Record fast-cycling accelerator magnet based on high temperature superconductors

Four decades ago development of high-current superconducting NbTi wire cables revolutionized the magnet technology for energy frontier accelerators, such as Tevatron, RHIC and LHC. The NbTi based magnets offered advantage of much higher fields B and much lower electric wall plug power consumption if operated at 4.5 K but relatively small ramping rates dB/dt<<0.1 T/s. The need for the accelerators of high average beam power and high repetition rates have initiated studies of fast ramping SC magnets, but it was found the AC losses in the low-temperature superconductors preclude obtaining the rates in the excess of (1- 4) T/s. Here we report the first application of high-temperature superconductor magnet technology with substantially lower AC losses and report record high ramping rates of 12 T/s achieved in a prototype dual-aperture accelerator magnet.

2 resistive power loss in the conductor and magnetization loss in the magnetic cores steel make them prohibitively power inefficient. Operation of the world's largest accelerator complex at CERN requires about 180 MW electric power and three smaller, low-energy normal conducting RCS's altogether boosting the proton energy beam from 50 MeV to 450 GeV consume more electric power than much larger 6500 GeV SC LHC collider ring 15 . Fast cycling SC magnets face great challenges due to the AC losses -energy dissipation in the conductor caused mostly by the magnetization of the superconducting filaments and due to coupling currents between the filaments in the strands. State-of-the-art cryogenic systems require 930 W of wall plug power to provide 1 W of cooling capacity for NbTi SC magnets at 1.8 K, and 230 W/W at 4.5 K 17 and that poses very stringent limits on the allowable AC losses in the low-temperature superconducting (LTS) magnet accelerators. To-date, the highest ramping rates achieved in the operational LTS accelerator magnets are about 4 T/s 18,19 . When comes to the accelerator magnet technology, the high temperature superconductors (HTS) 20 have triple advantage against the LTS based on the NbTi or Nb3Sn superconductor-(i) much higher critical current densities and fields, (ii) lower AC losses and (iii) higher operational temperatures. In this Letter we report the 12 T/s ramp rates achieved in a dual-bore accelerator HTS magnet prototype.
The AC losses in a SC magnet are proportional to the mass of the conductor and depend on total current I, frequency f, maximum field B, and temperature T. The physical mechanisms and scaling of the AC losses in HTS tapes are different for the magnetization losses and for the transport-current losses are discussed in detail in 21,22 . Of importance for the magnet design is that high current density of the HTS superconductors allows to strongly minimize the mass of the conductor and that the AC losses are significantly enhanced by the magnetic field components perpendicular to the tape surface. For typical rapid cycling operation expected in future accelerators the inductive loss component due to self-fields induced by the AC transport current will dominate the AC power loss. The inductance L scales as N 2 with the number of turns and the minimal one can be achieved with a single-turn cable but for the required field B in the gap of the magnet very high current I conductor may be needed as I~B/L. For example, 2 T field has been achieved in the 20 mm double-gap superferric DC magnet with 100 kA current in a NbTi-based single-turn power cable 23 . For the fast-cycling operation, the power supply voltage V grows with dI/dt and can be prohibitively large, therefore, to optimize technical feasibility and cost of such power supply some compromise between the magnet current and the magnet inductance is required.
The rapid-cycling magnet design developed, tested and reported below -see figure 1 -has three novel features: a) it uses the high-current density HTS conductor; b) the conductor is placed inside the steel core of the magnet such that the magnetic field in the conductor is minimal; and c) its two vertically aligned beam gaps are energized by one conductor. The choice of a 3-turn conductor allows to operate the magnet with three times lower current than needed for a single turn option for the same field in the gap: (here g the gap size, μ0 is magnetic permeability of the vacuum) at the expense of acceptable 9fold increase of the inductance. In such arrangement -conceptually proposed in 9 -the magnetic fields in the upper and lower gaps are of the same value but of opposite polarities that makes it uniquely beneficial for simultaneous acceleration of two beams at once -either beams of opposite charge particles (e.g., electron and positrons, positive and negative muons, protons and antiprotons) circulating in the same direction in each of the gaps, or two beams of the same 4 particles circulating in opposite directions. Also of importance for the particle acceleration application is that in such design, the ever existing particle beam losses and decay products which have lower energy than the primary beams will be bent out and away from the HTS conductor, thus, minimizing its highly undesirable heating and therefore, greatly easing the requirements for the particle collimation and radiation protection systems 24 .
FIG. 1. A conceptual design of a vertical dual-bore HTS based accelerator magnet.
The quench propagation velocity in the HTS superconductor is very slow and that makes the quench detection and protection difficult 20,25 . Operation of the HTS conductor at the temperatures much lower than the critical one allows efficient use of the temperature-based quench detection system. This is achieved by having the total cross-section of the HTS superconductor sufficiently large to carry the design transport current up to, e.g., 30 K, so at the operational temperature set to 5 K there is a wide quench safety margin of about 25 K. For example, according to Eq. (1), B=1T 5 field in g=40 mm gap can be achieved with the total transport current of I·N=36 kA that requires the superconductor cross-section of only about 1 cm 2 for the 30 K operations. That surely will be more than enough to operate at 5 K. Figure 2 shows the magnetic field simulation 26  In our previous study 33 , the cryogenic power losses for the cable constructed of twenty 4.2 mm × 0.25 mm YBCO strands exposed to the ramping external fields of dB/dt= (4 -20) T/s at 6.5 K were reliably measured. Also measured for comparison were the losses in the NbTi-based SC cable constructed to carry the same critical current at the same temperature. Figure 6