superconducting rf cavity preparation and testing...srf cavity testing: the challenge 1. cool down...
Post on 05-Aug-2020
6 Views
Preview:
TRANSCRIPT
1Matthias Liepe April 18, 2008
Matthias LiepeMatthias LiepeMatthias LiepeMatthias LiepeMatthias LiepeMatthias LiepeMatthias LiepeMatthias Liepe
Department of Physics, CLASSEDepartment of Physics, CLASSEDepartment of Physics, CLASSEDepartment of Physics, CLASSEDepartment of Physics, CLASSEDepartment of Physics, CLASSEDepartment of Physics, CLASSEDepartment of Physics, CLASSE
Cornell UniversityCornell UniversityCornell UniversityCornell UniversityCornell UniversityCornell UniversityCornell UniversityCornell University
Superconducting RF Cavity Preparation and Testing
2Matthias Liepe April 18, 2008
Outline
• Why do we test superconducting RF cavities?
• How do we test SRF cavities?
• How do we make and prepare SRF cavities?
• What do we find?
3Matthias Liepe April 18, 2008
Why do we test superconducting RF
cavities?
11
10
9
8
0 25Accelerating Field
Ideal
10
10
10
10
Q0
50 MV/m
1970s
4Matthias Liepe April 18, 2008
• Superconducting RF cavities look simple…
• … but making a good cavity is not simple at all
– Took 30+ years to learn how to prepare the surface of Niobium cavities for highest RF fields
– Fabrication and surface preparation involve a long list of critical steps = “recipe”
“Understanding” SRF Cavities
5Matthias Liepe April 18, 2008
Science and Art
• How did we arrive at this “recipe”?
– Science: Understanding superconductors in
high fields at microwave frequencies (GHz)
– Art: working in clean rooms…
– Persistence: Performance tests of 100’s of
cavities to find out what we got…
– Luck: found that drying cavities at 100 C not
only helps to save time, but also reduces the RF
surface resistance at high fields dramatically…
6Matthias Liepe April 18, 2008
SRF Cavity Performance Tests• Goal: Measure RF surface resistance of the cavity
wall as function of RF field gradient and temperature
• Typical results:
1.5 GHz
Exponential drop
Residual resistance
Q0 = intrinsic quality factor ∝∝∝∝ 1 / Rs
good1.6 K
2.1 K
Bsurface, peak [mT]
Q0
7Matthias Liepe April 18, 2008
How do we test SRF cavities?
8Matthias Liepe April 18, 2008
SRF Cavity Testing: The Challenge
1. Cool down SRF cavity below TC to make it superconducting (usually ≤≤≤≤ 2K)
2. Couple RF power into the cavity to excite RF fields in the cavity at certain frequency (“mode”)
3. Keep field amplitude constant to measure power dissipated in the cavity walls at this field level (this gives as the intrinsic Q0)
4. Increase power to increase cavity field, measure dissipated power again…
9Matthias Liepe April 18, 2008
Cavity Test Stand
• After fabrication and surface treatments, a
SRF cavity is mounted on a test stand,
evacuated and immerged into LHe.
10Matthias Liepe April 18, 2008
Field Excitation
• RF power from CW or pulsed power sources is
coupled into the cavity to excite EM fields in
the cavity (at GHz frequencies, 10’s of MV/m).
RF Power
11Matthias Liepe April 18, 2008
Keeping the Field Stable…
1,299,999,500 1.3e9 1,300,000,100
cavi
ty fi
eld
[arb
. uni
ts]
drive frequency [Hz]
• SRF cavities are oscillators with extremely high quality factors
of 1010 to 1011 !
• Width of resonance curve at GHz frequencies is 0.1 to 0.01 Hz!
• To keep field amplitude constant, need to drive oscillator on
resonance ⇒⇒⇒⇒ drive needs to follow cavity resonance!
Lorentz forced on the cavity walls
deform cavity and thereby shift the
resonance frequency ∆∆∆∆f∝∝∝∝ E2
12Matthias Liepe April 18, 2008
…with a Feedback Loop
• The cavity is driven with constant amplitude (RF power)
• The drive frequency is adjusted to follow the natural cavity frequency
• This is done by a phase-looked-loop (PLL)
13Matthias Liepe April 18, 2008
The Phase-Locked-Loop
Cavity field signal
Forw
ard power signal
IF signalMixer as phase detector:
X x)sin( 222 ϕω +tV)sin( 111 ϕω +tV
{ }[ )()(cos2
1212121 ϕϕωω −+−= tVVIF
{ }])()(cos 2121 ϕϕωω +++− t
Filter high f component
ω1 = ω2 (driven oscil.)
{ })(cos2
12121 ϕϕ −= VVIF
tan ϕϕϕϕ = 2Q ∆∆∆∆f / f
-10 0 10-100
-50
0
50
100
∆ f [Hz]
∆ϕ[d
eg]
14Matthias Liepe April 18, 2008
Temperature Mapping• 100’s of temperature sensors are used to map the
distribution of the losses in the cavity walls with mK resolution.
15Matthias Liepe April 18, 2008
T-Mapping: A powerful Tool
• Allows to distinguish field limiting effects
• Gives “local” Q(E) curves
1200 mK0
Local defect ⇒⇒⇒⇒ quench
Electron field emission
J. Knobloch et al.
16Matthias Liepe April 18, 2008
Quench Location Detection with Second Sound in superfluid Helium (I)
• Second sound waves in superfluid Helium: The normal and superfluid components oscillate in counter flow leaving stationary (to first order) the center of mass.
