Chameleon Gravity and Density-Dependent Screening

• Chameleon gravity is a scalar-tensor theory where a light scalar field acquires an effective mass that increases in high-density environments, effectively screening extra forces.
• The theory uses a conformal coupling, A(φ)=exp(βφ/Mₚₗ), which mediates a fifth force that is competitively suppressed via a thin-shell mechanism in astrophysical bodies.
• Its applications span f(R) gravity, galactic dynamics, and cosmology, with tests from laboratory, Solar System, and cluster observations providing tight constraints.

Chameleon gravity is a class of scalar–tensor modifications of gravity in which a light scalar field acquires an environment-dependent effective mass through its coupling to matter, becoming heavy in high-density regions and light in low-density regions. In the Einstein-frame formulation, matter couples conformally to the metric through a factor such as $A(\phi)=\exp(\beta\phi/M_{\rm Pl})$, so that the scalar mediates a fifth force $F_5=-(\beta/M_{\rm Pl})\nabla\phi$, while the effective potential $V_{\rm eff}(\phi;\rho)=V(\phi)+\rho\,A(\phi)$ makes the range and amplitude of that force depend on ambient density (Burrage et al., 2017). This screening mechanism was developed to reconcile light scalars with laboratory and Solar-System bounds, yet it has also been used in $f(R)$ gravity, cluster phenomenology, galactic dynamics, stellar streams, binary pulsars, and cosmology, where the same density dependence can generate qualitatively different behavior across astrophysical environments (Lombriser, 2014).

1. Fundamental formulation

In its standard Einstein-frame form, chameleon gravity is defined by an action of the type

$$S=\int d^4x\sqrt{-g}\Bigl[\frac{M_{\rm Pl}^2}{2}R-\frac12(\partial\phi)^2-V(\phi)\Bigr]+S_m[\psi_m,A^2(\phi)g_{\mu\nu}] ,$$

with $M_{\rm Pl}\equiv(8\pi G)^{-1/2}$, a scalar self-interaction potential $V(\phi)$, and a conformal matter coupling $A(\phi)$, often chosen as $A(\phi)=\exp(\beta\phi/M_{\rm Pl})$ (Anderson et al., 2012). In the nonrelativistic limit, the scalar equation becomes

$$\nabla^2\phi = V'(\phi)+\frac{\beta}{M_{\rm Pl}}\rho ,$$

or, equivalently, $F_5=-(\beta/M_{\rm Pl})\nabla\phi$0 with

$F_5=-(\beta/M_{\rm Pl})\nabla\phi$1

so the field’s local minimum and curvature depend explicitly on ambient density (Morris, 2014).

The same structure appears in several equivalent or related formulations. In metric $F_5=-(\beta/M_{\rm Pl})\nabla\phi$2 gravity one begins from

$F_5=-(\beta/M_{\rm Pl})\nabla\phi$3

defines $F_5=-(\beta/M_{\rm Pl})\nabla\phi$4, and obtains the quasi-static weak-field system

$F_5=-(\beta/M_{\rm Pl})\nabla\phi$5

which implies $F_5=-(\beta/M_{\rm Pl})\nabla\phi$6 (Stark et al., 2016). In the Einstein-frame representation of chameleon $F_5=-(\beta/M_{\rm Pl})\nabla\phi$7 gravity one may introduce $F_5=-(\beta/M_{\rm Pl})\nabla\phi$8 through $F_5=-(\beta/M_{\rm Pl})\nabla\phi$9, so that the matter coupling is fixed to $V_{\rm eff}(\phi;\rho)=V(\phi)+\rho\,A(\phi)$0, corresponding to $V_{\rm eff}(\phi;\rho)=V(\phi)+\rho\,A(\phi)$1 (Zhang, 2022).

Jordan-frame formulations are also used. A generalized Brans–Dicke construction with a potential $V_{\rm eff}(\phi;\rho)=V(\phi)+\rho\,A(\phi)$2 and matter coupling $V_{\rm eff}(\phi;\rho)=V(\phi)+\rho\,A(\phi)$3 yields an effective potential $V_{\rm eff}(\phi;\rho)=V(\phi)+\rho\,A(\phi)$4, but the realization of a true chameleon behavior then depends sensitively on the chosen $V_{\rm eff}(\phi;\rho)=V(\phi)+\rho\,A(\phi)$5 and $V_{\rm eff}(\phi;\rho)=V(\phi)+\rho\,A(\phi)$6 (Bisabr, 2014). A purely Jordan-frame analysis of metric $V_{\rm eff}(\phi;\rho)=V(\phi)+\rho\,A(\phi)$7 gravity similarly derives the density-dependent effective mass directly from the trace equation, without using the scalar–tensor identification in the Einstein frame (Negrelli et al., 2020).

