{"id":"262a23c3-643b-4924-9be9-a3f2577b29e5","arxiv_id":"2411.15369","paper_version":1,"verdict":"REJECT","confidence":"MODERATE","novelty_score":3.0,"correctness_risk":"high","formal_verification":"none","parameter_count":7,"one_line_summary":"A heuristic cost model plus company data is used to claim that Global Thermostat's honeycomb solid-sorbent DAC can reach below $50 per tonne, matching the lowest-cost case in a 2019 NAS study.","lead":"This paper argues that Direct Air Capture could cost under $50 per tonne of CO2 using solid-sorbent honeycomb contactors, and that Global Thermostat's technology already demonstrates the key performance requirements. It matters because cost and energy uncertainty is a major barrier to scaling carbon removal, and the authors use this cost path to call for an immediate global mobilization.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Paper's own text concedes no integrated unit, and the $25/tonne target rests on an NAS lower bound that NAS itself calls 'perhaps not realistically achievable.'","rationale":"I read the paper in good faith and find real value in the component-level measurements: the pressure-drop data are consistent with laminar channel theory, the uptake curve is internally coherent, and the steam partial-pressure argument for suppressing water desorption is thermodynamically plausible. The reader's chosen weakest assumption—water co-desorption—is therefore not the most decisive point. The decisive problem is the inference from component demonstrations to the integrated $25/tonne claim. The paper explicitly concedes that the stages have not been integrated into a single operating unit, and the NAS lower bound it invokes is one that NAS itself flags as 'perhaps not realistically achievable.' Even if every input assumption were granted, the claim that GT has demonstrated the properties of the $25/tonne approach is unsupported by the presented evidence. Because the verdict semantics treat an unsupported central claim as a critical red flag, REJECT remains the appropriate outcome; my analysis therefore does not change the reader's verdict.","tokens_in":19191,"tokens_out":5179,"duration_ms":49486,"concrete_test":"Run an integrated cycle test on a GT monolith: at least 100 full adsorption–steam-regeneration–cooling cycles at v=5 m/s and ~400 ppm CO2, measuring per-cycle CO2 captured, steam and electricity consumption, water balance, and sorbent capacity. Compute annualized tonnes CO2 per kg contactor and GJ per tonne. If productivity falls below ~1 t/kg/yr or specific heat exceeds ~4 GJ/tonne, the $50/tonne cost chain fails.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim—that GT has demonstrated the properties of the $25/tonne NAS approach—is contradicted by the paper's own admissions. In the Scaling section, it states: 'The performance demonstrated for each stage of the DAC process, together with heat management, has not yet been demonstrated in a single operating unit.' The $25/tonne figure is an integrated-process cost from NAS; component-level demonstrations (pressure drop, uptake, plug flow) cannot establish it. Moreover, Appendix 1 quotes NAS: 'Disregarding the lower bound as perhaps not realistically achievable'—NAS itself does not treat $18–25/tonne as achievable, yet the paper adopts that lower bound as the 'learning cost limit' and claims GT matches it. The economic chain then relies on unvalidated inputs: 1 tonne CO2/kg contactor/year, $15/kg device cost, 0.25 learning rate, 400 kWh/tonne electricity, and under 4 GJ/tonne steam. The water co-desorption issue is real, but the more fundamental gap is that the integrated performance required by the NAS cost model has not been demonstrated.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper argues that direct air capture (DAC) can reach costs below $50 per tonne of CO2, and possibly $18–25 per tonne at the learning-curve limit, by using low-pressure-drop monolith contactors with fast kinetics, steam-based regeneration, and manufacturing analogs such as truck or diesel-engine production. The authors present a simplified cost model, experimental data on pressure drop, CO2 uptake, and plug flow for Global Thermostat's contactors, and an appendix claiming that the National Academy of Sciences (NAS) study independently identified GT's patented approach as the lowest-cost DAC pathway. The paper concludes that GT has demonstrated the properties of the $25 per tonne NAS approach and that scaling should begin immediately, despite admitting that the stages have not yet been integrated into a single operating unit.","tokens_in":19337,"tokens_out":4027,"duration_ms":36665,"significance":"If the central claim were substantiated, the