CEPNZ webinar

Picking the right heat pump refrigerant: R290 v CO2

Using simulation analysis to assess what refrigerant performs best for your project

Presented byAlastair McDowell, EnergyAE
Date28 September 2026
AudienceCEPNZ members and guests
StatusEducational · v1
Overview

Where we are going

R290 or CO2, for a commercial hot water system: how do you decide, and which features of the application make the call?
01How the machines are built
→
02CO2 and R290, compared
→
03Three live New Zealand projects
→
04Ring mains, side by side
→
05Sizing from the load profile
→
06Where the crossover falls

Finding the critical ratio that determines the optimal refrigerant choice

01
Part one

Heat pump water heater fundamentals

Understanding heat pump components, commercial design variants, distribution systems and refrigerants.

Fundamentals

Introduction to heat pump water heaters

one outdoor heat pump unit ambient air in cooler, drier air out Evaporator takes heat from air Compressor raises pressure Condenser gives heat to water Expansion valve refrigerant loop heated water P pump hot cold Stratified tank to the load cold mains
Figure 1. Basic heat pump water heater system.
Fundamentals · live

Watch it run

Draw 10 L/min
Speed
06:00
one outdoor heat pump unit ambient air in cooler, drier air out Evaporator takes heat from air Compressor off Condenser gives heat to water Expansion valve refrigerant loop heated water pump Stratified tank sensor 60°C tap closed cold mains
HP off
at tap62°C
COP–
drawn0 L
Live simulation, illustrative.
Fundamentals

Common commercial designs and variants

TANKS IN PARALLEL HP P the pump is told to hold a fixed flow a fixed outlet temperature or a fixed rise across it Both tanks sit at the same temperature. You gain volume, not a cold end to draw from. TANKS IN SERIES HP P cold hot load Water passes one tank then the next, so the store keeps a cold end and a hot end rather than one average temperature. DIRECT HP TANK The stored water is the water that goes through the machine. Simplest, and the machine sees exactly what the tank holds. INDIRECT HP TANK coil A coil keeps the two waters apart. The coil can sit on the heat pump side, as drawn, or on the load side where it heats water on its way out.
Figure 2. Commercial design variants.
Fundamentals · live

Build the arrangement

Heat pumps always parallel
Tanks
Tanks are
Coupling
Machine
Distribution
Pump holds
Figure 3. Build your own commercial system schematic
Fundamentals

Dead-leg and ring-main variants

DEAD LEG heat pump tank outlets cold mains Every litre drawn is replaced by cold mains. Nothing circulates, so there is no standing loss and no loop to keep hot. Cost: a long wait at a distant outlet. RING MAIN heat pump tank supply, 60°C or more outlets P return, 55°C or more Hot water circulates continuously so every outlet runs hot on demand. The loop loses heat all year and hands warm water back to the plant. Cost: a standing load, and a warmer tank.
Figure 4. Dead-leg vs ring-main
Fundamentals

Natural refrigerants in commercial water heating

RefrigerantGWP FlammabilityToxicityWater temperature
Propane R290 3 A3 highly flammable Low To about 75°C
Carbon dioxide R744 1 A1 non-flammable Low Up to about 90°C
Ammonia R717 0 B2L lower flammability B toxic To about 80°C
Against the synthetic refrigerants:
  • Very low GWP, so no phase-down pressure
  • Good upper temperature range: 75°C and above
  • Each has a safety issue: R290 flammable, ammonia toxic, CO2 high pressure
Fundamentals

Synthetic refrigerants in commercial water heating

RefrigerantGWP FlammabilityToxicityWater temperature
R-32 difluoromethane 677 A2L mildly flammable Low To about 65°C
R-513A blend 573 A1 non-flammable Low To about 60–65°C
R-1234yf / ze HFO < 1 A2L mildly flammable Low To about 65–70°C
R-134a 1,300 A1 non-flammable Low To about 65°C
R-410A blend 1,924 A1 non-flammable Low To about 60°C
Against the natural refrigerants:
  • Less flammable and less toxic, so easier to site
  • Higher GWP, and above about 700 being phased down
  • Lower maximum water temperature: few go much past 70°C
02
Part two

