Progress on investigation of dynamic vacuum (RF structures only) Sergio Calatroni With Cedric...

Progress on investigation of dynamic vacuum(RF structures only)

Sergio CalatroniWith Cedric Garion, Chiara Pasquino, Pedro Costa Pinto, Mauro Taborelli

Outline

• Requirements from beam dynamics• Static vacuum• Dynamic vacuum

– Breakdowns– Dark current

• Experimental approach– ESD measurements– Direct measurement

• Outlook and conclusions

IWLC 2010 Sergio Calatroni - 21.10.2010

Requirements from beam dynamics

• The dynamic vacuum threshold for preventing fast ion beam instability (essentially do to direct field ionization and not the usual impact ionization) is:

H2 and CO2 partial pressures < 10-9 Torr

• This happens for practically all the main LINAC length, inside the RF accelerating structures

• For details see G. Rumolo, J-B Jeanneret, D. Schulte and C. Garion (this workshop)


http://indico.cern.ch/contributionDisplay.py?contribId=0&confId=85692

http://ilcagenda.linearcollider.org/contributionDisplay.py?contribId=222&sessionId=83&confId=4507

Tools for dynamic vacuum calculations

• Several tools have been developed for the calculation of dynamic vacuum:

– Thermal analogy implemented in FEM by Cedric Garion– Monte-Carlo simulations implemented in CASTEM (FEM

code) by Cedric Garion (CLIC09 Workshop)– Electrical analogy implemented in PSpice, with conductance

of single elements calculated by Monte-Carlo by Pedro Costa Pinto (TS Workshop 08)

– Analytical models by Volker Ziemann (Uppsala University)

• All tools have been crosschecked and agree within 10%


http://indico.cern.ch/contributionDisplay.py?contribId=154&confId=45580

https://edms.cern.ch/document/936540/1

CATIA Model

Calculations are based on the equivalence of vacuum and thermal conductances, which are evaluated by FE analysis, and using the following outgassing values (H2O):

10-10 mbar.l.sec-1cm-2 for 100h pumping

10-11 mbar.l.sec-1cm-2 for 1000h pumping

Does not take into account re-adsorpion

Static vacuum – thermal analogy


J INRAlexandre.Samochkine @ cern.ch

33

CLIC MODULES

High Order Mode damping

TOWARDS TO FINAL DESIGN

3D CATIA Model

From: A. Samoshkin

FE code model

Pumping surfaces

One word on FE analysis based on thermal/vacuum analogy (1D model)

Static vacuum

Tkqt

Tcp

2

2

x

Pca

t

PS

Heat transfer equation:

Gas flow equation (1D):

Static vacuum - Results


8x10-9 ÷ 1x10-8 mbar

Based on outgassing (H2O): 10-10 mbar.l.sec-1cm-2 for 100h pumping

2.1012 H2 or CO molecules released during breakdown (in a baked system)

Data measured in DC “spark test” reported in PRST-AB12, 092001 (2009)

Dynamic vacuum I - Breakdown


FE code model

Pumping surfaces

Calculated with thermal analogy model (and Monte-Carlo model)

Four manifolds

Longitudinal pressure distribution in the cells:

Uniform after ~3 ms

Calculated with Monte-Carlo and thermal analogy model

Maximum pressure vs time:

20 ms to reach 10-9 mbar for H2

20 ms to reach 5x10-8 mbar for CO

(Note: CLIC repetition rate = 50 Hz

Duty cycle = 20 msec)

Dynamic vacuum I – Results

1.00E-06 1.00E-05 1.00E-04 1.00E-03 1.00E-02 1.00E-011E-10

1E-09

1E-08

1E-07

1E-06

1E-05

1E-04

1E-03

1E-02H2

Thermal 1 Manifold Thermal 4 Manifolds

Time [s]

Pre

ssu

re [m

ba

r]


Monte-Carlo

H2 CO

10-3

10-2

10-1

100

101

100 101 102 103 104 105

500 l/s1000 l/s2000 l/s

time[

s]

n x Qspark

10-3

10-2

10-1

100

101

100 101 102 103 104 105

500 l/s1000 l/s2000 l/s

time[

s]

n x Qspark

Dynamic vacuum I – Effect of a manifold• Calculated recovery time to 10-8 mbar with PSpice for a quadrant type structure in a

CTF2 type tank (old data)• For one sparks outgassing Qspark, or multiples of it equal to n x Qspark

• Two regions: first is determined by the filling time of the cavity + tank volume• Second is determined by the applied pumping speed


Dynamic vacuum II – Dark currents• Dynamic vacuum due to field emission: the problem

– Electrons are field-emitted from high field regions in absence of a breakdown (dark current).

