Insight into the high reactivity of commercial Fe–Si–B amorphous zero-valent iron in degrading azo dye solutions

Abstract

Improving intrinsic reactivity is one of the key requirements in applying zero-valent iron in the field. As a new kind of zero-valent iron, iron based amorphous alloys were recently found to be capable of rapidly remediating wastewater. However, the mechanisms for the rapid degradation have not yet been fully understood. In this study, commercial Fe–Si–B amorphous alloy ribbons (Fe–Si–B^AR) were used to degrade azo dyes (Direct Blue 6 and Orange II) to study the reaction kinetics, pathway and mechanism behind the high reactivity of these iron based amorphous alloys. The results show that, under the same conditions, the surface normalized reaction rate constants for the decomposition of Orange II and Direct Blue 6 by Fe–Si–B^AR could be 1300 and 60 times larger respectively than those obtained by using 300 mesh iron powders. Through UV-vis spectrophotometry and mass spectrometry, it is found that the intermediate products of the azo dyes degraded by Fe–Si–B^AR are similar to those produced in degradation by iron powders. However, the controlling step of the degradation reaction by Fe–Si–B^AR turns out to be the diffusion process rather than the surface chemical reaction found in the reaction by iron powders. Further analysis indicates that the high degradation efficiency of Fe–Si–B^AR results from its amorphous structure and the metalloid additions, which could enhance the catalytic effect and promote the formation of a non-compact and easily detached oxide layer on the surface. The experiments under different environmental conditions show that the factors that influence the degradation efficiency of crystalline iron powders affect that of Fe–Si–B^AR in a similar way, but Fe–Si–B^AR is capable of efficiently degrading wastewater under broader conditions than the crystalline iron powders. The results indicate that Fe–Si–B^AR is a promising environmental catalyst for wastewater treatment.

---

Introduction

Since its application in permeable reactive barriers, zero-valent iron (ZVI) has been successfully used to decompose pollutants in soil/groundwater, because ZVI is non-toxic, abundant, and cheap, and its degradation process requires little maintenance.^1,2 Over the past two decades, ZVI has been reported to be efficient in the reductive decomposition of chlorinated organic compounds,^3,4 nitroaromatic compounds,⁵ azo dyes,⁶ heavy metals,⁷ and so on. However, since iron itself is a mild reductive reagent (E₀ = −0.44 eV), a series of research studies has been conducted to improve the reactivity of ZVI, such as by increasing the specific surface area of iron powders (nanoscale ZVI, NZVI^8–10), doping with noble metals (such as Pd,^4,11 Ni,¹² Cu,¹³ and so on. Although all these methods and their combination have proved to be effective, their field applications are limited by their inherent drawbacks,^14,15 i.e., aggregation (NZVI), the high cost (NZVI, Pd) and the potential toxicity (NZVI, Ni). Among all the studies of NZVI, Liu^16,17 found that partially crystallized Fe–B particles prepared by chemical co-precipitation exhibit higher degradation efficiency than commercially available Fe particles and suggested that the amorphous structure is the reason that Fe–B particles could exhibit their catalytic effect in the dechlorination of trichloroethene under an H₂ atmosphere. These reports^16,17 indicate that iron-based amorphous alloys (AAs) might be promising highly active catalytic ZVI materials.

Iron-based AAs have been widely studied since the 1970s due to their exceptional soft magnetism, corrosive resistance and mechanical properties.^18–23 With the development of the fabrication technique based on melt spinning, iron-based AA ribbons with thicknesses around 30 μm have been commercialized with low cost and have been utilized as the coils in transformers.²⁴ The metastable state of AAs makes them thermodynamically more reactive than their crystalline counterparts and the homogenous structure of AAs favours a uniform reaction on the surface.^22,25 All of these merits enable the iron based AAs to exhibit higher degradation efficiency than the widely used crystalline ZVI (elemental iron scraps or powders, CZVI) at a low cost. The results reported very recently have confirmed that the degradation rate of Direct Blue 6 (DB6) with laboratory-prepared iron based AAs is obviously higher than that seen with their crystalline counterparts or iron powders.^26,27 Our previous work²⁸ showed that the surface normalized reaction rate constant of Fe–B binary AA ribbons could be 1.8 and 89 times those of its crystalline ribbons and the 300 mesh commercial iron powders, respectively. This suggests that iron based AAs might be a good candidate for replacing CZVI in industrial applications. However, to date, several issues including reaction pathway, degradation mechanisms, and the effects of metalloids and commercial impurities on the degradation efficiency have not yet been clearly elucidated.

The purpose of our studies is to understand the relationship between the amorphous zero-valent iron and the decomposition kinetics, reaction pathway and mechanism during the degradation of azo dyes. The practical goal is to propose new ways to design new materials for processing contaminants. Based on the experimental results, the kinetics and mechanism of the decomposition of azo dyes in the presence of low cost (∼%2 per kg) commercial Fe–Si–B amorphous alloys are discussed, which may accelerate their application under various environmental conditions.

