Download this is cooling!

Cooling for Hadron Beams
Yaroslav Derbenev
Thomas Jefferson National Accelerator Facility
The 4th EIC Workshop
Hampton, May 19-23, 2008
Cooling for HB
Outline
•
•
•
•
•
•
•
•
Introduction
Stacking positive ions
Electron Cooling for eRHIC
EC for ELIC
Optical Stochastic Cooling
Coherent Electron Cooling
Thinking on prioritizing the cooling techniques
Conclusions
Ya. Derbenev /TJNAF/ . 4th EIC Workshop, Hampton, May 2008
Preamble: Interaction Region as a Microscope
•Beam emittance:      --invariant
•Luminosity (p-e round beams):
 f f
 * *
 00
Beam-beam tune shift:
*
*
2 
•Requirement to bunch length:  z  *  b  F

 2f
•
•
 *  F f  F
Small transverse and longitudinal beam emittance
allows one to design and use a strong final focus:
b* about 5 mm or even shorter can be designed!
Chromaticity F  Fp / p can be an obstacle, but it can be
compensated (we know exactly how to do so!)
The (6D) emittance is a subject to change by cooling!
Ya. Derbenev /TJNAF/ . 4th EIC Workshop, Hampton, May 2008

f
What is Beam Cooling?
• One needs a media with which the beam should
interact incoherently
• An individual charged particle creates an effective
charge polarization of the media (image charge)
• In response, particle motion is effected by the image
field, that decelerates the particle relatively the
beam frame - this is cooling!
• Feedback with the right receiver - kicker phasing
can be used in order to amplify the response
• Influence of image fields from particle neighbors can
only decrease the cooling effect (shield effect)
• The achievable cooling rate and equilibrium is limited
by noise in the media and amplifier
Ya. Derbenev /TJNAF/ . 4th EIC Workshop, Hampton, May 2008
What cooling does for luminosity?
• Decrease of beam emittances (up to reaching the
beam-beam or intra-beam space charge limit)
• Preventing the beam blow-up by IBS and other slow
heaters
• Small transverse emittance allows one to design
a tight star-focus (more beam extension is available
in final quadrupoles)
• Short bunches allow one to:
- use the designed low beta-star (no “hour glass”)
- implement the crab-crossing beams – hence,
increase the bunch collision rate
• Beam stacking after linac (works for all species!)
Ya. Derbenev /TJNAF/ . 4th EIC Workshop, Hampton, May 2008
Cooling techniques and ideas
Maxwell-Lorentz demon
Radiation Cooling (RC)
1950th
works in e± colliders, can be used in LHC..
Shrinking before plague (muons…)
Ionization Cooling (IC)
1966/1981
Electron Cooling (EC)
1966
Stochastic Cooling (SC)
under development towards the
Neutrino Factory and Muon Collider!
Thermostat of the relativistic engineer
-used at low and medium energies
-under development for colliders
van der Meer’s demon (RF feedback)
used
1968
Coherent Electron Cooling (CEC)
1980
Spoiled hybrid of EC and SC
development just started…
Optical Stochastic Cooling (OSC)
1993
Max’s demon
(RC+optical feedback)
development to start…
Ya. Derbenev /TJNAF/ . 4th EIC Workshop, Hampton, May 2008
Electron Cooling:
The thermostat of a relativistic engineer
Do not renounce
from prison and
money bag
Landau liked to call me
“The relativistic engineer”,
I am very proud of that.
Gersh Budker
Kinetic equation (plasma relaxation) was derived by Landau
in 1937. But… can it work for beams? It does! Yet very
interesting and important phenomena have been discovered
(magnetized cooling, super-deep cooling, christaline beams…)
Ya. Derbenev /TJNAF/ . 4th EIC Workshop, Hampton, May 2008
Stochastic cooling
The van der Meer’s demon
“Is n’t it the Maxwell’s demon?”
(G.Budker)
( c ) 
N
2f
It works!!
•Works well for a coasted low current,
large emittance beams.
•Can it work for bunched beams? Hardly… but
demonstrated by M.Blaskewitz for lead at RHIC!
•May help ELIC (eRHIC?) (stacking and precooling of the coasted beam)…?
