Adsorption-Desorption Behavior of Arsenate Using Single and Binary Iron-Modified Biochars: Thermodynamics and Redox Transformation.

Arsenic (As) is a dangerous contaminant in drinking water which displays cogent health risks to humans. Effective clean-up approaches must be developed. However, the knowledge of adsorption-desorption behavior of As on modified biochars is limited. In this study, the adsorption-desorption behavior of arsenate (AsV) by single iron (Fe) and binary zirconium-iron (Zr-Fe)-modified biosolid biochars (BSBC) was investigated. For this purpose, BSBC was modified using Fe-chips (FeBSBC), Fe-salt (FeCl3BSBC), and Zr-Fe-salt (Zr-FeCl3BSBC) to determine the adsorption-desorption behavior of AsV using a range of techniques. X-ray photoelectron spectroscopy results revealed the partial reduction of pentavalent AsV to the more toxic trivalent AsIII form by FeCl3BSBC and Zr-FeCl3BSBC, which was not observed with FeBSBC. The Langmuir maximum AsV adsorption capacities were achieved as 27.4, 29.77, and 67.28 mg/g when treated with FeBSBC (at pH 5), FeCl3BSBC (at pH 5), and Zr-FeCl3BSBC (at pH 6), respectively, using 2 g/L biochar density and 22 ± 0.5 °C. Co-existing anions reduced the AsV removal efficiency in the order PO4 3- > CO3 2- > SO4 2- > Cl- > NO3 -, although no significant inhibitory effects were observed with cations like Na+, K+, Mg2+, Ca2+, and Al3+. The positive correlation of AsV adsorption capacity with temperature demonstrated that the endothermic process and the negative value of Gibbs free energy increased (-14.95 to -12.47 kJ/mol) with increasing temperature (277 to 313 K), indicating spontaneous reactions. Desorption and regeneration showed that recycled Fe-chips, Fe-salt, and Zr-Fe-salt-coated biochars can be utilized for the effective removal of AsV up to six-repeated cycles.

Introduction

Arsenic (As) is included as a group 1 carcinogenic chemical[1] which occurs naturally in the Earth’s groundwater. Natural sources such as weathering and dissolution of As-enriched minerals, volcanic emissions, and biological reactions and anthropogenic sources like mining and smelting operations, wood preservation activities, pesticides use in agriculture, and discharge from tannery and battery industries, all make major contributions to the release of As into the environment.[2−4] The As in water is commonly present as inorganic oxyanions of trivalent arsenite (AsO33– and As3+) and pentavalent arsenate (AsO43– and As5+).[2,5,6] It is reported that more than 200 million people in 107 countries are adversely affected by a range of health-related issues caused by the consumption of elevated levels of As-tainted drinking water greater than the World Health Organization (WHO) provisional guideline value of 10 μg/L.[7,8] Chronic exposure to As can cause cancer of the bladder, skin, and lungs and other impacts to the central nervous system, IQ impacts in children, skin pigmentation, cardiovascular systems, hypertension, and endocrine disruption.[9−11] Therefore, an efficient As removal technology from water is paramount to protect human health and the environment.

Over the last few decades, various removal strategies such as oxidation, precipitation/co-precipitation, coagulation, ion exchange, reverse osmosis, membrane separation techniques, and adsorption have served to either remove or diminish As concentrations below the WHO provisional guideline value.[12−19] Among these options, adsorption is one of the efficient techniques for the removal of As from contaminated water as this technique is simple, cost-effective, and produces less waste products.[20−22] Various adsorbents such as clay, modified clay, and clay supported nano zero valent iron (nZVI),[23−26] alumina,[27−29] activated carbon,[30−32] graphene,[33] and biochar (BC)[34−36] have been used to remove As from contaminated waters.

In contrast to activated carbon, BC has emerged as an affordable remediation substrate for the removal of environmental pollutants in natural and wastewaters.[37−40] However, the adsorption efficiency of pristine BCs can be improved by single or binary metal loadings on to BC surfaces.[41−46] Natural and synthetic Fe (hydr)oxides and Fe-based materials are promising to remove As from aqueous solutions effectively.[47−50] Iron-based adsorbents are commonly being used for remediation techniques because (1) generally, these adsorbents are relatively favorable to As adsorption and are environmentally friendly[51,52] and (2) As is mostly co-precipitated with Fe, and the As-removal efficiency depends on the Fe/As ratio during co-precipitation.[53,54] Superior As removal efficiency was reported by the higher Fe/As ratio.[55,56] However, As mobilization is controlled by pH, oxidation number, and elements like Fe and Mn.[49,57,58]

Several studies have demonstrated that strong adsorption of As occurs from aqueous solution on the surface of binary metal oxides including Fe–Al,[59] Fe–Cu,[60] Fe–Mn,[49,61,62] Fe–Ni,[63,64] and Fe–Zr.[65] In recent years, utilization of bimetals such as zirconium (Zr) and Fe for the modification of BC exhibits a much higher adsorption capacity for AsV compared to individual Fe- and Zr-modified BC.[46] Very limited study has been reported on the removal of AsV using Zr–Fe-modified BC.[46] In addition, Ren et al. (2011) reported the high adsorption capacity of the Fe–Zr binary oxide adsorbent in removing both AsV and AsIII from water.[66] Additionally, the practical application of adsorbents in removing As can be evaluated by desorption and reusability testing. After performing As adsorption, adsorbents in the suspension are needed to be regenerated using acids or alkali to evade secondary contamination. Therefore, additional management is required to minimize the risk of regenerated As-containing solutions before disposal.[67,68]

Under oxic environments, BC may potentially reduce AsV to AsIII by donating electrons from the surface of BC.[14,69] It was reported that partial reduction of AsV occurred on the BC surface under oxic conditions with BC produced at 300 and 700 °C, respectively.[14] Nevertheless, the redox transformation of AsV and the bonding chemistry by modified BCs are still largely unexplored. Hence, it is crucial to investigate adsorption–desorption behavior of AsV by bimetal (Fe and Zr)-modified BC and to explore whether modified BC reduces AsV to AsIII during adsorption. However, there is little evidence in the literature that modified BCs may lead to AsV reduction. Thus, additional research is required to investigate the modification of BCs using single Fe and Zr–Fe bimetals for the removal of As from aqueous solutions.

