1 Introduction

1.1 Contaminants in urban soil

Urban soils are specifically threatened by the contaminants such as trace metals (X. Li et al. 2013) and de-icing salt (Cunningham et al. 2008). These contaminants in urban soil can be toxic to soil microbes and invertebrates, trees, and human (Luo et al. 2012; Jim 1998). Mobile metals in urban soil can also contaminate underlying groundwater (Clark and Pitt 2007).

The mobility of trace metals in the soil is a function of metal characteristics, total concentration of metals in the soil, and various soil properties (Brümmer 1986; Linde et al. 2007; Waterlot et al. 2013). Some authors found a positive correlation between Cu retention and pH, soil organic matter (SOM), cation exchange capacity (CEC), and the sum of basic cations (mainly exchangeable Ca) (Covelo et al. 2007; Sherene 2010; Matos et al. 2001).

Zinc retention can be dependent on soil pH and SOM. According to Güngör and Bekbölet (2010), an increase in Zn release to the soil solution at high pH values can be attributed to the decrease in intra- and inter-molecular hydrogen bonds in humic acid molecules that make them more open and linear in shape and increase in repulsive forces between the dissociated functional groups. Zinc dissolution can also increase at low soil pH. The reason might be the dissolution of Zn bound to different soil constituents, such as calcite and Fe/Al oxides, which are positively charged, and probable H+–Zn2+ exchange processes occurring on the mineral surfaces. For a given pH, soils with higher SOM content can provide a greater sorption capacity than soils with less content of SOM (Díaz-Barrientos et al. 2003).

According to Rashti et al. (2014), soil features including pH, SOM, and CEC are negatively associated with Cd desorption. Various studies (Covelo et al. 2007; Matos et al. 2001) reported that Pb desorption from soil was negatively related to clay content, pH, and the content of organic matter. According to Strawn and Sparks (2000), by increasing SOM content, the rate of Pb sorption decreases. With higher levels of SOM in soils, there is often an increase in DOC. Dissolved organic carbon has the ability to form Pb complexes (Shahid et al. 2012). However, Strawn and Sparks (2000) observed a decrease in Pb desorption as the amount of SOM increased. They attributed the slow desorption reaction to the SOM fraction of the soil, and to the larger activation energy that desorption reaction requires as compared with that required for sorption.

Many studies indicated the high vulnerability of leaching trace metals in urban soils as a consequence of exposure to high concentration of NaCl as de-icing salt (F. Li et al. 2015). According to Nelson et al. (2009), the mobilization of trace metals under the effect of de-icing salt can happen through three main mechanisms including cation exchange, colloid dispersion, and Cl complex formation. When Na+ is present at high concentrations, it displaces naturally occurring Ca2+ and Mg2+. Exchange of Ca2+ and Mg2+ with Na+ in soil exchange sites and deflocculating soil particles can increase colloid mobility and therefore associated trace metal mobility (Green et al. 2008).

1.2 Organic Amendment of Urban Soil

It is commonly assumed that amending urban soil with organic materials such as compost can help the soil adsorb contaminants including trace metals and de-icing salts (Wuana and Okieimen 2011). Owing to a spontaneous microbial oxidation of numerous raw sources such as green wastes, compost can be produced as a biologically stable, humified organic matter end-product (Beesley et al. 2014). Pardo et al. (2014) indicated that amending soil with compost prepared from solid olive-mill waste significantly decreases the mobile concentrations of Cd, Cu, Pb, and Zn in contaminated mining loamy sand. Compost retains metals by raising soil pH, cation exchange, complexation, sorption, the presence of P, Al compounds, and other inorganic minerals or a combination of them (Karami et al. 2011; Paradelo et al. 2011; Bolan et al. 2014).

