© 1993 IEEE. Personal use of this material is permitted.

However, permission to reprint/republish this material

for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers
or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE.

Experimental Demonstration of Plasma Lens Focusing

G. Hnimpetinn, P. Davis. C. E. Clayton. C. Joshi
Depnrtment of Electrical Engineering, UCLA. 405 Hilgnrd Avenue, Los Angeles. CA 90024 USA

S. Hnrtmnn. C. Pellegrini
Depnrtment of Physics, UCLA, 405 Hilgard Avenue, Los Angeles. CA 90024 USA

T. Kntsoulcns
Department of Eleclrical Engineering, USC, Los Angeles, CA 90089 USA

Abstract
The magnetic self-focusing of :I rckltivistic electron beam
propagating through a plasm:1 is demonstmled. The plnsmn
which is produced by an RF dischnrgc in n glass tube with no
externally applied m:rgneGc field, focuses n 3.5 McV. 25 ps
(FWHM) long electron beam from an initial size of 2.5 mm
(FWHM) to about 0.5 mm (FWHM) at n focal length of 18
cm.
I. INTRODUCTION
A relativistic electron beam propagating through :1pl:rsma
can self pinch due to its self generated nzitnulhal magnetic
field[l]. In vncuum. the bcnrn generated Lorcntz force
counteracts the radial space chnrgc force and the bcntn
propagates with nn equilibriutn radius. As the barn enters Ihc
plasma, the plasma clcclrons re-distribute thcmsclves lo
charge neu~r:rlize the beam. If Ihc beam radius is small (r <
c/w, = collisionless skin depth). tnosl of the pl;~nn return
currenfs will flow oufsidc of the IXX~ITI.As n rcsull. Ihe bcatn
magnetic field will not bc rcduccd apprecinbly within the
beam, and it will focus under its own radial Lorcnlz force.
This mech:rnism an be used to gcncmtc focusing gl-adienls
exceeding MG/ctn which is several orders of magnitude
greater Ihan the strength of conventional qundrupolc
mngnets[2].
While a large number of Ihcoreticnl studies[2].[?] of lhis
effect have been performed, there have been only two
cxperimenlal sludies[4],[5]. In (hc first cxpcritnent[4], Ihc
plasma lcnglh (35 cm) was much longer than the focal length
(8 cm), and the spar size was measured outside of the plasma
column. In the second experitncnt[S], ihe beam focused
outside of the plasm;t colutnn but the amount of radial
focusing was very litnited (in some cases the spot size
vnrintions were wilhin rhc experimental errors ). In this paper,
we report on a plasma lens experiment at LJCLA, where a 3.5
MeV elcclron bcntn was focused from an inilinl spot size of
2.5 mm (FWHM) lo about 0.S mm, well outside of Ihc plasma
column.
II. EXPERIMENTAL SETUP
The expcritnent was performed with the UCLA’s 4.5
McV. laser driven RF gun[(i]. The gun is capable of producing
very short (7 < 10 ps) clcclron bunches (Q > 1 nC, Ipeak >
100 .4) at 1 Hz. In order to simplify the diagnostics, the
clcclron bunches were stretched to about 25 ps (FWHM) by
illutnina(ing the cathotlc with :I longer laser pulse. The photo
electroils arc nccelcmrcd to nbou~ 3.S Mcv in the I .S cell RF
0‘fun and trnnsporlcd usin,‘7 n focusing solenoid and four
steering tnagncts to the plasma chamber 2 m downstream of
the gun (Figure I). The plasma chnmbcr which is filled with
Ar gas up lo n pressure of Xl tnTorr is connected to the RF
gun (p < 5 x IOwx Torr) Ihrough n windowless, Iwo stage
diffcrentinl pumping systctn (Figure I). Each pumping stage
consists of ;I low conductance tube followed by n
turbomolccul:~r pump. The first stage maintains n pressure
diffcrcntial of Ihrec orders of magnitude; while the second, a
diffcrenti:rl of two orders of magnitude. The electron beam
which inilially contains up to I .S nC of charge, is scraped by

