Laboratory Measurement and Boundary Conditions for the Water Vapour Resistivity Properties of Typical Australian Impermeable and Smart Pliable Membranes

: The duo of better insulated and more air-tight envelopes without appropriate consideration of water vapour diffusion and envelope moisture management has often demonstrated an increased potential of moisture accumulation, interstitial condensation, and mould growth within the building envelope. To inform a resilient, energy efﬁcient, and healthy building design, long-term transient hygrothermal modelling are required. Since 2008, concern has been raised to the Australian building regulators regarding the need to establish the vapour diffusion properties of construction materials, in order to develop a hygrothermal regulatory framework. This paper discusses the results from laboratory testing of the vapour diffusion properties of two common reﬂective pliable membranes, and one smart pliable membrane. The two reﬂective pliable membranes are often used within the exterior walls of Australian buildings. The smart pliable membrane is a relatively new, internationally available product. The three membranes were tested as per ISO 12,572 at 23 ◦ C and 50% RH. To establish if the vapour resistivity properties were constant, under different relative humidity conditions, the membranes were further tested at 23 ◦ C and relative humidity values of 35%, 65%, and 80%. The results of the three pliable membranes show that the vapour resistivity properties varied in a non-linear (dynamic) manner subject to relative humidity. In conclusion, this research demonstrates that regardless of the class, each of the tested membrane types behaved differently under varying relative humidity and pressure gradients within the testing laboratory.


Introduction
The aim of this research was to investigate whether the single point vapour resistivity test method as described in ISO 12572 and ASTM E 96m provides adequate data to inform building envelope hygrothermal simulation. The current standard only requires a single point measurement for materials tested at 23 • C and 50% relative humidity (RH). The incorporation of high-quality material property inputs in the hygrothermal simulation has been identified by many as critical, which may significantly impact the moisture and mould risk calculations. This article reports on the observed water vapour resistivity properties of two impermeable pliable membranes, and one smart pliable membrane, when tested under different relative humidity conditions (humidity-dependent), in order to plot multiple point, rather than single point, hygrothermal boundary curves for each of these materials.
The combination of thermal insulation and airtightness without appropriate consideration of the external envelope's ability to manage water vapour diffusion and moisture have been identified as key contributors to moisture, moisture accumulation, and mould permeable reflective pliable membranes are tagged specimen D and E, respectively. These membranes are commonly applied to the external face of the timber or steel frame (outside the insulation layer), of low-rise buildings, to improve air tightness (thermal performance), as a water barrier (façade durability), and as a vapour control layer (condensation and mould) in walls and roofs of buildings. The two types of impermeable membrane were reflective foil products from different manufacturers. They both had similar polymer characteristics and were coated with aluminium foil on the surface of one side. The smart membrane, which is commonly used in walls and roofs is a polyethylene copolymer product with varying water vapour diffusion properties, subject to the air moisture content. The total measurement period for all pliable membrane testing took fourteen months, which involved four experiments during which the temperature was maintained at 23 • C and the relative humidity conditions of the hygrothermally-controlled room were maintained at 35%, 50%, 65%, and 80%, respectively. Since the three membrane types discussed here are not readily open to the water vapour diffusion process, it took an average of three and a half months to establish a moisture equilibrium state for each specimen for each period of relative humidity conditioning.

