Proton Radiography of a Laser-Driven Implosion
A.J. Mackinnon,1P.K. Patel,1M. Borghesi,2R.C. Clarke,4R.R. Freeman,4H. Habara,4S.P. Hatchett,1D. Hey,1
D.G. Hicks,1S. Kar,2M.H. Key,1J.A. King,3K. Lancaster,4D. Neely,4A. Nikkro,5P.A. Norreys,4M.M. Notley,4
T.W. Phillips,1L. Romagnani,2R.A. Snavely,1R.B. Stephens,5and R.P.J. Town1
1Lawrence Livermore National Laboratory, Livermore, California 94550, USA
2Queen University, University Road, Belfast BT7 1NN, Northern Ireland, United Kingdom
3Department of Applied Science, University of California, Davis, California 95616, USA
4Rutherford Appleton Laboratory, Chilton, Oxon OX11 0QX, United Kingdom
5General Atomics, 3550 General Atomics Court, San Diego, California 92121, USA
(Received 25 March 2005; published 24 July 2006)
Protons accelerated by a picosecond laser pulse have been used to radiograph a 500 ?m diameter
capsule, imploded with 300 J of laser light in 6 symmetrically incident beams of wavelength 1:054 ?m
and pulse length 1 ns. Point projection proton backlighting was used to characterize the density gradients
at discrete times through the implosion. Asymmetries were diagnosed both during the early and stagnation
stages of the implosion. Comparison with analytic scattering theory and simple Monte Carlo simulations
were consistent with a 3 ? 1 g=cm3core with diameter 85 ? 10 ?m. Scaling simulations show that
protons >50 MeV are required to diagnose asymmetry in ignition scale conditions.
DOI: 10.1103/PhysRevLett.97.045001 PACS numbers: 52.50.Jm, 52.40.Nk, 52.40.Mj, 52.70.Kz
The conditions leading to ignition of inertial confine-
ment fusion (ICF) capsules are extremely difficult to real-
ize. In conventional ICF, cryogenic deuterium tritium fuel
is compressed from an initial diameter of 2 mm to a 60 ?m
diameter hot spot, resulting in particle densities (?) rang-
ing from 300–600 g=cm3around the cold fuel-hot spot
interface and an areal density (?R) of 0:3 gcm?2inside
the hot spot. To achieve ignition the compressed core must
be spherically symmetric, with a relatively smooth inter-
face between the cold fuel and the hot spot . Diagnosing
density perturbations and general deformity of the core
shape in these conditions is a challenge to conventional
radiographic techniques. The ideal diagnostic would pene-
trate densities up to 600 gcm3and resolve features of less
than 5–10 ?m. Imaging of neutron emission from an
igniting core is capable of high resolution imaging during
ignition shots , however it is unclear whether a target
that does not reach ignition will produce enough neutrons
to form a good image.
Advanced diagnostic techniques that can characterize
dynamically evolving conditions in highly compressed
media are clearly desirable. One possible method, cur-
rently under development is x-ray imaging in the range
of 15–30 keV . An alternative, perhaps complementary,
technique utilizes laser-driven proton beams of 30–
100 MeV to probe density perturbations in highly dense
materials. Multiple scattering arising from density struc-
tures in the probed object modulates the proton beam, thus
allowing micron scale structures to be imaged onto the film
plane [4,5]. Proton probing experiments of electromag-
netic fields in low-density laser-produced plasmas have
been carried out , however, the utility of proton radiog-
raphy in dense laser-compressed materials created by
laser-driven implosion has yet to be studied. Protonsbeams
with substantial particle fluxes at 40–50 MeV have been
produced by focusing a petawatt (PW) laser pulse on a
thin foil target . These protons can penetrate ? ?
500–1000 gcm3and so picosecond laser-driven proton
radiography is a potential candidate as a probe of dynamic
events in extremely dense matter.
