Journal of Porous Materials

, Volume 19, Issue 1, pp 79–86

Adsorption of Imazamox herbicide onto Filtrasorb 400 activated carbon

  • Catherine Morlay
  • Etienne Quivet
  • Michaela Pilshofer
  • René Faure
  • Jean-Pierre Joly
Article

DOI: 10.1007/s10934-011-9450-4

Cite this article as:
Morlay, C., Quivet, E., Pilshofer, M. et al. J Porous Mater (2012) 19: 79. doi:10.1007/s10934-011-9450-4

Abstract

Imazamox is an imidazolinone herbicide, a new class of pesticides, which can exist as cationic, anionic or neutral species in water. The adsorption isotherms of Imazamox onto Filtrasorb 400 (F400) activated carbon were determined varying the pH and the ionic strength of the aqueous medium. The results show that ionic strength has no significant effect on Imazamox uptake, contrary to pH, and that F400 has a high affinity for Imazamox. Moreover, it is found that Imazamox adsorbs onto F400 as its neutral form. The best fit of the experimental points is obtained with the Langmuir–Freundlich model, consistent with surface site heterogeneity. Finally, calculating Langmuir–Freundlich isotherms for various constant pH values, it is shown that the two plateaus observed in the experimental isotherms obtained at free pH are due to the variation of the pH along the isotherms.

Keywords

Filtrasorb 400 activated carbon Imazamox herbicide Adsorption Langmuir–Freundlich equation 

1 Introduction

According to the European Directive relative to drinking water quality [1], the maximum concentration of individual pesticides in drinking water has to be lower than 0.1 μg l−1 and the total concentration of all pesticide species should not exceed 0.5 μg l−1. These requirements can be most of the time achieved using activated carbon in the drinking water production process to adsorb a number of hazardous compounds present in natural waters. However, a good understanding of the adsorption mechanism helps choosing the appropriate operative conditions of treatment.

Imidazolinones are a new class of pesticides which are used in order to control a wide spectrum of broad-leaf weeds and grasses [2]. Imidazolinones act as branched chain amino acid synthesis inhibitors [3]. The studied imidazolinone herbicide is Imazamox (2-[4,5-dihydro-4-methyl-4-(1-methylethyl)-5-oxo-1H-imidazol-2-yl]-5-(methoxymethyl)-3-pyridinecarboxylic acid; C15H19N3O4). Imazamox is specifically used as a post emergence weed control for corn, soya, peanuts and other edible plants [4]. To our knowledge, no quantitative study has yet been published on the adsorption of Imazamox onto activated carbon.

The activated carbon chosen for this study is Filtrasorb 400 (F400). This Chemviron Carbon (European Operations of Calgon Carbon Corporation, USA) product is widely used for drinking water production and it is specially recommended for treatment aimed to remove organic micropollutants such as pesticides. F400 is often cited in the literature and thus it is an interesting material for data comparison purposes.

The aim of this work was to get an insight into the mechanism of Imazamox adsorption. This was achieved by studying the uptake of the herbicide from aqueous solution onto F400 activated carbon varying Imazamox concentration, pH and ionic strength of the aqueous medium in specifically designed experiments.

2 Experimental section

2.1 Materials

Imazamox was purchased from Chem Service, Inc (USA) with a 99.0 ± 0.5% purity. The herbicide as well as its 250 mg l−1 (8.2 × 10−4 mol l−1; Mw = 305.3 g mol−1) aqueous stock solution were kept in stopped flasks, at 6 °C, in the dark as imidazolinone compounds are UV light sensitive [5, 6]. The diluted solutions of the herbicide were prepared just before use. The solubility of Imazamox in water was found in the literature and refers to the technical product (purity ≥ 95.9%): 4.4 g l−1 at 20 °C [7].

F400 activated carbon is produced from bituminous coal and is physically (steam) activated. Granular F400 was grinded and the resulting powder had an average particle size of 40 μm. F400 did not undergo any further treatment.

Sodium nitrate, sodium hydroxide, nitric acid and phosphoric acid (Merck products) of analytical grade as well as acetonitrile “HPLC grade” (SDS product) were used. All the solutions were prepared with ultra pure water of 18 MΩ cm minimal resistivity produced by an Elga water purifier (Elgastas UHQ PS).

