Re: Mercury in a piston and cylinder.

Since mercury expands when heated, how much work could be done by allowing mercury in a cylinder to push a piston as the ambient temperature increases?

Could significant, useful work be done by allowing such a device, or a series of devices, to be cooled and heated by the fluctuations of daily temperature?

Do the laws regarding the conservation of energy say that the work output cannot exceed the heat energy going into the system?

To determine whether such a machine is worth the effort of designing it (not to mention the danger associated with mercury, which is a cumulative nerve toxin and best not to play around with), we need the thermal coefficient of (linear) expansion of mercury, which I don't have in any of my references.

However, the coefficients of thermal linear expansion a, defined by

Contrariwise, gases (such as steam) all have a about equal to 4´10^-3 K^-1, 10 to 100 times larger than those of condensed materials.

You could (possibly) design a system which derived energy from daily temperature fluctuations, but you would need a working fluid which boiled at around 25°C. The system design would probably be similar to that of an ocean thermal energy system.

Mercury doesn't fill the bill at all. And again,

DANGER     DANGER

Mercury is a cumulative nerve toxin. Do not play with it. Do not use it except under carefully controlled conditions, with the mercury completely contained. A cautionary tale may be found here.

Information in this answer was obtained from the Handbook of Chemistry and Physics, 61st Edition.