Catalytic performance of PC catalysts

The typical catalytic performance of pelagic sediments (PC#1), examined at the initial concentrations of 4-NP and NaBH₄, were 0.1 mM and 25.0 mM, respectively. The color change was observed from light yellow to bright yellow (Fig. S2), and the absorption peak was red-shifted from ~317 nm to ~400 nm after adding NaBH₄ (Fig. 1a). This was due to the fact that the solution pH increased from neutral to ~10.0, and the 4-NP molecules changed into 4-nitrophenolate ions with the addition of NaBH₄ (Reference Jiang, Liu, Yuan, Feng, Ji, Wang, Losic, Yao and ZhangJiang et al., 2020; Reference Mohammadnezhad and AriaeinezhadMohammadnezhad & Ariaeinezhad, 2021). The absorbance peak at ~400 nm remained constant, and solution color remained bright yellow in the absence of PC, proving that 4-nitrophenolate ions were not reduced if the PC particles were not added (as shown in Figs S2b, S5a). In contrast, the UV–Vis absorbance at λ = 400 nm decreased quickly (Fig. 1b), and the color of 4-NP changed to colorless when PC particles were added to the solution and used as catalysts. A video recording of the in situ change of the UV–Vis spectra (link at the end of this paper), demonstrates the outstanding performance of the PC within 720 s. Along with a reduction of the absorbance at 400 nm, a new peak appeared at 300 nm which represented the formation of 4-AP. The degradation of 4-NP was not prominent during the first 2.0 min (Fig. 1c), revealing that the catalytic reaction required a so-called induction time (λ, usually < 2.0 min). After this induction period, the reaction was accelerated and the intensity of the peak at 400 nm declined rapidly, suggesting the degradation was rapid after the induction period. Almost all of the 4-NP was converted to 4-AP within ~480 s, which were confirmed by the color change from bright yellow to colorless (Fig. S2). This result suggested that PC was very efficient at the catalytic reduction of 4-NP.

Fig. 1 a UV–Vis spectra of 4-NP solution before and after addition of NaBH₄; b UV–Vis spectra showing reduction of 4-NP with time after addition of PC; c plots showing the conversion efficiency of 4-NP to 4-NP with added PC (#1–#4); and with d Fermi’s function fitted to the plot of conversion of 4-NP vs. time which was catalyzed by pelagic clays (PC #1). Dosages for the catalysis reaction were 0.10 mM 4-NP, 25.0 mM NaBH₄, and 0.20 g/L PC

As the catalytic reaction progressed, the brown-colored PC particles turned black which indicated that magnetic clay (m-PC) particles were produced (Fig. S2e,f). The SLM images also illustrated an increase in the number of magnetic black particles (Fig. S3). The black m-PC particles, however, were unstable and easily transformed back into brown-colored when the m-PC was treated by washing/drying procedures, especially when particles were exposed to air. The conversion percentage reached 100% within 780 s when samples PC #1–#4 were used as fresh catalysts (Fig. 1c). In case of PC#5, only 10.0% conversion of 4-NP was achieved within 20.0 min (Fig. S4). The catalytic efficiency of the five samples followed the order #1 > #2 > #3 > #4 > #5, suggesting that the pelagic clays with larger larger Fe- and Mn contents showed better catalytic performance.

Traditionally, the conversion percentages (η) of 4-NP by NaBH₄ is as described in Eq. 1. The pseudo-first order kinetics expressed as Eq. 2 described the reaction for metal catalysis, which was of zero-order by considering the concentration of 4-NP with respect to an excess amount of NaBH₄ (Reference Chen, Gao, Fu and YangChen et al., 2020; Reference Jiang, Liu, Yuan, Feng, Ji, Wang, Losic, Yao and ZhangJiang et al., 2020).

(1) $\eta (\%)=100(C_0-C_t)/C_0$

(2) $\ln \left(\frac{C_t}{C_0}\right)=-kt$

In Eqs. 1 and 2, C ₀ and C _t are the initial and instant concentrations of 4-NP at time t, and k is the apparent rate constant. This indicates that the 4-NP consumed in the form of C _t/C ₀ varied with time (t) exponentially. For most metal catalysts, the two-step pseudo-first order model was always used to determine rate constant k (Reference Wunder, Polzer, Lu, Mei and BallauffWunder et al., 2010). However, the kinetics curve of 4-NP degradation catalyzed by the PC in the present system was S-shaped, which varies slightly from that of metal catalysts (Reference Wunder, Polzer, Lu, Mei and BallauffWunder et al., 2010). The reason may be the simultaneous occurrence of 4-NP degradation that followed the ‘induction period’ during catalysis. Therefore, an alternative model, the Fermi function (Eq. 3), was proposed to fit the kinetics of degradation catalyzed by PC (Reference Minz, Garg and GuptaMinz et al., 2018):

(3) $\eta \left(\%\right)=100/(1+\exp (-k(t-t\ast )))$

where η is the conversion percentage of 4-NP, k is the apparent rate constant, and t* is the transition time. The conversion curves were generally fitted in a symmetrical style, and the position of t* determines the period of inflection point in the S-shaped curve, which corresponded to the maximum rate (µ _m ). Therefore, the period of t* is called the lag phase and was appointed as an exponential parameter. According to Reference Ware and PowerWare and Power (2017), the S-shaped kinetics was always modulated as bacterial growth to describe superior kinetics catalyzed by enzyme-processes. Thus, the PC catalysis in the present system was assumed to be an enzyme-like process with great efficiency. The degradation curve is further characterized by introducing the parameter λ as the induction time, which is the intersection of curve slope with the X-axis (Reference Zwietering, Jongenburger, Rombouts and van 't RietZwietering et al., 1990). The equations (Eqs. 4&5) were listed as follows:

(4) ${\mu }_m=k/4$

(5) $\lambda =t^{\ast }-2/k$

where λ is given by the intercept of the curve with the X-axis.

The equations above were used to fit the degradation of 4-NP for reaction conditions of 0.1 mM 4-NP, 25.0 mM NaBH₄, and 0.2 g/L PC#1. Parameters derived from the typical fitted curve (Fig. 1d) were k = 27.53 × 10^–3 s⁻¹, t* = 188.43 s, µ _m = 6.11 × 10^–3 s^–1, λ = 101.2 s, and R² = 0.9991. The model provided accurate quantification on both induction time (λ) and the maximum degradation rate (µ _m ). The coefficient of determination (R²) of catalysis using PC #1–#4 was in the range 0.9983–0.9991 (Fig. 1c, Table 2), suggesting a good match of Fermi’s function. As can be seen, the values for k and µ _m increased and λ decreased with increasing Fe and Mn contents.

Table 2 Parameters derived from Fermi’s function by using pelagic clays (PC #1–#4) as catalysts in the conversion of 4-NP

The PC dosage, pH, and initial concentrations of 4-NP and NaBH₄ in the reaction solutions affect 4-NP degradation performance (Table 3). The k values increased from 24.43 × 10^–3 s^–1 to 31.87 × 10^–3 s^–1 (Fig. 2a) with incremental dosages of PC particles from 0.1 to 0.3 g/L. The pH values ranging from 4.0 to 8.0 were examined to study the influence of pH on the reaction kinetics. The k value was 78.99 × 10^–3 s^–1 at a pH of 4.0, which was much greater than that of pH 7.0 (2Figs b, S5c). With increasing concentration of 4-NP from 0.05 to 0.2 mM, the kinetic constants (k) decreased from 43.34 × 10^–3 s^–1 to 9.31 × 10^–3 s^–1 (Fig. 2c), indicating that the reaction rates were very influenced by the initial concentration of 4-NP. When the initial concentrations of NaBH₄ increased from 10.0 to 75.0 mM, k values increased from 3.58 × 10^–3 s^–1 to 44.47 × 10^–3 s^–1 (Figs 2d, S5b)). The value of k increased, µ _m increased, and the induction time λ decreased, which is consistent with the accelerated kinetics. The range of R² was always found to be between 0.9985 and 0.9995, illustrating the accuracy of using Fermi’s equation in describing all reactions in the catalytic conversion of 4-NP when the PC particles were used as catalysts. These results also indicated that the initial concentration of NaBH₄ was excessive compared with 4-NP; the small variation in NaBH₄ concentration thus had less impact on the 4-NP conversion. The color of 4-NP solution always changed from light yellow to colorless (Fig. S2d) when the initial solution of 4-NP was acidic (pH 4.0). The leaching of metal ions in PC and hydrolysis of NaBH₄ were more susceptible in an acidic environment, which accelerated the enzyme-like catalysis by PC. The induction time decreased significantly (Fig. S5d). Considering the influence of factors above, a better dosage of PC#1 for conducting the catalytic reduction of 4-NP (0.1 mM 4-NP, 25.0 mM NaBH₄) should be 0.2 g/L, and the pH value of the reaction solution was 7.0. To illustrate the superiority of PC, a survey of literature for various catalysts used to catalyze 4-NP reduction by NaBH₄ was done. The catalytic ability of PC was superior to most of the reported catalysts (Table S1).

Table 3 Factors affecting degradation of 4-NP and derived parameters from Fermi’s function

Fig. 2 Plots of the conversion percentage of 4-NP versus time, illustrating various factors that affect the reaction kinetics, including a dosages of PC, b initial pH (4.0 or 7.0) of the aqueous solution of 4-NP + NaBH₄, c, d initial concentrations of 4-NP c and NaBH₄ d

To study the recyclability of the PC catalyst, the m-PC was retrieved by washing, then drying either with or without exposure to air. The PC particles were separated from the solution with a magnet and reused for the next run of catalysis. When samples were washed and dried in air, an increase in reaction time in the next run of catalysis was observed, i.e. the activity was inferior to the fresh sample (Table S2). For example, the conversion percentage of 4-NP during the 3^rd-catalytic reaction cycle was 83.38% with 60 min of reaction time (Fig. 3a). The magnetic particles (m-PC) transformed back to non-magnetic particles if the sample was air dried, but the kinetics curve of 4-NP during the cycling catalysis still remained S-shaped. If the sample was used without exposure to air, the degradation time of 4-NP increased gradually from 8.0 to 45.5 min after five consecutives reuses (Fig. 3b). Although the PC catalysts were deactivated during reuse, they could be reactivated by prolonged time in the mother solution of 4-AP + NaBH₄. The m-PC particles were kept in the mother solution for 5.0 h before reclaiming the PC, then removed with precautions to avoid oxidation to be reused in subsequent catalysis runs. The degradation time for complete conversion of 4-NP was 8.0 min, 6.5 min, 6.5 min, 7.8 min, and 11.7 min for the 1^st- to 5^th-cycles of PC catalysis (Fig. 3c, d). Certainly, the catalytic activity of the fresh sample remained relatively good (Fig. S6), indicating that the prolonged time for preserving PC stabilizes active sites in the reducing environment of the mother solution.

Fig. 3 Plots of the conversion percentage of 4-NP versus time during the reuse of PC particles as catalysts that were reclaimed via various methods: a the PC sample reclaimed after water-washing and air-drying; b PC was reused immediately after the previous run of catalysis without exposure to air; c PC was preserved in the 4-NP solution for ~5.0 h before the next run of catalysis; and d plots showing performance by comparison of the 2nd catalysis using PC reclaimed by the three means above

To perform the fixed-bed catalysis, a length (20.0 mm) of skim cotton and 2.0 mm thick PC were loaded in a glass tube (L = 20.0 mm, ϕ = 6.0 mm), and the catalysis was carried out by flowing a mixed solution of 4-NP and NaBH₄ (Fig. S7). The variation of UV–Vis absorption spectra and conversion percentages of 4-NP collected at different flow rates (Fig. 4) revealed that, with an increase in flow rate from 2.11 to 3.30 mL/min, the conversion efficiency of 4-NP declined from 100% to 99.0%, which meant a complete conversion of 4-NP through this fixed bed. The result indicated that the PC catalyst had the ability to catalyze the conversion of the 4-NP in a continuous flow process. When the flow rate was accelerated, the contact time between 4-NP molecules and PC particles decreased, resulting in incomplete conversion of 4-NP. Based on the experiment, the optimum flow rate for the fixed bed of 4-NP should be ≤2.50 mL/min in the fixed bed of the column.

Fig. 4 a UV–Vis spectra of solution of 4-NP and NaBH₄ collected from the fixed bed after catalyzing at various flow rates (2.11, 2.50, and 3.30 mL/min), and b conversion percentage of 4-NP solution tested at various flow rates

Material characterization

XRD patterns of the five raw pelagic clays illustrated that all samples were almost identical except for the relative intensity of diffraction peaks indexed to rock-forming minerals, e.g. quartz, feldspar, etc. (see Fig. S8). For PC#1 to #4, the main reflections appeared at 12.4º2θ (d ₀₀₁ = 7.13 Å) and 25.0º2θ (d ₀₀₂ = 3.56 Å), indicating the low crystallinity of kaolinite present in the PC. According to the XRD analysis (Fig. S8) and chemical compositions (Table 1), Fe and Mn are probably present primarily as amorphous phases which may be hydroxide phases or the so-called Fe-Mn nodules (Reference Banerjee, Roy, Dasgupta, Mukhopadhyay and MiuraBanerjee et al., 1999; Reference Sensarma, Chakraborty, Banerjee and MukhopadhyaySensarma et al., 2016). Well crystallized calcite was observed as in the main mineral in sample PC#5, which indicated that the geological environment of PC#5 was different from samples PC#1–PC#4.

The ex-situ variation of PC structures during catalysis was studied by using XRD. The diffraction peaks at ~25.0º2θ were broader (Fig. 5a), suggesting structural transformation from the poorly crystalline kaolinite to more amorphous features. The structural transformation seemed to be reversible when the sample was collected from the reaction solution and exposed to air. This could be explained by the fact that PC was reduced by NaBH₄, and the clay structure was slightly collapsed. When the PC was re-oxidized, the structure of PC was reconstructed after exposure to air. The FTIR spectrum of the PC sample after the catalytic reaction gained a new absorption peak at 1500 cm⁻¹, which was the amino vibration of 4-AP (Fig. 5b), indicating that the degradation of 4-NP occurred on the surface of PC, and the 4-AP products were partially adsorbed and blocked the active sites of PC particles.

Fig. 5 a The XRD patterns of the PC samples collected before (raw clay), during, and after the catalytic reduction of 4-NP; and b FTIR spectra of PC before and after catalytic reaction

SEM and TEM were used to characterize the morphology of the PC and the m-PC (Figs. 6, 7). As can be seen, mixed-layer illite-montmorillonite, feldspar, spherical Fe-Mn nodules, and clay layers were all observed (Fig. S9). In accordance with the XRD characterization, the mixed-layer illite-montmorillonite layer in PC was in the range 2.0–3.0 µm with rough surfaces, exhibiting broken and curled edges (Fig. 6). This morphological feature would expand the specific surface area, and would promote more active sites for degradation of 4-NP. The TEM images (Fig. 7a-7c) indicated the complicated composition/mineralogy of the PC. Although the surface curls and clusters of the PC were shape-altered as the cyclic reaction proceeded, most of the layered structures of the pelagic clays were preserved after the reaction (Fig. 7d–f), suggesting that the stability of clay minerals was significant during the catalysis.

Fig. 6 SEM images of a the raw pelagic clays, and b–d the reused pelagic clays after various numbers of cycles of use: b the 1^st, c 2^nd, and d 5^th run of catalysis

Fig. 7 TEM images of the pelagic clay samples used before a–c and after d–f the catalytic reduction of 4-NP

The surface chemistry of PC before and after the reaction was investigated by XPS. The Fe 2p band (Fig. 8a) implied that species of Fe were present both as Fe(III) and Fe(II). The XPS bands at 709.9 and 723.4 eV, and satellite peaks at 715.5 and 728.1 eV, were attributed to the Fe(II). Bands at 711.2 and 725.0 eV and the satellite peaks at 719.0 and 733.0 eV were attributed to Fe(III) (Reference Wan, Li, Wang, Chen, Lu and HuWan et al., 2015; Reference Wang, Zhang, Ning, Zhang, Jin, Wang, Zhang, Li, Wei and MiaoWang et al., 2021). The amount of Fe(II) increased with progress in the catalysis. Based on the integrated area of deconvoluted bands, ratios of the two oxidation states Fe(II):Fe(III) were evaluated to be 0.37:1, 0.52:1, and 0.80:1 from the beginning to end of the reaction. Both oxidation states of Mn(III) (642.3 eV) and Mn(IV) (641.3 eV) were also detected in raw PC (Reference Díaz-Arriaga, Baas-López, Pacheco-Catalán and Uribe-CalderonDíaz-Arriaga et al., 2020; Reference Wang, Zhang, Ning, Zhang, Jin, Wang, Zhang, Li, Wei and MiaoWang et al., 2021). A new characteristic band was observed at 640.4 eV, however (Fig. 8b), which can be assigned to the Mn(II) state, and the amount of this type of manganese increased progressively with reduction of 4-NP (Reference Grissa, Martinez, Cotte, Galipaud, Pecquenard and Le CrasGrissa et al., 2017). When the reaction was completed, the ratio of Mn(II), Mn(III), and Mn(IV) was determined to be 1.03:0.89:1. This might be due to the much lower valence states (Fe(II) and Mn(II)) produced in the PC with the presence of NaBH₄. This can also be used to explain the magnetic variation between PC and m-PC, and the appearance of the lag phase and S-shaped curve of the 4-NP conversion was found in the case of m-PC. No variations in the XPS bands from Si 2p and Al 2p were observed before or after the reaction (Fig. S10). This indicated that the Fe and Mn species played roles in the 4-NP reduction. Unfortunately, due to the poor crystallinity of Fe- and Mn-nodules, the XRD spectra could not give better evidence.

Fig. 8 a The high-resolution XPS spectrum and deconvoluted curves of Fe 2p, and b Mn 2p of pelagic clay before, during, and after the catalytic reduction of 4-NP using NaBH₄

The assay of free Fe²⁺ ions (Fe_fr) that might be dissolved from PC was tested using 1,10-phenanthroline as the indicator. Four types of mother solutions ‘PC + NaBH₄ + 4-NP’, ‘PC + NaBH₄’, ‘NaBH₄ + 4-NP’, and ‘PC + 4-NP’ were tested for time-released Fe ^x+ ions (tested as Fe²⁺, see SI). Only systems of ‘PC + NaBH₄ + PC’ and ‘PC + NaBH₄’ were found to contain Fe_fr (Fig. S11). This indicated that part of the Fe-species was released into solution with the presence of NaBH₄. The curve of the Fe²⁺ concentration of the whole reaction (Fig. 9) revealed that the change of Fe_fr concentration in solution was roughly S-shaped, which was similar to the 4-NP conversion curve. The Fe_fr concentration decreased gradually after the reaction, presumably due to the Fe_fr being fixed back into the clay particles, which could also be illustrated by the Fe_fr concentration at different time in different systems (Fig. S12). The raw PC was a superior enzyme-like catalyst in the conversion of 4-NP to 4-AP with the presence of NaBH₄.

Fig. 9 The Fe²⁺ concentration in the solution (PC + NaBH₄ + 4-NP) during degradation

