Synthesis of magnetic Fe₃O₄/polyamine hybrid microsphere using O/W/O Pickering emulsion droplet as the polymerization micro-reactor

Abstract

In order to obtain a magnetic polyamine microsphere, a magnetic oil–water-in-oil multiple emulsion droplet was used as the micro-reactor and template for polymerization. The influences of magnetite (NanoFe₃O₄) and the polymer precursor polyethyleneimine (PEI) on the properties of an emulsion were investigated. After using hydrophobic CaCO₃ nanoparticles (SM-CaCO₃) in the emulsification, the liquid paraffin-in-water emulsion stabilized by NanoFe₃O₄ and PEI was changed to an O/W/O Pickering emulsion. The investigation found that suspensions of 2 wt% SM-CaCO₃ in liquid paraffin should be emulsified with 1 wt% NanoFe₃O₄ in 20 wt% PEI solution at a ϕ_w of 0.5 in order to incorporate as much PEI as possible, in emulsion droplets, while avoiding the phase inversion of emulsion. After the crosslinking of PEI in the emulsion droplets by glutaraldehyde, a magnetic Fe₃O₄/polyamine microsphere with both hydrophilic and hydrophobic characteristics was obtained. The microsphere demonstrated an isoelectric point of 9.6. The magnetic capacity of NanoFe₃O₄ decreased as a result of being trapped in the polyamine matrix of the microsphere.

---

Introduction

Compared to a common microsphere, a magnetic polymer microsphere demonstrates the advantages of combining magnetic power with the physicochemical properties of a polymer for purposes such as, drug delivery,¹ purification of biological entities² and immobilization of proteins and enzymes.³ At the nanoscale, magnetic spheres or particles have the advantages of high specific surface area and ease of transport. However, the recovery or capture of magnetic particles with diameters in the nanometer range requires a special separator with a very high magnetic field gradient to overcome Brownian motion and the capture rate is usually low.^4,5 In contrast, magnetic microspheres of micron-scale size can be separated easily from the solution in a low strength magnetic field or with centrifugation at low speed.⁶ For instance, in industrial chromatography, micron-sized microsphere columns are more favorable than nano-sized particle columns due to the higher eluent flow rate, reduced pressure drops, and lower probability of column clog.

In contrast with the common emulsion, the Pickering emulsion is stabilized by solid particles that are self-assembled at an interface, showing excellent stability. Micron-sized droplets of Pickering emulsion have been used as micro-reactors for fabricating micron-sized hybrid polymer spheres and capsules combining the properties of inorganic particles and functional groups of the polymers.^7–9 Polyethyleneimine (PEI) has the highest cationic density of all the synthetic polymers currently available.¹⁰ PEI has been widely used in various fields such as gene therapy,¹¹ biosensors,¹² separation and purification of various proteins^13,14 and the improvement of stability and activity of enzymes.^15,16 PEI-containing supports were also developed for wastewater treatment,¹⁷ carbon dioxide removal¹⁸ and enzyme immobilization.^19,20 In the present work, using the Pickering emulsion as the polymerization template, a magnetic polyamine microsphere was prepared from the magnetic multiple O/W/O Pickering emulsion of PEI solution through a droplet-to-droplet reaction mode of crosslinking. The fabrication of the emulsion template and the morphology of the magnetic polyamine microsphere were analyzed. Unlike the microspheres with core-magnetic shell structures,^9,21 the magnetite in the magnetic polyamine microsphere was embedded in the polymer matrix, which reduces the direct exposure of magnetic particles to the surrounding medium.

Experimental section

Materials

Polyethyleneimine (PEI, M_w = 70 000 Da; 50 wt% aqueous solution) was purchased from Sigma-Aldrich Shanghai Trading Co. Ltd. Glutaraldehyde (25 wt% aqueous solution) and liquid paraffin were purchased from Sinopharm Chemical Reagent Co., Ltd., China. Stearic acid-modified calcium carbonate nanoparticles (SM-CaCO₃) with diameters of 40–90 nm were supplied by Shanghai Yaohua Nano-tech Co. Ltd., China. Other reagents used were of analytical grade and were used as received.

Preparation of Fe₃O₄ nanoparticles

Fe₃O₄ nanoparticles (NanoFe₃O₄) were prepared according to the method reported by Zhou et al.:³³ 0.05 mol FeSO₄·7H₂O and 0.1 mol FeCl₃·6H₂O were dissolved in 100 mL 0.2 mol L⁻¹ HCl solution. Then, 25 mL FeSO₄ solution was mixed with 25 mL FeCl₃ solution in a 250 mL flask. Under moderate stirring, 120 mL of 12 mol L⁻¹ NH₃·H₂O was added dropwise into the flask in a N₂ atmosphere. The reaction was carried out at 80 °C for 2 hours. The resultant black particles were collected, washed with water and dried at 45 °C in a vacuum. The size distribution of the nanoparticles was determined as 30–240 nm using a ZetaPlus zeta potential analyzer. The XRD analysis of NanoFe₃O₄ found that the peak positions were in accord with the Fe₃O₄ diffractive data in the PDF card (19-0629) in soft MDI Jade 5. The characteristic diffraction peaks are at 2θ = 30.21°, 35.59°, 43.25°, 57.10° and 62.73°, indicating that the NanoFe₃O₄ possesses a typical inverse spinel structure.

Preparation of Fe₃O₄/polyamine hybrid microsphere

0.4 g NanoFe₃O₄ was dispersed in 120 mL of 20 wt% PEI aqueous solution by homogenizing at 3000 rpm for 2 min using a XHF-D dispersator (Ningbo Xinzhi, Zhejiang, China). Then, a suspension of 1.44 g SM-CaCO₃ in 120 mL liquid paraffin was added and the mixture was homogenized at 3000 rpm for 2 min to form a magnetic Pickering emulsion of PEI. The emulsion type was identified by the drop test.

Under moderate stirring, 150 mL of 25 wt% glutaraldehyde aqueous solution was added dropwise into the magnetic Pickering emulsion at a speed of 0.2 mL min⁻¹ to cross-link the PEI. After 15 h of reaction at room temperature, the microspheres were collected by filtration and successively rinsed with diethyl ether, isopropyl alcohol and ultrapure water. Afterwards, the microspheres were reduced using 1.5 g NaBH₄ in 100 mL sodium bicarbonate buffer at pH 10 and room temperature for 2 h. The produced microspheres were collected by filtration and rinsed with ultrapure water. By soaking in the prepared 0.001 mol L⁻¹ HCl for 2 h, SM-CaCO₃ particles on the surface of the microspheres were dissolved and removed without affecting the NanoFe₃O₄ particles embedded in the matrix of the microspheres. Finally, the magnetic Fe₃O₄/polyamine microspheres were collected by filtration and washed with ultrapure water until the pH value of the washing water became neutral. The illustration of the procedure for preparing the magnetic polyamine microspheres is demonstrated in Fig. 1.

Fig. 1 The process for the synthesis of the Fe₃O₄/polyamine magnetic microsphere with the emulsion droplet as micro-reactor.

Characteristics of NanoFe₃O₄

The zeta potentials of NanoFe₃O₄ and milled microsphere (1 wt%) dispersed in water of different pH values were determined by a ZetaPlus zeta potential analyzer (Brookhaven instruments corporation, New York, USA) at 25 °C. The pH was adjusted by adding HCl or NaOH and the dispersions were left overnight at 25 °C before the test.

The X-ray diffraction (XRD) pattern of NanoFe₃O₄ was recorded by continuous scanning in a D8 Advance diffractometer (Bruker AXS, Germany) with Cu Ka radiation (λ = 0.1541 nm) at room temperature. The measurements were performed at 40 kV and 20 mA from 10° to 90° with a 2θ scanning rate of 5° min⁻¹ and a step of 0.02°.

The magnetization curves of NanoFe₃O₄ and the magnetic polyamine microsphere were obtained using a Lake Shore 7047 vibrating sample magnetometer at room temperature. The applied magnetic field was increased from 0 to 10 390 Oe, then reduced to −10 452 Oe, and then increased again to 10 964 Oe.

Viscosity and surface tension of PEI solution

Viscosities of PEI solutions at various concentrations were measured using parallel plates with a diameter of 60 mm in a Discovery Hybrid Rheometer (DHR-2, TA Instruments, DE, USA). The sample was placed between parallel plates at 25 °C and the gap between the two plates was set to 1.0 mm. The equipment was driven according to the Trios program (TA Instruments, DE, USA). The shear rate was increased from 0 s⁻¹ to 300 s⁻¹ and the viscosity was recorded.

The surface tension of PEI solution was measured using the du Noüy ring method carried out at 25.0 ± 0.1 °C on a Kruss K8 GMBH interfacial tensiometer equipped with a platinum–iridium ring. The ring was rinsed with a hydrochloric acid solution (4.0 mol L⁻¹), and then rinsed several times with ultrapure water before each measurement. All polymer solutions prepared were allowed to rest for at least four hours before measurement. Repeated measurements were made until equilibrium was observed and the variation in the surface tension was <0.2 mN m⁻¹.

Characterization of emulsion

Optical microscopic images of the emulsion droplet and microsphere were obtained using a VHX-1000 optical microscope (Keyence corporation, Japan) equipped with a digital camera (Fujifilm, Japan). Emulsion droplets (usually diluted with their continuous phase) were placed either in a wedge-shaped gap formed by a cover slip resting on a microscope slide with one edge propped up by a second cover slip or directly on a slide when the droplet size was large.²²

The emulsion type was identified by the drop test. A drop of emulsion was separately added to a small volume of the oil and the aqueous phase. An emulsion which dispersed in the aqueous phase but not in the oil phase was assessed as O/W emulsion and vice versa.²³

Characterization of microsphere

The morphology of the magnetic polyamine microsphere was examined by scanning electron microscopy (SEM). After drying for 48 h in vacuum, the microspheres were coated with a 2 nm layer of gold and imaged using an S-4800 scanning electron microscope (Hitachi Ltd., Tokyo, Japan).

The diameter and size distribution of the microspheres and emulsion droplets were determined by measuring at least 200 microspheres or emulsion droplets and analyzing the data using Excel.

The polydispersity index U, number average diameter (D_n) and weight average diameter (D_w) were calculated from the following equations:^24,25

U = D_w/D_n (3)

where n is the total number of measured microspheres or emulsion droplets, and D_i is the diameter of the determined microsphere or droplet.

Adsorption capacity of the magnetic microsphere for naphthalene was tested and used as the indicator of hydrophobicity.²⁶ A suspension of 0.05 g microspheres in 10 mL 0.1 M phosphate buffer (pH = 7.5) was added to a 40 mL sample vial and capped with a Mininert screw-cap valve (Sigma-Aldrich Co. Ltd.). For each sample, a control vial was set up following the same procedure, using 10 mL ultrapure water instead of the microsphere suspension, to account for the possible adsorption of naphthalene to the glass and caps. The head space in the vials was approximately 30 mL. Naphthalene-in-acetone stock solutions were then injected with a microsyringe such that the initial naphthalene concentration in the aqueous phase was within the range of 0.05 to 1.0 mg L⁻¹. The volume fraction of acetone in the aqueous phase in each vial was maintained at less than 0.002 to avoid possible cosolvent effects. The vials were rotated end-over-end at 20 rpm in the dark at room temperature. After 3 days, the concentration of naphthalene in the head space was analyzed by gas chromatography (GC) using a Shimadzu GC-2010 (Kyoto, Japan) coupled with flame ionization detector (FID). The concentrations of naphthalene in the aqueous phase were then calculated based on Henry's Law (with the Henry's Law constant for naphthalene at 25 °C = 0.0197) and the concentrations of naphthalene of gaseous phase.²⁶ Tests on all the samples were repeated in triplicate.

The FTIR spectra of the magnetic microspheres were obtained using a Nicolet Aratar 370 spectrophotometer (USA). The dry samples were crushed and mixed with KBr and pressed into pellets. Spectra were scanned in the range between 4000 and 400 cm⁻¹. All the spectra were obtained after subtracting KBr background, by averaging 30 scans.

The C, H, N content of the microspheres was measured with a Vario EL III elemental analyzer manufactured by Elemental Analysis in Germany.

The swelling ability of the microspheres was measured gravimetrically by swelling the microspheres in ultrapure water, alcohol, acetone and cyclohexane at 25 °C and measuring the changes in their weight during swelling. Prior to the measurement, the microspheres were dried at 45 °C in vacuum. A given number of microspheres was introduced into the medium under continuous stirring at 50 rpm. The swollen samples were removed periodically and the net weight was determined by blotting the surface with filter paper to remove medium adhered to the surface, and then immediately weighing on an electronic balance (Mettler, Model ME1004E, China). The swelling experiment was carried out until the net weight of the microsphere became constant. The degree of swelling S_w was calculated using the following equation:²⁷

where M_t is the weight of the swollen sample at time t and M₀ is the initial weight of the sample before swelling. The assays were performed in triplicate for each experiment and average values were considered for data analysis.

Results and discussion

Emulsifying ability of Fe₃O₄ nanoparticles and PEI

Polyethyleneimine (PEI) is a highly branched cationic polymer containing a high density of ionizable tertiary, secondary, and primary amino groups. As a water-soluble polymer, PEI not only increases the viscosity of solutions in water (Fig. S1A†), but also shows surface-activity by reducing the surface tension of water from 71.97 mN m⁻¹ to 53.01 mN m⁻¹ at 25 °C (Fig. S1B†). By homogenization, liquid paraffin can be dispersed in the PEI solution to form a stable O/W emulsion (Fig. 4). During homogenization, the dispersed droplets generate a newly formed interface, and the surface-active PEI spontaneously adsorbs at the interface and decreases the surface energy stabilizing the dispersed liquid paraffin droplets.²⁸ Because the increased continuous phase viscosity can significantly increase the stability of disperse systems,^29,30 the emulsion of liquid paraffin-in-PEI solution can be stabilized for more than 12 hours when the amount of PEI added is more than 1% due to the enhanced viscosity of the bulk phase (Fig. 2B). The magnified images of emulsion droplets in Fig. 2C show that the emulsion droplets are polydispersed from several μm to about 50 μm.

Fig. 2 Photographs of liquid paraffin-in-water emulsion stabilized by PEI taken after 1 h of preparation (A), 12 h after preparation (B) (from left to right, the concentration of PEI is 0.1, 1, 5, 10, 20 wt% of water phase; the volume ratio of oil to water was 1 : 1); (C) the optical micrograph of emulsion droplets (the concentration of PEI is 20 wt% of water phase; the volume ratio of oil phase to water phase is 1 : 1).

The liquid paraffin-in-water emulsion can also be stabilized by NanoFe₃O₄ (Fig. 3A). NanoFe₃O₄ prepared by the coprecipitation method carries many hydroxyl groups on the particle surfaces.^31,32 The inherently hydrophilic NanoFe₃O₄ particles can be used to prepare stable nonpolar oil-in-water emulsions.³³ When replacing the pure water phase with 20 wt% PEI aqueous solution, stable, magnetic O/W emulsions were produced (Fig. 3B). This is because PEI can not only be used as a stabilizer for liquid paraffin-in-PEI solution emulsion but can also interact with NanoFe₃O₄ surfaces through both electrostatic and non-coulombic interactions resulting in adsorption and increasing the hydrophilicity of the particles.³⁴ The enhanced positive potential of NanoFe₃O₄ in the presence of PEI is further evidence testifying the adsorption of PEI on NanoFe₃O₄ (Fig. S2†).

Fig. 3 (A) Photograph of liquid paraffin-in-water emulsion stabilized by NanoFe₃O₄ (the volume ratio of oil to water was 1 : 1); (B) photographs of liquid paraffin-in-PEI aqueous solution emulsion stabilized by NanoFe₃O₄ (from the left to right, the volume ratio of oil to solution was 3 : 2, 1 : 1, 3 : 7, 1 : 4, 1 : 9 respectively).

In order to produce magnetic polyamine microspheres from PEI, the obtained O/W magnetic emulsion droplets containing PEI and NanoFe₃O₄ were proposed as the polymerization template for conducting the crosslinking of PEI. Glutaraldehyde is one of the most effective crosslinking reagents containing aldehydic groups, which can react rapidly with the amine groups of PEI at neutral pH and room temperature to form imine linkages.^35,36 However, the crosslinking of PEI through directly adding the aqueous solution of glutaraldehyde dropwise into the magnetic emulsion merely resulted in sediments of amorphous particles or agglomerates. In this case, a large excess of amino groups of PEI in the stoichiometry led to a crosslinking reaction, which took place too rapidly for the emulsion droplets to form and maintain the globular shape during polymerization. Supposing that glutaraldehyde could react with PEI through a droplet-to-droplet mode, the intensity of crosslinking would be lessened through reducing the amount of glutaraldehyde and PEI available for each reaction. Because the dispersed droplets in the emulsion tend to coalesce in droplet-to-droplet mode, which is driven by the free energy minimization, it is reasonable to assume that the droplets of glutaraldehyde solution can coalesce and react with the droplets of PEI solution in a W/O emulsion system. In order to test this assumption, a magnetic W/O emulsion of PEI solution-in-liquid paraffin must be achieved.

Magnetic O/W/O emulsion of PEI stabilized by SM-CaCO₃

The hydrophobic SM-CaCO₃ is a nanoparticle modified with stearic acid by adsorption, which is a good stabilizer of W/O emulsion.³⁷ Fig. 4 demonstrates that a quite stable W/O Pickering emulsion of PEI solution-in-liquid paraffin was obtained by introducing hydrophobic SM-CaCO₃ into the emulsification. As the concentration of SM-CaCO₃ increased, the extent of sedimentation of the emulsion progressively decreased. A glue-like paste was obtained after emulsification when the concentration of SM-CaCO₃ reached 3 wt%. In order to endow the emulsion with magnetism, NanoFe₃O₄ in PEI solution suspension was used as the water phase. Then, a magnetic W/O Pickering emulsion in the presence of SM-CaCO₃ was prepared (Fig. 5). A surplus of NanoFe₃O₄ particles could be discerned in the continuous phase when the concentration of NanoFe₃O₄ reached 2 wt% of the water phase (Fig. 5D and E).

Fig. 4 PEI solution-in-liquid paraffin emulsion stabilized by SM-CaCO₃ after 1 h (A) and 12 h (B) of preparation (the concentrations of SM-CaCO₃ were 1, 1.5, 2, 2.5, 3 wt% of the oil phase, the volume ratio of oil to water was 1 : 1, the concentration of PEI was 20 wt%).

Fig. 5 Photographs (A) and optical micrographs (B–E) of NanoFe₃O₄ contained PEI-in-liquid paraffin emulsion stabilized by 2 wt% SM-CaCO₃ taken 6 hours after preparation. (From left to right, the wt% of NanoFe₃O₄ is 0, 0.5, 1, 2 wt%. The volume ratio of oil phase to water phase was 1 : 1.)

From enlarged optical micrographs of samples in Fig. 5B–E and 6B–F, we can clearly observe the structure of double emulsions with lots of close-packed oil droplets dispersed in the water droplets. This system of both W/O and O/W emulsion existing simultaneously is known as multiple emulsion, requiring the presence of lipophilic and hydrophilic emulsifiers.^38–40 In addition to common surfactants, Binks and co-workers found that the W/O/W multiple emulsions could be successfully stabilized by two types of particles having different hydrophobicity.²² In this study, because PEI and NanoFe₃O₄ are hydrophilic and SM-CaCO₃ is lipophilic, it is believed that SM-CaCO₃ functioned as the stabilizer of W/O outer emulsion droplets, while PEI and NanoFe₃O₄ functioned as the stabilizer of O/W inner emulsion droplets, forming the magnetic O/W/O emulsion. Careful examination of the emulsion droplets in Fig. 5 showed that some droplets had a deformed globular shape. This may be due to the assembly of NanoFe₃O₄ or PEI at some regions of the W/O interface distorting the bending of the W/O interface. In a Pickering emulsion, the solid particles are irreversibly adsorbed at the oil–water interface, acting as mechanical barriers against coalescence;^22,23,37 thus, the migration of particles is minimal after emulsion formation.

Fig. 6 (A) Photographs of PEI-in-liquid paraffin emulsion containing 1 wt% NanoFe₃O₄ stabilized by 2 wt% SM-CaCO₃ taken after 6 hours of emulsion preparation (from left to right, the volume ratio of oil phase to water phase is 4 : 1, 3 : 1, 2 : 1, 3 : 2, 5 : 4, 2 : 3, 1 : 2), and (B–F) the optimal micrographs of the multiple emulsion droplets (from B to F, the volume ratio of oil to water is 4 : 1, 3 : 1, 2 : 1, 3 : 2, 5 : 4).

The magnetic multiple O/W/O emulsion obtained can be stabilized for more than 6 hours, which satisfies the time requirement for accomplishing the polymerization of PEI (see later). However, it was found that the emulsion gelated when the concentration of SM-CaCO₃ in liquid paraffin and the concentration of NanoFe₃O₄ in the water phase reached more than 3 wt% and 1 wt%, respectively. This catastrophic change in the presence of the emulsion may be associated with floc structures in which many particles were connected by polymer chains of PEI and bound together.⁴¹ The other reason put forward is that the homogenization used in emulsion preparation may give rise to shear-induced gelation.⁴¹ The presence of strong hydrogen-bonding interactions between the hydroxyl groups of nanoparticles and the hydrogen atoms of the polymer chains can also cause cross-linking gelation of the polymer solution.⁴² The gelation of the emulsion is the factor impairing the crosslinking reaction by retarding the free diffusion of glutaraldehyde droplets added in the emulsion; thus, hindering them from encountering and reacting with PEI-containing emulsion droplets.

Investigation found that the diameters of emulsion droplets were reduced as a result of increasing either the amount of hydrophilic NanoFe₃O₄ or that of hydrophobic SM-CaCO₃ for emulsification (Tables S1 and S2†). It is suggested that both the particles can assemble at the interface of the external emulsion, leading to reduced size despite differences in their hydrophilicity.^23,43 The average diameter of emulsion droplets can also be decreased by reducing ϕ_w (Table S3†). According to Binks et al., the increased ϕ_w creates new interfaces, which may not be sufficiently covered by adsorbed particles, Resulting in drop growth by coalescence during emulsification.²³ When the volume ratio of the water to oil phase was at 1 : 1, the diameter of the emulsion droplets reached the maximal value. Fig. 6 shows that the extent of sedimentation decreased as ϕ_w increased. When ϕ_w was more than 0.5, catastrophic phase inversion occurred. In order to incorporate as much PEI as possible in the emulsion droplet for crosslinking, the size of the emulsion droplet should be as large as possible while avoiding phase inversion (Fig. 5A). Hence, the emulsion template for crosslinking PEI must be fabricated by homogenizing the mixture of 1 wt% NanoFe₃O₄ in 20% PEI solution and 2 wt% SM-CaCO₃ in liquid paraffin at a water-to-oil volume ratio of 1 : 1.

Magnetic Fe₃O₄/polyamine microsphere

Using the PEI and NanoFe₃O₄-containing magnetic O/W/O emulsion stabilized with SM-CaCO₃ as the polymerization template, an aqueous solution of glutaraldehyde was added dropwise to cross-link the PEI in the droplet-to-droplet mode. When glutaraldehyde solution (25 wt% in water) was added dropwise into the PEI-containing magnetic O/W/O Pickering emulsion, the drops of glutaraldehyde would diffuse through the bulk phase of liquid paraffin to encounter emulsion droplets of PEI solution driven by free energy minimization. Then, a polymerization or crosslinking reaction occurred between the amine groups of PEI and the aldehyde groups of glutaraldehyde to generate thermally and chemically stable imine linkages through the Schiff reaction. After hydrogenization of imine linkages with NaBH₄ (ref. 35) and the removal of SM-CaCO₃, magnetic Fe₃O₄/polyamine microspheres were obtained and are shown in Fig. 7. The rustycolored microspheres appear translucent. Clusters of NanoFe₃O₄ can be identified inside the microsphere. SEM images of microspheres demonstrated that there are small pores on the microsphere surface after the removal of SM-CaCO₃ (Fig. 7F). Observation of collapsed microspheres with dents on the surface indicates that the microspheres are hollow. The cross-section SEM of the microspheres shows Fe₃O₄ nanoparticles, in the form of single particles or agglomerates are dispersed across a shell layer of 20–30 μm in thickness (Fig. 7G–I). It is also clear that the shell layer has multi-hollow structures and that each pore is not interconnected (Fig. 7H).

Fig. 7 Optical micrographs and SEM of the Fe₃O₄/polyamine magnetic microspheres from (A to C), the addition amount of 25 wt% glutaraldehyde was 0.175, 0.35, 0.75 mL mL⁻¹ emulsion respectively; (D) SEM of microspheres; (E) the outer surface of microsphere with SM-CaCO₃; (F) the outer surface of microsphere after the removal of SM-CaCO₃; (G) SEM image of cross-section of the microsphere; (H and I) SEM images of shell layer of the microsphere.

Fujii et al. consider the Pickering emulsion to be an efficient polymerization template for the fabrication of materials with multiple hollows, with the droplets in the inner phase functioning as porogens.^42,44 In this study, the synergic effect of hydrophobic SM-CaCO₃, hydrophilic NanoFe₃O₄ and PEI promote the formation of O/W/O multiple emulsions in which small oil droplets are entrapped within larger water droplets which in turn are dispersed in a continuous oil phase. During polymerization, reactions occurred at the W/O interface of the external emulsion, which was stabilized mainly by SM-CaCO₃, consuming PEI molecules near the interface. The concentration gradient of PEI between the inside and the W/O interface of external emulsion droplets causes a PEI flow to the interface and leaves voids after polymerization. Thus, multi-hollow polyamine microspheres were fabricated. The multi-hollow structure of shell layers in microspheres indicates the presence of a W/O/W emulsion during the polymerization. Because no assembly of Fe₃O₄ nanoparticles can be found at the surface of the hollows, the O/W inner emulsion droplets may be stabilized primarily by PEI molecules.

Investigation into the number average diameters of the microspheres demonstrated that the average diameter decreased as a result of increasing the amount of glutaraldehyde crosslinker added (Table 1). Compared to the size distribution of emulsion droplets used as polymerization templates (Tables S1–S3†), the magnetic polyamine microspheres were smaller in diameter. This is due to the polycondensation of PEI in the micro-reactor of the emulsion droplet causing the molecules to adopt a more compact arrangement. Moreover, the reduction in the occupied volume after polymerization also indicates that there was only transfer of glutaraldehyde molecules to cross-link PEI when the glutaraldehyde droplets made contact with the emulsion droplets. If coalescence of water had occurred between the glutaraldehyde droplets and the emulsion droplets, the emulsion droplets would have been enlarged or broken. The assembly of particles on the interface of emulsion droplets was robust enough to prevent the coalescence of water in the glutaraldehyde droplets with that of the emulsion droplets. According to Table 1, the polydispersity index of the microspheres decreased as the amount of added crosslinker was increased.

Table 1 The influence of the amount of cross-linker added on the size distribution of the microspheres

Glutaraldehyde added (mL mL^−1 emulsion)	Number average diameter (μm)	Diameter distribution range (μm)	U
0.175	79.37 ± 3.62	55.99 ± 3.13–139.97 ± 8.67	1.38
0.35	62.44 ± 5.62	34.02 ± 2.21–106.63 ± 6.15	1.20
0.75	48.68 ± 3.31	32.17 ± 3.53–84.25 ± 6.16	1.18

---

Compared to other methods,^19,45–48 the preparation of magnetic polyamine microspheres using the magnetic O/W/O Pickering emulsion of PEI as the polymerization template is simple and more effective. The amount of PEI used in the polymerization can be as high as 20 wt% of the water phase, which is significantly higher than the amount reported by other papers.^19,46–48

FTIR and TG analysis of the microsphere

FTIR analysis results for the magnetic Fe₃O₄/polyamine microspheres and PEI are shown in Fig. 8. IR (KBr) cm⁻¹: 3200–3400 (–N–H stretching), 1550–1650 (–N–H in-plane deformation vibration), 1046 (–C–N stretching), 900–650 (–N–H out-of-plane deformation vibration), 2850–2929 (–C–H stretching), 1465 (–C–H asymmetric bending vibration). Unlike PEI, the FTIR spectrum of the microsphere displays an absorption peak at 575 cm⁻¹, which overlaps the peak at 650 cm⁻¹ and is ascribed to the stretching band of Fe–O in NanoFe₃O₄. It was also found that the strength of the peak for the –C–H asymmetric bending vibration at 1465 cm⁻¹ significantly increased after polymerization. This may be caused by the alkyl chains of glutaraldehyde cross-linking with PEI.

Fig. 8 Infrared spectra of PEI (A) and microspheres (B).

Thermal-stabilities of PEI and the magnetic Fe₃O₄/polyamine microspheres are compared in Fig. 9. The PEI lost 40 wt% of the initial weight in the range 50 °C to 135 °C, and decomposed completely in the range 270 °C to 390 °C. In contrast, the magnetic polyamine microsphere lost 95.1 wt% of the total weight in the range 50 °C to 550 °C (Fig. 9A). The decomposition residue is about 1.9 wt%, which is ascribed to NanoFe₃O₄. According to the DTG curves (Fig. 9B), the polymerization shifted the maximal decomposition rate from 335 °C (PEI) to 388 °C (the microspheres). Both PEI and the polyamine microspheres present an initial mass loss in the range from 50 °C to 128 °C. This is possibly due to the loss of adsorbed water and CO₂. CO₂ is easily absorbed by the amino groups of alkaline PEI, impregnated or covalently coated PEI materials.^49–52

Fig. 9 TG (A) and DTG (B) thermograms of the microsphere.

Elementary analysis, zeta potential and hydrophobicity of the magnetic microsphere

Elementary analysis results for magnetic microspheres prepared using different amounts of cross linker are shown in Table 2. When the amount of glutaraldehyde was enhanced, the content of N in the microsphere reduced, while the content of C in the microsphere increased. Because the quantity of PEI in the emulsion template was fixed, the increased supply of glutaraldehyde for crosslinking definitely increased the level of C content in the microsphere.

Table 2 The elementary analysis of the magnetic microspheres prepared by different amount of cross linker

The amount of 25% glutaraldehyde for crosslinking (mL)	N (wt%)	C (wt%)	H (wt%)
3.5	12.04 ± 0.28	38.12 ± 1.29	8.31 ± 0.21
7	10.58 ± 0.03	44.52 ± 0.89	8.72 ± 0.44
15	9.23 ± 0.11	48.22 ± 0.19	8.95 ± 0.62

---

Zeta potential measurement showed that the isoelectric point (pI) of the magnetic polyamine microsphere is 9.6 (Fig. 10A). The amount of naphthalene adsorbed by the microsphere increased with the increase in amount of naphthalene added (Fig. 10B). This indicates that hydrophobic parts existed in the magnetic microsphere. This result is attributed to the long alkyl chain of the cross-linker, glutaraldehyde, which increases the hydrophobicity of the microsphere. This observation is similar to that of glutaraldehyde cross-linked chitosan microspheres, in which hydrophobicity was found to be in direct proportion to the degree of cross-linking.²⁷

Fig. 10 Zeta potential (A) and naphthalene adsorption capacity (B) of the magnetic microspheres.

Degree of swelling and magnetic behavior of the magnetic polyamine microsphere

The magnetic polyamine microspheres were soaked in ultrapure water, ethanol, acetone and cyclohexane in order to study the swelling behavior, and the results are shown in Table 3. The microspheres can be swollen in both water and non-polar cyclohexane, and the swelling ability of the microspheres decreased as the amount of crosslinker added was increased. The inverse relationship between the swelling ability of the microsphere and the degree of crosslinking is due to the enhanced compactness of matrices, which contain a greater degree of crosslinking inhibiting the permeation of solvent.⁵³

Table 3 The swelling ability of microspheres for different solvent

25% glutaraldehyde solution used (mL mL^−1 emulsion)	Sw (%)
	Water	Ethanol	Acetone	Cyclohexane
0.175	41.48 ± 0.23	30.32 ± 0.36	23.67 ± 0.11	22.82 ± 0.11
0.35	39.85 ± 0.09	28.92 ± 0.12	21.18 ± 0.12	18.71 ± 0.15
0.75	38.37 ± 0.15	26.43 ± 0.14	19.41 ± 0.09	14.13 ± 0.21

---

The comparison of magnetization of NanoFe₃O₄ and the magnetic polyamine microspheres as a function of the magnetic field at room temperature is demonstrated in Fig. 11. The saturation magnetization, coercivity and remanent magnetization values for NanoFe₃O₄ are 46.077 emu g⁻¹, 27.747 Oe and 0.09 emu g⁻¹, respectively. An extremely small hysteresis loop is exhibited with a low coercivity. This was characteristic of superparamagnetic particles.⁵⁴ As for the microspheres, the hysteresis loop indicates a saturation magnetization of 11.1 emu g⁻¹, a coercivity of 121.7 Oe and a remanent magnetization of 2.05 emu g⁻¹ at room temperature. The decrease in saturation magnetization might be due to the protective layer of polymer matrix formed around NanoFe₃O₄. The magnetic dead layer of polymer on the surface of nanoparticles prevents the coupled dipoles from aligning along the magnetic orientation, decreasing the saturation magnetization of NanoFe₃O₄.⁵⁵ Because many magnetite particles dispersed in the microspheres are in the agglomerate form (Fig. 7), the dipolar interactions between NanoFe₃O₄ particles may cause hysteresis when anisotropic energy is increased, resulting in the increase in coercivity.⁵⁶ Fig. 12 demonstrates that the microspheres in water can move towards the outer field set outside the vessel. This sensitive magnetic response promises applications related to the magnetic field such as bioseparation and biomedical applications.

Fig. 11 Hysteresis loop of NanoFe₃O₄ (A) and the magnetic Fe₃O₄/polyamine microsphere (B).

Fig. 12 Photographs of the magnetic microspheres in water attracted by a magnet.

Conclusion

Both NanoFe₃O₄ and PEI are good stabilizers for the O/W emulsion of liquid paraffin-in-PEI solution; crosslinking of PEI through directly adding glutaraldehyde solution dropwise into the O/W emulsion merely resulted in sediments of amorphous particles or agglomerates. Thus, hydrophobic nanoparticles of SM-CaCO₃ were introduced to form a O/W/O Pickering emulsion. During the polymerization, the droplets of glutaraldehyde coalesced with emulsion droplets of PEI, driven by free energy minimization, and crosslinked the PEI in the droplet-to-droplet mode. Thus, the Fe₃O₄/polyamine hybrid microsphere was obtained. The magnetic polyamine microspheres obtained have both hydrophilic and hydrophobic characteristics, showing great potential in the fields of separation, CO₂ fixation and biocatalysis.

References

1. Y. Wu, J. Guo, W. Yang, C. Wang and S. Fu, Polymer, 2006, 47, 5287–5294.
2. J.-M. Nam, C. S. Thaxton and C. A. Mirkin, Science, 2003, 301, 1884–1886.
3. J. Huang, B. Han, W. Yue and H. Yan, J. Mater. Chem., 2007, 17, 3812–3818.
4. Q. A. Pankhurst, J. Connolly, S. K. Jones and J. Dobson, J. Phys. D: Appl. Phys., 2003, 36, R167.
5. I. Safarik and M. Safarikova, J. Chromatogr. B: Biomed. Sci. Appl., 1999, 722, 33–53.
6. M. Tudorache, M. Co, H. Lifgren and J. Emneus, Anal. Chem., 2005, 77, 7156–7162.
7. C. Wang, C. Zhang, Y. Li, Y. Chen and Z. Tong, React. Funct. Polym., 2009, 69, 750–754.
8. A. Schmid, J. Tonnar and S. P. Armes, Adv. Mater., 2008, 20, 3331–3336.
9. Q. Gao, C. Wang, H. Liu, Y. Chen and Z. Tong, Polym. Chem., 2010, 1, 75–77.
10. L. Kouisni and D. Rochefort, J. Appl. Polym. Sci., 2009, 111, 1–10.
11. E.-H. Lin, H.-Y. Chang, S.-D. Yeh, K.-Y. Yang, H.-S. Hu and C.-W. Wu, Nanomedicine: Nanotechnology, Biology and Medicine, 2013, 9, 1293–1303.
12. Y. Zhang, M. A. Arugula, M. Wales, J. Wild and A. L. Simonian, Biosens. Bioelectron., 2015, 67, 287–295.
13. U. Dissing and B. Mattiasson, Bioseparation, 1998, 7, 221–229.
14. C. Holler, D. Vaughan and C. M. Zhang, J. Chromatogr. A, 2007, 1142, 98–105.
15. M. M. Andersson and R. Hatti-Kaul, J. Biotechnol., 1999, 72, 21–31.
16. C. Garcia-Galan, O. Barbosa and R. Fernandez-Lafuente, Enzyme Microb. Technol., 2013, 52, 211–217.
17. M. Amara and H. Kerdjoudj, Talanta, 2003, 60, 991–1001.
18. E. R. Monazam, J. Spenik and L. J. Shadle, Chem. Eng. J., 2013, 223, 795–805.
19. F. Wang, Z. Gu, Z. Cui and L. Liu, Bioresour. Technol., 2011, 102, 9374–9379.
20. C. Cui, Y. Tao, L. Li, B. Chen and T. Tan, J. Mol. Catal. B: Enzym., 2013, 91, 59–66.
21. A. Vilchez, C. Rodriguez-Abreu, J. Esquena, A. Menner and A. Bismarck, Langmuir, 2011, 27, 13342–13352.
22. B. P. Binks and S. O. Lumsdon, Langmuir, 2000, 16, 2539–2547.
23. B. P. Binks and C. P. Whitby, Langmuir, 2004, 20, 1130–1137.
24. B. B. Weiner, in Modern Methods of Particle Size Analysis, ed. H. G. Barth, Wiley, New York, 1984, vol. 73, ch. 3, p. 111.
25. L. M. Gan, N. Lian, C. H. Chew and G. Z. Li, Langmuir, 1994, 10, 2197–2201.
26. Y. Xiao and M. R. Wiesner, J. Hazard. Mater., 2012, 215–216, 146–151.
27. K. C. Gupta and F. H. Jabrail, Carbohydr. Polym., 2006, 66, 43–54.
28. D. Langevin, Adv. Colloid Interface Sci., 2000, 88, 209–222.
29. O. Behrend, K. Ax and H. Schubert, Ultrason. Sonochem., 2000, 7, 77–85.
30. S. Tesch and H. Schubert, J. Food Eng., 2002, 52, 305–312.
31. C. Yee, G. Kataby, A. Ulman, T. Prozorov, H. White, A. King, M. Rafailovich, J. Sokolov and A. Gedanken, Langmuir, 1999, 15, 7111–7115.
32. G. Kataby, M. Cojocaru, R. Prozorov and A. Gedanken, Langmuir, 1999, 15, 1703–1708.
33. J. Zhou, X. Qiao, B. P. Binks, K. Sun, M. Bai, Y. Li and Y. Liu, Langmuir, 2011, 27, 3308–3316.
34. R. Meszaros, L. Thompson, M. Bos and P. De Groot, Langmuir, 2002, 18, 6164–6169.
35. I. Migneault, C. Dartiguenave, M. J. Bertrand and K. C. Waldron, BioTechniques, 2004, 37, 790–802.
36. C. P. Govardhan, Curr. Opin. Biotechnol., 1999, 10, 331–335.
37. Z. G. Cui, C. F. Cui, Y. Zhu and B. P. Binks, Langmuir, 2012, 28, 314–320.
38. J. Jiao, D. G. Rhodes and D. J. Burgess, J. Colloid Interface Sci., 2002, 250, 444–450.
39. A. Desforges, M. Arpontet, H. Deleuze and O. Mondain-Monval, React. Funct. Polym., 2002, 53, 183–192.
40. D. Štefanec and P. Krajnc, Polym. Int., 2007, 56, 1313–1319.
41. Y. Saito, Y. Hirose and Y. Otsubo, Colloids Surf., A, 2011, 384, 40–46.
42. S. Fujii, Y. Eguchi and Y. Nakamura, RSC Adv., 2014, 4, 32534–32537.
43. B. P. Binks, J. Philip and J. A. Rodrigues, Langmuir, 2005, 21, 3296–3302.
44. H. Zhou, T. Shi and X. Zhou, J. Appl. Polym. Sci., 2014, 131, 39761.
45. J. M. Bolivar, J. Rocha-Martin, C. Mateo, F. Cava, J. Berenguer, R. Fernandez-Lafuente and J. M. Guisan, Biomacromolecules, 2009, 10, 742–747.
46. Y. Murakami, K. Sugo, M. Hirano and T. Okuyama, Talanta, 2011, 85, 1298–1303.
47. W. Tong, C. Gao and H. Mohwald, Polym. Adv. Technol., 2008, 19, 817–823.
48. M. S. Niepel, D. Peschel, X. Sisquella, J. A. Planell and T. Groth, Biomaterials, 2009, 30, 4939–4947.
49. J. Wang, H. Chen, H. Zhou, X. Liu, W. Qiao, D. Long and L. Ling, J. Environ. Sci., 2013, 25, 124–132.
50. S. Sung and M. P. Suh, J. Mater. Chem. A, 2014, 2, 13245–13249.
51. E. P. Dillon, C. A. Crouse and A. R. Barron, ACS Nano, 2008, 2, 156–164.
52. P. Li, B. Ge, S. Zhang, S. Chen, Q. Zhang and Y. Zhao, Langmuir, 2008, 24, 6567–6574.
53. E. Uguzdogan, E. B. Denkbas, E. Ozturk, S. A. Tuncel and O. S. Kabasakal, J. Hazard. Mater., 2009, 162, 1073–1080.
54. X. Luo, S. Liu, J. Zhou and L. Zhang, J. Mater. Chem., 2009, 19, 3538–3545.
55. A. A. Novakova, V. Y. Lanchinskaya, A. V. Volkov, T. S. Gendler, T. Y. Kiseleva, M. A. Moskvina and S. B. Zezin, J. Magn. Magn. Mater., 2003, 258–259, 354–357.
56. Z. L. Liu, Z. H. Ding, K. L. Yao, J. Tao, G. H. Du, Q. H. Lu, X. Wang, F. L. Gong and X. Chen, J. Magn. Magn. Mater., 2003, 265, 98–105.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra00997a

---