Manganese Adsorption onto Permanganate-Modified Bamboo Biochars from Groundwater

Abstract

Potassium permanganate-modified bamboo biochar (MBB) was used to adsorb manganese from simulated groundwater and its performance was compared to that of unmodified bamboo biochar (BB), activated carbon, and manganese greensand. The adsorption kinetics, adsorption isotherms, and manganese fractions were investigated. The Langmuir model was the best fit for manganese adsorption by MBB and BB at the maximum adsorption capacities of 21.277 and 0.803 mg g⁻¹, respectively. The heat of adsorption from the Temkin model indicated that manganese adsorption occurs via an ion exchange process for MBB and a physical adsorption process for BB. The sequential extraction results revealed that manganese was strongly bound to the iron/manganese oxide fraction, in accordance with the chemical adsorption established in pseudo-second order kinetic data records.

1. Introduction

Many water utilities strive to remove manganese (Mn) from sources of drinking water. Mn exhibits a range of valencies and colors depending on its oxidation state; Mn(II) is soluble in water and has a pale pink color, while Mn(IV) is insoluble and dark brown [1]. It is found naturally in the Earth’s crust; the geological characteristics of the Earth have led to the natural contamination of groundwater supplies. Mn removal from groundwater is chemically difficult; it accumulates in distribution systems and affects the taste and color of drinking water even at low concentrations [1]. Various studies have reported the negative impacts of Mn. Environmental Mn can affect verbal processing functions in adults [2], and another study revealed that school-aged children exposed to higher concentrations of Mn in drinking water had lower IQ levels [3]. In southern Quebec, Canada, groundwater sources for tap water have become contaminated with a high Mn content of 34 μg L⁻¹ [3]. Further, high Mn consumption in southern China has significantly impacted people’s health, ultimately resulting in chronic illnesses and symptoms related to brain function [4]. In 2022, the World Health Organization (WHO) recommended that the Mn concentrations in safe drinking water be less than 0.08 mg L⁻¹, while the maximum allowable concentration of Mn in groundwater for drinking purposes in Thailand was 0.3 mg L⁻¹ [5]. Thus, groundwater contaminated with Mn to levels above those recommended must be treated to make drinking water safe for consumption.

Mn removal from groundwater can be carried out via physical, chemical, biological, and physicochemical processes [6]. The oxidation [7] and filtration [8] processes are currently popular due to their high efficiencies in Mn removal [9,10]. However, ozone oxidation is expensive, and chlorine oxidation often produces hazardous trihalomethane byproducts, which are carcinogenic agents. Thus, filtration is suitable for household Mn removal. Activated carbon (AC) is commonly used as filter media [11], however, it has no capacity with regard to oxidation. Thus, manganese greensand—sand coated with manganese oxide—has been employed to remove Mn. Recently, adsorbents coated with various oxidizing agents, such as zeolite tuff [12], kaolin [13], and manganese oxide-coated zeolite [14], have been developed for Mn removal. However, commercial adsorbents such as AC and greensand are expensive. Therefore, biochar and other natural materials have recently been studied more thoroughly for heavy metal removal. Biochar has high porosity and similar adsorption properties to AC, and serves as a low-cost alternative sorbent or adsorbent for the removal of metals, including Mn, from groundwater. Biochar from poultry and farmyard manure has served as an adsorbent for the removal of Mn from groundwater [15], and bamboo biochar (BB) has been used for the adsorption of lead(II) from aqueous solutions [16]. Biochar is a product of the thermal decomposition of biomass. Three different methods have been reported for thermal decomposition, namely, pyrolysis, hydrothermal carbonization, and microwave-assisted carbonization, in which pyrolysis temperature is between 400–700 °C. As biochar has low surface area and porosity, physical and chemical activation methods have been applied both pre- and post-carbonization [17]. For example, biochar coated with potassium permanganate has proven effective in the removal of cadmium [18,19] and lead [20] from aqueous solutions. Piispanen and Sallanko [21] reported that the addition of manganese oxide-coated media in filters can improve Mn removal to achieve Mn concentrations of less than 0.02 mg L⁻¹. Furthermore, Chen et al. [22] reported that MnO₂ can be produced at the surface of bamboo biochar (BB) for effective uranium(VI) adsorption. Among the biochars derived from biowastes and biomass, bamboo is a good choice, as it is abundantly available, has a high growth rate, and is easy to process. Moreover, the need for a cost-effective solution for Mn pollution control, particularly in developing countries and rural areas where advanced technologies such as ozone oxidation are not possible. The use of permanganate-modified bamboo biochar has previously been proposed as a novel technology. Although many adsorbents are used to remove heavy metals, they may be capable of both storing and releasing them. The chemical fractionation of Mn on the adsorbent can be used to determine the metal released from the adsorbent. The sequential extraction processes presented by Tessier [23] consist of numerous steps, resulting in the following fractions: water soluble, exchangeable, bound to carbonates, bound to iron and manganese oxide, bound to organic matter, and residual. The highly soluble and exchangeable fractions present the risk of heavy metals leaching into the environment. The mobility of Mn is often increased at low pH or redox potentials. Previous studies have examined the Mn fractions in water, soil, and sediments; for example, in soil amendments with sewage sludge, biochar reduced Mn mobility in contaminated sites, decreasing the acid-soluble fractions by 70.38% [24]. Another study found that dissolved and particulate Mn in the Seine River estuary were bound to calcium carbonate, leading to the stabilization of the Mn(II) form [25]. Furthermore, the intraparticle diffusion, pseudo-first order (PFO) and pseudo-second order (PSO) can be used to describe adsorption mechanisms and explain the uptake rate of adsorbate on biochar [26]. Therefore, it is important to refer to the results of such studies as a guide for effective heavy metal removal.

