CLARA - Science and Technology Facilities Council · FELs and have strategically decided that CLARA should have a longer term vision and be targeted at proving concepts which will

CLARA Conceptual Design Report

July 2013

The Council does not accept responsibility for loss or damage arising from use of information contained in any of its reports or in any communications about investigations.

© Science and Technology Facilities Council.Enquiries about copyright and reproduction should be addressed to:

The LibrarianScience and Technology Facilities CouncilDaresbury LaboratorySci-Tech DaresburyDaresbury, WarringtonWA4 4AD

CLARA Conceptual Design Report

J. A. Clarke1,2, D. Angal-Kalinin1,2, N. Bliss1, R. Buckley1,2, S. Buckley1,2, R. Cash1,P. Corlett1,2, L. Cowie1,2, G. Cox1, G. P. Diakun1,2, D. J. Dunning1,2, B. D. Fell1,

A. Gallagher1, P. Goudket1,2, A. R. Goulden1,2, D. M. P. Holland1,2, S. P. Jamison1,2,J. K. Jones1,2, A. S. Kalinin1,2, W. Liggins1,2, L. Ma1,2, K. B. Marinov1,2, B. Martlew1,

P. A. McIntosh1,2, J. W. McKenzie1,2, K. J. Middleman1,2, B. L. Militsyn1,2,A. J. Moss1,2, B. D. Muratori1,2, M. D. Roper1,2, R. Santer1,2, Y. Saveliev1,2,E. Snedden1,2, R. J. Smith1,2, S. L. Smith1,2, M. Surman1,2, T. Thakker1,2,N. R. Thompson1,2, R. Valizadeh1,2, A. E. Wheelhouse1,2, P. H. Williams1,2,

R. Bartolini3,4, I. Martin3, R. Barlow5, A. Kolano5, G. Burt2,6, S. Chattopadhyay2,6,7,8,D. Newton2,7, A. Wolski2,7, R. B. Appleby2,8, H. L. Owen2,8, M. Serluca2,8, G. Xia2,8,S. Boogert9, A. Lyapin9, L. Campbell10, B. W. J. McNeil10, and V.V. Paramonov11

1STFC Daresbury Laboratory, Sci-Tech Daresbury, Warrington, UK2Cockcroft Institute, Sci-Tech Daresbury, Warrington, UK

3Diamond Light Source, Oxfordshire, UK4John Adams Institute, University of Oxford, UK

5University of Huddersfield, UK6University of Lancaster, UK7University of Liverpool, UK

8University of Manchester, UK9John Adams Institute at Royal Holloway, University of London, UK

10University of Strathclyde, UK11Institute for Nuclear Research of the RAS, Moscow, Russian Federation

ii CLARA Conceptual Design Report

CLARA Conceptual Design Report iii

Acknowledgments

The authors gratefully acknowledge the support we have received from the international accelerator andFEL communities in the development of the conceptual design of CLARA. Special thanks are due tocolleagues at the Paul Scherrer Institute for their very helpful advice and input to this report and theirsupport of CLARA in general. The authors are also extremely grateful to colleagues at the Laboratoire del’Accélérateur Linéaire for sharing their experience of RF photoinjectors. Finally, we would like to thankthe University of Strathclyde for kindly providing the RF photoinjector, klystron and other equipment forthe VELA project.

iv CLARA Conceptual Design Report

CLARA Conceptual Design Report v

Contents

1 Introduction and Motivation 1

2 Benefit to VELA 7

3 FEL Design 113.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113.2 Parameter Selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113.3 Operating Modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

3.3.1 Seeding Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123.3.2 SASE Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133.3.3 Ultra-Short Pulse Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133.3.4 Multibunch Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143.3.5 Electron Beam Stability Requirements . . . . . . . . . . . . . . . . . . . . . . . . 14

3.4 FEL Layout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153.4.1 Modulator Section . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153.4.2 Radiator Section . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153.4.3 Afterburner . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

3.5 FEL Schemes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163.5.1 SASE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163.5.2 Generation of Short Pulses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173.5.3 Improving Temporal Coherence . . . . . . . . . . . . . . . . . . . . . . . . . . . 213.5.4 Afterburner Schemes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

3.6 Scaling to Short Wavelengths . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

4 Accelerator Design 294.1 Layout Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

4.1.1 Phase Space Linearisation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 304.1.2 Energy at Magnetic Compressor . . . . . . . . . . . . . . . . . . . . . . . . . . . 304.1.3 Variable Bunch Compressor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 314.1.4 Diagnostic Sections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

4.2 Beam Dynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 324.2.1 Electron Source . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 324.2.2 Optimisation of Seeded Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . 344.2.3 Other Operating Modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

4.3 Tolerance Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

vi CLARA Conceptual Design Report

4.3.1 Beam Based Alignment Strategy . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

5 Accelerator Systems 395.1 Electron Source . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

5.1.1 Baseline Electron Source . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 395.1.2 Advanced Electron Source . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

5.2 Radio Frequency Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 425.2.1 Linac Accelerating Structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . 445.2.2 High Power RF Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 465.2.3 Low Level RF System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46

5.3 Electron Beam Diagnostics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 505.3.1 Bunch Charge Monitors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 505.3.2 Strip Line BPMs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 505.3.3 Cavity BPMs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 515.3.4 Screen Diagnostic Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 515.3.5 Beam Arrival Monitors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 515.3.6 Bunch Compression and Temporal Profile Monitors . . . . . . . . . . . . . . . . . 525.3.7 Laser Arrival Monitors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52

5.4 FEL Output Diagnostics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 525.4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 525.4.2 Spectral Diagnostics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 535.4.3 Temporal Diagnostics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54

5.5 Optical Timing and Synchronisation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 555.5.1 Synchronisation Targets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 555.5.2 Timing System Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 555.5.3 RF Master Oscillator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 565.5.4 Laser Master Oscillator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 575.5.5 Optical Clock Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 575.5.6 Laser-To-Laser Synchronisation . . . . . . . . . . . . . . . . . . . . . . . . . . . 575.5.7 Beam Arrival Monitors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 575.5.8 Laser Arrival Monitors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 575.5.9 Referencing of LLRF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

5.6 Lasers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 585.6.1 Seed Lasers for FEL Modulation . . . . . . . . . . . . . . . . . . . . . . . . . . . 585.6.2 Photoinjector Laser System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 595.6.3 Lasers for FEL Photon Diagnostics . . . . . . . . . . . . . . . . . . . . . . . . . 605.6.4 Laser Synchronisation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60

5.7 Undulators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 605.8 Control System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61

5.8.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 615.8.2 Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 625.8.3 Controls Hardware . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 625.8.4 Timing and Synchronisation System . . . . . . . . . . . . . . . . . . . . . . . . . 625.8.5 Interlock Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 635.8.6 Feedback Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63

6 Radiation Safety 656.1 Shielding Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

6.1.1 Radiological Classification of Areas . . . . . . . . . . . . . . . . . . . . . . . . . 656.1.2 Source and Material Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 656.1.3 Shielding Calculations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

6.2 Personnel Safety System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

CLARA Conceptual Design Report vii

6.2.1 Purpose . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 666.2.2 Design Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 666.2.3 Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

7 Potential Upgrades and Future Exploitation 697.1 Plasma Accelerator Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 697.2 Ultrafast Electron Diffraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 707.3 Compton Photon Production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 707.4 Dielectric Wakefield Acceleration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 717.5 Nonequilibrium Electron Rings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 717.6 Exotic Storage Ring Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 727.7 Industrial Exploitation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72

References 75

viii CLARA Conceptual Design Report

CLARA Conceptual Design Report 1

1Introduction and Motivation

Figure 1.1: CLARA – the Compact Linear Accelerator for Research and Applications

Free-electron lasers (FELs) have made hugeadvances in the past few years with the firstsuccessful demonstration of an X-ray FEL at LCLSin the USA in 2009 [1], followed by similar successat SACLA in Japan in 2011 [2]. New X-ray facilitiesare currently under construction in Germany [3],Switzerland [4], and elsewhere and soft X-ray FELs,such as FLASH in Germany [5] and FERMI@Elettrain Italy [6], are also operating for users routinely.Whilst the new X-ray FELs are remarkable intheir performance the potential for improvementsis enormous. Many suggestions have been made byFEL experts for improving the FEL photon output in

terms of temporal coherence, wavelength stability,increased power, intensity stability and ultra-shortpulse generation. Unfortunately, given the lownumber of operating FELs and the pressure todedicate significant time for user exploitation it isnot surprising that very few of these ideas havebeen tested experimentally. This Conceptual DesignReport describes the design of CLARA (CompactLinear Accelerator for Research and Applications),a dedicated flexible FEL Test Facility, which will beable to test several of the most promising of thenew schemes. Figure 1.1 shows a three-dimensionalrepresentation of CLARA. The successful proof of

2 CLARA Conceptual Design Report

Single-Spike SASE

Emittance Spoiling

Emittance Spoiling

Superradiant Pulse Shortening

Laser Slicing

Mode-Locking + Modelocked Afterburner

Pre-bunching + Short Radiator

Laser Seeding

Self Seeding

Cavity FELs - RAFEL and X-FELO

Harmonic Cascades and Echo-Enabled Harmonic Generation

High-Brightness SASE

HHG Seeding

UL

TR

AS

HO

RT

PU

LS

ES

TE

MP

OR

AL

CO

HE

RE

NC

E

Multicolour FELs

Mode-Locking

Echo-Enabled Harmonic GenerationTAIL

OR

ED

PU

LS

ES

All Seeding

Tapering

Cavity FELs - RAFEL and XFELO

STA

BIL

ITY

& P

OW

ER

IN ROUTINE USE

or WELL TESTED

PROOF-OF-PRINCIPLE

or NOT TESTED

Figure 1.2: Schematic representation of the FEL landscape in terms of potential improvements to particularoutput properties against progress to date.

principle demonstration with CLARA will be a vitalstepping stone to the implementation of any newscheme on an existing or planned FEL facility.

Of course CLARA will not operate in isolationand existing FELs are already dedicating machinedevelopment time for the testing of new ideas,such as self seeding or harmonic generation, andthis is sure to continue given the strong call fromusers for ever higher quality light output. We havecarefully assessed the short term focus of existingFELs and have strategically decided that CLARAshould have a longer term vision and be targetedat proving concepts which will not just have anincremental impact on FEL performance but takeFELs into a whole new regime. We believe thatthis will ensure that the international impact ofCLARA will be maximized and place the UK in avanguard position should it choose to develop its

own future FEL facility. The landscape showingthe potential improvements currently identified forFELs is presented schematically in Figure 1.2. It isclear that whilst significant progress has been maderecently in making FELs operational, there are manymajor areas where the potential of FELs remainsuntapped. We aim to release this potential to openup further new fields of science for investigation andexploitation.

Our vision for CLARA is that it should bededicated to the production of ultra-short photonpulses of high-brightness coherent light. ExistingX-ray FELs are already capable of generating pulsesof light that are only tens of femtoseconds induration (tens of thousands of optical cycles) butFEL experts have proposed several schemes whichhave the potential to generate pulses that are twoor three orders of magnitude shorter than this


(hundreds or tens of attoseconds) [7, 8, 9, 10, 11,12, 13, 14, 15] and a recent paper has even proposeda novel idea for sub-attosecond pulse generationfrom a FEL (a few optical cycles) [16]. The sciencewhich is enabled by ultra-short photon pulses wasdescribed in detail in the NLS Science Case [17]and reviews of science carried out using attosecondpulses are also available [18, 19, 20]. Many excitingapplications of attosecond pulses have already beendemonstrated, including coherent X-ray imaging,femtosecond holography, real-time observations ofmolecular motion and capturing the movement ofelectrons in atoms and molecules. Attosecond X-rayscience could revolutionise how we understand andcontrol electron dynamics in matter.

In order to achieve this vision for ultra-shortpulse generation, CLARA must be able toimplement advanced techniques, such as laserseeding, laser-electron bunch manipulation, andfemtosecond synchronisation. These can onlybe achieved by developing a state-of-the-artaccelerator with the capability to drive current FELdesigns. CLARA is therefore of direct relevance tothe wider international FEL community and will alsoensure that the UK has all the skills required shouldit choose to develop its own future FEL facility.This underpinning foundation requirement is shownin Figure 1.3.

In detail the goals, opportunities and benefits ofCLARA will be:

• Proof-of-principle demonstrations of ultra-short‘attosecond’ photon pulse generation (of orderof the coherence length or less, typically lessthan 100 optical cycles) using schemes which areapplicable to X-ray FELs (such as laser slicing,mode locking, or single spike Self-AmplifiedSpontaneous Emission (SASE)) and with extremelevels of synchronisation.

• The ability to test novel schemes for increasingthe intrinsic FEL output intensity stability,wavelength stability, or longitudinal coherenceusing external seeding, self seeding or throughthe introduction of additional delays within theradiator section.

• The ability to generate higher harmonics of a seedsource using Echo Enabled Harmonic Generation(EEHG), High Gain Harmonic Generation(HGHG), or other novel schemes.

• The generation and characterisation of verybright (in 6D) electron bunches and the

subsequent manipulation of their properties withexternally injected radiation fields, and the testingof mitigation techniques against unwanted shortelectron bunch effects.

• The development and demonstration of advancedaccelerator technologies, with many wideranging applications well beyond FELs, suchas a high repetition rate normal conductingRadio Frequency (RF) photoinjector, novelundulators, RF accelerating structures andsources, single bunch low charge diagnostics, andnovel photocathode materials and preparationtechniques. The potential to test the newgeneration of plasma-based accelerators asdrivers of FELs is also a significant consideration.

• The enhancement of VELA (Versatile ElectronLinear Accelerator) [21], in terms of energy, beampower, and repetition rate, enabling additionalindustrial applications of electron beams that arecurrently excluded.

• A flexible, high quality, electron test acceleratoravailable to the entire UK accelerator communityon a proposal-driven basis, enabling wide ranging,high impact accelerator R&D.

• The development and retention of vital skillswithin the UK accelerator community, includingproviding excellent opportunities for attractingthe best PhD students and early stage researchersto work on a world class accelerator test facility.

• The possibility to use the high quality brightelectron beam for other scientific researchapplications such as ultra-fast electron diffractionexperiments, plasma wakefield accelerators (asa witness bunch or a drive bunch), as thedrive beam for a Compton scattering source ofX-rays or gamma photons, and for other novelacceleration schemes such as dielectric wakefieldaccelerators and exotic storage rings.

We recognise the dynamic nature of FELresearch and aim not just to demonstrate andstudy the novel concepts of today but also be wellpositioned to prove the novel concepts of tomorrow.For these reasons we have planned a number ofdifferent CLARA operating modes, each of whichis designed to be appropriate for a different class ofFEL experiments. The complete set of experimentsdiscussed in this design report impacts on every areaof the future FEL landscape presented in Figure 1.2.


Cathode

Preparaon

Temperature

Stability

Simulaon

Codes

Underpinning

Infrastructure

Vibraon

Control

Environment

Control

FEA

Modelling

Vacuum

Technology

Transverse &

Longitudinal

Diagnoscs

Photoinjector

Laser

Normal Conducng

RF Systems

Magnets and

Undulators

Digital Low

Level RF

Timing

System

RF & Laser

Master Clock

Bunch

CompressionFeedback Systems Feed Forward

Systems

RF Phase and

Amplitude

Control

Trajectory

Straightness

Accurate

Alignment

Transverse

Matching to

FEL

High Peak CurrentsHigh Pulse

Energy Seed

Laser

Long and

Short Term

Stability

Low Emiance Electron

Bunches

Laser and

Electron

Overlap

Electron Energy Modulaon

with External Fields

Synchronisa-

on of Laser

& Electrons

Low Bunch

Timing JierHarmonic Generaon

Single-Shot

FEL Photon DiagnoscsLaser Seeding

ULTRA-SHORT PULSE GENERATION

UNDERPINNING

REQUIREMENTS

ULTIM

ATE

GOAL

Figure 1.3: To achieve the primary goal of CLARA will require the mastery and understanding of manyother techniques and technologies.


This flexibility of approach also extends to advancedtechnological aspects, such as retaining the abilityto implement and test alternative RF systems inthe future, for example C-band linac structures oreven superconducting RF cavities, as required by theneeds of the UK accelerator programme.

CLARA will be well placed to test emerging ideasfor fifth generation light sources based on laser orelectron driven plasma accelerators, such as thoseput forward by the FACET-II proposal [22]. Theexperimental testing of such advanced ideas, thatscale well to larger projects, can result in leapsforward in capability.

The ability to simulate coherent light sourcesdriven by plasma accelerators requires differentcomputational models from those used to simulatelight sources driven by conventional accelerators.A new simulation code has been developed byASTeC in collaboration with researchers at theUniversity of Strathclyde [23]. This is the only such

code currently available and is generating significantinternational interest resulting in collaborationsdeveloping with international groups that wish touse both laser and electron beam driven plasmaaccelerators as FEL drivers.

These collaborations offer the potential forCLARA to develop into an international focalpoint for such research. The embedded excellenceof ASTeC staff in the design of electron beamtransport systems and start-to-end modellingcapability would ensure that any plasma drivenresearch towards a FEL would be internationallyleading.

Since CLARA is intimately linked to the existingVELA facility, much of the essential infrastructurefor the project already exists. This will significantlyreduce the time required to implement CLARA.We believe that within 3 years of funding wecould procure and install all of the equipment andcommence beam commissioning.

International Context

Currently there are only four dedicated single pass FEL test facilities worldwide, one in the US (NLCTA),one in Asia (SDUV-FEL) and two in Europe (SPARC and MAX). We have assessed the capabilities andprogrammes of each of these facilities in turn and have confirmed that CLARA will offer unique capabilitiesand have a complementary programme to the other test facilities.

The NLCTA at SLAC is a low energy (120 MeV) test accelerator deliberately focussing on near termResearch and Development (R&D). The main goal of this facility is the development and optimization ofEEHG for the improvement of temporal coherence, indeed this facility was the first to demonstrate EEHGexperimentally. The primary objective now appears to be the establishment of very high order harmonicgeneration with the EEHG scheme.

The SDUV-FEL in Shanghai is the only dedicated test facility in Asia. They first observed SASE lasingin 2009 and this has been followed by seeded FEL experiments, EEHG, and other techniques for generatinghigh harmonics from a seed laser, such as HGHG.

The SPARC test FEL at Frascati currently operates at 150 MeV with a planned upgrade to take it to250 MeV. SASE lasing has been demonstrated and characterised, and direct High Harmonic Generation(HHG) seeding is being studied. SPARC has also carried out some short pulse experiments including energychirped electron beams and undulator tapering and seeded superradiance.

The MAX-lab test FEL in Sweden will close at the end of 2013 after carrying out a final experimentexploring seeding by an HHG source at lower harmonics.

In summary, there are currently four FEL test facilities with similarities to CLARA. Improved temporalcoherence using various seeding techniques dominate the programmes but one of the facilities has alsocarried out some promising experiments aimed at short pulse generation.

Two other pioneering facilities, previously dedicated to FEL development, have now evolved into largerprojects due to the high demand for their output from users. The FLASH facility in Germany operates at1.25 GeV, generating light down to 4 nm, and has operated for user experiments since 2005. SCSS in Japanoperates at 250 MeV and routinely delivers 50–60 nm light for user experiments. Other accelerator testfacilities, such as the ATF at Brookhaven, have broader programmes not concentrated on FEL development.



2Benefit to VELA

Figure 2.1: Photograph of the existing VELA facility at Daresbury Laboratory.

The upgrade of the VELA facility [21] (shownin Figure 2.1) to incorporate CLARA will beof considerable benefit to VELA itself. TheCLARA project has been carefully designed tominimize disruption to the operation of VELAduring the installation and commissioning phasesby making a number of key design choices. First,

the CLARA components will be assembled ontoindividual girder modules offline in the EngineeringTechnology Centre at Daresbury Laboratory. Thisapproach was used successfully by VELA and otherprojects and has been shown to be an efficientmethod of assembly, allowing easy access to eachmodule for hardware installation and commissioning


Figure 2.2: Engineering layouts of (top) the existing VELA facility and (bottom) VELA and the front endof CLARA once CLARA is fully installed.

and enabling accurate alignment of individualcomponents on each module. Second, CLARA willbe permanently installed parallel to VELA, spacedapproximately one meter from the current VELAbeam axis. This will mean that, with carefulscheduling of installation activities, VELA will beable to continue to operate for users in parallelwith CLARA installation. Indeed, the majority ofCLARA is well away from VELA and this can beinstalled simultaneously to VELA operations withno disruption at all. There will, of course, haveto be a scheduled shutdown of VELA for finalinstallation which will include: the reconfigurationof the radiation shielding into a single tunnel; thetransfer of the photoinjector gun cavity across tothe CLARA front end, along with modifications tothe associated RF waveguide and photoinjector lasertransport systems; breaking of the VELA machinevacuum to join the two facilities together with ashort dogleg electron beam transport line betweenthe CLARA beam axis and the existing VELA axis.This last aspect is another deliberate design choicewhich will mean that the VELA user areas willbenefit from enhanced electron beam parameterspost-CLARA installation.

Figure 2.2 shows the current VELA engineeringlayout and the layout once the installation ofCLARA is complete. Both facilities will thenbe served by a single photoinjector, mounted onthe CLARA beam axis. Initially this will be theexisting VELA gun but will later be replaced bya more advanced gun, capable of a much higherrepetition rate (400 Hz compared with 10 Hz).For VELA operation the dogleg transfer line will beenergised and beam delivered to either the straightahead or perpendicular user beam areas. Clearlythis operating scheme does not allow simultaneousoperation of both VELA and CLARA but sinceVELA is not scheduled to run 24 hours a day itis likely that both facilities can operate successfullyin this mode with intelligent scheduling of beamtime. However, if the user demand for both facilitiescannot be met, it would be relatively straightforwardto implement a dedicated electron gun for VELAin the space vacated by removing the existingphotoinjector cavity. In this case it is likely thata thermionic electron gun would be implemented.The option for the highest quality bunches fromthe CLARA photoinjector would still be available asrequired. The added benefit of transferring electron


bunches from CLARA to VELA is that they are ableto undergo further acceleration before the doglegand so higher energy electrons will be available toVELA users once CLARA is installed. At presentVELA will deliver up to 5 MeV electrons. Followingthe CLARA installation VELA will be able to offerup to 25 MeV electrons. This will increase thebeam power available by a factor of five. Thesubsequent implementation of the high repetitionrate photoinjector will increase the beam power by

another factor of 40. Hence, VELA users will seea factor of 200 beam power increase following thefull installation of CLARA. If the additional upgradeof a thermionic gun is implemented later the beampower will be increased even further due to the veryhigh number of bunches per macropulse that will bepossible. Note that increased beam power has beenidentified during an extensive market survey as oneof the key parameters from which users of VELAwould benefit most [24].



3FEL Design

3.1 Introduction

As discussed in Chapter 1 CLARA has been designedto be a dedicated flexible FEL test facility, which willbe able to test several of the most promising of thenew FEL schemes. The successful proof of principledemonstration with CLARA will be a vital steppingstone to the implementation of any new scheme onan existing or planned FEL facility.

In this chapter we discuss the main parametersfor CLARA, before detailing a number of differentCLARA operating modes each of which is designedto be suitable for a different class of FELexperiments. We then summarise the layout of theFEL systems before giving details of a selectionof the FEL schemes we will be able to study onCLARA, including some predictions of the FELoutput properties. We broadly divide the researchtopics into two main areas, each of which is intendedto demonstrate new schemes for the improvementof FEL output beyond that available from theSASE process [25, 26]. The first area is thegeneration of ultra-short pulses and the second areais improvement of temporal coherence. The physicsof the FEL mechanism is not described in this reportbut is covered in a number of texts [27, 28, 29]. Anumber of review articles give a clear overview ofthe worldwide FEL status and future prospects [30,31, 32].

CLARA has been designed and optimised so thatthe FEL wavelengths are in the visible and VUV.This has tremendous advantages for the operationof the FEL and diagnosis of the output. Ouremphasis for the short pulse schemes is to generatepulses with as few optical cycles as possible withdurations of the order of, or shorter than, the FEL

cooperation length. These are our figures of meritand we aim to study the essential physics of theschemes which can often therefore be independentof the FEL wavelength.

Of course ultimately many of the schemes westudy are intended for application at X-ray FELs.In this case the absolute pulse durations will scalein some way with the wavelength and in the finalsection of the chapter we indicate the potentialof the schemes we study when applied at shorterwavelengths. Of course at these wavelengths therecould be technological issues, tolerance criteria ordisruptive effects which scale with absolute electronbunch length or beam energy. In the final sectionwe also discuss these issues to understand thewavelength scalability of the research we plan toundertake.

3.2 Parameter Selection

The wavelength range chosen for the CLARA FELis 400–100 nm, appropriate for the demonstrationof advanced FEL concepts on a relatively low energyaccelerator. Key drivers for this choice are theavailability of suitable seed sources for interactingwith the electron beam, as required by many ofthe FEL topics we propose to study, and theavailability of single shot diagnostic techniques forthe characterisation of the FEL output.

We propose to study short pulse generation overthe wavelength range 400–250 nm, where suitablenonlinear materials for single shot pulse profilecharacterisation are available. For schemes in thisregime it is often necessary to produce a periodicmodulation in the properties of the electrons along


the bunch such that it can be arranged (using avariety of methods, depending on the scheme) thatonly some sections of the beam can lase. Often thescale length of each independent FEL interactionis given by the slippage in one gain length, knownas the cooperation length lc , so the modulationof the electron bunch properties should have aperiod λM significantly longer than this to enablean independent temporally-separated interaction foreach period of the modulation. This means werequire λM lc . For the lasing wavelength range400–250 nm for the short pulse schemes this leadsto the requirement that to cover most of theresearch topics the modulating laser should havewavelength λ ' 20–50 µm with the possibility ofextending the wavelength to 120 µm later on.

For schemes requiring only spectralcharacterisation (for example producing coherenthigher harmonics of seed sources, or improvingthe spectral brightness of SASE) the operatingwavelength range will be 266–100 nm, with thepossibility of output down to 80 nm at a laterstage. For these schemes the most appropriate seedsource (if required) is an 800 nm Ti:sapphire laserfrom which we can generate coherent harmonics upto the 8th harmonic, or 10th harmonic later on.

To operate at shorter wavelengths than 80 nmwould currently offer little advantage, but manycomplications. The wavelength would in fact haveto be significantly shorter than 80 nm to enablepulse profile characterisation and this would haveto be done using photoionisation in gases as thenonlinear medium for the cross-correlation. Sucha technique is complex and would not providesingle-shot measurements, so is not acceptable.The photon diagnostic beamlines would also bemore complex and require more floor space becauselow incidence angle optics would be necessary.Finally, the CLARA energy would need to beupgraded and the undulator beamline extended.

The required electron beam energy andundulator parameters depend on the required tuningrange. The FEL wavelength λr is given by

λr =λw2γ2

(1 +

K2

2

)where λw is the period of the FEL undulator and γis the electron energy in units of its rest mass. Kis the undulator parameter which is proportional tothe on-axis field which depends on the undulatorgap. As the gap is extended the field decaysuntil K falls below a level that gives useable FEL

interaction. This sets a soft limit for the shortestwavelength. As the gap is reduced the field increasesuntil the minimum gap imposed by the vacuumvessel is reached. This sets the longest wavelength.For CLARA we set the minimum useful undulatorparameter to be K ' 1 (in common with other FELfacilities) and the minimum undulator gap to 6 mm.For a hybrid planar undulator, and a tuning rangeof 400–100 nm, this uniquely defines the requiredelectron beam energy to be E = 228 MeV and theundulator period to be λw = 27 mm. However,we have designed the CLARA accelerator to providea maximum beam energy of 250 MeV. This allowssensible contingency in three areas. First, it allowsthe full wavelength tuning range to be achieved ata slightly reduced linac gradient. Second, it allowsthe linac cavities to be operated further off-crestfor added flexibility. Third it allows us to pushthe FEL wavelength to around 80 nm with onlya slight reduction in undulator parameter enablingus to generate even higher harmonics of the seedsources.

3.3 Operating Modes

The approach we have adopted is to design aflexible, well-diagnosed facility for testing a varietyof advanced FEL concepts. This flexibility willbe built into the accelerator itself, as discussed inChapter 4 and also incorporated into the systemsspecific to the FEL. We recognise the dynamicnature of FEL research and aim not just todemonstrate and study the novel concepts of todaybut also be well positioned to prove the novelconcepts of tomorrow. For these reasons we haveplanned a number of different CLARA operatingmodes, each of which is designed to be appropriatefor a different class of FEL experiments. Thecomplete set of experiments discussed in this designreport spans every area of the future FEL landscapepresented in Chapter 1. The parameters forthe different operating modes are summarised inTable 3.1.

3.3.1 Seeding Mode

This mode is designed for any FEL scheme wherea seed source interacts with the electron beam.For schemes at 100 nm wavelength the acceleratorhas been optimised to produce an electron bunchwith a relatively flat-top current profile of duration


Operating Modes

Parameter Seeding SASE Ultra-short Multibunch

Max Energy (MeV) 250 250 250 250Macropulse Rep Rate (Hz) 1–100 1–100 1–100 1–100Bunches/macropulse 1 1 1 16Bunch Charge (pC) 250 250 20–100 25Peak Current (A) 125–400 400 ∼1000 25Bunch length (fs) 850–250 (flat-top) 250 (rms) <25 (rms) 300 (rms)Norm. Emittance (mm-mrad) ≤ 1 ≤ 1 ≤ 1 ≤ 1rms Energy Spread (keV) 25 100 150 100Radiator Period (mm) 27 27 27 27

Table 3.1: Main parameters for CLARA operating modes.

'250 fs. The specification for the flatness ofthe current within this region is σI/I ≤ 7%, asdiscussed in Section 3.3.5. The reason for the flattop is to make the FEL performance insensitiveto up to ±100 fs temporal jitter between theelectron bunch and seeding laser. The requiredpeak current to reach saturation at 100 nm in asufficiently compact undulator section, taking intoaccount the expected emittance and energy spreadin the beam, is 400 A. These requirements leadto a necessary bunch charge of 250 pC of whichapproximately 100 pC is within the flat top regionand the remainder is in the head and tail.

For schemes where the beam is modulated withthe long wavelength seed the bunch length must belonger. This is because of the slippage that occursbetween the seed field and the electrons in themodulator undulator—every period of the undulatorthe electrons slip back one wavelength. For a 50 µmseed interacting in a 4-period modulator the slippagetime is '650 fs so the flat region of the bunch mustbe at least this duration. Adding contingency for±100 fs temporal jitter the flat top region must be&850 fs duration. Fortunately, the gain length isshorter at longer wavelengths so the peak currentto reach saturation in the same length of undulatoras required for 100 nm operation is only 125 A. Thecharge in the flat top region is therefore 100 pC asbefore, so a total charge of 250 pC is well matchedto the requirements.

In this mode it is important to keep the energyspread small so that the ratio between the energymodulation applied by the laser and the beam energyspread is as high as possible to enable strong

bunching to be generated at high harmonics.

