Removal of dissolved heavy metals from stormwater by filtration with granular recycled glass and mussel shell with and without microalgae biofilm

Environmental Technology & Innovation 18 (2020) 100662

Contents lists available at ScienceDirect

Environmental Technology & Innovation journal homepage: www.elsevier.com/locate/eti

Removal of dissolved heavy metals from stormwater by filtration with granular recycled glass and mussel shell with and without microalgae biofilm Courtenay Bremner, Thomas A. Cochrane, Peter McGuigan, ∗ Ricardo Bello-Mendoza Department of Civil and Natural Resources Engineering, University of Canterbury, Private Bag 4800, Christchurch 8140, New Zealand

article

info

Article history: Received 7 September 2019 Received in revised form 6 February 2020 Accepted 7 February 2020 Available online 13 February 2020 Keywords: Microalgae Copper Lead Zinc Stormwater treatment

a b s t r a c t Stormwater treatment technologies are routinely used to reduce urban stormwater contaminant loads entering surface water and groundwater systems. However, in most cases, their treatment capabilities are limited to particulate contaminants. Research into alternative filtration materials for stormwater treatment is necessary to provide sustainable and environmentally friendly options for the removal of dissolved heavy metals. In addition, periphytic biofilms are known to develop on substrate materials and little is known about their effect on the sorption of dissolved heavy metals. A laboratory study was conducted to assess the effectiveness of recycled glass, mussel shell, and freshwater periphytic biofilm to remove dissolved metals (copper, lead and zinc) from stormwater. Synthetic stormwater containing typical concentrations of heavy metals (Cu, 154 µg/L; Pb, 41 µg/L; Zn, 280 µg/L) was used for all experiments. The stormwater was recirculated through 1-L columns filled with granular recycled glass or mussel shell either colonised with periphytic biofilm (mainly microalgae) or not. The biofilm demonstrated an ability to sorb high concentrations of the dissolved metals. However, its contribution to heavy metal removal was much lower than that of the substrates on their own. In some cases, the biofilm actually reduced the ability of the substrates to remove Cu and Pb presumably by occupying adsorption sites on the filter media. The recirculation of a fixed volume of synthetic stormwater over a 48-hour period resulted in a stable removal efficiency for the three metals with the trend of removal being lead (>97%) ≥ zinc (97%) > copper (89%) for mussel shells and lead (96%) > zinc (86%) > copper (73%) for recycled glass. Both the composition and surface texture of the substrate seem to influence the removal of dissolved heavy metal from aqueous solutions. The use of waste products such as recycled glass and mussel shells for stormwater treatment might contribute to waste minimisation and to increase the value associated with the extended lifecycle of these materials. © 2020 Elsevier B.V. All rights reserved.

1. Introduction Stormwater transports urban contaminants at concentrations frequently above regulatory limits, resulting in the degradation of local waterways when discharged into these environments. The most common contaminants found in ∗ Corresponding author. E-mail address: [email protected] (R. Bello-Mendoza). https://doi.org/10.1016/j.eti.2020.100662 2352-1864/© 2020 Elsevier B.V. All rights reserved.

2

C. Bremner, T.A. Cochrane, P. McGuigan et al. / Environmental Technology & Innovation 18 (2020) 100662

stormwater runoff include sediments, organic micropollutants (e.g. polycyclic aromatic hydrocarbons and polychlorinated biphenyls), pathogenic microorganisms, nutrients, and heavy metals (Barbosa et al., 2012). These contaminants build up on urban surfaces because of anthropogenic activity and are washed off in stormwater during precipitation. Heavy metals, particularly zinc, copper and lead, are found ubiquitously in stormwater typically above regulatory guidelines and are a significant concern because of their bioavailability and toxicity to aquatic ecosystems (Seelsaen et al., 2006). Dissolved heavy metals are considered bioavailable and can be readily assimilated by aquatic organisms causing acute toxicity; they are also difficult to treat because of their dissolved nature. The removal of a dissolved metal ion requires that either the ion be precipitated from the solution or adsorbed to the surface of a solid. In addition to physicochemical conditions such as pH, each metal species has a differing affinity to solid surfaces, which results in differences in their dissolved concentration within stormwater. Studies show that as much as 90% of the total zinc concentration in stormwater is found in the dissolved form; while copper can be found in both particulate and dissolved form, the latter of which may comprise 40%–80% of the total concentration in stormwater (Pennington and Webster-Brown, 2008; Charters et al., 2016). It is therefore crucial that stormwater treatment systems target the dissolved and most ecotoxic form of heavy metals to effectively reduce environmental degradation. Stormwater Treatment Systems (STS) have been developed as a way of returning an area’s post-development hydrology to near natural conditions (Ahiablame and Shakya, 2016). STS can include contaminant filtration, retention, evapotranspiration, sorption, precipitation, biodegradation, phytoremediation, and percolation processes as a means of treating and controlling stormwater. Recently, readily abundant and recyclable materials such as mussel shells and recycled glass have been investigated for use in stormwater treatment. Mussel shell is an abundant, alkaline, waste product from the shellfish industry that has been shown to successfully remove heavy metals from stormwater through pH buffering (Good et al., 2014). Mussel shells increase the pH of the stormwater preventing the dissolution of particulate-bound heavy metals and facilitating the sorption of dissolved heavy metal ions to particulates. This ultimately allows the particulates, and associated heavy metals, to be filtered out through the STS media. Recycled glass is a municipal waste product that has demonstrated potential to be utilised in wastewater treatment and potable water treatment. Published studies on the use of recycled glass for stormwater treatment are limited to the work of Seelsaen et al. (2006) who found that fine recycled glass had zinc and copper removal capabilities of up to 70% and 40% respectively. Another potential way to enhance removal of heavy metals is through the promotion of the development of biofilms on the filtering media or substrate. Biofilms are collections of microorganisms attached to moist abiotic surfaces and enclosed in a matrix of extracellular polymeric substances (EPS) (Donlan, 2002; Singh et al., 2006). Biofilms offer mechanical strength to microorganisms, which is not available to their free-living counterparts, which protects the biofilm community from shear forces, nutrient deprivation, and other environmental changes (e.g. pH) (Singh et al., 2006). Biofilms possess a range of properties that make them highly effective at removing heavy metals from metal-contaminated water (Feder et al., 2015). Together, the EPS and the microorganisms within the biofilm can facilitate the sorption of heavy metals through passive (biosorption) and active (bioaccumulation) cellular processes. These processes are possible because of the functional groups found in both the EPS and in the cell walls of microorganisms (e.g. amino, carboxylic, hydroxyl, phospholipids, proteins, and polysaccharides) (Hu et al., 2005; Feder et al., 2015). While soil, mulch, and vegetation in STS contribute to the removal of particulate heavy metals from stormwater by sedimentation and filtration, dissolved heavy metals can remain untreated and are discharged in stormwater outflows at high concentrations. The typically short hydraulic retention times in STS means that other processes such as adsorption to soil or uptake by plants are not very effective at removing dissolved heavy metals. Biofilms have a natural affinity for dissolved metals and thus could offer a means to improve the bioremedial properties of bioretention systems by reducing concentrations of dissolved heavy metals in stormwater outflows. Experiments by Feder et al. (2015) showed that biofilm growth in gravel filters significantly enhanced metal removal by 29%. Another study by Ancion et al. (2010) showed that biofilms can accumulate metals quickly and to high concentrations, with the most significant accumulation occurring during the first 7 days. As noted before, it is crucial that STS target the dissolved and most ecotoxic form of heavy metals to effectively reduce and prevent environmental degradation. Research has demonstrated the ability of some non-conventional STS materials to adsorb heavy metals from aqueous environments, and current understanding would suggest that the presence of biofilms in STS may improve the removal of dissolved heavy metals from stormwater. However, there is a need to understand the role of periphytic biofilms in the sorption of dissolved heavy metals from stormwater in an effort to improve the efficiency of STS. The aim of this research was thus to evaluate the potential use of recycled glass and mussel shell for dissolved heavy metal treatment and to assess whether the growth of periphytic biofilms (mainly microalgae) on these substrates can improve the removal of dissolved heavy metals in stormwater treatment systems. 2. Material and methods To ascertain the effectiveness of microalgae biofilms and alternative substrates (i.e. recycled glass and mussel shell) in removing heavy metals from stormwater, a series of laboratory-scale column experiments were conducted.

