Radiation Effects on Silica-Based Preforms and Optical Fibers—I: Experimental Study With Canonical Samples

Abstract

Prototype samples of preforms and associated fibers have been designed and fabricated through MCVD process to investigate the role of fluorine (F) and germanium (Ge) doping elements on the radiation sensitivity of silica-based glasses. We characterized the behaviors of these canonical samples before, during and after 10 keV X-ray irradiation through several spectroscopic techniques, to obtain global information (in situ absorption measurements, electron paramagnetic resonance) or spatially-resolved information (confocal microscopy, absorption and luminescence on preform). These tests showed that, for the Ge-doped fiber and in the 300-900 nm range, the radiation-induced attenuation (RIA) can be explained by absorption bands associated with the following radiation-induced point defects: Ge(l); Ge-NBOHC and GeX. Other defects such as GeE' Ge(2); and Ge-ODC are generated but do not contribute in this spectral domain. For the F-doped sample, the different point defects identified, SiE', Si-NBOHC and Si-ODC(II), are unable to reproduce the RIA spectra for energies lower than 4 eV. We suggest that the radiation-induced absorption in this part of the spectrum is due to chlorine-related species, probably CI⁰ radiolytic groups that absorb at around 3.5 eV. The comparison between the sensitivities of the preform and the fiber reveals the influence of the drawing process on the glass response. Its effect is strongly dose-dependent for the germanosilicate glass. The drawing process seems to be responsible for the main part of the defects generated at low doses (<1 Mrad).

Full text

Page 1

IEEE TRANSACTIONS ON NUCLEAR SCIENCE, VOL. 55, NO. 6, DECEMBER 2008 3473
Radiation Effects on Silica-Based Preforms and
Optical Fibers—I: Experimental Study With
Canonical Samples
Sylvain Girard, Member, IEEE, Youcef Ouerdane, Giusy Origlio, Claude Marcandella, Aziz Boukenter,
Nicolas Richard, Member, IEEE, Jacques Baggio, Member, IEEE, Philippe Paillet, Senior Member, IEEE,
Marco Cannas, Jean Bisutti, Jean-Pierre Meunier, Member, IEEE, and Roberto Boscaino
Abstract—Prototype samples of preforms and associated fibers
have been designed and fabricated through MCVD process to
investigate the role of fluorine (F) and germanium (Ge) doping
elements on the radiation sensitivity of silica-based glasses. We
characterized the behaviors of these canonical samples before,
during and after 10 keV X-ray irradiation through several
spectroscopic techniques, to obtain global information (in situ
absorption measurements, electron paramagnetic resonance) or
spatially-resolved information (confocal microscopy, absorption
and luminescence on preform). These tests showed that, for the
Ge-doped fiber and in the 300–900 nm range, the radiation-in-
duced attenuation (RIA) can be explained by absorption bands
associated with the following radiation-induced point defects:
Ge(1); Ge-NBOHC and GeX. Other defects such as GeE’ Ge(2);
and Ge-ODC are generated but do not contribute in this spectral
domain. For the F-doped sample, the different point defects iden-
tified, SiE’, Si-NBOHC and Si-ODC(II), are unable to reproduce
the RIA spectra for energies lower than 4 eV. We suggest that the
radiation-induced absorption in this part of the spectrum is due
to chlorine-related species, probably ???radiolytic groups that
absorb at around 3.5 eV. The comparison between the sensitivities
of the preform and the fiber reveals the influence of the drawing
process on the glass response. Its effect is strongly dose-dependent
for the germanosilicate glass. The drawing process seems to be
responsible for the main part of the defects generated at low doses
(
? ????).
Index Terms—Absorption, confocal microscopy, EPR, lumines-
cence, optical fibers, point defects, X-rays.
I. INTRODUCTION
A
of silica-based optical glasses to evaluate their vulnerability to
LARGE number of experimental studies has been de-
voted to the characterization of the radiation responses
Manuscript received July 11, 2008; revised September 15, 2008. Current ver-
sion published December 31, 2008.
S. Girard, C. Marcandella, N. Richard, J. Baggio, and P. Paillet are with CEA
DIF, F91297 Arpajon Cedex, France (e-mail: sylvain.girard@cea.fr).
Y. Ouerdane, A. Boukenter, and J.-P. Meunier are with Laboratoire Hubert
Curien, UMR-CNRS 5516, F42000 Saint-Etienne, France (e-mail: ouer-
dane@univ-st-etienne.fr).
G. Origlio iswith the LaboratoireHubert Curien, UMR-CNRS 5516, F42000
Saint-Etienne, France and also with the Dipartimento di Scienze Fisiche ed As-
tronomiche dell’Università di Palermo, I-90123 Palermo, Italy.
M. Cannas and R. Boscaino are with Dipartimento di Scienze Fisiche ed
Astronomiche dell’Università di Palermo, I-90123 Palermo, Italy (e-mail:
cannas@fisica.unipa.it).
J. Bisutti is with CEA DIF, F91297 Arpajon Cedex, France and also with the
Laboratoire Hubert Curien, UMR-CNRS 5516, F42000 Saint-Etienne, France.
Color versions of one or more of the figures in this paper are available online
at http://ieeexplore.ieee.org.
Digital Object Identifier 10.1109/TNS.2008.2007297
various harsh environments [1]. Researchers showed that radi-
ations generate, at the microscopic scale, point defects in the
amorphous silica ( -
) glass network through ionization or
“knock-on” processes [2]. These point defects, or color centers,
induce the appearance of new energy levels located inside the
band gap of the dielectric. As a consequence, the defect-con-
taining glass absorbs a more important part of the transmitted
signal leading, at the macroscopic scale, to an increase of the
attenuation of the fiber waveguide. This absorption increase is
called Radiation-Induced Attenuation (RIA). Some of these de-
fects can also be excited by the signal transmitted along the
fiber and then emit parasitic light that superposes on it: Radi-
ation-Induced Emission (RIE). The amplitude and time kinetics
of these macroscopic changes depend on the nature, concentra-
tions and stability of point defects [2]. As a consequence, the
radiation responses of the fibers will depend on the composi-
tion of the silica glass used for their elaboration (choice of core
and cladding dopants, impurity levels, stoichiometry) and also
on the glass physical characteristics (fictive temperature, strain,
…) that strongly depend on the fabrication parameters (preform
depositionanddrawing process)[1]–[5]. These intrinsicoptical
fiberparametersaregenerallynotaccessiblefortheengineersin
charge of their radiation-tolerance study as they are mostly con-
sidered as confidential by fiber manufacturers. This difficulty
explains that up to now, the quantitative influence of each of
these parameters on the silica-glass radiation response is still an
open problem whereas we have a good knowledge of their qual-
itative influences.
In this paper, we present the preliminary experimental results
of an approach that we recently developed to study radiation-in-
duced effects with the help of specifically-designed fiber sam-
ples (refered to hereafter as canonical samples). The prototype
samples presented in this work have been elaborated for the
studyoftheinfluenceofGe-andF-dopingontheglassradiation
sensitivity. We present this new approach and some selected ex-
amples of the experimental results that are accessible through
the collaboration between our different research groups.
II. EXPERIMENTAL DETAILS
An important part of the interest in our approach is based on
the choice for the designs of canonical samples. These sam-
ples have to be representative of commercial fibers that have
already been tested and will be used in future facilities [6], [7].
Theyhavealsotoofferaneasierinterpretationoftheirresponses
thanks to their custom designs. Previously, two different studies
0018-9499/$25.00 © 2008 IEEE
Authorized licensed use limited to: CEA DAM. Downloaded on January 21, 2009 at 03:14 from IEEE Xplore. Restrictions apply.

