The raw earthen material is a beige clayey soil extracted from the locality of Pabré (012˚31.397'N, 001˚34.373'W, and altitude of 297 m). Its lumps were crushed to release the elementary particles. It was dried and sieved to remove aggregates larger than 5 mm, in order to facilitate homogeneous mixing and interaction with the binder. The binder and stabilizer is calcium carbide residue (CCR), an industrial by-product. The CCR was collected from Burkina Industrial Gas (BIG) located in Kossodo (012˚25.935'N, 001˚29.374'W, alt. 301 m). It was milled and sieved to collect the passing on 125 μm. This CCR was previously found to contain more than 40% of calcium hydroxide [18]. Industrial cement, CEM IIA 42.5, commercially produced by CIMBURKINA was used as control stabilizer of CEB, as reference [6] [8]. Other reference properties of the different conventional materials have been taken from the literature.

The mineral composition was analyzed by the X-ray diffraction technique using the Brucker D8-Advance Eco 1.5 kW diffractometer equipped with a copper anode (Cu Kα λ = 1.54060 Å, 40 kV, 25 mA) and the Lynxeye xe detector in coupled 2θ/θ mode. Qualitative and semi-quantitative analysis of the spectra was performed using Diffrac.Eva V4.11 and Topas V5 software of Brucker based on the rietveld refinement method.

Different formulas of CEB were elaborated by adding various mass percentages of CCR (0%, 5%, 10%, 15%, 20% and 25%) with respect to the dry mass of earthen material. CEB stabilized with 8% cement, which is commonly used for stabilization of CEB in the area of Ouagadougou, were produced as reference. Dry materials (earth + CCR/cement) were mixed until achieving homogeneous mixtures. Dry mixtures were mixed with appropriate amount of water, equivalent to the optimum water content (OWC of 16% - 20% for 0% - 25% CCR), required for achieving the maximum density for each formula. Specimens of CEB were produced by manual compression of the wet mixtures in 295 × 140 × 95 mm³ mold and cured in polymeric bags to minimize the carbonation and loss of moisture necessary for the reaction to take place. The curing took place at ambient temperature in the laboratory (30˚C ± 5˚C) for 45 days for the specimens stabilized with the CCR and 28 days for those stabilized with cement, the period required to achieve more than 95% of the reaction [20].

Cured CEB were dried at 40˚C ± 2˚C until reaching constant mass. Apparent density of the dry CEB was determined through hydrostatic weighing and Equation (1) according to NF P18-459 [21]. Where, ρ_app is the apparent density of the dry sample (kg·m⁻³); M_dry is the dry mass of the dry sample (g); ρ w a t is the density of water (kg·m⁻³); M_sat.wat is the mass of the saturated sample weighed in water (kg); M_sat.air is the mass of the saturated sample weighed in air (kg).

ρ a p p = M d r y × ρ w a t M s a t . a i r − M s a t . w a t (1)

Thermal effusivity, specific heat capacity, thermal conductivity, thermal diffusivity and depth of penetration of heat flux are the main thermal properties measured on the CEB. They are determined using the DEsProTherm (Dispositif d’Estimation des Propriétés Thermiques) system presented in Figure 1. The measurement of thermal effusivity (E, J·m⁻²·K⁻¹·s^−1/2) was performed on single size samples (6 × 4 × 3 cm³). The thickness of 3 cm does not allow the heat to cross through the sample. Moreover, the measurement of the heat capacity, equivalent of the specific heat (Cp, J·kg⁻¹·K⁻¹), was carried out on samples of size 6 × 4 × 1 cm³ allowing the heat flux to cross through the sample [22]. Once these two parameters were obtained, the thermal conductivity (λ, W·m^˗1·K⁻¹), thermal diffusivity (a, m²·s⁻¹) and depth of penetration of the heat flux (δp, m) for the period of time T (s) of 24 hours were determined using Equations (2), (3) and (4), respectively.

λ = E ² ρ a p p × C p (2)

a = λ ρ a p p × C p (3)

δ p = a × T π (4)

The optical properties (emissivity, solar absorptance, visible absorptance) were estimated based on the mineral composition of the earthen material and CCR. Equation (5) presents the proportionality used to estimate the optical property (Op) of CEB with respect to the percent content (Ci) and optical property of the reference mineral (Opi) taken from the literature and libraries from Open studio and Energyplus [23] - [28]. The values obtained from this approximation were comparable with the values reported in the literature for the earthen materials [26] [28]. In addition, solar reflectance and visible transmittance values are assumed to be substantially the same [26] [29]. This approach was adopted given that the change in formula of CEB (different content of CCR) should logically have an impact on their optical properties. Thus, it would not be appropriate to consider the same value for all formulas of CEB. Other studies previously proposed alternative methods for determining the optical properties of material, but they require some specific equipment for measurement [30].