Proposed reaction mechanisms

The catalytic efficiency depends on both the activated species and surface adsorption of the substrate molecules/ions according to the Langmuir–Hinshelwood pathway (Reference Wunder, Polzer, Lu, Mei and BallauffWunder et al., 2010). In the catalytic reaction, the surface activation during the induction period was assumed to be the initial step, i.e. a reduction of the Fe-Mn amorphous nodules with assistance of NaBH₄. Some low-valence ions in the catalyst would be transferred to active sites at the surface, and ~2.75% of the Fe-species was released into the solution/dispersion. The layered structure of clay minerals provided an ideal support for activation of the clay surface and adsorption of 4-nitrophenol and borohydrides. This type of structure would facilitate electron transfer as well as enhance the catalytic performance (Reference Menumerov, Hughes and NeretinaMenumerov et al., 2016). Based on the characterization of PC samples, the present authors concluded that the induction time was related to dynamic surface activation. This activation of Fe-Mn active sites and adsorption of borohydride and 4-nitrophenol ions were indicated to take place on the PC surface (Reference Minz, Garg and GuptaMinz et al., 2018; Reference Wunder, Polzer, Lu, Mei and BallauffWunder et al., 2010). Based on the characterizations above, a proposed reaction mechanism could be postulated (Fig. 10). With addition of NaBH₄, the 4-NP molecules were converted rapidly to 4-nitrophenol ions; meanwhile the clay minerals were converted to black magnetic m-PC (Fe(0)) by NaBH₄. At the same time, ions of BH₄ ^– lost electrons and converted to HBO₃ ^– which can be described as Eqs. 6–9. The Fe(0) and Mn(II) obtained from the PC during the reduction would participate in the degradation process (4-NP → 4-AP), but these low-valence Fe- or Mn-species would be re-oxidized to Fe(II) and Mn(III) (Fig. 10a) again when the suspension was exposed to air. For the subsequent cycles of reusing PC, the activated PC particles were able to catalyze the reaction directly under condition that the reclaimed m-PC was not oxidized (Fig. 10b). Otherwise, the re-oxidized PC particles have to be activated again in the reductive environment. With assistance of the activated m-PC, the 4-nitrophenol ions and BH₄ ^– ions were adsorbed on the Fe-Mn sites, and the electronic transfer would take place. Following the Langmuir–Hinshelwood pathway, the reduction of 4-NP was indicated to involve collection and transport of electrons, and the continual reduction and immobilization of Fe and Mn species were of great importance. The PC with greater Fe and Mn contents exhibited greater catalytic activity. Therefore, our PC behaved as an efficient enzyme-like catalyst. The activation of PC and its simultaneous catalysis were ascribed to the efficiency that was described by an S-shaped kinetics plot.

Fig. 10 The proposed reaction pathway catalyzed by a raw PC, and b m-PC for conversion from 4-NP to 4-AP

(6) $\mathrm{BH}{{}_4}^{-}+3H_{2}O+2\mathrm{Fe}\left(\mathrm{III}\right)\to 2\mathrm{Fe}\left(0\right)+6H^{+}+H_2{\mathrm{BO}}_3{}^{-}+H_2$

(7) $\begin{array}{c}\mathrm{BH}{}_4{}^{-}+{3H}_{2}O+\mathrm{Mn}\left(\mathrm{IV}\right)+2\mathrm{Mn}\left(\mathrm{III}\right)\to 3\mathrm{Mn}\left(\mathrm{II}\right)+{4H}^{+}+H_2{\mathrm{BO}}_3{}^{-}+{2H}_2\end{array}$

(8) $\begin{array}{c}4-\mathrm{NP}+3\mathrm{Fe}\left(0\right)+{6H}^{+}\to 3\mathrm{Fe}\left(\mathrm{II}\right)+2H_{2}O+4-\mathrm{AP}\end{array}$

(9) $\begin{array}{c}4-\mathrm{NP}+4\mathrm{Mn}\left(\mathrm{II}\right)+6H{}^{+}\to 2\mathrm{Mn}\left(\mathrm{III}\right)+2\mathrm{Mn}\left(\mathrm{IV}\right)+2H_{2}O+4\mathrm{AP}\end{array}$

Comparison sample study

More crystalline BCS clays, as described in the Materials and Methods section, were also tested for their catalytic activity for a comparison purposes. The BCS contains a certain amount of Fe (Fe₂O₃, 9.40%). Two other types of clay minerals (Ilt and Mnt) with smaller Fe and Mn species contents were also used for comparison. The Fe and Mn contents in these clay minerals was less than those in PC. To increase the Fe content, commercial metallic iron powders (Fe(0)) and synthesized iron oxide (Fe₃O₄) nanoparticles were incorporated into BCS or other minerals to prepare composites, e.g. Fe(0)/BCS. The synthesis method for Fe₃O₄ is described in the SI. Only the composite of Fe(0)/BCS had certain catalytic performance, with a conversion efficiency of 89.97% after 360 min (Figs 11, S13). This result suggests that Fe(0) was the catalytic species. The curve could be fitted by first-order kinetics with k = 0.11 × 10^–3 s^–1, which was smaller than the PC. The results above confirmed the high catalytic ability of PC, which originates mainly from the amorphous Fe and Mn species in the pelagic clays.

Fig. 11 a Plot of the degradation of 4-NP using Fe(0)/BCS as the catalyst, and b plot of ln(C _t/C ₀) versus time, for the reduction of 4-NP using NaBH₄ as the reducing agent and Fe(0)/BCS as the catalyst