Introduction

The constant rapid growth of population, urban civilization, and industrialization in the world has been dramatically deteriorating the global climate and natural ecosystem owing to waste generation, which contain various pollutants including heavy metal ions (Moghaddam et al.2016), toxic organic compounds, pharmaceuticals, organic and inorganic dyes, and dissolved inorganic solids (Taghavi et al. 2018a, b).

Among the inorganic pollutants, arsenic and fluoride have attracted great attention towards the secure drinking water predominantly due to two major reasons. One is their chronic deadly effects on human health caused by high intake and the other is their natural abundance in drinkable water sources. Smedley and Kinniburgh (2002) had reported the abundance of high arsenic concentration (100–5000 μg L−1) in surface and groundwater in sulfide mineralization areas. Underground aquifers in south eastern Asia (India, Bangladesh, Philippines, China), South and North South America, and part of the Europe (Romania, Hungary, Serbia, and Croatia) are also contaminated with high arsenic (> 10 μg L−1) (Smedley and Kinniburgh 2002). Less mobile arsenate occurring naturally in soil reduces to more mobile and toxic arsenite by organic matters at optimum geochemical environment, which leads to the contamination of groundwater aquifers (Mohan and Pittman 2007; Wang et al. 2015). In accordance to the US Environmental Protection Agency (Wang et al. 2015), the allowance concentration limit of arsenic(total) for the safe drinking water is ≤ 10 μg L−1 (World Health Organization 2011).

Besides arsenic, presence of the fluoride at high proportions in drinkable water sources was identified in 1937 at Nellore district (Andhra Pradesh) in India (Siddiqui 1955). Fluoride intake beyond the maximum allowance level (1.5 mg L−1) of international regulation (Wang et al. 2015) may effect in progressive incapacitate disorder known as fluorosis (Fallahzadeh et al. 2018; Miri et al. 2018). Many countries including India still depend upon the groundwater due to economic viability and well accessibility. But existence of the high arsenic or fluoride in groundwater makes the water unhygienic for drinking which gains paramount attention for decontamination of arsenic/fluoride from water.

Accounting the attenuation of these contagions several water purification techniques have been studied (Pendergast and Hoek 2011; Crittenden et al. 2012; Gupta et al. 2012; Ali 2012; Li et al. 2016; Saha et al. 2016). Among them, adsorption technique is unique due to ease of operation with high purification efficiency in a low expense (Babel 2003). Multiple materials as adsorbent had been examined for arsenic and fluoride removal from aqueous phases. Fe2O3 (Zhong et al. 2006), Fe3O4 (Yavuz et al. 2006; Paul et al. 2015), FeOOH (Li et al. 2011a; Wang et al. 2012), Carbon nanospheres (Li et al. 2015),TiO2 (Zhong et al. 2008), CeO2 (Cao et al. 2010), CuO (Yu et al. 2012), titanium lanthanum oxide-activated carbon (Jing et al. 2012), Fe2O3-Silica (Yang et al. 2014), AlOOH-Fe3O4 (Yang et al. 2013), ZrO2-carbon nanotubes (Addo Ntim and Mitra 2012), and Fe3O4-graphene (Chandra et al. 2010) were employed for the arsenic removal. The materials such as ZrO2 (Biswas et al. 2007a), rare earth metal oxide (Raichur and Jyoti Basu 2001), manganese coated alumina (Maliyekkal et al. 2006),coal based sorbents (Sivasamy et al. 2001), fly ash (Chaturvedi et al. 1990), waste carbon slurry (Gupta et al. 2007), oxide minerals (Rongshu et al. 1995), bone char (Medellin-Castillo et al. 2007), zeolites (Mayadevi 1996), spent catalyst (Lai and Liu 1996), and activated alumina (Ku and Chiou 2002) were employed for the fluoride removal. Many of them are efficient even at high ionic concentration (Fan 2003), but some of them are expensive such as activated carbon and rare earth (Tian et al. 2011). Among them, ferric oxide was found to be cheaper as well as it has moderate affinity towards both arsenic (Zhong et al. 2006) and fluoride (Chen et al. 2012). Improved adsorption affinity of iron oxide towards arsenic and fluoride were achieved by surface modification incorporating some multivalent metal ions (Ghosh et al. 2006; Gupta et al. 2008, 2010; Chen et al. 2012). Till date very few research reports available for scavenging both arsenic and fluoride (Tian et al. 2011; Li et al. 2011a, b). Anticipated that incorporation of calcium ion inside the iron(III) oxide could be an effective cheap material for scavenging arsenite and fluoride from the aqueous phase. Hence, we report the synthetic calcium ion-incorporated hydrous iron(III) oxide (CIHIO) as a better candidate for removal of arsenite and fluoride from water. Adsorption capacity as well as mechanism has also been reported with modeling of kinetics and equilibrium data.

