International Journal of Refrigeration 28 (2005) 1140–1148

www.elsevier.com/locate/ijrefrig
Review
Carbon dioxide—new uses for an old refrigerant
Andy Pearson*
Star Refrigeration Ltd, G46 8JW, Glasgow, UK
Received 5 September 2005; received in revised form 12 September 2005; accepted 13 September 2005
Available online 2 November 2005

Abstract

Carbon dioxide has been used as a refrigerant in vapour compression systems of many types for over 130 years, but it is only
in the last decade that inventive minds and modern techniques have found new ways to exploit the uniquely beneficial properties
of this remarkable substance. This paper traces the development of the old carbon dioxide systems, considers the technical,
commercial and social reasons for their slow development and subsequent decline and examines the recent renaissance across a
surprisingly broad range of applications, from trans-critical car air conditioners to low temperature industrial freezer plants. The
paper then concentrates on industrial refrigeration systems, which were the basis of early developments in the period 1865–
1885, but which have been somewhat overlooked in the current renaissance. The paper concludes with a review of possible
future developments, indicating the areas of research and product development required to maximise the potential of the
only non-toxic, non-flammable, non-ozone-depleting, non-global-warming refrigerant available for Rankine cycle vapour
compression systems in the 21st century.
q 2005 Elsevier Ltd and IIR. All rights reserved.

Keywords: Refrigeration; Air conditioning; History; Review; CO2; Technology; Recommendation; Research

Dioxyde de carbone: nouvelles utilisations d’un vieux frigorigène

Mots clés : Froid ; Conditionnement d’air ; Historique ; Enquête ; CO2 ; Technologie ; Recommandation ; Recherche

1. Introduction and evaporation temperature limit of 0 8C place severe

There are five substances generally recognised as
‘natural refrigerants’ in modern refrigeration. Air is used
in a variety of gas cycles, with no change of phase, and can
achieve reasonably low temperatures, but the low theo-
retical efficiency of the Brayton cycle and the difficulty of
getting close to that ideal have limited its use. Water vapour
has been used with large centrifugal and axial turbines in
open systems but the low pressures, large swept volumes
* Tel.: C44 141 638 7916; fax: C44 141 638 8111.
E-mail address: apearson@star-ref.co.uk.
0140-7007/$35.00 q 2005 Elsevier Ltd and IIR. All rights reserved.
doi:10.1016/j.ijrefrig.2005.09.005
restrictions on its use and make it fundamentally unsuited to
smaller air conditioning systems and industrial cooling and
freezing applications. Ammonia, carbon dioxide and
hydrocarbons have a broader range of application, and are
used in much more conventional systems. Despite a
generally excellent safety record there is a strict limit on
the allowable charge of hydrocarbon systems, which makes
them unsuitable for use in large water chillers and industrial
systems unless relevant safety standards can be applied. In
many ways ammonia is ideal for large industrial systems
where its mild flammability, pungent smell and low
threshold limit value do not present problems. It is, however,
clearly unsuited to domestic, automotive and small
 A. Pearson / International Journal of Refrigeration 28 (2005) 1140–1148
commercial refrigeration and heat pump systems. This
leaves carbon dioxide as the only natural refrigerant to find
favour across the broad spectrum of automotive, domestic,
commercial and industrial refrigeration and air-conditioning
systems (Pettersen [11]).
2. Historical perspective—early steps
The first steps towards modern carbon dioxide refriger-
ation systems were taken in the 18th century by two Scots
physicians, Dr William Cullen and Dr James Black. Cullen,
who practiced medicine in Glasgow, was also professor of
medicine at Glasgow University and, in 1748 established the
department of Chemistry in the university. He is credited
with the discovery of latent heat through his experiments
with water in 1755, and also observed that the boiling point
of water could be reduced by lowering the pressure below
atmospheric. This led to experiments with various other
volatile fluids, such as sulphuric ether, although in these
early systems the fluid was exhausted to atmosphere, not
recirculated. Black was one of Cullen’s medical pupils, and
succeeded him as professor of chemistry at Glasgow in
1755. Black’s experiments heating ‘magnesia alba’ (mag-
nesium carbonate) led him to the discovery of carbon
dioxide, which he called ‘fixed air’. Further experiments
proved that this unusual gas was involved in many familiar
processes, including burning and breathing. Black correctly
predicted that ‘fixed air’ would be present in small quantities
in the atmosphere, although it was many years until the level
of 0.03% was confirmed. Neither Cullen nor Black was
primarily interested in thermodynamics or refrigeration, and
their ideas were not developed for nearly a century.
