Evaporation from Porous Building Materials and Its 3 Cooling Potential

5 Abstract: Evaporative 3 cooling is a traditional strategy to improve summer comfort, which has gained renewed relevance in the context of the 6 transition to a greener economy. Here, the potential for evaporative cooling of two common porous building materials, natural stone and 7 ceramic brick, was evaluated. The work has relevance also to the protection of built heritage becauseevaporation underlies the problems of 8 dampness and salt crystallization, which are so harmful and frequent in this heritage. It was observed that the drying rate of the materials is, in 9 some cases, higher than the evaporation rate of a free water surface. Surface area measurements by a three-dimensional optical technique 10 suggested, as probable cause of this behavior, that surface irregularity gives rise to a large effective surface of evaporation in the material. 11 Surface temperature measurements by infrared were performed afterward during evaporation experiments outside during a hot summer day in 12 Lisbon. Their results indicate that ordinary building materials can be very efficient evaporative media and, thus, may help in achieving higher 13 energy efficiency while maintaining a simultaneous constructive or architectural function. DOI: 10.1061/(ASCE)MT.1943-5533.0001174. 14 © 2014 American Society of Civil Engineers.

Water is a constant presence in the intricate pore network of tradi-18 tional building materials, such as mortar, stone, or ceramics.It may 19 have very harmful effects because it functions as a catalyst for 20 deterioration mechanisms, such as sulphate attack, biodeteriora-21 tion, or salt decay.But it can also have positive effects, for example, 22 when it enables the evaporative cooling of environments.23 The evaporative drying of porous materials involves liquid 24 transport toward an evaporation front and vapor transport from that 25 front outward (Sherwood 1929;Scherer 1990).Two main regimes 26 are in general considered, which for a material drying from satu-27 ration, correspond to the following main stages (Fig. 1): in Stage I, 28 also called the constant drying rate period (CDRP), there is liquid 29 continuity across the material, and the wet front is located at their 30 surface; the drying rate is constant because drying proceeds under 31 steady-state conditions.Stage II, also called the falling drying rate 32 period (FDRP), starts when the moisture content in the material, 33 and therefore the liquid flow, decreases to a point where it is no 34 longer able to compensate the evaporative demand, and the wet 35 front begins receding toward the interior of the material.
During the CDRP, the drying rate is at its highest value.This value is often assumed to be equal to that of a free water surface, which would be explained by the presence of a liquid film covering the whole surface of the material.However, this idea has been contradicted by researchers, such as Hammecker (1993), Jeannette (1997), Tournier et al. (2000), Rousset-Tournier (2001), and Diaz Gonçalves et al. (2012).These researchers observed that the evaporation rate from porous stones and other building materials during the CDRP was not necessarily equal to the evaporation from a free water surface and, in fact, could even be higher than that.A similar phenomenon was observed by Tang and Etzion (2004) who noticed that, with a low wind velocity, the rate of evaporation from a pond was greater when it was covered with wet tissue.
The possible enhancement of the CDRP drying rate of a porous material in comparison to a flat water surface has a wide range of implications.Indeed, evaporation is often used as a boundary condition in numerical models for moisture transport in porous media, and the most current reference for the CDRP is the evaporation rate of a flat water surface.The study of evaporative processes is also extremely important from more practical perspectives, such as the protection of the architectural heritage.The evaporation rate determines, for example, the height of capillary rise (I'Anson and Hoff 1984), a chronic problem in historical buildings (Massari and Massari 1993).Also, salt decay, one very harmful degradation mechanism that affects this type of building (Charola 2000), happens precisely during evaporative processes by which the solutions increase their concentration until they saturate and eventually crystallize.The study of evaporative drying is therefore fundamental to understand and ultimately develop solutions for these degradation processes.Finally, such study is also relevant from the point of view of sustainability because it is the base of evaporative cooling, one of the oldest strategies for improving summer comfort in hot, dry environments.The wetting of ceramic floors, traditional in Mediterranean countries, such as Portugal, can be mentioned as an example.Evaporative cooling methods rely on the fact that the passage of water from liquid to vapor state involves energy consumption (Matias et al. 2007).These cooling methods have recently gained a new importance and are more and more often incorporated The materials used for determination of the CDRP drying rate are 87 rigid building materials that encompass six natural stones, a red ceramic brick, an air lime/sand mortar, and three calcium silicate materials (Table 1).They were chosen on an exploratory basis for being representative of those used in civil engineering and found in the built heritage and also because they cover a wide range of capillary porosity.This porosity, which corresponds to the natural capacity of the material to absorb water at atmospheric pressure, is also given in Table 1.The pore-size distribution of 10 of the materials is presented in Fig. 2.An Autoscan60 porosimeter from Quantachrome was used, with a pressure range between 0 and 320 MPa.  2).

