The Preparation of Amorphous Aluminum Oxide Modiﬁed g-C 3 N 4 to Improve Photocatalytic Performance in Contaminant Degradation Applications

: For the ﬁrst time, aluminum alloy was used as the main source to prepare aluminum oxide-modiﬁed carbon nitride with a melamine–cyanuric acid supramolecular complex. The introduction of amorphous aluminum oxide confers macroporosity to the skeletons of g-C 3 N 4 -AlO x . Its surface area increased to 75.5 m 2 g − 1 , about 1.5 times that of single g-C 3 N 4 . After modiﬁcation, the visible light response range was expanded, especially at 450~500 nm, while the band structure could be adjusted. Compared with g-C 3 N 4 , g-C 3 N 4 -AlO x has better photocatalytic performance, the adsorption rate for the dye rhodamine B (RhB) is about 2.1 times that of g-C 3 N 4 , and the RhB removal rate is 1.2 times that of g-C 3 N 4 .


Introduction
The rapid development of global industrialization and growth of population have led to increasing demand for energy. With non-renewable energy sources being increasingly consumed and rapidly depleted, the effective utilization and development of renewable energy sources has become extremely urgent and received wide attention [1][2][3].
As an important part of renewable energy, solar energy has competitive advantages, such as wide distribution, large energy reserves, cleanliness and safety [3,4]. In order to promote the usage of solar energy, research on photocatalytic materials with good stability and high activity becomes a key topic in material science [5][6][7][8][9]. Among all the materials, g-C 3 N 4 is eye-catching because of its low cost, low toxicity and wide light response range [10][11][12]. Since the first report about its use in water splitting in 2009 [13], related reports have continued. Compared with the classic photocatalyst TiO 2 , g-C 3 N 4 not only exhibits satisfying photocatalytic activity under ultraviolet light but also can be activated by part-visible light [5,[14][15][16]. This property makes g-C 3 N 4 appealing, as visible light accounts for about 43% of the solar spectrum [4,17,18]. To date, its application in photocatalysis includes water splitting [13,19], organic matter degradation [20,21], CO 2 reduction [22][23][24] and photocatalytic organic reactions [24,25].
However, g-C 3 N 4 still presents several disadvantages, such as its low specific surface area, low charge mobility, high excitation dissociation energy, high electron hole recombination rate and The surface area, porosity of g-C3N4, composite g-C3N4-AlOx and amorphous aluminum oxide were studied with the nitrogen adsorption-desorption method at 77 K. All samples were treated under vacuum at 423 K for 4 h prior to these measurements. It can be observed from Figure 2a that the N2 adsorption isotherms of g-C3N4, g-C3N4-AlOx and amorphous aluminum oxide increase slowly at low relative pressure and sharply at high relative pressure. These materials present type Ⅳ isotherms with hysteresis loops of type 3 (H3), revealing the samples are all mesoporous materials. Based on this information, the specific surface areas were calculated with the BET model. The BET surfaces area of g-C3N4, g-C3N4-AlOx and amorphous aluminum oxide were evaluated to be 51.1, 75.5 and 189.2 m 2 g −1 , respectively. The surface area of g-C3N4-AlOx was higher than that of g-C3N4, which would be helpful in catalysis. Its surface area is between that of g-C3N4 and that of amorphous aluminum oxide, suggesting the successful combination of the two components in the new material. The pore size distributions of the samples were calculated with the Nonlocal Density Functional Theory (NLDFT) method, as shown in Figure 2b. Compared with that of g-C3N4 and amorphous aluminum oxide, the porosity of g-C3N4-AlOx includes not only the mesoporous part but also macroporosity (60-130 nm). The appearance of new porosity displays a new structure in g-C3N4-AlOx, which means that the CM complex and AlOxHy colloid are not simply mixed together in the g-C3N4-AlOx but they form new material. This proves that the new preparation strategy is valuable for achieving new materials and expanding the material scope. Additionally, the material in this method very likely possesses A heterojunction structure, which is usually preferred in semiconductor catalysis.  