Experimental characterization of moisture transport in brick masonry with natural hydraulic lime mortar

1 EXPERIMENTAL CHARACTERIZATION OF MOISTURE TRANSPORT IN BRICK 1 MASONRY WITH NATURAL HYDRAULIC LIME MORTAR 2 R. Ramireza,*, B. Ghiassib, P. Pinedac, P.B. Lourençoa 3 a University of Minho, Institute for Sustainability and Innovation in Structural Engineering (ISISE), 4 Department of Civil Engineering, Campus de Azurém s/n, 4800-085 Guimarães, Portugal 5 b Centre for Structural Engineering Design and Informatics (CSEI), Faculty of Engineering, University 6 of Nottingham, University Park, NG7 2RD Nottingham, United Kingdom 7 c School of Architecture, Department of Building Structures and Geotechnical Engineering, University of 8 Seville, Avda. Reina Mercedes 2, 41012 Seville, Spain 9 * Corresponding author. Mail: Institute for Sustainability and Innovation in Civil Engineering, 10 University of Minho, Campus de Azurém, s/n, 4800-058 Guimarães, Portugal. E-mail: 11 rafael.alva[email protected]m. Telephone: +351 910905513 12 ABSTRACT 13 This article presents an experimental study on the hygric performance of brick masonry with a special focus 14 on the hydraulic characterization of the brick-mortar interface. The hygric behavior was studied at the level 15 of the constituent materials as well as at the composite scale. An extruded fired-clay brick and two types of 16 mortar –Portland cement mortar and natural hydraulic lime mortar– were chosen as target materials. The 17 hygric properties were determined following prescribed experimental procedures. The experimental 18 program included vacuum saturation tests, static gravimetric tests for adsorption/desorption isotherms, 19 dry/wet cup tests, capillary absorption tests, and isothermal one-dimensional drying. The results allowed 20 for a comprehensive identification of the main hygric features of the target materials. Fired-clay brick and 21 lime mortar revealed low hygroscopicity and a strong capillary-active behavior with negligible hysteresis, 22 whereas cement mortar showed activity in both hygroscopic and capillary ranges with a distinct moisture 23 storage hysteresis. Different curing conditions led to dissimilar properties between the lime mortar prepared 24 in molds and the same mortar cured in masonry bed joints. Water absorption in fired-clay bricks showed a 25 marked anisotropic behavior, likely derived from the extrusion process. The existence of an imperfect 26 hydraulic contact at the brick-mortar interface was demonstrated for water absorption. Conversely, the 27 interfacial effect on drying kinetics was not evident. 28 Keywords: Masonry; moisture transport; hygric characterization; porous building material; multi-layered 29 Revised Manuscript with No Changes Marked Click here to view linked References 2 material; brick-mortar interface 1 2 3 1. INTRODUCTION 1 Masonry has been traditionally used as the main construction material in historical structures all around the 2 world. Nowadays, brickwork is still commonly used in veneer walls within contemporary frame structures. 3 Due to their enveloping nature, masonry walls play a major role in the overall energy efficiency of buildings 4 and control of indoor conditions [1–3]. Moreover, the external walls are usually exposed to environmental 5 actions, such as solar radiation, wind-driven rain, moisture condensation and rising damp, among others. 6 Such environmental exposure consequently produces a degradation in each constituent material as well as 7 in masonry as a whole [4,5]. From a durability point of view, moisture-related processes are of special 8 interest since they may lead to significant damage, e.g. cracking due to hygric swelling and contraction, 9 spalling caused by freeze-thaw cycles and (crypto)efflorescence due to salt crystallization [6,7]. Hence, 10 prevention and reparation of moisture-related damage in masonry require a thorough understanding of its 11 hygric behaviour. 12 Due to the importance of moisture phenomena, different theoretical models and numerical tools have been 13 developed to account for moisture transport in porous media [8–10]. Regardless of the chosen approach, a 14 series of material parameters is always necessary to describe the hygric behavior of a given porous medium. 15 This basic set of material data encompasses both moisture storage and moisture transport properties 16 including both liquid and gas (vapor) phases. In the hygroscopic range (0-95% relative humidity, RH), 17 water vapor plays the key role and moisture transport occurs mainly through vapor diffusion. In the over- 18 hygroscopic or capillary range (95-100% RH), liquid transport by capillarity dominates the moisture 19 transfer. A set of basic experiments and derived material properties for the characterization of a porous 20 material over the entire moisture range was defined by Scheffler [11]. It includes general properties, such 21 as bulk density and open porosity, moisture storage capacity, vapor diffusion properties, and liquid transport 22 properties, such as liquid water diffusivity or permeability, which may be derived from water absorption 23 experiments and drying tests. 24 Several round-robin initiatives have been launched to define and collect hygric properties of commonly 25 used porous building materials, e.g. IEA Annex 24 [12] and EC HAMSTAD [13]. Nevertheless, 26 considerably large variations were reported for some materials, including certain types of masonry units 27 and mortar mixes. The reason behind such deviations is manyfold. On one hand, ceramic products and 28 mortars are hardly free of intrinsic variability and inhomogeneities. Even industrially produced bricks can 29 4 have highly variable properties depending on the raw materials and the forming and firing methods [14,15]. 1 Similarly, mortar mixes are in-situ prepared materials, and their final properties depend on the type and 2 proportion of components as well as on the curing conditions. Moreover, the registered variation present in 3 the literature may also come from the testing process itself, e.g. lack of a unified experimental setup, 4 operator expertise and familiarity with the procedure, neglected factors such as temperature, etc. While 5 intrinsic material variability can hardly be overcome univocally, the standardization and consistent 6 implementation of experimental procedures is still a work-in-progress to minimize testing errors. Indeed, 7 some works have focused on repeatability and reproducibility of the testing methods to determine hygric 8 properties [16–18]. Shared conclusions were the importance of applying a unified approach, the value of 9 providing a detailed report of testing conditions and procedures, and the need for a revision of the current 10 vapor permeability testing methods. In this context, a thorough material characterization contributes to 11 improving the quality of available material databases as well as increasing our confidence in the reported 12 values and test protocols. 13 Numerous works in the literature have documented the hygric properties of clay brick as a porous building 14 material, e.g. [11–13,19]. However, given the high intrinsic variability previously mentioned, the 15 extrapolation of literature data to specific real case scenarios is not always straightforward. Moreover, most 16 databases assume isotropic properties and very few references point at an anisotropic behavior, which has 17 been nonetheless demonstrated for both molded [20] and extruded [21] bricks. 