DIMETHYL ETHER AS AN R-12 REPLACEMENT Ben Adamson B.E., M.Eng.Sc., F.IEAust., M.AIRAH, M.ASHRAE Managing Director, Refrigeration Engineering Pty Ltd PO Box 550, Milsons Point NSW 2061, Australia

ABSTRACT Environmental concerns have prompted a search for replacements for CFC and HCFC refrigerants, and more recently, for HFCs. Hydrocarbons generally meet all environmental concerns and offer good drop-in replacements for R-22, R-502 and similar high-pressure refrigerants, although with the drawback of flammability. However, drop-in replacements for R-12 which can be used in large flooded refrigeration systems, without extensive system modification or cleanout are uncommon. Zeotropic blends are unsuitable, due to fractionation of components. This paper proposes dimethyl ether as a single-component, drop-in replacement for R-12 in flooded systems. Its saturation pressure/temperature relationship is very close to R-12, heat transfer properties are significantly better and energy costs are slightly lower. Compressor capacity is close to R-12 capacity at medium temperatures, but slightly lower than R-12 capacity at low temperature. It is non-toxic in normal usage, low cost, widely available and environmentally safe, although flammable. 1. INTRODUCTION The worldwide concerns about ozone depletion and global warming have thrown up a large number of new refrigerants being promoted as replacements for CFCs and HCFCs, and more recently for HFCs. Most of these new refrigerants address ozone depletion concerns, but many, such as HFCs, are still significant contributors to global warming. In addition, many also require extensive system modifications or cleaning in case of retrofit into existing systems. Hydrocarbons can be used as alternatives to CFCs, as they possess many desirable characteristics as refrigerants. Amongst these are zero ozone depletion potential (ODP) and small or zero global warming potential (GWP), as well as compatibility with common mineral oils. Due to flammability concerns, hydrocarbons are unlikely to enter widespread service in systems where the refrigerant charge exceeds a small amount - perhaps 5-50 kg. However, in applications where flammability is not a concern, such as in oil and gas process installations, hydrocarbons have been used for many years, and will continue to be used. Propane, and less commonly propylene, exist as good hydrocarbon drop-in replacements for R-22, R-502 and refrigerants operating at similar pressures. They have the advantage of being cheap (about US$1/kg or less), widely available, miscible and compatible with commonly used oils, and being single-component refrigerants, completely suitable for use in large flooded systems. However, easy hydrocarbon replacements for R-12 have been less common. Propane is generally not suitable as a drop-in replacement, as condensing pressure is typically about 400 kPa higher than R-12 (and propylene is similar). In addition, simply recharging an R-12 system with propane leads to a compressor capacity and power increase in the region of 40% over the same system charged with R-12. This usually then requires compressor and/or motor replacement, and other system modifications. Butane pressures are significantly lower than R-12, but system capacities are also much lower when R-12 is replaced with butane. Some butane/propane blends are being promoted as R-12 replacements, but these are generally zeotropic, and although suitable for small direct expansion systems, their zeotropic nature rapidly leads to fractionation within flooded IIR - Commissions B & E - Oslo, Norway - 1998

Adamson, B.M. - Dimethyl ether ... Page 1 of 7

systems. In the flooded evaporator the propane evaporates more readily, and the evaporator ends up rich in butane and operating at relatively low pressure. As a consequence the condenser ends up rich in propane and operating substantially at propane pressure. Such operation leads to poor system performance. Large industrial R-12 systems, typically of capacity 300 kW and above, are mainly flooded systems, and are therefore unsuited to zeotropic refrigerants as drop-in replacements. The hydrocarbon dimethyl ether (DME, C2H6O) possesses a range of desira0ble properties as a replacement for R-12. These include better heat transfer characteristics than R-12, a pressure/temperature relationship very close to R-12, compatibility with mineral oils, compatibility with residual R-12 contamination, low cost and ready availability. It is also highly environmentally friendly. Being free of halogens, its ODP is zero, and using a basis of GWP for R-12 = 1.0 the GWP of DME is 0.0024 (Propel), principally due to its very short atmospheric lifetime, only 6 days. For comparison purposes, on the same basis the GWP for CO2 is 0.00005. DME occurs naturally in small quantities in natural gas. However, worldwide demand outstrips natural supply, and so it is made commercially in many locations worldwide, normally from methanol feedstock. Its main use is as a propellant in aerosols, where it has been widely adopted as a replacement for R-12 and other CFCs. It is also manufactured as an intermediate component in some petrochemical processes. Others (DEA, 5) have considered DME for use in small systems such as domestic refrigerators. This paper looks at DME as an alternative or replacement for R-12 in large systems. 2. PRESSURE/TEMPERATURE PROPERTIES OF DIMETHYL ETHER DME is a colourless, virtually odourless, stable compound, with low viscosity and low surface tension. It has high solvency power with both polar and non-polar compounds (DEA, 4). The pressure-temperature relationship at saturation is very close to R-12, as shown in Table 1.

