Choice of the WLS fibres for the Tilecal calorimeter
M. David, A. Gomes, A. Maio, J. Pina, M. J. Varanda
LIP and FCUL, Lisbon, Portugal
We report on the criteria and measurements performed for the choice of the type of fibre that
would equip the Tilecal Calorimeter. There were three types of candidate WLS fibres each from a
different producer: BCF91A from Bicron, S250-100 from Pol.Hi.Tech and Y11(200)MSJ from
Kuraray. The three producers have a history of cooperation with the Tilecal/ATLAS group and some
of those fibres were used to instrument several prototype modules of the Tilecal. After many tests, the
fibres that were chosen to equip the Tilecal were the Y11(200)MSJ from Kuraray.
1 The requirements
In order to choose the WLS fibres for the Tilecal calorimeter, we had to
establish criteria to select one out of the several candidates proposed by Bicron,
Kuraray and Pol.Hi.Tech. These criteria take into account the optical and mechanical
properties of the WLS fibres as well as cost.
The physics requirements in terms of light yield, attenuation length, fibre to
fibre fluctuations, diameter, mechanical stress and radiation hardness are listed below.
i) Light yield and attenuation length are key factors that affect the calorimeter
response, and that should be maximised in order to get the best performance in the
beginning of the operation as well as later on, after degradation due to natural ageing
and radiation damage. Selected fibres were used as reference for the light yield. The
value of the light output at a distance of 140 cm from the photomultiplier tube (PMT),
I140, should be at least 80% of the value obtained with these reference fibres. The
value of attenuation length should be at least 250 cm.
ii) Fibre to fibre fluctuations of light yield and attenuation length affect
directly non-uniformity inside a cell , and there is no possibility of recovery using
calibration. This kind of non-uniformity contributes to the constant term of energy
resolution. The fibre to fibre fluctuations should be less than 7% both on light yield
and attenuation length, similar to the fluctuations expected for the tiles .
iii) Diameter and eccentricity need to be within very stringent restrictions,
imposed by the demands of the robotized insertion of the fibres into the profiles.
The diameter should be 1.00mm with a tolerance of 0.02mm. The eccentricity
tolerance is |Dmax-Dmin| < 0.03mm.
iv) Light loss due to mechanical stress has to be taken into account, since the
fibres follow curved paths in the routing to the PMT’s. The light loss is measured
bending the fibres with a full circle of 7cm diameter, the maximum bending imposed
to the fibres in the calorimeter, and should be less than 5%.
v) Radiation hardness plays an important role in the expected light loss during
operation due to the high radiation environment of the LHC. So radiation hard fibres
are required. After the irradiation with a total dose of about 150 krad, the light loss
measured at a distance equal to 140 cm from the PMT should be less than 15% after
one week of recovery.
2 Results of the QC of the fibres candidate to Tilecal
Each producer was asked to send a sample of 500 WLS fibres. These 2 m long
fibres should satisfy the technical specifications described in the previous section.
The three types of WLS fibres are the following:
Bicron BCF91A fibres produced in two preforms, one with 408 fibres
labelled lote 12070, and other with 92 fibre, lote 12071.
Kuraray Y11(200)MSJ fibres produced from a single preform.
Pol.Hi.Tech S250-100 fibres produced in two preforms, one with 272
fibres labelled pn 13434, and the other with 228 fibres, pn 13435.
The results are presented for the global production of each individual type of
fibre, as well as for the individual preforms. This individualisation of the preform was
introduced because large fluctuations from preform to preform were observed in
some cases. A brief description of the experimental conditions for the measurement
of each parameter is given. For each test, all types of fibres were measured and
handled in the same way. For the optical characterisation of the fibres, our usual
measurement test bench “Fibrometer”  was used.
2.1 Light output, attenuation length and fluctuations
The light output and attenuation length was measured for 50 fibres of each
preform of Bicron and Pol.Hi.Tech fibres, summing up to 100 fibres of each type, and
100 Kuraray fibres.
The light output values used as reference were taken from the mean I140 of a
sample of 45 Y11(200)MSJ fibres. Moreover, 16 of those fibres were used to check
the stability of the measurement setup during all the evaluation period, which took
place during the whole month of January 1999.