C. Bremner, T.A. Cochrane, P. McGuigan et al. / Environmental Technology & Innovation 18 (2020) 100662

3

Fig. 1. Left: A diagrammatic representation of the column setup for experiments. Right: Image of experimental set up in the laboratory. Each column had its own synthetic stormwater solution reservoir, with influent inflow at the bottom and effluent outflow at the 1000-mL mark.

2.1. Experimental setup and procedure Eight filtration columns were constructed in the laboratory for use in the experiments. The columns were constructed using 1000-mL Polymethylpentene (PMP) cylinders (internal diameter of 60 mm), and each column had a tube connector fixed into the base and an outflow pipe fitted at the 1000 mL level (Fig. 1). Substrate was added to the columns to a total depth of 300 mm, according to the specific treatment conditions, which is the recommended minimum media depth specified by the Christchurch City Council rain garden design construction and maintenance manual (CCC, 2016). Each column had its own 1-L solution reservoir. A 4-L batch of synthetic stormwater was made up as needed at the beginning of the experiments and divided evenly into the 1-L solution reservoirs of each column. At the start of the experiment the synthetic stormwater was pumped, using a peristaltic pump (5 mL/min), through the inflow valve at the bottom of the column and ‘treated’ effluent was drained back into the solution reservoir, thus circulating the synthetic stormwater (Fig. 1). Preliminary experiments revealed that the greatest pollutant removal occurred within the first 24 h, which dictated the frequency of sampling and physicochemical analysis in later experiments. Sampling from the reservoir and synthetic stormwater characterisation occurred every 2 h for the first 6 h, then 6 h after that (i.e. 12 h from the beginning of the experiment) and finally 24 and 48 h after the start of the experiments. The experiments ended after 48 h (2 days). Treated synthetic stormwater samples were taken from the respective column reservoir. For the actual experiments, four columns were filled with each substrate (i.e. recycled glass or mussel shell) to a depth of 300 mm (or approximately 830 mL), three columns of each substrate were used to grow the biofilm; while the remaining column for each substrate was a control (no biofilm). Additionally, a blank experiment was conducted using a column without substrate to ensure that the column walls and tubing themselves would not contribute to the removal of heavy metals; in all cases, heavy metal removal was negligible. Only the results of a final 48-h run are presented. 2.2. Substrate characterisation The recycled glass (RG) and mussel shell (MS) were sourced from local manufacturers. Before use, all materials were sieved, washed with deionised water (DI water), and dried. The RG had a particle size of ≤ 16 mm. The MS was crushed and sieved, and material passing the 2.36 mm sieve but retained on the 1.18 mm sieve was used in the experiments. The RG was used as is after it was washed and dried. The RG and MS substrates were analysed for background heavy metal concentrations, pH, hydraulic conductivity and porosity. A single composite sample of each substrate was soaked in DI water (mass to volume ratio of substrate to DI water of 1:2) for at least 24 h and the liquid was then analysed by Inductively Coupled Plasma Mass Spectrometry (ICP MS) for dissolved zinc, copper, and lead. The hydraulic conductivity was measured using a modified version of the constant-head saturated hydraulic conductivity method and porosity using the saturation method (Klute, 1965).

4

C. Bremner, T.A. Cochrane, P. McGuigan et al. / Environmental Technology & Innovation 18 (2020) 100662

Table 1 Recycled glass and mussel shell characteristics. Average soluble metal concentrations (µg/kg)

Substrate

Recycled glass (RG) Mussel shell (MS)

Zinc

Copper

Lead

1.095 1.201



1.507

pH

Hydraulic conductivity (m/h)

Porosity (%)

9.3 ± 0.1 9.3 ± 0.2

37.4 ± 5.0 21.5 ± 3.2

41.5 ± 1.2 55.5 ± 0.8

D.L. = Detection limit. Table 2 Individual chemical characteristics (mean ± SD Wt%) of the RG (n = 6) and MS (n = 3). Recycled glass (RG)

Mussel shell (MS)

