Status of the superconducting cavity development at RISP.

Gunn Tae ParkAccelerator division, RISP

May 9th. 2014

Friday, May 9, 2014

Contents

1. Introduction2. Design

3. Fabrication

Friday, May 9, 2014

1. Introduction

Friday, May 9, 2014

What is the accelerator?

• An accelerator is a machine that accelerates the charged particles by applying the electromagnetic fields.

• To transfer the energy to the particles, the electric field must be applied along the designated beam line.

• The accelerator is essentially the capacitor.

......

Conductors

E-field

gap

V

Friday, May 9, 2014

• The modern accelerators use the RF (radio frequency) technology and superconductors

• Superconductor introduces the cryogenic system into the accelerator.

V=V0sin(ωt+ϕ)

The frequency reaches radiofrequency (microwave frequency, in particuar) as the particle velocity increases

Exeremely low resistance of the superconductor enables the much more efficient acceleration with the smaller heat loss

Alternating accelerating voltage makes the high energy acceleration available.

Friday, May 9, 2014

Superconducting linac

Power couplerSolid state amplifier

Slow tuner

Helium vessel

Cavity

Beam axis

LHe @ 2 or 4 K

High vacuum @10-9 Torr

Friday, May 9, 2014

RAON: The heavy ion accelerator at RISP

• The design is based on the acceleration of the uranium U+33 and U+34 from 0.5 MeV/u to 200 MeV/u with the current 8.3 pμA.

• For efficient acceleration, charge stripper section is inserted in the midway, dividing driver linac into SCL1 and SCL2.

• For more efficient acceleration, SCL1,2 are further divied into the subsections of SCL11,SCL12 and SCL21, SCL22, respectively.

• SCL3 (Post Accelerator) has the same structure as SCL1.

Friday, May 9, 2014

Driver linac

Injector

SCL11 SCL12

Charge stripper

SCL21 SCL22

IF

• In each subsection, the ions are accelerated by a different kind of the SC with associated nominal beta as determined by beam

dynamics study.

SCL1 SCL2

18 MeV/u 200 MeV/u2.5 MeV/u 56.5 MeV/u0.5 MeV/u

subsection cav.type cav/cm cm no.

SCL11 QWR 1 21

SCL12 HWR 2 13

SCL12 HWR 4 18

SCL21 SSR1 3 23

SCL22 SSR2 6 23

Friday, May 9, 2014

QWR HWR

SSR1 SSR2

Parameters of the resonators

parameters QWR HWR SSR1 SSR2

f (MHz) 81.25 162.5 325 325

βg 0.047 0.12 0.3 0.53

Aperture (mm) 20 20 25 25

Epeak (MV/m) 35 35 35 35

Temp. (K) 4 2 2 2

Superconducting cavities of the RISP

Friday, May 9, 2014

2. Design of the HWR

Friday, May 9, 2014

Performance of the Superconducting cavity: Figures of merit

• Efficient machine, i.e., maximum accelerating gradient with the minimum power supplied.

• The efficiency is characterized by three quantities, i.e.

Q0, R/Q0, TTF

• Once the efficiency is established, one could power up to obtain the maximum gradient, but there is a limit

Epeak, Bpeak

Q0 =ωU

Pwall, R/Q0 = V 2

acc/ωU, T (β) =∫

#E · #vdt∫

#E · d#l

Friday, May 9, 2014

Electromagnetic design

• Beam dynamics study determines the approximate no. of cavities and the accelerating gradients. For example, SCL12 needs ~120

HWR with the accelerating voltage ~1.3 MV.

• The frequency roughly determines the heigh of the cavities

•Beta and the frequency roughly determines the gap to gap distance of the cavities

Hcav =λ

2

d =βλ

2

16 I. 3. Cavity design

3.1 Electromagnetic design

The RF performance of the cavity is characterized by a set of figures of meritand determined by the EM design, which determines the geometry of the cavityoptimizing the figures of merit. As a starting point of the design, one notices thatthe structure of the HWR is based on the coaxial cable modified to form a cavity byshorting the both ends. Thus the performance of the cavity can be approximatedby the analytic formula for coaxial cable.

3.1.1 Preliminary design

The schematic of the HWR cavity is shown as in Fig.2.1 with ZL = 0, ZI = ∞.The input impedance ZI at zi is given by

ZI = Z0ZL + iZ0 tan k(zl − zi)

Z0 + iZL tan k(zl − zi), (3.1)

where Z0 is the characteristic impedance of the cable, ZL is a load impedance at zl,and k is the wave vector of the AC voltage. With the (shorted) loaded impedancefor the cavity ZL = 0, (3.1) becomes

ZI = iZ0 tan k(zl − zi) (3.2)

(3.2) is maximized when h = zl − zi satisfies

h =λ

4(3.3)

Putting λ = 1.85 m, we have 2h = 0.92 m. The gap-to-gapp distance d is determinedso that as the particle moves,

d =βλ

2(3.4)

Putting β = 0.12, we have d = 11.1mm. Similarly for HWR-1 and HWR-2, wehave

h1 = 0.46m, d1 = 0.14m (3.5)

h2 = 0.46m, d2 = 0.25m (3.6)

3.1.2 Optimizations

3.1.2.1 Figures of merit

In the characterization of the efficiency of the cavity, the highest energy transfercan be characterized by the (unloaded) quality factor Q0, the shunt resistance(normalized by Q0) R/Q0, and time transit factor T .

