Bubble Conformal Complexes

• Bubble conformal complexes are local exact sequences of bubble finite element spaces that build conforming analogues of 3D conformal Hessian and elasticity complexes.
• They are constructed via a localized Bernstein–Gelfand–Gelfand framework, ensuring interior degrees of freedom, exactness, and trace continuity.
• These complexes support discrete TT tensors, York decompositions, and structure-preserving finite element methods for applications in elasticity, relativity, and fluid mechanics.

Searching arXiv for papers on “bubble conformal complexes” and adjacent terms. Calling arXiv search for exact phrase and related finite-element conformal complexes. Bubble conformal complexes are local exact complexes of bubble finite element spaces used to construct conforming finite element analogues of the three-dimensional conformal Hessian and conformal elasticity complexes. In the finite-element literature, they arise by applying the Bernstein–Gelfand–Gelfand framework locally to bubble de Rham, Hessian, elasticity, and divdiv complexes, rather than globally to conforming spaces, and they serve as building blocks for global spaces involving symmetric traceless tensors and higher-order operators such as $devhess$, $\symcurl$, $divdiv$, $devdef$, and the linearized Cotton–York operator $cott$ (Huang, 2 Aug 2025). Within the broader framework of finite element complexes with trace structures, exact bubble complexes on simplices of every dimension encode interior degrees of freedom, determine the discrete cohomology, and support locally $L^2$-bounded commuting projections (Hu et al., 28 Sep 2025).

1. Continuous conformal complexes and conformal tensors

The underlying continuous objects are two conformal complexes in three dimensions. The conformal Hessian complex is

${\rm CH}\xhookrightarrow{} H^2 \xrightarrow{\,devhess\,} H(\symcurl; \mathbb S\cap\mathbb T) \xrightarrow{\,\symcurl\,} H(divdiv; \mathbb S\cap\mathbb T) \xrightarrow{\,divdiv\,} L^2 \to 0,$

and the conformal elasticity complex is

$${\rm CK}\xhookrightarrow{} H^1(\mathbb R^3) \xrightarrow{\,devdef\,} H(cott; \mathbb S\cap\mathbb T) \xrightarrow{\,cott\,} H(div; \mathbb S\cap\mathbb T) \xrightarrow{\,div\,} L^2(\mathbb R^3) \to 0.$$

Here $\mathbb S$ denotes symmetric matrices, $\mathbb T$ traceless matrices, and $\symcurl$0 the conformal tensors, meaning symmetric and trace-free tensors (Huang, 2 Aug 2025).

The kernels at the left end are explicitly identified. The space

$\symcurl$1

is the space of quadratic polynomials whose Hessian is pure trace, while

$\symcurl$2

is the space of conformal Killing fields in $\symcurl$3 (Huang, 2 Aug 2025).

The differential operators are obtained by taking trace-free or symmetric parts of classical operators. For example,

$\symcurl$4

The Cotton–York operator is given by

$\symcurl$5

and maps $\symcurl$6 into $\symcurl$7 (Huang, 2 Aug 2025). In the 2023 conformal-complex construction for symmetric traceless tensors, the same target space $\symcurl$8 supports discrete TT tensors and York splits, and the continuous conformal deformation complex is written as

$\symcurl$9

(Hu et al., 2023).

2. Local bubble conformal complexes

On a tetrahedron $divdiv$0, bubble conformal complexes are local exact sequences of bubble spaces with prescribed vanishing traces. They are not merely bubble spaces, but local exact BGG-constructed complexes on bubble spaces (Hu et al., 2023).

A smoothness vector $divdiv$1 encodes vertex, edge, and face smoothness, with

$divdiv$2

For a tetrahedron $divdiv$3,

$divdiv$4

Tensor-valued variants impose additional normal or tangential boundary conditions, producing spaces such as $divdiv$5, $divdiv$6, $divdiv$7, and $divdiv$8 (Huang, 2 Aug 2025).

Two local exact sequences are central.

Complex	Local exact sequence
Bubble conformal Hessian	$divdiv$9
Bubble conformal elasticity	$devdef$0

For the conformal Hessian complex, the smoothness constraints are

$devdef$1

with polynomial degree $devdef$2. For the conformal elasticity complex, the constraints are

$devdef$3

with degree $devdef$4 (Huang, 2 Aug 2025).

