3.1. Powder X-Ray Diffraction Analysis
The XRD patterns of Cr
2O
3 samples calcined at 900/1000 °C and the precursor without calcination are depicted in
Figure 3. The absence of peaks in the precursor indicates a typical amorphous structure, which is different from the results in similar studies of solution combustion synthesis [
10,
11]. In this work, the additions of citrate acid, urea, and PEG 200 were optimized by multiple pre-experiments, and then stable combustion at lower temperatures was created. With the utilization of a flame nozzle in the combustion furnace, the precursor generated can merely stay in the high-temperature area of the flame for seconds before it is pushed out of the nozzle. The lower temperature of combustion and the short time exposed to the flame cannot guarantee the crystallization of Cr
2O
3; thus, the amorphous structure is kept in the precursor.
After calcination at 900/1000 °C for 1.0 h, XRD patterns of the undoped Cr
2O
3 samples (S1-900, S1-1000) are in precise agreement with Cr
2O
3 (eskolaite, PDF reference pattern: 01-072-1207).
Figure 3 also illustrates the XRD patterns of the doped Cr
2O
3 samples (S3, S6, S12) calcined at 1000 °C, where all the prominent diffraction peaks were indexed as Cr
2O
3 (eskolaite, PDF reference pattern: 01-072-1207), indicating that the addition of small amounts of Ti/Co/Fe does not significantly change the phase composition of Cr
2O
3. Besides, an emphasized view on the patterns from 32.5 to 37.5 degrees displayed a distinct shift of the peak toward a lower angle after Ti/Co/Fe was doped. From the periodic table of elements, it is easy to know that the atomic radius of Ti, Cr, Fe, and Co decreases sequentially. The same shift trend of diffraction peaks for the S3, S6, and S12 samples is probably because 4 mol% Ti among the dopants plays a major role during the doping process. Based on the Bragg equation, it could be inferred that the interplanar space increased when Cr was substituted by Ti, resulting in the left shift of diffraction peaks.
3.2. Morphological Analysis
Figure 4 gives information about the morphology of both precursors and calcined powders.
Figure 4a shows that the precursor has an amorphous structure, which is well matched with the XRD result. A series of pores are distributed randomly, and a plate-like structure is more likely to be formed at this stage.
Figure 4b, c displays the morphology of S1 calcined at different temperatures. It can be found that well-defined sub-micron crystalline grains with slight aggregation are observed in both two samples, while the grain size of S1-1000 is larger than that of S1-900, indicating that a higher calcination temperature is beneficial to accelerating the growth of Cr
2O
3 grains. When compared with Cr
2O
3 samples prepared by a hydrogen reduction of chromite ore [
1] and thermal decomposition of chromium hydroxide [
12,
13], the serious aggregation and grain size distribution are efficiently optimized. From
Figure 4d, we can know that the morphology of S12 is similar to that of S1-1000, where the grain size is estimated to be about 0.4 μm, implying that Fe/Co/Ti doping has no negative effects during this process. Therefore, the solution combustion synthesis is approved to be a good approach to obtain submicron Cr
2O
3 crystals with better crystallinity and uniform distribution.
3.3. Visible Near-Infrared Diffuse Reflectance Spectra Analysis
The diffuse reflectance spectra of Cr
2O
3 samples calcined at various temperatures and samples with Ti/Co/Fe doping have been measured, and the analysis results are shown as below. According to
Figure 5, in the NIR range (780–2500 nm) of the pure Cr
2O
3 sample, the reflectance of S1-1000 is larger than that of S1-900, while there is no significant difference between the two spectra in the range from 400 to 780 nm. It can be concluded that calcination at 1000 °C creates a better condition for improving the NIR reflectance of Cr
2O
3 samples. Therefore, the calcination temperatures of the remaining samples (S2 to S12) are fixed at 1000 °C. Compared with S1, significantly larger NIR reflectance can be achieved in Ti-doped Cr
2O
3 samples S2 and S3, which show promising potential to be used as cool pigments. Meanwhile, a small rise of reflection peak at about 540 nm can be found after Ti doping. Although reflectance in the region of 800–1300 nm is improved, diffuse reflectance spectra of Ti-doped Cr
2O
3 is still different from that of green plants. Further work is needed to lower the reflection peaks at around 550 nm and simulate the valley in the region of 1300–1500 nm caused by moisture absorption.
As shown in
Figure 6, NIR reflection in the range of 1200–1700 nm for Co-containing Cr
2O
3 samples (S4) is significantly lower than that of S1-1000. This change can be attributed to the characteristic absorption of Co
2+ according to previous studies [
8]. Although a characteristic absorption of Co
2+ (1200–1700 nm) cannot perfectly match with the first valley (1300–1600 nm) on the spectra of green plants, the gap of reflectance between Cr
2O
3 pigments and green plants is significantly narrowed, which is beneficial to reducing their recognition. Indeed, Co has already been used to produce Cr
2O
3 pigments used for camouflage by the Shepherd color company [
14], as an absorption band of Co
2+ is still the best choice to simulate the valley (1300–1600 nm) of spectra of green plants until now. Besides, Co doping decreases the reflection peak at 540 nm from 30% (S1-1000) to 22% (S4). For samples S5 and S6, in which Co content is fixed at 4 at.%, the NIR reflection can still be improved after Ti doping when compared with S4, and the overall shapes of absorption bands are quite similar. Herein, it makes sense that Ti and Co can modulate the spectra of Cr
2O
3 when the amounts of additives are tiny. This is probably because Ti and Co are uniformly doped into the Cr
2O
3 crystal lattice rather than forming grain boundary segregation and secondary phases. Furthermore, NIR platforms emerge on the reflectance spectra of Co-containing Cr
2O
3 samples (S4, S5, S6), and the height of the platform can be adjusted flexibly by changing the amount of Ti.