(Thévenot [13]).
Oliver Evans of Delaware proposed a closed cycle for
refrigeration in 1805, although no such systems existed at
that time, and this innovation did not progress until Evans’s
friend Jacob Perkins was granted British Patent number
6662 in 1834 in London for his ethyl ether machine. Perkins,
who was 68 by then, did not exploit his patent, and vapour
compression did not progress until Alexander Twining, a
professor at Yale, patented another ethyl ether-based system
in the USA in 1850. Twining made several efforts to
commercialise his system, including an ice plant installed in
1856 in Cleveland producing 2000 lb (909 kg) in 20 h, but
he did not achieve long term success. At the same time,
James Harrison in Australia developed an ethyl ether based
vapour compression ice machine, probably in complete
ignorance of the work of Perkins and Twining. Harrison
brought his system to London in 1856 to patent and develop
it, and gave several successful demonstrations to notable
scientists of the day, including Michael Faraday and John
Tyndall (Gosney [1]). Although these early systems may
appear to be novelties, they were certainly not trivial.
Twining’s Cleveland plant is said to have had a double
acting compressor with a 210 mm diameter piston and
1141
450 mm stroke. Harrison’s 1857 machine, described in
British Patent number 2362 had a 380 mm bore and a
770 mm stroke.
3. Diverse developments of rival technologies
Faraday’s interest in artificial cooling went back to 1824
when he demonstrated a form of absorption cooling using
ammonia and silver nitrate in a sealed U-tube. He used this
arrangement to demonstrate the liquefaction of several
common gases. Absorption systems using aqua-ammonia
were further developed by Ferdinand Carre in the 1850s and
immediately found widespread success in block ice making.
In 1867 (the year after the civil war ended) there were three
artificial ice plants in San Antonio, out of five in Texas and
only eight in total in North America. Harrison attempted to
be first to ship beef from Australia to England on the sailing
ship Norfolk in 1873. Believing that mechanical equipment
would not be acceptable on board ship, Harrison’s system
used a stock of ice and salt to chill brine, which was trickled
over pre-frozen meat wrapped in heavy waterproof canvas
sacking. This early marine venture failed, apparently
because leaks from the brine circulation system contami-
nated the cargo during the voyage. Harrison might have
been more successful if he had trusted his equipment; in
1876 Charles Tellier of France equipped an old British ship,
the Eboe, with three methyl ether compressors for the first
transatlantic shipment of refrigerated meat. Renamed
‘Frigorifique’ by Tellier, she sailed from Rouen to Buenos
Aires in 105 days with a small cargo of cattle and sheep
carcases, and returned with 25 ton of chilled beef. The
following year Ferdinand Carre equipped the SS Paraguay
with a marine version of his absorption machine, for the
shipment of 150 ton of frozen beef from Marseilles to
Buenos Aires and returned to France with a further 80 ton,
all reportedly ‘edible’ when unloaded in Le Havre [13].
Carre’s method of brine chilling was initially popular for
ice-making installations on shore, and was the basis of
Thomas Mort and Eugene Nicolle’s first proposals for the
shipment of meat to England from Darling Harbour, Sydney
in the 1870s. However, although Nicolle had constructed
several successful absorption ice plants for warehouses in
the Sydney area, he fitted the SS Northam with an air cycle
system in 1876, after permission to use ammonia on board
was refused. Unfortunately, owing to problems during the
commissioning of the system, the Northam sailed for
England without a cargo, and although the plant worked
satisfactorily throughout the voyage, no meat was carried.
Further misfortune followed, when the Northam was lost at
sea during the return voyage in 1877.
Other practitioners also favoured the air cycle, which had
been proposed in 1820 by Richard Trevithick, an employee
of J&E Hall in their pre-refrigeration era (Miller [8]), but
which was not demonstrated in a commercial machine until
25 years later. Dr John Gorrie, a Florida physician,
1142 A. Pearson / International Journal of Refrigeration 28 (2005) 1140–1148
constructed his ice plant using an air compressor in 1844,
prompted by the lack of ice for treatment of malaria patients
in his infirmary. Alexander Kirk, a Scottish oil engineer,
developed a much more effective air-cycle machine in 1862.
Unlike Gorrie’s ice maker, Kirk used a closed cycle based
on Rev Robert Stirling’s heat engine, and it is reported that
the first machine ran continuously for ten years! Joseph
Coleman, one of Kirk’s colleagues, further developed this
Stirling cycle machine, and in the mid 1870s corresponded
with Lord Kelvin of Glasgow University. Kelvin introduced
Coleman to Henry and James Bell, butchers in Glasgow.
