Could you exploit the temperature differentials between the either hot or cold surface, and the presumably in-between temperature at the bottom of a drill hole?
In theory the temperature underground should be the average of the surface temperature, so you could use that gradient to generate energy at night. Someone smarter than me would have to do the math on the energy density of a scheme like this, my gut says it's going to be fairly low and you will need a really big power plant to supply even a modest colony.