INTRODUCTION:

In Brazil, sugarcane area encompasses more than 5 million hectares. Farmers produce more
than 570 million tons of sugarcane yearly. The industry uses this crop to produce sugar,
alcohol and other derivates (^{ALCOPAR, 2012}). This
production level reflects the sugarcane favorable climate in many parts of Brazil.
However, those same environmental conditions also favor the growth of several weed
species, and growers use herbicides to manage them.

Sugarcane requires large amounts of water during its vegetative cycle and has a high
transpiration rate. Normally water is also the vehicle of herbicide uptake by the
plants. The herbicides used in sugarcane cultivation have high water solubility, long
soil half-life, and high accumulation potential in plants (^{TRAPP, 1995}; ^{COUSINS & MACKAY,
2001}).

Mathematical models predict plant herbicide concentrations (^{TRAPP, 1995}). Several models simulate any substance uptake by plants
(^{TRAPP & MATTHIES, 1995}; ^{FUJISAWA et al., 2002}; ^{TRAPP et al., 2003}; ^{TRAPP,
2007}; ^{PARAIBA & KATAGUIRI, 2008}).
Some researchers developed models to simulate specific substance uptake by leaves (^{TRAPP, 1995}), roots and tubers, (^{TRAPP et al., 2003}; ^{PARAIBA & KATAGUIRI, 2008}) or by roots and leaves (^{FUJISAWA et al., 2002}; ^{TRAPP,
2007}). However, none of these models estimate the herbicide bioconcentration
factor and uptake in sugarcane.

The bioconcentration factor of a substance (*BCF*) in an organism is a
coefficient that describes the increase in concentration of herbicides in the organism
in relation to concentration in the medium, estimated by the limit in time in the
chemical steady state equilibrium. In case of plants for food, the *BCF*
permits scientists to evaluate the human's daily ingestion of pesticide establishing
safe limits for concentration in the medium (^{PARAIBA
& KATAGUIRI, 2008}). This research used this model to estimate sugarcane
herbicide absorption from soil solution. This work evaluated the *BCF* of
herbicide in sugarcane through physical and chemical properties of herbicides and
physiological characteristics of the crop. We chose tebuthiuron as an indicator because
it is commonly used in sugarcane cultivation in Brazil and has high mobility (^{CERDEIRA et al., 2007}).

MATERIALS AND METHODS:

The *BCF* model

This paper's estimation of sugarcane herbicide bioconcentration relies on several
assumptions: degradation in the soil and its metabolism and dilution in the plant are
described by first order kinetic equation; the plant uses the water transpiration stream
to uptake the herbicide from the soil solution; soil solution herbicide present in
concentrations that are available for plant uptake; and the plant distributes the
herbicide throughout itself by transpiration. These assumptions reflect the results of
several studies (TRAPP, 2007; ^{REIN et al., 2011};
^{TRAPP & LEGIND, 2011}; ^{TRAPP & EGGEN, 2013}).

We obtained the *BCF* for the steady state chemical equilibrium, which we
estimated using the time limit of the quotient between the plant herbicide and the soil
solution herbicide concentration. We calculated the total balance of the herbicide's
mass through the following equations:

(^{TRAPP et al., 2003}; ^{PARAIBA, 2007}; ^{PARAIBA et al.,
2010}), where *M*
*
_{P}
* (kg ha

^{-1)}is the total plant fresh biomass,

*Q*(L day

^{-1}ha

^{-1)}is the plant transpiration rate,

*TSCF*

*is the herbicide concentration factor in the transpiration stream in relation to the soil solution,*

_{soil}*C*

*(mg L*

_{W}^{-1)}is the soil solution herbicide concentration,

*k*

*(day*

_{E}^{-1)}is the herbicide transformation rate in the plant,

*k*

_{G}(day

^{-1)}is the plant growth rate,

*C*

*(mg kg*

_{P }^{-1)}is the plant herbicide concentration, and

*K*

*(L kg*

_{PW}^{-1)}is the herbicide plant-water sorption coefficient or herbicide bagasse-water sorption coefficient measured as described in the sorption experiments.

The *TSCF*
*
_{soil}
* was estimated from the herbicide octanol-water partition coefficient using the
equation, given by (

^{BURKEN & SCHNOOR, 1998})

where log*K*
*
_{OW}
* is the logarithm of herbicide octanol-water partition coefficient and

*TSCF*is the concentration factor of pesticide in the transpiration stream without interference of soil elements. The present paper calculates the

*TSCF*

*is developed by*

_{soil}

(^{NICHOLLS, 1984}), where *K*
*
_{OC}
* (L kg

^{-1)}is the herbicide sorption coefficient in the soil organic carbon,

*f*

*(g g*

_{OC}^{-1)}is the soil organic carbon volumetric fraction,

^{-1)}is the total soil density, and (g g

^{-1)}is the soil water volumetric content.

