Transcript Fisher, ARD status meeting 2011-01

LHC
From LCLS to LHC:
Light from Electrons, Protons,
and even Lead Ions
Alan Fisher
ARD Status Meeting
2011 January 6
Terahertz Radiation from LCLS Electrons

Electrons passing through a metal foil emit intense, coherent
transition radiation (CTR) at wavelengths l  bunch length.





Highly compressed LCLS bunches: l  10 to 100 µm
Corresponding frequencies: 3 to 30 THz
Intense pulses with the time structure of the electron beam
Powerful diagnostic tool for the beam’s temporal profile
When focused:





Electric fields > 1 GV/m = 0.1 V/Å (depends on bunch charge, length)
Magnetic fields > 3 T
Well above other THz sources
Duration of tens of fs
Powerful enough to pump femtosecond chemistry and nonlinear
behavior in materials
Terahertz Collaborators

LCLS









Henrik Loos
Stefan Moeller
Jim Turner
Gene Kraft
Dave Rich
Rob McKinney
Roenna del Rosario
Frank Hoeflich
John Wagner

PULSE





FACET


Aaron Lindenberg
Dan Daranciang
John Goodfellow
Shambhu Ghimire
Ziran Wu
University of Maryland


Ralph Fiorito
Anatoly Shkvarunets
Calculated Field and Spectrum
Electric Field
Spectrum
At the THz focus, for a 1-nC, 20-fs bunch
Calculations by Henrik Loos
Terahertz Layout

Extract THz in the Undulator Hall




Pneumatic actuator, 30 m past the end of the undulator, inserts a thin
beryllium foil at 45° to the electron beam
Electrons and hard x-rays pass through
THz light goes downward through a diamond window to an optical
table below
Measurements




Joulemeter: Energy
Pyroelectric video camera: Size at focus
Michelson interferometer: Spectrum
Will soon install a 20-fs Ti:sapphire laser on the table



Electro-optic measurements of electron bunch
THz/laser pump/probe studies of materials
Proof of principle for a future THz/x-ray pump/probe upgrade

Requires a long THz transport line to the NEH
Beryllium Foil



Thickness: 2 µm
Diameter: 25 mm
Electrons go through with
little scattering or radiation
Transparent to hard x rays


Can be used parasitically
Absorbs x rays below 2 keV

For soft x-ray experiments,
foil must be pulled out
2-µm Be Foil at 45°
1.0
0.9
0.8
X-Ray Transmission

0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
0
1000
2000
3000
Photon Energy [eV]
4000
Beamline and Optical Table
Beryllium
foil
Pneumatic
actuator
Diamond
window
Track for
laser curtain
e−
Laser
chiller
Control rack
Laser Enclosure with Curtain
Foil in 6-way Cross
Pneumatic
actuator
Beryllium
foil
Bellows
Diamond
window
Optics for First Measurements
Initial Characterization of the Terahertz Radiation:
Energy and Profile at Focus
2010 October 12
Filter wheel Fluorescent card
Pyrocam
Off-axis
parabolic
mirror
Translation stage:
Move through focus
HeNe
Joulemeter
Si
Fluorescent card
Fluorescent card (flip up)
(Enclosure 1)
Elevation View
Plan View
Complete Optics Layout
Optics
enclosure
The optics will be enclosed for laser
safety and for a dry-air purge.
(Water has THz absorption lines.)
THz Energy versus Charge and Bunch Length



Electric field E of a relativistic bunch varies with bunch charge
q and duration t : E ~ q/t
Energy in the pulse then follows: E2t ~ q2/t
Compare a q2/t fit (open circles) to the measured THz (solid)
for two bunch-charge values
Transmission of GaAs versus THz Intensity


Translate a GaAs wafer through the THz focus (at z = 0)
Transmitted THz energy shows nonlinear absorption
Michelson Interferometer: THz Spectrum



