A.M. Sessler- Laser Accelerators

8/3/2019 A.M. Sessler- Laser Accelerators

1/7

eScholarship provides open access, scholarly publishing

services to the University of California and delivers a dynamic

research platform to scholars worldwide.

Lawrence Berkeley National Laboratory

Title:

LASER ACCELERATORS

Author:

Sessler, A.M.

Publication Date:

09-15-2008

Publication Info:

Lawrence Berkeley National Laboratory

Permalink:

http://escholarship.org/uc/item/2cc556rh
http://escholarship.org/uc/item/2cc556rhhttp://uc/search?creator=Sessler,%20A.M.http://escholarship.org/uc/lbnlhttp://escholarship.org/uc/lbnlhttp://escholarship.org/http://escholarship.org/http://escholarship.org/http://escholarship.org/


2/7

,

LBL-15237C':

Lawrence Berkeley LaboratoryUNIVERSITY OF CALIFORNIA

Accelerator & FusionResearch DivisionPresented a t the 1983 Pa r t i c l e Accelera to r Conference,Santa Fe, NM, March 21-23, 1983

LASER ACCELERATORS

A.M. Sess le r

Apr i l 1983

TWO-WEEK LOAN COpyThis "is a Library Circulating Copywhich may be borrowed for two weeks.For a personal retention copy, callTech. Info. Division, Ext. 6782.

Prepared for the U.S. Department of Energy under Contract DE-AC03-76SF00098


3/7

LBL-15237LASER ACCELERATORS*Andrew M. SesslerLawrence Berkeley LaboratoryUniversi ty o f CaliforniaBerkeley, CA 94720

Summary can ' t understand how i t could ever have appeared tobe complicated.Laser accelerators may be conveniently

characterized, by the i r mode of operation, intomedia, far-f ield, and near - fie ld acce lera to rs. Thef i rs t category -- media accelerators - - i nc lu de t heInverse Cherenkov Effect Accelerator, th e PlasmaFocus Accelerator, and th e Beat Wave Accelerator(BWA). The second category -- f ar- fie ld accelerators -- inc1ude the Two-Wave Devi ce and the InverseF ree E le ct ron Accelerator (IFELl. The th ird category -- near- fi e 1d accelerators -- i ncl udes conventi on al l in ac s scaled to small dimensions, dielectricsheets , small holes in dielectric cylinders, andgratings. Attention is devoted to an example fromeach category; namely (1 ) th e BWA, (2 ) th e IFEL, and(3 ) the linac scaled to small dimensions (about 30GHz) and powered by a free electron 1aser (FEll.Finally, special attention is given to gratingaccelerators.

IntroductionThe use of th e large electr ic f ie lds of lasersfo r th e acceleration of part ic les has fascinatedphysicists for many years. Last year, a f i r s t Workshop devoted to this topic was held and th e Proceedings of this Workshop make interesting reading. 1I t was concluded, in th e Brief Report on thisWorkshop, that "the potential of laser-driven accelera tor devices just i f ies devotion of resources forthe i r further study and experimental exploration."In th i s review paper we shall sketCh theconcepts and experience which le d to th e above

quoted conclusion. Of course this paper is no t asubsti tute fo r Ref. 1, bu t we hope i t wi 11 serve asan introduction to Ref. 1.

There ar e many papers in th e 1i terature on thelaser acceleration of particles, and two excellentrevi e \ papers have been presented a t th i s veryseries of conferences.2 ,3 They provide the background with which th e reader can r ead il y t ackl e thispaper; in fact, this review is simply an up-date ofth e review by Palmer. Attention is also called toth e fine review of th e Workshop which recentlyappeared in Physics Today.4

