CO₂ Refrigeration Lubricants: POE vs PAG for R-744 Systems
Lubricant selection for R-744 (carbon dioxide) refrigeration systems hinges on three factors: how the system returns oil to the compressor, the lubricant’s behavior in contact with CO₂, and its viscosity after CO₂ dissolution. Polyol ester (POE) is the most common choice in transcritical booster and direct-expansion systems, where the refrigerant itself carries the lubricant back to the compressor. Specially formulated polyalkylene glycol (PAG) is used in transcritical and pumped systems where its phase behavior and the oil-recovery design are compatible.
Subcritical cascade and pumped systems can recover lubricant in separate vessels, reducing the miscibility requirement. In all configurations, successful selection depends on two combined checks: reliable oil return and sufficient viscosity under actual operating pressure and temperature.
CO₂ refrigeration performance is determined by both pressure and temperature. At transcritical high-side pressures of 80–120 bar (1,160–1,740 psi), dissolved CO₂ and operating temperature both reduce lubricant viscosity — precisely when the compressor needs it most.
CO₂ has a critical temperature of 30.98 °C (87.8 °F). Whether a system runs subcritically or transcritically depends on gas-cooler outlet temperature, controlled high-side pressure, load, and control strategy — not solely on ambient temperature.
The suction-to-discharge pressure differential in a transcritical CO₂ system is roughly five times that of a comparable R-404A system. This increases compressor bearing loads, making residual lubricant viscosity a critical reliability factor, not a minor detail.
Published phase-behavior data for an evaluated POE-85 show full miscibility with liquid CO₂ across the common subzero evaporating range. The evaluated PAG-68 is only partly miscible, and below approximately −32 °C (−25.6 °F) saturated suction temperature, it becomes less dense than liquid CO₂ and floats above it in low-pressure vessels.
Both POE and PAG absorb moisture. The difference lies in what happens next: water breaks POE down into acids through hydrolysis, while PAG base stocks resist hydrolysis. Both chemistries require dry handling and moisture monitoring.
CO₂ lubricant selection balances reliable oil return against sufficient in-service viscosity. The right balance depends on the system architecture — no single chemistry excels in both aspects simultaneously.
CO₂ Refrigeration System Types and Lubricant Requirements
R-744 systems operate across subcritical and transcritical conditions, but lubricant requirements are determined by more than just pressure. Compressor type, circuit layout, oil separation, and the return path collectively dictate the required chemistry and in-service viscosity.
Why CO₂ is different. Most refrigerants have critical points well above typical refrigeration system temperatures, allowing them to condense normally. CO₂’s critical point—30.98 °C (87.8 °F) at 73.77 bar absolute (1,070 psia)—falls within ordinary summer conditions. Consequently, CO₂ systems must be designed for an operating regime that other refrigerants rarely encounter.
Subcritical vs Transcritical Operation
Below the critical point, CO₂ condenses conventionally: pressure determines the saturation temperature along the saturation curve. This is the mode of operation for cascade systems, where a CO₂ low stage rejects heat to a high stage using ammonia, an HFC, or a hydrocarbon refrigerant.
Transcritical operation occurs when the R-744 high-side process operates above the critical pressure and rejects heat without condensing. The heat exchanger functions as a gas cooler rather than a condenser, and high-side pressure is actively controlled for capacity and efficiency instead of being fixed by saturation temperature.
Transcritical high-side pressures typically range from 80–120 bar (1,160–1,740 psi), though specific operating and design pressures depend on the equipment. For instance, Copeland specifies a 120 bar (1,740 psi) design pressure for one compressor family and 135 bar (1,958 psi) for another, while certain R-744 valves are rated for a maximum working pressure of 140 bar (2,030 psi). Therefore, it is imperative that every component is selected based on its individual allowable pressure rating.
- The transition is dynamic. While ambient temperature influences the achievable gas-cooler outlet temperature, it does not solely determine the high-side state. Gas-cooler outlet temperature, controlled high-side pressure, load, and control strategy determine whether the system runs subcritically or transcritically—and most transcritical systems fluctuate between these modes throughout the year.
- What this means for the lubricant. The lubricant must lubricate and seal the compressor, withstand discharge temperatures, and reliably return from wherever the refrigerant carries it. These three requirements often conflict, and CO₂’s high pressures intensify this conflict more than any common refrigerant.
