High-Q, High Gradient Niobium-Coated Cavities at CERN Sergio Calatroni CERN, 1211 Geneva 23, Switzerland C. Benvenuti P. Darriulat M.A. Peck A.-M. Valente C.A. Van’t Hof
Outline of the talk The Nb/Cu technology and motivation for the study The coating procedure RF and material characterization Optimization of the residual residual resistance High gradient operation Conclusions Sergio Calatroni - CERN
Copper surface preparation 1 Copper surface preparation 1.5 GHz spun copper cavities have been used throughout the study. Hydroformed and electroformed cavities have also been tested. Electropolishing parameters: 55% vol. H3PO4 45% vol. butanol Current density: 0.2A/cm2 HP water rinsing Electropolishing has replaced chemical polishing (LEP standard) for surface preparation. Advantages: lower roughness, absence of defects Sergio Calatroni - CERN
Niobium sputtering Sputtering parameters: Discharge current stabilized at 3 A. Sputter gas pressure of 1.5x10-3 mbar, corresponding to ~ 360 V. Coating temperature is 150 °C. Thickness: 1.5 µm LEP “standard film” characteristics: RRR: 11.5 ± 0.1 Argon content: 435 ± 70 ppm Grain size: 110 ± 20 nm Tc: 9.51 ± 0.01 K Strain: a/a = 0.636 ± 0.096 % Sergio Calatroni - CERN
Change of the sputter gas Sputter gas ions are back-scattered by the cathode with an energy proportional to (MNb-Mgas)2/ (MNb+Mgas)2 (for normal incidence) Sergio Calatroni - CERN
Change of the sputter gas Argon-Neon mixtures allow to fine-tune the RRR between 5 and 15 Sergio Calatroni - CERN
Change of substrate Various substrates have been explored: Copper (covered with natural oxide) Oxide free copper (by sputter etching prior to coating) Niobium Oxide free niobium And various underlayers (sputter-coated): Copper Titanium Gold Aluminium Oxidised aluminium Niobium films on most substrates or underlayers behave as standard LEP films. However, some films belong to a second “family”. In the case of oxide free copper: - RRR: 28.9 ± 0.9 - Argon content: 286 ± 43 ppm - Grain size: >1 µm - Tc: 9.27 ± 0.09 K - Strain: a/a = 0.466 ± 0.093 % Sergio Calatroni - CERN
Standard films and films on oxide-free copper Difference of the texture of the films Sergio Calatroni - CERN
The surface resistance The surface resistance can be written in the form: Rs (HRF, Hext, T) = RBCS (HRF, T) + Rfl (HRF, Hext, T) + Rres (HRF) The dependence of RBCS (0,T) on has been verified by changing the sputter gas RBCS (HRF, T) has an intrinsic dependence of HRF Rfl (HRF, Hext, T) has a dependence on similar to RBCS (0,T) Sergio Calatroni - CERN
Theoretical and experimental BCS resistance at zero RF field 1 10 350 400 450 500 550 600 650 700 750 800 850 900 2 5 20 RBCS (4.2K) [n] 1+0/2 RBCS at 4.2 K Nb bulk: ~900 n Nb films: ~400 n RBCS at 1.7 K Nb bulk: ~2.5 n Nb films: ~1.5 n Sergio Calatroni - CERN
The RF field dependence of the BCS resistance RBCS (4.2K, HRF) for film and bulk has a similar slope, and can be written as R0BCS (T)•G(HRF). The G(HRF) term can be approximated with a quadratic dependence on the RF field. 100 1000 10 20 30 40 50 RBCS (4.2 K) [n] HRF [mT] Squares: Nb bulk Rounds: Nb films G(HRF)(1+5.10-4HRF2) or G(ERF) (1+0.01ERF2) Sergio Calatroni - CERN
The RF field dependence of the BCS resistance The value of RBCS (1.7K, HRF) for films can be calculated on the basis of the previous formulas. 10 9 11 5 15 20 25 30 35 40 Q (1.7 K) E acc [MV/m] RBCS 20 n at 35 MV/m Sergio Calatroni - CERN
