Report from MeVArc 2013 CLIC Meeting, 7 March 2014 Walter Wuensch, - - PowerPoint PPT Presentation

Walter Wuensch, CERN CLIC Meeting, 7 March 2014

Report from MeVArc 2013

Walter Wuensch, CERN CLIC Meeting, 7 March 2014

Stands for - Mechanisms of Vacuum Arcs

• Focuses on the fundamental physics of vacuum arcs – our main

performance limitation in CLIC.

• Theory, simulation and experiments.
• Multi-disciplinary - material science, surface physics, plasma

physics, high-voltage systems, radio frequency, etc.

• Multi-project – accelerators (rf (even a little superconducting),

kickers ), fusion, vacuum interrupters, satellite ion thrusters, fast switches

Walter Wuensch, CERN CLIC Meeting, 7 March 2014

Fourth in series:

• 2012, hosted by Sandia Laboratory

https://www.regonline.com/builder/site/Default.aspx?EventID=1 065351

• 2011, hosted by University of Helsinki

http://beam.acclab.helsinki.fi/hip/mevarc11/index.php

• 2010, hosted by CERN http://indico.cern.ch/event/75380/
• 2013, hosted by us and held from 4-7 November in Chamonix

http://indico.cern.ch/event/246618/overview

• 2015, to be hosted by University of Helsinki

Walter Wuensch, CERN CLIC Meeting, 7 March 2014

We now have contact with another bi-annual workshop series ISDEIV (International Symposium on Discharges and Electrical Insulation in Vacuum), and had good participation from them. We will alternate years with them from now on. Personal opinion now. We had around 55 participants. Some loss of US participation due to government shut-down.

Walter Wuensch, CERN CLIC Meeting, 7 March 2014

Primary organizers were me and Alexia. My sincere thanks to her for her incredible efficiency! Big efforts also from Flyura Djurabekova (University of Helsinki), Matt Hopkins (Sandia Laboratory) and Sergio Calatroni (CERN). Plus lots of background help from all the CERN-based students! Thanks, thanks, thanks!

Walter Wuensch, CERN CLIC Meeting, 7 March 2014

I will now review of content and ideas. Emphasis on newcomers and other areas. Broad, fast brush overview from presentations. We also had a half day poster session. Very good format for medium sized workshops. No rf today although a ½ day was spent at the workshop.

Micropropulsion and Nanotechnology Laboratory (MpNL)

Modeling Approaches to Vacuum Arc Plasma

Michael Keidar

Mechanical & Aerospace Engineering The George Washington University

Acknowledgement: NSF, NASA, AFOSR In collaboration with: I. I. Beilis, R.L. Boxman, M.B. Schulman,

• E. Taylor, P. Slade, T. Zhuang, A. Shashurin,

Micro-cathode arc thruster (µCAT)

Schematic of the μCT

Isolation Material Spring Cathode Anode Magnetic Coil Core Isolator

Feed Mechanism

10

The Free Bounda The Free Boundary Model ry Model

     

                               E j v B v B j E j B j v v

i i e i e i i i i

n en n n e kT n n n T T Z k m

• Assumptions

– Steady-state, fully ionized, collision dominated, quasi-neutral plasma – Anode acts as a passive current and particle collector – Cathode spots act as source of plasma at a specified jet angle and velocity – Cathode spots evenly distributed (no arc constriction) within a circular area – Magnetized electrons, unmagnetized ions – External magnetic field purely axial and uniform, self field purely azimuthal

• Numerical methods

– Iterative scheme for solving for the potential – Implicit second-order accuracy method to calculate the velocity, current density, and density from the potential

• Two approaches:

– self consistent solution – voltage is set by high-current column

 



  

r z

r d r j r r B 



Keidar et al, J. Phys. D, 1996



2

) ( ) ( 3 2 3 j V div P j div kT T k e j T V kN

e e e e e e

            

   

    1 1

2 2

( ) V V V V V V V

n n n

11

Comparison with Experiment

Experiment: V.M. Khoroshikh,

• Sov. Phys. Tech. Phys., 33(6)

723 (1988)

Keidar, J. Appl. Phys. 1998; Rev. Sci. Instr. 2000

cathode anode

Mikhail M Tsventoukh Lebedev Physical Institute of the Russian Academy of Science

