DOI: 10.1002/cssc.200900187

Composites of Molecular-Anchored Graphene andNanotubes with Multitubular Structure: A New Type ofCarbon ElectrodeXi Liu,[a] Yong-Sheng Hu,*[b] Jens-Oliver M�ller,[a] Robert Schlçgl,[a] Joachim Maier,[c] andDang Sheng Su*[a]

Introduction

The discovery of fullerene and carbon nanotubes (CNTs) re-vealed the exciting wrapping and rolling capability of gra-phene sheets resulting in low-dimensional carbon struc-tures.[1–14] Such architectures are used in many applications;one of these is the storage of Li between interlayer spaces,which is enhanced compared to planar graphene. Recently, alot of effort has been focused on the large scale production ofgraphene and graphene composites by facile wet-chemicalmethods, as these material have displayed promising potentialfor such applications.[6, 13–14] However, the influence of thechemical modification on the physical and chemical propertiesof the graphene composites has not been thoroughly investi-gated.[8] For example, the dimensional stability of graphitic ma-terials with respect to the interlayer distance of graphene isone of the key factors for the reversibility of lithium ion batter-ies. Although the objectives could be achieved by developinga synthetic methodology for the preparation of novel carbonmaterials, chemical modification could be a more facile and ef-ficient method for the alteration and improvement of the func-tional properties of graphite materials as even a simple oxida-tion treatment could lead to a significant increase in the inter-layer distance of graphite materials.[15]

Zhu and co-workers reported a method for the preparationof bi-tubular nanocarbon materials (carbon tube-in-tube, CTIT).The self-assembly of graphene was initiated under mild reac-tion conditions over CNTs, which acted as a template.[16] It wasnoted that isolated graphene sheets, dissolved from adventi-tious carbon occurring during CNT synthesis, were used asbuilding blocks for graphene composites in solution at atmos-pheric pressure and low temperature. Furthermore, the re-maining functionalized edge sites of the graphene can beused for additional chemical modification to improve the phys-icochemical properties of graphene composites.

In the present work, a method is developed that uses thecondensation of oxygen functionalities between different edgesites of adjacent graphene sheets to bind the sheets throughcovalent bonding. This method is analogous to the condensa-tion reaction between hydroxo-metal complexes formingmetal oxides known as the “olation” reaction.[17] The edge ofgraphene sheets or CNTs can be lightly etched with acids toafford oxygen-containing groups (phenolic hydroxides and car-boxylic acids), which facilitate covalent bonding.[18] The gra-phene sheets can be further modified by oxalic acid moleculesand then poly-condensed under the evolution of water viaoxalyl bonding (Figure 1). The graphene sheets assemble alongthe tubular axis of templating CNTs, forming layered co-axialtubes (>3). The difference in molecular structuring of the ini-tial two-dimensional continuous graphene units forming thetemplate, and the condensation product of graphene islandswith functionalized carbon–carbon bonds in between, givesrise to different interlayer arrangements. It can be assumed

Graphene–carbon nanotube nanocomposites that contain mul-titubular co-axial and hollow cavity microstructures are pre-pared. Nanometer-scale graphene sheets are anchored withoxalic acid and consequently linked to each other via oxalylbonding, thereby self-assembling into numerous outer tubeswith distinct borders and a homogeneous thickness along theinnermost pristine tube, which acts as a template. The result-ing interstitial inclusion of oxalic acid into the graphene stack-

ing modifies both the surface and the bulk properties of thenewly formed tubes. It is observed that the unique microstruc-ture of the modified graphene–carbon nanotube nanocompo-site significantly facilitates the insertion and extraction of lithi-um, demonstrating superior electrochemical performance asanodes for lithium-based batteries. This facile chemical ap-proach provides a new graphene architecture, showing superi-or stability, for use as anode material in lithium ion batteries.

