APPLIED PHYSICS REVIEWS Surface characterization and functionalization...

APPLIED PHYSICS REVIEWS

Surface characterization and functionalization of carbon nanofibersK. L. Klein,1,2 A. V. Melechko,2,3 T. E. McKnight,4 S. T. Retterer,2,5 P. D. Rack,1,2

J. D. Fowlkes,2 D. C. Joy,1,2 and M. L. Simpson1,2,a�

1Department of Materials Science and Engineering, University of Tennessee, Knoxville, Tennessee 37996,USA2Center for Nanophase Materials Sciences, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831,USA3Material Science and Technology Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831,USA4Engineering Science and Technology Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee37831, USA5Biosciences Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831, USA

�Received 12 November 2007; accepted 10 December 2007; published online 17 March 2008�

Carbon nanofibers are high-aspect ratio graphitic materials that have been investigated for numerousapplications due to their unique physical properties such as high strength, low density, metallicconductivity, tunable morphology, chemical and environmental stabilities, as well as compatibilitywith organochemical modification. Surface studies are extremely important for nanomaterialsbecause not only is the surface structurally and chemically quite different from the bulk, but itsproperties tend to dominate at the nanoscale due to the drastically increased surface-to-volume ratio.This review surveys recent developments in surface analysis techniques used to characterize thesurface structure and chemistry of carbon nanofibers and related carbon materials. These techniquesinclude scanning probe microscopy, infrared and electron spectroscopies, electron microscopy, ionspectrometry, temperature-programed desorption, and atom probe analysis. In addition, this articleevaluates the methods used to modify the surface of carbon nanofibers in order to enhance theirfunctionality to perform across an exceedingly diverse application space. © 2008 American Instituteof Physics. �DOI: 10.1063/1.2840049�

TABLE OF CONTENTS

I. INTRODUCTION. . . . . . . . . . . . . . . . . . . . . . . . . . . . 1II. SURFACE CHARACTERIZATION

TECHNIQUES. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3A. Scanning probe microscopy. . . . . . . . . . . . . . . 3

1. Scanning tunneling microscopy. . . . . . . . . 32. Atomic force microscopy. . . . . . . . . . . . . . 4

B. Infrared spectroscopy. . . . . . . . . . . . . . . . . . . . 4C. Electron spectroscopy. . . . . . . . . . . . . . . . . . . . 6

1. X-ray photoelectron spectroscopy. . . . . . . . 62. Ultraviolet photoelectron spectroscopy. . . . 73. Auger electron spectroscopy. . . . . . . . . . . . 8

D. Electron microscopy. . . . . . . . . . . . . . . . . . . . . 91. Scanning electron microscopy. . . . . . . . . . . 102. Energy dispersive x-ray microanalysis. . . . 103. Transmission electron microscopy. . . . . . . 114. Electron energy loss spectroscopy. . . . . . . 12

E. Secondary ion mass spectrometry. . . . . . . . . . 131. Dynamic SIMS. . . . . . . . . . . . . . . . . . . . . . 132. Static SIMS. . . . . . . . . . . . . . . . . . . . . . . . . 14

F. Temperature-programed desorption. . . . . . . . . 14

G. Atom probe. . . . . . . . . . . . . . . . . . . . . . . . . . . . 15III. SURFACE MODIFICATION TECHNIQUES. . . . 16

A. Chemical vapor deposition of thin filmcoatings. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161. In situ PECVD coatings and surface

modification. . . . . . . . . . . . . . . . . . . . . . . . . 162. CVD of oxides and nitrides. . . . . . . . . . . . 173. CVD of polymers. . . . . . . . . . . . . . . . . . . . 184. Atomic layer deposition. . . . . . . . . . . . . . . 18

B. Electro- or electroless plating. . . . . . . . . . . . . 19C. Chemical functionalization. . . . . . . . . . . . . . . . 20

1. Wet etch treatment. . . . . . . . . . . . . . . . . . . . 202. Plasma treatment. . . . . . . . . . . . . . . . . . . . . 203. Photochemical functionalization. . . . . . . . . 214. Thermal treatment. . . . . . . . . . . . . . . . . . . . 215. Addition of linkers and polymers. . . . . . . . 21

D. Biochemical functionalization. . . . . . . . . . . . . 221. Proteins and antibodies. . . . . . . . . . . . . . . . 222. DNA. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

IV. SUMMARY AND OUTLOOK. . . . . . . . . . . . . . . . 23

I. INTRODUCTION

The rapid surge in new synthetic nanomaterials demandscomplementary advancement in surface characterizationa�Electronic mail: [email protected].

JOURNAL OF APPLIED PHYSICS 103, 061301 �2008�

0021-8979/2008/103�6�/061301/26/$23.00 © 2008 American Institute of Physics103, 061301-1

Downloaded 31 Mar 2008 to 128.219.239.94. Redistribution subject to AIP license or copyright; see http://jap.aip.org/jap/copyright.jsp

http://dx.doi.org/10.1063/1.2840049

http://dx.doi.org/10.1063/1.2840049

techniques and modification methods in order to understandand utilize these materials. It is well known that surfaceproperties often deviate substantially from the bulk materialproperties due to a difference in physical structure and chem-istry. Moreover, the surface is a dynamic system that inter-acts with the environment, a characteristic which is exploitedfor many applications. The surfaces of nanostructured mate-rials are of special significance because of their enhancedrole in determining functional properties—a phenomenonwhich becomes more pronounced as the surface-to-volumeratio increases. At the same time the complex morphology ofnanostructured surfaces also creates new challenges for theircharacterization. To a large extent, the development of sur-face science of the last century has been based on the as-sumption that the sample presents a flat, well-defined sur-face, which is examined under ultrahigh vacuum. Thetranslation of traditional surface characterization techniquesto the study of complex three-dimensional functional sur-faces is nontrivial and the methods employed are often spe-cific to each particular family of nanostructured materials.

Among the multitude of nanomaterials, carbon nano-structures hold a special place due to their mechanicalstrength and chemical stability. In addition, the covalentchemistry of carbon with oxygen, hydrogen and nitrogenprovides facile routes for functionalization of carbon sur-faces with organic or biological molecules. In elementalform, carbon constructs allotropes with different kinds ofcarbon-carbon bonds, such as sp2-based graphite andsp3-based diamond, resulting from the variety of covalentbonding arrangements provided by orbital hybridization.1

Fullerenes are a class of sp2-based nanostructured materialsthat can be seen as close-cage derivatives of a hexagonalnetwork of carbon atoms.2 In its simplest form, such a hex-

agonal network terminated by hydrogen atoms is representedby a graphene sheet, Fig. 1�a�. A carbon nanotube �CNT�,more specifically a single-walled carbon nanotube�SWCNT�, can then be considered as a graphene sheet rolledinto a cylinder, where multiple concentric sheets create amultiwalled carbon nanotube �MWCNT, Fig. 1�b��. The in-troduction of five and seven member rings into the grapheneallows the formation of curved structures such as“buckyballs”3 and nanocones.4 Carbon nanofibers �CNFs�are a class of these materials that have curved graphene lay-ers or nanocones stacked to form a quasi-one-dimensionalfilament, whose internal structure can be characterized by theangle � between the graphene layers and the fiber axis �Fig.1�c��.5 One common distinction noted in the literature is be-tween the two main fiber types: “herringbone” �Fig. 1�d��,with dense conical graphene layers and large �, and “bam-boo” �Fig. 1�e��, with cylindrical cuplike graphene layers andsmall �, which is more similar to a MWCNT. However, inthe case of a true CNT � is zero.

Despite distinct differences in their internal structures,nanofibers are often called nanotubes as they can displaysimilar morphology to MWCNTs; however, their physicaland chemical properties are quite different. While nanotubesare reported to display ballistic electron transport6 and dia-mondlike tensile strength along their axis,7 nanofibers haveproven their robustness as individual, freestanding structureswith higher chemical reactivity and electron transport acrosstheir sidewalls, important for functionalization8–11 and elec-trochemical applications,8,12,13 respectively. In fact, earlystudies of highly oriented pyrolytic graphite �HOPG� andglassy carbon have shown that the edge planes of graphitehave electron transfer rates on the order of 105 times higherthan basal planes.14 Only recently has there been demon-

FIG. 1. �Color online� Schematic of carbon nanostructures: �a� single sheet of graphite, �b� CNT consisting of concentric graphene sheets, and �c� CNFcomposed of stacked graphene cones at an angle alpha with respect to the axis of the fiber. The two primary CNF structures: �d� herringbone-type CNF and�e� bamboo-type CNF. �f� Representative VACNF composed of a Ni catalyst nanoparticle at the tip and a graphitic fiber body. ��a�–�e�� Adapted withpermission from Ref. 5.

061301-2 Klein et al. J. Appl. Phys. 103, 061301 �2008�


strated control over the modulation of the internal graphiticstructure of CNFs, in turn modulating the density of edgeplane termination on the nanofiber surface.15

Vertically aligned carbon nanofibers �VACNFs, Fig.1�f��, synthesized by plasma enhanced chemical vapor depo-sition �PECVD�,5 are highly compatible with microfabrica-tion, thereby facilitating their incorporation as functionalnanostructured components of a large variety of devices.Demonstrated CNF applications include electron fieldemitters,16–19 charge and hydrogen storage media,20,21 com-posite materials,22,23 biosensors,8,9,24,25 gene deliveryarrays,26–29 synthetic membrane structures,30,31 electrochemi-cal probes,13,32,33 electrodes for neuronal interface,34,35 andscanning probe microscopy �SPM� tips.36–38 The structureand surface chemistry of the nanofibers play a crucial role inthe performance characteristics of these nanofiber-based de-vices. For many applications it is necessary to modify thesurface in order to change its properties and induce addi-tional functionality.39 The surfaces of CNFs can vary sub-stantially depending on synthesis and postsynthesis process-ing conditions such as those encountered duringmicrofabrication, and subsequent operations such as heattreatment or oxidation.40 For each of the many possible ap-plications of interest, but specifically for biological and com-posite applications, the goal is often to manipulate the sur-face chemistry in order to amplify the number of potentialattachment sites, maximize the specificity and selectivity ofadsorption processes, or just to maintain the stability of thesurface. In order to understand the benefits that surface modi-fications provide and to quantify their effectiveness, it is nec-essary to characterize the physical and chemical changes thatthey cause.

Biological applications have been one of the most sig-nificant examples of the successful implementation of carbonnanostructures, generating a swiftly growing appreciation ofthe surface functionality of these materials. Synthetic nanos-cale structures offer a particularly suitable means of interfac-ing with biological systems because they intervene at thescale where life processes proceed—the molecular level.CNFs are especially appropriate for biological interfacingbecause of their high surface area coupled with an abundanceof dangling bonds terminated in hydrogen or other functionalgroups. Consequently, CNFs have exhibited excellent speci-ficity and reversibility in binding DNA probes9 as well assuperior long term chemical stability even at elevated tem-peratures. The success of nanomaterials-enabled biology hasled to a renewed interest in surface characterization, as ex-emplified by an ample number of journal publications on theanalysis of CNF surfaces employing a variety of techniques.

This review focuses on the surface characterization andmodification of CNFs. Surface reviews on CNTs publishedthus far have been limited to methods of functionalizationonly, in which case the chemistry for graphite basal surfacesdiffers from nanofiber sidewall chemistry. Furthermore, thisreview examines the advantages of surface analysis by vari-ous techniques including scanning probe and electron mi-croscopies, infrared and electron spectroscopies, ion spec-trometry, temperature-programed desorption, and atom probe

analysis. In addition, substantial emphasis is placed on ex-ploring the recent methods of CNF surface modification forachieving a wide range of functionality.

II. SURFACE CHARACTERIZATION TECHNIQUES

This section organizes the experimental surface charac-terization work by technique, presenting a brief explanationof the method with its strengths and limitations, and illustrat-ing with several interesting examples of how the techniquehas been applied to CNF surface studies. While we cannotcover all surface techniques, we attempt to review the mostcommon, relevant, and novel techniques from the literature.

A. Scanning probe microscopy

SPM surveys small areas of surface with high lateralresolution, as opposed to techniques such as x-ray photoelec-tron spectroscopy �XPS� and infrared �IR� spectroscopy,which sample relatively large surface areas and yield meanvalues for surface properties. SPM techniques, includingscanning tunneling microscopy �STM� and atomic force mi-croscopy �AFM�, are capable of providing high surface sen-sitivity with atomic-level resolution; however, the yield ofacceptable images is low and image interpretation can bedifficult. When performing high-resolution scans, large num-bers of images must be taken to ensure the data is represen-tative.

1. Scanning tunneling microscopy

STM can map the surfaces of electrically conductive ma-terials with sub-Angstrom vertical resolution. Paredes et al.used STM observations to address the atomic-scale reorgani-zation of the CNF surface after oxidation.41,42 Figure 2shows STM images of a nanofiber surface before and after a5 min oxygen plasma treatment, from which they drew twomain conclusions about the surface structure. First, the un-treated CNF surfaces, Fig. 2�a�, displayed basal graphitecharacteristics on a local scale of several atom lengths, butlacked long range atomic order. Second, oxygen plasmatreatment resulted in the annihilation of the short-range basalgraphitic order �Fig. 2�b��, leading to a significant and uni-form increase in the atomic-scale disorder of the CNF sur-face. The results agree with previous studies of oxygenplasma etching of HOPG surfaces resulting in atomic-scaledefects.43 These atomic defects introduced in the plasma en-vironment ultimately led to the addition of functional oxygengroups to the CNF surface and may be resolved by atomicforce microscopy, XPS, IR, or other methods.

Surface graphitization can be improved by postgrowthheat treatment. Paredes et al. monitored the graphitizationprocess of CNF surfaces at various annealing temperaturesby using STM imaging.44 These images provided clear,atomic-sale evidence for the evolution of the CNF surfaceconsistent with spatially averaged characterization resultsobtained by x-ray diffraction �XRD� and Raman spectros-copy. In particular, STM revealed a nanoscale surface trans-formation from platelet morphology to larger ordered areasand finally to atomically flat terraces on faceted nanofibers at2800 °C. This corresponded in the atomic-scale images to a



conversion from small, highly defective crystallites coalesc-ing to form larger, defective crystallites and finally to truegraphitic domains in many areas. Endo et al. also observedsimilar graphitization results using STM imaging.45 AnotherSTM study by Yoon et al. found that graphitized plateletCNFs are actually assembled of carbon nano-rod units,where the edges of the graphene layers formed closed caps.46

2. Atomic force microscopy

The AFM has the inherent advantage in that it can gen-erate high lateral resolution images with superb z-height dis-crimination from all types of surfaces—even those that arewet or insulating. Furthermore, the AFM exhibits sensitivityto chemical changes via molecular recognition and frictionimaging, otherwise known as chemical force microscopy.Recently, an AFM-based approach was developed to resolvehydrophilic oxygen group regions on carbon surfaces.43 Thismethod is based on the detection of phase changes in thenoncontact tapping mode of the AFM. Paredes et al. appliedthis approach to detect and map the oxygen functionalizedcontent of CNF surfaces.42 However, phase image contrast isa relative measure within a given image, which limitssample-to-sample comparisons. Paredes et al. avoided thisissue by providing a common reference for evaluation of twonanofiber samples by placing them on a HOPG substrate.

Shown in Fig. 3 are line profiles from the noncontact tappingmode phase images of a CNF before and after oxygenplasma treatment.

