Hello. It would be far easier to provide you with a list of legal / valid characters / code points. This is because Unicode has 1,114,112 total possible code points, and 46,543 of them are valid for identifiers (SQL Server 2005 to at least SQL Server 2019). Hence, there are 1,067,569 illegal / invalid code points (not all of those have been allocated to actual characters yet, but more are being added each year).
I posted the list of legal / valid characters here:
Completely Complete List of Valid T-SQL Identifier Characters (for SQL Server 2005 through at least 2019)
Of course, "valid" / "legal" characters are ones that can be used in "regular" identifiers (non-delimited identifiers). For delimited identifiers (i.e. ones that require being enclosed in either square-brackets or double-quotes), I believe all code points are legal, except U+0000 and U+FFFF.