Activated Carbons Derived from Different Parts of Corn Plant and Their Ability to Remove Phenoxyacetic Herbicides from Polluted Water

Abstract

In this study, the adsorption of phenoxyacetic acid (PAA) and its chlorinated derivatives, including 4-chlorophenoxyacetic acid (4CPA) and 2,4-dichlorophenoxyacetic acid (2,4-D), on activated carbons (ACs) from corn kernels (AC-K), corn leaves (AC-L), and corn silk (AC-S) were investigated. The adsorption kinetics followed the pseudo-second-order model, and the film diffusion was the rate-limiting step. The adsorption rate increased in the order PAA < 4CPA < 2,4-D and was correlated with the porous structure (mesopore volume) of these ACs. The Langmuir isotherm models best fit the experimental data; PAA was adsorbed least and 2,4-D most preferentially. The observed trend (PAA < 4CPA < 2,4-D) was positively correlated with the molecular weight of the adsorbates and their hydrophobicity while being inversely correlated with their solubility in water. The adsorption for 2,4-D, according to the Langmuir equation, is equal to 2.078, 2.135, and 2.467 mmol/g and S_BET 1600, 1720, and 1965 m²/g, respectively. The results for other herbicides showed a similar correlation. The adsorption of phenoxy herbicides was strongly pH-dependent. The ACs produced from corn biomass can be an eco-friendly choice, offering sustainable products that could be used as efficient adsorbents for removing phenoxyacetic herbicides from water.

1. Introduction

Activated carbons are highly effective and universal adsorbents extensively used in various applications such as removing water pollutants, gas purification, separation or storage, catalysis, and many others [1,2]. Their porous structure, characterized by a sizeable specific surface area and pore volume and the presence of surface functional groups, defines the main properties of carbon adsorbents necessary for this particular application [1]. Although activated carbons are often chosen adsorbents, their widespread use is limited due to their relatively large cost. One of the ways to reduce the cost of producing carbon adsorbents is to use agricultural by-products and waste as low-cost precursors [3,4,5]. On a global scale, there is a growing need to find green and sustainable solutions to environmental issues.

Along with wheat and rice, corn is one of the three most important crops in the world. The United States remains the largest producer of corn. It is followed by China and Brazil. The European Union crop accounts for about six percent of the world production.

Corn is a versatile crop. It can be grown not only for use as a grain but as animal feed. A large number of farmers also grow it for silage. It is also used for the production of furfural [6], bioethanol [7], biogas [8], lactic acid [9], paper pulp [10], and building materials [11]. It also serves as an energy feedstock [12].

Ahmed et al. [13] report that corn residues such as cobs, straws, stalks, stovers, husks, silk fibers, and leaves have been used to obtain carbon adsorbents for purifying water from various chemicals. Cobs (31%), stalks (31%), and straw (28%) have been used most frequently to date, followed by husks (6%), silk (3%), and leaves (1%). These residues are used in crude [14], chemically modified [15], biocarbon [16] and composites [17]. However, the most applied adsorbents from corn residues are activated carbons (ACs) [18,19,20,21,22,23,24]. Producing activated carbons (ACs) from renewable resources helps maintain environmental balance and utilize agricultural waste. Using local residual biomass as a raw material can enhance sustainability. Opting for lignocellulosic materials from nearby sources reduces the need for long-distance transportation, thus lowering the carbon footprint of the entire treatment process. Activated carbons are mainly obtained from corn cobs by chemical activation using KOH, NaOH, H₃PO₄, ZnCl₂, or FeCl₃. These carbons have a strongly developed specific surface area (under 1000 m²/g) and can be successfully used for the removal of pollution from liquid and gas phases [25,26,27,28,29,30].

In contrast, in the present study, activated carbons produced from corn biomass (different parts of a corn plant) have been used as adsorbents to remove herbicides such as phenoxyacetic acid (PAA), 4-chlorophenoxyacetic acid (4CPA), and 2,4-dichlorophenoxyacetic acid (2,4-D) from aqueous solutions. Phenoxyacetic herbicides, especially 2,4-D, are the most popular and widely used herbicides today, next to glyphosate [31,32]. Unfortunately, their widespread use is associated with their frequent occurrence in the environment, which is highly undesirable due to their high toxicity. Among the many techniques for removing these pollutants from water, adsorption, especially on activated carbon, is considered the most popular and effective. This effectiveness largely depends on the physical and chemical properties of the adsorbent used. Therefore, the search for and production of new sustainable carbon materials as potentially effective adsorbents is currently a scientific trends.

