{"id":"ba1eab1e-7105-4700-af10-740d8551722c","arxiv_id":"2505.21654","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":8,"one_line_summary":"HALHF 2.0 presents a cost-optimized baseline for a hybrid, asymmetric plasma/RF Higgs factory with 48 plasma stages and a 5 km footprint.","lead":"This paper describes HALHF 2.0, an updated design for a plasma-based linear Higgs factory in which electrons are accelerated to 375 GeV by plasma wakefields and positrons to 42 GeV by conventional RF linacs. The authors explain how they used Bayesian optimization with a detailed cost model to choose the design parameters.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Near-optimality claim lacks uncertainty quantification: ABEL's luminosity estimate is explicitly lower-fidelity, and the 10% margin is not tested against plausible luminosity or cost perturbations.","rationale":"The paper is internally consistent: the updated parameters follow from the stated optimization and manual constraints, and the major design changes (separate linacs, 1 GV/m gradient, 48 stages) are logical. The reader's conditional verdict is appropriate. The most load-bearing question is not whether the design is achievable in principle but whether the cost-optimality claim survives realistic uncertainty in luminosity and costing; the paper itself flags the luminosity/power fidelity as future work. A targeted sensitivity analysis would settle whether the 10% claim is meaningful. The 12-parameter Bayesian optimization and reproducibility of the optimum are real evidence that the baseline is a local optimum within ABEL, but they do not validate ABEL's inputs. No change to the reader's verdict is needed; it is already conditional.","tokens_in":5511,"tokens_out":5302,"duration_ms":56592,"concrete_test":"Perform a sensitivity re-optimization with the same ABEL/Ax setup but with the calculated luminosity scaled by 0.5, 0.7, and 1.3, and separately with the major cost categories (driver linac, positron linac, plasma stages, civil engineering) perturbed by ±30%. Record the energy asymmetry and number of PWFA stages of the new optimum. If any of these optima moves more than 10% in Full Programme Cost away from, or changes the selected operating point by more than one stage or a unit of asymmetry, from the HALHF 2.0 baseline, the 'within 10%' near-optimality claim is not robust to the stated model uncertainty. A decisive variant would replace ABEL's luminosity with an independent beam-delivery/beamstrahlung calculation (e.g., Guinea-Pig) for the same beam parameters.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central claim is that HALHF 2.0 is within 10% of the Full Programme Cost minimum found by ABEL. That objective depends on luminosity through both the integrated energy term and the maintenance term (2 ab^-1 divided by luminosity); it also depends on construction costs extrapolated from ILC/CLIC. The justification for near-optimality therefore rests on the accuracy of ABEL's luminosity estimate and cost model. The authors state in the Future Work section that higher-fidelity start-to-end simulations are needed 'for more accurate estimates of luminosity and power usage', so the present simulation fidelity is explicitly admitted to be limited. No uncertainty band is attached to the 10% claim, and no end-to-end model of emittance growth, alignment jitter, final focus, or beamstrahlung is presented here. If the true luminosity were, say, 30% lower, the time to collect 2 ab^-1 and thus the integrated energy and maintenance costs would rise by roughly 30%, which can move the cost-optimum energy asymmetry and stage count substantially. The cost-model transfer to cool-copper and 48-stage plasma subsystems is another unquantified input, but the paper itself argues the plasma linac is not a dominant cost driver at 1 GV/m, making luminosity the more load-bearing unknown.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents the updated HALHF 2.0 baseline design for a hybrid, asymmetric linear Higgs factory. It summarizes the main changes from the original proposal: separate driver and positron RF linacs, a reduced plasma gradient of 1 GV/m with 48 stages, a polarized positron source, dual interaction points, and an energy asymmetry of 375 GeV electrons versus roughly 42 GeV positrons. The parameter set is motivated by a Bayesian optimization over 12 parameters using the ABEL start-to-end simulation framework and a 'Full Programme Cost' metric calibrated on ILC/CLIC costings. The authors report that the baseline is within 10% of the optimizer's solution, with additional manual tuning for physical constraints (e.g., integer multiples of the combiner-ring factor) and practical considerations such as length and power.","tokens_in":5769,"tokens_out":3466,"duration_ms":33535,"significance":"If the cost model and luminosity estimates are sufficiently accurate, the paper makes a credible case that HALHF 2.0 is an affordable and technically plausible Higgs-factory design, and it provides a transparent, reproducible parameter-optimization methodology that could be applied to other future-collider concepts. Strengths include a clear statement of the optimization