Ellis-Bronnikov Wormhole: Geometry & Physics

• Ellis–Bronnikov wormhole is a traversable, static, spherically symmetric GR solution sustained by a phantom scalar field that violates the null energy condition.
• Generalizations introduce asymmetry, nonzero masses, alternative coordinate forms, and higher-dimensional embeddings, expanding its theoretical and observational applications.
• Perturbation and rotation studies reveal that while the static configuration is radially unstable, slow and rapid rotations can stabilize the geometry and impact its lensing and shadow profiles.

The Ellis–Bronnikov wormhole is a traversable, static, spherically symmetric wormhole solution of General Relativity supported by a phantom scalar field, i.e. a scalar whose kinetic term has the opposite sign from the canonical one and therefore violates the null energy condition at the throat. In its most familiar massless form, it is horizonless, symmetric between two asymptotically flat regions, and has a throat at a minimum of the areal radius; in more general forms it admits asymmetry between the two ends, nonzero mass parameters, and alternative coordinate representations adapted to lensing, perturbation theory, or numerical construction (Li et al., 1 Feb 2026, Cai et al., 2023).

1. Classical geometry and coordinate representations

In the Morris–Thorne parametrization, a static spherically symmetric wormhole is written as

$$ds^2=-e^{2\Phi(R)}dt^2+\frac{dR^2}{1-b(R)/R}+R^2d\Omega^2.$$

The classic massless Ellis–Bronnikov solution sets

$\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$

with throat radius $R_0=a$, satisfying $b(R_0)=R_0$ and

$b'(R)=-\frac{a^2}{R^2}<1.$

An equivalent proper-radial-coordinate form uses $r\in(-\infty,+\infty)$ and

$R(r)=\sqrt{r^2+a^2},$

so that

$ds^2=-dt^2+dr^2+(r^2+a^2)d\Omega^2,$

with the throat at $r=0$ and $R(0)=a$ (Li et al., 1 Feb 2026).

Several later works employ equivalent 4D forms with different parameter conventions. One representative massive/asymmetric Bronnikov form is

$\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$0

with

$\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$1

and

$\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$2

In this representation the throat is at the minimum of $\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$3, located at $\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$4, and the spacetime has two asymptotically flat regions denoted Universe I and Universe II. The horizonless condition is $\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$5 (Cai et al., 2023).

Another widely used form is

$\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$6

with areal radius

$\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$7

In this gauge the throat occurs at $\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$8, and the two asymptotic masses differ in sign: $\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$9 The symmetric Ellis limit is recovered by setting $R_0=a$0, giving

$R_0=a$1

with throat at $R_0=a$2 (Ishkaeva et al., 2023).

A further form used in perturbation theory writes the static background as

$R_0=a$3

with

$R_0=a$4

where $R_0=a$5 controls asymmetry and $R_0=a$6 marks the throat (Azad et al., 2023).

These forms are coordinate or parameter re-expressions of the same basic idea: a regular throat connects two asymptotic regions without an event horizon, while the redshift and shape functions determine whether the configuration is symmetric, massless, or asymmetric. This suggests that “Ellis–Bronnikov wormhole” functions as both the name of the original massless solution and a broader label for a family of phantom-supported wormhole geometries continuously connected to it.

2. Phantom scalar support and energy-condition violation

The defining matter source is a phantom scalar field. In one normalization, the 4D action is

$R_0=a$7

with $R_0=a$8 for the phantom choice in signature $R_0=a$9 (Li et al., 1 Feb 2026). Equivalent conventions appear elsewhere as

$b(R_0)=R_0$0

(Azad et al., 2023), or

$b(R_0)=R_0$1

(Khoo et al., 2024). The sign difference relative to a canonical scalar is the essential feature.

For the massless EB geometry in proper-radial coordinates, a representative scalar profile is

$b(R_0)=R_0$2

which generates the stress-energy needed to sustain $b(R_0)=R_0$3 with $b(R_0)=R_0$4 (Li et al., 1 Feb 2026). In the static asymmetric background used in slow-rotation studies, the scalar is

$b(R_0)=R_0$5

(Azad et al., 2023). In the EB line element used for scalar-cloud and Q-ball studies, the background geometry is fixed and the support field is not re-derived, but the paper explicitly identifies the wormhole as the standard phantom-supported EB background (Blázquez-Salcedo et al., 2022).

