Progress in Plasma Booster R&D at FLASHFORWARD - Jens Osterhoff DESY. Accelerator Division Head of Plasma Accelerator R&D
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Progress in Plasma Booster R&D at FLASHFORWARD‣‣ Jens Osterhoff 7th Matter and Technologies Days Head of Plasma Accelerator R&D February 3rd, 2021 DESY. Accelerator Division
Acknowledgements
FLASHFORWARD‣‣ TEAM
Richard D’Arcy (Coordinator) Sarah Schröder
Stephan Wesch (Technical C.) Bridget Sheeran
Judita Beinortaite Gabriele Tauscher*
Jonas Björkland Svensson Jon Wood (PI)
Simon Bohlen Ming Zeng*
Lewis Boulton
James Chappell
Jimmy Garland (PI) THEORY GROUP
Pau Gonzalez
Maxence Thévenet
Julian Hörsch
Gregory Boyle
Alexander Knetsch*
Severin Diederichs
Carl Lindstrøm (PI)
Angel Ferran Pousa
Gregor Loisch (PI)
Alberto Martinez de la Ossa
Felipe Peña Asmus
Mathis Mewes
Kris Põder
Adam Scaachi
…and the technical groups
from the accelerator and particle physics devisions!
© Kyrre Ness Sjøbæk (2020)
Plasma
source
Driver (laser or charged-particles) Depleted driver
Witness Accelerated witness
(electrons)
Plasma accelerators are a centimeter-scale source of GeV beams © Kyrre Ness Sjøbæk (2020)
Plasma
source
Driver (laser or charged-particles) Depleted driver
Witness \
(electrons)
Electron beam can be externally injected
or formed from trapped plasma electrons (internal injection)
~10 – 100 µm
FBPIC simulation simulation by Ángel Ferran Pousa (2020)
Plasma wakefields can sustain accelerating fields of up to ~1-100 GV/m x1000 more than
with focusing gradients above ~1 MT/m RF technology
Our customers: high-energy physics and photon science
> High-energy physics and photon science demand high(est) energy at low cost.
> Solution: Plasma accelerators — signi cantly higher acceleration gradients.
> Simultaneously, particle colliders have strict demands for luminosity:
(FELs have similar demands for brightness)
High repetition rate High energy ef ciency
HD Pwall ηN
ℒ=
8πmec 2
βx βy ϵnxϵny
Luminosity distribution across collision energies.
Source: M. Boronat et al., Phys. Lett. B 804, 135353 (2020).
Low energy spread Low emittance
(luminosity spectrum, nal focusing) η = ηwall→DB × ηDB→WB
> Energy e ciency motivates use of beam-driven plasma acceleration. Beam-drivers are orders of magnitude more ef cient
than laser-drivers (for now)
| Jens Osterho | MT Days | February 3, 2021 Page 5
00
ff
ffi
fi
fi
fi
fi
Primary goal of FLASHFORWARD‣‣
Develop a self-consistent plasma-accelerator stage
with high-e ciency, high-quality, and high-average-power
High ef ciency High beam quality High average power
Transfer e ciency Energy-spread preservation High repetition rate
Driver depletion Emittance preservation
| Jens Osterho | MT Days | February 3, 2021 Page 6
00
ff
fi
ffi
ffi
FLASHFORWARD‣‣ utilizes FLASH superconducting accelerator
Plasma accelerator tightly integrated into facility and benefits from Free-Electron Laser beam quality
ACC1 BC2 ACC23 BC3 ACC45 ACC67
FLASH 1
ACC39
ACC → SCRF modules FLASH 2
Photo
cathode BC → Bunch compressors
> FLASH is an FEL user facility
- 10% of beam time dedicated
to generic accelerator research
