Retention of Copper and Zinc from Traffic Area Runoff by Topsoil of Vegetated Infiltration Swales Amended with Recycled Demolition Waste

1. Introduction

Traffic area runoff (TAR) constitutes a substantial source of pollution, transporting a diverse array of metals to the environment [1,2]. The metals copper (Cu), and zinc (Zn) from brake dust and tire wear, respectively, are expected to have the highest concentration in TAR [3,4]. An effective solution to manage metal contamination is the implementation of roadside infiltration swales, which can be integrated into stormwater management strategies and is the state of the technology in Germany [5].

Recently, there has been growing interest in using construction and demolition waste (CDW), which consists mainly of concrete and bricks, as an innovative, cost-effective material for runoff treatment [6,7]. CDW is one of the most significant waste streams in Europe, and its utilization could reduce CO₂ emissions and save landfill space [8,9].

In this context, the concept of technosols, or human-made soils created from the recycling of materials such as CDW, has gained attention as a key approach to promoting sustainability within the framework of the circular economy [10]. In landscaping, the development of local material flows is key to a sustainable construction sector, and local processing and use of CDW can contribute to this [11]. Repurposing CDW as a topsoil amendment in infiltration swales offers benefits: reducing the environmental burden of waste disposal and providing an efficient means of mitigating road runoff pollutants. However, the use of CDW is often hindered by issues related to acceptance and concerns about the release of potentially hazardous substances [12,13].

This three-year study investigates the potential of six CDW-derived technosols to remove Cu and Zn from genuine traffic runoff on a pilot scale for the first time compared to typical soil used in infiltration swales as a control. By examining the material composition-dependent leaching of substances based on batch tests and retention efficiency for Cu and Zn from real traffic area runoff under semi-real conditions, this research aims to evaluate their viability as an alternative to traditional filtration systems.

We hypothesize that topsoil from infiltration swales can be enriched with CDW without harming groundwater quality. In addition, we hypothesize that pollutant retention from traffic area runoff can be improved if the topsoil is enriched with CDW. We also hypothesize differences in the chemical binding of Cu and Zn in the technosols compared to typical soils used for infiltration swales.

To test the hypotheses, we investigated three-step phases:

Phase 1: Batch tests of different technosols (soil mixtures based on CDW) and the evaluation of the potential groundwater hazard

Phase 2: A field test of the technosols on the retention potential and efficiency of Cu and Zn when charged with real traffic area runoff

Phase 3: Sequential extraction using reference material BCR 701 (BCR-SEP) of Cu and Zn from the technosols with the highest CDW ratios after treatment with TAR to understand binding mechanisms.

The findings could contribute to developing sustainable solutions for urban pollution control while simultaneously addressing waste management challenges and advancing the circular economy.

2. Materials and Methods

2.1. Technosols from CDW

For the investigations, different technosols were produced based on two recycling mixtures (RCM; RCM1 and RCM2) that derive from building demolition and processing its CDW. The RCMs are crushed aggregates with grain sizes of 0 to 16 mm. RCM1 is enriched with brick material. It contains 60 % bricks and 40 % concrete, plaster and mortar. RCM2 is enriched with concrete, plaster, and mortar (70 %) and contains 30 % bricks (Figure 1). These RCMs have been blended with natural topsoil, subsoil, and compost to achieve comparable particle size distributions and organic matter contents (Table 1). The resulting six technosols ((a75, a50, a25, b75, b50, b25) containing different CDW ratios and compositions and one control soil (ctl) that did not contain CDW were produced.

2.2. Study Design

For lysimeter tests, seven vessels (Polyethylene, DN 776 mm, height = 500 mm) were buried 80 cm below ground in 2021 and filled with the six different technosols and ctl (see ). An extensive grass mix was then sown over the surface areas of all lysimeters to establish the vegetation. Boundary rings were placed over the surfaces of the vessels (High-Density Polyethylene, DN 800). A plastic hose connected the outlets (1/2”) at the bottom of the vessels to manholes where the seepage was quantified and collected using tipping counters (100 mL, polycarbonate, REED-Sensor, max. flow rate = 5 L/min).

The assessment considered two thresholds: German threshold values for seepage water (TV_1) and for soil solution (TV_2) according to the German Federal Soil Protection and Contaminated Sites Ordinance (Bundesbodenschutz- und Altlastenverordnung; BBodSchV) [15].

