Current technologies for water purification are limited by their contaminant-specific removal capability, requiring multiple processes to meet water quality objectives. Here we show an innovative biomimetic micellar nanocoagulant that imitates the structure of Actinia, a marine predator that uses its tentacles to ensnare food, for the removal of an array of water contaminants with a single treatment step. The Actinia-like micellar nanocoagulant has a core–shell structure and readily disperses in water while maintaining a high stability against aggregation. To achieve effective coagulation, the nanocoagulant everts its configuration, similar to Actinia. The shell hydrolyses into ‘flocs’ and destabilizes and enmeshes colloidal particles while the core is exposed to water, like the extended tentacles of Actinia, and adsorbs the dissolved contaminants. The technology, with its ability to remove a broad spectrum of contaminants and produce high-quality water, has the potential to be a cost-effective replacement for current water treatment processes.


Water scarcity is one of the most pressing global challenges of the twenty-first century due to its effects on public health, economies and the environment1,2,3. Projections show that by 2050 nearly two-thirds of people in the world will have limited access to adequate and safe water4. Water treatment and wastewater reuse are important approaches to meeting the demand for potable water5,6,7. However, with the growth of modern industry and agriculture, numerous contaminants are polluting water resources at non-trivial concentrations8,9,10, which has prompted the establishment of more rigorous water quality standards to protect human health and the environment3,11. As a result, conventional physicochemical and biological treatment methods, such as coagulation, sand filtration and activated sludge processes, cannot meet the demand for both adequate water supply and quality12,13,14.

Advanced water treatment techniques, such as advanced oxidation, adsorption and membrane filtration, can effectively remove emerging contaminants (for example, pesticides and pharmaceuticals) that are not removed by conventional water treatment processes15,16,17,18,19,20. Yet, such treatment techniques cannot independently achieve a comprehensive removal of contaminants, as the techniques often require conventional pretreatment practices to remove colloidal and suspended particles12. In practice, multiple techniques are combined for advanced water treatment, which considerably increases the process length, energy consumption, capital cost and land use21. There is an urgent need for new approaches to attain complete removal of a broad spectrum of contaminants in a single treatment phase.

Coagulation is one of the most common stages in water and wastewater treatment plants and is regarded as a simple, cost-effective and upgradable technology22,23. Coagulants destabilize and aggregate colloidal water contaminants into large aggregates (flocs) through charge neutralization, adsorption and enmeshment, and then settle and separate the colloidal contaminants from water24,25. However, conventional coagulants are ineffective in removing dissolved contaminants, which include emerging contaminants of increasing public concern26,27,28. Therefore, coagulation is mainly applied in advanced water treatment processes as a pretreatment step to remove suspended and colloidal particles as well as natural organic matter29,30,31.

A coagulant with the capability of removing multiple water contaminants simultaneously would allow for single-stage water treatment. However, introducing new components or functional groups in a novel coagulant could induce precipitation during its preparation and long-term storage32,33, and thus negate the coagulant’s ability to stably disperse in water and remove contaminants. This study aims to develop a multifunctional coagulant capable of removing a broad spectrum of contaminants while maintaining a high stability.

Actinia is a marine predator that uses its tentacles to ensnare food. The species has a cylindrical or spherical body with tentacles that are retracted while resting but are extended while preying. Inspired by this eversion behaviour of Actinia, we synthesized a biomimetic micellar nanocoagulant with a core–shell structure. The spherical Actinia-like micellar nanocoagulant (AMC) has a high stability against aggregation and a configuration that can be everted when added to water. The shell of the nanocoagulant hydrolyses into flocs and destabilizes and enmeshes colloidal matter and suspended particles, while the core captures small dissolved contaminants. The unique structure of the AMC provides it with an ability to effectively remove a range of contaminants, and thus enables its use for broader applications than conventional coagulants.

Nanocoagulant characteristics and performance

We employed an acid–base self-assembly method to synthesize a biomimetic eversible core–shell micellar nanocoagulant that simulates the configuration of Actinia (Fig. 1a). First, 3-(trimethoxysilyl)propyl-n-octadecyldimethylammonium chloride (TPODAC), which contains positively charged quaternary ammonium and a hydrophobic long carbon chain, was hydrolysed to silanetriol at pH 2.2. After dialysis, the silanetriol was allowed to react with AlCl3 at pH 3.5. As the condensation occurred, aliphatic quaternary ammonium–Si–Al complexes were obtained by regulating the ratio of Si to Al. By controlling the Al concentration, these conjugates self-assembled into core–shell micelles with an aliphatic quaternary ammonium core and a cationic Si–Al complexes shell. The product had a total aluminium concentration (AlT) of 0.1 M at pH 4.0.

Fig. 1: Synthesis and characterization of AMC.
Fig. 1

a, Schematic diagram of the synthesis of the AMC via an acid–base controlled self-assembly method. b, Molecular dynamic simulation of the AMC structure. Al, pink; O, red; Si, yellow; C, turquoise; N, blue; H, white. c, TEM image of the AMC. Scale bar, 200 nm. d, Size distribution of the AMC characterized by DLS at AlT = 0.1 M and pH 4.0.

The condensation of silanetriol and hydrolysed AlCl3 was identified by the vibration of Si–O–Al (719 cm−1) in Fourier transform infrared spectra (Supplementary Fig. 1)34. Molecular dynamics simulations suggest that the aliphatic carbon chains gather inside the AMC by hydrophobic attraction and positively charged hydrophilic quaternary ammonium–Si–Al complexes disperse outside by electrostatic repulsion, which resembles an Actinia-like core–shell structure with retracted tentacles (Fig. 1b). When the hydrophobic dye Sudan III was mixed with the AMC, the dye uniformly dispersed in the solution, but when mixed with Al2(SO4)3, the dye concentrated on the top of the solution. This behaviour indicates that the AMC forms a micellar structure with a hydrophobic core that binds the hydrophobic dye molecules (Supplementary Fig. 2).

