Prepared for submission to JINST

Lithium-loaded Liquid Scintillator Production for thePROSPECT experiment

The PROSPECT CollaborationJ.Ashenfelter,p A.B.Balantekin,m H.R.Band,p C.D.Bass, f D.E.Bergeron,g D.Berish,j

L. J.Bignell,a N.S.Bowden,e J. P.Brodsky,e C.D.Bryan,h C.Camilo Reyes,a S.Campos,a

J. J. Cherwinka,n T.Classen,e A. J.Conant,c D.Davee,o D.Dean,i G.Deichert,h R.DiazPerez,a M.V.Diwan,a M. J.Dolinski,b A.Erickson,c M.Febbraro,i B. T. Foust,p J.K.Gaison,p

A.Galindo-Uribarri,i,k C.E.Gilbert,i,k B. T.Hackett,i,k S.Hans,a,1 A.B.Hansell,j B.Hayes,a

K.M.Heeger,p J. Insler,b D.E. Jaffe,a D.C. Jones,j O.Kyzylova,b C.E. Lane,b T. J. Langford,p,2

J. LaRosa,g B.R. Littlejohn,d X. Lu,i,k D.A.MartinezCaicedo,d J. T.Matta,i R.D.McKeown,o

M.P.Mendenhall,e P.E.Mueller,i H.P.Mumm,g J.Napolitano,j R.Neilson,b J.A.Nikkel,p

D.Norcini,p S.Nour,g D.A.Pushin,l X.Qian,a E.Romero-Romero,i,k R.Roseroa,3 D.Sarenac,l

P. T. Surukuchi,d M.A. Tyra,g R. L. Varner,i B.Viren,a C.White,d J.Wilhelmi,j T.Wise,p

M.Yeh,a Y.-R. Yen,b A. Zhang,a,3 C. Zhang,a X. Zhangd

aBrookhaven National Laboratory, Upton, NY, USAbDepartment of Physics, Drexel University, Philadelphia, PA, USAcGeorge W.Woodruff School of Mechanical Engineering, Georgia Institute of Technology, Atlanta, GA USAdDepartment of Physics, Illinois Institute of Technology, Chicago, IL, USAeNuclear and Chemical Sciences Division, Lawrence Livermore National Laboratory, Livermore, CA, USAfDepartment of Physics, Le Moyne College, Syracuse, NY, USAgNational Institute of Standards and Technology, Gaithersburg, MD, USAhHigh Flux Isotope Reactor, Oak Ridge National Laboratory, Oak Ridge, TN, USAiPhysics Division, Oak Ridge National Laboratory, Oak Ridge, TN, USAjDepartment of Physics, Temple University, Philadelphia, PA, USAkDepartment of Physics and Astronomy, University of Tennessee, Knoxville, TN, USAlInstitute for Quantum Computing and Department of Physics and Astronomy, University of Waterloo,Waterloo, ON, Canada

mDepartment of Physics, University of Wisconsin, Madison, Madison, WI, USAnPhysical Sciences Laboratory, University of Wisconsin, Madison, Madison, WI, USAoDepartment of Physics, College of William and Mary, Williamsburg, VA, USApWright Laboratory, Department of Physics, Yale University, New Haven, CT, USA

E-mail: azhang@bnl.gov, rrosero@bnl.gov

1Also at Department of Chemistry and Chemical Technology, Bronx Community College, Bronx, NY, USA.2Also at Yale Center for Research Computing, Yale University, New Haven, CT, 06520.3Correspondng authors

Abstract: This work reports the production and characterization of lithium-loaded liquid scintilla-tor (LiLS) for the Precision Reactor Oscillation and Spectrum Experiment (PROSPECT). Fifty-nine90 liter batches of LiLS (6Li mass fraction 0.082%±0.001%) were produced and samples from allbatches were characterized by measuring their optical absorbance relative to air, light yield relativeto a pure liquid scintillator reference, and pulse shape discrimination capability. Fifty-seven batchespassed the quality assurance criteria and were used for the PROSPECT experiment.

Keywords: Scintillators, scintillation and light emission processes (solid, gas and liquid scintilla-tors); Neutrino detectors; Neutron detectors (cold, thermal, fast neutrons); Liquid detectors.