[a] Dr. X. Liu, Dr. J.-O. M�ller, Prof. R. Schlçgl, Dr. D. S. SuFritz-Haber-Institut der Max-Planck-GesellschaftFaradayweg 4–6,14195 Berlin (Germany)Fax: (+ 49) 30 8413 4401E-mail : dangsheng@fhi-berlin.mpg.de

[b] Dr. Y.-S. HuBeijing National Laboratory for Condensed Matter PhysicsInstitute of Physics, Chinese Academy of Sciences100190 Beijing (PR China)Fax: (+ 86) 10 82649046E-mail : yshu@aphy.iphy.ac.cn

[c] Prof. J. MaierMax-Planck-Institut f�r FestkçrperforschungHeisenbergstr. 1, 70569 Stuttgart (Germany)

Supporting Information for this article is available on the WWW underhttp://dx.doi.org/10.1002/cssc.200900187.

ChemSusChem 2010, 3, 261 – 265 � 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim 261

that the templating carbon forms an excellent electronic con-ductor and the patchwork-like condensation layers will exhibita storage space for Li with oxygen functionalities and an en-hanced interlayer thickness. The overall co-axial packaging ofthe two functions allows for good mesoscopic transport ofcharges and ions and maintains the overall dimensional stabili-ty. The many covalent carbon–carbon bonds within the gra-phene in perpendicular orientation to the mechanical strainduring insertion and depletion of Li are chemically non-reac-tive and will keep the overall composite stable.

A covalent bond between graphene sheets can be formedwhen the condensing species contain carboxylic acid and phe-nolic hydroxide functional groups.[18] An additional ether-likeconnectivity via a carbon–oxygen group may arise from thecondensation of two carboxylic groups of differing pK value(local environment). This is a much more frequent case underpreparation conditions.

Results and Discussion

Commercial CNTs purchased from Applied Science were usedas starting materials. The sample, CTIT, was prepared accordingto the procedure described in Ref. [16] ; this sample was modi-fied with oxalic acid to yield CTIT2 (see Experimental Section).The morphology and microstructure of pristine CNTs and CTIT1are shown in Figure 2. The pristine CNTs display a fish-bone mi-

crostructure combined with a poorly graphized carbon depos-it.[16] Figure 2 b shows the CTIT1 sample, which forms followingoxidation and reintegration of CNTs. The carbon deposit hasdisappeared and the tubes are overgrown with a bi-tubularco-axial microstructure with diameters ranging from 10 to20 nm. The growth mode of CTIT1 has been discussed in previ-ous work.[16]

The significant influence the addition of oxalic acid to CTIT1has on the morphology and microstructure are shown inFigure 3. The resulting composite CTIT2 displays a multitubular

co-axial microstructure, consisting of multiple tubes formedalong the pristine tube with a homogeneous thickness rangingfrom 10 to 50 nm. In addition, a clear cavity between newlyformed concentric carbon nanotubes was also observed. Thesignificant difference in the microstructure of CTIT1 and CTIT2suggests that the formation of nanocarbons with a multitubu-lar microstructure is attributed to the addition of oxalic acid tothe acidic graphene suspension.

The difference in microstructure and functionalized natureof prepared CTIT1 and CTIT2 was studied by thermogravimet-ric mass spectrometry (TG-MS). The TG plots of pristine andmodified nanocarbons are displayed in Figure 4 a. Comparedto the starting material, both CTIT1 and CTIT2 exhibit a weightloss starting at 200 8C and reaching about 10 % at 1000 8C. Theenhanced weight loss compared to the pristine CNTs is due tothe defective microstructure on the outer surface (see high res-olution transmission spectroscopy (HRTEM) images in Figure 5),which contains increased functionality compared to the pris-tine CNTs. A difference in the TG curves of CTIT1 and CTIT2was apparent above 700 8C. As it is shown in Table 1, the CTIT2sample lost about 2 % more weight at 1000 8C. Strong CO andCO2 desorption peaks were observed for both CTIT1 and CTIT2at 700 8C, which is associated with the weight loss (Figure S2).For CTIT2, an increased amount of CO desorption was ob-served, confirming that the extra desorption species are car-bonyl functionalities assigned to oxalic acid modification. Thisstrongly supports the proposed model where the graphenesheets in CTIT2 are covalently connected via oxylal bondingowing to the interstitial inclusion of oxalic acid into the gra-phene stacking. The chemical nature of the highly functional-ized graphene composites were also identified by X-ray photo-electron spectroscopy (XPS), which showed that the oxygenatom concentration on the surface of CTIT2 was 20 atom %,twice as much as that of CTIT1.

