Hydrogen-helium immiscibility boundary in planets

By employing large-scale molecular dynamics with machine learning potentials to map the hydrogen-helium immiscibility boundary, this study reveals that demixing temperatures are approximately 2000 K lower than previous estimates, suggesting that helium rain is plausible in Saturn but unlikely in Jupiter's warmer interior.

The Gist

Imagine the giant planets in our solar system, like Jupiter and Saturn, as massive, swirling balls of gas. For decades, scientists have wondered about a specific "weather event" happening deep inside them: Helium Rain.

Think of it like this: Deep inside these planets, the pressure is so crushing and the heat so intense that hydrogen and helium—the two main ingredients of these planets—should be mixing together like milk in coffee. But, under certain conditions, they might stop mixing and start separating, like oil and vinegar. When this happens, the helium forms droplets that are heavier than the surrounding hydrogen. These droplets fall toward the planet's core, creating a "rain" of helium.

This paper is about figuring out exactly when and where this helium rain starts.

The Problem: A Foggy Map

Scientists knew this rain might happen, but they didn't have a clear map.

• The Experiment Problem: You can't just build a giant pressure cooker to test this. The conditions inside Jupiter are so extreme (millions of times the pressure of Earth's atmosphere) that our current lab equipment struggles to measure it accurately.
• The Computer Problem: Previous computer simulations were like trying to predict the weather by looking at a tiny puddle. They used small groups of atoms (a few hundred) and didn't run long enough to see the big picture. They also had to make some guesses about how the atoms interacted, which led to conflicting results. Some said the rain starts at 6,000°C; others said 8,000°C. That's a huge difference!

The Solution: The "AI Weather Forecaster"

The authors of this paper built a new, super-smart tool to solve this.

1. The Teacher (DFT): They started with the best existing physics equations (called Density Functional Theory) to teach a computer how hydrogen and helium atoms behave. They used three different "teachers" (mathematical models) to make sure the lesson was solid.
2. The Student (Machine Learning Potentials): They trained an Artificial Intelligence (AI) to learn from these teachers. Instead of doing the heavy math for every single atom every time, the AI learned the "rules of the game."
3. The Simulation: Once the AI was smart enough, they let it run massive simulations with thousands of atoms (instead of just a few hundred) for longer periods. This is like going from looking at a puddle to watching a massive ocean storm.

The Big Discovery: It's Cooler Than We Thought

The results were surprising. The team found that helium rain starts at temperatures about 2,000 degrees Celsius lower than previous computer models predicted.

The Analogy: Imagine you thought you needed to boil water at 120°C to make it steam. Your new, better thermometer tells you it actually starts steaming at 100°C. This changes everything about how you cook your pasta. Similarly, knowing the "boiling point" of helium separation is lower changes our understanding of how these planets work.

What This Means for Jupiter and Saturn

Now that we have a better map, let's look at our two favorite gas giants:

• Jupiter (The Warm Giant): Jupiter is incredibly hot deep inside. Even with our new, lower temperature for helium rain, Jupiter's interior is still too hot. The hydrogen and helium are still happily mixed together like a smoothie. Conclusion: Jupiter probably doesn't have helium rain right now. This is a puzzle because Jupiter's atmosphere is missing some helium, and we thought the rain was the reason. Now, scientists have to find a different explanation for Jupiter's "missing" helium.

• Saturn (The Cooler Giant): Saturn is smaller and cooler than Jupiter. Our new map shows that Saturn's interior is just cold enough for the helium to stop mixing and start raining down.

  • Why Saturn is Shiny: Saturn is glowing brighter than it should be for its age. Scientists think the falling helium droplets act like a heater, releasing extra energy as they sink. Our new map confirms this is very likely happening in Saturn.
  • The Structure: As the helium rains down, it leaves behind a layer of hydrogen that is poor in helium. This creates distinct layers inside the planet, like a cake with different flavors, which might explain why Saturn's magnetic field is so weirdly perfect and symmetrical.

The Takeaway

This paper is like upgrading from a blurry, black-and-white photo of a storm to a high-definition, 3D video. By using AI to simulate massive systems, the authors have narrowed down the uncertainty.

They tell us:

1. Helium rain is real, but it happens at lower temperatures than we thought.
2. Saturn is likely experiencing this rain right now, which explains its extra heat and strange magnetic field.
3. Jupiter is likely too hot for the rain, meaning we need to rethink how Jupiter lost its helium.

This new understanding helps scientists build better models of not just our solar system, but the thousands of "super-Jupiters" and "super-Saturns" orbiting other stars in the galaxy.

Technical Summary

1. Problem Statement

The phase separation of hydrogen (H) and helium (He) in the deep interiors of gas giants (Jupiter and Saturn) is a critical phenomenon known as "helium rain." This process involves helium becoming immiscible in hydrogen under high pressure and temperature, precipitating as droplets that sink toward the core. This mechanism explains atmospheric helium depletion, excess planetary luminosity, and internal compositional stratification.

However, the precise location of the H/He immiscibility boundary (the $P-T-x_{He}$ conditions where phase separation occurs) remains highly uncertain due to:

• Experimental limitations: High-pressure experiments (e.g., laser-shock) are difficult to perform and currently provide only sparse data points.
• Computational limitations: Previous ab initio molecular dynamics (AIMD) simulations based on Density Functional Theory (DFT) are constrained by small system sizes (typically <100 atoms) and short simulation times (picoseconds). These constraints lead to significant finite-size effects and reliance on idealized entropy assumptions, resulting in large discrepancies (up to 2000–4000 K) in predicted demixing temperatures among different studies.
• Functional dependence: The impact of different exchange-correlation (XC) functionals (PBE, vdW-DF, HSE) on the immiscibility boundary has not been systematically resolved at the scale required for planetary modeling.