Materials and methods

Materials used

This section is given in Supplementary Information (SI).

Analytical tools and reagents preparation

This is given in SI section.

Experimental design

This is given in SI section.

Calcium ion-incorporated hydrous iron(III) oxide (CIHIO) preparation

Calcium ion-incorporated hydrous iron oxide (CIHIO) samples were prepared by mixing the salt solution calcium and ferric ions (mole/mole) from 0.2:1 to 1:2. NaOH solution (4 M) was dropped slowly onto the solution stirred mechanically till the supernatant pH rose up above 10.0 keeping reaction temperature 70 °C. The brown gel precipitate was separated from mother liquid after 48 h, which was washed with water till washing liquid drained out is neutral. The brown mass dried inside an oven at 80–100 °C had been treated with cold water, and re-dried at that temperature. Finally, the dried material was ground, sieved to particles size of range 52–100 mesh (149–290 μm) and, employed for arsenite and fluoride adsorption from their separate aqueous solutions. It is found that the material showed the increase of arsenite and also fluoride adsorption amount with increasing the amount of calcium ion (Figure S1). However, the CIHIO sample became hygroscopic when volume proportion of the calcium solution (0.05 M) was higher than that of the ferric solution (0.05 M). Therefore, CIHIO sample prepared from a mixture of 0.05 M ferric solution and 0.05 M calcium solution (1: 1, v/v), which showed good arsenite/fluoride adsorption capacity, was characterized and used systematically for conducting adsorption experiments.

Arsenic and fluoride samples analysis

The methods used for analysis of arsenic and fluoride samples are given inside SI section in details.

Adsorption capacity calculation

Adsorption capacity (q, mg g−1) had been calculated by the mass balanced equation (Eq. 1).

$$ q=\frac{\left({C}_i-{C}_f\right)}{m}\times \frac{V}{1000} $$
(1)

where Ci (mg L−1) is initial concentration of solute in solution. The q = qt for Cf (mg L−1) = Ct (adsorbate concentration at time, t; V = volume (mL) of the solute solution taken, m = mass (g) of CIHIO added. The q = qe for Cf = Ce (adsorbate concentration at equilibrium).

Results and discussion

Characterization

Figure 1 shows SEM images (a and b), TEM image (c), EDAX spectrum for the elements over surface (d), and bar graph showing elemental percentage (e), respectively. Image a demonstrated the shape of particles while the image b showed the presence of facture in particles. The factures should be beneficial allowing penetration of solute solution inside the particles for adsorption. The TEM image c (Fig. 1) shows the presence of tiny particles inside agglomerates of size about 50 nm, which are obviously nanoparticles (NPs) in dimension. From the EDAX spectrum (d) and the elemental composition (plot e) (Fig. 1), it can be concluded that the atomic percentages (in parenthesis) of the elements over CIHIO surface are Fe (35.21), Ca (20.76), Cl (4.4), and O (39.62); indicating the atomic percentage of Ca is about 0.59 times that of Fe. Low chloride percentage found has appeared from the reagents, which cannot be washed out owing to entrapping in lattice structure. Lower percentage of Ca than Fe in CIHIO is due to the higher solubility product value (Ksp) of calcium hydroxide (5.5 × 10–6) than iron(III) hydroxide (4 × 10–38). Figure 2 shows the AFM images of native hydrous ferric oxide (HFO) in 2-dimension including the particle size distribution and surface profile showing a clear smooth surface (image a). Graphical magnification of the dotted line of image b and extracted profile from 2D image c of HFO suggested the presence of particles size ~ 100 nm with highest probability (image d). Figure 3 shows the AFM images of CIHIO in 3D view (a and b) and 2D view including surface morphological pattern (c), indicating the rough and heterogeneous surface of the material. Extracted surface profile of the AFM image c (Fig. 3) shows the particles size of CIHIO is around 50 nm, which is smaller than the native HFO indicating increase possibility of surface reactivity of the material.

