Chiara Tarantini1 Steve Heald2 Peter Lee1 Michael Brown1 Arup Ghosh3 David Larbalestier1

1, Florida State University, Tallahassee, Florida, United States
2, Argonne National Lab, Argonne, Illinois, United States
3, Brookhaven National Laboratory, Upton, New York, United States

To meet Jc requirements for the Future Circular Collider (FCC) the high-field performance of Nb3Sn strands must improve. In the last decade the best in-field performance of Nb3Sn strands has been obtained by doping with Ti or Ta, however, these wires are far from homogeneous in phase or properties. It has been demonstrated that Ti doping can produce a more chemically homogeneous A15 layer in internal Sn wires and that it significantly narrows the distribution of properties (Tc, Hc2) [1]. However this occurs in a still significantly suppressed stoichiometry (23.1 and 23.4 at.%Sn in Ti and Ta doped wires, respectively). Moreover, the effects of doping is not fully understood: early studies suggested that both dopants preferentially substitute on the Nb site. But more recently, taking into consideration ternary compositions and the different effectiveness of Ti (1-2 at%) and Ta (3-4 at%) in changing the resistivity and obtaining the maximum Hc2, it was suggested that Ti might substitute on the Sn site. In order to resolve this conundrum and point to new ways to improve the Nb3Sn performance, we recently performed an Extended X-ray Absorption Fine Structure (EXAFS) study at the APS synchrotron in Argonne [2]. The EXAFS technique is sensitive to the local environment of a specific element in the structure and so it allows us to identify the site location of the dopants in modern high-performance Nb3Sn strands. We investigated Ti, Ta and Ta+Ti doped internal Sn wires and we found that, whereas Ti exclusively occupies the Nb site in the A15 structure, Ta unexpectedly occupies both sites. In particular we have determined that about 21-32% of Ta resides on the Sn site in the Ta-doped and the Ta+Ti doped wires. With such occupancies the (Nb+Ti):Sn ratio becomes strongly off-stoichiometry (~3.3-3.4) whereas the Ta and Ta+Ti doped samples become almost stoichiometric (~2.9-3.1), thus we have the paradox that the Ti-wires, which have the best performance of accelerator Nb3Sn wires, are the most off-stoichiometric.
Combining those findings with magnetic and specific heat characterization, we will discuss how the different occupancy affects the disorder and, as a consequence, the Hc2 behavior and the effect of charge doping in changing the A15 phase DOS. We will also discuss how this understanding could help to redesign an improved internal Sn strand for higher field application such as the 16T FCC project.
[1] C. Tarantini et al., Appl. Phys. Lett. 108, 042603
[2] S. M. Heald, C. Tarantini et al., in preparation.
APS supports: the US DOE-BES, the CLS and its funding partners, the Advanced Photon Source, the US DOE under Contract No. DE-AC02-06CH11357. NHMFL partial supports: the US DOE-SC, HEP under Award No. DE-SC0012083 and the NHMFL supported by NSF Coop. Agr. No. DMR-1157490 and the State of Florida.