Page 2

3474 IEEE TRANSACTIONS ON NUCLEAR SCIENCE, VOL. 55, NO. 6, DECEMBER 2008
used a set of homemade samples to understand the influence of
several process and composition parameters on the radiation re-
sponseofsingle-mode(SM)germanosilicateopticalfibersat1.3
and1.55
[3],[8].Duetothegoodknowledgeoftheirsample
characteristics and experimental design, E.J. Friebele et al. [8]
were able to obtain statistically significant correlations between
some of the -ray steady-state RIA fitting parameters and some
ofthefabricationparameters.Asanexample,theyfoundthatfor
dosesof2
at,theRIAlevelattheendoftheir-
radiationis correlatedwiththeGe contentin thefibercore (fora
Ge/Fdopedcladding).Thesecondstudywasdevotedtothetran-
sient X-ray radiation response of Ge-doped fibers and showed,
forexample,thatloweringthestandardpreformdepositiontem-
peraturefrom2000to1600
andthedrawingtensionfrom140
to 20 g slightly decreases the induced losses at both Telecom
wavelengths[4]andtheinfluenceofvariouscladdingcodopants
[3]. However, these two studies were limited by the difficulty to
obtain the samples with the characteristics needed to get unam-
biguous correlations.
For the present work, we design the structures to overcome
thesedifficulties.First,aquantitativeanalysisoftheinfluenceof
adopantcanonlybeachievedifallthedifferently-dopedglasses
have been made with strictly identical processes. From a prac-
ticalpointofview,duetothenon-negligibleinfluenceofMCVD
process parameters [4], this can not be achieved with the design
of several preforms and fibers. It must be done within a single
sample. Secondly, new spectroscopic techniques are accessible
that allow to spatially resolve the radiation-induced changes in
the fiber with micrometer resolution. For example, we show the
efficiency of the confocal microscopy of luminescence (CML)
tocharacterizetheradiation-inducedpointdefectsinpassive[9]
or rare-earth doped [10] optical fibers. These spatially-resolved
techniques enable the characterization of new sample designs
that would have not been possible in the past.
A. Tested Optical Fibers
We characterized the two first canonical samples developed
for our research. Both preform and associated fiber samples
were made through the Modified Chemical Vapor Deposition
(MCVD) processby iXFiber SAS, Lannion,France [11].About
50 mm of each prototype preform has been kept for analysis
whereas the other part of the preform has been drawn to ob-
tain several hundreds meters for each fiber. Opto-geometric and
optical characteristics of both preforms and fibers have been
checked to follow our specifications. Standard conditions of
fiber manufacturing (preform deposition and drawing process)
have been used for these waveguides.
1) Multi-StepStructures: Toquantitativelyinvestigatethein-
fluence of the dopant concentrations on radiation effects, the
structure of each fiber has been designed with several steps of
concentrationofonedopingelement(GeorF)inthecore.Using
spatially-resolved techniques [9], [10], [12] , we will then be
able to correlate, within the fiber or preform, the radiation re-
sponse to the dopant concentration for samples with strictly
identical MCVD process parameters.
2) Choice of Concentration Levels: Some of the dopant con-
centration values have been chosen to reproduce the classical
range of concentrations measured on commercial fibers (e.g.,
Fig. 1. Refractive-index profiles of (a) F-doped canonical sample and (b)
Ge-doped canonical sample. The refractive-index measurements have been
performed on fiber preforms at 633 nm.
TABLE I
CONCENTRATIONS OF DOPANTS IN THE DIFFERENT ZONES OF THE CANONICAL
SAMPLES (MEASURED BY EMPA)
from 2 to 12 wt.% for the Ge-doped fibers). The other ones
havebeendefinedinrelationwithourabinitiocalculationscon-
ducted on a 108 atoms silica-based supercell. This comparison
betweenourtheoreticalandexperimentalresultswillbedetailed
in another paper [13].
At the fabrication stage, the obtained multi-step radial dis-
tribution of the dopant along the fiber diameter can be roughly
estimated through measurements of the fiber and preform re-
fractive-index profiles (RIP). A more accurate estimation of the
concentration values of the dopants is obtained by electron mi-
croprobe analysis (EMPA) with a spatial resolution of one mi-
crometer.
In the two F- or Ge-doped samples, five different zones are
definedwithvariableconcentrationsofthestudiedchemicalele-
ment. The concentrations measured by EMPA for each zone are
given in Table I. Fig. 1 illustrates the measured refractive-index
(RI) profiles of F-doped and Ge-doped preforms.
3) F-Doped Canonical Sample: its core consists in three
zones (zones 1 to 3) with three different F-concentrations.
The F-incorporation inside -
in Fig. 1. Optical cladding (zone 4) corresponds to the zone
with the highest F-doping region (
enables the injected light signal to be guided into the core.
The outer cladding (zone 5) is made of pure-silica with low
concentration of hydroxyl groups (OH). Only a negligible part
decreases its RI as shown
). This structure
Authorized licensed use limited to: CEA DAM. Downloaded on January 21, 2009 at 03:14 from IEEE Xplore. Restrictions apply.