The null energy condition is violated near the throat. In the classic EB solution one has $b(R_0)=R_0$6 near the throat (Li et al., 1 Feb 2026). In the wormhole-adapted gauge used for higher-curvature generalizations, the matter-sector NEC along the radial direction becomes

$b(R_0)=R_0$7

so the NEC violation is purely radial (Li et al., 1 Feb 2026). In double-null dynamics, the corresponding null fluxes $b(R_0)=R_0$8 and $b(R_0)=R_0$9 are negative everywhere in the static massless configuration, with strongest violation at the throat (Xu et al., 10 Mar 2025).

The canonical massless EB wormhole has zero ADM mass and approaches Minkowski space on both ends (Li et al., 1 Feb 2026). More general EB solutions are asymmetric and assign different asymptotic masses to the two universes. In the Bronnikov form,

$b'(R)=-\frac{a^2}{R^2}<1.$0

(Cai et al., 2023). In the $b'(R)=-\frac{a^2}{R^2}<1.$1-gauge quoted above,

$b'(R)=-\frac{a^2}{R^2}<1.$2

(Ishkaeva et al., 2023). Thus, the massless symmetric EB wormhole is a special case of a broader asymmetric family in which one asymptotic region can appear as a negative-mass end.

A recurrent misconception is that traversability alone implies ordinary matter support. The supplied results consistently show the opposite for the classical 4D EB solution: the throat is maintained by a matter sector that violates the NEC, whether described as a phantom scalar, a ghost scalar, or an effective exotic fluid (Li et al., 1 Feb 2026, Azad et al., 2023, Takahashi et al., 1 Jun 2026).

3. Generalizations, higher-dimensional embeddings, and modified gravity

A substantial literature extends the EB geometry beyond its original 4D GR setting. One direction introduces generalized Ellis–Bronnikov shape functions labeled by an even parameter $b'(R)=-\frac{a^2}{R^2}<1.$3 or $b'(R)=-\frac{a^2}{R^2}<1.$4. In one formulation,

$b'(R)=-\frac{a^2}{R^2}<1.$5

where $b'(R)=-\frac{a^2}{R^2}<1.$6 recovers the standard EB geometry and larger even $b'(R)=-\frac{a^2}{R^2}<1.$7 yields increasingly cylinder-like throats (Souza et al., 2022, Furtado et al., 2022). In Morris–Thorne form another generalized shape function is

$b'(R)=-\frac{a^2}{R^2}<1.$8

with $b'(R)=-\frac{a^2}{R^2}<1.$9 reproducing $r\in(-\infty,+\infty)$0 (Nilton et al., 2022). A related presentation uses

$r\in(-\infty,+\infty)$1

recovering EB at $r\in(-\infty,+\infty)$2 (Crispim et al., 2024).

The generalized geometries alter flare-out details and matter requirements. In asymptotically safe gravity, the generalized EB family is analyzed by replacing Newton’s constant with a running coupling built from curvature scalars. For the standard EB shape $r\in(-\infty,+\infty)$3, the curvature invariants are

$r\in(-\infty,+\infty)$4

and the improvement functions are chosen as

$r\in(-\infty,+\infty)$5

For the original EB throat, the asymptotically safe models allow radial NEC, WEC, and DEC to hold at the throat for finite intervals of $r\in(-\infty,+\infty)$6, e.g. $r\in(-\infty,+\infty)$7 in the Ricci and squared-Ricci models, even though the effective equation-of-state parameter remains exotic (Alencar et al., 2021). For generalized EB wormholes in asymptotically safe gravity, only the $r\in(-\infty,+\infty)$8 case under Ricci-scalar improvement yields a window

$r\in(-\infty,+\infty)$9

in which ordinary matter is allowed at the throat, whereas the squared-Ricci and Kretschmann prescriptions continue to require exotic support (Nilton et al., 2022).

