Thermal bridges

Constructional and geometrical thermal bridges, linear and point thermal bridges, thermal bypass, internal and external psi-values, dealing with thermal bridges, thermal bridge calculation

The subject of thermal bridges is, of necessity, technically complex. It is therefore important to ensure that your professional team has sufficient understanding of thermal bridges and how to avoid them, in order to ensure that the design eliminates them (or, in retrofits, minimises them).

In buildings, thermal bridges (commonly known as ‘cold bridges’), occur when a material with relatively high conductivity interrupts or penetrates the insulation layer. A thermal bridge provides a ‘path of least resistance’, allowing heat to bypass the insulation and significantly reducing its performance.

There are three reasons to avoid thermal bridges:

•  In poorly insulated buildings, the additional impact of thermal bridges is, as a percentage of the total heat loss, relatively small. As insulation levels are increased, the relative impact of thermal bridges grows dramatically, such that it would not be possible to reach the Passivhaus standard without effectively addressing them. In a retrofit aiming for the EnerPHit standard, thermal bridges need particular attention: although they cannot generally be designed out, they can be minimised.

•  Those sections of the thermal envelope where thermal bridging occurs will have lower internal surface temperatures in cold weather, resulting in increased relative humidity (RH – the quantity of water vapour in a given volume of air) on the surface, which in turn increases the risk of mould or condensation. This is a common phenomenon in poorly insulated buildings and one where the causes are often incorrectly diagnosed.

•  Thermal bridges are difficult and time-consuming for the Passivhaus Designer to calculate (as we saw in Chapter 7, the PHPP does not do this automatically). Designing them out makes it much simpler (and more cost-effective) to create an accurate energy model of a building in the PHPP. Where a thermal bridge exists and no separate thermal bridge calculation has been made, the Passivhaus Certifier will make very conservative assumptions.

When do thermal bridges arise?

Thermal bridges can form wherever there is a junction between different building components or where there are corners in the thermal envelope (usually the external walls). The different types of thermal bridge are described below.

Constructional thermal bridges

Inexperienced Passivhaus designers may think that there are insurmountable conflicts between the need to build a structurally sound building and an energy-efficient one. It is true that materials that are commonly used for their structural properties, such as steel, concrete or even wood, are not generally good insulators, and in conventional structures they often penetrate insulation at building junctions, causing ‘constructional thermal bridges’.

The image overleaf, from a 1970s house before retrofitting work, shows a reinforced concrete lintel above a veranda door. The surface temperature (represented by colour in the thermographic picture) of the lintel is similar to that of the single-glazed window frame to its right, and is a good example of a thermal bridge. Other examples of constructional thermal bridges are window-to-wall joints, wall ties in a filled cavity wall, and timber in an insulated timber-frame construction.

Example of a constructional thermal bridge: a reinforced concrete lintel above a veranda door. Image: Infrared Thermal Imaging Surveys UK

Figure 8.1 opposite shows an architectural detail and thermographic images of a wall-to-floor junction where a mistake with the detailing has left a gap in the insulation, causing a constructional thermal bridge.

In a building based on structural timber, it is very easy to create thermal bridging unintentionally. It is essential, therefore, to plan the layout of the studwork so that such effects are minimised while ensuring that structural requirements are still met. A good example of where this was not done can be seen in the photograph opposite, showing an unnecessary clustering of studwork. In particular, careless use of prefabricated elements, such as structural insulated panels (SIPs), will create unnecessary thermal bridging.

Thermal bypass

Thermal bypass can sometimes be confused with other types of thermal bridge, which is why it is mentioned here. It is different from the other effects described in this chapter. The main practical point is that insulation must always be installed without gaps!

