A new perovskite-type NdFeO₃ adsorbent: synthesis, characterization, and As(V) adsorption

The TG and DTA diagrams of the gel precursor are presented in figure 5, which shows two discrete weight losses obtained around 185.29, 224.91, and 371.87 °C. The first major weight loss (24.51%) between 120 °C and 300 °C accompanied by exothermic peaks at 185.29 and 224.91 °C in the DTA curve is due to the decomposition of the nitrate and side-chain of PVA. The second weight loss (17.16%), accompanied by exothermic effects between the temperatures ∼300 and ∼600 °C with a maximum at 372 °C, might be associated with the oxidation decomposition of the PVA main chain, further decomposition of nitrate and the direct crystallization of NdFeO₃ from the amorphous component. Heating the samples at 600 °C resulted in the formation of the NdFeO₃ nanoparticles. It is in good agreement with the XRD results shown in figure 1.

Zoom In Zoom Out Reset image size

The samples of the synthesized NdFeO₃ under the optimum process conditions are the orthorhombic single phase of the perovskite type (from the XRD patterns in figures 1–4), which is identical to that reported in reference [23] using the coprecipitation method for the synthesis of NdFeO₃ nanoparticles. Wang et al [28] used the glycine combustion method for the preparation of some perovskites. The crystalline size of the calcined powders was calculated using Scherrer's equation. They lay within a narrow range of 15–20 nm.

The BET surface area of calcined powders, prepared by PVA-gel combustion, was found to be 20 m² g⁻¹ and the BJH (Barret–Joyner–Halenda) average pore diameter and total pore volume were found to be about 21 nm and 11 cm³ g⁻¹, respectively. The data imply that the as-prepared NdFeO₃ nanoparticles could be potentially applied in adsorption.

TEM and SEM micrograph images of the NdFeO₃ particles provide information on their size and morphology. The morphology of the samples with very homogeneous agglomerates made of nano-sized crystallines has been observed in the SEM images in figure 6 and shows a uniform grain size distribution and a fine powder size in the nanoscale range, between 30 and 50 nm. Figure 7 shows the TEM micrograph image of the NdFeO₃ nanoparticles at 100 nm and 20 nm scale. Sphere-like NdFeO₃ monodisperse particles with an average size of approximately 20 nm are shown, which agrees well with those calculated (21 nm) by the Scherrer equation using XRD data for the sample NdFeO₃. Such consistency implies that the NdFeO₃ nanoparticles are single-crystalline.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

EDS was performed to further confirm the composition of the prepared products. Figure 8 demonstrates the EDS spectra of NdFeO₃ nanoparticles before and after arsenic adsorption. The EDS spectra show the presence of neodymium and iron in the adsorbent before As(V) adsorption (figure 8(a)) and the presence of neodymium, iron, and arsenic in the adsorbent after As(V) adsorption (figure 8(b)). It is clear that arsenic presented on the adsorbent after adsorption (about 4.71 mass %). Carbon species presented with a significantly low content of about 2% (the presence of carbon species may also be confirmed by the IR spectra) and the NdFeO₃ adsorbent has an Fe/Nd atomic ratio of 0.86 as shown in table 1.

Zoom In Zoom Out Reset image size

The surface charge is usually calculated from the balance of protons sorbed and desorbed, with reference to the surface area. The characteristic parameter is the PZC, that is the pH value (pH_PZC) for which the surface charge equals zero, i.e. positive and negative sites have equal concentrations. The zeta potential results of the NdFeO₃ adsorbent in the absence and presence of As(V) as a function of pH are shown in figure 9. pH_PZC occurred at pH 6.1 for NdFeO₃ in 0.01 M NaCl. In the presence of 0.5 mM As(V), pH_PZC shifted to a lower pH value (approximately 4) and the zeta potential became more negative in the pH range from 4 to 9. The shifts in the pH_PZC were used as evidence of inner-sphere surface complex formation [29], because the formation of outer-sphere surface complexes cannot shift the pH_PZC (no chemical reaction between the adsorbent and the adsorbent surface). The same trend was observed by Dou et al [14] and Li et al [18].

