⚛️ nuclear experiments

The impact of nuclear uncertainties on the p-process nucleosynthesis in Supernovae

This study quantifies how uncertainties in nuclear level densities and photon strength functions, particularly those arising from local parameter variations in photoneutron emission rates, propagate to limit the precision of p-process nucleosynthesis predictions in Type-Ia and Type-II supernovae, identifying specific reactions involving stable or near-stable nuclei as primary targets for future experimental constraints.

Original authors: S. Martinet, S. Goriely, A. Choplin

Published 2026-07-13

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: S. Martinet, S. Goriely, A. Choplin

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, chaotic kitchen where stars are the chefs. Most of the ingredients (elements) are made by slowly adding neutrons to a pot, like a slow-cooked stew. But there's a rare, tricky set of ingredients called p-nuclides. These are the "neutron-deficient" spices of the cosmic pantry—heavy atoms that are missing some neutrons. They are so rare that in our Solar System's recipe book, they make up only about 1% to 0.1% of their heavier cousins.

For decades, astronomers have been trying to figure out exactly how these rare spices get cooked. The leading theory is the p-process, which happens during the explosive finale of massive stars (supernovae) or the violent death of white dwarfs (Type Ia supernovae). It's like a sudden, super-hot blast that blasts neutrons off existing heavy atoms, turning them into these rare p-nuclides.

But here's the problem: our recipe cards are full of holes. We don't know the exact "cooking times" (reaction rates) for these blasts because the ingredients involved are unstable and hard to study in a lab. This paper by Martinet, Goriely, and Choplin is like a team of detective chefs trying to figure out how much those missing recipe details mess up the final dish.

The Great Uncertainty Hunt

The team used a clever computer trick called the Backward–Forward Monte Carlo (BFMC) method. Imagine you are trying to guess the weight of a mystery box. Instead of guessing randomly, you first look at boxes whose weight you do know. You tweak your measuring scale until it perfectly matches those known boxes. Once your scale is calibrated, you use it to weigh the mystery boxes.

In this study, the "scale" is a complex nuclear physics model. The team tweaked the model's settings (parameters) until it perfectly matched known experimental data. Then, they used those same tweaked settings to predict the behavior of the unknown, unstable nuclei involved in the p-process. They ran thousands of these simulations to see how much the final amount of p-nuclides would wiggle if the recipe details were slightly off.

The Big Surprise: It's All About the Neutrons

The most exciting finding? The team discovered that the biggest source of confusion comes from just one type of reaction: photoneutron emission.

Think of the p-process as a game of musical chairs where atoms are the chairs and energy is the music. When the music stops (the explosion cools down), atoms need to lose neutrons to become p-nuclides. The paper shows that the uncertainty in the final amount of p-nuclides is almost entirely driven by how easily an atom can lose a neutron when hit by a photon (a particle of light).

The team explicitly ruled out other suspects. They tested whether the uncertainty in proton capture (adding a proton) or alpha capture (adding a helium nucleus) was the problem. The answer? Nope. In their simulations, changing the rates for these reactions barely made a dent in the final results. The uncertainty budget is overwhelmingly dominated by the neutron-loss reactions.

How Much Does It Matter?

When they ran their simulations for three different types of cosmic explosions (a standard massive star, a spinning massive star, and a Type Ia supernova), they found a consistent level of chaos.

Regardless of which star exploded, the uncertainty in the final amount of p-nuclides was about 0.7 dex. In plain English, that means the predicted amount could be off by a factor of five. If the model says you should have 10 cookies, the reality could be anywhere from 2 to 50 cookies, just because we aren't sure exactly how fast the neutrons fly off.

This huge swing happens even though the different stars have very different starting ingredients. The paper argues that this uncertainty isn't because our models of how the stars explode are wrong, but because our knowledge of the nuclear physics (the atomic rules) is incomplete.

The "Systematic" vs. "Statistical" Showdown

The authors also asked: "Is the problem that we picked the wrong type of model, or that the numbers inside our chosen model are fuzzy?"

They compared their main model against several other sophisticated nuclear models. They found that the differences between the models (systematic uncertainty) were tiny—only a few percent. However, the fuzziness within their own model (parameter uncertainty), caused by the lack of experimental data, was massive.

The verdict: The problem isn't that we're using the wrong theory; it's that we don't have enough data to pin down the numbers in the theory we are using. The "fuzzy numbers" allowed by current experiments are the real culprits.

The Treasure Map for Future Experiments

So, where should scientists look next? The team used a statistical tool to map out exactly which reactions are the "bosses" controlling the uncertainty.

They found that for a huge chunk of the p-nuclides, the most critical reactions are surprisingly close to home. Often, the reaction that controls the abundance of a specific p-nuclide is simply the photodisintegration of that same p-nuclide or its immediate neighbor on the periodic table.

This is great news for experimentalists! Many of these "boss" reactions involve stable or near-stable nuclei. This means they aren't impossible to study in a lab. The paper suggests that if we can measure the photodisintegration of these specific stable neighbors, we could shrink that factor-of-five uncertainty down to something much more manageable.

The Bottom Line

This paper doesn't claim to have solved the mystery of p-nuclides. Instead, it acts like a flashlight in a dark room, showing us exactly where the shadows are thickest. It tells us that while our astrophysical models of supernovae are doing a decent job, our nuclear physics data is the weak link.

The uncertainty isn't a mystery of the stars; it's a mystery of the atom. And the good news is that the key to unlocking it lies in measuring reactions on stable nuclei that we can actually reach in a laboratory. Until we do, our cosmic recipes will remain a bit too vague to perfectly predict the universe's rarest spices.

