Reflectivity Spectra for Commonly Used Reflectors

Abstract

Monte Carlo simulations play an important role in developing and evaluating the performance of radiation detection systems. To accurately model a reflector in an optical Monte Carlo simulation, the reflector's spectral response has to be known. We have measured the reflection coefficient for many commonly used reflectors for wavelengths from 250 nm to 800 nm. The reflectors were also screened for fluorescence and angular distribution changes with wavelength. The reflectors examined in this work include several polytetrafluoroethylene (PTFE) reflectors, Spectralon, GORE diffuse reflector, titanium dioxide paint, magnesium oxide, nitrocellulose filter paper, Tyvek paper, Lumirror, Melinex, ESR films, and aluminum foil. All PTFE films exhibited decreasing reflectivity with longer wavelengths due to transmission. To achieve $> 0.95$ reflectivity in the 380 to 500 nm range, the PTFE films have to be at least 0.5 mm thick—nitrocellulose is a good alternative if a thin diffuse reflector is needed. Several of the reflectors have sharp declines in reflectivity below a cut-off wavelength, including ${\rm TiO}_{2}$ (420 nm), ESR film (395 nm), nitrocellulose (330 nm), Lumirror (325 nm), and Melinex (325 nm). PTFE-like reflectors were the only examined reflectors that had reflectivity above 0.90 for wavelengths below 300 nm. Lumirror, Melinex, and ESR film exhibited fluorescence. Lumirror and Melinex are excited by wavelengths between 320 and 420 nm and have their emission peaks located at 440 nm, while ESR film is excited by wavelengths below 400 nm and the emission peak is located at 430 nm. Lumirror and Melinex also exhibited changing angular distributions with wavelength.

Full text

490 IEEE TRANSACTIONS ON NUCLEAR SCIENCE, VOL. 59, NO. 3, JUNE 2012
Reflectivity Spectra for Commonly Used Reflectors
Martin Janecek
Abstract—Monte Carlo simulations play an important role in
developing and evaluating the performance of radiation detection
systems. To accurately model a reflector in an optical Monte Carlo
simulation, the reflector’s spectral response has to be known.
We have measured the reflection coefficient for many commonly
used reflectors for wavelengths from 250 nm to 800 nm. The
reflectors were also screened for fluorescence and angular distri-
bution changes with wavelength. The reflectors examined in this
work include several polytetrafluoroethylene (PTFE) reflectors,
Spectralon, GORE diffuse reflector, titanium dioxide paint, mag-
nesium oxide, nitrocellulose filter paper, Tyvek paper, Lumirror,
Melinex, ESR films, and aluminum foil. All PTFE films exhibited
decreasing reflectivity with longer wavelengths due to transmis-
sion. To achieve
reflectivity in the 380 to 500 nm range,
the PTFE films have to be at least 0.5 mm thick—nitrocellulose
is a good alternative if a thin diffuse reflector is needed. Several
of the reflectors have sharp declines in reflectivity below a cut-off
wavelength, including
(420 nm), ESR film (395 nm), nitro-
cellulose (330 nm), Lumirror (325 nm), and Melinex (325 nm).
PTFE-like reflectors were the only examined reflectors that had
reflectivity above 0.90 for wavelengths below 300 nm. Lumirror,
Melinex, and ESR film exhibited fluorescence. Lumirror and
Melinex are excited by wavelengths between 320 and 420 nm and
have their emission peaks located at 440 nm, while ESR film is
excited by wavelengths below 400 nm and the emission peak is
located at 430 nm. Lumirror and Melinex also exhibited changing
angular distributions with wavelength.
Index Terms—Fluorescence, Lambertian reflection, reflection
coefficient, specular reflection.
I. INTRODUCTION
S
CINTILLATING crystals emit light in a broad range of
wavelengths, with many of them having their peak emis-
sions located between 375 and 480 nm [1]–[6]. In order to con-
vert this optical signal into a sizable electrical signal, the light
has to be directed into a photodetector by the means of sur-
rounding the scintillating crystal with a reflective material. This
reflector should maximize the light collection and the reflector
has to be chosen to provide high reflection coefficients at the
scintillator’s emission wavelengths.
Monte Carlo simulations play an important role in developing
and evaluating the performance of radiation detection systems.
The simulations offer a way of evaluating the system as a whole
Manuscript received September 23, 2011; revised December 14, 2011; ac-
cepted January 04, 2012. Date of publication March 12, 2012; date of current
version June 12, 2012. This work was supported by the Director, Office of Sci-
ence, Office of Biological and Environmental Research, Biological Systems
Science Division of the U.S. Department of Energy under Contract DE-AC02-
05CH11231.
The author is with the Lawrence Berkeley National Laboratory, Berkeley, CA
94720 USA (e-mail: mjanecek@lbl.gov).
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.2012.2183385
Fig. 1. Schematic of an integrating sphere. Note that the light enters at a slight
angle in relation to the examined sample
in order to be able to detect the
specular reflection from the sample. The inside surface of the integrating sphere
is coated with a diffuse reflector creating a uniform angular distribution at the
photodetector. To force the light to reflect multiple times within the integrated
sphere, baffles (i.e., light barriers) are located between the illuminated sample
and the photodetector as well between the in-port and the photodetector, elimi-
nating any direct optical paths.
by exploring system parameters without the added cost of con-
structing the system or any of its components. For these simula-
tions, the reflection’s angular distribution [7] and the reflection
coefficient (as a function of wavelength) has to be known for the
results to be accurate.
