The table scan occurs if the Engine determines that a table scan will be more efficient. If you have null values or empty strings in your column, they will be included in your non-clustered index if you create it.

You should test the creation of your non-clustered index to determine if the new index will help improve performance. One of the factors to consider is column selectivity and uniqueness of the data in that column. The query that is causing the table-scan may also need to be revised.