Measurement of CDRP Drying Rate
The CDRP drying rate was measured by means of drying tests (RILEM 1980).These tests followed a method similar to that de-    ).In these cases, it was considered reasonable to use only two specimens, as the preceding results were very homogeneous (Diaz Gonçalves et al. 2012).For the free water surfaces, three full petri dishes were always used.
The environmental conditions inside the drying box were continuously monitored by means of a Mikromec Multisens sensor positioned in its center (Table 3).The measurements started before the specimens and petri dishes were placed inside the box and proceeded until after they were removed from it.As seen in Fig. 3 25%, 44%, 66%, and 79%, respectively (Table 3).
The result of a drying test is a graph depicting the mass of the specimens as function of time (Fig. 4).The drying rate of the specimen, in g=h, is the slope of the mass-time function.This value is then divided by the area of the top surface of the specimen to obtain the amount of water evaporated per unit area, in g=ðm 2 • hÞ.

Measurement of Surface Area by Optical Method
The surface texture of 10 materials (those in Table 1, except the Maastricht limestone) was studied using the 3D optical measuring   1).Distilled water and 187 several incoherent materials (three grades of siliceous sand, saw-188 dust, and cellulose) were also tested to serve as reference (Table 2).The base of these boxes was perforated to allow capillary absorption, and a filter paper was put on the inside base to avoid material loss.
The materials were wetted by capillary absorption through the base by means of partial immersion in water for 48 h in a conditioned room (20°C and 50% RH).After this period, the samples were removed from immersion and its lower face immediately sealed with polyethylene film to ensure that drying would be unidirectional, taking place only through the upper surface.
The cubes and containers were fit in openings cut in a XPS board with dimensions of 800 mm × 800 mm × 50 mm, as shown in Fig. 5. Two specimens of each material were used, one wet and one dry, as well as two containers (AD1 and AD2) filled with distilled water.The dry materials, which served as a reference for subsequent interpretation of results, were previously dried in an oven during 24 h at 60°C, followed by 24 h in the conditioned room at

Assessment of the Emissivity
The infrared (IR) thermometer measures the amount of energy (radiance E, in W=m 2 ) emitted by an object.Then, based on an emissivity value entered by the operator, it calculates the surface temperature of that object through Eq. ( 2) which is based on Stephan Boltzmann law (Matias 2012) In    Despite these variation factors, there is an approximately linear relationship between the two quantities.This linearity means that Fick's law [Eq.( 3)] is obeyed, which corresponds to an essentially diffusive process (Fig. 8).It also means that the thickness of the stagnant air layer δ adjacent to the material does not vary with the RH, which happens because the drying tests were performed within a closed box and, thus, air velocity was always close to zero

303
The fact that in some cases the CDRP drying rate is lower than 304 for a free water surface confirms that there is not a liquid film cover-305 ing the total surface of the materials during the CDRP.If such a film 306 existed, assuming that the liquid possesses the same thermody-307 namic properties in the pores and in a free surface, the drying rate 308 of the material could perhaps be higher than for the free surface, 309 due to surface irregularity, but it could never be lower.These features indicate that the measured surfaces have fractal properties (Mandelbrot 1967(Mandelbrot , 1998)) Due to the topological complexity of the pore space, a higher effective surface of evaporation is therefore a likely explanation for the high CDRP drying rate depicted by some of the tested materials.
In Fig. 10, the CDRP drying rate is shown as a function of capillary porosity.Point (0,0) is attributed to a theoretical material with 0% porosity.Since it is admitted that surface irregularity derives from the presence of pores, this theoretical material would be totally flat.As can be seen in the figure, when the two calcium silicate materials with higher porosity are considered, the relationship between the CDRP drying rate and capillary porosity cannot be described by a linear function.Instead, a parabolic function may, for example, be used as a first-order approximation.This means that the CDRP drying rate will increase with increasing porosity but only up to a certain value.Any further rise in the porosity will result, rather, in a decrease of the drying rate.At a certain point, the situation of a free water surface is reached, which corresponds to the maximum possible porosity (P ¼ 1).This behavior is probably due to the mentioned variation of the complexity of the physical surface (which is null for the two extremes (P ¼ 0 and P ¼ 1) and higher for the intermediate situations (P0,1½).However, it is important to mention that a clear correlation between the CDRP evaporation rate and the RA values measured with the profilometer was not found.The reason could be the fact that RA varies with the measurement scale and the scale of interest is not necessarily the same for the different materials.Another reason could be that menisci curvature is also relevant in terms of effective surface of evaporation.These subjects clearly require further investigation.In the shade (Fig. 11), the wet building materials achieve lower surface temperatures than the free water surface (AD) and depict no significant difference in relation to, for example, cellulose (C), which is a product typically used in evaporative cooling devices.

