Effect of Atom Transfer Radical Polymerization Reaction Time on PCB Binding Capacities of Styrene-CMA/QMA Core-Shell Iron Oxide Nanoparticles

Abstract

Water pollution continues to be one of the greatest challenges humankind faces worldwide. Increasing population growth, fast industrialization and modernization risk the worsening of water accessibility and quality in the coming years. Nanoadsorbents have steadily gained attention as remediation technologies that can meet stringent water quality demands. In this work, core-shell magnetic nanoparticles (MNPs) comprised of an iron oxide magnetic core and a styrene based polymer shell were synthesized via surface initiated atom transfer radical polymerization (SI-ATRP), and characterized them for their binding of polychlorinated biphenyls (PCBs), as model organic contaminants. Acrylated plant derived polyphenols, curcumin multiacrylate (CMA) and quercetin multiacrylate (QMA), and divinylbenzene (DVB) were incorporated into the polymeric shell to create high affinity binding sites for PCBs. The affinity of these novel materials for PCB 126 was evaluated and fitted to the nonlinear Langmuir model to determine binding affinities (K_D). The K_D values obtained for all the MNP systems showed higher binding affinities for PCB 126 that carbonaceous materials, like activated carbon and graphene oxide, the most widely used adsorption materials for water remediation today. The effect of increasing ATRP reaction time on the binding affinity of MNPs demonstrated the ability to tune polymer shell thickness by modifying the reaction extent and initial crosslinker concentrations in order to maximize pollutant binding. The enhancement in binding affinity and capacity for PCB 126 was demonstrated by the use of hydrophobic, aromatic rich molecules like styrene, CMA, QMA and DVB, within the polymeric shell provides more sites for π-π interactions to occur between the MNP surface and the PCB molecules. Overall, the high affinities for PCBs, as model organic pollutants, and magnetic capabilities of the core-shell MNPs synthesized provide a strong rationale for their application as nanoadsorbents in the environmental remediation of specific harmful contaminants.

1. Introduction

Atom transfer radical polymerization (ATRP) is one of the most frequently applied techniques in the engineering of surfaces and interfaces with polymers brushes. ATRP is a powerful technique that allows the tuning of chemical and physical properties of a surface/interface due to its simple experimental set up, performance under mild reaction conditions, tolerance for a variety of functional groups, and compatibility with organic and inorganic solvents.^[1] There are two main strategies to graft polymer brushes onto a surface: ‘grafting to’ or ‘grafting from’ approach. Of particular interest are methods on the ‘grafting from’ approach, where a surface initiator is first anchored, and then in situ polymerization occurs to generate a polymer brush.^[2,3] An example of this approach is surface initiated atom transfer radical polymerization (SI-ATRP). SI-ATRP is a well-established technique that offer control over the polymer thickness and densities. Additionally, it allows for the growth of polymer brushes on virtually any surface, as long as the surface initiator is properly selected.

Surface initiated ATRP has been widely used to grow polymers from a variety of nanoparticle surfaces, such as Au, Ni, MnFe₂O₄, BaFe₂O₃, Fe₃O₄, among others.^[4-8] These types of ATRP synthesis give rise to core-shell nanoparticles, an ideal composite system that combines the advantages of the polymeric shell and the metallic core, that offers enhanced physical and chemical properties. Of particular interest is the formation of core-shell nanoparticles is the use of iron oxide nanoparticles (IO MNPs) to obtain magnetic nanoparticles. Herein, the core consists of magnetite (Fe₃O₄) or maghemite (γ-Fe₂O₃), which can be superparamagnetic, meaning that upon exposure to an external magnetic field the particles will rapidly aggregate together, yet able to redisperse back in solution once the magnetic field is removed.^[9-11] This characteristic allows the IO MNPs to be magnetically separated from solution with the use of a static magnetic field. Additionally, IO MNPs have the ability to respond to an alternating magnetic field (AMF) by converting magnetic work into internal energy, through magnetic relaxation processes, and dissipating it as heat.^[12,13] The polymeric coatings on the IO MNPs can improve the stability of the particles in solution, prevent their aggregation and protect them from oxidation. These functional polymer shells can provide the core-shell magnetic nanoparticles with desired functionalities to tailor their composition for specific applications. Polymer usually possess tunable porous structures, excellent mechanical properties, and a variety of functional groups. Because of this, core-shell nanoparticles have found application in a variety of areas like drug delivery, magnetic resonance, cancer treatment, rheology, energy storage, and environmental remediation, among others.^[14-20]

Core-shell magnetic nanoparticles have gained growing appeal in the environmental field for their versatility in polymer functionality, and core magnetic functions that allow for magnetic separation from the contaminated media. This grants the nanoparticles with a significant advantage over other remediation technologies: a fast and easy way to recover the sorbent material from raw environmental samples without the need of centrifugation or filtration steps.^[21] In addition, nanoadsorbents have a very high specific surface, high associated sorption sites, tunable porosity, and have been successfully been employed in environmental applications for pollutant mitigation and removal.^[22-24] The selection of the various functional monomers or crosslinkers to obtain the polymer shell is designed base on the final application. Studies have shown that the introduction of aromatic functional groups into the functional polymer shell increased the affinity of the core-shell nanoparticles for aromatic compounds.^[25,26] More specifically, styrene and divinylbenzene have been shown to be relatively selective for analytes with aromatic rings due to their specific π-π interactions, and have been used to remove aromatic pollutants from water.^[27] Given this information, it is expected that incorporating any aromatic rich molecule into the polymer shell will increase affinity for aromatic analytes. One such group is plant derived polyphenols, like curcumin and quercetin. These naturally occurring, complex heterocyclic molecules can be acrylated to produce functional monomers to be used in SI-ATRP of core-shell nanoparticles.^[28-30]

