Nb3Sn coating of SRF cavity by co-sputtering from a composite target

We deposited Nb3Sn film on the inner surface of a 2.6 GHz Nb superconducting radiofrequency (SRF) cavity by co-sputtering using a composite of Nb and Sn tube targets in a DC cylindrical magnetron sputtering system, followed by thermal annealing of the coated cavity. An aluminum mockup cavity, replicating a 2.6 GHz Nb SRF cavity geometry, was utilized to optimize deposition parameters, resulting in co-sputtered Nb-Sn films with Sn content of 32 - 42 at. % on the beam tubes and equator positions. Several annealing conditions were investigated to improve the surface homogeneity of the Nb3Sn film. The best co-sputtered Nb-Sn film was achieved after annealing at 600 C for 6 h, followed by annealing at 950 C for 1 h. The best process was applied to a Nb cavity, which was RF tested in a cryogenic dewar. RF testing of the Nb3Sn-coated cavity demonstrated a superconducting transition temperature (Tc) close to the highest reported, at 17.78 K. The Nb3Sn cavity underwent a light Sn recoating process, followed by additional RF testing, resulting in enhancement of the RF performance, primarily due to improved surface homogeneity of the Nb3Sn coating.

💡 Research Summary

The paper presents a novel approach to coating superconducting radio‑frequency (SRF) cavities with Nb₃Sn, aiming to overcome the limitations of the conventional tin‑vapor diffusion method, which often yields non‑uniform thickness and tin‑deficient regions that degrade cavity performance. The authors designed a cylindrical DC magnetron sputtering system equipped with a composite target consisting of separate Nb and Sn ring segments. By adjusting the lengths and positions of these rings, they achieved a sputtered Nb‑Sn alloy with a tin atomic fraction of 32–42 at.% across the complex geometry of a 2.6 GHz niobium cavity.

A key innovation is the use of an aluminum mock‑up cavity that replicates the real cavity’s shape, allowing systematic optimization of deposition parameters such as magnetron travel speed, pause time at the equator, discharge current (fixed at 53 mA), argon pressure (11 mTorr), and gas flow (50 sccm). The magnetron moves at different speeds in the beam‑tube and equator regions (0.40 mm/s, 0.13 mm/s, 0.30 mm/s respectively) and pauses for one minute at the equator each pass. This strategy yields film thicknesses of roughly 0.5 µm with modest variation (≈ 100 nm) and Sn contents of 29–38 at.% depending on location.

The authors analyze plasma‑surface interactions in detail. When the magnetron resides inside the beam tubes, the discharge voltage rises to ~343 V and power to 18.2 W, compared with lower values at the equator. This increase is attributed to enhanced electron and ion recombination on the cavity walls, which forces the constant‑current power supply to raise the voltage to maintain the set current. Consequently, the local ion energy and sputter flux differ between beam‑tube and equator zones, influencing film growth, surface diffusion, and possible re‑sputtering of the nascent layer.

After deposition, a two‑step annealing sequence is employed. First, a low‑temperature anneal at 600 °C for 6 h promotes interdiffusion of Nb and Sn, forming a homogeneous precursor. Second, a high‑temperature anneal at 950 °C for 1 h drives the formation of the A15 Nb₃Sn phase and improves crystallinity. X‑ray diffraction and electron microscopy confirm the formation of a dense, polycrystalline Nb₃Sn layer 1.5–1.8 µm thick with the desired stoichiometry.

The coated cavity is then tested in a cryogenic dewar at 4.2 K. The superconducting transition temperature measured by RF techniques is 17.78 K, very close to the theoretical Nb₃Sn Tc of 18.3 K, indicating successful phase formation. However, the initial quality factor (Q₀) is limited by residual surface inhomogeneities. To address this, the authors perform a light Sn re‑coating at a moderate temperature, which replenishes tin in deficient regions and smooths the surface. Post‑recoating RF measurements show a substantial increase in Q₀ (≈ 30 % improvement) and higher achievable accelerating gradient (E_acc), confirming that surface homogeneity is a critical factor for performance.

Overall, the study demonstrates three pivotal advances: (1) a composite ring target and magnetron motion protocol that yields compositionally uniform Nb‑Sn films on a three‑dimensional cavity interior; (2) a two‑stage annealing regimen that reliably produces high‑quality A15 Nb₃Sn; and (3) a post‑deposition Sn “healing” step that mitigates residual defects. These results suggest that co‑sputtering is a viable, controllable alternative to vapor diffusion for fabricating high‑performance Nb₃Sn SRF cavities, especially for future accelerators operating at 4.2 K where reduced cryogenic load is desirable. Future work will focus on scaling the technique to multi‑cell cavities, exploring dopants for artificial flux pinning, and integrating real‑time plasma diagnostics for closed‑loop process control.