LEGEND: Neutrinoless Decay Experiment

• The LEGEND Experiment is a staged search for neutrinoless double-beta decay in 76Ge, leveraging high-purity germanium detectors and a low-background liquid argon environment.
• It builds on GERDA and MAJORANA Demonstrator principles by achieving sub-3-keV energy resolution and employing advanced pulse-shape discrimination and active veto techniques.
• LEGEND aims to reach a half-life sensitivity of up to 10^28 years by scaling from a 200 kg phase to a ton-scale setup in a quasi-background-free regime.

The LEGEND experiment—Large Enriched Germanium Experiment for Neutrinoless Double-Beta Decay—is a staged search for neutrinoless double-beta decay, $0\nu\beta\beta$, in $^{76}$Ge using high-purity germanium detectors enriched in the isotope and operated in a low-background liquid-argon environment. It combines detector technologies, shielding concepts, and low-background methodologies developed in GERDA and the MAJORANA DEMONSTRATOR, with the immediate aim of reaching a $0\nu\beta\beta$ half-life sensitivity of order $10^{27}$ yr in LEGEND-200 and a ton-scale discovery sensitivity approaching or exceeding $10^{28}$ yr in LEGEND-1000. Observation of the decay would demonstrate lepton-number violation and establish that neutrinos are Majorana fermions (Acharya et al., 15 May 2025, Guinn et al., 2019).

1. Physics objective and theoretical framing

LEGEND targets the process

$$^{76}\mathrm{Ge}\rightarrow {}^{76}\mathrm{Se}+2e^{-},$$

with no neutrinos emitted. In the standard light-Majorana-neutrino interpretation, the inverse half-life is written as

$$\bigl(T_{1/2}^{0\nu}\bigr)^{-1}=G^{0\nu}\,\bigl|M^{0\nu}\bigr|^2\,\frac{m_{\beta\beta}^2}{m_e^2},$$

where $G^{0\nu}$ is the phase-space factor, $M^{0\nu}$ the nuclear matrix element, $m_e$ the electron mass, and $^{76}$0 the effective Majorana mass (Acharya et al., 15 May 2025, Guinn et al., 2019).

For $^{76}$1Ge, the signal appears at $^{76}$2 keV. The experimental strategy therefore relies on combining very small background indices with the intrinsic energy-resolution advantages of germanium detectors. The 2019 LEGEND status report states that covering the inverted-ordering region, with $^{76}$3 meV in the standard inverted-ordering scenario, corresponds to a half-life around $^{76}$4 yr for $^{76}$5Ge (Guinn et al., 2019). Earlier LEGEND design studies summarized the same target as effective Majorana masses of about $^{76}$6–$^{76}$7 meV for a tonne-scale experiment with excellent energy resolution and backgrounds at the level of $^{76}$8 count/(FWHM$^{76}$9t$0\nu\beta\beta$0yr) (Collaboration et al., 2017).

This physics framing determines the structure of the program. The half-life sensitivity improves with exposure, efficiency, resolution, and background suppression; in the background-free regime it scales linearly with exposure, whereas in the background-limited regime it scales approximately as $0\nu\beta\beta$1 (Collaboration et al., 2017, Saleh, 13 Mar 2026). LEGEND is designed explicitly to operate as close as possible to the quasi-background-free limit over large exposures.

2. Origins, predecessor experiments, and phased program

LEGEND was formed by combining the technological bases of GERDA and the MAJORANA DEMONSTRATOR. Those experiments established the main germanium-based design principles later adopted by LEGEND: detector-as-source operation, point-contact pulse-shape discrimination, ultra-clean materials, and aggressive active vetoing. The 2019 program summary states that GERDA and the MAJORANA DEMONSTRATOR had achieved the best intrinsic energy resolution and the lowest background rate in the signal region among all $0\nu\beta\beta$2 experiments, with GERDA BEGe detectors at $0\nu\beta\beta$3 keV at 2039 keV and background index $0\nu\beta\beta$4 counts/(FWHM$0\nu\beta\beta$5t$0\nu\beta\beta$6yr), and the MAJORANA DEMONSTRATOR at $0\nu\beta\beta$7 keV and background index $0\nu\beta\beta$8 counts/(FWHM$0\nu\beta\beta$9t$10^{27}$0yr) (Guinn et al., 2019).

