Copper electroplating for background suppression in the NEWS-G experiment
NEWS-G Collaboration, L. Balogh, C. Beaufort, A. Brossard, R. Bunker, J.-F. Caron, M. Chapellier, J.-M. Coquillat, E.C. Corcoran, S. Crawford, A. Dastgheibi Fard, Y. Deng, K. Dering, D. Durnford, G. Gerbier, I. Giomataris, G. Giroux, P. Gorel, M. Gros, P. Gros, O. Guillaudin, E. W. Hoppe, I. Katsioulas, F. Kelly, P. Knights, L. Kwon, S. Langrock, P. Lautridou, R. D. Martin, J.-P. Mols, J.-F. Muraz, X.-F. Navick, T. Neep, K. Nikolopoulos, P. O'Brien, R. Owen, M.-C. Piro, D. Santos, G. Savvidis, I. Savvidis, F. Vazquez de Sola Fernandez, M. Vidal, R. Ward, M. Zampaolo, S. Alcantar Anguiano, I. J. Arnquist, M.L. di Vacri, K. Harouaka, K. Kobayashi, K.S. Thommasson
CCopper electroplating for background suppressionin the NEWS-G experiment
L. Balogh a , C. Beaufort b , R. Bunker c , A. Brossard a , J.-F. Caron a , M. Chapellier a , E.C. Corcoran d ,J.-M. Coquillat a , S. Crawford a , A. Dastgheibi Fard b , Y. Deng e , K. Dering a , D. Dunford e , I. Giomataris f ,P. Gorel g,h,i , M. Gros f , P. Gros a , G. Gerbier a , G. Giroux a , O. Guillaudin b , E.W. Hoppe c , I. Katsioulas j ,F. Kelly d , P. Knights f,j, ∗ , L. Kwon d , S. Langrock h , P. Lautridou k , J.-P. Mols f , R. D. Martin a ,J.-F. Muraz b , X.-F. Navick f , T. Neep j , K. Nikolopoulos j , P. O’Brien e , R. Owen j , M.-C. Piro e , D. Santos b ,G. Savvidis a , I. Savvidis l , F. Vazquez de Sola Fernandez a , M. Vidal a , R. Ward j , M. Zampaolo b (NEWS-G Collaboration) S. Alcantar Anguiano c , I. J. Arnquist c , M.L. di Vacri c , K. Harouaka c , K. Kobayashi m,n , K.S. Thommasson c a Department of Physics, Engineering Physics & Astronomy, Queen’s University, Kingston, Ontario K7L 3N6, Canada b LPSC, Universit´e Grenoble-Alpes, CNRS/IN2P3, Grenoble, France c Pacific Northwest National Laboratory, Richland, Washington 99352, USA d Chemistry & Chemical Engineering Department, Royal Military College of Canada, Kingston, Ontario K7K 7B4, Canada e Department of Physics, University of Alberta, Edmonton, Alberta, T6G 2R3, Canada f IRFU, CEA, Universit´e Paris-Saclay, F-91191 Gif-sur-Yvette, France g Department of Physics and Astronomy, Laurentian University, Sudbury, Ontario, P3E 2C6, Canada h SNOLAB, Lively, Ontario, P3Y 1N2, Canada i Arthur B. McDonald Canadian Astroparticle Physics Research Institute, Queen’s University, Kingston, ON, K7L 3N6,Canada j School of Physics and Astronomy, University of Birmingham, Birmingham B15 2TT United Kingdom k SUBATECH, IMT-Atlantique, Universit´e de Nantes/IN2P3-CNRS, Nantes, France l Aristotle University of Thessaloniki, Thessaloniki, Greece m Kamioka Observatory, ICRR, University of Tokyo, Higashi-Mozumi, Kamioka, Hida, Gifu 506-1205, Japan n Kavli Institute for the Physics and Mathematics of the Universe, University of Tokyo, Kashiwa, Chiba 277-8582, Japan
Abstract
New Experiments with Spheres-Gas (NEWS-G) is a dark matter direct detection experiment that willoperate at SNOLAB (Canada). Similar to other rare-event searches, the materials used in the detectorconstruction are subject to stringent radiopurity requirements. The detector features a 140-cm diameterproportional counter comprised of two hemispheres made from commercially sourced 99.99% pure copper.Such copper is widely used in rare-event searches because it is readily available, there are no long-lived Curadioisotopes, and levels of non-Cu radiocontaminants are generally low. However, measurements performedwith a dedicated
Po alpha counting method using an XIA detector confirmed a problematic concentrationof
Pb in bulk of the copper. To shield the proportional counter’s active volume, a low-background elec-troforming method was adapted to the hemispherical shape to grow a 500- µ m thick layer of ultra-radiopurecopper to the detector’s inner surface. In this paper the process is described, which was prototyped at PacificNorthwest National Laboratory (PNNL), USA, and then conducted at full scale in the Laboratoire Souter-rain de Modane in France. The radiopurity of the electroplated copper was assessed through inductivelycoupled plasma mass spectrometry (ICP-MS). Measurements of samples from the first (second) hemispheregive 68% confidence upper limits of < . µ Bq / kg ( < . µ Bq / kg) and < . µ Bq / kg ( < . µ Bq / kg) onthe Th and
U contamination levels, respectively. These results are comparable to previously reportedmeasurements of electroformed copper produced for other rare-event searches, which were also found to havelow concentration of
Pb consistent with the background goals of the NEWS-G experiment.