Reflectivity measurements are generally performed with
an integrating sphere, where the entire angular distribution
of the reflected light is collected and thus contributes to the
photodetector response. As the light is reflected many times
within the integrating sphere before detection, it is important
to coat the inside of the sphere with a highly reflective “white”
material. The reflection data acquired using this technique do
not always provide sufficient information to perform an accu-
rate Monte Carlo simulation—some reflectors are fluorescent
or change their angular reflectance distributions with incidence
angle. Integrating sphere materials include barium sulfate
[8]–[14], magnesium oxide (MgO) [8], [11]–[17], and
polytetrafluorethylene (PTFE) based reflectors [18]–[20], and
these reflector materials have been extensively studied. Other
reflectors that are frequently used in optical systems include
titanium dioxide
paint [11], [16], [21], Tyvek
®
paper
[22], [23], ESR (Enhanced Specular Reflector) film [24], [25],
and Spectralon [19], [26]. In addition, some commonly used
reflectors have not been reported on in the literature, including
nitrocellulose, GORE
®
, Lumirror
®
, and Melinex
®
.
The aim of this work is to measure the reflection coefficients
as a function of wavelength for the most common reflectors used
in the radiation detection field, and to screen them for fluores-
cence and angular distribution changes with wavelength.
II. B
ACKGROUND
We previously measured the angular distribution of the
reflected light at a fixed wavelength (440 nm) [7] and found
0018-9499/$31.00 © 2012 IEEE

JANECEK: REFLECTIVITY SPECTRA FOR COMMONLY USED REFLECTORS 491
ESR film and aluminum foil to be specular reflectors. The
angular distributions of Spectralon, PFTE films, GORE
®
dif-
fuse reflector, magnesium oxide, titanium dioxide paint, and
nitrocellulose filter are best described as Lambertian (diffuse)
reflectors. All Lambertian reflectors that were measured had
a high
Lambertian component at low incidence
angles
, but a specular component appeared at high
incidence angles. Lumirror
®
, Melinex
®
, and Tyvek
®
paper
were measured to have reflection distributions that could not
be described by a linear combination of specular and diffuse
reflection distributions [7].
The angular reflection distribution measurements for com-
monly used reflectors [7] were performed at 440 nm, dictated
by the use of a 440-nm solid-state laser [CrystaLaser
®
, Reno,
NV], and we determined the reflection coefficient of the reflec-
tors by integrating the reflected light collection over the full
of solid angle. However, the reflection coefficient was only de-
termined at one single wavelength. In this work, we measure the
relative reflection coefficient for commonly used reflectors over
a large range of wavelengths. The results are normalized to the
reflectivity of white reflection standards made out of Spectralon.
There are several confounding effects when measuring the re-
flection coefficient as a function of wavelength for a reflector: 1)
the reflector can exhibit fluorescence, shifting some of the inci-
dent light to a longer wavelength, and 2) the angular distribution
can change with wavelength, so measuring the reflectivity at a
specific angle can give misleading results. Because of this, we
have in this work 1) screened all reflectors for fluorescence, 2)
compared the reflectors’ specular and diffuse reflection behavior
over wavelength, and 3) measured the reflectivity with an inte-
grating sphere (which is insensitive to changes in the angular
distribution of the reflectivity).
III. M
ETHODS
A. Reflectors
The reflectors we examined in this work are summarized in
Table I. The reflectors’ reflection coefficients at 440 nm were de-
termined from literature values or manufacturer data—if avail-
able—or from our previous measurements [7]. The thicknesses
of the reflectors were measured and are reported on in the table.
To achieve adequate thickness for the reflectivity measure-
ments, several of the reflectors were measured in multiple
layers, including ACE Teflon
®
tape, glossy PTFE tape,
Tetratex
®
film, and Tyvek
®
paper. The magnesium oxide
powder was painted in a 1-mm thick layer onto a black plastic
holder by first dissolving the powder with ethanol. The ni-
trocellulose reflector was examined in the same manner as
described in [7] (i.e., by measuring the angular distribution over
the full
of solid angle and integrating over the entire light
distribution), however, after the original paper had already been
published. The conclusions from these measurements were that
the nitrocellulose reflector is an excellent reflector with
of the light being reflected in a Lambertian light distribution,
and the reflection coefficient was determined to be 103% of the
reflection coefficient for four layers of ACE Teflon
®
tape. The
ESR film was examined for both the front and the back surfaces
after all protective films and glue had been removed (with
TABLE I
E
XAMINED
REFLECTORS
ethanol). The ESR film front-side is in this work defined as the
side of the film that originally had a protective film covering
it, but with no glue layer, and the backside is defined as the
side that originally contained a glue layer as well as a protec-
tive film. Measurements were performed with both VM2000
and VM3000 ESR films. The Spectralon sample, which was
bought from Ocean Optics as a white reflection standard for
this project, is manufactured by Labsphere
®
and has the same
model number as Labsphere’s SRS-99 (Spectralon Reflection
Standard) white reflection standard. The only physical differ-
ence between these two standards is their diameters—25.4 mm
for the WS-1-LS versus 50.8 mm for the SRS-99.
