status on superb effort daresbury, april 26, 2006 p. raimondi
TRANSCRIPT
Status on SuperB effort
Daresbury, April 26, 2006
P. Raimondi
Outline
• Basic Concepts (March-Sept,2005)• Parameters and layout optimization based on a
High-Disrupted regime (Nov, 2005)• Parameters and layout optimization for a
Minimal-Disrupted regime (Jan, 2006)• Layout for a Ring Collider with Linear Collider
Parameters (Mar, 2006)• Conclusions
Basic concepts• SuperB factories based on extrapolationg
current machines require:
• Higher currents
• Smaller damping time (weak function ^1/3)
• Shorter bunches
• Higher power
SuperB gets very expensive and hard tomanage, expecially all the problems relatedto the high current => look for alternatives
• Basic Idea comes from the ATF2-FF experiment
In the proposed experiment it seems possible to acheive spot sizes at the focal point of about 2m*20nm at very low energy (1 GeV), out from the damping ring
• Rescaling at about 10GeV/CM we should get sizes of about 1m*10nm =>
• Is it worth to explore the potential of a Collider based on a scheme similar to the Linear Collider one
Idea presented at the Hawaii workshop on Super-B factory on March-2005
LinearB scheme
IPLER Bunch compressor and FF
HER Bunch compressor and FF
LERHER
Overall rings lenght about 6Km,Collision frequency about 120Hz*10000bunch_trains=1.200MHzBunch train stays in the rings for 8.3msec, then is extracted, compressed and focused. After the collision is reinjected in its ring
LER injection Her injection
Scaling laws to optimize the IP parameters
• Disruption:
• Luminosity
• Energy spread:
yx
NL
2
yx
zN
D
zx
E
N
2
2
Decrease z + decrease NIncrease spotsize
Increase NDecrease spotsize
Increase z + decrease NIncrease spotsize
Contraddicting requests!
A lot of homework done in collaboration at SLAC and at the LNF for a few months.
Explored the parameters phase in order to maximize the luminosity per crossing
Laid down all the other possible advantages (e.g. less current through the detector, smaller beams, smaller beam pipe etc)
Leading to a workshop held on Nov,11-12 2005 in Frascati to investigate and optimize the scheme and the feasibility of the different subsystems
Horizontal Collision Vertical collision
Effective horizontal size during collision about 10 times smaller, vertical size 10 times larger
Simulation by D.SchulteFirst attempt
Horizontal phase after the collision
Vertical phase after the collision
IP Parameters set considered at the workshop causedlarge increase of the emittance due to the collision:
x_out/x_in=12 y_out/y_in=300M. Biagini studies
Linear Super B schemes with accelerationand energy recovery, to reduce power
e- Gun2GeVe+ DR IP
5GeV e+ SC Linac
e- Dump7GeV e+
4 GeV e-
4GeV e- SC Linac
2 GeV e+ injection
2 GeV Linac1.5 GeV Linac 1.5 GeV Linac
Linac
Damping Rings2 GeV
Linac
e+ Gun e- Gun
• Use SC linacs to recover energy• Use lower energy damping rings to
reduce synchrotron radiation• No electron damping ring• Make electrons fresh every cycle
– Damping time means time to radiate all energy
– Why not make a fresh beam if storage time is greater than 1 damping time
Progress in design optimization after the 1° SuperB workshop
Between December-2005 and March-2006 a lot of studies have been made in order to understand what are the sources of the blow-ups in the collision and how to minimize then.
Power requirements could be greatly reduced if collision is less disruptive
Search for a trade off between luminosity delivered in one collision and power spent for each collision
Search for the simplest and more economic solution
x=10m y=10nm z=250um
x=33mm y=1mm
x=3.3*10-9 y=10-13
Lnorm=Ld/log(y_out/y_in)=2.27 at Npart=2.5*1010
Introduced the luminosity merit function: Lnorm=L/log(ey_out/ey_in) => Luminosity per energy cost
Luminosity vs bunch charge for a given set of IP parameters
Outgoing/incoming emittances ratiovs bunch charge for a given set of IP parameters
Ey_after collision increases fast with current and 1/x: with Npart=2*10^10 Lnorm=2.27
Cure 1: decrease y from 1mm to 250um =>
y_in=4*10-13 y_out=23.4*10-13 y_out/y_in=6 instead of 24 Ld drops from 7.22*1032 to 5.6*1032 (hourglass) Lnorm=3.12 Cure 2: travelling focus (waist position is shifted by z wrt
the IP for the slice ‘z-z+dz’=> Ld goes back up to 7.01*1032 Lnorm=3.91
Cure 3: further decrease y from 250um to 100um and use the pinch to keep the beams from diverging rather than to disrupt them with overfocusing=>
BB FOCUS COMPENSATION CONCEPT !!!!!!!!!