This Letter reports on the first proof of principle study of
dynamic proton probing to diagnose density gradients in a
laser-driven spherical implosion. The experiment was car-
ried out using the 100 TW, 1 ps Vulcan laser pulse coupled
to a six-beam laser-driven implosion system . The com-
pressed core was characterized by protons and verified
using x-ray radiography; the experiments were simulated
by the 1D hydrocode (HYDRA ) and a 3D Monte Carlo
(MC) particle tracing code based on SRIM .
The protons were produced from a 1:054 ?m wave-
length laser with 1 ps pulse duration and 50 J energy,
focused by an F=3:5 off-axis parabola close to normal
incidence onto a 25 ?m thick tungsten target. The focal
spot (containing 30%–40% of the energy) was 10 ?m
full-width at half maximum (FWHM), giving a peak irra-
diance of 5 ? 1019W=cm2. This interaction produced a
proton beam with an exponential spectrum of mean energy
1.5 MeVand a high-energy cutoff around 10–15 MeV. The
proton producing target was protected from the plasma
surrounding the imploded target with a 6 ?m thick Al
filter, an arrangement that preserved the fidelity of the
proton beam while maintaining good spatial resolution in
the proton images .
The implosion was driven by six laser beams, each
1 ?m wavelength and 1 ns duration, focused onto a micro-
balloon at an irradiance of 1 ? 1013Wcm?2. These beams
were derived from the same laser oscillator as the short
pulse beam, thus eliminating any relative timing jitter be-
tween implosion beams and the proton backlighter. The
targets were plastic (poly-?-methyl styrene) microbal-
PRL 97, 045001 (2006)
28 JULY 2006
© 2006 The American Physical Society
loons, 500 ?m in diameter and with wall thickness either
3 or 7 ?m and no fill. Individual beam energy was in the
range of 50 J, giving a maximum energy on target
of 300 J and the beams were arranged such that they
illuminated the target tangentially from 6 orthogonal di-
rections, giving the best symmetry for a six-beam implo-
sion, as shown in Fig. 1. The synchronization of the 1 ns
heater beams to the picosecond backlighter beam was
measured to within accuracy of 100 ps using an optical
The proton detector used in this experiment consisted of
a multilayer film pack containing spatially resolving radio-
chromic (RCF) dosimetry film (types: MD-55 and HD-
810). This arrangement, which has been extensively used
in laser-driven proton acceleration experiments, gave a
diagnostic in which each layer was filtered by the preced-
ing layer, giving a series of images per shot, each with a
slightly different energy, ranging from 3 to 15 MeV. The
propagation of protons through the radiography object was
modeled using a MC simulation code based on SRIM .
SRIM calculates the final spatial and energy distribution of
ions passing through the object, taking into account ion-
ization energy loss by the proton into the target and energy
transferred to recoil atoms. The simulation code uses SRIM
to record the energy and position of each proton as it passes
through the target and the specific RCF film pack used in
the experiment. The program calculates the dose deposited
in each RCF layer by adding up the contributions from
each incident proton. This procedure provides a radio-
graphic projection of a transverse 2D slice of the target.
If the object has spherical symmetry then a three-
dimensional object can be obtained by superimposing a
large number of profiles rotated in small angles about the
A static test of point projection proton imaging was
obtained by backlighting an undriven microballoon with
diameter 500 ?m and wall thickness 7 ?m. The image of
the balloon in 7 ? 1 MeV protons shown in Fig. 2(a) was
obtained by projecting the proton beam onto the film pack
with a magnification of 10? (in this case the 3rd layer of
the film pack). A profile through the center of the shell is
shown in Fig. 2(b). It is clear that the balloon strongly
modulates the proton beam, with the highest modulation
occurring at the edges of the shell where the path length
through solid density material is highest. Also shown on
Fig. 2(b) is the output of 3D proton Monte-Carlo simula-
tions of 7 MeV protons from a point source propagating
through the microballoon. The simulations agree well with
the data, reproducing the shape of the edge of the balloon
and confirming that multiple small angle scattering ac-
counts for the image formation.