12 I. 2. Theoretical background

xZI ZL

xI xL

Saturday, April 19, 2014

Figure 2.1 The equivalent lumped circuit of the coaxial cable with some loads

Now consider a coaxial line with some boundary conditions imposed by theload impedances, ZL’s.

First when one end x = xL is connected (in parallel) to a load impedance ZL,with the other end being connected to the source, we have

V (xL)

I(xL)=

V +0 eγxL + V −

0 e−γxL

I+0 eγxL + I−0 e−γxL

= ZL (2.16)

We want to find the relation between V +0 and V −

0 in terms of Z0, ZL. Dividingnumerator and denominator by eγxL , and using (2.15), we obtain

→ V +0 + V −

0 e−2γxL

V +0 − V −

0 e−2γxLZ0 = ZL

→ V +0 = −V −

0 e−2γxL(ZL + Z0)

(Z0 − ZL)(2.17)

Then the voltage reflection coefficient Γ is defined to be

Γ =V −

0

V +0

= e−2γz0ZL − Z0

ZL + Z0(2.18)

Further more, we define return loss RL to be

RL = −20 log |Γ|(dB)

= −20 log

∣∣∣∣ZL − Z0

ZL + Z0

∣∣∣∣ + 40γz0(dB) (2.19)

When Γ = 0, i.e., ZL = Z0, the load is said to be matched and the transmissionline is said to be flat. This physically means the voltage is transmitted fully andthere is no reflected wave.

d

Friday, May 9, 2014

• EM design is done by 3D FEA (Finite element analysis) code that optimizes the figures of merit while sweeping the design

parameters

Rbottom sweep Rout sweep

Rtop sweep Rring sweep

Friday, May 9, 2014

Thursday, May 1, 2014

TTF vs. beta

Hcav sweep

Friday, May 9, 2014

design parameter value (mm)

Hcav 920

Router 120

d 100

g 35

Rtop 45

Rbottom 21

Rring 60

Rnose 60

Final specification of the HWR

Wednesday, April 30, 2014

Perspective view of the HWR Design parameters of the HWR

Friday, May 9, 2014

Wednesday, April 30, 2014

Longitudinal field distribution of the HWR

Wednesday, April 30, 2014

Wednesday, April 30, 2014

Electromagnetic fields of the HWR

Electric field distribution Magnetic field distribution

Friday, May 9, 2014

Thursday, May 1, 2014

E-field difference vs. longitudinal distance along the beam axis

• Asymmetry of the transverse (quadrupole) compoent of the E-field was investigated by obtaining Ez-Ey at 10mm away from the

beam axis, which must be zero if the transeverse field were symmetric. The difference is about 1% of the longitudinal

component.

Beam axis

Ez

Ey10 mm

Friday, May 9, 2014

figures of merit value

Q0 4.10E+09

R/Q0 316.2 Ohm

TTF 0.89

Ep 35 MV/m

Bp 52.2 mT

Vacc 1.4 MV

Pw 1.5 W

Figures of merit (HWR)

Optimal figures of merit of the HWR

The peak field values are sensitive to the meshing and thus determined by a largeer number of the meshes with the use of 3

symmetry planes.

Friday, May 9, 2014

Error study Axial component of the accelerating gradient

Beam axis

1% error

1% error

1% error

6 mm

2 mm

6 mm6 mm1 mm

Friday, May 9, 2014

Only one gap (RHS) is deformedTransverse(Vertical) component

of the accelerating gradient

0.1 mm

0.12 mm

10%

10%

Friday, May 9, 2014

Multipacting factor vs. gradient scale factor

Electron source

Multipacting electrons

Multipaction

Schematic of the multipaction

• The enlarging the flat region may spread the electrons disrupting the resonance

• Increased Rtop from 50 mmto 45mm.

Friday, May 9, 2014

Interface to the CouplerIn over-coupling, the power needed to maintain the constant

accelerating voltage is given by

where Ib is the beam current, ϕb is the accelerating phase.

With the bandwidth 2Δf=80 Hz, R/Q0=317,Qext=2.03e6, Ib~0.7 mA, and ϕb, the power is computed as

P=1.5 kW

Coupler antenna

Qext

-1.8 mm Penetratiin depthFriday, May 9, 2014

Mechanical design

Fabrication

Welding

BCP

Cool down

Tuner implementation

Trimming (in clamp-up)

Evacuation

Helium pressure fluctuation

Lorentz detuning

(plastic) Tuning Operation

Clamp-up

Tuesday, March 4, 2014

• As a RF device with a narrow bandwidth, SC is very sensitive to the mechanical deformation.

