Experiments For CP-Violation: A Giant Liquid Argon Scintillation, Cerenkov And Charge Imaging Experiment ?

Worth reading for the design synthesis, not for the feasibility proof. What is actually here: a 2004 Venice proceedings that pulls LAr TPC imaging, Cerenkov and scintillation readout, bi-phase charge amplification, and LNG tanker engineering into one concrete 100 kton single-tank concept. The genuinely useful parts are the water-versus-LAr comparison in Table 1 and the pion/muon separation estimate in Fig. 10 — the latter is a real number for beta-beam backgrounds, and the observation that a proton below water-Cerenkov threshold still ranges 80 cm in LAr is a clean way to say why imaging beats rings. The cryogenic bookkeeping in Tables 5–7 is honest: 60 kW heat load, 0.03% per day boiloff, 30 MW to fill in two years. The LNG analogy is the right one, and the author went to an actual tanker firm about it. Where it is soft: the stress-test note is right and the reader's framing is slightly off. The 2 ms lifetime is cited to published ICARUS work (ref. [25]), so that link holds. The gain is the hole. Section 7.3 says only that 'we have experimentally obtained encouraging results which show that a gain of 100–1000 is achievable' — no reference, no measured curve, no spark rate, and the acknowledgements show it is the author's own unpublished work with Picchi, Pietropaolo and Messina. Combined with the 1/150 attenuation across 20 m, that is not just 'needs R&D'; it is a uniformity and dynamic-range specification on 100,000 channels that nobody has demonstrated. The paper is candid about its other open items — the cavern study is 'under study', the 70 kton magnetized version is called an 'unsolved challenge', R&D results are 'expected in 2004' — so this is not overclaiming, it is one unhedged sentence in an otherwise careful document. Minor but real: Table 1 gives LAr scintillation as roughly 5×10³ photons per MeV while Section 7.3 says about 50,000; Wγ = 19.5 eV gives about 51,000, so Table 1 is off by a factor of ten, conservatively. Who it is for: people scoping the next giant detector, and anyone who wants the LAr case stated in one place with numbers attached. Send it to referees — the gain claim and the Table 1 number are both checkable and should be checked.

The paper's feasibility argument for a single 100 kton volume rests on the chain in §7.3 and Table 4: 20 m drift at 1 kV/cm (2 MV cathode), electron lifetime τ ≈ 2 ms, attenuation e^(−t_max/τ) ≈ 1/150, compensated by bi-phase extraction and proportional gain of 10^2–10^3 in pure argon. Two links in this chain have very different support. The lifetime is cited to a published ICARUS result (ref. [25], NIM A516 (2004) 68–79), so that number is supported. The gain is not: §7.3 states only 'we have experimentally obtained encouraging results which show that a gain of 100–1000 is achievable', with no reference, no measured gain curve, no spark-rate or stability data, and the acknowledgements indicate this is the authors' own unpublished work with Picchi, Pietropaolo and Messina. The paper itself notes the difficulty — amplification must run in pure argon because 'the addition of a quenching gas is not practical since it would pollute the liquid' — which is precisely the regime where breakdown and sparking are hardest. The consequence is sharper than a generic 'not yet demonstrated': because attenuation is a factor 150 across the volume, a minimum-ionizing track near the cathode arrives with ~1/150 of the charge of one near the anode, so the gain must be uniform to better than that margin and the front-end noise must keep the far-end signal above threshold at 3 mm pitch over 100,000 channels. A single spark at gain 10^3 can also disable a channel. If this fails, the fallback is the modular ICARUS-style approach the paper already calls feasible, so the specific single-tank claim — not the whole LAr programme — is what is at stake. A secondary internal inconsistency worth noting: Table 1 lists LAr scintillation as '≈ 5 × 10^3' γ/MeV while §7.3 states 'about 50000 γ/MeV'; the latter matches Wγ = 19.5 eV, so Table 1 appears to be off by a factor of ten (in the conservative direction).