> Superconducting accelerator based on ILC/XFEL technology
- ≲ 1.25 GeV energy with ~nC charge at few 100 fs bunch duration
- ~2 µm trans. norm. emittance
- ~10 kW average beam power, MHz repetition rate in 10 Hz bursts
- exquisite stability by advanced feedback/feedforward systems
> Unique opportunities for plasma accelerator science
| Jens Osterho | MT Days | February 3, 2021 Page 7
00
ff
FLASHFORWARD‣‣ utilizes FLASH superconducting accelerator
Plasma accelerator tightly integrated into facility and benefits from Free-Electron Laser beam quality
ACC1 BC2 ACC23 BC3 ACC45 ACC67
FLASH 1
ACC39
ACC → SCRF modules FLASH 2
Photo
cathode BC → Bunch compressors
25 TW laser
FLASHFO
RWARD‣
‣
COLLIMATOR
DIFFERENTIAL
ENERGY
PUMPING
CENTRAL
EXTRACTION & INTERACTION PHOTON
COMPRESSION CHAMBER
UNDULATORS DIAGNOSTICS
X-BAND
25 TW LASER (OUTSIDE THE TUNNEL) MATCHING AND FINAL FOCUSSING BROAD- AND NARROW-BAND TDS
SPECTROMETERS
R. D’Arcy et al., Phil. Trans. R. Soc. A 377, 20180392 (2019)
| Jens Osterho | MT Days | February 3, 2021 Page 8
00
ff
Advanced collimator system for longitudinal bunch shaping
FLASHFORWARD‣‣ beamline features innovative components and methods
(a) (b) (e)
Notch S. Schröder et al.,
collimator Tail J. Phys. Conf. Ser. 1596 012002 (2020)
collimator
Three energy collimators:
(1) Tail (high energy)
(2) Head (low energy)
Tail (3) Central notch (two bunches)
Head
collimator Head µm-precision movements
allows for precise bunch shaping
Energy
(in conjunction with
collimators FLASH compressors and 3.9 GHz cavity)
Electron Cavity
Toroid
bunch BPM Dipole
Dispersive – +
COLLIMATOR
(chirped) section spectrometer
DIFFERENTIAL
ENERGY
Transverse
PUMPING
Dipole CENTRAL
EXTRACTION & PHOTON
INTERACTION deflecting structure
UNDULATORS
COMPRESSION Matching CHAMBER DIAGNOSTICS
quadrupoles Plasma
cell Imaging X-BAND
25 TW LASER (OUTSIDE THE TUNNEL) MATCHING AND FINAL FOCUSSING BROAD- AND NARROW-BAND
quadrupoles TDS
(c) SPECTROMETERS
LANEX
screen GAGG:Ce
| Jens Osterho | MT Days | February 3, 2021
screen Page 10
00
ff
Two discharge capillaries provide density controllable plasma
FLASHFORWARD‣‣ beamline features innovative components and methods
J.M. Garland et al.,
Rev. Sci. Instrum. 92 013505 (2021)
High-voltage discharge
Sapphire capillaries
(50 mm and 195 mm long)
Gases: He, Ne, Ar, Kr, H (soon),
COLLIMATOR
DIFFERENTIAL
ENERGY
PUMPING
CENTRAL
EXTRACTION & INTERACTION PHOTON
COMPRESSION CHAMBER
UNDULATORS DIAGNOSTICS
X-BAND
25 TW LASER (OUTSIDE THE TUNNEL) MATCHING AND FINAL FOCUSSING BROAD- AND NARROW-BAND TDS
SPECTROMETERS
| Jens Osterho | MT Days | February 3, 2021 Page 11
00
ffTwo-BPM tomography enables accurate beam focus characterization
FLASHFORWARD‣‣ beamline features innovative components and methods
C.A. Lindstrøm et al.,
PRAB 23, 052802 (2020)
COLLIMATOR
DIFFERENTIAL
ENERGY
PUMPING
CENTRAL
EXTRACTION & INTERACTION PHOTON
COMPRESSION CHAMBER
UNDULATORS DIAGNOSTICS
X-BAND
25 TW LASER (OUTSIDE THE TUNNEL) MATCHING AND FINAL FOCUSSING BROAD- AND NARROW-BAND TDS
SPECTROMETERS
| Jens Osterho | MT Days | February 3, 2021 Page 12
00
ffPolariX cavity enables 6D phase space measurements
FLASHFORWARD‣‣ beamline features innovative components and methods
