Doug Sparks, president, M2N Technologies
The concept of unlimited green energy from fusion using water as the underlying fuel dates back over 60 years. The sun uses gravity to confine and compress hydrogen together to initiate fusion. Man-made fusion of hydrogen isotopes was first demonstrated in the 1930s using particle accelerators. In the 1940s, the Oppenheimer-led Manhattan Project team developed inertial compression of uranium and plutonium using chemical explosions to generate the critical mass needed for a fission-based atomic bomb. Members of this team later used a fission explosion to inertially confine deuterium when they developed the hydrogen bomb in the 1950s. Plasma-based fusion development was started at the Princeton plasma physics lab in this same timeframe, with the hope of generating controlled electrical power.
More than 30 different methods of using plasma and lasers to fuse deuterium and tritium have been explored around the world in the past 60 years. In December 2022, the National Ignition Facility (NIF) at the Lawrence Livermore National Laboratory (LLNL) in California broke the record for energy output from controlled fusion reactions, namely, a fusion event had created more energy than the laser beams had directed into the confinement chamber that compressed the nuclear fuel1 . The NIF uses 192 high-power lasers to study inertial confinement fusion reactions. Like the fission reaction of a hydrogen bomb, the lasers provide the energy to a shell of material around a heavy water core that compresses the deuterium and tritium together.
Figure 1 illustrates how the deuterium and tritium nuclei are compressed to an ultra-high density using laser energy. The deuterium and tritium are part of heavy water molecules. This heavy water is in the form of ice inside of a hollow diamond sphere. The sphere is compressed to one ten-thousandth its original volume, leading to nuclear fusion. Helium, an extra neutron, and a lot of energy are produced in the fusion event. Key to this happening is the almost perfectly spherical diamond shell holding the heavy water ice, and this is where silicon micromachining comes in.
Like it does for many common devices in the modern world, e.g., automobiles, gaming systems, smart phones and so much more, silicon micromachining plays a small but critical role in the fabrication of the heavy water ice fuel capsule at the heart of the inertial confinement fusion system. An almost perfectly rounded, smooth spherical diamond shell is needed to ensure uniform compression of the deuterium and tritium together during the fusion process2,3. Surface roughness and out-of-roundness of this hollow shell must be less than 10 nm. To achieve this, a roughly 1 to 2 mm diameter silicon sphere is first ground and polished. The solid silicon sphere, also known as a mandrel, forms the mould for the diamond shell. Figure 2 illustrates the silicon micromachining-based fabrication process used to form the nuclear fuel containment capsule.
The silicon micromachining process.
A silicon crystal is first sawn into thick wafers, the thick wafers are diced into cubes, and each of those cubes is then ground into a sphere and polished. The polished sphere must match the sub-10 nm surface smoothness requirement. The silicon wafer industry has spent decades using chemical mechanical polishing (CMP) to produce millions of polished silicon wafers such as those shown in figure 3. In CMP, a colloidal slurry of abrasive particles and silicon chemical etchants are used to remove surface roughness and obtain this nanometre grade topography on a wafer. To produce the silicon sphere, this silicon CMP process was merged with ball bearing manufacturing to achieve an exceptionally smooth mandrel that becomes the mould around which a diamond shell is fabricated.
Silicon wafers and spheres.
Next, another familiar microelectromechanical systems (MEMS)-like process known as chemical vapour deposition (CVD) is employed to coat the sphere. A 100 µm thick diamond, or dense carbon, film is deposited onto the surface of the polished silicon sphere4. This CVD diamond coating process uses methods and deposition equipment developed for machine tool coating as well as for manufacturing high-power, diamond integrated circuits (ICs) and graphene memory semiconductors. The CVD system rotates the sphere continuously during the deposition process to ensure a uniform coating thickness.
Forming a shell requires the removal of all the silicon lying underneath the 100 µm thick diamond layer. To minimise imperfections, a single, narrow channel is cut into the diamond layer. Microfluidic plasma and laser etching of the five micron or smaller diameter hole have been used to form the high aspect ratio via through the diamond coating.
Etching the silicon in a 1 to 2 mm diameter sphere is more challenging than etching a few microns of polysilicon on a wafer. The etching of the sphere is done through the single, narrow channel that is drilled through the diamond shell. A variety of silicon etch processes have been tested to dissolve the silicon sphere, namely, plasma, wet acidic and wet caustic silicon5. A variety of silicon etchants such as ethylene diamine pyrocatechol (EDP), potassium hydroxide (KOH), tetramethylammonium hydroxide (TMAH), EDP (ethylene diamine and pyrocatechol) and nitric-hydrofluoric (HF) acid solutions as well as plasmas using carbon tetrafluoride (CF4), sulfur hexafluoride (SF6) and xenon difluoride (XeF2) are all well developed for MEMS wafer processing. It is a challenge to get the required volume of etchant through a single, deep, high-aspect ratio, diamond channel and an even bigger challenge to flush out the silicon-based reactants from the hollowing shell. MEMS process engineers are accustomed to just etching a channel a few microns deep, or for cavities maybe 100 to 200 microns deep. Etching a 2 mm thick silicon sphere using a single, tiny channel required some creative process development.
