🔭 astrophysics

Non-LTE Ionization Modeling for Helium and Strontium in Neutron Star Merger Ejecta

This study employs non-LTE ionization modeling to determine that the 1 $\mu$ m absorption feature in the kilonova AT2017gfo can be explained by either approximately 1% helium or 1–10% strontium in the ejecta, with the latter implying low electron fraction and entropy conditions consistent with solar r-process abundances, while the former at very low electron fractions suggests the production of elements beyond the third r-process peak.

Original authors: Koya Chiba, Masaomi Tanaka, Shinya Wanajo, Sho Fujibayashi, Kyohei Kawaguchi, Kenta Hotokezaka

Published 2026-04-08

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Koya Chiba, Masaomi Tanaka, Shinya Wanajo, Sho Fujibayashi, Kyohei Kawaguchi, Kenta Hotokezaka

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

The Cosmic Crime Scene: Solving the "1 Micron Mystery"

Imagine two neutron stars (the super-dense, city-sized corpses of dead stars) colliding in a violent dance. This crash creates a massive explosion of debris, glowing with a radioactive afterglow called a kilonova.

In 2017, astronomers watched one such crash (GW170817) and saw a specific "fingerprint" in its light: a deep dip in the spectrum right around the 1-micron mark (a specific shade of near-infrared light).

The Mystery: What element caused this dip?
For years, scientists argued it was either Strontium (a metal used in fireworks) or Helium (the gas in balloons). The problem? They were trying to solve this puzzle using an old map that didn't quite fit the terrain.

The Old Map vs. The New GPS

The Old Map (LTE):
Previously, scientists modeled these explosions assuming the gas was in "Local Thermodynamic Equilibrium" (LTE). Think of this like assuming a crowded party is calm and everyone is chatting at the same volume. In this calm state, atoms behave predictably.

The New GPS (Non-LTE):
This paper argues that the kilonova isn't a calm party; it's a mosh pit.
Because the explosion is powered by radioactive decay, it's bombarded by high-energy electrons (like tiny, hyperactive bullets). These bullets knock atoms apart and excite them in ways that a "calm" model can't predict. This is called Non-LTE (Non-Local Thermodynamic Equilibrium).

The authors built a new, high-tech GPS model that accounts for these "bullet" electrons. They realized that under these chaotic conditions, atoms don't just sit there; they get ionized (stripped of electrons) much more aggressively than we thought.

The Two Suspects: Helium and Strontium

The team ran their new model to see how much of each element was needed to create that 1-micron dip.

The Strontium Suspect:
- The Twist: In the old "calm" model, you needed a little bit of Strontium to make the dip. But in the new "mosh pit" model, the radioactive bullets strip so many electrons from Strontium that it becomes "invisible" to the light.
- The Result: To get the dip to show up, you actually need 10 to 100 times more Strontium than previously thought.
- The Good News: When they calculated this new, higher amount, it matched perfectly with the amount of Strontium found in our own Solar System. This is a huge win! It suggests that neutron star collisions are indeed the main factories creating heavy elements in the universe.
The Helium Suspect:
- The Twist: Helium is tricky. In the hot, early days of the explosion, the "bullets" knock Helium atoms into a specific excited state that creates the dip.
- The Result: If the explosion was extremely chaotic (very low "electron fraction"), the dip could be caused by about 1% Helium.
- The Deep Mystery: If Helium is the culprit, it implies something even wilder happened: the collision created elements heavier than Lead (the "third peak" of the periodic table) which then decayed into Helium. Finding this Helium would be like finding a fossil that proves a dinosaur existed, even if you never saw the dinosaur itself.

The Verdict: What Does This Tell Us About the Crash?

By comparing their new model with computer simulations of how these stars crash, the authors drew some fascinating conclusions:

The "Too Hot" Problem: Some computer simulations of long-lasting neutron star remnants (stars that survive the crash for a while) predict the debris should be very "hot" and full of Helium and Strontium. But our new model says that would create a dip too strong to match what we saw.
The Solution: The crash must have happened under "cooler" and "denser" conditions than those simulations predicted.
The Implication: The debris was likely ejected very quickly, before the environment got too chaotic. It suggests the neutron star remnant didn't survive for long, or the physics of the crash was different than we thought.