The aims of the current study were to (1) synthesize Fe-coated BCs from two different commercial Fe-sources using metallic Fe-chips and Fe-salt (FeCl3·6H2O); (2) synthesize additional Zr–FeCl3-coated BC to assess the affinity of Zr–Fe bimetals on As binding;[65,70] (3) examine the adsorptive behavior of Fe in the presence of Zr; (4) evaluate the effectiveness of single and binary Fe- and Zr–FeCl3-coated BCs in removing AsV from aqueous solution using batch experiments; (5) explore the redox transformation of sorbed AsV on the BC surface by XPS; and (6) undertake a desorption study of As-loaded BCs for assessing regeneration and stability in practical applications of these adsorbents.

Results and Discussion

AsV Adsorption Experiments in Single Fe and Binary Zr–Fe-Modified BCs

Solution pH

The adsorption process is greatly affected by the pH-dependent surface protonation and deprotonation of metal oxide/hydroxides. First, protonation of the metal hydroxide surfaces occurs at low pH (pH < 5), and the deprotonation of metal hydroxide tends to increase with increasing pH (pH > 7). This result is due to the low affinity among oxyanions of As-species and adsorbents at high solution pH.[66,71−73]Figure A displays the percentage of AsV (10 mg/L) removal by all examined Fe-modified BCs. Figure A describes that the amount of adsorbed AsV sharply increased up to pH 5 and then gradually decreased with further increasing pH from 6 to 11 for both FeB and FeCl3B. Similarly, for Zr–FeCl3B, AsV adsorption increased up to pH 6 and then declined. The decreased adsorption at pH > 7 was due to the electrostatic repulsion of negatively charged AsV species, ligand displacement from hydroxide, and the negative surface charge of adsorbents.[74]

Figure 1

(A) Effect of pH on removal (%) of AsV, (B) effect of time on adsorption capacity of AsV (initial AsV concentration was 10 mg/L, BC dosage was 2 g/L, and temperature was 22 °C), and (C) adsorption capacity at equilibrium of FeB, FeCl3B (pH 5), and Zr–FeCl3B at pH 6 (initial AsV concentration was 5–300 mg/L, BC density was 2 g/L, pH was 5 for FeB and FeCl3B, pH was 6 for Zr–FeCl3B, and temperature was 22 ± 0.5 °C).

The zeta potentials of FeB and FeCl3B were positive in the pH range of 2–4, with a net zeta potential value of +22.4 to +9.34 mV at pH 2–4 and +12.23 to +7.66 mV at pH 2–4, respectively. The zeta potential decreased gradually with increasing pH. The Zr–FeCl3 coating possessed a net zeta potential of +25.02 mV at pH 2, decreasing to +4.78 mV at pH 6 (Table S1 in Supporting Information section). The point of zero charge (pHPZC) for FeB, FeCl3B, and Zr–FeCl3B were calculated as being 4.7, 4.9, and 6.3, respectively (Table S2 and Figure S1), which indicated that the BCs contained a net positive surface charge at pH < pHPZC. At low pH, the BC composites performed as weak acids and formed positive surface sites that were able to attract negatively charged As species, such as H2AsO4–, HAsO42–, and AsO43–. In addition, As adsorption was inhibited by electrostatic repulsion at high pH.[75] Similar findings have been reported by other studies for the adsorption of multi-protonated As species and other oxyanions such as CrVI toward metal oxides and bio-adsorbents.[46,75−77]

Reaction Time and Kinetic Modeling

Adsorption kinetics is critical to determine the efficacy and mechanisms of adsorbate removal processes. The maximum AsV removal efficiencies of FeB, FeCl3B, and Zr–FeCl3B were 78.25 (at pH 5), 87.57 (at pH 5), and 99.15% (at pH 6), respectively, after 48 h (the initial concentration of AsV was 10 mg/L). Virtually no further change was observed after this time (Figure B). Thus, AsV adsorption remained constant after 48 h reaction time. To determine the reaction rate-controlling step for AsV adsorption by Fe-coated BCs, four models were applied, including the pseudo-first-order, pseudo-second-order, Elovich model, and intraparticle diffusion model (Figure ). The fitting parameters of the models are described in Table .

Figure 2

Non-linear kinetic models: pseudo-first-order (A), pseudo-second-order (B), Elovich (C), and intraparticle diffusion model (D) (initial AsV concentration was 10 mg/L, BC density was 2 g/L, and pH was 6 at 22 ± 0.5 °C).

Table 1

Kinetic Models and Best–Fit Parameters for AsV Adsorption Data

		pseudo-first-order			pseudo-second-order			Elovich			intraparticle diffusion
BC	qe-exp (mg/g)	k1 (1/h)	qe-cal	R2	k2 (g/mg/h)	qe-cal (mg/g)	R2	β (mg/g)	α (mg/g h)	R2	kid (g/mg/h1/2)	C (mg/g)	R2	pH
FeB	3.26	0.12	3.16	0.98	0.05	3.50	0.99	2.81	0.15	0.81	0.33	0.64	0.85	5
FeCl3B	3.50	0.13	3.44	0.98	0.05	3.79	0.99	2.96	0.23	0.85	0.35	0.75	0.83	5
Zr–FeCl3B	4.02	0.11	3.93	0.97	0.03	4.39	0.99	2.67	0.24	0.91	0.41	0.68	0.89	6

The calculated AsV adsorption capacity (qe-cal) of FeB, FeCl3B, and Zr–FeCl3B was 3.16, 3.44, and 3.93 mg/g, which is lower compared to the experimental value (qe-exp) of 3.26, 3.5, and 4.02 (Table ). The regression coefficient (R2) values were 0.98, 0.98, and 0.97 of FeB, FeCl3B, and Zr–FeCl3B, respectively, for the pseudo-first-order kinetic model (Table ). However, the regression coefficients (R2) are closer to unity (R2 = 0.99) for pseudo-second-order reaction kinetics (Table and Figure B). The higher regression coefficient value (R2 = 0.99) indicating the AsV adsorption process was best fitted with the pseudo-second-order kinetic model (Table ). Furthermore, the calculated qe-cal values were 3.5, 3.79, and 4.39 mg/g, which agreed well with the experimental qe-exp values for AsV adsorption by FeB, FeCl3B, and Zr–FeCl3B (Table ). Thus, AsV adsorption is mainly affected by chemical interactions between AsV and the BC surface.

