Gl

AI

a

ARRAA

KCHAL

1

iehgtvbbtanatAbCrv

h0

Electrochimica Acta 146 (2014) 769–775

Contents lists available at ScienceDirect

Electrochimica Acta

j ourna l ho me page: www.elsev ier .com/ locate /e lec tac ta

raphitized stacked-cup carbon nanofibers as anode materials forithium-ion batteries

lberto Ramos ∗, Ignacio Cameán, Nuria Cuesta, Ana B. Garcíanstituto Nacional del Carbón, CSIC, Francisco Pintado Fe 26, 33011-Oviedo, Spain

r t i c l e i n f o

rticle history:eceived 28 July 2014eceived in revised form 26 August 2014ccepted 1 September 2014vailable online 21 September 2014

eywords:arbon nanofibersigh-temperature heat treatment

a b s t r a c t

The electrochemical performance as anodes in lithium-ion batteries (LIBs) of graphitized stacked-cup car-bon nanofibers (CNFs) with different degree of structural order is investigated by galvanostatic cycling.To this end, two types of commercial CNFs (PR19 and PR24) bearing this microstructure were heat treatedin the range of 1800-2800 ◦C. The discharge capacity provided at the end of cycling, the capacity retentionalong cycling (cyclability) and the cycling efficiency (charge capacity/discharge capacity) of these nano-materials improve significantly after heat treatment. Specifically, a progressive increase of the dischargecapacity with the growth of the crystallites sizes (Lc, La) and the decrease of the structural defects (ID/It)is observed. All heat-treated CNFs exhibit cyclabilities > 80% after 50 cycles and efficiencies > 90%. Par-

nodeithium-ion batteries

ticularly outstanding is the performance of PR19-2800 which provides a discharge capacity of 295 mAhg−1 with capacity retention up to 88%. The CNFs heat treated at 2800 ◦C also show good rate capabilities,keeping reversible capacities > 200 mAh g−1 at a 2C (744 mA g−1) rate, and > 100 mAh g−1 after prolongedcycling (500 cycles) at a 1C rate, clearly outperforming synthetic graphite having micrometric particlesize and much larger development of the crystalline structure.

© 2014 Elsevier Ltd. All rights reserved.

. Introduction

Since their introduction in the market by Sony in 1991, lithium-on batteries (LIBs) have become one of the most important mobilenergy storage systems in the world. Nowadays, the demand forigh-performance LIBs is rapidly increasing in line with the fastrowth of the market of portable electronic devices (mobile phones,ablets, laptop computers, etc.) and the development of the electricehicle. To satisfy the needs of this market, intensive research haseen carried out to improve the overall performance of the LIBs,y increasing both the energy density and power, and extendinghe lifetime, which are largely governed by the electrode materi-ls. In this respect, the development of new materials, particularlyanomaterials, to replace graphite and LiCoO2, traditionally useds the active materials in anodes and cathodes of LIBs, respec-ively, is receiving significant attention by the researchers [1–3].s regards the anode material, carbon nanofibers (CNFs) appear toe a promising candidate [4–20]. In general, the main advantage of

NFs over graphite or other micrometer-sized carbons is their highate capabilities, providing high reversible capacities even at ele-ated electric current densities due to their nanometric size which

∗ Corresponding author. Tel.: +(+34)985119004.E-mail address: alberto.ramos@incar.csic.es (A. Ramos).

ttp://dx.doi.org/10.1016/j.electacta.2014.09.035013-4686/© 2014 Elsevier Ltd. All rights reserved.

reduces the diffusion length of the lithium ions in the electrode,thus allowing faster insertion/de-insertion processes, i.e. charge-discharge rate [8,10,15,19,20].

Graphitization thermal treatment is a process consisting in theheat treatment of carbon materials at high temperature (HTT)under inert atmosphere to achieve their transformation intographite (graphitization) or graphitized materials (graphitizationheat treatment). The HTT has been widely applied to CNFs as well,generally leading to materials with higher degree of graphitic order,less structural defects, lower surface area and larger thermal andelectrical conductivities [21]. This fine-tuning of the structure andproperties of the CNFs often goes in line with an improvement oftheir electrochemical performance as anodes on LIBs [21]. In thisregard, methane-based graphitized CNFs with a high degree of crys-tallinity, comparable to commercially available micrometer-sizedgraphite, prepared by HTT in our research group were investigatedfor this application [18,22,23]. These materials showed excellentcapacity retention along cycling and cycle efficiency (> 99%), aswell as high discharge capacity (up to 310-320 mAh g−1 after 50cycles), which made them comparable or even superior to com-mercial graphite [18].

Taking these precedent results into account, the electrochemicalperformance as potential anodes in LIBs of graphitized stacked-cupCNFs is herein investigated by galvanostatic cycling at differentelectric current densities. To this end, two different commercial

7 imica

Cceag

msoc[a[btb

2

2

2(Asio1wa

2n

2fltftnP

2

tdbweffbcofcfswoaiw

70 A. Ramos et al. / Electroch

NFs bearing this microstructure with an outer sheath of pyrolyticarbon were heat treated in the range of 1800-2800 ◦C. The influ-nce of the structural (graphitic order as evaluated from XRDnd Raman parameters) and textural (porosity) properties of theraphitized CNFs on their anodic behavior is studied.

