Experimental Studies of Electron and Gas Sources in a Heavy-Ion Beam* A.W. Molvik 1,2 With contributions from F.M. Bieniosek 1,3, J.J. Barnard 1,2, D.A. Calahan 2, R.H. Cohen 1,2, A. Friedman 1,2, M. Kireeff Covo 1,2, J.W. Kwan 1,3, W.R. Meier 2, L. Prost 1.3, A. Sakumi 4, P.A. Seidl 1,3, G. Westenskow 1,2, S.S. Yu 1,3, 1 HIF-VNL, 2 LLNL, 3 LBNL, 4 CERN ECLOUD04 Napa, CA April 19, 2004
OUTLINE Introduction to Heavy Ion Fusion (HIF) Measurements of gas desorption & electron emission Measurements of electrons in quadrupole magnets Related Papers Jean-Luc Vay, Status report on the merging of ECE code POSINST with 3-D accelerator PIC code WARP Tuesday, pm Ron Cohen, Simulations of e-cloud for Heavy-Ion Fusion Wed. am Peter Stoltz, The CMEE Lib. for numerical modeling ECE Wed. am Hong Qin, Delta-f simulations of Electron 2-stream Instab. Wed. am Molvik, ECloud04, 2
Target Requirements establish accelerator requirements for power plant driver 3-7 MJ x ~ 10 ns ~ 500 Terawatts Ion Range: 0.02-0.2 g/cm 2 1-10 GeV Beam charge (3-7 MJ/1-4 GeV) few mcoul 0.7 cm 1.5 cm Molvik, ECloud04, 3
HIF Power Plant Driver Many high-current beams needed to deliver several Mjoules to target with GeV ions A = 209amu (Bi) q= +1 L = 2.9km E = 4.0GeV I = 94.A/beam T = 0.2µs Multibeam (120) Accelerator 120 beams E = 1.6MeV I = 0.63A/beam T = 30µs 120 beams Molvik, ECloud04, 4 E = 4.0GeV I = 1.9kA/beam T = 10 ns
Induction Acceleration can achieve 20-50% efficiency B Efficiency increases as current increases Multiple beams within single induction core Molvik, ECloud04, 5
System studies show that driver cost reduced at high fill factor [fill factor may be limited by ECE or desorption] Total driver cost, $B 9.0 8.0 7.0 6.0 5.0 4.0 3.0 2.0 1.0 0.0 IBEAM results: ~$1B Robust Point Design (2.8 B$) range being explored 20 30 40 50 60 70 80 90 100 Fill factor, % (fixed number of beams, initial pulse length, and quadrupole field strength) Fill factor = a max /R pipe a max Beam Clearance a avg R pipe Beam pipe
HIF-ECE distinguishing features Economic mandate to maximally fill beam pipe - ions scrape wall Linac with high line charge density (Beam potential φ b > 1 kv) Induction accelerator characteristics 0.2-30 µs - If beam head scrapes: gas desorbed (Γ 0 ~ 10 3-10 4 ) and secondary e - (Γ e ~ 100) trapped by rising φ b. Control of beam head is essential. - If beam flattop scrapes: gas desorbed, SEY not necessarily trapped. - If desorbed gas reaches beam: e - from ionized gas are deeply trapped by φ b, cold ions expelled. This is expected to be main e - source in HIF, especially near injection energies (10-100 kev/amu) where atomic cross sections peak (~10-15 cm 2 ). - Electrons are trapped for time to drift through 1 magnet, then expelled. Beam-induced multipactor not present Trailing-edge multipactor not an issue ( 0.2 s between pulses). Molvik, ECloud04, 7
Beam hitting gas or walls creates electrons and gas these can multiply Beam on gas, I b Beam loss to walls, I bw 1.0-1.8 MeV K + Beam γ e - γ K 0 K + Beam γ n 0 e - Fe n 0 K + 2-5 kv potential K 2+ i + These interaction products create rich opportunities for diagnostics along with problems for diagnostics and beams Molvik, ECloud04, 8
