Review of Single-Bunch Instabilities Driven by an Electron Cloud

1
Review of Single-Bunch Instabilities Driven by an
Electron Cloud

• experimental evidence
• simulation approaches
• analytical treatments
• similarities differences to impedance-driven
  instabilities
• synergetic effects
• countermeasures
• open issues

Napa Valley April 2004
2

• concerns
• beam loss
• emittance growth
• trajectory change
• (turn-by-turn or pulse-to-pulse)

3
single-bunch instability

• but in multibunch or multi-turn operation
• (in all/most cases e- are already present
• when bunch arrives)
• for long proton bunches as in PSR, e- density
• increases towards tail of the bunch due to
• trailing-edge multipacting tail
  becomes
• unstable first
• e- cloud can as well drive coupled-bunch
• instabilities (talk by K. Ohmi)
• also strong possibility of combined
  coupled-bunch
• head-tail instabilities! (talk by D. Schulte)

4
(1) observations
5
INP Novosibirsk, 1965, bunched beam
other INP PSR 1967 coasting beam instability
suppressed by increasing beam current fast
accumulation of secondary plasma is essential
for stabilization 1.8x1012 in 6 m
first observation of an e- driven instability?
coherent betatron oscillations beam loss with
bunched proton beam threshold 1-1.5x1010,
circumference 2.5 m, stabilized by feedback
(G. Budker, G. Dimov, V. Dudnikov, 1965).
V. Dudnikov, PAC2001
6
Argonne ZGS, 1965 bunched beam, h8 (J.H.
Martin, R.A. Winje, R.H. Hilden, F.E.
Mills) oscilloscope traces showing
coherent vertical instability. Sweep rate is
0.2 sec/cm top signal from vertical pick
up bottom beam current.
growth time 5-100 ms, threshold 2-8x1011 protons
distributed over 8 bunches, largest bunches are
most unstable bunches move independently from
each other threshold varies with horizontal
position range or memory of the blow up does
not extend for more than 70 feet around the
machine instability suppressed by wideband (100
MHz) transverse damper
7
10 ms/cm, 0.2 cm amplitude growth at 1.15x1012
protons, bunched beam
BNL AGS, 1965 (E.C. Raka) coherent
vertical betatron oscillations and beam
loss caused by a poor vacuum (gt10-5 mm Hg) in
a small portion (1/12) of the ring threshold
showed weak dependence on pressure but rise
time strongly pressure dependent
threshold around 4x1011 protons per
pulse growth rates 20-500 ms for n8, 9 modes,
slow compared with 8 ms synchrotron
period, instability suppressed by sextupoles
narrow-band feedback studied
also at Orsay pressure dependent instabilities
were observed, and attributed to nonlinear fields
introduced by electrons (H. Bruck, 1965)
8
Bevatron, 1971, coasting beam (H.A. Grunder, G.R.
Lambertson)
n3-10 mode number changed towards smaller
values as instability progressed electron oscill
ation frequency decreased as beam size grew
for 1012 protons/pulse, beam size doubled in 200
ms clearing field at pick ups decreased
oscillation signal by factor 2 instability not
very sensitive to octupoles gas pressure 2x10-6
Torr feedback stopped growth
9
ISR, coasting proton beam, 1972 (R. Calder, E.
Fischer, O. Grobner, E. Jones)
excitation of nonlinear resonances gradual beam
blow up similar to multiple scattering
beam induced signal from a pick up
showing coupled e-p oscillation beam current is
12 A and beam energy 26 GeV
2x10-11 Torr, 3.5 neutralization, DQ0.015
extensive system of electrostatic clearing
electrodes
10
PSR instability, 1988 (D. Neuffer et al, R. Macek
et al.)
beam loss on time scale of 10-100 ms above
threshold bunch charge of 1.5x1013,
circumference 90 m, transverse oscillations at
100 MHz frequency
beam current and vertical oscillations hor.
scale is 200 ms/div.
11
frequency spectrum of oscillation
PSR instability contd (D. Neuffer et al, R.
Macek et al.)
beginning of instability, t0
log. y scale
t100 ms
lower frequencies associated with lower
intensities,
t300 ms after beam loss
0
1 GHz
12
PSR instability contd (R. Macek et al., M.
Blaskiewicz et al.)

