamental fibre properties or an additional change to the photochemistry of the refractive index modulation forming the Bragg grating structure Dispersive Fourier transform spectroscopy Due to physical space restrictions in the radiation test facility, and in order to reduce the amount of each fibre required, it was necessary to use short ( 100 mm) lengths of fibre for this test. Figure 2. Typical configuration for DFTS: BS is a glass beamsplitter, BBS is the broadband source, HeNe laser (frequency stabilized) for optical path length calibration measurement. For such lengths of fibre, the dispersion cannot be measured using traditional pulse spreading techniques, so instead we use dispersive Fourier transform spectroscopy [9] to obtain both the effective refractive index and fibre dispersion (see schematic in figure 2). In this configuration the fibre is placed in one arm of a scanning Michelson interferometer illuminated with a low coherence source. The interferometer is balanced when the scanning path length is equivalent to the path length to the front face of the fibre, and an interferogram is recorded. A second interferogram is observed when the scanning path length is equivalent to the path length to the back face of the fibre; in this case the interferogram is modified by the fibre dispersion. Comparing these interferograms allows the dispersion to be measured [9]. The separation of these interferograms gives the effective refractive index. Typical plots of the interferogram are given in figure Sample preparation Fibres were cleaved to identical lengths using two cleavers (Fujikura) clamped to an optical table with the fibre stretched through the fibre guides, tensioned with a small mass (20 g) via a pulley. All sample fibres were cleaved to a length of mm ± 0.1 mm. The sample lengths were verified using a travelling microscope with 0.05 mm resolution. To ensure that the fibres were straight during this length measurement the samples were placed in a 300 µm wide slot machined into a plate of glass, this was sufficient to hold the fibre without imposing any loading on the fibre that might affect its length. Following characterization of fundamental dispersion and refractive index using the DFTS instrument (see section 3), samples were inserted into slots machined into borosilicate glass plates ( mm), figure 4. The slots were 260 and 300 µm wide and were cut to a depth equal to their width. The PCF fibres used here had a jacket OD of 270 µm and were inserted into the 300 µm channels. All other fibres were inserted into the 260 µm channels. The slight natural fibre curl is sufficient to prevent axial slippage of the fibre samples in the narrow channels thus protecting the cleaved fibre ends. A cover plate without slots closes the assembly. To avoid a dose build-up effect at the surface of the fibre, it is preferable to surround them with a homogeneous material of similar density: the size of single mode (SM) fibres is often small compared with the range of secondary particles generated by γ radiation interactions with the fibre material. Burying the fibre samples inside a glass carrier plate assists in the analysis of the deposited dose, at least for SM fibre. 1660

3 Dispersion and refractive index measurement for Ge, B-Ge doped and photonic crystal fibre Figure 3. Interferogram plots obtained from DFTS experiment. Top: total data set. Bottom: interferograms extracted from data set for front face (left) and back face with fibre dispersion (right). The back face interferogram is broadened due to the fibre dispersion. 3mm thick glass plates with 260 /300 µm slots 60 Co γ sources tie rods 150mm 60mm Figure 5. Photograph of Brigitte irradiation facility at SCK CEN showing the ten Co γ sources. The vacant position inside the ring of Co γ sources is the location in which the sample canister is placed. Figure 4. Fibre carrier: left, schematic; right, photograph after irradiation showing darkened glass plates. This scheme will ensure that γ induced secondary electron equilibrium (SEE) is achieved across the diameter of a SM fibre [11]. In the case of PCF, the core is by construction inhomogeneous in density in such a way that strict SEE cannot be verified, hence the dose deposited in PCF may be only a fraction of the dose deposited in standard SM fibre. It is nevertheless valid to study the results for SM and PCF in a comparative fashion and draw qualitative conclusions from these tests. It is therefore possible that a PCF fibre will have an added benefit where not only the pure SiO 2 core is expected to show low sensitivity to radiation, but also that the structure will reduce cross section for the generation of damage causing secondary particles and possibly shift the statistical distribution of the stopping distance for these particles out of the core region of the fibre. In total five plate assemblies (A to E), each containing three Corning SMF-28, three PCF (fabricated by Blaze Photonics/Bath University), plus additional samples of two types of Bo Ge co-doped photosensitive fibre (Fibercore PS1250/1500 and Nortel), were prepared at HWU and shipped to Mol for installation and irradiation in the Brigitte facility. 1661