Ya. Derbenev /TJNAF/ . 4th EIC Workshop, Hampton, May 2008
Stochastic Cooling rate as function of # pbars
(‘Fast Schottky’)
Fast V
Fast H
0.0
-0.5
-1.0
-1.5
2-4 GHz vertical
-2.0
-2.5
Np = 186e10
Np = 124e10
Np = 51e10
Cooling rate [  mm mrad per hour]
Cooling rate [  mm mrad per hour]
Fast H
0.0
-0.5
-1.0
-1.5
2-4 GHz horizontal
-2.0
-2.5
Np = 186e10
Np = 124e10
Cooling rate [  mm mrad]
Fast H
A 1/N-like dependence is clearly
visible for the 2-4 GHz system but
not for the other two systems
(although the 1-2 GHz system shows
some dependence on the # of pbars)
Np = 51e10
Fast V
0.0
-0.5
-1.0
-1.5
1-2 GHz horizontal
-2.0
-2.5
Np = 186e10
Ya. Derbenev /TJNAF/ . 4th EIC Workshop, Hampton, May
Fast V
Np = 124e10
Np = 51e10
400
3.5
350
3
300
2.5
250
2
200
1.5
150
1
100
0.5
50
0
0
Vertical e-beam offset, mm
Number of antiprotons, E10
EC in Fermilab’s Recycler ring
Electron cooling is used at the
Recycler for cooling 8 GeV pbars
4.34 MeV, 0.1 A DC electron beam
Primarily longitudinal cooling
To free space for the next
antiproton injection from
Accumulator ring
-0.5
0
5
10
15
20
140
14
120
12
100
10
80
8
60
6
40
4
20
2
0
0
0
5
10
15
Hor. emittance, mm*mrad
Long. emittance, eVs
Time, hr
20
Time, hr
Ya. Derbenev /TJNAF/ . 4th EIC Workshop, Hampton, May
Cooling strength is regulated by
a vertical offset of the electron
beam in the cooling section
Most of the time, e- beam is
shifted by ~2 mm to maximize
the life time
Moved on axis before
extraction to minimize
emittances of the extracted
beam
Forming high current ion beams
(agenda in general)
1.Stacking in pre-booster or
an accumulator ring
2. Accelerating in the pre-booster
and large booster
3. Pre-cooling in collider ring at
injection energy
4. Continuous cooling
in collider mode
Ya. Derbenev /TJNAF/ . 4th EIC Workshop, Hampton, May 2008
Stacking ions
• 1A ion current is required for high luminosity
• Stripping injection into pre-booster can be used for proton and
deuteron beam
• Accumulation of He and other ions only is possible at use of beam
cooling
• Stochastic cooling of 1A beam is too slow
• The DC EC (or CEC) is promising!
• There is an experience:
- Accumulating polarized protons from positive source in IUCF
Proton Cooler Ring
- Electron cooling of 3.10^12 pbar’s in Recycler ring of FNAL
• An issue for beam accumulation in pre-booster at 200 MeV/n: large
emittance caused by the space charge… This emittance cannot be
“managed” by Electron pre-Cooling in collider ring
• A solution seems to exist…
Ya. Derbenev /TJNAF/ . 4th EIC Workshop, Hampton, May 2008
Stochastic Cooling and Stacking of Ions
An optimistic scenario








Multi-turn (10 – 20) injection from
285 MeV SRF linac to pre-booster
Stochastic damping of injected
beam
Accumulation of 1 A coasted beam
at space charge limited emittance
RF bunching and accelerating to 3
GeV
Inject into large booster
Fill large booster, accelerate to 30
GeV
Inject into collider ring
Transverse stochastic cooling of 1
A coasted ion beam in collider ring
Stacking proton beam in prebooster with stochastic cooling
Beam Energy
MeV
200
%
1
Pulse current from linac
mA
20
Cooling time
min
1.5
Accumulated current
A
1
Stacking cycle /fill LB
duration
min
4/60
Beam emittance, norm.