J

pl

v

INITIATION of EXPLOSIVE ELECTRON EMISSION PULSES – ECTONS as INITIATION of VACUUM DISCHARGE STAGES – the BREAKDOWN, the SPARK, and the ARC

Gennady A Mesyats and Sergey A Barengolts Lebedev and Prokhorov Institutes of the RAS

RECENT INTEREST FOR ARCING IN FUSION

Two reasons for recent interest in arcing and, in general, in collective plasma-surface interactions Large transient energy flux (~1 MW/cm2) due to ELMs (edge localized modes)

MAST (megaamp. spherical tokamak)

Surface fine structure, i.e.

• W-deposited

films (ASDEX-Upgrade tiles)

• Layers of W-fuzz
• Liquid

Li films

• n

a capillary structure

INITIATION BY PLASMA ACTION

Numerical modeling has been performed for the plasma action

• nto

the wall having a microprotrusion with taking into account

• thermo-field-emission,
• 2D thermal balance,
• heating by incident plasma,
• sheath properties

[Uimanov 2003 IEEE Trans Plas Sci 31

822; Barengolts, Mesyats, Tsventoukh 2008 JETP 107 1039]

For plasma: 1020 cm-3, 4 eV new explosion within t ~ 10 ns has been shown numerically

   

2

         grad gradT div t T c

 

e j q q T grad T

TF e i

       



) (

   

TF e i

j e grad T         



) (

 

    grad div

INITIATION BY PLASMA ACTION

explosive

• verheating
• f

a surface microprotrusion by volume Joule energy release, whereas the surface fluxes likely being balanced

‘FINE-STRUCTURE’ EFFECT on ARCING

A new power threshold q threshold,1 << q0  Film- structure of surface absorbs the incident energy (positive)  The condition for explosive electron emission (arcing) arise at a lower power threshold (negative) q > q0 ~200 MW/cm2

Within factor 2 of our Sc limit!

17

1D PIC/DSMC computer modeling of near- cathode plasma layers and expansion of cathode plasma flare of vacuum arc cathode spot

Dmitry L. Shmelev

Institute of Electrophysics UB RAS, Ekaterinburg, Russia 106 Amundsen St., Ekaterinburg, 620016, Russia

• e-mail: shmelev@iep.uran.ru

MeVArc 2013

• D. Shmelev, shmelev@iep.uran.ru

Institute of Electrophysics UB RAS

18 18 Flux of the atoms from the cathode is calculated according to Hertz-Knudsen approximation Flux of the electron is calculated according to Murphy and Good approx.

2 ) ( T m T P nv

a a

   

    W

d T W N W E D nv

a

W e



 

  , , , Cu Cathode fixed T

c

U0 voltage e a e a i e a i Plasma

Ghost cell with parameters

• btained by

extrapolation from the gap

1D3V PIC/DSMC model of cathode layer

Problem geometry

Calculation domain - 5 µm Collisions

• D. Shmelev, shmelev@iep.uran.ru

Institute of Electrophysics UB RAS

19 19

1D3V PIC/DSMC model of cathode layer

Results Tc=4100 K, U0=15 V.

Electrons Ions Atoms

Distribution of particles returned to cathode Cathode sheath is collisional layer for ions

Walter Wuensch, CERN CLIC Meeting, 7 March 2014

Walter Wuensch, CERN CLIC Meeting, 7 March 2014

Walter Wuensch, CERN CLIC Meeting, 7 March 2014

Walter Wuensch, CERN CLIC Meeting, 7 March 2014

Walter Wuensch, CERN CLIC Meeting, 7 March 2014

Walter Wuensch, CERN CLIC Meeting, 7 March 2014

Walter Wuensch, CERN CLIC Meeting, 7 March 2014

Walter Wuensch, CERN CLIC Meeting, 7 March 2014

Walter Wuensch, CERN CLIC Meeting, 7 March 2014

Walter Wuensch, CERN CLIC Meeting, 7 March 2014

Walter Wuensch, CERN CLIC Meeting, 7 March 2014

Walter Wuensch, CERN CLIC Meeting, 7 March 2014

Early signals of breakdown through Stochastic modeling

Yinon Ashkenazy & Michael Assaf Racah Institute of Physics, Hebrew University, Jerusalem, Israel

Dislocation mediated – self organized criticality

Uchic, Shade & Dimiduk, Annual Review of Materials Research (2009). Dimiduk, Woodward, LeSar & Uchic: “Scale-Free Intermittent Flow in Crystal Plasticity.” Science (2006) 1188.