Contents

1 Introduction 2

2 LiLS production 22.1 Materials and quality assurance 22.2 LiLS production procedures and storage 4

3 LiLS quality assurance measurements 63.1 Relative optical absorbance 63.2 Relative light yield 83.3 Pulse shape discrimination 10

4 Summary 13

– 1 –

1 Introduction

The Precision Reactor Oscillation and Spectrum Experiment (PROSPECT) is a reactor antineutrinoexperiment at the Oak Ridge National Laboratory (ORNL) [1]. It is deployed to make a precisemeasurement of the energy spectrum of antineutrinos emitted from the High Flux Isotope Reactor(HFIR) at short baselines in the range from 7 to 13 meters. It also provides a probe of eV-scalesterile neutrino oscillations.

The PROSPECT detector is a ∼ 2.0 m × 1.6 m × 1.2 m active rectangular volume containingabout four tons of non-flammable liquid scintillator loaded with 6Li (LiLS) to a mass fraction ofapproximately 0.1% [2]. The detector operates with minimal overburden from the HFIR building.Thin specularly reflecting panels divide the LiLS volume into an 11×14 array of 154 opticallyisolated rectangular segments (14.5 cm × 14.5 cm × 117.6 cm each) viewed on both ends by 5-inch(12.7 cm) photomultiplier tubes (PMTs).

Including a 20% contingency, the production of approximately 5000 liters LiLS was required.LiLS was chosen because it allows efficient capturing of neutrons produced by the inverse betadecay (IBD) reaction in a compact detector. 6Li has a high capture cross section for IBD neutrons,and an alpha and a triton are produced from the capture with approximately 540 keV of visibleenergy in the scintillator [3, 4]. These particles are spatially localized, and the deposited energycan be distinguished from electron-like events by pulse shape discrimination (PSD).

The collaboration developed a LiLS that fulfills the requirements of the experiment in termsof light yield, PSD, and stability. The chosen LiLS formulation is the result of about three years ofdevelopment. The production of LiLS was carried out at the Chemistry Department of BrookhavenNational Laboratory (BNL). Fifty-nine 90 liter batches of LiLS were produced over a period of ninemonths, starting in Jan. 2017.

This paper summarizes the LiLS production and its quality assurance (QA) program. Criteriafor acceptance were based on comparison to the performance of an initial LiLS batch deployedin a 50 liter, two-segment prototype detector known as PROSPECT-50 [5]. PROSPECT-50 wasdesigned to replicate essential properties of the 154-segment PROSPECT and achieved PSD andlight yield of the LiLS satisfying the PROSPECT design requirements. Each LiLS production batchwas required to be consistent with or superior to that of PROSPECT-50 in terms of absorbance at420 nm of incident light relative to air, light yield relative to a liquid scintillator reference (linearalkylbenzene, LAB), and PSD.

2 LiLS production

2.1 Materials and quality assurance

The generic formula of the LiLS consists of a nonionic surfactant, a 9.98 mol/L aqueous solution oflithium chloride (LiCl) with 95% enriched 6Li by atom, 2,5-diphenyloxazole (PPO) and 1,4-bis(2-methylstyryl) benzene (bis-MSB) in a commercial, di-isopropylnaphthalene (DIN)-based scintillator(EJ-309)1. The surfactant is an ether-based glycol from DOW Chemical. The PPO and bis-MSBwere obtained from Research Product International.

1Certain trade names and company products arementioned in the text or identified in illustrations in order to adequatelyspecify the experimental procedure and equipment used. In no case does such identification imply recommendation or

– 2 –

The EJ-309 scintillator was purchased from Eljen Technology, it was delivered in 23 drums(4600 L in total). The absorbance of a sample of each EJ-309 drum relative to air was measuredusing a UV-vis spectrophotometer (Shimadzu UV-1800). Two cylindrical quartz glass cells (10 cmpath length, 25.4 mm diameter, 1 mm wall thickness) were used in the measurements, with oneholding a 48 mL liquid sample and the other left empty for reference. Each sample’s absorbancespectrum was measured by scanning from wavelength 200 nm to 1100 nm in 1 nm steps. Figure 1shows the measured relative absorbance spectrum of an EJ-309 scintillator sample from drum 10over the wavelength range from 350 nm to 600 nm where the PMTs are sensitive. Negative valuesin absorbance are attributed to the refractive index differences between the sample (scintillator)and reference (air) in the measurements. Figure 2 shows relative absorbance of all EJ-309 samplesat 420 nm, near the most sensitive wavelength for the PMTs. Drums 2-4 had a slightly higherabsorbance due to oxygen contamination in the manufacturer’s nitrogen purging line.

