---
title: Coupled Quintessence Dark Energy
url: https://www.emergentmind.com/topics/coupled-quintessence-dark-energy-model
type: topic
---

# Coupled Quintessence Dark Energy

Coupled quintessence is a class of interacting dark-energy models in which the late-time acceleration of the Universe is driven by a canonical scalar field \(\phi\), while cold dark matter (CDM) exchanges energy-momentum with that field through a non-gravitational interaction. In the standard Einstein-frame formulation, baryons and radiation are left uncoupled, the scalar sector has \(\rho_\phi=\dot\phi^2/2+V(\phi)\) and \(p_\phi=\dot\phi^2/2-V(\phi)\), and the dark sector obeys \(\dot\rho_c+3H\rho_c=-Q\) and \(\dot\rho_\phi+3H(\rho_\phi+p_\phi)=Q\), so total energy-momentum is conserved while the individual dark components are not [2112.07608][2606.22023]. The model generalizes uncoupled quintessence by replacing separate conservation with a coupling term \(Q\), or equivalently by allowing the dark-matter mass or effective metric to depend on \(\phi\), and it has been studied as a framework for scaling behavior, coincidence-problem alleviation, effective phantom-divide crossing, and modified structure growth [1009.5263][1109.2189].

## 1. Field-theoretic structure and background equations

A common formulation starts from Einstein gravity plus a canonical scalar with potential \(V(\phi)\), with CDM coupled either through a field-dependent mass term or through a conformal factor in the CDM action. Representative actions include \(S_c[\mathcal A^2(\phi)g_{\mu\nu};\psi_c]\) for conformally coupled CDM and \(m_{dm}(\phi)\bar\Psi\Psi\) for field-dependent dark-matter mass, while baryons remain minimally coupled to \(g_{\mu\nu}\) in order to evade local gravity constraints [2606.22023][1009.5263]. In this sense, coupled quintessence is frequently described as an interacting dark-sector model rather than a modification of the visible sector.

At the homogeneous level, the coupling is encoded in a source term \(Q\). For the generalized model with exponential scalar potential \(V(\phi)=V_a e^{-\mu\kappa\phi}\), the equations are
\[
\ddot{\phi}+3H\dot{\phi}+\frac{\partial V}{\partial\phi}=\frac{Q}{\dot{\phi}},\qquad
\dot{\rho}_{DM}+3H\rho_{DM}=-Q,
\]
\[
\dot{H}=-\frac{\kappa^2}{2}\left(\dot{\phi}^2+\rho_{DM}\right),\qquad
H^2=\frac{\kappa^2}{3}\left(\frac{\dot{\phi}^2}{2}+V+\rho_{DM}\right),
\]
with interaction \(Q=\kappa\beta\rho_{DM}C(\phi)\dot{\phi}\) in one widely used class [2112.07608]. Closely related conformal models give
\[
Q=\beta\,\rho_c\,\frac{\dot\phi}{M_{\rm Pl}},
\]
with \(\rho_c=\rho_{c0}a^{-3}\mathcal A(\phi)/\mathcal A(\phi_0)\), so the usual CDM dilution law is explicitly modified [2606.22023]. This change in the background matter evolution is one of the defining signatures of the framework.

Coupled quintessence also has a well-known scalar-tensor interpretation. In Brans–Dicke-inspired constructions, the Einstein-frame theory is equivalent to a quintessence field conformally coupled to dark matter, with interaction \(Q=\beta\rho_c\dot\varphi\), unstable matter domination, a dark-energy-dominated point, and, for one branch, a mixed scaling solution relevant to the coincidence problem [1109.2189]. This establishes coupled quintessence as both a phenomenological dark-sector model and an Einstein-frame description of broader scalar-tensor theories.

## 2. Interaction functions and model families

The defining freedom of the subject lies in the choice of coupling function. The standard constant-coupling model corresponds to \(C(\phi)=1\), giving \(Q=\kappa\beta\rho_{DM}\dot\phi\), but many generalizations have been studied in order to alter the asymptotics, improve the matter era, or generate effective late-time dynamics beyond those of the constant-coupling case [2112.07608].