Temperature °C -40 -20 0 20 40 60

Pressure R-12 (sat.) kPa(a) (2) 64 151 309 567 959 1521

Pressure DME (sat.) kPa(a) 49 124 267 512 896 1457

Table 1. Saturation pressure/temperature relationships of R-12 (ASHRAE, 1986) and DME

For DME in table 1, pressure was calculated from log10 P = 4.1651- 913.40/(T+244.29) where P = pressure, bar abs., T = temperature, °C (DEA, 4)

(Eq. 1)

The error between pressure calculated using eq. (1) and measured pressure is less than 0.5% over the range -20°C to +60 °C (DEA, 4).

IIR - Commissions B & E - Oslo, Norway - 1998

Adamson, B.M. - Dimethyl ether ... Page 2 of 7

3. REFRIGERATION PROPERTIES OF DME The performance of DME as a refrigerant can be considered in two parts - its performance in relation to compressors, and its performance in relation to heat exchangers. Basic physical properties of interest in these areas are listed in Table 2, for saturated liquid and vapour at 30°C .

Density, kg/m3 Specific heat (Cp), kJ/kg.K Thermal conductivity, W/m.K Viscosity, centipoise Latent heat of vaporisation, kJ/kg

R-12 liquid 1292.5 0.984 0.068 0.254 135.9

R-12 vapour 42.15 0.72 0.010 0.055

DME liquid 654.3 2.47 0.147 0.125 415.7

DME vapour 1.86 1.47 0.017 0.0096

Table 2. Saturation properties of R-12 (ASHRAE, 1976) and DME (PPDS) at 30°C.

3.1 Compressor capacity In looking at compressor performance in a retrofit situation, there are two main quantities of interest - refrigeration capacity and drawn power, and perhaps the combination of these commonly referred to as coefficient of performance or COP. In large industrial systems such as this paper is directed towards, the most commonly used compressors are positive displacement type (reciprocating or screw.) Centrifugal compressors are much less commonly used. Where they are used, change of refrigerant from R-12 to either R-134a or DME is likely to require a significant modification or replacement of the compressor. Pressure rise across a centrifugal compressor operating at a given speed is directly related to molecular weight of the vapour being compressed. Molecular weight of R-12 is 121, of R-134a is 102 and of DME is 46. To achieve a comparable “lift” on R-134a with a centrifugal compressor designed for R-12 will require a speed increase, but this may be possible relatively simply by a gear change on many integrally-geared compressors. Reusing an R-12 centrifugal compressor on DME would generally not be feasible by speed change - the speed increase would usually be beyond the R-12 machine’s design limits. Either addition of stages or complete replacement would be required. Due to the relative rarity of centrifugal compressors in this application, centrifugal compressors will not be considered further in this paper. The volumetric efficiency of a reciprocating compressor operating with different gases at similar compression ratios will be very similar. A DME system operating at -30°C evaporating and +40°C condensing will have a compression ratio of 11.2, compared to 9.6 for R-12 at the same conditions. At 0°C/40°C the compression ratio is 3.35 for DME and 3.11 for R-12. The volumetric efficiency for a given compressor on DME will therefore be similar to, but slightly less than for R-12. Theoretical capacity and isentropic power for a reciprocating compressor, calculated from refrigerant properties and typical compressor volumetric efficiency data, is shown in table 3. Where compression ratios are similar, at higher evaporating temperatures, compressor volumetric efficiencies are also similar and the properties of DME give slightly higher theoretical capacity and slightly better theoretical COP than R-12. At lower evaporating temperatures the lower pressure of DME begins to outweigh its apparently more favourable properties and capacity of the compressor on DME is less than on R-12. However, at the lower temperature theoretical COP for DME is slightly better than R-12. These conclusions at the lower temperature are supported by tests on IIR - Commissions B & E - Oslo, Norway - 1998