Among the different types of as-produced CNFs, the aboveentioned stacked-cup, which are commercially available at a rea-

onable price (< 0.40 $g−1) from Applied Sciences, Inc (ASI) arene of the most studied. Specifically, the structural and texturalhanges showed by these as-produced CNFs during graphitization24–31], as well as some potential applications, including tests asnodes for LIBs, have been previously reported by other authors4,7,12,19,32–34]. However, to the best of our knowledge, a com-ined study correlating the degree of graphitization after HTT withheir electrochemical performance as anodes in LIBs has not yeteen carried out.

. Experimental

.1. Materials

The as-produced CNFs, namely PR-19-XT-PS (PR19) and PR-4-XT-PS (PR24), were obtained from Pyrograf Products Inc.http://pyrografproducts.com), an affiliate of Applied Sciences, Inc.ccording to the supplier, both types of as-produced CNFs have atacked-cup microstructure with a hollow core and an outer chem-cally vapor deposited (CVD) turbostratic layer, which is minimalr negligible for PR24 (average diameters: ∼ 150 nm for PR19 vs ∼00 nm for PR24). For comparison purposes, synthetic graphite (SG)ith micrometric particle size that is currently commercialized for

node in LIBs was also selected.

.2. High-temperature treatment of the as-produced carbonanofibers

The graphitization experiments were carried out at 1800, 2200,500 and 2800 ◦C in a graphite electrical furnace for 1 h under argonow. The heating rates were 25 ◦C min−1 from room temperatureo 1000 ◦C, 20 ◦C min−1 in the range 1000-2000 ◦C and 10 ◦C min−1

rom 2000 ◦C to the target temperature. The graphitized materialshus prepared were identified by the as-produced CNFs desig-ation, followed by the treatment temperature, e.g. PR19-1800,R24-2200, etc.

.3. Characterization techniques

The diffractograms were recorded in an X-ray powder diffrac-ometer as described elsewhere [35]. For each material, threeiffractograms were obtained, using a different representativeatch of sample for each run. The mean interlayer spacing, d002,as evaluated from the position of the (002) peak applying Bragg’s

quation. The mean crystallite sizes, Lc and La, were calculatedrom the (002) and (110) peaks, respectively, using the Scherrerormula, with values of K = 0.9 for Lc and 1.84 for La [36]. Theroadening of diffraction peaks due to instrumental factors wasorrected with the use of a silicon standard. Typical standard errorsf the XRD parameters are < 4% and < 9% of the reported valuesor Lc and La, respectively; the d002 values are much more pre-ise, with standard errors of < 0.03%. La could not be calculatedor the as-produced CNFs, PR19 and PR24, due to the low inten-ity and signal-to-noise ratio of the (110) peaks. Raman spectraere obtained in a Raman microspectrometer as described previ-

usly [37]. The intensity I of the Raman bands was measured using mixed Gaussian–Lorentzian curve-fitting procedure. The relativentensity of the Raman D-band ID/It (It = ID + IG + ID′ ) was calculated

ith standard errors lower than 5%. The commercial SG was also

Acta 146 (2014) 769–775

analyzed for comparative purposes. These XRD and Raman parame-ters are used in this study to evaluate the degree of structural orderof the materials [38].

The textural properties of the materials were determined by N2adsorption-desorption at −196 ◦C in a volumetric adsorption sys-tem. Before measurements, the samples were degassed overnightat 120 ◦C. The specific surface areas, SBET, were calculated by apply-ing the Brunauer-Emmett-Teller (BET) method, taking 16.2 nm2 forthe cross-sectional area of the nitrogen-adsorbed molecule.

2.4. Cell preparation and electrochemical measurements

The electrochemical measurements were carried out in two-electrode Swagelok-type laboratory cells by galvanostatic cycling.Metallic lithium discs of 12 mm of diameter were used as counter-electrodes. To prepare the working electrodes, the active material(92 wt.%) and the polyvinylidene difluoride (PVDF) binder (8 wt.%)were mixed and stirred vigorously in the minimum amount of 1-methyl-2-pyrrolidone solution (NMP). The slurry was then driedunder vacuum at 120 ◦C for 24 h and the solid thus obtained wasground in an agate mortar. Afterwards, an small amount of thiscomposite (active material + binder) was deposited onto a copperdisc of 12 mm of diameter and a thickness of 25 �m (compositeloadings of 1.77 to 3.54 mg cm−2 for CNFs-based electrodes, and3.54 to 8.84 mg cm−2 for SG-based electrodes) and a few dropsof NMP were added. The slurry thus formed was homogeneouslyspread over the Cu disc with a metal bar and dried under vacuumat 120 ◦C for 2 h. Finally, the electrode was hydraulically pressedat a pressure ca. 90 kPa and the cell assembly was carried out in adry box under argon atmosphere with oxygen and water contentsbelow 0.1 ppm. Two glass micro-fiber discs, impregnated with a fewdrops of a solution of the electrolyte (1 M LiPF6 in EC:DEC, 1:1, w/w),were used as separators between the counter (metal lithium) andworking (CNFs-based or graphite-based) electrodes. The galvano-static cycling was carried out in the 2.1–0.003 V potential range atC/10 (C = 372 mA g−1) using a potentiostat/galvanostat. Moreover,the study of the rate capability of the PR19-2800 CNFs and SG wasalso conducted, cycling in the same voltage range at C/10, C/5, C/2,C, 2C and back to C/10 during 10 cycles for each one, as well as aprolonged cycling (500 cycles) at a 1C rate.

