Chemical activation and magnetization of onion waste derived carbon for arsenic removal

2.1 Pretreatment of onion and its carbonization

Waste onion bulbs collected from the local vegetable market was peeled off, cut into fine pieces and washed well with deionized water (DI) to remove any dust particle. It was then dried in oven overnight at 110 °C. Dried onion was then grind into fine powder and saved in polymeric vials. The dried biomass was then carbonized in a tube furnace under continuous nitrogen flow rate of 0.2 L/min for 1 h at 450 °C. After that carbonized sample was meshed into powder and then stirred with 10% HCl solution followed by washing with DI water to maintain its pH at 7. Finally, the sample was dried in oven overnight at 110 °C.

2.2 Activation of carbon by using different activators

Carbon obtained as discussed in Section 2.1 was divided into 3 equal parts and activated with different activators like HNO₃, H₃PO₄ and NaOH. For activation, carbon was soaked with 3 M solution of each activator overnight and then heated in a tube furnace under continuous nitrogen flow at the rate of 0.2 L/min for I h at 600 °C. After that the sample was washed with DI water to maintain its pH at 7, followed by drying in oven overnight at 110 °C. The activated samples were labelled as ACH, ACH3 and ACNa based on the activators used.

2.3 Magnetization of activated carbon

Magnetization of samples was carried out by impregnation of each sample with Fe₃O₄ using solvothermal process. In this method, 0.149 g of FeCl₂·4H₂O and 0.405 g of FeCl₃·6H₂O salts were stirred for 15 min in 120 mL DI water. After that 10 mL HCl was added followed by addition of 1 g of activated carbon sample just before the addition of precipitating agent (NH₄OH). These contents were then transferred to Teflon coated steeled line autoclave and heated at 180 °C for 2 h. Same procedure was adopted for other activated carbon samples. Finally, the samples were washed with DI water till their pH approaches 7 and the magnetized samples were labeled as MCH, MCH3 and MCNa.

2.4 Determination of As(III) in aqueous system by colorimetric method

A stock solution (100 mg/L) of As(III) was prepared by dissolving 0.132 g of As₂O₃ in 20 mL of 1 M sodium hydroxide. To neutralize the pH, 50 mL of 0.5 M hydrochloric acid was added. Afterward, the volume of solution was made up to 1000 mL by using HPLC grade water. 50 mL of different concentrations (100–10,000 μg/L) were prepared by dilution of stock solution. For each experiment, freshly prepared solution was kept in brown bottles in the dark to avoid any chance of oxidation. Analysis of As(III) was carried out by colorimetric method. For this purpose, a complexing-coloured reagent was prepared by mixing different solutions that include 4 mL ascorbic acid (10.8%), 4 mL antimony potassium tartrate (0.56%), 6 mL ammonium molybdate (3%) and 10 mL of sulphuric acid (14%). The green colour complexing reagent was found to be stable for 3 h below 30 °C (Dhar et al., 2004).

For preferably formation of antimony arseno-molybdate complex, As(III) was converted to As (V) by using oxidizing agent. Herein, 2 mmol/L KIO₃ solution was prepared by dissolving 0.0214 g of KIO₃ in 50 mL of 2% HCL solution. After that 0.5 mL of this solution was added in 5 mL of each concentration of As(III). Then the solution was stirred and sonicated at room temperature for complete oxidation of As(III) to As(V). After sonication 0.5 mL of coloured reagent was added in series of vials having different concentration of As(III) solution, stirred slowly and kept for 30 min for the formation of blue colour antimony arseno-molybdate complex. UV– Visible spectra of each sample was recorded by running the scan from 400 to 1100 nm wavelength range. Absorbance at λ_max (900 nm) was used to get calibration curve, straight line equation later used to find unknown concentration of arsenic.

2.5 Adsorption study

To find the efficiency of adsorbents optimum conditions like concentration, pH, time, temperature and amount of adsorbent were achieved. For concentration and temperature effect 10 mg of each adsorbent was added in series of flasks having 10 mL of different concentration of As(III) having pH 7. For batch adsorption each flasks was placed in shaker bath at the speed of 130 rpm for 4 h. Adsorption was performed with respect to temperature (293–308 K) after 4 h, immediate filtration of dispersion was carried out to cut off the contact time of adsorbate and adsorbent. The filtrate was then analyzed by colorimetric method as discussed in Section 2.4. Amounts adsorb and percent removal was calculated by Eqs. (1) and (2).

