Enhancement mechanism of field electron emission properties in hybrid carbon nanotubes with tree- and wing-like features

Article abstract of DOI:10.1016/j.jssc.2009.10.005

In this work, the tree-like carbon nanotubes (CNTs) with branches of different diameters and the wing-like CNTs with graphitic-sheets of different densities were synthesized by using plasma enhanced chemical vapor deposition. The nanostructures of the as-

Full text of DOI:10.1016/j.jssc.2009.10.005

ARTICLE IN PRESS
Journal of Solid State Chemistry 182 (2009) 3393–3398
Contents lists available at ScienceDirect
Journal of Solid State Chemistry
journal homepage: www.elsevier.com/locate/jssc
Enhancement mechanism of field electron emission properties in hybrid
carbon nanotubes with tree- and wing-like features
a,b
b,Ã
c
a,Ã
b
G.M. Yang , C.C. Yang , Q. Xu , W.T. Zheng , S. Li
a
Department of Materials Science, State Key Laboratory of Superhard Materials, and Key Laboratory of Automobile Materials of MOE, Jilin University, Changchun 130012, PR China
School of Materials Science and Engineering, The University of New South Wales, NSW 2052, Australia
b
c
Changchun Institute of Technology, Changchun 130021, PR China
a r t i c l e i n f o
a b s t r a c t
Article history:
In this work, the tree-like carbon nanotubes (CNTs) with branches of different diameters and the wing-
like CNTs with graphitic-sheets of different densities were synthesized by using plasma enhanced
chemical vapor deposition. The nanostructures of the as-prepared hybrid carbon materials were
characterized by scanning electron microscopy and transmission electron microscopy. The structural
dependence of field electron emission (FEE) property was also investigated. It is found that both of the
tree- and wing-like CNTs exhibit a lower turn-on field and higher emission current density than the
pristine CNTs, which can be ascribed to the effects of branch size, crystal orientation, and graphitic-
sheet density.
Received 17 June 2009
Received in revised form
4
October 2009
Accepted 11 October 2009
Available online 17 October 2009
Keywords:
Branch
Density
&
2009 Elsevier Inc. All rights reserved.
Graphitic sheets
RF-PECVD
1
. Introduction
distribute in the tree-like CNT composite, which depresses their
FEE properties. One way to solve these problems is to grow the
thick CNTs as supports [8,9]. The other way is to prepare the wing-
like CNTs [10,11]. However, up to date, the FEE properties of the
wing-like CNTs have not been investigated, and the mechanism
underlying the enhancement of FEE properties in the tree- and
wing-like CNTs is unclear. This has impeded the rapid develop-
ment of CNT-based field-emission devices.
In this work, both the tree- and wing-like CNTs were
synthesized using radio frequency (RF)-PECVD, where (1) the
branch diameters of the tree-like CNTs are controlled by changing
the quantity of ferrocene contents; and (2) the graphitic sheet
density of the wing-like CNTs is controlled by the deposition time.
The effects of the branch diameter, crystal orientation, and
graphitic sheet density on the FEE properties of the synthesized
CNTs were studied. The growth mechanism of the wing-like CNTs
is also investigated.
In the past two decades, carbon nanotubes (CNTs) as one of the
most promising candidates in field emission devices have drawn
great attentions due to their unique physicochemical properties,
small curvature radius, high thermal stability, high electrical
conductivity, etc. [1–3]. To develop CNT-based field emission
devices, low turn on field, high current densities, and the
uniformity over a large area need to be achieved. For this purpose,
a number of experimental efforts have been implemented to
synthesize the hybrid CNTs with the enhancing field electron
emission (FEE) properties [4–8]. For example, arc discharge [5],
plasma enhanced chemical vapor deposition (PECVD) and thermal
CVD [2], nano-agglomerate fluidized bed reactor and floating
catalytically CVD (NAFBR-FCCVD) [6], and direct current (dc)-CVD
[
7] have been widely used for the preparation of tree-like CNT
composites with CNT branches on the surface of the pristine CNTs.
However, most of these efforts are facing two challenging issues,
i.e., a low formation efficiency of hybrid structures and un-
controllable synthesis process, which may be due to that (1) the
aggregation of pre-loaded nanosized particles is difficult to be
eliminated completely, thus restraining the growth of the
branches on the CNTs; and (2) the prepared CNTs randomly
2. Experimental section
The tree-like CNTs are defined as thin CNTs grown on the trunk
or tips of pristine CNTs, whereas the graphitic sheets on CNTs are
named as wing-like CNTs in this work. The tree-like CNTs were
prepared by the RF-PECVD and Fig. 1 shows
illustration of the RF-PECVD apparatus used in our experiments
RF power: 13.56 MHz) and the vapor-phase transport process of
Fe nanoparticles decorated on vertically aligned CNTs. To
a schematic
Ã
Corresponding authors. Fax: +612 93855956 (C.C. Yang); +86 43185168246
(
W.T. Zheng).
E-mail addresses: ccyang@unsw.edu.au (C.C. Yang), wtzheng@jlu.edu.cn
W.T. Zheng).
(
(
0
022-4596/$ - see front matter & 2009 Elsevier Inc. All rights reserved.
doi:10.1016/j.jssc.2009.10.005