2.2 F400 characterization

For any determination, powdered F400 was dried in an oven (130 °C, at least for 24 h) before weighting. The elemental analysis was performed at the “Service Central d’Analyse” of the “Centre National de la Recherche Scientifique” (CNRS; Vernaison, France).

Temperature programmed desorption experiments followed by mass spectrometry analyses (TPD-MS) were performed in order to evidence the possible presence of oxygen groups. TPD-MS experiments were performed using a Setaram (TGA 92 16.18 model) thermobalance equipped with a Balzers (QMG 420 model) quadrupole mass spectrometer. Experiments were carried out under an argon stream (1.5 l h−1) at 1 atm. The heating rate was 10 °C.min−1, from ambient temperature up to 1,000 °C. H2O, CO and CO2 desorption rates were followed by monitoring the heights of the mass peaks m/e = 18, 28 and 44 amu, respectively. The mass spectrometer was calibrated by decomposing various weights of calcium oxalate.

The pH of the point of zero charge (pHPZC) of F400 was determined using the mass titration method proposed by Noh and Schwarz [8]. The concentrations of the suspensions were in the range 5–150 g l−1 and the initial pH values were set at 3.0–5.0 or 7.0–11.0 using 0.1 mol.l−1 HNO3 and NaOH solutions. The suspensions were left under stirring in stopped glass flasks at 25 ± 1 °C. The final pH of the suspensions was measured after a 48 h stirring.

F400 textural characterization procedures have been described in detail elsewhere [9]. F400 nitrogen physisorption isotherm was determined at 77.4 K for equilibrium relative pressures, P/P0, in the range 4 × 10−7–0.995; F400 sample was previously outgassed at 393 K under low pressure vacuum. This isotherm was analyzed using the BET, the αS (Sing), the Dubinin-Radushkevich (DR) and the Density Functional Theory (DFT) methods. The following characteristics of F400 activated carbon were assessed: total surface area (SBET), external surface area (Sext), micropore (Smicro) and ultramicropore (Sultramicro) surface areas, total porous volume (Vp), microporous volume (Vmicro) and pore size distribution. Macro-, meso- and micro-pores refer to the definitions given by IUPAC [10].

2.3 Aqueous solution analyses

pH values were measured using a Metrohm (702 SM Titrino model) pH meter equipped with a combined electrode (Ag/AgCl saturated with 3 mol l−1 KCl; Metrohm).

Imazamox concentrations in aqueous solutions were determined using an Agilent Technology (LC 1100 model) liquid chromatograph, equipped with a photodiode array UV detector and monitored by the Chemstation software. A Chrompack Omnispher C18 column (250 × 4.6 mm i.d., dp = 5 μm) preceded by a Chromosep pre-column (10 × 4.6 mm i.d., dp = 5 μm) was used in the isocratic elution mode with the following mixture as the eluent: ultra pure water at pH = 3.0 ± 0.1 (with H3PO4)/acetonitrile at 70/30 (v/v). The flow rate was set at 1 ml min−1 and the injection volume of the aqueous samples was 20 μl. UV detection wavelength was 250 nm. Calibration curves were linear within the concentration range considered (i.e. from 3.3 × 10−6 to 8.2 × 10−4 mol l−1).

2.4 Adsorption procedure

Imazamox solutions at initial concentrations ranging from 1 to 250 mg l−1 (i.e. from 3.3 × 10−6 to 8.2 × 10−4 mol l−1) were prepared in glass flasks. Each sample was stirred and its initial pH value was measured when stable.

Adsorption experiments were carried out in static conditions, at 25 ± 1 °C in the dark. The suspensions of 10 mg of powdered F400 in 25 ml of Imazamox solutions were stirred for 24 h. Previous kinetic experiments showed that the adsorption equilibrium is reached in 2 h. This time is relatively short compared to some literature data because the activated carbon is used as a powder in the present work. The suspensions were then membrane filtered (0.45 μm porosity). The equilibrium herbicide concentrations in solution, C, as well as the equilibrium pH values were determined on the filtrates. Imazamox equilibrium concentrations in F400, q (mg g−1), were deduced from the initial and equilibrium concentrations of the herbicide in solution. Both F400 and Imazamox blanks (5 mg l−1, i.e. 1.6 × 10−5 mol l−1) were included in each experiment and treated in exactly the same way as the other samples. All blanks were found to be negative.