3. Local BGG construction, reduction, and face complexes

The decisive construction is local. Instead of running BGG on global conforming spaces, one starts from bubble de Rham, Hessian, elasticity, and divdiv complexes on a tetrahedron or a face, and applies the BGG machinery there. This yields simpler and more tractable constructions than global BGG-based approaches (Huang, 2 Aug 2025).

Two reduction mechanisms are used. The first is a tilde reduction that shrinks intermediate spaces so that the image of one operator is exactly the kernel of the next. The second is the reduction from symmetric tensors to symmetric trace-free tensors via the splitting

$devdef$5

At the bubble level this leads to the decomposition

$devdef$6

which isolates the conformal part from the scalar trace component (Huang, 2 Aug 2025).

Face complexes are equally important. The construction includes face bubble complexes and, on triangles, a bubble conformal divdiv complex,

$devdef$7

This is how the face traces needed for $devdef$8, $devdef$9, and $cott$0 are encoded locally (Huang, 2 Aug 2025).

A recurring misconception is that bubble conformal complexes are only interior enrichment spaces. The 2025 construction shows instead that they are exact local complexes that govern conformity, trace continuity, and the decomposition of global spaces into vertex, edge, face, and element contributions (Huang, 2 Aug 2025).

4. Global finite element complexes, exactness, and supersmoothness

The local bubble complexes are assembled into global conforming finite element complexes. For the conformal Hessian case,

$cott$1

and for the conformal elasticity case,

$cott$2

Exactness on topologically trivial domains is proved by combining local bubble exactness, dimension counts based on the polynomial conformal complexes, and Euler’s formula for the mesh (Huang, 2 Aug 2025).

The 2023 construction already produced a discrete conformal complex on tetrahedral meshes,

$cott$3

and showed exactness on contractible domains for $cott$4 (Hu et al., 2023). In that setting, supersmoothness is structural rather than cosmetic: $cott$5 uses $cott$6 at vertices, $cott$7 uses $cott$8, $cott$9 uses $L^2$0, and $L^2$1 uses $L^2$2. The key surjectivity statement is that

$L^2$3

is not surjective for $L^2$4, but is surjective for $L^2$5 (Hu et al., 2023). This explains why higher vertex smoothness propagates through the complex.

These exact complexes furnish discrete TT tensors, discrete York decompositions, and structure-preserving discretizations for relativity, Cosserat elasticity, and fluid mechanics (Hu et al., 2023).

5. Trace structures, bubble exactness, and cohomology

A broader unifying interpretation is provided by finite element complexes with trace structures. In this framework, one specifies for each simplex $L^2$6 and degree $L^2$7 a space $L^2$8 and trace operators $L^2$9 satisfying a weakened composition condition,

${\rm CH}\xhookrightarrow{} H^2 \xrightarrow{\,devhess\,} H(\symcurl; \mathbb S\cap\mathbb T) \xrightarrow{\,\symcurl\,} H(divdiv; \mathbb S\cap\mathbb T) \xrightarrow{\,divdiv\,} L^2 \to 0,$0

rather than a strict trace-composition law. This accommodates extra smoothness, jets at vertices, and nonstandard edge and face traces (Hu et al., 28 Sep 2025).

From the trace structure one defines local bubble spaces

${\rm CH}\xhookrightarrow{} H^2 \xrightarrow{\,devhess\,} H(\symcurl; \mathbb S\cap\mathbb T) \xrightarrow{\,\symcurl\,} H(divdiv; \mathbb S\cap\mathbb T) \xrightarrow{\,divdiv\,} L^2 \to 0,$1

Because the top-degree local cohomology is represented by generalized currents, one replaces ${\rm CH}\xhookrightarrow{} H^2 \xrightarrow{\,devhess\,} H(\symcurl; \mathbb S\cap\mathbb T) \xrightarrow{\,\symcurl\,} H(divdiv; \mathbb S\cap\mathbb T) \xrightarrow{\,divdiv\,} L^2 \to 0,$2 by a modified space

