MnNH 4 P 2 O 7 -Based Coating for High Temperature Assessment on the Surfaces of Cement Composites

: This study examines the implementation of an MnNH 4 P 2 O 7 (ammonium manganese III pyrophosphate)-based coating on structural elements to obtain temperature information with color changes. Based on the MnNH 4 P 2 O 7 material, a coating was prepared and deposited on cement mortar surfaces. Heat experiments were then conducted to evaluate the thermochromism on the fabricated samples. The coated samples exhibited a superior irreversible thermochromic property at 400 ◦ C with a color change from dark violet to light grayish blue at the heated surface. The color changes were retrieved at each temperature using a digital camera, and the change in color properties was evaluated in the RGB and L * a * b * color spaces using image processing techniques. With increasing temperature from room temperature, the RGB values were almost constant until 200 ◦ C. At higher temperatures, the color changes started to accelerate until 400 ◦ C. The values showed a 167%, 567%, and 49% increase in R, G, and B values, respectively, at 400 ◦ C. In the L * a * b * color space, when the temperature was increased from room temperature to 400 ◦ C, the L * a * b * values showed an increase of 211%, a decrease of 94%, and an increase of 78%, respectively.


Introduction
Thermochromism is a phenomenon where the color changes as a response to variations in temperature. The color changes are based on the transition and/or transformation of the molecules when the materials are heated up or cooled. The materials can exist in different forms, such as metal oxides, polymers, solid-state semiconductors, and leuco-dyes [1,2]. Thermochromic materials can be classified as reversible and irreversible materials. Reversible thermochromic materials exhibit a shift in color when heated and a return to their original colors when cooled down to room temperature. These materials have been applied for solar reflectance in structures to provide a thermally comfortable indoor environment [3][4][5][6][7]. Kamalisarvestani et al. [8] proposed a thermochromic window coating that changes its color to block solar radiation. In addition, a temperature sensor based on a polydiacetylene material was proposed to show reversible color properties in solutions in response to temperatures between 30 and 70 • C [9]. Irreversible thermochromic materials experience a permanent color change at a certain temperature. Some studies were performed to utilize these materials as thermal indicating paints that can monitor the temperature change profiles and provide a damage warning on aero-engine components [10][11][12]. Moreover, Rabhiou et al. [13] proposed a phosphor-based irreversible thermal coating to measure the temperature in the range of 600 to 1000 • C. Most of the previous studies focused on temperature indicating systems on combustors used in gas turbine engines [14] and on the hot surfaces of metals [15]. On the other hand, few studies were designed to apply irreversible materials Table 1. Typical characteristics of MnNH 4 P 2 O 7 pigment [26].

Color Dark violet Chemical characterization
Ammonium manganese III pyrophosphate Density 2.7-2.9 kg/m 3 Bulk density 0.60 g/cm 3 Average particle size 2.30 µm pH value 2.5-4.7 Thermal decomposition >400 • C the pigment at room temperature. When heated to 370 °C from room temperature, the pigment particles changed from dark violet to reversible blue color, as shown in Figure 1b. At the discoloration temperature, the H2O molecules are driven off from the particles, which leads to the change in Mn2P4O13(NH3)2 and blue color as expressed in Equation (1). When exposed to the room atmosphere, Mn2P4O13(NH3)2 reacts with atmospheric moisture and returns to the original dark violet color [27].
At the higher temperature of 410 °C, the particles experience an irreversible color change to grayish yellow, as shown in Figure 1c. The application of further heat to the pigment causes the liberation of NH3 from the particles, which is associated with the oxidation of hydroxylamine to change Mn2P4O13(NH3)2 to Mn2P4O12 (manganous tetra-metaphosphate), as described in Equation (2). Then, the oxidized hydroxylamine (NH3OH) is broken into nitrogen (N2) and water (H2O), or the ammonia (NH3) is directly oxidized to nitrogen [27]. When cooled to room temperature, the particles retain the grayish yellow color and do not recover the original color.   When heated to 370 • C from room temperature, the pigment particles changed from dark violet to reversible blue color, as shown in Figure 1b. At the discoloration temperature, the H 2 O molecules are driven off from the particles, which leads to the change in Mn 2 P 4 O 13 (NH 3 ) 2 and blue color as expressed in Equation (1). When exposed to the room atmosphere, Mn 2 P 4 O 13 (NH 3 ) 2 reacts with atmospheric moisture and returns to the original dark violet color [27].
