Large-core silica optical fibers are expected to find many uses in the nuclear ... absorption, the other radiation-induced effect, luminescence, is rather poorly ...
IECL TRANSACTIONS ON NUCLEAR SCIENCE, VOL. 47, NO. 3, JUNE 2000
693
Radiation-Induced Absorption and Luminescence in Specially Hardened Large-Core Silica Optical Fibers A.L.Tomashuk’, K.M.Golant’, E.M.Dianov’, O.I.Medvedkov’, O.A.Plaksin’, V.A.StepanovZ, P.A.Stepanov*, P.V.Demenkov*, V.M.ChernovZ,S.N.Klyamkin’ ‘Fiber Optics Research Center, 38 Vavilov St., 117756 Moscow, Russia ’SSC R1;- Institute of Physics & Power Engineering, 249020 Obninsk, Russia
’Moscow State University, 119899 Moscow, Russia
Abstract Radiation-induced ahsorption and luminescence are measured in the visible spectral region in II,-loaded and asdrawn fibers with KU and KS-4V silicas in the core in the process of y-irradiation (60Co-source).The induced absorption in HJoaded fibers is shown to he much lower than in asdrawn fibers, at least up to 1.2 MGy. A ‘blue’ band prevails in the luminescence spectra of all the fibers. Its intensity is greater in a KS-4V fiber than in a KU fiber. It is also somewhat greater in H,-loaded fibers than in unloaded ones. The luminescence lifetime in a KU fiber has been measured under excitation with a pulsed reactor to be 100 ps at 1=488 nm and 60 ps at h=633 nni. It is concluded that the ‘blue’ radioluminescence cannot he attyibuted either to relaxation of the oxygen-deficient center or to Cerenkov emission.
A fabrication technology of radiation-hardened H,containing fibers with a hermetic aluminum coating is reported. Such fibers appear to be the hest candidates for various applications in radiation environments.
I. INTRODUCTION Large-core silica optical fibers are expected to find many uses in the nuclear industry, which include transmission of the plasma emission spectrum in fusion reactors [I] and delivery of optical signals from sensors installed inside or in the vicinity of fusion and fission reactors [2]. Preforms for fiberscopes, which are the key element of viewing systems 131, are produced by the same plasma outside deposition process (POD) 141 as preforms for large-core fibers; therefore, the conclusions concerning the optimum silica in the core and the hardening pre-treatment techniques of rad-hard fibers also apply to rad-hard fiberscopes. Under ionizing radiation, transmission of signals in silica fibers and fiberscopes is affected by radiation-induced absorption and luminescence, the latter being accompanied by Cerenkov emission. Radiation-induced absorption in the nearIR region is sufficiently small for many applications, whereas in the UV region it is too large for a fiber to remain functional even at kilogray doses. In the visible region, which is of particular importance for many applications and which is dealt
with in this paper, radiation-induced absorption can he minimized by optimizing the fiber design and technology and by applying hardening techniques to as-drawn fibers. Recently, I-IJoading of fibers was shown to reduce radiation-induced absorption by an order of magnitude and more at megagray y-doses [5]. H,-loaded fibers with a bigbOH low-Cl KU silica in the core and with a low-OH low-Cl KS-4V silica in the core demonstrated comparable induced absorption, whereas an untreated KU fiber was strongly outperformcd by an untreated KS-4V fiber [SI. It is H,-loaded KS-4V and KU fibers that were selected as the hest candidates for further research. However, the previous experimental results outlined above were based on post-irradiation measurements of induced absorption. Therefore, it was important to compare H,-loaded and unloaded fibers in-situ, during irradiation. The results of such a coniparison experiment are described in Section 11. The basic problem related to practical application of the HJoading technique is the need to prevent out-diffusion of H, molecules from thc fiber over the course of its operation under radiation. Because of the out-diffusion, a two-fold reduction of the H, concentration in the core of a polymer-coated fiber with the core and cladding diameters of 100 and 120 pm, respectively, occurs within 7 days at room temperature. A possible solution might consist in pre-irradiation of an H,loaded fiber [SI. After sach a pre-treatment, the fiber would feature a ninch lower induced absorption under subsequent irradiations even in the absence of H, molecules in the glass. However, a more radical solution is the fabrication technology of H,-containing fibers with a hermetic metal coating, which we report in Section IV. Although there is progress in lowering radiation-induced absorption, the other radiation-induced effect, luminescence, is rather poorly understood. Comparison studies of radioluminescence in different fibers are still lacking, while techniques to reduce radioluminescence are unknown. Because thc dcfcct structure of irradiated H,-loaded and unloaded fibers is different, one may expect that radioluminescence will also be differcnt. In fact, the blue radioluminescence band (- 450 nm) is often attributed to the oxygen-deficient centers, which are converted into H(1)centers by hydrogen [6]. The red luminescence is likely to be
694
due to non-bridging oxygen bole center (NBOHC) with the 630 nm absorption band [7]. NBOHC is also healed by hydrogen [5]. A comparison of radioluminescence in H,loaded and unloaded fibers under continuous y-irradiation with a cobalt source is described in Section 11.