Single crystal micro-pillar compression: Dislocation mediated intermittent flow - size effects, hardening. Dislocation density inside a plane as a controlling parameter.

Direct quantitative analysis of strain bursts (~20 micron). Intermittency characterized by a universal Power law burst PDF Acoustic emissions: Similar + space and time coupling between events (Weiss & Marsan, Scjence 2003 ) Earthquakes show similar PDF and spatio-temporal correlation (Kagan, Geopgysical J. (2007)

Using dislocation dynamics to reproduce PDF

Csikor, Motz, Weygand, Zaiser & Zapperi, “Dislocation Avalanches, Strain Bursts, and the Problem of Plastic Forming at the Micrometer Scale” . Science (2007)

• 3D dislocation dynamics reproduce strain burst

scaling

• where C is a normalization constant, τ is a

scaling exponent, and s0 is the characteristic strain of the largest avalanches.

• Intermittency – as a result of dislocation
• Interactions. Stochastic nature a result of

varying initial conditions.

• Avalanche is a 2D event, with an upper cutoff

due to structure and work-hardening. Strain is limited to about 10^-6 in a cm size sample.

• Recently (Chen, choi, papanikolaou & Sethna

2010 to 2013): scaling of structures using an advanced CDD code.

P s

( ) = Cs-t exp - s / s0

( )

2

é ë ù û

1 t RBD » exp E2DV / kBT

( )

é ë ù û

PRE-breakdown

• As the system approaches

the critical point. Fluctuation diverge.

• Observable through

standard deviation of the time correlation

• Or, more generally,

autocorrelation in the signal

R(k) = I(t)- < I >

( ) I(t + k)- <

I >

( )

t-k

ò

dt I(t)- < I >

( )

2 t-k

ò

dt

SD(t) = I(t)- < I >

( )

2 t- D t+D

ò

dt < I >

( )

2

Initial try – DC measurements

• Dark current measurements

(Varying field and gap distance)

• Low pass filter is clearly needed

(applied 0.2 GhZ)

Nick Shipman, Adar Sharon

HIP Computer simulations of Cu surface

behavior before and after a breakdown event

Flyura Djurabekova, Aarne Pohjonen, Avaz Ruzibaev, Stefan Parviainen, Riikka Ruuth, Johann Muszynski, Kai Nordlund

Helsinki Institute of Physics and Department of Physics University of Helsinki Finland

Flyura Djurabekova, HIP, University of Helsinki

39

More detailed analysis under the breakdown spots

 We analysed a number of breakdown spots, there are many vacancies and

small vacancy clusters as well as the dislocations under the impact spot

200 ions, crystal structure 500 ions, heated crystal structure

Different cooling rates

Electro-mechanical and thermal simulations of surface under electric field

2013

• V. Zadin, S. Parviainen, A. Aabloo, F. Djurabekova

Computer simulations in Chemistry and Physics

DFT Molecular dynamics Mesoscale modeling Finite Element Analysis

Distance

1Å 1nm 1μm 10nm 1mm femtosec picosec nanosec microsec seconds years

Time

Simulated systems

• Coupled electric, mechanical, thermal interaction

– Electric field deforms sample and causes emission currents – Emission currents lead to current density distribution in the sample – Material heating due to the electric currents – Electric and thermal conductivity temperature and size dependent – (Deformed) sample causes local field enhancement

h d

• Dc El. field ramped up to 10 000 MV/m
• Comsol Multiphysics 4.3b

— Nonlinear Structural Materials Module — AC/DC module

• Simulated materials:

— Soft copper — Single crystal copper — Stainless steel

Single tip deformation

Plastic deformation Necking

• Nanoscale tip under electric

field induced stress

• Simulations with FEM and MD
• Constant temperature
• No emission currents
• Linear ramping of el. field
• MD and FEM predict the same

location for plastic deformation

• Piece of material is removed

from the tip

• Plastic deformation in FEM
• Dislocations in MD
• Dislocations are carriers of

plastic deformation FEM overestimates plastically deformed area!