Another direction embeds EB/GEB wormholes in a warped 5D braneworld. With

$R(r)=\sqrt{r^2+a^2},$0

and warp factor

$R(r)=\sqrt{r^2+a^2},$1

a decaying warp factor can render the effective on-brane weak energy condition compatible even with the EB geometry, improving all other energy-condition diagnostics relative to the 4D case (Sharma et al., 2021). This suggests that extra-dimensional geometry can partly replace exotic matter from the four-dimensional viewpoint.

Higher-curvature extensions also preserve the EB structure while modifying its support. In quasi-topological gravity, the action

$R(r)=\sqrt{r^2+a^2},$2

admits higher-dimensional traversable wormholes supported by a phantom scalar, with second-order ODEs in the static spherically symmetric sector. In the GR limit $R(r)=\sqrt{r^2+a^2},$3, one recovers the EB system with

$R(r)=\sqrt{r^2+a^2},$4

(Li et al., 1 Feb 2026).

A distinct extension derives dyonic EB wormholes from a warped Kaluza–Klein reduction. In the resulting 4D Einstein-frame theory,

$R(r)=\sqrt{r^2+a^2},$5

the EB geometry with $R(r)=\sqrt{r^2+a^2},$6 is supported by a phantom dilaton, canonical axion, and dyonic Maxwell and Kaluza–Klein fields. The scalar profile is fixed by

$R(r)=\sqrt{r^2+a^2},$7

leading to

$R(r)=\sqrt{r^2+a^2},$8

(Lobo et al., 2 May 2026).

Taken together, these constructions show that the EB wormhole is not merely an isolated GR curiosity. It is a template geometry recurrently used in asymptotically safe gravity, quasi-topological gravity, higher-dimensional warped compactifications, and generalized shape-function families.

4. Rotation, perturbations, and stability

The static EB wormhole has a well-known radial instability in the $R(r)=\sqrt{r^2+a^2},$9 sector (Azad et al., 2023). A central recent question is whether rotation changes this conclusion.

In the slow-rotation expansion, the background metric is written to second order in a small rotation parameter $ds^2=-dt^2+dr^2+(r^2+a^2)d\Omega^2,$0 as

$ds^2=-dt^2+dr^2+(r^2+a^2)d\Omega^2,$1

with $ds^2=-dt^2+dr^2+(r^2+a^2)d\Omega^2,$2 (Azad et al., 2023). The $ds^2=-dt^2+dr^2+(r^2+a^2)d\Omega^2,$3 perturbation problem reduces to a Schrödinger-type equation

$ds^2=-dt^2+dr^2+(r^2+a^2)d\Omega^2,$4

and the imaginary part of the unstable frequency behaves as

$ds^2=-dt^2+dr^2+(r^2+a^2)d\Omega^2,$5

For the main unstable branch, $ds^2=-dt^2+dr^2+(r^2+a^2)d\Omega^2,$6 for all asymmetry parameters $ds^2=-dt^2+dr^2+(r^2+a^2)d\Omega^2,$7 studied, indicating a stabilizing tendency. The critical scaled angular momentum at which this principal unstable mode disappears is

$ds^2=-dt^2+dr^2+(r^2+a^2)d\Omega^2,$8

roughly half the limiting value approached as the wormhole tends toward extremal Kerr (Azad et al., 2023).

The same study identifies a second unstable $ds^2=-dt^2+dr^2+(r^2+a^2)d\Omega^2,$9 mode emerging from a static zero mode, with $r=0$0, which crosses the main branch near

$r=0$1

The authors regard this crossing as a limitation of the quadratic slow-rotation approximation and conjecture that nonperturbatively the instability disappears slightly earlier (Azad et al., 2023). Thus, the evidence supports rotation as a stabilizing mechanism, but not yet a complete proof of full linear mode stability.