3.3.2 SASE Mode

This mode is designed for operating the FEL inSASE mode. This is the ‘base’ FEL operation modebecause no seed fields or optics are required butwill be an essential mode of operation for validatingthe performance of the accelerator, the propertiesof the electron bunch and the alignment withinthe undulator sections. In addition, many of theschemes proposed for CLARA aim to improve thetemporal coherence and/or stability of SASE so it isimportant to be able to use SASE as a control. Thebunch charge and peak current have already beenset by the requirements of the Seeding Mode so weadopt the same maximum charge and peak currentfor SASE mode. There is however, no requirementfor a flat-top profile so this mode of operationwill be easier to demonstrate. The SASE FELperformance is quite insensitive to energy spread atthe 25–100 keV level so in this mode we allow theenergy spread to be higher than in the seeding mode.

3.3.3 Ultra-Short Pulse Mode

This mode is specific to research into ultra-shortbunch generation and transport, and the use of suchbunches for single spike SASE FEL operation, wherethe bunch length must be shorter than the typicalSASE spike separation of 2πlc . The cooperationlength lc depends strongly on the peak current,which itself of course depends on the bunch chargeand bunch duration. To generate and transport veryshort bunches is only possible if the charge is quite


modest. A study of the parameter space has led tothe conclusion that a bunch charge of 20-100 pC isappropriate for single-spike SASE FEL operation at100 nm wavelength with a peak current requirementof up to 1000 A.

3.3.4 Multibunch Mode

This mode is designed to allow research intoshort-wavelength oscillator FELs, a topic currentlyundergoing a resurgence of interest within the FELcommunity for short wavelengths, with interest inthe Regenerative Amplifier FEL (RAFEL) which isa low-feedback high gain oscillator [33, 34, 35, 36,37, 38], and also the high feedback low gain X-rayFEL Oscillator (X-FELO) [39, 40].

To operate the FEL in oscillator mode wouldrequire some rearrangement of components toinstall an appropriate optical cavity, but werecognise the potential importance of this researcharea and make cost-neutral design choices whichdo not unnecessarily exclude this mode. Theproposed scheme is to demonstrate a RAFELfor the generation of transform limited pulses atwavelengths as short as 100 nm. Such a systemrequires only a very low feedback optical cavity.Our initial studies show that the undulator lengthneeds only be one third of the SASE saturationlength, therefore even using only five of the sevenradiator undulators we can afford to increase thegain length significantly and reduce the peak currentto a modest 25 A. We also require that the electronbunch is longer than the slippage length givingrequired length of ≈ 300 fs rms and hence minimumbunch charge of 15 pC.

The bunch repetition rate must be matched tothe ≈ 17.5 m optical cavity round trip time (a17.5 m cavity assumes the first and last radiatorsare removed and replaced with cavity mirrors andelectron beam chicanes to divert the beam aroundthe mirrors) giving f = 8.5 MHz and bunch spacing120 ns. Typically 8 cavity round trips are requiredto reach saturation so the minimum macropulseduration is 1 µs. We therefore specify a macropulseduration of 2 µs to allow time to diagnose the FELoutput (and its stability) at saturation. The totalcharge in the macropulse is thus 250 pC whichis consistent with other operating modes. Theaverage power demands on the accelerator systemsare therefore no greater than in any other operatingmode.

3.3.5 Electron Beam StabilityRequirements

The stability of the FEL output will depend onthe stability of the electron bunch parameters,particularly the energy and the peak current. Todevelop a complete specification each FEL schemeneeds to be studied in detail. For now we specify,as an absolute minimum, that the energy jitterin the electron bunch should translate to an rmsFEL wavelength jitter of less than half the SASEbandwidth. This will allow practical characterisationof unseeded schemes and ensures that for seededschemes, for which the seed wavelength jitter shouldbe relatively small, the seed always falls within thegain bandwidth of the FEL. From time-dependentGENESIS 1.3 [41] simulations of SASE operationat 100 nm the FWHM bandwidth at saturationis ∆λ/λ ' 5 × 10−3 so the required wavelengthstability is σλ/λ ≤ 2.5× 10−3. From the resonancecondition, energy jitter and wavelength jitter arerelated by ∆E/E = ∆λ/(2λ) so for σλ/λ ≤ 2.5 ×10−3 we require σE/E ≤ 1.25× 10−3.

We also state, in line with the specificationproposed for the NLS [42], the requirement thatthe relative rms shot-to-shot FEL pulse energy jittershould be less than 10%. For SASE the pulseenergy will fluctuate anyway due to the intrinsicnature of the process, but for schemes aimingto improve the temporal coherence of SASE theintrinsic SASE fluctuation will be removed. TheFEL power scales with peak current as P ∝ I4/3

therefore to achieve 10% stability we require therms variation of the peak current to be better thanσI/I ' 7%. Equivalently, for a constant bunchcharge, we require the rms bunch length jitter tobe better than 7%.

For seeding schemes the temporal jitter inthe electron bunch will cause the seed to alignlongitudinally with different sections of the bunch.We therefore require that the temporal jitter is lessthan the flat-top region of the bunch, and thatwithin this flat-top region the current variation isagain less than σI/I ' 7%.

It is emphasised that study of the CLARAtolerances is at an initial stage. The stabilityspecification stated here is a minimum requirement,but as seen in Chapter 4 it is consistent with theinitial tolerance studies made of the acceleratorsystems. Further iteration will strengthen thestability specification consistently with the predictedperformance of the accelerator.


3.4 FEL Layout

The schematic layout of the FEL systems is shownin Figure 3.1. The proposed research topics requireinteraction with seed lasers in a modulator undulatorthen amplification in radiator undulators.

3.4.1 Modulator Section

Our initial studies have indicated that sufficientenergy modulation can be obtained at 50 µm usingonly four undulator periods, assuming the pulseenergy of the seed laser is 10–20 µJ. In order tomaintain a resonant interaction between a 250 MeVelectron beam and seed lasers covering the range120 µm to 800 nm the undulator period shouldbe around 200 mm with minimum gap 20 mm toallow the longest wavelength seed to be focussedat the centre of the modulator with an optimumRayleigh length equal to half the modulator lengthand negligible diffraction loss on the internal vacuumvessel aperture. Hence the minimum modulatorlength is 1.0 m (this allows an extra period at eachend for suitable termination of the field). Furtherstudy may indicate that at shorter wavelengths itwould be useful and practical to utilise more periodsin the modulator—in this case the design can berevised accordingly. A practical economic solutionis to use an existing multipole wiggler which waspreviously installed on the Synchrotron RadiationSource (SRS) at Daresbury. This has length 1 m,period 200 mm and minimum gap 20 mm, so isquite well matched to CLARA requirements. For thepresent layout we assume the use of this device andintend to study its performance in this role in moredetail. For schemes which require an additionalmodulator, for example EEHG which is discussedin Section 3.5.3, the first radiator can be replacedby a short modulator and chicane. We intendthe engineering design to allow such configurationchanges to be made quickly and conveniently.

The seeds will be injected onto the FEL axisvia an insertable mirror in the dog-leg immediatelyupstream of the modulator. The seed opticaltransfer lines will have the flexibility to focus theseed to a Rayleigh length of 0.35–1.25 m witha variable waist position. The optimum Rayleighlength, for strongest interaction between seedand electron beam, is approximately one half themodulator undulator length, so this flexibility allowsoptimum use of modulator undulators of length0.7–2.5 m.

In some of the schemes proposed it is necessaryto convert the imposed energy modulation into adensity modulation. This will be done using a4-dipole chicane immediately downstream of themodulator. The parameters for this chicane aredependent on the maximum required longitudinaldispersion (or R56). We anticipate that the EEHGscheme [43, 44] will place the strongest demand onthis chicane. A thorough study of EEHG applied toCLARA (Section 3.5.3) has shown the maximumR56 will be 40 mm, and this can be achieved witha four-dipole chicane of magnet and drift lengths of200 mm and peak field 0.7 T. The total footprintof this chicane (including magnet coil overhangs)is thus 1.5 m. Beyond the chicane are fourquadrupoles for matching the Twiss parameters intothe radiator section.

3.4.2 Radiator Section

The radiator section comprises seven 1.5 m longradiator undulators of planar hybrid design with thearrays aligned vertically (so that the magnetic fieldon-axis is horizontal). In this way the polarisation ofthe FEL output is vertical allowing much improvedtransmission using horizontal reflections to thephoton diagnostic beamline. The undulator modulesare arranged within a Focus-Drift-Defocus-Drift(FODO) lattice, with a single quadrupole betweeneach module. The gaps between undulator modules

RAD

IATO

RS

MO

DU

LATOR

MATC

HIN

G Q

UA

DS

CH

ICA

NE

DO

G-L

EG

AFTERBU

RNERS

LASER 5 m

Figure 3.1: FEL Layout


P [W

]109

108

107

106

105

104

103

102

z [m]151050

Gaussian bunchStart-to-end bunch

P [W

]

4 x108

3 x108

2 x108

1 x108

0

s [µm]140120100806040200

Figure 3.2: Results of 100 nm SASE simulations. The left hand plot shows the evolution of power alongthe undulator lattice for the s2e and gaussian bunches, and the right hand plot shows the photon pulselongitudinal profiles for the s2e bunch after 7 undulator modules.

are 1.1 m to allow sufficient space for diagnosticand vacuum components and a compact 0.3 melectron beam delay/phase shifter with a maximumdelay capability of 50 µm. The period of theFODO lattice is therefore 5.2 m. A substantialstudy has been made to optimise the FODOperiod. The conclusion is that ∼5 m is thelongest period that allows periodically matchedTwiss parameters over the full wavelength rangeof the FEL while maintaining high electron beamenergy. For longer FODO periods it becomesnecessary to reduce the electron beam energy tomaintain a matched solution for longer wavelengths,which means that the undulator parameter mustalso be reduced to maintain FEL resonance. Theseeffects have a doubly detrimental effect on theFEL performance—the efficiency of the FEL atconverting electron beam power to radiation poweris decreased because of the reduced K and the beampower itself is reduced because of the lower energy.

3.4.3 Afterburner

Beyond the final radiator we have allocated aspace of 5 m for the installation of afterburnerundulators. These can be used for generation ofshorter wavelengths than the FEL resonance, toinvestigate novel methods for polarisation control,or as part of an exotic scheme for short pulsegeneration. Possible applications are discussed inSection 3.5.4. Allocation of this dedicated spacewill also allow CLARA to be an excellent facilityfor testing advanced undulator designs, as discussedfurther in Section 5.7.

3.5 FEL Schemes

In this section we present further details of anumber of FEL research topics that could be studiedusing CLARA. We include simulation results ofthe schemes applied to CLARA to illustrate theflexibility of the accelerator and FEL systems andgive indications of the expected FEL output. Theschemes presented cover the full range of theFEL landscape as presented in Chapter 1. Ofcourse unknown schemes lie beyond the horizonand we intend to maintain the long-term flexibilityof CLARA to enable rapid study of future ideasproposed by the FEL community.

3.5.1 SASE

This is the base scheme, which we must be able todemonstrate in order to show the improvement overSASE available from the more advanced schemes.This scheme has been simulated, with the FELoperating at its shortest design wavelength of100 nm, using the well-benchmarked standard FELcode GENESIS 1.3. Electron bunch distributionsfrom the simulation codes used to model theaccelerator and beam transport have been importedinto the FEL simulation. The simulation is thereforea complete ’start-to-end’ (s2e) simulation.

The electron bunch properties are listed forthe SASE mode in Table 3.1. This bunchis non-uniform, having developed some structurethrough the processes used to generate andtransport it, so a comparative study has been madeof the FEL performance using an ideal bunch witha Gaussian current profile and uniform emittance


−3 −2 −1 0 1 2 3

445.5

446

446.5

447γ

time (ps)−0.5 0 0.5

0

10

20

30

40

50

60

70

80

90

time (ps)

pow

er (

MW

)

average

Figure 3.3: (Left) Energy modulation given to the electron beam by a 50 µm, 10 µJ, 500 fs seed laser.(Right) FEL power at saturation for 10 shot-noise seeds.

and energy spread along the bunch (with thesevalues set to match that of the s2e bunch at theposition of maximum peak current). The results ofthe simulations for these two electron bunches areshown in Figure 3.2. The left hand plot shows theevolution of power along the undulator lattice forthe s2e and gaussian bunches, and the right handplot shows the photon pulse longitudinal profilesfor the s2e bunch after 7 undulator modules. Itis seen that both the s2e bunch and gaussian bunchreach FEL saturation before the end of the 7thundulator module at a distance of 16.5 m fromthe start of the radiator section. These resultsdemonstrate that the electron bunch properties andundulator lattice design are sufficiently good. Manyof the FEL schemes to be studied on CLARA willreach saturation earlier in the radiator section, eitherbecause they will be tested at longer wavelengthswhere the gain length is shorter, or because theelectron bunch will enter the radiator pre-buncheddue to manipulations in the modulator section andchicane.

3.5.2 Generation of Short Pulses

The generation of short pulses is one of theimportant research themes for CLARA. Our aimis to generate pulses with as few optical cyclesas possible with durations of the order of, orshorter than, the FEL cooperation length lc . Thecooperation length is the slippage in a gain lengthand therefore an intrinsic scale length for the FELprocess. It is therefore difficult to generate highintensity FEL pulses of duration less than lc . The

SASE process generates spikes of duration ' πlcwith each SASE spike developing independentlyfrom the shot noise in the electron beam. Manyof the short pulse schemes aim to slice out, orisolate, a single SASE spike and the duration ofthe output pulse is therefore typically no shorterthan πlc . One gain length in a FEL is typicallyone hundred undulator periods, so the cooperationlength is therefore typically 100 wavelengths andpulse durations possible from slicing schemes areof this order. In the first two parts of this sectionwe discuss two schemes with this level of potential:Slicing and Single-Spike SASE. To generate shorterpulses than ' πlc requires fundamentally alteringthe FEL process. Two such schemes, Mode-Lockingand the Mode-Locked Afterburner are discussedlater in this section. For each scheme in thissection we indicate the predicted pulse durations,and lengths in µm, number of optical cycles andnumber of cooperation lengths. Of course otherschemes will be proposed in the future and theflexibility of CLARA will make us well placed toinvestigate these schemes in due course.

Slicing Schemes

One promising scheme for ultra-short pulsegeneration is the energy chirp plus taperedundulator scheme [13] and we have studied theimplementation on CLARA in some detail [45]following earlier assessment for implementation onthe UK’s NLS proposal [46]. In this scheme, afew-cycle seed laser is used to modulate the electronbeam energy to an amplitude greater than the


Figure 3.4: 100 nm Single Spike SASE Pulses: Power and Spectrum

natural bandwidth of the FEL. By tapering the gapof the undulator, only the sections of the electronbunch where the energy chirp is correctly matchedto the undulator taper will experience high FEL gain.The final FEL pulse therefore consists of a train ofindividual SASE spikes, each separated by the seedlaser period. The number of spikes can be controlledby varying the number of cycles in the seed laserand periods in the modulator undulator—by usinga single-cycle laser and a single period modulator asolitary spike can be generated which would be theultimate aim for implementing this scheme at shortwavelengths.

For CLARA it is not feasible to provide a singlecycle seed of appropriately long wavelength andpulse energy but we can demonstrate the physicsof the mechanism by generating a train of spikes.The scheme has been studied using a combination ofelegant [47] and GENESIS 1.3 assuming Gaussianelectron bunch distributions with parameters asgiven in Table 3.1.

Simulations were carried out for both 400 nmand 266 nm FEL wavelengths. For the 400 nmcase, a seed laser of 50 µm wavelength looksoptimal, with a 500 fs/10 µJ pulse providingsufficient energy modulation to restrict the growthof the SASE radiation spikes to the energy-chirpedregions. For the 266 nm case, a similar seedlaser pulse energy would be required, but thewavelength should be reduced to 40 µm in orderto give a good match between the length of theenergy chirps in the electron bunch and the widthof the SASE radiation spike (' πlc). Assumingthese conditions can be met, numerical simulationssuggest FEL pulses of the order 50 MW peakpower and 50 fs/15 µm/55 cycles FWHM can be

expected.Figure 3.3 shows the calculated energy

modulation given to the electron bunch, alongwith the FEL pulse profile at saturation for 10different shot-noise seeds. The taper in radiatorgap was applied in steps, with constant gap foreach undulator module.

The EEHG scheme discussed later can alsobe used to control the length of the FEL pulseby varying the length of the second seed laser,keeping the electron energy modulation amplitudefixed. Studies of this option using a 40 fs FWHMlaser indicate the scheme will work just as well aswhen using the 500 fs seed, generating a temporallycoherent FEL pulse of >100 MW peak power and25-40 fs FWHM duration at 100 nm.

Single Spike SASE

The idea of the single-spike SASE scheme is simple.In a SASE FEL the spacing between the SASEspikes is ∆s ≤ 2πlc . If the electron bunch Lb > 2πlcthen the number of SASE spikes N in the FELoutput will be N = Lb/2πlc [48]. If however theelectron bunch is of length Lb ' 2πlc then only oneSASE spike can develop, hence the term single-spikeSASE. For Lb < 2πlc then again only one SASEspike will develop, but the saturation power will bereduced because the radiation will slip out of thefront of the electron bunch before it can be fullyamplified.

We have investigated this scheme using anelectron bunch compressed via the velocity bunchingscheme and an initial simulation result using atracked bunch imported into GENESIS 1.3 is shownin Figure 3.4. The FWHM pulse length obtained at100 nm is 23 fs/7 µm/70 cycles.


600 800 10000

1

2

3

4

5

6

s ( µm)

P (

MW

)

260 270 2800

0.2

0.4

0.6

0.8

1

P(λ

)(a

.u.)

λr (nm)

200 250 300 350 4000

20

40

60

80

100

120

s ( µm)

P (

MW

)

96 98 100 102 1040

0.2

0.4

0.6

0.8

P(λ

)(a

.u.)

λr (nm)

Figure 3.5: Mode-locked output pulses and spectra using the standard CLARA configuration, at 266 nm(top) and 100 nm (bottom)

Mode-Locking

The Mode-Locking scheme [15, 49, 50] requires theuse of delay sections between the radiator undulatormodules to periodically delay the electron bunchwith respect to the radiation. This generates aset of sideband modes in the radiation spectrumwhich correspond in the time domain to a strongmodulation in the radiation pulse envelope with aperiod equal to the slippage in one undulator moduleplus the added delay. If a periodic modulationis added to the electron bunch current or energy,with a frequency equal to the spacing betweenthe sideband modes, then these modes developtheir own sidebands which overlap and phase lockwith the neighbouring modes. The process isanalogous to mode-locking in a conventional laser.

In the time domain the output then comprises atrain of evenly spaced phase-locked spikes, witha duration potentially far shorter than the FELcooperation length. In the frequency domain theradiation modes extend out to the full bandwidthof a single undulator module, which is far greaterthan the bandwidth of a normal SASE FEL.The mode-locking technique therefore has twoapplications of interest—generation of ultra-shortpulses and generation of simultaneous multi-colourFEL output with a wide and variable frequencyseparation.

The scheme has been studied forimplementation in two phases. In the first phase thescheme has been simulated for CLARA parameters,using the standard undulator lattice, with the delaysimparted via the phase-shifters between undulator


600 800 1000 12000

5

10

15

20

s (µm)

P (

MW

)

255 260 265 270 275 2800

0.2

0.4

0.6

0.8

1

P(λ

)(a.

u.)

λr (nm)

300 350 400 4500

2

4

6

8

s ( µm)

P (

MW

)

96 98 100 102 1040

0.2

0.4

0.6

0.8

1P

(λ

)(a

.u.)

λr (nm)

Figure 3.6: Mode-locked output pulses in the second phase of the research. Output at 266 nm withextra delays within standard undulators (top) and single-spike mode-locked output at 100 nm with 120 µmmodulation (bottom).

modules. We would aim to demonstrate initially at266 nm, where the single shot temporal diagnosticsare available and we would also characterise theoutput spectrally. We would then reduce thewavelength to 100 nm and rely on the spectraldiagnostics only, looking for equivalent spectralprofiles at 266 nm and 100 nm to prove thescheme. The simulation results at 266 nm and100 nm are shown in Figure 3.5. A 50 µm energymodulation has been added to the beam, and thedelays are set so that total slippage in undulatormodule and delay is also 50 µm. The resultsindicate that in this phase the scheme can beproven in principle. For the results shown here, theFWHM pulse lengths (of individual pulses within thetrain) are 43 fs/13 µm/50 cycles at 266 nm and18 fs/5 µm/50 cycles at 100 nm. For comparisonπlc ' 20µm at 266 nm and πlc ' 10µm at

100 nm. The spectra show that the output is alsodiscretely multichromatic over a full 7% bandwidthwhich is far broader than the SASE bandwidth.

In the second phase the aim would be toreduce pulse durations further to below theFEL cooperation length, using additional delayswhich would be inserted within the undulatormodules—from the scalings given in [15] this wouldenable a broader gain bandwidth supporting moremodes and thus the synthesis of even shorter pulseswithin the train. Simulation results of this setup,operating at 266 nm, are shown in Figure 3.6. Thepulse length obtained is 17 fs/5 µm/20 cycles, morethan a factor of two shorter than possible with thestandard configuration.

Another interesting possibility, also shown inFigure 3.6, is an extension scheme where it mightbe possible to generate isolated pulses. The


0 20 40 60 80 100 120 140 160 180 2000

5

10

15

20

t [fs]

P [M

W]

90 100 1100

2

4x 10

4

λ [nm]

P(λ)

[a.

u.]

3

Figure 3.7: Simulation results of the Mode-Locked Afterburner operating at 100 nm

modulation wavelength is longer here, at 120 µm.The motivation for this is by obtaining the correctratio between electron bunch length and modulationperiod it is possible to preferentially amplify only asingle pulse within the train. The pulse durationobtained here is 14 fs/4 µm/40 cycles. Furtherstudy is required of this concept but the hope isto eventually demonstrate the production of isolatedmode-locked pulses. It may in fact be possible to dothis with a 50 µm beam modulation if the electronbunch is compressed to the appropriate length.

Mode-Locked Afterburner

The Mode-Locked Afterburner scheme [16] is adevelopment of the mode-locked amplifier FELconcept and promises to deliver ultra-short pulses.It retains the baseline radiator stage of the CLARAFEL, with short pulses generated in a relativelyshort ‘afterburner’ comprising several few-periodundulators separated by chicanes. For CLARAthe electron bunch would be modulated withperiod 3 µm using an Optical Parametric Amplifier(OPA) driven by the Ti:sapphire laser. Simulationresults using GENESIS 1.3 are shown in Figure 3.7using the following afterburner parameters: eachundulator module has 8 periods and the chicanescomprise four 2.9 cm length, 0.25 T dipoles.FWHM pulse durations of 1.6 fs/0.5 µm/5 cyclesare predicted, with peak power reaching ' 20 MWin 10 undulator-chicane modules, so the totalafterburner length is just over 4 m. The spectrumshows clearly separated distinct wavelengths, over abroad bandwidth of ∼13%.

Summary of Pulse Durations

For each scheme in this section the predictedpulse durations are summarised in Table 3.2 inunits of fs, µm, number of optical cycles and

number of cooperation lengths. For referencethe cooperation length lc ' 7 µm at 266 nmand lc ' 3µm at 100 nm. It is seenthat the Slicing/Taper, EEHG and Single SpikeSASE schemes would produce pulses of around55–75 cycles FWHM or just over two cooperationlengths. The Phase I Mode-Locking scheme wouldgenerate pulses slightly shorter (but in a train) ofaround 50 cycles or just less than 2 cooperationlengths. With the Mode-Locking scheme upgradedin Phase II pulses would come down to around20 cycles (in the train) or less than one cooperationlength, or for the case where an isolated pulse couldbe produced about double this number of cycles.Finally, the Mode-Locked Afterburner may producepulses of only 5 cycles, nearly an order of magnitudeless than the cooperation length.

3.5.3 Improving Temporal Coherence

Currently the full potential of X-ray FELs is notrealised because they operate in SASE mode forwhich the temporal coherence is relatively poor.For this reason, their spectral brightness is typicallytwo orders of magnitude lower than that thatof a transform limited source. Improvementof SASE FEL temporal coherence would greatlyenhance scientific reach and allow access to newexperimental regimes. There are a number ofmethods for improving SASE coherence, many ofwhich have been tested or are already in routineuse. Existing methods fall into two classes. In thefirst class, an externally injected seed source of goodtemporal coherence ‘seeds’ the FEL interaction sothat noise effects are reduced. This seed fieldmay be either at the resonant radiation wavelength,where available, or at a subharmonic which isthen up-converted within the FEL. These methods,which include HGHG [51, 52, 53, 54] and EEHG [43,


FWHM Pulse Duration

Scheme Pulse Type Wavelength (nm) fs µm #cycles #lc

Slice/Taper Single 266 50 15 56 2.2EEHG Single 100 25 8 75 2.6Single-Spike SASE Single 100 23 7 70 2.3

Mode-Locking Phase ITrain 266 43 13 49 1.9Train 100 18 5.3 50 1.8

Mode-Locking Phase IITrain 266 17 5.1 20 0.7Single 100 14 4.1 41 1.4

Mode-Locked Afterburner Train 100 1.6 0.5 5 0.16

Table 3.2: Predicted pulse durations for CLARA Short Pulse Schemes.

44], rely on a synchronised external seed at theappropriate wavelength, pulse energy and repetitionrate. In the second class, the coherence is createdby optical manipulation of the FEL radiation itself,for example by spectrally filtering the SASE emissionat an early stage for subsequent re-amplificationto saturation in a self-seeding method [55, 56,57, 58], or via the use of an optical cavity [33,34, 35, 36, 37, 38, 39, 40]. Methods in thisclass rely on potentially complex material-dependentoptical systems which limit the ease and range ofwavelength tuning. If an optical cavity is used,the electron source repetition rate should also be inthe MHz regime to enable a practical cavity length.A third, more recently proposed class of methods,rely on artificially increasing the slippage betweenFEL radiation and electron bunch to slow downthe electrons which extends the coherence length[59, 60, 61] or even completely ’delocalises’ the FELinteraction allowing the radiation coherence lengthto grow exponentially [62].

Schemes in all three classes can be optimised,validated or even demonstrated for the first timeon CLARA, as discussed in the remainder of thissection.

Seeded Harmonic Cascade

When using injected seed sources to improve FELtemporal coherence, an important aspect is toextend the seeding effect beyond the wavelengthrange of available seed sources towards the muchshorter wavelengths achievable with FELs. Methodsto do this involve operating the FEL in stages, withharmonic up-conversion to shorter wavelength at

each stage. The use of multiple such stages istermed a ‘seeded harmonic cascade’, and is a keyunderlying technique of the recently commissionedFERMI@Elettra facility [6] and also for the proposedUK NLS project [42, 63].

There are several variants on the technique.Here we investigate the one proposed for theshortest wavelength of the UK NLS project whichinvolves two harmonic steps. To study this onCLARA we would seed with the 800 nm sourcein the first stage before conversion to the secondharmonic (400 nm) in the second stage, followed bya further fourth harmonic conversion for 100 nmFEL operation. The seed imprints a sinusoidalenergy modulation on the electron beam within themodulator undulator, then chicanes convert thisinto a density modulation (or ‘bunching’) containinghigher harmonic components. The induced energyspread must be sufficiently low to allow FEL lasingin the final stage.

A two stage harmonic cascade could bedemonstrated with the baseline CLARA layout byusing the first radiator undulator as a modulator.GENESIS 1.3 simulations were carried out and theresults are shown in Figure 3.8. An 800 nmseed with 200 fs FWHM duration and 0.5 MWpeak power was used in the modulator to applyan energy modulation ∆γ/σγ ≈ 1. The chicanewas set for R56 ≈ 200 µm, giving a bunchingfactor of ∼2% at 400 nm. This is deliberatelyless than optimum—the first radiator stage usedas a modulator resonant at 400 nm is relativelylong, so the relatively low bunching factor acts tokeep the induced energy spread sufficiently low inthis stage. Energy modulation of ∆γ/σγ ≈ 6 was


Figure 3.8: Simulation results for the seeded harmonic cascade scheme, showing peak power along theradiator compared to an equivalent SASE case (top), and the pulse properties at saturation (bottom).

applied, giving ∼20% bunching factor at 100 nm,using the inter-module electron delay chicanes toapply a relatively low compression (R56 ≈ 60 µm).The remainder of the radiator is resonant at 100 nmand the FEL saturates after only two undulatormodules with excellent temporal coherence andcontrast ratio over the SASE background, as shownin Figure 3.8.

Saturating so early in the radiator opens up thepossibility of investigating other FEL methods suchas undulator tapering [64] to enhance the outputpower, or superradiance [65, 66, 67], in whichsub-cooperation length pulses can be generated asthe FEL interaction proceeds deep into saturation.

Echo-Enabled Harmonic Generation

The EEHG scheme has been proposed as a way toimprove the temporal coherence of FEL pulses, andworks by combining two energy modulation stageswith two chicanes to induce a fine-structure densitymodulation in the electron bunch [43]. The electronbunch is then sent through a radiator sectionresonant at the density modulation wavelength,ultimately generating a temporally coherent FELpulse at a wavelength many times shorter than theinitial seed lasers. Initial studies of this scheme [68]assumed both modulators were identical, with65 mm period and 1 m total length but forimplementation on CLARA the first radiator would


0 50 100 150 2000

0.05

0.1

0.15

0.2

0.25

0.3

time (fs)

init

ial b

un

chin

g

0 50 100 150 2000

100

200

300

400

time (fs)

po

we

r a

t sa

tura

tio

n (

MW

)

97 98 99 100 101 102 1030

1

2

3

4

5x 10

11

wavelength (nm)

po

we

r a

t sa

tura

tio

n (

arb

. un

its)

0 50 100 150 200

−3

−2

−1

0

1

2

3

time (fs)

ph

ase

at

satu

rati

on

(ra

d)

SASE (0m)

EEHG (0m)

SASE (13.5m)

EEHG (6.5m)

SASE (13.5m)

EEHG (6.5m)

SASE (13.5m)

EEHG (6.5m)

Figure 3.9: EEHG: Comparing the final FEL pulse to a standard SASE simulation, the FEL pulse generatedusing EEHG shows a more uniform peak power at saturation, reduced bandwidth and constant radiationphase, combined with a substantial reduction in saturation length.

be replaced by an additional short modulator andweak chicane. We expect performance to be similarin this configuration.

The optimal seed laser wavelength forinvestigating EEHG on CLARA is 800 nm. Basedon the analytic equations derived in [44] and theabove modulator parameters, maximum bunchingat wavelengths from 100 nm down to 8 nm (8th to100th harmonics of the seed lasers) can be achievedusing seed lasers of up to 100 µJ pulse energywith 500 fs rms pulse duration. The maximumrequirements placed on the two chicanes would beR56 values of 41 mm and 0.5 mm respectively.

