LETTER You may also like

- A review of laser–plasma interaction

Diamond spheres for inertial confinement fusion physics of indirect-drive fusion
R K Kirkwood, J D Moody, J Kline et al.

To cite this article: J. Biener et al 2009 Nucl. Fusion 49 112001 - Nuclear spectrometry of 9.6 h 196Aum2 and
the reaction of 197Au with fast neutrons
K J Moody, N Gharibyan, D A
Shaughnessy et al.

- Modeling and projecting implosion

View the article online for updates and enhancements. performance for the National Ignition
Facility
D.S. Clark, C.R. Weber, A.L. Kritcher et al.

This content was downloaded from IP address 50.230.201.134 on 03/02/2023 at 20:41

IOP PUBLISHING and INTERNATIONAL ATOMIC ENERGY AGENCY NUCLEAR FUSION
Nucl. Fusion 49 (2009) 112001 (5pp) doi:10.1088/0029-5515/49/11/112001

LETTER

Diamond spheres for inertial confinement

fusion
J. Biener1,a , D.D. Ho1 , C. Wild2 , E. Woerner2 , M.M. Biener1 ,
B.S. El-dasher1 , D.G. Hicks1 , J.H. Eggert1 , P.M. Celliers1 ,
G.W. Collins1 , N.E. Teslich Jr1 , B.J. Kozioziemski1 , S.W. Haan1
and A.V. Hamza1
1
Nanoscale Synthesis and Characterization Laboratory, Lawrence Livermore National
Laboratory, Livermore, CA 94550, USA
2
Fraunhofer-Institut für Angewandte Festkörperphysik, Tullastraße 72, 79108 Freiburg,
Germany
E-mail: [email protected]

Received 3 July 2009, accepted for publication 27 August 2009

Published 30 September 2009
Online at stacks.iop.org/NF/49/112001
Abstract
The National Ignition Facility (NIF) will allow scientists to prove the feasibility of inertial confinement fusion
(ICF). The success of ICF experiments at NIF will critically depend on the availability of robust targets. Guided
by computer simulations, we generated a new target design that takes advantage of the extreme atomic density of
synthetic diamond, and developed a process that allows us to produce large quantities of these ultrahigh precision
diamond targets via a low-cost batch process. Computer simulations were used to assess the performance and the
robustness of these diamond targets. The results demonstrate that diamond has the potential to outperform other
target materials in terms of energy efficiency and implosion stability, thus making successful ignition more likely.

PACS numbers: 07.05.Tp, 81.05.Uw, 81.07.Bc, 81.15.Gh, 89.30.Jj

(Some figures in this article are in colour only in the electronic version)

Fusion, the sun’s energy source, is viewed as a potentially deuterium and tritium will fuse together to form helium, a
important option to secure the global energy demand in a neutron and 17.6 MeV of energy.
sustainable fashion [1–6]. The National Ignition Facility (NIF) Developing an ICF target (figure 1(b)) requires
will deliver up to 1.8 MJ of ultraviolet (UV) laser light that will understanding of the complex physical phenomena that occur
allow scientists to prove the feasibility of inertial confinement under the extreme pressure and temperature conditions of an
fusion (ICF) [7, 8]. However, the success of these experiments ICF experiment [11]. Our target design [12] is guided by
will critically depend on the availability of robust targets. the two-dimensional radiation hydrodynamics codes LASNEX
A typical NIF ICF target consists of a thin-walled, millimetre- and Hydra, which contain the physics of both laser-driven
sized, spherical shell (ablator) made of a low atomic number hohlraums and capsule implosions, including the conversion of
(Z) element—beryllium, boron or carbon—that contains a absorbed laser light into x-rays, the flow of those x-rays within
frozen layer of deuterium–tritium (DT) fuel [9, 10]. In the the hohlraum and their absorption in the diamond ablator, as
so-called ‘indirect-drive’ configuration, which is currently well as ablation and hydrodynamic instability growth [13, 14].
the preferred configuration for NIF, the laser is focused on the These simulations are computationally very expensive and
inside wall of a centimetre-scale can surrounding the target sometimes require weeks of supercomputing time. In the
thereby converting the initially UV laser light into soft x-rays case of diamond, the shock physics is further complicated
(figure 1(a)). Precisely timed bursts of soft x-rays will then by its tightly bound crystalline nature in combination with
drive the compression of the DT fuel to conditions similar to an extremely high melting temperature, and the existence
those found deep in our sun, that is, 50 times the density of gold of another high pressure solid phase of carbon called BC8
at temperatures in excess of 106 K. Under these conditions, [15, 16]. Specifically, it is important to ensure that the ablator
a Author to whom any correspondence should be addressed. does not spend time in a partially melted state in which

