{"id":"c9d98c9e-b3c5-4aa1-a0b7-c68ebbe64905","arxiv_id":"1908.04361","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For the Heisenberg group with the balanced metric, the paper proves a product splitting, transfers a known asymptotic Dirichlet theorem to build minimal graphs, and proves a new exterior Dirichlet problem with foliation.","lead":"This paper proves that the Heisenberg group with a 'balanced' metric, the sum of the left and right invariant metrics, splits as a product of a surface and a line, and then builds complete minimal surfaces by solving two boundary value problems. It matters because it gives explicit minimal surfaces in a metric Lie group that is not homogeneous.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The unproved claim that the collars Ω_r inherit the exterior circle condition is load-bearing; the full condition is false for Euclidean annuli, and no argument covers the outer-boundary supporting disks on which the barrier rests.","rationale":"The stress-test pass finds a structural gap rather than an evident false theorem. The proof of Theorem 3 hinges on the collars Ω_r having supporting geodesic disks at outer boundary points; this is asserted in one sentence and used to derive the bound (4). The reader's weakest assumption coincides with this, though our reading sharpens it: the exterior circle condition as written is likely false for Ω_r at the inner boundary, as the Euclidean annulus example shows, so the authors must have intended only the outer-boundary version. That weaker version is unproved for T and cannot be taken for granted because T has variable negative curvature. The remaining machinery—the explicit function f, the maximality argument with T_m, and the comparison-principle limit—is coherent once the geometric estimate is available. The theorem may well be true, and the gap appears repairable by adding a lemma on distance collars of geodesically convex sets in Cartan-Hadamard surfaces, so the verdict remains CONDITIONAL. We partial-agree with the reader because the identified risk is exactly that f(r(x)) may cease to be a subsolution, but the precise defect is that the global property asserted is stronger than what is used and the weaker property is missing.","tokens_in":8221,"tokens_out":18845,"duration_ms":208851,"concrete_test":"Take Ω to be a geodesic disk centered at e with radius ρ in T, which certainly satisfies the exterior circle condition. Using the explicit distance formula r(x,y,z) = √2√(x²+y²) and the geodesic parametrization from Theorem 1(d), construct the outer parallel ∂Ω_r for several values of r. For a dense sample of p ∈ ∂Ω_r, solve numerically for a point q such that the geodesic sphere centered at q and passing through p bounds a disk containing Ω_r, equivalently max_{x∈Ω_r} d(q,x) = d(q,p). If such q fails to exist for some r and p, Theorem 3's barrier construction is invalid. The analytical companion is to prove that for a C^{2,α} geodesically convex Ω in T every outer parallel point admits a supporting geodesic disk containing the collar; establishing that lemma would fill the missing step.","verdict_should_be":"UNCHANGED","load_bearing_attack":"In §4.1, after defining Ω_r = {x ∈ Λ : r(x) < r}, the paper states without proof that Ω_r satisfies the exterior geodesic circle condition and then, for each p ∈ ∂Ω_r\\∂Ω, takes a geodesic disk D_p through p containing Ω_r. This is the hinge of the proof: the curvature lower bound (4), and hence the subsolution property M[f(r)] ≥ 0, depends entirely on those supporting disks. The asserted global statement is not true in the model case: if Ω is a Euclidean disk, Ω_r is an annular collar, and at a point of the inner boundary any convex disk containing Ω_r would also contain the convex hull of Ω_r, so it cannot have p on its boundary. The proof only uses the outer-boundary case p ∈ ∂Ω_r\\∂Ω; that case is plausible (it holds in Euclidean for convex Ω), but it is not proved for T, where the curvature is negative but nonconstant. The normal exponential map from ∂Ω and the geometry of large geodesic spheres in T are never invoked to justify the existence of D_p. Thus there is a real gap between the geometric hypothesis on Ω and the construction of v_s(x) = f(r(x)).","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies Nil3 equipped with the balanced metric g, the sum of the left-invariant and right-invariant metrics. Theorem 1 establishes that Nil3 splits isometrically as a Riemannian product T×Z, where T is a totally geodesic, rotationally