What happens when a glycol solution freezes?

[fxb80]

I've been told by a knowledgeable engineer that when a glycol solution freezes, the water freezes first and increases the glycol concentration. It stands to reason this is true since a glycol solution is a mixture. I've seen a glycol solution with a 13F freezepoint that had ice crystals floating in it.

So, if a refrigerant-to-water heat exchanger is chilling a glycol solution to 25F and the evaporator temp is 0F and the glycol freeze point is 10F, wouldn't a pure water ice form on the evaporator? And if pure water ice forms on the evaporator, wouldn't the glycol solution have to reach 32F+ to thaw the ice?

Thanks,

CW

[jayguy]

i would believe that if the water seperated out when standing still (like oil and water...mixes ok when moving but seperates out when left standing). the water and the glycol stays mixed and doesn't freeze in parts...if that were true, then why use glycol at all? eventually it would accumulate in the evaporator and lock up the flow. but we know this does not happen.

[CraziFuzzy]

i would believe that if the water seperated out when standing still (like oil and water...mixes ok when moving but seperates out when left standing). the water and the glycol stays mixed and doesn't freeze in parts...if that were true, then why use glycol at all? eventually it would accumulate in the evaporator and lock up the flow. but we know this does not happen.

Glycol Water solutions are a solution, not a mixture. The glycol is bonded with the water, which is why it doesn't separate, and why it affects the boiling and freezing points of the water. This is also why it is so hard to 'recycle' glycol from the water - distillation does not work. It has to be done via a much more complex process (pervaporation - which is a combination of RO and distillation principles, with a chemical 'helper').

[fxb80]

Thanks, guys. Makes sense. But let me offer an observation.

There's a job where we have a glycol premix tank with 23% ethylene glycol. The tank is slightly open to atmosphere; there's a lid but some small openings not sealed. Knowing that the water will evaporate out of a glycol solution and leave the glycol behind, after it had been in a few weeks I checked the solution, expecting it to show a greater glycol concentration. Thing is, it showed less. Same refractometer, and I knew nobody had added water, etc. I stirred the tank and the percentage showed higher, I added a little water, stirred, and rechecked until it showed 23%. Which led me to believe that some of the glycol, being heavier than water, had migrated to the bottom. Sample was taken off the top of the tank. I repeated this observation a few weeks later to prove it wasn't a fluke. What would cause that to happen if the glycol bonds to the water?

[fxb80]

Well, I'm back to the original question because I got data from Dow that says ice forms when the solution temperature drops below freeze point. See attachment.

[bigtime]

Dont know if this answers your question, but I have put a glycol/water sample in a freezer to make sure it wouldnt freeze. It made a slushie, but didnt freeze hard enough to bust stuff, which I was concerned about.

[jayguy]

Glycol Water solutions are a solution, not a mixture....

solution is a better term for sure.

from the document provided by fxb, Dow says that the water freezes first...true, however, it is freezing first in a 'non-pumping system'. also, the water (in the glycol solution) is freezing first at its 'solution freeze point'...not 32F (unless it is 0% glycol of course).

as for your tank, the only thing that i can offer is that a refractometer requires that the lens and the sampled solution be at the same temperature for testing purposes. i have not performed any experiments to see the differences. i usually see it within a few degrees F of design or waaaay off. i don't usually have to be gnats-ass precise in my job so i haven't worried about it. perhaps there can be some slight differences in density depending on your tank height and how long it sits. glycol also has inhibitors and other chemicals...perhaps that is making a difference too. how far off is the non-stirred glycol compared to the stirred glycol?