HCX layout for ECE studies in magnetic quads QI-10 D2 e clearing D-end Electrodes e-suppressor (+10 kv) MA1 MA2 MA3 MA4 Slits, optical diagnostics Optical Diagnostics 1-17-04 Electrostatic transport magnetic transport Diagnostics on insert tube Gas-Electron Source Diagnostic (GESD) ECE experiments began with diagnostics mounted on insert tubes within magnetic quads MA3 & MA4. Later experiments removed insert tubes, added electron-suppressor after MA4 and clearing electrodes between magnets. Molvik, ECloud04, 9
Measure electron emission Γ e and gas desorption Γ 0 from 1 MeV K + beam impact on target Gas, electron source diagnostic (GESD) Suppressor grid Ion gauge Faraday cup Beam Beam Target, angle ~2 o -15 o Reflected ion collector Tiltable target Electron Suppressor Grid & target bias varied Measure coefficient of electron Γ e and gas emission Γ 0 per incident K + ion. Calibrates beam loss from electron currents to flush wall electrodes. Evaluate mitigation techniques: baking, cleaning, surface treatment Measuring scaling of Γ 0 with ion energy test electronic sputtering model Molvik, ECloud04, 10
GESD secondary electron yield (SEY) varies with cos(θ) -1, secondary energy T e = 30 ev Simple model gives cos(θ) -1 - Delta electrons pulled from material by beam ions (de/dx) - Electrons from depth > δ (δ~ few nm) cannot leave surface - Ion path length in depth δ is L. L = δ /cos(θ) Results depart from this near grazing incidence where the distance for nuclear scattering is < L 1 δ θ Molvik, ECloud04, 11 L L = δ /cos(θ) 1. P. Thieberger,A. L. Hanson, D. B. Steski, et al., Phys. Rev. A 61, 42901 (2000). Coefficient of electron emission Target current (ma) 150 100 50 SEY 6.06/cos SRIM(22A) 0 76 78 80 82 84 86 88 90 Angle Angle from of normal incidence (deg.) 100.0 10.0 1.0 T e = 30 ev 0.1-200 -100 0 100 200 Grid bias (V) Grazing incidence
Rough surface mitigates ion-induced electron emission, gas desorption, and ion scattering Surface roughened by glass-bead blasting (Inexpensive, but can warp surface) Angle of incidence: grazing ~60 o [from 1/cos emission] Sawtooth surface (CERN-SPS) more effective, but more expensive. Backscatter coefficient 1.0000 0.1000 0.0100 0.0010 Ion backscattering - SRIM X30 (Rough) X1000 (Sawtooth) Gas desorption coef. Electron emission coef. 150 100 50 0 10,000 5,000 Electron emission x10 Smooth 6.06/cos Rough 80 82 84 86 88 90 Angle from normal (deg) Gas desorption X2-3 Smooth Rough 0.0001 0 30 60 90 Angle from normal (deg.) 0 80 82 84 86 88 90 Angle from normal (deg) Molvik, ECloud04, 12
Electron studies in magnetic quads Initial studies with diagnostics mounted on 5.5 cm diameter tube in quad. 180 ma full beam scraped cylindrical diag. tube - Diagnostics difficult to interpret 15-25 ma apertured beam, mostly not scraping wall - Capacitive probes measure φ b (With apertured beam signals approximate expectations n e n b ) - Flush probes (right) measure secondary electron emission, from which we infer beam loss and gas desorption. Goal measure accumulation of electrons and gas - This may require diagnostics functioning with electrons / gas present. - Develop mitigation techniques to increase performance. Molvik, ECloud04, 13
