nano-drilled multiwalled carbon nanotubes: characterizations and application for lib anode materials

7
Nano-drilled multiwalled carbon nanotubes: characterizations and application for LIB anode materials Haryo S. Oktaviano, a Koichi Yamada b and Keiko Waki * ab Received 17th July 2012, Accepted 8th October 2012 DOI: 10.1039/c2jm34684b Nano-drilling of multiwalled carbon nanotubes (MWCNTs) has been conducted by taking advantage of CoO x as the oxidation catalyst at a relatively low temperature of 250 C. An increase of the I D /I G ratio of Raman spectroscopy, together with TEM visualization, justifies the presence of hole defects on the sidewalls, in comparison to the pristine and purified MWCNTs. Holes drilled on the basal plane of MWCNTs could increase access into the inner core as observed by the N 2 adsorption isotherm. Drilled MWCNTs (DMWCNTs) show the largest quantity of functional groups of 5.4 10 3 mol g 1 , which suggests that our drilling method also introduces functional groups on the edges. We also tested the electrochemical performance of our MWCNT samples as the anode material for lithium ion batteries. After 20 cycles at a current density of 25 mA h g 1 , the specific discharge capacities of pristine, purified and DMWCNTs are 267, 421 and 625 mA h g 1 , respectively. DMWCNTs exhibit the largest quantity of Li extraction from the GIC and inner core as shown by the dQ/dV discharge profiles. This suggests that increased access into the MWCNT inner core is proven to have positive effects on Li storage properties. 1. Introduction Owing to their high surface area, combined with good conduc- tivity, carbon nanotubes (CNTs) are considered to be promising materials for energy storage. 1–6 The most prominent structural characteristic of CNTs is their one dimensional, nanoscale cylinder cavities enclosed by the graphitic sheet(s). The avail- ability of the inner core is determined by the access into CNTs, either through an opened end-cap or sidewall defects. 7 In terms of lithium ion batteries (LIBs), the Li storage sites for CNTs have long been considered to be a graphitic intercalation compound (GIC), inner tube and functional group related sites. 8–11 A comparison study of Li intercalation into open and closed single wall carbon nanotubes (SWCNTs) has been conducted by Shi- moda et al. 12 The large reversible capacity observed on open SWCNTs was attributed to an enhanced Li diffusion into the interior of carbon nanotubes via the open ends and sidewall defects. A similar consideration was also reported for short multiwalled carbon nanotubes (MWCNTs) and vertically aligned MWCNTs as Li anode materials. 10,11 Both authors related the enhancement of their Li storage properties, particu- larly to the open access into the inner hole of tubes through defects and edges formation. Such defects facilitate the Li insertion and extraction, providing improved storage character- istics. These results are in agreement with the previous MD simulation carried out for nanotube-like structure in graphite, revealing that the Li ion intrusion sites are to be open interstices and/or void defects. 13 Considering the important features of the CNT inner core for various functional applications, many techniques have been developed to open the holes on CNTs. Electron irradiation and chemical oxidation treatments, whether by heating under a certain gas atmosphere or treating in acids, have been intensively studied to open the access into the inner surface of CNTs. 14–20 However, such treatments have their own limitations. Although the electron irradiation may produce well-controlled defects on a specific location for surface nano-patterning application, this technique will be difficult in preparing well-distributed open holes on the bulk of CNTs. On the other hand, chemical oxidation treatments may provide a plausible approach to make well-distributed defects of CNTs, particularly due to the fact that all samples have a uniform contact to the oxidation medium, such as a gas atmosphere or liquid acid. 17–20 Unfortunately, it is difficult to form hole defects using such strong unlocalized oxidation treatments, and the tube might deteriorate, which would result in low electrical conductivity. It is important to note that the gas oxidation treatment should be carried out at high temperatures (above 500 C), while acid-assisted oxidation treatment is usually conducted by heating at a prolonged expo- sure time. To overcome such drawbacks, our group has reported the possibility of controlling the defect formation on the surface of MWCNTs by solid-state reaction utilizing CoO x as an a Department of Energy Sciences, Tokyo Institute of Technology, 4529 Nagatsuta-cho, Midori-ku, Yokohama-shi, Kanagawa 226-8502, Japan. E-mail: [email protected]; Fax: +81 459 24 5614; Tel: +81 459 24 5614 b Center for Low Carbon Society Strategy, Japan Science and Technology Agency, 7 Gobancho, Chiyoda-ku, Tokyo 102-0076, Japan This journal is ª The Royal Society of Chemistry 2012 J. Mater. Chem., 2012, 22, 25167–25173 | 25167 Dynamic Article Links C < Journal of Materials Chemistry Cite this: J. Mater. Chem., 2012, 22, 25167 www.rsc.org/materials PAPER Published on 08 October 2012. Downloaded by Duke University on 01/10/2013 09:41:43. View Article Online / Journal Homepage / Table of Contents for this issue

Upload: keiko

Post on 13-Dec-2016

230 views

Category:

Documents


5 download

TRANSCRIPT

Page 1: Nano-drilled multiwalled carbon nanotubes: characterizations and application for LIB anode materials

Dynamic Article LinksC<Journal ofMaterials Chemistry

Cite this: J. Mater. Chem., 2012, 22, 25167

www.rsc.org/materials PAPER

Publ

ishe

d on

08

Oct

ober

201

2. D

ownl

oade

d by

Duk

e U

nive

rsity

on

01/1

0/20

13 0

9:41

:43.

