PandaX-4T: Multi-Purpose Xenon Detector

• PandaX-4T is a dual-phase liquid-xenon TPC that serves as a rare-event detector for dark matter, neutrino, and double-beta decay studies.
• The experiment integrates advanced cryogenic systems, triggerless readout, and stringent low-background controls to achieve high sensitivity and precise event reconstruction.
• Operational results include leading dark matter limits, first natural xenon double-beta decay measurements, and effective supernova burst detection capabilities.

PandaX-4T is the third-generation detector in the PandaX program and an underground rare-event experiment centered on a large dual-phase liquid-xenon time projection chamber at the China Jinping Underground Laboratory. It was developed as a four-ton-scale direct-detection instrument for weakly interacting massive particles, but operational papers describe it more broadly as a multi-purpose platform for dark matter, neutrino, and double-beta-decay physics. The design study described approximately 6 tons of total liquid xenon and 4 tons in the sensitive TPC volume, whereas operational papers describe 3.7 tonnes of liquid xenon in the sensitive region or active volume; in both formulations, the experiment is defined by multi-ton xenon mass, deep underground siting, water shielding, low-background construction, and event-by-event reconstruction from scintillation and ionization signals (Zhang et al., 2018, Meng et al., 2021, Collaboration et al., 2022, Collaboration et al., 2024, Pang, 24 Nov 2025).

1. Development, siting, and experimental scale

PandaX-4T was conceived as the next-stage PandaX dark matter direct-detection experiment, to be installed in the B2 hall of the second phase of the China Jinping Underground Laboratory, CJPL-II (Zhang et al., 2018). Operational papers place the experiment in Hall B2 of CJPL-II under about 2400 m of rock or marble overburden, and one atmospheric-dark-matter study quotes an overburden of 6700 m water equivalent with a cosmic-ray muon flux of $2.0\times10^{-10}\ \mathrm{cm^{-2}\,s^{-1}}$ (Meng et al., 2021, Ning et al., 2023). The detector is further enclosed by a large ultra-pure water shield; depending on the paper and context, this is described as a stainless steel water tank 10 m in diameter and 13 m deep, or equivalently as a shield 13 m high and 5 m in radius (Meng et al., 2021, Pang et al., 2024).

The experiment’s infrastructure was designed around a multi-ton xenon inventory. The cryogenics and xenon-handling system was commissioned for about 6 tons of liquid xenon, with a measured maximum cooling power of about $580~\mathrm{W}$ at $178~\mathrm{K}$, a filling rate of about $700~\mathrm{kg/day}$ with assisted liquid-nitrogen cooling, an average recuperation rate of about $440~\mathrm{kg/day}$, a maximum total purification speed of about $155~\mathrm{slpm}$ in two circulation loops, and large heat exchangers with measured efficiency $\sim 97.5\pm0.5\%$ (Zhao et al., 2020). Before Run 2, the detector underwent additional upgrades that included cathode replacement, redesigned PMT voltage-divider bases, new PTFE field-cage panels, conversion of the surrounding shield into an active water-Cherenkov veto with 270 8-inch PMTs, and a new triggerless readout based on custom 14-bit $500~\mathrm{MS/s}$ digitizers (Collaboration et al., 2 Jul 2026).

2. Detector architecture and operating principle

PandaX-4T is a cylindrical dual-phase xenon TPC bounded by 24 reflective PTFE wall panels. The operational detector is described with an opposite-panel distance of 1185 mm, a cathode-to-gate spacing of 1185 mm, and a gate-to-anode spacing of 10 mm; the sensitive xenon inventory is quoted as 3.7 tonnes in the active volume, while the total xenon mass is quoted as 5.6 tonnes in the commissioning dark-matter paper (Meng et al., 2021, Shen et al., 2024). Photosensor coverage is provided by 169 three-inch Hamamatsu R11410-23 PMTs on top and 199 on the bottom, and the commissioning configuration also included 105 one-inch Hamamatsu R8520 PMTs in the outer veto region (Meng et al., 2021).

