The issue is because the field is being tokenized using the standard analyzer. If you have a non-analyzed version you can search for an escaped substring.
Here my query is being tokenized so the slash gets removed, and the tokens match the tokens in the field, but not the slash (so as you mention, it will match more than you want)
Here I am searching on the non-analyzed version of the field, so I can search an exact substring. Using * in the query causes the query to not be analyzed to my slashes are preserved.
Here you can see the two fields are indexed differently:
Hope this helps!