General PrinciplesPrior to th e Workshop a few people, no doubt, hada clear understanding of what could, and could no t bedone with electr ic fields so as to accelerate par

t icles. Most people, however, had quite a cloudyview of this subject and th e l i terature, unfortunately, did l i t t l e to help for one could find in i tschemes which would work, schemes which would no twork, theorems which proved certain configurationswould no t work, and very complicated proposals someof which worked and some of which didn't work. Perhaps the most i ~ p o r t a n t output of th e Workshop is acommunity of physici sts who understand what can andcan no t be accomplished.As is often th e case in physics, once a si tuat ionis understood i t turns ou t to be very simple and one

To accelerate particles by th e oscillating fieldof an electromagnetic wave one must mainta in t he pa rt icles in synchronism with th e wave. (Otherwise th eacceleration wi 11 be cancelled ou t by deaccelerationsuch as is experi enced by a free electron subj ect toa pas si ng pul se of l ight , which one c an t hin k of asmany Fourier component s each of which does nothingon th e average.) In addition to synchronism onemust have, of course, an electr ic field componentalong th e d ir ec ti on o f th e par t ic le motion.

There ar e only two ways to achi eve these condit ions: ( l l either one slows th e wave down (and letsth e particle travel in a s t ra ight l ine) , or (2) onebends th e particle continuously and periodically (andlets th e l ight wave travel in a straight l ine) . (Isuppose one can combine these two concepts, bu t thaton ly comp 1i cates th e subj ect and, so- far, all proposed devices employ only one method or the other.)

1. Slow Electromagnetic Wave. Conventional1inacs work th l s way: One makes a slow wave structure in which th e electromagnetic field "bouncesaround" in a periodic structure or, as it is usuallydescribed, th e wave has a phase velocity less thanin free space and h en ce can resonate with a particle.

Thi s concept can be used at h igher f requenciesthan have been employed to date and one interestingconcept, whi ch we sha 11 di scuss at greater 1engthbelow, is "simply" (There ar e great technologicaldiff icult ies .) t he ext en si on of slow-wave l inacs to(say) 30 GHz.

I t is no t possible to construct a s low-wave device of th e usual (cylindrical) type a t very highfrequencies such as that of a 10 ~ C02 laser .However i t is possible to make planar slow wavestructures which can be periodic (gratings) ordielectric loaded (and no t neces sa ri ly per iodi c) .Of course, jus t as in ordi nary 1i nacs, one must be(roughly) within a wavelength of th e walls of such astructure. Thus devices of this class ar e calledNear Field Accelerators.There is another way in wh i ch one can slow anelectromagnetic wave down; namely to have it travelin a media. Devices of this type ar e called MediaAccelerators. In this case th e particle which is tobe acce1erated must a 1so travel in the medi um andhence is subject to scattering which l imits th e

length of such a device. The Inverse CherenkovEffect Accelerator is simply such an accelerator.(One can tr y to be clever and make a hole in themedium so th e particle is no t subject to scattering.Clearly, the hole c an not have a radius much largerthan a wavelength, and thus one is simply considering , a gain, a Near Field Accelerator; i . e . , a small

hole in a cylindrical die lec tr ic slow wave device.)The media, which slows th e electromagnetic wave,could be more responsive to the wave than simplybecoming polarized, as is typically the case when

* This work was supported by the Director, Office of Basic Energy Sciences Div is io n o f th e U.S. Depar tment ofEnergy under Contract No. DE-AC03-76SF00098.1


4/7

1.OO,-------r-----,----,-------,there is an ordi nary index of refraction. Thus ifth e medi a is a plasma, and fo r very intense 1i ghtpulses ionization will happen to any media, thereopens up th e very interesting possiDTTity of a colle ctiv e fie ld laser accelerator. The Beat WaveAccelerator, which will be discussed later, isexactly th is. 52. Periodic Particle Bending. Alternatively to

slowing th e e le ctr omagne ti c wave down, one canwiggle the particle so that th e motion is , part ial ly , along t he t ransve rs e electr ic f ield of th eaccelerating l ight wave. Of course, synchronismwith the e lec tromagnet ic wave is necessary and thisis simply accomplished by wiggling th e particleperiodically and having th e particle pass throughone peri od of th e structure whi 1e one peri od of th eelectromagnetic wave passes the particle. Devicesof this type do no t have to have the particle closeto a material boundary and, hence, ar e called FarField Accelerators.