result would be highly significant: DAC at the claimed cost would change the economic feasibility of gigatonne-scale carbon removal and would directly inform policy decisions. The paper has some genuine strengths: the laminar-flow pressure-drop correlation and diffusion-time arguments are textbook physics, and the reported pressure-drop and uptake data are consistent with those standard correlations. However, the central quantitative conclusion depends on a chain of unvalidated assumptions and on an interpretation of the NAS report that the report itself does not support. The authors are also candid about a major unverified premise, the water co-desorption behavior, and about the lack of an integrated demonstration. Because the load-bearing claims are not backed by integrated experimental evidence or independent cost validation, the paper's significance for the stated policy conclusion is currently limited.","major_comments":[{"comment":"The central claim that 'GT has demonstrated the properties of the $25 per tonne approach identified in the NAS study' is contradicted by the paper's own admission in the Scaling section: 'The performance demonstrated for each stage of the DAC process, together with heat management, has not yet been demonstrated in a single operating unit.' The NAS cost limit is an integrated-process cost, and component-level demonstrations (pressure drop in Fig. 3, uptake in Fig. 4, plug flow in Fig. 6, temperature profiles in Fig. 7) cannot establish that the integrated process will meet that cost target.","section":"Conclusion and Scaling section"},{"comment":"The paper quotes NAS's statement that 'disregarding the lower bound as perhaps not realistically achievable' and that the middle range of scenarios yielded capture costs of $88–228 per tonne for a generic solid sorbent DAC system, yet it adopts the $18–25 per tonne lower bound as GT's learning-curve limit. This selective reading of the NAS report does not support the claim that NAS identified GT's approach as the lowest-cost path; in fact, the report explicitly cautions against treating the lower bound as realistic.","section":"Appendix 1"},{"comment":"The annualized capital cost of $21 per tonne relies on the unvalidated analogy between Class 8 trucks and diesel engines on the one hand and DAC contactors on the other, giving $15 per kg of device, together with the assumption of 1 tonne CO2 captured per kg of contactor per year. The paper provides no evidence for this productivity or unit-cost figure, and the resulting $21 per tonne is load-bearing for the under-$50 per tonne conclusion. The truck-manufacturing analogy is asserted, not demonstrated.","section":"Analysis of GT DAC Technology (capital cost derivation)"},{"comment":"The steam energy cost of roughly $25 per tonne depends critically on the premise that water co-adsorbed on the sorbent does not desorb during regeneration because steam maintains a water partial pressure at or above the adsorption value. The paper argues this only qualitatively and explicitly concedes that if 10x the water desorbs, the heat requirement triples to 480 kJ/mol CO2. No integrated contactor-plus-regeneration experiment is presented to verify the premise, so the energy-cost claim is not empirically supported.","section":"DAC Cost Analysis, 'A significant issue is whether water co-adsorbed with CO2 must also be desorbed...'"},{"comment":"The assertion that the NAS's lowest-cost scenario 'is in fact the GT process' is made by the authors who developed and patented that approach, but the NAS report describes generic low-pressure-drop solid-sorbent configurations without identifying GT. This identification is therefore circular and is not independently established. The paper's conclusion that GT is on the lowest-cost learning curve follows only from this unsupported mapping.","section":"Appendix 1 (NAS comparison)"}],"minor_comments":[{"comment":"In the derivation of the diffusion timescale and the removal fraction, the text writes 'C = (S/2)^2/D' and 'η = (1 − y^{−θτ})', which appear to be typographical errors for τ_diff = (S/2)^2/D and η = 1 − e^{−θτ}.","section":"Capturing the CO2 in the Air"},{"comment":"The caption says 'contractors' instead of 'contactors' and includes the ambiguous phrase 'eg (9)', which likely refers to an equation number or reference but is not clear.","section":"Figure 3 caption"},{"comment":"The text reads 'the new electrical energy required is 10^19 joules or 27,80 TWh'; 10^19 J equals approximately 2,780 TWh, so the number appears to be a typographical error.","section":"Energy Use section"},{"comment":"In the scaling formula C_x2 = C_x1 × (T2/T1)^a, the surrounding text says 'N is less than 1, usually