0925-9635/02/$ - see front matter䊚 2002 Elsevier Science B.V. All rights reserved. PII: S 0 9 2 5 - 9 6 3 5 Ž 0 1 . 0 0 6 4 0 - 9
Growth mechanism and properties of the large area well-aligned carbon
nano-structures deposited by microwave plasma electron cyclotron
resonance chemical vapor deposition
Chao Hsun Lin*, Hui Lin Chang, Ming Her Tsai, Cheng Tzu Kuo
Department of Materials Science and Engineering, National Chiao Tung University, Hsinchu 300, Taiwan, ROCAbstract
Large area (4-inch diameter) well-aligned carbon nano-structures on Si substrate were successfully synthesized by using a catalyst-assisted microwave plasma electron cyclotron resonance chemical vapor deposition (ECR-CVD) system with CH as4
source gas.The catalysts include Fe, Co and Ni.The catalysts and the deposited nano-structures were characterized by scanning electron microscopy(SEM), transmission electron microscopy (TEM), Raman and field emission I–V measurements.Effects of process parameters on morphologies, structures and properties of the nano-structures were examined.The results show that the deposited nano-structures can include normal nano-tubes, split catalyst nano-tubes, seaweed-like nano-sheets and carbon film, depending mainly on substrate temperature and bias, catalyst materials and their application methods.Deposition mechanisms for different nano-structures, especially, the unique split catalyst nano-tubes and seaweed-like nano-sheets, were proposed.The differences in oxidation resistance and field emission properties between different nano-structures will be compared and discussed. 䊚 2002 Elsevier Science B.V. All rights reserved.
Keywords: Carbon; Electron cyclotron resonance(ECR); Chemical vapor deposition (CVD); Catalytic processes
1. Introduction
Since the discovery of fullerenes in 1985 w1x and carbon nano-tubes(CNTs) in 1991 w2x, materials science research and technology development on nano-structures have been intensively conducted in the world.The nano-structure materials have been synthesized from atoms and molecules of themselves, namely, self-assembly, to become scaled particles, tubes, nano-wires, nano-layers, etc.They possess many special magnetic, electrical, physical or chemical properties as compared with the bulk materials.Therefore, the nano-structures have high potential applications in areas of dispersions and coating, high surface area materials, functional nano-devices, etc. w3x.
Carbon nano-structures have been synthesized by a variety of methods, e.g. arc discharge w4x, laser ablation w1x, microwave plasma chemical vapor deposition
*Corresponding author.Tel.: q886-3-573-1949; fax: q886-3-572-1065.
E-mail address: [email protected](C.H. Lin).
(MPCVD) w5x, template-based CVD w6x, etc.A few deposition mechanisms for nano-structures have been proposed w7–9x.However, large area ()4 inch in diameter) carbon nano-tubes have not been successfully synthesized so far in the literatures, and the proposed deposition mechanisms could not successfully explain the formation of many different nano-sturctures.The purpose of this work was to synthesize various carbon nano-stuctures and to study their properties and forma-tion mechanisms, and to develop large area deposiforma-tion techniques for potential mass production researches in the future.It is known that the microwave plasma ECR-CVD has the advantages of high dissociation percentage of the precursor gas and high uniformity of plasma energy distribution, and was generally used for large area dry etching, surface cleaning, or diamond film deposition.Therefore, we adopted the advantages of ECR-CVD to synthesize large area carbon nano-struc-tures without using templates, though ECR-CVD was used to synthesize CNTs or carbon nitride nano-tubes with anodic alumina as templates w10,11x
Fig.1.SEM micrographs of catalysts:(a) Fe catalyst; (b) Co catalyst; (c) Ni catalyst and (d) Si substrate.The corresponding as-deposited
nano-structures:(e) Fe catalytic grown nano-sheets; (f) Co catalytic
grown CNTs;(g) Ni catalytic grown CNTs and (h) the nano-sheets
on Si substrate without catalyst application. 2. Experimental
Large area(4-inch diameter) well-aligned CNTs were synthesized on Si wafer by using catalyst-assisted ECR-CVD method with CH as gas source.The transition4
metals acted as catalysts include Fe, Ni and Co metals. The Fe, Co and Ni catalysts were applied on the substrate surfaces by spin coating a thin layer of iron nitrate wFe(NO ) x, cobalt nitrate wCo(NO ) x, or nickel3 2 3 2
acetate wNi(OOCCH ) x in 1 M water solutions, respec-3 2
tively.Then the spin-coated substrates were dried in air and reduced in 10-sccm flow rate of hydrogen plasma at approximately 600 8C substrate temperature for 15 min to become well-distributed nano-sized catalysts. The catalyst-coated substrates were then deposited with nano-structures in the same ECR-CVD system under 875-Gauss magnetic field strength by varying different deposition conditions.The substrate temperatures were varied by a substrate heater and plasma heating effect from 330 to 600 8C.The substrates were subjected to either 0 V or y100 V bias.The deposited nano-structures were characterized by FESEM(field emission scanning electron microscopy), TEM, HRTEM (high resolution transmission electron microscopy), Raman spectroscopy and I–V measurements.The TEM speci-mens were prepared by immersing and agitating the specimens in an ultrasonic acetone bath to disperse the nano-structures.Then the droplets with the dispersed nano-structures were dropped and dried on a carbon-coated Cu mesh for TEM examination.
