We need a lot more detail.   It appears likely that the data you have is not standard English characters, but just changing a datatype from varchar to nvarchar only provides the means to store unicode characters.   As the column was varchar at the time it was populated, if there were unicode characters being used, then such characters were likely converted, possibly implicitly, at the time the data was inserted or updated into the table, and one of the possible consequences is that information got lost in that conversion.   You'll need to provide more specific details and as much specific history of events as you can in order for someone to be able to help much.   One other possibility is that you just have foreign language characters, and you may just need to set the proper collation for that column, but that seems a lot less likely.