Benzyl alcohol, nitrates with crystalline waters, and other reagents were analytical grade. Gas chromatography (GC) analysis was performed on SHIMAZDU GC-2014 equipped with a HP-5 column (30 m × 0.32 mm × 0.25 um) and a flame ionization detector.

Typically, 3 mmol of benzyl alcohol, 3 mmol of naphthalene as an internal standard, and 15 mL of 1,4-dioxane as solvent were added to a 25 mL three-necked flask. Then the three-necked flask provided with a reflux condenser was evacuated using an aspirator and followed by the attachment of a nitrogen balloon. Subsequently, the solution was heated to the desired temperature. When the temperature of solution reached the set point, 2 mmol of ferric nitrate (Fe(NO₃)₃·9H₂O) was added into it. Samples were taken at appropriate intervals through a silicon septum using a hypodermic needle and were filtered with a membrane filter (PVDF) with 13 mm × 0.22 μm pore prior to GC analysis.

The analysis of benzyl alcohol and oxidation products was carried out on Aglient 7980 series with a HP-5 column and a flame ionization detector. The condition of GC for the HP-5 capillary column (30 m, DF = 0.25 mm, 0.25 mm i.d.), and temperature program was carried out (initial temperature = 100°C, 3 min; final temperature = 250°C, heating rate = 10°C min⁻¹, temperature of injector = 280°C, temperature of detector = 280°C). The quantitative results of products were based on the internal standard method, using naphthalene as an internal standard. The typical analytic procedure was as following: (1) 0.4 ml sample was taken from reaction solutions and was then filtered with a membrane filter (PVDF) with 13 mm × 0.22 μm; (2) then, the sample would be diluted 10-fold before GC analysis; (3) when the GC temperature reached the set points, the diluted solution was injected into this equipment to analyze products according to the different retention times and response peak area. The typical results of GC measurements and internal standard working curve were shown in Figure S1. The results reported as conversion and selectivity are expressed in mol%, based on the total benzyl alcohol intake. The calculation of the conversion and selectivity was as follows:

Figures 1A,B showed the oxidation results of benzyl alcohol to benzaldehyde in the presence of ferric nitrate and nitric acid. In N₂ condition, the conversion of benzyl alcohol catalyzed by ferric nitrate was 94.9% after 6 h. To ensure reproducibility, the experimental error-based three parallel experiments was shown in Figure S2 and the carbon balance during experiments was also evaluated (Table S1). The results showed that performance of Fe(NO₃)₃ had a good testing repeatability and was highly efficient. Interestingly, the benzyl alcohol conversion catalyzed by Fe(NO₃)₃ could be as high as 96.84% with 94.5% selectivity when replacing N₂ with He. Compared with HNO₃, ferric nitrate exhibited an excellent activity in the benzyl alcohol oxidation. The conversion of benzyl alcohol catalyzed by ferric nitrate was 46.2%—higher than that by nitric acid. Similarly, under aerobic conditions (O₂), the conversion of benzyl alcohol catalyzed by ferric nitrate was about 13% higher than that by nitric acid with similar selectivity. The results indicated the oxidation performance of ferric nitrate was markedly better than that of nitric acid, no matter under anaerobic or aerobic conditions.

Compared with other typical catalysts, the catalytic activity of Fe(NO₃)₃ was also outstanding. The conversion of benzyl alcohol catalyzed by Fe(NO₃)₃ with any oxidant (96.84%) was close to that by CNT-HNO₃ (Luo et al., 2012) (96.2%), as shown in Table 1 (Entry 2). Besides, the performance of Fe(NO₃)₃ surpassed the typical metal-based catalyst, such as Au/γ-Al₂O₃ (73.4%) (Ndolomingo and Meijboom, 2017), Fe/MCM41 (55%) (Cánepa et al., 2017), or Fe-N-C (78.0%) (Xie et al., 2017). Moreover, Fe(NO₃)₃ also exhibited comparable activity with transition-metal oxides, such as Co₃O₄/RGO-_N (93.9%) (Nie et al., 2013), MnO_x (72.7%) (Jing et al., 2007). The excellent catalytic activity of Fe(NO₃)₃ may be attributed to the combined action of Fe³⁺ and NO3-. On the one hand, NO3- could produce NO₂, and then NO₂ with H₂O was converted to HNO₂, which had been proved as a pivotal role in benzyl alcohol oxidation (Aellig et al., 2012; Luo et al., 2012). On the other hand, the valence change of Fe could catalyze benzyl alcohol oxidation according to the literatures (Miao et al., 2016; Hu et al., 2018). Coincidentally, the transformation between Fe³⁺ and Fe²⁺ was demonstrated by the K₃[Fe(CN)₆] solution. As shown in Figure S3, the Prussian blue precipitate appeared in the experimental process due to the existence of Fe²⁺. Therefore, Fe(NO₃)₃ exhibited the excellent oxidation activity via Fe³⁺ initiating a series of electron and proton transfer. It was noted that the anaerobic condition was beneficial to improve the oxidation performance of ferric nitric. The reason is discussed in detail in mechanism Part 3.4.