y_in=10*10-13 , y_out(slice)=15.0*10-13,
y_out/y_in=1.5 instead of the initial 24 Ld drops to 5.77*1032 Lnorm=14.1 !!
With travel focus
Without travel focus
Horizontal collisions in a round beam case
3 planes slice emittances after the collision (round beams case), each color is a different slice: red head of the bunch, black tail
Without Travel Focus
y_out/y_in=3
With Travel Focus
y_out/y_in=1.1
In summary, the small disruption regime requires:
small z (=> large e from compressor)
big x
small y (for luminosity) and y
BB-compensation by traveling focus all the requirements do fit togheter with the horizontal monocromator:
betatron x small, and made large by adding horizontal dispersion with opposite sign for the 2 beams. High energy particles of first beam collide with low energy particles of the sencond one
it simultaneneously enlarge x and decrease the luminosity energy spread moreover since the natural horizontal emittance is small,
the emittance ratio of about 0.5% ensures the small y
Single pass Collider
Ring with compressor
Ring without compressor
Sigx* m 30 (2 betatron) 18 (2 betatron) 2.67Etax mm +-1.5 +-5.0 0.0Sigy nm 12.6 63.2 12.6Betx mm 5.0 5.0 9.0Bety mm 0.080 0.500 0.080Sigz_IP mm 0.100 0.600 6.0Sige_IP 2.0e-2 3.5e-3 1.3e-3Sige_Lum 1.0e-3 1.0e-3 0.9e-3Emix nm
0.8 0.8 0.8
Emiy nm
0.002 0.008 0.002
Emiz m 2.0 2.0 8.0Cross_angle mrad Optional >2*12 2*25Sigz_DR mm 4.0 4.0 6.0Sige_DR 0.5e-3 0.5e-3 1.3e-3Np 10e10 7.0 2.0 2.3Nbunches 12000 12000 12000DR_length km 6.0 6.0 6.0Damping_time msec 20 20 20Nturns_betwe_coll 50 1 1Collision freq MHz 12.0 600 600Lsingleturn 1e36 1.5 1.7 1.2Lmultiturn 1e36 1.1 0.9 1.0
Single pass Collider with Np=7*1010
Nbunches=12000 (6Km ring)
Travel focus in vertical plane
x=1.5mm (opposite sign for the two beams), x*=30m
z=100m z=4mm in DR
e=100MeV e/e=2*10-2 e/e=5*10-4 in DR
e_Luminosity=7MeV
x=0.8nm x_norm=8m
y=0.002nm y_norm=20pm
z=2.0m Stored time between collision=1msec=50turns
Lmultiturn=0.9*1036 (Lsingleturn=1.5*1036)
FF with large energy spread tricky
Multiturn Simulation for flat case6Km ring,Np= 7*1010
coll_freq=1Khz*12000 Lmultiturn=0.9*1036
Compressor
Compressor Decompressor
DeCompressor
IP
OptionalAccelerationand deceleration
OptionalAccelerationand deceleration
FF FF
ILC ring with ILC FFILC compressorColliding every 50 turnAcceleration optionalCrossing angle optional
Now the acceleration is not needed anymore in order to reduce the power
Simplified layout in the Small Disruption Regime
• Equilibrium Emittance Vertical blowup about 60%• Blowup as function of beam currents almost linear• Blowup as function of damping time goes like Tau1/3
• Reducing the bunch charge by a factor 6 (1010), the blowup at the equilibrium decreases to 10%
• Reducing the damping by a factor 50 (collision every turn) equilibrium blowup increases by a factor 4 (501/3)
• Final Blowup in this case is about 40%• Geometric Luminosity decreases by a factor 36 due to
less charge and increases by a factor 50 for increased collision rate
• With the same parameters but colliding in the ring (bunch compressor and FF in the ring), we get:
L=1036 with Npart=1010 and L=4*1036 with N=2*1010 => overhead in luminosity
Possible to reoptimize the parameters in this scenario
Scaling the parameters to a Ring Colliding Machine
Ring Collider Compressed Bunches
With proper tunes choice (.503, .55) the BB compensation with travel focus is not needed
e=16MeV e/e=3.5*10-3 e_Luminosity=10MeV
x=0.8nm y=0.008nm z=2.0m