Page 3

GIRARD et al.: RADIATION EFFECTS ON SILICA-BASED PREFORMS AND OPTICAL FIBERS 3475
Fig. 2. Scheme of the experimental setup used for the RIA measurements under steady state 10 keV X-rays.
of the signal is guided in this part of the waveguide. As a conse-
quence, its contribution to the global RIA can be considered as
negligible. Doped parts of this fiber contain small amounts of
chlorine impurity (
typical of MCVD glasses.
4) Ge-Doped Canonical Sample: the fiber core diameter of
62.5
corresponds to zones 1 – 4 with Ge-doping varying
from
to. Ge-doping increases the RI of
- . Fiber cladding (zone 5) is made with the same pure
- as the outer cladding (zone 5) of the F-doped sample.
Doped parts of this fiber contain typical levels of chlorine im-
purity (
) and OH-groups (
These fibers exhibit pre-irradiation optical characteristics at
1.55
close to that of commercial fibers:
for the F-doped sample and
sample. The excess losses compared to commercial fibers can
mainly be explained by the non-optimized guiding properties
of these samples due to our choices for their refractive-index
profiles.
) and OH-groups ( ),
).
for the Ge-doped
B. Experimental Setups
A large variety of experiments has been performed on these
canonical samples. The canonical fibers and preforms have
radiation responses representative of commercial or prototype
MCVD optical fibers with the same composition. In this paper,
we present only the response of the fiber samples during and
after steady state 10 keV X-ray irradiation and some selected
results that can be obtained with the set of spectroscopic tools
provided by the collaboration between CEA, Saint-Etienne and
Palermo universities.
1) InSituRIAMeasurement: Theexcessoflossesinducedby
10 keV X-rays in our samples is evaluated using an ARACOR
machine [14]. This machine reproduces well the effects of
-photons or high-energy protons and offers flexible irradiation
conditions with a highly focused beam (
Our optimized setup, described in Fig. 2, enables the study
of radiation-induced effects on optical fibers in the ultraviolet
(UV) and visible part of the spectrum at temperatures varying
from
to 150 . All the measurements presented in this
paper were obtained at room temperature. During irradiation,
the deuterium-halogen source (DH2000 from Ocean Optics)
and spectrometer (Ocean Optics HR4000) are located inside the
ARACOR chamber and shielded against the scattered X-rays.
diameter).
The penetration range of the incoming X-rays is sufficient to
deposit the dose in the whole fiber cross-section. Two pigtails
of UV-transmitting high-OH pure-silica fibers are used to inject
the white light inside the tested fiber sample (
length) and to transmit the propagated signal to the spectrom-
eter. Dose rates of 9
for the Ge-doped sample were used to adjust the
fiber sensitivities to the setup dynamics.
2) SpectroscopicCharacterization of Radiation-InducedDe-
fects: InadditiontotheinsituRIAmeasurements;variousspec-
troscopic techniques have been used to identify the precursor
sitesinnon-irradiatedsamplesorroom-temperaturestablepoint
defects in irradiated samples:
• Electron paramagnetic resonance (EPR) measurements
are performed at Palermo University giving the structure
and concentration of paramagnetic defects in the pristine
and irradiated glasses at room-temperature. A more com-
plete description of this setup and preliminary results ob-
tained on the F-doped glass have been recently published
by G. Origlio et al. [15]. This technique gives global infor-
mation on the fiber behavior; differences in the response of
the five zones can not be distinguished.
• Confocal microscopy luminescence (CML) and Raman
(CMR) measurementsare performed at Saint-EtienneUni-
versity. The description and the advantages of this tech-
nique are detailed in [10]. Different lasers emitting in the
UV or in the visible domains are used to selectively excite
the different radiation-induced point defects. In this paper,
wepresentsomeCMLresultsobtainedwithgreenlaserex-
citation at 488 nm (2.55 eV). The power level of the probe
light is reduced to few hundreds of
bleaching effects.
• Spatially-resolved Absorption and Luminescence (SRA)
measurements have been done at Palermo University in
the different-doped zones of the preforms. These tests can
be performed before and after irradiation at several steps
of dose [12]. For SRA measurements, the studied range
of wavelengths depends on the preform composition with
a starting value around 150 nm (8.3 eV) for the F-doped
sample, and around
one). Preliminary results obtained with this technique on
the Ge-doped sample have been recently presented by G.
Origlio et al. [12].
in
for the F-doped fiber and
to avoid photo-
(6 eV) for the Ge-doped
Authorized licensed use limited to: CEA DAM. Downloaded on January 21, 2009 at 03:14 from IEEE Xplore. Restrictions apply.

