Surfactant modified tectosilicates and phyllosilicates for 2,4-D removal and slow release formulation

Abstract

The removal performance of 2,4-dichlorophenoxyacetic acid (2,4-D) in aqueous solution was investigated by an adsorption process on surfactant modified silicates. Surfactant loaded tectosilicates (clinoptilolite and zeolite Y) and phyllosilicates (bentonite and montmorillonite) have been evaluated as potential adsorption media for this purpose. Complete structural characterizations of surfactant loaded materials were performed by infrared spectroscopy and electron microscopy. Surface modification of selected adsorbents by loading surfactants results in a corresponding increase in the removal of 2,4-D from water. The adsorption data fitted well with the Langmuir adsorption isotherm. Based on the Langmuir adsorption model, the calculated maximum monolayer 2,4-D adsorption capacity values for hexadecyltrimethylammonium ion and dioctadecyldimethylammonium ion loaded montmorillonite (HLM and DLM) at 303 K are found to be 158.73 and 161.29 mg g⁻¹, respectively. The slow release studies have also been performed by a thin layer-funnel analytical test and by a soil column percolating system. The adsorption and desorption results obtained from this reported study prove the surfactant loaded silicates as potential adsorbents for 2,4-D (anionic herbicide) from aqueous solution and can be used as slow release herbicide medium.

---

Introduction

Low cost and good selectivity of 2,4-D make it one of the most widely used herbicides in the world which is most commonly used to control dandelions and other broadleaf weeds in both terrestrial and aquatic environments such as lawns, rangeland, grass hayfields, road sides, industrial areas, in agriculture and forestry. Agricultural uses include pasture land, wheat, corn, soybeans, barley, rice, oats, and sugar cane. It is also used for the control of aquatic weeds, some woody vegetations, and site preparation, conifer release in forests. In addition, 2,4-D is also used as an active ingredient in several formulations of herbicides recommended for the control of broadleaf weeds. As only a portion (∼5%) of the herbicides reaches the target weeds, the rest of these and their deadly residues move through the environment such as air, water and soil, finding their way in living tissues where they can bio-accumulate up in food chains into human diet. Because of the widespread use of 2,4-D, it is frequently detected in the environment in spite of its highly toxic and carcinogenic nature.¹ There is numerous evidence indicating its major health effects, from cancer to immunosuppression, reproductive damage to neurotoxicity. In aqueous media, 2,4-D predominantly exists in an anionic form which makes it of great concern as anionic contaminant in water, weakly retained by most soil and subsurface materials.^2–4 Therefore, the efficient removal of anionic herbicides from contaminated water is of great concern. Considerable research efforts have been made by a number of researchers to reduce the content of 2,4-D and organic compounds in water by physical, chemical and biological means either through retention or degradation.^5–10 Among these, adsorption has been proven to be an effective and attractive mechanism. Adsorption of organic pollutants on surfactant modified silicates and oxides is considered an economically and technically feasible remediation approach.^11–13 Boyd and co-workers^14,15 demonstrated in their studies that on a unit-weight basis surfactants or surfactant-modified clays are 10 to 30 times more effective than natural soil organic matter for sorbing nonpolar or weakly polar organic contaminants.

Organo-clays (surfactant-modified clays) have shown to be good sorbents for removing hydrophobic organic chemicals and polar herbicides from water. Hydrophobic and polar interactions between herbicides and alkyl chains of organo-clays and mineral surfaces, respectively, can reduce herbicide mobility in soil. Zhang et al.¹⁶ found that surface modification of paraquat-loaded zeolites encapsulates the guest molecules in the zeolite cages and release of paraquat by ion-exchange with sodium. Interactions of acidic herbicides bentazon and dicamba with organo-clays (octadecylammonium and hexadecyltrimethylammonium exchanged Arizona montmorillonite) has been studied by Carrizosa et al.¹⁷ According to their studies, hydrogen bonding between herbicides and a monosubstituted amino group of surfactant reinforces the strength of the hydrophobic interactions. A slow release formulation of the anionic herbicides sulfosulfuron was prepared by incorporating it in micelles of organic cation octadecyltrimethyl ammonium which adsorbs on the clay–mineral montmorilonite.¹⁸ El Nahhal et al.¹⁹ designed clay based formulations for hydrophobic herbicides in order to control environmental pollution. The primary alkylammonium saturated clays are suggested as suitable potential sorbents for slow release formulation which could be used to decrease bentazone loss by leaching and runoff.²⁰

The present study focuses on the evaluation of adsorption potential of surfactant modified tectosilicates (clinoptilolite and zeolite Y) and phyllosilicates (bentonite and montmorillonite) for 2,4-D in aqueous solution. Adsorption of 2,4-D with variation of surfactant loading, adsorbent dosage, pH of solution, humic acid concentration, equilibration time and temperature was examined, and subsequently, the equilibrium data were fitted to the Langmuir adsorption isotherm. Further, a thin layer analytical method and a soil column method were used for slow release studies.

