 One way or another, most energy we use on Earth comes from a giant nuclear fusion reactor in the sky. In my previous video, I gave a short summary about the challenges of making electrical power from fusion energy, and I recommend you watch that first. In this video, I will go into more depth about fusion reactions themselves and answer the most common question I've received. Why can't we make electricity by forcing nuclei together in a fuser or particle accelerator? I'll also briefly touch on NASA's lattice fusion discovery. Before I get to that, I have to cover something slightly technical which involves graphs, so buckle up. All forms of nuclear power rely on the fact that a significant amount of binding energy is released going from a random collection of neutrons and protons to a single compact nucleus. The more binding energy that has been released per nucleon, that is a proton or neutron, the more tightly bound and generally stable the nucleus A hydrogen atom, a single proton, has the lowest binding energy of zero because there is nothing to bind together. Nickel-62 has the highest binding energy per nucleon, making it the most stable, with the very common iron-56 close behind. Plotted on a graph, the binding energy curve shows how this quantity increases up to nickel-62 and then falls off again as the nuclei get more massive. Any reaction which has inputs lower than its outputs on this graph releases energy. The process of nuclear fission does this by breaking apart heavy nuclei, such as uranium, while fusion does it by combining lighter ones together. The thing that stands in the way of two lighter nuclei fusing together to release this binding energy is the fact that they are both positively charged and hence repel each other. One or both nuclei must have sufficient kinetic energy to overcome that initial repulsion until they are close enough to effectively combine into a single nucleus. Even if nuclei do have enough energy when they collide in this way, it doesn't necessarily mean that fusion will happen. Anything the size of an atom or smaller is bound by quantum mechanics, which are fundamentally probabilistic in nature. It is impossible to be certain if a collision will lead to fusion, but only give a likelihood that it will happen. The most common measure of the probability of fusion reactions taking place is the cross-sectional area for a given reaction. In simple terms, this is an area each nucleus sweeps out as it moves in which a given reaction is likely to take place. For a large number of nuclei all interacting simultaneously, the cross-section is a measure of the reaction rate. Although the fact is being equal, the cross-section determines how quickly energy will be given off. The cross-section is small when the energy of the nuclei is small because it is that much harder to overcome the repulsion. Due to quantum effects, the cross-section also becomes small at high energies where interactions decrease. There is a sweet spot in terms of energy where the cross-section is largest, but it varies from reaction to reaction. Let me take you through some important fusion cross-sections and that energy sweet spot of each one. I'll give the size of the cross-section in terms of a unit called a barn, which was jokingly named because in atomic terms you couldn't hit the side of a barn. I'll give the energy in terms of thousands of a unit called an electron volt, which I'll explain in a moment. What I really want you to take away though is the relative size. We want the cross-section to be large and the required energy, which we must supply to the nuclei, to be low. Juterium tritium fusion has the highest cross-section at about five barns and an energy of 64 kiloelectron volts. Juterium fusing with helium-3 has just under a barn at 250 kiloelectron volts. For the reaction between a proton and the common boron-11 isotope, it's 1.2 barns at 550 kiloelectron volts. These are the most probable reactions, but they are between relatively rare isotopes. The sun is mostly composed of single protons. The cross-section for any two of them to fuse is a whopping 10 septillion, or a factor of one, followed by 25 zeros, lower than Juterium tritium. Even though there is plenty of energy to be released by making anything up to nickel, the cross-section for those reactions is also very small. So let's take the most probable Juterium tritium reaction. The cross-section peaks at 64,000 electron volts. This unit of energy is so-called because this is the energy which an electron gains going through one volt of electric potential. In other words, if I put the terminals of a 9 volt battery in a vacuum, I could accelerate an electron to 9 electron volts of energy from the negative to the positive, or a hydrogen atom to the same energy going in the opposite direction. Given that electronics can maintain a constant voltage up to a few million volts, it is totally viable to accelerate nuclei to fusion relevant energies using electrodes in this way. The simplest design would be to have the tritium nuclei held fixed, either frozen solid or simply as a gas, and to use a constant voltage to accelerate a beam of Juterium atoms. This is referred to as beam target fusion. There's a small subtlety of center of mass versus lab frame I won't go into, but to be safe let's bump the accelerating voltage up to 