Catalytic reductions of nitroaromatic compounds over heterogeneous catalysts with rhenium sub-nanostructures

Nitroaromatic compounds (NACs) are key contaminants of anthropogenic origin and pose a severe threat to human and animal lives. Although the catalytic activities of Re nanostructures (NSs) are significantly higher than those of other heterogeneous catalysts containing NSs, few studies have been reported on the application of Re-based nanocatalysts for NAC hydrogenation. Accordingly, herein, catalytic reductions of nitrobenzene (NB), 4-nitrophenol (4-NP), 2-nitroaniline (2-NA), 4-nitroaniline (4-NA), and 2,4,6-trinitrophenol (2,4,6-TNP) over new Re-based heterogeneous catalysts were proposed. The catalytic materials were designed to enable effective syntheses and stabilisation of particularly small Re structures over them. Accordingly, catalytic hydrogenations of NACs under mild conditions were significantly enhanced by Re sub-nanostructures (Re-sub-NSs). The highest pseudo-first-order rate constants for NB, 4-NP, 2-NA, 4-NA, and 2,4,6-TNP reductions over the catalyst acquired by stabilising Re using bis(3-aminopropyl)amine (BAPA), which led to Re-sub-NSs with Re concentrations of 16.7 wt%, were 0.210, 0.130, 0.100, 0.180, and 0.090 min−1, respectively.


Results and discussion
The new heterogenous catalysts base on polymeric matrices with stabilizing functionalities derived from bis(3aminopropyl)amine (BAPA), 4(5)-(hydroxymethyl)imidazole (HMI), 1-(2-pyrimidyl)piperazine (PP), thiosemicarbazide (TSC), 2-amino-3-hydroxypiridine (AHP), 1-(2-hydroxyethyl)piperazine (HEP), 4(6)-aminouracil (AUr), 1,1′-carbonyldiimidazole (CDI) and 2-aminothiazol (AT).To the structure of these polymers, Re-reactive sites were introduced using two approaches, which are explained in Fig. 1.Briefly, the first approach involved the reduction of ReO 4 − using NaBH 4 as a reducing agent (Fig. 1A), while the second approach included the reduction of ReO 4 − via the transfer of an electron from the N atom of the amino functionality to ReO 4 − (Fig. 1B).The detailed discussion on the synthesis of catalysts is provided in Supplementary Information, Sect.S1.The catalyst samples were coded using abbreviated names of amines, with the prefix ext Re or Re added for the samples obtained with the external reducing agent and reduction-coupled adsorption, respectively.
Model reaction of 4-NP reduction.Herein, the reduction of 4-NP to 4-aminophenol (4-AP) is chosen for modeling catalysts' activity 25 .At first, the ext Re samples were examined.The corresponding sets of pseudofirst-order kinetic plots and UV/Vis spectra for the catalytic reduction of 4-NP using each NCat sample are depicted in Fig. 2.
The obtained data (Fig. 2) implied that the model catalytic reaction appropriately obeyed the pseudo-firstorder kinetics as correlation coefficients (R 2 ) ranged from 0.83 to 1 (Fig. 2).The fabricated NCat samples were divided into two groups.The first group of samples comprised ext ReBAPA, ext ReCDI, ext ReHMI, ext ReAHP, and ext ReHEP and demonstrated catalytic activities for the reduction of 4-NP to 4-AP.The highest pseudo-firstorder rate constant (k 1 ) (5.17 × 10 −2 min −1 ) was observed in the case of ext ReHEP, whereas the smallest k 1 was observed for ext ReAHP (9.01 × 10 −4 min −1 ).This simultaneously influenced the 4-NP conversion yield (%). ext ReHEP led to > 95% conversion of 4-NP within 300 min, whereas ext ReCDI (k 1 = 2.33 × 10 −3 min −1 ) and ext Re-HMI (k 1 = 1.64 × 10 −2 min −1 ) resulted in 95 and 93% conversions of 4-NP within 125 and 175 min, respectively.ext ReAHP was considerably less efficient as it led to only 27% conversion of 4-NP with k 1 = 9.01 × 10 −4 min −1 .The second group of samples consisting of ext RePP, ext ReTSC, ext ReAUr, and ext ReAT exhibited insufficient catalytic activities for the reduction of 4-NP to 4-AP.The k 1 values obtained for the second group of samples were approximately one order of magnitude lower than those achieved for the first group of samples and ranged from 4.84 × 10 −5 to 9.61 × 10 −5 min −1 .Furthermore, negligible catalytic conversions of 4-NP (< 4%) were acquired using the second group of samples.
