Conformal Normal Curvature (CNC)

• Conformal Normal Curvature (CNC) is a family of curvature measures that link normal vectors with conformal structure across varied geometric settings.
• In the fractional immersion framework, CNC is defined as a scale-invariant two-point energy of the Gauss map, controlling Sobolev regularity and compactness.
• For normal curves and Riemann coordinates, CNC captures deviation formulas and correction terms that govern curvature transformation under conformal maps.

Conformal Normal Curvature (CNC) is not introduced as a single explicit term in the cited literature; rather, it emerges as a reconstructed label for conformally meaningful curvature quantities tied to normals. In one setting, aligned with Schikorra’s analysis of conformal immersions, CNC is the scale-invariant fractional energy of the Gauss map. In another, developed for normal curves on immersed surfaces, CNC is the corrected difference between transformed and original normal curvature under an ambient conformal motion. In a third, more speculative normal-coordinate setting, CNC denotes curvature content in Riemann normal coordinates that is absorbed by a conformal factor. Taken together, these sources suggest that CNC is best understood as a family of normal-curvature constructions adapted to conformal geometry rather than a single standardized invariant (Schikorra, 2018, Lone, 2019, Siqueira, 2010).

1. Terminological scope and geometric setting

The most concrete formulation arises for conformal immersions $$\Phi \colon B \subset \mathbb{R}^2 \to \mathbb{R}^3$$, with $B$ a planar ball and $\Phi \in W^{1,2}(B,\mathbb{R}^3)$. Conformality means that, for almost every $x \in B$,

$$\langle \partial_i\Phi(x),\partial_j\Phi(x)\rangle = e^{2\lambda(x)}\delta_{ij}, \qquad i,j\in\{1,2\},$$

or equivalently there is an orthonormal tangent frame $(e_1,e_2)$ and a function $\lambda$ such that

$$\partial_\alpha \Phi = e^\lambda e_\alpha, \qquad \alpha=1,2.$$

The associated Gauss map is

$u = e_1 \wedge e_2 \colon B \to S^2.$

Under conformality, $\Phi$ is an immersion almost everywhere exactly when $B$0 almost everywhere, equivalently when $B$1 almost everywhere.

A distinct geometric setting is that of a unit-speed curve $B$2 lying on a smooth immersed surface $B$3. With Darboux frame $B$4 along the curve, the standard decomposition is

$B$5

$B$6

$B$7

Here $B$8 is the normal curvature, $B$9 the geodesic curvature, and $\Phi \in W^{1,2}(B,\mathbb{R}^3)$0 the geodesic torsion. In the terminology of the 2019 paper, a “normal curve” is a space curve whose position vector lies in its normal plane, equivalently $\Phi \in W^{1,2}(B,\mathbb{R}^3)$1.

A third usage appears only by inference in the 2010 normal-coordinate paper. There the phrase “Conformal Normal Curvature” is not defined, but the geometry combines normal coordinates, curvature coefficients, and an explicit conformal factor. This supports only an inferred, not explicit, CNC terminology.

2. Fractional CNC for conformal immersions

For $\Phi \in W^{1,2}(B,\mathbb{R}^3)$2, Schikorra’s framework introduces a two-point energy of the Gauss map,

$\Phi \in W^{1,2}(B,\mathbb{R}^3)$3

A coherent definition aligned with that paper is

$\Phi \in W^{1,2}(B,\mathbb{R}^3)$4

with $\Phi \in W^{1,2}(B,\mathbb{R}^3)$5 the Gauss map of the conformal immersion. Small CNC means $\Phi \in W^{1,2}(B,\mathbb{R}^3)$6 for the threshold in the main theorem (Schikorra, 2018).

The geometric content of this quantity comes from the identity $\Phi \in W^{1,2}(B,\mathbb{R}^3)$7 for unit vectors $\Phi \in W^{1,2}(B,\mathbb{R}^3)$8, where $\Phi \in W^{1,2}(B,\mathbb{R}^3)$9 is the angle between them. For $x \in B$0-valued maps one has

$x \in B$1

with a universal $x \in B$2. Thus the wedge product controls, and is controlled by, the spherical distance between normals. The energy is therefore an angular measure of oscillation of the Gauss map.

Its analytic significance is criticality. In dimension two, $x \in B$3 scales like $x \in B$4 under $x \in B$5, the kernel $x \in B$6 scales like $x \in B$7, and $x \in B$8 is scale-free. Hence $x \in B$9 is invariant under domain dilations. This is the critical, scale-invariant regime.

