Trade-Offs in Refrigerant Selections:
`Past, Present, and Future
`
`James M. Calm and David A. Didion
`
`REFRIGERANTS FOR THE 21st CENTURY
`
`ASHRAE/NIST REFRIGERANTS CONFERENCE
`
`National Institute of Standards and Technology
`
`October 6-7, 1997
`
`American Society of Heating, Refrigerating
`and Air-Conditioning Engineers
`
`0 1997 by the American Society of Healing, Refrigerating and Air-Conditioning Engineers, Incorporated.
`Figures I-4, 6, 9, and I0 previously copyrighted in i995 and 1997 by James M. Calm, Engineering Consultant.
`
`Page 1 of 16
`
`Arkema Exhibit 1112
`
`Arkema Exhibit 1112
`
`

`
`Trade-Offs in Refrigerant Selections:
`Past, Present, and Future
`
`James M. Calm, P.E.
`Fellow ASH RAE
`
`David A. Didion, P.E.
`Fellow ASHRAE
`
`ABSTRACT
`
`Recent attention to depletion of stratospheric ozone,
`by chemicals containing bromine and chlorine, resulted
`in an international accord to halt their production. The
`most widely used refrigerants are among them. Chemi-
`cal and equipment manufacturers mounted aggressive
`research and development programs to introduce alter-
`native and transition refrigerants, associated lubricants
`and desiccants, and redesigned equipment. The already
`difficult criteria became even more complex, with Slb-
`sequent linkage of chemical emissions lrom human ze-
`tivities to global climate change. The very successful
`response to protect the ozone layer has led some regu-
`lators and users to assume that ideal substitutes will be
`found. Such chemicals should be free of all environ-
`
`mental and safety concerns, be chemically and thermally
`stable, and perform efficiently. The analyses presented
`in this paper demonstrate that the outlook for discovery
`or synthesis of ideal refrigerants is extremely unlikely.
`Trade-offs among desired objectives,
`therefore, are
`necessary to achieve balanced solutions. The paper
`also shows that fragmented regulation of the chemicals
`involved,
`to address individual
`issues, jeopardizes the
`prospect of solving subsequently addressed problems.
`The paper reviews the history of refrigerants, their roles
`in ozone depletion and global climate change, and neces-
`sary trade-offs in refrigerant selections.
`
`INTRODUCTION
`
`The air-conditioning and refrigeration industry is in
`the midst of an unprecedented transition, catalyzed by
`environmental concerns with the impacts of refrigerant
`emissions. Compounds containing chlorine and, though
`less frequently used as refrigerants, bromine are being
`phased out. Production of chlorofluorocarbons (CFCs)
`ended in I995 in developed countries. Attention is now
`focused on the next issue, global warming, with a view
`toward reducing greenhouse gas emissions. Chemicals
`with long atmospheric lifetimes and high fluorine con-
`tents are targeted as potent greenhouse gases.
`Per-
`
`James M. Calm is an engineering consultant in Great Falls,
`VA, USA. David A. Didion is a Fellow in the Building and Fire
`Research Laboratory of the National Institute of Standards
`and Technology, Gaithersburg, MD, USA.
`
`fluorocarbons (compounds consisting solely of carbon
`and fluorine) already are perceived as doomed by the
`technical community, even though specific treaty and
`regulatory provisions are still being worked out.
