Eur. Phys. J. C manuscript No. (will be inserted by the editor)

Sensitivity of the DARWIN observatory to the neutrinoless double beta decay of 136Xe

F. Agostini1, S. E. M. Ahmed Maouloud2, L. Althueser3, F. Amaro4, B. Antunovic5,†, E. Aprile6, L. Baudis7, D. Baur8, Y. Biondi7, A. Bismark7,8, P. A. Breur9,‡, A. Brown7, G. Bruno10, R. Budnik11, C. Capelli7, J. Cardoso4, D. Cichon12, M. Clark13, A. P. Colijn9,§, J. J. Cuenca-Garc´ıa14, J. P. Cussonneau15, M. P. Decowski9, A. Depoian13, J. Dierle8, P. Di Gangi1, A. Di Giovanni10, S. Diglio15, J. M. F. dos Santos4, G. Drexlin16, K. Eitel14, R. Engel14, A. D. Ferella17,18, H. Fischer8, M. Galloway7, F. Gao6, F. Girard7, F. Gl¨uck14, L. Grandi19, R. Gr¨oßle14, R. Gumbsheimer14, S. Hansmann-Menzemer20, F. J¨org12, G. Khundzakishvili13, A. Kopec13, F. Kugera,8, L. M. Krauss21, H. Landsman11, R. F. Lang13, S. Lindemann8, M. Lindner12, J. A. M. Lopes4,¶, A. Loya Villalpando9, C. Macolino22, A. Manfredini7, T. Marrod´anUndagoitia12, J. Masbou15, E. Masson22, P. Meinhardt8, S. Milutinovic5, A. Molinario23, C. M. B. Monteiro4, M. Murra3, U. G. Oberlack24, M. Pandurovic5, R. Peres7, J. Pienaar19, M. Pierre15, V. Pizzella12, J. Qin13, D. Ram´ırezGarc´ıa8, S. Reichard7, N. Rupp12, P. Sanchez-Lucasb,7, G. Sartorelli1, D. Schulte3, M. Schumann8, L. Scotto Lavina2, M. Selvi1, M. Silva4, H. Simgen12, M. Steidl14, A. Terliuk20, C. Therreau15, D. Thers15, K. Thieme7, R. Trotta25,k, C. D. Tunnell26, K. Valerius14, G. Volta7, D. Vorkapic5, C. Weinheimer3, C. Wittweg3, J. Wolf16, J. P. Zopounidis2, K. Zuber27 (DARWIN Collaborationc) 1Department of and Astronomy, University of Bologna and INFN-Bologna, 40126 Bologna, Italy 2LPNHE, Universit´ePierre et Marie Curie, Universit´eParis Diderot, CNRS/IN2P3, Paris 75252, France 3Institut f¨urKernphysik, Westf¨alische Wilhelms-Universit¨atM¨unster,48149 M¨unster,Germany 4LIBPhys, Department of Physics, University of Coimbra, 3004-516 Coimbra, Portugal 5Vinca Institute of Nuclear Science, University of Belgrade, Mihajla Petrovica Alasa 12-14. Belgrade, Serbia 6Physics Department, Columbia University, New York, NY 10027, USA 7Physik-Institut, University of Zurich, 8057 Zurich, Switzerland 8Physikalisches Institut, Universit¨atFreiburg, 79104 Freiburg, Germany 9Nikhef and the University of Amsterdam, Science Park, 1098XG Amsterdam, Netherlands 10New York University Abu Dhabi, Abu Dhabi, United Arab Emirates 11Department of and Astrophysics, Weizmann Institute of Science, Rehovot 7610001, Israel 12Max-Planck-Institut f¨urKernphysik, 69117 Heidelberg, Germany 13Department of Physics and Astronomy, Purdue University, West Lafayette, IN 47907, USA 14Institute for (IKP), Karlsruhe Institute of Technology (KIT), 76344 Eggenstein-Leopoldshafen, Germany 15SUBATECH, IMT Atlantique, CNRS/IN2P3, Universit´ede Nantes, Nantes 44307, France 16Institute of Experimental Particle Physics (ETP), Karlsruhe Institute of Technology (KIT), 76344 Eggenstein-Leopoldshafen, Germany 17Department of Physics and Chemistry, University of LAquila, 67100 LAquila, Italy 18 arXiv:2003.13407v2 [physics.ins-det] 7 Sep 2020 INFN-Laboratori Nazionali del Gran Sasso and Gran Sasso Science Institute, 67100 LAquila, Italy 19Department of Physics & Kavli Institute for Cosmological Physics, University of Chicago, Chicago, IL 60637, USA 20Physikalisches Institut, Ruprecht-Karls-Universit¨atHeidelberg, Heidelberg, Germany 21The Origins Project Foundation, Phoenix, AZ 85020, USA 22Universit´eParis-Saclay, CNRS/IN2P3, IJCLab, F-91405 Orsay, France 23INFN-Laboratori Nazionali del Gran Sasso and Gran Sasso Science Institute, 67100 L’Aquila, Italy 24Institut f¨urPhysik & Exzellenzcluster PRISMA+, Johannes Gutenberg-Universit¨atMainz, 55099 Mainz, Germany 25Department of Physics, Imperial Centre for Inference and Cosmology, Imperial College London, London SW7 2AZ, UK 26Department of Physics and Astronomy, Rice University, Houston, TX 77005, USA 27Institute for Nuclear and Particle Physics, TU Dresden, 01069 Dresden, Germany Received: date / Accepted: date 2

