8/3/2019 L'Ancien d'Algrie, Jun 2011

1/3

Dynamic Growth Rate Behavior of aCarbon Nanotube Forest Characterizedby in Situ Optical Growth Monitoring

Do-Hyung Kim,*, Hoon-Sik Jang, Chang-Duk Kim, Dong-Soo Cho,

Hee-Sun Yang, Hee-Dong Kang, Bong-Ki Min, and Hyeong-Rag Lee*,

Department of Physics, Nanophyscis and Technology Laboratory,

Kyungpook National UniVersity, Daegu, 702-701, Korea

Received April 8, 2003

ABSTRACT

We characterize the dynamic growth rate behavior of a carbon nanotube (CNT) forest grown by means of optical interference phenomena. The

CNT growth rate increased with an increase in CNT length at the initial stage and became stabilized after the CNT length was about 2 m.Then the growth rate started to decelerate, passing the critical growth length in an almost linear manner. The termination length of the carbon

nanotube was also precisely estimated by fitting the data of growth rate of carbon nanotubes to time. It was found that the CNTs show a

transition from straight to curly nanotubes that is related to the decrease in the growth rate. The use of an in situ optical monitoring method

has made possible the delicate length control of carbon nanotubes independent of the growth rate.

Carbon nanotubes (CNTs) have shown a variety of applica-

tions such as scanning probes,1,2 field emitters,3-5 and

nanoelectronic devices.6 The nanoscale structuring of carbon

nanotubes with respect to growth7-9 has rapidly progressed.

The in situ control of CNT growth will be required in order

to optimize and realize their applications in the area of

vacuum electronics. In addition, the in situ control of thelength of CNTs is needed for optimal performance as well

as for wider applications involving nanobased device struc-

tures. However, in situ characterization and control methods

are difficult to apply during the CNT growth period due to

unfavorable conditions such as no vacuum, gas ambient

conditions, and high temperatures used in the chemical vapor

deposition (CVD). In this letter, we describe an in situ

method for monitoring the growth of aligned carbon nano-

tubes based on an optical interference technique. Using this

method, it was possible to investigate the growth behavior

of carbon nanotubes during the growth period. Furthermore,

the length of carbon nanotubes can now be precisely

controlled.A typical plasma-enhanced chemical vapor deposition

(PECVD) technique was used to grow the aligned CNTs.9

The substrate consisted of silicon wafers with 10 nm Ni

sputtered film. Flows of C2H2 and NH3 were kept constant

at 60 sccm and 180 sccm, respectively. A mixture of

acetylene (C2H2) and ammonia (NH3) was used as the gas

source at 2 Torr. CNT growth was performed at 700 C.

A -450 V bias was applied to the substrate in order to create

a dc plasma. We counted about 100 nanotubes per square

micrometer. In situ optical interferences were measured by

means of a focused 650 nm laser diode with a 1 mm beam

diameter and a photodiode detector. Tempered glass was usedfor the PECVD growth chamber in order to transmit the

coherent laser beam from the laser source to the substrate

and the reflected beam from the substrate to a detector. We

confirmed that the plasma had negligible influence on the

transmittance of the laser light to the photodiode. Measure-

ments and growth experiments were automatically controlled

by a personal computer and in-house prepared software.

Figure 1a shows the interference oscillations measured at

a 30 angle to the beam incidence angle during the growth

of nanotubes under previously optimized growth conditions.

The interference patterns can be varied by experimental

conditions such as laser wavelength, beam incidence angle,

and growth conditions, which also provides a very effective

way to investigate growth behavior. The interference oscil-

lations in Figure 1a represent interference phenomena

between the reflected beam from the top of carbon nanotubes

and from the surface of a nickel-coated substrate. The

diminution in the interference oscillations is mainly due to

the absorption of laser light through the CNTs. The reflec-

tance remains nearly unchanged during NH3 pretreatment

prior to CNT growth in the presence of C2H2 and NH3. The

starting intensities of the reflected laser beam were slightly

* Corresponding authors. E-mail: hrlee@nptl.knu.ac.kr; dhkim@nptl.knu.ac.kr

Kyungpook National University. Korea Basic Science Institute, Daegu branch. Instrumental Analysis Center, Yeungnam University, Kyongsan.

