Calibrating Forms for Minimal Graphs in Arbitrary Codimension

Abstract: We introduce a new family of closed differential forms naturally associated with minimal graphical submanifolds in Euclidean space, defined in arbitrary codimension. For each minimal graph, we construct an explicit closed form whose restriction coincides with the induced volume form. These forms admit a geometric interpretation as pullbacks, via the Gauss map, of tautological differential forms on the Grassmannian. In contrast to most known calibrations, they are generally not parallel and do not arise from special holonomy or symmetry considerations. The calibration problem is thus reduced to estimating the pointwise comass of the constructed forms. We show that the comass bound can be characterized in terms of explicit inequalities involving the singular values of the defining map of the graph, formulated via its two-dilations and we identify precise conditions ensuring that the comass is at most one. As a consequence, any minimal graph satisfying these conditions is calibrated and hence area-minimizing. This yields a broad class of new calibrated minimal graphs, extending the classical codimension-one theory, and provides an effective criterion for determining precisely where a given minimal graph is area-minimizing. As an application of our construction, we confirm a conjecture of Lawson and Osserman under two-dilation conditions, in arbitrary codimesnion.

• The paper introduces an explicit family of closed differential forms as calibrations to certify area-minimization for minimal graphs using singular value-based inequalities.
• The paper presents three equivalent formulations, including an SVD-based approach, linking the calibration property directly with the minimal graph PDE system.
• The paper demonstrates that sharp 2-dilation constraints resolve the Lawson–Osserman conjecture and ensure regularity for Lipschitz minimal graphs.

Calibrating Forms for Minimal Graphs in Arbitrary Codimension — Summary and Analysis

Introduction and Motivation

The central issue addressed is the characterization of area-minimizing minimal submanifolds in Euclidean space, particularly in the context of minimal graphs with arbitrary codimension. The established theory of calibrations—where closed differential forms with pointwise comass bounded by one provide area minimization certificates—has seen limited applicability. Classical calibrations mainly arise from parallel forms linked to special holonomy or high-symmetry situations, with general constructions for minimal graphs only well-understood in codimension one or specific integrable geometric contexts (e.g., special Lagrangian geometry).

This paper advances the theoretical framework considerably by constructing a new, explicit family of closed differential forms for minimal graphical submanifolds in arbitrary codimension. These forms serve as calibrations, and their applicability reduces to concrete, singular-value-based inequalities derived from the defining map of the graph. The approach generalizes and subsumes previous codimension-one results and provides new analytic and geometric tools for the area-minimizing property of minimal graphs.

Construction of Calibrating Forms

Given a smooth map $F:\Omega \to \mathbb{R}^m$, with its graph $\Sigma \subset \mathbb{R}^{n+m}$, the authors define a closed $n$-form $\Theta(F)$ on $\mathbb{R}^{n+m}$. This form is uniquely characterized by the property that its restriction to the graph coincides with the induced metric volume form. The construction leverages the Gauss map, pulling back universal tautological forms from the Grassmannian $Gr(n, n+m)$.

Importantly, the constructed calibrations are generally not parallel and do not originate from special holonomy, distinguishing them from nearly all previous constructions in the theory. The calibration property thus becomes an explicit estimate on the comass of these forms, making the problem accessible to analysis in terms of the singular values of $DF$.

Three equivalent formulations for $\Theta(F)$ are provided: a frame-independent (Gauss map) version, an SVD-based normal/tangent frame suitable for comass estimates, and a local coordinate-based expression facilitating PDE analysis.

Calibration and the Minimal Graph System

The closedness of $\Theta(F)$ is shown to be equivalent to the minimal graph system—the (nonlinear, elliptic) PDE that characterizes a minimal submanifold written as a graph—holding for $F$. Specifically, for Lipschitz or $\Sigma \subset \mathbb{R}^{n+m}$0 functions, the vanishing of $\Sigma \subset \mathbb{R}^{n+m}$1 corresponds exactly to the mean curvature vector vanishing (or, equivalently, the stationarity of the volume under ambient deformations).