Temporal evolution of the capsule density profile during
the implosion was studied by varying the delay between
the backlighter and the implosion beams. The temporal
resolution of these proton radiographs is determined by the
energy bandwidth of each detector layer and the time of
flight differences for these protons as they propagate from
the source foil to the target plane. This results in two
effects: (1) each film layer probes a slightly different
time in the implosion and (2) time smearing within each
layer of film is determined by the energy bandwidth of the
film. The time smearing effect can be calculated knowing
the center energy and film layer bandwidth. For instance
for the third layer of film, with center energy of 7 MeV,
1MeVbandwidth, and a source to object distance of 4 mm,
the temporal smearing is ?10 ps. This film layer is easily
able to freeze motion in inertial confinement fusion (ICF)
plasmas where time scales for motion are typically
>50–100 ps. The time difference between the arrival of
5 and 10 MeV protons at the target is 30 ps, so no addi-
tional information can be obtained by comparing different
film layers in this relatively slowly evolving plasma, how-
ever this feature has been used in other experiments to
investigate sheath field evolution on ps time scales .
With nominally symmetrical drive conditions the cap-
sule remains roughly spherical as the implosion proceeds,
as can be seen from Fig. 3(a). This 7 MeV proton radio-
graph, taken 2 ns after the start (defined as the time where
the heater beam intensity reaches half maximum) of the
implosion, shows that the capsule has retained roughly
spherical symmetry with the shell diameter reducing
from 500 to 300 ?m and capsule walls that are still largely
FIG. 1 (color).
Geometry for proton radiography of the 6 beam
-0.20 0.2 0.4
Dose ( gray)
Distance in Object plane (mm)
RCF data Ep = 7 MeV
Simulation Ep = 7 MeV
FIG. 2 (color).
500 ?m diameter, 7 um wall thickness shell. (b) Lineout through
data (red circles) and Monte Carlo simulation (blue line) of
proton propagation through shell and RCF film pack.
(a) 7 MeV Proton radiograph of undriven
PRL 97, 045001 (2006)
28 JULY 2006
intact. There is some residual asymmetry due to the im-
perfect synchronization and energy balance of the 6 heater
beams; the shell is slightly elliptical, having departed from
its original spherical shape. These features are consistent
with 300–500 ps synchronization errors that were present
in the relative timing of the heater beams. In contrast
Fig. 3(b) shows a proton image of a capsule where the
drive was very asymmetrical due to significant timing
difference between the drive beams. In this case there
were delays of up to 2 ns between the heater beams,
leading to large distortions from spherical symmetry.
Radiographs were also taken closer to the time of peak
density (stagnation) of the implosion of a 3 ?m wall
thickness shell. The 5 MeV proton radiograph (the highest
proton energy image available for this shot) in Fig. 4(a)
was taken at 3 ns, which is close to stagnation for this shell.
At this time a dense core has assembled just below the
center point of the original capsule, due to the higher drive
levels from the upper beams. This results in a core that is
slightly elliptical with a minor diameter of 120 ?m and a
major diameter of 140 ?m (FWHM). The radiograph
clearly resolves this asymmetry (it is important to note
that the primary goal of this experiment was not to achieve
a perfect implosion but to prove the capability and effec-
tiveness of protons as a radiography source to diagnose
departures from symmetry).
In order to quantify the core density, the MC simulation
codewas used to model the propagation of a protons with a
1.5 MeV slope temperature through a spherically symmet-
ric object assuming a Gaussian radial profile and then onto
the film plane with a magnification of 10. The simulation
accounted for scattering in the dense core and the film
pack to allow a direct comparison with the data. The
peak density and FWHM of the Gaussian profile were
treated as variable parameters until the best fit with the
observed radiographs was obtained, with the background
level of the simulations scaled to match the data back-
ground. Figure 4(b) shows a profile across the minor
diameter of the radiograph of the core, together with the
output radiograph from Monte Carlo simulations. The
Gaussian source function was varied from 1–4 gcm3while
the source diameter (FWHM) was varied from 120 to
75 ?m,respectively, (thus keeping the total mass constant)
until the simulated radiograph matched the measured one.
The best fit to the data for the source density was 3 ?
1 gcm3with a core diameter (FWHM) equal to 85 ?