18 The most common mortars used in masonry structures are lime- and cement-based mixes [22]. Lime 19 mortars are more frequently found in historical constructions. However, most hygric characterization 20 studies have focused on cement-based mixes. Similarly, the available studies on brick masonry are mostly 21 concern with cement-based or blended lime-cement mortars, and few cases tackle the combination of brick 22 and lime-based mixes, e.g. [23–26]. 23 Several authors have dealt with moisture transport in multi-layered materials such as masonry [27–35]. The 24 existing literature comprises experimental, analytic, and numerical works, which account for the interfacial 25 effects from different viewpoints. Most of the research is focused on liquid water uptake whereas drying 26 has been scarcely studied [27]. Wilson et al. [28,29] studied sorption phenomena in multi-layered building 27 materials, namely plaster and plaster/sand mixtures. The authors used analytic expressions based on the 28 sharp front model and assumed perfect hydraulic contact at the interface. Their experimental studies showed 29 5 an overall good match with the analytic solutions, but some deviations were identified. The differences 1 were explained as a consequence of the wet sharp front simplification. Hall and Hoff [30] later extended 2 this approach to account for imperfect hydraulic contact cases and introduced the interfacial resistance as 3 an additional cause to explain the observed discrepancies. 4 Considering the possible causes that may lead to the existence of a hydraulic resistance at the brick-mortar 5 interface, a commonly accepted explanation is related to the curing conditions of the bedding mortar 6 between bricks [27,31]. During application, bricks absorb water from the fresh mortar, which leads to a 7 drop in the water-binder ratio of the mixture. Furthermore, water extraction from the mortar is also 8 connected to the transport of fine binder particles towards the interface and a stratification of the mortar 9 across the joint thickness. Consequently, one would expect a denser, more compact mortar at the interface. 10 However, this is not always the case since a more porous mortar near the interface may be equally found in 11 real scenarios. To account for this fact, Groot and Larbi [32] hypothesized a reversed water flow from brick 12 to mortar after compaction and initial hydration of the mortar, with subsequent impact on the interface 13 development. Overall, the quality of brick-mortar interfaces seems to be highly dependent on the curing 14 history of the mortar. Derluyn et al. [33] and Janssen et al. [34] studied the absorption behavior of brick- 15 cement masonry composites with different interfacial configurations, namely perfect hydraulic contact 16 (kaolin layer), and wet- and dry-cured specimens. Their studies confirmed that the interfacial effects were 17 proportional to the water extraction from the mortar during curing. Further causes for the existence of a 18 hydraulic resistance at the brick-mortar interface can be the presence of air gaps resulting from damage 19 (cracks) or due to poor workmanship during the application of the fresh mortar [23]. Even if the application 20 and curing conditions were optimal, a certain hydraulic resistance is expected due to the pore structure 21 discontinuity between the materials [27]. 22 The present paper aims to provide a comprehensive hygric description of several porous building materials 23 traditionally used in masonry structures. First, the properties of single constituent materials –fired-clay 24 brick, cement-based mortar and natural hydraulic lime (NHL) mortar– are studied. Subsequently, multi- 25 layered masonry specimens made up of fired-clay bricks and lime mortar are analyzed. Thus, the existence 26 of interfacial phenomena is assessed by comparing the behavior of each material separately and in 27 combination with one another. The composite behavior of masonry elements is studied in terms of water 28 absorption and isothermal drying tests. Finally, special focus is set on contrasting the hygric properties of 29 NHL mortar cast in molds with the same mortar cured in masonry bed joints. The different curing conditions 30 6 for these two cases are usually neglected but it is proven that this assumption may lead to erroneous 1 estimations. 2 2. MATERIALS AND METHODS 3 This section presents a brief description of the studied materials, namely fired-clay brick (B), Portland 4 cement mortar (CM), NHL mortar –molded (LM) and from masonry bed joints (LMJ) –, and masonry 5 composites. Subsequently, the configuration of the specimens used throughout the study is introduced. 6 Finally, a detailed explanation of the different experimental procedures is given. 7 2.1. Materials 8 2.1.1 Extruded fired-clay bricks 9 The masonry units selected for this study were commercial fired-clay solid extruded bricks. The extrusion 10 process, as well as the drying and subsequent baking (or ‘firing’) were fully automated. According to the 11 specifications of the manufacturer, the bricks were baked in a flash oven at 850°C for 3 hours. It must be 12 noted that in this type of process, the actual baking temperature depends on the position of the units inside 13 the oven, which may result in a certain level of variability between bricks of the same batch. Uneven heating 14 of the brick mass is usually perceived within single units as well, which becomes apparent in the color 15 transition between the outer layer and the core (Figure 1). Clay particles become more sintered at higher 16 baking temperatures, resulting in a lower porosity and a more intense red color of the brick [27]. 17 (a) (b) Figure 1. Cross-sections of fired-clay brick units: (a) transversal section through the middle plane; (b) 18 longitudinal section through the middle plane. 19 2.1.2 Portland cement mortar 20 Portland CEM I – 42.5R, as classified by EN 197-1:2011 [36], was used as a reference for comparison with 21 the lime-based mortar used in this study. The cement mortar was prepared with a binder-aggregate ratio 1:5 22 by volume and water-binder ratio 1:3 by weight. After casting, the mortar was kept in controlled curing 23 7 conditions, 20ºC and 90% RH, for 48 hours. Then, the specimens were demolded and immersed in water 1 at 20ºC for more than a year. 2 2.1.3 Natural hydraulic lime mortar 3 The lime mixes analyzed in this study were prepared using a commercial pre-mixed NHL-based mortar, 4 NHL 3.5 (REABILITA Cal Consolidação). The mortar was prepared as suggested by the producer by 5 blending 1 kg of the dry powder provided by the manufacturer with 0.15 kg water. For molded specimens, 6 the freshly cast mortar was kept for 48 hours in controlled curing conditions, 20ºC and 90% RH. Then, the 7 specimens were demolded and maintained in laboratory conditions, 20ºC and 60% RH, for more than a 8 year. 9 2.1.4 Masonry 10 Masonry samples (brick + mortar) were extracted from a wall using a diamond coring wet drilling machine. 11 The wall was constructed using the same bricks and NHL-based mortar discussed above. The bricks were 12 docked in water before placement to prevent undesired suction of water from the fresh mortar. The average 13 thickness of the mortar joints was 12 mm. Right after the wall was built, it was covered with a polyethylene 14 sheet for 48 hours to prevent water evaporation. After that, the cover was removed and the wall was kept 15 in laboratory conditions, 20ºC and 60% RH, for more than a year before the cores were extracted. 16 2.2. Specimens 17 Fired-clay extruded bricks were used in different formats depending on the corresponding test setup. 