Such a wiggling device, when operated in th ereverse direction, and when the particle wiggling isaccomplished by a static magnetic f ield, is known asa Free Electron Laser; when operated in the forwarddirection i t becomes an I nver se F re e E le ct ron LaserAcee1erator.

.50

Fig. 1.

. - 1 2 ~ - - - - 2 t O - - - - 3 " 7 ~ . 5 ~ - - " 5 " ' O . az (c/,"p)

Longitudinal electric f ield as a function ofdistance l 120 wpl. S imul at ion p ar ameters ar e wI 10.6 wp, w2 9.6wp, th e in it ial laser field strength is5. 0 r.1cwp and Te = 10 keV is the plasmatemperature. (From Ref. 1, p. 63).24 Stage Laser Beat Wave Accelerator 20-50 GeV Total Gain

Beam intensi ty per beam -5 x 1015 W cm- 2 , volc -0.5Plasma density _1 017 cm'3

The wiggling of th e particle does no t have to bedone by a static magnetic field, bu t could be doneby an electromagnetic wave. If a microwave f ield isemployed fo r this purpose then the device is knownas a Two-Wave Accelerator.?

The Beat-Wave Accelerator (BWA)The Beat-Wave Accelerator (BWA) is a mediaaccelerator in which th e media is a p1asma. 5 ,8Two di fferent 1asers, . at frequency wI and w2 ar earranged to fire in th e same dir ec tio n into aplasma. When th e frequencies ar e such thatwl-w2 is th e plasma frequency wp then aresonance ta ke s p la ce and th e plasma bunches at th ebeat frequency. This bunch ing produce s a longitUdinal electric f ield which, then, can be employed

t o accel er at e part ic les .

Be lai l 7Injected particles

H+, e-

.cGe lacussingWI kl lens 1/ 7wo ko

Plasma fem peratureGain per s tage

PlasmachamberI

fJ pinch cai l15 cm long

-100 eV- 1 - 2 GeV

rBeam dump

One can think of th e process of plasma-laserinteraction as non-linear forward Raman scattering.This process can lead to a very large longitudinalelectric f ield which is oscil latory and moves with av el oc it y c lo se to , bu t less, than th e velocity ofl ight. To develop an acceptable accelerating f ieldone needs to dri ve th e plasma hard (so as to be inthe non- 1i near regime) and have a very hot plasma(so as to Landau damp the backscattered RamanWave). Although one can make some progress on thisproblem analytically, one-dimensional particlesimulations have provided the most i ns ig ht i nt o th esUbject. In Fig. 1, obtained by Sull ivan andGodfrey, can be seen how a 1arge acce1erati ng fi e1dis developed in th e 8WA.First experiments have been performed by Joshi,

e t a1., and support the theory, at least to th e extent of demonstra ting the acceleration of particlesand th e strong dependence on plasma temperature.Of course, many questions need to be answeredbefore a BWA is made into a practical device; someof these questions are l isted in Ref. 1. Nevertheless, i t is interesting to consider what a fullscale machine woul d be 1ike. Figure 2 shows such aconceptual design.

2

XBl8Z4_311!)

Fig. 2. A full-scale Beat Wave Accelerator for producing electrons of 20-50 GeV, with 24stages and each stage giving 1-2 GeV to theelectrons. (From Ref. 1, p. 25).Further progress demand, and will ob ta in , theoret ical and experimental work. The BWA is the mostspeculative of the l a se r acce le rato rs (Le . , i t mayno t be possible to construct a very high energyaccelerator with this approach), but holds out greatpromi se as i t combi nes th e b es t f ea tu re s of 1aseracceleration and collec t ive accelera t ion .

