This a very strange, and incorrect, conclusion you have come to. It doesn't matter that datetime fields are unique, as most of the time you are not searching for a particular date time, but a range. I.e. "show me all of rows created between date X and Y". In that case, the ordering of the index makes it efficient to do that query.

Furthermore, often times your query will have an "ORDER BY dateTimeCol" value on it (or, more commonly, ORDER BY dateTimeCol DESC). If your index is created correctly, it means it can return rows with a quick index scan instead of needing to implement a sort as part of the execution plan.