Dependence of polytetrafluoroethylene reflectance on thickness at visible and ultraviolet wavelengths in air
S. Ghosh, J. Haefner, J. Martín-Albo, R. Guenette, X. Li, A.A. Loya Villalpando, C. Burch, C. Adams, V. Álvarez, L. Arazi, I. J. Arnquist, C. D. R Azevedo, K. Bailey, F. Ballester, J. M. Benlloch-Rodríguez, F. I. G. M. Borges, N. Byrnes, S. Cárcel, J. V. Carrión, S. Cebrián, E. Church, C. A. N. Conde, T. Contreras, G. Díaz, J. Díaz, M. Diesburg, J. Escada, R. Esteve, R. Felkai, A. F. M. Fernandes, L. M. P. Fernandes, P. Ferrario, A. L. Ferreira, E. D. C. Freitas, A. Goldschmidt, J. J. Gómez-Cadenas, D. González-Díaz, R. M. Gutiérrez, K. Hafidi, J. Hauptman, C. A. O. Henriques, J. A. Hernando Morata, P. Herrero, V. Herrero, Y. Ifergan, B. J. P. Jones, M. Kekic, L. Labarga, A. Laing, P. Lebrun, N. López-March, M. Losada, R. D. P. Mano, A. Martínez, M. Martínez-Vara, G. Martínez-Lema, A. D. McDonald, F. Monrabal, C. M. B. Monteiro, F. J. Mora, J. Muñoz Vidal, P. Novella, D. R. Nygren, B. Palmeiro, A. Para, J. Pérez, M. Querol, A. Redwine, J. Renner, J. Repond, S. Riordan, L. Ripoll, Y. Rodríguez García, J. Rodríguez, L. Rogers, B. Romeo, C. Romo-Luque, F. P. Santos, J. M. F. dos Santos, A. Simón, M. Sorel, T. Stiegler, J. F. Toledo, J. Torrent, A. Usón, J. F. C. A. Veloso, R. Webb, R. Weiss-Babai, J. T. White, K. Woodruff, N. Yahlali
PPrepared for submission to JINST
Dependence of polytetrafluoroethylene reflectance onthickness at visible and ultraviolet wavelengths in air
The NEXT Collaboration
S. Ghosh, , a J. Haefner , a J. Martín-Albo, , b R. Guenette, X. Li, A.A. LoyaVillalpando, , c C. Burch, C. Adams, V. Álvarez, L. Arazi, I.J. Arnquist, C.D.R Azevedo, K. Bailey, F. Ballester, J.M. Benlloch-Rodríguez, F.I.G.M. Borges, N. Byrnes, S. Cárcel, J.V. Carrión, S. Cebrián, E. Church, C.A.N. Conde, T. Contreras, G. Díaz, J. Díaz, M. Diesburg, J. Escada, R. Esteve, R. Felkai, , , A.F.M. Fernandes, L.M.P. Fernandes, P. Ferrario, , A.L. Ferreira, E.D.C. Freitas, A. Goldschmidt, J.J. Gómez-Cadenas, , , d D. González-Díaz, R.M. Gutiérrez, K. Hafidi, J. Hauptman, C.A.O. Henriques, J.A. Hernando Morata, P. Herrero, V. Herrero, Y. Ifergan, , B.J.P. Jones, M. Kekic, L. Labarga, A. Laing, P. Lebrun, N. López-March, M. Losada, R.D.P. Mano, A. Martínez, M. Martínez-Vara, G. Martínez-Lema, , , e A.D. McDonald, F. Monrabal, , C.M.B. Monteiro, F.J. Mora, J. Muñoz Vidal, P. Novella, D.R. Nygren, , c B. Palmeiro, , A. Para, J. Pérez, M. Querol, A. Redwine, J. Renner, J. Repond, S. Riordan, L. Ripoll, Y. RodríguezGarcía, J. Rodríguez, L. Rogers, B. Romeo, , C. Romo-Luque, F.P. Santos, J.M.F.dos Santos, A. Simón, M. Sorel, T. Stiegler, J.F. Toledo, J. Torrent, A. Usón, J.F.C.A. Veloso, R. Webb, R. Weiss-Babai, , f J.T. White, , g K. Woodruff, N. Yahlali Department of Physics and Astronomy, Iowa State University, 12 Physics Hall, Ames, IA 50011-3160, USA Argonne National Laboratory, Argonne, IL 60439, USA Department of Physics, University of Texas at Arlington, Arlington, TX 76019, USA Institute of Nanostructures, Nanomodelling and Nanofabrication (i3N), Universidade de Aveiro, Campus deSantiago, Aveiro, 3810-193, Portugal Fermi National Accelerator Laboratory, Batavia, IL 60510, USA Nuclear Engineering Unit, Faculty of Engineering Sciences, Ben-Gurion University of the Negev, P.O.B.653, Beer-Sheva, 8410501, Israel Nuclear Research Center Negev, Beer-Sheva, 84190, Israel Lawrence Berkeley National Laboratory (LBNL), 1 Cyclotron Road, Berkeley, CA 94720, USA Ikerbasque, Basque Foundation for Science, Bilbao, E-48013, Spain a Corresponding author. b Now at the Instituto de Física Corpuscular (IFIC), Valencia, Spain. c Now at the California Institute of Technology, Pasadena, CA, United States. d NEXT Collaboration’s co-spokesperson. e Now at the Weizmann Institute of Science, Rehovot, Israel. f On leave from Soreq Nuclear Research Center, Yavneh, Israel. g Deceased. a r X i v : . [ phy s i c s . i n s - d e t ] S e p Centro de Investigación en Ciencias Básicas y Aplicadas, Universidad Antonio Nariño, Sede Circunvalar,Carretera 3 Este No. 47 A-15, Bogotá, Colombia Department of Physics, Harvard University, Cambridge, MA 02138, USA Laboratorio