• maximum number of protons scales linearly with
  rf voltage
• depends only weakly on bunch length!
• conditioning over time
• increases in pressure losses have marginal
  effect
• sustained coherent oscillations below loss
  threshold
• intense e- flux on the wall during bunch passage

• instability starts at bunch tail

instability e- production combined process!
13
AGS Booster, 1998/99 (M. Blaskiewicz)
time
5
beam current A
y power density
500 ms
-500 ms
0.2 GHz
coasting beam vertical instability growth time 3
ms
100 MHz downward shift as instability
progresses
14
KEKB e beam blow up, 2000 (H. Fukuma, et al.)
IP spot size
threshold of fast vertical blow up
slow growth below threshold?
beam current
15
KEKB e beam blow up, 1999 (H. Fukuma, E.
Perevedentsev, et al.)
KEKB witness bunch experiment bunch size depends
on its charge current of preceding bunches was
kept constant. Blow up has single- bunch
characteristics!
16
centroid motion bunch size tilt by
KEKB streak camera preliminary, October
2002 J. Flanagan, H. Fukuma, S. Hiramatsu, H.
Ikeda, T. Mitsuhashi
tail bunches blown up, slight evidence for tilt
17
PEP-II e beam blow up, 2000 (F.-J. Decker, R.
Holtzapple)
specific lumi
new
old
single beam
blow up due to combined effect of e-cloud
and beam-beam x blow up disappeared after change
in working point
colliding beam
18
CERN SPS with LHC beam, 2000
Intensity of 72-bunch LHC beam in the SPS vs.
time. batch intensity (top) and bunch intensity
for the first 4 bunches and last 4 bunches (where
losses are visible after about 5 ms) of the batch
(bottom)
(G. Arduini)
19
CERN SPS with LHC beam, since 2000
x coupled bunch instability y single-bunch
instability t50 turns
(K. Cornelis, G. Arduini,)

• suppressed by damper and high chromaticity
  (xy), possibly by linear coupling
• much improved after scrubbing, but residual blow
  up may occur
• interaction e- cloud impedance

20
tune vs oscillation amplitude for a bunch in the
tail of a train, sliding average over 32 turns
evidencing positive and negative detuning with
amplitude and sort of hysteresis indication
of nonlinear coupling between bunches in the tail
due to e- cloud
G. Arduini
21
calculated measured head-tail phase
difference for an LHC bunch train in the SPS
start of train
additional e- cloud wake field with wavelength
of 0.3-0.5 bunch length can reproduce measurement
end of train
K. Cornelis, 2002
22
central frequency 357 kHz, zero span
CERN PS, 2001 with LHC beam (R. Cappi, et al.)
adiabatic rf gymnastics for shorten the
bunch horizontal instability leading to
persistent oscillations w/o loss threshold
Nb4.6x1010 rise time 3-4 ms almost constant
above threshold but onset in time depends on
intensity for highest intensity bunches are
longer (sz is constant only over last 100 ms)
23
CERN PS, 2001 with LHC beam (R. Cappi, et al.)
instability rise time independent of x (up to
x0.5) marginal effect of octupoles introducing
HWHM tune spread of 0.5x10-4.
Fourier spectrum up to 10 MHz
signal at 357 kHz vs. time
Nb5.5x1010
instability visible only in the
horizontal plane (due to combined function magne
ts!?) no regular pattern along the bunch train
PS pickup before extraction
pickup in transfer line to SPS
24
BEPC e beam size blow up study (ZY. Guo et al,
APAC04 talk by J.Q. Wang)
sy
sy
Q -46
BPM bias -18
sy
0
600 V
2.0
0.2
solenoid -27
w/o BPM bias
head
tail
0
30
sy
octupole -34
with BPM bias
head
tail
1.0 A
0.0
25
90 consecutive bunches 30 bucket gap
DAFNE e ring, 2004 (M. Zobov, C. Vaccarezza, et
al.)