One important use of core-shell magnetic nanoparticles is as nanoadsorbents in water treatment. Water pollution is a worldwide problem that needs to be addressed. Many persistent organic pollutants (POPs), persist in the environment despite their production having been banned decades ago or being under strict regulations today. One such class is polychlorinated biphenyls (PCBs). Despite their production being banned in 1979 in the United States and in 2001 by the Stockholm Convention on Persistent Organic Pollutants, concentrations of PCB congeners can still be found in water, sediment, soil, aquatic biota and other animals throughout the world.^[31-33] PCBs have low volatilities and poor aqueous solubility, making their extraction from water and soil very challenging. Current remediation technologies are either too time consuming, not efficient enough at removal of the pollutant, and/or too costly.^[34,35]

In this work, core-shell magnetic nanoparticles were prepared using SI-ATRP to coat IO-MNPs with a styrene-based polymer shell crosslinked with acrylated plant derived polyphenols. Curcumin multiacrylate (CMA) and quercetin multiacrylate (QMA) were synthesized and incorporated in the core-shell magnetic nanoparticles to enhance their adsorption capacity for PCBs. Divinylbenzene (DVB) crosslinked systems were also studied as a comparison group. The effect of ATRP synthesis time on the shell thickness was studied at two different initial acrylated polyphenol or DVB loadings. The functionalized nanoparticle systems were characterized for size, shell coating percent, and stability. The binding isotherm for a model contaminant, PCB 126, was studied, and the binding constants for the four systems synthesized were evaluated using the Langmuir adsorption model.

2. Experimental details

2.1. Materials

Iron (III) chloride hexahydrate (FeCl₃ • 6 H₂O); iron chloride tetrahydrate (FeCl₂ • 4 H₂O); 2-bromo-2-methyl propionic acid (BMPA); 4,4’-dinoyl-2,2’-dipyridil (DNDP); copper (I) bromide (CuBr); copper (II) dibromide (CuBr₂); triethyl amine (TEA), acryloyl chloride; and potassium carbonate (K₂CO₃) were obtained from Sigma Aldrich (St Louis, MO). Ammonium hydroxide (NH₄OH) was purchased from EMD Chemicals (Gibbstown, NJ). Styrene (Sty) and divinylbenzene (DVB) were obtained from Polysciences INC. (Warrington, PA). Curcumin was purchased from Chem-Impex International, Inc. (Bensenville, IL), and quercetin was purchased from Cayman Chemicals (Ann Arbor, MI). 3,3',4,4',5-Pentachlorobiphenyl (PCB-126) in isooctane was purchased from Accustandard (New Haven, CT). All solvents (Isooctane, ethanol HPLC grade, xylene, toluene, tetrahydrofuran (THF); dichloromethane (DCM), acetonitrile (ACN)) were obtained from Fisher Scientific (Hannover Park, IL). All materials were used as received.

2.2. Iron oxide nanoparticle synthesis

Iron oxide magnetic nanoparticles (IO MNPs) were synthesized through a one-pot co-precipitation method.^[10] Iron chloride salts, FeCl₃ • 6 H₂O and FeCl₂ • 4 H₂O, were dissolved in 40mL of DI water in a 2:1 molar ratio of respectively, and combined in a sealed 3-neck flask under vigorous stirring and nitrogen flow to achieve an inert synthesis environment. The flask was heated to 85°C and 5 mL of NH₄OH (30.0 % v/v) was injected dropwise into the vessel. The reaction was carried out for 1 h. The nanoparticles were magnetically decanted and washed three times against DI water. Finally, the particles were re-suspended in 45 mL of DI water and dialyzed for 24 h. (100 kDA molecular weight cutoff)

2.3. BMPA initiator addition

The iron oxide nanoparticles were mixed in a 1:4 molar ratio with the 2-bromo-2-methyl propionic acid (BMPA) initiator in a 75-25 ethanol – DI water solvent. The mixture was stirred for 24 h at room temperature. Following this, the particles were magnetically decanted and washed twice times with ethanol, and twice with xylene. The initiator coated particles (BMPA MNPs) were kept suspended in xylene.

2.4. Surface initiated atom transfer radical polymerization

The core-shell nanoparticles were prepared by adapting a the method reported by Li et al.^[36] The BMPA MNPs suspended in xylene were mixed with the catalyst mixture. The amount of catalyst used was determined based on a styrene ratio of: 70:1.1 for DNDP, 70:0.3 for CuBr and 70:0.015 for CUBr₂. The solution containing the BMPA MNPs and the catalyst had a total volume of 120 mL. This solution was placed in a 3-neck flask, under nitrogen bubbling, 325 rpm, and heated to 135°C. The crosslinker, in this case the acrylated polyphenol (CMA or QMA), or DVB, was dissolved/mixed in 15 mL of xylene and injected into the reaction vessel at 110°C. CMA and QMA were synthesized following the protocol described by Patil et al ^[28,29] and Gupta et al ^[30], respectively. Two different crosslinker feed were studied: 5 mol% and 10 mol%. The reaction was carried out for a total of 24 hours. Samples of 25 mL were drawn out at 2, 6, 12 and 24 hours using a stainless steel syringe. Each sample collected was transferred into a 30 mL borosilicate amber glass vial, magnetically decanted and washed twice with xylene, twice with acetone, three times with a 50-50 % (v/v) ACN/DCM solution, and twice with a 50-50 % (v/v) ethanol/DI water solution. Finally, the particles were re-suspended in DI water.

2.5. Particle characterization

A Varian Inc. 7000e spectrometer with attenuated total reflectance FTIR (ATR-FTIR) was used to determine the surface functionalization of the sore shell nanoparticles. Dried samples were placed on the diamond ATR crystal and the spectrum was obtained between 700 and 4000 cm⁻¹ using 32 scans.

2.5.1. Thermogravimetric analysis

A Netzsch Instruments STA 449A system was used to quantify the mass percent of the coating on the nanoparticle systems. Approximately 5 mg of the dry sample was heated at a rate of 5°C per minute until a temperature of 120°C under constant nitrogen flow. The system was kept isothermal for 20 min to vaporize residual solvent and water vapors. The sample continued to be heated at 5°C per minute until a temperature of 600°C. The presented mass loss values are normalized to the mass after isothermal heating at 120°C.