`Although less visible, a parallel
`transition toward
`safer refrigerants, based on both flammability and toxic-
`
`Tablc 1: Historical introduction of refrigerants
`
`year
`
`refrigerant
`(/absorbent)
`
`18305
`
`caoutchoucine
`
`18405
`1850
`1856
`1859
`1866
`
`l 8605
`
`I 870
`
`l 875
`1878
`l 8705
`l 89]
`
`l900s
`
`1912
`
`l920s
`
`I922
`I923
`I925
`I926
`
`sulfuric (ethyl) ether
`
`methyl ether (R-E170)
`water / sulfuric acid
`
`ethyl alcohol
`ammonia / water
`
`chymogene
`
`carbon dioxide
`
`ammonia (R-7 l 7)
`methyl amine (R-630)
`ethyl amine (R-631)
`methyl formate
`(R-6] l)
`sulfur dioxide (R-764)
`methyl chloride (R-40)
`ethyl chloride (R-I60)
`blends of sulfuric acid
`
`with hydrocarbons
`ethyl bromide
`(R-160Bl)
`carbon tetrachoride
`
`water vapor (R-718)
`isobutane (R-600a)
`propane (R-290)
`dielene (R-l I30) “
`gasoline
`trielene (R-1 120)
`methylene chloride
`(R-30)
`
`chemical formula
`
`or makeup
`
`distillate of india
`rubber
`
`CH3—CH3-0-CH3-
`CH3
`CH3-0-CH3
`H30 / I-l3S04
`CH3-CI-I3-OH
`NH; / H30
`petrol ether and naph-
`tha (hydrocarbons)
`C03
`NH;
`CH3(NH3)
`CH3-CH3(NH3)
`HCOOCH3
`
`S03
`Cl l3Cl
`CH3-Clrl3Cl
`H2504, C4Hio, C5Hi2,
`(Cl-l3)3CH-CH;
`Cl'l3-CHQBT
`
`CCI4
`H30
`(CH3)3CH-CH3
`Cl-[3-CH3-CH3
`CHCl=CHC|
`
`hydrocarbons
`CHCl=CCl3
`CH3Cl3
`
`a blend of cis- and tran.s'- l ,2-dichloroethene isomers
`
`6
`9 213! Canmrv Prnceerlinas of the ASHRAE/NIST Refrkuerants Cnnferenrte Gaithershura MD USA October 6-7 1997
`
`

`
`'/'/‘at/e-()/is in Itaffirgc/‘uni Se/cclio/rs:
`
`/’u.s'I, I’re.s'uIiI, and /71/!uI't'
`
`ity, has been underway. Unfortunately, some of the
`obvious solutions to the environmental concerns raise
`
`the hydrocarbons
`
`flammability and/or toxicity concerns;
`and ammonia are examples.
`The concept of sequentially eliminating chemicals by
`class, defined by molecular composition,
`raises
`the
`question of what will be left as successive issues are
`addressed (Wucbb|es and Calm,
`I997). The answer
`begins with examination of the refrigerants now used.
`
`HISTORICAL REFRIGERANTS
`
`times, using
`Refrigeration goes back to ancient
`stored ice and a number of evaporative processes; they
`are outlined by Thévenot (1979). Oliver Evans pro-
`posed the use of a volatile fluid in a closed cycle to
`freeze water into ice (Evans,
`I805). He described a
`system that produced refrigeration by evaporating ether,
`under a vacuum, and then pumped the vapor to a water-
`cooled heat exchanger to be condensed for reuse. Al-
`though there is no record that he built a working im-
`chine, his ideas probably influenced both Jacob Perkins
`and Richard Trevithick. The latter proposed an air-
`cycle system for refrigeration in I828, but he also did
`not build one.
`
`Actual refrigerants were introduced in the l830s,
`with invention of the vapor-compression machine by
`Perkins. He designed the machine to use sulfuric
`(ethyl) ether as the refrigerant. His patent describes a
`cycle using a “volatile fluid for the purpose of producing
`the cooling and freezing
`and yet at the same time
`condensing such volatile fluids, and bringing them into
`operation without waste” (Perkins, I834). The first ma-
`chine actually used caoutchm/cine, an industrial solvent
`that Perkins apparently utilized in his business as a
`printer.
`It seems the first trade-off in refrigerants —-
`and one still driving selections —-— was based on avail-
`ability.
`Table 1 summarizes early refrigerants, namely those
`
`predating fluorinated chemicals. Downing (I988), Na-
`gengast (I989 and I996). and Thévenot (I979) present
`further details.
`
`The first century of refrigerant use was dominated
`by innovative efforts with lamiliar fluids in almost pro-
`totypical machines. The goals were to provide refrig-
`eration and,
`later, durability. Use of blends was at-
`tempted where single-compound solutions could not be
`found (Pictet, I885).
`As production increased following World War I, a-
`tention turned to safety and performance as well. Willis
`H. Carrier, known for his advances in psychromctrics
`and air conditioning, and R. W. Waterlill initiated one of
`the first documented systematic searches (Carrier and
`Waterfill, I924).