The application of biochar, as a porous, low-density carbon-rich material produced from biomass combustion under low-oxygen conditions, can also reduce the mobility of trace metals in contaminated soils and therefore decrease the risk of these metals being taken up by plants (Beesley et al. 2010; Beesley and Dickinson 2011). The potential of biochar for metal sorption can be related to the large surface area, CEC and porosity of biochar, and an increase in the soil pH after biochar addition (Houben et al. 2013a, b; Park et al. 2011; Kloss et al. 2014). Also, Biochar carries a negative charge on its surface; therefore, incorporation of biochar with soil can make the negative surface charge of the variable charge soils more negative. As a result, the adsorption affinity of the soil surface for cations such as trace metal cations will increase (Jiang et al. 2012).

Conversely, there is indication that biochar might not immobilize trace metals as efficiently as some other amendments. For instance, Hanauer et al. (2012) studied the sorption of Cd, Cu, and Zn in mining soil with neutral pH level and high content of clay amended with iron grit, natural zeolite, biochar, and Divergan (a scavenger) for 12 months. In this study, biochar was derived from peanut hull residues pyrolyzed at 480 °C and indicated the least effect on the sorption of the trace metals. The feedstock and pyrolysis conditions used for biochar production likely affect its trace metal sorption capacity (Cao et al. 2011; Uchimiya et al. 2010).

1.3 Sorption Capacity of Urban Soil

The mobility of metals in the soil environment is directly related to their partitioning between soil and solution, which is known as distribution coefficients (K d) (Evans 1989). The K d values indicate the capability of a soil for sorption and retention of a contaminant and the extent of its movement to the liquid phase (Reddy and Dunn 1986). A high K d value indicates high metal retention by soil solid phase, which results in low metal bioavailability. Similarly, a low K d value indicates that a higher proportion of the metal will be released to the soil solution (Shaheen 2009). According to Buchter et al. (1989), soil pH and CEC are the most important characteristics affecting log K d values of cation species. In general, high pH and high CEC soils retained greater quantities of the cation species than did low pH and low CEC soils.

This study is focused on determining how compost and biochar, used both singly and in combination, affect the chemical characteristics of urban soil (tree pit soil as a case study) and its sorption and retention of Cu, Zn, Cd, and Pb in the presence of de-icing salt (Na). The null hypothesis in this research is that neither compost nor biochar increased the tree pit soil sorption capacity for trace metals under the effect of de-icing salt. The objective of this research is related to the broader goal of designing a soil medium that can reduce groundwater pollution and promote a healthier and longer life for street trees.

2 Materials and Methods

2.1 Soil, Compost, and Biochar

In order to compare the ability of compost with biochar for decreasing the mobility of Na, Cu, Zn, Cd, and Pb in tree pit soil, sorption and desorption tests were conducted using soil mixtures with different percentages of compost and biochar. The fresh soil used as the basis for the soil mixtures was a sandy loam provided by the borough of Ville-Marie, Montreal, Quebec, Canada. The compost was derived from mixed green and table waste from the West Island region of Montreal. The biochar was made from maple wood pyrolyzed at 450 °C.

2.2 Preliminary Analysis of Soil, Compost, and Biochar

A preliminary analysis (pH, CEC, organic matter content, and texture) of the soil was compared with the standards for tree pit soil for the City of Montreal (Table 1) (City of Montreal 1995) and the analysis results of compost and biochar. Measurement of pH was done with an Accumet AR10 meter (Fisher Scientific Inc., Pittsburgh, PA) in a 1:2 soil-to-solution ratio (7 g soil/14 mL H2O) (Hendershot et al. 1993b). Exchange capacity for cations (Ca2+, Mg2+, K+, and Na+) was measured in soil mixtures, compost, and biochar extracted with 0.1 M BaCl2 following Hendershot et al. (1993a, b). Soil organic matter was measured by loss on ignition (Schulte et al. 1991). Soil texture was determined using a hydrometer, following Bouyoucos (1936).

