Effect of organic matter and iron oxides on quaternary herbicide sorption–desorption in vineyard-devoted soils

Journal of Colloid and Interface Science 333 (2009) 431–438

Contents lists available at ScienceDirect

Journal of Colloid and Interface Science www.elsevier.com/locate/jcis

Effect of organic matter and iron oxides on quaternary herbicide sorption–desorption in vineyard-devoted soils Mirian Pateiro-Moure a , Cristina Pérez-Novo a , Manuel Arias-Estévez a , Raquel Rial-Otero b , Jesús Simal-Gándara b,∗ a b

Soil and Agricultural Science Group, Plant Biology and Soil Science Department, Faculty of Food Science and Technology, University of Vigo, Ourense Campus, E32400 Ourense, Spain Nutrition and Bromatology Group, Analytical and Food Chemistry Department, Faculty of Food Science and Technology, University of Vigo, Ourense Campus, E32400 Ourense, Spain

a r t i c l e

i n f o

a b s t r a c t

Article history: Received 8 November 2008 Accepted 7 February 2009 Available online 12 February 2009

Herbicide soil/solution distribution coefficients (K d ) are used in mathematical models to predict the movement of herbicides in soil and groundwater. Herbicides bind to various soil constituents to differing degrees. The universal soil colloid that binds most herbicides is organic matter; however metallic hydrous oxides might also have some influence. The adsorption–desorption of three quaternary ammonium herbicides on soils with different chemical–physical characteristics was determined using a batch equilibration method before and after the following sequential selective dissolution procedures: removal of organic matter, and removal of organic matter plus free iron oxides. The experimentation involved paraquat (PQ), diquat (DQ) and difenzoquat (DFQ) herbicides. The distribution coefficients (K d ) of the molecules and their correlation to the soil components were determined and a significant negative correlation with organic carbon was highlighted (r < −0.610, p < 0.035, n = 12). All quats cations experiment high adsorption in the control soils with a Zeta potential at about −21 mV. The order of adsorption on soils (based on K d ) was the following: PQ > DQ  DFQ. The adsorption isotherms of these three herbicides on the natural and processed soils were satisfactorily fitted with the Freundlich equation, and a significant correlation with organic carbon was highlighted for quats K F (r < −0.696, p < 0.012, n = 12). The removal of organic matter from soils seems to leave free new adsorption sites for quats on the clay surface, which is no longer occluded by organic matter. This work shows that the amount and nature of the surface that remains available after the removal of single soil constituents is a critical parameter in determining the sorptive behavior of cationic contaminants. © 2009 Elsevier Inc. All rights reserved.

Keywords: Quaternary ammonium herbicides Organic matter Iron oxides Sorption/desorption Vineyard-devoted soils

1. Introduction Quaternary nitrogen herbicides were developed from the observation that quaternary ammonium germicides, like cetyl trimethylammonium bromide, desiccate young plants [1,2]. Table 1 shows the chemical structure, chemical and common names, together with the physical–chemical properties of the studied ammonium quaternary herbicides. Diquat (DQ) and paraquat (PQ) are included in a priority list of herbicides of potential concern established for the Mediterranean countries by the European Union (EU), due to their widespread usage in this area [3]. Consequently, they may be present as residues in environmental, food and biological samples [4,5]. Diquat and paraquat are extremely toxic and are often encountered in cases of poisoning [6,7]. The pyrazolium monocation difenzoquat (DFQ), the active ingredient in registered trademarks of American Cyanamide Co, is also used throughout the world as a

*

Corresponding author. Fax: +34 988 387001. E-mail address: [email protected] (J. Simal-Gándara).

0021-9797/$ – see front matter doi:10.1016/j.jcis.2009.02.019

©

2009 Elsevier Inc. All rights reserved.

selective herbicide for post emergence control of wild oats in barley and fall-seeded wheat. The adsorption and desorption of these compounds in heterogeneous soils govern their fate. Soil adsorption of bipyridinium herbicides (largely PQ and DQ) has been studied since the 1960s and 1970s. Tests have revealed that both are strongly adsorbed by soils and soil clays [8–11]. Also, according to Calderbank [12], some clays in soil dramatically reduce the herbicidal power of both compounds. More recent studies have shown that these herbicides have a high affinity for clay surfaces in relation to soil organic matter, especially as compared with inorganic cations [13–17]. Their interactions with clay particles depend on the particular type of clay. Thus, Weber and Scott [18] found PQ to bind to interlayer spacings in montmorillonites via Coulombic and van der Waals forces, and also to kaolinite via Coulombic forces alone. This work is part of an ongoing research project intended to increase available knowledge about the specific factors affecting adsorption–desorption processes, and hence pesticide mobility, in crop soils [19–21]. In this work, we studied the sorption behavior of the three quats (PQ, DQ and DFQ) on a range of different soils, H2 O2 -treated soils and dithionite–citrate-treated soils in order to

432

M. Pateiro-Moure et al. / Journal of Colloid and Interface Science 333 (2009) 431–438

Table 1 Characteristics of the quaternary herbicides studied. Common name

PQ

DQ

DFQ

Chemical structure CH3 SO− 4 Name CAS no . MWa Sb (g L−1 ) log P ow c Kocd Soil half-lifee a b c d e