The surface area, porosity of g-C 3 N 4 , composite g-C 3 N 4 -AlO x and amorphous aluminum oxide were studied with the nitrogen adsorption-desorption method at 77 K. All samples were treated under vacuum at 423 K for 4 h prior to these measurements. It can be observed from Figure 2a that the N 2 adsorption isotherms of g-C 3 N 4 , g-C 3 N 4 -AlO x and amorphous aluminum oxide increase slowly at low relative pressure and sharply at high relative pressure. These materials present type IV isotherms with hysteresis loops of type 3 (H3), revealing the samples are all mesoporous materials. Based on this information, the specific surface areas were calculated with the BET model. The BET surfaces area of g-C 3 N 4 , g-C 3 N 4 -AlO x and amorphous aluminum oxide were evaluated to be 51.1, 75.5 and 189.2 m 2 g −1 , respectively. The surface area of g-C 3 N 4 -AlO x was higher than that of g-C 3 N 4 , which would be helpful in catalysis. Its surface area is between that of g-C 3 N 4 and that of amorphous aluminum oxide, suggesting the successful combination of the two components in the new material. The pore size distributions of the samples were calculated with the Nonlocal Density Functional Theory (NLDFT) method, as shown in Figure 2b. Compared with that of g-C 3 N 4 and amorphous aluminum oxide, the porosity of g-C 3 N 4 -AlO x includes not only the mesoporous part but also macroporosity (60-130 nm). The appearance of new porosity displays a new structure in g-C 3 N 4 -AlO x , which means that the CM complex and AlO x H y colloid are not simply mixed together in the g-C 3 N 4 -AlO x but they form new material. This proves that the new preparation strategy is valuable for achieving new materials and expanding the material scope. Additionally, the material in this method very likely possesses A heterojunction structure, which is usually preferred in semiconductor catalysis. The surface area, porosity of g-C3N4, composite g-C3N4-AlOx and amorphous aluminum oxide were studied with the nitrogen adsorption-desorption method at 77 K. All samples were treated under vacuum at 423 K for 4 h prior to these measurements. It can be observed from Figure 2a that the N2 adsorption isotherms of g-C3N4, g-C3N4-AlOx and amorphous aluminum oxide increase slowly at low relative pressure and sharply at high relative pressure. These materials present type Ⅳ isotherms with hysteresis loops of type 3 (H3), revealing the samples are all mesoporous materials. Based on this information, the specific surface areas were calculated with the BET model. The BET surfaces area of g-C3N4, g-C3N4-AlOx and amorphous aluminum oxide were evaluated to be 51.1, 75.5 and 189.2 m 2 g −1 , respectively. The surface area of g-C3N4-AlOx was higher than that of g-C3N4, which would be helpful in catalysis. Its surface area is between that of g-C3N4 and that of amorphous aluminum oxide, suggesting the successful combination of the two components in the new material. The pore size distributions of the samples were calculated with the Nonlocal Density Functional Theory (NLDFT) method, as shown in Figure 2b. Compared with that of g-C3N4 and amorphous aluminum oxide, the porosity of g-C3N4-AlOx includes not only the mesoporous part but also macroporosity (60-130 nm). The appearance of new porosity displays a new structure in g-C3N4-AlOx, which means that the CM complex and AlOxHy colloid are not simply mixed together in the g-C3N4-AlOx but they form new material. This proves that the new preparation strategy is valuable for achieving new materials and expanding the material scope. Additionally, the material in this method very likely possesses A heterojunction structure, which is usually preferred in semiconductor catalysis.  