volatile gasses
OK, so it depends upon the coal your burning, and this explains some of the differences between the loco designs between the 4 grouping in the UK's steam hey day. But the amount of energy released from burning the off gassed volatiles normally exceeds the energy released from the burning of the bulk carbon of the coal. It is also easier to exploit since the combustion can occur high in the fire box or better still inside the fire tubes running through the boiler, which is actually where you need it.
Modern control systems might well allow you to get around some of the big bug bears of external combustions engines, the need to get the fuel into the fire 5+ minutes before it will be needed. In the past it was down to the skill of the driver and most especially the fireman, to know just when to feed the inferno so that it would be ready for the next section of the line. That should be possible to automate now.
But as others have pointed out, making a steam engine expire the Mallards run is going to be very difficult. Gressley wasn't exactly hot on the concept of elf'N'safety. Let's face it Stanier would let the LMS have a go at the record, and went ape when they did run Coronation upto 114MPH near Crewe, having forgotten the upcoming set of points!
Not be restricted by the UK loading gauge would allow much larger driven wheels, but you are still going to have massive reciprocating forces. The pistons and conrods are huge and so consequently weighty. The baring are going to be taking a massive loading.
As Chris says, a turbine such as Turbomotive is a much better approach.
This will still leave huge technical challenges.
Mines the one with the burn whole from the cinders and a copy of the Engineman's handbook in the pocket.