3. Results and discussion
3.1. Effect of catalyst types on carbon nano-structure growth
The growth mechanisms for CNTs and nano-sheets were examined by comparing different catalytic-grown nano-structures.The SEM micrographs for the Fe, Co and Ni catalyst-coated substrate surfaces and the Si substrate itself, before nano-structure depositions, are shown in Fig.1a–d.The corresponding as-deposited nano-structures are depicted in Fig.1e–h, respectively, under the same deposition conditions (20 sccm CH ,4
800 W microwave power, 600 8C substrate temperature, y100 V substrate bias, and 15 min deposition time). These figures indicate that either the Fe catalyst or no catalyst applications show nano-sheet-like nano-struc-tures and have no nano-tube formation (Fig.1e,h).The Fe catalyst particles in Fig.1a are approximately 30;45 nm in size, but the particles with relative shallow boundaries are embedded tightly like a film.This may relate to a higher sintering effect of Fe particles than other catalysts under the same catalyst reduction process at 600 8C.In other words, the tightly sintered particles will be more difficult to melt during deposition to
become nano-scaled liquid droplets, which are pushed upward individually to form nano-tubes, according to the catalyst-assisted nano-tube tip-growth mechanism w7x.The same reasoning can be applied to the case of no catalyst applications in Fig.1h.In these cases of the nano-sheet formation, the C2 (C–C bond) bonds in the
initial stage from the precipitated carbons andyor free carbons may form nucleus structures and be connected linearly with each other on the surface of the tightly sintered catalyst particles or the substrate without cata-lyst applications.The free carbons from the plasma atmosphere will further adhere on the existing structure to become a scale-like carbon sheet.The sheets are then connected with each other in wavy fashion to become seaweed-like nano-sheets.In an ECR plasma system with the substrate bias, the gas dissociation is mainly controlled by microwave power with assistance of ECR. The kinetic energy of positive ions is mainly manipu-lated by the negative substrate bias w12x.Furthermore, since the ECR-CVD system is working under lower pressure(2=10y3 torr), most of the ion motion
trajec-Fig.2.TEM micrographs:(a) Co and (b) Ni catalytic grown CNTs.
Fig.3.HRTEM micrograph of CNT lattice image around Co catalyst particle.
Fig.4.TEM micrograph of the ‘split Co catalyst nano-tubes’.
tory will be vertical to the substrates.Therefore, the nano-sheets(approx.20 nm in thickness) are perpendic-ular to the substrate surface.
