Mekong Forum Essays
Make your own free website on


Seminar on Environment and Development in Vietnam
Friday and Saturday, December 6-7, 1996. Common Room, University House, and J G Crawford Building, National Centre for Development Studies, Australian National University

Natural Replacements For Ozone-Depleting Refrigerants In Eastern And Southern Asia

Emma Aisbett and Tuan Pham School of Chemical Engineering and Industrial Chemistry University of New South Wales Tel: +61 (2) 9385-5267 Fax: +61 (2) 9385-5966 Email: Web:


A computer model has been written to predict the consumption of refrigerants for vehicle air conditioning in China, India, South Korea and South-East Asia, their effect on ozone depletion and global warming, and their costs. Both synthetic HFC (134a) and natural (hydrocarbons) refrigerants are considered. The use of hydrocarbons will lead to significant reductions in global warming potential and large savings in cost. The synthetic HFC option will incur costs exceeding a billion US dollars per year after the year 2005.


For nearly sixty years chlorofluorocarbon refrigerants (CFCs) have been used for refrigeration, foam blowing and spray cans. Their harmful effect on the Earth's protective ozone layer was first published by Rowland and Molina in 1974. Subsequently the ozone hole over the Antarctic was discovered and heightened world attention led to the Montreal Protocol of 1989.

Over the same period, it was discovered that CFCs also contributed significantly to the world's greenhouse warming problem. The global warming potential of CFC-12 is 8500 times that of carbon dioxide over one hundred years. The global warming implications of CFC substitutes is now a major issue.

The Montreal Protocol and subsequent, increasingly stricter versions stipulates that production of the main CFCs such as R11 and R12 are banned by 1996. For developing nations (those that use less than 0.3 kg per capita in 1989), a ten-year grace period is granted and special help is provided to ease the transition period.

The response of the chemical industry to the ban has been to develop new synthetic refrigerants without the offending chlorine atoms which cause ozone depletion. These are known as hydrofluorocarbons, or HFCs, of which the best known is 134a. The manufacture of these HFC are a high-tech process and consequently their cost is high, which would cause a burden on the poorer countries. Furthermore, their greenhouse warming potential is also significant, though less than that of CFCs (Table 1)

,td> No
Table 1: Environmental effect of refrigerants
Substance Natural ODP GWP 100yr GWP 20yr
CFC-12 No 1.0 8500 8500
HCFC-22 No 0.055 1700 4100
HFC-134a 0 1300 3100
HFC-143a No 0 4400
Isobutane Yes 0 3
Propane Yes 0 3 0
Ammonia Yes 0 0 0

In the last few years, interest is gathering over the use of naturally occurring hydrocarbons such as propane and butane as refrigerants . These substances are cheap and plentiful (they are components of natural gas and are already used as fuel), and furthermore they have zero ozone depletion potential and near zero greenhouse effect (see table). Their flammability has caused some concern, but all tests done so far indicate that they are quite safe in small applications such as domestic fridges and car air conditioners, due to the very small amounts involved (James & Missenden, 1991, Maclaine-cross, 1993).

The objective of this work is to quantify the environmental and economic implications of different CFC substitutes using relationships between income, population growth rate, equipment ownership rate to make projections for the next 35 years. Only vehicle air conditioning has been considered; this is known to be the major use for refrigerants, accounting for half of all CFCs used in the developed countries (UNEP, 1992).

The countries chosen for this survey belong mainly to East and South-East Asia, a populous and rapidly developing region which is of immense importance to the future of Australia and the world. India is also included due to its proximity to the region and its large population. Thus the countries considered consist of China, India, Indonesia, Korea, Thailand, Malaysia, Vietnam, the Philippines, Singapore, Myanmar, Laos, Papua New Guinea and Brunei. The combined population of these countries is 2.5 billion or about half the world's population, with China (1.2 billion) and India (0.9 billion) making up most of the total.


The method used to predict the effects of using natural and synthetic refrigerant consists of the following steps: (a) constructing a model to predict economic and population growth for the countries to be studied, (b) finding a correlation between per capita income and ownership rate and using it to predict the total number of units in operation, (c) calculating the refrigerant consumption rate and losses from a yearly material balance, taking replacement of obsolete units, introduction of new units and refrigerant leakage into account, and (e) calculate the resulting economic and environmental effects for different refrigerant replacement policies.

