Fluorescent Plastic Nanoparticles to Track their Interaction and Fate in Physiological Environments

This work aims to establish a production and characterization protocol for fluorescent plastic nanoparticles of poly(ethylene terephthalate) (PET), polypropylene (PP), and polystyrene (PS) that can be tracked in biological environments.


Section S1.2: Melt-Mixing Methods
Table S1: Material properties such as molecular weight (M w ) and density, dye mass (per 2.5 g of virgin polymer), and melt processing temperatures for each polymer.Values for the various material properties were obtained from the respective company that produced the virgin polymer pellets.The acronym n.r.indicates a value was not reported by the manufacturer.

Polymer
M w (kDA) Density (g/cm Virgin pellets, melt processed pellet controls, cryo-milled unlabeled microparticles, and cryo-milled fluorescent microparticles were placed into a DSC 2 Star System (Mettler Toledo, USA).Each sample of particles was subjected to two heating and two cooling cycles using heating and cooling rates of 10 °C per minute, with settings adjusted to account for differences in the anticipated transition temperatures of each material (Table S2).The same temperatures were utilized for labeled and unlabeled microparticle samples.

Section S2.2: DSC Results
The melt processing incorporation of the C1RG fluorophore and subsequent cryo-milling into plastic microparticles was possible for all materials utilized in the study.Upon completion of the melt S5 processing and cryo-milling for all materials, the impact of the melt processing treatment on the plastic microparticles' transition temperatures were assessed utilizing DSC (Figure S2).
The DSC traces of the virgin PS pellet, PS melt processed control, unlabeled PS microparticles, and the PSC1RG microparticles are virtually identical.As expected for this amorphous polymer, the only transition visible is the glass transition, which appears with a glass transition temperature (T g ) in agreement with values reported in literature to range from 100 ˚C to 107 ˚C [1] (e.g.near 102 ˚C for PSC1RG and unlabeled PS microparticles, near 108 ˚C for the melt processed PS control, and near 109 ˚C for the virgin PS).
The DSC traces of the virgin PP pellet, PP melt processed control, unlabeled PP microparticles, and PPC1RG microparticles are also practically identical.Crystallization temperatures (T c ) range from 115 ˚C for the virgin PP pellet to 116 ˚C (fist cooling) for the microparticles to 121 ˚C for the melt processed PP control.
Melting temperatures (T m ) range from 159 ˚C (second heating) for the PPC1RG microparticles to 162 ˚C for the unlabeled PP microparticles to 164 ˚C for both pellet controls.These values are in good agreement with the value of 117 ˚C for the T c and 160 ˚C for the T m reported in literature [2].No glass transition could be discerned for PP (according to literature around -10 ˚C [2]), presumably on account of the high crystallinity.
Interestingly, the DSC traces of the PET samples show noticeable differences.The first cooling/second heating traces of the unlabeled microparticles show a broad T c around 185 ˚C upon cooling, a T g around 80 °C, and a T m of 244 ˚C upon heating.The PETC1RG microparticles show a different crystallization behavior.The first cooling trace reveals that the crystallization peak is much sharper and T c is higher (205 °C) than in the unlabeled material, while the integration of the melting endotherm in the second heating reveals a slightly higher crystallinity than in the unlabeled material.A similar trend could be seen for the melt processed PET control; with a sharp crystallization peak at 205 °C and a T m of 251 °C.A S6 comparison of the DSC results for the microparticle and melt processed control samples with the DSC data for the virgin pellet revealed that there was a striking difference in the crystallization temperatures and peak shapes for the samples; with the virgin sample having a very broad T c peak at 150 °C and a T m of 248 °C.Thus, the peak sharpening and increase in the T c values for the milled and melt processed samples is likely the result of a reduction of the molecular weight of the polymer; a trend which was also reported by Romão et al. for recycled PET samples [3].Depending on the type of material processing the sample underwent prior to DSC measurements, two key sources of this damage should be considered.The first is a reduction in the molecular weight as a result of the mechanical forces the material is exposed to during the milling procedure; resulting in the increased T c observed for the unlabeled PET microplastic sample.
The second likely cause of the increase in the T c of the melt processed samples (i.e. the melt processed PET control pellet and the PETC1RG microparticles) is chain scission as the result of the thermo-mechanical forces that the material is subjected to during the melt mixing procedure; a phenomenon also reported by Spinacé et al. [4].However, in the study conducted by Spinacé et al. a maximum T c shift of 40 °C was observed; a difference which is likely the result of the pre-drying step conducted at 160 °C prior to their material processing [4].This additional step can be introduced into future protocols looking to work with PET samples to ensure that the damage to the polymer chains as a result of moisture present during the melt processing procedure [5] is mitigated.The impact of the materials' thermomechanical properties on the mechanical disintegration during milling and the morphology of the particles produced can clearly be seen in the SEM images (Figure S3).

