Refrigerating with air

Gallery opened 20 Apr 2018

Updated: 10 July 2018

New Haslam Machine
Refrigeration the safe way Back to Home PageBack to The Museum

Modern refrigerants
Historical refrigerants
The early history of refrigeration
Cold-air refrigeration
John Gorrie's cold air machine
William Siemens contribution
The Kirk machine
The Giffard machine
The Windhausen machine
The Sturgeon machine NEW
Hick Hargreaves machines
The J&E Hall machines Updated
The Haslam machines Updated
Bell and Coleman machines Updated
The Allen Dense Air Machine: 1881 NEW

Refrigeration may not be a glamourous area of technology but it is of vital economic importance. On the large scale it means frozen food can be transported around the world; on the small scale it means we don't have to go shopping for perishable food every other day.

The well-known refrigerating agents today are ammonia, for commercial operations, hydro-flurocarbon compounds (HFCs), hydrocarbons such as butane and pentane, and air and water. In case you are wondering about the last two, air is used as a refrigerant in air-conditioning machines for aircraft, fed by bleed air from the jet engines. Water is used as a refrigerant in static air-conditioning systems.

Ammonia is known to be dangerous. It is poisonous and explosive in air in some proportions. You might think that by this time the dangers of ammonia had been completely conquered, but you would be wrong. This accident in August 2013 in China killed 15 people and injured 26.

HFCs are used because they do not contain chlorine and so do not affect the ozone layer. A release into a confined space can cause asphxiation, as is true of any gases except air and oxygen.

Any sort of inflammable refrigerant presents a risk of explosion. There have been recent reports of domestic butane refrigerators exploding due to leaks.

All the dumped refigerators that I've seen recently have had 'Pentane' written on the back. Pentane has a flash point of -49 degC and ignites easily. Pentane is apparently being replaced by propane in this application.

Are many in number, including CFCs, sulphur dioxide, carbon dioxide, ethyl chloride, methyl chloride, methyl formate. Ethane (C2H6), propane (C3H8), butane (C4H10), iso-butane ((CH3)3-CH), methyl-ether (by Tellier) and diethyl-ether () have all been used as refrigerants in the past.

Into this category comes methyl chloride, and some of the hydrocarbons.

CFCs (Freons) are non-toxic but some of them have disastrous effects on the ozone layer when released and they are now banned.

Sulphur dioxide causes severe respiratory irritation and is toxic.

Old refrigeration textbooks are prone to say things like "carbon dioxide is completely safe" which is clearly untrue because of the suffocation hazard if a lot of it is released in a confined space

Any sort of inflammable refrigerant presents a risk of explosion. Into this category comes methyl chloride, and the hydrocarbons. Ethane, propane (C3H8), butane (C4H10), and iso-butane () have all been used as refrigerants in the past. There is a Wikipedia page on Refrigerants

Several early vapour compression machines used diethyl ether as a refrigerant; it is suitably volatile and as chemicals go was reasonably cheap, being made simply by the action of sulphuric acid on ethanol. Ether is of courser highly inflammable, and its heavy vapour tends to accumulate alarmingly in low places. But perhaps its worst characteristic is that of spontaneous explosion when ether peroxides are formed. The Museum page on ether as a working fluid in engines warns that:

Diethyl ether oxidises and polymerises in the presence of air, creating the interesting compound diethyl ether peroxide (-CH(CH3)OO-)n. This is a colorless oily liquid that is an extremely brisant (fast-exploding) and friction-sensitive explosive; less than 5 milligrams can damage chemical apparatus. The dangerous properties of ether peroxides are the reason that the use of diethyl ether and other peroxide-forming ethers like tetrahydrofuran (THF) or ethylene glycol dimethyl ether (1,2-dimethoxyethane) is carefully avoided in industrial chemical processes.

I am certainly not an expert on ether chemistry, but continuously reboiling ether sounds like it might be a good way to generate the peroxide. Peroxides generally have higher boiling points than the compounds they come from, so if an ether mixture is heated, the peroxide can become progressively more concentrated in the boiler and the risk of explosion increases rapidly.