TEM images of CTIT1 and CTIT2 after thermal treatment aredisplayed in Figure 4 b and 4 c and Figure S1. The images show

Figure 2. TEM and HRTEM images of pristine a) CNTs and b) CTIT1.

Figure 3. a, b) TEM and c, d) SEM images of CTIT2.

Figure 1. Oxalyl bridge between graphene sheets.

262 www.chemsuschem.org � 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemSusChem 2010, 3, 261 – 265

Y.-S. Hu, D. S. Su et al.

little influence of thermal treatment on the outer morphologyof the newly formed carbon in CTIT1, suggesting that the gra-phene units are polymerized via multiple carbon–carbonbonds being insensitive to thermal treatment. In contrast, theremarkable deconstruction of newly formed carbon in CTIT2 isassociated with the loose aggregation of graphene sheets aris-ing from oxygen species involved in the interconnections ofthe graphene units. A close inspection of the organization ofthe walls of the heat-treated composites revealed that the dif-ference between inner and outer graphene layers has disap-peared and that the walls now consist of randomly orientedgraphene basic structural units of ca. 2 nm in size with nolayer stacking. Some of these units are locally stacked in paral-lel and give rise to strong diffraction contrast (black dots).

It is obvious that both the edge sides and defects of isolatedgraphene sheets were modified by oxalic acid. Therefore, theinterlayer distance of aggregated graphene sheets should in-crease owing to the modification. The HRTEM images and theelectron diffraction patterns of CTIT1 and CTIT2 are shown inFigure 5. A highly oriented alignment of graphene in thenewly formed tube was observed for both CTIT1 and CTIT2.This was confirmed by the electron diffraction patterns, whichshowed spots rather than rings as expected for distributions ofstacking spaces. The interlayer distance of the newly formedtubes in CTIT2 was ca. 3.78 �, which is significantly larger thanthat for CTIT1 (3.66 �) and for the interlayer distance of pristineCNTs (3.38–3.40 �). The enlarged stacking distances, deter-mined by HRTEM, are consistent with the complex geometryof the interconnections of the graphene units arising from theattached oxygen groups. These oxygen groups contribute withtheir extra electrons to a bending of the graphene units. Thiscan be seen in the HRTEM images. The outer polymerized gra-phene units can occur in columns of about 2 nm large gra-phene units or are coupled in sequential order with no relativeorientation in stacking direction. In both cases, it can be clearlyseen that the template graphene units are much larger andhighly parallel in their relative alignment supporting thenotion of the inner “nanowire” covered by the outer “nano-sponge”. The images in Figure 5 also show the weakness ofthe synthetic concept being the space between the wire andthe sponge. Poor contact between atoms and charges will de-teriorate both the capacity and the charging kinetics of a bat-tery electrode made from these composites.

It can be observed in Figures 2 and 3 that the total thick-ness of the newly formed walls of the graphene CNT nano-composite in CTIT2 is remarkably higher than in CTIT1. This in-dicates that the polycondensation process of graphene unitsonly by neighboring oxygen functional groups is less effective

Figure 4. a) TG profiles of carbon nanotubes before and after modification.TEM images of b) CTIT1 and c) CTIT2 after thermal treatment.

Figure 5. HRTEM images and electron diffraction patterns of a) CTIT1 andb) CTIT2.

Table 1. Weight loss of CTIT1 and CTIT2 at different calcination tempera-tures.

Weight loss [%]Sample 200 8C 600 8C 800 8C 1000 8C

CTIT1 1.2 6.3 8.5 9.0CTIT2 0.7 6.7 9.6 11.0

ChemSusChem 2010, 3, 261 – 265 � 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemsuschem.org 263

Composites of Molecular-Anchored Graphene and Nanotubes

than the coupling via oxalic acidlinkers. However, it may be ex-pected that the overall compo-site tubular structure should bedimensionally stable and thusprovide good host propertiesfor a Li-battery anode.