At the higher temperature of 410 • C, the particles experience an irreversible color change to grayish yellow, as shown in Figure 1c. The application of further heat to the pigment causes the liberation of NH 3 from the particles, which is associated with the oxidation of hydroxylamine to change Mn 2 P 4 O 13 (NH 3 ) 2 to Mn 2 P 4 O 12 (manganous tetra-metaphosphate), as described in Equation (2). Then, the oxidized hydroxylamine (NH 3 OH) is broken into nitrogen (N 2 ) and water (H 2 O), or the ammonia (NH 3 ) is directly oxidized to nitrogen [27]. When cooled to room temperature, the particles retain the grayish yellow color and do not recover the original color.

Preparation of Specimens and Test Variables
The following procedures were followed to prepare the specimens, resembling the application of coatings in real buildings. Sample preparation consisted of two stages. The first stage involved the preparation of coating surfaces, and the second stage included the preparation of coatings and fabrication.
Cement mortar samples, 50 × 50 × 50 mm 3 in size, were prepared by mixing water, silica sand, and ordinary Portland cement (Type I), manufactured at a local company. The ratios of water to cement and sand to cement were 0.4 and 2, respectively. The prepared specimens were left under cover inside the laboratory for 24 h. They were then removed from the molds and cured in water in a constant temperature and humidity chamber (20 ± 2 • C, 60% ± 10%) until the test. Subsequently, the mortar specimens were dried for 24 h at 100 • C in an oven. Figure 2 shows the coating procedure on a cross section of the specimens. The mortar surfaces were cleaned using sandpaper, and the dust and dirt particles were removed to maintain a clean surface and allow adhesion of the applied coating to the mortar surface, as shown in Figure 3a. The first coating, called the prime coating, was dropped on the surface with silica-based binder and water at a 1:5 ratio. The prime coated samples were placed in an oven at 100 • C for 10 min and dried for 6 h at ambient temperature, as shown in Figure 3b. To prepare the coating solution, pigment and water were added at a 2:3 ratio and stirred for 10 min at room temperature with the slow addition of 10% silica-based binder under stirring. Finally, the coating solution was poured on the specimen surface and coated in a single layer, as shown in Figure 3c. The coated surfaces were left undisturbed at room temperature for seven days to cure and achieve complete color exposure on the surfaces. After curation, the specimens were placed in a furnace chamber and heated to the target temperatures of 100, 200, 300, 400, and 450 • C. At each target temperature, the specimen was maintained for around 30 min and taken out to photograph the surface of the specimen using a digital camera. After that, the specimen was placed again in the furnace and heated to the next target temperature.

Preparation of Specimens and Test Variables
The following procedures were followed to prepare the specimens, resembling the application of coatings in real buildings. Sample preparation consisted of two stages. The first stage involved the preparation of coating surfaces, and the second stage included the preparation of coatings and fabrication.
Cement mortar samples, 50 × 50 × 50 mm 3 in size, were prepared by mixing water, silica sand, and ordinary Portland cement (Type I), manufactured at a local company. The ratios of water to cement and sand to cement were 0.4 and 2, respectively. The prepared specimens were left under cover inside the laboratory for 24 h. They were then removed from the molds and cured in water in a constant temperature and humidity chamber (20 ± 2 °C, 60% ± 10%) until the test. Subsequently, the mortar specimens were dried for 24 h at 100 °C in an oven. Figure 2 shows the coating procedure on a cross section of the specimens. The mortar surfaces were cleaned using sandpaper, and the dust and dirt particles were removed to maintain a clean surface and allow adhesion of the applied coating to the mortar surface, as shown in Figure 3a. The first coating, called the prime coating, was dropped on the surface with silica-based binder and water at a 1:5 ratio. The prime coated samples were placed in an oven at 100 °C for 10 min and dried for 6 h at ambient temperature, as shown in Figure 3b. To prepare the coating solution, pigment and water were added at a 2:3 ratio and stirred for 10 min at room temperature with the slow addition of 10% silica-based binder under stirring. Finally, the coating solution was poured on the specimen surface and coated in a single layer, as shown in Figure 3c. The coated surfaces were left undisturbed at room temperature for seven days to cure and achieve complete color exposure on the surfaces. After curation, the specimens were placed in a furnace chamber and heated to the target temperatures of 100, 200, 300, 400, and 450 °C. At each target temperature, the specimen was maintained for around 30 min and taken out to photograph the surface of the specimen using a digital camera. After that, the specimen was placed again in the furnace and heated to the next target temperature.

