Journal ot Envimnmental Science and Health. B40:29-37, 2005 /""^ T I C rCnpyright iG Taylor & trancLS Inc. V*y ^ GISSN: 0360-1234 (Prinl): 1532-4109 (Online)DOI: I0.I08I/PFC-20OO342O5

Sorption and LeaohingPotential of Acidio Herbioidesin Brazilian Soils

Claudio A. Spadotto;* Arthur G. Hornsby,̂and Marco A. F. Gomes^'Embrapa Environment, Jaguariuna, Sao Paulo, Brazil^University of Florida, Gainesville, Florida, USA

Leacbing of acidic herbicides {2,4-D, flumetsulam, and sulfentrazone) in soils was es-timated by comparing the original and modified AF (Attenuation Factor) models formulti-layered soils (AFi). The original AFi model was modified to include the concept ofpH-dependence for K^ (sorption coefficient) based on pesticide dissociation and changesin the accessibility of soil organic functional groups able to interact with the pesticide.The original and modified models, considering soil and herbicide properties, were ap-plied to assess the leaching potential of selected herbicides in three Brazilian soils. ThepH-dependent Kd values estimated for all three herbicides were observed to be alwayshigher than pH-independent K^ values calculated using average Kne data, and there-fore the original AFi model overestimated the overall leaching potential for the soiisstudied.

Key Words: Acidic herbicide; Pesticide; Soil sorption; Soil pH; Leaching potential; Model;Attenuation factor.

INTRODUCTION

After the application of pesticides to a soil system, various physical, chemical,physico-chemical, and biological processes determine their behavior. An under-standing of pesticide transport processes in soil is essential to solve a numberof problems facing agricultural and environmental scientists.'^'

The main soil factors impacting pesticide movement as noted in the lit-erature are organic carbon content, acidity and alkalinity (expressed as pH),texture and physical structure, mineralogy, and associated soil hydraulic prop-erties. These factors influence pesticide half-life, soil sorption, and physicalmovement ofthe compounds through soils.

'Correspondence: Claudio A. Spadotto, Embrapa Meio Ambiente, Rodovia SP 340 Km127,5, Jaguariuna, SP 13820-000, Brazil; E-mail; spadotto@cnpma.embrapa.br.

30 Spadatto, Harnsby, and Gomes

Several authors have proposed screening methods for determining whethera pesticide is likely to leach. Some have attempted to set threshold values fora physical property or set of properties which, when exceeded, should indicatethat the pesticide will leach. Others have presented simple analytical or numer-ical models, which are run using measured or estimated properties of pesticideand soil, in order to predict the likelihood of leaching.

Jury et al.'^' recommended use ofa screening model that includes the influ-ence of mobility and biochemical half-life in assessing the potential for leaching.A similar approach was used by Rao et al.'^' in forming a simple index basedon vadose-zone residence time in the soil and biochemical half-life.

Rao et al.'^' incorporated pesticide degradation functions into expressionsfor travel time to yield the Attenuation Factor {AF) model, which is the relationbetween pesticide mass entering groundwater at some distance below the soilsurface and pesticide mass applied to the surface. Rao et al.'^' used the approx-imate actual distance to groundwater underlying a given soil in computing therelative leaching potential.

The purpose of this study was to estimate sorption and to predict transportof acidic herbicides in soils, comparing the original and modified AF models formultilayered soils (AFi). The original AFi model was modified by introducingthe concept of pH-dependence for Kd, and using equations developed and pa-rameters estimated as presented previously by Spadotto and Hornsby.'^' Theoriginal and modified AFi models, considering soil and pesticide properties,were applied to assess the leaching potential of selected herbicides in threedifferent Brazilian soils.

MATERIALS AND METHODS

Leaching potentials of three acidic herbicides, 2,4-D [C.A. name; (2,4-dichlorophenoxy) acetic acid; CAS RN: 94-75-71, flumetsulam [C.A. name: N-(2,6-difluorophenyI)-5-methyl [1,2,4] triazolo [1,5-a] pyrimidine-2-sulfonamide;CAS RN: 98967-40-91, and sulfentrazone {C.A. name: N-[2,4-dichloro-5-[4-(difluromethyl)-4,5-dihydro-3-methyl-5-oxo-lH-l,2,4-triazol-l-yl]phenyl]methanesulfonamide; CAS NR: 122,836-35-5}, were assessed for several soilsofthe Ribeirao Preto region, Sao Paulo State, Brazil. Herbicide properties arepresented in Table 1. Soil samples were obtained from the Espraiado Streamwatershed, which is located in an area where the risks of contamination tendto be high for groundwater because of aquifer vulnerability. The watershed ison a recharge area of an important aquifer which covers about 1,194,000 kmand is one ofthe world's largest.