Phase 1 – Leaching behaviour

To assess the leaching of substances for the pathway soil-groundwater in advance, laboratory batch tests according to DIN 19529 with a liquid-to-solid ratio of 2 L/kg batch aqua test were conducted with RCM1, RCM2, the six technosols, and the ctl in triplicate [16]. Obtained eluates were analyzed for selected parameters according to the threshold values TV_2 for soil solutions.

Besides total Cu and Zn concentrations in the eluate, the investigated parameter spectrum included the total concentrations of the metals Sb, As, Pb, B, Cd, Cr, Cr(VI), Co, Mo, Ni, Hg, Se, Tl, V, and Sn. Additionally, sulfate, chloride, fluoride, and dissolved organic carbon (DOC) were analyzed to evaluate potential risk for groundwater contamination (List of methods, limits of quantification (LOQ) and threshold values see Table A1).

Phase 2 – Retention of Cu and Zn from traffic area runoff

After backfilling the lysimeters, the six technosols and the ctl were exposed to natural weather conditions to ensure stabilization until the Cu and Zn discharges had normalized to avoid biasing seepage concentrations when charging with TAR and measuring background concentrations. The quantities of seepage water up to the start of the TAR charging were recorded using tipping counters and reported as L/S in the results section. Mean background concentrations of Cu, Zn, V, and DOC are based on the values of 4 random samples per technosol between June 2023 and February 2024.

From April 2024 to August 2024, the lysimeter vessels were charged 12 times (6 per rainfall intensity) with TAR, collected at a highly trafficked road (approx. 24,000 vehicles/day; [17]). The TAR was transported to the study site, stored in containers, and pumped to the test plots using gear pumps. The flow rate was measured using flow meters. Two rainfall intensities were applied according to measured data from the German Weather Service (KOSTRA data): RI_1: 121 L/(s·ha), volume per charging = 139 L, and RI_2: 221 L/(s·ha), volume per charging = 127 L. Charging volume and flow rate of TAR were calculated according to the German guideline DWA-A 138-1, assuming a ratio of 15 m² of asphalt surface (mean runoff coefficient = 0.9) to 1 m² of active infiltration area in a swale. Influent and total seepage of the lysimeters were sampled and analyzed for Cu and Zn concentrations after aqua regia (AR) digestion and compared to the threshold values TV_1 for seepage water

Phase 3 – Sequential extraction of Cu and Zn from soil

Soil samples from a75, b75, and ctl were taken in triplicate from the infiltration areas and, as a control, from the non-infiltration areas at depths of 0 - 25 mm. From these soil samples, Cu and Zn concentrations were further analyzed following the three-step optimized sequential extraction protocol (BCR-SEP) using BCR 701 (Community Bureau of Reference, European Commission) reference material according to Rauret, et al. [18]. All used utensils were made of borosilicate glass, polypropylene, or polytetrafluoroethylene (PTFE). The extractions were carried out in 80 mL centrifuge tubes (Herolab, Wiesloch, Germany), which were cleaned with 4 mol/L HNO₃ before use and then rinsed with ultrapure water [19]. From each sample, 1 ± 0.01 g was weighed into the centrifuge tubes three times. Additionally, as a reference BCR-701 sediment (European Commission, Joint Research Center, Belgium) was added to each batch for quality control. The individual steps of the procedure, including the different fractions and chemical solutions, are briefly described in Table 2. Calculated total Cu and Zn contents are the sum of fractions S1 to S4.

Due to difficulties with separation during centrifugation, the speed was increased from 3000 g to 4000 g. To compare the total concentrations from S1 to S4, an AR digestion was performed directly in each batch with 3 ± 0.1 g per technosol plus a BCR-701 reference. The Zn and Cu concentrations in the extracts were determined using inductively coupled plasma optical-emission spectrometry (ICP-OES).

In addition, all technosols’ dry weight and loss on ignition (LOI) were determined. For this purpose, 1 ± 0.01 g of each technosol in triplicate and one BCR-701 reference per batch was weighed and dried overnight at 105 ± 2 °C in an oven until a constant weight was achieved. These values were used as correction factors, and all analytical results in this work refer to the dry mass of the samples. To determine the LOI, the samples were heated in a muffle furnace at 550 °C for 2.5 h.