Boundary Conditions
The experiments are based on the principle of diffusion of water vapour from an area with higher partial pressure of water vapour to an area with lower partial pressure. In these experiments, the vapour drive was established as a result of the pressure differences between the relative humidity within the internal air space of the test dish and the relative humidity within the internal condition of the hygrothermal testing room. This creates partial vapour pressure differences which cause water vapour diffusion to occur, which is similar to what the external building envelope may be experiencing in real life. The production and control of the testing room for these experiments has been discussed in previous publications. To avoid both air temperature and humidity stratification within the test room, a fan was in operation throughout the entire measurement period. This fan generated an air velocity of approximately 0.2 m/s which stabilised the pressure within the room. The average air temperature of the room was maintained at 23 • C throughout the four periods of measurement with ±1 • C variation. The relative humidity and their vapour pressure differences, which represent the targeted boundary condition for the four periods of the experiment, are shown in Table 1. This table was adapted from ISO 12,572 with additional testing points that were identified based on common climatic condition adopted for the four testing periods in this research. This table is important as it provides input data for the calculation and tabulation of the vapour resistivity properties of materials involving multiple moisture-dependent variables, which are needed to plot the hygrothermal curves that reflect different boundary conditions. In each of the experiments, the relative humidity in the hygrothermally-controlled room was carefully controlled. The relative humidity was maintained between 35.0% to 36.9%, with an average of 36% in the first test period. In the second test period, the relative humidity was maintained between 49.8% to 50.8%, with the average humidity of 50.4%. The relative humidity was maintained between 64.5% to 65.2%, with an average relative humidity of 65.12% in the third test period. During the fourth test period, the relative humidity was maintained between 77.84% to 83.2% with an average relative humidity of 80. 29%. The details about the temperature and relative humidity performance of the chamber with respective variation has been reported in the previous publication [13].

Experimenal Procedure
The gravimetric procedure employed in this research followed the guidelines of the international standard ISO 12,572 [25]. The procedure involved completing both wet cup and dry cup gravimetric water vapour transmission measurement for samples from the three membranes. In each testing period, ten specimens were cut from the smart membrane (C), and another ten specimens from each of the two types of impermeable membrane (D). For each of the pliable membranes, five of these specimens were for wet cup testing and five specimens were for dry cup testing. Due to the amount of sample material provided, only six specimens were prepared for the second type of impermeable membrane (E), representing three specimens for wet cup and three specimens for the dry cup test. The test dishes comprised a round glass dish with 200 mm diameter and a 60 mm depth. The thickness of each specimen was measured by a digital micrometre screw gauge over ten different points on the surface of each specimen and the mean value was determined and recorded for later use during the water vapour diffusion calculation process. The specimens were precisely cut to the dish size (200 mm) such that they fit to the edge of the mouth of the dishes. The desired relative humidity within the cups were achieved by the use of dry and wet substrates. This research adopted silica gel as dry cup substrate because it was very stable at 3% relative humidity. The wet cup test achieved the desired relative humidity of 93% through the use of anhydrous ammonium dihydrogen phosphate solution at 23 • C. The substrates were gently poured into the dishes until an airspace of 20 mm was established between the top of the substrate and the top of the test dish. The specimens were then sealed to the edge of the test dishes with glue, followed by tightly wrapping the edge of the dishes with paper tape. To achieve adequate vapour seal, paraffin wax of 6:4 ratio and melting at 58-60 • C was gently applied around the paper tape with an artist's brush until the paper tape was no longer visible and the molten wax was allowed to harden. The gravimetric measurement began once a set of five specimens were completed for a particular test. Given that the temperatures inside the cup and outside the cup are the same, partial vapour pressure differential was achieved by the difference between the conditioned room's relative humidity and the wet or dry substrate within each dish, causing water vapour diffusion through the test specimen. The amount of water vapour diffusion was established by periodically measuring the weight of the cup assembly. The measurement of the dish weight continued until a steady state was reached. However, due to these membranes being vapour impermeable, after an initial period of regular measurements, the time between measurements was expanded to once a week, for a period of three months, and the diffusion properties of the specimens were determined through mathematical calculations.