Air can be a good insulator when it is static in a sealed space, if the space is narrow enough and if it is not perpendicular to the thermal envelope (e.g. in the sealed space of a double-glazed window). However, as soon as there is any air movement it becomes an effective carrier of heat, which can seriously compromise the performance of an insulation layer. Common examples include an unfilled cavity in a typical UK cavity wall or in poorly laid loft insulation. Thermal bypass occurs where heat is carried by convection or by ‘blow through’: the former is where air circulates within a space, the latter is where air passes through the space from the inside to the outside of the thermal envelope. Thermal bypass can reduce the expected performance of an insulated building element by 40-70 per cent.1

Thermal bridges arising from a building’s geometry

The corners of building elements cause thermal bridging simply because of their shape, unless they are intentionally designed to avoid them (as we see on pages 116-17, using the external psi-value convention, a junction is deemed to be ’thermal–bridge-free’ if its external psi-value is less than 0.01W/mK). Figure 8.2 overleaf shows a solution that meets the structural requirements of the building – in this case a single-storey structure. The dotted black line represents a flexible lining that abuts the neighbouring building and contains the cellulose insulation infill. It allows the I-beams to be recessed by 50mm. This, plus the positioning and limited number of the I-beams at the corner, eliminates thermal bridging at the building junction. Examples of geometrical thermal bridging all occur in situations where the external heat loss area is different (usually larger) than the internal heat loss area: in a wall-to-floor junction, a wall-to-roof junction, or the vertical corner joints between walls. Most geometrical thermal bridges have an element of constructional thermal bridging too.

Figure 8.1 An example of a thermal bridge on a wall-to-floor junction where a gap in the insulation is created by a brick plinth, and infrared modelling of before and after the remedy. (The remedy is likely to have been insertion of a load-bearing insulating block such as Foamglas® Perinsul.)

An example of poor timber framing, where the effect of over-designing the structural frame creates multiple thermal bridges.

Figure 8.2 Architectural detail of a corner junction of two walls that avoids thermal bridging. The photograph on the right shows how it was executed on-site. This detail relies on blown-in insulation – so consider where the access hole for the pump will be for each section of the wall.

Measuring thermal bridges

This section looks at the units of measure used to quantify a thermal bridge, as well as at the conventions for measuring them – both are important in understanding thermal bridges. The actual calculation of thermal bridges is referred to at the end of this chapter.

Point thermal bridges

Some construction details cause point thermal bridges, usually occurring in repeating patterns: for example, mechanical fixings through insulation, or structural connection points at a junction. The photo below shows thermally broken fixings, i.e. made to minimise point thermal bridging, to attach external wall insulation.

Terms explained

chi-value (χ) – the rate at which heat passes through a material that penetrates another material at a point, where the penetrating material conducts heat better than the surrounding material; measured in W/K. Used to measure heat loss in a point thermal bridge.

psi-value (ψ) – the rate at which energy passes through a length of material, measured in W/mK. Used to measure heat loss in a linear thermal bridge.

W/mK – watts per metre per degree kelvin. For linear thermal bridges, this is watts per metre length of the thermal bridge per degree kelvin temperature difference between inside and outside the thermal envelope. (W/mK is also used to measure thermal conductivity / lambda value.)

W/K – watts per degree kelvin [temperature difference between inside and outside the thermal envelope].

External insulation attached using thermally broken mechanical fixings (see insert, top right: a plastic sleeve allows the metal screw to be deeply counter sunk within the surrounding insulation).

The heat loss of a point thermal bridge is described by its chi-value (χ). Point thermal bridges are deemed by the PHI to be ‘thermal-bridge-free’ (i.e. insignificant – see overleaf) unless they are caused by steel or other penetrations with good conductivity. Where there are recurring steel thermal bridges, such as traditional metal cavity wall ties, additional thermal bridge modelling would be required and the results entered into the PHPP.

Point thermal bridges are not normally created purely as a result of geometry. One exception would be in a cone-shaped building – not a common building form!

Linear thermal bridges

Linear thermal bridges can be caused by construction details or may be formed as a result of a building’s geometry. They often occur where a length of one building element or material meets another – for example, where a wall meets a floor, roof or window. They can also exist within a wall or other building element – for example, where poor installation has left a gap between the edges of two pieces of insulation.

The former can be avoided by careful design of the building junction (see page 118); the latter by taking sufficient care to fit the insulation accurately on-site. When using board-based insulation, this means very careful cutting to ensure less than 3mm gaps. Softer roll-based and blown-in insulation requires less skill and time to achieve accurate installation. The building designer should take this into account when specifying.

The specific heat loss (i.e. the heat loss for a given length) of a linear thermal bridge is described by its psi-value (ψ).