Together they developed the Bell–Coleman air cycle to suit
marine transport, and patented it in 1877. In 1879 they
equipped the SS Circassian for the trans-atlantic run and the
SS Strathleven for the first successful shipment of frozen
meat from Australia to England. These set the pattern for
marine refrigeration systems for the next decade, although
both ships were stripped of their refrigeration plant after
only one voyage.
Paul Giffard in France and Franz Windhausen [14] in
Germany produced refinements of the Gorrie air cycle
design in the 1870s and licenced their technology to
companies on both sides of the Atlantic. J&E Hall of
Dartford, and Alfred Haslam of Derby took licences for air
machines during the late 1870s and supplied the marine
market from 1881 onwards. With a licence from Giffard in
1878, J&E Hall’s young owner, 20 year old Everard Hesketh
started a 10 year air compressor development programme
which turned the company from a ‘languishing, out-of-date’
engineering workshop into a leader in industrial refriger-
ation technology. The Haslam Foundry and Engineering
Company also started their air compressor programme in
1878, and equipped their first meat carrier, SS Orient in
1881, followed by the sailing ship Mataura the following
year. Fig. 1 shows a time-line of some refrigeration
developments. It is evident that, in the period 1845–1885,
there were eight major developments in the choice of
refrigerant or system, but from 1885 to 1925 there were
none. The 40 years of continual effort must have been
prompted, fundamentally, by a deep-seated dissatisfaction
with the ‘state-of-the-art’.
Fig. 1. Timeline of refrigeration development.
4. System rationalisation—the dominance of vapour
compression
A common trait of all types of early refrigeration
equipment was that the concepts outstripped the manufac-
turing capability of the day, and therefore progress was
erratic, because each new development was dependent upon
parallel innovation in related fields. For example, various
compressors were proposed from 1820 onwards, but could
not be commercialised until machining capability had
advanced sufficiently and suitable prime movers were
available. Although the machines he developed in Australia
seemed to work well, James Harrison was severely critical
of the standards of workmanship at that time. Early vapour
compression systems used a variety of naturally occurring
compounds, including ether, ammonia, carbon dioxide and
sulphur dioxide. Each had its own advantages and draw-
backs, and consequently rose and fell in popularity as
technical development opened up new possibilities.
In vapour compression, ethyl ether systems were the first
to be proposed, as early as 1834, perhaps because ether was
readily manufactured, already in use as a solvent and easy to
work with as it is liquid at room temperature and
atmospheric pressure. As ether is highly flammable and
required to operate below atmospheric pressure to create ice,
these systems were never sufficiently safe or reliable to
achieve commercial success, although James Harrison
constructed the first marine system in 1855 in Australia
and persevered with ether until his death in 1893.
Carbon dioxide was the next to make a breakthrough,
through the work of Thaddeus Lowe in Texas. Lowe was a
self-taught scientist with a passion for aeronautics, and was
responsible for founding the Union Army’s Observation
Corps in 1861. Lowe’s compressor was developed in 1860
for filling military observation balloons with hydrogen, and
he served as an observer for the Unionists throughout the
American Civil War. His compressor was adapted for CO2
in 1866 and then used for the manufacture of artificial ice.
Lowe was some 20 years ahead of other developers of
carbon dioxide systems, and it has therefore been suggested
that his systems made ‘dry-ice’ in an open system. However,
there is no doubt that his British Patent, number 952, of 1867
(Newton [9]) discloses a closed vapour compression cycle,
with compressor, condenser and evaporator. In 1869 he was
narrowly defeated by Henry Howard in the race to be the
first to ship frozen beef, by sea, from Texas to New Orleans,
supposedly because Lowe’s custom built refrigerated cargo
ship, the William Tabor, was too large to dock in New
Orleans harbour (Woolrich [15]). Unlike ether, carbon
dioxide was non-flammable, and essentially non-toxic, but
for a closed vapour compression cycle it required extremely
high pressures—much higher than those used in the steam
boiler plant of the day. This perhaps was the cause of the 20
year delay, letting ammonia and sulphur dioxide systems
become established first. Throughout the 1860s ammonia
had been used in absorption systems, but in 1872, the first
 A. Pearson / International Journal of Refrigeration 28 (2005) 1140–1148
ammonia compressor was developed by David Boyle in
Texas, closely followed by Prof Carl von Linde’s machine
in Germany in 1876. In 1874 Raoul Pictet produced a
system in Switzerland based on a sulphur dioxide
compressor, and a few years later, in 1878, methyl chloride
systems were also developed.
