Alcohol Engines

Updated: 9 Apr 2015

Cartwright engine added
More info on Alco-Vapor
Back to Home PageBack to The Museum

This page deals with engines that use alcohol vapour or a water/alcohol mixture as a working fluid, rather than simply as a fuel to burn.

Alcohol vapourises easily, but that does not mean it is more efficient than water as a working fluid. Quite the reverse. See: Carnot's Law on the thermodynamics page.


Left: A toy steam-engine driven by alcohol vapour, with the exhaust combusting.

This scheme comes from the book "Engines" by E N da C Andrade, published G Bell & Sons in 1928. (A very engaging book that manages to deal with the complexities of thermodynamics in a most approachable way)
The author points that the alcohol vapour exhaust can be lit, though it would be wise to make sure that all the air has been driven out of the plumbing before trying it. This underlines that a non-condensing (total-loss) alcohol vapour engine is a grotesquely inefficient and expensive proposition. I hasten to add that the author was not claiming otherwise, but goes on to deal with the concept of Carnot efficiency.


The French journal Nature, (21 June, 1890) has this to say: "Artwright proposed and constructed, in 1797, an alcohol-vapour machine which worked well enough, but which did not survive, very probably because of the very imperfect construction of the time; substantial leaks of a relatively costly vapour would have made the system very uneconomic."

I have long suspected that "Artwright" should be "Arkwright", but actually it looks as though the name was Cartwright; the entry in Knight's dictionary above refers to the Reverend Edmund Cartwright, but it only says he suggested an alcohol engine, not that he actually constructed one.

Left: The entry on alcohol engines from Knight's Dictionary

With a description of Howard's alcohol engine, patented in 1825.

From Knight's Practical Dictionary of Mechanics pub Cassell & Co, 1884.


Edmund Cartwright spent most of his labours on textile machinery, for which he is far better known, but in 1797 was granted a patent for an alcohol engine. The patent included the possibility of using water/alcohol mixtures and a rotary version of the engine.

Left: The Rev Edmund Cartwright's alcohol engine: 1797

The explicit intention was to reduce fuel consumption as the alcohol would evaporate at a lower temperature than steam; Cartwight planned to halve the fuel used. This apparently sensible notion is actually a thermodynamic fallacy.

The Cartwright engine was single-acting, the power cylinder sitting on top of a box of cold water used for condensing the exhaust vapour. The operation of the valves is somewhat obscure, but it appears that alcohol vapour from the boiler was admitted to the top of the cylinder when the piston impacted a valve in the top of the cylinder. It was a bash valve. The vapour then gets to the other side of the piston when a valve in it hits the bottom the cylinder; this does not sound like a recipe for optimal valve timing. The round item W is a double-walled condenser with water on both inside and outside. The liquid alcohol was then returned to the boiler by the pump below the power cylinder. h was an 'air-box' which allowed the escape of non-condensing gases (eg air) to escape when the float fell.

The real technological breakthrough here was only indirectly connected with the use of alcohol. It was the use for the first time of metallic piston rings to seal the pistons. This was driven by Cartwright's need to minimise the leakage of an expensive working fluid.

Left: The Rev Edmund Cartwright's alcohol engine: 1797

Image from: The Steam Engine: Comprising an Account of Its Invention and Progressive Improvement by Thomas Tredgold.


Howard's engine is mentioned in Knight's dictionary (above) and it appears to have been installed at the Rotherhithe Iron Works, though I doubt very much if it did any useful work there. There were several ironworks in Rotherhithe and the relevant company has yet to be identified. Some of the first steamships were built in Rotherhithe; the first sea-going iron ship, the Aaron Manby, was constructed in the local shipyards.

Above: Howard's alcohol engine: 1825

Howard appears to have been fully aware of the need to minimise the losses through leakage of an expensive working medium. He used hot oil as a liquid piston. (Early experiments seem to have beeen done with mercury, but the cost of this for a large engine would have been crippling. Breathing the mercury vapour would probably have been crippling too) In the diagram above a flat dish D can be seen floating on the hot oil A in the middle cylinder. The oil was kept hot by the four Argand burners underneath. Alcohol was sprayed into the hot dish by the narrow pipe just above it, propelled by pump R, and flashed into vapour. This pushed down the oil and pushed up the piston in cylinder B on the right.
On the return stroke the alcohol vapour was pushed into the condenser on the left via pipe S. The condenser tubes U were wrapped in wet flannel, with air blown over them by a fan mounted on vertical shaft Z.