Subterráneo de Canfranc, Paseo de los Ayerbe s/n, Canfranc Estación, E-22880, Spain LIBPhys, Physics Department, University of Coimbra, Rua Larga, Coimbra, 3004-516, Portugal LIP, Department of Physics, University of Coimbra, Coimbra, 3004-516, Portugal Department of Physics and Astronomy, Texas A&M University, College Station, TX 77843-4242, USA Donostia International Physics Center (DIPC), Paseo Manuel Lardizabal, 4, Donostia-San Sebastian,E-20018, Spain Escola Politècnica Superior, Universitat de Girona, Av. Montilivi, s/n, Girona, E-17071, Spain Departamento de Física Teórica, Universidad Autónoma de Madrid, Campus de Cantoblanco, Madrid,E-28049, Spain Instituto de Física Corpuscular (IFIC), CSIC & Universitat de València, Calle Catedrático José Beltrán, 2,Paterna, E-46980, Spain Pacific Northwest National Laboratory (PNNL), Richland, WA 99352, USA Instituto Gallego de Física de Altas Energías, Univ. de Santiago de Compostela, Campus sur, Rúa XoséMaría Suárez Núñez, s/n, Santiago de Compostela, E-15782, Spain Instituto de Instrumentación para Imagen Molecular (I3M), Centro Mixto CSIC - Universitat Politècnica deValència, Camino de Vera s/n, Valencia, E-46022, Spain Centro de Astropartículas y Física de Altas Energías (CAPA), Universidad de Zaragoza, Calle PedroCerbuna, 12, Zaragoza, E-50009, Spain
E-mail: [email protected] , [email protected] Abstract: Polytetrafluoroethylene (PTFE) is an excellent diffuse reflector widely used in lightcollection systems for particle physics experiments. However, the reflectance of PTFE is a functionof its thickness. In this work, we investigate this dependence in air for light of wavelengths 260 nmand 450 nm using two complementary methods. We find that PTFE reflectance for thicknesses from5 mm to 10 mm ranges from 92.5% to 94.5% at 450 nm, and from 90.0% to 92.0% at 260 nm.We also see that the reflectance of PTFE of a given thickness can vary by as much as 2.7% withinthe same piece of material. Finally, we show that placing a specular reflector behind the PTFEcan recover the loss of reflectance in the visible without introducing a specular component in thereflectance.Keywords: Time projection chambersArXiv ePrint: 1234.56789 ontents
Polytetrafluoroethylene (PTFE), often referred to by the brand name Teflon, is a commonly usedplastic in particle physics experiments (see, for example, Ref. [1–3]) due to its excellent diffusereflectance and affordability. The NEXT experiment [4], designed to search for neutrinoless doublebeta decay, uses PTFE as a diffuse reflector around the drift volume of the high pressure gaseousxenon time projection chamber (TPC) to improve light collection. The measurements describedbelow will inform the design decisions for the upcoming phases of NEXT, including NEXT-100[1, 4] and a future tonne-scale experiment [5].While PTFE is excellent for light collection, its radiopurity can be a concern for low-backgroundexperiments [6], especially at large scale where a significant quantity of material is needed. Moreover,PTFE can absorb gaseous xenon [7], leading to a loss of active detector volume. Minimizing theamount of PTFE in such experiments is therefore an attractive solution, provided that the lightcollection remains high.Noble gases scintillate primarily in the vacuum ultraviolet (VUV) range of the electromagneticspectrum, and some investigations of PTFE reflectance in this region have been done in the past.Reflectance of PTFE at 175 nm at a single 5 mm thickness was studied thoroughly in Ref. [8], andreflectance in liquid xenon for thicknesses from 1 mm to 9.5 mm was investigated in Ref. [9]. Further,reflectance as a function of angle was studied in Ref. [10]. However, as many experiments (including– 1 –EXT) use tetraphenyl butadiene (TPB) wavelength shifter to shift the 175 nm scintillation light ofxenon to the visible ( ∼