Transmission electron microscope (TEM) images show the AMC to be monodisperse, spherical nanoparticles (Fig. 1c), and dynamic light scattering (DLS) measurements indicate an average hydrodynamic diameter (Dh) of 50 nm (Fig. 1d). The zeta potential of the AMC (~35 mV) was significantly higher than those of conventional coagulants (Supplementary Fig. 3), which indicates its strong capacity for charge neutralization. The AMC is very stable, as the morphology and size show no change after one year of storage (Supplementary Fig. 4). The necessary conditions for the stable storage of the AMC are pH ≤ 4.0 and AlT ≤ 1.0 M; when the pH or the AlT was much higher, the stability of the AMC decreased due to hydrolysis, condensation and gelatinization of the Si and Al components (Supplementary Fig. 5).

Secondary wastewater effluent from a municipal sewage treatment plant was used as raw water (the characteristics are given in Supplementary Table 1) to test the coagulation performance of the AMC. Three conventional coagulants, Al2(SO4)3, FeCl3 and poly(diallyldimethylammonium chloride) (polyDADMAC) partially removed the turbidity (87–98%), dissolved organic carbon (DOC) (30–54%) and total phosphorus (8–68%) (Fig. 2a), consistent with previous results27,35,36. For the AMC, turbidity was effectively removed (>90%), but the removal efficiencies for DOC, total phosphorus and NO3 also exceeded 90%, which is a significant improvement compared to the industry-applied coagulants (Fig. 2a). This study demonstrates the removal of NO3 by the AMC, whereas traditional coagulants exhibited a negligible NO3 removal37,38.

Fig. 2: AMC coagulation performance.
Fig. 2

a,b, Removal efficiency of turbidity, DOC, NO3 and total phosphorus (TP) (a) and micropollutants (b) in secondary wastewater effluent at Al2(SO4)3, FeCl3 and AMC dosages of 120 mg l–1 and a polyDADMAC dosage of 2 mg l–1. The chemical structures and properties of the tested micropollutants are given in Supplementary Table 2. For DCF, IBP, SA, BEZ, Pyr, BaP and PFOA, the per cent removal that corresponded to half the detection limit was assigned because effluent concentrations were below the detection limit. Error bars in a and b represent the s.d. from triplicate experiments. c, Relative abundances of heteroatom class species in the dissolved organic matter of raw water (RW) and coagulated water. The x axis represents heteroatom class species and the y axis represents relative abundance. IBP, ibuprofen; SA, salicylic acid; SMXL, sulfamethoxazole; BEZ, bezafibrate; Flu, fluorene; Ant, anthracene; Flr, fluoranthene; Pyr, pyrene; BbF, benzo[b]fluoranthene; BaP, benzo[a]pyrene; PFOA, perfluorocaprylic acid.

We next investigated the ability of the AMC to remove emerging contaminants that are typically persistent in the environment, such as organic micropollutants and pharmaceuticals (Supplementary Table 2). Concentrations of these micropollutants in the wastewater sample were present in the parts per trillion to parts per billion range (Supplementary Table 1). The observed removal efficiencies for micropollutants by conventional coagulants were between 0 and 60%, which is consistent with previous reports39,40. However, the AMC achieved removal efficiencies of over 90% for all tested micropollutants (Fig. 2b), despite the variability in contaminant charge and hydrophobicity.

Ultrahigh-resolution Fourier transform ion cyclotron resonance mass spectrometry (FT-ICR MS) was used to characterize the dissolved organic matter removal by the AMC at a molecular level. We found that the dissolved organic components in the raw water mainly contained CHO, CHON, CHOS and CHONS species. The results were then classified by organic species according to the number of heteroatoms (O, N and S) in the molecules (Fig. 2c). We observed that the Ox and S1Ox class species were the most abundant, followed by N1Ox and N2Ox. After coagulation with conventional coagulants, minor removal was observed for some contaminant species. In contrast, the AMC significantly removed many species, particularly the most dominant ones (O6–O8, NxOx, S1O3 and S1O5–S1O7). The effluent concentrations of multiple contaminants were extremely low after coagulation using the AMC (Fig. 2b and Supplementary Fig. 6).

Coagulation behaviour

We compared the coagulation behaviour of the AMC and conventional coagulants in removing humic acid (as a representative large molecule) and diclofenac (DCF) (as a representative small molecule). The removal efficiencies of humic acid improved and the zeta potentials transitioned from negative to positive with increasing dosages of all four positively charged coagulants used (Supplementary Fig. 7a,c). Our results suggest that charge neutralization is the main removal mechanism for humic acid, consistent with previous coagulation studies41,42. The positively charged shell of Si–Al complexes of the AMC functions similarly to conventional coagulants. As for DCF, zeta potentials also increased with increasing coagulant dosages, but the AMC removed 91.7% of DCF whereas conventional coagulants removed only 15% (Supplementary Fig. 7b,d). Our results indicate that charge neutralization alone cannot explain the removal of small dissolved molecules by the AMC.

The molecular weight distribution of DCF in the supernatant solution after coagulation was determined using ultrafiltration membranes (Fig. 3a). Before coagulation, about 87% of DCF was distributed in the 0–1 kDa range, but after coagulation by the AMC, 91.7% of DCF was transferred to the solid phase (>0.2 μm). In contrast, after the coagulation by conventional coagulants, most residual DCF was distributed in the 1–30 kDa (65.1% for Al2(SO4)3, 68.6% for FeCl3 and 34.4% for polyDADMAC) and 30 kDa to 0.2 μm ranges (8.3% for Al2(SO4)3, 3.2% for FeCl3 and 21.2% for polyDADMAC). This result indicates that small molecules interacted with conventional coagulants to form larger, but still dissolved, complexes.