Compressor Applications and Lubricant Requirements
| Position | System Type | Compressor Type | Lubricant Requirement |
|---|---|---|---|
| Transcritical booster, medium temperature | Transcritical | Reciprocating, semi-hermetic | High miscibility for oil return; viscosity retention at 80–120 bar (1,160–1,740 psi). |
| Transcritical booster, low temperature | Transcritical | Reciprocating, semi-hermetic | High miscibility and low-temperature oil return from LT evaporators. |
| Parallel compression / flash gas | Transcritical | Reciprocating | Same requirements as the booster; sees the highest suction density. |
| Cascade low stage | Subcritical | Reciprocating or screw | Separation-based oil management; miscibility is not required. |
| Industrial pumped / overfeed | Subcritical | Reciprocating or screw | Oil recovery from the low-pressure drum governs selection. |
| Direct expansion, commercial | Subcritical or transcritical | Reciprocating | Miscibility supports oil return through the evaporator. |
| High-stage CO2 in a hybrid cascade | Transcritical | Reciprocating or screw | Transcritical lubricant requirements apply. |
| Immiscible-design industrial systems | Subcritical | Screw with separation | Low CO2 solubility and maximum viscosity retention. |
| Secondary CO2 circulation pumps | Subcritical | Pump, not compressor | Pump-bearing lubrication; the lubricant does not contact the refrigerant. |
Two observations follow from this table. Transcritical systems place the highest demands on oil return and viscosity retention because the refrigerant itself is the return mechanism. Subcritical systems support a broader range of lubricant chemistries because their vessels allow for the collection and recovery of oil.
POE vs PAG: Miscibility, Dilution, and Oil Return
POE and PAG interact differently with liquid and gaseous CO₂. Selection requires a balance between phase behavior, oil return, CO₂ dilution, and the viscosity that remains at the compressor.
How Miscibility Affects Oil Return
Two terms are central to this discussion, and they are not the same:
Describes whether liquid CO₂ and lubricant mix into a single liquid phase or separate into layers.
Describes how much CO₂ gas dissolves into the lubricant at a given pressure and temperature.
Dissolved CO₂ thins the lubricant. Solubility and oil temperature together determine the viscosity loss — more dissolved gas and hotter oil both thin the film. Miscibility determines whether the oil travels with the refrigerant or stratifies away from it. A lubricant can be highly miscible and heavily diluted, or partly miscible and comparatively thick — and that is exactly the trade-off between POE and PAG.
The central tension is this: the lubricant must maintain enough viscosity to protect the compressor, while the oil-management system must return whatever lubricant the refrigerant carries into the circuit. No single lubricant perfectly satisfies both lubricity and miscibility; every selection is a deliberate compromise matched to the system.
- Transcritical systems rely on the refrigerant for oil return. POE’s high miscibility across the common subzero evaporating range keeps the oil moving with the CO₂ through booster and direct-expansion circuits and back to the compressor. Purpose-formulated PAG is also used in transcritical direct-expansion and booster configurations — but only where its phase behavior, piping velocities, oil separation, and return path have been verified together. Transcritical operation alone does not dictate the chemistry; the complete oil-management architecture does.
- Subcritical systems can separate oil instead. Cascade low stages and industrial pumped systems include receivers, drums, and rectifiers — physical locations where lubricant collects and can be deliberately recovered. With that infrastructure in place, a partially miscible lubricant with higher diluted viscosity becomes a practical option, which is where PAG earns its position.
Partial Miscibility and Density Inversion
Published phase-behavior data for an evaluated POE-85 show full miscibility with liquid CO₂ across the common subzero evaporating range, with a limited immiscibility region at positive temperatures and certain CO₂-rich compositions — so “fully miscible” is accurate for the conditions that matter, not a universal description. The evaluated PAG-68 shows a much wider immiscible region, and its miscibility depends strongly on temperature and oil concentration.
Miscibility charts answer one question: one phase or two. They do not answer a second question that matters just as much in vessel-based systems: which layer sits on top.
Separate density data for the same fluids show that the evaluated PAG-68 becomes less dense than liquid CO₂ below approximately −32 °C (−25.6 °F) saturated suction temperature. Above this point, the PAG-rich phase sits at the bottom of a low-pressure vessel. Below it, the layers swap, and the PAG floats on the CO₂. This crossover, called density inversion, dictates the placement of a recovery connection. A bottom take-off designed to collect a PAG-rich layer will collect liquid CO₂ instead once the operating temperature drops through the inversion point.