To complement the analytical studies, numericalsimulations have been carried out at λFEL =100nm (8th harmonic) using elegant for the twomodulator and chicane sections, and GENESIS 1.3for the radiator section. The tracking in elegantwas carried out with and without coherent andincoherent synchrotron radiation emission in thechicane dipoles. In the simulations, a 250 fs longsection of the electron bunch was studied, withsufficient particles to account for the individualelectrons (>590M particles). The seed laser pulseenergy was 10 µJ. Comparing the final FEL pulseto a standard SASE simulation, the FEL pulsegenerated using EEHG shows a more uniform

peak power at saturation, reduced bandwidthand constant radiation phase, combined with asubstantial reduction in saturation length, as shownin Figure 3.9.

For the study of coherent bunching at veryhigh harmonics, i.e. at wavelengths shorter than100 nm where it will not be possible to demonstratelasing, methods for generating coherently enhancedradiation from the electron beam may be used as adiagnostic. This is a subject for further study.

RAFEL

Simulations have been done of a RAFEL drivenby CLARA operating in multibunch mode, withelectron bunch parameters given in Table 3.1.The one-dimensional FEL oscillator simulation codeFELO [69] was employed. Although this codedoes not model the radiation propagation inthe cavity, employing instead a simple radiationfeedback factor, its predictions of the longitudinalradiation characteristics and build up from shotnoise have been shown previously to agree wellwith fully three-dimensional simulations of RAFELsystems using GENESIS 1.3 and dedicated opticalpropagation codes [70].

Only five of the seven radiator undulatorsare required. The radiation feedback factor F


600 800 1000 12000

5

10

15

t (fs)

P (

MW

)

600 800 1000 1200−4

−2

0

2

4

φ

t (fs)

99 99.5 100 100.5 1010

0.2

0.4

0.6

0.8

P(λ

)(a

.u.)

λr (nm)

600 800 1000 12000

5

10

15

t (fs)

P (

MW

)

600 800 1000 1200−4

−2

0

2

4

φ

t (fs)

99 99.5 100 100.5 101

0

0.2

0.4

0.6

0.8

P(λ

)(a

.u.)

λr (nm)

Figure 3.10: Results of 100 nm RAFEL simulations. The left hand column shows the control case withoutan optical cavity, i.e. normal SASE, whereas the right hand column shows the effect of adding thefeedback—the output pulse is cleaned up both spectrally and temporally. The pulse is shown after 11cavity round trips.

to maximise the output power and longitudinalcoherence [38] is given by F = 25 exp(−

√3G)

where G = 4πρNw with ρ the usual FEL parameterand Nw the number of undulator periods. For theelectron bunch parameters as listed in Table 3.1 itis found that F = 1.8 × 10−3 showing that only avery small fraction of the emitted radiation needsto be fed back via the optical cavity to the start ofthe undulator. The scheme is in effect a self-seededFEL rather than an oscillator FEL because there isno requirement to build up a stable optical mode inthe resonator.

The simulation results are shown in Figure 3.10where the RAFEL has reached saturation after11 cavity round trips. The left hand plot showsthe control case for these parameters without an

optical cavity, i.e. normal SASE, whereas theright hand plot shows the effect of adding thefeedback—the output pulse is cleaned up bothspectrally and temporally with time bandwidthproduct (∆λ/λ2)c∆t = 0.58 which is close to thatof a transform limited pulse. Simulations for 266 nmand 400 nm output also show near transformlimited output using optimum feedback factors ofF = 4.5× 10−6 and F = 3.5× 10−7 respectively.

The emphasis here has been the use of aRAFEL to improve SASE temporal coherence, butit should be noted that theoretical work has alsoindicated the potential of RAFEL systems for thegeneration of stable attosecond pulses [71]. Thisis another research topic that could be investigatedexperimentally on CLARA.


Figure 3.11: HB-SASE: the spectra (in scaled frequency ω = (λ−λr )/2ρλr ) of four statistically independent100 nm SASE pulses on the top row, and the spectra of four statistically independent HB-SASE pulses onthe bottom row.

High-Brightness SASE Schemes

A recently proposed class of methods rely onartificially increasing the slippage between FELradiation and electron bunch to enhance the SASEFEL longitudinal coherence and hence improvethe spectral brightness. This can be doneby using chicanes to apply a series of unequaldelays to the electron bunch as it propagatesthough the undulator [59]. Variations of thescheme have been studied in some detail andnamed High-Brightness SASE (HB-SASE) [62] andImproved SASE (iSASE) [72]. For CLARA thedelays can be applied via magnetic chicanes betweenundulator sections, using the same hardware asrequired for the Mode-Locking scheme discussedin Section 3.5.2. A proof-of-principle experimentto demonstrate the concept has already been donesuccessfully at the LCLS using detuned undulatorsections as delays, but this only allowed a limitedslippage enhancement [73]. Another proposalcalled purified SASE (p-SASE) [61] suggestsusing subharmonic undulators as ‘slippage-boosted’sections. Theoretical work on HB-SASE and iSASEhas shown that by using magnetic chicanes to delaythe beam and optimising the delay sequence itmay be possible to significantly extend the radiationcoherence length. Further advantages are that theefficiency of a taper can be much enhanced and thatthe shot-to-shot stability of the output power in thepresence of electron beam energy fluctuation shouldbe improved compared to a self-seeding schemeusing optics [60].

The scheme has been studied forimplementation on the standard CLARA lattice,with delays based on a prime number sequence

[74]. The results indicate we could generatetransform limited pulses at 100 nm. Figure 3.11shows the spectra of four statistically independentSASE pulses on the top row, and the spectra offour statistically independent HB-SASE pulses onthe bottom row. The particular areas of researchthat could be done on CLARA are: study andoptimisation of the delay sequence; statisticalstudies of output stability; tolerance studies ofsensitivity to accelerator jitter; enhancing taperingefficiency. A later phase of research (whichcould run parallel to the extended research onmode-locking) would investigate: impact on FELoutput of extra in-undulator delays; demonstrationof novel compact isochronous delays [75]; use ofmixed non-isochronous and isochronous delays;study of a ‘correction chicane’ with negativeR56 [76].

3.5.4 Afterburner Schemes

A number of schemes which use afterburnerundulators could be tested:

Short Wavelength Generation This is done byexploiting the electron beam bunching at harmonicswhich are higher than the FEL resonance. Thisharmonic bunching occurs naturally within the beamdue to the FEL interaction. By propagating thebeam into a short afterburner which is tuned to thisharmonic the beam can be made to emit a pulseof coherently enhanced radiation at a wavelengthshorter than the FEL wavelength. This burst ofradiation undergoes a sharp initial power growth(quadratic with distance through the afterburner)


which rapidly saturates—the energy spread inducedin the beam by the previous FEL interactionprohibits exponential growth and lasing, but theemitted power can still be orders of magnitudegreater than incoherent spontaneous undulatoremission. Such afterburner schemes are of interestto many FEL facilities as a way of extractinguseful short wavelength radiation from an otherwise’spent’ beam. The research interest is in the use ofcompact novel undulators for this purpose.

Exotic Short Pulse Generation A proof ofprinciple experiment could be done for the’mode-locked’ afterburner concept, which is anew proposal for generating FEL pulses in thezeptosecond regime from existing X-ray FELs viaa compact afterburner extension. This idea wasexplained further in Section 3.5.2.

Polarisation Control To generate circularlypolarised radiation via the crossed undulatorscheme. This requires two undulators asthe final FEL radiators with orthogonal linearpolarisation [77]. Alternatively a single variablypolarised undulator could be used as the finalradiator to demonstrate high intensity variablypolarised emission from an electron beam stronglypre-bunched in the previous planar undulators [63].

3.6 Scaling to ShortWavelengths

The CLARA wavelength range, in the visible andVUV, has been chosen to enable easy diagnosis ofthe output. For many schemes we are concernedonly with the underlying physics of the mechanismwhich is often wavelength-independent—thewavelength only becomes an issue when consideringthose factors specific to the scheme in questionwhich may mitigate performance at shorterwavelengths. However, scalability to shorterwavelengths is important to understand because theR&D done at CLARA will be applied at FELs whichoperate at shorter wavelengths than CLARA canaccess. This will benefit UK scientists whether theyare current users of short-wavelength FEL facilitiesabroad or future users of a short-wavelength FELfacility in the UK.

In general the parameter tolerances are moredemanding as the wavelength decreases. The

required normalised emittance is proportional tothe wavelength, the tolerances on energy spreadand field quality scale with the FEL ρ parameterwhich is smaller at shorter wavelengths, and thetolerances on electron beam trajectory control scalewith the size of the electron beam which reduces dueto adiabatic damping at the higher beam energiesrequired for shorter wavelengths. These issues mustbe addressed for the FEL to lase even in SASE modeso do not directly affect the scalability of CLARAresults—the issues we are concerned with are onlythose relating to the extra manipulation requiredto convert SASE output to short pulse output ortemporally coherent output.

For example, in many schemes one or moremagnetic chicanes are used, either to longitudinallyshear the electron bunch or delay it with respectto the radiation. Often some wavelength-specifictolerances on the stability of these chicanes havebeen determined in the original papers (see forexample [15]). Sometimes more specific issueshave been raised which need further study, suchas: the effect of R51 leakage smearing theimparted modulation in EEHG; the degradingeffect of Coherent Synchrotron Radiation (CSR)emission in the chicanes; the washing out ofFEL induced microbunching during beam transport;the discrepancy between the EEHG model, whichassumes each harmonic has a δ-function linewidth,and reality where the harmonics all have a widthwhich depends on laser pulse length, chirp, laserphase noise and electron beam shot noise. All ofthese effects will become relatively more significantfor the finer beam manipulations and structuresrequired at shorter wavelengths. Using CLARAwe will be uniquely placed to actively study theimpact of these effects by progressively changing theinput parameters and carefully assessing the FELperformance. For example we can add errors tomagnet angles, add jitter to magnet power supplies,adjust the laser heater settings to control the beammicrobunching, propagate bunched beams overvariable distances before diagnosing the bunchingdegradation via Coherent Optical TransmissionRadiation (COTR) screens, adjust the laser chirpand pulse duration, and so on. In this way wecan obtain invaluable data which can be usedto prioritise schemes for future implementation atX-ray wavelengths.

As a last comment, we consider how the pulsedurations for the CLARA short pulse schemes wouldscale if implemented on an X-ray FEL. For the


Slice/Taper scheme and the Single Spike SASEscheme the pulse duration would be expected toscale with the cooperation time tc = lc/c . Anestimate for tc at 0.15 nm wavelength is tc ' 80 as,so from Table 3.2 which gives the pulses lengthsin units of the cooperation length we would expectpulse durations of order 180 as at 0.15 nm. For theMode-Locked schemes the pulse durations would beexpected to scale with the number of cycles, so at

0.15 nm we would expect FWHM pulse durationsof 10–25 as for Mode-Locking, and around 2.5 asfor the Mode-Locked afterburner. Both of theseestimates are in agreement with the published 3Dsimulations of these schemes at this wavelength[15, 16]. What is clear then is that if the schemes wetest on CLARA are one day applied in the X-ray theymay enable a transformative change in the utility ofthe free-electron laser as a scientific tool.


4Accelerator Design

4.1 Layout Overview

Bu

nch

Co

mp

ress

or

X-b

an

d C

av

ity

Gu

n

70 - 150 MeV

Lin

ac

1

Lin

ac

2

Dia

gn

ost

ics

1

Lin

ac

3

Lin

ac

4

~250 MeV

Mo

du

lato

r

Ra

dia

tor

1

Ra

dia

tor

2

Ra

dia

tor

3

Ra

dia

tor

4

La

ser

He

ate

r

Ra

dia

tor

5

Ra

dia

tor

6

Ra

dia

tor

7

Do

gle

g

VELA

Ou

tpu

t D

iag

no

stic

s

Beam DumpPI Laser Seed Laser

Tra

ns.

De

f. C

av

ity

1

Dia

gn

ost

ics

2

Tra

ns.

De

f. C

av

ity

2

Sp

ect

rom

ete

r 1

Sp

ect

rom

ete

r 2

Aft

erb

urn

er

Ex

tra

cte

d B

ea

m

Figure 4.1: CLARA layout overview

The design approach adopted for CLARA is tobuild in flexibility of operation, enabling as wide anexploration of FEL schemes as possible. To thisend a range of possible accelerator configurationshave been considered. A major aim is to be able totest seeded FEL schemes. This places a stringentrequirement on the longitudinal properties of theelectron bunches, namely that the slice parametersshould be nearly constant for a large proportion ofthe bunch length. In addition, CLARA should havethe ability to deliver high peak current bunches forSASE operation and ultra-short pulse generationschemes, such as velocity compressed bunches.This flexibility of delivering tailored pulse profileswill allow a direct comparison of FEL schemes inone facility.

An overview of the proposed layout of CLARAis shown in Figure 4.1. The S-band (2998.5 MHz)RF photocathode gun [78] is followed by Linac 1.

This is a short (∼2 m long) structure, chosensuch that it may be used in acceleration orbunching configurations. A spectrometer line whichalso serves as injection to VELA branches atthis location. Linac 2 follows which is ∼4 mlong and capable of accelerating up to 150 MeV.Space for a laser heater is reserved at this point.Initially this will not be installed however weexpect that in the ultra-short bunch mode thebeam properties will be degraded by microbunchinginstability (predominantly driven by longitudinalspace-charge impedance). This effect will bequantified and the case for installing the laser heaterdetermined in the future.

A fourth harmonic linearising X-band cavity(11994 MHz) [79] is situated before the magneticcompressor to correct for longitudinal phasespace curvature. A variable magnetic bunchcompressor is then followed by the first dedicated


Figure 4.2: (1) Laminarity (black/red) in the 10% of charge slice containing the peak current withcompression at 70 MeV/130 MeV (for an indicative layout). (2) Longitudinal phase space (black/red)with compression at 70 MeV/130 MeV.

beam diagnostics section, incorporating transversedeflecting cavity and spectrometer, enablingmeasurement of emittance, bunch length and sliceproperties. Linacs 3 & 4 (each ∼ 4 m long)accelerate to 250 MeV. These are followed bya second diagnostics section. It has also beenproposed to divert this high energy beam for otherapplications. The beamline then passes a dogleg,offsetting the FELs from the linacs transverselyby ∼ 50 mm to enable co-propagation of longwavelength laser seeds. Immediately following thedogleg is the FEL modulator undulator and chicane.A dedicated matching section ensures that periodicoptics is achievable in the radiators for the entirewavelength range. Seven FEL radiators and a spacefor a FEL afterburner complete the accelerator andthe beam is then dumped.

4.1.1 Phase Space Linearisation

Magnetic compression of electron bunches requiresthat some attention be given to the removal of thecurvature imposed on the longitudinal phase spaceof the bunch by the accelerating RF. For CLARA,a harmonic linearising cavity was compared withnon-linear magnetic correction in the chicane [80].The additional complication of a harmonic cavitywas shown to be justified by the ability to predictablytailor the longitudinal phase space.

4.1.2 Energy at Magnetic Compressor

The seeded FEL schemes to be demonstratedat CLARA require small correlated energy spread

at the undulators, therefore when magneticcompression is to be used the compressor mustbe situated at substantially less than full energy.This ensures that the chirp needed at compressioncan be adiabatically damped or suppressed throughrunning subsequent accelerating structures beyondcrest. This requirement must be balanced againstthe fact that compressing at low energy exacerbatesspace-charge effects. To quantify this we use thelaminarity parameter

ρL =

(I/(2IA)

εthγγ′√

1/4 + Ω2

)2

where I is the current in the slice underconsideration, IA is the Alfven current, εth isthe thermal emittance, γ

′= dγ/ds and Ω

is a solenoidal focusing field (zero in the caseconsidered).

When this parameter is greater than 1, weshould consider space-charge effects in the bunchevolution. To inform this we select two candidateconfigurations, one with magnetic compression at70 MeV and one at 130 MeV. An indicativebunch was tracked through both configurations,setting the machine parameters to produce a zerochirp 250 pC bunch of peak current 350 A at250 MeV. Tracking was carried out with ASTRA [81,82] to the exit of the first linac section toinclude space-charge, followed by elegant [47, 80]taking into account the effect of cavity wakefields,longitudinal space-charge (LSC) and CSR emittancedilution. Figure 4.2 shows the resultant laminaritiesand final bunch longitudinal phase spaces.


Value Unit

Energy at compressor 70 - 150 MeVMin. : Max. bend 0 : 200 mradBend magnetic length 200 mmMax. bend field 0.5 TMin. : Max. offset 0 : 300 mmSeparation of dipoles 1 & 2 1500 mmSeparation of dipoles 2 & 3 1000 mmMax. bellows extension 260 mmMin. : Max. R56 0 : -72 mmBeam size from δE (±6σ) 0 : 20 mmBeam size from βx (±6σ) 3.0 mm

Table 4.1: Specification of variable bunch compressor.

We see that in both cases space-charge shouldbe considered. However compression at 130 MeVdoes not allow us to subsequently de-chirp thebunch (for the same energy gain). A dedicatedde-chirping cavity at full energy has been consideredand rejected due to technical immaturity and lack ofspace. Note that we go no further than 30 beyondcrest in the final accelerating structures in order toavoid large jitter effects. As we wish the facility tobe flexible we select a nominal compressor energyof approximately 70 MeV, but we achieve this byreducing the gradient in the accelerating structurebefore the compressor (and increasing the gradientrequired in Linacs 3 and 4 to 23 MV/m). This givesus the option of compressing at higher energy inregimes where a de-chirped bunch is not required.

The elegant LSC model is used in all trackingand the achieved compression will be compared withother codes in further studies allowing differencesbetween 1D and full 3D space charge effects.

4.1.3 Variable Bunch Compressor

With the above considerations in mind thespecifications for the CLARA variable bunchcompressor are shown in Table 4.1.

For flexibility, the compressor has a continuouslyvariable R56 and is rated for maximum energyof 150 MeV. Physical movement of the elementstransversely allow for a small aperture beam pipe inthe central section. This enables energy feedbackon Linacs 1, 2 and the X-band lineariser via a highresolution cavity Beam Position Monitor (BPM)situated in the high dispersion region. A screen andcollimator are also situated in this section. Theability to set a straight-through path also allows

the use of velocity compressed bunches. Opticsare kept such that βx is minimised at the lastdipole, thereby minimising CSR. Residual dispersionarising from non-identical compressor dipoles willthen be corrected using trim coils informed by recentexperience at FERMI@Elettra and LCLS.

4.1.4 Diagnostic Sections

Dedicated beam diagnostics sections are situatedafter the bunch compressor at 70 − 150 MeV, andagain after Linac 4 at 250 MeV. The transverse andlongitudinal properties of the compressed bunch willbe analysed immediately post compression in thefirst section, and those at full energy in the secondsection.

Specifically we intend to measure the Twissfunctions, emittance, energy, energy spread, bunchlength, slice emittance and slice energy spread. Thisis achieved with a five quadrupole system togetherwith transverse deflecting cavity (TDC) andspectrometer line. This arrangement has been usedwith success at, for instance, FERMI@Elettra [83,84]. The large relative energy spread required atthe bunch compressor, of O(2%), dictates that thequadrupole integrated strengths in this section arelimited by chromatic aberration to kl ' 0.7 m−1.This in turn implies that these sections are relativelylong.

Figure 4.3 shows the optics of these diagnosticsections (starting at 0 m and 19 m respectively).βx is small at the start to minimise the CSRinduced emittance growth in the preceding bunchcompressor. Large βy at the TDC and verticalphase advance of ∼ 90 is chosen to maximisethe vertical size of the streaked beam on a screen


0 5 10 15 20 25

0

10

20

30

s HmL

Βx

,yHmL

TDC

Screen

TDC

ScreenΒy TW

Βy SW

Βx TW

Βx SW

Figure 4.3: Optics of CLARA from bunch compressor exit (at 24.5 m from the cathode) to dogleg entrance,comprising low and high energy diagnostics lines with Linacs 3 & 4 separating. Matches for both standingwave (SW) and travelling wave (TW) linac focussing are shown. Between each TDC and subsequent screenthere is a vertical phase advance of π/2.

situated after the fifth quadrupole, The followingfour quadrupole telescope ensures that CLARA candrive the FEL without changing optics from thediagnostic configuration. This ensures minimalintervention from running mode, and allows thepossibility of online diagnostics on the upgrade ofthe spectrometer dipole to a pulsed magnet.

We estimate the achievable slice resolutionfollowing the method of Craievich [85]. Figure 4.4shows that we are able to resolve 10 fs slices in thelow energy diagnostics line at 70 MeV with a TDCdeflecting voltage of 5 MV.

Figure 4.5 shows the long pulse bunch trackedthrough the low energy diagnostics section. Thebunch is sliced in 0.1 ps time slices at the TDC,and the screen image with TDC on and off is shown.Figure 4.6 shows the same information, but with thescreen now in the spectrometer line after the fifthquadrupole. The screen is at the same path lengthfrom the TDC as the non-dispersive screen with|ηx | = 0.5 m, allowing energy and energy spreaddetermination. With the TDC on, the resolution ofslice parameters is seen.

A similar analysis in the high energy diagnosticsline at 250 MeV produces similar results for a TDCwith deflecting voltage of 10 MV. The possibilityof moving (or inserting an additional) TDC anddiagnostics section after the FEL will be investigatedas an upgrade.

4.2 Beam Dynamics

The beam was simulated from the cathode until theexit of Linac 1 using ASTRA to include the effects ofspace-charge. Wakefield effects have not yet beenincluded. The rest of the machine was then trackedusing elegant [47]. Linac wakefields, longitudinalspace-charge and coherent radiation effects wereincluded.

4.2.1 Electron Source

The electron source of CLARA will initially be theVELA 2.5 cell S-band normal conducting gun, asdetailed in Chapter 5. A solenoid surrounds thegun cavity and an additional bucking coil cancelsthe magnetic field on the cathode plane. Thedesign peak field is 100 MV/m, which equatesto a maximum beam energy of 6.5 MeV. For allsimulations, an intrinsic transverse emittance fromthe copper photocathode is included as per LCLSmeasurements of 0.9 mm mrad per mm rms ofa flat-top laser spot [86]. A laser diameter of1 mm has been assumed with Gaussian longitudinallaser profile of duration 76 fs rms. This shortpulse length allows the gun to operate in theso-called ‘blow-out’ regime, where the bunch lengthexpands due to space-charge. For the long pulsemode a 250 pC bunch expands to 1.3 ps rmsin the gun and does not evolve afterwards untilthe magnetic bunch compressor. The blow-out


-100 -50 0 50 1000.0

0.2

0.4

0.6

0.8

1.0

1.2

Screen Coordinate HΜmL

Inte

nsi

tyHA

.U.L

Dx=50 ΜmDx=40 ΜmDx=30 ΜmDx=20 Μm Dx

A

B

1

æ æ

æ

æ

æ

ææ æ æ æ æ æ æ æ æ æ æ æ æ æ

à à à à

à

à

à

à

à

à

àà à à à à à à à à

ì ì ì ì ì ì

ì

ì

ì

ì

ì

ì

ì

ì

ìì

ìì ì ì

ò ò ò ò ò ò ò ò

ò

ò

ò

ò

ò

ò

ò

ò

ò

ò

òò

0 50 100 150 2000

20

40

60

80

100

Separation of Gaussians HΜmL

Inte

nsi

tyC

on

tras

tH%L

40 Μm

30 Μm

20 Μm

10 Μm

2

0 1 2 3 4 5 60

10

20

30

40

50

TDC Deflecting Voltage HMVL

Res

olve

dS

lice

leng

thHf

sL

DΨy=120°

DΨy=110°

DΨy=100°

DΨy=90°

3

Figure 4.4: (1) Two features on a screen modelled as Gaussians separated by ∆x with screen resolution of10 µm. Intensity contrast (IC) is then defined as (A−B)

A . (2) IC as a function of the separation of the twoGaussian features for 4 screen resolutions from 10 µm to 40 µm. (3) Resolved slice length in low energydiagnostics line as a function of TDC deflecting voltage assuming a screen resolution of 10 µm and IC of70%.

æææ

æææ

ææææ

æ

æææææææææ

æ

æ

ææææ

æ

æææ

æ

æ

ææ

æ

ææ

ææ

ææ

æ

ææææææææææææææ

æ

ææææææ

æææææææææææ

æææææ

æ

æ

æ

æ

æ

æ

æ

æ

ææææææ

ææææ

ææææææ

æ

æææææææææææææ

æ

æææææ

æ

ææ

æææææ

æ

æ

æææææææææææ

ææææ

æ

ææææææææææ

æ

æææææææ

æ

æ

ææ

æææææææææ

æ

ææææææ

æææææ

æææ

æ

æ

ææ

æ

æææ

ææææ

æ

æææææææ

ææææ

ææ

æ

æ

æææ

æ

æææææ

æ

ææææææææ

æ


æ

æ

æ

æ

ææ

ææ

æ

æ

æ

æ

æææ

æ

æææ

æææ

ææ

ææææ

æ


ææææææææ

æææææ

ææ

æ

æ

æææææææ

ææ

æ

æ

æ

æ

æ

ææ

ææææ

æ

ææææ

æ

æææ

æ

æ

ææ

æ

æææ

æ

æ

æ

æ

ææ

æ

æ

æ

æ

æ

ææ

ææ

ææææ

æ

ææ

æ

æææææ

æææ

ææææ

æææææ

æ

æ

ææ

æ

æ

ææ

æ

ææ

ææ

æ

æ

æ

ææ

æ

æ

ææ

ææææ

ææææææ

æ

ææ

ææææ

æææ

æææ

ææææ

ææ

æææææ

ææææ

æ

æ

ææ

ææ

æ

æ

æ

æææ

æ

æ

æææææ

æææ

æææ


æ

æ

ææ

æ

æææææ

æ

æ

æ

æææ

ææ

æ

æææ

æ

æ

æ

æ

ææ

ææ

æ

æ

ææ

æ

æ

æ

æ

æ

æ

æ

æ

æ

æææ

æ

æ

æ

ææ

ææ

æ

æ

æ

æ

æ

æ

æ

æ

ææææ

ææ

æææ

ææ

æææ

ææ

æ

æ

æ

æ

æ

æ

æ

ææ

ææ

æææ

æ

æææ

æ

æ

ææ

æ

æææææ

æ

æ

æ

æ

æ

æ

ææ

æ

æ

æ

æ

ææ

ææ

æ

ææ

æ

æ

æ

æ

æ

ææ

æ

ææ

æ

æææ

æ

ææ

æ

æ

æ

æ

ææ

ææ

æ

æææ

æ

æææææ

æ

ææææ

æ

æ

æ

æ

æ

æ

æ

æ

æ

æ

æ

æ

ææ

ææ

æ

æ

ææææææ

æ

æææææ

æ

æ

ææ

æ

æ

æ

æ

æ

æ

æ

ææ

æ

æææææ

æ

æ

æ

æ

æ

ææ

æ

ææ

ææ

æ

æ

æææ

æ

æ

æ

æ

æ

æ

æ

æ

æ

ææ

æ

æ

ææ

æ

æ

ææ

æ

ææ

æ

æ

ææ

æææ

æ

æææ

æ

æ

æ

æææ

ææ

æ

æ

ææ

æ

æ

æ

æ

æ

æ

æ

æ

ææ

æ

ææ

æ

æ

æ

æ

æ

æ

æ

æææ

æ

æ

æ

æ

æ

æ

æ

æ

æ

æ

ææ

æ

ææ

æ

æ

ææ

æ

æ

ææ

ææ

ææ

æ

æ

ææ

æ

æ

æ

ææ

æ

æ

æææææ

æææ

æ

æ

æ

æ

æææ

æ

æ

æ

ææææ

æ

æ

ææ

æ

æ

æ

ææææ

æ

ææ

æ

ææ

æææ

ææ

æ

ææ

æ

æ

æ

ææ

æææ

æ

æ

æ

æ

æ

æ

æ

æ

ææ

æ

æ

ææææ

æ

æ

æ

æ

æ

æ

æ

æ

æ

æ

æ

æ

ææ

æ

ææææ

ææ

æ

æ

æ

æ

æ

æ

æ

æ

æ

æ

æ

ææ

æ

ææ

æ

æ

æ

æ

æ

æ

æ

ææ

æææ

ææ

æ

æ

æ

æææ

æ

ææ

æ

æ

ææ

ææ

æ

æ

æ

æ

æ

ææ

ææææ

ææ

æ

ææ

æææææ

æ

æ

æ

ææ

æ

æ

æ

æ

ææææææ

æ

æ

æ

ææ

ææ

æ

æ

ææææ

æ

æ

æ

ææææ

ææ

æ

æ

æ

æææ

æ

ææ

ææ

æ

æ

æ

æ

ææ

æ

ææææ

æææ

ææ

æ

ææ

æ

ææ

ææ

æææ

ææ

æ

æ

æ

ææ

æ

æ

æææ

ææææ

æ

æææ

æ

ææ

æ

æ

æææ

æ

ææ

ææ

æææ

æææ

æ

æ

ææ

æ

æ

æ

æ

æ

ææ

ææææ

æ

æ

æ

ææ

æ

æ

æ

ææ

æ

æ

ææ

æ

æ

æ

æ

æ

ææ

æ

æ

æææ

æ

ææ

æ

æ

æææ

ææ

æææ

æ

æ

æ

æ

ææ

æ

æ

æ

æ

æ

æ

æ

ææ

ææ

æ

ææ

æ

æ

æ

æææ

æ

æ

æ

æ

æ

æ

æ

æ

æ

æ

æ

ææ

æ

æ

æææ

ææ

æ

æ

æææ

æ

æ

æ

æ

ææææ

æ

ææ

æ

æ

æ

æ

æ

æ

æææ

æ

ææ

æ

æææ

æ

ææ

ææ

ææ

ææ

æ

ææ

æ

æ

æ

æ

æ

ææ

æ

æ

ææ

æ

æ

æ

æ

æ

ææææ

æ

æ

ææ

æææ

æ

æ

æ

ææ

ææ

ææ

ææ

æ

æææ

æ

ææ

æ

æ

æ

æ

æææ

æ

ææææ

æ

æ

æ

ææ

æ

æææ

ææ

æ

æ

ææ

æ

ææ

æ

æææ

æ

æ

æ

æ

æ

ææ

ææ

æ

æ

ææ

æææ

æ

æ

ææ

æææ

æ

æ

ææ

ææææ

ææ

æ

ææ

æ

æ

æ

æ

æ

æ

æ

æ

æ

æ

æææ

æ

æ

ææ

æ

æ

æææ

æ

æ

æ

ææ

ææ

ææ

ææ

ææ

æææ

æ

æ

æ

æ

ææ

æ

æ

æ

æ

æææ

æææ

æ

æææææ

æ

ææ

æ

ææææ

æ

æææ

æ

æ

æ

æ

æ

æ

æææ

æææ

æ

ææ

ææ

ææ

ææ

æ

æ

æ

æ

æ

ææææ

æ

æ

æææ


æ

æææææ

æ

æ

æææ

æææ

æ

æ

æ

æææ

æ

ææææ

æ

æ

æ

æ

ææææ

æ

æ

æ

ææ

æ

ææææææææ

ææææ

ææææ

æ

æ

æ

æ

ææææææ

ææ

æ

æ

æ

æ

æ

æ

æ

ææææææ

æ

æ

æ

æ

æ

æ

æ

ææ

ææ

æææ

ææ

æææ

æ

æææ

ææ

æ

æ

æ

æ

æ

æ

æ

ææææ

æ

ææ

ææ

æ

æææææææææ

æææ

æææ

ææ

æ

æ

ææææææ

ææ

æææ

æ

ææ

æ

æææ

æ

ææ

ææ

æ

æ

æ

ææææ

æ

æ

æ

ææææ

æ

ææ

æ

æ

ææ

ææ

æææ

æ

æ

ææ

æ

ææææ

ææææ

-0.4 -0.2 0.0 0.2 0.498

99

100

101

102

t at screen HpsL

DppH%L

ææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææ

ææææææææææææææææææææææææææææææ

æææææææææææææææææ

ææææææ

ææææææææææææææææææææææææææ

ææææææææææææææææææææ

æææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææ ææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææ

-6 -4 -2 0 2 4 6

-6

-4

-2

0

2

4

6

x HmmL

yHm

mL

æææææææææææææææææææ

ææ

æææææææææ

ææææææææææææææææææææææææææææææææ

ææææææææææææææææææææææææææææææææææ

æææææ


æææææææææææææææææææææææææ

ææææææææææææææææ


æææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææ

æææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææ ææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææ

-6 -4 -2 0 2 4 6

-6

-4

-2

0

2

4

6

x HmmL

yHm

mL

ææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææ

æææææææ

æææææææææææææææææ

æææææ

æææ

ææææææ


æææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææ

-0.4 -0.2 0.0 0.2 0.4

-6

-4

-2

0

2

4

6

t at TDC HpsL

yHm

mL

1 2 3 4

Figure 4.5: Long pulse bunch streaked by TDC in low energy diagnostics line: (1) Long. phase space atscreen; (2) Screen image with TDC deflecting voltage = 0 MV; (3) Screen image with TDC deflectingvoltage = 5 MV; (4) Correlation between arrival time at TDC and vertical position on screen.