Abrupt changes in surface topography also lead to phasecontrast. In the aforementioned case, evidence of this effectwas manifest as a large change in phase at the edge of thenanofiber. Since such effects were only of topographical ori-gin and were not related to the surface properties, they weredisregarded. Thus, in this case, comparison was only madebetween the top of the nanofiber �where it is locally atomi-cally flat� and the HOPG surface, indicated by arrows in Fig.3. Typical phase profiles implied that the phase of the un-treated CNF was very similar �within 1°� to that of theHOPG substrate, while the phase of the plasma treated CNFwas consistently 3° higher than the HOPG substrate. This 3°shift had been previously reported for oxygenated planarHOPG surfaces as well.43

B. Infrared spectroscopy

IR spectroscopy, with the addition of the interferometer,is a robust and easy method for characterization of organicsurfaces. Surface sensitive IR techniques can probe depths ofa few centimeters to as shallow as 100 nm below the surface.Since CNFs are generally on the order of 20–200 nm indiameter, the spectra in this case convey “bulk” chemicalinformation. However, this can be useful to distinguishwhich surface groups are present. For example, Ros et al.compared the surface structure of untreated bamboo-type toherringbone-type CNFs using transmission IR.47 They hy-pothesized that herringbone fibers, in which the sidewallsterminate with a greater number of reactive graphitic edges,would be more susceptible to surface oxidation than bambooCNFs. By keeping the transmission levels from each samplethe same, the two fiber types are compared in Fig. 4. Overall,they concluded that the CNFs have a defect-rich structureand that carbon-hydrogen bonds are present on herringboneas well as bamboo fibers �stretching at 3012, 2947, 2917, and

FIG. 2. Atomic-scale STM images of a CNF surface �a� before and �b� afteroxygen plasma treatment. Reprinted with permission from Ref. 42, Copy-right 2003 American Chemical Society.

FIG. 3. Typical noncontact tapping mode line profile of a phase image takenin the direction perpendicular to the axis of �a� an untreated “fresh” fiber onHOPG and �b� a fiber exposed to 5 min of oxygen plasma on HOPG. Re-printed in part with permission from Ref. 42, Copyright 2003 AmericanChemical Society.



2846 cm−1�. However, only herringbone fibers showed evi-dence of carboxyl groups �1717-1712 stretch� and a loss ofmass as CO2 when annealed.

Additionally, by comparing spectra from before and afterprocessing treatments the change in surface structure can beevaluated. In many cases the surface of the fiber is oxidizedto create reactive carboxylic acid sites useful for furtherfunctionalization. Ros et al. surveyed several methods of sur-face oxidation and found that a mixture of concentrated nitricand sulfuric acids proved to be the most effective method forcreating oxygen-containing surface groups.48 Furthermore,the IR spectra revealed that these surface groups occurred atdefect sites on the fiber surface and that oxidation proceedsfirst from carbonyl groups and other oxides to carboxy andcarboxylic anhydride groups.

Ros et al. also investigated the immobilization �by cova-lent bonding� of metal complexes onto CNFs by first treatingthe oxidized fibers in anthranilic acid.49 The success of thistreatment is demonstrated by the IR spectra in Fig. 5. Afteroxidation, an additional peak appears at 1724 cm−1. Thisband originates from the CvO stretching vibration of car-boxyl and/or carbonyl groups. In the spectrum of CNF-AA,the carboxyl CvO stretching vibration at 1724 cm−1 ofCNF-OX all but disappeared, indicating that an amide bondbetween the carboxyl groups on the CNFs and the aminegroups of anthranilic acid was formed. Additionally, the aro-matic CuH bending mode at 746 cm−1 �signifying four ad-jacent H atoms� occurring in the spectrum of CNF-AA is agood indication for the presence of anthranilic acid on thesurface of the CNFs.

Li et al. derivatized CNFs with concentrated nitric acidat 140 °C for 4 h and then acylated the surface-oxidizedfibers using thionyl chloride for 24 h.50 Following acylationvarious amine compounds were then surface bound to thefibers. IR results showed that the graphitic backbone CvCstretching peaks remain unaffected by the nitric acid treat-ment. IR also verified that a new CvO stretching band ap-peared following the acid oxidation, which was then reducedby the acylation and amine treatments. Unfortunately thistechnique was not useful for verifying the expected carbonylstretching band at 1616 cm−1 for amide functional groupssince it overlaps with the intense graphene CvC peak at1578 cm−1.

Han et al. used Fourier transform IR �FTIR� to verifythat oxidized multiwalled CNTs made a covalent bond �CH3

to CH2 transition� with alkanethiolate-coated goldnanoparticles,51 showing the importance of linker moleculesin creating composite materials. Additionally, they haveshown via FTIR that the organic shells can later be removedfrom the gold nanoparticles on the surfaces of the MWCNTs

FIG. 4. TEM of untreated �a� fishbone �herringbone� and �b� parallel �bam-boo� CNFs with the corresponding IR spectroscopy results below. Adoptedwith permission from Ref. 47, Copyright 2002 Blackwell Publishing.

FIG. 5. IR spectra of untreated �CNF-U�, oxidized �CNF-OX�, anthranilic-acid-treated �CNF-AA� CNFs, and of a physical mixture of anthranilic acidand CNFs �AA-phys�. Adopted with permission from Ref. 49, Copyright2002 Blackwell Publishing.



by a thermal anneal at 300 °C. Efforts by Oh et al. utilizedFTIR to demonstrate the covalent grafting of polyetherke-tones onto the surfaces of both MWCNTs and CNFs treatedwith phosphoric acid.52

C. Electron spectroscopy

There are several ways of obtaining atomic compositionand chemical bonding information from the surface. Whenx-rays of sufficient frequency �energy� interact with an atom,inner shell electrons in the atom are excited to outer, emptyorbitals, or they may be ejected from the atom completely, soionizing the atom. These electrons ejected by incident x-raysare called photoelectrons and can be detected by XPS. Like-wise, an ultraviolet excitation source can also be used andthe generated photoelectrons are measured by ultravioletphotoelectron spectroscopy �UPS�. Once a surface atomloses a photoelectron the atom desires to return to a relaxedstate; thus the inner shell “hole” left by the photoelectronwill then be filled by electrons from outer orbitals and theirexcess energy must be given off in the form of either x-rayfluorescence �XRF� or an ejected Auger electron. The ejectedAuger electrons can be analyzed via Auger electron spectros-copy �AES�.

Thus there are two competing atomic relaxation mecha-nisms. For lighter elements, the probability of Auger electronemission is significantly higher than emission of an x-rayphoton. However, collection of characteristic x-ray photonsis more efficient for measuring the composition of heavierelements. XRF from photon excitation generally is not verysurface sensitive with contributions from a few microns tomillimeters deep within the sample, and is therefore beyondthe scope of this review. X-ray emission as a result of elec-tron induced fluorescence, however, will be discussed furtherin Sec. II D 2.

The major advantages of photoelectron and Auger spec-troscopy techniques are the high surface sensitivity and theease of depth profiling. By sputtering the sample surfacewith an ion beam between successive spectra collections,depth-resolved chemical information can be obtained.

1. X-ray photoelectron spectroscopy

XPS is considered the “workhorse” of surface analysisbecause of its robustness and versatility. Using an x-raysource with energy in the keV range, the escape depth ofcore-level photoelectrons is only several atom layers deepwith a lateral spatial resolution from �30 �m to a few mil-limeters. The data produced permit the detection of all ele-ments except hydrogen and helium with a sensitivity of bet-ter than 1 at. %. One drawback of this technique is that itgives spatially averaged information across a large samplingarea, as opposed to AES which can probe areas down to10 nm with a focused scanning electron beam. Overall, thereis a plethora of literature characterizing carbon nanostruc-tures by XPS. Thus the scope of the examples presented inthis section will be limited to selected articles that focus onCNF studies, mainly surface oxidation treatment and the sur-face adsorption of metals.

Hoshi et al. utilized XPS to verify the purity of untreatedCNFs with a C 1s peak of 284.7 eV.53 Since pure CuCXPS peaks of graphite reside at �284.5 eV and C—O bond-ing is the vicinity of 288 eV, they inferred that there was lessthan 5% oxygen binding on the surface.

According to Toebes et al., the most dominant effect ofacid oxidation is a surface area increase upon opening of aninner tubular channel along the length of the fibers.54 XPSwas again used to establish the amount of oxygen on the 2-3nm sub-surface of the CNF. They demonstrated that totaloxygen content, as well as the number of acidic groups, wasa function of the type of oxidizing agent and the treatmenttime.

Lakshminarayanan et al. used XPS to take an in-depthlook at what occurs during the HNO3 oxidation process.55

They found that maximum levels of �20% oxygen werereached after only a 10 min treatment. This level remainedfairly constant until oxidation times longer than a few hourslead to carboxyl group removal and replacement by ester,anhydride, quinoid, and phenolic hydroxyl groups.

Bubert et al. investigated the formation of functional sur-face groups on CNFs from plasma treatment using variousgases, pressures, and powers.56 Figure 6 displays an XPSsputter depth profile of a CNF showing most of the oxygen isin the outer 2 nm of the sample. Additionally, from the tabu-lated results, they concluded that the highest oxygen concen-tration �NF-6� is achieved at low gas pressures �1.5 mbars�.The kind of oxygen source gas, whether O2 or CO2, did notappear to make a significant difference.

Winter et al. used a combination of XPS and transmis-sion electron microscopy �TEM� to determine the percentage

FIG. 6. XPS sputter depth profile of a CNF treated in Ar /O plasma for10 min. The table lists XPS atomic composition results for six nanofibersamples treated with different plasma conditions. The values in parenthesesat the top of the table are the mean binding energies of the fit lines, given ineV. Reprinted with permission from Ref. 56, Copyright 2002 Elsevier.



of Co and Pd nanoparticles that had adsorbed to the innercore of the nanofiber rather than the external surface follow-ing an oxidative HNO3 treatment.57 TEM elucidated the av-erage particle size and the weight percent loading of particlesto fiber, but it could not distinguish which particles werelocated in the central cavity of the fiber as opposed to thoseadsorbed to the surface. However, metal particles located onthe interior of the CNFs gave no contribution in XPS signaldue to the short mean free path of photoelectrons in graphite��2.6 nm�. From these two sets of data they were able tocalculate the weight percent of metal particles located in theinterior of the nanofiber to be 10% for Pd and 28% for Co.

Xia et al. studied the process of depositing iron oxideparticles on CNFs for additional growth of secondary fibersusing XPS.58 The specific surface area of the nanocompositewas enhanced by the growth of secondary nanofibers, and itwas possible to tune the morphology of the nanofiber-nanofiber composites by the process parameters. First, CNFswere exposed to oxygen plasma to introduce oxygen-containing functional groups. Next, the chemical vapor depo-sition �CVD� of ferrocene was carried out under oxidizingconditions, yielding nanofiber-supported iron oxide nanopar-ticles. Secondary CNFs with diameters in the range from10 to 20 nm were subsequently catalyzed by the sinteredmetallic iron nanoparticles. XPS was applied to monitor thechemical changes of the surface composition and the sinter-ing of the metallic iron particles. The XPS results shown inFig. 7 were used to derive the surface composition of boththe as-synthesized iron oxide coated fibers and the heat

treated �sintered� sample. The surface concentrations of bothFe and O decreased after heat treatment, which indicatedsevere sintering and a loss of oxygen-containing surfacegroups on the CNF.

2. Ultraviolet photoelectron spectroscopy

UPS is similar to XPS but uses ultraviolet radiation�10–50 eV� to excite valence-level photoelectrons of muchlower kinetic energies. While the penetration depth and lat-eral spatial resolution are similar to XPS, the type of infor-mation obtained is quite different because UPS gives mo-lecular orbital bonding and electronic structure information.Additionally, the spectral resolution of UPS is greater thanXPS because the photon linewidth for ultraviolet is less thanthe linewidth for x rays. The UPS technique works best witha synchrotron source, which generates a beam of high inten-sity with a narrow spectral excitation width.

In the literature there is an absence of work specificallyon UPS characterization of CNFs; however, there are a fewarticles characterizing MWCNTs by UPS. Most studies fo-cused on comparing CNTs to graphite electronic configura-tion or looking at the effects of oxidation. The early UPSwork done by Chen et al. characterized MWCNTs producedby catalytic CO disproportionation.59 They found that theCNT valence band is basically the same as graphite withminor variations. The main differences in the spectra, shownin Fig. 8, are that the intensity for the 2p-� binding energy�2.0–7.6 eV� is lower and the 2p-� ��11.5 eV� is slightlyhigher compared to graphite, an effect explained by the cur-vature of the nanotubes and �-� hybridization. Analogousresearch by Umishita et al. compared MWCNT samples pro-duced by arc discharge before and after purification.60 UPSand XPS showed many similarities between the MWCNTsand graphite; however, they attributed the variations betweenthe spectra to the nonalignment of the MWCNT samples.

Ago et al. studied the work function and density of statesof MWCNTs affected by oxidation of the surface.61 PureMWCNTs were found to have a slightly lower work functionof 4.3 eV than graphite �4.4 eV� but after oxygen plasmatreatment the density of states was affected and the work

FIG. 7. XPS results: �a� Fe 2p and �b� O 1s spectra of iron oxide coatedCNFs, with �1� as-synthesized and �2� heated to 700 °C in hydrogen andhelium �followed by exposure to air�. Summary of surface compositions issummarized in the table. Reprinted in part with permission from Ref. 58,Copyright 2003 American Chemical Society.

FIG. 8. UPS He II valence band spectra of the CNTs �solid line� and graph-ite �dotted line�. Reprinted with permission from Ref. 59, Copyright 1999American Physical Society.



function increased up to 4.8 eV as shown in Fig. 9. Theinvestigations of Lim et al. on MWCNT oxidation effectsusing UPS and TEM led to similar results.62 They found thatwith increasing oxygen exposure the carbon 2p-� states at3 eV below the Fermi level diminished while the 2p-� statesaround 6 eV significantly increased. Furthermore, annealingappeared to increase the density of states around the Fermilevel.

3. Auger electron spectroscopy

As mentioned before, a fundamental advantage of AESis that the yield of Auger electrons is highest for the lighterelements such as C, Si, N, and O. In addition, the incidentelectron beam can be focused to a fine spot giving excellentlateral spatial resolution on the order of a few tens of nanom-eters. Furthermore, since Auger electrons have lower ener-gies, their escape depth is even shallower than that of x-rayinduced photoelectrons and so they describe a region just afew atoms deep. Auger spectroscopy must therefore be per-formed in ultrahigh vacuum to maintain a clean surface, andin addition, the sample must be a good electrical conductoror else charging induced by the incident beam will result in ashift of the energies of the characteristic emission edges.

Throughout the literature AES has been used to charac-terize the surface of CNFs for many types of studies includ-ing irradiation changes and the composition of surface coat-ings. One of the most fundamental studies, by Dementjev etal., looked at the relationship between the x-ray excited CKVV Auger line shape and the layered structure of graphite.63

A plot of the angular dependence of the Auger lineshape forseveral types of carbon is displayed in Fig. 10, where theAuger spectra can be divided into two groups: first theHOPG at grazing emission, fullerene, quarterphenyl andSWCNT, and second, HOPGn at normal emission and

MWCNT. The MWCNT sample has a regular interwall spac-ing of �3.4 Å so its spectrum is very similar to HOPGn.Dementjev et al. concluded that the C KVV Auger spectra ofsingle-layer �i.e., CNTs, fullerenes� and multilayer structures�i.e., MWCNTs, the most similar to CNFs� with sp2 bondsare significantly different and may be used as a fingerprintfor the single-layer and multilayer growth modes.

Zhu et al. investigated the interaction of MWCNTs withAr+ beam irradiation using in situ AES and XPS character-ization techniques.64 Following irradiation with the Ar+

beam, the kinetic energy of the C KLL lines in x-ray inducedAES decreased 1.6 eV in the MWCNT samples, becomingmore similar to graphite, as seen in Fig. 11. They hypoth-esized that the conjugated �-bonds in the tube structure ofthe MWCNTs were destroyed by irradiation and that thenanotube was converted into an amorphous carbon rod con-taining many dangling bonds but no conjugated �-bondseven though the carbon is sp2 hybridized.