Lmultiturn=0.84*1036 (Lsingleturn=1.7*1036) with Npart=2*1010
FF with large energy spread tricky
Bunch compressor (z 4mm=>0.6mm) in the ring tricky:
150MeV 1.3GHz
Betx_dyn=3.0mmsigx_dyn=2umsigxp_dyn=0.7mradEmix_dyn=1.5nmEtax_dyn=+-5mmSigx_eta=18umSigz_dyn=0.57mmSige_dyn=5*10-3Bety_dyn=0.5mmSigy_dyn=93.0nmSigyp_dyn=0.2mradEmiy_dyn=0.020nm
Betx=5.0mmsigx=2umsigxp=0.4mradEmix=0.8nmEtax=+-5mmSigx_eta=18umSigz=0.6mmSige=3.5*10-3Bety=0.5mmSigy=63.2nmSigyp=0.11mradEmiy=0.008nm
L0=1.7*10^36, Ldyn=0.84*10^36, sige_lum=11MeV
Npart=2*10^10, fcoll=600MHz
Compressor Compressor
Decompressor Decompressor
IPFF FF
ILC ring with ILC FFILC Compressor, 0.4GeV S-Band or 1GeV L-BandCrossing angle optional
Simplified layout in the Small Disruption Regime Collisions every Turn
Do we need to compress the bunches?
Sz
Sx
Both cases have the same luminosity,(2) has longer bunch and smaller x
1) Standardshort bunches
2) Crossing angle
Overlapping region
Sx
Sz
Overlapping region
Colliding every turn very promising but requires a bunch compressors and a decompressor in the ring
In principle not needed to compress the beams if we collide with a crossing angle such as:
z*xcross=24m (same projected horizontal size)
x/xcross=100m (same effective longitudinal interaction region)
y=12.6nm, y=80um like in the compressed case These parameters gives the same geometric luminosity like the
compressed case:
If z=4mm we need:
x_cross=6mrad, x=0.6um However now beam-beam worsened because the beams
see each other also at non-minimum betay locations
Scaling the parameters for an every-turn colliding machine, with Uncompressed Bunches
With large crossing angle X and Z quanties are swapped Very important!!!
Sz
Sx
z
x
Sx/
Sz*
e-e+
Easy way to decrease the ‘Long Range Beam Beam’ is to increase the crossing angle by a factor 4 at a direct luminosity cost of a factor 4:
xcrossing_angle=2*24mrad x=2.4m This still gives a luminosity of 1.2*1036 with
N=2*1010 if the vertical blowup were none
Unfortunately the blowup is still large and y at the equilibrium is about 6 times larger:
Lmutiturn=0.4*1036
Overall a factor 10 loss in luminosity w.r.t. full compressed bunches (with all the other parameters the same)
Second way to decrease the ‘Long Range Beam Beam’ is to apply the travel focus idea, but now has to be applied in the transverse plane since x and z are swapped.
Vertical waist position in z is a function of x: Zy_waist(x)=x/2Crabbed waist All components of the beam collide at a
minimum y => - the ‘hour glass’ is reduced - the geometric luminosity is higher - the bb effects are reduced
y blowup at the equilibrium less than a factor 2
Vertical waist has to be a function of x:
Z=0 for particles at –x (- x/2 at low current)
Z= x/ for particles at + x (x/2 at low current)
2Sz
2Sx
z
x
2Sx/
2Sz*
e-e+Y
Emittance blowup due to the crossing angle Colliding with no crossing angle and
x=100m, z=100m:
y (single pass)=4*10-4 L=2.1*1027
Colliding with crossing angle=2*25mrad and
x=2.67um, z=4mm (z*=100um, x/=104um):
y =4*10-3 (single pass) L=2.14*1027
Same geometric luminosity but 10 times more emittance blowup Adding the “Crab-waist”, Zy_waist(x)=x/2:
y =1.5*10-3 (single pass) L=2.29*1027
- the ‘hour glass’ is reduced, the geometric luminosity is higher: small effect about 5% more luminosity - the main effect: blowup due the the beam-beam is reduced,
about a factor 2.4 less y (3.8 times the nocross)
Colliding with an angle requires just the ILC DR and the ILC FF.