Influence of Initial As-Concentration and Adsorption Isotherms

Experimental adsorption data showed that the amount of AsV adsorption increased when the AsV concentration also increased (Figure C). However, AsV removal efficiency gradually declined at high initial AsV concentrations. This is because of the higher concentration difference between the adsorbents and the solution and the potential energy driving force.[75] In addition, the available active sites were fixed for Fe- and Zr–Fe-modified BCs, which would be saturated by AsV at higher concentrations.

To evaluate the AsV adsorption capacity of the single Fe-coating and binary Fe and Zr coatings BC, four adsorption isotherm models [Langmuir, Freundlich, Temkin, and Dubinin-Radushkevich (D-R)] were employed to fit the equilibrium adsorption data (Figure ). The R2 values ranged from 0.98–0.99, 0.98–0.98, 0.96–0.99, and 0.99–0.99 for the Langmuir, Freundlich, Temkin, and D-R models, respectively (Table ). The maximum AsV adsorption capacities achieved were 27.4, 29.77, and 67.28 mg/g by the Langmuir non-linear model fitting for FeB, FeCl3B, and Zr–FeCl3B, respectively (Table and Figure A). The KF values were calculated as 2.02, 3.21, and 7.37, while the 1/n values were 2.18, 2.45, and 2.11 for FeB, FeCl3B, and Zr–FeCl3B, respectively (Table ). The higher KF value indicates higher AsV adsorption capacity by Zr–FeCl3B, where 1/n value describes that the AsV adsorption process could be favorable.[78,79] The strong concentration gradient between AsO43– and the BC surface in the solution phase contributed to more AsV adsorption.[78] Thus, AsO43– anions migrated to the heterogeneous surfaces of BCs (Figure B).

Figure 3

Non-linear fitting of isotherm models: Langmuir (A), Freundlich (B), Temkin (C), and D-R (D) (the initial AsV concentration was 5–300 mg/L, BC density was 2 g/L, pH was 6, and temperature was 22 ± 0.5 °C).

Table 2

Adsorption Isotherm Models and Best–Fit Parameters for AsV Adsorption Data

		Langmuir model parameters					Freundlich model parameters				Temkin model parameters			Dubinin-Radushkevich model parameters
BC	qexp (mg/g)	qcal (mg/g)	qm (mg/g)	KL (L/mg)	RL	R2	qcal (mg/g)	KF (g/mg/h)	1/n	R2	b (J/mol)	A (L/g)	R2	qm (mg/g)	E (kJ/mol)	β	R2	pH
FeB	21.97	21.70	27.4	0.02	0.13–0.9	0.98	24.21	2.02	2.18	0.98	5.37	0.26	0.96	69.84	9.6	5.3 × 10–3	0.99	5
FeCl3B	26.28	25.45	29.77	0.03	0.09–0.86	0.99	29.13	3.21	2.45	0.98	5.84	0.41	0.99	71.74	10	4.6 × 10–3	0.99	5
Zr–FeCl3B	54.39	53.62	67.28	0.05	0.05–0.77	0.99	62.74	7.37	2.11	0.98	12.81	0.76	0.97	223	9.9	4.8 × 10–3	0.99	6

Meanwhile, the R2 values of the Temkin modeled data for FeB, FeCl3B, and Zr–FeCl3B were 0.96, 0.99, and 0.97, respectively (Table ). This model (Temkin) describes the heterogeneous BC surface structure with adsorption sites that have a range of binding energies during AsV adsorption (Figure C).[80]

The high R2 (0.99) values indicated that the best isotherm fit was with the D-R model (Figure D). The higher theoretical AsV adsorption of BCs could be ascribed to the greater micro-porosity and reduced pore diameter. This result also agreed with the greater specific surface area (SSA) of the BCs.

Characterization of BC

The pH values of FeB, FeCl3B, and Zr–FeCl3B were 5.41, 5.88, and 5.64 in H2O and 5.28, 5.45, and 5.17 in CaCl2, respectively, which indicated that modified BCs showed acidic characteristics, whereas raw BSBC (pH = 7.12) showed slightly basic characteristics(Table S2). The pore volume of BSBC, FeB, FeCl3B, and Zr–FeCl3B was 0.006, 0.007, 0.017, and 0.019 cm3/g; however, the pore size was 6.51, 5.75, 4.80, and 3.92 nm for BSBC, FeB, FeCl3B, and Zr–FeCl3B, respectively. The pHPZC also increased from 3.6 to 4.7, 4.9, and 6.3 when compared from BSBC to FeB, FeCl3B, and Zr–FeCl3B. The Fe content in raw BSBC was 100.69 mg/g, whereas after modification, the Fe content was increased to 176.2, 228.2, and 238.8 mg/g for FeB, FeCl3B, and Zr–FeCl3B, respectively. The physico-chemical characterization and elemental composition of all Fe-coated BCs are presented in Tables S2 and S3. The SSA of Fe-modified BCs, specifically FeB, FeCl3B, and Zr–FeCl3B, increased to 6.6, 24.02, and 25.51 m2/g, respectively (Table S2), compared to the pristine BC (4.64 m2/g) reported earlier.[46] Thus, increasing trends of SSA in BCs (FeB < FeCl3B < Zr–FeCl3B) affected the pore size and pore volume on the coated BC surface.[81] However, surface area increases with smaller particle sizes, and this explains the higher adsorption at lower particle size.[81] The average particle size for FeB, FeCl3B, and Zr–FeCl3B was determined to be 909, 249, and 235 nm, respectively. Also, SSA correlated with an increased pore volume (Table S2).[82]

The Zr particles precipitated on the BC surface, which resulted in a rough and heterogeneous surface during synthesis of the Zr–FeCl3B composite (Figure Ci). After reacting with As, the particle size of BCs became finer, and the morphologies of FeB, FeCl3B, and Zr–FeCl3B were transformed into non-regular shaped aggregates and/or rough surfaces with elongated shards [Figure A(ii),B(ii),C(ii)].

Figure 4

SEM micrographs of FeB, FeCl3–B, and Zr–FeCl3B, before A(i)–C(i) and after A(ii)–C(ii) As-adsorption; A(iii)–C(iii) represents SEM–EDS of the corresponding As-loaded BCs.