420 nm), studies of reflectance as a function of PTFE thickness in the visibleare also of high interest. While it is generally accepted that PTFE has around 99% reflectance [8] inthe visible, information on its variation as a function of thickness is also relevant.In this work, the reflectance variation of PTFE with thickness is investigated with two differentmethods. The first method uses an integrating sphere with a Cary 7000 Universal MeasurementSpectrophotometer. The integrating sphere is an Agilent Diffuse Reflectance Accessory . Thesecond method involves enclosed boxes made of PTFE of different thicknesses where the reflectance isstudied by measuring the signal as a function of distance from the end of the box. The measurementspresented here are made for 260 nm and 450 nm. In order to investigate the reflectance of PTFE, several 12 in ×
24 in sheets of different thicknesseswere purchased. All sheets for the main reflectance measurements were sourced from ePlastics .The thicknesses used are 4.76 mm, 6.53 mm, 7.94 mm, and 9.53 mm. For brevity, we refer to these,respectively, as 5 mm, 6 mm, 8 mm, and 10 mm.The pieces cut from the PTFE sheets were all treated and cleaned in the same way. One sideof each piece was sanded by hand with 300, then 1000, and finally 2000 grit sandpaper to makethe surface uniform and matte. Each piece was then cleaned in an ultrasonic bath in a solution ofdeionized (DI) water and Alconox for 15 minutes, finally followed with an ultrasonic bath in pure DIwater for 15 minutes. The pieces were air-dried overnight before measurement. All cleaning wascompleted in a single day, each step immediately following the last. Pieces were generally cleanedand measured within a few days of being cut, although sometimes remeasured several days or weekslater. Pieces were stored in a cabinet wrapped in tin foil. This is comparable to cleaning methodsused in particle physics experiments (cf. [9]). Other surface treatments for radon removal will beexplored in future works. The first method for determining reflectance makes use of a spectrophotometer (SPM), the AgilentCary 7000 Universal Measurement Spectrophotometer (Figure 1), owned and maintained bythe Harvard Center for Nanoscale Systems (CNS). The SPM is coupled to a reflecting sphere andallows measurement of the reflectance of a given sample relative to some reference, as a function ofwavelength.
The primary measurement in the spectrophotometer comes from averaging the reflectance of twopieces cut from two different locations on the same PTFE sheet. These are in fact the same piecesthat are used in the measurements described in Section 2.2, which are carried out on boxes built outof PTFE. As such the pieces are referred to as the "box back" and the "box side," as they are the – 2 – igure 1 : Left: The Cary 7000 spectrophotometer of the Harvard Center for Nanoscale Systems(CNS) during a measurement. The compartment containing the integrating sphere is covered with ablackout shroud to prevent light leaks. The PTFE disks used to assess systematic variation can beseen in the transparent sample boxes on the right. Right: A closer look at the PTFE disks.
250 300 350 400 450 500Wavelength (nm)9596979899100101102103 R e f l e c t a n c e ( % )
10 mm8 mm6 mm5 mm
Figure 2 : Example of reflectance measurements from the spectrophotometer for different PTFEthicknesses relative to the 10-mm sample. The purple and blue dashed lines represent the valuesof the LED used in method 2 (260 nm and 450 nm, respectively). Determination of values anduncertainties is described in Section 2.1.2. The differences only rise to the few percent level at most.respective sides and backs of boxes of a given thickness. Our error bars consisted of a componentestimating the variance measurement to measurement of the same piece, and variation betweendifferent pieces cut from the same sheet, added in quadrature. The variation between different piecescame from the variance of the box back versus box side measurements. To assess variations betweenrepeated measurements, 3-inch diameter disks (visible on the right side of Fig. 1) were cut from the– 3 –TFE sheets of different thicknesses and measured three times with the SPM. A difference of up to2.7% (relative to 1 cm reference) is observed when studying the reflectance of different pieces fromthe same sheet.The SPM measurements of the reflectance, relative to the reference sample of 10 mm, for thedifferent thicknesses are shown in Figure 2. The error bars are calculated as described previously.