3. Results and discussion

3.1. Anodic performance of as-produced and heat-treated CNFs ata C/10 rate

Initially, the as-produced and heat-treated CNFs were galvano-statically cycled (50 cycles) at a low electric current density (C/10rate). All the relevant electrochemical parameters thus obtainedare registered in Table 1. From a general perspective of this data,it can be concluded that the HTT of the CNFs leads to a remarkableimprovement of the capacity retention along cycling at even thelowest temperature of 1800 ◦C (49% for PR19 vs 87% for PR19-1800).Furthermore, the discharge capacity provided by these materialsafter 50 cycles tends to increase as the treatment temperatureincreases (175 mAh g−1 for PR19-1800 vs 295 mAh g−1 for PR19-2800) reaching values larger than that delivered by SG graphite(197 mAh g−1). Surprisingly, the heat treatment of the as-producedCNFs at 1800 ◦C results in a slight decrease of the discharge capacity(193 mAh g−1 for PR19 vs 175 mAh g−1 for PR19-1800). No signif-icant variations of the irreversible capacity (%) are observed after

the HTT of the as-produced CNFs. In any event, these values areunusually large as compared to other graphitic materials, such asfor example the SG graphite. These results will be further discussedin detail by studying the influence of the structural and textural

A. Ramos et al. / Electrochimica Acta 146 (2014) 769–775 771

Table 1Electrochemical parameters from the galvanostatic cycling at C/10 rate of as-produced (PR19, PR24) and heat-treated CNFs, and SG graphite.

Material Discharge capacity(1st cycle) (mAh g−1)

Discharge capacity(2nd cycle) (mAh g−1)

Discharge capacity(50th cycle) (mAh g−1)

Irreversible capacity(1st cycle) (mAh g−1) [%]

Capacityretention (%)

PR19 656 393 193 283 [43] 49PR19-1800 333 200 175 156 [47] 87PR19-2200 368 251 218 170 [46] 87PR19-2500 427 266 241 162 [45] 91PR19-2800 490 334 295 187 [38] 88PR24 586 360 206 267 [46] 57PR24-1800 378 238 192 169 [45] 81PR24-2200 442 244 215 229 [52] 88

249 346 [60] 81270 271 [54] 88197 106 [25] 57

pa

tvatFs(cadwccsdtCttvaacs

ltwta3mftfipwditioatPh2t

PR24-2500 581 306

PR24-2800 500 308

SG 427 345

roperties of the CNFs on the electrochemical performance asnode in LIBs.

Focusing on the lithium intercalation into these materials,wo different mechanisms can be distinguished by comparing theoltage-capacity plots of the first discharge-charge cycle of thes-produced (PR19 and PR24) and some of the corresponding heat-reated (PR19-2200,2500,2800 and PR24-2200,2500,2800) CNFs inig. 1. PR19 and PR24 show a progressive voltage drop with con-tant slope below 1.2 V (vs Li/Li+) and apparently negligible plateauxFig. 1a,b) as can be confirmed by the analysis of the differentialapacity vs potential plots of the first discharge of these materi-ls (Fig. 2a,b). In these plots, a more pronounced increase of theifferential capacity below 0.2 V can be observed; however, peakshich are attributed to different Li+ intercalation stages in graphitic

arbons are not present. This type of behavior is characteristic ofarbons presenting low degree of structural order or turbostratictructure [39,40], as supportd by the XRD and Raman spectroscopyata in Table 2 since values of ∼ 0.342 nm and ∼ 7 nm were, respec-ively, calculated for the d002 and Lc parameters of the as-producedNFs (PR19 and PR24), whereas the relative integrated intensity ofhe D Raman band, ID/It (owing to structural defects, especially athe surface level) is > 60%. These values are similar to those pre-iously reported by Endo et al. for related stacked-cup CNFs [29],nd are fully consistent with the microstructure of PR19 [25,27]nd PR24 [25] CNFs, both displaying an outer sheath of turbostraticarbon formed by pyrolytic deposition and an inner hollow coreurrounded by nested (stacked) cones.

