Temperature lift is the irreducible bill
An ideal reversible machine sets the upper bound. Real equipment is farther away because heat cannot cross a finite coil without a temperature difference.
The useful ceiling is below Carnot
For heating, ideal COP equals the hot-side absolute temperature divided by the temperature lift. With a 70°F room and 5°F outdoor air, the generous air-to-air Carnot reference is about 8.2. But the refrigerant must boil below 5°F to absorb heat and condense above 70°F to release it. The compressor is not reversible, motors and fans consume power, and pressure drops, cycling, frost, and controls add loss.
This is why a field COP near two in deep cold can be technically impressive while remaining far below the textbook ceiling. Carnot shows direction: reduce lift and performance rises. Lower the supply-water temperature, enlarge the heat exchanger, seal the ducts, or reduce the building load, and the machine can operate in a friendlier part of its map.
The practical floor is not a compressor number. It is the smallest lift that the building, emitters, and weather allow.
Evaporate outside
Cold refrigerant absorbs heat from outdoor air. It must be colder than the air, so the true source temperature is below the weather reading.
- Measure
- Evaporating temperature, coil approach, airflow
- Failure boundary
- Frost blocks airflow and insulates the coil
Where the frontier moves
Larger coils, better surfaces, and demand-based defrost.
Compress the vapour
The compressor raises pressure and temperature so heat can flow indoors. Variable-speed drives let the machine follow load instead of cycling on and off.
- Measure
- Isentropic efficiency, speed range, discharge temperature
- Failure boundary
- High lift raises work and compressor stress
Where the frontier moves
Enhanced vapour injection, economization, and broader modulation.
Condense indoors
The hot refrigerant gives up heat to indoor air or water. The hotter the supply temperature, the larger the lift and the lower the COP.
- Measure
- Condensing temperature, supply air or water temperature
- Failure boundary
- Legacy radiators can demand hot water that erases efficiency
Where the frontier moves
Larger emitters, lower-temperature hydronics, and better ducts.
Expand and repeat
An expansion device drops refrigerant pressure and meters flow back to the evaporator as outdoor conditions and compressor speed change.
- Measure
- Superheat, subcooling, stable control
- Failure boundary
- Poor charge or control starves or floods the coil
Where the frontier moves
Electronic expansion valves and model-based controls.
Defrost without losing the house
Below freezing, moisture can ice the outdoor coil. Many systems briefly reverse, consuming electricity while taking heat from indoors.
- Measure
- Defrost frequency, duration, energy penalty
- Failure boundary
- Timed defrost runs when it is not needed
Where the frontier moves
Sensors and controls that infer actual frost mass.
Deliver heat through a building
Ducts, fans, pumps, emitters, thermostats, backup heat, and the envelope determine what reaches occupied rooms.
- Measure
- Static pressure, duct leakage, room temperature, peak load
- Failure boundary
- The installation can overwhelm equipment gains
Where the frontier moves
Measured commissioning, envelope work, and thermal storage.
Two curves bend at once
As outdoor temperature falls, a conventional air-source heat pump usually loses both COP and maximum heat output while the building needs more heat. Cold-climate designs attack both curves.
DOE moved the cold-weather threshold
All eight participating manufacturers' prototypes passed laboratory validation; several demonstrated heating at −15°F. The Challenge also required staged backup heat, demand-response capability, and refrigerant GWP no greater than 750.
Defrost creates a hidden sawtooth
Cold dry air is not always the hardest condition. Around freezing, moist air can deposit frost on the outdoor coil. Frost blocks airflow and reduces heat transfer; defrost often reverses the cycle, briefly cooling the indoor side while consuming electricity. A test-point COP can miss the timing and severity of that penalty.
Variable-speed compressors, vapour injection, electronic expansion valves, and larger heat exchangers have improved the map. Smarter defrost matters because it avoids paying the reversal penalty on a clean coil. The control system becomes part of the thermodynamic machine.
The prototype passed. The system still varied.