18 Initially, whole brick units with nominal size 200 mm × 100 mm × 50 mm were tested. Subsequently, the 19 external surfaces of the units were removed by grinding. Thus, the outermost brick layer (thickness 5 mm) 20 was removed, and the resulting prisms were cut into smaller pieces for further testing (Figure 2). 21 Figure 2. Cutting layout of brick units and mortar prisms (average dimensions in mm). 22 8 As for the mortars, standard molds with dimensions 160 mm × 40 mm × 40 mm were used to cast the 1 different mixes. Initially, whole mortar prisms were used. Then, the prisms were cut into smaller pieces for 2 further testing (Figure 2). Additionally, lime mortar discs extracted from masonry bed joints (Figure 3) 3 were analyzed to determine differences with respect to the molded counterpart. 4 The multi-layered specimens extracted from the masonry wall consisted of cylinders with different stacking 5 arrangements, i.e. different B+LMJ configurations (Figure 3). The average diameter of the cylinders was 6 Ø=51.30 mm. 7 Figure 3. Lime mortar discs and multi-layered specimens extracted from masonry wall (schematic 8 representation) (average dimensions in mm). 9 The number and type of specimens used for the characterization of single materials are summarized in 10 Table 1. Likewise, the masonry specimens tested are summarized in Table 2. 11 Table 1. Number of single-material specimens tested. 12 Material Specimen Test and test direction (dimensions in mm) Vacuum saturation Sorption isotherm Water absorption 1D drying test Vapor permeability X Y Z X Y Z Z B 200×100×50±2 10 -- 5 (1) -- -- -- -- 40×40×40±1 15 (2) -- 5 (3) 5 (3) 5 (3) 5 (4) 5 (4) 5 (4) -- 50×50×20±1 -- -- -- -- -- -- -- -- 5 (2) 30×30×10±1 -- 5 (2) -- -- -- -- -- -- -- CM 160×40×40±0.5 4 -- 4 (3) -- -- -- -- 9 40×40×40±0.5 3 (5) -- 3 (3) -- -- 3 (4) -- -- -- 30×30×10±1 -- 3 (5) -- -- -- -- -- -- -- LM 160×40×40±0.5 4 -- 4 (3) -- -- -- -- 40×40×40±0.5 3 (5) -- 3 (3) -- -- 3 (4) -- -- -- 30×30×10±1 -- 3 (5) -- -- -- -- -- -- -- LMJ Disc Ø51.3, h=12±1 5 5 -- -- -- -- -- -- -- (1) 5-brick unit subset chosen from the initial 10 brick units tested for vacuum saturation (2) Cut from 5-brick unit subset (3) Same specimens used for vacuum saturation (4) Same specimens used for water absorption (5) Cut from mortar prism 1 Table 2. Number of masonry specimens tested. 2 Masonry Specimen Test and test configuration (dimensions in mm) Water absorption 1D drying test +Z -Z Drying from brick Drying from mortar M1 Cylinder Ø51.3, hB=45±1, hLMJ=12±1 8 8 (1) -- -- M2 Cylinder Ø51.3, hB=45±1, hLMJ =12±1 8 8 (1) -- -- D1 Cylinder Ø51.3, hB=22±2, hLMJ =12±1 -- -- 6 6 (1) Same specimens tested in the opposite direction 3 2.3. Methods 4 16 For the case of masonry cylinders, a transparent adhesive film was used to wrap the lateral surfaces. 1 Different configurations of masonry composites were studied (Figure 3). The first set was made up of 8 2 cylinders with 2 layers (1 interface), i.e. B+LMJ. The specimens were tested twice, once sucking up water 3 from the mortar side and the second time absorbing water from the brick. The second group of masonry 4 specimens was composed of 8 cylinders with 3 layers (2 interfaces), namely B+LMJ+B. Similarly, these 5 cylinders were tested twice, each time sucking up the water from a different brick. The purpose of these 6 tests was to verify the existence of a hydraulic resistance at the interface between the two materials. 7 2.3.6 Drying tests 8 Isothermal, one-dimensional drying tests were carried out following the recommendations of EN 16322 9 [48], i.e. subjecting initially capillary saturated specimens to 1-D drying from a single surface. These tests 10 were used to study the drying kinetics of the selected materials. Furthermore, drying results may be used 11 for the identification of transport properties, such as liquid permeability, by means of inverse modelling 12 [49]. 13 The tests were performed on a total of 15 brick cubes (5 specimens for each direction, i.e. X, Y, Z) and 3 14 mortar cubes for each mortar type (testing direction X) which were initially saturated by means of capillary 15 absorption tests. Once the capillary moisture content was attained, 𝑤𝑐𝑎𝑝, the specimens were kept partially 16 immersed in water and the upper face was sealed using neutral silicone (note that the lateral faces were 17 already sealed with the same material prior to capillary tests). Once the silicone was dry, the specimens 18 were placed in a climatic chamber with controlled environmental conditions, namely 23ºC and 55% RH, to 19 dry from the unsealed surface. The specimens were weighed periodically to establish the mass loss until a 20 plateau was reached. 21 Drying tests were also performed on masonry cylinders (Figure 3). The initial cylinders were made up of 22 three layers, namely LMJ+B+LMJ. Subsequently, they were cut in half, so two pieces LMJ+B were 23 obtained. The pieces were then prepared following the same procedure described for the single-material 24 cubes. From each pair, one was prepared to dry from the mortar face and the other one to dry from the 25 brick. A total of 12 specimens were tested, 6 per case. 26 3. RESULTS AND DISCUSSION 27 3.1. Results on single constituent materials 28 3.1.1 Vacuum saturation tests results 29 17 The properties derived from vacuum saturation tests are collected in Table 4. 1 Table 4. Results from vacuum saturation tests (CoV between parentheses). 2 Material 𝜌𝑏𝑢𝑙𝑘 𝜙𝑜 𝑤𝑠𝑎𝑡 [kg/m3] [-] [kg/m3] B 1900 (2.69) 0.280 (6.76) 280 (6.76) CM 2000 (0.15) 0.210 (1.05) 210 (1.05) LM 1990 (0.81) 0.255 (2.53) 255 (2.53) LMJ 2060 (1.32) 0.230 (6.93) 230 (6.93) Vacuum saturation tests performed on whole bricks and brick cubes extracted from those initial units 3 produced the same results in terms of bulk density and open porosity. Similarly, no difference was observed 4 between mortar prisms and mortar cubes cut out of them. Therefore, the use of smaller specimens is justified 5 and is indeed recommended to speed up the process and guarantee the saturation of the innermost spaces. 6 As it was expected, the properties of the lime mortar obtained from masonry bed joints differed from those 7 obtained from molded specimens even though the mixes were prepared with the same composition and in 8 the same controlled environment. Overall, LMJ showed higher bulk density and lower open porosity, and 9 consequently lower saturation moisture content. 10 3.1.2 Moisture isotherm 11 The data collected from static gravimetric tests are shown in Figure 5. The results showed clear similarities 12 between B and LM, whereas CM and LMJ followed a different trend. On one hand, B and LM hardly 13 showed moisture adsorption until high RH values, revealing a low hygroscopic response. Moreover, no 14 significant differences were found between adsorption and desorption. Conversely, hysteresis was observed 15 in CM and LMJ. CM showed considerable moisture adsorption for lower RH levels, confirming a markedly 16 hygroscopic nature. Likewise, LMJ showed moisture adsorption for lower RH levels, and therefore a more 17 hygroscopic response than its molded counterpart. 18 18 (a) (b) (c) 0.0 0.2 0.4 0.6 0.8 1.0 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 Adsorption Desorption Adsorption Desorption wg [kg/kg] RH [-] 0.0 0.2 0.4 0.6 0.8 1.0 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 Adsorption Desorption Adsorption Desorption wg [kg/kg] RH [-] 0.0 0.2 0.4 0.6 0.8 1.0 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 Adsorption Desorption Adsorption Desorption wg [kg/kg] RH [-] 19 (d) Figure 5. Sorption isotherms and fitting curves for the studied materials at 23°C: (a) B, and (b) LM, fitted 1 with Künzel’s model [50]; (c) CM, and (d) LMJ, fitted with Mualem’s model [51]. 