 Thermonuclear fusion occurs when the temperature of a particular fuel is so large that a significant number of particles are moving quickly enough to overcome their mutual repulsion. The thermonuclear approach is the only way to release energy by fusion, as discussed in a previous video. In practice, this means raising the fuel to millions of degrees long enough for all the nuclei to collide many times and eventually fuse. Based on our daily experiences, it might sound as if attaining such an enormous temperature is the main challenge of achieving controlled fusion. Here's a cartoon by a Soviet scientist joking about how larger thermometer one might need to measure such a high temperature for instance. In actuality, reaching a high temperature is not difficult in itself. Temperature is a measure of the average random kinetic energy of particles in a clump of matter. To get up to a large average, the power of a modest power plant can be channeled into just a few grams of material, as in most controlled fusion experiments, or the enormous power of a nuclear blast can be channeled into a few kilograms of material in a hydrogen bomb. In those cases, it is relatively straightforward to achieve a high temperature. The problem instead is keeping the material confined long enough to release enough energy to make the initial energy investment worthwhile. The hotter something is, the higher pressure it exerts. If we were to instantaneously create a thermonuclear plasma in the middle of empty space, it would rapidly disassemble itself due to the pressure. The nuclei would then be too spread out to continue reacting. On the other hand, if a plasma were kept in a solid pressure chamber, its enormous heat would be conducted to the walls until the temperature became too low for fusion or the walls melted. For this reason, most fusion research is focused on the confinement of plasma, keeping both the particles and the energy they carry from escaping the reaction region. There are three main methods to achieve this. The pressure in a star is precisely counterbalanced by the mutual gravitational attraction of the astronomical number of particles within it. For much smaller objects, the mass of the fuel itself can resist the at-wads pressure simply through its inertia long enough for fusion to occur, if set up correctly. The constituent atoms in a plasma are broken up into electrons and ions, both with a net electric charge and therefore influenced by electromagnetic fields. A sufficiently strong magnetic field can counteract the pressure. I will cover the first two of these methods now and leave magnetic confinement to the next video. Even though fusion reactions happen in a variety of plasmas with sizes from the microscopic to the astronomical, it is possible to make some generalizations. One metric which allows confinement performance to be compared is the triple product, the density multiplied by the temperature multiplied by the energy confinement time. Up to a point, increasing temperature increases the probability that collisions will lead to fusion. A higher density, defined as the number of particles in a given volume, increases the rate at which those collisions happen. The energy confinement time is usually defined as the average time it takes for a unit of energy to escape the plasma. The longer the energy confinement time, the lower the rate at which energy is lost, and the lower the amount of input energy which must be provided to keep the reaction going. I want to make it clear that the triple product is not an absolute measure of the fusion power output or the fusion gain, but rather it is merely a figure of merit allowing useful comparisons. For example, comparing two types of magnetic confinement devices, the best triple product achieved by a tokamak is 100 times higher than a stellarator. While this does not mean that tokamaks are strictly 100 times better than stellarators, they are superior for the moment. There are a couple of issues with all methods of plasma confinement including any others not yet discovered. A hot plasma is a messy, unpredictable beast. It is highly unstable. All that energy, which is good for causing fusion reactions, also makes the plasma chaotic, unpredictable and uncontrollable. Phenomena such as turbulence allow particles to escape or heat to be conducted out, thereby making the energy confinement time drastically low. In terms of controlled fusion research, understanding and overcoming myriad plasma instabilities has been the greatest issue so far. I will make reference to the terms in the triple product when I talk about the specific confinement methods in a moment. A nucleus must collide on average thousands of times before it fuses. Each collision causes acceleration among particles. In physics, an acceleration is any change in velocity, whether that be direction, speed or both. The problem is that accelerating charges emit photons by a process called Bremstrahlen, or breaking radiation in German. This is not noticeable for the kind of speeds we experience in everyday life, but it is significant for the kind of motion required for nuclear reactions. A plasma is constantly radiating away energy by Bremstrahlen, with everything from radio waves to X-rays. Once they are emitted, the photons usually, though not always, leave the plasma and energy is lost. The electrons being lighter and more mobile than ions are responsible for most of this emission. The power lost due to Bremstrahlen depends on the densities of ions, electrons and the temperature of the plasma. I have said that controlled fusion aims to maximise the triple product, but the issue of Bremstrahlen means that there is somewhat of a trade-off between the energy confinement time and the other two parameters. Any good design for a fusion power plant seeks to mitigate Bremstrahlen losses by limiting the operational temperature. The power loss is also proportional to the square of the charge of the constituent ions, which is very significant. For example, the boron ions, with a charge