The collaboration adopted a two-stage roadmap. LEGEND-200 reuses the GERDA infrastructure at LNGS and was designed around 200 kg of enriched germanium with a background goal below $10^{27}$1 counts/(FWHM$10^{27}$2t$10^{27}$3yr), corresponding to approximately $10^{27}$4 counts/(keV$10^{27}$5kg$10^{27}$6yr), and a projected sensitivity of order $10^{27}$7 yr with about $10^{27}$8 t$10^{27}$9yr exposure (Guinn et al., 2019, Brugnera, 17 Jan 2025). LEGEND-1000 is the ton-scale phase, designed around 1000 kg of enriched detectors, $10^{28}$0 t$10^{28}$1yr exposure, and a background target below $10^{28}$2 counts/(FWHM$10^{28}$3t$10^{28}$4yr) or $10^{28}$5 counts/(keV$10^{28}$6kg$10^{28}$7yr), depending on the study convention used (Collaboration et al., 2021, Brugnera, 17 Jan 2025).

Phase	Main parameters	Nominal reach
LEGEND-200	200 kg of enriched $10^{28}$8Ge in the existing GERDA infrastructure at LNGS; background goal $10^{28}$9 counts/(FWHM$$^{76}\mathrm{Ge}\rightarrow {}^{76}\mathrm{Se}+2e^{-},$$0t$$^{76}\mathrm{Ge}\rightarrow {}^{76}\mathrm{Se}+2e^{-},$$1yr)	Sensitivity $$^{76}\mathrm{Ge}\rightarrow {}^{76}\mathrm{Se}+2e^{-},$$2 yr with $$^{76}\mathrm{Ge}\rightarrow {}^{76}\mathrm{Se}+2e^{-},$$3 t$$^{76}\mathrm{Ge}\rightarrow {}^{76}\mathrm{Se}+2e^{-},$$4yr (Guinn et al., 2019)
LEGEND-1000	1000 kg of detectors enriched to $$^{76}\mathrm{Ge}\rightarrow {}^{76}\mathrm{Se}+2e^{-},$$5 in $$^{76}\mathrm{Ge}\rightarrow {}^{76}\mathrm{Se}+2e^{-},$$6Ge; background goal $$^{76}\mathrm{Ge}\rightarrow {}^{76}\mathrm{Se}+2e^{-},$$7 counts/(FWHM$$^{76}\mathrm{Ge}\rightarrow {}^{76}\mathrm{Se}+2e^{-},$$8t$$^{76}\mathrm{Ge}\rightarrow {}^{76}\mathrm{Se}+2e^{-},$$9yr)	Discovery sensitivity $$\bigl(T_{1/2}^{0\nu}\bigr)^{-1}=G^{0\nu}\,\bigl|M^{0\nu}\bigr|^2\,\frac{m_{\beta\beta}^2}{m_e^2},$$0 yr at 99.7% CL in the preconceptual design (Collaboration et al., 2021)

The phased structure is not merely logistical. It enables progressive validation of detector mass scaling, background rejection, radiopurity control, and cryogenic instrumentation before committing to the ton-scale configuration.

3. LEGEND-200 experimental configuration

LEGEND-200 operates at the Laboratori Nazionali del Gran Sasso. In first-year operation, the installed detector mass was 142.5 kg in 101 detectors arranged in 10 vertical strings, with 86.7 kg of inverted-coaxial point-contact detectors, 22.1 kg of p-type point-contact detectors, 19.0 kg of BEGe detectors, and 14.7 kg of coaxial detectors, enriched to 86–92% in $$\bigl(T_{1/2}^{0\nu}\bigr)^{-1}=G^{0\nu}\,\bigl|M^{0\nu}\bigr|^2\,\frac{m_{\beta\beta}^2}{m_e^2},$$1Ge (Saleh, 13 Mar 2026). Other overview papers summarize the phase as a $$\bigl(T_{1/2}^{0\nu}\bigr)^{-1}=G^{0\nu}\,\bigl|M^{0\nu}\bigr|^2\,\frac{m_{\beta\beta}^2}{m_e^2},$$2 kg enriched-Ge deployment in a low-background cryostat, reflecting the program-scale target rather than the initially installed mass (Acharya et al., 15 May 2025).

