Transcript ILC_SULItalk_CN

The International Linear Collider
Christopher Nantista
SLAC
SULI Lecture
July 22, 2008
Outline
•
•
•
•
Introduction
Some Accelerator Basics
Linear Colliders
ILC Anatomy
Introduction
Questions for the Universe
•
•
•
•
•
•
•
•
•
Are there undiscovered principles of nature?
How can we solve the mystery of dark matter?
Are there extra dimensions of space?
Do all the forces become one?
Why are there so many kinds of particles?
What is dark matter? How can we make it in the lab?
What are neutrinos telling us?
How did the universe come to be?
What happened to the antimatter?
Why a Linear Collider?
“The international particle physics community has reached concensus
that a full understanding of the physics of the Terascale will require a
lepton collider in addition to the Large Hadron Collider.”
– Particle Physics Project Prioritization Panel (P5)
Why leptons?
Electron-positron (or muon-antimuon) collisions are much cleaner than protonproton collisions because the former are elementary particles whereas the latter
are composed of quarks which share the energy.
Clearer results and more accurate measurements can thus be gleaned from
lepton annihilations than from hadron collisions.
Why linear?
Having much smaller mass than protons, electrons radiate more of their energy
into synchrotron radiation when bent around a curve.
The diameter of a circular electron accelerator must thus be scaled as the
energy squared and would be prohibitively large at this energy scale.
An Asside
Big e± linear accelerators (linacs) don’t
really do much accelerating.* † ‡
E  mc 2 ,

1
1 v2 / c2
 v / c  1  mc / E   1  0.511 MeV / E 
2
Ek (= E – mc2)
ILC main
linacs go
from here
to here
0
1 keV (103 eV)
10 keV
100 keV
1 MeV (106 eV)
10 MeV
100 MeV
1 GeV (109 eV)
10 GeV
100 GeV
1 TeV (1012 eV)
2
2
v/c _
0
0.0625
0.1950
0.5482
0.9411
0.9988
0.999987
0.9999999
0.999999999
0.99999999999
0.9999999999999
* but they do add energy to the
particles in the beam.
† Proton
and ion linacs do more
accelerating due to much larger
rest masses.
‡ This
is good because constant
speed simplifies accelerator
design.
Some Accelerator
Basics
Microwave Accelerators
Charged particles are (generally) accelerated by high oscillating electric
fields of electromagnetic waves stored or guided in evacuated metal
cavities or structures through which the bunched beam passes.
The electromagnetic frequency used is generally in the range of hundreds
of megaherz to tens of gigaherz (108–1011 cycles/s), generally refered to as
RF (radiofrequency) or microwaves.
MAXWELL’S
EQUATIONS:


B
E  
t

D  

 D 
H 
J
t

B  0

 

D  E, B  H

2



 E
    E     0   H     0 0 2
t
t

 1  2E
2
  E  2 2  0 wave equation in free space
c t
Waveguides
waveguide – a hollow metal tube for transporting power in
confined electromagnetic waves (RF).

 it ikz  ik  z  k t 

wave solution: E( z , t )  Ee
 Ee 
moving waveform
  2f
angular frequency:
free space

2

free space
wavelength
k0  
wavenumber:
c 0
A certain amount of transverse
cutoff
2f c
cutoff
frequency bending/variation of fields is needed
kc 
to meet boundary conditions at walls
wavenumber*:
c
of closed waveguide.
2

guide wavenumber:
What’s left of the free space
k g  k02  kc2 
g
Ey of TM10
mode in
rectangular
waveguide
wavenumber (in quadrature)
guide
goes into longitudinal variation.
wavelength
Ez of TM01
mode in
rectangular
waveguide
 
  
 
*determined by waveguide cross-section and mode, no wave propagation below fc.
Dispersion
Curve

