SLAC-PUB-7362

A Proposed Injector for the LCLS Linac

*Work supported by Department of Energy contract DE–AC03–76SF00515.

Stanford Linear Accelerator Center

Stanford University, Stanford, CA 94309

A.D. Yeremian, V. K. Bharadwaj, P. Emma, R.H. Miller,

D.T. Palmer, M.D. Woodley

November 1996

Presented at 17th IEEE Particle Accelerator Conference (PAC 97): Accelerator Science,

Technology, and Applications, May12–May16, 1997, Vancouver, B. C., Canada

A PROPOSED INJECTOR FOR THE LCLS LINAC

*

A. D. Yeremian, V. K. Bharadwaj, P. Emma, R. H. Miller,

D. T. Palmer, M. D. Woodley

Stanford Linear Accelerator Center

Stanford, California 94309

Abstract

The Linac Coherent Light Source (LCLS) will use the

last portion of the SLAC accelerator as a driver for a short

wavelength FEL. The injector must produce 1-nC, 3-ps

rms electron bunches at a repetition rate of up to 120 Hz

with a normalized rms emittance of about 1 mm-mrad.

The injector design takes advantage of the photocathode rf

gun technology developed since its conception in the mid

1980's [1], in particular the S-band rf gun developed by

the SLAC/BNL/UCLA collaboration [2], and emittance

compensation techniques developed in the last decade

[3,4]. The injector beamline has been designed using the

SUPERFISH, POISSON, PARMELA, and

TRANSPORT codes in a consistent way to simulate the

beam from the gun up to the entrance of the main

accelerator linac where the beam energy is 150 MeV.

PARMELA simulations indicate that at 150 MeV, space

charge effects are negligible.

1 BEAMLINE

The injector beamline for LCLS consists of a 1.6 cell

S-band photocathode rf gun, 4 SLAC type 3 m S-band

constant gradient traveling wave accelerator sections,

emittance compensation solenoids at the gun (after

accelerator section 2 and after accelerator section 3), a 14˚

achromatic, isochronous translation bend system (dog-leg)

into the main linac, and instrumentation for beam

diagnostics as shown in Fig. 1.

The rf distribution system is also shown in Fig. 1.

Two un-SLEDed 5045 klystrons will be used: one to

power the gun and the first two accelerator sections and

the other for the last two accelerator sections. High power

rf attenuators and phase shifters will allow independent

control of the rf power, phase and amplitude for the gun

and each of the first two accelerating sections.

Many diagnostics have been designed into the injector

to characterize the beam and to aid in tuning the injector.

In the drift section from the gun to the first accelerator

there will be a beam position monitor (BPM) a current

monitor, and a profile screen/Faraday cup pop-in device. A

mask in the laser beam into the photocathode can be used

to produce beamlets in various patterns to study emittance

trends from the gun, or just as a tuning aid to optimize

the steering. After each accelerator section there is a BPM,

* Work supported by Department of Energy contract

DE-AC03-76SF00515

a current monitor, and a screen. After the second

accelerator section there is a bunch length monitor which

is an x-band cavity outside the beamline near a ceramic

gap in the beamline [5]. While this monitor will not

characterize the longitudinal bunch profile, it will be a

useful tuning aid to minimize the bunch length.

A full transverse emittance diagnostic section is

included prior to the dog-leg which consists of four wire-

scanners [6] separated by 45˚ of x and y betatron phase

advance. The matched input beam produces an x and y rms

beam size of 67

µ

m at each wire. A fifth wire scanner in

the dog-leg allows the energy spread measurement.

Emittance and energy scrapers are also planned in this

region.

2 SIMULATION

Emittance growth and compensation in the LCLS

injector from the gun to the end of the fourth accelerator

section has been thoroughly studied using PARMELA

simulations. The electric field map of the rf gun was

obtained with SUPERFISH and directly used in

PARMELA. SUPERFISH was used to simulate the fields

in the traveling wave accelerator sections and space

harmonics were calculated for use in PARMELA. The rf

fields were assumed to be cylindrically symmetric. Much

care has been taken to symmetrize the dominant dipole rf

fields in the gun [2] thus assuming cylindrical symmetry

in the rf fields is not unreasonable.

A magnetic field map for the solenoid magnets at the

gun was produced using POISSON and passed to

PARMELA. Single coils were used to represent magnetic

fields from solenoids between accelerator sections.