NANO

LETTERS

2003Vol. 3, No. 6

863-865

10.1021/nl034212g CCC: $25.00 2003 American Chemical SocietyPublished on Web 05/17/2003

8/3/2019 L'Ancien d'Algrie, Jun 2011

2/3

different than the experimental setup and the initial laser

intensity. However, the unique interference oscillations were

shown in each run. The four curves in Figure 1a were

obtained from the growth of well-aligned CNTs with

identical growth conditions. However, the oscillations showslightly different behaviors, which means the growth always

brings the unintentional slight fluctuations of growth condi-

tions. Meanwhile, the as-grown CNTs do not show signifi-

cant differences, and the growth lengths were ruled by the

interference formula. Constructive interference by a thin film

is described by dsin ) n /2, where and are thewavelength and incidence angle of the laser beam, dthe film

thickness, and integral n () 0, 1, 2,...). Our experimental

results indicate that this simple equation can be used to

successfully describe the growth of the carbon nanotubes.

Figure 1b shows the average length of CNTs as function of

growth time. The average growth length of CNTs increased

in the initial stage, then gradually stabilized with increasinggrowth time.

The lengths of the carbon nanotubes were measured by

scanning electron microscopy (SEM) as shown in Figure 2,

and these data are in agreement with interference oscillations.

The lengths of the nanotubes are plotted as a function of the

interference oscillation in the inset of Figure 1. The growth

rate per oscillation was determined to be about 649 nm ( 3

nm, which coincides with the 650 nm wavelength of the laser

used here. Various shapes for the interference oscillations

were measured in order to confirm the consistency between

the measured oscillation patterns and laser source conditions

such as the beam incidence angles and the laser wavelengths.Figure 3 shows the growth rate as a function of the CNT

length calculated from the data in Figure 1. The growth rate

increases with an increase in CNT length at the initial stage,

which can be due to the involvement of a surface diffusion

process suggested by Louchev et al.10,11 The carbon species

collide with CNT surfaces and diffuse to the CNTs wall edge

where they incorporate directly into the wall or via diffusion

through a catalyst particle. These papers show that the growth

rate increases as long CNT length is smaller than the

diffusion length. In our experiment, the growth rate increase

continues up to a length of 2 m. After the maximumgrowth rate is reached, it becomes stabilized. The growth

rate becomes almost constant and then finally starts to

decrease at the critical CNT length in an almost linear

manner, as shown in Figure 3. This termination behavior is

attributed to the change of the critical growth condition

passing periods with stable growth rate. O. Zhou et al. 12 have

suggested that the termination is due to the encapsulation of

catalyst metal at the bottom of the nanotubes after an

extended period of growth. However, diverse experimental

parameters and lack of in situ methods to characterize CNT

growth have so far hindered a complete understanding of

the precise nature of the termination of CNT growth. Figure

Figure 1. (a) Interference oscillation behaviors as a function ofgrowth time. The oscillation periods were gradually increased withincreasing growth time. The inset shows the length of carbonnanotubes as a function of the oscillation period and the linear fitof experimental data has the relation Y) (0.649 ( 0.003)X. (b)Average length of a carbon nanotube forest as function of growthtime.

Figure 2. (a)-(c) Carbon nanotube images controlled at the pointsindicated by gray arrows in inset of Figure 1. (d),(e) Carbonnanotube images grown for 6 and 30 min, respectively. (f) TEMimage showing a bundle of nanotubes with the upper part straightand the lower part curled.

Figure 3. Growth rate of carbon nanotubes as function of thegrowth length, calculated from the data in Figure 1.

864 Nano Lett., Vol. 3, No. 6, 2003

8/3/2019 L'Ancien d'Algrie, Jun 2011

3/3

4 is the transmission electron microscopy image of the

sample. Our CNT was grown via a base model mechanism

with Ni particles at the bottom of CNTs as shown in Figure

4a. The Ni particle was not fully enclosed at the bottom of

the CNT, as shown in Figure 4b. It was also confirmed that

the additional growth of nanotubes per a quarter or half

periods was found for some CNTs growth, showing a drag

in the interference oscillations curve. This can indicate that

the encapsulation of catalyst metal can be processed after

finishing the growth. It may be difficult to estimate the

dynamic change during growth using a TEM measurement

taken after finishing the growth period. In our CNTs based

on the base growth mechanism, bottom parts of CNTs were

shown to be relatively curly whereas the upper parts of the

CNTs appear in an aligned manner, as shown in Figure 2d,e,f.