View Article Online / Journal Homepage / Table of Contents for this issue

Nano-drilled multiwalled carbon nanotubes: characterizations and applicationfor LIB anode materials

Haryo S. Oktaviano,a Koichi Yamadab and Keiko Waki*ab

Received 17th July 2012, Accepted 8th October 2012

DOI: 10.1039/c2jm34684b

Nano-drilling of multiwalled carbon nanotubes (MWCNTs) has been conducted by taking advantage

of CoOx as the oxidation catalyst at a relatively low temperature of 250 �C. An increase of the ID/IGratio of Raman spectroscopy, together with TEM visualization, justifies the presence of hole defects on

the sidewalls, in comparison to the pristine and purified MWCNTs. Holes drilled on the basal plane of

MWCNTs could increase access into the inner core as observed by the N2 adsorption isotherm. Drilled

MWCNTs (DMWCNTs) show the largest quantity of functional groups of 5.4 � 10�3 mol g�1, which

suggests that our drilling method also introduces functional groups on the edges. We also tested the

electrochemical performance of our MWCNT samples as the anode material for lithium ion batteries.

After 20 cycles at a current density of 25 mA h g�1, the specific discharge capacities of pristine, purified

and DMWCNTs are 267, 421 and 625 mA h g�1, respectively. DMWCNTs exhibit the largest quantity

of Li extraction from the GIC and inner core as shown by the dQ/dV discharge profiles. This suggests

that increased access into the MWCNT inner core is proven to have positive effects on Li storage

properties.

1. Introduction

Owing to their high surface area, combined with good conduc-

tivity, carbon nanotubes (CNTs) are considered to be promising

materials for energy storage.1–6 The most prominent structural

characteristic of CNTs is their one dimensional, nanoscale

cylinder cavities enclosed by the graphitic sheet(s). The avail-

ability of the inner core is determined by the access into CNTs,

either through an opened end-cap or sidewall defects.7 In terms

of lithium ion batteries (LIBs), the Li storage sites for CNTs have

long been considered to be a graphitic intercalation compound

(GIC), inner tube and functional group related sites.8–11 A

comparison study of Li intercalation into open and closed single

wall carbon nanotubes (SWCNTs) has been conducted by Shi-

moda et al.12 The large reversible capacity observed on open

SWCNTs was attributed to an enhanced Li diffusion into the

interior of carbon nanotubes via the open ends and sidewall

defects. A similar consideration was also reported for short

multiwalled carbon nanotubes (MWCNTs) and vertically

aligned MWCNTs as Li anode materials.10,11 Both authors

related the enhancement of their Li storage properties, particu-

larly to the open access into the inner hole of tubes through

defects and edges formation. Such defects facilitate the Li

aDepartment of Energy Sciences, Tokyo Institute of Technology, 4529Nagatsuta-cho, Midori-ku, Yokohama-shi, Kanagawa 226-8502, Japan.E-mail: [email protected]; Fax: +81 459 24 5614; Tel: +81 45924 5614bCenter for Low Carbon Society Strategy, Japan Science and TechnologyAgency, 7 Gobancho, Chiyoda-ku, Tokyo 102-0076, Japan

This journal is ª The Royal Society of Chemistry 2012

insertion and extraction, providing improved storage character-

istics. These results are in agreement with the previous MD

simulation carried out for nanotube-like structure in graphite,

revealing that the Li ion intrusion sites are to be open interstices

and/or void defects.13

Considering the important features of the CNT inner core for

various functional applications, many techniques have been

developed to open the holes on CNTs. Electron irradiation and

chemical oxidation treatments, whether by heating under a

certain gas atmosphere or treating in acids, have been intensively

studied to open the access into the inner surface of CNTs.14–20

However, such treatments have their own limitations. Although

the electron irradiation may produce well-controlled defects on a

specific location for surface nano-patterning application, this

technique will be difficult in preparing well-distributed open

holes on the bulk of CNTs. On the other hand, chemical

oxidation treatments may provide a plausible approach to make

well-distributed defects of CNTs, particularly due to the fact that

all samples have a uniform contact to the oxidation medium,

such as a gas atmosphere or liquid acid.17–20 Unfortunately, it is

difficult to form hole defects using such strong unlocalized

oxidation treatments, and the tube might deteriorate, which

would result in low electrical conductivity. It is important to note

that the gas oxidation treatment should be carried out at high

temperatures (above 500 �C), while acid-assisted oxidation

treatment is usually conducted by heating at a prolonged expo-

sure time. To overcome such drawbacks, our group has reported

the possibility of controlling the defect formation on the surface

of MWCNTs by solid-state reaction utilizing CoOx as an

J. Mater. Chem., 2012, 22, 25167–25173 | 25167

Page 2: Nano-drilled multiwalled carbon nanotubes: characterizations and application for LIB anode materials

Fig. 1 Schematic procedure for drilling MWCNTs.

Publ

ishe

d on

08

Oct

ober

201

2. D

ownl

oade

d by

Duk

e U

nive

rsity

on

01/1

0/20

13 0

9:41

:43.