Element

Wt%

Element

Wt%

Oxygen Silicon Sodium Calcium Magnesium

45.8 ± 0.7 33.6 ± 0.7 10.9 ± 0.3 5.3 ± 0.2 3.3 ± 0.4

Oxygen Calcium Sulphur Aluminium Silicon

69.4 ± 1.6 21.7 ± 0.7 3.7 ± 0.4 1.7 ± 0.3 1.7 ± 0.3

Both RG and MS substrates had low concentrations of background heavy metals, meaning they were unlikely to influence the effluent concentration of dissolved heavy metals (Table 1). Both substrates had similar pH values and high hydraulic conductivity values. This is likely the result of the highly angular nature of RG, however, MS displayed a greater porosity than the RG. Surface images of the RG and MS were taken using a scanning electron microscope (SEM), model JEOL JSM-IT300 at the University of Canterbury. The SEM utilised model Oxford 50 mm2 SDD detector with Aztec software for additional EDS analyses. Scanning electron microscope analyses (Supplementary information, Figures S2 & S3) showed the surface of the RG to be smooth and angular with particulates, and in some cases, ingrained striations; the cause of these striations is likely a result of manufacturing. The MS had porous surfaces, with particulates on the surface. Results of the energy-dispersive x-ray spectroscopy (EDS; Table 2) show the chemical characteristics (mean ± SD Wt%) of the RG and MS. 2.3. Biofilm growth and digestion Dense microalgae biofilm was grown on the column substrates over a 6-week period. Scrapings of natural periphytic biofilm from Haytons Stream (Christchurch, New Zealand) were collected and used to seed the columns to initiate the growth of biofilm. Water collected from the same stream was used as the primary solution to promote biofilm growth within the columns. Stream water samples were collected under baseflow conditions and upstream known sources of metals such as air-conditioning and stormwater discharges. Previous studies have shown background metals concentrations to be low (<15 µg/L) under these conditions (O’Sullivan et al., 2012). Dosing of the stream water solution with nutrients and carbon was necessary to accelerate biofilm growth. The columns were clear acrylic, allowing exposure to sunlight during normal daylight hours (approximately 9 h) to promote algae growth and establishment. Control columns were covered with foil to prevent the growth of microalgae biofilm. Experiments commenced 6-weeks after seeding, once biofilm growth was considered established in all columns; this stage was determined visually based on biofilm growth throughout each column. Algae populations and composition were determined through microscopic analyses. Images of the biofilm were taken using a Nikon Eclipse 80i compound microscope at 20 and 40 times magnification. Microscope images revealed a periphytic biofilm with a diverse composition. Dominant groups included blue–green algae (Cyanobacteria), green algae (Chlorophyta), and diatoms (Bacillariophyceae), as well as microscopic animals and bacteria (Supplementary Information, Figure S1). At the end of the column experiments, biofilm was removed from the recycled glass columns and digested (method 3030 E; APHA, 2005) to determine the concentration of heavy metals within the biofilm itself. All the recycled glass, with biofilm attached, of each column was tipped into a beaker and washed with DI water. The biofilm was detached from the recycled glass by vigorous shaking and the columns were scraped to remove any biofilm from the internal wall of the column. The collected liquid biofilm mixture was left to settle for 24 h and the supernatant was then poured for the settled biomass to be recovered. Three 100 mL subsamples were transferred to beakers and 5 mL of concentrated HNO3 was added. The samples were boiled until digestion was complete as indicated by a light coloured, clear solution. The digested solution was filtered into a volumetric flask and diluted to make a 100 mL solution. Portions of this solution were taken for total heavy metal analysis by ICP MS in accordance with method 3125 B (APHA, 2005). 2.4. Synthetic stormwater Experiments were conducted using a synthetic rather than natural stormwater. Since the focus of these experiments was on the three most common heavy metals found in stormwater (i.e. zinc, copper, and lead) it was preferred not to

C. Bremner, T.A. Cochrane, P. McGuigan et al. / Environmental Technology & Innovation 18 (2020) 100662

5

Table 3 Influent (hour = 0) synthetic stormwater characterisation for pH, specific conductance, alkalinity, and dissolved heavy metal concentrations. Values represent mean ± SD (n = 3). Parameter

Synthetic stormwater

Dissolved heavy metal concentration

pH

Specific conductance (µS/cm)

Alkalinity (mg/L as CaCO3 )

Zn (µg/L)a

Cu (µg/L)b

Pb (µg/L)c

5.8 ± 0.1

11.8 ± 0.3

14.7 ± 10.9

285 ± 18

109 ± 6

33 ± 3

Chemical sources: a Zinc chloride (ZnCl2 ). b Copper chloride (CuCl2 ). c Lead acetate (Pb(CH3 COO)2 ).

use natural stormwater given its composition variability. Using a synthetic stormwater allowed for more consistency and repeatability of the composition of the stormwater. Stock solutions of zinc, copper, and lead were prepared using reagent quality chemicals. A 4-L batch of synthetic stormwater was made up as required at the beginning of each round of experiments using the heavy metal stock solutions and DI water. Nutrients were also added to the synthetic stormwater to ensure the growth and survival of algal biofilm during the experiments. Contaminant and nutrient concentrations levels (Supplementary Information Table S1) approximated typical concentrations of constituents found in natural stormwater and used in other studies (Seelsaen et al., 2006; Blecken et al., 2010). Additionally, dissolved organic carbon (DOC) was added to the solution as a carbon source for biofilm growth. Both the specific conductance and alkalinity of the synthetic stormwater fall within levels recorded for natural stormwater (e.g. Datry et al., 2003). The pH value was slightly more acidic than reported for natural stormwater (e.g. Mosley and Peake, 2001), but is typical of the pH reported for precipitation (e.g. Pennington and Webster-Brown, 2008). Typical dissolved heavy metal concentrations were achieved in the synthetic stormwater (Table 3). 2.5. Water quality analysis Influent synthetic stormwater samples, for heavy metal analysis, were taken from each solution reservoir immediately prior to the commencement of the experiments; effluent samples (5 mL) were taken at each sampling time thereafter. All heavy metal samples were sampled following method 3030 B (APHA, 2005) and analysed for dissolved heavy metals using an ICP MS (Agilent 7500cx, Agilent Technologies, USA), in accordance with method 3125 B (APHA, 2005). Before analysis, the samples were filtered through a 0.45 µm PVDF syringe filter shortly after sampling and preserved with trace grade nitric acid to pH ≤ 2. The ICP MS method detection limit was determined to be 1 µg/L for all three metals. Calibrated EDT and YSI meters were used to measure pH and specific conductance respectively. Total alkalinity was measured at the beginning and the end of each experimental run. Where possible, a 200 mL sample of synthetic stormwater was taken and total alkalinity was determined using the APHA 2320 B titration method (APHA, 2005). Removal efficiency and normalised values were calculated from the concentrations measured in the synthetic stormwater influent and effluent. 3. Results 3.1. Changes in water quality parameters The pH of both RG and MS effluent was slightly alkaline whereas the influent (t = 0) was slightly acidic (pH = 5.8 ± 0.1) (Table S2 in Supplementary information). Furthermore, the MS biofilm treatment achieved a pH consistently above 9 (range = 9.1–9.8) compared to the neutral pH of the MS control (MS without biofilm) with an average pH of 7.3 (range = 6.6–7.7). The RG biofilm treatment showed some variation over time but maintained an alkaline pH (range = 7.4–8.9). As with the MS, the RG control (i.e. RG without biofilm) had a lower pH (range = 6.7–7.6) than the RG with biofilm treatment. Differences in pH were observed since the first sampling point (t = 2 h) for mussel shell and from the second sampling point (t = 4 h) for recycled glass. For example, in the mussel shell experiments, on average, the pH of the biofilm treatment was 2 standard pH units greater than the control after 2 h. After 4 h in the recycled glass the difference between biofilm and the control columns were as much as 1 standard pH unit. Specific conductance differed considerably between the RG and MS substrates, with the effluent of both substrates increasing in specific conductance compared to the influent (influent specific conductance = 11.8 ± 0.3 µS/cm). This increase was most noteworthy for the MS (without biofilm = 392.7 µS/cm; with biofilm = 394.7 ± 2.7 µS/cm). RG showed a modest increase in specific conductance more so in the biofilm treatment (79.1 ± 7.9 µS/cm) than the control treatment (42.7 µS/cm). No external carbonate alkalinity was added to the synthetic stormwater and as a result its initial total alkalinity was low (14.7 ± 10.9 mg/L as CaCO3 ). The RG substrate did not change the alkalinity of the effluent to a great extent (RG

6

C. Bremner, T.A. Cochrane, P. McGuigan et al. / Environmental Technology & Innovation 18 (2020) 100662

Fig. 2. Removal efficiency (%) of dissolved zinc shown for recycled glass (RG) and mussel shell (MS), with no biofilm (Control) and with biofilm in the column (Biofilm). Error bars show standard deviation values (n = 3) for RG and MS biofilm column results.