ARTICLE IN PRESS
3
394
G.M. Yang et al. / Journal of Solid State Chemistry 182 (2009) 3393–3398
synthesize the tree-like CNTs with different branch diameters,
high-purity ferrocene powders (0.005, 0.01, and 0.02 g,
respectively) were put into a ceramic tube with a diameter of
2 4
H gas is 80/20 sccm. After 20-min deposition, CH inlet was
shutoff and the system was cooled down to room temperature.
To grow the wing-like CNTs, Ti thin film with a thickness of
ꢀ20 nm was deposited on Si (100) substrate using magnetron
sputtering, and then the Co thin film with a thickness of ꢀ15 nm
as a catalyst was coated on the surface of Ti thin film. In order to
compare the catalyst activity, the Ni as catalyst is also used to
grow CNTs. The RF-PECVD chamber was evacuated to a pressure of
8.0 Pa, and then pure hydrogen (99.99%) at the flow rate of
20 sccm was then introduced into the chamber. After the Si
6
mm through a small hole at the center of one end of the tube.
The hole was then covered by a piece of Al plate tightened with Al
wires to avoid rapid decomposition of the ferrocene. The as-
prepared pure CNTs on Si (10 0) substrates were placed in the
downstream site along the vapor flow direction at a distance of
4
mm from the ceramic tube. During the deposition, a radio
frequency power of 230 W was utilized and the substrate
temperature was kept at 800 1C. The flow rate ratio of CH and
4
4
substrate was heated to 800 1C for 40 min, pure CH (99.99%) with
a gas flow rate of 80 sccm was purged into the chamber. A working
pressure of 800 Pa and a radio frequency power of 230 W were
utilized to grow the wing-like CNTs with the deposition time of
10, 15, and 30 min, respectively. After the deposition, the CH
4
inlet
was shutoff and the system was cooled to room temperature in a
H
2
atmosphere.
The as-prepared materials were characterized by scanning
electron microscopy (SEM) (JEOL JSM-6700F), Raman spectro-
scopy (Renishaw-inVia with 514 nm excitation wavelength),
transmission electron microscope (TEM) (Hitachi H-8100), and
high resolution transmission electron microscope (HRTEM) (JEOL
TEM-2010). The TEM and HRTEM samples were prepared by
scratching the surface of the as-prepared sample using a sharp
tungsten tip and depositing the collected material on a copper
TEM grid. The FEE properties were measured using a parallel plate
configuration, an indium-tin-oxide (ITO) anode and a cathode (the
sample) in a vacuum chamber maintained at a pressure of
1
.2 ꢁ10ꢂ7 Pa.
3. Results and discussion
Figs. 2(a–c) show the SEM images of tree-like CNTs with
branches having different diameters, from which the diameters
and lengths of pristine CNTs are about 15–70nm and 2 m,
Fig. 1. Vapor-phase transport process for Fe nanoparticles decorated on vertically
aligned pristine CNTs.
m
Fig. 2. SEM images of tree-like CNTs with branches having different diameters: (a) 4, (b) 6, and (c) 20 nm, respectively, and (d) the corresponding TEM image of branch and
CNT denoted by arrows in (b).