View Article Online

oxidation catalyst at a relatively low temperature (250 �C),resulting in nano-drilled structures on the sidewalls of

MWCNTs.21

In this work, we further simplify the previous method21 by

excluding the pre-Ar treatment and directly proceeding to the

oxidation treatment in air. We attempt to characterize the nano-

drilled MWCNTs by TEM, Raman, BET and TDS to confirm

the openings of MWCNTs and to understand the defective

structures. We further conducted electrochemical studies on our

sample as the LIB anode material. From these results, we

correlate the presence of the defects with the enhancement of Li

storage capacities.

2. Experimental

Drilling multiwalled carbon nanotubes

Pristine multiwalled carbon nanotubes (MWCNTs) used for the

present study were received from Showa Denko KK, Japan

(VGCFX, diameter 15 nm, length approximately 3 mm). Prior to

the preparation of drilled structure MWCNTs (DMWCNTs),

pristine MWCNTs were purified by acid treatment as described

elsewhere.22 The purified MWCNTs were then impregnated with

10 wt% Co dispersed in ethanol. Subsequently, the as prepared

Co/purified MWCNTs underwent oxidation under air atmo-

sphere at 250 �C for 25 minutes. In order to obtain the

DMWCNTs, cobalt oxides were removed in a boiled concen-

trated nitric acid for 1 hour. Finally the sample was filtered,

washed and dried overnight.

Analysis

The morphologies of the MWCNTs were examined by FE-TEM

(JEOL, JEM-2010F) operated at 200 kV. Thermal analysis of

MWCNTswas conducted by TG-DTA (RIGAKU, Thermo-plus

EVO, TD-DTA/8120ST). The measurement was conducted by

heating in air (100 mL min�1) up to 800 �C at a rate 10 �Cmin�1.

Raman spectra were taken under ambient conditions using an

NRS-3000 Series (JASCO) with an Ar-ion laser beam at an exci-

tation wavelength of 514.5 nm. Nitrogen adsorption isotherms

were measured at 77 K on a BEL JAPAN, BELSORP-mini II

instrument. The samples were degassed at 120 �C under vacuum

for 3 h prior to the measurement. The specific surface area (SBET)

was calculated from the 0.05–0.3 P/P0 region of the adsorption

isotherm using the standard Brunauer–Emmett–Teller (BET)

method, while pore size distribution was derived by the Barrett–

Joyner–Halenda (BJH) method. The O-functionalities of the

MWCNTs were evaluated by thermal desorption spectroscopy

(TDS). TDS (ESCO, EMD-WA1000S) measurement was per-

formed by heating the sample up to 1100 �C at a rate of 30 �Cmin�1 under ultrahigh vacuum (base pressure: 2.0� 10�7 Pa) and

detecting the evolved gas using a quadrupole mass spectrometer

system (QMG422). The quantization of CO andCO2 gas evolved

from TDS measurement was conducted by calibrating the sensi-

tivity of the detector using a H-implanted Si substrate.

Electrochemical measurements

The lithium storage properties were evaluated using 2032 coin

cells; Li foil was used as the reference and counter electrodes. The

25168 | J. Mater. Chem., 2012, 22, 25167–25173

MWCNT samples were used as the working electrode. To

prepare the electrode material for the electrochemical test, a

mixture of 92% active material and 8% polyvinylidene fluoride

(PVDF) was dissolved in N-methyl-2-pyrrolidone (NMP) and

cast onto Cu foil (50 mm thickness). Typically 0.3 mg of active

material was coated on the Cu foil. After eliminating the solvent

in a vacuum oven at 120 �C, Cu foil was punched and hydrau-

lically pressed under a pressure of 40 MPa. All coin cells were

assembled in an Ar-filled glove box with less than 0.1 ppm

moisture. The electrodes were separated by a Celgard 2500

battery membrane soaked in a liquid electrolyte consisting of 1.0

MLiPF6, dissolved in a 3 : 7 (v/v) mixture of ethylene carbonate–

diethyl carbonate (EC/DEC; Tomiyama Chemicals). The gal-

vanostatic charge and discharge were controlled between 0.02

and 3.0 V at ambient temperature using a TOSCAT 3100 (TOYO

System Co. Ltd, Japan).

3. Results and discussions

The preparation of the multiwalled carbon nanotube (MWCNT)

drilled structures is outlined in Fig. 1. This method involves cobalt

deposition onto purified MWCNTs as the oxidation catalyst for

the solid-state reaction to formholes on carbonnanotube surfaces,

followed by cobalt removal in acid solution, resulting in drilling of

MWCNTs (DMWCNTs). The morphological structures on

MWCNT samples were examined by transmission electron

microscopy (TEM), as depicted in Fig. 2. Originally, the exterior

walls of the purifiedMWCNTs showed hole-free andwell-ordered

graphitic crystallites, suggesting that our purification method had

no effect on drilling the basal plane of MWCNTs.

After the oxidation of Co/purified MWCNTs, the TEM of

CoOx/DMWCNTs (Fig. 2b) shows how CoOx nanoparticles

(indicated by dark circular objects) drilled through the walls of

the nanotubes, resulting in defective structures with some

nanoparticles still residing in the holes. After Co removal by acid

digestion, the dark particles disappear and the opened hole

structures on the walls of MWCNTs remain (Fig. 2c). The

formation of hole defects indicates that carbon was partially

removed during the solid-state reaction. It can be expected that

This journal is ª The Royal Society of Chemistry 2012

Page 3: Nano-drilled multiwalled carbon nanotubes: characterizations and application for LIB anode materials

Fig. 2 Transmission electron micrographs of MWCNT samples: (a) purified MWCNTs, (b) CoOx/DMWCNTs and (c) DMWCNTs.