A summary of the different reflectances measured with the SPM is presented in Table 1. The resultsare shown for 260 and 450 nm in order to directly compare with the second method (see Section 2.3),where LEDs of these wavelengths are used.The data points are the average of single measurements of box back and side panel pieces. Theuncertainties come from the standard deviation of triplicate measurements of PTFE disks of thecorresponding thicknesses, added in quadrature to the standard deviation between the box side andback measurements. The variation between subsequent disk measurements is generally around0.1%; the difference between box back and side is by far the dominant contribution. Please notethat different pieces are used to obtain the mean of the data points versus the error bars. The disksprovide our estimate of the variation between subsequent measurements of the same sample, butthey are not the pieces whose actual reflectances are reported, only their standard deviation. Thereason for not averaging in the means of the disk measurements is that degradation moved them farfrom the box reflectances, see Section 2.4.At 260 nm, the reflectance of all thicknesses measured is the same within errors, indicatingthat the UV reflectance of PTFE is not very dependent on the thickness of the material. At 450 nm,between 5 and 10 mm we measure a difference of 3.6 ± Table 1 : Relative reflectance (with respect to the reference sample of 10 mm) of PTFE box piecesfor various thicknesses at 260 nm and 450 nm measured using the spectrophotometer. Compare toFigure 2. See Section 2.1.2 for specifics of value and error bar determination.Thickness 260 nm 450 nm10 mm 100 ± ± ± ± ± ± ± ± The two significant disadvantages of the SPM method described in Section 2.1 are that themeasurements are relative and that the small area of the piece that is measured (approximately0.2 cm ) makes the measurements highly sensitive to potential reflectance variation across the pieceor the sheet from which they were cut. The alternative method described in this section, referred toas the box method , uses a PTFE cuboid of a total surface area of 749 cm . This method provides anabsolute reflectance and averages potential reflectance variations within the PTFE samples, as itcovers a much larger surface area. – 4 – eflecting boxSiPMsLED inside pinholeInsertion distance 25cm 7cm Figure 3 : Side-view schematic of the box setup. The inside dimensions of the box are7 cm × ×
25 cm. The board holding the LED and SiPMs system is inserted through the openend of the box. The amount of light collected is measured as a function of the distance of the boardinsertion to extract the reflectance of the PTFE box.
The box method uses several cuboids with inner dimensions of 7 cm × ×
25 cm made of PTFEpieces of a given thickness. The boxes are open at one end to allow the insertion of a printed circuitboard with a light-emitting diode (LED) and 4 silicon photomultipliers (SiPMs). The board ismounted on a sliding post such that the insertion distance can be changed. Measurements of the lightcollected as a function of the distance of insertion are used to extract the reflectance. A schematicview of the setup is shown in Figure 3.Two different ThorLabs LEDs are used: an LED260W with light emission peaked at 260 nm,and an LED450LW with light emission peaked at 450 nm. These were chosen to be as close aspossible to the light wavelengths of interest: the 175 nm VUV scintillation light of xenon, andthe 420 nm TPB-shifted light. In all measurements, the LED is outfitted with a small pinholemade of opaque black plastic to collimate the light, reducing the importance of near-field effectsand making the simulation of the setup easier. The SiPMs used for the main measurements arethe 3 × UV-sensitive Hamamatsu S13370-3050CN. A different model, the 3 × Hamamatsu S13360-3050CS SiPMs, which are designed for visible light and offer no sensitivityto UV light, were also used to cross-check the measurements as both these SiPMs offer similarefficiency at blue wavelengths. The SiPMs are mounted at different locations on the board in orderto average over any possible effects from non-uniformities in the PTFE, as shown in Figure 4.Measurements are taken for insertion distances ranging from 0 cm to 21 cm at intervals of 3 cm.The pulses used to power the UV LED had a frequency of 50 Hz with a maximum voltage of6.5 V and minimum voltage of 0 V. The ramp-up time was 5 ns and the pulse width was 200 ns. Thepulses used to power the blue LED had a frequency of 50 Hz with a maximum voltage of 3.1 V andminimum voltage of 0 V. The ramp-up time was 5 ns and the pulse width was 200 ns. We basedthese characteristics on the stated tolerances of the LEDS in their product sheets , and found that the – 5 – igure 4 : Left: Picture of the LED-SiPMs system mounted on a printed circuit board. The LED,located in the center, is covered with a pinhole of 2 mm diameter. The black tape is intended toreduce reflection of light from the board. Right: Picture of the box setup with the readout boardattached to a movable rod to allow variation of the insertion distance inside the box. The wholesetup is inside a black box to ensure that external light does not affect the measurement. Figure 5 : Example SiPM pulse for box measurement. Yellow curve is SiPM response, pink curve isLED trigger.LEDs performed stably under these parameters.The signals from the SiPMs are directly fed into an oscilloscope. When measuring the signalsfrom the SiPMs, about 1000 pulses are acquired at each position for each SiPM, and the averagepulse area for all the pulses is calculated. A typical pulse is shown in Figure 5. A data point, as seenin Figure 6, is then obtained by summing the average areas for each SiPM. As a single bias voltage isapplied in parallel across all four SiPMs, all detector signals were weighted by the inverse of theirovervoltage. The SiPMs were provided with documentation indicating results of measurements oftheir individual breakdown voltages. The statistical error on each point is the standard deviation ofthe ∼ √ · ns for all points.The systematic error is determined by repeating the measurement for each box four times, with the– 6 –epeats spaced apart by an hour or so each. The systematic error was of order 1 V · ns for all points. In order to extract the reflectances from the box measurements, the data are compared withsimulations of the setup for reflectances ranging from 80% to 99.5% in 0.5% steps. The simulationsare made using Geant4 [12], where the box geometry includes the PTFE with a diffuse reflectancecomponent only. Note that measurements with the spectrophotometer support the fact that anyspecular component to the reflectance of the PTFE is negligible, see Section 3. The LED simulation,described in Appendix A, is dependant on the exact emission spectrum used. Therefore, in order toaccount for potential inaccuracies in the simulation of the LED emission spectrum, two differentemission models, chosen to capture the most significant variation seen in the measured profile ofthe LED, were simulated, and the differences were used to assess a systematic uncertainty due tothe LED emission profile. The statistical uncertainties of the individual models were calculatedas (cid:112) n ( − nN ) , where n is the number of photons detected in the simulation and N is the totalnumber of photons injected. These were added in quadrature and divided by 2 to determine thestatistical uncertainty on the averaged simulation curve. In the blue LED with UV SiPM case, theuncertainty on the individual simulation curves was about 0.31 V · ns, with the averaging contributingan additional 10 − V · ns statistical and 0.46 V · ns systematic uncertainty. In the UV LED withUV SiPM case, the uncertainty on the individual simulation curves was about 0.72 V · ns, with theaveraging contributing an additional 10 − V · ns statistical and 0.93 V · ns systematic uncertainty. Theoutputs of the simulation are in the form of the number of photons emitted from the LED that reachthe SiPMs, based on the different PTFE reflectances assumed. The simulation for the differentreflectances was fitted to the data and the best fit was used as the reflectance value for that data set.More details on the fitting procedure are given below.In order to allow comparison of the data taken with the SiPMs (in V · ns) to the simulation(in photon counts), it is assumed that there is a constant conversion factor between the number ofphotons collected and the signal read out by the SiPMs. Therefore, the scale factor parameter ( α )is left free to vary in the fit. This α parameter is then the number of V · ns expected per simulationphoton, and soaks up effects such as the intensity of the LED and the efficiency of the SiPMs. The α scale factor should not change significantly between measurements of different reflectances, and thefit is performed simultaneously on all the data sets for the different reflectances using the same α parameter. Fig. 6 illustrates how the fit of the different simulated reflectances is performed on thedata set.Although we expect the α value to be relatively stable, there is some possibility of variationcoming in particular from the gain of the SiPMs changing in response to temperature. A cross checkwas run with a non-simultaneous fit allowing a different alpha value for each curve. Our results wereconsistent within error bars with those in Fig. 6. Measurements of reflectances using the boxes were done for thicknesses of 5 mm, 6 mm, 8 mm,10 mm. Figure 6 shows the results obtained at 