The Auger technique can also be useful for characteriz-ing species deposited on the surface of carbon nanostructuresduring growth or following postgrowth treatments. Wang etal. grew conical CNFs with Co catalysts by microwavePECVD with a high substrate bias.65 Similarly, Cui et al.grew conical CNFs from Fe-Pt catalysts using direct currentPECVD.66 Both authors used focusable, electron excitedAES to sample the chemical composition at the base and thetip of the cone and found significant levels of C, N, O, andSi. They attributed the high levels of silicon to sputtering ofthe substrate and redeposition on the sidewalls of the CNFsduring the growth process. The oxygen and nitrogen incor-poration onto the surface may have resulted from the sam-ple’s exposure to air.

An amorphous carbon layer can also be deposited undercertain growth conditions �discussed further in Sec. III C 1�.The work of Teo et al. looked at the effect of the gas ratioused for CNF growth on the deposition of noncatalytic car-bon on the substrate67 and fiber sidewalls.68 They found thatthe CNFs contain well-ordered graphitic carbon in contrastto the amorphous carbon by-product on the substrate. Byadjusting gas ratios during PECVD, the fibers range fromcylindrical to conical depending on the amount of carbonsidewall deposition, as conveyed in Figs. 12�a� and 12�b�.

FIG. 9. UPS spectra of �a� HOPG, �b� purified MWCNT film, �c� air oxi-dized MWCNT film and �d� plasma oxidized MWCNT film, with He II40.8 eV. Arrows indicate pi-derived density of states. Adopted with permis-sion from Ref. 61, Copyright 1999 Elsevier.

FIG. 10. C KVV Auger spectra of HOPGn at normal emission and MWCNT,SWCNT, fullerene, quarterphenyl, and HOPG at an emission angle of 5°.Reprinted with permission from Ref. 63, Copyright 2005 Elsevier.



Auger analysis in Fig. 12�c� of the tips and bodies of cylin-drical and conical CNFs demonstrates two things: first, thatthe body of the CNF does not contain detectable amounts ofNi and second, that only the cylindrical CNF contains a sig-nificant level of N, O, and Si. Additionally, AES depth pro-filing in Fig. 12�d� reveals that gas ratios greater than 30%C2H2:NH3 resulted in an amorphous carbon-nitrogen filmdetected on the substrate. In all cases, a 5–10 nm amorphousinterfacial layer was detected on the surface of the siliconsubstrate.

Auger spectroscopy can also be useful in observing post-growth coatings deposited on the CNFs. Wang et al. em-ployed AES and electron microscopy to characterize eachstep of a process whereby the catalyst tips of the CNFs wereremoved and then the fibers were coated with a phase-changealloy GeSbTeSn.69 The AES results showed that followingheat treatment of the phase-change alloy coated fibers, thecomposition of the alloy changed, presumably due to thefaster evaporation rates of Sb and Te than Ge. To obtain thedesired end-point alloy composition, the initial depositionratio must be modified.

D. Electron microscopy

Secondary electron microscopy �SEM� is perhaps themost frequently used method of characterizing the morpho-logical structure and topography of a sample. TEM and scan-ning transmission electron microscopy �STEM� convey in-formation about the nanostructure as a whole, but can alsoshow the presence of surface layers and reveal the atomicstructure of the surface interface by high-resolution TEM�HRTEM�. A common companion tool to both the SEM andthe TEM is an energy dispersive x-ray �EDX� spectrometer,which readily gives the elemental composition of the sampleand can also be useful in generating elemental maps. Elec-tron energy loss spectroscopy �EELS� is a valuable analyticaltool both for determining the composition of TEM specimensand for providing information about the valence state andelectronic structure of the material under examination. The

FIG. 11. TEM images of MWCNT samples �a� before and �b� after irradia-tion. �c� AES carbon KLL spectra from the nanotube samples is shown �A�before irradiation, �B� after 30 min sputter time, �C� after 210 min sputtertime, and �D� graphite carbon. Adopted with permission from Ref. 64,Copyright 1999 Elsevier.

FIG. 12. SEM images of cylindrical CNFs deposited using �a� 20%C2H2:NH3 gas ratio and �b� conical CNFs deposited using 75% C2H2:NH3

gas ratio. Auger chemical composition analysis is presented in �c�, where �1�is from the head of the cylindrical CNF in �a�, �2� is from the body of theconical CNF in �a�, �3� is from the head of the conical CNF in �b�, and �4�is from the body of the conical CNF in �b�. �d� Summary of Auger depthprofiles of the substrate surface at various C2H2:NH3 ratios. Adopted withpermission from Ref. 68.



atom’s response to electron bombardment is similar to itsrelaxation from UV or x-ray stimulation, discussed in Sec.II C. There are still two main competing relaxation mecha-nisms: characteristic x-ray emission and Auger electron ejec-tion, although for higher atomic number materials relaxationcan also efficiently occur by means of nonradiative mecha-nisms. Backscattered electron imaging �BSE� is generallynot very useful for surface analysis of CNFs because theescape depths of backscattered electrons are on the order of1 �m or greater, and is therefore not discussed here.

1. Scanning electron microscopy

Although scanning electron microscopy is the mostwidely used surface imaging technique, the depth fromwhich the relevant secondary electrons typically escape �usu-ally ranging from 5 to 50 nm deep depending on the mate-rial� results in the image containing both surface and bulkinformation. Image contrast and brightness can also be am-biguous and not quantitatively topographical; edges are oftenhighlighted and surface charging can result in large fluctua-tions in signal level as well as in distortions of the scanraster. Nevertheless, there are numerous examples of the useof SEM for studying surface morphology and process controlstudies related to CNFs. For the sake of brevity, only twostudies are described.

For example, the work of Xia et al., mentioned earlier inSec. II C 1, shows a morphological change in the fiber sur-face, depicted in Fig. 13.58 This morphological transitionfrom a smooth fiber in Fig. 13�a� to one with protrusionsfrom the surface in Fig. 13�b�, agrees with the processingsteps taken to grow secondary fibers from iron oxide par-ticles deposited on the oxidized CNFs. As a result the surfacearea of the fibers was significantly increased.

Similarly, Klein et al. used SEM to document a changein surface morphology of VACNFs with respect to catalystcomposition.70 Small levels of Cu in Ni-rich alloy catalystsresulted in split particles and branching nanofibers �Fig.14�a��; however, as the alloy composition became Cu-rich,the nanofibers became more conical as seen in Figs. 14�b�and 14�c�.

2. Energy dispersive x-ray microanalysis

EDX measures the energies of the characteristic x-raysgenerated from ionizations induced within the specimen inan electron microscope. Each element emits a unique finger-print of x-ray energies related to the difference in bindingenergies of the electron shells involved in the relaxation pro-

cess. A semiconductor diode detects each incoming photonand also measures its energy to provide a histogram displaydepicting the emitted x-ray emission spectrum from the irra-diated area of the sample. This spectrum can then be ana-lyzed to identify the elements present �“qualitative analysis”�and to determine the chemical composition of the material�“quantitative analysis”�. Therefore, EDX is a tool often usedalong with electron microscopy imaging to give complemen-tary chemical information. Depending on the energy of theelectron beam and the density and chemistry of the specimenthe lateral spatial resolution can range from micrometers totens of nanometers, while the depth resolution is a fraction ofthe incident beam range and so is on the scale of hundreds ofnanometers or less.

For example, the study by Klein et al., mentioned in theprevious section, investigated catalyst effects on CNF mor-phology using EDX to determine the composition of the fi-bers and location of catalyst material.70 From the EDX infor-mation, shown in Fig. 15, it was noted that the copper fromthe Cu-Ni catalyst particles segregated out of alloy phase toreside at the base of the nanofiber. Furthermore, they found

FIG. 13. SEM images of plasma-treated CNFs �a� before and �b� after a5 min growth of secondary nanofibers. Reprinted in part with permissionfrom Ref. 58, Copyright 2003 American Chemical Society.

FIG. 14. SEM images of CNFs grown from Cu-Ni alloys with �a� 81% Ni,�b� 39% Ni, and �c� 20% Ni. Adopted with permission from Ref. 70, Copy-right 2005 Elsevier.



that the fiber itself was encapsulated in a thick in situ depos-ited silicon-rich coating. These data lead the authors to con-clude that Cu-rich alloys were not highly active catalysts forcarbon growth, but rather the Ni portion segregated to cata-lyze a very thin CNF, which served as a scaffold for thecondensation of predominantly silicon species from theplasma.

In a second example, Weng et al. reported significantlyenhanced field emission �FE� properties from VACNFs on aSi substrate treated briefly with an argon plasma.71 Analysisby electron microscopy and EDX, shown in Fig. 16, suggeststhat a structural transformation of the fibers is a result of acosputtering/deposition process by energetic plasma ions, re-vealed by the Si counts increasing with longer Ar plasmatreatment. They attributed the field emission enhancement tothe combined effect of additional Si /C layer coverage, cata-lytic nanoparticle removal, and the physical sharpening ofthe CNFs tips. However, with longer treatment times��5 min�, structural damage effects were seen to decreasethe effectiveness of the plasma treatment as conveyed by theFE measurements in Fig. 16�c�.

Hang et al. found EDX mapping useful in their searchfor the best nanocarbon material for Fe /C composite air bat-tery anodes.21 Hollow- and herringbone-type CNFs exhibitedfavorable results, including improved conductivity andhigher redox currents of the Fe /C electrode. EDX revealedthat on the Fe/nanofiber surface, iron was more dispersedthan on Fe/graphite after cycling. They concluded that such ahigh dispersion of iron on nanofiber surfaces may improvethe electrochemical behavior of the iron redox species.

3. Transmission electron microscopy

The surface as well as internal structure of CNFs can beanalyzed using STEM or TEM. Both real space �“image”�and reciprocal space �“diffraction”� data, together withchemical and electronic analytical information �derived fromEDX or EELS�, can be obtained from the same nanoscalearea. The TEM and STEM therefore provide a wide and deeprange of data about the specimen of interest.

For instance, a study on the nanofiber structural effectsof different catalysts �Ni versus Pd� was performed by Omi-nami et al.72 It can be seen from the STEM images in Fig. 17that the Ni-catalyzed fibers have graphitic planes ending onsidewalls of the fiber in a herringbone fashion, whereas thePd-catalyzed fibers have multiwalled nanotubelike layers onthe outer walls of the fiber. Ominami et al. attributed theincreased electrical conductivity along the length of the Pd-catalyzed CNFs to this improvement in the graphitic struc-ture and lower interface resistance with the substrate.

Similarly, Naguib et al. used HRTEM to analyze thenanofiber wall structure in order to compare processingmethods for biological applications.40 They reported thatheat treatment of CNFs at 3000 °C formed graphitic loopson their surface �Fig. 18�a��, eliminating chemically activedangling bonds and leading to hydrophobic behavior. How-ever, CNFs that were pyrolytically stripped in carbon dioxideto remove polyaromatic hydrocarbon layers from their outersurface had significantly more disordered surfaces �Fig.18�b��, and exhibited hydrophilic behavior. As a result ofimproved wetting and increased protein binding capacity ofthe nanofiber surface, Naguib et al. concluded that antibodycoatings were more successful on the pyrolytically strippedCNFs.

The work by Han et al., previously described in the IRspectroscopy section,51 is also good example of the utility ofTEM for surface characterization. TEM imaging, such as in

FIG. 15. �Color online� EDX line scan showing the elemental composition along the length of a conical nanofiber �grown from Cu80Ni20 catalyst� shown inthe SEM image on the right. Reprinted with permission from Ref. 70, Copyright 1999 Elsevier.

FIG. 16. �Color online� EDX spectra �a� from the body of an argon treatedfiber �b�. FE measurements �c� of typical J-E curves for VACNFs treatedwith Ar plasma for various times. Adopted with permission from Ref. 71.



Fig. 19�a�, can give an idea of the degree of functional sitesavailable by the dispersion of gold nanoparticles on the sur-face of the fibers. Additionally, the HRTEM image in Fig.19�b� shows a closeup of how the graphitic planes are termi-nated by a gold nanoparticle.

4. Electron energy loss spectroscopy

EELS measures the energy spectrum of the electronbeam transmitted through the sample in the TEM or STEM.The spectrum contains chemical data, which are complemen-tary to those derived from EDX, but with much higher sen-sitivity for the lower atomic number elements �Z�20�. Inaddition, other phenomena such as plasmon excitations, ion-ization edge fine structure, and extended fine structure effectsprovide a detailed look at valence, bond structures, radialdistribution functions, and other descriptors of microstruc-ture on the atomic scale.

Hollow CNF wall structure investigations by Ye et al.revealed that the exterior surface was covered by a 5 Å hair-like layer in the TEM micorgraphs.73 Further investigationby EELS suggested that the “hairs” were functional groupscontaining oxygen and carbon, thus corroborating the hydro-philic behavior of the as-synthesized CNFs.

Chen et al. developed a method for synthesizing boronnitride �BN� coatings on the surface of CNFs without dam-aging the graphitic walls of the fiber.74 A reaction betweenboric acid and ammonia was initiated on the surface of the

fibers to form the coatings. HRTEM and EELS were utilizedto characterize the BN coating quality, thickness, and inter-face with the underlying CNF. An EELS chemical line scanis shown in Fig. 20�a�, along with a TEM image �Fig. 20�b��

FIG. 17. STEM images and cartoons of the nanofiber outer wall structure:��a� and �b�� using a Ni catalyst and ��c� and �d�� using a Pd catalyst. Thewhite arrows in the STEM images point towards the catalyst particle at thetip of the fiber. Adopted with permission from Ref. 72.

FIG. 18. HRTEM images showing the surface morphology differences be-tween �a� heat treated and �b� pyrolytically stripped, hollow CNFs. Adoptedwith permission from Ref. 40, Copyright 2005 Institute of Physics.

FIG. 19. �a� TEM image of Au-nanoparticle CNF composites after thermalactivation at 300 °C. �b� HRTEM image showing a single nanoparticle onthe CNF surface. Reprinted with permission from Ref. 51, Copyright 2004American Chemical Society.



illustrating there was about �20 nm thick BN layer coveringthe surface. In addition, they found that the surface structureof the CNF significantly influenced the morphology of theBN coating. If the surface of the hollow CNFs was relativelyfree of defects �i.e., dangling bonds�, then highly crystallizedBN sheaths would encapsulate the CNFs. However, if thesurface of the CNF was disordered due to pyrolytic stripping,for instance, then a thinner polycrystalline BN sheath wasproduced.

E. Secondary ion mass spectrometry

Secondary ion mass spectrometry �SIMS� is a powerfulcharacterization tool with outstanding surface sensitivity onthe scale of a few atomic layers. SIMS also provides detec-tion limits surpassing those attainable with XPS. The draw-backs of SIMS include sample charging and the complexityof the instrumentation and data analysis. In SIMS, energetic,heavy ions such as Ar+ are used to decompose the surfaceinto atoms or molecules. A small percentage ��10% � of thesurface fragments are ejected as charged particles or “sec-ondary ions,” which can be sorted on the basis of theircharge to mass ratio or their velocity. There are two types ofSIMS experiments, dynamic SIMS �D-SIMS� and staticSIMS �S-SIMS�.

1. Dynamic SIMS

In dynamic SIMS, a continuous, high-flux stream of pri-mary ions having an energy of 1–20 keV bombards the sur-face and fragments it as much as possible to maximize thegeneration of charged atomic species. Additionally, this in-tense bombardment serves a second purpose of eroding�sputtering� the sample’s surface for elemental depth profileinformation. D-SIMS has ppt to ppm sensitivity with about a2 nm depth resolution and 50 nm lateral resolution. For thisreason, SIMS is often used for doping experiments or todetect trace level elements. Uniquely, all elements, includinghydrogen, can be detected with this method and isotopes canalso be differentiated.

The work of Khare et al. appears to be the only applica-tion of D-SIMS for the surface analysis of carbonnanomaterials.75 In this work SWCNTs were irradiated byhydrogen ions and the effects evaluated by SIMS and FTIR,as shown in Fig. 21. The SIMS results showed that the hy-drogen concentration was much higher for the irradiatedsample than for the control. This was in agreement with theFTIR spectra, which confirmed the presence of CuHstretching modes only in the proton treated sample.

