Determination of nuclear deformations with an emulator for sub-barrier fusion reactions

This paper demonstrates that an emulator based on eigenvector continuation can significantly accelerate coupled-channels calculations for sub-barrier fusion reactions while accurately extracting nuclear deformation parameters, thereby providing a powerful tool for systematically exploring intrinsic nuclear shapes.

The Gist

The Big Picture: Predicting the Shape of a Mystery Ball

Imagine you are a detective trying to figure out the shape of a mysterious, invisible ball. You can't see it, and you can't touch it. The only way to learn about it is to throw other balls at it and watch how they bounce off or stick to it.

In the world of nuclear physics, scientists do exactly this. They smash small atomic nuclei (like Oxygen) into larger ones (like Samarium or Tungsten) to see how they fuse together. The way they fuse tells the scientists about the shape of the target nucleus. Is it a perfect sphere? Is it a football? Is it a weird, lumpy potato?

The problem is that calculating exactly how these nuclei interact is like trying to solve a massive, 1,000-piece puzzle every single time you want to test a new theory. It takes a supercomputer hours to solve just one puzzle. If you want to find the perfect shape, you have to solve that puzzle thousands of times, changing the shape slightly each time. This is too slow and too expensive.

The Solution: The "Magic Cheat Sheet" (The Emulator)

This paper introduces a new tool called an emulator. Think of the emulator not as a calculator, but as a brilliant student who has memorized the answers to a few key practice tests.

Here is how the "student" learns:

1. The Training: The supercomputer solves the hard puzzle for just a few specific shapes (let's say 5 or 9 different shapes). These are the "training points."
2. The Shortcut: The emulator looks at these few solutions and figures out the mathematical pattern connecting them. It builds a "cheat sheet" (mathematically called a reduced basis).
3. The Prediction: Now, if you ask the emulator, "What happens if the shape is this specific combination?" it doesn't need to solve the whole puzzle from scratch. It just looks at its cheat sheet and blends the previous answers together to give you an instant, highly accurate result.

It's like if you knew the weather on Monday, Wednesday, and Friday. You wouldn't need a supercomputer to predict Tuesday's weather; you could just guess based on the pattern of the days you know.

The Results: Speed and Accuracy

The authors tested this "cheat sheet" on three different nuclear reactions:

• Samarium-144: A nucleus that vibrates like a jelly.
• Samarium-154: A nucleus that is permanently stretched like a rugby ball.
• Tungsten-186: Another deformed nucleus.

The findings were amazing:

• Speed: The emulator was 200 to 400 times faster than the traditional method. A calculation that used to take hours now takes seconds.
• Accuracy: The results were almost identical to the slow, perfect method. The "cheat sheet" was just as good as doing the math from scratch.
• Discovery: Using this fast tool, they successfully extracted the exact shapes (deformation parameters) of these nuclei, matching what other experiments had found.

Why This Matters

Imagine trying to design a new car. If you had to build a full-scale prototype and crash it into a wall to test safety, it would take years and millions of dollars. But if you had a perfect virtual simulator that could predict the crash results in a split second, you could test thousands of designs instantly.

That is what this paper does for nuclear physics.

• Before: Scientists could only test a few shapes because the math was too slow.
• Now: They can systematically explore thousands of shapes to find the "true" shape of any nucleus.

The Takeaway

The authors have built a super-fast, high-precision shortcut for understanding the shapes of atomic nuclei. By using a mathematical trick called "Eigenvector Continuation," they turned a slow, heavy calculation into a quick, lightweight prediction. This opens the door to understanding the fundamental building blocks of our universe much faster and more accurately than ever before.

In short: They found a way to stop solving the whole puzzle every time and started using a smart pattern-recognition trick to solve it instantly.

Technical Summary

1. Problem Statement

• Context: Nuclear deformation significantly influences low-energy heavy-ion fusion reactions, particularly near the Coulomb barrier. Deformed nuclei exhibit a distribution of barrier heights based on their orientation, leading to enhanced fusion cross-sections.
• Challenge: Extracting nuclear deformation parameters (e.g., quadrupole $\beta_2$, hexadecapole $\beta_4$, octupole $\beta_3$) from experimental fusion cross-sections requires solving the coupled-channels (CC) equations repeatedly.
• Bottleneck: The standard CC calculation is computationally expensive. To find optimal deformation parameters, one must perform a multi-dimensional parameter search (minimizing $\chi^2$), which involves thousands of iterative CC calculations. This creates a prohibitive computational cost, limiting systematic exploration of nuclear shapes.

2. Methodology

The authors propose a surrogate model (emulator) based on Eigenvector Continuation (EC), a specific application of the Reduced Basis Method (RBM), to accelerate CC calculations while maintaining high accuracy.