Scan of Michelson delay gives an autocorrelation
Fourier transform of autocorrelation yields the power spectrum
Still being commissioned, but we have preliminary data
Pyroelectric
detector
Delay scan
Si beamsplitter
I(t)
Delay stage
Fourier transform
I(ω)
Interferometer Scans
e− bunch
characteristics
350 pC, 50 fs
350 pC, 115 fs
I(t)
I(ω)
200 nm steps, 10 shots/point, adjacent averaging
Power spectrum
Insight from Simulations
Measured I(ω)
Simulated electron form
factor
Simulated THz pulse
350 pC, 50 fs
1 nC, 20 fs
350 pC, 115 fs
1 nC, 60 fs
Summary of Spectral Findings





Short bunches show a broad spectrum peaking at 10 THz,
consistent with a single-cycle pulse.
Long bunches have nodes in the spectrum, corresponding to
ripples in the time domain from imperfect compression.
The width of the autocorrelation trace is consistently shorter
than the value from the electron bunch-length monitor
Better electronics are on order to greatly speed up the scans.
A laser synchronized with the beam will be installed in
February, allowing THz pump/optical probe experiments and
direct measurements of time-dependent E-field.
Transport to the NEH


A 100-m transport line is needed to relay the THz light from
the source to an NEH hutch.
Sequence of off-axis parabolic (OAP) mirrors, to alternate
between collimation and focusing to a waist




Diffraction demands large mirrors and frequent refocusing
For example, 400-mm-diameter mirrors with a focal length of 5 m,
spaced every 10 m
90° bend at each OAP lengthens the path
THz arrives tens of nanoseconds after the x rays

We want THz/x-ray pump/probe, but this gives probe/pump!
Two Electron Bunches


To get the THz radiation to a user before the x rays arrive,
we’ll need two electron bunches, separated by tens of ns
First bunch has a high charge, optimized for THz

That would lase poorly in the FEL, but we can also spoil the gain:



Choose an RF bucket for first pulse so that it arrives ps before x rays



THz “optical trombone” delay line then tweaks the arrival time
Second bunch makes x rays
A second THz pulse arrives tens of ns after pump/probe



Turn up the laser heater: Excessive energy spread
Or add a fast kicker: Orbit oscillation along the undulator
No problem for users because it is so late
If necessary, a fast kicker could force electrons to miss the foil
Test in July: 2 equal bunches, 8.4 ns apart

Both bunches were lasing in the FEL
2010-07-28: Two-Bunch Test in LCLS

2 bunches, 8.4 ns apart

Photodiode viewing 2 x-ray
pulses on YAG screen

10 ns/div

Both pulses, adjusted to have
slightly different energies, on
the SXR spectrometer

Upgrade of instrumentation
(BPMs, toroids) required to
measure individual bunches
Synchrotron-Light Monitors for the LHC

Five applications:


BSRT: Imaging telescope, for transverse beam profiles
BSRA: Abort-gap monitor, to verify that the gap is empty





Abort-gap cleaning
Longitudinal density monitor (in development)
Halo monitor (future upgrade)
Two particle types:



When the kicker fires, particles in the gap get a partial kick and might
cause a quench.
Protons
Lead ions (1 month a year): First ion run was in November/December
Three light sources:




Undulator radiation at injection (0.45 to 1.2 TeV protons)
Dipole edge radiation at intermediate energy (1.2 to 3 TeV)
Central dipole radiation at collision energy (3 to 7 TeV)
Consequently, the spectrum and focus change during the energy ramp
CERN Collaborators






Stéphane Bart-Pedersen
Andrea Boccardi
Enrico Bravin
Stéphane Burger
Gérard Burtin
Ana Guerrero






Wolfgang Hofle
Adam Jeff
Thibaut Lefevre
Malika Meddahi
Aurélie Rabiller
Federico Roncarolo
My work at CERN has been supported by SLAC through
the DOE’s LHC Accelerator Research Program (LARP).
Power Radiated by a Charge in a Dipole



A relativistic charge q = Ze with mass m, velocity bc  c, and
energy E = eEeV = gmc2 travels through a dipole field Bd
gb mc EeV
r

Radius of curvature of the orbit:
qBd
ZcBd
Lead ions: Since r and Bd don’t change, the maximum energy
must be Z = 82 times higher—574 TeV, or 100 J per ion!