Several representative model classes appear repeatedly in the literature.

| Interaction class | Representative form | Characteristic feature |
|---|---|---|
| Constant conformal coupling | \(Q=\kappa\beta\rho_{DM}\dot\phi\) | Standard coupled quintessence |
| Generalized field dependence | \(Q=\kappa\beta\rho_{DM}C(\phi)\dot\phi\) | Perturbed or unbounded coupling |
| Scale-factor parametrization | \(\rho_{dm}=\rho_{dm,0}a^{-3+\epsilon(a)}\) | Time-dependent phenomenological interaction |
| Dissipative coupling | \(Q=\Gamma\dot\phi^2\) | Warm-inflation-inspired energy transfer |
| Kinetic coupling | \(f(X)=\left(M_{\rm Pl}^{-4}X\right)^\alpha\) | Coupling through the scalar kinetic term |
| Curvature-modulated coupling | \(Q_0=\alpha\kappa\rho_m\dot\phi\left[1-\beta R^2/(6H^2)\right]\) | Late-time activation by curvature suppression |

In the generalized interaction model, two explicit choices are emphasized: a small perturbation around a constant,
\[
C(\phi)=1+\epsilon \sin(\alpha\kappa\phi),
\]
and an unbounded form,
\[
C(\phi)=(\kappa\phi)^n,\qquad n=1,2,3,\dots
\]
[2112.07608]. In the first case, the familiar scaling accelerated solutions survive, although perturbed; in the second, true scaling solutions disappear. A distinct phenomenological route replaces the field dependence by a scale-factor function \(\epsilon(a)\), with parameterizations \(\epsilon(a)=\epsilon_0 a^\xi\) and \(\epsilon(a)=\epsilon_0\exp(1-a^{-1})\), so that the interaction can vary explicitly over cosmic time [1009.3841].

Other constructions encode the interaction more directly in the CDM sector. In the “higher-order” model, the dark-matter mass is chosen as
\[
m_{dm}=m_0\exp\!\left[\left(\frac{V(\phi)}{\rho_{cr}^0}\right)^n\right],
\]
which yields a coupling \(\beta(\phi)\propto V(\phi)^n\) and a late-time accelerated attractor, while large \(n\) suppresses the coupling and moves the theory toward uncoupled quintessence [1009.5263]. In the linear-in-e-folds parametrization, \(\phi-\phi_0=\lambda N\) with \(N=\ln a\), the coupling is a constant conformal coupling \(\beta\), but the ansatz reconstructs an explicit potential
\[
V(\phi)=A e^{\left(-\frac{3}{\lambda}+\beta\right)\phi} + B e^{-\lambda\phi} + C e^{-\frac{3}{\lambda}\phi},
\]
giving analytic control over the background evolution [2104.14889].

The literature has also extended coupled quintessence beyond energy exchange of the standard conformal type. Kinetic couplings link CDM to the scalar kinetic term \(X\), momentum couplings modify the Euler equation without altering the CDM continuity equation, and scalar couplings to \(p\)-form kinetic sectors generate anisotropic dark-energy dynamics rather than an isotropic perfect-fluid limit [2207.13682][2309.04530][1910.10499]. These variants retain the essential idea of a canonical quintessence field but enlarge the phenomenology of the dark sector.

## 3. Dynamical systems, scaling solutions, and cosmic sequence

A central question in coupled quintessence is whether the model yields a realistic cosmic sequence: a matter-dominated era long enough for structure formation, followed by late-time accelerated expansion. Dynamical-systems analyses have made this issue particularly precise. For the standard exponential-potential model with constant coupling, the usual dimensionless variables
\[
x=\frac{\kappa \dot{\phi}}{\sqrt6 H},\qquad
y=\frac{\kappa\sqrt{V}}{\sqrt3 H},
\]
give the Friedmann constraint \(x^2+y^2+\Omega_{DM}=1\) and effective equation of state \(w_{\rm eff}=x^2-y^2\) [2112.07608]. In that standard case, the system admits scaling critical points, including an accelerating one, but the key drawback recalled in the literature is that these accelerated scaling solutions are not preceded by a sufficiently long matter-dominated era.

For generalized couplings, the phase-space structure changes qualitatively. When \(C(\phi)=(\kappa\phi)^n\), the compactified variable \(z=\arctan(\kappa\phi)\) is introduced, and the autonomous system contains a matter point \(C(0,0,0)\) with \(\Omega_{DM}=1\), \(\Omega_\phi=0\), \(w_{\rm eff}=0\), together with critical lines
\[
A(x,\sqrt{1-x^2},-\pi/2),\qquad
B(x,\sqrt{1-x^2},\pi/2),
\]
which are scalar-field dominated [2112.07608]. The striking result is that there are no true scaling solutions for this unbounded coupling. For \(\beta>0\), the late-time attractor is a dark-energy-dominated state on the \(B\) line, accelerating when \(\mu<\sqrt2\), with
\[
w_{\rm eff}=-1+\frac{\mu^2}{3}.
\]
For \(\beta<0\), the attractor becomes \(B(0,1,\pi/2)\), a pure dark-energy state with \(w_{\rm eff}=-1\), but the approach is extremely slow. During that long transient, \(\Omega_{DM}/\Omega_\phi\sim\mathcal O(1)\), so the evolution mimics an accelerated scaling solution without possessing a true scaling fixed point.