`'l‘hcy investigated a range of candidate
`refrigerants for suitability, for both positive displacement
`and centrifugal compression machines. These analyses
`closely examined ammonia, ethyl ether, carbon dioxide,
`carbon tetrachloride, sulfur dioxide, and water. They
`concluded, for example, that the performance of carbon
`dioxide would depend on the cycle and amount of liquid
`subcooling, but that it yielded the lowest predicted per-
`formance of the fluids analyzed. They also noted that
`ammonia and water would require excessive stages of
`compression for the conditions sought. and that water
`“gives a low efficiency of performance.” Sulfur dioxide
`was discarded for safety reasons and carbon tetra-
`chloride because it attacks metals, especially in the
`presence of water. They finally selected dielene (I,2-
`dichloroethene, R-II30) for the first centrifugal ma-
`chine, though an international search was needed to find
`a source for it (lngels, I952).
`
`Midgley Elements
`
`Nearly all of the early refrigerants were flammable,
`toxic, or both, and some also were highly reactive. Ac-
`cidents were common. For perspective, propane was
`marketed as the odorless safety refrigerant (CLPC,
`I922).
`
`brillle metals
`
`ductile metals
`
`||0WI
`melt
`
`l'10l'|fl18l8lliC
`
`-- lanthamdos (rare earth elements)
`
`.- actinides lransuranium elements)
`
`..4
`
`. R
`
`n‘-'-{_._.,
`
`;.
`
`Figure 1: Periodic table of the elements
`
`highlighting Midgley‘s selections
`
`Page 3 of 16
`
`...-
`
`,.‘- ._1
`
`The discovery of lluorinatcd refrig-
`crants began with a phone call to Tho-
`mas Midgley, Jr.,
`in I928. He already
`had established himself by finding tetra-
`ethyl
`lead, to improve the octane rating
`of gasoline. The caller stated that “the
`refrigeration industry needs a new rc-
`frigcrant
`if they expect
`to get any-
`where” (Midgley, I937).
`With his associates Albert L. Hennc
`
`R. McNary, Midgley
`and Robert
`scoured property tables to find chemi-
`cals with the desired boiling point. They
`restricted the search to those known to
`
`be stable, but not toxic or flammable.
`An error in the published boiling point
`for tetrafluoromethanc (carbon tetra-
`
`

`
`}. Calm and l). Didion
`
`fluoride) drew their attention to the organic fluorides;
`the correct boiling temperature later was found to be
`much lower. Nevertheless, the incorrect value was in
`the range sought and suggested consideration of fluori-
`nated chemicals. While fluorine was known to be toxic
`
`by itself, Midgley and his collaborators felt that com-
`pounds containing it could be nontoxic.
`Recognizing the deficiencies of the published litera-
`ture, Midgley turned to the periodic table of the eb-
`ments. He quickly eliminated those yielding insufficient
`volatility. He then eliminated those resulting in unstable
`and toxic compounds as well as the inert gases, based
`on their low boiling points. He was left with just eight
`elements:
`carbon, nitrogen, oxygen, sulfur, hydrogen,
`fluorine, chlorine, and bromine. They clustered at an
`intersecting row and column of the periodic table of the
`elements, with fluorine at the intersection. They are
`shown on a modern periodic table in figure 1; see
`Midg1ey(I937) for the arrangement then used.
`Midgley and his colleagues then made three inter-
`esting observations. First, flammability decreases from
`lefl to right for the eight elements as shown in figure 1.
`Second,
`toxicity generally decreases from the heavy
`elements at the bottom to the lighter elements at the top.
`And third, every known refrigerant at
`the time was
`made from combinations of those elements. This obser-
`
`vation can be verified by comparing the compositions of
`the historical refrigerants,
`listed in table 1, to figure l.
`The historical fluids actually comprised only seven of the
`eight Midgley elements, since there appears to be no
`record of prior use of refrigerants containing fluorine.
`The first publication on fluorochemical refrigerants
`shows how chlorination and fluorination of hydrocarbons
`can be varied to provide desired boiling points (Midgley,
`1930). This paper also shows how the composition i1-
`fluenoes relative flammability and toxicity. Commercial
`production of R-l2 began in 1931, followed by R-ll
`in
`1932 (Downing, I966 and 1984).
`-
`Other investigators have repeated Midgley’s search
`with newer methods and modern databases, but they
`have come to similar findings. McLinden and Didion
`(1987) documented an extensive screening of industrial
`chemicals. Of the chemicals meeting their criteria, all
`but two — both highly reactive and toxic — consisted
`of the Midgley elements.