  1. Model for economic and population growth It is commonly accepted that population growth rate changes with economic development. This belief is termed "transition theory" and states that during economic development population growth will first increase rapidly and then decrease to be around zero (Stanford, 1972; Kasun, 1988). The initial increase is due to the falling death rate while the subsequent decrease results from a falling birth rate, attributable to numerous social and economic factors such as increased education levels and female economic activity.
    Many other political, historical and cultural variables can also intervene, resulting in large variations. War, religion, political intervention (as in China) will cause positive or negative deviations from the simple transition theory model. Perhaps for this reason economic texts do not present any quantitative correlation for population growth rate vs economic development. Apart from China (which has a very stringent one-child policy), the countries considered share fairly similar cultural characteristics in that there are no strong religious influences and there has been no major war for the last twenty years, so it can be expected that the transition theory will be applicable.
    Figure 1 depicts the correlation between population growth rate (on a logarithmic scale) and GNP per capita for the countries of this study, with the exception of China due to its on-child policy. Japan is included to enable extrapolation to the higher incomes which will occur in the future. A linear regression of ln (growth) against GNP per capita is also plotted, with an R2-value of 0.84. The regression equation is

        ln (rG) = 8.80E-5 i + 0.887

    where rG is the population growth rate (% per annum) and i the GNP per capita. If the present GNP per capita i0 is known and rates of increase in total GNP rI for future years are assumed, then the GNP and population can be calculated for every future year, using an iterative calculation. Deviations from the regression line which are attributable to non economic factors are taken into account by assuming that the slope of the line is common to all countries, but the intercept is shifted up or down to fit each individual country.

    Figure 1: Relationship between population growth and per capita income

  2. Correlation between ownership rate and per capita income

    Projections of future vehicle air conditioning ownership are based on the rate of ownership of car and commercial vehicles and the fraction that have air conditioners.

    Car ownership is based on current passenger car ownership worldwide. Car ownership plotted against income follows an S curve, with the ownership rate being negligible in very poor countries, increasing rapidly once the average income is a substantial fraction of car prices, and levelling out as saturation is reached. Figure 2 shows the relationship worldwide while Figure 3 shows the low income countries. A broken line relationship is used to represent the data. There is a large scatter, particularly at intermediate incomes, which indicates that the transition to the rapid growth period is reached at different incomes for different countries. As far as the predictions of this study are concerned this means that underestimates in one five-year period will be compensated by overestimates in later years and vice versa, however long-term projections should not be affected.

    Figure 2

    Figure 3
    Commercial vehicle numbers was derived from the number of passenger vehicles. This relationship is based on data for the ESCAP region (ESCAP, 1995) and plotted on Figure 4. The ratio of commercial vehicles to passenger cars ranged from 0.55 to 0.61, and a constant value of 0.585 is assumed for this work.

    Figure 4

    The proportion of vehicles with air conditioning is assumed to vary with per capita income. It can be found by linear interpolation of the data in Table 2. The values in Table 2 are based on data for the world as a whole and North America, with consideration given to the generally hotter climes of Asia. The ownership of vehicle air conditioners at various levels of income is then given by the total number of vehicles, passenger plus commercial, multiplied by the fraction with air conditioning.

    Table 2. Vehicle air conditioning ownership vs income
    GNP per capita Passenger car per capita % cars with condit. Car air cond.
    < 200 0.00 NA 0.0000 0.0000
    1000 0.02 50 0.0075 0.0119
    3500 0.15 70 0.1050 0.1664
    14000 0.42 90 0.3780 0.5991
    32000 0.42 97 0.4074 0.6457

    The total number of vehicle air conditioning units in operation in a given year is predicted by multiplying the forecast population from (a) by the rate of vehicle air conditioning (VAC) ownership at the forecast per capita income found from (b).
    (c) Refrigerant consumption rate and losses
    For a given year, the net increase I in the number of units is calculated from the predicted total number for the present year minus that for the previous year, both found from (b). The number of new units, C, is equal to this net increase plus the number of obsolete units replaced, O:

       C = I + O

    The number O is a function of the age distribution of the units in operation. We made the simplifying assumptions that all units fail after exactly the average lifetime L (Kroeze, 1995); the consequence of this assumption is not expected to be serious since over time the errors will largely cancel out; so the number of units O that fail in the year Y is equal to the number of units CY-L introduced in the year Y-L. To calculated CY-L, however, requires that the number of obsolescences OY-L in the year Y-L be known; to avoid recursion, we make the further assumption that OY-L is equal to the number of units in use the previous year (Y-L-1), divided by the average lifetime (Greenpeace Japan, 1993).

