The research aims to establish the risk of interstitial condensation and moisture accumulation within four potential replacement infill panels for timber-framed buildings: a) traditional wattle and daub; b) expanded cork board; c) a composite of wood fibre and wood wool boards; and d) hempcrete. Thermal performance and moisture content are being monitored over a minimum of 2 years, with hygrothermal conditions compared to those favourable for fungi and insects known to endanger hardwood frames. In the future, measured results will also be compared to those arising from digital hygrothermal simulation using WUFI^® Pro (Fraunhofer Institute for Building Physics, Stuttgart, Germany). The results will be used to corroborate previous research by the authors using in situ monitoring [7] and laboratory testing [8].

Initially, digital interstitial hygrothermal simulations, using the software WUFI^® pro 5.3, were undertaken to investigate hygrothermal conditions created by a range of potential retrofit solutions suggested by guidance documents [15,16]. Simulations were undertaken for a range of orientations, and six distinct UK geographical locations where a significant number of surviving historic timber-framed buildings can be found. These were Suffolk, Essex, Kent, Herefordshire, Devon and a rare northern example in Cumbria. The results suggested that whilst the material properties of the replacement infill materials, orientation and climatic conditions, all had an impact on the resulting moisture content, no prolonged exposure to hygrothermal conditions favourable for biological attack, as defined in Table 1, were identified [16]. There was, however, uncertainty over the validity of the results due to the lack of material property data of some retrofit and historic materials, and the fact that these simulations represented idealised conditions with homogeneous layers well in the heterogeneous reality.

In 2017 the lead author undertook in situ monitoring at a 16th century historic timber-framed farmhouse in Suffolk UK [7]. The building had been cement rendered externally in the 1950s and had undergone a poorly considered energy retrofit in 2005, with the replacement of lath and plaster infill panels with rigid polyisocyanurate boards (PIR). Interstitial hygrothermal monitoring over a period of a year showed favourable conditions were being met for deathwatch beetle for almost 17,000 hours, accompanied by approximately 160 hours of conditions favourable for dry rot and cellar rot [7]. A smaller number of hours were also recorded of conditions favourable for other wood-boring insects.

In order to address some of the limitations of both digital simulations and in situ monitoring, three physical mock-up replacement infill panels were monitored under laboratory conditions [8]. The panels were installed to fill apertures in the dividing wall between two climatically controlled chambers at the University of Bath’s Building Research Park. The frames were constructed from reclaimed oak and the infill materials monitored were traditional wattle and daub, a composite of wood fibre and wood wool boards [14], and expanded cork board. All panels were finished on both sides in lime render. The interstitial temperature and moisture content were monitored in the centre of the panel and at the interface between the infill and a reclaimed oak frame at three depths, 10 mm, 50 mm and 90 mm, for a period of 3 weeks under steady state conditions (external chamber 5°C/80%, internal chamber 21°C/70% relative humidity [RH]). These conditions had been defined using Glaser calculations [17] as those likely to create interstitial condensation. Following these 3 weeks, a further 2 weeks were monitored using external temperatures following a diurnal cycle (5°C/94%–12°C/61%) that more closely replicated real-life conditions. The measured climatic data from the two chambers was also used in one-dimensional (1D) and two-dimensional (2D) digital interstitial hygrothermal simulations using WUFI^® Pro 5.3 and WUFI^® 2D. The in situ measurements showed that under forced steady state conditions interstitial condensation did occur in the wood fibre/wood wool composite panel, accompanied by an increase in moisture content towards the outer face of all panels (8). This was not, however, measured to occur during the 2-week period of cyclical conditions. Whilst both the 1D and 2D simulations did predict the interstitial condensation, the increase in moisture content towards the outer face of the panels was not anticipated. Significant disagreements between simulations and measured results and between 1D and 2D simulations were also encountered.

Due to technical and financial constraints, the previously described monitoring of physical mock-up panels was limited to a total of 5 weeks. In order to repeat the experiment over a longer timescale, with panels exposed to real climatic conditions, funding was sought and gained from Historic England’s Heritage Protection Commission. There follows the presentation of the design and implementation of this experiment, accompanied by analysis of the results of the first year of monitoring.

A test cell with a controlled internal environment (internal dimensions 3.5 wide × 1.9 deep × 2.2 high) was constructed at Cardiff University. Test panels form the north wall of the cell and are exposed externally to the Cardiff climate. The dimensions of the test panels were determined using the results of a study of a representative sample of 100 historic timber-framed buildings with exposed timber frames [6]. This showed 53% to be square framed (of approximately equal width and height) and 47% close studded (tall thin panels). The average size of the square panels was 785 mm × 950 mm (width × height) and the close studded 305 mm × 1830 mm (width × height), approximately 1 ft wide by 6 ft tall. Given the configuration of the test cell, it was decided to construct close studded panels, allowing the monitoring of eight adjacent panels all at the same height above ground level (Fig. 3). This array of eight panels was constructed using reclaimed oak, and allowed for the monitoring of pairs of panels of four different infill solutions (Fig. 4).