Time is an important parameter that affects the adsorption process. To determine the kinetics of adsorption, optimization of time required to reach equilibrium was done. For this stock solution was diluted to 1000 μg/L and then 10 mL was taken in series of flask followed by addition of 10 mg of each adsorbent. All flaks were kept in shaker bath with speed of 130 rpm at 298 K for different time intervals varying from 5 to 360 min. After definite time intervals, flasks were taken out from shaker bath followed by filtration. The route of analysis was the same as adopted earlier, amount adsorbed at time t and percent removal at different time intervals was calculated by Eqs. (3) and (4).

Similarly, the effect of adsorbents amount was studied by taking 10 mL of 1000 μg/L As(III) solution in series of flask and the amount of adsorbents (2–30 mg) was added in each flask. The flasks were kept in shaker bath with speed of 130 rpm at 298 K for 4 h. Afterwards, each flask was taken out from the shaker bath followed by filtration. To determine the concentration at time t and% removal of As(III), filtrate was then analyzed by colorimetric method. Arsenite (As(III)) exist as H₃AsO₃ when pH is less than 8 however it exists in different anionic form like H₂AsO₃^-, HAsO₃^2- and AsO₃^3- when pH is higher than 8. Therefore, for the maximum removal of As(III), optimization of solution pH is very necessary. To study pH effect 10 mL of 1000 μg/L As(III) was taken in series of flasks and each solution was maintained at pH range from 2 to 10. Batch adsorption was performed at 298 K at the speed of 130 rpm for 4 h. After filtration concentration of As(III) was determined by using colorimetric method.

2.6 Desorption study

For desorption study 10 mg/L As(III) solution was prepared and adjusted at pH 7. Then 30 mL of this solution was transferred in each flask having 100 mg of different adsorbents MCH, MCH3 and MCNa. Adsorption studies were performed in shaker bath at 298 K for 4 h. After filtration the adsorbents were dried and divided into two equal parts and further used for desorption studies using 0.5 M solution of each NaCl and NaOH as eluents. Desorption studies were also performed in shaker bath at 298 K for 4 h. The percent desorption by each of eluent was calculated by using the following formula. $$\%\mathrm{d}\mathrm{e}\mathrm{s}\mathrm{o}\mathrm{r}\mathrm{p}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}=(\mathrm{A}\mathrm{m}\mathrm{o}\mathrm{u}\mathrm{n}\mathrm{t}\mathrm{d}\mathrm{e}\mathrm{s}\mathrm{o}\mathrm{r}\mathrm{b}/\mathrm{A}\mathrm{m}\mathrm{o}\mathrm{u}\mathrm{n}\mathrm{t}\mathrm{a}\mathrm{d}\mathrm{s}\mathrm{o}\mathrm{r}\mathrm{b})\ast 100$$

3.1 Nitrogen adsorption/desorption isotherms

Nitrogen adsorption/desorption isotherms were obtained in NOVA 1200e (Quantachrome, Virginia, USA) instrument. Before analysis, the samples were degassed at 102 °C for 8 h. Fig. 1 represents that BJH isotherms include two regions; low pressure region corresponds to monolayer segment while high pressure region that shows steep rise in nitrogen uptake correspond to multilayer adsorption. The shape of isotherms is of mixed type (type I and IV) of IUPAC, that suggest a mixture of microporous and mesoporous material. The initial portion of the isotherms are of type I that relates to substantial adsorption in micropores at relatively low pressure. The isotherms are of type IV with hysteresis loop of type H4 at intermediate and high relative pressure, associated with monolayer to multilayer adsorption followed by capillary condensation through a narrow slit like pores. More nitrogen uptake was observed in the case of ACH3 and ACNa (Fig. 1b, c) while least uptake was observed in case of ACH (Fig. 1a). This indicates the development of uniform porous structure when H₃PO₄ and NaOH were used as activators while in case of HNO₃ deterioration of wall of micropores occurs. Widens hysteresis loop indicates that structure is more populated with mesopores as represented in pore size distribution curve (inset Fig. 1). Present materials developed highly porous structure and shows adsorption at entire pressure range, no plateau is reached indicates wide range of pore diameter (Kumar and Jena 2016). BET surface area of ACH, ACH3 and ACNa was found to be 36 m²/g, 148.7 m²/g and 203.5 m²/g with mesopore volume 0.01 cm³/g, 0.13 cm³/g and 0.24 cm³/g respectively. Moving from acid to base activator a pronounced increase in surface area was observed that confirmed NaOH a good activator. The open loop observed in case of ACH and ACNa is due to presence of narrow slit or bottle shaped pores through which adsorbate don’t desorb easily as reported in previous study (Bibi et al., 2023).