1,1 -Dimethyl-4,4 -bipyridinium dichloride 1910-42-5 257.2 620 −4.5 15 to 51 644

1,1 -Ethylene-2,2 -bipyridyldiylium dibromide 85-00-7 344 677 −4.6 164 to 134 3450

1,2-Dimethyl-3,5-diphenyl-1H-pyrazolium methyl sulphate 43222-48-6 362.4 740 0.3 23 to 36 6810

Molecular weight. Solubility in water 20 ◦ C. Octanol/water partition coefficient at 20 ◦ C. −1 or L kg−1 ). Partition coefficient normalized to organic carbon content (mL goc oc Aerobic soil half-life (Average, days). Data were obtained from Kenaga [45], Haag and Yao [46], and US-EPA [47].

determine the influence of the organic matter and iron oxides. The treatments seem to create high-affinity sites for quats on the fine clays, in good agreement with the findings by Moyer and Lindwall [22], who state that the environmental impact of paraquat is highly influenced by their interactions with the clay components of the soil. The objectives of this study are to: (1) assess the relative importance and contributions of various soil colloidal components to the sorption of quaternary herbicides, and (2) understand the role of organic matter and free Fe oxides associated with clay and their interactions on the sorption and desorption of quats.

Table 2 General properties of the control (a), H2 O2 -treated (b), and dithionite–citratetreated (c) soils.a Soil

pH(H2 O)

pH(KCl)

%Sand

%Silt

%Clay

%OC

CEC

1

a b c

5.3 6.4 6.7

3 .6 5 .3 5 .1

70.0

17.0

13.0

0.8 0.06 0.06

3.07

3.7 3.5 0.45

2

a b c

5.3 6.2 7.1

4 .0 5 .4 5 .7

64.0

21.0

14.0

1.9 0.32 0.22

3.96

5.0 8.8 0.86

3

a b c

5.5 6.9 6.7

4 .3 6 .4 5 .1

73.0

16.0

11.0

1.4 0.10 0.08

3.81

9.8 3.3 0.60

4

a b c

6.2 6.9 6.8

4 .9 6 .1 5 .3

40.0

38.0

22.0

1.1 0.13 0.08

4.31

26.7 32.4 0.23

2. Materials and methods 2.1. Soils Samples of soil were obtained from the 0–5 cm deep layer of vineyard-devoted soils developed on granite and schist materials in Galicia (NW Spain). Five replicates of each soil were collected within 0.5 m of each other and pooled. Once in the laboratory, the soils were air-dried at room temperature, passed through a 2 mm mesh sieve, homogenized and stored until analysis. Soils were treated with H2 O2 and dithionite–citrate according to the following protocols: 2.1.1. Peroxide-treated samples About 50 g of dry sample were saturated with 6% H2 O2 for 30 days in order to completely remove organic matter. Sample was then washed out with 200 mL of a 0.01 M CaCl2 solution for 60 min and centrifuged at 4000 rpm for 10 min. This washing process was repeated 8 times. Finally, samples were air-dried. This methodology was similarly to Agbenin and Olojo [23].

Fed

a OC: organic carbon; CEC: cation-exchange capacity (cmol(c) kg−1 ); Fed : dithionite–citrate extractable iron (g kg−1 ).

sand (2–0.05 mm), silt (0.05–0.002 mm) and clay (<0.002 mm) were determined by using the wet sieving and pipette methods [24]. Exchangeable cations were extracted with 0.2 M NH4 Cl and determined by atomic absorption spectroscopy (Cae and Mge ) or flame emission spectrometry (Nae and Ke ), using a Thermo Solar M series spectrometer (Austin, EE.UU). Exchangeable aluminum was quantified by displacement with 1 M KCl, followed by atomic absorption spectrophotometry. Finally, the cation exchange capacity, CEC, was determined as the combination of bases (Nae , Ke , Cae and Mge ) and exchangeable Al. The amount of dithionite–citrate extractable iron (Fed ) was determined according to Holmgren [25]. The characteristics of the soils are summarized in Table 2. 2.2. Quats extraction

2.1.2. Dithionite–citrate-treated samples At about 0.8 g of dry sample were mixed with a solution of 50 mL of sodium citrate (160 g L−1 ) plus 0.8 g sodium dithionite (Na2 S2 O4 ), shaked for 16 h, centrifuged at 4000 rpm for 10 min, and filtered. Sample was then washed out with 200 mL of a 0.01 M CaCl2 solution for 60 min and centrifuged at 4000 rpm for 10 min. This washing process was repeated 8 times. Finally, samples were air-dried. The pH of 1:2.5 (w/w) suspensions of soil in water or 0.1 M KCl was measured with a combined glass electrode. The organic carbon content was determined by elemental analysis on a ThermoFinnigan 1112 Series NC instrument (Austin, EE.UU). The proportions of

Paraquat, diquat and difenzoquat in the soils were extracted by a method developed by us with analyte recoveries of about 100% [26]. Mean recovery rates for the herbicides in soils ranged from 98 to 102%, whereas relative standard deviations were lower than 15%. Quantification limits in soils ranged from 10 μg kg−1 for DFQ to 20 μg kg−1 for DQ and PQ. The concentration of quats in the native soil samples was lower than detection limits, except for sample 1 with PQ at the level of 43 μg kg−1 . A soil sample of 5 g was extracted for 3 h with 30 mL of a 70:30 (v/v) mixture of methanol/5% (w/v) ethylenediaminetetraacetic acid disodium (EDTA), which was previously acidified

