20.2 Glycol and Brine Secondary Coolants (pH, freeze point, inhibitors)
Key Takeaways
- Food plants usually specify inhibited propylene glycol (PG) because it is less toxic; ethylene glycol (EG) transfers heat better and pumps easier but is toxic if ingested and must not contaminate food or potable water.
- Freeze point falls as glycol or CaCl2 concentration rises until the eutectic, then rises; use the inhibitor vendor’s chart and the correct EG vs PG refractometer (or a CaCl2 hydrometer), not a guessed percent.
- Inhibited industrial glycol is typically held slightly alkaline (often about pH 8.0–10.5 depending on the package) to protect steel; follow the inhibitor vendor, not a target of pH 7.
- Never mix EG with PG, automotive antifreeze with industrial inhibited glycol, or incompatible inhibitor chemistries; freeze-point tables and lab tests become unusable.
- A complete program trends freeze point, pH, inhibitor/reserve alkalinity, conductivity, and biological fouling on a written schedule—and after any large makeup or leak—not freeze point alone.
A secondary coolant (brine, glycol, heat-transfer fluid) is not a refrigerant in the vapor-compression sense. It does not boil in the room coils. A pump circulates a liquid that was chilled by ammonia—or by cascade CO2—in the engine room, and that liquid picks up heat in production coils and returns warmer. CIRO items in this area expect you to know which fluid, how freeze point tracks concentration, why pH and inhibitors matter, and what you must never mix.
Ethylene glycol versus propylene glycol
Ethylene glycol (EG) and propylene glycol (PG) are the two industrial glycols. Both are mixed with water by volume percent in most HVAC-R tables (always confirm whether a chart is volume or weight). Both depress freeze point. They are not interchangeable in a food plant or in an inhibitor package.
EG has lower viscosity and better heat transfer at the same concentration, so pumps and coils work easier. It is toxic if ingested (sweet taste, dangerous to animals and people). It belongs in closed industrial loops that cannot contaminate food, beverage, or potable water.
PG is less toxic. Food plants, beverage, and many inspected facilities specify inhibited propylene glycol, often with a food-grade / HT-1 heat-transfer listing from the vendor. The penalties are higher viscosity, poorer heat transfer, and more pump energy, especially below 0°F. At low temperature a 40–50% PG loop can look mysteriously short of capacity when the real problem is viscosity and film coefficient, not the ammonia chiller.
Automotive antifreeze is not industrial inhibited glycol. It is packed with silicates, phosphates, or organic-acid packages meant for aluminum engines. Those packages foul plate exchangers, drop out as sludge, and fight industrial inhibitor tests. Never dump leftover automotive product into a plant loop.
Never mix incompatible glycols. Do not mix EG with PG. Do not mix two inhibitor chemistries (silicate vs nitrite vs molybdate vs organic-acid) even if both jugs say propylene. Freeze-point tables become lies, inhibitors precipitate, and you cannot interpret a lab report. If a loop was contaminated, the honest fix is dump, flush, and recharge with one specified fluid—not a little of each to chase freeze point.
Freeze point versus concentration
Freeze point falls as glycol concentration rises, then rises again past the eutectic. More glycol is not always more protection. Pure EG freezes near 9°F; the water mixture is what gives you −30°F-class protection.
Approximate freeze points from widely published volume-percent tables (always use the inhibitor vendor’s chart and a glycol-specific refractometer—EG and PG scales differ):
| Volume % | Approx. EG freeze | Approx. PG freeze |
|---|---|---|
| 20% | ~18°F | ~19°F |
| 30% | ~7°F | ~9°F |
| 40% | ~−10°F | ~−7°F |
| 50% | ~−34°F | ~−28°F |
Design for freeze protection (still pumpable) at the lowest expected fluid temperature, with margin. Vendor practice often keeps freeze point well below the lowest operating temperature so you are not running in slush. Burst protection (a weaker mix that slushes but does not split a coil) is for idle hydronic systems, not for a freezer coil you still need to pump.
If someone adds water because the expansion tank was low, concentration and freeze protection fall together. Makeup water also brings hardness, chloride, and oxygen. Automatic city-water makeup on a glycol loop is how plants dilute themselves into a freeze-up.
A refractometer reads index of refraction, not inhibitor health. The scale must match EG or PG. Reading a PG loop on an EG scale (or the reverse) is a false freeze point. Handheld units also assume a temperature; a boiling-hot sample or a slushy sample will lie. Sample from a flowing header, not a dead leg.
