Topological ModMax–AdS Black Holes: Extremality, Charge Normalization, and Limits of WGC Diagnostics

Introduction

The Weak Gravity Conjecture (WGC) is motivated in part by the expectation that charged black holes should be able to discharge, but a classical metric does not determine whether the required charged states exist in the underlying quantum theory. A meaningful comparison requires the charge appearing in the conjecture to be canonically normalized, the candidate state spectrum to be specified, and the decay or extremality criterion to match the asymptotic setting [1, 2]. These requirements become particularly important in nonlinear electrodynamics, where the conserved electric charge is a flux of the displacement tensor and need not coincide with the integration constant in the field strength.

ModMax electrodynamics is a conformal, electromagnetic-duality-invariant nonlinear theory parameterized by a dimensionless coupling $\gamma$ [3, 4, 5]. On a purely electric background, the constitutive relation reduces to a Maxwell-like equation with a $\gamma$-dependent coefficient. This makes the background solution tractable, but it also creates a normalization trap: the field-strength amplitude, the displacement flux, and the charge measured by a canonically rescaled potential must not be assigned the same symbol without a conversion. A charge-to-mass ratio can otherwise acquire a spurious dependence on $\gamma$.

The topological ModMax/ModAMax–AdS family was studied by Eslam Panah, Hamil, and Rodrigues [6]. While the background solution and its basic thermodynamics were established in [6], the present work focuses on the charge normalization, the sensitivity of the extremality curve to the charge convention, the classification of all extremal roots, and the limitations of WGC diagnostics in this setting. A previous study by Gashti and collaborators examined ModMax black holes using horizon, weak-cosmic-censorship, and photon-sphere diagnostics [7]. More recently, Gashti, Pourhassan, and Sakallı derived a probe-based WGC criterion for selected asymptotically flat black holes by solving a charged-scalar radial problem and relating its near-horizon conformal weight to a relaxation-time condition [8]. That result uses a particular background and a global mode-matching argument. Earlier work by some of the present authors considered WGC diagnostics for charged rotating AdS black holes in a different matter environment [9]; those results likewise do not replace the normalization and spectrum checks required here. The CFT result cannot be transferred to the present topological AdS family merely by copying a near-horizon algebraic factor: the AdS boundary conditions, horizon topology, and complete radial problem must be reconsidered.

A separate strong-field analysis of black-hole decay reports that ModMax fails a necessary charge-growth condition in its asymptotically flat setting [10]. This result is not, by itself, a theorem for every AdS ensemble; it is, however, a reason not to advertise a larger unnormalized charge-to-mass ratio as evidence for WGC compatibility. We therefore keep three questions separate throughout: (i) what follows exactly from the classical extremality curve, (ii) what can be tested locally for an explicitly added probe, and (iii) what remains unknown without a charged spectrum and a controlled ultraviolet completion. Over the past few years, the study of ModMax black holes has expanded in several directions. For instance, their thermodynamic behavior, photon spheres, and weak cosmic censorship have been examined in various setups [7]. Similar thermodynamic analyses, including non-perturbative corrections to entropy and phase transitions, have also been carried out for other black hole solutions in AdS and dirty black hole backgrounds [11], and in other gravitational frameworks such as Kalb–Ramond gravity [12]. Topological black holes in anti-de Sitter space, with spherical, planar, and hyperbolic horizons, also have a long history, starting from the classic work of Birmingham [13]. However, a systematic look at how the charge normalization affects both the extremality curve and the WGC diagnostics in this specific topological ModMax–AdS family has been missing so far. The present work focuses on those questions and does not attempt to infer a complete stability theorem from thermodynamic data. We reconstruct the charge dictionary from the standard action, derive the first law with one convention throughout, classify every real positive extremal root in the spherical, planar, and hyperbolic sectors, and evaluate the limiting cases that change the physical interpretation. As an independent geometric check, we derive the unstable null circular orbit for the spherical extremal branch without turning it into an observational shadow claim. We then compute the local $\mathrm{AdS}_2$ BF index for a minimally coupled scalar used only as a probe. Finally, we state explicitly which swampland conjectures are and are not testable with a fixed ModMax parameter and a negative cosmological constant. The principal conclusions are deliberately limited to the calculations actually performed.