`
`R-410
`R-500
`
`ca
`
`06
`
`04
`
`2000
`
`4000
`
`6000
`
`3000
`
`" ODP (relative to R-11)
`
`I
`
`GWP (relative to C0,) '*
`
`Figure 2: ODP versus GWP
`for common refrigerants
`
`Stratospheric Ozone Depletion
`
`The first global environmental problem identified was
`depletion of stratospheric ozone. The problem arises
`from destruction of ozone molecules in the upper atmos-
`phere, primarily by bromine and chlorine from anthropo-
`genic chemicals. The chlorine and bromine react cata-
`lytically to destroy ozone molecules,
`thereby reducing
`the natural shield from incoming ultraviolet—B radiation.
`Mario Molina and F. Sherwood Rowland (I974) identi-
`fied CFCS as a source for the chlorine in the strato-
`
`sphere and the potential for more serious ozone deple-
`tion, with projected growth in use of these chemicals.
`The index used to indicate the relative ability of a re-
`frigerant or other chemical
`to destroy stratospheric
`ozone is the omne depletion potential (ODP) (Wuehbles,
`l98l). Figure 2 shows the ODPs of common refriger-
`ants and selected candidates for future use. The shaded
`
`tips on several of the ODP bars indicate the higher val-
`ues obtained by semi-empirical determination, as con-
`trasted to the modeled values shown in black (WMO,
`I994).
`Chlorinated and brominated refrigerants, along with
`similar solvents, foam blowing agents, aerosol propel-
`lants, fire suppressants, and other chemicals are being
`phased out under the Montreal Protocol, a landmark
`international treaty to protect the ozone layer (UNEP,
`I987 and 1997).
`
`ENVIRONMENTAL ISSUES
`
`Global Warming
`
`More than forty years passed, during which fluoro-
`chemicals became the dominant refrigerants, until they
`were connected with environmental concerns (Molina
`and Rowland, 1974). Other fluorochemicals also were
`introduced in this period, and most of the earlier refrig-
`erants were retired. A notable exception is R-717
`(ammonia), which remains the preferred refrigerant in
`some industrial applications.
`
`Concerns also have been raised with the prospect of
`global climate change. The average temperature at the
`surface of our planet results from an equilibrium,
`le-
`tween incoming solar energy and heat radiated back into
`space. Most of the latter is in the infrared range of
`emissions. Gases that absorb this infrared energy en-
`hance the greenhouse effect of our atmosphere, leading
`to warming of the Earth. While this mechanism is at-
`
`Page 4 of 16
`
`

`
`feedback
`cepted, scientific debate remains on natural
`mechanisms and both the timing and extent of the re-
`sultant warming,
`Refrigerants, most notably those with long atmos-
`pheric lifetimes and high numbers of carbon-fluorine
`bonds, have been identified as greenhouse gases. The
`measure most commonly used to quantify the degree of
`concern is the global warming potential (GWP). Figure
`2 shows GWP values alongside the ODPs for common
`refrigerants and candidates. The GWPS depicted are
`relative to the warming effect of a similar mass of car-
`bon dioxide for lOO year time frames. Shorter integra-
`tion periods emphasize near-term effects, while longer‘
`intervals better reflect
`the total
`impact of a release.
`The Intergovernmental Panel on Climate Change as-
`sessment (IPCC,
`l966) and Wuebbles (I995) discuss
`the influence of the integration period and present data
`values.
`Carbon dioxide is used as the reference chemical for
`
`GWPS because it is the one with the greatest net in-
`pact. Other chemicals, including most refrigerants, are
`more potent as greenhouse gases; the difference comes
`from the increasing abundance of carbon dioxide in the
`atmosphere. Most of the change results from increased
`use of fuels, in combustion processes, to meet our en-
`ergy needs.
`Air conditioners, heat pumps, and refrigeration tb-
`vices that use refrigerants also use energy. They con-
`tribute to global warming both by release of refrigerants
`and by emission of carbon dioxide and other greenhouse
`gases,
`in powering the devices. Detailed studies have
`shown
`that
`energy-related
`component,
`commonly
`dubbed the indirect eflect, is far greater than the direct
`effect from refrigerant releases for Inost applications
`(Calm, I993)
`One expression of the combined effects is the total
`equivalent warming impact (TEWI). Unlike ODP and
`GWP values, which can be determined from measure-
`ments of
`the compounds involved and other atmos-
`pheric data, TEWI determination also requires applica-
`tion data. Among those needed are the fuel or fuel mix
`to power the system, the conversion efficiencies, equip-
`ment efficiencies, loads, refrigerant release rates (manu-
`facturing,
`installation,
`leakage,
`service, disposal, and
`other losses), energy uses for heat rejection, related
`pump or fan energy for distribution systems, and others.
