The XLZD Design Book: Towards the Next-Generation Liquid Xenon Observatory for Dark Matter and Neutrino Physics

Abstract

This report describes the experimental strategy and technologies for a next-generation xenon observatory sensitive to dark matter and neutrino physics. The detector will have an active liquid xenon target mass of 60–80 tonnes and is proposed by the XENON-LUX-ZEPLIN-DARWIN (XLZD) collaboration. The design is based on the mature liquid xenon time projection chamber technology of the current-generation experiments, LZ and XENONnT. A baseline design and opportunities for further optimization of the individual detector components are discussed. The experiment envisaged here has the capability to explore parameter space for Weakly Interacting Massive Particle (WIMP) dark matter down to the neutrino fog, with a 3$\sigma$ evidence potential for spin-independent WIMP-nucleon cross sections as low as $3\times 10^{-49}\rm\,cm^{2}$ (at 40 GeV/c² WIMP mass). The observatory is also projected to have a 3$\sigma$ observation potential of neutrinoless double-beta decay of ¹³⁶Xe at a half-life of up to $5.7\times 10^{27}$ years. Additionally, it is sensitive to astrophysical neutrinos from the atmosphere, sun, and galactic supernovae.

1 Introduction

The nature of dark matter is one of the most important unsolved questions in physics today. Liquid xenon time projection chambers (LXe-TPCs) have been leading the search for Weakly Interacting Massive Particle (WIMP) dark matter candidates above a few GeV/c² for over a decade, building on many generations of experiments with ever-increasing target mass Alner:2007ja ; Lebedenko:2008gb ; Angle:2007uj ; Aprile:2016swn ; Aprile:2017iyp ; Akerib:2016vxi ; Xiao:2014xyn ; Wang:2020coa ; PandaX-4T:2021bab . The latest iterations of these experiments have demonstrated world-leading sensitivity and unprecedented low background levels with detector masses of several tonnes LZ_2024_DM_results ; XENON:2023sxq ; PandaX-4T:2024dm . The XENON-LUX-ZEPLIN-DARWIN (XLZD) collaboration brings together the collective expertise from these experiments and the leading R&D efforts to build a dual-phase LXe-TPC (here referred to as the XLZD experiment) with the capability to explore the WIMP parameter space down to cross sections where the expected number of neutrino events in the signal region would be equal to WIMP events - often referred to as the neutrino fog Billard:2013qya ; O'Hare:2015mda ; OHare:2021utq . The envisioned detector could readily accomplish a 200 t$\cdot$y fiducial exposure within its initial operating period while being designed to allow sensitivity improvement for up to 1000 t$\cdot$y exposure to deliver a 3$\sigma$ discovery at the background-limited neutrino fog. The XLZD experiment will also have competitive sensitivity to neutrinoless double-beta decay ($0\nu\beta\beta$) of ¹³⁶Xe and astrophysical neutrinos, making it the definitive rare event observatory with considerable impact on particle, nuclear and astrophysics Aalbers:2022dzr .

In this report, we present the experimental strategy and provisional design of the XLZD detector with 60–80 tonnes xenon target mass depending on the acquisition rate of the xenon. The detector design is driven by the requirement to minimize the particle detection threshold while optimizing the self-shielding of external backgrounds from the laboratory environment and detector construction materials, mitigation of internal trace contaminants, mostly radon and krypton, in the liquid xenon target, as well as instrument related accidental coincidence backgrounds. The mature technology of current-generation LXe-TPCs provides a reliable base for our design, profiting from the often complementary strengths and expertise of the merging collaborations. Existing screening techniques and purification mechanisms are in place or are currently being developed to ensure that backgrounds due to trace radioactivity fall below the level of irreducible neutrino backgrounds in the form of neutrino-electron and neutrino-nucleus scattering. Technical risks exist, as is to be expected for such an ambitious program Kopec:2023uii . However, these risks are well understood, and mitigation strategies such as early, long-term, and large-scale testing are planned.

2 Science Sensitivity and Design Drivers

Figure 1: The science channels of the proposed LXe observatory for rare events span many areas and are of interest to particle physics, nuclear physics, astrophysics, solar physics, and cosmology.

Advances in liquid xenon (LXe) detectors open the possibility to design, commission, and operate an observatory for several extremely rare processes with far-reaching impact for particle-, astro- and nuclear physics and associated fields Aalbers:2022dzr – see Fig. 1. Such a detector is ideal for probing some of the most popular dark matter models, with the potential for a momentous discovery allowing characterization of the particle nature of galactic dark matter. The same detector can also search for neutrinoless double-beta decay of ¹³⁶Xe to determine if the neutrino is its own antiparticle, i.e., a Majorana particle. It can also function as an observatory for several astrophysical neutrino sources, including solar pp and ⁸B neutrinos, atmospheric neutrinos, and neutrinos from a galactic supernova explosion. The detector proposed by the XLZD collaboration will thus be a pivotal observatory to tackle some of the biggest mysteries in physics.

2.1 Dark Matter

The gravitational effects of dark matter are evident on astrophysical and cosmic scales clowe2006direct ; aghanim2020planck , but the fundamental nature of dark matter remains a mystery. One of the most compelling hypotheses is that dark matter consists of new subatomic particles, the so-called Weakly Interacting Massive Particles (WIMPs)  gelmini2010dm ; Jungman:1995df . The relic abundance of dark matter is easily reproducible by adding a coupling to standard model particles at the electroweak scale Bertone:2010zza . Xenon, with an average atomic mass of $\sim$130 GeV/c², is an ideal target to search for kinematic collisions from such WIMPs. In the simplest case, the interaction would be a spin-independent (SI) interaction between a WIMP and a xenon nucleus. For the last two decades, experiments utilizing LXe-TPCs have led the search for WIMPs Alner:2007ja ; Lebedenko:2008gb ; Angle:2007uj ; Aprile:2016swn ; Akerib:2016vxi ; PandaX-4T:2021bab ; AKIMOV201214 ; Aprile:2018dbl , with current most-competitive results derived from the initial data of LUX-ZEPLIN (LZ), PandaX-4T and XENONnT LZ_2024_DM_results ; XENON:2023sxq ; PandaX-4T:2024dm . While these experiments have ruled out various WIMP candidates, many well motivated WIMPs remain, which can be realized in simple extensions to the SM. For example, electroweak multiplet DM Bloch:2024 and $Z^{\prime}$ mediated models Blanco:2019hah provide minimal WIMP candidates that satisfy experimental constraints. More complex extensions to the SM, such as supersymmetry, can also offer viable WIMPs ATLAS2024 ; Ellis_2023 . In these benchmark cases, significant parts of the remaining parameter space will be explored by XLZD (see Fig. 2 right), probing higher WIMP masses than accessible at the LHC.

XLZD will also have competitive sensitivity to spin-dependent WIMP-nucleon scattering via the naturally occuring ¹²⁹Xe (spin 1/2, 26.4% natural abundance) and ¹³¹Xe (spin 3/2, 21.2% natural abundance) isotopes. Current generation liquid xenon detectors have demonstrated leading sensitivities for both WIMP-neutron and WIMP-proton scattering LZ_2024_DM_results ; XENON:2023sxq ; LZ:2022ufs .

Though designed to target WIMP-nuclear scattering, these versatile detectors have exceptional sensitivity to other types of WIMP interactions (for example, WIMP-pion scattering Aprile:2018cxk , inelastic WIMP-nucleus scattering Baudis:2013bba , and Effective Field Theory analysis LZ:2023lvz ; Hoferichter:2016nvd ; Hoferichter:2018acd ) and additional well-motivated dark-matter candidates: sub-GeV Essig:2012yx , dark photon Pospelov:2008jk ; Essig:2013lka , axion-like particle Abbott:1982af ; raffelt2002axions ; preskill1983cosmology ; duffy2009axions , and Planck mass dark matter XENON:2023iku . Prominent examples are listed in Fig. 1 with sensitivity studies summarised in Ref. Aalbers:2022dzr .

Figure 2: Left: Projected $90$% C.L. upper limits on the spin-independent WIMP-nucleon cross section for 200 and 1000 tonne-year (t$\cdot$y) exposures of the XLZD detector, along with current upper limits LZ_2024_DM_results ; XENON:2023sxq ; PandaX-4T:2024dm ; Akerib:2015rjg . The blue shaded regions illustrate the “neutrino fog”OHare:2021utq . Right: Median evidence potential curves for 3$\sigma$ for both exposures. The example evidence contours for 20 and 80 GeV WIMPs are 1-,2- and 3-sigma (yellow, orange, red) confidence intervals with the 1000 t$\cdot$y exposure. The plot also illustrates example regions of interest for dark matter candidates where the overlaid shading (labelled as ATLAS) illustrates model regions excluded by ATLAS Bloch:2024 ; ATLAS2024 ; Ellis_2023 .

Figure 2 illustrates the decisive progress achievable with the XLZD detector in the flagship spin-independent WIMP-nucleon search channel. As the definitive WIMP discovery instrument, the experiment must reach into the neutrino fog Billard:2013qya ; O'Hare:2015mda ; OHare:2021utq , where its discovery potential becomes systematically limited by the coherent nuclear scattering of astrophysical neutrinos (CEvNS). In the nuclear recoil signal region, non-neutrino background events are maintained at the order of a single event expected within the entire exposure. However, the expected reach into the neutrino fog predicts more than one coherent elastic neutrino-nucleus scatter. The other backgrounds arise from neutrons, which can be reduced by low-background material selection, vetoed using the outer detector volumes, and discriminated against with potential multiple scatterings inside the detector. In addition, leakage from the more numerous electronic recoil (ER) backgrounds into the nuclear recoil signal region will need to be controlled. With a target suppression of ²²²Rn to $0.1\text{\,}\mu\mathrm{B}\mathrm{q}\mathrm{/}\mathrm{k}\mathrm{g}$ and ^natKr to 0.1 ppt, the processes dominating electron recoils will be those from solar (mostly pp) neutrinos scattering off electrons and naturally occurring ¹³⁶Xe in the LXe Schumann:2015cpa . Background reduction strategies to achieve these goals are further discussed in Section  5.

With a conservative 200 t$\cdot$y exposure, assuming a successful suppression of radioactive backgrounds below the irreducible neutrino backgrounds and negligible leakage of electron recoil events from solar neutrinos into the signal region, the experiment will deliver an order of magnitude improvement in exclusion sensitivity and discovery capability compared to current experiments. For a 40 GeV/c² WIMP it will reach 90% exclusion sensitivity down to a cross-section of $2\times 10^{-49}\rm\,cm^{2}$ and 3$\sigma$ evidence capability at a cross-section of $7\times 10^{-49}\rm\,cm^{2}$. In order to be the definitive xenon experiment, the detector must be capable of running up to a 1000 t$\cdot$y exposure without becoming limited by backgrounds from radioactive impurities, making it sensitive to a potential 3$\sigma$ evidence at a cross-section of $3\times 10^{-49}\rm\,cm^{2}$ for a 40 GeV/c² WIMP as shown in Fig. 2. In case of a discovery, XLZD could also shed light on the nature of the WIMP-nuclear responses Fieguth:2018vob .