Page 4

3476 IEEE TRANSACTIONS ON NUCLEAR SCIENCE, VOL. 55, NO. 6, DECEMBER 2008
Fig. 3. Spectral dependence of radiation-induced attenuation (RIA) during 10 keV X-ray irradiation of F-doped sample (dose rate of 9??? ?????). The inset
illustrates the RIA growth and decay at specific wavelengths for a dose rate of 9??? ????? as a function of time.
III. RESULTS
A. In Situ RIA Measurements During 10 keV X-Ray Irradiation
Fig.3givesthe10keV-X-rayRIAspectrameasuredatroom-
temperature during the steady-state irradiation at different dose
levels (for a dose rate of 9
Irradiation strongly increases the absorption levels in this
fiber, even at low irradiation doses. The RIA spectra reveal
the generation of several radiation-induced point defects with
strongly different growth kinetics in the studied range of
doses. For short irradiation times, a broad absorption around
(;
amplitude remains stable at higher doses. This is illustrated
in the inset of Fig. 3 with the saturation effect of the RIA
growth at 350 nm after 40 s (3.6
especially affects the fiber at low doses; increasing the dose
from 2.7
to 6.5
350 nm; 3.54 eV) by a factor of 3. At least another absorption
band peaking at lower wavelengths,
and continuously growths in this range of doses. At higher
wavelengths (
), other absorption bands seem to be
present but the dynamic of these X-ray measurements must be
improved to provide better characterization in this domain.
Fig. 4 gives the RIA spectra measured during the 10 keV
X-ray irradiation (
instead of 9
F-doped sample) at different dose levels for the Ge-doped
sample.
Our measurements provide evidence for a strong absorption
in the UV and visible range (
at
are generated by X-ray irradiation. The spectral
positions of the absorption bands responsible for these induced
losses cannot be detected from these measurements.
In Fig. 5, the comparison between the sensitivities of both
fibers at the same dose level (3.6
Ge-doped fiber exhibits higher RIA than the F-doped one, espe-
cially in the UV range. This is consistent with previous studies
) for the F-doped sample.
) appears and its
). This absorption
leads to a RIA increase (at
, appears
for the
). Defects absorbing
) shows that the
that showed the good radiation response of fluorine-doped
glasses compared to that of germanosilicate glasses [1]. How-
ever, the results between the two samples are not directly
comparable as different dose rates have been used for our tests.
Itwas alsopreviouslyshownthatdose rateinfluencesthesteady
state radiation responses of both fiber types [1], [2], [5].
B. EPR/CML Characterization of Canonical Samples
In this part, we give some examples of the experimental re-
sults obtained through spectroscopic analyses on pristine and
X-ray irradiated samples of the Ge and F-doped glasses. All
measurements were done at room temperature and several days
or months after irradiation. All the defects identified by these
techniques are room-temperature stable.
1) F-Doped Fiber and Perform: these samples contain con-
centrations of pre-existing paramagnetic defects lower than the
EPR system detection limit [15]. Measurements on -ray irra-
diated preforms and fibers showed that a dose of 9.1
leads to the generation of SiE’ centers (
denotes an unpaired electron and
bonds) at concentrations of
for the preform and fiber respectively
[15]. The EPR signature of the non-bridging oxygen hole
centers (NBOHC;
measurements although spatially-resolved CML measurements
on irradiated samples of F-doped glasses under 488 nm laser
excitation provide evidence for their red luminescence in the
core and cladding. A comparison between the responses of the
layers doped with different F-levels shows that the amplitude
of the red emission of Si-NBOHCs decreases when increasing
F-levels in the glass [15].
2) Ge-Doped Fiber and Perform: EPR tests have been
done on the pristine and X-ray irradiated Ge-doped preform
and fibers. We recently showed that
5 eV) laser exposure leads to the generation of the same
,
three Si-O regular
and
) was not revealed by our
-ray and UV (248 nm;
Authorized licensed use limited to: CEA DAM. Downloaded on January 21, 2009 at 03:14 from IEEE Xplore. Restrictions apply.

Page 5

GIRARD et al.: RADIATION EFFECTS ON SILICA-BASED PREFORMS AND OPTICAL FIBERS 3477
Fig. 4. Growth increase of 10 keV X-ray radiation-induced attenuation (RIA) for Ge-doped sample (dose rate of ?? ?????). The inset illustrates the RIA growth
at specific wavelengths for short irradiation times.
Fig. 5. Spectral dependence of radiation-induced attenuation (RIA) mea-
sured for the F-doped sample (9??? ?????) and for the Ge-doped sample
(?? ?????) at a dose of 3.6??? ???.
paramagnetic defects [12]. In the pristine samples, our mea-
surements provided evidence for several paramagnetic species
like GeE’ (
[16], [17]). The EPR spectra of
the irradiated sample pointed out the generation of additional
GeE’ centers and new defect structures: Ge(1) and Ge(2)
centers. Comparable paramagnetic species were obtained by
E.J. Friebele in a commercial Ge-doped fiber [16]. Ge(1) has
been described as an electron trapped in a fourfold coordinated
Ge atom [16] whereas the structure of Ge(2) is still debated.
Possible structures are a hole center [18] or a trapped elec-
tron center [19]. In the whole range of dose studied (up to
2
), we note a continuous increase of GeE’ defects up
to concentrations of
the concentrations of the Ge(1) and Ge(2) defects seem to
saturate at
and
for doses higher than
.
Fig. 6 presents some CML measurements with a laser probe
excitation at 488 nm (2.54 eV) on the Ge-doped fiber sample.
at 2 whereas
respectively
Fig. 6. Radial distribution of red emitting point defects in the pristine and ir-
radiated Ge-doped fibers. The inset illustrates the luminescence shape of the
luminescence acquired along the fiber diameter.
The luminescence spectra acquired along the fiber diameter
have similar shapes illustrated in the inset of Fig. 6. A strong
luminescence band centered at around 680 nm is measured in
bothsamplesanditsintensityincreaseswithradiations.Thisred
emission may be associated with the Ge form of non-bridging
oxygen hole centers (Ge-NBOHC;
spatial distribution of the defects responsible for this emission
under 488 nm excitation is illustrated in Fig. 6 for the pristine
and X-irradiated Ge-doped fibers through the radial distribution
of the integrated photoluminescence (PL) intensity. These
defects are mainly generated in the central part of the core
where the germanium concentration is the largest.
). The
IV. DISCUSSION
We focus our discussion on two main points. First, we at-
tempted to decompose the in situ RIA spectra with the help of
Authorized licensed use limited to: CEA DAM. Downloaded on January 21, 2009 at 03:14 from IEEE Xplore. Restrictions apply.

