Improvement of near-infrared (NIR) reflectivity and black color tone by doping Zn2+ into the Ca2Mn0.85Ti0.15O4 structure

Inorganic black pigments with thermal barrier characteristics, Ca2Mn0.85−xTi0.15ZnxO4−x (0 ≤ x ≤ 0.10), were synthesized using a conventional solid-state reaction method in order to improve the blackness without decreasing the near-infrared (NIR) reflectance of a Ca2Mn0.85Ti0.15O4 pigment, which was previously reported by our group. The composition was optimized to provide both high blackness and NIR reflection characteristics. As a result, the NIR solar reflectance value (RNIR) of Ca2Mn0.77Ti0.15Zn0.08O3.92 (RNIR = 74.6%) became larger than that of Ca2Mn0.85Ti0.15O4 (RNIR = 71.7%), and the black color tone of the former (L* = 23.2, a* = +2.81, b* = +0.83, C = 2.93) was improved in comparison with that of the latter (L* = 24.4, a* = +4.30, b* = +2.72, C = 5.09). This improvement is caused by the introduction of strain into the [MnO6] octahedra and a decrease in the manganese ion concentration. The RNIR value of the Ca2Mn0.77Ti0.15Zn0.08O3.92 pigment was also larger than those of the commercially available pigments (RNIR < 53.0%). Therefore, Ca2Mn0.77Ti0.15Zn0.08O3.92 has potential to be an inorganic black pigment for thermal shielding.


Introduction
The urban heat-island effect leads to the ambient temperature in an urban area being higher than that in the surrounding areas. 1 This effect oen generates adverse effects such as heatstroke, discomfort, and a large consumption of electricity by air conditioners in the summer season. Natural sunlight consists of 5% ultraviolet radiation (UV; 280-400 nm), 43% visible radiation (400-700 nm) and 52% near-infrared radiation (NIR; 700-2500 nm). 2 Since the 700-1300 nm wavelength region constitutes 80% of the total energy in the NIR region, sunlight in this range plays the most important role in generating heat. 3 For this reason, it is effective to shield NIR light in this region in order to prevent heat storage. Many studies have been reported on several colored pigments that can reect NIR light. [4][5][6][7][8][9][10][11][12][13][14][15][16][17][18] The NIR-reectance properties of variously colored pigments (e.g. white, yellow, and blue) are generally better than those of black pigments, because these pigments tend to reect not only visible but also NIR light. 19,20 However, black pigments such as carbon black basically absorb NIR as well as visible light to store heat. When the common black pigment on the outer walls and roofs of buildings absorbs sunlight, the temperature rises, and at the same time the amount of exhaust heat from the use of air conditioners increases. Additionally, the heat stored during the daytime is released at night, and this heat dissipation prevents night cooling. These phenomena promote the urban heat island. For this reason, application of NIR-reective black pigments to road surfaces, building roofs, and exterior walls has attracted attention. 20,21 Some compounds such as (Fe, Cr) 2 O 3 , Fe 2 TiO 4 , and YMnO 3 have been proposed to serve as NIR-reective black pigments. [21][22][23][24] However, (Fe, Cr) 2 O 3 contains toxic chromium, and NIR-reective properties of Fe 2 TiO 4 and YMnO 3 are not enough.
In our previous study, we found that a Ca 2 Mn 0.85 Ti 0.15 O 4 pigment was a promising novel inorganic NIR-reective black pigment. 25 But, unfortunately, this pigment exhibited slightly reddish black color. In this study, therefore, Zn 2+ was doped into the Mn 4+ site to improve the blackness without decreasing the NIR reectance, because Zn 2+ does not show optical absorption in the NIR region. Namely, Ca 2 Mn 0.85Àx Ti 0.15 Zn x -O 4Àx (0 # x # 0.10) samples were synthesized and the NIR reectance and color properties were characterized. Finally, the Zn 2+ concentration was optimized to meet both enough black hue and high NIR reectivity.

Materials and methods
The Ca 2 Mn 0.85Àx Ti 0.15 Zn x O 4Àx (0 # x # 0.10) samples were synthesized using a conventional solid-state reaction method. Stoichiometric amounts of CaCO 3 (FUJIFILM Wako Pure Chemical), MnO 2 (FUJIFILM Wako Pure Chemical), TiO 2 (FUJIFILM Wako Pure Chemical), and ZnO (Kishida Chemical) were mixed in an agate mortar. The mixtures were calcined in an alumina boat at 1200 C for 6 h under an air atmosphere. Finally, the samples were ground in an agate mortar before characterization.

