Direct Growth of Aligned Multiwalled Carbon Nanotubes on TreatedStainless Steel Substrates

Charan Masarapu† and Bingqing Wei*,‡

Department of Electrical and Computer Engineering and Department of Mechanical Engineering,UniVersity of Delaware, Newark, Delaware 19716

ReceiVed April 16, 2007. In Final Form: June 9, 2007

Growth of aligned carbon nanotubes (CNTs) on electrically conductive substrates is promising for many applications;however, the lack of complete understanding of the substrate effects on CNT growth poses a lot of technical challenges.Here, we report the direct growth of aligned multiwalled nanotubes (MWNTs) on chemically treated stainless steel(Type 304) using a chemical vapor deposition (CVD) process. A detailed X-ray photoelectron spectroscopy (XPS)analysis has been carried out for the various treated samples in order to better understand the correlation between thesurface properties of the substrates and the MWNT growth. The XPS studies revealed that the CNTs prefer to growon the enriched surface of iron oxides obtained by the chemical treatment rather than on the passive chromium oxidefilms present on the surface of the as-received stainless steel substrates. The density and alignment of the MWNTscould therefore be controlled by tuning the ratio of the iron oxides to chromium oxides through the chemical treatmenton the stainless steel surfaces. On the basis of this method, selective growth of CNT patterns on stainless steel hasalso been demonstrated.

Introduction

Carbon nanotubes (CNTs) with their exceptional properties1

show great potential in building new and improved devices withbetter performance. Utilizing CNT-based electrodes in electro-chemical devices such as supercapacitors has displayed phe-nomenal storage capacity and power handling capability2

compared to conventional capacitors. Rechargeable Li-ionbatteries utilizing CNT electrodes have exhibited reversiblestorage capacities in the range of several hundred milliamperehours per gram, even at high cycling rates.3,4Novel devices suchas transistors,5 flat panel displays,6,7 and gas discharge tubes8

utilizing CNTs have already been demonstrated.In most of the above-mentioned applications, the device

performance would be dramatically improved if the CNTs weredirectly synthesized on a conducting substrate. For example, infield emission applications, the cathode is prepared by mixingthe CNTs with organic binders and then coating on a metalsubstrate.6Similarly, in applications involving the electrochemicalintercalation of Li ions, the CNTs were mixed with a polymericbinder in order to increase the adhesion of the electrode materialto the current collector.2,4 Several intermediate steps wereassociated4 with preparing these coatings, and care should betaken in order for the CNTs to be uniformly distributed andproperly adhered to the substrate surface. These problems can

be overcome by directly synthesizing CNTs on conductingsubstrates, which prevent the usage of any binder that reducesthe weight of the cell, which in turn accounts for the improvedenergy density of the cell. Furthermore, the mechanical robustnesswill also be greatly enhanced with very low contact resistancebetween the CNTs and the conducting substrate. In addition, thealignment of the CNTs can be kept intact in the directlysynthesized process, which is not possible when a conventionalcoating process is employed. Earlier studies involving alignedCNTs grown directly on conducting substrates have shown betterstability and overall performance in the respective applications.9-11

In recent years, there has been a considerable increase in thedevelopment of new strategies for synthesizing CNTs onconducting substrates. For example, synthesis of CNTs onstainless steel substrates using microwave plasma chemical vapordeposition (MPCVD),12,13 radio frequency-powered plasmaenhanced chemical vapor deposition (PECVD),14 and flamesynthesis15,16 have been reported. In the above-mentionedmethods, either the CNTs were not well aligned on the substrateor the synthesis procedure was laborious. For instance, in thePECVD14 method, the pretreatment of stainless steel involvespolishing, etching in hydrofluoric acid, and hydrogen plasmatreatment to create catalyst particles, which are quite cumbersome.Recently, the CNTs were grown on Inconel 600 using a floatingcatalyst CVD method, where the catalyst is allowed to flowalong with the carbon source.17 However, no report has hithertobeen available on the mechanism behind the growth of CNTson stainless steel substrates, which is very important for bettercontrol of the CNT density and alignment.

* Corresponding author. E-mail: weib@udel.edu.† Department of Electrical and Computer Engineering.‡ Department of Mechanical Engineering.(1) Ajayan, P. M.Chem. ReV. 1999, 99, 1787.(2) Niu, C.; Sichel, E. K.; Hoch, R.; Moy, D.; Tennent, H.Appl. Phys. Lett.