2.3. Data Processing

For comparisons of means, p-values were calculated using the Mann-Whitney U test with Python’s scipy package [20]. The data was processed with Python’s pandas [21], and the plots were created with Python’s matplotlib [22] and seaborn [23].

3. Results

3.1. Leaching Potential

The concentrations, pH values, and EC of the leached substance in the batch tests are summarized in Table 3.

Most metal concentrations, also those for Cu and Zn, in the eluate of all CDW mixtures, the six technosols, and the control were low and under the threshold values of TV_2.

Exceptions were V, and Cr(VI). Here, the threshold values for TV_2 were exceeded. The highest V concentrations in the eluate were found for the CDW mixtures between 66.7 µg/L (RCM2) and 167 µg/L (RCM1), slightly lower concentrations for the six technosols (up to 43.3 µg/L for a75). Also, the V concentrations of 5 µg/L in ctl were higher than the threshold values of 4 µg/L in TV_2. The Cr(VI) concentrations were highest in the CDW mixtures with 25.0 µg/L (RCM 1) and 38.7 µg/L (RCM 2). In comparison, the threshold value TV_2 is 8 µg/L. The technosols and the ctl did not exceed the threshold value for Cr(VI).

Sulfate concentrations in the eluate of the two CDW mixtures RCM1 and RCM2 were the highest. However, in all technosols, sulfate concentrations were lower, ranging from 9.6 mg/L (b25) to 52.3 mg/L (a75). Fluoride concentrations in the eluate of all tested technosols, RCM1 and RCM2, and ctl were low and did not exceed the threshold values of TV_2 which is 1.5 mg/L. Additionally, chloride concentrations were also low. DOC concentrations in the technosols and ctl were low and ranged from 5.7 to 7.3 mg/L.

3.2. Retention and Water Quality

Figure 2 shows the retention performance for Cu and Zn from the traffic area runoff by the different technosols compared to ctl (cf. Table A3).

Mean influent concentrations of Cu were 73.4 µg/L at RI_1 and 84.7 µg/L at RI_2. While the Cu concentrations in the influent of the lysimeters were still above the threshold values of TV_1 (50 µg/L), in the leachate of all experiments (RI_1 and RI_2), the values are consistently below 20 µg/L. However, we also analyzed the seepage water between June 2023 and February 2024 for mean background concentrations, the background values (5 to 6.5 µg/L) were lower compared to the leachate concentrations (up to 19.4 µg/L for a50) after treatment with traffic area runoff. Based on Mann-Whitney U tests with a significance level of α=0.05, the effluents of a50, a25, and b50 at RI_1 showed significantly higher Cu concentrations than the control (p=0.019, 0.005, and 0.005, respectively). At RI_2, the effluents of the technosols showed no significant differences in Cu concentrations compared to the control.

Similar observations can be made for zinc, but differences between RI_1 and RI_2 (Figure 2). Mean influent concentrations of Zn were 247 µg/L at RI_1 and 68.1µg/L at RI_2, both below the threshold value TV_1 of 600 µg/L. Mean effluent concentrations of Zn of all technosols and the control were much lower, ranging from 9.4 µg/L (a25) to 16.2 µg/L (a50) at RI_1 and 15.7 µg/L (a50) to 38.5 µg/L (a25) at RI_2. RI_2 concentrations were higher compared to RI_1 and higher than the background values. Based on Mann-Whitney U tests with a significance level of α=0.05, effluent concentrations of Zn of all technosols including the control were significantly lower than in the influent (cf. Table A2) both at RI_1 and RI_2. The Zn discharges showed no significant differences compared to the control at RI_1. At RI_2, a25 showed marginally higher Zn concentrations in the effluent compared to the control (p = 0.041).

Mean influent concentrations of DOC were 7.4 mg/L at RI_1 and 10.4 mg/L at RI_2. Mean effluent concentrations of DOC were in the range of 9.4 (a25) to 15.4 mg/L (a75) at RI_1 and 13.6 (a25) to 21.6 mg/L (a75) at RI_2. DOC was significantly higher in the effluents of all technosols and the control compared to influent concentrations at RI_1 and RI_2. Compared to the control, effluent DOC was significantly higher in a75 and a50 at RI_1. At RI_2, effluent DOC was significantly higher in a75 and significantly lower in a25 compared to the control.