The factors to determine the CNT morphologies were studied by examining the differences in the Co and Ni catalytic-grown CNTs in Fig.1f,g, which are correspond-ing to the side view and top view of the tubes, respectively.It is noted that the sizes of the Co catalyst particles in Fig.1b are ranging from 30 nm to 80 nm, which are approximately the sizes of the nano-tubes in Fig.1f.The same conclusion is applied to the Ni catalytic grown CNTs in Fig.1g, where the tube diam-eters are in the range of 30–60 nm.This is in agreement with the reported statement that the sizes of the catalyst particles determine the diameters of the nano-tubes w8– 10x.As to the tube-number densities, a smaller particle size combining with a denser distribution of the catalyst particles on the substrate surface can result in a higher tube-number density.Under the present deposition con-ditions, the tube-number densities of the CNTs for Ni catalyst (55;60=10 tubesycm ) are greater than for6 2
Co catalyst (30;35=10 tubesycm ).Regarding the6 2
shapes of the tubes, Fig.2a,b reveals the TEM micro-graphs of Co and Ni catalytic grown CNTs.From Fig. 2, the catalysts at the tube caps are obvious, though some of the catalysts are missing due to TEM sample preparation.This implies that the growth mechanism in these cases is the tip growth mechanism w7x.The Co catalytic grown CNT tip lattice image by HRTEM around a catalyst particle is shown in Fig.3.It shows that the catalyst cap conforms closely to the nano-tube core and the multiple graphene layers are 0.34 nm in spacing, which is in agreement with the reported value for CNT wall spacing w13x.
3.2. Growth mechanism of the split catalyst nano-tubes According to the tip growth mechanism, the catalyst is pushed upward by the growing nano-tubes, and the
whole catalyst particle becomes the cap of the tubes.It is interesting to note that the Co catalysts are not only located at the caps of the tubes but also separately at few sites along the center of the tubes, as shown in Fig. 4.This may be reasoned by imagining the Co catalyst is split into few small pieces following a bigger mother catalyst at the tube cap.These types of nano-tubes will be called the ‘split catalyst nano-tubes’.They are more often found in the bigger catalysts or tubes.Regarding the splitting action, the capillary force from the tube center hole and the known gravity force of the catalyst will most probably elongate the liquid drop of the larger catalyst particle into the tube center during the tube growth.The tube growth direction may also most prob-ably be altered for a larger catalyst, because it is similar to an unstable one-dimension structure with overweight at one end.Furthermore, the change in growth direction will most probably cause the splitting of the elongated liquid drops.The split droplets will be left on the original sites by the capillary force.This is supported by the fact that the split positions are often at locations where the tube changes the growth direction.
3.3. Effects of other process parameters on carbon nano-structure formation
Besides catalyst type, particle size and shape, the nano-structures are also strongly affected by substrate temperature and bias.The results show that the required conditions to form three different nano-structures (nano-tubes, seaweed-like nano-sheets and carbon films) are
Fig.5.Raman spectra of the Co and Ni catalytic grown CNTs, the Fe catalytic grown nano-sheets, and the nano-sheets on Si substrate with no catalyst application, respectively(deposition time 15 min, 20 sccm
CH4).
as follows.For nano-tube formations, the required con-ditions include higher substrate temperature()560 8C), negative substrate bias and catalyst application.As to seaweed-like nano-sheets formation, the required con-ditions are the same as the nano-tubes except without the catalyst application.However, the carbon film will form if the substrate temperature is not high enough.It is obvious that both the substrate temperature and bias are necessary conditions to form nano-tubes and sea-weed-like nano-sheets.It is believed that the effect of substrate temperature is to minimize the thermal energy required to activate the catalysts to precipitate carbon atoms to form nano-tubes or to activate the substrate surface to accommodate carbon atoms to form carbon nano-sheets.Moreover, effect of the negative substrate bias is to attract and accelerate the positive species in plasma to stick into and form the nano-structures.The unique feature of the well-aligned tubes and nano-sheets perpendicular to the substrate surfaces may be a consequence of the substrate bias effect plus the action of magnetic field to generate ECR condition, as dis-cussed in the previous paragraph.