Fig. 3. Removal efficiency (%) of dissolved copper shown for recycled glass (RG) and mussel shell (MS), with no biofilm (Control) and with biofilm in the column (Biofilm). Error bars show standard deviation values (n = 3) for RG and MS biofilm column results.

without biofilm = 18.7 mg/L as CaCO3 ) but the MS substrate produced a large increase in effluent alkalinity (MS without biofilm = 115.0 mg/L as CaCO3 ). Nevertheless, the biofilm activity produced the largest increase in stormwater alkalinity after 48 h of treatment and in both cases the effluent alkalinity was comparable (RG with biofilm 234.9 ± 2.6 mg/L as CaCO3 and MS with biofilm = 252.1 ± 2.6 mg/L as CaCO3 ). 3.2. Heavy metal removal efficiency Both the RG and MS, with and without biofilm, demonstrated high dissolved metals removal efficiencies within the first 24 h. 3.2.1. Zinc removal The RG biofilm treatment showed considerably greater removal efficiency than the control treatment without biofilm between 2 and 24 h of experiments (Fig. 2 and Table S3 in Supplementary information). Within 6 h the biofilm treatment reached and maintained removal efficiencies of over 80%, with the greatest removal efficiency of 88% recorded at 12 h. By 48 h, both filters (control and biofilm) achieved similar percent removal (control = 82% and biofilm = 87%). The difference between MS treatments was less substantial than that seen in the RG experiments, yet, overall the MS achieved higher removal efficiency than the RG over time (Fig. 2 and Table S3 in Supplementary information). Initially, the biofilm treatment showed the best removal efficiency, reaching 91% removal after 6 h. After 12 h both MS treatments showed similar removal efficiencies, with the control treatment reaching 97% and the biofilm treatment reaching 98% and remained consistently high until 48 h. 3.2.2. Copper removal Both RG and MS substrates demonstrated ability to remove dissolved Cu. Positive removal trends could be seen for both RG filters (Fig. 3 and Table S4 in Supplementary information). The RG control consistently obtained a higher removal for dissolved copper than the RG biofilm treatment, with a maximum of 83% at 24 h. Similarly, both MS control and biofilm filters showed positive removal efficiencies over 48 h (Fig. 3 and Table S4 in Supplementary information). Interestingly, the removal efficiencies of both filters remained similar for the first six hours (control = 71% and biofilm = 75%). Thereafter the MS control surpassed the MS biofilm removal efficiency and continued to increase until it reached a maximum removal efficiency of 97% at 24 h. The biofilm treatment also continued to increase, to a lesser extent, and reached a maximum removal of dissolved copper (83%) after 48 h.

C. Bremner, T.A. Cochrane, P. McGuigan et al. / Environmental Technology & Innovation 18 (2020) 100662

7

Fig. 4. Removal efficiency (%) of dissolved lead shown for recycled glass (RG) and mussel shell (MS), with no biofilm (Control) and with biofilm in the column (Biofilm). Error bars show standard deviation values (n = 3) for RG and MS biofilm column results. Removal efficiencies for MS Biofilm and Control and for RB Control at 24 and 48 h are >97% according to detection limit of the ICP-MS analysis method. Table 4 Concentrations of Zn, Cu, and Pb within the biofilm (after digestion) and in the synthetic stormwater solution, and their respective enrichment factor values. Biofilm (mg/kg dry weight) Synthetic stormwater (mg/L) Enrichment factora

Zinc

Copper

Lead

620.7 0.28 2216

362.1 0.11 3292

65.5 0.03 2183

a

The enrichment factor was calculated from the concentration in the biofilm divided by the concentration in the synthetic stormwater.

3.2.3. Lead removal Dissolved lead was effectively removed by both substrates and achieved concentrations below 3 µg/L in all experiments after 48 h. The removal efficiency of both substrates reached over 90% after just 12 h (with one exception being the biofilm treatment in the RG) demonstrating that both substrates can efficiently remove dissolved lead from synthetic stormwater over time (Fig. 4 and Table S5 in Supplementary information). The addition of biofilm in the RG columns did not prove to be beneficial and overall the experimental control showed a higher removal efficiency than the biofilm treatment. The MS experiments showed excellent removal abilities with over 94% removal of dissolved lead achieved within the first 12 h (Fig. 4 and Table S5 in Supplementary information). Initial removal efficiency was greatest in the MS biofilm treatment; however, after 12 h the removal efficiency of both MS control and treatment experiments were effectively identical (control = 94% and average biofilm = 95%). 3.3. Biofilm digestion Following the experiments, biofilm was removed from the RG columns and tested for heavy metals. The freshwater biofilm grown for this study demonstrated an ability to sorb dissolved heavy metals from an aqueous solution with typical stormwater concentrations of dissolved zinc, copper, and lead (Table 4). Adsorbed zinc reached the highest concentration within the biofilm with 620 mg/kg of dry biofilm weight, followed by copper with 360 mg/kg of dry biofilm weight, and lastly lead with 70 mg/kg of dry biofilm weight. These concentrations were achieved after 48 h exposure to the synthetic stormwater. Initial biofilm growth occurred without known sources of heavy metal contributions and therefore initial concentrations of heavy metals are assumed to be negligible. 4. Discussion 4.1. Sorption of dissolved heavy metals by biofilms Exposure to natural light through the acrylic columns resulted in a biofilm which is greater in biomass, thickness, and species richness when compared to biofilm grown in dark conditions (Sekar et al., 2002). Natural biofilm collected from a filter drain will have a composition dominated by bacteria (Feder et al., 2015), meaning the biofilm grown in this study differed in composition from what may be expected to grow in STSs, because it contains the autotrophic microorganisms Chlorophyta and Bacillariophyceae. However, it is likely that biofilm growth on the inner substrate surfaces of the columns (not exposed to light) would have better replicated the dark environment of STS. Since metabolism is not an expected pathway for heavy metals removal, the identification of the bacteria was not regarded as critical for this study, although, it is highly probable that bacteria were present in the biofilm grown (e.g. Feder et al., 2015). The intention was to observe EPS production as a potential medium for metals sorption.