Wavelength (nm)350 400 450 500 550 600

Abs

orba

nce

0

0.02

0.04

0.06

0.08

0.1

Figure 1. The relative absorbance of a raw EJ-309 scintillator sampled from drum 10.

The LiCl solution was supplied by the National Institute of Standards and Technology fromenriched lithium carbonate (95±1% 6Li by atom) and analytical grade concentrated (37±1% bymass) hydrochloric acid, produced with the chemical reaction

Li2CO3 + 2HCl→ 2LiCl + H2O + CO2. (2.1)

The LiCl solution was filtered and passed through an anion exchange chromatography column(Bio-Rad AG 1-X4, 100 to 200 mesh) to remove colored (e.g., iron) impurities. A total of 86 Lof 9.98±0.02 mol/L purified LiCl solution was prepared in six individual batches. The averageabsorbance of the purified LiCl batches is shown in Figure 3.

endorsement by the National Institute of Standards and Technology, nor does it imply that the products are necessarilythe best available for the purpose.

– 3 –

EJ309 drum number0 5 10 15 20 25

Rel

ativ

e ab

sorb

ance

at 4

20 n

m

0.006−

0.005−

0.004−

0.003−

0.002−

0.001−

0

Figure 2. The relative absorbance at 420 nm of EJ-309 scintillator from different drums.

Wavelength (nm)350 400 450 500 550 600

Abs

orba

nce

0.012−

0.01−

0.008−

0.006−

0.004−

0.002−

0

hAbsEntries 901Mean 487.3RMS 68.22

Figure 3. The average absorbance curve for the LiCl solution. The error bars represent the standard deviationcalculated from the measured samples at each wavelength.

2.2 LiLS production procedures and storage

The LiLS was produced by first purifying the raw components and then mixing in stages in adouble-jacketed 90 L Chemglass reaction vessel, shown in Figure 4. The metal-loading principleof water-based liquid scintillator (WbLS) was applied to the synthesis of LiLS [6]. Since WbLSloading is a one-step, direct aqueous/organic mixing procedure, as shown in Figure 4, no loss of 6Liduring synthesis in LS is expected. The fraction of 6Li loaded in the scintillatorwas (0.082±0.001)%by mass [7].

The ether-based surfactant was purified by thin-film vacuum distillation while the EJ-309

– 4 –

was pre-purified by the manufacturer. The reaction vessel has several injection ports made ofpolytetrafluoroethylene (PTFE) for adding chemical materials. All the tubing, filtration system,liners, and the mixing system were pre-cleaned with ethanol (Ethyl alcohol 190 proof), rinsed withultrapure water (resistivity 18.2 MW×cm) and dried with nitrogen gas. The system was sealed inan inert environment until use.

Figure 4. LiLS production system at different stages in the 90 L Chemglass: injecting of raw material intothe reactor (left), mixing in progress (middle) and synthesis completion (right). The liquid is opaque duringthe synthesis and becomes clear after synthesis is done.

All purified raw materials were introduced through different ports into the reactor at differentsynthesis stages and mixed for 2 hours. After the synthesis was completed, the LiLS was dischargedthrough a 2 µm glass filter (Whatman) situated in a 316-stainless steel filtration system and storedin a 208 L lined drum. Each drum was equipped with a 5 µm thick perfluoroalkoxy alkane innerbag for liquid storage and a 5 µm thick outer polypropylene liner for secondary containment. Themaximum storage capacity of each drum was limited to 180 L, equal to two 90 L batches, for easeof handling and overflow prevention. Nitrogen cover gas was added to the rest of the volume.All materials in contact with the LiLS are compatible with LiLS. During batch production, twoone-liter-sized samples from every batch were taken for quality assurance measurements.

