Green Synthesis of Carbon-Encapsulated Magnetic Fe₃O₄ Nanoparticles Using Hydrothermal Carbonization from Rattan Holocelluloses

Abstract

How to design a simple and scalable procedure for manufacturing multifunctional carbon-based nanoparticles using lignocellulosic biomass directly is a challenging task. Based on the green chemistry concept, we developed a novel one-pot solution-phase reaction to prepare carbon-encapsulated magnetic nano-Fe₃O₄ particles (Fe₃O₄@C) with a tunable structure and composition through the hydrothermal carbonization (HTC) of Fe²⁺/Fe³⁺ loaded rattan holocelluloses pretreated with ionic liquids (EmimAc and AmimCl). The detailed characterization results indicated that the Fe₃O₄@C synthesized from the holocelluloses pretreated with ionic liquids (ILs) under alkaline conditions tends to have a higher saturation magnetization, probably due to the increased iron ions loading. Moreover, increasing the HTC temperature led to an increased abundance of hydroxyl groups on the surface of the synthesized particles and an elevated saturation magnetization. When EmimAc-treated holocelluloses were used as the carbon precursors, well-encapsulated Fe₃O₄@C nanoparticles were obtained with a maximum saturation magnetization of 42.6 emu/g. This synthetic strategy, coupled with the structure of the iron carbide-based composite and the proposed mechanism, may open a new avenue for the development of carbon-encapsulated iron oxide-based magnetic nanoparticles.

1. Introduction

Carbon-encapsulated magnetic nanoparticles (CEMNs) have distinctive structures and chemistry that make them have a high potential as absorbents [1,2] and electrodes [3,4] and in microwave adsorption [5], oxidative degradation [6], catalysis [7], and biomedical functions [8]. Unlike traditional activated carbon materials [9,10], abundant surface functional groups [11] are easily modified carbon shell and nano-size effect endowed CEMNs with excellent properties in the environmental, energy, and medicine fields. Until now, several approaches have been adopted in preparing carbon-coated nanosized metal particles, but these methods involve high energy consumption, require sophisticated procedures, and have difficulty producing high-purity materials in large quantities [12,13]. Therefore, exploring environmentally friendly and efficient strategies is urgently required in order to overcome these limitations.

Recently, glucose, sucrose, starch, etc., were used as carbon precursors to prepare CEMNs in a water reaction medium under self-generated pressure and moderate temperatures (180–280 °C). By a one-step hydrothermal carbonization (HTC) process of glucose and Fe(NO₃)₃·9H₂O followed by annealing under Ar, uniform and small Fe₃O₄ nanocrystals (about 9 nm) encapsulated in interconnected carbon nanospheres (about 60 nm) have been fabricated to produce high-rate Li-ion battery anodes [14]. Similarly, a one-pot simple procedure for the preparation of graphite carbon-coated magnetic Fe₃O₄ through a glucose hydrothermal reaction followed by high temperature pyrolysis have been developed [15]. Using a modified hydrothermal method, more complex magnetic carbon composites (Fe⁰/Fe₃C@CS) have also been synthesized [16]. More recently, lignin-based oligomer precursors were also used to prepare CEMNs. Novel Co@C core-shell nanoparticles were prepared by using a straightforward low-temperature carbonization process [17]. For these carbon precursors being homogeneously dispersed or dissolved in water, the prepared CEMNs achieved a complete core-shell structure and excellent application performance.

However, the next challenging step in the preparation of CEMNs is to directly exploit lignocellulosic biomass as a carbon precursor and find a novel synthesis route to prepare magnetic nanoparticles at a low temperature with higher yields and purities. Although considerable efforts have been directed toward the synthesis of carbon microspheres with insolubilized biomass in water using HTC [18,19,20,21], there are few reports concerning the use of lignocellulosic biomass as a carbon precursor in the fabrication of CEMNs. Its use as a carbon precursor is definitely more complex than that of simple sugars due to a higher degree of structural complexity and its insolubility in water. It is, therefore, the purpose of the present study to contribute towards extending HTC from complex biomass to the convenient synthesis of CEMNs. Inspired by a highly efficient biomass dissolving system, in this work, we designed a novel method for fabricating Fe₃O₄@C composites with complete core-shell structures and high saturation magnetization through a green wet chemical route under hydrothermal conditions. In this study, Fe₃O₄ cores were formed via the reaction of Fe²⁺/Fe³⁺ under an alkaline condition. At the same time, the carbon shells that were carbonized from rattan holocelluloses protected the Fe₃O₄ cores. The synthesized products may have promising applications in clinical diagnostics, therapeutics, molecular biology, bioengineering, catalysis, and magnetic storage devices.

2. Materials and Methods

2.1. Materials

The rattan cane (Daemonorops margaritae (Hance) Becc.) chips were ground to pass a 40-mesh sieve (aperture of 450 µm) to prepare rattan holocelluloses. The typical procedure is described as follows: rattan powder delignification in aqueous sodium chlorite/acetic acid reagent. The solid-to-liquid ratio was 1:26 (w/v), and the acetic acid and sodium chlorite loads were 0.1 mL/g and 0.3 g/g, respectively. The mixture was heated to 80 °C and maintained for 1.0 h. After delignification, the samples were carefully washed with deionized water until water pH became neutral. The analytic grade ferric chloride hexahydrate (FeCl₃·6H₂O) and ferrous chloride (FeCl₂·4H₂O) were purchased from Alfa Aesar Inc. (China). All chemical solutions were prepared using deionized (DI) water with a resistivity of 18.2 MΩ cm⁻¹, which was also used to rinse and clean the samples.

2.2. One-Pot Synthesis of Fe₃O₄@C Nanoparticles

Due to the excellent solubility properties of cellulose, 60 g of ionic liquids (ILs) (EmimAc or AmimCl, Dalian Institute of Chemical Physics, Chinese Academy of Science) was loaded into a 100 mL round-bottom flask and preheated to 110 °C to dissolve 3 g rattan holocelluloses under nitrogen atmosphere. As a clear light-yellow holocelluloses solution was obtained, a 3 mL solution with 1:1 M ratio of Fe²⁺/Fe³⁺ was dropped at the same time followed by vigorous stirring. The simultaneous iron ions deposition and cellulose regeneration process was proceeded by slowly pouring the homogeneous solution into a large amount of ultrapure water (80 °C) or alkaline solution (NH₃·H₂O), labelled as ACl/EAc_H2O and ACl/EAc_NH3, respectively. The precursors were filtered off. After drying in oven at 105 °C for 12 h, 3 g regenerated holocelluloses precursor was dispersed in 30 mL deionized water. The solution was then sealed in a 50 mL Teflon inlet and heated in an oven at 180 °C, 230 °C, and 280 °C for 8 h. Then, the obtained nanoparticles were washed with ultrapure water and ethanol several times, and then freeze-dried. By comparison, the mixtures of 3 g rattan holocelluloses and 3 mL solution with 1:1 M ratio of Fe²⁺/Fe³⁺ underwent the same treatment to produce the magnetic nanoparticles, denoted as W_H2O and W_NH3, respectively. The yields of synthesized nanoparticles were calculated by the following equation:

$$Yield (\%)=\frac{Weight \left(CEMNs, g\right)}{Weight \left(Iron ions loaded hollocelluloses precursor,g\right)}*100\%$$

(1)

2.3. Characterizations

Crystal structure and internal microstructure of CEMNs were identified using an X-ray diffractometer (Philips X’Pert PRO, PANalytical, Almelo, the Netherland) with a scattering angle ranging from 10° to 80° 2θ and at a scanning speed of 2°/min. X-ray photoelectron spectroscopy (XPS) was conducted on a Thermo Escalab 250 (Thermo Scientific, Waltham, MA, USA) with Al-Kα X-ray to investigate the chemical states of elements. A Shirley background was first subtracted followed by component fitting using Lorentzian-Gaussian functions with a 30% Lorentzian component. 633 nm-Raman spectroscope (Horiba Jobin Yvon, Kyoto, Japan) was used to investigate the graphitic degree. A 30 s integration time was chosen to achieve spectra with good counting rates, while the laser power of the incident beam was kept at 5 mW to prevent irreversible thermal damage to the specimen surfaces. Surface morphology of the Fe_-C products was observed using scanning electron microscopy (Carl Zeiss Merlin, Oberkochen, Germany) at 20 kV. Microstructural characteristics of the samples were characterized by HRTEM at 200 kV (JEOL-2010F, JEOL, Tokyo, Japan). Magnetization measurements of magnetic (Fe–C) nano-composition were performed at room temperature using vibration sample magnetometer (SQUID-VSM, Quantum Design, San Diego, CA, USA). Before all the measurements, the samples were vacuum dried. The thermal stability and degradation behavior for the synthesized materials were characterized by thermogravimetric analysis using a TA analyzer (Discovery TGA 55, TA, New Castle, DE, USA) at a heating rate of 10 °C/min in nitrogen atmosphere.

3. Results and Discussion

3.1. XRD

The XRD patterns of the Fe–C products prepared at 180–280 °C are shown in Figure 1. The broad peaks centered at 18° for the nanoparticles synthesized at 230 °C and 280 °C correspond to the amorphous structure of polymer-like carbons [22], indicating that these corresponding feedstocks have been carbonized as carbon. The typical XRD patterns of magnetite (2θ = 18.4°, 30.2°, 35.5°, 37.1°, 43.2°, 53.5°, 57.3°, and 62.9° corresponding to eight indexed planes (111), (200), (311), (222), (400), (422), (511), and (440)) were detected within the series of W_NH3 (Figure 1A), EAc_NH3 (Figure 1B), Acl_NH3 (Figure 1C), and EAc_H2O (Figure S1A) samples. Clearly, the Fe₃O₄@C nanoparticles were finally obtained after the regeneration of EmimAc-dissolving holocelluloses and were stable under different hydrocarbonization temperatures. However, the typical XRD patterns of magnetite are completely unable to be observed for AmimCl (Figure S1B). The phenomena could potentially be ascribed to the acidity or basicity of the ILs themselves. The EmimAc exhibited a certain alkalinity in favor of forming Fe₃O₄ nanoparticles in the dissolving process, which could have been simultaneously adsorbed on the fibers. Thereby, it is not surprising to observe the Fe₃O₄@C nanoparticles in this manner after they have been regenerated in hot water. On the contrary, the AmimCl creates an acidic condition, which could be unable to form Fe₃O₄ nanoparticles within the dissolution system, least of all the regeneration in hot water. The data of yields are listed in

Table 1. Yields, graphite degree, and magnetic parameters of the Fe₃O₄@C nanoparticles.

Samples	Yield (%) a	Raman b	Magnetic Parameters
		IG/ID	Ms (emu/g)	Mr (emu/g)	Hc (Oe)
WNH3-180	53.1	0.53	8.9	0.4	13.6
WNH3-230	43.1	0.65	11.5	1.0	15.9
WNH3-280	34.2	0.73	15.2	1.5	20.2
EAcNH3-180	85.9	0.62	13.1	0.2	3.1
EAcNH3-230	37.8	0.85	33.6	0.4	2.7
EAcNH3-280	37.9	0.83	42.6	0.6	4.1
EAcH2O-180	78.4	0.60	10.4	0.3	2.5
EAcH2O-230	33.8	0.65	13.8	0.4	1.9
EAcH2O-280	31.5	0.80	15.1	0.3	2.1
AClNH3-180	65.5	0.74	14.2	0.1	1.5
AClNH3-230	31.6	0.84	28.2	0.3	1.5
AClNH3-280	23.6	0.89	33.0	0.6	0.9
AClH2O-180	65.5	0.64	-	-	-
AClH2O-230	32.8	0.72	-	-	-
AClH2O-280	31.1	0.82	-	-	-

^a Weight% based on the starting material. ^b I_G: Integration of the G band; I_D: Integration of the D band.

, and it can be seen that the yields of the obtained carbon-encapsulated nanoparticles gradually decreased with the rising hydrocarbonization temperature. Moreover, the EAc_NH3 procedure was in favor of manufacturing samples with high yields, and, when hydrocarbonized at 280 °C, the EAc_NH3-280 nanoparticles displayed the highest yields of 37.9%.

3.2. Raman and XPS

Further information about the microstructural evolution of the carbon shell as a function of the HTC temperature was studied using Raman spectroscopy (Figure 2A,B). The typical Raman spectra of the Fe₃O₄@C core-shell structure showed characteristic wide D and G bands around 1340 cm⁻¹ and 1585 cm⁻¹, respectively, typical for disordered graphite and graphitic lattice vibration [23,24]. An increase in the I_G/I_D ratios for the series of Fe₃O₄@C demonstrates an enhancement of the degree of graphitization (Table 1). In comparison, the corresponding values for the typical amorphous carbon and highly ordered pyrolytic graphite are approximately 0.3 and 4.0, respectively [25]. This proves that the carbon shells in the Fe₃O₄@C consisted of a low degree of graphite carbon layers.

XPS has often been used for the surface characterization of various materials, and unambiguous results are readily obtained when the various surface components each contain unique elemental markers [26]. Here, in order to further analyze the carbon shell in the core/shell products, XPS was performed on the nanoparticles obtained by the EAc_NH3 process (Figure 3A and supporting materials Figure S2A,E). The high-resolution XPS C 1s spectra of nanoparticles, shown in Figure 3B and the supporting materials Figure S2B,F, were resolved into five individual component peaks that represent C=C (284.2 eV), C-C (284.8 eV), C-OH (285.3 eV), C-O-C (286.3 eV), O-C=O (288.4 eV), respectively [27]. Meanwhile, it can be observed that, at 180 °C, the Fe₃O₄@C is rich in oxygen-containing functional groups (carboxyl and carbonyl groups) (

Table 2. Experimental C 1s and O1s binding energy (BE eV)/chemical state assignments for the as-synthesized nanoparticles.

Samples	C1	C2	C3	C4	C5	O1	O2	O3	O4
	Carbidic C	C–C	C–O	C=O	O=C–O	C=O	C–O–C	C–O	O–C=O
EAcNH3-180		284.8/11.9%	285.3/21.4%	286.7/55.0%	288.0/11.7%	530.6/5.4%	-	-	533.1/95.6%
EAcNH3-230	284.2/26.6%	284.8/38.6%	285.3/22.4%	286.3/12.4%	-	530.2/15.1%	531.9/63.5%	-	533.3/21.4%
EAcNH3-280	284.2/21.2%	284.8/34.3%	285.3/32.3%	286.3/12.2%	-	530.2/23.0%	531.2/26.9%	532.2/28.4%	533.2/21.7%

). Increasing the HTC temperature leads to the increase of C-C (11.9–34.3%) and C-O (21.4–32.3%), which is largely due to intramolecular condensation, dehydration, and decarboxylation reactions. The slight loss of C-C bonds at 280 °C is most probably due to the less thermally stable aliphatic carbon chains that are normally present in hydrothermal carbon produced at lower temperatures. The O 1s spectra fitted to four component peaks (Figure 3C and supporting materials Figure S2C,G). Peak I (530.2 eV) corresponds to the C=O or quinine groups, Peak II (531.5 eV) to the C-O-C groups, peak III (532.2 eV) to the C-OH groups, and peak Ⅳ (533.2) to the -O-C=O groups [28,29]. The O-C=O contribution to the O 1s profile (peak Ⅳ) decreases significantly from 95.6% (EAc_NH3-180) to 21.7% (EAc_NH3-280), which is probably due to a reduction of the carboxyl group under acidic conditions. Clearly, the O 1s spectra are consistent with the C 1s profiles, corroborating that there was an increase in the hydroxyl groups on EAc_NH3-280. From the Fe2p high-resolution spectra (Figure 3D and the supporting materials in Figure S2C,G), there are two peaks at 724.4 and 711.4 eV corresponding to the Fe2p^1/2 and Fe2p^3/2 in Fe₃O₄ [30], respectively.

3.3. SEM and TEM

The as-synthesized Fe₃O₄@C hybrids were further characterized using SEM, TEM, and HRTEM images, and SAED patterns. The ILs’ dissolution process affects the morphology of the final products greatly. It can be seen that W_NH3-280 displayed non-spherical particles with irregular surface (Figure 4A) and aggregated strongly on the 2D plates (Figure 4B,C). By comparison, the nanoparticles EAc_NH3-280 and ACl_NH3-280 formed spherical aggregation clusters connected by carbonaceous materials (Figure 4D and supporting material Figure S3), which were further confirmed by the TEM image (Figure 4E). An individual particle EAc_NH3-280 at higher magnification clearly shows that the particle has a core-shell structure with an average core size in the range of 6 to 10 nm. It can be seen that the Fe₃O₄@C-EmimAc and Fe₃O₄@C-AmimCl consist of well-dispersed round/ellipsoidal shape nanoparticles. The carbon shell was amorphous, and the core exhibited a single crystalline with homogeneous diffraction rings (inset of Figure 4F). From the HRTEM image, the lattice fringe spacing of 0.252nm corresponds to the (311) plane of the face-centered structure Fe₃O₄, indicating that the as-synthesized products were well-crystallized with a high degree of crystallinity.

3.4. TGA

The TGA–DTA (thermogravimetry analysis and differential thermal analysis) profiles of the thermal behavior between 80 °C and 600 °C for selected samples are presented in Figure 5 and the supporting material Figure S4. A prominent weight loss occurred at a temperature ranging from 300 to 350 °C, which was observed in the Fe₃O₄@C nanoparticles hydrocarbonized at 180 °C. There was still ~35% and ~30% solid residue left at 600 °C, respectively. It has been confirmed that hemicelluloses started decomposition easily, with the weight loss mainly happening at 220~315 °C, and cellulose pyrolysis was focused at a higher temperature range (315~400 °C) [31]. Thus, due to the pyrolysis occurring under nitrogen protection, the dramatic weight loss for EAc_NH3-180 and ACl_NH3-180 was mainly due to the thermal degradation of the hemicelluloses and partial cellulose [32], though it displayed a thermal curve with a single weight loss dip when a one-step oxidative degradation process of crystalline carbon with oxygen occurred [33]. When the HTC temperature was higher than 230 °C, almost all the holocelluloses were converted into amorphous carbon, with a very high solid residue reaching 65% and 60% for the EAc_NH3-280 and ACl_NH3-280 samples. Obviously, a higher pretreatment temperature increases the thermalstability of the resultant materials.

3.5. VSM

Figure 6 shows the magnetic hysteresis loops of the W_NH3 and EAc_NH3; the corresponding Ms, Mr, and Hc values of each sample are tabulated in Table 1. It is clear that the series of EAc_NH3 nanoparticles produced by the simultaneous Fe₃O₄ deposition and holocelluloses regeneration process exhibited higher magnetic properties than that of the W_NH3 (supporting materials), mainly due to their higher Fe₃O₄ loading. Furthermore, increasing the HTC temperature was found to be beneficial in obtaining Fe₃O₄@C nanoparticles with higher magnetic performance, which is attributable to an increase in the crystallinity [34]. By mixing the iron solution with rattan sawdust or rattan charcoal, Hu et al., (2017) prepared a series of magnetic rattan biochar with the saturation magnetization value ranging from 1.52 and 27.11 emu g⁻¹ and found that magnetite dominates the crystal phases of the magnetic biochar pyrolyzed at 600–800 °C [35]. Comparatively, the synthesizing procedure in the present work has the potential to prepare magnetic carbon materials with higher saturation magnetization.

3.6. Proposed Growth Mechanism of Series Fe₃O₄@C Nanoparticles

Based on the above discussion and the characterization results of a series of nanoparticles, a possible mechanism of the formation process of the carbon-encapsulated iron-core nanoparticles is proposed (Figure 7). When holocelluloses are dissolved in ILs (EmimAc and AmimCl), the intra- and intermolecular H-bonds in the cellulose are completely cleaved [36,37]. During this process, the anion of the ILs is incorporated into the cellulose surface, attracting the surrounding iron ions. Subsequently, when NH₃·H₂O is added into the homogeneous systems, the iron ion absorbed on the holocellulose surface will react with the NH₃·H₂O to form Fe₃O₄, which is deposited on the surface of the holocellulose. Similar to the mechanisms of pure carbohydrates hydrocarbonization [38,39,40], regenerated rattan holocelluloses are hydrolyzed into sugars under acidic conditions obtained via the decomposition of H₂O when hydrothermally treated at 180 °C. Then, polymerization and condensation reactions may be induced by intermolecular dehydration, dehydrogenation, or aldol condensation. The Fe₃O₄ nanoparticles attached on the surface of the holocelluloses are then combined with small carbonaceous colloids through coulombic interactions with the surface functional groups (e.g., OH and C=O) and condense to form Fe₃O₄-in-C. This is then further assembled to Fe₃O₄@C core/shell nano-spheres via further intermolecular dehydration of the surface functional groups. Stable surface functional groups, such as ether or quinone, are also formed. Once the growth process is stopped, the nano-carbon shells, which cover nano Fe₃O₄ particles, are completed.

Actually, the formation of the Fe₃O₄ nanoparticles after the ILs dissolving process is divided into two parts: (1) the iron ions are absorbed on the holocelluloses surface by electrostatic attraction and complexation, and part of the Fe₃O₄ nanoparticles are formed under the basic condition created by the ILs themselves; (2) during the regeneration process, other parts of the Fe₃O₄ nanoparticles are formed by the reaction between the absorbed iron ions and OH^- from NH₃·H₂O. Clearly, the latter process plays a critical role in the formation of Fe₃O_4, evidenced by the excellent magnetization property of EAc_NH3-280 (Figure 6B) compared to that of EAc_H2O-280 (Figure S5B). In terms of the type of ILs, basic ILs (Figure S1A) favored the formation of Fe₃O₄ nanoparticles in the dissolving process, which could be simultaneously adsorbed on the surface of hollocelluloses. Under the same alkaline condition, the magnetic products from ILs-treated hollocelluloses, even in acidic ILs (Figure S5A), performed much higher magnetization than those dispersed in water (Figure 6A).

4. Conclusions

In conclusion, a very simple and scalable pathway towards the synthesis of Fe₃O₄@C nanoparticles has been explored. It has been demonstrated that the ILs dissolution and NH₃·H₂O regeneration procedure was in favor of manufacturing samples with high yields and saturation magnetization. Stable magnetic carbon-encapsulated nano-composites with an average core size in the range of 6 to 10 nm were obtained, which displayed a maximum saturation magnetization of 42.6 emu/g and the highest yields of 37.9%. This method has the potential to provide a simple, energy-efficient, and environmentally friendly approach for preparing carbon-coated nanomaterials, and the proposed mechanism may open a new avenue for the development of carbon-encapsulated iron oxide-based magnetic nanoparticles.