Modified de Donder Divergence

• Modified de Donder divergence is a gauge-fixing functional enhanced by trace, curvature, radial, or higher-derivative corrections applied in gravity and higher-spin theories.
• It provides essential modifications in diverse frameworks, including de Sitter propagators, AdS higher-spin dynamics, and quadratic gravity quantization.
• The term also appears in Hamilton–Jacobi and higher-order variational formalisms, highlighting a need for systematic disambiguation in the literature.

Searching arXiv for papers on modified de Donder gauge/divergence and related De Donder–Weyl usages. Search query: "modified de Donder gauge AdS higher spin (0808.3945)" Modified de Donder divergence is not a single universally standardized object across the literature. In one line of work it denotes a genuine modification of the harmonic or de Donder gauge functional, typically built from the densitized inverse metric divergence $F^\nu\equiv \partial_\mu(\sqrt{-g}\,g^{\mu\nu})$. In another line it refers more loosely to curvature-, radial-, or projector-corrected operators that play the role of de Donder constraints in de Sitter or AdS field theory. A separate and recurrent source of ambiguity is terminological: many papers with “De Donder” in the title concern the De Donder–Weyl covariant Hamiltonian formalism or higher-order De Donder forms rather than harmonic gauge. The topic therefore requires systematic disambiguation before any technical comparison is possible (Oda, 14 Jul 2025, 0808.3945, Riahi et al., 2019).

1. Terminological scope and competing meanings

Across the cited literature, “de Donder” appears in at least three technically distinct senses. In gauge-fixed gravity and higher-spin theory it refers to harmonic-type divergence conditions. In covariant Hamiltonian field theory it refers to the De Donder–Weyl formalism. In higher-order variational geometry it refers to a De Donder or Poincaré–Cartan-type form on jet bundles. This distinction is essential because papers in the latter two categories often contain total divergences, boundary terms, or polymomentum divergences, yet do not discuss a harmonic/de Donder gauge condition at all (Riahi et al., 2019, Śniatycki et al., 2019, Vey, 2014).

Context	Representative expression	Representative source
Harmonic/de Donder gauge	$\partial_\mu \tilde g^{\mu\nu}=0$ or $D^\mu h_{\mu\nu}-\frac12 D_\nu h=0$	(Oda, 2022, Miao et al., 2011)
Modified AdS higher-spin de Donder operator	$\bar C|\phi\rangle=0$, $C_{\rm con-sh}|\psi\rangle=0$	(0808.3945, Metsaev, 2013)
Higher-derivative de Donder gauge	$A^\mu{}_\nu F^\nu=0$ with $F^\nu=\partial_\mu\tilde g^{\mu\nu}$	(Oda, 14 Jul 2025)
De Donder–Weyl / higher-order De Donder form	$\partial_\mu S^\mu + H(\cdots)=0$, $\Theta=\pi_k^{2k-1*}A+\Xi$	(Riahi et al., 2019, Sniatycki et al., 2018)

This suggests that the phrase “modified de Donder divergence” is best treated as a family resemblance term rather than a uniquely fixed definition. In gauge-theoretic usage, the common core is a constraint built from the divergence of a field or field density, together with trace, curvature, radial, or higher-derivative corrections.

2. Ordinary de Donder gauge and field-weighted alternatives

A clean nonlinear form of the ordinary de Donder gauge is

$$\partial_\mu \tilde g^{\mu\nu}=0, \qquad \tilde g^{\mu\nu}\equiv \sqrt{-g}\,g^{\mu\nu}.$$

This is the exact gauge condition used in manifestly covariant quantization of scale-invariant scalar–tensor gravity equivalent to Brans–Dicke gravity after field redefinition, and the gauge-fixing/ghost sector is built from

$\partial_\mu \tilde g^{\mu\nu}=0$0

In that construction, the scalar field enters the classical Lagrangian, the field equations, and the ETCRs, but not the gauge-fixing function itself; the paper explicitly does not introduce a scalar-corrected or generalized de Donder divergence in the main formalism (Oda, 2022).

The same work is nevertheless important because it states explicit field-weighted alternatives that would be needed for local Weyl invariance: $\partial_\mu \tilde g^{\mu\nu}=0$1 These conditions are mentioned only as different gauge conditions, not as the gauge actually used. The paper also reiterates in its conclusions related $\partial_\mu \tilde g^{\mu\nu}=0$2-weighted expressions such as

$\partial_\mu \tilde g^{\mu\nu}=0$3

again as preferred alternatives for Weyl-invariant settings rather than developed ingredients of the quantization scheme (Oda, 2022).

A common misconception is therefore that coupling gravity to a dilaton or Brans–Dicke scalar automatically induces a modified de Donder divergence. In the cited scalar–tensor quantization, the technically correct statement is the opposite: the ordinary de Donder condition is retained unchanged, while the scalar modifies dynamics and operator structure elsewhere.

