SRF Surface Studies and the High Field Qslope Mystery

1
SRF Surface Studies and the High Field Q-slope
Mystery

• Alexander Romanenko
• Cornell University
• CLASSE

2
Outline

• High field Q-slope (HFQS) in niobium cavities
• Sharp decrease in Q-factor with field
• Empirically found cure baking at 100-120C for
  28-48 hours
• But explanation missing
• Interesting physics
• Understanding will significantly improve the
  current cavity treatment procedure
• Ways to find the explanation
• Cavity tests
• Surface studies
• Models studied
• Oxygen diffusion
• Magnetic field enhancement
• Interface tunnel exchange
• Newer ideas and experiments

3
Q-slope and baking
Real Nb cavities
Ideally
HFQS is eliminated (EP) or improved (BCP) by
vacuum baking _at_100-120C for 24-48 hours
Adapted from B. Visentin (Saclay)
4
High Field Q-slope - History

• Initially HFQS was called European headache
  since cavities chemically polished in Europe had
  it and electropolished and tested cavities at KEK
  (Japan) did not have it
• Difference was due to the process KEK used to
  improve vacuum after high pressure rinsing
  empirical HFQS cure was found by accident
• 100-1400C high vacuum annealing for 24-48 hours
• Removes HFQS in EP cavities
• Improves HFQS in BCP cavities

5
Temperature Mapping
Thermometry board
T-map
Nb cavity on the test stand
Thermometry boards mounted to the cavity
6
Temperature Mapping
dT, mK
log(T)
Hot
Cold
log(Epk)
Epeak 50 MV/m, Hpeak 123 mT

• Non-uniformity of heating
• Some regions (hot) have higher losses Why?

7
Two main directions

• Baking effect
• Study Nb samples treated similarly to cavities
• Heating non-uniformity
• Cut tested Nb cavities and analyze hot/cold
  regions

8
Cavity Measurements vs. Surface Studies

• Cavity measurements macro characterization
• Averaged over the whole surface characteristics
• Not possible to pinpoint the physical entity
  responsible for the HFQS
• Surface studies micro characterization
• Local probing of the surface
• Depth resolution of a few nm is needed

9
Nb Cavity Surface
Hydrocarbons
A few monolayers
Nb2O5
2-5 nm
NbOx, xlt2.5
Nb
40 nm London penetration depth _at_ 2K
10
Most popular model O
Adapted from G.Ciovati SRF07 talk
11
Secondary Ion Mass Spectrometry

• Very sensitive ppb detection possible
• Destructive depth profiling
• BUT
• Instrumental effects preferential sputtering of
  oxygen, roughness effect on signal, chemical
  information not reliable due to
  sputtering-induced ion production

12
ToF-SIMS results
G. Eremeev and J. Francis
5 nm
100 sec 1 nm
13
Arguments against O model

• No evidence of oxygen-enriched layer underneath
  oxide
• No evidence of oxygen diffusion at 100-120C
  baking temperature
• Oxygen depth profile does not change after baking

14
Magnetic field enhancement (MFE)
Chemically polished surface
Electropolished surface
Field enhancement region with enhancement factor
?, if ? H gt Hc the region becomes normal
conducting
15
Arguments against

• BCP and EP cavities behave similarly before
  baking
• But roughness is different
• No evidence of MFE at grain boundary steps from
  T-maps
• Possibly overestimated the field enhancement
  might be a negligible effect
• No difference in roughness between hot/cold spots

16
Oxide losses
Adapted from G.Ciovati SRF07 talk
One of the possible oxide-related loss mechanisms
17
X-Ray Photoelectron Spectroscopy

• Elemental composition within first few nm
  (except for H and He)
• Chemical state information
• Sensitivity limited to about 0.1-1 at.

18
No change in oxide structure
H.Tian, JLab, FNAL Material Workshop, 2007
19
XPS Oxide structure

• Is non-uniformity of heating caused by oxide
  structure differences? No!

Nb5
Hot
Nb0
Valence band
Hot
Cold
Cold
20
Arguments against oxide models

• Oxide structure change after baking is reversible
  
• Namely cancelled by air exposure
• Whereas cavity Q vs. E improvement is preserved
• No difference in the oxide between hot and
  cold spots

21
Other aspects to explore

• Crystalline orientation of individual niobium
  grains
• Deformation (stress)
• Vacancies and dislocations defects of the
  crystalline lattice

22
Electron Backscattered Diffraction
A tool to study crystalline microstructure

• Based on diffraction of backscattered electrons
• Information depth 20-100 nm
• Crystallographic orientation mapping
• Information on crystal defects distribution

23
EBSD Is grain orientation responsible? (Small
grain BCP cavity)
Hot
Cold
24
Recent ideas and methods

• Scattering off magnetic impurities (i.e. Nb12O39)
  T.Proslier et al., App.Phys.Let. in print
• Scanning Tunneling Microscopy (STM)
• Role of dislocations, niobium vacancies and Vac-H
  complexes
• Positron Annihilation Spectroscopy (PAS)
• Electron Backscattered Diffraction (EBSD)
• Detailed studies of the Nb/oxide interface
• Transmission Electron Microscopy (TEM)

25
Conclusion

• High field Q-slope is not yet understood
• Set of possible causes is narrowing down
• Several proposed mechanisms (roughness, grain
  orientation, oxide losses) have been eliminated