Space charge effects were included assuming cylindrical

symmetry. For these simulations, the thermal emittance

at the cathode and the emittance growth due to multipole

electric fields were ignored. The magnetic and electric

fields in the gun and the accelerator region were optimized

to minimize the emittance at 150 MeV at which point the

normalized emittance is virtually constant.

The peak electric field in the gun is 150 MV/m at the

cathode and the laser is injected at 23 degrees ahead of the

rf crest. In the first and second accelerator sections the

centroid of the beam is about 5 degrees ahead of the crest

and the gradient is 7 MV/m. This introduces about a 1.5%

energy spread for the full beam at the exit of the second

accelerator but helps slightly in emittance compensation.

This energy spread is removed by the third accelerator

section by phasing the rf such that the beam arrives

4–97

8253A1

Vac valves Attenuator

Phase ShifterSolenoid

BPM,

profile

monitor

BPM, current mon,

Faraday cup, screen

Steering

coils

Screen

Toroid

E scraper

Matching & Diagnostics

7 Quads,

BPMs,

steering

Main

linac

Klystron

Streak camera, screen,

BPM, Toroid

Beam stopper

Wire scanner

Toroid, BPM,

screen, BLM

RF

gun

K K

Figure 1. Beamline layout from rf gun to injection into the SLAC main linac. The ‘matching & diagnostics’ section

includes matching quadrupoles, 4 wire scanners for transverse diagnostics, 4 phase space scrapers, a toroid, BPMs and a

streak camera station for bunch length measurements. The two dog-leg inflector bends follow the diagnostic section.

slightly behind the crest. The gradient of the third and

fourth accelerator sections is 17 MV/m.

The magnetic field at the gun is generated by two

solenoids placed symmetrically up and downstream of the

cathode such that the axial magnetic field at the cathode is

zero. The field then rises sharply after the cathode to about

3 kG with an effective length of about 20 cm. The axial

magnetic field between accelerator sections is slightly

greater than 3 kG with an effective length of 70 cm. Fig.

2 shows the axial magnetic fields along the length of the

beamline where s = 0 is the cathode location.

Figure 2. Solenoidal magnetic field along the beamline

from the cathode (s = 0) to the 150 MeV point.

At the cathode, the edge beam radius is about 0.9 mm

dictated by emittance compensation—a balance between

space charge, magnetic field, and electric field effects. At

the entrance of accelerator-section-3, however, radial rf

fields have a more dominant effect on emittance growth

than space charge since the beam energy is nearly 50 MeV

In addition, the accelerating gradient in section-3 is more

than twice that of sections 1 and 2. Thus the beam radius

is drastically reduced at the entrance of the third accelerator

section to reduce the emittance growth due to radial

electric field effects.

The transverse distribution at the cathode is assumed to

be uniform, but the temporal shape is assumed to be a

truncated Gaussian achieved by using a Gaussian pulse

with an rms of 4.4 ps which is then truncated at ±2

σ

. The

overall rms length is then 3.8 ps. In fact, a uniform

temporal distribution is more desirable for optimal

emittance compensation, however a truncated Gaussian

results in a bunch-compression system which is much

less sensitive to injection timing jitter [7].

It should be noted here that simulations indicate a 40%

shorter beam at 150 MeV than at the cathode due to

bunching because the laser is fired 23 degrees ahead of the

rf. However there is experimental indication of bunch

lengthening in the 1.6 cell S-band rf gun at these current

densities which is not manifested in PARMELA

simulations [8]. Thus the expected bunch length at 150

MeV may be longer than indicated in simulation by

possibly 30-40%. Fig. 3 shows the temporal and energy

distributions and phase space as well as an x-y particle

scatter plot of the simulated bunch at 150 MeV.

Figure 3. Temporal distribution (a), x-y space (b),

longitudinal phase space (c), and energy distribution (d) at

150 MeV for Table 1 parameters.

Fig. 4 is a plot of the horizontal normalized emittance

(

βγε

x

, where

β

≡ v/c) along s for an optimized parameter

set of the injector beamline shown in Fig. 1. Bunch

parameters at the end of the fourth accelerator section are

given in Table 1. The core emittance (excluding a 7.7%

halo) at 150 MeV is 1.08 mm-mrad for the truncated

Gaussian temporal bunch shape.

If a uniform distribution with rise and fall of less than

1 ps is used at the gun, the emittance at 150 MeV is

0.95 mm-mrad (excluding a 3% halo). We have also

simulated the 1 nC, uniform transverse and Gaussian

temporal bunch truncated at ±4

σ

(i.e. nearly a full

Gaussian pulse). In this case the core emittance at 150

MeV is 1.07 mm-mrad with the exclusion of 8.1% halo.