Fig. 1
figure 1

The SEM mages at (a) low resolution and (b) high resolution; TEM image (c) (100 nm scale); (d) electron dispersive X-ray (EDX) spectrum with (d) surface elemental composition and (e) bar graph plots showing surface composition (e) of CIHIO material

Full size image
Fig. 2
figure 2

a Atomic force microscopic image, b extracted profile with 2D image, c graphical magnification of the dotted line from the picture b, and d particle size distribution of HFO

Full size image
Fig. 3
figure 3

Atomic force microscopic images of CIHIO in 3D a side view and b front view, c 2D image of CIHIO and extracted profile of a single particle

Full size image

The surface area and pore volume of any material are very important characteristics of surface science, which have been analyzed from N2(vapor) adsorption-desorption profile. Figure 4 shows the N2(vapor) adsorption-desorption profile (plot a) including pore volume versus pore diameter (plot b). It is revealed that the specific surface area of as-prepared CIHIO was estimated to be 269.61 m2 g−1, which is about 1.6 times larger than HFO(165.60 m2 g−1) (Clesceri et al. 1998), and it is consistent with the ~ 50% particle size lowering of CIHIO than HFO. The pore volume and pore diameter (Table 1) were also estimated from the plot b (Fig. 4), suggesting the material is mesoporous (pore diameter, 27.742–27.913 Å; pore width, 26.97 Å).

Fig. 4
figure 4

a Nitrogen (vapor) adsorption-desorption plot for BET surface area of CIHIO; b pore volume versus pore diameter plot; c powder X-ray diffraction pattern of CIHIO; and d FTIR spectral plot of CIHIO

Full size image
Table 1 Some important characteristic parameters of CIHIO
Full size table

Powder X-ray diffraction (XRD) pattern is a good guide for identification of the crystalline mineralogical phases, if present, in a material. Thus, the XRD pattern of CIHIO taken is shown in Fig. 4c. The peaks of moderate intensities at °2Ѳ = 27.13°, 35.76°, 39.79°, 46.95°, and 56.54° indicate the presence of crystalline Fe2O3 phase. The peaks at 34.73° and 36.08° confirm the presence of orthorhombic crystal of FeOOH in this hybrid oxide. The peaks of moderate intensities at °2Ѳ = 34.56°, 47.26°, and 52.31° had indicated the presence of CaO in CIHIO. The size of crystallite has been calculated using Scherrer equation as described by Taghavi et al. (2018a, b) and obtained ~ 10 nm. The FTIR spectrum d of CIHIO is shown in Fig. 4 (percent transmittance versus wave number (ν, cm−1) for in-depth characterization. The O–H bond stretching and bending modes obtained at ν = 3369.84 cm−1 and ∼ 1630 cm−1, respectively (Deschamps et al. 2005). The absorption bands of the spectrum at ν = 1395.94 and 678.91 cm−1, respectively, are vibration and bending modes of Fe–O bonds (Saha et al. 2015; Mukhopadhyay et al. 2017). Additional band at ν = 2374 cm−1 is probably for the unsymmetrical bending of Ca–O bond and the band at ν ~ 808.72 cm−1 is the stretching of Ca–O bond, respectively. The pHzpc of CIHIO had been analyzed using pH metric method as depicted by Babic et al. (1999) and the value is estimated to be 6.5 (Figure S2), which is very near to the neutral pH value.