3. Curvature-modified de Donder structures on de Sitter space

On de Sitter background, the exact de Donder gauge for the graviton perturbation is

$\partial_\mu \tilde g^{\mu\nu}=0$4

At propagator level this becomes

$\partial_\mu \tilde g^{\mu\nu}=0$5

and similarly in the second index group. Equivalently,

$\partial_\mu \tilde g^{\mu\nu}=0$6

The divergence is therefore not set to zero independently; it is tied to the derivative of the trace. This is the precise curved-space de Donder divergence relation used in the de Sitter propagator construction (Miao et al., 2011).

What is modified on de Sitter is not the formal gauge condition itself but its implementation in the reduced kinetic operator and propagator equation. After imposing de Donder gauge, the Lichnerowicz operator yields

$\partial_\mu \tilde g^{\mu\nu}=0$7

The graviton propagator equation then acquires curvature-shifted tensor and trace operators and additional derivative source terms built from a vector propagator. The paper emphasizes that the gauge condition is de Sitter invariant, but no fully de Sitter-invariant propagator exists; the symmetry breaking arises from the scalar structure functions rather than from any alteration of the de Donder condition itself (Miao et al., 2011).

A closely related projector formulation appears in the Weyl–Weyl correlator analysis. There the propagator is decomposed into spin-0 and spin-2 parts using de Sitter-covariant differential projectors. The spin-0 projector is

$\partial_\mu \tilde g^{\mu\nu}=0$8

while the spin-2 projector $\partial_\mu \tilde g^{\mu\nu}=0$9 is explicitly transverse and traceless,

$D^\mu h_{\mu\nu}-\frac12 D_\nu h=0$0

This realizes the de Donder constraint through background-dependent projectors containing shifted operators such as $D^\mu h_{\mu\nu}-\frac12 D_\nu h=0$1, $D^\mu h_{\mu\nu}-\frac12 D_\nu h=0$2, and $D^\mu h_{\mu\nu}-\frac12 D_\nu h=0$3 (Mora et al., 2012).

4. Modified de Donder operators in AdS higher-spin theory

For massless totally symmetric bosonic fields in $D^\mu h_{\mu\nu}-\frac12 D_\nu h=0$4, the modified de Donder gauge is introduced explicitly in a CFT-adapted $D^\mu h_{\mu\nu}-\frac12 D_\nu h=0$5 oscillator formalism. The gauge-fixing term is

$D^\mu h_{\mu\nu}-\frac12 D_\nu h=0$6

with

$D^\mu h_{\mu\nu}-\frac12 D_\nu h=0$7

$D^\mu h_{\mu\nu}-\frac12 D_\nu h=0$8

and modified de Donder gauge condition

$D^\mu h_{\mu\nu}-\frac12 D_\nu h=0$9

The ladder operators $\bar C|\phi\rangle=0$0 package the radial derivative, radial coordinate, and curvature contributions. Under this gauge the equations reduce to

$\bar C|\phi\rangle=0$1

or in components

$\bar C|\phi\rangle=0$2

which are decoupled and solvable in terms of Bessel functions. By contrast, the standard $\bar C|\phi\rangle=0$3-covariant de Donder operators are

$\bar C|\phi\rangle=0$4

whereas the modified ones are shifted by radial-oscillator terms,

$\bar C|\phi\rangle=0$5

The modification is therefore explicit: the ordinary divergence-minus-trace structure is corrected by AdS radial terms so as to simplify the gauge-fixed dynamics (0808.3945).

The fermionic higher-spin case has a parallel but distinctly spinorial structure. In the CFT-adapted formalism for massless fermionic fields, the gauge-invariant second-order equation is

$\bar C|\phi\rangle=0$6

with

$\bar C|\phi\rangle=0$7

and

$\bar C|\phi\rangle=0$8

$\bar C|\phi\rangle=0$9

The modified de Donder gauge condition is

$C_{\rm con-sh}|\psi\rangle=0$0

and it reduces the equations to

$C_{\rm con-sh}|\psi\rangle=0$1

Because $C_{\rm con-sh}|\psi\rangle=0$2-tracelessness is respected only on shell,

$C_{\rm con-sh}|\psi\rangle=0$3

the paper also introduces an off-shell projected operator

$C_{\rm con-sh}|\psi\rangle=0$4

In the AdS/CFT analysis, the same operator induces the boundary differential constraints $C_{\rm con-sh}|\psi\rangle=0$5 and $C_{\rm con-sh}|\psi\rangle=0$6 for currents and shadow fields (Metsaev, 2013).

5. Higher-derivative de Donder gauge in quadratic gravity

A more literal higher-derivative modification appears in manifestly covariant canonical operator quantization of quadratic gravity. The ordinary gauge functional is

$C_{\rm con-sh}|\psi\rangle=0$7

and the higher-derivative gauge is introduced through

$C_{\rm con-sh}|\psi\rangle=0$8

with

$C_{\rm con-sh}|\psi\rangle=0$9

The paper then states explicitly that

$A^\mu{}_\nu F^\nu=0$0

will be called the higher-derivative de Donder gauge. Using metric compatibility, the gauge functional is reduced to

$A^\mu{}_\nu F^\nu=0$1

or, equivalently,

$A^\mu{}_\nu F^\nu=0$2

Here the de Donder divergence survives as the base object $A^\mu{}_\nu F^\nu=0$3, but it is acted on by a curvature-dependent higher-derivative operator (Oda, 14 Jul 2025).