The heat-treated (≥ 2200 ◦C) CNFs clearly exhibit differentithium intercalation plateaux in the discharge profiles, apart fromhat at 0.8 V due to the SEI formation [41–43], which is consistentith the graphitic structure of these materials (Fig. 1). However,

he first discharge capacity of these heat-treated CNFs decreasess compared with the as-produced ones (Table 1). For example,33 mAh g−1 were calculated for PR19-1800 which is approxi-ately 49% of that for PR19. In any event, the extent of this capacity

all is much less remarkable by increasing the treatment tempera-ure of the CNFs (∼ 25% in the case of PR19-2800). In general, therst discharge capacity of the heat-treated CNFs seems to increasearallel to the temperature, with the only anomaly of PR24-2500,hich provides a higher value than PR24-2800. Going back to theischarge-charge plots for the first cycle of the heat-treated CNFs

n Fig. 1, in addition to the capacity associated with the SEI forma-ion (as seen by the plateau at 0.8 V in the discharge profile), anrreversible process in the potential range of 2.5-1.5 V can also bebserved in the profiles of PR19-2500 and PR24-2500 (Fig. 1a,b),nd even in that of the PR24-2800 with larger degree of struc-ural order (Fig. 1c). This process is of particular significance for

R24-2500, accounting for ∼ 75 mAh g−1, which could explain theigher first discharge capacity of this material with regard to PR24-800 that has been mentioned above (Table 1). Similar processes inhe first discharge of other CNFs [13] and carbon nanotubes[44,45]

Fig. 1. Profiles of the first discharge-charge cycle at C/10 rate of (a) PR19, PR19-2200and PR19-2500, (b) PR24, PR24-2200 and PR24-2500, and (c) PR19-2800, PR24-2800and SG (inset: zoom on the area below 0.3 V).

772 A. Ramos et al. / Electrochimica Acta 146 (2014) 769–775

Table 2Interlayer spacing (d002), crystallite sizes (Lc and La), Raman ratio (ID/It) and surface area (SBET) of as-produced (PR19, PR24) and heat-treated CNFs, and SG.

Material d002 (nm) Lc (nm) La (nm) ID/It (%) SBET (m2g−1)

PR19 0.3426 6.8 - 67.8 28PR19-1800 0.3435 7.0 12.5 38.4 12PR19-2200 0.3414 13.5 27.3 14.0 15PR19-2500 0.3391 16.2 38.8 12.9 14PR19-2800 0.3385 17.4 44.6 11.4 15PR24 0.3421 7.5 - 63.8 36PR24-1800 0.3432 7.4 14.6 37.9 32PR24-2200 0.3412 11.9 27.5 18.2 30

hrwhpt

FiP

PR24-2500 0.3394 13.8PR24-2800 0.3389 15.1

SG 0.3358 46.4

ave been documented in the literature, and attributed to theeduction of molecular oxygen or oxygenated functional groups

ith concomitant formation of Li2O. In this work, since the CNFsave been previously heat-treated at temperatures high enough toractically eliminate all the oxygen-containing functional groups,he first explanation (oxygen reduction) appears to be the most

ig. 2. Differential capacity vs potential plots for the first discharge of the CNFsn: (a) PR19, PR19-2200 and PR19-2500, (b) PR24, PR24-2200 and PR24-2500, (c)R19-2800, PR24-2800 and SG.

35.3 14.0 2941.1 11.6 2966.2 27.2 6

plausible one. After the plateau at 0.8 V, the potential decreasesgradually again to 0.2 V. Below this potential, different plateaux canbe observed (up to three, see inset in Fig. 1c for PR19-2800 andPR24-2800), accounting for the transitions between the differentstages of Li+ intercalation in graphite [46,47]. These intercalationstages are better appreciated in the differential capacity vs potentialplots for the first discharge (Fig. 2).

Thus, some of the heat-treated CNFs (2200-2500 ◦C) show twopeaks in the region below 0.2 V, at ∼ 0.07 and 0.09 V (Fig. 2a,b). Uponincreasing the treatment temperature up to 2800 ◦C, the intensity ofthe peaks increases and even a small third weak peak at ∼ 0.17 V canbe distinguished (Fig. 2c). However, the latter peak at the highestintercalation voltage is less intense than that of the SG of refer-ence (Fig. 1c, Fig. 2c), which also exhibits an additional peak at0.13 V (coexistence of the lithium intercalation stages 3 and 2L).This peak is typical of highly ordered graphite [46], therefore it wasnot observed in the profiles of the heat-treated CNFs with muchlower degree of crystallinity as shown by the values of the XRD andRaman parameters (Table 2).

Focusing now on the irreversible capacity in the first discharge-charge cycle (Cirr), in general terms, the heat-treated CNFs showmuch higher values than SG (Table 1), with lower BET surface area(Table 2). This is not surprising since the Cirr in carbon materialshas been reported to increase with this textural parameter [41,42].Moreover, as discussed above, the heat-treated PR24 CNFs, partic-ularly PR24-2500, exhibited an additional process in the potentialrange of 2.5-1.5 V during the first cycle (Fig. 1) which accounts forthe particularly high Cirr for this material (Table 1).