8

C. Bremner, T.A. Cochrane, P. McGuigan et al. / Environmental Technology & Innovation 18 (2020) 100662

Results from the biofilm digestion provided evidence that heavy metals accumulated within the biofilm to very high concentrations, confirming the biofilm’s ability to sorb heavy metals from aqueous environments. A variety of mechanisms can lead to the sorption of heavy metals from aqueous solutions by biofilms, which include active processes (bioaccumulation, biomineralisation, and biotransformation) and passive processes (biosorption, including physicochemical removal mechanisms) (van Hullebusch et al., 2003; Feder et al., 2015; Farag et al., 2007). In a study by Ancion (2010), concentrations of zinc, copper, and lead in a freshwater biofilm were reported to be 204.3, 74.6, and 153.4 mg/kg wet weight respectively after 21 days exposure to a highly-contaminated synthetic urban runoff. Enrichment factors given by the same study showed the biofilm growth accumulated very high concentrations of heavy metals, up to 500:1, 1500:1, and 6000:1 for zinc, copper, and lead respectively (Ancion, 2010). Those authors attributed these concentrations to a long exposure time (21 days), with an equilibrium (defined as the concentration of metals in the water and biofilm in a steady state) not reached before 7 to 14 days, suggesting a more permanent metal accumulation process than biosorption alone (Ancion, 2010). The biofilm in the present study had enrichment factors much greater than those reported by Ancion (2010), except for lead, despite being exposed for a significantly shorter period of time. This may be attributed to the differing removal processes occurring on different time scales. For example, biosorption is a rapid process which can reach equilibrium within an hour of exposure, compared to more permanent accumulation processes which can take much longer (Hu et al., 2003, 2005; Ancion, 2010). Additionally, the rate of biosorption can decrease with increasing heavy metal concentrations due to competition between ions for binding sites on the biofilm surface (Jang et al., 2001; Costley and Wallis, 2001; Feder et al., 2015). The substrates, particularly the mussel shells, produced an increase in the effluent pH. In addition to that, the biofilm activity also contributed to the increase in the effluent pH. The observed biofilm pH trend may be due to the phototrophic nature of the dense microalgal biofilm. Photosynthesis causes changes to the pH of a system as a result of carbon dioxide (CO2 ) uptake and release. In particular, it causes increases in the concentration of hydroxyl ions in a solution. When bicarbonate (HCO3 − ) is consumed and the concentration of hydroxyl ions increases so does the pH of a system (Hu et al., 2005; Beck et al., 2011; Feder, 2014). 4.2. Substrate heavy metal removal Despite the biofilm demonstrating an ability to sorb dissolved heavy metals, it did not have as big an impact on heavy metal removal as the substrates on their own. Furthermore, although high pH is known to increase the removal of dissolved cationic heavy metals (e.g. Blecken et al., 2011), it was not possible to isolate this effect in this study. When columns containing biofilm were compared to those without biofilm, the difference in removal efficiency after 48 h was minimal. That is, the biofilm did not have a significant impact on heavy metal removal relative to the substrate on its own. In some cases, the presence of biofilm appeared to inhibit the removal of dissolved heavy metals in both the recycled glass and mussel shell substrates. The rate of removal of heavy metal ions, however, depends on the type of metal being removed, the substrate the biofilm is grown on, and the timing. The MS consistently reached over 80% removal efficiency for all three heavy metals regardless of whether biofilm was present or not. Identical orders of efficiency were seen for the biofilm coated MS (Pb (>97%) > Zn (97%) > Cu (83%)) and the MS control (Pb (>97%) ≥ Zn (98%) > Cu (97%)). However, MS control showed a much higher removal efficiency of Cu (over 10 percent point greater removal) compared to the biofilm coated MS. The recycled glass showed good removal abilities but had greater variation within and between treatments. After 48 h, the biofilm demonstrated an order of efficiency as follows Pb (91%) > Zn (87%) > Cu (68%) and without biofilm, the removal efficiency had an order of efficiency of Pb (>97%) > Zn (82%) ≥ Cu (82%). In the columns without biofilm the removal efficiency of zinc decreased by approximately 5 percent points compared to the biofilm, while, the removal of copper increased by 14 percentage points when no biofilm was present. When examining our results over time and for specific metals, our results align with findings from other studies. Biofilm is known to effectively occupy and smother the reactive sites of the substrate resulting in decreased adsorption (Kulczycki et al., 2005; Anderson et al., 2006). In particular, the adsorption capacity of granite rock with no biofilm can be up to 88% greater than rock covered with biofilm (Anderson et al., 2006). However, a study by Feder et al. (2015) reported the opposite, whereby they found that biofilm grown on a gravel substrate immobilised up to 29% more heavy metals than those with no biofilm. Furthermore, they found notable increases in the removal of lead (+14%) and zinc (up to 8%) when biofilms were present, which they presented as evidence that microorganisms can sequester metals more readily than the gravel surface alone (Feder et al., 2015). The discrepancy in results may be attributed to the substrates used, the metals being examined, and the timing for assessing the biofilms effect. In our study, RG and MS provide greater opportunities for metal removal than other substrates, and the effect of biofilms was overshadowed at the end of the 48 h treatment. However, for a specific timeframe during our experiments, the biofilm had a positive effect on Zn removal for both RG and MS. 4.2.1. Recycled glass characterisation The recycled glass used in these experiments had an angular shape. This is consistent with the lower sphericity values (0.41–0.43) of granular glass compared to sand grains (0.70–0.76) as reported by Soyer et al. (2013). The results of EDS analyses are also comparable to those by Ibrahim et al. (2012), who described a similar chemical composition of recycled