Results
The result for the water vapour resistivity properties for each specimen tested was calculated from the mathematical equations provided by the ISO 12,572 [25]. The mathematical calculation and procedure for obtaining various iterative resistivity properties using specimen C 1 as an example is presented in the Appendix A. This calculation procedure was repeated for all the specimens, for all the boundary conditions (35%, 50%, 65%, and 80%). The detail results for the properties of each membrane tested and the statistical analysis for all the four testing periods are tabulated in Appendix C Tables A2-A13. Table 2, below, shows the average barometric pressure and the calculated vapour permeability of air for the four testing periods. Table 3 shows the summary of results for the water vapour permeability and permeance, in dry cup and wet cup scenarios, from the 26 tested materials. Similarly, the summary of the average water vapour resistance factor µ and diffusion air layer thickness S D for all the tested specimens is tabulated in Appendix B. The air gap resistance used for calculating the water vapour resistance factor for each of these membranes was calculated by multiplying the initial resistance factor by the mean thickness of each membrane to get the initial equivalent air layer thickness S D . This is then followed by deducting 20 mm air gap from the initial equivalent air layer thickness to obtain the final S D values. This final S D value is in turn divided by the mean thickness resulting into the final vapour resistance factor (also see Appendix A for details). Tables 4 and 5 shows the summary of hygrothermal water vapour resistance factor and equivalent air layer thickness across the boundary conditions over the average relative humidity for the three tested materials respectively, which were used to plot the hygrothermal boundary curves for specimen C, D, and E after harmonic adjustment.

Discussion
The aim of this research was to investigate the diffusion behaviour of water vapour impermeable, semi-impermeable, and permeable pliable membranes under different relative humidity boundary (moisture-dependent). This article focusses on the measured and calculated water vapour permeability values from two types of water vapour impermeable and one type of semi-impermeable membrane. A previous article reported the results from the measured and calculated water vapour diffusion properties of tested water vapour permeable membranes [13]. Firstly, it is important to state that this research successfully kept the temperature in the hygrothermal chamber stable at 23 • C with ±1 • C variation throughout all the four testing periods. Therefore, the shape of all the samples from the three tested pliable membranes observed was flat throughout the four testing periods, suggesting that there were no significant impact caused by temperature variation. Variation in temperature may alter the shape of membrane, which would affect the surface area calculations, which is very important when considering temperature-dependent measurement.
In Figure 1, the blue colour in the plot show the values for the water vapour resistance factor against the hygrothermal boundary conditions established from the average relative humidity conditions from the four laboratory measuring periods for Specimen C, which is classified as a class 2 (semi-impermeable membrane). In these measurements, the results indicate that the specimen was exposed to different relative humidities on both dry and wet sides. This difference in vapour pressure normally generates water vapour flow through the specimens, and as the humidity varies along the cross-section of the specimens, say from wet side to the dry, the vapour resistance also varies. Figure 1 shows that specimen C has different µ-values. From approximately 20% RH to 40% RH, the rate of water vapour diffusion varies significantly. Whilst from 40% RH to 80% RH, the water vapour diffusion rate is relatively constant. The vapour flow thus contains information about all the µvalues corresponding to the applied humidity interval as shown in Figure 2. This indicates that the water vapour diffusion resistivity properties of specimen C are non-linear as the amount of water vapour diffusion which passed through the specimen encounters different resistances, subject to the level of humidity. The graph also shows that the µ-values indicate an inverse relationship as the strength of the resistance factor decreased with increases in relative humidity.
Similar behaviour is observed from Figure 2, the blue colour is the plot of values of the equivalent air layer thickness against the established boundary conditions of the four experiment testing periods for specimen C. This also indicates that the water vapour flow across the cross-section of the specimen causes the equivalent air layer thickness to decrease in higher relative humidity, as the S D value is high at lower relative humidity and lower in higher relative humidity in an inverse and dynamic pattern. Even though this dynamic behaviour is consistent with previous findings, it further supports the suggestion that the hygrothermal properties of construction materials should be considered under different relative humidity conditions. In Australia, hygrothermal diffusion function of pliable membrane is not yet defined under the current AS4200:1, and determining the appropriateness of pliable membranes according to their diffusion properties in different relative humidity conditions has become contestable among design professionals and researchers. Moreover, there are no appropriate and recent hygrothermal data for construction materials, typically used in Australia. Therefore, the ability to determine the hygrothermal boundary conditions of specimen C, with the resistance factor and equivalent layer thickness, is essential for accurately fulfilling the hygrothermal modelling pathway for moisture management introduced to NCC in 2019.
