Introduction to e e LC IR Issues

1
Introduction to ee- LC IR Issues

• Tom Markiewicz / SLAC
• USPAS 2002 / Long Beach CA
• 13 November 2002

2
Outline

• Fighting the last war SLC/SLD and Accelerator
  Backgrounds
• Quadrupole Synchrotron Radiation (QSR)
• Muon Backgrounds from the Collimator System
• Bunch Structure
• Crossing Angle
• Magnet Technology
• Schemes for extracting beam from IP
• Overall Site Layout Energy Reach
• Beam-Beam Interaction IP Backgrounds
• ee- Pairs, Beamstrahlung Photons, Disrupted Beam
• IR Design and Masking
• Small Spots Control of final quad vibration

3
Luminosity vs. Backgrounds

• The Interaction Point sits in the middle of a
  colossal sensitive detector designed to see every
  single charged neutral particle than emanates
  from an interaction with 100 efficiency.
• In a LC every 8msec there is a new train of 1012
  e-/e that can, if misbehaved, swamp or even
  damage the detector. Circular colliders
  typically wait for 10 minutes so that the beam
  tails can disappear before the detector is turned
  on.
• In the past, attempts to increase LUMINOSITY were
  often limited by an increase in detector
  backgrounds.
• In the current generation of LC designs the
  sensitivity of the final magnetic lenses, that
  live inside the detector, is 1000x greater than
  before.
• The particle density in the current generation of
  designs is large enough to introduce a host of
  potentially troubling beam-beam effects that the
  detector must deal with.

4
SLD VXD,CDC,LAC,WIC
5
SLC
Z/hr
Y-IP
6
SLC Performance
7
Beam-Beam L Enhancement at SLC
8
Given e, sx sy from qx qy
SR Fans from Halo in Final Focus
9
Muon Backgrounds
Linear Ion Chamber Losses in Collimators
Intensity or Location Fluctuations
Muons from Collimators
10
Conclusions from 1995 Talk

•  The AVERAGE behavior of the standard
  backgrounds
• Synchrotron Radiation
• Muons
• Debris from collimators and masks
• has been anticipated and is often acceptable.
  However......
• UNSTABLE conditions OFTEN occur.
• They are due to BEAM TAILS
• Highly Variable,
• Poorly Understood,
• Non-Gaussian  
• from sources which are at any given moment
  unknown as they change in time depending on
  conditions at the source, linac, damping rings,
  and arcs.
• These tails are hard to measure, hard to model,
  and hard to protect against. They are the major
  experimental problem.
• The detector group needs to EXPECT the UNEXPECTED
  and to design the detector, trigger, and data acq
  system to handle it.

11
Typical Laments

• Damping Ring to Linac Kicker Jitter
• SawTooth Instability as Q raised
• Linac Wakefields as Q raised
• Temperature effects
• Phase Drift of reference for linac (Main Drive
  Line)
• BPM Readout
• Delicate phase cancellations
• Planned catastrophes
• Klystron trips
• Structure trips
• Halo Generation
• Dark current, Vacuum, Collimator wakefields, Junk
  from Damping Rings
• Tails Energy, Time (extra buckets), x,y,z
• Linac Girder vibrations
• Less than optimal performance of 104 aging parts
• What about never-before-seen multibunch effects?

Have Tor, Andrei, Brinkmann, etc. convinced you
that all tolerances are under control?
12
Large and Silicon Detectors(same scale)
Silicon Tracker w/ 5 Tesla Solenoid
TPC w/ 3 Tesla Solenoid
Both have 1.0cm radius VXD
13
Synchrotron Radiationfrom Beam Halo in Final
Doublet

• At SLD/SLC SR WAS a (THE) PROBLEM
• SR from triplet WOULD have directly hit beam-pipe
  and VXD
• Conical masks shadowed the beam pipe inner radius
• geometry set so that photons needed a minimum of
  TWO bounces to hit a detector
• Background rates consistent with flat halo
  model
• 0.1 - 1 of the beam filled the phase space
  allowed by the collimator setting.
• At NLC/TESLA
• Allow NO direct SR hits ANYWHERE near IP
• Collimate halo before the linac AND after the
  linac
• VXD MINIMUM RADIUS sets collimation depth as well
  as B_s
• Pain level goes at least as square of VXD min
  radius
• Muon production, collimator wakefields,
  collimator damage, radiation, .