B Marchetti et al., accepted for publication
in Scienti c Reports (2021)
P. Craievich et al.,
PRAB 23 112001 (2020)
PolariX X-band cavity (12 GHz)
for femtosecond resolution
Polarizable streak (any direction)
allowing 6D phase space measurements
COLLIMATOR
DIFFERENTIAL
ENERGY
PUMPING
CENTRAL
EXTRACTION & INTERACTION PHOTON
COMPRESSION CHAMBER
UNDULATORS DIAGNOSTICS
X-BAND
25 TW LASER (OUTSIDE THE TUNNEL) MATCHING AND FINAL FOCUSSING BROAD- AND NARROW-BAND TDS
SPECTROMETERS
| Jens Osterho | MT Days | February 3, 2021 Page 13
00
fi
ffTilt-curvature 2D scan
PolariX allows for diagnosis of head-to-tail beam tilts
FLASHFORWARD‣‣ beamline features innovative components and methods
Tilt-curvature 2D scan
55 > Head-to-tail centroid o sets are
sources of collective beam-
550
55.13
500
instabilities in plasma (“hosing”)
55.27
> Tweaking two magnets in the
55.4
450 FLASHForward beamline
55.53 controls and compensates for tilt
400
55.67
55.8
Quadrupole current (A)
350
Charge density (counts)
55.93
300
56.07
56.2 250
56.33
200
56.47
56.6 150
56.73
100 Hosing theory and control
• T.J. Mehrling et al., PRL 118, 174801 (2017)
56.87
57 50 • T.J. Mehrling et al., Phys. Plasmas 25, 056703 (2018)
• A. Martinez d.l.O. et al., PRL 121, 064803 (2018)
-75 -73 -71 -69 -67 -65 -63 -61 -59 -57 -55 -53 -51 -49 -47 -45
Sextupole current (A)
| Jens Osterho | MT Days | February 3, 2021 Page 14
00
ff
ff1.1 GeV energy gain and loss achieved in a 195 mm plasma module
Plasma accelerator essentials — demonstrating 6 GV/m field strength
Energy extraction ➞
Mean plasma
spectrometer beam
image
Imaging energy = 6 MeV
dump (+ efficiency) Energy doubling to 2.2 GeVimage
Spectrometer ➞ plasma booster
600
140
-3
-2 500
Charge density (counts)
-5 120
Charge density (counts)
100
-1 400
x (mm)
x (mm)
0 80
0 300
60
1
200
5 40
2
Initial energy: 1100 MeV Initial energy: 1100 MeV 100
First realization of deceleration to rest 20
3
5 10 15 20 25 30 35 40 1800 2000 2200 2400 2600 2800 3000
E (MeV) E (MeV)AVES AND DYNAMICS … 9, 011046 (2019)
PHYS. REV. X(d)
Optimal beam loading enables uniform and efficient acceleration
> Problem 1: Compared to RF cavities (Q ~ 104–1010), the electric
elds in a plasma decay very rapidly (Q ~ 1–10).
DIRECT OBSERVATION OF PLASMA WAVES AND DYNAMICS …
> The energy needs to be extracted very rapidly (e)
(b)
PHYS. REV. X 9, 011046 (2019)
—ideally within the rst oscillation.
(a1) (b)
(c)
(a2)
(c)
(a3) (d)
(a4) (e)
(d)
(a5)
Image source: M. F. Gilljohann et al., Phys. Rev. X 9, 011046 (2019)
wakefield. Left: (a) Raw shadowgrams showing electron-driven (e) plasma waves
FIG. 4. Ion-channel formation from a plasma wakefield. Left: (a) Raw shadowgrams showing electron-driven plasma waves
channels for (propagating
five consecutive shots.
to the right) and their trailing
| Jens Osterho The
ion channels
| MT Days | February 3, 2021 for fivedashed lines
consecutive shots. inlinesthe
The dashed in thelower shadowgram
lower shadowgram Page 16
00
fi
ff
fiOptimal beam loading enables uniform and efficient acceleration
> Problem 1: Compared to RF cavities (Q ~ 104–1010), the electric
elds in a plasma decay very rapidly (Q ~ 1–10).
