space · manufacturing · family: making one is easy. making a million is the problem
seventy thousand tonnesof liquid argon, underground
Building the world's largest liquid argon detector underground requires cryogenic engineering at unprecedented scale
Problem statement
The Deep Underground Neutrino Experiment (DUNE), one of the three construction projects the P5 report places in its highest-priority recommendation, requires building four liquid argon time projection chamber (LArTPC) modules, each containing about 17,000 tonnes of liquid argon (at -186°C), installed 1.5 km underground at the Sanford Underground Research Facility in South Dakota (the 2023 P5 report stages this as two modules in a first phase, a third in a second phase, and research toward a fourth). No detector of this scale has ever been built. The engineering challenges are threefold: (1) maintaining argon purity at <100 parts per trillion of oxygen-equivalent contaminants across the full volume, because even trace electronegative impurities capture the drifting ionization electrons that form the detector signal; (2) constructing wire planes or printed circuit board charge readout planes 12 m by roughly 60 m (the wire version is 12.0 m high and 58.2 m long, built from 6 m × 2.3 m anode plane assemblies with wire-plane spacing held to ±0.5 mm, from stainless-steel frames and beryllium-copper wires that contract by 2.7 and 2.9 mm/m respectively during cooldown); and (3) operating 384,000 readout channels per module at -186°C for 20 years with essentially zero access for maintenance, since the detector modules cannot be opened once filled.
Why this matters
DUNE will determine whether neutrinos and antineutrinos oscillate differently — a CP violation measurement that could help explain why the universe contains matter rather than equal amounts of matter and antimatter, one of the most fundamental open questions in physics. DUNE will also detect neutrinos from core-collapse supernovae throughout the Milky Way, providing a real-time window into the physics of stellar death, and search for proton decay in the p → K⁺ν̄ channel (the mode the 2020 design report's studies focus on, since a liquid argon TPC can track and identify kaons) with a sensitivity roughly twice the current Super-Kamiokande limit after a 400 kt·year exposure. The collaboration counts more than 1,000 scientists and engineers from 31 countries (2020 design report). The P5 report places completing DUNE's first phase in its highest-priority recommendation, under the "Elucidate the Mysteries of Neutrinos" driver in its "Decipher the Quantum Realm" section. If the detector technology challenges are not solved, the physics program is compromised proportionally — argon purity directly determines signal quality, and readout channel reliability determines how much of the detector is still working across the 20-year program.
What’s been tried and why it hasn’t worked
LArTPC technology was invented by Carlo Rubbia in 1977, and detectors of increasing size have been built: ICARUS (operated at Gran Sasso, Italy), MicroBooNE (operating at Fermilab when the 2020 design report was written), SBND (under construction at Fermilab at that time). ProtoDUNE — two prototype modules at CERN, ProtoDUNE-SP (single-phase, 419 tonnes of instrumented argon) and ProtoDUNE-DP (dual-phase) — demonstrated both readout designs at roughly 1/20 of a full DUNE module (the P5 report's glossary puts the prototypes at roughly one-tenth). ProtoDUNE-SP exceeded the required electron lifetime (the requirement is >3 ms, corresponding to <100 ppt O₂-equivalent contamination; ProtoDUNE-SP reached >20 ms, about 3.4 ppt) with the largest argon purification system built to date, but what the sources report stops at prototype scale or is still in progress: the purification pump turned over ProtoDUNE-SP's argon in about 4.5 days, and a full module is about twenty times the size; cold electronics (ASICs operating at -186°C) met their noise specifications with 0.2% dead channels, and the observed loss rate extrapolates to at most 0.5% of channels over 20 years against a 1% requirement, but the design report says the effect of failures that take out whole ASICs or boards (16 to 128 channels at once) is still being studied and that further ProtoDUNE-SP and SBND running will provide more information on the long-term stability of electronics immersed in argon; and the anode plane assemblies must hold wire-plane spacing to ±0.5 mm while their frames and wires contract by 2.7 to 2.9 mm/m during cooldown, which the design manages by cooling gradually. The engineering challenge is not any single component but the product of multiple demanding requirements simultaneously: extreme purity × cryogenic operation × mechanical precision × long lifetime × no maintenance access × underground construction constraints.
What would unlock progress
Scaled-up argon purification systems demonstrated to maintain <100 ppt O₂-equivalent in 17,000-tonne volumes — specifically, filtration and recirculation systems with redundancy for 20-year operation. Cold ASIC designs with demonstrated channel losses below the 1%-over-20-years requirement at -186°C, validated through accelerated lifetime testing of the kind the design report describes for commercial components, using a NASA-developed methodology. Modular anode plane assembly designs that can be tested, transported, and installed underground with sufficient mechanical tolerance to accommodate cryogenic contraction. Novel photon detection systems (ARAPUCA-type light traps) that provide calorimetric energy measurement independent of the charge readout, creating redundancy against charge collection degradation.