Fig. 3: Small contaminant species analysis.
Fig. 3

a, DCF distribution for various molecular weights before and after coagulation with Al2(SO4)3, FeCl3 and the AMC at dosages of 120 mg l–1, and polyDADMAC at a dosage of 2 mg l–1. b, ESI FT-ICR MS of DCF water coagulated by Al2(SO4)3 at a dosage of 120 mg l–1. Insets, Magnifications of parts of the spectra. The synthetic aqueous solution was used for the coagulation test and the coagulation pH was controlled at 7.5.

FT-ICR MS was used to analyse the residual DCF to confirm the formation of dissolved complexes. All mass spectra had a DCF molecular ion peak [M + Na]+ at m/z = 318 (Fig. 3b and Supplementary Figs. 8 and 9). This observation demonstrates that DCF was not fully removed by conventional coagulants and that some residuals were dissolved in water as single molecules or ions. Peaks were also found to represent DCF and Mx (polynuclear forms, where x is the number of Al, Fe or C8H16N+) complexes. For Al2(SO4)3, the peaks at m/z = 434 and 474 were assigned to the binding of DCF and Al2, m/z = 480 and 552 to the binding of DCF and Al3, and m/z = 536 and 558 to the binding of DCF and Al4 (Fig. 3b). For FeCl3, the peak at m/z = 478 was assigned to the complex of DCF and Fe2, m/z = 603 to the complex of DCF and Fe3, and m/z = 424 to the complex of DCF and Fe6 (Supplementary Fig. 8). For polyDADMAC, the peak at m/z = 251 represented the dimer of C8H16N+ (Supplementary Fig. 9). Different from the Al or Fe coagulants, polyDADMAC combined with DCF through ionic bonds to form large complexes, which manifested as separate peaks of polyDADMAC and DCF in the mass spectra.

We investigated the evolution of Dh during the first 300 s of coagulation using DLS. DLS measurements revealed that the particle size increased to 149 nm within the first 30 s of coagulation (Fig. 4a) and then increased continuously with time. The increased intensity of the diffraction pattern in the surface plasmon resonance microscopy images also show a growth in particle size during the coagulation (Supplementary Fig. 10)43. TEM images of the samples at coagulation times of 30 s, 120 s and 210 s further confirmed the formation of large and irregularly shaped aggregates (Supplementary Fig. 11). The above results suggest that when the stable AMC from the stock solution (pH 4.0) was added into water with a pH of 7.5, which is a typical pH of natural waters and wastewater, the shell of the Si–Al complexes hydrolysed and formed large aggregates and precipitates.

Fig. 4: AMC configuration eversion behaviour.
Fig. 4

a, Evolution of Dh during the coagulation as characterized by DLS. Error bars represent s.d. from triplicate experiments. b, EDX elemental mapping of the AMC in a stock solution of AlT 0.1 M at pH 4.0 for Al and C (left), Al (middle) and C (right). Scale bar, 20 nm. c, EDX elemental mapping of aggregates at 30 s of coagulation for Al and C elements (left), Al element (middle) and C element (right). Scale bar, 50 nm. The synthetic aqueous solution was coagulated for the DLS and EDX analysis at an AMC dosage of 120 mg l–1 and the coagulation pH was controlled at 7.5.

Spherical aberration corrected TEM equipped with energy-dispersive X-ray (EDX) spectroscopy was used to investigate the elemental distribution of the AMC in stock solution (pH 4.0) and the AMC-formed aggregates at 30 s of coagulation (pH 7.5). For the stock solution, the elemental mapping spectra demonstrated that Al is distributed in the exterior of the AMC, whereas C is distributed in the interior of the AMC, to form a core–shell structure (Fig. 4b). In contrast, after adding the AMC to water for coagulation, Al was trapped inside the aggregates and C was exposed outside, which clearly indicates that the AMC configuration was everted (Fig. 4c). The line-scanning spectra showed a similar result to that of the elemental mapping spectra (Supplementary Fig. 12a,b). Therefore, when the AMC is added into water at pH > 4.0, the shell hydrolyses and induces the eversion of the core–shell structure and exposes the aliphatic core.

Mechanistic investigation of small contaminant removal

We employed fluorescence microscopy to observe in situ the time-dependent aggregation of small molecules by the AMC. A fluorescent dye, 3,3′-dihexyloxacarbocyanine iodide (DiOC6(3)), was chosen as a small contaminant (Supplementary Fig. 13). In bright-field images, flocs were observed for both coagulation by the AMC and by Al2(SO4)3 (Fig. 5a and Supplementary Fig. 14a), indicating that the cationic Si–Al complexes in the AMC shell have the same function as the Al components in Al2(SO4)3: they destabilize large contaminants. In fluorescence images, however, no fluorescent aggregates were found for Al2(SO4)3 coagulation (Supplementary Fig. 14b), indicating that the low-molecular-weight dye cannot be incorporated into Al2(SO4)3 flocs. In contrast, fluorescent dots aggregated and adsorbed to the AMC flocs (Fig. 5b) as the dye was transferred from the liquid to the solid phase.

Fig. 5: Real-time fluorescence imaging of AMC coagulation.
Fig. 5

a,b, Bright-field images (a) and corresponding fluorescent images (b) of real-time DiOC6(3) coagulation by the AMC. Scale bars, 30 μm. The synthetic aqueous solution was coagulated at an AMC dosage of 120 mg l–1 and the coagulation pH was controlled at 7.5.

To investigate the removal mechanisms of characteristically different small contaminants, the interaction of the AMC with an inorganic anion (NO3) and an organic molecule (DCF) was analysed using molecular dynamics simulations. The results revealed that the binding patterns are quite different (Fig. 6a–h): NO3 was found near the surface of the AMC N atoms at the end of the simulation, whereas DCF was close to either the N atoms or the C chains of the AMC.