The experiment measured the herbicide plant-water partition coefficient
*K*
*
_{PW}
* (L kg

^{-1)}as follows (Table 1). Initially, it was determined the sorption coefficient of the herbicides in the bagasse (dry pulp that remains after juice extraction) for the development of the model that describes the mass balance of the herbicide in the soil-plant in sugarcane. We selected the following herbicides to evaluate the

*BCF*: acetochlor, ametryn, clomazone, diuron, hexazinone, metribuzin, picloram, simazine, sulfentrazone and tebuthiuron. For this experiment we first washed and processed the sugarcane cane in a manual mill to separate the juice from the bagasse. We then dried the resulting bagasse in an oven with air circulation at 60°C for 72h. After complete drying, we grounded the residue using a knife mill. Then we washed the bagasse with distilled water and filtered it through qualitative filter paper (Whatman n. 1), using a Büchner funnel. After washing, we again dried the residue under the same conditions described above and sieved in a 1mm sieve.

^{a}log *K*
*
_{OW}
*: partition coefficient octanol-water of herbicide (L
kg

^{-1)}(

^{TOMLIN, 2000}).

^{b}
*K*
*
_{PW}
*: partition coefficient bagasse-water of the herbicides
(measured).

^{c}
*Sol*: water solubility (g L^{-1)} (TOMLIN,
2000).

The soil solution herbicide concentration followed a first order kinetic equation given
by *C*
*
_{W }
*(

*t*) =

*C*

*(0) exp (-*

_{W}*k*

_{S}*t*), where

*C*

*(mg L*

_{W(0)}^{-1)},

*C*

*=*

_{W }*C*

*(*

_{W}*t*) (mg L

^{-1)}and

*k*

*(day*

_{S}^{-1)}are the initial soil solution herbicide concentration, the current soil solution herbicide concentration, and the herbicide dissipation rate in the soil, respectively. The equation

describes the plant herbicide concentration, where

(L kg day^{-1} ha^{-1)} and

(day^{-1)}. The *A* constant is the herbicide uptake rate by
plants and the *B* constant is herbicide dilution and metabolic rate in
the plants.

We calculated the bioconcentration factor for the steady state chemical equilibrium by the time limit of the quotient between the plant herbicide and soil solution herbicide concentrations, as follows

where *BCF* (L kg^{-1)} is the plant herbicide bioconcentration
factor and *k*
*
_{EGS }
*

*= k*

_{E}*+ k*

*-*

_{G}*k*

*(day*

_{S }^{-1)}is the herbicide dilution rate, metabolism and dissipation in the soil-plant system (

^{PARAIBA, 2007};

^{PARAIBA et al., 2010}).

Previous research provided the herbicide half-life time values and the soil sorption
coefficients (^{HORNSBY et al., 1995}). This
experiment uses the half-life time to estimate the degradation rate, *k*
*
_{S }
*values, through the expression

*t*

_{1/2}(day) is the soil herbicide half-life (

^{HORNSBY et al., 1995}). Although polar herbicides

*K*

*varies with soil pH with geographical regions, normally crop soils pHs are adjusted to near 6 worldwide, thus validating the possible use of literature for this estimation (*

_{oc}^{OLIVEIRA et al., 2001}).

Sorption kinetics and chromatographic analysis of the herbicides

The equation

determines the sorption coefficients, where*C*

*(mg µg*

_{P}^{-1)}is the concentration of the herbicides the bagasse sorbed,

*C*

*(mg mL*

_{W}^{-1)}is the concentration in water solution, and

*K*

*is the partition coefficient bagasse-water of the herbicides measured. The equation provides the coefficients, where*

_{PW}*X*(mg) is the amount of herbicide sorbed, and

*M*(µg) is the weight of bagasse and

*C*

*(mg mL*

_{fW}^{-1)}is the final concentration of herbicide in solution. The formula measures the amount of herbicide sorbed, where

*C*

*(mg mL*

_{iW}^{-1)}is the initial concentration of the herbicide in solution and

*V*(mL) is the volume of solution (Table 1) (

^{TRAPP et al., 2001}).

We performed sorption kinetics and isotherms curves for each individual herbicide. We
kept the rate of bagasse:solution at 1:15 (1g bagasse - 15mL of solution of 1µg
mL^{-1} of herbicide in water), and maintained the samples at a temperature
of 25°C, shaking at 185rpm. To quantify and determine the point on time of equilibrium
of herbicide absorbed, took the aliquots and filtered in 0.45μm sieve for herbicide at
regular intervals. The initial time of the experiment (*t *= 0) was when
we applied the herbicide to the media. We performed the analyses on HPLC using a UV-Vis
detector (Shimadzu, model SPD 10AVvp), Supelco-Lichosorb RP-18.5µm (250mm-4.6mm) column,
flow rate of 1.0mL min^{-1}, room temperature and injection volume of 20µL. Each
herbicide had a specific chromatographic conditions (mobile phase and UV wavelength).
All herbicides reached equilibrium within a period of 24 hours. For construction of the
isotherms, seven standard concentrations, ranging from 0.1 to 8.0µg mL^{-1}, in
duplicate were used.