This growth behavior indicates that CNTs in the constant

growth rate region were well aligned as well as straight, and

the curly CNTs were formed in the region of decreased

growth rate. It is clear that the growth termination brings

about the relatively curly form of CNTs under changed

growth conditions after the critical growth length is reached.

The straight part of the CNTs in Figure 2f is about 4 mlong, which corresponds to the point where the growth rate

decreases. More sophisticated studies are under way to unveil

the termination mechanism for the growth of carbon nano-

tubes.

An extrapolation of the linear fitting results in Figure 3indicates that the stabilized lengths of the carbon nanotube

have values ranging from 10 to 11.3 m under our optimizedgrowth conditions. The carbon nanotubes were allowed to

grow for 6 and 30 min for comparison, which results in a

length of about 10.5 m in both cases as shown in Figure2d,e, respectively. This is in agreement with the stabilized

length of a nanotube as calculated from linear fitting data.

In summary, we have presented an in situ optical monitor-

ing method for carbon nanotube growth. Growth rate

behaviors can be characterized and the length of carbon

nanotubes can be precisely controlled by in situ monitoring

of the interference oscillations. The termination length of

the CNTs was well predicted by the interpretation of

interference behavior. The upper portions of CNTs were

straight and the lower portions were curly , as related to the

decrease in the CNT growth rate based on a base growth

model. This in situ technique can be used as a method for

such growth control in future applications of nanotubes for

use in functional devices.

Acknowledgment. This study was performed under the

Daegu City Government 2002 Nano Project and was also

supported by the Korea Ministry of Information and Com-

munication.

References

(1) Dai, H.; Hafner, J. H.; Rinzler, A. G.; Colbert, D. T.; Smalley, R. E.Nature 1996, 384, 147.

(2) Wong, S. S.; Joselevich, E.; Woolley, A. T.; Cheung, C. L.; Lieber,C. M. Nature 1998, 394, 52.

(3) de Heer, W. A.; Chatelain, A.; Ugarte, D. Science 1995, 270, 1179.(4) Collins, P. G.; Zettl, Appl. Phys. Lett. 1998, 69, 1969.(5) Chung. D. S. et al. Appl. Phys. Lett. 2002, 80, 4045.(6) Tans, S. T.; Verschueren, A. R. M.; Dekker, C. Nature 1998, 393,

49.(7) Wei, B. Q.; Vajtai, R.; Jung, Y.; Ward, J.; Zhang, R.; Ramanath, G.;

Ajayan, P. M. Nature 2002, 416, 495.(8) Fan, S.; Chapline, M. G.; Flanklin, N. R.; Tombler, T. W.; Cassell,

A. M.; Dai, H. Science 1999, 238, 512.(9) Ren, Z. F.; Huang, Z. P.; Wang, D. Z.; Wen, J. G.; Xu, J. W.; Wang,

J. H.; Calvet, L. E.; Chen, J.; Klemic, J. F.; Reed, M. A. Appl. Phys.

Lett. 1999, 75, 1086.(10) Louchev, O. A.; Sato, Y.; Kanda, H. Appl. Phys. Lett. 2002, 80, 2752.(11) Louchev, O. A.; Sato, Y.; Kanda, H. Phys. ReV. E 2002, 66, 011601.(12) Bower, C.; Zhou, O.; Zhu, W.; Werder, D. J. S. Jin, Appl. Phys.

Lett. 2000, 77, 2767.

NL034212G

Figure 4. (a) Cross-sectional TEM image for the bottom of a well-aligned MWNT. (b) Magnified HRTEM image for the white circleregion of Figure 4a. The straight dotted line indicates the discon-tinuous carbon nanotubes at the bottom. The Ni particle was notfully enclosed at the bottom region marked in black dotted circle.

Nano Lett., Vol. 3, No. 6, 2003 865