${\rm CH}\xhookrightarrow{} H^2 \xrightarrow{\,devhess\,} H(\symcurl; \mathbb S\cap\mathbb T) \xrightarrow{\,\symcurl\,} H(divdiv; \mathbb S\cap\mathbb T) \xrightarrow{\,divdiv\,} L^2 \to 0,$3

The central local condition is exactness of the modified bubble complex

${\rm CH}\xhookrightarrow{} H^2 \xrightarrow{\,devhess\,} H(\symcurl; \mathbb S\cap\mathbb T) \xrightarrow{\,\symcurl\,} H(divdiv; \mathbb S\cap\mathbb T) \xrightarrow{\,divdiv\,} L^2 \to 0,$4

for every simplex ${\rm CH}\xhookrightarrow{} H^2 \xrightarrow{\,devhess\,} H(\symcurl; \mathbb S\cap\mathbb T) \xrightarrow{\,\symcurl\,} H(divdiv; \mathbb S\cap\mathbb T) \xrightarrow{\,divdiv\,} L^2 \to 0,$5 (Hu et al., 28 Sep 2025).

If geometric decomposition and compatibility with generalized currents hold, then the global discrete complex has the correct cohomology: ${\rm CH}\xhookrightarrow{} H^2 \xrightarrow{\,devhess\,} H(\symcurl; \mathbb S\cap\mathbb T) \xrightarrow{\,\symcurl\,} H(divdiv; \mathbb S\cap\mathbb T) \xrightarrow{\,divdiv\,} L^2 \to 0,$6 The same framework yields local, ${\rm CH}\xhookrightarrow{} H^2 \xrightarrow{\,devhess\,} H(\symcurl; \mathbb S\cap\mathbb T) \xrightarrow{\,\symcurl\,} H(divdiv; \mathbb S\cap\mathbb T) \xrightarrow{\,divdiv\,} L^2 \to 0,$7-bounded, commuting projections

${\rm CH}\xhookrightarrow{} H^2 \xrightarrow{\,devhess\,} H(\symcurl; \mathbb S\cap\mathbb T) \xrightarrow{\,\symcurl\,} H(divdiv; \mathbb S\cap\mathbb T) \xrightarrow{\,divdiv\,} L^2 \to 0,$8

with the local bound

${\rm CH}\xhookrightarrow{} H^2 \xrightarrow{\,devhess\,} H(\symcurl; \mathbb S\cap\mathbb T) \xrightarrow{\,\symcurl\,} H(divdiv; \mathbb S\cap\mathbb T) \xrightarrow{\,divdiv\,} L^2 \to 0,$9

This suggests a general principle: exact bubble complexes in every dimension remove local cohomology, leaving the skeletal complex to carry the topological content (Hu et al., 28 Sep 2025).

6. Other context-dependent uses of the expression

The phrase has also been used interpretively in several other areas, and this suggests that it is not a single cross-disciplinary term with one invariant meaning.

In Hele–Shaw dynamics, doubly connected bubbles are encoded by conformal maps from an annulus or from the exterior of the unit disk with a cut, and the dynamics becomes a flow on a complex-analytic parameter space of conformal data. In that setting, “bubble conformal complexes” refers to the representation of bubble shapes as points in that conformal parameter space, with the selected Taylor–Saffman bubble of speed $${\rm CK}\xhookrightarrow{} H^1(\mathbb R^3) \xrightarrow{\,devdef\,} H(cott; \mathbb S\cap\mathbb T) \xrightarrow{\,cott\,} H(div; \mathbb S\cap\mathbb T) \xrightarrow{\,div\,} L^2(\mathbb R^3) \to 0.$$0 appearing as the unique non-singular attractor (Vasconcelos et al., 2013).