B. Fluorescence
The excitation light for our fluorescence measurements
was produced by a 75W Xenon lamp [Oriel Research Arc
Lamp, Oriel Instruments, Stratford, CT], which has a usable
wavelength range of
to 2500 nm. A narrow band of
wavelengths was selected through a SP-2155 monochromator
[Princeton Instruments, Trenton, NJ] before being collected
at the exit-slit of the monochromator by 19 optical fibers (ar-
ranged in a vertical column). The monochromator was set to
entrance and exit slits, which is equivalent to
dispersion. The optical fibers were UV/VIS fibers with a 0.22
numeric aperture (NA). One of the optical fibers illuminated
a calibrated solid-state photodetector [S2281, Hamamatsu,
Japan], and this light was used to correct for the light intensity
variations across the wavelength spectrum of the Xenon lamp
and monochromator as well as any temporal variations. We
used a Keithley 6517A (Keithley Instruments, Inc., Cleveland,
OH) digital multimeter (DMM) to monitor the current on
the solid-state photodetector. The other 18 optical fibers are
bundled together and illuminated the sample, as illustrated in
Fig. 2. Since the light exiting the optical fibers will diverge
due to the limited numeric aperture (NA) of the optical fibers,

492 IEEE TRANSACTIONS ON NUCLEAR SCIENCE, VOL. 59, NO. 3, JUNE 2012
Fig. 2. Fluorescent measurement setup for reflector samples. The sample is po-
sitioned with stepper motors, and the reflected light from the sample is collected
and focused into the spectrometer with two quartz lenses. The reflected light is
measured over a
wavelength spectrum with a CCD. The figure is
not to scale.
a collimating lens was placed at the exit of the optical fiber
bundle to collimate the light onto the reflector.
The reflectors were attached to a flat surface, which was cov-
ered with black tape to minimize the back surface’s impact on
the fluorescence measurements. The tape was analyzed in the
setup and verified to be non-fluorescent. The collection beam
angle (in relation to the reflector) was equal to the incidence
angle. The light reflected off the sample and any fluorescent
signal that was produced was collimated into a parallel beam
by a 50.8-mm diameter, 50-mm focal length quartz lens, lo-
cated 50 mm from the sample. The parallel light beam was then
focused by a second quartz lens into a SP-2156 spectrometer
[Princeton Instruments, Trenton, NJ] with a 100B charge-cou-
pled device (CCD) [Acton Research Corporation, Acton, MA]
cooled to
. The setup is illustrated in Fig. 2. The oper-
ation of the setup and all data collection was controlled by a
LabVIEW program [National Instruments, Austin, TX].
We illuminated the reflector samples one wavelength at a
time, stepping from 220 nm to 600 nm in 5-nm steps. At each ex-
citation wavelength, an optical emission spectrum from 200 nm
to 1000 nm was collected. Since the CCD can only record
in a single acquisition, and since we could eliminate any
second or higher order reflections from the emission spectra by
using order-sorting filters, the spectrum was collected in three
separate spectra: 200–360 nm (blue), 360–620 nm (green), and
620–1000 nm (red). The blue spectra were acquired with no
order-sorting filters, while the green and red spectra used cut-off
filters located at 320 nm and 590 nm, respectively.
C. Reflectivity Measurements at Fixed Angles
To measure the reflectors’ specular and diffuse reflectivity
components as a function of wavelength, we modified the flu-
orescence setup described in the previous section. The illumi-
nation side of the setup was identical to the fluorescence setup,
with the exception of that the monochromator was set to
entrance and exit slits. These slit widths are equivalent to
dispersion. The reflectors were attached to a flat sur-
face, which was covered with black tape to minimize the back
surface’s impact on the reflectivity measurements. For the spec-
ular samples (i.e., the ESR films, Aluminum foil, and Tyvek
®
Fig. 3. Modified reflectivity measurement setup. The Xenon lamp’s light is
focused with two lenses into a monochromator, where a narrow band of wave-
lengths is selected. Most of the light is then guided to the reflector, where a
collimating lens (CL) focuses the light onto the reflector sample, and a smaller
portion of the light is monitored in a single strand of the optical fibers with
a photodiode for normalization purposes
. The sample is attached
to an Aluminum block, and the reflected light is measured in a second photo-
diode
. The photodiode currents are measured with digital multimeters
(DMMs). The figure is not to scale.
paper), the reflector sample was placed in relation to the inci-
dent light beam so the incidence angle (the angle between the
optical fibers’ direction to the normal of the reflector) was equal
to the reflection angle (the angle between the normal of the
reflector to the centerline of the collection lens), making sure
that all specular light was directed onto the center of a second
S2281 photodiode. This photodiode was read out by a Keithley
485 DMM (Keithley Instruments, Inc., Cleveland, OH). For the
diffuse samples, measurements were performed both with the
reflection angle equal to the incidence angle (i.e., a specular
setup), and a setup where no specular light was allowed to make
it onto the photodiode (i.e., a diffuse setup). In practice, the
sample was placed at 45
and 60 (relative to the incident beam)
for the specular and diffuse setups, respectively, with the inci-
dent light beam and the photodetector being placed 90
apart.
For the specular measurements, the photodetector was placed
at a distance that created a cone with a 6.0
half-angle for the
photodetector, which is equivalent to
of the solid
angle, and for the diffuse measurements the distance between
the sample and the photodetector was decreased, which led to
a19
photodetection half-angle, which is equivalent to a 5.5%
solid angle coverage.