Few studies have been conducted on the removal of Mn from water using bamboo biochar and modified bamboo biochar by permanganate impregnation. Thus, modified bamboo biochar may be a novel adsorbent for Mn adsorption. In addition, previous research has hardly investigated the comparison of adsorption performances of modified adsorbents with commercial greensand and activated carbon for Mn removal in real groundwater. Therefore, this study aims to investigate the removal efficiencies of manganese for bamboo biochar (BB) and potassium permanganate-modified bamboo biochar (MBB) as compared with those of AC and manganese greensand (MnG) in order to study the adsorption kinetics and isotherms and investigate the Mn fractionation in the adsorbents after the adsorption process using real groundwater with a certain Mn concentration.

2. Materials and Methods

2.1. Preparation of Simulated Groundwater

Groundwater was withdrawn from the NonNhon subdistrict, Warinchamrap district, Ubon Ratchathani province in Thailand. Simulated groundwater was prepared by adding MnSO₄.H₂O (analytical grade, 98%, KemAus™, Sydney, Australia) to the natural groundwater at specific concentrations of 1, 3, 5, 7, and 9 mg L⁻¹ for the batch experiments. The chemical characteristics of the groundwater are presented in

Table 1. The chemical characteristics of the groundwater.

Parameter	Analytical Method	Unit	Value
pH	Electrometric Method	-	6.78
Chloride	Argentometric Method	mg L−1	140
Total Hardness	EDTA Titrimetric Method	mg L−1 as CaCO3	420
Total Dissolved Solids	Gravimetric Method	mg L−1	625
Sulfate	Turbidimetric Method	mg L−1	25.28
Nitrate	Colorimetric Method	mg L−1	2.99
Ferrous	ICP-OES *	mg L−1	0.174
Manganous	ICP-OES *	mg L−1	1.098
Magnesium	EDTA Titrimetric Method	mg L−1 as CaCO3	12.91
Calcium	EDTA Titrimetric Method	mg L−1 as CaCO3	34.75

* Inductively coupled plasma-optical emission spectrometry.

.

2.2. Adsorbents

The adsorbents in this study included BB, MBB, AC, and MnG. BB was prepared via the carbonization of bamboo (Bambusa beecheyana) from Ubon Rachathani province, Thailand, in a mound kiln under oxygen-limited conditions and within the temperature range of 400–600 °C. The obtained BB was crushed to a uniform size in the range of 2.0–3.0 mm. The preparation of MBB was adapted from the method of Taffarel and Rubio [27] and Xuwen et al. [28]. Here, 10 g of the crushed BB was placed in a 500-mL beaker containing 400 mL of 5% (w v⁻¹) KMnO₄ solution (KMnO_4, analytical grade, 99%, KemAus™, Sydney, Australia) and then boiled and stirred at 90 °C for 1 h. The suspension was allowed to cool and then filtered and rinsed with tap water until the water washings became colorless. The biochar samples were then dried in an oven at 60 °C for 6 h. Commercial-grade AC and MnG were purchased from a domestic supplier (J.L. Intertrade Co., Ltd., Bangkok, Thailand) and were sieved to obtain particles 2.0–3.0 mm in size before use. All adsorbents were kept in closed containers and used for all the experiments.

2.3. Characterization of Adsorbents

The external surface structural morphology of all adsorbents was determined using scanning electron microscopy and energy dispersive X-ray spectrometry (SEM-EDS, JEOL Model JSM-7610F Plus, Tokyo, Japan) before and after Mn adsorption. The surface of the adsorbents was investigated using a surface area and pore size analyzer (BET, Quantachrome Model Quadrasorb evo, Anton Paar GmbH, Graz, Austria). The chemical bonds or functional groups in the adsorbent molecules were analyzed before and after adsorption by Fourier transform infrared spectroscopy (FT-IR, Thermo Scientific Nicolet 6700, Waltham, MA, USA). Elemental analysis of the adsorbent materials was conducted using a wavelength dispersive X-Ray fluorescence spectrometer (WDXRF, Model Rigaku ZSX Primus, Tokyo, Japan). The adsorption capacities of the adsorbents were tested by the adsorption of iodine following the ASTM D4607 standard.

