Covariant Dirichlet Helmholtz Projector

• The projector is defined as the orthogonal projection onto ker(D⋅) using the Dirichlet inverse of the covariant Laplacian, ensuring gauge-invariant decomposition.
• It plays a central role in non-Abelian gauge theory by enforcing Gauss law through the projection of the transverse velocity field, leading to effective rotor inertia computation.
• Similar constructions are employed in acoustic geometry and boundary integral methods, though only the gauge-theoretic version strictly uses Dirichlet boundary conditions.

A covariant Dirichlet Helmholtz projector denotes, in the most literal sense, a projection operator that extracts the covariantly divergence-free or exact component of a field by using a covariant differential operator together with the Dirichlet inverse of an associated Laplacian. In the literature considered here, that phrase is stated explicitly in non-Abelian gauge theory, where Gauss law is enforced by a projector built from the Dirichlet inverse of the covariant scalar Laplacian on a punctured three-ball (Ali, 6 Mar 2026). Closely related constructions appear in acoustic geometry, curved-surface Helmholtz–Hodge decomposition, boundary integral formulations of Helmholtz problems, infinite-lattice Calderón projections, and quasi-Helmholtz splittings for electromagnetic integral equations, but several of those works either use natural Neumann boundary conditions, trace-space Calderón projectors, or discrete loop/star projectors rather than an explicit Dirichlet covariant Helmholtz projector (Park et al., 5 Feb 2026, Betcke et al., 2020, Merlini et al., 2022).

1. Terminological scope and historical placement

The expression “covariant Dirichlet Helmholtz projector” is not uniformly established across the cited literature. Its most exact occurrence is in "Mass Without Mass from a Berry--Shifted SU(3) Holonomy Rotor" (Ali, 6 Mar 2026), where the object is a gauge-covariant projection onto $\ker D\!\cdot$ defined through the Dirichlet inverse of the covariant scalar Laplacian. In that setting, “covariant” refers to dependence on the non-Abelian covariant derivative $D_\mu=\partial_\mu+i g[A_\mu,\cdot]$, and “Dirichlet” refers to the scalar boundary condition $\phi|_{\partial D}=0$.

Other papers supply partial analogues rather than the same object. "Covariant Helmholtz-Hodge Decomposition: Resolving Spurious Vorticity via Acoustic Geometry" (Park et al., 5 Feb 2026) defines a covariant exact projector in an acoustic metric, but the numerical realization uses the natural Neumann condition

$(\nabla_\gamma\phi)^i n_i = u'^i n_i$

rather than $\phi|_{\partial\Omega}=0$. "Boundary element methods for Helmholtz problems with weakly imposed boundary conditions" (Betcke et al., 2020) studies the Helmholtz Calderón projector together with a Dirichlet-specific weak boundary operator, but on the Cauchy trace pair $$(u,\lambda)\in H^{1/2}(\Gamma)\times H^{-1/2}(\Gamma)$$, not as a covariant Hodge projector. "Laplacian Filters for Integral Equations: Further Developments and Fast Algorithms" (Merlini et al., 2022) develops basis-free projector-based Laplacian filtered quasi-Helmholtz decomposition, but the paper explicitly does not formulate a differential-geometric covariant Dirichlet Helmholtz projector.

This suggests that the phrase has at least three distinct usages in current research. In gauge theory it denotes a genuine covariant Dirichlet projection built from $D$ and $(D\!\cdot D)^{-1}$. In geometric acoustics it denotes a covariant Hodge projection whose Dirichlet specialization is mathematically natural but not the implemented closure. In computational electromagnetics and boundary integral analysis it often denotes the nearest discrete or trace-space analogue: Calderón projection, loop/star quasi-Helmholtz splitting, or basis-free filtered projectors.

2. Non-Abelian gauge-theoretic definition

In the SU(3) Yang–Mills setting of (Ali, 6 Mar 2026), the spatial domain is the punctured three-ball