Laser 2nd Low Conductance Tuhe

RF Gun
Coupling (5x1~11dia ., S cm long) Diagnostic Chamber
Bucking
Focusing Box
Solenoid
Solenoid I

1 I

1st Low Conduchnce Tube Ph.& Chamber

(Smm dia., 20 cm long)
I- l.Om -

Figure 1. Expcritncntnl Setup

0-7803.1203-l/93$03.00 0 1993IEEE 3543 PAC 1993

 the two low conductance tubes, and a maximum chnrge of 0.6 the antenna (n/Vn E 5 cm). The density directly underneath
nC is transported to the plasma chamber. Beam diagnostics the antenna cannot be measured because the Langmuir probe
include: phosphor screens both upstream and downstream of perturbs the RF discharge. The plasma density can be varied
the plasma chamber, retractable faraday cups. ;I current over an order of magnitude by varying the gas pressure and
transformer for non-destructive measurement of beam charge, the RF power.
and a Cerenkov radiator downstream of the plasma chamber. The plasma density has lo be high enough to chnrge
The plasma is produced by an RF discharge in a glass neutralize the beam in 3 time less than the electron bunch
tube (diam. = 1.7 cm, length = 10 cm) with no external length z = 25 ps: i.e. ~ctop >I where wp = (4rrnez/m)ln.
magnetic field. The RF amplifier (IO-20 MHz, 800 W) is Hence, the minimum required plasma density is n = 5 x 1011
connected via a capacitive tuning circuit to n helical ~IIPCIIII~ cm-j which can easily be attained by this source. The
wrapped on the outside of the glass tube filled with Argon transverse beam size is about 3 mm FWHM (< c/w,) at the
gas. The helical antenna (5 cm long, 12 turns) induces large entrance of plasma which im lies that the beam density is
azimuthal electric fields inside the tube where the Ar gas is ;lppt’OXil?IatCly Ilb < 4 X 10 I# cm-3
(nb < nplasma); i.e., the
ionized by impact ionization. Under optimum tuning, about his operates in “overdense” plasma regime. One can roughly
75% of the input RF power is coupled to the plasma. The RF estimate the focal length of the lens from the “thin” lens
plasma density is dingnoscd by n small cylindrical (diam. = approximation , which is given by[2]
0.76 mm, length = 2.5 mm) Lnngmuir probe. Figures 2(n) and
(b) show typical radial and axial density profiles. The radial f = (c/q,l,myII)
profile, measured at an axial position 0.5 cm away from the where f is ihc focal length, 1 is the plasma length, c is speed of
edge of the antenna, has ;I ll;it profile in the middle and light, and oq,,, is the bcnm plasma angular frequency. For n 5
decreases sharply near the tube walll. Axially, Ihc plnsm;r cm long plasma and n typical beam dcnsiiy of 3 xlOto cm-3,
density pNcaksunder the antenn;~ and falls r;ipidly away from the predicted focal length is about 27 cm. The focal length
a)
12, I I I I I I I
4 10 /“~1”‘1~(~1~~‘1~~“~~~,~‘~,‘~~,
•.o~o~o.&. 3000 I * I / I I I I I I i r f
00
m 0. 0
0 0 FWHM = 25 ps
E 0
=?7 lo’? 0
,x
.e 0 0
z 0
0”

2
a
2

1
lOI7

1012
i
l
“1
.

f”, I ’ c 8 ; ” /,‘I 1 I’# I”’

