They claim the temperature of the active zone to be about 2000°C. This is lower than most rocket exhaust. Still they claim twice the specific impulse of a chemical rocket; this means that the reactor is very-very lightweight.
I wonder how much radiation protection does it have, and whether the exhaust would be acceptable to use for a launch from Earth surface.
The specific impulse does not take the engine (or reactor) mass into account, but it's (approximately, if you neglect pressure effects) proportional to the exit velocity of the gas flying out of the nozzle. The exit velocity in turn is proportional to the square root of (temperature divided by molar mass of the exhaust gas).
Thus, if you use cooler hydrogen gas rather than hotter water+hydrogen, you can still make it a lot more efficient given that water has 9 times the mass of hydrogen per molecule.
One clarification: thrust-to-weight says something about an engine’s mass (by definition), but specific impulse only says something about the exhaust velocity of the propellant (i.e. the impulse per unit of mass), not the mass of the engine.
So the high Isp is due to the very light exhaust (hydrogen, not water or CO₂) which gives the higher velocity to the gas molecules at a lower temperature?
In a sense, yes. For a given temperature, hydrogen, with 1 atomic mass unit, will have a much higher velocity than water (18 amu, so √18 ≈ 4.2 times as slow) or carbon dioxide (44 amu, so √44 ≈ 6.6 times as slow).
Yes, having lighter exhaust also helps you raise your ISP. Scott Manley has a few videos on this topic, where he explains the science behind this much better than I ever could.
I believe a tank full of liquid hydrogen would be a reasonable neutron radiation shield. Generally things like water or concrete are used for neutron radiation shielding due to the large amount of light nuclei (hydrogen) in those.
You want to be careful to avoid heating your hydrogen fuel, it’s slowly evaporating through the tank even when near absolute zero, any significantly higher temperature is going to increase that rate.
I envision the tank being between the engine and the capsule, not wrapped around the engine or something like that. As another commenter points out, you only need a shadow shield. I don't think there would be much more heat transferred from the engine to the fuel tank than in traditional engines; the neutrons would warm up the fuel some but perhaps naively I assume that effect would be fairly small. Now that I think about it more, I'm not sure. Is it the neutrons that heat the water in a nuclear reactor?
On NTRs (and space reactors in general) you only need a "shadow shield" that casts a radiation shadow in the direction of thing to protect (crew, sensitive systems).
This makes the necessary shielding much less than a full sphere, reducing the mass to a mere fraction.
Thing get a bit hairy though when you need to dock with something or even when running multiple engines (the neutron balance would apparently be totally wrong).
I wonder how much radiation protection does it have, and whether the exhaust would be acceptable to use for a launch from Earth surface.