2.2 Neutrinoless Double Beta Decay

The lepton number violating neutrinoless double beta ($0\nu\beta\beta$) decay process is another critical signature for Beyond Standard Model physics. Its detection would reveal the nature of neutrinos, and under the assumption of light-neutrino exchange and Standard Model interactions, the half-life will yield insights into their mass ordering. With 8.9$\%$ abundance of ¹³⁶Xe in natural, non-enriched xenon, the observatory will instrument 5.3 tonnes of this isotope, aiming to observe $0\nu\beta\beta$-decays above the background spectrum shown on the left of Fig. 3. The corresponding half-life sensitivity projections for the $0\nu\beta\beta$-decay, derived using a Figure-of-Merit estimator Agostini:2017fom in corresponding optimal fiducial volumes, are shown on the right of Fig. 3 for the XLZD detector with either 60 or 80 tonnes target mass. The sensitivity is estimated for a range of detector performance parameters and background assumptions, as discussed below, represented by a band for each target mass between a nominal (lower bound) and an optimistic (upper bound) scenario.

Figure 3: Left: Energy spectra of a hypothetical $5\text{\times}{10}^{27}\text{\,}\mathrm{y}\mathrm{r}$ ¹³⁶Xe $0\nu\beta\beta$ signal (yellow) and the dominant backgrounds to this search: $\gamma$-rays from materials before (grey) and after (black) rejection by the veto systems (see text), $2\nu\beta\beta$-decays of ¹³⁶Xe (blue), ²²²Rn induced $\beta$-decays of ²¹⁴Bi (green), $\nu$-$e^{-}$ scattering of ⁸B neutrinos (purple), and $\beta$-decays of ¹³⁷Xe (red) for examples of different host laboratories. Right: Projected evidence sensitivity at $3\sigma$ significance to the $0\nu\beta\beta$ decay of ¹³⁶Xe as a function of exposure time for the two final target mass scenarios: $60\text{\,}\mathrm{t}$ and $80\text{\,}\mathrm{t}$, and an interim $40\text{\,}\mathrm{t}$ configuration further discussed in Section 3. For each target mass, the band represents the range of detector performance parameters and background assumptions between the nominal (lower limit) and optimistic (upper limit) scenarios, as discussed in the text. The right axis shows the sensitivity in effective Majorana mass $m_{\beta\beta},$ assuming a nuclear matrix element range of $M_{136\rm{Xe}}^{0\nu}\in[1.11,4.77]$ Agostini:2023 . The projected sensitivity of other proposed ¹³⁶Xe $0\nu\beta\beta$ experiments is shown for comparison Agostini:2023 ; nEXO:2021ujk ; NEXT:2020amj ; Chen:2016qcd .

The backgrounds to a $0\nu\beta\beta$-decay signal, shown on the left of Fig. 3, are caused by $\gamma$-rays emitted from detector materials and electron-induced signals, with the latter being uniformly distributed in the detector volume. Two of these uniform backgrounds are irreducible: the continuous spectrum induced by solar ⁸B neutrinos scattering off electrons and leakage of the $2\nu\beta\beta$ decay spectrum of ¹³⁶Xe into the $0\nu\beta\beta$ energy region of interest (ROI). The latter is highly suppressed due to the excellent energy resolution of $\sigma_{E}=0.65\%~{}Q_{\beta\beta}$ Pereira_2023 ; Aprile:2020yad (0.60% $Q_{\beta\beta}$) in the nominal (optimistic) scenario, which corresponds to a $Q_{\beta\beta}\pm 1\sigma$ ROI of $32\text{\,}\mathrm{k}\mathrm{e}\mathrm{V}$ ($29\text{\,}\mathrm{k}\mathrm{e}\mathrm{V}$) illustrated by the vertical yellow band on the left of Fig. 3. Additionally, two isotopes present in the target cause $\beta$-decays with energies extending over the $0\nu\beta\beta$ ROI: ²¹⁴Bi, a progeny of ²²²Rn, and ¹³⁷Xe, which is produced in the TPC by the capture of neutrons on ¹³⁶Xe. The rate of the latter scales with the muon flux and thus primarily depends on the depth of the host laboratory DARWIN:2023uje . We consider Laboratori Nazionali del Gran Sasso (LNGS), Italy (29.7 $\mu$ m^-2d^-1) and the Sanford Underground Research Facility (SURF), USA (4.6 $\mu$ m^-2d^-1) as representative of reasonable nominal and optimistic scenarios, respectively. A natural abundance xenon target limits ¹³⁷Xe production, and both facilities provide a sufficient muon flux reduction for it not to be a dominant background. Only ¹³⁷Xe production directly inside the TPC is considered an inevitable source of background: xenon activation outside the shielding water tank, e.g. during xenon purification, can be avoided by adequate shielding of the xenon handling infrastructure (discussed in Section 4.4). A low level of ²²²Rn in the target of $0.1\text{\,}\mu\mathrm{B}\mathrm{q}\mathrm{/}\mathrm{k}\mathrm{g}$ of natural xenon is a common requirement for the dark matter science goal Schumann:2015cpa ; Baudis:2013qla . Additionally, the tagging of the subsequent ²¹⁴Po $\alpha$-decay allows for vetoing ²¹⁴Bi decays with high efficiency, from 99.95% (nominal) to 99.99% (optimistic).

The projection of the external background in the nominal scenario assumes the material budget of a dimensionally-scaled LZ design and the radiopurity of materials already identified lz-xe136-sensitivity . In the optimistic scenario, a reduction of the external background by a factor of three is assumed.

Suppression of the ²⁰⁸Tl contribution in the external background (shown by the rate decrease from the gray to the black line in Fig. 3, left) requires an outer detector surrounding the cryostat, as well as instrumentation of the LXe region surrounding the TPC. The background from $\gamma$-rays emitted by radioactivity in the laboratory rock walls is assumed to be suppressed by a water shield ($3.5\text{\,}\mathrm{m}$ minimum thickness) to a negligible level compared to the detector material-induced $\gamma$-ray flux.

As shown on the right of Fig. 3, after an initial essentially background-free period, XLZD becomes background limited, with the uniform backgrounds (dominated by solar ⁸B neutrinos) and the external $\gamma$-ray backgrounds having similar contributions in the optimized fiducial volumes. Running the detector twice as long (20 years) would result in a 50% higher half-life sensitivity. The $0\nu\beta\beta$ science reach is primarily determined by the instrumented target mass, and within the sensitivity range bands, the reduction of the external background by improved material selection has the strongest impact on the achievable sensitivity. Partial enrichment in ¹³⁶Xe could be deployed at a future stage to confirm at 3$\sigma$ level a putative signal in XLZD or another experiment.

2.3 Astrophysical Neutrinos

The XLZD detector is a prime observatory for low-energy, MeV-scale, astrophysical neutrinos through nuclear and electronic recoil signatures of a few keV of energy. Primarily, several solar neutrino flux components can be measured. Current generation experiments have claimed first observations of solar neutrinos through coherent elastic neutrino-nucleus scattering (CE$\nu$NS) from ⁸B neutrinos Freedman:1973yd ; aprile2024measurementsolar8bneutrinos ; pandaxcollaboration2024indicationsolar8bneutrino . In such a channel, the XLZD detector expects an event rate of $\sim 90$ events per tonne-year above a nuclear recoil energy threshold of 1 $\rm keV_{nr}$ Baudis:2013qla ; Aalbers:2016jon , providing an independent measurement of the solar ⁸B neutrino flux. Similar to WIMP-nucleus scattering, neutrino-nucleus scattering can also be treated consistently in Effective Field Theories of the Standard Model Hoferichter:2020osn . In addition, by combining this measurement with neutrino-electron scattering data from other neutrino detectors, XLZD aims to constrain the $\nu_{e}$ survival probability in this energy range. The most critical background for this measurement comes from accidental coincidences (ACs), spurious events created by the incorrect pairing of detector signals, detailed in Section 5.4.

As for ER signals from solar neutrinos, XLZD aims to measure the pp solar neutrino spectrum via neutrino-electron scattering, improving the measurement of the neutrino luminosity of the Sun Aalbers:2020gsn ; PhysRevD.106.096017 . Furthermore, a high-statistics measurement of the solar pp neutrino flux in the XLZD detector will enable a direct measurement of the oscillation probability of the electron-type neutrinos emitted from the Sun in an energy range that is not accessible to any other experiment, as well as an independent measurement of the weak mixing angle, $\sin^{2}\left(\theta_{W}\right)$. Fig. 4 shows that with an exposure of 600 t$\cdot$y XLZD will constrain the low-energy survival probability to $\sim 5\%$. Such a measurement would test models of neutrino oscillations and probe exotic neutrino properties and non-standard interactions. The results in Fig. 4 were obtained considering a ²²²Rn concentration of $0.1\text{\,}\mu\mathrm{B}\mathrm{q}\mathrm{/}\mathrm{k}\mathrm{g}$, which is the main background for a pp-neutrino measurement. With the achievable exposure and energy resolution of the XLZD detector, it would be possible to conduct multiple measurements at uncertainties similar to the current best measurements but reaching to lower neutrino energies as illustrated in Fig. 4.

Figure 4: The $\nu_{e}$ survival probability versus neutrino energy, assuming the high-Z Solar Standard Model (SSM). The blue dots represent the solar measurements from Borexino Agostini:2018uly ; Agostini_2020 . The green (purple) point shows a measurement of ⁷Be (⁸B) from KamLAND (SNO) KamLAND:2011fld ; Aharmim:2011vm . The dark red dot indicates that XLZD, with 600 t$\cdot$y exposure, could enhance the precision of the $\nu_{e}$ survival probability in the low energy region to $\sim 5\%$, using solar pp neutrino events. The light red dots illustrate that multiple measurements could be conducted with XLZD with uncertainties of about $\sim 14\%$. The grey band represents the 1$\sigma$ prediction of the MSW-LMA solution Capozzi:2017 .

Galactic supernova neutrinos may also be detected in the XLZD detector, mainly through CE$\nu$NS. In contrast to other neutrino detectors, such detection is flavor-independent and allows the reconstruction of the total neutrino flux. A Type II core-collapse supernova at a distance of 10 kpc from the Earth would produce of $\mathcal{O}$(100) events in the detector within a 10 second window. As a result, XLZD can detect a supernova burst with 5$\sigma$ significant beyond the edge of the Milky Way and the Large Magellanic Cloud Lang:2016zhv . In this regard, XLZD will be able to actively contribute to the inter-experiment Supernova Early Warning System (SNEWS2.0) network Antonioli:2004zb ; Kharusi:2020ovw , responsible for the prompt follow-up of supernova events detection to the general astrophysics community. Unlike solar-neutrino detection, the signal of supernova neutrinos is highly concentrated in time ($\sim$ 10 seconds). The measurement will be limited by single and few-electron background rates, both sporadic and correlated with large events within the supernova burst duration.

2.4 Science Complementarity

Throughout the last decades, the search for dark matter has grown into an extensive interdisciplinary and international effort with three main thrusts: direct detection using ultra-sensitive detectors operating underground, production of dark matter at high-energy colliders, and indirect detection of particle fluxes caused by dark matter annihilation in the cosmos.