During the production period, one to three batches were produced weekly with 57 batchesproduced in the first six months. The production of the final two batches was delayed by threemonths due to a temporary shortage of raw scintillator. Batches 1 and 2 were used for prototypingand material compatibility tests. All other drums were stored in a temperature-controlled ware-house at BNL and later transported in temperature-controlled trucks to ORNL for the PROSPECTexperiment. The temperature was controlled between 20 ℃ and 30 ℃.

– 5 –

3 LiLS quality assurance measurements

In this section, the measurement methods, data analysis strategies, selection criteria, and resultsare described for qualifying the LiLS samples in terms of their optical absorbance, light yield, andPSD.

3.1 Relative optical absorbance

The relative optical absorbance of a sample from each batch was measured immediately after pro-duction. Figure 5 shows average absorbance for all the measured samples. Figure 6 presents thesamples’ absorbance at 420 nm, the absorbance for a PROSPECT-50 sample is included for compar-ison. A standard deviation of 0.003 for UV absorbance was determined from measurements of ninePROSPECT-50 samples and assigned as the systematic uncertainty of each sample measurement.The large variation in the first 10 batches is attributed to the oxygen contamination of EJ-309 asdescribed earlier. The absorbance for batches 11 and 46 was clearly higher than other samples andtherefore they were rejected. The other batches were all considered acceptable with absorbancecomparable to the PROSPECT-50 sample. The rejected samples were not subjected to light yieldand PSD measurements.

Wavelength (nm)400 450 500 550 600

Abs

orba

nce

0

0.02

0.04

0.06

0.08

0.1

hAbsEntries 901Mean 375.2RMS 79.99

Figure 5. Average of the relative absorbance of all the measured LiLS samples. The error bars represent thestandard deviation calculated from the measured samples at each wavelength.

The optical absorbance of the LiLS was stable from production to deployment. The absorbanceof repeated measurements of samples from a stored drum taken over a six month period showed novariations greater than the estimated 0.003 systematic uncertainty in the 400 nm to 600 nm range.Samples from drums after shipment from BNL to ORNL show similar behavior. All drums of LiLSshipped to ORNL were accepted for deployment in the PROSPECT detector.

A measurement of the oxygen quenching effect on absorbance is shown in Figure 7. The LiLSabsorbance degrades when oxygen is introduced by bubbling air through the scintillator. Sparging

– 6 –

LiLS batch0 10 20 30 40 50 60

Abs

orba

nce

at 4

20 n

m

0.004

0.006

0.008

0.01

0.012

0.014

0.016

0.018

0.02

0.022

0.024

0.026

0.028

Figure 6. Relative absorbance at 420 nm for every measured LiLS sample (black points) compared to thePROSPECT-50 sample indicated by the red dashed line. Batches 11 and 46 were rejected due to their highabsorbance. A common 0.003 uncertainty is assigned to each point as described in the text.

the LiLS with nitrogen gas removes the oxygen and improves the absorbance [8]. To minimizequenching, all the LiLS samples were sparged with nitrogen gas at 30 mL/min for 30 min prior tothe light yield measurements.

Wavelength (nm)400 450 500 550 600 650 700 750 800

Abs

orba

nce

0

0.005

0.01

0.015

0.02

0.025

0.03LiLS before air bubbling

LiLS after air bubbling (120 min)

LiLS after air bubbling (120 min) followed by nitrogen purge (60 min)

Figure 7. The relative absorbance spectra of LiLS samples before and after exposure to air, and subsequentnitrogen sparging demonstrating the oxygen quenching effect in the LiLS. The gas flow rates for air andnitrogen were set at 30 mL/min.

– 7 –

SN021401_spec0Entries 925164Mean 242.2RMS 210.1

Channel number0 100 200 300 400 500 600 700 800 900 1000

Cou

nts

0

200

400

600

800

1000

1200

1400 SN021401_spec0Entries 925164Mean 242.2RMS 210.1

Original Cs-137 spectrum

Smoothed spectrum

Figure 8. The measured 137Cs spectrum by the LS6500 for the LAB reference with the original spectrum(in black) and the smoothed distribution used for determining the Compton edge (in red).

3.2 Relative light yield

The light yield of a LiLS sample from each 90 L batch relative to a LAB sample was measured withan automatic liquid scintillator counter (Beckman LS6500). The counter is equipped with a 137Csgamma source, two photomultiplier tubes in coincidence mode, and a multichannel analyzer.