Three New Zealand case studies

Real systems implemented in NZ

Case study · Christchurch

Aged care decarbonisation, Christchurch

Before
  • 400 kW boiler on a 20% biodiesel blend
  • 8,000 L thermal store, 100 kW of resistance elements on off-peak
  • Erratic outlet temperatures, and outages through winter 2022
After
  • Hot water on a single-pass heat pump module
  • The ring main has its own dedicated reheat heat pump
  • Space heating: 4 × 90 kW variable-flow heat pump water heaters
  • R-32 throughout, plus 25 kWp of solar PV
Four Temperzone MAGNUS variable-flow heat pump water heaters installed outdoors at the site

The space heating machines as installed. Photo from the published report.

Heating emissions, monthly: fell 97% after the boiler came out.

George Manning, Heritage Lifecare. Project by Decarbonised Energy Solutions; solar PV by Sunshine Solar; report published by EECA, June 2023.
Case study · South Canterbury

A hospital that split by duty

Domestic hot water → CO2

  • 120 kW nominal, into cylinders
  • Electric elements for backup and Legionella cycles

Space heating → R-410A

  • 450 kW, delivering 65°C to the existing radiators
  • Holds output down to −2°C ambient
Timaru Hospital. Two 6 MW coal boilers replaced. 2,280 t of coal and 2,000 t CO2-e saved a year, NZ$4.56M. Consultant DETA Consulting, installer Airtech; case study published by RACE for 2030. The report states the refrigerants were chosen for their thermodynamic fit to each duty.
Case study · Auckland to Christchurch

Aged care on R290, Auckland to Christchurch

Before
  • Gas instantaneous units coupled to storage vessels
  • Plant at end of life; a carbon target to meet
  • Existing sites short of electrical capacity
After
  • Multi-pass R290 heat pumps, about 30 facilities
  • 65°C storage, no electric boost
  • One plant covers draw-off, recovery and maintenance
  • More storage, slower recovery: sized to the amps
Two outdoor R290 heat pump units on a concrete pad beside an aged care building

Two of the R290 units at one of the sites. Photo from the published project page.

Aged care rollout, about 30 facilities. Heat pumps supplied by Waterware, who published the write-up (June 2024). No measured energy or COP published.
Case studies · Compared

Three projects, three answers to the ring main

Christchurch aged care Timaru HospitalAged care rollout
Hot water refrigerant R-32 CO2 R290
Hot water machine Single-pass module CO2 heat pumps into cylinders Multi-pass, 65°C storage
Ring main Its own reheat heat pump Not published Carried by the same plant
What is published Measured emissions, part-year Project outcomes, not monitored Supplier write-up, no data
The question this session answers: give the loop its own heat pump, or choose a primary machine that can carry it. Sources: EECA (2023); RACE for 2030 (2025); Waterware (2024).
03
Part three

Comparing R290 and CO2 system architectures

Understanding how refrigerant choice impacts system design

Architectures

The R290 reference system

cold mains Multi-pass R290 HP +6 K Mixed store 70°C Mixing valve to building P loop return
Figure 6. Multi-pass R290 on a ring main. The loop return goes straight into the mixed store.
Architectures · live

The CO2 reference system

Loop return
Figure 7.
Architectures

R290 or CO2: pros & cons

Pros
Cons
R290subcritical, multi pass
  • Cheaper equipment, China-sourced
  • Low GWP
  • Good COP at high entering water
  • Easy to service
  • Flammable: charge limits, outdoor or ventilated plant
  • Outlet ceiling ~70–75°C
  • More storage on large loads
CO2transcritical, single pass
  • High temperatures easily: 65–90°C
  • Non-flammable, non-toxic, GWP 1
  • Strong in cold climates
  • Needs cold inlet water
  • Ring main: swing tank plus element
  • High pressures: specialist service
Architectures · live

Comparing test data for R290 and CO2 systems

Show
Metric
View
slide the air temperature
Air 7°C
Figure 8. Comparison of performance data for two typical R290 and CO2 systems.
04
Part four

System sizing principles

Peak draws set the storage. The time between peaks sets the machine.