M. Pateiro-Moure et al. / Journal of Colloid and Interface Science 333 (2009) 431–438

433

2.5. Sorption–desorption experiments

Fig. 1. Zeta potential (mV) for control soil 1 (sample a) at different pHs and H2 O2 treated sample (sample b). Vertical bars show 2 standard deviations around the mean.

with the addition of 2% (v/v) formic acid. The resulting extract was concentrated to 4 mL following pH adjustment to 9–10 with 3 N NaOH and cleaned up by passage through silica cartridges and elution with 10 mL of 70:30 (v/v) methanol/6.5 M HCl [27]. Finally, the extract was evaporated to dryness and filled up to a volume of 1 mL with a mixture of 100 mM ammonium formate adjusted at pH 3 with formic acid and 15 mM heptafluorobutyric acid (HFBA) before analysis by liquid chromatography.

2.5.1. Adsorption A 1 g sample of each soil was suspended in 10 mL of herbicide solution (1550–31100 μmol(+) kg−1 for PQ, 1160–23,260 μmol(+) kg−1 for DQ, and 69–2760 μmol(+) kg−1 for DFQ, all containing 0.01 M CaCl2 ). The individual herbicide solutions were prepared from 3.888, 2.907, and 1.380 mM stock solutions of PQ, DQ, and DFQ, respectively. Suspensions were shaken on a rotary shaker at 200 rpm at room temperature (20 ± 2 ◦ C) for 24 h and then centrifuged at 4000 rpm for 10 min. Solutions were added to pH 5.5. After equilibration between soil and adsorption solution, the solutions were still at such a pH (5.8 ± 0.3, for n = 510 measurements for many of the experiments with any of the quats). The resulting supernatant was analyzed by liquid chromatography according to the protocol described above. The amounts of PQ, DQ, and DFQ adsorbed were calculated as the differences between those initially present in solution and those remaining after centrifugation. All measurements were made in duplicate. The fitting calculations were performed using a non-linear parameter fitting algorithm, and the data were fitted to the Freundlich (2), Langmuir (3) and Linear (4) equations, which are described by: 1/n

2.3. Liquid chromatography determination Liquid chromatography system was a Dionex Corporation (Sunnyvale, USA) including a P680 quaternary pump, an ASI-100 autosampler, a TCC-100 thermostatted column compartment and a UVD170U detector. Separations were done on a Luna C18 column (150 mm long; 4.60 mm i.d.; 5 μm particle size) obtained from Phenomenex (Madrid, Spain) and a guard column (4 mm long; 2 mm i.d.; 5 μm particle size) packed with the same material. The mobile phases used were water containing 100 mM ammonium formate adjusted at pH 3 with formic acid and 15 mM HFBA (A), methanol (B) and isopropanol (C). The gradient was: 90% A and 10% B for 3 min, change to 10% A and 90% B in 2 min, hold 5 min, change to 10% A, 40% B and 50% C in 2 min and hold 4 min, change to 90% A and 10% B in 0.1 min and hold 10 min. The total analysis time was 26 min. The injected volume was 50 μL and the LC flow-rate 0.7 mL min−1 . The wavelengths used for detection were 258 nm for PQ and DFQ, and 310 nm for DQ.

Q e = KFCe ,

(2)

Q e = K C e Q m /1 + K C e ,

(3)

Q e = bC e + a,

(4)

where Q e is the concentration of sorbate sorbed at equilibrium (μmol(+) kg−1 ); C e is the concentration of sorbate in the aque(1−n)

ous phase at equilibrium (μmol(+) L−1 ); K F (Ln kg−1 μmol(+) ) and 1/n (dimensionless) are the Freundlich coefficients; K F and n are constants characterizing the adsorption capacity and intensity, respectively. For n = 1 a linear sorption isotherm is obtained and the sorption coefficient K F is identical with the distribution coefficient K d . K (L μmol−1 ) is a Langmuir constant related to the energy of adsorption, and Q m (μmol kg−1 ) is the maximum adsorption capacity of the sample. Both a (intercept) and b (slope) are the linear coefficients. Sorption partition coefficients (K d ) at a selected quat concentration are calculated by mean the following expression: adsorbed quat (μmol(+) kg−1 )

2.4. Clay extraction and electrokinetic measurements

Kd =

The Zeta potential of the clay fraction of soil 1 (50 g treated and not treated with H2 O2 ) was measured following extraction of this fraction by centrifugation for a time t calculated from the equation:

The data shown in the sorption isotherms and K d represent mean values calculated from duplicate sorption experiments. Basic and descriptive statistics, and linear correlation analyses were performed using SPSS v. 16.0 for Windows.

t=

9η L 2r 2 (ρs − ρl ) g

,

equilibrium quat concentration (μmol(+) L−1 )

.

(5)

(1)

where L is the depth of the medium above the pellet (7.4 cm), η the viscosity of water at 25 ◦ C (0.01008 dyn s cm−2 ), r the mean particle radius, ρs the density of the particles (2.6 g cm−3 ), ρl the density of water (1.0 g cm−3 ), and g the standard gravity (980 cm s−2 ). 1000 mg clay samples were suspended in 0.01 M NaNO3 , brought to a desired pH in the range 2–12 by addition of HNO3 or NaOH (final suspension volume 10 mL), and shaken for 1 h. A 10 mL sample was then pipetted out, diluted and run in a Malvern ZEN 3600 micro-electrophoresis apparatus. At least ten replicate runs were performed for each system. Zeta potential (mV) for a typical control soil is about −21 mV for the soil solution pH (Fig. 1).