æææ æ

æ

æ

ææ ææ

æ

æ

ææææææ æææ

æ

æææ

æ

æ

ææ

æ

æ

ææ

æ

æ

ææ

ææ

æ æ

æ

æææ

æææææ

ææ ææææ

æ

ææææ

æ ææ ææææ

æ ææææ

æ

æææææ

æ

ææ

æ

æ

æ

æ

æ

æ æææ ææææ

ææ

ææ æ

æææ

æ

ææææ ææ ææææææ

æ

æ

æææ

æ

æ

æ

æ æ

æææ

æ

ææ

æææ ææ

æææ

ææ

ææ

ææ

æ

æ

æ

ææ

æ æææ ææ

ææ

æ

æ

ææææ æ

ææ

æ

æ

æ

æææ

æ æ

æææ æ

æ

æ

ææ æææ

æææ

ææ

æ

ææ

æ

æ

ææ

æ

æææ

ææ ææ

æ

æ

ææ æææ

æ ææææ

ææ

ææ

æææ

æ

æææææ

æ

ææ

æææ æ

ææ

æ

æææ

ææ æææ

ææ

æ

æ

æ

æ

ææ

æ

æ

æ

æ

æ

æ

æææ

æ

æææ

æ ææ

ææ

æææ ææ

æ

ææææ

æææææ

ææææææææ

ææææææ

æ

æ

æææ

æ

æææ æ

ææ

æ

æ

æ

æ

æ

æ æ

ææ ææ

æ

æ æææ

æ

ææ

ææ

æ

æ

æ

æ

æææ

æ

æ

æ

æ

ææ

æ ææ

æ

æ

ææ

ææ

ææ

ææ

ææ

æ

æ

æ ææ

æ

ææ

ææ

ææææ

æææ

ææ

æ

æ

ææ

æ

æ

ææ

æ

ææ

æææ

æ

æ

æ

æ

æ

æ

ææ

æ æææ

ææææ

ææ

æ

ææ

æææ

æ

æææ

æææææææ

æææ

æææææææ

æ

æ

æ

ææ

ææ

æ

æ

æ

æææ

æ

æ

æ

æ

æ

ææ

æ

æææææ

æææ ææ

ææ

ææ

æ

æ

æ

ææ

æ

æææ

ææ

æ

æ

æ

æ

ææ

ææ

ææ

ææ

æ

æ

æ

æ

ææ

ææ

æ

æ

ææ

æææ

æ

æ

æ

ææ

æ

æ ææ

æ

æ

æ

æ æ

æ æ

æ

æ

æ

ææ

æ

æ

æ

ææææ

ææ

æ ææ

ææ

æææ

ææ

æ

æ

æ

æ

æ

ææ

æ

æ

ææ

æææ

æ

ææ

æ

æ

æ

ææ

æ

æææææ

æ

æ

æ

æ

æ

æææ

æ

æ

æ

æ

ææ

ææ

æ

ææ

æ

æ

æ

æ

æ

ææ

ææ

æ

æ

ææ æ

æææ

æ

æ

æ

æ

ææ

ææ

æ

æ ææ

æ

ææ

ææ æ

æ

ææææ

æ

æ

æ

æ

æ

æ

æ

ææ

æ

æ

æ

ææ

ææ

æ

æ

ææ

ææ

ææ

æ

æææææ

æ

æ

ææ

æ

æ

æ

æ

æ

æ

æ

ææ

æ

æææææ

æ

ææ

æ

æ

ææ

æ

ææ

ææ

æ

æ

æ

ææ

æ

æ

æ

æ

æ

æ

æ

æ

æ

ææ

æ

æ

ææ

æ

æ

ææ

æ

ææ

æ

æ

ææ

æ ææ

æ

æææ

æ

æ

ææææ

ææ

æ

æ

ææ

æ

æ

æ

æ

æ

æ

æ

æ

ææ

æ

ææ

æ

æ

æ

æ

æ

æ

æ

æææ

æ

æ

æ

æ

ææ

æ

æ

æ

æ

ææ

æ

ææ

ææ

æ æ

æ

æ

ææ

ææ

ææ

æ

æ

ææ

æ

æ

æ

ææ

æ

æ

æææææ

æææ

æ

æ

æ

æ

æææ

æ

æ

æ

ææææ

æ

æ

ææ

æ

æ

æ

æææ æ

æ

ææ

æ

ææ

æ ææ

ææ

æ

ææ

æ

æ

æ

ææ

æææ

æ

æ

æ

æ

æ

æ

æ

æ

ææ

æ

æ

ææææ

æ

æ

æ

æ

æ

æ

æ

æ

æ

æ

æ

æ

ææ

æ

ææææ

ææ

æ

æ

æ

æ

æ

æ

æ

æ

æ

æ

æ

ææ

æ

ææ

æ

æ

æ

æ

æ

æ

æ

ææ

æ ææ

ææ

æ

æ

æ

ææ

æ

æ

ææ

æ

æ

æ æ

ææ

æ

æ

æ

æ

æ

ææ

ææææ

ææ

æ

ææ

æ ææææ

æ

æ

æ

ææ

æ

æ

æ

æ

æ ææ æ ææ

æ

æ

æ

ææ

ææ

æ

æ

ææææ

æ

æ

æ

æææ

æ

ææ

æ

æ

æ

æ ææ

æ

ææ

ææ

æ

æ

æ

æ

ææ

æ

ææææ

ææ

æ

ææ

æ

ææ

æ

ææ

ææ

æææ

ææ

æ

æ

æ

ææ

æ

æ

æææ

æææ æ

æ

æææ

æ

æ æ

æ

æ

æææ

æ

ææ

ææ

æææ

ææ æ

æ

æ

æ

æ

æ

æ

æ

æ

æ

ææ

æ æææ

æ

æ

æ

ææ

æ

æ

æ

ææ

æ

æ

ææ

æ

æ

æ

æ

æ

æ æ

æ

æ

æææ

æ

æ æ

æ

æ

æææ

ææ

ææ

æ

æ

æ

æ

æ

ææ

æ

æ

æ

æ

æ

æ

æ

ææ

ææ

æ

ææ

æ

æ

æ

æææ

æ

æ

æ

æ

ææ

æ

æ

æ

æ

æ

ææ

æ

æ

æ ææ

ææ

æ

æ

ææ æ

æ

æ

æ

æ

ææææ

æ

ææ

æ

æ

æ

æ

æ

æ

ææ æ

æ

ææ

æ

æææ

æ

ææ

ææ

æ æ

æ æ

æ

ææ

æ

æ

æ

æ

æ

ææ

æ

æ

ææ

æ

æ

æ

æ

æ

ææææ

æ

æ

ææ

æææ

æ

æ

æ

ææ

ææ

ææ

æ æ

æ

æææ

æ

ææ

æ

æ

æ

æ

ææ æ

æ

ææææ

æ

æ

æ

ææ

æ

æææ

ææ

æ

æ

ææ

æ

ææ

æ

æææ

æ

æ

æ

æ

æ

æ æ

æ æ

æ

æ

ææ

æææ

ææææ

æææ

æ

æ

ææ

æ æææ

ææ

æ

ææ

æ

æ

æ

æ

æ

æ

æ

æ

æ

æ

æææ

æ

æ

ææ

æ

æ

æææ

æ

æ

æ

ææ

ææ

æ æ

ææ

ææ

æææ

æ

æ

æ

æ

ææ

æ

æ

æ

ææææ

æææ

æ

æ ææææ

æ

ææ

æ

ææææ

æ

æææ

æ

æ

æ

æ

æ

æ

æ

æ

æ

æææ

æ

ææ

ææ

ææ

æ æ

æ

æ

æ

æ

æ

ææ ææ

æ

æ

æææ


æ

æææææ

æ

æ

æææ

æææ

æ

æ

æ

æææ

æ

ææææ

æ

æ

æ

æ

ææ

æ æ

æ

æ

æ

ææ

æ

ææ

ææææææ

æ æææ

ææææ

æ

æ

æ

æ

ææææææ

ææ

æ

æ

æ

æ

æ

æ

æ

ææææææ

æ

æ

æ

æ

æ

æ

æ

ææ

ææ

æææ

æ æ

æææ

æ

æææ

ææ

æ

æ

æ

æ

æ

æ

æ

ææææ

æ

ææ

ææ

æ

æææ

ææææææ

æææ

æææ

æ æ

æ

ææ

æææ

æ æ

ææ

æææ

æ

ææ

æ

ææ æ

æ

ææ

ææ

æ

æ

æ

ææææ

æ

æ

æ

ææææ

æ

ææ

æ

æ

ææ

ææ

ææ

æ

æ

æ

ææ

æ

ææææ

ææææ

-3 -2 -1 0 1 2 398

99

100

101

102

t at screen HpsL

DppH%L

æææææææææææææææææææææææææææææææææææææææææ ææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææ

æææææææææææææææææææ

æææ

æææ

ææææææ


ææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææ

ææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææ æææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææ

ææææææ

æææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææ ææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææ

æææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææææ æææææææææ ææ æææ ææææææææææææææ ææææææææææææææææææææææ

-6 -4 -2 0 2 4 6

-6

-4

-2

0

2

4

6

x HmmL

yHm

mL

ææææææ æææææ ææææææææææææææææ æææææææ æææææææ ææææææ

æææææææææ

æææææææ æææææææææææ æææææ æ

ææ ææææ ææææææææææææææææææ æææææææææææ

ææææææææ

æææ

æææææ ææææææ

æææ

ææææææ

ææææææææææææææ ææææææ

æææææ

æææææææææ ææææææææææææææææ æææææ

æææææææ ææææææææ

ææææææ ææææ ææ

ææææ

æ ææææææææ ææææææææææææ æ ææ æææææ ææ

æææææ

ææææææ

ææææææææææææææææææææææææææææææææ ææææææææææææ ææ

æææ

æææææææææææ ææææææææ ææ

æ æææ æ æææææ ææ

ææ æææææææ ææ ææææææææææææææææææ

æ ææææ æææ

æ ææææ

æææ æææ æææ æææ

æææææææææææææææææææææææææææææææ æææææ ææææææ ææææææææ

ææææææææææææææææææ

æææææææ ææææææ ææ æææææ ææææ ææ æ

ææææææ æææ

ææææ ææ æææ ææææ

ææ ææææ

ææ

æææææææææææææææææææææææ ææ ææææææ æææ

æææææææ æææ ææææææ

ææ

æ

ææ æææ

æ ææ æææ æææ

æææææ

æ æææ ææææ æææ æææ ææ

ææææ ææ

ææææ ææææææ ææææææ ææ

æ æææ

ææ æ ææ ææææ

ææææææææ æææææææææææææææ

æ ææææ ææææææ æææ

æææ æææææ ææ æææææææ

æææ æææææ

ææææ ææ

æ æææ ææ æææææ ææææ

æææ ææææ

æ ææææ ææ æ

æ æææ æææææææ æææææ ææææææææ ææ

æ ææ æ æææ ææææææ ææ ææææææææ æææ ææææ

æææææææ ææææ æ ææ ææææ ææ ææææææ

æææ ææ ææææ ææææ ææ æææææ

ææææææææ æææ

æææ æææ ææ æææ æ æææææ ææææ æ

æææ

ææ æææ æææææ ææ

ææ

ææ ææ ææ æ

ææææææææ ææææ ææ

ææ æææ æææææ

ææææ

ææ ææ æææææææ ææ æææææææ

æ æææ æææææææ æææ ææ ææ ææææææ

ææææææææ

æææææ æ

ææææææ æææ ææ æææ

æææ ææ

ææææææææææææææ ææ æææ æææææ ææææææ ææææ æ

æ ææææææææ ææææ æææ æææææ ææææææææææææææ ææææ ææææææææææ

æææ

æææææ

æ

æ æææ ææææææ ææææ æææ ææ æææææææææ æææ ææææææææ æææ ææ

æ ææ ææ ææææ ææ

ææææ

ææ

æææ

æææææææææ æ

æ æææææ

æ æææææ æ æææ æ ææ

ææ ææææ æææææææææææ æææææ

æææ

æ æææ ææææ æææ ææ æ

æææææææææææææ æ æææ æææææææ æææ æææ

ææ

ææ æææ

ææææææ

ææææææææ

ææææ ææ

ææææ

æææ

æææ æ æææ

æææææææææ ææææææææææææææææææ æææ

æ ææ

ææ ææææææææ æ

ææææææææ ææ ææ

ææ ææææææ æææ ææ

æ ææææææ

æææææ æææ ææææææææ æææææ ææææææææ æ

ææææ ææææ ææ

ææææææææææ ææææææææææ ææææææææææææææ ææææææ

æææææææææ æææææ æææ

ææææ ææ æææææææ æææ ææææ

ææææææææ æææææ ææææææææææææ æææææ æææææææ ææ ææ æææ

æææææææææ æææææææ

ææ ææ ææ ææ ææææææ

ææ æææ

ææææææææææææ æææææ ææææææææææææææ æææ

æææ

æææ ææ æææ ææææ æææææææ æææ æ æææææ æææ ææ æææ ææææææææ

-6 -4 -2 0 2 4 6

-6

-4

-2

0

2

4

6

x HmmL

yHm

mL

æææ ææææ

ææææ æææ

æ æææ æææææ

ææ ææææ æ

ææ ææ æ æææ

ææ ææ ææ æææææææææææææææææ

æ ææ ææ æææ ææ ææ

ææææææææ

æ æææææ

æ ææ

æ æææææ æææ

ææ æææææ

ææ

ææ æææææ

æ ææææ

æ ææ

ææ æ

æ æææ ææ

æææ ææ

ææ æ

æææææ

æ

æ æææ

æ æ æææææ

æ æææ æææææ

ææææ ææææ æææææ

ææ ææ æ

ææææææææ

æ ææææ æ ææ æææ ææ æææ æææ æææææææææ æ ææ ææææ æææ

ææææ æ

æææ ææææ æææ ææ ææ ææ æææ ææææ

æ ææææ

ææ ææææææ

æ æææææææ æææ

æææ ææææ ææ ææææææ ææ ææææææææ ææ

æææ

æ æ æææ ææ æ

ææ æ

æææ ææææ æææ ææææ

æ ææ æææææææ æææææ æ æææ

ææææ ææ ææ ææ ææææææ

æ ææææææ ææææææ

æææ ææææææææææ æ

æ ææ æ æææææ æææææææ æææ

ææææ æææææ æææ ææææææ æ ææææææ ææææ ææææ ææ

æ ææææ æææ

æ æææ ææææ ææææ æææ æææ ææ æ

ææ ææææææ ææææ ææ æææ ææææ æææ

ææææ ææ ææææ æ ææææ æææ ææ ææ

ææ æ æ æææ

æ ææææææææææææ ææææ ææææææææ ææ

æ æ ææææ æææææ

ææææææææ ææææ

æææ

æ ææ æææææææ æææ

æææ æ æ

æææ æææ ææ æææ æææ æææ

æææææ

ææ

æ æææ æ æææ ææ ææææææ æææ

æææ

æ æææææææææ ææ æææææ ææææææææ ææææææææ ææ æææ ææ æ æææ

ææ æææ ææ æ ææ ææææ æææ æ

ææ ææææææ

ææ æææ ææææææ æææ æ

ææææ ææææ æææ ææææ æ

æææ ææææ æææ æææ ææ ææ æ

ææææææææææ ææ æ

æææææææ ææ æææ ææ æææææ æææ æææ æææææ

æ æ ææ ææ æææææ ææææææææ æææææææææ ææ ææææ ææ ææ ææææ æææ æ

æ ææ ææææææææ

ææææ æææ ææ ææ æææææ ææ ææ ææ

æ ææ

æææææææ æææææ

ææ

æææææææ

ææ

æ ææææ ææææ ææ

ææææ æææ ææ æææ ææææææ ææ æææææææææææ ææææ

ææææææææææ ææ æææææææ ææ æ ææ

ææææææ

æææ

ææææææ

æ ææ æææ æ ææææææ

ææææææææ

ææææ ææææææææææ æææææ ææææææææææ æææææææ æ ææ æææææææ æ ææ

æ æææææææææ æææ ææææ æææ ææææææ ææææ æææ æ

æ

æ ææææ

æ

ææ æ æææææ æææ æææææ ææ ææ ææææ æææ ææ æææ ææææææ ææææææææ ææ ææææææ æ

æææ æ

ææ

ææææææææ ææ æææ

æææ ææ æææææææææææ æææææææ æ æææææææ ææææææ ææ æææææ

æææ ææææ æææ æææææææææææ ææææ ææææ ææææææ æ æ ææ æææ ææ

ææ

ææææ ææ ææææ

ææ

ææææ ææ ææææææææææ

ææææ

ææ ææ ææ æ

æ ææææææææææ

æææææ ææ ææ

æ æ ææææ ææææææ

ææ ææ ææææææææ ææææ

æ ææ æææææ

æææ æ æææ æææææ æ

æææææ ææ

æææææææ ææ ææææ

ææ æææææææææææ ææ ææææ æææææææ ææææææ ææææ ææ ææ ææææ æ æææææææææ æææ æææææææææææ

æææææææææ æ

ææææææææ æææ ææ ææ ææææææ ææææææææææ æææææææ æææææææææææææææææ ææ æææææææ ææææææææææææææ æææ ææææææææææææ ææææææ

æææææ

ææææææææææ æææææææ æææææææ

æ ææææææ ææææææææææææ

ææ ææææææ ææ æææææ ææææææææææ ææ æææææææ ææææ æ

-3 -2 -1 0 1 2 3

-6

-4

-2

0

2

4

6

t at TDC HpsL

yHm

mL

1 2 3 4

Figure 4.6: Long pulse bunch streaked by TDC in low energy diagnostics spectrometer line: (1) Long.phase space at screen; (2) Screen image with TDC deflecting voltage = 0 MV; (3) Screen image withTDC deflecting voltage = 5 MV; (4) Correlation between arrival time at TDC and vertical position onscreen.


Figure 4.7: Seeded bunch longitudinal phase space, current profile, normalised slice emittance and sliceenergy spread.

regime has been demonstrated experimentally andcompared to simulations [87]. Full characterisationof this scheme and the beam properties achievablewith the CLARA gun will be tested using VELA.There is also the possibility of adding pulse stackersto stretch the laser pulse to the required lengthif the blow-out regime is found experimentallyto not provide the required beam properties. Asolenoid surrounding Linac 1 is used for transversefocussing and emittance compensation and Linac 1is operated off-crest to provide part of the chirprequired for magnetic bunch compression.

4.2.2 Optimisation of Seeded Mode

The seeded operating mode is the most challengingas it requires that the bunch slice properties at theentrance to the FEL are constant for a large fractionof the bunch length, i.e. 250 fs out of 500 fs FW,and that the peak current is above 400 A for thisfraction, without significant energy chirp. This is toensure that the FEL output performance is tolerantto jitter between the laser seed and the electronbunch.

To achieve this an optimisation [88] wasperformed on the longitudinal phase space of the

bunch at the FEL. At each step a transverserematch was necessary to preserve the projectedemittance. The optimisation variables are thevoltages and phases of the S-band linacs and X-bandlinac, and the bunch compressor bend angle. Theoptimisation constraints were that there exists a300 fs window where the mean current exceeds400 A, and in that window:

• the minimum current in a 20 fs slice is greaterthan 370 A,

• the standard deviation of the charges in a 20 fsslice is less than 20 pC and

• the chirp is no larger than 1% of the energy spreadin the 20 fs slice at the centre of the window.

In Figure 4.7 we show the optimised longitudinalphase space, current profile, slice emittance andslice energy spread at the FEL. The optimisationconstraints are broadly met. The standard deviationof the charges in the window is marginal. This maybe due to insufficient numerics in the simulationsthus far produced. The transverse matching alongthe bunch is also suboptimal and will be the subjectof further study.


Value Unit

Gun Gradient 100 MV/mGun φ −25

Linac 1 V 21.0 MeV/mLinac 1 φ −20

Linac 2 V 11.5 MeV/mLinac 2 φ −31

Linac X V 7.3 MeV/mLinac X φ −168

BC θ 95.0 mrad

Linac 3 V 22.5 MeV/mLinac 3 φ +0

Linac 4 V 22.5 MeV/mLinac 4 φ +0

Table 4.2: Optimised machine parameters for seeded bunch (gradients are specified at on-crest levels).

Table 4.2 shows the optimised machineparameters. The relatively low gradient in Linac 2and high gradients in Linacs 3 and 4 arise from therequirement to minimise the chirp. This will notbe the case in other modes, therefore flexibility inoperation of linac gradients is required.

4.2.3 Other Operating Modes

An alternative to magnetic compression is to usevelocity bunching in the low energy section of theaccelerator. Linac 1 is set to the zero crossing phaseto impart a time-velocity chirp along the bunch.The bunch then compresses in the following driftspace. Linac 2 then rapidly accelerates the beamand ‘captures’ the short bunch length. Solenoidsare required around Linac 1 to control the transversebeam size and prevent emittance degradation.

Simulations suggest that this mode ofcompression can produce a similar bunch profileto that produced by the magnetic compressionscheme if the fourth harmonic cavity is not usedfor linearization. Thus both compression modescan be used to meet the SASE mode of CLARAoperation.

Velocity bunching can also be used to drivethe ultra-short mode of operation. Preliminarysimulations suggest that for a 100 pC bunch, a peakcurrent ∼ 1 kA can be achieved, with a FWHMbunch of 100 fs. At the slice of the peak current,

the energy spread is ∼ 250 keV rms and the sliceemittance of ∼ 1 mm mrad. Further study andoptimisation of this mode is required, including thefull effect of wakefields at low energy.

4.3 Tolerance Studies

Parameter scans of the photoinjector (PI) laserproperties, RF voltages and phases, and bunchcompressor dipole field from the nominal values ofTable 4.2 are shown in Table 4.3. In each case wescan over the range quoted and state the maximalvariation in four observables, arrival time change,rms energy spread change, rms bunch length changeand mean energy change. To assess jitter sensitivitywe show the rms change from nominal of the fourobservables from 50 machines where all parametersare varied according to Gaussians with the standarddeviations quoted in Table 4.4. Note that these aredifferent from the ranges of the parameter scanswhich are chosen to be larger to establish trends inthe observables.

Table 4.5 shows the resulting rms bunchproperties. We see that at ∼ 60 fs, the rms arrivaltime jitter is significantly smaller than the ∼ 300 fsflat region of the long pulse bunch, ensuring reliableseed laser pulse overlap, and that the energy andbunch length (hence peak current) jitter satisfy thespecification given in Section 3.3.5.


Parameter Varied Range ∆ tarr (fs) ∆ δE (%) ∆σz rms (%) ∆E (%)

PI Laser ∆x ±0.1 mm 1.1 1 0.06 0.0004PI Laser ∆y ±0.1 mm 0.6 0.27 0.08 0.0004PI Laser ∆t ±300 fs 3 1.8 1.9 0.05PI Laser ∆σ ±10 % 15 50 9 0.06Bunch Charge 100− 300 pC 60 200 25 0.16

Gun ∆V ±1 % 110 4 4.5 0.06Gun ∆φ ±1 o 90 3.5 3.5 0.1

Linac 1 ∆V ±1 % 400 2.6 2.5 0.35Linac 1 ∆φ ±1 o 250 9 7 0.2

Linac 2 ∆V ±1 % 50 3 3 0.5Linac 2 ∆φ ±1 o 550 16 15 0.6

Linac X ∆V ±1 % 55 1 0.9 0.05Linac X ∆φ ±1 o 20 3 3 0.02

Linac 3 ∆V ±1 % 2.5 0.45 0.007 0.4Linac 3 ∆φ ±1 o 3 0.5 0.0035 0.3

Linac 4 ∆V ±1 % 0.9 0.4 0.0015 0.4Linac 4 ∆φ ±1 o 0.7 0.36 0.0025 0.3

BC ∆B ±0.01 % 10.5 7 6 0.5

Table 4.3: Tolerance data for the long pulse mode. Each parameter listed is varied over the range stated,change in arrival time, rms energy spread, rms bunch length and mean energy at the exit of Linac 4 aretabulated.

rms Value rms Value

σ∆σ laser 5% σφG 0.1

σ∆x laser 10% σV (LS,LX) 0.1%σ∆y laser 10 % σφLS 0.1

σ∆t laser 200 fs σφLX 0.3

σQbunch 2% σBC∆B 0.001%σV G 0.1%

Table 4.4: The rms values of the Gaussian distributions for the variables used to estimate jitter sensitivity.

Variable rms

∆ tarr (fs) 64.0∆ δE (%) 18.6

∆σz rms (%) 5.0∆E (%) 0.10

Table 4.5: The rms bunch properties of fifty jittered machines with respect to the nominal machine.


4.3.1 Beam Based Alignment Strategy

CLARA will require accurate alignment of theelectron beam to ensure the stringent beamproperties required for FEL operation are met.The operational strategy for correcting the beamposition to the centre of the magnetic elementsis based around a combination of Singular ValueDecomposition (SVD) trajectory correction, aswell as alignment to the magnet centres throughbeam-based alignment techniques. Trajectorycorrection on CLARA will be performed throughthe use of dedicated horizontal and vertical steeringdipole magnets. We assume these magnets to beindependent of each other but located at the samephysical location. The magnets use high-precisionpower supplies to provide accurate changes to thebeam angle at regular intervals through the lattice.The steering magnets are coupled with BPMs,spread throughout the machine, which provide beamtrajectory information. The strip-line BPMs areexpected to have . 20 µm resolution in the mainpart of the machine, with additional high-precisioncavity BPMs in the bunch compressor and betweenall undulators. The cavity BPMs are expected toprovide sub-micron resolution in single shot mode.The beam trajectory will be monitored and correctedvia the use of SVD of the trajectory responsematrix. This method provides a flexible and robusttrajectory correction system for the whole machine.Emittance spoiling in the main linacs can be aconcern for a low-emittance FEL facility. Thisis primarily induced by spurious dispersion throughthe accelerating structures. We assume someform of dispersion minimisation strategy for thelinac structures, based around a modified SVDalgorithm. Due to their narrow aperture design, and

concomitant tight position tolerances, alignment ofthe FEL undulator modules is expected to be criticalto FEL operation. We envisage improved BPMresolutions in the FEL section, primarily by utilisingcavity BPM structures, as well as a much stricterpre-alignment strategy to minimise the correctorrequirements and BPM offsets. We will use astandard alignment strategy based on powering ofquadrupole magnets and analysis of the resultingbeam motion on downstream BPMs. We will alsofurther investigate the possibility of photon-beamalignment, whereby downstream photon diagnosticsare used in place of electron BPMs and theundulator spontaneous signal can be used as a proxyfor the FEL output. The alignment strategy will relyon aligning the beam on an undulator by undulatorbasis. The beam optics in the FEL section is also ofparamount importance. This will be complicated bythe continuous FEL gap variability required, and assuch, we will require high accuracy beam diagnosticscreens throughout the FEL sections.

Finally, for continued stability of the electronbeam, we envisage the use of a shot-to-shotfeedback system, utilising the steering magnets andBPMs previously described, as well as feed-forwardfor the undulator gaps. The intention is to specifycomponents and controls such that shot-by-shotcorrection at 100 Hz is possible. Feed-back systemsfor the linac gradients will be performed in thebunch-compressor and dogleg structures, and willbe incorporated into the general SVD correctionalgorithm. Real-time analysis of beam frequencyspectra will be used to help diagnose unforeseennoise sources, and the use of feed-forward loops forcertain components is expected. Model IndependentAnalysis methods will be fully utilised throughout.



5Accelerator Systems

Figure 5.1: The VELA/ALPHA-X photoinjector gun.

5.1 Electron Source

The primary electron source of CLARA will be basedaround a normal conducting S-band photocathodegun with metal and/or alkali photocathodes. Thistechnology has been selected due to the desire todeliver low emittance beams with modest chargeand repetition rate. The VELA 2.5 cell gun operatedwith a fixed copper photocathode is implementedwith a maximum field of 100 MV/m and used asa baseline electron source. This gun fulfills therequirements of the seeded, SASE and ultra-shortpulse modes of CLARA.

Our vision for the further development of theelectron source will be concentrated on the designof a new highly stable, high repetition rate gun withinterchangeable photocathodes. This new design

would also allow us to increase the operationalfield to 120 MV/m. A gun with interchangeablephotocathodes allows for the use of both metaland alkali photocathodes. The removal of thephotocathode cleaning and preparation procedurefrom the gun will eliminate deterioration of the guncavity and, as a result, improve the stability of thegun operation. Experimentation of different types ofmetal photocathodes will be possible to reduce thebeam emittance and high quantum efficiency alkaliphotocathodes can also be implemented.