metric, explicit acknowledgment of the limitations of the present simulation fidelity, and honest reporting of the manual adjustments made to the optimizer output with pointers to detailed backup documentation (Refs. [18-20]). The main weakness is that the central near-optimality claim rests on an unquantified cost model and a lower-fidelity luminosity estimate, so the significance is conditional on future validation. The paper itself notes that higher-fidelity start-to-end simulations are needed for more accurate estimates of luminosity and power usage, which is the right caveat but is not currently reflected in the strength of the near-optimality claim.","major_comments":[{"comment":"The claim that the baseline is 'within 10%' of the cost-optimal solution is not accompanied by any uncertainty quantification. The Full Programme Cost depends on luminosity through the integrated energy cost and the maintenance cost (both scale roughly inversely with luminosity for a fixed 2 ab^-1 program). The paper itself states in the Future Work section that higher-fidelity simulations are needed 'for more accurate estimates of luminosity and power usage', so the present luminosity estimate is explicitly admitted to be lower fidelity. A plausible 20-30% luminosity overestimate would raise the integrated energy and maintenance terms by a comparable fraction and can shift the optimal energy asymmetry, stage count, and driver gradient. Please add a sensitivity analysis (e.g., one-at-a-time or Monte Carlo over the uncertain inputs: luminosity, overhead fraction, maintenance rate, carbon shadow cost, infrastructure markup) or rephrase the claim as 'within 10% under the nominal ABEL model'.","section":"Bayesian optimization"},{"comment":"See comment above.","section":"Bayesian Optimization of Collider Parameters"},{"comment":"The cost model's transferability to the novel subsystems is unquantified. The paper states that the plasma linac is not a significant cost driver below 1 GV/m, but this conclusion depends on the cost model's treatment of 48 plasma stages, drive-beam distribution, plasma cells, and the cool-copper linac, none of which are direct ILC/CLIC components. The paper should explicitly state which line items in Table 2 of Ref. [20] are direct ILC/CLIC extrapolations and which are new estimates, and give a rough uncertainty on each. This is load-bearing because the optimization minimizes this metric and the 10% near-optimality claim is defined relative to it.","section":"Cost model transferability"}],"minor_comments":[{"comment":"The phrase 'a electron–positron' appears twice; it should be 'an electron–positron'.","section":"Abstract and Introduction"},{"comment":"The positron energy is given as 41 GeV in the text ('375 GeV versus 41 GeV') but as 42 GeV in Figure 1 and elsewhere; please harmonize the value.","section":"Major changes and Fig. 1"},{"comment":"The caption does not state the values of the other parameters held constant, nor the exact baseline parameter values indicated by the dotted lines; adding these would make the plot self-contained.","section":"Fig. 2"},{"comment":"There is a duplicated article in 'supported by the the European Research Council'.","section":"Acknowledgements"},{"comment":"Reference [25] is cited as a conference presentation; please update to include the arXiv identifier or proceedings reference when available.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The paper is a summary of a design effort, with details in companion papers. I think it is appropriate for the journal if the near-optimality claim is made conditional on the nominal model. The central issue is not that the design is wrong, but that the evidence for near-optimality is weaker than the wording suggests. No concerns about citation patterns or novelty disclosure."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a genuinely useful overview of the HALHF 2.0 baseline, but the \"within 10% of optimal\" claim should be read as conditional on ABEL's own cost model, not as a robust physical statement.\n\nThe paper does what it claims: it summarizes the updated baseline and walks through the Bayesian optimization. The changes from v1 are clearly motivated—separate driver and positron linacs, lower plasma gradient, 48 stages, 375/42 GeV split—and the reasoning is easy to follow. The manual tuning discussion is refreshingly honest: they admit the optimizer's minimum wasn't chosen wholesale and explain why.\n\nThe main soft spot is the uncertainty attached to the near-optimality claim. Figure 2 has no error bars, and the cost metric depends on ABEL's luminosity, which the paper itself flags as needing higher-fidelity simulation. The stress-test point about a 30% luminosity change shifting the optimum is legitimate. That said, the paper doesn't hide this—it lists it as future work. So the limitation is disclosed, not concealed. I'd have liked a sensitivity scan over luminosity and cost-model assumptions, but its absence doesn't invalidate the design logic.\n\nThe cost-model provenance (ILC/CLIC costings) is a reasonable starting point, though the transfer to cool-copper and 48-stage plasma systems is unquantified. The paper's own argument that the plasma linac isn't the dominant cost driver softens this.