    Assuming no recycling is done, the total amount of refrigerant consumed will be the sum of refrigerant charge required for new equipment and refrigerant leakage for existing equipment. For new equipment, the refrigerant required is obtained by multiplying the number of new equipment units by the charge of refrigerant per unit. It is assumed that the liquid volume of refrigerant per unit is a constant, so that refrigerants with smaller liquid density will require less mass charge. For existing equipment, a leakage rate of 40% per year is assumed based on combined data from Greenpeace and UNEP. The total amount released into the atmosphere is the sum of leakage from operating equipment and obsolete equipment's charges. If a mixture of refrigerants is used, the calculation must be carried out for each individual refrigerant.

  3. Economic and Environmental Effects

    The economic and environmental effects considered were refrigerant supply cost, direct Global Warming Impact (GWI), and Total Equivalent Warming Impact (TEWI). All these are found from the amount of each refrigerant consumed and released as calculated in (c).

    Direct global warming impact (GWI) is the direct effect of released refrigerant emission in equivalent tonnes of carbon dioxide. The GWI is an annual figure obtained by multiplying annual emissions by the reported Global Warming Potential (GWP) of the species.

    Total equivalent warming impact per annum (TEWI) is the sum of the GWI and the carbon dioxide emitted in the production of energy to run the VACs. Carbon dioxide emissions to run VACs are calculated based on the efficiency of the vehicle, the efficiency of the air conditioner and the number of VACs in operation. The efficiency of the air conditioner will depend on the refrigerant used. Note that the conventional definition of TEWI measures the warming impact of an item over its whole life, but in this study it is defined as the warming impact of all items in operation over a single year.

Population and GNP (1987 $US) data from 1980 to 1994 were the base to which the previously described equations were applied. The only future input required by the combined model was the rate of growth in GNP. This really belongs to the realm of speculative futurology and there is no scientific way of making long term or even medium term projections. Pessimists argue that there is a limit to the world's resource and growth will eventually be replaced by catastrophic crashes, while optimists will argue that the region's economies will continue to growth at least until the average GNP of present day's industrial nations is reached. Political stability and government policies will also be important factors. Since we have no model to follow, the assumption will be made that present rates of GNP growth will hold for another ten years (up to 2005), then continue at an average of 6% per annum for all countries.

Since the aim of the project was to quantitatively compare use of natural and synthetic refrigerants, four scenarios were calculated. The first was a hypothetical `no change' scenario were CFC-12 is assumed to be the sole refrigerant in use during the entire period. The subsequent scenarios assume a transition away from CFC-12 for new equipment, beginning in 1990 and completed in 2000. Thus the second scenario assumes 100% of the CFC replacement market is supplied by HFC-134a; the third assumes 50% of the replacement market goes to each of HFC-134a and hydrocarbons; and the forth assumes 100% of the market is supplied by hydrocarbon refrigerants.

Predictions of refrigerant consumption, refrigerant supply cost, GWI, and TEWI were made for every fifth year from 1985 to 2030 inclusive. Using the "base" GNP growth rate (each country continuing at the present rate for ten years, then at 6% per annum after that), the number of VAC units and amount of refrigerant consumed in the region (in CFC equivalent) is expected to increase about 20 times between 1995 and 2030 (Table 3), an average growth rate of 9% per annum. About three quarters of refrigerant goes into replenishing existing units, the rest being for new units. As an indication of the magnitudes involved, the total consumption of the countries considered is expected to reach today's figure for the U.S.A. between 2010 and 2015.

