Introduction

Recently, numerous investigations connected with trace elements in soils have been conducted in order to assess their availability or to point out the contamination of soils. The trace elements contamination for soil is mainly from the mining activities, tanneries, atmospheric deposition, reuse of wastewater, disposal of tailings, sewage irrigation, and use of pesticides and fertilizers (Alloway 2013).

The existence of trace elements in the solid phase is a result of many processes with sorption being the most important (Evans 1989; Harter and Naidu 2001; Selim 2012; Alloway 2013). Thus, it is important to study the sorption behavior and mechanism of trace elements on soils, which may help to assess risks to the environment and manage remedial activities. Many studies on individual sorption and retention of trace elements have been conducted by different type of soil or different components such clay minerals, acid soils, organic matter, and oxides (McKenzie 1980; McLaren et al. 1998; Gomes et al. 2001; Bradl 2004; Papageorgiou et al. 2009). However, there are few studies focusing on competitive component sorption and retention (Selim 2012). Furthermore, individual sorption studies may not adequately predict the behaviors of multiple trace elements in soil (Fontes et al. 2000). It is, therefore, desirable to study the competitive sorption and desorption for better predicting the mobility, bioavailability, and fate of trace elements in soil.

Soil contains complex mixture of components and characterized by varied physicochemical properties that could be involved in trace elements sorption process. Soil pH, the amount and type of clay minerals, carbonate contents, organic matter, oxides, and eventually simultaneous presence of competing trace elements affect sorption processes and their fate in soils (McGrath et al. 1988; Harter and Naidu 2001; Veeresh et al. 2003; Bradl 2004; Cerquiera et al. 2011a, 2011b). Despite the sorptive aspect that gives these properties and components to the soil, trace elements can be desorbed from the surface of the soil colloid to the soil solution, and thus absorbed by plants or released to lower soil horizons and groundwater (Evans 1989; Backes et al. 1995; Selim 2012).

Sorption and desorption isotherms provide useful information about the soil retention capacity and show therefore whether sorption is reversible or irreversible (hysteresis). Vega et al. (2008) stated that K r is a useful parameter for comparing sorption and retention capacity and hysteresis of different soils or materials for trace elements. K r parameter has been used on previous studies of sorption and desorption of trace elements (Vega et al. 2009; Cerquiera et al. 2011a; Cerquiera et al. 2011b; Asensio et al. 2014).

The majority of the agricultural soils in Southern Tunisia are Aridisols type, characterized by rich-carbonate contents, poor organic matter (content is often lower than 1%), and sandy texture. Reuse of treated wastewater, spreading of different amendments like composts and sewage sludge, were used for striving water scarcity, and remediation to poor organic matter and poor-nutrients on southern Tunisian agricultural soils. Although these effluents generally display low concentrations of trace elements, their long-term use often results in the build-up of trace elements content in soils (Alloway 2013). Accumulation of these elements in agricultural soils constitutes a major environmental concern given that the trace elements translocation into crops and ultimately into the food chain has already been recorded (Bruemmer et al. 1986; Cheng 2003; Rattan et al. 2005; Solis et al. 2005; Zhao et al. 2014).

Therefore, studying their sorption and desorption became essential to assess their availability and fate in the soil.

Many trace elements, such as Pb, Zn, Cu, Ni, and Co, were more focused because of their accumulation in soil, contamination to groundwater, and endangerment to safety of agricultural products (Alloway 2013). These elements are highly toxic and cumulative, causing risks to humans and the ecosystem, even at low dosages. They also often occur simultaneously in contamination soil and compete with each other for sorption sites (Echeverria et al. 1998; Veeresh et al. 2003; Papageorgiou et al. 2009; Cerquiera et al. 2011a, 2011b).

The competitive sorption of trace elements in soil may influence their retention, mobility, and the ultimate fate in the environments.

While numerous studies has been conducted to understand competitive sorption and desorption of trace elements in clay minerals, organic matter, oxides, and natural soils, less is known about their competitive sorption and desorption in calcareous soils. Soil carbonates may control sorption and retention of trace elements in calcareous soils (Hussein et al. 2002; Maftoun et al. 2002; Jalali and Moharrami 2007).