Liquid scintillator samples of about 10.6 g each were filled in borosilicate glass vials withwhite polyethylene caps with PTFE inner liners (vial outer diameter 28.6 mm and height 61 mm).The LiLS samples were sparged with nitrogen at ∼30 mL/min for about 30 minutes and then closedfor measurement. No effort was made to remove air from the closed vials, and the light yieldsubsequently diminished due to oxygen quenching [8, 9].

The counter enabled automatic cyclical measurements of the reference and LiLS samples overan extended period of typically 70 hours. Each sample was measured for 15 minutes in each cycle.Three quantities were extracted from the data for each LiLS sample for the QA: the initial lightyield relative to the LAB reference and the fast and slow time constants. The initial relative lightyield is the main quantity used for LiLS acceptance. The fast time constant is attributed to oxygenquenching and the slow time constant indicates the LiLS performance stability.

There are artifacts in the observed spectrum that arise from the LS6500 counter [10], as canbe seen in Figure 8. The original spectrum is smoothed by applying the Gauss-Hermite quadratureand the smoothed spectrum is used to extract the Compton edge as a measure of the light yieldfollowing the differentiation method described in [11].

The identified Compton edge of the reference LAB sample used for the quality assurancemeasurements as a function of time is shown in Figure 9. The variation in the reference samplelight yield is consistent with the long term behavior of the LS6500 observed previously [10].

Each LiLS sample’s light yield was assessed using ∼70 hours data. The relative light yield is

– 8 –

03/01/17 05/01/17 07/01/17 08/31/17 10/31/17 12/31/17

Com

pton

edg

e (C

hann

el n

umbe

r)

630

640

650

660

670

680ComptonEdge_LAB

Entries 2432Mean 657.5RMS 9.323

Compton edge (Channel number)630 640 650 660 670 680 690

Co

un

ts

0

20

40

60

80

100

120

140

160

180 ComptonEdge_LAB

Entries 2432Mean 657.5RMS 9.323

Figure 9. The measured LAB Compton edge over a period of about 10 months (left) and its distribution(right). The error bars on the left are statistical.

Time (min)0 500 1000 1500 2000 2500 3000 3500 4000

Rela

tive lig

ht yie

ld (

LiL

S / L

AB

)

1.04

1.06

1.08

1.1

1.12

1.14

1.16

1.18

1.2

B2-S1

B21-S1

B21-S1r

B25-S1

B25-S2

B53-S1

Days

1−10 1 10 210

Rela

tive li

ght yi

eld

(LiL

S / L

AB

)

1.06

1.08

1.1

1.12

1.14

1.16 / ndf 2χ 150.8 / 68

A1 0.002093± 0.05045

1 τ 0.03621± 0.4731

A2 0.0006841± 1.1

2 τ 73.22± 2119

/ ndf 2χ 150.8 / 68

A1 0.002093± 0.05045

1 τ 0.03621± 0.4731

A2 0.0006841± 1.1

2 τ 73.22± 2119

Figure 10. Left: Relative light yield of a few representative LiLS samples during the ∼70 hours ofmeasurements. “B” and “S” denote batch and sample number respectively. Occasionally multiple samplesfrom the same batch were measured (B25-S1 and B25-S2) and the same sample was re-measured (B21-S1r).Each sample’s light yield is fitted with two exponentials (dashed lines) as described in the text. Right: Therelative light yield of the sample B18-S1 over 100 days. The fitted parameters are shown on the plot.

defined by the ratio of Compton edges between the LiLS sample and LAB reference. Figure 10shows the relative light yield of representative LiLS samples in the first 70 hours. All other sampleshave a light yield between the extreme curves shown in Figure 10. The relative light yield decreasesrapidly in the first ∼500 minutes and then undergoes a much slower decrease. Consequently, we fitthis time-dependence with a double exponential function

f (t) = A1e−t/τ1 + A2e−t/τ2 , (3.1)

in which A1,2 and τ1,2 are the corresponding relative light yields and time constants for the fast andslow components, respectively. The sum A1 + A2 describes the initial relative light yield of an LiLSsample and is the main quantity that is used in the QA process. Some samples were monitored fora few months and the fitted τ2 is at least four years. As an example, the fitted measurements for thebatch 18 sample are shown in Figure 10.