3.4. Raman spectra of the nano-structures
Fig.5 is the micro-Raman spectra of the different catalytic growth nano-structures on Si substrate.The excitation laser is a 514.5-nm Ar-ion laser. The D-line, 2nd order D-line and D9-line are from 1349–1354 cmy1, 2694–2702 cmy1and 3228–3244 cmy1, respec-tively.The corresponding G-line and the combination of D- and G-line are from 1585–1603 cmy1 and 2942– 2946 cmy1, respectively.Here, the Fe catalytic and non-catalytic grown nano-sheets have relatively higher wave numbers of D- and G-line than in the Ni and Co catalytic grown multi-walled nano-tubes.According to Cooper and Young w14x, the tensile stress in the nano-tubes can cause Raman peak positions shifting to a lower wave number side.This may imply that the nano-tubes suffer a tensile residual stress by comparing with
that in the nano-sheets, because the nano-sheets can be considered to consist of the graphene planes without bending stress.It can be reasoned by the fact that bending the graphene planes to a tube shape forms the nano-tubes.This is supported by the fact that the wave numbers of the Raman peaks for SWNTs(single-walled nano-tubes) are generally lower than that for MWNTs (multi-walled nano-tubes) due to a greater bending or higher tensile residual stress in SWNTs.The reported D-line and G-line peaks for nano-tubes are approximate-ly 1308–1349 cmyland 1532–1594 cmy1, respectively w14–17x.In other words, the D-line (1349 cm ) andy1
G-line (1585 cm ) peak positions of our nano-tubesy1
are closer to the higher wave number side.This means that our nano-tubes suffer a less tensile stress than the reported nano-tubes from the literatures.
Regarding I(G)yI(D) ratio, the ratios for the Ni and Co catalytic grown nano-tubes are 0.55 and 0.57, respec-tively.In contrast, the ratios for Fe catalytic grown and no catalytic grown nano-sheets are 0.59 and 0.45, respectively.This implies that the catalytic grown nano-tubes or nano-sheets possess more G–bonding than non-catalytic grown nano-sheets.In other words, the catalysts can enhance the G–bond formation in addition to assist-ing the tube formation.
3.5. Oxidation resistance of the nano-structures
Oxidation resistance of carbon nano-tubes was exam-ined under air atmospheres at 300–600 for 30 min.The Co catalytic grown nano-tubes are purified in structure at 300 8C, collapsed at 450 8C, and burned at 600 8C. In contrast, the Ni catalytic grown nano-tubes are puri-fied without structure collapse up to 450 8C, and also burned at 600 8C.In other words, the oxidation resis-tance of the Ni catalytic-grown nano-tubes is greater than the Co catalytic grown nano-tubes.This may imply that the Co catalytic grown nano-tubes contain more structure defects.This is supported by TEM examination in Fig.2a,b.
3.6. Field emission properties of the nano-tubes
The I–V curves of the 5th cycle test to represent the
field emission properties of the relatively stable state of the nano-structures are shown in Fig.6.It indicates that the Co catalytic grown nano-tubes possess the best properties with current density beyond the instrument capability.The results also show that the emission properties depend on the catalyst type.The order in magnitude of emission current density of the nano-structures for various catalyst applications is Co ()32 mAycm2) )Ni (19.8 mAycm ) )Fe (7.1 mAycm ) )2 2
no catalyst (2.5 mAycm ) at 10 Vymm.The turn-on2
voltages are in order of Ni(1.1 Vymm) -Co (3.0 Vy mm) -no catalyst (4.6 Vymm) sFe (4.6 Vymm).By
Fig.6.Field emission I–V curves of the Co and Ni catalytic grown CNTs, the Fe catalytic grown nano-sheets, and the nano-sheets on Si substrate with no catalyst application, respectively(deposition time
15 min, 20 sccm CH4).
comparing the Co and Ni catalytic grown nano-tubes, the better emission current density for the former one must relate to a combination effect of the tube mor-phologies.The Co catalytic grown nano-tubes possess a proper combination of tube diameter and tube-number density to minimize the screen effect among neighbor tubes to enhance the effective emission area, which may compensate insufficient in-field enhancement factor due to a lower aspect ratio of tube length-to-diameter.In other words, the best field emission properties can be improved by decreasing the tip radius of the tubes and manipulating a proper tube-number density.For Co catalytic grown nano-tubes, the current density can reach 32 mAycm at 7.6 Vymm; and the threshold voltage2
can go down to 6.6 Vymm.