Action, field equations, and charge normalization

We use $G=1$ and the standard ModMax Lagrangian

where

The action is

The sign $\eta=+1$ is the ordinary-sign ModMax branch, while $\eta=-1$ reverses the gauge-sector sign and is referred to as ModAMax. The boundary term depends on the chosen variational problem and does not change the local field equations.

The equations following from this action can be written as

where the metric-variation tensor is defined unambiguously by

An overall constant multiplying the constitutive equation is immaterial. For a purely electric configuration, $\mathcal{F}<0$ and $\mathcal{G}=0$, so

Thus $e^{+\gamma}$ multiplies the field-strength term. The $e^{-\gamma}$ in the lapse below appears only after rewriting the solution in terms of the conserved displacement charge.

For the static ansatz

let $q$ denote the field-strength integration constant and choose

The displacement field is proportional to $e^\gamma q$. For a spherical horizon we define the flux-normalized charge by

where the orientation is chosen so that $q\geq0$. Thus the integrated displacement flux is $4\pi q$ in this convention; $Q$ is the per-unit-horizon-volume charge density used below. For planar and hyperbolic horizons, all thermodynamic quantities below are likewise densities per unit horizon volume. The potential difference is therefore

The variables $A_t=-q/r$ and $Q=q/(4\pi)$ would be inconsistent with the lapse term $q^2e^{-\gamma}/r^2$ if used simultaneously; Eqs. (10)–(11) fix this ambiguity. On this purely electric background it is useful to introduce a Maxwell-normalized field and charge,

If a probe has covariant derivative $D_\mu=\nabla_\mu-ie_\Phi A_\mu$, its coupling in these variables is $\widehat{e}_\Phi=e^{-\gamma/2}e_\Phi$. This field redefinition concerns the purely electric background. The nonanalyticity of ModMax at $\mathcal{F}=\mathcal{G}=0$ means that it does not establish equivalence of the full theory, including generic dyonic perturbations, to Maxwell theory [3, 10].

Solution and thermodynamics

We focus on the AdS branch $\Lambda=-3/l^2<0$. With the normalization above, the exact pure-electric solution is [14, 15, 16, 17]

For $\eta=+1$ this is the ordinary-sign ModMax branch; for $\eta=-1$ it is the sign-reversed ModAMax branch. The scalar curvature is $R=4\Lambda=-12/l^2$, while the Kretschmann invariant is

For nonzero charge, the final term makes the curvature invariant diverge as $r\to0$. For a neutral solution with $m\ne0$, the mass term also diverges. The special case $m=q=0$ must not be described as a curvature singularity solely on the basis of Eq. (14). At large radius, $f(r)\sim r^2/l^2$ and the invariant approaches $24/l^4$, as required for asymptotically AdS geometry.

Solving $f(r_+)=0$ gives

The Hawking temperature is

Because the gravitational theory is Einstein gravity, the entropy obeys the area law. Per unit horizon volume, the thermodynamic variables are

For a spherical horizon these are densities divided by $4\pi$; the total quantities are $S_{\mathrm{tot}}=\pi r_+^2$, $M_{\mathrm{tot}}=m/2$, and $Q_{\mathrm{tot}}=q$. This distinction avoids mixing total ADM quantities with densities when comparing different horizon topologies.

In terms of $r_+$, the mass density is

Equivalently, using $r_+=2\sqrt{S}$ and $q=4\pi Q$,

At fixed $Q$, $\gamma$, and $l$, differentiation yields

Thus the non-extended first law is

The first law is an internal consistency check. It does not independently determine whether the gauge field has a unitary ultraviolet completion or whether the WGC holds. Figure 1 shows that the extremal radius depends

Exact extremality and topology dependence

General extremal equation

At a degenerate horizon, $f(r_e)=f'(r_e)=0$, equivalently $T(r_e)=0$.