The infill solutions were the three previously studied solutions: a) wattle and daub; b) expanded cork board; and c) composite wood wool and wood fibre; in addition to d) hempcrete, a construction technique originally developed in France in the 1980s specifically for the retrofit of historic timber-framed buildings [18] and recommended for such work in a number of publications [14,19]. Details of the infill solutions are shown in Fig. 5.

The first three panels were constructed in situ by Royston Davies Conservation Builders (Leominster, UK), whilst the hempcrete was installed by UK Hempcrete (Chesterfield, UK), both companies with a reputation for high-quality workmanship and conversant with working on historic buildings. The use of these professionals aimed to replicate as close as possible real-life scenarios. One of each pair of panels was finished internally and externally with a natural hydraulic lime plaster NHL 3.5 (Secil™ [Seciltek, Lisbon, Portugal]), whilst the other was finished in a non-hydraulic lime-hemp plaster (Ty Mawr Lime Ltd., Brecon, UK). All plasterwork was completed by a qualified plasterer who works exclusively in lime plasters.

Simulation of the test cell with the software DesignBuilder™ (Stroud, UK) showed a 1 kW heater would be sufficient to maintain an internal operational temperature of 21°C. This is provided during the heating season November–March by an oil-filled electric radiator thermostatically controlled via an InkBird^® (Shenzhen, China) ITC-306 temperature controller, with a set point of 21°C. When heating is in operation, humidification is also provided by a PurLine-Hydro 60™ (Climacity, Madrid, Spain) rotating drum cold water evaporation humidifier, controlled by an InkBird^® IHC-200 humidity controller, with a set point of 60% RH. A pedestal-mounted rotating fan is located behind both the heater and humidifier to circulate the air and avoid stratification. Outside of the heating season (i.e., April–October) the internal climatic conditions are free running, with no temperature or humidity control. This replicates the most common conditions within domestic buildings in the UK.

The interstitial hygrothermal conditions between materials were monitored at a total of 60 positions. These being, at three depths (interface of internal plaster and insulation, mid-depth, and interface of external plaster and insulation) at the centre of each panel, in the horizontal wall plate at the base of each panel, and halfway up the vertical stud at the junction with the panels finished in NHL 3.5.

Type T thermocouples were used to measure temperature (°C). Following a literature review [4, 20–23], electrical resistance was chosen as the measurement methodology for moisture content (%). This allowed continual measurement with minimal impact on the surrounding materials. A methodology based on the wood block/dowel methodology used by Dr Paul Baker at New Bolsover [3] and reported in Historic England Research Report 43-2016 [24] was followed. Pairs of stainless-steel screws were embedded at the monitoring points, set 20 mm apart along the grain. For monitoring points occurring at the junction between the infill panel and the timber frame, these were embedded directly into the oak frame (Fig. 6). For monitoring points within the depth of the panel, they were embedded in lengths of split oak lath (Fig. 7).

Insulated copper wire connects these back to a Campbell^® (Campbell Scientific, Loughborough, UK) CR1000 data logger via a AM16/32 multiplexer. Care was taken in the routing of the wires through the panels to avoid the creation of direct heat and moisture paths. The resistance between the two stainless steel screws is measured by comparing a voltage applied across the screws with that applied across a known resistance (100 kΩ resistor). A calibration exercise was undertaken comparing the resistance measured with gravimetric moisture content measurements of oak blocks at various states between saturated and oven dry. This gave the following equations (1 and 2) for the calculation of the moisture content:

(1) If   R < 0.31225   Then   M C = ( 0.1912 R ) − 0.192

(2) If   R > = 0.31225   Then   M C = ( 0.2263 R ) − 0.0271

where:

R = resistance

MC = moisture content %

The measurements must also be corrected for the effect of temperature using equation 3 [25]:

(3) M C K = ( M C + 0.567 − 0.0260 x + 0.000051 x 2 ) 0.881 ( 1.0056 ) x

Where:

MC = moisture content as measured %

MC_K = temperature corrected moisture content %

x = surface temperature +2.8°C

Measurements were recorded at 30-minute intervals. The internal temperature (°C) and RH (%) of the test cell are measured using a Campbell CS215. External temperature (°C), RH (%), precipitation (mm), air pressure (mbar), wind speed (m/s) and wind direction are measured using a Vaisala (Vantaa, Finland) Weather Transmitter WXT520 Series mounted on the roof of the test cell. Direct solar radiation (W/m²) incident on the test panels is measured using a Kipp and Zohnen (Delft, Netherlands) CM5 pyrometer.

