Effect of Continuous Capacity Rising Performed by FeS/Fe3C/C Composite Electrodes for Lithium‐Ion Batteries

Abstract FeS‐based composites are sustainable conversion electrode materials for lithium‐ion batteries, combining features like low cost, environmental friendliness, and high capacities. However, they suffer from fast capacity decay and low electron conductivity. Herein, novel insights into a surprising phenomenon of this material are provided. A FeS/Fe3C/C nanocomposite synthesized by a facile hydrothermal method is compared with pure FeS. When applied as anode materials for lithium‐ion batteries, these two types of materials show different capacity evolution upon cycling. Surprisingly, the composite delivers a continuous increase in capacity instead of the expected capacity fading. This unique behavior is triggered by a catalyzing effect of Fe3C nanoparticles. The Fe3C phase is a beneficial byproduct of the synthesis and was not intentionally obtained. To further understand the effect of interconnected carbon balls on FeS‐based electrodes, complementary analytic techniques are used. Ex situ X‐ray radiation diffraction and ex situ scanning electron microscopy are employed to track phase fraction and morphology structure. In addition, the electrochemical kinetics and resistance are evaluated by cyclic voltammetry and electrochemical impedance spectroscopy. These results reveal that the interconnected carbon balls have a profound influence on the properties of FeS‐based electrodes resulting in an increased electrode conductivity, reduced particle size, and maintenance of the structure integrity.


Introduction
The ever-growing demand for portablee lectronic devices and large-scale energy storages ystems has promoted the great successo fl ithium-ion batteries (LIBs) owing to their superior energy density,r eliable stability,a nd long cycle lifetime. Further improvement of energy/power density and long cycling behavior of LIBs requires technological innovation of the electrode materials. [1][2][3][4] Promising negative electrode materials are based on ac onversion mechanism. [5] This mechanism resultsi n significantly higherc apacities of more than 1000 mAh g À1 [6] compared with, for example, graphite with at heoretical capacity of 372 mAh g À1 . [7] Iron sulfidesh ave attracted much attention as electrodes for LIBs because of their high theoretical gravimetric capacities (609 mAh g À1 forF eS and 894 mAh g À1 for FeS 2 ), naturala bundance,a nd low cost. [8][9][10] However,i ron sulfide suffers from huge capacity decay owing to the sluggish electrode kinetics and drastic intrinsic volume expansion during the repeated lithiation and delithiation processes. It was reported that Li-ion insertion into the FeS material results in a2 00 %v olume change, leading to electrode pulverization and thus to ad rastic reduction of the electrical contactt ot he currentc ollector. [11] More deleteriously,i nsulating polysulfides Li 2 S x (2 < x < 8) are formed during lithium-ion insertioni nto iron sulfide materials. [12,13] Such polysulfides are partially soluble in organic electrolyte [14] and can gradually form an insulating layer on the surface of the electrode, whichs eriouslyd ecreases the conductivity among active materialp articles and hinders further electrochemical reactions. [15,16] To solve the aforementioned problems, some useful strategies have been implemented:o ne is to improvet he electrochemicalp erformance by downsizing the particle size or fabricating nanostructures; [17,18] another strategy is to control the composition with carbon materials (such as carbon nanotubes, [19] graphene, [20] conductive polymers, [21] and carbon fibers) [22] to improvee lectron conductivity and structure stability.T wo-dimensional (2D) structures, such as flakesa nd sheets can effectively increase the contact between active materials and electrolyte, buffer the volumetric fluctuations,a nd decrease the diffusionl ength of lithium ions ande lectrons during the lithiationa nd delithiation processes. [17,23] For example, Xu et al. [16] prepared carbon-coated FeS nanosheetsb yasurfactant-assisted solution-based synthesis, which deliver excellent Li storage properties (615 mAh g À1 at as pecific current of 100 mA g À1 ). The ultrathin FeS nanosheets can accommodate the volumee xpansion and shorten the diffusion paths.I ti s featured that the added carbon materialc an provide fast electron/ion transport and promote the formation of as table solid FeS-based composites are sustainable conversion electrode materials for lithium-ion batteries, combining features like low cost, environmental friendliness, and high capacities.H owever, they suffer from fast capacity decay and low electron conductivity.H erein, novel insights into as urprising phenomenono f this material are provided.AFeS/Fe 3 C/C nanocomposite synthesized by af acile hydrothermal methodi sc omparedw ith pure FeS. When applied as anode materials for lithium-ion batteries, these two types of materials show different capacity