Experimental

Materials

Four materials belonging to two different silicate families (tectosilicates and phyllosilicates) have been used for this study. Natural zeolite clinoptilolite and synthetic zeolite Y (sodium form) belonging to the tectosilicate family were obtained from HiMedia Laboratories Pvt. Ltd., India. Bentonite and montmorillonite, the members of phyllosilicates was supplied by Loba Chemie Pvt. Ltd., India and HiMedia Laboratories Pvt. Ltd., India. The surfactants selected for this study i.e. hexadecyl trimethylammonium bromide (CH₃(CH₂)₁₅N(Br)(CH₃)₃, HDTMAB) and dioctadecyldimethylammonium bromide ((CH₃(CH₂)₁₇)₂N(Br)(CH₃)₂, DODMAB) were obtained from Sigma-Aldrich whereas 2,4-dichlorophenoxyacetic acid (Cl₂C₆H₃OCH₂CO₂H, 2,4-D) was received from E Merck, India Limited. All these materials and chemicals were used without any further purification.

Surfactant loading on tecto- and phyllo-silicates

A known quantity (200 mg) of washed silicate samples (clinoptilolite, zeolite Y, bentonite and montmorillonite) was mixed with 100 mL of HDTMAB and DODMAB solutions (250, 500, 750 and 1000 mg L⁻¹) prepared in 1 : 1 ratio (v/v) ethanol–deionized water (DIW) solution. The water–alcohol solution has been used for improving the solubility of surfactants in water. The resulting suspensions are shaken at room temperature for 24 h, then centrifuged, washed with ethanol–DIW solution (50 : 50) until no Br⁻ has been detected by AgNO₃ test, and finally freeze dried. The residual surfactant concentration in the supernatant was determined spectrophotometrically and the loaded amount of the surfactant has been estimated from the difference of surfactant concentration in the aqueous phase before and after loading.²¹ Hexadecyltrimethylammonium ion loaded clinoptilolite, zeolite Y, bentonite and montmorillonite materials are classified as HLC, HLY, HLB and HLM, respectively. On the other hand, dioctadecyldimethylammonium ion loaded clinoptilolite, zeolite Y, bentonite and montmorillonite materials are named as DLC, DLY, DLB and DLM, respectively.

Characterization

Infrared spectra of unmodified and surfactant modified materials were obtained using a Fourier transform infrared (FT-IR) spectrophotometer (Thermo NICOLET Nexus spectrometer) over a range of 4000–400 cm⁻¹ at room temperature. In addition, surface morphological studies of tecto- and phyllosilicate particles were made using a scanning electron microscope (SEM, JEOL JSM-6390LV). The UV-Vis spectrophotometer (Systronics 114) was used to calculate the concentration of 2,4-D and surfactant in the aqueous phase. The cation exchange capacity (CEC) was determined for each material by treating 10 mL of its suspension (100 g L⁻¹) with 30 mL of 1 N CaCl₂ in three successive 10 mL washings. The excess salt was removed by washing the materials several times with deionized water until excess CaCl₂ was removed as indicated by a negative AgNO₃ test for chloride ion in the last three washings. The loaded Ca²⁺ cation on the adsorbents was then exchanged with Na⁺ cation by four washings with 10 mL portions of 1 N NaCl. The extracted solution thus produced was brought to a final volume of 100 mL with distilled water and its Ca²⁺ cation content was measured by ethylenediaminetetraacetic acid (EDTA) titration.

Batch adsorption studies

Due to the reliability and simplicity of the batch model, adsorption studies were conducted by a batch technique at 303 K. In a typical experiment, 200 mg of adsorbent material was placed in 150 mL conical flask containing 50 mL of 2,4-D solution of known concentration (552 mg L⁻¹) and pH 7. This mixture was then placed in a reciprocating thermostatic water bath shaker for 24 h. Solid and solution phases were separated by filtration using Whatman filter paper 42 followed by high speed centrifugation of filtrate to restrict the interference of any suspended nano particles. 2,4-D concentration in the filtrate was determined spectrophotometrically by measuring the absorbance at λ_max of 282 nm.¹ The amount of 2,4-D adsorbed by the adsorbent was calculated from the difference between the initial and final/equilibrium solution concentrations. Therefore, the amount of 2,4-D adsorbed on the adsorbent and solid-phase distribution coefficient of 2,4-D was computed from the following equation:

q_e = (C_o − C_e)V/W (1)

K_d = ((C_o − C_e)/C_e)V/W (2)

where q_e (mg g⁻¹) is the amount of 2,4-D adsorbed/weight of adsorbent. C_o and C_e represent the initial and equilibrium concentrations of 2,4-D solution (mg L⁻¹). V is the volume of the solution in mL and W is the weight of the adsorbent in g. To optimize the adsorption process, the influence of specific process parameters such as pH, adsorbent dosage, equilibration time, surfactant loading, temperature and presence of humic acid on 2,4-D adsorption has been investigated by changing one parameter and keeping the other parameters constant. All working solutions were prepared from 2,4-D stock solution by diluting with 95% ethanol and all experiments were carried out in duplicate.

Results and discussion

Surfactant loading and material characterization

The results of surfactant loading on clinoptilolite, zeolite Y, bentonite and montmorillonite materials indicate that the amount of surfactant loaded increases with increase of initial surfactant concentration in solution. The maximum adsorption of both surfactant cations was observed in the solution with an initial concentration of 1000 mg L⁻¹. At this particular concentration, the adsorbent surface seemed to be saturated with the surfactant. Detailed studies on the surfactant configuration and bi-layer formation mechanism on the adsorbent surface were discussed in our previous work.²¹ Therefore, organo-silicate materials with highest surfactant loading i.e. 1000 mg L⁻¹ of HDTMA (HLC, HLY, HLB and HLM) and DODMA (DLC, DLY, DLB and DLM) were used throughout the adsorption studies (except for the effect of surfactant loading experiments) and characterization.