Looking at these dire warnings it appears puzzling that anyone was able to operate an ether vapour-compression machine without blowing themselves to Kingdom Come. I suspect the answer is that in a refigerator you can arrange for every part to be under pressure, so air cannot seep in even if there are small leaks, and peroxides cannot form.

This is not remotely a comprehensive list. You can find a daunting array of refrigerants in Wikipedia; most of them are only relevant to vapour-compression cycles.

Table of Wikipedia links to the more common refigerants:
Methyl formate
Methyl chloride
Dimethyl ether
Diethyl ether
Methylene chloride (called carrene in the trade).

Bearing in mind all these various hazards, air and water as refrigerants have the attraction of being completely safe. Unfortunately they are also inefficient at refrigerating, and in the case of air its low heat capacity means that large volumes have to be handled. Air must be the ultimate safe refrigerant, (A water-based refrigerator could conceivably drown you) and that is one reason why it is worth looking at. Air has the further advantage that it wil not react with lubricants.

However, nothing is wholly risk free. A cold-air machine involves significant pressures, and something might be overstressed and fracture violently. Compressing air carrying a lot of lubricating oil could lead to a Diesel-type explosion.

The best-known refrigerating method is the vapour-compression refrigerator. This relies on a refrigerant like ammonia which liquefies easily when compressed and as a result is greatly cooled when passed through a throttling valve.

The prolific Ametrican inventor Oliver Evans described a vapour-compression refrigerator in 1805 but appears to have made no attempt to build one.

The first vapour-compression refrigerator was patented in 1835 by Jacob Perkins, but not commercially exploited. It used diethyl-ether or "some other volatile liquid" in a closed cycle.

The first commercial use of the vapour-compression refrigerator was achieved by James Harrison, whose first mechanical ice-making machine began operation in 1851 on the banks of the Barwon River at Geelong, in Australia.[3] His first commercial ice-making machine followed in 1854, and his patent for an ether vapor-compression refrigeration system was granted in 1855.


It is not possible to make a vapour-compression refrigerator using air because air cannot be liquefied at any reasonable temperature and pressure. (The conditions are quite different in machines specifically dsigned to liquify air) Therefore little or no reduction in temperature occurs on passing compressed air through a throttling valve. Instead work must be removed from the compressed air by making it drive a reciprocating engine or turbine. The use of an expanding engine is not merely a cunning way to reduce the power required to drive the process; it is essential for it to work at all. The work from the engine is used to help drive the compressor, improving the efficiency, but an air refrigerator can never be as efficent as vapour-compression refrigerator; it may not be 1/10 as efficient. Refrigeration engineers prefer to talk in terms of Coefficent Of Performance (COP) rather than efficency because it often comes out as greater than unity, and 'efficencies' of better than 100% have a long and dishonourable association with perpetual-motion nutcases.

The passages quoted below are from an article in Engineering for 4th Feb 1881, which was a report on a long and comprehensive paper given by A Mr Lightfoot, an employee of Messrs J & E Hall and Co, of Dartford- who were at the time much engaged in the manufacture of cold air machinery to be used for the importation of frozen meat from Australia, a very important trade. J & E Hall of Dartford was founded in 1785.

Cold-air machines had considerable advantages for the frozen meat trade. Suppose you used ammonia instead; a leak could be catastrophic in confined spaces, and if the refrigerant was lost there was no way to replace it and the entire cargo would rot. The latter problem was presumably eased when it became possible to store gases at high pressure in steel cylinders.


Dr John Gorrie of Florida in 1851 invented a cold-air machine for ice-making. The hot climate of Florida meant that ice was in much demand, and its distance from ice-exporting countries like Canada made ice expensive.

Left: Gorrie's cold-air machine: 1851

Air was compressed in a cylinder jacketed with water to keep it cool; water injection was also used to reduce the work of compression.. The air then passed through a worm P immersed in a tub of water for further cooling. The air was then expanded in another cylinder C and the resulting cold air used to make ice, after which it exhausted to atmosphere. This open cycle was very inefficient.