Equal to the kinetic energy of 82 mosquitos flying at 1 m/s
Or, a 1-mm grain of sand thrown at 40 km/h
2
2 2 e2
c
g4 2
Power emitted in synchrotron radiation: Ps  Z
3 4 0 c r


The factor of g 4 makes the radiated power substantial—except in the
LHC, where the protons have a g of 7500, but r = 6 km.
The factor of Z2g 4 means that ions radiate more than protons by a
factor of 826/2084 = 162
Synchrotron Power in Various Rings
Dipoles
Ring
Ring
Particle Beam Normalized Beam
CircumType Energy Energy
Current
ference
E
g
E
mc 2
[GeV]
Synchrotron Radiation
Radius Critical Power
Per
Average
of
Wavein
Particle
Power
Curvature length Dipoles
Per Turn
I
C
r
lc
I
Ps
ec
[A]
[m]
[m]
[nm]
[W/m]
2r
I
Ps
ec
2r
Ps
c
[W]
[eV]
460,000
912,000
SPEAR-3
e−
3.0
5,870
0.5
234
7.86
0.163
9,230
PEP-2 HER
e−
9.0
17,610
2.0
2,200
165
0.127
6,790 7,040,000 3,520,000
PEP-2 LER
e+
3.15
6,160
2.5
2,200
13.75
0.246
LHC
p
450
480
0.582
26,659
6,013
LHC
p
7,000
7,460
0.582
26,659
6,013
60.7
2,960 0.0061
26,659
6,013
968
LHC
208Pb
82+
574,000
18,300 1,580,000
228,000 8.2E-07
633,000
0.0309
0.0531
0.048
1,810
3,110
0.084
3,170
520,000
Spectrum and Critical Frequency
Spectrum near Critical Frequency
Critical Wavelength vs. LHC Beam Energy
Lead ions
Protons
Camera response


Normalized dipole emission, intergrated
over vertical angle y, versus energy E/Ec
Long tail for E < Ec



Visible light << Ec for electron rings
Rapid drop in emission for E > 10Ec
Peak is below Ec



Cameras respond from near IR to near UV
Proton emission wavelengths are too long to
see below ~1.2 TeV
Ion emission is too long below ~3 TeV

Ion energy given as “equivalent proton energy”:
Dipole set to the same field as for 3-TeV protons
Superconducting Dipole + Undulator


Add an undulator inside the
dipole’s cryostat
Dipole:



Undulator layout and field map


3.88 T
9.45 m
6013 m
1.57 mrad
Undulator:





Undulator inside cryostat
Field for 7 TeV
Length
Radius of bend
Bend angle
Peak field
5T
Periods
2
Period length
280 mm
Pole gap
60 mm
Wavelength for
609 nm
450-GeVprotons (injection)
Undulator Emission versus Beam Energy
Camera response
Lead ions
Protons


Undulator peak is red for injected protons, but moves into the
ultraviolet at 1 TeV. Dipole light is still in the infrared.
Injected ions can be seen only with the weak high-energy tail.

Radiation from a 2-period undulator has a broad bandwidth.
Short Dipoles and Edge Radiation



Radiation from the dipole’s edge fills the gap during the
energy ramp between the undulator and the central dipole
light.
If the dipole is short, the observer sees a faster “blip” of
radiation, which pushes the spectrum to higher frequencies.
Rapidly rising edge field of a (long) dipole has the same effect.
2010-01-18
Fisher — Imaging with Synchrotron
Light
28
Photoelectrons per Particle at the Camera
Protons
Combined
Lead Ions
Dipole center
Dipole center
Combined
Dipole edge
Undulator

Undulator
Dipole edge
In the crossover region between undulator and dipole radiation:


Weak signal
Two comparable sources: poor focus over a narrow energy range

Focus changes with energy: from undulator, to dipole edge, to dipole center

Dipole edge radiation is distinct from central radiation only for w >> wc
Particles per Bunch and per Fill
Protons
Lead Ions
Particles in a “pilot” bunch
5109
7107
Particles in a nominal bunch
1.151011
7107
Bunches in early fills
1 to 43
1 to 62
Bunches in a full ring
2808
592
Layout: Dipole and Undulator
Cryostat
70 m
194 mm
To arc
To RF
cavities
and IP4
1.6 mrad
420 mm
D4
10 m
D3
Extracted light sent
to an optical table
below the beamline
U
560 mm
26 m
937 mm
RF Cavities
BSRT for Beam 1
Door to
RF cavities
Beam 1
Undulator
and dipole
Beam 2
Optical Table
B1 Extraction mirror
(covered to hunt for a light leak)
Table Enclosure under Extraction Mirror
Beam 1
B1 Extraction mirror
Beam 2
Extraction Mirror
Protons—with their small heat load—can use a simple mirror without cooling.
Optical Table
Layout of the Optics
Extraction
mirror
Beam
Optical Table
Alignment
laser
Calibration light
and target
Shielding
PMT and 15% splitter for abort gap monitor
F1 = 4 m
F2 = 0.75 m
Intermediate
image
Cameras
Slit
Focus
trombone
Table Coordinates [mm]
LHC Beams at Injection (450 GeV)
Beam 1
Beam 2
Horizontal
1.3 mm
1.2 mm
Vertical
0.9 mm
1.7 mm
Light from undulator.
No filters.
Beam 1 at 1.18 TeV
Vertical Emittance
Synchrotron Light
Wire Scanner
Proton Energy

1.18 TeV has the weakest emission in the camera’s band.



Nevertheless, there is enough light for an adequate image.


Undulator’s peak has moved from red to the ultraviolet
Dipole’s critical energy is still in the infrared
Some blurring from two comparable sources at different distances
Vertical emittance growth before and after ramp

Comparing synchrotron light to wire scanner
LHC Beams at 3.5 TeV
Beam 1
Beam 2
Horizontal
0.68 mm
0.70 mm
Vertical
0.56 mm
1.05 mm
Light from D3 dipole.
Blue filter.
First Lead-Ion Images, 2.3 TeV
First ramp of one bunch of lead ions
2011 November 5
Calibration Techniques

Target



Wire scanners



Incoherently illuminated target (and
alignment laser) on the optical table
Folded calibration path on table
matches optical path of entering light
Compare with size from synchrotron
light, after adjusting for different bx,y
Only possible with a small ring current
Beam bump

Compare bump of image centroid with
shift seen by BPMs
5 mm
Emittance Comparisons at 450 GeV
Beam 1 Horizontal
Beam 1 Vertical
LHC Synchrotron Light
LHC Wire Scanner
From SPS
Beam 2 Horizontal
Time [h]
Nominal 
Beam 2 Vertical
Time [h]
Disagreement with Wire Scanners

The horizontal size—but not the vertical—measured with
synchrotron light is larger than the size from the wire scanners.


Beam 1: Factor of 2 in x emittance (2 in beam size)
Beam 2: Factor of 1.3 in x emittance

b beat isn’t large enough to explain this.

I tested the full system in the lab in 2009


Found distortions on the first focusing mirrors (old, perhaps left-overs
from LEP?)
Replacements arrived just before tunnel was locked: No time for tests
Nov 2010: Bench Tests with Duplicate Optics


I set up a new 4.8-m  0.8-m
table in the lab with a copy
of the tunnel optics.
Alignment of tunnel optics:

Images shift while focusing:
mirrors not properly filled




More diffractive blurring?
New alignment procedure
Entering light needs one more
motorized mirror
Camera and digitizer:

Fixed hexagonal pattern from
intensified camera


500-µm
line width
Increased magnification
reduces the effect
Digitizer grabs every other line
400-µm
line width
BSRA: Abort-Gap Monitor