4. Conclusions
The deposition conditions to synthesize different nano-structures, including normal CNTs, unique split catalyst CNTs and seaweed-like carbon nano-sheets, were examined.The catalyst types and their application methods, substrate temperature and bias are found to be the crucial parameters of determining the deposited nano-structures and their properties.The growth mech-anisms for different nano-structures, especially ‘split catalyst nano-tubes’ and ‘seaweed-like nano-sheets’, were proposed.The relationships between the nano-structures and properties were studied.The properties of different catalytic grown nano-tubes were compared. The properties including Raman I(G)yI(D) ratio, oxi-dation resistance and field emission were determined. The results show that the morphologies(i.e. tube length, diameter, shape and tube-number density) and the struc-ture defects of CNTs depend on catalyst type and its application method.The CNTs with less structure defects
may result in a higher oxidation resistance.This explains why the Ni catalytic grown nano-tubes have a better resistance than the Co catalytic grown tubes, As to field emission properties of the CNTs, the Co catalytic grown CNTs possess better properties (current density s32 mAycm at 7.6 Vymm, turn-on voltage s3.0 Vymm,2
threshold voltage s6.6 Vymm).The best field emission properties can be improved by decreasing the tip radius of the tubes and manipulating a proper tube-number density.
Acknowledgments
This work was supported by the National Science Council (contract no.: NSC89-2216-E-009-020 and-019), the Ministry of Education (contract no.: 89-E-FA06-l-4) and the ITRI-MRL of Taiwan (contract no. 90-3XS2542 or NCTU-C90098).
References
w1x H.W. Kroto, J.R. Heath, S.C. O’Brien, R.F. Curl, R.E. Smally,
Nature 318(1985) 162. w2x S.Iijima, Nature 354(1991) 56.
w3x R.W. Seigel, in: G.M. Holdridge(Ed.), Nano-structure Science
and Technology, Kluwer Academic Publishers, Netherlands, 1999, pp.1–13.
w4x T.W. Ebbesen, Carbon Nano-tubes Preparation and Properties,
CRC Press, New York, 1997.
w5x T.A.Habermann.Gohl, K.Janischowsky, D.Nau, M.Sta-mmler, L.Ley, G.Muller, in: Proc.11th Intl.Microelectronics Conference, 1998, p.200.
w6x A.Huczko, Appl.Phys.A 70(4) (2000) 365.
w7x X.H. Chen, S.Q. Feng, Y. Ding, J.C. Peng, Z.Z. Chen, Thin
Solid Films 339(1999) 6.
w8x C.J. Lee, D.W. Kim, T.J. Lee, Y.C. Choi, Y.S. Park, Y.H. Lee,
W.B. Choi, N.S. Lee, G.S. Park, J.M. Kim, Chem. Phys. Lett. 312(1999) 461.
w9x S.B.Sinnott, R.Andrews, D.Qian, A.M.Rao, Z.Mao, E.C.
Dickey, F.Derbyshire, Chem.Phys.Lett.315(1999) 25. w10x S.H. Tsai, F.K. Chiang, T.G. Tsai, F.S. Shieu, H.C. Shih, Thin
Solid Film 366(2000) 11.
w11x S.L. Sung, S.H. Tsai, C.H. Tseng, F.K. Chiang, X.W. Liu, H.C.
Shih, Appl.Phys.Lett.74(2) (1999) 197.
w12x S.W. Pang, in: A.N. Goldstein(Ed.), Handbook of Nano-phase
Materials, Marcel Dekker, New York, 1997, p.15.
w13x P.M. Ajayan, in: H.S. Nalwa (Ed.), Handbook of
Nano-structured Materials and Nano-technology, Academic Press, San Diego, 2000, p.377.
w14x C.A. Cooper, R.J. Young, in: D. Tomanek, R.J. Enbody(Eds.),
Science and Application of Nano-tubes, Kluwer Academic, New York, 2000, p.93.
w15x S.J. Chung, H.S. Kim, J. Jang, S.H. Lim, C.H. Lee, B.Y.
Moon, SID 01 DIGEST, Society of Information Display, San Jose, 2001, p.92
w16x J.M. Kim, W.B. Choi, N.S. Lee, J.E. Jung, Diamond Relat.
Mater.9(2000) 1184.
w17x R.R.Bacsa, Ch.Laurent, A.Peigney, W.S.Basca, Th.Vaugien,