0029-5515/09/112001+05$30.00 1 © 2009 IAEA, Vienna Printed in the UK

Nucl. Fusion 49 (2009) 112001 Letter

Figure 1. Diamond target design for indirect-drive ICF experiments on NIF: (a) The diamond shell containing the DT fuel sits in the centre
of a small cylinder called a hohlraum. The incident UV laser light will be focused on the inside wall of the cylinder where it will be
converted to soft x-rays which then compress the target by a rocket mechanism. (b) Optimized diamond target design with a predicted
energy gain of ∼15. The target consists of a 2.6 mm diameter diamond capsule with a wall thickness of 75 µm surrounding the 61 µm thick
DT ice fuel layer. (c) Actual targets made from NCD. (d) Simulation of the shock trajectory of diamond through the pressure–temperature
space of carbon during a NIF ICF experiment. Optimizing the laser pulse shape and thus the hohlraum temperature (inset, the numbers
indicate the shock sequence) allows us to avoid the BC8 phase and any partially molten state while keeping the overall fuel entropy tolerably
low. (e) Digital x-radiograph showing a diamond target with an inner layer of frozen DT fuel at 1.5 K below the DT triple point of 19.8 K. In
this example, the capsule’s inner diameter is 2 mm, and the fuel layer is ∼50 µm thick, with a surface roughness of less than 500 nm RMS.

density or velocity variations could arise and seed Rayleigh– and plastic ablators have been explored computationally
Taylor instabilities [17]. To better understand diamond’s and experimentally in more detail than diamond, and their
shock melting properties [18–20] we recently performed laser- fabrication has been developed [22–24]. Diamond, on the
shock compression experiments at the Omega laser at the other hand, is a relatively new candidate, with exciting
University of Rochester. These experiments demonstrate that prospects as described herein. The first experiments, beginning
diamond is partially melted at shock strengths between 650 in mid-2009, will use plastic. Future campaigns will consider
and 1050 GPa, and completely melted at higher pressures. diamond and explore its advantages. Our models predict that a
Other experiments on Omega were done to measure the diamond ablator is capable of adsorbing 10–20% more energy
perturbations that the nanocrystalline structure imprints on than ablator shells made from other low Z materials, thus
the shock front, and preliminary results suggest that the effectively increasing the laser energy by the same factor which
minimum imprint occurs when the first shock approaches the makes the diamond target more robust. The higher energy
melting curve, at about 600 GPa. Target simulations [12] efficiency of diamond targets is a consequence of diamond’s
have been done using equation-of-state (EOS) tables [21] higher initial density, which allows the ablator shell to be at a
that include liquid and two solid phases (diamond and BC8), larger diameter after shock compression, at the time of peak
consistent with the experimental melt data, as well as a energy delivery.
material-strength model for diamond. Optimizing the laser Besides its high atomic density, synthetic diamond has a
pulse shape and the corresponding hohlraum temperature and number of other advantages: the extremely high yield strength
shock sequence allows us to avoid the BC8 phase and any of diamond could enable room-temperature handling of filled
perturbations induced by partial melt, while keeping the overall targets in which the DT fuel would develop a pressure on
fuel entropy tolerably low (figure 1(d)). Our strategy is to the order of 1000 atm. Furthermore, diamond is chemically
keep the first shock in the regime indicated to be optimal extremely inert which allows one to use very effective
by the Omega experiments. A second shock is then used processing and cleaning procedures. Finally, synthetic
to melt diamond completely. (If the first shock were strong diamond is an environmentally friendly material that, contrary
enough to melt diamond completely, it would create too much to diamonds reputation as an expensive material, can be
entropy in the DT fuel and the fuel would not be sufficiently produced very cost-efficiently. Actually, the main ingredient
compressible to reach fusion conditions later in the implosion). necessary to synthesize synthetic diamond by chemical vapour
Assuming 1.3 MJ of UV laser light (0.35 µm), which heats deposition (CVD) is methane, and the material costs are only in
the hohlraum to a temperature of 270 eV, our optimized target the order of a few cents per target. In fact, it seems to be feasible
design (figure 1(b)) is calculated to produce a fusion yield of to produce large quantities of diamond targets for a commercial
∼24 MJ thus producing a gain of ∼15. Upon reaching ignition, ICF power plant (500 000 targets per day [25]) if polishing
deposition of alpha energy will totally dominate the energy can be avoided by growing ultra-smooth ultra nanocrystalline
content in the fuel. diamond (UNCD) films.
Several other candidate materials have been considered The main challenge in ICF target fabrication is precision
for NIF ignition target ablators. Specifically, beryllium as high target gain can only be achieved with shells that