The measurement of the point of zero charge (pH_pzc) was adapted from the method of Yang et al. [8]. A 0.005 M CaCl₂ solution (CaCl₂.2H₂O, analytical grade, 78%, Q RëC™, Auckland, New Zealand) was boiled for 15 min to remove CO₂ and then allowed to cool to room temperature. The initial pH (pH_I) of the solution was adjusted to 3, 5, 7, and 9 via the dropwise addition of 0.5 M HCl (HCl, analytical grade, 37%, Q RëC™, Auckland, New Zealand) or 0.5 M NaOH (NaOH, analytical grade, 99%, Q RëC™, Auckland, New Zealand). Subsequently, 1 g of the adsorbents was added to 30 mL of each pH solution in 50-mL screw cap test tubes, the mixtures were shaken at 150 rpm for 24 h at 28 ± 0.5 °C, and the pH value after adsorption (pH_F) was measured. The pH_F values were plotted against the pH_I ones to determine pH_pzc from the intersection point of the curves with the pH_I = pH_F line.

2.4. Adsorption Experiments

2.4.1. Study of Contact Time

Batch adsorption experiments were carried out to determine the equilibrium contact time of Mn adsorption. The simulated groundwater with an Mn concentration of 5 mg L⁻¹ was adjusted to pH 6.7 ± 0.1 with 1 N HNO₃ (HNO₃, analytical grade, 65%, Q RëC™, Auckland, New Zealand) and 1 N NaOH. Dosages of 50 g L⁻¹ of BB and 0.25 g L⁻¹of MBB, AC, and MnG were used. The mixtures were shaken for various times—24, 48, 72, and 96 h—at 100 rpm and 28 ± 0.5 °C. The suspensions were then filtered through Grade 42 Whatman filter paper and the Mn concentration of the solutions was analyzed via inductively coupled plasma-optical emission spectroscopy (ICP-OES, Perkin ElmerOptima 8000, Rodgau, Germany).

Then, the amount of Mn adsorbed on each adsorbent and the percentage removal efficiencies were calculated following Equations (1) and (2), respectively:

$$q_t=\frac{\left(C_o-C_t-C_{control}\right)}{m}\times V,$$

(1)

where $V$ is the volume of the solution (L), $C_o$ is the initial Mn concentration (mg L⁻¹), $C_t$ is the final Mn concentration at time $t$ (mg L⁻¹), $C_{control}$ is the Mn concentration of the blank control without the addition of an adsorbent, and m is the sorbent mass (g).

$$Removal efficiency (\%)=\frac{C_0-C_t}{C_0}\times 100$$

(2)

2.4.2. Adsorption Kinetics

The kinetics of Mn adsorption from simulated groundwater at an Mn concentration of 5 mg L⁻¹ were determined using BB, AC, MBB, and MnG in amounts of 50, 50, 0.25, and 0.25 g L⁻¹, respectively. The mixtures were shaken in an incubator shaker at 30 °C and 100 rpm at an initial pH of 6.7 ± 0.1. In addition, in order to investigate the influence of temperature at 25 and 35 °C on thermodynamic adsorption, the solutions were separated from the mixtures via filtration through Grade 42 Whatman filter paper before analysis of the Mn concentration. All experiments were conducted in triplicate. The kinetics of Mn adsorption were evaluated by applying two different models: the pseudo-first order and pseudo-second-order kinetic models [29]. The pseudo-first order kinetic model is expressed as in Equation (3):

$$log(q_e-q_t)=log q_e-\frac{k_1}{2.303}t,$$

(3)

where $q_e$ and $q_t$ are the amounts of Mn (mg g⁻¹) adsorbed at equilibrium and at time $t$, respectively, and $k_1$ is the Lagergren pseudo-first order adsorption rate constant (min⁻¹). The linear form of the pseudo-second order model can be represented by Equation (4):

$$\frac{t}{q_t}=\frac{1}{k_2 q_e^2}+\left[\frac{t}{q_e}\right],$$

(4)

where $k_2$ (g mg⁻¹ h⁻¹) is the pseudo-second order rate constant.

Intraparticle diffusion is the mechanism of heavy metal adsorption onto adsorbent particles. The most widely applied intraparticle diffusion equation for adsorption systems is provided by Cheung et al. [30] as Equation (5):

$$q_t=k_i{t}^{0.5}+c$$

(5)

where $k_i$ is the intraparticle diffusion rate constant (mgg⁻¹ min^1/2), $q_t$ is the amount of solute on the surface of the sorbent at time $t$ (mgg⁻¹), and $k_i$ is the intraparticle diffusion rate constant (mgg⁻¹ min^1/2) and the intercept of the plot of $q_t$ versus t^1/2. When film diffusion is present, the intercept is $c$, which provides insight into the boundary layer’s thickness.

