Thanks for the reply.
(1) Below is the extract from wikipedia:

For example, Given an employee table consisting of the columns:
job and

we could use the employeeID in combination with any or all other columns of this table to uniquely identify a row in the table. Examples of superkeys in this table would be {employeeID, Name}, {employeeID, Name, job}, and {employeeID, Name, job, departmentID}.
In a real database we don't need values for all of those columns to identify a row. We only need, per our example, the set {employeeID}. This is a minimal superkey that is, a minimal set of columns that can be used to identify a single row. So, employeeID is a candidate key.
Now, if employeeID is a candidate key then why not it is the superkey. Because employeeID can also uniquely identify the tuples.

(2) In your example why Roll number is not the superkey as it is uniquely identifying the tuples?

Any Unique key with some Non Unique key combination is called a super key of the relationship
Is it neccessary that Unique key has to be combined with some Non Unique key to be called as a super key.