Post on 28-Dec-2015
2/17/05 Don Hartill, MC05 1
200MHz SCRF cavitydevelopment
Don HartillDon Hartill
LEPP, Cornell UniversityLEPP, Cornell University
2/17/05 Don Hartill, MC05 2
H. PadamseeR. GengP. BarnesJ. SearsV.ShemelinJ. Kaufman
R. LositoE. ChiaveriH. PreisS. CalatroniE. Palmieri - INFNM. Pekelar - ACCELG. Wu - JLAB
2/17/05 Don Hartill, MC05 3
Contents
Fabrication and RF testsFabrication and RF tests
Performance: Eacc and QPerformance: Eacc and Q
Q-slopeQ-slope
Performance when HPerformance when Hext ext 0 0
Future work plan and statusFuture work plan and status
ConclusionConclusion
2/17/05 Don Hartill, MC05 4
Muon-based neutrino source
Acceleration starts after coolingFast acceleration required since
muon has a short life time
2/17/05 Don Hartill, MC05 5
Requirements to acceleration
The highest possible Eacc to minimize muon decayThe highest possible Eacc to minimize muon decay
Large transverse and longitudinal acceptancesLarge transverse and longitudinal acceptances
Both requirements favor the choice of SRF
SRF cavities have a high QSRF cavities have a high Q00
SRF can achieve high gradients with modest RF powerSRF can achieve high gradients with modest RF power
SRF cavities accommodate a larger aperture without a SRF cavities accommodate a larger aperture without a large penalty for the low R/Qlarge penalty for the low R/Q
0
2
)( QQREacc
Pd =
2/17/05 Don Hartill, MC05 6
200MHz SRF layout for Linac
Focusing Solenoid(2-4 T) 2-cell SRF cavity
2/17/05 Don Hartill, MC05 7
200MHz SRF parameter list
300 high gradient 200MHz cavities needed
2/17/05 Don Hartill, MC05 8
Why Nb-Cu cavities? Save material costSave material cost Save cost on magnetic field shielding (Rs of Nb-Cu less Save cost on magnetic field shielding (Rs of Nb-Cu less
sensitive to residual mag. field)sensitive to residual mag. field) Save cost on LHe inventory by pipe cooling (Brazing Cu Save cost on LHe inventory by pipe cooling (Brazing Cu
pipe to Cu cavity)pipe to Cu cavity)
1.5GHz bulk Nb cavity (3mm) material cost: ~ $ 2k/cell200MHz: X (1500/200)2 = 56 $ 112k/cellThicker material (8mm) needed: X 2.7 $300k/cell
Nb Material cost for 600 cells: 180M$ Cu (OF) is X 40 cheaper: 5M$
2/17/05 Don Hartill, MC05 9
First 200MHz Nb-Cu cavity
400mm BT
Cavity length: 2 m
Major dia.: 1.4 m
2/17/05 Don Hartill, MC05 10
Fabrication at CERN
Electro-polished half cell
Magnetron Nb film (1-2 m) sputtering
• DC voltage: 400-650 V• Gas pressure: 2 mTorr• Substrate T: 100 °C• RRR = 11• Tc = 9.5 K
2/17/05 Don Hartill, MC05 11
RF test at CornellCavity on test stand Cavity going into test pit
in Newman basement
Pit: 5m deep X 2.5m dia.
2/17/05 Don Hartill, MC05 12
Two-point Multipacting
• Two points symmetric about equator are involved• Spontaneously emitted electrons arrive at opposite point after T/2• Accelerated electrons impact surface and release secondary electrons• Secondary electrons are in turn accelerated by RF field and impact again• The process will go on until the number of electrons are saturated
MP electrons drain RF power A sharp Q drop
2/17/05 Don Hartill, MC05 13
Two-point MP at 3 MV/m
MULTIPAC simulationconfirmed exp. observation
Resonant trajectory of MP electrons
It was possible to process through MP barrier
2/17/05 Don Hartill, MC05 14
Performance of the cavity
• Eacc = 11MV/m• Low field Q = 2E10
Limited by RF coupler
• 75% goal Eacc achieved• Q-slope larger than expected
Q(Eacc) after combined RF and Helium processing
Q improves with lower T
FE not dominant
2/17/05 Don Hartill, MC05 15
Hext effect on cavity
200MHz cavity
SC Nb/Ti coil
2T solenoid
• 2T solenoid needed for tight focusing• Solenoid and cavity fitted in one cryostat• Large aperture (460 mm)• Q: Will cavity still work Hext > 0 ?