The science goals, as presented in the previous sections, complement the next-generation astrophysical searches like the Cherenkov Telescope Array (CTA), which will probe for a thermal WIMP annihilation gamma-ray signal in the $\sim$ 0.2-20 GeV/c² mass range Ong:2017ihp . When combined with lighter dark matter mass constraints (<200 GeV/c²) from the currently operating Fermi-LAT FermiLAT_2015 , CTA results could severely constrain the WIMP parameter space. Water Cherenkov detectors, like the future SWGO, complement CTA due to a wider field-of-view abreu2019_swgo .

The production of dark matter in the electroweak energy scale will be tested at the high-luminosity LHC Arduini:2016xsb . Several generations of LXe-based WIMP-search experiments have successfully demonstrated how this interplay of experimental approaches can work by ruling out new dark matter candidates in the GeV/c² to TeV/c² mass range beyond the reach of the LHC in mass or production cross-section LZ_2024_DM_results ; XENON:2023sxq ; PandaX-4T:2024dm ; Aprile:2018dbl ; Akerib:2015rjg . XLZD will continue this tradition and explore WIMP-nucleus scattering in the remaining well-motivated regions of parameter space down to the neutrino fog.

Besides XLZD, the liquid argon (LAr) community is proposing a future dark matter detector, ARGO Agnes_2021 , to also explore the WIMP cross-section space to the neutrino fog. Probing this phase space using both targets, LXe and LAr, would allow exploring degeneracies, as the comparison of recoil measurements performed on different mass nuclei allows determination of dark matter properties such as the dark matter mass, density, and velocity distribution Pato:2010zk ; Bozorgnia_2018 . In order to probe the same parameter space as XLZD, an LAr experiment will require significantly higher exposure due to the lower atomic mass and the higher energy threshold of LAr detectors, which would be an upscale by two to three orders of magnitude in target mass over existing detectors depending on the architecture that is chosen. XLZD and ARGO are also complementary in the sense that spin-dependent interactions can only be probed by LXe detectors due to the $\sim 50\%$ abundance of xenon isotopes with unpaired spins and that only LXe detectors can effectively search for new physics in low-energy electronics recoils Chepel:2012sj .

Neutrinoless double beta decay can occur in different isotopes, with different $Q$-values, natural abundances, and material properties. This results in different detection techniques and experimental challenges. Using a dual-phase (liquid/gas) TPC with natural xenon XLZD will be complementary to other experiments searching for $0\nu\beta\beta$ decay in ¹³⁶Xe with different technologies, such as liquid-xenon-only TPCs with an enriched xenon target (nEXO nEXO:2021ujk ), high-pressure gaseous TPCs (NEXT NEXT:2020amj ) or ¹³⁶Xe-loaded liquid scintillator (KamLAND-Zen Nakamura:2020szx ). Conversion from the measured decay half-life to the physically relevant Majorana mass ($m_{\beta\beta}$) requires knowledge of the nuclear matrix element (NME) for each isotope. NME predictions provided by different nuclear models can vary by a factor of more than four Agostini:2023 , resulting in significant uncertainties in $m_{\beta\beta}$ measurements. Complementarity between experiments using different isotopes (e.g., LEGEND legendcollaboration2021legend1000 using high-purity Ge detectors to search for $0\nu\beta\beta$ in ⁷⁶Ge) is therefore paramount to reduce NME uncertainties when covering the neutrino mass ordering phase space.

With its sensitivity to solar neutrinos, XLZD will provide complementary measurements to dedicated neutrino efforts Abdullah:2022zue . Current and future CE$\nu$NS experiments either use stopped-pion beams as a source of electron, muon, and antimuon neutrinos or nuclear reactors as sources of electron antineutrinos Akimov:2017ade ; Akimov:2020pdx ; Angloher:2019 . These experiments measure the CE$\nu$NS cross-section for energies of $\mathcal{O}$(10) MeV (stopped-pion beam) and $\mathcal{O}$(1) MeV (reactor). Solar neutrinos causing CE$\nu$NS in XLZD provide a unique source as the flux contains all three flavour components allowing XLZD to probe for flavour dependence beyond the tree level. The accessible energy range spans approximately 1-10 MeV, bridging the gap in energy between the other experiments Strigari:2023 . The current generation of detectors has just claimed the first measurements of CE$\nu$NS from ⁸B solar neutrinos aprile2024measurementsolar8bneutrinos ; pandaxcollaboration2024indicationsolar8bneutrino . XLZD is also sensitive to electron-neutrino scattering at lower energies than typically detected at neutrino observatories KamLAND:2011fld ; Smirnov:2015lxy ; Liao:2017awz and, therefore, will also provide complementary measurements in this interaction channel.

3 The XLZD Project

The XLZD project builds upon the legacy of several successful generations of liquid xenon dark matter searches. It brings together experts from the leading current-generation detectors and R&D efforts. In this section, the experimental strategy based on this mature detector technology is introduced.

3.1 Technical Heritage of Liquid Xenon TPCs

LXe-TPCs take advantage of several excellent properties of xenon Chepel:2012sj ; Aprile:2008bga . Its high atomic mass enhances the scattering cross-section for the simplest elastic interactions of WIMPs. Natural xenon is free from problematic long-lived radioisotopes, and its dense liquid phase offers excellent properties for particle detection and for realizing extremely low backgrounds. The detector technology exploits two signatures to reconstruct the location, energy, interaction multiplicity, and type of particle interactions: a prompt primary scintillation signal in the vacuum ultraviolet (VUV) at 175 nm and delayed electroluminescence, also in VUV, derived from the ionization electrons released at the interaction site, see Fig. 5. In the so-called ‘dual-phase’ detector configuration, a thin layer of xenon vapor tops the liquid phase that constitutes the target for particle interactions. The ionization electrons from the event site drift upwards through the liquid in an applied electric field, which is established between transparent electrodes. Once emitted into the gas region, these electrons produce electroluminescence light. Each interaction in the active region thus generates one primary (S1) and one secondary (S2) scintillation signal, both of which can be detected by photosensors above and below the TPC. Combining information from these two signatures allows the accurate reconstruction of the interaction site and the discrimination between electronic and nuclear recoil interactions. Owing to the large scintillation and ionization yields of liquid xenon Doke:2002oab ; Akerib:2016qlr ; Aprile:2017xxh , thresholds of order keV can be realized for both electronic and nuclear recoils XENON:2023sxq ; Aalbers:2022dzr .

Figure 5: Operating principle of the dual-phase (liquid/gas) xenon time projection chamber. Particle interactions in the active liquid xenon target produce prompt (S1) and delayed (S2) optical signals. These are detected by two arrays of VUV-sensitive photosensors located at the top and bottom of the TPC.

Combining the ability to spatially resolve interactions with the high density and high atomic number of the medium enables the self-shielding properties of liquid xenon. An inner ‘fiducial’ volume can be defined, which is largely free from external radiation. In the larger detectors operating now, the main remaining experimental backgrounds are thus from radioisotopes dispersed in the liquid bulk (²²²Rn, ⁸⁵Kr), and various techniques have been developed to effectively remove them (see Section 5). Finally, the ratio of S2 to S1 can be used to distinguish nuclear recoils from electronic recoils Akerib:2020lkv ; Aprile:2017xxh . Electronic recoil rejection levels of 99.9% at 50% nuclear recoil acceptance have been achieved right down to the detection threshold (see a summary of electron recoil background rejection in Szydagis:2022ikv ; Araujo:2020rwg ).

Many generations of such LXe-TPCs have successfully conducted searches for dark matter since the early 2000s through the ZEPLIN Alner:2007ja ; Lebedenko:2008gb , XENON Angle:2007uj ; Aprile:2016swn ; Aprile:2017iyp ; XENON:2023sxq , LUX/LZ LZ:2019sgr ; Aalbers:2022dzr and Panda-X Xiao:2014xyn ; Wang:2020coa ; PandaX-4T:2021bab programs, leading to three currently running multi-tonne LXe-TPCs (PandaX-4T, XENONnT and LZ). The PandaX collaboration plans to expand to $43\text{\,}\mathrm{t}$ active mass in a staged-approach experiment at China Jinping Underground Laboratory (CJPL) PandaX:2024oxq .

Since 2007, the direct search for dark matter of masses above a few GeV has been led by LXe-TPCs; while the target mass increased by over a factor of 1000, the achieved background levels decreased by four orders of magnitude. The competitive and incremental nature of these projects has been crucial to their success and places this technology in a unique position to probe the remaining parameter space well ahead of any other. Technical risks do exist – as is natural for such an ambitious experiment as planned by the XLZD collaboration – but they are now well understood as discussed in Section 6. Furthermore, the technical track record of the new collaboration, which operates two of the world-leading experiments of the day (shown in Fig. 6), LZ and XENONnT, is critical to the experiment’s success.

Figure 6: The LZ (left) and XENONnT (right) LXe-TPCs containing 7-tonnes and 5.9-tonnes of liquid xenon in the active detector volume, respectively. Both TPCs have a drift length of around 1.5 m, while the diameter of the sensitive target is 1.5 m in LZ and 1.3 m in XENONnT.

The LZ collaboration is operating a TPC with 7 tonnes of LXe in its active region (10 tonnes LXe in the full detector) LZ:2019sgr at the Sanford Underground Research Facility (SURF), South Dakota, USA. The detector features a segmented-polytetrafluoroethylene (PTFE) field cage capped by woven electrode grids, with optical readout by two arrays of 3-inch Hamamatsu R11410-22 photomultiplier tubes (PMTs). The LXe-TPC is housed in a titanium cryostat. The thin layer of LXe ($\sim$2 tonnes) between the TPC and the inner cryostat vessel, often referred to as LXe Skin, is instrumented with scintillation readout to function as a veto detector. The cryostat is surrounded by an Outer Detector containing 17 tonnes of Gd-loaded liquid scintillator in acrylic vessels, viewed by 120 8-inch PMTs. Both detector systems are immersed in a water tank with a 7.6 m diameter. Xenon cooling and purification are done separately; cold LXe is fed to a pipe manifold at the bottom of the detector; liquid is extracted from a weir system at the top and converted to gas for purification. Krypton removal was conducted using gas chromatography before deployment. LZ published world-leading results from its first science run in 2022 LZ:2022ufs .