Zoom In Zoom Out Reset image size

Figure 10 shows FTIR spectra of NdFeO₃ adsorbent before (figure 10 curve (a)) and after (figure 10 curve (b)) As(V) binding adsorption at a pH of 6.5 in the frequency range 400–4000 cm⁻¹. The obtained FTIR spectra of perovskite before As(V) adsorption is identical to that reported in [23]. The NdFeO₃ adsorbent before and after As(V) adsorption have bands at wave number values around 3200–3500 cm⁻¹, which represent the hydroxyl stretching vibration of O-H bonds, and those around 1622–1630 cm⁻¹, which represent the bending vibration of the physical water molecules [7, 13, 16, 18]. More broad peaks are observed after As(V) adsorption than before As(V) adsorption and the peaks around 1383 cm⁻¹ and 1367 cm⁻¹ can be assigned to the significantly small amount of the carbon species remaining (about 1%) [27, 30], which is reconfirmed by the results of the EDS spectra. A strong band at 1095 cm⁻¹ of NdFeO₃ before As(V) adsorption can be assigned to the bending vibration of hydroxyl groups on metal oxides (M-OH; M represents surface metal ions of Nd and Fe) [13, 16]. This observation demonstrates that the NdFeO₃ nanoparticles have high adsorption capacity to hydroxyl groups existing on their surfaces. After As(V) adsorption (curve (b) in figure 10) the FTIR spectra of NdFeO₃ clearly indicated that the M-OH bending band decreased. Some researchers have reported that the peak of the hydroxyl groups decreased or disappeared after the As(V) adsorption on the Fe-Mn [12], Fe-Ce [13], and Ce-Ti [18] adsorbents. These observations indicate that the hydroxyl groups on the adsorbent surface were involved in the arsenic adsorption [13, 18]. On the other hand, a new band appeared at 852 cm⁻¹ which corresponds to the stretching vibration of an uncompleted As-O bond (figure 10 curve (b)) [13, 14, 31, 32]. The formation of an As-OM bond indicates that the As(V) adsorption onto NdFeO₃ nanoparticles also follows the inner-sphere complex mechanism (see [8]). The FTIR study results indicate that the substitution of M-OH groups by arsenic species plays a key role in their adsorption. The peaks at 501 cm⁻¹ and 559 cm⁻¹ attributed to the M-O stretching vibration decreased to 455 cm⁻¹ and 546 cm⁻¹ after As(V) adsorption, suggesting a change of M-O groups after the adsorption.

Zoom In Zoom Out Reset image size

Raman spectroscopy is very useful to obtain further information on the As(V) adsorption on the NdFeO₃ adsorbent. Raman spectra of the NdFeO₃ adsorbent at pH = 6.0 ± 0.1 in the frequency range from 100 cm⁻¹ to 1500 cm⁻¹ are shown in figure 11. Curve (a) of figure 11 shows the Raman spectrum of the NdFeO₃ nanocrystals before As(V) adsorption. The bands at 141, 214, 286, 341, 439, 631, 857, and 1299 cm⁻¹, and the rather weak bands between 700 and 1000 cm⁻¹, are more similar to those reported in reference [33].

Zoom In Zoom Out Reset image size

The intensity of some peaks at 631 and 1299 cm⁻¹ in the Raman spectrum of NdFeO₃ before As(V) adsorption (curve (a) in figure 11) were decreased after As(V) binding adsorption (curve (b) in figure 11). This is due to the substitution of the M-O bond by an M-OAs bond. When oxygen is protonated, the M-O bond is strong, but the substitution of the –H atom by the arsenic species has weakened the M-O bond since the arsenic species attached to the M-O group is much more electronegative than the –H atom. This fact implies that an inner-sphere surface complex has been formed on the adsorbent surface [34]. Moreover, a new mode at around 857 cm⁻¹ (curve (b) in figure 11) may be attributed to the vibration of the As-O bond, indicating the presence of arsenic species in the NdFeO₃ sample.

According to the above presented results of EDS, pH_PZC, FTIR, and Raman spectra measurement, evidence of the presence of As(V) on NdFeO₃ nanoparticles can be confirmed.

