Unlike what some people would like you to believe, fusion fuel is neither free nor necessarily easy to make.
Last month, The Fusion Report started its series on some of the challenges that commercial fusion energy must overcome to be successful: concepts like tritium breeding, cryonic vacuum pumps, heat exchangers, and other similar items which are critical to fusion, but which fusion companies themselves may not be the developers of. In this article, we examine the complexity in manufacturing fusion fuel, and systems for injecting that fuel into commercial fusion machines. But first, let’s dispel with the idea that fusion (or more specifically, fusion fuel) is “free”.
The Fuel Costs for A 1GW Fusion Energy Power Plant
Contrary to popular opinion, the fuel for fusion energy machines is not free. For instance in a recent Wall Street Journal interview, Bob Mumgaard, CEO of Commonwealth Fusion Systems (CFS), made the statement that, “The affordability story is straightforward. Fusion doesn’t consume a fuel that costs something, so there’s no price tag for gas, coal or nuclear power”. While the annual fuel cost for a 1GW commercial fusion power plant might in the future be significantly less than that for a 1GW nuclear fission, gas, or coal power plant, it is non-zero even in the most optimistic future conditions (the only power sources that have zero recurring fuel costs are renewables such as solar power, hydroelectric power, wind power, or geothermal power). While these are not deal-breakers fusion energy, the implication that fusion is free does not accurately depict its real costs. Moreover, the upfront capital expenses, and the initial lithium-6 to stock a tritium breeder blanket is significant, as we will explore deeper below.
A hypothetical 1GW commercial deuterium-tritium (D-T) fusion power plant would burn roughly 100 grams of deuterium and 150 grams of tritium per day. On the cheap side, the cost of cryogenically frozen pellets of highly purified deuterium would be roughly $13/gram, or $475,000 annually for a 1GW power plant. Then there’s tritium, which is today neither cheap nor plentiful, costing roughly $30,000/gram, or $1.6B per year. Assuming the technology to breed tritium is conquered, the price of the tritium itself goes to zero, but you are trading the cost for lithium-6 required in the breeding blanket, and the increased capital cost. The lithium-6 consumed annually in a tritium breeder blanket for a 1 GW fusion power plant would run roughly $20M/year, only somewhat lower than the $50M-$70M in annual fuel cost for a 1GW nuclear fission reactor.
However, the initial cost of filling the breeder blanket with 40 to 100 metric tons of enriched lithium today potentially threatens the competitive advantage of fusion energy that Bob Mumgaard trumpeted. At today’s market cost of $40,000 per kilogram of lithium-6, that cost would be $1.6B if only 40 metric tons were required ($4B if 100 metric tons were needed). Even if techniques like advanced vapor laser isotope separation (AVLIS) can be perfected, the cost would still be $160 million to $400 million to fill a blanket for a 1GW fusion power plant. This is certainly a workable cost, but it is only one of the costs for building a 1GW fusion power plant, and is based on yet-unproven technologies. Other promising breeder technologies such as liquid lead-bismuth thermal blankets have the advantage of increased neutron production which allows the use of cheaper natural lithium (Li-7), but liquid lead-bismuth produces polonium-210 (highly radioactive and toxic), and increases corrosion issues.
The Differences In Fueling Requirements for Fusion Energy Machines
The way fusion fuel (in this case, we’ll speak about D-T fuel, but it’s similar for other fuels) is prepared and injected into fusion machines depends significantly on the type of fusion technology utilized, whether it is inertial fusion energy (IFE), magnetic fusion energy (MFE), or hybrid fusion machines; we will explore each below.
IFE Fuels and Fuel Injection Systems: Inertial fusion machines run like car engines, emitting periodic impulses (explosions) of power, typically at a rate of one impulse per second or faster. The fuel is formed into small spheres of 4mm-5mm in diameter (For a great discussion about the specific requirements for inertial fusion pellets, please see the recent press release from Inertia Fusion, the illustration below is what we are borrowing). Inside the pellet is D-T gas, surrounded by a thin layer of D-T ice, and finally an outside carbon shell. Inertia believes they can reduce the cost to roughly one dollar per pellet, which would be a cost of roughly $63 million per year at a 2 Hz average impulse rate. The relatively hard part is that the pellets have to be injected periodically, acquired with the laser aiming system, and successfully fired at.
MFE Fuels and Fuel Injection Systems: Magnetic fusion machines are built to run continuously, and to heat the fuel continuously. The fuel can be much simpler than it is in an IFE (it can simply be D-T ice or gas), but the process to heat the ‘ice’ and turn it into plasma is a little more complex. Generally, a small amount of D-T gas is injected into the vacuum vessel (whether a tokamak, stellarator, or a similar device); the gas is then turned into plasma by a high-voltage breakdown. The plasma is then heated by a strong electrical current to a temperature of roughly 10 million degrees K. At the same time, hydrogen atoms are injected into the plasma by neutral beam injection. Finally, radio frequency (RF)/microwave systems boost the energy in the chamber to a level where the plasma can be fused using a strong magnetic fields.
Hybrid Fusion System Fuels and Fuel Injection Systems: Hybrid systems, such as Helion’s Polaris fusion energy system, work mostly like MFE systems from a fuel standpoint; the only difference is that fusion involves a combination of magnetic fields and inertia which is achieved by accelerating two ‘clouds’ of plasma towards each other.
Conclusion: Fusion Fuel Systems Are Still A Work In Progress
Fusion fuel systems (even without including breeder blankets, which are really a separate component) still have a ways to go to produce electricity commercially. For IFE systems, continually injecting frozen pellets at a rate of 1Hz or greater is the challenge yet to be conquered. For MFE systems, the question is: how do you inject fuel through incredibly strong magnetic fields and into the plasma without destabilizing the plasma. Both are infinitely solvable with enough time, work, and money. The problem with the fuel itself, which is really a breeder blanket problem, is more difficult, but it is a problem that all fusion companies have to solve.
