🔭 astrophysics

Constraints on Horndeski Gravity with Phantom Crossing

This paper introduces the Asymptotic Cubic Galileon (ACG) subclass of Horndeski gravity models, demonstrating that they successfully fit key cosmological observations while enabling a phantom-crossing dark energy equation of state, thereby providing a robust Lagrangian foundation for the observationally preferred $w_0w_a$ CDM behavior.

Original authors: Krishna Naidoo, James Hallam, Tessa Baker, Sergi Sirera

Published 2026-06-23

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Krishna Naidoo, James Hallam, Tessa Baker, Sergi Sirera

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine the universe as a giant, expanding balloon. For decades, scientists have been trying to figure out what's inside that balloon pushing it to expand faster and faster. The standard story, called $\Lambda$ CDM, says there is a mysterious, unchanging force called "Dark Energy" (represented by the Greek letter Lambda, $\Lambda$ ) doing the pushing. It's like a constant, invisible wind that never changes its strength.

However, recent measurements of the universe's history suggest this "constant wind" story might be wrong. The data hints that the wind is actually changing speed. In fact, it seems the wind was blowing too hard in the past (a state physicists call "phantom," where the pressure is more negative than physics usually allows) and has since slowed down, crossing a critical threshold to become more normal.

This paper introduces a new, more sophisticated story to explain this changing wind, using a framework called Horndeski Gravity. Here is how the authors break it down:

1. The Problem with the Old "Phantom" Models

In the past, scientists tried to explain this "phantom" behavior using models based on a theory called the Cubic Galileon. Think of this like a simple machine with a single gear. It worked well for the early universe but had a fatal flaw: it couldn't naturally slow down and cross that critical threshold to match what we see today. It was like a car that could only drive in reverse and couldn't switch to forward gear.

2. The New Solution: The "Asymptotic Cubic Galileon" (ACG)

The authors propose a new version of this machine, which they call the Asymptotic Cubic Galileon (ACG).

The Analogy: Imagine the old machine had a rigid gear. The new ACG machine has a smart, adjustable gear.
How it works: In the early universe (when the balloon was small), this new gear acts just like the old, simple one. It creates the "phantom" behavior (the super-strong wind) that the data likes.
The Twist: As the universe expands and time passes, the gear slowly changes its shape (the authors call this "breaking shift symmetry"). This change allows the machine to naturally slow down and cross the critical threshold, matching the "phantom crossing" we see in the data.

They tested two specific versions of this smart gear:

Growing G( $\phi$ ): The "braiding" part of the machine gets stronger over time.
Decaying K( $\phi$ ): The "kinetic" part of the machine gets weaker over time.

Both versions successfully mimic the changing wind of the universe without breaking the laws of physics.

3. Testing the Theory

The authors didn't just build the theory; they tested it against the biggest datasets we have:

The Baby Picture (CMB): The Cosmic Microwave Background, which is the afterglow of the Big Bang.
The Ruler (BAO): Baryon Acoustic Oscillations, which act like a cosmic ruler to measure distances.
The Distance Markers (Supernovae): Exploding stars used to measure how far away things are.

The Result: The new ACG models fit the data just as well as the popular "changing wind" models (called $w_0w_a$ CDM) and much better than the old, unchanging "constant wind" ( $\Lambda$ CDM) model.

4. The "Ghost" in the Machine (ISW Effect)

There is a catch. When you change how gravity works, it can create weird ripples in the fabric of space-time that we can detect by looking at how galaxies and the Cosmic Microwave Background interact. This is called the Integrated Sachs-Wolfe (ISW) effect.

The Constraint: The authors found that if their new machine changes too drastically, it predicts a "negative" interaction that we don't see in the real universe.
The Fix: By adding a rule that says "the interaction must be positive," they forced their models to be more gentle. This ruled out the most extreme versions of their theory and narrowed down exactly how the "smart gear" must behave. It turns out the gear must change very gradually.