FIG. 21. �a� Surface hydrogen concentration as measured by D-SIMS and�b� FTIR spectra of a 1 MeV proton bombarded SWCNT film. Reprintedwith permission from Ref. 75, Copyright 2003 American Chemical Society.

FIG. 20. �a� EELS analysis of a BN-coated pyrolytically stripped hollowCNF and �b� HRTEM image of a heat treated BN-coated hollow CNF.Adopted with permission from Ref. 74, Copyright 2004 BlackwellPublishing.



2. Static SIMS

In contrast to dynamic SIMS, the goal in static SIMS isto maintain the molecular integrity of the surface fragmentsas much as possible. Therefore, a pulsed source is often used,the primary beam flux is reduced, and the beam is rastered togenerate larger charged sample fragments. Because many ofthe detected species are not distinctly identifiable, S-SIMS isonly qualitative. However, with this method isomers can usu-ally be distinguished because each fragment of the isomershould be measured in predictable levels in the mass spec-trum relative to a base peak.

Static SIMS has been utilized most often for character-izing linker molecules and polymer films bound to thenanofibers. For instance, Okpalugo et al. reported on surface-to-depth analysis of functionalized MWCNTs using SIMSand XPS.76 The CNTs in this work were first treated withnitric acid followed by dimethylformamide and further modi-fied by covalently bound 1-ethyl-3-�3-dimethylaminopropylcarbodiimide�. The functionalized nanotubes possessed sev-eral surface species with levels of OHu and CNu steadythroughout the depth profile, indicating a deeply penetratingchemical modification had occurred.

Chen et al. have described success with preparing polya-niline films on MWCNTs.77 Using time-of-flight S-SIMS,they determined that the primary functional group on thesurface of the modified CNTs was C6H6N. Similarly, He etal. applied S-SIMS to confirm the deposition of a carbon-fluorine polymer film onto CNFs by comparing the beforeand after spectra.78 As a final example, Shi et al. utilizedtime-of-flight S-SIMS to characterize a plasma-polymerizedpyrrole film deposited on the surface of hollow CNFs.79 Asmall amount of C6F14 was copolymerized to distinguish thepolymer coating from the hydrocarbon fiber. The results ofthe TEM and S-SIMS analysis are shown in Fig. 22. TheTEM image in Fig. 22�b� illustrates a thin pyrrole film

��7 nm� on the outer surface of the fiber and an ultrathinfilm ��1–3 nm� on the inner surface of the fiber. Since thefibers were open ended and hollow, some of the precursorgas was able to polymerize on the interior of the fiber. Themass spectra in Figs. 22�c� and 22�d� confirm that largecarbon-fluorine molecules were incorporated into the film.Judging from the size of the ionized polymer pieces, Shi etal. deduced that the polymer coating was highly branchedand cross-linked.

F. Temperature-programed desorption

Temperature-programed desorption �TPD�, also calledthermal desorption spectroscopy �TDS�, is a method of char-acterizing adsorbed surface species by heating the samplewhile under vacuum and simultaneously detecting the re-sidual gas by means of a mass spectrometer. As the tempera-ture rises, certain absorbed species will have enough energyto escape or desorb from the surface and will be detected asa rise in pressure for a particular mass component. A contin-ued rise in temperature will lead to a reduction in the amountof the species on the surface causing the pressure to dropagain, which manifests in a peak in the pressure versus timeplot. The temperature at the pressure peak maximum pro-vides information on the binding energy of the bound spe-cies, while the identity of the bound species is deduced fromits atomic mass. Additional information can be obtained ifsample weight is measured simultaneously with the desorp-tion spectra in a technique called thermogravimetricanalysis-mass spectrometry �TGA-MS�.

TPD is a commonly used technique for characterizationof surface oxygen complexes.80 These species are often cre-ated by reaction with oxidizing gases �O2, O3, and CO2� orby treatment with aqueous solutions of HNO3, H2SO4, orH2O2. Such treatments afford CNFs with surface functional-ity that can be utilized in biofunctionalization, catalyst sup-

FIG. 22. TEM images of hollow CNFs �a� before and �b� after pyrrole deposition. Time-of-flight SIMS spectra of �c� untreated and �d� polymer treated CNFs.Adopted with permission from Ref. 79.



port, or composite applications. Upon heating the oxidizedcarbon at a constant rate, the evolved gases of H2O, CO2,and CO are determined using a mass spectrometer. With anappropriate deconvolution procedure, TPD can be used tocharacterize both the type and amount of surface oxygencomplexes on CNFs. In general it is assumed that carboxylicacids and lactones result in CO2 peak only, while carboxylicanhydrides produce both CO and CO2 peaks.

Zhou et al. combined XPS and FTIR for disambiguationof TPD spectra obtained from gas and liquid-phase oxidizedCNFs.81 The TPD spectra, shown in Fig. 23, have overlap-ping peaks of desorbed carbon oxides decomposed from dif-ferent oxygen complexes. In their work they offer a decon-volution method that uses peak assignment �for similarlytreated nanofibers based on peak temperature, Tm�: including280 °C for carboxyl, 460 °C for carboxylic anhydride,520 °C for peroxide, 570 °C for hydroxyl, 620 and 720 °Cfor lactones, 660 °C for ether or carbonyl, 790 °C for car-bonyl, and 930 °C for pryrone-type structures. Other ex-ample work, by Toebes et al., used TGA-MS to demonstratethat liquid oxidation in mixtures of HNO3 and H2SO4

changes the surface by opening the inner tubelike cavities ofthe fibers.54 They also concluded that the total oxygen con-tent of the oxidized nanofibers is much higher than what canbe attributed to the surface alone, suggesting that some oxy-gen is bound 2–3 nm deep within the subsurface.

Another common application of TPD is the characteriza-tion of carbon nanostructures for hydrogen storage.20,82–84

Dillon et al. used TPD to show that hydrogen condensedinside SWCNTs under conditions that do not induce adsorp-tion within a standard mesoporous activated carbon, whichshows the potential of CNTs as effective media for hydrogenstorage.84 Chambers et al. applied TPD to confirm the pres-ence of chemisorbed hydrogen in CNFs after exposure tohydrogen at room temperature and high pressure.20 In thishydrogen storage study they found an observable differencebetween stored and recoverable volumes of hydrogen atroom temperature.

G. Atom probe

The atom probe is an instrument that combines both aprobe-aperture field ion microscope �with atomically highresolution� and a mass spectrometer �with single particle sen-sitivity�. Over the years, the atom probe has evolved to in-clude the unique capability of atomic layer-by-layer depthprofiling using time-of-flight mass analysis. The scanningatom probe �SAP� works by field evaporating surface atomsfrom the sample for mass analysis; thus for field enhance-ment purposes, protruding, high aspect ratio structures areideal. Only recently has there been atom probe study of car-bon nanostructures.85,86 Nishikawa et al. have found that car-bon specimens adsorb a large amount of hydrogen and mayserve as highly efficient reservoirs. In addition, they alsohave found that MWCNT and SWCNT samples exhibit asignificant difference in the adsorption and desorption char-acteristics of hydrogen.

Mass peaks from MWCNTs are fairly broad with a tailto the right, as seen in Fig. 24�a�, indicating that the fieldevaporated CuH cluster ions were dissociated while theions travel from the specimen surface to the ion detector,losing neutral hydrogen atoms along the way.86 The width ofthe peaks indicates the abundance of hydrogen. Clusters withodd numbers of carbon atoms are more common than even,reflecting the binding state in the specimen. After annealingthe MWCNT sample in vacuum for 10 min at 1000 °C �Fig.24�b��, the peaks become slightly sharper but the sample stillcontains large amounts of hydrogen and contaminants.

Mass peaks from a rod made of SWCNTs, measuredimmediately after introducing the sample into the vacuum�Fig. 24�c��, exhibit a wide tail to the right side of the peaks,again implying that the dissociation of CuH clusters oc-curred prior to entering the mass spectrometer and an abun-dance of hydrogen.86 However, after holding the sample inthe vacuum over a 40 h period �Fig. 24�d��, the initially highhydrogen level was significantly reduced, implying that theSWCNT specimen has considerably lower activation energyfor hydrogen adsorption and desorption as compared to theMWCNT sample. In fact, even though CVD diamond has ahigher hydrogen to carbon ratio, the CuH binding inSWCNTs proved weaker than that in diamond. Furthermore,Watanabe et al. coupled the mass spectra with FE data fromSWCNT samples to conclude that the work function of theSWCNTs is lower than that of other carbon materials andincreases two- to threefold with the removal of adsorbedhydrogen.85

FIG. 23. �Color online� Temperature-programed desorption spectra of fromCNFs treated with HNO3 for 12 h at room temperature for �a� CO and �b�CO2. The peak temperature �Tm� of the different types of surface oxygencomplexes used in fitting were 280 °C for carboxyl, 460 °C for carboxylicanhydride, 520 °C for peroxide, 570 °C for hydroxyl, 620 °C and 720 °Cfor lactones, 660 °C for ether or carbonyl, 790 °C for carbonyl, and 930 °Cfor pyrone-type structures. Reprinted with permission from Ref. 81, Copy-right 2007 Elsevier.



Improvement of the SAP technique in the future mayopen up new approaches to analyzing the properties of car-bon nanomaterials at the atomic scale, which cannot be ex-plored by conventional techniques such as AES and SIMS.

III. SURFACE MODIFICATION TECHNIQUES

Controlling the surface chemistry of CNFs is critical todefining their functionality. Whether being used for micro-fluidic or intracellular devices, the surface charge, hydropho-bicity, and chemical reactivity of CNFs can be alteredthrough both physical and chemical modifications, as de-tailed in this chapter. Contamination, poor solubility, andchemical inertness of carbon nanostructures have been ob-stacles to their use in applications ranging from biosensors tocomposite materials. Surface modification and functionaliza-tion techniques are necessary for resolving these issues. Inthe literature it can be seen that surface coatings �Sec. III Aand III B� not only improve the mechanical strength andchemical stability of CNFs but also add functionality such asvariable conductivity/electrical isolation or the ability to se-lectively activate certain regions on the surface through mi-crofabrication routes. A second method, covalent attachmentof functional groups �Sec. III C�, is commonly used to in-crease wettability, dispersiblility, and surface reactivity ofCNFs, enabling further biochemical functionalization �Sec.III D�.

A. Chemical vapor deposition of thin film coatings

Deposition of a thin film coating on the external surfaceof a CNF provides a means to impart additional propertiesand functionality to the nanofiber. These properties may befar beyond the scope of those attainable with the carbonnanostructure alone, thereby making coatings an extremelyuseful component. A diverse set of material coatings hasbeen deposited successfully on CNF substrates including di-electrics, metals, and polymers. Because of the high aspectratio and the limited line-of-site access to the CNF bases, themost common approaches to thin film coatings for nanofibersare CVD routes. This section discusses both in situ andpostsynthesis deposition of inorganic and polymer coatingsas they are applied to CNF surfaces.

1. In situ PECVD coatings and surface modification

The processes involved in the catalytic synthesis of car-bon nanostructures not only control the internal graphiticstructure,15 but also significantly affect the condition of thenanofiber surface. This is especially pronounced in thePECVD nanofiber growth process, as plasma activation ofgas species is frequently used for noncatalytic thin film depo-sition. Some degree of sidewall deposition is practically un-avoidable in PECVD processes involving high aspect ratiostructures, thus minimizing sidewall deposition requires sig-nificant effort if clean, graphitic surfaces are desired. Thereare essentially two types of thin films that get deposited on

FIG. 24. Mass spectrum of �a� MWCNT initially, �b� MWCNT after heating in vacuum to 1000 °C for 10 min, �c� SWCNT rod initially, and �d� SWCNT rodafter 40 h in vacuum. Na+, K+, and NaOH+ are adsorbed contaminants. Adopted with permission from Ref. 86, Copyright 2003 Japanese Journal of AppliedPhysics.



carbon nanostructure sidewalls: �1� carbon films from thesource gas,67,68,87,88 as mentioned in Sec. II C 3 and �2� com-pounds of redeposited substrate material, for instance silicon,which can react with the nitrogen from the ammonia etchantgas to form SixNy coatings on the surface.66,70,89

The sidewall deposition of carbon films was first re-ported by Chen et al. in a high temperature, high pressure,hot filament PECVD process.87 They observed formation ofconical structures, shown in Fig. 25, which had graphiticcarbon branches due to sidewall deposition during growth.Amorphous carbon coatings have also been realized on thenanofiber sidewalls and substrate.67,68,88 If the carbon sourcegas ratio is increased relative to the etchant gas, then carboncondenses on the walls of the fibers and substrate simulta-neously along with the catalytic vertical growth of the CNFas in Fig. 26. By changing the source/etchant gas ratio, thecone angle can be controlled.88 This effect is most pro-nounced for sparse arrays compared to dense forests ofnanofibers where geometric shielding becomes a factor.90

The second type of sidewall deposition occurs in a moreetching regime �higher NH3 flow�, in which amorphous car-bon is prevented from condensing. In this regime the sub-strate, unprotected by carbon film, is etched by the plasmaspecies and the etch products redeposit on the sidewalls ofgrowing CNFs. In the case of CNF growth on silicon sub-strates using a C2H2 /NH3 mixture it was determined thatSixNy deposits with x and y around 0.5.18,89 Thus there is adelicate balance to the gas ratios used in regard the desiredsurface condition; for CNFs without an amorphous carboncoating a C2H2:NH3 ratio of 20% or below must be used;68

however for ratios lower than this, sidewall deposition ofSixNy material becomes more favorable.

Other types of in situ surface modification include thecontrol of graphene edge termination groups or the doping ofthe graphene structure itself. In situ nitrogen doping has beenrecently reviewed by Ewels and Glerup.91 The doping can beachieved in a catalytic CVD process by incorporating anitrogen-containing gas, for example, Villalpando-Paez et al.used benzylamine �C7H9N� as part of ferrocene/ethanol/benzylamine mixture.92

2. CVD of oxides and nitrides

One of the most common coatings deposited on bothCNFs and CNTs via CVD methods is siliconoxide.24,29,32,81,93–97 Silicon oxide films deposited on CNFswill be commonly referred to here as SiOx since the filmstoichiometry following deposition on CNF surfaces has notbeen specifically quantified. Indeed, the literature regardingthe deposition of silicon dioxide on CNFs is limited, al-though SiO2 has been deposited on SWCNTs by the decom-position of silicon tetra-acetate using a PECVD technique.96

The surface coating composition was stoichiometric SiO2 asdetermined by XPS analysis of the deposited layer 2–8 nmin annular thickness.

Silicon oxide coatings deposited on CNF surfaces byPECVD have exhibited several advantageous physical prop-erties. Moon et al. coated MWCNTs with SiOx and foundthat the thin layer prevents high temperature oxidation andthermal degradation as well as improves the field emissionproperties.97 FTIR analysis of the SiOx-MWCNT interfacedetected intercalated SiuO, SiuC, and SiuOuC char-acteristic bonding.

Thicker SiOx coatings have also been shown to improvemechanical strength and electrically insulate CNFs from thesurrounding environment for electrochemical probe or bio-sensing applications.24,32 Freestanding, hollow SiOx nan-

FIG. 25. �a� Low magnification TEM image of a conical CNT showingnumerous graphite branches grown around the main tube; �b� a closer viewof the tip where the hollow tube and catalyst particle decorating the tip areclearly visible. Adopted with permission from Ref. 87, Copyright 2000Springer Science and Business Media.