Mechanical interactions of emitters

σMises, max=130 MPa σMises, max=60 MPa σMises, max << 1 Pa E=1V/m E=135 MV/m E=166 MV/m Nearby emitters interact The emitters repel due to the surface charge Elastic regime:

• Reversible deformation
• f the emitters

Plastic regime:

• Highest stress is at inner

side of the tip

• Limiting effect to the

density of emitters?

• Two closely located

emitters

• Emitter aspect ratio ~10
• Distance between the

emitters – 0.3H (H – height

• f the emitter)
• Linear ramping of el. field

Impact of dry ice cleaning on the enhanced field emission from flat Cu samples

• S. Lagotzky, G. Müller

University of Wuppertal, FB C – Physics Department, Wuppertal, Germany

• T. Muranaka, S. Calatroni

CERN, Geneva, Switzerland 1. Motivation and strategy 2. Measurement techniques 3. Samples 4. Field emission properties 5. Conclusions and Outlook 4th International Workshop on Mechanisms of Vacuum Arcs (MeVArc 2013) 06.11.2013

Acknowledgements: Funding by BMBF project 05H12PX6

Impact of dry ice cleaning on the enhanced field emission from flat Cu samples Stefan Lagotzky | 4th International Workshop on Mechanisms of Vacuum Arcs 46 von 17

DC FIELD EMISSION SCANNING MICROSCOPE (FESM)

sample anode piezotrans- lators electron gun ion gun

• Regulated voltage V(x,y) scans at fixed FE current (typ. I = 1nA) and gap ∆z

(Øanode = 300 µm, scan range ≤ 25ˣ25 mm2, tilt correct. ±1 µm within ±5 mm) → emitter position, number density N and localization of emitters

• Local U(z) & I(V) measurements of single emitters → Eon(1 nA), βFN, SFN
• Ion bombardement (Eion = 0 – 5 keV), SEM (low res.), heat treatments (< 1200°C)
• Ex-situ SEM & EDX: Identification of emitting defects (positioning accuracy ~100 µm)

Impact of dry ice cleaning on the enhanced field emission from flat Cu samples Stefan Lagotzky | 4th International Workshop on Mechanisms of Vacuum Arcs 47 von 17

EMITTER NUMBER DENSITY (N) AT DIFFERENT FIELD LEVELS ON SAMPLE 18E

• Field maps between 140 - 260 MV/m, 10 (20) MV/m steps for E > (<) 200 MV/m
• Scanned area: 5x5 mm², truncated cone anode (W, Ø= 300 µm), step size = 150 µm,

Δz = 25 µm (E ≥ 240 MV/m), 40 µm (180 – 240 MV/m) or 50 µm (E < 180 MV/m)

• 23 emission sites at Eact = 260 MV/m

→ Emitter number density : 92 cm-2

• EFE free region in the scanned area at E = 260 MV/m
• Activation field Eact > onset field Eon
• Eact = 260 MV/m, Eon =168 MV/m

Impact of dry ice cleaning on the enhanced field emission from flat Cu samples Stefan Lagotzky | 4th International Workshop on Mechanisms of Vacuum Arcs 48 von 17

SINGLE EMITTER CHARACTERISTICS

Measuring I(E)-curves and making SEM/EDX investigations correlated to the field maps Two Examples from 17E:

) exp( ) (

2 / 3 2 2

E B E AS E I

FN FN FN

     

β~20, S ~10-3 µm² β~13, S ~10-1 µm²

Al, Si

Impact of dry ice cleaning on the enhanced field emission from flat Cu samples Stefan Lagotzky | 4th International Workshop on Mechanisms of Vacuum Arcs 49 von 17

DRY ICE CLEANING SYSTEM

• Commercial DIC system (SJ-10, CryoSnow)
• Installed in cleanroom (class iso 5) at BUW
• Liquid CO2 (10 bar) and N2 (8 bar, propellant gas)

