🔭 astrophysics

A window for water-hydrogen demixing on warm metal-rich sub-Neptunes

This paper introduces the ATHENAIA framework to demonstrate that water-hydrogen demixing can occur in warm, metal-rich sub-Neptunes like TOI-270 d, challenging the assumption of fully miscible envelopes and suggesting that current models may underestimate bulk metallicities and the prevalence of molten mantle conditions.

Original authors: Caroline Piaulet-Ghorayeb, Daniel P. Thorngren, Eliza M. -R. Kempton, Justin Lipper, Leslie Rogers, Fernanda Correa Horta, Shi Lin Sun

Published 2026-03-17

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Caroline Piaulet-Ghorayeb, Daniel P. Thorngren, Eliza M. -R. Kempton, Justin Lipper, Leslie Rogers, Fernanda Correa Horta, Shi Lin Sun

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 a sub-Neptune planet as a giant, fluffy cloud of gas and water vapor swirling around a rocky core. For years, scientists believed that on "warm" sub-Neptunes (planets closer to their stars than Uranus or Neptune), the heat would keep everything mixed together like a perfectly blended smoothie. They thought the hydrogen gas and water vapor were so hot and energetic that they couldn't separate; they would stay as one uniform soup from the top of the atmosphere all the way down to the rocky ground.

This paper says: "Not so fast."

The authors, led by Caroline Piaulet-Ghorayeb, have built a new super-computer model called ATHENAIA (a fancy name for a detective tool) to look inside these planets. They discovered that even on warm planets, the hydrogen and water might not be a smoothie at all. Instead, they might be like an oil-and-vinegar dressing that has separated into layers.

Here is the breakdown of their discovery using simple analogies:

1. The "Oil and Vinegar" Problem

In a salad dressing, oil and vinegar don't mix well; the oil floats on top, and the vinegar sinks. In the deep, high-pressure hearts of these planets, hydrogen and water act similarly under certain conditions.

The Old View: Scientists thought the heat of the star kept the "dressing" shaken up, so the top of the atmosphere (where we look with telescopes) represented the whole planet's recipe.
The New View: The authors found a "Demixing Window." This is a specific zone of temperature and pressure where the hydrogen and water decide to separate. The heavy, metal-rich water sinks deep down, while the lighter hydrogen stays on top.

2. The Case of TOI-270 d

The team used a specific planet, TOI-270 d, as their test case. It's a warm sub-Neptune about 380 degrees Kelvin (hot, but not scorching).

The Clue: When we look at this planet's atmosphere with the James Webb Space Telescope (JWST), we see it is very "metal-rich" (full of water vapor and heavy elements).
The Twist: The authors' model suggests that because the planet is so rich in heavy elements, the deep interior gets cold enough (relative to the pressure) for the water to sink.
The Result: The atmosphere we see is actually "lighter" than the planet's true interior. The heavy stuff has sunk to the bottom. If we assume the atmosphere represents the whole planet, we are underestimating how much water and heavy material the planet actually holds. It's like judging the richness of a cake by tasting only the frosting, while the dense, chocolatey cake underneath is hidden.

3. Why Does This Happen? (The Greenhouse vs. The Lapse Rate)

This is the tricky part, but here is the analogy:

The Greenhouse Effect: Imagine a thick blanket (the atmosphere) trapping heat. The more "heavy" elements (water, metals) you add to the blanket, the better it traps heat. This makes the top of the atmosphere very warm.
The Lapse Rate (The Slope): However, adding all that heavy stuff also changes how heat travels downward. It makes the temperature drop off much more slowly as you go deeper.
The Conflict: The "Demixing" happens when the deep interior is cool enough for the water to separate from the hydrogen. The authors found that the heavy blanket traps so much heat at the top, but the slope of temperature change is so gentle, that the deep interior ends up in the "Goldilocks zone" for separation. It's cool enough for the water to sink, but not so cold that it freezes into ice clouds.

4. The "Molten Rock" Surprise

The paper also looked at what happens at the bottom of the atmosphere, where it touches the rocky core (the mantle).

