The idea that cesium is present in metallic form is chemically very dubious.
Cesium is extremely reactive, as is noted. In particular, it will readily reduce U(+4) to U(+3). Nuclear reactor fuel is primarily uranium dioxide, so there is ample material there for this putative metallic cesium to react with. Cesium is the most electropositive element, so it will give electrons to (reduce) almost anything.
The state of cesium in the vapor gap will be relatively volatile cesium compounds, like cesium iodide. The core temperature of a uranium dioxide fuel pellet greatly exceeds the normal boiling point of this salt.
> Cesium will be the primary radionuclide released in a nuclear waste accident because it is present in what is called the fuel-clad gap. This gap is the space between the fuel pellets and the inside wall of the metal tube that contains the fuel. This “gap cesium” can be released in any event where the cladding is breached. Cesium is a highly reactive metal and even a small break in the seal will release significant amounts of it. Cesium burns spontaneously in air, and will explode when exposed to water.
Obviously the "highly reactive" applies to elemental cesium and is meant to imply that a collision would be a serious problem because exposing it to air would cause a big fire and release a plume of radioactive material. If that isn't the case then it seems like the thesis of the paper is rubbish?