The energy-dispersive X-ray spectroscopy (EDS) spectra shows evidence for the presence of sorbed As including Fe in FeB and FeCl3B (Figure A(iii),B(iii)). Similarly, As-spectra along with both Fe and Zr peak were confirmed in Zr–FeCl3B by EDS analysis (Figure Ciii). The amount of sorbed As was estimated to be 1.0 wt % (0.24 atomic %), 3.65 wt % (0.87 atomic %), and 3.89 wt % (1.93 atomic %), by FeB, FeCl3B, and Zr–FeCl3B, respectively (Table S4). The amount of Fe was also determined to be 24.78 wt % (7.86 atomic %), 15.32 wt % (4.91 atomic %), and 46.24 wt % (30.69 atomic %) by As-loaded FeB, FeCl3B, and Zr–FeCl3B, respectively (Table S4). Additionally, Zr particles were determined to be 16.84 wt % (6.84 at. %) in the As-loaded Zr–FeCl3B composite (Table S4). Results suggested that a large amount of As was adsorbed by the Zr–FeCl3B composite, which may have been due to the combined effect of Fe and Zr. This could be due to the higher SSA from the Zr–Fe loadings on the surface of BC and increased positive surface charge (zeta potential) produced compared to unmodified BC and high Fe content in the Zr–FeCl3B composite.

The transmission electron microscopy (TEM) elemental mapping and EDS of As-adsorbed FeB, FeCl3B, and Zr–FeCl3B are depicted in Figure . Similar to scanning electron microscopy (SEM), the HTEM images described the presence of As and other major elements like Fe and O existing on the FeB, FeCl3B, and Zr–FeCl3B BC surfaces (Figure A–C). Meanwhile, the presence of Zr particles was observed in As-loaded Zr–FeCl3B composites (Figure C). The Zr (K-line) (mass 2.93%) was located heterogeneously on the Zr–FeCl3B surface and confirmed by the TEM–EDS spectrum at 15.74 keV. The presence of As (K-line) was confirmed by TEM–EDS, while the amount of the As was determined masswise to be 0.99% As, 1.04% As, and 1.24% As in FeB, FeCl3B, and Zr–FeCl3B, respectively, after adsorption with As at 10.53 keV (right-hand corner of Figure A–C). This finding agrees with the SEM results.

Figure 5

TEM elemental distribution of the As-loaded FeB (A), As-loaded FeCl3B (B), and As-loaded Zr–FeCl3B (C) and TEM–EDS of their respective As-loaded BCs (right side).

Before AsV adsorption, the Fourier transform infrared (FTIR) spectra showed similar peaks for all BCs except for the formation of a new peak at 3695 cm–1, which may explain the Zr–O and Zr–OH–Zr vibrations,[83,84] and confirmed the successful coating of Zr along with Fe (peak at 802 cm–1) in Zr–FeCl3B (Figures A and S2A). The FTIR spectra also revealed broad band peaks at around 3400–3450 and 1620–1650 cm–1 for all BCs. The peak close to 2300–2400 cm–1 assigned to the presence of atmospheric CO2.[85] These represent stretching and vibration peaks of −O–H and H–O–H, respectively, further indicating the bending and deformation of water molecules.[66]

Figure 6

FTIR spectra of BCs before and after As-reaction (A), XPS spectra of Zr in Zr–FeCl3B (B), XPS spectra of As in As-loaded FeB (C), FeCl3B (D), and Zr–FeCl3B (E).

After AsV adsorption, a new second peak at 782 cm–1 was observed close to 802 cm–1 after reacting with As (Figure S2A), which is attributed to symmetric vibrations between As–O– and As–O–Fe complexation, which was supported by past studies.[66,86] Another new peak was observed at 3748 cm–1 in the As-reacted Zr–FeCl3B BC (Figure S2B), and this may be caused by the formation of an inner As–O–Zr complex. The spectra at 3730 and 3745 cm–1 reported the presence of a bi-bridged hydroxyl group on Zr–O2.[84,87,88] Therefore, it can be assumed that the peak 3695 cm–1 (observed in Zr–FeCl3B) (Figure S2B) could be shifted to 3748 cm–1 (close to 3745 cm–1) after the interaction between As and Zr–O groups.

The observed peaks at 22.97, 24.28, 31.16, 36.95, and 53.38° in the XRD pattern represented different types of trigonal (hexagonal axes) quartz (silica) and graphite, in Fe-coated or Zr–Fe coated BCs (Figure S3A). No confirmed peaks corresponding to Zr and/or Fe coatings in the XRD spectrum of the modified BCs were observed (Figure S3A), which suggested that Fe- and/or Zr–Fe-associated BCs existed predominantly in the amorphous phases. The formation of amorphous Fe-oxide/hydroxides on BC and Fe-granular activated carbon composites are documented in the literature.[89−91] No As-related minerals were detected when employing XRD analysis in this study (Figure S3B).

XPS analysis of the survey profile revealed enriched amounts of C, O, N, and Fe on the BCs’ surface (Figure S3). The Zr peak was observed in two binding energies at approximately 182.64 eV (Zr 3d5/2) and 185.45 eV (Zr 3d3/2), which represented Zr-oxide[92,93] on the Zr–FeCl3B surface (Figure B). XPS survey analysis also confirmed the As 3d spectrum on the BC surface (Figure S4). Ren et al. (2011)[66] and Ding et al. (2000)[94] reported that the values of As 3d core level of AsV may move to between 45.31 and 45.6 eV during the adsorption of As. In this study, the XPS spectrum at bonding energies of 45.51, 45.51, and 45.54 eV correspond to AsV in the As 3d region for FeB, FeCl3B, and Zr–FeCl3B, respectively[95] (Figure C–E).

Influence of Interfering Ions

The anions Cl–, NO3–, CO32–, SO42–, and PO43– are commonly present in natural waters, which can potentially interfere with AsV adsorption. Chloride and NO3– had no effect on AsV adsorption, which may be due to their negligible affinities for the BC surface. The AsV removal efficiency was reduced by 15, 14, and 7% in FeB, FeCl3B, and Zr–FeCl3B in the presence of SO42–, which may have little affinities toward the BC surface. The Cl– and NO3– anions mainly form outer-sphere complexes with Fe-oxy compounds[96] and thus inhibitory AsV adsorption to a minor extent on Fe-modified BC surfaces.