2.5.2. Desorption Once the adsorption process was finished after 24 h, the sample was centrifuged at 4000 rpm for 10 min and the liquid phase in each centrifuge tube was removed. Afterward, centrifuge tubes were weighted in order to know the interstitial herbicide amount. Then, 10 mL of a 0.01 M CaCl2 solution (pH 5.5) without herbicide were added and the sample was processed as described in the previous subsection. The procedure was repeated five times (six steps in all). The concentrations used in the tests were 7780 and 15,550 μmol(+) kg−1 for PQ, 5810 and 11,630 μmol(+) kg−1 for DQ, and 550 and 1110 μmol(+) kg−1 for DFQ. Desorption data are given as percentages of the initial amounts adsorbed. All measurements were made in duplicate.

434

M. Pateiro-Moure et al. / Journal of Colloid and Interface Science 333 (2009) 431–438

Table 3 Fitting parameters (average ± standard deviations) obtained from the Freundlich equations as applied to PQ, DQ and DFQ adsorption together with K d (L kg−1 ) after incubation with a herbicide solution of 80 mg L−1 for PQ and DQ and a solution of 40 mg L−1 for DFQ, in the control (a), H2 O2 -treated (b) and dithionite–citrate-treated (c) soils. PQ

Soil

KF

DQ

DFQ

1/n

R2

Kd

KF

1/n

R2

Kd

KF

1/n

R2

Kd

1

a b c

– – 5038 ± 2637

– – 0.23 ± 0.08

– – 0.902

2516 3840 1783

3403 ± 211 4799 ± 562 3381 ± 599

0.21 ± 0.01 0.20 ± 0.02 0.24 ± 0.03

0.992 0.96 0.975

1245 4532 1783

165 ± 20 345 ± 40 210 ± 50

0.68 ± 0.03 0.46 ± 0.06 0.60 ± 0.07

0.977 0.968 0.984

35 107 67

2

a b c

4148 ± 991 4370 ± 1816 7401 ± 2208

0.26 ± 0.06 0.37 ± 0.14 0.17 ± 0.05

0.887 0.776 0.908

1520 2674 3512

2600 ± 225 4404 ± 538 3912 ± 730

0.24 ± 0.02 0.18 ± 0.02 0.21 ± 0.03

0.985 0.952 0.946

596 2684 3053

104 ± 20 208 ± 14 272 ± 52

0.74 ± 0.05 0.57 ± 0.03 0.57 ± 0.06

0.993 0.996 0.985

34 70 86

3

a b c

4456 ± 973 5176 ± 2027 7673 ± 2370

0.22 ± 0.05 0.28 ± 0.12 0.15 ± 0.05

0.887 0.773 0.891

1547 3423 3609

2957 ± 399 3717 ± 378 4000 ± 657

0.19 ± 0.02 0.19 ± 0.02 0.23 ± 0.03

0.937 0.968 0.957

708 1642 1633

124 ± 17 299 ± 77 313 ± 69

0.72 ± 0.04 0.49 ± 0.13 0.54 ± 0.07

0.996 0.902 0.981

46 85 95

4

a b c

4150 ± 929 4816 ± 341 7631 ± 1991

0.31 ± 0.06 0.18 ± 0.03 0.16 ± 0.04

0.906 0.943 0.932

1756 1833 2608

2662 ± 292 3599 ± 491 3265 ± 401

0.26 ± 0.02 0.22 ± 0.02 0.25 ± 0.02

0.978 0.949 0.978

692 1579 1389

29 ± 4 169 ± 50 142 ± 44

0.95 ± 0.03 0.53 ± 0.11 0.63 ± 0.08

0.999 0.943 0.979

23 38 43

a

a

– Sorption close to 100%.

3. Results and discussion 3.1. General characteristics of samples Soils were acid (pH(H2 O) 5.3–6.2); their textures were moving from loam (sample 4a in Table 2) to sandy loam (samples 1– 3a); and their organic carbon contents were in the range 0.8–1.9%, whereas the iron oxides contents were from 3.7 to 26.7 g kg−1 . All these values are typical from soils devoted to vineyards [28]. The treatment with H2 O2 reduced their contents in organic carbon to values lower than 0.32%, whereas pH(H2O) was increasing in 1 unit as an average. The treatment with dithionite–citrate produced a similar effect both in organic carbon and pH, but reduced also the free iron oxides content (<0.86 g kg−1 ). 3.2. Sorption of quats onto samples The values of the distribution coefficients were calculated by K d = Q e /C e (where Q e is the adsorbed quat concentration in μmol kg−1 , and C e is the concentration of the quat in solution at equilibrium in μmol L−1 ). These K d values found were either high or moderate for the different soils and treatments (Table 3), pointing the low mobility potential of these herbicides. According to K d values obtained the order of adsorption on control soils was the following: PQ (1520–2516 L kg−1 ) > DQ (596–1245 L kg−1 )  DFQ (23–46 L kg−1 ). Our previous findings [26] showed lower values for PQ (106–1280 L kg−1 ), similar values for DQ (418– 1010 L kg−1 ), and higher values for DFQ (28–1370 L kg−1 ). The distribution coefficients (K d ) of the molecules and their correlation to the soil components were determined and a significant correlation with organic carbon was highlighted (r < −0.610, p < 0.035, n = 12). Freundlich, Langmuir and Linear equations have been employed to describe the adsorption behavior of quats. Fittings with Freundlich equation were shown in Table 3. The sorption isotherms were mainly of the L-type (Figs. 2–4), which according to Giles et al. [29] and Limousin et al. [30] correspond to a decrease of sites availability as the solution concentration increases, what indicates saturation of the surfaces. For PQ, it generally shows high-affinity class H-type adsorption isotherms, an extreme version of the Ltype isotherms (Fig. 2), being its adsorption in many instances close to 100%, particularly for the lowest concentrations; R 2 was ranging between 0.77 and 0.94 (Table 3). Initially, the adsorption of paraquat onto the soil adsorbent is very high; this is then followed by a lower adsorption, and gradually approaches a plateau. However, Langmuir fittings were not satisfactory according to parameter associated errors and R 2 values, what is indicative of no