symmetric Hadamard surface, and it computes the geodesics, the distance function, and the curvature of T. Theorem 2 is stated as the asymptotic Dirichlet problem for the minimal surface equation on T, but it is a direct restatement of [15, Cor. 1.2] rather than a new proof. The main new result is Theorem 3: for a C^{2,α} bounded domain Ω⊂T satisfying the exterior geodesic circle condition, for every s≥0 there is a solution us of the exterior Dirichlet problem M[u]=0 in Λ=T\\Ω with us=0 on ∂Λ and sup_{∂Λ}|∇us|=s, and the graphs of these solutions foliate an open set of Nil3. The proof constructs barriers of the form f(r(x)), where r is the distance to ∂Ω, uses the comparison principle on bounded domains Ω_m={x∈Λ:r(x)<m}, and passes to a limit by elliptic compactness.","tokens_in":8472,"tokens_out":15102,"duration_ms":164359,"significance":"If the main result is correct, the paper gives a reasonably concrete construction of complete properly embedded minimal graphs in Nil3 for a non-homogeneous metric, including a foliation with prescribed boundary slope. The splitting theorem and the explicit curvature and distance formulas are useful and clearly presented. The paper is also honest about relying on the authors' prior work for the asymptotic Dirichlet problem and for the comparison and regularity tools. However, the proof of Theorem 3 rests on several unproved geometric assertions, most notably the inheritance of a global supporting-disk property by the parallel domains Ω_r and the control of the disk radii R(r). These issues are load-bearing, so the significance of the paper is conditional until they are resolved.","major_comments":[{"comment":"The assertion 'One may see that Ω_r satisfies the exterior geodesic circle condition' is load-bearing and is not proved. The subsequent construction of a geodesic disk D_p through p containing Ω_r for every p∈∂Ω_r\\∂Ω, and hence the curvature lower bound (4) and the subsolution property M[f(r)]≥0, all depend on this assertion. The full statement is false in the Euclidean model case of an annular collar: at a point of the inner boundary no Euclidean disk with that point on its boundary can contain the collar. The proof only uses p on the outer boundary, but even the restricted outer-boundary statement needs an argument in T, where geodesic disks are not Euclidean and the distance function to ∂Ω may develop focal points. Please supply a proof of the needed supporting-disk property, or modify the construction so that it does not require such a strong unproved geometric claim.","section":"§4.1, definition of Ω_r and the barrier construction"},{"comment":"The passage 'Take R(r)>0 such that ∪_{p∈∂Ω_r\\∂Ω} D_p ⊂ D_{R(r)} ... Since ∂Ω is compact it follows from the triangle inequality that lim_{r→∞} R(r)/r=1' is not justified. The triangle inequality bounds the distance from e to points of Ω_r by max_{∂Ω} d(e,·)+r, but it gives no control on the radii of the arbitrarily chosen supporting disks D_p or on the positions of their centers. The bound (4), in particular the numerator r−α, requires that the disks D_p can be chosen inside a ball centered at e with radius r+O(1). This asymptotic control is essential for the ODE barrier to work and must be proved, for example by explicitly constructing the supporting disks from the normal exponential map along ∂Ω.","section":"§4.1, paragraph following the definition of R(r)"}],"minor_comments":[{"comment":"There are typographical errors in the abstract ('se t', 'o f') and in Proposition 4 ('around around'), and the formula in Theorem 1(e) should be written as K=−4(r^2+12)/(r^2+8)^2 to avoid the ambiguous reading on the printed page.","section":"Abstract and Theorem 1(e)"},{"comment":"The displayed formula for M[u] and the sentence 'Since Δv_s is the curvature the ∂Ω_r\\∂Ω' are confusing: it is Δr, not Δv_s, that is the geodesic curvature of the level curves, and the displayed inequality appears to contain an extra factor f'(r). The intended sufficient condition is f''(r)+f'(r)(1+f'(r)^2)Δr≥0, and the text should be corrected accordingly.","section":"§4.1, formula for M[u]"},{"comment":"The exterior geodesic circle condition is phrased as 'given p∈Ω̄'; it should presumably be 'given p∈∂Ω', since for points in the interior the condition is not what is used in the proof.","section":"Theorem 3 statement"}],"recommendation":"major_revision","confidential_remarks":"The paper's genuinely new contribution is Theorem 3, and its proof has two unproved geometric steps that are essential: the inheritance of the exterior circle condition by the parallel domains Ω_r and the asymptotic control R(r)/r→1. These are not mere presentation issues; they are exactly the hypotheses that make the barrier function a subsolution. I recommend major revision rather than rejection because the overall strategy is plausible and the gaps appear to be fillable with a careful geometric lemma. I also note that Theorem 2 is a restatement of the authors' own cited result, so the novelty of the paper should be presented as Theorem 3 and the splitting theorem."