LEGEND-200: Neutrinoless Beta Decay Experiment

• LEGEND-200 is the first operational phase of a germanium-based experiment that searches for neutrinoless double beta decay, indicating Majorana neutrino properties.
• It employs bare, enriched high-purity germanium detectors in a liquid argon cryostat at LNGS to achieve high energy resolution and effective background suppression.
• The initial one-year results set a half-life lower limit of >0.5×10^26 yr, paving the way for improved techniques and the future tonne-scale LEGEND-1000 phase.

Searching arXiv for LEGEND-200 and related technical papers. LEGEND-200 is the first operational phase of the Large Enriched Germanium Experiment for Neutrinoless $\beta\beta$ Decay, a staged $^{76}$Ge program that combines the low-background methods of GERDA with the detector-development and radiopurity practices of the MAJORANA Demonstrator. It searches for neutrinoless double beta decay by operating enriched high-purity germanium detectors as bare diodes in instrumented liquid argon at the Laboratori Nazionali del Gran Sasso, and its first blinded physics result established the initial performance and half-life reach of the LEGEND program (Saleh, 25 Sep 2025).

1. Physics objective and decay signature

LEGEND-200 is designed to search for the decay

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

with

$Q_{\beta\beta}=2039.061(7)\ \mathrm{keV}.$

In the Standard Model, ordinary two-neutrino double beta decay is allowed, whereas neutrinoless double beta decay is forbidden. Observation of the neutrinoless mode would therefore demonstrate that neutrinos are Majorana fermions and establish lepton-number violation by two units, $\Delta L = 2$ (Saleh, 25 Sep 2025).

In the light-Majorana-exchange picture, the inverse half-life is written as

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

so an experimental limit on $T_{1/2}^{0\nu}$ constrains the effective Majorana mass $m_{\beta\beta}$ (Romo-Luque, 26 Aug 2025). The process is also discussed in connection with leptogenesis and the matter–antimatter asymmetry in the Universe (D'Andrea, 2019).

The germanium approach is central to the LEGEND strategy because the detector material is simultaneously source and detector, the isotope can be enriched to high $^{76}$Ge fraction, and HPGe technology provides energy resolution around $0.1\%$ FWHM at $^{76}$0 together with pulse-shape sensitivity to event topology (Saleh, 25 Sep 2025).

2. Position within the LEGEND roadmap

LEGEND is a two-stage program. LEGEND-200 is the current-generation phase, while LEGEND-1000 is the later tonne-scale phase. The design goal for LEGEND-200 is a $^{76}$1 discovery sensitivity for a $^{76}$2 half-life of about $^{76}$3 yr after $^{76}$4 ton yr exposure at a target background index of

$^{76}$5

whereas LEGEND-1000 aims at sensitivity beyond $^{76}$6 yr (Saleh, 25 Sep 2025).

This program was formulated by explicitly building on GERDA and the MAJORANA Demonstrator. GERDA had already demonstrated background-free operation at current exposures with a background index near $^{76}$7 of about $^{76}$8 cts/(keV·kg·yr), while the MAJORANA Demonstrator achieved an energy resolution of $^{76}$9 keV FWHM at $$^{76}\mathrm{Ge} \rightarrow {}^{76}\mathrm{Se} + 2e^-,$$0 keV (D'Andrea, 2019). A later review framed LEGEND-200 as a 200 kg deployment in the GERDA facility at LNGS with a target sensitivity of about $$^{76}\mathrm{Ge} \rightarrow {}^{76}\mathrm{Se} + 2e^-,$$1 yr at 90% C.L. with 1 t·yr exposure and a background reduction of roughly a factor of three beyond GERDA’s final level (D'Andrea et al., 2021).

A major enabling development for LEGEND-200 is the inverted-coaxial point-contact detector. In GERDA, five enriched IC detectors manufactured from material enriched to 87.7(5)% in $$^{76}\mathrm{Ge} \rightarrow {}^{76}\mathrm{Se} + 2e^-,$$2Ge reached resolutions at $$^{76}\mathrm{Ge} \rightarrow {}^{76}\mathrm{Se} + 2e^-,$$3 of about 2.1 keV FWHM in vacuum cryostat, and after 18 months of operation in liquid argon achieved a background index after analysis cuts of $$^{76}\mathrm{Ge} \rightarrow {}^{76}\mathrm{Se} + 2e^-,$$4 counts/(keV·kg·yr) around $$^{76}\mathrm{Ge} \rightarrow {}^{76}\mathrm{Se} + 2e^-,$$5, establishing their feasibility for LEGEND (Collaboration et al., 2021).

3. Apparatus and detector concept

After a commissioning period with 60 kg of Ge, LEGEND-200 began physics data taking in March 2023 at LNGS with a total Ge mass of 142.5 kg. In the first deployment, four detector geometries were used: 6 Coax and 28 BEGe detectors inherited from GERDA, 26 PPC detectors from the MAJORANA Demonstrator, and 41 newly produced inverted coaxial point-contact detectors. The detectors are enriched in $$^{76}\mathrm{Ge} \rightarrow {}^{76}\mathrm{Se} + 2e^-,$$6Ge beyond 86%, with enrichment fractions between 86% and 92% depending on detector type, and crystal masses range from roughly $$^{76}\mathrm{Ge} \rightarrow {}^{76}\mathrm{Se} + 2e^-,$$7 to $$^{76}\mathrm{Ge} \rightarrow {}^{76}\mathrm{Se} + 2e^-,$$8 (Saleh, 25 Sep 2025).