The XENONnT collaboration is operating a TPC with 5.9-tonnes of LXe in the active region (8.5 tonnes LXe in the full detector) XENON:2024wpa at the INFN Laboratori Nazionali del Gran Sasso (LNGS), Italy. It is the upgrade of XENON1T Aprile:2017aty , the first LXe-TPC with a target above the tonne scale, which was operated at LNGS from 2016-2018. XENONnT features a lightweight TPC made of thin PTFE walls, two concentric sets of field-shaping electrodes, and high-transparency electrode grids made of individual parallel wires. Two arrays of Hamamatsu R11410-21 PMTs provide the optical readout. Although the LXe Skin concept was pioneered by XENON100 Aprile:2011dd , it was not installed in XENON1T/nT to minimize backgrounds due to radon emanation from PTFE and maximize the active target. The TPC is housed in a stainless steel cryostat, placed in the center of a 9.6 m diameter water shield operated as Cherenkov muon veto. The neutron veto has an inner volume of 33 m³ around the cryostat, defined by highly-reflective PTFE walls and instrumented with 120 8-inch PMTs: after the first runs with demineralized water, Gd has been added to the shield to increase the neutron tagging efficiency. A diving bell controls the LXe level inside the TPC. For purification, LXe is extracted from the bottom of the cryostat and efficiently purified in the liquid phase Plante:2022 ; an additional gas purification system cleans the warmer gas phase. Krypton removal is done via a cryogenic distillation column XENON:2021fkt installed on-site, allowing for online distillation. A second cryogenic distillation system constantly removes radon atoms from the liquid and gaseous xenon target Murra:2022mlr . XENONnT published results for a search for new physics with a world-leading electronic recoil background level in 2022 XENON:2022ltv and for a first WIMP search in 2023 XENON:2023sxq .

Although the two experiments may appear similar, very different implementations have been adopted for most subsystems, with some differences highlighted above. These proven alternative implementations, already demonstrated at the multi-tonne scale in world-leading dark matter detectors, constitute a powerful tool for risk management. In each case, there are two solutions to choose from, and their performance is thoroughly evaluated in real dark matter search conditions. Adding this diversity of options to the long track record of this technology, it may be argued that the next step to the XLZD detector entails only modest technical risk – and although the proposed $\sim$10-fold mass scale-up is significant, the increase in linear dimensions is relatively modest (factor $\sim$2).

Figure 7: The XLZD nominal system features a LXe-TPC with a 1:1 aspect ratio for 60 tonnes of active mass (2.98 m in diameter and height) housed in a double-walled cryostat.

3.2 XLZD Detector: Strategy and Xenon Acquisition

The nominal XLZD detector features a dual-phase LXe-TPC with around 3 m inner diameter and height, containing 60 tonnes of active mass as depicted in Fig. 7. Such a detector enables dark matter searches down to the neutrino fog and a competitive search for $0\nu\beta\beta$ decay in ¹³⁶Xe using a natural abundance target (see Section 2 and a forthcoming sensitivity study paper). While the XLZD TPC diameter is fixed early in the detector design, adjusting the field cage height will allow for an interim detector of reduced mass (up to 40 tonnes) and the flexibility to achieve a larger target mass of 80 tonnes for the final detector. This strategy provides a central element of risk and opportunity management, allowing for early testing of detector components and adaptability to the xenon market.

The interim detector configuration of up to 40 tonnes at full field cage width and reduced height will be used for initial technical performance verification and risk mitigation. It also has the potential for competitive early science with a dark matter search exposure of up to 1000 days, leading to more than five times the exposure expected for current-generation detectors. The $0\nu\beta\beta$-decay sensitivity for a 40 tonnes detector running for up to 3 years is illustrated in Fig. 3. The interim configuration will also maximize the time XLZD can monitor for a supernova burst with competitive sensitivity Lang:2016zhv and enable an early measurement of the solar neutrino flux via CE$\nu$NS and neutrino-electron scattering Aalbers:2020gsn .

Simultaneously, we will continue the xenon acquisition program until we acquire sufficient xenon to deploy the full nominal height with 60 tonnes of active mass. At this point, all critical detector systems have been thoroughly tested in the running interim detector, and problems encountered can be addressed or components replaced. The 60 tonnes baseline detector delivers the full science case outlined in Section 2.

Should the xenon market conditions permit a higher acquisition rate after the initial period, or if our initial phase already sees a hint of a signal, a more ambitious instrument with 80 tonnes of active mass can be accommodated to maximize exposure. The larger target mass would require a taller field cage keeping the same diameter as the baseline scenario. It would also approach the best discovery sensitivity for $0\nu\beta\beta$ decay searches since, at this target mass, we benefit more significantly from the self-shielding of external backgrounds from 2.5 MeV $\gamma$-rays.

The outlined experimental strategy builds on a realistic xenon acquisition schedule based on conversations with the global rare-gas suppliers and the collaboration’s experience in successfully procuring large amounts of xenon (20 t) for currently running experiments. The strategy allows for xenon procurement to spread over several years as not to disrupt the market. In-advance purchase planning, multi-year contracts, and early availability of funding will be essential for a successful xenon acquisition campaign. It should also be noted that the xenon will be retained over the experiment’s lifetime and will remain an asset that does not deteriorate.

3.3 Siting

A suitable underground facility will be required to host this flagship observatory, featuring key characteristics (e.g. depth, space, accessibility, services, support) appropriate for such an ambitious project. Five underground laboratories located at 1,000 m.w.e. depth or more have expressed interest in hosting the detector planned by the XLZD collaboration; these are located either in deep mines or under mountains, and all offer atmospheric cosmic-ray muon fluxes attenuated by a factor of at least one million relative to the flux on the surface. These are the Boulby Underground Laboratory in the UK, the INFN Laboratori Nazionali del Gran Sasso in Italy, the Kamioka Observatory in Japan, the Sanford Underground Research Facility in the USA, and SNOLAB in Canada. All have experience hosting world-class dark matter and/or $0\nu\beta\beta$ decay research.

In addition to the all-important depth requirement, other parameters must be considered, including the availability of suitable amount and quality of spaces underground and on the surface, the schedule for occupation, accessibility for people and materials, the radiation and cleanliness environments underground, and availability of power and other services.

The XLZD collaboration is evaluating hosting options thoroughly, working towards an inclusive down-selection of sites that meet key physical and infrastructure requirements, i.e. those laboratories where such an experiment can deliver its science mission fully. The shortlisted sites either meet these requirements now or will meet them with proposed developments. An independent study by the DARWIN collaboration has explored the cosmogenic backgrounds at potential host sites DARWIN:2023uje .

3.4 The XLZD Collaboration

The XLZD collaboration was formed in September 2024 by a Memorandum of Understanding signed by over 100 senior scientists of the XENON, LUX-ZEPLIN, and DARWIN collaborations after working together as a consortium since June 2021. This is a global collaboration of scientists and engineers in 17 countries who have committed to working together to develop the definitive rare event observatory based on the world-leading liquid xenon technology. This endeavor benefits from two decades of expertise and a track record of the world’s prominent collaborations now operating experiments at the 10-tonne scale, as well as from the DARWIN team, which has been working on design studies for a next-generation experiment.

We anticipate that a construction project would receive substantial support from the major international funding agencies in the U.S.A., Europe (including the U.K. and Switzerland), Japan, and others. The science reach of this LXe observatory is very timely and has high priority in many major national and international roadmaps, and the collaboration member institutions have the track record required to deliver that science. Coordinated proposals will be submitted to fund the experiment’s design, construction, and operation.

4 Experimental Architecture

Operating for up to 1000 t$\cdot$y exposure requires the largest liquid xenon target yet and the ability to run stably for over a decade. Radioactive and electronegative impurities in materials must be minimized and controlled while maintaining adequate structural support and optimizing scintillation detection. Backgrounds from cosmic ray muon-induced neutrons must be mitigated to the level of one event over the entire exposure. The TPC design, mitigation of fast neutrons, detection infrastructure such as electric fields and photosensors, materials selection, and outer detectors for coincidence vetoes are vital to this goal. This section describes the components of the XLZD observatory that are needed to meet these stringent requirements.

4.1 The Xenon Time Projection Chamber

The primary science detector is the cylindrical TPC (similar to those illustrated in Fig. 6), containing the liquid xenon target. The TPC design needs to ensure high detection efficiency of VUV photons by choosing high reflectivity materials such as PTFE and efficient photosensors (see 4.1.2) and extraction of ionization electrons to the top of the detector using a homogeneous drift field established by field-shaping electrodes (see 4.1.1) connected by high-value ohmic resistors. To maintain thermal equilibrium between the liquid and gas phase (in which ionization electrons create electroluminescence), the TPC will be housed in a vacuum-insulated cryostat. Between the walls of the TPC and the cryostat, a thin surrounding layer of liquid xenon can be instrumented to veto gamma-ray backgrounds. This is further discussed in Sec. 4.3. Table 1 lists the nominal TPC dimensions and xenon requirements, along with a more ambitious detector if the opportunity arises. The required total mass of xenon assumes a 10 cm charge-insensitive reverse field region below the cathode to protect the bottom photosensor array, a LXe Skin of 7 cm between the TPC walls and the cryostat, and an extra 30 cm below the photocathodes of the bottom photosensor array.

Table 1: The liquid xenon active target dimensions and mass for the envisaged TPC. The total mass of liquid xenon includes that in the Skin veto and other volumes, as discussed in the text.

TPC parameters	Nominal	Opportunity
Target diameter [cm]	298	298
Target drift length [cm]	297	396
Target Mass [tonne]	60	80
Total Mass [tonne]	78	104

Figure 8: A finished LZ grid with crossed wires on the custom loom (left) and a XENONnT electrode with parallel wires (right).

4.1.1 Electrodes and High Voltage

A dual-phase LXe-TPC contains three electrodes that create the drift (cathode to gate) and extraction (gate to anode) regions, typically consisting of parallel wires or meshes mounted on ring frames. Two additional electrodes can be placed to serve as electrical protection for the top and bottom photosensor arrays, which is the design found in the XENONnT TPC. In the LZ TPC, only the bottom shielding electrode was placed, and a shield for the top array was omitted to optimize the electric field in the region around the anode.

Possible electrode designs are parallel wires, crossed wires, and etched sheets of hexagonal elements. While the XENONnT and LZ TPCs incorporated the first two designs, respectively, hexagonal meshes were disfavored for the current detector generation due to the technical challenge of their manufacturing scalability. Ongoing R&D aims to overcome this challenge by welding smaller sections Andrieu:2023buk . The crossed-wire grids in the LZ TPC were produced by a custom-built loom, as shown in Fig. 8, which provided uniform tension over wires. The materials for these meshes were selected after screening campaigns, such as annealed SS316 (XENONnT) and SS304 (LZ), and they have been shown to perform more stably after electropolishing and passivation of their surfaces Tomas:2018pny ; Linehan:2021qnb .

The drift field strength is usually selected to enable efficient discrimination between nuclear and electronic recoils and short drift times to avoid the pile-up of interactions. Values between 240 and $290\text{\,}\mathrm{V}\mathrm{/}\mathrm{c}\mathrm{m}$ for the drift field are reported to optimize the discrimination between the two types of recoil Akerib:2020lkv . The gas extraction field is typically held around $\mathrm{6}\mathrm{-}\mathrm{8}\,\mathrm{k}\mathrm{V}\mathrm{/}\mathrm{c}\mathrm{m}$ to ensure a large number of extracted electrons and secondary photons for energy and position reconstruction.