The same quantity is naturally compared with the fractional Sobolev seminorm. For $$\langle \partial_i\Phi(x),\partial_j\Phi(x)\rangle = e^{2\lambda(x)}\delta_{ij}, \qquad i,j\in\{1,2\},$$0 and $$\langle \partial_i\Phi(x),\partial_j\Phi(x)\rangle = e^{2\lambda(x)}\delta_{ij}, \qquad i,j\in\{1,2\},$$1,

$$\langle \partial_i\Phi(x),\partial_j\Phi(x)\rangle = e^{2\lambda(x)}\delta_{ij}, \qquad i,j\in\{1,2\},$$2

Because $$\langle \partial_i\Phi(x),\partial_j\Phi(x)\rangle = e^{2\lambda(x)}\delta_{ij}, \qquad i,j\in\{1,2\},$$3 for $$\langle \partial_i\Phi(x),\partial_j\Phi(x)\rangle = e^{2\lambda(x)}\delta_{ij}, \qquad i,j\in\{1,2\},$$4-valued $$\langle \partial_i\Phi(x),\partial_j\Phi(x)\rangle = e^{2\lambda(x)}\delta_{ij}, \qquad i,j\in\{1,2\},$$5, $$\langle \partial_i\Phi(x),\partial_j\Phi(x)\rangle = e^{2\lambda(x)}\delta_{ij}, \qquad i,j\in\{1,2\},$$6 is a natural angular variant of the fractional seminorm. A key proposition states that there exists $$\langle \partial_i\Phi(x),\partial_j\Phi(x)\rangle = e^{2\lambda(x)}\delta_{ij}, \qquad i,j\in\{1,2\},$$7 such that, if $$\langle \partial_i\Phi(x),\partial_j\Phi(x)\rangle = e^{2\lambda(x)}\delta_{ij}, \qquad i,j\in\{1,2\},$$8, then

$$\langle \partial_i\Phi(x),\partial_j\Phi(x)\rangle = e^{2\lambda(x)}\delta_{ij}, \qquad i,j\in\{1,2\},$$9

Thus small fractional normal curvature yields control of the standard fractional Sobolev regularity of the Gauss map.

This construction is lower order than an $(e_1,e_2)$0 bound on the second fundamental form, but it remains scale invariant. That balance between weaker differentiability and critical scaling is central to the role of CNC in compactness theory.

3. Compactness, rigidity, and gauge structure

The main compactness theorem considers smooth conformal immersions $(e_1,e_2)$1 with

$(e_1,e_2)$2

where $(e_1,e_2)$3 is an orthonormal frame and $(e_1,e_2)$4 is the Gauss map. If $(e_1,e_2)$5 weakly in $(e_1,e_2)$6 and

$(e_1,e_2)$7

then there is a dichotomy: either $(e_1,e_2)$8 is constant, or $(e_1,e_2)$9 is a conformal immersion almost everywhere. In the non-collapse case there exist $\lambda$0 and an orthonormal frame $\lambda$1 such that

$\lambda$2

The theorem is explicitly described as being in the spirit of results of T. Toro, S. Müller and V. Šverák, and F. Hélein, but under a weaker curvature hypothesis and with a correspondingly weaker conclusion (Schikorra, 2018).

The proof uses a fractional lifting theorem. When $\lambda$3 is small, one can rotate the tangent frame by a scalar angle $\lambda$4 to obtain a new orthonormal frame $\lambda$5 such that

$\lambda$6

and

$\lambda$7

This is presented as the fractional analogue of Hélein’s lifting for small Willmore energy, and it is proved using compensated compactness, commutator estimates, and fractional Wente-type inequalities.

The conformal factor is tied to the frame by the identity

$\lambda$8

After a gauge rotation,

$\lambda$9

Decomposing $$\partial_\alpha \Phi = e^\lambda e_\alpha, \qquad \alpha=1,2.$$0 with $$\partial_\alpha \Phi = e^\lambda e_\alpha, \qquad \alpha=1,2.$$1 harmonic and $$\partial_\alpha \Phi = e^\lambda e_\alpha, \qquad \alpha=1,2.$$2 yields

$$\partial_\alpha \Phi = e^\lambda e_\alpha, \qquad \alpha=1,2.$$3

A sharp fractional Wente inequality then controls $$\partial_\alpha \Phi = e^\lambda e_\alpha, \qquad \alpha=1,2.$$4 in terms of the fractional seminorms of the frame, and Moser–Trudinger gives exponential integrability for $$\partial_\alpha \Phi = e^\lambda e_\alpha, \qquad \alpha=1,2.$$5.