The fibres light output was measured between 10 and 200 cm far from the
PMT in steps of 5 cm. The attenuation length is taken from a single exponential fit
between 70 and 190 cm.
The results are summarised in table I, in which ´total´ refers to the results of
the 2 preforms combined.
Table I Light output at a distance of 140 cm from the PMT (I140) and
attenuation length (Latt) of the fibres of the three producers. The Y11(200)MSJref are
the fibres used as reference for I140.
TYPE I140/I140ref RMS(I140) Latt (cm) RMS(Latt)
BCF91A MC(12070) 88% 4.2% 261 6.2%
BCF91A MC(12071) 101% 4.8% 302 6.3%
BCF91A MC (total) 95% 8.3% 282 9.9%
S250 (13434) 79% 4.3% 230 5.3%
S250 (12435) 80% 5.5% 235 5.0%
S250 (total) 79% 5.0% 232 5.3%
Y-11(200)MSJ 103% 2.5% 303 4.2%
Y-11(200)MSJ ref 100% 2.2% 294 3.5%
Figure 1 - Attenuation length distributions for the 3 types of WLS fibres (in cm).
Figure 2 – Light output (I140/I140ref) distributions for the 3 types of fibres.
It can be seen in Table I that the fluctuations (RMS) of the BCF91A MC
fibres in the attenuation length and light output are higher than 8%. The respective
distributions are shown in Figures 1 and 2. A global inspection of the plots shows that
the large value of the fluctuations results from the different average values of the two
Bicron preforms, each of them giving origin to each of the two peaks in the top-left
plot of the figures. The difference in the attenuation length from one preform to the
other is 41 cm, which has also a substantial impact in the light output measured at 140
cm, where a difference of 13% was measured, as can be seen in Table I. So, the fibres
did not present values of RMS within the requested criteria, in spite of the average
light output and attenuation length fulfil the criteria.
From the values presented in Table I we can also see that the S250 fibres of
the 2 preforms present similar optical properties, with a light output 20% below the
reference fibres and attenuation length of 232 cm, 18 cm lower than the requested.
There was not found any significative difference between preforms. These fibres did
not satisfy the attenuation length requirements, and were on the edge of the minimal
request in the case of the light output average value.
The Y11(200)MSJ fibres presented values similar to the ones of the reference
fibres, as expected, because the reference fibres were also Y11(200)MSJ.
2.2 Diameter and eccentricity
The diameter of a fibre is measured at a given point using a micrometer with a
precision of + 0.005 mm. The fibre is rolled inside the micrometer to find the
maximum and minimum diameter at that same point.
Figure 3 Diameter and eccentricity are evaluated in several points of the fibres,
measuring Dmin and Dmax in each point.
The maximum diameter allowed is 1.02 mm, and the minimum 0.98 mm,
while the eccentricity defined by Dmax – Dmin (see figure 3) has to be less or equal to
0.03 mm. The measurement was performed in 5 fibres of each preform in 6 points per
fibre. For Pol.Hi.Tech, 8 fibres per preform were measured. Table II summarises the
results of the measurement, showing the number of points out of the tolerance. In this
table the numbers in parenthesis (n/m) represent the number n of points found out of
limits when m points were measured.
Table II Mechanical tolerances of the fibres of the three producers. The values
measured out of range are presented in % and in the format (n/m) where m is the
number of points measured and n the number of points found out of range.
Diam out Diam above Diam bellow Eccentricity out
of range upper limit lower limit of range
BCF91A MC(12070) 40% (12/30) 3% (1/30) 37% (11/30) 3% (1/30)
BCF91A MC(12071) 0% (0/30) 0% (0/30) 0% (0/30) 0% (0/30)
BCF91A MC (total) 20% (12/60) 1,6% (1/60) 18% (11/60) 1,6 (1/60)
S250 (13434) 75% (36/48) 75% (36/48) 0% (0/30) 4% (2/48)
S250 (12435) 100% (48/48) 98% (47/48) 0% (0/30) 17% (8/48)
S250 (total) 88% (84/96) 86% (83/96) 0% (0/60) 10% (10/96)
Y-11(200)MSJ 0% (0/30) 0% (0/30) 0% (0/30) 0% (0/30)
The S250 fibres presented also some defects like "bubbles" for which the
diameter can be as high as 1.14mm with a consequently high eccentricity well above
0.03mm. In result, the criteria on these parameters were not satisfied.