This engine, with its complexity and its interesting combination of alcohol and hot oil, seems wholly impractical. No wonder that according to Knight, it "...wearied out the patience or the means of its inventor...".


A binary vapour engine boils a mixture of liquids, usually water and some kind of alcohol. It doesn't appear to be a good idea, but that has not stopped a few people from trying it.


In 1885 a commission of US Navy engineers tried out a mixture of water and 15% wood alcohol, Methanol was presumably chosen for its low boiling point, though this is fallacious. (see below) Perhaps they thought its toxicity would deter the crew from drinking it. Historically this approach has been of questionable effectiveness.

Above: An experiment with a binary working fluid- in this case water and methanol.

The account reproduced here is taken from the British journal Engineering for 9th Jan, 1885. Wood alcohol is methanol (methyl alcohol)

Note that the board of US Navy engineers had grave (and entirely justified) doubts about the thermodynamics of low-boiling-point liquids as working fluids. (Methanol boils at 64.7 degC) The US Navy has not always distinguished itself in this field.

The major problem with the later large-scale test seems to have been leakage; not good when the vapour leaking is toxic, inflammable and relatively expensive. Whether methanol really does leak out more easily than steam I have not so far been able to determine; presumably it depends on the size of the vapour molecule.
Another pitfall was clearly the tendency of this binary mixture to change its composition when boiled, so that the condensed liquid in the hotwell was much richer in methanol than the liquid in the boiler. For an ethanol-water mixture, the vapour would not boil off in the same proportions as in the liquid in the boiler unless the methanol concentration was a lot higher than 15%; ethanol-water forms a constant-boiling mixture at a concentration of 95.6% ethanol at 78.1 degC. However it appears that methanol does not form a constant-boiling mixture with water; see The Methanol Manual, on page 5. I'll leave it to someone who knows more physical chemistry than I do to evaluate what effect this would have had on the tests.

To learn more about constant boiling mixtures see here. (external site)


Frank W Ofeldt is most closely associated with the naphtha launch, but when he left the Gas Engine and Power Company he produced the Alco-Vapor power plant for launches. This boiled a mixture of water and alcohol, not because the lower boiling-point of alcohol would make the engine more efficient (the exact opposite is true) but in not very impressive attempt to evade the USA regulations that required every steam launch to carry a licensed steam engineer. The naphtha launch was the first approach to this, and it seems to have worked in evading the law, at the cost of the inherent hazards in boiling petrol.

The evidence is comtradictory as whether 'alcohol' in this context means 'ethanol' or 'methanol', but given the problems that were experienced when methanol was used by the US Navy, as described above, its use seems likely. In the journal MotorBoating (Nov 1942, p46-48) the proportion used was described as "a merely nominal amount of alcohol" and if that is true and the question had arisen, it is hard to see how Ofeldt could have argued at all convincingly that he was not building steam engines.

Above: Basic layout of an alco-vapor launch: from Ofeldt's US patent 551,226

The convenient layout with the fuel at one end and the power-plant at the other was retained from the naptha launches, but the ingenious principle of burning the working fluid as fuel was abandoned, probably because of the low calorific value of alcohol when it burns. Heat was provided by burning what the patent calls "head-light oil", but in practice was kerosene. The middle drawing shows the folded condenser pipes running the length of the boat. The lower drawing shows that the vertical coil boiler has become a horizontal coil design but the integrated nature of boiler and engine is to some extent retained.

Frank W. Ofeldt's alco-vapor patent was US 551,226, granted 10th December 1895, and assigned to The Marine Vapor Engine Company. It states that Ofeldt prefers to use wood alcohol (methanol as it is the cheapest. The patent also suggests the use of choloroform or ether as volatile working fluids. The US Navy team had serious trouble with toxicity and eye irrtation when they used methanol, and this supports the suspicions of MotorBoating that there was a good deal more water than alcohol in an alco-vapor boiler.