evolution upon cycling.S urprisingly,t he composite delivers a continuousincrease in capacity instead of the expected capacity fading. This unique behavior is triggered by ac atalyzing effect of Fe 3 Cn anoparticles. The Fe 3 Cp hase is ab eneficial by-product of the synthesis andw as not intentionally obtained. To furtheru nderstand the effect of interconnected carbon balls on FeS-based electrodes, complementary analytict echniques are used. Ex situ X-ray radiation diffraction and ex situ scanning electron microscopy are employedt ot rack phasef raction and morphology structure. In addition, the electrochemical kinetics and resistance are evaluated by cyclic voltammetry and electrochemical impedance spectroscopy.T hese results reveal that the interconnected carbon balls have ap rofound influence on the properties of FeS-based electrodes resulting in an increasedelectrode conductivity,reduced particlesize, and maintenanceofthestructure integrity. electrolyte interphase (SEI) layer on the active material surface during the electrochemical processes. [9,24,25] However, ad etailed understanding of the kinetic phenomenaa nd resistive contribution of FeS-based materials during long-term cycling and charge/discharge at high specific current (i.e.,h igher than 1Ag À1 ) is stilln ot established. Moreover,t he design of Fe 3 Ccontaining composite materials was proposed as beneficial by differentr esearch groups. Su et al. [26] presented ac ore-shell Fe@Fe 3 C/C composite and attributed the observed extra capacity beyond the carbon component to reversible redox reactions of some SEI components.T hese reactions werep roposed to be catalyzed by Fe 3 Cn anoshells. In another work by Zhang et al., [27] as imilar positive contribution of Fe 3 Co nF e 3 O 4 @Fe 3 C core-shell nanoparticles was assigned to the stabilizationo f Fe 3 O 4 particle integrity.W hat's more, Chan et al. [28] prepared Fe/Fe 3 C/NPGC with highcatalytic activity for enhanced bioelectricity generation. Bothg roups intentionally added the Fe 3 C component to their composite.
In this study,w eh ave synthesized FeS nanosheets and FeS/ Fe 3 C/C nanocomposites consisting of well-dispersed FeS and Fe 3 Cn anoparticlesa nd interconnected carbon balls by af acile hydrothermalm ethoda nd as ubsequent sintering process. The Fe 3 Cn anoparticles were formed as ab yproduct but demonstrate ap ositive influence on the electrochemical performance. The FeS electrode continuously declines in capacity and exhibits at errible capacity retention, whereas the capacity of the FeS/Fe 3 C/C electrode demonstrates af luctuation prior to the 140th cyclea nd ac ontinuous increase during further cycling. This is the first time that this kind of interesting behavior is observed. This manuscript focuseso nacomprehensive and indepth investigation into the effect of the interconnected carbon balls-FeS and their property relationship.F or this purpose, as eries of electrochemical, physical, and morphological characterization techniques are employedt ou nderstand the influenceo ft he interconnected carbon balls on FeS-based electrodes.

Structural and morphological characterization
The phase fraction and crystal structure of FeS and FeS/Fe 3 C/C are evaluatedb yX -ray diffraction (XRD) and analyzed by the Rietveld refinement by using the FullProf softwarep ackage, as shown in Figure 1a    Cc omposite is composed of FeS and Fe 3 Cn anoparticles with as ize of 50-60nma nd interconnected carbon balls of 1-2 mm (Figure 2c,d). Herein, FeS nanoparticles are surrounded by interconnected carbon balls, which are expected to providet he paths for electronm ovement and effectively buffer the volume expansion upon repeated cycling. The correspondinge lemental mappingo fF eS/Fe 3 C/C ( Figure S1 d-g in the Supporting Information) shows that the nanosizedF eS nanoparticles are randomly scattered in the interconnected carbon ball matrix. To better understand the structure of FeS nanosheets and FeS/Fe 3 C/C composites, Raman spectroscopy of FeS and FeS/Fe 3 C/C were conducted and are shown in Figure S2 (in the Supporting Information). Both FeS nanosheets and FeS/Fe 3 C/C composites have two peaks located at 218 and 290 cm À1 ,w hich are attributed to the asymmetrica nd symmetric stretching modes of FeS. [29] In the FeS/Fe 3 C/C composites, two distinct peaks are presented at 1315 and 1600 cm À1 ,w hich are relatedt ot he Db and and Gb and of amorphous carbon with the intensity ration of I D /I G = 1.0, implying that the interconnected carbon balls have ah ighly disordered carbon structure. [30] The Db and is linked to disordered carbon atoms and defects, whereas the Gb and is due to the relative motion of sp 2 carbon atoms. [31,32] The carbon percentage in FeS/Fe 3 C/C is 55 %, which is calculated from the organic elemental analysis (OEA) measurements (see the Supporting Information, Ta ble S1). These results reveal that the introduction of interconnected carbon balls in FeS drastically affects the phase fraction, the morphology,a nd the particle size.