Entry points for student teams
A student team could characterize cold electronics noise and reliability by operating commercial or custom ASIC designs in a liquid nitrogen bath (~77K, accessible with standard cryogenic equipment) over extended thermal cycles, extrapolating the observed failure modes to 384,000 channels per module over the 20-year DUNE lifetime. A second door needs no hardware at all: the DUNE far detector technical design report (Volume IV: Far Detector Single-Phase Technology, arXiv:2002.03010) and the ProtoDUNE-SP performance paper (JINST 15 P12004, arXiv:2007.06722) publish measured electron lifetimes, noise figures, and channel-level performance, which is enough to build either a purity-versus-signal-attenuation model scaled to a 17,000-tonne volume or a channel-reliability extrapolation that tests the assumption the collaboration is currently making. A third door — a small-scale liquid argon purity monitor using UV laser ionization or radioactive source injection, measuring electron drift lifetime as a function of purification system parameters — is the most direct attack on the purity requirement but is not facility-free: it needs a liquid argon cryostat held at part-per-billion purity plus either a UV laser or a licensed radioactive source, access held by national laboratories and by the university particle-physics groups already on LArTPC collaborations. Relevant disciplines: cryogenic engineering, electrical engineering, mechanical engineering, nuclear/particle physics.
Genes
Structural cousins: stuck the same way, in other fields
Sources
"Exploring the Quantum Universe: Pathways to Innovation and Discovery in Particle Physics" (P5 Report), Particle Physics Project Prioritization Panel, 2023. accessed 2026-02-16. Also: DUNE Collaboration, "Deep Underground Neutrino Experiment (DUNE), Far Detector Technical Design Report," Volume I: Introduction to DUNE (), Volume II: DUNE Physics (), Volume IV: Far Detector Single-phase Technology (), 2020; DUNE Collaboration, "First results on ProtoDUNE-SP liquid argon time projection chamber performance from a beam test at the CERN Neutrino Platform," JINST 15 (2020) P12004, go to source 1 ↗ go to source 2 ↗ go to source 3 ↗ go to source 4 ↗ go to source 5 ↗
Working notes
Notes made while writing and checking this brief, shown as written. They use working shorthand.
- DUNE is one of the three construction projects in the P5 report's highest-priority recommendation (Recommendation 1), discussed under the "Elucidate the Mysteries of Neutrinos" driver in the report's "Decipher the Quantum Realm" section.
- The `failure:lab-to-field-gap` tag captures the scale-dependent emergence pattern: ProtoDUNE demonstrated technology at 1/20 scale, but argon purity, electronics reliability, and mechanical behavior may exhibit failure modes that appear only at full scale — the same sub-pattern seen in energy-grid-battery-scale-failure-prediction.
- The `failure:not-attempted` tag reflects that no 17,000-tonne LArTPC has ever been built or operated.
- The `domain:space` primary tag is used because DUNE falls within the physics/space science domain, though the engineering challenges are squarely in manufacturing and cryogenic systems.
- The `temporal:window` tag reflects that DUNE construction is underway at SURF, with the 2020 design report scheduling the first far detector module for 2024–2026 and the remaining modules through 2036. Technology decisions must be made now.
- Cross-domain connection: shares the scale-dependent-failure structure with energy-grid-battery-scale-failure-prediction (lab-scale testing doesn't predict large-system behavior) and the underground-facility-constraint with environment-critical-mineral-deposit-discovery (subsurface access limitations constraining scientific investigation).
- Note reconciled 2026-08-20: a note above argues for `failure:not-attempted`; the genome now carries `failure:lab-to-field-gap` after a taxonomy revision. The original note is kept verbatim as the tagging rationale of record.
Reconciliation 2026-08-21: Entry-point realism pass (panel C37 triage, score 2). The triage flag is upheld: the small-scale liquid argon purity monitor was written as a peer of the liquid-nitrogen arm, but it requires a liquid argon cryostat maintained at part-per-billion purity plus a UV laser or a licensed radioactive source — no facility-free team can start there. The liquid-nitrogen cold-electronics arm was correctly scoped and now leads. The purity monitor is kept, since it is genuinely the most direct attack on the binding constraint, but carries an explicit access line naming what it takes and who holds that access. A new desk door was added on two sources verified live in this session: the DUNE far detector technical design report Volume IV, "Deep Underground Neutrino Experiment (DUNE), Far Detector Technical Design Report, Volume IV: Far Detector Single-Phase Technology," arXiv:2002.03010 (https://arxiv.org/abs/2002.03010, FERMILAB-PUB-20-027-ND), and DUNE Collaboration, "First results on ProtoDUNE-SP liquid argon time projection chamber performance from a beam test at the CERN Neutrino Platform," JINST 15 (2020) P12004, arXiv:2007.06722 (https://arxiv.org/abs/2007.06722) — both open-access preprints, no application or membership required. A dedicated DUNE open-data portal was checked and not found: `opendata.dunescience.org` does not resolve, and a search of the CERN Open Data Portal (https://opendata.cern.ch/search?q=protodune) returns no ProtoDUNE records, so no such portal is named here and the desk door rests on the published papers instead. Genome Tags untouched.