Subsequently, the catalytic performance of other nitrates was also studied. The benzyl alcohol oxidation catalyzed by other nitrates including was investigated, as shown in Figure 2. The result proves other nitrates are also capable for converting the benzyl alcohol to benzaldehyde. In Al(NO₃)₃ system, the conversion of benzyl alcohol and the selectivity to benzaldehyde was 88.1% and 80% after 6 h, lower than those in Fe(NO₃)₃ system. Similarly, Cu(NO₃)₂ could also oxidize the benzyl alcohol with the conversion of 82.3%. While the selectivity to benzadehyde was only 70%, and the other 30% was benzoic acid from the excessive oxidation. The catalytic performance of Co(NO₃)₂ significantly decreased. While Mg(NO₃)₂ and Zn(NO₃)₂ had almost no catalytic activity. The order of activity of different nitrates is Fe(NO₃)₃ > Al(NO₃)₃ > Cu(NO₃)₂ > Co(NO₃)₂ > Mg(NO₃)₂ ≈ Zn(NO₃)₂. Yuvaraj et al. (2003) tested the decomposition temperature of these nitrates by TG/DAT. And they found Fe(NO₃)₃ and Al(NO₃)₃ had the lowest the decomposition temperature (130°C), followed by Cu(NO₃)₂ (227°C) and Co(NO₃)₂ (247°C), and Zn(NO₃)₂ (367°C) was the highest. Their decomposition products were metal oxide, NO₂ and O₂. NO₂ and H₂O together would convert into HNO₂ which could attack the benzyl alcohol, finally producing benzaldehyde. Hence, the nitrates with lower decomposition temperature could easily produce NO₂ and oxidize more benzyl alcohol. As expected, the rank of conversion of benzyl alcohol in different nitrate system was related to the decomposition temperature of these nitrates.

Fe(NO₃)₃ is an excellent catalyst among several nitrates; the effect of reaction condition and solvent on the performance of Fe(NO₃)₃ was investigated. As the results shown in Figure 3A, the higher temperature would markedly enhance the yields of benzaldehyde. Since high temperature would increase the risk of the over oxidation from benzaldehyde to benzoic acid, the optimum temperature was 80°C. Similarly, 2 mmol Fe(NO₃)₃ could selectively oxidize the benzyl alcohol to benzaldehyde with 91.5% yields shown in Figure 3B. Besides, the solvent had an important effect on the activity of Fe(NO₃)₃. The results in Figure 3C show the strong polar solvent could bring a mutual solution containing the aqueous (HNO₂) and organic phase (benzyl alcohol), which would be more favorable to form of benzyl nitrite, resulting in high conversion and selectivity.

Subsequently, to demonstrate the general applicability of Fe(NO₃)₃, selective oxidation of substituted benzyl alcohols with different functional groups was investigated. The results presented in Table 2 showed all these primary benzylic alcohols could be converted to corresponding aldehydes. The oxidation results were somewhat related to the substituent groups on the phenyl ring. The alcohols with electron withdrawing groups (–NO₂, –Cl) gained lower yields of products than those with electron donating groups (–MeO, –OH). The effect rule of substituent groups in Fe(NO₃)₃ system is consistent with that in HNO₃ system (Joshi et al., 2005). The formation of benzyl nitrite, a vital intermediate product, can be regarded as the electrophilic substitution reaction of benzyl alcohol. Thus, the electron-donating substituents would enhance the yield of product by increasing the electron density on the benzyl ring and vice versa. The results proved that the electron density on the aromatic ring played a critical role in the oxidation of benzylic alcohol.