Effect of pH

The surface reaction at solid-liquid interface is highly dependent upon the solution pH, and sometimes pH plays a key role for the adsorption phenomenon. Influence of the pHi (i = initial) on arsenite/fluoride adsorption by CIHIO was investigated in a pHi range 3.0 to 9.0. Figure 5 demonstrates the pH dependent adsorption amount of fluoride (plot a) and arsenite (plot b) at 303 K. It can be highlighted that amount of the fluoride adsorption was nearly equal in pHi range 3.0–7.0 showing a peak value at pHi between 4.0 and 6.0, but amount of the arsenite adsorption was well at pHi range 5.0–7.0 showing a peak value at pHi 5.0. Figure 5 also reflected that more than 70% (pHi = 3.0–7.0) of initial fluoride loading has transferred from liquid to solid surface, while more than 85% (pHi = 5.0–7.0) of initial arsenite loading has transferred from solution to solid surfaces. The removal amount of fluoride or arsenite by CIHIO, however, declined substantially at pHi > 7.0 due to change of the surface charge of the material (pHzpc = 6.5). This result is very close to the results reported previously (Mukhopadhyay et al. 2017). The mechanism apprehended of this adsorption reaction shall be described later on in this manuscript taking help of the experimental support.

Fig. 5
figure 5

Effects of the pHi towards adsorption of a fluoride and b arsenite by CIHIO at 303 K

Full size image

Adsorption kinetics

Kinetics of the heterogeneous adsorption process helps to predict a minimum contact time required to transfer the solute from solution to solid surface for the continuous flow system. Here, kinetics of the present adsorption reactions were monitored for Ci of solutes 5.0, 10.0, and 20.0 mg L−1at pHi 7.0 and 303 K; and also for the Ci = 10 mg L−1 at three different selected temperatures (288, 303, and 318 K) and pHi 7.0. Figure 6 shows variation of the experimentally acquired kinetic data for arsenite (plot a) and fluoride (plot b) at varying solute concentrations; and also for arsenite (plot c) and fluoride (plot d) at a fixed Ci (10.0 mg L−1) and the temperature 288, 303, and 318 K, respectively. It shows that about 80% of the initially loaded solutes have been transferred over CIHIO surface by 30 min from the starting of reaction (t = 0). Relatively a longer time was taken to transfer the residual 20% solute from solution to solid phase. It might be the consequence of steric and electrostatic repulsion between solute hosted and solute species in solution to be hosted over solid surfaces. The kinetic data displayed (Fig. 6) have been analyzed by the commonly used pseudo-first order (PFO) (Eq. 2) and pseudo-second order (PSO) models (Eq. 3) (Saha et al. 2015).

$$ {q}_{\mathrm{t}}={q}_{\mathrm{e}}\left(1-{e}^{- kt}\right) $$
(2)
$$ {q}_{\mathrm{t}}=\frac{{\left({q}_{\mathrm{e}}\right)}_2{k}_2t}{\left(1+{q}_{\mathrm{e}}{k}_2t\right)} $$
(3)
Fig. 6
figure 6

The plots of qt (mg g−1) versus time of reaction, t (min) for variation of concentration of a arsenite and b fluoride; and c the effect of temperature on reaction kinetics for arsenite and d fluoride adsorption on CIHIO. Solid line indicates pseudo-second order fitting and dotted line indicate pseudo-first order model fits of data

