Soil sampling

We collected samples of distrophic Red Umbric Latosol (^{Embrapa 2006}) of a clay texture at a depth of 0-30 cm. Soil had remains of roots that were manually removed, was air dried, broken down into smaller pieces and passed through a 2 mm sieve mesh. The determination of physicochemical characteristics of the soil (Table I) was conducted according to the methodologies described by ^{Embrapa (1997)} and ^{Tedesco et al. (1995)}.

Extraction of dissolved organic matter

Dissolved organic matter (DOM) was extracted from milk cattle dung, collected in a farm with free-stall system. The dung was dried and air-sieved in a 2 mm sieve mesh.

The extraction methodology used in the experiment was adapted from ^{Song et al. (2008)} and ^{Yongkui et al. (2008)}, by mixing cow manure and deionized water at a rate of 1:10 (p/v), in 200-rpm continuous rotation for 12 h at room temperature. The suspension was centrifuged at 5000 rpm for 30 min and filtered through a 0.45 µm membrane, being immediately packaged and frozen for later use.

The DOM extracted was detailed (Table II) with regards to the following parameters: pH, dissolved organic carbon (DOC), copper and zinc following the ^{APHA (1998)} methodology. Certain fractions of humic and fulvic acids, as well as humina present in DOM were also determined as per a methodology adapted from ^{Benites et al. (2003)}.

Sorption experiment

Two grams of soil were placed in 50 mL centrifuge tubes, mixed with 10 mL of a CaCl₂ (0.01 mol/L) solution containing 5, 10, 30, 60 and 120 mg/L of Cu²⁺ or Zn²⁺ (prepared using CuCl₂ or ZnCl₂). Using the CaCl₂ solution has the purpose of avoiding eventual structural destabilization of the soil (^{Corrêa et al. 1999}). Ten mL of DOM were added to the tubes, considering also a control treatment only with the mixture of soil.

The experiment was conducted by triplicate in natural soil pH. The tubes were shaken at 250 rpm for 12 h at room temperature, then centrifuged at 5000 rpm for 10 min. After this procedure, the resulting supernatant from centrifugation was collected and the samples were analyzed to determine the concentrations of Cu and Zn with an atomic absorption spectrophotometer.

The amount of Cu and Zn sipped in the soil was calculated based on the difference of concentration in the solution before and after the sorption experiment, according to the mass balance shown in equation 1:

where q _e is the sorbed metal concentration (mg/g), C ₀ the initial metal concentration (mg/L), C _e the equilibrium concentration at t time (mg/L), V the volume of solution (L), and m the mass of soil (g).

The experimental results were adjusted to mathematical models of Langmuir and Freundlich in their linearized forms, in accordance with the equations 2 and 3, respectively:

where q ₀ is the maximum adsorption capacity (mg/g), K _L is the adsorption constant (mg/L), and K _F and n are constants of the Freundlich isotherm for the ability of adsorption and adsorption intensity, respectively.

Leaching experiment

The soil used to fill the leaching columns was previously contaminated separately, using Cu solutions in the form of CuCl₂.2H₂O and Zn as ZnCl₂. The soil remained in contact with the solutions for 24 h on a shaker (200 rpm/25 ºC) and was then allowed to stand. The supernatant was discarded to remove the non-adsorbed metals and the contaminated soil was oven dried at 50 ºC for 24 h and then passed through a 2 mm sieve (^{Juwarkar et al. 2007}, ^{Franzetti et al. 2009}). Soil, after the contamination, and 3025.82 and 2402.94 presented mg/kg for Cu and Zn, respectively. The pH of the soil after the contamination was 4.07 for Cu and 4.68 for Zn.

The columns were made of PVC pipes, with 5.0 cm in diameter and 30 cm in length, fixed by vertical metal clamps in a universal mount. The total mass of soil added to the columns was of 766 g, and soil placed on the columns was compressed in order to hinder the formation of preferential paths. Aiming to expel the air contained in the pores of the soil with the purpose of avoiding possible structural destabilization of the soil, columns were saturated with a low-flow solution of CaCl₂ (0.01 mol/L) starting from the base of the column..

Once the columns were saturated and the condition of a medium flow of 8.5 mL/min was proven, maintained with the use of a peristaltic pump, the passage of DOM or distilled water (control) in a downward flow was initiated. Twenty 50-mL samples of leachate were collected for a total of 1000 mL. In each sample of leachate, both pH and concentrations of Cu or Zn were determined by a straightforward reading in an atomic absorption spectrophotometer.