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Paper gives a clean theorem (Theorem 1) on the balanced metric on Nil3: it splits as T×Z, with T a totally geodesic rotationally symmetric Hadamard surface, and explicit distance and curvature formulas. The curvature is negative, nonconstant, and the symmetry makes T a useful test space for minimal surface PDEs. Theorem 2 is the authors' own asymptotic Dirichlet result applied to T—not new, and they say so. The real new item is Theorem 3: an exterior Dirichlet problem on T with prescribed boundary gradient and a foliation conclusion. The PDE machinery is standard and mostly well handled: the barrier construction, subsolution comparison, C^{2,α} bounds, and the limiting argument all look sound. The catenoid example in Proposition 4 is a nice explicit surface.\n\nThe problem is the geometric hinge in §4.1. The proof says 'One may see that Ω_r satisfies the exterior geodesic circle condition' and then, for each p on the outer parallel boundary ∂Ω_r\\∂Ω, takes a geodesic disk D_p containing Ω_r with p on its boundary. That is exactly what makes the curvature lower bound (4) work. The assertion is not justified. The full condition for all of Ω_r cannot hold—at points of the original boundary ∂Ω, no disk containing the collar passes through that boundary point. For the only case that matters (p on the outer boundary), the Euclidean analogue is true, so the claim is plausible, but T is not Euclidean and the proof never gives an argument using the actual geometry of T, such as the normal exponential map or the behavior of large geodesic spheres. Also, the verification that T satisfies the hypotheses of the self-cited asymptotic Dirichlet result is skipped, and the R(r)/r→1 step is a one-line appeal to compactness. These are addressable gaps, not signs that the theorem is false.\n\nSerious thinker: yes. The paper is honestly written and the main existence claim is likely correct. For peer review, send it out; a good referee can fix the supporting-disk lemma or steer the authors to a replacement. Would I cite it? Probably not in my own work, but it deserves citation in the minimal-surfaces-in-Lie-groups subfield.","headline":"Solid niche paper: new splitting and curvature formulas for a balanced Heisenberg metric, plus an exterior Dirichlet theorem whose proof has a genuine but repairable geometric gap.","tokens_in":9011,"tokens_out":7301,"would_cite":false,"duration_ms":78600,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["53C42","35J93"],"pacs":[],"model":"deepseek-v4-flash","headline":"A balanced metric on the Heisenberg group splits the space and supplies minimal surfaces with any prescribed boundary slope.","keywords":["minimal surface equation","Heisenberg group Nil3","balanced metric","asymptotic Dirichlet problem","exterior Dirichlet problem","foliation by minimal graphs","totally geodesic surface","catenoids"],"falsifier":"Take $\\Omega$ to be a geodesic disk $D_R(e)$ in $T$, so $\\Omega_r$ is the annulus $D_{R+r}(e)\\setminus D_R(e)$. For a point $p$ on the inner boundary $\\partial D_R(e)$, a geodesic circle through $p$ whose disk contains $\\Omega_r$ would have to enclose the outer boundary at distance $R+r$ while passing through $p$ at distance $R$ from $e$. Using the explicit distance formula $r=\\sqrt{2}\\sqrt{x^2+y^2}$ and the rotational symmetry of $T$, one can compute for all $R>0$ and $r>0$ whether such a circle exists; if some pair admits no such circle, then $\\Omega_r$ fails the exterior geodesic circle condition and the proof's key estimate (4) lacks support for the simplest possible domain.","tokens_in":8026,"feed_emoji":"📐","tokens_out":17685,"duration_ms":155075,"temperature":0.7,"pith_summary":"The paper establishes that the Heisenberg group $\\mathrm{Nil}_3$ with the balanced metric (the sum of its left- and right-invariant metrics) is isometric to a Riemannian product $T\\times Z$, where $T$ is a totally geodesic surface and $Z$ is the center. It then proves that the minimal surface equation over $T$ is solvable in two complementary senses: for any continuous data on the boundary at infinity there is a unique solution, and for any bounded obstacle $\\Omega$ with the exterior geodesic circle property and any $s\\ge 0$ there is a solution of the exterior Dirichlet problem in $T\\setminus\\Omega$ with zero boundary values and boundary gradient magnitude $s$. The graphs of these exterior solutions are complete minimal surfaces in $\\mathrm{Nil}_3$, and the family is strictly ordered, so the graphs foliate an open subset of $\\mathrm{Nil}_3$. If the paper is right, the balanced metric makes $\\mathrm{Nil}_3$ a source of many embedded minimal surfaces controlled by boundary data, despite the metric being nonhomogeneous.","feed_headline":"Every prescribed slope gives a minimal surface in the Heisenberg group","feed_subtitle":"Splitting the group into a surface times its center turns minimal-surface problems into boundary data.","key_machinery":"The load-bearing identity is the isometry $\\Psi:T\\times Z\\to\\mathrm{Nil}_3$, $\\Psi((x,y,xy/2),(0,0,t))=(x,y,xy/2+t)$, which makes $T$ a totally geodesic surface and turns graphs over $T$ into surfaces in $\\mathrm{Nil}_3$. The proof of Theorem 3 is carried by radial barriers $v_s(x)=f(r(x))$, where $r(x)=\\operatorname{dist}(x,\\partial\\Omega)$; applying the minimal surface operator to $f\\circ r$ reduces the PDE to an ODE controlled by the curvature of the level sets of $r$. The explicit bounded function $f(r)=\\int_0^r \\frac{c e^{\\sqrt{2}\\alpha\\arctan((t+\\alpha)/2^{3/2})}}{(t+\\alpha)^2+8}\\,dt$ has $f(0)=0$ and $f'(0)=s$, and it is a subsolution precisely when the curvature lower bound (4) holds; that bound comes from the tangency principle, the Hessian comparison theorem, and the curvature formula $K=-4(r^2+12)/(r^2+8)^2$. The comparison principle then forces the approximating sequence of solutions to converge to $u_s$ with the required boundary gradient.","core_discovery":"The central claim is that the balanced metric makes $\\mathrm{Nil}_3$ rigid enough to split as $T\\times Z$, and that this splitting reduces the minimal surface equation to a problem on the surface $T$, with distance $r=\\sqrt{2}\\sqrt{x^2+y^2}$ and curvature $K=-4(r^2+12)/(r^2+8)^2$. Theorem 3 states that for any $C^{2,\\alpha}$ domain $\\Omega\\subset T$ satisfying the exterior geodesic circle condition and any $s\\ge 0$, the problem $M[u]=0$ in $\\Lambda=T\\setminus\\Omega$, $u|_{\\partial\\Omega}=0$, has a solution $u_s$ with $\\sup_{\\partial\\Lambda}\\|\\nabla u\\|=s$. The paper proves that $u_{s_1}<u_{s_2}$ in $\\Lambda$ when $s_1<s_2$, so the graphs of the $u_s$ form a foliation of an open subset of $\\mathrm{Nil}_3$, and $\\limsup_{r\\to\\infty}(u_{s_2}-u_{s_1})>0$. Theorem 2 gives a unique smooth solution for every continuous boundary function at infinity of $T$, so in particular there are infinitely many non-congruent foliations of $\\mathrm{Nil}_3$ by complete properly embedded minimal surfaces transverse to the center.","pith_inferences":["The product splitting $T\\times Z$ depends on using the two invariant metrics with equal weight; a natural next step is to test the same Dirichlet theorems for the one-parameter family with unequal weights, where the splitting should fail and the barrier construction may break down.","The only structural hypothesis in Theorem 3 is the exterior geodesic circle condition, so a valuable test is whether the curvature lower bound (4) can be derived directly from a growth condition on the geometry of $\\partial\\Omega$; if so, the same subsolution argument would apply to more general rotationally symmetric surfaces.","The steepening of $u_s$ at $\\partial\\Omega$ as $s$ grows suggests that the family may converge, after normalization, to the vertical cylinder over $\\partial\\Omega$ or to one of the explicit catenoids; identifying that limit would give a geometric description of the foliation's asymptotic end."],"forward_implications":["For every $s\\ge 0$ there is a complete minimal graph over $T\\setminus\\Omega$ that vanishes on $\\partial\\Omega$ and has boundary gradient magnitude exactly $s$, so the exterior Dirichlet problem has a continuum of ordered solutions.","Because $u_{s_1}<u_{s_2}$ for $s_1<s_2$, the graphs $\\{u_s\\}$ form a genuine foliation of an open subset of $\\mathrm{Nil}_3$, with every surface in the family embedded and disjoint from the others.","The unique solvability of the asymptotic