A similar asymptotic structure appears in the warm-inflation-inspired model with \(Q=\Gamma\dot\phi^2\). There the autonomous system has the standard matter point \(C(0,0,0)\), an unstable matter-like scaling point \(E\), and new critical lines at \(z=1\), but the attractors lie on those lines rather than on a stable mixed scaling point [2312.09171]. For \(\alpha<0\), the system can spend a long time in a scaling-type regime with appreciable \(\Omega_{DM}/\Omega_\phi\), yet asymptotically it still approaches dark-energy domination. This reinforces a recurring conclusion in the subject: long-lived quasi-scaling behavior need not correspond to a true asymptotic scaling attractor.

Other parametrizations recover scaling more explicitly. In the linear-in-e-folds model, the field tracks radiation at early times, with \(w_\phi=1/3\), enters a matter-era scaling attractor with
\[
w_\phi=-\frac{\beta}{\lambda+\beta},\qquad
\Omega_\phi=\frac{\lambda}{3}(\beta+\lambda),
\]
and then freezes at late times as the shallower exponential term dominates [2104.14889]. This suggests that coupled quintessence can realize the radiation-to-matter-to-acceleration sequence either through true scaling phases or through slow transient regimes, depending on the interaction function.

## 4. Effective equation of state and phantom-divide crossing

The intrinsic scalar field in coupled quintessence is canonical, so its equation of state obeys \(w_\phi\ge -1\). Nevertheless, several coupled models allow the effective dark-energy equation of state inferred from the background expansion to cross the phantom divide \(w=-1\) without introducing a wrong-sign kinetic term [2503.10806][2606.22023]. This distinction between the physical scalar equation of state and the observationally reconstructed effective equation of state is now one of the most important conceptual points in the subject.

In Yukawa-type coupled models, dark matter has a field-dependent mass through
\[
{\mathcal L}_{\rm int}=-f(\phi)\bar\psi\psi,\qquad
f(\phi)=\exp\!\left(\frac{\beta\phi}{\sqrt{8\pi}\,M_{\rm Pl}}\right),
\]
so that \(\rho_{\rm DM}\sim f(\phi)/a^3\) rather than \(a^{-3}\) [2503.10806]. In that framework,
\[
w_{\rm eff}=\frac{w_\phi}{1-x},\qquad
x=-\frac{\rho_{\rm DM}^0}{a^3\rho_\phi}\left(\frac{f(\phi)}{f_0}-1\right).
\]
If \(x>0\), then \(w_{\rm eff}\) is shifted below \(w_\phi\); thus \(w_{\rm eff}\) can cross \(-1\) even though \(w_\phi>-1\) at all times. With suitable parameter choices, the model crosses the phantom divide around \(z\sim 0.3\) to \(0.5\), in agreement with the DESI-motivated trend discussed in that work.

The sign of the coupling can also change dynamically. In the exponential-potential, exponential-coupling CQ-EXP model, the branch \(\beta<0\) allows the interaction force to oppose the potential force at early times, so the scalar can be driven up its potential and later roll back down, making \(\dot\phi\) change sign [2606.22023]. Since \(Q\propto\beta\rho_c\dot\phi\), the energy transfer reverses direction during evolution, and the effective dark-energy equation of state can cross \(w=-1\). By contrast, for \(\beta>0\), the scalar rolls monotonically, \(Q>0\) throughout, and the effective equation of state remains above \(-1\).

A different realization of sign switching appears in the \(\beta\varphi\)CDM model with
\[
\mathcal A(\varphi)=1+\frac{1}{2}\beta\varphi^2,\qquad
V(\varphi)=V_0 e^{\alpha\varphi}.
\]
Because \(\mathcal A'(\varphi)=\beta\varphi\), the coupling term changes sign when \(\varphi\) crosses zero, generating an early phase in which energy flows from dark matter to dark energy and a later phase in which it flows from dark energy to dark matter [2604.02204]. The scalar itself remains canonical, with \(w_\varphi>-1\), but the modified CDM evolution produces an effective \(w_{\rm DE}^{\rm eff}\) that can cross the phantom divide.