The reader's weakest_assumption identifies the right link in the argument: the 20 m drift combined with 2 ms lifetime and bi-phase gain of 100–1000 in §7.3 and Table 4. I agree with that placement and with the CONDITIONAL verdict, and I would not move it. My refinement is about where inside that link the support actually runs out. The lifetime is a cited, published ICARUS measurement, so under the usual standard it counts as supported; the gain is not cited at all and is described only as 'encouraging results', which is the one place in the feasibility chain where an experimental number is asserted without a source. That is a narrower and more actionable concern than the reader's framing, which treats lifetime and gain together as one unvalidated premise. I also add the quantitative consequence the reader did not spell out: a 150× attenuation across the volume is not just a gain requirement but a dynamic-range and noise requirement on every one of ~100,000 channels, and the paper does not present that budget. On the rest of the paper I find no circularity and no internal contradiction in the drift arithmetic: 20 m × 1 kV/cm = 2 MV, v_d ≈ 2 mm/µs gives 10 ms, e^(−10/2) ≈ 1/148 matches the quoted 1/150, and σ_D = sqrt(2·4 cm²/s·10 ms) ≈ 2.8 mm matches the 3 mm pitch, so the stated 20 m ceiling on drift is internally coherent. The cavern siting and the magnetized 70 kton variant are explicitly flagged as under study or unsolved in the paper, so those are acknowledged limitations rather than new findings. The one genuine internal inconsistency I found is the factor-of-ten discrepancy between Table 1 ('≈ 5 × 10^3' γ/MeV) and §7.3 ('about 50000 γ/MeV'); the §7.3 value is the one consistent with Wγ = 19.5 eV, so Table 1 should be corrected, but since the error is conservative it does not threaten the feasibility claim. The physics-rate numbers quoted in §7.6 (460 νµ CC per 10^21 POT at 130 km) are low enough that the 'ideal match for a superbeam' framing deserves a statistics study, but that is a beam-and-exposure question rather than a defect in the detector argument under review, and the paper defers the full physics case to ref. [27]. Overall: the central claim is a feasibility argument, it is presented as such, and the one step that carries the most weight without published support is the pure-argon gain. That is worth a targeted R&D check, not a rejection.

If the design holds together, a single 100 kton liquid argon volume with bubble-chamber-quality imaging, calorimetry and Cerenkov/scintillation readout would be a substantial step beyond existing liquid argon TPCs and would serve accelerator CP violation, atmospheric, solar and supernova neutrinos and proton decay with one instrument. The paper's engineering arithmetic is a genuine strength and is internally consistent: I checked the tanker volume and mass (Table 3), the heat load, evaporation rate and nine-year empty time (Table 5), the separation and liquefaction powers and the 30 MW and 6.2 MW totals (Tables 6 and 7), the Cerenkov yield of Eq. (24)–(25), the minimum beta-beam baseline of Eq. (10), and the 3σ charge-discrimination condition of Eq. (20), which follows correctly from the sagitta and multiple-scattering expressions of Eqs. (18)–(19). The paper is also honest about what is not done, stating in the abstract and in §7.5 that the underground cavern is under study. The value of the contribution is therefore as a well-posed feasibility argument that identifies its own open problems, not as a demonstrated design.

The paper asks how to build a neutrino detector large enough to measure the CP-violating phase in neutrino oscillations, and argues that a single 100-kiloton liquid argon volume with three simultaneous readouts — drifting charge, scintillation light, and Cerenkov light — is the right answer for a superbeam or a beta-beam. The author's target is not a new measurement but a feasibility argument: the technology demonstrated by the ICARUS program can, he claims, be scaled from hundreds of tons to a hundred kilotons in one tank rather than many modules. The claim matters because the size of the detector sets the statistics, and the quality of the event image sets the background rejection, for any attempt to see whether neutrinos and antineutrinos oscillate differently. The author also argues the same tank would collect atmospheric, solar and supernova neutrinos and search for proton decay, so the cost would buy a broad program rather than a single measurement.

