{"id":"fea815e6-feba-4f6e-b764-ffe6fdbc2bc8","arxiv_id":"2501.07900","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"In one dimension, under g(0)=0, g>=0 and convex sub-level sets, every minimizer of the free energy with prescribed mass is an interval and a minimizer always exists.","lead":"This paper proves that in one dimension, the shape that minimizes a crystal's surface plus potential energy is always an interval (a convex set), whenever the potential has convex sub-level sets. The result settles the one-dimensional version of a classical problem attributed to Almgren about when energy minimizers are convex.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The existence proof rests on a boundedness claim that is false: for g=1 away from 0 with g(0)=0, every translate of (0,m) has the same energy, so unbounded minimizing sequences exist. The theorem may still be true, but Theorem 1.1 is not established as written.","rationale":"The reader's weakest assumption identifies exactly the false boundedness claim in Section 2, and I agree it is the main load-bearing issue. I independently checked whether the counterexample satisfies all hypotheses of Theorem 1.1; it does, and it shows the proof's assertion that sup_k |a_k| < ∞ is demonstrably false. I also checked whether the theorem itself might be false: it is not. An unbounded minimizing sequence forces an attained minimizer at a=0 or a=-m, and Claim 2 proves that an interval of mass m has energy no larger than any competing set. Thus the central result is correct and the proof is repairable by a short argument, but the proof as written contains a false statement in the existence argument, so a CONDITIONAL verdict is appropriate. No circularity, data, or reproducibility issues were found.","tokens_in":6680,"tokens_out":12861,"duration_ms":130428,"concrete_test":"Run the following analytical check. Set g(x)=1 for x≠0, g(0)=0, fix m>0, and let I=(0,m). For every a, ∫_{a}^{a+m} g dx = m, so E(I+a)=2+m and a_k=k is a minimizing sequence with |a_k| → ∞; this directly disproves the boundedness assertion in Section 2 used to obtain a convergent subsequence. To confirm the minimal repair, verify the alternative argument: if a_k → +∞, then for large k, E(I+a_k) ≥ E(I) because g is nondecreasing on R_+, so inf = E(I) and a=0 is a minimizer; if a_k → -∞, similarly E(I+a_k) ≥ E((-m,0)) and a=-m is a minimizer.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The proof of Theorem 1.1 reduces existence to finding a minimizer of a ↦ E(I+a) for I=(0,m). Section 2 then asserts that for a minimizing sequence (a_k), 'monotonicity yields sup_k |a_k| < ∞'. This assertion is false. Let g(x)=1 for x ≠ 0 and g(0)=0; the sub-level sets {g<t} are {0} for 0<t≤1 and R for t>1, so the hypotheses of Theorem 1.1 hold. For every a, ∫_{a}^{a+m} g(x) dx = m, hence E(I+a)=2+m for all a. Thus a_k=k is a minimizing sequence with |a_k| → ∞. The attempted contradiction in the proof requires the strict inequality lim_l ∫_{I+a_{k_l}} g > ∫_0^m g; in this example the two integrals are equal for every a, so the argument collapses. Consequently, the extraction of a convergent subsequence and the existence part of Theorem 1.1 are not established as written. The theorem is very likely correct: if a minimizing sequence goes to +∞, monotonicity gives E(I+a_k) ≥ E(I) for large k, so a=0 already attains the infimum; similarly for -∞ one obtains a=-m. I also noticed a secondary slip in the final convexity argument in the case E⊂{g=0}, where 'since H0(∂E)≥2, E is an interval' is not a valid inference for arbitrary finite-perimeter sets, but Claim 2 repairs it immediately, so this is not the main obstruction.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the one-dimensional Almgren problem: minimize the free energy E(E) = H^0(∂E) + ∫_E g dx over sets E ⊂ R of prescribed measure m, where g ≥ 0, g(0)=0, and all sublevel sets {g < t} are convex. Theorem 1.1 claims that minimizers exist and are intervals. The proof first shows (Claim 1) that the sublevel-set condition is equivalent to g being nondecreasing on [0,∞) and nonincreasing on (−∞,0], then proves (Claim 2) a rearrangement inequality showing that for any set E, the interval with the same mass on the left and right of the origin has no larger potential energy. From this, the author reduces the problem to minimizing a ↦ E((0,m)+a), and then attempts to show that a minimizing sequence of translations is bounded. A separate optimal-transport proof of Claim 2 is also included, and Section 3 discusses identifying the optimal translation α.","tokens_in":7039,"tokens_out":4082,"duration_ms":44877,"significance":"If Theorem 1.1 is established, it settles the natural one-dimensional version of the Almgren convexity problem under very mild assumptions, and it gives a satisfying explanation of why convex minimizers should appear in one dimension. The paper's core rearrangement inequality (Claim 2) is correct, self-contained, and is the right key