Thus the five main refrigerants of the 19th century, ether,
carbon dioxide, ammonia, sulphur dioxide and methyl
chloride were introduced over a 25 years period and vied
with each other and with alternative technologies to
dominate the market. It is evident that ease of use was a
prime factor in system selection, followed to a lesser extent
by reliability, space required, installation cost, efficiency
and safety. All the systems in use were to some extent
hazardous, either because the refrigerant was highly
flammable (various ethers, naphtha, chemogene, methyl
chloride), or noxious (ethers, ammonia, sulphur dioxide) or
required high-pressure equipment (carbon dioxide).
Development tended to progress where there was a strong
commercial demand for refrigeration, and the local balance
of selection factors meant that different systems gained
popularity in different markets. Absorption was popular in
early ice plant because it was simple and relatively easy to
construct, requiring no compressor and no prime mover—
usually a steam engine in the nineteenth century. It lost
popularity because it was unreliable, possibly because
systems were relatively large and because they operated
intermittently, not continuously. Air cycle, like absorption,
was relatively simple and found favour for early marine
installations because the plant was relatively compact. In
systems based on Gorrie’s design there was no need for a
complex and messy brine system as the air could be used in
open systems, drawing from the hold, compressing, cooling
and expanding back into the refrigerated space. The
disadvantage of snow forming in the suction was to an
extent overcome by using suction superheaters and
improved valve designs. Ether was first choice in vapour
compression because it exists as a liquid at room
temperature, but this means that to chill brine or produce
ice the suction pressure is sub-atmospheric. This required
relatively large compressors and led to unreliability,
including the risk of explosions.
By the 1880s the capacity required of installations on
land made efficiency a more important factor in system
selection. Air cycles required 8–10 times the coal required
of an ammonia plant, and absorption systems required 60%
more fuel. In addition, most systems were cooled by river
water, and the absorption system was reckoned to require
two and a half to three times more water than an ammonia
compression plant. Ammonia was still not preferred at sea,
because the noxious smell posed a major hazard in the event
of a leak below decks. For small systems air continued to be
used, and carbon dioxide compressors improved sufficiently
to make them the preferred choice for larger systems. Franz
Windhausen patented an improved carbon dioxide com-
pressor in 1886 and this design was licensed and further
1143
improved by Everard Hesketh of J&E Hall who developed a
compound compressor to improve efficiency of carbon
dioxide systems in 1889. Over the next 6 years Halls
installed over 400 such systems, mainly on ships, although a
few were for dockside cold stores. Space requirements were
obviously important on ships, but less so on land; and the
same could be said for safety. A plant explosion or a major
leak of toxic gas could cause a ship to go down with all
hands. Although there were several disasters on shore with
ammonia plant, most notably the fire in the Columbian
exhibit at the World’s Fair in Chicago in 1893, which caused
17 deaths, owners and designers of plant were rather more
relaxed about safety. Thus land and marine designs
diverged. Ammonia compressors were typically larger, but
being a lower pressure design could be built quite cheaply.
They were generally reliable, and were kept substantially
leak-tight, owing to the nasty smell of ammonia. Carbon
dioxide compressors were much smaller, but of a heavy
construction to contain the pressures of 50 or 60 bar
required. In some cases pressures were even higher,
permitting supercritical operation. Usually the high pressure
was limited to a heavy walled cylinder, and the crankcase
was open, with shaft seals on the piston rod, rather than the
modern, sealed crankcase designs The driver, still usually a
steam engine, was much larger than the compressor.
5. Impediments to progress
It took about a century for refrigeration to progress from
a laboratory curiosity of no commercial value to the basis of
a fledgling industry, and a further 50 years to grow this new
industry into a thriving market. The slow progress is simply
explained by a complete lack of appreciation of the
commercial possibilities for artificial cooling. However,
even when the new technology clearly and easily met a
need, there were still barriers to be overcome. John Gorrie
was a well-respected figure in his local community and had
been mayor of his hometown, Apalachicola, in 1836. When
he built his artificial ice maker in 1844 he was afraid of
adverse comment from the local church leaders, so he did
not publish his remarkable feat. Writing under the pen-name
‘Jenner’ he predicted in the local press that it might be
possible to make ice in such a manner. The New York Globe
duly reported that ‘there is a crank, down in Apalachicola,
Florida, that thinks he can make ice by his machine as good
as God Almighty.’ Dr Gorrie did not attempt to
commercialise his invention until 1851 when his US patent
was published (Thévenot [13]).
