From plaster blocks standardized by NBR 16494 (57), eight test elements were produced, covering the different thicknesses for the configurations available between hollow and compact in the Brazilian market. The main characteristics are presented in Table 1.

Both chemical and physical characteristics used in manufacturing plaster blocks are shown in Table 2.

Every plaster block was manufactured in an industrial manner, including cubical metal molds (solid and hollow) with smooth surface and side fits. Subsequently, plaster blocks were dismounted after 60 min and taken to an oven at 40 ºC for 24 hours, according to NBR 16494 (57). It is worth mentioning that test elements are from the plaster block cuts provided by the industry. Therefore, their lateral dimensions were changed to a 42 cm square on the side, inserting the identification acronym and the Type K thermocouple temperature sensors in the central part on both sides of the test elements (Figure 2).

Figure 3 presents the mechanical properties of the plaster specimens. The flexural strength of the blocks meets the requirements of NBR 16494 (57). The compressive strength was determined by ABNT 12129 (60), using cubic specimens of 50 cm x 50 cm x 50 cm.

In order to carry out the thermal chamber behavior experiment, a thermal chamber was developed and built to provide the test element with the positioning between two environments with a temperature gradient between them (25, 28, 31). The test apparatus was built of wood, with insulation on the heated side and opening on the cold side (controlled room temperature). In addition, temperature sensors were inserted on both sides with a display on the control panel. A dimmer switch was also inserted to regulate the heating provided by an infrared lamp present in the central part of the hot side and the location for installing the digital thermometer with integrated data logger (Figure 4). Although there is no official standard, a proposed thermal chamber was developed according to Standard ASTM C1363 (61) provisions and studies by Kheradmand et al. (45), Pedreño-Rojas et al. (46), and Ferrari and Zanotto (62). The thermal chamber was used for temperature monitoring through thermal sensors. Additionally, elements’ surface temperature distribution was evaluated through infrared thermography.

The procedure of the experiment consists of controlling the temperature in the environment: 26 ± 1 °C, from an air conditioner. Then, the test elements are placed with the “outer face” turned to the inside the chamber. Strips of expanded polystyrene are placed on the edges the sensors are connected, the data logger is programmed, and heating is started. The heating lasts 360 minutes. Thermograms are recorded every 60 minutes on the cold side of the test elements and the temperatures are recorded every minute. After the warm-up period, the test elements cool for 120 minutes, with the same rate of temperature recording by the sensors. However, thermograms are made every 30 minutes and in various positions (cold side, hot side, lateral, and perspective) to observe the heat transition in the test element. The follow-up used the FLIR E-60 equipment, with main characteristics are shown in Table 3.

The emissivity values of the analyzed surfaces were found using the black layer method. Interaction was performed between the emissivity value of the tape - known - and that of the neighboring surface (material, whose emissivity is unknown) until the temperatures coincide. Values between 0.93 and 0.95 were found for the plaster. The distance to the object was always between 1.0 and 1.5 m. The reflected temperature was the same as that of the environment, with no interference in the results (no interference from sunlight, variation in lighting, or considerable temperature in the controlled environment).

The NBR 15220 standard (20) recommends the performance of tests to determine the thermal resistance of elements. One of the methods mentioned is the flow meter; the test is based on ISO 8301. For this research, the Netzsch heat flow meter - HFM 436/6 was used, with main characteristics presented in Table 4. Tests were performed under the steady-state heat transfer, using the absolute technique, detailed by Zhao et al. (64).

To carry out this test, the wells of the hollow blocks were closed with plaster paste, with the aid of a glass plate to prevent heat loss from the sides. It is noteworthy that the size of the test elements (42 cm x 42 cm) is compatible and still have a margin in relation to the detectable area of the heat flow transducer (25.4 cm x 25.4 cm) (Table 4). In addition to the tests on plaster blocks - all belonging to the medium density range as shown in Table 1 -, 3 plaster plates had the exact size of the equipment’s specimen (60 cm x 60 cm x 5 cm). The plates are D1 - 901.77 kg/m³, D2 - 1011.49 kg/m³ and D3 - 1165.93 kg/m³: two distant points but within the average range (between 800 and 1100 kg/m³) and one point within the high-density range (greater than 1100 kg/m³), according to NBR 16494 (57). The plasterboards were tested to have a greater range of results, considering that the plates were made specifically for the size of the equipment (60 cm x 60 cm). In addition, the plasterboards are made of a homogeneous material, which makes it possible to better infer thermal conductivity, based on the relationship between thermal resistance and the thickness of the test specimen.