14
1 h ‘I
0

16
0
n
20 40
j
60 80 100
0 2 4 6 8 10 12 Time (ps)
r (InIn)
(b) Figure 3. Temporal Profile of the electron beam.
SlO’~~/, (, I,, , , I / / 1, I
calculation is for a beam with uniform density in space and
time. In reality, the focal strength of the lens is not constant
and has both longitudinal and radial variations.
2 00
x.3,012 .O 0 III. EXPERIMENTAL RESULTS
.‘j
m;c” 4 IO’?- 7 .*O 00 0
c The time integrated electron beam size and the total
0
O charge were measured 22 cm upstream of the plasma. The
2 2 10” I-
bcnm is about 2.5 mm FWHM in the vertical dimension, and

5‘li.,
a

:,,,/,,,/.::..:
110

- 1-
O i
I
10
Axial Position (cm)
I

IS
I-
contains up to .A nC of charge. The electron bunch length is
measured 27 cm downstream of plasma source by streaking
the Cerenkov ligh from n 0.5 mm thick fused silica[7]. Figure
3 shows the temporal profile of the electron bunch. It has n
sharp rise and n fairly long tail. The electron beam propagates
through the plasma and hits the first phosphor screen 18 cm
downstream of the plasma. Figure 4(n) shows the time
Figure 2. (a) Radial profile of plasma density. (b) Axial integrated beam size nt the first phosphor screen with no
profile of plasma density.

3544
PAC 1993
 No Plasma Vertical Dimension (mm)
0 2 4 6 8 10
(a)
Plasma on
0.0 2.0 4.0 6.0 8.0 10.0
Vertical Dimension (mm)
Figure 3. (a) Unfocused beam image 18 cm downstrcnm of plasma. source. (b) Vertical line out of image (a). (c) Focused
beam image. (d) Vertical line out of image (c).
plasmn. The beam dimensions are measured by Inking either n
vertical or horizontal line-out through the pixel with the
highest intensity. As Figure 4 (b) shows. the vertical
dimension of the unfocused beam is about 3.2 mm FWHM.
When the plasma is turned on. the bcnm focuses down to
about 0.5 mm (Figs. 4 (c) and (d)). The spot size increases to
about I.1 mm nt a second phosphor screen. 22 cm
downstream of the first. The apparent plasma focal length (1X
cm) is shorter than that predicted from the thin lens
approximation (27 cm). This can be attributed in part to the
finite length of the plasma. As lhc beam prop:lg:ltes through
the plasma, it begins to focus within the lens, increases its
density, and shortens the lens focal length.
IV. FUTURE WORK
To learn more about the time dependent focusing of the
beam, the Ccrcnkov radiation from the focused elcclron beam
has been streaked using a fast streak camera (temporal
resolution = 3.5 ps). Currently. we are in the process of
analyzing these results which will he reported elsewhere. We
also plan to investigate the effects of beam and plnsma density
(b)
Cd)
Y (mm)
on the focusing. Furthermore, experimental results will be
compared with both particle simulation and analytical
cnlculnlion results.
VI. ACKNOWLEDGMENTS
The authors acknowledge helpful discussions with J.B.
Roscnzweig. J.J. Su. and H. Zwi. Also, expert technical
assistance from N. Barov. M. Fauver, M. Hogan. P. Kwok. S.
Park, G. Tmvish, and R. Zhang is greatly appreciated. This
work is supported by SDIO/lST through ONR Grant NOOO14-
90-J-1952 and US DOE Grant DE-FGO3-92ER-40493.
V. REFERENCES
111P. Chen, Part. Accel. 20, 171 (1987).
I21 J.J. Su et. al... Phys. Rev. A 41, 3321 (1990). and
Rcfcrences thcrcin.
[3]P. Chcn. Phys. Rev. A 45, r3398 (1992).
[4] J. B. Rosenzwieg et. al., Phys. Fluids B 2 (6), 1376 (1990).
[S] H. Nakanishi ct. al., Phys. Rev. Lett. 66, 1870 (1991).
[h) C. Pellegrini ct. al.. these proceedings.
[7] G. Hnirnpetian et. al., these proceedings.
354s
PAC 1993