Chemicals and materials
High-quality PS microplastic powder was obtained from Sigma-Aldrich, while various PS microplastics in different shapes were sourced from Thermo Fisher and Chem-Impex. We used X-ray photoelectron spectroscopy (PHI Versaprobe II) to assess the surface purity of these model PS microplastics before proceeding with further tests. Other materials included PE, PMMA, and PVC, which were also sourced from Sigma-Aldrich in micrometre-sized powder form; PET and PP powders were purchased from Goodfellow and Materials Science, respectively.
We obtained Rh800 from Sigma-Aldrich and DMPO from TCI America. Other chemicals like Triton X-100, methyl orange, and ThT were from Thermo Fisher. APF, as a 5 mM stock in N,N-dimethylformamide (DMF), was diluted with phosphate-buffered saline (PBS; pH 7.4) to create working concentrations, ensuring robust fluorescence quantification under microscopic observation. To visualize the distribution on the plastic surface, we utilized R-PE (from BioLegend) and WGA–Alexa Fluor 647 conjugate (from Thermo Fisher Scientific). The single-stranded DNA labeled with Alexa 660 (5′-Alexa 660-CAC GCG AAG ACT TAA GAT A-3′) was synthesized and purified by Integrated DNA Technologies.
All aqueous solutions were formulated using Milli-Q ultrapure water (18.2 MΩ·cm). Solvents like methanol (MeOH), ethanol (EtOH), acetonitrile (MeCN), and DMF were of HPLC quality. Standard buffers and salts (such as NaCl, FeCl2, NaOH, triethylammonium acetate, and trifluoroacetic acid) were primarily sourced from Sigma-Aldrich unless indicated otherwise.
We pre-cleaned microscopy consumables, including glass coverslips, using a piranha solution (3:1 H2SO4:H2O2) and rinsed them thoroughly with water. For measuring zeta potential, disposable cuvettes were employed with a Malvern ZS Nanosizer instrument.
Methyl orange degradation
A 20 µM methyl orange solution was prepared in various solvents (H2O, MeOH, EtOH, MeCN, and DMF), which were then mixed with PS powder (from Sigma-Aldrich) at a concentration of 10 mg/ml. The solution was stirred at room temperature at 900 rpm over 10 days. Samples were collected every two days to check for potential degradation of methyl orange. Suspensions were centrifuged at 16,000 rpm for three minutes to separate the supernatant. The absorbance of methyl orange was calculated as OD465nm–OD700nm, where the optical densities were recorded at 465 and 700 nm. The subtraction of OD700nm accounted for background signals, including any scattering from residual particles. Supernatants were tested in triplicate (100 µl per well) using a 96-well plate reader. Control samples of methyl orange prepared with the same solvents and concentration but without PS were monitored simultaneously under identical conditions.
ESR measurements
DMPO was utilized for trapping radicals for detection via ESR spectroscopy. To investigate the spontaneous formation of radicals at the PS–water interface, 200 mg of PS powder was combined with a 20 mM DMPO solution in water and left to incubate at room temperature for 20 minutes before filtration. The resulting filtrate was transferred to a Wilmad Glass quartz flat cell for ESR spectroscopy measurements, which were conducted at room temperature using a Bruker ELEXSYS E500 EPR spectrometer under specific conditions: microwave power of 20.23 mW, averaging 12 scans, centre field at 3,480 G, sweep width of 100 G, modulation amplitude at 1 G, and microwave frequency at 9.768037 GHz.
APF radical sensing assay
Confocal fluorescence imaging
APF (5 mM; Thermo Fisher) was diluted to 5 µM using PBS. We mixed 1 ml of this dilution with about 1 mg of microplastic particles (PP, PS, PE, PET, PVC, or PMMA) through vortexing. Aliquots of this mixture were placed between a glass chamber made with an imaging spacer that separated a treated cover glass from a glass slide. We performed confocal fluorescence imaging promptly after preparing the sample. The green fluorescence channel was excited using an argon laser at 488 nm, capturing emissions between 505–520 nm. Time-lapse images were collected every three minutes with optimized excitation power to reduce photobleaching during the in situ tracking of the fluorogenic APF signal.
