Lexical analyser uses the regular expressions as defined by the language specification to output a set of tokens for the synatax analyser whose output is a systax tree after matching the syntax.
So, answer is D.
Option B is not wrong as each token can be considered a string of tokens, but D is a better choice.