Conceptual design of a mid-energy spin rotator for continuous-wave operation based on a compact multi-pi rosetta magnet

The paper describes the conceptual design of a mid-energy spin rotator for continuous-wave operation based on a compact multi-pi rosetta magnet. The authors note that spin-rotators in the few-MeV range require large trajectory deflection angles to change the spin direction appreciably, because the spin-precession angle is given by ψ = γaϕ, where a = 1.159 × 10−3 is the gyro-magnetic anomaly of the electron and γ is the beam energy in units of the electron rest mass. At moderate energies, the small values of both a and γ require very large values of ϕ to achieve even moderate values of ψ.

The proposed system is a rosetta magnet inspired by the Rhodotron, which is based on a coaxial accelerating cavity and turn-around dipole magnets. The rosetta magnet omits the accelerating cavity and has all magnets excited equally. The geometry is governed by two parameters: the length L from the center to the entrance point into the magnet (chosen as L = 30 cm) and the half-angle ϕ0 of one leaf, specified by the number N (chosen as N = 4.5), related via ϕ0 = 45°/N. The number of magnets used is given by kmax = 2N − 1, leading to eight magnets. The bending radius is ρ = L tan ϕ0, and the magnet bends the trajectory by θ = 180° + 2ϕ0. For an electron momentum of p = 6 MeV/c, the rigidity is (Bρ) = 0.02 Tm, requiring a field of B0 = 0.378 T. The total accumulated angle is kmax(180° + 2ϕ0) = (2N − 1)(180° + 45°/N), which for the chosen parameters is 4.44 times a full turn.

For the electromagnetic design, a C-type magnet with a gap of h = 2 cm requires N I = B0h/2µ0 = 3000 ampere-turns. Using a conductor with a cross section of 5 × 5 mm2 with a water hole of 3 mm diameter (cross section A = 18 mm2) and a current density of 7 A/mm2, each conductor can carry up to 126 A, requiring 24 turns. A sandwich of five layers with five turns each suffices, though an even number of layers (e.g., 6 × 7) is preferable to reduce the current to about 72 A.

The beam optical properties are analyzed using the high degree of symmetry, allowing analysis of a single leaf which repeats 2N − 1 times. The turn-around magnet provides weak focusing in the horizontal plane, and pole-face rotation angles of magnitude ϕ0/2 are introduced to provide approximately equal focusing in the horizontal and vertical planes. The beta functions outside the dipole are close in both planes, indicating equally divided focusing. A sharp minimum of the horizontal beta function has a geometric origin: two horizontally displaced trajectories with the same radius of curvature cross halfway, corresponding to a focal point and a 180° phase shift. The magnitude of the beta functions at the start and end is approximately equal to the design parameter L.

The integrated bending angle of one rosetta (4.44 turns) only turns the polarization by ∆ψ = 4.44 × 360° × γa = 21.8°. To rotate the polarization by 90°, at least four rosetta magnets are needed. The exiting beam direction is deflected by 180° − 2ϕ0 = 160°, and combining multiple rosettas in a star-shaped formation (with R = 350 cm) allows the beam to traverse up to eight rosettas. Using four rosettas provides a spin rotation angle of 4 × 4.44 × 360° γa = 87.3°. This system works without injection or extraction and supports continuous beams.

Matching between rosettas is achieved using a doublet configuration of quadrupoles with focal lengths around f ≈ ±0.4 m, placed in the straight sections while avoiding the central region where beams cross. The beta functions show a distinct crossing at the center of the straight, which helps match to a generic FODO beamline. The matching section uses two cells with individually variable quadrupoles to adjust βx, βy, αx, and αy to the values in the center of the straight. The dual purpose of the doublet lattice is to act as a final focus system, demagnifying incoming beta functions by more than an order of magnitude to fit the small beta functions in the rosetta, and to connect one rosetta to the next with only four quadrupoles.