O p = ∑ ​ ( O p i × C i ) (5)

The building was modelled using SketchUP 2017 software. Once the model is elaborated, the properties of different construction elements, measured in the present study and collected from the literature, were defined. EnergyPlus software was used for thermal and energy simulation via the plug-in of Open Studio 2.8.0. XEsoView software was used to visualize the results of the simulation. The Givoni diagram was used for evaluation of the number of hours of thermal discomfort. This method was previously used for evaluation of the thermal comfort in the Sahelian and tropical contexts [23] [31] [32]. The Givoni diagram can be used to assess thermal comfort of occupants of a building by considering the data from the climate of the area, the activities of occupants inside the building and the expected indoor temperature and humidity [33] [34] [35].

The model building is a single-story building with four rooms, floor area of 185 m² and main facade toward the South. Figure 2 and Figure 3 respectively present the plan view and 3D view of the model building. The building has been modeled so that during simulations, only the characteristics of the masonry elements change from one variant of the model to another. The characteristics of the elements used in different variants of the model are summarized in Table 1. The dimensions and positions of the bays are presented on the plane view (Figure 1). The windows are made with single glazing and outer doors are made of steel. The windows have a height of 1.2 m, placed at 1 m above the ground, except those in bathrooms that have a height of 0.7 m, placed at 1.5 m. All the doors have height of 2.2 m. Depending on the room and its use, the widths of windows and doors range between 0.6 m and 2 m such as shown in Figure 2. The model has a 1.5 mm thick corrugated steel sheet for the roofing and a 5 mm thick plywood ceiling placed 0.5 m below the roof. Table 2 presents the assumptions for the different schedules for the occupation and internal loads in the building. Note that these configurations are very common in the Sahelian region such as Burkina Faso. The parameters and assumptions made for modelling the building are summarized in Table 3.

Note that for the present simulations, the supply of cooling by devices such as air conditioners, humidifiers, is initially neglected for evaluating the thermal comfort induced by each masonry type. Only interactions between external conditions and materials are considered. The weather file used is that for Ouagadougou. The infiltration airflow rate is set at 0.5 m³·s⁻¹ and mean velocity of the air is set at 0.3 m·s^˗1. The thermal resistance of the air gap (between the two steel sheets that constitute the external doors) is set at 0.16 W·m⁻¹·K⁻¹. The thermo-physical and optical properties of the conventional materials are given in Table 3.

The thickness of the glazing material, e is 0.3 cm, its thermal conductivity, λ is 0.93 W·m⁻¹·K⁻¹, its solar transmittance at normal incidence, g is 0.86, its visible transmittance at normal incidence, τl is 0.9. The front and back side infrared hemispherical emissivity, ε of the glazing material is 0.92, its front side solar reflectance at normal incidence FSSR index is 0.07 and its front side visible reflectance at normal incidence FSVR index is 0.08. The back side solar reflectance at normal incidence BSSR and back side visible reflectance at normal incidence BSVR indexes of this material are null. For the windows and frame of the door, the thickness is taken as 1.35 cm, its solar absorptivity and visible absorptivity indexes are set at 0.9. The front and back side infrared hemispherical emissivity value of the frame is 0.7. The width of the frame is 5 cm and its thermal transmittance Uf value is 3.3 W·m⁻²·K⁻¹.

The thermal comfort assessment was based on the analysis computer program EnergyPlus 9.1.0 [29] and ASHRAE Standard 55 [14]. EnergyPlus evaluates the outdoor air temperature based on the input weather file, the indoor relative humidity and temperature depending on the activity of the occupants and internal thermal loads. Then the number of hours of discomfort are assessed based on the diagram proposed by Givoni [37]. This diagram presents boundaries of comfort zones, which are based on air temperature and air relative humidity (Figure 4). The black boundary limits the comfort zone for a null value of the velocity of the air. Every point that is inside that boundary represents an hour at which 90% of the occupants feel comfortable, according to the Predicted Mean Vote (PMV) [14]. The blue boundary limits the comfort zone for a velocity of air assumed at 0.3 m·s⁻¹. Depending on the velocity of air, this zone may become thinner or larger.