2.5.2. X-ray diffraction

A Siemens D-500 X-ray spectrometer was used to determine the X-ray patterns of the nanoparticles using a with a CuKα radiation source (λ = 1.54 Ǻ) at 40 kV and 30 mA, using scanning speed of 1° per minute from 5° to 65°. The XRD patterns were used to estimate the particle’s crystal domain using the Scherrer equation:^[37]

$$\tau =\frac{K\lambda }{\beta cos\Theta }$$ (1)

where τ is the mean size of the ordered, crystalline domains, K is a dimensionless shape factor with a value close to unity (for iron oxide, K = 0.8396), λ is the X-ray wavelength, β is the line broadening at half the maximum intensity (FWHM) after subtracting the instrumental line broadening, and θ is the Bragg angle, in radians (17.72°). The XRD patterns were also used to confirm the magnetic crystal structure of the iron oxide nanoparticles.

2.5.3. Dynamic light scattering

A Malvern Zetasizer, Nano ZS90 instrument was used to obtain DLS measurements. Before analysis, the nanoparticle solutions were diluted to 200 μg/mL in DI water and probe sonicated for 10 minutes.

2.5.4. Ultraviolet (UV)-visible spectroscopy

A Cary Win 50 probe UV-visible spectrophotometer was used to study the stability of the nanoparticles. The magnetic nanoparticles were diluted to 200 μg/mL in DI water, and probe sonicated for 10 min. The samples were then placed in a quartz cuvette and their change in absorbance was read at 540 nm for a period of 12 h.

2.5.5. Alternating magnetic field (AMF) heating

Using a custom Taylor Winfield magnetic induction source the heating profiles of the nanoparticles were obtained. The temperature change in solution was recorded using a fiber optic temperature sensor (Luxtron FOT Lab Kit from LumaSense). A sample of 1.5 mL of the nanoparticles suspended in DI water at a concentration of 3 mg mL⁻¹ of iron oxide was placed in a microcentrifuge tube inside and in the center of the AMF induction coil. The alternating magnetic field source was operated at a field amplitude of approximately 55 kA m⁻¹ and a frequency of 300 kHz for 5 minutes. The specific absorption rate (SAR) values of the nanoparticles was calculated using the following equation:

$$SAR=\frac{{}^{C_{p,Fe}{m}_{Fe}+C_{p,H_{2}O}{m}_{H_{2}O}}dT}{m_{Fe}dt}$$ (2)

where C_p,Fe is the heating capacity of iron, m_Fe is the mass of iron, C_p,H2O is the heating capacity of iron, m _H2O is the mass of water, and dT (dt)⁻¹ is the initial slope of the heating profile of the system.

2.5.6. PCB 126 binding studies

In order to determine the binding capacity of the core-shell nanoparticles to PCB 126, equilibrium binding studies were conducted. The process followed has been described by our group in previous publications.^{[39, 40]} Briefly, 0.1 mg mL⁻¹ of the core-shell MNP systems suspended in a 99:1 DI water ethanol solvent were added in 3 mL borosilicate glass vials. Each sample was spiked with a known concentration of PCB 126 and sonicated for 10 minutes. PCB stocks were freshly prepared in ethanol to obtain 6 initial concentration ranging from 0.005 ppm to 1 ppm. The samples were placed in an orbital shaker (200 rpm, 25°C) for the duration of the study. Once the study finalized, the samples were magnetically decanted for ~ 10 min. The supernatant was collected and placed in a new vial for to extract the free PCB in solution using isooctane. After 24 h the organic phase, rich in PCB 126, was collected and placed in a gas chromatography vial. Here each sample was spiked with the internal standard, 5’-fluoro-3,3’,4,4’,5-pentachlorobiphenyl (F-PCB 126). All PCB 126 concentrations before and after the binding study were determined using an Agilent 6890N gas chromatography coupled to electron capture detection (GC-ECD), equipped with an Agilent HP-5MS UI column (30x0.25x0.25).

The MNP systems used for this studied were: styrene curcumin multiacrylate magnetic nanoparticles with 5 mol% initial acrylate loading at ATRP reaction times of 6h, 12h and 24h (Sty CMA MNPs_5%_6h, Sty CMA MNPs_5%_12h, Sty CMA MNPs_5%_24h); curcumin multiacrylate magnetic nanoparticles with 10 mol% initial acrylate loading at ATRP reaction time of 24h (Sty CMA MNPs_10%_24h); styrene quercetin multiacrylate magnetic nanoparticles with 5 mol% initial acrylate loading at ATRP reaction times of 6h, 12h and 24h (Sty QMA MNPs_5%_6h, Sty QMA MNPs_5%_12h, Sty QMA MNPs_5%_24h); quercetin multiacrylate magnetic nanoparticles with 10 mol% initial acrylate loading at ATRP reaction time of 24h (Sty QMA MNPs_10%_24h); and styrene divinylbenzene magnetic nanoparticles with 5 mol% initial crosslinker loading at ATRP reaction times of 6h, 12h and 24h (Sty DVB MNPs_5%_6h, Sty DVB MNPs_5%_12h, Sty DVB MNPs_5%_24h).

