Refrigerants: GWP, ODP, safety class & the great phase-downs

1 · What makes a good refrigerant

A refrigerant is not chosen for one property but for how a dozen of them line up — and they rarely line up neatly. The job is to boil at the cold end of the cycle and condense at the hot end, both at pressures a real compressor and real piping can handle. Get any one requirement badly wrong and the fluid is unusable, so refrigerant selection is an exercise in compromise, not optimisation. The tensions that matter:

• The right boiling point at practical pressures. The saturation temperature must fall inside the working range at pressures that are neither a deep vacuum (which lets air and moisture leak in) nor dangerously high. A fluid that boils too low runs the whole system at high pressure; one that boils too high runs the evaporator under vacuum. This single requirement is what separates a comfort-cooling refrigerant from a low-temperature or a high-temperature one.
• High latent heat of vaporisation. Most of the heat moves as the refrigerant changes phase, so a large latent heat means a small mass flow carries a large duty — smaller compressor, smaller pipes. Ammonia is the champion here, which is why it dominates big industrial plants.
• Chemical stability and material compatibility. The fluid must not decompose over years of cycling, and it must live happily with the compressor oil, the copper or steel piping, the elastomer seals and the motor windings. Ammonia attacks copper; CO₂ needs components rated for very high pressure; some HFOs are mildly reactive. Get this wrong and you get acid formation, sludged oil, or a burnt-out hermetic motor.
• Safety — toxicity and flammability. A leak into an occupied space must not poison or ignite. This is the axis ASHRAE 34 formalises, and it is where the environmentally best fluids (hydrocarbons, ammonia) carry the most handling risk.
• Low environmental impact. Historically the last thing considered; now often the first. A refrigerant that destroys ozone or forces the climate is a liability that regulation will eventually strand, whatever its thermodynamics.

These pull against each other. The most thermodynamically elegant fluids — the hydrocarbons and ammonia — carry safety penalties. The safest synthetic fluids tend to have high GWP. Low-GWP synthetics buy their environmental score with mild flammability. There is no free lunch; every refrigerant on the market is a point on that trade surface.

2 · The families

Refrigerants sort into a few chemical families, and the history of the industry is essentially the story of moving down this list as each generation is regulated out.

CFCs and HCFCs — the legacy, ozone-depleting fluids

Chlorofluorocarbons (CFCs) like R-12 and hydrochlorofluorocarbons (HCFCs) like R-22 were the twentieth-century workhorses: stable, non-flammable, cheap, excellent thermodynamics. Their fatal flaw is the chlorine atom, which reaches the stratosphere and catalytically destroys ozone. R-12 is long gone from new equipment; R-22 — for decades the air-conditioning refrigerant — has been phased out of new production under the Montreal Protocol, and surviving systems run on dwindling, expensive reclaimed stock. If you still meet R-22 in the field, you are looking at end-of-life equipment.

HFCs — no ozone damage, but high warming

Hydrofluorocarbons (HFCs) removed the chlorine, so their ODP is zero — they were the Montreal-era replacements and solved the ozone problem completely. But most are potent greenhouse gases. R-134a (automotive AC, chillers, domestic refrigeration), R-410A (the standard residential AC and heat-pump blend) and R-404A (commercial and low-temperature refrigeration) have GWPs of roughly 1,400, 2,000 and 3,900. That warming footprint is exactly what the Kigali generation of rules now targets. R-32 is a single-component HFC with a much lower GWP (~675) and is treated as a bridge fluid — lower-GWP than the blends it replaces, though mildly flammable.

HFOs and HFO blends — the low-GWP synthetics

Hydrofluoro-olefins (HFOs) add a carbon double bond that makes the molecule break down in days rather than decades in the lower atmosphere, collapsing GWP to single digits. R-1234yf (now standard in automotive AC) and R-1234ze (chillers, heat pumps) are the pure HFOs. Because pure HFOs have modest capacity, most equipment uses HFO/HFC blends tuned to hit a GWP target while keeping usable properties — R-454B and R-32 are the leading replacements for R-410A in new residential and light-commercial systems. Nearly all of these are mildly flammable (class A2L).