1997, 70, 1480.(3) Gao, B.; Kleinhammes, A.; Tang, X. P.; Bower, C.; Fleming, L.; Wu, Y.;

Zhou, O.Chem. Phys. Lett.1999, 307, 153.(4) Shin, H. C.; Liu, M.; Sadanadan, B.; Rao, A. M.J. Power Sources2002,

112, 216.(5) Tans, S. J.; Verschueren, A. R. M.; Dekker, C.Nature1998, 393, 49.(6) Wang, Q. H.; Setlur, A. A.; Lauerhaas, J. M.; Dai, J. Y.; Seelig, E. W.;

Chang, R. P. H.Appl. Phys. Lett.1998, 72, 2912.(7) Choi, W. B.; Chung, D. S.; Kang, J. H.; Kim, H. Y.; Jin, Y. W.; Han, I.

T.; Lee, Y. H.; Jung, J. E.; Lee, N. S.; Park, G. S.; Kim, J. M.Appl. Phys. Lett.1999, 75, 3129.

(8) Rosen, R.; Simendinger, W.; Debbault, C.; Shimoda, H.; Fleming, L.;Stoner, B.; Zhou, O.Appl. Phys. Lett.2000, 76, 1668.

(9) Croci, M.; Arfaoui, I.; Stockli, T.; Chatelain, A.; Bonard, J. M.Microelectron.J. 2004, 35, 329.

(10) Rao, A. M.; Jacques, D.; Haddon, R. C.; Zhu, W.; Bower, C.; Jin, S.Appl.Phys. Lett.2000, 76, 3813.

(11) Chen, J. H.; Li, W. Z.; Wang, D. Z.; Yang, S. X.; Wen, J. G.; Ren, Z.F. Carbon2001, 40, 1193.

(12) Wang, N.; Yao, B. D.Appl. Phys. Lett.2001, 78, 4028.(13) VanderWal, R. L.; Hall, L. J.Carbon2003, 41, 659.(14) Park, D.; Kim, Y. H.; Lee, J. K.J. Mater. Sci.2003, 38, 4933.(15) Lee, G. W.; Jurng, J.; Hwang, J.Carbon2004, 42, 667.(16) Pan, C.; Liu, Y.; Cao, F.; Wang, J.; Ren, Y.Micron 2004, 35, 461.(17) Talapatra, S.; Kar, S.; Pal, S. K.; Vajtai, R.; Ci, L.; Victor, P.; Shaijumon,

M. M.; Kaur, S.; Nalamasu, O.; Ajayan, P. M.Nat. Nanotechnol.2006, 1, 112.

9046 Langmuir2007,23, 9046-9049

10.1021/la7012232 CCC: $37.00 © 2007 American Chemical SocietyPublished on Web 07/17/2007

In this paper, we report a simple processing technique involvingcost-effective stainless steel (Type 304, composition is listed inTable 1) that can be used as a conducting substrate for the directgrowth of high-density aligned multiwalled nanotubes (MWNTs).The surface properties of the substrates were particularlyinvestigated, and the effect of treatment of the substrates on theCNT growth is critically analyzed. Such thorough understandingis a foundation for the site-selective growth of CNTs onconducting substrates for many prospective applications such asin nanoelectronics, field emission devices, and so forth. On thebasis of the understanding of the obtained X-ray photoelectronspectroscopy (XPS) results, here we report a conventionalphotoresist patterning technique for the selective growth of CNTpatterns on stainless steel substrates. This kind of photopatterningtechnique has been popularly used to synthesize CNT patternson Si-SiO2 substrates.18-22

Experiment

The as-received stainless steel foil (Type 304) with a thicknessof 25 µm was subjected to etching in 9 M sulfuric acid solution atroom temperature. The foil samples pretreated for different lengthsof time were loaded in the thermal CVD furnace for the CNT growth,and MWNTs were synthesized on these foils using a vaporizedmixture of ferrocene and xylene. Ferrocene acts as the catalystprecursor and xylene as the carbon source. The experiment wascarried out at a temperature of 700°C in an Ar/H2 gas atmospherefor 30 min to 1 h.