Preparation of Specimens and Test Variables
The following procedures were followed to prepare the specimens, resembling the application of coatings in real buildings. Sample preparation consisted of two stages. The first stage involved the preparation of coating surfaces, and the second stage included the preparation of coatings and fabrication.
Cement mortar samples, 50 × 50 × 50 mm 3 in size, were prepared by mixing water, silica sand, and ordinary Portland cement (Type I), manufactured at a local company. The ratios of water to cement and sand to cement were 0.4 and 2, respectively. The prepared specimens were left under cover inside the laboratory for 24 h. They were then removed from the molds and cured in water in a constant temperature and humidity chamber (20 ± 2 °C, 60% ± 10%) until the test. Subsequently, the mortar specimens were dried for 24 h at 100 °C in an oven. Figure 2 shows the coating procedure on a cross section of the specimens. The mortar surfaces were cleaned using sandpaper, and the dust and dirt particles were removed to maintain a clean surface and allow adhesion of the applied coating to the mortar surface, as shown in Figure 3a. The first coating, called the prime coating, was dropped on the surface with silica-based binder and water at a 1:5 ratio. The prime coated samples were placed in an oven at 100 °C for 10 min and dried for 6 h at ambient temperature, as shown in Figure 3b. To prepare the coating solution, pigment and water were added at a 2:3 ratio and stirred for 10 min at room temperature with the slow addition of 10% silica-based binder under stirring. Finally, the coating solution was poured on the specimen surface and coated in a single layer, as shown in Figure 3c. The coated surfaces were left undisturbed at room temperature for seven days to cure and achieve complete color exposure on the surfaces. After curation, the specimens were placed in a furnace chamber and heated to the target temperatures of 100, 200, 300, 400, and 450 °C. At each target temperature, the specimen was maintained for around 30 min and taken out to photograph the surface of the specimen using a digital camera. After that, the specimen was placed again in the furnace and heated to the next target temperature.

Image Acquisition and Post Processing
The color data were collected by photographing the surfaces at all temperature intervals. A digital camera (ILCE-TR, SONY, Tokyo, Japan) with a 7360 × 4912-pixel resolution was placed at a fixed position, approximately 30 cm above the specimens. All the images were recorded using the highest pixel count (most pixels per object area) to provide high quality images. The recorded photographs were imported to a computer and analyzed using an image processing technique. The color and brightness of the light source have a significant influence on the color of the surface. Photographs were taken in a consistent illuminance throughout at all temperatures to minimize the variation in color and brightness.
The color of the images on the surface of the samples was analyzed using MATLAB Version 9.6 programming to extract the RGB and L*a*b* color information. From the images obtained at different temperatures, the surfaces of the samples were taken from the background as the regions of interest, as shown in Figure 4a. That is, the sample surface of approximately 1000 × 1000 pixels was separated from the total image (7360 × 4912 pixels). Then, the mean and standard deviation of the RGB and L*a*b* color values were obtained from the separated sample surface at each pixel. Furthermore, the separated surface was divided into 4 × 4 regions to present more accurate color values, as shown in Figure 4b. The mean and standard deviation of RGB and L*a*b* color values were also calculated using the separate 4 × 4 regions defined as approximately 250 × 250 pixels for each region.
highest pixel count (most pixels per object area) to provide high quality images. The recorded photographs were imported to a computer and analyzed using an image processing technique. The color and brightness of the light source have a significant influence on the color of the surface. Photographs were taken in a consistent illuminance throughout at all temperatures to minimize the variation in color and brightness.
The color of the images on the surface of the samples was analyzed using MATLAB Version 9.6 programming to extract the RGB and L*a*b* color information. From the images obtained at different temperatures, the surfaces of the samples were taken from the background as the regions of interest, as shown in Figure 4a. That is, the sample surface of approximately 1000 × 1000 pixels was separated from the total image (7360 × 4912 pixels). Then, the mean and standard deviation of the RGB and L*a*b* color values were obtained from the separated sample surface at each pixel. Furthermore, the separated surface was divided into 4 × 4 regions to present more accurate color values, as shown in Figure 4b. The mean and standard deviation of RGB and L*a*b* color values were also calculated using the separate 4 × 4 regions defined as approximately 250 × 250 pixels for each region.