– 9 –

P50

-1B

1-S

1B

1-S

2B

2-S

1B

3-S

1B

4-S

1B

5-S

1B

6-S

1B

7-S

1B

8-S

1B

9-S

1B

10-S

1B

12-S

1B

13-S

1B

14-S

1B

15-S

1B

16-S

1B

17-S

1B

18-S

1B

19-S

1B

20-S

1B

21-S

1B

21-S

1rB

22-S

1B

23-S

1B

24-S

1B

25-S

1B

25-S

2D

12-S

1B

26-S

1B

27-S

1B

28-S

1B

29-S

1B

30-S

1B

31-S

1B

32-S

1B

33-S

1B

34-S

1B

35-S

1B

36-S

1B

37-S

1B

38-S

1B

39-S

1B

40-S

1B

41-S

1B

42-S

1B

43-S

1B

44-S

1B

45-S

1B

47-S

1B

48-S

1B

49-S

1B

50-S

1B

51-S

1B

52-S

1B

53-S

1B

54-S

1B

55-S

1B

56-S

1B

57-S

1B

58-S

1B

59-S

1

Initi

al r

elat

ive

light

yie

ld (

LiLS

/LA

B)

1.08

1.09

1.1

1.11

1.12

1.13

1.14

1.15

1.16

1.17

1.18

1.19

1.2

1.21

1.22

Mean 1.169RMS 0.0152

Initial relative light yield (LiLS/LAB)1.08 1.1 1.12 1.14 1.16 1.18 1.2 1.22

Cou

nts

0

1

2

3

4

5

6

7

8

9 Mean 1.169RMS 0.0152

PROSPECT LiLS

PROSPECT-50

Acceptance criterion

Figure 11. Initial relative light yield of each LiLS sample (top) and its distribution (bottom) from theCompton edge from a 137Cs source as described in the text. Batch 25 was measured twice (B25-S1 andB25-S2). The sample for batch 21 was remeasured (B21-S1r).

Figure 11 shows the initial relative light yield for all the LiLS samples as well as a sample fromPROSPECT-50 [5]. The PROSPECT-50 sample was used to set the light yield selection criterion:the initial relative light yield of LiLS was required to be greater than 95% of the PROSPECT-50sample; that is, A1+ A2 > 1.085 since the measured initial light yield for the PROSPECT-50 samplewas 1.142. As can be seen from Figure 11, all the measured samples show a satisfactory initiallight yield well above the required threshold.

3.3 Pulse shape discrimination

The same LiLS samples were used in the PSDmeasurements right after the light yield measurementwithout additional nitrogen sparging. The PSD capability of the LiLS samples was measured using

– 10 –

an 241Am-9Be neutron source (105 MBq) from which the rate of nuclear recoils in a ∼10 g samplelargely exceeded the thermal neutron capture rate. The PSD for nuclear recoils as a function ofvisible energy is measured and the result is used to estimate the PSD for thermal neutron capturesaccording to the measured response of the LiLS in reference [5].

Each sample was placed in a reflective PTFE collar on the face of a 1.5-inch (38.1 mm)Hamamatsu PMT (R9420-100) operated at about 5 × 105 gain (Figure 12). A Tektronix 3450oscilloscope was used to record 20000 waveforms at 5 GHz sampling rate, with a 2 µs window andan 80-mV threshold. Another 5000 waveforms were acquired with an additional 137Cs source (0.68MBq) to provide energy calibration. Figure 13 shows the measured 137Cs spectrum for the samplefrom batch 2.

Figure 12. Side view of the setup for the PSD measurements. The PMT is vertically mounted and facingup, the vial with LiLS rests on top of the PMT and it is surrounded by a PTFE cylinder to enhance lightcollection. The Am-Be neutron source shown mounted at the right side of the PTFE cylinder.