[fxb80]

It has been a couple of years since I found that problem with the glycol premix storage tank, but to my recollection the tank showed 2-3% less concentration before stirring. The refractometer was at the same temperature as the solution in the tank and is temperature compensated from 55F-85F. The tank is in a climate-controlled mechanical room. For the second measurement, that showed higher percentage after stirring, I did allow the tank to stop swirling before drawing a sample; stirring would introduce some air into the solution. There were no obvious bubbles, as if that means anything. The ethylene is Wintrex brand, which comes with inhibitors, and there is no water treatment being done. I did try to be as close to gnat's ass perfect as possible because this glycol and the percentage was specified by an engineer and is the only protection for an outdoor chiller, the pump doesn't run in cold weather, and the pump capacity was sized for that percentage so I didn't want to go heavy. It seemed to me that 23% was a little light so I had my records for personal protection. I recorded on the tank what the initial fill percentage was, and double-checked the percentage of the original fill after the tank had a day or two to settle. So it was very surprising to see an initial measurement the opposite of what I expected on two dates some weeks apart. The second time convinced me the glycol was increasing concentration on the bottom.

What I should have done to absolutely positively sure was measure bottom concentration. Duh. But I'm not gonna do that now because I added a valve and piping to turn the solution over once a day using the existing fill pump, measurement is taken every three months, and doesn't show lower percentage.

If the solution is moving that does change things, not because the freeze point changes but simply because you can keep the ice thawed as it forms.

Back to my hypothetical, there would be a point where if the evaporator is cold enough it could have ice on it even if there is flow. That's what I'm trying to think through.

The aggravating thing is I'm spending all this time thinking and asking questions because the powers that be for this job from hell, the brewery chiller that I detailed in the Commercial section, won't increase the glycol percentage. Simple me thinks if the evaporator is 10F the glycol freeze point must be no more than 10F, and definitely not the 20F they measured.

[skpkey9]

This is really in the wrong section, and should be in refrigeration.

I'm rushing out the door but will answer your question later, but in short water will separate out and form pure ice on the evap if not kept in check , if you are talking about a beer glycol system... Will elaborate more later when I have a moment...

[jayguy]

...but in short water will separate out and form pure ice on the evap if not kept in check...

it will be interesting to know what you mean by 'kept in check'. i have many chillers that run below 32F leaving SOLUTION temperature. they run for many, many days (sometimes weeks) without ever shutting down. I would imagine that they would have shown issues if the water separated out and froze on the inside of the evaporator. this would have caused the glycol to concentrate and 95% glycol (SR-1, EG) freezes at -3F. as it concentrates above a certain concentration, the freeze points rise.

in a flowing system, the water stays in solution as long as you are above the SOLUTIONS freeze point...otherwise, why use glycol at all if it is just going to separate out and freeze up?

[Nuclrchiller]

I wonder if condensation was laying on top of the solution in the tank, from humid air entering the equipment room...

[fxb80]

it will be interesting to know what you mean by 'kept in check'. i have many chillers that run below 32F leaving SOLUTION temperature. they run for many, many days (sometimes weeks) without ever shutting down. I would imagine that they would have shown issues if the water separated out and froze on the inside of the evaporator. this would have caused the glycol to concentrate and 95% glycol (SR-1, EG) freezes at -3F. as it concentrates above a certain concentration, the freeze points rise.

in a flowing system, the water stays in solution as long as you are above the SOLUTIONS freeze point...otherwise, why use glycol at all if it is just going to separate out and freeze up?

Are you saying you have chillers with an evaporator SST colder than the solution freeze point?

If flow is enough to keep the glycol in solution regardless of evap temp, is any flow enough? I did some measurements today of the evaporator that prompted this. After deducting approximate refrigerant tubing volume (I have an identical barrel that I cut the top off), the solution volume of the evaporator is 40 Gallons. Flow, corrected for temperature and % ethylene, is about 28 GPM. Evaporator is 2.5' from inlet to outlet. Average fluid velocity is 0.029 Feet/Second or 0.348 Inch/Second. I don't know what chiller normal is but that doesn't seem like much.

I, also, am interested to hear what skpkey9 has to say.

[fxb80]

I wonder if condensation was laying on top of the solution in the tank, from humid air entering the equipment room...

Sample was actually drawn from about 1" below the surface. I've never seen any condensation in this room, but understand that doesn't necessarily mean there isn't any; the room has never felt humid, room temperature tends to be above outdoor dew point.

[jayguy]

I wonder if condensation was laying on top of the solution in the tank, from humid air entering the equipment room...

that is an absolutely excellent point. i only had 1 chiller that was an open tank and that was a long time ago.