qX450 mrad qY270 mrad
qX32 mrad qY28 mrad
BE SURE YOU NEED IT!
14
HALO Synchrotron Radiation Fans with Nominal 240
mrad x 1000 mrad Collimation
(Similar plots for TESLA)
15
Collimation Designed so all Synchrotron Rad.
passes OUT of IR
107 Halo left after E_Slit
IP
Oct Off
Oct On
100
10
850mm
400mm
E_Slit
Collimation DepthSLC 5sx x 8syNow 19sx x 52sy
Spoilers Absorbers
16
Muon Backgrounds from Halo CollimatorsNo Big
Bend, Latest Collimation Short FF
FF
Energy
18m 9m Magnetized steel spoilers
Betatron
BetatronCleanup
If Halo 10-6, no need to do anything If Halo
10-3 and experiment requires lt1 muon per 1012 e-
add magnetized tunnel filling shielding Reality
probably in between
17
LD Muon Endcap Background e- Scraped to Make
1Muon
Bunch Train 1012
Engineer for 10-3 Halo
Efficiency of Collimator System is 105
Calculated Halo is 10-6
18
New Issues for the Next Generation LC
Bunch Structure ?Crossing Angle, Detector
Effects, Feedback Design, Extraction Beam-beam
effects ?IP Backgrounds, Pinch, Disruption,
Neutrons Small spot sizes ?Control position
motion of final quads and/or the beam
19
Beam Separation Required Before 1st Parasitic
Collision at ctB/2
qC 0 mrad for TESLA
IP
qC gt 4 mrad for NLC
Luminosity Loss vs. Crossing Angle for CLIC,
tB0.67 ns D. Schulte, LCWS 2000
20
NLC Detector MaskingPlan View w/ 20mrad X-angle
Large Det.- 3 T
Silicon Det.- 5 T
30 mrad
32 mrad
21
JLC IR8 mrad Design
Elevation View

• Iron magnet in a SC Compensating magnet
• 8 mrad crossing angle
• Extract beam through coil pocket
• Vibration suppression through support tube

22
TESLA IR
23
Luminosity Monitor DetailNon cylindrically
symmetric geometry for inner detectors
Is this important for Hermaticity requirement
of Detector? Under Study
24
Baseline Magnet Technology Choices

• Extraction outside QD0
• Permanent Magnets (NLC)
• Compact, stiff, few external connections, no
  flowing fluids
• Adjustment more difficult
• Extraction in QD0 Coil Pocket
• Conventional Iron in SC Tube (JLC)
• Adjustable, familiar
• Massive, SC tube shields magnet from solenoid
• Extraction through QD0 Bore
• Superconducting (TESLA)
• Adjustable, large aperture bore
• Massive and not stiff

25
TESLA SC Final Doublet QuadsMature LHC based
Design

• QD0
• L2.7m
• G250 T/m
• Aperture24mm
• QF1
• L1.0m

26
NLC Baseline Permanent Magnet Quadcompact,
stiff, connection free
27
SC Final Doublet Quad Optionsbased on BNL
magnets BUILT for HERA BEPC
Compact 5.7cm Radius Warm Bore Design
Brett Parker, BNL
28
Cold Bore NLC SC Quadrupole w/ Integrated
Sextupole Windings

Quadrupole Coil Layers
Thermal Shield and Cold Mass Support Structure
1 cm
Coil Support Tubes
LHe Flow Space
Sextupole Coil
29
LCD-L2 (3T) with 3.8m L OpticsSeparate
(Easier?) Extraction Line qC20 mrad
52 mrad Cal acceptance
32 mrad M1 acceptance
Calorimeter
SF1
QF1
M2
M1
SD0
QD0
Feedback BPM Kicker
6.3 mrad Lum-Mon acceptance
Low Z shield
Beampipe
Pair LumMon
1 mrad exit aperture
Support Tube
30
NLC Extraction Line150 m long with chicane and
common g and e- dump
X-Angle allows separate beam line to cleanly
bring disrupted beam to dump and allows for
post-IP Diagnostics
E_beam
E_lost
0.2 of beam 4kW lost _at_ 1 TeV 0-0.002 beam
0-20W lost _at_ 500 GeV
31
TESLA Vertical Extraction at 0º
Electrostatic separators at 20m Shielded septum
at 50m (ctB/2) Dipoles to e-/ dump at
z240m Calculated losses OK if beam
ideal Challenging problem No space for diagnostic
equipment
Photons to separate dump at 240m with hole for
incoming beam
32
TESLA Pre-IP Polarimeter and Energy Spectrometer
No TESLA plans for post IP diagnostics NLC plans
pre-IP diagnostics but no work yet begun No
detailed work on post-IP diagnostics done yet
33
Maximum Crossing AngleCrab Cavity