2. Screening, thin shells, and force suppression

The defining mechanism of chameleon gravity is the density dependence of the scalar mass,

$V_{\rm eff}(\phi;\rho)=V(\phi)+\rho\,A(\phi)$8

which becomes large in high-density environments and small in low-density environments (Farajollahi et al., 2012). For widely used power-law potentials such as $V_{\rm eff}(\phi;\rho)=V(\phi)+\rho\,A(\phi)$9, the minimum obeys

$f(R)$0

and one finds $f(R)$1 and $f(R)$2, so the scalar can be cosmologically light while remaining short-ranged in terrestrial or Solar-System settings (Pourhasan et al., 2011).

For a spherical body of radius $f(R)$3, Newtonian potential $f(R)$4, interior density $f(R)$5, and ambient density $f(R)$6, the thin-shell parameter is conventionally written as

$f(R)$7

or, in $f(R)$8-adapted notation,

$f(R)$9

If the body is thin-shelled, only a thin outer shell sources the exterior scalar profile, and the fifth force is strongly suppressed (Morris, 2014). In the standard summary of the mechanism, the exterior field is reduced by the factor $$S=\int d^4x\sqrt{-g}\Bigl[\frac{M_{\rm Pl}^2}{2}R-\frac12(\partial\phi)^2-V(\phi)\Bigr]+S_m[\psi_m,A^2(\phi)g_{\mu\nu}] ,$$0, while the force ratio becomes

$$S=\int d^4x\sqrt{-g}\Bigl[\frac{M_{\rm Pl}^2}{2}R-\frac12(\partial\phi)^2-V(\phi)\Bigr]+S_m[\psi_m,A^2(\phi)g_{\mu\nu}] ,$$1

with $$S=\int d^4x\sqrt{-g}\Bigl[\frac{M_{\rm Pl}^2}{2}R-\frac12(\partial\phi)^2-V(\phi)\Bigr]+S_m[\psi_m,A^2(\phi)g_{\mu\nu}] ,$$2 the screening radius (Burrage et al., 2017).

In $$S=\int d^4x\sqrt{-g}\Bigl[\frac{M_{\rm Pl}^2}{2}R-\frac12(\partial\phi)^2-V(\phi)\Bigr]+S_m[\psi_m,A^2(\phi)g_{\mu\nu}] ,$$3 models this screening can be expressed in terms of the scalaron $$S=\int d^4x\sqrt{-g}\Bigl[\frac{M_{\rm Pl}^2}{2}R-\frac12(\partial\phi)^2-V(\phi)\Bigr]+S_m[\psi_m,A^2(\phi)g_{\mu\nu}] ,$$4. In the Hu–Sawicki model, the quasi-static equations

$$S=\int d^4x\sqrt{-g}\Bigl[\frac{M_{\rm Pl}^2}{2}R-\frac12(\partial\phi)^2-V(\phi)\Bigr]+S_m[\psi_m,A^2(\phi)g_{\mu\nu}] ,$$5

show that in high-density regions the field sits at $$S=\int d^4x\sqrt{-g}\Bigl[\frac{M_{\rm Pl}^2}{2}R-\frac12(\partial\phi)^2-V(\phi)\Bigr]+S_m[\psi_m,A^2(\phi)g_{\mu\nu}] ,$$6, recovering GR, whereas in unscreened regions the effective Newton’s constant is enhanced by $$S=\int d^4x\sqrt{-g}\Bigl[\frac{M_{\rm Pl}^2}{2}R-\frac12(\partial\phi)^2-V(\phi)\Bigr]+S_m[\psi_m,A^2(\phi)g_{\mu\nu}] ,$$7 below the Compton wavelength (Lombriser et al., 2012). A halo-model treatment parameterizes the transition through a chameleon mass threshold $$S=\int d^4x\sqrt{-g}\Bigl[\frac{M_{\rm Pl}^2}{2}R-\frac12(\partial\phi)^2-V(\phi)\Bigr]+S_m[\psi_m,A^2(\phi)g_{\mu\nu}] ,$$8 and rapidity exponent $$S=\int d^4x\sqrt{-g}\Bigl[\frac{M_{\rm Pl}^2}{2}R-\frac12(\partial\phi)^2-V(\phi)\Bigr]+S_m[\psi_m,A^2(\phi)g_{\mu\nu}] ,$$9, with