Page 6

3478 IEEE TRANSACTIONS ON NUCLEAR SCIENCE, VOL. 55, NO. 6, DECEMBER 2008
TABLE II
CHARACTERISTICS OF RADIATION-INDUCED POINT DEFECTS IDENTIFIED IN
OUR PRISTINE AND IRRADIATED SAMPLES.
a set of Gaussian bands corresponding to the absorption bands
of defects identified by our different spectroscopic techniques.
Secondly, we compare the radiation-induced effects on the
fibers and their associated preforms to investigate the influence
of the drawing process on the glass sensitivity.
A. Nature of Radiation-Induced Attenuation (RIA) in Ge and
F-Doped Glasses
Through the different spectroscopic analyses, we have been
able to identify some of the different radiation-induced point
defects that are generated by 10 keV X-rays in the canonical
samples. We review these Ge and Si-related defects in Table II
and give their structures, main characteristics of their optical
absorption (OA) bands (spectral positions
and thetechniquethatallowedtheirdetection. We distinguishin
the table the point defects identified through our different spec-
troscopic measurements (ESR, CML, …) and those identified
through the decomposition of the RIA spectra.
Most of these techniques (EPR, CML, and SRA) can not be
performed in situ during the irradiation and the defects unstable
at room-temperature can only be studied through in situ OA
measurements. Due to the overlapping between the different ra-
diation-induced absorption bands in the UV and visible range,
some diamagnetic additional point defects may be generated by
X-rays. Furthermore, paramagnetic defects at low concentra-
tions may not be detected through EPR measurements due to
the detection limits of this system [15]. To understand the ori-
gins of the induced losses in this part of the spectrum, we de-
compose the obtained RIA curves with the help of
, and FWHM)
Gaussian
bands
with time dependent amplitude
and
The commonly used approach consists in adjusting the number
of these Gaussian bands and their characteristics
toachievethebestfitsoftheexperimentalcurvesandthento
discuss the color centers that can be associated to these bands.
However, as in our previous study [20], we choose to only use
the defects given in Table II to decompose the measured RIA
spectra. Additional point defects from literature will be used
onlyifthissetofdefectsisunabletoreproducetheexperimental
data.
1) Decomposition of RIA Spectra in Ge-Doped Samples:
In our spectral range of measurement (1.5 – 4 eV), the RIA
curve cannot be accurately reproduced using the set of defects
of Table II. Our analysis shows that the absorption bands of the
GeODC, GeE’ and Ge(2), present in the irradiated glass, do not
contribute to the RIA in this part of the spectrum. To correctly
fit the experimental curves, the addition of another absorption
band is necessary:
associated with the GeX defect [21], [22] which is diamagnetic
and has not yet been associated with any luminescence band.
Typically, this defect can not be studied through CML or EPR
techniques. Our decomposition agrees with the work of G.
Origlio et al. [12] who found a linear correlation between the
Ge(1) concentration and the RIA at 4 eV in the X-ray irradiated
preform, showing that at this energy, the contribution of the
defects with absorption bands peaking in the UV is negligible.
In Fig. 7, we illustrate on the same graph the measured RIA
at a dose of 3.6
, with the best fit (red curve) ob-
tainedwiththecombinationofthethreedefects(Ge(1),GeXand
Ge-NBOHC). We also represent the other defects that can the-
oretically contribute to losses in this spectral domain for infor-
mation, but they do not contribute to the fit. The same three de-
fects have been previously shown to be able to reproduce -ray,
14 MeV neutron induced attenuation in commercial Ge-doped
fibers [20]–[22]. This good agreement between our work and
previousstudiesoncommercialGe-doped glassesalso confirms
the ability of our canonical samples to represent commercial
glasses.
2) Decomposition of RIA Spectra in F-Doped Samples:
In Fig. 8, we try to decompose the RIA measured at doses
of 2.7
and 6.5
described in Table II. From our analysis, it is obvious that ad-
ditional defects, absent from Table II, have to be considered to
explain the measured RIA curves, especially at lower energies
(larger wavelengths). For energies from 4 eV to 6 eV (up to
300 nm), induced losses can be explained by a combination of
the three absorption bands at 4.8 eV; 5.03 eV and 5.8 eV that
are associated with Si-NBOHC, Si-ODC(II) and SiE’ centers
respectively. As our sample has been made through MCVD
process, it also contains chlorine impurity at concentrations
in the order of 1200 ppm. In pristine silica-based samples,
molecular
has been associated with an absorption band
around 3.78 eV (
, and
; . This band is
with the set of defects
) [23] but there is no
Authorized licensed use limited to: CEA DAM. Downloaded on January 21, 2009 at 03:14 from IEEE Xplore. Restrictions apply.

Page 7

GIRARD et al.: RADIATION EFFECTS ON SILICA-BASED PREFORMS AND OPTICAL FIBERS3479
Fig. 7. Decomposition of the radiation-induced attenuation (RIA) in the
Ge-doped fiber at a dose of 3.6???
lengths, the experimental RIA can be fitted (red curve) by three defects: GeX,
Ge-NBOHC and Ge(1).
???. In the studied range of wave-
firm evidence that
under irradiation. However, it has been shown that the chlorine
presence in the silica-based glass leads under irradiation to the
creation of
andmolecular ions [24], [25]. From the
decay kinetics of RIA after the irradiation (see inset of Fig. 3),
we suggest that induced bands due to Cl impurities are limited
tothe
bandas thisspeciesrapidlydecays(
the radiation is turned off whereas
room temperature [26].
products have been associated with
absorption band around 3.65 eV (
and to an absorption band around 3.26 eV (1.2 eV) [27], [28].
Additional measurements performed at
on the F-doped sample reveal the temperature dependence of
the chlorine-related absorption: its intensity decreases when
increasing temperature. Furthermore, these temperature mea-
surements confirm that several absorption bands are present
in this range of wavelengths. At this time, we do not have
performed specific EPR measurements in order to correlate
the absorption bands to the EPR signature of the different
Cl-related centers. This has to be done in the future. However,
the in situ RIA measurements showed a rapid bleaching of
these species at the end of the irradiation. As a consequence,
they may be not characterized by our EPR or CML tools.
Duetotheshortlengthsofthetestedsamplesandtheresulting
low RIA values, the quality of our measurements are limited at
shorter energies (
). However, we assumed that the
absorptionbandswillnotbesufficienttoreproducelossesin
thisrangeofwavelengths.Previousexperimentsonsuchglasses
[21], [29] suggest that losses in the visible-near infrared will be
due to two types of defects: the NBOHC and the Self-Trapped
Holes (STHs). The relative contribution of the two classes of
defects will change with respect to the dose rate, the dose of
the experiments [29], [30]. Further experiments with improved
resolution in the 0.5–3 eV have to be done to clearly identify
the defects at the origin of the induced losses in this spectral
domain.
is created in chloride-containing silicas
)when
species are stable at
) [24]
and
Fig. 8. Decomposition of the radiation-induced attenuation (RIA) in the
F-doped fiber at doses of (a) 2.7??? ??? and (b) 6.5??? ??? (dose rate of
9??? ?????). The Gaussian bands used for the fit are described in the text.
Fig. 9. Ge-concentration dependence of the total attenuation at 4 eV and 4.75
eV (??????????????? ? ?????????????????) measured at the different zones
of a X-ray irradiated Ge-doped preform.
3) Influence of the Dopant Concentration on RIA: The
multi-step structures of our canonical samples is used to
evaluate the effect of the dopant concentration versus the
defect concentration. An example of the possible accessible
results, we plotted in Fig. 9 the permanent total attenuation
(
lengths versus the Ge-concentration. These preliminary data
have been acquired through SRA measurement on a slice of
X-ray irradiated preform at Palermo. The spectral dependence
of the attenuation reveals that Ge(1) centers are mainly respon-
sibleforthe4eV(
)absorption[12]whereasGe-ODC
centers are mainly responsible for the 4.75 eV (
) at two single wave-
)
Authorized licensed use limited to: CEA DAM. Downloaded on January 21, 2009 at 03:14 from IEEE Xplore. Restrictions apply.