5.1.1 Baseline Electron Source

The baseline electron source for CLARA will be theVELA photoinjector [89]. The electron gun is a2.5 cell normal conducting S-band RF design, as


Cavity Design

Repetition Rate Gradient 1.5 Cell 2.5 Cell

100 Hz100 MV/m

Pulsed RF Power (MW) 5.7 10.0Average Power (kW) 1.7 3.0

120 MV/mPulsed RF Power (MW) 8.2 14.4Average Power (kW) 2.4 4.3

400 Hz100 MV/m

Pulsed RF Power (MW) 5.7 10.0Average Power (kW) 6.8 12.0

120 MV/mPulsed RF Power (MW) 8.2 14.4Average Power (kW) 9.8 17.3

Table 5.1: Operational power of the CLARA high repetition rate gun at different operation modes.

shown in Figure 5.1. This was originally intendedfor use on the ALPHA-X laser wakefield accelerationproject [90]. The gun is operated with a copperphotocathode driven by the third harmonic of aTi:sapphire laser (266 nm), which is installed in adedicated thermally stabilized room. The injectorwill be operated with laser pulses with an energyof up to 1 mJ, pulse duration of 76 fs RMS anda repetition rate of 10 Hz. At a field gradient of100 MV/m provided by a 10 MW klystron, the gunis expected to deliver beam pulses with energy ofup to 6.5 MeV. Modelling with ASTRA suggests thelength and emittance of the electron bunches atthe exit of the gun varies with charge from 0.1 psat 20 pC to 5 ps at 250 pC and from 0.2 to0.5 mm-mrad respectively. A solenoid surrounds thegun cavity with a bucking coil to zero the magneticfield on the cathode plane.

5.1.2 Advanced Electron Source

Further development of the CLARA electron sourcewill be concentrated on the design of a highly stable,high gradient and high repetition rate gun cavitywhich would allow operation at a field of 120 MV/mat 100 Hz repetition rate and 100 MV/m at400 Hz. Both 1.5 cell and 2.5 cell designs are underconsideration.

Operational Field and Power

One of the critical limiting factors which restrictsperformance of normal conducting RF guns is theRF power dissipated inside the cavity. Excess

power leads to detuning of the gun and parasiticmodulation of the amplitude and phase of theaccelerating field. We estimate the operationalpower of the gun on the basis of existing designs.For a 1.5 cell option we have selected a highrepetition rate design [91] with a well-developedcooling system. The proposed cavity can potentiallyoperate with a maximum field of 100 MV/m with3 µs RF pulses with a repetition rate of 1 kHz. Themaximum cathode field is restricted by the averagepower dissipated in the cavity which exceeds 17 kW.For a 2.5 cell cavity, the required power may beestimated on the basis of the existing VELA guncavity [90]. The RF power required to reach acathode field of 100 MV/m is 10 MW and at arepetition rate of 10 Hz average dissipated power is300 W.

The results of scaling these design parametersto the CLARA requirements of 100 Hz and400 Hz repetition rate at two gun cathode fieldsof 100 MV/m and 120 MV/m are summarized inTable 5.1. As may be seen the 1.5 cell design isable to cover all possible operation modes at bothpeak and average power of 10 MW and 10 kWrespectively which can be delivered by the existingVELA RF station. The 2.5 cell design is able tooperate in 100 Hz 100 MV/m mode only. All othermodes would require an upgrade of the gun RFstation.

Dark current generated by field emission is asource of concern for RF guns operating at highfield. The S-band gun at LCLS has shown that0.6 nC is produced over a 2 µs train [92]. Dryice cleaning was first demonstrated at PITZ, which


Figure 5.2: CST simulation of the electric field distribution in the VELA/ALPHA-X front-coupled 2.5 cellgun.

resulted in a dark current 10 times lower thana similar cavity cleaned with high pressure waterrinsing [93]. Collimators in the gun region may alsobe used to remove unwanted dark current.

Gun Stability

Seeded FEL experiments which require interactionbetween a short laser pulse and the electronbunch place extremely high demands on the RFgun stability. For example, the jitter of thelaunching phase of the beam in the magnetic bunchcompression mode should be less than 300 fs,which, in terms of the S-band RF phase, is 0.32.To provide such a phase stability the requiredcavity peak to peak temperature stability should bebetter than 0.028C. This is still below the currentstart-of-the-art performance of thermal stabilisationsystems which is 0.04C [94].

Cavity detuning which is due to a smallmechanical deformation caused by the RF heating isa dynamic process and the phase shift changes alongthe RF pulse following this detuning. For the singlebunch modes, the required phase stability may beachieved by proper selection of the bunch launchingtime within the RF pulse and the introduction ofslow feedback from arrival time monitors. Forthe multi-bunch operation mode, the situation iscomplicated by the rise of the cavity temperaturecausing detuning and consequent phase shift changealong the RF pulse. For the S-band cavities thesituation is even more complicated as the timerequired for deformation to propagate along the

cavity wall is comparable to the duration of the RFpulse and for cavity thermal analysis the steady stateapproximation, typically used for analysis of the longpulse L-band gun, is not applicable. Very carefulcoupled transient analysis should be performed tosatisfy the extremely high phase stability requiredfor CLARA.

The dynamics of pulsed cavity detuning is anew subject and is now under investigation incollaboration with the Institute for Nuclear Researchof the Russian Academy of Science. Preliminaryanalysis indicates that phase instability may becompensated with a feed forward phase correctionwith Low Level RF and other subsystems.

Cavity Design

The RF cavity of the photoinjector must operate inboth the 100 Hz 120 MV/m and 400 Hz 100 MV/mregimes, delivering bunches of up to 6 MeV energy.

Two designs are under consideration: a 1.5 cellcavity as seen in the conventional S-band injectordesign [95] or a 2.5 cell design, similar to the cavity[90] currently used on VELA. In order to providefeedback for stabilising the amplitude and the phaseof the RF field in the cavity it will be equipped withpick-up sensors. Initial design simulations suggestthat the gradient required to deliver a 6 MeV beamusing a 1.5 cell cavity may cause prohibitive heatingand resulting frequency instability. The impact ofthe lower energy delivered by the 1.5 cell gun needsto be investigated.

Two methods of RF power coupling are being


considered. One option is to use coaxial couplingwhich is advantageous because it conserves the axialsymmetry of the RF field in the cavity. This schememay however require a complex engineering designand is also susceptible to multipactor. A coaxialcoupler can either be implemented from the frontor back of the gun. Front coupling, as in the VELAdesign (see Figure 5.2), is an established technologybut may suffer also from heating issues.

The other possibility is to use side coupling witha rectangular waveguide. This has the advantageof simplicity, but compromises the field symmetry inthe cavity which is then transmitted to the beam andalso restricts the position of the focusing solenoid.In order to compensate the field asymmetry atwo-waveguide (as on the LCLS gun [96]) and evenfour-waveguide scheme may be considered.

Further evaluation is required to narrow downthe design choices in terms of numbers of cellsand coupling scheme. Alternative design optionswill also be investigated. Detailed simulationwork and optimisation using multiple codes willthen be required to find a design that satisfiesall the beam physics, RF, cooling, vacuum andmechanical requirements necessary to meet theCLARA specification.

Photocathode and Drive Laser

The gun photocathode in the initial stage will bedriven by an existing frequency-tripled Ti:sapphirelaser system with a repetition rate of up to 400 Hz.The maximum pulse energy at 266 nm, at theexit of the laser system, is 1 mJ and the pulseduration is 180 fs FWHM (∼76 fs RMS). Thepulse energy is sufficient to deliver 250 pC chargefrom a metal photocathode without relying on highquantum efficiency, and allowing for unavoidableenergy losses in the transport line.

Future developments will look at the use ofhigher quantum efficiency photocathode materialssuch as single-crystal Cu and metals such as Mg, Pb,Nb and eventually telluride based photocathodes.In order to transport photocathodes into the gunand easily interchange them, a load-lock system willbe used. Such systems were originally developedfor GaAs based DC photocathode guns then laterwere implemented into RF guns [97, 98, 99].The design of the photocathode plug will bechosen to provide reliable RF contact with thegun cavity. Presently two plug designs are widelyused: the DESY/INFN-LASA design developed in

the framework of the TESLA collaboration [99],now used at FLASH/PITZ, FNAL and LBNL, andthe CERN-CTF3 design [98] which has been chosenby PSI for the SwissFEL project. One of thesetwo plug designs will be adopted in order to becompatible with some of the groups listed.

The photocathode plug will be initially preparedin the photocathode preparation system basedon the upgraded Vacuum Generators ESCALABIIfacility being set up at the Accelerator Scienceand Technology Centre (ASTeC) at Daresbury.Preparation of metal photocathodes will include exvacuo chemical and in vacuo thermal cleaning andsurface processing with O3 plasma to remove carboncontaminants. Surface analysis of the preparedphotocathodes will be made with high resolutionX-ray photoelectron and Auger spectroscopy.Prepared photocathodes will be transported to theaccelerator hall with a vacuum transport systemwhich will be attached to the gun load-locksystem. This procedure will allow for maintainingultra-high vacuum conditions in the gun and enablereplacement of photocathodes in tens of minutes. Aload-lock system will also allow for investigation ofa broad range of photocathodes in order to obtainmaterials for optimal performance at CLARA andfuture FEL facilities.

5.2 Radio Frequency Systems

A review has been performed regarding thefrequency options for the CLARA RF system,which assessed the options for S-band, C-band andX-band for both the European and US frequencyoptions. The review assessed the availability ofhardware (klystrons and cavity designs) and alsoconsidered the various mixes of frequencies. Itwas concluded that European and US frequencieswould not be mixed so as not to cause phasestability issues and limit the buckets that canbe filled during operation, and that for hardwareavailability the most favourable option would be theEuropean S-band and X-band frequencies. Thusthe CLARA RF system consists of 4 types of RFcavity structures; an S-band photoinjector gun, 4S-band linac structures providing acceleration ofthe electron bunches up to 250 MeV, an X-bandcavity for the linearisation of the longitudinal phasespace, and 2 S-band TDCs for 6D emittancecharacterisation of the beam at low and high energy.

The initial photoinjector gun to be installed on


Figure 5.3: RF system layout

Linac 1 Linacs 2-4 4th Harmonic

Number of cells 43 122 72Frequency (MHz) 2998.5 2998.5 11994Nominal RF voltage (MV/m) 20 20 12Maximum RF voltage (MV/m) 25 25 16Nominal RF power (MW) 23 37 6Max RF power (MW) 32 57 29Operating Mode π/2 2π/3 5π/6

Repetition rate (Hz) 100 100 100Filling time (µs) <1.0 0.995 0.1Length Flange to Flange (m) 2.15 4.15 0.965Active Length (m) 2 4.07 0.75Quality Factor (Q0) ∼14000 ∼15000Shunt Impedance (MΩ/m) 82 62 68Nominal operating temp. (C) ∼25 ∼ 35 ∼ 35

Table 5.2: Design specification for Linacs 1–4 and 4th Harmonic Cavity

CLARA will be the current S-band RF gun installedon VELA, which is the ALPHA-X gun built byLaboratoire de l’Accélérateur Linéaire (LAL) andprovided by Strathclyde University.

The photoinjector is a 2.5 cell, 6 MeV gun whichoperates at repetition rates up to 10 Hz [90, 100]. Itis intended that a high repetition rate photoinjectorgun is designed and installed to meet the CLARArequirements and is discussed in Section 5.1.2.

For each of the individual cavity structuresa dedicated high power RF system is to be

installed to enable flexibility in providing the requiredaccelerating gradients for the individual CLARAdefined requirements as well as the amplitudeand phase stability necessary to meet the beamjitter specifications. Additionally, as there isa requirement for a 16 bunch macro-pulse of2 µs duration a SLED (SLAC Energy Doubler)compression has not been considered as part ofthe RF design option. The RF layout is shown inFigure 5.3.


Figure 5.4: Concept design for a 2 m long, 43 cell, high repetition rate linear accelerating structure.

Figure 5.5: Superfish simulation of accelerating cells.

5.2.1 Linac Accelerating Structures

Linac Section

The initial linac section is to be a 2 m S-bandaccelerating structure, required to accelerate theelectron beam up to 35 MeV, but will be operated inan off-crest mode (up to 30). A concept design hasbeen produced by Advanced Energy Systems INC(AES) which is a standing wave (SW) structure thatoperates in the π/2 mode to provide an acceleratinggradient of 25 MV/m with a 3 µs RF pulse at arepetition rate of 400 Hz. Table 5.2 shows thespecification for the structure which consists of43 accelerating cells and 42 on-axis coupling cells,with 3 RF pick-up loops as shown in Figure 5.4.Figure 5.5 shows the electrical field modelling workperformed on the accelerating cells. To ensure thatthe filling time is kept to less than 1 µs it will

be necessary to have an increased RF power atthe beginning of the pulse before reducing to itssteady state level. In addition AES have performeda thorough thermal analysis of the structure designto ensure the high average power levels can bedissipated. Further analysis of the requirements isto be performed to define whether a SW structuremeets the full requirements or whether a travellingwave (TW) would be more suitable, especially withrespect to phase stability.

The further 3 linac accelerating sections areto be kindly provided by PSI from the SwissFELInjector facility. The linac sections are 4.15 mlong from flange to flange, and are constant 2π/3

TW structures based on the mechanical design oflinac II DESY structures. To meet the SwissFELrequirements the shunt impedance was increasedto reduce the peak RF drive power required to


Figure 5.6: 3D model of the transverse deflecting cavity.

Number of cells 9Frequency (MHz) 2998.5Nominal RF voltage (MV) 5Nominal RF power (MW) 6Operating Mode π

Repetition rate (Hz) 10Nearest mode (MHz) 6.5Length flange to flange (m) 0.698Active length (m) 0.5Quality factor (Q0) ∼18000Shunt Impedance deflecting mode R (MΩ) 5.0Aperture pipe diameter (mm) 35 (Iris 32)Nominal operating temp. (C) 35Number of pickups 2 (1 pickup, 1 HOM)

Table 5.3: Specification for the transverse deflecting cavity at low energy.

provide the accelerating gradient and the RFsymmetry is maintained with a split input coupler.However, these structures have been designed for100 Hz operation, thus for operation at 400 Hzthe accelerating gradient will have to be reducedso as not to over dissipate the structures. Theparameters for the linac structures are defined inTable 5.2 and the designs of these structures arediscussed fully in the Swiss FEL Injector ConceptualDesign Report [101].

Fourth Harmonic Cavity Structure

The X-band, 11,994 MHz, harmonic cavity isrequired to linearise the longitudinal phase space

of the electron beam to correct for the collectiveinstability effects in the successive acceleratingstages caused by space charge and wakefields. Tomeet this requirement a deccelerating gradient of12 MV/m is needed. There have been a numberof X-band structures developed for the EuropeanX-band frequency, and a review of these structureshas shown that to meet this requirement it willbe necessary to use a TW structure comparedto a SW structure as these structures are lesssensitive to temperature variations. It is proposedto use a modified design for the European X-bandfrequency based on the SLAC H75 structure, whichis a constant gradient, 5π/6 phase advance TWstructure comprising of 72 cells [4, 102, 103]. A


Figure 5.7: Vertical particle trajectories for the TDC (left) without and (right) with beam correction.

5π/6 phase advance TW structure is preferred toa 2π/3 as these structures have proven greateroperational reliability. A review of the design willbe performed to ensure that the parameters definedin Table 5.2 are met. Additionally, an essential partof the design of this structure will be the ability toalign the cavity to within 10 µm to ensure accuratebeam steering and to enable accurate correction tobe applied to the beam.

Transverse Deflecting Cavities

Two S-band TDCs are required, one at low energy,70–150 MeV and the second at the full energy of250 MeV. These will be used to provide longitudinalslice emittance profiling of the electron beam andare required to provide deflecting voltages of 5 MVand 10 MV, respectively.

For VELA a 9-cell TDC has been designed [104](Figure 5.6) which provides a 5 MV transverse kick.Parameters are shown in Table 5.3. The cavity hasbeen designed with steering magnets at the entranceand exit and shortened end cells to reduce theeffect of the electrical field decay in the beampipes,so as to minimise the effect of the vertical kickon the electron beam trajectory (Figure 5.7). Amatching short opposite the input coupler has beenincluded to maintain the RF symmetry to minimisethe monopole component of the RF fields in thecoupling cell.

For the higher energy requirement it will benecessary to investigate whether the cavity iscapable of the higher gradient levels or whether itwill be necessary to design a higher gradient TWstructure.

5.2.2 High Power RF Systems

For the CLARA design, klystron modulators arebeing considered in order to provide the pulsed RF

power needed to obtain the required acceleratinggradients within each of the RF cavity structures.A K2 klystron modulator has been supplied byScandiNova (Figure 5.8) and successfully operatedon the photoinjector gun cavity on VELA.Additionally a second klystron modulator is beingsupplied by Diversified Technologies (Figure 5.9)for operation on the first linac cavity. Modulatorrequirements are outlined in Table 5.4. Typicallythe best stability that can be achieved at present is0.0001 and 0.1 [101] for the amplitude and phasestability respectively, and it will be necessary tocompensate for slow fluctuations with a low levelRF system. Presently there is ongoing developmentwork being pursued by the modulator manufacturersto achieve greater stability. Progress on this workneeds to be reviewed to ensure that this is in linewith CLARA requirements. The RF amplitude andphase stability are essentially defined by the stabilityof the klystron modulator voltage.

5.2.3 Low Level RF System

RF Reference and Timing System

The synchronisation system for CLARA will consistof a very stable, low noise RF reference oscillatorsystem that will be distributed to the accelerationsections (Figure 5.10). The RF referencedistribution will use a temperature elevated andstabilised environment system running the length ofthe machine to distribute timing and RF signals.This system will use copper RF cables to providea baseline system for both RF and timing signalsto a level of 50 fs. An automatic phase driftcompensation system will be used that should allowstability of less than 10 fs to be achieved [105].In addition a compensated optical fibre distributionrunning in parallel with the coax system will offer theultimate timing performance down to <10 fs. This


Figure 5.8: ScandiNova K2 klystron modulator

is explained in detail in Section 5.5. The decision touse both technologies side by side, although moreexpensive, should offer the greatest flexibility ininitially setting up the machine and achieving theultimate performance level.

At each acceleration section, low noise localphase locked loops and multiplier/dividers will beused to generate all the required RF frequencies andtiming signals locked to the reference system.

Low Level RF system

The purpose of the Low Level RF (LLRF) systemis to stabilise and regulate the RF fields inthe accelerating cavity. Errors in amplitude andphase will cause energy modulation of the beamparticularly if the accelerating cavities are runningoff-crest; this in turn causes timing jitter frompulse to pulse after the bunch compressors. Theability to feed back becomes very difficult, althoughtechniques are now emerging that seek to measureand correct for small errors on a pulse-by-pulse

basis [106]. The dominant part of the systemthat affects the RF fields seen in the cavity is theklystron modulator and drive system. The stabilityand repeatability of these systems to always producethe same RF field on each pulse is the critical andlimiting factor for the short term performance ofthe machine. The accelerating cavities are equippedwith field probes: these can be used to monitor thephase and amplitude inside the structure, providinginformation on the stability of each cavity overmultiple pulses. These signals in conjunction withthe forward and reflected power signals will providethe LLRF with information on the state of tune foreach cavity and the amount of phase slippage seenby the beam in each accelerating structure. Feedforward calculations can be made and applied toattempt to correct for these issues.

Environmental effects, temperature, humidityand pressure will cause changes to the machineand the RF measurement systems themselves.The accelerator hall, RF modulator areas and allmeasurement cabling will be held at a precise


Figure 5.9: Diversified Technologies klystron modulator system.

Figure 5.10: Basic RF Control System


Figure 5.11: An example of a commercial digital LLRF system from Instrumentation Technologies.

temperature (to ±0.1) to combat at least part ofthese issues, however these environmental changesare slow in nature, and they can be detected inthe operation of the machine. To further reducethe effect of these environmental issues, the LLRFsystem will be equipped with a calibration system[107], this will measure the path length phase andamplitude of the RF drive system on every pulse,monitoring slow changes in the system and applyingcorrections to the drive set points, this will bring thesystem long term stability up as high as possible. Inaddition the use of beam monitoring systems (beamarrival monitors and bunch compression monitors)will provide electron beam measurement and timingstability information that will be used to providefeedback to the RF phase and amplitude at a rate ofup to 1/10th of the machine repetition rate [108].

Due to the necessity to interface with calibrationsystems and electron beam monitoring equipment,digital based systems for RF control will beimplemented on CLARA. These systems provide avery flexible and robust solution that can be adaptedand upgraded as the machine is operated.

Digital LLRF systems down convert the RFsignals from each cavity to an intermediatefrequency (IF) of around 50 MHz. Thesignals are sampled using high speed 14/16 bitAnalogue-to-Digital Converters (ADCs) and passedto Field Programmable Gate Array (FPGA) systemsfor digital processing and adaptive feed forwardcompensation. The resulting control signals foramplitude and phase are up converted and appliedto the klystron modulator for amplification. Thehardware for 3 GHz and 12 GHz accelerating cavitieswill be the same apart from changes to the analogfront ends. An example of a commercially availableLLRF system is Instrumentation TechnologiesLibera LLRF system shown in Figure 5.11.

It is anticipated that commercially availableLLRF hardware will be purchased, that includeshighly stable front end systems, low cross talkbetween RF channels and direct ExperimentalPhysics and Industrial Control System (EPICS)interface. The hardware will allow programmingand simulation via mathematical program routines(MATLAB etc). By using this commercially

Photoinjector Linac 1 Linacs 2–4 4th Harmonic LE TDC HE TDC

Frequency (GHz) 3 3 3 12 3 3RF Peak Power (MW) 10 23–32 30–60 6–29 6 21RF Average Power (kW) 10 21 <9 4.5 0.2 0.7Klystron Voltage Range (kV) 170 280 260–350 331 156 225Klystron Current Range (A) 125 300 265–415 219 120 215Flat Top Pulse Length (µs) 3 3 1.5–3 1.5 3 3Repetition Rate (Hz) 10–400 10–400 10–400 10–100 10 10

Table 5.4: Modulator requirements for photoinjector cavity, Linacs 1–4, 4th harmonic cavity, low energy(LE) TDC and high energy (HE) TDC. Where ranges of values are given these represent the anticipatedrequirements over a variety of accelerator setups and operating modes.


Figure 5.12: Overview of beam diagnostics.

designed LLRF hardware we would then be ableto focus on the design and implantation ofcontrol algorithms in conjunction with the LLRFmanufacturer and will be able to update the LLRFhardware and software as the project develops.

It will be possible to use fast analog control loopsto monitor and feedback on the klystron modulator[109]. This control loop will be able to reducemodulator noise and remove phase droop during theRF pulse. This control loop requires the use of ahigh speed phase shifter and locating the electronicsvery close to the modulator to keep latency timesto a minimum.

The performance of the LLRF system in theshort term (based on a modulator stability of4.6× 10−5 over 1 min) will be 0.035% amplitudestability and 0.05 phase stability.

5.3 Electron BeamDiagnostics

5.3.1 Bunch Charge Monitors

Three types of monitors are planned to be usedfor bunch charge measurement. These will bewall current monitors (WCM) [110], IntegratingCurrent Transformers (ICT) and Faraday Cups (FC)with the FC providing precision and low chargemeasurements. Four WCMs and ICTs are to beinstalled after the gun (dark current measurement),after the X-band cavity, and in the low energy andhigh energy diagnostic sections. The FCs installedat the end of each spectrometer diagnostic section

and at the machine end will also serve as full beamdumps.

On the basis of CLARA simulations,shot-to-shot bunch charge stability of better than1% is required in the high charge operation modes.To provide reliable single bunch charge variationmonitoring, monitors of different types will beinstalled and the most reliable and sensitive monitorswill be identified operationally. The ICT is acommercial ‘in-flange’ device, and the WCM isa FNAL/STFC Daresbury design. The FC andits associated electronics is a further upgradeof an ASTeC design produced for the ALICE(Accelerators and Lasers in Combined Experiments)accelerator. The VELA test facility has enabledfull beam testing of FC developments prior tospecification for CLARA.

5.3.2 Strip Line BPMs

For routine bunch trajectory tracing, twentyfive BPMs are spread along the machine (seeFigure 5.12). The strip line pickups haveincorporated survey monuments and can be alignedwith respect to the quadrupole magnet axes. Theresidual zero offset (including BPM electronics zerooffset) is expected to be within 0.2 mm. Transversebeam size, varying along the machine, on averageis ∼0.2 mm. It is required that the BPMs have abunch charge range from 250 pC down to severalpC. At the lower range of charge, thermal noise willlimit strip line BPM accuracy to ∼30 µm, while aresolution of 20 µm is expected for the higher chargemodes.


It is intended to use a further development ofthe in-house designed EPICS based VME BPMselectronics. This system was originally madefor single-bunch/turn-by-turn measurements withbutton pickups on the EMMA (Electron Machinefor Many Applications) Accelerator [111, 112].Advanced testing of the CLARA version of theelectronics using strip line pickups will be carried outusing VELA. This will provide cross check data withthat previously measured from ALICE to confirm theachievable BPM resolution. Data Acquisition BPMsoftware already written for EMMA and ALICE willalso be used with CLARA.

5.3.3 Cavity BPMs

In the dispersive section of the bunch compressor,and in the inter-undulator/modulator regions, thebeam position accuracy and resolution requirementsare significantly more challenging than elsewherein the accelerator. The requirements here aresignificantly greater than can be provided by a stripline BPM design, and only cavity BPMs can providethe resolution and accuracy required.

A design programme in collaboration with RoyalHolloway University London (RHUL) has beeninstigated to develop a cavity BPM for CLARA,based upon work carried out on a cavity BPM for theInternational Linear Collider (ILC) final focus modelat ATF2 in Japan [113]. Further investigation is stillrequired to confirm that sub-micron resolution withCLARA’s low bunch charge is possible. Prototypingand development/testing will be carried out onVELA prior to cavity BPM production.

5.3.4 Screen Diagnostic Systems

Screen systems employed will incorporate bothCerium doped Yttrium Aluminium Garnet (YAG)and Optical Transition Radiation (OTR) screens asappropriate.

Beam transport systems will be instrumentedby a compact standard diagnostic unit based upondesigns from FLASH in Germany. These incorporatea number of selectable devices mounted on a linearactuator stage. Each screen ‘carriage’ that can beinserted orthogonally into the beam holds a 5 µmthick substrate mounted YAG screen or silicon OTRscreen.

An additional screen location allows theinstallation of a calibration graticule screento provide accurate camera focussing and a

dimensioning facility for images. Screens are viewedon a camera system via an offset mirror througha side port in the diagnostic unit. This versatiledevice is less than 20 cm in length and can alsoincorporate a wire scanner and button type BPM ifrequired, for example to overcome coherent OTRimage distortion [114]. Screen system diagnosticsin the intra-undulator sections will be requiredto image not only the electron beam, but alsothe photon beam as well. The co-existence ofthe electron and photon FEL beams requires anadvanced multi-beam screen system designed fortransverse size and profile measurement on both theelectron beam and the FEL radiation. Challengingdesign constraints at these locations include therequirement to suppress any present coherent OTR,seed laser and minimise any ionizing radiationdelivered to the undulators during beam imaging.A novel design for the FERMI@Elettra FEL hasbeen reported, with an optical design resolution of12 µm [115]. To provide this functionality a similarscreen system will be employed on CLARA.

For precise emittance measurement, theelectron beam size has to be measured with anerror better than 10% of the rms beam size. Thestandard camera system used by FLASH uses asimple 1:1 lens system and can provide a theoreticalresolution of 10 µm/pixel for a beam size ∼100 µm.Improved cameras for specific locations such asthose in the intra undulator sections where thephoton beam is to be imaged on the YAG screenswill be required. In conjunction with a distortionminimising lens system, the improvement in signalto noise of the system is such that a theoreticalresolution of 8 µm/pixel. In operation this shouldprovide resolutions better than 20 µm, which shouldbe sufficient for precise beam emittance and photonbeam measurements.

5.3.5 Beam Arrival Monitors

Beam Arrival Monitors (BAMs) will be implementedto improve understanding of the beam dynamicsand monitor synchronisation points such as the FELentrance. The BAMs measure the relative delaybetween the electron bunch centroid and the lasermaster oscillator. The electron bunch timing isdetected with electrodes in the beampipe which pickup the Coulomb field of the passing electron bunch.The output from the electrodes is then sampled byoptical clock pulses (see Section 5.5) by gating theiramplitude in an electro-optic crystal. Measurement


FEL Scheme Type Output Wavelength range (nm) Pulse Length Range (fs)

SASE 100–400 20–300Short Pulse 266–400 2–50High Temporal Coherence 100–400 30–300

Table 5.5: Output spectral ranges and pulse durations that CLARA is expected to deliver in each of thethree main types of FEL scheme.

of the sampled pulse amplitude then gives a measureof the electron time delay with respect to the masteroscillator. Such BAMs have been implementedat DESY and FERMI@Elettra with demonstratedresolutions of better than 10 fs [116].

5.3.6 Bunch Compression andTemporal Profile Monitors

The compression of the bunch will be monitoredthrough a pyroelectric detector measuring the CSRpower emitted in the final dipole of the compressor.This monitor is not envisaged as providing aquantitative measure of the bunch length, butrather as a means for stabilising the compressionat an operationally determined set-point, Forquantitative measurement of the temporal profile,a spectral upconversion [117] diagnostic will beused to measure the spectrum of the CSR atthe compressor. A second in-beamline spectralupconversion monitor will be located before the FELmodulators. The spectral upconversion systems willbe used for non-invasive monitoring and stabilisationof the accelerator and beam transport.

A TDC system is also present in the highenergy diagnostic section, and will be used for sliceemittance and sliced energy characterisation of thebunch. The spectral upconversion diagnostics willbe calibrated against the TDC.

5.3.7 Laser Arrival Monitors

All-optical Laser Arrival Monitors (LAMs) will beused for monitoring the photoinjector laser arrivalat the virtual cathode, and the FEL seed lasers at apoint close to the injection of the near-IR or mid-IRseed into the electron beamline. The LAMs will bebased on optical cross-correlation with the opticallydistributed RF timing signals (Section 5.5).

5.4 FEL Output Diagnostics

5.4.1 Introduction

The main aims of the CLARA FEL outputdiagnostic systems will be to measure the temporaland spectral properties of a single light pulse. Thepulse length and spectrum of CLARA FEL outputwill span a wide range of values depending on theFEL scheme employed (see Chapter 3). Table5.5 summarises the different FEL schemes and theanticipated spectral and temporal ranges of theoutput.

SASE operation will be the baseline againstwhich the schemes to improve temporal coherenceand deliver shorter pulses will be measured. SASEdelivers relatively long pulses with complex temporaland spectral profiles. The product of pulse durationand bandwidth will be considerably larger than thetransform limit. The diagnostics will give a detailedpicture of both the temporal and spectral structure.