The drift fields in meter-scale TPCs require the delivery of high voltage, of $\mathcal{O}(10-100)\textrm{kV}$, to the cathode. The high radioactivity in commercially available feedthroughs requires custom-made solutions for LXe TPCs Aprile:2011dd . The high voltage can be fed through from the top, side, or bottom of the TPCs, with different advantages and caveats. The routing of the high voltage conductor from the top, traversing the LXe volume, requires the insulation of the conductor with PTFE or Ultra-High Molecular Weight Polyethylene (UHMWP), which is cryofitted in between the conductor and a grounded conductive cylinder, as is the chosen design in the XENONnT TPC XENON:2024wpa . As the HV conductor increases the electric field in its vicinity, clearance with respect to the field cage is required. As LZ uses the space in between the TPC and cryostat as Skin veto, the high voltage is delivered to the cathode from the side of the TPC. Extensive R&D was conducted to find a mechanical and electrical robust solution LZ:2015kxe . A high-voltage cable is routed through the vacuum, which allows the use of a commercial cable connected at its end to a xenon-filled umbilical at room temperature outside the water shield. The lower part of the umbilical is filled with LXe. It accommodates a field grading structure that allows for the ground braid of the cable to terminate while the insulation and conductive center of the cable continue, which connects through a compliant spring to the cathode grid ring.

4.1.2 TPC Photosensor Arrays

The detector baseline design will employ two arrays of low-radioactivity PMTs, similar to the 3-inch diameter PMTs (Hamamatsu R11410-21/22) presently in operation in LZ and XENONnT and shown in Fig. 9. A total of 1182 3-inch PMTs would be needed in the top and bottom PMT arrays, respectively, to ensure the required light collection efficiency, which drives the detector’s low energy threshold. The top PMT positions will be selected to optimize the position reconstruction using the S2 signal. The currently used PMTs have an average quantum efficiency of 34% at 20^∘C and at a wavelength of 175 nm, with an average internal photoelectron collection efficiency of 90%. The quantum efficiency has been measured to show a relative improvement of up to  18% upon cooling to LXe temperatures LopezParedes:2018kzu .

The VUV-sensitive photomultipliers were jointly developed by Hamamatsu and the two collaborations for operation in liquid xenon. They were optimized for low radioactivity, low spurious light emission, and vacuum tightness in cryogenic conditions. All units currently in operation ($494\times 2$) were characterized at low temperatures LopezParedes:2018kzu ; Baudis:2013xva ; Barrow_2017 ; Antochi:2021wik in various test facilities before being installed in the LZ and XENONnT TPCs, respectively. The failure rate after LZ and XENONnT commissioning and subsequent two years of operation is less than 3%, which is consistent with the continued good performance of the TPC detector systems.

Figure 9: The LZ bottom (left) and XENONnT top (right) PMT arrays. Each experiment has 494 R11410 Hamamatsu PMTs installed.

The photosensors will be equipped with resistive voltage divider circuits (bases). The bases must comply with low radioactivity, high linearity (for the $0\nu\beta\beta$ search) and low power consumption requirements. The expected gains are around $3\times 10^{6}$ at a bias voltage of 1500 V, with a dark count rate of $\sim$12-24 Hz/PMT at liquid xenon temperatures Aprile:2017aty . Significant PMT channel wiring savings may be possible and are under study.

The support structures for the two arrays will be made from low-radioactivity copper or titanium, covered by a thin layer of PTFE to maximize VUV reflectance. The design will ensure that the mechanical stress induced by the thermal contraction of the different materials will not significantly affect the PMTs. The high-voltage and signal cables may use low-radioactivity coaxial cables deployed in XENONnT or LZ. The development of other types of transmission lines, such as striplines, may possibly further reduce the Rn emanation from the signal cables.

The materials used in the PMT variants employed in LZ and XENONnT were selected after extensive radio-assay campaigns. Specific activities for ²³⁸U and ²³²Th are $<13.3$ mBq/PMT and $<0.6$ mBq/PMT, respectively  XENON:2015ara ; Akerib:2020com . Further optimization of the PMT materials is required to meet the background requirements. Current efforts include screening of different iterations of the stem materials (low-activity glass stems, refined metal stems), with promising results, considering that the ceramic stems in the 21/22 model are dominating their U/Th/K content XENON:2015ara . Alternative photosensors are being evaluated based on radioactivity, dark counts, and quantum efficiency requirements.

4.2 Cryostat

The cryostat will be a double-walled structure centrally placed in the water tank. The inner vessel holding the TPC and liquid xenon will be nested in the outer vessel, which provides vacuum insulation and is held by the cryostat support.

As in XENONnT Aprile:2020vtw and LZ Mount:2017qzi , both vessels will have conventional cylindrical geometry with shallow torispherical heads. Alternatively, a flat floor cryostat would help to reduce the amount of xenon needed. R&D into such designs is ongoing Brown_2024 . To obtain a minimum required thickness of the walls, both vessels will be split by a pair of flanges serving as stiffening rings against buckling and providing the modularity necessary for a possible expansion of the target mass and the size of the detector. In addition, the shape of the inner vessel will be optimized to minimize passive volumes filled with LXe. The vessels’ design will comply with the pressure code the hosting country requires.

Due to its mass and direct contact with xenon, the cryostat is one of the main contributors to the experiment’s radioactivity budget. Therefore, a critical requirement for the cryostat is to limit its contributions from the bulk of the material and its surface to radioactive backgrounds in the sensitive region of the xenon detector. The baseline design for the cryostat assumes the use of commercially pure CP-1-grade titanium obtained from the cold hearth electron beam refining process, which follows a successful material search campaign by the LUX Akerib:2011rr and LZ LZ:2017iwn experiments.

To minimize the impact of radon-emitting surfaces in direct contact with xenon, they will be either electropolished or chemically etched, depending on the vessel’s material and the process’s suitability. A novel radon mitigation method showed that a suppression of the ²²²Rn emanation rate by up to three orders of magnitude could be achieved using micrometer-thick surface coatings Jorg:2022spz . Further details about the applicability of this technique are discussed in Section 5.2.

Alternative welding techniques, such as local vacuum electron beam welding, are being sought to mitigate possible radioactive contamination due to TIG welding with electrodes containing traces of radioactive elements.

Figure 10: Left: The LZ outer detector with acrylic tanks monitored by R5912 PMTs in the water tank. Right: The XENONnT water tank is instrumented as muon veto. In the center, the outside of the inner neutron veto is seen.

4.3 Veto Detectors

To enhance XLZD’s background rejection, the central LXe-TPC will be surrounded by additional veto detectors, such as an instrumented layer of LXe Skin, and dedicated neutron and muon detectors. When run in anti-coincidence with the TPC, veto detectors allow the rejection of neutron and $\gamma$-ray backgrounds, which may mimic a dark matter signal. The additional veto detectors permit the separation of backgrounds based upon timing: signals from $\gamma$-ray background and proton recoils from neutrons will occur within tens of nanoseconds of the TPC S1, while signals from neutron captures will be delayed by typically tens of microseconds.

As background events from materials primarily populate the outer regions of the LXe volume, an efficient veto system increases the fraction of LXe which can be fiducialized for the dark matter search. Veto detectors additionally provide a means for in-situ characterization of the experiment’s background environment. In particular, this information would demonstrate that an unexpected flux of background neutrons does not induce a putative dark matter signal. The efficient tagging of high- and low-energy $\gamma$’s suppresses backgrounds critical for other new physics searches as well, e.g., $0\nu\beta\beta$ of ¹³⁶Xe.

A LXe Skin veto is achieved by instrumenting the layer of xenon between the TPC field cage and the cryostat with PMTs. PTFE lining of the outside of the field cage and the inside of the cryostat increases the light collection efficiency in this region and makes it an effective gamma veto. For XLZD, the benefits of a potential Skin veto will be evaluated against radon emanation considerations and the design of HV delivery in this region.

The outermost veto systems for LZ and XENONnT are shown in Fig. 10. LZ has a passive water shield on the outside, followed by PMTs monitoring both another layer of water and liquid scintillator (LS) tanks for neutron detection. XENONnT has an instrumented water tank to serve as an active water shield, while it has an optically insulated inner neutron veto. Both systems veto energetic events, including muons and showers caused by muons in the cavern walls surrounding the detector, with their instrumented water layers. The optimum configuration for XLZD will be designed by choosing elements of the distinct LZ and XENONnT solutions.

Fast neutrons from detector material impurities are dangerous for nuclear recoil dark matter searches. This background can be significantly suppressed and characterized by surrounding the LXe volume with a neutron detector. Crucial design requirements for a high ($\gtrsim$95%) neutron tagging efficiency include $\sim 4\pi$ detector coverage around the LXe-TPC and minimizing inactive material between the LXe and the neutron-sensitive medium. The rate in the neutron detector must also be kept low to minimize losses in the dark-matter-search livetime, which drives radiopurity requirements for all materials in the outer portions of the experiment.

Hydrogen-rich liquids such as water or organic liquid scintillator are an elegant solution. They can be made very radiopure and doped with elements with a high neutron capture cross-section, such as boron or gadolinium. In the LZ experiment, gadolinium is dissolved into a LS with a high light yield, which detects energy deposited by the 3–5 $\gamma$-rays emitted (on average) after neutron capture on ^155,157Gd or the single 2.2 MeV $\gamma$ from capture on hydrogen. The high LS light yield permits a threshold as low as 30 keV for prompt vetos, such as proton recoils and $\gamma$-rays, and a higher threshold adapted for acceptable dead time, which is currently 200 keV in LZ.

The XENONnT experiment instead dissolves gadolinium directly into the water of the shield, relying on the Cherenkov light emitted from above-threshold (289 keV) electrons created by $\gamma$-rays from neutron capture on gadolinium and hydrogen. Although the energy threshold of this configuration is higher, it allows for hermetic coverage around the central cryostat without the need for additional liquid vessels.

An alternative option that bridges these two media is a water-based liquid scintillator (WbLS), where a scintillator is mixed with water via the addition of a surfactant “tail” to the liquid scintillator molecule Yeh:2011 . A lower threshold in energy is possible with WbLS than with water, and the threshold can be adjusted by varying the concentration of the scintillator. A spill of WbLS would cause less environmental upset than a spill of pure scintillator. In LZ, the liquid scintillator concentration in water has been required to be $<0.02\%$ by mass to permit water disposal in the SURF drainage system. For XLZD and its eventual site, the allowable concentration and disposal criteria for WbLS would be established. WbLS can be formulated underground through integration into the water circulation and purification system. However, careful evaluation and management of the background sources that can provide a rate of pulses above WbLS threshold will be required to address the dead time caused by accidental vetos. The use of WbLS or Gd-WbLS currently also requires a more complicated filtration and purification system than the other options, but there is ongoing research into simplifying this process. Near-future tests at the tens of tonnes scales with a demonstrator at Brookhaven National Lab Xiang:2024jfp , the BUTTON experiment at Boulby and EOS at LBNL Anderson:2022lbb will provide essential information and feedback on the use of WbLS.

Figure 11: The ReStoX I (left, 7.6 tonne capacity) and ReStoX II (right, 10 tonne capacity) on-site cryogenic storage systems for XENONnT have demonstrated high-purity storage of xenon gas and liquid XENON:2024wpa .