The AsV removal ability declined by 49, 45, and 35% in the presence of CO32–. Schmidt et al. (2020)[97] and Mendez and Hiemstra (2018)[98] reported that SO42– and CO32– can form inner-sphere complexes with Fe-oxide surfaces and thus affected AsV adsorption by Fe-modified BC. The PO43– ion was the greatest competitive anion, and it significantly inhibited AsV adsorption by 88, 85, and 75% for FeB, FeCl3B, and Zr–FeCl3B, respectively, compared to NO3– (Figure A–C). The adsorption capacity of AsV decreased from 7.62 to 0.91 mg/g (at solution pH 5), 7.98 to 1.19 mg/g (at solution pH 5), and 9.41 to 2.35 mg/g (at solution pH 6) in the presence of PO43– with FeB, FeCl3B, and Zr–FeCl3B, respectively, using 2 g/L BC doses rate and at 22 ± 0.5 °C. This substantial inhibitory effect could be due to the similar molecular structures of PO43– and AsO43–, which creates a strong competition for binding sites of BCs.[99] Moreover, PO43– is able to compete for the same adsorption sites of AsO43– forming inner-sphere complexes in the presence of Fe-oxy-hydroxides.[100,101]

Figure 7

Effect of interfering ions (A) FeB, (B) FeCl3B, and (C) Zr–FeCl3B (the initial AsV concentration was 20 mg/L, and BC density was 2 g/L at 22 ± 0.5 °C) and (D) temperature of AsV adsorption on BCs (the initial AsV concentration was 10 mg/L).

The influence of common cations such as K+, Na+, Mg2+, Ca2+, and Al3+ on AsV adsorption toward FeB, FeCl3B, and Zr–FeCl3B was assessed. No noteworthy effect was observed on AsV adsorption by FeB, FeCl3B, and Zr–FeCl3B in the presence of these cations (Figure A–C).

Effect of Temperature

The calculated thermodynamic parameters, including ΔG, ΔH, and ΔS, are listed in Table . The AsV removal efficiency rose as the temperature increased from 4 to 40 °C (277–313 K) (Figure D). The negative ΔG values indicate that the adsorption process is spontaneous at 4–40 °C (Tables and S5).[35] The mobility of AsV ions increases with increasing solution temperature; therefore, elevated AsV removal capacity was achieved by the BC composites.[102] The obtained ΔH values were positive, which demonstrated that the AsV adsorption process was endothermic for all BCs.[76]

Table 3

Thermodynamic Parameters for the Adsorption of AsV on BCs

	ΔG (kJ/mol)
BC	277 K	288 K	293 K	298 K	303 K	313 K	ΔH (J/mol)	ΔS (J/mol.K)	pH
FeB	–12.47	–12.96	–13.19	–13.42	–13.64	–14.09	13.17	45.06	5
FeCl3B	–13.1	–13.62	–13.86	–14.1	–14.33	–14.81	13.72	47.35	5
Zr–FeCl3B	–13.23	–13.76	–14.0	–14.24	–14.48	–14.95	13.6	47.82	6

The positive ΔS value means that the adsorption process was favorable but had increased randomness of AsV, leading to a more disordered state on the BC/solution interface.[35] In addition to this, ion solvation could play a role in increasing entropy (ΔS).[103,104] In this study, the higher positive change in enthalpy and increase in entropy collectively contributed to the strong spontaneous adsorption of AsV by all Fe-coated BCs.

Reusability of BCs

After successful AsV adsorption, the As-loaded FeB, FeCl3B, and Zr–FeCl3B were subjected to regeneration tests in order to determine reusability and stability of BCs. The experiment consisted of up to six adsorption–desorption cycles to examine repetitive usage of BCs under optimum conditions. Results showed that the adsorption capacities of all BCs slightly reduced after six adsorption–desorption cycles (Figure ). Very low desorption capacity was observed with MQ water for FeB (8.38–4.98%), FeCl3B (7.88–5.11%), and Zr–FeCl3B (5.77–4.72%), respectively (Figure A(i)–C(i)). The percentages of desorbed As ranged from 75.14 to 55.14, 78.11 to 73.17, and 80.43 to 74.87 up to six trials by FeB, FeCl3B, and Zr–FeCl3 BCs, respectively, when using (NH4)2SO4 (Figure Aii–Cii). The desorption efficiencies of FeB, FeCl3B, and Zr–FeCl3B decreased from 95.7 to 89.75, 96.96 to 91.4, and 99.1 to 95.73%, respectively, from the first to the sixth trial using NaOH (Figure A(iv)–C(iv)). Results showed that the desorbing solution of HNO3 followed almost the same pattern as NaOH; however, NaOH was a more efficient desorbing agent than HNO3 for all Fe-coated BCs. The leached Fe ions from FeB, FeCl3B, and Zr–FeCl3B were determined to be 0.63, 0.72, and 0.35 mg/L after the first regeneration cycle, while the concentration of Fe was 0.74, 0.86, and 0.44 mg/L after the six cycles at pH 5. Zr leached from Zr–FeCl3B was 0.08 and 0.13 mg/L after the first and sixth adsorption–desorption cycle, respectively, at pH 5.

Figure 8

Adsorption–desorption of As-loaded (A) FeB, (B) FeCl3B, and (C) Zr–FeCl3B (BC density was 2 g/L under optimum conditions).

The sequence for the examined materials’ reusability was Zr–FeCl3B > FeCl3B > FeB, respectively. Results suggested that all Fe-modified BC composites—FeB, FeCl3B, and Zr–FeCl3B—were effectively recycled and can be employed in repeated AsV adsorption batches at least six times with minimum loss in their adsorption capacities[35] using NaOH as the preferred desorption agent.