Figure 12 shows the effects of solution pH on the As(V) removal on the NdFeO₃ adsorbent. As can be seen, the As(V) removal was evidently dependent on pH, with the greatest adsorption occurring under acidic conditions and decreased with increasing solution pH. It was found that the As(V) sorption capacities were nearly constant, 20 mg g⁻¹ and 5 mg g⁻¹ for the initial As(V) concentration of 10 mg l⁻¹ and 5 mg l⁻¹, respectively, up to pH 6.5, and decreased from pH 6.5 to pH 10, which can be explained by the pH_PZC value (pH_PZC = 6.1) of NdFeO₃. According to the As(V) speciation at pH 7.0, all arsenate complexes are dissociated as ${{\rm{H}}}_{2}{{{\rm{AsO}}}_{4}}^{-}$ and ${{{\rm{HASO}}}_{4}}^{2-}$ species [35]. Those species will undergo an attraction/repulsion on the adsorbent surface, depending on the nature of its charge. The positive surface of NdFeO₃ binds the ${{\rm{H}}}_{2}{{{\rm{AsO}}}_{4}}^{-}$ and ${{{\rm{HASO}}}_{4}}^{2-}with$ Coulombic forces strongly up to the pH < pH_PZC. As solution pH was increased from pH 6.1, the negative charge on the adsorbent increased and the positive charge decreased. The surface of NdFeO₃ should transform to negative, and repel the ${{\rm{H}}}_{2}{{{\rm{AsO}}}_{4}}^{-}$ or ${{{\rm{HASO}}}_{4}}^{2-}$ species for likeness of charges, resulting in the observed decrease in As(V) adsorption. The same trend in the effect of pH on arsenic removal was observed by other researchers [16, 18, 32, 36]. The As(V) removal by NdFeO₃ adsorbent was maintained at the high adsorption capacity at pH levels up to 7.0. It is reasonable to assume that, in addition to electrostatic interaction, chemical interaction exists, which will be discussed below.

Zoom In Zoom Out Reset image size

The above results of the zeta potential and FTIR analysis suggested that possible adsorption reaction of As(V) onto NdFeO₃ nanoparticles could be realized through the replacement of the OH group of M-OH with arsenate for the formation of surface complexes.

Adsorption isotherms provide some of the most useful data for understanding the mechanism of adsorption, and the characteristics of isotherms are needed before the interpretation of the kinetics of the adsorption process. Many modes have been proposed to explain adsorption equilibrium, however, no general model has been found to fit the experimental data accurately under any given condition. Among various methods employed for analyzing the nature of adsorbate–adsorbent interaction, adsorption isotherms are the most significant. The present study used Langmuir and Freundlich isotherms, which are most often used to describe the equilibrium adsorption of metal ions and are usually used to describe the equilibrium adsorption data [32, 37].

The Langmuir equation is

$$\begin{eqnarray}&&{q}_{e}={q}_{m}\displaystyle \frac{{K}_{L}{C}_{e}}{1+{K}_{L}{C}_{e}}\cdot \end{eqnarray} \tag{ 1 }$$

For linearization of the data, it can be written in the form

$$\begin{eqnarray}&&\displaystyle \frac{{C}_{e}}{{q}_{e}}=\displaystyle \frac{1}{{K}_{L}{q}_{\mathrm{max}}}+\displaystyle \frac{{C}_{e}}{{q}_{\mathrm{max}}}\cdot \end{eqnarray} \tag{ 2 }$$

The Freundlich equation is

$$\begin{eqnarray}&&{q}_{e}={K}_{F}{C}_{e}^{1/n}\cdot \end{eqnarray} \tag{ 3 }$$

The above presented equation was linearized to determine the process parameters as given below

$$\begin{eqnarray}&&\mathrm{log}\;{q}_{e}=\;\mathrm{log}\;{K}_{F}+\displaystyle \frac{1}{n}\;\mathrm{log}\;{C}_{e},\end{eqnarray} \tag{ 4 }$$

where q_e is the amount of arsenic adsorbed (mg g⁻¹) at equilibrium, q_m and K_L are Langmuir constants related to monolayer adsorption capacity (mg g^–1) and adsorption equilibrium constant (l mg⁻¹), respectively, C_e is the equilibrium arsenic concentration (mg l⁻¹), and K_F and n are the Freundlich constants related to the adsorption capacity (mg g⁻¹) and adsorption intensity, respectively. The plots of C_e/q_e versus C_e, and logq_e versus logC_e, yielded straight lines. The values of q_m and K_L, K_F can be calculated from the slopes and intercepts, respectively. The results of As(V) adsorption studies relating to the Langmuir and Freundlich equations at different concentrations of As(V) for pH 6.5 on a fixed amount of adsorbent are presented in figure 13 and summarized in table 2.