S9
Additional images of the particles, obtained with cLSM, served two purposes.First, they allow for further analysis of the particle shape.Second, fluorescent and bright field scanning modes could be used to assess whether fluorophore incorporation was homogenous in the plastic microparticles, and whether the unlabeled plastics display any fluorescence of their own (i.e. as a result of the presence of additives such as optical brighteners).Particles composed of all unlabeled plastic types showed no fluorescence when imaged with an excitation laser wavelength of 488 nm.Particles which were melt-mixed with C1RG showed fluorescence at this wavelength.Additionally, the particles displayed this fluorescence in a homogenous manner, with a good correlation between the shape of the plastic microparticles in bright field images and the shape observed in the fluorescent images.Fluorescence was observed only for particles that were within the focal plane during imaging.from the remaining single bonds in the backbone, and peaks from 1450 cm -1 to 1600 cm -1 as a result of the C=C stretch in the aromatic ring [8].PPC1RG had a single, broad peak at 2950 cm -1 as a result of C-H stretching in the S15 alkane backbone and multiple sharper peaks between 1470 cm -1 and 1450 cm -1 as a result of C-H bending from the alkanes.Smaller peaks near 715-725 cm -1 are seen as the result of -CH 2 groups [8].The PET and PETC1RG spectra show sharp peaks near 1640-1670 cm -1 as a result of the alkene C=C stretch, a peak near 1020-1070 cm -1 and 1200-1275 cm -1 as a result of the C-O-C stretch of the ethers present with the aromatic rings, and multiple peaks present below 1,000 cm -1 as a result of the multiple para-aromatic rings present within the backbone [8].

Section S4.2: Zeta Potential Methods
All zeta potential (ζ-potential) measurements were obtained at room temperature with a 90Plus particle size analyzer (Brookhaven, USA).Particles were suspended in MilliQ water at neutral pH.

Section S4.3: Zeta Potential Results
The aggregates which could be seen in microscopy images highlighted the need to assess the colloidal stability and size of the particles in their suspended state.Thus, ζ-potentials for the nanoparticle dispersions were measured (Table S5).Unlabeled and C1RG labelled PET nanoparticles, with ζ-potential values near the +/-30 mV commonly reported as the indicator for good particle stability [9] when suspended in MilliQ water, required no additional treatment prior to DLS and DDLS measurements.
PSC1RG and PPC1RG nanoparticles rapidly sedimented upon dialysis completion.Thus, accurate ζpotential and light scattering measurements could not be obtained for these particles until they were restabilized with 1% SDS.ζ-potential values of -43 mV for PSC1RG and -36 mV for PPC1RG were obtained after surfactant stabilization (Table S5). to ϴ = 150° with 10 measurements of 30 seconds taken per angle.The incident beam was formed by a linearly polarized and collimated laser beam (Cobolt 05-01 diode pumped solid state laser, λ = 660 nm, P max.= 500 mW), and the scattered light was collected by single-mode optical fibers equipped with integrated collimation optics.The incoming laser beam passed through a Glan-Thompson polarizer with an extinction ratio of 10 -6 , and another Glan-Thompson polarizer with an extinction ratio of 10 -8 was placed in front of the collection optics (for polarized and depolarized light scattering) [10].To construct the intensity auto-correlation function , the collected light was coupled into two APD detectors via laser-line filters (Perkin Elmer, Single Photon Counting Module), and their outputs were fed into a multitau digital correlator (LSI Correlator LS Instruments AG, Switzerland).To improve the signal-to-noise ratio and to eliminate the impact of detector after-pulsing on at early lag times (<1 µs), necessary for the cumulant analysis, these two detectors were cross-correlated.The field auto-correlation function was obtained via the Siegert relation (eqn.1): .
Without any modification made, the photon count traces of one of the detectors were obtained through the same detection line as above, at a sampling rate of nearly 19 Hz.To maximize the available accuracy and precision, the data analysis-akin to the well-known cumulant analysis resulting in the estimation of the scattering intensity-weighted average hydrodynamic radius, also known as z-average, included an unbiased classification of data quality based on the statistical analysis of photon counts [11][12][13][14].
The plastic nanoparticles were also dispersed in cell culture medium (cDMEM supplemented with proteins; 1.2 μg mL -1 PSC1RG; 2.1 μg mL -1 PPC1RG; 10.2 μg mL -1 PETC1RG; 7.9 μg mL -1 unlabeled PET), and if the scattering intensity from the media itself was not less than 5% compared to that of the particles, polarized DLS spectra (10 measurements of 30 seconds at ϴ = 90˚) were analyzed as reported elsewhere [15].experiment was amorphous, the nanoparticles of this material did not scatter light in an anisotropic manner and could not be seen in DDLS.

