Pengbo
Wang
,
Xiaopeng
Zhang
,
Shuaijie
Jiang
,
Zheting
Dong
,
Ruyi
Lu
,
Yuangang
Xu
,
Pengcheng
Wang
* and
Guo-Ping
Lu
*
School of Chemistry and Chemical Engineering, Nanjing University of Science & Technology, Xiaolingwei 200, Nanjing 210094, P. R. China. E-mail: glu@njust.edu.cn; alexwpch@njust.edu.cn
First published on 16th June 2024
The discovery of ambient stable pentazolate salts has paved way for a new frontier in the chemistry of pentazoles, which represent a distinctive category of energetic materials. One challenge impeding their practical applications lies in achieving the scale-up of pentazolate synthesis via a facile and low-cost route. An efficient and recyclable iron single-atom catalyst with FeN4 sites (Fe1@NC-700) has been first developed using a pyrolysis-milling strategy for scaled-up pentazolate synthesis and achieves significant yields (38.4 g, 15% yield for NaN5; 77.6 g, 26% yield for CoN5) via a three-step, one-pot process with a low iron dosage (0.67 mol%). According to experimental and theoretical calculation results, the exceptional catalytic performance of Fe1@NC-700 can be attributed to its stable FeN4 enzyme-like catalytic active sites, which facilitate the formation of a Fe(IV)O intermediate, and it maintains its catalytic activity even under weakly acidic conditions and after five cycles.
At present, NaN53 is synthesized via the C–N bond cleavage of arylpentazole 2, which is produced from arylamine hydrochloride (2,6-dimethyl-4-aminophenol hydrochloride, 1) through a diazotization and cyclization reaction (Scheme 1). The yield of arylpentazole 2 is about 60%, while the yield of the C–N bond cleavage step is poor (∼10%); thus, the total yield is only ∼6% with respect to arylamine hydrochloride 1. In the C–N bond cleavage step, a tedious work-up procedure and excessive iron catalyst and oxidant usage are the norm, thereby making this approach unsuitable for large-scale synthesis.1a Hence, it is appealing and desirable to improve the synthetic process of NaN5 for the achievement of its efficient scaled-up synthesis.
Hang's group revealed the mechanism of the oxidative cleavage of the C–N bond of 2: Fe(Gly)2 reacts with mCPBA to form a Fe(IV)O intermediate, which can realize the oxidative C–N bond cleavage of 2 to form NaN5. DFT calculations indicate that the Fe(IV)O intermediate plays a crucial role in the reaction, since it greatly reduces the reaction activation energy.3 It is well known that some iron-based enzymes with a FeN4 structure (such as cytochrome P450 or hemin) can activate oxygen to form Fe(IV)O species for the oxidation of organic substrates under mild conditions.4 According to the bionic theory, a series of N-doped carbon-supported iron single-atom catalysts with enzyme-like FeNx sites have been developed for selective oxidation reactions.5 Hence, an N-doped carbon-supported iron single-atom catalyst with FeNx sites could exhibit good performance in the C–N bond cleavage of 2 for the effective preparation of NaN5.
In addition, the unstable compound 2 must be purified using acetone washing at −60 °C in the original preparation process, which greatly reduces the yield of 2 owing to the decomposition and dissolution of 2 in acetone. The separation and purification of NaN5 requires column chromatography, which is costly and inefficient, and makes the process difficult to scale up. According to these results, we propose three strategies for the upgrading of this synthetic process: (1) an iron single-atom catalyst with highly active FeNx sites could be applied instead of Fe(Gly)2 for the selective oxidative cleavage of the C–N bond to improve the activity of Fe(IV)O. (2) The synthesis of 3 could be directly completed via a one-pot, three-step method, which would avoid the decomposition of 2 during the work-up process. (3) Pentazolate ions could be converted into cobalt pentazolate precipitates (CoN5) via a coordination-precipitation strategy, which can be separated by simple filtration and transformed into other pentazolate salts through metathesis reactions.6
Along this line, we have developed a pyrolysis-milling approach for the synthesis of N-doped carbon-supported iron single-atom catalysts (Fe1@NC-x), which display much better performance than Fe(Gly)2 and hemin owing to their high activity and the stability of enzyme-like FeN4 single-atom sites. These materials can also achieve a one-pot, three-step process for the scaled-up synthesis of NaN5 (38.4 g) and CoN5 (77.6 g), providing several advantages, including a much lower dosage (0.67 mol%) of the recyclable iron catalyst (>5 runs), a simplified work-up procedure and higher pentazolate yields (Fig. 2). To the best of our knowledge, this is the first example of single-atom catalysis for scaled-up pentazolate synthesis.
