Simple SQL Query -- Perhaps You Can Help

I hope it's considered appropriate to post a SQL query question here; it's not I'll gladly post this question elsewhere.  That having been said, I've run into a simple but quite tricky SQL query and was wondering if the community might be able to help.  Here's what's going on:

I have a "Documents" table.  Let's call it D
I have a "Benefits" table.  Let's call it B.
There is a many-to-many relationship between D and B so I have a third table to store the relationships.  Let's call that table BD.

When I create a document it creates a row in D.  Let's call that D1, D2...DN. 
When I create a benefit it creates a row in B.  Let's call that B1, B2...BN

When I associate a document to a paritcular benefit, let's say that if D1 and B2 are associated, that I created row B2-D1 in my BD table.  Now here's what my challenge is.  Suppose I have these entries in my tables:

D: D1, D2, D3
B: B1, B2, B3

Table BD:
B1-D1
B2-D1
B2-D2

When the user goes to add documents to benefit B1, I want to show the user only documents that are not yet added to B1.  In this case, I want to show the user D2 and D3 but not D1 since that one's already added.  How do I write a SQL query to do this?

Here's what I've been using, but it shows too many results (e.g. included D1, D2, D3)

SELECT D.*, BD.*
FROM Documents AS d LEFT JOIN rel_BenefitsDocuments AS bd ON d.doc_id = bd.doc_id
WHERE (benefit_id <> @benefit_id OR benefit_id IS NULL)

Any thoughts?