Each color in the RGB space is represented by a combination of spectral components of red, green, and blue colors. The values of each RGB component are in the range 0 to 255. The measured RGB color values are related to the lighting conditions (color and brightness), colors of the objects being recorded, and the sensitivity of the recording image sensor. Thus, the RGB color space used in digital cameras is device-dependent. The color values of the pigment can correspond to predefined values of the ColorHexa codes. Hence, the captured images were adjusted by changing the illumination and settings of the camera to match the predefined ColorHexa codes [38,39]. ColorHexa is a six-digit color code providing information about the colors obtained from the images [40]. Therefore, to compensate for the variations due to the color and brightness of the light source, the RGB images were transformed into standardized, device-independent CIE1976 (L*a*b*) space using MATLAB programming. This resulted in values between 0 and 100 for the L* dimension, and −128 and 128 for a* and b* dimensions.

Thermal Response of Cement Composites and MnNH4P2O7-Based Coatings with Temperatures
MnNH4P2O7-based inorganic coatings were applied to a cement mortar specimen using pigment coating formulations and curing conditions. Figure 5 presents the surface color of the mortar specimen without the MnNH4P2O7 coating. At room temperature, the surface color of the cement Each color in the RGB space is represented by a combination of spectral components of red, green, and blue colors. The values of each RGB component are in the range 0 to 255. The measured RGB color values are related to the lighting conditions (color and brightness), colors of the objects being recorded, and the sensitivity of the recording image sensor. Thus, the RGB color space used in digital cameras is device-dependent. The color values of the pigment can correspond to predefined values of the ColorHexa codes. Hence, the captured images were adjusted by changing the illumination and settings of the camera to match the predefined ColorHexa codes [38,39]. ColorHexa is a six-digit color code providing information about the colors obtained from the images [40]. Therefore, to compensate for the variations due to the color and brightness of the light source, the RGB images were transformed into standardized, device-independent CIE1976 (L*a*b*) space using MATLAB programming. This resulted in values between 0 and 100 for the L* dimension, and −128 and 128 for a* and b* dimensions.

Thermal Response of Cement Composites and MnNH 4 P 2 O 7 -Based Coatings with Temperatures
MnNH 4 P 2 O 7 -based inorganic coatings were applied to a cement mortar specimen using pigment coating formulations and curing conditions. Figure 5 presents the surface color of the mortar specimen Coatings 2020, 10, 396 6 of 15 without the MnNH 4 P 2 O 7 coating. At room temperature, the surface color of the cement mortar exhibited a dark grayish color, as shown in Figure 5a. With increasing temperature, the surface color was very similar to the dark grayish color presented at room temperature. At high temperatures between 300 and 450 • C, the surface color of the mortar was slightly darkened, as shown in Figure 5d-f, but no significant color changes were observed by naked eye.
the violet-pattern color on the surface disappeared. That is, the color of the coating changed from very dark desaturated blue to grayish violet with the evaporation of water particles. As the temperature was increased to 400 °C, the surface coating displayed excellent thermochromic progress and turned to light grayish blue. The coatings achieved an extensive, irreversible color change at 400 °C, which was 10 °C lower than the temperature observed for the pigment. This might be due to the occurrence of pigments dispersed in the coating, which was manufactured by a 2:3 pigment-to-water ratio. Further heating to 450 °C resulted in a stable light grayish blue color. Figure 6e,f show the color changes of the samples with a MnNH4P2O7 coating at 400 °C and 450 °C, respectively. After cooling to room temperature, the samples retained the changed light grayish blue color, indicating irreversible thermochromic behavior. When applied to the surface of the concrete structures, the thermochromic MnNH4P2O7 coatings developed in the laboratory can provide a tool to determine the temperature, with a change in color defined as a function of temperature.  The thermochromic behavior of the MnNH 4 P 2 O 7 coating on the mortar is shown in Figure 6. The mortar sample coated with the MnNH 4 P 2 O 7 coating showed a dark violet color at ambient temperature, as illustrated in Figure 6a. Figure 6b,c show samples at 100 and 200 • C, respectively. At 100 • C, the coating sample displayed the same dark violet color observed at ambient temperature. At 200 • C, the sample exhibited a very dark, desaturated blue color. This was attributed to the evaporation of water particles, due to heating, moving towards the outer surface of the mortar specimen due to the increase in temperature. Figure 6d shows the sample heated to 300 • C, at which the violet-pattern color on the surface disappeared. That is, the color of the coating changed from very dark desaturated blue to grayish violet with the evaporation of water particles. As the temperature was increased to 400 • C, the surface coating displayed