Friday, May 9, 2014

Trimming/welding shrinkage

•Frequency shift rate=272 kHz/2mm

Trimming of the straight section

Frequency vs. trimmingFriday, May 9, 2014

Polishing

Polishing the inner surface

Frequency vs. etch depth

• frequency shift rate=48.4 kHz/0.1mm• The polishing is done by 0.15mm

Friday, May 9, 2014

Pressure sensitivity

The doubler The gussets

The stiffeners were introduced and

optimized for the minimum deformation against the pressure.

B.C: fixed ports

Frequency shift :0.27kHz/bar

Friday, May 9, 2014

Cool down

The deformation due to cooldown from room

temp. to 2K

Maximum deformation of 1.1mm @

toroids

The first principal stress due to cooldown from

room temp. to 2K

Maximum stress of

533MPa @ beam ports

Frequency shift: 2.7kHz

To overcome the thermal contraction difference between cavity (Nb) and the helium vessel (SS 304L), the bellows were introduced.

As an approximation, fixed b.c applied.

Friday, May 9, 2014

Interface to the tuning system

Friday, May 9, 2014

3. Fabrication of the HWR

Friday, May 9, 2014

Fabrication procedure

Welding

Polishing

Forming

Evacuation

Material Acceptance

Deep drawing, 3D measurement

Machining, (Part) Polishing, (Part) Welding,

Clamp-up test, Final welding, Leak check, RF

test

BCP, RF test, High temperature annealing, Light

etching, Rinsing, HPR

Assembling, Leak check, HPR, Evacuation, RF

test

RF test Low temperature baking, (plastic) RF tuning

Tuesday, March 4, 2014

Friday, May 9, 2014

Wednesday, January 15, 2014

Courtesy of M. Reece

Anomalous behavior of the cavity

• The origins of these anomalous behaviors trace mostly back to the fabrication imperfections.

Friday, May 9, 2014

Microscopic particles Geometrical defects

Pit diameter ~ 400 μm

N, O, S, Fe

Courtesy R.L. GengThe field emitters

Normal conducting impurity Geometrical defects (pit)

The quenchersChemical contaminant

Friday, May 9, 2014

Inspection

Tuesday, April 29, 2014

Grain structure of Nb

DI (deionized) water dipping Rust

Grain size <4 ASTM

RRR >300

Recrystallization 100

Specifications of Nb

Friday, May 9, 2014

Pressing the outer conductor

Forming

Pressing the re-entrant nose

Press jig for re-entrant nose Pressing jig for the upper toroid

Friday, May 9, 2014

Re-entrant nose Ring

Inner conductor

entrantentrant

Outer conductor

Formed parts of the cavities

Friday, May 9, 2014

EBW (Electron beam welding)

The ring fixed in welding jig Frontbead of the ring

Frontbead of the ring

Voltage Beam Size Frequency Focus Distance Feed rate

120 kV 1 × 0.5 4999 Hz -120 mA 641 mm 5 mm/s

Current of weldingCurrent of weldingCurrent of weldingCurrent of weldingCurrent of weldingCurrent of welding

Ø138 Radian (21 mA → 20.5 mA) Ø138 Radian (21 mA → 20.5 mA) Ø138 Radian (21 mA → 20.5 mA) Ø138 Radian (21 mA → 20.5 mA) Ø138 Radian (21 mA → 20.5 mA) Ø138 Radian (21 mA → 20.5 mA)

Welding condition for the ring welding

Frontbead of the outer housing

Backbead of the outer housing

Friday, May 9, 2014

RRR test after welding

9% Reduction at 2x10-5 [At lowest alue]

5% Reduction [Avg value]

22% Reduction [At lowest value] 13% Reduction [Avg value]

234

314

261

3mm Nb with RRR>300

Welding at 2x10e-6 Torr

Friday, May 9, 2014

Clamp-up test of Copeer QWR

Clamp-up w/o indium wire Resonant frequency measurement

Frequency (Mhz) vs. cavity height(mm)

Frequency shift=-71kHz/mm(Simulation)

Before trimming 80.734 MHzAfter trimming (6.5 mm) 81.272 MHzAfter Upper end welding 81.363 MHzAfter Lower end welding 81.32 MHz

Frequency shift=-83kHz/mm(Experiment)

Friday, May 9, 2014

BCP (Buffered chemical polishing)

HF:HNO3:H3PO4=1:1:2 (volume)

• Standard chemical composition

• Etching rate=1 micron/min @ 20C

•We plan to polish the surface by 20 μm before the welding, 150 μm for bulk polishing and again 10 μm for the light etching.

Spoke before BCP Spoke after BCP

Friday, May 9, 2014

4. Test in preparation......

Friday, May 9, 2014

Bakcups

Friday, May 9, 2014

Lorentz detuning

Friday, May 9, 2014

Brazed ports

Friday, May 9, 2014