Experimental

The Fe–Si–B amorphous alloys with a nominal composition of Fe₇₈Si₁₃B₉ (atomic ratio) were provided by Antai Co. (Beijing, China). The width of the ribbons was around 20 cm. To fit our equipment, the amorphous ribbons were cut into pieces of area 10 × 10 mm², and parts of them were vacuum-annealed at 873 K for 30 minutes to obtain the fully crystallized counterpart. The 300 mesh iron powders (purity > 99.5%) were purchased from Cuiboling Co. (Beijing, China). Orange II (C₁₆H₁₁N₂NaO₄S, CAS 633-96-5) and Methyl Orange (C₁₄H₁₄N₃SO₃Na, CAS 547-58-0) were procured from Sinopharm Chemical Reagent Beijing Co., Ltd. (Beijing, China). Direct Blue 6 (DB6, C₃₂H₂₀N₆S₄O₁₄Na₄, CAS 2602-46-2) was from Hailan Chemical Pigment Co. (Tianjin, China). All other reagents were of analytical reagent grade. Deionized water was used as the solvent throughout this study.

All experiments were conducted in 500 mL beakers, which were placed in a temperature controlled water-bath trough as has been done before.²⁷ During the reaction, the solution was rod-stirred at a fixed speed. At preset time intervals, 4 mL aliquots were removed with a syringe and filtered with 0.45 μm membranes. The filtered samples were pipetted out and subjected to UV-vis spectrum scanning. For the experiments with Orange II, 250 mL of 100 mg L⁻¹ Orange II solutions were reacted with 40 g L⁻¹ Fe–Si–B scrapes or 300 mesh iron powders at 25 °C without any addition of acid or alkali. The reaction of Fe–Si–B with DB6 was investigated with different conditions, including different initial DB6 concentrations, ribbon dosages, initial solution pH, reaction temperatures and concentrations of inorganic salt Na₂SO₄. The initial pH of the DB6 solution was adjusted by 1 N hydrochloric acid or sodium hydroxide, and determined by a digital pH meter (FE20, Mettler Toledo).

The morphology and structure of the ribbons and commercial powders before and after degradation were examined by a scanning electron microscope (SEM, LEO 1530) and an X-ray diffractometer (XRD, Rigaku D/max-RB) with Cu Kα radiation. The transmission electron microscopy (TEM) images of the Fe–Si–B amorphous ribbons were obtained using an FEI F20 TEM (Tecnai). The Brunauer–Emmett–Teller (BET) surface area analysis of the ribbons and powders was performed utilizing the nitrogen method with a surface analyser model (NOVA4000, Quantachrome Instruments, USA). X-ray photoelectron spectroscopy (XPS) was performed on a Thermo Scientific ESCALAB 250Xi instrument with Al-Kα radiation.

UV-vis spectra of azo dye solutions were recorded using an UV-vis spectrometer (Unico 2800, Unico, Shanghai, China) equipped with a quartz cell of 1.0 cm path length. Open circuit potential (OCP) was measured by using a saturated calomel electrode but all potentials are reported with reference to the standard hydrogen electrode. OCP measurement was carried out by the use of a CHI 660E electrochemical workstation (CH Instruments, Shanghai, China). Intermediates of azo dyes during the reduction by Fe–Si–B were analysed by mass spectrometry (MS) (Q-Exactive, Thermo Scientific, San Jose, CA).

Results and discussion

Characterization of all three materials

The measured physical and chemical properties of Fe–Si–B amorphous alloy ribbons (denoted as Fe–Si–B^AR), Fe–Si–B crystalline alloy ribbons (denoted as Fe–Si–B^CR) and 300 mesh iron powders (denoted as Fe^CP) are summarized in Table 1.

Table 1 Properties of Fe–Si–B^AR, Fe–Si–B^CR, and Fe^CP

Parameters		Fe–Si–B^AR	Fe–Si–B^CR	Fe^CP
a The OCP values were measured at 25 °C and 100 mg L^−1 Orange II, without addition of any other reagents.b The DB6 batch experiments were conducted at 25 °C, 200 mg L^−1 DB6 and 13.3 g L^−1 iron materials, without addition of any other reagents.c The Orange II batch experiments were conducted at 25 °C, 100 mg L^−1 Orange II and 40 g L^−1 iron materials, without addition of any other reagents.d The Methyl Orange batch experiments were conducted at 25 °C, 25 mg L^−1 Methly Orange and 10 g L^−1 iron materials, without addition of any other reagents.
Size (mm)		10 × 10 × 0.02	10 × 10 × 0.02	∼0.05
Crystallinity		Amorphous	Crystalline	Crystalline
SA (m^2 g^−1)		0.013	0.013	0.343
OCP (V)^a		−0.735	−0.471	−0.439
Composition (at.%) (Fe : Si : B : O)		1 : 1.70 : 0.57 : 6.12	1 : 6.72 : 1.91 : 18.55	1 : — : — : 2.49
DB6^b	kobs	0.115 min^−1	0.025 min^−1	0.051 min^−1
	kSA	0.66 L (min^−1 m^−2)	0.14 L (min^−1 m^−2)	0.011 L (min^−1 m^−2)
Orange II^c	kobs	0.238 min^−1	0.076 min^−1	0.005 mg (L^−1 min^−1)
	kSA	0.46 L (min^−1 m^−2)	0.15 L (min^−1 m^−2)	3.6 × 10^−4 mg (min^−1 m^−2)
Methyl Orange^d	kobs	0.103 min^−1	3.87 × 10^−3 mg (L^−1 min^−1)	7.37 × 10^−4 mg (L^−1 min^−1)
	kSA	0.79 L (min^−1 m^−2)	3.0 × 10^−2 mg (min^−1 m^−2)	2.1 × 10^−5 mg (min^−1 m^−2)