Electrochemical performance and kinetic processes
As huge disparities between the FeS nanosheets and FeS/Fe 3 C/ Cc omposites were observed by the previously described analytics, ad ecisively different electrochemical behavior is expected for the respective electrodes. To better understand the lithium-storage mechanism taking place in the electrodes, cyclic voltammetry (CV) measurements werec onducted at as can rate of 0.05 mV s À1 in the voltage range from 0.01 to 3.0 V( vs. Li + /Li)f or the FeS and FeS/Fe 3 C/C electrodes, respectively.F igure 3a,b shows the CV profileso ft he first five cycles.
The related equations from previous reports, [3,17,[33][34][35] representing state-of-the-art lithium insertion in FeS and Fe 3 Ca re shown below: The lithium storagem echanism for Fe 3 Ci sb ased on ac onversion mechanism;i ti sr eported that only 1/6 Li per unit can insert into the Fe 3 Cm aterial ( % 26 mAh g À1 ), [35] and the phase fraction of Fe 3 Ci nF eS/Fe 3 C/C is 23 %. It is expected that the capacityc ontribution from Fe 3 Cf or the FeS/Fe 3 C/C electrode is negligible. Considering the CV of the FeS electrode, three reductionp eaks appear at 1.7 V, 1.26 V, and 0.76 V, whereas only one oxidation peak appearsa t1 .88 Vd uring the first scan. According to previousr eports, [17,33,34] the small peak at 1.7 Vc orrespondst ot he formation of the intermediate phase Li 2 FeS 2 duringt he Li + insertion into the FeS bulk [Eq. (1)].T he sharp peak at around1 . 26 Vi sr elated to the conversion reaction belongingt ot he formation of metallic Fe nanocrystals and Li 2 S matrices [Eq. (2)]. [16] The broad peak at 0.76 Vi sa ssigned to the formation of asolid electrolyte interphase (SEI) on the electrode surface. [36] In the first anodic process, the oxidation peak at 1.88 Vc orresponds to the oxidation of metallic Fe to form Li 2Àx FeS 2 [Eq. (3)]. [23] In the subsequent cycles, the reduction peaks at 0.76 and 1.26 Vs hift to 0.79 and 1.42 V, respectively, and the oxidation peak at 1.88 Vs hifts to 1.91 V. These changes indicate that some irreversible reactions occur during the first electrochemical process. During the second to fifth cathodics cans, an ew broad reduction peak appearsa t2 .0 V and can be related to the phase transformation from Li 2Àx FeS 2 to Li 2 FeS 2 [Eq. (4)]. [15] The sharp peak at 1.42 Vc orresponds to the lithiationp rocess [Eq. (5)].C orrespondingly,t he oxidation peak at 1.91 Vi nt he second cycle accounts for the reversible delithiation process from Li 2 FeS 2 to Li 2Àx FeS 2 [Eq. (4)]. [16,17] Upon the first five CV scans,t he intensity of the redox peaks gradually decreases, indicating that the capacityd ecreases. This might result from unstable formation of the SEI layer and the sluggish reactionk inetics of pure FeS nanosheets. [3,37] The CV curveso ft he FeS/Fe 3 C/C electrode are similar to those of FeS except for the additional broad cathodic peak at 0.39 V, which is attributedt os ide reactionsb etween the FeS/Fe 3 C/C materiala nd the electrolyte and SEI formation. [2] The conversion reaction between Fe 3 Ca nd Li can be describea sE quation (6), with lessc apacity contribution. Moreover,t he intensity of the reduction/oxidation peaks are much weaker owing to the interconnected carbon balls. Comparing the CV curveso f FeS and FeS/Fe 3 C/C electrodes, the peak current densities and the peak potentials of the FeS/Fe 3 C/C electrode barely change after the first cycle. This points out that ab etter structural stability and good reversibility is accomplished for the FeS/Fe 3 C/C electrode. In addition, the polarization voltage of the FeS/Fe 3 C/ Ce lectrode is 0.46 V, which is lower than that of the FeS electrode (0.50 V). This demonstrates that the interconnected carbon ball morphology improves the conductivity of the FeS/ Fe 3 C/C electrode, leadingt oar educed electrode polarization. Figure 3c and dd isplays the lithiation/delithiation profileso f FeS and FeS/Fe 3 C/C electrodes at the first, second, fifth, 35th, 