Once the concentration values of Cu and Zn in the leachate were obtained, the mass balance of metals in soil columns was calculated and the accumulated mass curves of metals in leachate, in relation to the leached volume, were built. Moreover, with the accumulated mass relationship of metal versus leached volume,we obtained the equations of regression for each treatment. In the equations of regression a straight comparison statistical test was applied to the 5 % level of significance, in order to verify whether they had the same behavior.

Sorption of copper and zinc

The isotherms of adsorption for copper and zinc, which relate the equilibrium concentration in the aqueous phase (mg/L) and the amount of metal adsorbed by soil (mg/g), are presented in figure 1, which demonstrates that the addition of DOM increased metals adsorption in soil.

Fig. 1 Adsorption isotherms for the metals Cu (a) and Zn b) 

For both heavy metals, the control treatment showed isotherms of type L, which are frequent in adsorption studies. This type of isotherm is the result of a relatively high affinity between soil particles and metal, which demonstrates that with increasing concentration the most active sites for adsorption are occupied (^{Sposito 1989}, ^{Antonladis and Tsadilas 2007}). Thus, there is a gradual saturation of the adsorption sites until it reaches a plateau, which characterizes the maximum adsorption of ion or surface molecules of the adsorbent (^{Alleoni et al. 2009}). This condition is observed mainly for zinc in figure 1b, where there is a tendency to stabilize the amount of adsorbed metal from the initial concentration of 30 mg/L in aqueous solution.

In the presence of DOM, isotherms of type H were obtained for copper and zinc, suggesting a very high adsorption affinity in this system. This condition is normally produced by the formation of inner sphere complexes (in which there is no water molecules between the surface functional groups and the ion), or by the van der Waals interactions in the sorptive process, causing specific adsorption (^{Sposito 1989}, ^{Refaey et al. 2014}).

DOM is characterized by a large amount of humic and fulvic acids (Table II), which contributes positively to the adsorption. According to ^{Rose et al. (1979)}, the abundance of carboxyl and hydroxyl radicals gives humic and fulvic acids the adsorption, complexation and chelation properties of metals, that when dissolved, have negative charge. ^{Kiekens (1995)} states thet zinc is preferentially retained by electrostatic forces, while copper is more effective in forming specific linkages, being more stable. Therefore, for this study, the adsorption scheme seems to be the following: DOM has high affinity for metal and forms a complex with it, which binds itself to the soil (^{Sposito 1989}).

Another factor that may have influenced the best adsorption of copper and zinc in the presence of DOM is the pH, since in our tests pH was higher (between 6.63 and 6.22 in solving balance) in relation to the control treatment (between 4.16 and 5.13). The effect of pH on the adsorption of metal cations is mainly the result of charges on soil particles. When pH increases, the electrostatic attraction of the soil by cations also increases and therefore the adsorption rises (^{Sposito et al. 1989}, ^{Alleoni et al. 2009}).

The Langmuir model was not suitable to describe the experimental results for Cu in the presence of dissolved organic matter. According to ^{Alleoni et al. (2009)}, as well as the Freundlich model, the Langmuir equation best describes the adsorption on the surface of colloids and at low concentrations of the element. In addition, these models feature assumptions that are not fully applied to the soil system. As noted in figure 1a, b, the adsorption of copper and zinc was more effective in the presence of the DOM; they were chemically adsorbed, possibly involving the exchange of ligands, which resulted in specific adsorption.

In non-specific adsorption, metals are adsorbed by positively charged surfaces, being weakly retained in a diffuse layer (^{Hingston et al. 1974}), therefore they are best described by the equations of Langmuir and Freundlich. In this context, the models were suitable to describe the adsorption of Cu and Zn in the control treatment, which featured isotherms of type L.

The parameters of Langmuir and Freundlich adsorption for copper and zinc are presented in table III. Although both models are well suited for the study of adsorption for the control treatment, the Freundlich isotherm obtained a higher determination coefficient (R² = 0.986) for the copper, and for zinc the Langmuir model was more appropriate, showing a greater determination coefficient (R² = 1.000).

According to the parameters given in table III for the Langmuir isotherm, it can be observed that the maximum capacity of adsorption (q ₀ ) and the binding energy (K _L ) for zinc have increased considerably with the addition of the DOM, demonstrating the high affinity of the element with DOM.