The materials are found to exhibit a typical infrared spectroscopic pattern. The FT-IR spectra of clinoptilolite, zeolite Y, bentonite and montmorillonite with and without surfactant loading are presented in Fig. 1. The characteristic vibrational bands due to internal vibration of the primary unit of the tetrahedron structure TO_4/2 (T = Si or Al) have been observed at 950–1250 cm⁻¹ and 420–500 cm⁻¹ and these bands are insensitive to other structural vibrations. The strongest vibration at 950–1250 cm⁻¹ is assigned to T–O stretching and the next strongest band at 420–500 cm⁻¹ is assigned to T–O bending mode. The stretching bands are sensitive to Si : Al ratio of the framework while the bending vibrations are related to the linkages between the tetrahedrons. Furthermore, structural vibrations are sensitive to the overall structure and joining of the individual tetrahedral in the secondary structural unit as well as their existence in large pore openings. The band around 3500 cm⁻¹ is attributed to loosely bound water molecules and O–H stretching. The water molecules attached to the aluminosilicate framework show strong characteristic structure sensitive bands (shown by drawing ovals in Fig. 1) due to H–O–H bending vibration at 1650 cm⁻¹.²² The peaks below 550 cm⁻¹ are generally due to δ(O–T–O) bending and rotation mode. The peaks between 700–850 cm⁻¹ and 1000–1150 cm⁻¹ are assigned to symmetric and anti-symmetric T–O–T stretching vibration, respectively.²³ To affirm the surfactant loading and structural stability of the materials, the major peaks appearing in unmodified materials were compared to those of surfactant modified materials. The obtained FT-IR spectra (Fig. 1) reveal that, in contrast to unmodified materials, two intense bands around 2850 and 3000 cm⁻¹ have been observed in the infrared spectrum of HLC, DLC, HLY, DLY, HLB, DLB, HLM and DLM (shown by dotted lines in Fig. 1). These two bands are assigned to the asymmetric and symmetric stretching vibration of C–CH₂ of the alkyl chain, respectively.^24,25 Furthermore, the structure sensitive bands remain unaffected by surfactant treatment which predicts the structural stability of all adsorbent materials. The morphological studies of the reported silicates reveal that except zeolite Y, the rest of the silicates (clinoptilolite, bentonite and montmorillonite) particles exist in an agglomerated form (Fig. 2).

Fig. 1 FT-IR spectra of clinoptilolite, zeolite Y, bentonite, montmorillonite, and their HDTMA (HLC, HLY, HLB and HLM) and DODMA (DLC, DLY, DLB and DLM) loaded forms.

Fig. 2 SEM micrographs of clinoptilolite, zeolite Y, bentonite and montmorillonite.

The calculated values of cation exchange capacities for clinoptilolite, zeolite Y, bentonite and montmorillonite are found to be 1.24, 1.31, 1.68 and 1.87 mequiv. g⁻¹, respectively. This CEC data reveal that phyllosilicate montmorillonite has a maximum exchange capacity in comparison to other phyllosilicate (bentonite) and tectosilicates (clinoptilolite and zeolite Y).

Adsorption studies of 2,4-D

Effect of solution pH on 2,4-D adsorption. The adsorption behavior of 2,4-D on unloaded and surfactant loaded adsorbents at different pH (viz. 3, 5, 7, 9 and 11) of the solution was studied. The adsorption data presented in Fig. 3 reveal that the maximum adsorption of 2,4-D is achieved at neutral pH for all materials. The results show that the amount of 2,4-D adsorbed per gram of adsorbent increases with increase in pH up to pH 7. However further increase in pH results in a decrease of 2,4-D adsorption. The change in adsorption with changing the pH may be due to the change in the form and function of 2,4-D with pH changes.²⁶ In general, at high pH, 2,4-D molecules can easily dissociate (pK_a = 2.73) to form anionic form, and almost all existed in a dissociated form in solution. Therefore, the anionic form of 2,4-D shows greater electrostatic interactions with the positively charged surfactant loaded silicate surface (<pH 7).²⁷ Further, a more polar molecule is more likely to move closer to the charged surface, increasing the likelihood of van der Waals interactions.²⁸ However, due to its anionic form, the 2,4-D molecule would experience repulsion by the predominantly negatively charged (OH⁻) aqueous phase (above pH 7).²⁶ Therefore, the competition between the anionic form of 2,4-D molecule and OH⁻ will decrease the adsorption percentage at pH > 7.²⁹

Fig. 3 Effect of pH of the aqueous phase on the adsorption of 2,4-D onto different adsorbents (temperature 303 ± 1.0 K; weight of adsorbent 200 mg; initial 2,4-D concentration 552 mg L⁻¹; equilibration time 24 h).

Effect of the adsorbent dose. The amount of adsorbed 2,4-D with a varying amount of adsorbent presented in Fig. 4 reveals that increase in the adsorbent dosage results in increased percentage adsorption of 2,4-D for all adsorbents. A greater concentration of the adsorbent means more number of active sites for adsorption. Therefore, the increased amount of adsorbent facilitates higher adsorption sites for different adsorption, ion-exchange, binding processes etc. to take place on the adsorbent surface. Maximum adsorption is observed with 200 mg adsorbent dosage.