• Transverse RF cavities on each side of IP rotate
  the bunches so they collide head on
• Cavity power req. and relative voltage phase
  stability limit maximum crossing angle
• 2 DL/L when bunch overlap error Dx 0.4
  sx
• Since Dx (qC/2)Dz , at qC20mrad phase error Dz
  corresponds to 10mm
• 0.2 degree of X-Band phase

Dx
qC lt 40 mrad
34
Crossing Angle Considerations Interaction with
Detectors Solenoid

• Beam Steering before IP
• Transverse component of solenoid changes position
  and angle of beams at the IP
• 1.7 mm , 34.4 mrad at 1 TeV, L2m, Bs6 T,
  qC20mrad
• Dispersion and SR cause spot size blow up
• 3.1 mm added to vertical spot size
• Handle with clever upstream beam steering
  gymnastics and by moving QD
• So NOT a problem (unless SR term a (LBsqC)5/2
  grows too large)
• Beam Steering after IP
• Energy dependence of angle of extraction line
• Steering position (410 mm) angle (69 mrad)
  different from B0 case at 1 TeV
• Only run with solenoid ON and Realign extraction
  line when necessary

35
Program Flexibility _at_ NLC TESLA Both have
Minimal Angles to Primary IR ?20mrad to IR2
Bypass Lines 50, 175, 250 GeV
Length for 500 GeV/beam
32 km
Low Energy IR (0.09 1TeV)
High Energy IR (250 GeV to multi-TeV)
Energy Reach of Simultaneous Expts. Z-pole gg e-
e- Fixed target
Injector Systems for 1.5 TeV
36
TESLA Post-Linac Layout
37
Tunnel Constraints

• Enough length for magnets
• Bends per unit length low enough to keep
  SR-caused emittance growth below desired limit
  given maximum forseen beam energy
• IRs separated Ds enough so vibration is not a
  concern
• Ds 0.5 x DR circumference
• Minimize tunnel length for cost

Ds
FF
FF
LINAC
LINAC
BBCOLL
BBCOLL
COLL
COLL
FF
FF
38
NLC 2002 Layout
e
Skew Correction / Emittance Diagnostics
Interaction Region Transport (High Energy)
Interaction Region Transport (Low Energy)
Collimation / Final Focus (High Energy)
Collimation / Final Focus (Low Energy)
IP Separation ? 150 m
IR Hall (High Energy)
IR Hall (Low Energy)
IP2 ?c 30 mrad
IP1 ?c 20 mrad
Collimation / Final Focus (Low Energy)
Collimation / Final Focus (High Energy)
Interaction Region Transport (Low Energy)
Interaction Region Transport (High Energy)
IR2 FF Optimized for 500 GeV cms and Bends
softened to improve L at 1 TeV
Skew Correction / Emittance Diagnostics
e-
39
Luminosity versus Energy

• Over most of E range, Luminosity ? E
• At Highest Energies, Luminosity ? E-2.5
• Synchroton Radiation causes horizontal emittance
  growth
• Final Focus Bends (part of the Chromatic
  Correction System)
• Big Bend required to transport beam to a 2nd IR
• For constant De, length of Big Bend scales as
  E_max
• For fixed geometry, x-emittance scales as E6 so
  Luminosity scales as g-2.5
• For constant dq/ds, De scales as q3

40
Lum to Hi-E IR NLC 2002
LsScale Bends down
L0Geometric
LCalculated for lattice
Y. Nosochkov et al, SLAC Pub 9254
Improvements _at_ Egt1.5 TeV possible by reoptimizing
Final Doublet length
41
e is the short arm