SUGRA-based coupled quintessence yields an analogous mechanism. With
\[
V(\phi)=V_0\left(\frac{\phi}{M_{\rm Pl}}\right)^{-\alpha}\exp\!\left(\frac{\phi^2}{2M_{\rm Pl}^2}\right),
\]
the field can begin near the potential minimum; coupling to CDM then displaces it, induces a turnaround, changes the sign of \(Q\propto\beta\dot\phi\), and produces \(w_{\rm DE,eff}\) crossing from below to above \(-1\) [2605.17754]. This shows that phantom-divide crossing in coupled quintessence is not a unique consequence of one coupling ansatz, but a general effect of interaction-induced modification of the inferred dark-energy sector.

Not all coupled models predict eternal acceleration. In the phenomenological \(\epsilon(a)\) class, positive late-time coupling can drain enough energy from dark energy that the present accelerated epoch is only temporary, after which the Universe re-enters a future CDM-dominated era [1009.3841]. This is qualitatively different from standard \(\Lambda\)CDM and from the more common coupled-quintessence asymptotics in which dark-energy domination persists.

## 5. Perturbations, fifth forces, and structure formation

Coupled quintessence affects not only the homogeneous background but also linear and nonlinear structure formation. In the simplest conformal models, the matter-density contrast equation acquires both modified friction and modified source terms. One representative form is
\[
\delta_c''= -\left(\frac12-\frac12\Omega_r-\frac32(\gamma-1)\Omega_\phi -m\sqrt{3\gamma\Omega_\phi}\right)\delta_c'
+\frac32\left(1-\Omega_\phi(1+\gamma)-\Omega_r\right)(1+2m^2)\delta_c,
\]
which makes explicit that the coupling changes both the damping term and the effective gravitational strength [2307.03740]. In related treatments, the fifth-force interpretation is summarized by
\[
G_{\rm eff}=G(1+2\beta^2),
\]
so the coupling enhances or suppresses CDM growth depending on the sign and the background scalar evolution [1103.0694].

Kinetic couplings alter perturbations in a more distinctive way because the interaction depends on the scalar kinetic term rather than on \(\phi\) itself. For \(f(X)=\left(M_{\rm Pl}^{-4}X\right)^\alpha\), the model yields early-time scaling behavior, modified CDM continuity and Euler equations, and effective friction and gravitational couplings in the sub-horizon growth equations [2207.13682]. The paper reports measurable signatures in the CMB TT spectrum, the lensing potential auto-correlation, and the matter power spectrum; observationally, however, the coupling parameter is constrained to be of order \(10^{-4}\), and model selection does not reveal a statistical preference over \(\Lambda\)CDM.

At the nonlinear level, coupled quintessence modifies halo abundance and halo structure. Semi-analytic halo-mass-function calculations show that the effect depends strongly on the potential and on the direction of energy transfer. For the double-exponential potential, a weak positive coupling can enhance the abundance of very massive high-redshift halos, while Albrecht–Skordis and SUGRA models generally predict fewer such objects than \(\Lambda\)CDM [1103.0694]. The same study compares spherical-collapse and Newtonian prescriptions for the collapse threshold \(\delta_\star(z)\) and finds that they can differ substantially, especially at high redshift and strong coupling, although this uncertainty is secondary to parameter uncertainties and fitting-function uncertainties.

Momentum-exchange models provide yet another perturbative channel. In the Type-3 momentum coupling, the CDM continuity equation remains standard, but the Euler equation becomes
\[
(1+h_1)\dot{v}_i + (1+h_2)\mathcal{H}v_i + (1+h_3)\nabla_i\Psi = 0,
\]
so the coupling acts through a modified friction term and a modified effective gravitational force [2309.04530]. N-body simulations of this scenario find that, for certain quintessence potentials, a positive coupling can reduce small-scale structure, enhance somewhat the large-scale power, lower halo central densities, and increase halo velocity dispersions. This marks an important distinction between energy-exchange couplings and pure momentum-exchange couplings: the former alter the background matter density itself, whereas the latter leave the continuity equation unchanged and act directly on the dark-matter trajectories.

The literature has also extended coupled-quintessence phenomenology to galactic dynamical friction. In the uncoupled limit, standard quintessence does not generate dynamical friction at galactic scales, but once the scalar couples nonminimally to matter, the scalar-mediated fifth force produces a modified wake and a nonzero dynamical-friction force [2502.10676]. The Fornax application in that analysis shows that galactic-scale fits can imply significant departures from standard CDM, whereas cosmological limits on the coupling make the effect negligible. This illustrates a broader tension between local and cosmological constraints on dark-sector interactions.