PNNL's field validation installed pre-commercial Challenge units in occupied and test homes across the northern United States and Canada, then monitored them through heating seasons.
| Evidence | Condition | Efficiency | Capacity | Boundary |
|---|---|---|---|---|
| DOE Challenge laboratory | 5°F | COP ≥2.4 up to 48,000 Btu/h; ≥2.1 above | 100% of 47°F capacity | All eight manufacturers' prototypes passed before field deployment |
| PNNL Challenge field validation | 0–5°F bin | Median COP 1.9 | Varied by site and operation | 22 completed sites; only eight had ≥10 hours below 0°F |
| ENERGY STAR cold-climate floor | 5°F | COP ≥1.75 | ≥70% of 47°F capacity | Certification screen, not a promise of whole-house seasonal performance |
| Earlier NY/MA field study | Heating season | Seasonal COP 2.34 | Not evaluated | 11 ducted cold-climate systems, summarized by PNNL |
What the 1.9 does and does not prove
The field median at 0–5°F cleared ENERGY STAR's 1.75 cold-climate threshold but trailed the Challenge laboratory requirement. It does not show that every prototype “lost” the same fraction: the dataset covers different systems, homes, loads, controls, and operating points. It does show why a single certified point cannot stand in for delivered seasonal performance.
The coldest tail was also sparse. Only eight of 22 sites had at least ten hours of heating data below 0°F, and many sites experienced milder winters than normal. Evidence at −15°F remains thinner than evidence at 5°F. The correct conclusion is progress with an uncertainty band, not universal performance in every cold snap.
A heat pump is commissioned, not delivered in a box
The same outdoor unit can be excellent or disappointing depending on the system around it.
Load calculation
Oversizing reduces modulation and can worsen comfort; undersizing pushes the house onto resistance backup. Manual J assumptions must match the actual envelope.
Ducts and emitters
High static pressure, leakage, poor return paths, or radiators designed for very hot water can turn a good compressor into a bad system.
Refrigerant charge
The right amount of refrigerant and a clean, dry circuit are prerequisites. A nameplate cannot compensate for a leaking flare or contaminated line set.
Controls
Thermostat logic, switchover temperature, defrost, staging, and installer defaults decide when expensive resistance heat enters.
Electrical service
A compressor, resistance strips, water heater, range, and vehicle charger can turn a heating retrofit into a panel, service, and utility upgrade.
Installer capacity
PNNL cites research indicating 70–90% of air conditioners and heat pumps exhibit at least one performance-compromising installation or maintenance fault.
The envelope is a thermal component
Air sealing and insulation do more than save annual energy. They lower the design heat loss, allowing a smaller heat pump, reducing required supply temperature, extending compressor-only operation, and shrinking the resistance-backup peak. In PNNL's field report, the authors recommend weatherizing before replacement where practical and note that study homes varied in insulation and airtightness.
This creates an ordering problem. Equipment replacement is often urgent; envelope work takes assessment, contractors, and capital. A policy that rebates the box but not the load calculation, ducts, panel, and shell can purchase rated efficiency without purchasing field performance.
Annual efficiency and winter peak are different objectives
A heat pump can reduce annual energy while increasing the electric system's coldest-hour load, especially when resistance strips switch on together across many homes.
When does the backup coil take over?
Cold-climate performance has two independent axes: how efficiently the compressor moves heat and how much heating capacity it retains. The second axis often sets winter peak demand.
Whole-system COP 1.90
- Compressor and fans
- 6.3 kW
- Resistance backup
- 0.0 kW
- 20% better equipment COP
- 5.3 kW peak
- 20% lower building load
- 5.1 kW peak
kWth is heat delivered; kW is electrical input. Resistance backup is modeled at COP 1. Capacity retention applies to the unit's 47°F rated output. Distribution losses, cycling, defrost, crankcase heat, and service-panel diversity are excluded.
Calculation and interpretation
Available cold-weather capacity = rated capacity × retained capacity. Compressor power = heat delivered by the heat pump ÷ COP. Any remaining building load is met by electric resistance at one electrical kilowatt per thermal kilowatt.
The “20% better COP” case improves only compressor efficiency. The “20% lower building load” case represents envelope or air-sealing work. When resistance heat is active, cutting the building load can reduce peak more than improving compressor COP because it removes the least-efficient final kilowatts first.