Fig. 3 (a) Synthesis route of Fe1@NC-x; HRTEM images of Fe1@NC-700 (b), Fe1@NC-800 (c), Fe1@NC-900 (d); AC-HAADF-STEM images of Fe1@NC-700 (e), Fe1@NC-800 (f), and Fe1@NC-900 (g). |
According to Raman spectra (Fig. S2†), the D (defect) and G (graphite) bands of Fe1@NC-x were observed at about 1336 cm−1 and 1582 cm−1, respectively. The ID/IG value of Fe1@NC-700 is higher than that of the other two catalysts, indicating that Fe1@NC-700 has the most defects.8 The Fe contents of Fe1@NC-700, Fe1@NC-800 and Fe1@NC-900 are 0.82 wt%, 0.73 wt% and 0.65 wt%, respectively (Table S1†). The BET surface area and aperture of the Fe1@NC-x catalysts are shown in Fig. S3† and Table S2.† Fe1@NC-700 has the highest BET surface area (68.41 m2 g−1) and largest average aperture (38.56 nm) among these catalysts, which is consistent with Raman results.9
The surface chemical states of these catalysts were measured using X-ray photoelectron spectroscopy (XPS). The N 1s spectrum of Fe1@NC-x could be deconvoluted into four peaks centered at binding energies of 398.4, 399.4, 400.5 and 403.2 eV, which were assigned to pyridinic N, Fe–N, pyrrole N and N-oxide, respectively (Fig. S4†).10 The N content of the three catalysts was calculated by semi-quantitative analysis of XPS results (Table S3†). Among the three catalysts, Fe1@NC-700 has the highest N content, which further explains the reason for its high defect degree and corresponds to Raman and BET characterization data.11 The defects on the carbon material are of great significance for capturing N species, which can anchor more iron species to generate FeNx sites. Moreover, defects enlarge the surface area of the catalysts, which is beneficial to expose more active sites and thus increase interaction between the catalyst and reactive species. Furthermore, Fe1@NC-700 has the highest Fe–N content, which has a positive correlation with the number of active Fe sites, further elucidating the reason for its high catalytic activity.12
Aberration-corrected high annular angle dark-field scanning TEM (AC-HAADF-STEM) and X-ray absorption fine structure (XAFS) spectroscopy were also employed to study the Fe configurations of Fe@NC-x catalysts at the atomic level. As shown in Fig. 3e–g, isolated bright spots (marked by red circles) show good dispersion in these catalysts, proving that iron species exist in these catalysts as single atomic iron sites.
The energy absorption edge of Fe1@NC-700 lies between the energy absorption boundaries of Fe foil and Fe2O3, indicating that the Fe valence state is between 0 and +3 (Fig. 4a).13 Fourier-transformed k3-weighted extended X-ray absorption fine structure (FT-EXAFS) spectra are shown in Fig. 4b. Fe1@NC-700 exhibits only a main peak at 1.6 Å, which is attributed to the Fe–N bond, further confirming that the Fe species of Fe1@NC-700 mainly exist in the form of single Fe atom sites.14 According to the results of fitting using the program FEFF6, the coordination number of the single Fe atom is about 3.9 (Table S4†), and the fitting curve of the FeN4 sites is basically consistent with this catalyst curve (Fig. 4c); thus, iron single-atom sites should mainly be FeN4 coordination structures.15
Fig. 4 (a and b) XANES and FT EXAFS spectra at the Fe K-edge of Fe1@NC-700, Fe foil, FePc and Fe2O3. (c) Corresponding fitting of the EXAFS spectrum of Fe1@NC-700 in R space. |
With these iron catalysts in hand, their catalytic activities were studied in a one-pot three-step process for the synthesis of NaN5 (Table 1), whose structure was confirmed using MS, IR and ion chromatography (Fig. S5–S7 and S11a†). Fe(Gly)2 could not achieve this three-step, one-pot process (entry 1), since its Fe ions would dissociate from the ligand under acidic conditions. Hemin, with its stable FeN4 sites, could achieve the selective oxidation and cleavage of C–N bonds to synthesize NaN5 under acidic conditions, with a yield of 10%. When FeCl2 was used as the catalyst, no desired product was obtained, indicating that the Fe ions without N coordination have no catalytic activity for the reaction (entry 3). The yields of Fe1@NC-x are much higher than those of Fe(Gly)2, hemin and FeCl2 (entries 1–6), since their Fe sites are anchored into NC to form stable FeN4 single-atom sites, which exhibit excellent catalytic oxidation performance.16 In addition to the FeN4 sites, the porous structure of Fe1@NC-700 promotes the reaction to a certain extent, because it is beneficial for the interaction between reactants (or intermediates) and catalysts.