Full size image

where k (min−1) and k2 (g mg−1. min−1) are rate constants for PFO and PSO, respectively. The significance of qe and qt are given elsewhere. The kinetic parameters estimated for arsenite and fluoride with non-linear model fits (Fig. 6) of the data are listed in Table 2 (different temperatures) and Table 3 (different concentrations), respectively. The weaker fit of kinetic data with the PFO equation (Eq. 2) than the PSO equation (Eq. 3) as the qe(model) values obtained from the PFO equation are away from the qe(exp)values; which has also been evident from lesser R2 and larger χ2 values. That is, adsorption kinetics is more close to the PSO equation. Increase of the rate constant values (Table 2) with rising temperature indicates the increase of fastness of the adsorption reaction, because the solute species available has been activated to override the threshold energy for effective collision (Clesceri et al. 1998; Saha et al. 2015). Additionally, values of the qe for arsenite increased gradually indicating the reaction is endothermic, but those for fluoride decreased favoring the reaction is exothermic (Table 2). Quite similar trends for arsenic and fluoride adsorption had been reported very recently (Gupta et al. 2008, 2010). The kinetic parameters of concentration dependent of solutes (Table 3) show that the values of rate constant for arsenite and also for fluoride increased with the increase of solute concentrations. This increase of rate constant, i.e., rate of adsorption reaction with increase of the Ci, can be justified from the increase of availability of solute per unit surface sites of material (Gupta et al. 2008; Saha et al. 2015; Mukhopadhyay et al. 2017).

Table 2 The parameters estimated from the model fits of kinetic data of arsenite and fluoride adsorption reactions with CIHIO at pH 7.0 (± 0.2) and temperature 288, 303, and 318 K
Full size table
Table 3 The kinetic parameters estimated from the model fits of kinetic data of arsenite and fluoride adsorption reactions with CIHIO at pH 7.0 (± 0.2) and temperature 288 K
Full size table

Equilibrium modeling

Equilibrium study is helpful to predict the thermodynamic aspects of any reactions and adsorption mechanism. Figure 7 (qe versus Ce) shows the equilibrium data (as points) of arsenite and fluoride adsorptions by CIHIO at temperatures 288, 303, and 318 K at pH 7.0, which have shown the rapid enhancement of qe with increase of temperatures indicating increase of endothermic nature of the adsorption processes. These data have been modeled with most widely used isotherm equations namely Langmuir (Eq. 4) and Freundlich (Eq. 5) (Li et al. 2011a, b; Saha et al. 2015; Mukhopadhyay et al. 2017).

$$ {q}_{\mathrm{e}}=\frac{q_{\mathrm{m}}\times b\times {C}_{\mathrm{e}}}{1+b\times {C}_{\mathrm{e}}} $$
(4)
$$ {q}_{\mathrm{e}}={K}_{\mathrm{F}}\times {\left({C}_{\mathrm{e}}\right)}^{\frac{1}{n}} $$
(5)
Fig. 7
figure 7

The equilibrium data fit with the a Langmuir isotherm, b Freundlich isotherm for the arsenite adsorption; the c Langmuir isotherm and the d Freundlich isotherm for the fluoride adsorption by CIHIO at different temperatures

Full size image

Where qe and Ce have their usual implications and mentioned earlier, qm the maximum monolayer adsorption capacity (mg g−1), b the Langmuir isotherm constant (L g−1), KF the Freundlich isotherm constant (mg g−1) related to the adsorption capacity, and n is the adsorption intensity.

Figure 7 (a–d) shows the non-linear fitting of the equilibrium data with Langmuir and Freundlich isotherm models, respectively, in a and b for the arsenite and c and d for the fluoride. The isotherm parameters of related equations (4 and 5) which are estimated from the model fitting of data are given in Table 4. The parametric values inform that the fits of isotherm data were better with the Langmuir equation (R2 = 0.97–0.99) than the Freundlich equation (R2 = 0.90–0.97), implying adsorption sites of CIHIO are homogeneous and the tested solutes adsorption occur with monolayer surface coverage (Clesceri et al. 1998; Sivasamy et al. 2001; Addo Ntim and Mitra 2012). Goodness of the equilibrium data fitting with Langmuir isotherm has also been confirmed from the hybrid error function (HYBRID) and the Marquardt's percent standard deviation (MPSD) calculations (Table 4) using the relations as described by Sanchooli Moghaddam et al. (2016). The capacities of monolayer adsorption of Langmuir isotherm (qm, mg g−1) values are 25.57 and 29.07 at 303 K for fluoride and arsenite, respectively. With increase of the temperature, values of the qm of CIHIO are increased for both arsenite and fluoride, indicating endothermic nature of the reactions (Gupta et al. 2007, 2010; Ghosh et al. 2014). Thus, the accessible adsorption sites by solutes over CIHIO have increased with increasing temperature, and the adsorption capacities enhanced. The dimensionless parameter, KL which is also called separation factor highlights an obligatory feature of the Langmuir isotherm and can be calculated by the relation (Eq. 6).