The test was carried out according to the recommendations of the NBR 15520 standard (20) and the manufacturer’s recommendations (65). For the test with the 8 plaster block test elements, two main temperature values were used - Mean T of 24 and 40 °C - referring to a temperature close to the environment in part of Brazil and a higher one representing a peak during summer or artificial warming situations. On the other hand, for the plasterboard, in addition to these main temperatures, other two higher ones (50 and 60 °C) were added to expand the analysis of plaster, considering thermal conductivity (λ) under density conditions of apparent mass present in this work.

To obtain theoretical thermal properties for test elements, calculations of several parameters were performed according to prescriptions and tabulated input data values present in NBR 15220 (20). Among the input data values collected, the following stand out for the plaster: specific heat (c) of 0·84 kJ/kg·K, thermal conductivity (λ) of 0.35 W/m·K, and apparent mass density (ρ) of 875.0 kg/m³. The calculations were performed with the aid of spreadsheets programmed in the PTC Mathcad and MS Excel software. The calculated parameters were thermal resistance (R) (Equation [1] and Equation [2]), thermal transmittance (U) (Equation [3]), thermal capacity (C_T) (Equation [4]), and thermal delay (φ) (Equation [5]).

Here, e is the layer thickness (m); λ is the thermal conductivity (W/m·K); R_T is the total thermal resistance m²·K/W; R_j is the thermal resistance of each component layer (m²·K/W); R_SI is the internal surface thermal resistance (m²·K/W), and R_SE is the external surface thermal resistance (m²·K/W).

Here, λ_j is the thermal conductivity of each component layer (W/m·K); R_j is the thermal resistance of each component layer (m²·K/W); c_j is the material specific heat of each component layer (kJ/kg·K); ρ_j is the apparent mass density of each component layer (kg/m³), and e_j is the layer thickness of each component layer (m).

Here, R_t is the surface to surface thermal resistance (m²·K/W); λ is the thermal conductivity of the material (W/m·K); c is the material specific heat (kJ/kg·K); ρ is the apparent mass density of the material (kg/m³), and R_EXT is the thermal resistance of the component’s outer layer (m²·K/W).

Flow meter measured the thermal resistance and based on the thickness of the specimen, calculates the thermal conductivity (in case of homogeneous material). The values found for thermal conductivity corroborate with the value suggested in NBR 15220 (20), which is 0.35 W/m·K for a temperature of 27 °C (20). For densities D1, D2, and D3, the values for temperature 24 °C were 0.358 W/m·K, 0.354 W/m·K, and 0.364 W/m·K, respectively. The value suggested in the standard, therefore, can be used if testing is not available, according to the presented results. The values of R and λ for each of the main temperatures are in Table 5.

It should be noted that the variation in the conductivity value with temperature occurs in a more sensitive way when this difference is high in most materials (66).

For the nominal main temperature of 24 °C (SP#1 in Table 5), there is slight variation in conductivity in relation to the density of the material. However, for higher temperatures, an increase in thermal conductivity occurred with a decrease in the density of apparent mass. This situation differs from the results pointed out by Souza (27) who concluded in his study that the increase in the porosity of plaster specimens with the addition of sodium bicarbonate caused the decrease in thermal conductivity. The author also states that heat transfer through pores are slow processes and that the stagnant air, usually present inside the pores, is a bad heat conductor (λ = 0.02 W/m·K); when isolated, they make gas convection difficult (27).

The temperature affects the thermal conductivity of ceramic materials, the increase in temperature (in this case up to 60 °C) can explain the increase in thermal conductivity. On the other hand, porosity also has a direct influence, the increase in volume and/or number of pores reduces the thermal conductivity of ceramic materials (67, 68). In this case, based on the results presented, the higher temperatures of the test had more influence than the porosity on the thermal conductivity; however, a detailed study about the phenomena involved could better define this behavior.