80th, 140th,2 25th, and 500th cycle at as pecific current of 1Ag À1 .D uring the first lithiation of the FeS electrode, al ong potentialp lateau at around 1.3 Vand as hort potentialp lateau at 0.8 Va re observed, which correspond to the lithiation process forming Li 2 S, Fe, and the SEI layer formation, respectively. During the delithiation process, the long potentialp lateau at 1.8 Vi sr elatedt ot he formation of Li 2Àx FeS 2 .A ll thesep otential plateausa re in agreement with the peaks observed in the CV curves. The FeS electrode delivers af irst lithiation capacity of 900 mAh g À1 and ad elithiation capacity of 782 mAh g À1 with a coulombice fficiency of 86.9 %. An irreversible capacity of 118mAh g À1 at the first cycle results from the inevitable formation of the SEI film on the surfaceo ft he active materiala nd electrolyte decomposition. [37] In the second and fifth cycles, the reduction and oxidation plateauss hift to 1.4 and 1.9 V, respectively.T he reason is an increased electrode polarization. In the second cycle, the FeS electrode delivers ad elithiationc apacity of 763 mAh g À1 whereas the coulombic efficiency in-creasesto96.6 %. After the fifth cycle, it still delivers alithiation capacityo f7 60 mAh g À1 and ad elithiation capacity of 738 mAh g À1 with the coulombic efficiency of 97.1 %. Examining the subsequent cycles of the FeS sample (the 35th, 80th, and 140th), ad rop of the reduction plateau to lower potentials accompanied by ag raduald ecrease in delithiation capacity can be observed. This behavior is attributed to an enhancement of the electrode polarization during cycling. Furthermore, in the 225th and 500th cycles, no pronounced potentialp lateau is observed, implying structural changes or pulverization. After the 500th cycle, the voltage plateau disappearsa nd the electrode delivers av ery low delithiation capacity of 150 mAh g À1 .T his fact confirms that the structure of FeS nanosheets is destroyed. When comparing the FeS sample with the FeS/Fe 3 C/C one, the potentialp lateau of the FeS/Fe 3 C/C electrode is much shorter.D uring the first lithiation process, two plateausa re located at 1.3 and 0.8 Vc orresponding to the formation of Li 2 S, Fe, and SEI layers,r espectively.Abroad peak at 1.20 Vc orresponds to the oxidation of Fe 0 andavery short po-tentialp lateau at 1.8 Vr elated to Li 2Àx FeS 2 formation is observed in the first delithiation process.I nt he first cycle, the FeS/Fe 3 C/C electrode shows lithiation and delithiation capacities of 946 and 530 mAh g À1 ,r espectively,a nd ac oulombic efficiencyo f5 6%.T he huge irreversible capacity of 416 mAh g À1 is attributed to the SEI layer formation, electrolyte decomposition, and side reactions. In the subsequently cycles, the reduction plateau shifts from 1.3 to 1.4 Va nd the long slope disappears, implying that an irreversible reactiono ccurred. In the second cycle, it delivers ad elithiation capacity of 490 mAh g À1 and the coulombic efficiency increases to 93 %w hereas the fifth cycle showsalithiation capacity of 476 mAh g À1 and ad elithiation capacity of 463 mAh g À1 with ac oulombic efficiency of 97 %. It becomes apparent that the length of the potential plateau decreases upon cycling (the 5th, 35th,8 0th, and 140th cycles).A tt he 140th cycle, the potential plateau disappears and the electrode shows the lowest capacity of 395 mAh g À1 at this state. Interestingly,acapacity increaseo ft he FeS/Fe 3 C/C electrode is observed upon further cycling. It is important to note that the lithiation/delithiation profiles after the 140th cycle with no clear plateau are strongly different from that of the first five cycles. Moreover,t he specific capacity increases to 800 mAh g À1 .A ccording to the conversion mechanism for lithium storage, Fe 3 Cc an insert only 1/6 Li per unit (26 mAh g À1 ), which is negligible compared with the high capacity of 800 mAh g À1 . [35] This novel phenomenon has been observed for the first time and will be discussed hereafter in detail.