Bunches 25, 50, 70, 90
Bunches at the train end75, 80, 85,90
horizontal instability
positive x tune shift
probably linked to electron cloud but several
open questions
26
what is new after 40 years?
similar cures chromaticity, octupoles,
wide-band and/or narrow-band feedback, clearing
electrodes, better pumping new cures TiN or
getter coating clear identification as
e-cloud, better diagnostics, improved models,
computer simulations still lots of questions
27
(2) simulations
28
simulation approaches

• Microbunches (K. Ohmi, PEHT Y. Cai, ECI)
• Soft-Gaussian approximation (G. Rumolo, HEADTAIL
  v.0)
• discrete PIC codes (K. Ohmi, PEHTS HEADTAIL, G.
  Rumolo IHEP program)
• quasi- continuous PIC codes (QUICKPIC, USC)
• codes by M. Blaskiewicz, T.-S. Wang (centroids)
• df method for solving Vlasov-Maxwell equations
  (BEST code, H. Qin, R. Davidson)

29
no synchrotron motion
with synchrotron motion
after 100 turns
no synchr. motion
with synchr. motion
densities 2, 4, and 10x1011 m-3
TMCI
TMCI HT
BBU
Qx,y4,8
Qx,y0,0
PEHT
microbunches, multiple air bag model
(K. Ohmi, F.Z., PRL 85, 2000 )
30
ECI microbunch simulation for PEP-II (Y. Cai,
ECLOUD02)
slow emittance growth along bunch train below
TMCI threshold
e-cloud density for each bunch was obtained by
fit to independent simulation (M. Pivi)
31
simulation scheme for discrete PIC code
(G. Rumolo)
32
e- TMCI instability in PIC codeeffect of
synchrotron tune e- density
instability is suppressed by higher synchrotron
tune synchrotron tune required scales linearly
with density
PEHTS
(K. Ohmi, et al., PAC 2003)
33
scaling with r/Qs
(K. Ohmi, et al., PAC 2003)
PEHTS
this scaling works well for moderate e-
densities for largest densities there is a
different type of emittance growth (2 regimes,
see talk by E. Benedetto)
34
code comparison chromaticity dependence for KEKB
PEHTS (K. Ohmi)
HEADTAIL (G. Rumolo)
in PIC code Q acts stabilizing, HT-inst. not
seen (different from microbunch codes)
(G. Rumolo, F.Z.,PRST-AB 5, 121002, 2002)
35
e- density
beam density
quasi-static plasma code
e-cloud instability simulations using the plasma
code QUICKPIC CERN-USC collaboration T.
Katsouleas, A. Ghalam, G. Rumolo,
36
Contacts Persons for the Comparison of
Electron-Cloud Simulations Identified at ECLOUD02
Mike Blaskiewicz BNL
Yunhai Cai SLAC
Miguel A. Furman LBNL
Tom Katsouleas USC
Kazuhito Ohmi KEK
Mauro Pivi LBNL
Lanfa Wang KEK
Hong Qin PPPL
Giovanni Rumolo GSI/CERN
Tai-Sen Wang LANL
Frank Zimmermann CERN
G. Bellodi, RAL!
detailed comparisons between ECLOUD and POSINST

Build-up simulations highly successful 5 results
Mike Blaskiewicz, ECLOUD (F.Z./G. Rumolo), PEI
(Ohmi), POSINST (Pivi/Furman), CLOUDLAND (L.
Wang) less results for instability simulations!
37
Code Comparison after ECLOUD02
http//wwwslap.cern.ch/collective/ecloud02/ecsim/i
nstresults.html
benchmark case for instability simulations round
bunch in a round pipe 1e11 protons uniform
electron cloud with density 1e12 m-3 each bunch
passage starts with a uniform cloud chamber
radius 2 cm uniform transverse focusing for beam
propagation zero chromaticity, zero energy
spread no synchotron motion energy 20 GeV beta
function 100 m ring circumference 5 km betatron
tunes 26.19, 26.24 rms transverse beam sizes 2
mm (Gaussian profile) rms bunch length 30 cm
(Gaussian profile, truncated at /- 2 sigma_z)
no magnetic field for electron motion elastic
reflection of electrons when they hit the wall
38
1 mm
1.4 mm
ex,y
ex,y
PEHTS 1 IP, K. Ohmi
HEADTAIL 1 IP, G.Rumolo
5 ms
5 ms
0.06 mm
Post-ECLOUD02 Instability Code Comparison
- below TMCI threshold QUICKPIC gives a
rather different result!
ex,y
quasi-continuous QUICKPIC A. Ghalam, T. Katsouleas
need several/many IPs!?
4 ms
http//wwwslap.cern.ch/collective/ecloud02/ecsim/i
nstresults.html
39
discretized QUICKPIC with 1 IP
HEADTAIL with 1 IP
discretized QUICKPIC with 1 IP
HEADTAIL with 1 IP
another comparison of QUICKPIC-HEADTAIL for the
emittance growth in LHC here QUICKPIC was
discretized to model 1 IP for benchmarking
purposes both codes consider conducting
boundary conditions for rectangular pipe. (E.
Benedetto, A. Ghalam) no explanation for
difference yet.
40
transition between 2 regimes?
HEAD- TAIL
change from incoherent to coherent emittance
growth as IPs is increased no clear
convergence example HEADTAIL simulation for LHC
at injection re6x1011 m-3 (E. Benedetto, 2003)
41
effect of space charge
Simulated bunch shape after 0, 250 and 500 turns
(centroid and rms beam size shown) in the SPS
with an e- cloud density of re1012 m-3 without
(left) and with (right) proton space charge
(G. Rumolo, 2001)
42
ltygt
ey mm
suppression by chromaticity no HT in PIC
e-cloud broadband impedance tune spread
Evolution of centroid vertical position of an SPS
bunch over 500 turns for three cases e- cloud
broadband impedance, broadband impedance tune
spread, broadband impedance alone
Vertical emittance versus time for three
chromaticities.e-cloud, broad- band impedance
space charge.
(G. Rumolo, F.Z.,PRST-AB 5, 121002, 2002)
simulation results including effects of space
charge, broadband impedance and chromaticity
43
e- cloud effects in single-passsystem (LC beam
delivery)