The extended galvanostatic plots (discharge capacity vs cyclenumber) of the CNFs are presented in Fig. 3. Firstly, it shouldbe remarked that the cycling performance of these nanomate-rials improves significantly after the heat treatment in terms ofdischarge capacity provided at the end of the cycling, capacityretention along cycling (cyclability) and cycling efficiency (chargecapacity/discharge capacity). The heat-treated CNFs show good-to-excellent cyclabilities with capacity keeping values > 80% after 50cycles and cycling efficiencies of > 90% after the 1st discharge/chargecycle and, in most cases > 95% after the first 5 to 10 cycles, whenSEI formation is thought to be completed. Particularly outstand-ing is the performance of PR19-2800, which provides an averagedischarge capacity of 295 mAh g−1 with capacity retention up to88%.

The as-produced PR19 and PR24, despite exhibiting highdischarge capacity in the 2nd cycle (393 and 360 mAh g−1, respec-tively), it decreases continuously (almost in a linear fashion) alongcycling to values of 193 and 206 mAh g−1 in the 50th cycle, leading tocapacity retentions of only 49 and 57%, respectively (see Table 1 andFig. 3a,b). This type of behavior is typical of disordered carbons and

can be found in hard and soft carbons heat-treated about or below1000 ◦C [48]. The discharge capacity in the 2nd cycle of the materialsis plotted versus treatment temperature (for the as-produced CNFsa production temperature of 1100 ◦C was considered) in Fig. 4. As

A. Ramos et al. / Electrochimica

F1

c(b(a(Dif2bvoIcb

FP

ig. 3. Extended galvanostatic cycling at C/10 rate (37.2 mA g−1) for (a) PR19, PR19-800/2800 and (b) PR24 and PR24-1800/2800.

ompared with the as-produced CNFs, a significant fall at 1800 ◦C200 and 237 mAh g−1 for PR19-1800 and PR24-1800), followedy a progressive increase up to the final temperature of 2800 ◦C334 and 308 mAh g−1 for PR19-2800 and PR24-2800) can beppreciated. These data fit well to a third-order polynomial curveR2 = 0.926), which is reminiscent of that previously reported byahn and coworkers in a study about the dependence of the lithium

ntercalation on the crystal structure of soft carbons [39,40]. Theyound that the highest reversible capacities (> 372 mAh g−1) in thend cycle were reached by disordered carbons that had been heatedelow 900 ◦C. On increasing the temperature, a minimum capacityalue first appeared at ∼ 1500 ◦C and afterwards, the capacity wasbserved to increase progressively up to ∼ 300 mAh g−1 at 2800 ◦C.

n line with this, Endo et al. also reported a relation between therystallite thickness (Lc) and the reversible capacity of the car-on materials [49,50], concluding that very disordered carbons

ig. 4. Discharge capacity in the 2nd cycle (C/10 rate) vs treatment temperature forR19 and PR24 CNFs.

Acta 146 (2014) 769–775 773

(Lc < 3 nm), together with well-ordered ones (Lc > 20 nm), providedthe highest capacities, whereas minimum values were attainedfor carbon materials with Lc values near 10 nm. In this work, thistendency is not entirely applicable since, as mentioned, the capac-ities in the 2nd cycle measured for the as-produced CNFs (PR19,PR24) and the corresponding heat-treated at 1800 ◦C (PR19-1800,PR24-1800) having similar Lc values (∼ 7 nm, see Table 2) differsignificantly. In addition, according to the results of Endo et al., amuch lower capacity for the as-produced CNFs should be expected.However, it must be borne in mind that the dual nature of PR19and PR24 CNFs, with an outer layer of pyrolytic disordered carbon,might influence on their electrochemical behavior.

Significant differences, however, are observed in the cyclingbehavior of the heat-treated CNFs, specifically as regards thereversible capacity. By comparing the electrochemical data (Fig. 2,Table 1) and the XRD and Raman parameters of these materials(Table 2), it is evident that those with higher degree of graphiticstructural order deliver larger lithium storage capacity at the endof cycling. For example, final capacity values of 218 and 295 mAhg−1 were calculated for PR19-2200 and PR19-2800 with crystallitesizes, Lc, of ∼ 14 nm and ∼ 18 nm, respectively; and the corre-sponding interplanar distances, d002, are 0.3414 nm and 0.3385 nm.The variation with the relative intensity of the D Raman band fol-lows a similar tendency. It is well known that the electrochemicalintercalation of lithium in ordered (graphite-like) carbon materialsdepends on their crystal structure [23,35,49–51]. Specifically, thecrystal thickness, Lc, was reported to be the most important factoraffecting the extent of the reversible capacity in graphitic materi-als providing similar morphology and particle size. As stated in theliterature [24,25,27,29], the heat treatment of stacked-cup CNFs athigh temperatures (> 1500 ◦C) (i) reduces significantly the thick-ness of the pyrolytic layer until it becomes practically negligible at∼ 3000 ◦C, (ii) increases at the same time the graphitic order, of bothinner (stacked-cups) and outer (multi-wall nanotubes) layers and(iii) creates loops and multi-loops joining adjacent graphene edgesof the stacked cones from the inner and outer walls. These multi-loops are particularly relevant in this kind of low-diameter carbonsand cause structural strains that impede reaching high degrees ofgraphitization even for materials heat-treated at 2800 ◦C, namelyPR19-2800 and PR24-2800 (Table 2). Nevertheless, this does notseem to hamper Li intercalation/de-intercalation despite their rel-atively low Lc (17.4 and 15.1 nm for PR19-2800 and PR24-2800,respectively), both delivering capacities of ∼ 300 mAh g−1 after 50cycles.