step; the optimal transport formulation of that inequality is elegant and potentially useful beyond this setting. The manuscript also correctly identifies that coercivity of g is not necessary for existence. However, the proof of existence contains a false boundedness claim, so the main theorem is not established as written. The convexity conclusion is likely salvageable with a modest repair, but the current version needs substantial correction before the result can be accepted.","major_comments":[{"comment":"The assertion 'monotonicity yields sup_k |a_k| < ∞' is false. Let g(x)=1 for x≠0 and g(0)=0. Then the sublevel sets {g<t} are {0} for 0<t≤1 and R for t>1, so the hypotheses of Theorem 1.1 hold. For every a, ∫_a^{a+m} g(x) dx = m, hence E((0,m)+a)=2+m for all a. Therefore a_k=k is a minimizing sequence for inf_a E((0,m)+a) with |a_k| → ∞, contradicting the claimed uniform boundedness. The proof's contradiction uses the strict inequality lim_l ∫_{I+a_{k_l}} g > ∫_0^m g; in this example the two integrals are equal for every a, so the argument collapses. Since this boundedness is used to extract a convergent subsequence and prove existence, the existence half of Theorem 1.1 is not established as written. The claim can likely be repaired by treating subsequences tending to +∞ or −∞ separately, but the repair is not present.","section":"Section 2, paragraph beginning 'If a_k are numbers such that ...'"},{"comment":"The inference 'if E ⊂ {g=0}, then since H^0(∂E) ≥ 2, E is an interval' is not valid for arbitrary finite-perimeter sets. For instance, the union of two disjoint intervals of total length m has H^0(∂E)=4 and is not an interval. The conclusion that any minimizer is an interval can nevertheless be obtained directly from Claim 2, which shows that any disconnected E is strictly dominated by an interval of the same mass (once the existence of a minimizer is known). The proof should be restated to use Claim 2 instead of the invalid inference from the boundary count.","section":"Section 2, final paragraph of the proof of Theorem 1.1"},{"comment":"The derivation g(α+m)=g(α) from differentiating G(a+m)−G(a) presupposes differentiability (or at least absolute continuity) of G, which is not assumed in Theorem 1.1. Moreover, the examples impose strict monotonicity and continuity hypotheses that are not part of the theorem. The paragraph asserting 'one always finds a minimizer E_m with 0 ∈ E_m' and that 'E_m is a minimizer as well' is not proved in the stated generality and appears to conflict with cases where g vanishes on a half-line. These statements should either be proved under the actual assumptions or explicitly presented as heuristic.","section":"Section 3, first-order condition and examples"}],"minor_comments":[{"comment":"There are numerous typographical errors, including 'countin g' in the introduction, 'th e', 'suppo sing', and 'theo rem'. The text would benefit from a careful proofreading pass.","section":"Throughout"},{"comment":"The notation 'E0' is undefined; the text should read 'if E^0 = ∅' for the interior, and the subsequent inclusion should be written as 'E ⊂ ∂E ∪ E^0'.","section":"Section 2, first paragraph"},{"comment":"The dominated convergence argument requires that g be locally integrable (or at least that the relevant integrals are finite), but the theorem's hypotheses only state g≥0 with convex sublevel sets. The assumptions should explicitly include measurability and local integrability of g.","section":"Section 2, Claim 2 proof"},{"comment":"Claim 2 is stated and proved a second time with the same numbering as in Section 2; this duplicate numbering is confusing and should be renumbered.","section":"Section 2.2"},{"comment":"The notation 'H^0(∂(−|E_−|,|E_+|)) = 2' should be clarified with parentheses, e.g., 'H^0(∂((−|E_−|,|E_+|))) = 2'.","section":"Section 3"}],"recommendation":"major_revision","confidential_remarks":"The paper has a high density of self-citations and references to the author's own prior work; this is not inherently problematic, but the novelty relative to [Ind24a] should be stated more crisply. The central idea (Claim 2) is good, but the existence proof as written is not correct. I believe the theorem is likely true and repairable, but the repair is not a purely local edit: the boundedness argument needs to be replaced with a correct argument handling divergent minimizing sequences."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague—\n\nThe short version: this paper proves the natural 1D version of Almgren's problem—if g(0)=0, g≥0, and the sublevel sets {g<t} are convex, then the free energy E(E)=H0(∂E)+∫_E g admits a minimizer of mass m and every minimizer is an interval. The theorem is new relative to the prior literature (which mostly needs n=2, coercivity, or extra regularity) and the proof is mostly a clean rearrangement argument: convex sublevel sets are equivalent to g being non-decreasing on [0,∞) and non-increasing on (-∞,0], and then a simple layering argument shows intervals beat every other set. The optimal transport presentation of the key rearrangement step is a nice optional aside. The paper is self-contained and no circularity or fitted parameters are involved.