100 kilovolts. Okay, so we supply 100,000 electron volts of energy to each Juterium atom, but after it undergoes fusion, 18 million electron volts will be released. If every accelerated nucleus did so, this would be an energy gain of 180 times, more than enough to be engineered into a power plant. Here is the problem, however. Every time two atoms collide, they are much more likely to simply bounce off each other in a process called alternatively Rutherford or Coulomb scattering. Here are the fusion cross sections on a graph, showing how each one rises to a peak and then starts falling off again. On this logarithmic scale, for every division on the y-axis, the cross section changes by 10 times. Way up above the fusion cross sections is the Coulomb scattering one, which is several thousand times larger in our scenario. Let's be optimistic and say just one thousand. This means that for every collision which results in fusion, there are over a thousand which do not. In our simple setup, once an energetic Juterium nucleus scatters off a stationary tritium, the two will exchange energy. The Juterium will lose energy, making it even less likely to fuse, and it will eventually be brought to a stop by this type of scattering. We have basically divided our fusion gain by a factor of a thousand, to where it is less than one, and way too low to make electricity. The problem was that the stationary nucleus stole a fraction of the kinetic energy, so a better idea might be to have counter propagating beams colliding with one another as they do in research accelerators like the LHC. Scattering will still happen, and so the vast majority of our nuclei will be deflected out of the beams and go to the walls of the apparatus where they will lose all of their energy. A device called a fuser has spherical electrodes allowing nuclei to be accelerated inwards from all directions to collide and fuse. With this design, there isn't a tight beam for the nuclei to be scattered out of, but they will still exchange energy unfavorably amongst themselves. Now, there are other problems with each of these designs and other ways that energy is lost, but it is that thousand to one or worse ratio in cross sections which is the real killer for doing fusion this way, at least if you want to get useful energy out. To be perfectly clear, fusion does happen and it is detectable, it just doesn't happen nearly enough to be worthwhile. So, how could fusion ever be useful if the vast majority of the energetic nuclei end up losing their energy through collisions? Well, imagine if all the nuclei already had a large average kinetic energy and collided with one another frequently. This picture of particles randomly bouncing around is precisely the model of particles in a gas which may be familiar from high school. The higher the temperature, the larger the average energy, though I must stress that many of the particles have one which is much higher or much lower than average. While any single collision might transfer energy unfavorably, over time they will all average out. If the temperature is so large that many of the more energetic nuclei have enough energy to fuse, then this is called a thermonuclear reaction. This is how we all get our fusion energy, especially those of you in the comments who correctly but sarcastically say that we have a thermonuclear reactor just eight light minutes away. In practical terms, thermonuclear fusion requires temperatures in excess of 10 million degrees. On this graph, all that business of the random thermal motion with some slow and some fast nuclei has been folded into the cross section so that for every temperature we have a true measure of the reaction rate. The deuterium tritium reaction wins out over all the rest and doesn't need a particularly enormous temperature. For context, the center of the sun is at 15 million degrees, a thermonuclear bomb is in the range of 100 million degrees, and magnetic fusion experiments have reached peaks of 300 million at the joint European torus. The MIT arc proposal and PPPL both target this sort of temperature. The Wendelstein 7x stellarator has set a record for its class of device, but only at 40 million. Suppose we achieved good power output from a deuterium tritium reactor. The tritium is a problem because it's hard to make and it gives off neutrons, so we decide to replace the fuel with a deuterium helium-3 mix instead. This will absolutely clobber our power output when running at the same temperatures before. We could claw back some of that power if we up the temperature or density, but of course this increases the physics and engineering challenges. Note that in this case some neutrons would still be produced as deuterium nuclei would fuse with each other. The same kind of argument applies to the proton boron approach. The advantage is that this reaction releases no neutrons whatsoever and the fuel is commonplace. The major disadvantage is that the temperature must be higher and even then the reactor must be made larger or higher pressures must be applied to squeeze more nuclei into a given volume. This is the kind of challenge which companies like tri-alpha energy are facing when they promise to do proton boron fusion. That's not to say that these challenges can't be overcome, but the scientific community and most startups have taken the sensible