The reversibility of lithium-based batteries relies on the di-mensional stability of the hostmaterials during the insertionand extraction of Li. Significantexpansion and consequent exfo-liation and cracking of the elec-trode have been widely report-ed, resulting in a serious deacti-vation of the battery.[19–24] In ad-dition, the dimensional stabilityas a prerequisite requirementfor a good anode has beenpushed since propylene carbon-ate, an electrolyte with en-hanced safety, may cause severesolvent co-intercalation in di-mensionally less stable hosts,leading to deterioration of thegraphitic material and, as aresult, drastic lowering of thestorage performance. The miss-ing compatibility between elec-trode and electrolyte for conventional carbon hosts hinderedthe commercial application of propylene carbonate. A practicalsolution would be to improve the dimensional stability of agraphitic electrode by modifying the terminating atoms of thegraphene sheets such as to prevent solvent co-intercalation,while simultaneously allowing for fast Li insertion and extrac-tion. This approach could be realized with the present carbon–carbon composite, provided that sufficient chemical stabilitycan be built into the rich carbon–oxygen bonding structures.

Figure 6 a shows the discharge (Li insertion) and charge (Liextraction) curves of the CTIT2 electrode cycled in 1 m LiPF6

ethylene carbonate (EC)/dimethyl carbonate (DMC) (1:1 byvolume) at a rate of C/5 (one lithium per six formula units(LiC6) in 5 h). Two long flat plateaus at low voltages in the dis-charge and charge curves are observed as well as clear reduc-tion and oxidation peaks in the cyclic voltammetry (CV) curve(Figure 6 b). The features are ascribed to Li insertion and ex-traction into the graphene layers being in good agreementwith the graphene alignment seen in Figure 5. A large irreversi-ble capacity is observed in the first cycle, which is likely owingto the formation of the solid electrolyte interphase (SEI) as cor-roborated by the appearance of a reduction peak at around0.6 V in the first cycle and the disappearance of this peak inthe second cycle in the CV curve (Figure 6 b). The first chargecapacity is about 380 mA h g�1 and the reversible capacity sta-bilizes at 370 mA h g�1 after 20 cycles (Figure 6 c). CTIT2 showsthe best cycling performance with higher reversible capacity

when compared with the original CNT and CTIT1 (Figure 6 cand Figure S3). The superior stable Li storage performance ofCTIT2 could be attributed to the larger space between the gra-phene interlayer after the interstitial functionalization, favoringinsertion and extraction of lithium during the discharge andcharge processes. Furthermore, it is important to note that thecycling performance of CTIT2 is also excellent at a rate of C/5in a propylene carbonate-based electrolyte (Figure 6 d), whichis considered a safe and low temperature electrolyte, eventhough the propylene carbonate and solvated Li ions tend toco-intercalate into graphite accompanied by severe exfoliationof graphite.[25–27] The remarkable Li insertion and extractionperformance results from the unique structure of the CTIT2modified by oxalic acid.

Conclusions

A modification method was used to prepare graphene compo-sites with unique microstructure and electrochemical proper-ties. Oxalic acid molecules were immobilized on the grapheneand the functionalized graphene assembled into multitubularnanotubes. Consequently, interstitial inclusion of moieties in-creased the interlayer distance of graphene and improved ad-herence of graphene stacking. Simultaneously, the electro-chemical properties were significantly improved by facilitatinginsertion and extraction of lithium. The synthesis and intersti-tial functionalization of CTIT exploit an interesting and promis-

Figure 6. a) Galvanostatic discharge (Li insertion, voltage decreases)/charge (Li extraction, voltage increases)curves of CTIT2 cycled at a rate of C/5 in 1 m LiPF6 in EC/DMC solution. b) CV of CTIT2 at a scan rate of 0.1 mV s�1

in the voltage range of 0.01–3 V in 1 m LiPF6 in EC/DMC solution.c) Comparison of electrochemical performance ofCTIT1 and CTIT2 in 1 m LiPF6 in EC/DMC solution. d) Galvanostatic discharge/charge curves of CTIT2 cycled at arate of C/5 in 1 m LiClO4 in PC solution.