Define

The temperature condition is the quadratic equation

with roots

Only real, positive roots can represent finite-radius degenerate horizons. The extremal mass simplifies after use of Eq. (24):

The sign of $M_{\mathrm{ext}}$ therefore depends on topology and the root, not only on the sign of $\eta$.

ModMax: $\eta=+1$

For spherical horizons,

It is positive for nonzero charge. The formal zero-charge limit has $x_e\to0$, not a finite-radius extremal Schwarzschild–AdS black hole. For planar horizons,

This is likewise positive for $Q\ne0$ and tends to zero when the charge vanishes.

For hyperbolic horizons the positive root is instead

The discriminant has a plus sign. The alternative root, proportional to $1-\sqrt{1+192\pi^2Q^2e^{-\gamma}/l^2}$, is negative for nonzero charge and is not a physical radius. Hence the hyperbolic ModMax branch has a real positive extremal radius for every $Q$, rather than the upper-charge restriction stated in the original draft.

The extremal mass for $k=-1$ is

Since $r_e^2/l^2 \ge1/3$, this mass is negative for $1/3 \le u < 1/2$, zero at $u=1/2$, and positive for $u > 1/2$. The zero-mass threshold is

Thus it is incorrect to classify all hyperbolic ModMax extremal solutions as negative-mass solutions. A positive-mass WGC comparison is available only above the threshold in Eq. (31); the zero-mass point makes a charge-to-mass ratio singular and is not a standard WGC state. Figure 2 shows that the spherical and planar extremal masses are non-negative for positive effective charge. The hyperbolic ModMax branch behaves differently: its extremal mass changes sign at $\mathcal{A}=1/4$. This point corresponds to $r_e^2/l^2=1/2$, where the mass vanishes. A charge-to-mass ratio is singular there because its denominator vanishes; this singularity is not, by itself, evidence for a superextremal state. Similar ModMax black hole solutions with different horizon topologies and thermodynamic properties have been studied in [14, 15].

ModAMax: $\eta=-1$

For $k=+1$ and $k=0$, Eq. (24) has no positive solution for $\mathcal{A}>0$. Thus these branches have no finite-radius extremal black hole. For the hyperbolic case, real roots exist only if

and are

For a nonzero charge below the discriminant bound, both algebraic roots are positive. They are not physically interchangeable. Since

the smaller root has $f^{\prime\prime}(r_e)<0$ and is a local maximum of $f$, not the outer degenerate horizon. The physical outer extremal branch is the larger root $x_+$, for which $f^{\prime\prime}>0$, except at the endpoint where the two roots coincide. At

we have $f^{\prime\prime}(r_e)=0$; the horizon is more degenerate and the usual AdS$_2$ throat approximation below is not valid.

On the physical outer ModAMax extremal branch, $l^2/6<r_e^2\le l^2/3$ for nonzero charge. Consequently,

The negativity follows from the root range. It does not follow from the expression $-r_e^3/(4\pi l^2)$, which is not the correct extremal mass formula. In addition, the wrong-sign gauge kinetic sector is a separate obstruction to interpreting ModAMax as a healthy, unitary charged sector without a specified completion. The Figure 3 shows the two hyperbolic ModAMax roots merge

at the discriminant endpoint $a=1/12$, where $u=1/6$. For nonzero charge below this endpoint, both algebraic roots are positive, but they do not have the same geometric status. The smaller root satisfies $f^{\prime\prime}(r_e)<0$ and is not the outer degenerate horizon. At the merger, $f^{\prime\prime}(r_e)=0$, so the standard quadratic near-horizon expansion leading to an AdS$_2$ throat cannot be used.