μm
12
Momentum Spread
Laslett tune shift
At this stage, Ion beam is ready for
electron cooling
Ya. Derbenev /TJNAF/ . 4th EIC Workshop, Hampton, May 2008
0.03
Stacking positive ions by use of DC EC
Stacking of positive fully stripped polarized and unpolarized ions can
be realized in the accumulator-cooler ring (ACR) with electron cooling.
Such method has been successfully used for accumulating of
polarized beam in Proton Cooler Ring of IUCF
Accumulator Ring with DC EC
Design advancements
:
Circumference
Arc radius
Crossing straights length
Energy/u
Electron current
Electron energy
Cooling time for protons
Stacked ion current
Large norm.emittance
after stacking
Accumulation time
m
m
m
MeV
A
KeV
ms
A
µm
s
50
3
2 x 15
200
1
100
10
1
16
10
Ya. Derbenev /TJNAF/ . 4th EIC Workshop, Hampton, May 2008
Beam stacking in one o f two
circular modes matched with
solenoid of cooling section
This concept allows one to
overcome space charge
limit on 4D emittance,
hence, increase EC precooling rate in collider ring
Requirements for eRHIC
1. Protons (250 GeV):
Bunch intensity: 2e11
95%, normalized emittance: 6 mm
Rms bunch length: 0.2 m
Peak luminosity: 2.6e33
2. Au ions (100 GeV):
Bunch intensity: 1.1e9
95%, normalized emittance: 2.4 mm
Rms bunch length: 0.2 m
Peak luminosity: 2.9e 33
Requires pre-cooling:
Au ions: factor of 6 reduction in initial emittance
Protons: factor of 3-2 reduction in initial emittance
Such electron cooling is possible based on RHIC-II electron
cooler with 703 MHz SRF gun.
Ya. Derbenev /TJNAF/ . 4th EIC Workshop, Hampton, May 2008
Pre-cooling of Au ions (N=1e9)
Ya. Derbenev /TJNAF/ . 4th EIC Workshop, Hampton, May 2008
Pre-cooling of protons (N=2e11)
Ya. Derbenev /TJNAF/ . 4th EIC Workshop, Hampton, May 2008
Electron cooling section at RHIC 2 o’clock IP
RHIC triplet
solenoids
RHIC triplet
e-
100 m
`
IP
2
helical wigglers
e-
Each electron beam cools ions in Yellow ring of
RHIC then the same beam is turned around and
cools ions in Blue ring of RHIC.
Ya. Derbenev /TJNAF/ . 4th EIC Workshop, Hampton, May 2008
e-
ERL
10 m
Energy Recovery Linac (ERL) for RHIC-II
Cooling of Au ions at 100 GeV/n:
• 54.3 MeV electron beam
• 5nC per bunch
• rms normalized emittance < 4 mm
• rms momentum spread < 5×10-4
Courtesy D. Kayran
D. Kayran, PAC07
Ya. Derbenev /TJNAF/ . 4th EIC Workshop, Hampton, May 2008
R&D ERL at BNL
0.5 ampere SRF gun – under construction
703 MHz 5 cell SRF cavity
designed by BNL and
fabricated by Advanced
Energy Systems.
The cavity has reached its
specification of 20 MV/m
at a Q of 1×1010 .