264 www.chemsuschem.org � 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemSusChem 2010, 3, 261 – 265

Y.-S. Hu, D. S. Su et al.

ing approach to the fabrication and functionalization of nano-carbon materials.

Experimental Section

Commercial CNTs purchased from Applied Science were used asstarting materials (PS). The lengths of the nanotubes were up totens of micrometers and they exhibited a diameter distribution of50–170 nm. The raw CNT sample (1 g) was oxidized in concentrat-ed HNO3 (100 mL) for 20 h under refluxing condition with stirring.After filtration, the remaining solid was washed with deionizedwater and ethanol and dried in air for 12 h at 110 8C. To obtainCTIT1, the oxidized sample (50 mg) and concentrated H2SO4

(0.2 mL) was dispersed in tetrahydrofuran (50 mL) and then heatedunder reflux for 10 h to allow the full reintegration of graphene.The suspension was filtered, washed with ethanol, and dried in airfor 12 h at 110 8C to yield CTIT1.[16]

To obtain the modified CTIT, oxalic acid (0.5 g) and oxidized CNTs(50 mg) were dispersed in tetrahydrofuran (50 mL) and concentrat-ed H2SO4 (0.2 mL). The suspension was heated under refluxed for10 h. After filtration and washing, the obtained solid was dried inair for 12 h at 110 8C to yield CTIT2. The heating treatment wasconducted as following: a sample of CTIT1 or CTIT2 was heatedunder a helium atmosphere with a heating rate of 10 8C min�1 upto 600 8C or 800 8C and a holding time of 10 h. The total flow rateof helium was 70 mL min�1.TEM images were recorded on a Philips CM 200 LaB6 operated atan accelerating voltage of 200 kV or a Phillips CM200 FEG field-emission gun electron microscope operated at an acceleratingvoltage of 200 kV. The samples were prepared by suspending thesolid powder in ethanol or chloroform under ultrasonic vibration.One drop of the prepared suspension was brought onto holycarbon films on copper grids. SEM was performed with a Hitachi S-4000 apparatus operated at 15 kV and 25 kV acceleration voltages.Thermal decomposition behavior of the samples was examined byTG-MS with NETZSCH 4. The heating ramp was 10 8C min�1 and theflow rate of Ar was 70 mL min�1.Lithium insertion and extraction experiments were carried outusing two-electrode Swagelok-type cells. The working electrodeswere prepared by mixing the CTIT1, CTIT2, or CNTs and polyvinyldi-fluoride at a weight ratio of 90:10 and pasting the slurry on pureCu foil (99.6 %, Goodfellow). A glass fiber (GF/D) from Whatmanwas used as a separator. Pure lithium foil (Aldrich) was used as acounter electrode. The electrolyte consisted of a solution of 1 m

LiPF6 in EC/DMC (1:1 by volume) obtained from Ube Industries or asolution of 1 m LiClO4 in propylene carbonate. The cells were as-sembled in an argon-filled glove box. Electrochemical performan-ces were tested at different current densities in the voltage rangeof 0.01–3 V on an Arbin MSTAT battery test system. CV experimentswere performed on a VoltaLab 80 electrochemical workstation at ascan rate of 0.1 mV s�1.

Acknowledgements

The authors are indebted to the Max Planck Society and ac-knowledge support in the framework of the ENERCHEM project.

Keywords: batteries · graphene · lithium · nanocomposites ·nanotubes

[1] K. S. Novoselov, A. K. Geim, S. V. Morozov, D. Jiang, Y. Zhang, S. V. Dubo-nos, I. Grigorieva, A. A. Firsov, Science 2004, 306, 666.

[2] K. S. Novoselov, D. Jiang, F. Schedin, T. J. Booth, V. V. Khotkevich, S. V.Morozov, A. K. Geim, Proc. Natl. Acad. Sci. USA 2005, 102, 10451.

[3] A. K. Geim, K. S. Novoselov, Nature 2007, 6, 183.[4] E. Rotenberg, A. Bostwick, T. Ohta, J. L. McChesney, T. Seyller, K. Horn,

Nat. Mater. 2008, 7, 258.[5] S. Y. Zhou, G.-H. Gweon, A. V. Fedorov, P. N. First, W. A. de Heer, D.-H.