\n\nBottom line: this is a summary paper, not a new physics result. Its value is as a clear reference for the HALHF baseline and the optimization framework. It deserves peer review because it consolidates the design in a way the community will want to cite. I'd recommend acceptance after minor revisions asking for uncertainty bands on the cost curve and a brief sensitivity analysis.","headline":"Useful, honest summary of the HALHF 2.0 baseline; the near-optimality claim is conditional on ABEL's unvalidated cost and luminosity model.","tokens_in":6449,"tokens_out":1984,"would_cite":true,"duration_ms":20749,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"HALHF 2.0 is a roughly 5 km electron–positron Higgs factory design whose cost lies within 10% of the Bayesian optimum.","keywords":["HALHF","Higgs factory","plasma-wakefield acceleration","linear collider","cost optimization","Bayesian optimization","positron acceleration","cool-copper RF linac"],"falsifier":"Re-run the Bayesian optimization with plasma gradient and density as free parameters and with a luminosity model that includes drive-beam misalignment, synchronization jitter, and plasma-density stability; if the optimum moves by more than 10% in Full Programme Cost away from the manual 1 GV/m, $6\\times10^{14}\\,\\mathrm{cm^{-3}}$ choice, the near-optimality claim is falsified. A complementary check is a staged plasma-wakefield experiment at 1 GV/m and $6\\times10^{14}\\,\\mathrm{cm^{-3}}$ that measures the transformer ratio and final emittance to see whether the assumed per-stage luminosity holds.","tokens_in":5333,"feed_emoji":"⚛️","tokens_out":12943,"duration_ms":129614,"temperature":0.7,"pith_summary":"To build an affordable electron–positron Higgs factory, HALHF accelerates electrons in plasma wakefields—intense electric fields driven by an electron bunch passing through plasma—while accelerating positrons with a conventional RF linac, because plasma acceleration of positrons is not yet efficient enough. The updated baseline, HALHF 2.0, is a roughly 5 km facility in which 48 plasma stages at 1 GV/m take electrons to 375 GeV and a liquid-nitrogen-cooled 'cool-copper' RF linac takes positrons to 42 GeV, for 250 GeV center-of-mass collisions. The paper's central claim is that this layout is close to the cheapest credible design: a Bayesian optimizer minimizing a 'Full Programme Cost' (construction, operation, maintenance, and carbon) converged repeatedly on a solution whose cost the chosen baseline matches to within 10%. The design also replaces the original combined-function linac with separate driver and positron linacs, and lowers the plasma gradient to ease tolerances and cooling.","feed_headline":"Plasma-RF collider design shrinks Higgs factory to 5 km","feed_subtitle":"48 plasma stages push electrons to 375 GeV; a cool-copper linac takes positrons to 42 GeV, within 10% of cost-optimal.","key_machinery":"Three linked elements carry the argument: ABEL, the adaptable beginning-to-end linac simulation framework that turns a candidate layout into predicted luminosity, length, and power usage; the 'Full Programme Cost' metric, which balances construction cost, about 22% overheads, integrated energy cost, roughly 1%-per-year maintenance, and a carbon shadow cost; and the Bayesian optimizer that searches the 12-parameter space (energy asymmetry, bunch-train structure, driver and positron linac gradients, number of stages, transformer ratio, and others) to find the cost minimum. The physical lever that defines the new baseline is the plasma gradient: lowering it from 6.4 GV/m to 1 GV/m, with the plasma density dropping to $6\\times10^{14}\\,\\mathrm{cm^{-3}}$, relaxes synchronization, alignment, matching, beam-ionization, and cooling requirements while adding little to the cost, because the plasma linac is not a dominant cost driver at such gradients.","core_discovery":"The central claim is that the updated HALHF 2.0 parameter set is a viable, self-consistent, and nearly cost-optimal baseline for a 250 GeV electron–positron Higgs factory. The machine accelerates a 1.6 nC electron bunch through 48 beam-driven plasma-wakefield stages, each adding 7.8 GeV at a gradient of 1 GV/m, to reach 375 GeV, while a cool-copper RF linac at 40 MV/m and 3 GHz accelerates a 4.8 nC positron bunch to 42 GeV; the two beams collide at two interaction points. The authors justify the numbers by minimizing a 'Full Programme Cost' computed in ABEL, a start-to-end simulation that estimates luminosity, length, and power, with a Bayesian optimizer varying 12 parameters; the optimizer ran 80 iterations, converged to the same solution each time, and the manually tuned baseline lies within 10% of that solution. The deliberate deviations, such as choosing a 375/42 GeV energy split rather than the optimizer's cheaper 250/62.5 GeV split, are motivated by shortening the facility to about 5 km and reducing construction cost at a small penalty in the cost metric.","pith_inferences":["If the cost model's assumptions hold, the same optimization machinery could be applied to other plasma-based collider concepts once their