FIG. 26. Schematic representation of the growth of �a� a CNF using con-ventional thermal CVD, �b� a vertically aligned carbon nanostructure usingPECVD, and �c� a carbon nanocone formed due to additional precipitationof C on the outer walls during PECVD. Reprinted with permission fromRef. 88.



opipes have been produced by either partially or entirely re-moving the interior CNF scaffold by excavating the top ofthe oxide coating followed by reactive ion etching of theCNF in an O2 plasma.29,95 An example recipe for SiOx filmdeposition by PECVD includes growth at a total pressure of1 Torr under a dual gas flow of 125 cm3 /min SiH4 and125 cm3 /min NO2 at a substrate temperature of 400 °C.93

The resulting deposition rate under these conditions is�45 nm /min. Generally, the coating deposits in a conformalfashion on CNF surfaces over short PECVD growth times of1–10 min. Conformal deposition creates a smoother surfaceand retains the overall CNF geometry. In addition, the depo-sition rate is linear, making PECVD a simple and reproduc-ible process. However, for a longer silicon dioxide deposi-tion time, the final shape of the coated CNF diverges fromthe template CNF. Fowlkes et al. attributed this shape changeto the mobility of deposited species.93 The high growth tem-perature �T�400 °C�, coupled with plasma excitation, im-parts energy to deposited species, which drives their subse-quent thermal migration on the evolving coating surface; thesurface free energy is minimized as a result. Thick0.5–1 �m columnar SiOx coatings form on the exterior ofthe tapered, underlying CNF. In some instances, to minimizesurface free energy, clustered CNFs in close proximity ulti-mately merge into a single, vertical SiOx pillar for exces-sively long film growth times �t�30 min, thickness�1000 nm�.93

In this way, predominantly hydrophobic CNFs are con-verted to hydrophilic structures by coating the CNF surfaceswith SiOx.

94 Moreover, the deposition of silicon oxide onCNF surfaces enables silane-based chemical functionaliza-tion strategies, providing a link to biorelated applicationssuch as turning the CNFs into active arrays for deliveringbiomolecules or into passive arrays where biological speciesare delivered to the CNFs.

One of the early reports of successful thermal CVD coat-ing of carbon nanostructures was by Chen et al. who coatedhollow CNFs with boron nitride.74 They used a two-step pro-cess in which they initially infiltrated the CNFs with a satu-rated boric acid solution and subsequently nitridized the fi-bers in an ammonia ambient at 1100–1200 °C. Furthermore,they correlated the surface structure of the starting nanofibersthat were either pyrolytically stripped versus annealed at�3000 °C to the resultant BN coatings. They determinedthat the annealed fibers had a more ordered surface and re-sulted in BN coatings with better crystallinity.

Xia et al. introduced a process in which a homogenousdistribution of secondary CNFs could be grown on CNFs viaa totally gas-phase process.58 The initial step of the processwas an oxygen plasma treatment, which oxygen-functionalized the primary CNFs. Subsequently, a ferrocene-oxygen CVD process was performed and the resultant filmwas sintered to form FeOx nanoparticles on the primary CNFsurface. Finally, a uniform distribution of secondary “nano-branches” of CNFs was catalytically grown by thermal CVDfrom the FeOx nanoparticles in a reducing H2/cylcohexaneatmosphere.

As a final example, Abdi et al. recently coated CNFswith TiO2 using an atmospheric pressure CVD process.98

While the particular CVD processing details were not di-vulged, the specific application they were developing was aFE lithography source in which they used an elaborate pro-cess: �1� initially CNFs were grown using Ni catalysts, �2�they subsequently coated the CNFs with a conformal CVDTiO2 film, �3� then they polished the tips of the TiO2-coatedCNFs, and finally �4� etched the emancipated tips in an oxy-gen plasma. The coated and etched nanofibers were success-fully used to expose a line in an electron beam resist.

3. CVD of polymers

Several groups have also coated CNTs and CNFs withpolymer thin films using CVD routes. Shi et al. has used aplasma-induced polymerization process coupled with a flu-idized bed reactor to coat CNTs with ultrathin polymerfilms.79 Their initial work demonstrated pyrrole coatings onhollow CNFs, where careful control of the plasma param-eters enabled them to deposit coating thicknesses in therange of 2–7 nm. Later using a similar plasma polymeriza-tion process they coated VACNFs with a carbon-flourinepolymer.78 Dhindsa et al. uniformly coated patterned VAC-NFs with parylene and compared their electrowetting char-acteristics to aluminum oxide coated nanofibers described inSec. III A 4.99 Figure 27�a� shows a SEM image of a pat-terned CNF template which has been uniformly coated withparylene.

4. Atomic layer deposition

Atomic layer deposition �ALD� occurs similarly to stan-dard CVD, except that in an ALD process the CVD reactionis broken into two half-reactions, keeping the precursor ma-terials separate throughout the process. ALD has been usedby a few research groups to coat CNTs and CNFs.99–101 ALDis particularly suited for coating large area and high aspectratio structures because it allows atomic layer thickness con-trol over an entire surface by a sequence of self-limitingreactions. Lee et al. was first to report coating CNFs viaALD as they successfully coated bamboo-type nanofiberswith Al2O3 using trimethylaluminum and distilled water.101

They demonstrated uniform and controlled growth of amor-phous Al2O3 on the outside of the fibers and also demon-strated that the interior cavities of the fibers could also becoated to varying degrees. Later, Herrmann et al. also coatedCNTs with Al2O3 and Al2O3 /W bilayer via ALD.100 TheAl2O3 was deposited using a similar process to that of Lee et

FIG. 27. Scanning electron images of �a� parylene-coated and �b� Al2O3

VACNFs. Reprinted in part with permission from Ref. 99, Copyright 2006American Chemical Society.



al. and the tungsten ALD layer was accomplished by alter-nating cycles of WF6 and Si2H6. In both cases, exceptionalthickness control and uniformity were demonstrated and ap-plications of the coaxial CNT /Al2O3 /W /Al2O3 structure,such as a miniature patch clamp device, were suggested. Fur-thermore, the Al2O3 coating was functionalized with a hy-drophobic monolayer to demonstrate the utility of the ALDcoating for subsequent surface functionalization. These dif-ferent functionalization approaches are schematically illus-trated in Fig. 28 along with a TEM image of aCNT /Al2O3 /W /Al2O3 composite structure. As a last ex-ample, Dhindsa et al. also coated arrays of VACNFs with anALD Al2O3 layer and demonstrated controlled and reversibleelectrowetting on the coated CNF scaffold, shown in Fig.27�b�.99 Contact angle measurements and capacitance mea-surements were performed to illustrate the electrowetting be-havior.

B. Electro- or electroless plating

The advent of electrically addressable VACNF materialhas enabled a multitude of electrochemical modificationstrategies. Guillorn et al. utilized the electrodeposition ofgold to decorate and thereby validate that only the extremetips of sidewall-passivated nanofibers were electroactive, andthat the sidewalls and underlying interconnect structureswere adequately passivated �by methods mentioned in Sec.

III A 2�.32 Melechko et al. demonstrated the electrodeposi-tion of gold upon VACNF material that was deeply recessedinto a nanopipe structure.95 In this process, arrays of highaspect ratio tubular cavities were produced by coating VAC-NFs with silicon oxide and subsequent etching the carbon-VACNF material from within the pipe. Electrodeposition ofgold within the pipe demonstrated that the interior of thenanostructure was wetted and viable for electrochemicalmethods.

Electrochemical deposition of Ni has been demonstratedon the surface of herringbone-type carbon material to pro-duce a vertical nanotube heterojunction.102 Ni electrodeposi-tion was conducted in 0.10M sulfuric acid containing 0.10MNi�NO3�2 at −0.7 V versus a silver quasireference electrode.Ye et al. have employed the electrodeposition of molybde-num oxide onto aligned MWCNTs for the electrocatalyticreduction of bromate.103 Ngo et al. demonstrated that coppercould be preferentially electrodeposited in the interstitialspaces between vertically aligned MWCNTs on an addressedsubstrate, thereby providing mechanical support and im-proved heat dissipation for a MWCNT heat transfermaterial.104 “Gap filling” of the Cu was achieved by passi-vating the CNF material with polyethyleneglycol �PEG� dur-ing the electrodeposition of the metal.

Extremely high surface area metallic electrodes havebeen demonstrated by Metz et al. using electroless deposi-tion of gold onto herringbone-type VACNFs as shown in Fig.29.105 Exposed graphite planes of the vertical carbon mate-rial were first functionalized with carboxylic acid sites usingphotochemical capture and deprotect of the methyl ester ofundecylinic acid under nitrogen-purged 254 nm irradiation.The resultant uCOOH sites were then decorated with elec-trolessly deposited gold by tin sensitization, silver activation,and gold deposition. Gold coverage on the nanofibers rangedfrom sparse, �20 nm nanoparticles to dense spikelike nucle-ations based on immersion time in the gold electroless bath�1–22 h�.

FIG. 28. Schematic cross-section images of functionalized nanotubes coatedvia ALD: �a� Al2O3 ALD film, �b� multilayered Al2O3 /W /Al2O3 ALD film,and �c� functionalized monolayer on an Al2O3 ALD seed layer. �d� TEMimage of a multiwalled nanotube coated via ALD with a multilayer asshown in �b�. Reprinted with permission from Ref. 100.

FIG. 29. �Color online� �a� Schematic illustration of the steps involved inthe functionalization of CNFs and subsequent procedure for electrolessdeposition. �b� Chemical transformations involved in the nanofiber modifi-cation. Reprinted with permission from Ref. 105, Copyright 2006 AmericanChemical Society.



Electropolymerization of pyrrole has been utilized byseveral groups for dramatically different applications. Chenet al. demonstrated that polypyrrole �PPy� could be elec-tropolymerized from an acidic monomer solution as a con-formal, electronically conductive film over the entire lengthof vertically aligned carbon material for potential use as highperformance electrode in rechargeable batteries.106

Nguyen-Vu et al. employed this same modification onVACNF surfaces to both increase the mechanical stability ofherringbone nanofiber electrodes as well as to improve bio-compatibility with neuronal cells.34 Fletcher et al. have alsopolymerized PPy on VACNF electrodes as a means of con-trolling the spacing and pore structure of nanofiber-basedmicrofluidic structures.107 Lastly, McKnight et al. have dem-onstrated photoresist-based templating strategies for achiev-ing localization of polypyrrole along the vertical length ofnanofiber electrodes, ultimately for application to theirnanofiber-based gene delivery methodologies.11,26

C. Chemical functionalization

Graphitic materials, when they are in their pristine stateand free of surface oxygen groups, have a hydrophobic na-ture; however, following routine activation procedures thesematerials can exhibit some degree of hydrophiliccharacter.108 The chemical stability of CNFs allow their sur-face chemistry to be controlled with variable wettability;both hydrophobic40,109,110 and hydrophilic40,48,73,111 surfaceshave been demonstrated. Chemical, biochemical, and elec-trochemical functionalizations of VACNF material are oftenpreceded by an oxidation step in which amorphous materialis removed and oxygen-containing moieties are generated onthe surface; most commonly noted are carboxyl surface ter-minus species.10,26,28,48,49,51,111,112

1. Wet etch treatment

Based on the recalcitrance of low defect-harboring CNTmaterial, early attempts with carbon nanostructure function-alization employed aggressive acidic pretreatment methods.For instance, Ros et al. surveyed several surface oxidationmethods and concluded that a mixture of concentrated nitricand sulfuric acids was the most effective method for attach-ing oxygen-containing surface groups to defect sites on thenanofiber surface.48 IR study has shown that the creation ofoxygen-containing surface groups occurs at graphitic defectsites and that oxidation process proceeds via carbonyls andother oxides to carboxy and carboxylic anhydride groups.

Sato et al. used an aggressive purification/functionalization procedure to investigate the cytotoxic fac-tors affecting cell activation on functionalized herringboneCNFs.113 First the CNFs were pyrolyzed in air then intro-duced to 6M HCl to dissolve the Ni catalyst material, fol-lowed by an anneal in vacuum and finished with ultrasoni-cation in a 3:1 mixture of H2SO4 and HNO3. This lengthytreatment effectively removed all Ni and generated carboxylsurface groups, which were then functionalized with 3-��3-cholamidopropyl� dimethylammonio�-1-propanesulfonate�CHAPS�. While the water-soluble CNFs did not affect cell

viability, they found that the functional group on the surfaceof the nanofibers was a significant cytotoxic factor affectingcell activation.

Lim et al. noted that HNO3 treatment helped to recoverthe free edges previously closed by annealing platelet-type,graphitized CNFs.114 Similarly, Li et al. derivatized CNFs bytreatment in concentrated nitric acid at 140 °C for 4 h.22,50

However, it must be conveyed that exfoliation under suchharsh conditions not only leads to oxidation but can alsodegrade the electrical and mechanical properties of CNTs.Even for herringbone/bamboo-oriented carbon, Nguyen et al.noted that acidic pretreatments often compromise the me-chanical integrity of the vertically aligned material.115 There-fore they employed spin-on glass to enhance the mechanicalstability of their VACNF samples. Subsequent work by thisgroup with electrically addressable carbon arrays has alter-natively employed an electrochemical etch in 1M NaOH�1.5 V, 1–2 min� to provide both VACNF sidewall etchingand carboxylic acid sites.25,116

For MWCNT material with a low density of sidewalldefects, acidic pretreatment can be used to selectively gener-ate uCOOH and uOH groups at nanotube tips, due to ahigher defect density and reactivity at these sites. Taft et al.employed this selectivity to site specifically functionalizeMWCNT material with different oligonucleotides at the tipsversus on the sidewalls.110

2. Plasma treatment

Early stuides by Esumi et al. on surface treatments ofmesocarbon microbeads found that oxygen plasma oxidizedthe carbon surface, resulting in the formation of acidicgroups; in contrast, nitrogen and ammonia plasma treatmentsproduced basic functionalities.117 Plasma etching has com-monly been used to provide uCOOH sites for both chang-ing the hydrophobicity of nanofiber samples and for subse-quent functionalization. He and Dai reported the use of anacetic acid plasma to provide uCOOH functionality,118

while and Fletcher et al. and McKnight et al., among others,reported the use of an oxygen plasma etch prior to furtherfunctionalization steps.10,26 Paredes et al. observed reorgani-zation of the CNF surface after oxidation treatment and con-cluded that exposure to oxygen plasma leads to a significantand uniform increase in the atomic-scale disorder of thenanofiber surface.41

Bubert et al. conducted a comparative study investigat-ing the formation of functional surface groups on CNFs fromlow pressure plasma treatment using various gases, flowrates, pressures and powers.56 From XPS analysis of thenanofiber surface they inferred that there was an �1 nmfunctionalized layer composed of mainly hydroxyl, carbonyl,and carboxyl groups. Bubert et al. also noted that after theinitial functionalization of this layer, further plasma treat-ment did not enhance the oxygen content, but rather changedslightly the proportions of the functional groups. Addition-ally, they concluded that the kind of oxygen source gas,whether O2 or CO2, did not appear to make a considerabledifference in the resultant surface composition.

Chen et al. have developed a novel approach for chemi-cal modification of aligned CNTs by carrying out radio fre-



quency glow-discharge plasma treatment, and subsequent re-actions characteristic of the plasma-induced surfacegroups.119 In the first of two reaction schemes, they reportedthe surface immobilization of polysaccharide �amino-dextran� chains onto acetaldehyde plasma-activatedMWCNTs through the formation of Schiff-base linkages,which were further stabilized by reduction with sodium cy-anoborohydride. In a second approach, the nanofiber surfaceswere first aminated using ethylenediamine plasma, followedby Schiff-base formation with aldehyde groups of periodateoxidized Dextran-FITC. Again, the immobilized materialwas reductively stabilized with sodium cyanoborohydride.

3. Photochemical functionalization

A technique that allows for the UV-induced reaction ofuH groups on the nanofiber sidewalls with alkene termi-nated molecules has been developed by Baker et al.9 In theirwork CNFs having both COOu and NH3

+ functional groupswere produced using protected forms of acid and amine ter-minated molecules �methyl ester protected undecylenic acidand tBOC-10-aminodec-1-ene, where tBOC�tert-butyloxycarbamate�. Solutions of the protected moleculeswere incubated with freshly grown CNFs and illuminatedunder nitrogen-purged 254 nm light for 16 h. Deprotectionof the tBOC protected surfaces was carried out using a treat-ment of 1:1 trifluoroacetic acid: methylene chloride for 1 h,immersed in 10% ammonium hydroxide, and rinsed with de-ionized �DI� water. Deprotection of the methyl ester pro-tected surfaces was carried out by immersion in a slurry of250 mM potassium t-butoxide in dimethyl sulfoxide for6 min. This treatment was followed by a rinsing in 0.1Mhydrochloric acid and DI water. In these reactions it has beenreported that the use of fibers immediately following growthprovides optimal results. Baker et al. attribute this to the“highly reducing, hydrogen-rich” environment of the CNFgrowth chamber.