Role of Fe and Zr in AsV Adsorption

Oxides of Fe and Zr generally have strong binding affinities toward AsV. Arsenate could attract both Fe–OH and Zr–OH surface sites on the BC through the possible formation of FeAsO4 and Zr3(AsO4)4 surface precipitates.[65] The Zr–OH sites prefer to bind with AsV compared to Fe–OH sites. The reason for this may be due to the low solubility product constant of Zr3(AsO4)4 compared to FeAsO4 (Ksp = 1.47 × 10–9). Another reason could be the predominant formation of the Zr3(AsO4)4 phase rather than FeAsO4 phases, when free AsO43– concentrations were insufficient compared to additional Zr–OH or dissolved Zr4+ onto the BC surface.[65] Therefore, Zr–OH- or Zr-coated surface sites are much stronger than Fe–OH sites, which was also suggested by previous studies.[65] The maximum AsV sorption capacities of pristine and Zr-modified biosolid BC (Zr-BSBC) were reported to be 15.2 and 33.1 mg/g, respectively.[46] In the current study, binary metal Zr–Fe-coated BCs (67.28 mg/g) have shown improved AsV adsorption efficiency compared to single Fe coatings only (27.4 and 29.77 mg/g). This is attributed to the co-presence of Zr and Fe, which increased the SSA and bound more strongly with AsV. However, the surface area of Zr–FeCl3B (25.51 m2/g) was almost similar to FeB (24.02 m2/g). Previous studies showed that even single metal Zr-coated BC (75.9 m2/g) showed much higher surface area compared to binary Zr–Fe-chip-coated BC (27.9 m2/g).[46] This could be due to the geometric position (hindering effect) of the Zr and Fe atoms on the BC surface. Therefore, Zr–OH had the most active sites responding to AsV, and thus, AsV tends to bind more favorably to Zr–O moieties.[65] Despite the inability to detect Zr or As crystallinity by XRD, the SEM–EDS and TEM–EDS also support this outcome.

The presence of Fe was 6.4% Fe, 8.60% Fe, and 19.90% Fe, whereas the amounts of O in mass were 8.35% O, 12.9% O, and 19.95% O, respectively, in Fe-coated FeB, FeCl3B, and Zr–FeCl3B BC composites by TEM–EDS analysis. The ratios of O/Fe were determined to be 1.3, 1.5 and 1.0, while the ratios of O/As were 8.43, 12.4, and 16.08, respectively, in FeB, FeCl3B, and Zr–FeCl3B. However, the ratio of O/Zr was calculated to be 7.02 in the Zr–FeCl3B composite. Results revealed that at higher O/As, there was more As sorbed by BCs. The ratio of O/Fe in Zr–FeCl3B was comparatively low compared to FeB and FeCl3B BCs. This is because the additional presence of Zr with Fe in the Zr–FeCl3B composite reduced the O/Fe ratio due to the high affinity toward Zr and O rather than Fe and O (electronegativity of O, Zr, and Fe are 3.44, 1.33, and 1.83). These results clearly indicated that the role of O/Fe and/or O/Zr for enhancing the removal of As by Fe- and Fe–Zr-coated BC composites compared favorably to pristine BSBC (Table ). The XPS documented a similar result.

Table 4

Comparison of Removal Capacity of AsV Using Various Pristine and Modified BCs

BCs	AsV adsorption capacity, (mg/g)	References
BSBC	15.2	(46)
almond shell	3.6	(106)
ZnO-modified coffee husk	1.54	(107)
ZnO-modified corncob	25.9	(107)
corncob	13.06	(107)
rice straw	11.2	(108)
	0.55	(109)
red mud-modified rice straw	5.92
Fe2+- and Fe3+-modified rice straw	26.9	(108)
ZnCl2-modified crayfish shell	17.2	(110)
yak dung	1.05	(111)
FeCl2·4H2O + NaClO-modified yak dung	2.93
perilla leaf	2.95	(14)
Japanese oak wood	3.89	(37)
Fe-impregnated corn straw	6.80	(112)
FeCl3-modified Cassia fistula (golden shower)	1.07	(113)
nZVI-modified red oak	15.66	(114)
nZVI-modified switchgrass	6.48
chestnut shell	17.5	(41)
magnetic gelatin-chestnut shell	45.8	(41)
paper mill sludge	23.1	(115)
Ni/Fe-modified pinewood	6.52	(44)
rice husk	0.42	(116)
	2.59	(117)
	7.1	(118)
bismuth oxide-modified Wheat straw	16.21	(119)
empty fruit bunch	18.9	(118)
Fe2+- Fe3+-modified Water hyacinth	7.41	(120)
Zr-BSBC	33.1	(46)
Fe-chip-coated BSBC	27.4	this study
Fe-salt-coated BSBC	29.77	this study
Zr–Fe-salt-coated BSBC	67.28	this study

Redox Transformation of AsV to AsIII on BCs

Analysis of high-resolution XPS spectra showed a single peak of As 3d at the binding energies of 45.51 eV in FeB BC surface, after reacting with AsV. The peaks appearing at 45.51 eV confirmed the presence of AsV on FeB and FeCl3B BC (Figure C,D). Interestingly, two peaks of the As 3d region appeared at 45.51 eV and 43.77 eV in FeCl3B and 45.54 eV and 43.58 eV in Zr–FeCl3B BC solid surface after adsorption with AsV, which confirmed the co-existence of both AsV (45.51 and 45.54 eV) and AsIII (second small peak at 43.77 and 43.58 eV) on BC surfaces (Figure D,E and Table S6).[41,105] Results suggest that redox transformation of AsV (86.4 and 84.9%) to AsIII (13.6 and 15.1%) occurred onto FeCl3B and Zr–FeCl3B BC surfaces during adsorption despite the oxic reaction environment.

The XPS spectra revealed that the second peak produced at 185.45 eV in the Zr 3d3/2 region corresponds to the Zr–O2[93,121] (Figure B). Lewis acid–base definition classified Zr to be more basic than Fe species (Fe2+ or Fe3+) which also enables Zr to function as a reductant.

It is previously reported that FeIII could influence surface reduction from AsV to AsIII;[122] therefore, the presence of FeIII, reduction of AsV occurred in FeCl3B and Zr–FeCl3B BC surfaces. However, further research should be explored to understand the probable mechanism and the role of Fe and Zr behind this reduction. As the toxicity of AsIII is greater than AsV and it is extremely difficult to remove AsIII, this finding of As transformation has practical implications and needs more research to determine the extent of reduction from AsV to AsIII.

According to the literature, it can be assumed that a passive layer was formed on the Fe-based adsorbent surface during adsorption of CrVI, which could prevent the reduction of CrVI to CrIII.[123,124] Therefore, further studies should include the modification of BC materials to control the redox transformation of As during the adsorption process.

