PAN Xin, LIU Yang, WANG Xu-zhen,, ZHAO Zong-bin, QIU Jie-shan

(1.School of Chemistry, Faculty of Chemical, Environmental and Biological Science and Technology, Dalian University of Technology, Dalian116024, China;2.Carbon Research Laboratory & Liaoning Key Lab for Energy Materials and Chemical Engineering, School of Chemical Engineering, Faculty of Chemical, Environmental and Biological Science and Technology, Dalian University of Technology, Dalian116024, China;3.University of Jinan, Jinan250022, China)

Abstract: The performance of lithium ion batteries (LIBs) is highly dependent on the properties of the anode materials. Developing new carbon materials and metal oxides/sulfides with high capacities has attracted growing attention due to the limited theoretical capacity of commercial graphite. Pyrite (FeS2) is environmental benign, inexpensive and has a high theoretical capacity of 894 mAh g-1 because of its four-electron reduction by lithium, which make it promising for use as an anode material of LIBs. To improve the electrical conductivity and volume change of pyrite, hybrids made of pyrite confined in nitrogen-doped carbon nanotubes (FeS2/N-CNTs) were fabricated by floating catalyst chemical vapor deposition (FCCVD), followed by sulfidation. Results indicate that the original Fe/N-CNTs formed during the FCCVD have iron nanowires or nanorods inside the N-CNTs and their sulfidation with sulfur vapor at 400 ℃ for 1, 2 and 5 h leads to FeS2/N-CNTs with pyrite contents of 22.4, 45.2, and 55.8 wt%, respectively. The pyrite in FeS2/N-CNTs is in two forms, one is FeS2 nanowires confined in half-open N-CNTs and the other is FeS2 nanoparticles attached to the outer walls of the N-CNTs as a result of diffusion out of the inner tubes. A large number of defects on the N-CNTs tube walls is necessary for the diffusion of sulfur vapor into the tubes to make the iron species accessible to the sulfur. The FeS2/N-CNTs-45.2 wt% has the highest discharge capacity (996 mAh g-1 at 0.1 A g-1), good rate capability and stable cycling performance as an anode material for LIBs.

Key words: Lithium ion batteries; Carbon nanotubes; Nitrogen doping; Confinement; Sulfidation

1 Introduction

Lithium ion batteries (LIBs), important energy storage devices, can be applied in almost all current portable electronics[1]. The performance of LIBs is highly dependent on the property and morphology of cathode or anode materials[2-4]. In terms of anode materials, the commercial graphite with a low theoretical capacity of 372 mAh g-1can’t meet the practical demands[5]. To improve the capability of LIBs, many other materials like graphene[6,7], carbon nanotubes (CNTs)[8], silicon[9]，metal oxides[10]and sulfides[11-13]have been developed. Among these materials, iron pyrite (FeS2) has attracted growing attention as an anode material for its special properties[14,15]. Firstly, pyrite is an inexpensive earth-abundant mineral which is nontoxic and environmentally benign. More importantly, the four electron reduction of cubic pyrite by lithium ions (FeS2+ 4Li++ 4e-→ Fe + 2Li2S) provides a large theoretical specific capacity of 894 mAh g-1[16]. However, there are still some problems that prevent pyrite from being applied in LIBs. One problem is that FeS2has a poor electrical conductivity and always experience serious structural pulverization caused by the huge volume change during the insertion and extraction of Li+. And the iron sulfide batteries also face the same problem existing in Li-S batteries which called “shuttle” phenomenon[17-19]. It has been pointed out that sulfur is one of the charge products during the FeS2conversion reaction. The initial reduction of sulfur produces soluble and highly mobile polysulphide (Li2Sn) which can diffuse throughout the separator and deposit on the opposite electrode. It will not only cause the losing of active component but also prevent the further electrochemical reactions[20,21].