Old Idea: Scientists thought that if the atmosphere was heavy enough, the core would be solid rock.
New Idea: Because the heavy atmosphere traps so much heat (the greenhouse effect), the bottom of the atmosphere is actually much hotter than we thought. This means the rocky core might be molten (a magma ocean) even on planets we thought were cool enough to have solid rocks. It's like wearing a heavy winter coat that keeps you so warm you start sweating, even if the room feels cool.

Why Does This Matter?

This discovery changes how we understand the "family tree" of planets.

We've been underestimating the water: If the heavy stuff sinks, these warm planets might be much "wetter" and heavier than we thought.
We need new maps: We can no longer assume that what we see in the sky (the atmosphere) is the same as what is deep inside. We have to build new models that account for this "separation" layer.
Evolution: If the heavy stuff sinks, it changes how the planet cools down over billions of years, which affects how long it keeps its atmosphere.

In a nutshell:
Scientists used to think warm sub-Neptunes were like a well-mixed fruit punch. This paper shows they are more like a layered cocktail, where the heavy syrup has settled at the bottom. If we only taste the top layer, we miss the true flavor of the planet. This changes our understanding of what these planets are made of, how they formed, and whether their rocky cores are melting.

1. Problem Statement

Sub-Neptunes are the most common exoplanet demographic, yet their bulk compositions remain poorly constrained due to a degeneracy between envelope mass fraction and metallicity.

Current Paradigm: It was traditionally believed that "warm" sub-Neptunes (receiving high stellar irradiation, $T_{eq} \sim 300\text{--}450$ K) possess fully miscible envelopes where hydrogen and water are homogeneously mixed from the deep interior to the observable atmosphere. This assumption was based on the idea that high temperatures prevent water condensation and that supercritical water is miscible with hydrogen.
The Gap: Recent theoretical work and Density Functional Theory (DFT) simulations suggest a "solvus" (critical curve) exists in temperature-pressure space where hydrogen and water become immiscible, even at high pressures. While this was known to affect cold ice giants (Uranus/Neptune), it was assumed irrelevant for warm sub-Neptunes.
The Question: Do warm, metal-rich sub-Neptunes (like TOI-270 d) actually experience water-hydrogen demixing, leading to compositional gradients (stratification) where the deep envelope is more metal-rich than the upper atmosphere? If so, current models assuming full miscibility may significantly underestimate bulk envelope metallicities and core masses.

2. Methodology

The authors introduce ATHENAIA (ATmospHEre-iNterior bAyesian Inference frAmework), a coupled modeling framework designed to assess the thermodynamic stability of sub-Neptune envelopes against demixing.

Coupled Modeling:
- Interior Models: Uses 1D hydrostatic equilibrium models (based on Thorngren & Fortney 2019) with updated Equations of State (EOS) for H/He (Chabrier et al. 2019) and H $_2$ O (Mazevet et al. 2019). The interior is modeled as a convective adiabat extending from the mantle-envelope boundary (MEB) up to the radiative-convective boundary (RCB).
- Atmosphere Models: Uses the SCARLET radiative-convective code to generate self-consistent temperature-pressure ( $T$ - $P$ ) profiles. Crucially, these models account for composition-dependent opacity and greenhouse effects, solving for the RCB self-consistently rather than assuming a fixed pressure.
- Coupling: The framework links the atmosphere and interior by minimizing the $\delta TPR$ distance (temperature, pressure, and radius mismatch) at the link point (typically 1 kbar or the RCB).
Demixing Criterion:
- The framework compares the planet's $T$ - $P$ profile against hydrogen-water coexistence curves derived from recent DFT calculations (Gupta et al. 2024).
- Instability Condition: If the planet's temperature profile drops below the coexistence curve at any point within the envelope, the mixture is thermodynamically unstable, and demixing (phase separation) is expected.
Case Study & Grid:
- Case Study: TOI-270 d ( $M_p \approx 4.8 M_\oplus$ , $R_p \approx 2.2 R_\oplus$ , $T_{eq} \approx 380$ K).
- Parameter Space: A grid covering planet masses ( $2.9\text{--}20 M_\oplus$ ), envelope mass fractions ( $f_{env}$ ), and envelope metallicities ( $Z_{env}$ from solar to pure water).
- Inference: A Bayesian approach is used to map the posterior distribution of $f_{env}$ and $Z_{env}$ consistent with TOI-270 d's observed mass, radius, and atmospheric constraints.