In order to hy to interpret the origin ofradioluminescence, we have measured its lifetime under pulsed reactor irradiation. The results are discussed in Section 111.
11. c
The induced absorption expressed in decibels per meter was calculated as
a = (1/ L ) .IOlog,, (jZ/iJ),
(1)
where L is the length difference in meters between the irradiated parts of the sample and reference fiber; i and j are, respectively, the output signals of the sample and reference irradiation; and ~ ~ fibers under ~ ~ ~ , ~ J ~are, respectively, ~ the output signals of the sample and reference fibers before starting the irradiation.
~OF H ~~ AND uNLoADED J~ ~ FIBERS UNDER 'f-IRRADIATION
The luminescent capacity of the fibers was calculated as
A. Experimental
Four fiber types were used in the comparison experiment: as-drawn fibers with KU and KS-4V silicas in the core and their H,-loaded counterparts, The H,-loading procedure was performed at a pressure of 1.20 GPa and a temperature of 80 "C. The H, concentration in the fibers was determined from the amplitude of the 1.24 pm HZ absorption band in the fiber where loss spectrum [ 8 ] to be 1,1OZocm.'. Note that the silica glass output, densitv is about S.10*2atoms uer cubic centimeter.
'
500 nm,whereas at shorter wavelengths the asdrawn KS-4V fiber features the least absorption. The growth of loss in H,-loaded fibers in the blue region is due to the H(1)center. This center as well as NBOHC and other color centers show up already in the absence of H, molecules in the fibers. Therefore, to achieve maximal transparency of fibers during long-term operation at megagray doses, it is necessary to have as much molecular hydrogen dissolved in the fiber glass as possible.
appears to be larger in the as-drawn KS-4V fiber than in the as-drawn KU fiber; therefore, we cannot attribute it to NBOHC (cf. Figs. 1 and 2). Thus, unfortunately, the above comparison does not provide an explanation for the origin of radiation-induced luminescence and does not suggest a way to suppress this effect either. From the practical standpoint, we must conclude that H,-loading somehow promotes the ‘blue’ radioluminescence.
111. LUMINESCENCE LIFETIME ESTIMATION USING
C. Radiation-Induced Luminescence
PULSED REACTOR EXCITATION
Radiation-induced luminescence remained undetectable Temporal characteristics of radioluminescence in an asuntil a dose of 1 MGy and then grew monotonically with drawn KU fiber were investigated using a pulsed nuclear dose in all the four fibers. At 4.3 MGy, the most intense reactor (pulse half-width of 80 p, neutron fluence per pulse of up to 5.10” n/cm2,neutron energy E > 1 keV, and dose rate ofup to 105 ~ ~ 1 s ) .