For codimension two, the explicit calibration and the area-non-increasing (2-dilation $\Sigma \subset \mathbb{R}^{n+m}$2) condition provide a direct and verifiable criterion for the area-minimizing property of minimal graphs.

Sharp Singular Value (2-Dilation) Estimates

A significant portion of the analysis is devoted to estimating the comass of $\Sigma \subset \mathbb{R}^{n+m}$3 in terms of the singular values $\Sigma \subset \mathbb{R}^{n+m}$4 of the differential $\Sigma \subset \mathbb{R}^{n+m}$5. The precise combinatorial and analytic structure allows the authors to obtain the following main result: For rank $\Sigma \subset \mathbb{R}^{n+m}$6 (the essential supremum of the rank of $\Sigma \subset \mathbb{R}^{n+m}$7), the condition

$\Sigma \subset \mathbb{R}^{n+m}$8

guarantees the comass of $\Sigma \subset \mathbb{R}^{n+m}$9 does not exceed one. Thus, any minimal graph obeying these 2-dilation constraints is area-minimizing. This generalizes the classical maximum principle for codimension one ($n$0) and codimension two (previously considered by multiple authors) to all $n$1.

Sharper (but more technical) versions refine the $n$2 bound to $n$3 for some absolute constant, exploiting symmetric function inequalities and higher-order term estimates.

Consequences and Applications

Extension of Area-Minimizing Theory

The calibration criteria provide an effective, computable method for verifying when a minimal graph is area-minimizing, allowing for precise localization in terms of the singular values of $n$4. This not only recovers but extends previous results (Lawlor-Morgan, codimension one Bernstein-type theorems) and applies equally to graphs of Lipschitz maps, via the theory of real flat chains and currents.

Lawson-Osserman Conjecture

A striking corollary is the resolution, under 2-dilation assumptions, of the Lawson–Osserman conjecture: Any Lipschitz solution of the weak minimal graph system that satisfies the 2-dilation condition is in fact stationary (hence, smooth and truly minimal). This bridges a previously open gap in the regularity theory and connects the PDE and geometric-measure-theoretic formulations of minimality.

Regularity and Further Implications

By combining the calibration property with blow-up arguments and Allard regularity, all potential singularities (for solutions satisfying the 2-dilation condition) are excluded, yielding interior smoothness for weakly minimal Lipschitz graphs. This advances the Bernstein-type regularity theory to higher codimension under verifiable analytic constraints.

Theoretical and Practical Outlook

The new family of calibrating forms fundamentally modifies the toolkit available for minimal submanifold theory in arbitrary codimension. The explicit singular value inequalities directly tie the analytic properties of the defining map to geometric minimality and regularity. Practically, this enables precise criteria for minimizing currents, effective in both local and global settings even under mere Lipschitz regularity.

Theoretically, this work opens several directions:

• Further refinement of the comass bounds could yield sharper geometric results and possibly characterize area-minimizing cones or singularities in more detail.
• The construction suggests possible generalizations to non-Euclidean ambient spaces via the geometry of Grassmannians and tautological bundles.
• Connections to stability, uniqueness, and moduli of minimal submanifolds in arbitrary codimension are now analytically accessible.

Conclusion

This work introduces a fundamentally new and explicit calibration construction for minimal graphs in arbitrary codimension, reducing area-minimizing verification to analytic inequalities on singular values. The framework subsumes and extends existing calibration results, confirms a long-standing conjecture (Lawson–Osserman) under natural geometric constraints, and synthesizes the analytic, PDE, geometric, and measure-theoretic aspects of the minimal surface system in a unified setting. The methods developed have substantial potential for further extension in geometric analysis and the calculus of variations.

Reference:

"Calibrating Forms for Minimal Graphs in Arbitrary Codimension" (2604.04336)