To assess the thermal performance of the replacement infill panels both thermography and in situ U-value measurements were undertaken during the heating seasons of 2019/20 and 2020/21. Thermography was undertaken using a FLIR^® (Teledyne FLIR, Oregon, USA) B250 thermal imaging camera. This took place just before dawn, maximising the internal/external temperature difference and avoiding the influence of direct solar gain, on 19/02/20 and 19/11/20. In situ U-value measurements utilised Hukseflux (Delft, Netherlands) HFP01 heat flux plates and type-T thermocouples also connected to a Campbell Scientific CR1000 data logger with readings taken at 5-minute intervals.

The initial results (Figs 8–12) show an initial drying period followed by a series of wetting and drying cycles. To date, no evidence of interstitial condensation has been found, with these wetting cycles correlating with climatic measurements of wind-driven rain. The impact of two major storm events, Alex and Bella, are particularly prominent. Storm Alex was the event that saw a record-breaking wettest day on record [26]. A rapid increase in the moisture content at the interface between the external render and the infill material can be seen in all materials, with the lowest being that of wattle and daub (WDe), for those panels finished in hemp-lime (Fig. 8). It would appear that the lower moisture permeability of the other infill materials concentrates the moisture in the external render, whereas the more moisture permeable wattle and daub allows the moisture to penetrate deeper into the panel. A rise in moisture content being measured at the centre of the panel’s depth (WDc) and at the interface between the infill and internal plaster (WDi) can be seen to follow. Over time, the moisture content at the central position (WDc) becomes higher than the external position (WDe), with this being maintained until the following spring. A similar pattern of behaviour can be seen for the mid-panel positions of the panels finished in NHL 3.5 (Fig. 9).

The results for the monitoring positions located at the junction between the base of the infill panels and the reclaimed oak frame (Figs 10 and 11), show a reduced initial drying period for those panels finished in lime-hemp plaster. Following the storm-induced wetting events, a substantial increase in moisture content is recorded at the centre of the wood wool/wood fibre panel’s depth (WFc). This is followed by an increase at both the internal (WFi) then the external positions (WFe), which continues for the following months. At the centre the moisture content remains around 20%, or more in the case of the panel finished in NHL 3.5, until the spring. The perimeter detail of this panel includes a bitumen impregnated expanding strip and mastic sealant [14], which potentially traps the moisture at this point. This highlights an area for further research and underlines the challenge that this exposed junction presents for both design and workmanship.

The moisture contents at the vertical junction between panel and frame (Fig. 12) are overall lower than those for the horizontal junction (Figs 10 and 11), most probably due to gravity permitting drainage down the joint. The highest moisture content is measured at the centre of the expanded cork board’s depth. An expanding foam sealant is also used here and similarly may be trapping moisture. Further investigation is needed to compare the use of moisture permeable and non-moisture permeable solutions to this interface.

Monitoring is ongoing and it is hoped for a longer period with no major storm events that will allow analysis of the continued drying of the different infill materials.

The thermography undertaken on 19/02/20 took place between 6:40 am and 7:20 am. The conditions were as presented in Table 2. The results are compiled as a composite image, as presented in Fig. 13.

This demonstrates that the panels with the highest external surface temperature, and as such, the poorest thermal performance, are the wattle and daub, with the best performing being the expanded cork board. A small difference in external surface temperature can be seen within the pairs of panels, with those rendered in lime-hemp performing better than those rendered in NHL 3.5. The second round of thermography took place on 19/11/2020 and showed very similar findings, with no significant differences between the two rounds.

The results of the two rounds of in situ U-value measurements (Table 3) corroborated the results obtained by thermography. It had been hoped for a greater difference in moisture content between the first measurements undertaken during the initial drying period and those at the end of the year. However, as noted above, the wetting brought by the storm events resulted in very similar moisture contents. Notwithstanding, there can be seen to be an improvement in the thermal performance of the hempcrete panels over time.

Work has begun on comparing the measured hygrothermal conditions at each monitoring position with those favourable to the potential biological risks outlined in Table 1. Of the insects, the only significant risk so far identified is from the deathwatch beetle, most frequently towards the inner face of the panels. However, to date no conditions favourable to fungi have been found. As deathwatch beetles require timber to have previously been modified by fungi, the overall risk is lessened. A very small risk from house longhorn beetle and trace risks from powder post beetle and furniture beetle.