The apparatus is built around bare HPGe detectors immersed in liquid argon. The liquid argon volume is 64 m$$\bigl(T_{1/2}^{0\nu}\bigr)^{-1}=G^{0\nu}\,\bigl|M^{0\nu}\bigr|^2\,\frac{m_{\beta\beta}^2}{m_e^2},$$3 at about 88 K and provides both cooling and passive $$\bigl(T_{1/2}^{0\nu}\bigr)^{-1}=G^{0\nu}\,\bigl|M^{0\nu}\bigr|^2\,\frac{m_{\beta\beta}^2}{m_e^2},$$4 shielding. Scintillation light from the argon is read out by silicon photomultipliers coupled to wavelength-shifting fibers coated with TPB, forming an active veto against events depositing energy in the argon outside the germanium diodes. The cryostat is itself surrounded by a 590 m$$\bigl(T_{1/2}^{0\nu}\bigr)^{-1}=G^{0\nu}\,\bigl|M^{0\nu}\bigr|^2\,\frac{m_{\beta\beta}^2}{m_e^2},$$5 ultra-pure water tank instrumented with 63 PMTs, serving both as passive neutron/$$\bigl(T_{1/2}^{0\nu}\bigr)^{-1}=G^{0\nu}\,\bigl|M^{0\nu}\bigr|^2\,\frac{m_{\beta\beta}^2}{m_e^2},$$6 shielding and as an active Cherenkov muon veto. The underground location at LNGS, under $$\bigl(T_{1/2}^{0\nu}\bigr)^{-1}=G^{0\nu}\,\bigl|M^{0\nu}\bigr|^2\,\frac{m_{\beta\beta}^2}{m_e^2},$$7 m rock or $$\bigl(T_{1/2}^{0\nu}\bigr)^{-1}=G^{0\nu}\,\bigl|M^{0\nu}\bigr|^2\,\frac{m_{\beta\beta}^2}{m_e^2},$$8 m.w.e., reduces the cosmic-muon flux by $$\bigl(T_{1/2}^{0\nu}\bigr)^{-1}=G^{0\nu}\,\bigl|M^{0\nu}\bigr|^2\,\frac{m_{\beta\beta}^2}{m_e^2},$$9 to about $G^{0\nu}$0 muons/(m$G^{0\nu}$1h) (Saleh, 13 Mar 2026).

Detector geometry is central to the design. LEGEND inherited PPC and BEGe technologies and introduced large-mass inverted-coaxial point-contact detectors, which combine low capacitance and strong pulse-shape discrimination with larger single-detector mass. In the design literature, these detectors are described as combining large mass, typically $G^{0\nu}$2–$G^{0\nu}$3 kg per detector in LEGEND-200 planning, with excellent pulse-shape discrimination (Acharya et al., 15 May 2025).

Background mitigation is multi-layered. Principal background sources identified in program documents include $G^{0\nu}$4K from $G^{0\nu}$5Ar in liquid argon, U/Th-chain $G^{0\nu}$6 rays, degraded surface $G^{0\nu}$7 events, and cosmogenic isotopes such as $G^{0\nu}$8Ge and $G^{0\nu}$9Ge (Collaboration et al., 2017). Suppression methods include the active liquid-argon veto, the water-Cherenkov muon veto, point-contact pulse-shape discrimination, detector multiplicity cuts, ultra-clean materials, and nylon mini-shrouds around strings to limit $M^{0\nu}$0K deposition on detector surfaces (López-Castaño et al., 2019, Collaboration et al., 2017). The resulting architecture is explicitly optimized for rare single-site energy depositions at $M^{0\nu}$1.