Soil properties were determined based on methods presented by Camargoet al.'^' and EMBRAPA.'^' A total soil depth of 120 cm was considered, and soilparameters used in this study are presented in Tables 2-4. These properties

Sorption and Leaching Potential of Acidic Herbicides in Brazilian Soiis 31

Table 1: Properties of selected aoidic herbicides.

Herbicide

2,4-DFlumetsulamSulfentrazone

Dissociationconstant

2,84.66.6

(mug)

203510

Sorption coefficient

(10^ rtiL/g)

0.490,700,02

(10^'mL/g)

41,030,430,03

Half-life

(days)1060

200

°Data from Iiterature,'3.i3-17) and databases (PIPs/EXTOXNET; PPD/ARS/USDA).^Average /<oc volue at pH 6-7. considering no pH-dependence.'̂ Overall value estimated in previous study.* '̂'Considering that f'/2 cloes not depend on sorption.

Tabie 2: Properties of soii 1 (Quartzipsammentic Haplorthox).

Depth(cm)

0-2020-4040-6060-8080-100

100-120

pHCaCis

5.14.94.84.64.44.1

Ciay(%)

242630293234

Porosity(%155.862.154.954.456.551.7

Fieid capacity(%)

22,321.721.320,219.618.8

Buii< density(g/cm^)

1.291.221,211,281,211.16

Organic carbon(%1

0.720.500.340.250.220.20

Tabie 3: Properties of soil 2 (Typic iHaplorthox),

Depth(cm)

0-2020-4040-6060-8080-100

100-120

pHCaCi;

5.35.14.94.84.64.2

Ciay(%)

687473706763

Porosity(%)

58.257.760.461.665.264.7

Fieid capacity(%)

26.224.623.523.122.821.4

Bulk density(g/cm^)

1.371,381.321.191.161.18

Organic carbon(%)

1.381.050.870.670.640.60

Table 4: Properties of soil 3 (Typic Quartzipsamment).

Depth(cm)

0-2020-4040-6060-8080-100

100-120

pHCaCI;

5.35.25,04.84,54.2

Clay(%)

68

n87

10

Porosity(%)

44.662.561.042.357.661.7

Field capacity(%)

19.418.717.317.016.516.1

Buik density(g/cm^)

1,461.521,481.461.451.44

Organic carbon(%)

0.300.240.200.130.080,03

32 Spadctto. Hornsby. and Gomes

represent those of major soils in that region. The net recharge rate (50 cm/jT.)in the watershed was estimated as the difference between rainfall (150 cm/yr.)and evapotranspiration (100 cm/yr.). This assumes steady flow in the soil zoneconsidered.

The highly weathered Oxisols, with large areas in Brazil and other parts ofthe world, present very stable aggregates coated with oxides and organic matter,and the hydraulic conductivity of many clayey Oxisols is considerable.'*^' Oxisolshave variable charge because ofthe presence of pH-dependent minerals such askaolinite, and iron and aluminum hydroxides, as well as due to organic matter.The Quartzipsamments are sandy soils with low organic matter content andhigh hydraulic conductivity.

Leaching potentials ofthe herbicides were assessed for multi-layered soilsby means of the AFi model. The originally presented AF model for one-layersoil'̂ ^ was generalized for a multi-layer soil profile (AFi model) by Hornsby andRao.'^' According to Rao et al.,'^' AF is defined as the fraction ofthe amount ofpesticide applied at the soil surface that leaches through a given soil depth. Itis expressed as the following:

AF = expi-tn • ki) (1)

where tr is the travel time and k is the first-order rate constant for pesticidedegradation.

Pesticide half-life (iV2) is related to k by:

(2)

and the travel time is defined as follows:

f r - L R F — I (3)

where L represents the distance to groundwater, FC is the soil field capacity,and q stands for the net recharge rate. The retardation factor (RF) for pesticideflow is:

where BD stands for the soil bulk density, Foe is the fraction of organic car-bon, Koc is the organic-carbon pesticide sorption coefficient, AC is the air-filled porosity, and KH represents the dimensionless form of the Henry's lawconstant. The organic-carbon-normalized sorption coefficient {KQC) is definedas:

^oc = -^ (5)

Sorptian and Leaching Potentiai of Acidic Herbicides in Brazilian Soils 33

where Kd is the pesticide sorption coefficient. For pesticides with low volatility(i.e., low KH values), retardation factor can be simplified to:

(6)

Hornsby and Rao'^' presented a form of Eq. (1) for a multi-layered soil asfollows;

AFi = n,^(-tri • ki) (7)

in which fl stands for the product ofthe expression [expi-tri • k,} over all lay-ers, and the subscript i designates the ith layer; thus, i — 1 n, where nrepresents the number of layers, and the values for tr and k are unique to eachlayer.

