Influence of Air Space on Multi-layered Material Water Vapor Measurement Yang Wu 1, Mavinkal K. Kumaran 2, Paul Fazio 1 Abstract The effect of interface between material layers in building envelope is of great importance for characterization of moisture transport behavior. This investigation found that air space in multilayered material influences the water vapor transportation through it. A simplified dry cup test based on ASTM E96 standard is conducted on different configurations of gypsum specimens under 50% and 90% RH and at 23 C. The measured water vapor resistance (WVR) of multilayered gypsum board shows an unexpected difference from the theoretical calculation results. To understand the reason, further investigation is necessary. However, the measurement results show that neglect or incorrect consideration of air space effect during the simulation of moisture transmission will result in inaccurate water vapor transmission rate. 1. Introduction Building envelope is assembled from different layers of building materials, and the effect of interface between material layers on moisture transfer is relevant to the appropriate understanding of whole assembly performance. Most moisture transfer through building envelope involves interface phenomena that can not be ignored. Water vapor transport proceeds from layer to layer through the whole envelope. With the validated hygrothermal models and accurate measured material property data moisture transport through building envelope can be numerically predicted. However, most models treat the materials as individual layers in perfect hydraulic contact, i.e., the interface has no influence on moisture transport. In reality, this assumption can be violated by the fact of natural contact or air space between layers of materials. The knowledge about the continuity between layers was introduced by De Freitas (1991). Three kinds of contact configurations were concluded as follows: Hydraulic contact when there is interpenetration of both layers porous structure. Natural contact when there is contact without interpenetration of both layers porous structure. Air space between layers when there is an air pocket, a few millimeters wide. Hydraulic contact usually represents the perfect touch of layers. In this situation, the capillary or vapor pressures at the contact surfaces of both materials are equivalent. In case of natural contact, the hygric resistance in the interface can be found with the reference of the experimental investigations by De Freitas et al. (1996) and Qiu et al. 1 Department of Building, Civil and Environmental Engineering, Concordia University, Montreal, Canada. 2 Institute for Research in Construction, National Research Council Canada, Ottawa, Canada. - 1 -
(2003). When air space existed between layers, vapor or liquid transports by conduction or convection through it. The effect of air layer between layers on vapor transport is experimentally investigated in present work by the simplified dry cup method. 2. Theoretical study Water can exist in building materials in three states, i.e., solid (ice), liquid (water), and gas (water vapor). In practical conditions that buildings are operated, all these states may exist and are difficult to distinguish. The moisture transport mechanisms in building envelope, therefore, are complicated. It can be either in vapor phase or liquid phase with various driving potentials. In vapor phase, there are convective flow and gas diffusion for which the driving potentials are air pressure and water vapor pressure respectively. The processes of liquid phase moisture transport can be electrokinesis, hydraulic flow, gravitational, and capillary flow, which are driven by electrical voltage, total pressure, height, and suction pressure correspondingly. Theory on water vapor transport in building materials is originally based on Fick s diffusion law. It is generally written as WVT = δ p P v (1) dpv P v dx (2) Where WVT = water vapor flow rate (kg/m 2 s), δ p = water vapor permeability (kg/m s Pa), P v = partial water vapor pressure (Pa), x = distance along the flow path (m). According to Equation 1, the driving potential of pure water vapor diffusion is P v. 3. Cup method for water vapor permeability measurement Dry cup and wet cup methods are used widely to determine the water vapor permeability of building material (e.g., Galbraith and Mclean 1986, Hansen 1990, Burch 1992, Mclean et al. 1992, Kumaran 1998a). ASTM standard E96-00, Standard Test Method for Water Vapor of Materials (ASTM 2000), describes the test procedure in detail. The specimen is sealed to the mouth of an impermeable test dish containing water or desiccant and placed in a controlled environment. Figure 1 shows the cup assembly used at the Institute for Research in Construction, NRC Canada. The cup is made out of PVC, the specimen is sealed on the cup ring by molten wax. The molten wax is made of 60% beeswax and 40% paraffin wax warmed to 180 C. Dry cup is filled with desiccant (Calcium Chloride) for dry cup measurements, and distilled water for wet cup. Cups are placed in a climate-controlled chamber, and the assemblies are weighed at different intervals depending on materials. According to the analysis procedure established by (Kumaran 1998b), to guarantee the true steady state of each set of measurement, a linear least-squares analysis of the data on time versus mass change is used, and the linear regression coefficient - 2 -
should not be less than 0.999. Instead of measuring the water vapor transport using dry cup and wet cup methods under different RH conditions, a simplified test was adopted for this investigation. The vapor transmission rates of different layered gypsum composites in 50% and 90% RH environmental chamber were measured using dry cup method. The configurations of gypsum composites in this study are shown in Figure 2. By those configurations, the water vapor transport in multilayered gypsum board with air space was experimentally estimated. The controlled temperature and relative humidity chambers developed for ASTM E96 test procedure was used. Those chambers have the capability to maintain temperature and RH within the specified set point ±0.1ºC and ±1% for an indefinitly long period. The mechanical balance used for weighing the specimens and test assemblies satisfy the criteria specified in the ASTM E96 standard. Test Specimen PVC Ring PVC Cup Wall Wax Together PVC Cup Bottom Figure 1. Assembly of measuring cup III. x=d III, δ p = δ piii II. x=d II, δ p = δ pii I. x=d I, δ p = δ pi I. x=d I, δ p = δ pi II. x=d II, δ p = δ pii I. x=d I, δ p = δ pi (a) (b) (c) Figure 2. The gypsum composites consisting of (a) single layer, (b) double layers, and (c) triple layers. Air space exists between layers, and the lateral sides are sealed using molten wax. 4. Analysis of Test Results The measurement results were analyzed with the same principles used in ASTM Standard E96. The change of the cup weight is plotted against the time elapsed. A straight line observation, involving at least six properly spaced points, indicates the establishment of steady state water vapor transmission process. The slop of this straight line is water vapor transmission rate (WVT). The test results obtained in this study, when plotted and curve fitted, show a very clear straight line with an R-square value 0.999 or higher. - 3 -
The water vapor transmission rate (WVT) is calculated using the following equation. ( G / t) G WVT = = ta A (3) where, G = Weight change of desiccant or water (kg) t = time (s) G/t = slop of the straight line (kg/s) A = test area, i.e., cup mouth area (m 2 ), and Water vapor permeance where: S R 1 R 2 WVP WVT WVP = (4) S( R 1 R 2 ) = saturation vapor pressure at test temperature (Pa), = relative humidity at the moisture source expressed as a fraction (test chamber for desiccant method; in the dish for water method), = relative humidity at the vapor sink expressed as a fraction, and = water vapor permeance (kg/m 2 s Pa) Water Vapor The water vapor resistance (WVR) of a building component is expressed as the reciprocal of the water vapor permeance (WVP) of the same. 1 WVR = (5) WVP where: WVR = water vapor resistance (m 2 s Pa/kg) In addition the following corrections are applied to the test results. 1. Corrections for resistance due to the still air layers, and 2. Corrections due to resistance offered by the specimen surface. Due to Still Air Layer If the thickness of the still air layer present between the desiccant and specimen or adjacent layers of specimen is known, then the corresponding water vapor resistance can be calculated using the following equation of permeability, proposed by (Schirmer, 1938). Where δ a = water vapor permeability of the still air (kg/m s Pa), 5 1.81 2.306 10 P0 T δ a = (6) RvTPa 273.13-4 -
