Does neutron radiation form clouds?
Thermal neutrons capture on hydrogen and carbon with reasonable (i.e. not large, but significant) cross-sections (this is the delayed event detection methods of most organic liquid scintillator anti-neutrino detectors--i.e the one that don't dope their scintillator with Gadolinium).
So though a "cloud"--meaning a localized diffuse gas--of neutrons can develop in the neighborhood of a strong source (size of the cloud is driven by how far they go as they thermalise), their dissipation is driven by their mean capture time, not their half-life.
Confession: Here I am presuming that the mean capture time is significantly shorter than the half-life, but I haven't measured it in a "near the laboratory" setting. In organic liquid scintillator the capture time is on order of $200\text{ }\mu\text{s}$, but air has a lot less hydrogen and carbon in it. Note that the neutrons also go into the ground, the building, nearby vehicles and passers-by (if any) where they may find things to interact with.
At my grad-school we had a 2 Curie (i.e. huge) AmBe source. The source vault would register unusually high back-grounds on a survey meter for a few minutes after it was returned from the moderator tank to the shielded vessel, so that may be a rough measure of the time scale. It also says something about the strength of the radiation field: a few times the in-the-basement background level.
Shielding methodology for strong neutron sources generally incorporates a great deal of boron in various layers to help suck up the thermal neutron flux; not incidentally this means that most of the capture gammas are generated inside the shielding. Borated plastics are common as are borated concretes. These days Gadolinium is cheep enough that I imagine we'll start seeing it used in shielding design. The source vault in grad school was built of borated cinder block---two layers with a meter air-gap between.
Another not-very-quantitative story that might shed some light on this.
I was friends with one of the Radiation Safety guys at JLAB. Part of his job was monitoring the radiation level at the fence around the secure area with the accelerators, experimental halls, etc. Mostly they just put out general purposes detectors and compared the results with background reading from nearby, but early on they built a more sophisticated detector out there to understand the various contributions to the dose (probably trying to tune their Monte Carlos, those guys are really big into modelling). He told me two interesting things
- If they ran the accelerator at high current and high duty cycle they could about double the dose at the fence (i.e. the accelerator related dose was as big as the background at the fence).
- Neutron sky-shine was the single biggest contributor. Sky-shine means that the neutrons got out through the lightly shielded roofs of the halls (only 50 cm of concrete and 2 meters of packed earth), and their detectors saw radiation coming from the captures/decays that occurred above them.
The fence was about 40 meters from the beam dumps.