Puzzle solved: negative spike at end-of-pulse varies with bias on BPM, caused by SEY from beam loss Joel Fajans, UC Berkeley not confinement! Halo loss and scattered ions generate secondary e - ESQ currents (capacitive pickup) match di Rogowski /dt except for burst at end of pulse [L. Prost, Ph.D thesis] 0.8 0.6 QM1 electrode monitor (positive) Many e - to collector BPM B Few e - off collector Amplitude [V] 0.4 0.2 0-0.2 Theory Data -0.4-0.6-0.8-1 2 3 4 5 6 7 8 9 10 0.50 Time [µs] 0.25 QM6 electrode monitor (positive) 0.00-0.25 Amplitude [V] -0.50-0.75-1.00 Theory Data -1.25-1.50 Loss of 0.6% of beam can produce e - pulse at tail Molvik, ECloud04, 14-1.75 2 3 4 5 6 7 8 9 10 Time [µs]
Integrated charge to flush full-length collectors in quad magnets ok at head, but tail? Q beam λ(µcoul/m) l = Beam charge within magnet of length l Q head (0.8-1.0) Q Beam Q Tail 3 x Q Beam Why? Head: can t distinguish Capac. Pickup SEY from electrode Tail: can t distinguish Capac. Pickup e - to electrode Gradual growth of negative signal: Measured Q can t exceed Q beam unless electrons supplied from outside this beam tube. Molvik, ECloud04, 15
Progress towards high quality beam transport electron effects only part of picture Beam split into 3, going through a 5.5 cm diam. circular bore (Imaged on scintillator, after beam passes through a slit) Slight improvement from opening bore to 6 x 10 cm elliptical bore without suppressor. 3-shots shown: still not reproducible. Electron suppression added between quad. magnets and scintillator blocks secondary electrons trifurcation an ECE Scintillator image of beam through a slit is much cleaner Quad magnetic field errors: harmonics 1%, 1mm, 1 (?) Simulations predict retuning of electrostatic and magnetic quads will eliminate beam loss. Molvik, ECloud04, 16
Simulations: centering beam and minimizing envelope changes reduces halo growth* Phase-II diagnostics Elliptical-quad-magnet beam tube Diagnostic tube-ii Dashed red lines from envelope code, solid from XY PIC Code PIC shows larger excursions Retuning required upstream to match into magnets. *Steven Lund, private communication 2004. Molvik, ECloud04, 17
New tools: suppressor ring, clearing electrodes between quads Suppressor blocks electrons from quads improves beam quality Clearing electrodes work: upstream indep. of downstream changes Measure drift velocity of e -? v v e b 2I = e = I b 014. Current (V) 0.5 0-0.5 V(suppressor) = -10 kv Capac. (a) (b) ( c) -1 2.0E-6 7.0E-6 1.2E-5 Time (s) 100 Vc(a,b)=+9 kv, Vs=0 (a) (b) ( c) Capac Current (V) 0.5 0 Suppressor V(suppressor) = 0-0.5 Capac. (a) (b) ( c) -1 2.0E-6 7.0E-6 1.2E-5 Time (s) 100 Vc(a)=+9 kv, Vc( c) = 0, Vs=0 Capac. electrode: polarity varies with V s Can suppressor reduce e - to reproducible trickle? Current to clearing elec (mv) 0-100 -200-300 -400 (a) (b) ( c) 0 2 4 6 8 10 Bias on clearing electrode-c (kv) Current to clearing elec. (mv) 0-100 -200-300 -400-500 (a) (b) ( c) 0 2 4 6 8 10 Bias on clearing electrode-b (kv) Molvik, ECloud04, 18
Compare capacitive electrode (MA4) with timederivative of beam current (Faraday cup) Head Consistent with electrons 13% of beam charge. di ( b) dt Cap.-probe l CP It ( + 012. µ s) It ( ) l = ( + ) t t 2v where l CP 0.12 s (v b ) v b CP b Tail C-probe(MA4) & di/dt(fc-d2) Head 0.08 0.06 0.04 0.02 0-0.02-0.04-0.06 0.020-0.033 CapProbe(µC) <d/dt(farcup)> 2 4 6 8 10 12 Time (µs) C-electrode (µcoul) di/dt(fc) (µcoul) 0.023-0.022 Fraction e - 0.13? Molvik, ECloud04, 19