– They hit the cavity wall releasing gas by Electron Stimulated Desorption (ESD)

– Outgassed molecules are emitted with eV speeds, but after first collision they are thermalized and then travel <1 mm during the 200 ns RF pulse, thus cannot escape from RF structure.

– Timescale of the order of the RF pulse duration: outgassing at the beginning of the pulse may affect bunches within the same train.

– Very difficult to measure: we may have p>10-9 mbar locally during the 200 ns RF pulse (in all structures) in a small volume. When this is diluted and pumped through the conductance of the RF structure, it may not be measureble


Dynamic vacuum II – (over-)simplified model

• Only one cell is modelled• Electrons are field-emitted, then accelerated and distributed

uniformly in the cell• Dynamic vacuum by ESD: desorbed molecules fill the whole cell

volume• No pumping


e- on faraday cup

e- are uniformly distributed inside the cell

e- on faraday cup

Cell [mm] dimensions outer area beam area copper area cell volume [mm3] beam channel [mm3]

Ext diam 20 1086.99 28.27 1030.44 2293.36 206.40

Iris diam 6 cell volume [liters] beam channel [liters]

Length 7.3 2.29E-03 2.06E-04

Faraday cup measurements - T18_VG24_Disk_2 - KEK


40 50 60 70 80 90 100 110 1200.01

0.1

1

10

100

f(x) = 2.23364820812998E-17 x^9.03596372712817

f(x) = 1.31429070987924E-18 x^9.34977881350076

FC-UP [microA]Power (FC-UP [microA])FC-DN [microA]Power (FC-DN [microA])

Eacc [MV/m]

Fara

dy cu

ps cu

urre

nt [µ

A]

Pulse 173 ns

From: T. Higo

ESD data• No much data on unbaked copper(N. Hilleret, CAS Vacuum School 2006 - G. Vorlaufer CERN-Thesis (2002) )


Data


Total e- current [A] Pulse duration [ns] Total charge [C] Number of electronsSolid angle (one cell, one side)

Total electrons on copper Dose per pulse (e- /cm2)

1.00E-04 200 2.00E-11 1.25E+08 0.027439024 4.56E+09 4.42E+08

ESD coefficient for H2 (unbaked copper)

Total H2 molecules per RF pulse

Equivalent pressure at RT (total volume)

2.00E-01 9.11E+08 1.12E-08

ESD coefficient for CO2 (unbaked copper)

Total CO2 molecules per RF pulse

6.00E-02 2.73E+08 3.37E-09

ESD coefficient for H2 (copper baked 250 C)


1.30E-02 5.92E+07 7.29E-10

ESD coefficient for CO2 (copper baked 250 C)


6.00E-03 2.73E+07 3.37E-10

ESD coefficient for H2 (copper baked 300 C)


3.00E-03 1.37E+07 1.68E-10

ESD coefficient for CO2 (copper baked 300 C)


1.60E-03 7.29E+06 8.98E-11

107 pulses to start conditioning(~2 days at 50 Hz)109 pulses for 10 ESD reduction(200 days at 50 Hz)

10 times maximum allowed

3 times maximum allowed

For the dynamic vacuum of breakdowns we were considering 2x1012 molecules.

G. Vorlaufer CERN-Thesis (2002)

Benvenuti et al LEP2 94-21

MathewsonJVSTA 15 (1997) 3093

Dynamic vacuum I – Results


Pressure goes to < 10-9 mbar in less than 1 msec !This is faster than the sampling time of common vacuum gauges...

Same plot as for dynamic vacuum due to breakdowns (2x1012 molecules released)

Extrapolating to 1000 less molecules released due to ESD

1.00E-06 1.00E-05 1.00E-04 1.00E-03 1.00E-02 1.00E-011E-10

1E-09

1E-08

1E-07

1E-06

1E-05

1E-04

1E-03

1E-02H2

Thermal 1 Manifold Thermal 4 Manifolds

Time [s]

Pre

ssu

re [m

ba

r]

Experimental programme

• Get more precise data on ESD: higher electron energies, effect of surface treatments and of fabrication procedure

• Obtain e- trajectories maps for reasonable assumptions of distributions of field-emitters, and calibrate the intensities with known Farady-cup dark current measurements.