\n\nThe soft spot is real and exactly where the stress-test puts it. In Section 2, after reducing to intervals I+a=(a,a+m), the proof asserts that any minimizing sequence (a_k) must be bounded because 'monotonicity yields sup_k |a_k| < ∞'. This is false. Take g(x)=1 for x≠0, g(0)=0. The sublevel sets are {0} and R, so the hypotheses hold, and for every a the integral ∫_a^{a+m} g is m, so E(I+a)=2+m is constant. The sequence a_k=k is minimizing and unbounded. The displayed contradiction in the paper relies on the strict inequality lim_l ∫_{I+a_{k_l}} g > ∫_0^m g, which fails here because both sides equal m. So the existence part of Theorem 1.1 is not established as written.\n\nThat said, the flaw is repairable and the theorem is very likely correct. If a minimizing sequence goes to +∞, monotonicity gives E(I+a) ≥ E(I) for every a>0, so a=0 already attains the infimum; the case a→−∞ is symmetric and gives the interval (−m,0). The convexity of minimizers follows from the rearrangement inequality once existence is granted. There is also a minor slip in the final paragraph where 'since H0(∂E)≥2, E is an interval' is not valid for arbitrary finite-perimeter sets, but Claim 2 (the mass rearrangement inequality) immediately supplies the correct argument, so this is not a serious issue.\n\nWho is this for? People working in the calculus of variations, isoperimetric inequalities, or crystal shapes will find it a clean, quotable result in the one-dimensional base case. It deserves a serious referee: the theorem is new, the proof strategy is sound in its main line, and the existence gap is a few lines to fix. I'd send it to review, with a request that the author repair the boundedness argument and acknowledge the counterexample.","headline":"A clean one-dimensional theorem with a false step in the existence proof; the result is probably right and the paper is worth engaging.","tokens_in":7504,"tokens_out":2676,"would_cite":true,"duration_ms":24366,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["49Q20","49J45"],"pacs":[],"model":"deepseek-v4-flash","headline":"In one dimension, convex sub-level sets of the potential force every mass-constrained free-energy minimizer to be an interval.","keywords":["free energy minimization","crystal equilibrium shape","convex sub-level sets","one-dimensional variational problem","rearrangement inequality","optimal transport","interval minimizer","calculus of variations"],"falsifier":"Set $g(0)=0$ and $g(x)=1$ for $x\\neq 0$, and fix $m>0$. Then $E((0,m)+k)=2+m$ for every integer $k$, so the translations $a_k=k$ are minimizing with $|a_k|\\to\\infty$; this directly refutes the boundedness assertion used to prove existence. Any complete proof must add a coercivity assumption or replace that compactness step.","tokens_in":6506,"feed_emoji":"💎","tokens_out":10517,"duration_ms":98522,"temperature":0.7,"pith_summary":"This paper answers a one-dimensional version of a classical variational question: if a crystal of fixed mass minimizes free energy under a potential with convex sub-level sets, must the minimizing shape be convex? The author's theorem says yes in one dimension: under $g(0)=0$, $g\\ge 0$, and convex sub-level sets, the minimum is attained and every minimizer is an interval. The argument first shows that convex sub-level sets are equivalent to the potential being monotone on each half-line, then proves a rearrangement inequality that any candidate set has energy at least that of an interval of the same mass. This reduces the whole problem to choosing a translation, and the optimal translation $\\alpha$ satisfies $g(\\alpha+m)=g(\\alpha)$.","feed_headline":"Convex sublevels force 1D crystal shapes to be intervals","feed_subtitle":"With $g(0)=0$, $g\\ge 0$ and convex sub-level sets, the minimum energy is a translated interval of prescribed mass.","key_machinery":"Two claims carry the argument. Claim 1: the sub-level sets $\\{g<t\\}$ are convex if and only if $g$ is non-decreasing on $[0,\\infty)$ and non-increasing on $(-\\infty,0]$. Claim 2: for any set $E$ with $|E|<\\infty$, $\\int_{E_+} g\\,dx \\ge \\int_0^{|E_+|} g\\,dx$ and $\\int_{E_-} g\\,dx \\ge \\int_{-|E_-|}^0 g\\,dx$; therefore the energy of $E$ is at least the energy of the interval $(-|E_-|,|E_+|)$. The proof of Claim 2 is given twice: once by approximating $E_+$ with disjoint intervals and translating them leftward, and once by an optimal transport map that pushes the excess measure left while $g$ is monotone. The minimization then depends only on the one-variable function $a \\mapsto 2+\\int_a^{a+m} g(x)\\,dx$, whose critical points satisfy $g(\\alpha+m)=g(\\alpha)$.","core_discovery":"The paper's central claim is Theorem 1.1: for $n=1$, $m\\in(0,\\infty)$, $g(0)=0$, $g\\ge 0$, and convex