5. Why This Matters

The paper makes a crucial distinction between phenomenology (just describing what happens) and fundamental theory (explaining why it happens).

The popular "changing wind" models ( $w_0w_a$ CDM) are like a weather report: they describe the wind changing, but they don't explain the engine driving it.
The new ACG models provide the engine. They show that this changing wind can be the result of a specific, robust set of physical laws (a Lagrangian formulation).

6. What They Found (and Didn't Find)

Neutrinos: There is a known tension in physics where data sometimes suggests neutrinos (tiny particles) have a "negative mass," which is impossible. The authors checked if their new models fix this. They found that, unlike some other theories, their models do not fix the negative mass problem. The universe still seems to prefer a tiny bit of negative mass in the data, but the evidence isn't strong enough to be certain yet.
Voids: They also looked at empty spaces in the universe (voids). In some extreme versions of their theory, the math breaks down in these empty spaces (like a car engine stalling in a vacuum). However, the "gentle" versions of their theory (the ones that pass the ISW test) avoid these breakdowns.

Summary

The authors have built a new, mathematically robust engine (the ACG model) that explains why the universe's expansion seems to be changing speed. It fits the data as well as the best current theories but offers a deeper, more fundamental explanation rooted in the laws of gravity. It's a step toward understanding that the "Dark Energy" pushing our universe apart might not be a constant, but a dynamic force that evolves over time.

Technical Summary: Constraints on Horndeski Gravity with Phantom Crossing

Problem Statement
Recent cosmological observations, including data from the Dark Energy Spectroscopic Instrument (DESI), the Dark Energy Survey (DES), and the Planck cosmic microwave background (CMB), indicate a preference for dynamical dark energy (DE) where the equation of state (EoS), $w$ , crosses the phantom divide ( $w = -1$ ) at a redshift of $z \approx 0.5$ . While phenomenological models like the Chevallier-Polarski-Linder (CPL) parameterization can describe this behavior, they lack a fundamental Lagrangian formulation to predict impacts on other observables such as structure growth and weak lensing. Furthermore, standard Horndeski scalar-tensor theories with non-minimal coupling or specific shift-symmetric potentials often struggle to accommodate phantom crossing without conflicting with large-scale structure (LSS) observations, particularly the galaxy-Integrated Sachs-Wolfe (ISW) cross-correlation, which requires a positive signal.

Methodology
The authors introduce a new subclass of minimally coupled Horndeski models termed Asymptotic Cubic Galileon (ACG) models. These models are designed to replicate the Cubic Galileon behavior at early times (ensuring consistency with CMB physics) but diverge at late times to allow for phantom crossing.

Theoretical Framework: The ACG models break shift symmetry by introducing $\phi$ -dependence into either the kinetic term $K(\phi, X)$ or the kinetic-braiding term $G_3(\phi, X)$ , rather than modifying the potential. Two specific realizations are analyzed:
1. Growing $G(\phi)$ : A linearly growing kinetic-braiding term ( $G(\phi) = 1 + c_{g3}\phi$ ).
2. Decaying $K(\phi)$ : An exponentially decaying kinetic term ( $K(\phi) = \exp(-c_k\phi)$ ).
  Both models satisfy the luminal speed of gravitational waves constraint ( $c_T = c$ ) and are required to be ghost-free and gradient-stable.
Numerical Implementation: The analysis utilizes the Hi-COLA software package to solve the background evolution of the scalar field and expansion history. Linear perturbations are computed using the quasistatic approximation to derive the modified Poisson equation ( $\mu$ ) and lensing potential ( $\Sigma$ ).
Observational Constraints: The models are constrained using a custom Markov Chain Monte Carlo (MCMC) sampler comparing predictions against:
- Planck: A compressed likelihood focusing on the acoustic scale ( $\theta^*$ ), baryon density ( $\omega_b$ ), and matter density ( $\omega_{bc}$ ).
- DESI DR2: Baryon Acoustic Oscillation (BAO) measurements across multiple redshift bins.
- DES-Dovekie: Type Ia supernova distance modulus measurements.
- ISW Prior: A critical constraint is imposed to ensure the ISW-galaxy cross-correlation remains positive, effectively ruling out models that predict a sign flip in the time derivative of the lensing potential.
Model Comparison: The authors employ frequentist ( $\Delta\chi^2_{MAP}$ ) and Bayesian (evidence ratios, $\Delta \ln Z$ ) metrics to compare ACG models against the standard $\Lambda$ CDM and the phenomenological $w_0w_a$ CDM models.