3.2. Anodic rate performance of the heat-treated PR19-2800 CNFs

The cycling behavior of PR19-2800 CNFs at increasing electriccurrent densities (C/10 to 2C rates) is shown in Fig. 5. Among theheat-treated CNFs, this material was selected because of its excel-lent electrochemical performance during cycling at the lowest rateof C/10 (Fig. 3). For comparison, that of SG graphite is also included.

As can be seen, the increase of the cycling rate has little-to-moderate repercussion on the discharge capacity delivered byPR19-2800 with capacity losses ranging from only 39 mAh g−1

(C/10 to 1C) to 59 mAh g−1 (1C to 2C). Even at an elevated rateof 2C, this material still provides a reversible capacity slighly over200 mAh g−1. This is an essential property for consumer electronicapplications in which at least a 1C rate and reasonable energy den-sity are usually required. Furthemore, the initial discharge capacityat C/10 rate is basically recovered after returning to this cyclingrate from the 2C one. In contrast, SG graphite shows higher sensi-

tiveness to the increase of the rate, delivering similar dischargecapacities to those of PR19-2800 at C/10 (∼300 mAh g−1), butwith capacity losses of ∼200 mAh g−1 (C/10 to 1C) to 35 mAhg−1 (1C to 2C). Consequently, the discharge capacity of SG falls

774 A. Ramos et al. / Electrochimica

Fig. 5. Cycling performance of PR19-2800 CNFs and SG at C/10, C/5, C/2, 1C, 2C andC/10 rates.

Table 3Capacity retentions in % of PR19-2800 and SG at different rates.

Material C/10a C/5 C/2 1C 2C C/10

PR19-2800 95 99 100 100 97 100

btabt2liwtpe

PPcii4f(g

Fa

SG 93 97 92 66 74 86

a Calculated from the 2nd cycle to the 10th.

elow 100 mAh g−1 and 30 mAh g−1 at 1C and 2C rates, respec-ively. Moreover, PR19-2800 exhibits a remakably stable capacitylong cycling in the whole range of rates studied, even after cyclingack to the initial C/10 (93% against a value of 72% calculated forhe SG). For comparison, the capacity retention values of PR19-800 and SG are reported in Table 3. Since PR19-2800 has a much

ower development of the crystalline structure than SG (Table 2),t can be assumed that the nanometric particle size of these CNFs,

hich reduces the diffusion length of the lithium ions in the elec-rode and consequently, may promote faster insertion/de-insertionrocesses, i.e. charge-discharge rate, is responsible of their betterlectrochemical performance.

Finally, prolonged galvanostatic cycling, up to 500 cycles, ofR19-2800 and SG was carried out at 1C rate (Fig. 6). As expected,R19-2800 displays larger discharge capacity than SG all alongycling. Thus, after 500 cycles, the discharge capacity of PR19-2800s 112 mAh g−1 vs 33 mAh g−1 of the SG. In terms of overall capac-ty retention, both materials are comparable (49% for PR19-2800 vs5% for SG). However, as regards the discharge capacity, three dif-erent ranges are observed in the prolonged cycling of PR19-2800:

i) 2-60 cycles, it drecreases quickly (from 230 mAh g−1 to 176 mAh−1), (ii) 60-300 cycles, it remains almost constant (average value

ig. 6. Prolonged galvanostatic cycling (500 cycles) at 1C rate for PR19-2800 CNFsnd SG.

[[

[[[[[

[

Acta 146 (2014) 769–775

of 171 ± 6 mAh g−1) and (iii) 300-500 cycles, it slightly decreasesto the final value of 112 mAh g−1.