Adsorption isotherms were determined from two types of experiments: (1) Imazamox solutions containing 0.05 or 0.25 mol l−1 NaNO3 were used in order to study the influence of the ionic strength of the aqueous medium on the adsorption at free solution pH (no chemical was added to control the solution pH; initial pH in the range 3.5–6.4) and (2) Imazamox solutions containing 0.05 mol l−1 HNO3 (constant final ionic strength of the solution: 0.05 mol l−1) were used in order to study the influence of the acidification of the aqueous medium on the adsorption (equilibrium pH = 1.4).

A third type of experiment was carried out using solutions having a constant initial concentration of Imazamox (100 mg l−1, i.e. 3.3 × 10−4 mol l−1) and containing 0.05 mol l−1 NaNO3. In this experiment, the equilibrium pH of the suspensions was imposed using micro-additions of HNO3 and varied in the range 1.6–6.3. The micro-additions of HNO3 did not significantly modify the sample total volume and lead to a negligible increase of the solution ionic strength (less than 10%).

3 Results

3.1 F400 characterization

The results of F400 textural characterization have been presented and discussed in detail by the authors elsewhere [9]. The results of the elemental analysis are 84.2, 1.0 and 0.6% for C, H and N, respectively. The ash content is equal to 8.1%. The accuracy of the results is ± 0.3 on C, H, N and ash weight percentages.

TPD-MS analyses showed that F400 has a low content of oxygen groups; the amounts of CO and CO2 released during the thermal desorption of the sample are 280 and 85 μmol g−1, respectively. The oxygen content calculated from these results represents less than 1% in weight.

The three curves of the plot showing the final pH of the suspensions versus F400 concentration in solution reach a single plateau which corresponds to the pHPZC value (Fig. 1 not shown). F400 pHPZC was found to be 10.3 ± 0.2. This high pHPZC value is expected for a low oxygen content activated carbon.
Fig. 1

NLDFT pore size distribution of F400

The textural characterization of F400 showed that the total surface area SBET and the total porous volume Vp are 1,012 m2 g−1 and 0.57 cm3 g−1, respectively. In addition, F400 is a highly microporous activated carbon (pore size < 2 nm) having a large distribution of pore sizes which spreads from ultramicropores (pore size < 0.7 nm) up to narrow mesopores. Using the αS method, Vmicro, Vultramicro, Smicro and Sultramicro were found to be 0.37, 0.16 cm3 g−1, 907 and 488 m2 g−1, respectively. Figure 1 shows the NLDFT pore size distribution of F400; it can be seen from Fig. 1 that F400 has a large proportion of narrow micropores showing a maximum at 6 Å.

3.2 Imazamox adsorption

Figure 2 gathers the adsorption results obtained for the various experiments described above. All the isotherms exhibit a very steep part at low Imazamox concentrations which reflects the high affinity of F400 for the herbicide. The isotherms obtained at free solution pH show two pseudo-plateaus. According to Giles classification [11], these isotherms are of the H4 type. In very acidic medium (pH = 1.4), the lower pseudo-plateau disappears and the isotherm turns to H2 type. Finally, the height of the upper pseudo-plateau, obtained for the highest Imazamox concentrations used (up to 160 mg l−1), is independent from both the ionic strength and pH of the aqueous medium. It corresponds to 250 mg (i.e. 8.2 × 10−3 mol) of Imazamox adsorbed per gram of activated carbon.
Fig. 2

Adsorption isotherms of Imazamox onto F400 activated carbon and influence of the equilibrium pH value on Imazamox uptake. Lower left corner frame: Imazamox uptake by F400 versus solution equilibrium pH (Imazamox initial concentration: 100 mg l−1)

The isotherms obtained for different ionic strengths and at free solution pH are close to each other on the whole concentration range considered. The determination of each isotherm has been carried out at least twice; taking into account the dispersion due to experimental uncertainty, it can be stated that the little discrepancies observed between the different isotherms can not be considered to be significant. Therefore, one can conclude that varying the ionic strength of the solution in the range 0.05–0.25 mol l−1 has a negligible effect on Imazamox uptake by F400. Moreover, an increase of the solution pH was observed during Imazamox adsorption (equilibrium solution pH in the range 4.1–6.7; Fig. 2). In contrast, imposing a very acidic medium with HNO3 (solution pH = 1.4) lead to a much higher uptake at lower and intermediate Imazamox concentrations.