Adamson, B.M. - Dimethyl ether ... Page 3 of 7

domestic refrigerators and freezers (DEA, 5) which reported lower compressor capacity on DME than on R-12, but lower overall energy usage. Evaporating temperatures for these tests were approximately -10°C to -30°C. The tests also experimented with two blends of R-12 and DME, as well as tests on pure DME and pure R-12. In all cases the oil used was the original oil supplied in the units from the manufacturer, when they were originally charged with R-12. These tests, particularly the tests with DME/R-12 blends, show that residual traces of R-12 contamination should not be a problem in applications changing R-12 systems to DME. However, further tests in this area are desirable. Solubility and adsorption of refrigerant in the oil can have a significant effect on compressor performance through volumetric efficiency, particularly in screw compressors where large quantities of oil are injected into the compressor. DME has a high solubility and miscibility in oils, and DME vapour is expected to be readily absorbed into oils, although firm data on this is not yet available. However, as vapours of R-12, propane and similar refrigerants are readily absorbed and soluble in oils, the performance of DME in this regard is expected to be similar. Detailed performance data on screw compressors with DME is not currently available, but a large existing R-12 system in Australia (R-12 charge approximately 10 tonnes) is to be changed to DME, and the compressor changed from centrifugal to screw, and commissioned late in 1998. It is expected that this project will provide some data on this subject.

Evaporating temperature Condensing temperature Compression ratio Volumetric efficiency Refrigeration capacity, kW Isentropic power, kW Isentropic C.O.P.

DME 0°C +40°C 3.35 0.810 320 56.7 5.64

R-12 0°C +40°C 3.11 0.820 330 61.5 5.42

Evaporating temperature Condensing temperature Compression ratio Volumetric efficiency Refrigeration capacity, kW Isentropic power, kW Isentropic C.O.P.

DME -30°C +40°C 11.2 0.445 51.4 20.1 2.55

R-12 -30°C +40°C 9.6 0.515 64.0 25.8 2.48

No liquid subcooling, no suction superheat, no line losses Compressor swept displacement 200 L/s, volumetric efficiency from Ref. 7 Table 3 Compressor performance comparison

3.2 Heat exchangers Several heat exchangers were modelled using recognised exchanger design software (HTFS) and drawing refrigerant properties from the database accompanying the software. Identical exchanger designs were used for both DME and R-12, taken from exchangers used in an existing R-12 system,

IIR - Commissions B & E - Oslo, Norway - 1998

Adamson, B.M. - Dimethyl ether ... Page 4 of 7

with no attempt to optimise or otherwise change the R-12 exchanger designs for DME. Exchanger performance is summarised in Table 4. Actual printouts can be supplied by the author on request.

Exchanger E-63 TEMA BKU type, flooded R-12 or DME boiling in shell at -43°C, hydrocarbons condensing in tubes between -25°C and -40°C. R-12 boiling film coefficient 1406 W/m.K DME boiling film coefficient 2847 W/m.K Exchanger E-65 TEMA BXT type, flooded R-12 or DME boiling in shell at -42°C, hydrocarbons condensing in tubes between -6°C and -36°C. R-12 boiling film coefficient 842 W/m.K DME boiling film coefficient 1261 W/m.K Exchanger E-71 Condenser, TEMA AEM type, condensing R-12 or DME in shell on low-fin tubes at 39°C, water in tubes entering at 24°C and leaving at 30°C. R-12 condensing film coefficient 4419 W/m.K DME condensing film coefficient 6789 W/m.K Exchanger E-72 Shell & tube subcooler, TEMA BKU type, flooded R-12 or DME boiling in shell at -26°C, liquid R-12 or DME in tubes cooled from -7°C to -20°C. R-12 boiling film coefficient 808 W/m2.K R-12 convective film coefficient 1260 W/m2.K at 1.09 m/s DME boiling film coefficient 1610 W/m2.K DME convective film coefficient 1934 W/m2.K at 0.96 m/s Table 4. Heat exchanger performance with R-12 and DME

It can be seen that in all cases - boiling, condensing and convective - DME film coefficients are in the range 50-100% higher than for R-12 in exchangers originally designed for R-12.