No compression needed
Crabbed ywaist is achieved by placing a sextupole upstream the IP (and symmetrically downstream) in a place in phase with the IP in X and at /2 in Y (much easier than the longitudinal travel focus).
Only natural energy spread in the beams Angular divergences about 150rad in both planes Crossing angle so large makes the IR (and the FF)
design very easy Low energy spread makes the FF very easy Beam currents around 1.9Amps, possible better trade
off currentdamping time
Collisions with uncompressed beamsCrossing angle = 2*25mradRelative Emittance growth per collision about 1.5*10-3
yout/yin=1.0015
Horizontal Plane Vertical Plane
Y bb_scan
Y bb_scan with 40um horizontal separationY field linear and much smaller kick:1.5rad instead of 150rad
No X separation y-scan
Ring Collider Uncompressed Bunches
Crab focus on in vertical plane Xcrossing_angle=2*25mrad
z=6mm e=6MeV e_Luminosity=9MeV
x=0.8nm y=0.002nm z=8.0m Collision_frequency=600MHz Lmultiturn=1.0*1036 (Lsingleturn=1.2*1036) with Np=2.3*1010
Vertical tune shift like in PEP!!! (similar currents,100 times more luminosity, 100 times smaller betay) Projected Sigmax=Sigmaz*Cross_angle=100m, like in PEP! Possible to further increase the luminosity with: - more current: L=1.8*1036 with Np=4*1010 - further simultaneuos x and y squeeze
- z shortening
Vertical beam size vs crab focusK2=sextupole strength
Luminosity vs crab focusK2=sextupole strength
With k2=8 the vertical emittance blowup is < 20% Luminosity gain about 70%Vertical size rms reduction about a factor 2.5, large tails reduction
Luminosity in excess of 1e36 is achievable
Ohmi simulations
Beam-Beam Tails (D.N. Shatilov)
Without Crab Focus With Crab Focus
Tune shift and luminosity with crossing angle
Very small:
z*=100m
=25mrad
x=9mm
X-Z Coupling smaller then KeK:z*=100m=25mradx=9mm
Kicks that a particle receives while passing through the other beam
Multiturn simulations for uncompressed beams6Km ring, 12000 bunchesColl_freq=600MHz
Bunch_charge=2*1010 e=4mm, Xcrossing=2*25mrad
y0=2pm x0=400pm
IPFF FF
ILC ring & ILC FF
Simplified layout in the Small Disruption Regime Collisions every TurnUncompressed bunchesCrossing angle = 2*25 mradCrabbed Y-Waist
35m long FF
A.Seryi
Conclusions (1):- Rapid progress in the optimization of the
SuperB parameters and layout- First workable parameters set requires: - ILC damping ring, - ILC bunch compressor, - ILC Final Focus - No energy acceleration- Same parameters set but with increased
collision rate and reduced beam current gives more and more luminosity
- The optimal collision rate is every turn.
Solution with ILC DR + ILC FF and no compression seems extremely promising:
- Crossing angle of about 25mrad - Requires virtually no extra R&D - Uses all the work done for ILC - Ring and FF layouts virtually done, 6km (down to 2km should be
possible) circunference rings - 100% Synergy with ILC - IR extremely simplified - Beam stay clear about 20sigmas supposing 1cm radius beam pipe - Beam Currents around 2.0Amps - Background should be better than PEP and KEKB - Possibly to operate at the energy with L=1035
- Total cost less than half of the ILC e+ DRs (2 e+ 6km rings in ILC) - Power around 40MW, further optimization possible - Possible to reuse PEP RF system, power supplies, Vacuum pumps,
etc., further reducing the overall cost - Needs the standard injector system, probably a C-band 7GeV linac like
in KEKB upgrade (around 100ME)
Conclusions (2)
Conclusions (3)
• Possible fall back on the existing factories
• The crabbed waist potentially beneficial also for the current factories
• Possibility to simultaneously boost the performances of the existing machines and do SuperB R&D