taken about .6 or .7', but the exponent in the equation is written as 'a'; the notation should be made consistent.","section":"Learning Curve Cost Reductions"},{"comment":"References 2 and 3 are duplicates of the same source, and reference 8's title ('Flawed analysis of the possibility of air capture') does not match the cited PNAS article that is presumably intended.","section":"References"}],"recommendation":"reject","confidential_remarks":"The manuscript reads largely as a policy advocacy document for a specific company, authored by the technology's developers, with the central cost claims resting on component-level demonstrations and an over-reading of the NAS report. The lack of an integrated unit and the paper's own caution about the NAS lower bound are weighty enough that the central claim cannot be accepted in its present form. The scope of the journal may also be a consideration, as the core content is engineering economics and technology advocacy rather than a physics or social-physics contribution."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a Global Thermostat advocacy paper with a real experimental core and a cost model that does not support its own headline. The pressure-drop and plug-flow measurements are genuine and consistent with laminar theory, and the simplified cost decomposition is a useful way to separate fan energy, heat, and capital. The problem is that the $50/tonne (and especially $25/tonne) conclusions are assembled from assumptions, not measurements. The paper itself admits no single unit has demonstrated the integrated process, and the water co-desorption question—flagged as the most important issue—is answered by assertion that steam keeps water partial pressure high enough to avoid desorption. If that fails, the paper's own numbers triple the heat requirement. The Appendix 1 mapping of GT onto the NAS lowest-cost scenario is the authors' own interpretation, and they do not disclose that they are the technology's developers. NAS itself says the lower bound is 'perhaps not realistically achievable.' So the central claim—that GT has demonstrated the properties of the $25/tonne approach—is not established. I side with the reader's REJECT on that basis.\n\nWhat the paper does well: it makes the cost drivers concrete, uses standard fluid mechanics, and presents actual data that can be checked. The learning-curve discussion is conventional but clearly framed. If I were modeling DAC costs, I might borrow the cost decomposition and the GT data points, but not the cost conclusions.\n\nThis is a paper for people tracking DAC policy and cost claims: read it as a position statement from GT, not an independent validation. A serious editor should still send it to peer review—the stakes are high, the data deserve scrutiny, and the cost model deserves a technical referee. I would not cite it as evidence for achievable DAC costs.","headline":"A GT advocacy paper with real experimental data but a cost model that does not support its $50/tonne claim; the central NAS mapping is the authors' own, and the paper's own text admits the integrated process is not demonstrated.","tokens_in":19941,"tokens_out":3670,"would_cite":false,"duration_ms":32321,"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":"The paper argues that under-$50-per-tonne direct air capture is achievable, and that the assessed monolith-sorbent technology already matches the cheapest generic path identified by a national cost study.","keywords":["direct air capture","carbon dioxide removal","solid sorbent","monolith contactor","steam regeneration","cost learning curve","gigatonne scale","low-temperature heat"],"falsifier":"Run an integrated monolith contactor through repeated adsorption and steam-regeneration cycles while metering water and CO2 leaving the module; if the water-to-CO2 molar ratio in the desorbed gas approaches ten rather than near zero under the claimed 70-100 degree Celsius steam conditions, the regeneration energy triples and the under-$50-per-tonne cost path fails.","tokens_in":18861,"feed_emoji":"🌍","tokens_out":16154,"duration_ms":131988,"temperature":0.7,"pith_summary":"Direct air capture (DAC) is often dismissed as inherently expensive because CO2 in the atmosphere is so dilute. This paper argues the opposite: after air passes through a low-pressure-drop contactor, the dominant cost is regenerating the sorbent, and that cost stays near 4 gigajoules per tonne when steam keeps co-adsorbed water from evaporating. Combining measured fan, pressure-drop, uptake, and steam-sweep data with a simplified cost model, the authors claim a mature system can capture CO2 for under $50 per tonne, with a learning-curve floor around $25 per