Are you saying you have chillers with an evaporator SST colder than the solution freeze point?

If flow is enough to keep the glycol in solution regardless of evap temp, is any flow enough? I did some measurements today of the evaporator that prompted this. After deducting approximate refrigerant tubing volume (I have an identical barrel that I cut the top off), the solution volume of the evaporator is 40 Gallons. Flow, corrected for temperature and % ethylene, is about 28 GPM. Evaporator is 2.5' from inlet to outlet. Average fluid velocity is 0.029 Feet/Second or 0.348 Inch/Second. I don't know what chiller normal is but that doesn't seem like much.

I, also, am interested to hear what skpkey9 has to say.

typical minimum flow rates are around 3 feet per second minimum but i don't know what you have for a chiller barrel.

this whole thread seems to me to center around the water molecules separating out or freezing at 32F instead of the SOLUTION freeze point temperature. this is not true in a flowing system. the water molecules are going to stay in solution and not freeze out separately and collect in the coldest part of the system (evaporator).

i do not have any chillers that have a design evaporating temperature colder than the solution freeze point...that would be dumb and counter productive. i have run across systems that were improperly maintained and had an evaporating temperature lower than the solution freeze point (due to low flow rates) and they did occasionally make slush...but i believe that it was a solution slush, not a water slush and the glycol still a fluid.

[fxb80]

jayguy, I never said the water would freeze at 32F whether the solution is flowing or not flowing, or asked if it would. And we agree it's dumb to have a freeze point higher than the evaporator SST. But that's what I've got, a dumb situation. The system is about 120 gallons short on propylene (1800+ gallons system volume); the barrel was going to be replaced; I suggested then was a good time to not re-use the solution being drained; put straight propylene back in; it wasn't done. The man who made the decision said it would be easy enough to fix later. In the meantime, I think ice has formed on the evaporator.

Which really got me thinking. How do I prove there's ice? What will it take to thaw it? If it wasn't a glycol solution, or if it operated at normal HVAC temperature, it's easy enough to go by the SST if the refrigeration is idle and the evaporator is the coldest part. But, as long as the solution temp stays below 32F, and there's pure ice on the evaporator, the SST with the refrigeration idle would be less than or equal to the solution temperature. The owner won't tolerate letting the solution temperature rise to 32F if it can be helped. So I would turn off the pump, valve off one side of the barrel, and add heat from an external source (heat gun, etc.).

During this process, if the SST rises steadily, then plateaus at the freezing point of the solution, then rises again, that would mean the ice was returned to solution at the freezing point. If the SST plateaus at 32F, that would prove the ice must get to 32F to return to solution. So I can find out by experiment if there's ice and at what temperature it returns to solution.

Since the barrel is isolated, and if the evaporator had pure ice on it, and if the ice were to return to solution at the freezing point of the solution, the freezing point would rise during the thawing process due to the greater percentage of water in solution. Which adds another variable I might need to think through.

But if I could know with certainty that ice in a glycol solution must get to 32F+ to thaw, that's easier.

[double bogey]

I read a paper about glycol solutions a few years back as I was taking care of a dew ice storage systems. It is my understanding that as your temp drops to the solution freeze point, it makes a slurry, seperating some water, and causing the solution to be a higher concentration, therefore lowering the solutions freeze point. Still, if you start glazing ice on your chiller tubes, your situation is going downhill.

Just read the Dow paper, thats exactly what it says.

[jayguy]

i would not know how to go about proving your theory, although i would like to know. another factor to consider is the pressure inside the solution system. as the pressure goes up, the freeze point of water changes...so that is going to complicate things further.

[nique]

two points that I have talked to chemists about. first glycol that is below 30% reacts to temps different then would be expected, meaning that the glycol does have the ability to chemicly seperate. second point is that as the solution starts to slush (in the barrel) the flow will decrease in some of the tubes causing problems with the chiller pressures/temps as now there is little heat transfer at those tubes. just my opinion but I never run my machines at less then 30%. another factor to consider is the inhibitors at less then 30%