Combined with the experimental results and literature reports, the following catalytic mechanism hypothesis was preliminarily put. Both Fe³⁺ and NO3- may have an important effect on the oxidation process. Fe³⁺ provides catalytic function by the electron transferring, while the NO_x produced from the experimental process may have certain oxidation performance. To verify above possible speculations and analyze the catalytic mechanism, a series of experiments were designed, and the results were shown as follows.

To illustrate the role of Fe³⁺ and NO3-, Zn(NO₃)₂ or FeCl₃ was individually added to the reaction mixtures. The result showed benzyl alcohol conversion in both systems was very low (Figure 4), indicating that only Fe³⁺ or NO3- had poor oxidation activity. While in Zn(NO₃)₂ + FeCl₃ system, the conversion of benzyl alcohol was as high as 95.8% with 93.6% selectivity, which was similar to that in Fe(NO₃)₃ system with the same amount of Fe³⁺ and NO3-. The results proved the Fe³⁺ and NO3- together could oxidize the benzyl alcohol. To illuminate how Fe³⁺ and NO3- together catalyze the benzyl alcohol, introducing urea as a known HNO₂ scavenger into Fe(NO₃)₃ system. The results in Figure 4 showed the catalytic activity of Fe(NO₃)₃ was almost prevented. This phenomenon proved that the oxidation process of Fe(NO₃)₃ was mainly achieved by HNO₂ attacking benzyl alcohol. The intermediate benzyl nitrite detected by GC-MS in our experiments further verified the reaction between benzyl alcohol and HNO₂. Besides, when butylated hydroxytoluene (BHT) as a radical trapping agent was added into the Fe(NO₃)₃ system, the conversion still kept a high level of 86.3%, proving that the benzyl alcohol oxidation catalyzed by Fe(NO₃)₃ is not a radical-involved reaction. Interestingly, the conversion would increase obviously, and the selectivity also kept at a high value (95%) when Fe³⁺ instead of H⁺ with the same mole of NO3- as shown in Figure 4. Furthermore, the addition of Fe³⁺ into HNO₃ system also significantly enhanced the conversion of benzyl alcohol compared with the HNO₃ system. From these results, we speculated Fe³⁺ had a special catalytic function in the process of benzyl alcohol oxidation.

A series of experiments were conducted to further illuminate the catalysis of Fe³⁺. As the results shown in Figure 5A, the reaction rate increased as the amount of Fe³⁺ increased in FeCl₃-HNO₃ system. Meanwhile, the results shown in Figure 5B proved that the variation of selectivity was little. To quantify the effect of Fe³⁺, the turnover frequency (TOF) was calculated based on Fe content at the conversion lower than 25.0%, as shown in Figure 5C. The results proved the Fe³⁺ had a remarkable effect on the oxidation reaction. The TOF of Fe³⁺ in the FeCl₃-HNO₃ system further enlarged compared to that in the Fe(NO₃)₃ system. As shown in Figure 5D, the TOF of Fe³⁺ in the FeCl₃-HNO₃ system was almost unchanged (8.1 h⁻¹), even if the amount of Fe³⁺ increased from 0.5 to 1 mmol.

Furthermore, the change of Fe in benzyl alcohol oxidation was analyzed by experiments. The transformation between Fe³⁺ and Fe²⁺ was detected in situ by 1 mol/L KMnO₄ solution in the oxidation process. For comparison, the reaction solvent (1,4-dioxane) was added to the KMnO₄ solution, the solution was black-purple as shown in Figure 6. The color of KMnO₄ solution changed from black-purple to yellow when KMnO₄ solution was added into reaction solution at 2 h. This phenomenon showed the KMnO₄ was reduced by Fe²⁺, and the Fe²⁺ existed in the reaction process. However, the KMnO₄ solution changed from yellow to brown when it was added into the reaction solution at the reaction time of 6 h. The phenomenon illustrated that there were less Fe²⁺ in the reaction solution. Because the oxidation reaction of benzyl alcohol almost completed, most of Fe³⁺ did not convert anymore and Fe²⁺ would further oxidized to Fe³⁺ by the O₂ from the decomposition of Fe(NO₃)₃. So, the cyclic conversion of Fe³⁺ and Fe²⁺ really occurred in benzyl alcohol oxidation. According to the reported results, the valence transformation of metal iron such as Mn^III/Mn^II (Yang et al., 2014; Fei et al., 2017; Gurrala et al., 2018), Co^III/Co^II (Zhou et al., 2015; Cordoba et al., 2017; Li et al., 2017), Cr^III/ Cr^II (Thao et al., 2018) could catalyze the oxidation of alcohols. Especially, the transformation of Fe³⁺/Fe²⁺ possessed high catalytic activity by initiating a series of electron transfer (Hu et al., 2016, 2018; Miao et al., 2016). Combined with above results, Fe³⁺ could be regard as a catalyst and the electron transfer between of Fe³⁺ and Fe²⁺ would catalyze the oxidation process.