In the theory of branched complex projective structures, bubbling is a surgery that cuts along a bubbleable arc and inserts a copy of $${\rm CK}\xhookrightarrow{} H^1(\mathbb R^3) \xrightarrow{\,devdef\,} H(cott; \mathbb S\cap\mathbb T) \xrightarrow{\,cott\,} H(div; \mathbb S\cap\mathbb T) \xrightarrow{\,div\,} L^2(\mathbb R^3) \to 0.$$1, creating two simple branch points while preserving holonomy. For quasi-Fuchsian holonomy, a generic branched complex projective structure with two branch points is obtained by bubbling some unbranched structure, so one may view the resulting branched $${\rm CK}\xhookrightarrow{} H^1(\mathbb R^3) \xrightarrow{\,devdef\,} H(cott; \mathbb S\cap\mathbb T) \xrightarrow{\,cott\,} H(div; \mathbb S\cap\mathbb T) \xrightarrow{\,div\,} L^2(\mathbb R^3) \to 0.$$2-structures as bubbled conformal or projective chart complexes (Ruffoni, 2017).

In surface meshing, conformal parameterization is used to flatten a disk-topology surface, perform bubble packing in the plane, and pull back the Delaunay triangulation to the surface. The resulting object is a conformally controlled Delaunay simplicial complex generated by bubbles, and the improved method reports that computation time is cut by over 70% in surface triangulation examples while retaining high minimum-angle quality (Shangyu et al., 7 Aug 2025).

In $${\rm CK}\xhookrightarrow{} H^1(\mathbb R^3) \xrightarrow{\,devdef\,} H(cott; \mathbb S\cap\mathbb T) \xrightarrow{\,cott\,} H(div; \mathbb S\cap\mathbb T) \xrightarrow{\,div\,} L^2(\mathbb R^3) \to 0.$$3-geometry, sequences of associative Smith maps with bounded 3-energy may be conformally rescaled to yield finite bubble trees, and the limiting configuration is a tree-shaped complex of conformally parameterized associative maps. When the $${\rm CK}\xhookrightarrow{} H^1(\mathbb R^3) \xrightarrow{\,devdef\,} H(cott; \mathbb S\cap\mathbb T) \xrightarrow{\,cott\,} H(div; \mathbb S\cap\mathbb T) \xrightarrow{\,div\,} L^2(\mathbb R^3) \to 0.$$4-structure is closed, both the 3-energy and the homotopy are preserved in the bubble-tree limit (Cheng et al., 2019).

A further related viewpoint appears in the multi-bubble isoperimetric problem on $${\rm CK}\xhookrightarrow{} H^1(\mathbb R^3) \xrightarrow{\,devdef\,} H(cott; \mathbb S\cap\mathbb T) \xrightarrow{\,cott\,} H(div; \mathbb S\cap\mathbb T) \xrightarrow{\,div\,} L^2(\mathbb R^3) \to 0.$$5 and $${\rm CK}\xhookrightarrow{} H^1(\mathbb R^3) \xrightarrow{\,devdef\,} H(cott; \mathbb S\cap\mathbb T) \xrightarrow{\,cott\,} H(div; \mathbb S\cap\mathbb T) \xrightarrow{\,div\,} L^2(\mathbb R^3) \to 0.$$6, where minimizing clusters for $${\rm CK}\xhookrightarrow{} H^1(\mathbb R^3) \xrightarrow{\,devdef\,} H(cott; \mathbb S\cap\mathbb T) \xrightarrow{\,cott\,} H(div; \mathbb S\cap\mathbb T) \xrightarrow{\,div\,} L^2(\mathbb R^3) \to 0.$$7 have spherical interfaces and are spherical Voronoi clusters, with cells obtained as Voronoi cells of affine functions or as intersections of $${\rm CK}\xhookrightarrow{} H^1(\mathbb R^3) \xrightarrow{\,devdef\,} H(cott; \mathbb S\cap\mathbb T) \xrightarrow{\,cott\,} H(div; \mathbb S\cap\mathbb T) \xrightarrow{\,div\,} L^2(\mathbb R^3) \to 0.$$8 with convex polyhedra in $${\rm CK}\xhookrightarrow{} H^1(\mathbb R^3) \xrightarrow{\,devdef\,} H(cott; \mathbb S\cap\mathbb T) \xrightarrow{\,cott\,} H(div; \mathbb S\cap\mathbb T) \xrightarrow{\,div\,} L^2(\mathbb R^3) \to 0.$$9. Möbius geometry and conformal Killing fields are central there, and a conformal-complex interpretation of bubble configurations is natural, although the phrase is not introduced as a formal technical term in that work (Milman et al., 2022).