Magnetizing Polymer Particles with a Solvent-Free Single Stage Process Using Superparamagnetic Iron Oxide Nanoparticles (SPION)s

Magnetic polymer composites are used in a variety of applications in many industries. Their production methods are usually time-consuming and solvent-intensive as they are performed in liquid phase processes, such as emulsion polymerization or precipitation. In this work, a quick, easy, and solvent-free method is presented to coat polymer particles with a discrete, non-coherent coating of superparamagnetic nanoparticles. The results of the dry coating process are evaluated optically, by means of scanning electron microscopy (SEM), via powder X-ray diffraction and thermally by means of differential scanning calorimetry, before finally demonstrating the effectiveness of dry coating by means of a vibrating sample magnetometer.


Introduction
Magnetic polymer composites are being investigated and used for a wide variety of applications in the medical technology, for example magnetic resonance spectroscopy [1,2], pharmaceutical [3][4][5][6][7], and (bio-)chemical industries [8][9][10] as gas and humidity sensors [11,12] or for urea sensing [13]. The sensing abilities of iron oxides are always of interest, as recently iron oxide-based supra-particles were used to track the temperature [14] or to create a magnetic fingerprint [15]. Therefore, the functionalization and production of magnetic polymers has been the topic of research for several years. A number of processes have been developed, all taking place in the liquid phase, where the polymerization occurs in a liquid or emulsion containing the magnetic particles [16][17][18][19][20][21] as well as the deposition of nanoscale iron oxid within the liquid phase [22]. This not only entails the use of various solvents, which can be problematic for both humans and the environment, but this process route is always followed by a drying step, which also takes time and resources.
To create a magnetizable polymer composite, various substances are currently used; most of them are iron-based, since iron has the highest saturation magnetization at ambient temperatures [23]. Iron oxides, especially Fe 3 O 4 [21,[24][25][26][27][28][29][30][31][32] and γ-γ-Fe 2 O 3 , are the most popular. The benefits of iron oxides in comparison with other iron compounds are their stability, their environmental friendliness, and their inexpensiveness [33], which makes them easy-to-use and to prepare [34][35][36][37][38][39]. Their combination with polymers can yield completely new functionalities such as the recently reported magnetic indication of elapsed temperature events [33]. Another large group of magnetic particles with which the polymer is modified during polymerization are the carbonyl iron compounds such as shown in [40][41][42][43][44][45][46][47][48][49]. As well as some rarer compounds containing zinc [50], copper [51], or cobalt [52]. For an ideal random single-layer coating, as in Figure 1 (e-1), Yang et al. [53] gave the approximate additive content in weight percent: • 100 (1) For an ideal random single-layer coating, as in Figure 1 (e-1), Yang et al. [53] gave the approximate additive content in weight percent: with the diameter of the host particles d host and the guest particle diameter d g as well as the density of the host particles ρ host and the guest particles ρ g , as well as N (cf. Equation (2)), which represents the number of guest particles to achieve the surface coverage.
The dry coating process will therefore form a discrete core-shell coating around on the particle surface. When the process is optimized a uniformly distributed coating, which is near Figure 1 (e-3), is achievable. The difficulties in dry coating are primarily the de-agglomeration of the nanoparticulate GP, selecting the process time so that decoating does not occur, and observing the yield, since triboelectric effects have a massive influence on this.
The aim of this work is therefore to showcase a sustainable, solvent-free, scalable, and easy-to-use method to create discrete magnetic core-shell structures on polypropylene (PP) particle surfaces by use of dry particle coating. PP is a widely used thermoplast with high chemical resistance and good thermal properties which is used in additive manufacturing or injection moulding. A magnetisable PP-composite would, for example, be perfect to produce magnetically responsive sensors. The composite is analyzed by use of scanning electron microscopy to showcase the coating and the differences in the two additives used, as well as to evaluate the degree of coverage on the polymer surface. Furthermore, the powder is analyzed by powder X-ray diffraction to evaluate polymorphs in the coating material and differential scanning calorimetry to determine the crystallinity of the polymer-super paramagnetic iron oxide nanoparticle (SPION) composite. Finally, the magnetizability of the newly made composites is determined and results are compared to each other.

