Removal of dissolved organic matter in road runoff by granules prepared using sludge from waterworks Open Access

Abundant pollutants, such as suspended solids, nutrients, organics, and heavy metals, are detected in road runoff and have been deemed as predominant non-point pollution sources for the urban aquatic environment (Gilbert & Clausen 2006). Among these pollutants, dissolved organic matter (DOM) was the most detected. DOM has been of particular concern due to its widespread occurrence in gasoline leaks, vehicle exhaust, and tire wear (Howe & Clark 2002; Opher & Friedler 2010; Zhao et al. 2015). Previous studies have reported that DOM in road runoff contains a large number of aromatic and aliphatic moieties with alcoholic and phenolic hydroxyls, carboxyl, and methoxyl functional groups, and so on (Sun et al. 2005; Tan & Kilduff 2007; Zhao et al. 2018). Such properties of DOM present in road runoff result in its strong affinity for many contaminants (e.g. heavy metals, nutrients, colloids, endocrine disruptors, etc.), and then influence their transport and fate (Guan et al. 2006; Yang et al. 2013; Philippe & Schaumann 2014; Ding et al. 2019). Thus, it has been desirable to develop technologies removing DOM to simultaneously control these contaminants.

Researchers have tested various methods to remove DOM in drinking water. Conventional coagulants such as poly aluminum chloride (PAC) and polymerization ferric chloride (PFC) have been found capable of adsorbing DOM via ligand exchange between the hydrous oxide and metal ions in coagulants and negative surface charged DOM, thus forming hydrolysis products or hydroxide precipitate (Matilainen et al. 2010; Hussain et al. 2013; Zhou et al. 2017). Precipitate sludge is a cost-effective adsorbent owing to its high annual production. It has been reported that the annual production of sludge with a water content of 70% is 2 × 10⁷ tons in China (Guo et al. 2010), causing considerable concerns over its disposal and the associated cost. In order to reuse this sludge more effectively, numerous new sludge-based materials have also been found and tested in recent years. For example, a granular sludge-clay (GSC) adsorbent using the readily available sludge from waterworks was synthesized to investigate the removal of Cu(II), Zn(II), and Cd(II). The GSC demonstrated good effectiveness for the removal of heavy metals (Du et al. 2020). Therefore, the extensive use of sludge has become an important issue from the point of view of engineering applications.

In addition to the removal of heavy metals, sludge has been extensively investigated as an adsorbent for some specific pollutants, like nitrogen and phosphorus, for example (Yang et al. 2006, 2015; Zhou & Haynes 2011; Luo et al. 2013; Siswoyo et al. 2014). Due to the good DOM removal characteristics of coagulants, sludge containing coagulants is feasible to apply as a filler in the construction of some green infrastructures (e.g. wetlands, bioretentions, rain gardens, etc.) for removing DOM in road runoff. As such, the target of this study was to test the adsorption capacity of GSC on DOM in road runoff. To further elucidate the underlying adsorption mechanisms, the DOM fractions that lead to the highest removal rate of dissolved organic carbon (DOC) were investigated.

Brown wide-mouth bottles were used to collect road runoff from the ditch of Chegongzhuang Street in Beijing (39°55′N, 116°20′E) on March 20th, 2019. The site is an arterial street with approximately 10,000 cars per day in the downtown district. The collected samples were transported to the laboratory and stored as DOM stock solutions in the dark at 277 K after filtration with 0.45 μm membranes. Before each subsequent batch experiment, the DOC concentration of the stock solutions was measured by a total organic carbon analyzer (multi N/C3100, Jena, Germany), and then the stock solution was diluted to the desired DOC concentration.

The sludge and clay were obtained from the Third Water Purification Plant in Beijing, China, and from river bank soil in Shanghai, China. The sludge and clay were thoroughly mixed with a mass quantitative relation of 1:2, and then an appropriate volume of water was added in a mechanical mixer to mix well. Afterward, pellets with particle sizes ranging from 2 to 3 mm were prepared for the experiments, dried in an oven set at 380 K for 2 h, and then baked in a muffle furnace at 873 K for 1 h. The obtained pellets were then collected and denoted as GSC.

The DOM stock solution of road runoff was diluted to a solution with a DOC concentration of 15 mg/L using ultrapure water at pH 7.0. Based on pre-experiments, 0.5 g GSC was added to 40 mL of the solution, and then the mixture was shaken on an orbital shaker at 120 rpm at 298 K. Finally, samples of the mixture were gathered at defined times (t = 5, 10, 20, 30, 60, 120, 240, and 360 min) to determine DOC concentrations.

DOM stock solutions of DOC concentrations within 0–100 mg/L at pH 7.0 were prepared and 0.5 g GSC was added to each of these solutions. Then the mixtures were shaken on an orbital shaker at 120 rpm at 298 and 318 K. After being stirred for 12 h, the supernatants were filtered by 0.45 μm membranes, and the residual DOC concentration was determined.

To investigate the effect of pH, 0.5 g GSC was added to various 40 mL DOM solutions (pH 3.0–11.0) with a DOC concentration of 15 mg/L. The mixtures were shaken on an orbital shaker at 120 rpm at 298 K for 12 h, and then DOC concentrations were determined.