The morphology of g-C 3 N 4 -AlO x was investigated with SEM as shown in Figure 3. The SEM photos show that single g-C 3 N 4 displays tubular structures, with a diameter of about 500 nm, and a relatively smooth surface (Figure 3a). For amorphous aluminum oxide-decorated g-C 3 N 4 , the morphology changes greatly into an irregular "coralloid" structure with a rough surface (Figure 3b,c). Importantly, a large number of pores appear in the "coralloid" structure, whose existence is consistent with the results of the N 2 adsorption-desorption measurements discussed above. The morphology variation intuitively proves the change in material structure and influence of Al colloids. In order to reveal the element composition, g-C 3 N 4 -AlO x was analyzed with energy dispersive X-ray energy spectroscopy (EDS). The EDS mapping images (Figure 3d) suggest that the four main elements Al, C, N and O are extensively distributed in g-C 3 N 4 -AlO x . This even existence of amorphous aluminum oxide is probably the reason for the increased in surface area, in that they act as template agents during the skeleton formation process. It also shows that the method of mixing AlO x H y colloid with supramolecular complex CM is so effective that the amorphous alumina keeps close contact with g-C 3 N 4 . The close combination of two materials provides the possibility of a heterojunction structure, which may be good for further catalysis applications. The selected area in Figure 3b was analyzed, and the corresponding EDX results are given in Table 1. The EDX results reveal that the amorphous aluminum oxide makes up about 5.10 wt% of g-C 3 N 4 -AlO x (here, X is defaulted to 1.5).
Catalysts 2020, 10, x FOR PEER REVIEW 4 of 11 The morphology of g-C3N4-AlOx was investigated with SEM as shown in Figure 3. The SEM photos show that single g-C3N4 displays tubular structures, with a diameter of about 500 nm, and a relatively smooth surface ( Figure 3a). For amorphous aluminum oxide-decorated g-C3N4, the morphology changes greatly into an irregular "coralloid" structure with a rough surface (Figure 3b,c). Importantly, a large number of pores appear in the "coralloid" structure, whose existence is consistent with the results of the N2 adsorption-desorption measurements discussed above. The morphology variation intuitively proves the change in material structure and influence of Al colloids. In order to reveal the element composition, g-C3N4-AlOx was analyzed with energy dispersive X-ray energy spectroscopy (EDS). The EDS mapping images (Figure 3d) suggest that the four main elements Al, C, N and O are extensively distributed in g-C3N4-AlOx. This even existence of amorphous aluminum oxide is probably the reason for the increased in surface area, in that they act as template agents during the skeleton formation process. It also shows that the method of mixing AlOxHy colloid with supramolecular complex CM is so effective that the amorphous alumina keeps close contact with g-C3N4. The close combination of two materials provides the possibility of a heterojunction structure, which may be good for further catalysis applications. The selected area in Figure 3b was analyzed, and the corresponding EDX results are given in Table 1. The EDX results reveal that the amorphous aluminum oxide makes up about 5.10 wt% of g-C3N4-AlOx (here, X is defaulted to 1.5).   The samples were characterized by Fourier transform infrared spectroscopy (FTIR), as shown in Figure 4. g-C 3 N 4 -AlO x possesses the peaks most characteristic of g-C 3 N 4 in the FTIR spectrum. This shows that its main structure is similar to that of g-C 3 N 4 , which is a connected hepazine ring. The peak band around 3000-3450 cm −1 is attributed to the stretching vibration of N-H and hydroxyl (O-H) groups, and is derived from the process of supramolecular dehydration while forming g-C 3 N 4 . Compared with the peak band in the box of g-C 3 N 4 , the peak band of g-C 3 N 4 -AlO x near 3000-3450 cm −1 is broadened and the peak strength is slightly increased, indicating that there were more hydroxyl groups on the surface of g-C 3 N 4 -AlO x . This is possibly because AlO x H y colloid joins the supramolecular formation and changes the surface structure of the original g-C 3 N 4 . It is worth noting that g-C 3 N 4 -AlO x has an absorption peak near 2200 cm −1 , which is not shown in the graphs of single g-C 3 N 4 and amorphous aluminum oxide. This absorption peak is attributed to the peak characteristic of C≡N. The appearance of this new peak again illustrates that AlO x H y colloid affects the supramolecular formation and, later, the g-C 3 N 4 -AlO x synthesis. In the presence of AlO x H y , some triazine rings in the supramolecular formation do not condense into heptazine rings completely but result in C≡N on the surface of the material. Overall, this is consistent with the previous analysis showing that g-C 3 N 4 -AlO x is a new material from g-C 3 N 4 surface decoration.