2 The experimental measurements were fitted with common analytic expressions found in the literature. In 3 particular, the model proposed by Künzel [50] was used to describe the sorption behavior of B and LM. 4 For CM and LMJ, the model established by Mualem [51] was adopted. The expression proposed by Künzel 5 can be written as [50]: 6 𝑤(𝜑)=𝑤𝑐𝑎𝑝·(𝜓−1)·𝜑 𝜓−𝜑 (11) where 𝑤(𝜑) [kg/m3] is the moisture content as a function of relative humidity, 𝜑 [dimensionless], 𝑤𝑐𝑎𝑝 7 [kg/m3] is the capillary moisture content, and 𝜓 [dimensionless] is a fitting parameter. Künzel’s model 8 stems from a simplified form of the BET equation [52] and applies to sorption curves with a marked 9 exponential trend, i.e. non-hygroscopic, capillary-active materials. 10 On the other hand, the model proposed by Mualem [51] is extensible to hygroscopic materials. This model 11 is a constrained form of the more generalized equation proposed by van Genuchten [53]: 12 𝑤(𝑃𝑐)=𝑤𝑐𝑎𝑝·[1+(𝑎·𝑃𝑐)𝑛]−𝑚 (12) where 𝑤(𝑃𝑐) [kg/m3] is the moisture content as a function of capillary pressure, 𝑃𝑐 [Pa], and 𝑎 [1/Pa], 𝑛 13 [dimensionless] and 𝑚 [dimensionless] are fitting parameters. Capillary pressure and relative humidity are 14 related by 𝑃𝑐=𝜌𝑤·𝑅𝑣·𝑇·ln𝜑, known as Kelvin equation. Considering van Genuchten’s expression, 15 Mualem fixed 𝑚=1−1/𝑛 [51]: 16 0.0 0.2 0.4 0.6 0.8 1.0 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 Adsorption Desorption Adsorption Desorption wg [kg/kg] RH [-] 20 𝑤(𝑃𝑐)=𝑤𝑐𝑎𝑝·[1+(𝑎·𝑃𝑐)1/(1−𝑚)]−𝑚 (13) The curves obtained from these analytic expressions are plotted in Figure 5 together with the experimental 1 points. The corresponding fitting parameters for the different cases are presented in Table 5 together with 2 relevant criteria to evaluate the accuracy of the model. In this case, the accuracy was determined by the 3 adjusted coefficient of determination (𝑅𝑎𝑑𝑗 2) and the normalized root mean square error (%NRMSE) 4 between predicted and measured points. The adjusted 𝑅2 takes into account the effect of the number of 5 fitting parameters and is given as: 6 𝑅𝑎𝑑𝑗 2=(𝑁−1)𝑅2−(𝑀−1) 𝑁−𝑀 (14) where 𝑁 is the number of measured data and 𝑀 is the number of fitting parameters. 7 The %NRMSE is expressed as a percentage and is defined as follows: 8 %𝑁𝑅𝑀𝑆𝐸= 1 ∑𝑥𝑒𝑥𝑝,𝑖/𝑁 𝑁 𝑖=1 √1 𝑁∑ (𝑥𝑒𝑥𝑝,𝑖−𝑥𝑝𝑟𝑒,𝑖)2 𝑁 𝑖=1 (15) where 𝑥𝑒𝑥𝑝 and 𝑥𝑝𝑟𝑒 are the measured values from the experiments and the predicted values from the fitted 9 curves, respectively. A lower value of %NRMSE indicates a better match. 10 Table 5. Fitting parameters and accuracy of the modelled sorption isotherms. 11 Material Künzel (Eq. (11)) Mualem (Eq. (13)) 𝑅𝑎𝑑𝑗 2 %NRMSE B 𝜓𝑎𝑑𝑠=1.0055 -- 0.999 2.30% 𝜓𝑑𝑒𝑠=1.0070 -- 0.999 2.49% LM 𝜓𝑎𝑑𝑠=1.0052 -- 0.999 5.30% 𝜓𝑑𝑒𝑠=1.0066 -- 0.999 4.49% CM -- 𝑎𝑎𝑑𝑠 =9.40·10−7 [1/Pa] 𝑚𝑎𝑑𝑠=0.312 0.993 6.62% -- 𝑎𝑑𝑒𝑠 =4.00·10−6 [1/Pa] 𝑚𝑑𝑒𝑠=0.197 0.983 7.77% LMJ -- 𝑎𝑎𝑑𝑠 =2.76·10−6 [1/Pa] 𝑚𝑎𝑑𝑠=0.325 0.996 7.94% -- 𝑎𝑑𝑒𝑠 =2.00·10−6 [1/Pa] 𝑚𝑑𝑒𝑠=0.326 0.994 7.96% 21 1 3.1.3 Vapor permeability 2 The results from dry and wet cup tests performed on brick specimens gave 𝜇𝑑𝑟𝑦= 34.14 (CoV = 12.64%) 3 and 𝜇𝑤𝑒𝑡 = 13.50 (CoV = 17.85%), respectively. Vapor permeability was only tested in the bed direction 4 (Z) since the cutting of the original units after removal of the outermost layer (5 mm) did not allow to obtain 5 the specimens for the other directions (Figure 2). It must be noted that the results obtained from cup 6 measurements showed the highest scatter among the tests performed in this study. Low reliability for this 7 type of test has been indicated before, see e.g. [16–18,54]. 8 The water vapor resistance factor, 𝜇, indicates how many times lower the vapor diffusion in the material is 9 in comparison with vapor diffusion in still air. By definition, 𝜇= 1 for stagnant air. Expressing the water 10 vapor permeability through the vapor resistance factor by means of Equation (6) has the advantage of 11 assigning the temperature and pressure dependencies of water vapor diffusion to the empirical term 12 𝛿𝑎(𝑇,𝑃), so that the vapour resistance factor is independent of these variables and becomes a constant 13 characteristic for each material. Nonetheless, as the values for our brick specimens show, vapor diffusion 14 tests performed at different relative humidity levels, e.g. dry and wet cups, usually result in different 15 permeability values for the same material [12,13]. Consequently, vapor diffusion (expressed through either 16 vapor permeability or vapor resistance factor) must be a function of moisture content. There have been 17 considerable efforts in the literature to explain this phenomenon, e.g. [11,41,50], but a univocal standpoint 18 has not been reached. Most authors agree on the overall influence of the pore-saturation level since the 19 presence of water islets in the pore system modifies the effective space available for vapor diffusion [11]. 20 However, the debate appears when defining the effect of moisture condensation in the pores for high 21 moisture contents and subsequent reduction of the space accessible to vapor transport. Künzel [50] argued 22 that vapor diffusion might be either obstructed or –this is the most common assumption– enhanced through 23 the water islets depending on the local conditions of temperature and moisture content inside the material. 24 Conversely, Krus [41] introduced the phenomenon of surface diffusion, i.e. advective water transport along 25 the adsorbed liquid film on pore walls, to explain the differences observed experimentally between dry and 26 wet cup tests. Surface diffusion becomes noticeable at high humidity levels, but in practice, it cannot be 27 distinguished from vapor transport and therefore it is lumped together with diffusion, which could cause 28 the observed differences in apparent vapor permeability. The additional liquid transport superimposed on 29 diffusion would be likely to increase the observed vapor permeability, therefore providing a lower 30 22 resistance value for wet cup measurements. Indeed, this is in agreement with the observations made for the 1 brick specimens in our study. 2 3.1.4 Capillary absorption 3 Water uptake tests for material characterization were performed on whole units, brick cubes and mortar 4 prisms in different directions. The results are summarized in Table 6. 5 Table 6. Results from water absorption tests (CoV between parentheses). 6 Material 𝐴𝑤,X 𝐴𝑤,Y 𝐴𝑤,Z 𝐴𝑤 (1) (2) 𝑤𝑐𝑎𝑝 (2) [kg/m2/s0.5] [kg/m2/s0.5] [kg/m2/s0.5] [kg/m2/s0.5] [kg/m3] B units 0.118 (8.43) 0.090 (5.51) 0.068 (7.36) -- 250 (2.35) B cubes 0.104 (6.31) 0.089 (7.78) 0.061 (5.86) -- 240 (1.07) LM (3) 0.237 (1.62) 0.234 (1.59) 0.233 (1.46) 0.235 (1.57) 225 (2.05) CM (3) 0.059 (3.96) 0.059 (5.32) 0.060 (7.77) 0.059 (6.49) 180 (2.45) (1) Isotropic (2) Average of all tests (3) Sample of 4 molded prisms The capillary moisture content, 𝑤𝑐𝑎𝑝 [kg/m3], was determined as the average moisture content at the end of 7 the water absorption tests. It must be noted that in the literature, it is possible to find 𝑤𝑐𝑎𝑝 calculated as the 8 moisture content at the intersection between linear approximations of the first and second stages of the 9 capillary absorption test. However, for porous materials with a sharp transition between the two stages and 10 a well-developed plateau during the second stage –such as the ones presented in this study–, 𝑤𝑐𝑎𝑝 calculated 11 in one or the other way is equivalent. The capillary moisture content is always lower than the saturation 12 moisture content defined from open porosity, i.e. 