The fluorescence excitation emission spectra (EEM) of DOM solutions before and after adsorption by GSC were measured with a Hitachi F-7000 fluorescence spectrophotometer (F-7000, Hitachi, Japan). EEM spectra were collected with both emission (E_m) and excitation (E_x) wavelengths in the range of 200–550 with a step of 5 nm. The E_m and E_x slits were kept at 5 nm with a scan speed of 1,200 nm/min. An emission cutoff filter of 290 nm was used for the scans to eliminate Rayleigh scattering, and fluorescence spectrometry data from deionized water was eliminated by a blank control experiment. Origin 8.5 software (Origin Lab Co., MA, USA) was used to generate the contour map of E_x/E_m.

DOM solutions before and after GSC adsorption were isolated according to the method reported by Aiken (Aiken et al. 1992; Imai et al. 2002). According to the different adsorption characteristics of DOM on XAD-8 resin, MSC resin, and Duolite A-7 resin, the DOM in road runoff was classified into six fractions: hydrophilic acids (HiA), hydrophilic basic (HiB), hydrophilic neutral (HiN), hydrophobic acids (HoA), hydrophobic basic (HoB), and hydrophobic neutral (HoN). Then the DOC concentrations in six DOM fractions were measured. The organic carbon mass balance of resin isolation was assessed at 100 ± 10% in this study.

A sample 30 mL of each DOM fraction was lyophilized to powder, and then 1% of the powder was mixed with 99% dried spectrometry grade KBr. The powder mixture was ground and pressed to a film for a FTIR spectra scan with a Nicolet 6700 Thermo Fisher Scientific FTIR spectrometer covering a wavenumber range of 4,000–400 cm⁻¹.

It is inferred from the coefficient of determination (R²) values (Figure 1) that the pseudo-second-order model is more appropriate for simulating the adsorption kinetics of GSC (R² = 0.9934), indicating that chemisorption is the mode of DOM adsorption by GSC. This relates to valence forces generated by sharing or exchanging electrons between the DOM and GSC surfaces (Bouzid et al. 2008).

The data for the parameters obtained from the Langmuir, Freundlich and Temkin isotherm models are shown in Table 1. Apparently, the Langmuir model is more appropriate to interpret the adsorption isotherm of DOM for GSC at 298 and 318 K with R² values of 0.9905 and 0.9861, respectively. The observation suggests that road runoff DOM adsorption onto GSC was monolayer adsorption. Moreover, GSC provided the maximum adsorption capacities (q_m) of 4.466 and 4.325 mg/g at 298 and 318 K, respectively, which are significantly higher than those for rice husk ash, goethite, and magnetite (Imyim & Prapalimrungsi 2010; Safiur Rahman et al. 2013).

Generally, as a result of the dissociation of acidic functional groups of DOM (e.g. carboxylic and phenolic hydroxyl) in aqueous solutions, DOM mainly existed in electronegative forms. On the other hand, the numerous hydroxyl groups on the surface of GSC (e.g. Al-OH and Fe-OH) (Illés & Tombácz 2003; Du et al. 2020) would be protonated in an acidic solution or deprotonated in an alkaline solution, resulting in a positive surface charge or negative surface charge of GSC, respectively. The acid condition (pH < 4) is more favorable for DOM adsorption on GSC due to the electrostatic interaction between DOM and GSC. Because the electrostatic interaction could promote the ligand exchange reaction of DOM with the sorbents (Safiur Rahman et al. 2013), it is thus reasonable to suppose that the ligand exchange was another primary sorption mechanism (e.g. M-OH₂⁺ + R-COO⁻——M-OOC-R + H₂O, where M and R represent the functional group of GSC and DOM, respectively). Additionally, at pH = 7.0 (the pH_pzc of such GSC has been reported around 7.0 (Tony 2020)), the ligand exchange reactions may occur between the aromatic group and hydrophobic components of DOM (e.g. M-OH + R-COO⁻——M-OOC-R +OH⁻), which results in higher removal efficiency of DOM by GSC (Wu et al. 2008).

Because hydrophobic organics could be absorbed through ligand exchange reactions by the adsorbent (Wu et al. 2008), the higher removal efficiency of hydrophobic fractions further evidenced the ligand exchange reaction between the DOM and the GSC. Generally, the dominant hydrophobic organics in road runoff are related to the aromatic and aliphatic structures with higher molecular weights, while hydrophilic organics are associated with carbon hydrates with lower molecular weights (Wang et al. 2009; Zhang et al. 2020b). So the DOM fractions with high molecular weights and high levels of unsaturated structure are preferentially adsorbed, which agrees with the report by Zhang et al. 2020a). Specifically, the fulvic-like substances contained in HoA and HoN fractions (see peak A in Figure S1 in the Supplementary Material) related to the unsaturated and oxygen-containing functional aromatic structures were preferentially removed by GSC, leading to higher removal efficiency of HoA and HoN.

To control DOM in road runoff and make extensive use of GSC from waterworks, GSC was prepared to remove DOM. The resulting adsorption followed a pseudo-second-order kinetics model and the Langmuir isotherm model. The adsorption of road runoff DOM was observed to be highly dependent on pH, with the highest removal percentage at pH 4.0 and 7.0. EEM fluorescence spectroscopy analysis indicated that fulvic-like substances contributed substantially to road runoff DOM removal. HoA and HoN fractions which contained abundant fulvic-like substances were more preferentially removed by GSC. The adsorption of DOM by GSC was achieved through electrostatic attraction and ligand exchange reactions, i.e. between the carboxyl group of DOM and the hydroxyl group of GSC.

This work was supported by the National Natural Science Foundation of China (grant no. 51878024), the Beijing Outstanding Talent Project for Youth Talent Support Program, and the Pyramid Talent Training Project of Beijing University of Civil Engineering and Architecture (grant no. JDJQ20200302).

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict of interest.