`Calm (i993) and Fischer et al. (l99l and 1994) discuss
`these factors and consequent findings.
`Figure 3 illustrates the comparative magnitudes of
`warming impacts from emission of refrigerants and as-
`sociated energy use for water-cooled chillers. The effi-
`ciencies used in this example are the highest commer-
`cially available in I996. Table 2 presents these efficie n-
`cies, which were taken from an industry survey (ARI,
`I996). The indicators shown are Integrated Part Load
`Values (ll’LVs), expressed both as a coefficient of per-
`
`Page 5 of 16
`
`'1'/'u¢le—()_/is in lt’e_/i'1ge/‘alt!Sr’/cctiom':
`
`I’a.s‘t,
`
`/’I'c.s'w1t and Future
`
`TEWI(kgCO2lkW-yr)
`
`.
`
`I
`
`GMISSIOHS F.
`5:=’.‘l§ii:i%¥1
`i
`
`'
`
`‘
`R22 R123 R1343
`
`TEWI(lbC0,lton-yr)
`
`.
`
`4
`
`'
`7-.1!
`\.ii|'.‘: ,,0
`R22 R123 R1348
`
`1230 kW (350 ton)
`
`3500 kW (1000 ton)
`
`Figure 3: Total equivalent warming impact
`(TEWI) for the best available chillers
`
`Table 2: Best available efficiency (specific power)
`for water-cooled chillers in 1996 based on
`
`certified integrated part-load values (IPLVS)
`
`mmrmscr Wto
`refrigerant
`kW/kW
`kW/ton
`
`3501) 331 1 I000 tgnl
`kW/kW kW/ton
`
`s.enti:i.t.‘um1l
`R-22
`R-I23
`R-I 34a
`
`ggreyy
`R-22
`R-134a
`
`5.96
`7.03
`6.28
`
`6.39
`5.96
`
`NA =- not available
`
`formance (COP, kW/kW) and specific power (kW/ton).
`The survey was restricted to equipment with certified
`performance ratings, and the submissions were verified.
`While the data in table 2 provide an objective com-
`parison of the best chiller offerings then available, three
`caveats are necessary. First, the fiaction of products
`sold with the highest elliciencies offered is small. Com-
`petitive products are available with each of the refriger-
`ants shown at average performance levels, which are
`approximately 20% lower (~20% higher on a kW/ton
`basis). Second, additional options are now available; R-
`4l0A chillers are now marketed. And third, while the
`data are less than a year old,
`improvements have
`emerged in that time. The best available performance
`improved by 5%,
`to 8.l8 kW/kW (fallen to 0.43
`kW/ton), also for a certified IPLV, in the half year since
`the survey.
`lmprovements continue, particularly for R-
`123 and R-l34a. Nevertheless, the tabulated data pro-
`vide a consistent comparison of the highest performance
`options recently available.
`
`

`
`J. (‘aim and D. l)idion
`
`I refrigerant emissions
`El energy related
`
`
`
`
`
`comparativewarmingimpact0/.)
`
`
`
`
`
`average7'_~f.'
`
`R-11
`1985
`
`R-‘l1R123
`1990
`
`R-11 R123
`1992
`
`R-123
`1995
`
`R-123
`2000
`
`Figure 4: Progression and projection for chiller
`total equivalent warming impacts (TEWIS)
`
`The calculation method and remaining data used for
`the analyses summarized in figure 3 are consistent with
`those in Calm (I993). The refrigerant emission rates
`and equipment lives were updated to those presented in
`Calm et al. (I997) and new (JWP data (ll’CC, 1996)
`were used.
`lilectricity generation mixes and heat rates
`were updated to revised projections (NERC, 1995), and
`load profiles were decreased to correspond to an
`equivalent-full load level of I500 hr/yr.
`As shown in figure 3, the direct effect of refrigerant
`emissions amounts to only 2%-4% of the annual total for
`chillers using R-22, which has the highest GWP of the
`three refrigerants shown. These fractions drop slightly,
`to ?.%-3%,
`in chillers with average efficiencies. The
`direct effect
`in figure 3 is 2%-3% of the total for R-
`l34a chillers.