Metz et al. have employed this same photochemicalfunctionalization technique to achieve extremely high cover-age of uCOOH upon VACNF material as demonstrated inFig. 29.105 Herringbone nanofibers with high hydrogenationwere reacted with the methyl ester of undecylenic acid using254 nm irradiation for 16–18 h. Then they converted themethyl ester groups to carboxylic acid, yielding CNFs func-tionalized with carboxylic acid groups. The resultantuCOOH sites were electrolessly plated with gold as shownin the last step of Figs. 29�a� and 29�b�. They concluded thatthis method combining photochemical oxidation of the fiberswith electroless gold deposition increased tenfold the electri-cally active surface area of the CNF template, amounting toa hundred times higher than the planar substrate.

4. Thermal treatment

Annealing at high temperatures in inert environmentshas traditionally been used to remove oxygen complexes anddefects to increase the graphitic order of CNFs.44,81,114,120

However, thermal treatment in oxidizing atmospheres hasalso been reported for removal of amorphous material andoxidation of graphitic material. Combustion of thermal CVD

carbon nanofiber products has been used by Sato et al. as apurification step,113 while Wang et al. heated a MWCNTarray to 400 °C in air in order to increase hydrophilicity viageneration of surface oxides and defect sites on the stericallyaligned material.121 Chen et al. and Naguib et al. comparedpyrolytically stripped CNFs to those that had been annealedat 3000 °C.40,74 The pyrolytically stripped CNFs possessed adisordered, turbostratic surface due to the thermal etching,whereas the high temperature annealed fibers formed ener-getically and chemically stable closed graphitic edges ontheir surface.74 Furthermore, CNFs that were pyrolyticallystripped in carbon dioxide exhibited hydrophilic behaviorwhile the elimination of chemically active dangling bonds byvacuum annealing led to hydrophobic behavior.40

5. Addition of linkers and polymers

The covalent addition of linker molecules and polymersis another way to enhance functionality. Li et al. derivatizedCNFs with concentrated nitric acid at 140 °C for 4 h andthen acylated the surface-oxidized fibers using thionyl chlo-ride for 24 h.50 After acylation, various amine compoundswere then surface bound to the fibers. In subsequent work,CNFs were functionalized to render hydrophobic or hydro-philic graphitic CNF polymer brushes.22 Following oxidationin nitric acid at 400 °C, thionyl chloride was used to acyl-chlorinate the surface, which was in turn reacted with either�4-hydroxymethyl�benzyl-2-bromopropionate �HBBP� or2-hydroxyethyl-2�-bromoprionate �HEBP� to provide surfacebound radical initiators. Atom transfer radical polymerization�ATRP� was then employed with acrylate and methacrylateesters to generate hydrophobic composite nanofiber-polymerbrush materials. Subsequent ester hydrolysis of the lattercould also be used to hydrophilize the material, demonstrat-ing tailored control of the material hydrophobicity withoutsignificant alteration of the structural integrity of the under-lying nanofiber material.

Zhong et al. similarly activated nitric acid-etched nanofi-bers with thionyl chloride for reaction with3–4�-oxydianiline �ODA� to provide pendant free aminogroups.23 Extension of the ODA linker molecule with a re-active diluent, butyl glycidyl ether, and incorporation intoepoxy resin provided nanofiber/epoxy nanocomposites withenhanced flexural strengths, exemplifying the utility of linkermolecules in creating composite materials. Wang et al. alsoemployed thionyl chloride with acid oxidized nanofibers togenerate hydroxylated nanofiber feedstock, by reaction of theacylchloride functionalized nanofibers with glycol at 120 °Cfor 48 h.122 Biodegradable poly�caprolactone� �PCL� wasthen covalently grafted onto the CNF-OH feedstock using insitu ring opening polymerization �ROP�. Oh et al. also uti-lized a covalent grafting technique to attach polyetherketonesto the surfaces of both MWCNTs and CNFs treated withphosphoric acid.52 Additional work, mentioned in Sec. II B,demonstrated the covalently bonding of oxidized MWCNTsto alkanethiolate-coated gold nanoparticles, where the or-ganic coating was later removed from the gold nanoparticle-MWCNT composites by a thermal anneal at 300 °C.51



D. Biochemical functionalization

Preparation of VACNFs for bioapplications requires ahigh level of control over their surface chemistry. Whetherbeing used to modify molecular transport or act as a platformfor biomolecule immobilization, the surface charge, hydro-phobicity, and chemical reactivity of CNFs can be tailoredthrough both chemical and physical modifications, as de-tailed in Sec. III A–III C. Most initial strategies for biomol-ecule immobilization focus on the decoration of surfaceswith prescribed densities of COOu or NH3

+ functionalgroups. This provides researchers with reasonable controlover surface charge and chemical reactivity, allowing eitherpassive adsorption or covalent cross-linking strategies to beused for biomolecule immobilization.

In addition to changing the morphology of nanofibersand nanofiber assemblies, physical modification of the CNFsurfaces via coating with gold �Sec. III B� or silicon dioxide�Sec. III A 2� can be an effective strategy for making fibersamenable to specific chemical modification strategies usingwell-established thiol and silane chemistries. However, somesensing applications benefit from using native CNF graphiticsurfaces for biomolecular immobilization. In these cases,control of surface chemistry can be achieved during growth�Sec. III A 1� or through the use of postgrowth modificationtechniques �Sec. III C�.

1. Proteins and antibodies

Treatment of CNFs with an oxygen plasma or wet etchproduces COOu groups on the surface, which can then bemodified using conventional carbodiimide strategies andcommercial reagents. As an alternative to oxygen plasmatreatment, the native uH groups that terminate grapheneedges on the fiber surface have been modified using photo-chemical and electrochemical reactions to yield COOu andamine terminated surfaces that can be modified using con-ventional cross-linking and adsorption strategies.8,9,24,105

Such techniques have been used for both DNA and proteinimmobilization. However, contaminates and materials thatredeposit on the nanofiber sidewalls can limit the degree towhich functional groups are distributed over the surfaces.

Protein attachment can provide an efficient way to evalu-ate the biologically accessible area of a sample. Baker et al.took advantage of avidin’s high affinity for the small mol-ecule biotin to quantify VACNF surface activity.9 By bindinga large molecule in solution �avidin, �50 Å diameter�, to asmaller molecule �biotin, �10 Å�, which is covalently teth-ered at an excess of binding sites on nanofiber surface, thequantity of large molecules binding to the surface from so-lution becomes limited by their packing density. Initially thebiotin was attached to the VACNFs using a linker containinga disulfide bond, thus after the biotin groups had bound onemonolayer of avidin molecules, the disulfide bond wascleaved, releasing the avidin into solution �where it can bequantitatively measured�. Although the exact size of thesurface-bound avidin molecule and the surface area of theVACNFs is not known precisely, this method gives a goodmeasure of the relative surface areas of different samples andyields an absolute measure of protein molecules per planar

area of the surface. Baker et al. stated that this method �formeasuring the active surface area using the self-terminatingadsorption of avidin onto biotin-modified VACNFs� agreedwith their DNA hybridization experiments, demonstrating aneightfold higher biologically accessible surface area as com-pared to planar substrates.

Another study by Baker et al. involved the adsorption ofcytochrome C, a redox-active protein, to treated VACNFsurfaces.8 It was found that COOu terminated surfaces werethe most favorable for protein adsorption as demonstrated bysignificant reduction peaks in cyclic voltammograms takenfrom the addressable “treated” CNF electrodes. It is also con-ceivable that photolithographic masking could be used tocarry out selective VACNF functionalization allowing for thefunctionalization of individual CNF chips with a collectionof different biofunctional molecules.

Naguib et al. have demonstrated how differences in thesurface structure of heat treated versus pyrolytically strippedCNFs significantly affect the adsorption of phycoerythrin–conjugated anti-CD3 antibody.40 The pyrolytically strippedCNFs with disordered surfaces exhibited hydrophilic behav-ior. Thus as a result of improved wetting and increased pro-tein binding capacity, Naguib et al. concluded that antibodycoatings were more prolific on the pyrolytically strippednanofibers. They also established that protein adsorptioncould be appreciably enhanced by pretreating dispersions ofCNFs with poly-L-lysine.

Protein attachment to CNFs has also been utilized for invivo work. Nguyen-Vu et al. combined the electropolymer-ization of pyrrole films with the incorporation of extracellu-lar matrix proteins, to promote the attachment of mammaliancells.34 They noted that neuronal cells only proliferated onthe VACNF arrays after coating them with collagen proteins.

2. DNA

DNA modification of VACNFs has been implementedfor both sensing and gene delivery applications. Afluorescence-based nucleic acid sensor employing bambooVACNFs was first described by Nguyen et al. in 2002.115 Thevertically aligned CNF material was embedded withinspin-on glass to provide mechanical stability to subsequentoxidative etching of the carbon material. This etching com-prised thermal treatment at 430 °C in air for 1 h followed byacidic treatment. Resultant uCOOH groups on the exposedCNFs were then used to capture both dye-labeled amine-terminated 21-mer oligonucleotides and unlabelled 11-merpeptide nucleic acid �PNA� oligos using the water solublecoupling reagents 1-ethyl-3�3-dimethyl aminopropyl� carbo-diimide hydrochloride �EDC� andN-hydroxysulfosuccinimide �sulfo-NHS�. Hybridization ofcomplementary dye-labeled oligos to the bound PNA wasverified using laser fluorescence microscopy. Li et al. ad-vanced this effort using electrically addressable VACNF ar-rays to transduce DNA complementation.25,116 A less harshoxidative treatment was employed to generate carboxylicacid binding sites for subsequent EDC chemistry whereinoxide sidewalls insulating the CNF electrodes were electro-chemically etched in 1M NaOH at 1.5 V for 1–2 min. Thistreatment both tailored the emergent length of the exposed



carbon electrode material and introduced uCOOH anduOH groups to the surface for subsequent DNA capture.Subattomole DNA detection limits were observed forcomplementary unlabeled targets using �Ru�bpy�3�2+ medi-ated guanine oxidation.

Parallel work by He and Dai demonstrated DNA hybrid-ization detection using the oxidation signal of ferrocene�FCA� on FCA-labeled target strands.118 The approach takenwas similar; however, as opposed to electrochemical oxida-tion of the electrode in NaOH, an acetic acid plasma treat-ment was used to generate uCOOH binding sites.

Taft et al. investigated the immobilization of amine-terminated oligos onto the tips and sidewalls of large-diameter CNTs templated using an aluminum oxide nano-porous membrane.110 A two part functionalization strategywas employed to provide tip-specific and sidewall-specificalteration of the vertically aligned material. Acid pretreat-ment �50% sulfuric/nitric acid, 3:1 in H2O, 1 h� introduceduCOOH groups to defect sites predominantly located at theopen tips. Amine-terminated DNA was condensed to thesesites using an organic phase cross-linking reaction �TFTU/DEAI in DMF�. Subsequently, the less-reactive sidewalls ofthe MWCNTs were functionalized by absorption of pyrene-terminated oligos. SEM imaging of these structures, follow-ing incubation with complementary oligos labeled with dif-ferent sized gold nanoparticles, indicated that there was littlecrosstalk between the two functionalization schemes, attrib-uted again to the presence of defect sites at the open tips ofthe CNT material.

In pursuit of orthogonal or sequential functionalizationof spatially separated nanofiber bundles Lee et al. used dia-zonium chemistry, individually addressable electrodes, andselective reduction of nitro to amine groups in order to dem-onstrate the functionalization of particular nanofiber bundleswith different DNA sequences.123 The graphene sidewalls ofaddressed nanotube/nanofiber arrays were first modified witharomatic nitro groups by 24 h immersion of the array in4-nitrobenzenediazonium tetrafluoroborate and sodium dode-cyl sulfate. Electrochemical reduction was then employed atdiscrete electrodes to reduce nitro groups to primary amines.Thiolized oligonucleotides were subsequently capturedat aminated electrodes using the heterobifunctionalcross-linker, sulfosuccinimidyl 4-�N-maleimidomethyl�cyclohexane-1-carboxylate �SSMCC�. Successful attachmentof distinct oligo probes at various electrodes was validatedby capture of fluorescently tagged complimentary oligo-nucleotides. In additional work by the same group, Baker etal. carried out the sequential, electrochemically addressedfunctionalization of CNFs by reducing nitrophenyl groups ondifferent electrodes and exposing the amines to sulfo-SMCCand thiol modified DNA strands.9 Throughout their work thegroup relied both on IR spectroscopy techniques and quanti-tative fluorescence measurements to confirm fiber modifica-tion and deprotection. Their measurements show that bothphotochemical and diazonium functionalization routesyielded DNA-modified CNFs exhibiting excellent specificityand reversibility in binding to DNA probes. In addition, theeightfold greater hybridization of DNA to the nanofiber

samples as compared to planar substrates demonstrated thatmost of the DNA was bound to the sidewalls of the nanofi-bers.

Using dense arrays of VACNFs, Fletcher et al. illustratedrelatively homogenous functionalization of the tips and side-walls of oxygen plasma etched samples following EDC-condensed capture of amine terminated oligos, as shown inFig. 30.10 Confocal microscopy following incubation withcomplementary dye-labeled oligos presented fluorescent re-sponse along the entire length of 4 �m tall nanofibers, puta-tively due to the presence of uCOOH and oligo capturealong the entire length of the herringbone-structured fibers.

McKnight et al. demonstrated the capture and transcrip-tional activity of larger �5081 bp� double stranded DNA se-quences on VACNF arrays.26,28 Periodic arrays of VACNFsat 5 �m pitch were oxygen plasma etched and functionalizedusing an overnight incubation of plasmid DNA and EDC inacidic buffer �MES, 2-morpholinoethanesulfonic acid, pH4.5–5.0�. Following extensive rinsing, nanofibers remainedfunctionalized with covalently bound, active full-lengthpromoter/gene sequences, as evidenced by expression offluorescent proteins encoded by these genes, following pen-etration of the nanostructured arrays into mammalian cells.Quantitative analysis of the amount and transcriptional activ-ity of tethered DNA was subsequently documented by Mannet al. using quantitative polymerase chain reaction �PCR�and in vitro transcription bioassay.27

IV. SUMMARY AND OUTLOOK

The control of material properties at the nanoscale lies atthe very crux of nanotechnology, and for the many applica-

FIG. 30. �Color online� Illustration of biomolecular functionalization of aCNF with an amine-terminated oligonucleotide, first four bases shown of5�-amino-c6-G-GGG… �courtesy of M. Fuentes-Cabrera�. Attachment uponthe nanofiber is provided by an amide bond, such as that resulting from anEDC condensation reaction, at a putative nanofiber-COOH site.



tions involving interfaces, the critical behaviors are driven bysurfaces. Detailed characterization of surface structure andcomposition are an integral part of understanding the prop-erties of CNFs and how they can be modified to enable di-verse applications. This literature review examined carbonnanostructure characterization via various microscopy, spec-troscopy, and spectrometry techniques. Examples from thecurrent literature pertaining to CNF surface studies were dis-cussed in depth. It has been shown that each surface charac-terization technique offers unique advantages and disadvan-tages and we emphasize the importance of couplingcomplementary information from both spatially averagedand high-resolution techniques whenever possible to get aclear representation of the surface structure and properties.Additionally, in situ and postsynthesis CNF functionalizationmethods were reviewed, from CVD thin film coatings to bio-molecule attachment. This control over the surface chemistryof CNFs has significantly broadened the application space.