The binding capacity of the nanoparticles was calculated using the following equation:

$$q_e=\frac{(C_0-C_e)V}{m}$$ (3)

where q_e is the equilibrium binding capacity (mg g^−, C₀ and C_e are the initial and equilibrium concentrations (mg L⁻¹), respectively, V in the total volume of the solution (L), and m is the mass of the adsorbent (g). The obtained data was fitted to the Langmuir isotherm model, as it is the most useful model to represent adsorption of polycyclic aromatic hydrocarbons from water onto adsorbents.^[41] The Langmuir model best represents monolayer adsorption on homo generous surfaces, where there is a set number of binding sites that are all energetically equivalent and no interactions between adsorbed molecules occurs.^[42] The Langmuir model is represented by the following equation:

$$q_e=\frac{B_{max}{K}_D{C}_e}{1+K_D{C}_e}$$ (4)

where qe (mg g⁻¹) represents the adsorption capacity at equilibrium, Ce (mg L⁻¹) is the equilibrium concentration of the adsorbate, K_D (L mg⁻¹) is the adsorption coefficient of the sorbent related to the energy of adsorption, and B_max (mgg⁻¹) is the maximum binding capacity of the sorbent.

3. Results and Discussion

Core-shell magnetic nanoparticles were successfully prepared via surface initiated atom transfer radical polymerization. The MNP synthesis can be broken down into 3 mains steps: the preparation of the IO MNPS, the functionalization of the IO MNPs surface with an anchoring group for ATRP, in this case BMPA, and finally the SI-ATRP occurs under inert atmosphere. A schematic representation of this process and obtained core-shell nanoparticles is depicted in Figure 1 FTIR analysis confirms the successful SI-ATRP reaction and the formation of a polymer shell on the IO MNPs. Figure 2a shows the spectra for the CMA containing MNPs at two different initial loadings, 5 mol% and 10 mol%, both after a complete 24h ATRP reaction time. The presence of three main peaks between 1800 cm⁻¹ and 1400 cm⁻¹ are attributed to the symmetric ring vibrations of the benzene rings present in CMA. The presence of a peak at 1100 cm⁻¹ in both Sty CMA MNPs can be attributed to the ether C-O stretching of CMA. Additionally, peaks are seen at approximately 1000 cm⁻¹ and 950 cm⁻¹ corresponding to the enol (C-O-C) peak and C-H benzoate vibrations of the aromatic rings. Because styrene also presents aromatic rings, the peaks between 1600 cm⁻¹ and 1400 cm⁻¹ also provide evidence for the presence of styrene on the IO MNP polymer shell. Similar results are observed in Figure 2b, where the acrylated polyphenol present in QMA. Again, the aromatic ring vibration of the benzene are observed by broader peaks between 1800 cm⁻¹ and 1400 cm⁻¹, as well as a peak at 1200 cm⁻¹ for ether stretching. Once again, the presence of the styrene on the Sty QMA MNPs can also be inferred from the aromatic peaks, and by the presence of small peaks around 830 cm⁻¹ corresponding to aromatic ring bending. Figure 3c shows the spectra for the core-shell MNPs made without polyphenols. In this case, divinylbenzene was used as the crosslinker. The presence of both monomers can be seen in the appearance of peaks corresponding to the C-H deformation vibrations of the benzene ring around 1000 - 800 cm⁻¹.

Figure 1.

Schematic representation of the synthesis of the core-shell magnetic nanoparticles (MNPs): a) The co-precipitation synthesis of iron oxide nanoparticles (IO MNPs), shown inside the red rectangle b) Surface functionalization of the IO MNPs with 2-bromo-2-methyl propionic acid (BMPA) to obtain BMPA MNPs, c) Atom transfer radical polymerization reaction with styrene (Sty) and curcumin multiacrylate (CMA) to obtain core-shell Sty-CMA MNPs.

Figure 2.

FTIR spectra of the synthesized magnetic nanoparticles. A) Sty CMA MNPs, B) Sty QMA MNPs and C) Sty DVB MNPs.

Figure 3.

Mass loss profile with increasing temperature of the synthesized core-shell magnetic nanoparticles at different ATRP reaction times, A) Sty CMA MNPs at 5% initial loading, B) Sty CMA MNPs at 10% initial loading, C) Sty QMA MNPs at 5% initial loading, D) Sty QMA MNPs at 10% initial loading, and E) Sty DVB MNPs at 5% initial loading.

To begin to study the effect of ATRP reaction time on the growth of the shell on the IO MNPs, the determination of the amount of polymer grown needed to be determined. Thermogravimetric analysis was used to burn off the polymer shell over a selected temperature range, leaving the IO MNPs as residue. The TGA curves for the MNP systems are presented in Figure 3. It can be seen that for all the MNPs the amount of polymer coating, or shell growth, increases as the ATRP reaction time increases. This was expected given that SI-ATRP has been shown to allow for precise control of polymer density, molecular weight, and shell thickness.^[1,43] For the 5 mol% of initial crosslinker/functional monomer initial loading (Figure 3a, c, e), the increase in polymer composition appears to be relatively the same for all three MNP systems where after 24h of reaction, the polymer shell represents close to 20% of the total mass. On Figure 3b and d the initial amount of functional monomer, CMA or QMA, was increased to 10 mol%. In both cases an increase in weight loss is seen in the thermogram, indicating a higher polymer composition in the resulting MNPs. After 24h, the polymer coating on the Sty CMA MNPs from 5 mol% to 10 mol% of initial loading has increased from 21.82% to 38.06%. Likewise, the polymer coating on the Sty QMA MNPs from 5 mol% to 10 mol% of initial loading has increased from 20.42% to 55.8%. In both cases the polymer mass has almost doubled, or in fact doubled, its mass compared to its counterpart at 5 mol%. This increase in polymer mass, when the initial functional monomer is of 10 mol%, continues to be seen at lower reactions times, but it is less pronounced the shorter the reaction time. For example, the Sty CMA MNPs at 5 mol% have a polymer growth of 6.12% at 2h, 10.42% at 6h and 12.22% at 12 h, compared to the Sty CMA MNPs at 10 mol% with a polymer growth of 9.36% at 2h, 16.8% at 6h and 38.06% at 12 h. This increase in polymer shell growth with increasing functional monomer loading has been observed by other groups, where the thickness obtained through SI-ATRP was dependent on the molecular weight of the monomer, and the amount of monomer present in solution.^[43-45]

In order to verify the magnetic iron oxide nanoparticle core remained unchanged throughout the synthesis process, X-ray diffraction was performed. The XRD patterns for the prepared MNP systems, seen in Figure 4, are in agreement with the JCPDS card (19-0629) associated with magnetite. Furthermore, these XRD patterns present broad diffraction lines suggesting the nano-crystallite nature of the magnetite particles.^[41,45] The sharp peaks observed in the diffractograms indicate the formation of a crystalline structure, where the highest peak observed at 35.5° (2θ) corresponds to the (3 1 1) reflection plane of the iron oxide crystalline structure This information can be used in conjunction with the Scherrer equation to calculate the crystallite size of the core-shell MNPs. The iron oxide crystal size obtained from MNP for each system can be seen in Table 1.

