Sustainable Low-Concentration Arsenite [As(III)] Removal in Single and Multicomponent Systems Using Hybrid Iron Oxide-Biochar Nanocomposite Adsorbents-A Mechanistic Study.

Rice and wheat husks were converted to biochars by slow pyrolysis (1 h) at 600 °C. Iron oxide rice husk hybrid biochar (RHIOB) and wheat husk hybrid biochar (WHIOB) were synthesized by copyrolysis of FeCl3-impregnated rice or wheat husks at 600 °C. These hybrid sorbents were characterized using X-ray photoelectron spectroscopy, X-ray diffraction, scanning electron microscopy (SEM), SEM-energy-dispersive X-ray spectroscopy, Fourier transform infrared spectroscopy, transmission electron microscopy, physical parameter measurement system, and Brunauer-Emmett-Teller (BET) surface area techniques. Fe3O4 was the predominant iron oxide present with some Fe2O3. RHIOB and WHIOB rapidly chemisorbed As(III) from water (∼24% removal in first half an hour reaching up to ∼100% removal in 24 h) at surface Fe-OH functions forming monodentate ≡Fe-OAs(OH)2 and bidentate (≡Fe-O)2AsOH complexes. Optimum removal occurred in the pH 7.5-8.5 range for both RHIOB and WHIOB, but excellent removal occurred from pH 3 to 10. Batch kinetic studies at various initial adsorbate-adsorbent concentrations, temperatures, and contact times gave excellent pseudo-second-order model fits. Equilibrium data were fitted to different sorption isotherm models. Fits to isotherm models (based on R 2 and χ2) on RHIOB and WHIOB followed the order: Redlich-Peterson > Toth > Sips = Koble-Corrigan > Langmuir > Freundlich = Radke-Prausnitz > Temkin and Sips = Koble-Corrigan > Toth > Redlich-Peterson > Langmuir > Temkin > Freundlich = Radke-Prausnitz, respectively. Maximum adsorption capacities, Q RHIOB 0 = 96 μg/g and Q WHIOB 0 = 111 μg/g, were obtained. No As(III) oxidation to As(V) was detected. Arsenic adsorption was endothermic. Particle diffusion was a rate-determining step at low (≤50 μg/L) concentrations, but film diffusion controls the rate at ≥100-200 μg/L. Binding interactions with RHIOB and WHIOB were established, and the mechanism was carefully discussed. RHIOB and WHIOB can successfully be used for As(III) removal in single and multicomponent systems with no significant decrease in adsorption capacity in the presence of interfering ions mainly Cl-, HCO3 -, NO3 -, SO4 2-, PO4 3-, K+, Na+, Ca2+. Simultaneous As(III) desorption and regeneration of RHIOB and WHIOB was successfully achieved. A very nominal decrease in As(III) removal capacity in four consecutive cycles demonstrates the reusability of RHIOB and WHIOB. Furthermore, these sustainable composites had good sorption efficiencies and may be removed magnetically to avoid slow filtration.

Introduction

Arsenic contamination in groundwater is a global issue.[1,2] Arsenic is introduced into groundwaters naturally,[3] while mining[4,5] and nonferrous smelting[6] constitute anthropogenic sources. Arsenic was classified as a human carcinogen by the International Agency for Research on Cancer[7] and the National Research Council, DC, USA.[8] The World Health Organization and Bureau of Indian Standards recommend permissible arsenic concentrations of 10 and 50 μg/L in drinking waters, respectively.[9] Hyperpigmentation, keratosis, skin and internal cancers, and vascular diseases are some major health impacts of arsenic in humans.[6] In the environment, arsenic can exist in −3, 0, +3, and +5 oxidation states. Groundwater arsenic exists primarily as arsenite [As(III)] and arsenate [As(V)] in the pH range of 6–9.[10,11] The predominant As(III) species include uncharged H3AsO3, while primary arsenate species include monovalent H2AsO4– and divalent HAsO42–. As(V) mainly exists in oxic waters, while As(III) predominates in anoxic waters. As(III) and As(V) also coexist in both types of waters. As(III) is more toxic and difficult to remove from water versus As(V).[11]

Adsorption,[11] ion exchange,[12] membrane filtration,[6] and oxidation and reduction[13] are available methods for aqueous arsenic remediation.[11] Adsorption is easy to operate and cost-effective, minimizes sludge production, and makes adsorbent regeneration possible.[14,15] Iron-loaded biochar,[16] alumina,[17] ion-exchange resins,[11,18] iron–zirconium binary oxide,[19] schwertmannite,[20] and biomaterials from agricultural wastes[21,22] have been explored to adsorb arsenic. Goethite, magnetite, maghemite, and hematite have been used over different pH ranges for arsenic removal.[11,22] Many other adsorbents used for arsenic removal were reviewed.[11,23]

Biochar was successfully applied for the remediation of heavy metals,[24,25] fluoride,[26] pharmaceuticals,[27,28] phenols,[29,30] and other emerging contaminants[31] because of its low cost and local availability.[32] Application of iron oxide–biochar composites for arsenic removal is a relatively new approach.[33,34] The biochar provides a porous surface to disperse and hold the iron oxide adsorbent and increases iron oxide’s surface area.

The adsorption of phosphate,[35−37] arsenite,[38−40] and arsenate[20,40,41] on various iron oxides has been studied. Akaganéite has a tunnel-like microporous structure giving a higher surface area per particle size versus goethite (α-FeOOH), maghemite (γ-Fe2O3), hematite (α-Fe2O3), or magnetite (Fe3O4), but it is less robust.[42] The ability to adsorb phosphate per unit surface area is similar for these oxides, so the chemically stable magnetite with its macroscopic magnetic moment makes it an attractive candidate as an adsorbent if its surface area can be increased.[36] Magnetite nanoparticles are readily produced from aqueous Fe2+ and Fe3+-containing solution while adding the base. These particles readily aggregate, so a high surface area carrier such as biochar can be employed to dispense the particles while providing overall adsorbent particle sizes that are useful for column flow experiments and facile magnetic separation.[29,43,44]

Most methods efficiently remediate arsenite as arsenate due to the fact that arsenite [As(III)] predominantly exists as uncharged H3AsO3 at pH < 9.2. Thus, arsenic removal is usually carried out by first oxidizing arsenite [As(III)] to arsenate [As(V)], followed by As(V) removal.[45] Several iron-based adsorbents have shown good arsenic adsorption capacities. Iron–graphene oxide nanocomposites have been synthesized and characterized for the successful removal of As(III) from water.[46] Similarly, a novel material obtained by coating activated carbon with zirconium oxide and manganese dioxide[47] was utilized for removing As(III). In the presence of zirconium dioxide, As(III) was oxidized to less toxic As(V) and adsorbed over manganese oxide nanoparticles through inner-sphere complexation.[47] Several metal oxide–biochar composites were synthesized through biomass pretreatment and were used for treating arsenic in water. AlCl3-blended biochar, Cu(OH)2-blended biochar, FeSO4-blended biochar, KCl-blended biochar, and MgCl2-blended biochar composites were synthesized and utilized for As(III) removal.[48]

Nanocomposites offer a plausible solution to the shortcomings of using pure nanoparticles.[49] Application of nanomaterials exposed on host surfaces takes advantage of both the hosts and the available impregnated functional nanoparticles.[50] Hosts mitigate the release of nanoparticles into the environment and improve the stability, dispersion, and reusability.[50] Very recently, magnetic Fe3O4/Douglas fir biochar composites (MBCs) were prepared with a 29.2% wt Fe3O4 loading and successfully used to treat As(III)-contaminated water.[51] This present study has paved the way for advancing water nanotechnology from labs to larger-scale field/industrial applications, especially since recycling to remove arsenic was established.

In the present study, iron oxide–biochar hybrids were synthesized by copyrolysis at 600 °C for 1 h of FeCl3-impregnated rice or wheat husks under nitrogen. Rice husk iron oxide biochar (RHIOB) and wheat husk iron oxide biochar (WHIOB) composites were then sieved to particle sizes of 0.3–0.5 mm (30–50 mesh size). Then, these composites were used for As(III) removal. Most previous arsenic remediation studies were carried out at high As(III or V) concentrations (>5 mg/L). Some adsorbents capable of remediating arsenic at high concentrations have failed to work in low concentration range. Thus, studies here were conducted at low As(III) concentrations (50–1000 μg/L) to demonstrate sorptive As(III) removal at concentrations usually present in the actual water bodies.