Adsorption of ReO 4 − on anion-exchange resins more efficiently proceeds on amines with complex structures [26][27][28] .In this context, Certainly, there is a relationship between the observed catalytic activities of NCats and the structures of the amines incorporated into the polymer matrices of NCats (Fig. 1C).Samples containing amines with complex structures, namely, a clam-like structure (CDI), samples comprising substituents based on an aliphatic chain (HMI and HEP), or samples characterised by a long-chain aliphatic structure (BAPA) demonstrated significantly higher catalytic activities as compared to those of the samples containing short substituents (PP, TSC, AHP, AUr, and AT) (Fig. 1C).This indicated that the functionalities derived from CDI, HMI, HEP, and BAPA might provide a synergistic effect between the polymer matrix and the ReNSs, and thus, NCats with high catalytic activities would be obtained.This synergistic effect may not be the only reason for the differences observed between the apparent k of the catalytic reductions of 4-NP to 4-AP performed using different NCats.Several other factors, including the concentration of Re in the samples, average sizes of the fabricated ReNSs, and the oxidation state of Re in the synthesised ReNSs, might also influence the catalytic activities of NCats.The effects of these factors on the catalytic activities of NCats are discussed later in the manuscript.
The catalytic activities of the samples obtained using second synthesis approach (Fig. 1B) were also analysed for the model reduction reaction, and the corresponding results are shown in Fig. 3.The highest catalytic activity was observed in the case of ReBAPA, which converted 97% 4-NP within 40 min with k 1 = 0.012-0.21min −1 (Fig. 3).The same 4-NP conversion yield was achieved within 47 min using ReHMI.However, in this case, k 1 was 0.077 min −1 .The remaining samples ReCDI, ReHEP, and ReAHP resulted in 97, 83, and 32% 4-NP conversions with k 1 values of 0.064, 0.61 × 10 −3 , and 0.27-4.1 × 10 −2 min −1 , respectively.These data are consistent with the results obtained for the catalysts comprising ext Re as the apparent catalytic activities of ReBAPA, ReCDI, ReHMI, ReHEP, and ReAHP (the first group of samples) are linked with the functionalities derived from BAPA, CDI, HMI, HEP, and AHP, respectively.RePP, ReTSC, ReAUr, and ReAT (the second group of samples) exhibited negligible or no catalytic activities for this reaction, similar to the cases of their equivalents acquired using NaBH 4 (Figs. 2 and 3).This suggested that the conclusion regarding the synergistic effect between amino functionalities and ReNSs was valid.
To make comparing the data easier, Table S4 located in the Supplementary Information, Sect.S1.4 sets obtained rate constants, together with Turnover Frequencies (TOF) and maximum yields of NACs conversions for all of the samples.Based on Figs. 2 and 3 and Table S4 significant differences were observed between the catalytic activities of the NCat samples obtained using the two methods (Fig. 1A,B).Generally, the catalytic activities of the Re samples were substantially higher than those of the ext Re samples.For instance, the k 1 value calculated for ReBAPA (0.21 min −1 ) was an order of magnitude higher than the corresponding value evaluated for ext ReBAPA (1.78 × 10 −3 min −1 ).Similar phenomena were noticed for ReCDI and ext ReCDI (k 1 = 0.064 vs. 2.33 × 10 −3 min −1 ), ReHIM and ext ReHIM (k 1 = 0.077 vs. 1.64 × 10 −2 min −1 ), and ReHEP and ext ReHEP (k 1 = 0.041 vs. 3.25 × 10 −2 min −1 ), respectively.These observations, combined with received TOF parameters (Table S4) suggest that this tendency is associated with the method applied for the syntheses of NCats.Thus, the origin of this tendency can be hypothesised to be the differences between the physiochemistries of ReO 4 − reductions .This could be the reason for the higher catalytic activities of the Re samples than those of the ext Re samples.Therefore, the above-mentioned issue is comprehensively discussed in the next section.