Many methods especially forming carbon/FeS2composites have been tried to solve the above mentioned problems[22]. For example, Liu et al[23]increased the cycling and rate performance of FeS2cathode material through carbon coating. While Wu et al[20]and Qiu et al[24]synthesized FeS2-embedded carbon microsphere and FeS2-nitrogen doped graphene composite respectively to improve the capability of iron sulfide anode materials. However, the fabrication of pyrite and carbon composite mentioned in these papers all used hydrothermal way and further annealing procedure was needed to improve the crystalline, which limited their production and application. Herein, pyrite/nitrogen doped carbon nanotube nanocomposites (FeS2/N-CNTs) were prepared through a facile vulcanization process with nitrogen doped carbon nanotubes filled with iron (Fe/N-CNTs) as a precursor. And their lithium storage abilities were tested as anode materials for LIBs.

As we all know that the confinement effect of CNTs has been long investigated[25-28]. Many nanomaterials like (C60)[29], NiO[30]nanoparticles or CeF3[31]nanowires etc. have been filled into the hollow inner cores of CNTs. Among these filled materials, iron or its derivatives attracted great attentions owing to their low cost and various applications in the areas of catalysis and energy storage. But as reported by Bao et al[32], the oxidation of iron filled inside nanotubes was harder than iron outside due to the protection role of carbon nanotube walls. So the transformation of iron filled inside was difficult. Here an unique structure of Fe/N-CNTs was designed and produced. The tube walls showed an interesting semi-confinement effect due to the doped nitrogen that results in a large amount of defects on the N-CNT walls, which made the transformation of iron wrapped inside tube cavities easy. There is strong adhesion between the in-situ produced pyrite and carbon nanotubes, which not only improves the electron transfer ability of anode materials, but also inhibits the shuttle effect of iron sulfide. What’s more, the transformation of iron filled inside nanotubes also provides us a new sight to study and utilize the confinement effect of CNTs.

2 Experimental

2.1 Synthesis of Fe/N-CNTs and Fe/CNTs

Fe/N-CNTs were synthesized by a floating catalyst chemical vapor deposition (FCCVD) method using anhydrous ferric chloride (FeCl3) as the catalyst precursor and melamine as the carbon source as reported previously[33]. In brief, the same mass of FeCl3and melamine were separately put into two porcelain boats (length 80 mm), then pushed into the quartz tube side by side, reaching the entrance of furnace. They would evaporate together in furnace entrance whose temperature could reach over 400 ℃ above the boiling point of melamine and anhydrous ferric chloride when the furnace was heated to 800 ℃. And the vapor mixture came into the furnace center with the carrier gas of 200 sccm argon. Fe/N-CNTs would grow on the center of quartz tube for 30 min and were collected at room temperature.

The comparative sample of the iron-filled carbon nanotubes without nitrogen doping (Fe/CNTs) was synthesized using the same method only with melamine replaced by ethylene. The flow rate of ethylene was controlled at 20 sccm and the deposition lasted for 30 min until all FeCl3evaporated.

2.2 Sulfidation of Fe/N-CNTs and Fe/CNTs

The obtained Fe/N-CNTs or Fe/CNTs were put into a small quartz tube (inner diameter of 0.8 cm and length of 3.2 cm) with one end closed. Then the samples were transferred into a glove box filled with argon (water and oxygen concentration less than 0.110-6) and sealed in a 14 mL stainless steel vessel with 1 g sublimed sulfur power placed at its bottom. The reaction of Fe/N-CNTs and sublimed sulfur was carried out at 400 ℃ for different times. Then obtained products were heated at 300 ℃ for 1 h under the protection of 100 sccm argon to eliminate the extra sulfur.