Key Results

Phantom Crossing Viability: The ACG models successfully generate a phantom crossing at $z \sim 0.7-1.0$ (slightly earlier than the $z \sim 0.5$ preferred by pure phenomenological fits) while maintaining a positive ISW signal. The positive ISW prior significantly restricts the parameter space, favoring models with moderate scalar field contributions ( $f_\phi \approx 0.5$ ) and shallower EoS evolution.
Parameter Constraints:
- The Growing $G(\phi)$ and Decaying $K(\phi)$ models show consistent constraints on the Hubble constant ( $H_0$ ) but prefer slightly higher baryon density and lower CDM density compared to $\Lambda$ CDM.
- The models are consistent with DESI BAO and DES supernova data, predicting a suppression in the volume-averaged BAO distance at low redshifts compared to $\Lambda$ CDM.
Model Comparison Metrics:
- Both ACG and $w_0w_a$ CDM models provide a better fit to the data than $\Lambda$ CDM, with $\Delta\chi^2_{MAP} \approx -11$ for ACG and $\approx -13$ for $w_0w_a$ CDM.
- In terms of Bayesian evidence, ACG models show a modest preference over $\Lambda$ CDM ( $\Delta \ln Z \approx 2.09$ with ISW prior), comparable to or slightly better than $w_0w_a$ CDM. The authors note that while $w_0w_a$ CDM fits the data slightly better, it is penalized for being less predictive (broader parameter volume) compared to the Lagrangian-constrained ACG models.
Structure Growth and Screening:
- The growth of structure ( $f\sigma_8$ ) predicted by ACG models is broadly consistent with $\Lambda$ CDM and observational data, showing only a slight preference for higher amplitude at $z < 0.5$ .
- Analysis of the Vainshtein screening factor ( $X/\delta$ ) reveals that both ACG models may exhibit pathological solutions (imaginary screening factors) in voids ( $\delta < 0$ ) at low redshift. The authors suggest that adopting more gradual functional forms for $K(\phi)$ or $G(\phi)$ could alleviate this issue.
Neutrino Mass Tension: Unlike dynamical DE models which can alleviate the preference for negative effective neutrino mass found in $\Lambda$ CDM, ACG models retain a preference for negative neutrino masses. This is attributed to the lack of an early phantom phase in ACG models (where $w \to -1$ at high $z$ ). However, the significance of this preference is low ( $<1\sigma$ ), and Bayesian evidence favors fixing the neutrino mass to the standard value ($0.06$ eV) for these models.

Significance and Claims
The paper claims that the ACG models provide a robust Lagrangian formulation for the observationally preferred dynamical DE behavior (phantom crossing) without resorting to non-minimal coupling or complex potentials. By grounding the phenomenological $w_0w_a$ CDM behavior in a specific scalar-tensor theory, the models enable predictions for nonlinear scales and other observables that purely phenomenological models cannot.

The authors emphasize that the ISW constraint is a powerful discriminator, confining viable ACG models to a well-defined region of the Horndeski landscape characterized by gradual evolution and shallow phantom behavior. While the models currently face potential pathologies in voids under the Vainshtein-only approximation, the work establishes a foundation for further testing these theories against Stage IV survey data (e.g., Euclid, DESI) and nonlinear simulations. The study concludes that while ACG models are not definitively favored over $w_0w_a$ CDM by current data, they offer a theoretically motivated alternative that is consistent with current expansion history and perturbation constraints.