---

The XRD spectra of Fe–Si–B^AR, Fe–Si–B^CR and Fe^CP are shown in Fig. 1a. The crystalline α-Fe, Fe₂B and Fe₃Si phases and α-Fe phase are observed in the XRD spectra of the Fe–S–B^CR and Fe^CP samples, respectively. For the Fe–Si–B^AR sample, except for a broad hump centered at around 45°, no detectable crystalline peak could be observed, indicative of its amorphous structure. This result is further confirmed by the HRTEM and SAED image of Fe–Si–B^AR (Fig. 1b), in which neither crystalline lattice fringes nor crystalline diffraction spots could be observed. Fig. 1c and d show the SEM images of Fe–Si–B^AR and Fe^CP. Fig. 1c shows that the thickness of the amorphous ribbons is around 20 μm, and their surface (the inset) is rather smooth. In contrast, for 300 mesh iron powders, the average diameter is around 50 μm and the surface is quite rough (Fig. 1d). It is well known that specific surface area is essential to heterogeneous reactions in chemistry. To identify the specific surface area, a BET analysis with nitrogen gas was applied. For the 300 mesh iron powders, the specific surface area was measured to be 0.343 m² g⁻¹. However, for the ribbons, because their specific surface area is too low and out of the range of the BET analysis with nitrogen, other methods needed to be applied. Considering that the ribbons are quite smooth, it is reasonable to assume that they are cuboids, as was done in previous studies.^26,28 Based on this assumption, the specific surface area of the ribbons was calculated to be 0.013 m² g⁻¹.

Fig. 1 XRD spectra, TEM and SEM images of three iron materials: (a) the XRD patterns; (b) the TEM image of Fe–Si–B^AR; (c) the cross section and surface (inset) of Fe–Si–B^AR; (d) the surface morphology of Fe^CP.

Since the reaction of azo dyes and zero-valent iron materials involves an electron transfer reaction at the water–iron interface, the change of OCPs of iron materials during the reaction could be used as a “marker” to estimate their reactivity; that is, a lower potential of iron is necessary to ensure a sufficient decomposition rate.²⁹ In this study, the changes of OCP of iron materials in Orange II solution during the reaction were recorded and the stable OCP values are presented in Table 1. The results show that the OCP of Fe–Si–B^AR is much lower than those of Fe–Si–B^CR and Fe^CP, indicating that, as a reductant, Fe–Si–B^AR is more reactive than Fe–Si–B^CR and Fe^CP in degrading Orange II.

Moreover, to obtain the qualitative and quantitative elemental information from the outermost atomic layers, XPS measurement was implemented to identify the surface chemical states of all three iron materials (Fig. 2). Using non-linear least squares curve-fitting with a mixed Gaussian–Lorentzian function in the XPSPEAK4.1 program, the components of each sample were resolved from their XPS spectra and the detailed component results of spectra analysis are provided in S1. The XPS data show that the surfaces of the three kinds of iron materials are covered with complex passive layers. Through reference to previous works,^30–32 the possible compositions are shown below. For Fe–Si–B^AR, the surficial layer should be composed of a mixture of Fe⁰ (706.6 eV), Fe₂B (707.6 eV), FeO_x (represented by Fe²⁺, 710.3 eV), Fe₂O₃ and FeOOH (Fe³⁺, 711.1 eV), Si⁰ (99.3 eV), SiO (Si²⁺, 102.4 eV), B⁰ (187.7 eV), B_xO_y (192.2 eV). For Fe–Si–B^CR, the surface layer should consist of Fe₂B (707.8 eV), Fe₃Si (708.6 eV), Fe₂O₃ and FeOOH (Fe³⁺, 711.3 eV), Si_xO (Si^2−δ, 100.7 eV), SiO₂ (Si⁴⁺, 103.8 eV), B_xO_y (192.5 eV) and B₂O₃ (193.7 eV). Lastly, for Fe^CP, the covered film should be the mixture of FeO_x (Fe²⁺, 710.2 eV) and FeOOH (Fe³⁺, 711.5 eV). Actually, the real state of these surface oxide films is rather complex, but several phenomena could be observed as is shown below. In comparison with Fe–Si–B^AR, the surface composition of Fe–Si–B^CR after the annealing process is devoid of Fe⁰, Si⁰ and B⁰, but contains further oxidized species such as Fe₃Si, SiO₂ and B₂O₃. When Fe–Si–B^CR is compared with Fe^CP, due to the existence of Si and B, relatively lower oxidation states of iron such as Fe₂B and Fe₃Si can be identified in Fe–Si–B^CR. Moreover, according to the atomic ratio of the different elements (Table 1), as is found in the results of Fe–B amorphous alloys,^31,32 there is a trend that the metalloids aggregate on the surface and the annealing process (to create Fe–Si–B^CR) further aggravates the aggregation. The results show that, under the same conditions, Fe–Si–B^AR is more resistant to oxidation than Fe–Si–B^CR and Fe^CP due to its homogeneous amorphous structure, which provides fewer defects for oxygen to attack. Furthermore, the reductive Fe⁰, B⁰ and Si⁰ found on the surface might act as the electron donors during the degradation process.