Thus, the objectives of the present study are to evaluate and compare the Pb2+, Zn2+, Ni2+, Cu2+, and Co2+ competitive sorption and desorption capacity and sorption hysteresis in three Tunisian Aridisols. The relationship between soil properties and the sorption capacities of these elements were also studied.

Materials and methods

Soil sampling

Soil samples from the surface layer (horizon A) were collected from three areas of Tozeur Province, southwestern Tunisia (Fig. 1). The three soils were classified as Aridisols Calcorthids soils. They were analyzed and extensively characterized to assess the influence of their components and properties on Pb2+, Zn2+, Ni2+, Cu2+, and Co2+ sorption and retention. Using an Eijkelkamp sampler, the samples were collected from the surface horizon of each soil and then stored in polyethylene bags. The samples were pooled, air dried, passed through a 2 mm sieve, homogenized, and stored until analysis. Subsamples were used for determining soil proprieties and for sorption and desorption experiments.

Fig. 1
figure 1

Location of sampling areas (AR1, AR2, and AR3) in Tozeur province (Southwestern Tunisia)

Full size image

Soil characterization

Soil pH was determined with a pH meter in 2:1 water/soil extracts (Guitián and Carballas 1976). Particle-size distribution was determined following oxidation of organic matter with hydrogen peroxide; the fraction down to 50 mm was separated by sieving, and the lower fraction was characterized as Kroetsch and Wang (2008). The carbonate content was determined with a LECO-100 analyzer connected to a LECO-CN-2000. Effective cation exchange capacity (CECe) was determined by the method of Hendershot and Duquette (1986).

Total organic carbon (TOC) was determined in a module for solid analysis (SSM-5000, Shimadzu, Japan) coupled with a TOC analyzer (TNM-1, Shimadzu, Japan).

Oxides were determined using the method of Mehra and Jackson (1960); samples were shaken in a solution of sodium hydrogen carbonate and sodium citrate. Fe, Al, and Mn contents were determined in the extracts by inductively coupled plasma optical emission spectrometry (ICP-OES) in a Perkin-Elmer Optima 4300DV apparatus.

Clay mineralogy was determined by powder X-ray diffractometry in a SIEMENS D-5000 Bragg–Brentano (θ/2θ) apparatus with a Cu anode, using 0.05 steps and 10 s per step (Brindley and Brown 1980).

Total dissolved trace elements were determined using the method of Houba et al. (2000) by calcium chloride extraction. Available trace elements were extracted with diethylenetriaminepentaacetate (DTPA) as described by Lindsay and Norwell (1978). Total contents were determined following acid digestion with a mixture of concentrated nitric, hydrochloric, and hydrofluoric acids (Marcet et al. 1997). Trace elements concentrations on the extractions were determined by ICP-OES.

Sorption and desorption experiments

Batch equilibrium experiments were constructed using the methods of Fontes et al. (2000), as modified by Harter and Naidu (2001). Sorption experiments were evaluated using single-trace element solutions containing equal mass concentrations of Pb2+, Zn2+, Ni2+, Cu2+, and Co2+ nitrates (0.05, 0.1, 0.2, 0.4, 0.5, 1, 2, 3, 4, and 5 mmol L−1). Competitive sorption was evaluated using multi-trace element solutions of Pb2+, Zn2+, Ni2+, Cu2+, and Co2+ nitrates, each element had the same concentration (0.05, 0.1, 0.2, 0.4, 0.5, 1, 2, 3, 4, and 5 mmol L−1).

In both experiments, the solutions contained 0.01 M NaNO3 as background electrolyte. In each experiment, triplicate suspensions of 3 g of soil were equilibrated in 50 mL of sorption solution in polyethylene tubes. The solutions were shaken in a rotary shaker for 24 h at 25 °C in order to ensure sorption equilibrium. After equilibration, the samples were centrifuged at 4000 rpm for 5 min and the supernatant was filtered through Whatman 42 paper. Trace element concentrations in the supernatant were determined by ICP-OES. The pellet was set aside for use in the desorption stage of the experiment. The amount of each element sorbed was calculated from the difference between its concentration in solution before making contact with the soil and after equilibration.