The consequence is clear: for typical medium-temperature pumped systems, both evaluated fluids work. At approximately −40 °C (−40 °F) saturated suction temperature, the evaluated POE-85 is preferred with a conventional rectifier, because the reversed PAG layer position defeats that recovery arrangement. This conclusion holds for these two fluids and that vessel design, not for every POE and PAG pairing in every system.
The rectifier itself is why any of this matters. Oil separators are never 100% efficient, so some lubricant always reaches the low-pressure drum with the liquid CO₂. A rectifier recovers it by evaporating an oil-rich CO₂ stream and directing the concentrated lubricant back toward compressor suction. That recovery design is matched to one lubricant’s layer behavior. It is not a property that transfers when the lubricant is swapped — changing chemistry means reviewing the vessel arrangement.
Measured: NEXT 744-POE and NEXT 744-PAG under CO₂
NEXT ran both R-744 lubricants on pure CO₂ at two points: NEXT 744-PAG at ISO VG 68 and NEXT 744-POE at ISO VG 85 — deliberately the same grade pairing as the published evaluation discussed above, so the results read side by side with it.
| Condition | Lubricant | Dissolved CO2 | Operating Viscosity | Mixture Density |
|---|---|---|---|---|
| MT crankcase — 60 °C (140 °F), 35 bar (508 psi) abs | NEXT 744-PAG-68 | 8.95 wt% | 14.2 cSt | 0.957 g/ml |
| MT crankcase — 60 °C (140 °F), 35 bar (508 psi) abs | NEXT 744-POE-85 | 9.09 wt% | 10 cSt | 0.948 g/ml |
| Evaporator — −40 °C (−40 °F), 10 bar (145 psi) abs | NEXT 744-PAG-68 | 55.44 wt% | 6.2 cSt | 1.089 g/ml |
| Evaporator — −40 °C (−40 °F), 10 bar (145 psi) abs | NEXT 744-POE-85 | 50.37 wt% | 7.33 cSt | 0.989 g/ml |
Three readings follow from the table.
- The crankcase point reproduces the published comparison. At 0 °C (32 °F) saturated suction temperature — 35 bar of CO₂ over 60 °C oil — the NEXT calculation returns 14.2 cSt for the PAG against 10 cSt for the POE, where the published evaluation of the same grade pairing reported approximately 15 against 10. The two chemistries absorb nearly the same amount of CO₂ at this point; the difference is what each retains while holding it.
- Evaporator Performance and Lubricant Mobility. At −40 °C under saturated CO₂, both lubricants absorb roughly half their weight in refrigerant. Fluids that are effectively immobile when neat flow at 6 to 7 cSt. The dissolution that thins the film at the crankcase is the same mechanism that mobilizes oil for return from an evaporator.
- Densities are consistent with layer behavior. The diluted PAG mixture calculates to 1.089 g/ml at −40 °C — below liquid CO₂’s approximately 1.12 g/ml at the same temperature. This is consistent with published findings that the PAG-rich phase floats in low-pressure vessels below approximately −32 °C (−25.6 °F). The POE mixture’s layer position is less critical because POE stays mixed with the liquid CO₂.
Both runs are on pure R-744; calculations are extrapolated from measured solubility data on NEXT products and apply specifically to NEXT products.
Oil Management by System Type
Miscibility is one input to the oil-management design. The required phase behavior depends on piping, refrigerant velocity, oil separators, receivers, rectifiers, and the engineered return path.
Direct Expansion
The refrigerant expands into the evaporator and leaves as superheated vapor. There is no low-side vessel to collect lubricant, so return depends entirely on phase behavior, refrigerant velocity, piping geometry, and complete evaporation before the compressor. Lubricant that remains in the evaporator forms an oil film that degrades heat transfer and capacity. Both POE and purpose-formulated PAG are used in CO₂ direct expansion; the selected formulation must demonstrate reliable return and sufficient diluted viscosity across the operating envelope.
Transcritical Booster Systems
Medium- and low-temperature loads converge on a common suction line, with flash gas managed at the receiver — often by a parallel compressor. Oil separators sit at the discharge, but the low side is a distribution network, not a collection point. Lubricant return relies on the refrigerant’s carrying capacity, following direct-expansion principles, with the added complication of seasonally swinging high-side pressures.