Near-term upgrades to ECE experiments on HCX Mid-FY04: New octagonal diagnostic tubes approximate elliptical shape to pass larger beams without scraping walls study full beam without aperturing. present: 54 mm new: 79 mm Later-FY04: Addition of induction cores between magnets: can accelerate electrons in gap to energy E e > φ b. They will be lost to wall in upstream magnet. Induction core quadrupole diagnostics Support rail Molvik, ECloud04, 20
HIF-ECE Experimental Summary/conclusions ECE (mostly from desorption) likely to influence allowable fill factor, and therefore cost of HIF Driver for power plant. Gas desorption Γ 0 large testing electronic sputtering model Rough surface reduces emission, desorption, & scattering. Beam transport through 4 magnetic quads, with high fill factor ok. Progress in understanding diagnostic signals. Simulation plays significant role in improving performance. Electron suppressor necessary at magnet exit. Clearing electrodes remove electrons in drift region. new tools for ECE in linacs Molvik, ECloud04, 21
Backup material Molvik, ECloud04, 22
HIF-ECE distinguishing features Economic mandate to maximally fill beam pipe - ions scrape wall Linac with high line charge density (Beam potential φ b > 1 kv) Induction accelerator characteristics 0.2-20 µs - Long (ish) pulse duration 0.2-20 µs [Time for desorbed gas to reach beam and be ionized? But no beam-induced multipactor] - 5 Hz rep. rate [time to pump desorbed gas?] - >50% of length at injector occupied by quadrupoles, v e-drft < v e-thermal - Ionized gas e - are born trapped, e - from wall may not be trapped - Multiple beams and frequent accel gaps [Pump gaps or cold bores?] - Large neutral desorption coefficients at pipe wall (Γ o ~10 3-10 4 ) - Injection energies near peak atomic cross-sections [10-100 kev/amu] Heavy-ions stripping cross sections σ E -0.5, σ v E 0 ; don t win at high energy like proton accelerator where σ E -1 Molvik, ECloud04, 23
In search of a mechanism for gas desorption SEY = secondary emission coef. Γ 0 = Gas desorption coef. Γ 0 scales with de/dx(elec) for electronic sputtering Improved background subtraction for 300 kv_a [Compare open vs. solid green diamonds] Experiments and analysis continuing SEY 150 100 50 50 KV 70 KV 140 KV 200 KV 200 KV a 300 KV 400 KV 300kV_a K+ on SS 0 78 80 82 84 86 88 90 Angle from normal (deg) 1E+2 14 250 de/dx (MeV/mg/cm2) 1E+1 1E+0 1E-1 1E-2 Nuclear Electronic Desorption coef (1000-particles) 12 10 8 6 4 2 DesorpCoef 82 deg DesorpCoef 88 deg Secondary electron emission coef. 200 150 100 50 80dE/dx_Elec SEY-82 deg SEY-88 deg 1E-3 1E-3 1E-2 1E-1 1E+0 1E+1 1E+2 1E+3 K+ energy (MeV) 0 0 200 400 600 800 1000 1200 K+ Beam energy (kev) 0 0 200 400 600 800 1000 1200 K+ Beam energy (kev) Molvik, ECloud04, 24
Current-Voltage characteristics of GESD Faraday cup and target, indicate reliable current measurements Negative Faraday cup measures beam current into GESD. Positive Faraday cup measures electrons from ionization of desorbed gas. Suppressor (-150 V) Beam Faraday cup Electron emission Use Faraday cup to measure beam current Current (ma) 1.000 0.500 0.000-0.500-1.000-1.500 GESD Faraday Cup I-V Characteristic, Vs=-200 V, Vg=-150 V, Data 10/30/02. Electron emission Saturation Desorbed gas ionization -200-100 0 100 200 Faraday cup (Target) voltage (V) Saturation of target current indicates reliable measurement of electron emission. Electron emission coefficient is ratio of electron emission current to incoming ion-beam current from Faraday cup. Molvik, ECloud04, 25