Fig. 6: Molecular dynamic simulations of AMC–NO3 and AMC–DCF interactions.
Fig. 6

a,b, Snapshots of the initial state (a) and final state (b) of the interaction between the AMC and NO3. c,d, Enlargements of interaction sites at the final state of the AMC and NO3 interaction viewed from the top (c) and from the side (d) (distance in Å). e,f, Snapshots of the initial state (e) and final state (f) of the interaction between the AMC and DCF. g,h, Enlargements of the interaction sites at the final state of the AMC and DCF interaction at the end of carbon chains (g) and at the N atom surface (h). O, red; C, light blue; N, dark blue; H, white. i, Computed RDFs of Nq and On, and of Nq and Cl ions. j, Computed atom density distributions of Cb and Oc; here, ρ(z) represents the atom density and r represents the atom coordinate along the vertical (z) direction. a.u., arbitrary units.

By computing the radial distribution functions (RDF) of a water H atom as well as the H2O ligancy surrounding a NO3 O atom and a DCF carboxyl O atom, we found the hydration of NO3 and DCF changed for the AMC. For NO3, the primary peak value of the water H atom RDF in the AMC–NO3 system was less (1.7) than that in the NO3 system (2.6) (Supplementary Fig. 15a), and the H2O ligancies of NO3 also decreased (Supplementary Fig. 15b). This indicates the hydrogen bonds N–OH–O were weakened. For DCF, hydration was only slightly less in the DCF–AMC system compared to the DCF system; the water H atom RDF peak decreased from 2.75 to 2.5 and H2O ligancies from 28 to 26 (Supplementary Fig. 15c,d). This finding suggests the degree of dehydration by the AMC is dependent on molecular structure.

We further analysed the interaction patterns and driving forces for small contaminant removal by the AMC. Both the diffusion coefficient and mean squared displacement of the AMC Cl atoms increased after NO3 ions were added (Supplementary Table 3), which indicates that the Cl atoms had a higher mobility and the interaction between the Cl atoms and the AMC nitrogen atoms (Nq) weakened in the presence of NO3. By computing the RDF of the Cl atoms and NO3 oxygen atoms (On) that surround Nq, the On RDF exhibited a strong peak, whereas the Cl RDF exhibited a peak with half the intensity of the On peak. The On peak position was also closer to Nq than the Cl peak (Fig. 6i). This demonstrates that, in comparison with Cl, On has a stronger interaction with Nq. Combined with the dehydration of NO3 by the AMC, we conclude that dehydration caused the exposure of NO3 oxygen atoms, and then NO3 was removed through the electrostatic attraction between the positively charged N atoms of the AMC and the exposed O atoms of NO3.

The interaction between the AMC and DCF is demonstrated by the density distribution of atoms (Fig. 6j). Compared with a C atom in a hydrophobic benzene ring (Cb), the O atom in the DCF hydrophilic carboxyl (Oc) had a peak closer to Nq, which reveals that the electrostatic interaction between the DCF carboxyl and the AMC Nq is one of the key driving forces in DCF removal. However, as the DCF dehydration effect was not noticeable, water that surrounds Oc had a hindering effect on contact between Oc and Nq. As a result, only a small proportion of the DCF approached the AMC N atom surface, which indicates that their electrostatic attraction makes a minor contribution to DCF removal. Compared with the Nq surface, the peaks around the end of the carbon chains were stronger such that more DCF was attracted to the carbon chains (Fig. 6f–h). This is an indication that hydrophobic attraction was the dominant mechanism for DCF removal. The Cb peak was closer to that of carbon chains than the Oc peak, which suggests that the hydrophobic interaction was primarily between the AMC carbon chains and DCF benzene rings.

Our molecular dynamics simulations suggest that the AMC has varying dehydration effects for different dissolved contaminants. For species that undergo a strong dehydration, the AMC interacted with them mainly by electrostatic attraction, but for species that undergo a weak dehydration, the removal was primarily due to hydrophobic interactions. For contaminants without hydrophilic groups, such as anthracene, the AMC removed them only by hydrophobic interactions (all anthracene molecules were close to carbon chains (Supplementary Fig. 16)). Therefore, the degree of electrostatic attraction and hydrophobic interaction with the AMC core is dependent on the structure of the dissolved molecule because this controls the binding to the sites on the AMC core.

Comparison with conventional technologies

We compared coagulation with AMC with other conventional and state-of-the-art water treatment processes in terms of the removal of various contaminant types (Supplementary Table 4). Conventional coagulation can effectively remove particles, colloids, high-molecular-weight hydrophobic organic macromolecules (for example, humic acid) and some oxyanions (such as phosphate). Coagulation as it is currently practiced functions well as a pretreatment process for microfiltration/ultrafiltration (MF/UF), activated carbon adsorption and advanced oxidation processes (AOPs), which are either unable to remove particles and colloids or are subject to particles and colloids that cause process-limiting behaviour, such as membrane fouling. MF/UF can effectively remove suspended particles, whereas activated carbon adsorption and AOPs remove most dissolved organics. Combining conventional coagulation with MF/UF, activated carbon adsorption or AOPs significantly enhances the removal of organics. However, AOPs can lead to the formation of toxic by-products, and activated carbon adsorption has a low removal efficiency for hydrophilic organics. A common problem among conventional coagulants, MF/UF, activated carbon adsorption and AOPs is that they have a limited ability to remove small dissolved organic contaminants and small ionic contaminants, such as NO3.

By integrating membrane technologies, such as microfiltration, ultrafiltration, nanofiltration or reverse osmosis with conventional coagulation, the previously discussed limitations to removing various contaminants are mostly eliminated, but such treatment practices require long process trains and are complex to operate. The nanofiltration/reverse osmosis concentrate also needs treatment or disposal. In comparison, coagulation with the AMC has significant advantages over other treatment methods, which arise from its process simplicity, ease of operation and ability to thoroughly remove a broad spectrum of contaminants in a single treatment step.