Keywords:
Dark matter, Direct detection, Rare Event, Electroforming, Copper, Low background
Preprint submitted to Nucl. Instrum. Methods Phys. Res. A August 10, 2020 a r X i v : . [ phy s i c s . i n s - d e t ] A ug igure 1: Schematic diagram of the NEWS-G detector andshielding. The ∅
140 cm spherical proportional counter isshown at the centre, surrounded by 3 cm of archaelogicallead, followed by 22 cm of low radioactivity lead in a stainlesssteel skin. The outer-most part of the shielding comprises40 cm of high-density polyethylene (HDPE).
1. Introduction
Direct searches for dark matter (DM) and neu-trinoless double-beta decay [1, 2, 3, 4] have strictrequirements on the experimental background toachieve their targeted sensitivities. While such ex-periments are generally carried out in undergroundlaboratories and in specifically designed shieldingto suppress backgrounds from external sources, oneof the main remaining sources arises from radioac-tive decays in the detector’s construction materials,including the gaseous target. The effort to procurematerials with the lowest possible radioactivity hasdriven significant improvements in the techniquesand facilities used to assay and prepare radiopurematerials [5, 6, 7, 8].A common choice for a high-purity material iscommercially sourced copper [9, 10, 11], because itis readily available and there are no long-lived Curadioisotopes—with a half-life of 61 . Cu is the longest-lived. For this reason, theNEWS-G collaboration [13] chose 4N (99 .
99% pure)copper to construct a ∅
140 cm spherical propor-tional counter [14], which will be housed in the com-pact shielding shown in Fig. 1, to perform a directDM search at SNOLAB, Canada. Even withoutlong-lived Cu radioisotopes, a copper sample will ∗ Corresponding author
Email address: [email protected] (P. Knights) Now at Waseda Research Institute for Science and Engi-neering, Waseda University, 3-4-1 Okubo, Shinjuku, Tokyo169-8555, Japan Procured from Aurubis AG, Hovestrasse 50, 20539 Ham-burg, Germany
Po21884Pb21482Rn22286 Bi21483 Po21484Pb21082 Bi21083 Po21084Pb20682
U23892
Th23490
24 d
Pa234m91
U23492
250 ky
Pa91
Th23090
24 d
Ra22688
Figure 2:
U decay chain. All daughters are solid at roomtemperature and pressure except
Rn, which is a gas. Onlydecays with a branching fraction greater than 0 .
05% areshown [12]. have some (non-Cu) radiogenic contamination re-sulting from cosmogenic activation and industrialproduction processes. For example, cosmic-ray neu-trons interacting with copper through the ( n , α ) re-action can produce Co. The half-life of the pro-duced Co is approximately 5.3 years, making it along-lived background relative to the typical timescale of direct DM detection experiments. At thesurface of the Earth, the added activity due to Cois approximately 0 . µ Bq / kg / day [15]. Other cos-mogenic contaminants with shorter half-lives arealso produced, e.g. Fe. These contributions canbe suppressed by minimising the copper’s exposureto cosmic rays. Other radiocontaminants primar-ily originate from the
U and
Th decay chains.The
U decay chain is shown in Fig. 2. This con-tamination is inherent to the raw material and aresult of the manufacturing and handling processes.An established technique is to directly measure theuranium and thorium levels with inductively cou-pled plasma mass spectrometry (ICP-MS), whichhas been demonstrated to have sensitivity betterthan 30 fg / g to these contaminants [6, 16, 17]. Theprogeny activities can also be inferred and used toestimate background contributions to experiments,under the assumption of secular equilibrium.However, Rn, which is part of the
U de-2ay chain, is a gaseous isotope. As a result,
Rnmay deposit its decay products on the copper sur-face or into the copper bulk at the raw-ore stageor during manufacturing. This contribution addsto the contamination and may break the secular-equilibrium assumption. The longest-lived isotopein the
Rn decay chain is
Pb with a half-life of22 . Pb from
Rndeposits can result in experiment backgrounds thatcannot be inferred by ICP-MS measurements of the
U progenitor. One method to assess this contam-ination is by directly measuring the 5 . α par-ticles from the Po decays [19, 20, 8], using a high-sensitivity XIA UltraLo-1800 spectrometer, whichhas a sensitivity of 0 . α/ cm / hour [21]. TheXMASS collaboration has established a method toestimate very low Pb contamination in copperbulk, having demonstrated the ability to distin-guishing the contamination in bulk from that onthe surface [20]. For oxygen-free copper (at least99 .