The XPS studies were performed with Kratos Axis-165 X-rayphotoelectron spectroscope in a high-vacuum environment of 10-11

Torr. An Al-KR X-ray source with pass energy of 40 eV was usedto obtain the spectra. Assuming a three-layer model23the compositionof the surface films was determined by applying curve fits to theXPS spectra after satellite and background removal.24

Results and Discussion

The stainless steel samples with different acid treatment timesof 1, 5, and 10 min along with the as-received sample (as thecontrol sample) were considered to elucidate the effect of surfaceproperties on the CNT growth. Figure 1 shows the scanningelectron microscope (SEM) micrographs of MWNTs synthesizedon the stainless steel foils treated for different times. It isobserved that the density and alignment of the CNTs stronglydepend on the duration of the acid etching time. The as-receivedstainless steel (Figure 1a) barely had any CNTs on its surface,whereas the sample etched for 1 min (Figure 1b) had randomlyscattered CNTs. The density of the CNTs gradually increasedon the substrates as the etching time increased, and highly alignedCNTs were obtained on the substrate etched for 10 min (Figure1d).

To understand the substrate effect on the CNT growth, all thetreated substrate samples along with the control substrate werefirst heated in the CVD furnace up to the CNT growth temperatureof 700 °C and kept at that temperature for 30 min in an Aratmosphere without any chemical flow. These samples wereallowed to cool to room temperature in the Ar atmosphere, andsurface property analysis was carried out using the XPS technique.The survey spectra of all the samples revealed that the peakswere mainly contributed by the compounds of iron and chromium.Nickel remained undetectable on all the surfaces of both theas-received and the treated samples. Figure 2a,b shows the XPSspectra of Cr 2p3/2 and Fe 2p3/2 for the as-received and 1, 5,and 10 min treated stainless steel samples. An intense chromiumsignal is observed on the surface of the as-received stainlesssteel foil, with almost no detectable iron peak. This is due to thepresence of the chromium-rich passive oxide film on the surfaceof the stainless steel.25 Generally, the chromium-rich passiveoxide film is believed to protect stainless steel from corrosion.As the sample is treated in the 9 M sulfuric acid, a gradualdecrease in the intensity of the chromium signal and an increasein the iron signal are noticed with the increase in the treatmenttime from 1 min to 10 min. This suggests that chromium is beingdepleted from the surface of the treated samples.

The composition of the surface films was analyzed byperforming a detailed curve fitting of the XPS spectra for all thesamples. Figure 2c,d shows typical curve fittings of Cr 2p3/2and Fe 2p3/2 spectra for the 5 min treated sample. Chromiumgave three distinguishable peaks at 575.5, 576.6, and 578.1 eV.The peak at 575.5 eV (blue curve in Figure 2c) corresponds tothe CrN due to the absorbed nitrogen from the atmosphere. Thepeak at 578.1 eV (magenta curve in Figure 2c) corresponds tothe Cr6+ state of chromium. The dominant peak is contributedby the Cr3+ located at 576.6 eV (Figure 2c). For the Fe 2p3/2spectra, there is a small peak at 709.3 eV (blue curve in Figure2d) corresponding to the Fe2+ state of iron, and the dominantpeak at 710.7 eV corresponds to the Fe3+ state (red dashed curvein Figure 2d). Another developed predominant peak at 711.0 eVwas observed as the sample treatment time was increased thatalso corresponded to the Fe3+ state (red dotted curve in Figure2d). The summary of the dominant peaks of Cr 2p3/2, Fe 2p3/2,and O 1s spectra (spectra not shown) for all the samples istabulated in Table 2. This reveals that, out of all the compoundsof chromium, Cr2O3is the prevailing oxide present on the surfaces.Similarly, Fe2O3 is the dominant compound of iron present onthe surfaces of the acid-treated samples. The presence of oxygenpeaks at 530.1( 0.1 eV gave additional confirmation for theexistence of either Cr2O3 and/or Fe2O3.21 All the peak positionsclosely matched the standard results in the Handbook of X-rayPhotoelectron Spectroscopy.26