$M_{\rm Pl}\equiv(8\pi G)^{-1/2}$0

so that screened and unscreened regimes can be interpolated directly in nonlinear structure calculations (Li et al., 2011).

A notable conceptual development is the electrostatic analogy: in the thin-shell regime, the field outside a screened body obeys $M_{\rm Pl}\equiv(8\pi G)^{-1/2}$1 with nearly constant $M_{\rm Pl}\equiv(8\pi G)^{-1/2}$2 inside, mathematically analogous to the potential outside a conducting sphere held at fixed potential. This permits a method-of-images construction and yields a repulsive chameleonic self-force for a test object near a screened spherical source (Pourhasan et al., 2011).

3. Model classes and theoretical variants

The most widely studied chameleon models use runaway bare potentials together with exponential matter couplings. Representative choices are

$M_{\rm Pl}\equiv(8\pi G)^{-1/2}$3

or the related form $M_{\rm Pl}\equiv(8\pi G)^{-1/2}$4 used in stellar-stream analyses (Naik et al., 2020). These models were motivated by the possibility that a scalar field could be relevant to dark energy, dark matter, or neutrino physics while remaining compatible with fifth-force bounds through environmental screening (Pourhasan et al., 2011).

A major subclass is chameleon $M_{\rm Pl}\equiv(8\pi G)^{-1/2}$5 gravity, especially the Hu–Sawicki model. In the large-curvature regime it may be written as

$M_{\rm Pl}\equiv(8\pi G)^{-1/2}$6

with $M_{\rm Pl}\equiv(8\pi G)^{-1/2}$7 acting as the scalar degree of freedom (Lombriser et al., 2012). In this setting $M_{\rm Pl}\equiv(8\pi G)^{-1/2}$8 controls the background scalar amplitude and, therefore, the screening scale. The corresponding cosmological and cluster phenomenology is often cast directly in terms of $M_{\rm Pl}\equiv(8\pi G)^{-1/2}$9 rather than $V(\phi)$0 (Stark et al., 2016).

Not all scalar–tensor extensions realize the chameleon mechanism equally well. In generalized Brans–Dicke models with exponential or power-law $V(\phi)$1 and $V(\phi)$2, the existence of a density-dependent minimum is not sufficient by itself; local-gravity constraints can drive the model parameters into highly fine-tuned or inconsistent regions. This suggests that chameleon behavior in such theories is not generic and depends significantly on the forms attributed to the potential and the coupling functions (Bisabr, 2014).

A distinct theoretical response to early-Universe difficulties is the quartic chameleon, defined by

$V(\phi)$3

Because the potential is scale-free, the field oscillates after inflation around the minimum of its effective potential and only generates high-energy perturbations at comparably high temperatures. In this construction the theory can remain a well-behaved effective field theory at nucleosynthesis for $V(\phi)$4 and $V(\phi)$5, in contrast to the runaway potentials that suffer catastrophic excitation near Big Bang Nucleosynthesis (Miller et al., 2016).

A more recent extension proposes chameleon gravity on galactic scales as an alternative to dark matter. In that construction, a static spherically symmetric halo solution yields $V(\phi)$6, a scalar-field mass profile $V(\phi)$7, and hence

$V(\phi)$8

in the flat-rotation-curve region (Zaregonbadi et al., 18 Jul 2025). This suggests that, in at least one model realization, the chameleon sector can be organized to mimic an effective $V(\phi)$9 halo density profile.

4. Astrophysical and cosmological phenomenology

Chameleon screening has been studied extensively in nonlinear structure formation. Systematic $A(\phi)$0-body simulations using ECOSMOG for a generic four-parameter family $A(\phi)$1 found that linear theory overpredicts deviations from $A(\phi)$2CDM even on scales as large as $A(\phi)$3, and that the chameleon mechanism is significantly more efficient than the dilaton and symmetron in high-density regions and at early times (Brax et al., 2013). Under local-screening-motivated conditions $A(\phi)$4, the models considered there produce $A(\phi)$5 deviations in $A(\phi)$6 on scales $A(\phi)$7 for most parameter choices (Brax et al., 2013).