Gorrie attempted to raise money to manufacture his machine, but this failed in part owing to the death of his partner. He died in relative obscurity (as inventors so often do) at the age of 52.

Image from US patent 8,080, 6 May 1851


Sir William Siemens was one of the great pioneers of refrigeration, though he did not design a machine himself. In July 1857 Siemens completed an analysis of Gorrie's cold air machine. He pointed out that Gorrie's machine would more efficient if the exhaust cold air was used to cool the air going into the compression cylinder, and took out provisional patent No 2064. (1857) Siemens is regarded as the inventor of the counter-flow heat exchanger.

Siemen's second criticism of the Gorrie machine was the presence of a throttling valve before the expansion cylinder; this would have given little or no drop in temperature, but by reducing the air pressure reduced the amount of work that could be removed from the air by the expansion cylinder.

Siemen's third criticism related to the problems that would be caused by moisture in the cooled air freezing.


"Kirk's machine consists in principle of a single cylinder in which air is compressed at one end and expanded at the other. The beat caused by compression is partially carried off through the cylinder cover, which is water-jacketted, and the cold from expansion is used to abstract heat from a current of brine or other medium, circulating over the cover at the expansion end. Between the two ends is a regenerator, formed of several thicknesses of wire gauze. Through this both the hot compressed air and the cold expanded air pass, on their way from one end of the cylinder to the other; so that there is a continual alternate compression and expansion of the air, and a continual beating and cooling of the regenerator."

Left: Advert for Kirk's cold-air machine in 1872

Note the emphasis on safety.

Dr Alexander C Kirk was the engineering manager of a paraffin-oil works at Bathgate, in West Lothian, Scotland; this appears to have been E W Binney & Co, whose dominant partner was James Young Paraffin and other hydrocarbons were extracted from sandstone or distilled from cannel coal (oil shale) in a retort. The owners wanted a method of cooling for paraffin production, presumably to cool the distilled vapour so that the lighter hydrocarbons were not lost. A vapour-compression machine was tried but proved to be a fire hazard.* Given the date, the refrigerant was probably diethyl ether, which would indeed present a serious fire hazard in the event of any leak; not perhaps the best machine to install in a paraffin factory. Kirk therefore built a cold-air machine, based on Gorrie's work. Several similiar machines were built for ice-making, by John Norman and Co.

* From Refrigeration In America, by Oscar Edward-Anderson Jr

Left: Dr Alexander C Kirk, 1830-1892

The Stirling hot air engine was developed between 1825 and 1840. Kirk's machine, introduced in 1861, was the Stirling cycle reversed.

Kirk has a Wikipedia page, on which it is suggested that the original refrigerant was indeed diethyl ether.


"The Giffard cold air machine consists of one single-acting water-jacketted compression cylinder, and one single-acting expansion cylinder, both worked from cranks on an overhead shaft. The compressed air is led from the cylinder into the air cooler, which is merely a cluster of small tubes placed vertically in a case. The cooling water passes upwards outside the tubes, and thence goes to the compression cylinder jacket; the air is admitted into a casing below the ends of the tubes, passes up through them, and is taken off from the top to a wrought iron reservoir. A pipe from this reservoir supplies the air to the expansion cylinder; the admission and exhaust being controlled by two independent steel mitre valves in the cylinder bottom, worked by cams from the shaft. In this machine no attempt is made at drying the air; all the moisture taken into the compression cylinder is discharged in the form of snow from the expansion cylinder, with the exception of the portion deposited in the air cooler owing to the partial cooling of the compressed air."

Left: The Giffard cold-air machine: 1873

This cold–air machine was patented in France by Paul Giffard in 1873; do not confuse him with his older brother Henri Giffard, the inventor of the injector and the first to build a powered and steerable airship. Paul Giffard is best known for his work on gas-powered guns.