Technical Summary: The Impact of Nuclear Uncertainties on the p-Process Nucleosynthesis in Supernovae

Problem Statement
The origin of the 35 stable neutron-deficient isotopes (p-nuclides) heavier than iron remains a significant challenge in nuclear astrophysics. While the p-process, driven by photodisintegration reactions ( $\gamma, n$ ), ( $\gamma, p$ ), and ( $\gamma, \alpha$ ) on pre-existing seed nuclei in supernova environments, is the leading explanation, current models struggle to reproduce solar system abundances for specific isotopes (notably $^{92,94}$ Mo and $^{96,98}$ Ru). These discrepancies arise from uncertainties in both astrophysical conditions and nuclear physics inputs. Specifically, the calculation of photodisintegration rates relies on the Hauser-Feshbach statistical model, which requires Nuclear Level Densities (NLDs) and Photon Strength Functions (PSFs). These inputs are subject to large uncertainties, particularly for neutron-deficient, unstable isotopes where experimental data is scarce. The extent to which these nuclear uncertainties propagate to final abundance predictions, and which specific reactions dominate this uncertainty budget, has not been systematically quantified with correlated inputs.

Methodology
The authors extend the Backward–Forward Monte Carlo (BFMC) approach, previously applied to the i-process, to the p-process regime. The methodology involves three primary stages:

Generation of Correlated Rate Sets: Using the Hauser-Feshbach code TALYS, the authors computed radiative capture rates ( $(n, \gamma)$ , $(p, \gamma)$ , $(\alpha, \gamma)$ ) and their inverse photodisintegration rates for approximately 2000 neutron-deficient isotopes. To quantify parameter uncertainties, they varied four local parameters governing the NLD (scaling factor $\alpha$ and energy shift $\delta$ ) and PSF (centroid energy $\delta_E$ and width $\delta_\Gamma$ ) within the D1M+QRPA and HFB+comb microscopic frameworks.
BFMC Filtering: Thousands of parameter combinations were sampled. In the "backward" step, only sets reproducing known experimental Maxwellian-averaged cross sections (MACS) at 30 keV were retained. In the "forward" step, these filtered sets were applied to unmeasured reactions. Crucially, for each nucleus, the $(n, \gamma)$ , $(p, \gamma)$ , and $(\alpha, \gamma)$ rates were derived from the same parameter set, preserving physical correlations between channels.
Propagation and Sensitivity Analysis: The resulting correlated extreme rate sets were propagated through full p-process nucleosynthesis networks for three astrophysical sites: the W7 Type Ia supernova (SNIa) model, a non-rotating 25 $M_\odot$ core-collapse supernova (CCSN) at solar metallicity, and a rotating 25 $M_\odot$ CCSN at low metallicity. To identify key drivers, the authors employed regularized linear-response modeling, stability analysis, and a decomposition of contributions using SHAP (SHapley Additive exPlanations) values to distinguish between direct effects and interaction effects.

Key Results

Dominance of Parameter Uncertainties: The study finds that uncorrelated parameter uncertainties (local variations of NLD and PSF parameters allowed by current experimental constraints) dominate the uncertainty budget. Systematic uncertainties arising from the choice of different microscopic nuclear models (e.g., comparing D1M+QRPA to SMLO or RMF+cQRPA) are relatively small (typically <10%) compared to the factor-of-five spread introduced by parameter variations.
Channel Sensitivity: The uncertainty in final p-nuclide abundances is overwhelmingly controlled by the $(n, \gamma)$ reaction rates (and their inverses, $(\gamma, n)$ ). Variations in $(p, \gamma)$ and $(\alpha, \gamma)$ rates, while physically correlated, contribute negligibly to the total abundance spread. This is attributed to the fact that NLD and PSF uncertainties have a significantly stronger impact on neutron-capture rates in the proton-rich region relevant to the p-process.
Magnitude of Uncertainty: Across all three astrophysical sites, the propagation of nuclear uncertainties results in abundance variations of approximately 0.7 dex (a factor of five) between minimum and maximum predictions. This level of uncertainty is consistent regardless of the specific supernova model or seed distribution.
Identification of Core Drivers: Statistical analysis identifies a specific set of "core driver" reactions controlling the abundance of each p-nucleus. For a large fraction of p-nuclides (from $^{74}$ Se to $\sim^{152}$ Gd), the dominant uncertainty stems from either the photoneutron emission of the p-nucleus itself or the $(\gamma, n)$ reaction on a neighboring isotope in the same isotopic chain. These drivers are physically located near independently determined photodisintegration branching regions where $(\gamma, n)$ competes with $(\gamma, p)$ or $(\gamma, \alpha)$ .
Site Independence: While the absolute yields differ between SNIa and CCSN models due to different seed distributions, the specific nuclear reactions driving the uncertainties are largely common across the different astrophysical sites.

Significance and Conclusions
The paper concludes that the primary limitation in reducing p-process abundance uncertainties is the lack of sufficiently constraining nuclear data in the relevant neutron-deficient region, rather than the choice of theoretical nuclear models. A significant finding is that many of the identified "core driver" reactions correspond to stable or near-stable nuclei, including the photodisintegration of the p-nuclei themselves. Consequently, the authors assert that a substantial fraction of the reactions controlling p-process abundance uncertainties are, in principle, experimentally accessible through direct measurements or indirect constraints (such as the Oslo method). Improving experimental constraints on these specific photoneutron emissions is identified as the most effective path toward reducing the uncertainty in p-process nucleosynthesis predictions.