Catalysts 2020, 10, x FOR PEER REVIEW 5 of 11 The samples were characterized by Fourier transform infrared spectroscopy (FTIR), as shown in Figure 4. g-C3N4-AlOx possesses the peaks most characteristic of g-C3N4 in the FTIR spectrum. This shows that its main structure is similar to that of g-C3N4, which is a connected hepazine ring. The peak band around 3000-3450 cm −1 is attributed to the stretching vibration of N-H and hydroxyl (O-H) groups, and is derived from the process of supramolecular dehydration while forming g-C3N4.
Compared with the peak band in the box of g-C3N4, the peak band of g-C3N4-AlOx near 3000-3450 cm −1 is broadened and the peak strength is slightly increased, indicating that there were more hydroxyl groups on the surface of g-C3N4-AlOx. This is possibly because AlOxHy colloid joins the supramolecular formation and changes the surface structure of the original g-C3N4. It is worth noting that g-C3N4-AlOx has an absorption peak near 2200 cm −1 , which is not shown in the graphs of single g-C3N4 and amorphous aluminum oxide. This absorption peak is attributed to the peak characteristic of C≡N. The appearance of this new peak again illustrates that AlOxHy colloid affects the supramolecular formation and, later, the g-C3N4-AlOx synthesis. In the presence of AlOxHy, some triazine rings in the supramolecular formation do not condense into heptazine rings completely but result in C≡N on the surface of the material. Overall, this is consistent with the previous analysis showing that g-C3N4-AlOx is a new material from g-C3N4 surface decoration. The details of g-C3N4-AlOx regarding the chemical composition and local connectivity motifs were revealed by X-ray photoelectron spectroscopy (XPS, Figure 5). In the C1s curve of single g-C3N4, the main signal peak appears at 288.3 eV, which is attributed to sp 2 C in triazine rings (N − C = N2). (The other peak around 284.8 eV is recognized as adventitious carbon from impurities.) After the decoration of AlOx colloids, the C1s peak in g-C3N4-AlOx becomes shifted to a higher binding energy area (288.4 eV), indicating its chemical environment has changed and there is close interaction between amorphous aluminum oxide and g-C3N4 (Figure 5a). The N1s spectrum of g-C3N4-AlOx was similar to that of g-C3N4, which shows that the main framework of g-C3N4-AlOx is still g-C3N4 ( Figure  5b). Compared with those for amorphous aluminum oxide, the positions of the O1s peak and Al2p peak also exhibit variation in g-C3N4-AlOx. In amorphous aluminum oxide, the positions of O1s and Al2p locate at 531.3 and 74.4 eV, which are shifted to 532.1 and 74.7 eV in g-C3N4-AlOx (Figure 5b,c). The details of g-C 3 N 4 -AlO x regarding the chemical composition and local connectivity motifs were revealed by X-ray photoelectron spectroscopy (XPS, Figure 5). In the C1s curve of single g-C 3 N 4 , the main signal peak appears at 288.3 eV, which is attributed to sp 2 C in triazine rings (N-C=N 2 ). (The other peak around 284.8 eV is recognized as adventitious carbon from impurities.) After the decoration of AlO x colloids, the C1s peak in g-C 3 N 4 -AlO x becomes shifted to a higher binding energy area (288.4 eV), indicating its chemical environment has changed and there is close interaction between amorphous aluminum oxide and g-C 3 N 4 (Figure 5a). The N1s spectrum of g-C 3 N 4 -AlO x was similar to that of g-C 3 N 4 , which shows that the main framework of g-C 3 N 4 -AlO x is still g-C 3 N 4 (Figure 5b). Compared with those for amorphous aluminum oxide, the positions of the O1s peak and Al2p peak also exhibit variation in g-C 3 N 4 -AlO x . In amorphous aluminum oxide, the positions of O1s and Al2p locate at 531.3 and 74.4 eV, which are shifted to 532.1 and 74.7 eV in g-C 3 N 4 -AlO x (Figure 5b,c). This change illustrates strong chemical interaction between the two components, which suggests the successful preparation of g-C 3 N 4 -AlO x with an interactive interface.