The total charge, qtot , measured by a waveform is computed from an interval of [-10, 160] nsabout the minimum of the negative-going signal waveform. The Compton edge of the 137Cs data isfitted to estimate the energy scale. The PSD is defined as

PSD = (qtot − qf )/qtot, (3.2)

where qf is the charge in the [-10, 20] ns interval about the waveform minimum. The PSDdistribution in intervals of charge is fitted with two Gaussians to evaluate the PSD for electronicand nuclear recoils (Figure 14). The figure-of-merit (FOM) is defined as

FOM = (µn − µe)/√

FWHM2n + FWHM2

e, (3.3)

where µn(e) and FWHMn(e) are the mean and full width at half maximum of the Gaussianscorresponding to nuclear (electronic) recoils.

– 11 –

Entries 8314

/ ndf 2χ 57.86 / 24

p0 0.28± 67.92

p1 0.361± 8.613

p2 3.6± 189.1

Charge (pC)0 20 40 60 80 100 120 140 160

Cou

nts

0

50

100

150

200

250

300

350

400Entries 8314

/ ndf 2χ 57.86 / 24

p0 0.28± 67.92

p1 0.361± 8.613

p2 3.6± 189.1

Figure 13. 137Cs spectrum of the LiLS sample from batch 2. The red curve is fitted with a complementaryerror function. The fitted Compton edge in this example is at 67.92 pC (p0).

Charge (pC)0 50 100 150 200 250 300 350 400

PS

D

0.1−

0

0.1

0.2

0.3

0.4

0.5 hPSDvsQtot_B2S1

Mean x 127.7Mean y 0.1843RMS x 100.2RMS y 0.06452

0

5

10

15

20

25

30

35hPSDvsQtot_B2S1

Mean x 127.7Mean y 0.1843RMS x 100.2RMS y 0.06452

hPSD_Qtot_60_100pCEntries 3742

/ ndf 2χ 58.14 / 60const1 6.1± 186.1 mean1 0.0005± 0.1255 sigma1 0.0004± 0.0185 const2 6.1± 237.8 mean2 0.0005± 0.2474 sigma2 0.00036± 0.02319

PSD0 0.1 0.2 0.3 0.4 0.5

Cou

nts

0

50

100

150

200

250hPSD_Qtot_60_100pC

Entries 3742 / ndf 2χ 58.14 / 60

const1 6.1± 186.1 mean1 0.0005± 0.1255 sigma1 0.0004± 0.0185 const2 6.1± 237.8 mean2 0.0005± 0.2474 sigma2 0.00036± 0.02319

Figure 14. Left: PSD versus total charge for the LiLS sample from batch 2. The upward trend at high chargeis due to signal saturation during measurement. Right: The fitted PSD distribution for total charge between60 and 100 pC. For this sample a total charge of 77 pC corresponds to 540 keV visible energy.

The measured FOM as a function of energy is linearly interpolated to estimate the FOM ofthe neutron capture process 6Li(n, α)3H at ∼540 keV electron equivalent which is used to assessthe LiLS PSD quality. The higher the FOM value, the better the separation between nuclear andelectronic recoils. The FOM of all measured samples exceeds that of the PROSPECT-50 samples asshown in Figure 15, consequently the samples are all accepted. The increase in FOM for batch ≥12is attributed to the better quality of EJ-309 scintillator from the later drums as described earlier.

– 12 –

P50

-m1

P50

-m2

P50

-m3

P50

-m4

P50

-m5

B1-

S1

B1-

S2

B2-

S1

B3-

S1

B4-

S1

B5-

S1

B6-

S1

B7-

S1

B8-

S1

B9-

S1

B10

-S1

B12

-S1

B13

-S1

B14

-S1

B15

-S1

B16

-S1

B17

-S1

B18

-S1

B19

-S1

B20

-S1

B21

-S1

B22

-S1

B23

-S1

B24

-S1

B25

-S1

B25

-S1r

B25

-S2

D12

-S1

B26

-S1

B27

-S1

B28

-S1

B29

-S1

B30

-S1

B31

-S1

B32

-S1

B33

-S1

B34

-S1

B35

-S1

B36

-S1

B37

-S1

B38

-S1

B39

-S1

B40

-S1

B41

-S1

B42

-S1

B43

-S1

B44

-S1

B45

-S1

B47

-S1

B48

-S1

B49

-S1

B50

-S1

B51

-S1

B52

-S1

B53

-S1

B54

-S1

B55

-S1

B56

-S1

B57

-S1

B58

-S1

B59

-S1

FO

M

1.4

1.5

1.6

1.7

1.8

1.9

2

2.1

2.2

2.3

Mean 2.037

RMS 0.1576

FOM1.4 1.5 1.6 1.7 1.8 1.9 2 2.1 2.2

Cou

nts

0

2

4

6

8

10

12

14 Mean 2.037

RMS 0.1576LiLS samples

PROSPECT-50 samples

Figure 15. The measured FOM (n, 6Li) for the LiLS samples (top) and the distribution of the FOM values(bottom). The results for the PROSPECT-50 samples are also presented.