Left: Two views of the Marine Vapor Engine Company's engine

The engine is a three-cylinder radial mounted directly under the boiler; in the front view the top two cylinders can be seen. In the side view a circular cylinder-head can be seen just below the boiler.

The handle seen in the front view is a combined throttle/cutoff/reversing control.

Left: The front cover of the Marine Vapor Engine Company's catalogue

This is taken from page 4 of the Marine Vapor Engine Company's 1897 catalogue:

"As we use alcohol expansively for power, it may be asked, wherein does it differ and in what respect is it better than the use of naphtha? In answer to the first query, we will state that the alcohol has the peculiar property of expanding a greater number of times and at a lower temperature than naphtha. John W Nystrom, Civil Engineer, a recognised authority, gives the boiling point of naphtha at 320 degF (160 degC), water at 220 degF (100 degC), and alcohol at 173 degF (78 degC). So it will appear quite obvious that should we begin to create pressure at 173 degF, by the time we have increased the temperature to the point of boiling water we have fifteen pounds, and before we reach the boiling point of naphtha we have a pressure of 100 pounds."

"In fact, owing to the low boiling point of the alcohol, it is not necessary in starting the Alco engine to labouriously work the liquid out of the cylinders by turning the engine by hand as in the case of the naphtha, for immediately we have 25 lbs pressure- our engine starts without aid, and owing to the small amount of heat required one can fearlessly place their hand upon the casing of the engine when running at 100 ilbs pressure. But our intelligent critic may say, admitting all the advantage you claim, is not alcohol dangerous? Our answer is no; for unlike naphtha, it combines with water in all proportions, and should any be allowed to escape, the fact that there is always a small quantity of water in the bilge eliminates any likelihood of fire. This could be demonstrated by the simple experiment of taking a spoonful of alcohol, igniting it, and then pouring it into a glass of water. The flame will be immediately extinguished, while it is a well-known fact that naphtha burns more intensely upon the surface of the water than elsewhere."

"As we carry but one-tenth as much alcohol as is usually carried of naphtha by naphtha or gasoline launches, and use the same over and over with but little attention during the entire season, the convenience and safety of our system will be readily seen and appreciated."

I inserted the Celsius temperatures. The boiling point of 173 degF (78 degC) given for 'alcohol' is for ethanol; methanol boils at 148 degF (64 degC)

Some comments on this passage; it clearly implies that the working fluid is just alcohol, rather than a mixture of alcohol and water. The miscibility of alcohol and water, and the much reduced fire risk is a good point. This would apply to a leak of the working fluid, but not to a leak of the kerosene fuel. The bit about carrying "one-tenth as much alcohol as is usually carried of naphtha by naphtha or gasoline launches" seems dubious as the power-plants look about the same size; perhaps this comparison included the naphtha used as fuel, but ignored the kerosene carried by the alco-vapor boats. If so it was dishonest.

Note how Ofeldt (or his agents) are bad-mouthing the naphtha launches that he had himself popularised.

I had my doubts about John W Nystrom, but there he is in Wikipedia. Unfortunately he seems to have been a bit of a nutter: Nystrom is best known for his proposal in 1859 to switch from base-10 to base-16 number systems. By 1875 he had changed his mind and was advocating base-12. This sort of thing does not inspire confidence. At first I assumed he was called on as a consultant as a result of his book A new treatise on steam engineering, physical properties of permanent gases, and of different kinds of vapor (Philadelphia: London: J.B. Lippincott & Co. 1876) noting the "different kinds of vapor" in the title.

But... Wikipedia has his dates as 18251885 so he did not make old bones. More to the point, he had been dead for 12 years when The Marine Vapor Engine Company were quoting him as an expert. Clearly my doubts were well-founded.

Olfeldt's personal history can be found on the page devoted to the The Petrol-Vapour Motors.


  • US 538,694 F W Ofeldt: Compound vapor engine with two HP-LP units arranged in a V-format.
  • US 540,757 F W Ofeldt: Vertical coil boiler and three-cylinder radial vapor engine
  • US 551,226 F W Ofeldt: Vapor launch patent

Back to Home PageBack to The Museum EntranceTop of this page