𝑤𝑐𝑎𝑝 <𝑤𝑠𝑎𝑡, due to air entrapment in the pore structure 13 during imbibition [55]. For prolonged imbibition or immersion times, 𝑤𝑐𝑎𝑝 will approach 𝑤𝑠𝑎𝑡 as the air 14 trapped in the pores dissolves into the water [11]. Therefore, 𝑤𝑐𝑎𝑝 is more a fuzzy limit than a fixed value. 15 For the studied materials, the comparison between saturation and capillary moisture contents gave 𝑤𝑐𝑎𝑝 ≈ 16 80%−90% 𝑤𝑠𝑎𝑡, which is an indication of the quantity of pores that can be classified as capillaries with 17 23 respect to the overall open porosity. 1 The results obtained for whole bricks are shown in Figure 6 for the three studied directions: a) header or 2 extrusion direction, labelled X; b) stretcher direction, labelled Y; c) bed direction, labelled Z. Note that 3 each set of tests ended up in a different plateau since the exposed surface was not the same for each case; 4 nonetheless, the moisture content at the end of all tests was equivalent. The rate of absorption per unit area 5 was fastest in the extrusion direction (X) and slowest in the bed direction (Z). Considering the bed direction 6 as a reference –it is usually the most reported one–, we obtained 𝐴𝑤,𝐵−𝑋≅170% 𝐴𝑤,𝐵−𝑍 and 𝐴𝑤,𝐵−𝑌≅ 7 130% 𝐴𝑤,𝐵−𝑍. This entails that the absorption behavior of a brick wall might be considerably 8 underestimated if the properties of the bed face are generalized to other directions, e.g. lateral wetting from 9 wind-driven rain. Note that the different cases analyzed here made use of the same specimens and the 10 geometrical parameters do not influence the absorption process –water inflow normalized by area and 11 heights too small to produce any significant pressure head–. Thus, the different behavior in the three 12 directions can only be explained by an anisotropic pore system or by the existence of imperfections, such 13 as hollows, fissures or cracks within the brick mass. 14 Figure 6. Water absorption results for brick units in the 3 studied directions. 15 The results of the water uptake tests in brick cubes are shown in Figure 7 for the three studied directions. 16 The same overall trends observed in full-size bricks hold for the case of brick cubes: the capillary absorption 17 rate was faster in the extrusion direction (X), followed by the stretcher direction (Y), and finally, bed 18 direction (Z) was the slowest. Using the bed direction as a reference, 𝐴𝑤,𝑏−𝑋≅170% 𝐴𝑤,𝑏−𝑍 and 𝐴𝑤,𝑌 ≅ 19 150% 𝐴𝑤,𝑏−𝑍. In this occasion, the exposed area and height were equal for all specimens, so any 20 geometrical influence was discarded. Consequently, the results point at an anisotropic structure of the 21 0100 200 300 400 500 600 700 0 10 20 30 40 50 60 Water uptake inflow [kg/m2] Time [sqrt(s)] X Y Z 24 material, the presence of internal imperfections, or a combination of both. 1 Figure 7. Water absorption results for brick cubes in the 3 studied directions. 2 By careful examination of the specimens, the main cause for the anisotropic behavior was associated with 3 the existence of rifts or canals within the bulk mass of the brick. The reader is referred to Figure 1 where 4 these imperfections are visible, with a more noticeable presence in the longitudinal direction. Therefore, 5 the forming process of the bricks –in this case, extrusion– is deemed to be the reason behind the anisotropic 6 behavior for water absorption. Furthermore, the impact of internal canals on the water uptake process may 7 be observed in brick cubes B2X and B2Y (individuated in Figure 7). For these two cases, signs of absorbed 8 water were spotted on the upper part of the specimens very early on in the process, starting from a particular 9 location and spreading from there to the rest of the surface. This could be linked to the presence of a 10 discontinuity across the specimen (bottom-to-top) from where the water was sucked up faster. In fact, the 11 absorption curves for B2X and B2Y show a distinct non-linear trend with a faster absorption rate in 12 comparison with the rest of the cases. Such behavior can be explained by the so-called “free-water” 13 transport through canals (pores with diameter greater than 100 μm), which is faster than water suction 14 through capillaries [56–58]. 15 Overall, the results obtained from full-size bricks and brick cubes are comparable, as expected. However, 16 the capillary moisture content and the water absorption coefficients (all 3 directions) are slightly higher for 17 the whole bricks. As no size effect is expected in the capillary test results [59], these differences can be 18 attributed to variations in the microstructure of the superficial layer of the bricks (which was removed in 19 the cubes) compared to the core mass [18]. The existence of an external layer may be associated with the 20 different temperatures reached during firing and consequently, different degrees of sintering of the clay and 21 density/porosity values. Moreover, local defects associated with the manufacturing process, as mentioned 22 040 80 120 160 200 240 0 2 4 6 8 10 12 B2Y Water uptake inflow [kg/m2] Time [sqrt(s)] X Y Z B2X 25 before, can be another reason for the observed difference in the capillary absorption coefficients obtained 1 from brick units and brick cubes. Note that the brick cubes that showed evidence of canals were discarded 2 for the calculation of average values although their curves are shown in the figures. 3 The results of the water absorption tests for CM and LM are shown in Figure 8. In contrast to the bricks, 4 the results did not show significant deviations for different directions, confirming the isotropic behavior of 5 these materials. The two mortars showed a noticeably distinct behavior: the absorption process in LM was 6 the fastest, even faster than the bricks, whereas CM showed a very slow water uptake rate. Taking as a 7 reference the water absorption coefficient for the bed direction of whole bricks, 𝐴𝑤,𝐿𝑀≅335% 𝐴𝑤,𝐵−𝑍 and 8 𝐴𝑤,𝐶𝑀 ≅85% 𝐴𝑤,𝐵−𝑍. Provided that these properties were directly applicable to multi-layered masonry 9 walls, the expected absorption performance would vary considerably depending on the brick-mortar 10 configuration. For combinations B+LM, the fastest path would be the mortar joint, whereas the joints in 11 B+CM cases would act as a retardant barrier. 12 Figure 8. Water absorption results for cement mortar (CM) and lime mortar (LM) prisms. 13 Due to the small thickness of the specimens, the absorption tests on LMJ did not render reliable results in 14 terms of water absorption coefficient. However, a clear plateau was reached at the end of the tests, from 15 where capillary moisture content was calculated, 𝑤𝑐𝑎𝑝,𝐿𝑀−𝑗𝑜𝑖𝑛𝑡 = 185 kg/m3 (CoV = 6.47%). This value is 16 in agreement with the relation 𝑤𝑐𝑎𝑝 ≈80%−90% 𝑤𝑠𝑎𝑡, observed for the other materials. 17 Comparison of water absorption rates between different materials has to be done with caution. It must be 18 borne in mind that the water absorption coefficient is a phenomenological parameter and does not describe 19 any real physical property. After all, different pore systems may result in the same water absorption 20 coefficient, e.g. a small group of coarse pores may produce the same water absorption rate as a greater 21 040 80 120 160 200 240 0 2 4 6 8 10 12 Water uptake inflow [kg/m2] Time [sqrt(s)] CM LM 32 M2-4(a-to-b) and M2-5(b-to-a) (individuated in Figure 12), presented a considerable absorption rate drop 1 at the first interface, most probably associated with local defects. Conversely, a more noticeable deviation 2 was generally observed between the mortar joint and the uppermost brick. This confirms the directional 3 behavior of the interface and points at a microstructure change between B and LMJ. The pore quality 4 difference between both materials is therefore assumed to be the main factor affecting the suction capacity 5 of the following layer. In this case, the porous medium with finer pores tends to withhold water due to the 6 higher capillary suction and thus the liquid transfer to a material with bigger pores is impeded. 