$D:=B_R\setminus \mathrm{Tub}(\Gamma),$

with

$$\partial D=\partial B_R\cup \partial \mathrm{Tub}(\Gamma),$$

and with

$D_\mu=\partial_\mu+i g[A_\mu,\cdot]$0

Adjoint-valued one-forms satisfy relative boundary conditions

$D_\mu=\partial_\mu+i g[A_\mu,\cdot]$1

while adjoint-valued scalars satisfy Dirichlet conditions

$D_\mu=\partial_\mu+i g[A_\mu,\cdot]$2

The relevant operators are the covariant divergence

$D_\mu=\partial_\mu+i g[A_\mu,\cdot]$3

the covariant scalar Laplacian

$D_\mu=\partial_\mu+i g[A_\mu,\cdot]$4

and its Dirichlet inverse

$D_\mu=\partial_\mu+i g[A_\mu,\cdot]$5

The covariant Dirichlet Helmholtz projector is then defined by

$D_\mu=\partial_\mu+i g[A_\mu,\cdot]$6

with complementary longitudinal projector

$D_\mu=\partial_\mu+i g[A_\mu,\cdot]$7

Acting on a one-form $D_\mu=\partial_\mu+i g[A_\mu,\cdot]$8,

$D_\mu=\partial_\mu+i g[A_\mu,\cdot]$9

If $\phi|_{\partial D}=0$0 solves

$\phi|_{\partial D}=0$1

then $\phi|_{\partial D}=0$2 (Ali, 6 Mar 2026).

Its mathematical properties are those expected of a true orthogonal Helmholtz projector. The paper states that $\phi|_{\partial D}=0$3 is the orthogonal projector onto $\phi|_{\partial D}=0$4, with

$\phi|_{\partial D}=0$5

and bounded $\phi|_{\partial D}=0$6. Idempotence follows from the Dirichlet inverse relation, and the range satisfies

$\phi|_{\partial D}=0$7

The associated decomposition is

$\phi|_{\partial D}=0$8

unique modulo covariantly harmonic relative $\phi|_{\partial D}=0$9-forms. The paper also gives the distance formula

$(\nabla_\gamma\phi)^i n_i = u'^i n_i$0

so the projected field is the $(\nabla_\gamma\phi)^i n_i = u'^i n_i$1-closest covariantly divergence-free field to $(\nabla_\gamma\phi)^i n_i = u'^i n_i$2 (Ali, 6 Mar 2026).

The operator is central rather than decorative. Along a slow path $(\nabla_\gamma\phi)^i n_i = u'^i n_i$3, Gauss law

$(\nabla_\gamma\phi)^i n_i = u'^i n_i$4

with $(\nabla_\gamma\phi)^i n_i = u'^i n_i$5 reduces to

$(\nabla_\gamma\phi)^i n_i = u'^i n_i$6

hence

$(\nabla_\gamma\phi)^i n_i = u'^i n_i$7

The electric field is therefore precisely the projected transverse velocity, and the effective rotor inertia is

$(\nabla_\gamma\phi)^i n_i = u'^i n_i$8

This is the mechanism by which the paper connects a Dirichlet covariant Helmholtz projection to Gauss-law enforcement, gauge-representative independence, and the resulting nonzero spectral scale (Ali, 6 Mar 2026).

3. Metric-covariant Helmholtz–Hodge realizations

A second major line of development appears in acoustic geometry. In (Park et al., 5 Feb 2026), the spatial metric is

$(\nabla_\gamma\phi)^i n_i = u'^i n_i$9

Velocity fluctuations are converted to a metric-dual one-form

$\phi|_{\partial\Omega}=0$0

and the acoustic component is defined by exactness of that one-form,

$\phi|_{\partial\Omega}=0$1

In the appendix, the full covariant Hodge splitting is

$\phi|_{\partial\Omega}=0$2

The practical exact projector is not written as formal operator notation in the paper, but it is clearly implied by the Laplace–Beltrami solve

$\phi|_{\partial\Omega}=0$3

with reconstructed acoustic velocity

$\phi|_{\partial\Omega}=0$4

(Park et al., 5 Feb 2026).

The implemented boundary condition is explicitly Neumann: $\phi|_{\partial\Omega}=0$5 and for periodic domains the gauge is fixed by

$\phi|_{\partial\Omega}=0$6

The paper therefore does not literally define a covariant Dirichlet Helmholtz projector. A mathematically natural extension, explicitly identified in the supplied details as an inference rather than an author-stated formula, is the Dirichlet specialization

$\phi|_{\partial\Omega}=0$7

That extension is structurally consistent with the appendix’s variational identity

$\phi|_{\partial\Omega}=0$8

but it is not the boundary treatment used in the paper’s numerical demonstrations (Park et al., 5 Feb 2026).

The significance of the covariant metric choice is physical and numerical. Euclidean post-processing misclassifies thermal refraction and shock bending as vorticity, whereas the covariant splitting keeps the residual at the numerical noise floor, typically

$\phi|_{\partial\Omega}=0$9

including near the sonic horizon after regularization (Park et al., 5 Feb 2026). The conceptual lesson is that “covariant Helmholtz projector” may mean that the exact/coexact split is taken with respect to a non-Euclidean metric and the exact part is the exact part of the metric-dual one-form, not the Euclidean gradient of a vector field.