Regarding the test elements from plaster blocks, Table 6 shows results and details for the apparent mass density, with specific value for each tested element. Results confirmed the trend of less thick blocks (GS 50 +, GS 70 + and GS 70 -) to have the lowest R values. Among the blocks of 100 mm thick, those massive (GH 100 + and GS 100 +) presented lower values than the hollow ones (GS 100-) and even lower values than both 76 mm hollow blocks (GS 76 - and GS 76 =). It highlights the significant contribution of the air layer (alveoli) inside these blocks to the increase of thermal resistance, even with the reduction of the total thickness. Similar results were reported when thermal resistance for hollowed-block walls was evaluated (44, 69). The values of the standard 100 mm compact, water repellent compact, and standard hollow blocks ranged between 0.25 and 0.30 m²·K/W, while both 76 mm blocks performed with R around 0.29 m²·K/W.

The values of heat flow for the plaster block test elements are also graphically presented in ascending order for the two main temperatures of 24 °C and 40 °C (Figure 5). Thermal resistance is not significantly modified by temperature (up to 40 °C). However, Mansour et al. (70) points out that higher temperatures (above 100 °C) may cause the plaster work as a firebreak.

Table 7 presents the results of the thermal performance parameters calculated as indicated in NBR 15220 (20) for the 8 plaster block test elements.

Analyzing the thermal resistance value, calculated from the presence of the alveoli in the plaster blocks, it is observed that this causes an increase in the thermal resistance value when compared to the compact ones (solid). An increase of 14% is observed between the 70 mm blocks while it’s only 3% for the 100 mm blocks. This difference in magnitude in the increase between the 70 and 100 blocks is due to the greater influence of the alveoli size in relation to their total thickness in the less thick block - since the alveoli have the same size. However, this behavior does not happen between elements with 76 mm plaster blocks, where there is a decrease in the resistance for the block with the largest alveolus (GS 76 =) - around 3% (Figure 6).

The “removal” of solid material from the blocks proved to be an alternative to increase the total resistance, as seen in the results of the flow meter (Table 6 and Figure 5). Evidence shows less thick hollow blocks with results superior to other thicker blocks. As mentioned by Zhang et al. (71), hollow blocks improve thermal insulation properties in walls (mainly thermal resistance) and reduce energy consumption. This is caused by thermal resistance of air contained in the block. However, the amount of reduction in plaster thickness presented in the “popular block”, material for the “GS 76 =” element, resulted in a decrease in thermal resistance since the calculation considers a fixed value for the resistance of the air layer: if it is increased, the only practical effect is to reduce the thickness of the solid material and, consequently, decrease its contribution to the element thermal resistance.

Thermal transmittance (U) and thermal capacity (C_T) are two criteria analyzed by the building performance standard NBR 15575 (21) for external vertical sealing systems. This work points out that usually compartmentalization of environments, as to thermal aspects, is needed. Therefore, it is important to understand the properties of the blocks as one of the components in sealing systems, both external and internal (the latter more common for plaster blocks). Figure 7 presents a joint graph of these parameters.

For the bioclimatic zone 8 (20) minimum values of C_T are not required, while in the other zones (1 to 7) at least 130 kJ/m².K is imperative for minimum thermal performance, according to NBR 15575 (21). In addition, for U appreciation it is necessary to know the absorption to solar radiation (α), which is related to the last outer layer of the External Vertical Sealing System. Therefore, for the plaster block test elements (without coating), this analysis, according to the normative requirements, is not relevant. However, it is interesting to note that the test elements in solid plaster blocks (GS 100 + and GH 100 +) presented the best combination of results: low U values - around 2.2 W/m².K - and C_T - about 82 kJ/m²·K. All test elements with thicknesses from 76 mm show U values within the requirements of any of the bioclimatic zones for walls (U≤ 2.5), even without coatings, which significantly improve these values (blue line in Figure 7). The minimum C_T value (130 kJ/m²·K) for zones 1 to 7 corresponds to the orange line (Figure 7).

