Portland versus alkali-activated cement wall panels containing mine tailing as aggregate: one-story house thermal performance simulation in a Brazilian and Portuguese hot and humid climate

Alkali-activated materials (AAM) compared to Portland cement (PC) may significantly reduce the carbon dioxide emissions, as well as the consumption of nonrenewable natural resources in civil engineering applications. Further environmental advantages are possible if natural aggregates used for mortars and concretes are replaced with residues and wastes from industrial or mining activities. This paper compares the performance of PC with AAM as binders in cementitious wall panels for external cladding in hot and humid climate. Three different cementitious matrices are proposed, consisting of either 100% Portland cement (PC), 100% alkali-activated metakaolin (MK) or 80/20 alkaliactivated Metakaolin/Blastfurnace slag (80/20 MK/BFS). Mortars were produced with the addition of tailing from iron-ore mining activities in the state of Minas Gerais, Brazil, at an aggregate to binder ratio of 1.0 for all matrices. The thermal property determined for the three mortars was Thermal Conductivity using a heat flow meter (HFM) apparatus according to ISO 8301 (1999); their apparent density was also measured. After that, onestory house building simulation was carried out using the Energy Plus Software. The main room annual operative temperature provided by different panels used as cladding was compared to the adaptive comfort range established on ASHRAE Standard 55/2013 for a Brazilian and Portuguese hot and humid climate. According to the Climate Zone Definitions of ANSI/ASHRAE Standard 169/2006, Belo Horizonte (Brazil) and Funchal (Portugal) were selected as a sample of 2A zone that presents a hot and humid climate. Partial results of this research were presented in this paper. Results show that building simulations can effectively contribute to validate the selection of materials in the production of sustainable wall panels that provide suitable thermal conditions to the users in hot and humid climate.


Introduction
The diversity of concrete applications in construction is directly related to its wide range of variations in their physical characteristics according to their composition. Several studies have analysed the physical and mechanical behaviour of alkali-activated materials (AAM). The results point out that these materials are an alternative that not only presents a mechanical performance equivalent to the Portland cement (PC) but also advantages such as greater chemical and thermal resistance. In addition, the growing interest in AAM is directly related to their low environmental impact when compared to PC-based materials. The main environmental advantage of AAM over PC is the significantly lower CO2 emission rate in the former, up to 80% in some cases [1].
AAM are considered inorganic polymers obtained by the reaction between an aluminosilicate and an alkaline solution (activator solution). They can be produced by the activation of commercial materials such as metakaolin, fly ash, blast furnace slag and active silica [1][2][3][4]. The use of waste and tailings from mining processes as substitutes for natural aggregates has also been investigated around the world. Employment of mining by-products in AAM arises due to the significant increase in their generation in recent decades [5][6][7]. In Brazil, most of the wastes that derive from the extraction and processing of ores are located in dams, which can cause risk such as infiltration of solutions containing heavy metals and cyanides in the soil beside the control of the pH of these solutions. The geological-geotechnical control of these structures is of great importance, in order to maintain stability and avoid possible disruptions [8][9].
Certainly, one of the main obstacles to the AAM utilization in the construction industry is now standardization, rather than technological restrictions. Nevertheless, some important information on the thermal properties of geopolymers is still unavailable [1]. The description of building materials thermal properties could assist the design decisions. It is the case of the Brazilian thermal performance of residential buildings regulation NBR 15220/2005 and also on ISO 10456/2007, both of which present different materials and their properties for designers [10,11]; AAM are not part of the list of materials available for designers. Yet, studies show that AAM may have lower specific heat and thermal conductivity values than PC, which may enhance the material thermal performance in some climate conditions, since the thermal diffusivity is directly proportional to the thermal conductivity [12][13][14].
The present study aims at measure the thermal conductivity and apparent density of three different matrices containing mine tailing as aggregate (two AAM and a reference PC matrix). Furthermore, it intends to evaluate their thermal performance as wall panels through building simulation case study for a one-story house located in hot and humid climate in Brazil and Portugal.

