Defrost Dynamics: The Latent Heat Penalty of Air-Source Heat Pumps in Maritime Climates

The coefficient of performance printed on a heat pump datasheet is a snapshot of a machine in a controlled chamber. The coefficient of performance achieved on a damp December morning in Manchester, Cardiff or Glasgow is a different quantity, shaped by a process that the datasheet barely mentions: frost. Every air-source heat pump cools outdoor air as it extracts heat, and in the wrong combination of temperature and humidity that cooling deposits ice on the evaporator. The ice insulates the coil, restricts airflow and starves the refrigerant circuit, so the machine must periodically stop heating the building and melt the ice. This article examines the thermodynamics of that process, quantifies the energy it costs, and audits the extent to which the standard test methods capture it, for clean-heat specifiers, building services consultants and microclimate deployment analysts who need to model real winter behaviour in the maritime UK.

The Micro-Physics of Frost Formation in Maritime Air

A heat pump evaporator is a finned heat exchanger through which a cold, low-pressure refrigerant evaporates, absorbing heat from air blown across the fins by a fan. To transfer heat at a useful rate the refrigerant must be colder than the air, and the practical temperature difference between them is typically 5 K to 10 K at the design condition. When outdoor air is at 5°C, the evaporating refrigerant temperature is therefore around -5°C to 0°C, and the fin surface temperature, which sits between the two, is frequently below 0°C. The surface does not need to be cold all over for frost to form, since the coldest portions near the refrigerant inlet will frost first and the frozen region will spread from there.

Whether the fin surface collects moisture depends on the dew point of the incoming air. Air at 5°C and 90 per cent relative humidity has a dew point of around 3.5°C. Any surface cooler than that will condense water vapour from the air, and if the surface is also below 0°C, the vapour deposits directly as frost (desublimation) or condenses and then freezes on contact. The result is a growing layer of ice and, in the early stages, a low-density, porous frost with air trapped in its structure. Early frost can actually increase surface area and slightly improve heat transfer, but as the layer thickens and densifies, thermal conductivity worsens and the airflow passages between fins narrow.

The amount of water available matters as much as the temperature. The moisture content of air is a steep function of temperature. Saturated air at 5°C carries around 6.8 grams of water vapour per cubic metre, while saturated air at -5°C carries only around 3.3 grams, and at -15°C only around 1.4 grams. Cold, dry continental air therefore contains very little moisture to freeze, and heat pumps in such climates may frost lightly even at low temperatures. Maritime air, which has crossed a relatively warm ocean, arrives humid, with relative humidity often above 85 per cent for days at a time. In the band of ambient temperature from roughly 2°C to 7°C, the air holds a large quantity of moisture, yet the evaporator surface is cold enough to be below freezing, so the coil is simultaneously sub-zero and exposed to a heavy moisture load. This is the worst-case operating boundary, and it describes much of the British winter.

The mass flow rate is substantial. A domestic heat pump moves air in the region of 2,000 to 5,000 cubic metres per hour across the coil. If the coil removes just 1 to 2 grams of water per cubic metre from this air stream, the unit collects several kilograms of ice per hour. The latent heat of fusion of water is about 334 kilojoules per kilogram, and the heat of desublimation, from vapour directly to ice, is about 2,830 kilojoules per kilogram, so every kilogram of frost that forms also adds a significant thermal load to the evaporator, which is initially helpful (the latent heat release warms the surface) but rapidly becomes obstructive as the frost thickens. The system also experiences a progressive drop in evaporating pressure and temperature as the coil becomes less effective, which in turn lowers the capacity of the unit and increases the compressor's work for each unit of heat moved.

Micro-climate adds another layer of variability. Local conditions such as proximity to rivers, coastlines, low-lying cold air pools, or dense vegetation influence humidity and temperature at the unit. A unit sited in a damp, enclosed courtyard may see persistent high humidity, and one mounted close to a wall may recirculate cold, moist discharge air. Wind speed and direction affect the supply of fresh air to the coil. Defrost frequency can therefore vary substantially between two apparently similar installations in the same region, which is a major limitation on generalising laboratory results to a specific site.