3.3 Material compatibility DME is compatible with most materials commonly found in refrigeration systems. It is suitable for use with ferrous metals, copper and copper-based alloys, and aluminium. It is also suitable for prolonged exposure to PTFE (teflon), viton, EPDM, buna-N, polypropylene and polyurethane.

IIR - Commissions B & E - Oslo, Norway - 1998

Adamson, B.M. - Dimethyl ether ... Page 5 of 7

Compatibility with other elastomers is covered in literature available from manufacturers of DME and some equipment (DEA, 4; Aerofako; Haskel). 4. SAFETY ASPECTS There are three primary safety aspects for consideration in relation to any refrigerant. Two of these relate to safety of people in the immediate vicinity - flammability and toxicity - and the other relates to safety of people worldwide - environmental safety, which was covered in Section 1. 4.1 Flammability DME, being a hydrocarbon, is flammable. Its flammable limits are between 3.4% and 27%. (NFPA, 1994) Its hazard level in this regard is little different from propane, which is widely used as a refrigerant in the petrochemical, oil and gas industries. The use of propane is now (with appropriate safety precautions) increasing in air conditioning and non-industrial refrigeration applications, particularly in small systems. The range of use of DME outside the petrochemical, oil and gas industries is likely to be acceptable in similar applications to propane. 4.2 Toxicity Due to its widespread use as an aerosol propellant, extensive testing has been done on effects of DME on humans. The most common form of ingestion of most refrigerants (and aerosols) is inhalation. DME is considered to have a low order of acute, subacute and subchronic inhalation toxicity, and no adverse effects from contact with the vapour on the skin. (DEA, 4.) Normal refrigerant handling safety precautions should be observed when handling the liquid to prevent injuries due to freezing. 5. CONCLUSION DME has a range of physical and thermodynamic properties favourable for use as a drop-in replacement for R-12 in flooded refrigeration systems where flammable refrigerants are acceptable. System modifications and cleanout required for changing R-12 systems to DME are minimal, and environmental and toxicity risks are very low compared to CFCs, HCFCs and HFCs. Tests in small systems have produced results which support theoretical performance predictions, and further tests in large systems are desirable. REFERENCES 1.

Aerofako BV, Netherlands. DME product data.

2.

ASHRAE - Thermodynamic properties of refrigerants (publication THPRSI), 1986

3.

ASHRAE - Thermophysical properties of refrigerants, 1976

4.

DEA Mineraloel AG, Wesseling, Germany. Dimethyl ether product information

5.

DEA Mineraloel AG, Wesseling, Germany. Research on household refrigerators and deep freezers using DME and R-12 (unpublished).

6.

Haskel International Inc. USA. Fluid compatibility guide.

7.

National Fire Protection Association, USA. Fire hazard properties of flammable liquids, gases and volatile solids. NFPA publication 325, 1994

IIR - Commissions B & E - Oslo, Norway - 1998

Adamson, B.M. - Dimethyl ether ... Page 6 of 7

8.

Physical Property Data Service. Physical property database PPDS2 V1.1

9.

Private communication with compressor manufacturer, confidentiality requested by source.

10.

Propel (Aerosols Australia Pty Ltd) - Dimethyl ether product information

11.

UK Atomic Energy Authority - Heat Transfer & Fluid Flow Service (HTFS) M-TASC V3.03 Shell & tube exchanger software and DIPPR V1.0 database

IIR - Commissions B & E - Oslo, Norway - 1998

Adamson, B.M. - Dimethyl ether ... Page 7 of 7