To evaluate the effect of interconnected carbon balls on FeS-based electrodes, rate performances wereapplied at different specific currents in the voltage range from 0.01 to 3.0 V (vs. Li + /Li)f or the FeS and FeS/Fe 3 C/C electrodes and are shown in Figure 4a andb ,r espectively.O nt he one hand, the FeS electrode delivers 874, 819, 748, 674, 624, and 460 mAh g À1 at specific currents of 0.1, 0.2, 0.5, 1, 2, and 5Ag À1 ,r espectively.W hen the specific current is set back to 0.1 Ag À1 ,t he specific capacity reaches 788 mAh g À1 .O nt he other hand, the FeS/Fe 3 C/C electrode delivers 815, 676, 610, 572, 532, and 457 mAh g À1 at specific currentsof0 .1, 0.2, 0.5, 1, 2, and 5Ag À1 ,r espectively.A st he specific current returns to 0.1 Ag À1 ,t he specific capacity returnst o7 24 mAh g À1 .F igure 4c directly compares the capacity values of FeS and FeS/ Fe 3 C/C electrodes at the various specific currents.B etween 0.1 and 2Ag À1 ,t he FeS electrode displays higherc apacities with respect to the FeS/Fe 3 C/C electrode. However,a tt he highest specific current (5 Ag À1 ), both electrodes deliver the same specific capacity of 460 mAh g À1 .F urthermore, the long-term cycling performance of FeS and FeS/Fe 3 C/C electrodes were investigated at specific current of 1Ag À1 for the potentialr ange 0.01-3.0V (vs. Li + /Li;F igure 4d). Both electrodes display ad ifferent behavior upon cycling. Despite the FeS electrode initially showing am uch higherc apacity compared with the FeS/ Fe 3 C/C electrode, its capacity rapidly fades to 130 mAh g À1 . Apart from that, the FeS/Fe 3 C/C electrode shows ac apacity fluctuation during the first 140 cycles.A fter that, the capacity increases to 800 mAh g À1 after 500 cycles. Apossible reason for such extra capacity could be the catalytic activation of Fe 3 C. It is reported that Fe 3 Cp lays the role of catalyst to decrease the ChemSusChem 2020, 13,986 -995 www.chemsuschem.org 2020 The Authors. Publishedb yWiley-VCH Verlag GmbH &Co. KGaA, Weinheim concentration of SEI components and boost the reversible formation/decomposition of the SEI layer upon cycling. [26,35] In addition, the FeS/Fe 3 C/C particles shrink owing to pulverization, leadingt oh igher electrochemical efficiency.A dditionally,s maller particles can lead to higherc apacities owing to less inactive materialp arts. Most plausible, the abovementioned trends result from the pseudo-capacitive behavioro f the material and therefore improvethe electrochemical kinetics.
To prove the transformation of diffusion-controlled behavior to ap seudo-capacitive energy storage process after 140 cycles, CVs measurement were conducted at variouss can rates. This experiment should allow insights into the storage mechanism during the initial cycling. Figure 5a and bs hows the CV profiles of the FeS and FeS/Fe 3 C/C electrodesa t scan rates between 0.05 and 10 mV s À1 .A se xpected, the peak currenti ncreases with the increase of sweep rate. The current (i)i sr elated to the scan rate (v)t hrough the relation: i ¼ av b .G enerally, b = 0.5 implies ad iffusion-controlled process, whereas b = 1 represents ac apacitive process. [38] The cathodic and anodic peaks in Figure 5a and bw ere chosen to calculate the b value by using the equation log (i) = b log (v) + log (a). [39] Based on the value of b,w ec an distinguish if the lithiation and delithiation are diffusion or surfacec ontrolled. Figure 5c,d presents the linear relationship between the log (i)a nd log (v)a t cathodic and anodic peaks for FeS and FeS/Fe 3 C/C electrodes, respectively.A fter linear fitting, the calculated b value of the cathodica nd anodic peaks for the FeS electrode are 0.40 and 0.46, respectively;w hereas those for the FeS/Fe 3 C/C electrode are 0.53 and 0.64, respectively.A se xpected, this analysis confirms that the ion-diffusion behavior controls the electrochemical process in both FeS and FeS/Fe 3 C/C electrodes for the initial cycles. The diffusion-controlled mechanism can explain the initial capacity fading. Enhanced stress is applied to the FeS active material by diffusion compared with as urface-controlled process. Therefore, an initial pulverization and phase amorphization of both samples is expected, leading to ac ontact loss between the particles, which resultsi na ni nferior percolation and increased resistance. We expect that the interconnected carbon ball matrix of the FeS/Fe 3 C/C sample buffers the pulverization anda llows as table change to a pseudo-capacitive mechanism during cycling. In this case, the interconnected carbon balls undertake the role of ac onductive matrix with pulverized FeS nanoparticles distributed on it. These particles exhibit pseudo-capacitive behavior,w hich is indicated by the loss of the plateaus and the simultaneous capacity increase. To prove this interpretation, furtherc ycling experiments and ex situ measurements were conducted and are presented in the following section.

Phase fraction, morphology,a nd electrochemical performanceoft he electrode after cycling
The above resultsh ighlight that the lithium storage mechanism strongly changes during cycling. To un-  derstandt his mechanism,t he FeS and FeS/Fe 3 C/C electrodes were analyzed after the 140th galvanostatic cycle by complementaryt echniques such as CV,e xsitu XRD, and SEM. To understandt he differences betweenf resh and cycled electrodes, CV measurements were performed after different cycle numbers and are shown in Figure 6a and c. It can be observed that after the 140th cycle at 1Ag À1 the redox peaks become weaker forb oth the FeS (Figure 6a)a nd the FeS/Fe 3 C/C electrode (Figure 6c). Thisd emonstrates an unstablec ycling feature resulting from irreversible phase transitions and pulverization. It is noted that in the FeS/Fe 3 C/C electrode, the oxidation peak becomes broader and shifts (from 1.89 to 2.05 V) after 140 cycles, which indicates that the FeS/Fe 3 C/C electrode has higher resistance upon cycling. This result was analyzed in depth by electrochemical impedance spectroscopy (EIS) measurements and will be shownl ater (Figure 10 b,d). The oxidation/reduction peaks nearly disappear for the FeS electrode, implyingt hat the faradaic reactiond oes not exist anymore. More interestingly,t he cathodic peak (located at 0.84 V) of the FeS/Fe 3 C/C electrode appearsa tt he fifth CV cycle of the fresh electrode owing to the formation of the SEI. However,t here is av ery broad cathodic peak in the fifth CV recorded after 140 cycles, which is relatedt os ome irreversible reactions. Figure 6b and dp resent the CVs of FeS and FeS/Fe 3 C/C electrodes at different sweep rates after the 140th cycles. Cycled electrodes show no clear peaks comparedw ith that of fresh ones (Figure 5a,b). This confirms that the amorphous phase exists in the cycled electrode.