Surface Temperature during Drying
Under the sun (Fig. 12), the situation is different.The brick (T), the coarse sand (AG), the mortar (A), and the sawdust (S) provide, in this case, surface temperatures of 1.2°C to 6.3°C higher than the free water surface (AD).However, one of the limestones (M), the calcium silicate (CS), and two sands (AM and AF), achieve surface temperatures of 1.7°C to 2.8°C lower than the free water surface, although about 2.0°C to 3.1°C higher than cellulose (C).
These results indicate, therefore, that ordinary building materials have interesting evaporative cooling potential.In the future, it would be useful to investigate how much the heat capacity of the materials and their coefficient of solar absorption contribute to the differences among them.It would also be important in the future to

1 National
Laboratory for Civil Engineering (LNEC), Materials Dept., Av. do Brasil 101, 1700-066 Lisbon, Portugal (corresponding author).E-mail: teresag@lnec.ptsolutions because they can help meet 75 energy-efficiency needs.76 In this study, the drying rate of porous building materials during 77 the CDRP is experimentally analyzed.The objective was to inves-78 tigate the possible enhancement of the CDRP drying rate in com-79 parison to a flat water surface.To support the interpretation of 80 the results, surface area measurements by a three-dimensional 81 (3D) optical technique were carried out.Afterward, the evaporative 82 cooling potential of some of the materials was assessed through the 83 measurement of surface temperature with an infrared thermometer 84 Materials and Methods 85 Materials 86 The measurements were performed according to ASTM D4404-10 standard (ASTM 2010) and were always replicated.The pore-size distribution of the remaining material, Maastricht limestone, can be found elsewhere (De Clercq et al. 2007).Incoherent materials like sand, sawdust, and cellulose, were used as reference in the evaporative cooling tests (Table

F1: 1 Fig. 1 .
A schematic representation of the two main drying stages of a F1:2 porous material

Fig. 2 .
Pore-size distribution as determined by mercury instrusion porosimetry (MIP) small cubes with a 24-mm edge.The mor-117 tar cubes were made using metallic molds.The other specimens 118 were sawed from larger stone blocks, ceramic bricks, calcium 119 silicate brick, or insulation boards.All the cubes were brushed 120 to remove as much stone or brick powder as possible from their 121 surfaces.Then, they were cleaned in an ultrasonic cleaner (model 122 B1200 E-1, Branson Ultrasonics Corporation, United States).123 Finally, they were laterally sealed with epoxy.124 For every condition, the materials and free water surfaces were 125 tested simultaneously.The specimens were first saturated by partial 126 immersion in pure water during three days.Afterward, their bottom 127 surface was sealed with polyethylene film (PE).They were then left 128 to dry and were periodically weighted during 8 h.129 Three test specimens of each kind of material were used at every 130 condition, except for four of the previously tested materials (B, 131 MB, CA, and T 5 , both the initial RH and the final RH are similar to the nominal RH.This shows that when the wet materials and petri dishes are not inside the box, the actual RH (eventually) assumed the values expected in each case.The localized perturbations of the RH inside the box were due to the periodic opening of the box and removal of the specimens to weight them.The actual RH considered for this work was the average of the RH measured during the CDRP in each test: instrument Talysurf CLI 1000, by Taylor Hobson.The instrument was equipped with a (noncontact) white light CLA gauge with a vertical range of 3 mm, vertical resolution of 100 nm, lateral resolution of 5 μm and measuring slope of 13°(Taylor Hobson 2009).The measurements were carried out in 3D with the highest possible resolution, which corresponds to a spacing of 5 μm in both the X and Y directions.A velocity of 2 mm=s was chosen because it is the highest possible at the selected resolution.Areal parameter Sdr (ISO 2012) was calculated after the measurements using the Talymap Gold software.Sdr is the developed interfacial area ratio and expresses the percentage of additional surface area contributed by the texture, as compared to the projected area.Using the resampling operator of the Talymap software, Sdr could be calculated for different measurement scales: 5 μm (the original measurement step), 10 μm, 20 μm, 50 μm, 100 μm, and 200 μm.From the Sdr value, the relative area (RA) could be