$$ {K}_{\mathrm{L}}=\frac{1}{1+b\times {C}_i} $$
(6)
Table 4 Non-linear isotherm model fit parameters for arsenite and fluoride adsorption by CIHIO at pH 7.0 (± 0.2) at temperatures 288, 303, and 318 K (Ci range, 5.0–80.0 mg L–1)
Full size table

where b is the Langmuir constant (L mg−1) and the significance of Ci is given elsewhere. If the value of KL is (a) ranged between 0.0 and 1.0, the reaction is favorable; (b) greater than unity, the reaction is unfavorable, and (c) KL = 1, the reaction is irreversible (Hall et al. 1966). Here, values of the KL ranged between 0 and 1.0 even at Ci = 0.01 mg L−1 of the tested ions, indicating the highly favorable adsorption processes.

In order to compare the removal efficiency, values of the qm for arsenite and fluoride adsorption by CIHIO have been compared with other iron based materials in Table 5 (Manna et al. 2003; Dey et al. 2004; Ghosh et al. 2006; Mohan and Pittman 2007; Biswas et al. 2007a, b; Gupta et al. 2008; Kumar et al. 2009; Biswas et al. 2009, 2010; Basu et al. 2010; Basu and Ghosh 2011). This comparison of qm has a little significance unless the experimental parameters such as concentration range of solute, adsorbent dose, and agitation speed are identical. However, values of the qm of CIHIO are quite high than some of the materials reported such as iron(III) oxide and iron(III) based mixed oxides (Table 5).

Table 5 A comparative list of monolayer capacity of CIHIO for adsorption of arsenite and fluoride with some similar type of iron-based adsorbent material at a temperature 303 K
Full size table

Thermodynamics parameters

Evaluation of the thermodynamic parameters is particularly important to estimate the spontaneity of adsorption reactions. Thus, the parameters such as the changes of Gibb’s free energy (∆G0), entropy (∆S0), and enthalpy (∆H0) under the standard states can be estimated by the relations (Eqs. 7–9) (Viswanathan et al. 2009; Saha et al. 2014; Ghosh et al. 2014; Li et al. 2017; Mukhopadhyay et al. 2017) below:

$$ \varDelta {G}^0=- RTln{K}_{\mathrm{c}} $$
(7)
$$ \varDelta {G}^0=\varDelta {H}^0- T\varDelta {S}^0 $$
(8)
$$ \mathit{\ln}{K}_{\mathrm{c}}=\frac{\varDelta {S}^0}{R}-\frac{\varDelta {H}^0}{RT} $$
(9)

where R is an ideal gas constant (8.314 J mol−1 K−1), Kc, the equilibrium constant at a temperature T, can be calculated by \( {K}_c=\frac{q{}_e}{C_e} \) where \( \frac{q{}_e}{C_e} \) is called the adsorption affinity which was calculated as calculated by Frantz et al. (2017), Milonlic (2007), and Cadaval Jr et al. (2015). Assuming ∆S0and ∆H0are to be constant at the range of working temperature, the values calculated from slope and intercept of the plot of ln (qe/Ce) against 1/T (Figure S3) were used to calculate the ∆G0 at temperatures studied using the Eq. 7. The values obtained for ∆G0, ∆S0, and ∆H0 are recorded in Table 6. The positive value of the ∆H0 has thus supported that the arsenite and fluoride adsorption by CIHIO inclined to endothermic. Again, the positive values of ∆S0 for arsenite and also fluoride adsorption by the present material confirm the increase of chaotic nature at solid-liquid interface owing to the release of aqua molecules when hydrated solute was hosted by the solid surface. Influence of the entropy change is so high that the influence of enthalpy change has been overturned to have the value of ∆G0 negative and the adsorption reactions become spontaneous. Increase of the temperature has increased the magnitude of ˗∆G0, indicating the increase of spontaneity of the adsorption processes (Gupta et al. 2007, 2010).