excellent thermochromic progress and turned to light grayish blue. The coatings achieved an extensive, irreversible color change at 400 • C, which was 10 • C lower than the temperature observed for the pigment. This might be due to the occurrence of pigments dispersed in the coating, which was manufactured by a 2:3 pigment-to-water ratio. Further heating to 450 • C resulted in a stable light grayish blue color. Figure 6e,f show the color changes of the samples with a MnNH 4 P 2 O 7 coating at 400 • C and 450 • C, respectively. After cooling to room temperature, the samples retained the changed light grayish blue color, indicating irreversible thermochromic behavior. When applied to the surface of the concrete structures, the thermochromic MnNH 4 P 2 O 7 coatings developed in the laboratory can provide a tool to determine the temperature, with a change in color defined as a function of temperature.   The quantitative difference of colors on the surface of the coating samples at ambient temperature and those at critical temperature when the color completely changed were analyzed  The RGB values decreased by 8%, 9%, and 11% at 300 • C to show the mean (124, 118, 109), respectively, corresponding to dark grayish orange (Hex code #7c766d). After that, the RGB values exhibited relatively similar values of (123, 119, 113) at 400 • C and (123, 118, 113) at 450 • C, which correspond to dark grayish orange (Hex codes #7b7771 and #7b7671, respectively). The cement mortar with some pores on the surface provided some variations in the RGB values, which led to the standard deviation in the range of 20 to 24.   The quantitative difference of colors on the surface of the coating samples at ambient temperature and those at critical temperature when the color completely changed were analyzed  The quantitative difference of colors on the surface of the coating samples at ambient temperature and those at critical temperature when the color completely changed were analyzed using the images obtained at each temperature. The variation in the thermochromic color of the coating samples with increasing temperature was expressed in three-dimensional RGB color space. The mean and standard deviation of the RGB values were calculated at each pixel from the sample surfaces. Figure 8 presents the RGB mean values with the standard deviation obtained from the surfaces with increasing temperature. The mean RGB values were interpreted to determine their corresponding ColorHexa codes and color description at various temperatures, as shown in Table 2. surfaces. Figure 8 presents the RGB mean values with the standard deviation obtained from the surfaces with increasing temperature. The mean RGB values were interpreted to determine their corresponding ColorHexa codes and color description at various temperatures, as shown in Table 2.

Color Changes Obtained in the RGB Color Space from Digital Images with Temperatures
The RGB values at ambient temperature were (76, 31, 157), which represent a dark violet color (Hex code #4b1f9c). These values were relatively constant and stable until 200 °C. As the temperature was increased to 300 °C, the RGB values increased to (159, 146, 187), and the color changed to grayish violet (Hex code #9E91BB). At 400 °C, the thermochromic coating lost its chroma and transformed from dark to light grayish blue (Hex code #cad0e9) with an increase in all the RGB values to (202,209,233). Compared to those at room temperature, the RGB values increased by approximately 167%, 567%, and 49%, respectively. Few changes in the RGB values were observed when the temperature was increased to above 400 °C. The standard deviations of the RGB values, as shown in Figure 8, were in the range of 13 to 30 from room temperature to 200 °C. The deviations of the RGB values revealed some variations at each pixel of the image mainly due to uneven coatings and some pores on the surface. After the coating surface changed completely to light grayish blue at 400 °C, the deviations were relatively low, in the range of 10 to 11.  The images were divided into 4 × 4 regions on the surface, and the mean RGB values were extracted over the divided regions. The values extracted at 4 × 4 regions were plotted in the threedimensional diagrams shown in Figure 9.   The RGB values at ambient temperature were (76, 31, 157), which represent a dark violet color (Hex code #4b1f9c). These values were relatively constant and stable until 200 • C. As the temperature was increased to 300 • C, the RGB values increased to (159, 146, 187), and the color changed to grayish violet (Hex code #9E91BB). At 400 • C, the thermochromic coating lost its chroma and transformed from dark to light grayish blue (Hex code #cad0e9) with an increase in all the RGB values to (202,209,233). Compared to those at room temperature, the RGB values increased by approximately 167%, 567%, and 49%, respectively. Few changes in the RGB values were observed when the temperature was increased to above 400 • C. The standard deviations of the RGB values, as shown in Figure 8, were in the range of 13 to 30 from room temperature to 200 • C. The deviations of the RGB values revealed some variations at each pixel of the image mainly due to uneven coatings and some pores on the surface. After the coating surface changed completely to light grayish blue at 400 • C, the deviations were relatively low, in the range of 10 to 11.