Close

Fig. 1

Nitrogen adsorption/desorption isotherm and pore size distribution curve for (a) ACH (b) ACH3 and (c) ACNa.

Export to PPT

3.2 X-ray diffraction analysis (XRD)

The diffraction intensity data for activated and magnetic carbon samples were recorded in JDX-3532 X-ray diffractometer (JEOL-Japan) operated at 30 mA and 40 kV. Cu-Kα radiations of wavelength 1.54 Å were used for this purpose and the 2θ values were measured from 10 to 80° with a step size of 0.30° and step length of 5°/s. XRD pattern (Fig. 2) of activated carbon samples show amorphous character with two wide peaks around 2θ values 23° and 43° corresponds to (0 0 2) and (1 0 0) planes. The successful in situ synthesis of magnetite nanoparticles within amorphous activated carbon was confirmed from XRD pattern with diffraction at 30.3°, 36.1°, 43.1°, 53.3°, 59.7° and 62.1° attributed to planes (2 2 0), (3 1 1), (4 0 0), (4 2 2), (5 1 1) and (4 4 0) (Rafli et al., 2021).

Close

Fig. 2

X-ray diffractogram of (a) ACH, MCH (b) ACH3, MCH3 and (c) ACNa, MCNa.

Export to PPT

MAC samples exhibit a large peak at 2θ value of 36.1° attributed to (3 1 1) diffraction plane characteristics of single crystalline phase of inverse cubic spinel structure of magnetite nanoparticles (Anyika et al., 2017a, 2017b). The average crystallite size was calculated with the help of Scherrer’s formula given in Eq. (5).

3.3 Fourier transform infrared (FTIR) analysis

FTIR spectra of the synthesized materials was recorded on Nicolet 6700 (USA) in spectroscopic wavelength range of 400–4000 cm⁻¹ (Fig. 3). The FTIR spectroscopy gives the idea of functional groups present in the material. The different band appears in spectra of activated carbon reveals the presence of various functional groups. The band at 1486 cm⁻¹ with relatively high intensity corresponds to C—C stretching. The strong and broad spectral band at 3450 cm⁻¹ corresponds to stretching vibration of OH group present on the surface of activated carbon. Moreover, a short band around 2934 cm⁻¹ is assigned to C—H aliphatic symmetric and asymmetric stretching vibration of CH₂ and CH₃ groups. Furthermore, the C⚌C stretching vibration in the aromatic ring is often present in carbonaceous material responsible for absorption band in vicinity of 1656 cm⁻¹, while the band around 1450 cm⁻¹ attributed to C—H bending vibrations. Strong band around 1051 cm⁻¹ attributed to C—O—C stretching vibration (Chukwuemeka-Okorie et al., 2018). Change of activators results in increase or decrease in intensity of some band give the similar chemical structure. These observation suggest the presence of oxygen rich functional groups present on the surface of carbon, participate in formation of Fe₃O₄ on the surface. In magnetite incorporated activated carbon samples the band around 588–593 cm⁻¹ corresponds to Fe—O bond vibrations at octahedral and tetrahedral sites.

Close

Fig. 3

FTIR spectra of (a) ACH, MCH (b) ACH3, MCH3 and (c) ACNa, MCNa.

Export to PPT

3.4 Thermogravimetric analysis (TGA)

Fig. 4 displays the thermogram of activated and magnetic carbon samples. The sample mass change as a function of temperature was determined using the Thermal Analyzer (model Perkin Elmer USA). During this analysis, a small mass of sample is heated in furnace at temperature between 25 and 2200 °C attached with sensitive balance. As illustrated in Fig. 4, all the samples display good thermal stability. Initial weight loss observed in ACH, ACH3 and ACNa is 4.7, 9.6 and 10.4% respectively due to entrapped moisture content into porous network. Moreover, activated carbon samples were found stable from 300 to 1200 °C and a few percent weight loss was observed due to the breakdown of low volatile organic compounds. However, initial weight loss observed in MCH, MCH3 and MCNa due to desorption of physically adsorbed water molecules was 8.5, 8.3 and 6.4% respectively (Anyika et al., 2017a, 2017b).

Close

Fig. 4

Thermogram of (a) ACH, MCH (b) ACH3, MCH3 and (c) ACNa, MCNa.

Export to PPT

In case of magnetized sample stability is shift toward high temperature (1500 °C) that is more pronounced in case MCNa (Vinayagam et al., 2022). Overall, thermal analysis suggests that incorporation of magnetite enhances the thermal stability of activated carbon.