Calcium chloride brine
Calcium chloride (CaCl2) brine is the classic secondary for ice plants and older low-temp rooms. It is cheap, it has good thermal conductivity, and the freeze curve goes very low:
| CaCl2 weight % | Approx. freeze point |
|---|---|
| 10% | ~20°F |
| 20% | ~−4°F |
| 25% | ~−21°F |
| 29.8% (eutectic) | ~−67°F |
Past the eutectic, freeze point rises. A stronger brine is not automatically safer. Check concentration with a hydrometer / specific gravity table for CaCl2, not a glycol refractometer.
Uninhibited CaCl2 is aggressive to carbon steel, and a leak attacks concrete. Keep air (oxygen) out of the loop, maintain the inhibitor package the vendor specified, and hold pH in the slightly alkaline range the brine inhibitor requires. Older chromate inhibitors are a hexavalent-chromium problem; do not reintroduce them. Sodium chloride brine exists but its eutectic is only around −6°F, so it is the wrong tool for a −20°F room.
Food plants that still have CaCl2 often wish they had PG. A brine leak is salty, corrosive, and a sanitation headache. PG is the usual modern choice where toxicity and food contact drive the spec.
pH, inhibitors, conductivity, biology
pH of inhibited industrial glycol is typically slightly alkaline—commonly about 8.0–10.5 depending on the package—because that range passivates steel. Follow the inhibitor vendor, not a generic idea that pH 7 is neutral so it must be fine. Acid drift means reserve alkalinity is gone: the buffers that eat residual acidity are spent, and steel (and often iron in the chiller) starts to dissolve. Very high pH can attack aluminum, zinc, and some gaskets. Yellow-metal protection (azoles) is a separate part of the package; a steel-only nitrite program can still eat brass.
Corrosion inhibitors are consumed. Heat, oxygen in-leakage, galvanic couples, and chloride all spend the package. You cannot see inhibitor level on a freeze-point refractometer. A loop can have a perfect −20°F freeze point and still be eating the plate chiller.
Conductivity reflects dissolved ions: inhibitors, plus contamination. A sudden drop often means dilution. A sudden rise can mean chloride or hardness from makeup, a process leak into the glycol, or inhibitor overfeed. Some food plants use conductivity as a glycol-to-product leak detector. Trend it; do not chase a single number without the vendor’s band.
Biological fouling is not only a cooling-tower story. Closed glycol loops, especially dilute PG (a carbon source, more biodegradable than EG), grow slime. Biofilm on plate exchangers and coil tubes is a ΔT penalty that looks like the plant needs a bigger chiller. Use closed-loop design (air separator, no open sump), keep concentration in the specified band, and apply only biocides the inhibitor vendor allows. The wrong biocide wrecks the inhibitor.
Testing you can defend on an audit
A competent program, documented in the plant’s mechanical-integrity / water-treatment SOP:
- Weekly to monthly: visual (color, haze, particles, oil sheen), freeze point with the correct EG or PG refractometer, pH.
- Quarterly: inhibitor or reserve-alkalinity kit per vendor, strainer inspection, expansion-tank level, pump strainer ΔP.
- Semi-annual to annual: full lab (iron, copper, chloride, biological dip-slides or ATP as specified, glycol percent). Corrosion coupons if the program uses them.
- After any large makeup, leak, or unexplained capacity loss: do not wait for the calendar.
Log results against the vendor’s control band. If pH has collapsed or chlorides are high, adding a drum of glycol does not fix chemistry. You may need a dump and recharge. If freeze point is barely below room temperature in a freezer loop, you have a freeze-up waiting, not a chemistry trivia item.
When makeup is required, add premixed inhibited glycol at the specified concentration, not raw water plus a guess of neat glycol. Record gallons. A loop that needs frequent makeup has a leak; find it. Glycol on a floor or in a sanitary drain is both a concentration problem and, in a food plant, a product-hold problem.
A USDA-inspected ready-to-eat plant is replacing a secondary coolant. Which fluid choice matches common industrial practice and why?
A mechanic pours leftover ethylene glycol into a propylene-glycol food-plant loop to restore freeze point. What is the correct CIRO-level response?
An inhibited steel-pipe glycol loop shows a correct freeze point on a PG refractometer, but pH has fallen into the acidic range. What does that tell you?
Which statement about freeze point versus concentration is correct for CIRO plant operation?