• Theoretical Framework:

  • Coupled-Channels Formalism: Utilizes the CCFULL code framework. The Hamiltonian includes the relative motion, intrinsic excitation, and coupling potentials (nuclear and Coulomb) dependent on deformation parameters ($\beta_\lambda$).
  • Discrete Basis Method (DBM): The continuous Schrödinger equation is discretized using a finite mesh. The authors employ the Kohn variational principle combined with DBM to transform the scattering problem into a matrix eigenvalue problem.
  • Numerov Method: To enhance stability and convergence without excessive mesh refinement, the authors implement the modified Numerov method for the kinetic energy operator instead of the standard three-point finite difference formula.

• The Emulator Construction (Eigenvector Continuation):

  1. Snapshot Generation (Offline Phase): High-fidelity solutions (scattering wave functions $\Psi_i$) are computed for a small set of selected training parameters (e.g., specific values of $\beta_2, \beta_4$).
  2. Subspace Construction: These solutions form a low-dimensional subspace.
  3. Projection: For any new set of parameters, the wave function is approximated as a linear superposition of the training snapshots:
     $$\Psi_{EC} = \sum c_i \phi_i + \text{boundary terms}$$
  4. Efficient Calculation: The coefficients $c_i$ are determined by solving a small linear system derived from projecting the Hamiltonian onto the reduced basis. This avoids solving the full high-dimensional matrix equation for every new parameter set.

• Workflow:

  1. Define training points in the parameter space.
  2. Generate high-fidelity snapshots (wave functions).
  3. Construct the EC emulator.
  4. Perform rapid $\chi^2$ minimization against experimental data to extract optimal deformation parameters.

3. Key Contributions

• Extension to 3D Scattering: While previous EC applications were limited to 1D models or simple potentials, this work extends the method to realistic 3D heavy-ion fusion reactions involving complex channel couplings.
• Integration with DBM and Kohn Variational Principle: The paper successfully integrates EC with the discrete basis formalism and the Kohn variational principle, providing a rigorous mathematical foundation for scattering emulators.
• Systematic Parameter Extraction: Demonstrates a robust workflow for extracting nuclear shape parameters (both static deformations and vibrational amplitudes) directly from fusion cross-section data using the emulator.
• Numerical Optimization: Introduces the use of the Numerov method within the DBM framework to balance computational efficiency with numerical accuracy, allowing for coarser meshes without sacrificing convergence.

4. Results

The emulator was applied to three specific reaction systems:

• Case 1: $^{16}\text{O} + ^{144}\text{Sm}$ (Octupole Vibration)

  • Goal: Extract the octupole deformation parameter $\beta_3$.
  • Result: The emulator identified the minimum $\chi^2$ at $\beta_3 = 0.21$, perfectly matching the value derived from experimental $B(E3)$ data and exact CC calculations.
  • Accuracy: The fusion cross-sections and barrier distributions calculated by the emulator were indistinguishable from the exact results.

• Case 2: $^{16}\text{O} + ^{154}\text{Sm}$ (Static Deformation)

  • Goal: Extract quadrupole ($\beta_2$) and hexadecapole ($\beta_4$) parameters in a 2D search space.
  • Result: The emulator found optimal parameters $\beta_2 = 0.31$ and $\beta_4 = 0.06$. These values are consistent with exact calculations and independent experimental probes (e.g., $\alpha$-inelastic scattering).
  • Correlation: The emulator reproduced the strong negative correlation ($r \approx -0.82$) between $\beta_2$ and $\beta_4$ found in exact calculations.

• Case 3: $^{16}\text{O} + ^{186}\text{W}$ (Static Deformation)

  • Goal: Extract $\beta_2$ and $\beta_4$ for a nucleus with negative hexadecapole deformation.
  • Result: The emulator yielded $\beta_2 = 0.29$ and $\beta_4 = -0.02$, consistent with exact results and neutron scattering data.
  • Observation: While the optimal means were accurate, the emulator slightly overestimated the anti-correlation between parameters near the boundaries of the training grid, suggesting a need for optimized snapshot selection in future iterations.

• Computational Performance:

  • Speedup: The emulator achieved a speedup factor of 200 to 400 compared to exact DBM calculations for systems with 2–4 channels.
  • Convergence: Numerical errors decreased by roughly an order of magnitude with each additional basis vector ($N_{EC}$).
  • Break-even Point: The emulator becomes computationally advantageous after approximately 5 evaluations (accounting for the offline cost of generating snapshots).

5. Significance and Future Outlook

• Efficiency: The method transforms a computationally prohibitive inverse problem (parameter extraction) into a tractable one, enabling systematic exploration of nuclear shapes.
• Reliability: The emulator maintains high fidelity to the underlying physics, accurately reproducing fusion cross-sections and barrier distributions across the entire energy range.
• Scalability: The framework is readily extensible to more complex deformations (triaxial $\gamma$, higher-order multipoles) and heavier systems (e.g., $^{238}\text{U}$).
• Broader Impact: This approach bridges low-energy nuclear reaction physics with high-energy heavy-ion collision studies (where nuclear shapes are also probed via flow observables), offering a unified tool for understanding intrinsic nuclear properties.
• Future Work: The authors plan to incorporate algorithmic optimization for snapshot selection and Principal Component Analysis (PCA) to further reduce the basis size and handle higher-dimensional parameter spaces.