As a rule of thumb, a junction in a Passivhaus is considered to be thermal-bridge-free if, on examination of the construction, the insulation can be seen to be continuous at a minimum of two-thirds the thickness of the insulation surrounding the junction (assuming all the insulation has the same thermal conductivity). Quantified, a junction can be declared thermal-bridge-free if the external psi-value (ψe) is 0.01W/mK or lower.

Internal versus external psi-values

There are two conventions for measuring a linear thermal bridge. They produce different psi-values in thermal bridges caused by geometry, e.g. of the type found in junctions between building elements, illustrated in the simplified wall-to-floor junction opposite (Figure 8.3). The general UK convention (outside Passivhaus) is to use internal dimensions to measure the thermal envelope and calculate heat loss, based on U-values multiplied by areas. In this case, the psi-value should also be calculated on the same assumption. Thus, internal dimensions of the junction are taken to calculate an internal psi-value (ψi). In Passivhaus, the convention is to use the external dimensions of the junction to calculate an external psi-value (ψe), which is why it is important to enter the external dimensions of the thermal envelope in the PHPP.

In all geometry thermal bridges, such as the one illustrated in Figure 8.3, there are heat losses associated with the two elements (in this example, the wall and the floor) and also the junction (in this case the corner – shown shaded in orange); these are measured using their respective U-values. In the example illustrated, the extra heat loss associated with the thermal bridge relates to the corner.

Figure 8.3 Simplified drawing of a wall-to-floor junction, showing (left) internal dimensions used to calculate an internal psi-value (ψi ) and (right) external dimensions used to calculate an external psi-value (ψe ).

The difference between measuring the external dimensions and internal dimensions is critical, because the former includes the heat loss relating to the corner and the latter excludes it.

If the corner junction contains only the same materials as the wall or the floor, its external psi-value (ψe) will be a negative figure. This negative value can be entered into the PHPP and will trim the annual heat demand; this is useful only if you are looking for savings on a project that has just missed the 15kWh/m2.a limit and where other options are too impractical or expensive.

As we see later in this chapter (page 120), once the heat loss factors described above are known, a separate calculation is needed to convert these into either internal (ψi) or external (ψe) psi-values.

The above may seem a rather laboured and perhaps counter-intuitive mathematical point, but it is raised here because the choice of convention has an important practical implication. If the general UK convention (using internal dimensions) is used, the psi-values of all building junctions have to be determined and incorporated into the energy-performance calculations in order to model the building accurately. As UK Building Regulations are updated such that psi-values need to be incorporated into SAP calculations, clearly it would be impractical to require so many psi-value calculations to be undertaken for each building scheme. The alternative is to use accredited junction details that have known internal psi-values (ψi). By contrast, if the Passivhaus external psi-value convention is used, well-designed building junctions, an example of which is shown in Figure 8.4 on page 119, will effectively be thermal-bridge-free (if their ψe is 0.01W/mK or lower. As noted above, ψe can even be negative).

Calculating psi-values, whether internal or external, is complex and time-consuming. It is therefore better to use a convention that, where junctions have been appropriately designed, minimises or eliminates the need for these values to be calculated separately. This is why Passivhaus chooses to use external psi-values.

Assumed thermal bridge value at window edges in a Passivhaus

As we saw in Chapter 7, an installation psi-value (ψinstallation) has to be entered into the PHPP to reflect the additional losses between the window frame and the wall. (This is discussed further in Chapter 11.) If no thermal bridge calculations have been done, and provided the window installation detailing can be shown to have been adequately addressed, the Passivhaus Certifier will require a conservative ψinstallation of 0.04W/mK to be entered into the PHPP. If the detail looks good, the Passivhaus Designer may be able to argue for a lower assumed value. However, it is better to design assuming a ψinstallation of 0.04W/mK and to aim for an overall annual specific heat demand of 12kWh/m².a or 13kWh/m².a. This avoids the risk of having to do additional thermal bridge calculations to try to scrape through at just below the 15kWh/m².a limit for Passivhaus Certification.