Natural refrigerants — the fluids nature already made

The naturals occur in the environment already, so their GWP is effectively zero or one and no future regulation can strand them. They are the long-term destination for much of the industry, at the price of handling challenges:

• R-744, carbon dioxide (CO₂). GWP of 1, non-flammable, non-toxic. It runs at very high pressure and often in a transcritical cycle (rejecting heat above its critical point rather than condensing). It shines in cold-climate applications and commercial refrigeration — supermarket racks and heat-pump water heaters — where its behaviour actually favours a large temperature glide on the high side.
• R-717, ammonia (NH₃). The efficiency champion, with unmatched latent heat and near-zero cost. It is toxic and mildly flammable, and it attacks copper, so it lives in industrial plants — cold stores, food processing, ice rinks — with steel piping, trained operators and gas detection. Superb thermodynamics, serious safety engineering.
• R-290 propane and R-600a isobutane — hydrocarbons. Excellent refrigerants with GWPs of about 3. Their problem is obvious: they are flammable (class A3). That confines them to systems with small charges — domestic fridges (R-600a), heat-pump monoblocs and self-contained commercial units (R-290) — where the total quantity that could leak stays below an ignition hazard.

3 · Two environmental metrics — ODP and GWP

Two numbers decide a modern refrigerant's regulatory fate, and it is worth being precise about what each one means because they are often confused.

Put them side by side for the fluids you actually meet in the field. Note how the industry has walked diagonally down this table over forty years — first killing ODP, now chasing GWP down toward the naturals at the bottom.

4 · The phase-downs — a timeline

Refrigerant history is written by treaties. Two distinct problems drove two distinct waves: first ozone (a phase-out on ODP), then climate (a phase-down on GWP). Follow the sequence and today's A2L transition makes complete sense.

1. 1987 — Montreal Protocol. The landmark ozone treaty. It phased out CFCs first (R-12 and relatives), then set a schedule to eliminate the less-damaging HCFCs (R-22). Regulated purely on ODP. Widely regarded as the most successful environmental agreement ever signed — the ozone layer is now measurably recovering.
2. 1997 — Kyoto Protocol. The first climate treaty. It listed HFCs among the basket of greenhouse gases to be controlled, but set no refrigerant-specific schedule. It flagged the problem; it did not solve it. HFC use kept climbing precisely because Montreal had pushed the industry onto them.
3. 2016 — Kigali Amendment (to Montreal). The turning point. It brought HFCs under the Montreal Protocol's enforcement machinery and mandated a global phase-down on GWP — cutting HFC production and consumption by roughly 80–85% by 2047 on a stepped schedule. This is the treaty driving the whole current transition to low-GWP fluids.
4. EU F-Gas Regulation. The EU's implementation, and long the world's most aggressive. It works by a quota system that shrinks the total CO₂-equivalent of HFCs placed on the market year on year, plus outright bans on high-GWP fluids in specific new equipment. The 2024 revision (Regulation 2024/573) tightened the schedule toward a near-complete HFC phase-out by 2050 and set explicit dates for switching sectors to low-GWP alternatives.
5. US AIM Act (2020). The American Innovation and Manufacturing Act gave the EPA authority to implement Kigali: an 85% HFC phase-down by 2036 via allowances, plus "technology transition" rules that set GWP limits for new equipment by sector — the rules pushing US residential AC and heat pumps onto R-454B and R-32 from 2025.
6. The result — low-GWP HFOs, naturals, and the A2L shift. The GWP ceilings leave two paths: synthetic low-GWP fluids (HFOs and their blends) or natural refrigerants. Most of the low-GWP synthetics are mildly flammable (A2L), so new equipment, standards and service practice are being rebuilt around handling flammable refrigerant safely — the defining engineering change of this decade.

5 · Safety classification — ASHRAE 34

As the industry moves to flammable low-GWP fluids, the ASHRAE Standard 34 safety class (mirrored in ISO 817) becomes the number that governs installation. It is a two-part code: a letter for toxicity and a digit for flammability.

• Toxicity — A or B. A = lower toxicity (no harm identified up to 400 ppm exposure); B = higher toxicity. Ammonia is the common B.
• Flammability — 1, 2L, 2 or 3. 1 = no flame propagation; 2L = mildly flammable, with a low burning velocity (≤ 10 cm/s) and high ignition energy; 2 = flammable; 3 = higher flammability (the hydrocarbons).

Combine them and you get the classes you see on every cylinder and nameplate:

The practical consequence of the A2L transition is charge limits: standards (IEC 60335-2-40 for heat pumps, EN 378 for systems) cap how much flammable refrigerant may be installed for a given room size, and require measures such as leak detection, restricted pipe routing, and the removal of potential ignition sources. For an A3 hydrocarbon the limits are tighter still, which is why propane and isobutane are largely confined to self-contained appliances. For service technicians it means new tooling and procedures: A2L-rated recovery machines and gauges, no open flames near a charged system, and brazing only after proper evacuation and purging. None of this makes A2Ls dangerous in normal use — a low burning velocity means an A2L is very hard to ignite and does not sustain a fast flame — but it does mean the old "any tech, any fluid" habits no longer apply.