C. Bremner, T.A. Cochrane, P. McGuigan et al. / Environmental Technology & Innovation 18 (2020) 100662

9

glass, which was dominated by identical elements but differing in Wt%, as follows: silicon (70 Wt%), sodium (14 Wt%), calcium (11 Wt%), and magnesium (3 Wt%). The recycled glass used in this study showed a very high hydraulic conductivity of 37.4 m/h compared to the 0.1 m/h hydraulic conductivity reported in the literature for other recycled glass studies (Disfani et al., 2012). The porosity of the recycled glass (42%) is comparable to the values reported in other studies (e.g. 49% porosity reported by Soyer et al., 2013); however, it is at the top end of porosities measured for typical stormwater treatment system substrates (Fetter, 2000). This is unsurprising given that recycled glass is a highly angular substrate that could be considered more free-draining than other natural aggregates utilised in stormwater treatment systems (Disfani et al., 2012). This is a crucial element as a higher porosity can limit the ability of the recycled glass to filter particulates (Soyer et al., 2013). 4.2.2. Heavy metal removal efficiencies by recycled glass A minimum exposure time of 24 h would maximise removal for RG. Dissolved copper had the lowest removal efficiency of the three heavy metals and a longer exposure time did not increase its removal efficiency, such that a maximum removal efficiency occurred after 24 h exposure which did not increase after 48 h. Lead consistently achieved the highest removal by recycled glass, achieving over 90% removal efficiency in both columns and a maximum removal efficiency of >97% at 24 h. Few studies have investigated the use of recycled glass for the treatment of heavy metals in stormwater, but the few that do support the removal of heavy metals by recycled glass. For example, Seelsaen et al. (2006) found that recycled glass had a removal efficiency of 16%–69% and 26%–39% for zinc and copper respectively, depending on size fraction. Meanwhile, Petrella et al. (2009) recognised an additional advantage of recycled glass in that it successfully removed lead ions from wastewater. Interestingly, both studies identified different removal mechanisms for heavy metal ion removal by recycled glass. Petrella et al. (2009) identified adsorption, specifically ion exchange, as the key removal mechanism for lead ion removal due to the equivalent release of sodium ions to the liquid phase, which were detected in the effluent. However, Seelsaen et al. (2006) attributed the removal of zinc and copper ions to precipitation which was greater in the fine glass fraction due to higher dissolved organic carbon (DOC) leaching rates. This is because DOC can change the solution pH thus increasing the precipitation of heavy metal ions from a solution. It is important to note that the recycled glass used by Petrella et al. (2009) in their research was porous. Recycled porous waste glass is the result of the further manufacturing of the raw glass material whereby it is heated and expanding agents are added to create a light and porous structure. Adsorption is a surface-based process in which dissolved heavy metal ions are transported to the porous surface of an absorbent (the solid on to which the ions are absorbed) by diffusion and are then adsorbed to the extensive surface area of the absorbent (Xu and McKay, 2017). Therefore, the porous nature of the recycled glass used by Petrella et al. (2009) would support their ion exchange theory. Precipitation, on the other hand, occurs in aqueous solutions when a change in geochemical conditions converts dissolved heavy metal ions into an insoluble solid phase via a chemical reaction with a precipitant agent (Lewis, 2017). Typically, the heavy metal precipitated from the solution is in the form of a hydroxide, but it can also be in the form of chlorides, sulfates or sulfides, and carbonates (Kurniawan et al., 2006; Lewis, 2017). Precipitation is a pH dependent process such that at higher pH values, such as the one afforded by the recycled glass (influent pH = 5.5 ± 0.5, average pH across all experiments = 7.9 ± 0.5), the heavy metal ions can change between dissolved and particulate form via precipitation (Mosley and Peake, 2001). Once in particulate form, the substrate can filter the particulates out of the solution. This is crucial given that stormwater treatment systems typically rely on contaminant removal through the filtration of particulate-bound heavy metals. The smooth surface of the recycled glass used in the present study supports precipitation as the main heavy metal removal mechanism. The EDS results also show small amounts of potassium and sodium present in the composition of the recycled glass; these ions are known to be exchangeable cations that are exchangeable with other cations in solutions such as lead and zinc (Erdem et al., 2004). Ion exchange may thus be working as a secondary removal mechanism in the present study. The high removal efficiencies of lead and zinc, but not copper, further support this hypothesis. Poor copper removal efficiency may be attributed to the presence of organic matter within the recycled glass systems. Recycled glass is known to leach DOC (Seelsaen et al., 2006) and dissolved copper is known to correlate significantly with DOC concentrations in natural stormwater (Mosley and Peake, 2001). Copper has a strong affinity for organic matter and can form stable Cu-DOC complexes mobilising the complex into solution, which ultimately results in higher concentrations of copper in the effluent (Antoniadis and Alloway, 2002; Blecken et al., 2011). Despite higher concentrations of copper being present in the effluent, compared to zinc and lead, it is merely speculation as to the role of DOC in the current study, given DOC was not measured in any of the experiments. Furthermore, DOC can influence the pH of an aqueous solution directly because of its acid properties, but also indirectly by affecting the buffer systems regulating pH in an aqueous environment (Erlandsson et al., 2010). As a result, the presence of DOC may influence the precipitation of heavy metals through changes in pH, and because DOC can be preferentially adsorbed to the surfaces of solids effectively competing with the heavy metal ions for binding sites, thus reducing metal binding (Antoniadis and Alloway, 2002).

10

C. Bremner, T.A. Cochrane, P. McGuigan et al. / Environmental Technology & Innovation 18 (2020) 100662

4.2.3. Mussel shell characterisation The composition of mussel shells used in this study was comparable to compositions that are reported in the literature for mussel shells in other places (e.g. Abdulkarim et al., 2013). Results from the hydraulic conductivity and porosity testing, 22 m/h and 56% respectively, indicated that the crushed mussel shell used in this study could be considered a well-draining substrate. Weber et al. (2015) reported a hydraulic conductivity of 3.6 m/h for mussel shell, as well as a porosity of 72%, which is much higher than that recorded during this study and of those porosities reported for common stormwater treatment system substrates. However, it is difficult to make comparisons between the conductivity and porosity results of the mussel shell as reported in this study and those reported in other studies because the size fraction used in other studies was often different to the fraction used in this study. For example, often the mussel shell is only roughly crushed before use which would support the much higher porosity reported by Weber et al. (2015). Whereas, in the current study a specific size fraction was tested resulting in different hydraulic conductivity and porosity measurements when compared to the literature. 4.2.4. Heavy metal removal efficiencies by mussel shell Previous studies have demonstrated the usefulness of mussel shell as a pH amendment for the treatment of acid mine drainage (AMD) and stormwater (e.g. Daubert and Brennan, 2007; McCauley et al., 2009; Good et al., 2014). A key study by Good et al. (2014) incorporated mussel shell into the substrate of bioretention systems, which significantly increased the heavy metal removal efficiency of the systems. For example, the removal efficiency of zinc increased from 55% to 80% in the presence of mussel shell, and copper saw a similar increase from 27% to 47% (Good et al., 2014). However, lead saw no significant difference in removal efficiency in the presence of mussel shell; and given that its partition coefficient is higher than that of other metals, it is consistently removed by stormwater treatment systems with a lesser need for pH amendments compared to other metals. Additionally, Good et al. (2014) found that the systems with mussel shell amendments had lower concentrations of dissolved heavy metals, which they attributed to the increased pH caused by the mussel shell. This is because dissolved heavy metal ions shift to particulate-bound forms at higher pH values (Dempsey et al., 1993). The current study recorded much greater removal efficiencies for zinc and copper and equally high lead removal that may be attributed to the lack of any other substrate in the experimental systems, which would add further variables to be considered (e.g. Li and Davis, 2008). Studies by McCauley et al. (2009) and Daubert and Brennan (2007) have both shown the benefit of using mussel shell for the treatment of heavy metals in AMD. For example, in the study by Daubert and Brennan (2007), they recorded a decrease in dissolved iron, aluminium, and manganese concentrations as pH increased in the presence of chitin derived from crab shell. As a result, the concentration of dissolved iron and aluminium dropped to less than 0.03% of the starting concentration (Daubert and Brennan, 2007). Using mussel shell, McCauley et al. (2009) found that the systems with the greatest mussel shell content showed the highest heavy metal (iron and aluminium) removal. In all of the studies mentioned above, heavy metal removal was attributed to both adsorption and precipitation (and thus filtration), and it is likely that either or both processes occurred in the current study. The composition of the mussel shell facilitates both adsorption and precipitation processes for dissolved heavy metal removal. Mussel shell is predominantly composed of minerals in the form of calcium carbonate (CaCO3 ; 52%) and chitin (38%) (Abdulkarim et al., 2013). EDS results showed that the mussel shell used in this study had a similar composition to the other studies with high concentrations of oxygen and calcium measured. Chitin is a natural biopolymer which has excellent adsorption potential because it contains high concentrations of amino and hydroxyl functional groups (Daubert and Brennan, 2007; Bhatnagar and Sillanpää, 2009). These functional groups are highly reactive, enhancing the ability of chitin to adsorb heavy metals, furthermore, adsorption is further enhanced up to a pH of 7 (Bhatnagar and Sillanpää, 2009). Chitin has an order of affinity for heavy metal ions as follows: Cu > Pb > Zn (Bhatnagar and Sillanpää, 2009). The second component of mussel shell is CaCO3 . The dissolution of CaCO3 results in increased alkalinity, subsequently resulting in increased pH (Daubert and Brennan, 2007; McCauley et al., 2009; Good et al., 2014). Alkalinity, and to a lesser extent pH, were considerably higher in the mussel shell systems than in the recycled glass systems, with an average alkalinity of 195.5 ± 55.8 mg/L as CaCO3 . Similar alkalinities have been recorded in the literature (e.g. Daubert and Brennan, 2007). When pH reaches a value above 7, precipitation of the heavy metal ions occurs, after which the substrate simply filters out the newly particulate metals (Dempsey et al., 1993). Therefore, we can speculate that at a pH value below 7 adsorption will be the predominant dissolved heavy metal removal mechanism, while at pH values above 7 precipitation may dominate as the key removal mechanism. 5. Conclusions The freshwater biofilm grown in this study sorbed high concentrations of dissolved heavy metals from the synthetic stormwater. Biofilms, however, did not have as big an impact on dissolved heavy metal removal as the substrates on their own after an experimental period of 48 h, despite the biofilm demonstrating an ability to sorb dissolved heavy metals to high concentrations and to increase pH. In fact, results suggest that the presence of biofilm may inhibit the removal of dissolved Cu in both the recycled glass and mussel shell substrates after a certain timeframe. From this it can be concluded that biofilm growth on RG and MS added no heavy metal removal benefit in the longer term, but may have some benefit in a shorter timeframe for selected heavy metals (Zn). From a cost-benefit point of view, encouraging the growth of biofilm