Current advanced water treatment technologies are hampered by their contaminant-specific removal ability and therefore a space exists for process innovation. Inspired by the configuration of Actinia, we synthesized a biomimetic micellar nanocoagulant with a core–shell structure. The coagulant structure allows it to be stable in storage but also to evert its configuration when added to water at pH > 4.0, similar to Actinia’s ability to transform its shape (Supplementary Fig. 17). The shell of Si–Al complexes hydrolyses into flocs and destabilizes and enmeshes particles and colloids, whereas the aliphatic quaternary ammonium core captures dissolved organic and inorganic contaminants by electrostatic and hydrophobic interactions. The synergistic effect of the core and shell components allow the AMC to remove simultaneously a wide variety of contaminants with a greater than 90% efficiency. The coagulated water can also meet different water quality objectives by regulating the coagulant dosage. Due to its ability to remove a wide range of contaminants, the AMC presents the opportunity to expand the application of coagulants as a treatment technology and for significant process advances in water purification. One-step coagulation using the AMC can simplify the water treatment process as well as reduce treatment time, land use and capital cost.



Raw water was collected from a secondary wastewater treatment effluent of a municipal sewage plant in Beijing. Sudan III, DCF sodium, humic acid (molecular weight ranged from 600 Da to 6 kDa), ferric chloride (FeCl3·6H2O), aluminium sulfate (Al2(SO4)3·18H2O), polyDADMAC (40 wt% in water), and fluorescent dye DiOC6(3) were purchased from Sigma-Aldrich. TPODAC was purchased from Bailingwei.

Synthesis of the AMC

TPODAC was diluted in a mixture of ethanol, hydrogen chloride and water with a volume ratio of 1:30:10−2:7. The solution was stirred for 24 h to hydrolyse TPODAC to silanetriol with an aliphatic quaternary ammonium (3-((n-octadecyldimethyl)quaternaryammonium)propyl silanetriol). Dialysis was then performed in a hydrogen chloride solution (pH 2.2) to remove the solvent, reactants and reaction by-products with a 100 Da molecular weight cutoff membrane (Spectrum Labs). The dialysate was replaced three times during the first 3 h and twice during the following 6 h. The hydrolysed TPODAC was then mixed in an AlCl3 solution (Si:Al molar ratio of 0.1) and condensation took place by adjusting to pH 3.5 via the addition of 0.5 M NaOH. The resultant solution was stirred for 12 h and the total aluminium concentration was adjusted to 0.1 M (pH 4.0). With ageing at room temperature for 12 h, nanoparticles with a core–shell structure self-assembled. Experimental conditions for obtaining stable micelle nanocoagulants, such as the ratio of Si to Al, the pH during hydrolysis and condensation, and the concentration of AlT, were optimized by a series of pre-experiments.

Coagulation experiments

Coagulation experiments were performed using a program-controlled TA6-1 jar test apparatus with six paddles. The coagulants (Al2(SO4)3, FeCl3, polyDADMAC or the AMC) were added to water samples at various dosages. The samples were mixed at 250 r.p.m. for 1.5 min, stirred at 50 r.p.m. for 15 min and then settled for 30 min. Then, samples were drawn from 2 cm under the top surface and filtered through a 0.22 μm membrane for the analysis of residual contaminants. Wastewater effluent raw water, spiked with humic acid or DCF, was used to study the variation of the zeta potential during the removal of humic acid and DCF; water samples were drawn after 1.5 min of rapid stirring for the zeta potential analysis. The synthetic aqueous solution (deionized water, 3 mg l–1 DCF and 2 mM NaHCO3) was coagulated for ultrafiltration fractionation experiments and for the analysis of the coagulant–DCF complexes formed. A synthetic aqueous solution without DCF was coagulated for the EDX analysis after 30 s of coagulation and for DLS analysis at 30 s intervals during the coagulation experiment. The pH of the synthetic test water during the coagulation was fixed at 7.5.

Analytical methods

Turbidity was determined by a turbidimeter (2100 N Turbidimeter, Hach). The concentration of humic acid was measured using ultraviolet absorbance at 254 nm (UV254). UV254, NO3 and total phosphorus were detected by an ultraviolet–visible 8500 spectrophotometer (Techcomp). DOC was analysed using a 5000A total organic carbon analyser (Shimadzu). The concentrations of DCF, ibuprofen, salicylic acid, sulfamethoxazole, bezafibrate and perfluorocaprylic acid were measured by a HPLC tandem mass spectrometer (Agilent liquid chromatography 1290/6460 triple quadruple mass spectrometer). Fluorene, anthracene, fluoranthene, pyrene, benzo[b]fluoranthene and benzo[a]pyrene concentrations were analysed using HPLC with a photodiode array detector (Agilent Technologies 1200 Infinity HPLC). The Supplementary Information provides the sample extraction and analytical details for the HPLC tandem mass spectrometer and HPLC measurements.

Fourier transform infrared spectrum measurements of the AMC were recorded on an infrared spectrometer (Nicolet iS50). Zeta potentials were determined using a laser diffraction instrument (Zetasizer 2000). Particle Dh values were evaluated by DLS with a particle size analyser (Brookhaven Instruments). The morphology of the AMC and aggregates were characterized by a TEM (Hitachi, H-9000NAR) operating at 300 kV. Elemental mapping was carried out by a spherical aberration corrected TEM equipped with EDX spectroscopy (FEI Titan Cubed Themis G2 300) at 80 kV. Elemental line scanning was carried out by a FEI Tecnai F30 field emission TEM at 300 kV. Surface plasmon resonance microscopy imaging was conducted by a high numerical aperture objective (1.49) and an inverted microscope (Nikon ECLIPSE Ti). The Supplementary Information gives detailed information for the surface plasmon resonance microscopy experiment.