Gated photomultiplier receives ~15% of collected light

PMT is gated off except during the 3-s abort gap:




Beamsplitter is before all slits or filters, to get maximum light
Gap signal is digitized in 30 100-ns bins


High gain needed during gap
Avoid saturation when full buckets pass by
Summed over 100 ms and 1 s
Requirement: Every 100 ms, detect whether any bin has a
population over 10% of the quench threshold

Integration over 1 s is needed where PMT signal is weak



Protons near 1.2 TeV
Worst case signal-to-noise is 10 for 1-s integration with a population of
10% of quench threshold
No PMT signal observed for an ion bunch at injection energy


Calculations said there should be a signal
Does PMT not extend into the near IR as far as the datasheet claims?
Protons/100ns at the Quench Threshold
Protons in a pilot bunch
Original specification
Model for BSRA (Q4 quench)
(M. Sapinski)
General quench model
(B. Dehning)

Original thresholds, specified only for 0.45 and 7 TeV, were too generous



Must detect levels well below a pilot bunch
BLM group provided improved models: using Sapinski’s calculation
Ion threshold is scaled from proton threshold:


Ion fragments on beam screen
Deposits same energy as Z protons at same point in ramp
Calibration of BSRA






Inject a pilot bunch
Charge measured by bunchcharge and DC-current electronics
Attenuate light by a factor of
 bunch charge / quench threshold
Move BSRA gate to include the
pilot bunch
Find PMT counts per proton
(adjusted for attenuation) as a
function of PMT voltage and
beam energy
Turn RF off (coast) for 5 minutes
to observe a small, nearly uniform
fill of the gap

Last bunch in fill
First bunch in fill
Pilot bunch
Abort gap
After coasting briefly,
bunch spreads out
Useful to test gap cleaning…
Time [100-ns bins]
2009 Dec 16: Test of Abort-Gap Cleaning

Injected 4 bunches into Beam 2




Poor lifetime, but not important for this experiment
Turned off RF, and coasted for 5 minutes
Abort-gap monitor detected charge drifting into the abort gap
Excited 1 µs of the 3-µs gap at a transverse tune for 5 minutes

How well did this work? Look inside the gap...
RF off
RF on
(poor lifetime)
Beam charge
(injection)
0
Gap cleaning
Beam dumped
Total PMT signal
(negative going)
in all 30 bins
25 minutes
Charge in Abort Gap
Abort gap (3 µs)
Beam dumped
Time (s)
Excitation had ringing
on the trailing edge
(improved in January)
Cleaning started in 1-µs
region: Immediate effect
Charge drifting from
first bunch after gap
RF off: coasting beam
Position in fill pattern (100-ns bins)
2010 Sep 30: Test with Improved RF Pulse

Injected 2 bunches (1 and 1201), 3 µs apart; no ramp



Another test at 3.5 TeV took place on 2010 Oct 22
Cleaning pulse applied between the bunches at the x or y tune
RF voltage reduced in steps; debunched protons drift into gap
Cleaning resumed.
Time (s)
Cleaning off
Voltage lowered by another step.
Cleaning turned off.
Protons repopulate cleaning region.
RF voltage lowered.
Low/high momentum protons drift
into gap from bunches on left/right.
Encounter cleaning region.
3-µs Gap (in 30 100-ns bins)
Darker = More protons
Longitudinal-Density Monitor

Monitor is being developed by Adam Jeff


Photon counting using an avalanche photodiode (APD)
Fiber collects 1% of the BSRT’s synchrotron light





Sufficient signal: 103 photons/bunch at 1 TeV, and 3106 for E ≥ 3.5 TeV
Measure time from ring-turn clock to photodiode pulse
Accumulate counts in 50-ps bins
One unit has been installed on Beam 2 for testing
Modes:

Fast mode: 1-ms accumulation, for bunch length, shape, and density


Slow mode: 10-s accumulation, for tails and ghost bunches down to
5105 protons (410-6 of a nominal full bunch)