Sizing · live

System sizing principles

How peaky is the day?
where practice sits
45%
Hours to recover
5 h
Daily hot water load
1000 MJ
· ring main assumed: +35% on the day’s energy for loop losses
·
Storage, litres
set by the peak draw
·
Heat pump, kW
set by the day’s volume
·
Recovery, L/h
formula 16.6
Peakiness moves the storage and leaves the machine alone. The recovery window moves the machine and leaves the storage alone.
Figure 9. Basic sizing method: size the load → storage volume → heat pump capacity.
05
Part five

Simulation of ring-main systems

Understanding the factors that trade off between R290 and CO2 in commercial hot water.

The crossover

TRNSYS simulation

Weather Air temperature and humidity, every hour of a typical year, for each site Heat pump Measured EN 14511 map: output and power looked up from air and entering water, every step Controls A tank sensor switches the machine and pump on and off, as the real thermostat does Element Tops the tank up when the heat pump cannot hold temperature; every kWh of it is counted Stratified tank Split into layers, so hot sits on cold and the machine sees the real temperature at its inlet Hot water load Hourly draw profile scaled to the daily load, mixed down to delivery temperature at the tap HOW IT RUNS Each icon is a component, each line passes values between them. The solver steps the whole system every 72 seconds for a full year and totals the electricity.
Figure 10. Illustration of basic heat pump water heater in TRNSYS simulation software.
The hypothesis

Factors that influence heat pump performance

THE PHYSICS CO2 is best on cold water and cold air A big lift in one pass. Warm inlet water costs it heavily. R290 copes with warm inlet water A small lift, many passes. Most other conditions favour it. WHAT THE TANK GETS BACK Load dominates cold mains loop return Plenty of cold make-up. CO2 runs on cold water and stands out. loop losses ÷ hot water delivered CO2 favoured R290 favoured a crossover somewhere in between (the next slides find it) Losses dominate 55–60°C return cold mains CO2 can’t use it: an element tops up at COP 1. R290 runs on it, with little element. WHAT MOVES IT Ring-main size Longer, bigger loop: more standing loss → towards R290 Thermal load More hot water delivered: more cold make-up ← towards CO2 Climate Colder mains and air: CO2’s best conditions ← widens the CO2 side
Figure 11. The hypothesis, before any numbers: what the tank gets back decides which machine wins, and the building sets that balance.
The crossover · live

Which one is cheaper to run

Climate cold warm Auckland
Ring main 150 m
Daily load 800 MJ
Day shape flat to peaky
The answer

The critical loss-to-load ratio

The crossover loss-to-load ratio is 0.3–0.4, higher for cold climates. Note: this result would change with different equipment, sizing decisions, or system design, but illustrates the principles directionally.
Conclusion

What to take away

  • Energy modelling earns its place on commercial hot water
  • R290 and CO2 are both suited to it, and neither is the default
  • CO2 wins on colder mains and low entering water
  • R290 wins once the loop is large enough to need real maintenance heat
  • Efficiency is only one part of the equation: capital cost, maintenance, safety and other considerations also influence it

Thank you for listening. Time for questions

Alastair McDowell · Director, EnergyAE
alastair.mcdowell@energyae.com · energyae.com

Energy Analysis & Engineering Ltd · HPWH compliance modelling, TRNSYS simulation, scheme registration.
Educational session. Modelled units are anonymised and banded, and comparisons are between thermodynamic architectures rather than between products. Case study material is reproduced from publicly published reports with credit to the parties that did the work.
→ advance  ·  F fullscreen  ·  N notes
Speaker notes