Publ

ishe

d on

08

Oct

ober

201

2. D

ownl

oade

d by

Duk

e U

nive

rsity

on

01/1

0/20

13 0

9:41

:43.

View Article Online

the size and depth of hole defects are highly related to the size of

Co catalysts and how the catalyst attached on the MWCNTs.

With the average size of about 4.9 nm for CoOx particles, we can

find many holes with similar dimensions. However, TEM visu-

alizations are only limited to some specific 2D projection loca-

tion. Controlling the size and preventing the agglomeration of

CoOx particles are the keys to control the defect structures.

Thermogravimetric analysis (TGA) was conducted to study

the oxidation profile of the cobalt assisted carbon oxidation for

defect formation (Fig. 3a). It is clear that Co/purified MWCNTs

completely oxidized at a much lower temperature than pristine

and purified MWCNTs. This demonstrates that the cobalt oxide

eased the oxidation of carbon by providing lattice oxygen in

promoting the C–O formation to produce CO gas as shown from

the reaction below.21,23

(i) 6CoO + O2 / 2Co3O4

(ii) Co3O4 + C / 3CoO + CO

Fig. 3 (a) TGA profiles and (b) Raman spectra obtained for MWCNT

samples.

This journal is ª The Royal Society of Chemistry 2012

The CoOx catalyst offers the possibility of forming holes on

the carbon surface at such a low oxidation temperature because

only carbon in contact with CoOx can be removed during the

solid-state reaction, leaving the other parts unaffected. TGA

curves also revealed that cobalt oxide particles were effectively

removed from DMWCNTs after acid washing, because of its

similar carbon oxidation profile to the purified MWCNTs. The

weight loss observed in the low temperatures range (<100 �C)could be attributed to the evaporation of the physisorbed water.

In the above mentioned temperature range, DMWCNTs have

faster weight loss than pristine and purified ones. This behavior

may be related to the presence of holes on the DMWCNTs,

which can absorb more water when in contact with the air.

Fig. 3b compares the Raman spectra of the three different

MWCNT samples. The Raman spectra show two strong peaks at

�1353 and�1581 cm�1, which are responsible for the disordered

sp2 carbon ‘‘D-band’’ and the graphite peak ‘‘G-band’’, respec-

tively.24 The D-band intensity relative to that of the G-band

(ID/IG) is often used as a measure of the nanotube quality. The

DMWCNTs showed the highest ID/IG ratio, followed by the

purified and pristine MWCNTs, respectively. This confirms that

DWMCNTs have the largest amount of graphitic edges and

disordered structures that come from the holes on the sidewalls.

The nitrogen adsorption isotherms of pristine, purified and

drilled MWCNTs are shown in Fig. 4. The isotherm of pristine

MWCNTs (Fig. 4a) is of type II in IUPAC classification and has

a sharp condensation step as the relative pressure approaches

unity,7,25,26 while the nitrogen isotherms of purified MWCNTs

and DMWCNTs are close to type IV with an obvious hysteresis

loop, indicating the main mesoporous characteristic of both

samples. Unlike the purified MWCNTs and DMWCNTs, pris-

tine MWCNTs demonstrate the largest saturated adsorption

amount at P/P0 ¼ 1, which corresponds to the multilayer filling

of large pores.26 Cheol-Min Yang et al. reported a similar

phenomenon on their pore structure study of purified and

unpurified SWCNTs.25 This suggests that pristine MWCNTs

may contain some kinds of metallic impurities, carbonaceous

agglomerates, and/or amorphous carbon forming incompact

structures, which are responsible for large quantity adsorption of

nitrogen gas at high relative pressure. We presume that our

pristine MWCNTs could be a promising gas adsorbent material,

however this is beyond the scope of this publication.

A clear understanding of porosity can be deduced through the

pore size distribution analysis, derived from the BJH method.

Fig. 5 reveals typical pore distribution among three different

MWCNT samples. It is shown that pristine MWCNTs have a

broad distribution of pore sizes, ranging in macropores and

J. Mater. Chem., 2012, 22, 25167–25173 | 25169

Page 4: Nano-drilled multiwalled carbon nanotubes: characterizations and application for LIB anode materials

Fig. 4 N2 adsorption isotherms on MWCNT samples: (a) pristine

MWCNTs, (b) purified MWCNTs and (c) DMWCNTs.

Fig. 5 Pore size distribution of the MWCNT samples.

Publ

ishe

d on

08

Oct

ober

201

2. D

ownl

oade

d by

Duk

e U

nive

rsity

on

01/1

0/20

13 0

9:41

:43.

