A word at the beginning: DictionaryForMIDs runs on several platforms, including traditional phones, Android, PCs and modern web browsers.
For dictionaries with non-latin characters, typically an UTF encoding is used for the character set, with UTF-8 being the most frequentl (UTF-16 would be fine, too).
PCs, modern web browsers should be able to display all Unicode characters out of the box; I also would assume this applies to Android. Traditional phones (= Java ME technology) usually are not capable to display all Unicode characters; for this reason DictionaryForMIDs supports the bitmap fonts on Java ME. Bitmap fonts allow to display on traditional phones all characters of a dictionary even if this is not directly supported by the traditional phone.
The Java language supports the Unicode characters. I just looked up, the Unicode code for the glottal stop is U+0294, there seems to be an U+02C0 variant and some others.
After that long explanation: can you post your inputdictionaryfile, or an extract from it, plus the file DictionaryForMIDs.properties and also the complete generated JAR file (or post a link to where I can donwload it) ? And please point to a section where that glottal stops occurs ?