As in other dual-phase xenon TPCs, an interaction in the liquid xenon produces prompt scintillation $S1$ and ionization electrons; the electrons drift upward, are extracted into the gas, and produce delayed electroluminescence $S2$. The $580~\mathrm{W}$0-$580~\mathrm{W}$1 time difference reconstructs $580~\mathrm{W}$2, and the top-array $580~\mathrm{W}$3 light pattern reconstructs the transverse position. A widely used PandaX-4T energy estimator is

$580~\mathrm{W}$4

where PDE is the photon detection efficiency, EEE the electron extraction efficiency, and $580~\mathrm{W}$5 the single-electron gain for the bottom-array $580~\mathrm{W}$6 (Meng et al., 2021). This combined scintillation-ionization formalism is reused across low-energy dark-matter analyses, MeV-scale double-beta-decay measurements, and solar-neutrino studies, with analysis-specific corrections and calibrations (Collaboration et al., 2022, Collaboration et al., 2023, Collaboration et al., 2 Jul 2026).

3. Cryogenics, readout, and low-background control

The PandaX-4T readout and data-acquisition system marked a major architectural change relative to PandaX-I and PandaX-II. The commissioning-era system introduced triggerless readout, 32 CAEN V1725 digitizers with 16 channels each, 14-bit resolution at $580~\mathrm{W}$7, a $580~\mathrm{W}$8 input range, and an average single-photoelectron recording efficiency of 96% for the dominant 3-inch PMTs. The DAQ reached a maximum bandwidth of $580~\mathrm{W}$9, and the summary reports performance above $178~\mathrm{K}$0 (Yang et al., 2021). Offline event building then reconstructs physical $178~\mathrm{K}$1-$178~\mathrm{K}$2 events from channel-level self-triggered waveform fragments, a design chosen to avoid efficiency loss from a global hardware trigger (Yang et al., 2021).

Low-background operation relies on a dedicated materials and cleanliness program. PandaX-4T developed and applied HPGe gamma spectroscopy, ICP-MS, NAA, radon emanation systems, a krypton assay station, and alpha detection for material screening and background control (Qian et al., 2021). The screening paper estimates total material background rates in the $178~\mathrm{K}$3 region of $178~\mathrm{K}$4 for electron recoil and $178~\mathrm{K}$5 for nuclear recoil, and estimates $178~\mathrm{K}$6Kr in the detector to be $178~\mathrm{K}$7 ppt (Qian et al., 2021). Operational analyses nonetheless show that residual internal backgrounds remained important in early data: the commissioning WIMP search identified tritium as a major and unexpected background, likely residual from PandaX-II end-of-run calibration, with average concentration about $178~\mathrm{K}$8 (Meng et al., 2021). This combination of very low intrinsic material background and analysis-dependent internal background control became a defining feature of the experiment’s physics program.

4. Dark-matter search program

The core PandaX-4T mission is direct dark-matter detection through low-energy xenon recoils. The first commissioning-run WIMP result used an exposure of $178~\mathrm{K}$9, identified 1058 candidate events in an approximate nuclear-recoil energy window between 5 and 100 keV, found no significant excess over background, and set a 90% C.L. spin-independent dark-matter–nucleon limit with a lowest excluded cross section of $700~\mathrm{kg/day}$0 at $700~\mathrm{kg/day}$1 (Meng et al., 2021). This early result already demonstrated that commissioning data from a multi-ton xenon detector could yield world-leading spin-independent WIMP limits (Meng et al., 2021).

PandaX-4T rapidly expanded beyond standard elastic WIMP scattering. A dedicated search for fermionic dark-matter absorption on xenon nuclei through a neutral-current process used the 95.0-day commissioning data, treated the signal as a monoenergetic nuclear-recoil-like feature from