Catalysts 2020, 10, x FOR PEER REVIEW 6 of 11 This change illustrates strong chemical interaction between the two components, which suggests the successful preparation of g-C3N4-AlOx with an interactive interface. The g-C3N4, g-C3N4-AlOx and amorphous aluminum oxide were studied with solid UV-vis absorption spectroscopy. As shown in Figure 6a, pristine g-C3N4 exhibits a light absorption range from ultraviolet light to visible light, and its absorption edge wavelength is at around 450 nm. After decoration with AlOxHy colloid, the absorption edge of g-C3N4-AlOx was redshifted to around 500 nm, which is an improvement for wider solar light spectrum absorption. The redshift of adsorption usually comes from the enlargement of conjugations. Here, the variation of adsorption may imply that amorphous aluminum oxide takes part in the conjugations in the new system. In the g-C3N4-AlOx, the conjugation may contain p-π conjugation, instead of only π-π conjugation from g-C3N4. As amorphous aluminum oxide only responds to ultraviolet light, the better light response performance of g-C3N4-AlOx is indeed due to the decoration effect of aluminum oxide on g-C3N4, instead of simple mixing. The band gap of those materials can be calculated with the Tauc plot method. The bandgap of g-C3N4-AlOx is calculated to be 2.51 eV, which is a little narrower than that of g-C3N4, at around 2.66 eV. One possible reason for this result is that the addition of AlOxHy colloid leads to lower crystallinity ( Figure 1b) and more crystal defects in g-C3N4-AlOx. In addition, in Figure 6, it is also clear that g-C3N4-AlOx has two components in its absorbance edge, which implies that the sample is not monodisperse in size. In this protocol, amorphous aluminum oxide was successfully doped but probably not evenly dispersed. To further determine the valence band level, g-C3N4-AlOx samples were characterized by ultraviolet photoelectron spectroscopy (UPS, Figure 6c). The maximum values of the g-C3N4 and g-C3N4-AlOx valence bands were −5.96 and −5.36 eV vs. vacuum or 1.46 and 0.86 eV vs. NHE (Normal Hydrogen Electrode); E (eV) = −4.5 − ENHE (V) [37]. Considering the width of the The g-C 3 N 4 , g-C 3 N 4 -AlO x and amorphous aluminum oxide were studied with solid UV-vis absorption spectroscopy. As shown in Figure 6a, pristine g-C 3 N 4 exhibits a light absorption range from ultraviolet light to visible light, and its absorption edge wavelength is at around 450 nm. After decoration with AlO x H y colloid, the absorption edge of g-C 3 N 4 -AlO x was redshifted to around 500 nm, which is an improvement for wider solar light spectrum absorption. The redshift of adsorption usually comes from the enlargement of conjugations. Here, the variation of adsorption may imply that amorphous aluminum oxide takes part in the conjugations in the new system. In the g-C 3 N 4 -AlO x , the conjugation may contain p-π conjugation, instead of only π-π conjugation from g-C 3 N 4 . As amorphous aluminum oxide only responds to ultraviolet light, the better light response performance of g-C 3 N 4 -AlO x is indeed due to the decoration effect of aluminum oxide on g-C 3 N 4 , instead of simple mixing. The band gap of those materials can be calculated with the Tauc plot method. The bandgap of g-C 3 N 4 -AlO x is calculated to be 2.51 eV, which is a little narrower than that of g-C 3 N 4 , at around 2.66 eV. One possible reason for this result is that the addition of AlO x H y colloid leads to lower crystallinity (Figure 1b) and more crystal defects in g-C 3 N 4 -AlO x . In addition, in Figure 6, it is also clear that g-C 3 