The unicode character U+0080 (which is a control character: PAD) is encoded as two bytes in UTF-8, which are 0xC2 and 0x80.
The byte stream containing only the byte 0x80 does not represent a valid UTF-8 encoding of anything on its own (i.e. that byte alone is a misformed UTF-8 stream).
Basically each byte in a UTF-8 stream can be classified as one of three different types:
- single-byte sequences: bytes in the range 0x00-0x7F (0-127) represent a single Unicode Codepoint on their own (this part is equivalent to the old US-ASCII encoding)
- leading bytes: bytes in the range 0xC0-0xFD (192-253) start a multi-byte sequence and indicate how long that sequence must be *
- continuation bytes: bytes in the range 0x80-0xBF (128-191) are the rest of a multi-byte sequence.
0x80 is a continuation byte, so it can't stand on its own (it has to be preceded with a leading byte and possibly some other continuation bytes to be valid).
The Wikipedia article on UTF-8 has some very extensive documentation with good examples.
* Note that some leading bytes can never appear in valid UTF-8 for various reasons, so the range of actually used leading bytes is smaller. Similarly 0xFE and 0xFF are just never used so can also never appear in valid UTF-8.