Patents could also act as impediments to progress, as
competitors tried to establish themselves in new markets.
J&E Hall and the Haslam company were both embroiled in
legal actions instigated by Bell and Coleman over a patented
method of removing moisture from suction gas in air
compressors. The dispute was only resolved when Haslam
purchased the patent rights from Bell and Coleman, and
1144 A. Pearson / International Journal of Refrigeration 28 (2005) 1140–1148
Everard Hesketh of Halls designed an alternative moisture
separator (Miller [8]). Both the technical and the
commercial considerations in this type of dispute must
have acted as a major distraction from the development and
testing of these new systems. Much innovative engineering
may have been ‘wasted’ in circumventing existing patent
restrictions rather than developing new concepts.
Lack of commercial ability may also have restricted
progress in artificial cooling, as many of the innovators were
unschooled mechanics from a rural background or pro-
fessionals from medicine, academia, publishing, and even
theology. Dr Gorrie was unfortunate because when he
finally agreed to develop his icemaker, his principal backer,
a Boston businessman died within a few months. The project
never recovered, and Gorrie himself died a few years later in
1855. It is notable that David Boyle, producer of the first
ammonia compressor only produced 20 compressors a year
and went out of business after 10 years. The company
formed by Prof Linde 4 years after Boyle’s produced over
750 compressors for breweries alone in the first 15 years of
its existence (Hård [2]), and it is still a major player in the
refrigeration and process gas markets worldwide nearly 130
years later. However, for every success, including Frick,
Vilter, York, Sabroe, Halls and Sulzer, there were many
more who failed to survive.
Lack of appropriate machines, materials and manufac-
turing techniques was another brake on progress. In some
cases, developments in related fields such as the construc-
tion of steam engines and internal combustion engines
provided the technical insight required to move refrigeration
technology forwards. Thaddeus Lowe seems to have been
successful in manufacturing ice with a closed cycle carbon
dioxide circuit in 1867, but he neither developed the system
further, nor licensed his technology to others, perhaps
because the pressures required were too great for the
available machinery. Fifteen years later carbon dioxide was
‘rediscovered’ by Raydt (1881), Linde (1882) and Wind-
hausen (1886). The development work of J&E Hall, using a
Windhausen machine as the basis, finally established carbon
dioxide as a viable technology from 1887 onwards.
Sometimes new techniques had to be developed for the
burgeoning refrigeration industry, for example the electric
welding of brine pipes, which was pioneered by J&E Hall in
1890 for their carbon dioxide installation on SS Highland
Chief.
Lack of scientific data must also have been a handicap.
Given the lack of understanding of thermodynamics and the
lack of physical information, it is remarkable that any
progress at all was made. This is a tribute to the powers of
observation and meticulous experimental practices of these
early pioneers, who designed and constructed working
systems without any hard information on refrigerant
properties. The full thermo-physical properties of carbon
dioxide were not in fact issued until Rudolf Plank’s tables
were published in 1929.
6. The decline and fall of carbon dioxide
Carbon dioxide gained favour as a refrigerant for the
marine market in the 1880s because it was substantially
more efficient than the open-circuit air cycle systems used
up until then, and it was also more reliable. Raydt’s 1884
British Patent (number 15,475 in the name of H Lake [5])
and Windhausen’s of 1886 (number 2864 [14]) list several
advantages of ‘liquid carbonic acid’, including being
‘already much cheaper than nearly all chemicals used as
yet in ice-machines’, and being ‘a much more intense
vehicle of cold than the gases heretofore used’. It was also
stated that ‘cold, of almost any low degree, can be
produced’, and ‘in case of leakage, no more or less
unpleasant gases which are deleterious to health enter the
work-room’. Contemporary accounts of the trials of these
early carbon dioxide systems report ‘unparalleled’ quality
levels for cargoes, together with coal consumption only one-
fifth that required for an equivalent size of air cycle
machine. Open air systems relied on fans to circulate air
around the hold and this could lead to warm spots within the
cargo, but the carbon dioxide system used brine grids on the
walls to provide exceptionally even temperatures through-
out. Although more expensive to construct, the system was
cheaper to run, and quickly dominated the marine market.