2.4.3. Adsorption Isotherms

The adsorption isotherms were determined using simulated groundwater at Mn concentrations of 1, 3, 5, 7, and 9 mg L⁻¹ at an adjusted pH of 6.7 ± 0.1. Sorbent doses of 50 g L⁻¹ of BB and 0.25 g L⁻¹ of MBB, AC, and MnG were used. The mixtures were shaken in an incubator shaker at 100 rpm and 30 °C until the equilibration time had been reached (48 h for BB and AC; 72 h for MBB and MnG). Thereafter, the suspensions were filtered through Grade 42 Whatman filter paper and the Mn concentration of the solutions was analyzed by ICP-OES. All experiments were performed in triplicate. The amount of Mn adsorbed onto the adsorbent was calculated using Equation (1):

The isotherm data were then subjected to the most widely used empirical models: the Langmuir, Freundlich [31], and Temkin [32] isotherms and the Dubinin–Raduskevich (D-R) isotherm [33]. The linearized form of the Langmuir model is presented in Equation (6):

$$\frac{1}{q_e}=\frac{1}{q_m}+\left[\frac{1}{K_L q_m}\right]\frac{1}{C_e}$$

(6)

where $q_m$ is the Langmuir maximum adsorption capacity (mg g⁻¹), $C_e$ is Mn concentration at the equilibrium time (mg L⁻¹), and $K_L$ is the Langmuir affinity constant (L mg⁻¹). $K_L$ is an important tool in the calculation of the equilibrium parameter ($R_L$) that explains the favorability of the adsorption process; $R_L$ is calculated using Equation (7):

$$R_L=\frac{1}{\left(1+K_L{C}_0\right)}$$

(7)

The linearized form of the Freundlich model is presented in Equation (8):

$$ln q_e=ln K_F+\frac{1}{n}ln C_e$$

(8)

where $K_F$ is the Freundlich constant (mg g⁻¹) (mg L⁻¹)^1/n and $n$ is a dimensionless constant related to the adsorption intensity. The Temkin model is presented in Equation (9):

$$q_e=\left(\frac{RT}{b}\right)1n(A_T)+\left(\frac{RT}{b}\right)1n\left(C_e\right),$$

(9)

where $A_T$ is the Temkin isotherm equilibrium binding constant (L g⁻¹), $R$ is the universal gas constant (8.314 J mol⁻¹ K⁻¹), $T$ is the temperature (K), and $b$ is a constant related to the heat of sorption (J mol⁻¹).

The Dubinin–Radushkevich isotherm model (D-R), which proposes that pore filling mechanisms explain adsorption, is widely employed to differentiate between chemical and physical adsorption of metal ions. The linear form of the Dubinin–Radushkevich isotherm is defined by Equations (10)–(12):

$$ln q_e=ln q_s-\beta {\epsilon }^2$$

(10)

$$\epsilon =RT ln(1+1/c_e)$$

(11)

$$E=1/\sqrt{2\beta }$$

(12)

where $\beta$ is the Dubinin–Radushkevich constant (mol² kJ⁻²), q_s is the theorical maximum adsorption capacity (mg g¹), and ε describes the Polanyi potential, with R being the universal ideal gas constant (8.314 j mol⁻¹ K⁻¹) and T the absolute temperature (K). From this model, we can determine the mean adsorption energy E (kJ mol⁻¹); when the value of E is <8 kJ mol⁻¹, the process is physical adsorption, and when E is between 8 kJ mol⁻¹ and 16 kJ mol⁻¹, the process is chemical adsorption [34].

2.5. Thermodynamics Calculations

In determining the effect of temperature on Mn adsorption onto adsorbents, thermodynamic parameters provide further information on the inherent energy changes associated with the adsorption process. The most typical equations for thermodynamic parameters include the change in free energy (ΔG°), enthalpy (ΔH°), and entropy (ΔS°). ΔG° is calculated as in Equation (13) [35]:

ΔG° = −RTlnK

(13)

where $R$ is the universal gas constant (8.314 Jmol⁻¹ K⁻¹), $T$ is the temperature (K), and $K$ is the equilibrium constant. If $K_L$ has units of L mg⁻¹, then $K=K_L$ × M_adsorbate × 1000 × 55.5 [35], where $K_L$ is the Langmuir affinity constant (L mg⁻¹), M_adsorbate is the molecular weight of Mn (54.94 g mol⁻¹), and 55.5 is the molar concentration of water. Here, the values of ΔH° and ΔS° were respectively obtained from the slope and the intercept of the plot between lnK and 1/$T$ by considering the following Equation (14):

$$lnK=-\left(\frac{\mathsf{\Delta }{H}^{\circ }}{R}\right)\left(\frac{1}{T}\right)+\left(\frac{\mathsf{\Delta }{S}^{\circ }}{R}\right)$$

(14)

2.6. Chemical Fractionization

Here, 2 g of adsorbent for the pre- and post-adsorption stages were analyzed using sequential extraction techniques [22]. The chemical forms of Mn are classified into six fractions—water soluble (F1), exchangeable (F2), bound to carbonates (F3), bound to iron and manganese oxide (F4), bound to organic matter (F5), and residual (F6)—following the procedures shown in

Table 2. Sequential extraction procedures.

