Controlled radical polymerization of styrene with magnetic iron oxides prepared through hydrothermal, bioinspired, and bacterial processes

Abstract

Controlled/living radical polymerization was examined with the use of magnetic iron oxide (Fe₃O₄) prepared through various processes, including hydrothermal synthesis, a bioinspired process, and magnetotactic bacteria. Prior to the use of various types of Fe₃O₄, commercially available Fe₃O₄ was employed as a heterogeneous catalyst for styrene polymerization in conjunction with an alkyl halide as an initiator. Under appropriately optimized conditions, with the addition of Ph₃P in a solvent mixture of toluene and DMF, the polymerization proceeded in a well-controlled manner. In addition, solid Fe₃O₄ was recovered using a magnetic approach and was reused for a polymerization reaction. The effects of stirring on the polymerization rate suggest that the surface of Fe₃O₄ is responsible for the catalysis in polymerization. Fe₃O₄ samples prepared through various processes were then used for styrene polymerization and showed different activities depending on the preparation process. In particular, Fe₃O₄ prepared through hydrothermal synthesis exhibited a much higher activity compared with commercial Fe₃O₄.

---

Introduction

Metal catalysts have been developed in various fields of chemistry to enhance reaction rates, activate specific substrates, and improve stereoselectivity. In addition to high performance in catalysis, added values that satisfy recent demands, such as low environmental load and high cost-effectiveness, provide a competitive advantage to catalysts. Among the metal elements employed for catalysts, iron is the most promising element that meets these demands because of its abundance in nature, low cost, nontoxicity, and ready availability. Recent developments in precision polymer synthesis have depended partly on iron catalysts designed for new techniques, including transition metal-catalyzed living radical polymerization and atom transfer radical polymerization.^1–5 Moreover, even simple iron halides exhibit excellent activity in controlled cationic^6–8 or radical^9–20 polymerization reactions when combined with suitable additives or ligands.

Iron oxides are common materials that are readily available from substances such as rust and iron sand and exhibit various interesting properties. Notably, heterogeneous catalysis of iron oxides is a distinct feature that is not attained with soluble iron catalysts. The merits of reactions using heterogeneous catalysts include stereoselectivity and easy removal of the catalyst. Because iron oxides exhibit Lewis acidity and redox properties, various reactions, such as Friedel–Crafts reactions,^21,22 reductive dechlorination of carbon tetrachloride,^23,24 and reduction of nitrobenzene,²⁵ have been conducted with iron oxides. In polymer synthesis, the Lewis acidity of iron oxides, including α-Fe₂O₃, γ-Fe₃O₄, and Fe₃O₄, has been harnessed for living cationic polymerization of vinyl ethers in a heterogeneous system.^26,27 Moreover, iron oxides prepared through a bioinspired process or synthesized by magnetotactic bacteria have been demonstrated to catalyze polymerization.²⁸ These specially prepared iron oxides had possibility to exhibit superior activity in polymerization compared with commercial iron oxides. On the other hand, iron oxides composed of divalent iron species have been found to induce controlled radical polymerization of styrene and methyl methacrylate.²⁹ In the past study, the polymerization reactions have been primarily investigated with FeO, which is an iron oxide consisting of a divalent iron species. Although polymerization using Fe₃O₄, a main component of iron sand, also proceeds in a relatively controlled manner, detailed research has not yet been conducted. In addition, radical polymerization has only been examined with commercial materials. Specially prepared iron oxides, such as Fe₃O₄, obtained through a hydrothermal synthesis,³⁰ synthesized via a bioinspired process,³¹ synthesized by bacteria,^32–34 or coated with oleate,³⁵ are expected to exhibit activity superior to that of commercial iron oxides.

In this study, we aimed to construct a suitable initiating system for the controlled radical polymerization of styrene using Fe₃O₄ in a heterogeneous system. The reaction conditions were first optimized by investigating the effects of additives and solvents on the polymerization behavior. Subsequently, Fe₃O₄ samples prepared via various processes were employed for the radical polymerization of styrene. Fe₃O₄ prepared through a hydrothermal synthesis exhibited much higher activity compared with commercial Fe₃O₄.

Experimental section

Materials

Styrene (Kishida, 99.5% or Nacalai Tesque, 99%), methyl methacrylate (MMA, TCI, >99.8%), ethyl α-bromophenylacetate (EBPA; Aldrich, 97%), nBu₃N (Wako, >98%), and N,N,N′,N′-tetramethylenediamine (TMEDA; ACROS, 99%) were distilled over calcium hydride under reduced pressure. N,N-Dimethylformamide (DMF; Kanto, 99.5% or Wako, 99.5%, super dehydrated) was distilled over calcium hydride under reduced pressure or used as received. A commercial Fe₃O₄ (Aldrich, 99.99%), Ph₃P (Aldrich, 99%), nBu₃P (Kanto, >98%), tricyclohexylphosphine (Cy₃P; Aldrich), 1,2-bis(diphenylphosphino)ethane (dppe; Kanto, >98%), 1,3-bis(diphenylphosphino)propane (dppp; TCI, >98.0%), 1,4-bis(diphenylphosphino)butane (dppb; Aldrich, 98%), nBu₄NBr (TCI, >99.0%), and triphenylamine (Ph₃N; TCI, >98.0%) were used as received. Toluene (Kanto, >99.5%, H₂O < 10 ppm or Wako, 99.5%) was purified by passage through solvent purification columns (Glass Contour).

Preparation of iron oxides

Hydrothermal synthesis³⁰. First, 5 mmol of FeCl₂·4H₂O (Kanto, >99.0%) and 5 mmol of picolinic acid (Kanto, >98.0%) were dissolved in 20 mL of distilled water with stirring. After the formation of a transparent solution, 10 mL of 4 M NaOH (Wako, >97.0%) aq. was added to the solution, and a suspension was formed. The prepared suspension was stirred vigorously and then sealed in a Teflon-lined stainless steel autoclave. The autoclave was heated at 200 °C for 6 h and was subsequently cooled to room temperature. Black precipitates were collected from the mixture by magnetic separation and were sequentially rinsed three times. The final product was obtained after air drying at room temperature.

Bioinspired process³¹. An aqueous solution of disodium–dihydrogen ethylenediamine tetraacetate (EDTA, Kanto, 99.5%) was prepared in a polypropylene vessel using purified water. Then, 25 mM of FeCl₂·4H₂O (Kanto, 99.0%) was dissolved in the aqueous EDTA solution or purified water. The pH of the aqueous solution was adjusted to 11 or 13 using 10 M NaOH (Junsei, 96.0%). The precursor solutions were maintained at 60 or 90 °C for 1 day without stirring. The precipitates were centrifuged and washed with purified water, and the resultant powder was dried at 60 °C.

Bacterial process. Bacterial magnetite (Fe₃O₄) was isolated from magnetotactic bacteria (Magnetospirillum magneticum AMB-1) and was purified by sequentially washing with a surfactant, a buffer solution, ethanol, and toluene. The isolated and dried Fe₃O₄ was heated at 100 °C under reduced pressure for 10 min immediately before use.

Fe₃O₄ coated with oleate³⁵. A solution of sodium oleate (0.0761 g, 0.250 mmol; TCI) dissolved in distilled water (24.7 mL) was added to an aqueous solution (25.0 mL) of FeCl₂·4H₂O (0.994 g, 5.00 mmol; Aldrich, 99%). Hydrazine monohydrate (0.30 mL; TCI, 98%) was added to the resulting white suspension, and the mixture was vigorously stirred. The solution was placed in a polytetrafluoroethylene vessel and heated at 200 °C for 3 h. After cooling to room temperature, the mature Fe₃O₄ nanoparticles were collected by centrifugation. The product was then poured into hexane and centrifuged to remove precipitated byproducts. Subsequently, ethanol was added to the supernatant and the generated aggregates were collected and redispersed in hexane. Finally, hexane was evaporated to obtain Fe₃O₄ coated with oleate.

Polymerization procedures

Polymerization was conducted under dry nitrogen in a baked glass tube equipped with a three-way stopcock. A typical polymerization reaction is described as follows: to the glass tube containing Fe₃O₄ and Ph₃P, toluene, DMF, styrene, and a stock solution of EBPA in toluene were added. The tube was immersed in a thermostatic oil bath at 80 °C. During the polymerization reaction, the heterogeneous reaction mixture was stirred using a magnetic stirrer (Koike Precision Instruments HE-20 GA; ∼1350 rpm) with a small handmade stirring bar consisting of a thin glass tube and a wire. A small amount of the reaction mixture was extracted several times to determine the monomer conversion, and the reaction was terminated by cooling and exposure to air. Monomer conversion was determined from the concentration of residual monomer measured by gas chromatography with toluene or DMF as the internal standard. The quenched reaction mixtures were diluted with toluene, washed with water, and evaporated to dryness under reduced pressure to give the product polymers.

Measurements

The molecular weight distribution (MWD) of the polymers was measured by gel permeation chromatography (GPC) at 40 °C with polystyrene gel columns [Shodex KF-805L (exclusion limit molecular weight = 4 × 10⁶; bead size = 10 μm pore size: 200–1000 Å; 8.0 mm I.D. × 300 mm; eluent: tetrahydrofuran) × 2 or TSKgel GMH_HR-M × 2 (exclusion limit molecular weight = 4 × 10⁶; bead size = 5 μm; column size = 7.8 mm I.D. × 300 mm; eluent: chloroform); flow rate = 1.0 mL min⁻¹] connected to a JASCO PU-2080 precision pump, a JASCO CO-2065 column oven, a JASCO UV-2075 ultraviolet detector, and a JASCO RI-2031 refractive-index detector (for the Shodex columns) or to a Tosoh DP-8020 pump, a CO-8020 column oven, a UV-8020 ultraviolet detector, and an RI-8020 refractive-index detector (for the Tosoh columns). The number-average molecular weight (M_n) and the polydispersity ratio [weight-average molecular weight/number-average molecular weight (M_w/M_n)] were calculated from the chromatographs with respect to 16 polystyrene standards (Varian: M_p = 575–2.783 × 10⁶; M_w/M_n = 1.02–1.10 or Tosoh: M_n = 577–1.09 × 10⁶, M_w/M_n ≤ 1.1) or 10 poly(MMA) standards (Varian: M_p = 875–1.677 × 10⁶; M_w/M_n = 1.02–1.23). ¹H NMR spectra were recorded on a JEOL ECS-400 spectrometer.

Results and discussion

Commercial Fe₃O₄ for controlled radical polymerization of styrene: optimization of reaction conditions

Fe₃O₄ was shown to induce radical polymerization of styrene and methyl methacrylate in a previous paper, but the reaction conditions were not optimized, in sharp contrast to the detailed research performed for polymerization using a different iron oxide, FeO. Preliminary results²⁹ show that the polymerization of styrene in conjunction with EBPA as an initiator (Scheme 1A) proceeded at a rate much smaller than that using FeO to yield polymers with unimodal but slightly broad MWDs (entry 1 in Table 1). The use of nBu₄NBr, which is a suitable additive for controlled polymerization using FeO, was not as effective in the case of Fe₃O₄ (entry 14). Thus, radical polymerization of styrene using Fe₃O₄ was investigated in detail in this study, particularly focusing on the effects of additives, solvents, and stirring on the controllability and reaction rate.

Scheme 1 (A) Radical polymerization of styrene using Fe₃O₄ and (B) recovery of Fe₃O₄ after polymerization reaction.

Table 1 Radical polymerization of styrene and MMA using commercial Fe₃O₄^a

Entry	Monomer	Additive	Tol/DMF (v/v)	Time (h)	Conv. (%)	Mn × 10^−3 (calcd)	Mn × 10^−3^b	Mw/Mn^b
a [Monomer]0 = 4.0 M, [EBPA]0 = 40 mM, [Fe3O4]0/[EBPA]0 = 2, [additive]0 = 20 mM, in toluene/DMF (2/1 or 8/1 v/v) or toluene alone at 80 °C.b By GPC (polystyrene calibration for entries 1–14 or poly(MMA) calibration for entries 15–19).c A product with a bimodal MWD was obtained. The values of the higher and the lower peaks are shown in the cells.
1	Styrene	None	2/1	411	83	8.9	8.4	1.53
2			8/1	339	80	8.6	8.7	1.49
3		Ph3P	2/1	411	78	8.4	10.1	1.30
4			8/1	292	85	9.1	9.3	1.13
5		nBu3P	2/1	456	74	8.0	21.2	1.39
6		Cy3P	2/1	339	63	6.8	12.7	1.28
7		dppe	2/1	263	80	8.5	20.4	1.33
8		dppp	2/1	263	78	8.4	29.0	1.53
9		dppb	2/1	413	78	8.4	38.0	1.83
10		Ph3N	2/1	456	62	6.7	7.4	1.45
11			8/1	267	80	8.5	8.6	1.55
12		nBu3N	2/1	75	64	6.9	8.6	2.40
13		TMEDA	2/1	150	84	9.0	13.0	1.85
14		nBu4NBr	2/1	196	29	3.2	4.0	1.49
15	MMA	Ph3P	8/1	29	66	6.8	8.4	1.44
16			Toluene alone	72	89	9.1	10.6	1.38
17		None	Toluene alone	72	53	5.6	121, 2.7^c	2.76, 1.53^c
18		nBu3P	Toluene alone	72	87	9.0	16.5	1.59
19		Cy3P	Toluene alone	72	78	8.1	15.7	1.63

---

In an investigation of the effects of various additives on polymerization behavior, Ph₃P was found to be superior to other additives with respect to controllability (entry 3). The polymerization of styrene using the Fe₃O₄/Ph₃P system proceeded at a rate similar to that in the absence of additives and yielded polymers with narrower MWDs. The M_n values of the products agreed relatively well with the calculated values, indicating that the polymerization proceeded in a controlled manner. In contrast to the beneficial effect of arylphosphine on polymerization, trialkylphosphines (nBu₃P and Cy₃P) and bidentate arylphosphines (dppe, dppp, and dppb) did not improve the controllability but induced less-controlled polymerization reactions (entries 5–9). The products obtained in the presence of these phosphine additives had M_n values greater than the calculated values, suggesting that the initiation efficiency declined under these conditions. The nitrogen-containing counterpart of Ph₃P, i.e., Ph₃N, was also effective for the generation of well-controlled polymers, although the controllability was slightly inferior to that of Ph₃P (entry 10). As in the case of alkylphosphine additives, alkylamines (nBu₃N and TMEDA) were not suitable for controlled reactions.

In addition to the use of a suitable additive, an appropriate ratio of toluene/DMF solvent was important for obtaining control over the polymerization. A decrease of the ratio of DMF from 2/1 to 8/1 v/v resulted in an improvement of the controllability to yield polymers with narrower MWDs both in the absence and presence of Ph₃P (entries 2 and 4). In particular, a small M_w/M_n ratio of 1.13 was attained in the reaction when using Ph₃P in toluene/DMF at 8/1 v/v. A ¹H NMR analysis of the products revealed that the functionality of the carbon–bromine bond at the propagating chain end (ω-end) depended on the solvents employed. In a solvent mixture with a larger DMF ratio, the carbon–bromine bond at the propagating chain end gradually decomposed through the elimination reaction of hydrogen bromide to give an olefin structure. In a past study, an elimination reaction of hydrogen bromide from 1-phenylethyl bromide was reported to occur to generate styrene in polar solvents, such as acetonitrile and nitromethane, upon heating.³⁶ DMF also promoted the elimination reaction in a similar way. The solvent nucleophilically attacks the carbon atom attached to the bromine atom to promote the elimination reaction. DMF was found to be indispensable for controlled polymerization because the reaction in toluene alone resulted in products with negligible carbon–bromine ends and with a high olefin ratio. The carbon–bromine ends were heterolytically cleaved to generate carbocationic species through the abstraction of the bromide anion on the Lewis acidic site of Fe₃O₄ because the suppression of the Lewis acidity was insufficient in the absence of DMF, which is a relatively strong Lewis base. Such heterolytic cleavage was followed by the β-proton elimination reaction to give the olefin chain end.

The present system was also effective for controlled radical polymerization of MMA. The polymerization of MMA smoothly proceeded to yield a polymer with a unimodal MWD in the presence of Ph₃P in toluene/DMF (8/1 v/v; entry 15; see Fig. S1 in the ESI† for the MWD curves), i.e., under the conditions that are suited for the polymerization of styrene. However, DMF was not necessary unlike the case of styrene (entry 16; Fig. S1†), potentially because the carbon–bromine bond at the propagating end derived from MMA does not heterolytically cleave due to the non-polymerizability of MMA in a cationic mechanism. In addition, the reaction in the absence of Ph₃P resulted in a worse-controlled reaction to give a product with a bimodal MWD (entry 17). Moreover, polymers with MW much higher than the theoretical values were obtained in the presence of trialkylphosphines (entries 18 and 19), in a manner similar to the polymerization of styrene. The results indicate that the reaction conditions need to be suitably tuned depending on monomers employed.

The reusability of Fe₃O₄ was examined under the most appropriate reaction conditions, as described above (Fig. 1). After the first use of Fe₃O₄ for the polymerization of styrene (red symbols), Fe₃O₄ was isolated from the product and residue solution with the use of a magnet (Scheme 1B). Subsequently, the Fe₃O₄ was washed with solvents and dried under reduced pressure for reuse in the polymerization reaction. The recovered Fe₃O₄ retained the activity to induce polymerization of styrene (blue symbols) in a manner similar to the reaction using fresh Fe₃O₄. The M_n values of the product polymers increased linearly along the theoretical line, suggesting the occurrence of highly controlled polymerization. Furthermore, the third use of Fe₃O₄ (green symbols) exhibited polymerization results comparable with those of the fresh and once-recovered counterparts. The slightly lower polymerization rates at the initial stages of the reaction with the recovered catalysts stem from the partial loss of activity on the Fe₃O₄ surface due to oxidation during the recovery steps.

Fig. 1 Reuse of Fe₃O₄ in radical polymerization of styrene: (A) time–conversion curves, (B) M_n and M_w/M_n values, and (C) MWD curves for the polystyrenes obtained with the fresh Fe₃O₄: [styrene]₀ = 4.0 M, [EBPA]₀ = 40 mM, [Fe₃O₄]₀/[EBPA]₀ = 2, [Ph₃P]₀ = 20 mM, in toluene/DMF (8/1 v/v) at 80 °C.

The effect of stirring on the reaction rate suggests that the surface of Fe₃O₄ is responsible for the catalysis in the radical polymerization reaction. The reaction without stirring was much slower compared with the reaction in a mixture that was vigorously stirred with a stirring bar throughout polymerization, as shown in Fig. 2 (red symbols). Therefore, frequent contact between the propagating chain end and the Fe₃O₄ surface through vigorous stirring is required for an efficient polymerization reaction. Furthermore, when the Fe₃O₄ was removed from the reaction mixture during polymerization, the reaction no longer proceeded (blue symbols in Fig. 2). In addition, the separated supernatant solution gradually turned orange, indicating that some iron species leached out of the Fe₃O₄ during polymerization and changed to colored species,³⁷ such as trivalent iron complexes. These soluble species alone, however, cannot catalyze polymerization in the absence of Fe₃O₄. The UV-vis spectrum of the supernatant differed from the spectra of FeBr₂/Ph₃P and FeBr₃/Ph₃P mixtures.

Fig. 2 Polymerization of styrene with Fe₃O₄/Ph₃P: the polymerization without stirring (red) or the removal of Fe₃O₄ during the reaction (blue): [styrene]₀ = 4.0 M, [EBPA]₀ = 40 mM, [Fe₃O₄]₀/[EBPA]₀ = 2, [Ph₃P]₀ = 20 mM, in toluene/DMF (8/1 v/v) at 80 °C.

Given the polymerization results described above, Fe₃O₄ is thought to function both as the activator and the reducing agent during the polymerization reaction (Scheme 2). The low valent Fe^II species on the surface of Fe₃O₄ (Fe^II[surf]) activates the carbon–bromine bond to generate a carbon radical and a high valent Fe^III species attached to a bromine atom (Br–Fe^III[surf]) through homolytic cleavage (eqn (1) in Scheme 2). A possible deactivation step is the reverse process of the activation, in which Br–Fe^III[surf] gives the bromine atom to a radical species. In addition to these processes on the surface of Fe₃O₄, soluble iron species that leach out of Fe₃O₄ may participate in the polymerization reaction. Because polymerization did not proceed with the separate supernatant alone, the leached species with orange color may be high valent Fe^III species (Fe^III[sol]). Fe^III[sol] species accumulate through processes such as two-molecular irreversible termination reactions. In the presence of Fe₃O₄, however, Fe^II[surf] species reduce high valent Fe^III species in solution (Fe^III[sol]) to active Fe^II species at a low valent state (Fe^II[sol]) during the polymerization reaction (eqn (3)). The Fe^II[sol] species may activate a dormant species (eqn (2)). In addition, the polymerization rate using Fe₃O₄ was much smaller than that using FeO, divalent iron oxide, under similar reaction conditions,^29,38 which suggests that the Fe^II species are definitely responsible for the generation of the propagating radical species through the direct activation or the reduction of soluble Fe^III species. In such polymerization mechanisms, additives coordinate to iron species on the surface and/or in solution to affect each reaction step. A beneficial additive, such as Ph₃P, makes the activation–deactivation reactions smoother to allow for even propagation reactions for each propagating chain, resulting in narrower MWDs of the products compared with those obtained in the absence of additives.

Scheme 2 Plausible mechanisms of radical polymerization using Fe₃O₄.

Controlled radical polymerization of styrene with the use of various Fe₃O₄ samples prepared through hydrothermal, bioinspired, and bacterial processes

The radical polymerization of styrene was next examined with the use of various Fe₃O₄ samples prepared through different processes. The iron oxides employed include Fe₃O₄ prepared through a hydrothermal synthesis, Fe₃O₄ obtained through a bioinspired process, Fe₃O₄ synthesized by magnetotactic bacteria, and Fe₃O₄ coated with oleate. These Fe₃O₄ particles have shapes that differ from commercial Fe₃O₄ (Fig. 3). In the hydrothermal synthesis,³⁰ picolinic acid is employed as an additive to generate Fe₃O₄ with high-index facets. In the bioinspired process,³¹ various iron oxides, including α-Fe₂O₃, α-FeOOH, γ-FeOOH, and Fe₃O₄, are generated with changes in the EDTA concentration, pH, and temperature. The obtained iron oxides have hierarchically organized structures consisting of nanoscale objects or mesocrystal structures. In addition, magnetotactic bacteria produce Fe₃O₄ with a well-defined shape.^32–34 Moreover, unlike the Fe₃O₄ obtained above, Fe₃O₄ coated with oleate³⁵ is easily dispersed in organic solvents because of the oleophilic outer layer. Radical polymerizations of styrene using these iron oxides were conducted using Ph₃P in conjunction with EBPA as an initiator in toluene/DMF (8/1 v/v) at 80 °C. The polymerization results are summarized in Table 2. The iron oxides all exhibited catalytic activity, although the degree of activity differed.

Fig. 3 (A) SEM images of Fe₃O₄ prepared through hydrothermal synthesis, (B) SEM images of Fe₃O₄ obtained through bioinspired process, (C) tunneling electron microscopy (TEM) image of Fe₃O₄ produced by bacteria, (D) photograph of Fe₃O₄ coated with oleate in hexane, and (E) SEM images of commercial Fe₃O₄.

Table 2 Radical polymerization of styrene using various Fe₃O₄^a

Entry	Fe3O4		Time (h)	Conv. (%)	Mn × 10^−3 (calcd)	Mn × 10^−3 ^b	Mw/Mn^b
	Method	Form
a [Styrene]0 = 4.0 M, [EBPA]0 = 40 mM, [Fe3O4]0/[EBPA]0 = 2, [Ph3P]0 = 20 mM, in toluene/DMF (8/1 v/v) at 80 °C.b By GPC (polystyrene calibration).c Conditions for Fe3O4 synthesis: (i) [EDTA]0 = 0 mM, pH = 11, at 60 °C; (ii) [EDTA]0 = 15 mM, pH = 11, at 60 °C; (iii) [EDTA]0 = 0 mM, pH = 13, at 60 °C; and (iv) [EDTA]0 = 0 mM, pH = 11, at 90 °C.d Fe3O4 with a lot number different from that used in Table 1.
1	Hydrothermal synthesis	Crystal	42	85	9.1	12.3	1.60
2	Bioinspired process^c (i)	Mesocrystal	214	24	2.7	3.0	1.28
3	Bioinspired process^c (ii)	Mesocrystal	214	17	2.0	2.2	1.29
4	Bioinspired process^c (iii)	Mesocrystal	327	52	5.7	6.3	1.26
5	Bioinspired process^c (iv)	Mesocrystal	335	18	2.2	2.8	1.19
6	Bacterial process	Crystal	196	31	3.5	3.5	1.26
7	Coated with oleate	Dispersed	313	70	7.6	13.0	1.24
8	Commercial^d	Sintered-like	309	64	6.9	7.0	1.15

---

Fe₃O₄ prepared through hydrothermal synthesis exhibited the highest activity (Fig. 4). The polymerization rate was much higher than that of commercial Fe₃O₄. The polymerization started without an induction period to reach a monomer conversion of 85% in 42 h. The M_n values of the products increased with increasing monomer conversion, although the results were slightly larger than the calculated values. Moreover, the polymerization proceeded even after the addition of a fresh feed of monomer at a later stage of the reaction (Fig. S2†), indicating the long-lived nature of the propagating species. In addition, Fe₃O₄ retained activity after the reaction and could be reused for a second polymerization reaction. The polymerization using the reused Fe₃O₄ proceeded at a similar rate, giving products with unimodal MWDs. In addition, stirring during polymerization affected the polymerization rate (Fig. 5), similar to the case of commercial Fe₃O₄. The reaction without stirring proceeded more slowly than the reaction with vigorous stirring (red symbols in Fig. 5). When stirring was stopped during the polymerization, the reaction rate decreased immediately (blue symbols in Fig. 5). Moreover, the resumption of stirring resulted in a reacceleration of the reaction. Thus, the surface of Fe₃O₄ is responsible for the activity in the polymerization.

Fig. 4 Radical polymerization of styrene using Fe₃O₄ prepared through a hydrothermal synthesis: (A) time–conversion curves, (B) M_n and M_w/M_n values, and (C) MWD curves for the polystyrenes: [styrene]₀ = 4.0 M, [EBPA]₀ = 40 mM, [Fe₃O₄]₀/[EBPA]₀ = 2, [Ph₃P]₀ = 20 mM, in toluene/DMF (8/1 v/v) at 80 °C.

Fig. 5 Effects of stirring on radical polymerization of styrene using Fe₃O₄ prepared through a hydrothermal synthesis: [styrene]₀ = 4.0 M, [EBPA]₀ = 40 mM, [Fe₃O₄]₀/[EBPA]₀ = 2, [Ph₃P]₀ = 20 mM, in toluene/DMF (8/1 v/v) at 80 °C.

A scanning electron microscopy (SEM) analysis of Fe₃O₄ before and after polymerization indicated some uncertainty regarding the active sites on Fe₃O₄. The shape and size of the hydrothermally synthesized Fe₃O₄ changed after being used twice for polymerization (Fig. 3A), although the polymerizations with the fresh and reused Fe₃O₄ proceeded at similar rates. The change in shape and size may have resulted from partial leaching of Fe₃O₄ during polymerization in a manner similar to the case for commercial Fe₃O₄. The crystal phase of Fe₃O₄, however, was unchanged after use for polymerization (confirmed by X-ray diffraction). In addition, high-index facets partially remained on the iron oxides. These facts indicate that the surfaces exposed after leaching exhibited activity similar to that of the fresh surfaces. Alternatively, reactions using Fe₃O₄ with activity greater than a certain degree would proceed at similar rates. In such cases, a reaction pathway in which the surface of Fe₃O₄ is not involved may be the rate-determining step in polymerization.

The polymerization reactions using Fe₃O₄ obtained through a bioinspired process (entries 2–5 in Table 2) or synthesized by magnetotactic bacteria (entry 6) also proceeded in controlled manners, but the reaction rates were smaller compared with commercial Fe₃O₄. In addition, various Fe₃O₄ samples prepared under different pH values, EDTA concentrations, and temperatures (entries 2–5) exhibited similar activity levels, irrespective of the preparation conditions.

The polymerization of styrene with Fe₃O₄ coated with oleate occurred in a dispersed state to give polymers with unimodal MWDs. The reaction rate was comparable with that for commercial Fe₃O₄. Because the surface of Fe₃O₄ is not completely covered with oleate, the exposed parts on Fe₃O₄ are responsible for the catalysis in polymerization.

The difference in polymerization activity among the Fe₃O₄ samples prepared through various processes resulted from the difference in surface conditions. However, the Fe₃O₄ synthesized through a bioinspired process had a larger BET surface area (21–47 m² g⁻¹) compared with the Fe₃O₄ prepared through hydrothermal synthesis (0.5 m² g⁻¹) and the commercial Fe₃O₄ (3.0 m² g⁻¹), but it exhibited a slower polymerization rate. Thus, particular iron species located on special surfaces, edges, or corners would be responsible for the catalysis in polymerization. Although the details are not fully clear at the present stage, the results obtained with various Fe₃O₄ samples suggest that the catalytic activity can be tailored through the preparation conditions of the iron oxide.

Conclusion

In conclusion, radical polymerization of styrene was demonstrated to proceed in a highly controlled manner with the use of an Fe₃O₄/Ph₃P initiating system. An appropriate ratio of toluene/DMF as the reaction solvent was indispensable for controlled polymerization through the retention of high functionality of the carbon–bromine end. The effect of stirring on the polymerization rates indicated that the surface of Fe₃O₄ is responsible for the catalysis in polymerization. Moreover, Fe₃O₄ samples prepared through various processes exhibited different activities in polymerization. In particular, Fe₃O₄ prepared through hydrothermal synthesis exhibited a higher activity compared with commercial Fe₃O₄. A more elaborate design of Fe₃O₄ will allow for fine-tuning of the catalytic activity to obtain better control over the radical polymerization of various monomers.

Acknowledgements

This work was partially supported by Grant-in-Aid for Scientific Research on Innovative Areas of “Fusion Materials” (no. 2206) from MEXT, Japan.

References and Notes

1. M. Kamigaito, T. Ando and M. Sawamoto, Chem. Rev., 2001, 101, 1721–1723.
2. M. Ouchi, T. Terashima and M. Sawamoto, Chem. Rev., 2009, 109, 4963–5050.
3. K. Matyjaszewski and J. Xia, Chem. Rev., 2001, 101, 2921–2990.
4. F. di Lena and K. Matyjaszewski, Prog. Polym. Sci., 2010, 35, 959–1021.
5. K. Satoh, M. Kamigaito and M. Sawamoto, in Polymer Science: A Comprehensive Reference, ed. K. Matyjaszewski and M. Möller, Elsevier B. V., Amsterdam, 2012, vol. 3.13.
6. A. Kanazawa, Y. Hirabaru, S. Kanaoka and S. Aoshima, J. Polym. Sci., Part A: Polym. Chem., 2006, 44, 5795–5800.
7. R.-F. Yan, A.-R. Guo, Q. Liu and Y.-X. Wu, J. Polym. Sci., Part A: Polym. Chem., 2012, 50, 3383–3392.
8. H. Aoshima, K. Satoh and M. Kamigaito, Macromol. Symp., 2013, 323, 64–74.
9. T. Ando, M. Kamigaito and M. Sawamoto, Macromolecules, 1997, 30, 4507–4510.
10. K. Matyjaszewski, M. Wei, J. Xia and N. E. McDermott, Macromolecules, 1997, 30, 8161–8164.
11. G. Moineau, P. Dubois, R. Jérôme, T. Senninger and P. Teyssié, Macromolecules, 1998, 31, 545–547.
12. S. Zhu and D. Yan, Macromol. Rapid Commun., 2000, 21, 1209–1213.
13. J. Louie and R. H. Grubbs, Chem. Commun., 2000, 1479–1480.
14. M. Teodorescu, S. G. Gaynor and K. Matyjaszewski, Macromolecules, 2000, 33, 2335–2339.
15. V. C. Gibson, R. K. O'Reilly, W. Reed, D. F. Wass, A. J. P. White and D. J. Williams, Chem. Commun., 2002, 1850–1851.
16. C. Uchiike, T. Terashima, M. Ouchi, T. Ando, M. Kamigaito and M. Sawamoto, Macromolecules, 2007, 40, 8658–8662.
17. K. Satoh, H. Aoshima and M. Kamigaito, J. Polym. Sci., Part A: Polym. Chem., 2008, 46, 6358–6363.
18. M. Ishio, M. Katsube, M. Ouchi, M. Sawamoto and Y. Inoue, Macromolecules, 2009, 42, 188–193.
19. H. Aoshima, K. Satoh, T. Umemura and M. Kamigaito, Polym. Chem., 2013, 4, 3554–3562.
20. K. Nishizawa, M. Ouchi and M. Sawamoto, Macromolecules, 2013, 46, 3342–3349.
21. J. O. Morley, J. Chem. Soc., Perkin Trans. 2, 1977, 2, 601–605.
22. N. B. Shrigadi, A. B. Shinde and S. D. Samant, Appl. Catal., A, 2003, 252, 23–35.
23. M. L. McCormick and P. Adriaens, Environ. Sci. Technol., 2004, 38, 1045–1053.
24. K. M. Danielsen and K. F. Hayes, Environ. Sci. Technol., 2004, 38, 4745–4752.
25. C. A. Gorski and M. M. Scherer, Environ. Sci. Technol., 2009, 43, 3675–3680.
26. A. Kanazawa, S. Kanaoka and S. Aoshima, J. Am. Chem. Soc., 2007, 129, 2420–2421.
27. A. Kanazawa, S. Kanaoka and S. Aoshima, J. Polym. Sci., Part A: Polym. Chem., 2010, 48, 916–926.
28. A. Kanazawa, S. Kanaoka, N. Yagita, Y. Oaki, H. Imai, M. Oda, A. Arakaki, T. Matsunaga and S. Aoshima, Chem. Commun., 2012, 48, 10904–10906.
29. A. Kanazawa, K. Satoh and M. Kamigaito, Macromolecules, 2011, 44, 1927–1933.
30. J. Sato, M. Kobayashi, H. Kato, T. Miyazaki and M. Kakihana, Journal of Asian Ceramic Societies, 2014, 2, 258–262.
31. Y. Oaki, N. Yagita and H. Imai, Chem. – Eur. J., 2012, 18, 110–116.
32. R. P. Blakemore, Science, 1975, 190, 377–379.
33. A. Arakaki, H. Nakazawa, M. Nemoto, T. Mori and T. Matsunaga, J. R. Soc., Interface, 2008, 5, 977–999.
34. A. Arakaki, J. Webb and T. Matsunaga, J. Biol. Chem., 2003, 278, 8745–8750.
35. K. Katagiri, K. Ohta, K. Sako, K. Inumaru, K. Hayashi, Y. Sasaki and K. Akiyoshi, ChemPlusChem, 2014, 79, 1631–1637.
36. K. Matyjaszewski, K. Davis, T. E. Patten and M. Wei, Tetrahedron, 1997, 53, 15321–15329.
37. Because some iron species leached out of Fe₃O₄ during the polymerization, crude products need to be washed with water to obtain colorless polymers.
38. Polymerization using FeO under conditions similar to those for the reaction using Fe₃O₄: conv. = 83% in 51 h, M_n = 10.8 × 10³, M_w/M_n = 1.41; [styrene]₀ = 4.0 M, [EBPA] = 40 mM, [Fe₃O₄]₀/[EBPA]₀ = 5, [Ph₃P]₀ = 20 mM, in toluene/DMF (8/1 v/v) at 80 °C.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra09149g

---