260 nm and at 450 nm. The simulation fit to the dataprovides the reflectance for each thickness. – 7 – .52.02.53.03.5 V n s / mm / = 6.68 = 2940 MC fit 91.0 +1.532.56 %MC fit 91.5 +1.321.73 %MC fit 91.5 +1.671.20 %MC fit 90.0 +1.901.82 %Data 1 cmData 8 mmData 6 mmData 5 mm c m mm mm mm V n s / mm / = 3.94 = 6776 MC fit 94.0 +0.841.14 %MC fit 93.0 +1.260.98 %MC fit 92.5 +1.021.27 %MC fit 91.5 +1.391.19 %Data 1 cmData 8 mmData 6 mmData 5 mm c m mm mm mm Figure 6 : Results for box measurements and fits. Upper row: The y-axis for the 260 nm UV LEDwith UV SiPM (left) and 450 nm blue LED with UV SiPM (right) gives the amount of light measuredper unit photosensitive area of the SiPMs with pulse area measured in V · ns. Error on the data pointsincludes statistical uncertainties of the SiPM pulses and systematic uncertainties from repeating themeasurement four times. Error on the simulations includes statistical uncertainties and systematicuncertainty evaluated by varying the LED emission profile using two distinct measurements. Lowerrow: These plots show the residuals from the fits to data in the upper row. The y-axis gives thedifference between a data point and its corresponding simulated point (simulation minus data), andthe x-axis gives the distances as in the upper row.For both plots in Fig. 6, it can be seen that the simulations reproduce the data relatively well.The agreement is worse for the UV LED results at small insertion, where the simulation producesa slower decrease in the number of photons observed. This could be explained by the effect ofthe simulated UV LED emission profile (see Appendix A), which could affect results where thephotons emitted at larger angles would have a greater chance to be absorbed when undergoing ahigher number of reflections. This effect would be less important when the number of reflections islower and when the insertion distance is larger.For both wavelengths, the extracted reflectances of the PTFE samples are the same within errorbars. This conclusion mostly comes from the large error bars from the fit of the simulations to thedata, which could point towards inaccuracies in the simulations. The data points at 260 nm are close– 8 – able 2 : Summary of PTFE reflectances at 260 nm and 450 nm for the box and photometermeasurements. The photometer relative values are taken from Table 1, then normalized to the10 mm measurements from the box results for easy comparison. Error bars on the 10 mm valuefrom the photometer represent rescaled standard deviation between measurements.260 nm 450 nmThickness Box SPM Box SPM10 mm 91.0% + . − . ± + . − . ± + . − . ± + . − . ± + . − . ± + . − . ± + . − . ± + . − . ± Reflectances of PTFE were assessed with two different methods. The method using the SPM providedrelative reflectances with smaller errors, and the box method fitted with simulations provided absolutereflectances with larger errors. In order to compare both methods, the photometer measurementvalues were scaled such that the results of the 10 mm thick PTFE pieces, used as the reference forthe relative measurements of the SPM, matched with the 10 mm results of the box measurements.Table 2 shows the comparison of both methods, where the results are in good agreement with eachother in both the UV and the blue.
The box reflectance measurements were repeated at a later date to investigate the possibility ofreflectance variation over time. Figure 7 compares the results of these measurements to the originalones. While reflectances with the UV LED appeared to decrease slightly, reflectances with the blueLED were slightly increased. In both cases, however, the original and new fit reflectances are withinuncertainty of one another.There was also a 3% difference observed between the reflectances in the UV obtained usingthe SPM of the disks and of the box pieces. This may be attributed to the fact that while the boxeswere kept covered when they were not being used, the PTFE disks were left exposed to light andlikely degraded through UV exposure from ambient light. The timescale for this degradation wasrelatively short, around three weeks. Further investigation is required to confirm and quantify thedegradation as the cause of this disagreement. – 9 – cm 8 mm 6 mm 5 mmThickness86889092 R e f l e c t a n c e ( % ) lateearly R e f l e c t a n c e ( % ) lateearly Figure 7 : Comparison of fit reflectances of the PTFE box measurements two months apart - "early"(circles) was taken in October 2019 and "late" (squares) was taken in December 2019. Left:Comparison for the UV LED/UV SiPM configuration. Right: Comparison for the blue LED/UVSiPM configuration. The early measurements were used for the results presented in the previoussections. Fit reflectances are all either within or nearly within uncertainties of each other.