7 Considering the average behavior of all the cases (black curves in Figure 12), the response of masonry 8 specimens M2 showed no significant hydraulic resistance at the first interface (water moving from B to 9 LMJ), whereas the second interface (water moving from LMJ to B) involved a certain flow reduction. 10 Furthermore, the match between the averages of the two tested configurations proves that the behavior of 11 the composite was independent of the analyzed direction. 12 3.2.2 Drying tests 13 The 1-D drying tests performed on two-layer (1 interface) masonry composites are analyzed in the 14 following section. Figure 13 collects the results obtained for the two tested configurations, namely cylinders 15 drying from the LMJ surface, labelled D1-M, and cylinders drying from the brick layer, labelled D1-B. The 16 results are plotted as gravimetric moisture content with respect to the dry mass of the whole specimen 17 (B+LMJ). For each case, the location of the interface is estimated from the capillary saturation level 18 calculated from the capillary moisture content and the average volume of the exposed layer. 19 Figure 13. Drying test results for masonry specimens D1 (B+LMJ). Grey curves represent test results and 20 black curves are the corresponding average. The location of the interfaces is estimated from the average 21 volume of the exposed material layer. 22 0200 400 600 800 1000 0.00 0.03 0.06 0.09 0.12 0.15 D1-M D1-B Gravimetric moisture content [kg/kg] Time [h] Interface D1-B Interface D1-M 33 The curves in Figure 13 show two distinct drying trends depending on the configuration of the composite. 1 The specimens drying from the brick layer experienced a long initial phase (Stage I with linear mass loss) 2 and a quite fast second stage. The same trend was observed for the drying tests of B and LM cubes (Figure 3 9, Figure 10). This response means that the stacked materials have the capacity to transport the liquid water 4 effectively towards the surface where evaporation takes place. Conversely, the cylinders drying from LMJ 5 showed a shorter initial drying phase and a prolonged, gradual second stage. This trend contrasts with the 6 drying behavior of B and LM and is more related to the behavior of CM (Figure 9, Figure 10). Therefore, 7 the LMJ on top rapidly exhausts its liquid transport capacity and then the drying behavior slows down 8 considerably, shifting to vapor diffusion as the main transport mechanism for drying. This behavior may 9 be explained by the presence of smaller pores in LMJ since finer pores withhold water due to their higher 10 capillary pressure. Additionally, the presence of smaller pores is supported by the residual moisture content 11 at the end of the experiments, which indicates the existence of a hygroscopic component, which was already 12 anticipated by the sorption isotherms for LMJ (Figure 5d). 13 When comparing both cases, it is evident that the LMJ layer on top limits the drying rate of B below, which 14 would be naturally faster as witnessed in the D1-B configuration. Considering the sharp front 15 approximation to be extensible to drying processes, a receding drying front may be assumed [64]. Taking 16 into account this receding moisture front, the existence of an interfacial resistance was expected to appear 17 as a discontinuity in the mass loss curve, as it was observed for the water absorption cases in multi-layered 18 compounds. Nonetheless, the mass loss figures reveal no clear influence of the interface. In general, the 19 curves are smooth and no significant change is perceived beyond the transition from Stage I to Stage II. 20 The effect of the mortar-brick interface in the drying process was also studied by Brocken [27] for brick- 21 cement masonry specimens. Similarly, the author could not qualify the nature of the interface as having 22 perfect or imperfect hydraulic contact. 23 Without further certainties about the role of the interface, we may assume a perfect hydraulic contact. In 24 this scenario, two causes for the observed trends may be mentioned. On one hand, it seems conclusive that 25 B has higher liquid transport capacity as well as higher vapor permeability. This would explain the longer 26 Stage I and a considerably rapid Stage II for the configuration D1-B. Conversely, considering the D1-M 27 cases, the liquid transport capacity of LMJ wears out quicker and a likely lower vapor permeability further 28 limits the drying process of the brick layer behind. The low liquid permeability of LMJ observed in these 29 drying tests contrasts with the absorption capacity of the same material observed in the previous water 30 34 uptake experiments. The disparity between absorption and drying confirms the hysteretic behavior observed 1 in moisture storage for LMJ (Figure 5d). 2 4. CONCLUSIONS 3 An extensive experimental campaign was conducted to determine the hygric response of masonry materials. 4 First, the hygric properties of constituent materials were assessed individually. Then, masonry was studied 5 as a multi-layered composite material. The experimental program included vacuum saturation tests, static 6 gravimetric tests for adsorption/desorption, dry/wet cup tests, capillary absorption tests, and 1-D isothermal 7 drying. The hygric properties determined following the prescribed experimental procedures make up a 8 consistent dataset and allow us to draw the following conclusions: 9 a) The results from the different experimental procedures are consistent and present acceptable variation. 10 Nonetheless, vapor permeability measurements present poorer reliability, which has been documented 11 by other authors as well, e.g. [16–18,54]. 12 b) Extruded fired-clay bricks show a distinct anisotropic hygric behavior, more noticeable for faster 13 processes such as water absorption. This behavior is explained by the existence of imperfections within 14 the mass of the bricks derived from the extrusion process. 15 c) Differences are observed between the hygric response of the two studied molded mortars. This is most 16 evident when comparing the moisture storage curves. LM shows low hygroscopicity and a strong 17 capillary-active behavior with negligible hysteresis. Conversely, CM shows hygric activity in both the 18 hygroscopic and capillary ranges with a marked hysteresis. Thus, the behavior of LM is much closer to 19 that of the studied fired-clay brick. 