For end-point quantification of the APF assay, mixtures containing 5 µM APF and approximately 1 mg of microplastics were dispersed and incubated in the dark at room temperature for 90 minutes to allow the fluorogenic reaction at the plastic–water interface. After incubation, aliquots from each sample were mounted in imaging chambers made of poly-l-lysine coated coverslips and a glass slide separated by a spacer. Consistent laser power, PMT voltage, detector gain, and pinhole size were maintained across all measurements for quantitative reliability. We imaged multiple overlapping fields of view for each type of polymer to consider potential spatial variability in fluorescence signals.
Quantification of APF fluorescence was carried out by creating maximum-intensity Z projections from the confocal stacks. Regions of interest around the particles were segmented using corresponding stimulated Raman scattering (SRS) images as references. We extracted the fluorescence intensities as area-weighted averages from these segmented regions of interest to represent the relative reactivity of each type of microplastic.
SRS imaging
SRS imaging was utilized to characterize microplastics while performing confocal imaging analysis. This was conducted using a confocal laser-scanning microscope (FV1200, Olympus) with integrated customized laser sources. A commercial laser system (PicoEmerald S, Applied Physics & Electronics) provided both pump and Stokes beams. The Stokes wavelength remained fixed at 1,031.8 nm and modulated at 20 MHz, while the pump beam was tunable at 784.2 nm to excite C–H vibrational modes that are characteristic of microplastics. Both beams had approximately 2-ps pulse widths and 80-MHz repetition rates, focused tightly onto the sample with a ×25 water-immersion objective (XLPlan N, 1.05 NA MP, Olympus). The transmitted beams were collected using an oil-immersion condenser with a numerical aperture of 1.4, with the Stokes beam removed using a high-optical-density bandpass filter. The transmitted pump beam was detected by a silicon photodiode under 64-V bias, with its output processed through a load and demodulated using a high-frequency lock-in amplifier at 20 MHz, leading to SRS images being generated.
HPLC analysis
An Agilent 1200 series system equipped with a Vydac 218TP-C18 column was utilized for HPLC analysis. The gradient was ramped from 5% to 95% solvent B over 15 minutes (solvent A comprised of water with 0.1 M triethylammonium acetate; solvent B was acetonitrile with 0.1% trifluoroacetic acid). The detection unit monitored the reaction by measuring absorbance at 488 nm. Peaks corresponding to APF and its oxidation product fluorescein were identified based on retention times from a standard solution.
To study the conversion of APF in water upon PS powder addition, we suspended 100 mg of PS powder in 1 ml of a 20 µM APF solution in water with and without 0.1% Triton X-100. This suspension was stirred in darkness for 16 hours at room temperature before undergoing filtration. A positive control was created by mixing 20 µM APF with 20 mM H2O2 and 2 mM FeCl2 in water, and this was also incubated for the same duration prior to HPLC analysis.
Surface localization imaging assay
Surface localization of charged fluorophore
PS particles (1 mg/ml) were suspended in 1 µM Rh800 prepared in various solvent conditions, including H2O, 0.1% Triton X-100 in H2O, and EtOH. These suspensions underwent a dark incubation at room temperature for 10 minutes prior to confocal fluorescence imaging. The far-red channel was excited using a 635-nm laser, with emission collected between 655–755 nm.
Surface localization of protein
Plastic particles (10 mg/ml) were combined with protein in PBS (0.02 mg/ml) to visualize protein adsorption to the surface via confocal fluorescence imaging. Proteins with different net charges within the PBS environment were used: R-PE (from Biolegend) being negatively charged and WGA (from Thermo Fisher, W32466) being positively charged at pH 7.4. The mixture was vortexed and kept in the dark for 15 minutes before being placed in a glass chamber similar to the previous setup. WGA conjugated with Alexa Fluor 647 was imaged in the far-red channel using a 635-nm laser, with fluorescence emission collected between 655 nm and 755 nm. Simultaneous imaging in two channels revealed different distribution patterns of oppositely charged proteins on the plastic surfaces.
