The HTML entity "™" (referring to Unicode U+2122, the superscript TM 
trademark symbol) isn't supported by Doxygen.  It looks like Doxygen intended 
to support this symbol, but did so with the entity string "&tm;", which isn't 
a recognized entity name in HTML at all.  Using "&tm;" in doc comments just 
results in the literal characters "&tm;" appearing in the rendered HTML (in 
Internet Explorer 6, anyway).

I was able to fix this problem in my private copy of Doxygen with two one-line 
changes:

In docparser.cpp, insert the following before line 1274:

  else if (symName=="™") return DocSymbol::Tm;

[this is in DocSymbol::decodeSymbol()]

In htmldocvisitor.cpp, change line 102 as follows:

    case DocSymbol::Tm:      m_t << "&trade;"; break;

[this is in HtmlDocVisitor::visit(DocSymbol*)]

I noticed a related bug in the database: bug #162041, opened by George Rhoten, 
asking for full support of all HTML entities and numeric character 
references.  That would be a more complete solution than what I'm suggesting 
above, and I'd like to second that request.  But the change I propose here 
will at least solve my biggest problem while we're all waiting for a more 
complete solution.