Results and Discussion

Characterization

Ultimate analysis was performed to find carbon, hydrogen, sulfur, and nitrogen.[23] Percent oxygen was calculated by difference (Table ). RHIOB has a higher percent carbon (38.4 wt %) than WHIOB (36.8 wt %). Both biochars have large ash contents (RHIOB: 53.5 wt % and WHIOB: 50.5 wt %), which include Fe (RHIOB: 16.02 wt % and WHIOB: 19.03 wt %) and Si (RHIOB: 27.45 wt % and WHIOB: 12.06 wt %) as the most abundant components. Qualitative analysis was carried out by inductively coupled plasma mass spectrometry (ICP-MS) (Tables and S1). The composites were ashed (6 h at 650 °C) to find the percent iron loading and other inorganic constituents. Ash was then dissolved in the mixture of HNO3 and HF (3:1 v/v %). Conc. HCl was added to stabilize Fe and Al in solution. Digested samples were then analyzed by ICP-MS. RHIOB and WHIOB hybrid composites contained 16 and 19 wt % Fe loadings, respectively (Table ). In contrast, pristine biochars from rice and wheat husk pyrolysis alone (RHB and WHB) and their respective precursor husk biomasses (RH and WH) contain very little iron content (Table S1). Other elements (Al, Ca, K, Mg, Mn, Na, and Zn) were also present in minor quantities. This was also confirmed from their respective intense peaks in scanning electron microscopy–energy-dispersive X-ray spectroscopy (SEM–EDX) graphs (Figure a–d).

Table 1

Elemental Analysis and Surface Area Properties of Slow Pyrolysis of RHIOB and WHIOB

properties	RHIOB	WHIOB
Ultimate Analysis
C (%)	38.39	36.84
H (%)	1.21	1.19
N (%)	0.29	0.39
S (%)	0.22	0.14
O (%)a	6.62	11.07
ash (%)	53.49	50.51
H/C ratio	0.0315	0.0323
O/C ratio	0.173	0.300
Quantitative Analysis (%)
Al	1.51	0.76
As	0.01	BDL
Ca	0.56	0.46
Cd	BDL	BDL
Co	BDL	BDL
Cr	0.04	0.16
Cu	0.01	0.02
Fe	16.02	19.03
K	0.15	0.05
Mg	0.1	0.18
Mn	0.16	0.13
Na	0.06	0.03
Ni	0.01	0.01
Pb	BDL	BDL
Si	27.45	12.06
Zn	0.03	0.02
Surface Area Characterization
SBET (m2/g)	300.0	339.0
external surface area (m2 g–1)	12.71	15.03
Dubinin–Radushkevich micropore volume, W0 (cm3 g–1)	18.07	18.72
average pore diameter (Å)	0.12	0.14

Calculated by difference.

Figure 1

SEM–EDX spectra of the hybrid biochars before and after As(III) adsorption: (a) unloaded RHIOB, (b) unloaded WHIOB, (c) As(III)-loaded RHIOB, and (d) As(III)-loaded WHIOB biochar composites.

SEM–EDX analyses of RHIOB and WHIOB before and after arsenic adsorption were carried out (Figure ). C, O, Si, Cl, and Fe were the major elements observed in the SEM–EDX spectra of RHIOB and WHIOB before As(III) adsorption (Figure a,b and Table S1). Na, K, Mg, Ca, Al, Mn, Cu, Rb, Mo, W, and Pb were also observed before As(III) adsorption. Silicon content was higher in RHIOB than in WHIOB (Figure ). Traces of arsenic were also observed in both RHIOB and WHIOB samples only after arsenic adsorption had occurred, corresponding to the arsenic observed in capacity measurements (Figure ).

Adsorbent ash contents were determined by dry combustion for 6 h at 750 °C. The higher RHIOB ash content versus WHIOB (Table ) is attributed to higher RHIOB silica content than that in WHIOB. The H/C ratio is the degree of carbonization based on the original biomass’s hydrogen content. Both RHIOB and WHIOB have similar H/C ratios (Table ), indicating comparable extents of carbonization. The biochar O/C ratio reflects oxygen from biomass functional groups that have not yet been lost during carbonization.[52,53]

SEM micrographs of the RHIOB and WHIOB morphologies for pre- and post-As(III) adsorption samples are provided in Figures and 3, respectively. RHIOB samples at 200× and 500× magnification exhibit conical arrays of raised surface features lined up in rows, which are characteristic of original rice husk morphology (Figure a,c). Pores at the edges of RHIOB were visible at 2k× magnifications (Figure b,d), which also reflect their original morphological features. RHIOB exhibited smooth surfaces before As(III) adsorption (Figure a,b), which have roughened after adsorption. Some partially clogged pores are seen after As(III) adsorption (Figure c,d). Similarly, WHIOB surfaces appear to have adjacent parallel canals leftover from the wheat husk morphology (Figure a). Honeycomb-like pores with an average diameter of 22 μm were observed at 2k× magnifications (Figure b). At 1k× and 3k× magnifications after As(III) adsorption (Figure c,d), this surface morphology roughened because of surface deposits formed during adsorption (Figure a,b).

Figure 2

SEM micrographs of RHIOB at different magnifications: (a) 200× and (b) 2k× before As(III) adsorption and (c) 500× and (d) 2k× after As(III) adsorption.

Figure 3

SEM micrographs of WHIOB at different magnifications: (a) 500× and (b) 2k× before As(III) adsorption and (c) 1k× and (d) 3k× after As(III) adsorption.

X-ray photoelectron spectroscopy (XPS) is a surface-sensitive technique for identifying elements, their oxidation states, and approximate surface region abundances. It analyzes ejected electrons while probing to about 100 Å below the surface, with increasing sensitivity the closer these elements are to the surface.[33] XPS analyses were carried out using X-ray monochromators in combination with Al Kα radiation (hν = 187.85 eV; power = 24.6 W; and beam spot size = 100 μm). High-resolution deconvoluted XPS spectra for Fe, C, O, and As are shown for RHIOB (Figure ) and WHIOB (Figure ). Deconvolution was done using Gaussian–Lorentzian sum function in OriginLab 2018(b) software. Comparative wide-scan spectra of pristine rice husk and wheat husk biochars and their respective iron oxide-loaded RHIOB and -WHIOB after arsenic adsorption are shown in Figure S1. Pristine rice husk and wheat husk biochars show no Fe 2p, As 3s, and As 3d peaks. RHIOB and WHIOB, in contrast, displayed peaks in these regions. This confirms successful loading of iron-(oxides) and adsorption of As in RHIOB and WHIOB, which otherwise is not present in their respective pristine biochars. For a more detailed study, high-resolution deconvoluted scans for Fe, O, C, and As were obtained for RHIOB and WHIOB after As(III) adsorption.

Figure 4

Full range XPS spectra of RHIOB after As(III) adsorption. High-resolution deconvoluted XPS spectra of RHIOB: (a) Fe 2p, (b) O 1s, (c) C 1s, and (d) As 3d.

Figure 5

Full range XPS spectra of WHIOB after As(III) adsorption. High-resolution deconvoluted XPS spectra of WHIOB: (a) Fe 2p, (b) O 1s, (c) C 1s, and (d) As 3d.

The iron peak at ∼706.7 eV was assigned to the Fe 2p3 peak. Both RHIOB and WHIOB displayed peaks at 710.17 and 710.15 eV, respectively, belonging to deposited iron oxide[51] (Figures and 6). Fe 2p spectra of RHIOB and WHIOB were deconvoluted to three components, differentiating peaks at 710.15 and 723.71 eV, which correspond to Fe 2p3/2 (RHIOB) and Fe 2p1/2 (WHIOB), respectively. These two peaks correspond to Fe(III) and Fe(II) iron atoms occurring in Fe3O4[54−56] and/or FeOOH[4,55] (Figures a and 5a). The small satellite peak at 715.5 eV that is due to charge transfer between Fe(II) and Fe(III) illustrates mixed oxide formation as in Fe2O3 [RHIOB (Figure a); WHIOB (Figure a)].[57]

Figure 6

TEM micrograph of RHIOB after As(III) adsorption (a) at 100k× magnification, (b) lattice fringes of Fe3O4, and (c) SAED pattern of the 220 plane of Fe3O4. TEM micrograph of WHIOB after As(III) adsorption (d) at 100k× magnification, (e) lattice fringes of Fe3O4, and (f) SAED pattern of the 220 plane of Fe3O4.