of plus 5 in the anutronic proton boron reaction, would lead to 25 times higher losses than, say, a vegetarian plasma where the ions have a charge of plus 1. With all that in mind, let's look at confinement techniques. Stars like the Sun are massive enough to keep plasma confined under gravity, even though the temperature and pressure are high enough for fusion, at least in the centre. It is not only the gravitational force, but the matter itself which is important for the quality of confinement of a star. If a strong gravitational well were somehow synthesised and used to confine plasma in lab, as some science fiction fans have suggested, it would rapidly cool down through Bremstrahlen losses. The large volume of plasma in a star traps radiation which would otherwise escape the plasma forever. The opacity, the measure of how opaque a substance is, determines the degree to which Bremstrahlen is trapped. A paper published early on in the history of fusion research was poetic in its description. As soon as the stars were formed, capacities became one of the basic subjects determining the structure of the physical world in which we live. And more recently, with the development of nuclear weapons operating at temperatures of stellar interiors, capacities become as well one of the basic subjects determining the process by which we may all die. Due to the opacity, some of the energy radiated by one blob of plasma is absorbed by its neighbour, which in turn radiates energy as well. Energy therefore becomes trapped as photons bounce around back and forth. The larger the size of the star, the greater this trapping effect is. The greater the temperature in the centre, in the same way that putting on more layers will keep you warmer in the winter. Nonetheless, energy leaks out of the sun as evidenced on Earth during the hours of daylight. Particles of plasma are also lost from the sun in the form of the solar wind. Most of the matter in the universe is in the form of single protons, so this is the primary nuclear fuel available to main sequence stars. I will release a follow-up video about stars which mostly fuse heavier elements, such as silicon, about a trillion years from now when it becomes relevant. The cross-section for two protons to fuse is extremely low, so the reaction proceeds very slowly. If you could actually create a miniature sun with the same conditions and energy density, it would be a very poor source of energy. You would need a reaction vessel the size of the Titanic to match a modestly sized power plant. This is a good thing overall, because it means the sun has been able to keep going for the billions of years during which life has evolved, but it makes doing proton-proton reactions on Earth prohibitively difficult. Once two protons fuse together into deuterium, subsequent reactions quickly form a very stable helium-4 nuclei. Groups of four protons can also form helium-4 by another type of process. The CNO cycle, short for carbon-nitrogen-oxygen, involves heavier nuclei absorbing protons. Beginning with carbon-12, a nucleus fuses with a proton, decays, fuses with another two protons, decays, and finally, after fusing with a fourth proton, splits into a carbon-12 and helium-4. The key is that in the end, the heavier nucleus returns to its original state, so it is ready to repeat the process again. It plays a similar role to a catalyst in chemical reactions. Due to the quirks of nuclear cross-sections, this roundabout process actually proceeds faster than protons fusing directly at high temperatures. There are other similar cycles involving other nuclei too. CNO cycles dominate in stars larger than the Sun, which have more plasma to keep radiation confined, and as a result have hotter cores. Using this type of process in an artificial reactor is also prohibitively difficult, due to the low cross-sections and Bremstrahlung losses. So let's look at fusion approaches which do work on Earth. The deuterium tritium reaction has the highest reaction rate, requiring a reasonably low temperature. It also minimizes Bremstrahlung losses, as the nuclei only have a charge of plus one. It is therefore the main reaction being pursued in fusion research and the reaction which I will talk about from now on. Let me briefly mention the fuel cycle for this reaction. Deuterium is reasonably common on Earth, accounting for one end every six and a half thousand hydrogen atoms. Tritium, on the other hand, is radioactive and does not occur naturally on Earth. Any supply must be synthesized. When deuterium and tritium fuse, they produce a helium-4 nucleus carrying about a fifth of the resulting energy and a neutron carrying off the rest. The neutron can be absorbed by lithium, which will decay into a tritium nucleus and another helium. Any artificial fusion device must be engineered to breed tritium in this way. Of its two most common isotopes, it is the lighter lithium-6, which has a higher cross-section for neutron capture than lithium-7. Weapons researchers have perfected a method for separating the much more reactive lithium-6 from naturally occurring minerals which have a mixture of both isotopes. So this is not a major problem. I will go further into the problems of neutronics and tritium breeding in a future video. Without the gravitational force of an entire star's worth of mass to confine the plasma, the next best approach is to achieve the conditions for fusion and release as much energy as needed before the plasma disassembles itself. This is referred to as inertial confinement because it is the mass of fusion reactants themselves and usually a dense surrounding material which resists the outwards pressure just long enough to release energy. This is the only technique thus far by which humans have practically released fusion energy, a thermonuclear bomb. The detailed physics are of course classified and the design varies between the different nuclear armed nations. Here are the basics. The aim of such a weapon is to utilize deuterium tritium fusion to release devastating amounts