Extremal charge-to-mass diagnostics

For any admissible extremal root with $M_{\rm ext}\ne0$, Eqs. (23) and (26) give the displacement-charge ratio

This formula is useful algebraically but should not be compared with a canonical WGC threshold before the charge convention is fixed. Using the canonical charge $\widehat{Q}=e^{-\gamma/2}Q$ gives

The explicit factor $e^{\gamma}$ has disappeared. If the canonical charge is held fixed, $u$ is also independent of $\gamma$ on the purely electric solution. The parameter dependence of Eq. (37) at fixed displacement charge therefore cannot be read as a physical enhancement of the WGC without making the same field redefinition for the particle sector.

In the asymptotically flat spherical limit, $l\to\infty$, the ModMax extremal solution obeys

The final equality is the normalization-invariant Reissner–Nordström value for this pure-electric sector. The preceding ratio differs because it uses displacement charge rather than the canonically normalized charge. This does not make the displacement charge incorrect; it means that both sides of a conjectural inequality must use the same convention.

For the planar branch, the displacement ratio becomes

whereas the canonical ratio is $e^{-\gamma}\mathcal{R}_{\mathrm{disp}}$. Its dependence on the charge and AdS scale is another reason that a universal threshold equal to one cannot be imposed on every topological AdS density. In the hyperbolic ModMax branch, Eq. (37) can diverge at the zero-mass point, but that divergence is a zero of the denominator, not evidence for a superextremal state. Figure 4 shows that, at fixed displacement charge, the ratio $Q/M_{\mathrm{ext}}$ changes with $\gamma$ because the metric depends on the combination $Q^{2}e^{-\gamma}$. The ratio formed with the rescaled charge, $\widehat{Q}=e^{-\gamma/2}Q$, follows a different curve. These results must be interpreted along the specified parameter path: holding $Q$ fixed does not hold $\widehat{Q}$ fixed. At fixed canonical charge, the purely electric background and its extremality curve are independent of $\gamma$.

Numerical benchmark and limiting cases

For a numerical benchmark, set $l=1$, $Q=0.1$, and $\eta=+1$. The exact root formulas above give the values in Table 1. The charge-to-mass entries use both the displacement convention and the Maxwell-normalized charge $\widehat{Q}=e^{-\gamma/2}Q$. The latter is the appropriate variable for comparison with a canonically normalized Maxwell sector. Neither column is a WGC pass/fail test without a specified particle spectrum and decay criterion. At fixed displacement charge, increasing $\gamma$ decreases $\mathcal{A}=16\pi^2Q^2e^{-\gamma}$ and shrinks the extremal radius. The raw ratio can consequently grow. At fixed canonical charge $\widehat{Q}$, the lapse is independent of $\gamma$ in the pure-electric sector, so this change is a normalization-path effect rather than an independent weakening of gravity.

$\gamma$	$k$	$x_e=r_e^2$	$r_e$	$M_{\mathrm{ext}}$	$Q/M_{\mathrm{ext}}$	$\widehat{Q}/M_{\mathrm{ext}}$
0.0	+1	0.577750	0.760099	0.130379	0.766993	0.766993
0.0	0	0.725520	0.851774	0.098354	1.016731	1.016731
0.0	−1	0.911084	0.954507	0.062450	1.601292	1.601292
0.5	+1	0.422437	0.649951	0.095420	1.048003	0.816186
0.5	0	0.565035	0.751688	0.067598	1.479335	1.152107
0.5	−1	0.755770	0.869350	0.035389	2.825761	2.200705
1.0	+1	0.303888	0.551260	0.070530	1.417841	0.859964
1.0	0	0.440050	0.663363	0.046459	2.152419	1.305508
1.0	−1	0.637221	0.798261	0.017434	5.736049	3.479090

The limiting values are equally important because formal charge-to-mass ratios can become singular at zero mass or where the extremal radius shrinks to zero. Table 2 records the directly evaluated cases. Values of $M/l$ use the same density normalization as above.