Characterization
The samples synthesized were characterized by X-ray powder diffraction (XRD; Rigaku, Ultima IV) with Cu-Ka radiation (40 kV and 40 mA). The sampling width and the scan speed were 0.02 and 6 min À1 , respectively. The sample compositions analyzed using X-ray uorescence spectroscopy (XRF; Rigaku, ZSX Primus) were in good agreement with the stoichiometric compositions of the starting mixtures. The lattice parameters and volumes were calculated from the XRD peak angles, which were rened using a-Al 2 O 3 as a standard and using CellCalc Ver. 2.20 soware. The morphology of the Ca 2 Mn 0.85Àx Ti 0.15 Zn x O 4Àx (x ¼ 0 and 0.08) particles was investigated by using eldemission-type scanning electron microscopy (FE-SEM; JEOL, JSM-6701F). The size distribution and the average particle size were estimated by measuring the diameters of 200 particles from the FE-SEM photographs.
The optical reectance spectra were measured with an ultraviolet-visible-near-infrared (UV-Vis-NIR) spectrometer (JASCO, V-770 with an integrating sphere attachment) with barium sulfate for the visible light region and polytetrauoroethylene (PTFE) for the NIR region as references. The total (R Tot , 300-2500 nm) and NIR (R NIR , 700-2500 nm) solar reectance was calculated in accordance with the American Society for Testing and Materials (ASTM) Standard G173-03, and was expressed as the integral of the product of the observed spectral reectance and the solar irradiance divided by the integral of the solar irradiance, both integrated over the each range as in the formula where r(l) is the spectral reectance obtained from the experiment and i(l) is the standard solar spectrum (W m À2 nm À1 ). The color property was evaluated in terms of the Commission Internationale de l'Éclairage (CIE) L*a*b*C system using a colorimeter (Konica-Minolta, CR-300). The L* parameter indicates the brightness or darkness of a color on relation to a neutral gray scale, and the a* (the red-green axis) and the b* (the yellow-blue axis) parameters express the color qualitatively. Chroma parameter (C) represents the color saturation of the pigments and is calculated according to the following formula:

Results and discussion
X-ray powder diffraction (XRD) and eld-emission-type scanning electron microscopic (FE-SEM) image The lattice parameters (a, c, and V) and the c/a ratios of all samples synthesized in this study were calculated from the XRD peak angles. These results are summarized in Table 1, where the numbers in parentheses indicate standard deviations. The cell volume increased as the Zn 2+ concentration increased in the range of 0 # x # 0.08, indicating that some Mn 4+ (ionic radius: 53.0 pm) 26 ions were partially substituted with the larger Zn 2+ (ionic radius: 74.0 pm) 26  Considering the charge compensation, the electroneutrality will be maintained by the generation of either higher valence Mn 5+/7+ or the formation of oxide anion vacancies, when Zn 2+ is doped into the Mn 4+ site. However, Mn 5+ is unstable in oxides Paper and tends to transfer into more stable Mn 4+ and Mn 7+ . 29 When Mn 7+ (ionic radius: 46 pm) 26 is generated in the structure for charge compensation, the lattice volume should increase nonlinearly, but that is not the case. Therefore, it is reasonable to consider that the electrical neutrality of Ca 2 Mn 0.85Àx Ti 0.15 Zn x -O 4Àx is maintained by the production of oxide anion vacancies.
The lattice parameters a and c also increased with increasing the Zn 2+ content. However, the increase rates of the former and the latter were different. The crystal structure of Ca 2 (Mn, Nb)O 4 was investigated by Taguchi. 30 When the Nb 5+ ions were introduced into the Mn 4+ site in Ca 2 MnO 4 , the c/a ratio decreased with increasing the Nb 5+ content, and the distortion of the [MnO 6 ] octahedra was alleviated. 30 On the contrary, in the case of the Ca 2 Mn 0.85Àx Ti 0.15 Zn x O 4Àx (0 # x # 0.10) samples synthesized in this study, the c/a ratio increased with increasing the amount of Zn 2+ in the range of 0 # x # 0.05, as seen in Table  1. Therefore, the distortion of the [MnO 6 ] octahedra was increased by the partial substitution of Mn 4+ with Zn 2+ . Fig. 3 shows the FE-SEM images and size distributions of the Ca 2 Mn 0.85Àx Ti 0.15 Zn x O 4Àx (x ¼ 0 and 0.08) samples. The faceted particles were observed in both samples. These particles were thermally fused due to the high-calcination temperature at 1200 C. In both samples, the average particle size was 0.97 mm and there was no signicant change in particle size, size distribution, and morphology. These results indicate that the changes in the optical and color properties of both samples were caused by the Zn 2+ doping. Fig. 4(a) depicts the UV-Vis and NIR reectance spectra of the Ca 2 Mn 0.85Àx Ti 0.15 Zn x O 4Àx (0 # x # 0.10) samples. All samples strongly absorbed visible light at a wavelength of 700 nm and shorter and reected NIR light, due to small bandgap energies around 1.77 eV. 25 An enlarged view of the reectance spectra from 300 to 750 nm was shown in Fig. 4(b). Optical reectance from 600 to 750 nm corresponding to the red light was decreased by the Zn 2+ doping. As a result, the color of the samples changed from slightly reddish black to more vivid black and the redness of the samples was reduced.