The harder the bend to get Dx, the more e-growth
e- is the long arm
x-emittance increases 20 for 30 mrad qC
42
Approximate LEIR Luminosity Relative to HEIR
NLC 2002
In NLC 2001 example to right, LEIR length/optics
optimized for 500 GeV cms with an 80 mrad bend
0.92
43
Engineering Constraints

• Magnet vacuum apertures set by lowest beam
  energy
• Since ex and ey scale as 1/E, you can only go
  down in E with b constant until beam spot fills
  the beam pipe, then you must scale the beta
  functions by 1/E. This keeps divergence
  constant, but Lum falls faster than before
• Hard to meet Power Supply
  sensitivity requirements
  (10-5) without limiting range
• X4 in Energy range
  has been rule of thumb

44
BackgroundsWell Studied by ALL GROUPS Not a
Problem

• IP Backgrounds
• Beam-Beam Interaction
• Disrupted primary beam
• Extraction Line Losses
• Beamstrahlung photons
• e,e- pairs from beams. gg interactions
• Hadrons from beams. gg interactions
• Radiative Bhabhas

Good, scale with luminosity 1) Transport them
away from IP 2) Shield sensitive detectors 3)
Detector Timing

• Machine Backgrounds
• Synchrotron Radiation
• Muons Production at collimators
• Direct Beam Loss
• Beam-Gas
• Collimator edge scattering
• Neutron back-shine from Dump

Bad, get nothing in exchange 1) Dont make
them 2) Keep them from IP if you do
45
Backgrounds and IR Layouts

• Most important background is the incoherent
  production of ee- pairs.
• pairs scales with luminosity and is equal for
  both designs.
• Detector occupancies depend on machine bunch
  structure and relevant readout time
• GEANT and FLUKA based simulations indicated that
  in both cases occupancies are acceptable and the
  CCD-based vertex detector lifetime is some number
  of years.
• IR Designs are similar in the use of tungsten
  shielding, instrumented masks, and low Z material
  to absorb low energy charged and neutral
  secondary backgrounds

46
Beam-Beam InteractionSR photons from individual
particles in one bunch when in the E field of the
opposing bunch

• Beams attracted to each other reduce effective
  spot size and increase luminosity
• HD 1.4-2.1
• Pinch makes beamstrahlung photons
• 1.2-1.6 g/e- with E3-5 E_beam
• Photons themselves go straight to dump
• Not a background problem, but angular dist. (1
  mrad) limits extraction line length
• Particles that lose a photon are off-energy

47
Luminosity Loss vs. Position Angle Jitter

• Beam-Beam Luminosity Enhancement (Hd)
• Lost quickly for small jitter
• Beam attraction helps for large Dy

NLC Hd1.4, Dy 14 sensitivity 0.3s 1nm
TESLA Hd2.1, Dy 25 sensitivity 0.1s 0.5nm
48
Pair ProductionPhotons interact with opposing
e,g to produce e,e- pairs and hadrons
500 GeV designs

• Pair PT
• SMALL Pt from individual pair creation process
• LARGE Pt from collective field of opposing bunch
• limited by finite size of the bunch

49
e,e- pairs from beams. g interactions
pairs scales w/ Luminosity 1-2x109/sec
BSOL, L, Masks
50
Pairs as a Fast Luminosity Monitor
TESLA
Also, Pair angular distribution carries
information of beam transverse aspect ratio
(Tauchi/KEK)
51
Pair Stay-Clear from Guinea-Pig Generator and
Geant
52
e,g,n secondaries made when pairs hit high Z
surface of LUM or Q1
High momentum pairs mostly in exit beampipe
Low momentum pairs trapped by detector solenoid
field
53
Design IR to Control ee- Pairs

• Direct Hits
• Increase detector solenoid field
• Increase minimum beam pipe radius at VXD
• Move beampipe away from pairs ASAP
• Secondaries (e,e-, g,n)
• Point of first contact as far from IP/VXD as
  possible
• Increase L if possible
• Largest exit aperture possible to accept
  off-energy particles
• Keep extraneous instrumentation out of pair
  region
• Masks
• Instrumented conical M1 protrudes at least 60cm
  from face of PAIR-LumMon
• Longer more protection but eats into EndCap CAL
  acceptance
• M1,M2 at least 8-10cm thick to protect against
  backscattered photons leaking into CAL
• Low Z (Graphite, Be) 10-50cm wide disks covering
  area where pairs hit the low angle W/Si Pair
  Luminosity monitor