The application of carbon adsorbents in the removal of water pollutants is much more effective than the use of other methods due to adsorption’s more convenient operating units in terms of convenience and simple design. On the other hand, as precursors, activated carbons were taking into account agricultural waste because of the much lower cost of carbon adsorbent production in comparison with the commonly used coal or wood methods. Agricultural wastes are considered a good choice, taking into account their availability, low cost, and the economic advantage of their useful exploitation, which also serves as a means of disposal. The method of activated carbons produced from different parts of the corn plant and their physical and chemical characteristics were described and discussed in earlier work [20].

The aim of this work was to attempt to provide a sustainable solution to several environmental protection problems at the same time. The use of waste biomass was chosen to provide raw material for the production of activated carbon, which was then to be used as a low-cost adsorbent to purify water from organochlorine herbicides and their derivatives.

2. Materials and Methods

2.1. Reagents

The herbicide standards—phenoxyacetic acid (≥98%, PAA), 4-chlorophenoxyacetic acid (98%, 4CPA), and 2,4-dichlorophenoxyacetic acid (≥99%, 2,4-D)—were received from Acros Organics (Geel, Belgium). The structural formulas of these compounds and their main physicochemical properties are given in

Table 1. Physicochemical properties of the phenoxyacetic herbicides.

Herbicide	CAS No.	Molecular Formula	Molar Mass (g/mol)	Water Solubility * (g/L)	log P	pKa
PAA	122-59-8		152.15	10	1.48	3.70
4CPA	122-88-3		186.59	0.96	1.85	3.10
2,4-D	94-75-7		221.04	0.89	2.37	2.98

* at 20 °C.

. The other high-purity chemicals and reagents were obtained from Chempur (Piekary, Śląskie, Poland).

2.2. Preparation and Characterization of the Activated Carbons

The raw materials for activated carbon production were separated anatomical parts of the corn plant: corn kernels (AC-K), corn silk (AC-S), and corn leaves (AC-L). ACs were obtained from the above-mentioned various plant parts in a two-step process as follows: pyrolysis/carbonization in an oxygen-free atmosphere by heating to 600 °C (in stable conditions for 1 h) and activation with KOH (mass ratio of 1:4) in an argon atmosphere at 750 °C for 15 min. The obtained activated carbons were washed with 2% HCl acid and then by deionized water to achieve a neutral pH. Detailed conditions are as shown in previous work [20].

The porous structure parameters of obtained ACs were calculated from determined N₂ adsorption–desorption isotherms at 77 K (ASAP 2020, Micromeritics, Norcross, GA, USA). The specific surface areas (S_BET, m²/g) were calculated according to the Brunauer–Emmett–Teller method at relative pressure p/p_o ≈ 0.05–0.2. The total pore volume (V_t, cm³/g) was determined from the adsorption isotherm at relative pressure p/p_o ≈ 0.95. The micropore volume (V_mi, cm³/g) and mesopore volume (V_me, cm³/g) were also determined using the Barrett–Joyner–Halenda and t-plot methods. The average pore diameter D were calculated from the following formula: 4 V_t/S_BET. Scanning electron microscopy (SEM) studies were performed on a ZEISS EVO 40 (Carl Zeiss Microscopy GmbH, Jena, Germany) scanning microscope. Activated carbons were applied to double-sided carbon stickers and then sputtered with gold. To characterize the surface chemistry of prepared activated carbons, the contents of oxygen functional groups, acidic as well as basic, were determined according to Boehm’s method. Additionally, the point of zero charge (pH_PZC) of each activated carbon was determined by the drift method. All the procedures and conditions were as shown in earlier work [33]. Another method of characterizing ACs’ surface chemistry was thermogravimetric analysis using STA 449 F5 Jupiter-QMS of the NETZSCH (Burlington, MA, USA). The conditions were as follows: temperature increase—5 °C/min, helium flow rate—25 mL/min, sample mass—10 mg ± 1 mg.

2.3. Bath Adsorption Experiments

All adsorption studies were performed using the batch method at 23 °C. Into glass Erlenmeyer flasks, 20 mL of each adsorbate solution at the appropriate concentration and 0.01 g of activated carbon (its dose 0.5 g/L) were added. The mixtures were then shaken at 200 rpm for the appropriate time (kinetic studies) or 8 h. The solutions were then filtered and analyzed for herbicide content. The experiments included the study of adsorption kinetics, adsorption isotherms, and the effect of solution pH. For the study of adsorption isotherms, initial herbicide concentrations ranged from 0.5 to 2.0 mmol/L, while adsorption kinetics and the effect of pH were investigated for solutions with a concentration of 1.0 mmol/L. Kinetic and isothermal studies were carried out in solutions at the original (natural) pH (~3). Studies on the effect of pH on adsorption were carried out in the pH range from 2 to 10. Solutions were adjusted to the desired pH by adding small amounts of 0.01 mol/L HCl and/or NaOH, and the pH was monitored using a pH meter (model CP-505, Elmetron, Zabrze, Poland).