Abstract The DARWIN observatory is a proposed Experiments can observe a certain decay rate in a next-generation experiment to search for particle dark detector. The corresponding half-life is inversely pro- 2 matter and for the neutrinoless double beta decay of portional to hmββi , 136Xe. Out of its 50 t total natural inventory, 40 t will be the active target of a time projection chamber 1 hm i2 which thus contains about 3.6 t of 136Xe. Here, we show = ββ G0ν |M 0ν |2, (1) T 0ν m2 that its projected half-life sensitivity is 2.4 × 1027 yr, 1/2 e using a fiducial volume of 5 t of natural xenon and assuming that the decay is mediated by the exchange 10 yr of operation with a background rate of less than of a light Majorana . m is the mass of the 0.2 events/(t · yr) in the energy region of interest. This e , G0ν is the phase space factor, and M 0ν is the sensitivity is based on a detailed Monte Carlo simula- nuclear matrix element. Recent experimental limits on tion study of the background and event topologies in T 0ν and hm i are of the order T 0ν ≥ (1025-1026) yr the large, homogeneous target. DARWIN will be com- 1/2 ββ 1/2 and hm i ≤ (0.06 − 0.17) eV, using a variety of nuclei parable in its science reach to dedicated double beta ββ and detector technologies [2,3]. decay experiments using xenon enriched in 136Xe. A particularly suitable isotope to search for the 0νββ-decay with is 136Xe, with Qββ=(2457.83±0.37) keV [4]. Current experiments 1 Introduction use liquid xenon either in its pure form, EXO-200 [5], or xenon dissolved in liquid scintillator, KamLAND- are the only known elementary particles Zen [6], and provide competitive constraints on the that are Majorana fermion candidates, implying that half-life. Future detectors that use xenon operated they would be their own antiparticles. The most sen- at high pressure, NEXT [7,8] and PandaX-III [9], will sitive probe for the Majorana nature of neutrinos is add tracking capabilities for improved background an extremely rare nuclear decay process called neu- rejection, while nEXO [10] proposes to operate a total trinoless double beta decay (0νββ), where a nucleus of 5 t of isotopically enriched liquid xenon. with mass number A and charge Z decays by emitting DARWIN [11] is a proposed observatory using 40 t of only two and changes its charge by two units liquid natural xenon (LXe) in a time projection cham- (A,Z)−→(A,Z+2) + 2e−. The observation of this decay ber (TPC) with the primary goal of searching for parti- would mean that lepton number is violated by two units cle . Here, we demonstrate that DARWIN and, in the standard light Majorana neutrino exchange has a similar reach to dedicated future neutrinoless dou- scenario, would yield information about the neutrino ble beta decay experiments. This is due to its large, ho- mass scale via the effective neutrino Majorana mass mogeneous target, and its ultra-low background, cou- 2 pled to the capability of the TPC to simultaneously hmββi = |ΣiU mi|. The sum is over the neutrino mass ei measure the location, energy, particle type and multi- eigenstates, mi, and Uei, the corresponding entries in the lepton mixing matrix, which are complex numbers. plicity of an event [12]. The two-neutrino double beta decay mode (2νββ) is The paper is organized as follows: in Sect.2 we pro- allowed in the Standard Model and has been observed vide a brief review of the baseline design of the DAR- in more than 10 nuclei [1]. In this case, the summed WIN detector and describe the detector model utilized energy of the two electrons is a continuum, while for in our simulation study. Sect.3 addresses the signal the 0νββ-decay the distinct signature is a peak at the topology and how it is used to reject background events. Q-value, the mass difference between the mother and In Sect.4 we discuss the expected background sources, daughter nuclei. while the resulting background spectra and rates are presented in Sect.5. We discuss DARWIN’s sensitivity [email protected] to 0νββ-decay in Sect.6 and give a summary and an [email protected] outlook in Sect.7. [email protected] †Also at University of Banja Luka, Bosnia and Herzegov- ina 2 The DARWIN Observatory ‡Now at SLAC, Menlo Park, CA 94025, USA § Also at Institute for Subatomic Physics, Utrecht Univer- DARWIN is a next-generation dark matter experiment sity, Utrecht, Netherlands ¶Also at Coimbra Polytechnic - ISEC, Coimbra, Portugal that will operate a 40 t active (50 t total) liquid xenon kAlso at SISSA, Data Science Excellence Department, Tri- TPC with the main goal to probe the entire experimen- este, Italy tally accessible parameter space for weakly interacting 3 massive particles (WIMPs) as dark matter candidates. Other physics goals include the search for the 0νββ- decay, the real-time detection of solar pp neutrinos via electron scattering, the observation of supernova and solar 8B neutrinos via coherent neutrino nucleus scat- tering and the search for solar axions, galactic axion-like particles and dark photons. The DARWIN detector is described in detail in [11]. In the baseline scenario, the detector is a cylindrical, two-phase (liquid and gas) xenon TPC with 2.6 m di- ameter and 2.6 m height. The TPC will be placed in a low-background, double-walled cryostat surrounded by an instrumented water tank to shield it from the en- vironmental radioactivity and to record the passage of cosmic muons and their secondaries as well as for neu- Fig. 1: Drawing of DARWIN’s double-walled cryostat tron thermalization. and TPC, showing all components considered in the Interactions in the TPC will give rise to a prompt simulation. signal (S1) from photons and a delayed, proportional scintillation signal (S2) from electrons transported by previously-measured specific activities of cryostat ma- a homogeneous drift field and extracted into the gas terials [13,14] showed that a cryostat made of titanium phase. Both signals will be detected by photosensor ar- yields a lower background rate than a stainless steel rays (made of photomultiplier tubes (PMTs), silicon cryostat of equal mechanical properties. photomultiplier (SiPM), or new types of sensors), pro- The inner cryostat contains the liquid xenon volume viding the x-y-z-coordinates of an interaction, as well as and the TPC. The TPC walls are formed by PTFE its energy with < 1% 1 σ resolution for MeV energy de- reflectors of 3 mm thickness with high reflectivity for positions. Interactions separated by more than 15 mm the vacuum ultra-violet (VUV) scintillation light, sur- are assumed to be individually identified in event re- rounded by 92 cylindrical copper field shaping rings. construction. This allows for separation between sin- The structure is reinforced with 24 PTFE support pil- gle scatters (as expected from 0νββ-decays and dark lars. Titanium frames at the bottom and top of the matter particle interactions) and multiple scatters (as TPC support the electrodes to establish drift and ex- expected from many sources of backgrounds), as well traction fields. Two photosensor arrays are located at as the definition of an inner (fiducial) volume with re- the top and bottom of the TPC cylinder, consisting of duced background levels. The high density of the liquid a structural copper support, a PTFE reflector disk, the xenon (∼3 g/cm3) ensures a short attenuation length VUV-sensitive photosensors and the sensors’ cold elec- for γ-rays. tronics. Because the final sensor type is yet to be chosen The final location of the DARWIN experiment is for DARWIN and R&D on light sensor options [15,16, yet to be decided. A good candidate is the Gran Sasso 17,18] is ongoing, the top and bottom sensors have, for Underground Laboratory (LNGS) in Italy. We will use the majority of simulations, been simplified to two disks its overburden in this study. which properly account for the material budget and the associated activities of radioactive isotopes. This allows for a direct comparison between a baseline scenario with 2.1 Monte Carlo model of the detector PMTs and an alternative based on SiPMs. All the major components included in the simula- For the Monte Carlo event generation and particle prop- tions are listed in Table1. The assumed radioactivity agation in geant4 we use a realistic model of the DAR- levels of the materials are discussed in Sect.4 and listed WIN detector. Its details are described in the following. in Table2. The TPC is enclosed within the outer and inner ti- tanium cryostat (shown in Fig.1), including torispher- ical domes, flanges and stiffening rings to minimize the 3 0νββ signal events in liquid xenon amount of material. A dome-shaped pressurizable ti- tanium vessel is placed on the inner cryostat floor to In a 0νββ-decay, the energy Qββ is released mainly in reduce the volume to be filled with liquid xenon while the form of kinetic energy of the two electrons. In liquid keeping the material budget low. A study based on xenon, the electrons thermalize within O(mm) result- 4