Table 1 Preliminary results of soil analysis as compared to the standards for tree pit soil for the City of Montreal and the analysis results for compost and biochar
Full size table

2.3 Soil Mixture Preparation

To prepare the soil mixtures for the sorption and desorption test, the unamended, uncontaminated soil was dried at room temperature for 14 days and passed through a 2-mm metal sieve to remove large particles. Different percentages of compost (maximum 15 % by weight) and biochar (maximum 10 % by weight) were combined with the uncontaminated soil. A total of nine combinations of soil, compost, and biochar were used in triplicate in this study (Table 2) according to a central composite rotatable design.

Table 2 Compost and biochar percentages (by weight) in soil mixtures defined according to central rotatable composite design
Full size table

2.4 Sorption and Desorption Test

The sorption test was performed on the soil mixtures for 10 days in order to provide the soil mixtures sufficient time to equilibrate with the contaminants, as indicated by stable electrical conductivity (EC) values in the leachate measured by CDM 83 conductivity meter (Radiometer, Copenhagen, Denmark). Using a programmable vacuum extractor, columns containing 15 g of soil each were gradually leached at a constant rate (2–3 mL h−1) with 20 mL of solution per day. The solution used for contaminating the soil mixtures contained the concentrations shown in Table 3. For the desorption test, the leached soils obtained from the sorption test were dried at room temperature, weighed, and re-leached with 400 μM CaCl2-CaSO4 following the same procedure as in the sorption test.

Table 3 The component concentration of percolating solution in sorption and desorption tests
Full size table

2.5 Chemical Analysis

The contaminated soil mixtures (after sorption test) and the soil leachates (collected on the 10th day of desorption test) were analyzed for Na, Cu, Zn, Cd, and Pb using an inductively coupled plasma–mass spectrometer (820-MS, Varian, Melbourne, Australia) (Hendershot et al. 2008). The certified reference material used as the quality control of the analysis of total concentration of trace metals was SED 98-4 provided by the Environment Canada Proficiency Testing program. The recovery percentages of the SED 98-4 were as follows: 165 % for Na, 107 % for Cu, 100 % for Zn, 126 % for Cd, and 107 % for Pb. Exchangeable base cations (Ca2+, Mg2+, K+, and Na+) of soil mixtures after adsorption test were extracted with 0.1 M BaCl2 following Hendershot et al. (1993a, b).

The pH values and electrical conductivity of the soil leachates after desorption test were determined using an Accumet AR10 meter (Fisher Scientific Inc., Pittsburgh, PA) and CDM 83 conductivity meter (Radiometer, Copenhagen, Denmark), respectively. Dissolved organic carbon (DOC) was also determined in the soil leachates after desorption test using a Sievers Innovox TOC analyzer (GE Analytical Instruments Inc., CO, US).

2.6 Statistical Analysis

Using the concentrations of CaCl2-CaSO4 extractable metals in the soil treatments, a bivariate correlation analysis (Spearman correlation coefficient) was done between Na, Cu, Zn, Cd, and Pb in the soil samples and some selected soil characteristics (e.g., pH, CEC, EC, and DOC). The partitioning coefficient (K d), defined as the ratio of the total bound contaminant (mg kg−1) to that in soil leachate after equilibrium (mg L−1), was calculated for the studied metals. To identify the importance of compost and biochar in modifying the partitioning coefficients of the metals of interest in the soil mixtures, multiple linear regression was applied using SPSS statistics (V.21, IBM). The regression input data were first standardized using standard score (Z-scores) following Eq. 1, where x is the data value, μ is the average of the data, and \( \delta \) is the standard deviation of the data. The normality of the dataset was then confirmed using the Kolmogorov- Smirnov test:

$$ Z=\frac{x-\upmu}{\updelta} $$
(1)

The K d values were then used for calculating the steady state concentration of metals in the soil. The steady state concentration is the maximum concentration of metals retained in soil mixture, as the concentration of the metals in the soil leachates equals the concentration of the metals in runoff water. In this case, steady state concentration of metals can be calculated following Eq. 2:

$$ S={K}_{\mathrm{d}}\times \mathrm{Conc} $$
(2)

where S is the steady state concentration retained in soil mixture (mg kg−1), K d is the distribution coefficient (L kg−1), and Conc is the concentration of metals in runoff water (mg L−1). Because of the lack of the measurements for trace metal and de-icing salt concentrations in runoff water collected from streets in downtown Montreal in existing studies, the average concentration of the metals in runoff water collected from traffic areas in Germany, UK, Netherlands, and USA derived from Göbel et al. (2007) was used as Conc in Eq. 2. (Table 4). In order to find the optimum percentage of compost and biochar in soil mixture for sorption and retention of these metals, the mean values of the steady state concentrations of the metals in different soil mixtures were compared using the Duncan test in SPSS statistics.