The final dichotomy depends on the harmonic part $$\partial_\alpha \Phi = e^\lambda e_\alpha, \qquad \alpha=1,2.$$6. If its negative part blows up, then the interior energy of $$\partial_\alpha \Phi = e^\lambda e_\alpha, \qquad \alpha=1,2.$$7 decays and the limit collapses to a constant. If it remains controlled, one obtains uniform interior $$\partial_\alpha \Phi = e^\lambda e_\alpha, \qquad \alpha=1,2.$$8 bounds on $$\partial_\alpha \Phi = e^\lambda e_\alpha, \qquad \alpha=1,2.$$9, compactness of the frames, and passage to the limit in the conformal system. This identifies CNC as a rigidity threshold for weak limits of conformal immersions.

The same paper relates the Gauss map directly to classical curvature. If $u = e_1 \wedge e_2 \colon B \to S^2.$0 is the second fundamental form and $u = e_1 \wedge e_2 \colon B \to S^2.$1 its components in the orthonormal tangent frame, then

$u = e_1 \wedge e_2 \colon B \to S^2.$2

and

$u = e_1 \wedge e_2 \colon B \to S^2.$3

Hence

$u = e_1 \wedge e_2 \colon B \to S^2.$4

This makes the fractional CNC energy a lower-order analogue of the classical $u = e_1 \wedge e_2 \colon B \to S^2.$5 curvature control behind Willmore-type compactness.

4. CNC for normal curves under ambient conformal maps

In the 2019 treatment of normal curves on surfaces, the central problem is how normal curvature and related position components transform under a conformal map $u = e_1 \wedge e_2 \colon B \to S^2.$6 with differential

$u = e_1 \wedge e_2 \colon B \to S^2.$7

where $u = e_1 \wedge e_2 \colon B \to S^2.$8 is the local dilation factor and $u = e_1 \wedge e_2 \colon B \to S^2.$9 is a local rotation. On a surface chart $\Phi$0, the first fundamental form transforms by

$\Phi$1

The Christoffel symbols change by explicit correction terms $\Phi$2 depending on derivatives of $\Phi$3 and the metric coefficients, equivalently by the standard conformal-change formula for Levi–Civita connections (Lone, 2019).

For a curve $\Phi$4, the normal curvature is

$\Phi$5

while the geodesic curvature is given by the Beltrami formula displayed in the source data. The paper derives explicit deviation formulas for the transformed curve $\Phi$6 on the conformal image surface $\Phi$7.

The normal component deviation is

$\Phi$8

or, in the paper’s abbreviated notation,

$\Phi$9

Here $B$00 is the normal curvature on $B$01, $B$02 is the original normal curvature, and $B$03 is the connection-type correction coming from the conformal change of the metric. This motivates the inferred definition

$B$04

The paper does not explicitly name this expression “CNC,” but it is exactly the bracket that governs the conformal behavior of the normal component of the position vector.

The tangential component deviation is likewise given explicitly:

$B$05

where $B$06 represent $B$07 in the chart basis and $B$08 collect conformal metric derivatives. Proposition 1 gives the corresponding transformation law for geodesic curvature,

$B$09

with $B$10 determined by the $B$11 corrections. In the homothetic case, where $B$12 is constant, all correction terms vanish and $B$13 and $B$14 scale by $B$15.

Several invariant statements then follow. For isometries, $B$16, geodesic curvature is invariant and the Christoffel symbols are identical. For homotheties, the correction terms $B$17, $B$18, $B$19, and $B$20 vanish, so the deviation formulas reduce to pure curvature-scaling relations. The paper also gives an invariant-sufficient condition for the conformal image of a normal curve to remain normal.

A frequent misconception is to read “normal curve” here as a geodesic-normal condition such as $B$21. The paper explicitly uses the space-curve notion: a normal curve is one whose position vector lies in its normal plane, equivalently $B$22. In this setting, CNC concerns the conformal behavior of normal curvature along such curves, not the geodesic condition.

5. Normal coordinates, conformal factors, and inferred CNC densities

The 2010 paper on normal coordinate transformations does not introduce the term “Conformal Normal Curvature,” but it develops a geometry in which such a notion can be inferred. The setting is an $B$23-dimensional pseudo-Riemannian metric written in Riemann normal coordinates around an origin where the normal-coordinate map is “well-behaved in the origin and in its neighborhood.” On the hypersurface $B$24, the line element is expressed exactly as

$B$25

and then rewritten in the conformal form

$B$26

with

$B$27

Here

$B$28

This scalar $B$29 is the curvature-dependent quantity that determines the conformal factor (Siqueira, 2010).