Ultrafiltration fractionation experiment

Ultrafiltration fractionation experiments were conducted to determine the molecular weight distribution of DCF in the coagulation samples. The coagulated water was filtered using a 0.22 μm membrane before fractionation. The ultrafiltration filtrate was subdivided using membranes in a sequential order of molecular weight cutoffs of 100, 50, 30, 10, 5 and 1 kDa in stirred Amicon ultrafiltration cells. The pressure during the ultrafiltration experiments was maintained at 0.1 MPa by nitrogen gas. The permeate from each membrane was collected and filtered with each subsequent membrane, which resulted in seven fractions with nominal molecular weights of >100, 50–100, 30–50, 10–30, 5–10, 1–5 and <1 kDa. Filtration was terminated when the volume of the retentate for any membrane was 0 ml.

Electrospray ionization FT-ICR MS analysis of organic components and coagulant–DCF complexes

A Bruker Apex ultra FT-ICR MS equipped with a 9.4 T superconducting magnet was used to analyse the water organic components and identify the coagulant–DCF complexes. For dissolved organic components, the extracted samples were ionized by electrospray ionization (ESI) in the negative mode. For coagulant–DCF complexes, the coagulation samples were ionized by ESI in the positive mode. All samples for the FT-ICR MS experiments were diluted in methanol solution and injected into the electrospray source at 250 μl h–1 using a syringe pump. The parameters of the FT-ICR MS for different ionization modes are shown in Supplementary Table 5. A total of 128 scans for the negative-ion ESI and 32 scans for the positive-ion ESI were conducted to enhance the signal-to-noise ratio. Compounds that contained P and Cl were excluded because of their low presence in samples (<0.5%). The Supplementary Information gives the procedures for sample extraction, analysis of dissolved organic components and identification of coagulant–DCF complexes.

Real-time fluorescence imaging of coagulated water

Lipophilic green-fluorescent dye DiOC6(3) was selected as a representative low-molecular-weight contaminant. A synthetic aqueous solution (5 mg ml–1 DiOC6(3) dissolved in a mixture of dimethyl sulfoxide and deionized water at a volume ratio of 1:10 and 2 mM NaHCO3) was coagulated for the fluorescence imaging, and the pH during the coagulation was controlled at 7.5. A fluorescence microscope (Olympus IX73) was used to observe the aggregation of DiOC6(3) over the coagulation process. Specimens were examined with a ×60 oil immersion objective lenses with a numerical aperture of 14; fluorescence images were obtained using a blue excitation filter (450–490 nm).

Molecular dynamics simulations

Atomic electrostatic potential fitted charges were derived by first-principle calculations based on density functional theory at B3LYP/RECP~6-31G* level using the Gaussian 09 program44,45. The resulting atomic charges were then used for classic molecular dynamics simulations. Water molecules were described by the simple point charge model46. The OPLS 2005 force field was used for the AMC simulation47, the general AMBER force field for nitrate48,49 and the consistent-valence force field for both DCF and anthracene50. Excess charges were neutralized by Cl ions and Na+ ions. For the simulation of the AMC structure, 16 AMC monomers and 300 water molecules were added to the system. To simulate the interaction between the AMC and small dissolved contaminants, the initial system configuration contained 3,000 water molecules and 16 AMC monomers. During the first stage of the simulation, NPT (constant number, pressure and temperature) time integration was performed at 1 atm for 1 ns, and the system was allowed to reach equilibrium. After 1 ns at NPT equilibrium, 6 NO3 ions and 6 Na+ ions were introduced into the AMC–NO3 system, 4 DCF anions and 4 Na+ ions into the AMC–DCF system and 4 anthracene molecules into the AMC–anthracene system. Then an NVT (constant number, volume and temperature) time integration was performed for 2 ns to record trajectories for the postanalysis. Interactions that exceeded 1 ns were considered stable and used for the analysis.

All molecular dynamics simulations were performed using DL_POLY 4.08 (ref. 51). The VMD18 program was employed for trajectory visualization and analysis52. The velocity Verlet integration algorithm with a time step of 1 fs was used in the molecular dynamics simulation. The Ewald summation method was applied for long-range Coulomb interactions53, whereas van der Waals forces and short-range interactions were truncated at 9 Å. During the simulation, the temperature was maintained at 298 K using the Nosé–Hoover thermostat. Periodic boundary conditions were applied in the x, y and z directions. The Supplementary Information supplies the molecular dynamics simulation data analysis methods of mean squared displacement, RDF and spatial density distribution.

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Additional information

Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.


  1. 1.

    Shannon, M. A. et al. Science and technology for water purification in the coming decades. Nature 452, 301–310 (2008).

  2. 2.

    Elimelech, M. The global challenge for adequate and safe water. J. Water Supply Res. Technol. AQUA 55, 3–10 (2006).

  3. 3.

    Qu, X. L., Alvarez, P. J. J. & Li, Q. L. Applications of nanotechnology in water and wastewater treatment. Water Res. 47, 3931–3946 (2013).

  4. 4.

    Service, R. F. Desalination freshens up. Science 313, 1088–1090 (2006).

  5. 5.

    Grant, S. B. et al. Taking the ‘waste’ out of ‘wastewater’ for human water security and ecosystem sustainability. Science 337, 681–686 (2012).

  6. 6.

    Ali, I. & Gupta, V. K. Advances in water treatment by adsorption technology. Nat. Protoc. 1, 2661–2667 (2006).

  7. 7.

    Hering, J. G., Waite, T. D., Luthy, R. G., Drewes, J. E. & Sedlak, D. L. A changing framework for urban water systems. Environ. Sci. Technol. 47, 10721–10726 (2013).

  8. 8.

    Schwarzenbach, R. P. et al. The challenge of micropollutants in aquatic systems. Science 313, 1072–1077 (2006).

  9. 9.

    Li, J. China gears up to tackle tainted water. Nature 499, 14–15 (2013).

  10. 10.

    Morel, F. M. M. & Hering, J. G. Principles and Applications of Aquatic Chemistry (Wiley, New York, 1993).

  11. 11.

    Richardson, S. D. & Temes, T. A. Water analysis: emerging contaminants and current issues. Anal. Chem. 90, 398–428 (2018).

  12. 12.