The Takeaway

Think of this paper as upgrading the detective's toolkit. By realizing the "crime scene" (the kilonova) was more violent and energetic than we thought, the scientists were able to:

Correct the evidence: Realize we need way more Strontium than we thought.
Confirm the theory: Prove that neutron star collisions are the cosmic factories making the heavy stuff in our universe.
Refine the timeline: Suggest that the debris was flung out faster and under different conditions than our current best computer models predicted.

It's a reminder that in the universe, things are rarely "calm," and sometimes you have to account for the mosh pit to understand the music.

1. Problem Statement

The kilonova AT2017gfo, associated with the binary neutron star (BNS) merger GW170817, exhibits a prominent absorption feature around 1 µm in its early-phase spectra. Two primary candidates have been proposed to explain this feature:

Strontium (Sr II): Specifically the transition around 1.03–1.09 µm.
Helium (He I): Specifically the transition at 1.08 µm ( $2^3S - 2^3P$ ).

Previous studies have struggled to quantify the contributions of these elements simultaneously due to two main limitations:

LTE Assumption: Early-phase spectral modeling has largely relied on Local Thermodynamic Equilibrium (LTE). However, the low density and high-energy radiation environment of kilonova ejecta (driven by radioactive decay) induce significant non-thermal ionization, causing deviations from LTE.
Lack of Simultaneous Modeling: Previous works either modeled He or Sr in isolation, or did not systematically explore the full range of mass fractions required to reproduce the observed feature.

The core problem is to determine the true abundances of He and Sr in the ejecta by developing a self-consistent Non-LTE (Non-Local Thermodynamic Equilibrium) ionization model that accounts for ionization by high-energy electrons.

2. Methodology

A. Spectral Fitting and Physical Conditions

The authors first analyzed VLT/X-shooter spectra of AT2017gfo at four epochs (1.43, 2.42, 3.41, and 4.40 days post-merger). They fitted the continuum and the 1 µm feature using a relativistic blackbody and a P-Cygni profile.

Key Finding: To reproduce the depth of the 1 µm feature, a Sobolev optical depth of $\tau \approx 1$ (specifically $0.5 \le \tau \le 2$ ) is required at the photosphere.
Parameters: Photospheric velocities ( $v_{ph}$ ) range from $0.21c$ to $0.14c$ , and temperatures ( $T_0$ ) range from $4500$ K to $2600$ K.

B. Non-LTE Ionization Framework

The study develops a rate-equation solver for both He and Sr, incorporating non-thermal ionization caused by high-energy electrons produced by radioactive decay ( $\beta$ -decay and $\gamma$ -rays).

Helium (He) Model:
- A detailed atomic model with 21 states (19 bound states of He I up to $n=4$ , plus ground states of He II and He III).
- Includes radiative and collisional (electron-impact) processes for excitation, de-excitation, ionization, and recombination.
- Crucial Mechanism: The 1.08 µm transition originates from the metastable $2^3S$ state. In typical kilonova temperatures ( $<5000$ K), thermal excitation cannot populate this state efficiently. The model relies on non-thermal ionization followed by recombination to populate the $2^3S$ state.
Strontium (Sr) Model:
- Due to the complexity of heavy elements, a simplified ionization model is used, tracking ionization states up to Sr XX (though only up to Sr VII are significant).
- Assumes a Boltzmann distribution for excited states of Sr II (responsible for the 1 µm feature).
- The ionization balance is governed by the competition between thermal photoionization (diluted blackbody) and non-thermal ionization ( $\Gamma_{nt} = \dot{q}/w_i$ ).
- Key Mechanism: High-energy electrons cause over-ionization, stripping electrons from Sr II to higher ionization states (Sr III, IV, etc.), thereby depleting the population of Sr II.

C. Abundance Constraints

The authors iteratively solve the rate equations to find the mass fractions ( $X_{He}$ and $X_{Sr}$ ) required to achieve the target optical depth ( $\tau \approx 1$ ) at each epoch, considering the self-consistent electron density derived from the ionization states.