96% pure by weight ) the Pb contamination isestimated to be in the range of 17–40 mBq / kg [20].The estimation of Pb contamination in the 4Ncopper procured by the NEWS-G collaboration isdiscussed in Section 2.This level of
Pb in the bulk of the NEWS-Gdetector’s copper sphere would dominate the ex-pected background and define the experimental sen-sitivity. An approach to suppress the backgroundfrom
Pb contamination is to grow a layer ofultra-radiopure copper onto the inner surface ofthe detector sphere. This layer acts as an internalshield to suppress backgrounds, e.g. from β -decaysof Pb and accompanying X-rays and Auger elec-trons, and its progeny
Bi, originating from thebulk of the commercially sourced 4N copper. Itis estimated by means of a Geant4 [22] simula-tion that a 500 µ m-thick layer of ultra-radiopurecopper will suppress the background in the en-ergy region of interest by approximately 62% to1 . ± .
38 count / keV / kg / day [15].A method to deposit ultra-radiopure copper ispotentiostatic electroforming [20, 23]. This methodtakes advantage of electrochemical properties toproduce copper with reduced impurities. The pro-cess is discribed in Section 3. This method waspreviously used to produce a variety of detectorcomponents, including those requiring extreme ra-diopurity such as for the Majorana Demonstra- Japanese Industrial Standard, JIS:C1020 tor [24]. Internal fittings were fabricated from elec-troformed copper with
U and
Th levels lessthan 0.099 and 0.119 µ Bq / kg at 68% confidence,respectively—limited by the ICP-MS assay preci-sion [17]. In order to apply this process to a hemi-spherical surface, a scale model was produced andused to determine the operating conditions. Thisis described in Section 4. The electroplating proce-dure used on the NEWS-G detector and the resultsof a subsequent radioisotope assay of the producedcopper are detailed in Section 5 and Section 6, re-spectively.
2. Assessment of the
Pb Contaminationin NEWS-G Copper
To assess the level of
Pb contamination in theC10100 copper used to produce the detector, sam-ples were taken from the same batch of copper aftercasting. The α particles from Po decays weremeasured using an XIA UltraLo-1800 [21] ionisa-tion chamber, which uses an active veto to obtaina second complementary signal arising from caseswhere the α particle does not originate from thesample under test. This is used to suppress back-ground coming from the detector’s own construc-tion materials. The sample is placed in the detec-tor which is flushed with argon gas to minimise Rn contamination. In this measurement, the
Po content of the bulk of the copper sample is ofinterest. The observable energy of 5 .
30 MeV α par-ticles emerging from the bulk of the copper samplewas estimated with a Geant4 simulation. An en-ergy window of 2 . . α particle originating from a depthof approximately 2 µ m to 8 µ m. This improvesthe signal-to-noise ratio for selecting bulk α par-ticle events. The conversion factor for measuredcounts to bulk activity was estimated from Geant4to be 2 . × (Bq / kg) / ( α/ cm / hour) [20]. Po has a half-life of approximately 138 days,which is significant shorter than the approximately22 years of the progenitor
Pb. As a result, theactivities of
Po and
Pb may be different dueto different contamination amounts at the produc-tion phase. Therefore, the activities of the two iso-topes may be out of secular equilibrium; however,the
Po activity in a sample will evolve over time For
U, 1 µ Bq / kg ≈ .
081 pg / g. For Th,1 µ Bq / kg ≈ .
244 pg / g Pb. Therefore, multiplemeasurements of the
Po activity over time arerequired to accurately infer the activity of
Pbin the copper. Four measurements of the α par-ticles from the sample were made over the courseof approximately one year, each lasting between 12and 23 days. Table 1 shows the results of the fourmeasurements. Table 1: Measurements of the α particles in a 2 . . Po decays in aC10100 copper sample.