sub-level sets $\\{g<t\\}$, the infimum of $E(E)=\\mathcal H^0(\\partial E)+\\int_E g(x)\\,dx$ over sets of measure $m$ is attained, and every minimizer is convex. The author's argument concludes that a minimizer must be one of the four intervals $(0,m)+\\alpha$, $[0,m)+\\alpha$, $(0,m]+\\alpha$, or $[0,m]+\\alpha$, with $\\alpha$ determined by the first-order condition $g(\\alpha+m)=g(\\alpha)$. The interval reduction comes from the rearrangement inequality in Claim 2: the potential energy of any set is no smaller than that of the interval $(-|E_-|,|E_+|)$, which has boundary cost $2$. The theorem is presented as an extension of prior results that required coercivity, radial symmetry, or higher regularity of the potential.","pith_inferences":["For a potential that is zero at the origin and positive and flat elsewhere, such as $g(0)=0$ and $g(x)=1$ for $x\\neq 0$, the interval energy is the same at arbitrarily large translations, so the boundedness step in the existence proof fails; the classification of minimizers as intervals still holds whenever a minimizer exists.","The optimal-transport proof of the rearrangement estimate suggests a quantitative stability statement: the energy gap between a set and its interval rearrangement should be controlled by how much $g$ grows under the leftward transport, giving control on the distance from near-minimizers to intervals.","In a thin strip with a potential that is large outside the strip, the one-dimensional theorem plausibly survives as a limit: as the strip width goes to zero, minimizers converge to the intervals described here."],"forward_implications":["Any minimizer is an interval of length $m$, so in one dimension the equilibrium crystal has exactly two boundary points.","The optimal translation is computable from the potential: it solves $g(\\alpha+m)=g(\\alpha)$; for even monotone potentials it is $\\alpha=-m/2$.","The energy comparison reduces to one variable, making the constrained minimization a root-finding problem rather than a shape problem.","When $g$ is coercive ($g(x)\\to\\infty$ as $|x|\\to\\infty$), the compactness step needed for existence is valid, so the full theorem applies in that class."],"supporting_citations":[{"why":"It supplies the founding principle that crystal equilibrium shape minimizes free energy under a mass constraint.","marker":"[Gib78]"},{"why":"It gives an independent early formulation of the same free-energy minimization principle.","marker":"[Cur85]"},{"why":"It provides background and the statement of the convex-sublevel problem, as well as the small-mass regime context.","marker":"[FM11]"},{"why":"It establishes the two-dimensional convexity result under extra assumptions and exhibits a convex potential with no minimizer, motivating the one-dimensional question.","marker":"[Ind24a]"},{"why":"It proves connectedness of drops in convex potentials under regularity and coercivity assumptions that the present theorem relaxes.","marker":"[DPG22]"},{"why":"It gives equilibrium shapes for planar crystals in an external field under symmetry and coercivity; the present translation condition refines the role of the potential's zero level.","marker":"[McC98]"}],"fun_headline_variants":["Convex sublevels mean 1D crystals are intervals","1D crystal: convex sublevels guarantee interval minimizers","In 1D, convex sublevels force interval minimizers","1D crystal shape proven to be an interval under convex sublevels"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"In the existence part of the proof, Section 2 assumes that any minimizing sequence of interval translations has bounded positions; this fails for potentials that are zero at the origin and positive and constant elsewhere, because intervals far away achieve the same energy, so the compactness argument needs an added coercivity assumption or an alternative mechanism.","fun_headline_variants_meta":{"raw":{"variants":["Convex sublevels mean 1D crystals are intervals","1D crystal: convex sublevels guarantee interval minimizers","In 1D, convex sublevels force interval minimizers","1D crystal shape proven to be an interval under convex sublevels"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001249,"raw_usage":{"total_tokens":5059,"prompt_tokens":822,"completion_tokens":4237,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":438,"completion_tokens_details":{"reasoning_tokens":4167}},"tokens_in":438,"tokens_out":4237,"duration_ms":28389,"temperature":1.0,"reasoning_tokens":4167,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T20:31:41.584033+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Set $g(0)=0$ and $g(x)=1$ for $x\\neq 0$, and fix $m>0$. Then $E((0,m)+k)=2+m$ for every integer $k$, so the translations $a_k=k$ are minimizing with $|a_k|\\to\\infty$; this directly refutes the boundedness assertion used to prove existence. Any complete proof must add a coercivity assumption or replace that compactness step.","supporting_citations":[],"review_version":1}