4. Conclusions

The anodic performance of the stacked-cup PR19 and PR24 CNFsin terms of discharge capacity provided at the end of the cycling,capacity retention along cycling (cyclability) and cycling efficiency(charge capacity/discharge capacity) improves significantly afterHTT (1800-2800 ◦C) due to the increase of the degree of graphiticstructural order. In fact, a progressive increase of the dischargecapacity delivered by the heat-treated CNFs with the growth ofthe crystallites sizes (Lc, La) as well as the decrease of the struc-tural defects (ID/It) was observed. Remarkably outstanding are theresults obtained with the heat-treated CNFs at 2800 ◦C, particu-larly PR19-2800, showing a discharge capacity of ∼ 300 mAh g−1

and capacity retention of ∼ 88% after 50 discharge-charge cycles at aC/10 rate. As compared to the as-produced PR-19 CNFs, these valuesaccount for improvements of ∼ 53% and 80%, respectively. Further-more, the nanometric particle size of heat-treated CNFs reducethe diffusion length of the lithium ions in the electrode, allow-ing faster intercalation/de-intercalation processes, as denoted bythe experiments at higher current rates (up to 2C), in which theyclearly outperformed the micrometer sized SG graphite. Therefore,stacked-cup CNFs heat treated at 2800 ◦C could be potential candi-dates for anodes of LIBs, especially for high-power applications.

Acknowledgements

Financial support from the Spanish Ministry of Economy andCompetitiveness MINECO (under Projects ENE2008-06516 andENE2011-28318) is gratefully acknowledged. A. Ramos and N.Cuesta, respectively, thank the Spanish Research Council for Sci-entific Research (CSIC) for a JAE-Doc contract, co-funded by theEuropean Social Fund (ESF), and the Spanish Ministry of Economyand Competitiveness (MINECO) for a pre-doctoral grant (BES-2012-052711).

Appendix A. Supplementary data

Supplementary data associated with this article can be found,in the online version, at http://dx.doi.org/10.1016/j.electacta.2014.09.035.

References

[1] K. Ozawa, Lithium Ion Rechargeable Batteries: Materials, Technology, and NewApplications, Wiley-VCH, Weinheim, 2009.

[2] M. Yoshio, R.J. Brodd, A. Kozawa, Lithium-Ion Batteries: Science and Technolo-gies, Springer, New York, 2009.

[3] Y. Abu-Lebdeh, I. Davidson, Nanotechnology for Lithium-Ion Batteries,Springer, 2013.

[4] W. Lu, D.D.L. Chung, Carbon 39 (2001) 493–496.[5] Q. Wang, H. Li, L. Chen, X. Huang, D. Zhong, E. Wang, J. Electrochem. Soc. 150

(2003) A1281–A1286.[6] T. Doi, A. Fukuda, Y. Iriyama, T. Abe, Z. Ogumi, K. Nakagawa, T. Ando, Elec-

trochem. Commun. 7 (2005) 10–13.[7] G.F. Ortiz, R. Alcántara, P. Lavela, J.L. Tirado, J. Electrochem. Soc. 152 (2005)

A1797–A1803.[8] H. Habazaki, M. Kiriu, H. Konno, Electrochem. Commun. 8 (2006) 1275–1279.[9] C. Kim, K.S. Yang, M. Kojima, K. Yoshida, Y.J. Kim, Y.A. Kim, M. Endo, Adv. Funct.

Mater. 16 (2006) 2393–2397.10] V. Subramanian, H. Zhu, B. Wei, J. Phys. Chem. B 110 (2006) 7178–7183.11] G. Zou, D. Zhang, C. Dong, H. Li, K. Xiong, L. Fei, Y. Qian, Carbon 44 (2006)

828–832.12] D. Burton, P. Lake, G.G. Tibbetts, M.L. Lake, SAMPE J. 43 (2007) 36–40.13] D. Deng, J.Y. Lee, Chem. Mater. 19 (2007) 4198–4204.

14] L. Ji, X. Zhang, Nanotechnology 20 (2009) 155705 (7pp).15] C. Li, X. Yin, L. Chen, Q. Li, T. Wang, J. Phys. Chem. C 113 (2009) 13438–13442.16] L. Ji, Y. Yao, O. Toprakci, Z. Lin, Y. Liang, Q. Shi, A.J. Medford, C.R. Millns, X. Zhang,

J. Power Sources 195 (2010) 2050–2056.17] J. Zhou, H. Song, B. Fu, B. Wu, X. Chen, J. Mater. Chem. 20 (2010) 2794–2800.

imica

[

[[

[[

[

[

[[

[[

[

[

[

[

[

[

[[[[

[

[[[[

[

[[[

[

[

A. Ramos et al. / Electroch

18] I. Cameán, A.B. García, I. Suelves, J.L. Pinilla, M.J. Lázaro, R. Moliner, J. PowerSources 198 (2012) 303–307.

19] R. Lawrence, J. Gnanaraj, X. Geng, J. Liang, Sci. Adv. Mat 4 (2012) 877–880.20] P. Suresh Kumar, R. Sahay, V. Aravindan, J. Sundaramurthy, W.C. Ling, V. Thavasi,

S.G. Mhaisalkar, S. Madhavi, S. Ramakrishna, J. Phys. D Appl. Phys. 45 (2012)265302 (5pp).