6 · Glide & zeotropic blends

A pure refrigerant boils and condenses at a single temperature for a given pressure. Many modern refrigerants are not pure — they are blends of two or three components chosen to hit a GWP or capacity target. And here a subtlety appears that trips up field work: how a blend changes phase depends on whether it is azeotropic or zeotropic.

• Azeotropic blends behave like a single pure fluid — the components boil together at one temperature, and the vapour has the same composition as the liquid. (The old R-500-series.) No glide.
• Zeotropic blends have components with different boiling points, so as the mixture evaporates or condenses its temperature slides across a range at constant pressure. That range is the temperature glide — often 4–7 K for common blends. R-407C, R-454B and R-404A are zeotropic to varying degrees.

Glide has two real consequences an engineer must respect:

• Evaporator/condenser matching and superheat. With glide, there is no single saturation temperature — you must use the dew point and bubble point when calculating superheat and subcooling, or you will misread the charge. Glide can be used deliberately, matching the refrigerant's sliding temperature to the temperature change of the air or water it exchanges with, for a small efficiency gain in a well-designed counter-flow exchanger.
• Charge as liquid, never vapour. Because the components have different volatilities, a zeotropic blend will fractionate if you charge it as a vapour — the more volatile component boils off the cylinder first, so the blend you put in the system no longer has the intended composition. Always charge a zeotropic blend from the cylinder as a liquid (cylinder inverted, or via a dip tube), metering it into the low side so it flashes on the way in. The same reasoning means a large leak can shift the composition of whatever remains, so a badly leaked zeotropic system should be recovered and recharged, not merely topped up.

7 · Charging & leaks

Whatever the fluid, a system only performs at its rated COP if it holds the correct charge. As the opening guide to this series set out, the two field measurements that tell you the charge are superheat and subcooling — the degrees of vapour above its boiling point at the compressor inlet, and the degrees of liquid below its condensing point at the condenser outlet. Subcooling is the primary charge indicator on most systems: low subcooling points to undercharge, high subcooling to overcharge or a condenser that cannot reject its heat. A technician trims the charge to hit the manufacturer's target superheat/subcooling for the measured conditions — and with a zeotropic blend does so against the dew and bubble points, not a single saturation temperature.

Which makes leaks the quiet enemy on two fronts at once:

• Efficiency. A leak lowers the charge, and an undercharged system loses capacity and COP — the evaporator starves, superheat climbs, the compressor works harder for less useful cooling or heating. A slow leak shows up first as a rising energy bill and a machine that can no longer hold setpoint on a hot or cold day, long before it fails outright.
• Climate. Every kilogram that leaks is a direct emission, and for a high-GWP fluid the arithmetic is brutal: one kilogram of R-404A leaked is climate-equivalent to almost four tonnes of CO₂. Over the life of a leaky commercial system, those direct refrigerant emissions can dwarf the carbon footprint of all the electricity the machine ever consumed. This is precisely why the low-GWP transition matters so much, and why F-Gas and the AIM Act mandate leak checks, leak-detection systems and recovery rather than venting.

So refrigerant management is not a side task — it is where reliability and sustainability meet. Tight systems, correct charge, prompt leak repair, and proper recovery at end of life protect the COP and the climate at the same time. Streamed continuously, superheat, subcooling, and refrigerant-side pressures are exactly the signals a condition-monitoring platform trends toward an early leak or charge alarm — the subject of the final guide in this series.

Key takeaways

• A good refrigerant is a compromise — boiling point at practical pressures, high latent heat, stability, material compatibility, safety, and low environmental impact all pull against each other.
• Know the families: CFC/HCFC (ozone-depleting, gone), HFC (zero ODP but high GWP), HFO and blends (low GWP, mostly A2L), and the naturals — CO₂, ammonia, hydrocarbons.
• Two metrics rule: ODP (ozone, R-11 = 1) drove Montreal; GWP (climate, CO₂ = 1) drives Kigali, F-Gas and the AIM Act. A fluid can have zero ODP and still a huge GWP.
• Safety class governs installation: ASHRAE 34 pairs toxicity (A/B) with flammability (1/2L/2/3). The move to A2L low-GWP fluids brings charge limits, leak detection and new service practice.
• Zeotropic blends have glide — use dew and bubble points for superheat/subcooling, and always charge them as liquid to avoid fractionation.
• Leaks hurt twice — undercharge cuts COP, and for high-GWP fluids the direct emissions can outweigh the electricity's entire footprint. Keep systems tight and recover, never vent.

Continue the series