Fraction	Extractable Solvent	Volume (mL)
Water—soluble (F1)	Deionized water	20
Exchangeable (F2)	1M Ammonium acetate, pH 7 (Analytical grade, 98%, Q RëC™, Auckland, New Zealand)	20
Bound to Carbonate (F3)	1M Ammonium acetate, pH 5	20
Bound to Iron and Manganese Oxide (F4)	0.04 M Hydroxylamine hydrochloride in 25%, acetic acid (v v−1), pH 3 (Hydroxylamine hydrochloride, analytical grade, 99%, Q RëC™, Auckland, New Zealand and acetic acid, 100%, MercK, Darmstadt, Germany)	20
Bound to Organic Matter (F5)	30% Hydrogen peroxide (adjust pH = 2, HNO3) (Analytical grade, 30%, Q RëC™, Auckland, New Zealand	15
	+3.2 M Ammonium acetate in 20% HNO3 (v v−1) (Analytical grade, 98%, Q RëC™, Auckland, New Zealand),	5
Residual (F6)	7 M HNO3	20

.

3. Results and Discussion

3.1. Characteristics of Biochar

The external surface structural morphology from the SEM-EDS images shows the porous structure of BB and AC (Figure 1a,b), with BB having longitudinal pores. According to the International Union of Pure and Applied Chemistry (IUPAC) pore size classification, it was found that BB and AC contained mostly macropores (>50 nm) and few mesopores (2–50 nm) and micropores (<2 nm), respectively. The high amount of macropores is suitable for removal of pollutants in liquid phase while providing liquid mass transfer, whereas mesopores provide solute diffusion to the active sites [36]. The pores of BB and AC range from micro- to macropore sizes (10–100 μm) [29]. The vascular bundles of the raw biomass are the origin of the large pores. As shown in Figure 1c, the Mn nanoparticles were successfully doped onto the surface of MBB, which has the same microparticle characteristics as the MnG surface in Figure 1d. The Mn composition on the MBB adsorbent surface increases to 50.07%, which is higher than the value for MnG (14.27%).

The BET surface area of the adsorbents is presented in

Table 3. Characteristics of adsorbents.

Adsorbents	pHDI	pHpzc	Iodine No.	BET Surface Area	Pore Volume	Pore Size
			mg g−1	m2 g−1	cc g−1	Å
AC	7.52	7.43	271.96	806.4	0.470	12.38
BB	8.62	8.10	67.55	176.2	0.143	15.21
MBB	3.70	7.85	69.13	104.2	0.079	15.33
MnG	3.02	4.72	6.14	0.147	0.001	223.10

. BB has a pore volume of 0.143 cc g⁻¹, which correlates with the value of 0.14 cc g⁻¹ reported by Sahoo et al. [37]. MBB has a BET surface area of 104.2 m² g⁻¹, smaller than the original BB (176.2 m² g⁻¹), which is due to Mn filling the pores, as shown in Figure 1c. The EDS spectral data reveal an Mn content of 50.07%. MnG is mainly composed of silicon dioxide, and has a BET surface area of 0.147 m² g⁻¹, smaller than the value reported by Xuwen et al. [28] (8.4 m² g⁻¹). In addition, Outram et al. [38] characterized MnO₂-coated silica (quartz) as having a low surface area (<0.5 m² g⁻¹). Similarly, the iodine numbers for MBB (69.13 mg g⁻¹) and BB (67.55 mg g⁻¹) are higher than that of MnG (6.14 mg g⁻¹), as shown in Table 3; all are lower than the value of 271.96 mg g⁻¹ for AC.

The pH_pzc value is another useful measure; it indicates the pH at which the charge of the adsorbent surface is zero as compared to pH_DI, where pH_DI is the pH value of DI water with adsorbents added. The pH_pzc of the adsorbents ranges from 7.43–8.10 for AC, BB, and MBB. The surface of an adsorbent is positive at pH_DI < pH_pzc, neutral at pH_DI = pH_pzc, and negative at pH_DI > pH_pzc; therefore, the adsorption of Mn(II) increases with an increase in pH due to the electrostatic attraction between the positively charged Mn(II) ions and the negatively charged surface of the adsorbent when pH_DI > pH_pzc. In the pH_DI range of 6 to 9, dissolved Mn(II), a divalent cation, can be removed from solution via sorption onto a solid surface, most often manganese oxide [39]. Mn(II) adsorption onto MnO_x surfaces is rapid [40] and is accompanied by the release of H⁺, as is the case with cation adsorption onto oxide surfaces, shown in Equation (15) [41]:

Mn²⁺(aq) + MnO(OH)₂(s) → MnO₂MnO(s) + 2H⁺(aq)

(15)