The images were divided into 4 × 4 regions on the surface, and the mean RGB values were extracted over the divided regions. The values extracted at 4 × 4 regions were plotted in the three-dimensional diagrams shown in Figure 9. Figure 9 represents a surface plot of the RGB values at room temperature and 400 • C. The mean R values obtained over 4 × 4 regions were between 60 and 93 at room temperature and 200 and 207 at 400 • C. The mean G values in the 4 × 4 regions were between 17 and 47 at room temperature and between 205 and 211 at 400 • C. The B values at room temperature in the 4 × 4 regions ranged from 142 to 176, and from 231 to 236 at 400 • C. As discussed with respect to the standard deviations, the RGB values obtained in 4 × 4 regions had some variations at room temperature but similar values at 400 • C. Some changes at room temperature might have been due to variations in the coating density of the dark and light particles on the relatively small surface. At 400 • C, the color change was obtained uniformly throughout the surface, which minimized the variations.
where ΔR, ΔG, and ΔB represent the color changes, in which R1, G1, and B1 are the mean values at room temperature, and R2, G2, and B2 are the mean values at 400 °C. The total color change ΔD showed an increase of intensities from room temperature to 400 °C, at which point ΔD reached approximately 231.
(a) (b) (c)  In addition, this study analyzed the total color changes of the mean RGB values obtained at room temperature and 400 • C. The total color change is defined as the Euclidean distance (∆D) in the RGB color space shown in Equations (3)-(6) where ∆R, ∆G, and ∆B represent the color changes, in which R 1 , G 1 , and B 1 are the mean values at room temperature, and R 2 , G 2 , and B 2 are the mean values at 400 • C. The total color change ∆D showed an increase of intensities from room temperature to 400 • C, at which point ∆D reached approximately 231.

Color Changes Obtained in the L*a*b* Color Space from Digital Images with Temperatures
The RGB images obtained were converted to CIE 1976 (L*a*b*) color space to provide more standardized results. The lightness (L*) is given a value of 100 for white and 0 for ideal black. The a* and b* parameters are in the range of −128 and 128; a* indicates greenness and redness, respectively, and b* indicated blueness and yellowness, respectively. Figure 10 presents the changes in the L*a*b* values on the mortar surface with increasing temperature. The L*a*b* values at room temperature were (57, 0.2, 4), which remained almost constant until 200 • C. At 300 • C, a slight decrease and increase in the L* and a* values was obtained to show the L*a*b* values of (49, 0.9, 5). At the temperatures of 400 • C and 450 • C, the L*a*b* values were very similarly (50, 0.6, 4) and (50, 0.7, 4), respectively. Thus, little change was found in the L*a*b* values on the surfaces of the mortar specimen with increasing temperature. The deviations obtained in the L* value were in the range of 8 to 9, and those in the a* and b* values were in the range of 1 to 2.