Zoom In Zoom Out Reset image size

Based on either the statistical error (χ²) or the correlation coefficient (R²) values, it could be suggested that adsorption of As(V) by NdFeO₃ nanoparticles was best modeled with the Langmuir equation (1) (χ² = 0.145, R² = 0.999), and better than the Freundlich equation (2) (χ² = 3.203, R² = 0.932). The maximum adsorption capacity of As(V) (q_m = 126.58 mg g⁻¹) for NdFeO₃ was given in table 3 for comparison with other adsorbents. It shows that NdFeO₃ has a higher maximum As(V) adsorption capacity than that of the different adsorbents reported earlier, except [13].

In this section the thermodynamic and kinetic aspect of the As(V) adsorption process will be considered. The experimental data obtained at different temperatures were used in calculating thermodynamic parameters such as Gibbs free energy (ΔG⁰), enthalpy (ΔH⁰), and entropy (ΔS⁰) using the standard relations available in literature [39]:

$$\begin{eqnarray}&&{\rm{\Delta }}{G}^{0}={\rm{\Delta }}{H}^{0}-T{\rm{\Delta }}{S}^{0},\end{eqnarray} \tag{ 5 }$$

$$\begin{eqnarray}&&\mathrm{ln}\;K=\;\mathrm{ln}\;\displaystyle \frac{{q}_{e}}{{C}_{e}}=-\displaystyle \frac{{\rm{\Delta }}{H}^{0}}{RT}+\displaystyle \frac{{\rm{\Delta }}{S}^{0}}{R},\end{eqnarray} \tag{ 6 }$$

where K is the equilibrium constant obtained from Langmuir isotherms at different temperatures and R is the universal gas constant, ΔH⁰ and ΔS⁰ were obtained from the slope and intercept of the plot of ln(q_e/C_e) versus 1/T (figure 14), namely: ΔH^o = +63.916 kJ mol⁻¹ and ΔS^o = +0.249 kJ mol⁻¹ K^-1, ΔG^o = −6.551 kJ mol⁻¹, −9.041 kJ mol⁻¹, −11.531 kJ mol⁻¹, and −14.021 kJ mol⁻¹ at 283 K, 293 K, 303 K, and 313 K, respectively. The negative values of ΔG^o obtained from equation (5) reflect a spontaneous adsorption process of As(V), while the positive value of ΔH indicates that the adsorption reaction is endothermic and the As(V) adsorption is more effective at higher temperatures.

Zoom In Zoom Out Reset image size

Kinetic studies determined the contact time of the adsorbent with adsorbate and evaluate reaction coefficients. Figure 15 shows the As(V) removal by NdFeO₃ adsorbent as a function of contact time with different initial As(V) concentration of 5 mg l⁻¹ and 10 mg l⁻¹, and the maximum uptake was observed within 100 min. Kinetic models have been used to test experimental data to investigate the kinetic adsorption mechanism. The kinetic models commonly used include the first-order rate equation and pseudo-second-order equation.

Zoom In Zoom Out Reset image size

The pseudo-first-order model for the sorption of a liquid/solid system based on solid capacity is usually written as follows [40, 41]:

$$\begin{eqnarray}&&\mathrm{log}({q}_{e}-{q}_{t})=\;\mathrm{log}\;{q}_{e}-\displaystyle \frac{{k}_{1}t}{2.303},\end{eqnarray} \tag{ 7 }$$

where q_e and q_t are the amounts of As(V) adsorbed (mg g⁻¹) at the equilibrium time and time t (min), respectively, and k₁ is the rate constant of adsorption (l min⁻¹). The slope and intercept of log(q_e-q_t) versus time t were used to determine the first-order-rate constant as shown in figure 16(a). The linear fit gives the correlation coefficient R² values of 0.945 and 0.972 for initial As(V) concentration of 5 mg l⁻¹ and 10 mg l⁻¹, respectively.