The pH values were comparable stable and were between 7.5 and 7.9 for all tests and technosols and the ctl.

The mean influent EC was 2,588 µS/cm at RI_1 and 352 µS/cm at RI_2. Mean effluent EC ranged from 2,270 µS/cm (b75) to 2896 µS/cm (a25) at RI_1 and from 643 µS/cm (a25) to 703 µS/cm (b25) (ctl 576.5 ± 257.4) at RI_2. At RI_1, the EC of effluents of all technosols, including the control, were not significantly different from influent EC. At RI_2, the EC of effluents of all technosols, including the control, was, on the other hand, significantly higher than the influent EC. Compared to the control, no technosol effluent showed significant differences in EC at RI_1 and RI_2.

We evaluated the retention performance of the technosols for Cu and Zn for the tests depending on the rainfall intensities (Figure 3) based on the total loads (cf. Table A4). At RI_1, mean retention efficiencies for Cu ranged from 78.3 % (a25) to 94.8 % (a75) (ctl: 90.2 %) and from 50.3 % (a75) to 80.6 % (ctl 78.7 %) at RI_2. At RI_1, mean retention efficiencies for Zn ranged from 93.7 % (a25) to 98.0 % (b25) (ctl: 93.2 %) and from 48.2 % (a25) to 83.3 % (b50) (ctl 75.5 %) at RI_2. Overall mean retention efficiencies of Cu were highest in b25 (85.6 %) and lowest in a75 with 72.5 % (ctl 73.9 %). Overall mean retention efficiencies of Zn were highest in b50 with 90.3 %) and lowest in a25 with 71.0 % (ctl 84.3 %).

RI_1 and RI_2, showed no significant differences in Cu retention of technosols compared to the control soil. At RI_1, there were no significant differences in Zn retention of technosols compared to the control. At RI_2, only a25 showed significantly lower Zn retention efficiency than the control (p= 0.03). Total retention efficiencies of technosols showed no significant differences compared to the control. However, across all soils, retention efficiencies for Cu and Zn were significantly reduced at higher rainfall intensity (p = 0.003 and p < 0.001, respectively).

3.3. Sequential Extraction of Cu and Zn

After treatment with TAR, Cu and Zn accumulation was measured in the technosols a75 and b75 and the control (Figure 4).

Based on the sequential extraction, the residual Cu fraction was the dominant component in all samples and always accounted for more than 60% of the total content (Figure 5). In the untreated substrates, the residual Cu content is higher in ctl compared to a75 and b75. The following order was determined in the treated soils: ctl_treat > b75_treat > a75_treat.

The second largest Cu fraction in all substrates was the oxidizable, accounting for approximately 30% in the untreated soils, 22.3 % in ctl_treat and 35.9 % in a75_treat. Again, in the untreated substrates, the Cu contents were higher in ctl compared to a75 and b75. A similar pattern was seen in the treated substrates: ctl _treat > a75_treat > b75_treat.

The reducible and acid-exchangeable Cu fractions were < 1 mg/kg in all samples, and each represented less than 2% of the total contents. An exception was a75, where 0.87 mg/kg was measured for the acid exchangeable Cu fraction, corresponding to 5.45 %.

Considering the relative frequencies of mobilizable Cu (sum of S1-S3), a75_treat and b75_treat showed increased mobile phases of Cu compared to ctl_treat.

In summary, the proportion of the residual Cu fraction increased after treatment while the mobile Cu phases (S1-S3) decreased, except for a75_treat. The mobile Cu phase increased slightly, shifting from reducible to oxidizable fractions.

Of all treated soils, ctl_treat had the highest total Zn content, followed by a75_treat and b75_treat. However, the content of Zn in the untreated substrates was also higher in ctl than in a75 and b75. In the untreated soils, the fractionation obtained from BCR-SEP for a75 and b75 was as follows: residual >> reducible >> acid exchangeable > oxidizable, while for ctl, the order was: residual >> reducible > oxidizable > acid exchangeable. In the treated substrates, the fractionation in ctl_treat was identical to ctl, while in a75_treat and b75_treat, it was: residual >> reducible > acid exchangeable > oxidizable.