From the diagram, out of the comfort zones, it can be inferred four zones of discomfort. Over the upper limit of the comfort zone is the humid discomfort zone. A point localized in that zone is considered non-comfortable hour because of high rate of humidity in the air. At the right side of the comfort zone is the warm (hot) discomfort zone. A point localized in that zone is considered non-comfortable because the temperature of the air is too high. Those situations, can cause health problems and harm the occupants body [38]. At the opposite,

below the lower and left limit of the boundary of the comfort zone, are respectively located the dry and cold zone of discomfort. Those areas are also non-comfortable because of low rate of humidity in the air and low temperature of the air. The objective is to get most of the points that represents the hours of the year in the comfort zone considering a mean velocity of the air of 0.3 m·s⁻¹. The other points should preferably be localized next to the comfort zone. The relative proximity of hours of discomfort to the comfort zone is assessed through the Hygrothermal Index (HI). This index, considering the combined effect of temperature and humidity, indicates whether a point is closer to the comfort zone or not. The lower the HI, the closer the points are to the comfort zone and the lesser aggressive is the discomfort.

The evaluation of the cost was carried out on the energy that would be consumed when Air Conditioning System (ACS) is used to keep the indoor temperature not higher than 28˚C in each variant of the model building. This calculation is based on the output amount of energy needed to cool down the indoor environment assessed from EnergyPlus. The cost was estimated based on the value of the energy needed for cooling at efficiency of the ACS assumed at 83% and price of 96 CFA francs (0.17 USD) for 1 kWh of electricity [39].

Table 4 presents the mineral composition of the earthen material and CCR. The earthen materials mainly contain quartz (63%) and kaolinite (30%) while the CCR mainly contains portlandite (43%), calcite and other carbonate minerals (50%). The optical properties such as the emissivity, solar absorption and visible absorption of CEB, approximated through the proportionality method, are presented in Table 5. It shows that the addition of CCR improves the optical properties of the CEB. The solar transmittance, thermal absorptivity (emissivity of the material) and visible absorptivity of the CEB decreased respectively from 0.82, 0.82 and 0.82 for the non-stabilized (0%) CEB to 0.80, 0.64 and 0.64 for 25% CCR stabilized CEB (Table 5).

Table 5 furthermore shows that the CCR-stabilized CEB have lower apparent density than the non-stabilized CEB and cementitious materials. The apparent density varies between 1820 kg·m⁻³ and 1600 kg·m⁻³ for 5% - 25% CCR-stabilized CEB compared to more than 2100 kg·m⁻³ for non-stabilized CEB (0% CCR). This phenomenon can be partly explained by the fact that the specific density of CCR particles (2.49 g·cm⁻³) is less than that of the earthen particles (2.66 g·cm⁻³). Moreover, the decrease of the apparent density of CEB with CCR stabilization can be related to the increase of the OWC (16% - 20%) for production of CEB stabilized with 0% - 25% CCR. This resulted in increase of the total porosity of CEB from 30% to 40% for 5% - 25% CCR compared to 20% for 0% CCR stabilized CEB. The apparent density and thickness of the envelope materials are very relevant parameters for the thermal comfort of the building.

Additionally, the stabilization of CEB with CCR resulted in the decrease of the thermal conductivity from 1 W·m⁻¹·K⁻¹ for the 0% CCR-CEB to 0.5 W·m⁻¹·K⁻¹ for 25% CCR-CEB. These values are much lower than those for materials containing cement (2.5 W·m⁻¹·K⁻¹ for concrete). Furthermore, the heat capacity of CCR (0% - 25%) stabilized CEB ranged in 800 - 900 J·kg⁻¹·K⁻¹ compared to 900 J·kg⁻¹·K⁻¹ for concrete (Table 5). It is important to remind that the thermal properties “e, λ, ρ, and Cp” have been measured in the laboratory and the optical properties “αth, αs and αv” have been calculated through the method explain in paragraph II.4

Figure 5 shows the evolution of the percentage of number of hours of thermal discomfort, warm, humid, other types, for the reference model (M1) and other variants. The total number of hours of thermal discomfort is about 64% for all models without air conditioning system. The predominant type of this discomfort is humid (wet), then warm and other types (dry and cold). In the area of Ouagadougou, it is the hot temperature of the air that worries more when referring to the thermal comfort of the buildings. The interest of representing the planning of occupation (Table 2) that reflects the habits of most people in the study area is to reflect the thermal discomfort that approaches the realities of the area. Compared to M1, out of all the 10 models, the model M9 (20% CCR-CEB masonry) has the lowest number of hours of warm discomfort.

While Figure 5 shows the cumulative number of discomfort hours, Figure 6 shows the relative number of hours of discomfort of different models with respect to the model M1. Negative values are favorable in a sense that they indicate that the percentage of hours of discomfort decreased. The thermal comfort can only be evaluated by the appreciation of the discomfort because it is a subjective sensation. The lower the relative discomfort induced from each model, the better the masonry behaves in term of the thermal performance comparative to the model M1. The model M9 (20% CCR CEB) has the lowest number of hours of warm discomfort. It records 21.2% lesser number of hours of warm discomfort compared to M1 (cement blocks), equivalent to a total of about 400 hours of warm discomfort in the year (Figure 5).