4.4 Xenon Handling, Storage and Recuperation

The xenon inventory will be delivered to the experimental site in standard high-pressure gas cylinders designed for transport. When the xenon arrives at the experimental site, it will be transferred to a small number of larger monolithic storage vessels; otherwise, an unmanageable number (more than 1000) of gas cylinders will be required to contain the xenon.

The on-site storage system will be a vital component of the xenon handling system, and a suitable design will optimize various factors, including size, cleanliness, safety, and ease of use. The storage vessels will be cryogenic and insulated such that xenon can be loaded via cryo-pumping. The vessels must also be rated for high pressure so that they can return to room temperature, if necessary, while loaded with xenon. The XENONnT experiment has demonstrated safe cryogenic storage for 10 tonnes of xenon with the ReStoX system, shown in Fig. 11. ReStoX is considered a prototype for the XLZD storage system. The size of the final storage vessels will be impacted by site access (horizontal or vertical), fabrication considerations such as the availability of underground welding and cleaning capabilities, and the dimensions of the underground space.

A simple one-pass removal of electronegative impurities will be done during the initial loading of the Xe into the storage vessels during xenon procurement activities. This pre-purification step, which will be the first in the purification campaign of the experiment, can be done in the gas phase without incurring any phase change energy cost. Conventional and inexpensive purification technologies may be suitable for this treatment, even those that emanate radon, such as Oxysorb Altenmuller_2021 .

Online purification of electronegative impurities during detector operation will be required to achieve an electron lifetime of better than 10 ms, largely exceeding the expected electron drift time from the bottom of the detector to the liquid surface. Purifying by the non-evaporate-getter (St707), the LZ experiment has already achieved up to 8 ms electron lifetime with recirculating gas. XENONnT has operated the liquid recirculation system shown in Fig. 12 and performed purification with 2 LPM (8.3 tonnes/day) in the liquid phase, achieving an electron lifetime better than 10 ms. The system can be scaled up as needed.

The storage system will be designed to be connected to the complete purification system, including one or more Kr and Ar removal systems as discussed in Section 5.2, allowing for the commissioning of the purification systems. The storage system will have the functionality to continuously circulate the xenon inventory in a closed loop for purification purposes as part of the purification campaign. Both the storage and the purification systems will be built and commissioned early so the xenon can be purified during the xenon procurement phase.

The system to recover xenon from the cryostat into the storage vessel must be robust to protect the xenon investment with multiple layers of safeguards. The XENONnT experiment has demonstrated the safe transfer of liquid xenon driven by gravity alone. The ReStoX storage vessel is located at a lower elevation than the XENONnT cryostat to facilitate this recovery mode. Gravity-driven recovery is attractive because it does not require mechanical pumps or energy for the phase transition. The storage system will be designed to accommodate this type of recovery if the underground site architecture allows it.

Gaseous recovery has been used in many experiments and may also play an important role in the XLZD recovery system. The LZ experiment developed high-purity, high-pressure gas compressors for gaseous circulation and recovery (Fig. 12). These are required to transfer the xenon inventory back into transport cylinders at the conclusion of the experiment.

4.5 Cryogenics

The primary purpose of the cryogenic systems is to provide sufficient cooling power to liquefy gaseous xenon and maintain xenon detector thermodynamics by offsetting the intrinsic heat loads of the system. To date, xenon-based experiments have used bulk liquid nitrogen or refrigerators to provide the cooling power and specialized arrangements (e.g. nitrogen-based thermosyphon systems or a cold finger) to deliver targeted cooling LZ:2019sgr ; XENON:2024wpa ; Aprile:2017aty ; LUX:2012kmp . Cryocoolers (or cryogenerators) are being used by both LZ and XENON to ultimately reject heat from the experiment.

The same systems can be reasonably scaled up for a larger xenon payload. Once other design choices are finalized, a careful engineering analysis will determine the exact specifications for the cryogenic systems. Details of the xenon circulation and purification systems will have the biggest impact on the requirements for this system. The xenon purity goals (see Section 4.4) will require faster xenon re-circulation in both the gas and liquid phases and, therefore, greater cooling power to overcome inefficiencies in heat-exchange between xenon gas and liquid and to sub-cool the liquid xenon below saturation temperatures. Other components of the system will also need to be accounted for: radiative heat losses from the xenon vessel, heat from the PMT arrays, heat influx through cables and access ports, increased xenon purge flow through cable conduits, etc. A total operating heat budget of a 100 t xenon system is estimated to be comfortably within the range of commercially available cryogenerator units. An additional consideration will be the cooling power required to initially condense the xenon payload on a reasonable timescale. For example, condensing 78 t of xenon, the total required xenon inventory for the 60 t detector, in 1 month of constant operation, will require $\sim$ 6.1 kW of concentrated cooling. While cryogenerators can easily provide this power, the delivery mechanisms and cold-heads need to be appropriately sized.

The cryogenic systems are also responsible for the distribution of liquid nitrogen for operational activities requiring cryogenic temperatures (e.g., cryopumping, xenon gas sampling) or nitrogen gas that can be used for purging detector vessels during commissioning and/or operation. Such distribution systems are in operation for current generation experiments, and can be readily implemented again for the new experiment. The nitrogen gas supply can be sourced by capturing boil-off liquid nitrogen or utilizing systems that generate nitrogen gas from a compressed air supply.

As with all auxiliary detector systems, redundancy must be embraced to minimize the risk of detector failure or downtime. For the cryogenic systems, this will motivate backup cryocooler unit(s), a liquid nitrogen cooling vessel (either integrated with a cryocooler or as a separate backup vessel), and at least two gaseous nitrogen distribution mechanisms.

Figure 12: Left: The XENONnT liquid purification system currently installed at LNGS. Right: The high-purity high-pressure Fluitron gas compressors used for LZ circulation and recovery.

4.6 Calibration Systems

A thorough calibration of the detector is a crucial step in ensuring accurate interpretation of its data. To this end, the calibration program will focus on extensively calibrating XLZD’s core TPC volume and veto detectors. The calibration campaign will build upon the experience gained from successful calibration efforts in previous experiments, including XENON1T, LUX, XENONnT, and LZ Aprile:2019bbb ; Akerib:2017vbi ; Aalbers_2024_calibration .

TPC calibrations will serve two main purposes: correcting the position and possibly time-dependent variations for the S1 and S2 signals, such as the light collection, electron lifetime, electron extraction efficiency, and field distortions, and measuring the light and charge yield for specific particle interactions, such as electronic and nuclear recoils and their respective detection efficiencies and energy calibrations. Due to the large size of the detector, external sources are not able to calibrate the detector besides the limited volume near the walls. As such, the extensive use of dispersed sources, such as gaseous radioactive sources injected into the calibration system and mixed into the TPC, will be an essential component of the calibration process. Potential calibration sources are ³⁷Ar Boulton:2017hub , ^83mKr Manalaysay:2009yq , ^131mXe, ²²⁰Rn Aprile:2016pmc ; Lang:2016zde ; J_rg_2023_Rn220 , ²²²Rn J_rg_2023_Rn222 and CH₄-based sources (e.g. ¹⁴C, ³H) Akerib:2015wdi . These sources will enable corrections to both the S1 and S2 signals with respect to the position of the interaction and the response of the detector to electron recoils induced by $\beta$ and ${\gamma}$ particles.

Calibration of the nuclear recoil response will still require an external source, as there is currently no efficient and secure method for dispersing them in the liquid. This calibration is typically performed using neutrons from a pulsed D-D generator external to the xenon detector Akerib:2016mzi or sealed radioactive sources exploiting reactions such as ²⁴¹Am($\alpha$, n)Be, ${}^{88}Y$($\gamma$, n)Be, or spontaneous fission ²⁵²Cf sources delivered in calibration tubes placed outside the TPC Collar:2013xva . While these calibrations are limited to the outer regions of the TPC due to the small interaction length of the neutrons in LXe, they will still be invaluable in verifying our modeling of nuclear recoils in liquid xenon Szydagis:2022ikv . In combination with the information on detector response gained from dispersed source electron recoil combinations, the nuclear recoil response in the detector can be precisely predicted.

A D-T neutron generator, emitting neutrons with an energy of 14.1 MeV, can be used to calibrate the response of nuclear recoil up to 400 keV, a critical requirement for an Effective Field Theory analysis LZ:2023lvz . Additional novel calibration methods are being investigated.

Optical calibration of the light sensors will also be conducted periodically to monitor stability in their gains. Optical calibration options include using LEDs permanently mounted on the top and bottom PMT support arrays or by injecting the LED light into the TPC by means of several optical fibers.

Besides TPC calibrations, the veto detectors (described in Section 4.3) will also be thoroughly calibrated, mainly for their external background tagging efficiency for vetoing those backgrounds. Optical calibration of the PMTs in the veto detectors will also be conducted. Currently running detectors have demonstrated the calibration of the neutron detector’s tagging efficiency with AmBe or AmLi sources situated outside the cryostat  XENON:2023sxq ; Aalbers:2022dzr .

4.7 Electronics and Data Acquisition

The individual signals from the light sensors of the TPC and the outer detector need to be digitized, time-stamped, and stored for further analysis. Online reconstruction of peaks (i.e., excursions from the baseline on individual channels) to events induced by particle interactions is required for immediate feedback on the detector performance and for fast data analysis. A single, highly parallelized data acquisition (DAQ) system for all sub-detectors will ensure that the signals are read out in a time-synchronized way, while different digitization speeds might be chosen for the different systems. The baseline digitization frequency for the large TPC is 100 MHz with a 40 MHz input bandwidth to prevent aliasing effects. However, the use of autoencoders on the digitizer’s FPGA (Field Programmable Gate Array) to exploit compressive sensing is an interesting option and will be explored. To achieve the lowest possible light (S1) threshold, the digitization threshold for the TPC photosensors needs to be well below a single photoelectron.

The currently employed DAQ systems aalbers2024data ; XENON:2019bth ; XENON:2022vye generally fulfill these requirements and can be straightforwardly scaled up for the planned detector. To achieve the lowest possible threshold, the XENON DAQ system XENON:2019bth ; XENON:2022vye operates without a global trigger, i.e., every channel is digitized individually if the signal on it exceeds 0.1 photoelectron. Its currently employed commercial firmware for peak identification on the FPGA caen_dpp-daw could be optimized and better matched to offline hit finding. While such operation is desirable for dark matter searches, the huge amount of data complicates operation during calibration campaigns where large event rates in non-interesting parameter spaces (location, energy) accumulate. Intelligent online veto systems (anti-triggers) that analyze the incoming data stream using machine-learning methods implemented on fast FPGAs and that are adapted to the used digitizers reject events from these parameter spaces and thus ensure that the fraction of relevant signals is highly increased during calibration.

4.8 Slow Control and Long-Term Stability

The slow control (SC) system will provide real-time monitoring and control over all subsystems and allow offline access to the data being monitored. The main goal of the system is to ensure the safety of both the personnel operating the detector and the hardware, including the xenon gas itself, while also maintaining optimal detector performance and enabling quick detection and resolution of any issues that may arise. This will allow the detector to run in a stable mode for extended periods without interruptions. The system will adhere to all relevant laboratory regulations and standards.