View Article Online

mesopores. Surprisingly, after purification and drilling treat-

ment, pore sizes larger than 50 nm are significantly reduced. This

suggests that impurities and amorphous carbon, which are

responsible for large multilayer N2 filling structures at high

relative pressure (Fig. 4), were washed away during such

25170 | J. Mater. Chem., 2012, 22, 25167–25173

treatments. In purified MWCNTs and DMWCNTs, the absence

of macropore sites led to a shift of the pore size distribution to the

mesopores range. According to the TEM presented in Fig. 2, our

MWCNT has an outer diameter of about 15 nm and an inner

core diameter in the range of 3–8 nm. Since an inner core

diameter larger than 10 nm was not observed by TEM, the pore

size distribution observed in the range above 10 nm (Fig. 5) could

be related to the intertube mesopores, while pores with diameter

less than 10 nm were related to intratube mesopores. This is

supported by Inoue et al. who found two deviations of a high-

resolution as plot of the N2 adsorption isotherm of MWCNTs,

describing two different mesopore characteristics.27 The inter-

tube mesopores are formed between nanotube criss-crossing,

particularly due to the aggregation of individual MWCNTs. It is

noticeable that purified MWCNTs have a greater participation

of intertube mesopores than that of DMWCNTs. This could be

associated with the strong van der Waals forces of DMWCNTs,

consolidating the intimate contact between nanotubes allowing

for denser packing.28 On the other hand, the intratube mesopores

can be attributed to the inner core and defects on the sidewalls of

carbon nanotubes. A narrow pore size distribution was observed

with the peak at around 4 nm. As we expected, DMWCNTs have

shown the highest feature for such pore size, followed by purified

and pristine MWCNTs. It is evident that by drilling the sidewalls

of MWCNTs, we could create holes and access into the inner

core, in contrast to the pristine and purified ones. This has also

been confirmed by TEM observation (Fig. 2). The specific

surface area was calculated by the BET method (SBET), using

nitrogen adsorption data in the linear BET range (P/P0, 0.05–

0.3). SBET of pristine, purified and drilled MWCNTs are listed in

Table 1. SBET shows an increasing trend by applying purification

and drilling treatments as compared to the pristine MWCNTs.

Fig. 6 shows thermal desorption spectroscopy (TDS) of

DMWCNTs. TDS has been reported as one of the most

commonly used techniques to identify the surface functional

groups of CNTs, which decompose at different temperatures

releasing CO, CO2 and water.29 TDS offers accurate qualitative

and quantitative analysis of surface functionalities by heating the

samples up to a certain temperature and monitoring the evolved

gas by an online mass spectrometer. Generally, the groups of

phenols, carbonyls, quinones, and pyrone-type groups release in

the form of CO, while lactones and carboxylic acids release in the

form of CO2 (and H2O) under heating, with the exception that

carboxylic anhydrides release both CO and CO2. DMWCNTs

exhibit a broad distribution of O-functionalities on the surface,

but are mainly categorized as oxygen double bond functional

groups (carbonyl, quinone, anhydride, carboxyl, etc.). The

presence of a carboxyl hydrophilic group enhances the wetta-

bility nature, which is important for good dispersion in polar

solvents, allowing for the application of CNTs as advanced

energy materials, such as fuel cell electrocatalysts. It is widely

known that the oxidation of carbon not only results in graphitic

edges and disordered defects, but it also introduces functional

groups on the edges. The TDS quantization results demonstrate

that the O-functionalities and oxygen content are increased

noticeably after subsequent purification and drilling treatments,

when compared to pristine MWCNTs (Table 1). Owing to large

defects as observed by the highest value of ID/IG ratio,

DMWCNTs contained the largest amount of functional groups.

This journal is ª The Royal Society of Chemistry 2012

Page 5: Nano-drilled multiwalled carbon nanotubes: characterizations and application for LIB anode materials

Table 1 The surface area and functional groups’ characteristics of MWCNT samples

Sample SBET (m2 g�1)P

CO (mol g�1)P

CO2 (mol g�1)

PO-functionalities

(mol g�1) O (wt%)Estimated Faradaiccapacitya (mA h g�1)

Pristine MWCNTs 270 1.3 � 10�4 6.9 � 10�6 1.4 � 10�4 0.2 4Purified MWCNTs 305.4 3.0 � 10�3 9.9 � 10�4 3.9 � 10�3 7.9 105DMWCNTs 313.8 4.3 � 10�3 1.1 � 10�3 5.4 � 10�3 10.4 144

a The estimated Faradaic capacity is calculated by the assumption that one functional group can react with one Li, as described by the reaction C]O +Li+ + e� 4 C–OLi.

Fig. 6 Thermal desorption profiles of DMWCNTs.

Publ

ishe

d on

08

Oct

ober

201

2. D

ownl

oade

d by

Duk

e U

nive

rsity

on

01/1

0/20

13 0

9:41

:43.

View Article Online

To further study the hole effects on electrochemical perfor-

mance, we tested ourMWCNT samples as the anode material for

LIB. The electrochemical characterizations were conducted by

galvanostatic charge–discharge measurement of the as-prepared

coin cells. In this paper, we refer lithium extraction as the

discharge process and Li insertion as the charging process.

Fig. 7a presents the 20th cycle of charge–discharge curves for the

represented MWCNT samples at a current density of 25 mA g�1.