2.3 Characterization

The morphology of products was characterized by scanning electron microscopy (SEM, JMS-6360 LV) and transmission electron microscopy (TEM, JEM-2000 EX). The crystalline phases of the products were examined using a Rigaku D/MAX-2400 diffractometer (XRD) equipped with a rotating anode using a CuKαradiation source (λ=0.154 18 nm). X-ray photoelectron spectroscopy (XPS) analysis was performed using a Thermo ESCALAB 250 spectrometer employing an Al-KαX-ray source. Raman spectroscopy (JY LabRam HR800, excited by 532 nm laser) was used to confirm the structural defects.

2.4 Electrochemical tests

At first, a slurry was prepared by mixing 70 wt% active electrode material (FeS2/N-CNTs), 20 wt% acetylene black and 10 wt% polyvinylidene fluoride (PVDF) binder. Then working electrodes were fabricated by casting the slurry on Cu foils. The electrodes were dried under vacuum at 120 ℃ overnight with a mass loading of active materials about 1 mg cm-2. Coin type cells were assembled in a glove box under argon atmosphere (water and oxygen concentration less than 0.1×10-6), which consisted of an as-prepared electrode as the working electrode, polypropylene separator, and lithium foil as the counter electrode. The electrolyte used in this experiment was 1 mol/L LiPF6salt dissolved in ethylene carbonate (EC) and diethyl carbonate (DEC) in a 1∶1 volume ratio. The amount of electrolyte is about 40 μL for each cell. The galvanostatic charge/discharge tests were carried out in a Land CT2001A battery test system between 0.01-3.0 V using 2016 coin-type cells. The cycle voltammograms (CV) were tested using a CHI660D electrochemistry workstation between 0.01-3.0 V at a scan rate of 0.1 mV s-1. The electrochemical impedance spectroscopy (EIS) measurements were carried out using a CHI660D workstation by applying an AC amplitude of 5 mV over the frequency range of 100 kHz to 0.01 Hz.

3 Results and discussion

Fig. 1 illustrates the whole process for the synthesis of FeS2/N-CNTs nanocomposite. The Fe/N-CNTs were produced by the FCCVD method with anhydrous ferric chloride and melamine as starting materials as previously reported[33]. From the measurement by TGA and XPS, the iron filling ratio and the nitrogen content was 37.0 wt% and 3.79 at% (proportion of nitrogen atoms in the total number of carbon and nitrogen atoms) respectively. The conversion of iron filled inside to FeS2was carried out at 400 ℃ in the atmosphere of sulfur vapor. The structure changes and lithium storage ability test will be discussed below.

Fig. 1 A schematic representation of the fabrication of FeS2/N-CNTs.

The iron filled inside nanotubes exists in two forms (a-Fe and Fe3C) as confirmed by XRD analysis. The reactions of iron and iron carbide with sulfur steam are: Fe + 2S(g) →FeS2; Fe3C + 6S(g) → 3FeS2+ C. However, the protection role of nanotube walls makes the conversion complicated. To investigate the sulfidation process in details, different reaction times were employed at the same temperature. XRD patterns (Fig. 2) were used to confirm the transformation of crystal structures and peaks of pyrite started to appear after ion species reacted with sublimed sulfur for 1 h as shown in the curve (a). By comparing the curve (a) with the curve (b), it can be seen that the peaks of iron and iron carbide weakened a lot while the pyrite peaks became stronger when the reaction lasted for 2 h. The peak of iron carbide disappeared and only a weak a-Fe peak remained after the reaction for 5 h. This demonstrated that the transformation degree of iron increased with reaction time. The content of FeS2in different samples employing conversion times of 1, 2 and 5 h was estimated to be 22.4 wt%, 45.2 wt%, and 55.8 wt% respectively based on the quality increment through a simple equation (The marking method is only used to distinguish the samples):

WhereWis the mass fraction of FeS2in obtained samples,miis the increased mass after sulfidation,mtis the total mass of obtained samples, andMFeS2andMsrepresent the molar mass of FeS2and elemental S, respectively. Different from the reduction of iron peaks, the graphite peaks nearly kept unchanged for

various samples, suggesting the sulfur didn’t react with carbon.