High-resolution O 1s XPS spectra were obtained for RHIOB and WHIOB after As(III) adsorption. The O 1s peaks were deconvoluted to three components. The peaks at 528.91 eV (RHIOB) and 528.73 eV (WHIOB) are attributed to O2– species. The second peak at 531.7 eV in both RHIOB and WHIOB is assigned to carbonyl oxygens or C=O, while the third peak from ∼532.0 to 533.7 eV is attributed to −OH groups (in phenolic hydroxyls and H2O of the biochar and surface Fe–OH groups of Fe3O4). The binding energies of O2– species shown in Figure b (RHIOB) and Figure b (WHIOB) correspond to iron-bound oxygen species present in Fe–O (FeO, Fe3O4, FeOOH).[58,59] The O 1s binding energies corresponding to carbonyl oxygen (C=O) and phenolic hydroxyl oxygens (−OH) result from incomplete biomass deoxygenation and partial carbonization. They are present at higher binding energies than at the oxygen binding energies of the iron oxide oxygens.[60,61] The O 1s binding energies of phenolic hydroxyl groups and other oxygen of biochar cannot be distinguished from small amounts of adsorbed water.[61,62] The iron oxide oxygen peak area in WHIOB is greater than in RHIOB, which is in agreement with the greater metal oxide (Fe–O) loading in WHIOB determined in elemental analysis.

The C 1s spectra were deconvoluted to three components, differentiating into peaks at 284.3 and 285.15 eV and a broadband from ∼286 to 289 eV in both RHIOB (Figure c) and WHIOB (Figure c). The peak at 284.3 eV represents C–H and C–C stretch.[55,63,64] The peak at 285.15 eV represents C–OH.[63,65,66] The presence of sp2 contributions is in accord with the formation of partial graphene-like structures.[66] This is an expected pattern for biochars.[67] The broad flat peak ranging from 286 to 289 eV includes −CO2H, −CO2R, and carbonates, which occur at the high binding energies above the −C=O region. This broad peak and the way the deconvolution was performed make it difficult to determine if a detectable C=O peak exists.

The high-resolution XPS spectra of As 3d in RHIOB (Figure d) and WHIOB (Figure d) were deconvoluted. A single component peak appears at 44.1 eV for RHIOB and WHIOB. This proves that As(III) was adsorbed by both RHIOB and WHIOB, as discussed in Section and Schemes and 2.[55,56,68,69] The As 3d peak for WHIOB may have a higher binding energy peak based on the shoulder near 45 eV, but this is hard to confirm. The absence of clearly deconvoluted components at higher binding energies makes it difficult to know if As(V) oxidation has occurred. As(V) is known to appear at higher binding energies in the species AsO43– (44.9 eV), HAsO42– (45.5 eV), and H2AsO4– (46.7 eV).[56] This “regular peak” for As 3d for WHIOB (Figure d) shows a possible shoulder at ∼44.8 eV. This might indicate that a As(V) peak occurs that was not caught in the deconvolution. Thus, we are unable to demonstrate that oxidation of As(III) to As(V) had occurred, but in the case of WHIOB, it cannot be ruled out.

Scheme 1

Dissociative and Associative Paths for Monodentate As(III) Complex to a Monodentate Corner-Sharing 2C or 1V Complex Formation

Scheme 2

Dissociative and Associative Paths for a Monodentate As(III) Complex Conversion to a Bidentate Edge-Sharing 2C or 1V Complex

Transmission electron microscopy (TEM) micrographs of RHIOB and WHIOB after As(III) adsorption are shown in Figure a–f. Magnetite (Fe3O4) nanoparticles are clearly observed for RHIOB (Figure a) and WHIOB (Figure d). Magnetite primary particle diameters range from 2 to 20 nm. Some primary Fe3O4 particles appear spherical, while others are hexagonal on both RHIOB and WHIOB and form loose sheet-like or clustered networks (Figure a,d). Lattice fringes of 0.291 Å width were captured for Fe3O4 on RHIOB (Figure b) that correspond to the 220 plane of Fe3O4.[54] The selected area electron diffraction (SAED) pattern of Fe3O4 on RHIOB (Figure c) showed fringe widths of 1.52, 2.08, 2.56, and 2.98 Å confirming Fe3O4 crystallinity.[70] A lattice fringe of 0.268 Å was found for Fe3O4 on WHIOB (Figure e). The Fe3O4 particles on WHIOB gave a SAED diffraction pattern, which showed fringes confirming a crystalline structure (Figure f).

X-ray diffraction (XRD) patterns of RHIOB and WHIOB are presented in Figure , and their 2θ (degrees), spacing (Å) values, and assignments appear in Table . Changes due to iron impregnation and 600 °C pyrolysis are observed by comparing the XRD patterns for the precursor rice husk (RHB) and wheat husk (WHB) biochars with RHIOB and WHIOB, respectively (Figure ). The XRD patterns of RHIOB and WHIOB were also compared to those of Fe2O3 and Fe3O4. RHB and WHB had a broad peak from 2θ from 19 to 29°. RHIOB and WHIOB also contained this same broad diffraction where the peak maximum was shifted to slightly higher 2θ values. These represent an amorphous range of interaromatic plane distances of the partially graphitized biochars. The wide diffraction peaks at 26.5 and 44.7° correspond to the (002) and (100) planes of graphite (JCPDS card no. 41-1487).[71−73] These peaks correspond to a combination of carbon in sp2 (26.5°; d spacing 3.35 Å) and sp3 (44.7°; d spacing 2.06 Å) hybridization states.[72,73] The C(002) plane represents parallel and azimuthal orientation of the aromatic lamellae, while C(100) is indicative of the aromatic lamina size. Greater peak sharpness indicates a more aromatic lamellae formation and progressive aromatic ring condensation.[73,74] Other XRD peaks common to RHB, WHB, RHIOB, and WHIOB belonged to the mineral oxides present[74] (SiO2, CaCO3, CaO, and MnO2), often found in biochars. These peaks occurred at 26.7, 38.4, 50.2, 64.9, 68.1, 76.6, and 83.2° and are assigned specifically to SiO2 (quartz), CaO (lime), CaCO3 (calcite), and MnO2 in Table .[72,75−77]

Figure 7

XRD spectra of (a) RHB, (b) WHB, (c) RHIOB after As(III) adsorption, (d) WHIOB after As(III) adsorption, (e) RHIOB before As(III) adsorption, and (f) WHIOB before As(III) adsorption and pure magnetite (i) akagenite (β-FeOOH), (ii) hematite (α-Fe2O3), (iii) pure magnetite (Fe3O4), and (iv) different mineral oxides (Table ).