luminosity models reach comparable fidelity, yielding similarly parameter-driven baselines.","Because the paper fixes plasma gradient and density outside the optimizer, re-running the optimization with these as free parameters could shift the optimum toward higher gradients if jitter and misalignment tolerances prove less restrictive than assumed.","The inclusion of a carbon shadow cost means the chosen baseline's near-optimality is partly a bet on sustainability being priced at 800 EUR/ton; if carbon costs rise or operational efficiency is valued more, a lower-power, lower-luminosity variant might become the best buy.","A natural extension would be to vary the center-of-mass energy: nothing in the method ties the optimization to 250 GeV, so the same cost-based search could map the price of a 350 or 500 GeV Higgs factory and identify where the plasma-asymmetric approach loses its advantage."],"forward_implications":["A 250 GeV Higgs factory built on this baseline would occupy about 5 km, much shorter than conventional RF collider designs of similar energy.","Separate driver and positron linacs allow the driver to run at high current and low gradient (8 nC bunches, 4 MV/m, 1 GHz) and the positron linac at high gradient and lower current (40 MV/m, 3 GHz), removing the need for a novel combined-function linac.","The lower plasma gradient and density make the plasma arm more conservative, with relaxed timing, alignment, beam-ionization, and cooling constraints, at a small cost penalty.","The energy asymmetry of 375 GeV electrons versus 42 GeV positrons minimizes the overall collider length and construction cost when all subsystems are included, even though it is not the pure cost minimum.","The dual interaction point and dual beam-delivery system allow two detectors to run simultaneously at the same collider."],"supporting_citations":[{"why":"The original HALHF proposal whose baseline this paper updates; supplies the hybrid, asymmetric concept.","marker":"[15]"},{"why":"Reports the updated HALHF 2.0 baseline in detail; the source of the 48-stage, 1 GV/m, 375/42 GeV parameters.","marker":"[18]"},{"why":"Provides the detailed cost model, in standardized collider cost units, underlying the Full Programme Cost.","marker":"[20]"},{"why":"Supplies the cool-copper RF linac technology and 40 MV/m, 3 GHz parameters used for the positron arm.","marker":"[21]"},{"why":"ABEL, the start-to-end simulation framework that computes luminosity, length and power for each candidate design.","marker":"[25]"},{"why":"Supplies the Bayesian optimization algorithm used for the 12-parameter search.","marker":"[26]"}],"fun_headline_variants":["HALHF 2.0: hybrid plasma-RF Higgs factory in 5 km, cost-tuned","5-km Higgs factory splits beams: plasma for electrons, RF for positrons","Updated HALHF baseline: 48 plasma stages, 5 km, near cost-optimal","Cost-optimized HALHF design: asymmetric plasma-RF Higgs factory"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the luminosity numbers from the start-to-end simulations and the cost model built from earlier large-collider costings are accurate enough that the design ranked cheapest really is the cheapest.","fun_headline_variants_meta":{"raw":{"variants":["HALHF 2.0: hybrid plasma-RF Higgs factory in 5 km, cost-tuned","5-km Higgs factory splits beams: plasma for electrons, RF for positrons","Updated HALHF baseline: 48 plasma stages, 5 km, near cost-optimal","Cost-optimized HALHF design: asymmetric plasma-RF Higgs factory"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000212,"raw_usage":{"total_tokens":1413,"prompt_tokens":938,"completion_tokens":475,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":554,"completion_tokens_details":{"reasoning_tokens":382}},"tokens_in":554,"tokens_out":475,"duration_ms":5670,"temperature":1.0,"reasoning_tokens":382,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T13:25:00.800747+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Re-run the Bayesian optimization with plasma gradient and density as free parameters and with a luminosity model that includes drive-beam misalignment, synchronization jitter, and plasma-density stability; if the optimum moves by more than 10% in Full Programme Cost away from the manual 1 GV/m, $6\\times10^{14}\\,\\mathrm{cm^{-3}}$ choice, the near-optimality claim is falsified. A complementary check is a staged plasma-wakefield experiment at 1 GV/m and $6\\times10^{14}\\,\\mathrm{cm^{-3}}$ that measures the transformer ratio and final emittance to see whether the assumed per-stage luminosity holds.","supporting_citations":[{"cited_title":"A “Cool” route to the Higgs boson and beyond. The Cool Copper Collider","cited_arxiv_id":null,"evidence_quote":"Supplies the cool-copper RF linac technology and 40 MV/m, 3 GHz parameters used for the positron arm."},{"cited_title":"ABEL: The adaptable beginning-to-end linac simulation framework","cited_arxiv_id":null,"evidence_quote":"ABEL, the start-to-end simulation framework that computes luminosity, length and power for each candidate design."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the Bayesian optimization algorithm used for the 12-parameter search."}],"review_version":1}