Component Material Mass Outer cryostat Titanium 3.0 t Inner cryostat Titanium 2.1 t Bottom pressure vessel Titanium 0.4 t LXe instrumented target LXe 39.3 t LXe buffer outside the TPC LXe 9.0 t LXe around pressure vessel LXe 270 kg GXe in top dome + TPC top GXe 30 kg TPC reflector (3mm thickness) PTFE 146 kg Structural support pillars (24 units) PTFE 84 kg Electrode frames Titanium 120 kg Fig. 2: Simulated energy deposition (color scale) of two Field shaping rings (92 units) Copper 680 kg different 0νββ-events in the x-y-plane. Left: The two Photosensor arrays (2 disks): electrons thermalize along two non-resolvable back-to- Disk structural support Copper 520 kg Reflector + sliding panels PTFE 70 kg back tracks. The emitted Bremsstrahlung photons yield Photosensors: 3” PMTs (1910 units) composite 363 kg detached energy depositions. Right: A O(400 keV) pho- Sensor electronics (1910 units) composite 5.7 kg ton Compton scatters 8 mm from the position of the de- caying 136Xe nucleus and travels more than 2 cm with- Table 1: List of detector components included in the out energy loss before absorption. The circles indicate geant4 geometry model of DARWIN stating their ma- the boundaries of individually resolvable clusters as- terial composition and total mass. suming a separation threshold  = 15 mm.

ing in a single site (SS) signal topology, as shown in Fig.2 (left). Bremsstrahlung photons emitted during electron thermalization travel some distance without energy deposition before scattering or being absorbed. Abundantly emitted low energy photons are likely to deposit their energy close to the decay position and remain unresolved in the DARWIN detector. Photons with energies above 300 keV have a mean free path of more than 15 mm and might travel larger distances be- fore interacting. This can result in an energy deposition which is spatially separable and can cause a false iden- tification as a multi site (MS) event, Fig.2 (right). Energy depositions are therefore spatially grouped using a density-based spatial clustering algorithm [19]. An energy deposition is considered as a new cluster if Fig. 3: Efficiency of 0νββ signal acceptance and back- its distance to any previous energy deposition is larger ground rejection as a function of the spatial separation than our selected separation threshold . Fig.3 shows threshold . The three signal lines (blue) compare differ- the efficiency for signal acceptance and background re- ent energy and angular distributions based on a back- jection for photons and electrons with an energy of Qββ to-back electron emission, a mass mixing (MM) mech- as a function of . anism and a right-handed current (RHC) model. The The distribution of energy per electron and the an- background rejection efficiency is shown for γs (red) gle between the two depend on the yet unknown decay and electrons (green) with E = Qββ. The vertical line mechanism. We assume a mass mixing mechanism and (grey) corresponds to the value of  assumed in this the most probable decay where the electrons are emit- study. Bands indicate ±2 σ uncertainties. ted back-to-back, each with a kinetic energy of Qββ/2. This assumption is compared in Fig.3 to the predicted energy and angular distributions in the mass mixing (MM: 88.7%, RHC: 86.6%) as SS events. Background (MM) model and a right-handed current (RHC) model events from electrons and photons with Qββ energy are presented in [20]. rejected as MS with an efficiency of 17.7% and 85.1%, We assume that a spatial separation between energy respectively. A smaller separation threshold  results in depositions of  = 15 mm can be resolved in the DAR- a larger fraction of misidentified 0νββ-decays. Simulta- WIN TPC. This results in a signal acceptance of 90.4% neously, electrons from β decays and γ-ray events are 5 more efficiently identified as MS. As we will discuss in – The 2νββ decay spectrum of 136Xe has been sim- Sect.7, a lower spatial threshold can increase the sen- ulated assuming the measured half-life of T1/2 = sitivity to 0νββ-decays. The decrease in signal accep- (2.165 ± 0.061) × 1021 yr [25]. For the analytic tance is overcompensated by the improved background spectrum we use the non-relativistic Primakoff- rejection. Rosen approximation for the interaction between nuclei and electrons in the parametrization dis- cussed in [26]. This approximation is conservative 4 Background events in the 0νββ energy range as it overestimates the rate around the spectral end point. We discuss all background sources which contribute – 222Rn in LXe is assumed to be reduced by on- events within the energy range of [2.3 - 2.7] MeV around line cryogenic distillation [27] and stringent ma- Qββ. We consider intrinsic background events from ra- terial selection to a concentration equivalent to dioactive decays (radiogenic) and those induced by cos- 0.1 µBq 222Rn activity per kg of xenon. Being cru- mic neutrinos, muons and their secondaries (cosmo- cial for the WIMP search, significant efforts are be- genic). Intrinsic events are homogeneously distributed ing undertaken to reach this design goal. The domi- in the liquid xenon. Likewise we study radiogenic back- nant intrinsic background contribution for the 0νββ 214 ground radiation from external sources emanating into search originates from the β-decay of Bi (Qβ = the target. 3.27 MeV). In 19.1% of the cases it decays to the 214Po ground state without γ-emission, which ren- ders the rejection based on spatial topology rather 4.1 Homogeneously distributed intrinsic background inefficient, as discussed in Sect.3. The short half-life 214 of the decay daughter Po (T1/2 = 164.3 µs), how- The intrinsic background sources originate from noble ever, allows for BiPo event tagging and suppression gas isotopes or from interactions of cosmogenic particles with more than 99.8 % efficiency [28]. with the xenon target:

– 8B solar neutrinos are an irreducible background 4.2 External radiogenic background sources source. The expected rate of ν-e− scatterings is de- 8 rived assuming a B-ν flux of φ = (5.46 ± 0.66) × Long-lived radionuclides are present in each detector 6 −2 −1 10 cm s [21]. The calculation of scattering cross material. Their decays, as well as the subsequent de- sections follows [22]. The electron neutrino survival cays of their daughter isotopes, might introduce back- probability is conservatively estimated to be Pee = ground in the target. Activity levels for all materials are +5% 0.50−30% for neutrinos with Eν > Qββ. listed in Table2 and based on reports from previous or 137 – Xe from cosmogenic activation: muon-induced ongoing experiments [13,14]. neutrons produced in the liquid xenon can thermal- ize and be captured on a 136Xe nucleus, producing – The natural decay chains of 238U, 232Th and 235U 137Xe, as measured by EXO-200 [23]. This isotope yield a background contribution primarily from γ- − 214 208 decays via a β process with Qβ = 4.17 MeV and a rays emitted by Bi- (Eγ = 2.45 MeV) and Tl- half-life of 3.82 min. Assuming the depth of LNGS decays (Eγ = 2.61 MeV). The former two chains and previous simulations of the muon-induced neu- were split into their early and late component at tron flux underground [24], we estimate the muon- 226Ra and 228Th, respectively, to account for radio- induced 137Xe production rate in DARWIN to be genic non-equilibrium. (6.9 ± 0.4) atoms/(t·yr). Neutrons produced in the – 60Co β-decays dominantly (99.95%) via the two ex- solid materials contribute about 5% of this rate. Ac- cited states of 60Ni. The de-excitation is tempo- tivation of 136Xe due to radiogenic neutrons from rally non-resolvable and spatial coincidences of the the TPC materials has been found to be subdom- 1.17 MeV and the 1.33 MeV γ-events contribute to inant by more than two orders of magnitude. Ac- the background. tivation of xenon in the non-shielded environment – Among the radio-isotopes from cosmogenic material of the purification loop is non-negligible, but can activation at sea level [29], 44Ti in the cryostat ma- be efficiently suppressed by a delayed re-feed of the terial is the most relevant, due to its long half-life 44 LXe into the detector. Suppression by three orders (T1/2 = 59.1 yr) and the subsequent decay of Sc of magnitude adds an additional 225 kg to the total with γ-emission at 2.66 MeV. xenon budget when cycling 1000 standard liter per – 222Rn contamination in the non-instrumented minute. xenon surrounding the TPC can contribute to the 6

Material Unit 238U 226Ra 232Th 228Th 60Co 44Ti Reference Titanium mBq/kg <1.6 <0.09 0.28 0.25 <0.02 <1.16 [13] PTFE mBq/kg <1.2 0.07 <0.07 0.06 0.027 - [14] Copper mBq/kg <1.0 <0.035 <0.033 <0.026 <0.019 - [14] PMT mBq/unit 8.0 0.6 0.7 0.6 0.84 - [14] Electronics mBq/unit 1.10 0.34 0.16 0.16 <0.008 - [14]

Table 2: Assumed activity levels for the simulated materials and isotopes.

214Bi-induced γ-background. The rejection based on BiPo tagging described above cannot be applied since the subsequent alpha decays are not observed.

5 Analysis and background results

The background sources discussed in Sect.4 are sim- ulated with the geant4 particle physics simulation toolkit [30], using the detector model presented in Sect. 2.1. The equivalent of at least 100 years of DAR- WIN run time has been simulated for each material and isotope. In this section, we discuss the methods applied for event selection. The analytical background model, used for the profile-likelihood analysis in Sect. 6.2, is Fig. 4: Spatial distribution of external background also described, and the background results are dis- events inside the instrumented volume for 100 years cussed. of DARWIN run time. The colored lines indicate the contours of the optimized fiducial volumes containing different LXe target masses. The 5 t fiducial volume is 5.1 Monte Carlo data processing and event selection used for the sensitivity estimate presented below.

