Japan is in the midst of a slow-motion nuclear meltdown. Each new day brings word of further problems. At this point three reactors have been flooded with seawater, and appear contained (at least for the time being). The news reports are incoherent and conflicting, and nobody seems to really know what’s happening. This may be because the information is not public. Or it could be because the situation on the ground is fundamentally incoherent. You can’t exactly walk up to reactor #2, open the door, and take a peek inside. Amazingly, the best up-to-date resource appears to be wikipedia (which incorporates the useful summary tables from the Japan Atomic Industrial Forum).

The earthquake happened at 2:26pm. Two minutes later, the Fukushima-Daiichi nuclear plant went into SCRAM mode, and shut down. The control rods were inserted. The diesel generators fired up. Everything worked to plan.

The Fukushima-Daiichi plants are boiling water reactors. In simplest terms, this is just a pile of radioactive material (generally uranium) which gets hot (literally hot, not just radioactive). You run water over it, generate superheated water and steam, drive a turbine, and produce electricity. Instead of burning coal, you use radioactive decay as the source of heat, but otherwise the basic mechanism is surprisingly similar to a conventional power plant. You turn off a nuclear reactor by inserting control rods, which absorb a lot of the neutrons, and inhibit further fission reactions. So, two minutes after the quake, the control rods were inserted, and the reactors were no longer undergoing nuclear fission. However, one of the peculiarities of nuclear power is that even after the reactor is shut “off” there is still a significant amount of residual radioactive material. This material continues to decay, generating significant heat (>10 megawatt; by now [almost a week later] it’s ~1 megawatt, enough to power a thousand homes). Thus, even after turning a reactor off, it still generates significant power for a few weeks, and the resulting heat needs to be removed and the radioactive core kept cool . And to do this, you need to pump in a lot of water (ideally thousands of gallons/min) at high pressure. And this requires a fair amount of power.

The plant was working perfectly for roughly 30 minutes after the earthquake. The tsunami was on its way, but the plant operators were blind to it. Had they known, they could have depressurized the nuclear cores in anticipation. But they were focused on riding out the earthquake, which they did admirably. And then the tsunami hit. Just a few years ago, after the tsunami in the Indian Ocean, the Fukushima-Daiichi plant was upgraded to deal with a worst-case, 5.3 meter tsunami. The wave that hit the plant last Friday was roughly 10 meters high. It swamped the diesel generators, as well as the fuel tanks and the switching station. The system was “live” because of the SCRAM, and the local electrical grid got fried. Fortunately there were backup batteries, which lasted another 9.5 hours. At around midnight the batteries ran out of power, and the plant was no longer able to cool its reactor cores. At this point, the Troubles began.

As the core starts to heat up, it boils off the surrounding water. Eventually the fuel rods are exposed to the air. This causes the core to heat up even faster, and also causes a reaction with the zirconium cladding (which holds the uranium fuel pellets in place), generating hydrogen gas. Without any cooling, the fuel gets hot (> 1500 K/2200 F), and starts to melt. The hydrogen gas collects, and eventually explodes (think Hindenberg). This happened in reactor #1 on Saturday, blowing the roof off of the reactor building, but leaving the containment vessel (which is ~1 meter thick steel) intact. On Monday a similar explosion happened to #3, and on Tuesday there was an explosion at #2. Both of their containment vessels were probably compromised. Rupturing a containment vessel is very bad. So long as most of the radioactive material is contained, the damage to the outside world is similarly contained (modulo venting of various radioactive gas, which has been happening, but not at profoundly dangerous levels). Once a containment vessel is ruptured, the radioactive material can end up anywhere; the sky’s the limit. Chernobyl did not have a containment vessel.

The current situation seems to be that seawater is being pumped into all three broken reactors (#1–3), and they are in thermal control. It seems likely that all three sets of fuel rods are partially melted and damaged. It also seems likely that the containment vessels in #2 and #3 have been compromised, although probably not severely. There are some concerns about spent fuel rods in pools near reactors #3 and #4. So long as the rods are covered in sufficient water, they are stable. If the rods are exposed, they heat up. And when they get hot, they start to burn through their cladding, and emit radioactive material. These pools are not within containment vessels, and therefore they are potentially even more dangerous than the cores of active reactors. Their radioactive emission goes directly into the surroundings. But so long as there is water in the pools, they should be fine. The latest claim (by the Chairman of the United States Nuclear Regulatory Commission) is that the storage pool at the #4 reactor has little to no water. If true, this is a very ominous development.

This is by far the most dire situation on the planet at the moment. It has the world’s attention. We’ve had almost a week. Why can’t we just fix it? There are a number of serious complications. First, there’s the issue of radiation. People are unable to walk up to most of the buildings and see what’s going on, lest they get immediate and severe radiation poisoning. There are remote sensors and cameras, but fundamentally everyone is guessing as to what’s happening inside. Even if we knew exactly how things looked, it’s still a major engineering feat to get the appropriate amount of water running through these highly complex systems to do the cooling. There have been explosions, there are stuck valves, there are broken pumps, there are ongoing fires. The world’s resources are focused on this problem. Millions of lives potentially depend upon the outcome. And, thus far, progress has been haphazard and halting, despite heroic efforts on the part of the Japanese crew. The engineering challenges may simply be too great.

The worst-case scenario for the Daiichi reactors plays out something like this: 1. the storage pool at #4 is indeed dry. Because it’s uncontained, the radiation levels in the area get very high. Everyone needs to evacuate the complex. 2. Without anyone manning the cooling systems, the cooling stops. Everything overheats. 3. There are various explosions, resulting in a breach to a containment vessel. 4. There is a subsequent steam explosion, and a plume of radioactive material is generated. 5. Wind carries the plume in the direction of Tokyo (world’s largest metropolis), a mere 140 miles (225 km) away. We can’t even contemplate trying to evacuate and treat a city of 35 million people. As far as I can tell, things do not appear to be headed in this direction. But such an outcome is unfortunately not outside the realm of possibility, and just contemplating this should freak you out. But, to reiterate, it’s very unlikely, and a lot of things would have to go catastrophically wrong. I’d love to quantify just how unlikely, but cannot. My guess is that nobody can, since there are too many uncertainties, and we’re fundamentally in uncharted territory.

The best-case scenario, and probably most likely, is that the Fukushima-Daiichi plant will limp along, but without any catastrophic events (such as a major Chernobyl-style radioactive explosion and fire). The fuel will continue to cool, the fires will be put out, the amount of radiation will subside, and eventually the entire site will be entombed and become a testament to human hubris.