`It is less than 0.2% for R-I23, which also
`offers the lowest energy—related impact, based on the
`highest available efficiency. From a TEWI perspective,
`phaseout of the liydrochIorofluorocarbons (HCl-‘Cs) and
`hydrofiuorocarbons (Hl“Cs) provides only small gain for
`chillers with low refrigerant releases.
`There is far
`greater opportunity by efficiency improvement. The
`same
`conclusion holds
`true
`for
`iriost other
`air-
`
`conditioning and refrigeration products as well. Two
`exceptions are mobile air conditioners and supennarket
`refrigeration systems, which still have high loss rates.
`Figure 4 summarizes the recent progression in effi-
`ciency and reductions in refrigerant releases. The data
`shown are for 1750 kW (500 ton) R-I l and R-I23 chill-
`ers. More than half of the centrifugal chillers installed in
`the years shown used these refrigerant. Similar conch-
`sions can be drawn for R-l2 and R-134a for the same
`
`time frame, although the Tl-LWls both start and end at
`higher levels.
`R-l l dominated in centrifugal chillers almost since its
`introduction. Refrigerant
`losses were high; they often
`exceeded l5% of the total charge each year, as shown
`for I985. Small efficiency gains appeared by 1990, but
`more significant gains were introduced in system tight-
`
`Page 6 of 16
`
`cning and improved purge technologies. Figure 4 shows
`the resultant drop in impacts from emissions. R-I23 use
`also had begun, though the first machines — essentially
`R-I I designs with materials changes for compatibility - -
`yielded l4%-l6% lower efficiency. The direct warming
`impact of R-I23 emissions is barely visible, owing to a
`98% lower GWP.
`
`Release reductions continued, both by equipment
`tightening (e.g., minimization ofthe number of joints and
`replacement of mechanical fittings with brazed connec-
`tions) and improved service practices. By I992, the last
`year in which R-ll chillers were manufactured for it)-
`mestic use in the United States, net emissions for nu-
`chines of comparable capacities, were half those in 1985
`and before. These release reductions offer several
`
`benefits beyond reduced global warming. They also
`lower the impact on ozone depletion.
`()DP and GWP
`only characterize chemical releases. Refrigerant
`that
`does not escape, and is recovered for reuse or safe dis-
`posal, does not harm the environment. Reduced losses
`also eliminate the need for makeup, thereby saving other
`resources and lowering costs,’improve safety, and avoid
`performance losses from insufficient refrigerant charge.
`Figure 4 also shows the dramatic improvements
`made in R-123 chiller optimization by 1992,
`leading to
`higher practical efficiency than available with the rctiied
`R-ll designs. This achievement, and subsequent fur-
`thcr gains, are all the more impressive since R-ll holds
`a theoretical efficiency advantage over R-I23, as shown
`below.
`
`Further improvements in performance followed and
`are expected to continue, but the pace will slow as gains
`approach theoretical limits. The best efficiencies avail-
`able, reflected in figures 3 and 4. are double those for
`many old machines still
`in operation. Coupled with
`emission reductions and R-l23‘s very low GWP,
`the
`best chillers in 1995 reduced net global warming impacts
`by more than 40% compared to typical machines a dec-
`ade earlier.
`
`stratospheric ozone depletion and global
`While
`warming are distinct plienomcna, they are linked in sev-
`eral ways. First,
`thc increase in carbon dioxide ~ a
`greenhouse gas with the highest not impact —- cools the
`stratosplicre. That results in formation of ice crystals,
`which,
`in turn, increases the efficiency of bromine and
`chlorine attack on ozone.
`Second, ozone itself is a
`greenhouse gas;
`reduction of its average equilibrium
`concentration reduces its warming contribution. Third,
`efficiency improvement in energy systems is a key cp-
`tioii
`for long-term reduction of greenhouse gas emis-
`sions. Bliininating the most eliicient refrigerants (for
`example, R-I23),
`to protect the ozone layer removes
`one of the most cost-effective options to reduce global
`warming.
`Uncertainties remain in global warming timing, off-
`sets, and therefore magnitude, but
`it
`is clear that the
`
`

`
`warming issue will be much more difficult to address
`than ozone depletion. That heightens the need for bal-
`anced, rather than fragmented, solutions to global ci-
`mate change issues.