In the Bicron BCF91A fibres the same kind of defects were found, but with
much smaller values.
The Y11(200)MSJ fibres are the only ones that fully satisfy both parameters.
2.3 Mechanical stress
Five fibres of each preform of Bicron and Pol.Hi.Tech and 6 from Kuraray
were used in this test. The light output of the fibres was measured with the fibres
“straight” and was measured again after they were bent with a 7-cm diameter loop, as
shown in Figure 4, and the results were compared.
10 20 30 130 140 150 160 170 180
Before bending Fibre
10 20 30 130 140 150 160 170 180
PM to the ones before
7 cm diameter
Figure 4 Geometry used for the measurement of the effect of the mechanical stress.
The light output at 30 cm is used to normalise the ratio of
I(140)bent/I(140)straight. This ratio is a measure of the loss in transmission through the
bent region of the fibre, due to mechanical stress. Table III summarises the results,
showing the ratio R(140)= I(140)bent/I(140)straight obtained just after bending (0 days)
and after 6 days of bending.
Table III Light loss due to mechanical stress just after bending the fibres with a 7cm
diameter loop and 6 days later. The values shown in the table are the ratios
0 days 6 days
BCF91A MC(12070) 97% 98%
BCF91A MC (12071) 97% 97%
BCF91A MC (total) 97% 97%
S250 (13434) 91% 86%
S250 (12435) 83% 54%
S250 (total) 88% 70%
Y-11(200)MSJ 96% 96%
The light loss due to mechanical stress for the BCF91A MC and
Y11(200)MSJ fibres is less than 4%.
Some of the S250 fibres presented light losses higher than 30% just after
bending. Those extreme losses increased to 50-60% six days after bending. It is clear
that the S250 fibres do not satisfy our criterium on this parameter.
2.4 Radiation damage
Five fibres of each Bicron and Pol.Hi.Tech preforms and 10 fibres of the
Kuraray preform were irradiated with a total dose of 175 Krad at a dose rate of 2.4
The ratio of I140 after and before irradiation is shown just after the end of
irradiation and after 7 days of recovery (Table IV).
Table IV Results of the radiation damage test. Ratio of the light output I140
Time of recovery
0 days 7 days
BCF91A MC (12070) 0.856 0.864
BCF91A MC (12071) 0.833 0.838
S250 100 (13434) 0.604 0.823
S250 100 (13435) 0.599 0.791
Y11(200)MSJ 0.875 0.906
The BCF91A MC fibres show a light output loss of 16% due to ionising
radiation without recovery after one week.
S250 fibres present 40% light loss just after the end of the irradiation, with a
recovery of 21% after one week, resulting in a 19% loss after one week.
The Y11(200)MSJ fibres presented just after irradiation a light output loss of
13% recovering 4% after one week, resulting on a 9% loss after one week.
Both the Y11(200)MSJ and BCF91A MC fibres satisfy the requirements for
this parameter. The S250 fibres do not satisfy the criterium just by a small margin.
The S250 fibres failed to the request on the mean value of the attenuation
length, diameter, eccentricity, mechanical stress and radiation damage.
The BCF91A MC fibres failed to satisfy the fluctuation of the average
light output and attenuation length, diameter and eccentricity.
The Y11(200)MSJ fibres satisfied all the criteria defined for the tendering.
So, the final choice was to equip the TILECAL/ATLAS Calorimeter with
Y11(200)MSJ fibres, because those fibres were the only type which fulfilled the
requirements relative to all the parameters.
 A. Gomes et al., “Routing of the fibres in module 0”, ATLAS Internal
 ATLAS Tile Calorimeter TDR, CERN/LHCC/96-42