3.5 Field emission scanning electron microscopy (FESEM) and energy dispersive X-ray (EDX) analysis

Surface morphology of activated carbon before and after magnetization was analyzed by FESEM. The micrographs in Fig. 5a, c and e clearly show the rough and porous structure of activated carbon with irregular channels and hollow cavities for ACH, ACH3 and ACNa respectively. The porosity of synthesized activated carbon was found to be comparable with commercial activated carbon as reported earlier (Rodrigues et al., 2020). At higher magnification development of random mesopores and macropores (2–100 nm) on the surface of activated carbon can be seen. The porosity in activated carbon is developed due to cellular structure of its source, activators and temperature conditions. Incorporation of magnetite resulted in filling of cavities with clear boundaries as can been seen in micrograph Fig. 5b, d and f for MCH, MCH3 and MCNa respectively. Relatively smooth surface shows the successful incorporation of magnetite into cavities that resulted in the formation of more functional group on surface making it more efficient for removal of As(III).

Close

Fig. 5

FESEM of (a) ACH (b) MCH (c) ACH3 (d) MCH3 (e) ACNa and (f) MCNa.

Export to PPT

Fig. S1 shows the EDX spectra of activated and magnetic carbon. EDX analysis of all the activated carbon samples (ACH6, ACH3 and ACNa) shows C and O as a major constituent with N, P and Na as minor constituents that comes from the activators used (Fig. S1a, c and e). The elemental composition shows high contents of C ranging from 76.7 to 86.7 wt% in activated carbon samples followed by oxygen 10.7–21.3%. Besides these elements in magnetized activated carbon (MCH, MCH3 and MCNa) Fe is also present as a major constituent. The Fe content was found in the range of 7.46–12.50 wt% (Fig. S1b, d and f) confirming the effective incorporation of magnetite into porous network of activated carbon.

3.6 Zeta potential

The zeta potential of each adsorbent was measured to determine the isoelectric point (IEP) and surface charge with respect to pH of medium. It is well documented that below pH_IEP the adsorbent develops positive charge while moving above pH_IEP surface of adsorbent gets negatively charged. IEP for ACH, ACH3 and ACNa was found to be at pH 2.12, 1.93 and 3.03 respectively (Fig. 6a–c) which is close to those reported in literature for activated carbon (Niksirat et al., 2019).

Close

Fig. 6

Zeta potential of (a) ACH and MCH (b) ACH3 and MCH3 (c) ACNa and MCNa.

Export to PPT

It is interesting to note that after incorporation of magnetite into pores of activated carbon, a shift in pH_IEP towards higher values 5.37, 4.42 and 5.78 was detected for MCH, MCH3 and MCNa respectively. These results are in correspondence with the literature value reported for magnetite particles at around 6.3 (Manrique et al., 2019). The shift in pH_IEP towards higher pH value corresponds to development of positive functional group on the surface of activated carbon. This shift in pH_IEP contributes to high adsorption capacity of material toward As(III) in pH range of 5–8 as explained in section 3.7.1. It is important to note that incorporation of magnetite shifts the zeta potential to more positive value for a wide pH range.

3.7 Adsorption studies

3.7.1

3.7.1 Effect of different parameters

To optimize the parameters that affect the percentage removal of arsenic by different adsorbent, the effect of various parameters was studied (Fig. 7). The percentage removal of As(III) shows a sharp increase for the initial 30 min, beyond this there was a gradual increase up to 120 min. Further increase in contact time showed negligible effect on percentage removal ensuring the attainment of equilibrium within 240 min. The initial fast adsorption rate was attributed to availability of large number of vacant sites. However, after this percentage removal goes ahead with slower rate due to less available sites that get finally saturated (Sahu et al., 2021). As evident from the graph, adsorption of As(III) on all the adsorbent surface is continuous and smooth till the saturation point indicating monolayer surface coverage (Fig. 7a).

Close

Fig. 7

Effect of different parameters (a) contact time (b) Amount of adsorbent (c) pH of medium and (d) initial concentration of As(III).