Based on experience from current generation detectors LZ:2019sgr ; XENON:2024wpa , the SC will be designed as a distributed system of programmable automation controllers (PACs), allowing each subsystem to operate independently while communicating over a dedicated network. To avoid interruptions, all equipment connected to the SC must adhere to high industrial standards.

The central SC interfaces will include a supervisory control and data acquisition (SCADA) system, alert and notification service, as well as an operator log, online monitoring tools, authentication and authorization system, and SC data archiving. A SC data stream will be integrated with the DAQ system to allow easy access to the data for analysis. Additional design requirements include a high level of network security, expandability, backup network, heartbeat monitoring, and redundancy.

4.9 Software and Computing Engineering

Storing, processing, and analyzing data, as well as efficiently simulating interactions in the detector, are essential to ensure that the extensive information collected can be transformed into scientific results. The size and scope of XLZD create strong computing and software design requirements to achieve this goal. These requirements are on scales more similar to other large particle physics experiments than previous dark matter experiments, given the pace of growth in our field. This is explored in detail in the whitepaper of Roberts:2022ezy .

The XLZD software packages benefit from the experience of the current generation detectors. The existing event reconstruction frameworks of LZ and XENON can be effectively scaled to meet our scientific goals. The success of this scaling largely depends on effective integration with the computing infrastructure available. Simulations serve as our blueprint, illuminating the nature of prospective data and the detector’s sensitivity, all thanks to the intensive modeling efforts in progress. These simulations, pivotal from the detector’s design phase to the conclusive data analysis, rely on tools like Geant4 GEANT4:2002zbu and NEST Szydagis:2011tk ; Szydagis:2013sih . However, hurdles remain, especially in simulating intricate detector responses. For instance, with Geant4, it is imperative for the experiment to incorporate and validate updated low-energy models. Geant4-based simulations from earlier projects Akerib:2021ap will be adapted, and co-processors integrated to facilitate optical modeling at this magnitude.

A primary challenge for XLZD computing infrastructure is the management of the expected data volume with current generation detectors having petabyte-scale datasets. Although XENONnT and LZ currently employ HEP tools such as Rucio Rucio_2019 , DIRAC DIRAC_2014 , and Globus Globus_2011 , tailoring these to the expanded needs of this endeavor necessitates robust community dialogue, crafting best practices, and integrating technologies like containers. This will guarantee efficient data management and reproducibility across various computing systems.

Therefore, the core foundations of the software and computing project are present and scalable. Still, an R&D effort is required to determine how to scale these tools to the experiment’s requirements using existing computing infrastructure.

5 Control and Mitigation of Backgrounds

An ultra-low background environment is essential to the success of XLZD. The background mitigation strategies discussed in this section aim to ensure that backgrounds due to radioactivity fall below the expected level of irreducible neutrino backgrounds in the form of neutrino elastic scattering and neutrino-nucleus scattering.

Controllable backgrounds in the detector are those that arise from radioactivity present in the detector materials, in particular from the ²³⁸U and the ²³²Th primordial decay chains. Given the precedence of radio-purity improvements in the field, accomplished with the aid of state-of-the-art material screening techniques, it is assumed that a reduction by up to a factor of three in material backgrounds compared to LZ/XENONnT can be reached. Despite the greater cryostat and photosensor mass, with these radioactivity estimates and considering the self-screening of the monolithic xenon volume, gamma-ray backgrounds will be negligible for a WIMP search with a detector of this target mass. Such a reduction in material backgrounds will also allow us to maximize XLZD’s sensitivity to ¹³⁶Xe $0\nu\beta\beta$ decays as shown in Section 2.2.

Impurities in the xenon target itself, such as ⁸⁵Kr and ²²²Rn, can cause beta decays without accompanying gamma ray (often referred to as “naked”). The resulting electron recoils can leak into the WIMP signal region. Developments in chromatography and distillation techniques to remove trace krypton and argon will allow the experiment to meet a requirement of 0.1 ppt of ${}^{\text{nat}}$Kr thus rendering beta backgrounds from ⁸⁵Kr insignificant. ²¹⁴Pb decays from the ²²²Rn chain are numerically the most significant background and, therefore, a major focus in present-generation experiments. For future R$\&$D, there are a number of planned and achieved advancements that can be capitalized on to suppress the ²²²Rn activity to a projected 0.1 $\mu$Bq/kg.

Finally, cleanliness protocols will be adopted modeled on the successful measures undertaken during the LZ construction campaign, which were shown to prevent the plate-out of long-lived ²¹⁰Pb on potential xenon-wetted surfaces. This, in turn, has lowered the impact of so-called wall backgrounds from misreconstructed ²⁰⁶Pb recoils on WIMP searches.

5.1 Low-Background Screening for Detector Materials

All detector materials contain trace amounts of naturally occurring radioactive isotopes. For the detector, the key isotopes to consider are those of uranium (²³⁵U and ²³⁸U) and thorium (²³²Th) and their subsequent decays within the decay series. Additionally, the isotopes ⁴⁰K, ⁶⁰Co, and ¹³⁷Cs are often present and emit gamma rays which drive the fiducial volume choice. Materials are preferentially chosen to limit this residual radioactivity and thus reduce backgrounds within the detector which necessitates a comprehensive screening campaign prior to the detector construction.

Different technologies are uniquely sensitive to different species, and so a multifaceted approach is necessary. For example, in the case of the uranium and thorium series, various techniques are used. To examine the elemental levels of the parent isotopes, precise inductively-coupled plasma mass spectrometry (ICP-MS) techniques can provide sub-ppt g/g sensitivity LaFerriere:2015 . This destructive technique involves the chemical treatment of samples before injecting them into the machine, which houses a plasma and mass spectrometer. ICP-MS was used heavily in the screening campaigns of LZ and XENON, assaying many hundreds of samples this way. The fast turnaround of samples (within a day) also provided a useful tool for removing more highly contaminated samples before embarking on more time-intensive assays  Akerib:2020com . A Thermo Fisher Element2, Agilent 7900 (as used for LZ) DOBSON201825 , as well as an Agilent 8900, will be available for a future assay campaign. They offer exceptional throughput and are considered more than sufficient to reach ppt sensitivity.

Complementary gamma spectroscopy can then provide individual component activities further down the decay chains, where secular equilibrium can be broken due to longer half-lives within the series. Indeed, screening campaigns within LZ Akerib:2020com and XENONnT XENON:2021mrg had access to some of the worlds leading gamma spectroscopy devices such as those at BUGS Scovell:2018ap , BHUC mount:black_2017 , Gator Baudis:2011am ; Araujo:2022kip , GeMSE Garcia:2022jdt , the GeMPI spectrometers Heusser:2006 , and GIOVE Heusser:2015ifa . During the construction of LZ and XENON1T/nT, hundreds of samples were assayed within gamma counters. Over more than six years, different materials were examined, forming two of the largest and most detailed assay campaigns for a low background experiment ever conducted. The resulting database of materials, including vendor information, is a valuable asset on which future efforts will build. In addition to the decays in the radon and thorium chains, these detectors are also sensitive to the other and usually unexpected gamma-emitting isotopes referred to earlier in this section.

Where there is a greater need for increased sensitivity, such as for components close to the active detector, or where the masses of such materials are very large, it may be worth surveying additional mass spectroscopy systems that are not heavily used in the rare event field. Neutron activation analysis (NAA), which several groups in XLZD have access to, can be used in combination with gamma screening to achieve ultra-low background results. Routine NAA sensitivity for K, Th, and U is possible at 1 ppb, 1 ppt, and 1 ppt, respectivelyAkerib:2020com . However, for critical samples, fine-tuned analyses and sample preparations have resulted in even greater sensitivity levels, such as Th and U reaching 0.02 ppt (EXO cryogen)LEONARD2017169 ; LEONARD2008490 . In addition, activation measurements have achieved the sensitivity required for LZ, detecting ²³²Th at concentrations ranging from 10 to 20 ppt, and placing limits on ²³⁸U concentrations ranging from 0.4 to 4.6 ppt, respectively Akerib:2020com . Another interesting technique is the Accelerator Mass Spectroscopy (AMS) which is orders of magnitude more sensitive than ICP-MS for long-lived radionuclides (e.g. ²¹⁰Pb) stenstrom2020identifying .

5.2 Control of Krypton and Radon

Small amounts of radioactive noble gases are mixed in the liquid xenon target and can neither be removed by hot getters in the xenon purification system nor by fiducialization. Krypton contains the anthropogenic beta emitter ⁸⁵Kr and is typically present at the few ppb levels in commercial xenon. Therefore, the xenon has to be purified either by gas chromatography Ames:2023 or by cryogenic distillation XENON:2016bmq to the required purity level of less than 100 ppq. Purification systems of current generation detectors are shown in Fig. 13. There are no sources of ⁸⁵Kr in the detector, so in principle, only one purification cycle is needed. However, tiny air leaks or outgassing from plastic components (mostly from PTFE) in the detector or an operation failure may increase the Kr concentration. A variation in the krypton level can only be discovered by regular monitoring with ppq sensitivity. This can be realized by a commercial RGA analyzer with enhanced sensitivity for krypton Dobi:2011vc or by a dedicated rare gas mass spectrometer Lindemann:2013kna . Both techniques require an efficient removal of the bulk xenon prior to the measurements, which may be realized with a custom-made gas chromatography system, by crogenic distillation, or by cryo-trapping of xenon at a well-defined temperature. The two former technologies are already available within XLZD, and the latter is being explored. Should the krypton level rise above a certain threshold, online xenon purification for krypton removal as demonstrated in XENON1T XENON:2021fkt will be applied. During this online cryogenic distillation, krypton is concentrated in the offgas, which is taken out of the system. To keep the corresponding xenon losses for such a large detector minimal a dual stage distillation system with very little offgas losses (e.g. $10^{-5}$) needs to be constructed. Therefore, krypton removal phases will not disturb the continuous data taking.

Radon, in particular ²²²Rn, is considered to be the more challenging dispersed contaminant because ²²²Rn sources will unavoidably exist in the detector materials due to traces of ²²⁶Ra from the natural uranium decay chain. Therefore, ²²²Rn will continuously emanate from detector materials into the xenon target. Avoiding ²²²Rn sources is thus the best strategy. This will be addressed by a thorough material screening program similar to the one applied for bulk trace radioactivity. Samples will be placed in gas-tight containers surrounded by an initially radon-free carrier gas. After a couple of days, the emanated radon will be determined with an appropriate ²²²Rn detector. Electrostatic radon monitors, scintillation counting, and ultralow background proportional counters Akerib:2020com ; XENON:2020fbs are available with sensitivities down to 20 $\mu$Bq. While most emanation measurements are done at room temperature, a setup to study the radon emanation at cryogenic temperature is also being developed. The Cold Radon Emanation Facility (CREF), based at the Rutherford Appleton Laboratory in the UK, aims to measure the emanation rates for large samples at cryogenic temperatures to better than 0.1 mBq sensitivity, furthering the knowledge of these processes. However, given the ubiquitous nature of ²²⁶Ra, not all ²²²Rn sources can be removed by screening. Further mitigation strategies include online radon removal, special surface treatments, radon tagging during data analysis, and TPC design.