Pumped and Overfed Industrial Systems
Liquid CO₂ is pumped to the evaporators and collects in a low-pressure drum. The drum is a deliberate collection point: lubricant is recovered through a rectifier rather than carried back by refrigerant velocity. This architecture tolerates — and in some arrangements prefers — a partially miscible lubricant, because the design already expects the oil to separate. In an evaluated ejector-based overfeed arrangement, PAG-68 remains denser than liquid CO₂ at the reported suction-accumulator condition of −6 °C (21 °F) and intermediate-receiver condition of 5 °C (41 °F). The PAG-rich phase settles at the bottom of these vessels, which simplifies recovery. POE-85 stays mixed with the liquid CO₂ under the same conditions and requires rectification or another engineered recovery method.
Cascade Low-Stage Systems
These run subcritically with a condensing condition fixed by the high stage. Oil management follows conventional practice for the compressor type, which lowers the miscibility requirement.
Two design details deserve attention before any chemistry is selected:
Suction Line Velocity. Oil return in refrigerant-carried systems depends as much on gas velocity as on chemistry. Even with the correct lubricant, an oversized suction line can still starve the compressor of returned oil.
Drum-Based Recovery. In vessel-based systems, the recovery arrangement must match the specific lubricant’s layer behavior, including its density inversion point.
Moisture, Materials, and Compatibility
Discussions of moisture in CO₂ lubricants are usually reduced to “POE is hygroscopic.” While true, this misses the point: both chemistries absorb water. The difference lies in what the water does once absorbed.
Water Absorption: POE vs PAG
Water absorption depends on PAG structure, end-capping, formulation, temperature, relative humidity, and exposure time. Both POE and PAG absorb moisture and require controlled handling.
POE Hydrolysis vs PAG Water Retention
This is where the two chemistries diverge completely.
- POE hydrolyzes. Water breaks the ester bonds in POE, producing organic acids and alcohols—the reverse of the reaction that created the lubricant. These acids drive corrosion and sludge formation.
- PAG base stocks resist hydrolysis. They contain no ester bonds to break. While additives and other formulation components still require moisture control, the base stock itself does not degrade like an ester.
The relevant distinction is hydrolytic stability, not a water-absorption ratio. Moisture is a controlled contaminant in both chemistries, and the control measures are identical: charge from sealed containers, minimize open exposure during service, evacuate properly, maintain the filter drier, and measure water content instead of assuming it.
Moisture Control in CO₂ Systems
Neither chemistry eliminates the need for moisture control. POE requires moisture control because water drives hydrolysis. PAG requires moisture control because water affects the complete formulation and system materials.
For both, the control measures remain consistent: charge from sealed containers, minimize open exposure during service, ensure proper evacuation, maintain the filter drier, and actively monitor water content rather than making assumptions.
Discharge temperature and in-service viscosity
At identical suction conditions, discharge pressure, speed, and geometry, the lubricant does not change the theoretical CO₂ discharge temperature.
However, it does change the actual lubrication conditions inside the compressor. At an oil temperature of 60 °C (140 °F) and a saturated suction temperature of 0 °C (32 °F), published data for the evaluated fluids show a residual viscosity of approximately 15 cSt for PAG-68 versus approximately 10 cSt for POE-85.
NEXT’s own calculation under the same conditions returns 14.2 cSt for NEXT 744-PAG-68 versus 10 cSt for NEXT 744-POE-85, reproducing the comparison from NEXT’s measured solubility data (see the measured section above). PAG retains more viscosity while both chemistries absorb a similar amount of CO₂ at this point.
This difference impacts oil-film thickness, sealing, friction, wear, and oil circulation, which is why two compressors at the same nominal conditions do not necessarily run at the same measured temperatures. Fluid analysis across the operating envelope—diluted viscosity, water content, total acid number, wear metals, deposits—reveals what a datasheet cannot.
Materials and seals
Two compatibility issues in CO₂ service are often overlooked and are unrelated to miscibility:
- PAG and Material Compatibility: PAG is incompatible with certain seal materials and paints. A system that has run on POE or a hydrocarbon lubricant since commissioning has an elastomer population conditioned by that lubricant, so a conversion to PAG requires a materials review, not just an oil change. Seal specifications belong in the conversation with the compressor manufacturer before the conversion, not after.