20 d) The properties of the NHL-mortar prepared in molds differ from the same mixture cured in masonry 21 bed joints. The discrepancies are associated with the different water-binder ratio resulting from water 22 extraction by the bricks upon placement. The vacuum saturation tests on LMJ reveal lower open 23 porosity and therefore lower saturation moisture content. The moisture storage curves obtained for LMJ 24 show a more hygroscopic behavior than its molded counterpart, with non-negligible hysteresis. Water 25 absorption tests performed on masonry cylinders M1-M show slower water uptake than expected from 26 LM, thus a lower water absorption coefficient is assumed for LMJ. Furthermore, isothermal drying tests 27 on masonry specimens drying from LMJ reveal mass loss curves with a short initial drying stage, which 28 35 is indicative of an enhanced water retention capacity and the presence of fine pores. Finally, the 1 dissimilar trends observed for water uptake and drying processes confirm the hysteresis in moisture 2 storage for LMJ. 3 e) In masonry specimens, the existence of an imperfect hydraulic contact at the brick-mortar interface is 4 demonstrated for water absorption. The imperfect contact may be assumed as a hydraulic resistance and 5 its effect varies from a slight reduction to considerable retardation of the water flow. The interfacial 6 effect is mostly detected when the water movement goes from LMJ to B. Thus, discontinuity in the pore 7 structure of the materials (finer pores in LMJ) is assumed as the main reason for the interfacial effect, 8 which may become more pronounced due to the presence of local imperfections. The interfacial effect 9 on drying kinetics is not evident. 10 f) Hygric properties obtained individually from testing the constituent materials separately cannot be 11 directly extrapolated to multi-layered masonry cases. Two main phenomena must be taken into 12 consideration in order to avoid erroneous estimations, namely: (a) for water absorption cases, there is a 13 hydraulic interfacial resistance derived from the dissimilar pore structure between the materials in 14 contact; (b) for water absorption and drying cases, the impact of curing conditions on the properties of 15 mortars cast in masonry joints must be considered. 16 The current study provides a comprehensive insight into the hygric properties of masonry materials by 17 relatively simple means. MIP results would be useful to confirm our conclusions related to the pore structure 18 of the materials. Additionally, more complex methods, such as NMR, CT scan or X-Ray, could further 19 improve our understanding since they provide moisture profiling inside the specimens. However, such 20 techniques are costly and not always available. Provided that the sharp front approximation is applicable, 21 the methodology presented in this paper provides all necessary data for a clear depiction of moisture transfer 22 processes in multi-layered porous materials such as masonry. 23 24 Acknowledgements 25 The first author would like to thank his fellow PhD students at the University of Minho, in particular Meera 26 Ramesh and Giorgos Karanikoloudis, for providing fundamental support with the experimental work. 27 28 36 Funding 1 This work was financed by national funds through FCT - Foundation for Science and Technology, under 2 grant agreement SFRH/BD/117114/2016 attributed to the first author. 3 4 References 5 [1] M. Harrestrup, S. Svendsen, Full-scale test of an old heritage multi-storey building undergoing 6 energy retrofitting with focus on internal insulation and moisture, Build. Environ. 85 (2015) 123– 7 133. https://doi.org/10.1016/j.buildenv.2014.12.005. 8 [2] E. Vereecken, L. Van Gelder, H. Janssen, S. Roels, Interior insulation for wall retrofitting – A 9 probabilistic analysis of energy savings and hygrothermal risks, Energy Build. 89 (2015) 231– 10 244. https://doi.org/10.1016/j.enbuild.2014.12.031. 11 [3] M. van Aarle, H. Schellen, J. van Schijndel, Hygro-thermal simulation to predict the risk of frost 12 damage in masonry – Effects of climate change, Energy Procedia. 78 (2015) 2536–2541. 13 https://doi.org/10.1016/j.egypro.2015.11.268. 14 [4] F.M. Fernandes, Clay bricks, in: B. Ghiassi, P.B. Lourenço (Eds.), Long-Term Perform. Durab. 15 Mason. Struct., Elsevier, 2019: pp. 3–19. https://doi.org/10.1016/B978-0-08-102110-1.00001-7. 16 [5] M. do Rosário Veiga, A. Santos Silva, Mortars, in: B. Ghiassi, P.B. Lourenço (Eds.), Long-Term 17 Perform. Durab. Mason. Struct., Elsevier, 2019: pp. 169–208. https://doi.org/10.1016/B978-0-08- 18 102110-1.00006-6. 19 [6] J.J. Brooks, Moisture Movement of Clay Brick Masonry, in: Concr. Mason. Movements, 20 Elsevier, 2015: pp. 223–253. https://doi.org/10.1016/B978-0-12-801525-4.00008-X. 21 [7] M.S. Sciolti, M.A. Aiello, M. Frigione, Effect of thermo-hygrometric exposure on FRP, natural 22 stone and their adhesive interface, Compos. Part B Eng. 80 (2015) 162–176. 23 https://doi.org/10.1016/j.compositesb.2015.05.041. 24 [8] J.F. Straube, E.F.P. Burnett, Overview of Hygrothermal (HAM) Analysis Methods, in: H.R. 25 Trechsel (Ed.), Moisture Anal. Condens. Control Build. Envel. – ASTM Man. Ser. MNL 40, 26 American Society for Testing and Materials, Philadelphia, 2001: pp. 81–89. 27 37 [9] H.M. Künzel, A.N. Karagiozis, A.H. Holm, A Hygrothermal Design Tool for Architects and 1 Engineers (WUFI ORNL/IBD), in: H.R. Trechsel (Ed.), Moisture Anal. Condens. Control Build. 2 Envel. – ASTM Man. Ser. MNL 40, American Society for Testing and Materials, Philadelphia, 3 2001: pp. 136–151. 4 [10] H. Hens, Modeling the Heat, Air, and Moisture Response of Building Envelopes: What Material 5 Properties are Needed, How Trustful Are the Predictions?, J. ASTM Int. 4 (2007) 1–11. 6 https://doi.org/10.1520/JAI100460. 7 [11] G. Scheffler, Validation of hygrothermal material modelling under consideration of the hysteresis 8 of moisture storage, Ph.D. dissertation, Dept. Civil Engineering, Dresden University of 9 Technology, 2008. 10 [12] K. Kumaran, IEA Annex 24: Heat, Air and Moisture Transfer in Insulated Envelope Parts. Final 11 Report, Volume 3, Taks 3: Material Properties, 1996. 12 [13] S. Roels, J. Carmeliet, H. Hens, O. Adan, H. Brocken, R. Czerny, C. Hall, A. Hamilton, K. 13 Kumaran, Z. Pavlik, L. Pel, R. Plagge, F. Tariku, HAMSTAD Work Package 1: Final Report - 14 Moisture Transfer Properties and Materials Characterisation, 2003. 15 [14] G. Cultrone, E. Sebastián, K. Elert, M.J. de la Torre, O. Cazalla, C. Rodriguez–Navarro, 16 Influence of mineralogy and firing temperature on the porosity of bricks, J. Eur. Ceram. Soc. 24 17 (2004) 547–564. https://doi.org/10.1016/S0955-2219(03)00249-8. 18 [15] F.M. Fernandes, P.B. Lourenço, F. Castro, Ancient clay bricks: manufacture and properties, in: 19 D.M. Bostenaru, R. Přikryl, A. Török (Eds.), Mater. Technol. Pract. Hist. Herit. Struct., Springer, 20 2010. https://doi.org/10.1007/978-90-481-2684-2. 21 [16] S. Roels, J. Carmeliet, H. Hens, O. Adan, H. Brocken, R. Cerny, Z. Pavlik, C. Hall, K. Kumaran, 22 L. Pel, R. Plagge, Interlaboratory Comparison of Hygric Properties of Porous Building Materials, 23 J. Therm. Envel. Build. Sci. 27 (2004) 307–325. https://doi.org/10.1177/1097196304042119. 24 [17] C. Feng, H. Janssen, Y. Feng, Q. Meng, Hygric properties of porous building materials: Analysis 25 of measurement repeatability and reproducibility, Build. Environ. 85 (2015) 160–172. 26 https://doi.org/10.1016/j.buildenv.2014.11.036. 27 38 [18] C. Feng, A.S. Guimarães, N. Ramos, L. Sun, D. Gawin, P. Konca, C. Hall, J. Zhao, H. Hirsch, J. 1 Grunewald, M. Fredriksson, K.K. Hansen, Z. Pavlík, A. Hamilton, H. Janssen, Hygric properties 2 of porous building materials (VI): A round robin campaign, Build. Environ. 185 (2020) 107242. 3 https://doi.org/10.1016/j.buildenv.2020.107242. 4 [19] Y. Aït Oumeziane, A. Pierre, F. El Mankibi, V. Lepiller, M. Gasnier, P. Désévaux, Hygrothermal 5 properties of an early 20th century clay brick from eastern France: Experimental characterization 6 and numerical modelling, Constr. Build. Mater. 273 (2021) 121763. 7 https://doi.org/10.1016/j.conbuildmat.2020.121763. 8 [20] R.J. Gummerson, C. Hall, W.D. Hoff, Water movement in porous building materials II. Hydraulic 9 suction and sorptivity of brick and other masonry materials, Build. Environ. 15 (1980) 101–108. 