The energy depositions generated by geant4 per Constraining the target to a super-ellipsoidal- event undergo a density-based spatial clustering algo- shaped fiducial volume (FV) allows us to exploit the rithm [19] to topologically distinguish signal-like single excellent self-shielding capabilities of liquid xenon. To site (SS) from background-like multi site (MS) events, compensate for the reduced shielding power in the as discussed in Sect.3. We assume a separation thresh- xenon gas phase, the FV is shifted slightly downwards old  = 15 mm for the DARWIN TPC. This compara- from the center of the instrumented volume. The fidu- tively coarse clustering inevitably results in a fraction of cial volume is optimized for each FV mass indepen- γ-accompanied β-decays from background events, e.g., dently. We use the lifetime-weighted combined external 214Bi decays which frequently occur with higher multi- background, after the selection of single site events, en- plicity, being falsely identified as SS and consequentially ergy and spatial resolution smearing. Only events with contributing to the background. an energy inside the 0νββ-ROI of [2435 - 2481] keV, de- To account for the finite energy resolution of the de- fined as the full width at half maximum (FWHM) range tector, the combined energy deposited inside each clus- of the expected signal peak, are considered. The spatial ter is smeared according to a resolution of distribution of external background events inside the σ a active volume is shown in Fig.4. E = + b, (2) E pE[keV] with a = (0.3171 ± 0.0065) and b = (0.0015 ± 0.0002). 5.2 Background model and fiducial mass dependence At E = Qββ this corresponds to σE/E = 0.8%, as demonstrated in the XENON1T TPC [31]. The cluster The selection of events within a fiducial volume removes position is smeared to account for the detector’s spatial all α- and β-contributions originating from external resolution which is conservatively assumed to be σx,y = sources. The γ-background is shown in Fig.5 (bottom) σz = 10 mm above 2 MeV. for the 20 t fiducial volume. In the 0νββ-ROI the back- 7 ground is composed of the absorption peak from 214Bi volume mass range of interest, it can be considered neg- at EBi = 2.45 MeV and Compton scattered photons, ligible. The largest background contribution in the ROI 208 mainly from the Tl line (ETl = 2.61 MeV). Comp- is induced by the absorption peak of 2.45 MeV γ-rays ton scatterings inside the fiducial volume with the sub- emitted by 214Bi decays in the detector materials. The sequent escape of the scattered lower energy γ-ray are contribution from 214Bi decays in the non-instrumented strongly suppressed by fiducialization. The continuous LXe around the TPC accounts for approximately 0.1% background is dominated by photons that undergo an of the total material-induced background. undetected Compton scatter outside the detector fol- The relative contributions to the γ-background in lowed by their absorption in the fiducial volume. the ROI are shown per material of origin in Fig.5 (top). The similar contribution of cryostat-induced events from the walls and the combined PMT and electronics background originating from the top and bottom sensor array is a result of the optimization of the fiducial vol- ume, which is properly balancing the r- and z-extent. The spectral shape of the material-induced γ-back- ground is modelled with a Gaussian peak and an expo- nentially decreasing continuum for each line, as shown in Fig.5 (bottom). We consider the 2 .61 MeV 208Tl peak, the 2.66 MeV 44Sc peak and each contribution of 214 214 Bi with Eγ > 2.0 MeV. The ratio between the Bi and the 44Sc peaks to the 208Tl peak intensity is es- tablished using Monte Carlo data in fiducial volumes sufficiently large to provide high statistics. Similarly, each continuum contribution is tied to its correspond- ing peak intensity and a fixed relation between the three slope parameters is found. The only remaining free pa- rameters of the combined model are the total intensity of the 208Tl peak and one common slope parameter. The model is tested and confirmed using a χ2 goodness- of-fit test on the combined external background in the fiducial mass range ≤ 20 t. The intrinsic background from 8B neutrinos is as- sumed to be flat. The spectra corresponding to 137Xe and 222Rn are approximated linearly falling in the [2.2- 2.8 MeV] range. The slopes are obtained from Monte Carlo studies. The 2νββ spectrum is convolved with the Gaussian energy resolution. The suppression of the external background with de- Fig. 5: Composition of the material-induced external creasing fiducial mass is shown in Fig.6, together with background in the 20 t fiducial volume. Top: Relative the target mass independent intrinsic contributions. contribution to the background in the 0νββ-ROI by material and isotope. Bottom: Background spectra by isotope with the corresponding model fits. The relative 5.3 Background rates in the 0νββ-ROI contributions and spectral shapes are representative for smaller fiducial volumes. 0ν The fiducial volume is optimized for T1/2 sensitivity, as discussed in detail in Sect. 6.1, and yields 5 t. The The continuous contribution from γ-rays emitted in resulting background spectrum from intrinsic and ex- 44Sc decays accounts for less than 1% of the external ternal sources is shown for this fiducial mass in Fig.7. 214 background. Bi decays with Eγ > Qββ contribute The intrinsic background in the ROI is dominated with a similarly subdominant level. Spatial coincident by the gently falling β−-spectrum of 137Xe decay. Sub- absorption of both 60Co gammas accounts for only ap- dominant contributions are the electron scattering of proximately 10−3 of the total material background at solar 8B neutrinos and β−-events from 214Bi-decays E = 2.51 MeV in the 30 t fiducial volume. In the fiducial which are not vetoed by BiPo tagging. The 2νββ spec- 8

Background source Background index Rate Rel. uncertainty [events/(t·yr·keV )] [events/yr] External sources (5 t FV): 214Bi peaks + continuum 1.36 × 10−3 0.313 ±3.6% 208Tl continuum 6.20 × 10−4 0.143 ±4.9% 44Sc continuum 4.64 × 10−6 0.001 ±15.8% Intrinsic contributions: 8B(ν − e scattering) 2.36 × 10−4 0.054 +13.9%, −32.2% 137Xe (µ-induced n-capture) 1.42 × 10−3 0.327 ±12.0% 136Xe 2νββ 5.78 × 10−6 0.001 +17.0%, −15.2% 222Rn in LXe (0.1 µBq/kg) 3.09 × 10−4 0.071 ±1.6% Total: 3.96 × 10−3 0.910 +4.7%, −5.0%

Table 3: Expected background index averaged in the 0νββ-ROI of [2435 - 2481] keV, corresponding event rate in the 5 t FV and relative uncertainty by origin.

driven) input parameters. The dominant contributions are the νe survival probability and the neutrino flux (8B ν-e− scattering), the 136Xe neutron capture cross- section (governing the 137Xe production rate) and the half-life of 136Xe (2νββ decay).