Date Measurement[10 − α/ cm / hour]Jul. 2 - 25, 2018 2 . ± . . ± . . ± . . ± .
3A joint likelihood fit of all measurements was per-formed and is shown in Figure 3 along with the mea-surements. From this fit, it was estimated that the
Pb activity in the sample is 29 +8+9 − − mBq kg − ,where the statistical and systematic uncertaintiesare given seperately. This is consistent with othercopper samples with similar purity [20]. Time [years] B u l k A c t i v i t y [ m B q / k g ] Measured
PoModelled
PoModelled Pb Figure 3: Measurements of the α particles from the decay of Po in a sample of C10100 copper used in the production ofthe NEWS-G detector. Time is measured from the estimatedproduction date of the copper. The purple (green) line showsthe fitted
Po (
Pb) activity over time, with the bandsshowing the ± σ region.
3. Electroplating
Electroplating is carried out through the use ofan electrolytic cell, which consists of an anode anda cathode separated by an electrolyte, as illustrated in Fig. 4. A current is used to supply electrons to
ElectrolyteAnode CathodeReductionOxidation
Figure 4: Schematic diagram of a simple electrolytic cell.Arrows indicate the motion of ions, which are released intothe electrolyte by oxidation reactions at the anode and thendeposited on the cathode in reduction reactions. the cathode where an ion undergoes a reduction re-action (gain of electrons) to form an atom depositedon the surface, while oxidation reactions (loss ofelectrons) occur at the anode. The reactions occur-ring at each of the electrodes will be of the generalform: A ( y + z )+ + ze − (cid:11) A y + , (1)where A is the molecular species, y is its ioniccharge and z is the number of electrons requiredfor the reduction reaction (reading left-to-right) orthe number of electrons released in the oxidationreaction (reading right-to-left). Reading this equa-tion in one direction gives the “half-cell reaction”,where the anode and cathode half-cell reactions arenot necessarily the same; e.g., in the case whereone species is oxidized at the cathode but a differ-ent species is reduced at the anode.A current I flows through the circuit and elec-trolyte. As the reduction reactions require elec-trons, the rate of electroplating is proportional tothe total supplied charge: Q ( t ) = (cid:82) I d t . The num-ber of moles n of ions reduced in time t is given by n ( t ) = Q ( t ) zF , (2)where F = eN A is the Faraday constant, and e and N A are the elementary charge and the Avo-gadro constant, respectively. The resulting de-posited mass as a function of time is M ( t ) = m r n ( t ) , (3)4here m r is the molecular mass of the depositedspecies. When the current is reversed the processis called electropolishing, which is a technique usedto remove material from a surface.There will be several species of ions in the elec-trolyte available to electroplate to the cathode. Thetendency of an ion species to be reduced is quan-tified by the reduction potential E . Examples areshown in Table 2 for copper and radioisotope con-taminants. A greater value of E indicates a speciesthat is more easily reduced. Each half-cell reactionwill have its own reduction potential. The standardcell potential E of the electrolytic cell is definedas the difference between the reduction potentials ofthe half-cell reactions at the anode E A and cathode E C : E = E − E . (4)For E <
0, additional energy will be required forthe reaction to proceed [25]. For E ≥
0, the re-action is spontaneous (or in chemical equilibrium inthe case of equality). For a given species being ox-idized at the anode, the reaction will only proceedwhen the cathode half-cell reaction has a higher re-duction potential. In the case of a copper anode be-ing oxidized, only ion species in the electrolyte witha reduction potential greater than that of copperwill reduce. The relatively high reduction potentialof copper compared to many radioisotopes meansthat it is purified during electroplating. However,other factors, such as mass transport of contami-nant ions, can cause species with lower reductionpotentials to be deposited with the copper in smallamounts [23].
Table 2: Reduction potential for copper and possible radio-contaminants.
Reductants Oxidants E (V)Cu + 2 e − (cid:11) Cu +0 .
34 [26]Pb + 2 e − (cid:11) Pb − .
13 [27]U + 3 e − (cid:11) U − .
80 [28]Th + 4 e − (cid:11) Th − .
90 [28]K + + e − (cid:11) K − .