> The energy needs to be extracted very rapidly
—ideally within the rst oscillation.
> Solution: Beam loading
The trailing-bunch wake eld “destructively interferes” with
the driver wake eld—extracting energy.
> Problem 2: to extract a large fraction of the energy, the beam will
cover a large range of phases (~90 degrees or more).
> Large energy spread is induced.
> Not (easily) possible:
Dechirping
R. D'Arcy et al.,
PRL 122, 034801 (2019) Image credit: M. Litos et al., Nature 515, 92 (2014)
| Jens Osterho | MT Days | February 3, 2021 Page 17
00
fi
ff
fi
fi
fiOptimal beam loading enables uniform and efficient acceleration
> Problem 1: Compared to RF cavities (Q ~ 104–1010), the electric
elds in a plasma decay very rapidly (Q ~ 1–10).
> The energy needs to be extracted very rapidly
—ideally within the rst oscillation.
> Solution: Beam loading
The trailing-bunch wake eld “destructively interferes” with
the driver wake eld—extracting energy.
> Problem 2: to extract a large fraction of the energy, the beam will
cover a large range of phases (~90 degrees or more).
> Large energy spread is induced.
> Solution: Optimal beam loading
The current pro le of the trailing bunch is precisely tailored
to exactly atten the wake eld.
> This requires extremely precise control of the current pro le.
> FLASHForward provides the tools to do that.
Image credit: M. Tzoufras et al., Phys. Rev. Lett. 101, 145002 (2008)
| Jens Osterho | MT Days | February 3, 2021 Page 18
00
fi
ff
fl
fi
fi
fi
fi
fi
fiKicker
magnet
High-resolution plasma wakefield sampling demonstrated
Electron beam
(chirped)
Dispersive
Transverse
deflecting cavity
Opens a pathway Energy
to targeted and precise field manipulation
section Dipole
collimators
> Beam itself acts as a probe YAG
Dipole
➞ measures in-situ (under actual operation conditions) the e ective eldscreen
acting on beam with µm / fs resolution
Notch
Matching
b collimator quadrupoles Toroid – +
S. Schröder et al., Nat. Commun. 11, 5984 (2020)
Cavity
Tail BPM
collimator Discharge
Cavity
c plasma cell
BPM
Head Imaging
collimator Tail quadrupoles
Dipole
Head spectrometer
LANEX
screen
| Jens Osterho | MT Days | February 3, 2021 Page 19
00
ff
ff
fiLoading the wakefield and beam shaping flattens the gradient
Direct visualization of electric-field control by wakefield sampling
(a) C.A. Lindstrøm et al., PRL 126, 014801 (2021)
Charge density (pC mm -1 MeV -1 )
1
x (mm)
400
0 Trailing
Driver
bunch Plasma off
-1 300
2
Plasma on
1 200
x (mm)
0
-1 100
-2
0
Spectral density (pC MeV -1 )
Plasma off Single-shot statistics:
150
Plasma on Accel. gradient (peak): 1.28 GV/m
(single shot) Transformer ratio: 1.26
Nearly flat field Plasma on, driver Energy-transfer efficiency: 39%
100
(imaging scan)
0.16% 0.13%
FWHM FWHM
50
0
990 1000 1010 1020 1030 1040 1050 1060 1070 1080
Energy (MeV)
(b) 0
> Accelerating gradient of 1.3 GV/m
Plasma off 40
1000 > Energy gain 45 MeV (over
1500 3.5 cm distance) of 100 pC witness,
ronological order)
Shots / 0.02%
35
(c)
density (pC MeV -1 )
with energy spread of 1.4 MeV FWHM and no charge loss
1000
30
500
2000 > Few-percent-level wake
0
eld attening demonstrated
25
0 0.1 0.2 0.3 0.4