• e- can build up along bunch train and reach
  densities up to 1014 m-3
• blow up of IP spot size for densities above
• threshold of 1011 m-3
• two effects breakdown of I in CCS and
• direct focusing effect at IP

phase advance change
direct focusing effect
(D. Chen, et al., 2003)
44
e- cloud effect in NLC beam delivery
IP beam size and central electron density 100 m
upstream of IP, vs. position along the bunch
IP beam size vs electron density, revealing
threshold at 1011 m-3
(D. Chen, A. Chang, M. Pivi, T. Raubenheimer,
2003)
45
(3) analytical treatments
46
interaction of beam and electron cloud

• electrons accumulate near beam center (pinch)
• tune spread, nonlinear fields, dynamic beta
• incoherent growth
• electrons follow transverse perturbations
• in bunch shape with delay
• wake field
• resulting net cloud response can drive
  instabilities
• beam break up (tltlt Ts)
• TMCI or strong head-tail (tTs)
• head-tail instability (tgtgtTs)
• exotic plasma instability? (e.g., monopole type)
• incoherent growth?

47
analytical estimates of equilibrium electron
density, wake field and coherent tune shift
e- density due to space charge and thermal energy
after S. Heifets, ECLOUD02
e- density due to charge neutralization F.Z.,
LHC Project Report 95, 1997
SB and CB wake of e- cloud K. Ohmi F.Z., PRL
85, 3821, 2000, G. Rumolo F.Z., APAC 2001,
Beijing
coherent tune shift due to e- cloud K.O. S.H.
F.Z., APAC2001, Beijing
48
analytical estimates for single-bunch instability

• adapt FBII theory
• (F.Z., CERN-SL-Note-2000-004)

BBU
(2) 2 particle model with length (K. Ohmi
F.Z.,PRL 85, 3821, 2000)
BBU
Head-Tail instability
TMCI threshold
49
(3) approximate wake by broadband resonator (K.
Ohmi, F.Z., E. Perevedentsev, PRE 65, 016502,
2001)
Green function wake damped oscillation
resonator frequency electron oscillation
frequency
low Q nonlinear force, variation of
lattice, variation of beam line density
shunt impedance
50
then apply standard instability analysis
(3a) TMCI threshold for long bunches (G. Rumolo
et al, PAC2001)
implicit equation since Rs/QR and wR depend on
Nb!
(applying conventional formula from B. Zotter,
CERN/ISR-TH/82-10, 1982)
(3b) threshold of fast blow up (K. Ohmi, F.Z.,
E. Perevedentsev, PRE 65, 016502, 2001)
implicit equation since Zeff and wR depend on Nb!
(applying conventional formalism by R.D. Ruth
Wang, IEEE Tr. NS-28 no. 3, 1981 P. Kernel,
et al., EPAC 2000 Vienna D. Pestrikov, KEK
Report 90-21, p. 118, 1991)
51
(3c) coasting beam instability threshold (E.
Perevedentsev, ECLOUD02 K. Ohmi, ECLOUD02, A.
Chao)
for mode near peak of resistive e-cloud impedance
no Landau damping for
left side of (1) scales as
(4) multiparticle models for combined effect of
e-cloud and beam-beam and/or space
charge (weak-strong G. Rumolo F. Zimmermann,
TWOSTREAM01 KEK strong-strong K. Ohmi A.
Chao, ECLOUD02)
52
various approaches to instability in PSR and/or
SNS