Fig. 4 Effect of the adsorbent dose on the adsorption of 2,4-D onto different adsorbents (temperature 303 ± 1.0 K; pH of solution 7; initial 2,4-D concentration 552 mg L⁻¹; equilibration time 24 h).

Effect of equilibration time on 2,4-D adsorption. Kinetic studies for 2,4-D adsorption on all unloaded and surfactant-loaded tecto- and phyllosilicates have been performed to find out the optimum equilibration time for maximum adsorption. Adsorption data presented in Fig. 5 demonstrates that the amount of 2,4-D adsorbed per gram of adsorbent increases with increase in equilibration time. The plot of the amount of 2,4-D adsorbed per gram of adsorbent versus the contact time between adsorbate–adsorbent indicates that the adsorption becomes asymptotic to the time axis after 12 h of equilibration and represents nearly an equilibrium pattern. Since the maximum adsorption for all adsorbents is observed at 24 h of equilibration, therefore, 24 h is considered as the optimum equilibration time for other adsorption studies.

Fig. 5 Effect of the equilibration time on the adsorption of 2,4-D onto different adsorbents (temperature 303 ± 1.0 K; pH of solution 7; initial 2,4-D concentration 552 mg L⁻¹; weight of adsorbent 200 mg).

Effect of surfactant loading on 2,4-D adsorption. The adsorption of 2,4-D on clinoptilolite, zeolite Y, bentonite, montmorillonite loaded with different concentrations of HDTMA and DODMA is presented in Fig. 6. It is evident from Fig. 6 that surface modification greatly enhances the adsorption potential of the material for 2,4-D adsorption. Furthermore, the amount of adsorbed 2,4-D increases with increase in surfactant loading and maximum adsorption is observed for DLB, DLM, HLB and HLM. The data also reveal that surfactant loaded forms of both phyllosilicates (bentonite and montmorillonite) show better adsorption potential for 2,4-D than that of reported tectosilicates. The greater adsorption of 2,4-D on surfactant loaded montmorillonite and bentonite may be due to their high cation exchange capacity because the loading of the surfactant and its configuration on the material surface depend on CEC or external CEC of the material.²¹ Similar kinds of results have been reported for the adsorption of other acetic herbicides on surfactant loaded material.^20,30 The retention of herbicide by surfactant modified silicates is mainly attributed to hydrophobic interactions between the bulky herbicide molecule and alkyl-chains of the quaternary cation of the surfactant. This mechanism was supported by the reversibility observed in adsorption–desorption isotherms which contributes to this increase in adsorption of 2,4-D with increase in surfactant loadings.^17,31

Fig. 6 2,4-D adsorption by clinoptilolite, zeolite Y, bentonite, montmorillonite loaded with different concentrations of HDTMA and DODMA (temperature 303 ± 1.0 K; weight of adsorbent 200 mg; pH of solution 7; initial 2,4-D concentration 552 mg L⁻¹; equilibration time 24 h).

Effect of temperature on 2,4-D adsorption. According to the adsorption theory, adsorption decreases with rise in temperature, and molecules adsorbed earlier on a surface tend to desorb from the adsorbent surface at elevated temperatures. But for activated carbon, inorganic and organic ion exchanger different results are noticed where decreasing viscosity and increasing molecular motion at higher temperature allow the uptake of molecules into the pores more easily, which causes adsorption to increase with increasing temperature. Thus, system temperature is considered to be a crucial parameter in the adsorption reaction. Therefore, the purpose of this study is to ascertain the effect of temperature on the adsorption of 2,4-D on clinoptilolite, zeolite Y, bentonite, montmorillonite and their surfactant loaded forms.

The obtained results of temperature effect have been presented in Fig. 7 and indicate that an increase in temperature from 303 to 323 K results in a decrease in 2,4-D adsorption. This decrease in adsorption may be attributed to change in the energy of adsorption or weakening of the van der Waals forces of attraction between 2,4-D molecule and the adsorbent surface, causing decrease in physical adsorption.^32,33 The other reason for the decrease in 2,4-D adsorption with increase of system temperature is change in the solubility of herbicides with change in temperature.^34,35 Results of the thermal study confirm that all reported adsorbents show a negative response to an increase in temperature for 2,4-D adsorption, and 303 K has been considered as optimum temperature for the adsorption studies.

Fig. 7 Effect of the system temperature on the adsorption of 2,4-D onto different adsorbents (weight of adsorbent 200 mg; pH of solution 7; initial 2,4-D concentration 552 mg L⁻¹; equilibration time 24 h).

Effect of the presence of humic acid on 2,4-D. The effect of dissolved humic acid on the adsorption of 2,4-D by unmodified and modified silicates was studied by taking 250 mg L⁻¹ of humic acid in the adsorption system and keeping the other parameters the same as described above. The reported results (Fig. 8) indicate a slight decrease in 2,4-D adsorption in the presence of humic acid. Low 2,4-D adsorption results from the interaction of humic acid with surface modified materials which causes low availability of the adsorption site on modified materials for 2,4-D to get adsorbed. The study made by Zhan et al.³⁶ also represents the affinity of surface modified materials for humic acid adsorption. Furthermore, there is also a possibility of interactions between 2,4-D and humic acid which can result in low accessibility of 2,4-D to get adsorbed on surfactant loaded materials.