Herbicide concentrations in solution were determined using UV-Vis spectrophotometry (Varian Carry 3E series, Palo Alto, CA, USA). Absorbance was measured at analytical wavelengths of 268, 278, and 283 nm for PAA, 4CPA, and 2,4-D, respectively. Calibration curves (generated in the concentration range from 0.1 to 1.25 mmol/L) were linear (R² ≥ 0.997).

The herbicide amount that was adsorbed on the activated carbon after a given time t (q_t, mmol/g) and at equilibrium (q_e, mmol/g) was calculated from the following dependencies:

$$q_t={\displaystyle \frac{(C_0-C_t)V}{m}}$$

(1)

$$q_e={\displaystyle \frac{(C_0-C_e)V}{m}}$$

(2)

where C₀, C_t, and C_e (mmol/L) are the initial concentration of PAA, 4CPA, or 2,4-D, the concentration of the herbicide after time t, and the concentration at equilibrium, respectively; m is the mass of the activated carbon (g); and V is the volume of the solution (L).

All adsorption experiments were performed in duplicate, and the average value was used for further calculations. All calculations were performed using Microsoft Office 365 Excel software (version 2407).

3. Results

3.1. Characteristics of the Activated Carbons

The N₂ adsorption–desorption isotherms on the ACs determined at −196 °C are presented in Figure 1.

The porous structure parameters of the activated carbons AC-K, AC-L, and AC-S that were obtained and calculated based on Figure 1 are listed in

Table 2. Textural properties of the corn biomass-derived activated carbons.

Sample	SBET (m2/g)	Vt (cm3/g)	Vmi (cm3/g)	Vme (cm3/g)	Vmi/Vt	D = 4 Vt/SBET (nm)
AC-K	1965	1.112	0.932	0.180	0.838	2.264
AC-L	1720	1.286	0.869	0.417	0.676	2.989
AC-S	1600	1.144	0.798	0.346	0.697	2.860

.

Figure 2 shows images of activated carbons taken using scanning electron microscopy (SEM).

The surface chemistry parameters of prepared activated carbons, contents of acidic and basic functional groups, and pH_PZC are presented in

Table 3. Chemical properties of the surface of the activated carbons from corn biomass.

Sample	Σ of Acidic Groups (mmol/g)	Σ of Basic Groups (mmol/g)	pHPZC
AC-K	2.15	0.77	6.85
AC-L	1.93	1.83	7.10
AC-S	2.25	0.48	6.60

. The AC-S sample surface has the strongest acidic properties, and the AC-L surface has the weakest acidic properties.

Results of another method of ACs surface chemistry characterization—thermogravimetric analysis—are summarized in

Table 4. Mass loss for activated carbon samples obtained from TG analysis.

Sample	Mass Loss (%) in Temperature Range (°C)
	200–400	400–600	600–750	750–950
AC-K	3.2	5.5	5.7	5.6
AC-L	3.0	2.3	4.8	5.5
AC-S	3.6	5.6	5.2	5.4

as mass loss values in temperature ranges corresponding to the thermal decomposition of specific types of functional groups.

The highest mass losses in the temperature ranges of 200–400 °C, 400–600 °C, and 600–750 °C can be observed for the AC-S sample, and the smallest can be observed for AC-L. This applies to the thermal decomposition of functional groups such as carboxyl, lactone, or hydroxyl. This confirms the data given in Table 3.

3.2. Adsorption Studies

3.2.1. Adsorption Kinetics

The adsorption kinetics of PAA, 4CPA, and 2,4-D on the three activated carbons tested are shown in Figure 3 as a plot of q_t versus time. Adsorption was fast during the first 30 min and slowed down, reaching equilibrium after about 2 h.

The experimental data have been described by various kinetic models, which can help interpret the results obtained in more detail. Two of the most popular kinetic models [34] were used—pseudo-first-order (PFO) and pseudo-second-order (PSO), whose linear forms can be expressed as follows:

$$\log (q_{\mathrm{e}}-q_{\mathrm{t}})=\log q_{\mathrm{e}}-{\displaystyle \frac{k_1}{2.303}}t$$

(3)

$${\displaystyle \frac{t}{q_{\mathrm{t}}}}={\displaystyle \frac{1}{k_2{q}_{\mathrm{e}}^2}}+{\displaystyle \frac{1}{q_{\mathrm{e}}}} t$$

(4)

where k₁ is the PFO adsorption rate constant (1/min) and k₂ is the adsorption rate constant obtained from the PSO model (g/mmol∙min).

The constants k₁ and k₂ were calculated from the slope and intercept of the linear relationships log (q_e − q_t) = f(t) and t/q_t = f(t), respectively. The results are presented in

Table 5. The PFO and PSO kinetic modeling data for the adsorption of phenoxyacetic acids on corn biomass-derived ACs.

