Removal of triazine-based herbicides on specific polymeric sorbent: fixed bed column studies

Materials

The following chemicals were obtained from commercial sources: divinylbenzene (80%) (DVB), benzoyl peroxide (BPO), sodium chloride, polyvinyl alcohol 88% hydrolyzed (PVA), ethanol, toluene, n-heptane purchased from Sigma-Aldrich. Monomer was purified by distillation before use. Deionized water (Millipore) was used for making the aqueous phase and solutions during sorption studies. Analyzed herbicides: simazine, atrazine, propazine, terbuthylazine and metamitron were obtained from Warsaw Institute of Industrial Organic Chemistry. All purities were ≥99% unless stated otherwise.

Preparation of polymer adsorbent

Polymer beads of poly(divinylbenzene) were synthesized in radical suspension polymerization. The continuous water phase comprised 2% w/w sodium chloride and 1% w/w PVA (calculated for organic phase). The dispersed organic phase contained monomer (DVB), initiator – BPO (0.5% w/w calculated for monomer) and solvents: toluene and n-heptane (1:7 w/w). Prepared polymer beads were modified with maleic anhydride in Diels-Alder reaction as described in [28]. First, maleic anhydrate was dissolved in toluene. Then previously prepared poly(divinylbenzene) beads were added. Modification was carried out at 110°C for 48 h. After that, base hydrolysis was performed using 3 M sodium hydroxide. Later the products were placed in ion-exchange columns and washed with water (for having carboxyl groups in sodium form on the polymer surface), hydrochloric acid and again with water (for having carboxyl groups in acidic form). Detailed synthesis and modification were described in [10].

Materials characterization

Sorption studies

Dynamic conditions were realized by the column filtration method. Fixed-bed sorption was done on 1.65 cm³ polymer bed (equivalent to 0.4 g of dry polymer) placed in a microcolumn (0.9 cm in diameter) by passing downwards a solution which was 10 ppm ethanol:water (1:9 v/v) herbicide solution. The feeding solution was pumped down-flow through the column by a peristaltic pump with flow rate 0.3 mL min⁻¹. The effluent was collected at regular time intervals from the bottom of the column and was then analyzed by UV/VIS spectrophotometry using Jasco V-630 apparatus. Wavelength was set at 222.0 nm for simazine, atrazine and propazine, 225.0 nm for terbuthylazine and 307.0 nm for metamitron, respectively. Flow of solution lasted each time until the moment of bed exhaustion, i.e. at the point where the concentration in the discharge became equal to that of the initial solution.

In desorption process ethanol was used as the eluent. It was passed through the adsorbent bed with constant flow equal to (0.3 mL min⁻¹) and the sampling time was 15 min. Desorption was carried out until maximum elution of the herbicides from the bed. Reusability of adsorbent was determined by its adsorption performance in three consecutive sorption/desorption cycles. All of the experiments were performed at room temperature (23±1°C).

Analysis of experimental data

The herbicide adsorption breakthrough curves were obtained by plotting the triazines concentration in the effluent (C) versus the effluent volume (V). The calculations were made based on the analysis included in the publication [29] and were carried out using the diagram presented in Fig. 1. Total adsorptive capacity P_t (mg g⁻¹) of specific material was calculated from the following formula, Eq. (2):

Fig. 1:

Supporting drawing for calculating total and usable adsorptive capacities C=C_o – point of bed exhaustion, C=C_p – breakthrough point [29].

(2)Pt=Ot/M

were M is the mass of bed prior to sorption process (g). The total amount of triazine retained in the column, O_t, was calculated from the Eq. (3):

(3)Ot=PDFBA–PDFB=PDFBA–∫DFf(x)

where P_DFBA is the area representing the quantity of compounds introduced to sorption system (the point of bed exhaustion, C=C_o=10 mg L^-1), P_DFB is the area representing the quantity of compounds not retained on the bed of adsorbent (the point of bed exhaustion, C=C_o=10 mg L⁻¹).

The usable adsorptive capacity P_u (mg g^-1) was calculated as follows, Eq. (4):

(4)Pu=Ou/M

The amount of triazine retained in the column to the breakthrough point O_u, was calculated from the Eq. (5):

(5)Ou=PDGHA–PDGI=PDGHA–∫DGf(x)

where P_DGHA is the area representing the quantity of compounds introduced to the sorption system (the breakthrough point, C=C_p=0 mg L^-1), P_DGI – the area representing the quantity of compounds not retained on the bed of adsorbent (the breakthrough point, C=C_p=0 g L⁻¹).

The mass transfer zone was calculated using the equation of Michaels and Treybal, Eq. (6):

(6)Ho=H⋅(te–tb)/(te–(1–φ)⋅(te–tb))

where: H_o – the adsorption front height, cm, H – the adsorbent bed height, cm, t_e – bed operation time until exhausted, min, t_b – bed operation time until breakthrough, min, φ – coefficient of sphericity of exit curves was calculated by dividing the surface area of the rectangle HBIJ−∫GFf(x)dx by the surface area of the rectangle HBIJ.