However, this ascendancy was restricted to marine systems,
where ammonia was generally not acceptable. For land-
based systems an ammonia plant for brine chilling or ice
making could be constructed more cheaply but run more
efficiently. Prime movers were usually steam engines, so
low efficiency translated into high coal consumption, which
was immediately evident to plant owners. At this time heat
rejection was usually to sea water or river water, so in
temperate climates like Britain and the northern parts of the
USA it was possible to run carbon dioxide systems sub-
critical in a traditional Rankine cycle. Water usage,
however, was as apparent as coal usage, and the introduction
of atmospheric condensers by the De La Vergne company
and L Sterne&Company in the 1880s was quickly followed
by many other ammonia installers. This development, a
large, open-air pipe grid sprayed with water, acted as a
natural circulation evaporative condenser. It greatly reduced
water consumption and allowed plants to be built further
away from river or lake water. Condensing temperatures
tended to be a bit higher, and either ruled out carbon dioxide
completely, or made it much less efficient, as the latent heat
available close to the critical point reduced significantly.
Ships continued to use carbon dioxide because it was safer,
and provided the sea water temperature was below 20 8C it
was reasonably effective, but even proponents of the carbon
dioxide systems had to introduce ammonia plant to their
product range. By 1910 J&E Hall had been established as
the pre-eminent marine refrigeration builder in the world for
20 years, but felt compelled to introduce a range of ammonia
compressors to satisfy the home market for cold storage,
brewing, ice making and skating rinks.
 A. Pearson / International Journal of Refrigeration 28 (2005) 1140–1148
As the 20th century progressed and manufacturing
standards improved, ammonia’s safety record also began
to improve. Pressure relief valves for ammonia plant were
introduced in the New York city safety code in 1915, and
other authorities quickly followed suit. Improved welding,
the use of electric motors instead of steam engines and the
introduction of smaller, faster running compressors all made
ammonia more feasible for carbon dioxide’s traditional core
market on board ship. Several innovations attempted to stem
the tide. The Haslam company patented a novel economiser
system for carbon dioxide reciprocating compressors in
1923 [3] in an attempt to match the efficiency of ammonia
systems as condensing temperatures tended towards the
critical point (Fig. 2).
In 1932 the Frick Company, in response to ongoing
safety concerns about large ammonia charges, started
installing a hybrid system which used carbon dioxide for
the low temperature stage, with a much smaller ammonia
plant providing the necessary refrigeration to condense the
carbon dioxide at moderate temperatures and pressures [4].
What Frick called the ‘split-stage’ system, shown in Fig. 3
was identical in principle to the modern carbon dioxide/
ammonia cascade systems which have been reported over
the last 10 years or so. Even in 1932, however, this concept
was already 65 years old, having been first proposed by
Tellier in 1867 [12], as a method of casting calcium
carbonate replicas of lifesized marble statues. It would
Fig. 2. Haslam’s Patent liquid pre-cooler.
1145
appear that, in Tellier’s time, cascade systems were deemed
to be too complex, and plants were not required to operate at
the low temperatures now demanded of industrial freezers.
By the 1930s when Kitzmiller designed Frick’s ‘split-stage’
system, the lower temperatures were necessary, but most
operators seemed to be willing to accept the hazards
associated with running a large ammonia plant, so there was
no need to go to the extra expense of installing a cascade
heat exchanger.
Neither Haslam’s economiser, nor Frick’s cascade was
able to reverse the movement away from carbon dioxide in
industrial systems. Even at sea, ammonia plant was
preferred for its higher efficiency under tropical conditions.
From 1930 until the 1950s industrial refrigeration did not
really see any startling developments. The major thrust of
refrigeration research was the proliferation of refrigerators
for the domestic and light commercial market. New
synthetic refrigerants—the chlorofluorocarbons—had been
produced as a result of a specific, market-driven research
program led by General Motors and DuPont. The major
battle for technical supremacy was between GM’s vapour
compression refrigerators and Electrolux’s absorption
systems, and the weight of research funding behind GM’s
program won the day conclusively. Having produced the
fluids, it was then necessary to develop suitable compres-
sors, condensers, controls and evaporators to exploit CFC’s
potential to the full. Small hermetic compressors for
1146 A. Pearson / International Journal of Refrigeration 28 (2005) 1140–1148

Fig. 3. Frick’s ‘split-stage’ ammonia/carbon dioxide system.