→ Flat (12x3 mm) or round (Ø = 10 mm) jet of CO2 snow particles

• Cleaning of (grounded) samples with handgun (d ~ 5 cm) typically for 5 min
• Samples are treated 2.5 min under 90°/ 45°and 3 x rotated in 90°steps

hand gun nozzle Clean room environment control panel inlet CO2 inlet N2

Walter Wuensch, CERN CLIC Meeting, 7 March 2014

Walter Wuensch, CERN CLIC Meeting, 7 March 2014

Walter Wuensch, CERN CLIC Meeting, 7 March 2014

Manufacturing of void-filled cathodes for breakdown experiments

Anders Korsbäck Department of Physics, University of Helsinki

Background

 Theoretical and simulational work at the

University of Helsinki indicates that the presence of voids (diameter: a few nm) near the surface of a cathode are a significant contributor to the mechanism

• f vacuum breakdown

 To validate these results

experimentally, we are manufacturing samples with voids in Helsinki to be used for breakdown experiments in the DC spark setups at CERN

4.11.2013 Anders Korsbäck 54

Pohjonen et al, University of Helsinki

Irradiation

 Samples are 12 copper cathodes manufactured at CERN:

Annealed, diamond-turned and solvent-cleaned

 Irradiation (30 keV He+ beam) was

simulated using SRIM:

 A dose of ~30 atom% is known as

the limit where blistering of the surrounding copper matrix happens

 Hence, the dose was chosen to give

5 atom% at most common stopping depth

 Half of each sample was irradiated while the other half was

covered to provide a clean, un-irradiated reference for comparison

4.11.2013 Anders Korsbäck 55

SRIM simulation of He ion implantation into Cu sample, ion energy of 30 keV

500 1000 1500 2000 2500 3000 2 4 6 8x 10

4

Depth (Å) Density per fluence (cm)

Positron annihilation spectroscopy

 Two samples were sent to the positron lab, measurement

took place ~ 3 months after irradiation

 Overview of PAS:  Result: Clear difference

between irradiated and reference side, vacancy clusters already detected at ~25 nm depth. Size of clusters at least ~10 missing atoms, possibly a lot larger

4.11.2013 Anders Korsbäck 56

MD simulations of Fe precipitates in Cu

Simon Vigonski1, Vahur Zadin1, Alvo Aabloo1, Flyura Djurabekova2

1 Institute of Technology, University of Tartu 2 University of Helsinki

MeVArc 05.11.2013

57

Surface geometry

• Precipitate close to

the surface.

• A depression forms
• n the surface due to

lower deformability of Fe compared to Cu.

• Cu-Fe-vacuum

interface facilitates atom evaporation

58

Depression above the precipitate at 112 ps after reaching maximum electric field strength. Color: Cu centrosymmetry parameter. Atom types: blue – Cu; red – Fe

DC Spark Experiments

Nick Shipman, Sergio Calatroni, Roger Jones, Anders Korsbeck, Tomoko Murunaka, Walter Wuensch

07/03/2014

60

What is the High Rep Rate Circuit?

The picture above shows the HRR circuit. The metal box housing the switch is placed as close as possible to the vacuum chamber to minimise stray capacitance. The HRR circuit uses a solid state switch to supply high voltage pulses (up to 10kV) at a rep rate of up to

• 1kHz. The energy is stored on a

200m/1us long coaxial cable.

BDR vs E 40um gap

Both the power law model and the stress model fit the data well. Going to a lower BDR in the future should help distinguish between them. The exponents obtained for the power law model are very similar to those obtained in high power RF tests of accelerating cavities. The fitted exponent tends to decrease for a larger gap.

62

Measured Turn on Times

63

The Swiss FEL turn on times are much longer than in the DC case and the variation is much greater, this is keeping with other RF breakdown turn on time measurements.

Falling edge duration - KEK

Test Measurement Result

Simulation 0.25ns New DC System Voltage Fall Time ~7ns TBTS (X-Band) Transmitted Power Fall Time 20-40ns KEK (X-Band) Transmitted Power Fall Time 20-40ns Swiss FEL (C-Band) Transmitted Power Fall Time 110-140ns

The summary table on the right suggests the characteristic size of the system breaking down may govern the turn on time.

Walter Wuensch, CERN CLIC Meeting, 7 March 2014

New perspectives