The readout of the DMMs and the movement of the
monochromator were controlled by a LabVIEW program, and
the acquired data was saved to a text-file for post-processing.
Each current measurement was calculated as an average of
10 individual current measurements, and the value was only
accepted if the standard deviation within the ten samples was
below 1% of the average value. If the standard deviation of the
current was above 1%, indicating transient currents and not
a steady-state current, the current was re-measured until the
condition was met.
We illuminated the reflector samples one wavelength at a
time, stepping from 230 nm to 800 nm in 2-nm steps. No order-
sorting filters (to remove higher order light) were used. We esti-
mate the absence of order-sorting filters to produce an error for
our results of
.
The measured reflected light intensity was normalized at
each wavelength against a white standard (SRS-99), i.e., a
sample that has a well-defined reflection coefficient at certain
wavelengths, enabling us to translate our relative reflectivity

JANECEK: REFLECTIVITY SPECTRA FOR COMMONLY USED REFLECTORS 493
Fig. 4. Control panel (screen shot) for the reflectivity measurements performed
with the Lambda 950 UV/VIS spectrophotometer and an integrating sphere. See
article text for details.
values to absolute values across the wavelength spectrum.
Since the reflection coefficient for the white standard is defined
from 250 nm to 2500 nm, we only report on our measured
reflection coefficients down to 250 nm. We had access to two
white standards, and we compare the results obtained from
both these standards. Both white standards are made out of
Spectralon, a polytetraflouroethylene (PTFE) based Lambertian
reflector [19], [26]. Each reflection spectrum measurement was
performed at least five times (with different illumination spots
on the sample) to evaluate consistency.
Since these measurements do not collect all the light (i.e., the
full
of solid angle), the reflection coefficient curves were
normalized to literature values at 440 nm. If the normalized
specular and the diffuse measurement curves were equal, we as-
sumed that the angular distribution does not change with wave-
length.
D. Reflectivity With an Integrating Sphere
We also measured all the reflector samples in a Lambda 950
UV/VIS spectrophotometer [PerkinElmer, Inc., Waltham, MA].
The screen shot of the control panel illustrating the setup is
shown in Fig. 4. A deuterium lamp is used for the shorter wave-
lengths
and a Tungsten lamp is used for the longer
wavelengths when illuminating the reflector sample. The en-
trance and exit slits on the monochromator were set to pro-
duce 1.0 nm dispersion. The sample is placed at
incidence
angle, and all the reflected light is collected by the integrating
sphere to a photomultiplier tube. A second beam, created from
the incidence beam with a beam splitter, is used to correct for
temporal intensity fluctuations in the measurements. The reflec-
tion coefficient was measured every 2 nm. The measurements
were normalized to the white standard (SRS-99). Each measure-
ment was performed three times to evaluate consistency.
IV. R
ESULTS
A. Fluorescence
The only reflectors listed in Table I that exhibited fluores-
cence were the ESR films, Lumirror
®
, and Melinex
®
. All of
these three reflectors exhibited strong fluorescence and their flu-
orescence spectra are shown in intensity plots in Fig. 5. Fig. 5(a)
shows the fluorescence plot for the ESR film, Fig. 5(b) shows
fluorescence for the glue that was originally attached to the
backside of the ESR film, and Fig. 5(c) shows the fluorescence
for Melinex
®
, respectively. All the ESR films exhibited very
similar fluorescence, and the only noticeable difference in all
the ESR film measurements was that the VM3000 film had a
slightly stronger fluorescent signal for excitations below 280 nm
compared to the VM2000 film. There was virtually no differ-
ence between the front- and backsides of each ESR film. The
Lumirror
®
and Melinex
®
fluorescence were virtually identical,
and only the Melinex
®
fluorescence plot is therefore displayed.
A fluorescence signal was detected from both sides (and for
both versions) of the ESR film. The emission peak for the ESR
film’s fluorescent signal is located at 430 nm, and is produced
by wavelengths shorter than 400 nm. The glue that was removed
from the backside of the ESR films was measured separately
for fluorescence on a (non-fluorescent) black tape. The glue’s
fluorescent signal has its emission peak located at 290 nm, and
the glue is excited by wavelengths between 250 and 285 nm.
Lumirror
®
and Melinex
®
have their emission peaks located at
440 nm, and are excited by wavelengths between 320 and 420
nm. Profiles through the emission (and excitation) maxima are
shown in Figs. 6 and 7; Fig. 6 shows the profiles for the ESR film
and the ESR glue, and Fig. 7 shows the profiles for Lumirror
®
and Melinex
®
.
B. Specular Reflectors
The reflection coefficients as a function of wavelength are
displayed in Fig. 8 for the specular reflectors, i.e., the ESR
films and the Aluminum foil, as well as for the Tyvek
®
paper.
The protective films from the ESR films and the glue layer
from the ESR backsides were removed (using ethanol) prior to
any reflection (or fluorescence) measurements. As can be seen
in Fig. 8, the VM3000 front-side and the VM2000 backside
are virtually identical. In the same way, the VM3000 back-
side and the VM2000 front-side are virtually identical above
380 nm—below 380 nm the VM2000 front-side has a lower
reflection coefficient. The difference between the front- and
backsides for both films above 380 nm is minimal, though we
did measure a 5 to 6 nm difference in the cut-off reflection
wavelength between the two sides (at
).