The Reverse-Cycle Defrost Mechanism

A heat pump is a closed refrigerant loop with four principal components: a compressor, a condenser, an expansion device and an evaporator. In heating mode, the compressor raises the refrigerant to high pressure and temperature, the hot gas gives up heat in the condenser (the plate heat exchanger connected to the building's water circuit), the liquid is expanded to low pressure, and the cold two-phase refrigerant absorbs heat from the outdoor air in the evaporator. A four-way reversing valve allows the direction of refrigerant flow to be switched.

When the control system decides to defrost, it actuates the reversing valve. The hot discharge gas from the compressor is now routed to the outdoor coil, which acts as a condenser, and the heat it releases melts the ice. The plate heat exchanger on the water side now acts as the evaporator, and it absorbs heat from the water in the heating circuit. The unit has effectively become a small chiller for the duration of the defrost. The outdoor fan typically stops during this phase, so that the heat is retained at the coil. Meltwater drains away, the coil warms through 0°C, and when the coil temperature or pressure indicates that the ice is cleared, the valve reverses again, the fan restarts, and heating resumes after a short stabilisation period.

The energy accounting of this event has three components. The first is the energy needed to melt the ice, which is the mass of ice multiplied by the latent heat of fusion. The second is the sensible energy needed to heat the ice to its melting point and then to warm the coil mass itself, which comprises aluminium fins and copper tubes with a combined mass that can reach several kilograms and which must be raised from below zero to a few degrees above zero. The third is the loss during the transition: the period before and after the defrost when the heat pump runs at a reduced capacity because the coil is partially iced, and the recovery period when the evaporator refrigerant temperature falls back to its operating value. Together these typically amount to a few megajoules per event, though this varies with unit size and frost load, and a defrost may last from three to ten minutes. During this time, the building receives no heat from the heat pump, and in fact loses it.

The heat that melts the ice comes from the hydronic circuit. The circuit's water volume is the thermal reservoir, and every kilojoule drawn to melt the evaporator is a kilojoule removed from the water in the radiators, underfloor pipes, buffer vessel and cylinder. The temperature drop is the energy extracted divided by the product of the water mass and its specific heat capacity of 4.19 kilojoules per kilogram per kelvin. In a system with 200 litres of water, extracting 2 megajoules lowers the water temperature by around 2.4 K. In a low-volume system with 40 litres, the same extraction would drop the temperature by nearly 12 K, and the flow temperature seen by the radiators falls sharply. In practice the control system may protect itself by limiting or aborting the defrost if the return temperature drops too far, but this risks an incomplete defrost and a quicker return to the iced condition.

This explains why the water volume of the circuit is a design parameter, not an afterthought. Systems with insufficient volume, those with zoned radiators that close down most of the circuit, or those with undersized buffer vessels, will experience noticeable temperature dips in the room during each defrost, and the heat pump will have to work harder afterwards to recover, which consumes extra electricity. Some units mitigate this with a dedicated defrost heater, or by drawing from a buffer vessel or the domestic hot water cylinder as an alternative thermal store, and some use hot gas bypass defrost, which diverts a portion of the discharge gas to the coil without fully reversing the cycle, but the reversing valve method remains the most common in domestic products.

The control logic that triggers defrost is itself a source of efficiency loss. The simplest approach is a timer, which defrosts at fixed intervals regardless of actual frost accumulation. This can waste energy by defrosting a clean coil, or fail to defrost a heavily iced coil. More sophisticated demand-defrost controls use sensors: the temperature difference between the air and the evaporator, the evaporating pressure, the fan current, or a combination, to infer ice build-up and defrost only when needed. Their effectiveness depends on calibration and on the quality of the sensors. An incorrectly calibrated system may defrost too early, too late, or too frequently, and the resulting loss can be a significant contributor to poor field performance.

