I'm being nitpicky about this - but let's avoid using the word PK in this case. Its function as Primary Key has no bearing whatsoever on the page splits. The clustered index is the only thing driving that decision. The fact that they're both on the same field is incidental only. The fact that SQL server allows us to use the same index for both is regrettable IMO, but again - that's not the question of the moment.

If your primary goal is to prevent page splits, then actually - an identity is a GOOD choice for a clustered index, since all of the inserts will have at the "end", and not in the middle (already full) pages. Now - you don't get it to leverage the clustered index in range seeks/scans that way - but again - that's a choice based on what's most important.

Also - Heaps don't have page splits. They're a byproduct of having a clustered index.