Figure 4.

XRD patterns of the synthesized core-shell magnetic nanoparticles. Iron oxide nanoparticle XRD pattern included for reference.

Table 1.

Size analysis from XRD diffractograms using the Scherrer equation; hydrodynamic size analysis via dynamic light scattering of the synthesized core-shell MNPs (mean ± std dev. for three independent batches and three samples from each batch); and SAR values from AMF heating.

MNP System	Hydrodynamic size (nm) [PDI]				XRD crystal size (nm)	SAR (W mgFe−1)
Reaction time	2h	4h	12h	24h
Sty CMA MNPs_5%	173.9 ± 6.3 [0.06]	262.9 ± 5.5 [0.1]	356.9 ± 19.5 [0.1]	479.5 ± 33.6 [0.11]	12.5 ± 0.7	295.5 ± 10.5
Sty CMA MNPs_10%	340.0 ± 0.7 [0.07]	440.0 ± 0.95 [0.1]	503.9 ± 12.2 [0.11]	969.8 ± 29.0 [0.1]	10.2 ± 0.9	148.7 ± 16.8
Sty QMA MNPs_5%	164.3 ± 2.7 [0.05]	176.5 ± 3.8 [0.1]	210.01 ± 5.4 [0.1]	474.6 ± 32.7 [0.1]	11.4 ± 1.0	297.8 ± 24.2
Sty QMA MNPs_10%	260.5 ± 5.7 [0.04]	279.0 ± 5.5 [0.1]	420.1 ± 3.6 [0.11]	1558.7 ± 36.8 [0.12]	9.8 ± 0.8	93.4 ± 9.8
Sty DVB MNPs_5%	262.6 ± 1.0 [0.06]	278.2 ± 4.6 [0.1]	272.4 ± 7.3 [0.1]	285.5 ± 3.2 [0.1]	10.6 ± 1.1	309.5 ± 22.9

The hydrodynamic size of the core-shell magnetic nanoparticles was determined using dynamic light scattering (DLS) and reported as Z-average, with the variability in particle size within the batches being quantified by the polydispersity index (PDI), as presented in Table 1. An increase in hydrodynamic size is observed for all MNP systems as the ATRP reaction time increases. Although it is know the nanoparticles aggregate in solution, the significant size increase observed in most cases would indicate an increase in the MNPs size as well. These results are in agreement with what was observed for the TGA results, where the increase in ATRP reaction time in fact increase the amount of polymer shell growth, hence increasing the size of the MNP system.

The stability of the core-shell MNPs in aqueous environment becomes an important factor for their application as nanoadsorbents. In order to maximize the binding capacity, it is ideal for the particles to remain suspended in solution so that all their surface area is available to interact with the contaminant of interest. In order to evaluate this, the MNP systems were suspended in DI water their change in absorbance was recorded for 12h (Figure 5). It can be seen that in Figure 5 a - d, the ATRP reaction time increases the MNP system becomes less stable in solution. This behavior can be explained by the increasing hydrophobicity of the MNPs as the polymer shell increases with increasing ATRP reaction time. Moreover, these systems show increasingly bigger hydrodynamic sizes as time progresses, suggesting aggregation is also occurring and most likely becoming a factor that pulls the MNPs out of solution. Accordingly, the use of mechanical agitation is necessary to make sure the MNPs remain suspended for the duration of the binding studies, and for their ultimate application as nanoadsorbents in water remediation.

Figure 5.

Normalized absorbance ( at 540 nm) of the core-shell MNPs in DI water for 12 hours using UV-visible spectroscopy.

Figure 5e shows the stability for the Sty DVB MNPs with initial 5 mol% loading. Here the stability of the MNPs does not seem to be affected by the increasing ATRP reaction time. Even though TGA data has confirmed the growth of a polymeric shell over time, DLS data suggests these particles are more stable in DI water and do not appear to aggregate as much, which would explain why they remain stable over a period of 12h.

A unique property of IO MNPs is their ability to generate heat upon exposure to an alternating magnetic field (AMF). This heat dissipation can be used as a regeneration mechanism of the spent sorbent after a binding cycle.^[46,47] However, the thickness of the polymeric shell coating the IO MNPs can negatively affect their heat dissipation ability. Thus the MNP systems in solution were exposed to and AMF for 5 minutes in order to obtain their heating profile and determine their specific absorption rate (SAR) values. The SAR values are reported in Table 1 and indicate the energy being produced per gram of iron oxide. Thought the SAR values for the core-shell MNPs vary significantly between them, all MNP systems are still able to generate localized heat upon exposure to an AMF.

The binding capacity of the core-shell MNP systems for PCB 126 was studied at equilibrium conditions. The loading of the MNP systems utilized was of 0.1 mg mL⁻¹ in a 99:1 DI water ethanol solvent, and six different PCB 126 concentration were used. The MNP systems studied were: Sty CMA MNPs_5%_6h, Sty CMA MNPs_5%_12h, Sty CMA MNPs_5%_24h, Sty CMA MNPs_10%_24h, (The CMA MNP systems); Sty QMA MNPs_5%_6h, Sty QMA MNPs_5%_12h, Sty QMA MNPs_5%_24h, Sty QMA MNPs_10%_24h, (The QMA MNP systems); and Sty DVB MNPs_5%_6h, Sty DVB MNPs_5%_12h, Sty DVB MNPs_5%_24h, (The DVB MNP systems). The adsorption isotherms for all the studied MNP systems based on total nanoparticle mass are presented in Figure 6.