Values of AFi were also calculated for each herbicide using a modified model(AFi*), where Kd values are pH-dependent according to the equation presentedby Spadotto and Hornsby'"*':

^ = l<c-/^) + Kc-r)t-^oc-0m (8)

where fraction of molecular pesticide form (0^), and accessibility of humin indissociated and nondissociated organic matter (f^ and /"") are estimated re-spectively with Eqs. (9-11):

^M- ' (9)

The values ofthe empirical constant, n, and the apparent dissociation con-stant for organic matter, pKom, used in this study, Eq. (10), were in accordancewith expected values reported by McBride,'^' that is n - 2 , and Bohn et al.,'^^'and Sparks,'^^' pKom — 5.

The K'^ and K'^^ coefficients are reference values for the magnitude of sorp-tion, while the estimates ofthe accessibility of humin in dissociated and nondis-sQciated organic matter {f^ and /"") are expected to change with pH. They,along with (pm, will define the shape ofthe K^. vs. pH curve. Thus, Eq. (8) andparameters {K'^^ and K'^^ values) that were estimated in previous study,''*' andare presented in Tabie 1, were introduced to calculate Kd values as a functionofpH.

34 Spadotta, Harnsby, and Games

RESULTS AND DISCUSSION

Herbicide sorption coefficient (Kd) values were calculated from literature dataof sorption coefficient normalized for soil organic carbon content (Koc), and alsoestimated by Eq. (8) and values of reference prime sorption coefficients nor-malized for nondissociated and dissociated organic carbon fractions {K'^ and K^respectively) for 2,4-D, fiumetsulam, and sulfentrazone (Table 1) in differentlayers of each soil.

The pH-dependent Kd values estimated for all three herbicides were alwayshigher than pH-independent Kd values calculated using average Koc data. TheKd values, as expected, were the lowest for Soil 3, and highest for Soil 2. Soil3 has the lowest organic carbon (OC) content, while Soil 2 has the highest OCcontent.

It is worth noting that the pH-independent Kd values follow primarily thedecrease of OC with depth. The pH-dependent Kd values for the three herbicidesevidenced a large variation from layer to layer following changes in OC and pHfor the different soil depths. That is, Kd depends on a combined effect ofthesetwo soil parameters. When OC decreases, Kd tends to decrease; on the otherhand, lowering pH tends to increase Kd.

For all three soils, OC and pH exhibit an overall decrease with depth. How-ever, for Soils 1 and 2, OC contents do not vary largely below 60 cm; thus,decreases in pH result in increasing Kd values for 2,4-D and fiumetsulam. ForSoil 3, OC content continues to decrease with depth below 60 cm, and 2,4-D andflumetsulam Kd values drop concurrently.

The pH effect on sulfentrazone sorption is shown to be very interesting.Because of its high PKA value (6.6), sulfentrazone Kd reaches a "maximum"value at higher pH values when compared with 2,4-D and flumetsulam. Thus,sulfentrazone Kd values do not increase when pH drops below 60 cm in all threesoils.

The AFi, expressed as percentage of pesticide mass applied at soil surface,was calculated for each soil. Overall AFi values for each soil profile are pre-sented in Figure 1. Despite differences between the pH-independent Kd andthe pH-dependent Kd values, the AFi values for 2,4-D, calculated by the origi-nal multi-layered-soil model and by the modified model (AFi*), were similarlylow for all three soils (Fig. 1), mostly due to the short half-life of 2,4-D.

The pH-dependent AFi values for flumetsulam were always much lowerthan values calculated by the original multi-layered model. Therefore, the pHindependent model appears to overestimate leaching potential of flumetsulam.The AFi values for sulfentrazone calculated by the original and the modifiedmodels were similarly high for all three soils, despite the differences in Kdvalues. The long half-life of sulfentrazone mostly contributed to the similarhigh values of AFi for the three different soils.

Sorption and Leactiing Potential of Acidic Herbicides in Brazilian Soils 35

Figure 1: Overall herbicide Attenuation Foctor values estimated by the ariginalmulti-loyered model (AF/) and by the modified model (AF/*) for Soil 1 (QuartzipsammenticHaplorthox), Soil 2 CTypic Haplorthox), and Soil 3 (Typic Quartzipsamment).

Overall AFi values showed large differences for sulfentrazone when cal-culated by the original and by the modified model (AFi*)—Figure 1, owing toits high AF value for each layer. Thus, the original AFi model would seem tomarkedly overestimate the leaching potential for sulfentrazone, as well as forflumetsulam for these soil conditions.