Although PTFE maintains good reflectance at relatively low thicknesses, it cannot be made arbitrarilythin and continue to function well as a reflector. It was investigated whether the mass of PTFE couldbe further reduced by introducing a backing specular reflector, without sacrificing reflectance. Thispossibility was addressed by looking at 3-inch diameter disks of PTFE backed with 3M DF2000MAspecular reflective foil. The foil is 38 microns thick. It was physically held in place by the samelever arm than holds the sample in place for the integrating sphere. This setup is shown in Figure 8.There was interest in both the resulting total reflectance and the amount of specular reflectance. Aspecular component would represent different optical properties for PTFE.In order to investigate lower thicknesses than were presented in the previous sections, twoadditional thicknesses of PTFE were sourced from Nationwide Plastics . The main differencebetween the two PTFE providers is that the Nationwide Plastics sheets are skived from PTFEbillets, while those from ePlastics are molded from PTFE resin. The two different types of PTFEare thus not directly comparable, but can still be compared to themselves before and after backingwith foil.The results of the 3M foil measurements are shown in Table 3. A study to understand thepotential introduction of a specular component to the diffuse PTFE reflector was also performed.The specular component (which should be 0 for a diffuse reflector) can be defined as the "diffuseexcluded" component. In order to measure it, a measurement of the full reflectance is comparedto a measurement where the specular reflection in the integrating sphere of the SPM is blockedout with black metal. This functionality is included with the integrating sphere. This gives the"diffuse" component of the reflectance; subtracting this component from the full measurement givesthe specular component of the reflectance. In all cases, the specular component is zero within error – 10 – igure 8 : Left: 3M foil being installed at the back of a disk of PTFE at the mouth of the integratingsphere. Right: 3M foil held on the back of a PTFE disk by the SPM lever arm. Table 3 : Comparison in spectrophotometer of reflectance at 260 nm and 450 nm between PTFEwith and without 3M backing. All measurements relative to 10 mm (unbacked) PTFE disk. Thetotal reflectance of the 3M foil alone at 450 nm (260 nm) is 101.6% (40.7%) relative to the 10 mmPTFE disk, and the specular component of the 3M foil alone is 76.9% (16.2%) relative to the 10 mmPTFE disk. Error bars come from standard deviation of repeated measurements of the same sample,done in triplicate, each time removing and replacing the sample. Samples from 5 mm to 10 mm aresourced from ePlastics, whereas 1.6 mm and 2.4 mm are Nationwide.260 nm 450 nmThickness PTFE PTFE + foil PTFE PTFE + foil10 mm 100% 99.9 ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± Measurement of the variation in reflectance of PTFE with variations in thickness and wavelength(summarized in Table 2) was presented as well as the impact of backing PTFE with specular reflectivefoil (summarized in Table 3). It was found that the variation in reflectance at 450 nm is measurable,ranging from 92% to 95% for PTFE of thickness 5 mm to 10 mm at 450 nm. At 260 nm, thevariation ranged from 91% to 92% but within errors for all thicknesses. A significant degradation ofthe PTFE reflectance in the UV is observed in PTFE disks, potentially due to exposure to ambientlight, although more investigation is required to make a definitive and quantitative statement of themagnitude or timescale of cause and effect. There are slight changes in the reflectances of the boxesover time; likewise a more focused investigation is warranted. It also found that the reflectance ofPTFE can vary up to 2.7% when measuring different locations within a single sheet of material,making careful selection of PTFE important when constructing a detector. This also emphasizesthe importance of having a method of measuring reflectance that can average over variations inreflectance over the surface of the PTFE, such as our box method. Note that an absolute differenceof 2.7% between two measurements gives a standard deviation of 2 . /√ = . Acknowledgments
The NEXT Collaboration acknowledges support from the following agencies and institutions: theEuropean Research Council (ERC) under the Advanced Grant 339787-NEXT; the European Union’sFramework Programme for Research and Innovation Horizon 2020 (2014–2020) under the GrantAgreements No. 674896, 690575 and 740055; the Ministerio de Economía y Competitividad andthe Ministerio de Ciencia, Innovación y Universidades of Spain under grants FIS2014-53371-C04,– 12 –TI2018-095979, the Severo Ochoa Program grants SEV-2014-0398 and CEX2018-000867-S, andthe María de Maeztu Program MDM-2016-0692; the Generalitat Valenciana under grants PROME-TEO/2016/120 and SEJI/2017/011; the Portuguese FCT under project PTDC/FIS-NUC/2525/2014and under projects UID/04559/2020 to fund the activities of LIBPhys-UC; the U.S. Departmentof Energy under contracts No. DE-AC02-06CH11357 (Argonne National Laboratory), DE-AC02-07CH11359 (Fermi National Accelerator Laboratory), DE-FG02-13ER42020 (Texas A&M) andDE-SC0019223 / DE-SC0019054 (University of Texas at Arlington); and the University of Texas atArlington (USA). DGD acknowledges Ramon y Cajal program (Spain) under contract number RYC-2015-18820. JM-A acknowledges support from Fundación Bancaria “la Caixa” (ID 100010434),grant code LCF/BQ/PI19/11690012. Finally, we thank Brendon Bullard, Paolo Giromini and NeerajTata for helpful discussions and assistance with preliminary measurements.