Fig. 6: Background rate in the ROI versus fiducial mass. External contributions are combined. Fiducial volume independent intrinsic sources are shown per contribu- tion. Bands indicate ±1 σ uncertainties. At 5 t, the ex- ternal sources contribute at the same level as the com- bined intrinsic background. Fig. 7: Predicted background spectrum around the 0νββ-ROI for the 5 t fiducial volume. A hypothetical trum overlaps negligibly with the ROI, but dominates 0ν signal of 0.5 counts per year corresponding to T1/2 ≈ the background toward lower energies. 2 × 1027 yr is shown for comparison. Bands indicate The model-estimated background indices for all con- ±1 σ uncertainties. tributions are summarized in Table3. To validate the analytic model introduced in Sect. 5.2, we compare the background model estimate with the values derived by weighted event counting in the 5 t fiducial mass data from Monte Carlo. Both results are in agreement within 6 Sensitivity Calculation the statistical errors. The model-derived uncertainty on the background, however, is a factor of 4 lower than We use the background rates predicted in Sect. 5.3 to the Poissonian statistics error in the simple counting derive a limit on the half-life sensitivity at 90% confi- approach. The uncertainties on intrinsic background dence level (C.L.) as well as the 3 σ discovery potential sources account for statistical errors, the variation of for the 0νββ-decay. The latter is defined as the mini- 0ν the overlap with the 0νββ-ROI based on the energy mal value of T1/2 required to exclude the null hypothesis resolution and systematic uncertainties from (theory- with a median significance of 99.7% C.L. 9

0ν 27 6.1 Half-life sensitivity estimation We obtain a T1/2 sensitivity limit of 2.4 × 10 yr for a 10 year exposure with 5 t fiducial mass. The corre- Based on the figure-of-merit estimator proposed in [32] sponding 3 σ discovery potential after 10 years exposure we calculate the half-life sensitivity at 90% C.L. as: is 1.1 × 1027 yr. √ 0ν  fROI α NA Mt T1/2 = ln 2 √ , (3) 1.64 MXe B∆E 7 Discussion with  = 0.9 being the detection efficiency of a single site 0νββ-decay event, fROI = 0.76 the fraction of signal The DARWIN observatory will reach a sensitivity to the 136 covered by the ROI, α = 0.089 the abundance of Xe neutrinoless double beta decay of 136Xe of 2.4 × 1027 −1 in natural xenon, NA the Avogadro number in mol , years T1/2 exclusion limit (90% C.L.) and a discovery MXe the molar mass number of xenon in t/mol, M the 27 sensitivity (3 σ) of T1/2 = 1.1 × 10 years after 10 years fiducial mass in tons, t the exposure time in years, B of exposure. the background index in t−1yr−1keV−1, and ∆E the In the baseline scenario discussed above, the as- width of the ROI in keV. The value 1.64 is the number sumptions on radio-purity and detector performance of standard deviations corresponding to a 90% C.L. are considered realistic or even conservative. In an op- Following Eq. (3) and using the background index timistic scenario, the external background could be re- for the 5 t fiducial mass (Table3), we obtain a half-life duced by a factor of three or more. The required mea- sensitivity of2 .0 × 1027 yr(1 .3 × 1027 yr) after 10 (4) sures include the use of less radioactive PMTs (with years of exposure. reduced mass of ceramic [33]) and/or low radioactivity This figure-of-merit estimation is an established tool SiPMs, more stringent material selection to reach lower to directly compare 0νββ sensitivities of different ex- levels of radio-activity for PTFE [34], copper [35] and periments using common statistical methods and as- titanium, as well as more radio-pure electronics. sumptions. It also allows for a straightforward assess- Intrinsic backgrounds, dominated by the muon- ment of the sensitivity as a function of different param- induced activation of 136Xe, are difficult to mitigate eters, such as the fiducial mass. It does not, however, assuming the muon flux at 3500 meter water equiva- consider background uncertainties, but assumes perfect lent (mwe) depth of LNGS. A time- and spatial- muon knowledge of the background rates. veto might allow for suppression by up to a factor of two at an acceptable exposure loss. The 137Xe contribution 6.2 Frequentist profile-likelihood analysis would, however, become subdominant in a sufficiently deep laboratory. A total intrinsic background suppres- To account for and effectively constrain the background sion by a factor of five or even eight could then be uncertainties, we apply a profile-likelihood analysis reached assuming a reduced BiPo tagging inefficiency based on the background model discussed in Sect. 5.2. of 0.1% and 0.01%, respectively. Assuming a factor five The inserted signal is a Gaussian peak with Qββ and reduction in external sources the latter scenario leads 8 σE(Qββ) according to Eq. (2), which is scaled by the to a solar B neutrino dominated background. 136Xe atoms in the target volume, an activity corre- The sensitivity could be increased by further ex- 0ν sponding to T1/2 and the detection efficiency, as shown ploitation of the SS/MS discrimination, discussed in in Fig.7. Sect.3. Despite increased signal rejection, the gain in Background uncertainties from the model are background reduction dominates for spatial separation treated as nuisance parameters with Gaussian con- thresholds down to  = 3 mm. The cluster separation straining terms in the likelihood. For external back- in the x-y-plane would benefit from a higher granular- ground contributions, their variances are obtained ei- ity photosensor top array, featuring e.g. SiPMs. The z- ther by the model fit on the spectrum corresponding position reconstruction is already more accurate and a to 5 t FV (208Tl peak intensity and slope parameter) combined three dimensional charge signal analysis will or extrapolation of the model parameters from larger optimize the separation. fiducial volumes into the low fiducial mass range (214Bi The largest sensitivity increase can be achieved with / 208Tl peak ratio, 208Tl continuum / 208Tl peak inten- a combination of the above mentioned measures. Fig.8 sity). The uncertainty on the subdominant contribution shows the fiducial volume mass dependency (top) and from 44Sc has been neglected. For the intrinsic contri- time evolution (bottom) of the DARWIN half-life limit butions, the variances correspond to the square of the sensitivity (90% C.L.) calculated with the figure-of- errors listed in Table3. The corresponding slope uncer- merit estimator (see Sect. 6.1) for the baseline and dif- tainties are negligible. ferent optimistic scenarios. The latter assume reduced 10 spatial separation threshold , intrinsic and external The objective of detecting particle dark matter with background rates. Fig.9 translates the half-life limit a sensitivity down to the neutrino floor requires the sensitivity to the effective Majorana neutrino mass mββ DARWIN observatory to be an ultra-low background 136 using Eq. (1), where the mββ range corresponds to the experiment. It additionally features a high Xe tar- range of published nuclear matrix elements [36]. Un- get mass, excellent energy resolution and single site der the conservative baseline assumptions, DARWIN discrimination capability. In the presented baseline sce- reaches a mββ limit of [18-46 meV]. The neutrino dom- nario DARWIN will reach a sensitivity that approaches inated scenario yields a limit in the [11-28 meV] range. that of the tonne-scale proposed 0νββ experiments. Un- Future dedicated neutrinoless double beta decay exper- der more optimistic assumptions, requiring adaptations iments using either 136Xe or other isotopes are aiming to the baseline design, DARWIN will explore the full for a similar science reach as DARWIN, as shown for inverted hierarchy and will compete with the most am- comparison in Table4 and in Fig.8 (bottom). bitious proposed 0νββ projects.