## 6. Observational constraints, degeneracies, and current status

A persistent result of observational work is that coupled quintessence is difficult to infer uniquely from background data because the dark-energy equation of state and the interaction are degenerate in the expansion history. A model-independent reconstruction based on Gaussian processes shows that, for a fixed \(H(z)\), one cannot independently determine both \(w(z)\) and the interaction \(f(z)\) from background data alone; a prior on one is needed to reconstruct the other [2006.14511]. Using cosmic chronometers, BAO, Type Ia supernovae, and Ly\(\alpha\) BAO, that analysis finds only a mild \(1\sigma\) hint of nonzero interaction at low redshift and a \(2\sigma\) hint at high redshift, with the latter driven mainly by the Ly\(\alpha\) BAO point at \(z=2.34\). The same work notes that allowing coupling weakens the reconstructed Swampland bound \(C1\) at \(z=0\) from \(C1\lesssim 1.23\) in the uncoupled case to \(C1\lesssim 4.44\) at \(95\%\) CL.

Fits to specific coupled-quintessence parametrizations generally favor small couplings, but the preferred sign can be dataset dependent. In the linear-in-e-folds model, Planck 2018 slightly prefers \(\beta>0\), corresponding to energy transfer from dark energy to dark matter, whereas KiDS-450 slightly prefers \(\beta<0\), corresponding to the opposite direction [2104.14889]. In a broader 2023 observational analysis with CMB distance priors, BAO, Pantheon+, cosmic chronometers, megamasers, growth data, and SH0ES, the preferred interaction remains very small, \(m_i\sim 10^{-3}\), and the model behaves close to uncoupled quintessence in practice; AIC and BIC favor the coupled model over \(\Lambda\)CDM, but the reduced-\(\chi^2\) criterion favors the simpler model [2307.03740]. This divergence between information criteria and reduced-\(\chi^2\) is an early example of a recurring model-selection ambiguity in the literature.

DESI-era analyses have sharpened attention on effective phantom-divide crossing. A 2025 Yukawa-type study shows that a canonical coupled scalar with \(w_\phi>-1\) can still reproduce the DESI-preferred trend in \(w_{\rm eff}(z)\), typically crossing near \(z\sim0.3\) to \(0.5\), with several parameter choices lying inside DESI or DESI+Union3+Planck \(2\sigma\) contours [2503.10806]. A 2026 CQ-EXP likelihood analysis with Planck CMB, DESI BAO, and DES-Dovekie supernovae reports that, when \(\sum m_\nu=0.06\) eV is fixed, both \(\beta>0\) and \(\beta<0\) branches prefer nonzero coupling, with the \(\beta<0\) branch favored more strongly and separated from zero by more than \(4\sigma\); however, allowing a free signed effective neutrino-mass parameter weakens the coupling preference and makes the \(\beta>0\) and \(\beta<0\) branches nearly identical in \(\chi^2_{\min}\), leaving CQ-EXP statistically indistinguishable from \(w_0w_a\)CDM with the data considered there [2606.22023]. This directly exhibits the degeneracy between dark-sector coupling and late-time neutrino-sector freedom.

The strongest DESI-era claim in the supplied literature comes from coupled SUGRA quintessence. Using DESI BAO, DES-Dovekie SNIa, and Planck CMB data, that analysis finds the coupling parameter differs from zero by more than \(4\sigma\), and the preferred branch is the one in which the interaction changes sign and the effective equation of state crosses the phantom divide [2605.17754]. Yet even there the best-fit coupled SUGRA branch is described as statistically indistinguishable from the CPL parametrization, with only a very small difference in \(\chi^2_{\min}\). A plausible implication is that present data may be more sensitive to the broad phenomenology of evolving dark energy than to the specific microphysical origin of that evolution.

Model-selection controversies remain active. In the 2025 Planck+DESI DR2 study of coupled quintessence with an additional warm-dark-matter sector, Deviance Information Criterion strongly prefers the quintessence model coupled with warm dark matter, while Bayesian evidence favors \(\Lambda\)CDM for all dataset combinations considered [2509.09624]. The literature therefore supports two simultaneous conclusions: coupled quintessence is observationally viable and can reproduce several late-time signatures now under discussion, but the statistical evidence for a nonzero dark-sector coupling is highly sensitive to parametrization, auxiliary sector freedom, and the model-selection metric itself.

Source: https://www.emergentmind.com/topics/coupled-quintessence-dark-energy-model