Things dropping below ambient air temperature is extremely common. The simplest examples would include nighttime ground surfaces when the sky is clear. There's a reason why the edges of fields and the areas near walls and trees is the last to get frosted up
In essence, the temperature of an object depends on the heat transfer between itself and its surroundings. This transfer can be of conduction, convection, or radiation. For objects that are not fluids, that is limited to conduction and radiation. Now, air is a pretty good insulator and a poor fluid for the transfer of heat. Noting that the effective temperature of a clear sky at night is in the negative 40 degrees or lower, it then becomes fairly easy to see that the object in question will radiate a lot heat in that direction without a balancing flow of heat back. If that object were in an enclosed room, then the room surfaces would be transferring heat to the object at the same rate the object would be transferring heat to the surfaces - no change in temperature. There is heat transfer between the air and the object by conduction, but that is a very inefficient process by comparison especially when the air is still - and heavy frosts happen more easily with still air.
Telescopes can regularly reach a few degrees below ambient and then become cold enough for dew to condense on the optics and the tube structures, because the optics can be exposed to the cold sky and the optics are not being warmed effectively by the surrounding air.
Whether one can power a steaming engine with that, well if you have a working fluid in the right temperature range of course you can. The available power though would be very small given the delta-T present. It may be possible to measure a voltage with a very large Peltier setup sinking heat by radiative cooling only, but not enough current to be practical and especially for the cost of setup..
Theoretically this would be a version of Solar, where daytime heating and night time cooling is the driving force.