Kinetic Model	Adsorbate
	PAA	4CPA	2,4-D
AC-K
qe (exp) (mmol/g)	0.930	1.197	1.441
pseudo-first-order
k1 (1/min)	0.0207	0.0182	0.0318
qe1 (cal) (mmol/g)	0.894	0.895	0.813
R2	0.991	0.974	0.976
χ2	0.188	0.163	0.417
pseudo-second-order
k2 (g/mmol∙min)	0.0219	0.0292	0.0353
qe2 (cal) (mmol/g)	1.093	1.318	1.556
R2	0.996	0.998	0.998
χ2	0.021	0.016	0.015
AC-L
qe (exp) (mmol/g)	0.844	1.077	1.284
pseudo-first-order
k1 (1/min)	0.2140	0.0193	0.0246
qe1 (cal) (mmol/g)	0.634	0.683	0.834
R2	0.936	0.961	0.978
χ2	0.388	0.596	0.487
pseudo-second-order
k2 (g/mmol∙min)	0.0424	0.0496	0.0554
qe2 (cal) (mmol/g)	0.933	1.151	1.350
R2	0.997	0.999	0.999
χ2	0.017	0.008	0.003
AC-S
qe (exp) (mmol/g)	0.809	0.975	1.224
pseudo-first-order
k1 (1/min)	0.0205	0.0209	0.0222
qe1 (cal) (mmol/g)	0.624	0.6094	0.831
R2	0.976	0.944	0.954
χ2	0.205	0.406	0.398
pseudo-second-order
k2 (g/mmol∙min)	0.0493	0.0563	0.0605
qe2 (cal) (mmol/g)	0.883	1.043	1.286
R2	0.998	0.998	0.999
χ2	0.006	0.014	0.004

. The coefficient of determination (R²) and chi-square test values were used to assess the suitability of a given model. Higher R² values (closer to unity) and lower χ² values indicate a better fit of the applied theoretical model to the experimental data description. Both parameters were determined using the following equations:

$$R^2={\displaystyle \frac{{\sum }_{i=1}^n{(q_{e\left(\mathrm{c}\mathrm{a}\mathrm{l}\right)}-\overline{q_{e\left(\mathrm{e}\mathrm{x}\mathrm{p}\right)}})}^2}{{\sum }_{i=1}^n{(q_{e\left(\mathrm{c}\mathrm{a}\mathrm{l}\right)}-\overline{q_{e\left(\mathrm{e}\mathrm{x}\mathrm{p}\right)}})}^2+{\sum }_{i=1}^n{(q_{e\left(\mathrm{c}\mathrm{a}\mathrm{l}\right)}-q_{e\left(\mathrm{e}\mathrm{x}\mathrm{p}\right)})}^2}}$$

(5)

$${\chi }^2={\sum }_{i=1}^n{\displaystyle \frac{{(q_{e\left(\mathrm{e}\mathrm{x}\mathrm{p}\right)}-q_{e\left(\mathrm{c}\mathrm{a}\mathrm{l}\right)})}^2}{q_{e\left(\mathrm{c}\mathrm{a}\mathrm{l}\right)}}}.$$

(6)

where q_e(exp) and q_e(cal) are the experimental and theoretical values predicted by the kinetic model, respectively.

The adsorption process consists of several steps [34]:

The transport of the adsorbate molecules from the solution to the boundary layer surrounding the adsorbent’s particle.

The transfer of adsorbate molecules through the boundary layer to the surface of the adsorbent (film diffusion).

The transfer of adsorbate molecules inside the adsorbent and into its pores (macro-, meso-, and micropore diffusion).

Binding the adsorbate molecules to the active sites of the adsorbent.

The first and last stages occur very rapidly and are not thought to be significant in affecting adsorption. The slowest occurring step is the one that determines the rate of the overall adsorption process. This is, therefore, either the film diffusion or the intraparticle diffusion.

PFO and PSO models describe the rate of the overall adsorption process but do not give any information on the adsorption mechanism. Therefore, to know the mechanism of herbicide adsorption on corn-derived ACs and to identify the limiting step of the whole process, different kinetic diffusion models (mass transfer models) were used [34].

The intraparticle diffusion model, also known as the Weber–Morris model, is expressed by the following formula:

$$q_{\mathrm{t}}=k_{\mathrm{i}}{t}^{0.5}+C_{\mathrm{i}}$$

(7)

where k_i is the intraparticle diffusion rate constant (mmol/g·min^−0.5) and C_i is the model constant.

The plots of q_t versus t^0.5 are shown in Figure 4. This model assumes that if the relationship q_t = f(t^0.5) is linear over the whole range and additionally passes through the origin, then the intraparticle diffusion is the only rate-limiting step. However, in Figure 4, it can be seen that this relationship is not linear (two stages can be clearly identified). Furthermore, none of the lines pass through the origin. A nonlinearity and intercept (C_i) different from zero indicate that the adsorption rate is affected by both steps (both intraparticle diffusion as well as film diffusion) and that pore diffusion is not a rate-limiting step.

