Lance Cooley Monday morning meeting 27 September 2010 Review of 1st International Workshop on Superconducting Science and Technology of Ingot Niobium Lance Cooley Monday morning meeting 27 September 2010
Workshop Overview 2 Days plus tutorial 3rd day About 40 attendees CBMM, Tokyo Denkai, Plansee, Cabot , Niowave were represented MSU, NC State, FSU, ODU JLab, Fermilab, IHEP, BARC, DESY, KEK, TRIUMF, NIST 1 World View session 3 Science and Technology sessions 1 Panel Discussion Motivation: Large-grain (LG) niobium may: Be cheaper to use Reduce tendencies for interstitial atom uptake (esp. H) Provide superior thermal conductivity (esp. phonon peak). September 27, 2010 Review SSTIN
Improvement of LG Ingot at Tokyo Denkai Ingot No.1 Ingot No.2 Ingot No.3 Ingot No.3 Single Crystal Improvement of crystallization at Tokyo Denkai Grain grow column like in EBM Make a seed plate for single crystal ingot Ultrasonic image by H.Umezawa Tokyo Denkai
Miscellaneous LG sheets – are we convinced that slicing is a proven technology? September 27, 2010 Review SSTIN
Comments / Take-Aways – Thermal conductivity The correlation between RRR and thermal conductivity is not strong (Chandrasekaran, MSU) High RRR measures a conduction pathway that does not include phonon scattering terms. Is a magnetic pair-breaker important? Phonons scatter off of normal fluid component Wiedemann-Franz “Phonon peak” September 27, 2010 Review SSTIN
MSU Thermal conductivity studies Factor of 5 reduction in b4 RT vs 800 °C Factor of 2 reduction 600 vs 800 °C b1 increases only 2x between RRR 150 and 400 Comment: b3 is sensitive to normal fluid component
Important comments / take-aways - Hydrogen Usually LG cavities have less Q-slope than FG cavities Do we agree? Less hydrogen in LG? Evidence for magnetic transitions in niobium hydride at ~120 K, in particular when Nb is tetrahedrally coordinated with H “NbH4” Hydrogen electromigration: electron beam attracts H to weld (HAZ is hydrogen rich) Different probes and different labs’ data now converge: 10 to 40 percent hydrogen at Nb surface Recoil spectroscopy Induction spectroscopy September 27, 2010 Review SSTIN
Comments / Take-Aways Niobium Pentoxide is a good diffusion barrier for hydrogen Evidence: implantation experiments Implication: must heat above 600 C to dissolve oxide and let hydrogen out from surface (explains why 800 C bake works better) Implication: oxide makes a good passivation layer against hydrogen Implication: niobium is “naked” to hydrogen uptake whenever active fluroine or chlorine is around Electropolishing should repel hydrogen Nb anode is (+) Hydrogen uptake might then come after current is off and HF is present (again, Nb is “naked”) September 27, 2010 Review SSTIN
Why is hydrogen such a big deal? My observations… IIT (not this workshop): electron spin resonance indicates Nb4+ as the active spin leading to magnetic scattering in point-contact tunneling experiments So far we have thought this is associated with a defect in the oxide Is this spin in the metal, especially at niobium (IV) hydride? Nb4+(H-)4 Plausible: Nb = [Kr] 4d4 5s1; Nb4+ = [Kr] 4d1 magnetic Implication: magnetic niobium hydride precipitates Magnetism breaks cooper pairs; dislocations and normal precipitates typically do not. Precipitates therefore increase the normal fluid component Does everything boil down to presence or absence of a certain niobium hydride? Q disease: the whole metal is full of precipitates Q slope: certain regions, e.g. cores of dislocations, have precipitates Thermal conductivity: phonon peak appears when precipitate is absent Implication: removal of hydrogen leads to NO Q drop! Is this what has been seen in ALD coated cavities? Implication: removal of hydrogen may improve Q0 too September 27, 2010 Review SSTIN
Coherent hydride precipitates – further observations A coherent precipitate shares an arrangement of one sub-lattice of atoms with the other phase across its boundaries Several hydrides are possible with very little lattice strain 1/2 cell plane Base 100 plane Tetrahedral void @ 1/4 cell A15 structure NbH3 ½ filling NbHx with x = 2,3,4,6,12 should be possible 110 planes Nb lattice = 3.3 Å; Nb dia = 2.8 Å Tetrahedral void ~ 0.6 Å September 27, 2010 Review SSTIN
Why then does 120 °C bake work? More observations… Romanenko: internal misorientation is reduced after bake, implying that dislocations are healed. Nb vacancies may become mobile to assist dislocations Vacancy mobility is due to breaking of H-vac binding Tetrahedral void is not quite large enough at room temperature to accommodate H Thus, H has to reside in vacancies, dislocation cores, grain boundaries => binding of vacancies H can hop over a limited distance => high enough local concentration to form magnetic precipitates Upon heating: Thermal expansion opens Nb lattice Attempt frequency exp (-E/kT) changes by orders of magnitude Result: very fast turn-on of hydrogen mobility out from vacancies at a modest temperature (~200 C is plausible) Implication: vacancies, dislocations are reservoirs of hydrogen and precipitate formation is likely in nearby regions unless mild baking permits H to move away. September 27, 2010 Review SSTIN
To recap… and plan what we should do about this Nb oxide is a hydrogen diffusion barrier Good if H is outside: prevents penetration Bad if H is inside: traps H underneath skin Concentrated H may form into magnetic coherent precipitates Q disease: precipitates everywhere Q slope: precipitates localized Possible loss of phonon peak Grain boundaries and dislocations are gaps in barrier, permit entry / exit of H (but in local region) Forged sheet has many more opportunities for hydrogen uptake than LG slices New strategy? Form Anneal (recrystallize) Weld Anneal (drive H away from weld) Bulk EP / BCP / tumble Passivation rinse? Fluorine taper? Final EP passivation? Or dry fluorine? Bake 800 C Dry passivation? HPR & Assy 120 bake should not be needed Result: higher Q0 and no Q slope September 27, 2010 Review SSTIN