What medium to use to drive steam engines?

I want to create a Steampunk setting, but I'd like to introduce another medium than water to do the inner-engine work for me. Of course, water is a fine medium for steam engines, because:

• it expands quite a bit when heated,
• it doesn't start to boil at room temperature,
• it is safe to handle,
• it doesn't pose harm to the environment when expelled,

Are there any other fluids that would make a good medium for use in steam engines? What would be the up- and downsides of using something else? How would this substance behave at room temperature, and what would we need to do to turn it into steam? Would the expansion into gaseous form be as effective as that of good ol' H2O?

• Water is also a liquid at room temperature, which some potential alternatives may not be. – Starfish Prime Nov 25 '19 at 19:29
• The abundance of water is an advantage that really can't be over stated. – Trevor Nov 25 '19 at 19:29
• Geothermal binary-cycle powerplants often use a hydrocarbon such as butane or pentane. – Arkenstein XII Nov 25 '19 at 20:08
• You forgot the most important quality of water (for this purpose): it has a very high heat capacity, much higher than any other practicable working fluid. Steam is capable of carrying a lot of energy around; alcohol vapor, very much less. (And I don't understand what you mean by "expands quite a bit when heated". Actually, it doesn't expand all that much, not as long as it remains a liquid. Steam is a gas, it will expand to fill any available volume, as all gases do.) – AlexP Nov 25 '19 at 22:40
• Well, water is safe to handle, but steam definitely isn't. The cloud above boiling water in a pan isn't steam; steam is actually transparent, which is another reason why it's quite dangerous. – Luaan Nov 26 '19 at 8:00

If you don't mind having to run a closed system engine (they're most efficient, anyway, a true Rankine cycle), pretty much any fluid that has been used in refrigeration can also be used in a vapor engine. There was actually a serious proposal back in the 1970s to build a steam powered car that used R-12 Freon as its working fluid (this was several years before that particular chlorofluorocarbon was banned due to its ozone layer damaging chemistry, but the same system could run on a mixture of propane and butane).

As another answer has noted, water has the powerful advantages of being almost universally available, non-toxic, and carrying an immense amount of energy (compared to virtually any other substance) in its phase changes. You simply cannot beat that combination of qualities with any other material -- and most of the other working fluids that have been tried in "steam" engines have to be kept under pressure to maintain them as liquid on the "cold" side of the engine.

An alternative that has much better potential than some alternative liquid in a Rankine cycle is to use a different heat engine cycle. The Stirling cycle can use nearly any gas, though Stirling engines are most commonly made to use air as their working fluid. They tend to have less power output than comparably sized steam engines, until you consider the size and mass of the boiler. Stirling engines can also be built to operate on very low temperature difference -- there are videos of model size Stirling engines running well from the heat of a cup of tea, or from the cold of a single ice cube sitting on one exchange plate, even from the heat of sitting on someone's palm in a comfortable room.

Stirling engines can be used in any climate (no water to freeze, no issues with being in a hypothetical desert too hot for the condenser to operate) as long as you don't exceed the heat tolerances of the construction materials, and can produce a suitable temperature differential (in Stirling engines made to actually power something, this is usually done by putting the hot end directly into a flame, though concentrated sunlight also makes a good heat input).

To make a more efficient steam engine (or more generically, a Rankine Cycle heat engine), there are two main things to do:

1. Maximise the temperature differential between the 'hot' and 'cold' ends of the engine;
2. Minimise energy lost to the enthalpy of vaporisation of the fluid. This is best achieved by using a supercritical fluid.

The maximum temperature achievable for 1 is limited by the maximum pressure your boiler can withstand, so ideally you want a fluid which becomes supercritical at a low enough pressure that you can operate the engine entirely within that range. That wikipedia article lists a number of organic compounds, all of which have significantly better properties than water in this regard. Any of the chemicals with critical pressures in the 4-6MPa range would be achievable with late-Victorian steam technology. I'd suggest carbon dioxide, ethane or nitrous oxide would be good candidates, as their critical temperature is well below the normal 'cold' operating range, so there's no risk of the fluid coming out of supercritical state at the low end.