Fig. 8 Effect of humic acid presence on the adsorption of 2,4-D onto different adsorbents (temperature 303 ± 1.0 K; weight of adsorbent 200 mg; pH of solution 7; initial 2,4-D concentration 552 mg L⁻¹; equilibration time 24 h).

Adsorption isotherm

The design of an adsorption system requires equilibrium isotherm data and in this study we consider the most commonly used Langmuir models to explain the same. The Langmuir isotherm is based on the premise that all active sites have an identical energy strength. Using the method of linear regression, the adsorption data have been fitted to the Langmuir adsorption model to describe the adsorption processes between the solid–liquid interfaces:³⁷

C_e/q_e = 1/(K_Lq_m) + C_e/q_m (3)

where K_L and q_m are the Langmuir coefficients.

Furthermore, the feasibility of the adsorption process is evaluated by a method suggested by Weber and Chakraborti et al.³⁸ One of the essential characteristics of the Langmuir equation could be expressed by a dimensionless constant called a separation factor, R_L which can be calculated as:³⁹

R_L = 1/(1 + K_LC_o) (4)

where C_o is the maximum initial 2,4-D concentration (mg L⁻¹).

Fig. 9a shows a typical fit of linearized Langmuir adsorption isotherm for the experimental adsorption data. The values of the Langmuir constant (K_L), the monolayer adsorption capacity of adsorbent (q_m), and R² are listed in Table 1. The obtained R² values confirm that the adsorption equilibrium data fit well to the Langmuir model. This also indicates the uniform adsorption and strong 2,4-D–adsorbent interactions over the adsorbent surfaces. But the deviation from linearity (Fig. 9a) and the low correlation coefficient (R²) values obtained for unloaded tectosilicates and phyllosilicates indicate that adsorption data for these adsorbents does not fit well into the Langmuir isotherm. The Langmuir parameters, q_m and K_L summarised in Table 1 were determined from the slope and intercept of the plots. The fitting of the adsorption data to the Langmuir adsorption isotherm reveals uniform adsorption and strong 2,4-D–adsorbent interactions over the surfaces of the surfactant loaded tectosilicates and phyllosilicates. The Langmuir monolayer sorption capacity, q_m is found to be higher for 2,4-D adsorption on HLM (158.7 mg g⁻¹) and DLM (161.3 mg g⁻¹) as compared to other adsorbents. A more reliable indicator for adsorption i.e. R_L parameter (Table 1) for 2,4-D adsorption on unloaded and surfactant loaded adsorbents are found to be positive and less than unity indicating thereby highly favorable adsorption.

Fig. 9 Adsorption isotherms: (a) Langmuir and (b) Freundlich, for 2,4-D adsorption studies (temperature 303 ± 1.0 K; weight of adsorbent 200 mg; pH of solution 7; equilibration time 12 h).

Table 1 Langmuir and Freundlich isotherm constants for adsorption of 2,4-D on unloaded and surfactant loaded tectosilicates and phyllosilicates

Isotherm constants		Clinoptilolite	HLC	DLC	Zeolite Y	HLY	DLY	Bentonite	HLB	DLB	Montmorillonite	HLM	DLM
Langmuir	qm (mg g^−1)	42.02	94.34	75.19	51.81	116.28	120.48	121.95	129.87	133.33	78.13	158.73	161.29
	KL (L mg^−1)	0.005	0.019	0.160	0.003	0.012	0.010	0.002	0.007	0.007	0.003	0.006	0.010
	R^2	0.93	0.99	0.98	0.88	0.99	0.99	0.78	0.99	0.98	0.87	1.00	1.00
	RL	0.266	0.088	0.011	0.377	0.130	0.148	0.534	0.213	0.196	0.353	0.240	0.157
Freundlich	n	2.38	2.96	1.39	1.92	1.91	1.72	1.36	0.94	1.24	1.65	0.75	1.06
	KF	6.56	335.27	3.35	1.65	46.43	25.46	0.29	0.13	2.56	2.63	0.01	0.80
	R^2	0.90	0.90	0.97	0.94	0.89	0.78	0.96	0.97	0.87	0.88	0.98	0.80

---

The Freundlich adsorption isotherm is the second most widely used mathematical description to characterize the adsorption process. This isotherm also provides an expression encompassing the surface heterogeneity and exponential distribution of active sites and their energy;

ln q_e = ln K_F + 1/n ln C_e (5)

where K_F and n are the Freundlich constants.

The Freundlich constants K_F and n were calculated from the equilibrium data for adsorption of 2,4-D on unloaded and surfactant loaded silicates at 303 K, and these are reported in Table 1 along with the correlation coefficient (R²) values. No satisfactory results were found on fitting the obtained adsorption data to the Freundlich adsorption isotherm model for 2,4-D adsorption at the desired temperatures (Fig. 9b). The values of the correlation coefficients for each plot reported in Table 1 demonstrate non-linearity of the Freundlich plots. So the calculated values of the Freundlich constants could not be considered authentic. For 2,4-D adsorption process the Langmuir adsorption isotherm model showed a better fit than the Freundlich isotherm.