-
14.3MGy
I
Under such high-dose-rate excitation, luminescence was observed already during the first neutron pulse (recall that under excitation with a cobalt source luminescence revealed itself only at doses of 1 MGy). By taking into account the temporal shape of the neutron pulse and that of the luminescence signal, we determined the luminescence 100 ps in the blue spectral region relaxation time to be (h=488 nm) and 60 ps in the red spectral region (h=633 nm). The blue luminescence lifetime of the oxygen-deficient centers is known to be much longer (- 30 ms) [12], whereas Cerenkov emission decays practically instantaneously [13]. Therefore, the time constants obtained in our experiments are not so easy to explain. It is possible that the light emission observed is a combination of Cerenkov emission, prevailing near the peak of the neutron pulse, and luminescence of some radiation-excited defects in the fiber network. It should also be noted that the luminescence mechanisms can be different in the case of high-dose-rate and low-dose-rate excitation.
-
i
460
500
600
7
860
wavelength, nm
Figure 3: Luminescent capacity of H,-loaded and unloaded fibers measured under y-irradiation at a dose rate of 4.5 Gyis and a dose of 4.3 MGy.
-
-
IV. METAL-COATED H 2
-
CFIBERS ~ ~
As was mentioned above, a polymer coating is sufficiently transparent for 13, molecules; therefore, it is impossible to hold all the H,molecules inside the fiber and to ensure low induced absorption in the course of long-term operation. A hermetic metal coating would prevent the escape of H, molecules. We have developed a technology for fabrication of H,-containing AI-coated fibers. Fig. 4 shows the induced absorption spectra measured after y-irradiation (1.7 MGy, 5.0 Gyis) in three KS4V fibers drawn from the same preform. Two of these fibers contained the same H, concentration at the moment of starting It is surprising that H,-loading appears to have no the irradiation (5.10’’ cm-’), hut differed in the coating type influence on the structure of luminescence. Moreover, the 480 (polymer or aluminum). The third fiber taken for comparison nm band intensity is larger in H,-loaded fibers than in asdrawn ones. The 480 n m band is also larger in the as-drawn did not contain molecular hydrogen at all. In Fig. 4 we see that KS-4V fiber than in the as-drawn KU fiber, the latter fiber the Al-coated fiber featured a lower induced absorption than showing the least luminescence intensity throughout the the polymer-coated fiber. This is because a noticeable share of visible region. As to the ’red’ luminescence, its intensity H, molecules escaped from the latter fiber during the four
luminescence was observed at the output of the H,-loaded KS4V fiber (-10.’’ W). Fig. 3 gives a comparison of the luminescent capacities of the fibers measured at a dose of 4.3 MGy and calculated by (2). A blue luminescence is dominant in all the fibers. In addition to a 480 nm band, a smaller satellite band at about 430 nm reveals itself. Lastly, one can also see a ‘red’ luminescence, which is several orders of magnitude less intense.
~
~
~
ACKNOWLEDGEMENTS We are grateful to SAVasiliev and Yu.V.Karavanin for their help in staging the experiments. Work by the first author was supported by the Russian Foundation for Basic Research (project No. 99-02-16856).
REFERENCES [I] wavelength, gm
Figure 4: Induced absorption measured in different KS-4V fibers after y-irradiation with a cobalt sourcc ( I .7 MGy, 5.0 Gyis).
O.Deparis, P.Megret, M.Decreton, M.Blondel, K.M.Golant. A.L.Tomashuk. “ Rad-hard fibres for diagnostics of experimental fusion reactors”, in Diagnostics f o r Experimental Thermonuclear Fusion Reactors 2, edited by Stott et al., Plenum Press, New York, 1998, pp. 291-295.
days of irradiation and therefore did not participate in healing radiation-induced color centers.