vapour diffusion rate is relatively constant. The vapour flow thus contains information about all the μ-values corresponding to the applied humidity interval as shown in Figure 2. This indicates that the water vapour diffusion resistivity properties of specimen C are nonlinear as the amount of water vapour diffusion which passed through the specimen encounters different resistances, subject to the level of humidity. The graph also shows that the μ-values indicate an inverse relationship as the strength of the resistance factor decreased with increases in relative humidity. Figure 1. Plot of resistance factor against relative humidity prior harmonic curve adjustment for the three specimens. Figure 1. Plot of resistance factor against relative humidity prior harmonic curve adjustment for the three specimens. Similar behaviour is observed from Figure 2, the blue colour is the plot of values of the equivalent air layer thickness against the established boundary conditions of the four experiment testing periods for specimen C. This also indicates that the water vapour flow across the cross-section of the specimen causes the equivalent air layer thickness to decrease in higher relative humidity, as the SD value is high at lower relative humidity and lower in higher relative humidity in an inverse and dynamic pattern. Even though this dynamic behaviour is consistent with previous findings, it further supports the suggestion that the hygrothermal properties of construction materials should be considered under different relative humidity conditions. In Australia, hygrothermal diffusion function of pliable membrane is not yet defined under the current AS4200:1, and determining the appropriateness of pliable membranes according to their diffusion properties in different relative humidity conditions has become contestable among design professionals and researchers. Moreover, there are no appropriate and recent hygrothermal data for construction materials, typically used in Australia. Therefore, the ability to determine the hygrothermal boundary conditions of specimen C, with the resistance factor and equivalent layer thickness, is essential for accurately fulfilling the hygrothermal modelling pathway for moisture management introduced to NCC in 2019. Similarly, the plot of the vapour resistance values against the relative humidity boundary condition for specimen D and E were indicated by the red colour and green, respectively, in Figure 1. The red and the green colour in Figure 2 also show the plot of equivalent air layer thickness against the humidity boundary conditions of specimen D Similarly, the plot of the vapour resistance values against the relative humidity boundary condition for specimen D and E were indicated by the red colour and green, respectively, in Figure 1. The red and the green colour in Figure 2 also show the plot of equivalent air layer thickness against the humidity boundary conditions of specimen D and E respectively. Generally, impermeable pliable membranes usually have a very high vapour resistance factor and air layer equivalent thickness values as observed in Figure 2. However, it is essential to point out that even these graphs further document dynamic behaviour, similar to what this research has observed with other pliable membranes under varying relative humidity boundary conditions. The expectation was that these reflective impermeable pliable membranes would have a constant high resistance to water vapour diffusion. On the contrary, this did not occur as these reflective foil products have a rather unusual water vapour diffusion behaviour as the relative humidity conditions increased. The reason for the dynamic patterns observed in these graphs has resulted from the changes in air permeability and vapour pressure that has occurred under the four different relative humidity conditions. This level of water vapour diffusion resistance will cause these impermeable pliable membranes to have narrow usability potential for moisture-permeable construction systems.
As previously mentioned, measurements were usually taken weekly because of high water vapour diffusion resistance of these specimens. In this scenario, this research observed that when the specimens from the two types of reflective membranes were tested at 80% RH, the specimen were swelling within a week. Additionally, by the second week of weighing, there was a visible formation of mould growth on the reflective surface of all the specimens (see Figures 3 and 4). The program for this research involved testing specimens from all the classes of pliable membranes at 80% RH during the same period. The yellow-brownish colouration of mould formation on these reflective materials was not observed on any of the other materials tested.