Table 4 Average concentrations of metals in runoff water of traffic areas (Göbel et al. 2007)
Full size table

3 Results and Discussion

3.1 Metal Immobilization in Amended Soil Mixtures

The comparison of sorption capacity (K d values) (Table 6), calculated from total metal concentration in soil and in soil solution (Table 5), in different soil mixtures indicated the following order for mobility of metals: Na>Zn≥Cu>Pb>Cd. This result contradicts Öborn and Linde (2001), who observed higher mobility for Cd as compared with Zn, Cu, and Pb in urban soil leachates. The mobility of metals can be dependent on total metal concentration in the soil (Adhikari and Singh 2003; Luo et al. 2012) and their properties such as their electronegativity and charge-to-radius ratio (Paradelo et al. 2011). Mainly originated from de-icing salt (NaCl), Na has low electronegativity and low charge-to-radius ratio, which results in higher mobility of Na in soil leachates as compared to Cu, Zn, Cd, and Pb. Different studies observed high susceptibility of roadside soils to leaching of trace metals as a result of exposure to high concentration of NaCl as de-icing salt (F. Li et al. 2015). One of the mechanisms through which Na makes the roadside soil vulnerable to the leaching of trace metal such as Zn and Cd is cation exchange (Norrström 2005; F. Li et al. 2015). This can be confirmed by negative correlation of Na with Cd and Zn in the soil leachates collected after desorption test (Table 7).

Table 5 Chemical characteristics (pH, CEC, DOC, EC), sorbed metal concentrations, CaCl2-CaSO4 extractable metal concentration from control soil (SL), compost (CP), biochar (BC), and soil mixtures (T1 to T9)
Full size table

In soils exposed to high Na concentrations, soil aggregates break up. Dispersion of soil aggregates can release the organic matter and clay that complex with metal species. This in turn promotes the mobility of metals such as Cu via colloid-assisted transport (Nelson et al. 2009; Kluge et al. 2014). The significant positive correlation between EC and Cu (Table 7) in the leachate of the soil mixtures also corroborates this possibility. The positive correlation of high EC values of soil solution and the use of de-icing salts and elevated leaching concentrations of trace metals such as Cu in roadside soils is in line with other studies (Kluge and Wessolek 2012).

Soil conditions (Tables 5 and 7) play a crucial role in determining the mobility of metals. According to Strobel et al. (2001) and Shaheen (2009), the release rate of metals like Cu and Pb into the soil solution can be related to pH, as the positive correlation between Cu and Pb and pH in our results also confirmed that (Table 7). If the mechanism of holding Cu and Pb in the soil is associated with exchange sites on soil particles, they will be easily extractable and sensitive to pH, because of the solubility of Cu and Pb hydroxide species (L. Y. Li 2006). Increase in soil pH, however, can increase the retention of cationic metals such as Cd to soil surfaces via adsorption, inner-sphere surface complexation, and/or precipitation (Shaheen 2009). The negative correlation between soil pH and Cd concentration in soil solution (Table 7) also validates the increased Cd sorption with increased soil pH.

The mobility of metals like Cu and Pb in the soil leachates might also be attributed to the formation of soluble complexes with oxygen-containing functional groups of dissolved organic matter (Beesley and Dickinson 2011). This was corroborated by the positive correlation between Cu and Pb concentrations and DOC in the soil leachates (Table 7). The lower mobility of Pb as compared to Cu may be attributed to stronger Pb fixation onto soil organic matter and Pb affinity to other adsorbing surfaces such as Mn and Fe oxyhydroxides and clay minerals in the soil mixtures (Moreno et al. 2006; Trakal et al. 2011).