4 W N MacPherson et al Table 1. Radiation dose for plate assemblies A to E. Fibre carrier plate A B C D E Dose (kgy (H2 O)) Table 2. Experimental results for dispersion and group refractive index for irradiated samples. Dispersion is in ps nm 1 km 1 ; refractive index values are in brackets. Dose Nortel Fibrecore (kgy) SMF28 PCF (B-co-doped) (B-co-doped) (1.4686) (1.4735) (Plate A) (1.4682) (1.4727) (1.4680) (1.4707) (Plate B) (1.4681) (Plate C) (1.4683) (1.4729) (1.4681) (Plate D) (1.4687) (1.4711) (Plate E) (1.4684) (1.4735) (1.4679) The samples were held at a constant temperature of 55 C in dry nitrogen atmosphere. 3. Experimental results Figure 6. Left: exterior of sample canister for use in Brigitte. Right: the interior of the canister, with multiple shelves for sample storage. The canister is sealed against water ingress and maintained at constant temperature during the experiment, with a constant pressure dry nitrogen atmosphere Irradiation facility In this work the underwater gamma irradiation facility BRIGITTE of SCK CEN (Belgian Nuclear Research Centre) was used to irradiate the fibre samples (see figure 5 and [12]). The test chamber consists of a pressurized canister (figure 6) in which the samples are placed. This is then lowered into the irradiation facility such that the samples are surrounded by the ten 60 Co sources. In this configuration, the maximum dose-rate available is 25 kgy h 1. The sample canister contains the test samples and the irradiation starts when the can is immersed in the irradiation position, 8 m below the water surface. In the experiments reported here, the five glass plates (see section 2.2) were fixed at a position where the γ dose-rate was 20 kgy h 1. Plates were removed from the irradiation facility after a predetermined time interval and the total dose received for each plate was recorded. Table 1 lists the total irradiation dose received by each plate. The dose-rate fluctuation as a function of position in Brigitte is better than 15% when considering plate to plate variation and negligible within samples belonging to the same glass plate. The uncertainty in the dose measurement is better than 10%. Measurement of the separation of the two interferograms in conjunction with the known sample length allows the group refractive index at 1550 nm to be measured. This wavelength is chosen since it is a popular choice for fibre communication systems, and many fibre sensors currently operate around this wavelength due to the availability of low cost components. The central position of front and back face interferograms was defined as the point with maximum signal intensity. For interferogram spacing of samples, and centre point error determination of 50 samples, this gives a measurement error better than 0.02%, which indicates that the error of 0.1% in the sample length dominates, thus giving a measurement accuracy of 0.1%. In this case we can measure refractive index to Although it is less straightforward to determine the accuracy of the dispersion measurement, a repeatability 0.08 ps nm 1 km 1 was nevertheless achieved. DFTS measurements of dispersion and refractive index were repeated following irradiation. The time span between removal of the fibre samples and analysis varied between 3 days and 2 weeks partially due to shipping of fibre samples between Mol and HWU. The results (see table 2 and figures 7 and 8) do not show any systematic changes above our noise limits for any of the fibres tested. This is a significant result because our measurement sensitivity is similar to the level of refractive index change/dispersion change that could pose significant difficulties for FBG and interferometric sensors. These results are in general agreement with refractive index measurements made in situ in which the refractive index varied by <

5 Dispersion and refractive index measurement for Ge, B-Ge doped and photonic crystal fibre -1-1 km Dispersion/ps nm error of dispersion measurement ~0.08 ps nm-1 km-1 (equivalent to symbol size) B/Ge co-doped (Fibrecore) Irradiation dose/kgy (H2 0) PCF fibre SMF-28 B/Ge co-doped (Nortel) Figure 7. Results from experimental measurement of dispersion post-irradiation for PCF, SMF-28 and photosensitized SM fibres. Refractive index Irradiation dose/kgy (H 2 0) PCF B/Ge (Fibercore) B/Ge (Nortel) SMF-28 Figure 8. Results from experimental measurement of refractive index at 1550 nm post-irradiation for PCF, SMF-28 and photosensitized fibres. for similar radiation doses [13]. It is important to note that our experimental results are for measurements several hours after the samples have been removed from the irradiation facility, as opposed to in situ measurements. This is significant because in situ measurements of transmission loss in conventional fibre [14] indicate that once samples are removed from the radiation environment relaxation effects take place, which allow the fibre to partially recover from the effects due to the irradiation. This may account for the slight differences between our observations reported here and the observations in [13]. 