Reconciliation 2026-09-30: Round 7 read-through against the sources. Read in full on 2026-09-30: the P5 report (https://doi.org/10.2172/2368847, resolving to the OSTI PDF at www.osti.gov/servlets/purl/2368847, 166 pages, full text searched); the DUNE Far Detector Technical Design Report Volume I (arXiv:2002.02967), Volume II (arXiv:2002.03005) and Volume IV (arXiv:2002.03010, FERMILAB-PUB-20-027-ND), each fetched as the arXiv PDF and searched in full (Volume III, arXiv:2002.03008, was fetched but not needed); the ProtoDUNE-SP performance paper (arXiv:2007.06722, journal reference JINST 15 (2020) P12004 on the arXiv record), full text; and the CERN Open Data Portal search named in the note above (https://opendata.cern.ch/search?q=protodune, still zero records). Changes: (1) The long title called DUNE the world's largest particle detector. No source says that; the ProtoDUNE-SP paper calls its cryostat and cryogenics "the largest LAr systems ever constructed" and the design report puts a full module at about twenty times a ProtoDUNE, so the title now says the world's largest liquid argon detector (old: "Building the World's Largest Particle Detector Underground Requires Liquid Argon Engineering at Unprecedented Scale"; new: "Building the World's Largest Liquid Argon Detector Underground Requires Cryogenic Engineering at Unprecedented Scale"). The display title, seventy thousand tonnes, is four modules of about 17.5 kt each per Volume I and is unchanged. (2) The text called DUNE "the P5 report's highest-priority construction project" and said P5 "designated DUNE as its top-priority project for the 'Decipher the Quantum Realm' science driver." The P5 report's Recommendation 1, its highest priority, is to complete three construction projects (HL-LHC, DUNE Phase I with PIP-II, and the Rubin Observatory) and does not rank DUNE above the other two; DUNE is discussed under the "Elucidate the Mysteries of Neutrinos" driver inside the report's "Decipher the Quantum Realm" section. Both sentences and the first Source Notes bullet now say that. (3) "Requires building four ... modules" is what the 2020 design report says ("The full DUNE FD requires four modules"), but the 2023 P5 report stages the far detector as two modules in Phase I, a third in Phase II and R&D toward a fourth; a parenthesis now says so. (4) The purity requirement was described as "more stringent than semiconductor-grade gases"; no source makes the comparison, and the clause was removed. (5) The readout planes were said to span "12m × 14m active areas" inside vessels that "contract by ~4mm/m." Volume I gives 12.0 m high, 58.2 m long anode planes built from 6 m by 2.3 m assemblies (the dual-phase anode plane is 12.0 m by 60.0 m); Volume IV gives a ±0.5 mm wire-plane spacing tolerance and thermal contraction to 87 K of 2.7 mm/m for stainless steel and 2.9 mm/m for the beryllium-copper wire. The sentence now carries those figures. (6) "150,000+ readout channels" appeared three times (problem statement, the ProtoDUNE extrapolation and the entry point). Volume IV gives 384,000 TPC channels per single-phase module; all three now say 384,000 per module. "20+ years" became "20 years," the nominal lifetime in Volume IV. (7) Proton decay sensitivity "10× beyond current limits" is not what Volume II says: for p → K⁺ν̄ the projected limit of 1.3 × 10³⁴ years for 400 kt·year is "roughly twice the current limit" from Super-Kamiokande, and for p → e⁺π⁰ DUNE would need 800 kt·year to match Super-Kamiokande's current limit. The sentence now says roughly twice, in the p → K⁺ν̄ channel, which Volume II says its studies focus on because of "the unique ability to track and identify kaons in a LArTPC." (8) "A $3+ billion international investment with contributions from 35 countries": no source gives a cost, and Volume I says "more than 1000 scientists and engineers from 31 countries." The sentence now gives the Volume I figures and no cost. (9) "Readout channel reliability determines detector uptime" was softened to how much of the detector is still working, which is how Volume IV frames the dead-channel requirement (degradation of physics performance, not uptime). (10) The tonnages given for ICARUS (760), MicroBooNE (170) and SBND (260) appear in no source and were removed; SBND is "under construction" in Volume I, not "under commissioning," and the text now says so with the date. (11) ProtoDUNE was described as "two 770-tonne prototype modules" that demonstrated "the novel vertical drift design." Neither 2020 source mentions vertical drift; the second prototype was the dual-phase