Table 2

XRD Peaks and Corresponding Planes Identified in RHIOB and WHIOB

2θ (deg)	2θ (deg)
RHB and WHBa	RHIOB and WHIOBa	plane (h k l)	spacing (Å)	JCPDS card no.	Fe oxide(s)	others	ref.
	24.2	0 1 2	3.68	84-0310	γ-Fe2O3		(71)
	25.5	1 1 0	3.49	84-0310	γ-Fe2O3		(78, 79)
26.5	26.5	0 0 2	3.36	41-1487		C-sp2	(3, 72)
26.7	26.7	1 0 0	3.35	46-1045		SiO2	(75)
	31.8	2 1 0	2.93	34-1266	FeOOH		(80, 81)
	33.1	1 0 4	2.70	84-0310	γ-Fe2O3		(71)
	35.7	3 1 1	2.52	19-0629	Fe3O4		(34)
	37.1	2 1 −2	2.48	34-1266	FeOOH		(80, 81)
37.4	37.4	1 0 0	2.40	011-1160		CaO	(76)
	41.7	4 0 0	2.08	19-0629	Fe3O4		(34)
44.7	44.7	1 0 0	2.06	41-1487		C-sp3	(72, 74)
50.2	50.2	1 1 2	1.82	46-1045		SiO2	(75)
	53.4	4 2 2	1.71	19-0629	Fe3O4		(34)
	54.4	1 1 6	1.69	84-0310	γ-Fe2O3		(71)
	57.1	5 1 1	1.61	19-0629	Fe3O4		(34)
	62.5	4 4 0	1.48	19-0629	Fe3O4		(34)
	64.0	3 0 0	1.45	84-0310	γ-Fe2O3		(71)
64.9	64.9	0 0 2	1.44	44-0141		MnO2	(72)
68.1		2 0 3	1.37	46-1045		SiO2	(75)
	70.9	6 2 0	1.33	19-0629	Fe3O4		(34)
	72.3	1 1 9	1.31	84-0310	γ-Fe2O3		(71)
76.6	76.6	4 2 0	1.24	05-0586		CaCO3	(77)
	78.9	4 4 4	1.21	19-0629	Fe3O4		(34)
83.2		2 1 2	1.16	44-0141		MnO2	(72)

The same peaks were seen for both RHB and WHB.

The same peaks were seen for both RHIOB and WHIOB.

The sharp peaks in RHIOB and WHIOB, but absent from RHB and WHB, belonged to iron oxide crystalline phases formed from FeCl3 during pyrolysis, cooling, and exposure to air. The observed peaks clearly demonstrated that two iron oxide species formed. Diffraction peaks centered at 35.7° (311), 41.7° (400), 53.4° (422), 57.1° (511), 62.5° (440), 70.9°(620), and 78.9° (444) all correspond to Fe3O4 (magnetite) peaks (JCPDS card no. 19-0629).[34] Peaks at 24.2° (012), 25.5° (110), 33.1° (104), 54.4° (116), 64.0° (300), and 72.3° (119) are attributed to Fe2O3 (maghemite) (JCPDS card no. 84-0310).[71,78,79] Two extra peaks observed at 31.8° (210) and 37.1° (21–2) could possibly belong to small amounts of FeOOH (goethite) (JCPDS card no. 34-1266).[80,81]

Fourier transform infrared (FTIR) spectra before and after As(III) loading were obtained for both RHIOB and WHIOB (Figure ). Several sharp peaks observed from 3935 to 3510 cm–1 correspond to O–H stretching vibrations of absorbed water molecules[54,82] (Figure ). The broad peak at 3430–3460 cm–1 in RHIOB and WHIOB spectra, before and after As(III) adsorption, correspond to O–H stretching modes of surface Fe–OH groups on the magnetite and maghemite particles[83,84] on the surface of rice and wheat husk biochars (Figure ). Organic hydroxyls (especially phenolic hydroxyls) present in biochar would occur from 3600 to 3400 cm–1. The peak at 2360 cm–1 is due to O=C=O vibration in CO2.[82] The broad peak region from 1750 to 1650 cm–1 is due to various saturated and unsaturated carbonyl functional groups in biochars.[25] Distinct peaks at 1115–1057 cm–1 depict both sp3- and sp2-hybridized C–O vibrations in C–O–C and C–O–H functions.[34,83,84] The sharp peak at 669 cm–1 is due to C–C–C angle bending.[25] The two sharp peaks at 584 and 456 cm–1 are Fe–O stretching vibrations of iron atoms occupying octahedral and tetrahedral sites, respectively, in Fe3O4[83] in good agreement with reported values.[30] One key absorbance at 835 cm–1 appeared only after As(III) adsorption [depicted in (red) Figure a,c] in both RHIOB and WHIOB. This is due to the As–O stretching vibration of As–O–Fe.[85]

Figure 8

FTIR spectra of (a) RHIOB after As(III) adsorption, (b) RHIOB before As(III) adsorption, (c) WHIOB after As(III) adsorption, and (d) WHIOB before As(III) adsorption.

Because magnetic Fe3O4 phases were confirmed on both adsorbents, saturation magnetization curves were measured (Figure S2). RHIOB exhibited a magnetic moment of ∼3 emu/g at 300 K, whereas WHIOB had a magnetic moment of 8 emu/g at 300 K. However, at 10 K, magnetic moments of 6 and 12 emu/g were observed for RHIOB and WHIOB, respectively (Figure S2). The saturation magnetization values suggest Fe3O4 crystallinity. This confirms that iron loading over precursor rice husk and wheat husk biochars resulted in magnetization because of formation of iron oxide nanoparticles.[86,87]

The point of zero charge, the pH where the sorbent’s net surface charge is zero,[88,89] provides information on ionization of adsorbent surface groups and the interaction with existing solution’s adsorbate species.[88,90] The pHpzc was determined at 25 °C from the ΔpH versus pHinitial plot (Figure S3). The pHpzc values of RHIOB, ∼5.5, and WHIOB, ∼7.0, were lower than those of RHB, 6.8, and WHB, 8.3, respectively.

Sorption Studies/pH Dependence

Sorption studies were conducted at various initial pHs, adsorbent–adsorbate concentrations, temperatures, and contact times to optimize the As(III) sorption process. The As(III) speciation (Figure ) and surface charge experienced by the adsorbate species are greatly affected by solution pH.[82,91] H3AsO3 greatly predominates until pH 8.0 and constitutes above half of As(III) in solution at pH 9.2 with H2AsO3– the other half. HAsO32– appears as pH exceeds 11 and reaches a maximum at pH ≈ 12.8. The effect of pH from 2.0 to 10.0 was investigated for As(III) adsorption on RHIOB, WHIOB, RHB, and WHB at an initial As(III) concentration of 100 μg/L and an adsorbent dose of 2.0 g/L (Figure ). Over this pH range, only H3AsO3 needs to be considered up to slightly above pH 7.0 and beyond that both H3AsO3 and H2AsO3– could adsorb. High As(III) removal occurred over this entire pH range with WHIOB (∼97%) and RHIOB (∼85%) (Figure ) versus modest removal with WHB (∼20%) and RHB (∼12%) at their highest points (Figure ). Thus, iron oxide surfaces adsorb most of As(III) in WHIOB and RHIOB, whereas RHB and WHB can only uptake As(III) on the biochar and its surface mineral ash contents. WHIOB removed more As(III) than RHIOB over the entire pH range (Figure ), most likely because of its higher Fe3O4 loading (23.8 wt % in WHIOB vs 18.2 wt % in RHIOB). The higher WHIOB Fe3O4 loading provides more surface area for adsorption. The As(III) adsorption efficiency increased as pH rose from 2.0 to 7.0 and then dropped slightly. However, substantial adsorption occurred over the entire pH range. Equilibrium pH values were also recorded (shown in Figure S4) and rose for WHIOB, RHIOB, WHB, and RHB in the pH range from 3.0 to 4.0 (Figure S4). A very small equilibrium pH rise occurred for WHIOB and RHIOB in the pH range of 4.0–8.0. At high pH (>8.0), the equilibrium pH values using RHIOB and WHIOB increased to ∼5.0 (Figure S4). Further kinetic and equilibrium experiments were conducted in the pH range of 6.5–7.5.

Figure 9

pH dependence of As(III) on RHIOB and WHIOB (adsorbent dose = 2.0 g/L; initial As(III) concentration = 100 μg/L; agitation speed = 100 rpm; temperature = 25 °C; contact time = 24 h) and fractional composition curves for As(III) speciation.

