INTRODUCTION

With the rapid consumption of traditional fossil fuel and growing concern of global environmental problem, it is urgent to develop cost-effective, large-capacity and high performance energy storage devices and technologies [1, 2]. Due to the high energy density, high energy efficiency, long cycle life and flexible design [3], lithium-ion batteries (LIBs) have drawn great attention from modern society, and become increasingly important power source for portable electronic devices as well as electric vehicles [4].

One of the key parameters that determine the performance of LIBs is the type of anode materials used [5]. From this point, numerous kinds of negative materials such as lithium alloys [6], silicon composites [7, 8], carbonaceous materials [9, 10] and metal compounds [11–14] have been developed, among which carbonaceous materials are proved to be the most effective hosts for Li-ions due to their advantages of structural stability, high safety, good cycle ability, relative low-cost and abundance in source [15]. In the case of carbonaceous materials, natural and artificial graphites, carbon nanofiber and mesocarbon microbead (MCMB) are commercially used in the practical production at present [16]. Nevertheless, the fact of limited theoretical capacity (372 mA h g–1) and poor Li-ions diffusion kinetics through the highly oriented carbon structure cannot satisfy the increasingly strict requirements of large energy density and high power density for the modern electrical devices. In order to overcome these problems, various special carbonaceous materials have been designed and proposed, among which pyrolysis carbon (also called hard carbon) has attracted great attention due to its high specific theoretical capacity, cycling stability and rate capability [17, 18].

Generally, pyrolysis carbon, which mainly consists of highly turbostratic nanostructure, can be prepared from pyrolyzing polymers such as resins [19–21] and biomass [22, 23]. The depicted structure can be tailored by individual pyrolysis conditions or alloying with other compounds [24, 25]. Unlike graphite, Li-ions can be not only intercalated into the scattered graphene sheets but also adsorbed at the edges of nanocrystallites or stored in the pores during the charge process [26], endowing pyrolysis carbon good electrochemical performance that is particularly desirable for power tools. However, due to some drawbacks such as high-cost synthetic process, low reversible capacity and initial columbic efficiency, the widespread application of pyrolysis carbon in commercial LIBs is still limited. Thus to make practical use of pyrolysis carbon, not only a high initial performance is important, but also a low-cost is necessary.

In this work, a facile method is reported for large-scale fabrication of novel pyrolysis carbon by using commercial epoxyphenolics (EPN) as the raw material. EPN is selected as the precursor due to its vast availability in chemical industry, low market price, high carbon yield (near 50% in weight) and adjust ability through chemical modification. However, the direct pyrolysis of EPN usually results in low yield and initial columbic efficiency [27]. To approach this concern, a “structural doping” strategy is adopted, in which the EPN, curing agent and doping source (phosphoric acid) are mixed and co-solidified firstly, and then pyrolyzed together to form the final products. The obtained carbons with optimal structure could display promising lithium storage capability of large initial reversible capacity (up to 420 mA h g–1), high initial coulombic efficiency (more than 80%), excellent rate performance and good cycling stability. The results are extremely encouraging to realize the high-value utilization of the staple epoxy product and provide an alternative solution for high performance pyrolysis carbon.

EXPERIMENTAL

Materials Preparation

Commercial EPN obtained from Jiangyinnanhui trading company (Jiangsu, China) was used as the main raw material. Besides, AR grade phosphoric acid (HP, 85 wt % in water, Aladdin) and phthallic anhydride (PA, Aladdin) were used as the dopant and curing agent to harden the epoxy resin prior to pyrolysis, respectively. The schematic diagram of the preparation process is shown in Fig. 1. In a typical run, about 3000 g EPN was re-melted by oil bath in a 5 L cylindrical stainless (316 L) container under stirring. Then 40 wt % curing agent was added to the epoxy and mixed at 80°C until a homogenous mixture was obtained. After this, 10 wt % HP was quickly added into the mixture with vigorous stirring. Then the mixture was further curing at 130°C for 4 h in a drying oven. After the curing, the sample was broken up into chunks of about 1 cm3 in size and then precarbonized at 600°C for 1 h in a pit type furnace with N2 as the flushing gas. The obtained precarbonized sample was then ball-milled into powders with the maximum particles size less than 50 μm using a vertical mixing mill (SY-10L Ball Miller) with zirconia beads. Finally, the obtained powders were pyrolyzed at 900, 1100, and 1300°C in an industrial controlled atmosphere furnace and at 2800°C in an Acheson furnace for 2 h, respectively, with UHP-grade argon (99.999% pure) used as the flushing gas. The derived carbon samples were labeled as EPN900, EPN1100, EPN1300 and EPN2800, respectively.