The Bangham model can be presented as a relation as follows:

$$\log \left(\mathrm{l}\mathrm{o}\mathrm{g}{\displaystyle \frac{C_0}{C_0-q_{t}M}}\right)=\log \left({\displaystyle \frac{k_{0}M}{2.303V}}\right)+\alpha \log t$$

(8)

where k₀ and α are Bangham’s model constants and M is the mass of the adsorbent in 1 L of the adsorbate solution (g/L).

The Bangham model plots for the adsorption of the herbicides on corn-derived ACs are presented in Figure 5. The plot of $\log \left(\mathrm{l}\mathrm{o}\mathrm{g}{\displaystyle \frac{C_0}{C_0-q_{t}M}}\right)$ versus log t gives a straight line if the adsorption process is controlled by pore diffusion. However, the plots are nonlinear, as can be seen in Figure 5, confirming the conclusions from the Weber–Moris model that intraparticle diffusion plays a secondary role and is not the step that controls the rate of the whole process.

3.2.2. Adsorption Isotherms

The adsorption isotherms of PAA, 4CPA, and 2,4-D from water on ACs prepared from corn biomass are depicted in Figure 6.

Two of the most popular models, the Langmuir and Freundlich [35], were used to describe the experimental isotherms.

The Langmuir isotherm assumes that the adsorbate can form a monolayer on the adsorbent surface, that the adsorbent surface is energetically homogeneous, and that lateral interactions between adsorbed molecules of the adsorbent are ignored [35]. Langmuir equations in general form (9) and in linear form (10) are as follows:

$$q_e={\displaystyle \frac{q_m{K}_L{C}_e}{1+K_L{C}_e}}$$

(9)

$${\displaystyle \frac{C_{\mathrm{e}}}{q_{\mathrm{e}}}}={\displaystyle \frac{1}{q_{\mathrm{m}}}}{C}_{\mathrm{e}}+{\displaystyle \frac{1}{q_{\mathrm{m}}{K}_{\mathrm{L}}}}$$

(10)

where q_m is the Langmuir’s maximum adsorption capacity (mmol/g) and K_L (L/mmol) is the Langmuir constant.

Both q_m and K_L values can be determined from the linear relationship C_e/q_e as a function of C_e.

The Freundlich model describes adsorption on heterogeneous surfaces and assumes that multilayer adsorption can occur and that adsorbate molecules can interact with each other [35]. The following formula can express the following general Freundlich equation:

$$q_e=K_F{C}_e^{1/n}$$

(11)

where K_F is the Freundlich constant ((mmol/g) (L/mmol)^1/n) and 1/n is the Freundlich exponent.

Equation (11) can be converted to the following linear form:

$$\ln q_e={\mathrm{l}\mathrm{n} K}_{\mathrm{F}}+{\displaystyle \frac{1}{\mathrm{n}}}{\mathrm{l}\mathrm{n} C}_{\mathrm{e}}$$

(12)

The K_F and n isotherm parameters were calculated from the slope and intercept of the linear plot of ln q_e versus ln C_e. The parameters of the Langmuir and Freundlich equations that were obtained are summarized in

Table 6. The adsorption isotherm constants for the adsorption of phenoxyacetic acids on corn biomass-derived ACs.

Isotherm Model	Adsorbate
	PAA	4CPA	2,4-D
AC-K
Langmuir
qm (mmol/g)	1.891	2.188	2.467
KL (L/mmol)	1.580	2.288	7.001
R2	0.993	0.998	0.992
χ2	0.009	0.011	0.007
ΔG0 (kJ/mol)	−28.0	−28.9	−31.7
RL	0.209–0.514	0.186–0.477	0.188–0.447
Freundlich
KF ((mmol/g)(L/mmol)1/n)	1.230	1.576	2.403
1/n	0.613	0.504	0.356
R2	0.981	0.968	0.922
χ2	0.123	0.235	0.152
AC-L
Langmuir
qm (mmol/g)	1.763	1.911	2.135
KL (L/mmol)	0.886	1.355	2.937
R2	0.995	0.995	0.994
χ2	0.008	0.017	0.055
ΔG0 (kJ/mol)	−29.2	−27.6	−29.6
RL	0.221–0.531	0.207–0.511	0.211–0.484
Freundlich
KF ((mmol/g)(L/mmol)1/n)	0.886	1.114	1.860
1/n	0.706	0.589	0.483
R2	0.993	0.980	0.958
χ2	0.381	0.133	0.251
AC-S
Langmuir
qm (mmol/g)	1.559	1.822	2.078
KL (L/mmol)	1.203	3.033	4.187
R2	0.993	0.993	0.995
χ2	0.029	0.030	0.031
ΔG0 (kJ/mol)	−30.0	−29.6	−30.4
RL	0.243–0.562	0.215–0.523	0.216–0.490
Freundlich
KF ((mmol/g)(L/mmol)1/n)	0.821	1.020	1.713
1/n	0.658	0.548	0.411
R2	0.991	0.934	0.935
χ2	0.157	0.815	0.711