[2] S.Coenen, M.Decreton, “Feasibility of optical sensing for robotics in highly radioactive environments”, IEEE Transactions on Nuclear Science, vol. 40, No. 4, pp. .. 851856 (1993). V. CONCLUSION [3] K.Obara, S.Kakudate, K.Oka, E.Tada, Y.Morita, MSeki, In the process of y-irradiation with a “CO source, “Development of optical components for in-vessel viewing radiation-induced absorption is strongly suppressed in fihers systems used for fusion experimental proc, spIE, containing H, molecules. The in-situ behavior of H,-loaded 2425, pp, 115-122 (1994), fibers h m e d out to he in a good agreement with the preceding [4] A.S.Biriukov, E.M.Dianov, K.M.Golant, R.R.Khrapko, results obtained ex-situ [SI. A.V.Koropov, A.N.Perov, A.V.Shakbanov, S.A.Vasiliev, ccSynt~lesisof fluorine.doped silica ,,lass by In the course of irradiation of H,-containing fibers, the of an H(1)-center as well as other color centers start to develop as outside deposition technique using a microwave plasma soon as the II, reservoir in the fiber glass has been exhausted. sovietLighhvnve c ~ 3, N pp, ~ Therefore, to achieve the maximal hardening effect, it is (1993), necessary to have as much molecular hydrogen dissolved in the fiber glass as possible. Secondly, it is necessary to prevent [’I A.L.Tomashuk, B.M.Dianov, K.M.Golant, the escape of H, molecules from the fiber, The problem of A.O.Rybaltovskii, “y-radiation-inducedabsorption in pureholding HZ molecules inside the fiber is solved by the silica-core fibers in the visible spectral region: the effect of H,-loading”, IEEE Transactions on Nuclear Science, vol. of H,-containing fibers with a hemetic metal coating, which was reported in this paper. Fibers and 45, 3, Part 3, pp. 1576-1579 (1998). fiberscopes fabricated by this technology with KS-4V or KU [6] A.V.Amossov, A.O.Rybaltovsky, “Radiation color center silica in the core appear to be the best candidates for plasma formation in silica glasses: a review of photo- and thermodiagnostics in fusion reactors and for fiber-optic viewing chemical aspects of the problem”, 1 Non-Crystalline Solid.v, vol. 179, pp. 226-234 (1994). systems. The radiation-induced luminescence spectrum in H,- [7] E.J.Friebelc, D.L.Griscom, M.J.Marrone, “The optical absorption and luminescence hands near 2 eV in irradiated loaded and unloaded fibers is similar, a ‘blue’ band being dominant. We have found that the blue hand intensity is larger and drawn synthetic silica”, J. Non-Crystalline Solids, vol. in H,-loaded fibers; however, the mechanism of the H, effect 71, pp. 133-144 (1985). on the blue luminescence intensity as well as the origin of the K,Nog,,chi, ~ , ~ h i N b ~, u~ ~ ,y,Negishi, ~ ~ ~ ~ ,c ; L ~ ~ ~ increase for optical fibers exposed to hydrogen blue luminescence itself are still to be understood. The blue luminescence lifetime measured under pulsed reactor atmospllere”,J , Lightwave Techno[.,vol. 3, No. 2, pp. 236excitation amounted to 100 ps, a value which is much smaller 243 (1985), than the lifetime of the blue luminescence of excited oxygendeficient center and much greater than the lifetime of [9] DLGriscom, K.M.Golant, A.L.Tomashuk, D.V.Pavlov, Yu.A.Tarabrin, “y-radiation resistance of aluminum-coated Cerenkov emission. It appears that further research should be all-silica optical fibers fabricated using different types of aimed at assessing the luminescence effect on the silica in the core”, Appl. Phys. Lett., vol. 69, pp. 322-324 of each specific fiber-optic system as a whole and at finding (1996). possibilities to use a chopped light signal in combination with synchronous detection.
~
698
[ 101 O.Deparis, P.Megret, M.Decreton, M.Blonde1, ”Gamma
radiation tests of potential optical fiber candidates for fibroscopy”, IEEE Trans. Nucl. Sei.,vol. 43, pp. 30273031 (1996).
[I 11 D.L.Griscom, “Visihleilnfra-Red Absorptioll Study in Fiber Geometry of Metastable Defect States in I-Iigh-purity Fused Silicas”, Proc. 13”’International Conf: on Defects in Insulating Mnterinls, Wake Foresf University, WinstonSalem, NC, July 1996.
[ 121 R.Tohmon, YShimogaichi, H.Mizuno, Y.Ohki, K.Nagasawa, Y.Hama, “2.7-eV luminescence in asmanufactured high-purity silica glass”, Phys. Rev. Lett., vol. 62, N 12, pp. 1388-1391 (1989). [ 131 W.Schncidcr, U.Babst, “Radiation-induced light emission in silica core fibers”, Proc. SPIE, vol. 506, pp. 189-195 (1984).