54
TESLA IRNLC/JLC/CLIC Similar
55
Detector Occupancies are Acceptable fcn(bunch
structure, integration time)
LCDL2 Hit Density/Train in VXD TPC vs. Radius
TESLA VXD Hits/BX vs. Radius
56
Photons
Beampipe _at_ r2.2cm, B4T
Beampipe _at_ r1.0cm, B3T
TESLA g/BX in TPC vs. z
Extraction Line (6m) shines thru shield here,
but not here
57
Detector Occupancyfrom ee- Pairs _at_ 500 GeV
bunches integrated in given detectors ReadOut
Time
TESLA
NLC
58
ee-? ee- gg ? ee- Hadrons

• NLC Analysis began Spring 2001
• CAIN simulation plus JETSET
• Need to integrate 190 bunches
• Doesnt appear to be a problem but one detector
  element with good time resolution will help if it
  is

For NLC, needs Study!! Mostly go forwardNo
problem for 1 bunch X
gg ? ee- Hadrons Energy Distributions
Endcap
Mask
Barrel
59
Neutron BackgroundsThe closer to the IP a
particle is lost, the worse

• Off-energy e/e- pairs hit the Pair-LumMon,
  beam-pipe and Ext.-line magnets
• Radiative Bhabhas Lost beam ltx10
• Solutions
• Move L away from IP
• Open extraction line aperture
• Low Z (Carbon, etc.) absorber where space permits

• Neutrons from Beam Dump(s)
• Solutions Geometry Shielding
• Shield dump, move it as far away as possible, and
  use smallest window
• Constrained by angular distribution of
  beamstrahlung photons
• Minimize extraction line aperture
• Keep sensitive stuff beyond limiting aperture
• If VXD Rmin down x2 Fluence UP x40

60
Energy Distributions
Tesla 500 GeV
NLC-1 TeV
61
Neutrons from Lost Pairs and Rad. Bhabhas
Neutrons which reach the IP are produced close to
the IP, mainly in the luminosity monitor
62
Neutrons from the Beam Dump
Neutrons per Year
Limiting Aperture
Integral
Geometric fall off of neutron flux passing 1 mrad
aperture
1.0
0.5
Radius (cm)
z(m)
63
Neutron BackgroundsSummary
Neutron hit density in VXD NLC-LD-500 GeV
NLC-SD-500 GeV Tesla-500 GeV Beam-Beam
pairs 1.8 x 109 hits/cm2/yr 0.5 x
109 hits/cm2/yr O(109 hits/cm2/yr) Radiativ
e Bhabhas 1.5 x 107 hits/cm2/yr no hits
lt0.5x108 hits/cm2/yr Beam loss in
extraction line 0.1 x 108 hits/cm2/year 0.1
x 108 hits/cm2/year Backshine from dump 1.0
x 108 hits/cm2/yr 1.0 x 108 hits/cm2/yr
negligible TOTAL 1.9 x 109
hits/cm2/yr 0.6 x 109 hits/cm2/yr
Figure of merit is 3 x 109 for CCD VXD
64
Colliding Small Beam Spots at the IP
Q1
Q1
Relative Motion of two final lenses
e
e-
sy 3 nm Dy sy/4 1 nm

• Control position motion of final quads and/or
  position of the beam to achieve/maintain
  collisions
• Get a seismically quiet site
• Dont screw it up Pumps, compressors, fluids
• Good magnet and detector engineering Light,
  stiff Q1 in a rigid detector
• Tie to bedrock get lenses outside detector as
  soon as possible

65
Luminosity Stabilization

• Performance of ALL LCs based on feedback systems
  such as that developed at SLC
• SLOW feedback based on machine rep rate f and
  can handle motion of frequencies up to f /20 to
  f /60
• 0.1-1 Hz at TESLA where f 5 hz
• 2-5 Hz at NLC where f 120 Hz
• TESLAs long (2820) train of widely (337ns)
  spaced bunches allows the extension of the
  technique to frequencies up to 100 kHz and
  should handle all correlated noise sources with
  minimal luminosity loss and little impact to the
  detector
• NLC relies on a variety of techniques to
  stabilize the collisions against jitter above the
  2-5 Hz range