Although the thermal transmittance (U) values are higher than those reported by Bianco et al. (72) (0.56-0.8 W/m²·K) and Asdrubali et al. (73) (0.23-0.33 W/m²·K) testing conditions and material settings were different. In previous studies, tests were carried out in situ and coating was included.

For thermal delay (φ), test elements with greater thickness and less voids present, in general, the highest values, as reported by Simões et al (74) and Tadeu et al. (75). This statement is even clearer when observing the test elements of the same thickness, with a difference only in the presence and size of the alveoli. Among the test elements with 100 mm, there is a 44% reduction between the massive one (3.5 h) and the hollow sample (2.3 h); for the 70 mm (GS 70 - and GS 70 +) blocks, there is 1.9 h for the hollow and 2.5 h for the similar compact. Finally, for the 76 mm elements, the difference was only 0.3 h: 1.5 hrs for GS 76 = and 1.8 hrs for GS 76 -; the 50 mm block had a thermal delay slightly higher than the GS 76 =, with 1.7 h.

The average temperature evolution of both the cold and warm environments in the thermal chamber is shown in Figure 8. Values were collected every 60 minutes through the display located on the controller board (Figure 4); the room temperature sensors were positioned close to the surface of the test elements on both sides.

Figure 8 shows that increasing temperature also causes an increase in standard deviation, resulting in greater variation, especially on the hot side of the chamber. Additionally, it was possible to compare the thermal behavior of the test elements in plaster block, since the variation occurred within an acceptable range, with a maximum value of 2.6 °C in the final minutes of heating. The graph for a 360-minute heating period for the 8 test elements is shown in Figure 9. In this graph, the control of the initial temperature is observed with values for all curves always within the range 26 ± 1 °C. The parameter most related to the thermal chamber behavior test is the thermal transmittance (U). To relate this parameter to the assay curves, Figure 10 presents a diagram where the test elements are on the sides in decreasing order for U value. In the central part, they are presented in the order of the curves for every 60 minutes.

However, one of the test elements, the GS 70 - (purple), shows less congruence in relation to the theoretical U value and its behavior during the test. Near 240 minutes, it presents a curve with very reduced acceleration and, therefore, almost no growth during the next 120 minutes (Figure 10). GS 76 - showed similar behavior. That happened especially between 180 and 300 minutes, with a possible state of balance between the room temperature on the cold side and the heat flow coming from the hot side. This situation may have been, in both cases, due to a cooling of the environment by the prolonged absence of operators, external climate, or even by distortions of the air conditioner thermostat. These situations were also indicated by Ferrari and Zanotto (62). Considering a trend in the curves of these two elements before the different points, in the final 60 minutes the relationship between U and the temperature on the cold side would probably be even clearer.

Relevant behavior is shown by the set of 100 mm blocks. The curves of the test elements GS 100 -, GH 100 + and GS 100 + remain isolated from the others between 40 and 260 minutes, returning to find the other hollow blocks (GS 70 -, GS 76) and approach the GS 76 = around 300 minutes. Behavior in 76 mm test elements should also be highlighted. It is important to note here that the difference between them is only due to the thickness of the alveoli, which is about 25% greater in GS 76 =. Figure 10 shows the first 80 minutes of heating, both show a similar behavior; however, after that time, the curve of the block with the largest number of voids and the highest U value maintains the same acceleration, while the GS 76 - presents a deceleration over a period of 200 minutes. It will increase again only in the final stage, after 320 minutes. Both in the measured values of thermal resistance by the flowmetry method, as well as by the theoretical calculation, the thermal parameters of the GS 76 - proved to be superior to those of the GS 76 = and. In the test, such values were corroborated, despite the greater temperature difference between them: 2.5 °C for about 260 minutes of heating and at the end of the process, over 1 °C.

The GS 50 + showed little capacity to retain the heat passage during the test, in accordance with the theoretical values of its low thermal resistance, R = 0.143 and R = 0.163, calculated and measured, respectively. At the end of the process, there was almost 10 °C of temperature difference for the GS 100 +, which showed the best behavior, and 6 °C of difference for the second with the worst performance, the GS 70 +. The use of this block is restricted to decorations, cabinets, and small closings. It is rarely used for closing masonry. One possibility of use would be associated with a back layer to the ventilation layer in double walls, as it presents the characteristics common to plaster, such as flatness suitable for finishing, low density of apparent mass; it presents greater thickness and self-supporting capacity than plasterboard: drywall boards common thickness is 12.5 mm, and they need metal profiles for support.