Here the compression and expansion cylinders are mounted in the base of the nachine, each side of the flywheel. The vertical tank with a pressure gauge on top is presumably the air cooler as it has a small diameter pipe suitable for cooling water running into it. The function of the tank at the right is unknown; it appears to have a safety valve mounted on it.

Left: The Giffard cold-air machine: 1877

This does not look very different from the 1873 model.

A must be the compression cylinder because it has spring-loaded self-acting valves. B is the expansion cylinder and requires valves driven from the crankshaft to operate. C is the heat exchanger which cools the air before it goes to the expansion cylinder.


"Windhausen's machine expands air from its ordinary atmospheric pressure under a piston; the cooled and expanded air being discharged much below the atmospheric pressure, either through tubes surrounded externally by brine, or into a hermetically sealed chamber, where the objects to be frozen are placed. After this process the air is again compressed to atmospheric pressure, cooled, and re-expanded. The disadvantages of this machine are the large size of the cylinders, &c, necessitated by the very low pressure employed, and the fact of its entirely depending for its action on the production of a partial vacuum."

Left: The Windhausen cold-air machine: 1883

By Franz Windhausen of Brunswick. He took out US patent 236,471 for an ice-making machine in 1881, that worked by evaporating water in a vacuum.

"Sturgeon's refrigerator is a horizontal machine with some novel arrangements as regards the construction of its air-valves and pistons. The compressed air is first cooled partially by being passed through tubes sunounded by cooling water, an then passed through charcoal or some other absorbent of moisture, before being admitted to the expansion cylinder. If the charcoal or other material is properly changed and renewed when necessary, this may form a dry air process; but, as already stated, the introduction of a chemical drier is in the author's opinion undesirable, except under special conditions."

Sturgeon and his refrigerator may have been known to Mr Lightfoot but they are unknown to Google. No illustration of the machine has so far been found.


Left: Advert for cold-air machine: 18

"Messrs. Hick, Hargraves, and Co, of Bolton, manufacture cold air machines of horizontal form, in which the Corliss cut-off gear is applied to the admission valves of the expansion cylinder. The air is compressed in a double-acting cylinder, into which cooling water is injected at each stroke; it then passes through a series of receivers, in which the water mechanically carried over is deposited, and is finally admitted to the expansion cylinder, and expanded to atmospheric pressure. So far as the author knows, no attempt is made at drying the air, which passes to the expansion cylinder fully saturated for its temperature and pressure; but a large snow-box, consisting of a series of baffles, abstracts the bulk of the snow from the cooled air, after expansion and before its introduction to the chamber. In a machine of this description which the author has seen, the snow had to be cleared out from the exhaust valves every few hours."

So far attempts to find out more about the refrigeration activities of Hick, Hargraves & Co have met with very little success. This advertisement confirms that they were in the refrigeration business, but that's about it so far.


Messrs J&E Hall and Co were the employers of Mr Lightfoot, who wrote the comments in quote marks on this page. He later branched out on his own.

Left: A J&E Hall cold-air machine: 18??

Nothing specific is known about this daunting lump of machinery, but some things may be deduced.

The cylinder on the far left is the compression cylinder, because it is of larger diameter than the other (expansion)cylinder. The pipe running into it labelled 'suction' confirms this.

The cylinder on the far right is the expansion cylinder; not only is the pipe coming from it labelled 'delivery' but its mechanically-operated valve-gear can be seen just in front of it, operated by a cam on the crankshaft.

The chest between the two cylinders is the air cooler; a small pipe for cooling water can be seen going into the top of it. The bolted access plates are to allow cleaning of the pipes in the cooler.

The two small horizontal cylinders above the cooling chest are the steam engine that drives the machine. Note the two cranks are set at 90 degrees to avoid dead-centre problems.

Left: A J&E Hall cold-air machine: 18??

C is the air compression cylinder. E is the air expansion cylinder, with the steam power cylinder on the same piston-rod..

Left: A J&E Hall cold-air machine No 6: 1886

This photograph clearly shows a different and later design compared with the two just above.