Strategies for dealing with thermal bridges

Clearly, then, it makes sense to design out or minimise thermal bridges. As explained above, achieving this requires careful design or ‘architectural detailing’ of building junctions. Currently, however, there are very few, if any, off-the-shelf architectural details that designers can use to create thermal-bridge-free junctions. The Austrian publication Passivhaus-Bauteilkatalog [Details for Passive Houses], also known as the IBO book (see Resources), which is written in both English and German, provides helpful guidance and many examples. These details show the psi-values associated with them, which can be useful when assessing whether your own details are broadly acceptable. However, some of them are for types of construction that are not currently common in the UK and other English-speaking countries. And, even if they were, the book does not provide a ‘copy-and-paste’ solution for designers. In reality, there are many variables specific to any given project, which would often make straight copying of solutions impracticable. There are also useful example details contained in the AECB Gold Standard design guidance (see Resources).

Thermal bridging in retrofits

In most retrofits, some thermal bridging will be unavoidable. Where the building is to be externally insulated, there will be thermal bridging at the floor–wall junction, because there will always be a gap between the under-floor insulation and the external wall insulation. There is nothing that can be done about this, short of adopting extreme measures such as propping up the walls and replacing sections around the floor perimeter with an insulation material that has suitably high compressive strength. In most projects, this would make no sense financially.

Thermal bridging is also likely to be a problem at the roof–wall junction. In both cases, a bespoke architectural detail will be needed to minimise it. Figure 8.4 opposite, from the Totnes Passivhaus retrofit project, which is illustrated in the on-site photograph on page 120, show how the potential roof–wall junction thermal bridge was eliminated. In this case, the roof was new, the building having been extended upwards; however, the wall construction was a continuation of the existing twin-leaf cavity wall (using concrete block work).

(The above diagram into is divided into two parts to make the annotations more legible.)

Figure 8.4 Architectural detail of roof–wall corner junction at the Totnes Passivhaus.

View of roof-wall corner junction from above, shown in the architectural detail on the previous page.

In a retrofit, the conflict between structural demands and the need to minimise thermal bridging is even greater than in a new build, because the building will not have been designed with thermal bridge avoidance in mind. A retrofit not only demands more of the architect and the Passivhaus Designer but also requires a pragmatic approach. This is reflected in the PHI’s EnerPHit standard (see Chapter 1).

Calculating thermal bridges

If your design necessitates the calculation of a thermal bridge, there are a few eye-wateringly expensive software packages on the market designed to model even the most complex thermal bridge junctions in 3D. However, these are beyond the scope of this book and are intended for use by specialist mechanical and electrical (M&E) consultants; in any case, such software is overkill for the types of thermal bridging problems faced in most domestic- and community-scale building projects.

For architects and Passivhaus Designers wishing to do their own thermal bridge calculations (bravely or foolishly, depending on your view!), there is currently one quite old but free piece of software called THERM,2 which can be used to calculate figures from which psi-values can then be calculated. It is used in the USA for calculating the thermal performance of windows, but it can be applied to other building elements. Although the software is free, it is wise to spend money on a one-day course3 to learn how to use the output from THERM to calculate an accurate psi-value in accordance with the relevant EN standards. Although the software was designed very well in terms of its efficient use of computing power, it is not at all intuitive, so the course will save time spent on a slow and frustrating learning curve.

Once a psi-value has been calculated, this can be entered into the PHPP along with the length of the thermal bridge in metres; this allows the additional heat loss associated with the thermal bridge to be incorporated into the PHPP’s energy model.


Thermal bridges are caused when a gap in insulation allows heat to ‘short-circuit’ or bypass it. This occurs when a material with relatively high conductivity interrupts or penetrates the insulation layer. Most buildings have multiple thermal bridges because of the need to build for sufficient structural strength (constructional thermal bridges) or because of corners (geometry thermal bridges). However, with care, in a new build, it is possible to build a structurally sound building and also to eliminate thermal bridging. In retrofits, while elimination of all thermal bridging will be impractical, the same knowledge can be used to minimise it.

Most thermal bridges (whether caused by the building’s construction or its geometry) are linear and are thus quantified by a psi-value. Some thermal bridges are measured as points (quantified by a chivalue): these are almost always caused by construction rather than by geometry.

Once the concept of thermal bridging is understood, the habit of addressing construction details to avoid it can quickly become part of normal working practice. With experience, you will develop knowledge of acceptable compromises where elimination is impossible. In the end, designing without thermal bridges is simply good common sense and is good practice in terms of avoiding moisture issues (by eliminating cold surfaces) as well as making the building thermally efficient. It also makes it easier to model a building’s energy use in the PHPP accurately, as no separate calculations are needed.