Materials
The AAM produced were obtained by the activation of metakaolin (MK) and blast furnace slag (BFS). The chemical composition and X-Ray diffraction pattern (XRD) of those aluminosilicates and aggregate (mine tailing) are presented in Table 1 and Figure 1, respectively. MK is composed mainly of SiO2 and Al2O3, and it is mostly amorphous; XRD peaks for quartz, illite and muscovite indicates impurities in the kaolin source, whereas the peaks for kaolinite show some incomplete dehydroxylation of kaolin during the calcination process.
BFS is rich in calcium (CaO = 32.74%), which will change the composition of the alkali-activated matrix; XRD peaks for gehlenite and akermanite are typical of slags from pig iron. The iron-ore tailing employed as aggregate results from the mining activities in the state of Minas Gerais, Brazil; it is composed of mainly quartz and hematite, with some XRDs peaks for kaolinite probably resulted from contamination. Figure 2 shows the particle size distribution of this aggregate, which has specific gravity of 2.95 and fineness modulus of 0.14.

Formulation design and specimens preparation
The mix design parameters used for the preparation of the AAM are shown in Table 2. A [SiO2]/[Al2O3] molar ratio of 3.0 was adopted for both AAM matrices. The alkaline solutions consisted of a mixture of sodium silicate (Na 2SiO3) and sodium hydroxide (NaOH). The solution to binder ratio and Na2SiO3 to NaOH ratio (wt.) varied in each mix, but the [H2O]/[Na2O] molar ratio was kept constant and equals to 13.6. All mortars were prepared with an aggregate to binder ratio of 1.0. The alkali-activated mortars were named as follows: PC for the reference PC mortar; 100 MK for the 100% MK AAM; 80/20 MK/BFS for 80% MK / 20% BFS AAM mortar ( Table 2). For AAM mortars the alkaline solutions were mixed together on the day of casting to prepare the liquid component of the mixture. When BFS was employed, the two binders (MK+BFS) were firstly dry-mixed. The alkaline activator was then added to the dry mix and the matrix was mixed for 3 minutes. Finally, the aggregate was gradually added while mixing until the mortar presented a homogeneous appearance. A similar procedure was employed in the manufacturing of the reference mortar (PC). For this last one, a commercial superplasticizer, previously mixed in water, was employed at a dosage of 2% by mass of cement, so that the overall water to cement ratio was reduced (0.38).
The mortars were cast in 300x300x20 mm moulds for the determination of thermal conductivity and 100x50 mm cylindrical moulds for compressive and determination of the apparent density. The specimens were exposed to ambient curing conditions until they reached the testing age (28 days).

Specific heat and Thermal conductivity
The ASTM 351-b details a specific calorimeter to define the specific heat of the materials. However, for a preliminary study, it was adopted the specific heat found in the literature proper to AAM and PC concrete (Table  3).

Solar absorptance
Solar absorptance data was obtained by calculation from the sample measurements made by ALTA II reflectance spectrometer, from Vernier Software & Technology. This equipment performs measurements corresponding to radiation in eleven different wavelengths between 470 and 940 nm, seven in the visible region and four in the near-infrared region. The analyses were performed as described by Sangoi et al. [16].

Apparent dry density
Apparent dry density was measured by dividing the mass of oved dried samples by their volume (cylindrical samples).

Simulation
The thermal comfort conditions of a single-family house (24m²) were tested for a hot and humid climate defined as 2A zone by ANSI/ASHRAE Standard 169/2006 [17]. Belo Horizonte and Funchal were chosen as representative cities of Brazil and Portugal, respectively. A 3D model (Figure 3a) was created using SketchUp software associated to Euclid extension that create and modify the geometry inputs for building energy models. The building envelope was designed according to the 2A climate zone requirements established in ASHRAE Standard 90.1/2013 [18]. Figure 3b shows that the roof is composed of thermo-acoustic metal tile (0.43mm) filled in with polyurethane (30mm), aluminium foil fiberglass as under-coverage (63mm), an air space greater than 5cm, a concrete slab (10cm), another air space greater than 5cm and a glass-wool ceiling (35mm) which achieve 0.217 W/(m²K) of thermal transmittance (U) and 21 hours of thermal lag (φ). The building walls may be composed of three options (PC, 100MK and 80/20 MK/BFS panels) produced with 2cm of thickness for inner and outer panels filled up with 8cm of EPS (expanded polystyrene) ( Table 4). The building floor (Figure 3c, U=0.598W(m².K)) is composed of concrete slab (5cm), glass-wool (63mm) and laminate wood flooring (7mm). The glass material properties adopted for the windows present U=3.629W(m². K), solar factor SF=0.517, 6mm of thickness. The main room window present 1