The negative correlation of Zn and Cd concentrations with exchangeable cations, such as Ca and Mg, in the soil leachate (Table 7) supported the idea that cation exchange was the main mechanism controlling Zn and Cd mobility (Cavallaro and McBride 1978). Positive correlation between exchangeable Ca and Mg and compost percentages in the soil mixtures (Table 7) corroborates that exchangeable Ca and Mg originate from organic amendments such as compost in the soil.

Our results indicated a low mobility for Cd. According to Kluge and Wessolek (2012), the low mobility of Cd in roadside soil can be related to high sorption capacity of soil and high pH values of roadside soil under the effect of de-icing salt, the influence of road building materials and alkaline dust from road surface abrasion. However, our result is in contrast to several other studies that indicated high mobility of Cd in urban roadside soil. F. Li et al. (2015) attributed the mobility of Cd to the formation of Cd-Cl complexes, which can penetrate to groundwater and surface water in areas exposed to de-icing salt.

3.2 Comparison of the Ability of Compost and Biochar to Improve Trace Metal Sorption in Soil Mixtures

The determination coefficients (R 2) of multiple linear regression (Table 8) based on the K d values (Table 6) indicated that the variation in partitioning of Zn and Pb between soil and soil solution was dominantly related to compost percentages in the soil mixtures. Different studies indicated the positive effect of compost on trace metal sorption in the soil. For instance, Paradelo et al. (2011) added compost from municipal solid waste at the rates of 3 or 6 % dry weight to a cropped soil artificially contaminated with Pb and Zn. They observed the reduction for 99 % of Pb and for 80 % of Zn solubility in 0.01 M CaCl2 extracts of soil amended with compost (6 % dry weight).

Table 6 K d values of control soil (SL), compost (CP), biochar (BC), and soil treatments (T1 to T9) for Na, Cu, Zn, Cd, and Pb
Full size table

The decrease in Pb and Zn availability in the soil amended with compost can be attributed to different factors. Increasing the strength of metal bonds to the soil matrix by displacing Pb from exchange sites to less labile forms such as metal complexes with humic and fulvic acids provided by organic matter in the soil can decrease Pb (Paradelo et al. 2011). This can be corroborated with the observed positive correlation between DOC and Pb in the soil solution after desorption test and also positive correlation between DOC and compost percentages in the soil mixtures (Table 7). According to Karaca (2004), the positive effect of compost on metal sorption may also be due to high CEC of compost. The positive correlation of compost with CEC and negative correlation of CEC and Zn concentration in the soil leachates (Table 7) can confirm the results of Karaca’s study (2004).

Table 7 Spearman correlation coefficients between trace metal concentrations in soil leachates and soil chemical characteristics after desorption test
Full size table

Comparing the steady state concentrations of Zn (Fig. 1c) and Pb (Fig. 1e) in different soil mixtures (T1 to T9), control soil (SL), compost (CP), and biochar (BC) indicated that CP, T2, T6, T8, and T9 for Zn and CP, T2, T3, T8, and T9 for Pb indicated significantly higher sorption capacity as compared to the other treatments. The comparison of the compost percentages in CP (100 % w/w), T2 and T3 (12.8 % w/w), T8 (15 % w/w), and T6 and T9 (7.5 % w/w) with other soil mixtures indicated that only in treatments with the percentages higher than 7.5 % w/w, a significantly positive effect of compost was observed on Zn and Pb sorption. However, no significant difference in sorption capacity was observed between treatments containing compost higher than 7.5 % w/w. Therefore, it was concluded that 7.5 % w/w can be selected as the optimum percentage of compost in soil mixture in order to increase sorption capacity of soil for Zn and Pb.