Based on Madrid and Diaz-Barrientos (1992), desorption isotherms were constructed using the pellets obtained in the sorption experiments. The pellets were dried at 45 C, weighed, resuspended in 100 mL of 0.01 M NaNO3 solution, and shaken in a rotary shaker for 24 h in polyethylene tubes at 25 C. The solutions were centrifuged at 4000 rpm for 5 mn, and the supernatant was filtered through Whatman 42 paper. Finally, the concentration of Pb2+, Zn2+, Ni2+, Cu2+, and Co2+ in the supernatant was analyzed by ICP-OES in a Perkin-Elmer Optima 4300DV apparatus.

The quantities of each element that had been retained on the soil sample were calculated from the difference between the concentrations sorbed in the sorption experiments phase and the concentrations of the trace element in solution following desorption. Each sorption and desorption experiment was performed in triplicate.

Isotherms were constructed for each element by representing the amount of trace element sorbed or retained upon desorption by the soils (μmol/g of dry soil), and the concentrations in the equilibrium solution at each added concentration (μmol L−1).

Sorption and desorption isotherms were classified, wherever possible, as belonging to one or other of the profile types described by Giles et al. (1974).

Sorption and retention capacities and hysteresis

Sorption and retention capacity of soils was estimated as the slope K r in the regression equation C s,i = K r C p,i (Vega et al. 2008), where C s,i is sorbed element i (μmol) per gram of soil at equilibrium, and C p,i is potentially sorbable element i, i.e., the amount of element i in the solution before contact with the soil, likewise per gram of soil.

As proved by Vega et al. (2008) and mentioned by Cerquiera et al. (2011a, 2011b), K r is a useful parameter for comparing the sorption and retention capacities of different soils or materials for trace elements under the same experimental conditions.

K r is dimensionless and ranges from 0 to 1; otherwise, K r adopts a null value for a soil that liberates all sorbed trace elements to the value 1 corresponding to a soil that releases no element during the desorption experiment phase. On other hand, a high K r value indicates a high trace element sorption and retention by soil components; similarly, a low K r value indicates that a high trace element amount is in the solution.

The irreversibility of sorption was obtained by calculating the hysteresis index (HI) defined as the ratio between the K r values for retention and sorption (Vega et al. 2008). HI varies from 1 for a totally irreversible sorption and rather 0 for a non-irreversible sorption.

$$ \mathrm{H}\mathrm{I}={K}_{\mathrm{r}}\mathrm{retention}/{K}_{\mathrm{r}}\mathrm{sorption} $$

Statistical analysis

All results were statistically analyzed using the software SPSS v. 20.0 for Windows (SPSS, Inc., Chicago, IL). The statistical significance of differences was determined by one-way analysis of variance (ANOVA) and the least-significant difference (LSD) test. The influence of soil properties on sorption and retention capacity and on HI was investigated by means of pairwise Pearson correlation analysis.

Results and discussion

Characterization of the soils

Basic chemical and physical properties of the soils are shown in Table 1. The three soils have contrasting physical and chemical characteristics.

Table 1 Chemical and physical properties of the studied soils
Full size table

The highest clay percentage is detected in AR1 soil, whereas AR2 has poor clay contents and has the lowest sand and carbonate contents. AR3 has the highest sand percentage, whereas clay fraction is inexistent. For this reason, mineralogical analysis was realized only for AR1. The mineralogical analysis of the clay fraction of AR1 showed that the most abundant mineral is halite, and contains in addition, moderate kaolinite content and small quartz and mica contents. In general, all soils have a low CECe; the AR1 exhibits slightly higher values than AR2 and AR3. The sand content is significantly different between soils and varied between 90 % (AR1) and 60 % (AR2). The TOC is not detected for the three studied soils.

The pH of the soils is alkaline, and they have very similar pH values. Carbonate content, the major defining characteristic of Aridisols, is the most abundant component in the studied soils, and there are no significant differences between AR1 and AR3 contents.

The AR1 soil has the highest proportion of Fe and Mn oxides. Fe oxide concentrations differ significantly between soils, whereas Mn and Al oxides concentrations are very low.

Total concentration of Pb2+, Zn2+, Ni2+, Cu2+, and Co2+ is within the range of the limits established in some reference guides (ICRCL 1987). Trace elements concentrations are significantly different among all studied soil, and the highest values for Zn2+, Ni2+, and Co2+ are detected in AR1 while the ones of Pb2+ and Cu2+ are in AR2 (Table 1). The contents extracted with DTPA and CaCl2 methods are zero or close to zero.