It shows that DMWCNTs have the largest charge–discharge

capacity. The specific discharge capacities observed for pristine,

purified and drilledMWCNTs at the 20th cycle were 267, 421 and

625 mA h g�1, respectively. Both purified MWCNTs and

DMWCNTs have a higher Li extraction capacity than the

theoretical capacity of graphite (372 mA h g�1, LiC6). This

suggests that purification and drilling treatment play an impor-

tant role in increasing the Li storage capacity. The cyclability for

both the specific discharge capacity and coulombic efficiency

(CE) is presented in Fig. 7b. The CE was derived from the ratio

of discharge to charge capacity on each cycle number. At the first

cycle, the CE of pristine, purified and drilled MWCNTs was 34,

38 and 40%, respectively. The rest of the capacity on the first

cycle is related to the irreversible processes, mainly contributed

from the reductive decomposition of the electrolyte and forma-

tion of Li organic compounds forming the so-called solid elec-

trolyte interface (SEI).2–4 Such a low initial CE characteristic has

been found in various nano-carbon materials because the large

surface area will result in a volume of SEI. Following subsequent

cycling, although the CE of all the samples increased and reached

a stable value, still some irreversible capacities remained. This

indicates that the SEI formed on the initial cycle may not

perfectly prevent the surface from further decomposition of

This journal is ª The Royal Society of Chemistry 2012

electrolyte. It is interesting to note that all MWCNT samples

reached a similar constant CE value of about 93%. This suggests

that the defects not only increase the specific reversible capacity,

but they also increase the electrochemical active surface area,

which is responsible for the increase of irreversible capacity.

All MWCNT samples exhibited a constant specific discharge

capacity over the cycling process (Fig. 7b). In contrast, graphene

materials have also been reported to have Li storage capacity

larger than graphite, however their extraction capacity fading

performances are poor even during a short cycling number.30–32

For instance, the oxidized graphene nanoribbons showed the

specific discharge capacity of 820 mA h g�1 at the first cycle,

which is then decreased to be about 550 mA h g�1 on the 15th

cycle.31 Due to these performance characteristics, CNTs are an

attractive Li anode material in comparison to the recent, widely

studied graphene materials. The stable and high conductivity of

CNTs comes from their rigid structure, which may be the reason

for the high cyclability.

So far many possible lithiation mechanisms in the carbon

materials have been proposed, such as Li intercalation in GIC

structure, Li hold in nanosize voids/cavities and Li bonded with

surface functionalities. Among them, GIC (graphitic intercala-

tion compound) is the most well known one, where the Li ions

are intercalated into the graphitic interlayers. Li insertion and

extraction occur at a very low potential (�0.1 V vs. Li+/Li) with

the maximum capacity of 372 mA h g�1 (LiC6). However, only

the GIC formation cannot explain the fact that many other

carbon materials have larger Li storage capacity than LiC6

value.8 To explain the large reversible capacity, the idea that Li

can be stored in small closed spaces in the carbon has been

proposed.33–36 One of the models has been constructed based on

the random stacking of small domain graphene layers, forming

small voids in between the carbon sheets (‘‘house-card’’

model).35,36 In spite of the different proposed void structures, all

explanations can be subscribed to the Li storage into a small

space in the carbon. Later, it has also been confirmed by low

temperature Li-NMR that lithium stored in the small space, like

cavities and voids, is in a metallic cluster form.37,38

Considering that carbon materials may also contain

functional groups on their edges and defects,8,9 another idea

based on the redox reaction of Li and functional groups has been

introduced. A reversible reduction and oxidation reaction

(C]O + Li+ + e� 4 C–OLi) between lithium ions and surface

functionalities, with the peak of 3.0 V vs. Li+/Li, has been

reported in the study of functional groups for high performance

batteries.39

In order to get better insight into the Li insertion and extrac-

tion process for our samples, the differential capacity (dQ/dV) of

J. Mater. Chem., 2012, 22, 25167–25173 | 25171

Page 6: Nano-drilled multiwalled carbon nanotubes: characterizations and application for LIB anode materials

Fig. 7 The electrochemical characteristics of representedMWCNT samples as Li anode materials: (a) charge–discharge obtained from the 20th cycle of

the differentMWCNT samples at a current density of 25 mA h g�1, (b) cyclability performance of differentMWCNT electrode materials, (c) dQ/dV plot

of the 20th cycle charge (Li insertion) curves (inset figure is the closed-up dQ/dV plot of the 1st and 20th cycle charge curves, showing the SEI formation of the

DMWCNTs) and (d) dQ/dV plot of the 20th cycle discharge (Li extraction) curves for the represented MWCNT samples [black line: pristine MWCNTs;

red line: purified MWCNTs and blue line: DMWCNTs].

Publ

ishe

d on

08

Oct

ober

201

2. D

ownl

oade

d by

Duk

e U

nive

rsity

on

01/1

0/20

13 0

9:41

:43.

View Article Online

the 20th cycle from the charge–discharge curves is plotted in

Fig. 7c and d. Fig. 7c shows that the insertion of Li mainly

occurred at a potential below 0.5 V vs. Li+/Li. For the first cycle,

a large and broad feature can be seen at 0.9 V, corresponding to

the SEI formation (inset, Fig. 7c). However, a small feature is still

left after continuing the cycles, which is responsible for the

irreversible capacities for each cycle, as we discussed before. For

the Li extraction, we found three features (Fig. 7d), which is in

agreement with the previous reports.8–12 The peak around 0.1 V is

associated with the extraction of Li ions from the GIC structure,

while the peak at about 1.2 V can be attributed to the Li

extraction out of the inner core of MWCNTs. A comparison

study of Li intercalation into open and closed SWCNTs has been

conducted by Shimoda et al.12 They found that Li extraction

from the inner core only occurs in open SWCNTs, as evidenced

by the presence of the peak at a similar potential on its dQ/dV of

the Li extraction process. Furthermore, the reversible capacity

increase with an increase of pores with sizes in the range of 3–8

nm is also shown in the BJH plot (Fig. 5). Thus, it seems

reasonable to consider that the peak at about 1.2 V is related to

the Li extraction from the inner core of MWCNTs. Nevertheless,

further investigation is required to clarify such Li species.