ProtoDUNE-DP. The paper gives 419 t of instrumented argon for ProtoDUNE-SP (Volume I calls it a 700 t prototype). The sentence now names both prototypes as single-phase and dual-phase, gives the 419 t instrumented mass, and keeps "roughly 1/20" (Volume I: "approximately one-twentieth the size"), noting that the P5 glossary says "roughly one-tenth." (12) The electron lifetime was given as ">10 ms, corresponding to <100 ppt." The requirement is >3 ms, which is what corresponds to <100 ppt O₂-equivalent (Volume IV); ProtoDUNE-SP measured >20 ms, equivalent to 3.4 ± 0.7 ppt (paper, Section 8). (13) The three "challenges identified" by ProtoDUNE were not in the sources as written: no source says the purification flow must scale 20× (the paper gives a pump of about 7 t/hr and a 4.5-day turnover; Volume I gives the twenty-fold size ratio); the cold-electronics failure rate does not extrapolate to "unacceptable signal loss" but to "at most 0.5% of the readout channels" over 20 years against a 1% requirement (Volume IV, Section 4.3), with the open items being failures of whole ASICs or boards and long-term stability to be learned from further ProtoDUNE and SBND running; and no source reports cooldown distortions of the anode plane assemblies (Volume IV discusses the ±0.5 mm tolerance, the contraction figures and gradual cooling to avoid tension spikes). The passage now says what the sources say. (14) "Failure rates below 0.1%/year" and "protocols adapted from space electronics qualification" were replaced by the Volume IV requirement (under 1% of channels over 20 years) and its wording ("accelerated lifetime testing, a methodology developed by NASA," applied to commercial components). (15) "Far detector installation planned for the late 2020s" in the notes now carries Volume I's schedule (first module 2024–2026, remaining modules through 2036). (16) The Source line named the design report volumes and the ProtoDUNE results without addresses; it now carries the arXiv addresses and the titles copied from the arXiv records for Volumes I, II and IV and for the ProtoDUNE-SP paper (the volumes the figures above come from), so that the numbers in the text trace to the pages they were read from. No source was added that the write-up did not already name. Confirmed in the sources and left as written: the Source line citation (title and "Report of the 2023 Particle Physics Project Prioritization Panel" on the OSTI PDF; the four TDR volumes dated 2020 on arXiv); 1.5 km underground at SURF in South Dakota; about 17,000 tonnes per module (Volume I: 17,000,000 kg; 17.5 kt total); -186°C (87 K); <100 ppt O₂-equivalent and the electron-capture mechanism; no access once the cryostat is filled; the CP-violation and matter-antimatter framing and the Milky Way core-collapse supernova window (P5); Rubbia 1977 as the LArTPC origin (Volume I reference 43); ICARUS at Gran Sasso, MicroBooNE at Fermilab; ARAPUCA photon detectors with a light response linear in energy (paper); the two desk-door papers and their identifiers; and the liquid-nitrogen and purity-monitor entry points as described. Not checked because no source addresses them: the "real-time" wording for supernova detection and the redundancy language in the unlock section. A second reader went over these repairs on the same day (2026-09-30), fetched the same sources independently (the P5 report through the Internet Archive copy of the OSTI PDF, the three TDR volumes and the ProtoDUNE-SP paper from arXiv) and confirmed each change against them; none was undone. Five passages were reworded to stay closer to the source wording: the readout-plane sentence now separates the wire version (12.0 m high, 58.2 m long, stainless-steel frames and beryllium-copper wire) from the dual-phase anode plane (12.0 m by 60.0 m) and gives the contraction as 2.7 and 2.9 mm/m rather than "about 3 mm/m"; the proton-decay sentence names the kaon mode as the one the design report's studies focus on rather than DUNE's "best channel"; the ProtoDUNE passage says a full module is about twenty times the size (Volume I: "approximately one-twentieth the size") rather than holding twenty times the argon, and that further ProtoDUNE-SP and SBND running "will provide additional information" on long-term stability rather than that stability resting on it; the three open items are introduced as what the sources report stopping at prototype scale or being still in progress, rather than as open questions the sources themselves pose; and MicroBooNE's operating status is dated to the 2020 design report as SBND's is. Genome Tags untouched.