    Shon, H. K., Vigneswaran, S. & Snyder, S. A. Effluent organic matter (EfOM) in wastewater: constituents, effects, and treatment. Crit. Rev. Environ. Sci. Technol. 36, 327–374 (2006).

  13. 13.

    Luo, Y. L. et al. A review on the occurrence of micropollutants in the aquatic environment and their fate and removal during wastewater treatment. Sci. Total Environ. 473, 619–641 (2014).

  14. 14.

    Jin, L. Y., Zhang, G. M. & Tian, H. F. Current state of sewage treatment in China. Water Res. 66, 85–98 (2014).

  15. 15.

    Zodrow, K. R. et al. Advanced materials, technologies, and complex systems analyses: emerging opportunities to enhance urban water security. Environ. Sci. Technol. 51, 10274–10281 (2017).

  16. 16.

    Westerhoff, P., Alvarez, P., Li, Q. L., Gardea-Torresdey, J. & Zimmerman, J. Overcoming implementation barriers for nanotechnology in drinking water treatment. Environ. Sci. Nano 3, 1241–1253 (2016).

  17. 17.

    Mauter, M. S. et al. The role of nanotechnology in tackling global water challenges. Nat. Sustain. 1, 166–175 (2018).

  18. 18.

    Moreira, F. C., Boaventura, R. A. R., Brillas, E. & Vilar, V. J. P. Electrochemical advanced oxidation processes: a review on their application to synthetic and real wastewaters. Appl. Catal. B 202, 217–261 (2017).

  19. 19.

    Bolisetty, S. & Mezzenga, R. Amyloid–carbon hybrid membranes for universal water purification. Nat. Nanotech. 11, 365–371 (2016).

  20. 20.

    Werber, J. R., Osuji, C. O. & Elimelech, M. Materials for next-generation desalination and water purification membranes. Nat. Rev. Mater. 1, 16018 (2016).

  21. 21.

    Pintilie, L., Torres, C. M., Teodosiu, C. & Castells, F. Urban wastewater reclamation for industrial reuse: an LCA case study. J. Clean Prod. 139, 1–14 (2016).

  22. 22.

    Bratby, J. Coagulation and Flocculation in Water and Wastewater Treatment (IWA, London, 2006).

  23. 23.

    Wang, J. P., Yuan, S. J., Wang, Y. & Yu, H. Q. Synthesis, characterization and application of a novel starch-based flocculant with high flocculation and dewatering properties. Water Res. 47, 2643–2648 (2013).

  24. 24.

    Stumm, W., Morgan, J. J. & Black, A. P. Chemical aspects of coagulation [with discussion]. J. Am. Water Works Assoc. 54, 971–994 (1962).

  25. 25.

    Stumm, W. & O’Melia, C. R. Stoichiometry of coagulation. J. Am. Water Works Assoc. 60, 514–539 (1968).

  26. 26.

    Bolong, N., Ismail, A. F., Salim, M. R. & Matsuura, T. A review of the effects of emerging contaminants in wastewater and options for their removal. Desalination 239, 229–246 (2009).

  27. 27.

    Edzwald, J. K. Coagulation in drinking water treatment: particles, organics and coagulants. Water Sci. Technol. 27, 21–35 (1993).

  28. 28.

    Krasner, S. W. & Amy, G. Jar-test evaluations of enhanced coagulation. J. Am. Water Works Assoc. 87, 93–107 (1995).

  29. 29.

    Huang, B. C., Guan, Y. F., Chen, W. & Yu, H. Q. Membrane fouling characteristics and mitigation in a coagulation-assisted microfiltration process for municipal wastewater pretreatment. Water Res. 123, 216–223 (2017).

  30. 30.

    Fan, X. J. et al. Performance of an integrated process combining ozonation with ceramic membrane ultra-filtration for advanced treatment of drinking water. Desalination 335, 47–54 (2014).

  31. 31.

    Bolto, B., Dixon, D., Eldridge, R. & King, S. Removal of THM precursors by coagulation or ion exchange. Water Res. 36, 5066–5073 (2002).

  32. 32.

    Wang, X. M. et al. Preparation and evaluation of titanium-based xerogel as a promising coagulant for water/wastewater treatment. Environ. Sci. Technol. 50, 9619–9626 (2016).

  33. 33.

    Zeng, Y. & Park, J. Characterization and coagulation performance of a novel inorganic polymer coagulant–poly-zinc-silicate-sulfate. Colloid Surf. A 334, 147–154 (2009).

  34. 34.

    Tzoupanos, N. D., Zouboulis, A. I. & Tsoleridis, C. A. A systematic study for the characterization of a novel coagulant (polyaluminium silicate chloride). Colloid Surf. A 342, 30–39 (2009).

  35. 35.

    Haberkamp, J., Ruhl, A. S., Ernst, M. & Jekel, M. Impact of coagulation and adsorption on DOC fractions of secondary effluent and resulting fouling behaviour in ultrafiltration. Water Res. 41, 3794–3802 (2007).

  36. 36.

    Park, T., Ampunan, V., Lee, S. & Chung, E. Chemical behavior of different species of phosphorus in coagulation. Chemosphere 144, 2264–2269 (2016).

  37. 37.

    Lacasa, E., Canizares, P., Saez, C., Fernandez, F. J. & Rodrigo, M. A. Removal of nitrates from groundwater by electrocoagulation. Chem. Eng. J. 171, 1012–1017 (2011).

  38. 38.

    Gabaldon, C. et al. Biological nitrate removal from wastewater of a metal-finishing industry. J. Hazard. Mater. 148, 485–490 (2007).

  39. 39.

    Suarez, S., Lerna, J. M. & Omil, F. Pre-treatment of hospital wastewater by coagulation—flocculation and flotation. Bioresour. Technol. 100, 2138–2146 (2009).

  40. 40.

    Liu, J. L. & Wong, M. H. Pharmaceuticals and personal care products (PPCPs): a review on environmental contamination in China. Environ. Int. 59, 208–224 (2013).