10 https://doi.org/10.1016/0360-1323(80)90015-3. 11 [21] K.J. Krakowiak, P.B. Lourenço, F.J. Ulm, Multitechnique Investigation of Extruded Clay Brick 12 Microstructure, J. Am. Ceram. Soc. 94 (2011) 3012–3022. https://doi.org/10.1111/j.1551- 13 2916.2011.04484.x. 14 [22] J.E. Lindqvist, K. van Balen, B. Bicer-Simsir, L. Binda, C. Bla¨uer, J. Elsen, E. Hansen, R. van 15 Hees, F. Henriques, E.E. Toumbakari, T. von Konow, J.E. Lindqvist, P. Maurenbrecher, B. 16 Middendorf, I. Papayanni, S. Simon, M. Subercaseaux, C. Tedeschi, M. Thompson, J. Valek, 17 M.R. Valluzzi, Y. Vanhellemont, R. Veiga, A. Waldum, Rilem TC 203-RHM: Repair mortars for 18 historic masonry. Testing of hardened mortars, a process of questioning and interpreting, Mater. 19 Struct. 42 (2009) 853–865. https://doi.org/10.1617/s11527-008-9455-x. 20 [23] C. Groot, J. Gunneweg, The influence of materials characteristics and workmanship on rain 21 penetration in historic fired clay brick masonry, Heron. 55 (2010) 141–154. 22 [24] C. Nunes, L. Pel, J. Kunecký, Z. Slížková, The influence of the pore structure on the moisture 23 transport in lime plaster-brick systems as studied by NMR, Constr. Build. Mater. 142 (2017) 395– 24 409. https://doi.org/10.1016/j.conbuildmat.2017.03.086. 25 [25] J.M.P.Q. Delgado, A.C. Azevedo, A.S. Guimarães, Drying Kinetics in Building Materials and 26 Components: The Interface Influence, Springer, 2019. https://doi.org/10.1007/978-3-030-31860- 27 4. 28 39 [26] K. Calle, T. De Kock, V. Cnudde, N. Van den Bossche, Liquid moisture transport in combined 1 ceramic brick and natural hydraulic lime mortar samples: Does the hygric interface resistance 2 dominate the moisture transport?, J. Build. Phys. 43 (2019) 208–228. 3 https://doi.org/10.1177/1744259119857762. 4 [27] H.J.P. Brocken, Moisture transport in brick masonry: the grey area between bricks, Ph.D. 5 dissertation, Dept. Built Environment, Technische Universiteit Eindhoven, 1998. 6 [28] M.A. Wilson, W.D. Hoff, C. Hall, Water movement in porous building materials—XIII. 7 Absorption into a two-layer composite, Build. Environ. 30 (1995) 209–219. 8 [29] M.A. Wilson, W.D. Hoff, C. Hall, Water movement in porous building materials—XIV. 9 Absorption into a two-layer composite (SA < SB), Build. Environ. 30 (1995) 221–227. 10 [30] C. Hall, W.D. Hoff, Water Transport in Brick, Stone and Concrete, 2nd ed., CRC Press, London, 11 2009. 12 [31] C. Groot, The characterisation of brick and mortar considering mortar/brick bond, in: M. Wu, Y. 13 Qian (Eds.), Proc. 11th Int. Brick/Block Mason. Conf., ECS (Engineering Construction 14 Standardization), Shanghai, 1997: pp. 50–58. 15 [32] C. Groot, J. Larbi, Influence of water flow (reversal) on bond strength development in young 16 masonry, Heron. 44 (1999) 63–78. 17 [33] H. Derluyn, H. Janssen, J. Carmeliet, Influence of the nature of interfaces on the capillary 18 transport in layered materials, Constr. Build. Mater. 25 (2011) 3685–3693. 19 https://doi.org/10.1016/j.conbuildmat.2011.03.063. 20 [34] H. Janssen, H. Derluyn, J. Carmeliet, Moisture transfer through mortar joints: A sharp-front 21 analysis, Cem. Concr. Res. 42 (2012) 1105–1112. 22 https://doi.org/10.1016/j.cemconres.2012.05.004. 23 [35] X. Zhou, G. Desmarais, P. Vontobel, J. Carmeliet, D. Derome, Masonry brick–cement mortar 24 interface resistance to water transport determined with neutron radiography and numerical 25 modeling, J. Build. Phys. 44 (2020) 251–271. https://doi.org/10.1177/1744259120908967. 26 [36] BS EN 197-1:2011, Cement. Composition, specifications and conformity criteria for common 27 40 cements, (2011). 1 [37] BS EN ISO 12570:2000 + A1:2013, Hygrothermal performance of building materials and 2 products - Determination of moisture content by drying at elevated temperature, (2000). 3 [38] C. Feng, H. Janssen, C. Wu, Y. Feng, Q. Meng, Validating various measures to accelerate the 4 static gravimetric sorption isotherm determination, Build. Environ. 69 (2013) 64–71. 5 https://doi.org/10.1016/j.buildenv.2013.08.005. 6 [39] RILEM TC 25-PEM, Recommended tests to measure the deterioration of stone and to assess the 7 effectiveness of treatment methods (Test No. I.1 Porosity accessible to water), Matériaux Constr. 8 13 (1980) 176–179. 9 [40] BS EN ISO 12571:2013, Hygrothermal performance of building materials and products – 10 Determination of hygroscopic sorption properties, (2013). 11 [41] M. Krus, Moisture Transport and Storage Coefficients of Porous Mineral Building Materials: 12 Theoretical Principles and New Test Methods, Fraunhofer IRB Verlag, Stuttgart, 1996. 13 [42] ASTM E96/E96M-10, Standard Test Methods for Water Vapor Transmission of Materials, 14 (2010). https://doi.org/10.1520/E0096. 15 [43] BS EN ISO 12572:2016, Hygrothermal performance of building materials and products – 16 Determination of water vapour transmission properties – Cup method, (2016). 17 [44] EN 15803:2009, Conservation of cultural property - Test methods - Determination of water 18 vapour permeability (δp), (2009). 19 [45] J.L. Monteith, M.. Shatleworth, Principles of Environmental Physics, 4th ed., Elsevier Ltd., 20 London, 2013. 21 [46] BS EN ISO 15148:2002+A1:2016, Hygrothermal performance of building materials and products 22 – Determination of water absorption coefficient by partial immersion, (2016). 23 [47] C. Feng, H. Janssen, Hygric properties of porous building materials (II): Analysis of temperature 24 influence, Build. Environ. 99 (2016) 107–118. https://doi.org/10.1016/j.buildenv.2016.01.016. 25 [48] EN 16322:2013, Conservation of cultural property - Test methods - Determination of drying 26 41 properties, (2013). 1 [49] Z. Zhang, G. Scherer, Determination of water permeability for a moisture transport model with 2 minimized batch effect, Constr. Build. Mater. 191 (2018) 193–205. 3 https://doi.org/10.1016/j.conbuildmat.2018.09.194. 4 [50] H.M. Künzel, Simultaneous heat and moisture transport in building components: One- and two- 5 dimensional calculation using simple parameters, Ph.D. dissertation, Dept. Building Physics, 6 Fraunhofer Institute of Building Physics, 1995. 7 [51] Y. Mualem, A new model for predicting the hydraulic conductivity of unsaturated porous media, 8 Water Resour. Res. 12 (1976) 593–622. 9 [52] S. Brunauer, P.H. Emmett, E. Teller, Adsorption of Gases in Multimolecular Layers, J. Am. 10 Chem. Soc. 60 (1938) 309–319. 11 [53] M.T. van Genuchten, A closed form equation predicting the hydraulic conductivity of unsaturated 12 soils, Soil Sci. Soc. Am. J. 44 (1980) 892–898. 13 [54] H. Hens, Applied Building Physics: Ambient Conditions, Building Performance and Material 14 Properties, 2nd ed., Ernst & Sohn, Berlin, 2016. https://doi.org/10.1002/9783433607114. 15 [55] F. Descamps, Continuum and discrete modelling of isothermal water and air flow in porous 16 media, Ph.D. dissertation, Building physics laboratory, KU Leuven, 1997. 17 [56] C. Groot, J. Gunneweg, Two views to deal with rain penetration problems in historic fired clay 18 brick masonry, in: J. Válek, J.J. Hughes, C.J.W.P. Groot (Eds.), Proc. 2nd Conf. Hist. Mortars - 19 HMC 2010 Final Work. RILEM TC 203-RHM, RILEM Publications SARL, Dordrecht, 2010: pp. 20 1005–1014. 21 [57] S. Roels, K. Vandersteen, J. Carmeliet, Measuring and simulating moisture uptake in a fractured 22 porous medium, Adv. Water Resour. 26 (2003) 237–246. https://doi.org/10.1016/S0309- 23 1708(02)00185-9. 24 [58] J. Carmeliet, J.-F. Delerue, K. Vandersteen, S. Roels, Three-dimensional liquid transport in 25 concrete cracks, Int. J. Numer. Anal. Methods Geomech. 28 (2004) 671–687. 26 https://doi.org/10.1002/nag.373. 27