The future challenge of surface characterization lies inunderstanding the functionality of nanomaterials. This re-quires tools that not only probe nonflat and nonperiodicstructures, but also probe the operational environment inwhich the nanomaterial functions. More specifically, for abetter understanding of CNFs and their interactions with bio-logical systems �such as proteins and nucleic acids�, there isstill a need for high-resolution imaging at the nanoscale us-ing analysis tools that allow for straightforward samplepreparation and in situ observation.

A central issue relating to CNF research involves under-standing how changes in the atomic-level structure of nanofi-bers affect their surface chemistry and their electrochemicalproperties. In particular, for CNFs, the ratio of edge planesversus basal planes and the degree of graphitic order is im-portant to obtaining a better understanding of the physicaland chemical properties of this diverse group of nanomateri-als. One important challenge that remains is the implemen-tation of controlled synthesis methods that allow for the se-lection of appropriate graphitic structure specific for thedesired function. In addition, ongoing development of newchemical functionalization methods will continue to advancethe integration of nanofibers with a variety of molecular, bio-molecular, and electrocatalytic materials.

ACKNOWLEDGMENTS

M.L.S. and A.V.M. acknowledge support from the Divi-sion Material Sciences and Engineering of the DOE Office ofScience. K.L.K., J.D.F., S.T.R., and P.D.R. acknowledge sup-port from the Center for Nanophase Materials Sciences.T.E.M. was supported in part by the National Institute forBiomedical Imaging and Bioengineering Grant No.R01EB006316. A portion of this research was conducted atthe Center for Nanophase Materials Sciences, which is spon-sored at Oak Ridge National Laboratory by the Division ofScientific User Facilities.

1M. S. Dresselhaus, Encyclopedia of Materials: Science and Technology�Elsevier Science, New York, 2001�.

2M. Terrones and H. Terrones, Philos. Trans. R. Soc. London, Ser. A 361,2789 �2003�.

3H. W. Kroto, J. R. Heath, S. C. Obrien, R. F. Curl, and R. E. Smalley,Nature �London� 318, 162 �1985�.

4A. Krishnan, E. Dujardin, M. M. J. Treacy, J. Hugdahl, S. Lynum, and T.W. Ebbesen, Nature �London� 388, 451 �1997�.

5A. V. Melechko, V. I. Merkulov, T. E. McKnight, M. A. Guillorn, K. L.Klein, D. H. Lowndes, and M. L. Simpson, J. Appl. Phys. 97, 041301�2005�.

6A. Javey, J. Guo, Q. Wang, M. Lundstrom, and H. J. Dai, Nature �London�424, 654 �2003�.

7M. F. Yu, O. Lourie, M. J. Dyer, K. Moloni, T. F. Kelly, and R. S. Ruoff,Science 287, 637 �2000�.

8S. E. Baker, P. E. Colavita, K. Y. Tse, and R. J. Hamers, Chem. Mater. 18,4415 �2006�.

9S. E. Baker, K. Y. Tse, E. Hindin, B. M. Nichols, T. L. Clare, and R. J.Hamers, Chem. Mater. 17, 4971 �2005�.

10B. L. Fletcher, T. E. McKnight, A. V. Melechko, M. L. Simpson, and M.J. Doktycz, Nanotechnology 17, 2032 �2006�.

11T. E. McKnight, C. Peeraphatdit, S. W. Jones, J. D. Fowlkes, B. L.Fletcher, K. L. Klein, A. V. Melechko, M. J. Doktycz, and M. L. Simpson,Chem. Mater. 18, 3203 �2006�.

12T. E. McKnight, A. V. Melechko, D. W. Austin, T. Sims, M. A. Guillorn,and M. L. Simpson, J. Phys. Chem. B 108, 7115 �2004�.

13T. E. McKnight, A. V. Melechko, B. L. Fletcher, S. W. Jones, D. K.Hensley, D. B. Peckys, G. D. Griffin, M. L. Simpson, and M. N. Ericson,J. Phys. Chem. B 110, 15317 �2006�.

14R. J. Rice and R. L. McCreery, Anal. Chem. 61, 1637 �1989�.15A. V. Melechko, K. L. Klein, J. D. Fowlkes, D. K. Hensley, I. A. Merku-

lov, T. E. McKnight, P. D. Rack, J. A. Horton, and M. L. Simpson, J. Appl.Phys. 102, 074314 �2007�.

16M. A. Guillorn, M. L. Simpson, G. J. Bordonaro, V. I. Merkulov, L. R.Baylor, and D. H. Lowndes, J. Vac. Sci. Technol. B 19, 573 �2001�.

17M. A. Guillorn, X. Yang, A. V. Melechko, D. K. Hensley, M. D. Hale, V.I. Merkulov, M. L. Simpson, L. R. Baylor, W. L. Gardner, and D. H.Lowndes, J. Vac. Sci. Technol. B 22, 35 �2004�.

18X. J. Yang, M. A. Guillorn, D. Austin, A. V. Melechko, H. T. Cui, H. M.Meyer, V. I. Merkulov, J. B. O. Caughman, D. H. Lowndes, and M. L.Simpson, Nano Lett. 3, 1751 �2003�.

19W. I. Milne, K. B. K. Teo, M. Mann, I. Y. Y. Bu, G. A. J. Amaratunga, N.De Jonge, M. Allioux, J. T. Oostveen, P. Legagneux, E. Minoux, L. Gan-gloff, L. Hudanski, J. P. Schnell, L. D. Dieumegard, F. Peauger, T. Wells,and M. El-Gomati, Phys. Status Solidi A 203, 1058 �2006�.

20A. Chambers, C. Park, R. T. K. Baker, and N. M. Rodriguez, J. Phys.Chem. B 102, 4253 �1998�.

21B. T. Hang, M. Eashira, I. Watanabe, S. Okada, J. I. Yamaki, S. H. Yoon,and I. Mochida, J. Power Sources 143, 256 �2005�.

22L. Li and C. M. Lukehart, Chem. Mater. 18, 94 �2006�.23W. H. Zhong, J. Li, C. M. Lukehart, and L. Y. R. Xu, Polym. Compos. 26,

128 �2005�.24S. E. Baker, K. Y. Tse, C. S. Lee, and R. J. Hamers, Diamond Relat. Mater.

15, 433 �2006�.25J. Li, H. T. Ng, A. Cassell, W. Fan, H. Chen, Q. Ye, J. Koehne, J. Han, and

M. Meyyappan, Nano Lett. 3, 597 �2003�.26T. E. McKnight, A. V. Melechko, D. K. Hensley, D. G. J. Mann, G. D.

Griffin, and M. L. Simpson, Nano Lett. 4, 1213 �2004�.27D. G. J. Mann, T. E. McKnight, A. V. Melechko, M. L. Simpson, and G.

S. Sayler, Biosens. Bioelectron. 97, 680 �2007�.28T. E. McKnight, A. V. Melechko, G. D. Griffin, M. A. Guillorn, V. I.

Merkulov, F. Serna, D. K. Hensley, M. J. Doktycz, D. H. Lowndes, andM. L. Simpson, Nanotechnology 14, 551 �2003�.

29A. V. Melechko, T. E. McKnight, M. A. Guillorn, V. I. Merkulov, B. Ilic,M. J. Doktycz, D. H. Lowndes, and M. L. Simpson, Appl. Phys. Lett. 82,976 �2003�.

30B. L. Fletcher, E. D. Hullander, A. V. Melechko, T. E. McKnight, K. L.Klein, D. K. Hensley, J. L. Morrell, M. L. Simpson, and M. J. Doktycz,Nano Lett. 4, 1809 �2004�.

31L. Zhang, A. V. Melechko, V. I. Merkulov, M. A. Guillorn, M. L. Simp-son, D. H. Lowndes, and M. J. Doktycz, Appl. Phys. Lett. 81, 135 �2002�.

32M. A. Guillorn, T. E. McKnight, A. Melechko, V. I. Merkulov, P. F. Britt,D. W. Austin, D. H. Lowndes, and M. L. Simpson, J. Appl. Phys. 91, 3824�2002�.

33T. E. McKnight, A. V. Melechko, M. A. Guillorn, V. I. Merkulov, M. J.Doktycz, C. T. Culbertson, S. C. Jacobson, D. H. Lowndes, and M. L.Simpson, J. Phys. Chem. B 107, 10722 �2003�.

34T. D. B. Nguyen-Vu, H. Chen, A. M. Cassell, R. Andrews, M. Meyyap-



http://dx.doi.org/10.1098/rsta.2003.1262

http://dx.doi.org/10.1038/318162a0

http://dx.doi.org/10.1038/41284

http://dx.doi.org/10.1063/1.1857591

http://dx.doi.org/10.1038/nature01797

http://dx.doi.org/10.1126/science.287.5453.637

http://dx.doi.org/10.1021/cm0609000

http://dx.doi.org/10.1021/cm051024d

http://dx.doi.org/10.1088/0957-4484/17/8/041

http://dx.doi.org/10.1021/cm052680g

http://dx.doi.org/10.1021/jp037987m

http://dx.doi.org/10.1021/jp056467j

http://dx.doi.org/10.1021/ac00190a010

http://dx.doi.org/10.1063/1.2786710

http://dx.doi.org/10.1063/1.2786710

http://dx.doi.org/10.1116/1.1358855

http://dx.doi.org/10.1116/1.1633768

http://dx.doi.org/10.1021/nl0346631

http://dx.doi.org/10.1002/pssa.200566101

http://dx.doi.org/10.1021/jp980114l

http://dx.doi.org/10.1021/jp980114l

http://dx.doi.org/10.1016/j.jpowsour.2004.11.044

http://dx.doi.org/10.1021/cm051720d

http://dx.doi.org/10.1002/pc.20085

http://dx.doi.org/10.1016/j.diamond.2005.08.019


http://dx.doi.org/10.1021/nl049504b

http://dx.doi.org/10.1088/0957-4484/14/5/313

http://dx.doi.org/10.1063/1.1544058


http://dx.doi.org/10.1063/1.1490142

http://dx.doi.org/10.1063/1.1448671

http://dx.doi.org/10.1021/jp034872+

pan, and J. Li, Small 2, 89 �2006�.35Z. Yu, T. E. McKnight, M. N. Ericson, A. V. Melechko, M. L. Simpson,

and B. Morrison, Nano Lett. 7, 2188 �2007�.36H. Cui, S. V. Kalinin, X. Yang, and D. H. Lowndes, Nano Lett. 4, 2157

�2004�.37M. A. Guillorn, A. V. Melechko, and M. L. Simpson, United States Patent

#7,151,256 �2005�.38Q. Ye, A. Cassell, H. Liu, K.-J. Chao, J. Han, and M. Meyyappan, Nano

Lett. 4, 1301 �2004�.39A. Hirsch, Angew. Chem., Int. Ed. 41, 1853 �2002�.40N. N. Naguib, Y. M. Mueller, P. M. Bojczuk, M. P. Rossi, P. D. Katsikis,

and Y. Gogotsi, Nanotechnology 16, 567 �2005�.41J. I. Paredes, A. Martinez-Alonso, and J. M. D. Tascon, Carbon 40, 1101

�2002�.42J. I. Paredes, A. Martinez-Alonso, and J. M. D. Tascon, Langmuir 19,

7665 �2003�.43J. I. Paredes, A. Martinez-Alonso, and J. M. D. Tascon, Langmuir 18,

4314 �2002�.44J. I. Paredes, M. Burghard, A. Martinez-Alonso, and J. M. D. Tascon,

Appl. Phys. A: Mater. Sci. Process. 80, 675 �2005�.45M. Endo, Y. A. Kim, T. Takeda, S. H. Hong, T. Matusita, T. Hayashi, and

M. S. Dresselhaus, Carbon 39, 2003 �2001�.46S. H. Yoon, S. Y. Lim, S. H. Hong, I. Mochida, B. An, and K. Yokogawa,

Carbon 42, 3087 �2004�.47T. G. Ros, A. J. van Dillen, J. W. Geus, and D. C. Koningsberger, ChemP-

hysChem 2, 209 �2002�.48T. G. Ros, A. J. van Dillen, J. W. Geus, and D. C. Koningsberger, Chem.-

Eur. J. 8, 1151 �2002�.49T. G. Ros, A. J. van Dillen, J. W. Geus, and D. C. Koningsberger, Chem.-

Eur. J. 8, 2868 �2002�.50J. Li, M. J. Vergne, E. D. Mowles, W. H. Zhong, D. M. Hercules, and C.

M. Lukehart, Carbon 43, 2883 �2005�.51L. Han, W. Wu, F. L. Kirk, J. Luo, M. M. Maye, N. N. Kariuki, Y. H. Lin,

C. M. Wang, and C. J. Zhong, Langmuir 20, 6019 �2004�.52S. J. Oh, H. J. Lee, D. K. Keum, S. W. Lee, D. H. Wang, S. Y. Park, L. S.

Tan, and J. B. Baek, Polymer 47, 1132 �2006�.53F. Hoshi, K. Tsugawa, A. Goto, T. Ishikura, S. Yamashita, M. Yumura, T.

Hirao, K. Oura, and Y. Koga, Diamond Relat. Mater. 10, 254 �2001�.54M. L. Toebes, E. M. P. van Heeswijk, J. H. Bitter, A. J. van Dillen, and K.

P. de Jong, Carbon 42, 307 �2004�.55P. V. Lakshminarayanan, H. Toghiani, and C. U. Pittman, Carbon 42, 2433

�2004�.56H. Bubert, X. Ai, S. Haiber, M. Heintze, V. Bruser, E. Pasch, W. Brandl,

and G. Marginean, Spectrochim. Acta, Part B 57, 1601 �2002�.57F. Winter, G. L. Bezemer, C. van der Spek, J. D. Meeldijk, A. J. van

Dillen, J. W. Geus, and K. P. de Jong, Carbon 43, 327 �2005�.58W. Xia, D. S. Su, A. Birkner, L. Ruppel, Y. M. Wang, C. Woll, J. Qian, C.

H. Liang, G. Marginean, W. Brandl, and M. Muhler, Chem. Mater. 17,5737 �2005�.

59P. Chen, X. Wu, X. Sun, J. Lin, W. Ji, and K. L. Tan, Phys. Rev. Lett. 82,2548 �1999�.

60K. Umishita, Y. Ochiai, K. Iwasaki, and S. Hino, Synth. Met. 121, 1159�2001�.

61H. Ago, T. Kugler, F. Cacialli, K. Petritsch, R. H. Friend, W. R. Salaneck,Y. Ono, T. Yamabe, and K. Tanaka, Synth. Met. 103, 2494 �1999�.

62S. C. Lim, C. S. Jo, H. J. Jeong, Y. M. Shin, Y. H. Lee, I. A. Samayoa, andJ. Choi, Jpn. J. Appl. Phys., Part 1 41, 5635 �2002�.

63A. P. Dementjev, K. I. Maslakova, and A. V. Naumkin, Appl. Surf. Sci.245, 128 �2005�.

64Y. F. Zhu, T. Yi, B. Zheng, and L. L. Cao, Appl. Surf. Sci. 137, 83 �1999�.65W. H. Wang, Y. T. Lin, and C. T. Kuo, Diamond Relat. Mater. 14, 907

�2005�.66H. Cui, X. Yang, H. M. Meyer, L. R. Baylor, M. L. Simpson, W. L.

Gardner, D. H. Lowndes, L. An, and J. Liu, J. Mater. Res. 20, 850 �2005�.67K. B. K. Teo, M. Chhowalla, G. A. J. Amaratunga, W. I. Milne, D. G.

Hasko, G. Pirio, P. Legagneux, F. Wyczisk, and D. Pribat, Appl. Phys.Lett. 79, 1534 �2001�.

68K. B. K. Teo, M. Chhowalla, G. A. J. Amaratunga, W. I. Milne, G. Pirio,P. Legagneux, F. Wyczisk, J. Olivier, and D. Pribat, J. Vac. Sci. Technol. B20, 116 �2002�.