The last important feature on the Li extraction dQ/dV curve is

observed in the voltage range of 1.5–3.0 V. This pronounced

feature has been attributed to the extraction of Li ions bonded

with oxygen functionalities, due to some general facts that the

carbon materials are decorated by the functional groups.8–10

However, it was also reported that even after removing the

functional groups by vacuum annealing, such Li extraction in

the high potential range can still be observed.4 Nevertheless, the

origin of such a feature is not well discussed and is still unknown.

From the dQ/dV of Li insertion (Fig. 7c), it can be seen that there

is no obvious feature in the potential range of 1.5–3.0 V. This

indicates that a high potential is required to extract the Li

25172 | J. Mater. Chem., 2012, 22, 25167–25173

inserted at the potential below 0.5 V such that the feature cannot

be explained by the reversible redox reaction as reported

before.39 Additionally, it is clear for our MWCNT samples that

the amount of O-functionalities itself is too small to explain the

large storage capacity (Fig. 7a), as evidenced by the estimated Li

capacity derived from the TDS quantization of functional groups

(Table 1). We conclude that the corresponding extraction feature

is not related to the extraction of Li ions bonded with the

carbonyl groups as proposed in the literature.8–10

To explain a large Li storage capacity for conventionally

prepared carbon material by pyrolizing epoxy resins, the closed

volumes surrounded by a small domain size of graphene layers

was proposed to hold Li clusters.35–38 Due to the high tempera-

ture treatments, it is possible for the non-graphitizable carbon to

have closed cavities, surrounded with small but highly crystalline

graphene layers. The available space for storing Li metal clusters,

together with the less defective graphitic layers, may result in a

large reversible capacity with a relatively low potential of Li

insertion–extraction below 1.0 V. However, this is not the case

for our samples. For CNTs with long dimension rolled graphene

layers, it can be expected that a number of defects may form in

the plane due to the highly stressed structures.

Hence, we propose for CNTs that the Li might be trapped by

defective structures located in the plane of rolled graphene layers,

resulting in Li extraction at high potential. These in-plane defects

are probably located in the discontinuous graphene layers and not

accessible to the electrolyte, which may be formed in the prepa-

rationofCNTs and/or during the oxidation treatments, because of

the highly stressed structure of CNTs. Nevertheless, the origins of

the in-plane defects are not clarified. Further studies are required

to understand the defective structures with dangling bonds,

hydrogen terminated carbon (C–H), heptagon–pentagon rings

and other functional groups, to understand the mechanism and

improve the potential characteristics of Li insertion–extraction.

This journal is ª The Royal Society of Chemistry 2012

Page 7: Nano-drilled multiwalled carbon nanotubes: characterizations and application for LIB anode materials

Publ

ishe

d on

08

Oct

ober

201

2. D

ownl

oade

d by

Duk

e U

nive

rsity

on

01/1

0/20

13 0

9:41

:43.

View Article Online

4. Conclusions

In summary, we have prepared a drilled structure ofMWCNTsby

utilizing CoOx as the carbon oxidation catalyst. The Li storage

capacity, obtained from our drilled MWCNTs, is almost twice as

large as the theoretical LiC6 structure. This emphasizes that the

openings on the sidewalls increase the access into storage sites,

contributing to their large reversible capacity. While most of the

Li insertion occurs at a potential below 0.5 V, the large extraction

of Li at a high potential was explained by the idea of Li trapping

sites in graphene layers. In this study, not only a high reversible

capacity but also a better understanding of the Li storage mech-

anisms were achieved through the structural control of CNTs.

Acknowledgements

We acknowledge Prof. R. Kanno in the Department of Elec-

tronic Chemistry at Tokyo Institute of Technology (TIT) for

providing access for electrochemical measurements. We also

thank Dr K. Nakajima in the Department of Electronic Chem-

istry at TIT for Raman characterization. This work was con-

ducted in a collaborative research project between Tokyo

Institute of Technology and the Japan Science and Technology

Agency (JST). We greatly appreciate financial support by JST.

References

1 X. Li, H. Zhu, L. Ci, C. Xu, Z. Mao, B. Wei, J. Liang and D. Wu,Carbon, 2001, 39, 2077–2079.

2 E. Frackowiak, S. Gautier, H. Gaucher, S. Bonnamy and F. Beguin,Carbon, 1999, 37, 61–69.

3 G. Maurin, C. Bousquet, F. Henn, P. Bernier, R. Almairac andB. Simon, Chem. Phys. Lett., 1999, 312, 14–18.

4 A. S. Claye, J. E. Fischer, C. B. Huffman, A. G. Rinzler andR. E. Smalley, J. Electrochem. Soc., 2000, 147, 2845–2852.

5 D. N. Futaba, K. Hata, T. Yamada, T. Hiraoka, Y. Hayamizu,Y. Kakudate, O. Tanaike, H. Hatori, M. Yumura and S. Iijima,Nat. Mater., 2006, 5, 987–994.

6 K. H. An, W. S. Kim, Y. S. Park, J.-M. Moon, D. J. Bae, S. C. Lim,Y. S. Lee and Y. H. Lee, Adv. Funct. Mater., 2001, 11, 387–392.