Tyvek
®
paper displays high reflection coefficients for all mea-
sured wavelengths, staying above 95% for wavelengths above
355 nm, and above 90% for wavelengths above 300 nm. Alu-
minum foil displays a flat reflectivity curve, with reflection co-
efficients between 70% and 80% above 265 nm.
C. Diffuse Reflectors
The reflection coefficients as a function of wavelength for
Spectralon (WS-1-LS), GORE
®
diffuse reflector, MgO powder,
paint, and nitrocellulose filter paper are displayed in

494 IEEE TRANSACTIONS ON NUCLEAR SCIENCE, VOL. 59, NO. 3, JUNE 2012
Fig. 5. Fluorescence for A) ESR film (VM2000 front-side is shown), B) ESR glue (from VM2000) and C) Melinex
®
. The excitation wavelengths are displayed on
the vertical axis, and the resulting intensity in the emission is shown along the horizontal axis. A reflection line is clearly visible in all plots, where the excitation
and emission wavelengths are equal. The highest intensity in each plot (excluding the reflection line) has been used to normalize the intensity scales.
Fig. 6. Excitation and emission profiles for ESR film and ESR glue. The pro-
files are taken from data from Figs. 5(a) and 5(b) and pass through the maximum
value in the fluorescence spectra.
Fig. 7. Excitation and emission profiles for Lumirror
®
and Melinex
®
. The pro-
files are taken from data from Fig. 5(c) and the Lumirror
®
data (not shown) and
pass through the maximum value in the fluorescence spectra.
Fig. 9. A zoom-in image of the wavelength range from 360 to
500 nm is also shown in the figure. The GORE
®
diffuse reflector
exhibits an excellent reflectivity
for wavelengths be-
tween 250 and 500 nm, while MgO and Spectralon—two
materials commonly used as white reflectors—drop below 90%
reflectivity for wavelengths shorter than 280 nm. Nitrocellulose
and
paint exhibit sharp declines in their reflectivity for
wavelengths below 330 and 420 nm, respectively.
The reflection coefficients as a function of wavelength
for PTFE tapes (glossy PTFE tape, ACE Teflon
®
tape, and
Tetratex
®
film) are displayed in Fig. 10. As described in de-
tail in the following paragraph, all films have a decreasing
reflectivity with increasing wavelength due to transmission.
The transmission for a single layer of glossy PTFE film is also
presented in the figure, as well as the sum of the reflected and
transmitted signal, i.e., 1 - absorbed. The transmission was
measured by placing the reflector in the incident beam’s path
into the integrating sphere.
The PTFE reflectors are bulk reflectors (as opposed to sur-
face reflectors) and longer wavelengths penetrate deeper into (or
though the material). For instance, a single layer of the
thick glossy PTFE tape allows a significant amount of light
Fig. 8. Reflection coefficient for the front- and backsides of two versions of
ESR film, Tyvek
®
paper, and Aluminum foil as a function of wavelength. The
insert shows a zoom-in of the reflection data between 360 and 500 nm. The data
were acquired in the specular setup of the
Reflectivity measurements at fixed
angles setup and normalized to the reflectivity values in Table I. The maximum
error was measured to be
, and the average error for all the data points
presented above was measured to be
.
Fig. 9. Reflection coefficient for several diffuse reflectors as a function of wave-
length. The insert shows a zoom-in of the reflection between 360 and 500 nm.
The data were acquired in the Reflectivity with an Integrating Sphere setup. The
measurements have not been normalized to the reflectivity values in Table I since
the Lambda 950 instrument is calibrated.
transmitted for wavelengths longer than 380 nm. By
increasing the number of layers, the reflectivity increases, and
at for instance 440 nm, goes from a reflection coefficient of 85%
(1 layer) to 92.6% (2 layers), to 94.4% (4 layers), and to 96.2%
(8 layers). In order to achieve at least 95% reflectivity for scin-
tillator light emissions (i.e., 380–500 nm), the examined PTFE

JANECEK: REFLECTIVITY SPECTRA FOR COMMONLY USED REFLECTORS 495
Fig. 10. Reflection coefficient for PTFE tapes as a function of wavelength. The
insert shows a zoom-in of the reflection between 360 and 500 nm. The data were
acquired in the
Reflectivity with an Integrating Sphere setup. The measurements
have not been normalized to the reflectivity values in Table I since the Lambda
950 instrument is calibrated.
Fig. 11. Reflection coefficient as a function of wavelength for Lumirror
®
for
a variety of reflection angles. The 45
-setup measured the specular component
of the reflectivity spectra, while the other angle-setups measure various compo-
nents of the diffuse spectra. The data were acquired in the Reflectivity measure-
ments at fixed angles setup. The measurements have not been normalized to the
reflectivity values in Table I.
films will need to be at least 0.5 mm thick. A good alternative,
if a thin diffuse reflector is needed, is to use nitrocellulose.
D. Changing Angular Distributions
The Lumirror
®
and Melinex
®
reflection coefficients as a func-
tion of wavelength were measured to be different in the spec-
ular and diffuse reflection measurements. Because of this, we
decided to measure the reflection over a larger range of reflec-
tion angles. The results from these measurements are displayed
in Fig. 11 for Lumirror
®
. Melinex
®
exhibited similar behavior.