Figure 6.

Adsorption isotherms for PCB 126 of the core-shell MNP systems in terms of total mass at room temperature. A) Sty CMA MNP systems, b) Sty QMA MNPs systems and c) Sty DVB MNPs systems. PCB 126 initial concentrations from 0.005 – 0.1 ppm fitted using the Langmuir model.

The adsorption isotherms for the CMA MNP systems studies are presented in Figure 6a. It is seen that for all the systems the amount of PCB 126 bound increases as the free concentration of PCB increased until reaching a plateau, or maximum binding capacity, at different values. The same behavior is seen in for the QMA MNP systems in Figure 6b. The DVB MNP systems behave in a similar manner, but the initial increase in the amount of PCB 126 bound as the free concentration of PCB increases has a lower slope. To better understand the adsorption phenomenon, the experimental data is fitted to the Langmuir model to obtain the maximum adsorption capacity (Bmax) and Langmuir adsorption coefficients (K_D) for each system (presented in Table 2). The binding isotherms for the CMA MNP systems and the QMA MNP systems (Figure 6a and b, respectively) show higher binding at all free PCB concentrations in comparison to the DVB MNP systems. Previous works have demonstrated the importance of π-π interactions at the aromatic surface in the sorption of hydrophobic organic chemicals, such as PCBs, to aromatic-carbon based materials.^[48-50] In Addition, PCB 126 is a planar molecule which allows it to closely approach the approach the sorption sites of the adsorbent material and form favorable π-cloud interaction between the aromatic groups in the adsorbent and the PCB aromatic rings.^[51,52] Given the additional aromatic groups present in CMA and QMA in comparison to DVB, the binding isotherms indicate that the presence of the acrylated polyphenol groups enhance the binding of PCB 126.

Table 2.

Langmuir binding constants for the binding isotherm of PCB 126 for the core-shell MNP systems synthesized in terms of total mass (n = 9 _{independent samples}). Confidence Intervals obtained from nonlinear regression using GraphPsd Prism

MNP system	Bmax (mg g−1)	95 % CI	Kd (nM)	95 % CI	R2
Sty CMA MNPs_5%_6h	172.8	167.6 to 178.6	2.13	2.03 to 2.24	0.998
Sty CMA MNPs_5%_12h	211.1	204.9 to 218.0	0.77	0.72 to 0.82	0.998
Sty CMA MNPs_5%_24h	223.7	211.0 to 240.7	0.27	0.25 to 0.30	0.992
Sty CMA MNPs_10%_24h	201.7	197.5 to 206.3	0.61	0.60 to 0.62	0.998
Sty QMA MNPs_5%_6h	138.3	131.4 to 146.9	2.00	1.87 to 2.16	0.978
Sty QMA MNPs_5%_12h	257.2	240.2 to 279.0	1.59	1.38 to 1.88	0.998
Sty QMA MNPs_5%_24h	207.6	206.0 to 209.3	0.19	0.18 to 0.20	0.994
Sty QMA MNPs_10%_24h	204.6	203.4 to 205.8	0.63	0.62 to 0.63	0.997
Sty DVB MNPs_5%_6h	145.1	137.4 to 154.8	5.03	4.58 to 5.63	0.997
Sty DVB MNPs_5%_12h	167.9	152.7 to 190.2	4.77	4.04 to 5.99	0.992
Sty DVB MNPs_5%_24h	155.2	147.3 to 165.0	2.99	2.74 to 3.34	0.995
Sty DVB MNPs_10%_24h	265.3	227.9 to 330.5	5.21	4.08 to 7.50	0.994

The maximum binding capacity for all the core-shell MNP systems can be seen in Table 2. The variation in the maximum capacity of the Sty DVB MNPs, with 5 mol% of initial crosslinker loading, remains relatively constant as the ATRP reaction increases, as seen from the confidence intervals. This behavior is the same for both the Sty CMA MNPs and the QMA MNPs, with 5 mol% of initial functional monomer loading. Focusing on the 5% initial loading after 24h of ATRP reaction, the confidence intervals obtained from a nonlinear regression in GraphPad Prism, indicate that the value for B_max for the Sty DVB MNPs (155.2 mg g⁻¹) is significantly lower than that for the Sty QMA MNPs (207.6 mg g⁻¹) and Sty CMA MNPs (223.7 mg g⁻¹). Again, this result suggests that the presence of the additional aromatic moieties in both the CMA and QMA allow for greater binding sites based on formation of π-π interactions at the surface with the PCB molecules. Maximum capacities for engineered magnetic nanomaterials and plastics have been reported for use in adsorption of organic pollutants in agreement with the values shown in Table 2.^[20,50-55] Likewise, the obtained values for B_max are also similar to those previously reported for some carbon-based materials, yet still a couple orders of magnitude lower than activated carbon..^[53,56-58] Once the initial functional monomer molar% increased from 5% to 10%, there was no significant change on the binding capacity of Sty CMA MNPs or Sty QMA MNPs, meaning that at the conditions studied, there is no significant effect on the binding capacity of the system.