Fig. 2 X-ray photoelectron spectra of the Fe–Si–B^AR, Fe–Si–B^CR and Fe^CP surfaces before the decomposition process.

Degradation kinetics

UV-vis spectroscopy was implemented to determine the kinetics of the decomposition reaction and the type and amount of intermediate products in the decomposition process. In this study, two kinds of azo dyes were used, namely, Direct Blue 6 (DB6) and Orange II. DB6 is an industrial azo dye which could be used to simulate the real wastwater in application. However, due to DB6's complex structure, it is hard to analyse the intermediates during degradation. Therefore, a widely used azo dye, Orange II, was used to study the reaction pathway. For DB6, as is shown in Fig. 3a, only peaks at 571 nm could be identified in the UV-vis spectra. By measuring their intensity, the concentration variation of DB6 against time was obtained and is presented in Fig. 3b. In contrast, for Orange II (Fig. 3c), several absorbance bands could be identified:²⁹ the one at 484 nm is due to the n–π* transition of the azo bond, and other characteristic bands observed at 310, 228 nm and the doublet at 255 and 261 nm are assigned to π–π* transitions related to aromatic rings. Using the absorption intensity at λ_max = 484 nm, the concentration change of Orange II against time was obtained (Fig. 3d). Through regression analysis, it is found that the degradation of DB6 with all three kinds of materials follows the pseudo-first-order kinetic model (C = C₀ exp(−k_obst), R² > 0.95). In contrast, for Orange II (Fig. 3d), the decomposition by either 40 g L⁻¹ Fe–Si–B^AR or 40 g L⁻¹ Fe–Si–B^CR still follows the pseudo-first-order kinetic model, but the kinetics of 40 g L⁻¹ Fe^CP transfer into the pseudo-zero-order kinetic model (C = C₀ − k_obst). All the calculated reaction rate constants (k_obs) can be seen in Table 1.

Fig. 3 UV-vis spectroscopy and normalized concentration of DB6 and Orange II during the decomposition process at 25 °C, initial pH = 7: (a) UV-vis spectra of 200 mg L⁻¹ DB6 degraded by 13.3 g L⁻¹ Fe–Si–B^AR; (b) the normalized concentration of 200 mg L⁻¹ DB6 after reaction with 13.3 g L⁻¹ Fe–Si–B^AR, Fe–Si–B^CR and Fe^CP; (c) UV-vis spectra of 100 mg L⁻¹ Orange II degraded by 40 g L⁻¹ Fe–Si–B^AR; (d) the normalized concentration of 100 mg L⁻¹ Orange II after reaction with 40 g L⁻¹ Fe–Si–B^AR, Fe–Si–B^CR and Fe^CP. All experiments were carried out in duplicate and vertical bars represent the standard deviations of the means.

For surface processes, the reaction rate (r) can be described by the Langmuir–Hinshelwood equation (shown in Eq. (1)), where the rate is proportional to the surface coverage (θ).²⁹

r = kθ = kKC (1 + kC) (1)

where k is the reaction rate constant, K is the adsorption constant, and C is the concentration of the contaminants.

With a fast reaction on the surface and low solution concentration, only a small number of reactive sites are occupied (kC ≪ 1) and the reaction rate should be described by a pseudo-first-order reaction: r = kKC = k′C. In the case of saturation of reactive sites, θ is close to 1 and the reaction rate shows a constant value that is characteristic of a pseudo-zero-order reaction, which is favorable at low temperature and low reaction rate. Experimental data in degrading azo dye solutions show that the higher degradation rate of Fe–Si–B^AR in comparison to Fe^CP is mainly due to its higher reactivity, which is supported by the OCP measurements. Since the reaction mainly takes place on the surface, to better evaluate the reaction rate of the different iron materials, according to the study by Nam,³³ the surface area normalized rate constant, k_SA, could be used. k_SA can be obtained by the equation k_SA = k_obs/ρ_a, where ρ_a is the iron surface area concentration. As is shown in Table 1, the k_SA value for the degradation of DB6 by Fe–Si–B^AR could be around 5 and 60 times larger than those of Fe–Si–B^CR and Fe^CP, respectively, while in the degradation of Orange II by 40 g L⁻¹ Fe–Si–B^AR, it could be around 1300 times larger than that of 40 g L⁻¹ Fe^CP.