Export to PPT

Fig. 7b represents the effect of amount of adsorbent on percent removal of As(III). An increase in adsorbent amount (0–10 mg) causes a sharp increase in As(III) removal until it reaches maximum value for 20 mg of adsorbent. Beyond this increase in adsorbent amount have shown negligible effect on percent removal of As(III). Almost 99% of As(III) was removed by using 30 mg of MCNa which was only 50% in case of MCH. Low adsorption capacity in case of MCH attributed to availability of a smaller number of active sites on surface, because most of magnetite nanoparticles goes into large cavities clear from FESEM image. It was noted that for all adsorbent systems initially, there is larger increase in percent removal due to greater exchangeable sites of adsorbent. An onwards increase in adsorbent amount has shown no considerable effect on percent removal that might be due to unavailability of active sites because of agglomeration of adsorbent particles (Mondal et al., 2008).

The pH of medium has greatly influenced the surface charge of adsorbent as well as ionization of adsorbate specie that in turn affected the adsorption of As(III). The difference in adsorption capacities as a function of pH can be explained by zeta potential. It is cleared from zeta potential values (Fig. 6) the IEP for all the adsorbent systems lies in the range of 4–6 above this value, the surface charge of adsorbents gets negative. As(III) generally exist as uncharged specie like H₃AsO₃ upto pH 8 however, for pH values ˃8 it exists in different anionic form like H₂AsO₃^-, HAsO₃^2- and AsO₃^3-_. Fig. 7c shows that percent removal of As(III) increases gradually until it reaches maximum value at pH 7 and then decreases sharply at higher pH due to repulsion between oxyanions and negative charge on adsorbent surface (Alam et al., 2018). At low pH adsorption capacity of material slightly decreases due to protonation of surface hydroxyl groups. From this study, pH 7 was found suitable for As(III) adsorption experiments. Charged neutral state of As(III) was found to be most suitable for adsorption, hence maximum removal efficiency was observed at pH 7. The effect of initial concentration of adsorbate on adsorption capacity was shown in Fig. 7d. It is obvious that as initial concentration varies from 100 to 10,000 μg/L, the percentage of arsenic removal reduces. For lower concentration (100–1000 μg/L) there is almost 100% removal of As(III), while by increasing As(III) concentration, competition for active adsorption sites increases which results in gradual decrease of percentage removal of arsenic (Babaee et al., 2018).

3.7.2

3.7.2 Kinetic modelling

To figure out the feasibility of adsorption process, it is important to know how long it takes for the system to attain equilibrium. Due to mass transfer phenomenon in an adsorption system, equilibrium is attained gradually but not instantly. The kinetics of adsorption is progression of adsorption process over time to obtain time concentration profile. Different phenomenon involving bulk diffusion, external diffusion, intraparticle diffusion and physical or chemical adsorption control the adsorption at solid/liquid interface (Sahoo and Prelot 2020). Various models were used to comprehend the specific properties of As(III) adsorption on magnetic activated carbon. In the first 30 min of adsorption process the Pseudo 1st order kinetic model hold true by assuming that the solute uptake rate is proportional to equilibrium concentration difference (Azizian 2004).

When adsorption follows diffusion or physisorption, Pseudo 1st order kinetics was applicable. In its linear form pseudo 1st order kinetic model can be expressed as:

(6)

$$\mathit{ln}{(q}_e-q_t)=ln{q}_e-k_{1}t$$

Fig. 8a represents the linear plot of $\mathit{ln}{(q}_e-q_t$ ) vs t. The slope of plot gives value of 1st order rate constant while q_e (μg/g) can be computed from intercept. The statistical analysis was also performed and has shown in Table 1. The regression coefficient value was found less than 0.90 which suggest that Pseudo 1st order is not a well fitted model for adsorption of As(III) by magnetic activated carbon. Pseudo 2nd order kinetic model is often regarded as adsorption controlled kinetic model that predicts the behavior across the whole adsorption rang and take into account the creation of surface complexes as a rate limiting step (Bullen et al., 2021). In this situation solute sorption relies on unoccupied adsorbent sites rather than adsorbate concentration and reaction rate depends on quantity of adsorbate on surface. Linearized form of Pseudo 2nd order is given below:

(7)

$$\frac{t}{q_t}=\frac{1}{k_2{q}_{e^2}}+\frac{t}{q_e}$$

Close

Fig. 8

Kinetic data of As(III) adsorption on MCH, MCH3, MCNa fitted with (a) Pseudo 1st order (b) Pseudo 2nd order (c) Elovich model and (d) Intraparticle diffusion model.

Export to PPT

Table 1 Kinetic parameters for different adsorbent.