1
kg 
02  c2 
3
2.5
2
tan-1vg
1.5
c
0
c
1
0.5
0
phase velocity:
group velocity:
vp 

tan-1vp
0
c
0.5
1
kg
1.5
k0
kg
d
vg 
 c2 / vp  c
dk g
2
2.5
3
k
speed at which wave crests travel
speed at which power pulse travels
Since charged can’t move faster than c, they can’t keep up with the wave crests and
thus can’t normally experience sustained net energy gain from a waveguide mode.
We need to slow down the confined or guided waves.
This is usually done by means of introducing periodicity.
Accelerator Structures
Disk Loaded Circular Waveguide or Coupled Cavity Chain
beampipe
p
By introducing irises or corrugations to produce a periodic
structure, we can slow down the wave to the speed of light.
Floquet’s Theorem: At a given frequency, in a mode of a periodic structure, the
field at positions seperated by one period differ only by a complex constant.
Space Harmonics: E z (r , z ) 

i t   n z 


A
J

r
e
 n 0 n
n  
 n   / c 2   n2 ,

n  0 
2n
p
0p = phase
advance per cell
speed of light
0
beam bunches synchronous
with this component
2

0
0p 
2 p
Traveling Wave Structure
input and output waveguides
RF pulse travels through, losing power to walls and beam
remainder is discarded in a load.
fill time Tf = L/vg
Standing Wave Structure (Cavity)
input waveguide only
fields build up uniformly, with forward and backward waves
Reflected and discharged power goes back out waveguide to load
p mode is generally used to get peak field in each cavity
Structure Parameters
Q0 
U
Pd
V2
R
Pd
R V2

Q U
a

(unloaded) quality factor, U=stored energy, Pd = wall dissipated power
shunt impedance, V=voltage seen by speed of light beam
a geometrical characterization independent of wall losses
iris radius normalized to RF free-space wavelength,
affects group velocity/cell coupling and wake fields
Traveling Wave
phase advance per cell
p
vg
group velocity
L
dz
Tf  
v ( z)
0 g
  0L ( z )dz
fill time, time for front of RF pulse to move from input to output.
attenuation parameter, wall losses attenuate fields traveling
through structure by e-.
Standing Wave
characterizes external coupling, Pe is power emitted
  Pe / Pd
into waveguide
U Q0
Qe 

external Q
Pe

Q0
U
f
QL 


loaded Q
Pe  Pe 1   f FW HM
Tc 
2QL

cavity time constant
Beam Loading
A linear collider beam consists of a many bunches in a long train for each pulse,
seperated in time by an integer number of RF cycles.
As bunches traverse a structure, they remove energy (beam loading).
To make sure all bunches get the same energy, the structure fields have to be
replenished at the same rate as they are depleted.
Traveling Wave
Standing Wave
Set external coupling and timing such that
rise of input RF induced voltage is canceled
by beam-loading induced voltage.
Shape the input pulse to “pre-load” the
structure. As beam-loading builds up,
the ramp flows out, to be replaced by
flat-top.
50
Gradient (MV/m)
PRF
60
beam
arrives
40
30
20
beam loading
10
Tf
t
no beam
0
Tf
0
0.2
0.4
0.6
0.8
1
Time (ms)
1.2
1.4
1.6
Wake Fields
Bunches of charged particles traversing a cavity/structure, in addition to taking
energy from the fundamental accelerating mode, leave energy behind in other
RF field modes called higher order modes or HOM’s.
These fields give kicks to following bunches, and their buildup and affect must be
controlled by:
•Damping – lets the power from these modes flow out to absorbing loads
through waveguides or couplers which don’t couple to the accelerating
mode.
•Detuning – subtly varying the dimensions of the cells so that HOM
frequencies are different from cell to cell. The bunches then experience
them at various phases, which tends to cancel their cumulative affect.
Klystrons
HEP particle accelerators generally get their RF power from amplifiers called klystrons.
An electron gun, powered by a DC pulse from a high-voltage modulator, produces a
high-current, unbunched beam.
An input cavity driven by a moderate power drive signal imposes periodic
energy/velocity variations along the beam.
Consequently, the beam then bunches as it drifts through the beam tube.
The bunched beam then resonantly excites fields in the output cavity. These fields
decelerate the bunches, sucking power out of the high-voltage beam and sending highpower RF out the output waveguide.
Input
Output
borrowed from Wikipedia
Magnetic Focusing
Without focusing angular divergence (spread of particle directions)
would cause the beam to spread out.
quadrupoles – quadrupole magnets create a transverse magnetic field pattern
that focuses in one dimension and defocuses in the other.
focusing in x
defocusing in y
S
N
focusing in y
defocusing in x
N
S
y
x
N
drift
S
S
N
FODO Array: net effect can be focusing in both x and y.
o
c
u
s
e
f
o
c
u
s
Quads are inserted at intervals along linacs between
structures/cavities, forming the focusing lattice or optics, in
which phase space is traded back and forth between beam
size and divergence.
F
D
F
D
z
Emittance
An important beam parameter, emittance () is the area of the
particle distribution in phase space.
x’ = dx/dz
x’
angular divergence
focus
 x2