• centroid equations, transverse uniform
  distribution, one-pass two stream
• effect, Lorentzian energy distribution,
  instability always grows quasi-
• exponentially, growth rate is a function of both
  space and time eventual
• damping by proton frequency spread electron
  oscillation frequency
• spread causes spatial damping but not temporal
  damping
• T.S-. Wang et al, PRST-AB 6, 014204 (2003)
• semi-analytical model linear proton space
  charge, longitudinal dynamics
• by a square well potential (boxcar
  distribution) to reduce dimension
• of eigenvalue problem coasting-beam estimate
  simulations including
• space charge prediction that SNS will be stable
• M. Blaskiewicz et al., PRST-AB 6, 014203 (2003)
• electron oscillation amplitude much larger than
  proton amplitude (factor
• 20-50), phenomenological theory for the nonlinear
  regime, e- give driving
• force, slower linear or logarithmic growth in
  time, head of bunch carries
• memory possible cure drive head at frequency
  different from b sideband
• P. Channel, PRST-AB 5, 114401 (2002) c.f. S.
  Heifets, SLAC-PUB-7411
• nonlinear Vlasov-Maxwell equations, 3D
  perturbative df particle
• simulation, noise much treduced (df/f)2, noninear
  space charge, nonlinear
• growth phase H. Qin et al., PRST-AB 6, 014401
  (2003)

53
characteristic features of e-cloud for intense
long p bunches

• (see M. Blaskiewicz, next ICFA newsletter)
• nonlinear space charge is important varied
  opinion exists
• e- oscillation frequency depends on local beam
  current and local e- density which strongly
  increases near bunch tail
• self-consistent treatment of instability and e-
  generation likely necessary

54
long vs. short bunches
no. of oscillations over 2sz
ngtgt1 long nltlt1 short bunch
(G. Rumolo, ICFA NewsL4/2004)
Ring Type of particles Typical sz/c (ns) n Z/(Ag)
DAFNE Positrons 0.083 0.6 1.88
SPS (LHC) Protons 1 1.1 0.036
LHC (inj) Protons 0.45 1.4 0.0021
KEKB LER Positrons 0.013 1.4 0.27
LHC (coll.) Protons 0.25 1.6 1.3x10-4
RHIC Au79 ions 2.5 2.7 0.0037
PS (store) Protons 2.5 2.7 0.036
SIS18 U73 ions 17 6.5 0.25
ISIS Protons 23 13.3 0.54
PSR Protons 54 48 0.54
55
TMCI threshold vs n
Transverse Green f. wake
n1/4
if nlt1/4, wake has same sign over 4 sz
(E. Metral, ECLOUD02, for conventional resonator
wake field)
according to this definition, all our bunches are
long!
56
long bunch
TMCI threshold increases as
where
power spectra and real imaginary broadband
impedance
short bunch
threshold slowly increases with chromaticity
(E. Metral, ECLOUD02, for conventional broadband
impedance)
57
one difference between e-cloud and a
conventional impedance is the evolution of the
e-cloud density during the bunch passage
(pinch) see Mondays talk by E. Benedetto
58
simulated e- distribution during bunch passage
electron pinch
snapshot of horizontal and vertical e- phase
space (top) and their projections onto the
position axes G. Rumolo
59
electron pinch
Simulated electron distribution after bunch
passage in PEP-II (left) and e-
density enhancement alongthe bunch (right)
(M. Furman, A. Zholents, PAC 99)
60
density enhancement at beam center during LHC
bunch passage
electron pinch
modulation reflects linear rotation in phase
space
(E. Benedetto, 2003)
61
incoherent vertical tune spread at KEKB
solenoids off (e- cloud)
solenoids on (less e- cloud)
(T. Ieiri, H. Fukuma, 2001)
62
nonlinear tune spread and resonance excitation in
simulation
tune footprint obtained by applying frequency map
analysis on HEADTAIL simulation
with frozen-field approximation for LHC at
injection (E. Benedetto, Y. Papaphilippou, PAC
2001) a tune spread of only 0.002 is expected
for unperturbed uniform cloud
less emittance growth than for dynamic 2-stream
case!
multitude of excited resonances (0,3) (1,-4),
10th order,
63
in TMCI calculation pinch effect acts stabilizing!
no incoherent tune shift DQ0
real part
imaginary part
incoherent tune shift DQ(/-sz)/-2.5Qs
(E. Perevedentsev, ECLOUD02 see also V. Danilov
et al., PRST-AB, 1998)
head-tail mode tunes in units of synchrotron
tune vs. the cloud density in units of 1012 m-3
at Nb1011
64
Wake Field Calculation
(G. Rumolo)
displace 1 slice calculate either field on axis
or average force on subsequent slices normalize
to charge and offset of displaced slice
65
average wake
wake on axis
factor 20 difference! dependence on z!
(G. Rumolo, F.Z., PRST-AB 5, 121002, 2002)
66
different e- distributions yield different
wakes for the same average density
(G. Rumolo, EPAC2002)
67
electron-cloud wake impedance
1) wake not strictly linear (nonlinear force) 2)
wake depends on intensity, beam size bunch
length 3) no translational invariance
(pinch, varying beam line density) 4)
superposition principle does not apply
(nonlinear forces, e- memory) 5) wake depends
on transverse position 6) dependence on IPs
(different from conv. wake)
conventional formalism must be applied with great
care and cross-checked with simulations!
only Point 3) addressed so far tune change was
included in 3 and 4 particle models G. Rumolo,
F.Z., 2-Stream 2001 in ECLOUD02 Proc. exact
analytical treatment by E. Perevedentsev using
generalized impedance
68
generalization of transverse impedance