The FT-IR spectra of the adsorbents were recorded between 500 and 4000 cm⁻¹ using a spectrometer. As depicted in Figure 2a, b, the FT-IR spectra of BB, AC, MBB, and MnG all display peaks at 3445.19, 3445.41, 3446.34, and 3448.29 cm⁻¹, respectively, indicating the presence of O–H stretching vibrations [18]. BB shows peaks at 1625.26, 875.62, and 571.61 cm⁻¹, attributed to C=C stretching, C–H bending, and C–H bending, respectively [38]. After adsorption, the respective peaks are shifted to wave numbers of 1602.42, 861.52, and 577.91 cm⁻¹. The functional groups of BB and Mn²⁺ ions may have become coordinated to produce the newly observed peak at 489.39 cm⁻¹ [42]. AC shows peaks at 1642.04, 1400.06, and 607.08 cm⁻¹, assigned to C=C stretching into the carbonyl/carboxyl groups, C–H bending, and aromatic C–H bending, respectively; the functional groups and chemical binding after adsorption lead to peaks at 1641.88, 1401.96, and 584.35 cm⁻¹, respectively.

The FT-IR spectra of MBB and MnG in Figure 2c, d show peaks at 1388.72 cm⁻¹ (MBB) and 1399.83 cm⁻¹ (MnG), arising from -OH stretching vibrations, while the peaks at 520.41 and 441.40 cm⁻¹ indicate Mn–O stretching [43]. The spectral peaks are shifted to 3448.26, 1383.97, and 464.02 cm⁻¹for MnG and 3447.97, 1400.09, and 503.22 cm⁻¹ for MBB, indicating the binding of Mn to the O-H group.

Table 4. The elemental composition of adsorbents by WDXRF.

Composition	Weight (% w w−1)
	AC	BB	MBB	MnG
Si	25.200	7.930	3.570	58.80
Ca	21.900	7.780	5.600	1.780
Fe	16.100	0.318	-	2.320
S	12.100	2.870	1.340	12.00
P	8.560	7.060	3.410	0.302
Al	7.270	0.161	0.081	4.260
Cl	2.280	5.620	1.020	-
Ti	1.960	-	-	-
Zn	1.290	0.327	0.111	-
Mg	1.090	2.690	1.750	0.689
Cu	0.895	-	-	-
K	0.654	64.50	24.80	11.20
Ni	0.541	-	-	-
Cr	0.160	-	-	-
Mn	-	0.700	58.30	8.610

presents the elemental composition of the adsorbents. Silica (Si) was the main component of AC, while that of BB was potassium (K). BB contained an Mn content of 0.70%, while Mn was not detected in AC. However, higher levels of other cationic metals were present in AC compared to BB. BB had high contents of K (64.5%), Si (7.93%), and Ca (7.78%), with compositional values comparable to those reported by Liu et al. [44]. MBB had lower metal contents than BB, and contained no iron (Fe) due to the leaching of metals during preparation. MBB had an Mn content of 58.30%, approximately six-fold higher than that of MnG, which is because MnO₂ is able to penetrate the micropores of MBB, whereas MnG does not possess a porous structure.

3.2. Batch Experiments

The contact time significantly influences the removal of metal ions from aqueous solutions. The Mn adsorption capacities (mg g⁻¹) and percentage removal efficiencies at different times for each adsorbent are shown in Figure 3a, b, respectively. Adsorption of Mn by all adsorbents reaches the equilibration time at 72 h. However, MnG shows the maximum adsorption at 24 h and then plateaus, indicating that Mn was rapidly adsorbed onto active sites. This phenomenon corresponds to the study of Köse et al. 2020 [45], and is caused by the collision of adsorbent, implying that chemical adsorption occurs at the active sites. Reaching sorption equilibrium within a short contact time is an indication of the efficiency of a material, and short contact times are preferred for efficient metal removal [46]. We found that when using an adsorbent dosage of 5% w v⁻¹ for BB and AC we could achieve rates of 87.18% and 96.85%, respectively, whereas for MBB and MnG the rates were 51.07% and 31.48%, respectively, with the same adsorbent dosage of 0.03% w v⁻¹. Thus, BB and MBB could be used as low-cost adsorbents for Mn removal to replace AC and MnG, respectively. Furthermore, these results suggest that chemical adsorption may be stronger than physical adsorption thanks to its shorter reaction time [47]. MBB initially exhibits good removal efficiency because of its empty surfaces, which adsorb manganese(IV) oxide (MnO₂). According to Equation (16), Mn²⁺ in groundwater is adsorbed onto the adsorbent’s surface as follows [48]:

Mn²⁺(aq) + MnO₂(s) → MnO₂(s) − Mn²⁺

(16)

Our investigation of contact time revealed that pH_DI is significantly lower than pH_pzc, which is due to positive discharge of Mn²⁺, resulting in lower Mn adsorption.

3.3. Adsorption Kinetics

The PFO and PSO kinetic models were applied to interpret the mechanisms of Mn adsorption onto the adsorbents. The kinetic parameters are shown in

Table 5. Adsorption kinetics parameters.