7 Z. Li, Z. Pan and S. Dai, J. Colloid Interface Sci., 2004, 277, 35–42.8 J.-Y. Eom, D.-Y. Kim and H.-S. Kwon, J. Power Sources, 2006, 157,507–514.

9 S. Klink, E. Ventosa, W. Xia, F. La Mantia, M. Muhler andW. Schuhmann, Electrochem. Commun., 2012, 15, 10–13.

10 X. X. Wang, J. N. Wang, H. Chang and Y. F. Zhang, Adv. Funct.Mater., 2007, 17, 3613–3618.

11 D. T. Welna, L. Qu, B. E. Taylor, L. Dai and M. F. Durstock, J.Power Sources, 2011, 196, 1455–1460.

This journal is ª The Royal Society of Chemistry 2012

12 H. Shimoda, B. Gao, X. P. Tang, A. Kleinhammes, L. Fleming,Y. Wu and O. Zhou, Phys. Rev. Lett., 2001, 88, 015502.

13 K. Moriguchi, S. Munetoh, M. Abe, M. Yonemura, K. Kamei,A. Shintani, Y. Maehara, A. Omaru and M. Nagamine, J. Appl.Phys., 2000, 88, 6369–6377.

14 C. Mik�o, M. Milas, J. W. Seo, E. Couteau, N. Bari�si�c, R. Ga�al andL. Forr�o, Appl. Phys. Lett., 2003, 83, 4622–4624.

15 P. S. Spinney, D. G. Howitt, S. D. Collins and R. L. Smith,Nanotechnology, 2009, 20, 465301.

16 A. V. Krasheninnikov and K. Nordlund, Jpn. J. Appl. Phys., 2010,107, 071301.

17 P. M. Ajayan, T. W. Ebbesen, T. Ichihashi, S. Iijima, K. Tanigaki andH. Hiura, Nature, 1993, 362, 522–525.

18 Y. Yamada, O. Kimizuka, K. Machida, S. Suematsu, K. Tamamitsu,S. Saeki, Y. Yamada, N. Yoshizawa, O. Tanaike, J. Yamashita,F. Don, K. Hata and H. Hatori, Energy Fuels, 2012, 24, 3373–3377.

19 O. Byl, J. Liu and J. T. Yates, Langmuir, 2005, 21, 4200–4204.20 J.-Y. Eom and H.-S. Kwon,Mater. Chem. Phys., 2011, 126, 108–113.21 D.-H. Kim and K. Waki, J. Nanosci. Nanotechnol., 2010, 10, 2375–

2380.22 K. Matsubara and K. Waki, Electrochem. Solid-State Lett., 2010, 13,

F7–F9.23 D. Eder, Chem. Rev., 2010, 110, 1348–1385.24 A. C. Dillon, T. Gennett, K.M. Jones, J. L. Alleman, P. A. Parilla and

M. J. Heben, Adv. Mater., 1999, 11, 1354–1358.25 C.-M. Yang, K. Kaneko, M. Yudasaka and S. Iijima, Nano Lett.,

2002, 2, 385–388.26 K. Kaneko, J. Membr. Sci., 1994, 96, 59–89.27 S. Inoue, N. Ichikuni, T. Suzuki, T. Uematsu andK. Kaneko, J. Phys.

Chem. B, 1998, 102, 4689–4692.28 Q.-H. Yang, P.-X. Hou, S. Bai, M.-Z.Wang andH.-M. Cheng,Chem.

Phys. Lett., 2001, 345, 18–24.29 S. Kundu, Y. Wang, W. Xia and M. Muhler, J. Phys. Chem. C, 2008,

112, 16869–16878.30 E. Yoo, J. Kim, E. Hosono, H. Zhou, T. Kudo and I. Honma, Nano

Lett., 2008, 8, 2277–2282.31 T. Bhardwaj, A. Antic, B. Pavan, V. Barone and B. D. Fahlman, J.

Am. Chem. Soc., 2010, 132, 12556–12558.32 Z.-S. Wu, W. Ren, L. Wen, L. Gao, J. Zhao, Z. Chen, G. Zhou, F. Li

and H.-M. Cheng, ACS Nano, 2010, 4, 3187–3194.33 A. Mabuchi, K. Tokumitsu, H. Fujimoto and T. Kasuh, J.

Electrochem. Soc., 1995, 142, 1041–1046.34 M. Nagao, C. Pitteloud, T. Kamiyama, T. Otomo, K. Itoh,

T. Fukunaga, K. Tatsumi and R. Kanno, J. Electrochem. Soc.,2006, 153, A914–A919.

35 J. S. Xue and J. R. Dahn, J. Electrochem. Soc., 1995, 142, 3668–3677.36 Y. Liu, J. S. Xue, T. Zheng and J. R. Dahn, Carbon, 1996, 34, 193–

200.37 K. Tatsumi, J. Conard, M. Nakahara, S. Menu, P. Lauginie,

Y. Sawada and Z. Ogumi, J. Power Sources, 1999, 81-82, 397–400.38 J. Conard and P. Lauginie, Tanso, 2000, 191, 62–70.39 S. W. Lee, B. M. Gallant, Y. Lee, N. Yoshida, D. Y. Kim,

Y. Yamada, S. Noda, A. Yamada and Y. S. Horn, Energy Environ.Sci., 2012, 5, 5437–5444.

J. Mater. Chem., 2012, 22, 25167–25173 | 25173