To further investigate the phase fraction and phase transition during the repeated lithiation/delithiationp rocesses, ex situ XRD of the FeS/Fe 3 C/C and FeS electrodes at the 9th, 140th, and 500th cycles was performed and are shown in Figure 7a,b. The FeS and FeS/Fe 3 C/C show sharp and clear XRD reflection patterns (Figure 1a,b), whereas XRD reflection patterns of the cycled FeS/Fe 3 C/C electrode show broad peaks, corresponding to Li 2 Sa nd Li 2Àx FeS 2 .T he reason behindt his is that during repeated lithiationa nd delithiationp rocesses, the material becomes amorphous with small crystallite size. In contrast, the XRD reflection of the FeS electrode becomes sharper (such as 308,3 2 8, and 378), indicating that the particles ize increases upon cycling.
To further demonstrate the reason behind the electrochemical performance differenceb etween FeS and FeS/Fe 3 C/C electrodes, the morphological changes of the cycled FeS and FeS/Fe 3 C/C electrodes are investigated by ex situ SEM and are shown in Figure 8. Figure 8a-c shows the ex situ SEM of the FeS electrode at the ninth, 140th, and 500th cycles. Compared with the fresh FeS pristine material ( Figure 2a,b), the morphology of the cycled FeS particle undergoes an irreversible change. Comparing the ex situ SEM images of the FeS electrode at different cycles, it is interesting that some small clusters of FeS particles tend to agglomerate andf orm al arge bulk, especially in the regionsI ,I I, and III. The ex situ SEM image of the ninth cycle is composed of nanoparticles and many holes (region I, Figure 8a), whereas the FeS particlesc rowd together and the holes disappear after the 140th cycle (region II, Figure 8b). Finally,t he particles further agglomerate, forming al arge bulk at the 500th cycle (region III,   Figure 8c). The nanoparticle agglomeration is not beneficialf or the reaction between the active materiala nd electrolyte during cycling;t his can explain the specific capacity decrease of the FeS electrode upon cycling (Figure 4d). Moreover,o ne can see that the ex situ SEM image of the FeS/Fe 3 C/C particle at the ninth (Figure 8d)c ycle shows the large clusters agglomerate, which are wrappedw ith SEI films similart ot he FeS electrode. The particles transform into smaller ones (Figure 8e,f, at the 140th and 500th cycles)a nd tend to interconnect with each otheru pon cycling. These smaller particles are equally distributed with the interconnected carbon balls, which can enlarget he contact between the active materiala nd electrolyte, thus resulting in the high efficiency of the electrochemical reaction, which is the mostp robable reason for the capacity increaseupon cycling.