Pathway

To identify the intermediate products and determine the pathway of the reduction of azo dyes by Fe–Si–B^AR, as is shown in Fig. 4, both the UV-vis spectroscopy of the solution after subtraction of the spectra of the remaining Orange II, and the mass spectra of the solution have been utilized. The expected first step of the reduction of Orange II, as proposed by several authors,^29,33 is:

Fig. 4 UV-vis spectra (a) after subtraction of the spectra of remaining Orange II in solution presented in Fig. 3c, and (b) the MS of Orange II in the reaction with Fe–Si–B^AR in Fig. 3c.

Therefore, in this study, sulfanilic acid and 1-amino-2-naphthol were investigated as the first possible intermediates. According to the research done by Mielczarski,²⁹ in the spectrum of the solution after 2 min (Fig. 4a, inset, red line), the formed band at 228 nm with a shoulder at 249 nm is due to the presence of 1-amino-2-naphthol. By comparison with the standard spectrum of sulfanilic acid shown in Fan's paper,³⁴ the spectrum of the solution after 60 min (Fig. 4a purple line), which exhibits bands at 198 and 248 nm is due to the existence of sulfanilic acid which could be further identified through the mass spectra shown in Fig. 4b. However, as is shown in Fig. 4b, there is no sign of 1-amino-2-naphthol, which could be due to its auto-oxidation. This result shows that the reduction of the azo dyes by Fe–Si–B^AR is just like the reaction catalyzed by crystalline iron powders; that is, the Fe–Si–B^AR acts as an electron donor which provides electrons for the electron-deficient azo bonds to decompose the azo dyes into two parts by either auto-oxidation or biological decomposition.

Besides the identification of the products of Orange II, the status of Fe–Si–B^AR was monitored and the sediments' composition was identified. As is shown in Fig. 5a, the XRD spectrum shows that the Fe–Si–B^AR remains amorphous after the degradation reaction. The SEM results reveal that the surface morphology of Fe–Si–B^AR is quite rough and that some parts of the passive films were already exfoliated from the samples. In Fig. 5b, the sediments are identified to be γ-FeOOH and Na₂SO₄. γ-FeOOH should be the result of the oxide exfoliation from the Fe–Si–B^AR surface. Since no sulfate is added in the solution, the Na₂SO₄ should be the decomposition product of sulfanilic acid as has been suggested before.²⁹ No sign of boron or silicon compounds could be found in the XRD pattern, which means that either the metalloids in the matrix do not participate in the reaction, or as was suggested by Liu,¹⁷ these metalloids dissolve into the solution.

Fig. 5 (a) SEM image and XRD spectrum (inset) of Fe–Si–B^AR after the decomposition process; (b) XRD spectrum of sediments.

Although the structure of Fe–Si–B^AR remains amorphous after the degradation process, according to the XPS results shown in Fig. 6, the surface chemical status is different from the original one. For Fe–Si–B^AR after the degradation process, as is shown in S1, the surficial layer is mainly composed of FeO_x (represented by Fe²⁺, 710.5 eV), γ-FeOOH (Fe³⁺, 711.5 eV), Fe₂(SO₄)₃ (Fe³⁺, 713.4 eV), Si⁰ (99.9 eV), SiO₃²⁻ (Si²⁺, 102.4 eV) and BO₃³⁻ (B³⁺, 192.4 eV). In contrast with Fe–Si–B^AR before the degradation process, there is no sign of Fe⁰ and B⁰ on the surface of Fe–Si–B^AR after the degradation process. Moreover, through quantitative analysis, the relative atomic ratio of Fe : Si : B : O after degradation is 1 : 0.44 : 0.18 : 2.93, which shows that the contents of silicon and boron are far less than before. This result shows that, during the degradation process, the rate of the exfoliation of silicate and borate from the surface is much higher than that of iron oxide.

Fig. 6 X-ray photoelectron spectra of the Fe–Si–B^AR surfaces after the decomposition process.

Together these experimental observations allow us to propose the detailed reaction pathways. As is shown in Fig. 7, the first step of the degradation is the adsorption of the azo dyes on the surface of Fe–Si–B^AR through the –SO₃⁻ molecule group. Then the adsorbed molecules undergo reductive decomposition to form 1-amino-2-naphthol and sulfanilic acid. Due to auto-oxidation, the formed 1-amino-2-naphthol soon goes through a further decomposition without forming any aromatic species. With increasing oxide layers, the rest of the adsorbed sulfanilic acid is either desorbed and goes back into the solution or experiences further decomposition and forms smaller molecules like SO₄²⁻.

Fig. 7 Schematic of the degradation pathway of Orange II by Fe–Si–B^AR.

Effect of environmental variables

It is widely known that environmental variables, such as solution pH, contaminant concentrations and iron dosage, have a great impact on the degradation efficiency of elemental iron.³⁴ As far as we know, there are no detailed reports about the effect of different environmental conditions on the decomposition with iron based AAs. So, in this study, different variables, including different temperature, initial solution pH, Fe–Si–B^AR usages, initial DB6 concentrations, and Na₂SO₄ concentrations, are used to explore their effects on the degradation efficiency of DB6 by Fe–Si–B^AR (shown in Fig. 8). The results show that, similar to the previous study on CZVI,³⁴ the degradation efficiency of DB6 by Fe–Si–B^AR was enhanced with the increase in ribbon dosage, but decreased with increasing initial solution pH, DB6 and sodium sulfate concentrations.