Table 3: Consumption of units and refrigerant, all countries

,td> 0.76
Year 1985 1990 1995 2000 2005 2010 2015 2020 2025 2030
Million VAC units 9.6 17.9 29.9 48.3 94.4 149.3 216.7 316.5 453.2 617.9
000 t. 4.9 8.9 14.8 24.3 49.1 73.0 108.1 159.6 222.9 305.2
Refrigerant Service 0.74,/td> 0.75 0.74 0.76 0.76 0.73 0.76 0.73 0.76

Table 4: Environmental effects, all countries EFFECTS 12.9 135.3 195.9 268.1 HFC 3.7 7.1 8.5 6.4 3.3 0.0 0.0 0.0 0.0 0.0 HC 3.7 7.1 8.5 6.4 3.3 0.0 0.0 0.0 0.0 0.0 GWP, million tonnes p.a. CFC 31.2 60.5 101.9 169.9 310.4 517.2 789.7 1150.2 1665.3 2278.4 HFC 31.2 60.2 76.5 71.0 67.7 72.8 111.2 161.9 234.4 320.7 HC 31.2 60.1 72.3 54.8 28.0 0.1 0.1 0.2 0.2 0.3 TEWI, million tonnes p.a. CFC 56.7 108.1 181.5 298.3 561.5 914.5 1366.2 1992.3 2871.3 3922.4 HFC 56.7 107.8 158.5 207.0 333.7 493.7 721.8 1053.9 1511.7 2062.1 HC 56.7 107.7 147.9 170.4 254.0 357.7 519.0 758.1 1085.6 1479.9 COST, million $US p.a. CFC 29.2 53.7 89.0 145.8 294.7 438.3 648.7 957.6 1337.3 1831.4 HFC 29.2 54.7 177.6 471.4 1062.1 1613.3 2388.0 3525.2 4923.0 6741.7 HC 29.2 53.3 57.2 29.3 20.1 17.7 26.2 38.7 54.1 74.0 Figure 5 shows the number of VACs in use and for China, India, and the total of other countries. The table shows clearly the importance of the two large countries. Figure 5: Number of vehicle air conditioning units (millions) The refrigerant supply cost is nearly one hundred times higher when using HFC-134a compared to hydrocarbons. This is the result of larger refrigerant charge mass required combined with a far higher unit price. Supplying HFC-134a would cost the region several billion dollars per year more than supplying hydrocarbons. Perhaps more importantly, this cost would be in foreign dollars as HFC production is complex and not practiced in many developing countries. In comparison, plant for the production of high quality hydrocarbon refrigerant is relatively cheap, and can be viable for small scale production. The higher TEWI of HFC-134a can be seen to be largely due to its significantly higher GWI. The cumulative difference in TEWI between the 100% HFC-134a and 100% hydrocarbon scenarios for the region as a whole is 8.9 billion tonnes by 2030. This is roughly equivalent to England's total CO2 emissions over 7 years. Recycling The effect of recycling all CFCs and HFCs from 1995 onward is simulated. While this may be considered unrealistic for this area, intermediate scenarios can be deduced by interpolation. Recycling CFC will eliminate ODS emission about five years earlier. Recycling HFC will have no effect on ODP, but will reduce GWP by about 15-20%, and will lead to a major savings in cost of manufacture. Table 5: Environmental effect with total recycling of CFCs and HFCs. Year 2000 2005 2010 2015 2020 2025 2030 ODP, thousand tonnes p.a. HFC or HC 4.1 0.6 0.0 0.0 0.0 0.0 0.0 Decrease 2.3 2.6 0.0 0.0 0.0 0.0 0.0 GWP, million tonnes p.a. HFC 51 44 67 96 137 203 272 Decrease 20 24 6 16 25 32 49 TEWI, million tonnes p.a. HFC 187 310 488 706 1029 1480 2013 Decrease 20 24 6 16 25 32 49 COST, million $US p.a. HFC 464 1043 1560 2244 3298 4629 6290 Decrease 7 19 54 144 228 294 452 CONCLUSIONS Many assumptions have been made in this work, in particular the relationships between population growth and per capita GNP and that between ownership rates of cars and per capita GNP. The latter is not expected to have big effects on the model: the S-shape of the relationship, where a country transitions from a low-ownership society to a high-ownership one within a few years, is a widely observed phenomenon for many kinds of goods. Thus any uncertainty will only cause a small shift - probably about five years or less - in the time scale. The former relationship (between population growth and per capita GNP) depends critically on India and China, which made up more than 80% of the total. Since the growth rate of China depends on the one-child policy, future political events there will play a critical role. The biggest uncertainty in these scenarios is the projected economic growth. It is virtually