Section S4.6: Light Scattering in Cell Culture Media
Prior to exposing the particles to cell cultures, deeper understanding of the behavior of plastic nanoparticles in cell culture media containing serum was required.It was determined that DDLS measurements were not suitable for characterization of the highly amorphous PSC1RG nanoparticles (Figure S7) as they have no crystalline regions; a feature that is a necessary requirement to obtain the anisotropic scattering needed to size with DDLS.Thus, DLS measurements were utilized to measure the hydrodynamic radii of the plastic nanoparticles dispersed in cell culture media.As the concentration and size of the proteins within the cell culture medium is relatively constant, the scattering of the proteins and other components present within the cell culture medium are assumed to remain static even after the addition of the nanoparticles.This assumption relies on the considerably higher concentration of proteins present in cell culture medium (3.0-4.5 g L -1 [16]) compared to the concentration of the nanoparticles added (0.01021 g L -1 for PETC1RG particles, 0.0012 g L -1 for PSC1RG particles, and 0.00214 g L -1 for PPC1RG particles).At such high concentration differences, the amount of proteins interacting with the surface of the plastic nanoparticles, resulting in the formation of a protein corona, is negligible.Therefore, this interaction is assumed to have no effect and the overall behavior and composition of the cell culture media and the scattering of the proteins can be treated as static noise that is accounted for by subtracting the corresponding autocorrelation function [17].S20 The change in crystal frequency, , is given by eqn.3: is a constant value of 1.25 ng Hz -1 that accounts for the crystal area and various physical properties of   the quartz [18]. is the change in the mass after deposition of the particles.is the crystal frequency ∆  0 prior to deposition, with being the final stable frequency value (Figure S9).

Figure S1 :
Figure S1: The NanoWitt-Lab mill (FREWITT SA, Switzerland) and chemical structure of the polymers and the

Figure S2 :
Figure S2: Differential Scanning Calorimetry plots recorded for virgin plastic pellets, melt processed plastic pellet controls, unlabeled plastic microparticles, and

Figure S5 :
Figure S5: Chemical fingerprints obtained for all materials used in the study.Fingerprints were obtained after each

Figure S6 :
Figure S6: Histograms of the sizes measured for the plastic nanoparticles during scanning electron microscopy

S18Section S4. 5 :
Figure S7: Plots of the decay constant (Γ) of the particles against their squared momentum transfer (Q 2 ) which were

Figure S8 :
Figure S8: Field auto-correlation functions obtained for all particles dispersed in cDMEM.For each measurement,

FigureS4. 8 :
Figure S9: A representative quartz crystal microbalance plot obtained for a single concentration measurement of

Table S2 :
Temperatures utilized for differential scanning calorimetry measurements of plastic pellets and microparticles.

Table S3 :
A summary of the transition temperature ranges measured with DSC for each polymer type.

Table S4 :
Final average contact angle measured for thin films of each plastic microparticle type compared to literature values.
S13Figure S4: Representative images of 2 µl water droplets utilized to obtain contact angle measurements for thin films of each plastic particle type.

Table S5 :
A summary of the ζ-potential values for all stable particle suspensions * Indicates particle stabilization with 1% SDS was required to obtain accurate measurements.and without stabilization with 1% SDS, at constant temperature (21 °C) on a commercial goniometer instrument (3D LS Spectrometer, LS Instruments AG, Switzerland).Scattering angles ranged from ϴ = 30°