A consistent inferred terminology, explicitly labeled as such in the source data, has two versions. The first is a tensorial version:

$B$30

in a Riemann normal frame for $B$31, with the Cotton tensor used in dimension three. The second is a scalar density version:

$B$32

The first isolates conformal curvature content in the normal frame; the second measures, in the paper’s own variables, how much curvature is absorbed by the conformal factor.

The same paper further claims that, under its normal-coordinate hypothesis, the general metric is conformal to a metric of constant curvature and can be embedded in a hyper-cone of an $B$33-dimensional flat manifold. The embedding is given by

$B$34

with

$B$35

A point of tension is explicit in the source material itself. The same data block records the standard conformal-geometry facts that local conformal flatness in dimension $B$36 requires vanishing Weyl tensor, in $B$37 requires vanishing Cotton tensor, and in $B$38 is automatic. The paper does not discuss Weyl/Cotton conditions, even though it asserts local conformal flatness under its “well-behaved” normal-coordinate hypothesis. This makes the tensorial CNC interpretation especially important: it distinguishes conformal-curvature content that, in standard conformal geometry, obstructs conformal flatness.

6. Comparative interpretation, limitations, and open directions

Across these sources, CNC has three structurally different meanings. In the fractional immersion theory it is a scale-invariant two-point energy of the Gauss map. In the theory of normal curves on surfaces it is a corrected transformed-minus-original normal curvature. In the normal-coordinate setting it is an inferred conformal-curvature content, either tensorial or scalar. These are not equivalent objects. A plausible implication is that CNC should be regarded as a contextual designation for curvature quantities coupling normals to conformal structure, rather than as a single invariant.

The fractional formulation is distinguished by compactness theory. Its hypotheses are local on planar domains, the arguments crucially use $B$39, and the exact threshold $B$40 is not explicit. At $B$41 one recovers the classical setting; at the borderline $B$42, the paper mentions a possible substitute hypothesis, such as a uniform $B$43 bound on suitable extensions of $B$44, but not a direct $B$45-smallness criterion. The gain is weaker, lower-order, scale-invariant curvature control; the loss, relative to the $B$46 curvature theory of Toro, Müller–Šverák, and Hélein, is that one gets an almost everywhere conformal immersion rather than a bilipschitz parametrization, and in general no uniform $B$47 control of $B$48 without additional Lorentz-type bounds (Schikorra, 2018).

The normal-curve formulation is explicitly local and chart-based, assumes smooth immersed surfaces and regular curves, and does not provide a closed-form global transformation law for the second fundamental form or the shape operator under a general conformal map. Its strength lies in precise deviation formulas for normal and tangential components and in an invariant-sufficient criterion for preservation of the normal-curve property. Its limitations are exactly those of a local surface calculation: twice-differentiable charts, explicit dependence on Christoffel corrections, and a notion of “normal curve” that is extrinsic rather than geodesic (Lone, 2019).

The normal-coordinate formulation is likewise local. It assumes sufficient differentiability for Taylor expansions and a region where the exponential map is regular, geodesics do not mix, and no conjugate points occur. The paper does not pursue global issues. Its open interpretive problem is the relation between the claimed conformal flattening and the standard Weyl/Cotton obstructions already noted in the source data.

The open directions stated or suggested in the sources are correspondingly diverse. For fractional CNC, the energy is described as vaguely reminiscent of tangent-point energies and Menger-curvature-type surface energies studied by Strzelecki, von der Mosel, and collaborators; the stated hope is that the analysis can be used to define weak immersions with this kind of energy bound. For curve-based CNC, an open direction is a fully intrinsic characterization in terms of the second fundamental form and the conformal factor, together with extensions to discrete conformal maps and computational pipelines. For the normal-coordinate framework, the central unresolved issue is how its conformal rewriting interacts with standard conformal-curvature invariants.

In summary, CNC names a conformally organized normal-curvature quantity whose precise meaning depends on context. In conformal immersion theory it is

$B$49

and smallness yields the dichotomy between almost everywhere immersion and collapse. In the geometry of normal curves it is naturally reconstructed as

$B$50

which governs the deviation of the normal component of the position vector under conformal motion. In the normal-coordinate literature it is only inferred, either as conformal-curvature tensor content in a normal frame or as the scalar $B$51 driving the conformal factor. The common thread is the same: normals, curvature, and conformal change are linked through quantities that are local, structurally geometric, and sensitive to the distinction between intrinsic metric transformation and extrinsic curvature behavior.