Branch / limit	$a$	$u$	$r_e/l$	$M_{\mathrm{ext}}/l$	Interpretation
ModMax, $k=+1$, $Q \to0$	0	0	0	0	No finite-radius neutral extremal horizon.
ModMax, $k=0$, $Q \to0$	0	0	0	0	The charged extremal radius shrinks to zero.
ModMax, $k=-1$, $Q \to0$	0	$1/3$	0.577350	$-0.015309$	Neutral hyperbolic extremal limit.
ModMax, $k=-1$, $M_{\mathrm{ext}}=0$	$1/4$	$1/2$	0.707107	0	$Q^2 e^{-\gamma}=l^2/(64\pi^2)$; $Q/M$ is singular.
ModAMax, $k=-1$, triple root	$1/12$	$1/6$	0.408248	$-0.021651$	$f^{\prime\prime}(r_e)=0$; the usual quadratic $\mathrm{AdS}_2$ throat is absent.

For another explicit check, take ModAMax with $l=1$, $Q=0.01$, and $\gamma=0.5$. Then $a=\mathcal{A}/l^2\simeq0.00957795<1/12$, and the roots are $u_-=0.00987021$ and $u_+=0.32346312$. The smaller root has $f''<0$ and is not an outer extremal horizon. The larger root gives $r_e\simeq0.568738$, $M_{\mathrm{ext}}\simeq-0.0159797$, and $L_2^2\simeq0.343825$. These values follow from the exact root and mass formulas; they are not fits to observational data.

Thermodynamic character of the zero-mass point on the hyperbolic ModMax branch

The zero-mass point identified in Eq. (31) deserves a separate thermodynamic reading. The charge-to-mass diagnostic of Eq. (37) diverges there, while the horizon radius and entropy density remain finite and nonzero. We show that this divergence is a pole of the ratio, not a signal of a superextremal state, and that the zero-mass point is a smooth zero crossing of the extremal mass along the connected extremal branch.

It is convenient to write the extremal mass (30) and the entropy density (17) in terms of $u=r_e^2/l^2$. For $r_e=l\sqrt{u}$, they take the compact forms

The zero of the mass occurs at $u=1/2$, i.e. $r_e^2=l^2/2$, in agreement with Eq. (31). At that point

so the vanishing of the extremal mass does not coincide with a vanishing horizon radius or a vanishing entropy density. The Bekenstein–Hawking area law gives a finite, nonzero entropy density.

The derivative of the extremal mass with respect to $u$,

is finite and positive at $u=1/2$:

Thus $M_{\mathrm{ext}}$ crosses zero with a nonvanishing, finite slope. The zero is a simple root of the extremal mass function, not a degenerate point of the mass function itself. Equivalently, starting from the general mass function $M(S,Q)$, the response to the displacement charge is

which at the extremal point $r_{+}=r_{\mathrm{e}}$ and $u=1/2$ evaluates to

This value is finite and nonzero at the zero-mass point, whereas the ratio $Q/M_{\mathrm{ext}}$ diverges. The divergence of the charge-to-mass diagnostic at $u=1/2$ is therefore entirely due to the vanishing of its denominator $M_{\mathrm{ext}}$. The numerator and the thermodynamic response of the mass remain regular.

Figure 5 summarizes these results visually: Figure 5 (a) shows the finite-slope zero crossing of the extremal mass, Figure 5 (b) shows the continuous and monotonic entropy density, and Figure 5 (c) shows the pole in the ratio $Q/M_{\mathrm{ext}}$. Table 3 records the corresponding numerical behaviour for $l=1$ and $\gamma=0.5$. The entropy density varies smoothly and monotonically across $u=1/2$, the extremal mass changes sign, and only the ratio $Q/M_{\mathrm{ext}}$ exhibits a pole. The negative- and positive-mass segments belong to the same connected extremal family, and the zero-mass point has a finite horizon radius and entropy density.