Lee, F. Guinea, A. H. Castro Neto, A. Lanzara, Nat. Mater. 2007, 6, 770.[6] A. A. Balandin, S. Ghosh, W. Z. Bao, I. Calizo, D. Teweldebrhan, F. Miao,

C. N. Lau, Nano Lett. 2008, 8, 902.[7] X. Wang, L. J. Zhi, K. M�llen, Nano Lett. 2008, 8, 323.[8] S. Gilje, S. Han, M. Wang, K. L. Wang, R. B. Kaner, Nano Lett. 2007, 7,

3394.[9] S. Watcharotone, D. A. Dikin, S. Stankovich, R. Piner, I. Jung, G. H. B.

Dommett, G. Evmenenko, S. E. Wu, S. F. Chen, C. P. Liu, S. T. Nguyen,R. S. Ruoff, Nano Lett. 2007, 7, 1888.

[10] S. Stankovich, D. A. Dikin, G. H. B. Dommett, K. M. Kohlhaas, E. J.Zimney, E. A. Stach, R. D. Piner, S. B. T. Nguyen, R. S. Ruoff, Nature 2006,442, 282.

[11] T. A. Land, T. Michely, R. J. Behm, J. C. Hemminger, G. Comsa, Surf. Sci.1992, 264, 261.

[12] A. Nagashima, K. Nuka, H. Itoh, T. Ichinokawa, C. Oshima, Surf. Sci. 1993,291, 93.

[13] D. Li, M. B. Muller, G. Scott, R. B. Kaner, G. G. Wallace, Nat. Nanotechnol.2008, 3, 101.

[14] Y. Xu, H. Bai, G. Lu, C. Li, G. Shi, J. Am. Chem. Soc. 2008, 130, 5856.[15] H. K. Jeong, Y. P. Lee, R. J. Lahaye, M. Park, K. H. An, I. J. Kim, C. W. Yang,

C. Y. Park, R. S. Ruoff, Y. H. Lee, J. Am. Chem. Soc. 2008, 130, 1362.[16] Z. P. Zhu, D. S. Su, G. Weinberg, R. E. Jentoft, R. Schlçgl, Small 2005, 1,

107.[17] M. Henry, J. P. Jolivet, J. Livage, Struct. Bonding (Berlin) 1992, 77, 155.[18] M. Sano, A. Kamino, J. Okamura, S. Shinkai, Science 2001, 293, 1299.[19] J. R. Dahn, T. Zheng, Y. Liu, J. S. Xue, Science 1995, 270, 590.[20] M. Winter, J. O. Besenhard, M. E. Spahr, P. Nov�k, Adv. Mater. 1998, 10,

725.[21] H. Zhou, S. Zhu, M. Hibino, I. Honma, M. Ichihara, Adv. Mater. 2003, 15,

2107.[22] E. Yoo, J. Kim, E. Hosono, H. S. Zhou, T. Kudo, I. Honma, Nano. Lett.

2008, 8, 2277.[23] F. Bonino, S. Brutti, P. Reale, B. Scrosati, L. Gherghel, J. Wu, K. M�llen,

Adv. Mater. 2005, 17, 743.[24] K. T. Lee, J. C. Lytle, N. S. Ergang, S. M. Oh, A. Stein, Adv. Funct. Mater.

2005, 15, 547.[25] G. H. Wrodnigg, T. M. Wrodnigg, J. O. Besenhard, M. Winter, Electrochem.

Commun. 1999, 1, 148.[26] Y. S. Hu, W. H. Kong, H. Li, X. J. Huang, L. Q. Chen, Electrochem.

Commun. 2004, 6, 126.[27] Y. S. Hu, W. H. Kong, Z. X. Wang, L. Hong, X. J. Huang, L. Q. Chen, Electro-

chem. Solid-State Lett. 2004, 7, A442.

Received: July 31, 2009Revised: October 13, 2009Published online on January 7, 2010

ChemSusChem 2010, 3, 261 – 265 � 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemsuschem.org 265

Composites of Molecular-Anchored Graphene and Nanotubes