The detectors are operated as bare diodes in a $$^{76}\mathrm{Ge} \rightarrow {}^{76}\mathrm{Se} + 2e^-,$$9 liquid argon cryostat at about $Q_{\beta\beta}=2039.061(7)\ \mathrm{keV}.$0. The argon acts simultaneously as coolant, passive shield, and active veto medium. The active veto uses wavelength-shifting fiber barrels read out by SiPMs to detect argon scintillation light and provide Ge–LAr anticoincidence. Outside the cryostat is a $Q_{\beta\beta}=2039.061(7)\ \mathrm{keV}.$1 ultra-pure water tank instrumented with PMTs, which provides additional passive shielding and functions as a Cherenkov muon veto. The LNGS overburden of 1400 m of rock, corresponding to almost 3500 m water equivalent, suppresses the cosmic muon flux to about $Q_{\beta\beta}=2039.061(7)\ \mathrm{keV}.$2 muons/$Q_{\beta\beta}=2039.061(7)\ \mathrm{keV}.$3 (Saleh, 25 Sep 2025).

The liquid-argon light instrumentation is itself a dedicated subsystem. Its front-end electronics, installed in July 2021, were designed to read out the SiPM response to argon scintillation while satisfying stringent radiopurity constraints and achieved a very low overall level of electrical noise of $Q_{\beta\beta}=2039.061(7)\ \mathrm{keV}.$4V peak-to-peak (Costa et al., 2022). Near the detectors, further background control is implemented through low-background PEN support structures: molded PEN components were qualified as active holders with a light yield of $Q_{\beta\beta}=2039.061(7)\ \mathrm{keV}.$5 and an estimated self-veto threshold of roughly 50–250 keV depending on position (Manzanillas et al., 2022).

Support infrastructure around the detector array was also optimized for calibration and materials control. Custom low-neutron-emission $Q_{\beta\beta}=2039.061(7)\ \mathrm{keV}.$6Th calibration sources were produced for LEGEND-200, with a measured neutron flux of

$Q_{\beta\beta}=2039.061(7)\ \mathrm{keV}.$7

approximately one order of magnitude below that of commercial sources, making the neutron-induced background rate negligible compared to other background sources in LEGEND-200 (Baudis et al., 2022). Detector production for the program required dedicated enriched-germanium processing; one campaign reported hydrogen reduction of a batch of 23 kg isotopically enriched Ge with an average yield of 99.85%, followed by zone refining yielding an overall Ge yield of 99.05% (Gradwohl et al., 2020).

4. Calibration, reconstruction, and blinded analysis

LEGEND-200 relies on weekly $Q_{\beta\beta}=2039.061(7)\ \mathrm{keV}.$8Th source calibrations to monitor the energy scale, resolution, and detector stability. A dedicated calibration-and-performance analysis reported a combined average resolution of $Q_{\beta\beta}=2039.061(7)\ \mathrm{keV}.$9 keV at

$\Delta L = 2$0

with weekly variation of calibration peak positions below $\Delta L = 2$1 keV for energies up to 2614.5 keV. The same study reported explicit corrections for residual non-linearities and energy bias in the region of interest (Collaboration et al., 21 May 2026).

The first $\Delta L = 2$2 search used a strictly blind analysis in the region

$\Delta L = 2$3

For statistical extraction of the background and the half-life limit, the fit window was

$\Delta L = 2$4

excluding the known $\Delta L = 2$5 lines at $\Delta L = 2$6 keV from $\Delta L = 2$7 and $\Delta L = 2$8 keV from $\Delta L = 2$9 (Saleh, 25 Sep 2025).

Event selection combined multiple rejection layers. The muon veto and Ge-detector multiplicity cut each have efficiencies above 99.9% at $$(T_{1/2}^{0\nu})^{-1} = G^{0\nu} |M^{0\nu}|^2 \frac{m_{\beta\beta}^2}{m_e^2},$$0. The liquid-argon veto reaches

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

and pulse-shape discrimination has

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

depending on detector type. This sequence strongly suppresses Compton-scattered $$(T_{1/2}^{0\nu})^{-1} = G^{0\nu} |M^{0\nu}|^2 \frac{m_{\beta\beta}^2}{m_e^2},$$3 backgrounds and other non-signal topologies (Saleh, 25 Sep 2025).

The data were partitioned into two categories with different expected background conditions. The “golden dataset” comprises 48.3 kg yr from BEGe, PPC, and Mirion-produced ICPC detectors. The “silver dataset” comprises 12.7 kg yr from Coax and ORTEC-produced ICPC detectors (Saleh, 25 Sep 2025).