Kinetic models	Kinetic Parameters	Adsorbent
		MCH	MCH3	MCNa
Pseudo 1st order model	k1 (min−1)	0.0299	0.027	0.025
	qe (cal) (μg/g)	446.6	405.9	286.6
	qe (exp) (μg/g)	1155	894	899
	R2	0.788	0.872	0.681
	RMSE	0.78912	0.5248	0.9431
	Reduced Chi-Sqr	0.62271	0.27541	0.88943
Pseudo 2nd order model	k2 (min−1)	1.72 × 10-4	2.52 × 10-4	3.56 × 10-4
	qe (cal) (μg/g)	1176	909	909
	qe (exp) (μg/g)	1155	894	899
	R2	0.999	0.999	1
	RMSE	0.00236	0.00119	0.02702
	Reduced Chi-Sqr	5.581E-6	1.413E-6	7.300E-4
Elovich model	β (g/μg)	0.0222	0.013	0.022
	α (μg/g min)	2.32 × 1010	36,600	8.75 × 107
	R2	0.9686	0.9079	0.9006
	RMSE	11.0927	37.62911	19.29079
	Reduced Chi-Sqr	123.047	1415.949	372.134
Intra particle diffusion model	Ci	1009.5	631.94	775.57
	kid (μg/g min1/2)	8.8475	16.995	7.5916
	R2	0.8662	0.6904	0.7364
	RMSE	22.913	77.349	29.921
	Reduced Chi-Sqr	525.039	5982.918	895.296

Fig. 8b represents the Pseudo 2nd order model, from which the value of q_e can be computed to calculate 2nd order rate constant (Table 1). There is close agreement between q_{e (cal)} and q_{e (exp)} and R² values are close to 0.999 for all the adsorbents suggest that adsorption of As(III) on magnetic activated carbon truly follows Pseudo 2nd order kinetics. The Elovich kinetic model interprets the kinetics of adsorption by assuming that the surface is energetically heterogeneous. It offers a good understanding of the chemisorption nature of kinetics. This model assume that adsorption rate decreases exponentially with the solute concentration (Kajjumba et al., 2018). Linear form of Elovich model can be expressed as:

(8)

$$q_t=\frac{1}{\beta }\mathrm{l}\mathrm{n}(\alpha \beta )+\frac{1}{\beta }lnt$$

This model is shown in the form of Fig. 8c. The value of activation energy of chemisorption β can be used to calculate value of initial sorption rate α from intercept. The value of R² greater than 0.90 suggests that along with Pseudo 2nd order Elovich model equally fitted to the experimental data. To identify diffusion mechanism linear form of intra particle diffusion model (Eq. (9)) was also applied to the experimental data.

(9)

$$q_{t=k_{\mathit{id}}{t}^{1/2}+C_i}$$

Plot of q_t vs t^1/2 (Fig. 8d) shows several linear segments that point to different steps involved in sorption process. As can be seen, the smooth curve followed by linear segment suggest that initial kinetics rate was controlled by surface reaction followed by intraparticle diffusion (Zhang et al., 2010).

3.7.4

3.7.4 Adsorption modelling

The link between isotherms at equilibrium has been the subject of study when different equilibrium models were constructed. Adsorption modelling is used to describe the experimental data to evaluate the effective interaction between adsorbate and adsorbent (Chen et al., 2022). Fig. 9 represents the applicability of Freundlich, Langmuir and Temkin models on the experimental adsorption data obtained by adsorbents at different temperatures. Langmuir isotherms describe the adsorption to take place on identical and energetically equivalent adsorption sites. It assumes monolayer adsorption without solute–solute interaction. Herein, the non-linear and linear forms of Langmuir model were applied by using following equations.

(10)

$$q_e=\frac{{\mathrm{q}}_{\mathrm{m}}{\mathrm{k}}_{\mathrm{L}}{\mathrm{C}}_{\mathrm{e}}}{1+{\mathrm{k}}_{\mathrm{L}}{\mathrm{C}}_{\mathrm{e}}}$$

(11)

$$\frac{C_e}{q_e}=\frac{1}{q_m{K}_L}+\frac{C_e}{q_m}$$ where C_e is the equilibrium concentration of As(III) (μg/L), q_e is the amount of As(III) adsorbed at equilibrium (μg/g), q_m is the monolayer adsorption capacity (μg/g), k_L is the adsorption equilibrium Langmuir constant (binding energy constant). As represented in Fig. 9a plot of $\frac{C_e}{q_e}$ vs q_e is a straight line with regression coefficient of 0.99 suggest that adsorption follows Langmuir model. This is a direct indication that adsorption takes place through surface complexation and ionic/covalent interaction leading to formation of monolayer. The monolayer adsorption capacity (q_m) calculated from slope of the plot is listed in Table 2. Dimensionless constant R_L is an essential feature of Langmuir model and can be expressed as:

(12)

$${\mathrm{R}}_{\mathrm{L}}=\frac{1}{1+{K_L\mathrm{C}}_{\mathrm{o}}}$$ where K_L is the Langmuir constant related to energy of adsorption and useful to calculate value of R_L that can be used to predict adsorption feasibility. If R_L ˃ 1 adsorption is unfavorable, 0 < R_L less than 1 adsorption is favorable, R_L = 1 linear adsorption while for R_L = 0 adsorption is irreversible (Ayawei et al., 2017). In the present study the R_L values lies between 0 and 1, indicating the favourable adsorption of As(III) (Table 2).