Competitive sorption and retention isotherms

Figures 2 and 3 show the amounts of Pb2+, Zn2+, Ni2+, Cu2+, and Co2+ sorbed by AR1, AR2, and AR3 and retained upon desorption, as well as the concentrations in the equilibrium solution at each added concentration. Sorption and retention isotherms were described by the four general types of isotherms (L, H, S, and C) using the classification of Giles et al. (1974). Most are irregular because of difference on the soils’ physicochemical properties and between-element competition for binding sites, especially in the case of retention isotherms on AR1 for the cations Zn2+, Ni2+, and Co2+.

Fig. 2
figure 2

Competitive sorption isotherms of Pb2+, Zn2+, Ni2+, Cu2+, and Co2+ on AR1 (a), AR2 (b), and AR3(c)

Full size image

The sorption and retention isotherm shapes of Pb2+, however, were identifiable for all soils. In most cases except for retention on AR1, they were almost vertical lines overlapped y-axis and close to C-type isotherms. Sorption and retention of Pb2+ increased vertically with very low initial concentration in the sorption solution, whereas this fact did not occur with Zn2+, Ni2+, and Co2+ which increased almost constantly and linearly with their initial concentration in the sorption solution and thus presented L curve type, except AR1. The constant slope, suggesting constant partitioning of the adsorbate between the solid interface and soil solution, characterized the Zn2+, Ni2+, and Co2+ sorption and retention isotherm on the studied soils except for AR1 retention case (Fig. 3).

Fig. 3
figure 3

Competitive retention isotherms of Pb2+, Zn2+, Ni2+, Cu2+, and Co2+ on AR1 (a), AR2 (b), and AR3 (c)

Full size image

The competitive sorption and retention isotherms of Cu2+ produced in almost all of the cases a C-type isotherm. The characteristic steep slope of the Cu2+ and Pb2+ sorption and retention isotherms in contrast to the linear and gentle slope of the Zn2+, Ni2+, and Co2+ sorption and retention isotherms indicated a stronger affinity for Cu2+ and Pb2+ than Zn2+, Ni2+, and Co2+. The steep slope of Pb2+ and Cu2+ isotherms could have been produced by inner-sphere complexation (Sposito 1989).

Sorption and desorption capacities

Table 2 lists the equation regression, calculated K r of sorption and retention and the corresponding coefficient of determination (R 2) for Pb2+, Zn2+, Ni2+, Cu2+, and Co2+ for the three studied soils.

Table 2 K r for sorption and retention of Pb2+, Zn2+, Ni2+, Cu2+, and Co2+
Full size table

Regarding the studied soils, K r values reveal that both AR1 and AR3 sorbed and retained larger amounts of all cations than AR2 does. By far, a high cations amount was sorbed and retained by AR3 which contain the highest carbonate contents (up to five times more than AR2) and despite the fact that it was characterized by the lowest CECe (0.73 cmol kg−1) (Table 1).

The three studied soils sorbed and retained Pb2+ and Cu2+ more than twice as the lowest fixed cation (Ni2+). However, the AR1 had the highest sorption and retention capacity for Pb2+ and Cu2+, followed by AR3 and AR2 (Table 2). This trend was consistent essentially with the carbonate contents, and less markedly the percentage of clay and CECe (Table 1). Obviously, sorption of Pb2+ and Cu2+ increased with increasing clay, carbonate contents, and CECe of the soil, which was in accordance with the result obtained by Echeverria et al. (1998) for soils with a range of chemical and physical properties.

Following McGrath et al. (1988) and Veeresh et al. (2003), in carbonated and alkaline soils and with no or little contents of Fe, Al, and Mn oxides, clays fraction and TOC, sorption, and retention are regulated by three different mechanisms: surface adsorption of colloids, complexation with surface functional groups, and the precipitation with hydroxides and carbonates. For Pb2+, fixation process is mainly regulated by strong covalent bonds (Bradl 2004) as well as the probability of formation of PbCO3 (Lindsay 1979). Eventually, AR1 and AR3 had a sufficient amount of carbonate, clay, and Fe oxide to be able to form strong bonds that permitted the sorption and retention of Pb2+ regardless of the presence of the other cations.