$u$	$r_e$	$M_{\mathrm{ext}}$	$S_{\mathrm{ext}}$	$Q(u)$	$Q/M_{\mathrm{ext}}$
0.40	0.6325	−0.0101	0.1000	0.0289	−2.87
0.45	0.6708	−0.0053	0.1125	0.0406	−7.60
0.49	0.7000	−0.0011	0.1225	0.0490	−44.0
0.50	0.7071	0.0000	0.1250	0.0511	pole
0.51	0.7141	+0.0011	0.1275	0.0531	+46.7
0.55	0.7416	+0.0059	0.1375	0.0611	+10.4
0.60	0.7746	+0.0123	0.1500	0.0708	+5.74

Two points are worth emphasizing. First, the zero-mass point is not a WGC state: a charge-to-mass comparison requires a nonvanishing mass, and the divergence of $Q/M_{\mathrm{ext}}$ at $u=1/2$ comes from the denominator, not from a superextremal ratio. This is in line with the warning in § 8 that a raw charge-to-mass ratio is not a WGC test [1, 2, 18]. Second, the zero-mass point is a smooth zero crossing of the extremal mass along the connected extremal branch, with finite horizon radius and entropy density. The entropy is continuous and monotonic, the temperature vanishes along the branch, and the mass crosses zero with finite slope. A similar picture emerges in other ModMax black hole backgrounds, where thermodynamic transitions and singular ratios have been studied in [19, 20]. We stress that these results hold for the fixed-coupling, purely electric background considered throughout. They do not address the spectrum, the ultraviolet completion, or the global perturbation problem [21], and they do not change the conclusion of § 8.

Spherical null circular orbit as a geometric check

For the spherical sector ($k = +1$), the null-geodesic effective potential per unit squared angular momentum is $V_{\rm null}(r) = f(r)/r^2$. A circular null orbit satisfies $V'_{\rm null}(r_{\rm ph}) = 0$, or

For the ordinary-sign ModMax lapse this reduces exactly to

provided the discriminant is non-negative. The cosmological term cancels in Eq. (47), although the AdS scale remains in $m$ and $\mathcal{A}$ along an extremal family. The outer circular orbit corresponds to the plus root. At a circular orbit, the second derivative of $V_{\rm null}$ is

for the outer root with nonzero discriminant, so that orbit is unstable in the radial direction.

On the spherical extremal ModMax branch, $m=2r_e(1+2u)$ and $\mathcal{A}=r_e^{2}(1+3u)$, where $u=r_e^{2}/l^{2}$. Substitution into Eq. (48) gives the simple exact relation

Thus the outer null circular orbit lies outside the extremal horizon. Null geodesics and photon spheres in ModMax black holes have also been examined in [22]. For the numerical values in Table 1, the resulting radii are shown in Table 4. These radii are geometric diagnostics; an observational shadow prediction would additionally require an observer prescription and a physical environment, and is not claimed here. At fixed displacement charge the radius decreases as $\gamma$ increases in this benchmark because the effective geometric charge $\mathcal{A}$ decreases. At fixed canonical charge, $\mathcal{A}=(4\pi\hat{Q})^{2}$ is unchanged, and so is this purely electric geometry. This orbit calculation is independent of the WGC verdict and cannot replace a charged-spectrum or decay analysis.

$\gamma$	$r_e$	$r_{\rm ph}$	$r_{\rm ph}/r_e$
0.0	0.760099	4.155082	5.466502
0.5	0.649951	2.947282	4.534620
1.0	0.551260	2.107650	3.823328

Near-horizon geometry and a charged-probe test

A double root with $f^{\prime\prime}(r_e)>0$ has a local near-horizon geometry $\mathrm{AdS}_2 \times\Sigma_k$. Expanding the lapse gives

with

This formula applies only when $k+6r_e^2/l^2>0$. It applies to the ordinary ModMax extremal branches with a positive finite radius and to the physical outer ModAMax hyperbolic root away from the triple-root endpoint. At that endpoint $f^{\prime\prime}=0$ and this quadratic $\mathrm{AdS}_2$ approximation fails.

The electric field at the horizon, in the gauge convention of Eq. (11), is

For the ModMax branch, extremality gives the useful identity

where the ratio is evaluated on the regular branch for which $L_2^2>0$.