4 Discussion

4.1 Imazamox solution

Imazamox molecule contains four centers that can be protonated or not, depending on the pH of the aqueous medium. These centers are the three nitrogen atoms and the carboxylic acid function. Figure 3 shows the structure of the neutral form of Imazamox (H2L) and the acid–base equilibriums leading either to the cationic forms H3L+ and H4L2+ or to the anionic forms HL and L2−. This scheme is proposed for Imazamox by analogy with that given by Duda [12] for Imazapyr, another imidazolinone herbicide. Imazamox and Imazapyr have similar chemical structures except that Imazapyr molecule has no methoxymethyl substituent on the pyridine ring. The pKa values mentioned on Fig. 3 (pKa: < 1; 2.3; 3.3; 10.8) were found in the literature and refer to the technical product [7]. Figure 4 shows the distribution curves of the different forms of Imazamox in aqueous solution we calculated from the pKa values given above.
Fig. 3

Imazamox successive protonation-deprotonation equilibriums

Fig. 4

Distribution curves of the different forms of Imazamox in aqueous solution

4.2 Imazamox adsorption

Generally speaking, adsorption in aqueous medium results from an interplay of electrostatic (i.e. coulombic) and non-electrostatic interactions between the solutes and the carbon surface [13]. Nevertheless, it is usually considered that electrostatic interactions are preponderant when these two types of interactions are possible. For all the experiments described above, it is worth noting that equilibrium pH values were lower than 7. Consequently, the neat charge of the carbon surface was largely positive for all of the experimental conditions considered in this study (solution equilibrium pH < pHPZC = 10.3).

For the experiments carried out at free solution pH, equilibrium pH values were in the range 4.1–6.7 (Fig. 2). Considering the distribution curves of the different forms of Imazamox in aqueous solution (Fig. 4), this implies that the neutral (H2L) and anionic (HL) forms of Imazamox coexisted in solution, the anionic form being the preponderant one (pKa H2L/HL = 3.3). At this stage, the adsorption of the anionic form of Imazamox (HL) on the positive surface of the activated carbon could thus be assumed. However, we observed that an imposed decrease of the solution pH down to about 3 (third type experiment; (+) on Fig. 2), which favors the neutral form H2L, strongly increased Imazamox adsorption. An anionic adsorption mechanism is thus unlikely.

On the other hand, when the suspension equilibrium pH was maintained constant and equal to 1.4 by means of nitric acid, we observed a higher uptake of Imazamox at lower and intermediate concentrations in solution, leading to the disappearance of the first pseudo-plateau of the isotherm. For this pH value, the neutral (H2L) and cationic (H3L+) forms of Imazamox coexist in solution, the cationic form being the preponderant one (pK H3L+/H2L = 2.3). However, considering the fact that the activated carbon surface is positively charged, the adsorption of the cationic form is unlikely.

Finally, the plot of Imazamox uptake by F400 versus solution equilibrium pH (third experiment; lower left corner frame in Fig. 2) evidences a maximum adsorption in the pH range 2.3–3.3 where Imazamox neutral form is preponderant in solution (Fig. 4). This pH range of maximum adsorption is centered on pH = 2.8 (average value for pK H3L+/H2L = 2.3 and pK H2L/HL = 3.3), which is the pH of the highest percentage of the H2L form in solution (i.e. 61%; Fig. 4). This result is consistent with the fact that Imazamox adsorption does not depend on the ionic strength of the aqueous medium, which strongly suggests that the adsorption mechanism does not involve any ionic adsorbate.

