Umm, not exactly...

If there is no air in the system, then the coolant simply expands due to heat, creating pressure. The thermal expansion coefficient is dependent to a degree on the mixture of water/coolant. Glycol has an expansion coefficient about 2.5 times that of water. The system reaches an equilibrium pressure that is determined by the relationship between the heat input to the system from the engine, and the heat output from the system via the radiator and heater if used. When the heat input rises above that equilibrium, the pressure is relieved by lowering the system volume, i.e. coolant being pushed into the recovery tank. As long as the system remains sealed, once it cools sufficiently, contraction of the liquid lowers the system pressure to below atmospheric, and atmospheric pressure in the recovery tank forces liquid back into the cooling loop.

Now, if there is air in the system, the overall pressure will he *higher* than if the system were filled totally with liquid. The pressure due to liquid expansion will be the same, but the increase in vapor pressure in the coolant will cause the liquid to outgas into the air layer/pockets. The air in the system will  expand due to heat, creating higher pressure, and the water/glycol vapor contained in the air will expand due to heating as well. So the pressure in the system will be controlled by the partial pressure of the air *plus* the partial pressure of the water/glycol vapor in the air, and the reduction in the air/vapor volume due to compression from the liquid expansion. Air will not decrease the pressure, it will increase it for any given overall heat gain. In decades of developing steam sterilization cycles, I've measured this effect many times.

It is also not quite accurate to say that the heat transfer capability of steam is less than water. For a given pressure, the specific heat of steam is less than water, by a couple of orders of magnitude, but creating the steam to start with requires a huge amount of energy transfer (the latent heat of evaporation) to change phases from liquid to gas. That energy would be given up again when it condensed in the radiator, if it could get there. If it weren't for this latent heat of evaporation property, air conditioning wouldn't work. 

Boiling, however, is another creature. When the water boils at the interface between the coolant and the engine, dissolved gases are liberated from the liquid, not just steam generation, but dissolved oxygen, nitrogen, etc. These gases create an insulating layer between the liquid and the engine surface causing very high localized engine temps. When those bubble collapse, they create shock waves that you hear as "thunks" and "bumps" from when they're cooled too rapidly. The same shock waves that damage cavitating pumps, although the cause of the bubbles is a bit different.

Keith Hughes
'86 Westy GenV (Marvin)
 
 ------------------------------
 
 Date:    Sun, 21 May 2017
 12:52:10 -0400
 From:    David Beierl <dbeierl@ATTGLOBAL.NET>
 Subject: Re: What can go wrong.....
 
 I'm sure that "not" was intended. 
 However it took me some headscratching
 to see how that would affect pressure
 buildup, as vapor pressure of coolant
 would be constant at a given
 temperature -- until I realized that there are
 two different mechanisms involved.
 
 Initial pressure results from straight
 expansion of liquid coolant against
 the confines of the cooling
 system.  Sixteen liters of 50% glycol coolant
 expands about half a liter going from
 30C to 100C.  Any air in the system
 would have a lower "spring rate" than
 the hoses, and would thus reduce the
 overall pressure -- until boiling,
 micro or otherwise, began; at which
 point the ultimate pressure would be
 controlled by the temperature in the
 hottest submerged section of the
 system, as vapor would be generated
 (quickly or slowly) until the
 equilibrium pressure was reached and the
 boiling stopped.
 
 All heat-generating portions of the
 system must be submerged to avoid
 severe local heating, as steam has
 greatly inferior heat transfer ability
 compared to liquid.  Same problem
 with micro boiling, it interposes a
 low-conductivity layer between the heat
 source and coolant.
 
 Yrs,
 d
 
 On Sun, May 21, 2017 at 11:45 AM,
 George Gaudette <gcgaudette@gmail.com>
 wrote:
 
 > Can someone say a few more words
 about this?:
 >
 > "If the engine is completely
 filled with coolant the system will not build
 > pressure and micro boiling and
 vaporizing will do damage to the cylinder
 > walls and head studs"
 >
 > I'm thinking the word "not" might
 have been accidentally left out ("If the
 > engine is **not**
 filled...").  If so, not further explanation
 requested.
 >
 > -George
 >
 
 ------------------------------