P 0 P a T R v = standard atmospheric pressure (101325 Pa), = is ambient air pressure (Pa), = the temperature (K), and = gas constant for water (461.5 J/K kg). The resistance offered by still air, where, l = thickness of air layer (m) l AR = (7) δ a Due to Specimen Surface The surface resistances, i.e. inside and outside surfaces of the specimen, have been approximated using Lewis relation (Pedersen, 1990). For the cup method, the total surface resistance offered by two surfaces is judged to be approximately 4 x 10 7 Pa s m 2 /kg. Water Vapor of the Material WVR of the specimen = (WVR from Equation 5) (resistance offered by still air (Equation 7) for a known thickness in the cup and specimen surfaces (i.e., 4 x 10 7 Pa s m 2 /kg)), i.e., 1 l WVRcorrected = S R (8) WVP δ a where: S R = resistance offered by specimen surfaces (Pa s m 2 /kg) WVP of the specimen = 1/( WVR of the specimen), i.e., 1 WVPcorrected = WVRcorrected (9) water vapor permeability (μ) of the material (kg/m s Pa) = ( WVR of the specimen) x (thickness of the specimen), i.e. μ corrected = WVP corrected d (10) where μ = water vapor permeability (kg/m s Pa) d = thickness of the specimen (m) 4. Measurement Results The investigation was carried out through three steps. Totally nine single layer gypsum specimens were cut. During the first step, all single layer specimens WVP in a 50% RH chamber were measured using the dry cup method. At the second step, three double layered specimens were prepared by sealing the lateral sides of two single specimens using wax, and then same dry cup tests were conducted on them. An air space - 5 -
existed between the two layers. In the third step, another layer was added. 50% RH dry cup tests were carried out on three triple layered specimens in which two layers of air space were enclosed. The whole procedure was repeated in a 90% RH chamber, in a second series. The selected test specimens used for the measurement in this work were cut from a 4 X 8 sheet of a commercial product with a nominal thickness of 12.7 mm. The commercial Gypsum Board conforms to CAN/CSA-A82.27. Nine specimens were marked as A to I in alphabetical order. The basic material properties of the single layer gypsum specimen and measurement results are listed in table 1. The bulk density of investigated gypsum material is (592 ± 5) kg/m 3. The measurement results were analyzed as described in the previous section. Table 1. Material properties and measured water vapor permeability of single layer gypsum board (a) 87.8% RH and 23 ºC Specimen Thickness Diameter A 12.48 143.83 5.89E-03 3.02E-09 3.31E+08 3.77E-11 B 12.41 143.27 6.01E-03 3.10E-09 3.23E+08 3.85E-11 C 12.53 142.96 6.26E-03 3.27E-09 3.06E+08 4.09E-11 D 12.40 143.22 5.83E-03 2.98E-09 3.31E+08 3.70E-11 E 12.51 143.13 5.89E-03 3.02E-09 3.35E+08 3.78E-11 F 12.38 143.19 5.90E-03 3.03E-09 3.30E+08 3.75E-11 G 12.46 144.01 5.87E-03 3.01E-09 3.32E+08 3.75E-11 H 12.55 142.64 6.14E-03 3.18E-09 3.14E+08 4.00E-11 I 12.48 143.29 6.23E-03 3.25E-09 3.08E+08 4.05E-11 (b) 50.2% RH and 23 ºC Specimen Thickness Diameter A 12.48 143.83 3.20E-03 2.85E-09 3.51E+08 3.56E-11 B 12.41 143.27 2.89E-03 2.51E-09 3.98E+08 3.12E-11 C 12.53 142.96 3.16E-03 2.80E-09 3.57E+08 3.51E-11 D 12.40 143.22 3.35E-03 3.02E-09 3.31E+08 3.75E-11 E 12.51 143.13 3.16E-03 2.81E-09 3.56E+08 3.52E-11 F 12.38 143.19 3.01E-03 2.64E-09 3.78E+08 3.27E-11 G 12.46 144.01 3.08E-03 3.67E-09 3.71E+08 3.40E-11 H 12.55 142.64 2.94E-03 2.54E-09 3.94E+08 3.17E-11 I 12.48 143.29 2.94E-03 2.54E-09 3.93E+08 3.21E-11 After the dry cup tests on single layer specimen, all the specimens were removed from the measurement cups carefully. The dry cup tests on double and triple layered - 6 -