- Ester contamination in ammonia circuits. In ammonia/CO₂ cascade systems, the CO₂ side commonly runs POE while the high stage runs ammonia with a hydrocarbon lubricant. Polyol esters react with ammonia to form amides — solids, viscous liquids, or sludge that block orifices and foul heat-transfer surfaces. The CO₂ side of the cascade heat exchanger often operates at the higher pressure during normal operation, which creates a credible leak path toward the ammonia circuit; the pressure differential during startup, shutdown, standstill, and defrost must be verified before the expected leak direction is defined. This is a design-review item rather than a lubricant-selection item, but it belongs on the same checklist — and it is one more reason to know exactly which chemistry runs in each circuit.
While this is primarily a design review item rather than a lubricant selection item, it should be included on the same checklist. It also underscores the importance of knowing the specific chemistry used in each circuit rather than making assumptions.
POE or PAG? Selecting the Right R-744 Refrigeration Lubricant
NEXT positions POE and PAG chemistries for different R-744 system architectures. The correct product follows from the required oil-return mechanism, the diluted viscosity, material compatibility, and the operating envelope.
| Factor | PAG | POE | Recommended NEXT Product |
|---|---|---|---|
| Liquid CO2 phase behavior | Partly miscible; strongly dependent on temperature and oil concentration. | High miscibility across common subzero conditions; limited immiscibility region at positive temperatures for the evaluated POE-85. | NEXT 744-POE where liquid-phase transport returns the oil; NEXT 744-PAG where separation and recovery are engineered. |
| Transcritical booster / DX | Adequate oil return after complete CO2 evaporation, subject to verified return design. | Widely used; phase behavior supports refrigerant-assisted return. | NEXT 744-POE or NEXT 744-PAG, selected from the oil-return design and formulation-specific PVT data. |
| Medium-temperature pumped / overfeed | Partly miscible; separates from liquid CO2 in the evaluated arrangement, simplifying recovery. | Stays mixed with liquid CO2; requires rectification. | NEXT 744-PAG where vessel-based separation is used. |
| Ejector-based overfeed | Evaluated PAG-68 settles below liquid CO2 at −6 °C (21 °F) and 5 °C (41 °F). | Evaluated POE-85 remains mixed under the same conditions. | NEXT 744-PAG, subject to verified formulation-specific density data. |
| Low-temperature pumped, near −40 °C (−40 °F) | Evaluated PAG-68 floats above liquid CO2 below approximately −32 °C (−25.6 °F), defeating a conventional bottom take-off. | Evaluated POE-85 works with a conventional rectifier. | NEXT 744-POE for conventional recovery; NEXT 744-PAG only with a recovery design built for the reversed layer position. |
| Residual viscosity at the published comparison point | Approximately 15 cSt at 60 °C (140 °F) oil temperature and 0 °C (32 °F) SST. | Approximately 10 cSt under the same conditions. | NEXT 744-PAG where higher verified residual viscosity is required. |
| Moisture and hydrolysis | The base stock resists hydrolysis; moisture still affects the formulation and system materials. | Susceptible to hydrolysis, acid formation, and deposits when wet. | NEXT 744-POE with standard moisture control through sealed charging, driers, and monitored water content; NEXT 744-PAG where moisture exposure cannot be reliably controlled. |
| ISO VG range | 46–150 | 55–170 | Select the grade from the calculated in-service viscosity requirement, not from the fresh ISO grade alone. |
POE dominates transcritical commercial systems because its phase behavior allows refrigerant-assisted oil return across common evaporating conditions. Purpose-formulated PAG holds a strong position in transcritical and pumped systems where its higher diluted viscosity is beneficial and the recovery design matches its layer behavior.
Neither is universal, and any recommendation that ignores whether the system is transcritical or subcritical has overlooked a fundamental question. OEM also governs chemistry selection.
The NEXT CO₂ Refrigeration Lubricant Range
| Product | Base Stock | ISO VG | Miscibility With CO2 | Position |
|---|---|---|---|---|
| NEXT 744-POE | Polyol ester | 55–170 | High across common subzero conditions; formulation-specific. | Transcritical booster, direct-expansion, commercial, and conventional rectifier-based systems. |
| NEXT 744-PAG | Polyalkylene glycol | 46–150 | Partial; formulation-specific. | Transcritical booster and DX systems; pumped and overfeed systems with active oil management. |
What type of oil is used in CO₂ refrigeration compressors?
Polyol ester (POE) and polyalkylene glycol (PAG), and the choice follows the system type rather than the refrigerant.
Transcritical systems depend on the lubricant returning with the refrigerant through the circuit, which requires miscibility. POE is fully miscible with liquid CO₂ and is the normal answer there.