Reectance spectra
As discussed above in Table 1, the [MnO 6 ] octahedron was signicantly distorted by the dissolution of Zn 2+ . This lattice distortion increased with increasing the Zn 2+ concentration, and the symmetry of the [MnO 6 ] octahedron decreased. Thus, the reduction in red light reection was caused by the dd transition absorption of Mn 4+ . This transition is essentially forbidden but has been partially allowed due to the loss of symmetry. On the other hand, the Mn 4+ content of the sample was decreased by the Zn 2+ substitution. In other words, enhancement of the optical absorption by Mn 4+ in the red-light region and decrease of the amount of Mn 4+ responsible for this absorption are in the relationship of trade-off. Accordingly, the optical reectance in the red-light region almost unchanged in the range of 0.03 # x # 0.08.  On the other hand, the optical reectance in the NIR region was increased by the Zn 2+ doping, as seen in Fig. 4(c). The optical absorption in the NIR region was caused by the allowed charge transfer transition between Mn 4+ and Mn 3+ ions, 31 and this absorption intensity depended on the concentration of manganese. Unfortunately, it is difficult to conrm how much the CaO and ZnO impurity phases are involved in improving the NIR reectivity. Since CaO and ZnO can strongly reect the visible light as well as the NIR light, the optical reectance in the visible light region shall also be increased by the Zn 2+ doping, when the effect of these impurities is large. However, the reectance in the visible light region did not increase, but rather the reectance of red light decreased. Therefore, the increase in the NIR reectivity is dominantly due to the decrease in manganese ions in the sample.

Chromatic properties and NIR solar reectance
The color coordinate data and total (R Tot ) NIR solar reectance (R) of the Ca 2 Mn 0.85Àx Ti 0.15 Zn x O 4Àx (0 # x # 0.10) pigments are summarized in Table 2. The photographs of these pigments are also displayed in Fig. 5. All pigments synthesized in this study showed low L* values and were black as seen in Fig. 5. The a*, b*, and C values of the Zn 2+ -doped pigments were almost the same, but lower than those of the undoped Ca 2 Mn 0.85 Ti 0.15 O 4 pigment (x ¼ 0). As already discussed with respect to the results in Fig. 4(b), this was due to the decrease in the optical reection in the red-light region (600-750 nm). The R value increased conversely with the Zn 2+ doping, because the relative number of manganese ions decreased and the reectance in the NIR region was increased as seen in Fig. 4(a) and (c).
For an achromatic color such as black, the C value should be as small as possible in the L*a*b*C system. As recognized in the C and R NIR values of the Zn 2+ -doped samples in Table 2, the color tone became blacker and the NIR solar reectance was improved. Among the Ca 2 Mn 0.85Àx Ti 0.15 Zn x O 4Àx (0 # x # 0.08) pigments synthesized in this study, Ca 2 Mn 0.77 Ti 0.15 Zn 0.08 O 3.92 showed a relatively low C value and the highest R NIR value. Therefore, it was evidenced that Ca 2 Mn 0.77 Ti 0.15 Zn 0.08 O 3.92 has high performance as an inorganic black pigment with thermal barrier characteristics.

Comparison with commercially available pigments
The UV-Vis-NIR reectance spectrum and the color parameters of the Ca 2 Mn 0.77 Ti 0.15 Zn 0.08 O 3.92 pigment was compared with those of the commercially available black pigments such as Black 6350 (iron and chromium oxide, Asahi Kasei), Black 6301 (manganese and bismuth oxide), MPT-370 (calcium, manganese, and titanium oxide, Ca(Ti, Mn)O 3 , Ishihara Sangyo), and carbon black (Wako Chemical), as shown in Fig. 6 and summarized in Table 3. The photographs of these pigments are also displayed in Fig. 7. As evidenced from these results, the present Ca 2 Mn 0.77 Ti 0.15 Zn 0.08 O 3.92 pigment showed higher reectance in the NIR wavelength region and signicantly higher R NIR value than did the commercial pigments. Furthermore, the present pigment showed sufficiently low L* and C

Chemical stability test
The chemical stability of the Ca 2 Mn 0.77 Ti 0.15 Zn 0.08 O 3.92 pigment was also evaluated. The powder sample was soaked into 4% acetic acid and 4% ammonium bicarbonate aqueous solutions. Aer leaving them at room temperature for 24 h, the samples were washed with deionized water and ethanol, and then dried at room temperature. The NIR-reectance properties and color of the samples aer the chemical stability test was evaluated using the UV-Vis-NIR spectrometer and the colorimeter. Unfortunately, the color degradation and the decrease of the R Tot and R NIR values were observed by leaching the sample in both acetic acid and basic ammonium carbonate solutions, as seen in Table 4. Therefore, it is suggested that surface coating with a stable compound such as silica is necessary to suppress the deterioration.

Conflicts of interest
There are no conicts to declare.