Considering that (1) protons are consumed during the adsorption at free solution pH, for equilibrium pH values in the range 4.1–6.7 (initial pH values in the range 3.5–6.4), (2) the acidification of the suspensions favors the adsorption for equilibrium pH values in the range 2.8–6.3 (third type of experiment) and (3) the acidification of the suspensions lowers the adsorption for equilibrium pH values in the range 1.6–2.8 (third type experiment), we propose the mechanism, shown on Fig. 5, expressing the solution equilibrium displacement accompanying the adsorption of Imazamox onto F400 activated carbon.
Fig. 5

Mechanism of Imazamox adsorption onto F400

According to this mechanism, when F400 activated carbon is added to Imazamox solutions having an initial pH in the range 3.5–6.4 (adsorption at free solution pH) then the adsorption of the neutral form H2L shifts Equilibrium 1 to the right, thus leading to the consumption of some protons and consequently to an increase of the pH. This effect adds to the shift of pH due to the basicity of the carbon surface.

On the other hand, the decrease from 6.3 to 2.8 of the equilibrium pH by acidification of the suspension with HNO3 (third type of experiment) brings protons to the system and consequently increases H2L form concentration in solution by shifting Equilibrium 1 to the right. This results in a higher uptake of Imazamox by F400. Then, decreasing the equilibrium pH of the suspensions to values lower than 2.8 renders H3L+ form preponderant in solution (Equilibrium 2) which results in a decrease of Imazamox uptake.

According to the mechanism proposed above, stating that the neutral H2L form is responsible of the adsorption of the herbicide onto F400, we plotted the adsorption isotherm as the amount of Imazamox adsorbed, q, versus H2L form equilibrium concentration in solution. This last concentration was obtained from the data used to draw Fig. 4, knowing the experimental adsorption equilibrium pH. The result is shown in Fig. 6. It is seen that all the experimental points from Fig. 2 (including free pH adsorptions, adsorptions carried out at fixed pH = 1.4 and adsorptions performed adding various amounts of HNO3), gather in a single curve without any inflection point. This corroborates the conclusion that Imazamox adsorbs through its neutral form.
Fig. 6

Amount of Imazamox adsorbed, q, versus H2L equilibrium concentration in solution. Models: L Langmuir, S Sips, T Temkin, F Freundlich. Experimental points: Opened triangle free pH; filled triangle pH = 1.4; + microadditions of HNO3

Langmuir, Freundlich, Temkin and Sips. [14, 15] models have been used to fit experimental data shown in Fig. 6; these models, respectively correspond to the following equations:
$$ q = q_{m} {\frac{KC}{1 + KC}} $$
(1)
$$ q = KC^{{{1 \mathord{\left/ {\vphantom {1 n}} \right. \kern-\nulldelimiterspace} n}}} $$
(2)
$$ q = K\ln \left( {aC} \right) $$
(3)
$$ q = q_{m} {\frac{{KC_{{}}^{{{1 \mathord{\left/ {\vphantom {1 n}} \right. \kern-\nulldelimiterspace} n}}} }}{{1 + KC_{{}}^{{{1 \mathord{\left/ {\vphantom {1 n}} \right. \kern-\nulldelimiterspace} n}}} }}} $$
(4)
where C and qm, denote the equilibrium concentration in solution and the amount adsorbed at surface saturation, respectively. K, n (with n > 1) and a are other parameters involved in the models. These equations have been applied to Imazamox neutral form by letting C equal to CH2L, the concentration of this form in solution. Table 1 provides the computed values of the parameters, as well as the minimized value of Σ defined by Eq. 5:
Table 1

Results of fittings with various models

Parameter

Model

Langmuir

Freundlich

Temkin

Sips

qm (mg g−1)

231

322

K (units)

106 (l g−1)

211 (mg g−1)(l mg−1)0.087

16.7 (mg l−1)

2.14 (l mg−1)0.23

n

11.54

4.39

a (l mg−1)

463,667

Σ (mg2 g−2)

24,171

6,053

7,206

4,284

$$ \Upsigma = \left( {q_{\exp } - q_{cal} } \right)^{2} $$
(5)

Comparing the values of Σ, it is seen that the best fit is obtained with the model of Sips. This model is also called “Langmuir–Freundlich” model because, contrary to the other models used in this paper, it has been designed to take into account both surface heterogeneity and adsorbent saturation at high solute concentrations [14, 15].