To improve the shell fabrication process, it would be advantageous to decrease the silicon sphere diameter and create a thinner shell with a wider or multiple vias cut into it. However, the physics of fusion suggests that a perfect shell with no etched via is ideal, so there is a push to shrink the diameter of the etch via in the diamond layer. The physicists that model the fusion process also desire a thicker, defect-free shell. A thicker shell with a smaller diameter, single via through it makes the formation of such a high aspect ratio channel very difficult to fabricate. These are some of the silicon micromachining challenges facing this project.
Microfluidics continues to play a role in the filling of the hollow sphere. A tiny tube is glued over the channel cut through the diamond shell. Numerous iterations in the tube mounting process are required to select the best adhesive with the optimum rheology to attach to the tube and to survive the cold gaseous filling process. This hair-like tube is essentially glued to the diamond shell. Considerable assembly development is required to attach the tube without clogging it or the via with the adhesive. Finally, prior to use, the heavy water ice fill process is undertaken.
The target of the 192 laser beams is not the micromachined diamond shell of the fuel capsule. Rather, as illustrated in figure 4, the ultraviolet (UV) laser light strikes the walls of a gold coated fixture in the vacuum chamber called a hohlraum, which holds the fuel capsule. The lasers heat the walls of the hohlraum, creating an ultra-hot, dense plasma, which generates X-rays that uniformly strike the deuterium and tritium ice containing diamond shell. The diamond shell is blown off the compressed heavy water ice in the centre of the chamber, resulting in nuclear fusion.
The laser ignition and compression of the micromachined fuel capsule.
The December 2022 experiment was the first man-made fusion event in which the energy released by fusion was more than the energy going into the hohlraum1. However, considerably more energy was needed to power the 192 lasers than was produced by the fusion of the heavy water ice fuel capsule. The system will be very useful for studying the nuclear fusion process and improving computer modelling of fusion reactions as well as being a significant step in the direction of one day being able to use water to supply unlimited, carbon-free energy.
Since the nuclear fusion fuel capsules used at LLNL’s NIF have a carbon shell, an interesting timeframe analogy can be made between coal-fired and inertial confinement-based nuclear fusion power generation. Coal, a fossilised form of carbon, was first burnt on wood campfires for heating over 3,000 years ago in northeastern China. It was also used for heat in Britain, Greece and Rome around 2,000 years ago. It was key to the industrial revolution where it was used to power steam engines and eventually generate electrical power from the 1880s. The first steam turbine had a thermal efficiency of just over 1 percent. Shovels were used to supply early steam engines with coal. Switching to conveyor belts enhanced power generation capability as did the use of pulverised coal. After 140 years of improvements, modern coal-fired power plants have efficiencies of just over 40 percent.
While the equivalent of two units of energy went into the NIF’s hohlraum and three units of energy came out, about 300 units of energy were required to power the 192 lasers. In terms of the coal-fired power generation development timeline, loading one carbon shell fuel capsule at a time into the inertial confinement chamber once a day roughly equates to throwing a single piece of coal onto a wood campfire. Only heat is being generated, not electricity. It is estimated that to produce a continuous supply of electricity, an improved inertial confinement system would need to laser blast 900,000 carbon shell fuel capsules every day. This would be analogous to the conveyor belt-fed, pulverised coal-fired power plant. Many technological hurdles and years of development remain before inertial confinement-based nuclear fusion power generation is able to match that of coal-fired and be used to power the electric grid.
Laser vapourising 900,000 fuel capsules per day is a daunting challenge, but it can be overcome. Take, for instance, ASML’s latest extreme ultraviolet (EUV) stepper for patterning the most advanced semiconductor wafers. To generate the 13.5 nm wavelength of light used by the EUV stepper, 50,000 tin droplets per second are shot into a vacuum chamber where they are vapourised by CO2 lasers to generate the stream of UV radiation6. This mechanism sounds similar to what LLNL or a fusion startup will need to develop when it comes to continuously loading fuel capsules into a fusion reactor. It took decades of development and billions of dollars for ASML to develop its lithography technology. An inertial confinement system must laser vapourise hundreds of thousands of heavy water ice-filled carbon spheres every day to practically generate electricity. There are several nuclear fusion power startups leveraging this and other government fusion research projects. One of these could be the unicorn that allows for the realisation of fusion power generation in the not-too-distant future.
M2N Technologies
References
1Nilsen, E. (2022) Nuclear fusion breakthrough a milestone for the future of clean energy, US officials say [press release]. CNN.
Available at: https://bit.ly/3OlJYyF
2Biener, J. et al. (2009) Diamond spheres for inertial confinement fusion. Nuclear Fusion, volume 49, issue 11, p.112001.
Available at: https://bit.ly/3rYDRsD
3Ross, J.S. et al. (2015) High-density carbon capsule experiments on the national ignition facility. Physical Review E, volume 91, issue 2, p.021101.
Available at: https://bit.ly/3OlSY6V
4Funer M., Wild C. and Koidl, P. (1998) Novel microwave plasma reactor for diamond synthesis. Applied Physics Letters, volume 72, issue 10, pp.1,149-1,151.
Available at: https://bit.ly/44QPqAt
5Kong, C. et al. (2019). Pressure cycle leaching of high-density carbon capsules. 23rd Target Fabrication Meeting, Annapolis, Maryland, US.
Available at: http://bit.ly/3DHqe3F
6Versolato O.O. (2019) Physics of laser-driven tin plasma sources of EUV radiation for nanolithography. Plasma Sources Science and Technology, volume 28, issue 8, p.083001.
Available at: http://bit.ly/47reM9P