The Inverse Free Electron Laser Accelerator (IFEL)This device, which is a far-field accelerator,consequently has import an t a dvantages over representatives of th e other classes of accelerators.First ly , th e particles and the laser beam do no t gothrough a medium and thus problems of scattering andbreak-down of th e media are avoi ded. Se condl y, th eaccelerated beam is not constrained to be near amaterial boundary and thus problems of transverseemittance limitations and, much more importantly,beam loading limitations ar e greatly alleviated.


5/7

Thi rd ly , the acceleration region is removed from theboundaries and, hence, electric breakdown of thewalls and heating (or even destruction) of the wallsis avoided.keeping the magnetic field of the wiggler at aconstant amplitude and having the wiggler periodbecome longer down the accelerator. (Of course, onecould mix-up these two extremes.)

A picture of one element of this device is shownin Fig. 3, which is taken from the article by Pellegrini in Ref. 1. 9 A summary of Workshop activityon the IFEL can be found in this very article.The second option has the advantage that theelectron synchrotron radiation loss only increasesas y2 (vs y4 in the first option). Up to a fewGeV, which is the case in one accelerating region,either option can be employed.

-" lX8l 833-8516

One of the results of the Near-Field WorkingGroup at the Workshop was the real izati on that manyadvantages could be achi eyed if ali nac wereoperated at (say) ten times the frequency of SLACi .e . , at a free-space wavelength of (aboutll cm.l uIn particular, i t was thought possible to obtain agradi en t of 400 MeV1m at th i s frequency. (Reca 11that for the SLC, SLAC will operate at a gradient of17 MeV/m.)

The Two-Beam Accelerator (TBA)

Typical parameters , based on a laser of 2 x10 14 Wand a pulse length of 1 ns are given inRef. 1. One can acce1erate currents of up to 5 kAto energies of 4.0 GeV in an accelerator of 40 m.

The basic physics of an IFEL has been demonstrated in the work being done at LANL, MSNW, and TRW onthe FEL. Thus the way woul d seem to be clear for amajor experimental study of the IFEL. Only in thisway, can one really learn what limitations there arein an IFEL. Of course, multiple acceleration stations is e ssen ti al f or a high-energy accelerator andth i s probably requi res the capabil it y of transportand refocusing of 1aser beams. The need for suchcapability is common to a number of laser accelerators and should be pursued.

In th i s frequency range there are no adequatehigh-peak-power sources, except, possibly, a freeelectron laser (FEL). The Two-Beam Accelerator (TBA)employs an FEL to produce radiation which then powersa con venti ona 1 structure. One of the beams is i ntense (- 1 kA), low energy (- 3 MeV), and long(- 30 m). The other beam -- the particles we arereally: interested in -- consi st s o f a few particles( - 1011) in a short bunch (- 1 mm) which are takento very high energies (- 375 GeV) in (about) 1.5Ion. Operati on at a repititi on rate of 1 kHz -which would be quite feasible for the FEL -- yieldsa luminosity of 4 x 1032 cm- 2 sec-I. A conceptual sketch of a TBA is shown in Fig. 5.

Fig. 5. Conceptual design of a TBA showing thesteady state FEL wi th its hi gh current beamand the high-gradient structure whichaccelerates particles to very high energy.

FOCUSING LENS

LASERAMPLIFIER-"

SEMITRANSPARENTMIRROR/'

Fig. 3. Schematic representation of an IFEL acceleration region. A laser beam and an electronbeam traverse an undulator magnetic field oflength L. The laser beam profile is assumedgaussian with a waist radius woo (FromRef. 1, p. 151l.