Table 6 Estimated thermodynamic parameter for arsenite and fluoride adsorption by CIHIO
Full size table

Effect of some groundwater occurring ions

Field trial is impractical for decontamination of the target ions from water without acquiring knowledge of the adverse influence of major ions occurring in the ground water. Thus, the arsenite and fluoride adsorption reactions with CIHIO were investigated separately in presence of PO43−, HCO3−, SiO32−, Cl−, and SO42− . The results are shown in plot a (arsenite) and plot b (fluoride) in Fig. 8. It shows sharp decline of removal efficiency of both arsenite and fluoride in the presence of PO43− ion (Viswanathan et al. 2009; Li et al. 2017). Like the PO43−, HCO3−, and SO42− ions also reduced the removal efficiency of this material but did this job somewhat less efficiently than PO43− (Mukhopadhyay et al. 2017). The chloride ion has no significant adverse influence on the adsorption reactions of arsenite and fluoride with CIHIO. The arsenite removal efficiency of CIHIO is thus reduced as PO43− > HCO3− > SiO32− > SO42− > Cl− when each was at a level of 10 mM per L. The order mentioned above in removing the arsenite has been modified somewhat as PO43− > SO42− > HCO3− > SiO32− > Cl− when the ions level was 1 mM per L. The results showed that the phosphate has parallel competition for the solid surface with arsenite, and drastically cuts the adsorption sites of solid surface for arsenite or fluoride adsorption (Kumar et al. 2009; Basu and Ghosh 2011; Ghosh et al. 2014). Like the arsenite, the fluoride adsorption by CIHIO was also notably cut in the sequence as PO43− > SO42− > HCO3− > SiO32− > Cl− when the foreign ion concentration was at a level of 1.0 mM or 10 mM per L with fluoride solution. The adverse effect of phosphate on the fluoride removal had also been reported for adsorbents such as granular ferric oxide (Kumar et al. 2009), chitosan bead (Biswas et al. 2010), and Al-Ce mixed oxide (Liu et al. 2010). It is fortunate that the phosphate concentration in groundwater is much lower than 1.0 mM per L (Ghosh et al. 2014). The sulfate ion did the same job but somewhat less strongly than the phosphate ion.

Fig. 8
figure 8

The effects of co-occurring ions on a arsenite and b fluoride adsorptions by CIHIO

Full size image

Regeneration of spent adsorbent

Recycling of an adsorbent material is dependent upon regain its activity via regeneration with a suitable reagent of the spent material. Hence, the adsorption activity regain by spent CIHIO was investigated through desorption of the adsorbed species from the adsorbent surface. Figure 9 shows the results as percent desorption of arsenic (plot a) and fluoride (plot b) against molar concentration of the NaOH solution. It shows the increase of percent desorption with increasing molar strength of NaOH solution. The desorption percentages increased noticeably to 80–85% with increasing molar strength of NaOH to 2.0 M. This high desorption percentage (80–85%) is a strong indication signal in favor of weak interaction between adsorbate and adsorbent (Gupta et al. 2008). As the adsorption reactions of arsenite and fluoride with CIHIO were endothermic indicating the low activation energy for desorption and possibility of high desorption percentage. It is found experimentally that the CIHIO has regained its adsorption efficiency by ~ 60–65% for both arsenic and fluoride adsorption even after the fifth stage of desorption of the spent material, indicating possibility of commercial viability of CIHIO.

