*Article* **Polyimide-Derived Carbon-Coated Li4Ti5O<sup>12</sup> as High-Rate Anode Materials for Lithium Ion Batteries**

**Shih-Chieh Hsu <sup>1</sup> , Tzu-Ten Huang <sup>2</sup> , Yen-Ju Wu <sup>3</sup> , Cheng-Zhang Lu <sup>4</sup> , Huei Chu Weng 5,\* , Jen-Hsien Huang <sup>6</sup> , Cai-Wan Chang-Jian 7,\* and Ting-Yu Liu 8,\***


**Abstract:** Carbon-coated Li4Ti5O<sup>12</sup> (LTO) has been prepared using polyimide (PI) as a carbon source via the thermal imidization of polyamic acid (PAA) followed by a carbonization process. In this study, the PI with different structures based on pyromellitic dianhydride (PMDA), 4,40 -oxydianiline (ODA), and *p*-phenylenediamine (*p*-PDA) moieties have been synthesized. The effect of the PI structure on the electrochemical performance of the carbon-coated LTO has been investigated. The results indicate that the molecular arrangement of PI can be improved when the rigid *p*-PDA units are introduced into the PI backbone. The carbons derived from the *p*-PDA-based PI show a more regular graphite structure with fewer defects and higher conductivity. As a result, the carbon-coated LTO exhibits a better rate performance with a discharge capacity of 137.5 mAh/g at 20 C, which is almost 1.5 times larger than that of bare LTO (94.4 mAh/g).

**Keywords:** Li4Ti5O12; polyimide; carbon coating; lithium ion battery; rate performance

## **1. Introduction**

With the continuous depletion of fossil fuels and associated increasing air pollution, it is essential to raise the proportion of renewable energy supplies. However, renewable energy sources such as wind and solar are intermittent and cannot be stockpiled without energy storage systems (ESSs). ESSs based on lithium ion batteries (LIBs) are emerging as one of the key solutions to effectively integrate high shares of variable renewable energy due to their high energy density, zero memory effect and long lifespan [1]. By introducing an EES into the power generation system, it can smooth the output of wind or solar power generation, and reduce the impact on the power grid. Unfortunately, most of the anode and cathode materials suffer from low electronic conductivity and poor ionic diffusivity in the lattice resulting in poor rate capability. Therefore, the rate performance of the active materials must be further improved for use in voltage regulation and frequency modulation.

**Citation:** Hsu, S.-C.; Huang, T.-T.; Wu, Y.-J.; Lu, C.-Z.; Weng, H.C.; Huang, J.-H.; Chang-Jian, C.-W.; Liu, T.-Y. Polyimide-Derived Carbon-Coated Li4Ti5O<sup>12</sup> as High-Rate Anode Materials for Lithium Ion Batteries. *Polymers* **2021**, *13*, 1672. https://doi.org/10.3390/ polym13111672

Academic Editor: Vito Di Noto

Received: 25 April 2021 Accepted: 19 May 2021 Published: 21 May 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

To address the above issue, alien ion doping [2], the building of a nanoporous structure [3,4] and surface coating [5–7] have proven to be effective for improving the rate performance by narrowing the band gap, shortening migration paths for Li<sup>+</sup> ions and reduction of interfacial resistance, respectively. Among these approaches, the surface coating is a feasible strategy to improve the LIB performance due to the multi-functional advantages such as enhancement of electric conductivity [8], structural stability [9] and offering a physical protection to avoid side reactions with the electrolyte [10]. Several compounds have been proposed to be an efficient coating layer such as carbon-based materials [11,12], metal oxides/hydroxides [13,14], phosphide-based materials [15] and glass-based materials [16]. Among these candidates, carbon coating is one of the most effective and facile ways to improve the electrochemical performance of LIB materials due to its excellent electron conductivity, low cost, and superior chemical/electrochemical stability. Compared with the metal oxide coating, carbon coating can easily form a smooth and uniform thin layer with high surface coverage on the active materials. Moreover, the carbon coating layer can also serve as a buffer layer to accommodate the dimensional variation of the active material during the lithiation and delithiation process leading to improved structural stability [17].

Recently, polyimide (PI) has been found to be a high-quality carbon source for the synthesis of graphene [18,19] and highly conductive carbon [20], because of the abundant hexagonal crystalline carbon within the imide structures. The obtained PI derived graphene reveals remarkable conductivity and electrochemical properties, which are useful for various applications e.g., supercapacitors [21], sensors [22], electrocatalysts [23] and electrothermal heaters [24]. Therefore, polyimide is expected to be suitable for forming the conductive carbon layer of LIB materials. It has been reported that the carbon films derived from PI with different structures can show quite different physical properties and graphite microstructure. However, the effect of carbon coating derived from different PIs on the LIB performance is still unclear.

In this study, two different PIs have been prepared from pyromellitic dianhydride (PMDA), 4,40 -oxydianiline (ODA) and *p*-phenylenediamine (*p*-PDA). By introducing the rigid planar moiety (*p*-PDA) into the main chain of classical PMDA/ODA PI (PO-PI), the PMDA/ODA/*p*-PDA PI (POP-PI) shows an increased crystallinity and orientation degree. Here, we used the PO-PI and POP-PI as carbon sources to modify the Li4Ti5O<sup>12</sup> (LTO) anode material, followed by thermal treatment to obtain the carbon-coated LTO materials. The results indicate that the uniformly coated carbon layer on LTO can reduce its resistance and polarization leading to better rate performance and the corresponding electrochemical properties. Moreover, with incorporation of the rigid *p*-PDA segment, the carbon derived from the POP-PI exhibits better molecular packing and less defect. As a result, the carbon-coated LTO prepared from the POP-PI can display an even better kinetic performance than that obtained from the PO-PI.

#### **2. Experimental Section**

#### *2.1. Material*

The monomers ODA (97%), *p*-PDA (98%) and PMDA (97%) were purchased from Jinyu Co., Ltd. (Kaohsiung, Taiwan). The TiO<sup>2</sup> (85%, anatase, Hombikat 8602) and Li2CO<sup>3</sup> (≥99%, Aldrich) used for the synthesis of LTO were purchased from World Chem Industries Co., Ltd. (Taipei, Taiwan) and Sigma Aldrich (St. Louis, MI, USA), respectively.

#### *2.2. Preparation of Pyromellitic Dianhydride/4,4*0 *-Oxydianiline Polyimide (PMDA/ODA PI) and PMDA/ODA/p-Phenylenediamine (p-PDA) PI*

In this study, the PI precursor, poly(amic acid) (PAA) was obtained through the two-steps synthesis method from its monomers. First, the ODA monomer or mixed ODA/*p*-PDA (1:1) were dissolved in *N*-methyl-2-pyrrolidone (NMP). Then, equimolar PMDA was added in the diamine solution under continuous stirring at 25 ◦C with a solid content of 15 wt% to produce the PAA solution. Here, the PAAs prepared from PMDA/ODA and PMDA/ODA/*p*-PDA are

denoted as PO-PAA and POP-PAA, respectively. The PAA was then cast on glass substrate by doctor blade coating, s followed by soft baking at 80 ◦C for 60 min. The soft-baked precursor films were thermally imidized (150 ◦C for 30 min, 250 ◦C for 30 min, 350 ◦C for 60 min and 400 ◦C for 30 min) under N<sup>2</sup> atmosphere to obtain the PMDA/ODA PI and PMDA/ODA/*p*-PDA PI denoted herein as PO-PI and POP-PI, respectively.

## *2.3. Preparation of Carbon-Coated Li4Ti5O<sup>12</sup> (LTO)*

The carbon-coated LTO powders were prepared by spray drying precursor solution of lithium titanium peroxide, followed by solid-state calcination. Firstly, the TiO<sup>2</sup> and Li2CO<sup>3</sup> were dispersed in de-ionized water with a Li:Ti molar ratio 4:5 and a solid content of 15 wt%. The solution was ball milled for 12 h to produce the homogenous slurry, which was fed to a pilot spray dryer (OHKAWARA KAKOHKI, model L-8i, No. 145874). The spherical LTO precursor obtained was added to the PMDA/ODA or PMDA/ODA/*p*-PDA PAA solution, which was then stirred for 60 min and subsequently centrifuged. The centrifuged powders were dried in vacuum at 80 ◦C. Finally, the samples were thermally annealed via a stepwise process (150 ◦C for 30 min, 250 ◦C for 30 min, 350 ◦C for 60 min, 400 ◦C for 30 min, 500 ◦C for 60 min and 800 ◦C for 120 min). The modified LTO prepared from PO- and POP-PI are denoted herein as PO-LTO and POP-LTO, respectively. For the preparation of the bare LTO sample, the spray-dried precursor was directly annealed with the same heating program.

## *2.4. Characterization*

The crystal structure of the sample was characterized by X-ray powder diffraction (XRD, Philips X'Pert/MPD instrument, El Dorado County, CA, USA). The morphologies were monitored by scanning electron microscopy (SEM, JEOL JSM-6701F, Tokyo, Japan) and transmission electron microscopy (TEM, JEOL 2010, Tokyo, Japan). The absorption measurement was taken using a Cintra 2020 (GBC scientific equipment, Australia) spectrophotometer. The differential scanning calorimetry (DSC) was measured on DSC 2500 (TA instrument, Lukens Drive, New Castle, DE, USA). Thermal gravimetric analysis (TGA) was carried out using a TGA 8000 (PerkinElmer, Boston, MS, USA). Functional group and chemical composition were characterized by using Fourier transform infrared (FTIR, PerkinElmer, Boston, MS, USA) spectroscopy and X-ray photoelectron spectroscopy (XPS, ULVAC-PHI, Tokyo, Japan). Raman spectrum of the as-prepared samples was measured using a Raman microscope (HR800, HORIBA, Tokyo, Japan).

## *2.5. Electrochemical Analysis*

The energy storage performance of the samples was evaluated by fabricating the coin cells in an argon filled glovebox. The working electrode was prepared by mixing active material, polyvinylidene fluoride (PVdF), KS4 and Super P with a ratio of 8:0.5:0.5:1 in *N*-methyl-2 pyrrolidone (NMP). The as-prepared slurry was cast onto a Al foil and then dried in a vacuum oven at 120 ◦C overnight. The electrolyte comprises 1.0 M LiPF<sup>6</sup> dissolved in a mixture of ethylene carbonate and dimethyl carbonate at a volumetric ratio of 1:1. Cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS) were carried out on a potentiostat (PARSTAT 4000 Potentiostat Galvanostat). The galvanostatic charge/discharge performance was examined between 1 and 2.5 V (vs. Li+/Li) at various C rates (1 C = 175 mA/g).

#### **3. Results and Discussion**

In general, the molecular packing of PI can be improved by using rigid *p*-PDA units to replace the rotatable ODA moiety. It has been reported that the regular PI with higher orientation degree of molecular chains can produce better graphite structure after a carbonization process [25]. Therefore, in this study, we prepared the carbon-coated LTO powders using the less-ordered PO-PI and more ordered POP-PI as carbon sources. POand POP-PI were synthesized through a conventional two-step polycondensation of PMDA dianhydride with ODA and *p*-PDA diamines and the reaction process is shown in Figure 1a. The ultraviolet–visible (UV-vis) absorption spectra of the PO- and POP-PAA solutions are

shown in Figure 1b. The inset in Figure 1b presents the digital photograph of the PO- and POP-PAA solutions. The PO-PAA solution exhibits strong absorption in the UV region with an absorption edge of 410 nm. With the incorporation of the PDA moiety into the polymer backbone, the absorption profile of POP-PAA tends to be shifted toward the long wavelength region. The red-shift of the absorption spectrum of POP-PAA can be explained by the formation of a charge transfer complex (CTC) between the alternating electron-acceptor (dianhydride) and electron-donor (diamine) moieties [26]. The C–O–C bond in ODA units can separate the chromaphoric centers and cut the electronic conjugations [27]. As a result, the PO-PAA shows a lighter color compared with that of POP-PAA due to its reduced intra-/intermolecular CTC formation. These results indicate that the PDA diamine units are successfully introduced into the copolymer. To study the effect of the *p*-PDA moiety on the imidization process, the enthalpy of imidization for the PO-PAA and POP-PAA was recorded by DSC measurement. As shown in Figure 1c, both the PAA samples exhibit a distinctive endothermic peak centered at around 175.5~179.8 ◦C in the first run which originated from the imidization reaction. In the second run, the DSC curves show a smooth profile without any peak, indicating the imidization process is complete. In addition, the enthalpy of imidization integrated from the DSC curve (first run) was calculated to be 228.6 and 235.6 J/g for PO-PAA and POP-PAA, respectively. The similar peak temperature and enthalpy of imidization for the two PAAs suggest that the incorporation of the *p*-PDA unit cannot alter the imidization behavior and the corresponding chemical conversion. *Polymers* **2021**, *13*, x FOR PEER REVIEW 7 of 13

**Figure 1.** (**a**) The process of preparation of PMDA (pyromellitic dianhydride) and ODA (4,4′-oxydianiline) (PO-) and PMDA/ODA/*p*-phenylenediamine (*p*-PDA) polyimide (POP-PI) with PMDA, ODA, and *p*-PDA; (**b**) the absorbance spectrum of PO- and POP-PAA solution and (**c**) the differential scanning calorimetry (DSC) profile of the PO- and POP-PAA (poly(amic acid)). **Figure 1.** (**a**) The process of preparation of PMDA (pyromellitic dianhydride) and ODA (4,40 -oxydianiline) (PO-) and PMDA/ODA/*p*-phenylenediamine (*p*-PDA) polyimide (POP-PI) with PMDA, ODA, and *p*-PDA; (**b**) the absorbance spectrum of PO- and POP-PAA solution and (**c**) the differential scanning calorimetry (DSC) profile of the PO- and POP-PAA (poly(amic acid)).

The FTIR spectra of the as-prepared PAA and its corresponding PI are shown in Figure 2a. Both the two PAAs exhibit a broad band located at 1630 cm−<sup>1</sup> , which can be assigned to amide C=O stretching mode [28,29]. After thermal treatment, this peak disappeared in the spectra of PO-PI and POP-PI (Figure 2b). Moreover, a new peak due to the C–N–C stretching vibration appears at 1365 cm−<sup>1</sup> [30]. Meanwhile, the C=C double bond in the benzene ring at around 1500 cm−<sup>1</sup> remains unchanged after the imidization process. These results indicate that the two PAAs were successfully converted into PO-PI and POP-PI after the imidization process. The peak at 1234 cm−<sup>1</sup> is characteristic of C–O–C stretching vibration belonging to the ODA moiety [25]. It is worth noting that the POP-PAA and POP-PI reveal a much weaker intensity of C–O–C bond than their counterparts, confirming the successful incorporation of PDA moiety into the polymer backbone. The XPS O 1s and N 1s spectra of PO-PI are also shown in Figure 2c,d. After deconvolution, the O 1s XPS spectrum of the PO-PI can be fitted with two peak components assigned to C=O and C–O–C species at 531.9 and 533.2 eV, respectively [31]. For the N 1s spectrum of the PO-PI, only one characteristic peak contributed from O=C–N specie at 399.7 eV can be obtained after deconvolution [32]. These results indicate that the PAA has been successfully converted into the PI structure after the thermal treatment. *Polymers* **2021**, *13*, x FOR PEER REVIEW 8 of 13

**Figure 2.** The Fourier transform infrared (FTIR) spectra of (**a**) PO-PAA and POP-PAA and (**b**) PO-PI and POP-PI; the peak deconvolution of the (**c**) O 1s and (**d**) N 1s X-ray photoelectron spectroscopy (XPS) spectra of the as-prepared PO-PI. **Figure 2.** The Fourier transform infrared (FTIR) spectra of (**a**) PO-PAA and POP-PAA and (**b**) PO-PI and POP-PI; the peak deconvolution of the (**c**) O 1s and (**d**) N 1s X-ray photoelectron spectroscopy (XPS) spectra of the as-prepared PO-PI.

The microscopic molecular packing state of the PO- and POP-PI films was investigated by XRD measurement. As shown in Figure 3a, the XRD pattern of PO-PI is blunt without obvious diffraction signals, suggesting its amorphous nature. The amorphous state of

**Figure 3.** (**a**) The XRD patterns of PO-PI and POP-PI; (**b**) the TGA profile of the PO-PI and POP-PI.

PO-PI originated from the flexible ether linkage within the ODA unit which loosens the chain packing of the PO-PI. In contrast, incorporation of the *p*-PDA unit in the polymer backbone can significantly enhance the polymer chain stacking. The regularly ordered structure of POP-PI is attributed to the rigid and planar skeletal structure of *p*-PDA leading to the better crystallinity [33]. Figure 3b compares the TGA curves of the PIs to investigate the thermal stability. Both the PI samples exhibit a major thermal decomposition ranging from 530 to 720 ◦C. The 5 wt% thermal decomposition temperature (*T*d) of the POP-PI is found to be 584 ◦C, which is higher than that of PO-PI (576 ◦C), suggesting its better thermal stability of POP-PI. The improved thermal stability of POP-PI can be interpreted by the presence of the *p*-PDA group, which increases the intra- and interpolymer chain interactions, resulting in tight polymer chain packaging. **Figure 2.** The Fourier transform infrared (FTIR) spectra of (**a**) PO-PAA and POP-PAA and (**b**) PO-PI and POP-PI; the peak deconvolution of the (**c**) O 1s and (**d**) N 1s X-ray photoelectron spectroscopy (XPS) spectra of the as-prepared PO-PI.

*Polymers* **2021**, *13*, x FOR PEER REVIEW 8 of 13

**Figure 3. Figure 3.**  ( (**aa** ) The XRD patterns of PO-PI and POP-PI; ( ) The XRD patterns of PO-PI and POP-PI; (**b b** ) the TGA profile of the PO ) the TGA profile of the PO-PI and POP-PI. -PI and POP-PI.

Figure 4a displays the XRD pattern of the carbon materials derived from PO- and POP-PI. The resultant carbon materials reveal two peaks located at 23.5 and 43.8◦ , indicating the development of (002) and (100) planes, respectively [34]. These results suggest that both the two PI samples can be converted into the hexagonal structures of the carbonized materials after the thermal treatment. Moreover, the POP-PI derived carbon exhibits much stronger XRD peaks than that derived from PO-PI, indicating its higher degree of graphitization. Figure 4b shows the Raman spectra of the two carbonized materials to further provide atomic-scale structural information. The Raman spectrum of the two samples shows the characteristic D band (1342 cm−<sup>1</sup> ) and G band (1582 cm−<sup>1</sup> ), which correspond to the sp<sup>3</sup> and sp<sup>2</sup> carbon, respectively [35,36]. In general, the intensity ratio between the D band and G band (ID/IG) can be used to determine the graphitization degree of the carbonized materials. The ID/I<sup>G</sup> values of the PO- and POP-PI derived carbons are calculated to be 1.04 and 0.75, respectively. The lower ID/I<sup>G</sup> ratio of POP-PI derived carbon also suggests its more ordered structure and the increased sp<sup>2</sup> content. The C 1s XPS spectra of the POand POP-PI derived carbons are shown in Figure 4c,d to further characterize their chemical composition. Both the samples show four peaks with binding energies of 284.5, 285.4, 286.5 and 288.0 eV, which correspond to the functional groups of C=C, C–C, C–O and C=O, respectively. Based on the XPS analysis, the proportions of C=C for PO- and POP-PI derived carbons are 63.3% and 66.4%, respectively. Conversely, the proportion of oxygen-containing functional groups (C–O and C=O) of POP-PI based carbon (6.98%) is lower than that of the PO-PI based one (10.77%). In addition, the electronic conductivity of the PI derived carbons was directly measured with a four-point probe. The values of PO-PI and POP-PI derived carbons are 89.5 and 100.1 S/cm, respectively. These results further confirm that it is crucial to ensure the increased level of the structural arrangement of the initial PI precursor to prepare high-quality carbon materials with graphite-like structure [37].

**Figure 4.** (**a**) The X-ray diffraction (XRD) pattern of the carbon materials derived from PO- and POP-PI; (**b**) Raman spectra of the carbon materials derived from PO- and POP-PI; peak deconvolution of the C 1s XPS spectra of (**c**) PO-PI derived carbon and (**d**) POP-PI derived carbon. **Figure 4.** (**a**) The X-ray diffraction (XRD) pattern of the carbon materials derived from PO- and POP-PI; (**b**) Raman spectra of the carbon materials derived from PO- and POP-PI; peak deconvolution of the C 1s XPS spectra of (**c**) PO-PI derived carbon and (**d**) POP-PI derived carbon.

The XRD pattern of the spray-dried LTO powders is exhibited in Figure 5a. All the peaks between 5 and 70◦ can be assigned to the spinel LTO structure without any impurity phases. The particle size distribution of the as-prepared LTO is shown in Figure 5b. The particle size of the LTO powders ranges between 1.88 and 27.4 µm with a mean particle size of 10.7 µm (*D*50). The micro-sized particles can be favorable for the powder packing during the electrode fabrication leading to better energy density. The surface area of the LTO was also evaluated by the Brunauer–Emmett–Teller (BET) analysis as shown in Figure 5c. Based on the nitrogen adsorption/desorption isotherms, the surface area of the LTO was calculated to be 11.3 m2/g. Figure 5d displays the Ti 2p core level XPS spectrum of the as-prepared LTO samples. There are two pairs of Ti 2p peaks observed at 464.3, 458.5 and 462.2, 456.4 eV for Ti4+ 2p1/2, Ti4+ 2p3/2 and Ti3+ 2p1/2, Ti3+ 2p3/2, respectively. The partial reduction of the Ti ions from Ti4+ to Ti3+ originates from the generation of oxygen vacancies during the thermal annealing in N<sup>2</sup> ambience [38]. The microstructure of the LTO was investigated by SEM observation. The low-magnification SEM examination as shown in Figure 5e reveals that the morphology of LTO sample is perfectly preserved as highly uniform microspheres. In addition, the enlarged SEM image (Figure 5f) shows that the surface of the microspheres is composed of a primary nanoparticle with an average size of around 80 nm leading to a porous surface. The porous structure can facilitate the ionic transport during the charge/discharge process.

The morphologies of the pristine and carbon-coated LTO are monitored by SEM and TEM as shown in Figure 6. It can be observed that the whole sample reveals similar SEM morphology, indicating the spherical structure can be maintained after the carbon coating process. Moreover, the high-resolution TEM images of the PO-LTO and POP-LTO show a carbon layer with a thickness of around 2 nm was uniformly deposited on the LTO surface. In contrast, no carbon layer can be observed for bare LTO. It has been reported that the carbon layer can offer a conductive pathway for the electron transport. In addition, the

randomly distributed defects and vacancies within the carbon layer also can improve the Li<sup>+</sup> ion migration [6]. As a result, the kinetic balance between the electronic and ionic transport can be established leading to better rate performance. The XPS survey of the POP-LTO and Raman spectrum of the three samples are also provided in the Supporting Information. *Polymers* **2021**, *13*, x FOR PEER REVIEW 10 of 13 *Polymers* **2021**, *13*, x FOR PEER REVIEW 10 of 13

**Figure 5.** (**a**) The XRD pattern of the spray-dried Li4Ti5O12 (LTO) powder; (**b**) the particle size distribution of the LTO; (**c**) the nitrogen adsorption/desorption isotherms of the LTO; (**d**) peak deconvolution of the Ti 2p XPS spectrum of LTO powder; (**e**,**f**) scanning electron microscopy (SEM) images of the LTO with different magnifications. **Figure 5.** (**a**) The XRD pattern of the spray-dried Li4Ti5O<sup>12</sup> (LTO) powder; (**b**) the particle size distribution of the LTO; (**c**) the nitrogen adsorption/desorption isotherms of the LTO; (**d**) peak deconvolution of the Ti 2p XPS spectrum of LTO powder; (**e**,**f**) scanning electron microscopy (SEM) images of the LTO with different magnifications. **Figure 5.** (**a**) The XRD pattern of the spray-dried Li4Ti5O12 (LTO) powder; (**b**) the particle size distribution of the LTO; (**c**) the nitrogen adsorption/desorption isotherms of the LTO; (**d**) peak deconvolution of the Ti 2p XPS spectrum of LTO powder; (**e**,**f**) scanning electron microscopy (SEM) images of the LTO with different magnifications.

**Figure 6.** The SEM morphology of (**a**) bare LTO; (**b**) PO-LTO and (**c**) POP-LTO; the transmission electron microscopy (TEM) image of (**d**) bare LTO; (**e**) PO-LTO and (**f**) POP-LTO. **Figure 6.** The SEM morphology of (**a**) bare LTO; (**b**) PO-LTO and (**c**) POP-LTO; the transmission electron microscopy (TEM) image of (**d**) bare LTO; (**e**) PO-LTO and (**f**) POP-LTO. **Figure 6.** The SEM morphology of (**a**) bare LTO; (**b**) PO-LTO and (**c**) POP-LTO; the transmission electron microscopy (TEM) image of (**d**) bare LTO; (**e**) PO-LTO and (**f**) POP-LTO.

The electrochemical properties of the bare and carbon-coated LTO samples were studied by measuring their CV profiles with a scan rate of 1 mV/s between 1.0 and 2.5 V as

shown in Figure 7a. All the samples reveal a pair of sharp redox couple which corresponds to the transition between Li4Ti5O<sup>12</sup> and Li7Ti5O12. The separation between the anodic and cathodic peaks (∆E) of the LTO would be dramatically reduced with the incorporation of carbon coating, indicating the reduction of polarization. The ∆E of LTO, PO-LTO and POP-LTO is found to be 0.7, 0.63 and 0.52 V, respectively. The lower polarization is mainly attributed to the enhancement of electric conductivity by the interfacial carbon modification. As shown in Figure 4, the POP based carbon layer shows a more regular graphite structure with a higher conductivity (100.1 S/cm) than that of PO based carbon layer (89.5 S/cm). As a result, the POP-LTO reveals a much lower polarization (0.52 V) due than that of PO-LTO (0.63 V). Furthermore, the EIS spectra of the three electrodes were taken to investigate the interfacial impedance in the electrodes. Figure 7b shows the Nyquist plots of LTO, PO-LTO and POP-LTO electrodes with a frequency range from 10<sup>5</sup> Hz to 10−<sup>2</sup> Hz at an amplitude of 10 mV. All the plots exhibit a semicircle with a sloping line. The charge transfer resistance (*R*ct) determined from the size of the semicircle in the high frequency is around 117, 43.7 and 21.5 Ω for LTO, PO-LTO and POP-LTO, respectively. The *R*ct of POP-LTO is much lower than that of the bare one due to the deposition of the conductive carbon layer. The rate performance of the three electrodes evaluated at various C rates is displayed in Figure 7c. As expected, the POP-LTO with the modification of the POP-PI derived carbon layer can deliver the best rate capability due to its low polarization and *R*ct. The POP-LTO shows a high capacity retention of 83.2% (137.5 mAh/g) at a high rate of 20 C, which is higher than 70.7% (115.3 mAh/g) and 58.5% (94.4 mAh/g) for PO-LTO and pristine LTO, respectively. The corresponding charge/discharge profiles of the POP-LTO with different C rates are also provided in Figure 7d. The comparison of LTO performance with different carbon coating is also summarized in the Supporting Information. *Polymers* **2021**, *13*, x FOR PEER REVIEW 11 of 13

**Figure 7.** (**a**) The cyclic voltammetry (CV) curves of bare LTO, PO-LTO and POP-LTO stepped between 1.0 and 2.5 V with a scan rate of 1 mV/s; (**b**) Nyquist plots of bare LTO, PO-LTO and POP-LTO in the frequency range of 10<sup>5</sup> Hz to 10−2 Hz; (**c**) rate capability of the as-prepared samples at C rates between 0.1 C and 20 C; (**d**) the corresponding charge/discharge profile of the POP-LTO with various C rates. **Figure 7.** (**a**) The cyclic voltammetry (CV) curves of bare LTO, PO-LTO and POP-LTO stepped between 1.0 and 2.5 V with a scan rate of 1 mV/s; (**b**) Nyquist plots of bare LTO, PO-LTO and POP-LTO in the frequency range of 10<sup>5</sup> Hz to 10−<sup>2</sup> Hz; (**c**) rate capability of the as-prepared samples at C rates between 0.1 C and 20 C; (**d**) the corresponding charge/discharge profile of the POP-LTO with various C rates.

trinsic electrical conductivity and poor Li<sup>+</sup> diffusion coefficient.

agreed to the published version of the manuscript. **Funding:** This research received no external funding.

In summary, a facile and scalable method has been developed for synthesizing car-

nano-carbon layer on LTO can improve both the electronic and ionic conductivities. As compared to pristine LTO, results show that the POP-LTO exhibits substantially improved cell performances particularly in the rate capability. A high initial discharge capacity of 165.1 mAh/g is delivered at 0.1 C and the specific capacity still maintain 137.5 mAh/g even at 20 C. It is believed that the developed method also can be extended to improve the electrochemical performances of other alternative LIB materials with low in-

**Author Contributions:** writing—original draft preparation, S.-C.H.; writing—review and editing, T.-T.H. and T.-Y.L.; resources, Y.-J.W.; resources, C.-Z.L.; supervision, H.C.W.; conceptualization, J.-H.H.; investigation, C.-W.C.-J. and T.-Y.L.; funding acquisition, T.-Y.L. All authors have read and

**4. Conclusions**

#### **4. Conclusions**

In summary, a facile and scalable method has been developed for synthesizing carboncoated LTO, using PI as the carbon source. After carbonization treatment at 800 ◦C in N<sup>2</sup> ambience, the PI coating can be transferred into the high-conductive carbon layer. The nano-carbon layer on LTO can improve both the electronic and ionic conductivities. As compared to pristine LTO, results show that the POP-LTO exhibits substantially improved cell performances particularly in the rate capability. A high initial discharge capacity of 165.1 mAh/g is delivered at 0.1 C and the specific capacity still maintain 137.5 mAh/g even at 20 C. It is believed that the developed method also can be extended to improve the electrochemical performances of other alternative LIB materials with low intrinsic electrical conductivity and poor Li<sup>+</sup> diffusion coefficient.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/article/10 .3390/polym13111672/s1, Figure S1: The XPS survey of the POP-LTO and the corresponding atomic percentages, Figure S2: Raman spectrum of the pristine LTO, PO-LTO and POP-LTO, Table S1: The comparison of LTO performance with different carbon coating.

**Author Contributions:** Writing—original draft preparation, S.-C.H.; writing—review and editing, T.-T.H. and T.-Y.L.; resources, Y.-J.W.; resources, C.-Z.L.; supervision, H.C.W.; conceptualization, J.-H.H.; investigation, C.-W.C.-J. and T.-Y.L.; funding acquisition, T.-Y.L. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Acknowledgments:** We are grateful to the Ministry of Science and Technology (MOST 104-2113-M-152-001-MY2, MOST 105-2320-B-038-014-) and CPC Corporation (105-3011). for financial support. We would also like to thank the research funding from Taipei Medical University (TMU102-AE1-B02, TMUTOP103004-2). The research endeavors at Ming Chi University of Technology were supported in part by the Ministry of Science and Technology (MOST 103-2221-E-131-002-MY2), by the Chang Gung Memorial Hospital, Linkou, Taiwan (CMRPG3E2091), and by the Academia Sinica Research Project on Thematic Project (AS-104-TP-A11).

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


## **Superparamagnetic, High Magnetic** α**-Fe &** α 00**-Fe16N<sup>2</sup> Mixture Prepared from Inverse Suspension-Polymerized Fe3O4@polyaniline Composite**

**Yen-Zen Wang <sup>1</sup> , Yu-Wei Cheng <sup>2</sup> , Lin-Chia Ho <sup>3</sup> , Wen-Yao Huang 4,\* ,† , Ko-Shan Ho 5,\* ,† and Yu-Ting Syu <sup>5</sup>**


**Abstract:** Oleic acid (OA)-modified Fe3O<sup>4</sup> nanoparticles were successfully covered with polyanilines (PANIs) via inverse suspension polymerization in accordance with SEM and TEM micrographs. The obtained nanoparticles were able to develop into a ferrite (α-Fe) and α 00-Fe16N<sup>2</sup> mixture with a superparamagnetic property and high saturated magnetization (SM) of 245 emu g−<sup>1</sup> at 950 ◦C calcination under the protection of carbonization materials (calcined PANI) and other iron-compounds (α 00-Fe16N<sup>2</sup> ). The SM of the calcined iron-composites slightly decreases to 232 emu g−<sup>1</sup> after staying in the open air for 3 months. The calcined mixture composite can be ground into homogeneous powders without the segregation of the iron and carbon phases in the mortar without significantly losing magnetic activities.

**Keywords:** polyaniline; ferrite; α 00-Fe16N<sup>2</sup> ; superparamagnetic; inverse suspension polymerization

#### **1. Introduction**

Recently, with the advance of technology, the requirement for fast and wireless charging becomes more and more urgent. In particular, 3C mobile electronic products need to charge within several hours using portable power sources that are usually heavy and take one or two to recharge. Therefore, we also need a fast wireless power source to solve the problem. The effective charging distance of wireless charging [1] and large-scale charging development [2], thus, becomes very important. Wireless charging can be carried out by absorbing electromagnetic (EM) waves that emit from the power generators. Additionally, we also require a stable EM wave absorbing material to transform it into power, which can be fulfilled by superparamagnetic materials with high magnetization. There are other interesting application fields related to the calcined carbon (polyaniline) coated magnetic particles [3–6] such as nanoelectronics, catalysis, optical application, biosensors, environmental remediation, energy, hydrogen storage, drug transport, magnetic resonance imaging and cancer diagnosis [7]. Metamaterial theory is also used for the applications of the magnetic materials [8].

The regular material that meets these requirements is magnetite (Fe3O4), which can be obtained from the sol-gel method [9] conducted at room temperature. Although it demonstrates a superparamagnetic property, its low magnetization (around 100 emu g−<sup>1</sup> ) shortens the wireless charging distance. Other possible candidates with a higher magnetization than Fe3O<sup>4</sup> are cementite (Fe3C) [10], ferric nitride (FexNy) [11,12] and ferrite

**Citation:** Wang, Y.-Z.; Cheng, Y.-W.; Ho, L.-C.; Huang, W.-Y.; Ho, K.-S.; Syu, Y.-T. Superparamagnetic, High Magnetic α-Fe & α00-Fe16N<sup>2</sup> Mixture Prepared from Inverse Suspension-Polymerized Fe3O4@polyaniline Composite. *Polymers* **2021**, *13*, 2380. https://doi.org/10.3390/ polym13142380

Academic Editor: Patrick Ilg

Received: 11 June 2021 Accepted: 15 July 2021 Published: 20 July 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

(α-Fe) [13]; they all demonstrate saturated magnetization over 150 emu g−<sup>1</sup> and even over 200 emu g−<sup>1</sup> for α-Fe and α <sup>00</sup>-Fe16N2. However, α-Fe, which is the pure Fe element, is vulnerable to the O<sup>2</sup> in the open air to become iron oxide.

Consequently, we need to provide some solid protection on the α-Fe, and α 00-Fe16N<sup>2</sup> itself can also provide additional protection to keep it from oxidation and maintain its high magnetization in the air.

Usually, PANIs that provide carbon and nitrogen sources can be prepared in affluent water [14–18] easily using water-soluble anilinium monomers. The Fe3O<sup>4</sup> particles prepared from the sol-gel method are usually hydrophilic materials with some –OH groups on the surfaces. If it stays with anilinium monomers in the water, the prepared PANI molecules cannot cover most of the Fe3O<sup>4</sup> particles that become very mobile in the water [19–26] due to the hydrophilicity. The obtained PANIs are not able to protect Fe3O<sup>4</sup> particles during calcination, since most of them will stay on the surface of PANI molecules, not inside.

Inverse suspension polymerization (ISP) is usually applied to cover functional particles with polymers that can be polymerized in the water phase [27–31]. In this study, we are trying to prepare the protecting Fe3O<sup>4</sup> particles by the ISP of PANIs. The sol-gel-prepared Fe3O<sup>4</sup> particles usually own surface –OH, which can be esterificated with OA to attach some hydrophobic tails to the particles [32,33] and allow them to stably stay inside the micelles in the toluene system. Moreover, the additional aliphatic tails that come from the attaching OA can also stabilize the micelles or polymer droplets during the polymerization of anilines. In other words, hydrophilic OA-modified Fe3O<sup>4</sup> (Fe3O4(OA)) was first mixed with the anilinium monomers in the water and became micelles in the hydrophobic toluene solvents after stirring. Eventually, the Fe3O4(OA) particles can be surrounded with long polyaniline molecules after water soluble initiators such as APS (ammonium persulfate) are introduced. The composites prepared via ISP can then be subject to calcination in the argon to transform into other iron-compounds with high magnetization.

#### **2. Experimental**

#### *2.1. Preparation*

## 2.1.1. Synthesis of Fe3O4(OA)

In a beaker, 7.08 g of ferric chloride hexahydrate (FeCl3·6H2O, J. T. Baker, NJ, USA) and 2.58 g of ferrous chloride tetrahydrate (FeCl2·4H2O, J. T. Baker, NJ, USA) were mixed with 40 mL of deionized water by a magnetic stirrer. The homogenized solution was transferred to a round bottom three-necked flask equipped with a water condenser in one of the mouths. One of the two remaining mouths was purged with high-purity nitrogen to prevent the oxidation of the reaction mixture at a high temperature, the other one behaved as the exhaust release outlet. The temperature of the reaction solution was ramped up to 80 ◦C in a silicone oil bath and kept with purging nitrogen for ten minutes. Then, 2 mL of OA (Hitachi Astemo Ltd, Tokyo, Japan) and some ammonia water (Fisher Sci., Bridgewater, NJ, USA) was introduced to tune the solution to become alkaline and then the reaction was started. The reaction continued for 30 min and was finished by stopping the stirring of the magnetic stirrer, followed by attaching a powerful magnet on the bottom of the reactor to separate the magnetic precipitate. The precipitate was washed several times with deionized water, and the clear, upper layer of the solution was discarded. The isolated black precipitate was placed in an ultrasonic oscillator for 20 min and then dried in an oven for 12 h at 60 ◦C, the Fe3O4(OA) was available.

## 2.1.2. Synthesis of PANI/Fe3O4(OA) Nanocomposite

An amount of 3 g (0.091 mol) of n-dodecylbenzenesulfonic acid (DBSA: Tokyo Kasei Kogyo Co., Tokyo, Japan) was dissolved in 50 mL of de-ionized water, the mixture was slowly stirred until a homogeneous solution was obtained, followed by the addition of 9 g (0.0968 mol) of aniline monomer (Tokyo Kasei Kogyo Co., Tokyo, Japan) and the solution was stirred to be clear. Eventually, the Fe3O4(OA) obtained from the previous experiment was added and the mixture was stirred again to become homogeneous [32,33].

A comparison emulsion polymerization prepared PANI(EB)/Fe3O4(OA) [29–31] was obtained in water in the absence of toluene. The resultant composite was named as PANI(EB)-Em/Fe3O4(OA).

#### 2.1.3. Calcination of PANI Nanocomposites

PANI(EB)/Fe3O4(OA) prepared in Section 2.1.2 was calcined in a tube furnace, ramping up from RT to 600–950 ◦C at most and staying for 30 min in the argon atmosphere. The obtained N, C-doped iron composites are named as FeNCs. When samples were kept at room temperature for two or three months in the air, they are named as FeNC-2 and FeNC-3, respectively. A sample prepared at 600 ◦C is named FeNC-600, etc.

#### *2.2. Characterization*

#### 2.2.1. Fourier Transform Infrared Spectroscopy (FTIR)

The main functional groups of neat Fe3O4(OA) and calcinated FeNC were assigned in accordance with the FTIR spectra recorded on an IFS3000 v/s FTIR spectrometer (Bruker, Ettlingen, Germany) at room temperature with a resolution of 4 cm−<sup>1</sup> and 16 scanning steps.

#### 2.2.2. Ultraviolet and Visible, Near-IR Spectroscopy (UV–Vis–NIR)

The UV–Vis–NIR spectra of the PANI(ES) (PANI without dedoping by NH4OH(aq)) in the PANI(ES)/Fe3O4(OA) nanocomposites were obtained from a Hitachi U-2001 and DTS-1700 NIR Spectrometer (Nicosia, Cyprus). The wavelength ranged from 300 to 1600 nm.

#### 2.2.3. TGA (Thermogravimetric Analysis)

The mass loss percentages of neat PANI(EB) and PANI(EB)/Fe3O4(OA) upon calcination (thermal degradation) were monitored and recorded using TGA (TA SDT-2960, New Castle, DE, USA) thermograms.

#### 2.2.4. Scanning Electron Microscopy (SEM)

The sizes and morphologies of neat Fe3O4(OA), non-calcinated PANI(EB)/Fe3O4(OA), and calcinated FeNCs were characterized using SEM (field emission gun scanning electron microscope, AURIGAFE, Zeiss, Oberkochen, Germany).

#### 2.2.5. Transmission Electron Microscopy (TEM)

Samples, of which photos were taken using the field emission transmission electron microscope, HR-AEM (HITACHI FE-2000, Hitachi, Tokyo, Japan), were first dispersed in acetone and put on carbonic-coated copper grids dropwise before subjecting to the emission.

#### 2.2.6. Raman Spectroscopy

The Raman spectra of calcinated PANI(EB)s and FeNCs treated at different temperatures were obtained from a Raman spectrometer (TRIAX 320, HOBRIA, Kyoto, Japan).

#### 2.2.7. Powder X-ray Diffraction (Powder XRD)

A copper target (Cu-Kα) Rigaku X-ray source (Rigaku, Tokyo, Japan) generating X-ray with a wavelength of 1.5402 Å after electron bombarding was used to create the diffraction patterns of neat Fe3O4(OA) and FeNCs. The scanning angle (2θ) that ranged from 10◦ to 70◦ with a voltage of 40 kV and a current of 30 mA, operated at 1◦ min−<sup>1</sup> .

#### 2.2.8. X-ray Photoelectron Spectroscopy (XPS)

The binding energy spectra of Fe 2p of FeNCs treated at different temperatures were used to characterize the characteristic crystallization planes of α–Fe, Fe3C, FeNx and, Fe3O<sup>4</sup> after calcination, and were obtained from an XPS instrument of Fison (VG)-Escalab 210 (Fison, Glasgow, UK) using an Al Ka X-ray source at 1486.6 eV. The pressure in the chamber maintained was under 10−<sup>6</sup> Pa or lower during performance. Tablet samples were prepared by pressing in a stapler with a ring mold.

#### 2.2.9. Superconductor Quantum Interference Device (SQUID)

The paramagnetic properties of neat Fe3O4(OA) and various FeNCs were measured from a SQUID of Quantum Design MPMS-XL7 (San Diego, CA, USA)

#### **3. Results and Discussion**

#### *3.1. FTIR Spectra*

The hydrophilicity of Fe3O4(OA) nanoparticles comes from the hydroxylated surfaces created during the sol-gel process, assigned at ~3300 cm−<sup>1</sup> according to Figure 1. Some of the hydroxyl groups are still present after esterification with OA, as described in Figure 1 as well, indicating that some of the –OH groups remained intact after esterification and polymerization. The remaining hydrophilicity of the nanocomposite makes it still dispersible in micelles before the polymerization mixing with anilinium monomers or staying in polymer droplets after polymerization, which was randomly dispersed in the hydrophobic toluene matrix. The polymerization of anilinium monomers after the addition of water-soluble initiator APS in the presence of Fe3O4(OA) did not destroy the carbonyl groups of the ester that link the OA onto the Fe3O<sup>4</sup> surface of the nanoparticles either, illustrating that the long hydrocarbon tails of the OA are still firmly connected to the Fe3O<sup>4</sup> nanoparticle's surface. It provided the nanocomposites with some hydrophobicity and affinity to the toluene and the large polymer droplets were still able to suspend in the solvent after polymerization. The polymer droplets need to de-emulsify with acetone before collecting the nanocomposite products through filtration. The symmetric and asymmetric stretching mode of aliphatic methylene and the methyl groups of OA tails assign at 2920 and 2840 cm−<sup>1</sup> , respectively. The additional peak around 587 cm−<sup>1</sup> reveals the presence of the Fe–O bonding of both the neat Fe3O4(OA) and the nanocomposite, revealing that Fe3O4(OA) nanoparticles are staying inside the nanocomposite, even after de-emulsification and filtration. The assignments of the main functional groups of polyaniline and Fe3O4(OA) are listed in Table 1 and illustrated in Figure 1. *Polymers* **2021**, *13*, x FOR PEER REVIEW 5 of 19 **Table 1.** Assignments of FTIR spectra. –OH 3400 cm−<sup>1</sup> –CH3, –CH2 (stretching) 2900, 3000 cm−<sup>1</sup> –C=O 1750 cm−<sup>1</sup> –C=C– 1650 cm−<sup>1</sup> Quinoid ring 1570 cm−<sup>1</sup> Benzoid ring 1476, cm−<sup>1</sup> –CH3, –CH2 (bending) 1400, 1350 cm−<sup>1</sup> –C–N– 1307 cm−1 –C–OH 1250 cm−<sup>1</sup> –B–NH–B– 1135 cm−<sup>1</sup> Fe–O 590 cm−<sup>1</sup>

**Figure 1.** FTIR-spectra of neat Fe3O4(OA) and PANI(EB)/Fe3O4(OA) composite. **Figure 1.** FTIR-spectra of neat Fe3O<sup>4</sup> (OA) and PANI(EB)/Fe3O<sup>4</sup> (OA) composite.

*3.2. λmax of PANI in the Nanocomposite Obtained from UV–Vis–NIR Spectra* 

If most of the Fe3O4(OA) in the nanocomposite is completely covered or embedded in the PANI matrix, it can also induce the shifting of the λmax of PANI in the UV–Vis spectra due to the interaction. Figure 2 illustrates the flattening effect on the λmax in the NIR region, the so-called free carrier-tail [14–18] in this region for the PANI(ES)-Em prepared via emulsion polymerization. It also represents a nanofibrous morphology that contrib-

accordance with Figure 2, the PANI(ES) prepared via ISP does not demonstrate any free carrier-tail, but the significant λmax peak and the morphology is not fibrous either, which will be confirmed in the TEM micropicture. The λmax of PANI(ES) prepared by ISP is around 860 nm, which is also in the near-IR region, indicating that its conjugation chain length is still long, and the chain is still extended. After the Fe3O4(OA) nanoparticles were introduced before the beginning of the polymerization, the λmax of PANI(ES) blue-shift from 860 to 800 nm with increasing Fe3O4(OA) nanoparticles and curve became more bent. The blue shift and the presence of the bended curves of the nanocomposites reflected the shortening of the conjugation of the PANI molecules whose extensions were slightly recoiled. Since the λmax is still high at 810 nm, a random-coil morphology (λmax = 780 nm) is not thought to be present. The attracting force that causes the recoiling of the PANI(ES) molecules is believed to stem from the formation of the H-bonding between the left –OH


**Table 1.** Assignments of FTIR spectra.

The obtained PANI(ES)/Fe3O4(OA) nanocomposites were then dedoped into PANI(EB)/ Fe3O4(OA) in NH4OH(aq) and its FTIR-spectrum is also demonstrated in Figure 1. The feature functional groups of PANI(EB) can be found for PANI(EB)/Fe3O4(OA) in Table 1 and Figure 1, revealing that the polymerization did not change the characteristic functional groups of PANI in the presence of Fe3O4(OA).

## *3.2. λmax of PANI in the Nanocomposite Obtained from UV–Vis–NIR Spectra*

If most of the Fe3O4(OA) in the nanocomposite is completely covered or embedded in the PANI matrix, it can also induce the shifting of the λmax of PANI in the UV–Vis spectra due to the interaction. Figure 2 illustrates the flattening effect on the λmax in the NIR region, the so-called free carrier-tail [14–18] in this region for the PANI(ES)-Em prepared via emulsion polymerization. It also represents a nanofibrous morphology that contributes to the carrier-tail due to the highly extended conjugation chain length. However, in accordance with Figure 2, the PANI(ES) prepared via ISP does not demonstrate any free carrier-tail, but the significant λmax peak and the morphology is not fibrous either, which will be confirmed in the TEM micropicture. The λmax of PANI(ES) prepared by ISP is around 860 nm, which is also in the near-IR region, indicating that its conjugation chain length is still long, and the chain is still extended. After the Fe3O4(OA) nanoparticles were introduced before the beginning of the polymerization, the λmax of PANI(ES) blue-shift from 860 to 800 nm with increasing Fe3O4(OA) nanoparticles and curve became more bent. The blue shift and the presence of the bended curves of the nanocomposites reflected the shortening of the conjugation of the PANI molecules whose extensions were slightly recoiled. Since the λmax is still high at 810 nm, a random-coil morphology (λmax = 780 nm) is not thought to be present. The attracting force that causes the recoiling of the PANI(ES) molecules is believed to stem from the formation of the H-bonding between the left –OH groups and the amino groups of PANI(ES). The blue-shifting is enhanced when more Fe3O4(OA) is introduced, referring to Figure 2. The ISP approach effectively polymerizes the anilinium monomers inside micelles where lots of Fe3O4(OA) nanoparticles are already present. It is believed that some of the H-bonds are already formed before the addition of water-soluble APS initiator and most of the Fe3O4(OA) nanoparticles inside micelles can be surrounded and protected by both monomers before polymerization and polymers after polymerization, which is described in Scheme 1. The presence of the long aliphatic/long tail of Fe3O4(OA) can improve the stability of the micelles in the toluene solvents as well and less CTAB is necessary to create the inverse micelle (W/O). Moreover, the hydrophobic tails of Fe3O4(OA) are extended to the toluene phase and can also immobilize the Fe3O<sup>4</sup> inside water micelles, allowing the growing PANI molecules to entangle around them.

around them.

groups and the amino groups of PANI(ES). The blue-shifting is enhanced when more Fe3O4(OA) is introduced, referring to Figure 2. The ISP approach effectively polymerizes the anilinium monomers inside micelles where lots of Fe3O4(OA) nanoparticles are already present. It is believed that some of the H-bonds are already formed before the addition of water-soluble APS initiator and most of the Fe3O4(OA) nanoparticles inside micelles can be surrounded and protected by both monomers before polymerization and polymers after polymerization, which is described in Scheme 1. The presence of the long aliphatic/long tail of Fe3O4(OA) can improve the stability of the micelles in the toluene solvents as well and less CTAB is necessary to create the inverse micelle (W/O). Moreover, the hydrophobic tails of Fe3O4(OA) are extended to the toluene phase and can also immobilize the Fe3O4 inside water micelles, allowing the growing PANI molecules to entangle

When PANI is prepared via common emulsion polymerization, most of the Fe3O4(OA) particles will randomly distribute in the water and the fibrous PANI is created inside the micelles where less water is present. Eventually, the obtained nanofibrous PANI molecules can only attract lots of Fe3O4(OA) particles on the surface through Hbonding, which will be seen in the SEM micropicture. The Fe3O4(OA)-covered fibrous PANI surely cannot form a stable/high magnetic FeNC compound after calcination, since the formed high magnetic materials cannot be protected by the PANI during calcination.

**Figure 2. Figure 2.**  UV–Vis–NIR spectra of PANI(ES) and various PANI(ES)/Fe UV–Vis–NIR spectra of PANI(ES) and various PANI(ES)/Fe3O4(OA) composites. <sup>3</sup>O<sup>4</sup> (OA) composites.

**Scheme 1.** Preparing diagram of FeNCs via inverse suspension polymerization and calcination. **Scheme 1.** Preparing diagram of FeNCs via inverse suspension polymerization and calcination.

*3.3. TGA Thermogram* The mass loss and rate of the calcination at a high temperature can be monitored by the weights vs. temperatures in the thermogram demonstrated in Figure 3. The neat PANI(EB) experienced significant weight loss after 400 °C, which originated from the When PANI is prepared via common emulsion polymerization, most of the Fe3O4(OA) particles will randomly distribute in the water and the fibrous PANI is created inside the micelles where less water is present. Eventually, the obtained nanofibrous PANI molecules can only attract lots of Fe3O4(OA) particles on the surface through H-bonding, which will be seen in the SEM micropicture. The Fe3O4(OA)-covered fibrous PANI surely cannot form

crosslinking of the neighboring molecules [18], and the weight loss continued gradually until 600 °C, when the crosslinking finished, and carbonization started. Almost no mass

temperature chosen was 600 °C, after which the weight loss was entirely contributed from the PANI(EB)-covered iron-compounds not from PANI(EB) only, in accordance with Figure 3. If we check the difference of the residue weights of neat PANI(EB) after 600 °C and weight of PANI(EB)/Fe3O4(OA) at 600 °C from Figure 3 and Table 2, it is 22.4 wt% (35.9% − 13.5% = 22.4%). It means there is about 22.4 wt% of iron-compound in the composite, which is thermally stable until 600 °C. There is only 1.7 wt% loss (37.5% − 35.8% = 1.7%) after 600 °C for the composite, as seen in the inset of Figure 3, which is all contributed from the thermal degradation of the FeNCs from 600 to 950 °C. The actual degrading % for FeNC is around 1.7/22.4 ≈ 7.6 wt%. The composition variation and the types of atoms (Fe, N, C, or O) that are lost during the calcination of the FeNCs from 600 to 950 °C are strongly related to the magnetic activity, which will be characterized from the powder X-

ray diffraction patterns, XPS, and SQUID spectra.

a stable/high magnetic FeNC compound after calcination, since the formed high magnetic materials cannot be protected by the PANI during calcination.

#### *3.3. TGA Thermogram*

The mass loss and rate of the calcination at a high temperature can be monitored by the weights vs. temperatures in the thermogram demonstrated in Figure 3. The neat PANI(EB) experienced significant weight loss after 400 ◦C, which originated from the crosslinking of the neighboring molecules [18], and the weight loss continued gradually until 600 ◦C, when the crosslinking finished, and carbonization started. Almost no mass loss occurred after 600 ◦C for PANI(EB), which is the reason why the lowest calcination temperature chosen was 600 ◦C, after which the weight loss was entirely contributed from the PANI(EB)-covered iron-compounds not from PANI(EB) only, in accordance with Figure 3. If we check the difference of the residue weights of neat PANI(EB) after 600 ◦C and weight of PANI(EB)/Fe3O4(OA) at 600 ◦C from Figure 3 and Table 2, it is 22.4 wt% (35.9% − 13.5% = 22.4%). It means there is about 22.4 wt% of iron-compound in the composite, which is thermally stable until 600 ◦C. There is only 1.7 wt% loss (37.5% − 35.8% = 1.7%) after 600 ◦C for the composite, as seen in the inset of Figure 3, which is all contributed from the thermal degradation of the FeNCs from 600 to 950 ◦C. The actual degrading % for FeNC is around 1.7/22.4 ≈ 7.6 wt%. The composition variation and the types of atoms (Fe, N, C, or O) that are lost during the calcination of the FeNCs from 600 to 950 ◦C are strongly related to the magnetic activity, which will be characterized from the powder X-ray diffraction patterns, XPS, and SQUID spectra. *Polymers* **2021**, *13*, x FOR PEER REVIEW 8 of 19

**Figure 3.** TGA thermograms of PANI(EB) and PANI(EB)/Fe3O4(OA). **Figure 3.** TGA thermograms of PANI(EB) and PANI(EB)/Fe3O<sup>4</sup> (OA).

**Table 2.** Various properties of FeNCs vs. temperature. **Table 2.** Various properties of FeNCs vs. temperature.


(a) Saturation magnetization at various temperatures obtained from SQUID; (b) Weight percentages at various temperatures obtained from TGA thermograms; (c) Intensity of D-band over G-band at various temperatures, obtained from Raman (a) Saturation magnetization at various temperatures obtained from SQUID; (b) Weight percentages at various temperatures obtained from TGA thermograms; (c) Intensity of D-band over G-band at various temperatures, obtained from Raman spectra.

spectra.

chains of OA and become the huge cake-like morphology in Figure 4a after being treating with OA via esterification. The PANI(EB) prepared in the presence of Fe3O4(OA) particles via common emulsion polymerization demonstrates rib-like, juxtaposed nanofibers fully covered with lots of Fe3O4(OA) particles on their surfaces in Figure 4b, which is commonly seen in the traditional emulsion polymerization of PANI [31,34]. These exposed Fe3O4(OA) particles on the fibrous PANI(EB) certainly can transform or convert to other iron-compounds with higher magnetic activity. However, the formed iron-compounds would be directly exposed to the O2 in the atmosphere at RT without any protection. Therefore, these iron-compounds would easily oxidize and recover to iron-oxide, whose magnetic force is far below that of the FeCx, FeNx, or α-Fe obtained after calcination. Consequently, an ISP system is designed to encapsulate the un-protected Fe3O4(OA) particles

*3.4. SEM Micropicture* 

#### *3.4. SEM Micropicture*

Unlike the neat Fe3O<sup>4</sup> particles that demonstrate a pearl-like morphology, the Fe3O4(OA) particles are found to be able to coagulate by the interactive long, aliphatic chains of OA and become the huge cake-like morphology in Figure 4a after being treating with OA via esterification. The PANI(EB) prepared in the presence of Fe3O4(OA) particles via common emulsion polymerization demonstrates rib-like, juxtaposed nanofibers fully covered with lots of Fe3O4(OA) particles on their surfaces in Figure 4b, which is commonly seen in the traditional emulsion polymerization of PANI [31,34]. These exposed Fe3O4(OA) particles on the fibrous PANI(EB) certainly can transform or convert to other iron-compounds with higher magnetic activity. However, the formed iron-compounds would be directly exposed to the O<sup>2</sup> in the atmosphere at RT without any protection. Therefore, these iron-compounds would easily oxidize and recover to iron-oxide, whose magnetic force is far below that of the FeCx, FeNx, or α-Fe obtained after calcination. Consequently, an ISP system is designed to encapsulate the un-protected Fe3O4(OA) particles with PANI(EB) that can develop into a dense, strong protecting carbon layer after calcination at high temperature through crosslinking and carbonization. *Polymers* **2021**, *13*, x FOR PEER REVIEW 10 of 19

**Figure 4.** SEM micropictures of (**a**) Fe3O4(OA), (**b**) PANI(EB)-Em/Fe3O4(OA), (**c**) PANI(EB)/Fe3O4(OA), (**d**) FeNC-650, (**e**) FeNC-700, (**f**) FeNC-800, (**g**) FeNC-900, (**h**) FeNC-950.  **Figure 4.** SEM micropictures of (**a**) Fe3O<sup>4</sup> (OA), (**b**) PANI(EB)-Em/Fe3O<sup>4</sup> (OA), (**c**) PANI(EB)/Fe3O<sup>4</sup> (OA), (**d**) FeNC-650, (**e**) FeNC-700, (**f**) FeNC-800, (**g**) FeNC-900, (**h**) FeNC-950.

The –OH groups of the sol-gel-prepared Fe3O4(OA) stay inside the micelles before polymerization. Additionally, after the polymerization in the inverse suspended system, most of the PANIs are formed in the micelles and become polymer droplets that then coalesce and develop the morphology seen in Figure 4c when the ISP system was eventually de-emulsified by the addition of acetone (the breaking of polymer droplets) and the fibrous morphology is never seen.

Upon calcination at high temperatures, the Fe3O4(OA) that covered with PANI(EB) would convert to other magnetic materials with a higher magnetic activity. Combining the increasing magnetic attractive forces and thermal energy provided by the high temperature calcination, these Fe3O4(OA) nanoparticles were able to transform and to merge into bigger particles due to the increasing magnetic attraction forces with increasing temperatures, as seen from Figure 4d–h. The sizes of the new particles are smaller when the calcination temperatures are below 700 ◦C. Bigger cake-like ensembles are perceivable when temperatures reach 800 ◦C. These cake-like ensembles even impinge further into huge slabs around 900 ◦C, revealing the occurrence of extremely high magnetic attractive forces after 900 ◦C. The types of magnetic materials created when Fe3O4(OA) are calcined inside of the PANI(EB) matrix at high temperatures can be studied by checking their X-ray diffraction patterns and XPS spectra.

## *3.5. TEM Micropicture*

The pearl-like chain morphology of the TEM micropicture in Figure 5a expresses the neat Fe3O4(OA) particles that are connected to each other by the inter-entangled or inter-digitized long alkyl tails of the attached OA. Most of the Fe3O4(OA) particles were covered by the PANI(EB) after the ISP at RT, as seen in Figure 5b, where only some tiny ones can be seen in the margins of the big ensemble. The pretty dark and homogeneous color seen in Figure 5b reveals that Fe3O4(OA) particles are uniformly distributed in the PANI(EB) matrix. When the temperature reached 600 ◦C, most of the PANI(EB) were thermally degraded and only 13.5 wt% were left, according to Figure 3 and Table 2, the sample became more transparent in Figure 5c, and some tiny Fe3O4(OA) nanoparticles transformed and coalesced into bigger, dark particles after 600 ◦C. Whether they still remain in the form of Fe3O<sup>4</sup> and what kind of new iron-compounds formed after 600 ◦C, can be analyzed using X-ray or XPS spectra. The particle-assembly phenomena were enhanced with the calcination temperature and huge grain-like particles developed from 600 to 950 ◦C, in accordance with Figure 5c–h. The growing size of the dark particles resulted from the increasing magnetic attractive forces with temperature, which originated from the formation of some iron-compounds with a high magnetic activity. There are N, C, and O atoms inside the nanocomposite, except Fe. Additionally, the FeNCs are all derivatives of Fe3O4(OA) via calcination. The increasing magnetic forces compared to the neat PANI(EB)/Fe3O4(OA) or the neat Fe3O4(OA) can be attributed to the newly formed FeCx, or α-Fe, even the α 00-Fe16N<sup>2</sup> compounds that own much higher magnetic forces than the neat Fe3O4(OA). Furthermore, the obtained Fe3C (cementite) is the most stable FeCx compound. The variation % of these atoms and types of iron-compounds in the composite with increasing temperatures can be understood from either the XRD patterns or the XPS for each compound.

**Figure 5.** TEM micropictures of (**a**) Fe3O4(OA), (**b**) PANI(EB)/Fe3O4(OA), (**c**) FeNC-600, (**d**) FeNC-650, (**e**) FeNC-700, (**f**) FeNC-800, (**g**) FeNC-900, (**h**) FeNC-950. **Figure 5.** TEM micropictures of (**a**) Fe3O<sup>4</sup> (OA), (**b**) PANI(EB)/Fe3O<sup>4</sup> (OA), (**c**) FeNC-600, (**d**) FeNC-650, (**e**) FeNC-700, (**f**) FeNC-800, (**g**) FeNC-900, (**h**) FeNC-950.

The juxtaposed fibrous PANI(EB), fully covered with Fe3O4(OA) nanoparticles (PANI(EB)-Em/Fe3O4(OA)), can also be seen in Figure 6a. These fibers, which merged into a huge ensemble, gradually became hard stone in Figure 6b and cannot be easily broken into small pieces by simply grinding them in the mortar. It is believed that the surface Fe3O4(OA) nanoparticles developed into hard covering materials and the calcined PANI(EB) (950 °C) mostly remain inside the composites. In contrast, the ISP-prepared composite allowed the inclusive Fe3O4(OA) nanoparticles to develop into individual small magnetic particles inside the FeNCs, as seen in Figure 6c, which can be easily ground into tiny particles in the mortar, as depicted in Scheme 1. Amazingly, some highly crystallized The juxtaposed fibrous PANI(EB), fully covered with Fe3O4(OA) nanoparticles (PANI(EB)- Em/Fe3O4(OA)), can also be seen in Figure 6a. These fibers, which merged into a huge ensemble, gradually became hard stone in Figure 6b and cannot be easily broken into small pieces by simply grinding them in the mortar. It is believed that the surface Fe3O4(OA) nanoparticles developed into hard covering materials and the calcined PANI(EB) (950 ◦C) mostly remain inside the composites. In contrast, the ISP-prepared composite allowed the inclusive Fe3O4(OA) nanoparticles to develop into individual small magnetic particles inside the FeNCs, as seen in Figure 6c, which can be easily ground into tiny particles in the mortar, as depicted in Scheme 1. Amazingly, some highly crystallized stone powders

stone powders can be released from the cracked particles after grinding, as seen in Figure

can be released from the cracked particles after grinding, as seen in Figure 6d, which is also illustrated in Scheme 1. Actually, the breaching by grinding occurs following the carbonized PANI(EB) boundaries. It means we are able to fabricate any shapes of high magnetic stones by further sintering these magnetic powders in different shapes of molds at a temperature far below the melting points of the magnetic stones. Moreover, the magnetic powders can also be protected from oxidation during sintering by the carbonized coverings. 6d, which is also illustrated in Scheme 1. Actually, the breaching by grinding occurs following the carbonized PANI(EB) boundaries. It means we are able to fabricate any shapes of high magnetic stones by further sintering these magnetic powders in different shapes of molds at a temperature far below the melting points of the magnetic stones. Moreover, the magnetic powders can also be protected from oxidation during sintering by the carbonized coverings.

**Figure 6.** TEM micropictures of (**a**) PANI(EB)-Em/Fe3O4(OA), (**b**) PANI(EB)/Fe3O4(OA), (**c**) FeNC-**Figure 6.** TEM micropictures of (**a**) PANI(EB)-Em/Fe3O<sup>4</sup> (OA), (**b**) PANI(EB)/Fe3O<sup>4</sup> (OA), (**c**) FeNC-950, (**d**) Ground FeNC-950.

#### 950, (**d**) Ground FeNC-950. *3.6. Raman Spectra*

*3.6. Raman Spectra*  The ratio (ID/IG) of –C–C– single bonds (sp3, D-band) with –C=C– double bonds (sp2, G-band) demonstrated in the Raman spectrum can be used to monitor the degree of the thermal degradation of organic compounds. When it increases with temperature (more G-bands thermally destroyed to become D-band carbons), it means a rougher surface structure is created after heating and vice versa. The ID and IG peaks assigned at 1348 and 1575 cm−1 in the Raman spectrum represent the D- and G-band of the covalent-bonded carbons, respectively. The surface structures that vary with the calcination temperatures for PANI(EB) and FeNCs are monitored using the Raman spectra in Figure 7 and their ID/IG values are listed in the last two columns of Table 2. The ID/IG of PANI(EB) in Figure 7a decrease with the calcination temperature due to the crosslinking and some degree of the ordering of the graphene lattice, referring to the PANI(EB) matrix structure of the FeNC composites that is actually becoming more and more smooth with the formation of plane sp2 (C=C) bonding. However, the entire composite does not follow that trend with increasing temperatures when Fe3O4(OA) is present. The degradation of OA tails and the thermal transformation of the iron-compounds, which accompany the formation of bonds between Fe and N, or the C atoms of the PANI(EB), can destroy more C=C bonding too since it owns more active π-bonds. Moreover, the iron-compound itself would experience The ratio (ID/IG) of –C–C– single bonds (sp<sup>3</sup> , D-band) with –C=C– double bonds (sp<sup>2</sup> , G-band) demonstrated in the Raman spectrum can be used to monitor the degree of the thermal degradation of organic compounds. When it increases with temperature (more G-bands thermally destroyed to become D-band carbons), it means a rougher surface structure is created after heating and vice versa. The I<sup>D</sup> and I<sup>G</sup> peaks assigned at 1348 and 1575 cm−<sup>1</sup> in the Raman spectrum represent the D- and G-band of the covalent-bonded carbons, respectively. The surface structures that vary with the calcination temperatures for PANI(EB) and FeNCs are monitored using the Raman spectra in Figure 7 and their ID/I<sup>G</sup> values are listed in the last two columns of Table 2. The ID/I<sup>G</sup> of PANI(EB) in Figure 7a decrease with the calcination temperature due to the crosslinking and some degree of the ordering of the graphene lattice, referring to the PANI(EB) matrix structure of the FeNC composites that is actually becoming more and more smooth with the formation of plane sp<sup>2</sup> (C=C) bonding. However, the entire composite does not follow that trend with increasing temperatures when Fe3O4(OA) is present. The degradation of OA tails and the thermal transformation of the iron-compounds, which accompany the formation of bonds between Fe and N, or the C atoms of the PANI(EB), can destroy more C=C bonding too since it owns more active π-bonds. Moreover, the iron-compound itself would experience a crystallographic transformation at high temperatures in accordance with the Fe–C phase diagram after 900 ◦C. All these possible newly formed bonds and transformations at high temperatures play significant roles in the eventual structures of the FeNCs obtained at

a crystallographic transformation at high temperatures in accordance with the Fe–C phase

different temperature calcination, resulting in the increasing ID/I<sup>G</sup> with temperature in Figure 7b and Table 2. Certainly, the iron-compounds obtained at various temperatures all demonstrate much higher magnetic forces than the starting material PANI(EB)/Fe3O4(OA), which will be illustrated later. different temperature calcination, resulting in the increasing ID/IG with temperature in Figure 7b and Table 2. Certainly, the iron-compounds obtained at various temperatures all demonstrate much higher magnetic forces than the starting material PANI(EB)/Fe3O4(OA), which will be illustrated later.

**Figure 7.** Raman-spectra of (**a**) neat PANI(EB) (**b**) FeNCs. **Figure 7.** Raman-spectra of (**a**) neat PANI(EB) (**b**) FeNCs.

#### *3.7. Powder XRD Patterns 3.7. Powder XRD Patterns*

The powder XRD patterns of neat Fe3O4 and FeNC treated at the various temperatures displayed in Figure 8a significantly demonstrates the variation of the characteristic crystalline plane-peaks from Fe3O4 to FeO, Fe3C, Fe3N, and α-Fe (α″-Fe16N2) with temperature. In particular, when the temperature was close to 900 °C, α-Fe and α″-Fe16N2 became the dominant core product due to the migration of O, N, and C atoms out of the core The powder XRD patterns of neat Fe3O<sup>4</sup> and FeNC treated at the various temperatures displayed in Figure 8a significantly demonstrates the variation of the characteristic crystalline plane-peaks from Fe3O<sup>4</sup> to FeO, Fe3C, Fe3N, and α-Fe (α 00-Fe16N2) with temperature. In particular, when the temperature was close to 900 ◦C, α-Fe and α <sup>00</sup>-Fe16N<sup>2</sup> became the dominant core product due to the migration of O, N, and C atoms out of the core materials.

materials. Other type of hard, stable iron-compounds such as Fe3C or Fe3N with lower

Other type of hard, stable iron-compounds such as Fe3C or Fe3N with lower magnetic activities formed the covering materials protecting the inner α-Fe from oxidation and losing magnetic activity in the atmosphere at RT. The presence of α 00-Fe16N<sup>2</sup> in the core also provided additional protection for the formed α-Fe. The characteristic diffraction patterns of α-Fe(110) and α 00-Fe16N2(220) illustrated in Figure 8b remain almost intact after staying in the air for 3 months, which will otherwise become iron oxides in less than 1 month without any protection for neat α-Fe (ferrite). It again illustrates the necessity and importance of covering iron-materials with PANI(EB) via inverse suspension polymerization. importance of covering iron-materials with PANI(EB) via inverse suspension polymerization. According to FWHM, the crystallite sizes for FeNC-900 and -950 are 1.15 and 3.75 nm, respectively (calculated from the X-ray diffraction patterns of Figure 8b α-Fe(110) peak), which are much smaller than TEM (Figure 6c) The size of α-Fe (or α″-Fe16N2) crystallites also increased with temperature from 900 to 950 °C. The size of Fe3C based on Fe3C (031) plane in Figure 6c is 3.65 nm, which is also much smaller than the particles illustrated in TEM (Figure 6c).

and losing magnetic activity in the atmosphere at RT. The presence of α″-Fe16N2 in the core also provided additional protection for the formed α-Fe. The characteristic diffraction patterns of α-Fe(110) and α″-Fe16N2(220) illustrated in Figure 8b remain almost intact after staying in the air for 3 months, which will otherwise become iron oxides in less than 1 month without any protection for neat α-Fe (ferrite). It again illustrates the necessity and

*Polymers* **2021**, *13*, x FOR PEER REVIEW 14 of 19

**Figure 8.** X-ray diffraction patterns of (**a**) FeNCs and (**b**) FeNC-950 staying in the open air for **Figure 8.** X-ray diffraction patterns of (**a**) FeNCs and (**b**) FeNC-950 staying in the open air for months.

months. According to FWHM, the crystallite sizes for FeNC-900 and -950 are 1.15 and 3.75 nm, respectively (calculated from the X-ray diffraction patterns of Figure 8b α-Fe(110) peak), which are much smaller than TEM (Figure 6c) The size of α-Fe (or α 00-Fe16N2) crystallites also increased with temperature from 900 to 950 ◦C. The size of Fe3C based on Fe3C (031) plane in Figure 6c is 3.65 nm, which is also much smaller than the particles illustrated in TEM (Figure 6c).

#### *3.8. XPS Spectra*

Typical XPS spectra of FeNCs recording % of Fe, N, C, and O atoms in Figure 9a,b provide the % variation of each atom at different temperatures. Additionally, the Fe(2p3/2) of the composites calcined at various temperatures was revealed in Figure 9c. The characteristic peaks of α 00-Fe16N<sup>2</sup> (706.8 eV), α-Fe (707.7 eV), Fe3C (708.5 eV), Fe3O<sup>4</sup> (710.8 eV), and FeO (709.4 eV) [35–39], respectively, were illustrated in Figure 9c as well. Referring to the magnetic data obtained from Figure 10, we are able to construct Figure 9d, which illustrates how magnetic activity and each type of atom % varied with the calcination temperatures. The positions of the Fe(2p3/2) patterns, which ranged from 700 to 714 MeV, can be applied to understand what kind of iron-compounds the ISP-prepared PANI(EB)/Fe3O4(OA) composite became during calcination. Additionally, the satellite peaks that shift from ~715 up to ~725 eV when the thermal treatment temperature increases from 600 to 950 ◦C correspond to the ferrous compounds [40,41]. According to Figure 9c, α-Fe was not the dominant compound until the calcination temperature raised over 800 ◦C, but Fe3C is already created above 700 ◦C by a reaction between the included Fe3O4(OA) nanoparticles and the surrounding PANI(EB) matrix seen in Figure 9d and exists for all calcination temperatures. In other words, Fe3C, which is a very stable, hard magneton, dominates in the composite in the beginning of calcination and α-Fe formed later by driving some C atoms out of the core area at a temperature higher than 800 ◦C. Therefore, we can find a clear and sharp increase in the Fe % and a deep decease in the C % when the temperature was over 800 ◦C in Figure 9d, which was also accompanied with a sudden increase in magnetic forces. The formation of an α-Fe core and an Fe3C shell provides a facile way to fabricate stable magnetons with ultrahigh SM values, which will be discussed in the following section.

#### *3.9. SQUID Spectra*

Ferrite (α-Fe) and α <sup>00</sup>-Fe16N<sup>2</sup> are two of the magnetic materials with SM over 200 emu g−<sup>1</sup> , α <sup>00</sup>-Fe16N<sup>2</sup> even reaches 300 emu g−<sup>1</sup> . Limited to its vulnerable structure, α-Fe easily fuses with O, C, or N atoms to become iron oxide, Fe3C, and FeNx, respectively. Figure 10a clearly indicates that the SM of ISP-prepared PANI(EB)/Fe3O4(OA) varies with calcination temperatures from 600 to 950 ◦C. The SM increases from 10 to 249 emu g−<sup>1</sup> when calcination increases from RT to 950 ◦C, in accordance with Figure 10a, due to the transformation from Fe3O<sup>4</sup> to α-Fe and α <sup>00</sup>-Fe16N2, which has already been proven using the X-ray and XPS spectra discussed in the previous sections. However, the SM does not monotonously increase with temperature since different iron-compounds with a higher or lower SM formed at different calcination temperatures. The SM slightly increased from 10 to 67 emu g−<sup>1</sup> at 650 ◦C and fell back to 27 emu g−<sup>1</sup> after the temperature increased to 800 ◦C. Briefly, the SM does not exceed 70 emu g−<sup>1</sup> if calcination maintains below 800 ◦C. The X-ray patterns in Figure 8a and the XPS spectra in Figure 9c demonstrated the presence of mixtures of Fe3O4, FeO, and Fe3C and their SMs are well below 250 emu g−<sup>1</sup> theoretically. Until 900 ◦C is reached, the SM abruptly raises to 230 emu g−<sup>1</sup> and then 245 emu g−<sup>1</sup> at 950 ◦C. Again, their X-ray patterns and XPS spectra illustrate the formation of affluent α-Fe and α <sup>00</sup>-Fe16N<sup>2</sup> at this stage by driving other atoms out of the core area with the help of high thermal energy. Certainly, there might be the presence of the lattice transformation of the iron-compounds after 900 ◦C, which also propels N, C, and O atoms to the outer area of α-Fe and α <sup>00</sup>-Fe16N<sup>2</sup> to become protecting shell materials. Moreover, α <sup>00</sup>-Fe16N<sup>2</sup> is able to prevent α-Fe from oxidation as well. According to the common Fe–C diagram, BCC-Fe would convert to FCC-Fe after 912 ◦C. Therefore, it is very possible for the iron compounds to undergo a significant atom rearrangement when the temperature is over 900 ◦C.

**Figure 9.** XPS spectra of (**a**) FeNC-800, (**b**) FeNC-950, (**c**) Fe (2p3/2) of FeNCs, (**d**) specific magnetization (Ms) and various **Figure 9.** XPS spectra of (**a**) FeNC-800, (**b**) FeNC-950, (**c**) Fe (2p3/2) of FeNCs, (**d**) specific magnetization (Ms) and various atom % of FeNCs at different temperatures.

Under the protection of these stable iron-compounds, the SM of a mixture of α-Fe and α″-Fe16N2 decays only to 232 emu g−1 after staying in the air for 3 months, referring to Figure 10b. The slight decrease in the SM of FeNC-950 can be attributed to the possible surface oxidation of the powders during grinding, which largely created the surface area.

The ground FeNC-950 sample in the mortar retains the high SM of 245 emu g−1, as shown in Figure 10a. The FeNC prepared from the calcination of ISP-prepared PANI(EB)/Fe3O4(OA) is unlike other magnetic materials with a high SM, it can be easily broken into small pieces by a low degree of milling without losing the SM. The breaching

PANI(EB) boundaries that are affluent with carbon material, as illustrated in Scheme 1. In

All the FeNCs demonstrate superparamagnetic behaviors, referring to Figure 10a.

atom % of FeNCs at different temperatures.

other words, FeNC is composed of small particles made of α-Fe and α″-Fe16N2 covered

with carbon materials, which is very similar to the TEM micropicture in Figure 6c.

**Figure 10.** Saturated magnetization of (**a**) FeNCs and (**b**) various FeNC-950. **Figure 10.** Saturated magnetization of (**a**) FeNCs and (**b**) various FeNC-950.

**4. Conclusions**  Fe3O4 nanoparticles were successfully and fully covered by polyanilines via inverse suspension polymerization in accordance with the SEM and TEM micrographs and the nanoparticles were able to develop into magnetons protected by the carbonization mate-Under the protection of these stable iron-compounds, the SM of a mixture of α-Fe and α <sup>00</sup>-Fe16N<sup>2</sup> decays only to 232 emu g−<sup>1</sup> after staying in the air for 3 months, referring to Figure 10b. The slight decrease in the SM of FeNC-950 can be attributed to the possible surface oxidation of the powders during grinding, which largely created the surface area. All the FeNCs demonstrate superparamagnetic behaviors, referring to Figure 10a.

rials developed from polyaniline at different calcination temperatures. The saturated magnetization of the calcined iron-composites slightly increased from RT to 700 °C first and depressed continuously until 800 °C. A surprising jump of the saturated magnetization was found during 800~900 °C calcination. Based on the spectra of the X-ray diffraction and the XPS of iron-compounds calcined at temperatures higher than 900 °C, we understand a mixture of α-Fe and α″-Fe16N2 was formed in the core area and protected by the The ground FeNC-950 sample in the mortar retains the high SM of 245 emu g−<sup>1</sup> , as shown in Figure 10a. The FeNC prepared from the calcination of ISP-prepared PANI(EB)/ Fe3O4(OA) is unlike other magnetic materials with a high SM, it can be easily broken into small pieces by a low degree of milling without losing the SM. The breaching points or parts of FeNC by grinding or milling are believed to follow the carbonized PANI(EB) boundaries that are affluent with carbon material, as illustrated in Scheme 1. In other

surrounding hard iron-compounds such as cementite (Fe3C). The composite obtained from calcination at 950oC slightly lost its saturated magnetization from 245 to 232 emu g−<sup>1</sup>

words, FeNC is composed of small particles made of α-Fe and α <sup>00</sup>-Fe16N<sup>2</sup> covered with carbon materials, which is very similar to the TEM micropicture in Figure 6c.

#### **4. Conclusions**

Fe3O<sup>4</sup> nanoparticles were successfully and fully covered by polyanilines via inverse suspension polymerization in accordance with the SEM and TEM micrographs and the nanoparticles were able to develop into magnetons protected by the carbonization materials developed from polyaniline at different calcination temperatures. The saturated magnetization of the calcined iron-composites slightly increased from RT to 700 ◦C first and depressed continuously until 800 ◦C. A surprising jump of the saturated magnetization was found during 800~900 ◦C calcination. Based on the spectra of the X-ray diffraction and the XPS of iron-compounds calcined at temperatures higher than 900 ◦C, we understand a mixture of α-Fe and α <sup>00</sup>-Fe16N<sup>2</sup> was formed in the core area and protected by the surrounding hard iron-compounds such as cementite (Fe3C). The composite obtained from calcination at 950 ◦C slightly lost its saturated magnetization from 245 to 232 emu g−<sup>1</sup> after staying for 3 months in the air. Moreover, unlike the common high magnetic materials, the calcined high magnetic product is not too hard to break by grinding or milling and does not cause the loss of saturated magnetization either. The easily broken properties do not originate from the presence of the cementite, since its hardness is also very high, but from the weak surrounding carbonized materials developed from polyaniline. For the time being, the application of inverse suspension polymerization to cover iron-magnetic materials with polymers before subjecting them to calcination to prepare the high magnetic materials proves to be very successful.

In the future, sintering approaches will be applied to shape the ground magnetic powders into various shapes of magnetons with high saturated magnetization in specific molds for different purposes.

**Author Contributions:** Conceptualization, Y.-Z.W. and W.-Y.H.; methodology, L.-C.H.; formal analysis, Y.-T.S.; writing—original draft preparation, Y.-W.C.; writing—review and editing, K.-S.H. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the Ministry of Science and Technology in Taiwan, grant number MOST 105-2622-E-151-012-CC3, MOST108-2221-E-992-037, and MOST 109-2221-E-992-083.

**Acknowledgments:** We appreciate the use of the soft matter-TEM equipment belonging to the Instrument Center of National Cheng Kung Univ. (NCKU), Ministry of Science and Technology in Taiwan, ROC.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**