The pesticide sorption effect on degradation rate was not considered inthis study. However, Ogram et al.'^^' suggested that microbial degradation of2,4-D occurs only in the solution phase, and not when the pesticide is sorbed.Fontaine et al.'^^' reported that the degradation rate of flumetsulam was fasterin soils with lower sorption capacities. If a general suppressive effect of sorptionon degradation was recognized, an increase in pesticide AF values estimatedby the modified model would be expected when pH diminishes and sorptionconsequently increases.

Greater sorption might lead to somewhat longer persistence of the pesti-cides in soils. This is of interest because (a) it provides still another meansof testing both the sorption-prediction model and the "nondegraded-if-sorbed"theory, and (b) it also provides some interesting opportunities for decision-makers, based on decreased mobility vs. increased persistence of a givenpesticide.

36 Spadotto, Hornsby. ond Gomes

CONCLUSION

Incorporating the developed sorption model made possible the improvementof an existing pesticide transport model. The pH-dependent Kd values esti-mated for all three herbicides were observed to always be higher than pH-independent Kd values calculated using average Koc data, and therefore theoriginal AFi model overestimated their overall leaching potential for the soilsstudied.

REFERENCES

1. Davidson, J.M.; Rao, P.S.C; Green, R.E.; Selim, H.M. Evaluation of conceptual pro-cess models for solute behavior in soil-water systems. In Agrockemicals in Soils; Banin,A., Kafkafi, U, Eds.; Int. Irrigation Information Center, Pergamon Press: Oxford, Eng-land 1980; 241-251.

2. Jury, W.A.; Spencer, W.F.; Farmer, W.J. Behavior assessment model for trace organicsin soil: I. Model description. J. Environ. Qual. 1983,12, 558-564.

3. Rao, P.S.C; Hornsby, A.G.; Jessup, R.E. Indices for ranking the potential for pesticidecontamination of groundwater. Soil Crop Sci. Soc. EL Proc. 1985, 44, 1-8.

4. Spadotto, C.A.; Homshy, A.G. Soil sorption of acidic pesticides: modeling pH effects.J. Environ. Qual. 2003, 32, 949-956.

5. Camargo, O.A.de; Moniz, A.C; Jorge, J.A.; Valadares, J.M.A.S. Metodos de AndliseQuimica, Mineralogica e fisica de Solos {In Portuguese); Boletim Tecnico 106, InstitutoAgronoomico: Campinas, Brazil, 1986.

6. EMBRAPA (Empresa Brasileira de Pesquisa Agropecuaria) servifo nacional delevantamento e conserva^ao de solos. Manual de Metodos de Andlise de Solo. (InPortuguese); Ministerio da Agricultura; Rio de Janeiro, Brazil, 1979.

7. Sanchez, P.A. Properties and Management of Soils in the Tropics; John Wiley & Sons:New York, NY, 1976.

8. Homshy, A.G; Rao, P.S.C. Post Publication Addendum to Indices for Ranking thePotential for Pesticide Contamination of Groundwater [Online], 1998. Available athttp://muck.soils.ufl.edu/ (posted 21 Jul. 1998; verified 16 Eeh. 2004).

9. McBride, M.B. Environmental Chemistry of Soils; Oxford Univ. Press: New York,NY, 1994.

10. Bohn, H.L.; McNeal, B.L.; O'Connor, G.A. Soil Chemistry, 2nd Ed.; John Wiley:New York, NY, 1985.

11. Sparks, D.L. Environmental Soil Chemistry; Academic Press: San Diego, CA, 1995.

12. Ogram, A.V.; Jessup, R.E.; Ou, L.T.; Rao, P.S.C. Effects of sorption on biologicaldegradation rates of (2,4-dichIorophenoxy) acetic acid in soils. Appl. Environ. Microhiol.1985, 49, 582-587.

13. Fontaine, D.D.; Lehman, R.G; Miller, J.R. Soil adsorption of neutral and an-ionic forms of sulfonamide herbicide, flumetsulam. J. Environ. Qual. 1991, 20, 759-762.

14. Hornsby, A.G.; Wauchope, R.D.; Hemer, A.E. Pesticide Properties in the Environ-ment; Springer-Verlag: New York, NY, 1996.

Sorption and Leaching Potential ot Acidic Herbicides in Brazilian Sells 2>7

15. Jury, WA.; Spencer, W.F.; Farmer, WJ. Behavior assessment model for trace or-ganics in soil: III. Application of screening model. J. Environ. Qual. 1984, 13, 573-579.

16. Tomtin, CD.S., Ed. The Pesticide Manual, 12th Ed.; The British Crop ProtectionCouncil: UK, 2000.

17. Grey, T.L.; Walker, R.H.; Wehtje, G.R.; Hancock, H.G. Sulfentrazone adsorption andmobility as affected by soil and pH. Weed Sci. 1997, 45, 733-738.