saturation of the sorption surfaces even at high PQ concentrations. Given the PQ curve shapes linear fitting was obviously not necessary to be performed. DQ sorption curves can be found in Fig. 3. DQ behavior was similar to that of PQ and gave satisfactory fittings with Freundlich equation (R 2 between 0.94 and 0.99; Table 3). As for PQ, Langmuir and linear equation fittings were not satisfactory due to the large errors associated to the estimated parameters and low values of R 2 . However, DFQ sorption curves were different and showed a trend to linearity (sample 4a; Fig. 4). DFQ fittings were satisfactory both for Freundlich equation (R 2 between 0.90 and 0.99; Table 3) and for the linear equation (R 2 between 0.78 and 1.00). The Freundlich coefficient K F for PQ, DQ and DFQ in control soils was, respectively, >4000, 2600–3500, and 30–165 (Table 3: samples “a”). K F coefficient for PQ in the soils increased after H2 O2 treatment (“b” samples, with values ranging 4300–5200), but this increment was still higher after dithionite–citrate treatment (“c” samples, with values such as 7400–7700). K F coefficients for DQ and DFQ generally increased for both treatments (samples b and c) with regards to control soils (K F 3200–4800 and 140–350 for DQ and DFQ, respectively; Table 3). The lower sorption in the control soils for quats could be a consequence of the liberation of adsorption sites of the clay surface coating with the organic matter (OM) extraction, increasing then sorption with regards to native soils, since a significant correlation with organic carbon was highlighted for quats K F (r < −0.696, p < 0.012, n = 12). According to the work of Cavallaro and McBride [31] on cationic metals sorption by and acid soil clay, possibly, organic matter removal effected more efficient dispersion of the clays and oxides, exposing additional adsorption sites. This stresses the importance of organic matter occluding clay adsorption sites for quats, especially for those least sorbing quats (DQ and DFQ) in soil. The Freundlich coefficient n for PQ, DQ and DFQ in all samples followed the sequence DFQ > PQ  DQ (Table 3). For DFQ, the n values are higher for control samples “a”, without treatment, than for those obtained with “b” and “c” treatments. The maximum sorption reached for an input of 3110 μmol L−1 was observed at 20,000–25,000 μmol(+) kg−1 for PQ (almost the 80% of added PQ), and 15,000–18,500 μmol(+) kg−1 for an input of 2326 μmol L−1 for DQ (almost the 80% of added DQ), whereas for DFQ it was not found the sorption maximum even at 2500 μmol(+) kg−1 for an input of 276 μmol L−1 (Figs. 2–4). In any case, maxima sorption reached for all quats are about 70– 80% at the highest levels added to soils. For low levels added, it is also clear that a low percent (5%) of no sorption is reached at much higher concentrations for PQ (1555 μmol(+) L−1 ), followed by DQ (1163 μmol(+) L−1 ), and at much lower concentrations for DFQ (7 μmol(+) L−1 ), what is a clear indication of the different

M. Pateiro-Moure et al. / Journal of Colloid and Interface Science 333 (2009) 431–438

435

Fig. 2. Adsorption isotherms for paraquat in the control (a), H2 O2 -treated (b) and dithionite–citrate-treated (c) soils.

Fig. 3. Adsorption isotherms for diquat in the control (a), H2 O2 -treated (b) and dithionite–citrate-treated (c) soils.

soil binding-strength affinity of all quats. The process of sorption is greatly enhanced in expanding-lattice clay by the ability of the much more planar dication quat molecules (PQ/DQ) to become intercalated between the lattice layers and then be held by strong Coulombic forces [32]. The study of the factors controlling quats mobility is a substantial contemporary issue to guarantee environmental quality. The soils studied were selected for their diversity in physicochemical properties which could influence the sorption of quats. Soil 2 has high organic matter content (Table 2). Soil 4 has a large proportion of fine particles such as silt and clay (38 and 22%, respectively) vs. sand (40%). The low organic matter content (Table 2) of soil 1 may explain its highest measured adsorption potential, since quats ad-

sorption sites on the clay surface are not occluded by the organic matter, and clay resulted to be the main factor controlling quats adsorption onto soils. It could be possible that, at higher pHs, the carboxylic moieties of the humic acids (HA) were deprotonated and stronger cation attraction could occur (Fig. 1). Moreover, such an effect is more important on clays (i.e., on soils treated with H2 O2 to remove organic matter; Fig. 1). Those treatments for organic matter removal alone or with iron oxides provoked an increase of pH in 1–2 units (Table 2). However, the effect of increasing pH within the range 4.7–11.4 on quats sorption onto soils (Table 4) was not important and dependent on the quat: DQ is unstable in aqueous solution at pH >6.5 and a colored reaction product is formed,