y2
 ( x, x ' ) 
x’
z 2 
N  e  2 x2  2 y2  2 z2 

 ( x) 
e
2  x
At upright
points in lattice:
drift
x
area
conserved
x

N e
2 x x '
e
x2
 2
2 x
e
x '2
 2
2 x '
(bi-Gaussian
distribution)
 x ~  x x '
Same for y phase space.
For longitudinal emittance, z = bunch length and E replaces divergence.
Damping rings reduce the emittances to minimum values.
Growth through the rest of the machine must then be carefully controlled.
Radiation Damping
In damping rings, bend magnets and wigglers (periodic magnet arrays that wiggle
the beam) cause the charged particles to emit energy in light known as
synchrotron radiation.
RF driven accelerating cavities restore the lost energy.
The net effect is the gradual damping of the beam emittance as illustrated below.
• Electron (positron) radiates energy and momentum in all dimensions.
• Energy is restored in acceleration by adding longitudinal momentum.
x
x
photon
momentum
bend
x’
x’
z
particle momentum
in x-z plane
z
x = 0, x’ = 0
x constant
E reduced
x
accelerate
x’
z
x = 0, x’ < 0
x reduced
E restored
Luminosity
The other crucial deliverable of a linear collider, along with center-of-mass energy,
is luminosity. It determines the rate at which events with given cross-sections will
occur, and hence the rate of useful data collection by the detector.
number of e+/e-’s per bunch
number of bunches per pulse
disruption enhancement factor
L
N N 