The residual Zn fraction was dominant in all samples. The reducible Zn fraction was the second largest in all samples. The highest reducible Zn fraction was found in substrate a75, followed by ctl and b75. The treated samples showed the following order: a75_treat > b75_treat > ctl_treat. The substrates a75 and b75 differ from ctl in terms of Zn fractionation. In the control soil, the oxidizable Zn fraction ranked third in the untreated and treated. In a75 and b75, the acid exchangeable Zn fraction was the third largest.

The lowest Zn contents of ctl were found in the acid exchangeable fraction, both treated and untreated. For a75 and b75, the oxidizable Zn fraction was the smallest in treated and untreated. In summary, the residual fraction increases after TAR treatment while the mobile phase decreases, except in a75_treat. There, the proportion of the mobile Zn phase remains nearly unchanged, with a shift of the reducible fraction towards the acid-exchangeable fraction.

Comparison of the soils after TAR treatment

To assess shifts in the chemical bindings of Cu and Zn after treatment with TAR, the treated substrates’ mean Cu and Zn contents were subtracted from those of the untreated soils (Table 4). The most pronounced shifts of Cu contents were found for a75, where oxidizable Cu was nearly as high as residual Cu, showing a decrease in reducible Cu on the other hand. b75 also showed a relative increase in oxidizable Cu, albeit less pronounced. In ctl, most Cu accumulation occurred in the residual phase. Analogous, the most pronounced shifts of Zn were found in a75, where a significant accumulation of Zn was found in the acid exchangeable phase, and only half of the Zn accumulation was found in the residual phase. In b75, most of the Zn accumulation occurred in the residual phase. In ctl, all highly mobile phases stayed the same or decreased, and only the residual phase increased.

4. Discussion

4.1. Risk of Substance Release

Except for Cr(VI) and V, the leaching of substances obtained from batch tests at L/S 2 was largely unremarkable compared to the German threshold values for soil solution TV_2. Concentrations of Cr(VI) in both RCM1 and RCM2 exceeded the threshold value substantially. However, Cr(VI) concentrations were below the threshold value in the six technosols, which will be normally applied in practice. The concentration of V in RCM1 exceeded the threshold value TV_2 and was substantially higher than in RCM2. Hence, while V concentrations were lowest in the control, in the technosols, we see a dose-response of V with an increasing ratio of RCM1, which, in contrast, is vaguer for RCM2. Crushed bricks release substantially more sulfate, Cr, and V than crushed concrete [24]. Vanadium release from CDW that contains bricks can be three times higher than in CDW based solely on concrete [13]. Our data confirm that an increased proportion of bricks in the CDW increases the release of some elements like Cr and V [13,24,25,26]. While in German legislation, an L/S-dependent decay behavior is assumed to evaluate recycled building materials for Cr discharges, there is no generalizable decay behavior for V [27]. While the high pH of mixed CDW is related more to its concrete than its brick phase [24], in our study, the pH of both CDW mixtures derived from the batch tests was high (> 10, cf. Table 3). Over time, the increased specific surface area and increased air and water supply during storage facilitate the carbonation of crushed CDW [13], resulting in a pH between 8 and 9 [28,29,30,31] while partly carbonated CDW is suggested to have a pH 10.5 to 11.9 [29,32]. This indicates that the CDW mixtures were not yet carbonated as they were sampled. Carbonation is associated with increased leaching of anionic compounds when the aging of concrete leads to the release of sulfate and its substituting ions, chromate, and vanadate [13,33]. Consequently, in our study, the sulfate leaching from the mixed CDW, the technosols, and the control correlated strongly with the leaching of Cr(VI) (r² = 0.81) and V (r² = 0.71). Furthermore, V release in soils is supposed to correlate to higher pH [34]. Therefore, we attribute the reduced release of V from the soil to a reduction in the pH value by adding soils with a neutral pH and to a reduced release of sulfate in the technosol mixtures. For EC, we observed a dose-dependent response, increasing with higher ratios of both RCM 1 and RCM 2.

4.2. Retention and Binding

The traffic area runoff TAR concentrations of Cu and Zn used as influent in this study were comparable with concentrations reported for urban roads with AADT > 15,000 vehicles [1]. The Zn/Cu ratio at RI_1 was 3.4, which concurs with the literature [1]. However, Zn concentrations and thus Zn/Cu ratio of TAR used for RI_2 (0.8) were lower than expected. We attribute this to the different sampling periods. The TAR for RI_1 was sampled in late winter/early spring, whereas the TAR for RI_2 was sampled during the summer months. In winter, metal loadings in traffic area runoff are expected to be higher, which is often caused by the application of de-icing salts and increased corrosion rates [1,35,36].