4. Signal readout, calibration, and event reconstruction

The LEGEND-200 signal chain was designed under simultaneous low-noise and radiopurity constraints. The front-end is a resistive-feedback charge-sensitive amplifier split into two stages. The first stage, the Low Mass Front End (LMFE), sits only a few millimeters from each HPGe crystal and operates immersed in liquid argon. It consists of an in-die JFET (Moxtek MX11), an amorphous-Ge thin-film feedback resistor with $M^{0\nu}$2–$M^{0\nu}$3 G$M^{0\nu}$4 at 87 K, and patterned stray-trace capacitances on an ultra-clean Suprasil substrate with titanium-gold traces. The second stage is a low-noise differential amplifier located 30–150 cm away, based on the GERDA design and built from surface-mount components on clean Kapton circuit boards. Signals exit the cryostat on four low-mass picocoax cables and are transmitted over a $M^{0\nu}$5 m differential line to the digitizers (Willers, 2019).

The electronics targets were defined in terms of energy resolution, timing, linearity, and radiopurity. For LEGEND-200, the electronic-noise goal was an equivalent noise charge corresponding to $M^{0\nu}$6 keV FWHM, with bench tests around $M^{0\nu}$7 keV FWHM at optimized shaping times; the energy-resolution target was $M^{0\nu}$8 keV FWHM at $M^{0\nu}$9 keV; bandwidth extended to $m_e$0 MHz for rise times below 100 ns; and linearity was maintained to at least 10 MeV so that $m_e$1-induced background peaks could be monitored and vetoed (Willers, 2019). The total ENC was modeled through

$m_e$2

and the background index in the ROI through

$m_e$3

making explicit the coupling between readout performance and half-life sensitivity (Willers, 2019).

Digitization and calibration procedures in operational LEGEND-200 were subsequently described in detail. Data are digitized at 62.5 MHz with 16-bit FADCs. The digital signal-processing chain applies pole-zero correction, followed by trapezoidal, cusp, or zero-area cusp shaping, with a linear charge-trapping correction. Weekly $m_e$4Th calibration runs establish the energy scale, and a two-stage calibration corrects long-term non-linearity across multiple peaks from 583.2 keV to 2614.5 keV. The optimized energy reconstruction achieves an exposure-weighted, array-averaged

$m_e$5

while weekly peak-position variations remain below 0.05 keV for energies up to 2614.5 keV (Collaboration et al., 21 May 2026).

Event reconstruction for the first unblinded search combined hardware and software rejection layers. Quality cuts removed pulser signals, forced-trigger events, and periods of hardware instability; Ge-Ge multiplicity and muon-veto cuts enforced single-detector topology and rejected triggers within $m_e$6s of a water-Cherenkov signal; the LAr veto rejected events with $m_e$7 photoelectrons or $m_e$8 SiPM hits in a $m_e$9s window around a Ge trigger; and pulse-shape discrimination used A/E and late-charge variables, with an artificial neural network for coaxial detectors (Saleh, 13 Mar 2026). These procedures implement the experiment’s core assumption that a $^{76}$00 event should appear as a localized single-site energy deposition in one detector with no coincident activity in the surrounding active media.

5. First data, achieved performance, and reported limits

LEGEND-200’s first reported $^{76}$01 search used one year of stable operation from March 2023 to February 2024. The data set comprised 85.5 kg$^{76}$02yr collected, of which 61.0 kg$^{76}$03yr remained after quality and stability cuts for the $^{76}$04 analysis (Saleh, 13 Mar 2026). The first 2025 publication summarized the search as based on 61 kg yr, with over half of the exposure coming from the highest performing detectors, including newly developed inverted-coaxial detectors (Acharya et al., 15 May 2025).

After all cuts, 11 events remained in the unblinded $^{76}$05 keV window: 7 in the “golden” data set from ICPC/BEGe/PPC detectors and 4 in the “silver” data set from coaxial/Ortec ICPC detectors. The measured background indices in the ROI were

$^{76}$06

and

$^{76}$07

with the highest-performing detectors also described as having an estimated background level of $^{76}$08 cts/(keV ton yr) in the signal region (Saleh, 13 Mar 2026, Acharya et al., 15 May 2025).

