Well, it is a complex phonological environment.
This is what I wrote on englishforums for a similar question.
In the following cases, the first V is stressed.
1. VtVn, t is nasally released.(cf. button, mutton, glutton, Manhattan, written, beaten, Clayton, etc)
2. VntVn, t is nasally released (cf. Trenton, NJ; Scranton, NJ; Mountain View, CA; sentence, etc). The first V is nasalized. 3. Vrtn, t is nasally released (certain, Morton, hearten, Dumbarton bridge, Wharton school, Barton, etc)
4. Vltn: in some words, t is nasally released; in some others, it is not.
4a. No nasal release, instead there is a schwa between t and n in these words: Hilton, Milton, Dalton, Elton, Shelton, CT, Bolton, Dolton, IL, etc. MW transcribes Dalton with a nasal release, but when you hear the Dalton's sound file there, it is not nasally released. However, LPD transcribes with a nasal release and the sound file matches with it.
4b. Nasal release in these words: Walton (from MW), sultan, etc.
How about words like gluttony? Here, you see sonorant gemination. gluttony = glutton (pattern 1) + nee