Fig. 8 Effect of environmental variables on the decomposition of DB6 by Fe–Si–B^AR: (a) initial solution pH; (b) Fe–Si–B^AR dosage; (c) sodium sulfate concentration; (d) initial DB6 concentrations. (Beside the studied variables, others were set at 25 °C, 200 mg L⁻¹ DB6, initial pH = 7, 13.3 g L⁻¹ Fe–Si–B^AR, and 0 g L⁻¹ NaSO₄.)

Solution pH has been experimentally proved to be an important factor in iron-contaminant systems.³⁴ For most dye wastewater, the pH is in the range of 6–10. In this work, therefore, the effect of initial solution pH on the decomposition of DB6 by Fe–Si–B^AR was studied in a pH range of 2–10, and the results are shown in Fig. 8a. With the initial pH at 2, 4, or 7, DB6 solutions are completely decomposed within 30 minutes. As the initial pH increases up to 10, the final degradation efficiency is less than 95% after 60 minutes. This could stem from the fact that at lower pH (<pH_pzc, pzc stands for the point of zero charge), the surface of Fe–Si–B^AR is positively charged and the dye molecules with a sulfuric group are negatively charged, which is favourable for their adsorption onto the iron surface. When the solution pH is above the isoelectric point, the oxide surface becomes negatively charged and the surface could be easily covered by corrosion products which will decrease the number of reactive sites and inhibit the electron transfer. However, even at an initial pH of 10, more than 90 percent of DB6 could be decomposed within 60 min. This result has a very practical meaning in that it shows that there is no need to add any acid solution into the reaction system to ensure acid conditions for complete degradation.

Moreover, with increasing the Fe–Si–B^AR dosage from 1.67 to 13.3 g L⁻¹ (Fig. 8b), the observed reaction rate constants increase from 0.017 to 0.115 min⁻¹. Considering the low cost of Fe–Si–B^AR, it is practical to use a high dosage which could increase the reactive sites and thus the degradation rate. In contrast, with the addition of sodium sulfate from 0 to 10 g L⁻¹ (Fig. 8c), within 60 minutes, the degradation efficiency of Fe–Si–B^AR drops from 100% to 87%. This result could be due to the competition of the sulfate ions with DB6 to occupy the adsorptive and reactive sites on the surface of Fe–Si–B^AR. Lastly, with the increase of DB6 concentration from 100 to 300 mg L⁻¹ (Fig. 8d), unlike the effect of lowering degradation efficiency observed with high concentration azo dyes by nano-scale iron powders,³⁴ all the contaminants could be fully decomposed by Fe–Si–B^AR within 50 min. All these results indicate that Fe–Si–B^AR is a good candidate for field application in the degradation of azo dye containing wastewater.

Mechanism

Due to the sensitivity of chemical reactions to temperature, the determination of reaction rates under different temperatures is very useful in providing insight into the mechanism. For example, the effect of temperature on the rates of heterogeneous reactions can be used to distinguish the rate-limiting step, which may involve a chemical reaction at the surface, or the diffusion of a reactant;³⁵ diffusion controlled reactions in solution have relatively low activation energies (∼8–21 kJ mol⁻¹), whereas the surface controlled chemical reactions have large activation energies (>29 kJ mol⁻¹).

Experiments on the effect of temperature were conducted to evaluate the activation energy for the degradation of DB6 with Fe–Si–B^AR. As shown in Fig. 9a, as the temperature was increased from 25 to 55 °C, the decomposition efficiency within 10 minutes increased from 67.7 to 89.3%. Reaction rate constants at different temperatures were obtained by data fitting with the pseudo-first-order model. According to the Arrhenius-type equation, the activation energy can be derived from the change of temperature, with eqn (3):^34,35

ln k_T = −ΔE/RT + ln A (3)

where k_T is the kinetic rate constant at different temperatures (T), ΔE is the activation energy, R is the gas constant and A is a constant.

Fig. 9 Effect of temperature on the degradation of DB6 solution: (a) plot of (C_t/C₀) vs. time by Fe–Si–B^AR; (b) the Arrhenius plot for the estimation of activation energy of degradation of DB6 by Fe–Si–B^AR and Fe^CP (200 mg L⁻¹ DB6, initial pH = 7, 13.3 g L⁻¹ Fe–Si–B^AR or Fe^CP, and 0 g L⁻¹ NaSO₄).