The Langmuir adsorption coefficients obtained for the Sty CMA MNPs and Sty QMA MNPs range from 0.19 nM to 2.19 nM, all which are smaller than those obtained for the Sty DVB MNPs (5.03 nM at 2 hour, 4.77 nM at 6h and 2.99 nM at 24 h). These smaller K_D values indicate greater binding affinities of the CMA and QMA systems for PCB 126. Once again, the core-shell systems containing the acrylated polyphenol moieties are shown to enhance the binding affinity for PCB 126. This difference in affinity could be based on structural differences of the polymer shell formed with the DVB versus the CMA/QMA functional monomers. The accepted structure-binding relationships for PCBs in protein and antibodies has been explained as a docking mechanisms which is can be highly selective.^[59] The presence of different side groups around the docking site have the ability to allow or impede the binding to occur. Extrapolating this to the core-shell MNPs, the CMA and QMA contain other functional groups within their molecular structure that could be aiding in the creation of better or higher affinity binding sites for PCB 126. Still, it is important to highlight here that the Langmuir adsorption coefficients obtained for all the synthesized MNPs are in the same order of magnitude as the binding affinity of the monoclonal antibody S2B1 presents for PCB126 (2.5 ± 0.01 nM), which demonstrates the high affinity the core-shell MNPs possess for this contaminant.^[60] Moreover, these K_D values are all lower than reported values in literature for the adsorption of PCB 126 by activated carbon (6.12 nM), the most used adsorbent in water remediation for non-specific adsorption of organic contaminants, and micron sized charcoal (15.2 nM), another commonly used material for environmental remediation.^[56,61,62]

Looking more closely at the effect of reaction time on B_max it is seen that for all the systems (Sty CMA MNPs, Sty QMA MNPs and Sty DVB MNPs) there is an increase in the maximum binding capacity of the systems with increasing reaction time. As reaction time increases, so does the growth of the polymer shell on the nanoparticle, resulting in particles having a greater fraction of their mass being the polymer coating, which in turn leads to a higher binding capacity. However, as the ATRP reaction time increases, it appears that the binding affinity for PCB 126 also increases (lower K_D values). Since we expect the composition of the polymer coating to not change significantly as the reaction process occurs, the observed increase in K_D with increasing reaction time for each system could be an artifact of the model fit where the total mass of the system was used to normalizing the data. To examine this further, the binding isotherms were also analyzed on a per polymer shell mass basis in the following.

Figure 7 shows the binding results normalized to the polymer mass of each system as reaction time increases. By normalizing the binding data to polymer mass, it is seen that the binding isotherms for the nanoparticle systems significantly collapse onto each other. By normalizing to the polymer shell mass, it can be seen that the maximum binding capacity of each nanoparticle system becomes more similar to each other. In Figure 7c, can be seen that for the Sty DVB MNPs at the different reaction times the Langmuir curves in fact collapse onto each other and the B_max for each reaction time falls within the confidence intervals of each other, as seen in Table 3, meaning there is no significant difference in their values. The same effect on K_D values is observed, where they all are in error of each other (as seen in Table 3 in the confidence intervals). In this specific case, the Sty DVB MNPs appear to bind less than the Sty CMA MNPs and Sty QMA MNPs, as observed in the values of free PCB in solution on Figure 7. Due to their lower affinity, the binding isotherm data spans a larger range of free concentrations resulting in the Langmuir model being able to fit a larger range of data than in the CMA and QMA systems and thus representing the system with a high level of confidence. In contrast, some of the data in the Sty CMA MNPs and Sty QMA MNPs has a much smaller range for the concentration of free PCB in solution, resulting in the Langmuir model fit and prediction of B_max and K_D resulting from a limited range of concentrations, which might not accurately represent the system’s behavior. For both, Sty CMA MNPs and Sty QMA MNPs, the amount of PCB bound increases as reaction time increases, reducing the range of free PCB in solution and further impacting the accuracy of the model.

Figure 7.

Adsorption isotherms for PCB 126 of the core-shell MNP systems in terms of polymer mass at room temperature. A) Sty CMA MNP systems, b) Sty QMA MNPs systems and c) Sty DVB MNPs systems. PCB 126 initial concentrations from 0.005 – 0.1 ppm fitted using the Langmuir model.

Table 3.

Langmuir binding constants for the binding isotherm of PCB 126 for the core-shell MNP systems synthesized in terms of polymer mass (n = 9 _{independent samples}). Confidence Intervals obtained from nonlinear regression using GraphPad Prism

MNP system	Bmax (mg g−1)	95 % CI	Kd (nM)	95 % CI	R2
Sty CMA MNPs_5%_6h	1902	1802 to 2025	1.67	1.52 to 1.88	0.996
Sty CMA MNPs_5%_12h	1728	1651 to 1818	0.77	1.70 to 0.85	0.998
Sty CMA MNPs_5%_24h	1048	941 to 1233	0.27	0.28 to 0.41	0.987
Sty CMA MNPs_10%_24h	880	811 to 982	0.69	0.64 to 0.78	0.994
Sty QMA MNPs_5%_6h	1438	1397 to 1484	1.31	1.24 to 1.39	0.997
Sty QMA MNPs_5%_12h	1471	1371 to 1587	1.33	1.18 to 1.53	0.999
Sty QMA MNPs_5%_24h	932	888 to 984	0.63	0.61 to 0.67	0.996
Sty QMA MNPs_10%_24h	580	556 to 608	0.62	0.59 to 0.65	0.997
Sty DVB MNPs_5%_6h	1170	1099 to 1264	3.97	3.56 to 4.56	0.997
Sty DVB MNPs_5%_12h	1169	1113 to 1422	4.01	3.54 to 4.57	0.991
Sty DVB MNPs_5%_24h	967	899 to 1063	3.03	2.67 to 3.58	0.996
Sty DVB MNPs_10%_24h	1485	1236 to 1947	3.91	3.18 to 4.85	0.998

To further examine the binding isotherms, the data was normalized to nanoparticle surface area. The surface area of each systems at the different reaction times was calculated assuming a perfect sphere and additive densities of the nanoparticle components (iron oxide nanoparticle core and polymer shell constituents – Sty, DVB, CMA, and QMA). Figure S1 presents the Langmuir isotherms for the styrene nanoparticle systems based on surface area. Here, the curves for each nanoparticle system are visibly different from each other, suggesting that as reaction time increases so does the amount of PCB bound per surface area. This apparent increase in affinity and capacity suggest that the adsorption of PCB 126 to the styrene-based nanoparticles is not just an effect of surface area, given that the total surface area of the particles (seen in Table S1) decreases with increasing reaction time. Again, the phenomenon occurring during the binding studies appears to consist of more than just surface interactions between the nanoparticles and the PCB 126 molecules.