Improvements on basic SASE are the corepurpose of CLARA. In schemes aimed specificallyat delivering shorter pulses (such as Single-SpikeSASE and Mode-Locking), the diagnostics willultimately have to measure pulse durations downto the femtosecond level. Such schemes will alsodeliver significantly enhanced temporal coherenceand spectral and temporal measurement will showhow far the pulses deviate from the transformlimit. The schemes aimed specifically at deliveringthe highest temporal coherence (seeding, EEHGetc) will not deliver such short pulses but accuratetemporal and spectral measurement will be essentialto quantifying how successful the schemes are.

Developing the techniques for achieving a highconsistency in FEL output from shot to shot is alsoa goal of CLARA as this is of critical importance forscientific exploitation of FEL sources. This requiresall the output diagnostics to be capable of workingon a ‘single-shot’ basis, ideally at the full repetitionrate of the FEL.


Wavelength (nm) 100 266 400

HorizontalRMS size (µm) 131 243 223Position (m) 15.3 11.1 10.9M2 1.25 1.73 1.69

VerticalRMS size (µm) 112 257 241Position (m) 14.9 11.5 11.0M2 1.15 1.62 1.66

Table 5.6: Source parameters derived from a second moment analysis of the simulated output of theCLARA FEL.

The nature of the FEL source needs to beconsidered when designing the diagnostics. Somesort of focusing of the source will be requiredand the size and position of the source thereforeneeds to be known. Source size, position and M2

factor can be deduced from FEL simulations usinga second-moment analysis of the output field filespropagated to different distances. Table 5.6 showsthe results from simulations of the FEL outputin SASE mode with 2.5 m long radiator modules.The position is given relative to the start of thefirst radiator module. It is expected that thesewill still be realistic values for operating modes inwhich the temporal properties are modified as thespatial properties are not expected to change. Thesimulations show that the source and position aredifferent in the horizontal and vertical directions andvary with wavelength. It is likely that an adaptiveoptical scheme will be required to give the correctfocusing for the diagnostics.

5.4.2 Spectral Diagnostics

The aim will be to record the complete spectrumof a single photon pulse from the FEL. This willbe accomplished with a flat-field spectrograph anda multi-channel array detector. The spectrographwill be either a commercial instrument or anin-house design using a concave grating withvaried-line-spacing to achieve a suitable flat-field.The detector will be a commercial instrument.

The basic requirement for the spectrograph is forthe spectral width of the flat-field to be large enoughto capture the spectrum of the shortest pulses andthe spectral resolving power to be high enoughto resolve the longest pulses or spectral structurewithin SASE pulses. These two requirements arenot completely compatible, and some compromisewill be necessary if only a single instrument is

employed. The priority will be in capturing the fullspectrum of ultra-short pulses.

For a temporally coherent pulse, the spectrumand pulse length are linked through the transformlimit. For a pulse with a Gaussian temporal profile,the spectrum has a Gaussian profile and the FWHMof the spectral and temporal profiles in nm and fsare related by

∆λFWHM∆tFWHM = 1.472× 10−3λ2

To accurately measure the pulse spectrum, theflat-field of the spectrograph will have to spansubstantially more than the FWHM of the spectrum.A flat-field width of 5 times the rms width of thespectrum has been chosen as being a sensible designtarget. Figure 5.13 shows the relation between pulselength and this target spectral width for wavelengthsof 100, 266 and 400 nm. The calculation is for pulselengths down to 1 fs, but such short pulses will onlybe feasible for the shortest wavelengths and will notbe delivered in the first phase of CLARA operation.Nevertheless, the figure demonstrates the enormousvariation in the spectral width of the CLARA outputover the possible range of output wavelengths andpulse lengths.

The suitability of various commercialspectrographs has been examined in an internalreport [118]. This report also looks at the effect ofthe spectrograph resolving power and detector arraysize on the measurability of the longest pulses (i.e.those with smallest bandwidth). The most suitableinstrument considered is the Jobin Yvon VTM300vacuum spectrograph [119], which operates overthe wavelength range 60 to 300 nm with a flat-fieldwidth of approximately 60 nm. It will be able torecord the 5σ width of a 1 fs FWHM pulse at70 nm, a 1.4 fs pulse at 170 nm and a 3.6 fs pulseat 270 nm. Suitable CCD array detectors can bebought commercially, for example the Greateyes


Figure 5.13: Required spectral measurement width for pulses at 100, 266 and 400 nm wavelength

GE 1024 256 BI UV1 or UV3 1024 by 256 pixelcameras [120].

5.4.3 Temporal Diagnostics

The ultimate aim will be to develop and implementsingle-shot diagnostics with the capability ofextracting the full spectral phase information ofa photon pulse with duration as short as 1 fsFWHM at a wavelength near 266 nm. Measuring(and generating!) such short pulses will be verychallenging so the first target will be to reliablymeasure pulses with durations down to ∼10 fs andthen to improve the technique to measure shorterpulses. Single-shot measurements of pulses ofthis duration at shorter wavelengths will not be inthe baseline specification unless such an extensionproves to be trivial.

The diagnostic approach will be derived fromthe established Frequency Resolved Optical Gating(FROG) technique, which allows the completespectral phase information of the pulse to berecovered. However, even without the full analysis,the technique is a powerful window on thecorrelation of spectral and temporal information inthe FEL pulse because the FROG measurementgives an immediate picture of how the spectrumin the pulse changes along the pulse length. Forexample, it will be possible to see the spectrum of

each spike in a SASE pulse.

FROG has already been used to measure pulsesof 8 fs duration at 270 nm [121]. However,for CLARA with pulses potentially as short as1 fs, some modifications and developments to thetechnique are required. The very large relativebandwidth contained in sub-10 fs UV pulses meansthat propagation through even short air paths willstretch the pulse through dispersion. For example,even a ‘long’ pulse of 10 fs FWHM duration at266 nm would be stretched to about 30 fs afterpassing through 1 m of air. Because the lightis generated in an accelerator, the beam path onCLARA will be several meters in length, and so theentire beam path will have to be in vacuum.

Dispersion will also stretch the pulses iftransmissive optics (windows, lenses, polarisers,beam splitters etc) are inserted in the beam path.For example, transmission through 1 mm of fusedsilica will stretch a 10 fs pulse at 266 nm to∼55 fs. Therefore beam focusing will have tobe done with mirrors and regions of ultra-highand medium vacuum will have to be separatedby differential pumping rather than with windows.The latter point is easily satisfied as the lattice ofnarrow-gap undulator vessels is expected to providean adequate buffer to the parts of the machinerequiring the highest vacuum. More significantlyhowever, the non-linear beam mixing that is at the


heart of the FROG technique (or any optical auto-and cross-correlation technique) cannot be donewithin any practical transmissive medium withoutchanging the pulse as it is being measured. It istherefore proposed to exploit non-linear interactionsat surfaces.

A key challenge for the diagnostic on CLARA isthat the initial operation in the mode-locked schemewill produce a pulse train of ultra-short pulseswith spacing of just 167 fs. Unless the detectoris gated at an impossibly high rate of 6 THz,even a ‘single-shot’ autocorrelation technique willgive an average result over all pulses in the trainbecause both the probe and measured beams havean identical pulse structure [122]. To measure eitherone pulse or a short sequence of pulses in the trainwithout this averaging requires the probe beam tohave a single pulse per macro pulse of the FEL sothat the probe acts as the detector gate.

It is conceptually feasible to take the partof the FEL beam to be used as the probe andcondition it to a single pulse by gating withfast switching mirrors, accomplished by exploitingtransient reflectivity triggered by another laser.This approach will be technically very challengingand probably quite ‘lossy’, and thus may leaveinsufficient probe beam intensity. So it is thereforeproposed to use cross-correlation with the probebeam derived from the Ti:sapphire laser used togenerate seed pulses for the FEL, which will benaturally synchronised and have one pulse per FELmacro-pulse.

The disadvantages of the cross-correlationapproach are that the probe pulse will be longer thanthe measured pulse and that the temporal phaseof the probe must be accurately known before thecross-correlation signal can be unravelled to extractthe temporal phase of the FEL pulses. However, byusing the probe at the native 800 nm or frequencydoubled to 400 nm we can use conventional FROGto characterise the beam.

There will need to be some experimentalwork undertaken to develop the cross-correlationtechnique for these very demanding measurements.The most important aspect is to decide thenature of the non-linear interaction that will beexploited and to find a suitable surface. We arecurrently conducting experiments using the VELAphotoinjector laser to see if we can exploit differencefrequency generation (DFG) between 800 nm and266 nm light. The advantage of DFG over sumfrequency generation (SFG) is that the correlation

signal will be in the visible and therefore thedetection will be more straightforward and cheaper.

5.5 Optical Timing andSynchronisation

The timing and synchronisation system for CLARAwill fulfil two main functions. The first is to generateand deliver a stable clock signal to all the acceleratorsubsystems that require synchronisation, and thesecond is to use that reference to monitor andor improve the accelerator timing stability at thesubsystem. A combination of RF and optical timingtechniques will be adopted to achieve the requiredtiming and control tolerances.

5.5.1 Synchronisation Targets

To meet CLARA’s aims in beam and photon outputstability the synchronisation of various subsystemsis critical. In particular, synchronisation at thefollowing locations will have the most stringenttiming targets.

• RF cavities to electron beam: 0.02.

• Seed laser pulse to electron beam: 200 fs rms

• FEL output to user experiments when in seedingmode: 10 fs rms

Certain diagnostics also have tight timing tolerancessuch as electron beam arrival monitors and laserarrival monitors. These diagnostics will primarily aidin the understanding of CLARA’s performance, butcould also potentially be used as feedback signals toadvance stability.

5.5.2 Timing System Architecture

The baseline timing system architecture is shownin Figure 5.14. The system will operate froma Laser Master Oscillator (LMO) which will bephase locked to a RF Master Oscillator (RFMO)to reduce drift. Timing jitter between the LMOand the RFMO is expected to be <35fs [123].Both the RFMO and the LMO, along with theoptical distribution and stabilization boards, willbe situated in a common temperature stabilised(±0.1C) EMC shielded synchronisation room. Theoptical distribution cables running the length of theaccelerator will also use a temperature elevated and


Figure 5.14: Architecture of the timing and synchronisation system.

stabilised environment system. Distribution linesfrom the LMO will supply five types of end stations:

1. Laser-to-laser locking (L2L)

2. BAM

3. LAM

4. LLRF cavity controls

5. Triggered systems

In all except the triggered systems, which hasthe most relaxed timing tolerances, monitoringand feedback techniques will be used to maintainsynchronisation of the relevant subsystem with thedelivered clock and hence with the oscillators andother subsystems.

The system will use the LMO as the commonreference for the LLRF stations, the L2L of thelaser systems and time critical diagnostics suchas the LAMs and BAMs. Each of these willreceive its reference through parallel optical fibrelinks split from an optical distribution board nearthe LMO. This prevents the accumulation of jitterin the links and will be actively stabilized againsttiming delays that can arise from thermal drifts andvibrations along its distribution. BAMs and LAMswill be supplied with synchronised and stabilizedlinks that are length adjusted to match the nominalpropagation time of photons and electron beamsalong CLARA. This will reduce the effect of anyphase instabilities in the LMO by correlating allmeasurements for a single bunch to the same sourcepulse. Thus all the BAMs will be able to be

considered as time-of-flight monitors with respect tothe first LAM. Signals for the triggered systems willbe distributed through conventional copper cables.

Similar optical timing systems withdemonstrated femtosecond level performance havelargely followed two types of schemes. Pulsedschemes which stabilise the group delay of thedistribution link such as that at DESY [124], andcontinuous wave (CW) schemes which stabilisephase delay such as that at LBNL [125]. Whileboth have their merits, CLARA will adopt a pulsedscheme because it delivers short optical pulses whichenable direct implementation of established arrivaltime monitoring techniques with high fidelity. It isnoted that some facilities such as FERMI@Elettraand SwissFEL have adopted a hybrid scheme using alower cost CW distribution for the higher toleranceRF stations. However, due to the smaller size ofCLARA with only 7 LLRF stations, the reducedcost of supplying these with a CW distribution isconsidered to be outweighed by the need to developand implement a second stabilised distributionscheme.

5.5.3 RF Master Oscillator

The RFMO for CLARA will consist of a 10 MHzsource with a Global Positioning System (GPS)reference that will provide a stability of < 2×10−12

(100 seconds Allen variance) and a yearly stabilityof < 5× 10−10. This RF source will be used with aphase locked oscillator and multipliers to produce afrequency for locking to the LMO.


5.5.4 Laser Master Oscillator

The LMO will be a passively mode-locked fibrering laser with a pulse repetition rate at asub-harmonic of the cavity RF frequency. Theselasers are obtainable commercially and provideexcellent jitter performance at high frequencies(>10 kHz). However at low frequencies (<10 kHz),the LMO is prone to drifts in cavity length and willneed to be phase locked to the RFMO to reducedrifts in the LMO repetition rate.

5.5.5 Optical Clock Distribution

The optical reference clock will be distributedacross CLARA through single mode fibre. Thepropagation time through the fibre, which isvulnerable to temperature and humidity fluctuationsand vibrations, will be stabilized by reflecting aportion back from the end of the distributionlink and comparing its return time with referencepulses maintained in the synchronisation room.The comparison is achieved with a balancedoptical cross-correlator (BOXC) which uses aperiodically-polled KTiOPO4 crystal (PPKTP) togenerate a second harmonic signal from the overlapbetween the reference and the return pulses, thusmeasuring their relative timing [126]. The BOXCsignal will be used to feedback on delay controlsin each of the links compensating for any delaychanges.

The fibre and propagation lengths joining thestabilized clock signal to the required end station willnecessarily be outside the stabilisation loop. Theselengths will be kept to a minimum and housed intemperature controlled casing.

5.5.6 Laser-To-Laser Synchronisation

Laser-to-laser locking will be the most important ofthe synchronisation end stations. It will ensure thatthe photoinjector laser, which instigates electronbunch timing, is synchronised to the rest of theaccelerator as a whole. L2L locking will alsobe implemented for the FEL seed laser and theexternal laser to enable synchronisation of theFEL photon output when operating in the seededmode. The output of the slave laser will bereferenced to the delivered optical clock pulses witha two-colour BOXC, which uses sum frequencygeneration between the laser and the clock pulses ina similar configuration to the link stabilisation[127].

The error signal from the BOXC will be used in afeedback loop to adjust the slave laser cavity andlock its timing with respect to the LMO.

5.5.7 Beam Arrival Monitors

BAMs will be implemented to improveunderstanding of the beam dynamics and monitorsynchronisation points such as the FEL entrance.The BAM locations are shown in Figure 5.12.Further details on BAM operation are describedin Section 5.3.

5.5.8 Laser Arrival Monitors

Arrival monitors will be implemented just beforethe gun and at the laser seed entry point. Thepurpose of the first arrival monitor will be toassess the laser jitter level of the source as wellas provide a ‘zero’ time reference for all BAMsdownstream. The seed LAM will be used inconjunction with a BAM preceding the FEL toenable and monitor synchronisation of the bunchseeding. Nonlinear materials will be implementedin a cross-correlator arrangement to detect relativedelay between the delivered optical clock and thesignal to be monitored.

5.5.9 Referencing of LLRF

The LLRF systems will stabilise and regulate theRF fields in the accelerating cavities to reduceamplitude and phase errors. Such errors in thecavity fields would cause an energy modulation onthe electron beam which would convert to timingerrors after the bunch compressor. The RF fieldswill be measured through probes in the acceleratingcavities and stabilised in amplitude through thelocal LLRF feedback system. The LLRF will alsostabilise the measured phase by referencing it tothe transmitted clock through a phase locked loop.Details of the LLRF system are given in Section5.2.3.

The pulsed optical links will need to convertthe delivered optical pulses into an RF signal ontowhich the LLRF systems can be phase locked. Thelaser to RF clock conversion will be done througha low noise photodiode and filtered at 3 GHz forthe LLRF stations. The optical to RF conversionprocess is expected to introduce very little jitter initself, although some may be introduced throughamplification of the required harmonic. The optical


links supplying the LLRF stations will have highertolerances than the timing diagnostics stations. Assuch, it may be considered sufficient for these pulsedlinks to be stabilised through a more cost-effectiveRF harmonic comparison scheme [128].

5.6 Lasers

5.6.1 Seed Lasers for FEL Modulation

Several operating modes call for seed laser pulseswith mid-IR to far-IR wavelengths. It is proposedto generate these from a combination of opticalrectification of a Ti:sapphire system, and differencefrequency mixing of intermediary wavelength laserpulses generated in Ti:sapphire driven OpticalParametric Amplifiers (OPA). Separate experimentsrequire an 800 nm wavelength ultra-short pulse forthe seed, which will be provided from the sameTi:sapphire system used for the long wavelengthgeneration.

The FEL Mode-Locking scheme specifies aseed wavelength between 50-30 µm, as well aspotential experiments at a longer wavelength of120 µm. EEHG experiments are targeted at a seedwavelength of 800 nm.

Generation of 120 µm Wavelength Seed

For modulation at a wavelength of 120 µm,corresponding to a frequency of 2.5 THz, pulsesenergies of 10 µJ or greater are specified(see Chapter 3). Demonstrations of µJ levelTHz generation based on optical rectificationhave been reported, although the focus hasgenerally been on obtaining the highest electricfield strengths in quasi-single-cycle pulses. Incontrast, the requirement of the CLARA seed isfor quasi-monochromatic pulses at approximately2.5THz, with the number of cycles in the pulsematching or exceeding the number of undulatorperiods.

Efficient THz generation by optical rectificationin LiNbO3 is achievable with 800 nm Ti:sapphiretilted pulse front pumping. Pulse energies of 10 µJwere obtained when pumping with 20 mJ of 800 nmpulses [129], although the central frequency of0.5THz (λ = 300µm) is considerably lower thanrequired for the CLARA seed.

Even higher pulse energies have beendemonstrated, with 50µJ reported by Stepanov etal [130], and 125µJ achieved by Fulop et al. [131].

Despite these demonstrations of high power beingat the relatively low frequency of approximately0.5 THz, tunability through the pulse front tilt offerssome potential for 2.5 THz LiNbO3 sources withonly moderately reduced efficiency [132]. High pulseenergy THz pulses with bandwidth extending out to3 THz have been produced by optical rectificationin ZnTe, for example 1.5 µJ THz pulse energy foran excitation with 48 mJ pump energy has beenreported [133].

While the above systems have the advantageof being able to be directly pumped by Ti:sapphirepulses at 800 nm, the limitations on bandwidth(LiNbO3) or efficiency (ZnTe) place these assecondary solutions that would require furtherdevelopment before being considered base-linesources.

An alternative approach is the use of theorganic crystals DAST [134] or OH1 [135], togetherwith an OPA for conversion of the Ti:sapphire tomore optimum wavelength of 1.5 µm. Hauri etal [134] report THz pulse energies of 20 µJ forapproximately 1 mJ pump energy at a wavelengthof 1.5 µm. The spectrum of the THz pulsesextends over 0.5–5.0 THz, with a peak atapproximately 2.1 THz. Whilst this demonstrationwas targeted at short single-cycle THz pulses withhigh field strengths, the demonstrated efficiency(>2% efficiency in energy conversion) and largespectral bandwidth highlight its potential forquasi-monochromatic cycle generation in the targetregion of 2.5 THz.

The baseline 2.5 THz (50 µm) seed systemwill involve THz generation in DAST, pumped by1.5 µm wavelength pulses. The 1.5 µm laserwill itself be generated by an OPA pumped by aTi:sapphire that can deliver up to 20 mJ at the400 Hz repetition rate of CLARA. To provide theTHz central frequency and bandwidth tunability, achirped pulse beating (CPB) arrangement will beused. In the CPB, the broad bandwidth 1.5 µmpulse is linearly chirped to approximately 2 ps, splitinto two pulses which are then recombined with afixed time delay [136, 137]. The time delay betweenthe two pulses provides a constant frequency shiftbetween them, which is converted into the centralTHz frequency when driving the optical rectificationprocess. The bandwidth of the THz pulse can alsobe tuned by the choice of chirped pulse duration.


Generation of 50µm-30µm Wavelength Seed

The seed-modulated Mode-Locked FELexperiments described in Section 3 are based on aseed wavelength of 50 µm, 10 µJ and approximately3 cycles, or 500 fs, in duration. It is also possiblethat with further investigation these experimentsmay be carried out at shorter wavelengths, down toapproximately 30 µm.

While an approach similar to that describedfor optical rectification will be used for the seedgeneration, changes are required as the non-linearmaterial phase-matching differs significantly atthese wavelengths. Phonon resonances, generallyin the region of 6-10THz, give rise to both THzabsorption, and a THz refractive index varyingrapidly with THz frequency. The transverse-opticalresonance in ZnTe lies at 5.3 THz, and rulesout the use of this material for generation of the6 THz seed. A potential 6 THz source proposed byStepanov et al [138] involves optical rectificationof 800 nm pulses in a multilayer DAST/SiO2

structure. While the 800 nm and THz pulsesare not phase matched in DAST, the alternatinglayers of SiO2 allow the different wavelengths tore-phase in a non-active medium, before re-enteringthe subsequent DAST layer for further coherent,in-phase, generation of THz. An efficiencyapproaching 1 % has been calculated. While thisapproach appears promising, further investigationsand experimental validation is required to confirmthe capability.

For higher frequencies (shorter wavelengths)the generated radiation lies above the phononresonances, avoiding problems of THz absorption,and opening further opportunities for differencefrequency mixing. Extremely high electric fieldstrengths exceeding 100MV/cm, over a tunablerange of 10–72THz (λ=30–4µm) have beendemonstrated by difference frequency mixing inGaSe [139]. In this demonstration, differencefrequency generation between two distinctwavelengths of ultra-fast lasers, the first at1100 nm, the second tunable from 1100-1500 nm.Pulse energies of 1.7 µJ at 30 THz (λ = 10µm)have been reported for relatively modest 0.3 mJpulse energy in each of the OPA output laserspulses.

800nm Seed

For the FEL experiments involving sub-50 fs seedpulses, the Ti:sapphire output at a wavelength of

800 nm will be used. The pulse energy requirementwill be significantly below that needed for the far-IRconversion schemes.

Seed Transport

Under the FEL schemes to be tested with CLARA,it is foreseen to only require a single wavelength tobe provided at a given time. The seed transport willbe a single vacuum transport line, where if needbethe optics will be interchangeable between near-IR(800 nm) and its harmonics, and for the transportof the long wavelength sources. To allow forin-situ electro-optic characterisation of the far-IRTHz seed, facility will be made for the simultaneoustransport of the µJ level 800 nm pulses.

5.6.2 Photoinjector Laser System

The photoinjector laser system for CLARA will bethe existing Ti:sapphire system used for VELA,with modifications and enhancements as requiredto satisfy CLARA operating modes. The existingsystem, supplied by Coherent Inc, delivers 11 mJpulse energy at a wavelength of λ = 800 nm, ata repetition rate up to 400 Hz. High conversionefficiency to the λ = 266 nm third harmonic isachieved by using short 45 fs FWHM pulses inthe conversion stage, with UV pulse energy upto 2 mJ possible. The UV pulses have a broadbandwidth exceeding 1.5 nm, and a pulse durationof approximately 180 fs FWHM.

The VELA system does not include temporalpulse shaping. To provide this capability forCLARA, a UV fourier-plane-filtering system similarin concept to that developed by FERMI@Elettra isplanned.

To prevent degradation of the transverse beamquality through non-linear effects, as much of the'14 m long transport path as possible is in vacuum,with final entry into the accelerator beam transportthrough a differential pumping section rather thanthrough a window. The laser transport includesrelay imaging with reflective optics.

Burst Mode for RAFEL

The RAFEL mode of CLARA requires a train of∼20bunches, of up to 25 pC per bunch, with the spacingof approximately 120 ns set by the cavity round triptime. The total bunch charge required is withinthat deliverable with the Ti:sapphire and a copper


cathode. One option for delivering this multi-bunchstructure is using an arrangement of beam splittersand optical delays lines. Maintaining a stable opticalpulse train with a sequence of optical delay linesof (round trip) length 35 m will have significantengineering challenges. Controlling the transversediffraction of the beams will require an imagingsystem within the delay lines. Whether achieving therequired stability and intra-pulse variation is feasiblein practise requires further investigation.

If the delay line approach proves not to befeasible, the alternative would be to provide adifferent laser system that operates in the required8.5 MHz burst mode. Laser systems have beendemonstrated that provide confidence that suchan option is technically feasible. On the ALICEenergy recovery linac, a Nd:YVO4 system operatesat 81.25 MHz, and is able to provide pulse trainsof length from single pulse, up to 100 µs, andwith repetition rate selectable from 81.25 MHzto 1.2 MHz (in discrete steps of 81.25/n, n =

1 . . . 64). The FLASH and XFEL photoinjectorsystems at DESY demonstrate the production ofhigher pulse energy trains at a repetition rate of 1MHz. High pulse energies and ps pulse durations,in trains with approx 6 MHz repetition rate, are alsobeing developed for laser wire applications. Thechoice of approach for a burst mode laser systemwill be tightly coupled to the choice of photocathodematerial, and cannot be made at the current time.

5.6.3 Lasers for FEL PhotonDiagnostics

The temporal characterisation of the FELUV output will be via difference frequencycross-correlation with ultra-fast Ti:sapphire pulses.Initially the laser pulses at a wavelength of 800 nmor its harmonics will be split from the Ti:sapphiresystem that generates the FEL seed radiation, withpulse energies of less than 1 mJ expected to berequired.

5.6.4 Laser Synchronisation

The photoinjector and seed laser system willbe locked to the accelerator RF thorough theoptical timing distribution system. The timingof the 1.5 µm wavelength optical pulses that aredistributing the RF clock, and the pulse train of the75-100 MHz oscillators of Ti:sapphire systems willbe locked with a balance optical cross-correlator,

and a feedback loop acting on the Ti:sapphireoscillator cavity length.

The selection of individual pulses from theRF-locked oscillator train for amplification requiresonly ns stable triggering. However, additional pslevel drifts and timing noise can be introduced fromthe laser amplifiers and the laser transport. LAMs,based on optical cross-correlation between theamplified pulses and the optical timing distribution,will be located near the gun and the seed injectionpoint. The LAMs will be used to correct drifts inthe timing of the amplified laser pulses.

5.7 Undulators

As explained in Chapter 3 there are threedistinct undulator systems planned for CLARA; themodulator, the radiators, and the afterburner. Ourstudies so far have focussed on the modulator andradiator sections as the afterburner is not a corepart of the facility and will likely not be satisfiedby just one device type but will itself be changedto suit the particular experiment being proposed.The specifications for the modulator and radiatorundulators are given in Table 5.7. The modulatormust cover a wide range of seed input wavelengths,leading to a large K requirement. Fortunately oneof our existing insertion devices, available from thedecommissioned Daresbury SRS, seems very wellsuited to meeting these requirements and we arecurrently assuming that this device can be used forthis purpose, further FEL studies will be carried outto confirm this in the future. The SRS device [140]has a period length of 200 mm and 4 periods andwill be able to achieve the required K value easily.One possible issue which will be considered furtheris that to operate at 800 nm the magnet gap willbe rather large (∼200 mm) and so fringe fieldsmust be carefully assessed and magnet clamp platesretrofitted if necessary.

Another option which will be considered in thefuture is to operate the modulator over a morerestricted wavelength range but to then seed on ahigher harmonic of the device. In order to interactwith the 800 nm input laser the modulator could betuned to a fundamental of 4000 nm, for example,and the seed interact at the 5th harmonic. Ademonstration of the principle has previously beenmade at UCLA [141].

The radiator undulators will be planar hybridtypes utilising NdFeB permanent magnets and steel


Modulator Radiator

Type Planar Planar, Vertical Linear PolarisationNumber of Modules 1 7Module Length (m) ∼1.0 1.5Period (mm) ∼200 27Number of Periods 4 53Minimum Gap (mm) 20 6Minimum Wavelength (nm) 800 100Maximum Wavelength (nm) 120000 400Maximum K ∼25 3.3

Table 5.7: Specifications for the modulator and radiator undulators.

poles. In order to optimise the optical transmissionof the FEL photon output into the diagnostic area,which requires large angle horizontal reflections,the radiators will be designed to generate verticallinear polarisation. In practice this means that themagnet arrays will generate a horizontal magneticfield and the gap adjustment will also be in thehorizontal plane. It is not envisaged that thisarrangement will cause any particular engineeringdifficulty when compared to the more commonvertical gap adjustment mechanisms. Note thatthe modulator polarisation is independent of thisspecific radiator polarisation requirement since itis the electron modulation effect caused by thelaser-modulator interaction which is required, andnot the actual radiation generated in that section.We are assuming a modulator with horizontal linearpolarisation at present.

The radiator undulators are separated by theinter-undulator sections. These sections mustcontain several pieces of equipment to ensure theoptimum operation of the FEL, the total spacerequired for each section is presently assumedto be 1.1 m. Each section will contain aquadrupole to match the electron beam propertiesalong the length of the FEL, a phasing unit toensure the radiation phase from each module canbe independently adjusted, horizontal and verticalsteerers to maintain good electron beam orbitcontrol, an electron and photon beam screendiagnostic, and a cavity BPM. Space must alsobe included in the design for vacuum pumps andgauges. This space allowance is consistent withother FEL projects.

Detailed resistive wall wakefield studies must becarried out for the different bunch types interactingwith the narrow gap undulator vessel. Initial

simulations for the Ultra-short mode, which isexpected to be the worst case, with a peakcurrent of over 1500 A and 100 pC bunchcharge show that the induced energy spreadis significant and could impact on the FELperformance at short wavelength. Further studieswill be carried out to assess more carefully theenergy spread change in detail as an inducedlinear chirp could be compensated by undulator gaptapering, for example. Mitigation strategies existif the FEL performance is disrupted significantly.The undulator vessel size could be increased, forexample, as this is entirely compatible with 100 nmoperation.

As well as requiring undulators for the FEL,CLARA will be an excellent facility for testingadvanced undulator designs. ASTeC has a strongtrack record in developing short period, narrow gap,permanent magnet undulators [142] and also inhigh-field superconducting undulators [143]. Exoticideas, such as RF or laser driven undulators, whichwould be too risky for an operating user facility totry, can be tested on CLARA.

5.8 Control System

5.8.1 Introduction

The control system plays a central role in allaccelerator facilities. The CLARA control systemwill be a facility wide monitoring and control systemintegrating all parts of the CLARA accelerator.The control system will extend from the interfaceof the equipment being controlled through to theoperator, technical expert or physicist. It will includeall hardware and software between these bounds


including computer systems, networking, hardwareinterfaces, programmable logic controllers (PLCs)and fieldbuses.

The Personnel Safety System (PSS) will beinterfaced to, and monitored by, the CLARA controlsystem. Timing and synchronisation of the lasersand RF systems is covered in Section 5.5 and isnot considered to be part of the control system.A separate event synchronisation system will beimplemented as part of the control system to ensuresynchronisation of critical control operations and toprovide a common high resolution time-stampingcapability.

The CLARA control system will benefit fromexisting control system equipment standards andknowledge gained from the recently implementedVELA facility.