Since there are radon sources inside the detector, any radon removal strategy requires continuous operation during the lifetime of the experiment. The radon removal from xenon is done either via gas chromatography, making use of the different diffusion times of radon and xenon in cold charcoal filters Arthurs_2021 , or using a cryogenic distillation column Murra:2022mlr ; Aprile:2017kop profiting from the factor 10 lower vapor pressure of radon with respect to xenon. In contrast to cryogenic online distillation for krypton removal, for radon removal, there is no offgas loss because the radon is kept in the system until disintegration Murra:2022mlr . The experience from previous detectors showed that, typically, the majority of radon sources are located in the gaseous xenon part of the experiment. By a smart design of the xenon flow paths, it will be possible to selectively purify those parts that contain only a small fraction of the entire xenon mass.

At XENONnT the cryogenic online distillation system has demonstrated a radon reduction factor of two extracting xenon gas from the most relevant points. In addition, radon already inside the active LXe volume can be removed by online purification with LXe extraction if the purification time for the detector’s entire xenon mass is of the same order of magnitude as the ²²²Rn half-life Murra:2022mlr . This can be met by cryogenic distillation as demonstrated by XENONnT, which operates such an online radon removal system running at a typical purification speed of nearly 2 tonnes per day. This reduces the Rn concentration by another factor of two and yields a Rn concentration below 1 µBq/kg XENON:2024wpa . The larger XLZD detector requires a challenging xenon purification speed of $\mathcal{O}$(10 tonnes per day). R&D for constructing a larger system with efficient use of heat pumps to obtain the enormous cooling power is underway.

Figure 13: The online krypton XENON:2016bmq (left) and radon removal systems Murra:2022mlr (middle) developed for the XENON1T/nT experiments and the LZ offline krypton removal system located at SLAC (right).

For the cryostat containing the liquid xenon itself, novel surface treatment techniques are under development. A thin, clean, and tight coating may block recoil- and diffusion-driven ²²²Rn emanation. It was recently demonstrated that electro-deposited copper coating on a small stainless steel sample can reduce the radon emanation rate by more than a factor of 1500 Jorg:2022spz . To become applicable in XLZD, up-scaling is necessary, as well as strict purity control, as tiny traces of ²²⁶Ra in the coating would spoil the achieved reduction. Another promising method to reduce the radon level is an almost gas-tight TPC design, which separates the active target from the more Rn-contaminated outer volume Dierle_2023 ; Sato2020development . Finally, the remaining radon-induced background may be reduced by smart data analysis techniques XENON:2024qvh . The approach makes use of characteristic coincidences in the radon decay chains, which may be used to tag the event if the convective tracks of radon daughters in the detector are sufficiently well understood.

5.3 Surface Contaminants

Despite thorough material screening and selection efforts, surfaces may be contaminated during parts machining and the subsequent detector construction. This is predominantly from two sources: firstly, the plate-out of long-lived radon daughters such as ²¹⁰Pb ($\tau_{1/2}=22.3\,\textrm{yrs}$) jacobi_activity_1972 and from the deposits of dust and debris onto the detector parts. In the latter case, the dust can then become mobile within the LXe volume, which can cause instabilities in the HV should this dust later attach to electrodes or feed-throughs. In addition, this dust also emanates radon, thus adding to the total detector background.

In the case of radon daughters that have been plated out upon the surfaces of detector materials during detector construction, there are multiple contributions to detector backgrounds. Firstly, ²¹⁴Pb contributes a naked beta component, which is almost impossible to veto, and hence is present during all science searches. This is especially problematic in the case where daughters of radon become mobile, and hence, decays occur within the fiducial volume. Additionally, the alpha decays of ²¹⁰Po within the ²²²Rn chain can undergo ($\alpha,n$) processes within the detector walls, which, depending on the interaction topology of the neutron, can be indistinguishable from WIMPs. To mitigate this as much as possible, the experiment can utilize and further develop techniques used in previous experiments. For example, LZ used a reduced radon cleanroom in which air is passed through a cold carbon filter to reduce the quantity of radon in the air by several orders of magnitude Akerib:2020com . This will be particularly important within an underground setting where radon levels can be higher than on the surface. Air-ionizer units were additionally utilized to prevent charge accumulation on PTFE, limiting the plate-out of charged daughters onto the detector materials Akerib:2020com . Chemical surface treatment of individual detector parts like etching, pickling, leaching, and passivation could also be used to remove radon daughters that have plated out. This process includes placing components into acid to remove a thin layer of material (with any contamination) on the surface. This was utilized in both XENONnT XENON:2021mrg and LZ LZ:2019sgr . It has also been shown that the bagging of detector parts within nylon or storing in nitrogen-flushed or evacuated containers for extended periods of time can further reduce any plate-out XENON:2021mrg ; meng_new_2020 . While these mitigations limit the plate-out, any exposure should still be tracked during construction in order to help form an estimated background rate. For example, in LZ, the average plate-out was calculated to be ($158\pm 13$) $\mu$Bq/m², less than a third of the requirement Akerib:2020com . Measurement of the plate-out radon progeny on material surfaces such as that of PTFE and electrodes is possible, typically by examining the surface alpha activity from the ²¹⁰Po isotope measured via XIA UltraLo-1800 or Si-PIN detectors XENON:2021mrg . It is also possible to measure the plate out in detectors such as BetaCage, which is a neon drift chamber that is extremely sensitive to low energy electrons and alpha particles schnee_screening_2007 . The ²¹⁰Pb plate-out rate can be inferred by repeated measurements over an extended time period while parts are exposed to the detector construction conditions.

To limit dust deposition, the detector must be constructed within a cleanroom environment. Both XENONnT and LZ used class 1000 (ISO-6) clean rooms Akerib:2020com ; XENON:2021mrg . During operations, the levels of dust need to be monitored while working within the space. This can be performed using sensitive dust counters and witness coupons, which can estimate the deposition and hence provide an indication of the contamination. It was shown with previous experiments that the dominant contribution was from personnel working within the clean spaces by bringing in dust on clean room suits or equipment and by disturbing the laminar air flow Akerib:2020com . Therefore, methods to reduce quantities of dust and debris brought into the clean spaces will be examined more closely. The use of the air-ionizer fans also ensured that the dust was not attracted to charged surfaces, further limiting this deposition. Deposition levels of (214$\pm$22) ng/cm² were estimated from measurements on LZ, less than half of the target contamination Akerib:2020com .

Figure 14: Existing large test facilities at surface labs (top left: Freiburg Brown_2024 , top right: Zurich Baudis:2021ipf ), and underground facility (bottom left: XMASS at Kamioka Abe_2013 ) within the XLZD collaboration. The LZ and XENONnT (bottom right) experimental infrastructures at SURF and LNGS offer additional testing capabilities after completing their science programs.

5.4 Accidental Coincidence Backgrounds

Although not a radioactive background, accidental coincidence events from lone-S1 and S2 signals of low intensity are observed in the XENONnT, LZ, and other similar LXe-TPC detectors. ‘Lone’ refers to the observation of the S1 (S2) part of an event without the accompanying S2 (S1) signal. Typically, the lone-S1 pulses are from the charge insensitive volume in or below the target and dark counts from PMTs. Lone-S2 pulses come from multiple sources such as low energy events with S1s too low to be detected, low energy $\beta$ decays from the cathode Aprile:2019xxb and delayed electrons XENON:2021myl attached on impurities or trapped below the liquid surface. Uncorrelated lone-S1 and S2 signals can combine to form accidental coincidence events, forming a background located largely in the low-energy region relevant to the study of low-mass WIMPs, CE$\nu$NS, and axions. The current generation detectors XENONnT and LZ, with their low backgrounds, are the best resources available to characterize these backgrounds in detail. With a larger and taller target volume in XLZD, critical R&D will be carried out to reduce the lone-S1 and lone-S2 pulses to control and reduce the accidental coincidence background to improve the sensitivity of the detector to do interesting science at the lowest energy threshold possible.

6 Critical R&D, Risk Mitigation, and Opportunities

The next-generation liquid xenon observatory XLZD will adopt most of the mature technologies that have already been demonstrated and rigorously tested in the LZ and XENON experiments. Further R&D efforts will provide solutions to lessons learned from the current generation experiments. New and redundant technology choices provide risk mitigation and alternative solutions during the design stage. A complete program of quality assurance and control for the production hardware and testing operational parameters or background mitigation strategies to employ during detector running, likely utilizing R&D infrastructure, will be required to meet the needs of scaling up and longevity of the experiment. In addition, opportunities for upgrades and improved performance to increase physics sensitivity shall be explored.

The areas of critical R&D study link back to the fundamental operations of the liquid xenon TPC, as well as the veto systems: reducing external and internal backgrounds to unprecedented levels, enhancing light and charge collection to lower energy thresholds, and aiding recoil discrimination. These primary physics goals lead to studies of low background materials, radon reduction and removal, photosensors and surface reflectivity, high voltage and electrodes, xenon purification, reduction of spurious charge and light noises, Cherenkov and scintillator veto. Although in situ calibrations can never be replaced, continued studies of xenon microphysics, the charge and light yield, and their physical dependencies will also aid the larger project.

Many of these R&D studies will be based on experience and mature technologies from the current generation of running experiments, which can be further improved at individual institutions, harnessing the established expertise of the collaboration. A key development for this scaled-up experiment will be larger test facilities both at the surface and underground. Large and multiple testing facilities allow for testing the integration of multiple subcomponents, as well as accessing the unknowns that arrive in scaling up to mitigate the risks. Maintaining a suitable system to test issues that may arise in the main detector operations will also provide key information for the operational team. Large testing facilities within the collaboration (Fig. 14) already exist at surface and underground laboratories. Surface testing facilities will allow performance verification of critical R&D components. Dedicated underground facilities will be used to investigate and certify the scaled-up components in a low background environment before installing them in the final detector.

Finally, the XLZD program will allow opportunities to improve the physics sensitivity in the search for light and low-mass dark matter by doping light elements in the bulk liquid and/or in a single-phase operation, where the proportional S2 signal is produced in the liquid xenon phase. Photosensors with higher photon detection efficiency and lower intrinsic radioactivity would further enhance the detector’s capability with lower threshold, higher position and energy resolutions, and enhanced background discrimination to continue improving the sensitivities to dark matter. Filling the detector with a xenon target enriched in ¹³⁶Xe would improve the sensitivity to neutrino-less double beta decays.

7 Summary

In this report, we present the experimental strategy and technology choices of a next-generation liquid xenon observatory for dark matter and neutrino physics with a nominal 60-80 tonnes active liquid xenon target mass. Thanks to the development and extended operations of the current generation experiments XENONnT and LZ, we now have mature, and in some areas redundant, technology choices to embark on such a massive next-generation experiment. Further improvement and R&D will continue to push the technological frontier, such as reducing the trace radioactivity background below the irreducible neutrino backgrounds to maximize the sensitivity for physics. Modest risks will be mitigated through extensive tests at existing large-scale testing facilities at surface and underground labs.