Results obtained from the thermal chamber enable understanding thermal dynamic behavior and verifying theoretical values for tested plaster blocks. Previous studies have also established these thermal chamber advantages for plaster compounds (45, 46).

To complement the data collected by a digital thermometer, the thermograms made it possible to observe the distribution of heat by the test element during the heating period (Figures 11 and 12). By the test element arrangement in the thermal chamber (bottom on the base of the chamber and the upper part with a gap for the top), there is a greater pre-disposition of heating in the upper part, aggravated by the lightness of the hot air that rises (76). The heat transfer to the base of the chamber by conduction promotes this lower temperature at the base of the test element, as seen in Figures 11 and 12. Although the expanded polystyrene strips have shown effectiveness in reducing the exchange of air between the hot and cold parts through the side and top cracks during the test, a possible improvement for the insulation is to make it closer to all the edges of the test elements, including the base.

The thermograms in Figures 11 and 12 corroborate the need to only use contact thermocouples in the central part of the test element, when heated by a point source, as mentioned in similar works (28, 30).

Figures 11 and 12 also indicate that the presence of the vertical alveoli (Figure 11) causes a more vertical distribution of heat due to the transfer by convection that occurs within the air layer. On the other hand, in the solid element (Figure 12), a more radial distribution of the heat focus is clearly seen in the center due to the position of the heating lamp. Unlike the hollow element, where heat tends to rise between the alveoli, conduction allows the heating to be better distributed in the captured area, including the base (see part “6h” Figure 12). This situation can even explain the cause of the behavior of the GS 70 -; in the simulation it differs from the expected, when observing the values ​​of its thermal parameters (Tables 6 and 7). Since the alveoli are open at the top through the gap between the test element and the roof of the thermal chamber, the exchange of hot air, even though it is hampered by the lateral closing and the use of polystyrene strips, occurred, and may have “part the heat flow” perpendicular to the face of the element for that point. Since the GS 70 - is the least thick hollow element, this circumstance affected its behavior more than the other hollow elements, as in Figure 10, and in the behavior scheme (Figure 9).

The possibility of visualizing the heat distribution and transmission by the thickness of the element over time was impossible due to the shape and opacity of the walls in the thermal chamber. In view of this, after the removal of the chamber, this visualization will become possible and allow relevant analyzes on the test.

The side view of the plaster block test element GS 100 + is shown in Figure 13 in digital image (upper left corner) and in thermal images during cooling. Four points were selected: 2 at the ends and 2 in the central part. It was possible to observe how the heat transfer occurs in the block by conduction and from it to the environment by convection, predominantly. When the test element was removed from the chamber, it was subjected to room temperature, around 26 °C - colder than any of the exposed faces. Besides, the test element itself had heat transfer still occurring, mostly by conduction. The flow occurs from the point of highest temperature to the lowest; therefore, heat transfer to the environment and between the thermally different points of the test element itself. In Figure 13, point SP1 is closest to the hot (heated) side and SP4 to the cold side.

Regarding the thermogram of 0 min, a lighter stain appears, indicating a warmer region. This is because this thermogram is made as soon as the test element has been removed from the thermal chamber, right after the influence of the heat source. In the following thermograms, it is possible to see that this stain “moves” towards the side center of the test element. Then, in the 90 min thermogram, a symmetrical temperature balance occurs, with the most heated points in the center. With the heat convection, the cooling tends to occur in a similar way in the two main directions. As cooling time happens, the temperatures tend to present a “normal” curve shape, considering, for this test, that room temperatures on both main faces are equal (temperature of the same environment). It is noteworthy that the cooling of the walls usually occurs under different room temperature values. Consequently, the heat flow tends to be different with greater temperature difference; it can become balanced when the external and internal temperatures are equal. Previous studies have also used infrared thermography to study plaster thermal during heating and cooling cycles (77, 78). Just like in these studies, infrared thermography allowing dynamic thermal evaluation, and identification of thermic resistance to heat fluxes were observed.