N 4 -AlO x has two components in its absorbance edge, which implies that the sample is not monodisperse in size. In this protocol, amorphous aluminum oxide was successfully doped but probably not evenly dispersed. To further determine the valence band level, g-C 3 N 4 -AlO x samples were characterized by ultraviolet photoelectron spectroscopy (UPS, Figure 6c). The maximum values of the g-C 3 N 4 and g-C 3 N 4 -AlO x valence bands were −5.96 and −5.36 eV vs. vacuum or 1.46 and 0.86 eV vs. NHE (Normal Hydrogen Electrode); E (eV) = −4.5 − E NHE (V) [37]. Considering the width of the forbidden band, the minimum values of the corresponding conduction band are calculated to be −1.20 and −1.65 V vs. NHE, respectively. This result proves this method is effective in band structure adjustment, which provides g-C 3 N 4 -AlO x with completely different possibilities in photocatalysis.

Photocatalytic Activity
The photocatalytic activity of g-C3N4-AlOx was investigated with a rhodamine B (RhB) degradation experiment under visible light irradiation. In a typical degradation test, 5.0 mg of powder was added as a catalyst in 10.0 mL of RhB aqueous solution (10 mgL −1 ). After 30 mins of stirring in the dark, the whole system was subjected to white light illumination (380-840 nm). During the whole degradation process, the remaining RhB concentration was tested every 20 min by obtaining UV-vis spectra at the wavelength of 554 nm as shown in Figure 7. Amorphous aluminum oxide can absorb a small amount of dyes in the dark but can hardly degrade dyes in visible light. g-C3N4-AlOx displays better photocatalytic performance than g-C3N4, successfully eliminating 92.77% of the RhB within 80 min (Figure 7a). For better intuitiveness, the RhB degradation performance of g-C3N4-AlOx and single g-C3N4 is compared in terms of photodegradation first order profiles in Figure  7b. Those curves are in good agreement with the pseudo-first-order correlation of kinetics: ln(C0/Ct) = kt + ln(C0/C1),

Photocatalytic Activity
The photocatalytic activity of g-C 3 N 4 -AlO x was investigated with a rhodamine B (RhB) degradation experiment under visible light irradiation. In a typical degradation test, 5.0 mg of powder was added as a catalyst in 10.0 mL of RhB aqueous solution (10 mgL −1 ). After 30 mins of stirring in the dark, the whole system was subjected to white light illumination (380-840 nm). During the whole degradation process, the remaining RhB concentration was tested every 20 min by obtaining UV-vis spectra at the wavelength of 554 nm as shown in Figure 7. Amorphous aluminum oxide can absorb a small amount of dyes in the dark but can hardly degrade dyes in visible light. g-C 3 N 4 -AlO x displays better photocatalytic performance than g-C 3 N 4 , successfully eliminating 92.77% of the RhB within 80 min (Figure 7a). For better intuitiveness, the RhB degradation performance of g-C 3 N 4 -AlO x and single g-C 3 N 4 is compared in terms of photodegradation first order profiles in Figure 7b. Those curves are in good agreement with the pseudo-first-order correlation of kinetics: Catalysts 2020, 10, 1036 8 of 11 photocatalytic performance. Amorphous aluminum oxide provides more sites for g-C3N4 electrons to combine, promotes electron-hole separation and improves photocatalytic activity. Besides, g-C3N4-AlOx possesses a large number of macroporous structures, in which incident light is more prone to multiple scattering, increasing the light utilization rate of the material. The doping of amorphous alumina confers a smaller grain size and larger porosity to the final composite material.