Close

Fig. 9

Langmuir, Freundlich and Temkin model for adsorption of different concentration of As(III) on (a,b,c) MCH (d,e,f) MCH3 and (g,h,i) MCNa respectively.

Export to PPT

Table 2 Adsorption parameters calculated for adsorption of As(III) by different adsorbent.

Adsorption Isotherm Model	Adsorption parameters	Adsorbents
		MCH			MCH3			MCNa
		298 K	303 K	308 K	298 K	303 K	308 K	298 K	303 K	308 K
Langmuir	qm (μg/g)	725	685	613	1818	1695	1075	1469	1892	1703
	KL (L/μg)	0.023	0.035	0.020	0.011	0.010	0.0095	0.005	0.003	0.002
	R2	0.997	0.999	0.998	0.998	0.998	0.999	0.789	0.813	0.699
	RL	0.042	0.028	0.048	0.083	0.090	0.095	0.17	0.25	0.33
Freundlich	KF (L/μg)	449	414	339	1014	710	490	416	273	133
	1/n	0.053	0.059	0.068	0.058	0.093	0.084	0.140	0.217	0.278
	R2	0.958	0.959	0.960	0.860	0.949	0.936	0.993	0.965	0.867
Temkin	B (J/mol)	31.41	34.23	35.73	84.40	125.64	78.78	169.94	286.82	282.47
	KT (L/μg)	1.04 × 106	0.09 × 106	0.005 × 106	0.082 × 106	0.053 × 103	0.072 x 103	0.705	0.097	0.035
	R2	0.930	0.943	0.941	0.828	0.931	0.928	0.799	0.865	0.766

Freundlich model on the other hand explains the exponential relation between amount of solute adsorbed (q_e) and equilibrium concentration (C_e). According to this model, surface adsorption is multilayer on heterogeneous sites. In this study non-linear and linear form of Freundlich model was applied to our experimental data can expressed by equations given below.

(13)

$$q_e=k_f{C_e}^{1/n}n>1$$

(14)

$$\mathit{ln}{q}_e=\frac{1}{n}ln{C}_e+ln{K}_f$$

Freundlich constant 1/n and K_f represent the adsorption intensity (heterogeneity factor) and adsorption capacity respectively (Table 2). The greater value of 1/n implies the more interaction between adsorbate and adsorbent. In present study values of 1/n were found less than 1 indicating the favourable adsorption process (Mandal et al., 2013). Besides Langmuir and Freundlich, Temkin model was also applied to experimental data that is very important to estimate the heat of sorption. This model implies that there is uniform distribution of binding energies upto some maximum binding energy. Both linear and non– linear Temkin model were applied in the form:

(15)

$${\mathrm{q}}_{\mathrm{e}}=\frac{\mathrm{R}\mathrm{T}}{{\mathrm{b}}_{\mathrm{T}}}\mathrm{l}\mathrm{n}({{\mathrm{k}}_{\mathrm{T}}\mathrm{C}}_{\mathrm{e}})$$

(16)

$${\mathrm{q}}_{\mathrm{e}}=\frac{\mathrm{R}\mathrm{T}}{{\mathrm{b}}_{\mathrm{T}}}\mathrm{l}\mathrm{n}{{\mathrm{k}}_{\mathrm{T}}+\frac{\mathrm{R}\mathrm{T}}{{\mathrm{b}}_{\mathrm{T}}}\mathrm{l}\mathrm{n}\mathrm{C}}_{\mathrm{e}}$$

Binding constant k_T (L/μg) gives information regarding binding energy while $\mathrm{B}=\frac{\mathrm{R}\mathrm{T}}{{\mathrm{b}}_{\mathrm{T}}}$ expresses the heat of adsorption (J/mol).