3. Key Contributions

Development of a Self-Consistent Non-LTE Model: This is the first study to simultaneously model the ionization of both He and Sr in kilonova ejecta, explicitly accounting for non-thermal ionization by high-energy electrons.
Quantification of Non-LTE Effects:
- For Sr: Non-LTE effects significantly suppress the population of Sr II (by factors of $10^2$ – $10^4$ ) compared to LTE predictions due to over-ionization. Consequently, the required mass fraction of Sr to reproduce the feature is much higher in Non-LTE than in LTE.
- For He: The population of the metastable $2^3S$ state is highly sensitive to temperature and non-thermal ionization rates.
Systematic Abundance Constraints: The paper provides constraints on the mass fractions of He and Sr across the velocity range of the ejecta ( $v \sim 0.15c$ ), rather than just single representative values.

4. Results

A. Required Mass Fractions

To reproduce the 1 µm feature in AT2017gfo at $v \sim 0.15c$ :

Helium Scenario: Requires a mass fraction of $X_{He} \approx 1\%$ .
Strontium Scenario: Requires a mass fraction of $X_{Sr} \approx 1\% - 10\%$ .
- Note: The required Sr mass fraction in the Non-LTE case is roughly two orders of magnitude higher than in LTE models.
- The derived Sr mass fraction ( $1-10\%$ ) aligns remarkably well with the solar r-process abundance pattern.

B. Nucleosynthesis Implications

By comparing the derived abundances with parametric nucleosynthesis calculations (S. Wanajo 2018), the authors constrain the physical conditions of the ejecta:

Electron Fraction ( $Y_e$ ): The results favor a low electron fraction ( $Y_e \lesssim 0.35$ ).
- High $Y_e$ ( $>0.35$ ) leads to $\alpha$ -rich freeze-out, overproducing both He and Sr, which would create an absorption feature too strong to match observations.
Entropy ( $s$ ): The results favor low entropy ( $s \lesssim 30 k_B/\text{nucleon}$ ). High entropy also exacerbates $\alpha$ -rich freeze-out.
Two Possible Origins for the 1 µm Feature:
1. Intermediate $Y_e$ ( $0.15 \lesssim Y_e \lesssim 0.35$ ): The feature is dominated by Sr. This supports the hypothesis that BNS mergers are the primary source of r-process elements in the universe.
2. Very Low $Y_e$ ( $Y_e \lesssim 0.15$ ): The feature is dominated by He. In this regime, He is produced via $\alpha$ -decays of trans-Pb nuclei. If confirmed, this would provide an indirect signature of the production of elements beyond the third r-process peak.

C. Mass Ejection Mechanism

Comparing their constraints with long-term viscous hydrodynamics simulations (specifically the DD2-135-135 model, which leaves a long-lived neutron star remnant):

The standard long-lived remnant models tend to overproduce both He and Sr because the post-merger winds have high $Y_e$ and high entropy.
The authors suggest that the significant mass ejection in GW170817 must have occurred under lower $Y_e$ and lower entropy conditions than predicted by current long-lived models.
This implies the mass ejection might happen on a shorter timescale (before $Y_e$ rises too high), consistent with recent magnetohydrodynamics (MHD) simulations.

5. Significance

Resolving the 1 µm Feature: The study clarifies that the 1 µm feature can be explained by either Sr or He, but the required abundances are drastically different depending on the ionization physics. The Non-LTE treatment is essential; LTE models significantly underestimate the required Sr abundance.
Cosmochemical Constraints: The agreement between the derived Sr abundance and solar r-process patterns strengthens the case for BNS mergers as the dominant site of heavy element production.
Probing the Equation of State (EoS): The abundance of He serves as a potential probe for the lifetime of the remnant neutron star. If the feature is He-dominated, it implies a very low $Y_e$ environment, potentially constraining the nuclear EoS and the remnant's stability.
Future Directions: The paper highlights the need for more detailed modeling of He emission lines (e.g., at 2.06 µm) and consistent light curve modeling to break the degeneracy between the He and Sr scenarios.

In conclusion, this work establishes that Non-LTE ionization is critical for interpreting kilonova spectra. It provides robust constraints on the composition of GW170817's ejecta, suggesting a low-entropy, low-electron-fraction environment that may differ from current long-lived remnant simulations.