Rapidly rotating EB wormholes have been studied numerically at the level of quasinormal modes. The stationary axisymmetric line element is

$r=0$2

For symmetric static EB wormholes, axial-led, polar-led, and scalar perturbations are triply isospectral. Rotation breaks this triple isospectrality and produces three distinct branches—polar-led 1, axial-led, and polar-led 2—for each $r=0$3-led family (Khoo et al., 2024). In the sectors studied, $r=0$4, no instabilities were found; all computed quasinormal modes satisfy $r=0$5 (Khoo et al., 2024). This does not settle the radial sector, but it shows that rotation enriches the spectrum without automatically introducing new instabilities in the explored nonradial modes.

Scalar test-field analysis on the ultrastatic EB background reaches a complementary conclusion. For a free massive scalar

$r=0$6

with separation

$r=0$7

the radial equation becomes

$r=0$8

which can be rewritten as

$r=0$9

The exact solutions are expressed in terms of confluent Heun functions (Blázquez-Salcedo et al., 2022). Smooth localized free scalar clouds do not exist because a no-go identity makes the relevant integral strictly positive for $R(0)=a$0; patched configurations with derivative jumps at the throat would require an additional delta-source there (Blázquez-Salcedo et al., 2022). By contrast, with self-interaction

$R(0)=a$1

the background supports spherical and spinning Q-balls in the frequency window

$R(0)=a$2

(Blázquez-Salcedo et al., 2022).

These results narrow the stability picture. The static EB wormhole is radially unstable, slow rotation appears to suppress that instability, rapidly rotating backgrounds show no instability in the computed $R(0)=a$3 sectors, and nonlinear self-interacting scalar configurations can exist smoothly on the EB background.

5. Geodesics, lensing, shadows, and optical appearance

Null geodesics in EB geometries have been analyzed in several coordinate systems. For a generic static spherical metric

$R(0)=a$4

the impact parameter and closest approach satisfy

$R(0)=a$5

The general bending angle is

$R(0)=a$6

(Cai et al., 2023).

For the massless Ellis limit $R(0)=a$7, the exact weak-field deflection angle is

$R(0)=a$8

where $R(0)=a$9 is the complete elliptic integral of the first kind, with expansion

$\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$00

For the general massive EB wormhole, after introducing side-dependent rescaled impact parameters $\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$01, the weak-field expansion through $\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$02 is

$\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$03

Among the approximation schemes compared, isotropic-coordinate PPN reproduces the direct analytic expansion exactly to this order and performs best, followed by improved Gauss–Bonnet methods in isotropic gauge, then the Amore–Diaz formalism (Cai et al., 2023).

The general EB wormhole also admits photon rings and shadow-like structures. In the $\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$04-gauge,

$\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$05

the circular photon orbit lies at

$\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$06

with critical impact parameter

$\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$07

The paper distinguishes the shadow boundary from the throat silhouette produced by near-source emission; for the latter, the screen radius $\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$08 is determined implicitly by

$\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$09

For the parameters studied, both the shadow radius and the throat-silhouette radius exceed the corresponding Schwarzschild values at equal $\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$10 (Ishkaeva et al., 2023).

Accretion-image simulations sharpen the observational picture. In general-EB thin-disk imaging, the EB dark area is larger than the Schwarzschild one; for one example with $\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$11 and $\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$12, the EB central dark area is $\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$13 larger (Ishkaeva et al., 2023). In spherically symmetric synchrotron GRRT calculations, both EB and Schwarzschild images exhibit a central shadow and a bright photon ring, but the EB shadow interior and photon ring are brighter because the absence of a horizon allows emission from near and beyond the throat to contribute (Takahashi et al., 1 Jun 2026). Under the spherical-flow prescription used there, both EB and Schwarzschild models remain broadly compatible with current EHT constraints for M87* (Takahashi et al., 1 Jun 2026).

