You are attempting to plot a function handle.
The plot function expects data, not a function to evaluate.
Your definition of 'a' takes 'U' as a parameter but then ignores it. Also the syntax of
fsolve(f-0.5,0.1);
does not look correct. I don't have access to the fsolve function, but I assume it should be more like:
fsolve( f, -0.5, 0.1 )
if you are using -0.5 to 1 as your limits. Or rather, I assume you want U to be involved in there somewhere otherwise why pass it as an argument to your 'a' function?