tonne. If that claim holds, DAC would be economically viable at the million-tonne scale, well before the gigatonne scale needed for climate protection, so today's pilot costs should not be the basis for policy decisions.","feed_headline":"Direct air capture can fall below $50 per tonne","feed_subtitle":"A simplified cost model plus measured contactor and steam data put the low-cost path within reach.","key_machinery":"The load-bearing mechanism is a parallel-channel monolith contactor (a block of narrow channels whose porous walls carry the sorbent) run in a combined temperature-vacuum-concentration swing cycle with direct steam regeneration. The geometry gives laminar flow, a surface-area-to-volume ratio of thousands of square meters per cubic meter, and a channel-center-to-wall diffusion time under one second, so most incident CO2 reaches the sorbent at high air velocity with little pressure drop. Direct steam condensation delivers heat to the walls in seconds, acts as a sweep gas to keep the CO2 partial pressure low, and is the specific device that keeps water partial pressure at or above the adsorption value, preventing co-adsorbed water from being evaporated. That water-suppression step is what holds regeneration heat near 4 gigajoules per tonne; if ten times as much water as CO2 desorbed, the heat requirement would roughly triple. The fast cycling this enables also lets one regeneration station serve roughly ten adsorption panels, which is the step that turns capital cost into a mass-manufactured-product estimate near $150 per tonne of annual capacity.","core_discovery":"The central claim is that the cheapest generic DAC configuration from a major cost study is realizable: a parallel-channel monolith contactor whose porous walls hold the sorbent, capturing more than half the incoming CO2 from air moving at 3 to 5 meters per second with a pressure drop below 200 pascals, regenerated by direct contact with low-temperature steam. The steam heats the walls by condensing, sweeps the released CO2 out of the channels, and holds the water partial pressure high enough that co-adsorbed water does not have to be evaporated. From stage-level measurements, the paper estimates annualized capital cost of about $21 per tonne and total energy cost near $50 per tonne, with steam heat below 4 gigajoules per tonne and electrical demand around 400 kilowatt-hours per tonne; cogeneration with waste heat could bring the energy component to roughly $10 per tonne. Using a standard learning-curve law, it argues that $50 per tonne is reached within six to ten doublings of capacity, at a million-tonne-per-year scale, and that the relevant climate metric is therefore the learning-curve limit rather than current cost. The paper does not claim a commercial unit exists; it claims the critical performance parameters have been demonstrated separately and that no showstoppers are known.","pith_inferences":["The decisive unresolved test is the water balance: if integrated field operation shows a water-to-CO2 desorption ratio near ten rather than near zero, the energy cost triples and the under-$50 route fails; this should be the first measurement target for any scale-up program.","Because the paper prices capture only, the full cost of permanent net removal including compression, transport, and sequestration will be higher than $50 per tonne; the headline number should not be read as the total cost of net-negative removal.","The six performance parameters the paper names suggest a direct way to compare rival DAC approaches: measure contacting rate, kinetics, regeneration speed, regeneration energy, energy cost, and airflow resistance in a standardized test, independent of company-specific cost claims.","If the separately demonstrated stages can be combined in one integrated unit without performance loss, the paper's central claim would move from plausible to demonstrated; that integration is the missing experimental step."],"forward_implications":["At a million-tonne-per-year scale, capture cost can reach $50 per tonne, so DAC becomes commercially viable during the scaling phase rather than only after the full gigatonne build-out.","A 10-gigatonne-per-year DAC industry would need about 6 percent of projected 2050 global electricity generation, making low-carbon electricity supply a necessary complement to the technology itself.","Policy and investment decisions should treat the learning-curve limit of a DAC route, not its current pilot cost, as the relevant metric, because R&D can shift a technology onto a lower cost curve before capacity-driven learning begins.","With the current $130-per-tonne tax credit for sequestered CO2, a technology with capture cost below $50 per