4 Summary

A total of fifty-nine batches of 90 liter LiLS were produced for the PROSPECT experiment. Aone liter sample was collected from each batch for QA measurements. Two batches were rejecteddue to unsatisfactory absorbance, another two batches were used for prototyping and materialcompatibility tests, the remaining batches satisfied the acceptance criteria in absorbance, light yieldand PSD capabilities and were delivered for deployment in the PROSPECT detector at ORNL.

Acknowledgments

This material is based upon work supported by the following sources: U.S. Department of Energy(DOE) Office of Science, Office of High Energy Physics under Award No. DE-SC0016357 andDE-SC0017660 to Yale University, under Award No. DE-SC0017815 to Drexel University, underAward No. DE-SC0008347 to Illinois Institute of Technology, under Award No. DE-SC0016060to Temple University, under Contract No. DE-SC0012704 to Brookhaven National Laboratory,and under Work Proposal Number SCW1504 to Lawrence Livermore National Laboratory. This

– 13 –

work was performed under the auspices of the U.S. Department of Energy by Lawrence LivermoreNational Laboratory under Contract DE-AC52-07NA27344 and by Oak Ridge National Laboratoryunder Contract DE-AC05-00OR22725. This work was also supported by the Natural Sciences andEngineeringResearchCouncil of Canada (NSERC)Discovery programunder grant #RGPIN418579and Province of Ontario.

Additional funding was provided by the Heising-Simons Foundation under Award No. #2016-117 to Yale University. J.G. is supported through the NSF Graduate Research Fellowship Programand A.C. performed work under appointment to the Nuclear Nonproliferation International Safe-guards Fellowship Program sponsored by the National Nuclear Security Administration’s Office ofInternational Nuclear Safeguards (NA-241).

References

[1] J. Ashenfelter et al., (PROSPECT collaboration), The PROSPECT physics program, J. Phys. G 43(2016) 113001.

[2] J. Ashenfelter et al., (PROSPECT collaboration), The PROSPECT Reactor Antineutrino Experiment,Nucl. Instrum. Meth. A 922 (2019) 287-309.

[3] C. D. Bass et al., Characterization of a 6Li-loaded liquid organic scintillator for fast neutronspectrometry and thermal neutron detection, App. Rad. Isotop. 77 (2013) 130-138.

[4] B. M. Fisher et al., Fast neutron detection with 6Li-loaded liquid scintillator, Nucl. Instrum. Meth. A646 (2011) 126-134.

[5] J. Ashenfelter et al., (PROSPECT collaboration), Performance of a segmented 6Li-loaded liquidscintillator detector for the PROSPECT experiment, 2018 JINST 13 P06023.

[6] M. Yeh et al., A new water-based liquid scintillator and potential applications, Nucl. Instrum. Meth.A 660 (2011) 51-56.

[7] The synthesis team inadvertently produced LiLS with 6Li at 0.1% per volume instead of per mass, asintended.

[8] Xiao Hua-Lin et al., Oxygen quenching in a LAB based liquid scintillator and the nitrogen bubblingmodel, 2010 Chinese Phys. C 34 571.

[9] F. Verrezen, H. Loots, C. Hurtgen, A performance comparison of nine selected liquid scintillationcocktails, Appl. Rad. Isotop. 66 (2008) p. 1038-1042.

[10] L.J. Bignell et al., Measurement of radiation damage of water-based liquid scintillator and liquidscintillator, 2015 JINST 10 P10027.

[11] M. J. Safari et al., Differentiation method for localization of Compton edge in organic scintillationdetectors, [arXiv:1610.09185].

– 14 –