The ratios of the areas of Fe2O3 to Cr2O3 peaks of all thesamples are shown in Table 3. For the as-received sample, theratio is much less than 1, revealing that the surface is mainlycovered by the Cr2O3 film. Figure 2a shows that there is virtuallyno CNT on the surface of the as-received stainless steel. Thismay be due to the fact that the catalyst iron particles comingfrom the gaseous mixture are poisoned by the passive chromiumoxide film present on the substrate, possibly by forming a Fe-Cr-O compound. The surface is therefore devoid of active ironparticles acting as catalysts for CNT nucleation and growth. Thisis analogous to the selective growth mechanism of the CNTs onSi/SiO2 substrates.27 It has to be noted that, by passing only

(18) Ren, Z. F.; Huang, Z. P.; Xu, J. W.; Wang, J. H.; Bush, P.; Siegal, M.P.; Provencio, P. N.Science1998, 282, 1105.

(19) Fan, S. S.; Chapline, M. G.; Franklin, N. R.; Tombler, T. W.; Cassell,A. M.; Dai, H. J.Science1999, 283, 512.

(20) Kind, H.; Bonard, J. M.; Emmenegger, C.; Nilsson, L. O.; Hernadi, K.;Maillard-Schaller, E.; Schlapbach, L.; Forro, L.; Kern, K.AdV. Mater. 1999, 11,1285.

(21) Wei, B. Q.; Vajtai, R.; Jung, Y.; Ward, J.; Zhang, R.; Ramanath, G.;Ajayan, P. M.Nature2002, 416, 495.

(22) Wei, B. Q.; Vajtai, R.; Jung, Y.; Ward, J.; Zhang, R.; Ramanath, G.;Ajayan, P. M.Chem. Mater.2003, 15, 1598.

(23) Asami, K.; Hashimoto, K.Langmuir1987, 3, 897.(24) Sherwood, P. M. A. InPractical Surface Analysis; Briggs, D., Seah, M.

P. Eds.; Wiley: New York, 1983; Appendix 3, p 445.

(25) Yang, W. P.; Costa, D.; Marcus, P.J. Electrochem. Soc. 1994, 141, 111.(26) Moulder, J. F.; Stickle, W. F.; Sobol, P. E.; Bomben, K.; Chastain, J.

Handbook of X-ray Photoelectron Spectroscopy, 2nd edition; Perkin-Elmer:Waltham, MA, 1992.

Table 1. Chemical Composition of the Most Commonly UsedAustenitic Stainless Steels (Type 304)

grade (wt.%) C Mn Si P S Cr Ni Fe

304 0.08 2.0 1.0 0.045 0.03 18-20 8-10.5 rest

Direct Growth of MWNTs on Treated Stainless Steel Langmuir, Vol. 23, No. 17, 20079047

ethylene carbon source, no nanotubes were observed on thestainless substrate without a catalyst precursor. The stainlesssteel substrate acts as a support for the nanotube growth but isdevoid of any catalyst sources such as free metallic iron particles.

The surface treatment process as elucidated by XPS studiescan be described as follows: The surface passive layer of theas-received sample gives initial protection to the stainless steelfrom corrosion. When the stainless steel foil is subjected to a

treatment in 9 M sulfuric acid, the passive film weakens on thesurface with time, and etching of the stainless steel is observedby dissolution of chromium in the acid. The etching processstarts slowly at the beginning and gradually increases as thepassive film is completely removed from the surface. When thestainless steel foil is removed from the acid, a mesoporous ironoxide film forms on the surface of the stainless steel, which isbelieved to be favorable to the CNT growth. For the foil treatedfor 1 min, the ratio of iron oxide to chromium oxide is about 0.9,indicating that only partial passive Cr2O3 film is removed fromthe surface, and so the CNTs are not dense on the substrate, asobserved in Figure 1b. As the etch time is increased, the formationof the iron oxide film on the foil surface is complete and thereis an increase in both the alignment and the density of the CNTs(see Figure 1c,d). It has to be noted that the oxide ratio presented(27) Jung, Y. J.; Wei, B. Q; Vajtai, R.; Ajayan, P. M.Nano Lett.2003, 3, 561.

Figure 1. SEM micrographs of the MWNTs synthesized on various stainless steel samples: (a) as-received substrate, (b) substrate treatedfor 1 min (c) substrate treated for 5 min, and (d) substrate treated for 10 min.