Optical appearances become strongly side-dependent in asymmetric EB wormholes. In the metric

$\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$14

with $\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$15, the throat occurs at

$\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$16

and the unstable circular null orbit at

$\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$17

A key relation derived in this context is

$\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$18

so if observer and disk lie on opposite sides of the throat, the Euclidean aiming distance $\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$19 and the relativistic impact parameter $\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$20 differ by the asymptotic normalization factor (Cui et al., 29 Jun 2026). For an observer on the opposite side, the direct, lensing, and photon-ring order can appear inverted, producing an “internal/external inversion” in radial image structure and brightness profiles (Cui et al., 29 Jun 2026). Small $\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$21 values with the observer on the $\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$22 side can mimic EHT images to some extent, whereas large $\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$23 or observers on the $\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$24 side are disfavored (Cui et al., 29 Jun 2026).

The optical and lensing literature therefore identifies several EB-specific signatures: absence of a leading $\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$25 term in the massless limit, side-dependent asymptotic normalization, larger dark-region size than Schwarzschild at fixed mass scale, potential visibility of emission from the opposite side of the throat, and richer ring morphology in thin or optically thin flows.

6. Dynamical evolution, accretion, and source realizations

Nonlinear evolution studies show that the static EB wormhole is dynamically fragile. In double-null coordinates

$\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$26

the throat is characterized by the coincidence of

$\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$27

Perturbing the massless BE/EB wormhole with ingoing pulses yields two distinct outcomes. A normal scalar pulse drives collapse into a black hole, with a spacelike singularity at $\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$28 hidden behind horizons. A phantom pulse instead drives inflationary expansion, splitting the horizons into cosmological-type horizons and causally disconnecting the two asymptotic regions (Xu et al., 10 Mar 2025). Increasing pulse amplitude accelerates both instabilities, while increasing the wormhole mass parameter delays them (Xu et al., 10 Mar 2025). A strategically timed outgoing phantom pulse colliding with an ingoing normal scalar pulse can temporarily restore the coincidence of $\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$29 and $\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$30, delaying collapse, but does not produce permanent stabilization (Xu et al., 10 Mar 2025).

Accretion studies provide a complementary dynamical perspective. For a barotropic fluid $\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$31 accreting onto the isotropic-coordinate EBWH,

$\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$32

with

$\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$33

the ADM mass is $\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$34 and the throat radius is

$\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$35

For the massless case $\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$36, the Misner–Sharp mass is

$\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$37

giving $\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$38 at the throat (Yusupova et al., 25 Jun 2026). The paper derives closed-form profiles for radial velocity, density, and mass variation and concludes that, within its comparison framework, the mass of the Schwarzschild black hole increases under both phantom and non-phantom accretion, whereas the EBWH mass decreases. The massless EBWH shows the same accretion patterns as the massive one, so accretion cannot distinguish the two cases (Yusupova et al., 25 Jun 2026).

Generalized EB geometries also admit exact source constructions beyond a pure phantom scalar. A recent result shows that the generalized geometry

$\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$39

is an exact solution of GR when supported by a phantom scalar plus a magnetic or electric nonlinear electromagnetic source. The scalar profile obeys

$\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$40

which reduces at $\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$41 to

$\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$42

the standard EB field (Crispim et al., 2024). This result shows that generalized EB wormholes need not be interpreted only as phenomenological deformations; they arise as exact Einstein solutions with explicitly identified field sources.

A different stationary generalization starts from the vacuum ring wormhole and then “dresses” it with the unique bounded scalar

$\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$43

in oblate spheroidal coordinates. In that construction the vacuum Ernst sector fixes $\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$44 and $\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$45, while the scalar dressing modifies only the spatial block and cancels the ring singularity, yielding a globally regular spinning wormhole with two asymptotically flat regions (Volkov, 2021). The small-rotation relation

$\Phi(R)=0,\qquad b(R)=\frac{a^2}{R},$46

reflects the memory of the extended ring source even after scalar screening (Volkov, 2021).

Across these dynamical and source-based studies, a consistent picture emerges. The EB wormhole is easy to write down but hard to stabilize; it can collapse or inflate under perturbations, respond to accretion in ways distinct from black holes, and admit exact extensions with electromagnetic, axionic, higher-curvature, or higher-dimensional sectors. This suggests that the EB solution remains the canonical laboratory for isolating which properties of traversable wormholes are tied specifically to phantom support and which survive under more elaborate ultraviolet or matter-sector completions.