5 cell cavity successfully processed
Ya. Derbenev /TJNAF/ . 4th EIC Workshop, Hampton, May 2008
EC Test bed for RHIC-II and eRHIC
Commissioning starts February 2009
Ya. Derbenev /TJNAF/ . 4th EIC Workshop, Hampton, May 2008
Circulated Electron Cooling of ELIC
Max/min energy of e-beam
MeV
125/8
Electrons/bunch
1010
1
Number of bunch revolutions in CCR
solenoid
ion bunch
Current in CCR/ERL
A
3/0.03
MHz
1500/15
CCR circumference
m
80
Cooling section length
m
20
Circulation duration
ms
27
Bunch length
cm
1-3
Energy spread
10-4
1-3
T
2
Beam radius in solenoid
mm
1
Cyclotron beta-function
m
0.6
Thermal cyclotron radius
mm
2
Beam radius at cathode
mm
3
Solenoid field at cathode
KG
2
Bunch repetition rate in CCR/ERL
circulator
cooler
ring
(CCR)
electron
bunch
Cooling
section
Solenoid field in cooling section
kicke
r
kicke
r
energy
recovery
path
SRF
Linac
100
Laslett’s tune shift in CCR at 10 MeV
Time of longitudinal inter/intra beam heating
electron
injector
dum
p
0.03
ms
200
• CCR makes 100 time reduction of beam current from
injector/ERL
• Fast kickers operated at 15 MHz repetition rate and 2
GHz frequency bend width are required
Ya. Derbenev /TJNAF/ . 4th EIC Workshop, Hampton, May 2008
Cooling Time and Ion Equilibrium
Cooling rates and equilibrium of proton beam
Multi-stage cooling scenario:
• 1st stage: longitudinal
cooling at injection energy
(after transverses stochastic
cooling)
• 2nd stage: initial cooling after
acceleration to high energy
• 3rd stage: continuous cooling
in collider mode
Parameter
Unit
Energy
GeV/MeV
Valu
e
Value
30/1 225/123
5
Particles/bunch
1010
0.2/1
Initial energy spread*
10-4
30/3
1/2
Bunch length*
cm
20/3
1
Proton emittance, norm*
mm
1
1
Cooling time
min
1
1
mm
1/1
1/0.04
Equilibrium bunch
length**
cm
2
0.5
Cooling time at
equilibrium
min
0.1
0.3
0.04
0.02
Equilibrium emittance
, **
x / y
Laslett’s tune shift
(equil.)
* max.amplitude
** norm.,rms
Ya. Derbenev /TJNAF/ . 4th EIC Workshop, Hampton, May 2008
Electron Cooling of Ions in ELIC
Staged cooling
 Start electron cooling (longitudinal) in collider ring at injection energy,
 Continue electron cooling (in all dimension) after acceleration to high
energy
Sweep cooling
•
•
dN p (t  0)
dN p (t )
dv||
dv||
After transverse stochastic
cooling, ion beam has a small
transverse temperature but large
longitudinal one.
Use sweep cooling to gain a factor
of longitudinal cooling time
0
F(v
)

v e (t )
Dispersive cooling
Compensates for lack of transverse cooling rate
at high energies
Flat beam cooling
Based on flattening ion beam by reduction of coupling
IBS rate at equilibrium becomes reduced compared to cooling rate
Ya. Derbenev /TJNAF/ . 4th EIC Workshop, Hampton, May 2008
v
Injector and ERL for EC in ELIC
•
•
•
•
ELIC CCR driving injector
 30 mA@15 MHz, up to 125 MeV energy, 1 nC bunch charge, magnetized
Challenges
 Source life time: 2.6 kC/day (state-of-art is 0.2 kC/day)
 source R&D, & exploiting possibility of increasing evolutions in CCR
 High beam power: 3.75 MW  Energy Recovery
Conceptual design
 High current/brightness source/injector is a key issue of ERL based light
source applications, much R&D has been done
 We adopt light source injector as initial baseline design of ELIC CCR driving
injector
Beam qualities should satisfy electron cooling requirements (based on previous
computer simulations/optimization)
SRF modules
solenoids
500keV
DC gun
buncher
quads
Ya. Derbenev /TJNAF/ . 4th EIC Workshop, Hampton, May 2008
Fast Kicker for Circulator Cooling Ring
•
•
•
•
•
Sub-ns pulses of 20 kW and 15 MHz are
needed to insert/extract individual bunches.
RF chirp techniques hold the best promise
of generating ultra-short pulses. State-of-Art
pulse systems are able to produce ~2 ns, 11
kW RF pulses at a 12 MHz repetition rate.
This is very close to our requirement, and
appears to be technically achievable.
Helically-corrugated waveguide (HCW)
exhibits dispersive qualities, and serves to
further compress the output pulse without
excessive loss. Powers ranging from up10
kW have been created with such a device.