tonne is economically attractive even while it is still scaling.","The same monolith hardware can capture CO2 from flue gas at even lower cost, and the combined mode can remove roughly twice the CO2 emitted by the natural gas that powers it."],"supporting_citations":[{"why":"The detailed cost study whose lowest-cost scenario, around 25 dollars per tonne, the paper claims its technology matches.","marker":"15"},{"why":"The learning-curve projection that concludes DAC stays above 100 dollars per tonne and that the paper argues omits R&D-driven cost-curve shifts.","marker":"4"},{"why":"The analysis that poses the buying-down question using the solar learning curve as the comparison the paper extends.","marker":"2"},{"why":"The earlier treatment of DAC as a dilute separation problem with energy concentrated in regeneration, which frames the simplified model.","marker":"8"},{"why":"Supplies the mass-transfer expression used to benchmark contactor capture fraction and pressure drop.","marker":"10"},{"why":"Supports the assumed 1-2 cents per kilowatt-hour renewable electricity cost at scale, which drives the fan-energy estimate.","marker":"11"},{"why":"Provides the ammonia-plant capital-cost analog used to bound the alternative liquid-solvent capital path.","marker":"12"},{"why":"The original learning-curve law used to compute capacity doublings to reach 50 dollars per tonne.","marker":"28"},{"why":"Supplies the distinction between R&D learning and capacity-driven learning that underlies the claim that R&D can lower the learning-curve limit.","marker":"5"}],"fun_headline_variants":["DAC below $50 per tonne is achievable, study shows","Direct air capture cost drops to $50 via learning curve","Study: $50/tonne DAC within reach after scale-up","Path to cheap DAC mapped via demonstrated performance","Learning curve puts $50/tonne direct air capture in sight"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole cost projection rests on steam holding water on the sorbent during regeneration so that co-adsorbed water does not have to be evaporated; the paper itself notes that if ten times as much water as CO2 came off, the required heat would triple.","fun_headline_variants_meta":{"raw":{"variants":["DAC below $50 per tonne is achievable, study shows","Direct air capture cost drops to $50 via learning curve","Study: $50/tonne DAC within reach after scale-up","Path to cheap DAC mapped via demonstrated performance","Learning curve puts $50/tonne direct air capture in sight"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000519,"raw_usage":{"total_tokens":2600,"prompt_tokens":1113,"completion_tokens":1487,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":729,"completion_tokens_details":{"reasoning_tokens":1404}},"tokens_in":729,"tokens_out":1487,"duration_ms":11930,"temperature":1.0,"reasoning_tokens":1404,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T14:22:27.342433+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run an integrated monolith contactor through repeated adsorption and steam-regeneration cycles while metering water and CO2 leaving the module; if the water-to-CO2 molar ratio in the desorbed gas approaches ten rather than near zero under the claimed 70-100 degree Celsius steam conditions, the regeneration energy triples and the under-$50-per-tonne cost path fails.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"The detailed cost study whose lowest-cost scenario, around 25 dollars per tonne, the paper claims its technology matches."},{"cited_title":"Considering technology characteristics to project future costs of direct air capture","cited_arxiv_id":null,"evidence_quote":"The learning-curve projection that concludes DAC stays above 100 dollars per tonne and that the paper argues omits R&D-driven cost-curve shifts."},{"cited_title":"J., & Eisenberger, P","cited_arxiv_id":null,"evidence_quote":"The earlier treatment of DAC as a dilute separation problem with energy concentrated in regeneration, which frames the simplified model."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the mass-transfer expression used to benchmark contactor capture fraction and pressure drop."},{"cited_title":"(2021, March 25)","cited_arxiv_id":null,"evidence_quote":"Supports the assumed 1-2 cents per kilowatt-hour renewable electricity cost at scale, which drives the fan-energy estimate."},{"cited_title":"(2018, January 19)","cited_arxiv_id":null,"evidence_quote":"Provides the ammonia-plant capital-cost analog used to bound the alternative liquid-solvent capital path."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"The original learning-curve law used to compute capacity doublings to reach 50 dollars per tonne."}],"review_version":1}