Our group recently developed nanoadsorbent materials containing these functional acrylated monomers, CMA and QMA, as part of a core-shell structure to be used in environmental remediation.^[38] These core-shell MNPs were developed in a similar manner as to those described in this paper, with the exception of the use poly(ethylene glycol) 400 dimethacrylate (PEG400DMA) as part of the polymer shell, instead of styrene. The CMA and QMA containing MNP systems were synthesized for a period of 24h and an initial loading of 10 mol%, resulting in magnetic core-shell nanoparticles of uniform distribution with a polymer shell of roughly 10% of the total weight. Three systems were produced: CMA MNPs, QMA MNPs, and PEG MNPs (where the shell consisted of only a PEG400DMA). The binding capacity of these MNPs was also evaluated for PCB 126 under equilibrium conditions. The values obtained for the maximum binding capacity for the CMA MNPs and QMA MNPs was of 1.06 mg g⁻¹, and of 1.91 mg g⁻¹for the PEG MNPs. Comparing these values to those presented in Table 2 for the styrene based MNPs on total mass, it becomes clear that by using a hydrophobic monomer like styrene in place of PEG400DMA, the maximum binding capacity of the core-shell MNPs was drastically increased. Styrene is an organic monomer that can produce polystyrene polymers with a hydrophobic surface and high surface area per gram of material when crosslinked with other hydrophobic molecules like DVB or, in this case CMA or QMA.^{[24,25,50,63]} These styrene based polymers have shown to be particularly useful for the adsorption of molecules with aromatic rings because of the strong π-π interactions they can have. The Langmuir adsorption coefficients obtained for the CMA MNPs, QMA MNPs and PEG MNPs are 2.72 nM, 5.88 nM, and 8.42 nM, respectively. For the Styrene based systems synthesized under the same conditions (5 mol% initial loading and 24 h ATRP reaction), the K_D values are 0.27 nM, 0.19 nM and 2.99 nM for the Sty CMA MNPs, Sty QMA MNPs and Sty DVB MNPs respectively. Again, the values obtained indicate higher binding affinity for the styrene based MNPs compared to the PEG based MNPs further demonstrating the importance of the polymer shell composition for the targeting of PCB 126 removal. The use of a hydrophobic, aromatic rich molecule as styrene within the polymeric shell provides more sites for π-π interactions to occur between the MNP surface and the PCB molecules. This in turn, increases the maximum binding capacity of the MNP system for PCB 126. These results provide a strong rational for the use of our magnetic core-shell nanoparticle systems to be used as high affinity adsorbents in the environmental remediation of specific harmful contaminants.

4. Conclusions

This study reports the successful synthesis of core-shell magnetic nanoparticles (MNPs) using surface initiated atom transfer radical polymerization (SI-ATRP). Herein, the magnetic core is comprised of iron oxide nanoparticles which endow the MNP systems with magnetic decantation capabilities. The polymeric shell is composed by styrene and a crosslinker. Three different crosslinkers were used, all containing additional aromatic ring moieties to enhance pollutant binding capacity. Two of them were acrylated plant derived polyphenols, curcumin multiacrylate (CMA) and quercetin multiacrylate (QMA), and the third was divinylbenzene (DVB). The effect of ATRP reaction time was studied on the properties of the MNPs. Equilibrium binding studies were conducted at six different PCB 126 concentration, and binding isotherms were obtained. The Langmuir model was used to obtain the binding coefficients and the maximum binding capacity of the core-shell MNPs. The binding isotherms obtained showed that the CMA and QMA containing MNPs presented higher binding affinities and capacities. Despite this difference, all MNPs have higher binding affinities for PCB 126 that carbonaceous materials, like activated carbon and graphene oxide, the most widely used adsorption materials for water remediation today. And the binding affinities for all the Sty CMA MNPs and Sty QMA MNPs were similar to those observed for antibodies. The increase in ATRP reaction time increases the binding capacity of the MNPs given that as the polymer shell grows so does the available sites for π-π interaction to occur with the PCB molecules. The effect of increasing ATRP reaction time on the binding affinity and capacity of the MNPs for PCB 126 was further examined, and specifically, the data was analyzed for different normalization factors (total mass, polymer shell mas and surface area) to fit the Langmuir model. These results suggest the phenomenon occurring during the binding studies is not limited to a surface interaction between the nanoparticles and the PCB 126 molecules in solution. Finally, the importance of the polymeric shell composition was demonstrated by comparing the Langmuir coefficients obtained in this work to previous work done by our group with similar materials. It was seen that the use of a hydrophobic, aromatic rich molecule like styrene within the polymeric shell provides more sites for π-π interactions to occur between the MNP surface and the PCB molecules, increasing the binding capacity almost 200 fold in some cases and increasing the binding affinity of the MNPs as well. Overall, we have developed magnetic core-shell nanoparticle systems with high affinities for PCBs in aqueous media with tunable shell thickness for optimal affinity, that can be magnetically decanted from solution with the use of a static magnetic field, and has the potential to be regenerated upon the exposure to an alternating magnetic field, for their use as nanoadsorbents in the environmental remediation of specific harmful contaminants.

Highlights:

• Styrene based core-shell magnetic nanoparticles were synthesized via surface initiated atom transfer radical polymerization for use as nanoadsorbents in the environmental remediation of specific harmful contaminants.

• Surface initiated atom transfer radical polymerization was used to tune the shell thickness of the core-shell nanoparticles for optimal affinity

• The magnetic core-shell nanoparticle systems have high affinities for PCBs in aqueous media

Data Availability

The raw/processed data required to reproduce these findings cannot be shared at this time due to technical or time limitations.