Figure 4. Normalized emittance (solid) and rms beam size

(dash) along the beamline from cathode (s = 0) to 150 MeV.

The step at s ~ 16 m is a 7.7% halo cut (in simulation only).

Table 1. Simulated electron beam parameters at 150 MeV.

Longitudinal distribution — Gaussian

Transverse distribution — uniform

Bunch charge nC 1.0

Halo population % 7.7

rms pulse length psec (mm) 2.3 (0.69)

Energy MeV 150.5

rms relative energy spread % 0.13

Norm. rms core emittance mm-mrad 1.08

3 DOG-LEG INFLECTOR

The function of the dog-leg inflector is to transport the

150 MeV electron beam from the new injector linac into

the existing SLAC linac. It is possible to design the dog-

leg as a first bunch compression stage, but this

necessitates a large incoming correlated energy spread of 1-

2% and the chromaticity of the quadrupole magnets within

the dog-leg—required for linear achromaticity—generate

large second order dispersion which requires sextupole

compensation. Due to this, and also the need to tune the

momentum compaction of the compressor (not natural in

a dog-leg), the dog-leg is designed as a simple transport

line, not a compressor.

The design requirements of the line are: 1) should

provide a horizontal beamline inflection of ~1 meter over

a few meters distance, 2) should not alter the bunch length

(isochronous), and 3) should introduce no significant

transverse emittance dilution. The inflection may also be

made in the vertical plane or a rolled plane. However,

there is no strong motivation to do so. A simple system

which satisfies these conditions is composed of two dipole

magnets of opposite strength separated by a +I optical

transformer (seven quadrupoles) to produce a linear

achromat. The dipoles are rectangular bends.

The beamline is 4 meters long with two 14˚, 20-cm

long bends providing the ~1 meter inflection. This

produces a momentum compaction, R

56

, of 4 mm. An

extreme electron which is off energy by 1% will move

axially by 40

µ

m which is small compared to the ~1 mm

bunch length. The effect of the second order momentum

compaction, T

566

, is even less. Note, the incoming rms

energy spread from the injector is 0.13%. The system is

therefore, for all practical purposes, isochronous with no

significant chromatic emittance dilution.

4 SUMMARY

The injector design for the LCLS takes advantage of

the present development of photocathode rf gun

technology to produce 150 MeV, 1 nC, 3 ps rms electron

bunches with a normalized rms emittance of 1.1 mm-

mrad. The necessary beamline and rf systems have been

defined to produce such beams and the design includes

sufficient diagnostic components to aide in the production

and characterization of the electron bunches.

5 ACKNOWLEDGMENTS

We are grateful to Dr. Luca Serafini of INFN-Milan in

Italy and Dr. James Rosenzweig of UCLA for many

helpful discussions on the production and maintenance of

the high brightness beams from rf photocathode guns.

6 REFERENCES

[1]. J. S. Fraser, et. al., "Photocathodes in Accelerator

Applications", Proceedings of the IEEE Particle

Accelerator Conference, Wash. DC, 1987, p 1705.

[2]. D. T. Palmer, et. al., "Microwave Measurements of

the BNL/SLAC/UCLA 1.6 Cell Photocathode RF

Gun", Proceedings of the 1995 IEEE Particle

Accelerator Conference, Dallas, 1995, p982.

[3]. B. E. Carlsten, "New Photoelectric Injector Design

for the LANL XUV FEL Accelerator", NIM, A285,

1989, p313.

[4]. L. Serafini and J. B. Rosenzweig, "Envelope

Analysis of Intense Relativistic Quasi-Laminar

Beams in RF Photoinjectors: A Theory of Emittance

Compensation", to be published in Phys. Rev. E.

[5]. R. H. Miller, "Proposed Bunch Monitor", SLAC

TN-63-65, August 1963.

[6]. M. C. Ross, et. al., "Wire Scanners for Beam Size

and Emittance Measurements at the SLC",

Proceedings of the IEEE Particle Accelerator

Conference, San Francisco, 1991, p 1201.

[7]. V. K. Bharadwaj, et. al., “Linac Design for the

LCLS Project at SLAC”, these proceedings.

[8]. D. T. Palmer, et. al., "Emittance Studies of the

BNL,/SLAC/UCLA 1.6 cell Photoacthode rf gun",

these proceedings.