refrigerators and air-conditioners provided the means to
achieve this, and once they were established the focus
returned to the industrial sector. First R12, then R22, then
R502 were introduced to industrial systems over the period
from 1950 to 1970, almost completely supplanting carbon
dioxide in the marine market, and seriously threatening
ammonia in the land market. Here was a family of chemicals
able to provide the efficiency and flexibility of ammonia
with the safety and reliability of carbon dioxide. In parallel,
new compressors running at previously unknown speeds of
up to 600, 800 and even 1000 rpm were developed, making
plants smaller, lighter, cheaper and easier to maintain—
although not necessarily more efficient. The proliferation of
cheap electricity had shifted the emphasis in plant design
well away from efficiency by this time, and it could be
argued that it has not yet returned.
7. Reappraisal
The rapid decline of CFC systems in the late 20th
century has resulted in a tremendous increase in refriger-
ation research, as ‘new’ alternatives are sought. This search
has included a return to some old techniques including
ammonia and carbon dioxide. Carbon dioxide was identified
as a practicable option for various refrigeration cycles in

Fig. 4. Presentations about CO2 at IIR conferences and congresses.

 A. Pearson / International Journal of Refrigeration 28 (2005) 1140–1148
several areas almost simultaneously. In 1990 Prof Gustav
Lorentzen published a patent application for a trans-critical
carbon dioxide system for automotive air-conditioning [6].
In 1991 Dr Forbes Pearson submitted patents in Britain,
France, Germany and the USA on the use of carbon dioxide
as a volatile secondary refrigerant, including a novel hot gas
defrost system [10]. At much the same time, Stal AB
developed the use of carbon dioxide as a volatile secondary
refrigerant in supermarket systems for the Swedish market,
and the Liquid Carbonic Corporation published patents in
Spain and the USA covering a configuration similar to the
Frick ‘split-stage’ system of 1932, applied to spiral freezers.
This seems like a rapid rediscovery, but most of the research
effort applied to this topic to date has in fact been in the
commercial, air conditioning, automotive and heat pump
markets. Until recently, the industrial sector, which was the
only market for carbon dioxide systems in the 19th century,
has been somewhat neglected. In 1994 Prof Lorentzen
inaugurated the series of IIR conferences which now bear
his name, on ‘New Applications of Natural Working Fluids
in Refrigeration and Air-Conditioning’. Fig. 4 shows the
number of CO2 papers presented at these conferences and at
the IIR’s Congresses since then. The figures above each
column are the percentage of the total number of papers at
that conference that were related to carbon dioxide.
Lorentzen’s address to the first conference was also
summarised in an article in the International Journal of
Refrigeration [7] which covered all major sectors of the
refrigeration market.
It can be seen that, within the Gustav Lorentzen
conferences from 1994 to 2004 the number of papers on
carbon dioxide in one form or another has risen from six to
50, and the proportion of the total conference has risen from
7 to 48%. Likewise at the Congress in the Hague in 1995
there was only one paper on carbon dioxide out of four
hundred, but in Sydney in 1999 there were 16 (3.5%). In
Washington, DC in 2003 there were 31 papers (7.2%).
However, although this increase is dramatic, it has been
concentrated on the commercial and air conditioning
markets. From all these conferences there are only 10
papers specifically on industrial topics out of nearly 200 on
carbon dioxide.
In response to this lack of development in the industrial
field a group of European contractors, end-users, academics
and manufacturers formed an ‘interest group’—a forum for
the exchange of ideas, experiences and needs. The carbon
dioxide interest group (c-dig) met for the first time in
Switzerland in July 2000 with nine organisations rep-
resented. Initially the group was structured to avoid
commercial conflict, but it quickly became apparent that
information should be disseminated as widely as possible to
encourage installations in as many countries and market
segments as possible. To date the group has met on 12
occasions, and is now structured on more formal lines.
Topics investigated have included the testing of compres-
sors and lubricants, evaluation of cascade performance,
1147
investigation of trans-critical systems, heat exchanger
design and oil separator design. Related topics have
included the physiology of carbon dioxide and the effects
of water in CO2 systems. These presentations have been
extremely useful in generating a rapid field development,
but they have been of a pragmatic, practical nature, and in
general they have not been written up for presentation in a
more formal academic context. During this time several
industrial carbon dioxide systems have been commissioned,
and site visits have been conducted for c-dig by members.
These have included a visit to a Swiss ice rink in Bern,
converted by W Wettstein AG from ammonia to carbon
dioxide, a visit to the coffee freeze drying plant installed by
Star Refrigeration for Nestle at Hayes, and visits to various
Dutch installations by York and Bort de Graaf. Meetings
have also been held at the laboratories of DTI in Aarhus,
Denmark, TNO in Apeldoorn, The Netherlands, TU-
Dresden and ACRC in Urbana-Champaign. At the same
time, other companies, mainly also in Europe, have also
been applying carbon dioxide to industrial systems. These
experiences have shown carbon dioxide to be eminently
suited to the requirements of modern industrial systems,
whether used as the low temperature fluid in a cascade
system, as an evaporating secondary refrigerant or as the
refrigerant in a transcritical plant.