Materials and Methods
Polypropylene, type Coathylene PD0580 (Axalta, Bulle, Switzerland) with a density of 0.907 g cm −3 and particle sizes of x 10.3 = 32 ± 3 µm, x 50.3 = 93 ± 5 µm and x 90.3 = 182 ± 4 µm (measured by laser diffraction (Mastersizer 2000, Malvern, UK)) is used. Magnetic γ-Fe 2 O 3 powder, 98%, (NanoArc ® , APS Powder, S.A, Singapore) with a particle diameter between 20-40 nm and a specific surface area (BET) of 32.188 m 2 g −1 is used as an additive and compared with self-synthesized Fe 3 O 4 , with a mean particle size between 10 to 20 nm and a specific surface area (BET) of 79.138 m 2 g −1 , stabilized with oleic acid according to [14]. The synthesis of Fe 3 O 4 is achieved by a co-precipitation reaction. Therefore, FeCl 3 6 H 2 O (total: 10.80 g, 40 mmol, Sigma Aldrich, Darmstadt, Germany, >99%) and FeCl 2 4 H 2 O (total: 3.98 g, 20 mmol, Fluka-obtained by Sigma Aldrich, Darmstadt, Germany, >99%) were dissolved in deionized water (225 mL) at room temperature (ca. 20 • C) and mixed with 30% aqueous ammonia solution NH 3 (aq.) (25 mL). After 60 s of stirring, the black precipitate was magnetically separated, and the overlaying water decanted. In this manner, the particles were washed with deionized water (250 mL) three times before dispersing them in water (250 mL). The self-synthesized nanoparticle dispersion is dried with a vacuum oven and pre-crushed before use via mortar and pestle. The commercial γ-Fe 2 O 3 nanoparticles as well as the self-synthesized Fe 3 O 4 nanoparticles are displayed in Figure 2. It is important to note, that the Fe 3 O 4 powder is difficult to deagglomerate and most Fe 3 O 4 particles are present in agglomerates larger than 100 nm. itate was magnetically separated, and the overlaying water decanted. In this manner, the particles were washed with deionized water (250 mL) three times before dispersing them in water (250 mL). The self-synthesized nanoparticle dispersion is dried with a vacuum oven and pre-crushed before use via mortar and pestle. The commercial γ-Fe2O3 nanoparticles as well as the self-synthesized Fe3O4 nanoparticles are displayed in Figure 2. It is important to note, that the Fe3O4 powder is difficult to deagglomerate and most Fe3O4 particles are present in agglomerates larger than 100 nm.

Dry Coating
A Turbula mixer (T2f, Willy A. Bachofen AG, Muttenz, Switzerland) is used for the coating experiments. The polymer coating takes place in borosilicate snap-on lid glasses with a volume of 25 mL. To ensure reproducibility, each experiment is performed 3 times. The mixer is operated in all experiments at 49 rpm with a coating time of 60 min.
In each experiment 3 g of polymer powder were used. To achieve deagglomeration of the guest particles, 9 g of glass spheres (Sili S, Sigmund Lindner GmbH, Warmensteinach, Germany) with particle size ranging between 1.0-1.25 mm and a bulk density of 1.5 g cm −1 were added as mixing aids. The glass beads were separated with a 0.8 mm sieve after the coating process.

Nitrogen Sorption
Nitrogen sorption measurements (77.4 K) were performed to determine the properties of the nanoparticles using a NovaTouch LXII (Anton Paar, Graz, Austria). Before each measurement, the samples were gassed out at 200 °C for 12 h at vacuum and weighted in an inert atmosphere.

Dry Coating
A Turbula mixer (T2f, Willy A. Bachofen AG, Muttenz, Switzerland) is used for the coating experiments. The polymer coating takes place in borosilicate snap-on lid glasses with a volume of 25 mL. To ensure reproducibility, each experiment is performed 3 times. The mixer is operated in all experiments at 49 rpm with a coating time of 60 min.
In each experiment 3 g of polymer powder were used. To achieve deagglomeration of the guest particles, 9 g of glass spheres (Sili S, Sigmund Lindner GmbH, Warmensteinach, Germany) with particle size ranging between 1.0-1.25 mm and a bulk density of 1.5 g cm −1 were added as mixing aids. The glass beads were separated with a 0.8 mm sieve after the coating process.