D - - - - - - - - , % = ~ ; ; i t - ; : - ~ - - - - - ~ = ~ ~ : = : - = - : : : : : - --e-8EAM \ONDULATORSOURCE

X8l 333-8517

Although significant acceleration can beachieved in one accel erati on regi on, a hi gh-energyaccelerat ion requi res many acceleration regions.This might be accomplished either by refocussing thelaser beam (a technology which has not ye t beendeveloped, but may be amenable to attack) or byusing multiple laser beams (which could become quiteexpensive). The second option, which surely can beaccomplished, is shown in Fig. 4.

MASTER LASER

Fig. 4. Conceptual design of an IFEL acceleratorusi ng mul tiple laser beams and accelerationregions. (From Ref. 1, p. 152).In an IFEL one must preserve the resonancecondition, which involves the electron energy,between the wi ggler peri od and the 1aser frequencyas the electron is accelerated. This can be accomplished either (1) by keeping the period of thewiggler constant and having the wiggler magneticfield increase down the accelerator or (2) by

(2 Wo lASER BEAM PROfilE" jil jtt=::

3


6/7

The field of laser driven accelerators has beenput upon a fi rm foundati on and is "worthy of seri ousattention by the accelerator community" .12

Beam loading was examined and shown to impose aserious limit on grating accelerators. (Essentiallythis is because the beam is, necesarily, so close tothe conduct ing surface.) In particular, at 10 ~ m one can probably only have 105 electrons in abunch.

Transverse stabil it y of beams was studied andshown to be present in an alternating gradientversion which appears to be achievable. Also,attention was given to the tolerances on gratingsand shown not to impose unusually stringent requirements on the construction of gra ting accelerators.

Much effort, at the Workshop was devoted tograting accelerators.11 Scaling laws with wave1ength were deri ved and a compari son between conventional linacs and gratings was developed. In particular, i t was seen that a grating accelerator willneed very l i t t le RF energy.

Conclusion

much more work is demanded andgrating accelerators, but gratinghold ou t the promise of accelerat ingvery high energies.Clearly,needed--onacceleratorsparticles to

Considerable attention was devoted to the limitson accelerating graidents . (Such things as surfacedestruction by heating, " short ing-out " o f the grating by the plasma created by a very intense laserpulse, and electric breakdown.) I t was shown thatat (about) 1 cm wavelength operation one can expectgradi ents of 400 MeV 1m, but that beyond that thegrowth is small (perhaps "only" reaching 800 MeV/mat 10 ~ at the surface heating limit).

The TBA is really a power transforming device;i e., i t ta kes power from the power 1i nes to aninduction accelerator, to a low energy beam, then toradiation (via a wiggler) which powers a highgradient accelerat ing structure, and, finally, tothe few high-energy particles one is interested in .Thus th ere are a number of components of a TBA whichneed to be studied (especially before optimizing thewhole system).Some components are, in light of the work of thelast decades, rather straight-forward. We would,fo r example, put the induction accelerator in th i sclass. Other parts of the TBA, however, requirefurther development, such as , for example, the FELand, most particularly, a "s teady state FEL"; i .e . ,one that neither gains nor loses energy of itspowering beam fo r many kilometers.

Grating AcceleratorsThe popular conception of a "real" laser accelerator is a grating accelerator. (See Fig. 6) This

Some parts of the TBA need consi derab1e study,such as the "pipes" which take the microwave powerfrom the FEL to the accelerating column; Le. , thiscoupling needs to be investigated and optimized.Another part of the TBA which needs work is the highgradient accelerat ing column which is very smalland, yet, is required not to break down electrically.All in a 11, however i t was fe I t by many at theWar kshop that 1earni ng how to buil d and operate an

accelerator at 1 cm would be a good step fromcurrent experience at 10 cm towards a C02 1aser at10 ~ m I t may be possible to obtain this experiencewi thout the TBA; i . e., wi th some more conventi ona 1power source, but if i t could be made to work theTBA has a certain neatness and compactness which ismost attracti ve.