impossible to predict what will happen in the next 30 years: will growth continue at the same rate, or will environmental factors and finite physical resources eventually prevail and impose a limit on growth? Will there be a war or other cataclysmic political event? It is not the main purpose of this work to make precise predictions, but to work out different the consequences of different scenarios. To that effect the computer model that we have written will be a very useful tool. Only vehicle air conditioning has been considered so far by us. For developed countries it is the biggest source of refrigerant consumption, making up to 50% of the total. However, for developing countries other sources are likely to be much more important, since an (air conditioned) car is about the last possession that the average person can afford. Domestic refrigeration, air conditioning of dwellings, offices and shops, commercial and industrial refrigeration (especially food processing and storage) are likely to be far greater refrigerant consumers during the early years of development. These will be the next areas of studies. We also intend to include the effect of recycling in a future version. REFERENCES Economic and Social Commission for Asia & the Pacific (ESCAP), Statistical Yearbook for Asia and the Pacific, United Nations (1995). Greenpeace Japan, CFC Substitutes will not Save the Earth: the environmental impacts of refrigerants and insulation foaming agents for household refrigerators in the developing countries. Greenpeace Japan, Tokyo (1993). Kasun, J., The War Against Population: the economics and ideology of population control, Ignatius Press, San Franscisco (1988). Kroeze, C., Flourocarbons and SF6: Global emission inventory and options for control, RIVM National Institute of Public Health and Environmental Protection, Bilthoven, Netherlands (1995). Maclaine-cross, I., Fireball: A brief report on pilot experiments to measure the insurance risk of hydrocarbon refrigerants in motor cars. School of Mechanical and Manufacturing Engineering, UNSW, Sydney (1993). Missenden, J., James, R., Wong, A., Propane for Systems with Small Refrigerant Charge. Anglo-Swedish Conference, Sophia Antipolis, France (1990). Molina, M., Rowland, F., Stratospheric Sink for Chlorofluoromethanes: CHlorine-Atom Catalyzed Destruction of Ozone." Nature, no. 249 (1974). Stanford, Q., The World's Population: problems of growth, Oxford University Press, Toronto (1972). UNEP IE/PAC, Protecting the Ozone Layer: Vol 1. Refrigerants. United Nations, Oxford (1992). APPENDIX: VIETNAM SCENARIOS The main part of this paper considered the region as a whole and thus have to make simplifying assumptions for individual countries. For Vietnam, provided that deep structural problems are overcome (admittedly a big proviso), it might be not over-optimistic to assume a mean growth rate of 10% for the whole period considered (1995-2035), in view of the experience of other East Asian countries. Tables A1 to A3 show the resulting projections: Table A1: Projected population and GNP growths (Vietnam) Year 1985 1990 1995 2000 2005 2010 2015 2020 2025 2030 GNP/cap, US$ 171 188 251 372 552 821 1227 1844 2794 4283 GNP, US$billion 10.1 12.4 18.5 30.5 50.3 82.9 136.6 225.3 371.4 612.3 Population, m 59 66 74 82 91 101 111 122 133 143 Population 2.21 2.20 2.19 2.16 2.13 2.07 1.99 1.87 1.70 1.47 growth GNP growth 0.02 0.04 0.08 0.10 0.10 0.10 0.10 0.10 0.10 0.10 factor Table A2: Refrigerant consumption (Vietnam) Year 1985 1990 1995 2000 2005 2010 2015 2020 2025 2030 Million VAC 0.0 0.0 0.1 0.2 0.5 0.9 2.9 7.8 16.3 28.4 units 000 t. 0.0 0.0 0.0 0.1 0.2 0.5 1.6 4.0 8.0 14.0 refrigerant Service/total 0.80 0.80 0.49 0.64 0.68 0.69 0.53 0.66 0.71 0.74 Table A3: Environmental effects (Vietnam) Year 1985 1990 1995 2000 2005 2010 2015 2020 2025 2030 ODP, thousand tonnes p.a. CFC 0.0 0.0 0.0 0.1 0.2 0.4 0.9 2.8 5.9 11.6 HFC 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 H
Year 1985 1990 1995 2000 2005 2010 2015 2020 2025 2030
ODP, thousand tonnes p.a.