`Simplistic elimination based on
`ODP and GWP alone, without distinction between open
`and closed uses (highly emissive versus contained appli-
`cations), defeats this goal. Likewise, any GWP criterion
`that
`ignores emissions of associated, energ/-related
`greenhouse gases is likely to exacerbate the problem.
`And, the possibility of yet unforeseen environmental is-
`sues cannot be eliminated. Distinction between short-
`
`and long-lived chemicals increases the options to re-
`spond to future issues, without a long recovery period
`from prior releases.
`
`TRADE-OFFS
`
`In addition to having the desired thermodynamic
`properties, an ideal refiigerant would be nontoxic, non-
`flammable, completely stable inside a system, environ-
`mentally benign even with respect
`to decomposition
`products, and abundantly available or easy to manufac-
`ture.
`It also would be self-lubricating (or at least com-
`patible with common lubricants), compatible with other
`materials used to fabricate and service refrigeration
`systems, easy to handle and detect, and low in cost.
`It
`would not require extreme pressures, either high or low.
`There are additional criteria, but no current refriger-
`ants are ideal even based on the partial
`list. Further-
`more, future discovery of ideal refiigerants is extremely
`unlikely. The discussion that follows illustrates conflicts
`in desired molecular makeup and properties, which vir-
`tually preclude the possibility that ideal refrigerants exist
`or can be synthesized.
`
`Flammability, Toxicity, and Atmospheric
`Lifetime
`
`Increasing the hydrogen content of a compound gen-
`erally decreases its atmospheric lifetime, but increases
`its flammability; the fonner is desirable in a refrigerant,
`while the latter is undesirable. The CFCS, which contain
`no hydrogen, have long atmospheric lives, but are not
`flammable.
`In contrast, the hydrocarbons tend to have
`short lifetimes, but are highly flammable.
`I-ICFCs and
`HFCs fall
`in between. Those with high hydrogen con-
`tent, such as R-l52a, tend to be more flammable (Corr
`et al., 1995). Those with slightly lower hydrogen con-
`tent, R-141b, R-142b, and R-143a, exhibit lower fla m-
`mability. Those with low hydrogen content, such as R-
`22, R-23, R-123, R-I24, R-I25, and R-134a are not
`flammable under normal conditions. Refrigerants gen-
`erally are marginally flammable when the number of
`hydrogen atoms constitutes half of the total atoms con-
`nected to the carbons. Higher fractions become ii-
`creasingly flammable. This observation can be verified
`by examination the lower-flammability limits of refriger-
`
`Page 7 of 16
`
`Trade-0_[}3' in Ref/ivgcrwzl Se/act/'orrs'
`
`l’a.\'l, l’re.s'eHl, and I":/lure
`
`ants (for example, see Richard and Shankland, I992).
`The second numerical digit from the right,
`in standard
`fluorochemical designations for molecules with one to
`four carbon atoms (the methane, ethane, propane, and
`butane series), indicates the hydrogen content. The ac-
`tual hydrogen count is one less than that digit. Using R-
`l34a as an example, the hydrogen atom count is 3 minus
`I equals 2.
`Increasing the chlorine content tends to increase the
`normal boiling point temperature (McLinden and Didion,
`1987).
`Increasing the fluorine content (indicated by the
`right-most digit
`in the Iluoroehemical designation sys-
`tem) tends to reduce toxicity (Clayton, 1967).
`Increas-
`ing the fluorine by displacement of hydrogen tends to
`reduce flammability (Dekleva 1994, and Smith and
`Tufts, 1994), while doing so by displacement of chlorine
`increases the atmospheric lifetime. Comparison of per-
`chlorinated and perfluorinated chemicals, R-10 (CCI4)
`versus R-l4 (CF4) for example, reveals a change from
`42 to 50,000 years.
`Indeed, the perfluorocarbons (R-I4,
`R-I l6, R-218, R-C3l8, R-3l-l0, R-41-12, and others)
`tend to have exceptionally long lives.
`summarized these
`McLinden and Didion (1987)
`trade-offs as shown in figure 5, where the top of the
`triangle represents hydrocarbons (hydrogen and carbon
`only). The two bottom vertices represent perehlorinated
`(chlorine and carbon only) and perfluorinated (fluorine
`and carbon only) chemicals.