Entry | Catalyst | Catalyst (g) | mCPBAb (g) | Yieldc (%) |
---|---|---|---|---|
a Reaction conditions: step 1: 87 mmol 2,6-dimethyl-4-aminophenol hydrochloride (15.1 g), 96 mmol HCl, 90 mmol NaNO2, THF, −5 °C, 0.5 h; step 2: 90 mmol NaN3, MeOH, −40 °C, 2 h; step 3: catalyst, mCPBA, MeCN, −15 °C, 12 h. b m-Chloroperbenzoic acid. c Yields were calculated based on arylamine hydrochloride 1. | ||||
1 | Fe(Gly)2 | 42 | 33 | 0 |
2 | Hemin | 56 | 33 | 10 |
3 | FeCl2 | 4 | 33 | 0 |
4 | Fe1@NC-800 | 4 | 33 | 16 |
5 | Fe1@NC-900 | 4 | 33 | 14 |
6 | Fe1@NC-700 | 4 | 33 | 18 |
7 | Fe1@NC-700 | 2 | 33 | 13 |
8 | Fe1@NC-700 | 3 | 33 | 16 |
9 | Fe1@NC-700 | 5 | 33 | 18 |
10 | Fe1@NC-700 | 4 | 16.5 | 10 |
11 | Fe1@NC-700 | 4 | 24.8 | 16 |
12 | Fe1@NC-700 | 4 | 41.3 | 18 |
13 | Fe1@NC-700-2 | 4 | 33 | 17 |
14 | Co1@NC-700 | 4 | 33 | 1 |
15 | NC-700 | 4 | 33 | 0 |
16 | — | 4 | 33 | 0 |
Among the three Fe1@NC-x catalysts, Fe1@NC-700 exhibits the best performance because it has the most iron single-atom sites (entries 4–6). The dosages of Fe1@NC-700 and mCPBA were also optimized, and the combination of 4 g Fe1@NC-700 and 33 g mCPBA in the 87 mmol scale reaction was the best option (entries 6–12). Fe1@NC-700-2 and Co@NC-700 were synthesized via the same preparation methods using FeCl3 and CoSO4·7H2O as metal precursors. Fe1@NC-700-2 displays similar catalytic activity to Fe1@NC-700 (entries 6 vs. 13), while only trace NaN5 can be obtained in the case of Co1@NC-700 (entry 14), since it cannot form a Co(IV)O intermediate. Blank experiments verify that the iron sites are the catalytically active center of this reaction (entries 15 vs. 16).
To further highlight the potential of this catalyst, we increased the reaction scale 20 times (1.74 mol, 302 g of 1). Fe1@NC-700 maintained good catalytic activity, and the yield of NaN5 reached 15% (Fig. 5a). Nevertheless, the work-up procedure for the separation and purification of NaN5 is tedious, including multiple room-temperature vacuum rotary evaporations for water removal and column chromatography, which results in an ultra-long post-treatment time, the use of excessive organic solvents, high energy consumption, and loss and decomposition of NaN5 (Table 2, entry 2).