As(III) Removal Mechanism

Both adsorbent surfaces are net positively charged at pH < pHpzc (Figure S3), leading to increased electrostatic attraction of oxyanion As(III) species to the surface. However, both RHIOB and WHIOB adsorbents have hybrid surfaces made up of Fe3O4 phases, where most adsorption occurs, and biochar phases where less adsorption occurs. The local PZC values at the iron oxide and at the biochar are not measured and do not correspond to the net measured PZC values. The PZC value of Fe3O4 is independently known to be 7.4. The local pHZPC of the iron oxide surfaces mostly influences As(III) adsorption; these sites compete for neutral H3AsO3 and H3O+ ions in acidic media[4] and H3AsO4, H2AsO3–, and OH– in basic pH. As pH increases, the amount of both biochar surface and iron oxide protonation drops and the local net surface charge approaches zero at the iron oxide’s pHpzc (Figure ). The greatest total As(III) removal was achieved in the pH range (6.5–7.5) where the surface Fe–OH groups on magnetite are not highly protonated or deprotonated. This suggests that H3AsO3 reacts with Fe–OH to generate Fe–O–As(OH)2 chemisorbed groups. This process could be either acid- or base-catalyzed. As H2AsO3– becomes more abundant (pH 8.5 → 10), it could also react with Fe–OH surface groups, but it would be electrostatically repelled from the deprotonated Fe–O– surface sites. However, Fe–O– sites do not repel H2AsO3 even at pH 10.0. As pH continues to rise, the magnetic surface increasingly deprotonates and becomes more negatively charged. Also, H2AsO3– and eventually HAsO32– predominate and experience more surface repulsion. This could account for the drop in As(III) removal at the higher pH ranges. The neutral As(III) species, H3AsO3, most likely dominates adsorption over the highest adsorption capacity range.

Magnetite contains both Fe(II) and Fe(III) oxidation states. FeIII ions occupy octahedral sites, whereas FeII occupies both octahedral and tetrahedral sites.[92,93] At lower solution pH(s), the iron oxide surface Fe–OH groups are protonated to generate positively charged Fe–+OH2 sites (see Scheme ).[34] Conversely, at high pH, Fe–OH groups are deprotonated to Fe–O– species.[70,82] The pHZPC of magnetite is close to neutral (7.4),[82,93] and the maximum uptake of As(III) on both RHIOB and WHIOB was achieved at ∼pHpzc of magnetite, in agreement with the previous suggestion of Fe–OH sites reacting with H3AsO3.

Substantial As(III) removal occurs over pH 2 to 10, so it seems clear that H3AsO3, H2AsO3–, and HAsO32– (the latter two at surface Fe–OH sites) can each be adsorbed depending on pH.[4,90,94,95] At pH < pHPZC of iron oxide (∼7.4), Fe–OH surface groups are increasingly protonated to generate Fe–+OH2 groups (Figure ) and can react with the abundant H3AsO3. Oxyanions H2AsO3– and HAsO32– will not be available at acidic pH values, but some H2AsO3– might be available at pH 8.0 and a small amount of Fe–+OH2 may also exist. If these reacted with a fast enough rate constant, this reaction might compete. Scheme summarizes possible mechanisms to monodentate As(III) chemisorbed surface species. H3AsO3 might react with iron at Fe–+OH2 sites by an associative mechanism, followed by loss of water. Alternatively, loss of water from a surface iron may occur first after protonation, followed by the attack of H3AsO3 or H2AsO3– at that iron (e.g., a dissociative mechanism). Both routes are shown in Scheme . These processes form a monodentate bond from As(III) to through an oxygen to magnetite. Monodentate complexes could proceed further to form bidentate corner-sharing, 2C, and/or edge-sharing, 2E, complexes (Scheme ), analogous to the behavior of arsenate tetrahedra, previously confirmed by extended X-ray absorption fine structure spectroscopy and ab initio calculations.[96] In this analogy to As(V), monodentate As(III) complexes (1V) would be unstable and 2C bidentate complexes would predominate. The 2C complexes of As(V) were predicted to be 55 kJ/mol more stable than the “edge-sharing” bidentate complexes (2E) where arsenic is chelated to a single iron atom.[96] The relative stabilities of 1V complexes of As(III) versus As(V) are not known, but the monodentate 1V complexes of As(III) are likely to either dissociate chemisorbed As(III) back to solution or form more stable 2C bidentate complexes to adjacent iron surface sites. Whether associative or dissociative mechanistic pathways are occurring is unknown, and this distinction may be pH-dependent.

Because biochar surfaces still contain some organic and inorganic −OH groups (from alkali and alkaline hydroxides resulting from the ash fraction) after heating 1 h at 600 °C, some As(III) uptake will occur on the biochar surface. This amount is much lower than occurs after magnetite has been deposited on the surface (Figure ). The biochar primarily serves to disperse Fe3O4 and other iron oxides, where most As(III) adsorption occurred. Also, biochars from rice and wheat husks, RHB and WHB, may have chemically different surfaces than the biochar portions of RHIOB and WHIOB because the pyrolyses that formed RHB and WHB were not performed with the imbibed FeCl3 in them at 600 °C. Therefore, RHB and WHB surfaces cannot be used as reliable surface adsorption models for estimating the As(III) uptake by the biochar portions of RHIOB and WHIOB.

As(III) Adsorption Kinetics and Modeling

The pseudo-second-order equation best described As(III) sorption on RHIOB and WHIOB (Figures S5 and S6; Table ). Sorption equilibrium was established in 12 h. The first-order fits were very poor, so these data are not reported. The experimental linear pseudo-second-order qe values also agreed with the calculated qe values, and the linear pseudo-second-order R2 values were excellent (Table ). Second-order kinetics is consistent with rate-determining As(III) chemisorption over magnetite surface sites.[82] As(III) uptake at different adsorbent dosages was higher on WHIOB than on RHIOB (Table ) because of WHIOB’s higher iron loading (Table ). As As(III) concentrations rise, the time-dependent adsorption efficiencies for RHIOB and WHIOB also rose.

Table 3

Pseudo-Second-Order Rate Constants Obtained from the Linear Equation and Comparison of Experimental qe Values with Their Corresponding qe Calculated Second–Order Values Obtained at Different Adsorbent Dosages and As(III) Concentrations

	second-order linear plot		As(III) amount adsorbed (qe)
adsorbent dose (a) and adsorbate As(III) concentration (b)	second-order rate constant, k2 (g/μg h)	R2	qe, obtained experimentally (μg/g)	qe, calculated using second-order linear equation (μg/g)	half-life, t50 (h)
(a) At Different Adsorbent Dosages (g/L) at [As(III)]= 100 μg/L
RHIOB
0.5	0.003	0.919	142.0	142.9	14.5
1.0	0.001	0.992	109.9	125.0	7.5
2.0	0.006	0.994	58.6	62.5	5.5
WHIOB
0.5	0.002	0.979	111.8	125.0	8.5
1.0	0.013	0.998	50.5	52.6	2.5
2.0	0.032	0.998	27.2	27.8	0.5
(b) At Different As(III) Concentrations (μg/L) at an Adsorbent Dose of 2 g/L
RHIOB
50	0.035	0.996	8.49	8.3	0.25
100	0.014	0.992	26.1	27.0	3.5
200	0.006	0.994	58.6	62.5	4.5
WHIOB
50	0.035	0.993	15.0	15.9	0.5
100	0.010	0.999	53.1	54.6	2.5
200	0.009	0.998	69.9	70.5	3.0

Adsorption half-life, t1/2, is the time when the reactant concentration decreases by half, that is, 50% of total adsorption.[25] The t1/2 values for As(III) adsorption on RHIOB and WHIOB at different initial adsorbent dosages and concentrations are listed in Table . The As(III) adsorption half-life values obtained for RHIOB and WHIOB dropped as adsorbent dosages rose (from 0.5 to 2.0 g/L). Half-life, t1/2, values increased with a rise in As(III) concentrations (50–200 μg/L) for RHIOB and WHIOB (Table ).

Bt values were obtained for each F value from Reichenberg’s table[97] at different concentrations. Bt versus time plots for RHIOB and WHIOB at different As(III) concentrations were constructed (Figure S7a,b). RHIOB and WHIOB gave linear plots at all the concentrations. However, at lower As(III) concentration (50 μg/L), the linear Bt versus time plot does not pass through origin. Thus, particle diffusion is the rate-limiting step. However, at higher concentrations (100 and 200 μg/L), the Bt versus time lines pass through origin. Therefore, film diffusion is the rate-limiting step at these As(III) concentrations. Effective diffusion coefficients were calculated from Bt versus time plots using eq . These are summarized in Table . Di values of RHIOB and WHIOB rose as the As(III) concentration increased from 50 to 100 μg/L and then fell with a further concentration rise from 100 to 200 μg/L. This drop in D may result from a mobility decrease of As(III) ions because of the increase in retarding force acting on diffusing ions.