To test a specific local stability condition, add a minimally coupled complex scalar as a probe; it is not part of the background action:

Let $\lambda_k$ be an eigenvalue of the negative Laplacian on the unit horizon section. On a sphere $\lambda_k=\ell(\ell+1)$; on a planar section it is the squared wave number, and for a compact hyperbolic quotient it belongs to the scalar spectrum of that quotient. The near-horizon index is

The local charged-scalar Breitenlohner–Freedman condition is

A violation indicates a local BF instability of this probe in the near-horizon throat. Satisfaction of Eq. (57) is not a proof of full stability: it does not solve the global radial problem with AdS boundary conditions, coupled metric–gauge perturbations, or possible superradiant modes [21]. The choice $\lambda_k = 0$ below is a local constant-mode benchmark; global normalizability on noncompact planar sections, and the spectrum on compact horizon quotients, must be imposed separately.

As a reproducible numerical check, set $l = 1$, $Q = 0.1$, $\gamma= 0.5$, $\lambda_k = 0$, $\mu_\Phi= 0.2$, and $e_\Phi= 1$. Table 5 gives the horizon data and BF index calculated from Eqs. (52)–(56). The critical probe coupling is defined by $\nu_0^2 = 0$:

$k$	$L_2^2$	$\mathcal{E}_h$	$e_{\Phi,\mathrm{BF}}$	$\nu_0^2$
+1	0.119514	1.804268	2.340794	0.208282
0	0.166667	1.348923	2.253455	0.206122
−1	0.213819	1.008493	2.358059	0.212054

These numbers are an internal consistency check for a deliberately chosen probe, not a prediction for a known particle. They do not supply a WGC bound because $e_\Phi$ and $\mu_\Phi$ are inputs rather than data derived from the spectrum of the theory.

Parameter sensitivity and the meaning of fixed charge

The parameter $\gamma$ appears in the metric only through $Q^2e^{-\gamma}$ when the displacement-charge convention of Eq. (10) is used. Rewriting Eq. (19) in terms of the Maxwell-normalized charge $\widehat{Q}=e^{-\gamma/2}Q$ gives

The canonical mass function has no explicit $\gamma$-dependence in this purely electric sector. At fixed $S$ and $Q$, however, the chain rule yields

whereas

The same statement holds for a smooth extremal branch because $\partial M/\partial S=T=0$ at extremality. Thus the derivative at fixed $Q$ measures the change in the canonical charge along that path; it is not an invariant measure of a physical correction to the extremality bound.

This observation is deliberately restricted to the pure-electric background. It does not say that the full nonlinear constitutive law, generic dyonic solutions, or quantum fluctuations are independent of $\gamma$. It also shows why a variation of $\gamma$ at fixed noncanonical charge should not be presented as an independent WGC test.

What can be concluded about swampland conjectures?

Electric WGC

The exact extremality curve is one necessary ingredient in a black-hole WGC discussion, but it is not sufficient to establish the conjecture. A particle-level comparison would require at least: (i) the normalization of the compact gauge group and its canonical coupling, (ii) the masses and charges of the allowed states, and (iii) a decay or discharge criterion compatible with the topology, ensemble, and AdS boundary conditions. None of these spectral or global ingredients follows from the static metric.

Two recent results are relevant but do not settle the present AdS problem. Gashti and collaborators studied ModMax with horizon, photon-sphere, and weak-cosmic-censorship diagnostics [7]; such diagnostics are useful characterizations of a background but do not by themselves identify a superextremal particle. Gashti, Pourhassan, and Sakallı later derived a probe/CFT criterion for selected asymptotically flat black holes, including a purely electric Einstein–ModMax branch, by solving the charged-scalar radial problem and imposing a relaxation-time condition [8]. The latter is a global argument in its stated setting; only the local BF calculation is repeated here. We do not import that CFT threshold into topological AdS without solving the corresponding radial problem and boundary-value problem. Other aspects of ModMax black holes, such as their scattering properties and rotating solutions, have also been explored in [23, 24].