Requires corrections for APD deadtime and for photon pile-up
Only 1 photon every 200 turns
May require 2 APDs: APD for slow mode gated off during full bunches
Testing the Longitudinal-Density Monitor
1 turn
1 train
28 ns
One bunch, but
protons are also in
neighboring buckets
Adam Jeff
Observing the Solar Corona

Lyot invented a coronagraph in the 1930s to image the corona



Huge dynamic range: Sun is 106 times brighter than its corona
Block light from solar disc with a circular mask B on image plane
Diffraction from edge of first lens (A, limiting aperture) exceeds corona


Circumferential stop D around of image of lens A formed by lens C
Can we apply this to measuring the halo of a particle beam?
Bernard Lyot, Monthly Notices of the Royal Astronomical Society, 99 (1939) 580
Beam-Halo Monitor

Halo monitoring is part of the original specification for the
synchrotron-light monitor


But the coronagraph needs some changes:



SLAC’s involvement, through LARP, in both light monitors and
collimation makes this a natural extension to the SLM project
The Sun has a constant diameter and a sharp edge
The beam has a varying diameter and a Gaussian profile
An adjustable mask is needed
Fixed Mask with Adjustable Optics
Halo
image
Source
Steering
mirrors
Zoom lens
Masking mirror

SLM images a broad bandwidth: Near IR to near UV


Reflective zoom is difficult compared to a zoom lens
Bandwidth is a problem for refractive optics


Limited by need for radiation-hard materials
But a blue filter is used for higher currents: Fused silica lenses could work
Halo Monitor with Masking Mirror
Alignment
laser
Calibration light
and target
Masking mirror
PMT and 15% splitter for abort gap monitor
F1 = 4 m
F2 = 0.75 m
Cameras
Intermediate image
Diffraction stop
Slit
Zoom
lens
Focus
trombone
Table Coordinates [mm]
During halo measurements:
Insert zoom lens, masking mirror, and return mirror.
Digital Micro-Mirror Array
1024  768 grid
of 13.68-µm
square pixels
Mirror array mounted on a
control board, which is tilted
by 45° so that the reflections
are horizontal.
Pixel tilt toggles
about diagonal
by ±12°
Digital Micro-Mirror Array

Advantages:

Flexible masking due to individually addressable pixels



Adapts well to flat beams in electron rings
But the LHC beams are nearly circular
Disadvantages:

The pixels are somewhat large for the LHC



F1 is far from source: Intermediate image is demagnified by 7
RMS size: 14 pixels at 450 GeV, but only 3.4 pixels at 7 TeV
Reflected wavefront is tilted


DMA has features of a mirror and a grating
Camera face must tilt by 24° to compensate for tilt


Known as Scheimflug compensation
Will test this first at SPEAR
Halo Monitor with Digital Mirror Array
Alignment
laser
Calibration light
and target
DMA
PMT and 15% splitter for abort gap monitor
F1 = 4 m
F2 = 0.75 m
Intermediate
image
Diffraction stop
Cameras
Slit
Focus
trombone
Table Coordinates [mm]
During halo measurements:
Insert DMA and return mirror; rotate camera.
Summary: LHC Synchrotron Light

Synchrotron light is in routine use for observing the beam size
and for monitoring the abort gap.




Improvements proposed for better beam images
Two tests of abort-gap cleaning have been successful, with the
abort-gap monitor showing changes in the gap population.
A longitudinal-density monitor is being developed.
Now considering how to add a halo monitor. First measure the
halo at SPEAR, where access to the light is easy.
Summary: LCLS Terahertz Source


First THz light measured in October
Measurements so far:




Titanium-sapphire laser to be installed this winter:



Scans of THz energy versus charge and bunch length
Spectrum using interferometer
Nonlinear absorption in GaAs
Mapping temporal profile of the THz electric field using polarization
rotation in an electro-optic crystal
THz/laser pump/probe
Later, design a transport line to the NEH