Comparing the result of these three membranes with each other, specimen C, which is semi-impermeable and belongs to Class 2, appears to be more open to vapour diffusion process with higher vapour resistance values when the humidity was less than 40%. Specimen D and E, which are classified as Class 1, impermeable pliable membranes, have very high vapour resistance factors and equivalent air layer thickness despite the increase in relative humidity conditions. behaviour, similar to what this research has observed with other pliable membranes under varying relative humidity boundary conditions. The expectation was that these reflective impermeable pliable membranes would have a constant high resistance to water vapour diffusion. On the contrary, this did not occur as these reflective foil products have a rather unusual water vapour diffusion behaviour as the relative humidity conditions increased. The reason for the dynamic patterns observed in these graphs has resulted from the changes in air permeability and vapour pressure that has occurred under the four different relative humidity conditions. This level of water vapour diffusion resistance will cause these impermeable pliable membranes to have narrow usability potential for moisture-permeable construction systems. As previously mentioned, measurements were usually taken weekly because of high water vapour diffusion resistance of these specimens. In this scenario, this research observed that when the specimens from the two types of reflective membranes were tested at 80% RH, the specimen were swelling within a week. Additionally, by the second week of weighing, there was a visible formation of mould growth on the reflective surface of all the specimens (see Figures 3 and 4). The program for this research involved testing specimens from all the classes of pliable membranes at 80% RH during the same period. The yellow-brownish colouration of mould formation on these reflective materials was not observed on any of the other materials tested.
Comparing the result of these three membranes with each other, specimen C, which is semi-impermeable and belongs to Class 2, appears to be more open to vapour diffusion process with higher vapour resistance values when the humidity was less than 40%. Specimen D and E, which are classified as Class 1, impermeable pliable membranes, have very high vapour resistance factors and equivalent air layer thickness despite the increase in relative humidity conditions.

Harmonic Adjusment of Hygrothermal Boundary Curve
The cup measurements are used to determine the boundary condition of the water vapour diffusion characteristics of a material by plotting a curve after harmonic adjustment of the data from dry and wet cup gravimetric measurement. This is because there is the need to establish the different values of vapour diffusion at any given specified relative humidity ϕ, as the cup measurements do not directly provide these values. During the gravimetric measurement, the specimen is exposed to different relative humidities on both sides, for example, 50% on one side and 80% on the other side. The humidity varies continuously along the cross-section of the specimen, from 50% RH on one side to 80% RH on the other side, with all intermediate values also occurring somewhere in the specimen. Furthermore, the material in different parts of the specimen also has different μvalues, and all μ-values from μ (50%) to μ (80%) occur simultaneously somewhere in the specimen. However, the cup measurement can only tell us the strength of the vapour flow which diffused through the specimen. The vapour flow thus contains information about all the μ-values corresponding to the applied humidity interval, which are encoded in a single number. The task now is to determine μ-values for individual humidities. This can be done by combining the results of cup measurements using different humidity ranges. Hence, analysing gravimetric cup measurements as if the material had a constant μ-value, provides the effective μ-value of the specimen in such states of moisture. Therefore, the effective μ-value for a given humidity range is the harmonic mean of the μ(ϕ)-curve, taken over the humidity range. This is why it may be misleading to draw a μ(ϕ)-curve by simply plotting the effective μ-values against the mean applied humidity ranges from purely gravimetric measurement. Figure 5 shows a μ(ϕ)-curve (grey) and a series of mean values (orange bars) of this curve, taken over humidity ranges with the same midpoint but with increasing widths. First, due to the curvature of the μ-curve, all those mean values are lower than the curve itself at the midpoint. The curve point is the desired μ-value corresponding to the midpoint humidity. The mean values correspond to the results of cup measurements performed by applying the respective humidity ranges. The results of all cup measurements are lower than the μ-curve point corresponding to the midpoint humidity, so it would be inappropriate to simply substitute a cup measurement result for this μ-value. Secondly, the mean values for humidity ranges with the same midpoint are lower for wider ranges. This means that when cup measurements with different wide humidity ranges are combined into one diagram (plotting them against the midpoint humidity), a misleading zigzag-shape of the diagram may result, even though the underlying true μ(ϕ)-curve may be continuously falling for increasing humidity.