436

M. Pateiro-Moure et al. / Journal of Colloid and Interface Science 333 (2009) 431–438

Fig. 5. Desorption percent in the control (a), H2 O2 -treated (b) and dithionite–citratetreated (c) soils.

Fig. 4. Adsorption isotherms for difenzoquat in the control (a), H2 O2 -treated (b) and dithionite–citrate-treated (c) soils.

whereas PQ and DFQ sorption does not seem to be affected by pH. Non-specific interactions such as π –π interactions, London–van der Waals forces, or hydrophobic adsorption effects by soil organic matter can play a role in quats retention. And this type of interactions can be stronger in the case of clays: the presence of activated C–H bond from the methyl groups in quats structure can form hydrogen bonds (as C–H· · ·O bonding) with oxygens of the siloxane surface of silicate clays [33]. Independently of soil pH, despite the presence of negative charges on the organic surfaces, sorption of quats clearly increased with the organic matter removal from soils. In such a way, humification of OM has a direct influence on the physical, chemical, and biological properties of the organic constituents that may interact with exogenous organic chemicals [34].

The dynamics of OM transformation may affect the location of these interactions in the soil and the accessibility of the sorption sites [35]. Since vineyard soils have been identified as the most erodible agricultural soils [36], the transport of organic matter-rich particles by erosion can seriously affect the retention/mobilization of quats in those soils. All this helps to support previous evidence that the major factors governing the sorption of PQ are both the solid state organic fraction [37] and the clay mineral content [38]. It was seen that DQ was found to be unstable in aqueous solution at pH >6.5. Florêncio et al. [39] investigated the stability of diquat in aqueous solutions: in the 3–8 pH range no significant variation occurred, but a pronounced decreased for DQ was found at higher pHs. It was also observed that the color of diquat solutions changes to yellow in strong alkaline medium [40]. This change corresponds to an irreversible degradation process, since once the pH of the solution is lowered again, the absorption spectrum is not the same as the original. However, the appearance of the color in the solutions occurs even when the dissolved oxygen is carefully removed by purging with N2 the solution previous to the addition of diquat, indicating an initial process different to oxidation. Anyway, by different voltammetric techniques, one oxidation wave appears at potential values near to 0.3 V. 3.3. Desorption of quats from samples The strong binding of paraquat to the studied soils was indicated by its little amount desorbed from them that ranged in 2.1– 2.4% and 10.5–14.3% at 100 and 200 mg L−1 (Fig. 5: sample “a”),

M. Pateiro-Moure et al. / Journal of Colloid and Interface Science 333 (2009) 431–438

437

Table 4 PQ, DQ and DFQ adsorption (μmol(+) kg−1 ) in the control soils as a function of pH amended with Ca(OH)2 0.005 M and Ca(OH)2 0.01 M. Soil

CaCl2 0.01 M

Ca(OH)2 0.005 M

Ca(OH)2 0.01 M

Ce (μmol(+) L−1 )

C ads (μmol(+) kg−1 )

pH

Ce (μmol(+) L−1 )

C ads (μmol(+) kg−1 )

pH

Ce (μmol(+) L−1 )

C ads (μmol(+) kg−1 )

pH

PQ

1 2 3 4

7 .1 9 .5 10.9 6 .7

8068.3 8001.5 7256.2 7352.6

4.20 4.86 4.70 5.35

9 .3 15.6 18.0 13.7

7350.8 7335.2 7258.4 7344.3

7.73 6.86 7.04 6.86

14.6 21.1 20.1 16.1

8101.6 8039.7 8072.1 8090.6

11.46 10.18 8.37 10.90

DQ

1 2 3 4

10.2 20.1 15.7 12.4

6226.6 6132.8 6052.6 6120.8

4.24 4.85 4.73 5.26

–a – – –

– – – –

7.15 6.65 6.95 6.64

– – – –

– – – –

11.25 10.22 7.88 10.57

DFQ

1 2 3 4

21.6 22.5 17.6 30.3

766.7 766.2 821.7 693.7

4.74 5.00 5.84 6.10

14.4 22.2 18.6 32.7

6.53 6.60 6.60 6.99

20.1 25.9 19.8 36.1

a

840.3 838.4 840.0 666.6

768.1 685.5 710.9 527.9

11.41 9.78 8.03 10.51

– Sorption at 100%.