The morphologies and microstructures of Fe/N-CNTs precursor and the resulting FeS2/N-CNTs composite with different pyrite contents have been characterized by SEM and TEM observation (Fig. 3a-j).

Fig. 2 XRD patterns of FeS2/N-CNTs samples with different FeS2 contents (a) 22.4 wt% (react for 1 h), (b) 45.2 wt% (react for 2 h) and (c) 55.8 wt% (react for 5 h).

Fig. 3 (a, b) SEM and (c) TEM images of Fe/N-CNTs, (d) SEM and (g-h) TEM images of FeS2/N-CNTs-22.4 wt%, (e) SEM and (i) TEM images of FeS2/N-CNTs-45.2 wt%, (f) SEM and (j) TEM images of FeS2/N-CNTs-55.8 wt%.

As shown in Fig. 3(a, b), SEM images of Fe/N-CNTs indicate the as-grown N-doped CNTs have a high purity and the average diameter is about 200 nm. From the TEM image in Fig. 3c, it can be seen that the inner cores of the CNTs are filled with long iron nanowires whose length is several micrometers. Fig. 3d shows the SEM image of the sample derived from Fe/N-CNTs sulfidation for 1 h. Although it is hard to see any difference between pristine sample (Fig. 3a) and the vulcanized one (Fig. 3d, the formed pyrite is at a low content about 20 wt%), TEM observations (Fig. 3g, h) confirm that the sulfur element really has reacted with the iron nanorods and an obvious volume expansion has happened. As the degree of vulcanization increases, many white nanoparticles or nanorods appear in SEM images (Fig. 3e, f) of the samples after sulfidation for 2 and 5 h. Since the volume expansion effect of FeS2, the wrapping parts of N-CNTs are damaged and two different FeS2/N-CNTs structures appear (scheme as Fig. 1). The one is FeS2nanorods attached to the half-opened CNTs as displayed in Fig. 3i, the other is FeS2nanoparticles loaded on tube inner walls as shown in Fig. 3j.

The elemental mapping (Fig. 4) of the typical sample FeS2/N-CNTs-45.2 wt% (vulcanized for 2 h) confirms a large amount of sulfur and iron existed in the sample, suggesting the white parts in SEM image (Fig. 3e) could be ascribe to iron sulfide. The elemental mapping also shows that nitrogen atoms were distributed evenly in the sample. And the nitrogen content was 4.34 at% in the forms of pyridinic, pyrrolic and graphitic nitrogen as confirmed by XPS (Fig. 4f). We believe that the iron filling ratio and the defect level of tube walls are two essential factors that determine the finally structure. The long iron nanowires wrapped in the weak parts may unzip the tube walls while the short nanorods or nanoparticles filled in the strong part may come from the comparatively weak region.

Fig. 4 (a) SEM images of FeS2/N-CNTs-45.2 wt%, (b) C-, (c) N-, (d) S-, (e) Fe- elemental mapping of the entire region of (a), and (f) XPS of N1s spectrum of FeS2/N-CNTs-45.2 wt%. There are three peaks in the spectrum which belong to pyridinic nitrogen (398.5 eV), pyrrolic nitrogen (400.1 eV) and graphitic nitrogen (401.3 eV).

The above-mentioned characterization verified sulfur steam could react with iron confined inside carbon nanotube cavities. This result seems exotic and is different from the experiments which always affirm the protective effect of CNTs[32]. It was reported that nitrogen doping process generates defects in the walls and even Li+ions could diffuse into interwall space through them. So the presence of defects and vacancies in tube walls introduced by nitrogen doping were thought to be the main reason that brought about diffusion of the sulfur steam into the tube walls.

Fig. 5 (a) XRD pattern and (b) TEM image of carbon nanotubes filled with iron (Fe/CNTs), (c) TEM images of Fe/CNTs after sulfidation, (d) Raman spectra of Fe/CNTs and Fe/N-CNTs (excitation at 532 nm).