. The suitability of the isotherm equations to the experimental data were assessed by comparing the determination coefficients and the chi-square values. As in the case of kinetic studies, higher R² values and lower χ² values indicate a better fit of the isotherm model to the experimental data. From the data analysis, it can be concluded that both isotherm models used described the adsorption of phenoxyacetic acids on corn biomass-derived Acs quite well. However, slightly higher R² values were observed for the Langmuir model (≥9.992) than for the Freundlich model (from 0.922 to 0.993). Lower χ² values were also found for the Langmuir model, indicating that the Langmuir equation better describes the adsorption. The assumptions of this model suggest that the adsorption is monolayer and that the adsorbent surface is homogeneous.

The Langmuir equation can also be used to determine the nature of the adsorption process. Using the K_L constant, the Gibbs free energy of change (ΔG^o) can be determined according to the following formula [36]:

$$∆G^o=-RT\mathrm{l}\mathrm{n}\left(55.5K_L\right)$$

(13)

where R is the universal gas constant (8.314 J/mol·K), T is the temperature in Kelvin (296.15 K), and K_L is the Langmuir equation constant.

The calculated ΔG^o values are shown in Table 6. The Gibbs free energy of change is a measure of the spontaneity and feasibility of an adsorption process. A negative ΔG^o value confirms a spontaneous process, whereas a positive ΔG^o value indicates a non-spontaneous process. As shown in Table 6, the ΔG^o values varied from −27.6 to −31.7 kJ/mol, indicating the spontaneous nature of the adsorption.

The K_L constant of the Langmuir equation can be used to determine the so-called separation factor (R_L), which provides information on the favorability and nature of the adsorption process:

$$R_L={\displaystyle \frac{1}{1+K_L{C}_0}}$$

(14)

If R_L is equal to 0, the adsorption is irreversible; if 0 < R_L < 1, the adsorption is favorable; if R_L is equal to 1, the adsorption is linear; and finally, a value of R_L > 1 indicates an unfavorable character of adsorption [35]. The values of the calculated R_L parameters are presented in Table 6 and, as can be seen, are in each case greater than 0 but less than 1 (1 > R_L > 0), indicating the favorable nature of the adsorption process. The favorable nature of the adsorption is also indicated by the values of the parameter 1/n calculated from the Freundlich equation. It is assumed that the adsorption is unfavorable if 1/n is greater than 1, irreversible if 1/n is equal to 1, and favorable if 1/n is between 0 and 1 (0 < 1/n < 1) [35]. The values of 1/n (0.562 > 1/n > 0.186) shown in Table 6 support the favorable nature of the adsorption process.

3.2.3. Effect of Solution pH

The pH of the solution is an important parameter affecting adsorption as it determines the charge that accumulates on the adsorbent surface and the form in which the adsorbate exists (dissociated or not). This defines the nature of the electrostatic adsorbent–adsorbate interactions (repulsive or attractive). The effect of the solution pH on the adsorption of phenoxyacetic acids on the ACs was studied using pH ranges from 2 to 10, and the results are presented in Figure 7.

Adsorption was most efficient in an acidic environment (at pH 2) and then decreased rapidly with increasing pH from 3 to about 7. Further increases in pH did not result in changes in adsorption, which stayed more or less constant in the range from about 7 to 10. The behavior of all three phenoxyacetic acids is almost identical. This is not surprising as their physicochemical properties, especially pK_a, are comparable. The value of the pK_a (2.98–3.70) determines the form in which the adsorbate will exist in the solution. Thus, at pH < pK_a, the adsorbate molecules will be undissociated, and at pH > pK_a, the adsorbate will be in a dissociated, anionic form. The pH of the solution also affects the surface charge of the adsorbent. The values of the point of zero charge (pH_pzc) for ACs range from 6.6 to 7.1 (Table 3). In an acidic environment, if the pH of the solution is below pH_pzc, a positive charge will accumulate on the surface of the adsorbent. The adsorbent surface becomes negatively charged in a solution with pH > pH_pzc.

