I hit a situation similar to this a ways back--tried to reindex the database, and it triggered a database autogrow that failed due to lack of space. The exact details are fuzzy just now, but basically what's going on is that, when a table is being worked over, a second copy is being made (of it, the indexes, the clustered index... it'll say somewhere in BOL), and you need to store that in your database.

Essentially, SQL is making a sorted copy of table/index object, and only when that's done is the original dropped. So, for a brief stretch, you've got "double data", and odds are your database needs to grow to hold that much.

(I don't know but that "sortintempdb" might ease the pain...)

   Philip