10 ?m with the assumption of a Gaussian density profile.
These results agreed well (within 25%) to the density and
core size measured using K?radiography of this implosion
Simple analytical estimates of multiple scattering from
Highland’s formula, were used as an approximate check on
the detailed MC simulations . Particles of momentum
p (in units of MeV), velocity, v and charge z, traversing a
material length L (in units of gcm?2) are scattered with
?1=e?rad? ? z?Es=p?c??L=Lrad?1=2,
Es ? constant ? 15, Lradis the radiation length, a Z de-
pendent constant for each material , and ? ? v=c. For
5 MeV protons traversing 80 ?m of carbon at 3 g=cm3,
z ? 1,
0:024 gcm?2, Lrad? 43:4 gcm?2for carbon, giving
?1=e?rad? ? 0:035 ? 2?. For an object to film plane dis-
tance, d ? 40 mm, a scattering angle of 2?would lead to a
blur of d??1=e? 1400 ?m. Dividing by the magnification
(M ? 10) gives 140 ?m at the object plane. The approxi-
mate size of the image at the object plane would then be
convolution of the initial source size and the blurring
??802? 1402?1=2? 160 ?m. This simple calculation
agrees well (to within 30%) to the core size of 120 ?m
observed in the radiographs. In general, blurring resolution
can be improved by using more energetic protons to mini-
mize multiple scattering. For example, 15 Mev protons
gives ?1=e? 0:6?for these object conditions, reducing
the blurring from 140 to 35 ?m.
The feasibility of proton radiography for diagnosing
density uniformity in very dense objects such as a com-
pressed ICF shell was investigated by using the COG MC
p?c ? 10 MeV,
L ? 80 ?m ? 3 g=cm3?
FIG. 3 (color).
quasisymmetric implosion, taken 2 ns after the peak of the drive
beams. (b) Proton radiograph taken of highly asymmetric im-
plosion, caused by mistimed heater beams. Beams colored in
blue arrived 1–2 ns earlier than beams colored in red.
(a) Proton radiograph (7 MeV protons) of
Object Plane distance (microns)
Proton dose (gray)
FIG. 4 (color).
500 ?m diameter microballoon with a 3 ?m wall at a time close
to stagnation (To ?3 ns). (b) Radial lineout taken through center
of proton radiograph data (blue circles) through the minor axis
and Monte Carlo simulation output for variable peak density and
varying core size (FWHM) while keeping the total mass con-
stant. The blue curve shows the best fit of 3 g=cm3peak and
FWHM of 83 ?m. Yellow: 1 g=cm3, 120 ?m FWHM; Green:
2 g=cm3, 95 ?m FWHM; dashed red: 4 g=cm3, 75 ?m FWHM.
(a) Proton radiograph, in 7 MeV protons, of a
PRL 97, 045001 (2006)
28 JULY 2006
model  to simulate radiographs of a density profile
generated by a 1D HYDRA simulation of a subignition
national ignition facility implosion. Figure 5 shows a
HYDRA profile of a ‘‘failed’’ indirect drive national igni-
tion facility implosion, where some of the shocks used to
assemble the high-density fuel have been mistimed, to-
gether with simulated proton radiographs at three ener-
gies. Density conditions inside this nonigniting core are
still extremely high, with a peak core density up to
350 g=cm3. The profile shows a high-density shell sur-
rounding a hollow core at stagnation. In this case the
peak core density of 350 g=cm3is located at a radius of
50 ?m, surrounding a central region with density
50 g=cm3. Simulations of a point projection proton image
of the object using monochromatic 50, 120, or 200 MeV
protons are shown in Fig. 5. In these simulations the film
was filtered with 6 mm of gold to slow down the high-
energy protons before encountering the first RCF layer
(multiple scattering in the gold filter is significant for the
50 MeV protons, but it is mitigated by the closeness of the
scattering material to the detection plane ). It can be
seen that the density peak at 50 ?m radius is clearly
resolved using 120 or 200 MeV protons but only margin-
ally with 50 Mev protons. Proton beams with 50 MeV have
already been produced by PW laser systems , however,
the generation of sufficient >120 MeV protons would
require a significant improvement in the high-energy cutoff
of laser-driven proton beams.