approach of doing the easiest reaction first and then perhaps moving to a more challenging and slower reaction. If we wanted to do proton-proton fusion as in the sun, we would need to compensate for the septillion times lower reaction rate either by packing in the protons to astronomical densities or by making the reactor quite literally astronomical in size, so basically the sun. When matter is heated to extreme thermonuclear temperatures, bonds between molecules break down, molecules themselves are torn apart, collisions cause practically all the atoms to be stripped of their electrons and become ions. Ions and electrons flow past one another as fluids. This is a state of matter called a plasma. The reason for the name is that Irving Langmuir, an early pioneer in the field, made the comparison with blood plasma. The biological term refers to this amorphous fluid with various electrolytes and proteins just as the physical plasma has multiple types of ions, electrons and so on. Fusion takes place in plasmas and the study of fusion is primarily concerned with plasma physics. There is one final example of beam target fusion which I want to mention. I had some comments that it would yield great results if only the target itself were a hot plasma. Actually this happens very regularly in experiments. In magnetic fusion devices, both in Tokomax and Stellarators, deuterium plasma is confined at thermonuclear temperatures. Deuterium-deuterium fusion takes place, not enough to generate electricity but in significant amounts which are measured with high accuracy. Outside the reactor, deuterium ions are accelerated by a voltage of between 100,000 and 1 million volts and then have an electron added to make the ion neutral again. This way they can pass undeflected through the magnetic field and into the plasma. As they collide with other nuclei in the plasma, some do indeed undergo fusion to the point that an accurate accounting of the reaction rate must take this beam target component into consideration. However, both simulations and experiments agree that this is only a minor addition to the fusion power output. It is the fact that the beam particles collide and give up their energy to heat the plasma, which makes the thermonuclear fusion rate as high as it is. I hope that this proves my point that beam target fusion or variations on the idea could not produce electrical power on their own. One important reason to keep this fact in mind is that many proposals for fusion energy are basically beam target fusion in disguise. You could use a big gun to fire frozen chunks of deuterium into a target or two guns or lasers or some other scheme, but it is doomed never to produce power output if there is not a significant thermonuclear component. This is my initial thought when looking at the lattice fusion discovery made by NASA. The general problem was that the repulsion by positive nuclei made the fusion cross section low and the scattering cross section high. The NASA led team have reportedly been able to screen some of the positive charge of the nuclei, meaning that the cross sections became more favorable for fusion. However, to enable fusion reactions to happen, the team had to supply gamma rays each with 2.5 million electron volts of energy. Remember the simple beam target idea from before could make fusion reactions happen by supplying a mere 0.1 million electron volts. There is also a big efficiency cost to generating gamma rays, maybe in space where gamma radiation is common this could be viable. For now, this very recent result must first be reproduced and the hard numbers must be crunched for whether or not useful electrical energy could be generated, or if this is just a better fuser. The latter may also be valuable. As the summary of the NASA article states, process scale up using an energy efficient linear accelerator may lead to a new means of generating or boosting medical and industrial isotope production. The linear accelerator there is for producing the gamma rays. Let me finally state that there is no credible proposal to achieve cold fusion, that is overcoming the repulsion between nuclei in a fundamentally different way than either by having something beam-like or by means of thermal motion. Those demonstrations which do show signs of fusion reactions rely on a hidden variation of the beam target approach. If it were that easy to do fusion reactions and not bother with creating and maintaining a hot plasma, every DeLorean in the world would have such a reactor. So if you think you can do it, go ahead and prove me wrong, but what I've stated is the scientific consensus. For the record, I am also in full agreement with the scientific consensus that human activity is rapidly altering the climate of our planet. I don't recommend subscribing to this channel if you think otherwise. Thank you for watching. In the next video, I will address the most important subject for fusion research, how to raise matter to thermonuclear temperatures and keep the heat and particles confined. I'd also like to pre-empt any comments that I was not accurate to 15 decimal places that I used a figure of 60 rather than 61.2 or whatever. I intentionally used round numbers to make things clearer. I am trying my best to explain these topics to a wide audience, not to graduate students or experts in the field.