93 [29]In this work, a copper anode is used to provideCu ions to the electrolyte. For Cu ions reduc-ing at the cathode, the system will have E = 0 V.Thus, to drive the reaction and overcome energyloss mechanisms in the system, the electrodes arekept at a potential difference of 0 . relative to the bulk elec-trolyte. This slows down the rate of electroplat-ing and affects the properties of the deposited cop-per [31]. The waveform of the time-varying poten-tial difference allows this region to be replenished byallowing diffusion from the bulk electrolyte when novoltage is applied and by reintroducing more ionsfrom the surface during the reverse-voltage part ofthe waveform. Also, differences in current den-sity can arise due to differences in the distance be-tween the anode and cathode surfaces (e.g. a surfacerough point). High current density regions of theelectrolyte are more depleted of Cu than lowerdensity regions. When no voltage is applied, ionscan diffuse between two such regions and thus leadto a more uniform overall current density, whilethe reverse-voltage part of the waveform preventsa thick layer forming in the high current densityregions [31]; both effects promote more uniformgrowth of the electroplated copper layer. The re-versing of polarity also allows for release of contam-inant ions that may have been entrapped during thehigh mass transport portion of the forward plating.The waveform used for the electroplating is shownin Fig, 5. Note that while the potential is appliedit is potentiostatic at a level that favors the oxida-tion/reduction of copper. PotentialDifference
Figure 5: Waveform used in the electroplating. The negativeterminal was attached to the cathode.
4. Scale Model
The copper electroplating procedure describedin the previous section is a well-established andsuccessful method that has existed for over adecade [23]. However, this method must meet fairly5igid operational conditions to produce optimal ma-terial. Failure to meet these conditions can not onlyproduce copper of poor radiopurity but often resultsin deposits with poor physical properties as well.For NEWS-G the initial loading of copper into so-lution would need to be generated from an initialelectropolishing step, because commercially avail-able copper sulfate is not sufficiently pure. How-ever, for traditional electroforming, the amount ofcopper required in the electrolyte is too great toachieve through electropolishing. As a result, theplating conditions for the NEWS-G hemispheres re-quired a major deviation in the concentration ofcopper sulfate (CuSO ) in the electrolyte.Not all parameters have a well-studied effecton growth, especially when multiple parametersare outside of their established optimal operat-ing ranges. Prior experience has shown that elec-trolyte with a low copper-ion concentration can pro-duce dendritic copper deposition. In the absenceof accurate deposition models it was necessary torun a scaled experiment prior to plating the full-sized ∅
140 cm sphere underground in LaboratoireSouterrain de Modane (LSM). Key growth parame-ters were identified and an experiment was designedbased on those that could be adjusted in situ atLSM and projected onto a scale model.The key independent and adjustable variableswere determined to be the concentration of copperand overall conductivity of the electrolyte, and thecurrent based on the limiting set of voltage condi-tions. Control of the CuSO concentration is lim-ited by the amount of copper that can be dissolvedduring an initial electropolishing step, which servestwo purposes: a) expose the underlying bulk crystalstructure to prepare the copper surface for electro-plating; and b) load the electrolyte with copper.During this step, the ∅
140 cm hemisphere will actas the anode and careful control of the potentialis not as important, whereas subsequently copperwill be plated to the ∅
140 cm hemisphere whichwill then be serving as the cathode. During thelatter step, the voltage control and deposition rateare critical. As a result,establishing how the plat-ing responds to small changes in CuSO is crucial.As such, three variations of CuSO concentration,three conductivities, and three voltage settings wereidentified for experimentation on the scale model.A stainless steel spherical float with a diameterof 30 cm was cut in half and used as a stand-infor the full-scale ∅
140 cm copper hemisphere. Asmaller hemisphere was machined from aluminum (a) (b)(c) (d)Figure 6: (a) CAD model of the small-scale setup; (b) theassembled scale-model experiment; (c) copper plated ontothe scale model’s stainless steel hemisphere; and (d) the finalscale-model growth of copper. and plated with copper to serve as the anode afterthe initial electropolishing step. Figure 6 shows theexperimental setup of the scale model. Althoughthe transport dynamics involved are not fully un-derstood, previous experience has shown that theelectrode gap (path length) has an effect on plat-ing, regardless of CuSO concentration and con-ductivity. So, while the spacing between the twoelectrodes was scaled, the impedance needed to bematched to that of the full-scale setup. This re-quired the electrolyte conductivity to be reducedto compensate for the reduced electrode spacing inthe model.A bath, shown in Fig. 6(a) and Fig. 6(b), wasdesigned to hold and stabilize the stainless steelhemisphere, and several iterations of plating wereperformed to cycle through the plating variationsand determine the optimal electroplating condi-tions. Based on these trials, the parameters chosenfor plating copper onto the full-scale hemispheresare a CuSO concentration of 0.03 M, a conduc-tivity of 91.9 mS/cm (corresponding to a full-scaleconductivity of 300 mS/cm), and a potential of0.35 V. Using these parameters, the estimated timeto electroplate each full-scale hemisphere is ∼ a) (b)Figure 7: A detector hemisphere following (a) initial cleaningwith detergent and (b) chemical etching with an acidifiedhydrogen peroxide solution. to attain a thickness of 500 m. The resulting growthfor the small-scale model is shown in Fig. 6(c) andFig. 6(d).