Fig. 9
figure 9

The desorption percentages of a arsenite and b fluoride as bar graphs against the molarity of NaOH solution

Full size image

Adsorption mechanism

Figure 10 shows the FTIR spectra of pristine CIHIO, fluoride adsorbed and arsenite adsorbed CIHIO separately. The fluoride adsorption by the material takes place with increase of equilibrium solution pH, supporting the replacement of OH- from the solid surface with F-adsorption (Gupta et al. 2008; Patel et al. 2009; Liu et al. 2010; Ghosh et al. 2014), which has been confirmed from disappearance of the M–O stretching band at a wave number (υ) 1400 cm−1. Instead, a weak band appeared at υ = 1406 cm−1 for M–F stretching, which is similar to the report that was made by Liu et al. (2010). The weak band at 1358 cm−1 in fluoride adsorbed material is possibly for existence of the trace M–OH group still retained with a shift of nearly 40 cm−1 which is possibly due to hydrogen bonding with fluoride (Fornaro et al. 2015). The remaining bands between 1400 and 1628 cm−1 are retained with very small shift in the fluoride adsorbed material. However, the retention of more intense FTIR stretching band of hydroxyl group at 3400–3500 cm−1 of the fluoride adsorbed material (Li et al. 2010) instead of disappearing is possibly due to hydration of surface metal ions via co-ordination from the aqueous solution.

Fig. 10
figure 10

The FTIR spectra (% transmission versus wave number) of fluoride-adsorbed and arsenite-adsorbed CIHIO, and b selected portion of fluoride-adsorbed and c arsenite-adsorbed spectra in a comparing with that of a CIHIO spectra

Full size image

However, the adsorption of arsenite by CIHIO takes place somewhat differently from that of fluoride. The stretching band in FTIR spectra of pristine material at a υ value 3378 cm−1 shifted to 3386 cm−1, which moves slightly to the higher υ value with intensity increase of the band. This might be for the enhanced physically adsorbed water (Li et al. 2010). The bending O–H band at a υ = 1622 cm−1 became more intense in arsenite adsorbed material compared to the pristine material. However, the M–O stretching band of pristine material at υ = 1400 cm−1 split up to three bands (1337, 1461, and 1507 cm−1) with decrease of band intensity in arsenite adsorbed material which is different from the report made by Li et al. (Rongshu et al. 1995; Li et al. 2010). This is presumably due to the chelation of As(OH)3 on the metal oxide surface and, thus the characteristic As–OH band appears around 800 cm−1 in the arsenite adsorbed IR spectra. This is not prominently mentionable due to noisy overlapping of bands in the lower side of wave number region (Gupta et al. 2010; Fornaro et al. 2015). Based on the above analyses of equilibrium solution pH and FTIR spectroscopy of spent materials, the mechanism of arsenite/fluoride adsorption has been depicted in Scheme 1 below:

Scheme 1
scheme 1

Proposed surface reaction mechanism for arsenite/fluoride uptake by CIHIO

Full size image

Conclusion

Analytically calcium-incorporated hydrous iron(III) oxide (CIHIO) has been established as a material of microcrystalline and mesoporous nature with pHzpc ~ 6.5 and irregular surface morphology. The surface area is notably larger than hydrous ferric oxide only. This material when used for the adsorption of arsenite/fluoride with change of contact time from their aqueous solution at optimized pH shows the reaction kinetics of pseudo-second order type. The adsorption equilibrium agrees well to the Langmuir monolayer adsorption mechanism with encouragingly high monolayer capacity for both arsenite and fluoride. Highly spontaneous adsorption processes are the consequence of favorable entropy changes. High desorption (80–85%) percentages with 2.0 M of aqueous NaOH from the spent CIHIO surfaces has indicated reuse possibility of CIHIO. Repeating the use of this material shows that high (60–65%) adsorption efficiency is retained even after fourth stage of recycle. The phosphate ion even at a level of 1 mM per L is capable to reduce the adsorption efficiency of CIHIO. The adsorption mechanism for the arsenite was established to be surface complex reaction while that for the fluoride was ion-exchange type.