I think part of the problem arises from trying to wrap up the whole process too tightly. An 'apparatus and method' patent should be easier to obtain. This is where you just describe the critical hardware configuration you want to protect and then say ' when containing hydrogen or one of its isotopes at a pressure between x and Y and a temperature between a and b the following effects are observed.' And then you claim the 'effects' - radiation, electricity, whatever without attempting to patent physics itself.
Call it a 'novel muon generator' if you like.