5. Electroplating NEWS-G Detector
The electroplating was conducted at LSM at adepth of 4800 m water equivalent to reduce cosmo-genic activation. The detector outer shell is com-prised of two ∅
140 cm hemispheres, produced by aspinning technique using 4N copper. The result af-ter cleaning with commercial detergent is shown inFigure 7(a). The hemispheres were then sanded toproduce a smooth surface and subsequently chemi-cally etched using an acidified hydrogen peroxidesolution [32]. The result of this preparation isshown in Figure 7(b).A smaller 4N copper hemisphere was producedto act as the anode for electroplating and wascleaned in the same way as the detector hemi-spheres. It was suspended inside the detector,separated by an electrolyte comprised of deionizedwater (18 Mohm), Optima ® grade sulphuric acid(Fisher Scientific), and copper sulphate producedby a previous electroplating. A pump providedmechanical mixing with a filter removing partic-ulates greater than 1 µ m in size from the elec-trolyte. The anode and cathode were connected toa pulse-reverse power supply (Dynatronix, Amery,WI, USA), which could supply up to 80 A. Thewhole set-up was contained in a temporary purpose-constructed cleanroom to prevent particulates en-tering the electrolyte and subsequently providingnucleation sites for nodule-like copper growth [33].The setup is shown in Fig. 8.Prior to electroplating, each hemisphere was elec-tropolished to remove a layer of material from thesurface. This exposes the underlying crystalline structure and provides an ultraclean surface priorto deposition. Furthermore, this process enhancesthe amount of Cu in the electrolyte. A highervoltage was used for this process to extract allspecies from the surface. During electropolishing,(21 . ± . µ m and (28 . ± . µ m were removedfrom the first and second detector hemispheres,respectively. This was estimated from the inte-grated current and Eq. 3, assuming uniform pol-ishing. Following this process, the electrolyte cir-culated through the filter for several days prior toelectroplating to remove particulates released fromthe copper surface.The electroplating procedure used the reverse-pulse plating waveform shown in Fig. 4. The cur-rent and voltage were monitored throughout, andthe conductivity and temperature were recorded us-ing a HACH inductive conductivity sensor. Elec-troplating continued for a total of 19 . (a) Detector HemisphereInner HemisphereElectrolyteStainlessSteel Ring To Power Supply To Power SupplyTo Pump and Filter Conductivity Probe (b)Figure 8: (a) Electroplating setup showing the detec-tor hemisphere, anode, support structures, and fixtures.(b) Schematic diagram of the setup. . . ± . µ mand (539 . ± . µ m were plated onto the firstand second detector hemispheres, respectively. Theachieved plating rate corresponds to approximately1 . / year. A photo of the finished plating isshown in Fig. 10.After removing the hemispheres from the setup,they were rinsed with deionized water and the sur-face passivated with a 1% citric-acid solution to pre-vent surface oxidation [32]. P l a t ed T h i ck ne ss [ mm ] First HemisphereSecond Hemisphere
Figure 9: Estimated thickness of the electroplated copperfor both detector hemispheres.
6. Radioisotope Assay Results
Samples of the electroplated copper were used toassess its
U and
Th concentrations. Sampleswere taken from copper plated on the stainless steelring, shown in Fig. 8, to avoid damaging the detec-tor cladding. These samples originate from near theelectrolyte-air interface and from the stainless steelsurface; thus, they represent a worst case scenariowith respect to contamination. The samples col-lected from each hemisphere are shown in Fig. 11. (a) (b)Figure 10: (a) The inner surface of the second hemisphereafter electroplating and (b) a close-up of the surface.Figure 11: Samples of electroplated copper, taken from thestainless steel ring shown in Fig. 8.
The samples were shipped to Pacific NorthwestNational Laboratory and analysed using ICP-MSfollowing the methods described in Refs. [6, 16].The results are summarized in Table 3, along withrepresentative examples of electroformed and com-mercially sourced (machined) copper. A substantialimprovement over the latter is observed, with ra-diopurity levels comparable to previously measuredelectroformed copper. The measurement sensitivityfor the two hemispheres is limited by the mass ofthe available samples.