Halo and cluster structure can be modeled semi-analytically. In virialized cluster simulations of Hu–Sawicki $A(\phi)$8 gravity, the standard NFW halo density profile and the radial power law for the pseudo phase-space density provide equally good fits for $A(\phi)$9 clusters as they do in the Newtonian scenario, and analytic profiles for $A(\phi)=\exp(\beta\phi/M_{\rm Pl})$0, $A(\phi)=\exp(\beta\phi/M_{\rm Pl})$1, and $A(\phi)=\exp(\beta\phi/M_{\rm Pl})$2 track simulations to $A(\phi)=\exp(\beta\phi/M_{\rm Pl})$3 with fitting parameters $A(\phi)=\exp(\beta\phi/M_{\rm Pl})$4, $A(\phi)=\exp(\beta\phi/M_{\rm Pl})$5, and $A(\phi)=\exp(\beta\phi/M_{\rm Pl})$6 (Lombriser et al., 2012). A complementary halo-model description captures the nonlinear transition by modifying $A(\phi)=\exp(\beta\phi/M_{\rm Pl})$7 across the screening threshold, predicting a “pile-up” of halo abundance around $A(\phi)=\exp(\beta\phi/M_{\rm Pl})$8 and corresponding structure in the nonlinear power spectrum (Li et al., 2011).

Galaxy clusters supply some of the cleanest observational tests because the fifth force affects hot gas but not weak lensing in conformally coupled chameleon models. In the Coma analysis, the chameleon modification is tested through the discrepancy between hydrostatic and lensing mass profiles, leading in the $A(\phi)=\exp(\beta\phi/M_{\rm Pl})$9 case to $$\nabla^2\phi = V'(\phi)+\frac{\beta}{M_{\rm Pl}}\rho ,$$0 (Terukina et al., 2013). A stacked analysis of 58 clusters from the XMM Cluster Survey and CFHTLenS, using a joint MCMC fit to X-ray and lensing profiles, likewise found no fifth-force signal and obtained $$\nabla^2\phi = V'(\phi)+\frac{\beta}{M_{\rm Pl}}\rho ,$$1 (95% CL) in the $$\nabla^2\phi = V'(\phi)+\frac{\beta}{M_{\rm Pl}}\rho ,$$2 case (Wilcox et al., 2015). On larger cluster samples, the phase-space escape-edge method compares high-mass screened clusters with low-mass unscreened clusters through the ratio

$$\nabla^2\phi = V'(\phi)+\frac{\beta}{M_{\rm Pl}}\rho ,$$3

and forecasts that a DESI Bright Galaxy Survey-like sample can distinguish $$\nabla^2\phi = V'(\phi)+\frac{\beta}{M_{\rm Pl}}\rho ,$$4 from GR at $$\nabla^2\phi = V'(\phi)+\frac{\beta}{M_{\rm Pl}}\rho ,$$5 and $$\nabla^2\phi = V'(\phi)+\frac{\beta}{M_{\rm Pl}}\rho ,$$6 at $$\nabla^2\phi = V'(\phi)+\frac{\beta}{M_{\rm Pl}}\rho ,$$7 (Stark et al., 2016).

At galactic scales, the environmental nature of screening leads to equivalence-principle-violating phenomenology. In the outskirts of NFW haloes, the electrostatic image-mass construction implies a repulsive self-force on satellites near a thin shell, and intermediate-mass satellites can be slower than their larger or smaller counterparts by as much as $$\nabla^2\phi = V'(\phi)+\frac{\beta}{M_{\rm Pl}}\rho ,$$8 close to a thin shell (Pourhasan et al., 2011). Stellar streams provide a related probe: if main-sequence stars self-screen while the surrounding dark matter does not, then leading and trailing streams become asymmetric. Simulations of streams with apocentres between $$\nabla^2\phi = V'(\phi)+\frac{\beta}{M_{\rm Pl}}\rho ,$$9 and $F_5=-(\beta/M_{\rm Pl})\nabla\phi$00 kpc indicate attainable constraints at the level of $F_5=-(\beta/M_{\rm Pl})\nabla\phi$01, while $F_5=-(\beta/M_{\rm Pl})\nabla\phi$02 or even $F_5=-(\beta/M_{\rm Pl})\nabla\phi$03 are described as plausible if environmental screening of the satellite is accounted for (Naik et al., 2020).