Macroscopic electric fields driven by pressure gradients
in the compressed core will also deflect protons. For an ig-
nition scale implosion the electric fields are driven by
pressure gradients in the core of order of the ignition tem-
perature T, over the core size r; T=r ? 10 keV=30 ?m ?
3 ? 106Vcm?1. The deflection angle can be estimated
from the change in transverse momentum of the protons
induced by the field as it passes through the core. For a
constant field acting over a distance, x, the deflection angle
??rad?? eEx=?2Ekin?, where e is the proton charge, E is the
electric field, x is the core radius, and Ekinis the particles
kinetic energy. For E ? 3 ? 106Vcm?1, x ? 30 ?m and
50 MeV protons, the deflection angle, ? < 0:01?. This
would not be detectable compared to multiple scattering
?1=e? 1?for these conditions. From these estimates it is
clear that in the high-density core of a stagnating implo-
sion, typical of laser-driven ICF, multiple scattering would
always tend to dominate over deflections from pressure
gradient driven electric fields.
In conclusion proton radiography of a laser-driven im-
plosion with 10 ps time resolution has been demonstrated.
MC simulations through a core with a Gaussian density
profile agreed well with the experimental data. MC simu-
lations also showed that protons >50 MeV would be re-
quired to resolve core symmetry in highly dense plasmas
such as those predicted at ignition scale.
This work was performed under the auspices of the U.S.
Department of Energy by the Lawrence Livermore
National Laboratory under Contract No. W-7405-ENF-48
and EPSRC Grants. A.J.M. acknowledges discussions
with O. Landen and T.E. Cowan and partial support from
the I.R.C.E.P., Queens University, Belfast.
 J. Lindl et al., Phys. Plasmas 11, 339 (2004).
 M.J. Moran et al., Rev. Sci. Instrum. 74, 1701 (2003); 75,
 J. Koch et al., Rev. Sci. Instrum. 74, 2130 (2003).
 M. Borghesi, et al., Plasma Phys. Controlled Fusion 43,
A267 (2001); , Phys. Plasmas 9, 2214 (2002).
 J. Cobble et al., J. Appl. Phys. 92 1775 (2002).
 R. Snavely et al., Phys. Rev. Lett. 85, 2945 (2000).
 C.N. Danson et al., J. Mod. Opt. 45, 1653 (1998).
 M. Marinak, Phys. Plasmas 5, 1125 (1998).
 http://www.SRIM.org; J.F. Zeigler, The Stopping and
Range of Ions in Matter (Pergamon, New York, 1985),
 A.J. Mackinnon et al., Phys. Rev. Lett.. 86, 1769 (2001);
M. Roth, Phys. Rev. ST Accel. Beams 5 061301 (2002).
 L. Romagnani et al., RAL Annual report 01/02, 2002, 26;
S. Kar et al., RAL Annual report 03/04, 2004, 26.
 J. King et al., Appl. Phys. Lett. 86, 191501 (2005).
 V.L. Highland, Nucl. Instrum. Methods 129, 497 (1975).
 T.A. Lasinski et al., Particle Data Group, Rev. Mod. Phys.
45, S1 (1973).
 COG can be obtained from ETSC, P.O. Box 1020 Oak
Ridge, TN 37831-1020, USA.
 Multiple scattering of 50 MeV protons in 6 mm gold
gives a resolution element of ?0:8 mm at the film plane,
resulting in spatial resolution of ?20 ?m for 40?
0 0.10.20.3 0.4
Proton dose on film (arbitary units)
Hydra density profile
Core Density (g/cm3)
from density peak
FIG. 5 (color).
200 MeV protons propagating through a density profile of a
mistimed National Ignition Facility implosion obtained from 1D
Hydra simulation (blue dashed line).
Simulated dose profiles taken with 50, 120, and
PRL 97, 045001 (2006)
28 JULY 2006