Adsorbents	${\mathit{q}}_{\mathit{e}\mathit{x}\mathit{p}}$ mg g−1	PFO		R2	PSO		R2	Intraparticle Diffusion
		${\mathit{k}}_1$	${\mathit{q}}_{\mathit{e}}$		${\mathit{k}}_2$	${\mathit{q}}_{\mathit{e}}$		${\mathit{k}}_{\mathit{i}}$	$\mathit{c}$
		h−1	mg g−1		g mg−1h−1	mg g−1		mg g−1 min−1/2
AC	0.088	0.036	23.02	0.796	16.741	0.092	0.999	0.0002	0.0134
BB	0.076	0.063	9.743	0.949	2.725	0.080	0.991	0.0001	0.0062
MBB	4.700	0.019	0.198	0.918	0.015	6.954	0.956	0.0059	0.1165
MnG	2.120	0.006	0.017	0.129	11.085	3.579	0.986	0.0114	0.3019

. In the case of the pseudo-second order model, the theoretical $q_e$ values for all adsorbents are close to the experimental $q_{exp}$ values. The pseudo-second order model was the best fit for the kinetic data of Mn adsorption based on the values of the correlation coefficient (R²) and $q_e$. This implies that the rate-limiting step comprised chemical adsorption involving valence forces by the sharing or exchanging of electrons between the adsorbent and adsorbate [49], revealing a chemisorption relationship generated by both the adsorbent and the solution [50].

Table 5 displays the $k_i$ values, which were derived from the slopes of the linear portions of the curves. A linear line can be seen on the intra-particle diffusion plot, which indicates that intra-particle diffusion has an impact. This shows that diffusion occurs within the adsorbent’s pores and controls adsorption [51]. The highest diffusion rate is represented by the $k_i$ value of 0.0114 mg g⁻¹ min^−1/2 for MnG, followed by MBB at 0.0059 mg g⁻¹ min^−1/2. AC and BB are similar, and have lower diffusion rates. The higher diffusion rates of MnG and MBB may be due to several processes on adsorbents, such as adsorption, oxidation, and precipitation.

3.4. Adsorption Performance and Isotherm Fitting

The linear patterns of the Freundlich and Langmuir models were used to represent the equilibrium adsorption isotherm data.

Table 6. Parameters and constants for selected isotherms.

Adsorbents	Langmuir Isotherm			RL	Freundlich Isotherm		R2	Temkin Isotherm		R2	D-R Isotherm		R2
	${\mathit{q}}_{\mathit{m}}$	${\mathit{K}}_{\mathit{L}}$	R2		${\mathit{K}}_{\mathit{F}}$	$\mathit{n}$		$\mathit{b}$	${\mathit{A}}_{\mathit{T}}$		$\mathit{E}$	$\mathit{\beta }$
	mg g−1	L mg−1			(mg g−1) (mg L−1)1/n			kJ mol−1	L g−1		kJ mol−1	mol2 kJ−2
AC	0.620	0.812	0.959	0.135	0.174	1.912	0.740	0.114	0.8471	0.829	3.774	0.0351	0.945
BB	0.803	0.308	0.990	0.292	0.160	1.313	0.945	0.128	0.8651	0.998	3.026	0.0546	0.998
MBB	21.277	0.378	0.966	0.252	5.384	1.346	0.956	12.946	0.0147	0.835	1.671	0.1790	0.797
MnG	24.450	0.229	0.990	0.357	4.305	1.288	0.980	8.503	0.0178	0.964	1.697	0.1737	0.887

shows the constant parameters and correlation coefficients for each isotherm model. For all adsorbents, the Langmuir model leads to better correlation values (R² > 0.959) than the others. The value of $R_L$ indicates the separation factor, showing that the behavior of Mn²⁺ adsorption is favorable (0 < $R_L$ < 1). The $q_m$ value of BB is slightly higher than that of AC, and the $q_m$ value of MBB is higher than that of MnG. The Temkin plot shows that the adsorption energies of Mn for BB and AC are less than 8 kJ mol⁻¹; thus, adsorption is a physical process in this case. On the other hand, adsorption by MBB and MnG involves energies between 8 and 16 kJ mol⁻¹, indicating that the adsorption process proceeds via ion exchange [48].

Adsorption is assessed via the D-R isotherm model in terms of energy [52]. If $E$ is between 8–16 kJ mol⁻¹, the process is chemical adsorption, while if it is between 1–8 kJ mol⁻¹ it is physical adsorption. In this study, we discovered that a value of $\mathrm{E}$ in the range of 1.67–3.773 kJ mol⁻¹ indicates a physical process for all adsorbents. D-R isotherm descriptions of vapor and gas adsorption on microporous sorbents such as activated carbon and zeolites typically use this model [53]. As a result, the D-R model might not be appropriate for describing the lower R² of Mn adsorption on MBB and MnG [54]. In general, it is less appropriate for metal adsorption.

Table 7. Adsorption capacity of Mn compared to other adsorbents.