Entry | Route | Purification | Cat. Fe (g) | Solventb (L) | PTc (h) | Yieldd (%) | Puritye (%) |
---|---|---|---|---|---|---|---|
a Reaction conditions: step 1: 1.74 mol 2,6-dimethyl-4-aminophenol hydrochloride (302 g), 1.92 mol HCl, 1.80 mol NaNO2, THF, −5 °C, 1 h; step 2: 1.80 mol NaN3, MeOH, −40 °C, 2 h; step 3: catalyst, 660 g mCPBA, MeCN, −15 °C, 12 h. b Post-processing solvent usage. c Post-processing time. d Yields were calculated based on arylamine hydrochloride 1. e Product purity was determined using ion chromatography. | |||||||
1 | Two-pot | Column chromatography | Fe(Gly)2 (840) | 33.0 | 26 | 7 (NaN5) | 70 |
2 | One-pot | Column chromatography | Fe1@NC-700 (80) | 31.0 | 20 | 15 (NaN5) | 90 |
3 | One-pot | Precipitation and filtration | Fe1@NC-700 (80) | 6.5 | 7 | 26 (CoN5) | 98 |
To eliminate these issues, a coordination-precipitation strategy has been developed. Organic acid salts in the reaction solution are first converted into organic acids by adjusting the pH and further removed by extraction. Then, pentazolate anions (N5−) cooperate with Co2+ salts to yield CoN5 precipitate ([Co(H2O)4(N5)2]·4H2O), which can be separated by simple filtration (Fig. 5b, S8–S10, S11b; Tables S5 and S6†). Furthermore, CoN5 can replace NaN5 as the basic raw material for other pentazolate salts.6 A 26% yield of CoN5 (77.6 g) was afforded through the coordination-precipitation strategy.
To verify the stability and recyclability of Fe1@NC-700, the scaled-up synthesis of CoN5 was chosen as the model reaction. The substrate (arylamine hydrochloride) was completely converted into pentazolate and other by-products. Fe1@NC-700 should be able to adsorb some by-products generated during the reaction, such as phenols, mCPBA and m-chlorobenzoic acid. After the reaction was completed, the weight of the filtered recovered catalyst increased by 9% after drying, which confirmed our hypothesis. Therefore, ethanol washing is necessary to remove adsorbates during the catalyst recovery process. However, pentazolate salts have good water solubility, so they should not be adsorbed on the catalyst. As shown in Fig. 5c, Fe1@NC-700 could still maintain high catalytic activity after five runs. A comparison of different synthetic routes for the scaled-up pentazolate synthesis is provided in Table 2. Compared to the other two routes, the one-pot CoN5 synthetic route has a higher yield and purity and shorter post-processing time and uses less Fe catalyst and post-processing solvent.
To confirm the Fe(IV)O intermediate in this process, a low-temperature (77 K) electron paramagnetic resonance (EPR) experiment was carried out. Rhombic signals were observed after the reaction of hemin and Fe1@NC-700 with excess mCPBA, thereby demonstrating the existence of a Fe(IV)O intermediate.17 The EPR singlet intensity of Fe1@NC-700 is higher than that of hemin, suggesting that Fe1@NC-700 is more conducive to the generation of Fe(IV)O than hemin. No singlet was detected in the case of FeCl2, which means that it failed to react with mCPBA to afford Fe(IV)O. This can explain why FeCl2 has no catalytic performance in this reaction.
The Gibbs free energy profiles for the reaction of iron catalysts (hemin, Fe(Gly)2 and FeN4) with mCPBA are also illustrated in Fig. 6a and S12–S14.† The energy barrier for the reaction of FeN4 (the catalytically active site of Fe1@NC-700) with mCPBA to generate Fe(IV)O is the lowest (3.49 kcal mol−1). This further proves that the formation of Fe(IV)O over Fe1@NC-700 is easier than for other catalysts (hemin and Fe(Gly)2). Fig. 6b shows the free energy diagram for the selective oxidative cleavage of the C–N bond. According to the Gibbs free energy profiles of these two reactions, the energy barriers of iron-catalyzed reactions for NaN5 synthesis follow the order Fe1@NC-700 (6.23 kcal mol−1) < Fe(Gly)2 (9.85 kcal mol−1) < hemin (11.33 kcal mol−1).
According to above presented results, the superior catalytic performance of Fe1@NC-700 can be attributed to following aspects. (1) It has stable FeN4 enzyme-like catalytic active sites, which promote the formation of a Fe(IV)O intermediate and maintain catalytic activity even under weakly acidic conditions and after multiple cycles. (2) It has a higher surface area with more defects and the FeN4 sites than Fe1@NC-800 and Fe1@NC-900, all of which can boost its catalytic activity.
Footnote |
† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4ta03312d |
This journal is © The Royal Society of Chemistry 2024 |