To confirm this point, a comparative experiment has been conducted using iron filled carbon nanotubes without nitrogen doping (Fe/CNTs) as pristine material under the same sulfidation condition. The Fe/CNTs was also synthesized through the FCCVD method as described above with melamine replaced by ethylene that was used to as the carbon source. XRD pattern (Fig. 5a) and TEM image (Fig. 5b) verify that there are also iron nanowires filled inside in the forms of a-Fe and Fe3C. After sulfidation treatment, it can be seen clearly from TEM image (Fig. 5c) that iron nanowires are stilled confined inside without alteration, implying the difficulty of sulfidation of Fe in Fe/CNTs.

Raman spectra are used to identify the structural differences between the two kinds of CNTs. From Fig. 5d, it can be seen thatDandGbands are around 1 350 and 1 575 cm-1respectively. TheDband is a double-resonance Raman mode, which can be understood as a measurement of structural disorder coming from any defects[34]. TheGband originates from the in plane stretching vibrations of the C—C bonds in graphitic materials. The intensity ratio of theDandGband (ID/IG) can be used to estimate the defect density of tube walls. TheID/IGvalue of Fe/N-CNTs is 0.92 much higher than the 0.46 of Fe/CNTs, proving that the nitrogen doped tube walls have more defects. Thus, the Fe/N-CNTs would be more easily sulfided than Fe/CNTs and results in the formation of FeS2/N-CNTs with an efficient lithium storage performance.

The electrochemical lithium storage properties of various FeS2/N-CNTs composites were tested as anode materials for LIBs. Fig. 6a shows the cyclic voltammograms (CVs) of FeS2/N-CNTs-45.2 wt% as an anode at a scanning rate of 0.1 mV/s in the 0-3.0 V voltage window. For the first cycle, a sharp reduction peak is observed at approximately 1.4 V, corresponding to the lithiation of pyrite described by the following reactions[35]:

FeS2+ 2Li++ 2e-→Li2FeS2

(1)

Li2FeS2+ 2Li++ 2e-→ 2Li2S + Fe

(2)

At ambient temperature, these two reactions proceed simultaneously due to the relatively slow diffusion of Li+ions into pyrite, which account for why only one peak is observed. Another broad peak centered at 0.7-0.9 V suggests the formation of solid electrolyte interface (SEI) layer on the electrode surface, which disappeared at subsequent cycles[34]. There are two oxidation peaks located at 1.9 V and 2.5 V, which were ascribed to the following reactions[35,36]:

Fe + 2Li2S → Li2FeS2+ 2Li++ 2e-

(3)

Li2FeS2→ Li2-xFeS2+xLi++xe-(0.5

(4)

Li2-xFeS2→ 0.8ortho-FeS2+ 0.2FeS8/7+ 0.175S + (2-x)Li++ (2-x)e-

(5)

The position of reduction and oxidation peaks changes slightly from the second cycle, especially one additional small reduction peak appeared at 2.0 V. It means a change in Li+/FeS2reaction system related to the production of charge products like sulfur and so on[37].

Fig. 6 (a) Cyclic voltammograms (CVs) of FeS2/N-CNTs-45.2 wt% at a scan rate of 0.1 mV s-1 between 0.01 and 3.0 V, (b) Charge and discharge voltage profiles of FeS2/N-CNTs-45.2 wt%, (c) Comparative cycling performance of FeS2/N-CNTs-22.4 wt%, FeS2/N-CNTs-45.2 wt% and FeS2/N-CNTs-55.8 wt% tested at a current density of 0.1 A g-1, (d) Rate capability of FeS2/N-CNTs-45.2 wt% tested at various current densities ranging from 0.2 to 2 A g-1, and (e) Cycling performance of FeS2/N-CNTs-45.2 wt% at a current density of 0.5 A g-1. The inset in panel (e) shows the electrochemical impedance spectra of FeS2/N-CNTs-45.2 wt% and FeS2/N-CNTs-55.8 wt% by applying a sine wave with an amplitude of 5.0 mV over the frequency range from 100 kHz to 0.01 Hz.