Table 4

Effective Diffusion Coefficients for As(III) Adsorption on RHIOB and WHIOB at Different As(III) Concentrations

	adsorbents
	RHIOB			WHIOB
parameters	50 μg/L	100 μg/L	200 μg/L	50 μg/L	100 μg/L	200 μg/L
R2	0.9926	0.9809	0.9884	0.9953	0.9866	0.9848
radius of the adsorbent particle, r0 (nm)	7.68	7.68	7.68	6.95	6.95	6.95
effective diffusion coefficient, Di (m2 s–1)	54.5 × 10–16	71.1 × 10–16	52.9 × 10–16	88.7 × 10–16	94.2 × 10–16	82.9 × 10–16
Boyd number, B (s–1)	0.091	0.119	0.088	0.148	0.158	0.139

As(III) Adsorption Equilibrium Studies

Sorption isotherm experiments were conducted at 10, 25, and 40 °C between pH 6.5 and 7.5 and initial As(III) concentrations varying from 50 to 1000 μg/L (Figure ). This concentration range was selected based on arsenic concentrations reported worldwide in ground and surface waters. Equilibrium As(III) adsorption amounts on RHIOB and WHIOB increased going from 10 to 45 °C (Table S2). Thus, adsorption is endothermic. With rising temperature, the adsorptive capacity rose with increasing adsorbent mobility.[34] WHIOB’s adsorption capacity was higher than RHIOB’s, in accord with WHIOB’s modestly higher surface area and significantly higher iron content (Table ).

Figure 10

Langmuir nonlinear adsorption isotherms of As(III) adsorption by (a) RHIOB and (b) WHIOB at different temperatures (pH = 7.5; concentration = 100 μg/L; temp = 25 °C; agitation = 100 rpm; particle size = 0.3–0.5 mm (30–50 B.S.S. mesh size).

Freundlich,[98] Langmuir,[99] Temkin,[100] Sips or Langmuir–Freundlich,[101] Redlich–Peterson,[102] Radke and Prausnitz,[103] and Toth[104] isotherms (Supporting Information, eqs S8–S14, respectively) were used to fit the sorption equilibrium data and to determine the adsorption behavior, capacity, and parameters used in fixed-bed reactor design. The Freundlich isotherm[98] describes (eq S5) the adsorption equilibrium on heterogeneous surfaces. The Langmuir isotherm (eq S6) assumes a uniform sorbent surface sites with identical energies,[99] where adsorption occurs as a single surface adsorbate layer. The Sips isotherm[101] is a combination of the Langmuir and Freundlich isotherms (eq S7). The Redlich–Peterson model (eq S8)[102] is a three-parameter model, describing equilibrium on heterogeneous surfaces and contains a heterogeneity factor. The Toth isotherm,[104] an empirical model, assumes an asymmetric quasi-Gaussian distribution of site energies (eq S9). This model describes an improved fit versus Langmuir isotherm model and is often used to describe heterogeneous systems. The Temkin model is described by eq S10. The three-parameter Radke and Prausnitz isotherm,[103] derived from thermodynamic considerations (eq S11), is capable of describing data over a wide concentration range. The Koble–Corrigan isotherm[105] is also a three-parameter model (eq S12).

Langmuir fittings of As(III) sorption equilibrium data for WHIOB and RHIOB are given (Figure ) with the fittings for the other isotherm models (Figures S8–S14). Parameters and regression coefficient from all eight of these nonlinear isotherm fittings are summarized in Table S2. The fits at 25, 35, and 45 °C for all RHIOB and WHIOB adsorption data were ranked based on their R2 and χ2 values. At 45 °C, the Redlich–Peterson (R2 = 0.9909; χ2 = 0.003) equation fits the RHIOB equilibrium data best, while the Sips (R2 = 0.9258; χ2 = 0.0053) and Koble–Corrigan (R2 = 9881; χ2 = 0.0069) gave equal best fits to WHIOB data. The Redlich–Peterson isotherm constant β values for RHIOB at 45 °C are very close to 1 (Table S2), reducing As(III) adsorption to the Langmuir isotherm,[106,107] suggesting that RHIOB has a homogeneous adsorption surface. This agrees with predominant adsorption on the iron oxide surfaces of the hybrid adsorbent. WHIOB data give best fits by the Sips and the Koble–Corrigan isotherms, which combines both Langmuir and Freundlich. Hence, at lower As(III) concentrations, diffusion dominates, whereas at higher As(III) concentrations, monomolecular adsorption dominates.[24,29]

Monolayer adsorption capacities (Q0) are reported from the Langmuir adsorption isotherm (Figure ; Table ). These maximum adsorption capacities were 13.14 μg/g (10 °C), 94.34 μg/g (25 °C), and 96.14 μg/g (45 °C) for RHIOB and 70.35 μg/g (10 °C), 108.0 μg/g (35 °C), and 110.73 μg/g (45 °C) for WHIOB (Table , Figure ). These values are very close to the experimental capacities of 11.63, 91.91, and 94.34 μg/g for RHIOB and 72.46, 97.09, and 103.09 μg/g for WHIOB, respectively (Table S2). Adsorption capacities increase with an increase in temperature for both RHIOB and WHIOB, agreeing with endothermic As(III) adsorption. This As(III) adsorption increase with a rise in temperature may be due to both endothermic chemical adsorption and an increase in adsorbent ion mobility. Both would raise adsorption efficiencies as the temperature rises.[34] The Langmuir monolayer adsorption capacities of RHIOB and WHIOB are comparable to that of many adsorbents (Table ). High sorption capacities of many adsorbents are due to the high As(III) concentrations taken in performing the batch experiments. Some adsorbents capable of remediating arsenic at high concentrations have failed to work in low concentration range. Our present studies were purposely conducted at low As(III) concentrations (50–1000 μg/L) where less data exist and because many environmental sites have concentrations in this range. Therefore, these studies may result into low adsorption capacities. RHIOB and WHIOB successfully remediated As(III) at concentrations usually present in the actual water bodies (Table ).

Table 5

Comparison of Adsorption Capacities of RHIOB and WHIOB vs Other Adsorbents Used for As(III) Removal from Water

			adsorption study parameters
s. no	raw material	BET surface area (m2/g)	pH	temp (°C)	concentration range (mg/L)	Langmuir adsorption capacity Q0 (mg/g)
1	pine wood biochar[23]	2.73	3.5	25	0.075–37.46a	1.2
2	oak wood biochar[23]	2.04				5.85
3	pine bark biochar[23]	1.88				12.15
4	oak bark biochar[23]	25.4				7.40
5	carbon F-400[23]	984				0.204
6	magnetite–maghemite mixture[119]	12	6.0	25	1.0–7.0a	2.9
7	magnetic wheat straw[120]		7.0–9.0		2–30a	3.89
8	sulfate-modified iron oxide-coated sand[121]	2.9–7.9	7.2	27	0.5–3.5a	0.14
9	iron oxide-coated sand 2[122]	10.6	7.6		100a	0.041
10	iron hydroxide-coated alumina[123]	95.7	6.0–8.0		0.75–3.0a	7.64
11	iron oxide-coated sand[124]		7.5	27	0.1–0.8	0.029
12	uncoated sand[124]		7.5	27	0.1–0.8	0.006
13	iron oxide-coated cement[125]		6.7	27	5.0–40.0a	0.69
14	iron chitosan granules[126]		7.0		0.05	2.32
15	iron oxide-coated cement 2[127]		3.2–12.0	15	0.7–13.5a	0.73
16	ferrihydrite[40]		4.2			0.58
17	agricultural residue “rice polish”[128]	452.0	7.0	20	0.1–10.0a	0.14
18	iron-coated sea weeds[129]		7.0	20	0.01–0.05	4.2
19	iron-impregnated charred GAP[130]	52.2	5.0–9.0	25	0.05–2.0	3.25
20	unmodified alumina[131]	189.12	6.5–7.2		0.84–2.02a	0.92–2.16
21	iron-modified zeolite[132]		6.0–9.0		0.5	0.1
22	rice husk iron oxide composite biochar (RHIOB), this study	300.0	7.5	45	0.05–0.2b	0.096a
23	wheat husk iron oxide composite biochar (WHIOB), this study	339.0	7.5	45	0.05–0.2b	0.113a

Studies were carried out at high arsenic concentrations.