Regarding the parameters found in the Freundlich isotherm (Table III), the values of K _F and 1/n increased in the presence of DOM for copper and zinc. According to ^{Sposito (1989)} the value of n is qualitatively related to the distribution of energy sites in soil colloids. In this work, the value of 1/n was greater than 1 (1/n > 1) for copper, indicating the presence of highly reactive binding sites. According to ^{El Haddad et al. (2012)}, when (1/n > 1) the change in adsorbed concentration is greater than the change in the solute concentration.

Leaching of copper and zinc

The metals copper and zinc were leached from the beginning of the experiment as noted in figure 2a, b. It was observed, for the two metals under study, greater leaching up to approximately 450 mL, from which occurs a tendency to stabilize the amount of metal carried in the soil.

Fig. 2 Accumulated mass curves of metals Cu (a) and Zn (b) in leachate, in relation to the volume of leachate 

The addition of DOM increased the copper leaching in soil, as demonstrated in figure 2a and confirmed by statistical analysis of regression lines (Table IV), where there was a significant differences at 5 % between the treatment with DOM and the control treatment. However, the mass balance for the metal in the column revealed that only 1.04 and 3.02 % were eluted for control and DOM, respectively, which demonstrates low Cu movement in the soil profile.

For zinc no differences were found between treatments (Table IV and Fig. 2b), which corroborates with the results obtained by ^{Ashworth and Alloway (2004)}. The mass balance for this metal showed that there was an elution of 32.42 and 33.73 % in control treatments and with DOM, respectively, indicating greater miscible displacement of Zn in relation to Cu.

The relative affinity of metals retained in the soil varies according to type of soil, amount of organic matter, CTC, pH, mineralogy, and clay quantity, as well as the properties of the metals (^{Mclean and Bledsoe 1992}). In this study, there was greater retention of Cu in the soil compared to Zn. ^{Alleoni et al. (2009)} reported that metals are specifically adsorbed in the preferred order: Pb > Cu > Ni > Cd > Zn. However, the formation of complexes between metals and the dissolved form of organic ligands (DOM) can affect the adsorption in soil and thus the mobility. The extent of complexation between a metal and soluble organic matter depends on the competition between the metal and the binding sites of the surface and organic ligand (^{Sposito 1989}). In this sense, metals that rapidly form stable complexes with DOM, as Cu, are likely to be more mobile in the soil. ^{Amrhein et al. (1992)} demonstrated that metals such as Cd and Zn, which are not highly stable complexes with organic matter, are not as greatly affected by the presence of dissolved organic matter in the soil. In this way, the type of percolator solution used in this experiment did not influence on leaching of Zn, being able, possibly, to shift the metal present in the soluble fraction of the soil.

In a study by ^{Li et al. (2016)} in red soil in the presence of different phosphates, the authors observed that the concentration of Cu in leachates continuously increased during the whole leaching process for 10 days, while the concentration of Zn increased firstly and then decreased. The pH of the leachate increased continuously with the leaching time.

The pH of the soil system is a very important parameter, directly influencing sorption/desorption, precipitation/dissolution, complex formation, and oxidation-reduction reactions. Because of the complexity of the soil-waste system, with its myriads of surface types and solution composition, such a generalization may not be true. For example, cationic metal mobility has been observed to increase with increasing pH due to the formation of metal complexes with dissolved organic matter (^{Draszawka-Bolzan 2017}).

During the leaching experiments for Cu, pH values in the leachate ranged from 4.84 to 5.34 and from 4.37 to 5.67 for the control and the treatments with DOM, respectively. For Zn, they varied from 4.69 to 6.74 and from 5.08 to 5.64 for control and DOM, respectively. All treatments showed a correlation between the pH and the amount of metal leached (in mg) at a 5 % level of significance. The copper control treatment showed the highest coefficient of determination (R² = 0.600), and the zinc treatment with DOM showed the highest coefficient of determination (R² = 0.773).

According to ^{Harter (1983)}) and ^{Georgiev et al. (2017)}, the transport and precipitation of pollutants depend strongly on the pH and the content of dissolved organic compounds in the soil solutions, which affect various retention mechanisms of metals in soils. ^{McBride (1989)} suggested that high values of pH promote the dissolution of soil organic matter and increase the concentration of total dissolved Cu by the formation of Cu-organic complexes, also influencing the leaching.