Fig. 9: Effective Majorana neutrino mass vs. lightest neutrino mass. The sensitivity reach after 50 t×yr of exposure is shown for the baseline and the optimistic neutrino dominated scenario. The horizontal bands stem from the range of nuclear matrix elements [36]. Global sensitivity according to [38], oscillation param- eters from [39,40].

0ν Fig. 8: DARWIN median T1/2 sensitivity at 90% C.L. as a function of fiducial volume mass for 10 years of expo- sure (top) as well as of the exposure time for the opti- Acknowledgements mized fiducial volume (bottom). The baseline design is compared with different optimistic scenarios. The latter This work was supported by the Swiss National Sci- assume a reduction of the external (ext.) and the intrin- ence Foundation under grants No 200020-162501 and sic (int.) backgrounds and improved spatial separation No 200020-175863, by the European Union’s Hori- threshold of 10 mm (red, blue) or 5 mm (green). Sen- zon 2020 research and innovation programme un- sitivity projections for future 136Xe 0νββ experiments der the Marie Sklodowska-Curie grant agreements are shown for comparison [8,9,10,37]. No 674896, No 690575 and No 691164, by the Eu- ropean Research Council (ERC) grant agreements No 742789 (Xenoscope) and No 724320 (ULTI- MATE), by the Max-Planck-Gesellschaft, by the 11

Experiment Isotope Sensitivity limit (90% C.L.) Exposure time Reference 0ν T1/2 [yr] mββ [meV] [yr] DARWIN (baseline) 136Xe 2.4 × 1027 18-46 10 this work DARWIN (ν dominated) 136Xe 6.2 × 1027 11-28 10 this work KamLAND2-Zen 136Xe 6 × 1026 37-91 5 [37] PandaX-III 136Xe 1 × 1027 28-71 3 [9] NEXT-HD 136Xe 3 × 1027 16-41 10 [8] nEXO 136Xe 9.2 × 1027 9-23 10 [10] SNO+-II 130Te 7 × 1026 20-70 5 [37] AMoRE-II 100Mo 5 × 1026 15-30 5 [37] CUPID 130Te / 100Mo (2-5)×1027 6-17 10 [37] LEGEND-1000 76Ge 1 × 1028 11-28 10 [37]

0ν Table 4: Comparison of T1/2 and mββ sensitivity limits (90% C.L.) between DARWIN and future 136 0νββ experiments. For experiments using Xe the mββ ranges are calculated with the nuclear matrix element ranges from [36], those using other isotopes are taken from [37].