The cylinder that forms the base of the machine is the air cooler.

The 9 feet 4 inches written on the photograph appears to be the overall length of the machine.


Left: Haslam refrigerating machine: 1882

This Haslam cold air machine was designed to produce 2000 cubic feet (57,000 litres) of cold air per hour. On the left is a vertical single-cylinder double-acting steam engine. An eccentric on the crankshaft actuates the steam slide valve. The air cylinders are at the right, with the compression cylinder mounted above the expansion cylinder; their pistons are attached to a common piston rod. The compression cylinder has connections for its water jacket.

There must also be somewhere a heat exchanger, and water pump. I suspect the heat exchanger is the grey rectangular box behind the connecting rods.

Air at atmospheric pressure is drawn into the compression cylinder through an intake valve. As it is compressed, its temperature rises. The hot compressed air then passes to the water-cooled heat exchanger. The cooled air is admitted to the expansion cylinder through mechanically-operated expansion and cut-off valves, operated by the two eccentrics at the right end of the crankshaft. As the piston rises, the cut-off valve stops the admission of compressed air to the cylinder, which reduces the pressure, and thus the temperature, of the air as it does work. Later, near top dead centre, the expansion valve opens and, on the downward stroke of the piston, the cold air is expelled.

The temperature of the delivered air is controlled by the timing of the opening of the cut-off valve. The cycle theoretically follows the reversed Brayton cycle, comprising isentropic compression, constant pressure cooling, isentropic expansion and, finally, constant pressure discharge to either refrigerant coils or discharge into a cold room. An isentropic process is a reversible adiabatic process, in other words a reversible process having no heat exchange with the working fluid.

All three cylinders have 250 mm stroke. Cylinder bores are: steam 120 mm, compression 235 mm, expansion 155 mm. The expansion cylinder has wood lagging (the orange material in the photo) over a straw-like insulating material. There is a flywheel at each end of the crankshaft. This example was made 1882.

This machine is at the Museum of Applied Arts & Sciences in Sydney, Australia, where it is Exhibit H10484.

Left: Haslam advert: 1891

Left: Haslam cold-air machine: 1880

It is not immediately obvious why this machine was built on an incline.

The air cooler is presumably behind the flanged access plate, but beyond that it's pretty much impossible to work out what's what in this pile of machinery. Note the rocking beam driven from the crankshaft, which appears to be connected to a water pump and a Mysterious Cylinder.

Left: Haslam cold-air machine: 1880

This Haslam machine has a more conventional horizontal layout. The compression cylinder is presumably the large one to the left, as it is fitted with spring-loaded inlet valves. The two cylinders in the foreground are presumably the expansion and steam cylinders, as both appear to have valvegear driven from the crankshaft.

The air cooler is the rectangular box that forms part of the base of the machine.


Bell and Coleman were one of the leading suppliers of air refrigrators.

"The Bell-Coleman refrigerator consists of an ordinary machine for producing cold air by compression, cooling, and expansion, combined with an apparatus for depositing a portion of the moisture before the air is admitted to the expansion cylinder.

In this system the air is partially cooled during compression by the actual injection of cooling water into the compressor, and by causing the current of compressed air flowing from the pumps to come in contact with a spray of water. From the pumps the mixed air and water is led by pipes into a chamber or chambers with perforated diaphragms, which catch a portion of the suspended moisture. The air, still in its compressed state, and cooled to within 5 or 10 degrees of the initial temperature of the cooling water, is then led to the expansion cylinder through a range of pipes, or other apparatus, with extended metallic surfaces, cooled externally to a lower temperature than that of the cooling water; so as to induce a further reduction in temperature and consequent deposition of moisture.

This extra cooling of the compressed air is effected either by allowing the cold expanded air, before it reaches the chamber to be cooled, to come in contact with the range of pipes, or by exposing these pipes to the spent air passing from the cold chamber. The author then considered at considerable length the objections to which the machine was open. It should howwever, be stated that he very frankly added that these machines have been successfully worked in cases where a large amount of cooling water of low temperature is available, as, for instance, on board an ordinary Atlantic steamer. There is no doubt that moderately dry air would be obtained wherever a sufficient supply of water at 46 deg. or 50 degF can be had."