The plot of ln k_T versus 1000/RT is shown in Fig. 9b. Obviously, a good linear relationship can be obtained (R² > 0.99), and the calculated activation energies for the degradation of DB6 by Fe–Si–B^AR and Fe^CP are approximately 17.98 and 31.98 kJ mol⁻¹, respectively. The lower activation energy of Fe–Si–B^AR than that of Fe^CP means that Fe–Si–B^AR exhibited a catalytic effect in decomposing azo dyes similar to that observed for nano-scale Fe–B particles with amorphous structure.^17,28 Liu et al.¹⁷ found that the amorphous structure could activate hydrogen and catalytically promote the reductive hydrogenation of oxidative pollutants. Compared to the typical noble metal catalysts (Pd) for hydrogenation, the Fe–Si–B^AR has two roles. Firstly, the zero-valent iron is used as a reaction participant to promote the production of hydrogen from water. Meanwhile, the zero-valent iron is also used as a reductant for degrading the azo dyes. This is advantageous over the noble metals, which are poor electron donors. Secondly, the Fe–Si–B^AR catalyzes the dissociative adsorption of the formed hydrogen, in a similar way to the noble metals.

Moreover, the activation energy of Fe–Si–B^AR is in the range 8–21 kJ mol⁻¹, which shows that the rate-limiting step in the degradation reaction is the diffusion of reactants. Under the condition with rod stirring, the diffusion process caused by the inhomogeneous distribution of reactants is restricted, and the passive film formed on the iron materials' surface should be the major obstacle in the electron transfer between metallic iron and azo bonds.

Through the study of the reaction pathway, we could see that, during the degradation process, the surface of Fe–Si–B^AR is covered with a layer of mixture including γ-FeOOH, SiO₃²⁻, BO₃³⁻ and so on. According to the previous study on the Fe–B composites, due to its lower potentials than iron (E₀ for B and Si are −0.869 eV and −0.78 eV, respectively), metalloids should play a central role in the passive corrosion behavior.³⁶ According to the far lower content of silicon and boron on the Fe–Si–B^AR after the degradation process than originally, it can be inferred that the formed incompact passive film is inhomogeneous and easy to exfoliate to expose the underlying reductive Fe⁰ and B⁰ atoms. So, unlike the widely used Fe/Pd, whose catalysts Pd and Ni may lose activity due to the dense oxide film covering, our further work with Fe–Si–B amorphous powders showed that the high decomposition efficiency of Fe–Si–B^AR particles could be preserved after 5 cycles, similar to the results found by Zhang et al.²⁶

To further confirm the degradation ability of the Fe–Si–B^AR, Orange II, another widely used azo dye, and Methyl Orange (widely used as pH indicator) were also utilized in this degradation study (see ESI Fig. S2†). The calculated reaction rate constants are also listed in Table 1. The results show that, under the same conditions, the k_SA value of the Fe–Si–B^AR is around 25 and 37 000 times larger than those of Fe–Si–B^CR and Fe^CP, respectively. Furthermore, in the UV-vis spectra of Methyl Orange degraded by Fe–Si–B^AR, a peak at 248 nm could be identified, which shows the existence of the degradation product, sulfanilic acid. This result indicates that the degradation of azo dyes by Fe–Si–B^AR is through the reductive decomposition of azo dyes.

According to the above discussion, the high degradation rate of Fe–Si–B^AR should stem from both the catalytic effect and the incompact passive film. In contrast, Fe–Si–B^CR, which possesses a similar crystalline structure and lower iron content than Fe^CP, has a higher surface normalized reaction rate constant than Fe^CP. This can also be ascribed to the incompact passive film, which is easy to detach due to the borate and silicate formed during the degradation process. The present results show that the metalloid elements play an important role in enhancing the degradation rate of zero-valent iron, and Fe–Si–B^AR, as a new kind of zero-valent iron, is promising in the remediation of azo dye containing wastewater.

Conclusions

Commercial Fe–Si–B amorphous alloy ribbons with a thickness of around 20 μm were used to degrade DB6 and Orange II water solutions. It has been found that low dosages of Fe–Si–B amorphous ribbons could completely decompose 200 mg L⁻¹ DB6 and 100 mg L⁻¹ Orange II solutions within 30 min. Under the same conditions, the surface normalized reaction rate constants of the decomposition of DB6 and Orange II by Fe–Si–B^AR were 60 and 1300 times larger than those by 300 mesh iron powders, indicating that the degradation efficiency of iron-based AAs in decomposing azo dyes is much higher than that of the CZVI. The degradation mechanism of the amorphous Fe–Si–B is reductive degradation, the same as that of CZVI. It is found that the high degradation efficiency of the Fe–Si–B amorphous ribbons results from its amorphous structure and the metalloids' addition, which could enhance the catalytic effect and inhibit the formation of a thick and compact oxide layer on the surface. The experiments with different environmental conditions further proved the superior efficiency of Fe–Si–B amorphous alloys in the degradation of azo dyes. The present study indicates that commercial Fe–Si–B amorphous alloy ribbons possess high reductive activity and can serve as a new kind of ZVI material to remediate wastewater with oxidative contaminants.

Acknowledgements

We are grateful to Mr Xiaoyun Gong for his assistance with the MS experiment and the related analysis. The research is financially supported by the National Natural Science Foundation of China (no. 51271095).