Fig. 1
figure 1figure 1

Steady state concentrations of Na (a), Cu (b), Zn (c), Cd (d), and Pb (e) in contaminated soil mixtures. The y-axis is scaled logarithmically. *SL, 100 % control soil; CP, 100 % compost; BC, 100 % biochar; T1, 2.2 % compost and 1.5 % biochar; T2, 12.8 % compost and 8.5 % biochar; T3, 12.8 % compost and 1.5 % biochar; T4, 2.2 % compost and 8.5 % biochar; T5, 7.5 % compost; T6, 7.5 % compost and 10 % biochar; T7, 5 % biochar; T8, 15 % compost and 5 % biochar; T9, 7.5 % compost and 5 % biochar (the percentages are by weight). Different letters on bar graphs indicate a significant difference

Full size image

Compost did not indicate any significant relationship with Na, Cu, and Cd retention in soil mixtures (Table 8). Comparing the steady state concentrations of Na and Cd in different soil mixtures after desorption tests (Fig. 1a, d) with the Duncan test indicated that there is no significant difference between soil mixtures (T1 to T9) and control soil (SL). This suggested soil as the main adsorbent of Na and Cd in the soil mixtures.

Table 8 The relationship of compost and biochar percentages with partitioning coefficient of Na, Cu, Zn, Cd, and Pb in soil mixtures based on Z-scores
Full size table

The wide range of compost percentages (2.2 to 12.8 % w/w) in the treatments with highest sorption capacity for Cu (T1, T2, T3, T5, and T9) and the lack of significant difference between the steady state concentration of Cu in these treatments and control soil (SL) indicated that soil may be a more important adsorbent of Cu as compared to compost (Fig. 1b).

Our results did not indicate any significant relationship between Na, Cu, Zn, Cd, and Pb partitioning coefficients and biochar percentages in the soil. Comparing steady state concentrations of Na, Cu, Zn, Cd, and Pb (Fig. 1) in soil mixtures (T1 to T9), control soil (SL), compost (CP), and biochar (BC) indicated significantly lower concentration of all the studied metals in biochar (BC). This finding contrasts with studies in which leachate concentrations of Cu, Zn, Cd, and Pb decreased in biochar-amended soils (Beesley et al. 2010; Karami et al. 2011; Houben et al. 2013a, b; Trakal et al. 2011).

A few studies have indicated that biochar may not always decrease the mobility of trace metals. For instance, Kelly et al. (2014) showed that in mine tailings with high concentrations of toxic elements, biochar made from pine wood did not reduce the leachate concentrations of Cu, Zn, Cd, and Pb. Kloss et al. (2014) also reported that the application of woodchip-derived biochar had no impact on Cu, Cd, and Pb concentration in leachates from agricultural soils. According to Hanauer et al. (2012), the sporadic behavior of biochar in retaining trace metals might be related to the feedstock and conditions of manufacture. Kloss et al. (2014) attributed the increase of metals in leachates from biochar-amended soils to the possibility of non-homogeneity of physicochemical properties of the soil-biochar mixture in the interaction with the percolating solution in the leaching process.

4 Conclusion

This study demonstrated that Na, Cu, Zn, and Pb in tree pit soils are vulnerable to leaching when exposed to a high NaCl concentration. In order to increase the sorption and retention of these metals in tree pit soil, compost and biochar indicated different function. The results revealed that soil amended with compost (7.5 weight percentage as an optimum percentage) had a significantly higher capacity for sorption of Zn and Pb than did biochar-amended soil. The positive effect of compost on metal sorption and retention may be attributed to its high cation exchange capacity and positive effect on soil pH. For Na, Cu, and Cd, soil appeared to have the higher sorption capacity as compared to compost. The presence of biochar did not improve the ability of the soils to retain contaminants. The properties of biochar vary widely depending on the feedstock from which it is produced and the pyrolysis conditions. Although compost indicated a positive effect on sorption and retention of some trace metal during short-term sorption-desorption tests, long-term field studies would help to understand the functional sustainability of compost by monitoring the quality of the soil, the soil solution, and the groundwater.