A separate strong-field study reports that ModMax fails a necessary charge-growth condition for black-hole decay in an asymptotically flat setting [10]. Its result should not be asserted as a theorem for every AdS topology, but it prevents a blanket claim that the ModMax theory is WGC-compatible solely because some raw charge-to-mass ratios exceed unity. Our conclusions are therefore intentionally limited: the classical extremality diagnostic is established; a WGC pass/fail verdict is not.

When it comes to the distance and de Sitter conjectures [25, 26], things are even more out of reach in our current setup. These conjectures are deeply tied to the global structure of the moduli space and the sign of scalar potentials, which means they demand information about the ultraviolet completion that a purely electric, fixed-$\gamma$ background simply cannot offer. As pointed out in recent reviews [27], testing such ideas usually requires either a controlled string compactification or a well-defined asymptotic boundary problem—and neither of those is available here.

The sign-reversed branch

For $\eta=-1$, the gauge-sector action has the opposite sign. The corresponding Maxwell-like kinetic term around the purely electric branch has negative energy relative to the ordinary-sign theory. The classical solutions may still be analyzed mathematically, but they do not provide a conventional unitary WGC test bed without a separate completion that resolves the wrong-sign sector. The existence of a formal degenerate horizon does not remove this issue. For related studies on the thermodynamic and optical properties of ModMax black holes in various backgrounds, see [28, 29, 30].

Conclusions

We have re-examined four-dimensional topological AdS black holes sourced by ModMax electrodynamics and by its sign-reversed counterpart, consistently distinguishing the field-strength amplitude from the conserved displacement charge. With the standard ModMax Lagrangian, the pure-electric field-strength term is proportional to $e^{+\gamma}$; the $e^{-\gamma}$ factor in the metric appears after the solution is expressed in terms of displacement charge. This convention fixes the potential and yields the first law with the same variables throughout.

The extremality equation is quadratic in $x = r_{\mathrm{e}}^2$. Its correct roots show that the hyperbolic ModMax branch has no upper-charge discriminant bound. Its extremal mass is negative for $1/3 \le r_{\mathrm{e}}^2/l^2 < 1/2$, vanishes at $Q_{\mathrm{e}}^2 e^{-\gamma} = l^2/(64\pi^2)$, and is positive above that threshold. For ModAMax, nonzero-charge extremality is absent for spherical and planar horizons; the hyperbolic outer branch exists only below its discriminant bound and has negative mass. At the endpoint, the root is more degenerate and the usual quadratic AdS$_2$ throat expansion does not apply. We have also examined the zero-mass point on the hyperbolic extremal branch. Although the charge-to-mass ratio diverges at this point, the horizon radius and entropy density remain finite and nonzero, while the extremal mass crosses zero smoothly. This shows that the divergence of the charge-to-mass ratio results from the vanishing denominator and, by itself, does not indicate a superextremal configuration or provide evidence for the weak gravity conjecture. For the pure-electric background, the rescaled charge $\widehat{Q} = e^{-\gamma/2}Q$ removes the explicit $\gamma$-dependence of the Reissner–Nordström-like extremality curve. The apparent growth of $Q/M_{\mathrm{ext}}$ at fixed displacement charge is consequently not, on its own, a normalization-independent strengthening of the WGC. We also evaluated the local charged-scalar BF index for an explicitly stated probe. The chosen parameter set satisfies the local bound on all three ModMax topologies, but this numerical example does not establish stability of the full system.

The supported conclusion is narrower than a claim of WGC consistency: the classical extremality curve and the local probe BF condition are derived, while a genuine WGC test still requires a canonical quantum charge normalization, a charged spectrum, and the relevant black-hole decay calculation. The sign-reversed ModAMax sector has a wrong-sign gauge kinetic term and is not a healthy WGC candidate without a completion. The distance conjectures and the de Sitter conjecture cannot be quantified from a fixed coupling and a negative cosmological constant alone. No claim of full perturbative stability or of cosmic censorship is made.