Page 8

3480 IEEE TRANSACTIONS ON NUCLEAR SCIENCE, VOL. 55, NO. 6, DECEMBER 2008
Fig. 10. Comparison of the ratio between concentrations of Ge-related defects
(Ge(1), Ge(2) and GeE’) and SiE’ defects in Ge-doped sample (square) and
F-doped sample (circle) at different doses.
absorption. The attenuation levels in the pure-silica cladding
are also plotted for comparison although different point defects
are at the origin of the losses.
B. Influence of the Drawing Process on the Fiber Radiation
Sensitivity
Up to now, only few studies investigated the influence of
the drawing process on the fiber radiation response by com-
paring the behaviors of fiber and preform. Most of them have
characterized the effect of varying the drawing tension, speed
or temperature on the radiation response of optical fibers [4],
[8], [31]. H. Hanafusa et al. [31] showed that in the 700 to
1600 nm range of wavelengths, RIA decreases with increasing
drawing speed (from 120 m/min to 860 m/min) and with de-
creasing furnace temperature (from 2255
study, all our canonical samples have been made with standard
MCVD drawing conditions. By comparing the concentrations
of room-temperature stable paramagnetic defects at the same
doses in a preform and its corresponding fiber, we can estimate
the global influence of the drawing process on the glass sensi-
tivity. The obtained results are illustrated in Fig. 10 for the Ge
and F-doped samples.
Our results clearly show that the influence of the drawing
process on the generation of Ge-related centers is dose-depen-
dent. Similar dose-dependence is observed for the three para-
magnetic defects: Ge(1), Ge(2) and GeE’. The impact of the
drawingprocessisstronglynegativeat2
nearly negligible at higher doses for the two types of glass. De-
spite the limited number of experiments on F-doped glass, the
fiberbehaviorseemslessaffectedbythedrawingprocess.These
results show that at lower doses (space, military applications)
the drawing process governs the generation of paramagnetic de-
fects in Ge-doped glass whereas the glass composition seems to
bethemostinfluentialparameterforhigh-doseapplications(nu-
clear power plants, high energy physics).
A possible explanation for this dose dependence is that
drawing strongly increases the number of defect precursors
to 2000 ). In our
andbecomes
such as GeODC(II) and GeODC(I) which can turn into Ge(1),
Ge(2) or GeE’ under irradiation following the processes [32],
[33]:
(1)
(2)
At low doses, the contribution of defects generated from these
precursor sites to the total concentration of defects is predom-
inant whereas it may become less important at higher doses
due to defect generation via other mechanisms. Our CML
measurements on the Ge-doped fiber with UV excitation pro-
vided evidence for the presence of GeODC(II) in both pristine
and irradiated samples with an increase of the GeODC(II)
concentration after irradiation [20]. By the way, Ge(1) and
Ge(2) defects may be created from the preexisting and X-ray
radiation-induced GeODC(II). G. Origlio et al. [12] showed
that the Ge(1) and Ge(2) concentrations saturate at higher
doses (
) whereas GeE’ concentrations continuously
increase up to 2
providing evidence for several gener-
ation mechanisms for this defect. An interesting point is that the
work by H. Hanafusa et al. [31] was done using -ray irradiated
samples at a dose of
, which corresponds to the dose
range of strong influence of the drawing process on the glass
sensitivity. However, the authors found no correlation between
the measured RIA at 0.71
(1.75 eV) and the concentration
of paramagnetic Ge-related species measured by EPR. As a
consequence, they attributed the losses to Drawing-induced
Defects (possibly Ge-O-O-Si defects) created by irradiation.
Our results, illustrated in Fig. 7, agree with this work and
suggest that GeX centers, another diamagnetic defect, seem
mainly responsible for RIA at these wavelengths.
For the F-doped glass, it has been demonstrated that the ad-
dition of fluorine to the glass decreases the detrimental effect of
the drawing [34]. For this type of glass, the number of precursor
sites created during the drawing for the SiE’ generation is lim-
ited compared to pure-silica core fibers [35] and can explain the
ratio of
between the concentrations measured in both the
preform and the fiber. For example, it has been established that
F-doping reduces the number of strained bonds
that leads to thecreation of NBOHC-E’ pairs through the mech-
anism:
(3)
Our CMR measurements on the pristine sample proved that the
F-doping reduces the concentration of strained bonds that act as
precursor sites whereas our CML measurements on the irradi-
ated one confirmed the decrease of the red luminescence asso-
ciated with NBOHC with increasing fluorine content [15].
Additional tests have to be done to fully understand the
drawing influence on the fiber radiation response. Comple-
mentary ESR measurements have to be done at lower doses
to investigate this effect for low-dose environments like Laser
Megajoule [6]. Furthermore, different fiber samples have to be
drawn from the same Ge-doped preform to determine the most
favorable drawing conditions for this kind of optical fiber.
Authorized licensed use limited to: CEA DAM. Downloaded on January 21, 2009 at 03:14 from IEEE Xplore. Restrictions apply.