2
Nucl. Fusion 49 (2009) 112001 Letter

Figure 2. Fabrication and characterization of ultrahigh precision diamond shells for ICF experiments on NIF: (a) deposition of uniform,
∼100 µm thick nanocrystalline diamond films on silicon spheres by microwave PACVD. A batch of diamond-coated Si spheres is shown on
the right. (b) Polishing of the outer shell surface to a nanometre-level surface finish. The one-dimensional surface profile shown on the right
was measured around the equator of the shell with an atomic force microscope and reveals a sub-10 nm RMS surface roughness finish. Note
that the out-of-roundness is only in the order of 10 nm. (c) Fabrication of 5 µm diameter holes through the diamond film using a diode
pumped, q-switched Nd : YAG laser with a 4 kHz pulse repetition rate at an average power of 10 W. Only volatile products are formed. Note
that the density variation produced by the 5 µm hole is very small as indicated by the weak contrast of the radiograph image shown on the
right. (d) Etching of the Si mandrel using a self-developed ultrasonically assisted HF/HNO3 wet etch process. The etch rate exhibits only a
very weak diameter dependence of d 1.25 (where d is the hole diameter) indicating that the mass transport is facilitated by ultrasound.

are nearly perfectly spherical and defect free. This requires removed through a 5 µm sized laser-machined hole using a wet
complete control of the shell sphericity over six orders of etch process (figure 2). To obtain void free and approximately
magnitude, from the nanometre length scale of the surface 100 µm thick diamond coatings we use a microwave plasma-
finish to the millimetre length scale of the shell diameter. assisted chemical vapour deposition (PACVD) technique in a
Any deviation seeds instabilities during the compression phase reactor optimized for long-term plasma stability [26]. Uniform
that can ultimately cause intermixing of capsule material with coatings are achieved by continuously rotating the Si mandrels
the DT fuel thereby reducing the efficiency, and in the worst during diamond CVD by using a rotating sample platform
scenario, suppressing ignition. Specifically, a shell thickness capable of holding large batches of the mandrels (figure 2(a)).
uniformity of better than 10−4 is required which translates into It takes approximately 50 h to deposit the required film
a sub-10 nm RMS surface roughness. To achieve this precision thickness of 100 µm using a feed gas mixture of 1% methane in
we had to carry out intensive R&D work in diamond deposition hydrogen at a deposition temperature of 700–900 ◦ C. For the
and processing, including seeding, grain size engineering, ICF application, the use of nanocrystalline diamond (NCD)
polishing and high-aspect-ratio micromachining. or UNCD is advantageous over coarse-grained material as
In short, our process uses solid silicon spheres as mandrels its isotropic character reduces the risks associated with the
for diamond deposition via a CVD process. After polishing anisotropy of diamond in the context of shock physics [27].
the resulting Si/diamond core–shell structure, the Si mandrel is As previously demonstrated by our group, thick NCD films