Surface localization of DNA
Alexa 660-labeled DNA (500 nM; 5′-Alexa 660-CAC GCG AAG ACT TAA GAT A-3′ from Integrated DNA Technologies) was added to plastic particles (around 10 mg/ml) for imaging surface adsorption using confocal fluorescence imaging. This mixture was vortexed and incubated in the dark for 15 minutes before sampling. The sample was then positioned in a glass chamber with piranha-solution-treated cover glass for confocal imaging, using the far-red channel excited at 635 nm and capturing emissions at 655–755 nm.
ThT assay for protein aggregation analysis
ThT serves as a standard for identifying amyloid fibril formation, both in vivo and in vitro. In this study, a 0.1 mg/ml lysozyme solution stained with 20 μM ThT in PBS was mixed with plastic particles at 1 mg/ml concentration. The blend underwent vortexing before being placed in a glass chamber and imaged via confocal fluorescence. ThT fluorescence was excited at 488 nm and measured in the 505–520 nm range. Time-lapse imaging was performed on selected particles at five-minute intervals.
For quantification, fluorescence images were captured after incubating the mixture in darkness for two hours. Z-stack images were compiled for each area of interest to analyze ThT fluorescence at the particle surface. We quantified fluorescence signals following maximum intensity projection by segmenting the particles in reference to corresponding SRS images, recording area-weighted average fluorescence for the designated particle regions of interest.
Haemolytic assay
Whole blood was drawn from wild-type mice (BALB/cJ, 20 weeks old) following an approved institutional protocol (AC-AABN0554) from Columbia University’s Animal Care and Use Committee (IACUC). Approximately 10 mM EDTA was added to prevent coagulation. Haemolysis was assessed per an established protocol. Blood samples were centrifuged at 1,700 g for five minutes, and the supernatant collected before washing thrice with PBS (pH 7) until the solution was clear. The resultant pellet was diluted in PBS (1:100) to create an erythrocyte suspension. To evaluate haemolytic activity, 150 μl of erythrocyte suspension was mixed with 2 mg of plastic particles and incubated for an hour at 37°C without agitation. The same erythrocyte suspension without plastic particles served as a control. The haemolytic activity was quantified by measuring the optical density at 405 nm of the supernatant using a 96-well plate reader.
Zeta-potential measurement
The electrophoretic mobilities of PS particles in varying conditions were evaluated using a Malvern ZS Nanosizer instrument. Both pristine PS microparticles and carboxylic-acid-functionalized PS microparticles were prepared at 1 mg/ml concentration in a 1 mM NaCl aqueous solution to check their surface electrostatics sensitivity against different concentrations of Triton X-100. Pure water and ethanol were also employed to suspend the microparticles at the same concentration to analyze solvent sensitivity. The zeta-potential values were computed from the electrophoretic mobility using a specific formula. All measurements took place at 25°C using disposable plastic cuvettes.
Theory and simulation
Non-reactive and reactive force fields
This study employed the OPLS-AA force-field combination for the polymer analysis along with the SPC/E water model. However, for VIE calculations, a more advanced ReaxFF/C-GeM model was utilized to depict water, OH−, and Na+. Previous findings indicate that classical fixed-charge OH− force fields can oftentimes show overcoordinated hydration environments. Conversely, ab initio simulations and experiments suggest hydroxide is better described as a hypercoordinated species with roughly three to five hydrogen bonds. Thus, reactive force-field methods like ReaxFF/C-GeM offer a more accurate physical representation by taking into account charge redistribution and proton-transfer dynamics. The ReaxFF/C-GeM model has been shown to accurately depict the energetics and dynamics involved with hydronium and hydroxide ion species, specifically addressing proton-transfer events critical for hydroxide diffusion.