• Couple the above data as input for Monte-Carlo simulations of trajectories and densities of outgassed molecules


ESD measurement system


Sample Holder plate

Bias Filament

Cu Sample1

2

B-A gaugeRGA

Bias Sample

Electrometer

Butterfly valve

Pumping group

Conductance

Gas Injection Line

Filament

Filament Power Supply

Transport rod

Full Range Gauge

ESD system pictures


First results (dummy Cu specimen)


The system is operational (Diploma thesis of Chiara Pasquino)

1.00E+14 1.00E+15 1.00E+161.00E-05

1.00E-04

1.00E-03

1.00E-02

1.00E-01

H2

CH4

H2O

CO+N2

C2H6

CO2

Integrated dose [electrons/cm2]

Yiel

d [m

olec

ules

/ele

ctro

n]

Rationale

• We want to study the effect of surface treatments and fabrication on ESD

• Large number of samples have been prepared (will be tested also by “spark-test”, SEM, XPS)


w/o etch

Passivation

SLAC etch

w/o etch

Passivation

SLAC etch

w/o etch

Passivation

SLAC etch

w/o etch

Passivation

SLAC etch

w/o etch

Passivation

SLAC etch

w/o etch

Passivation

SLAC etch

w/o etch

Passivation

SLAC etch

w/o etch

Passivation

SLAC etch

w/o etch

Passivation

SLAC etch

CERN 2 2 2 2 2 2 2 2 2 2 2 2 24

Bodycote 2 2 2 2 2 2 2 2 2 18SLAC 2 2 2 2 2 4 14

2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 4 56

Hydrogen (1 bar)

elliptical

Vacuum REFERENCE Vacuum Argon (mbar) H2 (mbar) H2 (1 bar) Vacuum Argon (mbar) Hydrogen (mbar)

main steps: Machining Cleaning Heat

treatment ESD

Heat treatment under hydrogen• Rationale for the calculations:

– H2 bonding introduces hydrogen in the copper, subsequently outgassed in vacuum annealing.

– Any H2 leftover may influence ESD


0 100 200 300 400 500 600 7000

300

600

900

1200

2020 20

1040 1040

800

Time (minutes)

T (°

C)

Bonding at 1040 °C with 1 bar of H2

Annealing at 650 °C for 10 days1x10-8 mbar vacuum

Calculated data (assume Cu sheet 1 cm thick)


1 bar H2

1 bar H2650 °C 10-8 mbar vacuum

-0.5 0.0 0.51.194x10-4

1.196x10-4

1.198x10-4

1.200x10-4

C (

% w

t)

x (cm)

1309 K 1305 K 1301 K 1297 K 1293 K 1289 K 1285 K 1281 K 1277 K 1273 K 1269 K

Ramp down, 4K/min

-0.5 0.0 0.51.1782x10-4

1.1784x10-4

1.1786x10-4

1.1788x10-4

C (

% w

t)

x (cm)

1073 K

H in Cu @ 1073 K

-0.5 0.0 0.51.40x10-9

1.45x10-9

1.50x10-9

1.55x10-9

1.60x10-9

C (

% w

t)

x (cm)

10 d 5 d 2 d 1 d 10 h 8 h

Heat treatment after bonding

-0.5 0.0 0.5

1.1993573x10-4

1.1993574x10-4

C (

% w

t)

x (cm)

H in Cu, 1h @ 1313

1 bar H2

Direct measurements, anyway


F-Cup+Isolator

Stand alone Test Stand in CTFI I

Concrete block

girder girder girder

ACC

pump

Pump Tee

load

coupler

Circular WG

Mode converter

WG valve

0.5 m

0.5 m

Courtesy: S. Doebert

From: K.M. Schirm

Zone to be instrumented with fast gauges

Outlook 1

• Static vacuum seems to be achieved only marginally with present design– Need more precise data on water re-adsorption (sticking

probability depends on coverage)

• Dynamic vacuum due to breakdowns seem to be under control (recovery time ≤ pulse repetition)– However, data from RF tests are needed for further cross-

checking


Outlook 2

• Dynamic vacuum due to dark currents: still open question

• Experimental programme:– ESD data on unbaked copper at high e- energy from CERN– Dark current simulations from SLAC – ACE3P– Introduce these into MC+FEM models and get gas

distribution– Direct measurements will be attempted in 12 GHz test

bench (although feasibility is questionable)




Molecule speed

Atomic mass

Molecule speed @ 300 K [m/s]

Molecule displacement in RF pulse [mm]

H2 2 1579 3.16E-01

H2O 18 526 1.05E-01

CO 28 422 8.44E-02

CO2 44 336 6.73E-02

Assuming a molecular speed of 300 K = 0.026 eV


10-9

10-8

10-7

10-6

-1 0 1 2 3 4 5

CTF2 pressure burst

100xQspark sim

10xQspark sim

1xQspark sim

pre

ssu

re[T

orr]

time[s]

Tentatively crosschecked with CTF2 data + Montecarlo simulations

Date post:	18-Dec-2015
Category:	Documents
Upload:	cecilia-harrington
View:	218 times
Download:	0 times

Progress on investigation of dynamic vacuum (RF structures only) Sergio Calatroni With Cedric...

Documents