specimens were conducted after assembling the specimens as double or triple layers. The measurement results are listed in Table 2 and 3 respectively. Air space thickness in Table 2 and 3 is calculated from the measured multilayer specimen thickness subtracting the sum of corresponding single layer specimen. Table 2. Material properties and measured water vapor permeability of double layered gypsum board (a) 89.4% RH and 23 ºC Specimen Thickness Air layer AD 25.14 0.14 3.45E-03 1.65E-09 6.38E+08 3.98E-11 BE 24.99 0.18 3.36E-03 1.60E-09 6.53E+08 3.84E-11 CF 24.97 0.07 3.41E-03 1.63E-09 6.38E+08 3.90E-11 (b) 50.3% RH and 23 ºC Specimen Thickness Air layer AD 25.14 0.14 1.92E-03 1.55E-09 6.45E+08 3.90E-11 BE 24.99 0.18 1.87E-03 1.51E-09 6.63E+08 3.77E-11 CF 24.97 0.07 1.92E-03 1.55E-09 6.44E+08 3.88E-11 Table 3. Material properties and measured water vapor permeability of triple layered gypsum board (a) 90.3% RH and 23 ºC Specimen Thickness Air layer ADG 38.16 0.70 2.75E-03 1.26E-09 8.27E+08 4.66E-11 BEH 38.11 0.75 2.72E-03 1.24E-09 8.32E+08 4.59E-11 CFI 37.64 0.26 2.78E-03 1.27E-09 8.09E+08 4.64E-11 (b) 50.3% RH and 23 ºC Specimen Thickness Air layer ADG 38.16 0.70 1.45E-03 1.13E-09 8.83E+08 4.32E-11 BEH 38.11 0.75 1.42E-03 1.14E-09 8.77E+08 4.35E-11 CFI 37.64 0.26 1.47E-03 1.19E-09 8.43E+08 4.47E-11 Figure 3 plots the measured water vapor transmission rates (WVT) versus the number of specimen layers at 50% and 90% RH. It is affected by the number of layers, and this effect can be clearly seen in Figure 3. As can be predicted, the more layers of the - 7 -
material, the less water vapor will transfer through it. An interesting phenomenon is that WVT is exponentially decreased with number of layers increases instead of linearly. In the following section, it is further discussed. 8.00E-03 6.00E-03 50% RH 90% RH WVT (g/s m²) 4.00E-03 2.00E-03 0.00E+00 1 2 3 Number of Layers 4 Figure 3. Water vapor transmission rate (WVT) versus the number of specimen layers 5. Discussions In the aforementioned analysis method, for the multilayered specimen, the corrected vapor resistance from measurement should include the resistance and surface coefficient between each layer. It can be expressed as the sum of corrected WVR of single layer material, resistance offered by still air (Equation 6) and surface coefficients between materials, i.e., WVRn + AR 2 corrected = WVR + L+ AR corrected,1 + WVR n 1 + S R,1 + S R,2 + correctet,2 LS R, n 1 + L+ WVR corrected, n + AR where, WVRn = water vapor resistance of multilayered specimen (m 2 s Pa/kg), 1,2,, n = number of layers The corrected WVR of multilayered specimen are calculated from corrected single layer resistance using Equation 11. In this calculation, it can be found that the effect of air gap on whole specimen s permeance is small. As would be expected, WVR shows an almost linear relationship with the number of layers. The calculated and experimental WVR versus number of specimen layers under 50% and 90% are plotted in Figure 4 and 5 respectively. It can be seen that WVR of 2 and 3 layers specimen derived from measurement data shows different value from calculated resistance under both RH conditions. A logarithmic trend line can be found in experimental resistance with increasing of specimen layer number. The difference between calculated and 1 (11) - 8 -
experimental data increases with the increment of number of specimen layers, and this difference is more significant under 50%RH than 90% RH. 1.40E+09 1.20E+09 Calculated Experimental WVR (m2 s Pa/kg) 1.00E+09 8.00E+08 6.00E+08 4.00E+08 2.00E+08 1 2 3 Number of Layers Figure 4. Water vapor resistance (WVR) with different number of specimen layers under 50% RH 1.20E+09 1.00E+09 Calculated Experimental WVR (m2 s Pa/kg) 8.00E+08 6.00E+08 4.00E+08 2.00E+08 1 2 3 Number of Layers Figure 5. Water vapor resistance (WVR) with different number of specimen layers under 90% RH - 9 -