Harmonic Adjusment of Hygrothermal Boundary Curve
The cup measurements are used to determine the boundary condition of the water vapour diffusion characteristics of a material by plotting a curve after harmonic adjustment of the data from dry and wet cup gravimetric measurement. This is because there is the need to establish the different values of vapour diffusion at any given specified relative humidity φ, as the cup measurements do not directly provide these values. During the gravimetric measurement, the specimen is exposed to different relative humidities on both sides, for example, 50% on one side and 80% on the other side. The humidity varies continuously along the cross-section of the specimen, from 50% RH on one side to 80% RH on the other side, with all intermediate values also occurring somewhere in the specimen. Furthermore, the material in different parts of the specimen also has different µ-values, and all µ-values from µ (50%) to µ (80%) occur simultaneously somewhere in the specimen. However, the cup measurement can only tell us the strength of the vapour flow which diffused through the specimen. The vapour flow thus contains information about all the µ-values corresponding to the applied humidity interval, which are encoded in a single number. The task now is to determine µ-values for individual humidities. This can be done by combining the results of cup measurements using different humidity ranges. Hence, analysing gravimetric cup measurements as if the material had a constant µ-value, provides the effective µ-value of the specimen in such states of moisture. Therefore, the effective µ-value for a given humidity range is the harmonic mean of the µ(φ)-curve, taken over the humidity range. This is why it may be misleading to draw a µ(φ)-curve by simply plotting the effective µ-values against the mean applied humidity ranges from purely gravimetric measurement. Figure 5 shows a µ(φ)-curve (grey) and a series of mean values (orange bars) of this curve, taken over humidity ranges with the same midpoint but with increasing widths. First, due to the curvature of the µ-curve, all those mean values are lower than the curve itself at the midpoint. The curve point is the desired µ-value corresponding to the midpoint humidity. The mean values correspond to the results of cup measurements performed by applying the respective humidity ranges. The results of all cup measurements are lower than the µ-curve point corresponding to the midpoint humidity, so it would be inappropriate to simply substitute a cup measurement result for this µ-value. Secondly, the mean values for humidity ranges with the same midpoint are lower for wider ranges. This means that when cup measurements with different wide humidity ranges are combined into one diagram (plotting them against the midpoint humidity), a misleading zigzagshape of the diagram may result, even though the underlying true µ(φ)-curve may be continuously falling for increasing humidity. A better approach than plotting the cup results against the midpoints of the applied humidity ranges is to determine a curve which has the properties that the harmonic mean values provide, taken over the humidity ranges which were used for the cup measurements, and are identical with the cup results. Figure 6 illustrates the procedure using data from WUFI for an adaptive vapour retarder, with the green bars representing a set of cup measurements, the grey curve being the adjusted test curve, and the thin orange bars being the mean values of the test curve, taken over the indicated ranges. The test curve is adjusted until the orange bars coincide as well as possible with the green bars. The above principle was applied to the three materials measured in this research, to plot the final hygrothermal boundary curve for the equivalent air layer thickness for specimen C, D, and E (see Figure 7) after the harmonic adjustment. These harmonically balanced values could be applied to the material properties within the transient hygrothermal simulation software to provide a more accurate simulation of the water vapour diffusion through an external envelope. A better approach than plotting the cup results against the midpoints of the applied humidity ranges is to determine a curve which has the properties that the harmonic mean values provide, taken over the humidity ranges which were used for the cup measurements, and are identical with the cup results. Figure 6 illustrates the procedure using data from WUFI for an adaptive vapour retarder, with the green bars representing a set of cup measurements, the grey curve being the adjusted test curve, and the thin orange bars being the mean values of the test curve, taken over the indicated ranges. The test curve is adjusted until the orange bars coincide as well as possible with the green bars. A better approach than plotting the cup results against the midpoints of the applied humidity ranges is to determine a curve which has the properties that the harmonic mean values provide, taken over the humidity ranges which were used for the cup measurements, and are identical with the cup results. Figure 6 illustrates the procedure using data from WUFI for an adaptive vapour retarder, with the green bars representing a set of cup measurements, the grey curve being the adjusted test curve, and the thin orange bars being the mean values of the test curve, taken over the indicated ranges. The test curve is adjusted until the orange bars coincide as well as possible with the green bars. The above principle was applied to the three materials measured in this research, to plot the final hygrothermal boundary curve for the equivalent air layer thickness for specimen C, D, and E (see Figure 7) after the harmonic adjustment. These harmonically balanced values could be applied to the material properties within the transient hygrothermal simulation software to provide a more accurate simulation of the water vapour diffusion through an external envelope. The above principle was applied to the three materials measured in this research, to plot the final hygrothermal boundary curve for the equivalent air layer thickness for specimen C, D, and E (see Figure 7) after the harmonic adjustment. These harmonically balanced values could be applied to the material properties within the transient hygrothermal simulation software to provide a more accurate simulation of the water vapour diffusion through an external envelope.