Thermodynamic parameters

Evaluation of the thermodynamic parameters was made to assess the spontaneity of the adsorption process.^40,41 The influence of the temperature variation was examined on the adsorption of 2,4-D at fixed concentration of 552 mg L⁻¹, equilibration time of 24 h, aqueous : solid phase ratio of 1 :  4 (mL mg⁻¹) and temperature 303 K to 323 K. The Gibbs free energy change (ΔG^o) of the adsorption process is related to the equilibrium constant by the classic van't Hoff equation:

ΔG^o = −RT ln K_d (6)

According to thermodynamics, Gibbs free energy change is also related to the entropy change (ΔS^o) and the heat of adsorption (ΔH^o) at constant temperature by the following equation:

ΔG^o = ΔH^o − TΔS^o (7)

On combining eqn (6) and (7), we get

ln K_d = −ΔG^o/RT = (ΔS^o/R) − (ΔH^o/RT) (8)

where T is the absolute temperature (K) and R the universal gas constant (8.314 J mol⁻¹ K⁻¹). The values of heat of adsorption (ΔH^o) and entropy change (ΔS^o) can be determined from the slope and intercept, respectively, of the linear van't Hoff plot (eqn (8)).

The thermodynamic parameters (ΔG^o) varying with temperature were calculated using the distribution constant (K_d). Table 2 represents the values of the apparent thermodynamic parameters for adsorption of 2,4-D on twelve different adsorbents and these values were calculated using the van't Hoff plots (Fig. 10) having very good correlation coefficient values (R², presented in the same figure corresponding to each legend). The negative values of ΔG^o for all adsorbents indicate that the adsorption process is spontaneous with high adsorption affinity for the 2,4-D molecule. Furthermore, the negative values of ΔH^o confirm an exothermic nature of the 2,4-D/surfactant loaded silicate adsorption process. Hence, by increasing the temperature, the degree of adsorption will decrease, therefore low temperature is favorable for the present 2,4-D adsorption studies. The value of ΔS^o indicates randomness or disorder of the solid–solution interface during 2,4-D adsorption.

Table 2 Gibbs free energy change, enthalpy change and entropy change values for 2,4-D adsorption on unloaded and surfactant loaded adsorbents

Adsorbent	ΔG^o (kJ mol^−1)			ΔH^o (kJ mol^−1)	ΔS^o (K J mol^−1)
	303 K	313 K	323 K
Clinoptilolite	−10.54	−10.44	−10.21	−8.12	7.32
HLC	−14.63	−15.03	−15.27	−4.94	31.97
DLC	−14.98	−15.27	−15.29	−7.47	24.85
Zeolite Y	−10.71	−10.66	−10.63	−8.54	6.77
HLY	−14.94	−15.33	−15.43	−7.52	24.59
DLY	−14.95	−15.26	−15.28	−9.94	16.71
Bentonite	−12.18	−12.42	−12.59	−5.99	20.48
HLB	−16.14	−16.47	−16.68	−7.95	27.04
DLB	−16.34	−16.61	−16.54	−9.25	23.43
Montmorillonite	−11.42	−11.61	−11.52	−7.77	12.11
HLM	−15.84	−16.14	−16.36	−7.98	25.98
DLM	−15.97	−15.99	−16.07	−14.44	5.02

---

Fig. 10 Variation of ln K_d with temperature (1/T) for the adsorption of 2,4-D on clinoptilolite, zeolite Y, bentonite, montmorillonite, and their HDTMA (HLC, HLY, HLB and HLM) and DODMA (DLC, DLY, DLB and DLM) loaded forms (van't Hoff plot).

Adsorption mechanisms

These adsorption studies of 2,4-D clearly demonstrate that the studied tectosilicates and phyllosilicates show better adsorption after surfactant loading. Furthermore, the surfactant loading studies, as reported earlier in this manuscript, reveal that the higher the loaded surfactant concentration on the silicate surface, the more 2,4-D adsorption will be. To describe the adsorption mechanism, a schematic diagram is presented in Fig. 11. The higher adsorption tendency of surfactant loaded adsorbents for 2,4-D is mainly due to a change in the functionality of the adsorbent surface. On surfactant treatment of tectosilicates and phyllosilicates with HDTMA or DODMA, there is a layer formation of surfactant molecules on these silicate surfaces. Therefore, surfactant loading on the silicate surface leads to a change in the hydrophobic property of the same. In case of minerals, the surfactant loading also helps in neutralizing the negative surface charge because the existence of a negative charge on the adsorbent surface can repel the anionic group of 2,4-D molecule.⁴² Hence, 2,4-D adsorption on surfactant loaded silicates (HLC, DLC, HLY, DLY, HLB, DLB, HLM and DLM) can be governed by electrostatic attractions between the anionic 2,4-D molecule and the positively charged surfactant molecules. Moreover, hydrophobic interactions between the bulky herbicide molecule and the long alkyl-chains of the surfactant molecule also play a key role in the adsorption process. Other driving forces for 2,4-D adsorption on the organo-silicate surface is van der Waals interactions.⁴ Therefore, the amount of surfactant loaded on silicate nanoparticles plays a key role in controlling the adsorption of 2,4-D. Hermosín and Cornejo³⁰ have also reported in their study that the herbicide 2,4-D was adsorbed on surfactant loaded clays prepared from primary alkylammonium cations by a combination of hydrophobic, polar, and ionic interactions. In our studies, polar interactions should be significantly less important due to higher hydrophobicity of the tetra-substituted amino group of HDTMA and DODMA but an arrangement of alkyl chains of the surfactant loaded on a silicate surface also effects the 2,4-D adsorption significantly.³¹ The adsorption will be higher if the organic cations arrange themself in a vertical position which results in high basal spacing of the resultant organo-silicates in comparison to those having a horizontal arrangement of the alkyl chains of cationic surfactant molecules. Although we don't have any experimental proof, but from the high adsorption capacity value of surfactant loaded silicates in comparison to unmodified silicates, we can expect that in our case there will be a vertical arrangement of the organic cations in HLC, DLC, HLY, DLY, HLB, DLB, HLM and DLM. This vertical arrangement of the surfactant cations makes the interlayer space more accessible for adsorbate molecule. Besides the interlayer thickness, the nature and amount of surfactant loaded on the surface of the adsorbent also have an influence on 2,4-D adsorption process.^31,43