Notes and references

1. J. D. Gallinati, S. D. Warner, C. L. Yamane, F. S. Szerdy, D. A. Hankins and D. W. Major, Groundwater, 1995, 33, 834–835.
2. F. Fu, D. D. Dionysiou and H. Liu, J. Hazard. Mater., 2014, 267, 194–205.
3. R. W. Gillham and S. F. Ohannesin, Groundwater, 1994, 32, 958–967.
4. P. X. Wu, C. M. Liu, Z. J. Huang and W. M. Zhang, RSC Adv., 2014, 4, 25580–25587.
5. A. Agrawal and P. G. Tratnyek, Environ. Sci. Technol., 1996, 30, 153–160.
6. E. J. Weber, Environ. Sci. Technol., 1996, 30(2), 716–719.
7. J. P. Gould, Water Res., 1982, 16, 871–877.
8. C. B. Wang and W. X. Zhang, Environ. Sci. Technol., 1997, 31, 2154–2156.
9. W. X. Zhang, J. Nanopart. Res., 2003, 5, 323–332.
10. X. Li, F. Y. Gai, B. Y. Guan, Y. Zhang, Y. L. Liu and Q. S. Huo, J. Mater. Chem. A., 2015, 3, 3988–3994.
11. C. Grittini, M. Malcomson, Q. Fernando and N. Korte, Environ. Sci. Technol., 1995, 29, 2898–2900.
12. H. Ma, Y. P. Huang, M. W. Shen, D. M. Hu, H. Yang and M. F. Zhu, et al., RSC adv., 2013, 3, 6455–6465.
13. L. M. Ma, Z. G. Ding, T. Y. Gao, R. F. Zhou, W. Y. Xu and J. Liu, Chemosphere, 2004, 55, 1207–1212.
14. D. O’Carroll, B. Sleep, M. Krol, H. Boparai and C. Kocur, Adv. Water Resour., 2013, 51, 104–122.
15. C. Lee, J. Y. Kim, W. I. Lee, K. L. Nelson, J. Yoon and D. L. Sedlak, Environ. Sci. Technol., 2008, 42, 4927–4933.
16. Y. Q. Liu, S. A. Majetich, R. D. Tilton, D. S. Sholl and G. V. Lowry, Environ. Sci. Technol., 2005, 39, 1338–1345.
17. Y. Q. Liu, H. Choi, D. Dionysiou and G. V. Lowry, Chem. Mat., 2005, 17, 5315–5322.
18. C. Suryanarayana and A. Inoue, Int. Mater. Rev., 2013, 58, 131–166.
19. R. Hasegawa and R. Ray, J. Appl. Phys., 1978, 49, 4174–4179.
20. F. J. Liu, K. F. Yao and H. Y. Ding, Intermetallics, 2011, 19, 1674–1677.
21. J. F. Li, Y. Shao, X. Liu and K. F. Yao, Chin. Sci. Bull., 2015, 60, 396–399.
22. F. E. Luborsky, Amorphous metallic alloys, Butterworth and Co. Ltd., 1983, p. 534.
23. K. F. Yao and C. Q. Zhang, Appl. Phys. Lett., 2007, 90, 061901.
24. Y. Yoshizawa, S. Oguma and K. Yamauchi, J. Appl. Phys., 1988, 64, 6044–6046.
25. R. Hasegawa, in Glassy metals: magnetic, chemical and structural properties, Boca Raton, CRC Press, 1983.
26. C. Q. Zhang, H. F. Zhang, M. Q. Lv and Z. Q. Hu, J. Non-Cryst. Solids, 2010, 356, 1703–1706.
27. J. Q. Wang, Y. H. Liu, M. W. Chen, G. Q. Xie, D. V. Louzguine-Luzgin and A. Inoue, et al., Adv. Funct. Mater., 2012, 22(12), 2567–2570.
28. Y. Tang, Y. Shao, N. Chen and K. F. Yao, RSC Adv., 2015, 5, 6215–6221.
29. J. A. Mielczarski, G. M. Atenas and E. Mielczarski, Appl. Catal., B, 2005, 56(4), 289–303.
30. R. Jain, N. S. Saxena, K. V. R. Rao, D. K. Avasthi, K. Asokan and S. K. Sharma, Mater. Sci. Eng., A, 2001, 297, 105–110.
31. S. P. Chenakin, G. G. Galstyan, A. B. Tolstogouzov and N. Kruse, Surf. Interface Anal., 2009, 41, 231–237.
32. S. K. Sharma and S. Hofman, Appl. Surf. Sci., 1991, 51, 139–155.
33. S. Nam and P. G. Tratnyek, Water Res., 2000, 4(6), 1837–1845.
34. J. Fan, Y. H. Guo, J. J. Wang and M. H. Fan, J. Hazard. Mater., 2009, 166(2–3), 904–910.
35. H. L. Lien and W. X. Zhang, Appl. Catal., B, 2007, 77, 110–116.
36. S. M. Ponder, J. G. Darab, J. Bucher, D. Caulder, I. Craig and L. Davis, et al., Chem. Mater., 2001, 13, 479–486.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra02870a

---