The results in Fig. 11 have not been normalized to the reflec-
tion coefficient at 440 nm and hence show the intensity vari-
ations with reflection angle. Since the incident beam and the
photodetector are place 90
apart, the 45 -setup is a specular
measurement, while all the other angle-setups measure various
components of the diffuse spectra. As can be seen, the slope
above 400 nm, as well as the intensity of the reflection peak at
260 nm (in relation to the baseline of the reflection spectrum)
varies with reflection angle. Since these two reflectors are fluo-
rescent, a measurement in the integrating sphere produces arti-
ficially high reflection coefficients in the 320 to 420 nm wave-
length range, where the quantum efficiency for the photomulti-
Fig. 12. Reflection coefficient as a function of wavelength for Lumirror
®
and
Melinex
®
. The data were acquired in the Reflectivity with an Integrating Sphere
setup, and the fluorescence contribution to the signal between 320 and 420 nm
(shown in grey shading) produces an artificially high signal since the quantum
efficiency for the photomultiplier tube is higher for the emitted wavelength than
for the incident wavelength. The measurements have not been normalized to the
reflectivity values in Table I.
Fig. 13. Reflection coefficient as a function of wavelength for the two white
standars, WS-1-LS and SRS-99 (both Spectralon). Note that the “Reflectivity”-
axis scale is from 0.85 to 1.0 (and not 0 to 1) in order to enhance the differences
between the curves. The data for the SRS-99 were supplied by the manufacturer.
The data for the WS-1-LS were supplied by the manufacturer and acquired in the
Reflectivity with an Integrating Sphere setup. The measurements have not been
normalized to the reflectivity values in Table I, since the Lambda 950 instrument
is calibrated.
plier tube is higher for the fluorescent light compared to incident
light, see Fig. 12. Lumirror
®
and Melinex
®
have a decreased re-
flectivity below a cut-off wavelength of 325 nm.
E. White Standards
The manufacturers’ reflection coefficient data for the two
white standards are displayed in Fig. 13. The measured reflec-
tion coefficients for the WS-1-LS standard, when normalized
to the SRS-99 standard, are also displayed in the figure.
V. D
ISCUSSION
The high reflectance value of nitrocellulose presented in
Table I indicates that the reflectivity of the ACE Teflon
®
tape
sample—the reflector the nitrocellulose sample was normalized
against—is lower than reported values [18]–[20]. We do not be-
lieve that this is due to improper care of the samples (they were
stored at stable temperatures and out of sunlight, the surfaces
were clean, contact with chemicals and exposure to UV light
was avoided, etc.) but attribute it instead to transmission of light
through even thick layers of PTFE tape. The measurements
presented in Fig. 10 show that there is a very large amount of
transmission through PTFE films, and that the values reported
in the literature must assume very thick or dense samples. The

496 IEEE TRANSACTIONS ON NUCLEAR SCIENCE, VOL. 59, NO. 3, JUNE 2012
reflection coefficients at 440 nm were in this work measured
to be 0.944 for ACE Teflon
®
tape and 0.962 for nitrocellulose,
respectively, which gives nitrocellulose a 1.9% higher reflection
coefficient at 440 nm than four layers of ACE Teflon
®
tape.
The intensity scales shown on the right of each fluorescence
plot in Fig. 5 should not be used as absolute scales, as the size of
the reflectors and the reflectors positioning in relation to the in-
cident beam and the light collection lens play a very significant
role in the detected light intensity. Although the fluorescence
setup can be used to measure the reflection coefficient by fil-
tering out any unwanted fluorescence and higher order signals,
the alignment of the collimated light from the optical fibers, the
sample, and the “sweet spot” from which the light collection
lenses focus the light into the spectrometer (see Fig. 2), is a
very delicate operation and small variations in the positions of
any of these three parameters easily lead to inconsistent (i.e.,
inaccurate) results in intensity. For this reason, although this
setup is able to detect and characterize fluorescence emission
and excitation wavelengths, it should not be used for absolute
intensity measurements. For instance, Lumirror
®
and Melinex
®
exhibit nearly identical fluorescence spectra, and Lumirror
®
is
only slightly brighter compared to Melinex
®
under a black light
(excited at 365 nm), yet the measured intensity in Fig. 5(c) for
Melinex
®
exhibited a four times weaker signal than for the Lu-
mirror
®
fluorescence data (not shown).
Fluorescent light is isotropic in its nature, and the direction-
ality of the light will therefore be lost for the light that inter-
acts with the fluorescent reflector and is reemitted. Therefore,
the fluorescence exhibited by the ESR films, Lumirror
®
, and
Melinex
®
can be beneficial if the directionality of the light is not
an issue and if the photodetector has higher quantum efficien-
cies at the emission wavelengths compared to the incident wave-
lengths. This effect is clearly demonstrated for the Lumirror
®
and Melinex
®
reflectors, as shown in Fig. 12.