Fig. 11 Schematic diagram representing the adsorption mechanism.

Slow release formulation

Slow release formulation (SRF) of acidic herbicide (2,4-D) has been prepared by incorporating it into micelles of organic cation HDTMA and DODMA which adsorbed on materials used as described in previous publications.²¹ The slow release studies viz., thin layer-funnels analytical test and soil columns-analytical test have also been performed in the same manner as discussed earlier.²¹

The results presented in Fig. 12 clearly indicate that SRF can be used as a slow release supplier of 2,4-D. The surfactant loaded materials and their corresponding SRFs have a greater tendency to hold the anion/hydrophobic organic compounds and also release them slowly. Unmodified materials releases around 50% of 2,4-D after first irrigation and more than 90% of the applied 2,4-D is leached out after 6th irrigation in case of soil. On the other hand, SRF shows its slow release tendency with only 20–30% of leaching after the 1st irrigation and around 40–50% after the 6th irrigation, and still shows nearly 2–4% leaching after the 24th irrigation, which suggests its slow release capacity.

Fig. 12 Leaching studies of 2,4-D loaded on soil, unmodified and modified silicates using a thin layer funnels-analytical test (temperature: 303 ± 1.0 K, DIW as leachant).

A soil-column analytical test has been applied on 2,4-D loaded soil, clinoptilolite, zeolite Y, bentonite, montmorillonite, HLC, DLC, HLY, DLY, HLB, DLB, HLM and DLM. Before loading 2,4-D on the selected soil sample, leaching studies were made in the percolating system with the same soil. From this experiment we detected a very small amount of 2,4-D in the soil sample. As the concentration of 2,4-D is negligible, it can be neglected. Therefore, all measured concentrations of 2,4-D from leachates obtained from the percolating system having soil plus SRFs can be attributed to the source exclusively. It is clear from the results presented in Fig. 13 that at the start of the experiment (Day-1) 2,4-D desorption is maximum (51.5%) for soil whereas 40.6% for montmorillonite, 38.5% for zeolite Y, 38.3% for clinoptilolite and 31.2% for bentonite. For surfactant loaded slow release formulations 2,4-D release (i.e., 29.7% for HLC, 31.9% for DLC, 25.4% for HLY, 27.9% for DLY, 26.3% for HLB, 24.3% for DLB, 29.5% for HLM and 23.6% for DLM) is relatively low in comparison to their respective unmodified forms. Further, it is also clear from Fig. 13 that initial stage desorption of 2,4-D from all 13 formulations including soil occurs rapidly. In case of soil, clinoptilolite, zeolite Y, bentonite and montmorillonite there is a very small release of 2,4-D after 10 days. But on the other hand, even after 15 days there is a release of about 3% 2,4-D from the SRFs prepared from HLC, DLC, HLY, DLY, HLB, DLB, HLM and DLM.

Fig. 13 Leaching studies of 2,4-D loaded on soil, unmodified and modified silicates using soil column-analytical test (temperature: 303 ± 1.0 K, DIW as leachant).

Conclusions

The results of the batch experiment and slow release-analytical test showed that surface modified silicates can be served as a unique means for the removal of 2,4-D from aqueous solution and also as slow release formulations of 2,4-D. Data obtained by the adsorption of 2,4-D is fitted well with the Langmuir model. The high adsorption value indicates that these materials possess good adsorption potential for the removal of 2,4-D from aqueous solution. Slow-release formulations are excellent alternatives to reduce the use of soluble herbicides. A slow-release formulation is more convenient since less frequent application is required. The silicates after surfactant treatment with HDTMA and DODMA can be utilized as a slow release formulation and an efficient adsorbent for these agrochemicals, and the extent of adsorption increases with increasing surfactant loadings. The aluminosilicate materials show highest adsorption capacity after modification with the surfactant and indicate their possible use as controlled release formulations for herbicides.