4. Surface and finite-element realizations

On curved surfaces, (Chun, 2020) develops a high-order moving-frame discretization of covariant differentiation and then adapts Helmholtz–Hodge decomposition to surfaces with “Neumann boundary or no boundary.” For a tangent field $$(u,\lambda)\in H^{1/2}(\Gamma)\times H^{-1/2}(\Gamma)$$0, the decomposition is written as

$$(u,\lambda)\in H^{1/2}(\Gamma)\times H^{-1/2}(\Gamma)$$1

and on a $$(u,\lambda)\in H^{1/2}(\Gamma)\times H^{-1/2}(\Gamma)$$2-surface this is effectively

$$(u,\lambda)\in H^{1/2}(\Gamma)\times H^{-1/2}(\Gamma)$$3

where

$$(u,\lambda)\in H^{1/2}(\Gamma)\times H^{-1/2}(\Gamma)$$4

The scalar potentials solve

$$(u,\lambda)\in H^{1/2}(\Gamma)\times H^{-1/2}(\Gamma)$$5

The paper does not define projector notation explicitly, but it induces the obvious decomposition operators $$(u,\lambda)\in H^{1/2}(\Gamma)\times H^{-1/2}(\Gamma)$$6, $$(u,\lambda)\in H^{1/2}(\Gamma)\times H^{-1/2}(\Gamma)$$7, and $$(u,\lambda)\in H^{1/2}(\Gamma)\times H^{-1/2}(\Gamma)$$8. A Dirichlet version,

$$(u,\lambda)\in H^{1/2}(\Gamma)\times H^{-1/2}(\Gamma)$$9

is presented in the supplied details as a careful inference, not as an explicit contribution of the paper (Chun, 2020).

A complementary surface finite-element framework is provided by (Nestler et al., 2018). There the tangential projector is

$D$0

and tangential tensor fields are represented by ambient Cartesian components projected by $D$1. The exact Cartesian covariant derivative for a tangent vector field is

$D$2

and the covariant Helmholtz-type operator is

$D$3

The paper does not define a Helmholtz projector, but it gives the weak forms, integration-by-parts identities, and piecewise linear Lagrange surface finite elements needed to realize one on manifolds with boundary (Nestler et al., 2018).

The abstract functional-analytic counterpart is (Chen et al., 2016), which derives generalized Helmholtz decompositions from exact complexes and commuting diagrams. The key template is

$D$4

with zero-trace spaces such as $D$5, $D$6, and $D$7 providing the Dirichlet-compatible potentials. The paper does not explicitly define a “Helmholtz projector,” but the decomposition canonically induces bounded projections onto the summands. This suggests a broad abstract interpretation: a covariant Dirichlet Helmholtz projector can be understood as the bounded projection induced by a commuting exact complex onto the subspace generated by a zero-trace potential space (Chen et al., 2016).

5. Boundary-trace projectors and Calderón analogues

In boundary integral formulations, the closest exact analogue is the Helmholtz Calderón projector. For the exterior Helmholtz Dirichlet problem, (Betcke et al., 2020) defines the exterior projector on Cauchy data

$D$8

satisfying

$D$9

Here the trace space is

$(D\!\cdot D)^{-1}$0

and Dirichlet data are weakly imposed through

$(D\!\cdot D)^{-1}$1

with right-hand side

$(D\!\cdot D)^{-1}$2

This is not a covariant Dirichlet Helmholtz projector in the Hodge sense, but it is an exact projector-based variational framework for Dirichlet Helmholtz data on the natural trace/flux Sobolev pair (Betcke et al., 2020).

A discrete lattice analogue is developed in (Xia, 15 Jun 2026). On the infinite lattice, the Helmholtz lattice Green’s function defines an infinite-lattice discrete Calderón projection on the boundary strip $(D\!\cdot D)^{-1}$3. For strip data $(D\!\cdot D)^{-1}$4, the projector is

$(D\!\cdot D)^{-1}$5

where

$(D\!\cdot D)^{-1}$6

The paper proves idempotence,

$(D\!\cdot D)^{-1}$7

and characterizes its range as the trace space of interior lattice-Helmholtz solutions. The LGF depends only on $(D\!\cdot D)^{-1}$8 and is reusable across geometries (Xia, 15 Jun 2026).