were (57, 0.2, 4), which remained almost constant until 200 °C. At 300 °C, a slight decrease and increase in the L* and a* values was obtained to show the L*a*b* values of (49, 0.9, 5). At the temperatures of 400 °C and 450 °C, the L*a*b* values were very similarly (50, 0.6, 4) and (50, 0.7, 4), respectively. Thus, little change was found in the L*a*b* values on the surfaces of the mortar specimen with increasing temperature. The deviations obtained in the L* value were in the range of 8 to 9, and those in the a* and b* values were in the range of 1 to 2. Figure 11 presents the changes on the coating surface in the L*a*b* values with increasing temperatures. The mean and standard deviation of the L*a*b* values and their corresponding color charts are also summarized in Table 3. The mean L* value at room temperature was 26, which represents a dark appearance. The value remained relatively constant up to 200 °C. As the temperature was increased further to 300 °C, the L* value increased to 62. At 400 °C, the L* value rapidly increased to 84, and the color completely changed to a light color. Compared to that at room temperature, the L* value at 400 °C increased by 211%. The mean a* and b* values at room temperature were 48 and −60, respectively, which means redness (+a*) and blueness (−b*). From the temperature of 200 °C, the a* value gradually decreased with a concomitant increase in b* value. Compared to the a* and b* values at room temperature, the a* value at 400 °C decreased by 94% to a value of 2, and the b* value at 400 °C increased by 78% to −12. The standard deviations were also obtained at all pixels in the images at all temperatures. The standard deviations of the L*a*b* values between room temperature until 300 °C ranged from 5 to 8, 2 to 9, and 2 to 11, respectively. When the temperature reached more than 400 °C, the deviations of the L* values were in the range of 5 to 6, whereas those of the a* and b* values were in the range of 1 to 2. This means that the color change obtained at the critical temperature of 400 °C was uniform over the entire coating surface, which showed relatively small deviations.   Figure 11 presents the changes on the coating surface in the L*a*b* values with increasing temperatures. The mean and standard deviation of the L*a*b* values and their corresponding color charts are also summarized in Table 3. The mean L* value at room temperature was 26, which represents a dark appearance. The value remained relatively constant up to 200 • C. As the temperature was increased further to 300 • C, the L* value increased to 62. At 400 • C, the L* value rapidly increased to 84, and the color completely changed to a light color. Compared to that at room temperature, the L* value at 400 • C increased by 211%. The mean a* and b* values at room temperature were 48 and −60, respectively, which means redness (+a*) and blueness (−b*). From the temperature of 200 • C, the a* value gradually decreased with a concomitant increase in b* value. Compared to the a* and b* values at room temperature, the a* value at 400 • C decreased by 94% to a value of 2, and the b* value at 400 • C increased by 78% to −12. The standard deviations were also obtained at all pixels in the images at all temperatures. The standard deviations of the L*a*b* values between room temperature until 300 • C ranged from 5 to 8, 2 to 9, and 2 to 11, respectively. When the temperature reached more than 400 • C, the deviations of the L* values were in the range of 5 to 6, whereas those of the a* and b* values were in the range of 1 to 2. This means that the color change obtained at the critical temperature of 400 • C was uniform over the entire coating surface, which showed relatively small deviations.  Moreover, the images were extracted into 4 × 4 regions, and the L*a*b* information was analyzed, as shown in Figure 12. In Figure 12a, the mean L* values in the 4 × 4 regions were in the range of 22 to 33 at room temperature and 83 to 84 at 400 °C. The mean a* values obtained at room temperature were between 47 and 51 in the 4 × 4 regions, and those at 400 °C were between 2 and 3, as shown in Figure 12b. The mean b* values obtained in the 4 × 4 regions ranged from −61 to −58 at room temperature and −13 to −12 at 400 °C, as shown in Figure 12c.
The total color changes were also calculated based on the L*a*b* values using Equations (7)-(10). ∆ * = (∆ * ) + (∆ * ) + (∆ * ) ∆ * = * − * (8) where L1*, a1*, and b1* represent the color values at room temperature, and L2*, a2*, and b2* represent the values at 400 °C. With ΔL* = 57 of the coating surfaces indicating a light color, the change in a*, i.e., Δa*, was −46, indicating a light shade of green. The change in b*, i.e., Δb*, was 47, indicating a light shade of yellow. Thus, the total color change ΔD* of the coating surface was approximately 87.  Moreover, the images were extracted into 4 × 4 regions, and the L*a*b* information was analyzed, as shown in Figure 12. In Figure 12a, the mean L* values in the 4 × 4 regions were in the range of 22 to 33 at room temperature and 83 to 84 at 400 • C. The mean a* values obtained at room temperature were between 47 and 51 in the 4 × 4 regions, and those at 400 • C were between 2 and 3, as shown in Figure 12b. The mean b* values obtained in the 4 × 4 regions ranged from −61 to −58 at room temperature and −13 to −12 at 400 • C, as shown in Figure 12c.