Fig. 1.
figure 1

Schematic diagram of kilogram preparation process for the pyrolysis carbon.

Full size image

Materials Characterization

Thermogravimetric analysis (TGA) was carried out on a NETZSCHSTA449 F3 instrument from room temperature to 1000°C at a heating rate of 5°C min–1 under the nitrogen atmosphere. The morphology and elemental microanalysis were performed on a scanning electron microscope (SEM, FEI Quanta-200) coupled with an electron dispersive X-ray spectroscopy (EDX) detector (EDAX, Apollo XP model). Transmission electron microscopy (TEM) characterization was taken on a JEOL 2100F microscope. X-ray diffraction (XRD) patterns were performed using Cu Kα irradiation (λ = 1.54056 Å) with 2θ-angle recording from 10° to 90° (Bruker D8 Advance diffractometer). Raman spectra (Renishaw MKI-2000) were determined on a Labram HR800 with an argon-ion laser operating at 514 nm to investigate the microstructure of the samples. The electrical conductivity measurements were obtained at ambient temperature using the four-contact method. Nitrogen adsorption–desorption isotherms were performed at 77 K on Micromeritics ASAP 2460 analyzer to obtain the porosity parameters of the samples. Before measurements, the samples were degassed in vacuum at 473 K for at least 12 h. The specific surface area (SBET) was calculated by using conventional Brunauer–Emmett–Teller (BET) method, and the total pore volume (Vt) was estimated using a single point at the maximal relative pressure of 0.996.

Electrochemical Measurements

Electrochemical tests were conducted using CR2032 coin-type cells. The working electrodes were prepared by mixing the EPN-pyrolysis carbons, carbon black (Super-P, Timcal), and polyvinyl fluoride (PVDF, Arkema) binder in N-methyl-2-pyrrolidone (NMP, Aldrich) with a weight ratio of 85 : 12 : 3 to form a homogeneous slurry. Then, the slurry was coated onto copper foil current collector using a micrometer-adjust film applicator and dried at 120°C for 10 min. The loading mass of active material was controlled at about 6 mg cm–2. The as-prepared electrodes were pressed using a roll-pressing machine, punched into round pieces, and finally dried at 80°C overnight in a vacuum oven. The CR2032 coin-type cells were then assembled using the working electrode and lithium foil counter electrode. Porous polypropylene membrane (Celgard 2400) was employed as the separator. A solution of LiPF6 (1 M in a 1 : 1 V/V of EC : DEC) was used as the electrolyte. All the cells were assembled in a dry argon filled glove box. Galvanostatic charge and discharge tests were carried out on LAND CT2001A multichannel battery testing systems between 0.005–2.5 V (vs. Li/Li+). Cyclic voltammetry (CV, cut-off voltage: 0.005–2.5 V) and electrochemical impedance spectroscopy (EIS) measurements were conducted on an IM6ex electrochemical workstation (Zahner IM6ex). CV was recorded at a scan rate of 0.5 mV s–1, and EIS was measured over the frequency range from 100 kHz to 100 mHz with an amplitude of 5.0 mV. All the electrochemical tests were performed at room temperature.

RESULTS AND DISCUSSION

The pyrolysis behavior of the phosphorus doped-EPN is firstly determined by thermogravimetric (TG) analysis under nitrogen atmosphere, as shown in Fig. 2. According to the differential TG curve, the rapid weight loss range from 100 to 320°C is mainly attributed to the evaporation of water and other uncrosslinked low molecular weight compounds. While the weight loss between 350 and 650°C is mainly due to the removal of light molecules generated from the multiple and complex reactions (e.g., cracking, polymerization and condensation) [28]. As temperature reaches 750°C, the resins have already transformed into a coke-like solid state, and the weight loss could mainly result from the decomposition of some aromatic rings and partially escape of the unstable phosphide groups [29]. The formed phosphorus-doped structures would prevent the further decomposition by raising the decomposition temperature and result in a high carbon yield [30]. As a result, the pyrolysis yield at 1000°C arrives at 47.7 wt %, which is much higher than common biomass and other synthetic organic precursors [31], indicating a necessary advantage for the usage of EPN as possible precursor for anode materials in LIBs.