21] A. Ramos, I. Cameán, A.B. García, Carbon 59 (2013) 2–32.22] I. Cameán, A.B. García, I. Suelves, J.L. Pinilla, M.J. Lázaro, R. Moliner, J.-N.

Rouzaud, Carbon 50 (2012) 5387–5394.23] A.B. Garcia, I. Cameán, J.L. Pinilla, I. Suelves, M.J. Lázaro, R. Moliner, Fuel 89

(2010) 2160–2162.24] J.I. Paredes, M. Burghard, A. Martínez-Alonso, J.M.D. Tascón, Appl. Phys. A Mater.

80 (2005) 675–682.25] J.G. Lawrence, L.M. Berhan, A. Nadarajah, J. Nanopart. Res. 10 (2008) 1155–1167.26] M. Yoon, J. Howe, G. Tibbetts, G. Eres, Z. Zhang, Phys. Rev. B 75 (2007) 165402

(6pp).27] J.Y. Howe, G.G. Tibbetts, C. Kwag, M.L. Lake, J. Mater. Res. 21 (2006) 2646–2652.28] M. Endo, Y.A. Kim, T. Hayashi, Y. Fukai, K. Oshida, M. Terrones, T. Yanagisawa,

S. Higaki, M.S. Dresselhaus, Appl. Phys. Lett. 80 (2002) 1267–1269.29] M. Endo, Y.A. Kim, T. Hayashi, T. Yanagisawa, H. Muramatsu, M. Ezaka, H. Ter-

rones, M. Terrones, M.S. Dresselhaus, Carbon 41 (2003) 1941–1947.30] Y.A. Kim, T. Hayashi, Y. Fukai, M. Endo, T. Yanagisawa, Mol. Cryst. Liq. Cryst. 387

(2002) [381]/157-[385]161.31] M. Endo, B.J. Lee, Y.A. Kim, Y.J. Kim, H. Muramatsu, T. Yanagisawa, T. Hayashi,

M. Terrones, M.S. Dresselhaus, New. J. Phys. 5 (2003) 121.1–121.9.32] K. Nishimura, Y.A. Kim, T. Matushita, T. Hayashi, M. Endo, M.S. Dresselhaus, J.

Mater. Res. 15 (2000) 1303–1313.33] K. Oyaizu, A. Hatemata, W. Choi, H. Nishide, J. Mater. Chem. 20 (2010)

5404–5410.

[

[

Acta 146 (2014) 769–775 775

34] M. Endo, Y.A. Kim, T. Hayashi, K. Nishimura, T. Matusita, K. Miyashita, M.S.Dresselhaus, Carbon 39 (2001) 1287–1297.

35] I. Cameán, A.B. Garcia, J. Power Sources 196 (2011) 4816–4820.36] J. Biscoe, B.E. Warren, J. Appl. Phys. 13 (1942) 364–371.37] I. Cameán, P. Lavela, J.L. Tirado, A.B. García, Fuel 89 (2010) 986–991.38] A. Cuesta, P. Dhamelincourt, J. Laureyns, A. Martinez-Alonso, J.M.D. Tascon,

Carbon 32 (1994) 1523–1532.39] J.R. Dahn, A.K. Sleigh, H. Shi, J.N. Reimers, Q. Zhong, B.M. Way, Electrochim. Acta

38 (1993) 1179–1191.40] J.R. Dahn, T. Zheng, Y. Liu, J.S. Xue, Science 270 (1995) 590–593.41] R. Fong, U. von Sacken, J.R. Dahn, J. Electrochem. Soc. 137 (1990) 2009–2013.42] M. Winter, P. Novák, A. Monnier, J. Electrochem. Soc. 145 (1998) 428–436.43] K. Guerin, A. Fevrier-Bouvier, S. Flandrois, M. Couzi, B. Simon, P. Biensan, J.

Electrochem. Soc. 146 (1999) 3660–3665.44] E. Frackowiak, S. Gautier, H. Gaucher, S. Bonnamy, F. Beguin, Carbon 37 (1999)

61–69.45] W.X. Chen, J.Y. Lee, Z. Liu, Carbon 41 (2003) 959–966.46] J.R. Dahn, Phys. Rev. B 44 (1991) 9170–9177.47] T.D. Tran, J.H. Feikert, X. Song, K. Kinoshita, J. Electrochem. Soc. 142 (1995)

3297–3302.48] Z. Ogumi, M. Inaba, in: W. van Schalkwijk, B. Scrosati (Eds.),

Advances in Lithium-Ion Batteries, Springer, New York, 2002,pp. 79–101.

49] M. Endo, C. Kim, K. Nishimura, T. Fujino, K. Miyashita, Carbon 38 (2000)

183–197.

50] M. Endo, Y. Nishimura, T. Takahashi, K. Takeuchi, M.S. Dresselhaus, J. Phys.Chem. Solids 57 (1996) 725–728.

51] J.C. Arrebola, A. Caballero, L. Hernán, J. Morales, J. Electrochem. Soc. 156 (2009)A986–A992.