189
Cubic specimens of the rigid materials were used with 50-mm 190 edge.For the calcium silicate it was necessary to use samples with a 191 lower height, 35 mm, owing to the dimensions of the original 192 board.The four lateral sides of these specimens were sealed with 193 epoxy.The incoherent materials and water were placed in acrylic 194 boxes with internal dimensions of 50 mm × 50 mm × 50 mm.
20°C and 50% RH.After fitting all the materials and containers for water in the XPS board, the assembly was wrapped in PE film to prevent evaporation [Fig.5(a)].The water was placed in a closed container.The experimental device and the container with water were then transported and left in the selected place (in the sun) at 12 p.m.They remained in these conditions for 1 h to stabilize their temperature, after which the film was cut and the two acrylic containers filled with water.The surface temperature measurements [Fig.5(b)] began immediately and were repeated every 15 min for 1.5 h.The temperature was measured in the center of the top surface of the specimens, with the equipment positioned perpendicularly and at a distance of 350 mm from this surface.The environmental conditions (temperature and RH) were evaluated with a digital thermohygrometer.
this equation, ε (dimensionless) = the emissivity of the material, which represents the relation between the radiance of the body under examination and that of a black body (body that absorbs all radiation); σ = the Stefan-Boltzmann constant, which takes an absolute value of 5.67 × 10 −8 W=ðm 2 K 4 Þ; and T = the temperature (K).

F4: 1 Fig. 4 . 1 Fig. 5 .
Mass-time drying curves of two specimens of the MB stone, F4:2 and environmental conditions (temperature and RH) during the test F5:Infrared measurements: (a) measurement of surface temperature during evaporative cooling experiments performed outdoors; (b) materials F5:2 and water-filled container wrapped in PE film and even higher values with the presence of water in the material 260 [Fig.6(b)].The dispersion of results within each test was not very 261 expressive, as seen by the standard deviation values presented 262 in actual RH values measured during the CDRP.As seen, the dispersion of the experimental values is generally very low.The tendency to have slightly negative values of DR for RH ¼ 100% in some cases suggests that the RH av is a slight underestimate of the equivalent RH.

288 where J (ML − 2 TFig. 7 .
−1 ) = the mass flow of water vapor, i.e., the drying 289 rate of the porous material; π (T) = the vapor permeability of the 290 layer that water vapor has to cross; dp=dx (MT −2 L −2 ) = the uni-291 directional vapor pressure gradient across that layer; and p 0 = the 292 saturated vapor pressure.293 Another relevant observation from Fig. 7 is that the CDRP dry-294 ing rate of the materials is not necessarily equal to the evaporation 295 rate from the free water surface tested under the same environmen-296 tal conditions.It can be significantly lower, as it happens for ex-297 ample with the CC limestone in the majority of the conditions, but it Drying 12 rate (DR) from materials during the CDRP and from free water surfaces as a function of the actual relative humidity (RH av ); the error F7:2bars correspond to one standard deviation above and one below the average higher, as seen in Fig.7(i).A CDPR drying rate higher 299 than the evaporation rate from a free water surface had already been 300 observed for different types of materials by several authors 301 (Hammecker 1993; Jeannette 1997; Tournier et al. 2000; Rousset-302 Tournier 2001; Tang and Etzion 2004; Diaz Gonçalves et al. 2012).

310
Tournier et al. (2000) attributed the high CDRP drying rate of 311 porous materials to their surface roughness.This broadly encom-312 passes the fact that in the pores curved menisci are formed (rather 313 than flat water surfaces), as well as the geometrical irregularity of 314 the material surface.However, the concept of (geometrical) surface 315 roughness is quite slippery when applied to porous materials: 316 straightforward extrapolation of what happens with other simpler 317 types of surfaces, such as metals or plastics, is not possible.318 Table 6 depicts the values of the relative area obtained with the 319 optical instrument at different scales, i.e., for different measuring 320 steps.As seen, the RA values vary with the measurement scale.321Thisvariation of RA with the measurement scale is graphically de-322 picted in Fig.9for the two materials (CS.B and CC) with the larger 323 and smaller RA, respectively.These curves show that RA increases 324 exponentially with scale.They also show that there is a vertical 325 asymptote at point zero of the X-axis.