$700~\mathrm{kg/day}$2

and set a strongest 90% C.L. limit of $700~\mathrm{kg/day}$3 at $700~\mathrm{kg/day}$4 over the scanned range $700~\mathrm{kg/day}$5 (Gu et al., 2022). An ionization-only commissioning analysis selected unpaired $700~\mathrm{kg/day}$6 signals between 60 and 200 PE, corresponding to mean nuclear-recoil energies of 0.77 to 2.54 keV and electronic-recoil energies of 0.07 to 0.23 keV, and set the most stringent limits in several low-mass intervals, including $700~\mathrm{kg/day}$7 to $700~\mathrm{kg/day}$8 for point-like dark-matter–electron interaction (Collaboration, 2022). Later, a 259-day Run0+Run1 analysis with effective exposures of $700~\mathrm{kg/day}$9 for ionization-only data and $440~\mathrm{kg/day}$0 for paired data reported the most stringent constraints on spin-independent dark-matter–nucleon interactions within $440~\mathrm{kg/day}$1 to $440~\mathrm{kg/day}$2, spin-dependent neutron-only interactions within $440~\mathrm{kg/day}$3 to $440~\mathrm{kg/day}$4, and spin-dependent proton-only interactions within $440~\mathrm{kg/day}$5 to $440~\mathrm{kg/day}$6, while improving dark-matter–electron limits for both heavy- and light-mediator scenarios (Zhang et al., 16 Jul 2025).

The experiment has also been used to test nonstandard production and mediation mechanisms. PandaX-4T reported the first search for atmospheric light dark matter from $440~\mathrm{kg/day}$7-meson cascades in the atmosphere, introducing a boosted-dark-matter propagation framework with both elastic and quasi-elastic Earth attenuation and deriving a lowest excluded cross section of $440~\mathrm{kg/day}$8 at $440~\mathrm{kg/day}$9 and $155~\mathrm{slpm}$0, together with a lowest upper limit of $155~\mathrm{slpm}$1 on $155~\mathrm{slpm}$2-to-dark-matter decay branching ratio (Ning et al., 2023). A solar-boosted-dark-matter search using $155~\mathrm{slpm}$3 of low-energy ER data reported a strongest 90% C.L. upper limit of $155~\mathrm{slpm}$4 at $155~\mathrm{slpm}$5 over $155~\mathrm{slpm}$6 to $155~\mathrm{slpm}$7, stated as a 23-fold improvement over the previous experimental result from CDEX (Shen et al., 2024). A separate commissioning-run reinterpretation of a UV-complete lepton-portal model found 1058 observed events consistent with an expected $155~\mathrm{slpm}$8 background and reported strong exclusions for Dirac fermion dark matter, with substantially weaker direct-detection sensitivity for Majorana dark matter because only the anapole interaction survives (Collaboration, 2024). This suggests that PandaX-4T’s dark-matter role is best understood as a family of search channels spanning elastic scattering, absorption, boosted fluxes, and UV-complete mediator structures rather than a single WIMP observable.

5. Neutrino and double-beta-decay measurements

PandaX-4T has also been used as a MeV-scale rare-event detector with natural xenon. Its first $155~\mathrm{slpm}$9Xe two-neutrino double-beta-decay measurement used 94.9 days of data, a $\sim 97.5\pm0.5\%$0Xe exposure of 15.5 kg-year, and a central cylindrical fiducial volume of $\sim 97.5\pm0.5\%$1 kg natural xenon, yielding

$\sim 97.5\pm0.5\%$2

described as the first $\sim 97.5\pm0.5\%$3Xe double-beta half-life measurement using natural xenon (Collaboration et al., 2022). The detector was later used to search for double beta decay of $\sim 97.5\pm0.5\%$4Xe in a 656 kg fiducial volume containing 68.7 kg of $\sim 97.5\pm0.5\%$5Xe, establishing 90% C.L. lower limits of $\sim 97.5\pm0.5\%$6 yr for $\sim 97.5\pm0.5\%$7 and $\sim 97.5\pm0.5\%$8 yr for $\sim 97.5\pm0.5\%$9, both described as the best limits to date (Collaboration et al., 2023).