The XPS peaks for the Fe 2p shell in the Fe 2p3/2 (the more intensive of two peaks Fe 2p) and Fe 2p1/2 subshells (without extra excitation) and satellite structure is observed for different Fe compounds. This is due to the multiple splitting and electron shake up. The three Fe 2p3 spectra at binding energies of 711.76, 715.76, and 720.06 eV (Figure ) were detected on the surface of the FeB, FeCl3B, and Zr–FeCl3B BCs, respectively, before reaction with As (Table S6). The spectra at binding energies 711.76, 711.41, and 711.19 eV indicated the presence of Fe3+ (Figure ) as Fe2p3/2 (FeCl3 or Fe2O3),[121,125] whereas the spectra at 715.76, 714.89, and 714.46 eV represent the multiplet spectra of Fe3+ as Fe 2p3/2 (Figure ) (FeCl3 or Fe2O3) for FeB, FeCl3B, and Zr–FeCl3B, BCs.[125−127] Another XPS satellite structure of Fe3+ as Fe 2p3/2 (Figure ) was detected at 720.06, 719.42, and 719.23 eV, respectively, for FeB, FeCl3B, and Zr–FeCl3B BCs, which demonstrated aspects of overlapping undissolved oxidized Fe3+ (Fe2O3) and metallic Fe.[125,126,128] Added to these, three Fe 2p1/2 (Fe 2p1/2A, Fe 2p1/2B, and Fe2p1/2C) spectral peaks were observed at the 725–733 eV region, which describes the precipitation of combined Fe2+ and Fe3+ species on the BC surface during synthesis of Fe coatings.[126]

Figure 9

High-resolution XPS spectrum of Fe 2p before A(i)–C(i) and after A(ii)–C(ii) reaction with As by FeB, FeCl3B, and Zr–FeCl3B, respectively.

Similar XPS spectra of Fe 2p3/2 and Fe 2p1/2 peaks were detected in As-reacted BCs (Figure A(ii)–C(ii)). However, the peak intensity and BE positions of Fe 2p3/2, Fe 2p1/2, and satellite structures were reduced. The atomic percentages of Fe 2p3/2A, Fe 2p3/2B, and Fe 2p3/2C were reduced to 3–1.8, 1.02–0.52, and 0.86–0.42% for FeB, 4.02–2.48, 1.51–0.78, and 1.23–0.53 for FeCl3B, and 3.44–1.8, 1.32–0.59, and 0.98–0.4 for Zr–FeCl3B, respectively, after reacting with As (Table S6). Therefore, a minor shift in the Fe 2p and Zr 3d species occurred while reacting with As. In addition, the O1s XPS spectra divided into three forms with binding energy corresponding to 530.23–530.27 eV (lattice O2–), 531.52–531.55 eV (−OH), and 532.86–532.86 eV (C=O)[125] in FeB and Zr–FeCl3B, respectively, after reacting with As (Figure S5). However, two O 1s spectra at 531.57 eV (−OH) and 532.88 eV (C=O) were found in As-reacted FeB (Figure S5). All these results confirm the strong bonding of As with Zr and Fe species via the formation of inner sphere As–O–Fe in FeB and FeCl3B BCs;[129] and As–O–Fe and As–O–Zr complexation was evident in the Zr–FeCl3B BC. Previous studies from Peng et al. (2022)[121] and Fang et al. (2021)[130] also supported this.

Conclusions

The study demonstrated that Fe and combined Zr–Fe can be utilized to modify BCs to maximize the adsorption capacity of AsV. The adsorption capacity of Zr–Fe-modified BC was increased by 4.43, 2.45, and 2.3 folds when compared to the BSBC, Fe-BSBC, and Zr-BSBC, respectively (Table ). The adsorption of AsV is pH dependent, and the maximum adsorption was obtained under acidic conditions. The Langmuir maximum AsV removal efficiencies were found to be 27.4, 29.77, and 67.28 mg/g by FeB (at pH 5), FeCl3B (at pH 5), and Zr–FeCl3B (at pH 6), respectively. Zeta potential and pHPZC value are increased by bimetal Zr–Fe coatings (Tables S1 and S2). The presence of Zr resulted in the greatest AsV removal from solution. This may be due to the enhanced SSA with bimetal Zr–Fe coatings on the BC surface, and an increase in the positive surface charge produced compared to pristine and single Fe-modified BC. Among the anions tested, PO43– greatly competed in the adsorption process that reduced the AsV removal of 88, 85, and 75% with FeB, FeCl3B, and Zr–FeCl3B, respectively, due to the strong affinity of AsV active sites. Thermodynamic investigations demonstrated that the adsorption process is favored, and the AsV removal capacity of all Fe-coated BCs increased when the temperature increases from 277 to 313 K. Fe-modified BC is promising for AsV removal from aqueous solution as the Fe-modified BC composites are economical and efficient in removing AsV (>85%) from contaminated waters in repeated six adsorption–desorption cycles. The XPS spectra confirmed the transformation of AsV (86.4 and 84.9%) to the more toxic species AsIII (13.6 and 15.1%) and during adsorption with single FeCl3B and binary Zr–Fe-coated BSBC. Further research is required to confirm the redox transformations of As species and how to reduce such reduction to abate the potential toxicity of AsIII during adsorption. In addition, the possible role of shared charge in PO43–, SO42–, and CO32– competitive adsorption with As species needs to be explored further. This could provide useful information for practical and sustainable usage of the proposed materials.

Materials and Methods

Preparation of BC

The BS biomass (BSBM) was collected from Winmalee sewage treatment plant in Winmalee, NSW, Australia. The BSBM was stored at approximately 24 °C after being oven dried at 80 °C for 24 h. Slow pyrolysis was employed for producing pristine BC from BSBM at a peak temperature of 300 °C for 30 min at a heating rate of 7 °C min–1 as per Rahman et al. (2021c).[131] Briefly, 50 g of air-dried and ground (<1 mm, 50 mesh) biomass was employed in a ceramic crucible covered with a lid and heated in a muffle furnace under a N2 atmosphere. The resulting BSBC samples were allowed to cool at room temperature inside the furnace. Afterward, the BSBC was removed from the furnace, stored in airtight plastic containers, and preserved in a desiccator for further modifications.