Charge and discharge voltage profiles of FeS2/N-CNTs-45.2 wt% electrode are shown in Fig. 6b at a current density of 0.1 A g-1in the voltage range of 0.01-3.0 V. There is an obvious plateaus located at 1.3-1.6 V in the first discharge profile in according with the first reduction peak in CV curve. And two plateaux located at 1.8-2.0 V, 2.4-2.6 V in the charge profile correspond to the two oxidation peaks mentioned above. The electrode delivers a high initial discharge capacity of 913 mAh g-1and a reversible capacity of 653 mAh g-1with a first cycle coulombic efficiency of 71.5%. The irreversible part of 28.5% could be ascribed to the formation of SEI film and electrolyte decomposition which are common to other anode materials[38]. An interesting phenomenon is observed that the capacity of FeS2/N-CNTs-45.2 wt% electrode decreased in the first 10 cycles and then increased afterwards as shown in Fig. 6c. As discussed in the CV curve, the lithium storage activity of FeS2mainly happened when the voltage is above 0.8 V. So the capacity below 0.8 V could be mostly ascribed to N-CNTs. From the discharge curves, it can be seen that the capacity of 2nd, 50th, 100thcycles are 808, 672, 996 mAh g-1, respectively. Among them, the capacity generated above 0.8 V are 437, 258, 449 mAh g-1, and 371, 414, 547 mAh g-1below 0.8 V. Based on the data statistics, it can be deduced that the increment of the capacity attributes to the activity improvement of both FeS2and N-CNTs, which could be explained by the following two main reasons. Firstly, the initial discharge induces a sudden volume expansion of the active layer due to a large amount of Li+ions inserted into the electrode materials, which inhibits the further Li+ions transfer from electrode to electrolyte[28,37,39].Thus, a part of Li2S would be trapped in the electrode material, leading to a decrease of capacity. After a number of cycles, the Li2S could be re-exposured to the electrolyte and release Li+ions as the structure and morphology change of FeS2/N-CNTs, and electrode would be fully infiltrated by the electrolyte in the long activation process, resulting in the capacity growth[40,41].This phenomenon is normally observed in transition metal oxide-based anode materials for LIBs[37,40,42], which is attributed to the reversible formation of a polymeric gel-like film originating from kinetic activation of the electrode. In addition, nitrogen doping in the tube walls and the volume expansion of iron during the sulfidation process have generated many defects and chinks on the tube walls through which Li+ions could penetrate into the inner tube space and increased the lithium storage capacity. The capacity increment phenomenon can also be observed in the FeS2/N-CNTs-22.4 wt% and FeS2/N-CNTs-55.8 wt% samples but their discharge capacities are 607 and 820 mAh g-1respectively, lower than the capacity of FeS2/N-CNTs-45.2 wt%. N-CNTs and FeS2both contribute to the capacity. The low capacity of FeS2/N-CNTs-22.4 wt% is easy to understand as there are a large amount of unreacted Fe and Fe3C nanoparticles, which have very low lithium storage capacities[8]. But the FeS2/N-CNTs-55.8 wt% sample has a higher FeS2content but a lower capacity. To find the reasons, electrochemical impedance spectroscopy (EIS) measurements were taken (inset in Fig. 6e). The diameter of the semicircle for FeS2/N-CNTs-45.2 wt% electrode in the high-medium frequency region is smaller than that of FeS2/CNTs-55.8 wt%, meaning that the former electrode has a lower contact and charge-transfer impedance thus a higher capacity is expected[43]. In addition, there is a synergistic effect between FeS2and CNTs. The CNTs could not only improve the electron transfer ability of anode materials, but also inhibit the shuttle effect of lithium sulfides. However, their sulfur confinement ability cannot be reflected with a too much loading of FeS2. The FeS2nanocrystals have two forms of structures. The one is FeS2nanorods attached to the half-opened CNTs, the other was FeS2nanoparticles loaded on tube walls. Different contents of FeS2result in different ratios of the two forms of structures in different samples. FeS2nanoparticles are more easily infiltrated by the electrolyte and provide a larger capacity. Generally, the discharge capacity of pristine FeS2microspheres decreases quickly to less than 100 mA h·g-1within 50 cycles[37]. However, in this work, due to the strong adhesion force between FeS2nanoparticles and carbon nanotube walls, the discharge capacity didn’t decrease quickly. The FeS2nanoparticles encapsulated in CNTs with a high specific capacity of 525 mAh g-1after 1 000 cycles at 2 000 mA g-1has been demonstrated[28]. Thus FeS2nanorods attached to the half-opened CNTs exhibit a very stable capacity retention.