8. Future possibilities
The limiting factor for most carbon dioxide systems is
currently pressure. This does not appear to be a long-term
impediment, and compressors, pumps, valves and controls
are already on the market suited to operation at 40 bar g.
This is sufficient for cascade operation with an intermediate
temperature of about 0 8C, but it is not quite high enough to
enable an effective hot gas defrost to be engineered. The key
development required in the near future for industrial
systems is therefore compressors capable of operation at
50 bar g. for cascades, or 100 bar g. for trans-critical
systems. The latter have not yet been applied to industrial
applications, but this may follow, provided compressors are
available, and appropriate control devices can be devised.
These systems will be particularly appropriate where there
is a need for high-grade heat recovery. The use of
economised circuits will also gain importance as cascade
system intermediate pressures increase since the percentage
benefit of economising increases as the critical temperature
is approached. These developments in the industrial field
need not only apply to very large systems. As compressor
development continues and smaller machines become
available, it will be possible to engineer packaged cascade
plant comprising a semihermetic carbon dioxide compressor
of, say 50 kW capacity with a suitable ammonia or propane
compressor, perhaps also semihermetic, using brazed plate,
plate and shell or microchannel heat exchangers to give a
low charge, virtually leak-proof, compact installation using
1148 A. Pearson / International Journal of Refrigeration 28 (2005) 1140–1148
only ‘natural’ refrigerants. The system could be arranged to
provide reverse cycle defrosting of low temperature CO2
evaporators, making it eminently suitable for small freezers
and low temperature cold stores. Efficiencies for this system,
should be at least as good as for economised single-stage
ammonia plant, and better than a typical HFC installation of
this size, and capital cost will depend primarily on the unit
cost of the components. This, in turn, will be primarily
dictated by the size of the market.
9. Conclusion
Carbon dioxide system development was driven in the
19th century by the shortcomings of the alternatives; air
cycle, ether, absorption and ammonia. It was impeded by
lack of knowledge, lack of commercial awareness, lack of
manufacturing capability and lack of concern for safety.
Carbon dioxide was probably the cheapest available
refrigerant. One system patent even describes it as a by-
product of the production of calcium chloride, used as the
brine for the ice-maker. In the 21st century it is no longer
necessary to make your own brine or carbon dioxide.
However, many of the drivers for development—short-
comings in the alternatives—have come to the fore again.
This time round, manufacturing is easily able to cope with
the requirements, and an increased level of safety awareness
backed by appropriate international codes and legislation
will help to make carbon dioxide a preferred choice for
industrial systems in the near future.
References
[1] W. Gosney, James Harrison and the birth of mechanical
refrigeration, Proc Inst Refrigeration 90 (1994) 12–22.
[2] M. Hård, Machines are frozen spirit, Campus Verlag,
Frankfurt, 1994. p. 134.
[3] A.S. Haslam, Improvements in compression refrigerating
apparatus, British Patent No. 165929; 1921.
[4] F. Kitzmiller, CO2 split stage, Power (1932) (19 January 1932).
[5] H. Lake, Improvements in machines or apparatus for the
manufacture of ice, British Patent No. 15,475, London; 1884.
[6] G. Lorentzen, Trans-critical vapour compression cycle device
Patent No. WO/07683; 1990.
[7] G. Lorentzen, Revival of carbon dioxide as a refrigerant, Int J
Refrigeration 17 (1994) 292–300.
[8] H. Miller, Halls of Dartford 1785–1985, Hutchinson Benham,
London, 1985.
[9] W. Newton, Refrigerating, British Patent No. 952, London;
1867.
[10] S.F. Pearson, Cooling method and apparatus, British Patent
No. 2258298, London; 1995.
[11] J. Pettersen, CO2 as a primary refrigerant, Institute of
Refrigeration Centenary conference, Inst Ref London; 1999.
[12] C. Tellier, L’ammoniaque dans l’industrie, Paris; 1867: p.
277–279.
[13] R. Thévenot, A History of Refrigeration, IIF/IIR, Paris; 1979.
[14] F. Windhausen, Improvements in apparatus for refrigerating
purposes, British Patent No. 2864, London; 1886.
[15] W.R. Woolrich, The men who created cold: a history of
refrigeration, Exposition Press, New York, 1967.