4. Discussion

Studies of the porous structure (Table 2) showed that the value of every parameter, including S_BET, V_t, V_mi, and V_me, varies significantly for each of the three tested activated carbons. The order of values is consistent for S_BET and V_mi (maximum value for AC-K and minimum value for AC-S). In the case of other porous structure parameters V_t and V_me, the order of their values is different (maximum value for AC-L and minimum value for AC-K). All ACs were characterized by a high specific surface area S_BET (above 1600 m²/g) and a high total pore volume V_t (above 1.11 cm³/g). The AC-K had the strongest developed S_BET surface area. The corn biomass yielded carbon with varying proportions of micro- and mesopores. AC-K had the highest proportion of micropores, while AC-L had the lowest.

The presence of an extensive pore system were confirmed by SEM analysis (Figure 2). In the case of AC-K, most of the grains are up to 20 μm in size, while some of them are 100–150 μm in size. The shape of the grains is angular, their edges are sharp, and their surfaces are smooth. In the case of AC-L, most of the grains are up to 20 μm in size, while some up to 100 μm in size. Their shape is irregular, their edges are jagged, and their surfaces are irregular. In the case of AC-S, the grains are large, 100–300 μm in size, and finer grains are 50–100 μm in size. Their structure is porous, with large pores with a diameter of approx. 10 μm, and their surfaces are smooth.

Based on the surface chemistry parameters of the prepared activated carbons presented in Table 3 and Table 4, it can be noted that the use of various anatomical parts of corn as a precursor produces activated carbons with very diverse properties. Comparing the properties of the obtained activated carbons with those obtained in other works is quite difficult because they generally used one of the anatomical parts of the plant as a precursor, e.g., silk, leaves, or cobs. One activating factor was most often used there, e.g., CO₂ or KOH. Also, the conditions (temperature or time) were different.

Through analyzing the adsorption kinetic data presented in Table 5, it can be seen that significantly higher R² values (≥0.996), as well as lower χ² values (≤0.0175), were obtained for the pseudo-second-order model. A higher agreement between the theoretically calculated adsorption capacity (q_e2(cal)) and the experimental value (q_e(exp)) was also observed for the PSO model. These results indicate that PAA, 4CPA, and 2,4-D adsorption kinetics on all three corn-derived activated carbons follow the pseudo-second-order model. This is in general agreement with the literature data. As shown in reviews [5,31,32], the adsorption of phenoxyacetic herbicides on activated carbons (but also other non-carbon adsorbents) mostly followed the PSO model.

When comparing the adsorption rates of the individual herbicides on ACs, it can be observed that the k₂ values increase in the order PAA < 4-CPA < 2,4-D. In previous studies on the adsorption of these three phenoxyacetic acids on lignite [34], the opposite trend in the adsorption rate was observed (2,4-D < 4-CPA < PAA). This indicates that the adsorption rate of a specific adsorbate is an individual matter and depends on the type and physicochemical properties of the adsorbent used. The surface chemistry of the adsorbent and its porous structure determine the rate and adsorption efficiency. All three herbicides were adsorbed slowest on AC-K and fastest on AC-S (AC-K < AC-L < AC-S). This observed tendency seems to be closely related to the porous structure of these ACs (Table 2). The adsorption rate increases with the increased mesopore volume (V_me) of the activated carbons. The mesopores are mainly responsible for the transport of the adsorbate molecules into the micropores, and, in general, their greater number (greater volume and greater proportion of the total porosity) accelerates the adsorption process.

The intraparticle diffusion models, including the Weber–Morris and Bangham, suggest that the adsorption rates of PAA, 4CPA, and 2,4-D on ACs obtained from corn biomass are influenced by both pore diffusion and film diffusion, but the diffusion in the film is the rate-limiting step that controls the whole adsorption process.

The experimental adsorption isotherms were described using the Langmuir and Freundlich [35] models, and it was found that the Langmuir isotherm described the equilibrium data better than the Freundlich isotherm model. The adsorption capacity determined from the Langmuir and Freundlich equations makes it possible to compare the adsorption efficiency of each adsorbate on the adsorbent surface and also to compare the adsorption capacities of the adsorbents used. Thus, by analyzing the data in Table 6 (both q_m and K_F values), it can be seen that all three herbicides used were adsorbed best on the AC-K and least on the AC-S. The adsorption capacity of these activated carbons towards these adsorbates increases in the order AC-S < AC-L < AC-K and, as in the kinetic studies, seems to be closely correlated with the porous structure of the ACs used (Table 2). The adsorption efficiency increases with increasing values of the BET surface area as well as the micropore volume of the studied activated carbons.