nb f rep H D
4 x y
Gaussian dimensions
of distribution at IP
repetition (pulse) rate
Linear Colliders
Parts of a Linear Collider
• Electron Gun – produces beam electrons
• Injector – pre-accelerates and shapes beam (e.g. collimation, bunch compression)
• Positron production – uses electron beam to produce positrons (undulator, target)
• Damping rings – reduce emmitance of beams
• Main linacs – accelerate up to desired collision energy while preserving emittance
• Final focus – collimate and focus beams for smallest cross-sections at IP
• Interaction Point (IP) – collide beams, surrounded by detector
• Dump – discard spent beams, absorbing enormous energy
Detector: massive, multi-layered high-tech instrument surrounding IP that
senses and tracks particles coming from collisions using various technologies,
identifies interesting events, and stores data for later analysis.
Requiring different expertise outside “accelerator physics”, it is usually treated
as separate from the collider, developed in parallel, and given its own name.
Which is more important? Obviously the linear collider and the detector have
a symbiotic relationship in which either one is useless without the other.
Linear Collider History (A)
SLC (Stanford Linear Collider)
1st and only (so far) linear collider
• began construction in 1983,
operated from 1989-1998.
•Used upgraded SLAC 2-mile linac
• e-’s & e+’s share linac, bent
through separate arcs for collision
• single bunch, NC TW structures,
S-band (2.856 GHz)
• CofM energy ~90-100 GeV
• Polarized source added in 1992
• Allowed detailed studies of Z0
particle (a carrier boson of the weak
force)
Linear Collider History (B)
The Competition (1985?- 2004)
TESLA
(TeV Energy Superconducting Linear Accelerator) – DESY (Germany)based, superconducting SW cavities, L-band (1.3 GHz)
S-Band
– most straightforward extension of 2.856 GHz SLC technology to
larger machine
C-band – KEK alternate approach, innovative 5.712 GHz choke-mode cells.
NLC (Next Linear Collider) – SLAC-based X-band (11.424 GHz), NC TW, promises
higher gradient, required development of RF pulse compression, and wakefield
damping/detuning, Fermilab increasingly involved
JLC (Japan Linear Collider) – KEK-centered X-band design, collaborative R&D with
NLC, later redubbed GLC (Global Linear Collider) for greater pan-Asian
participation.
VLEPP – Russian Ku-band (14 GHz) design.
CLIC (Compact Linear Collider) – CERN (Europe)-based, 30 GHz NC TW, two-beam
approach with higher energy reach.
Linear Collider History (C)
A United Front
Beyond a certain point, it is not sustainable, in terms of funding and manpower, to
continue to pursue multiple designs. The physics community agreed to let an
international group of distinguished, unbiased experts referee a shoot-out between
the leading contenders for linear collider technology:
TESLA L-Band Superconducting
SW Cavities
NLC/GLC X-Band Copper
TW Structures
After visiting the labs to assess R&D status and considering multiple factors:
August 19, 2004: ITRP (International Technology Recommendation Panel)
recommends superconducting technology for a 0.5-1 TeV linear collider:
“…both technologies can achieve the goals presented in the charge. Each had
considerable strengths.”
“…recommending a technology, not a design. ”
ILC (International Linear Collider) program is born.
The accelerator community accepts and rallies behind decision.
SLAC wraps up X-band development, rapidly adjusts and gets on board to
play a leading role in the design of a cold (superconducting) L-band machine.
Why International?: Cost of project would require more resources than one
country could afford.
3 Regions: Americas, Europe, Asia
GDE (Global Design Effort):
International team of >60 experts leading
the effort and steering the coordinated R&D program, headed by Barry
Barish of Cal Tech, with a leader for each of the three regions.
August, 2007:
RDR (Reference Design Report) published, baseline design.
ILC is currently in the
TDP (Technical Design Phase): reduce cost, optimize design, prove technology
ILC Anatomy
Machine Layout
not to scale
injector
(5 GeV)
damping
rings
RTML transport line
undulator
e+ production
e+ injector
main linac (e-)
photocathode
electron gun
detector
IP
final
focus
31 km (19 ¼ miles)
main linac (e+)
Parameters
PARAMETER
NOMINAL VALUE
center-of-mass energy
500 GeV
peak luminosity
21034 cm-2s-1
average beam current in pulse 9.0 mA
pulse rate
5 Hz
beam pulse duration
0.97 ms
charge (particles) per bunch
3.2 nC (21010)
number of bunches per pulse
2,625
bunch spacing
369 ns (480 buckets)