Offline waveform-quality control is also an explicit component of the LEGEND-200 analysis chain. A semi-supervised Affinity Propagation plus Support Vector Machine method developed on the Full Chain Test is reported as being used to accelerate data-cleaning development for LEGEND-200 and as the main data cleaning method in the Juleana secondary software stack; the quoted maximum sacrifice of physics events is

$$(T_{1/2}^{0\nu})^{-1} = G^{0\nu} |M^{0\nu}|^2 \frac{m_{\beta\beta}^2}{m_e^2},$$4

(León et al., 2024)

5. First-year physics result

In about one year of running, LEGEND-200 accumulated a total physics exposure of 61 kg yr usable for the $$(T_{1/2}^{0\nu})^{-1} = G^{0\nu} |M^{0\nu}|^2 \frac{m_{\beta\beta}^2}{m_e^2},$$5 analysis (Saleh, 25 Sep 2025). After all cuts, 11 events remained in the analysis window: 7 in the golden dataset and 4 in the silver dataset. The measured background indices were

$$(T_{1/2}^{0\nu})^{-1} = G^{0\nu} |M^{0\nu}|^2 \frac{m_{\beta\beta}^2}{m_e^2},$$6

for the golden dataset and

$$(T_{1/2}^{0\nu})^{-1} = G^{0\nu} |M^{0\nu}|^2 \frac{m_{\beta\beta}^2}{m_e^2},$$7

for the silver dataset (Saleh, 25 Sep 2025).

The statistical interpretation was carried out in a frequentist framework, with a Bayesian analysis reported as giving compatible results. No evidence for a $$(T_{1/2}^{0\nu})^{-1} = G^{0\nu} |M^{0\nu}|^2 \frac{m_{\beta\beta}^2}{m_e^2},$$8 signal was found, and the first LEGEND-200 lower limit was

$$(T_{1/2}^{0\nu})^{-1} = G^{0\nu} |M^{0\nu}|^2 \frac{m_{\beta\beta}^2}{m_e^2},$$9

at 90% C.L. The median exclusion sensitivity was

$T_{1/2}^{0\nu}$0

Using ranges of phenomenological nuclear matrix elements, the corresponding upper limit on the effective Majorana mass was quoted as

$T_{1/2}^{0\nu}$1

(Saleh, 25 Sep 2025)

The collaboration also performed a combined analysis with the preceding germanium-based experiments. A joint fit to GERDA (127.2 kg yr), MAJORANA Demonstrator (64.5 kg yr), and LEGEND-200 (61 kg yr) found no signal and yielded

$T_{1/2}^{0\nu}$2

at 90% C.L., with a median exclusion sensitivity of

$T_{1/2}^{0\nu}$3

A conference-style summary emphasized that this combined result provided the best exclusion sensitivity yet achieved in any $T_{1/2}^{0\nu}$4 search (Saleh, 25 Sep 2025, Romo-Luque, 26 Aug 2025).

6. Background excess, detector subsets, and longer-term significance

A central technical result of the first year was that the observed background was somewhat higher than expected from prior radioassay-based predictions, including in the region around $T_{1/2}^{0\nu}$5. According to the first-results paper, the background-model fit to the less selective spectrum suggested that the excess could be attributable to contamination from the $T_{1/2}^{0\nu}$6 chain. This motivated a screening and cleaning campaign and a redeployment of the array in May 2025, with post-redeployment data expected to verify the background reduction (Saleh, 25 Sep 2025).

The two analysis categories encapsulate an important detector-performance distinction. The lower background in the golden dataset reflects the better-performing detector classes and cleaner subset, while the silver dataset contains the detector populations associated with a higher residual rate (Saleh, 25 Sep 2025). The same pattern appears in energy performance: BEGe, PPC, and ICPC detectors broadly meet the LEGEND energy-resolution goal, whereas the older coaxial detectors have poorer resolution, as already known from GERDA, and are not expected to be relevant for LEGEND-1000, where only ICPC detectors are planned (Saleh, 25 Sep 2025).

In programmatic terms, LEGEND-200 is both a stand-alone search and the technological bridge to the later tonne-scale phase. Earlier design papers defined it as a 200 kg deployment in reused GERDA infrastructure, with MAJORANA electroformed copper, an upgraded liquid-argon veto, improved front-end electronics, and new ICPC detectors as the main ingredients for reaching the $T_{1/2}^{0\nu}$7-yr sensitivity class (D'Andrea, 2019). The first-year result did not yet realize the ultimate LEGEND-200 background target, but it validated stable low-background operation of enriched HPGe detectors in instrumented liquid argon at substantially increased scale and placed the experiment in direct continuity with the cumulative $T_{1/2}^{0\nu}$8Ge program (Saleh, 25 Sep 2025).

LEGEND-200 therefore occupies a precise place in the contemporary neutrinoless double beta decay landscape. It inherits GERDA’s bare-detector-in-LAr architecture, MAJORANA’s ultra-clean detector and component development, and the ICPC scaling path established in precursor studies, while producing its own first half-life limit and strengthening the combined $T_{1/2}^{0\nu}$9Ge constraint. Its technical trajectory remains defined by the same coupled quantities that structured the original roadmap: exposure, energy resolution, signal efficiency, and above all background index.