Fix missing mapping for deuteron in RichMCTruthTool

Also improve how it looks up the entry in the PDG code to type map, to avoid needing to make the map mutable. I improved the implementation of MCTruthTool::mcParticleType that previously, when an MCParticle was found with a PDG code not in the map the method would create a default entry in the map and return that. Now the code explicitly checks to see if the map has a entry in it, and if not returns Rich::Unknown. This does alter a little the ID types returned for these 'unknown' particles, but also prevents m_localIDgrowing (indefinitely) which each new PDG code it gets.

Will change a little some of the MC monitoring tables in the Brunel logs. Will submit MRs to fix these next.