Usually for multilayered building material, it should be expected that the measured WVR is greater than the sum of single layers with additional air gap and surface coefficient that the material be composed of, because there are some additional unknown resistances. However, in this study, it shows the opposite characteristic in all cases. The reason of that has not been clearly recognized yet. Several factors can affect the result, and those factors could be the incorrect magnitude of surface coefficient, two dimensional water vapor flows or inaccurate correction made for WVR, etc. To find out the reasons further investigation is considered necessary. From the analysis, it can be said that the air space between specimen layers obviously influence the water vapor transmission through multilayered building material. In the simulation of moisture transport through building envelope, the small air space between each layer of building material usually dose not account. However, it might result in over- or under- predicting the amount of water that transports between indoor and outdoor. 6. Concluding Remarks Through the simple experiment of water vapor permeability based on ASTM E96 method, the effect of air space on water vapor transmission across multilayered material is investigated. The following conclusions can be drawn from this study: The dry cup tests on gypsum board with different configurations under 50% and 90% RH were carried out. The measured WVT of multilayered gypsum board has an exponential relationship with the number of layers. The measured WVR of multilayered gypsum board shows an unexpected difference with the theoretical calculation results. To understand the reason of that, further investigation is necessary. The air space in multilayered specimen can not be neglected, and it is important to the accuracy of moisture transport measurement. It is hoped that this study will inspire researchers to generate more information and discussions about the water vapor transmission characteristics of composite building materials. Reference ASTM Standard E96-00, 2000, Standard Test Method for Water Vapor of Materials, Annual Book of ASTM Standard, 14.02:878-885. Burch, D.M., Thomas, W.C., and Fanney A.H., 1992, "Water vapor permeability measurements of common building materials", ASHRAE Transactions, 98(2), pp. 486 De Freitas, V. P., (1991), Moisture diffusion in building envelope-influence of the contact between two layers. CIB W40 Meeting, Lund, Sweden, lo-12 September (1991). - 10 -
De Freitas, V.P., Abrantes, V. and Crausse, P. (1996). Moisture Migration in Building Walls Analysisof the Interface Phenomena, Building and Environment, 31(2): 99 108. Galbraith, G.H., Mclean, R.C., 1986, Realistic vapor permeability values, Building Research & Practice, 14(2), 98-103 Hansen, K. K. and Lund, H. B., 1990, Cup Method for Determination of Water Vapor Properties of Building Materials. Sources of Uncertainty in the Methods, Proceedings of the 2nd Symposium, Building Physics in the Nordic Countries, Trondheim, pp. 291-298. Kumaran, M.K., 1998a, Interlaboratory comparison of the ASTM standard test methods for water vapor transmission of materials (E 96 95), ASTM Journal of Testing and Evaluation (USA). Vol. 26, no. 2, pp. 83-88 Kumaran, M.K., 1998b, An Alternative Procedure for the Analysis of Data from the Cup Method Measurements for Determination of Water Vapor Properties, Journal of Testing and Materials, Volume 26, No. 6, pp575-581 Mclean, R.C., Galbraith, G.H., Sanders, C.H., 1992, Moisture transmission testing of building materials and the presentation of vapor permeability values, Building Research and Practice, 18(2), 82-91 Pedersen, C. R., Combined Heat and Moisture Transfer in Building Constructions, Ph.D thesis, Thermal Insulation laboratory, The Technical University of Denmark, 1990. Qiu, X., Haghighat, F., Kumaran, M.K., (2003), Moisture transport across interfaces between autoclaved aerated concrete and mortar, Journal of Thermal Envelope & Building Science, v. 26, no. 3, pp. 213-236 Schirmer, R., ZVDI, Beiheft Verfahrenstechnik, Nr. 6, S. 170, 1938. - 11 -