ReNSs in the polymer matrices.First, it was verified, whether the differences in the observed catalytic activities might have been caused by unequal concentrations of Re in the synthesized samples.The determined concentration of Re and detailed discussion is provided in Supplementary Information, Sect.S1.2 and Table S2.Briefly, the concentrations of Re in the Re samples were higher than those in the ext Re samples.This explains why the Re samples demonstrated significantly higher catalytic activities than the ext Re samples (Figs. 2 and 3).Further, the NCats demonstrated different concentrations of Re, ranging from 0 to 7.1% in the cases of the ext Re samples and from 0 to 18.4% in the cases of the Re samples (Table S2).Among all the applied amines, AUr and TSC evidently prevented or suppressed the reduction of Re(VII) (C Re values for these samples were ~ 0%, Table S2).Additionally, despite relatively high Re concentrations of ext RePP, RePP, ext ReAT, and ReAT (7.1, 17.0, 4.6, and 12.5%, respectively), these NCats exhibited negligible catalytic activities (Figs. 2 and 3); in contrast, other NCats  S2) demonstrated outstanding catalytic activities.These observations confirm the synergistic effect between ReNSs and amino functionalities and allow us to conclude that the applied amine precisely regulates the synthesis routes of ReNSs irrespective of the method used.Second, despite differences in Re concentrations, it was verified how amines, and the methods of syntheses influenced the morphology of the ReNSs.At first, the morphologies of the polymer samples were assessed via ultra-high resolution scanning electron microscopy (UHR-SEM); additional assessments of the polymer crosssections (achieved using Ga-focused ion beam (FIB) and Xe-plasma FIB) and elemental mapping acquired using energy dispersive X-Ray spectrometer (EDX) were also performed.Figure 4 shows the SEM and UHR-SEM images of the representative samples ReBAPA, ext ReBAPA, and ext RePP.These samples were chosen because of the following reasons.The findings obtained for ReBAPA and ext ReBAPA may explain why the two different methods of ReNS syntheses resulted in significantly different Re concentrations in the NCats.Moreover, combining these observations with those acquired for ext RePP may clarify why ext RePP exhibits negligible catalytic activity despite its higher Re concentration (Table S2) and why ext ReBAPA outperforms it.ext RePP is covered with a large number of agglomerates derived from Re (as indicated by the EDX map) (Fig. 4C1).Although a closer observation (Fig. 4A2,B2, and C2) revealed the presence of ReNS agglomerates on the surfaces of all samples, the surfaces of ReBAPA and ext ReBAPA had considerably less agglomerates.Observations of the polymer cross-sections revealed another phenomenon.The areas below the surfaces of the polymer grains were almost entirely clear in the cases of ReBAPA and ext ReBAPA (Fig. 4A3,B3, respectively), whereas those in the case of ext RePP comprised large particles and their agglomerates (Fig. 4C3).Simultaneously, the EDX mapping of Re on the cross-sections demonstrated that ReBAPA had a uniform distribution of Re from the surface to the centre of the polymer grain (Fig. 4A4), while within ext ReBAPA (Fig. 4B4) the higher density of Re was observed in close proximity to the ext ReBAPA surface.The absence of particles in the cross-sections of ReBAPA and ext ReBAPA (Fig. 4A3,B3) implied that Re produced very small NSs, and UHR-SEM was unable to observe them.Thus, high-resolution transmission electron microscopy analysis (HRTEM) was conducted.Figure 5 depicts the HRTEM images of the cross-sections of ReBAPA, ext ReBAPA, and ext RePP as the representative samples.
The observation of the NCat samples was difficult, even using HRTEM.The images of ext RePP demonstrated some structures near the polymer grain surface (first panel) and in the inner part of the polymer grain (right panel) (Fig. 5C).According to the EDX spectra (Fig. 5), all these structures were based on Re.The particles situated near the polymer grain surface were bigger (sizes: ~ 2-10 nm), whereas the particles in the inner part of the polymer grain were smaller (sizes: 0.83 ± 0.52 nm, dark spots in Fig. 5C, right panel).The same phenomenon was observed for ext ReBAPA (Fig. 5B).Nevertheless, in this case, the number of NSs was considerably lower than that of ext RePP (Fig. 5B,C).The difference between the observed catalytic activities of ext ReBAPA and ext RePP led to further uncertainties as ext ReBAPA outperformed ext RePP in the catalytic reduction of 4-NP (Fig. 2).This suggested the primary role of BAPA in the syntheses and performances of NCats; however, a possible reason for this phenomenon could be deduced by investigating both samples containing BAPA functionalities ( ext ReBAPA, ReBAPA).ext ReBAPA comprised a small number of NSs, whereas ReBAPA contained no NSs (Fig. 5A).Some phase contrast was noticed near the polymer grain surface in Fig. 5A, which indicated the production of very small structures; nevertheless, HRTEM was not a suitable technique to examine these structures.The significantly high catalytic activity of ReBAPA than those of all other NCats further validates the hypothesis that some Rebased structures must have formed in ReBAPA (Figs. 2 and 3).This finding necessitates further investigation of ReBAPA.Thus, additionally, HRTEM in the scanning-transmission (STEM) mode using large magnifications and a high angle annular dark field detector (HAADF) was performed.Figure 6 shows the corresponding images.
Large magnifications, along with the use of the STEM-HAADF mode, allowed the observation of single Re atoms (bright spots in the images depicted in Fig. 6).These atoms occasionally grouped in sub-NSs, implied by the local increase in Re density (larger bright areas).Single NSs were observed for ext ReBAPA and ext RePP (Figs. 5B,C and 6B,C, left panels), whereas the situation was completely different in the case of ReBAPA, which consists of large areas entirely loaded with Re atoms that tended to group on the << 1 nm scale were noticed (Fig. 6A, right panel).The corresponding images might have been acquired at the very beginning of the formation of these structures, which would have been "frozen" by the polymer matrix itself.This was ascribed to the effective stabilisation of ReO 4 − by amines, which might have prevented the growth of reduced Re forms.According to these findings, it was concluded that although BAPA prevented the formation of NSs, it facilitated the production of Re-sub-NSs, which were catalytically active and substantially boosted the catalytic activities of catalysts in the reduction of 4-NP.This conclusion might be extended to other samples that exhibited higher catalytic activities than those of other samples.Thus, CDI, HMI, and HEP functionalities must also contribute to the formation of Re-sub-NSs.
Re oxidation states in the polymer matrices.The selected area energy diffraction (SAED) patterns obtained during HRTEM (Fig. 5) indicated amorphous morphologies of the investigated materials.These observations could indicate the existence of Re-Re bonds in ReNSs 21,31,32 .Nevertheless, Re can form various stable species 16,17 , and the Re-sub-NSs observed by STEM-HAADF exhibited amplitude contrasts originating from the lack of long-range orderings of the particles.Therefore, SAED was not sufficient to draw any binding conclusions in this case.Therefore, X-Ray photoelectron spectroscopy (XPS) analysis was performed to determine the types of Re species in the synthesized samples.
Surface concentrations of chemical bonds achieved by fitting the XPS data for all analyzed samples are presented in Table S3.Furthermore, the methodology for determining specific bonds is systematically described in Sect.S2 of the supplementary information.Moreover, all the acquired spectra are provided in S3.1 (survey scans) and S3.2 (high-resolution spectra).Re 4f spectra (Section S1.2) were fitted with up to three doublet structures (doublet separation f 7/2 -f 5/2 = 2.43 eV) with the first 4f 7/2 line centred at 42.4 eV, which implied the presence of Re 4+ similar to that in ReO 2 .The second 4f 7/2 line centred at 44.0 eV demonstrated the existence of Re 6+ similar to that in ReO 3 , and the last 4f 7/2 line at 46.1 eV indicated the presence of Re 7+ similar to that in Re 2 O 7 33,34 .Based on these results, it was concluded that the reduction of ReO 4 − was indeed successful.Additionally, note that Re 7+ in Re 2 O 7 was different from Re 7+ in NH 4 ReO 4 (precursor).
Atomic% Re concentrations provided in Table S3 are consistent with those determined in Table S2.Generally, the concentrations of Re in ext Re samples were lower than those in the Re samples.Moreover, the atomic% concentrations of Re 4+ , Re 6+ , and Re 7+ were typically higher in the second group of samples (Table S3).All these findings agreed with the observed k values of 4-NP reduction for the Re samples; as evidenced in the model reaction, the Re samples outperformed the ext Re samples (Figs. 2 and 3).This almost three-fold difference between the k m values for 4-NP and 2,4,6-TNP was proportional to the number of -NO 2 groups present in the structures of 4-NP (one group) and 2,4,6-TNP (three groups).Nevertheless, this observation did not explain why the catalytic activity of ReBAPA was slightly lower for 2,4,6-TNP than those in the cases of other NACs.Consequently, the yields (%) of NAC reductions, along with the TOF values at 20, 50, 70, and 80% NAC reduction, were examined and detailed in Supplementary Information, Sect.S1 and Fig. S1.Based on these, it might be suggested that ReBAPA revealed similar catalytic activity towards each NAC (the detailed discussion is provided in Supplementary Information, Sect.S1.4).
