Borohydride-Assisted Surface Activation of Co3O4/CoFe2O4 Composite and Its Catalytic Activity for 4-Nitrophenol Reduction.

Surface activation of catalysts is known to be an efficient process to enhance their activity in catalytic processes. The activation process includes the generation of oxygen vacancies, changing the nature of the catalyst surface from acidic to basic and vice versa, and the reduction of catalyst surface by H2. On the other hand, magnetically separable catalysts are highly beneficial for their utilization in water or biological fluid-based catalytic processes, as they can be easily guided to the target site and recovered. Here, we present the fabrication of CoFe2O4 and composites of Co3O4/CoFe2O4/α-Fe2O3 and Co/CoFe2O4/α-Fe2O3 through solution combustion process to utilize them as catalysts for 4-nitrophenol (4-NP) reduction. Although none of the as-prepared CoFe2O4 and Co3O4/CoFe2O4 was seen to be active in 4-NP reduction reaction, the surface of the composite gets activated by borohydride (NaBH4) treatment to act as a highly active catalyst for 4-NP reduction. X-ray photoelectron spectroscopy of the composite revealed the formation of metal-hydroxide (M-O-H) species of both Co and Fe at its surface due to borohydride treatment. The mechanism of the surface activation and the dynamics of 4-NP reduction of the surface-activated composite have been studied, proposing a possible pathway for the reduction of 4-NP.

Introduction

Transforming nitroarenes into anilines is of great importance as anilines are among the most important intermediates used for the production of dyestuffs and pharmaceuticals.[1] 4-Nitrophenol (4-NP) is a toxic nitroarene, generated as a byproduct in the production of pesticides, synthetic dyes, leather preservative, and as intermediate in the production of paracetamol.[2,3] Furthermore, the reduction of 4-NP is a well-known model system commonly used to test the catalytic behaviors of metals and other materials that can reduce organic molecules.[4] To achieve 4-NP reduction, noble and non-noble metals have been frequently utilized as catalysts.[2,3,5] The inconveniences of using metals in these chemical reactions are the lack of selectivity in the reduction of nitroarenes and the high cost of noble metals (e.g., Au, Pt, Ru, Pd).

To address these inconveniences, metal/metal oxide composites have been investigated for the reduction of −NO2 group in nitroarenes. High selectivity of Ni/TiO2 or Au/TiO2 composites for the reduction of nitroarenes has also been demonstrated.[6] The high selectivity was attributed to the bonding of −NO2 group of 4-NP with metal oxides.[6] Likewise, hydrogenated MoO (H-MoO) loaded with bimetallic Pt–Sn (Pt–Sn/H-MoO) has also been utilized to hydrogenate 4-nitrostyrene.[7]

However, very few cobalt ferrite (CoFe2O4) or composites based on CoFe2O4 have been tested so far as catalysts for the reduction of 4-NP.[8,9] Some works claim that CoFe2O4 reduce the 4-NP, but the signal-to-noisy ratios in their X-ray diffraction (XRD) patterns were low, therefore, there is a possibility that the undetected impurities present in a small amount might be responsible for their catalytic activity.[10−12] One advantage of using CoFe2O4-based composites as catalysts is that they are magnetically separable. Cobalt ferrite (CoFe2O4) is an interesting ferrimagnetic material that has a high saturation magnetization.[13] Although metal oxides such as NiO and CuO have been utilized by several research groups as catalysts for the reduction of 4-NP,[14] rarely the catalysts have been analyzed by X-ray photoelectron spectroscopy (XPS) before and after their utilization in 4-NP reduction. The high-resolution spectra obtained by this technique provide relevant information regarding the chemical composition, oxidation states, adsorbed ions, and other characteristics of the oxide surfaces that can shed light on the nature of the catalyst surface required for the reduction of 4-NP.

On the other hand, during the synthesis of cobalt ferrite (CoFe2O4) nanoparticles by the solution combustion reaction, frequently small amounts of superparamagnetic (e.g., Co nanoparticles), nonmagnetic (e.g., α-Fe2O3, Co3O4, CoO), and amorphous byproducts are generated over their surfaces.[15] These byproducts might also be catalytically active in the reduction of 4-NP. In particular, the purity and size of the cobalt ferrite particles obtained using a solution combustion process depend on several factors such as: (i) oxidizing and reducing agent ratio,[16] (ii) annealing temperature, and (iii) annealing time. Apart from α-Fe2O3, CoO, and Co3O4, the generated byproducts can also contain the clusters or particles of cobalt, iron oxohydroxydes (e.g., FeOOH), amorphous materials, etc.,[10,12] the catalytic behaviors of which are unknown. Furthermore, a small amount of some of these byproducts may be detected through their catalytic activities, which is a relevant issue for applications in which pure CoFe2O4 is required. The utilization of CoFe2O4 as a heterogeneous catalyst has been reported in the literature,[17] and the strategy of using CoFe2O4 as a catalytic support is of great advantage due to its high chemical stability.

In the present work, we demonstrate that phase pure CoFe2O4 particles are not at all active as catalyst for the reduction of 4-NP. However, some byproducts such as Co3O4 generated during the preparation of CoFe2O4 particles by the solution combustion reaction manifest high catalytic activity after activating their surface with NaBH4 for about 4 h. The Co3O4-supported CoFe2O4 composites could be prepared by a low-temperature solution combustion process, controlling the nitrate ions/glycine (N/G) ratio in the reaction mixture. We demonstrate that the magnetic composite can be activated by NaBH4 during the reduction of 4-NP to 4-aminophenol, through the formation of metal hydroxides at its surface. The mechanism of surface activation and the reduction of 4-NP have been extensively studied utilizing X-ray photoelectron spectroscopy.

Results and Discussion

XRD Results

Figure presents powder X-ray diffraction (XRD) patterns of the cobalt ferrites prepared at four different N/G ratios. As can be seen in Figure a,b, CoFe2O4 prepared at N/G ratios 6 and 4.5 are in the phase pure spinel structure, and all their diffraction peaks coincide in intensity and position with the standard values reported in the PDF #4-006-4147. However, for the sample prepared at N/G ratio 3.0, there appeared a small peak at 2θ = 33° associated to the (104) plane (PDF # 00-033-0664) of α-Fe2O3. On the other hand, the sample prepared at a N/G ratio 2.0 revealed several diffraction peaks associated to α-Fe2O3 and Co3O4 (Figure d), confirming the formation of α-Fe2O3 and Co3O4 impurities during the synthesis of CoFe2O4 at lower N/G ratios. The two small diffraction peaks located at 2θ = 31.27 and 36.84° (indicated with the diamond symbols in Figure d) coincide in position with the (220) and (311) planes of Co3O4 in the cubic phase (PDF #04-005-4386). Formation of these two impurities is associated to the higher temperature and intense flame generated during the combustion reactions occurred for the N/G ratios 3.0 and 2.0. Unfortunately, owing to the fluorescence in the XRD patterns generated by the cobalt containing compound, it was not possible to estimate the particle/grain size of these byproducts from their XRD peaks.

Figure 1

XRD patterns of CoFe2O4 particles synthesized at different N/G ratios. (a) N/G: 6, (b) N/G: 4.5, (c) N/G: 3.0, and (d) N/G: 2.0.

Figure depicts the X-ray patterns of selected samples in the 2θ = 30–53° range with an enhanced signal-to-noise ratio (compared to Figure ). As can be noticed, the presence of cobalt nanoparticles could be detected in the CoFe2O4-3.0 sample from its diffraction pattern. The position of the diffraction peaked located around 44.30° matches perfectly with the position of the most intense diffraction peak of metallic cobalt (PDF # 00-015-0806). In addition, the diffraction peak located at 2θ = 36.97° in Figure associated to the (222) plane of Co3O4 is better resolved than its peak appeared in Figure . For comparison, the XRD pattern of Co3O4 is presented in Figure S1.

Figure 2

Slow-scanned XRD patterns of CoFe2O4 particles synthesized at different N/G ratios: 6.0, 3.0, and 2.0 in the 2θ = 30–53° range. XRD patterns of the impure samples prepared with the ratios 3.0 and 2.0 are also included after being used as a catalyst.

Scanning Electron Microscopy (SEM)

Typical SEM images of the CoFe2O4 particles prepared at four N/G ratios are presented in Figure . It can be seen that the CoFe2O4-6.0 and CoFe2O4-4.5 samples are composed of sub 100 nm particles, highly agglomerated due to the post-growth annealing at 600 °C. Interestingly, the CoFe2O4-3.0 and CoFe2O4-2.0 samples contain a good fraction of particles larger than 400 nm that were not found in the SEM images of the samples prepared at N/G ratios 6.0 and 4.5. These larger particles were generated in these samples as a result of the decomposition of NO3–, and glycinate (H2NCH2COO–) ions at high temperature reached inside the mixture under the intense flame. Consequently, the particles prepared at N/G ratios 3.0 and 2.0 grown faster during their annealing than the particles in samples prepared at N/G ratios 6.0 and 4.5. A common feature of these larger particles (Figure c,d) is their smooth facets.

Figure 3

Typical SEM images of the (a) CoFe2O4-6.0, (b) CoFe2O4-4.5, (c) CoFe2O4-3.0, and (d) CoFe2O4-2.0 samples.

Elemental mapping of the samples CoFe2O4-6.0, CoFe2O4-4.5, CoFe2O4-3.0, and CoFe2O4-2.0 are shown in Figures S2–S5 in the Supporting Information. The results confirmed that the constituent elements are distributed homogeneously in CoFe2O4. However, in Figure S5a, there appeared a zone in the sample (indicated by an arrow) where the amount of cobalt is lower than expected (Figure 4Sb). That zone could be a region of the sample containing the byproduct α-Fe2O3 previously detected in the XRD pattern.

N2 Adsorption–Desorption Isotherms and Surface Area

N2 adsorption–desorption isotherms of the cobalt ferrite samples recorded at 77 K are shown in Figure . As can be seen, although the sample CoFe2O4-6.0 exhibits type-II isotherm, the isotherms of the samples CoFe2O4-4.5, CoFe2O4-3.0, and CoFe2O4-2.0 are the mixture of type-II and type-IV. Although type-II isotherms are obtained for nonporous or macroporous adsorbents, type-IV isotherms contain hysteresis loops and are characteristic of mesoporous materials.[18] As we can see, for the CoFe2O4-6.0 sample, the adsorption and desorption branches of the isotherm almost coincide, which is the characteristic of a macroporous solid. In fact, some macropores in the CoFe2O4-6.0 sample were observed in its SEM image (Figure a). On the contrary, the isotherms for the CoFe2O4-4.5, CoFe2O4-3.0, and CoFe2O4-2.0 samples revealed hysteresis loops, which are associated to the capillary condensation inside the mesopores. The capillary condensation might have occurred in the voids of aggregated particles. Only for the CoFe2O4-4.5 and CoFe2O4-3.0 samples, the adsorption–desorption branches split at low pressure (p/p0 = 0.3), indicating the initiation of the capillary condensation in these samples at lower pressure.

Figure 4

N2 adsorption–desorption isotherms of cobalt ferrites synthesized at N/G ratios 6.0, 4.5, 3.0, and 2.0.

The specific surface area of the samples was calculated using the Brunauer–Emmett–Teller (BET) equationwhere Vm is the monolayer volume, L is the Avogadro constant, and σ is the cross-sectional area of the N2 molecule (0.162 nm2). Although the largest surface area was estimated for CoFe2O4-6.0, the second largest area was obtained for the CoFe2O4-4.5 sample (see Table ). On the other hand, the CoFe2O4-3.0 and CoFe2O4-2.0 samples had very low surface areas, as expected for the samples containing larger particles (Figure c,d). The large difference in surface area is due to the intense flame generated during the combustion in the synthesis of the later samples. As can be recalled, although there appeared no or low flame during the combustion for the samples prepared at N/G ratios 6.0 and 4.5, intense flames were generated for the samples prepared at N/G ratios 3.0 and 2.0. Consequently, the nitrate species were not fully decomposed during the combustion reaction in the case of N/G ratios 6.0 and 4.5. Instead, they were decomposed during the after-growth annealing, which is a slower growth and sintering process.

Table 1

Texture Parameters of CoFe2O4 Samplesa

sample	as (m2/g), BET	C	total pore volume (cm3/g), BET	pore volume (cm3/g), BJH-ads	pore area (cm2/g), BJH-ads
CoFe2O4-6.0	14.422	86.384	0.160	0.159	14.116
CoFe2O4-4.5	7.556	169.47	0.124	0.123	6.053
CoFe2O4-3.0	0.841	376.24	0.00857	0.00840	0.6057
CoFe2O4-2.0	2.064	39.476	0.0277	0.0275	2.043

as represents the specific surface area of the samples. C is approximately equal to exp{(E1 – EL)/RT}, where E1 = heat of adsorption of the first N2 layer, EL = heat of liquefaction of N2, R = constant of the gases, and T = temperature.[19]

Pore volume and pore area in the samples were also estimated by the Barrett–Joyner–Halenda (BJH) method using the adsorption branch of the isotherms and presented in Table . It can be noted that the total pore volume determined using both the BET and BJH methods is very similar for all of the samples.

Catalytic Activity

The catalytic reduction of 4-NP was tested for all of the four samples. The degradation of 4-NP was not observed for the CoFe2O4-6.0 and CoFe2O4-4.5 samples within the first 10 cycles (results presented in Figure S6 for the CoFe2O4-4.5 sample). Since for these two samples, only the CoFe2O4 phase was detected through the XRD analysis, we can conclude that the phase pure cobalt ferrite does not reduce 4-NP. However, the CoFe2O4-3.0 sample reduced all of the 4-NP to 4-AP in the first cycle and also in its subsequent reusability tests. The reduction of 4-NP to 4-AP could be affirmed from the progressive reduction of the intensity of the absorption band at 400 nm, characteristic of 4-NP, after the addition of NaBH4, and a subsequent growth of the 300 nm absorption band associated to the 4-aminophenol (4-AP) (Figure , cycle 1). During the reusability test of this sample, the reduction of 4-NP was accomplished in only 27 min (cycle 2 shown in Figure ). That is, half of the time needed in the cycle 1. Results similar to that presented in cycle 2 were obtained in the subsequent reusability tests (see cycles 3–5 in Figure S7).

Figure 5

UV–vis absorption spectra corresponding to the progressive reduction of 4-nitrophenol to 4-aminophenol using CoFe2O4-3.0 sample as a catalyst. Cycle 2 corresponds to the first reusability test.

Since the reduction reaction of 4-NP follows the Langmuir–Hinshelwood mechanism,[16,20,21] with a pseudo first-order kinetics, the expression ln(C/C0) = ln(A/A0) = −kt was used to determine the reaction rate constant (k). C and A stand for the concentration and absorbance of 4-NP at a given time (t) at 400 nm, and the subscript 0 represents the time zero. The ln(A/A0) versus time plots presented in Figure show that the k values for the reusability cycles 2–5 vary in-between 0.09 and 0.14 s–1.

The results of the first seven cycles of 4-NP degradation by the sample CoFe2O4-2.0 are presented in Figures and S8. As can be noticed, the degradation of 4-NP did not take place during the first 4 cycles. However, in the 5th cycle (after about 40 min), a slow reduction of 4-NP commenced. Despite of recording the UV–vis spectra of the solution during the first 100 min, only a small amount of 4-NP was reduced in the cycle 5. In cycle 6 (after about 60 min), a higher amount of 4-NP was reduced to 4-AP; the intensity of the 400 nm absorption band decreased roughly 50%. However, as can be noticed from Figure , in cycle 7, the 4-NP was fully reduced within 12 min, in the presence of the catalyst. Despite that the complete degradation of 4-NP occurred only until the cycle 7 for the CoFe2O4-2.0 catalyst, the time needed for the complete reduction was much shorter (12 min) than the time needed (27 min) for the CoFe2O4-3.0 catalyst. This difference was probably due to the presence of a larger amount of Co3O4 in the CoFe2O4-2.0 sample (see Figure ).

Figure 6

UV–vis absorption spectra correspond to the progressive reduction of 4-nitrophenol to 4-aminephenol using CoFe2O4-2.0 sample as a catalyst, during cycles 1–7. Apparent reaction rate constants (k) are included for the selected cycles.

Notably, during the catalytic reduction of 4-NP by CoFe2O4-2.0, a huge amount of bubbles evolved from the surface of the particles from cycle 5; these bubbles have the ability of propelling the aggregates of particles in the solution, even upward, as shown in the video in the Supporting Information. The observation that the 4-NP degradation by CoFe2O4-2.0 starts only until the 5th cycle strongly suggests that the chemical composition of the surface of one of the impurities changed and became active for 4-NP degradation. To verify that, X-ray photoelectron spectra of the CoFe2O4-3.0 and CoFe2O4-2.0 samples were recorded before and after their utilization in 4-NP degradation. These results are discussed in the Section . Regarding the evolution of bubbles, a flame test on the gas generated at the top of the vial indicated the formation of H2 gas due to the reaction. This agrees with the reported observation that Co3O4 is an efficient catalyst for NaBH4 hydrolysis to generate hydrogen.[22]

The ln(A/A0) versus time plots presented in Figure show that the k values for the reusability cycles 7–10 vary in-between 0.25 and 0.34 s–1. These k values for the CoFe2O4-2.0 catalyst are roughly twice of the same for the CoFe2O4-3.0 catalyst.

X-ray Photoelectron Spectroscopy (XPS)

Figure a presents the XPS survey spectra for the CoFe2O4-3.0 sample before and after its use in the reduction of 4-NP. The revealed peaks correspond to the emissions from Co, Fe, and O and some of their Aüger emission lines. In addition, there appeared an emission associated to the adventitious carbon, adsorbed at the surface of the sample. In both the spectra (before and after use in 4-NP reduction), the peaks of the C 1s emission were set at 284.5 eV to compare the binding energy (BE) of the remaining orbitals with their literature-reported values. Moreover, there appeared an emission peak around 191.76 eV for the CoFe2O4-3.0/catal (used in 4-NP reduction) sample, associated to borate anions such as [B4O7]2– or BO2– generated by the reaction of NaBH4 with the surface of cobalt ferrite, the traces of α-Fe2O3, or an amorphous material present in the sample. High-resolution spectra for boron and sodium elements are presented in Figure S9 of the Supporting Information. The appearance of the emission associated only to borate anions without any signal from Na+ ions in the CoFe2O4-3.0/catal sample indicates the adsorption or formation of chemical bonds with borate anions at the surface of the catalyst.

Figure 7

XPS spectra of the CoFe2O4-3.0 and CoFe2O4-3.0/catal (used in 4-NP reduction) samples. Survey (a) and high-resolution spectra of selected orbitals: (b) Co 2p1/2 and Co 2p3/2, (c) Fe 2p1/2 and Fe 2p3/2, (d) O 1s, (e) O 2s, and (f) C 1s.

The core-level spectra of Co 2p1/2, Co 2p3/2, Fe 2p1/2, Fe 2p3/2, O 1s, O 2s, and C 1s orbitals for the CoFe2O4-3.0 and CoFe2O4-3.0/catal samples are depicted in Figure b–f. Binding energy (BE) positions of the corresponding peaks are presented in Table . For comparison, survey and core-level spectra of these orbitals for the CoFe2O4-6.0, CoFe2O4-6.0/catal, CoFe2O4-4.5, CoFe2O4-4.5/catal, and Co3O4 samples are shown in Figures S10–S12. From Table and the XPS peaks associated with O, Co, and Fe (Figure b–e), it is evident that all of the peaks in the CoFe2O4-3.0/catal sample are shifted toward higher BE compared with the same in the CoFe2O4-3.0 sample. For example, the positions of the Co 2p3/2 and Fe 2p3/2 peaks were shifted roughly about 1 eV; similarly, the positions of the peaks corresponding to O 1s and O 2s orbitals were shifted about 2 eV. The observed shifts can be attributed to a considerable increase in −OH groups bonded to the cobalt and iron cations at the surface of the sample.[23] It should be noted that the cobalt shake-up lines depicted in Figure b for the Co 2p1/2 and Co 2p3/2 orbitals coincide in positions for both the samples. The maximum of the Fe 2p3/2 XPS peak in Figure c for the sample CoFe2O4-3.0 is centered at 710.8 eV, indicating that the chemical state of iron is Fe3+. Interestingly, the maximum of the Fe 2p3/2 XPS peak in Figure c for the sample CoFe2O4-3.0/catal is located at 711.5 eV, which coincides with the reported BE for the FeOOH phase.[23] The results of the deconvolution of the spectra correspond to Co 2p1/2 orbital for the CoFe2O4-3.0 and CoFe2O4-3.0/samples are presented in Figure S13 and Table S1. For comparison, the spectrum for pure cobalt ferrite (i.e., CoFe2O4-6.0) was also included. A discussion of the fit peak components of the Co 2p1/2 orbital for these samples is included in the Supporting Information.

Table 2

Binding Energies (BEs, eV) of the Representative XPS Emissions for the Cobalt Ferrite Samples

sample/orbital	Co 2p3/2	Co 2p1/2	Fe 2p3/2	Fe 2p1/2	O 1s	O 2s
CoFe2O4-6.0	779.77	795.46	710.56	724.01	529.62
CoFe2O4-6.0/catal	779.37	795.27	710.13	723.75	529.26
CoFe2O4-4.5	779.77	795.73	710.56	724.05	529.70
CoFe2O4-4.5/catal	779.66	795.38	710.38	723.84	529.50
CoFe2O4-3.0	780.17	795.73	710.78	724.32	529.72	21.65
CoFe2O4-3.0/catal	781.15	797.04	711.53	724.89	531.17	23.78
CoFe2O4-2.0	780.06	795.61	710.56	724.06	529.46	21.34
CoFe2O4-2.0/catal	780.91	796.87	711.49	724.89	531.14	23.56
Co3O4	779.42	794.76			529.22
Co3O4/catal	780.57	796.53			530.86

The core-level O 1s spectrum for the CoFe2O4-3.0 and the other cobalt ferrites samples is presented in Figures d and S14. The deconvolution of the O 1s peak, corresponding BEs and areas (%), is presented in Figure S14 (Supporting Information) and Table . The deconvolution corresponding to CoFe2O4-3.0 sample generated four peaks labeled as fit peaks 1–4, two intense peaks located at 529.5 and 530.0 eV, and two low-intensity peaks located at 530.8 and 531.5 eV. Fit peaks 1 and 2 located at BEs 529.5 and 530.0 eV correspond to M–O–M {M = Co(II), Co(III), Fe(III)} chemical bonds present in the CoFe2O4 lattice. On the other hand, fit peaks 3 and 4 located at BEs 530.8 and 531.5 eV correspond to O 1s orbital in M–O–H chemical bonds. Although the area of the fit peak 1 plus fit peak 2 accounts for ∼69% of the area of O 1s peak, the area of the fit peak 3 plus fit peak 4 accounts for the remaining 31%.[23] This deconvolution allows us to conclude that on the surface of the CoFe2O4-3.0 sample, roughly 69% of the oxygens atoms is forming chemical bonds with cobalt and iron cations in CoFe2O4, α-Fe2O3, and roughly 31% is oxygens atoms corresponding to metal hydroxides (M–OH).

Table 3

Binding Energy (BE, eV), Full Width at Half-Maximum (FWHM), and Area (%) of the Components of the O 1s Peaks

	CoFe2O4-6.0			CoFe2O4-6.0/catal			CoFe2O4-4.5			CoFe2O4-4.5/catal
peaks	BE	FWHM	area	BE	FWHM	area	BE	FWHM	area	BE	FWHM	area
peak 1	529.5	1.2	62.6	529.2	1.3	62.8	529.5	1.1	42	529.2	1	35
peak 2	529.9	1.1	25.3	530.1	1.5	12.7	529.9	1.1	42.2	529.7	1	40.5
peak 3	530.8	1.5	6.3	531.2	1.5	18.6	530.9	1.1	5.6	530.7	1.2	8.8
peak 4	531.6	1.5	5.8	532.3	1.2	5.9	531.8	1.6	10.2	531.9	1.8	15.8

On the other hand, the O 1s emission band of the CoFe2O4-3.0/catal sample could be deconvoluted in three components (fit peaks) with binding energy positions of the component bands at 529.8, 531.1, and 532.3 eV (see Figure S14f and Table ). Although the first component (fit peak 1 at 529.8 eV) corresponds to M–O–M {M = Co(II), Co(III), Fe(III)} chemical bonds in CoFe2O4, the second and third components (fit peaks 2 and 3 located at 531.1 and 532.3 eV) coincide with the BEs of O 1s orbitals in metallic hydroxides, for which typical reported values vary in-between 530.9 and 532 eV.[23] As the reported BE value for O 1s orbitals in borates (e.g., Na2B2O7) is around 531.1 eV,[24] which coincides with the estimated BE value of the second component band (fit peak 2), the BE values of fit peaks 2 and 3 (531.1 and 532.3 eV) were also compared with the reported BE values for the O 1s orbital in Co(OH)2 hydroxide, α-FeOOH oxohydroxide, and borate anions, which could be formed on the surface of the sample after its reaction with NaBH4.

In the case of Co(OH)2, the peak corresponding to the O 1s orbital was fitted by Biesinger et al., to two components with BE values 531.07 (86%) and 532.25 eV (14%).[25] These components were attributed to hydroxide (OH–) groups and H2O, respectively.[25] Values in parenthesis in this section correspond to the percentage of the area in the XPS peak. In the case of α-FeOOH oxohydroxide, the O 1s peak in the reported spectrum was deconvoluted in three components with binding energies 529.99 (39%), 531.20 (49%), and 532.50 eV (12%).[25] These three peaks were assigned to O2– ions in the lattice, hydroxide, and H2O, respectively. These analyses allow to state that the fit peaks shown in Figure S14f located at 529.8 eV (26.2%), 531.1 eV (44.1%), and 532.3 eV (29.7%) correspond to O2– anions in the lattice of CoFe2O4, iron or cobalt hydroxide, and H2O absorbed by the capillary condensation in the pores, respectively. These hydroxides in the CoFe2O4-3.0/catal sample were formed due to the partial reduction of the sample surface by NaBH4.

Since the surface of the CoFe2O4-3.0/catal sample contains borate anions and Fe(III) cations (probably making chemical bonds between them, i.e., Fe–O–B), we compared the BE of the O 1s and B 1s orbitals in this sample with their reported BEs in iron–sodium borate glasses. The binding energies of O 1s and B 1s orbitals in a glass composed of B (21.8%), Na (12.2%), Fe (4.2%), Al (6.7%), and O (55.1%) are reported to be at 531.0 and 191.8 eV,[24] respectively. These BE values coincide well with the corresponding peak positions in the CoFe2O4-3.0/catal sample (see Figure S14f for O 1s orbital and Figure S9 for B 1s orbital), indicating a probable formation of Fe–O–B or Co–O–B chemical bonds at the surface of the cobalt ferrite sample after its utilization in 4-NP degradation.

Figure depicts the survey and high-resolution XPS spectra of selected orbitals for the CoFe2O4-2.0 and CoFe2O4-2.0/catal samples. The shapes and positions of the emission bands revealed for these samples are very similar to those discussed for the CoFe2O4-3.0 and CoFe2O4-3.0/catal samples, respectively (Figure ). In the CoFe2O4-2.0 sample, α-Fe2O3 and Co3O4 byproducts made a small contribution to the area of the O 1s peak around 530 eV (fit peak 2 in Figure S14g). The reported BE values for the O 1s orbital in α-Fe2O3 are ∼530,[23] 529.8,[26] and 529.88 eV.[25] On the other hand, although Dupin et al. deconvoluted the emission band of O 1s orbital emanating from Co3O4 in two components with BEs 529.5 (72%) and 531.5 eV (28%),[27] Biesinger et al. deconvoluted this XPS band in three components with BEs 529.95 (53%), 530.84 (41%), and 532.66 (6%).[25] Although the component at 529.95 eV was associated to O2– anions in the Co3O4 lattice, the component at 530.84 eV was attributed to cobalt hydroxide, hydrated or defective Co3O4.[25] The discussion presented in the following two paragraphs further supports the claim that the surfaces of the two samples (CoFe2O4-3.0 and CoFe2O4-2.0) that exhibited a good catalytic activity after borohydride activation were partially transformed into hydroxides of cobalt and iron. For comparison, the emission band of the O 1s orbital in CoFe2O4-6.0 and CoFe2O4-4.5 is shown in Figure S14a–d (Supporting Information).

Figure 8

XPS spectra of the CoFe2O4-2.0 and CoFe2O4-2.0/catal samples. Survey (a) and high-resolution spectra of selected orbitals: (b) Co 2p1/2 and Co 2p3/2, (c) Fe 2p1/2 and Fe 2p3/2, (d) O 1s, (e) O 2s, and (f) C 1s.

On the other hand, the Fe 2p3/2 orbital for a cobalt ferrite sample could be deconvoluted in three peaks (Figure S15 and Table S2) centered around 709.5 (35%), 710.7 (42.6%), and 712.4 (22.4%) eV. The latter peak is attributed to the LMM Aüger line of the cobalt atom, and the earlier two peaks account for the Fe3+ cations in the CoFe2O4 lattice.[23]

Fourier Transform Infrared Spectroscopy (FTIR)

FTIR spectra of the samples before and after use in catalysis are shown in Figure S16. As can be seen in Figure S16i, the spectra of all of the samples exhibit an IR band around 3450 cm–1, associated to O–H vibrations of M–O–H bonds or H2O physisorbed at their surface. All of the samples revealed an intense IR band at 584 cm–1, which has been associated to the M–O vibration in CoFe2O4.[28] An amplification (zoomed-in) of the Figure S16i in the lower frequency region is shown in Figure S16ii. As can be noticed, there appeared a week shoulder around of 530 cm–1 for the CoFe2O4-2.0 and CoFe2O4-2.0/catal samples that can be associated to α-Fe2O3 impurity.[29] In fact, the IR absorption band of the α-Co(OH)2 phase has been reported to be at 634 cm–1.[29] However, we could not detect any band around 634 cm–1 (Figures S16ii,iii) in the samples to confirm the presence of α-Co(OH)2 in them either before or after utilization in 4-NP reduction.

Mechanism of 4-Nitrophenol (4-NP) Reduction

The results in the present investigation clearly demonstrate that CoFe2O4 does not reduce 4-NP. To test whether the byproduct α-Fe2O3 is able to reduce 4-NP, a sample of nickel ferrite was prepared using the same procedure utilizing N/G ratio 3 in the reaction mixture, to obtain α-Fe2O3/NiFe2O4 composite. The purpose of obtaining NiFe2O4 particles was to have a cobalt free sample. The composite was also treated thermally at 600 °C for 2 h. The XRD pattern recorded for this composite confirmed the presence of both α-Fe2O3 and NiFe2O4 since the XRD peaks coincide in position and intensity with their standard values (PDFs 04-006-6579 and 00-054-0964, Figure S17). Testing this α-Fe2O3/NiFe2O4 composite in 4-NP degradation revealed its inability for 4-NP reduction (results not presented); confirming that hematite (α-Fe2O3) is not active for the degradation of 4-NP. Therefore, the species responsible for 4-NP degradation observed for the CoFe2O4-3.0 sample are small cobalt nanoparticles, which were detected in its XRD spectrum (Figure ). Previous studies have also reported the reduction of 4-NP with cobalt nanoparticles.[30,31] Note that the diffraction peak located at 44.30° persists in the CoFe2O4-3.0 sample even after its use in catalysis (Figure ). In fact, cobalt NPs are the known active species for the reduction of 4-NP, which can be formed in solution combustion reactions when the N/G ratio is low.[32] For instance, mixing Co(NO3)2 and glycine at a N/G ratio of 0.75 in the solution combustion process, Khort et al. obtained a mixture of Co (6%), CoO (56%), and Co3O4 (38%).[32] On the other hand, 4-NP reduction ability of α-Co(OH)2 has been demonstrated by Khan et al. for a sample containing α-Co(OH)2, although the signal/noise ratio of the X-ray diffraction peaks of this phase was extremely low.[33] Finally, the possible contributions of amorphous CoFe2O3.66 and amorphous Fe40Co60O phases in 4-NP reduction cannot be discarded as both of them have been utilized successfully in the catalytic oxidation of water and oxygen reduction reactions revealing superior performances than their respective crystalline counterparts.[34,35]

In the case of the CoFe2O4-2.0 sample, the Co3O4 byproduct is responsible for 4-NP reduction. Some previous works also concluded that Co3O4 reduces 4-NP to 4-AP, which further supports the results found in this work.[27,36−38] Based on the XPS results discussed in this work and the fact that the 4-NP reduction started only until cycle 5, the possible mechanism of 4-NP reduction is depicted schematically in Figure . In step 1, the cobalt–water moieties at the surface of the Co3O4 are transformed to CoO(OH) through the reaction among the hydride anion (H–) provided by the NBH4 and the chemisorbed H2O, followed by redox reactions. In other words, the Co(III) cations at the surface of the Co3O4 particles are reduced by NaBH4 to Co(II), as indicated in step 1. This could be possible as the redox potential of NaBH4 is high (approximately −0.6 V).[39] Also, the nitro group (−NO2) of 4-nitrophenolate ion bonds to the cobalt cations in the Co3O4. Then, in the step 2, a highly reactive hydride anion (H–) is transferred from NaBH4 to the nitrogen atom in the 4-nitrophenolate ion and, simultaneously, occurs in the N–O single bond breaking. After that a second H– anion reacts with the N=O moiety and, hence, induces the N=O bond cleavage, generating 4-(hydroxylamino)phenol. In step 3, the 4-(hydroxylamino)phenol is reduced to 4-AP in the presence of H2.[40] Finally, the remaining B–H bonds are hydrolyzed with water to generate H2 and Co–O–B(OH)3 species at the surface of the reduced Co3O4 (step 4).

Figure 9

Proposed mechanism of 4-nitrophenol (4-NP) reduction inspired from Figure 8 in ref (41).

Conclusions

In summary, we demonstrate the fabrication of both phase pure and impure CoFe2O4 particles through the solution combustion process by controlling the nitrate ions/glycine (N/G) ratio in the reaction. Although phase pure CoFe2O4 particles of sub-micrometer sizes can be fabricated using N/G ratios of 6.0 or 4.5 in the reaction mixture, at N/G ratio 3.0, CoFe2O4 particles with traces of α-Fe2O3 and cobalt are generated. On the other hand, at N/G ratio 2.0, CoFe2O4 particles are generated along with large quantities of α-Fe2O3 and traces of Co3O4 byproducts. The generation of the intense flame and high temperature during the solution combustion process at lower N/G ratios such as 3.0 and 2.0 induces the formation of bigger particles (>400 nm) in the samples. Although phase pure CoFe2O4 particles are not active as catalysts for the reduction of 4-NP, the Co3O4 formed over CoFe2O4 (in the sample synthetized with a N/G ratio of 2.0) is capable of reducing 4-nitrophenol (4-NP) to 4-aminophenol (4-AP) in the presence of NaBH4. The required times for the complete reduction of 4-NP for the CoFe2O4 samples prepared at N/G ratios 3.0 and 2.0 were 27 and 12 min, respectively. We demonstrate that the Co3O4 supported on FeO2O4 particles is a low-cost and magnetically separable catalyst useful for the degradation of 4-NP. However, it requires a prolonged (∼4 h) activation in the aqueous solution of NaBH4. 4-NP starts to be reduced only until a considerable amount of bubbles are generated at the surface of the catalyst. At the surface of the catalytically active cobalt ferrite samples, oxides {M–O–M; M = Co(II,III), Fe(III)} are reduced to hydroxides by NaBH4. Borate anions generated from the NaBH4 oxidation in water chemically bond to the surface of the cobalt ferrite.

Experimental Section

Reagents and Equipment

The reagents used for the synthesis of cobalt ferrite particles and 4-NP degradation tests performed in this work were cobalt nitrate hexahydrate (Co(NO3)2·6H2O, Sigma, 99.99%), iron nitrate nonahydrate (Fe(NO3)3·9H2O, Sigma, 99.99%), glycine (H2NCH2COOH, Aldrich, 99%), diluted nitric acid (HNO3, J.T. Baker, 66%), 4-nitrophenol (O2NC6H4OH, Sigma, ≥99%), sodium borohydride (NaBH4, Sigma, 99%). All of the reagents were utilized as received, without further processing or purification. Deionized water from a Millipore water purification system (ρ > 1018 Ohm.cm) was utilized to dissolve the reagents and the 4-nitrophenol to carry out its catalytic tests. XRD patterns of the samples shown in Figure were recorded in a Bruker D8 Discover X-ray diffractometer, equipped with a LynxEye detector, using an acrylic sample holder. The spectra were recorded at 0.02°/step and 0.3 s acquisition time at each point. Fluorescence was diminished in the diffractogram by adjusting the discrimination lower level at 0.18 V in the detector of the equipment.

Also, the XRD patterns shown in Figure in the range of 2θ = 30–53° were recorded in a Bruker D2 Phaser diffractometer equipped with a LynxEye detector, using a low-noise silicon sample holder. The measurement conditions in the D2 diffractometer were 0.01°/step and 3 s acquisition time at each step. The fluorescence was not removed. A JEOL JSM-7800F field-emission scanning electron microscope (SEM) operating at 3.0 kV was utilized for the morphology evaluation of the particles. Nitrogen adsorption–desorption isotherms of the CoFe2O4 samples were measured at 77 K using a texture analyzer instrument (BELSORP Mini-II System). The outgassing of the samples was done at 10 Pa pressure and 300 °C temperature for 12 h. For catalytic tests, a Shimadzu UV-3101PC double-beam spectrophotometer was utilized. An X-ray photoelectron spectrometer (XPS, Thermo Scientific) with Al Kα (1486.6 eV) radiation source was utilized to analyze the surface composition of the composites. The deconvolution of the O 1s core-level emission bands was performed using Pseudo-Voight2 functions with 70% Gaussian and 30% Lorentzian components, after subtracting the Shirley-type background. FTIR spectra were recorded in a Perkin-Elmer Frontier spectrometer over the cold-pressed pellets made of 2% ferrite samples in 98% dry KBr powder. ATR-FTIR spectra of the powder samples were recorded in a Perkin-Elmer Spectrum One spectrometer, using the GladiATR accessory (PIKE Technologies).

Synthesis of Cobalt Ferrites

In a typical synthesis of CoFe2O4 particles, 3 mmol of Co(NO3)2·6H2O, 6 mmol of Fe(NO3)3·9H2O, and 12.9 mmol of glycine were dissolved in 70 mL of deionized water in a 600 mL beaker under magnetic stirring. Then, 1 mL of HNO3 was added to the mixture, similar to our previously reported work.[16] The volume of HNO3 was varied to attain nitrate ions/glycine (N/G) ratios of 6, 4.5, 3, and 2. The prepared mixture solutions were heated at 85 °C (under magnetic stirring) to evaporate the water. On evaporating all of the water from the mixture, an ignition occurred, and in some cases, a flame inside the beaker could be observed. The combustion reaction produced a black powder. The powder samples obtained in the solution combustion process were annealed in an air atmosphere at 600 °C for 2 h, inside a tubular furnace, using a heating ramp of 2 °C/min. A similar procedure was adapted for the synthesis of Co3O4 particles with a (N/G) ratio of 2.

Catalytic Reduction of 4-Nitrophenol (4-NP) to 4-Aminophenol

4-NP solution (4 mL, 0.2 mM) was mixed with 12 mL of deionized water in a 35 mL vial. Then, 15 mg of NaBH4 was added to the solution, and its UV–vis absorption spectrum was recorded. After that 10 mg of the cobalt ferrite sample was added to the solution under shaking. The absorption spectra of the reaction mixture were recorded every 2 or 3 min, isolating the catalyst from the reaction mixture by a bar magnet. After the first cycle of the catalytic test, the sample was held at the bottom of the vial with a magnet, and the liquid (supernatant) inside the vial was removed (discarded). Then, another 4 mL of 4-NP solution (0.2 mM) was mixed with 12 mL of deionized water and 15 mg of NaBH4. After that a UV–vis absorption spectrum was recorded. Finally, the solution was added to the 35 mL vial containing the used catalyst, and the UV–vis absorption spectra were recorded on the aliquots extracted from the reaction mixture every 2 or 3 min.