References

1. Z. Aksu and E. Kabasakal, Sep. Purif. Technol., 2004, 35(3), 223–240.
2. Y. F. Chao, P. C. Chen and S. L. Wang, Appl. Clay Sci., 2008, 40(1–4), 193–200.
3. A. D. Carter, Weed Res., 2000, 40(1), 113–122.
4. M. C. Hermosín, R. Celis, G. Facenda, M. J. Carrizosa, J. Ortegacalvo and J. Cornejo, Soil Biol. Biochem., 2006, 38(8), 2117–2124.
5. J. M. Brixie and S. A. Boyd, J. Environ. Qual., 1994, 23, 1283–1290.
6. J. S. Hayworth and D. R. Burris, Environ. Sci. Technol., 1997, 31, 1277–1283.
7. T. B. Boving and M. L. Brusseau, J. Contam. Hydrol., 2000, 42(1), 51–67.
8. M. Rebhun, S. Meir and Y. Laor, Environ. Sci. Technol., 1998, 32, 981–986.
9. M. J. Carrizosa, M. C. Hermosín, W. C. Koskinen and J. Cornejo, Soil Sci. Soc. Am. J., 2003, 67, 511–517.
10. T. Badr, K. Hanna and C. de Brauer, J. Hazard. Mater., 2004, 112(3), 215–223.
11. T. M. Holsen, E. R. Taylor, Y. C. Seo and P. R. Anderson, Environ. Sci. Technol., 1991, 25, 1585–1589.
12. S. P. Nayyar, D. A. Sabatini and J. H. Harwell, Environ. Sci. Technol., 1994, 28, 874–1881.
13. S. Sun and P. R. Jaffe, Environ. Sci. Technol., 1996, 30, 2906–2913.
14. S. A. Boyd, W. F. Jaynes and B. S. Ross, Immobilization of organic contaminants by organo-clays: application to soil restoration and hazardous waste containment, in Organic substances and sediments in water, ed. R. S. Baker, Lewis Publication, Chelsea, MI, USA, 1991, pp. 181–200.
15. S. Sheng, G. Xu and S. A. Boyd, Adv. Agron., 1997, 59, 25–62.
16. H. Zhang, Y. Kim and P. K. Dutta, Microporous Mesoporous Mater., 2006, 88(1–3), 312–318.
17. M. J. Carrizosa, W. C. Koskinen and M. C. Hermosín, Soil Sci. Soc. Am. J., 2004, 68, 1863–1866.
18. Y. G. Mishael, T. Undabeytia, O. Rabinovitz, B. Rubin and S. Nir, J. Agric. Food Chem., 2003, 51, 2253–2259.
19. Y. El Nahhal, S. Nir, C. Serban, O. Rabinovitch and B. Rubin, J. Agric. Food Chem., 2000, 48, 4791–4801.
20. M. J. Carrizosa, M. J. Calderón, M. C. Hermosín and J. Cornejo, Sci. Total Environ., 2000, 247(2–3), 285–293.
21. D. Bhardwaj, M. Sharma, P. Sharma and R. Tomar, J. Hazard. Mater., 2012, 227–228, 292–300.
22. P. Sharma, P. Rajaram and R. Tomar, J. Colloid Interface Sci., 2008, 325, 547–557.
23. W. Mozgawa, J. Mol. Struct., 2001, 596, 129–137.
24. J. Warchoł, P. Misaelides, R. Petrus and D. Zamboulis, J. Hazard. Mater., 2006, 137, 1410–1416.
25. Ö. Gök, A. S. Özcan and A. Özcan, Desalination, 2002, 220, 96–107.
26. C. D. S. Tomlin, The pesticide manual: A world compendium, eleventh edition, Survey, British Crop Protection Council, UK, 1997.
27. S. Bakhtiary, M. Shirvani and H. Shariatmadari, Microporous Mesoporous Mater., 2013, 168, 30–36.
28. K. M. Spark and R. S. Swift, Sci. Total Environ., 2002, 298, 147–161.
29. K. S. Reddy and R. P. Gambral, Agric. Ecosyst. Environ., 1987, 18, 231–241.
30. M. C. Hermosín and J. Cornejo, J. Environ. Qual., 1993, 22, 325–331.
31. R. Celis, M. C. Hermosín, J. Cornejo and M. J. Carrizosa, Int. J. Environ. Anal. Chem., 2002, 82, 503–517.
32. E. Barriuso, V. Baer and R. Colvet, J. Environ. Qual., 1992, 21, 359–367.
33. T. Stutzmann and B. Siffert, Clays Clay Miner., 1977, 25, 392–406.
34. J. Brücher and L. Bergström, J. Environ. Qual., 1997, 26(5), 1327–1335.
35. M. J. Sánchez-Martin, J. M. Gonzales-Pozuelo and M. Sánchez-Camazano, Weed Res., 1993, 33, 479–486.
36. Y. Zhan, Z. Zhu, J. Lin, Y. Qiu and J. Zhao, J. Environ. Sci., 2010, 22(9), 1327–1334.
37. I. Langmuir, J. Am. Chem. Soc., 1918, 40, 1361–1403.
38. T. W. Weber and R. K. Chakraborti, AIChE J., 1974, 20, 228–238.
39. K. R. Hall, L. C. Eagleton, A. Acrivers and T. Vermenlem, Ind. Eng. Chem. Res., 1966, 5, 212–223.
40. D. Bhardwaj, P. Sharma and R. Tomar, Indian J. Chem., Sect. A: Inorg., Bio-inorg., Phys., Theor. Anal. Chem., 2007, 46, 1796–1800.
41. P. Sharma, D. Bhardwaj, R. Tomar and R. Tomar, J. Radioanal. Nucl. Chem., 2007, 274(21), 281–286.
42. X. Yunfei, M. Mallavarapu and R. Naidu, Appl. Clay Sci., 2010, 49, 255–261.
43. M. C. Hermosín and J. Cornejo, Chemosphere, 1992, 24(10), 1493–1503.

---