Tebuthiuron sugarcane uptake experiment

To conduct the experiment we used containers of 40×10^{3}cm^{3 }capacity
with organic sugarcane so as to avoid interference with other chemicals watered to
maintain at field capacity, IAC2480 variety. The experiment was conducted at pH 5.9
(CaCl_{2}), soil organic carbon of 0.012g g^{-1}, total soil density
of 1.3kg L^{-1}; and soil moisture of 0.28g g.^{-1} The herbicide
tebuthiuron was applied at of 5.0kg a.i. ha^{-1}, a rate that although higher
than the recommended application rate did not cause phytotoxicity to the plant. We used
higher rates to simulate a worst case scenario. We harvested the sugarcane every three
months after application in order to quantify the herbicide until 20 months old.

RESULTS AND DISCUSSION:

We applied the model to the sugarcane with the following soil and plant parameters: soil
organic carbon=0.012g g^{-1}; total soil density = 1.3kg L^{-1}; soil
humidity =0.28g g^{-1}; total plant fresh biomass = 80,000kg ha^{-1};
average plant transpiration rate = 32,000L day^{-1} ha^{-1}; and average
relative plant growth rate = 0.05 day^{-1} (^{CABRAL et al., 2003}). The plant herbicide metabolism rate *k*
*
_{E}
* was calculated by

*k*

_{E}*= k*

_{S}*/16*(

^{JURASKE et al., 2008}).

The data (Table 1) were used to produce two
curves correlation of herbicide properties such as water solubility, partition
coefficient octanol-water, and sorption coefficient in bagasse. The equations
log* K*
*
_{PW}
*

*=0*.45+0.46xlog

*K*

*, with P<0.001,*

_{OW}*n*=8, correlation coefficient

*R - s*=79.75%, standard deviation=0.15 (Equation 1) and log

*K*

*= 0.29+0.046xlog*

_{PW}*Sol*+0.54x log

*K*

*, with P<0.001,*

_{OW}*n*= 8, correlation coefficient

*R-s*=82.93%, standard deviation =0.155 (Equation 2), were developed.

Equations 1 and 2 show the relationship of the sorption coefficient plant/water or
bagasse/water with the coefficient octanol-water. Due to the high solubility of the
herbicides picloram and acetochlor, the analytical method did not work as for the other
herbicides. For this reason, we used the equations to estimate sorption coefficient
bagasse/water for the herbicides: acetochlor (*Sol*=0.233g L^{-1}
and log*K*
*
_{OW }
*=4.14) and picloram (

*Sol*=0.430g L

^{-1}and log

*K*

*=19). Sorption coefficient bagasse-water (*

_{OW }*K*

*) of acetochlor was 230 (Equation 1) and 313L kg*

_{PW}^{-1}(Equation 2). Picloram was 21 (Equation 1) and 19L kg

^{-1}(Equation 2). The Pesticide Manual provided the values of

*Sol*(soubility) and log

*K*

*(octanol-water partition coefficient) of the herbicides (*

_{OW}^{TOMLIN, 2000}). The values obtained from Equation 1 for picloram and acetochlor were used to feed the model.

Sugarcane herbicide *BCF* varied between 0.0081L kg^{-1}
(acetochlor) and 0.8570L kg^{-1} (sulfentrazone) (Table 2). Based on the *BCF *values measured (Table 2), we concluded that the herbicides that
would most probably be found in plant are sulfentrazone > picloram > tebuthiuron
> hexazinone > metribuzin > simazine> ametryn > diuron > clomazone
> acetochlor. Herbicides with small *K*
*
_{PW}
* and

*K*

*and high*

_{OC}*TSCF*

*are the ones with the highest BCF, such as sulfentrazone, picloram, tebuthiuron, hexazinone, and metribuzin (Table 2). The herbicides ametryn, diuron, clomazone, and acetochlor have the lowest*

_{soil}*BCF*and have in common high

*K*

*,*

_{OC}*K*

_{OW}*,*

*K*

*, and low*

_{PW}*TSCF*

*(Table 2).*

_{soil}

log *K*
*
_{OW}
*: herbicide octanol-water partition coefficient;

*K*

*: herbicide sorption coefficient in the soil organic carbon;*

_{OC}*K*

*: plant-water partition coefficient; GUS : Groundwater Ubiquity Score; BCF: sugarcane herbicide bioconcentration factor;*

_{PW}*TSCF*

*: soil-plant transpiration stream concentration factor;*

_{soil}^{a}Data from

^{HORNSBY et al., (1995}).