The total color changes were also calculated based on the L*a*b* values using Equations (7)- (10).
where L 1 *, a 1 *, and b 1 * represent the color values at room temperature, and L 2 *, a 2 *, and b 2 * represent the values at 400 • C. With ∆L* = 57 of the coating surfaces indicating a light color, the change in a*, i.e., ∆a*, was −46, indicating a light shade of green. The change in b*, i.e., ∆b*, was 47, indicating a light shade of yellow. Thus, the total color change ∆D* of the coating surface was approximately 87.

Summary
The fabricated samples with the MnNH4P2O7-based coating were found to respond to increasing temperature. Up to 100 °C, the coating sample displayed the same dark violet color obtained at ambient temperature. As the temperature was increased to 200 °C, the sample showed a very dark desaturated blue color mainly due to the evaporation of water particles. The violet-pattern color disappeared and changed to grayish violet at 300 °C. At 400 °C, the coating surface completely turned to light grayish blue. Further heating to higher temperatures and cooling to room temperature resulted in a stable light grayish blue color, which indicates irreversible thermochromic behavior.
The RGB values at ambient temperature, which represent a dark violet color, were almost constant and stable to 200 °C. At 400 °C, when the thermochromic coating lost its darkness and completely changed to light grayish blue, the RGB values increased by 167%, 567%, and 49%, respectively. The RGB values also exhibited relatively small deviations at 400 °C, which means that a relatively uniform color change was achieved throughout the entire surface. The total change in the RGB values, which is defined as the Euclidean distance, showed an increasing trend with increasing temperature and reached approximately 231 at a critical temperature of 400 °C. A small change in the RGB values was obtained when the temperature was increased to above 400 °C.

Summary
The fabricated samples with the MnNH 4 P 2 O 7 -based coating were found to respond to increasing temperature. Up to 100 • C, the coating sample displayed the same dark violet color obtained at ambient temperature. As the temperature was increased to 200 • C, the sample showed a very dark desaturated blue color mainly due to the evaporation of water particles. The violet-pattern color disappeared and changed to grayish violet at 300 • C. At 400 • C, the coating surface completely turned to light grayish blue. Further heating to higher temperatures and cooling to room temperature resulted in a stable light grayish blue color, which indicates irreversible thermochromic behavior.
The RGB values at ambient temperature, which represent a dark violet color, were almost constant and stable to 200 • C. At 400 • C, when the thermochromic coating lost its darkness and completely changed to light grayish blue, the RGB values increased by 167%, 567%, and 49%, respectively. The RGB values also exhibited relatively small deviations at 400 • C, which means that a relatively uniform color change was achieved throughout the entire surface. The total change in the RGB values, which is defined as the Euclidean distance, showed an increasing trend with increasing temperature and reached approximately 231 at a critical temperature of 400 • C. A small change in the RGB values was obtained when the temperature was increased to above 400 • C.
In the L*a*b* color space, the L* value was essentially dark at room temperature but rapidly increased toward the lightness intensity at 400 • C. The a* and b* values demonstrated rich redness and blueness, respectively, at room temperature. The L*a*b* values at room temperature remained relatively constant up to 200 • C. At 400 • C, when the coating surface completely turned to light grayish blue, the L* and b* values at 400 • C increased by 211% and 78%, respectively, and the a* decreased by approximately 94%. Similar to those of the RGB values, the standard deviations of the L*a*b* values were also relatively small, indicating that the color change obtained at a critical temperature of 400 • C was uniform over the entire coating surface. The total change in the L*a*b* values of the coating from the room temperature to the critical temperature of 400 • C was approximately 87.

Conclusions
This study presented a method for the quick and exact assessment of the temperature profiles after an abnormally high temperature was applied to a structure element. A thermochromic coating based on MnNH 4 P 2 O 7 was fabricated and successfully applied to the mortar surface. The applied coating provided the excellent color changing process from dark violet to light grayish blue when heated to 400 • C. In addition to the change in the visual color, the color properties in the RGB and L*a*b* spaces were evaluated as a function of temperatures.
This study successfully demonstrated the feasibility and applicability of the thermochromic MnNH 4 P 2 O 7 coating as a temperature detection technology. When applied to cementitious materials and structures, the thermochromic MnNH 4 P 2 O 7 coating could provide a tool for determining the temperature information with the color information defined as a function of temperature. In addition, the results of the study indicate that the developed coating holds great potential for applications in electric appliances, jet engines, and automotive sectors in areas of research and development, maintenance, and non-destructive testing.