3. Key Contributions

ATHENAIA Framework: A novel, open-source tool that self-consistently couples radiative-convective atmosphere models with interior structure models to evaluate phase stability.
Identification of the "Demixing Window": The paper defines a specific region in parameter space (Mass, $T_{eq}$ , $Z_{env}$ , $f_{env}$ ) where warm sub-Neptunes are predicted to have stratified envelopes, challenging the "fully miscible" paradigm.
Re-evaluation of TOI-270 d: Provides a rigorous assessment of TOI-270 d, showing that its observed properties are compatible with a demixing scenario, implying a bulk envelope metallicity significantly higher than the upper atmosphere.
Impact of Methane: Preliminary analysis suggests that the presence of methane (CH $_4$ ) creates shallower adiabatic gradients, potentially expanding the demixing window for metal-rich planets.
Mantle State Re-evaluation: Demonstrates that self-consistent atmosphere models alter the predicted state of the silicate mantle (molten vs. solid) at the MEB, making envelope mass fraction the dominant factor rather than metallicity alone.

4. Key Results

Existence of a Demixing Window: Contrary to previous assumptions, the authors find that water-hydrogen demixing is possible for warm sub-Neptunes ( $T_{eq} \approx 330\text{--}450$ $T_{e q} \approx 330 - 450$ K) if they have high bulk metallicities ( $Z_{env} \sim 150\text{--}700\times$ $Z_{e n v} \sim 150 - 700 \times$ solar).
- The "window" is maximal at metallicities around $Z_{env} \approx 0.8$ ( $\sim 300\times$ solar).
- Demixing is favored by higher planet masses and lower equilibrium temperatures.
TOI-270 d Implications:
- The inferred bulk properties of TOI-270 d overlap significantly with the demixing window.
- If demixing occurs, the upper atmosphere (observed by JWST) would be metal-depleted relative to the bulk envelope.
- Correction: The true bulk envelope metallicity could be up to $Z_{env} \sim 0.85$ (much higher than the $\sim 200\text{--}300\times$ solar measured in the upper atmosphere), implying a less massive core and a more massive envelope than previously thought.
Population-Level Impact:
- Demixing is predicted across the entire mass-radius diagram for sub-Neptunes with $Z_{env} \gtrsim 150\times$ solar.
- Using upper-atmosphere metallicities to infer bulk composition leads to underestimates of envelope mass fractions and overestimates of core masses.
- This introduces a new degeneracy: a metal-poor, fully-mixed envelope can mimic the radius of a metal-rich, demixed envelope.
Mantle Melting: The competition between the enhanced greenhouse effect (warming the deep atmosphere) and shallower adiabatic gradients (cooling the deep interior) restricts the conditions for a molten mantle. For TOI-270 d, a solid mantle is only possible if the envelope mass fraction is $>7\%$ and the mean molecular weight is $\gtrsim 5$ amu.

5. Significance

Paradigm Shift: This work challenges the prevailing view that warm sub-Neptunes have homogeneous envelopes. It suggests that compositional stratification is a common feature for metal-rich sub-Neptunes, similar to the solar system's ice giants but occurring at much higher temperatures.
Interpretation of JWST Data: It implies that transmission spectroscopy (probing the upper atmosphere) may systematically underestimate the total metal content of sub-Neptunes. Future evolutionary models must account for metallicity gradients to accurately predict cooling rates and mass loss.
Formation and Evolution: The findings suggest that the "building blocks" of sub-Neptunes (accreted volatiles) may be more abundant than current atmospheric measurements suggest. It motivates new evolutionary models that track the onset of metallicity gradients over Gyr timescales.
Future Directions: The authors call for evolutionary modeling to determine if these planets cool sufficiently to enter the demixing window and for further study on the role of other volatiles (like methane and carbon) in driving demixing.

In summary, the paper establishes that water-hydrogen demixing is a viable and likely mechanism for warm, metal-rich sub-Neptunes, necessitating a fundamental revision of how we interpret their internal structures, bulk compositions, and evolutionary histories.