Finally, the function of NO_x was investigated by a series of auxiliary experiments. As we know, nitric oxide (NO) would be immediately oxidized to nitrogen dioxide (NO₂) by O₂ (Wang et al., 2005; Shen et al., 2015, 2019; Miao et al., 2016; Hu et al., 2018; Dong et al., 2019; Zhao et al., 2019). O₂ was flowed into the reaction system in the experimental process to detect the NO. Brown fumes immediately occurred (Figure S4), proving the presence of NO in the inert atmosphere. Subsequently, NO was prepared by the reaction between Cu and 35 wt % nitric acid (3Cu + 8HNO₃ → 3Cu(NO₃)₂ + 2NO + 4H₂O). The oxidation of benzyl alcohol by NO was conducted and the detailed experimental process was shown in Figure S5. The conversion of benzyl alcohol by NO was only 1% at 6 h with 99% selectivity, as shown in Figure 7. The results indicated NO itself could not oxidize benzyl alcohol. Hence, the NO₂ from the decomposition of Fe(NO₃)₃ played the key role in oxidation process by forming HNO₂.

According to the above investigation, the mechanism of the benzyl alcohol oxidation in N₂ condition was put forward and shown in Scheme 1. At the suitable reaction temperature, the Fe(NO₃)₃ decomposes into Fe₂O₃, NO₂, and O₂ (Yuvaraj et al., 2003) by Equation 1. Subsequently, the HNO₂ and HNO₃ can be formed by the reaction between NO₂ and H₂O, see Equation 2. Then, the oxidation reaction successfully gets into the propagation stage and HNO₂ attacks the benzyl alcohol (PhCH₂OH) and gives benzyl nitrite (PhCH₂ONO). Benzyl nitrite decomposes into benzaldehyde (PhCHO) and HNO at experimental temperature by Equation 4. In the termination stage, as shown in Equation 6, Fe(NO₃)₂ can be formed by the reaction between the HNO₂ and Fe₂O₃ with consuming O₂. At the same time, Fe(NO₃)₂ is oxidized to Fe(NO₃)₃ by O₂. Consequently, from Reaction 1 to Reaction 7, NO3- can be considered as the real oxidant, and the Fe ion acts a catalyst through the transformation between Fe²⁺ and Fe³⁺.

From the reaction mechanism, when the system was filled with O₂, the decomposition of Fe(NO₃)₃ would be hindered, leading to low conversion compared with anaerobic condition (N₂ or He). Interestingly, the O₂ was double-edged gas since it could oxidize the Fe²⁺ to Fe³⁺ and eliminate NO Equation 7. In the presence of adequate oxygen, the Fe²⁺ was completely oxidized into Fe³⁺, and oxidative state of Fe remained the same before and after the reaction. The conversion of benzyl alcohol in Fe(NO₃)₃-O₂ system was relatively low (Figure 8A), but it was still higher than that in HNO₃ system. Moreover, in Fe(NO₃)₃ system, the conversion of benzyl alcohol could reach 82% when O₂ replaced N₂ after the reaction continuing 2 h (Figure 8A). As expected, the conversion of benzyl alcohol in air (N₂ + O₂) condition was close to that in N₂ condition (Figure 8B). In the aerobic conditions, NO3- would be regenerated and the Fe²⁺ was completely oxidized to Fe³⁺, indicating O₂ was the actual oxidation. Though the O₂ would be not good for the high conversion, it could build a green cyclic oxidation process via removing NO and regenerating Fe(NO₃)₃. Hence, the benzyl alcohol conversion would be relatively high in air atmosphere, which also meet the need of green synthesis due to removing the NO.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.