What another 20% equipment improvement buys
When the compressor serves the entire load, a 20% COP improvement cuts its electrical input by one-sixth. If backup heat is already active, that gain touches only the compressor portion. Improving capacity retention, lowering the building load, or staging backup can remove COP-1 kilowatts and therefore outperform the same percentage gain in compressor efficiency at the system peak.
NREL's electrification work finds the largest peak-timing shifts in the Northeast and Midwest. Its ResStock analysis also finds high-efficiency cold-climate heat pumps produce materially lower peak demand than single-stage ENERGY STAR units because they rely less on backup heat. The grid benefit is therefore not “electrification versus no electrification”; it depends on which equipment, which envelope, and which controls.
Thermal storage changes when the compressor works
A water tank, phase-change store, or preheated building can shift compressor work away from the most constrained hour. Hybrid systems can preserve an existing fuel source for rare peaks, reducing electrical upgrades at the cost of two systems and continued combustion. Grid-interactive controls can curtail compressors temporarily, but comfort recovery and backup staging determine whether curtailed load simply returns as a rebound.
Lower climate impact, new installation rules
The working fluid is both an efficiency choice and an emissions risk if it leaks.
EPA's Technology Transitions program sets a GWP limit of 700 for new residential and light-commercial air-conditioning and heat-pump systems. R-410A has a rule value of 2,088; R-32 is 675 and R-454B is 465. Lower-GWP A2L refrigerants are mildly flammable, so the transition changes equipment, tools, training, charge limits, transport, and service practice.
Direct refrigerant emissions are not captured by COP. A low-leak, repairable system using a lower-GWP fluid can outperform a nominally efficient system that loses its charge. Conversely, rushed installation and an unfamiliar service base can create leakage and cost. Refrigerant choice is therefore another case where the component frontier shifts work onto the installer.
The 2025 shipment decline is the transition showing up in the market
The causal reading is Editorial inference: the R-410A to A2L transition under the post-January 1, 2025 refrigerant mandate, tariffs, high interest rates, the end of the federal tax credit, and inventory destocking arrived together, and a policy-and-technology transition can stall adoption even when cold-climate COP is good enough. The longer series still points the other way: heat pumps outsold gas furnaces by 11% in 2025, took 47% of cooling-equipment sales, and outsold one-way air conditioners for the first time in October and December 2025 (Measured, AHRI).
The optimistic view, with conditions
The compressor is no longer the whole story
Cold-climate equipment can now do useful work below 0°F. The next curve is the share of installations that deliver that performance without oversized backup, comfort failures, panel surprises, or excessive refrigerant loss.
Flatten the cold curve
Broader modulation, vapour injection, larger coils, smarter defrost, and lower-GWP fluids improve capacity and COP across conditions.
Reduce the lift and load
Air sealing, insulation, ducts, and low-temperature emitters move the operating point instead of asking the compressor to overcome it.
Coordinate the peak
Staged backup, thermal storage, service planning, and grid-responsive control make electrified heat a manageable load rather than a synchronized spike.
There is more than one finish line
- Cold-capableThe compressor runs at the local design temperature.
- Capacity-sufficientIt meets most or all design load without resistance heat.
- Field-efficientDucts, charge, controls, and defrost preserve the laboratory advantage.
- Grid-compatiblePeak demand, backup staging, and service capacity are planned.
- ScalableInstallers can repeat the result at acceptable cost with low refrigerant leakage.
Sources, method, and boundaries
Laboratory targets, certification floors, and field measurements are kept separate. COP values describe equipment or measured system performance at stated conditions; HSPF2 and seasonal COP are not interchangeable. The peak calculator is a steady-state editorial model and excludes distribution losses, diversity, thermal mass, cycling, and defrost.
- COP
- Heat delivered divided by electrical input at a stated operating condition.
- Capacity retention
- Low-temperature maximum heating output relative to rated output at 47°F.
- Seasonal performance
- Energy delivered across a heating season divided by total electrical input over that period.
- Field performance
- Measured operation of equipment embedded in a real building, including its controls and distribution.