Page 9

GIRARD et al.: RADIATION EFFECTS ON SILICA-BASED PREFORMS AND OPTICAL FIBERS3481
V. CONCLUSION
We present a new approach to investigate radiation-induced
effects on the transmission properties of silica-based preforms
and their corresponding optical fibers. We have developed and
characterized two specially designed waveguides to study the
effects of fluorine and germanium-doping on the sensitivity
of the fibers. Preliminary analysis allows the identification of
some of the defects responsible for the fiber degradation: SiE’;
GeE’, Ge(1), Ge(2), Ge-ODC, GeX, Si-ODC, Si-NBOHC and
Ge-NBOHC. The point defects at the origin of the radiation-in-
duced attenuation in the Ge-doped sample have been identified
and three defects are responsible for the signal decrease in
the 300 to 900 nm range: Ge(1), GeX and GeNBOHC. The
response of the F-doped glass seems more complex. Losses in
the range of wavelengths from 200 to 320 nm can be explained
by a combination of SiODC, Si-NBOHC and SiE’ defects. At
lower energies (higher wavelengths) several chlorine-related
species seem to explain the increase of loss during irradiation
of the F-doped fiber. Additional measurements combining
temperature treatments and
-loading will be done as well as
new tests on Fluorine-doped samples containing low levels of
the chlorine impurity.
The comparison between the fiber and preform radiation re-
sponse shows that the drawing process increases the fiber sensi-
tivity to radiation. Its influence seems strongly dose dependent
andwillaffectmorethebehaviorofthefiberatlowdoses.Addi-
tional tests on several fibers drawn from the same preform with
different drawing parameters (temperature, tension, speed) will
be done to improve our knowledge on the mechanisms respon-
sible for this dose dependence.
ACKNOWLEDGMENT
The authors would like to acknowledge T. Robin, B. Cadier,
and P. Crochet from iXFiber SAS for their efficiency to develop
the canonical samples, V. Ferlet-Cavrois and O. Flament from
CEA for fruitful discussions about this study. The authors also
acknowledge D. L. Griscom from ImpactGlass Research Inter-
national for his aid in interpreting and fitting the radiation-in-
duced attenuation spectra.
REFERENCES
[1] B. Brichard and A. F. Fernandez, “Radiation effects in silica
glass optical fibres,” in Proc. RADECS Short Course, 2005,
pp.95–137.
[2] D. L. Griscom, “Nature of defects and defect generation in optical
glasses,” in SPIE Proc., 1985, vol. 541, pp. 38–59.
[3] S. Girard, “Analyse de la Réponse des Fibres Optiques Soumises à
Divers Environnements Radiatifs,” Ph.D. dissertation, Université de
Saint-Etienne, , France, 2003.
[4] S. Girard, Y. Ouerdane, A. Boukenter, and J.-P. Meunier, “Transient
radiationresponsesofsilica-basedopticalfibers:InfluenceofModified
Chemical Vapor Deposition process parameters,” J. Appl. Phys., vol.
99, p. 023104, 2006.
[5] E. J. Friebele, “Photonics in the space environment,” in Nucl. and
Space Radiation Effects Conf., San Diego, CA, 1991, Short Course,
pp. III-1–III-25.
[6] S. Girard, J. Baggio, J.-L. Leray, J.-P. Meunier, A. Boukenter, and Y.
Ouerdane, “Vulnerability analysis of optical fibers for Laser Mega-
joule facility: Preliminary studies,” IEEE Trans. Nucl. Sci., vol. 52, pp.
1497–1503, 2005.
[7] M. Van Uffelen, “Modélisation de Systèmes d’Acquisition et
de Transmission à Fibres Optiques Destinés à Fonctionner en
Environnement Nucléaire,” Ph.D. dissertation, Univ. Paris XI,
Paris, France, 2001.
[8] E. J. Friebele, C. G. Askins, C. M. Shaw, M. E. Gingerich, C. C. Har-
rington, D. L. Griscom, T.-E. Tsai, U.-C. Paek, and W. H. Schmidt,
“Correlation of single-mode fiber radiation response and fabrication
parameters,” Appl. Opt., vol. 30, no. 15, pp. 1944–1957, 1991.
[9] S. Girard, J.-P. Meunier, Y. Ouerdane, A. Boukenter, B. Vincent, and
A. Boudrioua, “Spatial distribution of the red luminescence in pristine,
gamma-rays and ultraviolet-irradiated multimode optical fibers,” Appl.
Phys. Lett., vol. 84, pp. 4215–4217, 2004.
[10] B.Tortech,S.Girard,E.Régnier,Y.Ouerdane,A.Boukenter,J.-P.Me-
unier, M. Van Uffelen, A. Gusarov, F. Berghmans, and H. Thienpont,
“Core versus cladding effects of proton irradiation on erbium-doped
optical fiber: micro-luminescence study,” IEEE Trans. Nucl. Sci., vol.
55, pp. 2223–2228, 2008.
[11] iXFiber Website, [Online]. Available: http://www.ixfiber.com
[12] G. Origlio, S. Girard, M. Cannas, Y. Ouerdane, R. Boscaino, and
A. Boukenter, “Paramagnetic Germanium-related centers induced by
energetic radiation in silica devices for optical transmissions,” in ???
and Advanced Dielectrics, Saint-Etienne, Jun. 2008, J. Non-Cryst.
Solids.
[13] S. Girard, N. Richard, Y. Ouerdane, G. Origlio, A. Boukenter, L.
Martin-Samos, P. Paillet, J.-P. Meunier, M. Cannas, and R. Boscaino,
“Radiation effects on silica-based preforms and optical fibers – II: ab
initio simulations,” in NSREC 2008, Tucson, AZ, 2008, vol. IEEE
Trans. Nucl. Sci.
[14] P. Paillet, J. R. Schwank, M. R. Shaneyfelt, V. Ferlet-Cavrois, R. L.
Jones, O. Flament, and E. W. Blackmore, “Comparison of charge yield
inMOSdevicesfordifferentradiationsources,”IEEETrans.Nucl.Sci.,
vol. 49, pp. 2656–2661, 2002.
[15] G. Origlio, A. Boukenter, S. Girard, N. Richard, M. Cannas, R.
Boscaino, and Y. Ouerdane, “Irradiation induced defects in fluorine
doped silica,” Nucl. Instr. Methods in Phys. Res. B, vol. 266, pp.
2918–2922, 2006.
[16] E. J. Friebele, D. L. Griscom, and G. H. Sigel Jr., “Defect centers in a
germanium-doped silica-core optical fiber,” J. Appl. Phys., vol. 45, p.
3424, 1974.
[17] H. Hosono, Y. Abe, D. L. Kinser, R. A. Weeks, K. Muta, and H.
Kawazoe, “Nature and origin of the 5-eV band in ??? :???
glasses,” Phys. Rev. B, vol. 46, pp. 11445–11451, 1992.
[18] M. Fujimaki, T. Watanabe, T. Katoh, T. Kasahara, N. Miyazaki,
Y. Ohki, and H. Nishikawa, “Structures and generation mecha-
nisms of paramagnetic centers and absorption bands responsible for
Ge-doped ??? optical-fiber gratings,” Phys. Rev. B., vol. 57, no. 7,
pp. 3920–3926, 1998.
[19] J. Nishii, K. Kintaka, H. Hosono, H. Kawazoe, M. Kato, and K. Muta,
“Pair generation of Ge electron centers and self-trapped hole centers in
??? -??? glasses by KrF excimer-laser irradiation,” Phys. Rev. B,
vol. 60, pp. 7166–7169, 1999.
[20] S. Girard, J. Baggio, and J. Bisutti, “14-MeV neutron, ?-ray, and
pulsed X-ray radiation-induced effects on multimode silica-based
optical fibers,” IEEE Trans. Nucl. Sci., vol. 53, pp. 3750–3757, 2006.
[21] D.L.Griscom, “Gamma-ray-induced
Ge-doped-silica fiber image guides,” J. Appl. Phys., vol. 78, no.
11, pp. 6697–6704, 1995.
[22] E. V. Anoikin, V. M. Mashinski, V. B. Neustruev, and Y. S. Sidorin,
“Effects of exposure to photon of various energies on transmission of
germanosilicate optical fiber in the visible to near IR spectral range,”
J. Non-Cryst. Solids, vol. 179, pp. 243–253, 1994.
[23] L. Skuja, “Optical properties of defects in silica,” in Defects in ???
and Related Dielectrics, G. Pacchioni, Ed.
2000, pp. 73–116.
[24] D. L. Griscom, E. J. Friebele, and S. P. Mukherjee, “Studies of radia-
tion-inducedpointdefectsinsilicaaerogelmonoliths,”Cryst.Latt.Def.
and Amorph. Mat., vol. 17, pp. 157–163, 1987.
[25] D. L. Griscom, “Optical absorption of Cl2- hole type centers in irra-
diated alkali halide-alkali borate glasses,” J. Chem. Phys., vol. 51, pp.
5186–5187, 1969.
[26] D. L. Griscom, Private Communication 2008.
[27] E. M. Dianov, L. S. kornienko, A. O. Rybaltovskii, and P. Chernov,
“Unstable radiation colour centres in pure silica fibres: The nature and
properties,” in Proc. SPIE, 1994, vol. 2425, pp. 148–154.
opticalattenuation in
Boston, MA: Kluwer,
Authorized licensed use limited to: CEA DAM. Downloaded on January 21, 2009 at 03:14 from IEEE Xplore. Restrictions apply.