The paper also explores advanced options where the beam traverses the same magnet multiple times. These configurations, characterized by different values of N (e.g., N = 8/5 − 0.03 = 1.5700, N = 11/7 − 0.02 = 1.5514, N = 14/9 − 0.01, and N = 23/9 − 0.07), accumulate deflection angles of 9.88, 13.8, 17.87, and 28.84 turns, respectively, with weaker fields (e.g., 0.122 T for the first). These configurations have well-separated in and out sides of the magnets, allowing edge focusing in the vertical plane, but the separation becomes progressively narrower with higher turn counts, making manufacturing more difficult. These advanced options are deferred to a later report.

In conclusion, the system accumulates large deflection angles in limited floor space, supports continuous-wave operation, and can be combined in a star-like structure to achieve the desired 90° spin rotation. The rosetta shown accumulates 4.44 turns in a compact space, and four rosettas suffice to approximately achieve 90° rotation, with the possibility of operating at a different momentum to exactly reach 90°.

Improvements to AI Systems Based on This Paper:

1. AI-Driven Magnet Design Optimization

• Improvement: Train a reinforcement learning or Bayesian optimization model to automatically search the parameter space (N, L, ϕ0, number of turns, field strength) for spin rotators, minimizing floor space and power consumption while achieving exact spin rotation angles (e.g., 90°) under continuous-wave constraints.

• Capability: The AI can propose novel rosetta geometries (e.g., non-integer N, multi-pass configurations) that outperform the manual design, including trade-offs between manufacturing difficulty and beam optics.

2. Autonomous Beam Dynamics Tuning

• Improvement: Implement a neural network surrogate model that predicts beta functions, phase advances, and focal points for arbitrary rosetta configurations, replacing time-consuming particle tracking.

• Capability: The AI can instantly evaluate thousands of designs, identify optimal quadrupole doublet settings (focal lengths, positions) for matching between rosettas, and auto-correct for energy variations to maintain exact spin rotation.

3. Real-Time Spin-Polarization Control

• Improvement: Develop a model-predictive controller using the analytic spin-precession formula (ψ = γaϕ) and measured beam energy, adjusting magnet currents or beam momentum dynamically to compensate for drifts.

• Capability: The AI can maintain a precise 90° spin rotation in a live accelerator, even with fluctuating beam energy, by predicting required field changes and issuing corrections in milliseconds.

4. Multi-Objective Layout Planning

• Improvement: Use a graph-based AI planner to arrange multiple rosettas in star-shaped or other compact formations, optimizing for minimal footprint, beam-crossing avoidance, and serviceability.

• Capability: The AI can generate and rank floor plans for up to eight rosettas, automatically placing quadrupoles and straights while respecting physical clearance and magnetic interference constraints.

5. Automated Manufacturing Feasibility Assessment

• Improvement: Train a vision or rule-based classifier on magnet fabrication data to predict the manufacturability of advanced multi-turn configurations (e.g., N = 23/9 − 0.07) based on pole-face separation and coil winding complexity.

• Capability: The AI can flag designs that are impractical to build, suggest alternative N values with similar spin rotation but easier construction, and estimate cost/performance trade-offs.

6. Digital Twin for Commissioning and Fault Detection

• Improvement: Create a digital twin of the rosetta system using the paper’s analytical formulas (e.g., beta function minima, phase shifts) and train an anomaly detection model on simulated beam loss or field errors.

• Capability: The AI can predict misalignment or coil failure from beam position monitor data, isolate the faulty magnet, and recommend corrective actions without human intervention.

7. Generalizable Spin-Rotator Design Assistant

• Improvement: Fine-tune a large language model (like me) on this paper and similar accelerator physics literature to answer design queries, generate parameter tables, and explain trade-offs in natural language.

• Capability: The AI can serve as an interactive consultant for engineers, instantly providing design options (e.g., “What N gives 90° with four rosettas at 6 MeV?”) and citing relevant equations and constraints.

Spin-rotators in the the few-MeV range require large trajectory deflection angles in order to change the spin direction appreciably. We describe a magnet system that accumulates adequately large bending angles in a small footprint and, at the same time, supports continuous-wave operation.