trations sorbed. At the lowest concentrations sorbed both PQ and DQ experimented very little reductions in their already reduced desorption potential with any of the treatments. In any case, desorption is already so low for both that it was found a practically “zero” slope for a linear relationship between desorption (%) and corresponding K d (Fig. 6). Therefore, the observed constraint in desorption of paraquat and diquat suggests that ion exchange is not a major mechanism for their retention in the 0.01 M CaCl2 at pH 5. This is usually indicative of chemisorption reported by McBride [41]. The sorption of these dication species can be reasonably speculated to occur in two steps: (i) transfer of sorbate from the aqueous solution to the sites on the sorbent; and (ii) chemical complexation at these sites [42]. Instead, the largest variability in DFQ desorption (%) is someway related to K d , what indicates that its desorption is controlled by the affinity with which it was previously sorbed. This kind of adsorption–desorption relationships have been described for metals such as Cd [43,44]. 4. Conclusions

Fig. 6. Desorption percent in the control and treated soils vs. their quat K d .

respectively, with regards to DQ (6.2–6.7% and 19.4–20.8% at 100 and 200 mg L−1 , respectively) and DFQ (25.8–46.7% and 35.8–55.4% at 20 and 40 mg L−1 , respectively). The order of desorption on soils was following the same trend than adsorption: PQ < DQ  DFQ. Most of the samples had the highest percents of desorption after the removal of the organic matter at the highest quats concen-

Herbicide pollution in crop soils is produced by their application inputs. The mobility of quats depends on their retention by soil and on their stability in soil solution. It is difficult to describe the complexity of the interactions among cation molecules, such as quats, and soils, since different soil attributes and sorption mechanisms are acting simultaneously. However, this research cleared the contribution of clay to the sorption of quats in such soils, in the surface layers. It was found that quats adsorption to soils (K d ) decrease in the following order: PQ > DQ  DFQ. Anyway, quats cations experiment high adsorption in these soils with negative Zeta potential. While the best fits for quats adsorption to soils were reached with the Freundlich equation, Langmuir fittings were associated to much higher errors and were discarded. There was a high correlation amongst all quats sorption (K d and K F ) onto the soils. Moreover, quats sorption onto the soils was negatively correlated with their OC% (both K d and K F ). DQ was the quat clearly showing to be unstable in aqueous solution at pH >6.5, and a colored reaction product is formed. This change corresponds to an irreversible degradation process, since once the pH of the solution is again lowered, the absorption spectrum is not the same as the original. Regarding quats sorption, PQ adsorption to soils is maxima (≈100%) under any conditions, but DQ and DFQ adsorption (K F ) increases when removing organic matter and organic matter + Fed in a similar manner. Therefore, iron oxides are not considered to be relevant for quats sorption in arable soils with neutral to slightly acidic pH. The removal of organic matter from soils seems to leave free new adsorption sites for quats, since the increase in soil pH

438

M. Pateiro-Moure et al. / Journal of Colloid and Interface Science 333 (2009) 431–438

does not provoke a higher quats adsorption. The opposite for quats adsorption to soils occurs with quats desorption. Quats desorption to soils (%) decrease in the following order: DFQ  DQ > PQ. In any case, desorption increase with quats sorbed concentration, and this effect is higher for PQ and DQ (PQ > DQ). Clay can be considered an important sorbent for organic environmental chemicals such as quaternary ammonium herbicides and should be investigated further in this respect since it could prevent the herbicides from leaching and subsequent contamination of groundwater. Therefore, most of the sorption capacity in the soil was related to clay, which indicates that sorption processes are magnified by changes in organic compound composition due to weathering. Moreover, addition of organic matter-based fertilizer amendments may enhance the potential for off-site leaching of recently applied cationic pesticides. In subsurface soils (with much lower organic matter content), Fe oxides could play an important role in overall sorption, since it dictates the net electrical charge as well as the electric potential of the soils.

[13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30]

Acknowledgments This work was supported by Spain’s Ministry of Science and Technology under contract AGL2007-62075. M. Pateiro-Moure and R. Rial-Otero were funded, respectively, by the María Barbeito and Isidro Parga Pondal research programs from the Galicia’s Council of Innovation and Industry.

[31] [32] [33] [34] [35] [36]

References [1] C. De Liñán y Vicente, Farmacología Vegetal, Compendio de las materias activas más interesantes en fitopatología. Su toxicología y empleo, Agrotécnicas, Madrid, 1997. [2] C. De Liñán y Vicente, Vademecum de productos fitosanitarios y nutricionales, Agrotécnicas, Madrid, 1997. [3] D. Barceló (Ed.), Environmental Analysis Techniques, Applications and Quality Assurance, Elsevier, Amsterdam, 1993, p. 149. [4] M.T. Galcerán, M.C. Carneiro, L. Puignou, Chromatographia 39 (1994) 581. [5] J.M. Zen, S.H. Jeng, H.J. Chen, Anal. Chem. 68 (1996) 498. [6] A.M. Tsakis, K. Perakis, E. Koumantakis, Veterinary Human Toxicol. 38 (1996) 113. [7] S.M. Pond, in: L.R. Goldfrank, N.E. Flomenbaum, N.A. Lewin, R.S. Weisman, M.A. Howland, R. Hoffman (Eds.), Goldfrank’s Toxicologic Emergencies, Appelton & Lange, Stanford, 1998, p. 1475. [8] J.B. Weber, P.W. Perry, R.P. Upchurch, Soil Sci. Soc. Am. Proc. 29 (1965) 678. [9] G.E. Coats, H.H. Funderburk, J.M. Lawrence, D.E. Davis, Weed. Res. 6 (1966) 58. [10] B.A.G. Knight, T.E. Tomlinson, J. Soil Sci. 18 (1967) 233. [11] G. Basile, M. Arienzo, G. Celerio, Agrochimica 34 (1990) 181. [12] A. Calderbank, Adv. Pest Control Res. 8 (1968) 127.