As given in Table 2, positive value of B suggest that heat absorbed during the sorption phenomenon (Inyinbor et al., 2016). Adsorption and kinetic modelling suggest that adsorption of arsenic by these adsorbents is physicochemical and surface complexation is preferred mechanism than electrostatic attraction. Chemical activation and magnetization not only enhanced arsenic adsorption capacity as compared to other low-cost adsorbent (Table 3), but also ensure its magnetic recovery to prevent secondary pollution and reutilization.

Table 3 Comparison of q_m (mg/g) for removal of As(III) by low cost biosorbents.

Adsorbent	qm (mg/g)	References
MCNa	1.89	Present study
MCH3	1.69
MCH	0.72
FeMnC	0.72	(Chen et al., 2021)
Fe3O4/Sugar can bagasse activated carbon	6.69	(Joshi et al., 2019)
Magnetic palm kernel shell activated carbon	0.048	(Anyika et al., 2017a, 2017b)
Wheat husk iron oxide composite biochar (WHIOB)	0.113	(Singh et al., 2020)
Rice husk iron oxide composite biochar (RHIOB)	0.096
Fe-UTAC	6.83	(Mahmood et al., 2018)
Thioglycolated sugarcane carbon	0.085	(Roy et al., 2013)
Pine biochar (PB)	0.265	(Wang et al., 2015)
Hematite modified biochar (HPB)	0.429
Fe@AC from Tectona grandis sawdust	0.679	(Sahu et al., 2021)
Fe-Zr@AC from Tectona grandis sawdust	1.206

3.8 Thermodynamics and spectroscopic investigation of adsorption/desorption

To acknowledge the adsorption mechanism, thermodynamics parameters like Gibb’s free energy ( $\mathrm{\Delta }{G}^o)$ , enthalpy $\mathrm{\Delta }{H}^o$ and entropy $\mathrm{\Delta }{S}^o$ are of worth importance that can be evaluated by Van’t Hoff equation.

By utilizing Fig. 10d $\mathrm{\Delta }{H}^o$ and $\mathrm{\Delta }{S}^o$ can be evaluated from slope and intercept of linear plot of $\mathit{lnK}$ vs 1/ $T$ respectively (Al-Salehin et al., 2019). The positive values of $\mathrm{\Delta }{G}^o$ depict that reaction is nonspontaneous while negative values are indicative to spontaneous reaction. Value of $\mathrm{\Delta }{G}^o$ can be calculated by using Eq. (18) (Wu et al., 2019), outcomes of the study are given in Table 4.

Close

Fig. 10

Comparison of FTIR spectra before and after adsorption of As(III) by (a) MCH (b) MCH3 (c) MCNa. (d) plot of lnK vs 1/T (e) Desorption of As(III).

Export to PPT

Adsorption of As(III) by MCH was found to be endothermic and nonspontaneous, moving towards spontaneity at high temperature. The positive $\mathrm{\Delta }{H}^o$ and $\mathrm{\Delta }{S}^o$ reflected that adsorption is unfavorable and entropy driven process. For the other two system MCH3 and MCNa adsorption was found to be exothermic and non– spontaneous at high temperature and hence, the adsorption was found to be enthalpy driven.

Fig. 10a–c are concerned with spectroscopic investigation of As(III) adsorption which confirmed As(III) uptake. After As(III) adsorption a strong band around 619 cm⁻¹ and 674 cm⁻¹ appears due to As-O-Fe bidentate or monodentate complexes at different surface coverages (Taleb et al., 2019). Beside this, As-OH free bond in protonated or deprotonated form may be present depending on pH of medium. A small band around 889 cm⁻¹ correspond to As-O stretching vibration (Yu et al., 2019). Further, intensity and sharpness of C-O stretching vibrations band enhanced after adsorption of As(III).

To find the influence of eluent on the As(III) desorption from solid adsorption residue desorption study was performed (Fig. 10e). For this purpose, two different eluent systems including 0.5 M NaOH (pH = 13) and 0.5 M NaCl (pH = 6) were used (Di Natale et al., 2013). Present study suggests NaCl as a better eluent compared to NaOH with maximum of 26% elution in case of MCH3 which was only 12.4% in case of MCNa.

Since predominant adsorption mechanism is chemisorption and complex formation, therefore very less desorption of arsenic in the range 12.4%-26% was observed. This study demonstrated that desorption strongly depended on pH, moving from basic to neutral and then acidic% desorption increases but in case of magnetite incorporated carbon acidic media result in leaching of iron into desorption media (Štefušová et al., 2010). Keeping in view the pH effect, it was concluded that NaCl can be used as a better eluent for As(III), from the surface of magnetite incorporated activated carbon.