Estimated parameters for the kicker
Beam energy
MeV
125
Kick angle
10-4
3
Integrated BdL
GM
1.25
Frequency BW
GHz
2
Kicker Aperture
Cm
2
Peak kicker field
G
3
Kicker Repetition
Rate
MHz
15
Peak power/cell
KW
10
Average power/cell
W
15
Number of cells
20
20
Collaborative development plans
include studies of HCW,
optimization of chirp techniques,
and generation of 1-2 kW peak
output powers as proof of
concept.
kicker
Kicker cavity design will be
considered
Ya. Derbenev /TJNAF/ . 4th EIC Workshop, Hampton, May 2008
kicker
Coherent Electron Cooling
or
how the van der Meer’s demon can spoil
thermostat of the relativistic engineer
Ya. Derbenev /TJNAF/ . 4th EIC Workshop, Hampton, May 2008
History of CEC idea (1980-91-95-2007)
Coherent electron cooling (CEC) was proposed 27 years ago
• General idea: amplify response of ebeam to an ion by a micro-wave instability
of the beam
• A few instabilities have been shown
• CEC advantages/disadvantages
compared to:
EC :
Gain in cooling rate
Complicate BT
SC :
Very large FB (30 GHz – optics)
Precise phasing required
OSC : Effective in a wide energy range
Small signal delay
Intense e-beam required
Signal gain is limited
What changed in last 10 years?
• Relativistic DC EC realized (FNAL)
• ERL realized (JLab)
• SASE FEL realized (UCLA - DESY)
• ERL-based HEEC on the way (BNL)
And more…
Ya. Derbenev /TJNAF/ . 4th EIC Workshop, Hampton, May 2008
FELs and HEEC
And so, my fellow FELers, ask
not
what storage ring can do for
FELs:
Ask what FELs can do
for your storage rings!
Vladimilr Litvinenko
29th International FEL Conference
August 26-31, 2007, BINP, Novosibirsk
Ya. Derbenev /TJNAF/ . 4th EIC Workshop, Hampton, May 2008
And so, my fellow Americans,
ask not what your country
can do for you;
ask what you can do
for your country.
Optical stochastic cooling (OSC)
Max’s demon (by Max Zolotorev, 1993)
----Proton radiates light in undulator (“receiver”)
----optical amplifier------“proton interacts with the amplified light in
the end undulator “kicker”)
2000 - OSC considered for cooling of halo particles
in the VLHC, by Barletta, Chattopadhyay, Zholents, Zolotorev
Ya. Derbenev /TJNAF/ . 4th EIC Workshop, Hampton, May 2008
Why we need OSC if we have electron
cooling
1.5
10
Cooling rate
Beam profile
8
.
1
OSC
6
4
0.5
.
Electron cooling
0
0
2
4
6
8
2
10
0
4
2
0
2
4
Optical stochastic cooling can effectively reduce number of the
particles in the beam tails. This will increase lifetime of the
beam and reduce detectors background.
Custom beam shaping might be possible.
Ya. Derbenev /TJNAF/ . 4th EIC Workshop, Hampton, May 2008
Optical Stochastic Cooling
Horizontal and longitudinal
cooling time 1hr for the full gold
beam requires 16W of power
l = 12 mm
Bandwidth ~ 3 THz
LDRD test at ATF for Optical
Parametric Amplification
Gold beam
Optical amplifier on 3.5 cm CdGeAs2
crystal with d14=236 pV/m
s
p
p
s
Pump laser
i = p - s;
kp = k s + ki
Amplified
signal
Ya. Derbenev /TJNAF/ . 4th EIC Workshop, Hampton, May 2008
Freshly grown crystals at
Lockheed Sanders, NH
Optical Parametric Amplifier OPA
3 cm length crystal
→ intensity gain 3 105
Parametric process is photon interaction in which one
high frequency photon is annihilated and two lower
frequency photons are created, i.e.: (pump) =(signal)
+(idler) where photon energy is conserved. In
addition, photon momentum is conserved by the wave
vectors: k(pump) =k(signal) +k(idler) .