A comparison of the adsorption of phenoxyacetic acids on the ACs shows that PAA was adsorbed least and 2,4-D most preferentially. This observed order (PAA < 4CPA < 2,4-D) appears to be positively correlated with the molecular weight of the adsorbates and their hydrophobicity while inversely correlated with their solubility in water (Table 1). The adsorption efficiency increased with the number of chlorine atoms attached to the aromatic ring of the molecule. The Cl substituents reduce the electron density of the benzene ring, which increases the molecule’s hydrophobicity and reduces its solubility in water. The more hydrophobic compound has more affinity for the hydrophobic surface of the activated carbon. Indeed, this is the case; the best-adsorbed herbicide was 2,4-D, which is the most hydrophobic compound—it is the most poorly soluble in water and has the highest octanol–water partition coefficient (2.37). In contrast, the least adsorbed compound was the most hydrophilic one (PAA), which is the most soluble in water and has the lowest log p value (1.48). These findings are in general agreement with the results reported in other papers. The same adsorption sequence (PAA < 4CPA < 2,4-D) was observed on goethite [37], lignite [33], and various commercial activated carbons [38]. However, it should be noted that among these phenoxyacetic herbicides, 2,4-D is by far the most commonly studied, while the other two (PAA and 4CPA) are rarely investigated. For this reason, the available comparative material is inadequate.

Based on numerous literature data [5,31,32], it is known that the mechanism of 2,4-D adsorption on the surface of activated carbons includes van der Waals forces, electrostatic interactions, n-π and π-π interactions, and hydrogen bond formation. Since the physicochemical properties of PAA and 4CPA are very similar to those of 2,4-D, it can be speculated that their adsorption mechanism is analogous.

Table 7. Comparison of PAA, 4CPA, and 2,4-D adsorption on various adsorbents.

Adsorbent	SBET (m2/g)	Adsorption Capacity (mmol/g)			Ref.
		2,4-D	4CPA	PAA
AC from corn kernels (AC-K)	1965	2.467	2.188	1.891	this paper
AC from corn leaves (AC-L)	1720	2.135	1.911	1.763	this paper
AC from corn silk (AC-S)	1600	2.078	1.822	1.559	this paper
raw lignite	0.91	0.035	0.027	0.020	[33]
SX2 AC (Norit)	870	1.211	1.135	1.098	[38]
F-400 AC (Chemviron)	995	1.596	1.475	1.349	[38]
Corncob biochar	298	0.170	-	-	[39]
Corn stalk biochar (BC)	523	0.039	-	-	[40]
AC from corncob	1274	1.358	-	-	[26]
AC from coconut shell	991	1.054	-	-	[41]
AC from coconut endocarp	1068	1.065	-	-	[41]
AC from sugarcane bagasse	547	0.696	-	-	[41]
AC from willow (AC-W)	1280	2.310	-	-	[42]
AC from hemp shives (AC-H)	1324	2.446	-	-	[42]
AC from miscanthus (AC-M)	1420	2.577	-	-	[42]
AC from flax (AC-F)	1587	2.682	-	-	[42]

shows the adsorption capacities of various activated carbons and other carbon materials reported in the literature as adsorbents for the removal of PAA, 4CPA, and 2,4-D from water. Most of the adsorbents produced from corn biomass available in the literature were cited. A comprehensive and more detailed comparison of different adsorbents can be found in recently published reviews [5,31,32].

The pH dependence of adsorption observed in Figure 7 indicates that the process is most efficient at pH 2, where the interaction of undissociated herbicide molecules with the positively charged AC surface is most favored. Above pK_a, more and more herbicide molecules become anions, resulting in a gradual decrease in adsorption efficiency. Finally, in the pH range from about 7 to 10, adsorption does not change and stabilizes at a low level. The relatively low adsorption efficiency observed in this range results from repulsive electrostatic interactions between the herbicide anions and the negatively charged surface of the activated carbons. A similar effect of pH value on the adsorption of 2,4-D was reported on lignite [33], carbon black [43], and other activated carbons [42,44].

5. Conclusions

This paper describes the adsorptive removal of three phenoxyacetic acids, including PAA, 4CPA, and 2,4-D, from their aqueous solutions using activated carbons produced from corn biomass (corn kernels (AC-K), corn leaves (AC-L), and corn silk (AC-S)). Adsorption kinetics, adsorption isotherms, as well as the effect of solution pH were investigated. The kinetics were fitted with the pseudo-first-order, pseudo-second-order, Weber–Morris, and Bangham models. It was found that the adsorption kinetics of the herbicides was better represented by the PSO equation and was controlled by film diffusion. The equilibrium adsorption data were fitted with the Langmuir and Freundlich equations, and results showed that it was slightly better described by the Langmuir isotherm model. The adsorption of the herbicides on all the ACs increased in the order PAA < 4CPA < 2,4-D, which correlates well with the respective increase in the hydrophobicity of the adsorbates. Adsorption kinetics as well as adsorption at equilibrium were closely correlated with the porous structure parameters of the activated carbons. Adsorption was pH-dependent and was favored in the acidic environment (pH 2). This study showed that the activated carbons prepared from corn biomass have high surface areas and great adsorption capacities, seeming to be very effective adsorbents for the removal of phenoxyacetic herbicides from water.