The statistical inference was a frequentist unbinned profile-likelihood fit in the energy window [1930–2190] keV, excluding the known $^{76}$09 lines at $^{76}$10 keV from $^{76}$11Tl and $^{76}$12 keV from $^{76}$13Bi (Saleh, 13 Mar 2026). No signal peak was observed at $^{76}$14. For LEGEND-200 alone, the first-year analysis set

$^{76}$15

at 90% CL, with a median exclusion sensitivity of $^{76}$16 yr (Saleh, 13 Mar 2026). A combined fit with GERDA and the MAJORANA DEMONSTRATOR yielded a 90% CL exclusion sensitivity of $^{76}$17 yr and an observed lower limit

$^{76}$18

with the corresponding effective-mass constraint quoted as $^{76}$19–$^{76}$20 meV, depending on the adopted nuclear matrix element (Acharya et al., 15 May 2025).

Two interpretive clarifications are important. First, “LEGEND-200” denotes the phase target and infrastructure configuration, not the mass present in the first physics data set; the first-year result used 142.5 kg installed and 61.0 kg$^{76}$21yr selected exposure (Saleh, 13 Mar 2026). Second, the widely quoted $^{76}$22 yr limit is the combined GERDA + MAJORANA Demonstrator + LEGEND-200 result, whereas the first-year LEGEND-200-only limit was $^{76}$23 yr (Acharya et al., 15 May 2025, Saleh, 13 Mar 2026). This distinction is methodological rather than controversial, but it is essential for correct comparison across publications.

6. LEGEND-1000 design goals and ongoing R&D

LEGEND-1000 is the ton-scale continuation of the program. In the preconceptual design report, it consists of 1000 kg of germanium detectors enriched to more than 90% in $^{76}$24Ge, operated in a liquid-argon active shield at a deep underground laboratory. The design groups the detectors into 120 strings divided into four identical cryostat modules of about 250 kg each, immersed in a common liquid-argon bath of about 70 m$^{76}$25 at 87 K, with an external 2 m thick water tank instrumented with PMTs for muon tagging (Collaboration et al., 2021). The target total background index is $^{76}$26 cts/(FWHM$^{76}$27t$^{76}$28yr), the energy-resolution goal is FWHM $^{76}$29 keV at 2039 keV, and the projected discovery sensitivity is

$^{76}$30

at 99.7% CL in 10 yr of live time, corresponding to an effective Majorana mass range of 9–21 meV (Collaboration et al., 2021).

Meeting that goal requires further reductions in near-detector mass and radioactivity. One R&D direction is ASIC-based cold readout. Tests of the XGLab CUBE preamplifier as a single-chip charge-sensitive amplifier for LEGEND-1000 reported a chip size of $^{76}$31, mass $^{76}$32 mg, intrinsic rise times as fast as 15 ns with no detector load, baseline noise of 655 eV FWHM at $^{76}$33s, and an energy resolution of 2.3 keV FWHM at $^{76}$34 keV. Assay-based background estimates gave a combined contribution below $^{76}$35 counts/(keV$^{76}$36kg$^{76}$37yr), well below the $^{76}$38 design goal, while A/E-based pulse-shape discrimination retained a 90% DEP acceptance and reduced the SEP survival fraction to $^{76}$39 (Edzards et al., 2020).

Another R&D direction addresses the $^{76}$40Ar/$^{76}$41K background in liquid argon. A GEANT4 study of detector encapsulation with ultra-pure plastic, especially PEN, found that for 3.525 MeV electrons originating on the PEN surface, the residual background fraction relative to an unshielded configuration fell to 0.5% for a 2 mm shell, 0.1% for a 4 mm shell, and $^{76}$42 for an 8 mm shell (Mirza, 2022). Program documents also identify underground-sourced argon, improved LAr light collection, large-mass ICPC-only deployments, and continued materials assay as central to LEGEND-1000 background control (Brugnera, 17 Jan 2025, Collaboration et al., 2021).

Taken together, these developments show that LEGEND is not a single apparatus but a coordinated experimental program. Its present phase has established stable operation, sub-3-keV resolution, and background levels near the design trajectory for a germanium-based $^{76}$43 search, while its future phase is structured around extending those same performance principles—radiopurity, single-site discrimination, active vetoing, and narrow ROI spectroscopy—to the tonne scale.