Studies were carried out at low arsenic concentrations usually present in water.

Because almost all As(III) is adsorbed on the iron oxide particle surfaces, the adsorption capacity based on the weight of iron oxides deposited on the biochar was calculated for samples 22 and 23 as 0.83 and 0.82 mg As(III)/g of Fe3O4.

Multicomponent Sorption Studies

Single and multicomponent As(III) sorption on RHIOB and WHIOB was examined in the presence and absence of chloride, nitrate, sulfate, bicarbonate, phosphate, sodium, potassium, and calcium. All the sorption studies were carried out at an adsorbent dose of 2.0 g/L for RHIOB and 1.0 g/L for WHIOB at 25 °C and an initial pH of 7.5 (Figure ). Chloride, nitrate, sodium, potassium, and calcium when present individually (at 100 and 200 mg/L) do not affect the As(III) adsorption, while sulfate, bicarbonate, and phosphate significantly inhibited As(III) adsorption (Figure ). Furthermore, the As(III) adsorption decreased more when sulfate, bicarbonate, and phosphate concentrations increased from 100 to 200 mg/L. Negligible chloride and nitrate influence on As(III) adsorption may be due to their tendency to form outer-sphere complexes with iron (oxy)hydroxide nanoparticles present in RHIOB and WHIOB.[50,108,109] Similarly, sulfate and bicarbonates have been reported to form more stable complexes with iron (oxy)hydroxides, thereby competing with arsenic oxyanions for adsorption.[110] Phosphates showed drastic inhibition in As(III) adsorption (Figure ) on RHIOB and WHIOB. This is due to the fact that phosphate belongs to the same family as that of arsenic and therefore showed close structural similarities.[109] It forms stronger inner-sphere complexes with iron (oxy)hydroxides, offering strong competition, electrostatic repulsion, and hindrance for arsenic oxyanion adsorption on RHIOB and WHIOB.[109]

Figure 11

As(III) adsorption on (A) RHIOB and (B) WHIOB in the absence and presence of individual interfering ions at 100 and 200 mg/L [adsorbent dose: 2.0 g/L for RHIOB and 1.0 g/L for WHIOB; temp.: 25 °C, initial pH: 7.5, agitation speed: 100 rpm, contact time: 24 h].

As(III) (100 μg/L) adsorption in the presence of combined interfering ions [Cl–/NO3–/SO42–/HCO3–/PO43–/Na+/K+/Ca2+: 100:100:100:100:100:100:100:100 mg/L and Cl–/NO3–/SO42–/HCO3–/PO43–/Na+/K+/Ca2+: 200:200:200:200:200:200:200:200 mg/L) was significantly decreased in RHIOB and WHIOB (Figure ). The decrease in As(III) adsorption increased with a rise in individual ion concentration from 100 to 200 mg/L in the multicomponent solution. Furthermore, As(III) adsorption in the multicomponent system was also studied by taking all the above-mentioned interfering ions except phosphate. Phosphate was excluded due to the fact (Figure ) that it significantly decreased As(III) adsorption when As(III) and phosphate were taken together in a binary system [(As(III): 100 μg/L + PO43–: 100 mg/L] and [As(III): 100 μg/L + PO43–: 200 mg/L]. It is clear from Figure that As(III) adsorption on RHIOB and WHIOB does not significantly decrease when all the interfering ions except phosphate are present in solution. The slight decrease in As(III) adsorption in a multicomponent system (except phosphate) is due to the interference caused by bicarbonate and sulfate ions. Phosphates do not occur in natural water.[111] It can only be present when water is contaminated with some anthropogenic source (such as fertilizers). Thus, RHIOB and WHIOB can be successfully used for As(III) removal in single and multicomponent systems with no significant decrease in the presence of many interfering ions.

Figure 12

As(III) adsorption on (A) RHIOB and (B) WHIOB in the absence and presence of combined interfering ions at Cl–/NO3–/SO42–/HCO3–/PO43–/Na+/K+/Ca2+: 100:100:100:100:100:100:100:100 mg/L and Cl–/NO3–/SO42–/HCO3–/Na+/K+/Ca2+: 100:100:100:100:100:100:100:100 mg/L [adsorbent dose: 2.0 g/L for RHIOB and 1.0 g/L for WHIOB; temp. = 25 °C; initial pH = 7.5; agitation speed = 100 rpm; contact time = 24 h].

Desorption and Regeneration of Adsorbents

Simultaneous As(III) desorption and regeneration of the exhausted adsorbent was performed using 50 mL of a wash solution prepared by mixing 0.1 N NaOH and 0.1 N NaCl.

First, batch adsorption studies were carried out at a pH of 7.5 and an initial As(III) concentration of 100 μg/L at 25 °C. The suspension was filtered, and the exhausted adsorbent was collected. The adsorbent was washed using double distilled water and oven-dried. The dried adsorbent was dispersed in 50 mL of the wash solution (0.1 N NaOH + 0.1 N NaCl) and kept for 24 h. The suspension was filtered. The filtrate was analyzed for As(III) concentration desorbed from the adsorbent. This process was repeated for three consecutive cycles to find the decrease in recycled adsorbent efficiency. The percent As(III) adsorption and desorption during four consecutive cycles is shown in Figure . As(III) adsorption on RHIOB decreased from 90% (in the first cycle) to 86% (in the fourth cycle). Similarly, As(III) removal on WHIOB was reduced from 92% (in the first cycle) to 90% (in the fourth cycle) (Figure ). This nominal decrease in As(III) removal clearly shows the reusability of RHIOB and WHIOB for many more cycles.

Figure 13

Percent As(III) adsorption/desorption in four consecutive cycles for RHIOB and WHIOB [adsorption studies were conducted at adsorbent dose = 2.0 g/L; initial As(III) concentration = 100 μg/L; initial pH = 7.5; agitation speed = 100 rpm; temperature = 25 °C; and contact time = 24 h; desorption studies were carried out using 50 mL of wash solution prepared by mixing 0.1 N NaOH + 0.1 N NaCl].

Conclusions

Iron oxide modification of widely available rice husk and wheat husk biochars was achieved by pyrolysis of the FeCl3-preimpregnated husks at 600 °C (1 h). These hybrid adsorbents successfully remediated As(III) from low concentration mostly solutions on the exposed Fe3O4 and Fe2O3 particle surfaces. Little As(III) adsorption occurred on nonimpregnated RHB or WHB. The biochars act primarily as iron oxide particle dispensing agents for arsenic sorption, but the biochar phase can adsorb other organic and heavy metal pollutants. Fe3O4 and Fe2O3 particles were the major As(III) adsorbers. Reactions at the Fe–OH surface sites formed Fe–O–As(III) linkages, chemisorbing As(III). This is analogous to the binding of arsenate to iron oxide surfaces. No obvious oxidation of As(III) to As(V) was observed. RHIOB and WHIOB exhibited Langmuir maximum capacities of Q0 = 96.14 and 110.73 μg/g at 45 °C, respectively, in the pH range of 6.5–7.5 with excellent adsorption over the pH range of 3.0–10. Uptake rates were fast and followed pseudo-second-order kinetics. As(III) adsorption on RHIOB and WHIOB was endothermic. At lower concentrations (50 μg/L), particle diffusion is the rate-limiting step, whereas at higher concentrations (100 and 200 μg/L), film diffusion controls the rate. These low-cost hybrid sorbents have very fast sorption dynamics based on biochar’s large surface area, which caused the nanosized iron oxide particles to be well dispersed, exposing a large iron oxide surface area to the aqueous As(III) solutions. Thus, from low As(III) concentrations, adsorption capacities were achieved comparable to other adsorbents used at higher concentrations for As(III) remediation.

This research demonstrated the synthesis of low-cost dispersed iron oxide nanocomposites. It avoided some negative factors of employing pure nanoparticles and avoided the low arsenic sorption efficiency of pure biochar. Instead, both phases were combined to generate hybrid nanocomposites. Development of biochar-based nano iron oxide dispersed composites in the present study permitted reasonable liquid flow through columns and faster filtration together using a low-cost widely available agricultural waste.