3
Nucl. Fusion 49 (2009) 112001 Letter

can be grown by controlling the secondary nucleation rate

by adding oxygen and/or nitrogen to the CH4 /H2 gas phase
chemistry [28].
A critical factor in target performance is the quality of
the inner surface of the diamond shell. This is particularly
true in that this surface finish cannot be improved at a later
stage of target fabrication. The quality of the inner surface is
mainly controlled by two factors: mandrel precision (sphericity
and defect density) and the nucleation density of diamond.
To minimize the number and size of the mandrel defects
which are directly replicated in the diamond film, we use
ultrahigh precision silicon spheres with a surface roughness
of less than 3 nm RMS. The nucleation density, on the other
hand, affects the porosity of the silicon–diamond interface,
and a fully dense interface requires a nucleation density in
the order of 1014 m−2 . Unfortunately, the nucleation of
diamond on non-diamond substrates such as Si is difficult, Figure 3. Stability behaviour of diamond ablator targets: density
and achieving high nucleation densities requires some sort of contours at ignition using the experimentally determined
one-dimensional surface roughness shown in figure 2(b) as input.
surface pretreatment, for example, by mechanical scratching The red line marks the central hot spot region, at radius of
[29, 30]. This, however, negatively affects the quality of the approximately 50 µm, where the D–T fuel ignites
inner surface by introducing mandrel defects. The challenge (temperature > 1 keV, density < 0.5 peak density, energy
is to optimize the nucleation density while keeping the number production > 1). Only small perturbations of the spherical
of surface defects created by the surface preparation treatment symmetry are observed, thus demonstrating the excellent
confinement that can be achieved with the diamond ablator design
at a minimum. We achieved nucleation densities as high shown in figure 1.
as 1014 m−2 and surface roughness values of the inner-shell
surface of less than 20 nm RMS by seeding the Si mandrels in
products (carbon monoxide and carbon dioxide) are released
an ultrasonically agitated nanocrystalline diamond–methanol which eliminates the need for additional cleaning steps to
suspension. Recently, we started to evaluate a new seeding remove debris (figure 2(c)). The Si mandrel can then
technique based on using monodisperse diamond nanoparticle be removed through this hole within a few days using an
colloidal solutions with particle sizes less than 10 nm [31]. ultrasound-assisted wet etch process developed by us for this
Preliminary results suggest that the roughness of the inner purpose (figure 2(d)). Here, the role of ultrasound is to
surface can be further reduced by using this seeding technique increase the etch rate by supplying energy and facilitating
to less than 5 nm RMS. mass transport through the narrow hole. After removing the
Diamond CVD typically results in the formation of a very mandrel, the hollow diamond shell can then be filled with the
rough surface morphology (for coarse-grained materials, the DT fuel at cryogenic temperatures using a micrometre-sized
peak-to-valley surface roughness is on the order of 10% of fill tube glued to the hole (figure 1(e)).
the film thickness), and consequently a substantial fraction Finally, we assessed the robustness of our targets by
of the deposited material has to be removed by polishing to computer simulations using the experimental metrology data
achieve the required nanometre surface finish. The challenge shown in figure 2(b) as input in addition to other data such
is to remove the material sufficiently fast while avoiding as DT ice roughness characterization. These simulations
uneven material removal or grain pullout. To overcome these confirm the overall favorable stability behaviour of diamond
challenges, we continuously rotate the diamond-coated Si NIF ICF targets which is a consequence of diamond’s
spheres during polishing by using a sample holder with a high strength. The high density reduces the coupling of
circular groove that serves as a race track for the inserted the DT ice surface and the ablator/ice interface from the
spheres (figure 2(b)). The extreme hardness of diamond unstable ablation front, thereby reducing the growth of the
requires us to use a diamond grinding wheel, and our surface perturbations seeded by ice roughness and inner-
experience shows that sub-10 nm surfaces finishes can only shell roughness. Our simulations indicate that a diamond
be achieved by using high-precision CVD diamond-coated ablator implosion can tolerate approximately twice as rough
grinding discs with a microcrystalline grain structure. In a DT ice surface as with a beryllium ablator. Consequently,
view of polishing, the use of NCD as ablator material is only small perturbations of the spherical symmetry develop
advantageous as it avoids the problems arising from the in our simulations demonstrating the excellent confinement
anisotropic polishing behaviour of diamond [32]. In the future, which can be achieved by using diamond as an ablator
it may be possible to skip the polishing step altogether as the material (figure 3). On top of these advantages, the diamond
deposition of NCD films with a mirror-like surface finish has shells produced by the technology described above exhibit
recently been demonstrated [33]. an unsurpassed sphericity (this is a consequence of using a
The final steps in target fabrication are to remove the solid mandrel, which is a unique feature of the diamond target
Si mandrel and to fill the capsule with DT fuel through fabrication process) and surface finish which further reduces
a micrometre-sized (5–10 µm diameter) hole which can be the final amplitude of Rayleigh–Taylor instabilities.
drilled by laser ablation. Laser hole-drilling in diamond Although our preliminary experiments confirm the
is a very fast and easy-to-automate process as only volatile promising behaviour of diamond ablator targets for ICF