The same paper shows that the equal-time commutators among all time derivatives of the metric vanish both in the conventional and higher-derivative de Donder gauges: $A^\mu{}_\nu F^\nu=0$4 and derives the spacelike commutator

$A^\mu{}_\nu F^\nu=0$5

for the metric tensor. The argument proceeds by imposing consistency of the higher-derivative gauge condition on commutator ansätze and showing that the would-be nonzero coefficients vanish. In the paper’s interpretation, this makes the metric behave “as if it were not a quantum operator but a classical field” and renders micro-causality manifest in the metric sector (Oda, 14 Jul 2025).

6. De Donder–Weyl and higher-order De Donder constructions

A major source of confusion is that several mathematically relevant “De Donder” constructions have nothing to do with harmonic gauge. In the Hamilton–Jacobi analysis of general relativity, “De Donder” refers to the De Donder–Weyl covariant Hamiltonian formalism. The general DW Hamilton–Jacobi equation is

$A^\mu{}_\nu F^\nu=0$6

with embedding condition

$A^\mu{}_\nu F^\nu=0$7

For general relativity the paper works in Gaussian coordinates,

$A^\mu{}_\nu F^\nu=0$8

not in harmonic coordinates, and the only divergence-like term central to the derivation is a total spatial divergence of the DW eikonal current $A^\mu{}_\nu F^\nu=0$9 whose integral is discarded under boundary assumptions. The paper explicitly does not introduce or discuss a modified de Donder gauge condition (Riahi et al., 2019).

In second-order gravity, the De Donder form becomes a higher-order Poincaré–Cartan-type object. For a second-order Lagrangian $F^\nu=\partial_\mu\tilde g^{\mu\nu}$0, the paper defines Ostrogradski momenta

$F^\nu=\partial_\mu\tilde g^{\mu\nu}$1

so the lower momentum is modified by a total divergence of the higher momentum. The associated De Donder form is a global $F^\nu=\partial_\mu\tilde g^{\mu\nu}$2-form on $F^\nu=\partial_\mu\tilde g^{\mu\nu}$3 for diffeomorphism-invariant second-order gravity Lagrangians (Śniatycki et al., 2019). Closely related higher-jet work constructs a boundary form $F^\nu=\partial_\mu\tilde g^{\mu\nu}$4 and the higher-order De Donder form

$F^\nu=\partial_\mu\tilde g^{\mu\nu}$5

with

$F^\nu=\partial_\mu\tilde g^{\mu\nu}$6

Here $F^\nu=\partial_\mu\tilde g^{\mu\nu}$7 is the exact correction that geometrizes the repeated integration-by-parts divergence terms, while the Euler–Lagrange equations are recovered from

$F^\nu=\partial_\mu\tilde g^{\mu\nu}$8

for vertical vector fields $F^\nu=\partial_\mu\tilde g^{\mu\nu}$9 (Sniatycki et al., 2018).

Related DW and multisymplectic papers reinforce the same distinction. In covariant analytic mechanics, the standard first DW equation is corrected for differential-form fields to

$\partial_\mu S^\mu + H(\cdots)=0$0

while the divergence-type second equation

$\partial_\mu S^\mu + H(\cdots)=0$1

is retained (Nakajima, 2016). In Palatini/vielbein gravity, the DW equations likewise involve spacetime divergences of polymomenta, but not a de Donder gauge condition (Vey, 2014). This suggests that search results for “modified de Donder divergence” often mix gauge-fixing literature with covariant Hamiltonian and higher-order variational literature solely because of the shared historical name.

7. Conceptual synthesis

The literature supports three main technical patterns. First, the ordinary harmonic functional $\partial_\mu S^\mu + H(\cdots)=0$2 can be retained unchanged even in nontrivial scalar–tensor models, with field-dependent alternatives merely proposed rather than used (Oda, 2022). Second, the de Donder condition can be implemented through curvature- or projector-corrected operators without changing its underlying role; this is the case for de Sitter graviton propagators and de Sitter-covariant projector constructions (Miao et al., 2011, Mora et al., 2012). Third, the divergence itself can be genuinely modified by higher-derivative or radial operators, as in quadratic gravity and AdS higher-spin theory, where the modified operator is designed to achieve BRST compatibility, exact gauge fixing, or decoupled equations (Oda, 14 Jul 2025, 0808.3945, Metsaev, 2013).

A corresponding terminological caution follows. In the harmonic-gauge sense, a modified de Donder divergence is a gauge-fixing functional obtained from the standard divergence by adding trace, curvature, radial, or higher-derivative structure. In the De Donder–Weyl and higher-jet sense, however, “modification” refers instead to total-divergence corrections of polymomenta or boundary forms. Treating these as interchangeable obscures the mathematical content of both traditions.