Sunlight naturally allows this thermochemical cycle to occur, assuming that the device is allowed to cool at night. With other heat sources(geothermal, etc), you would need to remove the device to cut off the heat source for the "off" period of the duty cycle. Also, that makes this device less appealing than it might seem, as this "19%" efficiency cited doesn't mean you only get even 19% of the heat power of your source put in as power out, but thats only during the "on" period of the duty cycle. This may allow more flexible, if even a bit less efficient ways of converting heat to energy in "constant heat source" applications produce higher average power output from the same source.
That is probably why these researchers are pushing it for solar power applications, as its strengths(fair decent efficiency for the cost) stand but its [obvious] negatives(cycling) are a built in limitation of all solar power systems.