This construction is not classical Dirichlet/Neumann trace calculus. The closest Dirichlet-like data are the scalar strip values $(D\!\cdot D)^{-1}$9, especially on the interior and exterior sublayers $D:=B_R\setminus \mathrm{Tub}(\Gamma),$0 and $D:=B_R\setminus \mathrm{Tub}(\Gamma),$1. The capacity-matrix realization is

$D:=B_R\setminus \mathrm{Tub}(\Gamma),$2

This suggests a discrete “thickened Dirichlet trace” interpretation: the projector maps arbitrary strip data to the component consistent with an interior lattice Helmholtz field, while the complement is the outgoing exterior-consistent trace (Xia, 15 Jun 2026).

6. Quasi-Helmholtz electromagnetic analogues and conceptual boundaries

In computational electromagnetics, several projector families are close in function but not identical in meaning. The basis-free construction in (Merlini et al., 2022) acts on the RWG current space through filtered projectors onto loop and star sectors,

$D:=B_R\setminus \mathrm{Tub}(\Gamma),$3

with, in the standard setting,

$D:=B_R\setminus \mathrm{Tub}(\Gamma),$4

$D:=B_R\setminus \mathrm{Tub}(\Gamma),$5

For simply connected geometries,

$D:=B_R\setminus \mathrm{Tub}(\Gamma),$6

while on non-simply connected geometries a harmonic correction is required,

$D:=B_R\setminus \mathrm{Tub}(\Gamma),$7

The paper explicitly describes these as basis-free projector-based Laplacian filtered quasi-Helmholtz decompositions and explicitly does not formulate a Dirichlet boundary-value Helmholtz projector (Merlini et al., 2022).

Their role is preconditioning of the electric field integral equation. The paper states

$D:=B_R\setminus \mathrm{Tub}(\Gamma),$8

and introduces wavelet-like band projectors

$D:=B_R\setminus \mathrm{Tub}(\Gamma),$9

In the reported numerical results, condition numbers after low-frequency stabilization alone were $$\partial D=\partial B_R\cup \partial \mathrm{Tub}(\Gamma),$$0, $$\partial D=\partial B_R\cup \partial \mathrm{Tub}(\Gamma),$$1, and $$\partial D=\partial B_R\cup \partial \mathrm{Tub}(\Gamma),$$2 for $$\partial D=\partial B_R\cup \partial \mathrm{Tub}(\Gamma),$$3, $$\partial D=\partial B_R\cup \partial \mathrm{Tub}(\Gamma),$$4, and $$\partial D=\partial B_R\cup \partial \mathrm{Tub}(\Gamma),$$5 unknowns, and became $$\partial D=\partial B_R\cup \partial \mathrm{Tub}(\Gamma),$$6, $$\partial D=\partial B_R\cup \partial \mathrm{Tub}(\Gamma),$$7, and $$\partial D=\partial B_R\cup \partial \mathrm{Tub}(\Gamma),$$8 after adding the new Laplacian filters (Merlini et al., 2022). This is strong evidence for the effectiveness of projector-based spectral separation, but it remains a discrete quasi-Helmholtz technology rather than a Dirichlet covariant one.

A closely related PMCHWT stabilization is given in (Chhim et al., 2020), where the projectors are the standard loop/star quasi-Helmholtz projectors

$$\partial D=\partial B_R\cup \partial \mathrm{Tub}(\Gamma),$$9

used to stabilize three eddy-current-relevant regimes: (i) frequency decreases with constant conductivity, (ii) conductivity increases at fixed frequency, and (iii) frequency decreases while keeping the product of the frequency and the conductivity constant. These operators are discrete boundary Helmholtz-type projectors on tangential surface-current unknowns, but the paper does not use Dirichlet or covariant differential-form language (Chhim et al., 2020).

The main conceptual boundary is therefore sharp. A covariant Dirichlet Helmholtz projector, in the strict sense, is a projection defined through covariant differential operators and a Dirichlet inverse, as in (Ali, 6 Mar 2026). Metric-covariant exact projectors with Neumann realization, as in (Park et al., 5 Feb 2026), are close but not Dirichlet. Calderón projectors, as in (Betcke et al., 2020) and (Xia, 15 Jun 2026), are exact projectors on trace data rather than Hodge projections on fields. Quasi-Helmholtz projectors, as in (Merlini et al., 2022) and (Chhim et al., 2020), are discrete algebraic splittings of current spaces into loop/solenoidal and star/non-solenoidal sectors. A common misconception is to treat these objects as interchangeable. They are related by function—separation of admissible from non-admissible components—but not by definition, boundary treatment, or operator calculus.