Electrostatic embedding model for ReaxFF/C-GeM
This research implemented molecular dynamics simulations of an interfacial system with aqueous NaOH alongside three microplastic systems in slab geometry using an electrostatic embedding method. This method modeled both water and NaOH using the reactive ReaxFF/C-GeM force field and described the non-reactive polymer atoms via the OPLS-AA force field.
ReaxFF/C-GeM atoms encompass all water and NaOH treated with the reactive force field.
All polymer atoms are exclusively defined by a non-reactive force field.
The non-bonded interactions between ReaxFF/C-GeM and OPLS-AA regions, which include van der Waals and electrostatic interactions, further bridge the two force-field representations.
This strategy for electrostatic embedding modeling has been successfully applied and validated in previous studies involving neutral and charged reverse micelles. The force-field combination was benchmarked against various experimental observables related to interfacial proton-transfer phenomena.
Further benchmarking of force-field combinations
We also assessed interaction energies and intermolecular distances related to hydrogen bonds for polymer–water model systems. The corresponding data confirms that both OPLS-AA/SPC/E and OPLS-AA/ReaxFF-C-GeM reproduce expected qualitative trends in interaction strength across different systems, with the latter demonstrating closer agreement with density functional theory concerning binding energies and intermolecular distances, especially for PMMA. This suggests the OPLS-AA/ReaxFF-C-GeM hybrid approach offers a reliable depiction of polymer–water interactions at the interface.
Simulation details for OPLS-AA/SPC/E used for structural analysis
The plastics and Triton X-100 surfactant molecules were modeled with the all-atom OPLS-AA force field and water using the SPC/E model, constructing a system with a single long polymer chain in a slab geometry, totaling around 40,000 polymer atoms and 9,000 water molecules. Non-bonded interactions were truncated at 10 Å; long-range electrostatics were computed using a particle-particle particle-mesh Ewald method with a specific convergence criterion. Molecular dynamics simulations were carried out using LAMMPS with a 1-fs timestep and applied SHAKE constraints to SPC/E water molecules, maintaining the temperature at 300 K and constant pressure in the z direction.
Simulation details for OPLS-AA/ReaxFF-C-GeM used for structural analysis
We carried out molecular dynamic simulations with the electrostatic embedded model using a modified version of LAMMPS for PE, PVC, and PMMA. Each system consisted of a long polymer chain in slab geometry with around 8,000 polymer atoms and 2,000 water molecules for ReaxFF/C-GeM, using a 0.25-fs timestep. The temperature and pressure were maintained consistently as detailed previously. The initial polymer slabs underwent equilibration before adding water, followed by further relaxation and equilibration before production runs for sampling configurations.
Simulation quantities for the microplastics
Our analysis involved evaluating both polymer density and the distribution of surface-exposed functional groups at the interface for the OPLS-AA simulations. The equilibrated microplastic systems exhibited physically plausible bulk densities using a standard protocol. Surface-exposed functional groups were identified as those positioned close to the Willard–Chandler interface, displaying stable distributions throughout the simulation. Moreover, comparisons between OPLS-AA/SPC/E and OPLS-AA/ReaxFF-C-GeM suggest that the interfacial functional group exposure remains steady regardless of the water model, indicating that the interfacial configuration doesn’t undergo significant reconstruction over the simulation timescale.
VIE calculations and sampling details
In the VIE calculations, four NaOH molecules were incorporated into the water phase, with the surface charge density following previous reports. The VIEs were computed at the ωB97M-V level of theory, with explicit solvent included within 6 Å of the hydroxide ion, with hydrogen bonds defined based on specific distance criteria. The results indicate a net charge imbalance in the interfacial region, suggesting that the particle carries excess negative charge. Consequently, all Na+ ions were removed to simulate this charge imbalance adequately.