The number of hours of humid discomfort is relatively higher with CEB comparative with cement blocks (up to 14.6% higher for the model M9 vs model M1). The increasing humidity in the indoor environment can be explained by the relative high hygroscopic inertia of the masonry blocks. This was previously reported that the moisture content is closely related to the hygroscopic inertia of the building material in the studies which tried to show the relation between the hygroscopic inertia of the envelop and moisture content in a room [40] [41] [42].

To prevent or minimize the humid discomfort, natural and cost effective means exist such as having humidity-controlling salt deposited in the room. However, the warm discomfort can only be dealt with by bringing a vector of cooling such as air conditioner. In addition, according to the analysis of the Givoni diagram, models M1 and M9 have the same hygrothermal index value of 1.9, which reflects a relative proximity of discomfort hours to the comfort zone compared to other model with higher index, i.e. 2.0 for concrete.

Figure 7 shows the differences between the temperatures of the indoor environment for the models M1 and M9 from the 1^st to the 3^rd day of the month of April (the first 72 hours of the hottest month in Ouagadougou). The results show that the average indoor temperature for the model M9 is lower than that of the model M1 in most of the time (51% of the time calculated on the base of the data plotted on Figure 7). From this difference (“Delta M9-M1”), it can be inferred that the model M9 gets hotter slower than model M1 during the heating phases, but it reaches higher temperature. The same analysis can be made on the cooling phases, but this time, the model M9 is more suitable. These imply that the masonry in model M9 (CEB with 20% CCR) absorbs and releases the heat slower than that in model M1 (hollow cement blocks). It is important to note that the cooling phases are longer than the heating phases. Furthermore, from the heat capacities given in Table 3 and Table 5, it can be inferred that CEB stabilized with CCR have higher heat capacities than cement blocks (938 J·kg^˗1·K⁻¹ for 20% CCR-CEB and 600 J·kg⁻¹·K⁻¹ for hollow cement blocks). This implies that CEB can stock 56% more heat than cement blocks.

Figure 8 further shows that the depth of the penetration of heat flux of the model M9 (9.2 cm) is lower than that of the model M1 (14 cm). The depth of penetration indicates, on a time period of 24 hours, how deep the heat flux would penetrate into the wall. This implies that the flux of heat takes more time to get through the CEB than the cement blocks. This further explains the lag observed on the peak temperature due to the slow process of heating/cooling of the CEB (Figure 7). It additionally explains the fact that even if the 20% CCR-CEB has a better thermal behavior than the cement blocks, the model M9 reach higher indoor temperatures when the outdoor temperature is at its highest level. With lower external temperatures, this behavior of the CEB becomes favorable for the building, the CEB stores more coolness and deliver it on a longer time.

On these bases, different formulas of CEB stabilized with CCR are more suitable for maintaining the temperature set in the building much longer than the conventional materials such as cement blocks and concrete, and thus would economize the energy consumption if any cooling systems were used. Moreover, the results suggest that stabilization of CEB with CCR not only improved their mechanical performances [18] but also their thermal behavior compared to non-stabilized CEB. Indeed, the previous study [18] reported the positive effects of stabilizing a clayey soil with the same CCR on the mechanical performances of CEB. The results from the present study show the improvement of the thermo-optical properties of CEB.

Figure 9 shows the cost of the energy consumed in a year if the air conditioning system (ACS) were used to maintain the indoor temperature not higher than 28˚C during the presence of the occupants. It shows that the model M3 (concrete wall model) represent the highest cost of energy consumed through the year (6830 USD). The reference model M1 comes second in term of cost of energy consumption (6156 USD) while model M9 (CEB with 20% CCR) induces the lowest cost (5566 USD) on energy consumption. This further confirms that the use of CEB stabilized with CCR improves thermal behavior of the buildings, and the higher the CCR content, the better the CEB behaves.

Figure 10 shows the economy made with each model with respect to the model M1. The economy of up to 9.6% that equals to 535 USD (310,000 CFA francs) can be achieved if the building in model M9 (20% CCR˗CEB) is air conditioned instead of model M1 (hollow cement blocks). This economy subsequently shows the advantage of using CEB stabilized with CCR instead of the conventional hollow cement blocks. Furthermore, on the environmental point of view, CEB are much more ecological than cement blocks and concrete. By using a by-product as stabilizer, CCR-stabilized CEB would basically become more ecological than those stabilized with cement.