Page 10

3482 IEEE TRANSACTIONS ON NUCLEAR SCIENCE, VOL. 55, NO. 6, DECEMBER 2008
[28] H. Mori, Y. Suzuki, and M. Hirai, “Selective photobleaching of radia-
tion-induced absorption in a-??? ,” Nucl. Instr. Methods in Phys. Res.
B, vol. 91, pp. 391–394, 1994.
[29] D. L. Griscom, “Fractal kinetics of radiation-induced point defect for-
mation and decay in amorphous insulators: Apllication to color cen-
ters in silica-based optical fibers,” Phys. Rev. B., vol. 64, pp. 174201-
1–174201-14, 2001.
[30] S. Girard, D. L. Griscom, J. Baggio, B. Brichard, and F. Bergh-
mans, “Transient optical absorption in pulsed-X-ray-irradiated
pure-silica-core optical fibers: Influence of self-trapped holes,” J.
Non-Cryst. Solids, vol. 352, pp. 2637–2642, 2006.
[31] H.Hanafusa,Y.Hibino,andF.Yamamoto,“Drawingconditiondepen-
dence of radiation-induced loss in optical fibres,” Elect. Lett., vol. 22,
no. 2, pp. 106–108, 1986.
[32] M. Fujimaki, T. Kasahara, S. Shimoto, N. Miyazaki, S. Tokuhiro, K. S.
Seol, and Y. Ohki, “Structural changes induced by KrF excimer laser
photons in ? -loaded Ge-doped ??? glass,” Phys. Rev. B., vol. 60,
no. 7, pp. 4682–4687, 1999.
[33] V. B. Neustruev, “Color centres in germanosilicate glass and optical
fibres,” J. Phys. Condens. Mater., vol. 6, pp. 6901–6936, 1994.
[34] A. A. Abramov, M. M. Bubnov, E. M. Dianoc, L. A. Kol’Chenko, S.
L. Semjonov, A. G. Shebunajev, A. N. Gurajnov, and V. F. Khopin,
“Influence of fluorine doping on drawing-induced losses,” Elect. Lett.,
vol. 29, no. 22, pp. 1977–1978, 1993.
[35] K. Khalilov, K.-F. Klein, and A. V. Amosov, “Influence of fiber
drawing on optical UV-attenuation of all-silica fibers with undoped
core,” in Proc. SPIE, 1998, vol. 3262, pp. 166–172.
Authorized licensed use limited to: CEA DAM. Downloaded on January 21, 2009 at 03:14 from IEEE Xplore. Restrictions apply.