Subcritical and industrial systems have receivers, drums and rectifiers — physical places where lubricant can be collected and returned deliberately. Where that infrastructure exists, a partly miscible PAG becomes viable and retains more viscosity under dissolved CO₂.
Some industrial systems are designed to keep the lubricant out of the circuit entirely using separation and low solubility, in which case a polyalphaolefin (PAO) is appropriate.
The compressor manufacturer’s specification governs in every case.
Why does CO₂ reduce lubricant viscosity so much more than other refrigerants?
Because of how much dissolves and how thin the dissolved phase is.
CO₂ viscosity is two to three orders of magnitude lower than lubricant viscosity, so dissolving it into the oil can reduce viscosity by as much as an order of magnitude. The amount dissolved rises with pressure, and transcritical systems run 80 to 120 bar on the high side with components designed for 130 to 140 bar.
A NEXT calculation for a PAG at 70 °C shows 23.8 cSt neat falling to around 11 cSt with CO₂ dissolved — more than half, at the point where the compressor carries its highest loads.
This is why in-service viscosity is calculated per operating point rather than read from a chart.
Does dissolved CO₂ ever help?
Yes, and the same mechanism does both jobs at opposite ends of the circuit.
A POE at −40 °C is effectively immobile as a neat fluid — a NEXT calculation puts it at 126,862 cSt. Dissolve 12 % CO₂ into it at 50 bar and it becomes a 35 cSt fluid that flows. The refrigerant is what makes oil return physically possible from a low-temperature evaporator.
At the discharge end the same dissolution is the threat. That is the central tension in CO₂ lubricant selection: maximum oil return and maximum film strength sit at opposite ends of one axis, and the system design decides which end to favour.
Is POE unsafe in CO₂ systems because it is hygroscopic?
No, but moisture control is not optional — and the common framing has it backwards.
PAG absorbs roughly four times more water than POE: approximately 10,000 ppm against 2,500 ppm at saturation, with mineral oil around 25 ppm for comparison. So PAG is the more absorbent chemistry.
The difference is what happens next. PAG holds water by hydrogen bonding to the polyether chain and drains it out with the lubricant, with no chemical reaction. POE hydrolyses: water breaks the ester bonds into organic acids and alcohols, and those acids are implicated in corrosion and sludge.
Hydrolysis is also conditional rather than automatic. It requires sufficient water at elevated temperature, above roughly 80 °C, and is insignificant at ambient conditions and low moisture content. Transcritical discharge temperatures reach that window, which is why moisture is treated more seriously in CO₂ service than in HFC service running the same chemistry.
What is density inversion, and when does it matter?
It is the point at which liquid CO₂ and a partly miscible lubricant swap positions in a vessel, because their relative densities cross over as temperature falls.
Published industrial data places it below approximately −32 °C saturated suction temperature for a PAG-68. Below that point, an oil recovery arrangement designed to draw from a particular layer draws the wrong fluid — so the recovery strategy stops working at exactly the temperatures where oil accumulation matters most.
It is relevant to low-temperature industrial and pumped systems and not to commercial medium-temperature duty. Where a plant runs below −32 °C saturated suction and uses a partly miscible lubricant, the recovery arrangement should be reviewed against the specific fluid rather than assumed.
Can POE contaminate the ammonia side of a cascade system?
Yes, and the pressure difference makes it a one-way risk.
Ammonia attacks the ester bond and depolymerises POE, producing solids, viscous residues and sludge that foul heat transfer surfaces and can block orifices. In an ammonia/CO₂ cascade the CO₂ circuit commonly runs POE and normally sits at the higher pressure, so a cascade heat exchanger leak carries ester toward the ammonia circuit rather than away from it.
That belongs on the design review checklist alongside the refrigerant leak assessment, not only on the lubricant order.
Can one CO₂ lubricant be replaced with another without flushing?
It depends on the chemistry families involved, and documentation rather than assumption should decide it.
NEXT holds a documented conversion from Fuchs Reniso C85E to NEXT 744-POE-85 supporting changeover with no flush and top-off use, based on mixture testing. Compatibility documentation for other CO₂ lubricants is available on request.
Two limits apply. PAG and hydrocarbon-based lubricants including PAO are not compatible with one another and require a full drain, flush and filter change to change between. And POE and PAO, while chemically compatible, represent different oil management strategies — switching between them is a system decision rather than a product substitution.
One point specific to CO₂: any service intervention is a moisture exposure event. On these systems the changeover procedure matters as much as the compatibility result.