In addition, the rate capability of FeS2/N-CNTs-45.2 wt% at various current densities from 0.2 to 2 A g-1was examined and the result is shown in Fig. 6d. Benefited from the unique hybrid structure, discharge capacity of 340 mAh g-1could still be achieved even cycled at a high current density of 2 A g-1. When the current density restored to 0.2 A g-1, the discharge capacity arose to 665 mAh g-1. More importantly, FeS2/N-CNTs-45.2 wt% exhibits a good cycling performance at a current density of 0.5 A g-1as shown in Fig. 6e. At the initial 80 cycles, the capacity slowly decreases due to the trapping of Li2S in the electrode material. Then the capacity quickly increases to 756 mAh g-1from 80 to 180 cycles and comparatively stable cycling could be retained from then on. The FeS2/N-CNTs performs a high discharge capacity (996 mAh g-1/0.1 A g-1after 100 cycles) and the final capacity was 815 mAh g-1after 400 cycles, much higher than the capacities reported using FeS2/C, FeS/C or other sulfide/C composites as anode materials[14,23,28,37,39]. The superior lithium storage ability of FeS2/N-CNTs could be ascribed to the synergistic effect of many factors. Firstly, the iron sulfide nanowires and nanoparticles firmly combined with carbon tube walls, which improve their electrical conductivity and inhibit their agglomeration during charge and discharge process. Secondly, nitrogen doping introduces defects to the tube walls, which provide paths for the sulfur steam into the tube cavities in synthesis, and also diffusion paths for Li+ions into the interwall space. And the doped nitrogen atoms especially pyridinic N and pyrrolic N (Fig. 4f) have the ability to form SxLi…N interactions via the N lone-pair electrons thus alleviate the dissolution of lithium polysulfides in the electrolyte, which are found in the Li/S cells. Finally, cracking or unzipping of tube walls induced by the volume expansion of FeS2during lithiation shorten the path of Li+diffuse into the inner tube walls or laden FeS2.

4 Conclusions

The FeS2/N-CNTs hybrid structure has been formed through a facile vulcanization process with Fe/N-CNTs as the raw materials. The introduction of nitrogen atoms into tube walls generates defects through which the sulfur atoms could penetrate into the tube cavities and react with the iron filled inside. The obtained hybrid could be used as anode materials for LIBs. A high capacity of 815 mAh g-1has been obtained at a current density of 500 mA g-1after 400 cycles. And a high capacity of 340 mAh g-1could still be achieved at a high current density of 2 A g-1. The excellent performance should be attributed to a variety of synergic effect like the firm binding of FeS2with N-CNTs, nitrogen doping and cracking or unzipping of tube walls. The unusual semi-confinement effect of N-CNTs makes the transformation of element confined inside tube cavities easy and opens a new way to fabricate novel anode materials for high performance LIBs.