Modern Nb3Sn strands can
now exceed 3000 A mm−2 critical
current density Jc
at 4.2 K and 12 T within the non-copper area. However, the
aggressive reaction used to achieve this performance causes the
Nb3Sn
filaments to coalesce into a single large, continuous ring of superconductor, and
also allows tin to penetrate through diffusion barriers and alloy with the copper
stabilizer. This results in a lack of adiabatic stability, due to the combination of high
Jc
and large superconductor diameter, and a strong reduction of dynamic stability, due to the
reduction of the copper's thermal conductivity. Under these circumstances, flux jumps at
low fields are inevitable, and the associated heat release could propagate along the
conductor in a quench. In magnets, this means that quenches could be initiated in low-field
regions at currents well below the designed operating current. We show that by limiting the
final reaction duration, it is possible to keep the quench current density above
Jc, thus ensuring flux-jump recovery along the entire magnet load line. For the example
studied, keeping the residual resistivity ratio above
ensures safe operation. This was achieved for final reactions of 40 h or less, instead of the
typical 72–200 h. Surprisingly, the performance penalty was small: a 24 h final reaction reached
>90% of the
highest Jc
obtained. Energy-dispersive spectroscopy in a SEM did not reveal any detectable tin in the
copper for stable strands, but in unstable strands as much as 4% Sn was found in
the copper between sub-elements, suggesting that the contamination is rather
local. The thermal conductivity of the stabilizer should then vary strongly with
distance from the sub-element pack to the strand perimeter, complicating stability
analyses.