The temporal behavior of the fluorescent emissions was mea-
sured in an IBH FluoroHub [HORIBA Jobin Yvon Inc, Edison,
NJ] by exposing the reflectors to pulsed LED light close to the
reflectors’ maximum excitation wavelengths. The pulsed LEDs
we had access to have emission peaks located at 268 nm (used
for the ESR glue), 311 nm (ESR film), and 370 nm (Lumirror
®
and Melinex
®
). The fluorescent light emission as a function of
time was measured for each reflector through a monochromator
in which the signal was filtered into a narrow bandwidth (1 nm)
centered at the emission peak at 430 nm (ESR film) or 440 nm
(Lumirror
®
and Melinex
®
). The ESR glue was measured at 300
nm (instead of 290 nm) to minimize the light contributions from
the LED’s emissions. Each measurement was performed until
we accumulated 10,000 counts in the peak bin. The count rate
for the ESR glue signal was several orders of magnitude smaller
due to lower efficiencies of the light and the photodetector
and due to the lower fluorescence, and only 5,000 counts were
acquired for this material. Since the background signal is pro-
portional to the acquisition time, the ESR glue also exhibited
a much lower signal-to-noise ratio (by
). The temporal
emission curves are displayed in Fig. 14. The ESR film’s flu-
orescence was measured to have a half-life of 14 ns, the ESR
glue’s fluorescence was measured to have a half-life of 7 ns,
while the Lumirror
®
and Melinex
®
fluorescent half-lives were
measured to be
.
Fig. 14. Temporal behavior of the fluorescent light emitted from the reflectors
that exhibited fluorescence. Note that the vertical scale is logarithmic.
Several reflectors exhibited “cut-offs” for the reflectivity
for shorter wavelengths, including
(420 nm), ESR film
(395 nm), nitrocellulose (330 nm), Lumirror
®
(325 nm) and
Melinex
®
(325 nm). The lower reflection coefficients below
the cut-off wavelengths have to be taken into consideration
when pairing up a scintillator with a reflector, by taking into
account the scintillator’s emission spectrum and the reflector’s
reflection coefficients at these wavelengths.
Our measurements showed great repeatability, where the re-
flection coefficient between several runs in the Lambda 950
UV/VIS spectrophotometer were typically within 0.1% of each
other, and never more than 1%, and the accuracy in our mea-
surements are thus closely tied to the accuracy of the reference
standard. The data in Fig. 13 indicate that the white reflection
standards often used for these measurements can be a source
of errors, as their true reflectivity can differ significantly from
the calibration provided by the manufacturer. For example, the
measured reflectivity of WS-1-LS is 2% lower at 380 nm and
5% lower at 300 nm than the values provided by the manufac-
turer. The manufacturer-provided reflectivity values for SRS-99
contained a small
dip between 260 and 290 nm that
appears to be an artifact. Using these standards with their man-
ufacturer-provided reflectivity values leads to
reflec-
tivity for wavelengths below 300 nm (WS-1-LS), or a small peak
between 260 and 290 nm (SRS-99). Both reflection standards
were “certified reflection standards”, although only the SRS-99
was provided with a “reflection calibration certificate” (and cal-
ibrated against a NIST traceable standard within the last year).
Other groups have measured decreased reflectivity in Spectralon
samples over time [27], including samples that have not been
exposed to high illumination levels. In fact, Spectralon samples
stored in darkness have shown to degrade over time, and this is
the most likely cause for the results presented in Fig. 13.
VI. C
ONCLUSIONS
We have measured the reflection coefficient for several
commonly used reflectors. The reflectors were also screened
for fluorescence and changing angular distribution with
wavelength. The highest reflectivity for short wavelengths
was measured for the PTFE based reflectors, with
the GORE
®
diffuse reflector having the highest reflectivity over
the greatest wavelength range. PTFE based reflectors were the

JANECEK: REFLECTIVITY SPECTRA FOR COMMONLY USED REFLECTORS 497
only examined reflectors that had reflectivity for wave-
lengths below 300 nm, but all PTFE films exhibited decreasing
reflectivity with increasing wavelength due to increased trans-
mission for longer wavelengths. To achieve
reflectivity,
the PTFE films have to be at least 0.5 mm thick. If a thinner
diffuse reflector is needed, nitrocellulose is a good alternative.
Several of the reflectors have sharp declines in reflectivity
below a cut-off wavelength, including
(420 nm), ESR film
(395 nm), nitrocellulose (330 nm), Lumirror
®
(325 nm), and
Melinex
®
(325 nm). Lumirror
®
, Melinex
®
, and ESR film ex-
hibited strong fluorescence, and Lumirror
®
and Melinex
®
also
exhibited changing angular distributions with wavelength.
A
CKNOWLEDGMENT
The mechanical construction of the setups presented in
this paper was done by David Wilson. Dr. Jacob C. Jonsson,
Johan Borglin, and Ilya Zorikhin-Nilsson assisted in the fluo-
rescence and reflection measurements. Dr. William W. Moses
contributed with invaluable discussions and suggestions.
R
EFERENCES
[1] M. J. Weber and R. R. Monchamp, “Luminescence of
Bi4Ge3O12—spectral and decay properties,” J. Appl. Phys., vol. 44,
pp. 5495–5499, 1973.
[2] C. L. Melcher, J. S. Schweitzer, C. A. Peterson, R. A. Manente, and
H. Suzuki, “Crystal growth and scintillation properties of the rare earth
orthosilicates,” in Proc. Inorganic Scintillators and Their Applications
Conf., 1996, pp. 309–315.
[3] E. V. D. Van Loef, P. Dorenbos, C. W. E. Ven Eijk, K. W. Kramer,
and H. U. Gudel, “Scintillation properties of
crystals:
Fast, efficient and high-energy-resolution scintillators,” Nucl. Instrum.