At lower recoil energies, PandaX-4T has entered the solar-neutrino regime through coherent elastic neutrino-nucleus scattering. A combined Run0+Run1 analysis of paired and unpaired $500~\mathrm{MS/s}$0-only data collected exposures of 1.20 and 1.04 tonne$500~\mathrm{MS/s}$1year, respectively, and reported a best-fit $500~\mathrm{MS/s}$2B signal of 3.5 events from the paired data and 75 events from the US2 sample, with the background-only hypothesis disfavored at $500~\mathrm{MS/s}$3 and an inferred solar $500~\mathrm{MS/s}$4B neutrino flux of $500~\mathrm{MS/s}$5 (Collaboration et al., 2024). A later Run 2 electronic-recoil analysis, combined with Run 0, reported a solar $500~\mathrm{MS/s}$6 neutrino flux of $500~\mathrm{MS/s}$7, consistent with the Standard Solar Model, with a significance of $500~\mathrm{MS/s}$8 above background and identified this as the first positive indication of solar $500~\mathrm{MS/s}$9 neutrino–electron scattering below an electronic-recoil energy of 165 keV (Collaboration et al., 2 Jul 2026). These results establish PandaX-4T as both a nuclear-recoil and an electronic-recoil neutrino detector.

6. Low-energy electronic-recoil spectroscopy and nonstandard particles

Low-energy ER data have been used to search for a broad set of weakly coupled particles and interactions beyond standard WIMP phenomenology. Using a $S1$0 exposure below 30 keV, PandaX-4T searched for solar axions, neutrinos with anomalous magnetic moment, axionlike particles, dark photons, and light fermionic dark matter. No significant excess was observed, and the analysis reported $S1$1 for ABC solar axions, $S1$2 for $S1$3Fe solar axions, $S1$4 for Primakoff axions, and $S1$5 for anomalous neutrino magnetic moment, while stating that the fermionic dark-matter–$S1$6 conversion limits are the most stringent for most of the mass range $S1$7 (Collaboration et al., 2024).

A separate search for MeV-scale axionlike particles and dark photons used PandaX-4T’s electron-recoil response to look for monoenergetic absorption peaks. That analysis reported no signal excess and stated that PandaX-4T established the most stringent exclusion limits for most ALP and dark-photon masses across $S1$8 to $S1$9, with an average improvement factor of 3.5 in the $S2$0 interval compared with previous results (Collaboration et al., 2024). In combination, these ER programs show that PandaX-4T is not restricted to nuclear-recoil searches; it also functions as a precision low-background spectrometer for absorption-like and scattering-induced ER signals.

7. Supernova burst detection and broader observatory role

Because CE$S2$1NS in xenon is flavor inclusive and produces low-energy nuclear recoils, PandaX-4T has been developed as a supernova burst monitor. A dedicated 2024 study predicted that a Galactic core-collapse supernova at 10 kpc would yield 6.6 to 13.7 observed events in 10 s in the stricter “golden” trigger configuration, or 7.7 to 15.9 in the “silver” configuration, depending on progenitor and model, with negligible backgrounds in the selected burst window (Pang et al., 2024). That work also built two specialized alarms based on a 10 s time window and multiplicity threshold $S2$2, intended for real-time monitoring (Pang et al., 2024).

This capability was later implemented in a GPS-synchronized real-time monitoring system. The deployed system uses an effective target mass $S2$3 tonnes for the supernova analysis, a dead-time-cut efficiency of 89%, an overall detection efficiency of 13% for the reference $S2$4 Garching model, a post-cut candidate background rate of $S2$5, and a trigger condition of $S2$6 within 10 s, producing a false-alert rate of about twice per month and an end-to-end latency of about 6 minutes (Pang, 24 Nov 2025). For a supernova at 10 kpc, the expectation is about 8.1 events for the $S2$7 model and 3.7 events for the $S2$8 model, and the paper states that PandaX-4T reaches nearly 100% supernova detection probability at 10 kpc for the $S2$9 Garching model (Pang, 24 Nov 2025). The same work describes the method as directly scalable to PandaX-20T, with projected nearly 100% efficiency out to 30 kpc (Pang, 24 Nov 2025). A plausible implication is that PandaX-4T’s historical significance lies not only in its dark-matter limits, but also in demonstrating how a multi-ton xenon TPC can operate simultaneously as a dark-matter detector, a solar-neutrino instrument, a double-beta-decay experiment using natural xenon, and a network-ready supernova observatory.