Metals from TAR were found to bind to soil particles by precipitation or adsorption, forming strong bonds that are difficult to change under typical environmental conditions [1,37]. Copper occurs in topsoils in oxic conditions, mostly in oxidation state II, and is not very exchangeable at pH > 5; its solution concentration is mainly determined by adsorption and desorption, depending on pH and soluble, organic or inorganic complexing agents [38]. We assume that due to the high carbonate content of the soils analyzed, Cu is present in the form of carbonate and organometallic complexes [38]. In addition, we assume that the microbial degradation of the substrate compost releases soluble organic complexing agents and promotes the mobilization of adsorbed Cu [38]. On the other hand, the content of exchangeable Zn is very low at pH > 6.5 as Zn²⁺, Zn(OH)⁺, and ZnCO₃ are mainly present here, of which Zn(OH)⁺ in particular increases proportionally with increasing pH and is specifically adsorbed and fixed by oxides [38]. Furthermore, metals transported by the TAR to roadside soils are bound to particles, which in turn are largely retained by physical mechanisms, particularly in the upper soil layers [2,39]. Therefore, metal mobilities were determined in roadside soils depending on pH and organic matter [40,41,42].

Accordingly, Zn retention in our study was higher than Cu retention, comparable to findings reported from the US [43]. Our study found no correlations between Cu and Zn discharge in the effluents and its pH, EC, or DOC. Although DOC concentrations in the effluents were higher than in the influents (presumably due to the addition of compost), we suggest that due to the high charging rates applied in our study, retention times in the soil were not long enough to promote the complexation of Cu and organic matter. In addition, we can rule out unfavourable redox conditions due to the high oxygen content in the effluents. As the Cu and Zn retentions measured for RI_2 were significantly lower than RI_1, we consider it rather likely that the mechanical filtering effect of the coarse-grained soils with high water permeability was no longer present at the high charging rate. Hence, we assume that retention of particulate-bound Cu and Zn was impaired. This assumption is supported by the fact that there was no backwater in the plots during the irrigations, even at a high charging rate. Nevertheless, the Cu and Zn retentions of the studied technosols were not significantly lower than that of the control. The effluent concentrations of Cu and Zn of all analyzed technosols and the control were below the threshold of the German BBodSchV.

After the treatment with TAR, the direct AR pseudo-total analysis was generally unable to extract the same metal content as the sum of the BCR-SEP fractions (S1-4). However, we found accumulations compared to the non-treated soils for Cu and Zn. The relatively lower concentrations in the pseudo-total can also be reflected in the pseudo-accuracy of the reference material BCR-701, with the residual fraction slightly overestimated and the direct AR pseudo-total rather underestimated (cf. Table A5). The large difference cannot be fully explained. Possible reasons could be that the samples were not ground before direct digestion, leading to incomplete extraction of metal concentrations and that the AR digestion technique is more accurate for analyzing the residue than the raw material [19]. Davidson, et al. [44] and Rommel, Stinshoff and Helmreich [19] reported comparable deviations of summarized SEP fractions and pseudo-total digestion from sequential extractions of soil and filter media, respectively. Also, for these reasons, the recoveries are sometimes very poor (cf. Table A5). They are in the 109-234% range, which means that the sum of the fractions of metal concentrations is sometimes more than twice as high as the pseudo-total. Ideally, the values should be between 80-120% [45].

For Cu and Zn, the residual fraction increases slightly after treatment (approx. 2-11%), except in a75_treat, where a small increase in the mobile phase was observed. The residual fraction is >53% in all substrates. Due to the pseudo-accuracy calculated for this fraction, we attribute discrepancies to overestimating the residual fraction. Sutherland, et al. [46] reported lower residual fractions of Cu (35 %) and Zn (21%) in road-deposited sediments (RDS). In contrast, Pérez, López-Mesas and Valiente [37] found residual fractions of up to 80% for Cu in RDS. Rommel, Stinshoff and Helmreich [19] have found comparably high residual fractions in carbonate sand used to treat TAR from the same origin we used in this study. Zn and Cu are found predominantly in particulate form in traffic area runoff TAR, making the upper soil layer critically important. In this layer, metals are primarily retained mechanically by sedimentation and filtration, while chemical processes such as adsorption and binding to organic matter also contribute significantly to their immobilization [1,2]. Rommel, Stinshoff and Helmreich [19] sampled the TAR used in this study at the same sampling site as in this study and found concentrations of total Cu and Zn approx. 7 times higher than dissolved Cu and Zn.