5.8.2 Architecture

The CLARA control system will use the EPICSsoftware toolkit. EPICS has been successfullyapplied at Daresbury in the ALICE, EMMA andVELA facilities and on numerous acceleratorprojects worldwide. EPICS is a proven softwaretoolkit with well-defined interfaces at both the clientand server and will enable fast integration anddevelopment. Due to its collaborative nature, usingEPICS enables us to take advantage of work doneat other laboratories.

The various elements of the control system willbe connected to a high speed Ethernet local areanetwork. This will consist of several class C subnetsconnected to the main site network. Access willbe provided to the control system data via EPICSgateway systems running access security.

Whenever possible, wireless access will beprovided in the accelerator complex to enable quickcommissioning and maintenance to be carried outusing laptop/tablet PCs.

The hardware interface layer of the controlsystem will provide the connection to the underlyingsub-systems. This will consist of a modularhardware solution of commercial rack mount PCsand VME systems running the Linux operatingsystem connected to a range of I/O types directlyor via selected fieldbuses.

Application software will be installed on centralfile servers to ensure consistent operation from anyconsole on the control system network. The controlsystem will be implemented in such a way that itcan be incrementally expanded and upgraded as

the project proceeds. The control system must beable to be extended to new hardware and softwaretechnology during the project phase and later.

A number of EPICS Channel Access clientinterfaces will be supported for controls and physicsapplication development in high-level languages.These will include C#/VB via .NET, Matlab,Mathematica and Python.

5.8.3 Controls Hardware

The network hardware and commercial PC hardwareof the CLARA control system will use standardsdefined by the STFC Network and IT Supportdepartments.

As far as possible controls hardware used on theVELA facility should be preferred to take advantageof existing expertise and proven hardware standards.This hardware includes:

• Commercial rack mount PC systems runninga Linux operating system with real-timekernel: Existing EPICS support is available forinterfacing many I/O types including analogue,digital and serial (RS232/422/485) along withstatus/interlocking via Ethernet connected PLCs.

• Omron CJ series PLCs: Status and interlocksystems for digital control and componentprotection of vacuum subsystems, magnets,power supplies etc.

• Motion control systems: Fieldbus based systemsinterfaced to the control system via EtherCAT.These shall be jointly developed in collaborationwith PSI for use also on the SwissFEL facility.

Any special control requirements would be solved byusing dedicated real-time controllers (for exampleVME based or FPGA based). Examples ofsuch systems could include geographically isolateddevices or systems that require the computing speedof a dedicated CPU. Any dedicated controllerswould be integrated into the EPICS environmentand software build processes.

5.8.4 Timing and SynchronisationSystem

One of the major challenges with a networkdistributed control system is the synchronisationand accurate time-stamping of events spread acrossseveral physically and logically separated systems.


An example of this is the simultaneous measurementof beam diagnostic information along an entirebeam transport system. This operation will typicallyinvolve collecting data from a number of front-endserver systems.

Commercially available synchronisationhardware, as used on SLS, LCLS and other projects,is available for a number of platforms and providestiming outputs with a sufficient resolution forCLARA applications. It will be important to ensurethat the master clock used by the control system islocked to the RF and laser systems of CLARA.

5.8.5 Interlock Systems

The control system will provide a comprehensiveinterlock and protection system ranging from theenforcement of sensible operating limits rightthrough to the protection of the whole acceleratoragainst excessive beam loss. This section excludesthe PSS which is described in Section 6.2. Notethat full monitoring of the PSS will be provided bythe control system.

High Integrity

This will be used to provide high-speed andfail-safe local protection in situations where seriousdamage to equipment is likely to occur. A typicalexample would be protection of undulator magnetarrays from excessive beam losses. Embeddedmicro-controllers could be used for this purpose.

Routine

This level of protection is intended to preventminor damage to individual machine componentsor sub-systems. The Omron CJ series PLC basedinterlock system developed for the VELA facility willbe used for this purpose.

Both interlock systems will operateindependently of EPICS with monitoring and control

requests marshalled to the interlock systems viaEPICS for processing via internally programmedlogic.

In addition, but not described here, there willalso be local systems that supervise individualcomponents such as RF structures.

5.8.6 Feedback Systems

Digital feedback will be required in a number ofplaces on CLARA. The primary requirements are asfollows:

Global Transverse Trajectory Stabilisation

This will take the transverse beam positioninformation from strip-line BPMs spread throughoutthe machine along with high-precision cavity BPMsin the bunch compressor, and derive correctionsettings to remove the effects of thermal drift,mechanical vibrations, electrical noise and othersources of disturbance.

Undulator Transverse Trajectory Stabilisation

The stability requirements within the undulatorswill be particularly demanding. Transverse beamposition must be stabilised in order to maintain FELgain. Cavity BPMs will be used to provide positionalreference information.

Low latency feedback systems will be requiredto satisfy the transverse stability requirements ofCLARA on a shot-to-shot basis.

Stabilisation of the RF Fields within theAccelerating Structures

The requirements for a proposed digital LLRFsystem are described in Section 5.2.3. Thenecessary control loops for RF stabilisation will beimplemented within the LLRF system and will betightly integrated into the control system.



6Radiation Safety

6.1 Shielding Requirements

6.1.1 Radiological Classificationof Areas

A Radiological Classification of Areas considers allidentified areas within a given facility and specifiesa radiation zone for each area based on discussionswith the designers of the facility, with regard to thefollowing key drivers from the perspective of dosecontrol:

• Expected occupancy levels,

• Prompt radiation levels during electron beamoperations,

• Residual radiation levels following beamshutdown.

Minimisation of exposure to ionising radiationis aided by the classification of facility layoutsaccording to the radiation levels expected in eacharea and the access or restriction requirements.These classifications assist in the production of arational shielding design and dose saving strategythat will facilitate normal operational dose control.At this stage of the CLARA project, a fullRadiological Classification of Areas has not beendeveloped. In order to determine the bulk shieldingrequirements annual whole body effective dosesof 1 mSv within the site boundary and 0.3 mSvoutside of the site boundary have been assumed inaccordance with the relevant STFC Safety Code.Taking account of reasonable occupancy levels tothe side and on the roof of CLARA the followingdose rate criteria have been assumed:

• 1 µSv/hr at the cold-side of the shieldedenclosure walls;

• 2.5 µSv/hr at the cold-side of the shieldedenclosure roof.

6.1.2 Source and Material Data

All of the source terms are directly derived fromthe operating modes described in Chapter 3. Threescenarios have been assessed as these are thoughtto be those which represent realistic worse casesituations. These are:

• Beam Dump: 250 MeV electrons, 25 pC×20,100 Hz, 12.5 W,

• Beam Dump: 100 MeV electrons, 250 pC,400 Hz, 10 W,

• Beam Collimator: 130 MeV electrons, 25 pC×20,100 Hz, 6.5 W.

In order to perform shielding calculations it isnecessary to specify the isotopic compositions anddensities of the shielding materials employed. Allstandard elemental compositions (e.g. Iron) havebeen taken from the FLUKA code [144]. Thecomposition of concrete assumed is a pessimistic,dry composition, with a density of 2.32 g/cm3.A simple, non-optimised, beam dump has beenassumed which is an iron cylinder of 15 cm radiusand 55 cm length with a 20 cm recess where theelectrons first strike the material. This iron is thenencased within a concrete cylinder of 45 cm radiusand 85 cm long.


6.1.3 Shielding Calculations

Dose rates have been assessed on the exterior of theshield surface where the maximum dose rate wouldbe experienced whilst CLARA is operating. Alldose rates have been calculated using the FLUKAMonte Carlo code. Independent cross-checks ofcalculations have been performed using MCNPX [145]and these demonstrate good agreement with theFLUKA calculations.

The results for the beam dump scenarios showthat the 250 MeV case is bounding when comparedwith the 100 MeV case. Therefore, bulk shieldingrequirements will be driven by the 250 MeV case.The results indicate that 150 cm of concrete wouldbe adequate to reduce dose rates to less than1 µSv/hr (i.e. through the walls), or 125 cmto reduce the dose rate to less than 2.5 µSv/hr(i.e. through the roof). Beam dump optimisationcould result in further reductions in bulk shieldingthickness.

For the collimator scenario an impact angle of11.3 was assumed, corresponding to a failure ofa dipole magnet and the electron beam strikingthe steel vacuum vessel in an uncontrolled manner.Note that this scenario assumes all of the beamis lost so it would be a fault condition and not along term operational scenario. In this case thecalculations indicate that 160 cm of concrete wouldbe required for the walls and 150 cm for the roof.If a continuous beam loss of 10% is assumed asopposed to 100%, the shielding thicknesses reduceto 130 cm (walls) and 115 cm (roof). Note that noadditional local shielding has been assumed in thesecollimator calculations and so these remain ratherpessimistic.

In summary, for the scenarios considered sofar, the simulations suggest that a concrete wallthickness of 150 cm and roof thickness of 125 cmshould ensure that the dose rates are withinacceptable limits. Future studies will considerfurther optimisation of the beam dump and otherlocations where fractional continuous beam lossis expected, such as the collimator sections. Inaddition, the personnel labyrinth locations andgeometries will need to be modelled and verified aswell as the penetrations for cabling, waveguide, etc.Consideration will also be given to the anticipatedlevels of dark current and their sources to ensurethat this additional source does not impact on thebulk shielding definition.

6.2 Personnel Safety System

6.2.1 Purpose

The purpose of the PSS for CLARA is toprovide engineering control measures and interlocksthat work together with radiation shielding andadministrative procedures to ensure the safety of allpersonnel working on, or in close proximity to, thefacility. The high-level functionality of the systemmust meet the legal requirements laid down inthe Ionising Radiation Regulations, 1999 (IRR99).Also, an Approved Code of Practice (ACOP) toIRR99 provides guidance and advice on the design,operation and test of compliant systems.

6.2.2 Design Requirements

In order to meet the requirements of IRR99 andits ACOP a number of key design features needto be present in the CLARA PSS. These can besummarised as:

• Warning devices giving a clear indication of theoperating mode,

• Fail-safe Interlocks to ensure that persons cannotgain access to areas where exposure to ionisingradiation is possible,

• A monitored search procedure to ensure that noone is trapped in the area when ionising radiationis being generated,

• Emergency Off buttons,

• Radiation Monitors to ensure that radiation levelsare not significantly above expected design levels,

• Key activated permit switches to provide a meansof easily authorising or disabling the facilityor selected sub-systems. Also, over-ride keyswitches to permit, for example, low power RFto be enabled while trained staff are performingmeasurement and test within the machine.

Several additional design features coverfunctionality that is not required by legislation butwhich significantly improve operational usability:

• Integration with the CLARA control system sothat the state of all interlocks and permits can bemonitored, displayed and recorded,

• Use of permanently installed radiation monitorsto ensure that measured radiation levels do notsignificantly exceed pre-determined limits,


• Use of modular hardware to simplify futureexpansion, fault-finding and maintenance,

• Provision of ‘Limited Access’ operation to allowcontrolled access to one or more machine areasunder safe conditions without compromising theintegrity of the search process.

6.2.3 Implementation

In order to meet these requirements the PSSfor CLARA will use a Commercial-Off-The-Shelf

safety-rated PLC system to provide a modular,programmable and expandable system. Thistechnology has been available for a number ofyears and is widely used in manufacturing industryto implement high-integrity safety systems forproduction lines, chemical processes and dangerousmachinery. Although the use of such systems forradiation protection is relatively new this approach iscurrently being used on VELA and initial experiencesare very encouraging so it is expected that a suitablyscaled system will be a good solution for CLARA.



7Potential Upgrades and Future Exploitation

The primary motivation for CLARA is firstly asan FEL test facility, but the availability of brightbunches of electrons at a range of energies mayalso enable other challenging areas of acceleratorresearch to be developed which would otherwise notbe possible. This chapter puts forward a numberof areas which have come to our attention duringthe design of CLARA and that are worthy of furtherstudy. There may well be other applications possiblewhich could be proposed in the future, so theideas described here should be treated as a sampleof possible applications, and not as a well-definedpriority list or roadmap.

To enable these opportunities to be feasiblein the future, space has been reserved within theCLARA shielding area for a 250 MeV beam transportline with beam extraction from CLARA, close toor possibly part of the high-energy spectrometerline. The shielded enclosure has room for at leastone beamline parallel to the FEL and the option oftaking additional beamlines into a new shielded areais also practical within the confines of the existingbuilding.

7.1 Plasma AcceleratorResearch

High-gradient plasma accelerators hold the promiseto minimize the scale and cost of futureaccelerators: it is possible to achieve acceleratinggradients from 1 GV/m to 100 GV/m usingrelativistic plasma waves. Beam-driven plasmawakefield acceleration has achieved many significantbreakthroughs [146, 147]. Proton-driven wakefieldacceleration is proposed in the AWAKE project

[148] to demonstrate high gradients for futurehigh-energy colliders, as well as proposals suchas LUNEX5 [149] which seeks to demonstratelasing using a laser wakefield accelerated beam.This ongoing research will give significant boost toplasma accelerator research.

The beam parameters of CLARA are ideal todemonstrate plasma wakefield acceleration. APlasma Accelerator Research Station (PARS) basedat the CLARA facility will allow us to study thekey issues in electron-driven wakefield acceleration.This will be the first UK-based beam drivenplasma wakefield accelerator facility and will yieldinformation essential for longer-term developmentof such facilities [150].

Preliminary studies show that a 250 MeVelectron will obtain energy doubling with a ∼20 cmplasma cell. One-bunch electron-beam drivenwakefield acceleration could be demonstrated inany of the three operating regimes of CLARA:long pulse (bunch length: 250–800 fs); shortpulse (250 fs); and ultra-short pulse (30 fs), albeitusing different bunch charges. The diverse beamoperating modes of CLARA will allow the testingof scaling laws of plasma wakefield acceleration. Itwill also be possible to test many other interestingconcepts such as: two-bunch experiments (craftingtwo bunches using a collimator, one for drivingthe plasma wakefield, and the other for samplingthe wakefield); beam shaping studies for hightransformer ratio (beam density profile shapingby shaping the laser pulse—multiple bunchesor hard-edged beams can be near-ideal drivers);self-modulation of a long electron bunch canprovide inputs to the CERN proton-driven PWFAexperiment, AWAKE; use of CLARA as an electron


injector for laser-driven wakefield acceleration(combined with LWFA research).

The beam will be switched to PARS at a fullenergy of 250 MeV. The proposed dogleg beamline design using π transform between the dipoles(using two FODO doublets) limits the transversebeam emittance blow up which arises from coherentsynchrotron radiation; studies have assumed a250 pC bunch charge and 250 fs bunch length,or 20 pC at 30 fs. The possibility of usingan additional take-off line at lower angle ('6)from the 30 dipole for a high-energy diagnosticsspectrometer is also being investigated. Theproposed beamline is contained within the CLARAshielding area with a transverse centre-to-centreoffset of approximately 1.5 m. In addition toPARS, this beamline could also be used for otherexperiments mentioned in this section as well asother future exploitation of the high-quality beamfrom CLARA.

7.2 Ultrafast ElectronDiffraction

Electron beams with sub-angstrom wavelengthand sub-100 fs bunch lengths, available fromphotoinjectors such as the CLARA gun, providea cost-effective alternative to X-ray FELs fortime-resolved structure determination. Ultrafastelectron diffraction (UED) provides spatialinformation with atomic resolution on the time scaleof the making and breaking of chemical bonds.

Similar photoinjector based acceleratorsfor UED are being developed at DESY(Germany) [151], UCLA (US) [152] and OsakaUniversity (Japan) [153] amongst others, togetherwith the requisite detection and diagnosticsinstrumentation. The development of such a facilityformed a core part of the NLS science case [17]where the accelerator provided the probe electrondiffraction beam to monitor structural change as itevolved.

The initiation of change is with a pump laserderived from the photoinjector laser system andtherefore a fundamental requirement is to obtainsufficiently high quality diffraction data in a singleshot, as in most cases the sample is irreversiblychanged by the pump laser. This can be achievedwith only ∼1 pC charge.

The basic layout for UED is a sample chamberlocated at a position where the electron beam has

low energy (few MeV), low divergence (< 50µrad)and short bunch length (<100 fs). The chamberallows a pump laser to be focussed onto thesample such that it is coincident with the electronbeam. An electron collimator before the sample willallow beam shaping to provide a beam diameterof ∼0.5 mm at the sample, smaller than thediameter of the pump laser spot. This collimatorwill be sufficiently thick to prevent X-rays generatedwithin it emerging and contaminating the diffractionpattern on the downstream fluorescent screen; thisscreen is viewed with an image-intensifying CCDcamera that allows single electron imaging [154].

The CLARA design allows the first linac to beused for velocity bunching, and the sample chambercan be located before Linac 2, which would be usedas a drift section for the diffracted beams to betransported to the detection screen; the requireddiagnostics include beam arrival time monitoring.Key developments (which are also being pursuedelsewhere, see for example Byrd [155]) are to reducethe RF jitter which limits the ultimate temporalresolution.

7.3 Compton PhotonProduction

It will be possible to generate photons withvery high energies by Compton-scattering anintense longer-wavelength laser with the high-energyCLARA electron bunches: the naturally-shortbunches will mean that, even in a head-on scatteringgeometry, the generated photons will be confinedto pulses as little as 100 fs in length. 800 nmphotons from a Ti:sapphire laser scattering fromCLARA electrons at up to 250 MeV will generate amaximum photon energy around 1.5 MeV. By usinga longer wavelength laser or lower energy electrons,the photon energy can be easily reduced to the keVregion.

Dedicated Compton scattering facilities, suchas ThomX in France [156] and ELI-NP [157] inRomania, are now being developed, demonstratingthat there is a demand for short pulses of photonsin the keV to MeV region. The electron beambrightness of CLARA is very competitive with theseprojects and so it will be capable of generatingsimilar quality output.


7.4 Dielectric WakefieldAcceleration

Research into dielectric wakefield acceleration(DWA) has steadily increased over the pastdecade. This interest is driven by the prospect ofdeveloping advanced accelerators with acceleratinggradients greater than ∼ 1 GV/m that willdramatically reduce the size and cost of highenergy linear colliders, and that will also lead tothe development of compact light sources. DWAconcepts compare favourably with traditional metalaccelerating structures, which are typically limitedto gradients less than 100 MeV/m.

The wakefield in a DWA structure is generatedby a drive bunch of relatively low energy (tensor hundreds of MeV), short length (< 1 ps) andlarge bunch charge (many nC). A trailing bunchwith the correct phase with respect to the drivebunch will then be accelerated by a high gradientwakefield [158]. The DWA structures currentlybeing investigated are of simple cylindrical or planardesigns with sub-mm apertures and fundamentalfrequencies in the THz range. Recent experimentalprogress in DWA research is impressive. Breakdownlimits of 14 GV/m on the dielectric surfaceof a cylindrical dielectric structure have beendemonstrated [159]. Trailing bunches have beenaccelerated in planar DWAs, albeit so far withmodest ∼10 MeV/m gradients [160, 161]. Strongwakefield-induced energy modulation of the longerbunches has been demonstrated, the aim beingto generate density-modulated electron beams onthe sub-ps scale for production of coherent THzradiation [162]. A number of other experimentaland theoretical DWA-related studies have beenconducted. New enhanced DWA structuresare under development and several experimentalprogrammes are proposed, including those to beconducted at AWA (Argonne), FACET (SLAC)and FLASH (DESY). Near-term R&D encompassesthe development of accelerator systems that canprovide independent pairs of drive/witness bunchesfor DWA experiments, methods of shaping thedrive bunch to increase the so-called transformerratio (dual-frequency linacs, emittance exchangemechanisms etc.), and the construction andcharacterisation of novel DWA structures. Forexample, DWA structures are being investigated atAWA [163] as an alternative to metal RF structuresin two-beam accelerator schemes such as CLIC.

The CLARA design parameters (up to 250MeVbeam energy, sub-ps bunch lengths, low transverseemittance) and its flexibility in beam manipulationare ideal for conducting DWA-related R&D. Sincethe majority of the near-term DWA studies willbe proof-of-concept experiments, the sub-nC bunchcharges from CLARA should not be a particularlydetrimental factor as long as the wakefield effectscan be properly characterised. There are presently alimited number of suitable test facilities worldwide atwhich DWA research can be conducted. Thereforethe initiation of DWA R&D on CLARA wouldpotentially attract collaborations both nationally(with for example the Cockcroft and John Adamsinstitutes) and internationally. Dielectric wakefieldacceleration is a new and exciting area ofresearch with far-reaching practical applications; itcomplements the mainstream FEL-related studiesintended on CLARA, thus increasing this test facilityefficiency and elevating its profile in the internationalaccelerator physics community.

7.5 NonequilibriumElectron Rings

Although 4th-generation sources of synchrotronradiation are primarily considering the use of FELsto provide ultra-bright pulses of photons to users,there remains a strong demand for radiation withother properties such as high average intensity,or which give other wavelengths. Indeed, theUK’s first 3rd-generation source - DIAMOND -was only brought into operation six years ago,and is just entering its final phase of insertiondevice installation before it reaches full capacity.A programme of studies is underway to explorethe possibility to upgrade the DIAMOND latticewith completely new magnets to give a so-called’ultimate storage ring’ with very small emittancefor users. Similarly, several UK-led projects haveproposed construction of facilities (such as DAPS,4GLS and NLS) to service the VUV and softX-ray research communities with radiation sourcesoptimised for their needs.

Recent work carried out by ASTeC and theCockcroft Institute [164] has suggested a newapproach using nonequilibrium (NEQ) rings. Ratherthan allowing electron bunches to reach equilibriumas occurs in a normal electron storage ring,nonequilibrium rings store bunches for only a fewthousand turns and use ultrafast kickers (already


under intensive development for use in particlephysics collider damping rings) to inject and extractthem from the ring. With sufficient kicker speedand a good electron injector with high repetitionrate, currents of up to 100 mA or more may inprinciple be achieved, the ring giving intense beamsof photons with smaller emittances than would beachievable from a storage ring of the same size,but without the beam power problems inherent inhigh-energy ERLs. For example, ring designs similarto the Swedish MAX-II (90 m) or MAX-III (36m) designs may be used to generate photons ofenergy ∼ 2 keV but with 1 nmrad emittance. Evenshorter wavelengths are possible by incorporatingsuperconducting dipoles into the ring optics.

CLARA’s initial energy and repetition rateare insufficient to deliver high currents or shortwavelengths, but a first demonstration of thenonequilibrium principle could be conducted usingthe 250 MeV electron energy; a future upgrade ofCLARA to energies of 500 MeV or greater, andrepetition rates great than 100 Hz, would allow afull test of a non equilibrium ring for synchrotronradiation production. A test electron ring couldalso be used for other purposes, as detailed in thefollowing section.

7.6 Exotic Storage RingConcepts

A number of interesting demonstrations wouldbe enabled by a small electron ring ∼50 mcircumference that used CLARA as an injector.Firstly, an electron storage ring with electron energyof a few hundred MeV, in which electron polarisationnaturally develops in the stored beam over a fewhours, could be used as a calculable source ofsynchrotron radiation in the UV/VUV photon rangeby standards bodies such as the National PhysicalLaboratory; similar rings such as the GermanMetrology Light Source [165] have recently beenconstructed for just this purpose. The electronenergy may be measured to high precision using anow-standard method to synchronously depolarisethe beam.

Secondly, an electron ring of modest energycould be used to provide the first experimentaldemonstration of optical stochastic cooling (OSC)in an electron beam [166, 167]. Also proposedas a method to improve the beam brightness inhadron colliders, OSC could be used as a method to

reduce the beam emittance in electron storage ringsto improve the brightness of synchrotron radiation[168]. NEQ operation may be used to alleviatethe increased Touschek scattering and short lifetimebrought about by such small beam emittances, butanother method which may be explored is the recentwork to design the so-called Integrable Optics TestAccelerator (IOTA) [169, 170]. Proposed as anadd-on to the Fermilab ASTA project, IOTA wouldutilise special nonlinear magnetic lattice elements togive a storage ring with very large dynamic aperture,allowing scatted electrons to survive to damp backdown into the beam core. Combined with beamcooling this could be an alternative approach toproviding good beam quality and long beam lifetimein a low-energy electron ring. It is possible that asingle electron ring design could be used to test theNEQ, OSC, and IOTA principles at CLARA, whilstalso being used as a standards radiation source.

7.7 Industrial Exploitation

Whilst fundamental enabling research into nextgeneration light sources remains the core driverof the CLARA facility, it will also be possible toharness the unique characteristics of the facility forindustrial exploitation. Extensive market researchfor VELA [24] has identified that a high-powerbeam at energy up to 100 MeV would be optimum.Operation at 100 MeV, rather than 250 MeV, allowsthe maximum repetition rate to be increased from100 Hz to 400 Hz without placing excessive thermalload on the accelerating structures. The bunchcharge can be up to 250 pC to maximise the beampower if required and the other parameters can beoptimised to meet the requirements of individualusers. CLARA will build on the successful industrialengagement processes which are currently beingimplemented for first exploitation of the VELAaccelerator facility in the summer 2013, wherebya variety of technology development groups haveidentified an effective route to utilise the flexibleelectron-beam delivery facility in order to validatenew technologies or processes. Being able tooffer similar levels of beam performance at higherbeam energies and intensities will be the primaryadvantage of CLARA for industrial exploitation.

The fundamental accelerator technologysystems required for the majority of the industrialaccelerators globally can be effectively developedand demonstrated utilising the VELA/CLARA


facilities: from electron beam sources, beamacceleration and diagnostics systems, magnetsand UHV vacuum components, laser optics, beamcontrol and feedback systems. The VELA MarketResearch Survey [24] identified a number ofapplication areas where the increased electronbeam energy of CLARA may enable access tomarkets beyond those currently enabled throughVELA. These include a further enhancement ofcargo scanning applications, enabling access to agreater range of treatment-applicable radioisotopes,high-energy medical sterilisation and a broaderapplicability of electron beams to environmentalclean-up and purification challenges, all of whichrequire accelerator technology sub-systems whichcan be developed and demonstrated using theCLARA experimental facility.

Whilst the exploitation of the photon outputfrom CLARA for commercial use is not coveredwithin the scope of the project, the inherentflexibility of the facility layout and infrastructuredoes not preclude future commercial photonexploitation work, should a suitable application andbusiness case arise. From the broader perspectiveof the ASTeC facility portfolio, establishing theCLARA facility will further enhance the capabilitiesof the existing VELA facility by driving performanceimprovements to the equipment installed in theshared accelerator hall, and through the transferof electron beam optimisation and developmentwork to the new facility, thereby enabling agreater proportion of the VELA running time to bededicated to industrial exploitation.



References

[1] P. Emma et al. First lasing and operation of an angstrom-wavelength free-electron laser. Nat.Photon., 4:641 – 647, 2010. http://www.nature.com/nphoton/journal/v4/n9/abs/nphoton.2010.176.html.

[2] H. Tanaka et al. A compact X-ray free-electron laser emitting in the sub-angstrom region. Nat.Photon., 6:540–544, 2012. http://dx.doi.org/10.1038/nphoton.2012.141.

[3] M. Altarelli et al (eds). XFEL: The European X-ray free-electron laser technical design report.Technical Report DESY 2006-097, DESY, 2007.

[4] SwissFEL Conceptual Design Report. Technical Report PSI Bericht Nr. 10-04, Paul Scherrer Institut,April 2012. ftp://ftp.psi.ch/psi/SwissFEL_CDR/SwissFEL_CDR_V20_23.04.12.pdf.

[5] W. Ackermann et al. Operation of a free-electron laser from the extreme ultraviolet to the waterwindow. Nat. Photon., 1:336–342, 2007. http://www.nature.com/nphoton/journal/v1/n6/suppinfo/nphoton.2007.76_S1.html.

[6] C.J. Bocchetta et al. FERMI@Elettra Conceptual Design Report. Technical Report ST/F-TN-07/12,Sincrotrone Trieste, 2007.

[7] P. Emma et al. Femtosecond and Subfemtosecond X-ray Pulses from a Self-Amplified SpontaneousEmission-Based Free-Electron Laser. Phys. Rev. Lett., 92:074801, 2004.

[8] P. Emma et al. Attosecond X-Ray Pulses in the LCLS Using the Slotted Foil Method. In Proc. ofFEL 2004, 2004.

[9] E. L Saldin et al. Terawatt-scale sub 10-fs Laser Technology - Key to Generation of GW-levelAttosecond Pulses in X-ray Free Electron Lasers. Opt. Commun., 237:153–164, 2004.

[10] E. L. Saldin et al. A New Technique to Generate 100 GW-level Attosecond X-ray Pulses from theXray SASE-FELs. Opt. Commun., 239:161–172, 2004.

[11] A. A. Zholents and G. Penn. Obtaining Attosecond X-ray Pulses using a SASE-FEL. Phys. Rev.Spec. Top. Accel. Beams, 8:050704, 2004.

[12] A Zholents. Methods of Attosecond X-Ray Pulse Generation. In 2005 Particle AcceleratorConference, pages 39–43. Ieee, 2005. http://ieeexplore.ieee.org/lpdocs/epic03/wrapper.htm?arnumber=1590353.

http://www.nature.com/nphoton/journal/v4/n9/abs/nphoton.2010.176.html


http://dx.doi.org/10.1038/nphoton.2012.141

ftp://ftp.psi.ch/psi/SwissFEL_CDR/SwissFEL_CDR_V20_23.04.12.pdf

http://www.nature.com/nphoton/journal/v1/n6/suppinfo/nphoton.2007.76_S1.html

http://www.nature.com/nphoton/journal/v1/n6/suppinfo/nphoton.2007.76_S1.html

http://ieeexplore.ieee.org/lpdocs/epic03/wrapper.htm?arnumber=1590353

http://ieeexplore.ieee.org/lpdocs/epic03/wrapper.htm?arnumber=1590353


[13] E. Saldin, E. Schneidmiller, and M. Yurkov. Self-amplified spontaneous emission FEL withenergy-chirped electron beam and its application for generation of attosecond x-ray pulses. PhysicalReview Special Topics - Accelerators and Beams, 9(5):1–6, May 2006. http://link.aps.org/doi/10.1103/PhysRevSTAB.9.050702.

[14] A. A. Zholents and M. S. Zolotorev. Attosecond X-ray pulses produced by ultra short transverseslicing via laser electron beam interaction. New Journal of Physics, 10:025005, 2008.

[15] N. R. Thompson and B. W. J. McNeil. Mode locking in a free-electron laser amplifier. Phys. Rev.Lett., 100:203901, May 2008. http://link.aps.org/doi/10.1103/PhysRevLett.100.203901.

[16] D. J. Dunning, B. W. J. McNeil, and N. R. Thompson. Few-Cycle Pulse Generation in an X-RayFree-Electron Laser. Phys. Rev. Lett., 110:104801, Mar 2013. http://link.aps.org/doi/10.1103/PhysRevLett.110.104801.