Layout of Linear Accelerator for source
Cavity test in the presence of an Hext
2/17/05 Don Hartill, MC05 16
Hext effect on cavity
Cavity stays intact up to Hext = 1200 Oe
2/17/05 Don Hartill, MC05 17
Hext effect on cavity
• Nb is a type-II SC
• Mixed state above Hc1
• Magnetic flux penetration
• Normal cores cause Rs
• Onset Hext for loss increase consistent with Hc1 of Nb• Msmts at higher Eacc needed: Hext + HRF; resistive flux flow
2/17/05 Don Hartill, MC05 18
Q-slope of sputtered film Nb cavities
Q-slope is a result of Q-slope is a result of material properties of film material properties of film NbNb
The Cu substrate (surface) The Cu substrate (surface) has some influencehas some influence
The exact Q-slope The exact Q-slope mechanism is not fully mechanism is not fully understood understood
Sputtered Nb
Bulk Nb
2/17/05 Don Hartill, MC05 19
Nb-Cu cavities
350MHz LEP cavities 400MHz LHC cavities
Despite Q-slope, sputtered Nb-Cu cavities have achieved a 15MV/m Eacc at 400MHz
Q0(X
1E9)
2/17/05 Don Hartill, MC05 20
Expected performanceProjecting LHC 400MHz to 200MHz
Empirical frequency dependence of Q-slope
200MHz
Measured Q-slope of 200MHz cavity is10 times too steep than projected
2/17/05 Don Hartill, MC05 21
Q-slope: impact angle effect
100mm
R67mm
• CERN explored low 350MHz cavities• With the same cathode geometry, lower low
Impact angle of Nb atom:
2/17/05 Don Hartill, MC05 22
Q-slope: impact angle effect
Correlation: lower lower steeper Q-slope
2/17/05 Don Hartill, MC05 23
Q-slope: impact angle effect A smaller impact angle results in pronounced shadowing A smaller impact angle results in pronounced shadowing
effect and poor film quality (open boundaries, voids, effect and poor film quality (open boundaries, voids, dislocations)dislocations)
The cathode used to sputter 200MHz cavity was recycled The cathode used to sputter 200MHz cavity was recycled from sputtering system for LEP2 cavitiesfrom sputtering system for LEP2 cavities
Due to an increase in equator radius, a smaller impact Due to an increase in equator radius, a smaller impact angle is evident for 200MHz cavityangle is evident for 200MHz cavity
Cavity returned to CERN for recoating with improved Cavity returned to CERN for recoating with improved geometry - expect completion in March - retest 4/05 geometry - expect completion in March - retest 4/05
2/17/05 Don Hartill, MC05 24
Other techniques for Nb film deposition
Bias sputteringBias sputtering Energetic deposition in vacuumEnergetic deposition in vacuum Vacuum arc depositionVacuum arc deposition Electron cyclotron resonance sputtering Electron cyclotron resonance sputtering
2/17/05 Don Hartill, MC05 25
Nb Sputtering Variation
• Standard films have rod like form• Avoid oxide formation • More uniform and larger grains
Standard Films
Oxide-free
2/17/05 Don Hartill, MC05 26
Reducing Q-Slope
Study Nb film with 500MHz cavities (less LHe) with Study Nb film with 500MHz cavities (less LHe) with existing LEPP infrastructure developed for CESR SRFexisting LEPP infrastructure developed for CESR SRF
Seamless Cu cavities to simplify fabrication (Italy)Seamless Cu cavities to simplify fabrication (Italy)
2/17/05 Don Hartill, MC05 27
500 MHz Progress
ACCEL Etching Facility
ACCEL Etching Facility
2/17/05 Don Hartill, MC05 28
500 MHz
ACCEL Sputtering
Setup
ACCEL Sputtering
Setup
2/17/05 Don Hartill, MC05 29
500 MHz Progress
ACCEL Nb Coated Cavity before Final Water Rinse
ACCEL Nb Coated Cavity before Final Water Rinse
2/17/05 Don Hartill, MC05 30
500 MHz
Final Water Rinse after Nb Sputter Coating at ACCEL
Final Water Rinse after Nb Sputter Coating at ACCEL
2/17/05 Don Hartill, MC05 31
2/17/05 Don Hartill, MC05 32
2/17/05 Don Hartill, MC05 33
Recent Program• 500 MHz cavity from ACCEL, assembled and tested twice to 4MV/m with heavy field emission and quench.• Recoat 200 MHz cavity #1 at CERN in 3/04 - peeling observed - recoated again still bad - recoat again and hope to have by early spring.• Use Auger surface analysis system and SIMS to further characterize Nb sputtered surfaces.• Explore effectiveness of Atomic Force Microscopy in characterizing good Nb RF surfaces.
2/17/05 Don Hartill, MC05 34
Near term Program• Electron Cyclotron Resonance Coating R&D work at JLAB under way.•Incorporate the results from these studies into the 500 MHz cavity program.•Spin two 500 MHz cavities from explosion bonded Nb-Cu sheet. Single cell 1300 MHz cavity spun from this material has achieved 40 MV/m accelerating gradient.•Spin two 500 MHz cavities from hot isostatic bonded Nb-Cu sheet.•Bias Sputter coat a spun Cu single cell 500 MHz cavity at ACCEL.