For Cu, the oxidizable fraction is the second largest (approx. 22 to 36 %), while for Zn, it is negligible (about 5 to 11 %). In a75, treatment with TAR caused a shift of Cu from the reducible fraction (from 5.5 to 0.45 %) to the oxidizable fraction (from 29.2 to 35.9 %). Bacon and Davidson [47] point out that Cu expected in the oxidizable can also be released in the reducible phase, making interpretation difficult. Similar contents in the oxidizable fraction (approximately 26 %), which also correlate with organic carbon, were observed by Sutherland, et al. [48] and Kartal, et al. [49] in studies of RDS in Honolulu, Hawaii, and Kayseri, Turkey (cf. Table A6).

a75_treat also shows a shift of Zn, with a decrease in the reducible fraction (from 25.8 to 20.4 %) and an increase in the acid-exchangeable fraction (from 10.3 to 17.3 %). The latter, the most mobile and potentially harmful fraction, is strongly influenced by pH. Metals become more mobile the more acidic the soil. This behaviour is opposite to the pH in the substrates, especially for a75_treat and a75_treat. Bacon and Davidson [47] also found that a phase shift can occur during the extraction of Zn, which can distort the results. The findings of this study differ from other studies, where higher fractions of acid-exchangeable Zn were reported, ranging from 25 to 38 % [37,46,48,49].

A comparison of only the mobile fraction (S1-S3) shows a similar pattern in fractionation for Zn in the substrates a75_treat and b75_treat as in other studies [46,48,49]. However, the proportions of the individual fractions vary considerably, with overall higher Cu and Zn contents, especially in the acid-exchangeable fraction reported in these comparative studies.

No comparable pattern emerges for Cu. However, [37] reported a similar sequence of fractions, with the residue even accounting for about 80 %. In contrast, Kartal, Aydın and Tokalıoğlu [49] and Sutherland, Tack and Ziegler [46] reported the following order for Cu: acid-exchangeable < oxidizable < reducible. In contrast to b75_treat and ctl_treat, in a75_treat was an increase of Cu in both the mobile and residual phase.

b75, which contains higher proportions of concrete, mortar, and plaster, shows favourable distribution ratios than a75. This finding is in concordance with Pallewatta, et al. [50], who have found better retention in concrete-based waste than in masonry-based waste. Barrett, Katz and Taylor [3] point out that Portland cement concrete in roadside soils fosters metal retention mainly by controlling pH and precipitation. Furthermore, we hypothesize that in a75, where higher iron contents due to a higher brick ratio are to be expected, the presence of humic-coated Fe-oxide colloids combined with also higher Ca contents inherent in the CDW promotes the transport of Cu [51]. On the other hand, high Cu retention [6,7] and Zn retention [52] were found for brick or brick-dominated aggregates.

However, despite overall comparable Cu and Zn retentions when charged with TAR, both CDW-amended technosols showed unfavourable binding properties compared to the control. This is due to relative increases of mobile Cu and Zn phases which could be released by various environmental influences (e.g., changes in pH, de-icing salt input, enrichment of organic matter).

5. Conclusions

Technosols amended with brick-dominated CDW showed an dose-dependent increased leaching potential of V, which correlates with sulfate emissions. Therefore, V and sulfate content should always be analyzed and high contents should be avoided when bricks are used to amend topsoils of infiltration swales to not harm groundwater quality. However, CDW-amended technosols can be used in infiltration swales regarding the retention of metals from traffic area runoff. They had comparable Cu and Zn retentions as the control soil with the same particle size distribution. Considering the chemical binding of Cu and Zn, the amendment of concrete-dominated CDW is favourable to brick-dominated CDW. Still, both showed an increase of mobile Cu and Zn phases compared to the control, implying an increased potential for desorption and remobilization. Considering sustainability, using recycled construction and demolition waste in infiltration swales could be a sustainable solution for reducing CO₂ emissions and resource consumption.