69W. H. Wang, K. M. Chao, I. J. Teng, and C. T. Kuo, Surf. Coat. Technol.200, 3206 �2006�.

70K. L. Klein, A. V. Melechko, P. D. Rack, J. D. Fowlkes, H. M. Meyer, andM. L. Simpson, Carbon 43, 1857 �2005�.

71C. H. Weng, K. C. Leou, H. W. Wei, Z. Y. Juang, M. T. Wei, C. H. Tung,and C. H. Tsai, Appl. Phys. Lett. 85, 4732 �2004�.

72Y. Ominami, Q. Ngo, A. J. Austin, H. Yoong, C. Y. Yang, A. M. Cassell,B. A. Cruden, J. Li, and M. Meyyappan, Appl. Phys. Lett. 87, 233105,�2005�.

73H. H. Ye, N. Naguib, Y. Gogotsi, A. G. Yazicioglu, and C. M. Megaridis,Nanotechnology 15, 232 �2004�.

74L. L. Chen, H. H. Ye, and Y. Gogotsi, J. Am. Ceram. Soc. 87, 147 �2004�.75B. Khare, M. Meyyappan, M. H. Moore, P. Wilhite, H. Imanaka, and B.

Chen, Nano Lett. 3, 643 �2003�.76T. I. T. Okpalugo, P. Papakonstantinou, H. Murphy, J. McLaughlin, N. M.

D. Brown, and T. McNally, Fullerenes, Nanotubes, Carbon Nanostruct. 13,477 �2005�.

77W. Y. Chen, C. Y. Chen, K. Y. Hsu, C. C. Wang, and Y. C. Ling, Appl.Surf. Sci. 231–232, 845 �2004�.

78P. He, D. L. Shi, J. Lian, L. M. Wang, R. C. Ewing, W. van Ooij, W. Z. Li,and Z. F. Ren, Appl. Phys. Lett. 86, 043107 �2005�.

79D. L. Shi, J. Lian, P. He, L. M. Wang, W. J. van Ooij, M. Schulz, Y. J. Liu,and D. B. Mast, Appl. Phys. Lett. 81, 5216 �2002�.

80H. P. Boehm, Carbon 40, 145 �2002�.81J. H. Zhou, Z. J. Sui, J. Zhu, P. Li, D. Chen, Y. C. Dal, and W. K. Yuan,

Carbon 45, 785 �2007�.82P. Piseri, E. Barborini, M. Marino, P. Milani, C. Lenardi, L. Zoppi, and L.

Colombo, J. Phys. Chem. B 108, 5157 �2004�.83A. Zuttel, P. Sudan, P. Mauron, T. Kiyobayashi, C. Emmenegger, and L.

Schlapbach, Int. J. Hydrogen Energy 27, 203 �2002�.84A. C. Dillon, K. M. Jones, T. A. Bekkedahl, C. H. Kiang, D. S. Bethune,

and M. J. Heben, Nature �London� 386, 377 �1997�.85M. Watanabe, T. Yagyu, O. Nishikawa, T. Yamaguchi, N. Choi, H. Toku-

moto, and S. Nakano, Jpn. J. Appl. Phys., Part 1 41, 7469 �2002�.86O. Nishikawa, M. Watanabe, T. Murakami, T. Yagyu, and M. Taniguchi,

New Diamond Front. Carbon Technol. 13, 257 �2003�.87Y. Chen, L. P. Guo, S. Patel, and D. T. Shaw, J. Mater. Sci. 35, 5517

�2000�.88V. I. Merkulov, M. A. Guillorn, D. H. Lowndes, M. L. Simpson, and E.

Voelkl, Appl. Phys. Lett. 79, 1178 �2001�.89A. V. Melechko, T. E. McKnight, D. K. Hensley, M. A. Guillorn, A. Y.

Borisevich, V. I. Merkulov, D. H. Lowndes, and M. L. Simpson, Nano-technology 14, 1029 �2003�.

90V. I. Merkulov, A. V. Melechko, M. A. Guillorn, D. H. Lowndes, and M.L. Simpson, Appl. Phys. Lett. 80, 476 �2002�.

91C. P. Ewels and M. Glerup, J. Nanosci. Nanotechnol. 5, 1345 �2005�.92F. Villalpando-Paez, A. Zamudio, A. L. Elias, H. Son, E. B. Barros, S. G.

Chou, Y. A. Kim, H. Muramatsu, T. Hayashi, J. Kong, H. Terrones, G.Dresselhaus, M. Endo, M. Terrones, and M. S. Dresselhaus, Chem. Phys.Lett. 424, 345 �2006�.

93J. D. Fowlkes, B. L. Fletcher, E. D. Hullander, K. L. Klein, D. K. Hensley,A. V. Melechko, M. L. Simpson, and M. J. Doktycz, Nanotechnology 16,3101 �2005�.

94J. D. Fowlkes, E. D. Hullander, B. L. Fletcher, S. T. Retterer, A. V. Me-lechko, D. K. Hensley, M. L. Simpson, and M. J. Doktycz, Nanotechnol-ogy 17, 5659 �2006�.

95A. V. Melechko, T. E. McKnight, M. A. Guillorn, D. W. Austin, B. Ilic, V.I. Merkulov, M. J. Doktycz, D. H. Lowndes, and M. L. Simpson, J. Vac.Sci. Technol. B 20, 2730 �2002�.

96J. D. Whittaker, M. Brink, G. A. Husseini, M. R. Linford, and R. C. Davis,Appl. Phys. Lett. 83, 5307 �2003�.

97J. S. Moon, P. S. Alegaonkar, J. H. Han, T. Y. Lee, J. B. Yoo, and J. M.Kim, J. Appl. Phys. 100, 104303 �2006�.

98Y. Abdi, S. Mohajerzadeh, H. Hoseinzadegan, and J. Koohsorkhi, Appl.Phys. Lett. 88, 053124 �2006�.

99M. S. Dhindsa, N. R. Smith, J. Heikenfeld, P. D. Rack, J. D. Fowlkes, M.J. Doktycz, A. V. Melechko, and M. L. Simpson, Langmuir 22, 9030�2006�.

100C. F. Herrmann, F. H. Fabreguette, D. S. Finch, R. Geiss, and S. M.George, Appl. Phys. Lett. 87, 123110 �2005�.

101J. S. Lee, B. Min, K. Cho, S. Kim, J. Park, Y. T. Lee, N. S. Kim, M. S.Lee, S. O. Park, and J. T. Moon, J. Cryst. Growth 254, 443 �2003�.

102A. M. Cassell, J. Li, R. M. D. Stevens, J. E. Koehne, L. Delzeit, H. T. Ng,Q. Ye, J. Han, and M. Meyyappan, Appl. Phys. Lett. 85, 2364 �2004�.

103J. S. Ye, H. F. Cui, Y. Wen, W. D. Zhang, G. Q. Xu, and F. S. Sheu,Microchim. Acta 152, 267 �2006�.

104Q. Ngo, B. A. Cruden, A. M. Cassell, G. Sims, M. Meyyappan, J. Li, andC. Y. Yang, Nano Lett. 4, 2403 �2004�.



http://dx.doi.org/10.1002/smll.200500175

http://dx.doi.org/10.1021/nl070291a

http://dx.doi.org/10.1021/nl048740j

http://dx.doi.org/10.1021/nl049341r

http://dx.doi.org/10.1021/nl049341r

http://dx.doi.org/10.1002/1521-3773(20020603)41:11<1853::AID-ANIE1853>3.0.CO;2-N

http://dx.doi.org/10.1088/0957-4484/16/4/038

http://dx.doi.org/10.1016/S0008-6223(01)00255-X

http://dx.doi.org/10.1021/la020880q

http://dx.doi.org/10.1021/la0115280

http://dx.doi.org/10.1007/s00339-004-3109-9

http://dx.doi.org/10.1016/S0008-6223(01)00019-7

http://dx.doi.org/10.1016/j.carbon.2004.07.022

http://dx.doi.org/10.1002/1439-7641(20020215)3:2<209::AID-CPHC209>3.0.CO;2-S

http://dx.doi.org/10.1002/1439-7641(20020215)3:2<209::AID-CPHC209>3.0.CO;2-S

http://dx.doi.org/10.1002/1521-3765(20020301)8:5<1151::AID-CHEM1151>3.0.CO;2-#

http://dx.doi.org/10.1002/1521-3765(20020301)8:5<1151::AID-CHEM1151>3.0.CO;2-#

http://dx.doi.org/10.1002/1521-3765(20020703)8:13<2868::AID-CHEM2868>3.0.CO;2-C

http://dx.doi.org/10.1002/1521-3765(20020703)8:13<2868::AID-CHEM2868>3.0.CO;2-C


http://dx.doi.org/10.1021/la0497907

http://dx.doi.org/10.1016/j.polymer.2005.12.064

http://dx.doi.org/10.1016/S0925-9635(00)00476-3



http://dx.doi.org/10.1016/S0584-8547(02)00101-5


http://dx.doi.org/10.1021/cm051623k

http://dx.doi.org/10.1103/PhysRevLett.82.2548

http://dx.doi.org/10.1016/S0379-6779(00)01146-2

http://dx.doi.org/10.1016/S0379-6779(98)01062-5

http://dx.doi.org/10.1143/JJAP.41.5635

http://dx.doi.org/10.1016/j.apsusc.2004.10.002

http://dx.doi.org/10.1016/S0169-4332(98)00372-9

http://dx.doi.org/10.1016/j.diamond.2004.10.013

http://dx.doi.org/10.1557/JMR.2005.0106

http://dx.doi.org/10.1063/1.1400085

http://dx.doi.org/10.1063/1.1400085

http://dx.doi.org/10.1116/1.1428281

http://dx.doi.org/10.1016/j.surfcoat.2005.07.011


http://dx.doi.org/10.1063/1.1815062

http://dx.doi.org/10.1063/1.2137873

http://dx.doi.org/10.1088/0957-4484/15/1/042

http://dx.doi.org/10.1021/nl034058y

http://dx.doi.org/10.1081/FST-200039448

http://dx.doi.org/10.1063/1.1846957

http://dx.doi.org/10.1063/1.1527702

http://dx.doi.org/10.1016/S0008-6223(01)00165-8


http://dx.doi.org/10.1021/jp037146f

http://dx.doi.org/10.1016/S0360-3199(01)00108-2

http://dx.doi.org/10.1038/386377a0

http://dx.doi.org/10.1143/JJAP.41.7469

http://dx.doi.org/10.1023/A:1004889500361

http://dx.doi.org/10.1063/1.1395517

http://dx.doi.org/10.1088/0957-4484/14/9/318

http://dx.doi.org/10.1088/0957-4484/14/9/318

http://dx.doi.org/10.1063/1.1433905

http://dx.doi.org/10.1166/jnn.2005.304

http://dx.doi.org/10.1016/j.cplett.2006.04.074

http://dx.doi.org/10.1016/j.cplett.2006.04.074

http://dx.doi.org/10.1088/0957-4484/16/12/063

http://dx.doi.org/10.1088/0957-4484/17/22/021

http://dx.doi.org/10.1088/0957-4484/17/22/021

http://dx.doi.org/10.1116/1.1515306

http://dx.doi.org/10.1116/1.1515306

http://dx.doi.org/10.1063/1.1636267

http://dx.doi.org/10.1063/1.2384795

http://dx.doi.org/10.1063/1.2168039

http://dx.doi.org/10.1063/1.2168039

http://dx.doi.org/10.1021/la061139b

http://dx.doi.org/10.1063/1.2053358

http://dx.doi.org/10.1016/S0022-0248(03)01203-X

http://dx.doi.org/10.1063/1.1794356

http://dx.doi.org/10.1021/nl048506t

105K. M. Metz, K. Y. Tse, S. E. Baker, E. C. Landis, and R. J. Hamers,Chem. Mater. 18, 5398 �2006�.

106J. H. Chen, Z. P. Huang, D. Z. Wang, S. X. Yang, W. Z. Li, J. G. Wen,and Z. F. Ren, Synth. Met. 125, 289 �2001�.

107B. L. Fletcher, T. E. McKnight, J. D. Fowlkes, D. P. Allison, M. L.Simpson, and M. J. Doktycz, Synth. Met. 157, 282 �2007�.

108R. T. K. Baker, Encyclopedia of Materials: Science and Technology�Elsevier Science, New York, 2001�, p. 932–941.

109T. Werder, J. H. Walther, R. L. Jaffe, T. Halicioglu, F. Noca, and P.Koumoutsakos, Nano Lett. 1, 697 �2001�.

110 B. J. Taft, A. D. Lazareck, G. D. Withey, A. J. Yin, J. M. Xu, and S. O.Kelley, J. Am. Chem. Soc. 126, 12750 �2004�.

111 Y. Gogotsi, J. A. Libera, A. Guvenc-Yazicioglu, and C. M. Megaridis,Appl. Phys. Lett. 79, 1021 �2001�.

112 M. P. Rossi, H. H. Ye, Y. Gogotsi, S. Babu, P. Ndungu, and J. C. Bradley,Nano Lett. 4, 989 �2004�.

113 Y. Sato, K. Shibata, H. Kataoka, S. Ogino, F. Bunshi, A. Yokoyama, K.Tamura, T. Akasaka, M. Uo, K. Motomiya, B. Jeyadevan, R.Hatakeyama, F. Watari, and K. Tohji, Mol. Biosyst. 1, 142 �2005�.

114 S. Lim, S. H. Yoon, I. Mochida, and J. H. Chi, J. Phys. Chem. B 108,1533 �2004�.

115 C. V. Nguyen, L. Delzeit, A. M. Cassell, J. Li, J. Han, and M. Meyyap-pan, Nano Lett. 2, 1079 �2002�.

116 J. Li, J. E. Koehne, A. M. Cassell, H. Chen, H. T. Ng, Q. Ye, W. Fan, J.Han, and M. Meyyappan, Electroanalysis 17, 15 �2005�.

117 K. Esumi, M. Sugiura, T. Mori, K. Meguro, and H. Honda, Colloids Surf.19, 331 �1986�.

118 P. G. He and L. M. Dai, Chem. Commun., issue 3 �2004�, p. 348.119 Q. D. Chen, L. M. Dai, M. Gao, S. M. Huang, and A. Mau, J. Phys.

Chem. B 105, 618 �2001�.120L. J. Ci, H. W. Zhu, B. Q. Wei, C. L. Xu, J. Liang, and D. H. Wu, Mater.

Lett. 43, 291 �2000�.121K. Wang, H. A. Fishman, H. J. Dai, and J. S. Harris, Nano Lett. 6, 2043

�2006�.122K. Wang, W. W. Li, and C. Gao, J. Appl. Polym. Sci. 105, 629 �2007�.123C. S. Lee, S. E. Baker, M. S. Marcus, W. S. Yang, M. A. Eriksson, and R.

J. Hamers, Nano Lett. 4, 1713 �2004�.



http://dx.doi.org/10.1021/cm061563y

http://dx.doi.org/10.1016/S0379-6779(01)00404-0

http://dx.doi.org/10.1016/j.synthmet.2007.03.004

http://dx.doi.org/10.1021/nl015640u

http://dx.doi.org/10.1021/ja045543d

http://dx.doi.org/10.1063/1.1391228

http://dx.doi.org/10.1021/nl049688u

http://dx.doi.org/10.1021/jp036819r

http://dx.doi.org/10.1021/nl025689f

http://dx.doi.org/10.1002/elan.200403114

http://dx.doi.org/10.1021/jp003385g

http://dx.doi.org/10.1021/jp003385g

http://dx.doi.org/10.1016/S0167-577X(99)00276-1

http://dx.doi.org/10.1016/S0167-577X(99)00276-1

http://dx.doi.org/10.1021/nl061241t

http://dx.doi.org/10.1002/app.26285

http://dx.doi.org/10.1021/nl048995x

APPLIED PHYSICS REVIEWS Surface characterization and functionalization...

Documents

Transcript of APPLIED PHYSICS REVIEWS Surface characterization and functionalization...