A Simulating the LED
It was found that simulating the LED as a point source fit the data poorly. To account for an extendedsource, at least 2 intensity profiles at different positions were required, as any single profile could beproduced by a point source. The LED was measured at two downstream locations using a beamprofiler (Thorlabs BP209-VIS), one near the LED and one far. The blue LED was measured at 0.1and 0.3 inches from the face of a ThorLabs beam profiler. The UV LED was measured at 0.1 and0.2 inches from the face of the profiler. The profiles measured were 1 dimensional in x and y, withthe overall profile taken as the product of the x and y profiles. I n t e n s i t y ( a . u . )
1e 4 Near y LED dist.Near y before matchNear y after match
Figure 9 : Near y profile simulations with 10,000 hits. The generated distribution before matchingis included to indicate that some of the failure to match the empirical distribution comes fromfluctuations in the generation of the hits on each plane.In order to generate simulation photons, 10,000 hits were generated on both the near and farplane, according to the measured distributions. Hits were then matched, where a "match" consists of– 13 –
50 100 150 200Insertion (mm)4000600080001000012000 p h o t o n s d e t e c t e d UV LED simulation comparison p h o t o n s d e t e c t e d Blue LED simulation comparison
Figure 10 : Comparison of simulation curves obtained with iteration A and iteration C LED profilemeasurements for the UV LED (left) and the blue LED (right). The changes in the profiles do notsignificantly impact the simulation results.a near hit and a far hit where the line between them lands correctly on the source. Some amount ofdistortion arises from the fact that generally not all hits will be able to find a match. However, thisnumber of unmatched hits is reasonably small (10% for 10,000 hits), and thus the output profiles arereasonably good matches to the physical LED.To check the effectiveness of this procedure, the output profile at the near and far planes wascompared to the measured profile by taking the RMS difference point to point between the respectivehistograms and scaling relative to the maximum height of the profile. Doing so gave around 5%disagreement in the near x and y profiles, and 3% difference in the far profiles. This held regardlessof the LED being simulated.Another cross check was to measure a third profile, then use photons generated from the firsttwo profiles to try to recover the information on the third ("interpolated") profile. This profilesat between the other two, at 0.2 in in the blue case, and 0.15 in the UV. No information on theinterpolated profile is input in the simulation. The disagreement between the interpolated profileswas, in all cases, around 10%. A sample of the near profiles can be seen in Figures 9.In order to understand the possible impact on reflectance fits, the profile measurements wererepeated for four iterations (labeled A through D). These profiles were visually inspected for bothLEDs to see how they differed, and then a pair was chosen (iteration A and iteration C) whichmaximally differed from one another. One set of simulations was run using each, and the resultswere compared (Fig. 10). As the simulations did not appear to be significantly impacted by thechange in profiles, the box fits were performed using an average of the iteration A and iteration Csimulation curves, propagating those error bars into the uncertainties.
References [1] NEXT collaboration, V. Álvarez et al.,
NEXT-100 Technical Design Report (TDR): Executive Summary , JINST (2012) T06001, [ ].[2] M. Auger et al., The EXO-200 detector, part I: Detector design and construction , JINST (2012)P05010, [ ]. – 14 –
3] LZ collaboration, D. Akerib et al.,
LUX-ZEPLIN (LZ) Conceptual Design Report , .[4] NEXT collaboration, J. MartÃŋn-Albo et al., Sensitivity of NEXT-100 to Neutrinoless Double BetaDecay , JHEP (2016) 159, [ ].[5] NEXT collaboration, C. Adams et al., Sensitivity of a tonne-scale NEXT detector for neutrinolessdouble beta decay searches , .[6] NEXT collaboration, P. Novella et al., Radiogenic Backgrounds in the NEXT Double Beta DecayExperiment , JHEP (2019) 051, [ ].[7] NEXT collaboration, L. Rogers et al., High Voltage Insulation and Gas Absorption of Polymers in HighPressure Argon and Xenon Gases , JINST (2018) P10002, [ ].[8] C. Silva, J. Pinto da Cunha, A. Pereira, V. Chepel, M. Lopes, V. Solovov et al., Reflectance ofPolytetrafluoroethylene (PTFE) for Xenon Scintillation Light , J. Appl. Phys. (2010) 064902,[ ].[9] J. Haefner et al.,
Reflectance dependence of polytetrafluoroethylene on thickness for xenon scintillationlight , Nucl. Instrum. Meth. A (2017) 86–91, [ ].[10] S. Kravitz et al.,
Measurements of Angle-Resolved Reflectivity of PTFE in Liquid Xenon with IBEX , Eur.Phys. J.
C80 (2020) 262, [ ][11] Ch. Geis et al.,
Optical response of highly reflective film used in the water Cherenkov muon veto of theXENON1T dark matter experiment , JINST (2017) P06017, [ ][12] J. Allison et al., Recent developments in Geant4 , Nucl. Instrum. Meth. A (2016) 186–225.(2016) 186–225.