Cosmological analyses have also treated the chameleon directly as a homogeneous field. In a flat FLRW model with $F_5=-(\beta/M_{\rm Pl})\nabla\phi$04, both power-law and exponential potentials can explain current acceleration through a low-mass chameleon field, while the effective mass increases with redshift and the power-law potential is reported to be in better agreement with the observational data considered there (Farajollahi et al., 2012). Cosmographic reconstructions for a model with $F_5=-(\beta/M_{\rm Pl})\nabla\phi$05 and matter coupling $F_5=-(\beta/M_{\rm Pl})\nabla\phi$06 derive algebraic relations between $F_5=-(\beta/M_{\rm Pl})\nabla\phi$07 and the cosmographic parameters $F_5=-(\beta/M_{\rm Pl})\nabla\phi$08, with the jerk $F_5=-(\beta/M_{\rm Pl})\nabla\phi$09 singled out as especially important once $F_5=-(\beta/M_{\rm Pl})\nabla\phi$10 are partially specified (Salehi, 2020).

5. Solar-System, geodetic, and strong-field tests

Solar-System viability is a central constraint on chameleon gravity. In $F_5=-(\beta/M_{\rm Pl})\nabla\phi$11 language, a Jordan-frame treatment of compact objects in a static spherically symmetric spacetime shows that a screened regime requires $F_5=-(\beta/M_{\rm Pl})\nabla\phi$12, in which case the PPN parameter satisfies $F_5=-(\beta/M_{\rm Pl})\nabla\phi$13, compatible with Cassini (Negrelli et al., 2020). Applying this criterion, Starobinsky models with $F_5=-(\beta/M_{\rm Pl})\nabla\phi$14, Hu–Sawicki models with $F_5=-(\beta/M_{\rm Pl})\nabla\phi$15, and the exponential model are found to pass, whereas the logarithmic “MJWQ” model fails (Negrelli et al., 2020).

Satellite geodesy has been proposed as a direct probe of the Earth’s screening. In a monopole approximation, the exterior potential of the Earth takes the form

$F_5=-(\beta/M_{\rm Pl})\nabla\phi$16

with $F_5=-(\beta/M_{\rm Pl})\nabla\phi$17 the Earth’s thin-shell factor (Morris, 2014). Using the Cassini bound $F_5=-(\beta/M_{\rm Pl})\nabla\phi$18, one obtains $F_5=-(\beta/M_{\rm Pl})\nabla\phi$19, and with $F_5=-(\beta/M_{\rm Pl})\nabla\phi$20 this implies $F_5=-(\beta/M_{\rm Pl})\nabla\phi$21, or $F_5=-(\beta/M_{\rm Pl})\nabla\phi$22 (Morris, 2014). LAGEOS bounds on a Yukawa perturbation can be translated into $F_5=-(\beta/M_{\rm Pl})\nabla\phi$23 to $F_5=-(\beta/M_{\rm Pl})\nabla\phi$24, implying $F_5=-(\beta/M_{\rm Pl})\nabla\phi$25 to $F_5=-(\beta/M_{\rm Pl})\nabla\phi$26 (Morris, 2014). The same work suggests that a comparison of ground-based and space-based multipole moments of the geopotential could reveal a chameleon effect (Morris, 2014).

The Pioneer anomaly provided an early Solar-System application. Because a small spacecraft with no thin shell can experience a more pronounced anomalous acceleration than a large compact body with a thin shell, the chameleon mechanism reproduces the qualitative difference between spacecraft and planets. Quantitatively, however, using $F_5=-(\beta/M_{\rm Pl})\nabla\phi$27 for the Sun gives $F_5=-(\beta/M_{\rm Pl})\nabla\phi$28 and $F_5=-(\beta/M_{\rm Pl})\nabla\phi$29 at $F_5=-(\beta/M_{\rm Pl})\nabla\phi$30–$F_5=-(\beta/M_{\rm Pl})\nabla\phi$31 AU, far below the observed $F_5=-(\beta/M_{\rm Pl})\nabla\phi$32, so the chameleon contribution is negligible (Anderson et al., 2012). Even the model-independent Cassini-based constraint $F_5=-(\beta/M_{\rm Pl})\nabla\phi$33 allows at most a few percent contribution, and the chameleon jerk is likewise suppressed (Anderson et al., 2012).