Conclusions
This research has investigated through laboratory measurement the water vapour diffusion characteristics of two types of water vapour impermeable reflective pliable membranes and one type of smart membrane. The research included measuring their water vapour diffusion behaviour under varying boundary conditions, with specific attention to different relative humidity conditions.
The results from the cup measurements show that the two vapour impermeable pliable membrane products are not open to vapour diffusion even at higher relative humidity conditions, as their resistance factor and air layer thickness are too high to allow for vapour diffusion. On the other hand, the result from the smart pliable membrane product indicates that vapour diffusion is possible at higher relative humidity, as the resistance factor and air layer thickness decreases with increases in relative humidity. Contrary to the single point test method applied typically to construction materials, the graphs for all the three types of pliable membrane material show that they behave in a dynamic nonlinear manner subject to the relative humidity conditions. This finding is similar in nature to the previously published findings which reported on the non-linear vapour diffusion properties for tested vapour permeable pliable membranes [13]. Alarmingly, the composite materials that make up the vapour impermeable pliable foil-faced membranes supported mould growth during the test period, whilst the non-foil polyethylene copolymer products did not support mould growth. Further analysis based on the harmonic adjustment approach was then employed to plot the hygrothermal boundary curves for each of

Conclusions
This research has investigated through laboratory measurement the water vapour diffusion characteristics of two types of water vapour impermeable reflective pliable membranes and one type of smart membrane. The research included measuring their water vapour diffusion behaviour under varying boundary conditions, with specific attention to different relative humidity conditions.
The results from the cup measurements show that the two vapour impermeable pliable membrane products are not open to vapour diffusion even at higher relative humidity conditions, as their resistance factor and air layer thickness are too high to allow for vapour diffusion. On the other hand, the result from the smart pliable membrane product indicates that vapour diffusion is possible at higher relative humidity, as the resistance factor and air layer thickness decreases with increases in relative humidity. Contrary to the single point test method applied typically to construction materials, the graphs for all the three types of pliable membrane material show that they behave in a dynamic non-linear manner subject to the relative humidity conditions. This finding is similar in nature to the previously published findings which reported on the non-linear vapour diffusion properties for tested vapour permeable pliable membranes [13]. Alarmingly, the composite materials that make up the vapour impermeable pliable foil-faced membranes supported mould growth during the test period, whilst the non-foil polyethylene copolymer products did not support mould growth. Further analysis based on the harmonic adjustment approach was then employed to plot the hygrothermal boundary curves for each of these pliable membranes. This is because the results from the gravimetric cup measurement may not be enough to determine the effectiveness of a materials' moisture behaviour along the cross-section of these materials, which is needed for hygrothermal modelling.
In conclusion, this research demonstrates that regardless of the water vapour diffusion class, each of the tested pliable membrane types behaved differently under different relative humidity conditions and vapour pressure gradients within the testing laboratory. This may indicate that the current single point value for construction material vapour diffusion properties used in hygrothermal simulation may be inadequate and may provide inaccurate guidance regarding moisture, moisture accumulation, and mould growth.  The vapour flux g is calculated = G A kg/s.m 2 . Where A is the arithmetic mean of the exposed area of the test specimen in m 2 , for specimen C1, the diameter of the specimen after sealing is 190 mm: The water vapour permeance is then calculated as: To calculate ∆P V = P satwetside − P satdryside The water vapour resistance is calculated as follows: