Introduction

The textile industry occupies a very important place in the world economy. To satisfy customers and increase the turnover, very stable synthetic dyes are used to preserve the color of the textiles against the effect of ultraviolet sun irradiation and the oxidation from detergent products. These synthetic dyes which are generally high molecules of molar mass, and containing several benzene nuclei (Song and Li 2009), can create harmful impact on the environment (El Bouraie and El Din 2016), particularly on surface waters and soils, because they are resistant to biological degradation (Imran et al. 2015). Several textile industries from the developing countries which are unable to treat process wastewater often get rid of them directly, thus polluting the surface waters (Dellamatrice et al. 2016; Rajasimman et al. 2017).

In order to treat the textile effluents and degrade the harmful dyes, research efforts have concerned the treatment techniques based on advanced oxidation processes at both laboratory and industrial scales (Baldrian et al. 2006; Yu et al. 2016). These processes such as photocatalysis using TiO2/UV, electro-oxidation, catalytic ozonation, sono-chemistry and homogeneous Fenton are based on the production of very oxidant radical species able to reduce the contaminants in biodegradable compounds or even in CO2 and H2O (Ahmadi et al. 2017; Garza-Campos et al. 2016). Among these processes, several are expensive, while the Fenton process is financially more attractive because it requires for its implementation only the use of Fe2+ or Fe3+/H2O2 reagents to generate OH· hydroxyl radicals (De Lima Perini et al. 2013; Wang et al. 2014). In order to avoid the recovery of iron ions in the treated effluent, studies have focused on the supporting of iron nanoparticles on porous materials (silica, zeolite, graphene oxide, clay, activated carbon, etc.) (Diao et al. 2017; Guo et al. 2017; Liu et al. 2013; Rache et al. 2014) or on the use of materials (laterite, graphite waste) (Cuiping et al. 2012; Khataee et al. 2015; Wang 2011) naturally rich in iron.

Recent work has been devoted to the development of heterogeneous catalysts by co-precipitation of iron oxide nanoparticles on porous materials to form iron oxide deposit (Hassan et al. 2016; Kakavandi et al. 2016; Yu et al. 2016). This iron oxide deposit may considerably reduce the specific surface area and pore volume of the support (Yu et al. 2016). In fact, the low accessibility of the dyes to the pores by adsorption in the Fenton reaction can slow down the degradation speed (Yu et al. 2016). By contrast, a rational occupation of the pores of support and a good dispersion of the iron nanoparticles could raise the efficiency of the heterogeneous catalyst (Duarte et al. 2012; Liu et al. 2017). In an ideal catalyst, the surfaces defined by the micropores and mesopores must be accessible both to the hydrogen peroxide (H2O2) for the production of hydroxyl radicals and to pollutants submitted to the degradation. The development of an appropriate supported catalyst makes possible the reduction in the distance between the adsorbed pollutant and the produced hydroxyl radicals which have only a half-life time of a few nanoseconds and a migration distance of about one micrometer (Bačić and Mojović 2005).

Some authors have recently reported the efficient catalytic activity of iron oxide nanoparticles highly dispersed within porous materials for the removal of pollutants (Castro et al. 2009; Mahy et al. 2014; Mazilu et al. 2017). Among the porous materials able to support iron nanoparticles, activated carbon has interesting properties in terms of high surface area, surface chemistry, stability in acidic or basic media and very wide spectrum of pore size distribution (micropores, mesopores and macropores) according to its activation process type (El-Shafey et al. 2016). For example, the work of Nidheesh and Rajan (2016) reported on the degradation of Rhodamine B in an aqueous solution by the heterogeneous Fenton process using commercial activated carbon to support iron oxide nanoparticles. Generally, activated carbon as a catalytic support is used as a powder and crushing into smaller fraction damages some of the pores already formed during pyrolysis, as underlined by Yacob et al. (2008). Therefore, it is critical to know specifically the type of the pores where these iron oxide nanoparticles would be accessible and more active, which remains very little addressed in the work of heterogeneous catalysis on this type of materials.

In this study, an activated carbon has been prepared in the optimum conditions from the banana spike to develop a high micromesoporosity. Further, it has been used to support iron oxide nanoparticles at different rates of wet impregnation to obtain three heterogeneous catalysts. As-prepared heterogeneous catalysts were characterized by nitrogen and carbon dioxide adsorption–desorption measurements, atomic absorption, X-ray photoelectron spectroscopy (XPS) and transmission electron microscopy (TEM). The catalysts were tested in an aqueous solution in a batch reactor for the heterogeneous Fenton degradation of Rhodamine B, a xanthene dye, used as a model dye because of its high stability and known to undergo a natural anaerobic degradation in streams yielding the production of potentially carcinogenic aromatic amines (Gao et al. 2015; Rochat et al. 1978). The influence of the dispersion of iron nanoparticles within the pores of the heterogeneous catalysts obtained at different impregnation rates was studied on the removal of Rhodamine B by adsorption and Fenton degradation. The influence of the Fenton degradation parameters, such as initial concentrations of H2O2 and Rhodamine B, catalyst amount, pH and temperature, was studied on the best catalyst. Removal efficiency of Rhodamine B was evaluated by the rate of discoloration of the solution, the reduction in total organic carbon (TOC) and total nitrogen (TN).

Materials and methods

Reagents

The reagents used are of analytical grade: Rhodamine B (≥ 97%, Sigma-Aldrich, chemical formula C28H31ClN2O3, absorption wavelength λ = 554 nm, color index number 45,170 and dimensions: length 1.5 nm; width 0.98 nm; and thickness 0.43 nm (Canning et al. 2014)); H2O2 (30 wt%, Sigma-Aldrich); FeSO4,7H2O (99.5 wt%, Merck); NaOH (Sigma-Aldrich, 0.1 mol/L); Na2SO3 (≥ 99%, Sigma-Aldrich); HCl (37 wt%, Sigma-Aldrich); H3PO4 (85 wt%, Sigma-Aldrich). pH of the Rhodamine B solutions were adjusted using 0.1 mol/L HCl or NaOH aqueous solutions. Distilled water was used to prepare the synthetic solutions of Rhodamine B as well as the other analytical solutions.

Preparation of catalysts

The micromesoporous activated carbon (AC) used to support the iron nanoparticles was obtained by chemical activation (Briton et al. 2018). Thus, three catalysts were prepared by the impregnation of AC (4 g) in iron sulfate aqueous solutions (200 mL). The iron mass percentages of the solutions in which AC was dispersed were of 3%, 7% and 11%, respectively. Each dispersion was stirred for two hours in closed Erlenmeyer flasks at 300 rpm to adsorb the Fe2+ ions in the AC pores. Then, the dispersions were dried both using a rotary evaporator (Büchi Rotavapor R-200) at 100 °C and in an oven for 1 h at 105 °C. They were further heat-treated (under oxygen limited atmosphere) at 350 °C in a muffle furnace for 2 h (heating rate of 10 °C/min). The catalysts were crushed and sieved of at 125 µm. The catalysts prepared were designated as AC@Fe/1, AC@Fe/2 and AC@Fe/3 according to the increasing iron impregnation ratios: 3%, 7% and 11%.

Characterization of the catalysts

Specific surface area and porosity of the heterogeneous catalysts were determined by using a “Micromeritics ASAP 2020” sorptometer through adsorption–desorption measurements of N2 at 77 K and CO2 at 273 K. The specific surface area was calculated according to the Brunauer–Emmett–Teller equation (BET) considering the relative pressure range of 0.01–0.05 as for microporous materials (Kaneko and Ishii 1992).

The pore size distributions (PSDs) were determined by fitting N2 adsorption isotherms at 77 K and CO2 adsorption isotherms at 273 K with, respectively, a finite slit pore model and an infinite slit pore model, calculated by the non-local functional density theory (NLDFT) (Jagiello and Olivier 2009). The volume of the supermicropores (diameter between 0.7 and 2 nm) and of mesopores was calculated from the N2 adsorption. The volume of ultramicropores (diameter below 0.7 nm) was obtained from the PSD from CO2 adsorption.

XPS was used to characterize the surface chemistry of the catalysts. The binding energy of the chemical elements, C, O, S, P and especially Fe, was determined in order to investigate the chemical nature of iron compound formed in the pores. XPS spectra have been obtained with an ESCALab 220i-XL (Fisons Instruments) using the AlKα monochromatic source and a hemispherical analyzer. Thus, high-resolution (0.1 eV) spectra have been obtained on the energy ranges 123–142 eV (P2p), 158–174 eV (S2p), 275.1–299.9 eV (C1s), 525.1–544.9 eV (O1s) and 700.1–739.9 eV (Fe2p) with a passing energy of 20 eV.

A mapping of the chemical elements of the iron-based AC@Fe/3 catalyst has been achieved using a ZEISS ULTRA 55 Gemini field emission gun scanning electron microscope (FEG-SEM) coupled to an EDAX-type energy-dispersive spectrometer (EDS).

The TEM images of the catalysts were obtained using an EM-002B (TOPCON Co.) microscope equipped with a LaB6 thermal filament at 120 kV acceleration voltage. For the TEM observations, the powder samples have been dispersed in water under ultrasound and then fixed on a copper grid.

DRX diagrams were obtained using the capillary powder method with the INEL diffractometer. The use of a position detector sensitive to the INEL CPS 120 curve allowed simultaneous recording of the diffracted intensity over a range of 3°–90° 2θ with a 0.03° pitch. The radiation used is a monochromatic beam CoKα1.

The iron content in the catalysts was determined by atomic absorption. The catalyst was firstly calcined and its ash dissolved in 5 mL of 37 wt% HCl. Then 20 mL of distilled water were added to the solution further boiled to one-third of its volume. After cooling, the solution filtered on ordinary paper and completed to 50 mL with distilled water was analyzed using a Varian SpectrAA-20 brand flame atomic absorption spectrometer. The total iron content is evaluated relative to the mass of AC@Fe according to the following formula:

$$\% {\text{Fe}} = \frac{{\left[ {{\text{Fe}}^{2 + } \& {\text{Fe}}^{3 + } } \right]V}}{{m_{{{\text{AC}}@{\text{Fe}}}} }} \times 100$$
(1)

where \(m_{{{\text{AC}}@{\text{Fe}}}}\) is the mass of AC@Fe before calcination (g), V the volume of calcined AC@Fe solution (mL) and [Fe2+&Fe3+] the concentration of total iron dissolved in acidic solution (mg/L).

Catalytic activity

Before testing the catalytic potential, adsorption tests were carried out to assess the Rhodamine B adsorption by the catalysts (AC@Fe/1, AC@Fe/2, AC@Fe/3). These tests were conducted without any H2O2 addition in the same conditions as described for the heterogeneous Fenton process.

The catalytic activity was tested through several series of experiments (operated in triplicate) in 200 mL of Rhodamine B solution at a given concentration under magnetic stirring (300 rpm) with the required doses of catalyst and H2O2 (30 wt%) in Erlenmeyer flasks. At the selected time, 1 mL was collected from the reaction solution and diluted in 4 mL of distilled water, and then, 0.3 mL of Na2SO3 (0.1 mol/L) was added to deactivate the hydroxyl radicals (Li et al. 2014). The solution was thus centrifuged with a centrifuge at 3000 rpm to recover the supernatant without fines particles and analyzed by UV–visible spectrophotometry (Cary scan 50) at 554 nm to determine the Rhodamine B residual concentration (C) (standard range from 0 to 7 mg/L).

First the AC@Fe/1, AC@Fe/2 and AC@Fe/3 catalysts were compared for the 50 mg/L Rhodamine B removal kinetics studied at dose of 0.2 g/L, with 4 mmol/L H2O2, at pH 4.2 and 28 °C. In a second set of experiments, the influence of the Fenton degradation parameters on the Rhodamine B removal kinetics has been studied by using the AC@Fe/3 catalyst. Reaction parameters have been varied, such as initial concentration of H2O2 (from 4 mmol/L to 10 mmol/L), catalyst dose (from 0.2 g/L to 0.4 g/L), pH (from 2 to 8), Rhodamine B initial concentration (from 30 mg/L to 100 mg/L) and temperature (20 °C, 28 °C and 40 °C).

To assess the Rhodamine B mineralization, TOC and TN contents were measured by using a Shimadzu TOC-VCSN analyzer calibrated by a potassium phthalate standard solution. The iron leaching concentration in the test solution was measured by atomic absorption (Varian SpectrAA-20).

To assess the catalyst stability and the possibility of its reusing, four successive experiments of Rhodamine B (50 mg/L) discoloration kinetics were conducted on AC@Fe/3 (0.27 g/L) with 8 mmol/L H2O2 (pH = 4.2 at 28 °C). After each cycle, the reaction mixture was placed in an oven at 40 °C for 10 h to accelerate the regeneration of the catalyst active sites and then filtered to recover the catalyst. The catalyst particles were further washed with distilled water and dried in an oven for 2 h at 105 °C before reusing.

Results and discussion

Characterization of AC and heterogeneous catalysts

The binding energies of the chemical elements (Table 1) obtained from the XPS spectra of the AC@Fe/1, AC@Fe/2 and AC@Fe/3 catalysts have confirmed the presence of carbon, oxygen, phosphorus, iron and sulfur. In fact, the sulfur and iron contained in the catalysts originate from the iron sulfate used to impregnate AC. The carbon, oxygen and phosphorus are initially present in the raw AC. In addition, the mass percentages of the atomic elements have also been determined (Table 1). The carbon and phosphorus contents decreased in catalyst compared to pristine AC from 78.61 to 71.25% and 6.89–5.29%, respectively, while iron gradually increased to a maximum of 2.45% (content in AC@Fe/3) on the AC surface.

Table 1 XPS analysis results of heterogeneous catalysts and AC showing binding energies of various chemical elements and their content
Full size table

The deconvolution of the Fe2p spectra (Fig. 1) of each catalyst yielded four peaks located around 711.7 eV, 715,72 eV, 724.76 eV and 728.27 eV, respectively. Thus, the major contributions at 711.7 eV and 724.76 eV are attributed to Fe2p3/2 and Fe2p1/2, respectively (Guo et al. 2019). For these three heterogeneous catalysts, the difference between the Fe2p3/2 binding energy and that of satellite peaks is around 4 eV. Indeed, these differences in binding energy are close to that obtained by Yamashita and Hayes (2008) to clearly confirm the presence of Fe2+. In addition, these satellites are frequently used as fingerprints to identify iron oxide phases (Pariente et al. 2008). This binding energy around of 715.72 eV indicates the presence of wüstite (FeO) (Abreu et al. 2014; Gota et al. 1999) and is typical of an iron (II) oxidation state.

Fig. 1
figure 1

Fe2p XPS spectra of AC@Fe/1, AC@Fe/2 and AC@Fe/3 catalysts

Full size image

The EDAX analyses coupled to SEM mapping (Fig. S1) of the chemical elements of an iron-based nanoparticle of AC@Fe/3 catalyst have been previously described (Briton et al. 2018). They have shown that iron can be present together with O and P so that oxo-phosphate of iron might also have been formed together with iron oxides. The P element present on the surface of AC originates from the reaction of phosphoric acid with carbon during the activation of the carbon which is known to produce P–O surface functional groups (Puziy et al. 2005).

The presence of iron-based nanoparticles in all catalysts is evidenced by TEM analysis (Fig. 2). For example, TEM image displays nanoparticles of size in the range 1–50 nm remarkably dispersed in the matrix of AC@Fe/1 and AC@Fe/3 catalysts. Moreover, smaller nanoparticle sizes have been observed in AC@Fe/1 (from about 1 nm to 10 nm) as in the iron-impregnated silica matrix prepared by Mahy et al. (2014). The variation of size observed in the AC@Fe/3 catalyst (from 1 to 50 nm) could be the result of different growth conditions because of the higher impregnation rate of this catalyst. The increase in the impregnation rate could promote the polycondensation of iron nanoparticles as reported by Karimzadeh et al. (2017). Clusters of nanoparticles are clearly highlighted in the AC@Fe/3 TEM images (at 200 nm scale). The sizes and the aspects of the observed nanoparticles are likely to suggest that they are localized in the micropores as well as in the mesopores.

Fig. 2
figure 2

AC@Fe/1 and AC@Fe/3 TEM images

Full size image

XRD diagrams (Fig. 3) of AC@Fe/1, AC@Fe/2 and AC@Fe/3 catalysts each have a diffusion halo around 26.7° generally observed in activated carbons (Kan et al. 2017). The absence of diffraction peaks related to iron oxide crystallites suggests that these iron nanoparticles have been incorporated or confined in the AC pores as isolated sites and/or deposited as highly dispersed oxide species (Mazilu et al. 2017). This dispersion of iron nanoparticles within the nanopores confers on these heterogeneous catalysts an amorphous structure (Chen et al. 2009; Mazilu et al. 2017). In addition, these formed nanoparticles would be too small because the low percentages of iron in these samples do not favor net reflection corresponding to crystalline phases as reported by other authors (Kadirova et al. 2013; Silva et al. 2017).

Fig. 3
figure 3

XRD spectra of heterogeneous catalysts AC@Fe/1, AC@Fe/2 and AC@Fe/3

Full size image

The nitrogen adsorption and desorption isotherms of inserted graph (Fig. 4) are of type IV according to the IUPAC classification and therefore typical of both microporous and mesoporous materials. Indeed, for all isotherms, a hysteresis loop is observed between the desorption and the adsorption curves, typical of the nitrogen capillary condensation in the mesopores. Indeed, at relative pressure greater than 0.1, the isotherm adsorption profiles show a constant slope increase related to the gradual filling of the mesopores. The hysteresis loops in the relative pressure range 0.452–0.995 are of H3 type according to the IUPAC classification (Sing et al. 1982). Moreover, the PSDs (Fig. 4) confirm the presence of more or less wide slit mesopores (Ø from 2 to 25 nm). Narrow-slit pores such as supermicropores (0.7 < Ø < 2 nm) and ultramicropores (Ø < 0.7 nm) are also observed (Lowell et al. 2004).

Fig. 4
figure 4

AC and heterogeneous catalysts PSDs from N2 adsorption at 77 K. Inset: N2 adsorption–desorption isotherms of the AC support and AC@Fe/1, AC@Fe/2 and AC@Fe/3 catalysts

Full size image

However, porosity of heterogeneous catalysts is found to be affected as the rate of iron impregnation increases. This reflects the presence of iron-based nanoparticles inside the pores causing the decreases in the specific surface area from 896 to 532 m2/g and in the total pore volume from 0.886 to 0.509 cm3/g for AC and AC@Fe/3, respectively (Table 2). Indeed, the iron mass contents of 1.6%, 2.2% and 3.3% in AC@Fe/1, AC@Fe/2 and AC@Fe/3, respectively, determined by dissolution and atomic absorption are in agreement with the ones found by XPS analysis. These results are in agreement with previous work showing a reduction in the specific surface area of the porous material in which nanoparticles were dispersed (Duarte et al. 2012; Jonidi Jafari et al. 2017; Liu et al. 2017).

Table 2 Textural properties of pristine AC and the prepared heterogeneous catalysts
Full size table

The histogram presented in Fig. 5 indicates the percentage of occupied porous volume with respect to the raw AC for all the catalysts depending on different pore size ranges. As the iron content increases, the filling rate of the supermicropores and mesopores estimated from the PSDs increases progressively. This is remarkable for the AC@Fe/3 catalyst with a high iron content where the reduction in supermicropore and mesopore volumes is strongly observed. This suggests that the iron nanoparticles are mainly accommodated in supermicropores and mesopores. It also suggests that iron nanoparticles are hardly formed in ultramicropores after calcinations despite high impregnation rates and that these ultramicropores are rather blocked up and not accessible in the catalysts.

Fig. 5
figure 5

Proportion of pore filling or blocking up (%) by the iron nanoparticles in the prepared heterogeneous catalysts AC@Fe/1, AC@Fe/2, AC@Fe/3 as a function of pore size range (ultramicropore Ø < 0.7 nm; supermicropore 0.7 nm < Ø < 2 nm; mesopore 2 nm < Ø < 50 nm) determined from the difference of the PSDs of each catalyst with respect of AC one

Full size image

Removal of Rhodamine B

As adsorption is an important step in the heterogeneous Fenton oxidation process (Jonidi Jafari et al. 2017, the ability of the heterogeneous catalysts to adsorb Rhodamine B in an aqueous solution has been investigated. This adsorption is generally influenced by the media porosity and the surface chemistry (Mohammadi et al. 2010). After 180 min of contact time, the discoloring rates by adsorption are 69.2%, 59.3% and 56.1% for AC@Fe/1, AC@Fe/2 and AC@Fe/3, respectively (Fig. 6).

Fig. 6
figure 6

Kinetics of Rhodamine B adsorption on AC@Fe/1, AC@Fe/2 and AC@Fe/3 catalysts (catalyst dose 0.2 g/L; pH 4.2; T = 28 °C; C0 = 50 mg/L)

Full size image

Rhodamine B can be present in either a cationic (positive charge on the nitrogen atom) or a zwitterionic form (positive charge on the nitrogen atom and negative charge on the oxygen of the carboxylic group) at pH value lower or higher than its pKa (3.7), respectively. At pH 4.2, the adsorption of the zwitterionic Rhodamine B on the surface of the catalysts could be favored by the presence of functional groups. The pHpzc of AC@Fe/1, AC@Fe/2 and AC@Fe/3 catalysts is 4.3, 3.1 and 2.8, respectively. Thus, AC@Fe/1 is assumed to have a nearly neutral surface, while the AC@Fe/2 and AC@Fe/3 surfaces are negatively charged. These differences in the charge of the surfaces could yield different electrostatic interactions with the zwitterionic Rhodamine B, explaining the increase in the adsorption capacity as pHpzc value decreases.

The strong correlation (R2 = 98.9%, inset in Fig. 6) between the adsorption capacity and the total volume of the pores reveals that the adsorption of Rhodamine B depends on the accessible porous volume. Moreover, because of its geometry and size, the Rhodamine B adsorption could occur in the mesopores and in the supermicropores of the catalysts. In conclusion, the adsorption capacities of the catalysts are related, on the one hand, to the pore volume (or the specific surface area) and, on the other hand, to the surface chemistry. Thus, as impregnation rate increases, the adsorption uptake decreases. The adsorption is quick owing to the fast diffusion of the Rhodamine B molecule into the porosity of the powdered materials, and the plateau of adsorption (quasi-equilibrium) is attained after 60 min whatever the catalyst.

Prior to any catalytic degradation test, an experiment was carried out using only H2O2 and the Rhodamine B solution in the same amount as for the catalytic tests (Fig. 7). This test has led to a reduction of 9.25% of Rhodamine B by oxidation after 180 min. For Fenton catalytic tests, the as-prepared catalyst and the hydrogen peroxide were added simultaneously in the Rhodamine B solution in order to generate hydroxyl radicals and to degrade the molecules of Rhodamine B adsorbed in the pore volumes. After 180 min, the discoloration rates were 78.5%, 87.4%, 93.0% for AC@Fe/1, AC@Fe/2 and AC@Fe/3, respectively. Thus, the discoloring of the Rhodamine B solution was quite accelerated by the catalytic oxidation combined to adsorption. The profile of the kinetics of removal of Rhodamine B clearly indicates the superimposition of two phenomena. The first one is clearly adsorption which is rapid as previously observed and for which equilibrium is nearly attained after 60 min (Fig. 7), and the second one is the degradation of the Rhodamine B via Fenton reaction which continues still after the adsorption equilibrium. Indeed, Ramirez et al. (2007) reported that the mesopores are more accessible to macromolecules such as dyes, while the micropores are easily accessible to the H2O2. In addition, previous studies have shown that catalyst having appropriate PSD can accelerate the degradation speed of the pollutant (Jonidi Jafari et al. 2017; Liu et al. 2017; Zhou et al. 2014). The efficiency of our prepared catalyst is thus attributed to its mesoporous/microporous PSD. As reported by Georgi and Kopinke (2005), the main degradation route is assumed to be the HO· attack on the organic fraction of dissolved contaminants in the pores of the catalytic support. In fact, the catalytic activity of iron-based nanoparticles and their dispersion within the pores appear to be the key factors for the Rhodamine B discoloration. The content in iron-based nanoparticles dispersed in the AC pores is also a critical point for the optimization of the Fenton catalyst. In previous works, the authors used relatively important doses of catalyst in order to improve the catalytic yield (Cuiping et al. 2012; Guo et al. 2017; Wang et al. 2014). Despite its low adsorption capacity, AC@Fe/3 has the best catalytic activity, due to its highest amount of iron nanoparticles dispersed in the porosity. Similarly, the Fenton catalytic efficiency of highly dispersed iron nanoparticles was also emphasized by Li et al. (2015) and then by Mazilu et al. (2017) in non-carbonaceous materials.

Fig. 7
figure 7

Kinetics of Rhodamine B removal by heterogeneous Fenton using AC@Fe/1, AC@Fe/2 and AC@Fe/3 catalysts (catalyst dose 0.2 g/L; H2O2 4 mmol/L; pH 4.2; T = 28 °C; C0 = 50 mg/L)

Full size image

Influence of reaction parameters

The influence of the parameters (H2O2 dose, catalyst dose, rhodamine B initial concentration, pH and temperature) on heterogeneous Fenton process has been studied using AC@Fe/3 as the best catalyst.

Thus, Fig. 8a shows that the Rhodamine B removal rate is increased from 90.1 to 96.4% after 180 min as the H2O2 dose is increased from 4 to 10 mmol/L. In fact, the increase in the H2O2 concentration led to a strong production of hydroxyl radicals. In addition, between 8 mmol/L and 10 mmol/L, the gap in terms of the degradation rate is of little significance. The reason is that excess of H2O2 reacts with HO· radicals transforming them into hydroperoxyl radicals (\({\text{HO}}_{2}^{ \cdot }\)) according to the following equations (Khataee et al. 2015; De Laat and Le 2006):

$${\text{H}}_{2} {\text{O}}_{2} + {\text{HO}}^{*} \to {\text{HO}}_{2}^{*} + {\text{H}}_{2} {\text{O}}$$
(2)
$${\text{HO}}_{2}^{*} + {\text{HO}}^{*} \to {\text{O}}_{2} + {\text{H}}_{2} {\text{O}}$$
(3)
Fig. 8
figure 8

Effect various parameters on the removal kinetics of Rhodamine B catalyzed by AC@Fe/3: a H2O2 concentration at pH = 4.2 and 28 °C, C0 = 50 mg/L, 0.27 g/L of catalyst, b effect of the catalyst dose at pH = 4.2 and 28 °C, C0 = 50 mg/L, 8 mmol/L of H2O2, c effect of pH at 28 °C, C0 = 50 mg/L, 0.27 g/L of catalyst, 8 mmol/L of H2O2, d effect of the Rhodamine B initial concentration at pH = 4.2 and 28 °C, 0.27 g/L of catalyst, 8 mmol/L of H2O2

Full size image

Those radicals [E0(\({\text{HO}}_{2}^{ \cdot }\)/H2O) = 1.7 V) are less reactive than the HO· ones (E0(HO·/H2O) = 2.8 V). Such reactions reduce the collision probability between Rhodamine B and HO· and decrease the \({\text{HO}}_{2}^{ \cdot }\) degradation speed (Guo et al. 2014). The 8 mmol/L concentration is appropriate to accelerate the Rhodamine B degradation and to minimize the H2O2 consumption cost.

The effect of the catalyst dose (from 0.20 to 0.40 g/L) studied on the discoloration of a 50 ppm Rhodamine B solution (Fig. 8b) displays the acceleration of removal kinetics as increasing the catalyst amount. Indeed, the obtained rates are 92.0%, 96.1%, 99.2% and 100% for a dose of 0.20 g/L, 0.27 g/L, 0.33 g/L and 0.40 g/L, respectively. Comparison with previous works shows that higher dose of 1 g/L of iron-loaded silicate (3 wt% Fe) (Gan and Li 2013) and 50 g/L iron-loaded graphite (15 wt% Fe) (Cuiping et al. 2012) were used to discolor 5 mg/L and 100 mg/L Rhodamine B solutions, respectively. This proves the efficiency of the AC@Fe/3 prepared catalyst. Indeed, as the catalyst dose increases, the number of active sites also increases to generate a significant amount of hydroxyl radicals attacking the Rhodamine B chromophore groups. This attack can be also favored by the adsorption owing to a large contact surface availability on the catalyst.

Figure 8c shows that the discoloring rate decreases at the beginning of the removal kinetics as the pH increases: Rates are 99.4%, 96.1%, 90.2% and 86.6% for pH 2.0, 4.2, 6.2 and 8.2, respectively. Thus, the pH increase yields a decrease in the adsorption capacity of the catalysts. Previous authors reported that the Rhodamine B adsorption on activated carbon was optimal at pH lower than the pKa (3.7) of Rhodamine B (Guo et al. 2005; Mohammadi et al. 2010) where the adsorbate global charge is positive. Thus, this molecule could be electrostatically attracted by the negatively charged surface of the catalyst (at pH > pHpzc where pHpzc = 2.8). At pH higher than 3.7, the deprotonation of the carboxyl group occurs giving the zwitterionic form of Rhodamine B (Gan and Li 2013; Guo et al. 2005) which may increase its aggregation in forms of dimers by electrostatic interactions between the xanthenes and the carboxyl groups of monomers (Gad and El-Sayed 2009; Guo et al. 2005). This dimerization reduces the adsorption on the catalyst active sites. Also, at pH above pKa (3.7), the emergence of COO− functional group in the zwitterionic form promotes some electrostatic repulsion in the presence of negatively charged catalyst (pHpzc = 2.8). The decomposition of hydrogen peroxide into molecular oxygen without formation of any significant amount of hydroxyl radicals could also explain the decrease in the degradation rate as pH rises (Chen et al. 2010; Guo and Al-Dahhan 2003). Though pH 2 gives the highest rate (99.4% in Fig. 8c), an iron leaching of 2.892 mg/L is observed. At pH 4.2, 6 and 8, the Fe leaching rates are 0.507 mg/L, 0.142 mg/L and 0.100 mg/L, respectively.

As increasing the initial concentration from 30 to 100 mg/L, the discoloration rate turned from 100 to 80% after 180 min (Fig. 8d). Indeed, the increase in the Rhodamine B concentration could cause a congestion of the active sites by hindering the contact of the hydrogen peroxide and the catalyst and yielding a less effective oxidation reaction of the Rhodamine B. In fact, the discoloring can be more and more delayed due to a large number of molecules of Rhodamine B and intermediate compounds to degrade (Cuiping et al. 2012; Murugananthan et al. 2008). 80% of removal at C0 = 100 mg/L demonstrates the catalytic potential of AC@Fe/3 which is expected to respond to any variation of dye concentration in the wastewater.

The kinetics of discoloration is increased together with the temperature in the range 20–40 °C (Fig. 9). The degradation rate varies from 76.2 to 99.04% after 180 min of treatment, as increasing the temperature from 20 to 40 °C. The temperature increase favors the production of hydroxyl radicals accelerating the Rhodamine B degradation. The value of the k constant rate (estimated from the slope of ln(C/C0) versus time plot) (Matouq et al. 2014) determined from the pseudo-first-order kinetics fits rises from 0.006 mn−1 at 20 °C to 0.030 mn−1 at 40 °C. According to the Arrhenius law (Eq. 4), the activation energy Ea is deduced (Hites 2017):

$$\ln k = \ln A - \frac{{E_{a} }}{RT}$$
(4)

where A is pre-exponential factor, Ea is the energy of activation (J/mol), R is the perfect gas constant (8.314 J/mol/K) and T is the absolute temperature (K).

Fig. 9
figure 9

Effect of temperature on the discoloration of the Rhodamine B and the Arrhenius law representation as inset (ln(K) versus 1/T): AC@Fe/3 0.27 g/L; C0 = 50 mg/L; pH = 4.2; 8 mmol/L H2O2

Full size image

The 55.2 kJ/mol calculated activation energy is roughly between 60 and 250 kJ/mol as for advanced oxidation reactions (Chen and Zhu 2007). Also, this agrees with the 82.5 kJ/mol activation energy reported by Gan and Li (2013) for the Rhodamine B oxidation catalyzed by iron supported on silica.

Mineralization and Rhodamine B spectral variation

The UV–visible absorption spectra (Fig. 10) of the solutions collected at different times during the Rhodamine B removal kinetics by using AC@Fe/3 present three absorption peaks. The high-intensity peak at 554 nm is assigned to the extended chromophore, comprising all the coupled aromatic cycles connected by the C=C and C=N double bonds, while the peaks at 347 nm and 252 nm reveal the presence of cyclic benzene structure of Rhodamine B (Hou et al. 2011). During the treatment, an intensity reduction affects all the peaks from 90 min until their quasi-total disappearance after 180 min. This reveals the destruction of chromophore groups and benzene nuclei concomitant to the Rhodamine B degradation. After 180 min, 64% TOC and 70.7% TN of the Rhodamine B solution have been converted into mineral compounds. These results are in agreement with previous work (Hadjltaief et al. 2013), showing that the discoloration is much faster than the total dye mineralization.