Present study versus other Re-based catalysts.Up to date, few studies have been reported on the application of homogeneous and heterogeneous Re-based NCats in the hydrogenations of NACs, which are summarised in Table 1.The results of this study confirmed that the heterogeneous NCat ReBAPP developed herein was more efficient than the homogeneous catalysts based on Re nanoclusters 19 and raw ReNSs 23 .Additionally, the calculated k 1 values were similar to those obtained using hybrid catalysts comprising Ag and AuNPs anchored on ReS 2 nanosheets 31 .Nevertheless, ReBAPP also outperformed PtNPs immobilised on ReS 2 in the reductions of 4-NP and 2-NA 31 .Although homogeneous NCats based on Re are rare, heterogeneous NCats based on Re are even rarer.When this manuscript was written (July 2023), a review of the literature revealed that few studies were reported on the application of Re-based heterogeneous NCats in the hydrogenations of NACs.In our previous studies, we attempted to optimise the structures of ReNSs by synthesising Re-based NMs with Re 022 or O-doped ReNSs 24 .Although the first approach resulted in Re 0 NPs with high catalytic activities, it did not offer catalyst stability as the Re 0 NPs interacted with O and thus lost their catalytic activities 22 .In contrast, the other approach based on reduction-coupled adsorption enabled the fabrication of O-doped ReNSs with slightly lower catalytic activities and higher stabilities than those of the abovementioned Re 0 NPs in the catalytic reduction of 4-NP 24 .

Conclusions
In this study, we propose unique NCats loaded with ReNSs and Re-sub-NSs that can efficiently reduce NACs under mild conditions.The findings herein revealed that the unique reduction-coupled adsorption of ReO 4 − on amino functionalities should be preferred for the fabrication of ReNSs rather than the approach involving the use Results indicated a synergistic effect between the ReNSs and the amino functionalities present in the polymer matrix.The amino functionalities with complex structures enabled efficient stabilisation of Re atoms, resulting in Re-sub-NSs.Moreover, a synergistic effect was noticed between NCats and the amines applied for their stabilisation.The amino functionalities with complex structures provided high catalytic activities to NCats.The mechanism of ReO 4 − reduction was simple and did not require additional preparation or strict control over the synthesis conditions.Consequently, ReO 4 − present in the polymer matrix formed Re 4+ , Re 6+ , and Re 7+ O-doped species, which exhibited outstanding catalytic activities when combined with a synergistic amine.The type of Re species generated did not influence the catalytic activities of NCats as the corresponding ReNSs and Re-sub-NSs were versatile.In this regard, the kind of amine was the only limiting factor affecting the catalytic activities of NCats.Using a selected amine, the sizes of the structures formed by Re can be regulated, facilitating the development of NCats with high catalytic activities.
This study provides further insight into these unique NMs.First, Re-based NCats can be applied to other reductions instead of only the model reaction as this type of NCats effectively reduce other NACs in addition to 4-NP.Second, because of the versatility of ReNSs, the forms in which Re-based NPs are produced do not influence the catalytic activities of NCats, and thus, ReNSs can be prepared via simpler procedures.Third, the application of an appropriate amine enables the formation of Re subnanometric structures that further boost the catalytic activities of Re-based materials.Fourth, owing to the synergistic effect between ReNSs and amino functionalities and the affinity of the polymer matrix to NACs, the catalytic activities of the developed NCats can be enhanced or tailored.Thus, based on the results of this study, a process for NAC reduction that can be performed in a flow mode with the simultaneous production of AAMs can be envisioned.

Material and methods
Detailed list of standard materials, instrumentation, and synthetic protocols is provided in Supplementary Information, Sect.S1.

Figure 1 .
Figure 1.Schematics of the syntheses of Re nanostructures (NSs) in the polymer matrix using (A) a reducing agent and (B) reduction-coupled adsorption and (C) structures of the amines present in anion-exchange resins.

Figure 2 .
Figure 2. Pseudo-first-order kinetic plots and ultraviolet (UV)/visible (Vis) spectra for the reduction of 4-nitrophenol (4-NP) using catalysts containing ReNSs synthesised using an external reducing agent.Initial 4-NP concentration: 0.1 mmol L −1 and NaBH 4 concentration: 0.1 mol L −1 (0.3 mL).The total reaction time and time intervals between subsequent UV/Vis spectra correspond to the data points shown in the pseudo-firstorder kinetics plots.

Figure 3 .
Figure 3. Pseudo-first-order kinetic plots and UV/Vis spectra for the reductions of 4-NP using different catalysts containing Re-active sites fabricated by reduction-coupled adsorption of ReO 4 − on amino functionalities.Initial 4-NP concentration: 0.1 mmol L −1 and NaBH 4 concentration: 0.1 mol L −1 (0.3 mL).The total reaction time and time intervals between subsequent UV/Vis spectra correspond to the data points depicted in the pseudo-first-order kinetic plots.