(0)

(BL 833-851BFig. 6. In Fig. 5a is shown a convent ional linac(slow wave disk loaded structure) whichcould be scaled down to centimeter waveleng th s, b ut not much further. In Fig. 6bis shown a gra ting l inac which can certainly

be scaled to micron wavelengths and can,1ike the conventi ona 1 1i nac, accelerateparticles. (From Ref. 1, p. 187).

(b )

I \ \ I \ \\ \ II .' \ I I I .' 1 I \\ \} I I ' , I .'I \} I \ ,} I\ I \ I \ I I\ / \A number of schemes, which have both solid theoretical basis and experimental proof of the concept,need considerable more theoretical and experimentalwork before one can determine whether -- or not -

they have advantages over conventional accelerators. In th i s report we have gone into three schemesand mentioned a few others. I t is much too early,however, to choose between schemes. It is alsoimportant to be open fo r new schemes, whi ch may bebetter than any put forward to date.On the other hand, also, the fi e1d needs to beadvanced by research and development on areas whichare device independent .12 These inc lude theoretical studies; material breakdown and damage studies;development of high-quality radiation sources; focusing, transporting, and manipulating high-power laserbeams; and nonlinear effects in media. A start onsome of these subjects has been made in the manycontributed papers which can be found in Ref. 1.

Ac know1 edgments"popular conception" may be distorted for, as wehave seen, th er e a re a number of other laser accelerators such as the BiolA, lFEL, and TBA which certainly work and may have attractive features. Butgrating accelerators cer ta in ly , a lso, work and theyhold out the promise of obtaining very high energies(albeit for only a few particles) with very smallenergy expenditure.

This work was supported by the Director, Officeof Basic Energy Sciences Division of the U.S. Depar tment of Energy under Contract No. DE-AC03-76SF00098.References

1. P. J. Channell, Editor, Laser Acceleration ofParticles, AlP Conference Proceedings No. 91,~ J e w York, 1982.

4


7/7

2. J. D. Lawson, IEEE Trans on Nuc 1. Sci. NS-26,No.3, 4217 (1979). - - - - -3. Robert B. Palmer, IEEE Trans on Nucl. Sci. NS-28,No.3, 3370 (1981).4. "Laser-Driven Electron-Position Colliders,"Physics Today 36 , No.2, 19 (February, 1983).5. T. Tajima and J . M. Dawson, IEEE Trans Nucl.Sci. NS-26, No.3, 4188 (1979); T. Tajima and J .

M. Dawson, Phys. Rev. Lett 43, 267 (1979).6. A. A. Kolmenski i and A. N. Lebedev , Sov. Phys.JETP 23, 733 (1966); R. B. Palmer, J . Appl.Phys.4'3, 3014 (1972); W. B. Colson and S. K.Ride, ~ p l . Phys. 20, 61 (1979); P . Spr angl e andC. M. Tang, IEEE l"'rans. Nucl. Sci. NS-28, No.3,3340 (1981). ------

7. R. H. Pantell and T. 1. Smith, StanfordUniversity Internal Report (1981).8. See Ref. 1 the articles by A. Sessler p. 10, C.Joshi, p. 28, D. J. Sullivan and B. B. Godfreyp. 43 , T. Tajima and J . M. Dawson p. 69, and R.D. Ruth and A. W. Chao p. 94.9. See Ref. 1, th e art ic le by C. Pellegrini p. 138.10 . See Ref. 1, the articles by R, Palmer p. 179,and A. Sessler p. 154.11. See Ref. 1, the articles by R. Palmer p. 179,Kwang-Je Kim and N. Kroll p. 190, P. Wilson p.199, T. Weiland, p. 203, N. Kroll p. 211, P.Csonka p. 216, N. Kroll p. 237, P. Channell p.244, and P. Csonka p. 248, 266.12. See Ref. I , the article by P. Channell et al. p.1.

Documents

A.M. Sessler- Laser Accelerators