Figs. 11 Fig. 8 . 1 Fig. 9 .
Figs. 11 and 12 present the surface temperatures measured during the evaporation experiments performed in the shadow and under the sun, respectively.The first obvious observation is that the wet materials depict surface temperatures well below those of the dry materials due to the evaporative cooling effect.
385 obtain more accurate emissivity data for the range of temperatures 386 of interest, as emissivity may also depend on temperature.Such 387 knowledge would allow developing a numerical model to support 388 the development of evaporative cooling systems.389 Conclusions and Perspectives 390 It was experimentally observed that the drying rate during the 391 CDRP from porous building materials, such as natural stone or 392 ceramic brick, can be very high.In some cases it may even over-393 come the evaporation rate from a free water surface subjected to 394 similar environmental conditions.This high drying rate is probably 395 due to the fractal character of the evaporating surface and to 396 menisci curvature.Both features enhance the effective surface of 397 evaporation in the material.

F10: 1 Fig. 10 .Fig. 11 .Fig. 12 .
Variation of the CDRP evaporation rate with the capillary porosity.The free water surface is considered a material with 100% porosity; point F10:2 (0,0) is attributed to a theoretical material with 0% porosity and, thus, perfectly flat Surface temperature (ºC) Surface temperatures measured in shadow high CDRP drying rate, porous building materials 399 reveal a high potential for evaporative cooling, which is in accor-400 dance with traditional uses in hot, dry climates, such as the wet-401 ting of ceramic tiles during summer in Mediterranean countries.402 This was confirmed experimentally, by means of evaporative 403 cooling experiments performed in the exterior, during a hot 404 summer day in Lisbon.The dry materials achieved high surface 405 temperature, especially in the sun where some materials reached 406 more than 50°C.However, during drying, their surface tempera-407 ture dropped on average as much as 10°C to 15°C.The surface 408 temperature of the wet materials achieved values similar (in the 409 sun) or even lower (in the shadow) than that of a free water 410 surface.411 This article is expected to contribute to a better assimilation by 412 civil engineering disciplines of the idea that the surface morphol-413 ogy of ordinary porous building materials, such as brick or natural 414 stone, has a fractal multiscale character that affects the way they 415 interact with the environment.Moreover, the obtained experimental 416 results confirm that these materials may be used in efficient evapo-417 rative cooling systems for low energy-consuming buildings.They 418 may, thus, help to meet the current needs for energy efficiency 419 while maintaining a simultaneous constructive or architectural 420 function.421 ISO.(2012)."Geometrical product specifications (GPS).Surface texture: Areal-Part 2: Terms, definitions and surface texture parameters."ISO 25178Surface temperatures measured under the sun

Table 2 .
Incoherent Materials Used as Reference in Evaporative Cooling Tests a According to EN 196-1 (CEN 2005).a Average value of the actual RH in the box during the CDRP.b Data from CEN (2001).c Data from ASTM (2007).F3:1 Fig. 3. RH inside the drying box during the experiments 177 Conditions 178 The evaporation experiments were undertaken in real outdoor 179 conditions during a hot summer day in Lisbon.One test was carried 180 out under direct sunlight and another in the shade.There were no 181 nearby buildings or heat sources in the location chosen for the test 182 (at LNEC 8 campus).Surface temperature was measured with an 183 infrared thermometer Raytek (model MX4PG).184 The tests were carried out on five rigid materials, namely 185 calcium silicate CS.L500, Maastricht M and CB limestones, lime 186 mortar A, and red ceramic brick T (Table

Table 5
, which represent a maximum of 7% in relation to the 263 average values.264 However, the repeatability of this method for determining 265 the emissivity was not good, as shown by Fig. 6(c).For this 266 reason, it was decided to adopt a single emissivity value for all 267 the materials (which in fact is the usual procedure).A value of 268 0.92 was chosen because it is close to the obtained experimental 269 values and also to those found in the literature for this type of 270 materials.271 Results and Discussion 272 CDRP Drying Rate and Surface Area 273 Fig. 7 depicts the variation of the CDRP drying rate (DR) of the 274 materials and water surfaces as a function of the RH av , which is an a Wooden plates.b Agglutinated.c Water.F6:1 Fig. 6.Comparison of the measured emissivity values: (a) between materials with or without PE film; (b) between dry and wet materials; (c) among F6:2 two identical tests on the same materials a Considering or not the distilled water.

Table 6 .
Relative Area of Tested Material Surfaces at Different