[37]

[38] [39] [40] [41] [42] [43] [44] [45] [46] [47]

G. Rytwo, S. Nir, L. Margulies, J. Colloid Interface Sci. 181 (1996) 551. G. Rytwo, S. Nir, L. Margulies, Soil Sci. Soc. Am. J. 60 (1996) 601. G. Rytwo, D. Tropp, C. Serban, Appl. Clay Sci. 20 (2002) 273. G. Rytwo, M. Tavasi, S. Afuta, S. Nir, Appl. Clay Sci. 24 (2004) 149. Y. Sequi, K. Yurdakocü, J. Colloid Interface Sci. 287 (2005) 1. J.B. Weber, D.C. Scott, Science 152 (1966) 1400. M. Arias-Estévez, A.C. Torrente, E. López-Periago, B. Soto-González, J. SimalGándara, J. Agric. Food Chem. 53 (2005) 5675. R. Rial-Otero, B. Cancho Grande, C. Pérez-Lamela, J. Simal-Gándara, M. AriasEstévez, J. Chromatogr. Sci. 44 (2006) 539. G.C. López-Pérez, M. Arias-Estévez, E. López-Periago, B. Soto-González, B. Cancho-Grande, J. Simal-Gándara, J. Agric. Food Chem. 54 (2006) 1797. J.R. Moyer, C.W. Lindwall, Can. J. Soil Sci. 65 (1985) 523. J.O. Agbenin, L.A. Olojo, Geoderma 119 (2004) 85. G.W. Gee, J.W. Bauder, in: A. Klute (Ed.), Methods of Soil Analysis, in: SSSA Book Ser. 471, vol. 5, ASA and SSSA, Madison, WI, 1986, pp. 383–411. G.G. Holmgren, Soil Sci. Soc. Am. Proc. 31 (1967) 210. M. Pateiro-Moure, C. Pérez-Novo, M. Arias-Estévez, E. López-Periago, E. Martínez-Carballo, J. Simal-Gándara, J. Agric. Food Chem. 55 (2007) 6219. T.M.P. Chichila, S.M. Walters, J. AOAC Int. 74 (1991) 961. M. Arias, E. López, D. Fernández, B. Soto, Soil Sci. 169 (2004) 796. C.H. Giles, T.H. MacEwan, S.N. Nekhwa, D. Smith, J. Chem. Soc. 3 (1960) 3973. G. Limousin, J.P. Gaudet, L. Charlet, S. Szenknect, V. Barthès, M. Krimissa, Appl. Geochem. 22 (2007) 249. N. Cavallaro, M.B. McBride, Soil Sci. Soc. Am. J. 48 (1984) 1050. R.H. Bromilow, Pest. Manage. Sci. 60 (2003) 340. Z.Y. Hseu, S.H. Jien, S.F. Cheng, J. Environ. Sci. Health B 38 (2003) 441. K. Haider, in: G. Stotzky, J.M. Bollag (Eds.), Soil Biochemistry, Marcel Dekker, New York, 1993, p. 55. G. Almendros, Eur. J. Soil Sci. 46 (1995) 287. C. Kosmas, N. Danalatos, L.H. Cammeraat, M. Chabart, J. Diamantopoulos, R. Farand, L. Gutierrez, A. Jacob, H. Marques, J. Martinez-Fernandez, A. Mizara, N. Moustakas, J.M. Nicolau, C. Oliveros, G. Pinna, R. Puddu, J. Puigdefábregas, M. Roxo, A. Simao, G. Stamou, N. Tomasi, N. Usai, A. Vacca, Catena 29 (1997) 45. M. Pateiro-Moure, J.C. Nóvoa-Muñoz, M. Arias-Estévez, E. López-Periago, E. Martínez-Carballo, J. Simal-Gándara, J. Hazard. Mater. (2008), doi:10.1016/ j.jhazmat.2008.08.071. K.M. Spark, R.S. Swift, Sci. Total Environ. 298 (2002) 147. M.H. Florêncio, E. Pires, A.L. Castro, M.R. Nunes, C. Borges, F.M. Costa, Chemosphere 55 (2004) 345. T. Galeano-Díaz, A. Guiberteau-Cabanillas, F. Salinas, Electroanalysis 12 (8) (2000) 616. M.B. McBride, in: G. McBride (Ed.), Environmental Chemistry of Soil, Oxford Univ. Press, New York, 1994, p. 324. W.T. Tsai, C.W. Lai, K.J. Hsien, Colloids Surf. A Physicochem. Eng. Aspects 224 (2003) 99. M. Arias, B. Soto, M.T. Barral, Agrochimica 45 (2001) 218. C.W. Gray, R.G. McLaren, A.H.C. Roberts, L.M. Condron, Eur. J. Soil Sci. 50 (1999) 127. E.E. Kenaga, Ecotoxicol. Environ. Safety 558 (1980) 26. W.R. Haag, C.C.D. Yao, Environ. Sci. Technol. 26 (1992) 1005. US-EPA, United States—Environmental Protection Agency, EPA 738-R-94-018 (1994).