Electrochemicali mpedance spectroscopy evolution
EIS was performed to examine the kinetics of Li + insertion/deinsertionp rocessesu pon cycling. Figure 9a and bs hows the Nyquistp lots of the FeS electrode at different cycles (1st, 50th, 100th, 150th,a nd 200th) at the bias potentialo f0 .86V (vs. Li + /Li)d uring lithiationa nd delithiation processes, respectively; correspondingly,t he Nyquist plots of the FeS/Fe 3 C/C electrode are shown in Figure 9c,d. The inset pictures in Figure 9d isplay the zoom in on the high-frequency area. It is found that the EIS dispersions present commonc haracteristics: (1) as mall semicircle in the high-frequency region,c orresponding to the passivation layer;( 2) ap artially overlapped semicircle in the intermediate-frequency region, corresponding to the chargetransfer process and chargea ccumulationa tt he electrical double layer;( 3) as lopingl ine in the low-frequency region, which corresponds to the solidd iffusion of lithium into the nanoparticle. [40,41] The Nyquit plots of FeS and FeS/Fe 3 C/C electrodes were fitted by using an equivalent circuit described as R el (R SEI C SEI )( R CT C CT )W in Boukamp's notation [42] and shown in Figure 9e. R el represents the electrolyte resistance (including separator and internal connections), R SEI and C SEI are assigned to SEI resistance and capacitance, R CT and C CT are related to the charge-transfer resistance and doublel ayer capacitance, and W (alfa = 0.5) is attributed to the Warburg impedance. It is worth notingt hat the overall impedance of the FeS and FeS/Fe 3 C/C electrodes show ar esistance increasei nb oth lithiation and delithiation conditions upon cycling. Comparing the Nyquist plots of the FeS and FeS/Fe 3 C/C electrodes at somes elected cycles (i.e.,t he 1st, 100th, and 200th) in lithiationa nd delithiation conditions ( Figure S3 in the Supporting Information), it is observed that the diameter of the semicircle for the FeS/Fe 3 C/ Ce lectrode is smaller than that for the FeS electrode. This demonstrates that the FeS/Fe 3 C/C electrode has rapid electrochemicalr eactionk inetics, which benefits from the addition of the interconnected carbon balls. Moreover,t he slope in the low-frequency region for the FeS/Fe 3 C/C electrode is larger than that for the FeS electrode, which implies faster Li + mobility in the FeS/Fe 3 C/C electrode. [43,44] Figure10d isplays R values as calculated with Relaxis 3s oftware in lithiationa nd delithiation conditions. The electrolyte resistance R el for the FeS/Fe 3 C/C electrode (11 W)i sa lmost unchanged upon cycling, whereas the R el for the FeS electrode first increases until the 100th cycle, then it remains quite stable at 8 W (Figure10a,d). The smalld ifferenceo ft he R el value between the FeS/Fe 3 C/C and FeS electrode is probably due to connections inside the cell. The calculated value of R SEI is shown in Figure 10 b,e. For the FeS/Fe 3 C/C electrode, R SEI slightly increases in both lithiation andd elithiation conditions upon cycling. Furthermore,t he R SEI in lithiation conditions is highert han that in the delithiation state, indicating the dynamic nature of the SEI layer,which grows in the lithiation process and partially decomposesi nthe delithiation process. [45,46] The R SEI of the FeS electrode drastically increases during the first 100 cycles,t hen it decreases but remains still higher than that of the FeS/Fe 3 C/C electrode. It is demonstrated that Fe 3 C exhibits great activity in promoting the partially reversiblef ormation/decompositiono ft he SEI layer. [27] R CT of the FeS/Fe 3 C/C electrode slightly decreases in lithiationc onditions upon cycling, and the R CT value remains almosts table in delithiation conditions. The rapid charge-transfer kinetics of FeS/Fe 3 C/C may benefit from ap artial nanoparticle reaggregation, which mostly occurs during the initial cycles ( Figure 8c)w hereas the morphology appears stabilized upon furtherc ycling. [47] In contrast, R CT of the FeS electrode sharply increases until the 100th cycle, corresponding to the terrible capacity decay ( Figure 4d); in subsequent cycles, R CT of the FeS electrode continuously declines but is still highert han that of the FeS/Fe 3 C/C electrode in lithiation conditions. In summary,t he interconnected carbon ball morphology can improveL i + /electron mobility and form a better protective SEI layer,t hus promoting the redox reaction. [48] Conclusion FeS nanosheets and FeS/Fe 3 C/C nanocomposites consistingo f well-dispersed FeS and Fe 3 Cn anoparticles and interconnected carbon balls were synthesized by af acile hydrothermal methoda nd as ubsequent sintering process. The interconnected carbon balls are found to have as ignificant impact on the electrochemical performance of the FeS-based electrodes. We highlight the catalytic activity of Fe 3 C, which was formed as a beneficial byproduct during the conducted synthesis. Owing to the unique formulation of the composite, the electrochemical cycling performance is significantly enhanced. This is accompanied by acontinuous increase in capacity.T ounderstand the different lithium storagem echanisms and evaluate the effect of interconnected carbonb alls on FeS-based electrodes, some techniques such as CV,e xsitu XRD, and SEM were applied. We discovered that the introductiono fi nterconnected carbon balls in FeS drastically affects the phase fraction, morphology, and particle size. More importantly,t he interconnected carbon balls have ap rofound influence on the kinetic process and crystal structure during cycling. Furthermore, such carbon balls change the diffusion-controlledb ehaviort oa pseudo-capacitive energy storage process. Indeed, the interconnected carbon balls improvet he electron conductivity, reduce the crystal size, and maintain the structuralintegrity.Especiallyf or long cycling procedures, the well-distributed FeS nanoparticles with small average diameters provide sufficient electrode-electrolyte contact areas for high lithium ion flux across the interface. Ar eduction of the lithium iond iffusion length duringc ycling significantly promotes the electrochemical processes, especiallyath igh specific current.