Nitrogen Sorption
Nitrogen sorption measurements (77.4 K) were performed to determine the properties of the nanoparticles using a NovaTouch LXII (Anton Paar, Graz, Austria). Before each measurement, the samples were gassed out at 200 • C for 12 h at vacuum and weighted in an inert atmosphere.

Scanning Electron Microscopy
The polymer particles have been characterized by scanning electron microscopy (SEM) using a Gemini Ultra 55 (Zeiss, Jena, Germany) device equipped with a SE2 detector. An acceleration voltage of 1 kV has been applied.

X-ray Diffraction
For structural analysis of the product, powder X-ray diffraction (XRD) was performed with an AXS D8 Advance diffractometer in the Bragg-Brentano geometry (Bruker, Billerica, MA, USA). The device is equipped with a VANTEC-1 detector and a Ni filter and uses Cu Kα radiation (154 pm). The step size for collecting the diffractograms was set to 0.014 • with a measuring time of 1 s per step for the range of 2Θ = 10-60 • .

Dynamic Scanning Calorimetry
The crystallinity and the crystallization temperature of the coated powders are determined by differential scanning calorimetry (DSC). For this purpose, a Polyma 214 (Netzsch, Selb, Germany) is used. The samples with a weight of 10 mg ± 0.1 mg are measured with covered aluminum pans type Concavus Lids (Al), NGB817526 (Netzsch, Germany) with dry nitrogen gas purging at 40 mL min −1 . As the melting temperature of the PP is at 186 • C, the temperature profile to measure the thermogram is as following: (1) Start at 20 • C, (2) heating to 200 • C by a gradient of 10 K min −1 , (3) isothermal step of 60 • s, (4) cooling by a gradient of 10 K min −1 .

Virbating Sample Magnetometer
Magnetic properties of the nanoparticles as well as the product particles were determined with a vibrating sample magnetometer (VSM) type VersaLabTM 3T from Quantum Design Inc., San Diego, CA, USA. Field dependent magnetization measurements were conducted between −30 to +30 kOe with a measurement speed of 50 Oe s −1 and 5 Oe s −1 between −5 and 5 kOe.

Degree of Coverage
The effect of dry coating can be observed visually in

Scanning Electron Microscopy
The polymer particles have been characterized by scanning electron microscopy (SEM) using a Gemini Ultra 55 (Zeiss, Jena, Germany) device equipped with a SE2 detector. An acceleration voltage of 1 kV has been applied.

X-ray Diffraction
For structural analysis of the product, powder X-ray diffraction (XRD) was performed with an AXS D8 Advance diffractometer in the Bragg-Brentano geometry (Bruker, Billerica, MA, USA). The device is equipped with a VANTEC-1 detector and a Ni filter and uses Cu Kα radiation (154 pm). The step size for collecting the diffractograms was set to 0.014° with a measuring time of 1 s per step for the range of 2Θ = 10-60°.

Dynamic Scanning Calorimetry
The crystallinity and the crystallization temperature of the coated powders are determined by differential scanning calorimetry (DSC). For this purpose, a Polyma 214 (Netzsch, Selb, Germany) is used. The samples with a weight of 10 mg ± 0.1 mg are measured with covered aluminum pans type Concavus Lids (Al), NGB817526 (Netzsch, Germany) with dry nitrogen gas purging at 40 mL min −1 . As the melting temperature of the PP is at 186 °C, the temperature profile to measure the thermogram is as following: (1) Start at 20 °C, (2) heating to 200 °C by a gradient of 10 K min −1 , (3) isothermal step of 60° s, (4) cooling by a gradient of 10 K min −1 .

Virbating Sample Magnetometer
Magnetic properties of the nanoparticles as well as the product particles were determined with a vibrating sample magnetometer (VSM) type VersaLabTM 3T from Quantum Design Inc., San Diego, CA, USA. Field dependent magnetization measurements were conducted between −30 to +30 kOe with a measurement speed of 50 Oe s −1 and 5 Oe s −1 between −5 and 5 kOe.

Degree of Coverage
The effect of dry coating can be observed visually in Figure 3. The surface coating noticeably changes the color of the powder to the color of the iron oxide used. Inside the red box (top row), the powders functionalized with Fe2O3 are displayed; in the blue box (bottom row) results of functionalization with Fe3O4 are displayed.  The evaluation of the particle surface by means of SEM shows, an increase in the degree of coverage with an increase of the SPION content. Figure 4 shows the change in the degree of coverage of Fe 2 O 3 (red box) and Fe 3 O 4 (blue box). The coating with Fe 3 O 4 is not as uniform as in the samples with γ-Fe 2 O 3 . The reason is found in the production process of the SPIONs by spray drying: The Fe 3 O 4 formulation has not been stabilized against aggregation of the primary particles while drying. Nevertheless, the images show a homogeneous, discrete distribution of the guest particles on the polymer surface. Due to the van der Waals forces acting between the GP and the HP, the two materials are almost inseparably bonded, which makes the process stand out. Due to the surface deposition of the GP on the HP, their effect, such as magnetization in this case, is not hindered by any polymer layers, as often occurs in liquid phase processes, where the GP preferentially accumulate inside a polymer matrix. In that case, the polymer would act as a spacer between the nano particles and the external magnetic field. Corresponding to the SEM-images from Figure 5, the degrees of coverage (see Figure 4) are determined. Image analysis via MATLAB, due to binarization of the SEM images by use of Otsu's method [65] to determine the threshold. After binarization, the guest particles are displayed in white and the polymer surface is displayed in black. The degree of coverage is the amount of white area within the image. It is easy to observe that the degree of coverage increases continuously and flattens around 2 wt.% additive content onwards. The lower degree of coverage of the Fe3O4 samples is also evident here. The higher degree Corresponding to the SEM-images from Figure 5, the degrees of coverage (see Figure 4) are determined. Image analysis via MATLAB, due to binarization of the SEM images by use of Otsu's method [65] to determine the threshold. After binarization, the guest particles are displayed in white and the polymer surface is displayed in black. The degree of coverage is the amount of white area within the image. It is easy to observe that the degree of coverage increases continuously and flattens around 2 wt.% additive content onwards. The lower degree of coverage of the Fe 3 O 4 samples is also evident here. The higher degree of coverage when using the γ-Fe 2 O 3 is related to the easier deagglomeration of the commercial product. The Fe 3 O 4 particles from our own production, were hydrophobized by the oleic acid but not functionalized against aggregation. Results indicate that the interparticulate forces between the Fe 3 O 4 primary particles are stronger than the forces needed to deagglomerate them sufficiently during the coating process, leading to a non-ideal coating result and even spread on the HP surface. the oleic acid but not functionalized against aggregation. Results indicate that the interparticulate forces between the Fe3O4 primary particles are stronger than the forces needed to deagglomerate them sufficiently during the coating process, leading to a nonideal coating result and even spread on the HP surface. To confirm the lower degree of coverage of Fe3O4, a thermo-gravimetrical experiment to determine the amount of SPIONs on the polymer powder has been carried out. For this purpose, around 1 g of polymer-iron oxide powder is weighed into a ceramic bowl. Afterwards, the polymer is burned at 900 °C for 10 min within an oven. Subsequently, the residual mass (the inorganic iron oxide) is weighed, and the measured iron oxide content is calculated. The results are shown in Figure 6. As displayed, the residual mass of Fe3O4 within the polymer-iron oxide composite is lower in comparison with the easily-dispersible commercial product.  To confirm the lower degree of coverage of Fe 3 O 4 , a thermo-gravimetrical experiment to determine the amount of SPIONs on the polymer powder has been carried out. For this purpose, around 1 g of polymer-iron oxide powder is weighed into a ceramic bowl. Afterwards, the polymer is burned at 900 • C for 10 min within an oven. Subsequently, the residual mass (the inorganic iron oxide) is weighed, and the measured iron oxide content is calculated. The results are shown in Figure 6. As displayed, the residual mass of Fe 3 O 4 within the polymer-iron oxide composite is lower in comparison with the easily-dispersible commercial product. the oleic acid but not functionalized against aggregation. Results indicate that the interparticulate forces between the Fe3O4 primary particles are stronger than the forces needed to deagglomerate them sufficiently during the coating process, leading to a nonideal coating result and even spread on the HP surface. To confirm the lower degree of coverage of Fe3O4, a thermo-gravimetrical experiment to determine the amount of SPIONs on the polymer powder has been carried out. For this purpose, around 1 g of polymer-iron oxide powder is weighed into a ceramic bowl. Afterwards, the polymer is burned at 900 °C for 10 min within an oven. Subsequently, the residual mass (the inorganic iron oxide) is weighed, and the measured iron oxide content is calculated. The results are shown in Figure 6. As displayed, the residual mass of Fe3O4 within the polymer-iron oxide composite is lower in comparison with the easily-dispersible commercial product.  It is not surprising that the measured iron oxide content is lower than the set content in this work, as losses always occur on the wall of the mixing device. For the purpose of this work and better comparability, further results are displayed in regard to the set iron oxide content.

X-ray Diffraction
The measurements are intensity min-max-normalized for all values obtained within a sample according to: In all diffractogramms (Figure 7), the increase in the reflexes, characteristic of the iron oxides used, can be clearly observed. As can be seen in the diffractograms from the reflexes 2Θ = 31 • , 2Θ = 38 • , 2Θ = 47 • , and 2Θ = 55 • , the iron oxide nanoparticles are not one pure polymorph, but exhibit both alpha (hematite) and gamma (maghemite) reflexes [66,67]. The reflexes correspond to those of the raw material used, which proves the gentle processing conditions during dry coating, as otherwise a change in the polymorphy [66] would have occurred. Although not as distinct, the same behavior can be seen in the polymer particles coated with Fe 3 O 4 . The polymorphy is detectable on the surface. The reason for this is that there is often a slight transformation on the nanoparticle surface where γ-Fe 3 O 4 converts to Fe 2 O 3 , but the Fe 3 O 4 is still present in the core [68]. It is not surprising that the measured iron oxide content is lower than the set content in this work, as losses always occur on the wall of the mixing device. For the purpose of this work and better comparability, further results are displayed in regard to the set iron oxide content.

X-ray Diffraction
The measurements are intensity min-max-normalized for all values obtained within a sample according to: In all diffractogramms (Figure 7), the increase in the reflexes, characteristic of the iron oxides used, can be clearly observed. As can be seen in the diffractograms from the reflexes 2Θ = 31°, 2Θ = 38°, 2Θ = 47°, and 2Θ = 55°, the iron oxide nanoparticles are not one pure polymorph, but exhibit both alpha (hematite) and gamma (maghemite) reflexes [66,67]. The reflexes correspond to those of the raw material used, which proves the gentle processing conditions during dry coating, as otherwise a change in the polymorphy [66] would have occurred. Although not as distinct, the same behavior can be seen in the polymer particles coated with Fe3O4. The polymorphy is detectable on the surface. The reason for this is that there is often a slight transformation on the nanoparticle surface where γ-Fe3O4 converts to Fe2O3, but the Fe3O4 is still present in the core [68].

Thermal Analysis
Thermal analysis by DSC shows the influence of the iron oxide nanoparticles on the crystallization temperature (Figure 8a) as well as the relative crystallinity (Figure 8b). Further processing, e.g., in additive manufacturing or injection moulding, strongly depends on the melting and crystallization behavior. Figure 8 shows the determined crystallization temperatures (Figure 8a) and the corresponding relative crystallinities (Figure 8b). The nanoparticles act as crystallization nuclei and, thus, shift the crystallization temperature as expected to higher temperatures [69][70][71]. This effect is further enhanced by the higher heat transfer ability of the iron oxides. The more rapid heat transfer also explains the reduced relative crystallinity, since colder temperatures are also passed.

Thermal Analysis
Thermal analysis by DSC shows the influence of the iron oxide nanoparticles on the crystallization temperature (Figure 8a) as well as the relative crystallinity (Figure 8b). Further processing, e.g., in additive manufacturing or injection moulding, strongly depends on the melting and crystallization behavior. Figure 8 shows the determined crystallization temperatures (Figure 8a) and the corresponding relative crystallinities (Figure 8b). The nanoparticles act as crystallization nuclei and, thus, shift the crystallization temperature as expected to higher temperatures [69][70][71]. This effect is further enhanced by the higher heat transfer ability of the iron oxides. The more rapid heat transfer also explains the reduced relative crystallinity, since colder temperatures are also passed.

Vibrating Sample Magnetometer
The results of the magnetometer measurements ( Figure 9) show the full spectrum of magnetization measurements of the γ-Fe2O3 (a-1) and Fe3O4 (b-1) samples, showing that the maximum magnetization of the commercial Fe2O3 SPIONs does not differ from that of the synthesized Fe3O4 SPIONs. In the enlarged plot of the data (a-2) for γ-Fe2O3 and (b-2) for Fe3O4 functionalization, it is shown that the magnetization by the commercially available SPIONs is higher than with the self-synthesized Fe3O4. In this case, this is in line with expectations, since it has already been shown in the determination of the degree of coverage ( Figure 4) that the coating process using Fe3O4 cannot be performed optimally. A stabilization of the SPIONs that keeps the Fe3O4 nanoparticles separated during the drying step will solve this problem. Since the functionalization, unlike in liquid-phase processes, only takes place on the surface of the polymer, the resulting value of the magnetization is comparatively low. During dry coating, much less iron oxide is required to functionalize the polymer. For example, in [26], in which the ration between polymer and SPIONs is around 1:3, or in [28] where SPION contents up to 80% are used. For the iron oxides used, the more additive is applied to the surface, the higher the magnetization.

Vibrating Sample Magnetometer
The results of the magnetometer measurements ( Figure 9) show the full spectrum of magnetization measurements of the γ-Fe 2 O 3 (a-1) and Fe 3 O 4 (b-1) samples, showing that the maximum magnetization of the commercial Fe 2 O 3 SPIONs does not differ from that of the synthesized Fe 3 O 4 SPIONs. In the enlarged plot of the data (a-2) for γ-Fe 2 O 3 and (b-2) for Fe 3 O 4 functionalization, it is shown that the magnetization by the commercially available SPIONs is higher than with the self-synthesized Fe 3 O 4 . In this case, this is in line with expectations, since it has already been shown in the determination of the degree of coverage (Figure 4) that the coating process using Fe 3 O 4 cannot be performed optimally. A stabilization of the SPIONs that keeps the Fe 3 O 4 nanoparticles separated during the drying step will solve this problem. Since the functionalization, unlike in liquid-phase processes, only takes place on the surface of the polymer, the resulting value of the magnetization is comparatively low. During dry coating, much less iron oxide is required to functionalize the polymer. For example, in [26], in which the ration between polymer and SPIONs is around 1:3, or in [28] where SPION contents up to 80% are used. For the iron oxides used, the more additive is applied to the surface, the higher the magnetization.
If the maximum magnetizations are extracted from the measurements ( Figure 9) and plotted against the set iron oxide content (Figure 10), an expected linear relationship can be seen as the magnetization scales with the amount of iron oxide available. Again, the maximum values of Fe 3 O 4 are slightly lower than those of Fe 2 O 3 , which is again due to the different coating quality. If the maximum magnetizations are extracted from the measurements ( Figure 9) and plotted against the set iron oxide content (Figure 10), an expected linear relationship can be seen as the magnetization scales with the amount of iron oxide available. Again, the maximum values of Fe3O4 are slightly lower than those of Fe2O3, which is again due to the different coating quality.

Conclusions
In this work, a solvent-free dry coating process for producing magnetic polymer particles is introduced. Seven different additive concentrations of γ-Fe2O3 and Fe3O4 were used to show the versatility of this process as it is possible to coat commercially available host particles quickly and sufficiently. The discrete coating is evenly distributed on the polymer particle surface and adjustable by the additive concentration. The additive content controls the magnetization directly. Dry coating is therefore a simple, sustainable coating method for the production of magnetizable polymer composites with commer- Figure 10. Relationship between the maximum magnetization as regards the set iron oxide content.

Conclusions
In this work, a solvent-free dry coating process for producing magnetic polymer particles is introduced. Seven different additive concentrations of γ-Fe 2 O 3 and Fe 3 O 4 were used to show the versatility of this process as it is possible to coat commercially available host particles quickly and sufficiently. The discrete coating is evenly distributed on the polymer particle surface and adjustable by the additive concentration. The additive content controls the magnetization directly. Dry coating is therefore a simple, sustainable coating method for the production of magnetizable polymer composites with commercially available products and a low need for resources.