C. Bremner, T.A. Cochrane, P. McGuigan et al. / Environmental Technology & Innovation 18 (2020) 100662

11

seems less attractive than using only recycled glass or mussel shell as filter media in a conventional STS and the former maybe be discouraged, particularly if the target is to remove copper and lead from stormwater. Mussel shell consistently showed the best dissolved heavy metals removal efficiencies, typically achieving over 90% removal for zinc, copper, and lead. Mussel shells are an abundant, alkaline, waste product from the shellfish industry and could readily be used for stormwater treatment. Recycled glass achieved very good removal efficiencies, but not to the same extent as the mussel shell substrates. Recycled glass is a safe versatile material, which had previously been thought to be inert, and thus disposed of readily into landfills. However, this research has shown that recycled glass has benefits for stormwater treatment, adding value to the life cycle of the recycled glass. RG and MS substrates have shown stormwater treatment potential by effectively reducing the concentration of dissolved heavy metals from a synthetic stormwater. However, it is important to note that the long-term efficiency of these substrates is unknown. Further research is needed to ascertain each substrates contaminant removal lifecycle which may also help encourage an increased interest in the recovery of waste products. By incorporating recycled glass and mussel shell into STS we can contribute to waste minimisation and increased value associated with the extended lifecycle of these products. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. CRediT authorship contribution statement Courtenay Bremner: Investigation, Validation, Data curation, Writing - original draft. Thomas A. Cochrane: Conceptualization, Methodology, Writing - review & editing, Supervision. Peter McGuigan: Methodology, Supervision. Ricardo Bello-Mendoza: Conceptualization, Methodology, Writing - review & editing, Supervision. Acknowledgment The authors thank Jan McKenzie and Paul Brody for their assisstance in microalgae identification, Robert Stainthorpe for ICP-MS analysis and Mike Flaws for scanning electron microscope and energy dispersive X-ray spectroscopy analyses. Appendix A. Supplementary data Supplementary material related to this article can be found online at https://doi.org/10.1016/j.eti.2020.100662. References Abdulkarim, A., Isa, M.T., Abdulsalam, S., Muhammad, A.J., Ameh, A.O., 2013. Extraction and characterisation of chitin and chitosan from mussel shell. Extraction 3 (2), 108–114. Ahiablame, L.M., Shakya, R., 2016. Modeling flood reduction effects of low impact development at a watershed scale. J. Environ. Manage 171, 81–91. Ancion, P.-Y., 2010. Use of Stream Biofilm Bacterial Communities and Associated Metals As Indicators of Urban Runoff Impact on Freshwater Ecosystems (Ph.D. thesis). The University of Auckland. Ancion, P.-Y., Lear, G., Lewis, G.D., 2010. Three common metal contaminants of urban runoff (Zn, Cu & Pb) accumulate in freshwater biofilm and modify embedded bacterial communities. Environ. Pollut. 158 (8), 2738–2745. Anderson, C., Pedersen, K., Jakobsson, A.-M., 2006. Autoradiographic comparisons of radionuclide adsorption between subsurface anaerobic biofilms and granitic host rocks. Geomicrobiol. J. 23 (1), 15–29. Antoniadis, V., Alloway, B., 2002. The role of dissolved organic carbon in the mobility of Cd, Ni and Zn in sewage sludge-amended soils. Environ. Pollut. 117 (3), 515–521. American Public Health Association (APHA), 2005. Standard Method for Examination of Water and Wastewater, twentyfirst ed. APHA, AWWA, WPCF, Washington D.C. Barbosa, A.E., Fernandes, J.N., David, L.M., 2012. Key issues for sustainable urban stormwater management. Water Res. 46 (20), 6787–6798. Beck, A.J., Janssen, F., de Beer, D., 2011. The influence of phototrophic benthic biofilms on Cd, Cu, Ni, and Pb transport in permeable sediments. Biogeochemistry 102 (1–3), 167–181. Bhatnagar, A., Sillanpää, M., 2009. Applications of chitin-and chitosan-derivatives for the detoxification of water and wastewater - a short review. Adv. Colloid Interfaces 152 (1), 26–38. Blecken, G.T., Marsalek, J., Viklander, M., 2011. Laboratory study of stormwater biofiltration in low temperatures: Total and dissolved metal removals and fates. Water Air Soil Pollut. 219 (1–4), 303–317. Blecken, G.T., Zinger, Y., Deletić, A., Fletcher, T.D., Hedström, A., Viklander, M., 2010. Laboratory study on stormwater biofiltration: Nutrient and sediment removal in cold temperatures. J. Hydrol. 394 (3–4), 507–514. Christchurch City Council (CCC, 2016. Rain Garden Design, Construction and Maintenance Manual. Christchurch City Council, Christchurch, New Zealand. Charters, F.J., Cochrane, T.A., O’Sullivan, A.D., 2016. Untreated runoff quality from roof and road surfaces in a low intensity rainfall climate. Sci. Total Environ. 550, 265–272. Costley, S.C., Wallis, F.M., 2001. Bioremediation of heavy metals in a synthetic wastewater using a rotating biological contactor. Water Res. 35 (15), 3715–3723. Datry, T., Malard, F., Vitry, L., Hervant, F., Gibert, J., 2003. Solute dynamics in the bed sediments of a stormwater infiltration basin. J. Hydrol. 273 (1), 217–233.