Methods Phys. Res. A, vol. 486, pp. 254–258, 2002.
[4] J. T. M. d. Haas and P. Dorenbos, “Advances in yield calibration of
scintillators,” IEEE Trans. Nucl. Sci, vol. 55, no. 3, pp. 1086–1092,
Jun. 2008.
[5] C. L. Melcher and J. S. Schweitzer, “Cerium-doped lutetium oxy-
orthosilicate—a fast, efficient new scintillator,” IEEE Trans. Nucl.
Sci., vol. 39, no. 4, pp. 502–505, Aug. 1992.
[6] B. C. Grabmaier, “Crystal scintillators,” IEEE Trans. Nucl. Sci., vol.
NS-31, no. 1, pp. 372–376, Feb. 1984.
[7] M. Janecek and W. W. Moses, “Optical reflectance measurements for
commonly used reflectors,” IEEE Trans. Nucl. Sci, vol. 55, no. 4, pp.
2432–2437, Aug. 2008.
[8] W. Budde, “Standards of reflectance,” J. Opt. Soc. Amer., vol. 50, p.
17, 1960.
[9] J. B. Gillespie, J. D. Lindberg, and L. S. Laude, “Kubelka-munk optical
coefficients for a barium sulfate white reflectance standard,” Appl. Opt.,
vol. 14, pp. 807–809, 1975.
[10] E. M. Patterson, C. E. Shelden, and B. H. Stockton, “Kubelka-munk
optical properties of a barium sulfate white reflectance standard,” Appl.
Opt., vol. 16, pp. 729–732, 1977.
[11] F. Grum and G. W. Luckey, “Optical sphere paint and a working stan-
dard of reflectance,” Appl. Opt., vol. 7, pp. 2289–2294, 1968.
[12] W. Erb, “Requirements for reflectance standards and the measurement
of their reflection values,” Appl. Opt., vol. 14, no. 2, 1975.
[13] Bureau Central de la Commission Internationale de l’Eclairage, Re-
view of Publications on Properties and Reflection Values of Material
Reflection Standards (Paris, 1979), CIE Publication 46 (TC-2.3).
[14] J. B. Schutt, J. F. Arens, C. M. Shai, and E. Stromberg, “Highly re-
flecting stable white paint for the detection of ultraviolet and visible
radiations,” Appl. Opt., vol. 13, pp. 2218–2221, 1974.
[15] F. Benford, G. P. Lloyd, and S. Schwarz, “Coefficients of reflection of
magnesium oxide and magnesium carbonate,” J. Opt. Soc. Amer., vol.
38, pp. 445–445, 1948.
[16] F. Vratny and J. J. Kokalas, “The reflectance spectra of metallic oxides
in the 300 to 1000 millimicron region,” Appl. Spectrosc., vol. 16, 1, pp.
76–184, 1962.
[17] A. Zecchina, M. G. Lofthouse, and F. S. Stone, “Reflectance spectra of
surface states in magnesium oxide and calcium oxide,” J. Chem. Soc.
Faraday Trans., vol. 1, 1975.
[18] V. R. Weidner and J. J. Hsia, “Reflection properties of pressed poly-
tetrafluoroethylene powder,” J. Opt. Soc. Amer., vol. 71, no. 7, pp.
856–861, 1981.
[19] B. Waldwick, C. Chase, and B. Chang, “Increased efficiency and per-
formance in laser pump chambers through use of diffuse highly reflec-
tive materials,” Proc. SPIE, vol. 6663, pp. 66630N.1–66630N.7, 2007.
[20] A Guide to Reflectance Coatings and Materials Labsphere, Reflections
Newsletter, 1998, pp. 8–11.
[21] A. L. Companion and R. E. Wyatt, “The diffuse reflectance spectra
of some titanium oxides,” J. Phys. Chem. Solids, vol. 24, no. 8, pp.
1025–1028, 1963.
[22] J. Tickner and G. Roach, “PHOTON an optical Monte Carlo code for
simulating scintillation detector responses,” Nucl. Instrum. Methods
Phys. Res. B, vol. 263, pp. 149–155, 2007.
[23] A. Chavarria, “A Study on the reflective properties of Tyvek in air and
Underwater,” Ph.D. dissertation, Duke University, Durham, NC, 2007.
[24] M. F. Weber, C. A. Stover, L. R. Gilbert, T. J. Nevitt, and A. J. Oud-
erkirk, “Giant birefringent optics in multilayer polymer mirrors,” Sci-
ence, vol. 287, no. 5462, pp. 2451–2456, 2000.
[25] D. Motta, C. Buck, F. X. Hartmann, T. Lasserre, S. Schonert, and U.
Schwan, “Prototype scintillator cell for an in-based solar neutrino de-
tector,” Nucl. Instrum. Methods Phys. Res. A
, vol. 547, no. 2-3, pp.
368–388, Aug. 2005.
[26] D. H. Goldstein, D. B. Chenault, and J. L. Pezzaniti, “Polarimetric char-
acterization of spectralon,” Proc. SPIE, vol. 3754, pp. 126–136, 1999.
[27] W. Möller, K.-P. Nikolaus, and A. Höpe, “Degradation of the diffuse
reflectance of spectralon under low-level irradiation,” Metrologia, vol.
40, pp. 212–215, Feb. 2003.