The main problem Victorian engineers had with building very high-pressure engines was balancing the need to get the water very close to the firebox in order to heat it efficiently, against the risk of catastrophic damage caused by any breach of the steam pipes into said firebox. One option for minimising these risks is to use an intermediate fluid to transfer heat from one to the other.

Liquid sodium is a common coolant fluid for nuclear reactors. It has a number of physical properties which make it particularly suitable for nuclear reactors, but it also has the benefit of a very large liquid temperature range (785 degrees between melting at just below the boiling point of water, up to its own boiling point at 1156 K). This makes it an extremely good coolant and transporter of heat, but an absolutely rubbish medium for the engine itself. However, when used as a heat exchange fluid it decouples the high pressure steam from the fire: the liquid metal system is not pressurised and completely sealed, so is much more reliable to cycle through the firebox (you can also pump metals with electromagnetic pumps which have no moving parts, but that's not very steampunk); then the heat exchange to the engine fluid can take place in a separate converter that's isolated from any thermal or mechanical vibrations from the firebox. Any liquid metal can be used, although sodium is convenient and easily available, and although it's nastily reactive with water, is much less reactive with CO2 or nitrous oxide.

• Let's not forget that nitrous oxide can thermally decompose into nitrogen and oxygen, with significant energy release. Combine that with liquid sodium and you're back to disaster. For that matter, sodium can pull oxygen out of carbon dioxide when in direct contact at high temperature (magnesium does this, all the alkali metals can as well, they're more reactive). And don't forget that liquid sodium, liquid potassium,and their eutectic, known as NaK, are highly corrosive to metals. – Zeiss Ikon Nov 25 '19 at 20:14
• The thermal decomposition of nitrous oxide will be impeded at high pressures; I can't access any of the relevant papers to be able to say how strongly it will be suppressed at ~50atm, but I expect quite substantially. – Stephen Nov 25 '19 at 20:24

It is difficult to meet all of your requirements, but if you want something a bit different why not try liquid nitrogen? It can be condensed into a liquid with equipment from the Victorian era (from 1883).

Its readily available everywhere, it expands a lot when heated and it doesn’t pose harm to the environment. Care is required when handling it but it’s not that dangerous, the odd splash won’t cause a problem just avoid drinking it or washing in it etc. The engine would have the reverse problem to a conventional steam engine in that it would require heating from the atmosphere rather than cooling. I believe one has been built in fact.

The only thing it falls down on is boiling at room temperature which it does, although it can be stored in a vacuum flask for a long time.

• it would require heating from the atmosphere rather than cooling actually, quite the reverse. It would need even more cooling. Remember, the cycle needs going from hot to cold and back to hot. With nitrogen, "hot" is the easy part; where do you get the cold? You need serious freezers that will need lots of power to keep your engine running. This just doesn't compare to the efficiency and relative simplicity of burning a bit of wood or coal in a boiler. – Luaan Nov 26 '19 at 8:04
• No the nitrogen would absorb heat from the atmosphere and would boil to form a highly compressed gas. This compressed gas would then be allowed to expand in the cylinder of the steam engine and then be exhausted into the atmosphere. The same way that steam is exhausted from a steam engine. The nitrogen would not be recycled on board. – Slarty Nov 26 '19 at 9:01
• Of course, but you need a cycle. Where do you get the liquid nitrogen? That's the hard part. – Luaan Nov 26 '19 at 9:23
• The question was regarding a steam engine (open cycle). The nitrogen would be liquefied by a ground installation and would be used as a working fluid and fuel combined. In the same way a steam engine has to be refuelled and watered from an external source. – Slarty Nov 26 '19 at 16:04
• You don't liquify nitrogen by magic. You need to cool it down and compress it. That takes lots of energy. You must account for that energy cost in the efficiency calculations of your engine! A steam engine is refuelled by water at ambient temperature (or even preheated by having it part of the insulation of the boiler). It doesn't need to be cooled two hundred degrees below ambient temperature :) It doesn't matter whether the steam engine is open or closed cycle (both were used in actual steam engines). – Luaan Nov 26 '19 at 19:57