Unless there’s a bug in the reader’s firmware, no. And even so, you’re only likely to crash it at best, not disable it permanently. The only one I know of is the Halo scanner that can’t handle EM4xxx UIDs with zeros - and that doesn’t kill or crash it: it just fails to report them.
Having said that, you raise an interesting point. It might be worth investigating for one reason: I have a feeling most engineers who coded such firmware probably didn’t expect the card to return malformed datagrams. Or said another way, most readers probably don’t treat data returned by a card as unstrusted input, like they would a regular user input.
So there’s a good chance that you might crash more than a few readers by returning oversized datagrams to cause a buffer overflow, or datagrams containing “impossible” values in key bitfields to trigger division-by-zero errors.
If you’re old enough to remember, there was one infamous such bug in Windows 3.11, 95 and NT: if you sent a very large packet to a Windows computer on port 139, you would crash the entire OS. Back in the days, there was a considerable number of script kiddies having fun crashing computers on IRC chatroom.
I’m willing to bet you probably can use the same trick to crash a good few RFID readers - and possibly lock up a few of them permanently, if they’re silly enough to write that malformed data, or something that derives from it, in flash.
Other devices that might be susceptible to exploits are readers with undocumented features, like special “programming” or “configuration” cards. These devices treat certain cards as “master keys” that unlock certain configuration options, or set certain things. If you manage to identify what makes a card special for such a device, you could possibly set the device in an impossible state, disabling it.