2. Methodology

The authors employed a multi-faceted computational approach to overcome the limitations of traditional AIMD:

• Machine Learning Potentials (MLPs): They trained high-fidelity MLPs on three distinct DFT functionals to capture atomic interactions accurately while enabling large-scale simulations:

  • PBE: A standard generalized gradient approximation.
  • vdW-DF: Includes van der Waals interactions, crucial for high-pressure mixtures.
  • HSE: A hybrid functional including exact exchange, known for superior performance in benchmark studies.
  • Training Strategy: They used a $\Delta$-learning strategy for HSE, where the MLP learns the energy difference between vdW-DF and HSE, making the computationally expensive HSE training tractable.
  • Architectures: They tested and validated multiple architectures (N2P2, MACE, CACE), confirming that N2P2 offered the best balance of efficiency and accuracy for large-scale sampling.

• Large-Scale Molecular Dynamics (MD):

  • Simulations were performed on systems with 3,456 atoms (significantly larger than previous AIMD studies), allowing for the observation of macroscopic phase separation.
  • The study covered a broad parameter space: Pressures (100–1000 GPa) and Temperatures (1000–12,000 K), spanning the interior conditions of gas giants.

• Thermodynamic Calculations:

  • Chemical Potentials: Instead of relying on particle insertion methods (which suffer from convergence issues), the authors used the $S_0$ method to compute chemical potentials ($\mu$) from static structure factors ($S(k)$) in the limit of infinite wavelength.
  • Free Energy: The mixing free energy ($\Delta G_{mix}$) was derived from the integrated chemical potentials.
  • Validation: MLP-derived free energies were corrected to the DFT level using Free Energy Perturbation (FEP) to ensure accuracy.
  • Nuclear Quantum Effects (NQEs): Path Integral MD (PIMD) simulations were conducted to assess the impact of NQEs, finding they slightly enlarge the immiscibility region.

• Thermodynamic Modeling:

  • The computed $\Delta G_{mix}$ data was fitted to a Redlich-Kister (R-K) regular solution model (3rd order) to rationalize the thermodynamic driving forces and create a predictive representation of the boundary.

3. Key Results

• Consistency Across Functionals: Despite theoretical differences, the three XC functionals (PBE, vdW-DF, HSE) yield consistent immiscibility boundaries. The boundaries are highly sensitive to helium concentration ($x_{He}$) and temperature but show relatively subtle dependence on the choice of functional (differences < 500 K).
• Lower Demixing Temperatures: The most significant finding is that the demixing temperatures predicted by this large-scale study are ~2000 K lower than previous AIMD studies using small system sizes.
  • Example: At 150 GPa and protosolar helium abundance ($x_{He} \approx 0.089$), the study predicts demixing at ~3500 K, whereas previous PBE AIMD studies predicted ~6000–7500 K.
• Mechanism of Discrepancy: The authors attribute the ~2000 K discrepancy with prior work to finite-size effects. Small systems in previous AIMD studies enforced complete mixing, failing to capture the true thermodynamic instability and the linear segment of $\Delta G_{mix}$ associated with phase separation.
• Atomic-Molecular Transition: The study confirms that helium stabilizes molecular hydrogen, slightly raising the H atomic-molecular transition temperature. However, the immiscibility boundary is strongly coupled to this transition; helium solubility drops drastically once hydrogen becomes atomic (metallized).
• Thermodynamic Drivers: The R-K model analysis reveals that the phase separation is primarily driven by the $\omega_1$ and $\omega_3$ interaction parameters at low temperatures. As temperature increases, the sign reversal of the $\omega_2$ term promotes miscibility, particularly for helium-poor compositions.

4. Implications for Planetary Models

The authors compared their new immiscibility boundaries with current interior models of Jupiter and Saturn:

• Jupiter:

  • Both adiabatic and modern evolutionary models (e.g., HMH24) suggest that Jupiter's interior is unlikely to experience helium rain.
  • The planetary isentrope lies above the immiscibility boundary. Even with cooler interior models, the peak helium abundance ($x_{He} \approx 0.09$) remains below the threshold required for demixing ($x_{He} > 0.15$) at relevant pressures.
  • Conclusion: Alternative mechanisms must explain Jupiter's atmospheric helium depletion.

• Saturn:

  • Saturn's cooler interior profile intersects the immiscibility boundary, making helium rain highly plausible.
  • The study suggests helium rain may have begun early in Saturn's history and persists, potentially reaching the core.
  • This process explains Saturn's anomalous luminosity (via gravitational energy release), atmospheric helium depletion, and the formation of compositional gradients that influence its magnetic field.
  • Note: The onset of demixing is predicted to be delayed compared to previous models due to the lower temperature boundaries found in this study.

5. Significance

• Resolution of Uncertainty: This work significantly narrows the uncertainty in the H/He phase diagram, resolving long-standing discrepancies between different theoretical approaches.
• Methodological Advancement: It demonstrates the power of combining Machine Learning Potentials with large-scale MD and advanced thermodynamic integration ($S_0$ method) to study complex phase transitions in extreme conditions where experiments are infeasible.
• Planetary Evolution: The results provide critical inputs for next-generation planetary models, forcing a re-evaluation of the thermal evolution, internal structure, and magnetic field generation of gas giants. Specifically, it confirms that helium rain is a dominant factor in Saturn's evolution but likely not in Jupiter's.
• Exoplanets: The refined phase boundaries offer a robust framework for characterizing the interiors and evolution of exoplanetary gas giants.