[17] NLS Project: Science Case and Outline Facility Design, (Ed. J Marangos, R Walker and G Diakun).STFC, 2009.

[18] P. B. Corkum and F. Krausz. Attosecond Science. Nature Physics. http://dx.doi.org/10.1038/nphys620.

[19] F. Krausz and M. Ivanov. Attosecond physics. Rev. Mod. Phys., 81:163–234, Feb 2009. http://link.aps.org/doi/10.1103/RevModPhys.81.163.

[20] H. Kapteyn et al. Harnessing Attosecond Science in the Quest for Coherent X-rays. Science, 317:775,2007.

[21] http://www.stfc.ac.uk/ASTeC/Programmes/EBTF/38426.aspx.

[22] FACET-II Facility for Accelerator Science and Experimental Test Beams-II. Technical report, SLAC,2013. https://portal.slac.stanford.edu/sites/ard_public/facet/Documents/FACET-II%20Proposal%20v6.pdf.

[23] L. T. Campbell and B. W. J. McNeil. Puffin: a three-dimensional, unaveraged free-electron lasersimulation code. Physics of Plasmas, 19:093119, 2012. http://dx.doi.org/10.1063/1.4752743.

[24] ASTeC–The Electron Beam Test Facility, Market Research. Inventya Ltd, 2012.

[25] A. M. Kondratenko and E. L. Saldin. Generation of coherent radiation by a relativistic electron beamin an ondulator. Particle Accelerators, 10:207, 1980.

[26] R. Bonifacio, C. Pellegrini, and L. Narducci. Collective Instabilities and High-Gain Regime in aFree-Electron Laser. Opt. Commun., 50:373, 1984.

[27] J. B. Murphy and C. Pellegrini. Laser Handbook Vol. 6, pages 9–69. North-Holland, 1990.

[28] R. Bonifacio et al. Physics of the high-gain free electron laser and superradiance. Riv. NuovoCimento, 13:1–69, 1990.

[29] E. L. Saldin E. A. Schneidmiller and M. V. Yurkov. The Physics of Free Electron Lasers. Springer,2000.

[30] W. A. Barletta et al. Free electron lasers: Present status and future challenges. Nucl. Instrum.Methods Sect. A, 618:69–96, 2010. http://www.sciencedirect.com/science/article/pii/S0168900210005656.

[31] B. W. J. McNeil and N. R. Thompson. X-ray Free-Electron Lasers. Nature Photonics, 4:814–821,2010. http://www.nature.com/nphoton/journal/v4/n12/abs/nphoton.2010.239.html.

http://link.aps.org/doi/10.1103/PhysRevSTAB.9.050702


http://link.aps.org/doi/10.1103/PhysRevLett.100.203901



http://dx.doi.org/10.1038/nphys620

http://dx.doi.org/10.1038/nphys620

http://link.aps.org/doi/10.1103/RevModPhys.81.163

http://link.aps.org/doi/10.1103/RevModPhys.81.163

http://www.stfc.ac.uk/ASTeC/Programmes/EBTF/38426.aspx

https://portal.slac.stanford.edu/sites/ard_public/facet/Documents/FACET-II%20Proposal%20v6.pdf

https://portal.slac.stanford.edu/sites/ard_public/facet/Documents/FACET-II%20Proposal%20v6.pdf

http://dx.doi.org/10.1063/1.4752743

http://www.sciencedirect.com/science/article/pii/S0168900210005656

http://www.sciencedirect.com/science/article/pii/S0168900210005656



[32] H. H. Braun. The Future of X-Ray FELs. In Proc. of IPAC2012, pages 4180–4183, 2012. http://accelconf.web.cern.ch/AccelConf/IPAC2012/papers/fryap01.pdf.

[33] B. W. J. McNeil. A simple model of the free electron laser oscillator from low into high gain. IEEEJ. Quant. Electron., 26:1124–1129, 1990.

[34] D. C. Nguyen et al. First lasing of the regenerative amplifier FEL. Nucl. Instrum. Methods Sect. A,429:125–130, 1999.

[35] B. Faatz et al. Regenerative FEL amplifier at the TESLA test facility at DESY. Nucl. Instrum.Methods Sect. A, 429:424–428, 1999.

[36] Z. Huang and R. Ruth. Fully Coherent X-Ray Pulses from a Regenerative-Amplifier Free-ElectronLaser. Phys. Rev. Lett., 96:144801, 2006.

[37] B. W. J. McNeil et al. A Design for the Generation of Temporally-Coherent Radiation Pulses inthe VUV and Beyond by a Self-Seeding High-Gain Free Electron Laser Amplifier. New Journal ofPhysics, 9:239, 2007.

[38] D. J. Dunning, B. W. J. McNeil, and N. R. Thompson. Short Wavelength Regenerative AmplifierFree-Electron Lasers. Nucl. Instrum. Methods Sect. A, 583:116, 2008.

[39] R. Colella and A. Luccio. Proposal for a free electron laser in the X-ray region. Opt. Commun.,50:41–44, 1984.

[40] K.-J. Kim et al. A Proposal for an X-Ray Free-Electron Laser Oscillator with an Energy-RecoveryLinac. Phys. Rev. Lett., 100:244802, 2008.

[41] S. Reiche. GENESIS 1.3: a fully 3D time-dependent FEL simulation code. Nucl. Instrum. MethodsSect. A, 429:243, 1999. Code available at http://genesis.web.psi.ch/.

[42] J. Marangos et al. New Light Source Conceptual Design Report. Technical report, 2010. www.newlightsource.org.

[43] G. Stupakov. Using the Beam-Echo Effect for Generation of Short-Wavelength Radiation. Phys.Rev. Lett., 102(7):1–4, 2009. http://link.aps.org/doi/10.1103/PhysRevLett.102.074801.

[44] D. Xiang and G. Stupakov. Echo-enabled harmonic generation free electron laser. Physical ReviewSpecial Topics - Accelerators and Beams, 12(3):1–10, March 2009. http://link.aps.org/doi/10.1103/PhysRevSTAB.12.030702.

[45] I. Martin. Preliminary Studies of the Tapered Undulator Short Pulse Scheme Applied to the CLARAFEL. Technical Report CLARA-FEL-REP-001/AP 500, STFC, 2012.

[46] I. P. S. Martin and R. Bartolini. Comparison of short pulse generation schemes for a soft x-ray freeelectron laser. Phys. Rev. Spec. Top. Accel. Beams, 14:030702, Mar 2011. http://link.aps.org/doi/10.1103/PhysRevSTAB.14.030702.

[47] M. Borland. ELEGANT - Electron Generation and Tracking, 2000. APS-LS-287.

[48] R. Bonifacio et al. Spectrum, Temporal Structure, and Fluctuations in a High-Gain Free-ElectronLaser Starting from Noise. Phys. Rev. Lett., 73:70, 1994.

[49] D. J. Dunning et al. Start-to-end modelling of a mode-locked optical klystron free-electron laseramplifier. Physics of Plasmas, 18:073104, 2011.

[50] E. Kur et al. A wide bandwidth free-electron laser with mode locking using current modulation. NewJournal of Physics, 13, 2011. http://dx.doi.org/10.1088/1367-2630/13/6/063012.

http://accelconf.web.cern.ch/AccelConf/IPAC2012/papers/fryap01.pdf

http://accelconf.web.cern.ch/AccelConf/IPAC2012/papers/fryap01.pdf

http://genesis.web.psi.ch/

www.newlightsource.org

www.newlightsource.org






http://dx.doi.org/10.1088/1367-2630/13/6/063012


[51] R. Bonifacio et al. Generation of XUV light by resonant frequency tripling in a two-wiggler FELamplifier. Nucl. Instrum. Methods Sect. A, 296:787, 1990.

[52] I. Ben-Zvi et al. Proposed UV FEL user facility at BNL. Nucl. Instrum. Methods Sect. A, 304:181,1991.

[53] L. H. Yu. Generation of intense uv radiation by subharmonically seeded single-pass free-electronlasers. Phys. Rev. A, 44:5178, 1991.

[54] L.-H. Yu et al. High-Gain Harmonic-Generation Free-Electron Laser. Science, 289(5481):932–934,2000. http://www.sciencemag.org/content/289/5481/932.abstract,.

[55] J. Feldhaus et al. Possible application of X-ray optical elements for reducing the spectral bandwidthof an X-ray SASE FEL. Opt. Commun., 140:341–352, 1997.

[56] E. L. Saldin et al. X-ray FEL with a meV bandwidth. Nucl. Instrum. Methods Sect. A, 475:357–362,2001.

[57] G. Geloni et al. A novel self-seeding scheme for hard X-ray FELs. J. Mod. Opt., 58:1391–1403,2011.

[58] J. Amann et al. Demonstration of self-seeding in a hard-X-ray free-electron laser. Nat. Photon.,6:693, 2012. http://dx.doi.org/10.1038/nphoton.2012.180.

[59] N. R. Thompson et al. Improved Longitudinal Coherence in SASE FELs. In Proc. IPAC2010, pages2257–2259, 2010. TUPE050.

[60] J. Wu, A. Marinelli, and C. Pellegrini. Generation of Longitudinally Coherent Ultra High Power X-RayFEL Pulses by Phase and Amplitude Mixing. In Proc. of FEL 2012, 2012. also SLAC-PUB-15348.

[61] D. Xiang et al. Purified self-amplified spontaneous emission free-electron lasers with slippage-boostedfiltering. Phys. Rev. Spec. Top. Accel. Beams, 16:010703, 2013.

[62] B. W. J. McNeil, N. R. Thompson, and D. J. Dunning. Transform-Limited X-Ray Pulse Generationfrom a High-Brightness Self-Amplified Spontaneous-Emission Free-Electron Laser. Phys. Rev. Lett.,110:134802, Mar 2013. http://link.aps.org/doi/10.1103/PhysRevLett.110.134802.

[63] D. J. Dunning, N. R. Thompson, and B. W. J. McNeil. Design Study of an HHG-Seeded harmoniccascade free-electron laser. Journal of Modern Optics, 58(16):1362, 2011.

[64] N. M. Kroll, P. L. Morton, and M.N. Rosenbluth. Free-electron lasers with variable parameterwigglers. IEEE J. Quantum Electron., 17:1436, 1981.

[65] R. Bonifacio and F. Casagrande. The superradiant regime of a free electron laser. Nucl. Instrum.Methods Sect. A, 239:36, 1985.

[66] L. Giannessi et al. Nonlinear pulse evolution in seeded free-electron laser amplifiers and in free-electronlaser cascades. J. Appl. Phys., 98:043110, 2005.

[67] T. Wanatabe et al. Experimental Characterization of Superradiance in a Single-Pass High-GainLaser-Seeded Free-Electron Laser Amplifier. Phys. Rev. Lett., 98:034802, 2007.

[68] I. Martin. Feasibility Studies of Echo-Enabled Harmonic Generation on CLARA. Technical ReportCLARA-FEL-REP-002/AP 500, STFC, 2013.

[69] B. W. J. McNeil, G. R. M. Robb, D. J. Dunning and N. R. Thompson. FELO: A One-DimensionalTime-Dependent FEL Oscillator Code. In Proceedings of FEL2006, page 59, 2006.

http://www.sciencemag.org/content/289/5481/932.abstract

http://dx.doi.org/10.1038/nphoton.2012.180



[70] B W J McNeil et al. A design for the generation of temporally-coherent radiation pulses in thevuv and beyond by a self-seeding high-gain free electron laser amplifier. New Journal of Physics,9(7):239, 2007. http://stacks.iop.org/1367-2630/9/i=7/a=239.

[71] B. W. J. McNeil and N. R. Thompson. Cavity resonator free electron lasers as a source of stableattosecond pulses. EPL (Europhysics Letters), 96(5):54004, 2011. http://stacks.iop.org/0295-5075/96/i=5/a=54004.

[72] J. Wu, C. Pellegrini and A. Marinelli. Generation of Longitudinally Coherent Ultra High Power X-RayFEL Pulses by Phase and Amplitude Mixing. In Proc. of FEL 2012, page 237, 2013. TUPD07.

[73] J. Wu and C. Pellegrini. private communication.

[74] R. Santer and N. R. Thompson. High-Brightness SASE Studies for the CLARA FEL. pages1274–1276, 2013. TUPEA061.

[75] J. K. Jones et al. A Compact, Modular Electron Beam Delay Line for Use in Novel Free-ElectronLaser Schemes. In Proc. IPAC2012, pages 1759–1761, 2012. TUPPP069.

[76] H. L. Owen and P. H. Williams. A modular path length corrector for recirculating linacs. Nucl.Instrum. Methods Sect. A, 662:12–20, 2012.

[77] H. Geng et al. Polarization Analysis for Seeded FELs in a Crossed-Planar Undulator. In Proceedingsof IPAC2010, 2010.

[78] D. Angal-Kalinin et al. Design of the Production and Measurement of Ultra-Short Electron Bunchesfrom an S-band RF Photoinjector. In Proc. of IPAC12, pages 1560–1562, 2012.

[79] S. Lebet et al. A Multi Purpose X Band Accelerating Structure. In Proc. of IPAC12, pages 70–72,2012.

[80] P. H Williams, J. K. Jones, and J. W. McKenzie. Comparison of Compression Schemes for CLARA.In Proc. of IPAC12, pages 1756–1758, 2012. TUPPP068.

[81] K. Flöttmann. Astra - a space charge tracking algorithm, 2000. http://www.desy.de/~mpyflo.

[82] J. W. McKenzie and B. L. Militsyn. A Velocity Bunching Scheme for Creating Sub-picosecondElectron Bunches from an RF Photocathode Gun. In Proc. of IPAC11.

[83] S. Di Mitri. Machine Design and Electron Beam Control of a Single-Pass Linac for Free ElectronLaser: the FERMI@Elettra Case Study. PhD thesis, U. Groningen, 2011.

[84] S. Di Mitri et al. Electron beam optics and trajectory control in the fermi free electron laser deliverysystem. Phys. Rev. ST Accel. Beams, 15:012802, Jan 2012. http://link.aps.org/doi/10.1103/PhysRevSTAB.15.012802.

[85] P. Craievich et al. In Proc. DIPAC 2007.

[86] D. Dowell. The limits of beam brightness from photocathode rf guns. In Proc. Free Electron LaserConference 2010, 2010. FEL 2010.

[87] J. T. Moody et al. Longitudinal phase space characterization of the blow-out regime of rfphotoinjector operation. Phys. Rev. ST Accel. Beams, 12:070704, 2009. http://prst-ab.aps.org/pdf/PRSTAB/v12/i7/e070704.

[88] R. Luus and T. H. I. Jaakola. Optimization by direct search and systematic reduction of the size ofsearch region. AIChE J., 19:760–766, 1973.

http://stacks.iop.org/1367-2630/9/i=7/a=239



http://www.desy.de/~mpyflo



http://prst-ab.aps.org/pdf/PRSTAB/v12/i7/e070704

http://prst-ab.aps.org/pdf/PRSTAB/v12/i7/e070704


[89] B. L. Militsyn et al. Photoinjector of the EBTF/CLARA facility at Daresbury. In Proc. of LINAC2012, 2012.

[90] F.B. Kiewiet. Generation of Ultra-short, High Brightness, Relativistic Electron Bunches. PhD thesis,Technische Universiteit Eindhoven, 2003. ISBN 90-386-1815-8.

[91] J.-H. Han et al. Design of a high repetition rate S-band photocathode gun. Nucl. Instrum. MethodsSect. A, 647:17–24, 2011.

[92] D. H. Dowell et al. Measurement and Analysis of Field Emission Electrons in the LCLS Gun. InProc. of PAC07, Albuquerque, New Mexico, USA, 2007.

[93] F. Stephan et al. New Experimental Results from PITZ. In Proc. of LINAC08, Victoria, BC, Canada,2008.

[94] T. Asaka et al. Stability Performance of the Injector for SACLA/XFEL at SPring-8. In Proc. ofLINAC12, Tel-Aviv, Israel, 2012.

[95] K. Batchelor et al. Development of a High Brightness Electron Gun for the Accelerator Test Facilityat Brookhaven National Laboratory. In Proc. 1st European Particle Accelerator Conference (Rome),pages 954–956, 1988.

[96] R. Akre et al. Commissioning the Linac Coherent Light Source injector. Phys. Rev. Spec. Top.Accel. Beams, 11:030703, 2008.

[97] J.-P. Carneiro et al. Beam Transport, Acceleration and Compression Studies in the FERMILABHigh-Brightness Photoinjector. In Proc. of LINAC98, pages 878–881, 1998.

[98] E. Chevallay et al. Photo-Cathodes for the CERN CLIC Test Facility. In Proc. of LINAC98, pages872–874, 1998.

[99] S. Schreiber et al. On the photocathodes used at TTF photoinjector. In Proc. of PAC03, pages2071–2073, 2003. see also http://wwwlasa.mi.infn.it/ttfcathodes.

[100] J. Rodier. Construction of The Alpha-X Photo-injector Cavity. In Proc. of EPAC13, pages 1277 –1279, 2006.

[101] Swiss FEL Injector Conceptual Design Report. Technical report, Paul Scherrer Institut, July 2010.

[102] M. Dehler et al. A Multipurpose X-band Accelerating Structure. In Proc. of IPAC2012, pages 70–72,2012.

[103] M. El Ashmawy. Electron Beam Dynamics in CERN-PSI-Elettra 5π/6 Travelling Wave X-bandLinearizer. In Proc. of IPAC2010, pages 4662–4663, 2010.

[104] G. Burt et al. A Transverse Deflecting Cavity for the Measurement of Short Low Energy Bunchesat EBTF. In Proc. of IPAC2012, pages 3335–3337, 2012.

[105] P. Kownacki. Reflectometer System for Active Phase Drift Compensation in Coaxial Cables. In Proc.of LLRF11. DESY, October 2011.

[106] M. Hoffman et al. Precision LLRF Controls for the S-Band Accelerator REGAE. In Proc. of IPAC2013, 2013.

[107] M. Milloch et al. LLRF Performance Results in Fermi@Elettra. In Proc. of LLRF11. DESY, October2011.

[108] T. Schilcher. First Results of the SwissFEL Injector Test Facility LLRF System. In Proc. of LLRF11.DESY, October 2011.

http://wwwlasa.mi.infn.it/ttfcathodes


[109] A. Gallo et al. Laser and RF Synchronization Measurements at SPARC. In Proc. of PAC07, page992, 2007.

[110] J. Crisp and B. Fellenz. Tevatron Resistive Wall Current Monitor. Journal of Instrumentation,6:T1100, 2011.

[111] A. Kalinin et al. Diagnostic System Commissioning of the EMMA NS-FFAG Facility at DaresburyLaboratory. In Proc. of IPAC10, August 2010.

[112] D. Angal-Kalinin et al. ALICE ERL Intra-Train Variation Investigation Using Bunch-by-Bunch BPMs.In Proc. of IPAC13, May 2013.

[113] D. Lipka. Cavity BPM Designs, Related Electronics and Measured Performances. In Proc. ofDIPAC09, May 2009.

[114] J. Bodewadt et al. A Compact Electron Photon Diagnostic Unit for a Seeded FEL. In Proc. ofDIPAC09, 2009.

[115] M. Veronese et al. Intra Undulator Screen Diagnostics for the FERMIElettra FEL. In Proc. ofIBIC2012, 2012.

[116] A. Angelovski et al. High bandwidth pickup design for Bunch Arrival-time Monitors for Free-ElectronLaser. Phys. Rev. Spec. Top. Accel. Beams, 15:112803, 2012.

[117] S. P. Jamison et al. Upconversion of a relativistic Coulomb field terahertz pulse to the near infrared.Appl. Phys. Lett., 96:231114, 2010.

[118] M. D. Roper. Spectral photon diagnostics for CLARA; General requirements abd suitability ofcommercial spectrometers. Technical Report CLARA-PHOT-DIAG-RPT-0002, STFC, 2012.

[119] http://www.horiba.com/uk/scientific/products/vacuum-uv-spectroscopy/monochromators-spectrographs/vtm300/.

[120] http://www.greateyes.de/greateyes_CCDCameras1024256Models_en.html.

[121] C. G. Durfee et al. Intense 8-fs pulse generation in the deep ultraviolet. Opt. Lett., pages 697–699,1999.

[122] M. D. Roper. Photon diagnostics for CLARA: Temporal diagnostics for mode-locked FEL scheme.Technical Report CLARA-PHOT-DIAG-RPT-0001, STFC, 2012.

[123] http://www.onefive.com/sync.html.

[124] S. Schulz et al. Progress towards a permanent optical synchronization infrastructure at FLASH. InProc. of FEL 2009, 2009.

[125] R. Wilcox et al. Stable transmission of radio frequency signals on fiber links using interferometricdelay sensing. Opt. Lett., 34:3050–3052, 2009.

[126] J. Kim et al. Long-term femtosecond timing link stabilization using a single-crystal balanced crosscorrelator. Opt. Lett., 32:1044–1046, 2007.

[127] S. Schulz et al. All-optical synchronisation of distributed laser systems at FLASH. In Proc. of PAC2009, 2009. TH6REP091.

[128] T. T. Ng and S. P. Jamison. Optical Clock Distribution at the ALICE Energy Recovery Linac. InProc. of IPAC 2011, 2011. MOPC148.

[129] K.-L. Yeg et al. Generation of 10µ ultrashort terahertz pulses by optical rectification. Appl. Phys.Lett., 90:171121, 2007.

http://www.horiba.com/uk/scientific/products/vacuum-uv-spectroscopy/monochromators-spectrographs/vtm300/

http://www.horiba.com/uk/scientific/products/vacuum-uv-spectroscopy/monochromators-spectrographs/vtm300/

http://www.greateyes.de/greateyes_CCDCameras1024256Models_en.html

http://www.onefive.com/sync.html


[130] A. G. Stepanov et al. Mobile source of high-energy single-cycle terahertz pulses. Appl. Phys. B,101:11, 2010.

[131] J. A. Fulop et al. Generation of sub-mJ terahertz pulses by optical rectification. Opt. Lett., 37:557,2012.

[132] J. Hebling et al. Tunable THz pulse generations by optical pulse generation of ultrashort laser pulseswith tilted pulse fronts. Appl. Phys. B. Lasers and Optics, 78:593, 2004.

[133] F. Blanchard et al. Generation of 1.5 µJ single-cycle terahertz pulses by optical rectification from alarge aperture ZnTe crystal. Opt. Express, 15:13212, 2007.

[134] C. P. Hauri et al. Strong-field single-cycle THz pulses generated in an organic crystal. App. Phys.Lett., 99:161116, 2011. http://dx.doi.org/10.1063/1.3655331.

[135] C. Ruchert, C. Vicario, and C. P. Hauri. Scaling submillimeter single-cycle transients towardmegavolts per centimeter field strength via optical rectification in the organic crystal OH1. Opt.Lett., 37:899, 2012.

[136] Z. Chen et al. Generation of high power tunable multicycle teraherz pulses. Appl. Phys. Lett.,99:071102, 2011. http://dx.doi.org/10.1063/1.3624919.

[137] J. R. Danielson et al. Intense narrow band terahertz generation via type-II difference-frequencygeneration in ZnTe using chirped optical pulses. Journal of Applied Physics, 104:033111, 2008.http://dx.doi.org/10.1063/1.2959846.

[138] A. G. Stepanov, L. Bonacina, and J.-P. Wolf. DAST/SiO2 Multilayer Structure for efficientGeneration of 6 THz Quasi-Single Cycle Pulses via Cascaded Optical Rectification. In Proceedings of37th International Conference on Infrared, Millimeter, and Terahertz Waves (IRMMW-THz). IEEE,2012.

[139] A. Sell, A. Leitenstorfer, and R. Huber. Phase-locked generation and field-resolved detection ofwidely tunable terahertz pulses with amplitudes exceeding 100 MV/cm. Opt. Lett., 33:2767, 2008.http://dx.doi.org/10.1364/OL.33.002767.

[140] J. A. Clarke et al. Final measurements of the SRS multipole wigglers. In Proc. of PAC 99, pages2656–2658, 1999.

[141] S. Ya. Tochitsky et al. Efficient harmonic microbunching in a 7th-order inverse-free-electron laserinteraction. Phys. Rev. Spec. Top. Accel. Beams, 12:050703, 2009.

[142] B. J. A. Shepherd and J. A. Clarke. Design, measurement and correction of a pair of novel focusingundulators for the ALPHA-X project. Nucl. Instrum. Methods Sect. A, 654(1):8 – 17, 2011. http://dx.doi.org/10.1016/j.nima.2011.05.029.

[143] D. J. Scott et al. Demonstration of a High-Field Short-Period Superconducting Helical UndulatorSuitable for Future TeV-Scale Linear Collider Positron Sources. Phys. Rev. Lett., 107:174803, Oct2011. http://link.aps.org/doi/10.1103/PhysRevLett.107.174803.

[144] A. Fasso et al. FLUKA-2011: a multi-particle transport code. Technical Report CERN-2005-10,CERN, 2005.

[145] MCNPX Version 2.7.0. Technical report, LANL, 2011.

[146] M. Hogan et al. Plasma Wakefield Experiment at FACET. In Proc. of IPAC2011, pages 715–717,2011.

[147] K. Efthymios et al. High-Gradient Plasma-Wakefield Acceleration with Two Subpicosecond ElectronBunches. Phys. Rev. Lett., 100:074802, 2008.

http://dx.doi.org/10.1063/1.3655331

http://dx.doi.org/10.1063/1.3624919

http://dx.doi.org/10.1063/1.2959846

http://dx.doi.org/10.1364/OL.33.002767

http://dx.doi.org/10.1016/j.nima.2011.05.029

http://dx.doi.org/10.1016/j.nima.2011.05.029



[148] A. Caldwell et al. Proton Driven Plasma Wakefield Acceleration. Nature Phys., 5:363, 2009.

[149] M. E. Couprie et al. LUNEX5: A FEL project towards the fifth generation in France. In Proc, ofFEL2011, pages 208–211, 2011.

[150] G. Xia et al. A Proposed Plasma Accelerator Research Station at CLARA Facility. In Proc. ofIPAC2013, page 1280, 2013. TUPEA064.

[151] S. Manz et al. Towards ultrafast electron diffraction and dynamic microscopy with REGAE. http://regae.desy.de/e135097/e135098/.

[152] P. Musumeci et al. High quality single shot diffraction patterns using ultrashort megaelectron voltelectron beams from a radio frequency photoinjector. Rev. Sci. Instrum., 81:13306–013309, 2010.

[153] Y. Murooka et al. Transmission-electron diffraction by MeV electron pulses. Appl. Phys. Lett,98:251903 –251905, 2011.

[154] R. K. Li. Imaging single electrons to enable the generation of ultrashort beams for single-shotfemtosecond relativistic electron diffraction. J. Appl. Phys., 110:074512 – 074521, 2011.

[155] J. Byrd. Strategies for achieving femtosecond synchronization in Ultrafast Electron Diffraction. InWorkshop on Ultrafast Electron Sources for Diffraction and Microscopy Applications, 2012. http://pbpl.physics.ucla.edu/UESDM_2012/Program.html.

[156] A. Variola. The THOMX Project. In Proc. of IPAC2011, pages 1903–1905, 2011.

[157] http://www.eli-np.ro/.

[158] W. Gai et al. Experimental demonstration of wake-field effects in dielectric structures. Phys. Rev.Lett., 61:2756–2758, 1988.

[159] M. C. Thompson et al. Breakdown limits on gigavolt-per-meter electron-beam-driven wakefields indielectric structures. Phys. Rev. Lett., 100:214801, 2008.

[160] S. Antipov et al. Experimental demonstration of wakefield effects in a THz planar diamondaccelerating structure. App. Phys. Lett., 100:132910, 2012.

[161] A. Andonian et al. Dielectric wakefield acceleration of a relativistic electron beam in a slab-symmetricdielectric lined waveguide. Phys. Rev. Lett., 108:244801, 2012.

[162] S. Antipov et al. Experimental observation of energy modulation in electron beams passing throughteraherz dielectric wakefield structures. In Proc. of IPAC2012, pages 595–597, 2012.

[163] M. E. Conde et al. A test-bed for future linear collider technology: Argonne Wakefield Acceleratorfacility (AWA). In Proc. of IPAC2012, pages 2778–2780, 2012.

[164] H. L. Owen, P. H. Williams, and S. Stevenson. Nonequilibrium electron rings for synchrotronradiation production. Phys. Rev. Lett., 110:154801, Apr 2013. http://link.aps.org/doi/10.1103/PhysRevLett.110.154801.

[165] R. Klein et al. Operation of the metrology light source as a primary radiation source standard.Phys. Rev. Spec. Top. Accel. Beams, 11:110701, Nov 2008. http://link.aps.org/doi/10.1103/PhysRevSTAB.11.110701.

[166] A. A. Mikhailichenko and M. S. Zolotorev. Optical stochastic cooling. Phys. Rev. Lett.,71:4146–4149, Dec 1993. http://link.aps.org/doi/10.1103/PhysRevLett.71.4146.

[167] M. S. Zolotorev and A. A. Zholents. Transit-time method of optical stochastic cooling. Phys. Rev.E, 50:3087–3091, Oct 1994. http://link.aps.org/doi/10.1103/PhysRevE.50.3087.

http://regae.desy.de/e135097/e135098/

http://regae.desy.de/e135097/e135098/

http://pbpl.physics.ucla.edu/UESDM_2012/Program.html

http://pbpl.physics.ucla.edu/UESDM_2012/Program.html

http://www.eli-np.ro/






http://link.aps.org/doi/10.1103/PhysRevE.50.3087


[168] E. G. Bessonov, M. V. Gorbunkov, and A. A. Mikhailichenko. Enhanced optical cooling system testin an electron storage ring. Phys. Rev. Spec. Top. Accel. Beams, 11:011302, Jan 2008. http://link.aps.org/doi/10.1103/PhysRevSTAB.11.011302.

[169] V. Danilov and S. Nagaitsev. Nonlinear accelerator lattices with one and two analytic invariants.Phys. Rev. Spec. Top. Accel. Beams, 13:084002, Aug 2010. http://link.aps.org/doi/10.1103/PhysRevSTAB.13.084002.

[170] http://arxiv.org/abs/1301.7032.





http://arxiv.org/abs/1301.7032

Head office Science and Technology Facilities Council, Polaris House, North Star Avenue, Swindon SN2 1SZ, UKEstablishments at Rutherford Appleton Laboratory, Oxfordshire; Daresbury Laboratory, Cheshire;UK Astronomy Technology Centre, Edinburgh; Chilbolton Observatory, Hampshire; Isaac Newton Group, La Palma;Joint Astronomy Centre, Hawaii.

Accelerator Science and Technology CentreDaresbury Laboratory, Sci-Tech Daresbury, Daresbury, Warrington, Cheshire WA4 4AD, UK

T: +44 (0)1925 603000 F: +44 (0)1925 603100

Rutherford Appleton Laboratory, Harwell Oxford, Didcot, Oxfordshire OX11 0QX, UKT: +44 (0)1235 445000 F: +44 (0)1235 445808

www.stfc.ac.uk/astec