Dirichlet problem yields infinitely many non-congruent foliations of $\\mathrm{Nil}_3$ by complete properly embedded minimal surfaces transversal to the center.","When $\\partial\\Omega$ is a geodesic circle, the explicit catenoids of Proposition 4 are concrete members of the foliation, connecting the abstract existence result to computable rotationally symmetric examples."],"supporting_citations":[{"why":"Provides the Riemannian-setting existence method for the exterior Dirichlet problem on which Theorem 3 is based.","marker":"[3]"},{"why":"Supplies the Euclidean exterior Dirichlet problem results that Theorem 3 adapts to the surface $T$.","marker":"[14]"},{"why":"Yields the asymptotic Dirichlet problem for complete minimal graphs on rotationally symmetric Hadamard surfaces, quoted as Theorem 2.","marker":"[15]"},{"why":"Supplies the comparison principle and the a-priori estimates used to extract convergent subsequences in the proof of Theorem 3.","marker":"[16]"},{"why":"Supplies the elliptic regularity theory that upgrades limiting solutions to $C^{2,\\alpha}$.","marker":"[5]"}],"fun_headline_variants":["Heisenberg minimal surfaces exist for every slope","Balanced metric splits Heisenberg into surface times center","Exterior and asymptotic Dirichlet solved on Heisenberg","Every boundary slope gives a Heisenberg minimal foliation","Heisenberg group foliated by minimal surfaces"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"In the proof of Theorem 3, the paper asserts with 'One may see' that the intermediate domains $\\Omega_r=\\{x\\in T\\setminus\\Omega:\\operatorname{dist}(x,\\partial\\Omega)<r\\}$ satisfy the exterior geodesic circle condition for every $r$; if this assertion fails, the curvature inequality (4) that makes the barrier $f(r(x))$ a subsolution has no support, and the proof of the theorem collapses.","fun_headline_variants_meta":{"raw":{"variants":["Heisenberg minimal surfaces exist for every slope","Balanced metric splits Heisenberg into surface times center","Exterior and asymptotic Dirichlet solved on Heisenberg","Every boundary slope gives a Heisenberg minimal foliation","Heisenberg group foliated by minimal surfaces"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000241,"raw_usage":{"total_tokens":1554,"prompt_tokens":1013,"completion_tokens":541,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":629,"completion_tokens_details":{"reasoning_tokens":479}},"tokens_in":629,"tokens_out":541,"duration_ms":6245,"temperature":1.0,"reasoning_tokens":479,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:47:11.224510+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take $\\Omega$ to be a geodesic disk $D_R(e)$ in $T$, so $\\Omega_r$ is the annulus $D_{R+r}(e)\\setminus D_R(e)$. For a point $p$ on the inner boundary $\\partial D_R(e)$, a geodesic circle through $p$ whose disk contains $\\Omega_r$ would have to enclose the outer boundary at distance $R+r$ while passing through $p$ at distance $R$ from $e$. Using the explicit distance formula $r=\\sqrt{2}\\sqrt{x^2+y^2}$ and the rotational symmetry of $T$, one can compute for all $R>0$ and $r>0$ whether such a circle exists; if some pair admits no such circle, then $\\Omega_r$ fails the exterior geodesic circle condition and the proof's key estimate (4) lacks support for the simplest possible domain.","supporting_citations":[{"cited_title":"do Esp ´ ırito-Santo, J","cited_arxiv_id":null,"evidence_quote":"Provides the Riemannian-setting existence method for the exterior Dirichlet problem on which Theorem 3 is based."},{"cited_title":"Ripoll: Some characterization, uniqueness and existence results f or Euclidean graphs of constant mean curvature with planar bou ndary, Pa- cific Journal of Mathematics, v","cited_arxiv_id":null,"evidence_quote":"Supplies the Euclidean exterior Dirichlet problem results that Theorem 3 adapts to the surface $T$."},{"cited_title":"Ripoll, M","cited_arxiv_id":null,"evidence_quote":"Yields the asymptotic Dirichlet problem for complete minimal graphs on rotationally symmetric Hadamard surfaces, quoted as Theorem 2."},{"cited_title":"Ensaios de Matem´ atica","cited_arxiv_id":null,"evidence_quote":"Supplies the comparison principle and the a-priori estimates used to extract convergent subsequences in the proof of Theorem 3."},{"cited_title":"Elliptic Partial Differential Equations of Second Order","cited_arxiv_id":null,"evidence_quote":"Supplies the elliptic regularity theory that upgrades limiting solutions to $C^{2,\\alpha}$."}],"review_version":1}