The Energy Penalty in Operation

The impact on performance can be separated into several effects. The first is the direct energy cost of the defrost cycle, in which the compressor consumes electricity while delivering negative useful heat to the building. The second is the loss of heating output during and after defrost, which has to be made up later at a potentially higher flow temperature, reducing the average COP. The third is the degradation of performance in the period of frost accumulation before defrost, when the evaporating temperature drops and the unit's capacity and COP decline progressively. The fourth is any backup heating that may be activated if the heat pump cannot meet the demand during recovery.

The magnitude depends heavily on the climate and the unit. In field studies and manufacturer data, defrost losses are commonly reported as lying in the range of a few per cent of annual heating energy in dry or cold climates, rising to figures in the order of five to ten per cent or more at the peak defrost conditions in humid climates around 0°C to 5°C, though specifiers should confirm figures from product test data and independent monitoring. The loss is not uniformly distributed across the year. It is concentrated in the weeks when the ambient temperature and humidity fall into the frost-forming band, and during those weeks the instantaneous COP can fall markedly below the datasheet figure for the same ambient temperature.

The distribution matters for the cost of heat as well as its quantity. Because the UK maritime winter spends a large proportion of its hours in the 0°C to 8°C band, and because the heating demand is substantial in those hours, the frost-forming conditions coincide with a large share of the annual energy. A site that appears to have a mild climate on the basis of mean temperature may still face a heavy defrost burden, as the mean conceals the dwell time in the critical band. Analysts should therefore examine the hourly distribution of temperature and humidity from a local weather file, not merely the heating degree days.

Auditing the Standard: EN 14825 and the Seasonal Performance Factor

Heat pump performance claims in Europe rest on a family of standards. EN 14511 defines the test conditions and measurement methods for capacity and COP at fixed operating points, and EN 14825 defines how those points are weighted across a heating season to produce a seasonal coefficient of performance (SCOP) and the related seasonal space heating efficiency used on energy labels. The seasonal calculation applies a bin method: the reference climate is divided into temperature bins, each with a number of hours and a load derived from a linear load profile, and the unit's performance is interpolated from test points at several outdoor temperatures and flow temperatures. For air-source units, the standard test points include an outdoor condition of 2°C dry bulb and 1°C wet bulb, which is a frost-forming condition and for which the test method requires that defrost cycles are included in the measurement.

The standard is a reasonable foundation, but several limitations prevent it from capturing field behaviour in a maritime climate, and the specifier should be aware of them:

• Limited humidity coverage. The standard prescribes a small number of humidity conditions. Real frosting depends on continuous variation in humidity, and a unit tested at one humidity at 2°C may behave very differently at 6°C and 95 per cent relative humidity, which is a common UK winter condition.
• Steady-state character. Test points are taken under controlled, quasi-steady conditions. Real weather fluctuates, and thermal lag in the building, the water circuit and the evaporator coil means the machine is rarely in the equilibrium assumed by the test.
• Climate assumption. The reference climate used for the headline seasonal figure is a single average European climate, and the weighting of temperature bins does not match the distribution in a specific maritime location. Standards also define warmer and colder reference climates, but few products publish results for all of them.
• Load profile. The bin method assumes that heating load falls linearly with outdoor temperature and ignores internal gains, solar gains and occupant behaviour. It also assumes the unit matches load perfectly or cycles in a prescribed manner, which does not capture real cycling losses, short-cycling when oversized, and control interactions.
• Defrost frequency. A defrost in the test is measured under the specific conditions of the test chamber, which may not reproduce the frost loading of a real coil, the effect of wind, recirculation or snow, and which does not exercise the manufacturer's demand-defrost logic across the full range of field conditions.
• System boundary. The standard concerns the heat pump unit itself, not the installed system. Hydraulic losses, buffer vessels, distribution temperatures, controls, pumps and domestic hot water cycles are excluded, yet they strongly influence the real seasonal performance factor (SPF) at the system level.
• Parasitic loads. Crankcase heaters, trace heating on condensate drains, standby consumption and the electrical demand of pumps and controls are treated unevenly, and they can add noticeably to winter consumption.