• must consider wake W1(z,z), not W1(z-z)

2-dimensional Fourier transform
(E. Perevedentsev, ECLOUD02)

• the wake W1(z,z) can be obtained from
• simulations

69
standard TMCI
(E. Perevedentsev, ECLOUD02 G. Rumolo)
70
generalized TMCI
(E. Perevedentsev, ECLOUD02 G. Rumolo)
71
(G. Rumolo)
extracting the 2-dimensional wake
72
2-dimensional impedance for SPS SIS
(G. Rumolo)
73
(4) some questions
74
is there a monopole instability?
axisymmetric instability mode found in
plasma simulations with cylindrical symmetry
in quasi-static approximation strong emittance
growth if arrival point is in front of the beam
center
(V. Lotov, G. Stupakov, EPAC 2002)
75
Simulations by HEADTAIL
for LHC
w/o bunch centroid motion
increasing IPs
symmetrized position of macroparticles (1 per
quadrant)
w/o slice centroid motion
(E. Benedetto, D. Schulte, et al., PAC2003)
emittance growth driven only by dipolar motion!
76

• two types of instability (see talk by
• E. Benedetto) - slow emittance growth??
• effect of lattice?
• more realististic electron distributions,
• e.g., longitudinal discontinuities?
• better approaches than PIC?
• (PIC is for the birds R. Talman)

77
longitudinal plasma waves?
78
longitudinal e- plasma waves- observed
calculated
measured
calculated
Tevatron Electron Lens
(V. Parkhomchuk)
79
FNAL TEL experiment 1 April 2004
(P. Lebrun, T. Sen, V. Shiltsev, X.-L. Zhang, F.
Zimmermann)
transverse position of TEL vs. p pbar location
during the scan
loss rates on 2D grid
80
FNAL TEL experiment 1 April 2004
(P. Lebrun, T. Sen, V. Shiltsev, X.-L. Zhang, F.
Zimmermann)
proton losses due to longitudinal shaving
losses decrease 1/(distance)3
note sometimes longitudinal shaving was
observed for LHC beam in the SPS when e-cloud
was present
(1) scattering off plasma fluctuations (2)
longitudinal coherent waves excited at entry and
exit of p bunch
81
Thanks to
G. Arduini, V. Baglin, E. Benedetto, M.
Blaskiewicz, K. Cornelis, V. Danilov, V.
Dudnikov, H. Fukuma, Y. Funakoshi, M. Furman, A.
Ghalam, Z.Y. Guo, S. Heifets, T. Ieiri, D.
Kaltchev, T. Katsouleas, R. Macek, E. Metral, K.
Ohmi, K. Oide, E. Perevedentsev, M. Pivi, T.
Raubenheimer, B. Richter, F. Ruggiero, G.
Rumolo, D. Schulte, V. Shiltsev, C. Vaccarezza,
J. Wang, R. Wanzenberg, A. Wolski, M. Zobov !