Figure 2. (a,b) XPS spectra of Cr 2p and Fe 2p signals for varioustreated samples. (c,d) Typical curve fitting of the Cr 2p and Fe 2pspectra for the 5 min treated sample. The red curves represent thecorresponding dominant oxide component in both of the plots.

Table 2. XPS Peak Analysis of Cr 2p3/2, Fe 2p3/2, and O 1sa

surfacecomponent

binding energy(eV)

valancestate

chemicalspecies

chromium 576.6( 0.1 3+ Cr2O3

578.1( 0.1 6+ CrO3

iron 710.7( 0.1 3+ Fe2O3

711( 0.1 3+ Fe2O3

711.5( 0.1 3+ FeOOHoxygen 530( 0.1 2+ oxide

530.2( 0.1 2+ oxide

a Only dominant peaks from the curve fitting data are tabulated.

Table 3. Ratio of Iron Oxide to Chromium Oxide Obtainedfrom the XPS Spectra Fits

sample etch time iron oxide/chromium oxide ratio

as-received (no etching) 0.11 min 0.95 min 2.310 min 8.0

9048 Langmuir, Vol. 23, No. 17, 2007 Masarapu and Wei

in Table 3 is a relative number and gives an idea of the dominantsurface component. This ratio can vary slightly for differentsamples etched for the same period of time, but the variation inthe trend is always the same with an increase in the etch timeof the samples.

Since the CNTs grow only on the treated stainless steel surfaces,by preventing some areas of the stainless steel surfaces frometching we obtained selective growth of CNTs on patternedstainless steel substrates. Figure 3a shows a stainless steel substrate

patterned with photoresist by conventional microfabricationtechnique. The characters “NMDL” on the substrate are exposedto the acid while the remaining area is protected by the photoresist.After the acid treatment, the substrate was cleaned thoroughlywith acetone to remove all the photoresist and was placed in theCVD furnace for CNT growth. An SEM image of the top viewof the sample after the CNT growth can be seen in Figure 3b.It is clear from this figure that the CNTs are present only on thecharacters NMDL and not on the remaining places that werecovered with photoresist.

The method reported here on the direct growth of alignedCNTs on the stainless steel foil is superior in various respectscompared to other techniques reported before.12-16For example,when compared with the PECVD14 method where multiplepretreatment steps of stainless steel were involved, the presentmethod has only one pretreatment step, and it is very easy toexecute. Also, the present method may be economically scaled-up to large-scale production, and the alignment and density ofCNTs could readily be controlled compared to MPCVD12 andthe diffusion flame synthesis method.15,16 There is no need forthe catalyst preparation step since the catalyst is directly sentalong with the carbon source into the CVD furnace. This methodgives a higher yield of CNTs with growth rates exceeding 10µm/min compared to the MPCVD method.12 The stainless steelfoil utilized is only 25µm thick and is very flexible to be cutinto any size and shape with regular scissors. It can be used tomake sturdy and light-weight electrodes for field emission andelectrochemical applications.

Conclusions

In conclusion, an efficient and cost-effective method ofsynthesizing highly aligned MWNTs directly on stainless steelsubstrates is reported. XPS analysis reveals that the surface ofthe as-received stainless steel is covered by a passive chromiumoxide film, which is not favorable for the growth of CNTs becauseof catalyst poisoning. Treating the stainless steel in a strongsulfuric acid for several minutes removes the passive layer and,by enriching the surface with iron oxide, very much supports theCNT growth. The density and alignment of the CNTs dependon the ratio of the iron oxides to the chromium oxides and thereforecan be controlled by tuning the oxide ratio through chemicaltreatment on the stainless steel surface. Controlled etching onthe substrate surface proves to be an effective method for thesite-selective growth of CNTs on electrically conductive sub-strates.

Acknowledgment. The authors are deeply grateful to theNational Science Foundation for financial support of the projectunder the NSF award number DMI-0457555.

LA7012232

Figure 3. Selective growth of MWNT patterns on the stainlesssteel foil. (a) Top view of the stainless steel patterned with photoresist.The characters NMDL are not covered by the photoresist. (b)Nanotube growth only in the exposed regions of letter patterns onthe stainless steel foil.

Direct Growth of MWNTs on Treated Stainless Steel Langmuir, Vol. 23, No. 17, 20079049