It would be inaccurate to say that the standard ignores defrost altogether, since measurements at the frost-forming test points do include defrost cycles. The fair characterisation is that EN 14825 captures defrost in a limited, standardised way at selected points, whereas real maritime winters present a continuous range of frost conditions, intermittent cycling and system-level interactions that a small set of test points cannot represent. Consequently, SCOP figures should be treated as a comparative index between products tested under the same method, not as a prediction of the running cost at a specific address. Independent field trials in the UK have repeatedly shown that measured seasonal performance varies widely between installations of the same model, and that the variation is explained far more by system design and commissioning than by the datasheet number.

Design Responses and Mitigation

The thermodynamic picture implies a set of practical measures. None eliminates frost, since frost is an inevitable result of extracting heat from humid air, but each reduces its cost.

• Provide adequate water volume. Ensure the circuit has sufficient volume, through emitter design, a correctly sized buffer or a dedicated defrost store, to absorb a defrost without a perceptible fall in room temperature or a defrost abort.
• Size the system sensibly. Oversized units cycle more and can hold the coil at a lower temperature for longer. A unit matched to the design load runs for longer at a lower evaporating depression, which reduces frost accumulation.
• Run low flow temperatures. A lower flow temperature means a lower condensing temperature and less compressor lift, which allows a higher evaporating temperature for the same capacity, and a smaller frost rate.
• Site the unit carefully. Provide generous clearances, avoid enclosed courtyards and cold air sumps, orient the unit to avoid recirculation, and ensure good condensate drainage so that meltwater does not refreeze beneath the unit.
• Check the defrost control settings. Where the controller allows, review the defrost initiation and termination parameters, and verify from monitoring data that the defrost frequency is consistent with the observed ambient conditions.
• Monitor the installation. Log electricity consumption, heat output, flow and return temperatures and ambient conditions, so that defrost behaviour can be analysed and an underperforming site can be identified.
• Use local weather data in the model. When forecasting running costs, use hourly temperature and humidity data for the site rather than relying on the manufacturer's SCOP alone, and apply a defrost allowance appropriate to the climate.

The Comparative Matrix

The matrix below characterises the principal operating bands for a typical domestic air-source heat pump in a maritime climate. The thresholds are indicative and depend on the design of the coil, the refrigerant circuit and the control strategy. The COP impacts are qualitative, since quantitative values vary by unit and site.

Engineering Verdict and Specification Guidance

The latent heat penalty is a physical consequence of extracting heat from humid air at temperatures near and just above freezing, and it is concentrated in precisely the conditions that dominate a maritime winter. It is not a defect of any particular product, and it cannot be designed out entirely. But it can be characterised, modelled and mitigated, and its cost can be reduced by good system design. The following principles are recommended for specifiers and consultants:

• Treat datasheet SCOP as a comparative index. Use it to rank products, but model site running costs from local hourly weather data and an explicit defrost allowance.
• Design the hydraulics for defrost. Verify the system water volume and buffer arrangement against the defrost energy draw, and confirm that the controller will not abort defrosts for lack of volume.
• Prefer low flow temperatures and correctly matched capacity. Both reduce frost formation and improve the efficiency of the unit across the season.
• Treat siting as a thermodynamic decision. Local humidity, airflow, recirculation and drainage affect the defrost burden as much as the choice of model.
• Insist on monitoring. A logged installation can reveal whether defrost frequency and energy consumption are consistent with expectations, and permits a tuned control strategy.
• Communicate the limitations to clients. Present expected running costs as a range, explaining that winter spells in the frost-forming band will produce visible defrost events and temporarily lower COP.

In summary, the defrost cycle is an unavoidable part of air-source heat pump operation in the UK, and a specialised understanding of its thermodynamics separates reliable forecasts from optimistic ones. The standardised laboratory methods remain a useful basis for comparison, but they are not a substitute for localised analysis. Specifiers who characterise the micro-climate, protect the hydraulic volume and monitor the result will deliver systems whose performance matches their promises, which is the essence of sound asset-optimisation in clean heat.