Fluxon-induced losses Fluxon-induced losses at 1.7 K are characterized as Rfl = (Rfl0 + Rfl1 HRF) Hext The minimum values are obtained using krypton as sputter gas: Rfl0 = 3n/G Rfl1 = 0.4 n/G/mT 1 10 100 1+0/2 Rfl0[n/G] (a) 1 10 Rfl1[n/G/mT] 1+0/2 (b) Triangles: bulk Nb Squares: coatings on oxide-free copper Circles: coatings on oxidized copper Sergio Calatroni - CERN
Fluxon-induced losses Niobium films can be divided into two families, according to the substrate Coatings using argon Oxide-free copper family Oxidized copper family Sergio Calatroni - CERN
Fluxon-induced losses Fluxon-induced losses could be a limitation for high RF-field operation of film cavities, unshielded from the Earth’s magnetic field. 10 9 11 5 15 20 25 30 35 40 Q (1.7 K) E acc [MV/m] Fluxon losses for krypton calculated for Hext = 0.5 G ~32 n at 35 MV/m BCS + Fluxon losses ~52 n at 35 MV/m Sergio Calatroni - CERN
The residual resistance The BCS surface resistance and the fluxon-induced losses have a well determined dependence on the electron mean free path . The residual resistance has no correlation with the mean free path , or any other superconducting parameter. The residual resistance has been seen to be influenced by: The quality of the copper surface: roughness and defects The oxide at the Nb/Cu interface The hydrogen content of the niobium film Sergio Calatroni - CERN
The residual resistance: definitions 200 400 600 800 1000 20 40 60 80 100 Rs [nOhm] RF field [mT] The residual resistance is calculated by subtracting the BCS contribution to the measured 1.7 K surface resistance Sergio Calatroni - CERN
The residual resistance: effect of roughness 5 10 15 20 25 30 40 60 80 100 R res 1 (n W m MV -1 ) Standard coatings using argon as sputter gas Circles: spun cavities Triangles: hydroformed cavities Crosses: electroplated cavities Spun cavities: 0.2 µm Hydroformed cavities: 0.8 µm Electroplated cavities: 0.2 µm Spun cavities Hydroformed cavities Rres0 17±3 n 28±6 n Rres1 5±1 n m MV-1 10±2 n m MV-1 Sergio Calatroni - CERN
The residual resistance: effect of electropolishing Standard coatings using Xe, Kr, Ar, Ne as sputter gas Average roughness of chemically polished spun cavities: 0.2µm chemically polished hydroformed cavities: 0.8µm electropolished spun cavities: 0.04µm Absence of defects (etching pits) Sergio Calatroni - CERN
High gradients: High Pressure Water Rinsing HPWR influences Rres0 Improved HPWR extends the operating range Sergio Calatroni - CERN
The new HPWR station 100 bar HPWR Closed cycle operation Before and after coating Sergio Calatroni - CERN
High gradient cavities Coatings performed using krypton Rinsed with upgraded HPWR installation Sergio Calatroni - CERN
Subtracting the BCS contribution, we see that this is not yet a limit Sergio Calatroni - CERN
High gradient cavities Subtracting the BCS contribution 10 9 11 5 15 20 25 Q E acc [MV/m] (1.7K) residual BCS Sergio Calatroni - CERN
Conclusions We built on the successful LEP technology. RBCS has an intrinsic slope, similar for bulk and for films, which may be a future limitation at high fields. Fluxon-induced losses can be described in terms of . Coatings produced using krypton provide the lowest losses. Rres is mostly influenced by the quality of the substrate, and very small values of Rres0 and Rres1 can be obtained by electro-polishing the copper substrate. High accelerating fields can be reached with a more effective HPWR. No evidence for any fundamental limitation has been found. Sergio Calatroni - CERN