Mass exchange zone moving rate, u (cm min⁻¹), was calculated from the following formula, Eq. (7):

(7)u=Ho/(te–tb)

where: H_o – the adsorption front height, cm, t_e – bed operation time until exhausted, min, t_b – bed operation time until breakthrough, min.

Adsorbent characterization

In order to obtain polymer beads the poly(divinylbenzene) was synthesized in radical suspension polymerization. Product in the form of regular microspheres with diameters in the range 50–400 μm was obtained. Chemical modification of polymer was done by Diels-Alder reaction in which poly(divinylbenzene) reacts as diene and maleic anhydride as dienophile. The efficiency of the modification reaction was 49%. Generation of the carboxyl groups (see Fig. 2) capable of specific interactions followed by ring opening of the maleic anhydride with the use of base hydrolysis. Cycloaddition reaction scheme and the basic hydrolysis are shown in [10]. Selected material having porous structure and acidic groups on the surface was used in sorption experiments of herbicides based on triazine. Characteristics of this adsorbent are presented in Table 1. The study showed that the tested adsorbent has a well-developed surface area, which has a direct influence on the adsorption capacity of the material. The content of acidic groups (2.80 mmol g⁻¹) indicates the presence of the carboxyl groups in the adsorbent structure that have been introduced by modification of the free vinyl groups in the poly(divinylbenzene). These groups are able to form hydrogen bonds, their position and distance between them allow for the formation of complexes with adsorbate molecules such as triazines. Such specific, directional interactions are responsible for better sorption and selectivity when compared to the traditional polymeric adsorbents [9].

Fig. 2:

Fragment of the modified surface structure of the investigated poly(divinylbenzene).

Adsorption studies

Poly(divinylbenzene) modified with maleic anhydride in Diels-Alder reaction was selected for the sorption experiments because molecular structures of triazines exhibit complementarity to the arrangement of functional groups present in its structure. Sorption studies were performed in a dynamic mode for triazine-based herbicides such simazine, atrazine, propazine, terbuthylazine and, for comparison, for metamitron. The adsorbate solutions continuously pass through a column packed with adsorbent with flow rate 0.3 mL min⁻¹. The pH values of the feed solutions were seven and the concentrations of herbicides were 10 ppm. The performance of fixed bed is usually described using the breakthrough curve. Plots of the ratio C/Co (outlet adsorbate concentration/adsorbate feed concentration) or adsorbate concentration in the effluent (C) versus effluent volume (V) or time (t) are denominated as breakthrough curves. On the basis of obtained results the breakthrough curves (isoplanes) were plotted and presented in Fig. 3. The shape of the breakthrough curve and the time for a breakthrough are very important characteristics for determining the operation and the dynamic response of the fixed bed column. The analysis of the breakthrough curves clearly indicates that the tested specific adsorbent prefers sorption of triazine-based herbicides. For metamitron the polymeric bed is rapidly saturated. It was found that the breakthrough time increases with increasing the molecular weight of triazine and is much higher for more hydrophobic triazines (the logarithm of octanol/water partition coefficient (logK_ow) is 2.3, 2.7, 3.9 and 3.4 for simazine, atrazine, propazine and terbuthylazine, respectively [30]). In the case of terbuthylazine an extended breakthrough curve can be observed, indicating that a higher volume of solution could be treated. Calculating the parameters of sorption on the basis of the breakthrough curve is an important issue because it provides the basic information for the design of the column adsorption system and can be used to determine the rational scale of the column adsorption for practical application. The calculations were carried out in accordance with the diagram presented in Fig. 1. The parameters that characterize sorption of investigated herbicides in dynamic conditions were calculated using Eqs. (2–7). The obtained breakthrough curves can be described using mathematical equations presented in Table 2. The correlation coefficient R² for breakthrough curves are around 0.98–0.99. The analysis of isoplanes allows for determination of the total adsorption capacity (P_t) – up to the point of exhaustion and useful adsorption capacity (P_u) – up to the point of breakthrough and also the bed operation time until breakthrough (t_b) and exhaustion (t_e). Results are presented in Table 3. The adsorptive capacities obtained under dynamic conditions were comparable with those determined in batch sorption system for the same triazines concentration used, as shown in Table 3. The total adsorption capacity (P_t) obtained in the dynamic conditions were 15.3 mg g⁻¹, 26.1 mg g⁻¹, 43.5 mg g⁻¹, 66.2 mg g⁻¹ and 3.2 mg g⁻¹ for simazine, atrazine, propazine, terbuthylazine and metamitron, respectively. The sorption capacity of metamitron is 5–20 times smaller than the sorption capacity of triazine-based herbicides, which confirms the presence of a specific interaction between the sorbent and investigated triazines that enhance sorption. Selectivity of the tested adsorbent towards herbicides from the group of triazines has been confirmed in the previous publication [9]. The introduction of carboxyl groups into polymer structure resulted in creating hydrogen bonds with triazines. Specific interactions are consisted of two hydrogen bonds: between hydroxyl hydrogen atom of carboxylic group from modified divinylbenzene and nitrogen atom containing free electron pair from triazine (O-H…N) and between hydrogen atom of triazine amino group and carbonyl oxygen atom from modified divinylbenzene (O…H-N). The highest values of adsorptive capacity are observed for terbuthylazine and propazine. In the case of propazine, which has the same molecular weight as terbuthylazine, but is more hydrophobic, the sorption efficiency is smaller. This results can suggest that for propazine only non-specific interactions are possible, while sorption efficiency for terbuthylazine is intensified by additional specific interactions like hydrogen bonds. In the propazine molecule there are two bulky isopropyl groups, which probably can cause the steric hindrance in the creation of hydrogen bonds.