4. Conclusions We have investigated possible radiation-induced changes in dispersion and refractive index in four different single-mode fibre samples at 1550 nm up to a γ -radiation level of 5 and 7 MGy at a dose rate of 20 kgy h 1 using a 60 Co source at the Brigitte facility at SCK CEN at Mol. Standard Corning SMF-28, two samples of boron co-doped photosensitive fibre and a PCF sample were tested. We are so far unable to detect any systematic trend indicating correlation between irradiation level and a change in dispersion in any of the fibre samples under investigation. An error analysis showed that the maximum error in the determination of the dispersion of the fibre samples is <1% and that the effective refractive index at 1550 nm could be determined to better than It must be noted that the absorbed radiation dose for PCF is very difficult to establish due to the absence of equilibrium conditions for secondary electrons in the core area of the fibre due to the presence of air holes surrounding the core. However, it is expected that the absorbed dose will be significantly lower than that observed in a conventional SM fibre at an identical location. Further experiments are planned to test this. Fibre samples used in the experiments were too short to allow meaningful attenuation measurements but a few longer length of fibre co-located with the dispersion analysis samples show low attenuation. However, on-line measurements are planned to verify this. From the data presented here we therefore have to conclude that mechanism(s) leading to radiation-induced transmission losses do not affect the refractive index or dispersion of conventional single mode fibres with Ge-doped core and B co-doping for increased photosensitivity. The dispersion and refractive index of PCF fibre at 1550 nm have been equally shown to be insensitive to γ irradiation up to a level of 7 MGy. Acknowledgments We would like to thank Fibercore Ltd and Dr Lin Zhang of Aston University for providing the samples of photosensitive fibre. The irradiation work involved in this study is partially financed by the European Fusion Development Agreement. WM would like to acknowledge the UK EPSRC for provision of funding under the Advanced Fellowship Scheme. References [1] Borgermans P 2001 Safety and operational monitoring of nuclear waste repositories with fibre optics sensing systems EC FP5-Euratom R&D project, 2/ , SCK CEN [2] Jensen F, Kakuta T, Shikama T, Sagawa T, Narui M and Nakazawa M 1998 Optical measurements of high temperatures for material investigations in nuclear reactor environments Fusion Eng. Des [3] Henschel H, Koerfer M, Wittenburg K and Wulf F 2000 Fiber optic radiation sensing systems for TESLA TESLA Report no [4] Vasey F, Arbet-Engels V, Batten J, Cervelli G, Gill K, Grabit R, Mommaert C, Stefanini G and Troska J 1998 Development of radiation-hard optical links for the CMS tracker at CERN IEEE Trans. Nucl. Sci [5] Brichard B, Van Uffelen M, Fernandez-Fernandez A, Berghmans F, Decréton M, Hodgson E, Shikama T, Kakuta T, Tomashuk A, Golant K and Krasilnikov A 2001 Round-robin evaluation of optical fibres for plasma diagnostics Fusion Eng. Des [6] Fernandez-Fernandez A, Brichard B, Berghmans F and Décreton M 2002 Dose-rate dependencies in gamma-irradiated in-fibre Bragg gratings IEEE Trans. Nucl. Sci [7] Knight J C, Birks T A, Russell P StJ and Atkin D M 1996 All-silica single-mode optical fiber with photonic crystal cladding Opt. Lett Knight J C, Birks T A, Russell P StJ and Atkin D M 1997 Opt. Lett (erratum) [8] Birks T A, Knight J C and Russell P StJ 1997 Endlessly single-mode photonic crystal fiber Opt. Lett

6 W N MacPherson et al [9] Gander M J, McBride R, Jones J D C, Mogilevtsev D, Birks T A, Knight J C and Russell P StJ 1999 Experimental measurement of group velocity dispersion in photonic crystal fibre Electron Letts [10] Gusarov A, Fernandez Fernandez A, Vasiliev S, Medvedekov O, Blondel M and Berghmans F 2002 Effect of gamma-neutron nuclear reactor radiation on the properties of Bragg gratings written in photosensitive Ge-doped optical fibre Nucl. Instrum. Methods B [11] ASTM E Standard terminology relating to radiation measurements and dosimetry [12] Fernandez Fernandez A, Ooms H, Brichard B, Coeck M, Coenen S, Berghmans F and Décreton M 2002 Test facilities at sckcen for radiation tolerance assessment: from space applications to fusion environments Proc. IEEE RADECS 2002 (Padova) [13] Fernandez Fernandez A, Brichard B and Berghmans F 2003 In situ measurement of refractive index changes induced by gamma radiation in germanosilicate fibers IEEE Phot. Technol. Lett [14] Liu D T H and Johnston A R 1994 Theory of radiation-induced absorption in optical fibers Opt. Lett