Synthesis of FeS/Fe 3 C/C compositematerial d-Glucose (630 mg, C 6 H 12 O 6 ,S igma-Aldrich, 99.5 %) was dissolved in DI water (40 mL). The above harvested FeS nanosheets (75 mg) were dispersed into the d-glucose solution. Subsequently,t he solution was transferred into a5 0mLT eflon-lined autoclave and heated at 180 8Cf or 12 h. After cooling down to room temperature, the product was poured out and washed by using rinse-precipitation cycles with DI water.T he product was dried at 80 8C. Finally,t he harvested black powder was kept in Ar/H 2 (Ar/H 2 = 95:5) at 600 8Cf or 5hwith aheating rate of 10 8Cmin À1 .

Materials characterization
The crystallographic information and phase fraction of the FeS and FeS/Fe 3 C/C were obtained from aS TOE STADI PC OMBI diffractometer (Mo Ka1 , l = 0.709300 )i nD ebye-Scherrer geometry.T he morphology and composition of the sample were observed by using a thermal field-emission scanning electron microscope (FESEM, Carl Zeiss SMT AG) equipped with energy-dispersive spectroscopy (EDS, Quantax 400 SDD, Bruker). Raman spectra were collected with a LabRam Evolution HR in ViaR aman spectrometer with al aser source (l = 532 nm, 10 mW,H ORIBA Jobin Yvon). The carbon percentage in the composite was analyzed by using organic elemental analysis (OEA, Vario Micro Cube, Elementar).

Electrochemical characterization
Electrochemical measurements were conducted by using threeelectrode Swagelok-type half cells assembled in an argon-filled glovebox (MBraun, O 2 and H 2 O 2ppm). The working electrodes were prepared by mixing the active material, carbon black (super PL i, Timcal Ltd.), and polyvinylidene fluoride binder (PVDF) with am ass ratio of 7:2:1i nN-methyl-2-pyrrolidone solvent (NMP) to form a homogeneous slurry.T he slurry was stirred overnight at room temperature and then coated on ac opper foil and dried at 80 8C. The circular working electrodes with ad iameter of 12 mm were punched out and dried under vacuum at 120 8Cf or 24 h. The mass loading of active materials was 1.8 mg cm À2 for FeS and 0.7 mg cm À2 for FeS/Fe 3 C/C. The FeS/Fe 3 C/C electrodes are normalized to the entire mass of the composite (FeS/Fe 3 C/C) as active material. Al ithium metal foil was used as the counter electrode and another as the reference electrode. The working and counter electrodes were separated by two 12 mm diameter vacuum-dried glass-fiber discs (Whatman GF/D). 1 m LiPF 6 dissolved in ethylene carbonate/dimethyl carbonate (LP30, EC/DMC = 1:1i nv olume, BASF,G ermany) was used as the electrolyte. Cyclic voltammetry (CV), galvanostatic cycling with potential limitation (GCPL), and electrochemical impedance spectroscopy (EIS) were conducted by using am ultichannel potentiostat (VMP3, Bio-Logic) in the voltage window range 0.01-3.0 Vv s. Li + /Li. All electrochemical measurements were performed in ac limate chamber (Binder) at 25 8C. CV was recorded at different scan rates in the range between 0.05 and 10 mV s À1 .G CPL was conducted at different specific currents be- ChemSusChem 2020, 13,986 -995 www.chemsuschem.org 2020 The Authors. Publishedb yWiley-VCH Verlag GmbH &Co. KGaA, Weinheim tween 0.1-5 Ag À1 .L ong-term cycling was performed at 1Ag À1 for 500 cycles. Ex situ X-ray radiation diffraction (XRD) and ex situ SEM measurements were performed on cycled electrodes (i.e.,c ycled at 1Ag À1 ,a fter the 9th, 140th, and 500th cycles) in the delithiation state. The cycled electrodes were disassembled and washed with dimethyl carbonate (DMC, Sigma-Aldrich, 99 %) in an argon-filled glovebox. The ex situ XRD was measured with aS TOE STADI Pd iffractometer (Cu Ka1 , l = 1.5406 )i nf lat-sample transmission mode. EIS experiments with aw orking electrode of 7mmd iameter were conducted at various selected potentials in the frequency range between 10 mHz and 500 kHz every 50 cycles. The cells were equilibrated at the desired potential for 3h before recording the EIS data. The impedance spectra were analyzed by using Relaxis 3software (rhd Instruments, Germany).
The EDS elemental maps of the pristine FeS material and FeS/Fe 3 C/ Cm aterial, Raman, and OEA of the pristine FeS material and FeS/ Fe 3 C/C material, the Nyquist plots of FeS and FeS/Fe 3 C/C electrodes at some selected cycles (1st, 100th, and 200th) in lithiation and de-lithiation conditions are given in the Supporting Information.