8/14/2019 kim 1993 0340

1/3

Performance Characteristics, Optimization, and Error Tolerances of a4nm FEL Based on the SLAC LinacK.-J. Kim, M. Xie

Lawrence Berkeley Laboratory, University of California, Berkeley, CA 94720E. T. ScharlemannLawrence Liver-more National Laboratory, Livermore, CA, 94550

C. Pellegrini and C. TravishDepartment of Physics, University of California, Los Angeles, CA 90024Abstract

A 4nm free electron laser (FEL) operating in SelfAmplified Spontaneous Emission (SASE), and using theSLAC linac as a driver has been extensively studiedusing the FRED3D[l] and TDA3D[2] codes. Using a 7GeV beam with a normalized rms emittance of 3 mm-mrad and a peak current of 2500 A, obtained bylongitudinal bunch compress ion, the FEL can provideabout 20 GWatt of peak power, in a subpicosecondpulse. The FEL saturation length is about 60 m. Strongfocusing in both planes is provided throughout theundulator by a FODO quadrupole system. W e havestudied the system gain, its optimization and FELtolerance to beam parameter changes, wiggler errors andmisalignments.

Table 1: The base set of parameters for the SLAC basedx-ray FEL.Y Energy (m&) 14000

IntroductionThe promise of producing bright, coherent, shortwavelength XUV and X-ray radiation has yet to befulfilled. Free electron lasers have long been touted asthe right tool for this task. Yet, in the nearly twentyyears since the first operation of the FEL, the shortwavelength challenge has not been met because of thelimitations on beam brightness. Now it seems possible toproduce copious amounts of short wavelength radiationusing technology developed in the last few years [3,4,5].The primary distinguishing feature of this device is theelectron beam. A high current, low emittance (highbrightness) beam produced by an RF photocathode gunis accelerated to high energy (multi GeV) using a portion

of the SLAC linac. This beam is what distinguishes thisdesign from other potential x-ray FEL schemes (61.A large parameter space was explored in order tooptimize the FEL. The constraints where to maximize theoutput peak power while restricting beam and undulatorparameters to state of the art. A three dimensionalanalytic model [7] was used to initially explore theparameter space while particle simulations where usedto refine the choices. Table 1 lists a set of baseparameters. Subsequent sections of this paper presentFEL performance as functions of beam and undulatorparameters. The ma@ objective here is to establish theFEL tolerances with respect to changes in beam andundulator parameters and alignment errors.

En Emittance normalized(mm-mrad)Peak Current (A)Pulse Length (fs)oE Uncorrelated energy spreadau Undulator parameterhu Undulator period (cm)

h Optical wavelength (nm)n FEL parameter

3 x 10-625001604 x 10-4

68.34

1.7 x 10-3

Beam Parameter StudiesThe sensitivity of the FEL output to input beamparameters is paramount. The results presented below

are given in terms of the power gain length, LgP(z) = Poe z/L, (1)where P is the power as a function of the distance downthe FEL, z, and PO is the input power. The effects of

electron beam and undulator parameters on the FELperformance have been described by a 1-D model [3] andby a full 3-D analysis (reviewed in Ref. (51).Emittance

The usual constraint on the (unnormalized)beam emittance is that it be smaller than the wavelengthof radiation divided by four pi (E

8/14/2019 kim 1993 0340

2/3

4.0 4 - --.i 1 -.-.wAz 3.5 - --..--.--..-..-..-..........~.........-.---.-.-.--.-------~----.----.-................- --l2.5'Zo 2.05Eg 1.5

1.01O'B IO

1: The power gain length for various emittances.the emittance is a log scale.

The energy spread is characterized in two ways:correlated and the uncorrelated energy spread. FELnce (gain length) is affected by the uncorrelatedspread which is primarily determined by thesource. The theoretical limit is that cr~

8/14/2019 kim 1993 0340

3/3

Field ErrorsSimulations have been performed using randomwalk models of undulator rms field errors. E rrors inexcess of 0.4% seriously degrade the output power (seeFigure 4). The specified 0.2% tolerance is consideredpresently achievable. In fact, undulators meeting thisrequirement have already been constructed [I 31. Thequestion still remains whether rms errors are a figure ofmerit for free electron lasers. It is possible to constructundulators with large field errors and still achieve goodlasing [ 141. A more detailed study of field errors needs tobe performed.

10

1060 .l .2 .3 .4 .5 .6 .7

UndulaIor Error (% rms)Figure 4: The output power as a function of theundulator field errors is plotted. A total (rms) steeringaccuracy of 30 pm is assumed.Steering Errors

Whereas field errors are a dubious measure ofthe quality of the undulator. the second integral of thefield is generally considered a good measure. Steeringerrors can be corrected by judicious undulatorconstruction as well as a combination of beam positionmonitors and coils.

lo+- L : ! ! t ! ! : i !0 20 40 60 60 100

Steering Error (rms) [microns]Figure 5: FEL output power as a function of the beammisalignment.Simulations have shown that steering errors -30 pm can be tolerated. Figure 5 shown the FEL outputfor various steering and alignment errors whencorrectors are placed every 2 meters and a perfectundulator (no errors) is assumed.

ConclusionsThe numerous simulations performed for theproposed SLAC based X-ray FEL have shown that theparameters chosen are stable to fluctuations in beamparameters and achievable with present state of the artaccelerator, mechanical and magnetic technology.Further theoretical work needs to be performed toextend the 1-D theory of the start up and saturation

regimes (see Ref. [5] for a review]. Simulations of theseregimes will require codes which include pulse length(time) effects.Acknowledgments

We Thank Mark Hogan for his help inperforming some of the simulation used for this paper.This work was supported by the US Depatment ofEnergy.References

[l] E. T. Scharlemann, et al., Jour. Appl. Phys 58 (1985)p. 2154.[Z] T. M. Tran and J. S. Wurtele, Comput. Phys.Commun., 54 263 (1989).[3] C. Pellegrini, A 4 to 0.1 nm FEL Based on the SLACLinac, Proceedings of the Workshop on FourthGeneration Light Sources (1992).[4] W. Barletta, A. M. Sessler and L.-H. Yu, Using theSLAC nYo-Mile Accelerator For Powering An FEL,Proceedings of the Workshop on Fourth Generation LightSources (1992).[5] K.-J. Kim and M. Xie, Proceedings of 14 FELConference, Kobe, Japan (1992).[6] H. Winick. et al., A 2-4 nm Linac Coherent LightSource (LCLS) Using the SLAC Linac, Proceedings thisconference.[7] Y. H. Chin, K.-J. Kim and M. Xie, Phys. Rev. A46,6662 (1992).[8] J. Seeman, K. Bane and T. Raubenheimer. ElectronTransport of a Linac Coherent Light Source (LCLS) Usingthe SLAC Linac, Proceedings of this conference.[9] See G. Travish and J. Rosenzweig, NumericalStudies of Strong Focusing in Planar Undulators.Proceedings of this conference.[lo] R. Tachyn, Permanent Magnet Edge-FieldQuadrupoles As Compact Focussing Elements ForSingle-Pass Paticle Accelerators, SLAC-PUB-6058.[ 1 ] The Advanced Test Accelerator at LLNL operated the25 m Paladin FEL.[12] R. B. Plamer, SIAC-PUB 4295 (1987).[ 131 The UCLA-Kurchatov Hybrid undulator has a periodof 1.5 cm a peak field - 7.25 kC and a peak field error of~0.25% . Third generation light sources have similartolerances for their insertion devices.[14] Private comunication, Cliff Fortgang (Los AlamosNational Laboratory).

1535PAC 1993