Results
The results of absorptance, apparent density and thermal conductivity measured for the 3 different wall panels are shown in Table 5. The 100 MK mortar had an absorptance of 0.694, apparent density of 1.77 g/cm 3 and 0.186 W/(m.K) thermal conductivity. In the 80/20 MK/BFS sample the absorptance was 0.910, the density 1.83 g/cm 3 and the thermal conductivity 0.188 W/(m.K). For the reference panel (PC), the absorptance, apparent density and thermal conductivity were, 0.816, 2.20 g/cm³ and 0.100 x W/(m.K), respectively. These results agree with those found in the literature, in which thermal conductivity is lower in AAM then in PC [12][13][14]. Through computer simulation it could be noticed that the 100 MK wall panel had the best performance for Belo Horizonte, presenting 86% thermal comfort hours, against 82% presented by the PC panel and 72% by the 80/20% MK/BFS. However, for Funchal city, the PC wall panel provided 77% thermal comfort hours, against 74% for 100 MK and 72% for 80/20 MK/BFS wall panels (Figure 4). that some changes could be performed on the building envelope in order to reduce this number of hours' discomfort. Perhaps the isolated sealing, the thermal lag greater than 5h for all options of wall panels, associated to the cement compositions that presents a thermal inertia potential (thermal capacity greater than 56kJ/m².K) combined with high solar absorptance (greater than 0.7) hindered heat loss when external temperatures were high. The employment of a panel lighter in colour, i.e. with a lower value of solar absorptance, could contribute to reduce the heat hours' discomfort. On the other hand, this special feature seems to contribute to maintain the indoor temperatures when the outdoor temperatures were low. Therefore, the cold discomfort did not take place. Similar comparison was performed for Funchal. Figure 6 indicates different performance of the wall panels through the year in comparison between Belo Horizonte. It is possible to note that the maximum temperatures significantly exceed the upper comfort limit, especially in summer and in the beginning of fall. The minimum temperatures were far below from the lower comfort limit, during the fall and winter. The discomfort hours at low temperatures (cold climate) were greater than at high temperature (heat), except for the 80/20MK panel. However, the maximum operative temperatures were more extreme for hot conditions, although the maximum DBT did not surpass the 28.8ºC while the operative surpassed the 32.0ºC, in some cases. The thermal inertia potential of the cement wall panel (herewith an insulate sealing) associated with a high thermal absorptance makes the building system accumulate a significant amount of heat during the hot seasons. For this city, some changes could also be made in order to enhance the comfort hours during the summer. In this case, a lighter colour on the external panel could contribute to reduce the heat hours' discomfort but, on the other hand, will increase the cold hour's discomfort. To reduce the heat discomfort hours without increase significantly the cold discomfort through the year, the thermal capacity of the wall panels should be increased.

Conclusions
The thermal conductivity, apparent density and solar absorptance of three different mortars, consisting of either 100% Portland cement (PC), 100% alkaliactivated metakaolin (MK) or 80/20 alkali-activated Metakaolin/Blastfurnace slag (80/20 MK/BFS) were performed. The Energy Plus building simulation of a one-story house unit enabled to evaluate the thermal behaviour of those materials as cladding for a hot and humid climate. Belo Horizonte and Funchal were chosen as representative cities of Brazil and Portugal, respectively, considering the 2A zone addressed on ANSI/ASHRAE Standard 169/2006. The results showed that those materials, in compliance with ASHRAE Standard 90.1/2013, could be adopted as a wall panel although a schedule of building use and occupation was not considered. Relevant comfort hours throughout the year could be noticed in comparison between the adaptive comfort ranges established by ASHRAE Standard 55/2013 (greater than 72% for both cities). Further studies could test different arrangements of the wall panels composition and colour made up of those materials. However, this research can contribute to foster new studies and to highlight thermal data of an alternative and more sustainable building material option in comparison to Portland cement panels.