Synthesis of Fe- and Zr–Fe-Modified BC

Iron chips (Fe) and iron-salt (FeCl3·6H2O) were separately used to synthesize Fe-BCs by employing an in situ precipitation method according to Rahman et al. (2021b).[85] To this end, 5.58 g Fe-chips [(Fe-chips (99.98% purity) was purchased from Sigma)] was dissolved in 100 mL of HCl (1:1) to prepare 0.1 M Fe-chips solution. Following this, 5.0 g of ground BSBC (<1 mm) was submerged into a 50 mL of Fe-chips or Fe-salt solution (0.1 M FeCl3·6H2O) (mass ratio of Fe to BC = 1:1) adjusted at pH 6.5 by adding 0.1 M NaOH. Thus, the resulting BC suspension was aged for 12 h at room temperature. A bimetal adsorbent-like Zr–FeCl3.6H2O BC composite (Zr to Fe molar ratio 1:5) was synthesized following the same method, as described by Rahman et al. (2021b).[85] All synthesized BC composites were rinsed 3–4 times with MQ water to remove any impurities followed by centrifugation (5000 rpm for 15 min) and finally dried in an oven at 80 °C. The produced BCs were preserved in a desiccator after being labeled as FeB, FeCl3B, and Zr–FeCl3B, respectively, for further experiments.

Adsorbent Characterization

The point of zero charge (pHPZC) and zeta potentials were determined using a NanoPlus HD analyzer (Micromeritics, USA). SSA, pore volume, and pore size distribution were determined by Brunauer–Emmett–Teller and Barrett–Joyner–Halenda using N2 adsorption (Tristar II 3020, Micromeritics, USA). A LECO TruMac C/N/S analyzer measured the elemental composition (C, N, and S). The surface crystallinity, morphology, and functional groups were investigated with XRD (Empyrean, PANalytical), an environmental scanning electron microscope (SEM, Zeiss Sigma, Germany), a Bruker EDS detector, and FTIR (Agilent Cary 600). Furthermore, the micromorphology was determined with high-resolution TEM (HRTEM, JEM-2100F, Japan) coupled with an EDS detector (JEOL-JED-2300). The concentration of As in the BC-aqueous phase was analyzed by inductively coupled plasma optical emission spectrometry (ICP-OES, PerkinElmer Avio 200, USA). The surface oxidation state and elemental compositions of As were detected utilizing XPS (ESCALAB250Xi, Thermo Scientific, UK). Details of each method are described in the Supporting Information section.

AsV Adsorption Experiments Using BCs

Kinetics studies were controlled at a ratio of 1 to 500 (BC to suspension) using 0.05 g of BC in 25 mL of solution of 50 mL centrifuge tubes containing 10 mg/L AsV for 7 days followed by centrifuging at 5000 rpm for 15 m. Following this, the supernatant was filtrated utilizing 0.22 μm pore size nylon membrane filters. Adsorption isotherms were conducted using the same method as the kinetics but employing various concentrations of AsV (1–250 mg/L) for a reaction period lasting 48 h. The pH versus adsorption edge experiment was carried out in the pH range 2–11 at an AsV concentration of 10 mg/L. The pH of each BC suspension was controlled by adding 0.1 M HNO3 and/or 0.1 M NaOH.

The adsorption capacity and removal efficiency (%) of AsV onto BCs were calculated, respectively, using eqs and 2.[14,46] These equations are written below as followswhere qe (mg/g) represents the adsorbed amount of AsV by the adsorbent, and Ci and Ce are the initial and equilibrium AsV concentration (mg/L), respectively, while V denotes the total volume (L) of the medium, and W stands for the weight (g) of BC.

Different ratios of BC to solution, specifically 1:100, 1:250, 1:500, 1:1000, and 1:1500 were maintained to optimize the adsorbent dosage on AsV adsorption. Competitive anions such as Cl–, NO3–, SO42–, CO32–, and PO43– and cations such as Na, K, Mg, Ca, and Al at concentrations of 0.1 M were also investigated for AsV adsorption. Various concentrations of electrolytes, these being 0.01, 0.1, 0.25, 0.5, and 1.0 M NaNO3, were employed to determine the influence of ionic strength toward AsV adsorption. The pH levels of BCs suspensions were maintained at 6.0 using HNO3 (0.1 M) and NaOH (0.1 M), preserving BC density (2 g/L) and AsV concentration (20 mg/L) at 22 ± 0.2 °C. All experiments (pH, kinetics, adsorption isotherms, and competitive ions) were conducted using 0.01 M NaNO3 as the background electrolyte, and BC density was 2 g/L at temperature of 22 ± 0.5 °C. All analyses were performed in triplicate, and the average values were recorded. These includes thermodynamic studies.

All the parameters in kinetics and isotherm models are obtained through fitting the experimental raw data. The intuitive way to fit the raw data is to do a non-linear fit with the expression directly from the kinetics and isotherms equation using OriginPro software (Version 19).

Effect of Temperature on AsV Adsorption

The thermodynamic parameters, namely, change of entropy (ΔS), enthalpy (ΔH), and the Gibbs free energy (ΔG) are important for documenting a reaction spontaneity calculated using the linearized van ’t Hoff equations as follows (eqs 3–5).[76]where R is the ideal gas law constant [8.314 × 10–3 kJ/(mol K)], T is the absolute temperature (K), and Kc is the distribution coefficient, which is the ratio of the equilibrium adsorption quantity (qe) to the equilibrium concentration (Ce) of AsV. The final equation can be written as

Based on eq , ΔH and ΔS parameters can be calculated from the slope and intercept, respectively, of the plot of ln Kc versus 1/T using eq

Six different temperatures including 4, 15, 20, 25, 30, and 40 °C were employed in order to investigate the influence of temperature on AsV adsorption by all adsorbents maintaining isobaric and isochoric thermodynamic phases at constant pressure (P) and volume (V), respectively. The calculated thermodynamic parameters including G were expressed as reaction G (the change in the free energy of the reaction).

Desorption Study and Reusability of BCs

The AsV-loaded BCs were regenerated by extraction with 25 mL of MQ water, HNO3 (0.1 M), NaOH (0.1 M), and ammonium sulfate [(NH4)2SO4] (0.05 M), after washing and oven drying in six adsorption–desorption cycles separately. This was followed by continuous shaking for 24 h at 22 ± 0.5 °C. The ICP–MS determined the desorbed AsV, while the desorption efficiency of BCs was calculated using equation (eq ).[46,80]where Cdes and Cads are, respectively, the desorbed amount (mg/L) of AsV in the solution and adsorbed amount of AsV by BCs.

Similar to the adsorption study, a 0.01 M NaNO3 solution was used as the background electrolyte, and BC density was 2 g/L at 22 ± 0.5 °C. An overview of experimental conditions are tabulated in Table S7 in Supporting Information section.