of energy and also neutrons for further fusion reactions. The fusion fuel is stored in a bomb long term in the form of a solid compound, lithium deuteride. The fusion reactions are initiated by a primary nuclear fusion explosion which ionizes the fuel into a plasma and raises both temperature and density to the point that fusion happens. A primary fission device is held in a specially shaped chamber called a Holraum, German for hollow space, together with the secondary fusion device. It is believed that such a Holraum is being shown off in this image by the son of a self-professed Korean golf pro. The Holraum is designed in such a way that the primary explosion fills the cavity with radiation in the form of X-rays. This intense radiation causes the outside of the secondary fusion device to ablate away, effectively blow outwards, simultaneously causing an inwards compression. The opacity of the fusion device in large part determines the compression. The lithium deuteride fusion fuel is typically surrounded by uranium-238, which has several purposes. It has a large opacity, therefore it efficiently captures the X-rays. It provides a lot of inertia, hence the term inertial fusion, to fight against the lithium deuteride trying to push itself apart. Uranium-238 can undergo fission, but only after absorbing very energetic neutrons, such as those released by the fusion reactions. This is therefore a way to release additional energy. In terms of the triple product, the energy confinement time of inertial fusion is short. The temperature is somewhat higher than the center of the sun. It is the enormous density which really raises the triple product. The material starts off at solid density and becomes compressed by thousands of times. Thermonuclear weapons obviously work, though they rely on fission to initiate them, and as a means to release more energy. Inertial fusion is also being pursued with the goal of controlled, peaceful generation of electrical energy. The goal is to follow the same basic principle as a fusion bomb, but to scale down the explosion so that it can be contained, and to remove the primary fission stage so that only fusion takes place. This is advantageous, because fusion leaves fewer long-lived radioactive isotopes than fission. Proposed substitutes for delivering the required energy are lasers, intense electric currents, beams from particle accelerators, among others. Research into inertial fusion for power generation began with underground tests, code named halite and centurion, among others. Details are still classified, but supposedly these tests, which still use the primary fission stage, demonstrated the feasibility of scaling down the secondary fusion stage, the point that could be used in a power plant. Since then, the laser approach has been the most successful. The National Ignition Facility in the US has recently demonstrated net fusion gain. The controlled approach to inertial fusion replaces the lithium deuteride with a microscopic pellet containing a frozen deuterium-tritium mix. Just as with a bomb, the fuel must be compressed to a very high density. Fusion begins in a hot spot in the middle. The energetic helium-4 products then heat up more and more of the surrounding material to thermonuclear temperatures, until perhaps half the fusion fuel is exhausted. The reason it's done this way goes back to what I said at the start. Hot plasma exerts a higher outward pressure. Inertial fusion must therefore dance a very fine line, compressing the fuel while it's relatively cold, but not too cold that it will never fuse, and then heating it up, but not too hot that it disassembles too quickly. Lasers can be focused onto the outside of a pellet, direct drive inertial fusion, or to illuminate the inside of a small hole realm, indirect drive. X-rays ultimately compress the pellet in the indirect drive scheme in a similar manner to that of a thermonuclear bomb. For this reason, the National Ignition Facility has taken this approach first, while also performing classified experiments to maintain the US nuclear stockpile. As you might expect, the lasers at the National Ignition Facility are state-of-the-art, blazing a trail for subsequent experiments. There are 192 total beams allowing uniform illumination of any targets, though not every experiment makes use of all the beams. The lasers are solid-state neodymium-doped crystals, which allow for very intense pulses at an infrared wavelength of 1,053 nanometers. For electromagnetic waves in general, the higher the frequency, the deeper interplasma they penetrate. Each infrared laser beam passes through a crystal which triples its frequency. Effectively, three infrared photons are combined into a single ultraviolet photon. This comes at a cost of some of the total energy of the pulse, but it is well worth it in terms of plasma performance. The most important factor for power generation is what I will call the fusion energy gain, denoted by the letter Q. This is defined as the energy given out by fusion reactions divided by the energy put in. A fusion power plant, regardless of the type of confinement, must have a Q above 10 or so to output electrical energy, or else it would be a glorified electric heater. This is because heat can only be converted to electricity at 60% efficiency at most. Heating up the plasma requires a large input power, and other plant systems also take electrical input. Inertial fusion by its nature is transient and repetitive, so a good deal of input energy must necessarily be applied to every pellet. For the laser approach, the input energy is usually defined as the beam energy entering the reaction chamber after it has been generated and frequency tripled. Lasers have efficiencies up to 10 or so percent going from electrical power to the final beams. Other inertial fusion approaches have roughly similar efficiencies. The amount of fusion energy released by a single pellet must therefore greatly exceed the input energy. The big roadblock to getting a large fusion energy gain out of an inertial