Fig. 10
figure 10

Spectral variation of Rhodamine B during treatment (C0 = 50 mg/L and 28 °C, pH = 4.2; AC@Fe/3 0.27 g/L; 8 mmol/L of H2O2)

Full size image

Stability of the catalyst

The catalyst stability study showed that its efficiency decreased from 96.1% in the first cycle to 73.3% in the fourth cycle (Fig. 11). Indeed, little decreases in efficiency of 10%, 11.3% and 0.8% were observed after the second, third and fourth cycles, respectively. This might be linked to the iron leaching in solution as highlighted by some authors (Kakavandi et al. 2016; Ramirez et al. 2007; Yang et al. 2015). The leached iron doses were 0.507 mg/L, 0.476 mg/L and 0.4 mg/L, after the first, second and third cycles, respectively, explained by the slight leaching of the iron nanoparticles within the pores (loss of 15.7% of iron between the first and the fourth cycles).

Fig. 11
figure 11

Repeated recycling of AC@Fe/3 (0.27 g/L) with H2O2 (8 mM) for the degradation of Rhodamine B (50 mg/L) (pH = 4.2, Temperature = 28 °C)

Full size image

The decrease in the catalytic activity is related to the loss of iron from the catalyst and could be also due to the deactivation of the catalytic sites. BET surface area of the used catalyst was found to be almost similar to the pristine (from 534 to 511 m2/g after the second cycle) revealing almost no change in the pore occupancy. Though the activity of the catalyst has decreased to some extent, it has a potential for reuse as its stability is strongly related to the dispersion of the iron oxide nanoparticles through the pores. Considering the work of Rubeena et al. (2018), this low concentration of leached iron during each cycle is not responsible for the degradation of Rhodamine B, but this degradation rather results from the iron nanoparticles dispersed within the pores. In addition, this dose of dissolved iron is slightly greater than the dose of iron in the tap water recommended (WHO 2011). Therefore, the rejection of this solution of Rhodamine B treated with this catalyst does not present an environmental pollution risk.

After the degradation of Rhodamine B, XPS analysis was performed on AC@Fe/3 catalyst already used to obtain information on the degree of chemical oxidation of the iron nanoparticles after generating hydroxyl radicals in the presence of peroxide. Thus, from the comparison of Fe2p spectra before and after catalytic tests (Fig. 12), it can be observed a modification on the spectrum marked by a displacement of the satellite peak toward Fe2p1/2. According to the literature, this satellite peak located at 718.42 eV reflects the presence of iron nanoparticles in the form Fe2O3 (Beswick et al. 2015; Wang et al. 2013). This indicates that a fraction of Fe(II) of the wüstite (FeO) was converted to Fe(III) during the heterogeneous Fenton reaction.

Fig. 12
figure 12

XPS spectra of AC@Fe/3 catalyst before and after the degradation of Rhodamine B

Full size image

In this study, catalyst stability and degradation reactions are mainly attributed to the dispersion of iron nanoparticles into micropores and mesopores. Indeed, micropores that are not very accessible to Rhodamine B, but rather to H2O2 molecules, guarantee the stability of heterogeneous catalyst to sustain the production of hydroxyl radicals. Thus, through the mesopores more accessible to Rhodamine B, the hydroxyl radicals generated in both micropores and mesopores will degrade the previously adsorbed dye. In parallel, a part of the hydroxyl radicals oxidizes in an aqueous media the dye located near the active sites. This leads us to propose a mechanism of the degradation of this dye mainly in the mesopores of the heterogeneous catalyst (Fig. 13).

Fig. 13
figure 13

Mechanism of Rhodamine B degradation using a heterogeneous catalyst in the presence of H2O2

Full size image

Conclusion

Some heterogeneous catalysts for Fenton oxidation were prepared by the iron sulfate solution impregnation at different ratios of an activated carbon from banana spike further pyrolyzed at 350 °C. AC@Fe/1, AC@Fe/2 and AC@Fe/3 catalysts have been obtained with different amounts of iron-based nanoparticles (3, 7 and 11 wt% of Fe/AC ratio, respectively), containing O and P, mainly localized in supermicropores and mesopores as observed by XPS, TEM, SEM and PSD from N2 and CO2 adsorption. These heterogeneous catalysts were tested for the oxidation of Rhodamine B solution by the Fenton process in the presence of H2O2. AC@Fe/3 is the most promising catalyst explained by the high dispersion of iron (II)-based nanoparticles in the activated carbon nanopores. In addition, the influence of factors such as the temperature, pH, catalyst dose, H2O2 concentration and initial dye concentration has been studied to define the optimized conditions for the Rhodamine B removal kinetics when using the AC@Fe/3 catalyst. Using this catalyst (0.27 g/L at pH 4.2) with 8 mmol/L H2O2, an important abatement rate of the target pollutant (Rhodamine B, C0 = 50 mg/L) was obtained such as a 96.1% discoloring rate and 64% TOC reduction after 180 min. The solution discoloration is due to the Rhodamine B degradation and its mineralization as 70.7% of nitrogen has been removed and the benzene rings observed through spectral variation have disappeared. After the second cycle, the catalytic performance remained almost constant and around 73%. After four successive tests, a 15.7 wt% iron loss of this catalyst was observed. The residual presence of iron indicates that the catalyst remains potentially active for further Fenton reactions. In fact, this catalyst could be used to treat some real textile effluents by their dye degradation.