Left: A Bell-Coleman air refrigeration machine: 1881

Air at atmospheric pressure is drawn into the compression cylinder through an automatic intake valve. As it is compressed, its temperature rises. The hot compressed air then passes to a water-cooled heat exchanger.

The cooled air is admitted to the expansion cylinder through mechanically-operated expansion and cut-off valves, operated by the two eccentrics at the right end of the crankshaft. As the piston rises, the cut-off valve stops the admission of compressed air to the cylinder, which reduces the pressure, and thus the temperature, of the air as it does work. Subsequently, near top dead centre, the expansion valve opens and, on the downward stroke of the piston, the cold air is expelled.

The unlagged cylinder on the left is the compression cylinder, because it does not have mechanically-operated valve driven from the crankshaft. The lagged cylinder in the middle does have such valves, and so is either the steam power cylinder or the expansion cylinder; whichever it is the other cylinder is behind it, and connected to the crank on the far side of the crankshaft. The expansion cylinder was always smaller than the compression cylinder, because it was handling colder air.

The original source of this magnificent image is Engineering for 28th Oct 1881, p314.

The first Bell-Coleman machine was used on the SS Strathleven, which brought the first successful cargo of frozen meat from Australia to London, arriving in February 1880. The machine illustrated above is described as "a new type of machine of small dimensions for the cooling of ships' provisions" rather than freezing the whole cargo.

Left: Bell-Coleman air refrigeration machines on ships: 1879

Referring to the top diagram:

  • A - Steam cylinder
  • B - Compressor cylinders (two)
  • C - Water tower (spray cooling of air)
  • D - Pipe to air cooler E E
  • E - Compressed air cooled by return air
  • G - Expansion cylinder
  • K - Return air to compressor
Important refinements were the use of water injection into the compression cylinder, to cool the air and reduce the work required for compression, and the cooling of the compressed air by the return air from the refrigerated space. At this point the water was condensed from the moist air and removed by automatic trap valves. The dry compressed air then passed to the expansion cylinder G.

This was a closed cycle with the return air going back to the compressor cylinders via pipe K.

The lower diagram shows a later installation on the P&O steamships Rome and Carthage.

From Engineering for 28th Oct 1881, p318. Note that the artist has omitted both connecting rods from the Strathleven machine.

Left: Bell-Coleman air refrigeration machines on ships: 1881 article

This extract from the 1881 Engineering article gives details of the original Bell-Coleman machine in the Strathleven. See diagram above.

Left: Bell-Coleman air refrigeration machine: 18??

The air cooler is built into the base of the machine.


The splendidly-named Allen Dense Air Machine was a relatively late entry into the air refigeration market. Little is known about its operating parameter; I can only assume that 'Dense Air' refers to working at higher air pressures than usual to make the machinery more compact.

Left: Allen Dense Air refrigerating machine: 1883

While this diagram looks complicated, it has the same basic components as the other machines; a steam engine to drive it, an air compression cylinder, and a smaller air expansion cylinder which helps with the driving. There are also small pumping cylinders for cooling water and make-up air.

Diagram from Audells Plumbers and Steam-Fitters Guide, Volume 3, p2342. Pub 1949, reprinted 1953. I bought my copy in Vancouver in Septembe 2007.

Left: Allen Dense Air ice machine: 1883

That's an impressive list of steam-yachts. Note the emphasis on avoiding chemicals by using air as the refrigerant.

Left: Allen Dense Air ice machine: 1883

This is clearly the same picture as at left in the advert above, but at a larger scale so some more detail is visible. The cylindrical thing on top is the air cooler. (heat exchanger)

Cold-air machines were largely displaced by ammonia machines during the 1880's.

Back to Home PageBack to The Museum EntranceTop of this page