According to ion activity production of M(OH)2, Cu may be precipitated as an anionic hydroxyl complexes (Cu(OH)2, Cu(OH)3 − and Cu(OH)4 2−) which are the dominant copper species in the soil solution with a pH above 8. Also, copper carbonate (CuCO3) is the dominant inorganic soluble form of Cu2+ in alkaline soils and high carbonate soils like Aridisols.

Similarly to Pb2+, the amount of Zn2+ sorbed and retained by AR1 and AR3 is greater than for AR2. The high capacity of these soils to sorb and retain Zn2+ could be also revealed to the high carbonate contents as described by Hussein et al. (2002). However, since the carbonate contents in AR1 and AR3 were significantly the same and were much higher than that in AR2, higher sorption and retention capacities of Zn2+ is therefore presented in these two soils. On the other hand, because AR3 clay fraction was very low and not detected (Table 1), its sorption and retention capacities of Zn2+ were lower than AR1. These results are quite consistent with those obtained by Hussein et al. (2002) in Zn2+ sorption tests on Sudanese Aridisol.

Also as explained by Alloway (2013), in alkaline soils, zinc hydrolyses and the hydrolyzed species are strongly adsorbed to soil surfaces. Zn sorption increases due to the higher surface charge density and precipitation of Zn2+ as Zn(OH)2 is more prevalent. These formed compounds are more soluble than those formed for Pb2+.

As to Co2+ and Ni2+, the three soils present a low affinity to the two cations, which might indicate that the binding sites of studied soils had a larger affinity and selectivity for Pb2+, Zn2+, and Cu2+ than for Co2+ and Ni2+ (Table 2).

Nickel and Co2+ does not form insoluble precipitates in soils, and their retention are, therefore, exclusively regulated through sorption processes. Dähn et al. (2003) and Kim et al. (2006) found that the greatest amount of Ni2+ and Co2+ in soils was associated with clays, Fe, and Mn oxides. However, since these contents in the three soils are low and approximately the same, the two cations are the lowest sorbed and retained and present the same order in the selectivity sequences (Table 2).

Overall the high sorption and retention capacities of studied soils containing low proportion of known components that most commonly influence trace elements fixation (i.e., oxides, organic matter, and clay fraction), we suggest that trace element fixation may be due to the formation of element-precipitates, since the pH of the soils is alkaline. These sorption characteristics are in good agreement with previous studies on sorption of trace elements on Aridisols which have approximately the same characteristics as our selected soils (Elkhatib et al. 1991; Plassard et al. 2000; Fontes et al. 2000).

In addition, Bruemmer et al. (1986) has demonstrated that not only the alkalinity of soils lead to metal precipitation but also the high trace element concentrations in the added dissolution used in batch experiment can also be the reason to the low element’s solubility and also the small number of specific sorption sites. In our case, studied soils are characterized by their alkalinity due essentially to the presence of carbonates (Table 1) and also high trace element concentrations in sorption phase were used eventually which reached 5 mmol L−1. Those both factors may have led to trace element carbonate precipitation reactions.

Bivariate correlation analysis was carried out to study the correlations between the properties and components of the soils and the competitive sorption and retention of Pb2+, Zn2+, Ni2+, Cu2+, and Co2+ (Table 3). As can be seen above, the sorption of Pb2+ Zn2+, Ni2+, Cu2+, and Co2+ was strongly correlated with carbonate contents indicating that this component is the major contributor on sorption and retention of these trace elements (Table 3). Also, significant correlations were found between the sorption and retention of Pb2+, Zn2+, Ni2+, Cu2+, and Co2+ and with CECe, except for Cu2+ sorption. Fe oxide presented a significant correlation mainly between all studied soils and retention capacity. The lack of a correlation between clay contents (except for Ni2+ and Co2+ retention) and Pb2+, Zn2+, Ni2+, Cu2+, and Co2+ sorption and retention capacities was probably related to the fact that the soil samples had poor clay contents (Table 1). Significant correlations have been established between the K r for the competitive retention of Zn2+, Ni2+, and Co2+ and halite, kaolinite, quartz, and mica. Significant correlations have been established between the K r for the competitive retention of Zn2+, Ni2+, and Co2+ and halite, kaolinite, quartz, and mica. The positive correlation (Table 3) established between the studied trace elements content with the halite, kaolinite, quartz, and mica contents demonstrates that these minerals have a high capacity to retain Zn2+, Ni2+, and Co2+, which agrees with the results obtained by several authors (Miranda-Trevino and Coles 2003; Agoubordea and Navia 2009).