7. Summary
The NEWS-G collaboration has utilized recentadvances in high-purity copper electroforming toproduce a layer of copper on the inner surface of the ∅
140 cm detector. This layer will act as a shield tomitigate background from
Pb in the bulk of thedetector’s commercially sourced 4N copper. This isthe largest surface to be plated with ultra-radiopurecopper in an underground laboratory. This opera-tion has demonstrated the feasibility of plating ontothe surface of a large hemisphere. The radiopurity8 able 3: ICP-MS results for
U and
Th contamina-tion in samples of the electroplated copper layer, along withrepresentative examples of electroformed and commerciallysourced copper [17]. These are quoted as 68% upper confi-dence limits, where the measurement sensitivity was limitedby the available sample mass.
Weight Th USample [g] [ µ Bq / kg] [ µ Bq / kg]C10100 Cu - 8 . ± . . ± . < . < . < . < . < . < . . / year was achieved, which is promising forfabrication of a fully electroformed copper spherein the future. Acknowledgments
A portion of this work was funded by PNNL Lab-oratory Directed Research and Development fundsunder the Nuclear Physics, Particle Physics, As-trophysics, and Cosmology Initiative. The PacificNorthwest National Laboratory is a multi-programnational laboratory operated for the U.S. Depart-ment of Energy (DOE) by Battelle Memorial Insti-tute under contract number DE-AC05-76RL01830.This project has received funding from the Eu-ropean Union’s Horizon 2020 research and inno-vation programme under the Marie Sk(cid:32)lodowska-Curie grant agreement DarkSphere (grant agree-ment No 841261). Support has been receivedfrom the Royal Society International ExchangesScheme. This research was undertaken, in part,thanks to funding from the Canada Excellence Re-search Chairs Program, the Canada Foundationfor Innovation, the Arthur B. McDonald Cana-dian Astroparticle Physics Research Institute, andthe French National Research Agency (ANR-15-CE31-0008). The authors would like to thank theXMASS collaboration for the use of their XIA de-tector through the NEWS-G/XMASS collaborativeagreement.
References [1] R. Agnese, et al. (SuperCDMS Collaboration), Pro-jected Sensitivity of the SuperCDMS SNOLAB experi-ment, Phys. Rev. D 95 (2017) 082002.[2] M. Agostini, et al., Background-free search for neutri-noless double- β decay of Ge with GERDA, Nature544 (2017) 47–52.[3] P. Agnes, et al. (DarkSide Collaboration), DarkSide-50 532-day dark matter search with low-radioactivityargon, Phys. Rev. D 98 (2018) 102006.[4] C. Alduino, et al. (CUORE Collaboration), First Re-sults from CUORE: A Search for Lepton Number Vio-lation via 0 νββ
Decay of
Te, Phys. Rev. Lett. 120(2018) 132501.[5] P. Scovell, et al., Low-background gamma spectroscopyat the Boulby Underground Laboratory, Astropart.Phys. 97 (2018) 160–173.[6] B. LaFerriere, T. Maiti, I. Arnquist, E. Hoppe, A novelassay method for the trace determination of Th andU in copper and lead using inductively coupled plasmamass spectrometry, Nucl. Instrum. Methods Phys. Res.A 775 (2015) 93–98.[7] W. K. Warburton, B. Dwyer-McNally, M. Momayezi,J. E. Wahl, Ultra-low background alpha particlecounter using pulse shape analysis, in: IEEE Sym-posium Conference Record Nuclear Science 2004., vol-ume 1, 2004, pp. 577–581. doi: .[8] R. Bunker, et al., Evaluation and mitigation of trace
Pb contamination on copper surfaces, Nucl. Instrum.Meth. A 967 (2020) 163870.[9] C. Bucci, et al., First results from the Cuoricino ex-periment, Nucl. Instrum. Methods Phys. Res. A 520(2004) 132–134. Proceedings of the 10th InternationalWorkshop on Low Temperature Detectors.[10] E. Armengaud, et al., First results of the EDELWEISS-II WIMP search using Ge cryogenic detectors with in-terleaved electrodes, Phys. Lett. B 687 (2010) 294–298.[11] C. E. Aalseth, et al., DarkSide-20k: A 20 tonne two-phase LAr TPC for direct dark matter detection atLNGS, Eur. Phys. J Plus 133 (2018).[12] National Nuclear Data Center, information extractedfrom the Chart of Nuclides database, 2020. Accessed:10-02-2020.[13] Q. Arnaud, et al., First results from the NEWS-G directdark matter search experiment at the LSM, Astropart.Phys. 97 (2018) 54–62.[14] I. Giomataris, et al., A Novel large-volume Spheri-cal Detector with Proportional Amplification read-out,JINST 3 (2008) P09007.[15] A. Brossard, Optimization of spherical proportionalcounter backgrounds and response for low massdark matter search, Ph.D. thesis, Queen’s University,Kingston, Ontario, Canada and Universit´e Paris-Saclay,Saint-Aubin, France, 2020. http://hdl.handle.net/1974/27649 .[16] I. Arnquist, M. di Vacri, E. Hoppe, An automated ultra-cleanion exchange separation method for the determina-tions of
Th and
U in copper using inductively cou-pled plasma mass spectrometry, Nucl. Instrum. Meth-ods Phys. Res. A 965 (2020) 163761.[17] N. Abgrall, et al., The Majorana Demonstrator radioas-say program, Nucl. Instrum. Methods Phys. Res. A 828(2016) 22–36.