Ya. Derbenev /TJNAF/ . 4th EIC Workshop, Hampton, May 2008
Critical studies on Cooling for Hadrons
• Pre-cooling for eRHIC: Non-magnetized, space charge
dominated e-beam transport
• CEC for eRHIC: all physics and design
• EC for ELIC: ERL - CCR super-fast kicker for e-beam
• OSC (of “beam halo”): Parametric optical amplifier
• SC for beam stacking and pre-cooling (ELIC):
Limits for coasted beams
• DC EC (and CEC) for beam stacking (alternative to SC):
Accumulator-cooler design and all beam physics
Ya. Derbenev /TJNAF/ . 4th EIC Workshop, Hampton, May 2008
My view of Current prioritization of potential
employment of cooling techniques at EIC
by taking into account :
Potential efficiency + Conceptual readiness + Proximity to realization
Service:
Stacking
Pre-cooling
Core Cooling
Tail cooling
I
DC EC
HE EC
CR EC
OSC
II
DC CEC
CR EC
HE CEC
HE CEC
III
SC
SC
IV
HE CEC
HE EC
EC – electron cooling; DC - direct current; HE - high energy (ERL based)
CR – cirulator ring; CEC - coherent EC;
SC - stochastic cooling;
OSC - optical stochastic cooling
Ya. Derbenev /TJNAF/ . 4th EIC Workshop, Hampton, May 2008
Conclusions
• The studies and designs of cooling for EIC are getting
mature. The projects are well profiled in the goals and
expected performance.
• The engineering works and test studies for 1 hundred
MeV scale ERL started at BNL. A prototype of ERL will
be built, which, as minimum, will serve pre-cooling of
ions to reach
luminosity in eRHIC.
• Conceptual development of 125 MeV ERL + Circulator
Ring scheme of electron cooling for ELIC will finalize
soon (in 1-1.5 year) at JLab. Could such scheme also be
implemented in eRHIC to raise the cooler efficiency?
Ya. Derbenev /TJNAF/ . 4th EIC Workshop, Hampton, May 2008
Conclusions (cont-d)
• Recent development of Optical Stochastic Cooling
brings a realistic promise to extend luminosity lifetime
and ease the EC design, by cooling beam tail. The OSC,
however, can work only in a relatively narrow beam
energy region of the collider ring.
• The new perspectives for cooling hadron beams in
colliders seem to be open with development of the
Coherent Electron Cooling, which starts at BNL. It invites
SASE FEL amplifier to cooling section of a collider.
• The efficiency of colliders can be frequently increased
by stacking positive ions in a small ring after linac. Here,
the EC or CEC on a non-relativistic DC e-beam could be
employed to accumulate 1 A level of ion current.
Ya. Derbenev /TJNAF/ . 4th EIC Workshop, Hampton, May 2008
We entering an exciting era of
the new generation of
high luminosity colliding beams
provided by the
Advanced cooling techniques!
Thank you!
Ya. Derbenev /TJNAF/ . 4th EIC Workshop, Hampton, May 2008
Making sense of the different cooling techniques
•Stochastic Cooling: beam error signal
is measured by a pickup, amplified by
RF amplifier and corrected in the kicker
Works over wide range of the ion beam
energies
Can be used for stacking
Limited by amplifier bandwidth ~5GHz
Possibly can cool only tails of the RHIC
beam
•
Optical SC: Similar to SC, but at
optical wavelength
Not limited by amplifier bandwidth
~3THz
Can cool whole RHIC beam in one hour
with16W of optical power
Favors large amplitudes,
Greatly reduces requirements on
electron current if used with EC
Works only over limited ~5-10% range
of the ion beam energies
Requires challenging RHIC lattice
modifications
No demo, yet
•Electron cooling: ion beam “heat”
transferred to “cold” electron beam
Works over wide energy range
Effective for the beam core and against
IBS (faster cooling for the colder beam)
Complimentary with OSC as it effective
against IBS (main growth in RHIC)
Requires ERL and long solenoid
High e-efficiency in the ERL+CCR
advance scheme
Can be used for stacking
Was not demonstrated at high energies
• Coherent Electron Cooling: response of
e-beam to an ion is amplified by a microwave instability of the e-beam (SASE FEL)
Strong image interaction
Very large FB (30 GHz – optics)
Effective in a wide energy range
Small signal delay
Requires high current ERL
Signal gain is limited
Needs more study, no demo yet
Ya. Derbenev /TJNAF/ . 4th EIC Workshop, Hampton, May 2008