| Jens Osterho | MT Days | February 3, 2021 Energy spread (% FWHM) Page 21
00
20
ff
fi
fl(single shot) Transformer ratio: 1.26
Spectral density (p
100 Plasma on, driver Energy-transfer efficiency: 39%
(imaging scan)
0.16% 0.13%
High-quality, efficient acceleration for sustainable applications 50
FWHM FWHM
Beam-loading facilitates 42% energy-transfer efficiency, 0.2% energy spread with full charge coupling
0
990 1000 1010 1020 1030 1040 1050 1060 1070 1080
Energy (MeV)
(a) C.A. Lindstrøm et al., PRL 126, 014801 (2021) (b)
Charge density (pC mm -1 MeV -1 )
0
1
x (mm)
400
0 Trailing 40
Driver Plasma off
bunch Plasma off
-1 300 1500
1000
Shots (chronological order)
Shots / 0.02%
35
2 (c)
Spectral density (pC MeV -1 )
Plasma on 1000
1 200
x (mm)
30
500
0 2000
-1 100 0 25
0 0.1 0.2 0.3 0.4
-2 Energy spread (% FWHM)
0 20
Spectral density (pC MeV -1 )
3000
Shots / 2.5%
150 Plasma off Single-shot statistics: 1500 (d) 15
Plasma on Accel. gradient (peak): 1.28 GV/m 1000
(single shot) Transformer ratio: 1.26 500 10
Energy-transfer efficiency: 39% 4000
100 Plasma on, driver 0
(imaging scan) 0 10 20 30 40 50 5
0.16% 0.13% Efficiency (%)
FWHM FWHM 5000 0
50
990 1000 1010 1020 1030 1040 1050 1060 1070 1080
Energy (MeV)
0 decelerated accelerated
990 1000 1010 1020 1030 1040 1050 1060 1070 1080 driver bunch witness bunch
Energy (MeV)
(b) 0
> Accelerating gradient of 1.3 GV/m > 0.2% energy spread (input 0.16%)
Plasma off 40
(improvement by factor 10 over state-of-the-art)
1000 > Energy gain 45 MeV (over
1500 3.5 cm distance) of 100 pC witness,
ronological order)
Shots / 0.02%
35
(c) density (pC MeV -1 )
with energy spread of 1.4 MeV FWHM and no charge loss
1000
> (42±4)% energy transfer ef ciency
30
500
2000 > Few-percent-level wake
0
eld attening demonstrated
25
(improvement by factor 3 over state-of-the-art)
0 0.1 0.2 0.3 0.4
| Jens Osterho | MT Days | February 3, 2021 Energy spread (% FWHM) Page 22
00
20
ff
fi
fi
flFLASHFORWARD‣‣ roadmap aims at 10 kW with high beam quality
Plan covers major plasma accelerator challenges
108
Accelerating gradient:
107 0.01 - 0.1 GV/m
0.1 - 1 GV/m LCLS-II
ILC ILC
LCLS-II
106
2018 > 1 GV/m
average power Pavg. / W
Plasma dechirper
• D’Arcy et al., PRL 122, 034801 (2019) 105 XFEL
2019 Energy depletion and energy doubling 104 10 kW FLASHFORWARD
KALDERA
103
Wake eld sampling
AWAKE FACET
• Schröder et al., Nat. Commun. 11 5984 (2020) 102 FLASHF
FL ORWARD
2020 Energy spread preservation by beam loading control 101
BELLA
LUX
kHz LOA
• Lindstrøm et al., PRL 126, 014801 (2021) 100 2
10 10 1 100 101 102 103 104 105 106 107
Emittance preservation bunches per second N / s 1
High overall e ciency and gain for sustainable operation
2022 10 kW stage with 50% e ciency
Detection of slice properties with fs resolution & beam quality conservation
kHz-to-GHz plasma response ➞ FLASH: increase FEL energies,
2024 10 kW avg. power operation
access oxygen K-edge
at 2.33 nm wavelength
2026
2030
| Jens Osterho | MT Days | February 3, 2021 Page 23
00
fi
ff
ffi
ffiFLASHFORWARD‣‣ roadmap aims at 10 kW with high beam quality
Plan covers major plasma accelerator challenges
2018 Plasma dechirper
• D’Arcy et al., PRL 122, 034801 (2019)
2019 Energy depletion and energy doubling