Materials
Al-Ga-In ternary alloy (99% aluminum), melamine (C3N3(NH2)3), cyanic acid (C3H3N3O3), methanol (CH3OH) and AlCl3·6H2O were all purchased from Aladdin (Shanghai, China) and used without further purification. The calculated pseudo-first-order rate constant K and correlation coefficient R are shown in Table 2. According to the value of the intercept ln(C 0 /C 1 ) on the vertical axis of the dynamic curve, it can be found that the adsorption rate of g-C 3 N 4 -AlO x for RhB is about 2.1 times higher than that of g-C 3 N 4 for RhB. This reveals that the amorphous aluminum oxide indeed helps with the promotion of photocatalytic performance. Amorphous aluminum oxide provides more sites for g-C 3 N 4 electrons to combine, promotes electron-hole separation and improves photocatalytic activity. Besides, g-C 3 N 4 -AlO x possesses a large number of macroporous structures, in which incident light is more prone to multiple scattering, increasing the light utilization rate of the material. The doping of amorphous alumina confers a smaller grain size and larger porosity to the final composite material. Table 2. Kinetic fitting equation, quasi-first-order rate constant k and correlation coefficient in the process of the degradation of RhB by each sample prepared.

Preparation of Photocatalyst
First, melamine (1.5 g) and cyanic acid (1.5 g) were mixed in deionized water for 6 h at room temperature to form a supramolecular complex, cyanuric acid-melamine (CM). Then, the CM was ultrasonically dispersed in 100 mL of methanol for 30 min to obtain Part 1. AlCl 3 ·6H 2 O was used to prepare an AlCl 3 solution (0.7 mol L −1 , 100 mL) and heated to 90 • C. At the same temperature (90 • C), 1 g of Al-Ga-In ternary alloy was added into the solution and stirred for 3 h until the reaction was completed. It is cooled to room temperature to form a colloid, and the liquid metal (Ga-In alloy liquid metal ball) present at the bottom was sucked out. Then, after removing the unreacted metal, 100 mL of methanol was added to the colloid and stirred for 2 h to obtain Part 2. The two parts activated by methanol were mixed and stirred for 2 h. Then, they were dried in a vacuum drying oven (80 • C). After grinding, the dried samples were heated to 600 • C in an argon atmosphere in a tube furnace for 4.5 h. The final product g-C 3 N 4 -AlO x was obtained. In addition, the part 1 and the part 2 of the different substance quantities were synthesized and tested, and the data were presented in the supplementary material.

Conclusions
In this article, a new preparation strategy was proposed for amorphous aluminum oxide-modified carbon nitride with aluminum alloy as an aluminum source. The introduction of amorphous aluminum oxide confers macroporosity to the skeletons of g-C 3 N 4 -AlO x . The material prepared by this method shows a smaller grain size, larger specific surface area, wider light response range, better photocatalytic performance and different band structure in comparison with g-C 3 N 4 . These changes help with the improvement of its photocatalytic performance.
Supplementary Materials: The following are available online at http://www.mdpi.com/2073-4344/10/9/1036/s1. Figure S1: XRD pattern of each synthetic sample: (a,b) before roasting and (c,d) after roasting; Figure S2: (a,c) N 2 adsorption-desorption isotherm and (b,d) pore size distribution of samples; Figure S3: TEM images of g-C 3 N 4 -AlO x ; Figure S4: FTIR absorption spectra of the samples; Figure S5: The XPS survey spectra of the samples; Figure S6: UV-vis absorption spectra of the samples; Figure

Conflicts of Interest:
The authors declare no conflict of interest.