Again, this is a yes and no answer. You can file a provisional patent application once you think you know what the invention will be, spend the next year working out all the details, and then file a non-provisional patent before the provisional expires. (Provisional patents are not examined, go abandoned at the end of 1 year, and are never published unless a provisional which relies on it is later published.) Basically this is a fairly cheap way to preserve your date of invention to the date you filed your provisional patent application, while keeping it secret if you decide later you don't want it published. However, the provisional patent application will only preserve the date of invention for the material that it actually contains! Any new material will only have a date of invention to the filing date of the non-provisional application. Since the provisional patent is not examined, you can spend much of the intervening year before filing the non-provisional on making and testing examples, basically just proving that your invention did exactly what you say it would do.
Let's take your catalytic example. You've done a lot of theoretical head scratching, and you've figured out that you should be able to make a nine-coordinate complex with perfectonium using idealium as the ligand. You've even worked out the reaction conditions it would take to make your catalyst. IF all of your theories and your understanding of the science is correct, your invention is now complete, and you can write up your patent application. This is called "constructive reduction to practice" - you don't actually have to have the invention in your hands, but you have to be able to write an application such that another skilled person in the field can take it and make the invention following your instructions. You write up your application using prophetic examples, which is perfectly reasonable, but must be written that way. You don't report yields or analytical data for reactions you haven't run, you just write that if idealium and perfectonium are mixed in a 9:1 ratio at an elevated temperature and allowed to cool, you will obtain a nine-coordinate complex useful for the catalytic reaction you described.
So now you go off into the lab and start working on your invention, and you find that perfectonium and idealium make a nice nine-coordinate complex just like you predicted, and you make a lot of examples under a lot of conditions, but it just isn't very active as a catalyst. You're about ready to give up on your invention and let your provisional go abandoned when one of your lab assistants gets the brilliant idea to play "Pink Floyd" loudly through headphones attached to the sides of the beaker, and presto! The ligands get twisted into just the right orientation to form a much more active catalyst. You add a method claim to your patent including the playing of "Pink Floyd", add your assistant's name to the list of inventors, throw in some composition claims about twisted ligands, and file your non-provisional patent application.
The trick here is that the priority date of your non-provisional application is the filing date of the provisional application, but the only material in the non-provisional application that is entitled to that priority is the material that was in the patent office on the date the provisional was filed. Suppose that in between the time you filed your provisional patent application and the time you filed your non-provisional patent application, somebody else filed an application covering the use of "Pink Floyd" on nine-coordinate perfectonium-idealium systems to form catalysts. You will still be entitled to your patent on the perfectonium-idealium catalyst because you made a lot of them and they do actually work as catalysts, but the other inventor has priority on the improvement of the catalyst system using "Pink Floyd". At this point you either cross-license the patents so that both of you can market the catalyst systems, or you end up in an epic patent litigation war.