The spectra of Fe and Ti co-doped samples (S8, S9, S10) are shown in
Figure 7. It is clear that the reflection peak at around 550 nm decreases gradually with the addition of Fe. Although Fe doping reduces the NIR reflectance in the range of 800–1300 nm, it is tolerable considering that the reflectance is still higher than that of most green plants. Therefore, Fe can be introduced into Ti and Co co-doped samples to make further efforts to lower the reflection peak at around 550 nm.
As
Figure 8 shows, the reflectance at around 550 nm of Ti, Co, and Fe co-doped samples is between 10% and 20%, so it matches with most green plants. Each spectrum of these samples owns an apparent NIR platform and an absorption band of Co
2+. Particularly, the reflection spectra (400–1600 nm) of S10, S11, and S12 are in good accordance with that of the green plants given in
Figure 1. In conclusion, our doping strategy is convinced to take effects while fabricating camouflage coatings.
Unfortunately, the diffuse spectra analysis shows that all the samples prepared fail to create the same characteristics of the waveband from 1600 to 2500 nm on the reflectance spectrum of green plants. In this research, no helpful element is found to simulate water absorption features at 1900 nm or 2500 nm. However, these pigments still have potential utilization in camouflage coatings for some objective reasons, which are as follows. First, only a small portion of lights in the region of 1600–2500 nm can reach the ground due to the radiation characteristics of solar and moisture absorption of the atmospheric layer [
15]. Moisture absorption occurs again before the lights reflected by green plants and camouflage coatings are detected by the visible light and near-infrared sensors. So, it becomes extremely difficult to make a distinction between natural green plants and artificial camouflage coatings when the analysis is conducted on a waveband from 1600 to 2500 nm. In this case, the waveband from 1600 to 2500 nm is rarely used as an operation band by most of the army’s detection equipment. Similar studies also feature discussions on this waveband [
16,
17]. All in all, the defects of Ti, Co, and Fe co-doped samples can be acceptable for practical applications. However, the authors still believe that some elements not mentioned in this article may be useful to create a perfect match, and many more studies still need to be done via the solution combustion synthesis method that we improved.
3.4. Chromatic Properties Analysis
Further study is conducted to evaluate the chromatic properties of some samples to make sure if they can be used to simulate the color of natural green plants. Ficus microcarpa, a kind of widely distributed evergreen tree of southern China as well as South and Southeast Asia, is chosen as a representative of natural green plants for its changeable green color. Diffuse reflectance spectra in the range of visible light (400–700 nm) of these samples as well as several Ficus microcarpa leaves (A to F) are shown in
Figure 9. Meanwhile, the similar color to every sample and leaf generated by the colorimeter are listed after the serial number.
Figure 9 shows that every Ficus microcarpa leaf tested has a reflection peak around 550 nm and the peak value ranges from 8% to about 25%. In contrast, S1 and S3 have reflection peaks around 540 nm, and their peak values are more than 30%. Obviously, the doping of Co lowers peak values, and then the spectra of S6 is close to the spectra of leaf A and B. As for Ti, Co, and Fe co-doped samples (S10, S11, S120), reflection peaks are getting even lower and the peak values range from nearly 16% to about 10%. Meanwhile, the positions of the reflection peaks shift from 540 nm (S1, S3, S6) to 550 nm (S10, S11) and 560 nm (S12) as the addition of Fe increases.
Figure 9 also shows a reflection peak around 410 nm, which is noticeable on the spectrum of pure Cr
2O
3 sample (S1), which is gradually smoothed out by the doping of Ti/Co/Fe. It can be found the spectra of sample S10 and leaf D are a good match in the range of 400–600 nm. Generally, samples S6, S10, S11, and S12 all have the potential to make green camouflage coatings considering that the average diffuse reflectance at 550 nm of most green plants is between 10% and 20%. However, it also needs to be pointed out that the doping of Ti/Co/Fe fails to simulate the valley around 680 nm on the reflectance spectrum of green plants, and further work still need to be done.
The L*, a*, and b* values of Cr
2O
3 samples as well as Ficus microcarpa leaves (A–F) are listed in
Table 2. For Ficus microcarpa leaves, the values of L* and b* decrease quickly, but the value of a* changes little as the color turns from yellowish-green to deep-green. It can be found the value of a* of the undoped Cr
2O
3 sample S1 is much lower than those of the Ficus microcarpa leaves tested in this study, and Co, Fe doping is effective to improve the value of a*. So, compared with the undoped Cr
2O
3 sample (S1), the Ti, Co co-doped sample (S6) and Ti, Co, Fe co-doped sample (S11) have values of a* that are much closer to those of the leaves. Besides, we also find that the values of a* of S11 and S12 are already too high, indicating that the addition of Fe needs to be cut down when further study is conducted. The color variation of doped Cr
2O
3 samples can also be found in the photographs of Cr
2O
3 tablets prepared for the color test shown in
Figure 10.