66
In 90 bunches and DL lt 10, bunches are
controlled to 0.1sy Intra-train Feedback at TESLA
67
Sensor Driven Active Vibration Suppression at NLC
Inertial Capacitive Sensors
QD
Carbon fiber stiffener
Piezo mover
Interferometric Sensors Optical anchor
FFTB style cam movers
Cantilevered support tube
68
First Inertial Sensor Tests
Block with Accelerometers/ Geophones
Electrostatic Pushers
x10-100 Jitter Suppression in Frequency Range of
Interest
69
Current NLC Vibration Stabilization Program
3m magnet and 5.5m support have same
mechanical properties (mass, w) as final design
70
Optical Anchor Interferometer to Stabilize
Quads w.r.to Tunnel
Sub-nm resolution measuring fringes with
photodiodes ? drive piezos in closed loop
71
UBC RD on Interferometers
72
UBC RD on Interferometers

• Test platform challenges
• clean system
• high natural frequency
• Resonances
• Bandwidth
• Algorithms
• Piezo stiffness

From 90 nm to 5 nm at 5 Hz
Stabilize mirror 10 Meters in Air RMS 25.4 nm
before, 0.27 nm after
73
Very Fast Intra-train IP Feedback
Latency 37 ns
74
NLCTA FONT InstallationFeedback On Nanosecond
Timescales
Magnet assembly and X-band BPM installed onto
NLCTA downstream of RF structures.
Beam direction
Feedback loop
BPM
Dipole/Kicker
75
FONT Results 27 Sept 2002x10 Improvement in
Position
BPM for 5 Dipole Settings
After 1 Latency Period Position corrected
independent of gain, then loop turns off
As above but loop ON Beam stays corrected until
end of pulse
76
LINX Engineering Test Facility Address x104
Luminosity Issue!!

• ee- collisions at SLC with 50nm beams
• Test stabilization techniques proposed for future
  linear colliders and demonstrate nanometer
  stability of colliding beams
• Investigate new optical techniques for control of
  beam backgrounds
• Provide a facility where ultra-small and
  ultra-short beams can be used for a variety of
  other experiments

NLC 250 GeV 300 / 2 360 / 3.5 same 0.11 mm
245/2.7nm 7.5E9

• Beam Energy 30 GeV
• DR emittances g?x,y1100/50E-8 m-rad
• FF emittances g?x,y1600/160E-8 m-rad
• IP Betas ?x8mm ?y0.1 mm
• Bunch length ? z0.1 - 1.0
  mm
• IP spot sizes ?x,y1500/55
  nm
• Beam currents N? 6E9

77
Nanometer Stability of Colliding Beams
Beam-Beam Deflection gives 1nm stability
resolution
BPMres
400nm

• Colliding beams provide a Direct
  Model-Independent Test of any engineering
  solution to the final doublet stability problem
• Not possible in FFTB

Project on-hold awaiting funds
Proposal written
78
Cavity BPMs with lt1nm Resolution
1995 FFTB Tests 25 nm resolution _at_ 6 x 109 /
bunch
Resolution Theoretically sub-nmPractical
Limits -Common Mode Leakage -Angle
Alignment -Losses, Drifts Phase matching in
Cables, Filters, Mixers -Digitizers and Pulse
Shapers not optimized
79
Post 1995 Developments May Preserve Intrinsic
Resolution

• Remark made at Nanobeam 2002 WorkshopIf we
  had nm BPMs we wouldnt need the beam-beam
  collision to verify nm-level vibration stability
  (LINX)
• Reply No problem in principle

4 C-Band ATF BPMs

• Improvements
• Digital Down Conversion Fast Scope (Drifts,
  resolution dynamic range)
• New BPM design w/couplers that remove Common Mode
• Angle Adjustments Hardware
• Improved S/N by using rad hard electronics on
  beamline

80
Single Beam Vibration Stability Test
Spot Size
Magnet
BPM
BPM
Artists conception
81
Conclusion

• Linear Collider IR design issues are common to
  all proposed machines.
• The proposed designs look more similar than
  different
• All projects have been actively collaborating to
  resolve issues through meetings constant
  communication, personnel exchange
• IR Design is well advanced and not a reason to
  delay consideration of a linear collider
• Lets choose a machine technology
• And get on with it!