Wake eld sampling
• Schröder et al., Nat. Commun. 11 5984 (2020)
2020 Energy spread preservation by beam loading control
• Lindstrøm et al., PRL 126, 014801 (2021)
Emittance preservation
High overall e ciency and gain for sustainable operation
2022
Detection of slice properties with fs resolution
kHz-to-GHz plasma response
2024 10 kW avg. power operation
2026
2030
| Jens Osterho | MT Days | February 3, 2021 Page 24
00
fi
ff
ffiSimulations play a crucial role for research at FLASHFORWARD‣‣
Long-time-scale plasma dynamics challenge current capabilities
> Accurate simulations are essential > Simulations for plasma accelerators require High-Performance Computing (HPC)
- to predict new phenomena - ~M core hour for single simulation with full particle-in-cell (PIC) scheme
- to prepare and plan experimental studies ➞ ensemble of simulations are (prohibitively) expensive
- to verify and analyze measurements - Development of specialized codes / ef cient and accurate algorithms critical
> HPC is a dynamic eld
- Performance portability on heterogeneous platforms required
- Inter-operability of HPC tools
- Advanced numerical methods and AI increasing in importance (➞ ACCLAIM)
Maxence Thévenet
> New group on Plasma Accelerator Theory and Simulations Group leader
State-of-the-art code development Enabling ensembles of S2E multi-physics simulations
- WarpX – full 3D electromagnetic, open-source, GPU (LBNL) - Adoption of the openPMD I/O standard (HZDR)
- FBPIC – quasi-RZ, Python, open-source, GPU (LBNL, UHH, DESY) - AI to improve productivity (UHH & DESY)
- Hipace – quasi-static, 3D, work in progress… (DESY, LBNL) - Capability to study long-time plasma dynamics
| Jens Osterho | MT Days | February 3, 2021 Page 25
00
ff
fi
fiUnderstanding ultimate repetition rate limits of plasma accelerators
Long-time-scale plasma dynamics challenge current capabilities
- Need to simulate
> 104 plasma oscillations to
investigate plasma recovery
- Requires new ideas and
new low-noise codes
- Critical to understand
energy dissipation, power
density limits, repetition
rate limits
- Will catalyze the experimental
progress at FLASHForward
and beyond
Maxence Thévenet
> New group on Plasma Accelerator Theory and Simulations Group leader
State-of-the-art code development Enabling ensembles of S2E multi-physics simulations
- WarpX – full 3D electromagnetic, open-source, GPU (LBNL) - Adoption of the openPMD I/O standard (HZDR)
- FBPIC – quasi-RZ, Python, open-source, GPU (LBNL, UHH, DESY) - AI to improve productivity (UHH & DESY)
- Hipace – quasi-static, 3D, work in progress… (DESY, LBNL) - Capability to study long-time plasma dynamics
| Jens Osterho | MT Days | February 3, 2021 Page 26
00
ffProgress in Plasma Booster R&D at FLASHFORWARD‣‣
Summary and outlook
Develop a self-consistent plasma-accelerator stage
with high-e ciency, high-quality, and high-average-power
High e ciency High beam quality High average power
✓Transfer e ciency ✓ Energy-spread preservation High repetition rate
Driver depletion Emittance preservation
• Impactful and exciting research programme will help advance plasma accelerators to application-readiness
ffi
ffi
ffiYou can also read