Fig. 3:

Adsorption breakthrough curves (isoplanes) obtained in the studied systems.

Analyzing the breakthrough curves one can notice that in spite of dosing a solution onto the column with the same initial concentration of the herbicide the breakthrough and exhaustion times were varied for the tested compounds. The longer time necessary to reach the breakthrough and exhaustion points (in this study observed for terbuthylazine and propazine) indicates much better filling of the polymer bed by triazines. Breakthrough time for both terbuthylazine and propazine is about 70 h, but exhaustion time is much higher for terbuthylazine (see Table 3). It means that the synthesized adsorbent has a much greater affinity for terbuthylazine. On the basis of the derived parameters the coefficients of sphericity of isoplanes (φ), the heights of adsorption fronts (H_o) and the mass exchange moving rates (u) were calculated (see Eq. 6–7). Results are presented in Table 4. In fixed bed experiments, the concentration of adsorbate in the mobile phase as well as in the solid phase changes with time and with the position in polymer bed. The mass transfer zone in a column moves from the top of the bed and proceed towards the bottom. In good sorption systems are small heights (H_o) of mass transfer fronts and low mass exchange moving rate (u). Small values of this parameters mean that the breakthrough curve is close to an ideal step with negligible mass-transfer resistance [29]. Analysis of the results shows that the best sorption parameters in dynamic conditions have been obtained for terbuthylazine and propazine. A much higher height of adsorption front was obtained for metamitron (7.01 cm). Results also show that only in the case of propazine and terbuthylazine the heights of mass transfer fronts are smaller than adsorbent bed height, which was 2.7 cm in performed experiments. The highest mass exchange moving rate for metamitron results probably from the weak affinity of this herbicide to the investigated polymer. For the same reason the breakthrough and exhaustion times have the smallest values and consequently the saturation of the adsorbent occurs earlier.

Desorption studies

Due to the good solubility of tested herbicides in ethanol it was decided to apply it in the desorption process as eluent. Based on the mass balance of herbicide adsorbed on the polymer bed and then desorbed the desorption process was investigated. The desorption efficiency, the desorption time and the volume of the eluent required to elute herbicides from the surface of the adsorbent were determined. The results are shown in Table 5. The use of ethanol as eluent assured the high recovery degree of adsorbed substances. Simazine can be completely desorbed. For other herbicides the desorption efficiency is in the range 88–97%. The desorption efficiency is related to the solubility of the herbicides in ethanol. For propazine and terbuthylazine the solubility in ethanol is 15 g L⁻¹. It was observed that the better solubility of the compound in the eluent, the time required to eluting substances from the adsorbent bed was shorter. Due to the low solubility of simazine in ethanol (0.57 g L⁻¹) its desorption time was the longest (1230 min). Despite only slightly higher solubility of metamitron in ethanol (1.1 g L⁻¹), elution was significantly faster (135 min), which may be caused by low binding affinity of metamitron to investigated sorbent. Desorption time for other triazines was comparable and was about 350 min. For maximum desorption of triazines showing the greatest affinity for the tested adsorbent the amount of needed eluent was about 100 mL. Furthermore, during the study it was observed that for all herbicides after passage a small volume of ethanol (ca. 4.5 mL) through the column already about 90% of investigated herbicides were desorbed (see Fig. 4), but the process was continued until complete elution of the herbicides from the bed. Thus, the effective elution of herbicides from polymer bed demonstrates that in industrial conditions already small volume of eluent provides sufficient regeneration of the bed.

Fig. 4:

Effect of eluent volume on the herbicides desorption efficiency.

Efficient removal of loaded herbicides from the adsorbent was necessary to ensure their long-term use for repeated sorption-desorption cycles. The tested adsorbent showed the greatest affinity for terbuthylazine, therefore, it was decided to investigate the ability to remove this triazine in three subsequent cycles. Ethanol was used as eluent. Desorption efficiency of investigated herbicide from modified poly(divinylbenzene) was evaluated based on the amount of terbuthylazine released to ethanol. In each cycle, after passing 100 mL of eluent, the polymer bed was washed using distilled water. Based on the results obtained after three sorption/desorption cycles it can be concluded that the removal of terbuthylazine is still effective. It was observed that the sorption capacities of regenerated modified poly(divinylbenzene) adsorbent in the first and second cycles are comparable (67.1 and 65.6 mg g⁻¹). In the third cycle the sorption capacity of the bed is reduced by 20% to 54.2 mg g⁻¹, but it is still a good result for effective removal of investigated triazine.