Materials, Instrumental Methods, and Sorption Study Methods

All the chemicals used were of either analytical reagent (AR) or guaranteed reagent (GR) grade. FeCl3·6H2O (>97.0%) was obtained from CDH. An As(III) stock solution (1000 mg/L) was prepared using NaAsO2 (>90.0%, Sigma-Aldrich) in double distilled water. KI (>99.0%), NaBH4 (>95.0%), and ascorbic acid C6H8O6 (>99.0%) were obtained from Merck, India. Solution pHs were maintained using HNO3 (0.1 N) and NaOH (0.1 N) and measured on a multiparameter ion meter (model Orion 5 star, Thermo Scientific). Adsorbent–adsorbate sample agitation was conducted with water bath shakers (model MSW-275, Macro Scientific India and RC51000, Scientech India) at 10, 25, and 40 °C. As(III) concentrations were analyzed using an atomic absorption spectrometer (AAS; model Aanalyst 400, PerkinElmer) equipped with a mercury hydride system and an arsenic electrodeless discharge lamp at 193.7 nm. Iron oxide–biochar composites were formed in a controlled atmosphere muffle furnace (model Thermolyne, Thermo Scientific). All sample filtrations were performed using a Whatman No. 1 filter paper. Magnetic collection of the adsorbent was also used in adsorption studies.

The iron oxidation states in iron oxide–biochar composites were determined using XPS (model Versa Probe II, PSI, FEI Inc.). The mineralogy, phase analyses, and crystallinity were identified on a powder X-ray diffractometer (model X’Pert PRO, PANalytical). Functional groups in the surface regions were identified using FTIR (model 7000, Varian). Surface morphology was studied by SEM (model EVO 40, Zeiss). XPS analyses were carried out using a monochromatic Al-X-ray source (hν = 187.85 eV; power = 24.6 W; and beam spot size = 100 μm). Elemental distribution in the surface/near-surface region in the biochar composites was determined by SEM–EDX (model Epsilon 5, PANalytical). Qualitative analysis was done using ICP-MS (ELAN-DRC-e). Crystallinity and diffraction patterns were obtained using a TEM (model JEM 2100F, JEOL) microscope. Specific surface areas and pore volumes were analyzed by the Brunauer–Emmett–Teller (BET) N2 analysis method using a Quantachrome Autosorb 1 Automated Gas Sorption System. The samples were degassed at 150 °C for 5.0 h. CHNS combustion analysis was performed using a vario MICRO CHNS instrument. Magnetic moments were calculated using a model T-415 Cryogenic USA system at 5 and 300 K from −5 to +5 T magnetic field strengths.

Iron Oxide–Biochar Composite Synthesis

The iron oxide–biochar composites were prepared by a proprietary copyrolysis (Indian Patent Application No. 201811010032)[112] of rice husks and wheat husks. Rice husks (∼70 g) and wheat husks (∼130 g) were soaked separately in 2.8 mol/L solutions of ferric chloride hexahydrate (FeCl3·6H2O) for 24 h. These samples were dried for 2 h at 80 °C and then slowly pyrolyzed in a muffle furnace by heating to 600 °C (10 °C/min) and holding for 1 h under N2. The hybrid iron oxide–biochar composites were washed with distilled water and then ethanol to remove impurities and soluble metal salts. The composites were oven-dried and stored in airtight containers until further use. The iron oxide-rice husk and iron oxide-wheat husk biochar hybrid composites were designated as RHIOB and WHIOB, respectively.

Equilibrium and Kinetic Studies

All sorption experiments were conducted in the batch mode. Adsorption experiments at pH values from 2.0 to 10.0 were performed with an initial As(III) solution at 25 °C for 24 h (adsorbent dose = 2.0 g/L; initial As(III) concentration = 100 μg/L; agitation speed 100 rpm). As(III) adsorption studies were conducted at 0.5, 1.0, and 2.0 g/L RHIOB and WHIOB dosages; 50, 100, and 200 μg/L As(III) concentrations; and contact times 0.5, 1.0, 2.0, 4.0, 6.0, 8.0, 10.0, 24.0, and 48.0 h at 25 °C. Sorption kinetics data were fitted to both pseudo-first-order (S5) and pseudo-second-order (S6) rate equations.[113] Sorption equilibrium studies were conducted at 10, 25, and 40 °C for 24 h. As(III) concentrations ranged from 50 to 1000 μg/L, and adsorbent dosages were 2.0 and 1.0 g/L for RHIOB and WHIOB, respectively. All As(III) equilibrium and kinetic studies were carried out in the optimum pH range of 7.0–7.5 for RHIOB and 6.0–6.5 for WHIOB. Other sorption study methods employed here are summarized in the Supporting Information.

Rate-determining step identification during adsorption is crucial for interpreting sorption/ion-exchange kinetics,[114] aiding material selection and conducting system design.[115] Rate depends on external, internal, or both types of diffusion. Adsorption generally occurs in three steps including film diffusion, particle diffusion, and exterior surface diffusion.[114] Adsorption on the exterior surface is rapid and cannot be considered as a rate-determining step.[116] Therefore, either film or particle diffusion is rate-controlling. Kinetic data were analyzed according to Boyd[117] and Reichenberg[97] to differentiate between particle and film diffusion using eqs –4orF is the fractional attainment of equilibrium at time “t” given by the following equation.B, the Boyd number, is given by eq Here, Di is the effective diffusion coefficient within the adsorbent phase, r0 is the radius of the adsorbent particle, and n is an integer defining an infinite series solution.

Isotherm studies were conducted using various nonlinear models mentioned in the Supporting Information. Fitting of the data was predicted based on the regression coefficient, R2, and Pearson chi-square, χ2, values. Pearson chi-square χ2 is the measurement of the degree of error between experimental and model predicted values.[118] It is given by eq where q and qm (μg g–1) are total adsorption at time t using experimental data and predicted model data, respectively. Bt values were obtained for each F value from Reichenberg’s table[97] at different concentrations.

Interfering ions including chloride, nitrate, sulfate, bicarbonate, phosphate, sodium, potassium, and calcium may affect the As(III) adsorption capacity of RHIOB and WHIOB. Thus, single and multicomponent As(III) sorption on RHIOB and WHIOB was examined. Multicomponent sorption experiments were carried out as follows: As(III) stock solution of 100 μg/L was prepared and used in all the multicomponent sorption experiments. Then, the corresponding amount of metallic salt [NaCl, KNO3, Na2SO4, NaHCO3, and Ca3(PO4)] was added to 100 mL of the previously prepared solution of 100 μg/L As(III). Competing ion concentrations of 100 and 200 mg/L were used. Three sets of batch experiments were performed to determine the As(III) (100 μg/L) adsorption in single and multicomponent systems. These include (1) the effect of individual interfering ion (100 mg/L) on As(III) (100 μg/L) adsorption, (2) the effect of individual interfering ion (200 mg/L) on As(III) (100 μg/L) adsorption, (3) the effect of combined interfering ions (Cl–/NO3–/SO42–/HCO3–/PO43–/Na+/K+/Ca2+: 100:100:100:100:100:100:100:100 mg/L) on As(III) (100 μg/L) adsorption, and (4) the effect of total interfering ions (Cl–/NO3–/SO42–/HCO3–/Na+/K+/Ca2+: 100:100:100:100:100:100:100:100 mg/L). All the sorption studies were carried out at an adsorbent dose of 2.0 g/L for RHIOB and 1.0 g/L for WHIOB at 25 °C and an initial pH of 7.5.

Desorption and Regeneration Studies

Simultaneous As(III) desorption and exhausted adsorbent regeneration was performed using 50 mL of a wash solution prepared by mixing 0.1 N NaOH and 0.1 N NaCl.

The adsorbent obtained from sorption studies was washed using double distilled water and oven-dried. The dried adsorbent was dispersed in 50 mL of the wash solution (0.1 N NaOH + 0.1 N NaCl) and kept for 24 h. The suspension was filtered. The filtrate was analyzed for As(III) concentration desorbed from the adsorbent. This process was repeated for four consecutive cycles to find the decrease in the efficiency (if any) of RHIOB and WHIOB.