Jaguar Clubs of North AmericaJCNA Home

Cooling Faq, Part 2: Coolants- Better Living Through Chemistry
by Michael Frank, Jaguar Touring Club

Home   Tech Index

It may come as a surprise that the theoretical maximum thermal efficiency of a gasoline motor is less than 60%. In practice, the efficiency of a real motor is much lower. It sounds grim, but your engine is really among the most efficient means of producing power. Compare this to photoelectric cells, which are about 7% efficient.

If only a fraction of the energy in your fuel is actually producing useful work, where does the rest go? Some goes out the tailpipe in the form of heat, noise and vibration, some is handled by the oil and lubrication system, and some goes to pumping loss internal to the engine. But most waste energy is handled by the cooling system.

The working fluid in your radiator is probably a combination of water and glycol antifreeze. In this FAQ, we will explore some of the chemistry involved in coolants. We will describe the physical properties of liquids that make them good heat transfer media, and the relative advantages and disadvantages of various mixtures.

Physical Properties of Liquids

Liquids have a few properties that need to be understood before we can proceed. Chemistry isn't everyone's favorite subject, so well keep it as simple as possible:

Specific Heat
Ethylene Glycol
Propylene Glycol
Latent Heat of Vaporization
Ethylene Glycol
195 cal/gm
Propylene Glycol
170 cal/gm
540 cal/gm

Boiling Point @ Sea Level
Ethylene Glycol 387
Propylene Glycol 370
Water 212

Stovetop Experiments

The standard old saw among mechanics is "water cools best". This even has some scientific basis, since water has the highest specific heat and latent heat among the available coolants. But on the other hand, water has the highest surface tension, lowest specific gravity,  and lowest boiling point, which would suggest that there may be problems with pure water as a coolant. I decided to conduct some controlled experiments to determine the facts.

Experimental design. The experiment was structured as follows. For each coolant mixture tested, a fixed volume of coolant was used (one pint). The stovetop was thermostatically controlled to 900F, and was allowed to reach its full operating temperature prior to each trial. The test coolant was placed in an aluminum pie tin, at 70F room temperature. The time required to heat the fluid to boil was observed.

BTU output of the stove was calculated based on  measuring the time required to boil a one pound sample of water, mixed with surfactant. This calculation is almost certainly incorrect, but since we are mainly interested in relative results, it's OK..

Since significant nucleate boiling occurs as soon as the pan comes in contact with the stovetop, a true boiling condition was determined when the temperature of the coolant reached its limiting temperature, i.e., when the entire volume of coolant was at the boiling point. A fast acting thermocouple and calibrated meter was used to determine this.

The coolants tested included pure water, pure ethylene glycol, 50/50 EGW, water modified with a surfactant, and 50/50 modified by a surfactant. These are the most common mixes.

Rationale: The pie tin is a fair model of a cooling system. The stovetop represents the heat generated by the engine, while the exposed surface of the coolant provides heat transfer to the atmosphere, emulating the radiator in a car. Using a standard volume of coolant reflects the way an actual cooling system is filled, and therefore includes the full systemic effect of specific gravity. The pie tin has minimal mass, so the heat transfer properties of the pan need not be considered. The longer the coolant can function without "boiling over", the more effective it will be in a real engine.

Shortcomings: In the experimental model, fluid remains static. The effect of circulation through a water pump may be significant. The systems is unsealed and unpressurized, which undoubtedly affects the results. In particular, results with pure water and aqueous glycol solutions are distorted, since significant evaporation occurs during the test.


Test 1. The above graph summarizes the first experiment, a trial of pure water vs. pure ethylene glycol vs. 50/50. It's clear that a 50/50 mix of water and ethylene glycol not only has a higher boiling point than water, but extends the heat carrying capacity of the system, which can be determined by calculating the area under the curve up to the boiling point. (Pure ethylene glycol was not tested to its boiling point, since it is very flammable.) The systemic heat carrying capacity is displayed in this graph:

Test 2. There are products called surfactants which improve the cooling ability of water by lowering it's surface tension. The next test attempts to measure the effectiveness of Redline Water Wetter, a commonly available surfactant. For comparison, a test was run with dishwashing liquid, another commonly available surfactant. The quantity of surfactant added to one pint of our test mix, per manufacturer's spec, was 1 fluid oz. Note that surfactants do not change the boiling point of the working fluid, but do affect it's ability to transfer heat.


Test 3. Various combinations of surfactant, ethylene glycol, and water were tested. The use of surfactant only applies to aqueous solutions.

Test4. Non-aqueous glycol. At least one manufacutrer makes a pure glycol coolant that is designed to be used without water. Since glycol has a relatively low specific heat, a concern with using such products is that operating temperatures would increase. To test this hypothesis, an measurement was made of the number of BTU's required to raise various coolant mixtures to a standard temperature (212F). The negative effect of surface tension is evident in the results for water vs. water+Water Wetter. The results for ethylene glycol are especially interesting, since they suggest a specific heat of .78, vs. an expected result of .57. This may mean that commercial coolants have a different composition than standard, or that there's some factor (like convection cooling) that isn't being measured in the experiment. Based on these results, pure glycol should not raise normal operating temperatures.

Summary of Results. Water does not cool best. But water with a surfactant is an excellent coolant. Pure ethylene glycol gives the best boilover protection, but should never be used in non aqueous solutions because it's inflammable. Given that most drivers are concerned with freeze protection as well as cooling, a 50/50 mix, plus a surfactant is the best choice.
Copyright©2002 Michael Frank

About the author - Michael Frank is Vice President and Webmaster of the Jaguar Touring Club. His red 2+2 E-Type is a familiar sight at JCNA events in the Northeast. Currently, he is building a reproduction of the 1963 Cunningham Lightweight E-Type .

Updated : 6/24/2002