18] M. Shamsuzzoha Basunia, Nuclear Data Sheets for A= 210, Nuclear Data Sheets 121 (2014) 561–694.[19] G. Zuzel, K. Pelczar, M. W´ojcik, Studies of surface andbulk
Po in metals using an ultra-low backgroundlarge surface alpha spectrometer, Appl. Radiat. Isot.126 (2017) 165–167. Proceedings of the 7th Interna-tional Conference on Radionuclide Metrology – Low-Level Radioactivity Measurement Techniques.[20] K. Abe, et al., Identification of
Pb and
Po in thebulk of copper samples with a low-background alphaparticle counter, Nucl. Instrum. Methods Phys. Res. A884 (2018) 157–161.[21] XIA Ultra-Lo1800, 2020. URL: , Accessed: 02-03-2020.[22] S. Agostinelli, et al. (GEANT4), GEANT4: A Simula-tion toolkit, Nucl. Instrum. Meth. A 506 (2003) 250–303.[23] E. W. Hoppe, et al., Use of electrodeposition for sam-ple preparation and rejection rate prediction for assayof electroformed ultra high purity copper for
Th and
U prior to inductively coupled plasma mass spec-trometry (ICP/MS), J. Radioanal. Nucl. Chem. 277(2008) 103–110.[24] N. Abgrall, et al. (Majorana Collaboration), The Majo-rana Demonstrator neutrinoless double-beta decay ex-periment, Adv. High Energy Phys. 2014 (2014) 365432.[25] W. Greiner, L. Neise, H. St¨ocker, Thermodynamicsand statistical mechanics, Classical theoretical physics,Springer-Verlag, 1995. URL: https://books.google.co.uk/books?id=V6zvAAAAMAAJ .[26] A. Bard, R. Parsons, J. Jordan, Standard Potentialsin Aqueous Solution, Monographs in Electroanalyti-cal Chemistry and Electrochemistr, Taylor & Fran-cis, 1985. URL: https://books.google.co.uk/books?id=fuJV1H18KtEC .[27] P. Atkins, Physical Chemistry, 6th edition, W.H. Free-man and Company, New York, 1997.[28] D. Lide, CRC Handbook of Chemistry and Physics,87th Edition, Taylor & Francis, 2006. URL: https://books.google.co.uk/books?id=yTTUQgAACAAJ .[29] W. Haynes, CRC Handbook of Chemistry and Physics,92nd Edition, CRC Press, 2011. URL: https://books.google.co.uk/books?id=pYPRBQAAQBAJ .[30] E. W. Hoppe, et al., Microscopic evaluation of contam-inants in ultra-high purity copper, J. Radioanal. Nucl.Chem. 282 (2009) 315.[31] M. Chandrasekar, M. Pushpavanam, Pulse and pulsereverse plating - conceptual, advantages and applica-tions, Electrochim. Acta 53 (2008) 3313–3322.[32] E. Hoppe, et al., Cleaning and passivation of coppersurfaces to remove surface radioactivity and preventoxide formation, Nucl. Instrum. Methods Phys. Res.A 579 (2007) 486–489. Proceedings of the 11th Sympo-sium on Radiation Measurements and Applications.[33] N. Overman, et al., Majorana Electroformed CopperMechanical Analysis, Technical Report PNNL-21315,Pacific Northwest National Laboratory, 2012. doi: ..