Fig. 2.
figure 2

TG and DTG curves of the resin after solidification reaction.

Full size image

After the curing, the obtained bulk product was immediately precarbonized at 600°C under nitrogen atmosphere and then milled into powders followed by high temperature pyrolysis under 900, 1100, 1300, and 2800°C, respectively. Figures 3a–3d show the morphology and texture of the obtained pyrolysis carbons. They are found to present irregular granular morphology with smooth surface. The carbon particles exhibit an average size of about 10 μm, with many fine grains dispersed inside. To be noted, the thermal treatment does not lead to particle aggregation. Figures 3e–3h illustrate the EDX patterns of the obtained carbons. It can be seen that carbon (C), oxygen (O) and phosphorus (P) are the main elements in the samples. The element distribution in Table 1 shows that the P content is 4.13, 2.44, 1.05 wt % and 0 for EPN900, EPN1100, EPN1300 and EPN2800, respectively, which is dramatically decreased with the increase of pyrolysis temperature. Clearly, the residual phosphorus can be easily tailored by pyrolysis regulation, which favors to derive various designed P-doped carbons.

Fig. 3.
figure 3

SEM images of (a) EPN900, (b) EPN1100, (c) EPN1300 and (d) EPN2800. (e), (f), (g) and (h) are EDX microanalyses of the samples of (a), (b), (c) and (d), respectively.

Full size image
Table 1. Elemental composition and physical parameters of the prepared pyrolysis carbons
Full size table

The nanostructures of the EPN based pyrolysis carbons were further characterized by TEM, as shown in Fig. 4. The obtained TEM images reveal a significant improvement of local structure ordering with the increase of pyrolysis temperature. For EPN900, only turbostratic graphitic structure (typical long-range disorder) can be observed (Fig. 4a), which matches the typical “house of cards” model for hard carbons [32]. While for EPN1100 and EPN1300, some local ordered structures containing well-arranged nano-graphene domains could be easily found (Fig. 4b), which develops in number and grows in length with increasing temperature (Fig. 4c), indicating an improvement of graphitization degree. Moreover, as the pyrolysis temperature reaches 2800°C, most of the disorder structure has transformed into layered structure (Fig. 4d), which is close to a typical long-range order graphite model [33]. According to previous research results [25, 34], the amorphous networks and graphene domains in pyrolysis carbon can both serve as effective Li+ storage sites, thus the EPN based carbons pyrolyzed at different temperatures may have various electrochemical performances.

Fig. 4.
figure 4

TEM images of (a) EPN900, (b) EPN1100, (c) EPN1300 and (d) EPN2800.

Full size image

Figure 5a shows the nitrogen adsorption–desorption isotherms of the obtained carbons. As the pyrolysis temperature increases from 900 to 1100°C, a typical I–V isotherm with distinct hysteresis loop at relative pressure P/P0 between 0.45 and 0.95 is observed, indicating the evolution of open mesopores. However, as the pyrolysis temperature successively increases to 1300 and 2800°C, the hysteresis loop subsequently shrinks to a minimal level, suggesting an obvious change of nanostructure under higher pyrolysis temperature. Combined with TEM results (Fig. 4), this change should be attributed to the construction of ordered nanostructure. Figure 5b shows the pore size distributions of the obtained samples. It can be found that the majority of the pores fall in size of ∼3 nm. The calculated pore parameters are displayed in Table 1. The specific surface area (SBET) is 2.7, 4.4, 2.4 and 2.8 m2 g–1 for EPN900, EPN1100, EPN1300 and EPN2800, respectively. To be noted, EPN1100 possesses the highest total pore volume (1.04 × 10–2 cm3 g–1) as well as the largest micropore volume (6.0 × 10–4 cm3 g–1), which is considered to benefit from the formation of open pores and proper decomposition of the functional groups (micropore formation could be ascribed to the release of gases [35, 36]).

Fig. 5.
figure 5

(a) Nitrogen adsorption/desorption isotherms and (b) the pore size distribution determined by the BJH model of EPN900, EPN1100, EPN1300 and EPN2800.

Full size image

X-ray diffraction (XRD) patterns are collected to confirm the structural features of the obtained samples. As shown in Fig. 6a, the intensities of (002) peak at 22° and (100) peak at 44° increase accordingly with increasing pyrolysis temperature, indicating the evolution of ordered nanostructure. Especially, after pyrolyzing at 2800°C, all the broad peaks develop to a sharp type, indicating a high degree of graphitization. This high graphitization degree is probably due to the catalytic effect of P fraction which facilitates defects removing and carbon structure rebuilding during the pyrolysis [37]. To be noted, no peak that belongs to P composites is detected in the XRD patterns, illustrating that the doped-P is homogenously distributed in the ultimate samples after pyrolysis. Based on the Bragg equation, the interlayer space d(002) is calculated to be 0.3741, 0.3716, 0.3676 and 0.3365 nm (as shown in Table 2) for EPN900, EPN1100, EPN1300 and EPN2800, respectively, which is considerably larger than that of ideal graphite (d(002) = 0.335 nm) [38].

Fig. 6.
figure 6

(a) XRD patterns and (b) Raman spectra for the prepared pyrolysis carbon EPN900, EPN1100, EPN1300 and EPN2800.

Full size image
Table 2. Structure parameters and electrochemical properties of EPN-based pyrolysis carbons
Full size table

Raman spectra are also determined to probe the microstructure of the obtained carbons. As shown in Fig. 6b, all the samples give the D-band at ∼1350 cm–1 and G-band peak at ∼1600 cm–1, corresponding to the disordered carbon band and the graphitic mode, respectively [39]. It is well established that integral intensity ratios of D to G peak (ID/IG) is indicative of order degree of the local carbon layers in carbon materials [40]. The ID/IG ratios are calculated to be 1.85, 1.49, 1.41, and 0.09 for EPN900, EPN1100, EPN1300 and EPN2800, respectively, which decreases accordingly with the increase of pyrolysis temperature, further suggesting the decreasing of defects and growing of local long-range ordered structure.

The electrochemical properties are investigated to understand the Li+ storage behavior in the obtained carbons. Figures 7a–7d show the cyclic voltammetry (CV) curves of the samples. Clearly, two oxidation peaks at around 0.32 and 0.95 V and two reduction peaks at about 0 and 0.56 V can be observed in the initial cycle for sample EPN900, EPN1100 and EPN1300. Generally, the noticeable irreversible reduction peaks at about 0.56 V in the first cathodic scan is ascribed to the generation of a solid electrolyte interphase (SEI) layer [41], which contributes to the irreversible capacities in the first cycle. In the subsequent cycles, the CV curves almost overlap with each other, indicating good cycle performance of the carbons. Moreover, with the increase of pyrolysis temperature, the wide oxidation peak around 0.32 V, which is assigned to the delithiation from interlayers of the graphitic nanocrystallities [42], gradually becomes sharp. While the wide oxidation peak around 0.95 V, corresponding to the delithiation from defect sites {including the dopant P/C composite [43]}, gradually weakens with the decrease of P content. For EPN2800, the reduction peak at 0.95 V disappears (Fig. 7d), resulting from the absence of P-containing defect sites and formation of more graphitic domains under 2800°C.

Fig. 7.
figure 7

(a–d) Cyclic voltammograms at 0.5 mV s–1 and (e–h) the first three cycle charge/discharge profiles of the prepared pyrolysis carbons.

Full size image

Figures 7e–7h show the galvanostatic charge/discharge profiles at 0.1 C (25 mA g–1) of the samples. All the samples display similar charge profiles containing a sloping region and a plateau region, corresponding to the lithium storage at defect sites and intercalation between graphene layers as well as Li+ adsorption in the micropores, respectively [37]. Clearly, the ratio of the lithiation capacity contributed by the slope region gradually decreases with the increase of pyrolysis temperature, corresponding to less defects at higher temperature. The galvanostatic test results of the initial cycle are summarized in Table 2. It can be found that EPN900, EPN1100 and EPN1300 deliver the charge/discharge capacities of 518.2/416.1, 389.5/320.2, and 279.7/232.4 mA h g–1, with an initial coulombic efficiency (ICE) of 80.3, 82.2 and 83.1%, respectively. As the pyrolysis temperature further increases to 2800°C, the initial reversible capacity increases to 342.9 mA h g–1 with an ICE of ∼86.0%. The elongated plateau region and the nearly disappeared slope region in EPN2800 make it more like a graphite sample.