Zoom In Zoom Out Reset image size

The pseudo-second-order equation based on adsorption capacity depends on time and is usually expressed as follows [42]:

$$\begin{eqnarray}&&\displaystyle \frac{t}{{q}_{t}}=\displaystyle \frac{1}{h}+\displaystyle \frac{t}{{q}_{e}},\end{eqnarray} \tag{ 8 }$$

$$\begin{eqnarray}&&h={k}_{2}{q}_{e}^{2},\end{eqnarray} \tag{ 9 }$$

where h (mg g⁻¹ min⁻¹) denotes the initial sorption rate, and k₂ (g mg⁻¹ min⁻¹) is the rate constant of adsorption.

The linear plot of t/q_e versus time t confirms that the sorption rate followed pseudo-second-order kinetics (figure 16(b)). The linear fit gives correlation coefficient values for R² of 0.999 for initial As(V) concentration of 5 mg l⁻¹ and 10 mg l⁻¹. Based on the square of the correlation coefficient R², the adsorption processes followed the pseudo-second-order kinetic model better (the R² is closest to 1) than the pseudo-first-order kinetic model. Table 4 presents the rate constants.

Due to the complexity of substances in natural water, there might be competition from other species which may largely deteriorate the arsenate removal performance by adsorption. The effect of coexisting anions on arsenate adsorption on the NdFeO₃ nanoparticles was examined. Here, the effect of coexisting anions on arsenate removal was assessed by investigating the influence of several ubiquitous anions, such as chloride, sulfate, phosphate, and fluoride ions on the arsenate adsorption at three levels (0.1, 1.0, and 5.0 mM), the pH of 6.5 ± 0.1, and the initial arsenate concentration of 10 mg l⁻¹. As shown in table 5, the removal of arsenate was obviously hindered by phosphate as compared to chloride, sulfate, and floride. This adverse effect may be attributed to preferential adsorption of phosphate ions by the NdFeO₃ surface competing with arsenate in solution because of the similar chemistry of phosphate and arsenate. Table 5 shows that chloride did not inhibit the arsenate adsorption process even though they coexisted at high concentration, suggesting that they had low affinity toward the adsorbent, and while a slight decrease of arsenate removal was observed when the floride and sulfate concentrations were as high as 5 mM, the total removal of As(V) was still over 92 and 98%, respectively. Similar adverse effects of these anions on metal oxide adsorbents were reported elsewhere [32].

Table 5.  Effect of competing anions for arsenate adsorption on the NdFeO₃ nanoparticles (initial As(V) concentration of 10 mg l⁻¹, loading NdFeO₃ of 0.05 g l⁻¹).

	% As(V) removal in the presence of
Competing anion concentration (mM)	Cl−	${{{\rm{SO}}}_{4}}^{2-}$	F−	${{\rm{H}}}_{2}{{{\rm{PO}}}_{4}}^{-}$
0	99.1	99.1	99.1	99.1
0.10	99.1	99.1	99.1	75.4
1.00	99.1	99.1	99.0	35.2
5.00	99.0	98.5	92.5	15.8

The regeneration and reuse of the spent NdFeO₃ adsorbent for As(V) removal in the successive sorption–desorption cycles were investigated by HCl, HNO₃, NaOH, NaCl, and Na₂CO₃ solutions. The 0.5 M NaOH solution was the best eluent solution for arsenic desorption from the preliminary experimental results (98.59, 97.38, 86.28, 55.90, and 24.52% for NaOH, HNO₃, HCl, NaCl, and Na₂CO₃ solutions, respectively). The desorption of arsenate from the spent adsorbent may be due to the electrostatic repulsion between the negative sites on the adsorbent surface and anionic arsenate in NaOH solution. After the regeneration in 0.5 M NaOH solution, the regenerated NdFeO₃ adsorbent was used in the next cycle, and the adsorption capacities of the adsorbent for As(V) in the successive five cycles decreased (the percentage of As(V) adsorption of 95.20, 90.60, 87.20, 85.10, and 82.50% for the first, second, third, fourth, and fifth cycles, respectively). Li et al [18] observed the same result when investigating arsenic removal from water by a Ce-Ti oxide adsorbent.