`
`ODP versus GWP
`
`There is no way to directly compare the demerits of
`ozone depletion and global warming, since they are dis-
`tinct phenomena. As illustrated in figure 6, increasing
`the chlorine content in refrigerant molecules generally
`increases the ODP. Compounds that contain no bro-
`
`HYDROGEN
`
`/ flammable
`
`CHLORINE
`
`FLUORINE
`
`long atmospheric lifetime
`(fully halogenated)
`
`Figure 5: Trade-offs in flammability, toxicity, and
`atmospheric lifetime with changes in molecular
`chlorine, fluorine, and hydrogen content
`in or-
`ganic refrigerants (McLinden and Didion, I987)
`
`

`
`J. Calm and D. Didion
`
`mine or chlorine have ODPS that are nearly zero.
`Likewise,
`increasing the fluorine count generally in-
`creases the GWP.
`In both cases, increasing the hydro-
`gen count tends to shorten the atmospheric lifetime.
`Compounds with very short lives will have low ODPs,
`since most emissions will decompose before reaching
`the stratosphere. They also will have low GWP values,
`since their atmospheric persistence will be compara-
`tively short in duration.
`As was shown in figure 2, CFCs generally have very
`high ODPs and GWPs. Most HCFCS have low ODP
`and GWP. HFCs have ODPs of almost zero, but
`GWPs that range from very low to very high. Rela-
`tively few fluorochemicals have both very low, or zero,
`ODP and very low GWP. Among them are R-123 and
`R-152a, both of which have short atmospheric lifetimes
`of 1.4 and 1.5 years, respectively. R-152a, however, is
`flammable.
`
`Even though ODP and GWP cannot be equated,
`some conclusions still can be drawn on trade-offs lic-
`
`tween them for specific compounds. One method e-
`quires determination of the chlorine-bromine loading
`(CBL) contribution. CBL is an indicator of the available
`chlorine and bromine reaching the tropopause,
`in turn
`suggestive of the maximal impact on ozone destruction.
`A detailed analysis for R-123 (Calm et al., 1997)
`shows that its use in chillers, at current emission rates
`for convened and new equipment, has a negligible iri-
`pact on the ozone layer.
`Its peak impact, with phaseout
`as scheduled under the Montreal Protocol, amounts to
`approximately 0.002% of the total CBL from all
`sources, natural and anthropogenic. Continued use of
`R-123 as a refrigerant would barely increase the
`0.002% peak, and the average CBL impact
`through
`2050 would be approximately 0.001%. Moreover, the
`contribution that coincides with the CBL peak from re-
`
`ozone depletion depends
`on chlorine (or bromine)
`content and delivery to the
`stratosphere
`
`direct global warming depends
`on infrared absprbence (e.g.,
`by camon-nuonne bonds) and
`longevity
`
`Figure 6: Chlorination and fiuorination impacts on
`ozone depletion and global warming potentials
`
`Page 8 of 16
`
`sidual CFC and halon effects is much smaller. These
`
`fractions are considerably lower than the variability in
`CBL from natural sources.
`
`Conversely, figures 3 and 4 point to a significant cp-
`portunity to reduce global warming by use of R-I 23. As
`discussed above,
`its short
`lifetime suggests a second
`benefit in the event that additional, but currently unfore-
`seen, atmospheric issues surface. R-123's high theo-
`retical and practical efficiencies also lead to immediate
`advantages in energy conservation and reduced use of
`other resources. Trade-off of a negligible impact on
`ozone depletion for significant warming avoidance, and
`other benefits, suggests a strong environmental rationale
`for allowing continued use of R-123.
`
`Importance of Efficiency
`
`As shown in figures 3 and 4 above, there is little cp-
`portunity for further TE,Wl reductions by elimination of
`the direct effect of refrigerant emissions, particularly for
`refrigerants with very low GWP.
`Future decreases,
`therefore, must come from containment, load reductions,
`and efficiency improvements. Figure 4 illustrated the
`substantial progress in equipment tightening for chillers.
`Table 3 summarizes efficiency limits in theoretical ty-
`cles for current and candidate refrigerants.
`Two cautions are warranted. First, the differences
`in efficiency limits for some of the primary candidates
`are small, and may be distorted by imprecision in the
`refrigerant properties used.
`Second,
`theoretical elf-
`ciency limits alone do not govern practical efficiencies.
`Other properties, such as viscosity and thus heat trans-
`fer coefficients, may have significant impacts on overall
`performance. Likewise, cycle design impacts perform-
`ance.
`Increased subcooling may reduce the efficiency
`distinctions in simple cycles, as seen by examination of
`the data presented in table 3. Use