My Theory of Computation text mentions a lot of unsolved problems. Some are in CS of course, but I was just editing an example about whether for all n there is a prime between n-squared and (n+1)-squared. That seems like a problem a US sophmore could appreciate.