Adsorbents	Initial pH	Type of Water	Initial Mn Concentration mg L−1	${\mathit{q}}_{\mathit{m}}$ mg g−1	References
Farmyard manure-derived biochar	6	Mn solution	2	6.652	[15]
Activated carbon modified by nitric acid	5–6	Mn solution	5	10.00	[55]
Activated carbon modified by persulfate	5–6	Mn solution	5	6.660	[55]
Pithecellobium dulce carbon	7.2	Mn solution	5	0.415	[56]
BB	6.78	Groundwater	5	0.803	This study
MBB	6.78	Groundwater	5	21.27	This study

compares the $q_m$ values from the Langmuir adsorption model of BB and MBB with those from other studies. MBB has a higher $q_m$ than other adsorbents when using an initial Mn concentration lower than or equal to 5 mg L⁻¹.

The commercial costs of AC, BB, and MnG in Thailand are within the respective ranges of 2.79–3.52, 0.88, and 5.29–7.34 USD kg⁻¹. The cost of MBB can be estimated from the cost of the chemicals use for preparation, notably the cost of permanganate at 2.35–4.14 USD kg⁻¹. Thus, MBB costs around 2.63 USDkg⁻¹. However, in the present study it was found that MBB can be used in lower amounts than BB, leading to a lower-cost adsorbent.

3.5. Thermodynamic Studies

Thermodynamic parameters are calculated as shown in

Table 8. Calculation of thermodynamic parameters.

Adsorbents	Temperature (K)	Thermodynamics
		ΔG°	ΔH°	ΔS°
		kJ mol−1	kJ mol−1	J mol−1K−1
BB	298	−33.89	52.88	290.41
	303	−34.65
	308	−36.81
AC	298	−33.57	309.40	1148.41
	303	−37.09
	308	−45.10
MBB	298	−35.62	−14.93	68.55
	303	−35.16
	308	−36.32
MnG	298	−37.21	103.63	466.44
	303	−33.90
	308	−42.00

. Adsorption in the case of all adsorbents was a spontaneous process according, as shown by the negative values of ΔG° [57,58]. The positive values of ΔS° demonstrate that the adsorption process increased the randomness at the interface of the solid/solution system [58].

3.6. Mn Fractionization of Adsorbents

The highest content of Mn in the pre-adsorption AC and BB was present in the fractions bound to organic matter (F5, 31.37%) and in those bound to iron and manganese oxide (F4, 25.17%), as shown in Figure 4. After adsorption, the main fraction of Mn was F4 in both AC (32.15%) and BB (26.34%). This finding corresponds to the findings of previous research by Inam et al. [59], who reported that the iron/manganese oxide bound fraction is a possible main source of Mn binding. The binding of Mn with iron and manganese oxide in this study was strong, in agreement with pseudo-second order kinetics data, leading to chemical adsorption. The behavior of Mn in binding to iron/manganese oxide in water is controlled by environmental conditions such as pH changes [60]. After adsorption by AC, the percentage of exchangeable Mn(F2) was above 10%, indicating that there were remaining portions that could be released. Metals exhibit thermodynamic instability in the F2 fraction, as they are sensitive to redox and pH changes [61].

The highest Mn fraction of MnG both pre- and post-adsorption was the F4 fraction at 76.70%. These was due to the main component of sand being oxides of silica and alumina; on the other hand, the main component of MBB is Mn oxide. When MBB contacted Mn ions in the solution, oxidation occurred and Mn was precipitated into its solid form [41]. Thus, the Mn fraction bound to carbonates (F3) increased from 21.25% to 44.6%.

According to the present study, the binding of Mn in the carbonate phase was poor [22], and dependent on the pH_DI of MBB and MnG. The properties and structure of biochar are critical to the immobilization process of the heavy metals on it. The pyrolysis conditions and feedstock type have a significant impact on the properties of biochar [62]. Physical properties such as the surface area, pore volume, and pore size favor the preliminary diffusion of elemental metals through the pores of the biochar surfaces [59]. The functional groups that appear on the surface of the sorbent are shown in Figure 2. Of these, the O-H group is important for binding of Mn to iron and manganese oxide(F4). Research has suggested that ion exchange processes involve manganese ions (Mn²⁺) that attach themselves to adjacent hydroxyl groups [62], resulting in the formation of complexes with OH in groundwater, such as Mn(OH)⁺, Mn(OH)₂, Mn₂(OH)₃⁺, Mn₂OH³⁺, and Mn(OH)₄²⁻ [63].

4. Conclusions

In this study, MBB displayed the maximum Mn adsorption capacity, which was due to adsorption in the biochar pores and Mn oxidation and precipitation on the surfaces. MBB is a low-cost adsorbent material, and can be used in smaller quantities than BB; moreover, it can be used as a substitute for MnG because it is cheaper. However, the use of MBB for the removal of low Mn concentrations in groundwater should be investigated further, as excess Mn from MBB may leach out. Thus, it is recommended to use a small amount of MBB and an up-flow adsorption filter. This could provide the reaction times and entrap the precipitated in the filter, thereby achieving both high removal efficiency and safety. Further research on Mn adsorption with biochar is needed in order to optimize the process and understand the mechanisms involved. Moreover, additional research on use of MBB as an adsorbent should be conducted in groundwater with high manganese concentrations and using column filtration.