  41. 41.

    Wei, J. C. et al. Comparison of coagulation behavior and floc structure characteristic of different polyferric-cationic polymer dual-coagulants in humic acid solution. Water Res. 43, 724–732 (2009).

  42. 42.

    Matilainen, A., Vepsalainen, M. & Sillanpaa, M. Natural organic matter removal by coagulation during drinking water treatment: a review. Adv. Colloid Interface Sci. 159, 189–197 (2010).

  43. 43.

    Wang, S. P. et al. Label-free imaging, detection, and mass measurement of single viruses by surface plasmon resonance. Proc. Natl Acad. Sci. USA 107, 16028–16032 (2010).

  44. 44.

    Hehre, W. J., Ditchfield, R. & Pople, J. A. Self-consistent molecular-orbital methods. XII. Further extensions of Gaussian-type basis sets for use in molecular-orbital studies of organic molecules. J. Chem. Phys. 56, 2257–2261 (1972).

  45. 45.

    Becke, A. D. Density functional thermochemistry. III. The role of exact exchange. J. Chem. Phys. 98, 5648–5652 (1993).

  46. 46.

    Mark, P. & Nilsson, L. Structure and dynamics of the TIP3P, SPC, and SPC/E water models at 298 K. J. Phys. Chem. A 105, 9954–9960 (2001).

  47. 47.

    Jorgensen, W. L., Maxwell, D. S. & TiradoRives, J. Development and testing of the OPLS all-atom force field on conformational energetics and properties of organic liquids. J. Am. Chem. Soc. 118, 11225–11236 (1996).

  48. 48.

    Jiang, W., Yan, T. Y., Wang, Y. T. & Voth, G. A. Molecular dynamics simulation of the energetic room-temperature ionic liquid, 1-hydroxyethyl-4-amino-1,2,4-triazolium nitrate (HEATN). J. Phys. Chem. B 112, 3121–3131 (2008).

  49. 49.

    Sprenger, K. G., Jaeger, V. W. & Pfaendtner, J. The general AMBER force field (GAFF) can accurately predict thermodynamic and transport properties of many ionic liquids. J. Phys. Chem. B 119, 5882–5895 (2015).

  50. 50.

    Asensio, J. L., Martinpastor, M. & Jimenezbarbero, J. The use of CVFF and CFF91 force fields in conformational analysis of carbohydrate molecules. Comparison with Amber molecular mechanics and dynamics calculations for methyl α-lactoside. Int. J. Biol. Macromol. 17, 137–148 (1995).

  51. 51.

    Smith, W. & Forester, T. R. DL_POLY_2.0: a general-purpose parallel molecular dynamics simulation package. J. Mol. Graph. 14, 136–141 (1996).

  52. 52.

    Humphrey, W., Dalke, A. & Schulten, K. VMD: visual molecular dynamics. J. Mol. Graph. 14, 33–38 (1996).

  53. 53.

    Darden, T., York, D. & Pedersen, L. Particle mesh Ewald: an log(N) method for Ewald sums in large systems. J. Chem. Phys. 98, 10089–10092 (1993).

Download references


The authors are grateful for financial support from the Major Program of the National Natural Science Foundation of China (grant no. 91434132), the Fund for Innovative Research Group of NSFC (grant no. 51721006) and the US National Science Foundation Graduate Research Fellowship awarded to R.M.D.

Author information


  1. Department of Environmental Engineering, Peking University, Beijing, China

    • Jinwei Liu
    • , Shihan Cheng
    •  & Huazhang Zhao
  2. The Key Laboratory of Water and Sediment Sciences, Ministry of Education, Beijing, China

    • Jinwei Liu
    •  & Huazhang Zhao
  3. Beijing Engineering Research Center of Advanced Wastewater Treatment, Beijing, China

    • Shihan Cheng
  4. College of Chemistry and Chemical Engineering, China University of Petroleum, Qingdao, China

    • Na Cao
    •  & Chunxiang Geng
  5. State Key Laboratory of Heavy Oil Processing, China University of Petroleum, Beijing, China

    • Chen He
    • , Quan Shi
    •  & Chunming Xu
  6. Beijing Innovation Center for Engineering Science and Advanced Technology, Peking University, Beijing, China

    • Jinren Ni
  7. State Key Laboratory of Plateau Ecology and Agriculture, Qinghai University, Xining, China

    • Jinren Ni
  8. Department of Chemical and Environmental Engineering, Yale University, New Haven, CT, USA

    • Ryan M. DuChanois
    •  & Menachem Elimelech


  1. Search for Jinwei Liu in:

  2. Search for Shihan Cheng in:

  3. Search for Na Cao in:

  4. Search for Chunxiang Geng in:

  5. Search for Chen He in:

  6. Search for Quan Shi in:

  7. Search for Chunming Xu in:

  8. Search for Jinren Ni in:

  9. Search for Ryan M. DuChanois in:

  10. Search for Menachem Elimelech in:

  11. Search for Huazhang Zhao in:


H.Z. conceived the initial idea and experimental design. M.E. and H.Z. supervised the study and experiments. J.L. performed the nanocoagulant synthesis and characterization experiments. N.C. and C.G. investigated the coagulation performance. J.N. and J.L. designed the coagulation behaviour experiments, and R.M.D. and J.L. carried out and analysed the experiments. S.C. carried out the molecular dynamics simulations. C.H., Q.S. and C.X. contributed to the data analysis of the FT-ICR mass spectra. All authors discussed the results and commented on the manuscript. M.E., R.M.D., H.Z. and J.L. wrote the paper with help from all authors.

Competing interests

The authors declare no competing interests.

Corresponding authors

Correspondence to Menachem Elimelech or Huazhang Zhao.

Supplementary information

  1. Supplementary information

    Supplementary Methods, Supplementary Figures 1–17, Supplementary Tables 1–5

About this article

Publication history






Further reading