2/17/05 Don Hartill, MC05 35
2/17/05 Don Hartill, MC05 36
2/17/05 Don Hartill, MC05 37
2/17/05 Don Hartill, MC05 38
2/17/05 Don Hartill, MC05 39
2/17/05 Don Hartill, MC05 40
PHI 660 Scanning Auger Micrcoscope (SAM + SIMS )
Sensitive to first 10 - 100 nm
SIMS resultsFor NbO/Nb
2/17/05 Don Hartill, MC05 41
Auger O-depth profile: no distinguishable difference for baked vs. unbaked Nb
Oxigen depth profile
0
10
20
30
40
50
60
-2 0 2 4 6 8 10 12 14 16
Oxygen concentration [%] in large grain samples vs. depth [nm]
2/17/05 Don Hartill, MC05 42
Oxygen Pollution Model
ratio of NbO/Nb signals (SIMS)
0
0.5
1
1.5
2
2.5
3
3.5
4
0 5 10 15 20depth (nm)
S4 BCP 20 min 8C
S1 EP
S3 EP 48 hour bake 100C
factor of 2 difference
SIMS AnalysisGo deeper !
BCP leaves natural oxide + oxygen rich layer
Baking dilutes oxygen rich layer
2/17/05 Don Hartill, MC05 43
2/17/05 Don Hartill, MC05 44
Blue circles – fresh BCPRed squares – after additional 100 C baking
Q-Slope Improvement with 100 C bake on a BCP Cavity
Russian Nb - 500 RRR, no HT, “smoother”
2/17/05 Don Hartill, MC05 45
2/17/05 Don Hartill, MC05 46
Improve Films with Energetic Deposition
2/17/05 Don Hartill, MC05 47
2/17/05 Don Hartill, MC05 48
Schematic drawing of the UHVCA system. The cathode(1) is mounted on a water cooled copper support (2) and inserted in a water cooled stainless steel anode (3). A niobium floating potential electrode(4) prevents the discharge from moving to the bottom part of the system. The magnetic coil (5) confines the hot spot on the flat cathode surface, while magnetic Helmotz coils (6) improve the deposition rate on the sample holder (7). The triggering is provided by a laser beam focused on the cathode through an optical window (8)
2/17/05 Don Hartill, MC05 49
2/17/05 Don Hartill, MC05 50
2/17/05 Don Hartill, MC05 51
Electron’s spiral path in external magnetic field superimposed with perpendicular periodic electrical field
2/17/05 Don Hartill, MC05 52
0.3275”CoolingWater passage
Coil 16 turns total 500-2000A / turn
Substrate or extraction grid
Nb+
RF In
E-beam Hearth
ECR plasma
168mm
ECR Ionization of Nb by Evaporation
2/17/05 Don Hartill, MC05 53
The electric field distribution inside the circular waveguide
2/17/05 Don Hartill, MC05 54
The AFM picture shows a flat, densely packed, niobium thin film on a sapphire substratewith 80 nm grain sizes
2/17/05 Don Hartill, MC05 55
2/17/05 Don Hartill, MC05 56
Optimize magnetic field for ECR inside an elliptical cavity
2/17/05 Don Hartill, MC05 57
Cryopump
magnet
E-gun
Nb grid
Yoke
2/17/05 Don Hartill, MC05 58
10-in conflat
RF window
WR284
Insulation (suggested by Larry)
2/17/05 Don Hartill, MC05 59
Conclusion First 200MHz SC cavities have been constructed.First 200MHz SC cavities have been constructed.
Test results for first cavity -> Eacc = 11 MV/m with QTest results for first cavity -> Eacc = 11 MV/m with Q00 = = 2E10 at low field.2E10 at low field.
MP barriers are present and can be processed through.MP barriers are present and can be processed through.
Cavity performance not affected by Hext < 1200 Oe.Cavity performance not affected by Hext < 1200 Oe.
Next 200 MHz test will include measurements on Hext Next 200 MHz test will include measurements on Hext effect at higher Eacc. effect at higher Eacc.
Making good progess on understanding Q-slope.Making good progess on understanding Q-slope.
Confident that we can build 200 MHz SCRF cavities with Confident that we can build 200 MHz SCRF cavities with Eacc > 17 MV/m.Eacc > 17 MV/m.
2/17/05 Don Hartill, MC05 60
Conclusion Because of diffusion of Cu into Nb, if it is essential to Because of diffusion of Cu into Nb, if it is essential to
have low temperature bake to delute the oxygen rich layer have low temperature bake to delute the oxygen rich layer then bonded Nb sheet to Copper is an attactive solution.then bonded Nb sheet to Copper is an attactive solution.
Cost of 1 mm Nb bonded to 4 mm of Cu is 1/3 that of 5 Cost of 1 mm Nb bonded to 4 mm of Cu is 1/3 that of 5 mm RRR 300 Nb sheet in small quantities. mm RRR 300 Nb sheet in small quantities.
Plan continued effort in developing sputter coated cavities Plan continued effort in developing sputter coated cavities after the end of the current NSF muon contract (9/1/05).after the end of the current NSF muon contract (9/1/05).
2/17/05 Don Hartill, MC05 61