What carries the argument. The load-bearing mechanism is the bi-phase liquid argon time projection chamber: charge drifts in the liquid, is extracted into the gas above it, and is amplified there before readout. This is what makes a single 20-meter drift length conceivable, because the attenuation expected over that distance (about a factor of 150 for a 2 ms electron lifetime) can in principle be bought back by gas-phase gain. The paper treats the amplification step as demonstrated only at the level of encouraging measurements in pure argon, and it explicitly notes the difficulty of amplifying without a quenching gas, since any additive would both pollute the liquid and absorb the scintillation light.

Core claim. The paper's central claim is that a giant liquid argon imaging detector is technically feasible and is an ideal match to a superbeam or $\beta$-beam CP-violation program. The proposed implementation is a single boiling cryogenic tanker of roughly 70 m diameter and 20 m height holding about 102,000 tons of argon, insulated with thick perlite rather than vacuum, drawing on industrial liquefied-natural-gas practice for vessels of comparable volume. Ionization electrons would drift upward as much as 20 m under a 2 MV cathode at 1 kV/cm, then be extracted into the gas phase and amplified by a factor of $10^2$ to $10^3$ before readout on two perpendicular wire views with 3 mm pitch. Immersed phototubes would separately record the 128 nm scintillation line and the weaker Cerenkov light, giving particle identification — for instance separating muons from pions near the 140 MeV Cerenkov threshold — that charge imaging alone does not provide. The author compares liquid argon with water throughout, noting that densities, radiation lengths, interaction lengths and Cerenkov angles are close enough that Super-Kamiokande-style reconstruction carries over, while tracking adds information water Cerenkov detectors cannot supply.

The premise it rests on. The design rests on being able to drift charge 20 meters through liquid argon while keeping the electron lifetime at roughly 2 milliseconds, and then recovering the resulting factor-of-150 loss with gas-phase amplification in pure argon at gain 100 to 1000 without sparking or contamination.

A single 100-kiloton argon volume would give roughly $6\times10^{34}$ nucleons of target, allowing proton-decay lifetimes around $10^{35}$ years to be reached in about ten years of running, with channels such as $p \to \nu K$ essentially background free.

The same tank would collect on the order of 10,000 atmospheric neutrino events per year, about 100 $\nu_\tau$ charged-current events per year from oscillations, some 324,000 solar neutrino events per year above 5 MeV electron recoil, and about 20,000 events from a supernova at 10 kpc.

Combined Cerenkov and charge readout would let a beta-beam experiment separate muons from pions down to the 140 MeV muon threshold, which is the dominant background for $\nu_e \to \nu_\mu$ appearance.

Because liquid argon and water have similar radiation and interaction lengths and similar Cerenkov angles, analysis techniques developed for giant water Cerenkov detectors transfer to this medium, so the two options can be compared on physics output rather than on reconstruction method.

The paper's own numbers imply a sharp engineering trade-off that it does not spell out: any impurity that shortens the electron lifetime below about 2 ms pushes the required gas gain above $10^3$, and the amplification stage is exactly where the design is least demonstrated, so purity and gain are coupled specifications rather than independent ones.

A 20-meter drift also sets a hard ceiling on position resolution, since diffusion alone spreads the charge cloud to roughly the 3 mm wire pitch; pushing to larger tanks would require finer pitch or shorter drift, which argues that 100 kilotons in one vessel is near a natural size limit rather than an arbitrary choice.

The paper notes that the magnetized 70-kiloton variant intended for a neutrino factory depends on a magnet and yoke whose feasibility it calls unsolved, so the charge-identification capability that a neutrino factory needs is a separate and harder problem than the drift and cryogenics argued here.

If the underground cavern turns out to be the binding constraint rather than the tank, the modular path the author rejects on complexity grounds may become the practical route, and the comparison with a megaton water detector would then turn on civil construction cost rather than on detector physics.

This review was written by a machine. No person has reviewed this paper on Pith yet. A scientist listed on Pith can review it for a fee, and their review is published here whatever it concludes.

Request a human review