• Measure second sound waves from heat at cavity quench location with oscillating superleak transducers (porous membrane is driven by the normal-fluid component of the wave)
0=+ nnss vvrr ρρ
17Matthias Liepe April 18, 2008
Quench Location Detection with Second Sound in superfluid Helium (II)
π−π−π−π−Mode 1.96K
-0.2
-0.1
0
0.1
0.2
0.3
0.4
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1
Time (s)
Vol
ts
@ T = 1.96 Kvs = 17.8 m/s
∆t = 13.2 ms∆l = 23.5 cm
∆t = 18.3 ms∆l = 32.6 cm
Time [s]
Tra
nsd.
sig
nal
oscillating superleak
transducers
D. Hartill, Z. Conway et al.
18Matthias Liepe April 18, 2008
How do we make and prepare SRF
cavities?
19Matthias Liepe April 18, 2008
The Fabrication of an SRF Cavity (I)
Sheets are scanned (eddy current; measures change of electric resistance) to check for foreign material inclusions (40 µm defect diameter sensitivity)
Electron beam melting on Niobium (done several times to purify Niobium)
Rolling, annealing, levering,
…gives Nb sheets
20Matthias Liepe April 18, 2008
The Fabrication of an SRF Cavity (II)
Electron beam
Half-cells are deep drawn and welded
together in vacuum with an electron beam
21Matthias Liepe April 18, 2008
1400 C Bake with Ti-Getter
• Thermal breakdown (quench) is usually triggered by a small normal conducting defect, when it heats the Nb above the critical temperature (100µµµµm defect sufficient!)
• Tolerate unavoidable defects but “neutralize” them by thermally stabilizing them.
�Improve the thermal conductivity of niobium.
�Improve puritypurity of the niobium.
22Matthias Liepe April 18, 2008
Thermal Breakdown
CuNo foreign materials found
Tungsten inclusion
copper inclusion
Breakdown field given by(very approximately):
Htb =4κ T (Tc − Tb )
rd Rd
κT: Thermal conductivity of NbRd: Defect surface resistanceTc: Critical temperature of NbTb: Bath temperature
Scales as κT
J. Knobloch et al.
23Matthias Liepe April 18, 2008
Thermal Breakdown
Scales as κT• After cavity is produced
– Heat in vacuum furnace to ~ 1400 C
– Evaporate Ti on cavity surface
– Use titanium as getter to capture impurities that diffuse to the surface
– Later etch away the titanium
– Doubles the purity
24Matthias Liepe April 18, 2008
Thermal Breakdown and Heat Treatments
DESY results
25Matthias Liepe April 18, 2008
Surface Preparation: Etching/Polishing
• Removes damaged surface layer (100 µµµµm)
Chemically etchingBCP = HF + HNO3 + H3PO4
Electro-polishingBCP = HF + HNO3 + H3PO4
26Matthias Liepe April 18, 2008
Chemically Etching vs. Electro-Polishing
• Electro-polished cavities each (often) higher field gradients (but not always)
• Difference from surface roughness? Likely not…
Chemically etched
Electro-polished
27Matthias Liepe April 18, 2008
High Pressure Rinsing and Clean Rooms
• All cavities and vacuum components are cleaned and assembled in clean rooms.
•• Dust particlesDust particles on the cavity surface are removed with up to 1000 psi ultra-pure water jets (High Pressure Rinsing)
28Matthias Liepe April 18, 2008
Electron Field Emission (I)
Micron size particles cause FE.
• Emission of e- (QM electron tunneling) from µm size defects
in high E-fields.
• All emission is associated with (conducting) microscopic
particles.
• Acceleration of electrons drains cavity energy.
• Impacting electrons produce heating of the surface.
29Matthias Liepe April 18, 2008
Electron Field Emission (II)
• QM tunneling theory predicts exponential Fowler−−−−Nordheim emission current density.
• Need GV/m fields!
• Fields in cavities are much lower than those
theoretically required for field emission.
• Electric field enhancement model (tip-on-tip)?
jFN = C1E2 exp −
C2
E
30Matthias Liepe April 18, 2008
Before and After High Pressure Rinsing
before
after
31Matthias Liepe April 18, 2008
High Power Processing
• In some cases
applying of high
power can cause the
destruction of field
emitters and improve
the cavity
performance.
• ���� Reduction of field
emission after the
cavity is installed in
the accelerator
before
after
32Matthias Liepe April 18, 2008
A final low Temperature Bake• In-situ baking of the cavity at low temperatures
(100 – 130 C) for 50 hours is good
– Reduces the low field BSC surface resistance by 50%
– Often allows to achieve higher maximum fields and lower surface resistance at high fields
• Why??? Many models…nothing conclusive
G. Ciovati et al. KEK results
33Matthias Liepe April 18, 2008
What do we find?
34Matthias Liepe April 18, 2008
State of the Art Cavity Performance
109
0 10 20 30 40E acc [MV/m]
Q0
1011
1010
Low field Q-dropMedium field Q-drop
Sharp drop of the quality
factor at B ≅≅≅≅100 mT
Recovered by low temp. (110 C) “in-
situ” baking
Linear BCS + residual
resistance. Hypersoundgeneration
and acoustic resonances?
Vortex penetration at
weak spots, grain boundaries,
surface roughness, oxide
diffusion, pollution layer ?
“Nonlinear”BCS
resistance. Heating and
non-equilibrium
effects?
35Matthias Liepe April 18, 2008
Wall Heating Distribution at High Fields
• See areas with increased heating / surface resistance at high magnetic RF fields (hot spots)
• ??? G. Eremeev et al.
36Matthias Liepe April 18, 2008
Recorded Intrinsic Quality Factor
37Matthias Liepe April 18, 2008
Record Field Gradient (2007 @ CU)
• Accelerating gradient = 60 MV/m
R.L. Geng et al.
top related