Deutsche Forschungsgemeinschaft (DFG) under GRK- 10. nEXO collaboration, J. B. Albert et al., Sensitivity and 2149, by the US National Science Foundation (NSF) Discovery Potential of nEXO to Neutrinoless Double Beta grants No 1719271 and No 1940209, by the Portuguese Decay, Phys. Rev. C97 (2018) 065503,[1710.05075]. 11. DARWIN collaboration, J. Aalbers et al., DARWIN: FCT, by the Netherlands Organisation for Scientific towards the ultimate dark matter detector, JCAP 1611 Research (NWO), by the Ministry of Education, Sci- (2016) 017,[1606.07001]. ence and Technological Development of the Republic of 12. M. Schumann, Dual-Phase Liquid Xenon Detectors for Serbia and by grant ST/N000838/1 from Science and Dark Matter Searches, JINST 9 (2014) C08004, [ ]. Technology Facilities Council (UK). 1405.7600 13. LZ collaboration, D. S. Akerib et al., Identification of Radiopure Titanium for the LZ Dark Matter Experiment and Future Rare Event Searches, Astropart. Phys. 96 (2017) 1–10,[1702.02646]. References 14. XENON collaboration, E. Aprile et al., Material radioassay and selection for the XENON1T dark matter 1. A. S. Barabash, Average and recommended half-life values experiment, Eur. Phys. J. C77 (2017) 890,[1705.01828]. for two neutrino double beta decay, Nucl. Phys. A935 15. L. Baudis, M. Galloway, A. Kish, C. Marentini and (2015) 52–64,[1501.05133]. J. Wulf, Characterisation of Silicon Photomultipliers for 2. R. Henning, Current status of neutrinoless double-beta Liquid Xenon Detectors, JINST 13 (2018) P10022, decay searches, Rev. Phys. 1 (2016) 29–35. [1808.06827]. 3. M. J. Dolinski, A. W. P. Poon and W. Rodejohann, 16. L. Baudis et al., Performance of the Hamamatsu R11410 Neutrinoless Double-Beta Decay: Status and Prospects, in cryogenic Xenon Environments, Annu. Rev. Nucl. (2019),[ 1902.04097]. JINST 8 (2013) P04026,[1303.0226]. 4. M. Redshaw, E. Wingfield, J. McDaniel and E. G. 17. F. Arneodo, M. Benabderrahmane, G. Bruno, Myers, Mass and double-beta-decay Q value of Xe-136, V. Conicella, A. D. Giovanni, O. Fawwaz et al., Phys. Rev. Lett. 98 (2007) 053003. Cryogenic readout for multiple VUV4 Multi-Pixel Photon 5. EXO-200 collaboration, G. Anton et al., Search for Counters in liquid xenon, Nucl. Instrum. Meth. A893 Neutrinoless Double-β Decay with the Complete EXO-200 (2018) 117–123,[1707.08004]. Dataset, Phys. Rev. Lett. 123 (2019) 161802, 18. D. Ferenc, A. Chang and M. Segedinˇ Ferenc, The Novel [1906.02723]. ABALONE Photosensor Technology: 4-Year Long Tests of 6. KamLAND-Zen collaboration, A. Gando et al., Search Vacuum Integrity, Internal Pumping and Afterpulsing, for Majorana Neutrinos near the Inverted Mass Hierarchy ArXiv (2017) , [1703.04546]. Region with KamLAND-Zen, Phys. Rev. Lett. 117 (2016) 19. M. Ester, H.-P. Kriegel, J. Sander and X. Xu, A 082503,[1605.02889]. density-based algorithm for discovering clusters in large 7. NEXT collaboration, V. Alvarez et al., NEXT-100 spatial databases with noise, in KDD Proceedings, Technical Design Report (TDR): Executive Summary, pp. 226–231, AAAI Press, 1996. JINST 7 (2012) T06001,[1202.0721]. 20. R. Arnold and et al., Probing new physics models of 8. J. J. Gomez-Cadena, Status and prospects of the NEXT neutrinoless double beta decay with supernemo, Eur. Phys. experiment for neutrinoless double beta decay searches, J. C 70 (2010) 927–943,[1005.1241]. ArXiv (2019) , [1906.01743]. 21. Particle Data Group collaboration, M. Tanabashi 9. X. Chen et al., PandaX-III: Searching for neutrinoless et al., Review of Particle Physics, Phys. Rev. D98 (2018) double beta decay with high pressure 136Xe gas time 030001. projection chambers, Science China Physics, Mechanics, 22. B. Dutta and L. E. Strigari, Neutrino physics with dark and Astronomy 60 (2017) 061011,[1610.08883]. matter detectors, Annu. Rev. Nucl (2019),[ 1901.08876]. 12

23. EXO-200 collaboration, J. Albert et al., Cosmogenic Backgrounds to 0νββ in EXO-200, JCAP 04 (2016) 029, [1512.06835]. 24. XENON1T collaboration, M. Selvi, Study of the performances of the shield and muon veto of the XENON1T experiment, PoS IDM2010 (2011) 053. 25. EXO-200 collaboration, J. B. Albert et al., Improved measurement of the 2νββ half-life of 136Xe with the EXO-200 detector, Phys. Rev. C89 (2014) 015502, [1306.6106]. 26. F. Boehm and P. Vogel, Physics of Massive Neutrinos. Cambridge University Press, 2 ed., 1992. 27. XENON collaboration, E. Aprile et al., Online 222Rn removal by cryogenic distillation in the XENON100 experiment, Eur. Phys. J. C77 (2017) 358,[1702.06942]. 28. L. Baudis et al., Neutrino physics with multi-ton scale liquid xenon detectors, JCAP 01 (2014) 044,[1309.7024]. 29. C. Zhang, D. M. Mei, V. A. Kudryavtsev and S. Fiorucci, Cosmogenic Activation of Materials Used in Rare Event Search Experiments, Astropart. Phys. 84 (2016) 62–69,[1603.00098]. 30. GEANT4 collaboration, S. Agostinelli et al., GEANT4: A Simulation toolkit, Nucl. Instrum. Meth. A506 (2003) 250–303. 31. XENON collaboration, E. Aprile et al., Energy resolution and linearity in the keV to MeV range measured in XENON1T, ArXiv (2020) , [2003.03825]. 32. F. T. Avignone, G. S. King and Yu. G. Zdesenko, Next generation double-beta decay experiments: Metrics for their evaluation, New J. Phys. 7 (2005) 6. 33. XENON collaboration, E. Aprile et al., Lowering the radioactivity of the photomultiplier tubes for the XENON1T dark matter experiment, Eur. Phys. J. C75 (2015) 546,[1503.07698]. 34. M. Auger et al., The EXO-200 detector, part I: Detector design and construction, JINST 7 (2012) P05010, [1202.2192]. 35. Majorana collaboration, N. Abgrall et al., The Majorana Demonstrator radioassay program, Nucl. Instrum. Meth. A 828 (2016) 22–36,[1601.03779]. 36. J. Engel and J. Menndez, Status and Future of Nuclear Matrix Elements for Neutrinoless Double-Beta Decay: A Review, Rept. Prog. Phys. 80 (2017) 046301,[1610.06548]. 37. A. S. Barabash, Possibilities of Future Double Beta Decay Experiments to Investigate Inverted and Normal Ordering Region of Neutrino Mass, Frontiers in Physics 6 (2019) 160,[1901.11342]. 38. M. Agostini et al., Probing Majorana neutrinos with double-β decay, Science 365 (2019) 1445–1448. 39. I. Esteban, M. C. Gonzalez-Garcia, A. Hernandez-Cabezudo, M. Maltoni and T. Schwetz, Global analysis of three-flavour neutrino oscillations: synergies and tensions in the determination of θ23, δCP , and the mass ordering, JHEP 01 (2019) 106, [1811.05487]. 40. NuFIT 4.1 (2019), www.nu-fit.org.