12

C. Bremner, T.A. Cochrane, P. McGuigan et al. / Environmental Technology & Innovation 18 (2020) 100662

Daubert, L.N., Brennan, R.A., 2007. Passive remediation of acid mine drainage using crab shell chitin. Environ. Eng. Sci. 24 (10), 1475–1480. Dempsey, B.A., Tai, Y.-L., Harrison, S., 1993. Mobilization and removal of contaminants associated with urban dust and dirt. Water Sci. Technol. 28 (3–5), 225–230. Disfani, M., Arulrajah, A., Bo, M., Sivakugan, N., 2012. Environmental risks of using recycled crushed glass in road applications. J. Clean. Prod. 20 (1), 170–179. Donlan, R.M., 2002. Biofilms: Microbial life on surfaces. Emerg. Infect. Dis. 8 (9), 881–890. Erdem, E., Karapinar, N., Donat, R., 2004. The removal of heavy metal cations by natural zeolites. J. Colloid Interfaces Sci. 280 (2), 309–314. Erlandsson, M., Cory, N., Köhler, S., Bishop, K., 2010. Direct and indirect effects of increasing dissolved organic carbon levels on ph in lakes recovering from acidification. J. Geophys. Res.-Biogeol. 115 (G03004). Farag, A.M., Nimick, D.A., Kimball, B.A., Church, S.E., Harper, D.D., Brumbaugh, W.G., 2007. Concentrations of metals in water, sediment, biofilm, benthic macroinvertebrates, and fish in the Boulder River watershed, Montana, and the role of colloids in metal uptake. Arch. Environ. Con. Toxiol. 52 (3), 397–409. Feder, M.J., 2014. Towards a Rational Design for Sustainable Urban Drainage Systems: Understanding (Bio) Geochemical Mechanisms for Enhanced Heavy Metal Immobilization in Filters. (Ph.D. thesis). University of Glasgow. Feder, M., Phoenix, V., Haig, S., Sloan, W., Dorea, C., Haynes, H., 2015. Influence of biofilms on heavy metal immobilization in sustainable urban drainage systems (SuDS). Environ. Technol. 36 (21), 2803–2814. Fetter, C.W., 2000. Applied Hydrogeology. Prentice hall, New Jersey. Good, J., O’Sullivan, A., Wicke, D., Cochrane, T., 2014. Ph buffering in stormwater infiltration systems—sustainable contaminant removal with waste mussel shells. Water Air Soil Pollut. 225 (3), 1885. Hu, Z., Chandran, K., Grasso, D., Smets, B.F., 2003. Impact of metal sorption and internalization on nitrification inhibition. Environ. Sci. Technol. 37 (4), 728–734. Hu, Z., Hidalgo, G., Houston, P.L., Hay, A.G., Shuler, M.L., Abruna, H.D., Lion, L.W., 2005. Determination of spatial distributions of zinc and active biomass in microbial biofilms by two-photon laser scanning microscopy. Appl. Environ. Microb. 71 (7), 4014–4021. van Hullebusch, E.D., Zandvoort, M.H., Lens, P.N., 2003. Metal immobilisation by biofilms: mechanisms and analytical tools. Rev. Environ. Sci. Biol. 2 (1), 9–33. Ibrahim, H.S., Ammar, N.S., Abdel Ghafar, H.H., Farahat, M., 2012. Adsorption of Cd (II), Cu (II) and Pb (II) using recycled waste glass: equilibrium and kinetic studies. Desalin. Water Treat. 48 (1–3), 320–328. Jang, A., Kim, S., Kim, S., Lee, S., Kim, I.S., 2001. Effect of heavy metals (Cu, Pb, and Ni) on the compositions of EPS in biofilms. Water Sci. Technol. 43 (6), 41–48. Klute, A., 1965. Laboratory measurement of hydraulic conductivity of unsaturated soil. In: Methods of Soil Analysis, Mono. 9, Part 1. Amer. Soc. of Agronomy, Madison, WI, pp. 253–261. Kulczycki, E., Fowle, D., Fortin, D., Ferris, F., 2005. Sorption of cadmium and lead by bacteria–ferrihydrite composites. Geomicrobiol. J. 22 (6), 299–310. Kurniawan, T.A., Chan, G.Y., Lo, W.-H., Babel, S., 2006. Physico–chemical treatment techniques for wastewater laden with heavy metals. Chem. Eng. J. 118 (1), 83–98. Lewis, A., 2017. Precipitation of heavy metals. In: Rene, E.R., et al. (Eds.), Sustainable Heavy Metal Remediation. Springer, Cham, Switzerland, pp. 101–120. Li, H., Davis, A.P., 2008. Heavy metal capture and accumulation in bioretention media. Environ. Sci. Technol.. 42 (14), 5247–5253. McCauley, C.A., O’Sullivan, A.D., Milke, M.W., Weber, P.A., Trumm, D.A., 2009. Sulfate and metal removal in bioreactors treating acid mine drainage dominated with iron and aluminum. Water Res. 43 (4), 961–970. Mosley, L.M., Peake, B.M., 2001. Partitioning of metals (Fe, Pb, Cu, Zn) in urban run-off from the Kaikorai Valley, Dunedin, New Zealand. New Zeal. J. Mar. Freshw. 35 (3), 615–624. O’Sullivan, A., Wicke, D., Cochrane, T., 2012. Heavy metal contamination in an urban stream fed by contaminated air-conditioning and stormwater discharges. Environ. Sci. Pollut. Res. 19 (3), 903–911. Pennington, S.L., Webster-Brown, J.G., 2008. Stormwater runoff quality from copper roofing, Auckland, New Zealand. New Zeal. J. Mar. Freshw. 42 (1), 99–108. Petrella, A., Petrella, M., Boghetich, G., Petruzzelli, D., Calabrese, D., Pace, L., Guastamacchia, M., 2009. Glass from recycling operations as thermal insulating material and sorbent in wastewater treatments. Spec. Top. Mater. Sci. Technol.: Ital. Panor. 303–312. Seelsaen, N., McLaughlan, R., Moore, S., Ball, J.E., Stuetz, R.M., 2006. Pollutant removal efficiency of alternative filtration media in stormwater treatment. Water Sci. Technol. 54, 299–305. Sekar, R., Nair, K., Rao, V., Venugopalan, V., 2002. Nutrient dynamics and successional changes in a lentic freshwater biofilm. Freshw. Biol. 47 (10), 1893–1907. Singh, R., Paul, D., Jain, R.K., 2006. Biofilms: implications in bioremediation. Trends Microbiol. 14 (9), 389–397. Soyer, E., Akgiray, Ö., Eldem, N.Ö., Saatçı, A.M., 2013. On the use of crushed recycled glass instead of silica sand in dual-media filters. CLEAN–Soil Air Water 41 (4), 325–332. Weber, P., Weisener, C., DiLoreto, Z., Pizey, M., 2015. Passive Treatment of ARD Using Mussel Shells-Part I: System Development and Geochemical Processes. In: Paper presented at the Proceedings of the 10th International Conference on Acid Rock Drainage and IMWA annual Conference, Santiago. Xu, M., McKay, G., 2017. Removal of heavy metals, lead, Cadmium, and zinc, using adsorption processes by cost-effective adsorbents. In: Bonilla-Petriciolet, A., et al. (Eds.), Adsorption Processes for Water Treatment and Purification. Springer, Cham, Switzerland, pp. 109–138.