4
Nucl. Fusion 49 (2009) 112001 Letter

applications, more work is necessary to better understand the [7] Miller G.H. et al 2004 Nucl. Fusion 44 S228
behaviour of diamond under extreme conditions, specifically [8] Miller G.H. et al 2004 Opt. Eng. 43 2841
[9] Lindl J. 1995 Phys. Plasmas 2 3933
with respect to material strength and fracture behaviour,
[10] Lindl J.D. et al 2004 Phys. Plasmas 11 339
hydrodynamic instability seeding and EOS. Further target [11] Haan S.W. et al 1995 Phys. Plasmas 2 2480
design studies will also allow more complete exploration of [12] Ho D.D. 2007 Ignition capsule design with a high-density
the operating space and the development of robust tuning carbon ablator for the National Ignition Facility 49th Annual
strategies. The performance of these targets will be studied Meeting of the American Physical Society (Orlando, FL)
[13] Zimmerman G.B. and Kruer W.L. 1975 Comment. Plasma
further on the Omega laser at the University of Rochester,
Phys. Control. Fusion 2 51
and on NIF. Although the targets described here have been [14] Marinak M.M. et al 1996 Phys. Plasmas 3 2070
developed for ICF experiments at Livermore’s National [15] Johnston R.L. and Hoffmann R. 1989 J. Am. Chem. Soc.
Ignition Facility, it is important to note that they can also be 111 810
used on the French megajoule laser, and smaller versions can be [16] Correa A.A. et al 2006 Proc. Natl Acad. Sci. 103 1204
[17] Sharp D.H. 1984 Physica D 12 3
used at the Omega laser. Even large-scale low-cost production
[18] Bradley D.K. et al 2004 Phys. Rev. Lett. 93 195506
of diamond targets for a commercial ICF power plant seems [19] Brygoo S. et al 2007 Nature Mater. 6 274
to be feasible. [20] Hicks D.G. et al 2008 Phys. Rev. B 78 174102
[21] Correa A.A. et al 2008 Phys. Rev. B 78 024101
This work was performed under the auspices of the US [22] Haan S.W. et al 2007 Eur. Phys. J. D 44 249
Department of Energy by the Lawrence Livermore National [23] Xu H.W. et al 2007 Fusion Sci. Technol. 51 547
[24] Chen K.C. et al 2006 Fusion Sci. Technol. 49 750
Laboratory under Contract DE-AC52-07NA27344.
[25] Goodin D.T. et al 2006 Phys. Plasmas 13 056305
[26] Füner M. et al 1998 Appl. Phys. Lett. 72 1149
References [27] Cook R.C. 2002 Fusion Sci. Technol. 41 155
[28] Kucheyev S.O. et al 2005 Appl. Phys. Lett. 86 221914
[1] Hoffert M.I. et al 2002 Science 298 981 [29] Arnault J.C. et al 1999 Eur. Phys. J. B 11 327
[2] Rebut P.H. et al 1992 Nucl. Fusion 32 187 [30] Lee S.T. et al 1999 Mater. Sci. Eng. R 25 123
[3] Nuckolls J. et al 1972 Nature 239 139 [31] Williams O.A. et al 2007 Chem. Phys. Lett. 445 255
[4] Nuckolls J.H. 1982 Phys. Today 35 24 [32] El-Dasher B.S. et al 2006 Appl. Phys. Lett. 88
[5] Key M.H. 2001 Nature 412 775 [33] Wiora M. et al 2009 Diamond Relat. Mater. 18 927
[6] Moses E.I. 2008 Energy Convers. Manage. 49 1795