The rate performance is determined to further evaluate the EPN based carbons. As shown in Fig. 8a, the charge capacities of EPN900, EPN1100 and EPN1300 under different rates gradually decrease with the increase of pyrolysis temperature. Similar capacities are obtained at the rate of 3 C for EPN1100 and EPN900. Notably, EPN2800 exhibits an inferior rate capability as the capacities dramatically decrease at the rate higher than 1 C. Capacities of 158.8, 160.1, 97.4, and 27.7 mA h g–1 are achieved at 3 C, corresponding to a retention ratio of 46.1, 55.0, 41.5, and 13.9% (RE1, as shown in Table 2) for EPN900, EPN1100, EPN1300 and EPN2800, respectively. This indicates that exorbitant pyrolysis temperature (e.g., 2800°C) is not conducive to improve the rate performance of EPN based carbons. As the rate reduces back to 0.1 C, the capacity can recover to 335.3, 277.2, 217.1, and 309.4 mA h g–1, with a retention ratio of 81.3, 89.2, 89.3 and 89.4% (RE2, as shown in Table 2) for EPN900, EPN1100, EPN1300 and EPN2800, respectively.

Fig. 8.
figure 8

(a) Rate performance from 0.1 to 3 C, (b) discharge profiles at 3 C, (c) Nyquist plots before cycling and (d) cycle performance at a current rate of 0.3 C of the pyrolysis carbons.

Full size image

Figure 8b presents the charge profiles of the obtained carbons at 3 C. It can be observed that only the slope region is retained, suggesting good diffusion kinetics of Li+ in the defect sites at high rate. Especially, EPN2800 exhibits large voltage drop and almost loses the original shape at 3 C, illustrating a severe electrochemical polarization due to the sluggish kinetics. The dynamic processes of the carbons before cycling can be revealed by electrochemical impedance spectroscopy (EIS, shown in Fig. 8c). The depressed semicircles in the high and medium frequency regions of EIS represent the total interfacial resistance from SEI and the charge-transfer resistance (Rct) [44].In this case, the smaller semicircle indicates lower Rct. Generally, lower Rct implies better electrode stability and lithiation kinetics [45]. EPN2800 exhibits the largest Rct possibly due to its high ordered structure. While EPN1100 has the lowest Rct which can be attributed to its rich mesoporous structures providing more channels for the Li+ diffusion [46].

Figure 8d shows the cycling performance of the obtained carbons at 0.3 C. Both EPN900 and EPN1100 display a rapid capacity fading in the initial 5 cycles due to SEI film formation and capacity consumption by active sites. In contrast, EPN1300 and EPN2800 exhibit quite stable cycle performance. After 100 cycles, the retained capacities are 277.8, 247.8, 206.2, and 298.9 mA h g–1 for EPN900, EPN1100, EPN1300 and EPN2800, respectively, with the corresponding capacity retention (against the initial reversible capacity) of 66.1, 77.4, 88.0 and 88.0%. According to the excellent high-rate performance, EPN1100 could be an attractive candidate for high power density application in LIBs. While the stable cycling performance combined with large capacity and high coulombic efficiency could make EPN2800 a promising anode material for high energy density storage systems.

CONCLUSIONS

Novel low-cost pyrolysis carbons are successfully fabricated from phosphorus-doped epoxyphenolics by a two-step method of solidification and pyrolysis. The content of the introduced phosphorus decreases gradually with the elevation of pyrolysis temperature. EPN900 exhibits a reversible capacity of 420 mA h g–1 at 0.1 C, which is higher than the theoretical capacity of graphite. While EPN2800 show a balanced lithium storage performance, which can deliver a relatively large reversible capacity of 343 mA h g–1, high initial coulombic efficiency (∼86%), and superior cycling performance (299 mA h g–1 after 100 cycles at 0.3 C). As indicated by structure analysis, the favorable electrochemical performance is attributed to the synergistic effect of phosphorus doping, composition of graphitic/amorphous microstructure and porosity. Considering advantages of simple synthetic process, high yield and high performance, the EPN based carbons may be promising anode materials for lithium-ion batteries.