Strong-field asymmetric systems give far tighter bounds in some realizations. In chameleon $F_5=-(\beta/M_{\rm Pl})\nabla\phi$34 gravity, the scalar charges $F_5=-(\beta/M_{\rm Pl})\nabla\phi$35 of neutron stars and white dwarfs modify both conservative and dissipative post-Keplerian observables, with dipole radiation making the orbital evolution typically faster than in GR for NS–WD binaries (Zhang, 2022). A Monte-Carlo analysis of three binaries yields

$F_5=-(\beta/M_{\rm Pl})\nabla\phi$36

at $F_5=-(\beta/M_{\rm Pl})\nabla\phi$37 CL, corresponding to $F_5=-(\beta/M_{\rm Pl})\nabla\phi$38, with PSR J1738+0333 identified as the most constraining test in that sample (Zhang, 2022).

6. Early-Universe behavior, constraints, and theoretical status

A longstanding issue for chameleon gravity is its behavior in the early Universe. For the usual runaway potentials, the scalar is very light when the trace of the stress-energy tensor is nearly zero, so Hubble friction keeps it away from the minimum of the effective potential. When particle species become nonrelativistic, the quantity $F_5=-(\beta/M_{\rm Pl})\nabla\phi$39 becomes temporarily nonzero and gives a sequence of “kicks” to the field (Erickcek et al., 2013). In that analysis, the velocity imparted by the kicks is large enough that the chameleon’s mass changes nonadiabatically as it passes the minimum, producing very high-energy perturbations through quantum particle production. Modes with momenta exceeding $F_5=-(\beta/M_{\rm Pl})\nabla\phi$40 can only be avoided for small couplings and finely tuned initial conditions, and for generic $F_5=-(\beta/M_{\rm Pl})\nabla\phi$41 the theory cannot be treated as a classical field theory at the time of Big Bang Nucleosynthesis (Erickcek et al., 2013).

The quartic model was introduced specifically to evade this problem. With $F_5=-(\beta/M_{\rm Pl})\nabla\phi$42, the amplitude of oscillations scales as $F_5=-(\beta/M_{\rm Pl})\nabla\phi$43, the effective mass scales as $F_5=-(\beta/M_{\rm Pl})\nabla\phi$44, and the largest excited momenta remain of order a few times the ambient temperature rather than becoming trans-Planckian at MeV temperatures (Miller et al., 2016). This suggests that not all chameleon models share the same early-Universe pathologies; the viability depends on the detailed form of the potential.

Across observational probes, the overall status is mixed but restrictive. A broad review concluded that commonly studied chameleon models are well-constrained, less commonly studied models have large regions of parameter space that are still viable, and the simplest Hu–Sawicki $F_5=-(\beta/M_{\rm Pl})\nabla\phi$45 models are well constrained by astrophysical probes (Burrage et al., 2017). The same review emphasized that laboratory bounds primarily exclude large couplings, while astrophysical bounds exclude small background field values for $F_5=-(\beta/M_{\rm Pl})\nabla\phi$46, making comparison across experiments easiest in a unified parametrization using $F_5=-(\beta/M_{\rm Pl})\nabla\phi$47 and $F_5=-(\beta/M_{\rm Pl})\nabla\phi$48 (Burrage et al., 2017). A plausible implication is that the remaining viable parameter space is increasingly model-dependent, with the sharpest tests coming from probes that exploit differential screening rather than simple fifth-force searches.

The contemporary literature also shows that “chameleon gravity” is no longer a single phenomenological program. It includes scalar–tensor dark-energy models, $F_5=-(\beta/M_{\rm Pl})\nabla\phi$49 realizations, nonlinear structure prescriptions, strong-field timing tests, geodetic proposals, and even attempts to reproduce dark-matter-like halo phenomenology (Zaregonbadi et al., 18 Jul 2025). What unifies these constructions is not a single potential or coupling, but the environmental suppression of the scalar interaction through a density-dependent effective mass and the associated thin-shell behavior.