fusion system has been dealing with numerous plasma instabilities. Inertial fusion relies on the implosion of the fuel pellet being almost perfectly symmetric. The compressive forces must converge at the center of a sphere. Plasma instabilities, such as the Raleigh Taylor instability, churn up the compressing plasma. Look for example at these simulations of experiments at the National Ignition Facility, showing temperature on the left and density on the right. Both profiles should be concentric rings in this cutaway view, indicating perfect spheres. They are clearly far from perfectly spherical however. Instead of a central hotspot, all of the fuel becomes merely tepid. Confinement is poor, fusion doesn't get a proper chance to get going. If that happens, not enough energetic helium 4 gets released to heat up the rest of the fuel. Ultimately, this means that each pellet outputs a small amount of energy and therefore the fusion energy gain Q is also small. As an example of how difficult this is to deal with, from a practical point of view, indirect drive experiments were expected to be successfully completed at the NIF within just a few years. The search for the titular ignition, Q greater than one, has in fact dragged on for the best part of a decade. In the meantime, the delays have killed off a pair of follow-on facilities, life in the US and hyper in Europe. While this saga is not scientific proof that inertial fusion is challenging to achieve in the lab, it shows that even an experienced scientific establishment, riding on the back of successful underground nuclear tests, can have major problems achieving a sufficiently high Q for a power plant. One thing that should not be underestimated is how difficult the instabilities make it to do the theory, simulations and analysis of inertial fusion. As you may know, physicists can only solve problems involving spherical cows in a vacuum. Churned up asymmetric pellets in the radiation field are much harder. Attempts to do inertial or pulsed power fusion without lasers are just as susceptible to instabilities. A Z-Pinch is a pulsed magnetic confinement device which was thought capable of producing a fusion gain until instabilities proved too detrimental. The startup First Light Fusion is attempting to use a gas gun to accelerate and implode pellets. The implosions are asymmetric, which they have claimed makes them resistant to instabilities, though in practice the latter is very doubtful. The private company General Fusion aims to use magnetized target fusion, whereby steam pistons compress deuterium tritium fuel surrounded by lithium. Although it uses magnetic fields, this scheme is otherwise very similar to inertial fusion. It relies on rapid compression of fuel in short pulses. In all cases, instabilities will degrade the fusion performance. Hot and cold plasma will become mixed up to the point that none of it is hot enough to fuse. If lithium or other impurities get into the plasma, they will radiate away the heat by Bram Strahlen. The other really major problem for all inertial and pulsed power schemes, which is not an issue for most magnetic confinement proposals, is just how fast they would have to get through pellets. If we built an entire fusion plant with all the vast complicated systems involved, only to give out a few kilowatts of power, it would never be economical. The minimum realistic target for a fusion power plant would be at least 100 megawatts of net electrical energy, or let's say 400 megawatts of fusion power. This is equivalent to the yield of 100 kilograms of TNT every second. The design of an inertial fusion power plant must choose between having a very high repetition rate, which is technologically challenging, and having a very high yield, which is difficult to contain. If we wait 10 seconds between fuel pellets, then every one of them would have the yield of a ton of TNT. Too much, I would argue, to be absorbed by a reasonably sized vacuum vessel. The repetition rate must therefore be at least one shot per second, and ideally closer to 10, essentially getting through pellets at the rate of a machine gun. Note that the National Ignition Facility currently does one or two per day. Even on paper, this is a tricky prospect. Every cycle, a fuel pellet must undergo fusion to achieve the required gain overcoming any instabilities. The reaction chamber must somehow be cleared of any residual plasma and debris, at least to the point that they don't get in the way of the next cycle. For laser fusion, pellets must be fired into the chamber at a sizeable velocity, and then detonate it precisely in the middle of the reaction chamber. All the power systems must be ready to discharge again, whether that be powerful lasers, capacitors, or whatever. Dealing with the cyclic stresses is also likely to be a major engineering challenge. I have mentioned some other approaches such as that of general fusion or first light fusion, which do not require lasers, but must nonetheless deliver these kinds of repetition rates effectively and consistently. A lot of thought has gone into solving these problems, at least for the laser approach. Research into lasers with high efficiencies pushing 10% and repetition rates of 10 to 100 shots per second is going quite successfully. This technology is probably furthest along, and of course this has other applications too. It is generally recognized that mass production of whole realms is not practical, so laser fusion would have to take this still unproven direct drive approach. For this, pellet manufacture in large amounts has been considered. Work has begun on accelerating pellets and tracking them inside a vacuum chamber so that the requisite repetition rate may ultimately be attained. Any future fusion power plant must include a tritium breeding blanket surrounding the thermonuclear plasma. It is likely that key progress in perfecting a blanket will be done by magnetic fusion experiments, which are overall further along. I will talk about them in the next video. Thanks for watching.