Table 3 Correlations between soil characteristics and the K r for sorption and retention of Pb2+, Zn2+, Ni2+, Cu2+, and Co2+
Full size table

Selectivity sequences of trace elements

Regarding the studied trace elements, Pb2+ and Cu2+ have been arranged in the first order of the sorption and retention sequences showing that their binding is much greater than coexisting elements binding and that Pb2+ and Cu2+ were less sensitive to the presence of the other cations. Abdelfattah and Wada (1981) reported that trace element affinity for 11 soil samples were in the order Cu ≈ Pb > Zn > Co. Results of the study on sorption onto quartz at different pH values revealed the following sequence of sorption: Pb > Zn > Ni > Cu (Honghai et al. 1999).

Jalali and Khanlari (2006) studied the sorption of trace elements on calcareous soils receiving spiked sewage sludge, and found that Pb2+ sorption was greater than coexisting elements.

Alternatively, an interchange was found in the order of Zn2+, Ni2+, and Co2+ in the obtained sequences (Table 2). Similar result was also reported by Jalali and Moharrami (2007), who measured the sorption of trace elements on ten Iranian calcareous soils and found that Cu2+ are always sorbed to a large extent than Zn2+ and Ni2+ which presented, as the sequences in our study, an interchange on their order on the sorption selectivity sequence based on K d distribution parameter. Furthermore, similarly to the obtained selectivity sequence in our study, some studies showed a Zn2+ and Ni2+ interchanging orders in the selectivity sequence of sorption, as indicated by the results of Antoniadis et al. (2007) for soils treated with sewage sludge and Fontes et al. (2000) for seven Brazilian soils.

In turn, McLaren et al. (1998) found that Co2+ was the most desorbed cations to soil solution on a binary sorption. According to the sequences reported by Backes et al. (1995), Co2+ is always desorbed to a large extent than coexisting elements on poor Fe and Mn oxide soils and soil clay minerals.

Hysteresis

Table 4 depicts the values of the hysteresis indices calculated from the sorption and retention capacity of the studied cations. Generally, the three soils presented high sorption irreversibility for Pb2+, Zn2+, Ni2+, Cu2+, and Co2+. The hysteresis of the sorption of Ni2+ and Co2+ on AR2 and AR3 is lower than on AR1, indicating their reversibility in these soils. On other hand, reversibility occurs in the soils with lowest clay fraction and oxides (Table 1). Respectively to the K r values, the soils which have the highest sorption and retention capacity present the higher hysteresis indices.

Table 4 Hysteresis indices HI
Full size table

A correlation analysis was carried out between the characteristics of the soils, and HI was studied (Table 5). The hysteresis of Pb2+, Ni2+, Cu2+, and Co2+ is correlated with pH, Fe oxide clay fraction, and carbonate contents meaning that these soils’ characteristics influence the irreversibility of the sorption of these cations. Apart from these soil characteristics, all the clay minerals present in it are also highly correlated with the irreversibility of the sorption of Pb2+, Ni2+, Cu2+, and Co2+ (Table 5).

Table 5 Correlations between soil characteristics and hysteresis indices
Full size table

Conclusion

In this study, sorption and desorption of Pb2+, Zn2+, Ni2+, Cu2+, and Co2+ in three Tunisian Aridisols were assessed with batch experiments at varying concentrations. The sorption and desorption isotherms, even the selective sequences of the trace element sorption based on the K r, showed that Pb2+ and Cu2+ are the most strongly sorbed trace elements by these soils, whereas Zn2+, Ni2+, and Co2+ were the least sorbed ones. The correlation analysis indicated that for the competitive sorption and retention, the soil properties that were most strongly related to trace element sorption were carbonate contents and CECe of soils. More studies, eventually microscopic, are required to understand which process is controlling the sorption on these soils.