horizontal beam size at IP
640 nm
vertical beam size at IP
5.7 nm
accelerating gradient
31.5 MV/m
RF pulse length
1.6 ms
beam power (per beam)
10.8 MW
total AC power consumption
230 MW
Electron Source
• redundant photocathode guns and laser systems
• normal conducting pre-accelerator followed by superconducting linac to 5 GeV
• polarized electron beam
Positron Source
e-’s wobbled by
magnets radiate
normal conducting
• helical undulator produces polarized photon beam from e- beam @ 150 GeV point
• collimated photon beam hits Ti alloy target wheel (spinning at ~100 m/s to limit
damage), spewing pair-created e-’s and e+’s.
• e-’s and e+’s are magnetically seperated, the former dumped and the latter
captured, accelerated, and injected into the damping ring.
Damping Rings
• each ring is 6.7 km in circumference.
• 6 straight sections: 4 for RF systems & wigglers,
2 for injection & extraction
• ~200 m of superconducting magnet wigglers
• 18 single cell SC 650 MHz CW cavities, total 24 MV.
• injector and extractor fast kickers must deflect one
bunch at a time without disturbing neighboring
bunches, due to >> bunch spacing in the linacs.
• incoming emittances must be greatly reduced
(by 5 orders of magnitude for positron beam y).
Superconducting RF
Certain materials, at temperatures close to absolute zero, enter a superconducting
state in which surface resistivity vanishes, although for RF a slight residual
resistivity remains.
For accelerators, SC cavities provide an efficient way to build up and store
accelerating fields  no RF pulse compression, long beam pulses.
Cryogenics systems (using liquid He) and well insulated cryomodules are required
to maintain cavities at operating temperature.
Accelerating gradient has a hard limit set by the maximum sustainable (in the SC
state) surface magnetic field.
Material purity and surface preparation also affect achievable gradient.
Accelerator Cavities
standing wave
-mode
superconducting
9-cell
‣
higher-order
modes damped
RF power in
Made with solid, pure niobium – it has the highest Critical Temperature (Tc = 9.2 K)
and Thermodynamic Critical Field (Bc ~ 1800 Gauss) of all metals.
‣
Nb sheets are deep-drawn to make cups, which are e-beam welded to form
cavities.
‣
Cavity limited to 9 cells (~1 m long) to reduce trapped modes, input coupler power
and sensitivity to frequency errors.
‣
Iris radius (a) of 35 mm chosen in tradeoff for low surface fields, low rf losses (~ a),
large mode spacing (~ a3 ), small wakes (~ a-3.5 ).
Cavity Parameters
nominal ideal waveforms
35
Gradient (MV/m)
30
25
20
15
10
5
0
0
0.5
1
1.5
Time (ms)
2
2.5
3
300
RF input power
Reflected Power (kW)
250
200
fill
150
beam
discharge
…
100
50
0
0
0.5
Tf
1
1.5
Time (ms)
2
2.5
3
RF Power Distribution
Cryomodules
8 or 9 cavities per cryomodule
SC quads in center of every 3rd one
Klystrons
BASELINE:
10 MW multi-beam klystrons* (MBK’s) with ~65% efficiency
Being developed by three tube companies in collaboration with DESY.
Thales
CPI
Toshiba
*operate at lower voltage yet with a higher efficiency
than simpler single round beam klystrons.
ILC Tunnel Layout
For baseline, developing deep underground (~100 m) layout
with 4-5 m diameter tunnels spaced by 7 m.
Accelerator Tunnel
main linac
cryogenic system
beamlines
penetrations
(every ~12 m)
RF waveguide
signal cables
HV & power cables
Service Tunnel
modulators
klystrons
support systems
RF Unit
1/2 1/3 1/4 1/5 1/6 1/7 1/8 1/9
One 10 MW klystron powers 26 cavities in 3 cryomodules.
Main Linac Layout
Rather than being “laser straight” the main linacs are curved in the vertical plane
slightly more than earth’s curvature to
1. Allow the beam delivery system (final focus) to be in a plane while
2. Keeping cryomodules close to following a gravitational equipotential for
cryogenic fluid distribution
*
*for both main linacs
Conclusion
The ILC is an ambitious project, of which I’ve attempted to paint a general
outline along with some accelerator physics background and history.
Many challenges remain, including:
• improving the cavity fabrication to increase the yield of units that reach
gradient spec.
• producing a robust klystron
• demonstrating the damping ring design concept
• improving expected availability (fraction of time all systems go)
• REDUCING COST
Politically/financially, the ILC has taken a hit recently in the UK and the US,
but the collaboration infrastructure remains in place, and we hope for
increased R&D support. Real momentum may have to await signals from the
LHC that the energy reach of this machine is indeed rich in physics.