Using Eq. 4 with the values of the parameters of the Langmuir–Freundlich model given in Table 1, on the one hand, and knowing the relation existing between CH2L and C, the total concentration of all Imazamox species in solution at a given pH according to Fig. 4, on the other hand, it is possible to draw theoretical isotherms corresponding to constant solution pH values by plotting q versus C. Figure 7 shows these theoretical isotherms as dotted lines. The experimental isotherms obtained either for a constant pH equal to 1.4 or for free pH values (with an ionic strength equal to 0.05 mol l−1 in both cases) are also shown on Fig. 7.
Fig. 7

Experimental (continuous lines) and theoretical (dotted lines) adsorption isotherms; q versus C, the total concentration of Imazamox in solution at equilibrium

Considering the experimental pH values reported close to the experimental points, it is seen that the calculated curves are in a good agreement with the experimental data. Moreover, it can be seen from Fig. 7 that the pseudo-plateau observed for q approximately equal to 140 mg g−1 on the isotherm at free pH is simply due to the change in the pH of the solution.

All these observations lead to the conclusion that the neutral form H2L is responsible for Imazamox adsorption onto F400. Hence, van der Waals interactions are supposed to dominate this adsorption phenomenon which probably involves a π-π interaction [13, 16] because of the aromatic character of the Imazamox molecule. This type of interaction has been largely documented in the case of phenol and its derivatives [17]. In addition, the large amount of Imazamox adsorbed at the highest solution concentrations considered (250 mg g−1) clearly indicates that Imazamox molecule penetrates the microporous system of F400. Considering a micropore which is broad enough to accommodate the Imazamox molecule, the narrower is the pore accepting this molecule, the stronger is the interaction between the adsorbate molecule and the pore walls of the carbon. Taking into account the fact that F400 has a broad pore size distribution, this probably implies the existence of a distribution of adsorption energies; hence, considering a Langmuir–Freundlich type adsorption isotherm model seems to be justified.

5 Conclusion

This work shows that Imazamox herbicide strongly adsorbs onto F400 activated carbon, thus revealing a high affinity interaction especially at low adsorptive concentration. Furthermore, the large amount of Imazamox adsorbed at the highest solution concentrations considered indicates that Imazamox molecule penetrates the microporous system of F400.

The results obtained lead us to the conclusion that the neutral form of the herbicide is likely responsible for its adsorption. The calculated adsorption isotherm of Imazamox neutral form can be satisfactorily modeled using the Langmuir–Freundlich equation. This equation is appropriate to reflect the heterogeneity in the interaction between the Imazamox molecule and the walls of micropores of different sizes.

Finally, the simulation of a set of adsorption isotherms of total Imazamox at constant pH values using the Langmuir–Freundlich model showed that the existence of the two pseudo-plateaus experimentally observed at free pH can be explained by the shift of pH along the isotherm curves.

Acknowledgments

The authors wish to thank Chemviron Carbon (European Operations of Calgon Carbon Corporation, USA) who kindly provided them with Filtrasorb 400 samples. The corresponding author also gratefully acknowledges kind encouragement from James Hickman, Director of the NanoScience Technology Center (NSTC, University of Central Florida, USA), during her stay at NSTC.

Copyright information

© Springer Science+Business Media, LLC 2011

Authors and Affiliations

  • Catherine Morlay
    • 1
    • 2
  • Etienne Quivet
    • 3
    • 4
  • Michaela Pilshofer
    • 1
    • 3
  • René Faure
    • 3
  • Jean-Pierre Joly
    • 1
  1. 1.Université de Lyon, Université Claude Bernard Lyon 1, CNRS UMR 5256 (IRCELYON)Villeurbanne CedexFrance
  2. 2.University of Central Florida, NanoScience Technology CenterOrlandoUSA
  3. 3.Université de Lyon, Université Claude Bernard Lyon 1, CNRS UMR 5180 (LSA)Villeurbanne CedexFrance
  4. 4.Universités d’Aix-Marseille I, II et III, CNRS UMR 6264 (LCP)Marseille Cedex 3France

Personalised recommendations