The *BCF* (Table 2) values permit
an estimation of the herbicide daily intake (*DI*), per body weight, from
sugarcane products consumption establishing acceptable limits. For example, a soil
solution containing 1.0mg L^{-1} (*C*
*
_{W}
*) of tebuthiuron leads to a sugarcane concentration (

*C*

*) of 0.4159mg kg*

_{P}^{-1 }(

*C*

_{P}*=*BCFx

*C*

_{W}*)*and a daily herbicide intake of 0.003 mg kg

^{-1}(mg of tebuthiuron per kg body weight) considering a 70kg person with a daily sugarcane juice consumption (

*DI*) of 0.5kg, measured by DI=0.5X

*C*

*/70.*

_{P}Besides being absorbed, herbicides could leach to groundwater. We can estimate the
herbicide leaching potential through the GUS index (^{GUSTAFSON, 1989}), given by* GUS*=(4-log *K*
*
_{OC}
*) x log

*t*

_{1/2}. Depending on the GUS index numerical value, we may classify the herbicide as a high leaching potential (

*GUS*≥2.8), non-leaching (

*GUS*≤1.8)or with undetermined leaching potential (transient) (1.8

*GUS*<2,8). Therefore, herbicides with low soil sorption coefficients and high water solubility, are classified as leaching pesticides because of their

*GUS*index values (Table 2). High soil herbicide sorption makes them adsorbed in soil matrix, and thus, unavailable for lixiviation or plant uptake (

^{TRAPP, 1995}). Six of the ten studied herbicides (60%) are potentially leaching. One (10%) is a non-leaching herbicide and others transient (Table 2). Overall herbicides with the highest

*BCF*or potential to be absorbed are also the ones with highest potential to leach to groundwater (Table 2).

Tebuthiuron bioconcentration

We applied Tebuthiuron to the soil at 5.0kg a.i. ha^{-1}, a rate higher than the
recommended, without causing phytotoxicity to the plant. We harvested the plants every
three months after application for herbicide quantification. We were able to measure and
detect the herbicide only during the first harvesting date, three months after
application, at a level of 0.5 mg kg^{-1}. Although, in our study, the herbicide
had a significant *BCF* of 0.42, it also had a high *GUS*
index rating, 5.36, indicating that most likely it would leach or break down (^{SILVA et al., 2010}) before the plant uptakes the
herbicide further.

Other authors found tebuthiuron residues in adult sugarcane plants cultivated in soils
treated with 2.2kg. a.i. ha^{-1} (recommended rate) and 11.2kg ha^{-1}
with herbicide residue concentrations of 0.074 and 0.026µg g^{-1}, respectively
(^{CAUX et al., 1997}). In this same experiment,
the authors' found in their sugarcane bagasse, juice and syrup (11.2kg ha^{-1)}
contained concentrations of 0.063, 0.076 and 0.193µg g^{-1}, respectively. They
found no residues in the sugar obtained from these plants. However, researchers detected
residues of three tebuthiuron metabolites in grasses and bushes grown in soils of a
semiarid region of Arizona, 11 years after herbicide application (^{JOHNSEN & MORTON, 1991}). We found no other literature regarding
experiments on the application of tebuthiuron to sugarcane plants similar to those
reported by ^{CAUX et al. (1997}) and ^{JOHNSEN & MORTON (1991}).

Another practical information generated by this study is related to herbicide efficacy
for weed control, since herbicides with high *K*
*
_{PW}
* such as measured in this study (Table
2) can be less efficient for weed control. These herbicides are pre plant soil
applied over a straw left by mechanical harvesting of sugarcane, are adsorbed by the
straw residues left and therefore failing to reach the weed seeds that should be
controlled (

^{CORREIA et al., 2007}).

CONCLUSION:

This study presents a model equation to estimate the *BCF*. This equation
depends directly on the herbicide concentration factor in the transpiration stream, the
transpiration stream volume and the plant-water partition coefficient; and indirectly on
transpiration stream volume, herbicide dilution rate, metabolism and dissipation in the
soil-plant system, the plant-water partition coefficient and the plant biomass. This
model allows identification of herbicides that might potentially bioconcentrate in
sugarcane and sugar products. The herbicides with highest *BCF* that
would most probably be found in plant are sulfentrazone > picloram > tebuthiuron
> hexazinone > metribuzin > simazine > ametryn > diuron > clomazone
> acetochlor. Overall herbicides with the highest *BCF* or potential
to be absorbed are also the ones with highest potential to leach to groundwater.