ARTICLE IN PRESS
G.M. Yang et al. / Journal of Solid State Chemistry 182 (2009) 3393–3398
3395
2
ꢂ1
respectively. After the CNTs are decorated with catalyst (Fe
nanoparticles), the nucleation and growth of thin CNTs around
the particles take place. The branches of CNTs are on the surfaces
or the tips of as-prepared CNTs, whose diameters are about 4, 6,
and 20 nm (Figs. 2(a, b, c), respectively), respectively. Fig. 2(d) is a
typical TEM image of the tree-like CNTs with branches having a
diameter of 6 nm. From SEM images, the branch diameters of CNTs
increase with increasing ferrocene content. In this work, when the
graphitic D-band at 1350 cm , and G-band related to in-plane sp
vibrations at 1580 cm [11,15,16]. The disorder D-band in CNTs is
primarily caused by the defects in the tube wall, such as bending
in the CNTs, the finite size of crystalline domains, sp -hybridized
ꢂ1
3
bonds, and functional groups created by the oxidation [17–19].
D G
The intensity ratio of D- over G-band, R=I /I , is estimated, which
characters the degree of disorder in the structures of
carbonaceous materials. The R values for the pristine CNTs, and
the tree-like CNTs with a branch diameter of 4, 6, and 20 nm,
respectively, are 0.65, 0.72, 0.81, and 0.96, respectively. It is
evident that the degree of disorder in the walls of CNTs
substantially increases with an increase in the branch diameters
or ferrocene content. All tree-like CNTs have higher disorder than
the pristine CNTs, which is attributed to the Fe nanoparticles
coated on the walls of CNTs in the vapor phase transport process.
The Fe atoms have a coordinative affinity to the surface of CNTs
and dissolve the walls of CNTs at high temperature. In addition,
the walls of CNTs are oxidated by oxygen atoms or clusters from
ferrocene decomposition. As a result, the surface of CNTs is
oxidated more severely with an increase in ferrocene content.
Moreover, the C–H bonds on the walls of CNTs could be deformed
by the hydrogen atoms existing in the chamber at a temperature
of 800 1C and the hydrogen plasma etching on the pristine CNTs
during the pyrolysis of ferrocene. This causes a hybridized carbon
4
CH content reaches to a critical point, carbon atoms quickly
become saturated for the Fe nanoparticles due to the solubility
limit of carbon in Fe and the segregation of carbon atoms occurs,
which determines the type of final composite products. In
the process of the segregation of carbon atoms, the CNTs
grow continuously [12,13]. It has been found that the hydrogen
atmosphere, carbon source, and catalyst play important roles
in the growth of tree-like CNTs [14]. In this work, although the
2 4 2
flow rate ratio of H /CH is high, which leads to that ionized H
etches the CNT branches severely due to the hydrogen plasma
irradiation, CNT branches still grow from the surface of CNT
trunks. This may be associated with that small catalyst Fe
particles are active.
Fig. 3 exhibits the normalized Raman spectra for pristine CNTs
ꢂ1
and tree-like CNTs in the range of 0–2000 cm
.
Two
characteristic bands can be found in each spectrum, i.e., disorder
2
3
transfer from sp to sp on the walls of CNTs during the deposition
process. Such a phenomenon also contributes to the higher R
values of tree-like CNTs.
Figs. 4(a–c) show the SEM images of the Co-catalyzed wing-
like CNTs with graphitic sheets having different densities grown
at 800 1C for 10, 15, and 30 min, respectively, in which the insets
in Figs. 4(a) and (c) show the lengths of the corresponding CNTs.
The diameter and length of the CNTs are about 30 nm and 2 mm,
respectively. With increasing deposition time, the density of
the graphitic sheets on CNTs increases, while the average length
of the CNTs is almost time-independent. No nanosheets grow
on the CNTs using Ni as a catalyst in this work although the
synthesis conditions are the same as those for depositing wing-
like CNTs for 30 min using Co as a catalyst, which is a consequence
of that the catalyst Ni is more active than Co. Thus the cata-
lysts may play an important role in the growth of graphitic
nanosheets.
Fig. 5 exhibits the normalized Raman spectra for wing-like
CNTs using Co as a catalyst grown for (a) 10, (b) 15, and (c) 30 min,
respectively. Two characteristic bands, D and G, can also be
observed in each spectrum, which is similar to those for tree-like
CNTs. The R values for the wing-like CNTs grown for 10, 15, and
30 mins, respectively, are 0.59, 0.82, and 0.97, respectively,
showing a substantial increase in the degree of disordering in
the walls of CNTs as deposition time increases. This may be caused
2
3
by the partial transfer of hybridized carbon from sp to sp on the
walls of CNTs during deposition. The difference in R-value
between the wing-like CNTs grown for 15 and 30 min is much
smaller than that between for 10 and 15 min, although the time
interval is longer for the former. This implies that the formation of
nanosheets on the walls of CNTs can protect CNT walls from a
further damage.
Figs. 6(a) and (b) are the TEM images for the wing-like CNTs
grown for 15 and 30 min, respectively, where well-dispersed
graphitic sheets grow on the sidewalls of CNTs. The nanosheets
emanate from the cylindrical surface of the CNTs and protrude up
to about 20 nm from the CNT surface at an angle of 20–1201 along
with the tube axis. Moreover, Fig. 6(b) evidences the hollow
structure of the as-prepared CNTs. Figs. 6(c) and (d) show the TEM
images for the nanosheets denoted by arrows in Fig. 6(b), from
which the nanosheets are roughly ordered and seem to connect
continuously with the walls of the CNTs.
Fig. 3. Raman spectra of (a) pristine CNTs (Curve A) and tree-like CNTs with
branches having different diameters: (b) 4, (c) 6, and (d) 20 nm, respectively
(Curves B, C, and D).

ARTICLE IN PRESS
3
396
G.M. Yang et al. / Journal of Solid State Chemistry 182 (2009) 3393–3398
Fig. 5. Raman spectra of wing-like CNTs grown for (a) 10, (b) 15, and (c) 30 min,
respectively.
impeding the further growth of CNTs. In this case, the carbon
radicals inserted into the side walls/tips of the tubes form more
and more graphite sheets, as shown in Figs. 4(b) and (c). On the
other hand, the high temperature is also one of the important
factors for forming graphitic nanosheets since carbon nanomater-
ials prefer to form amorphous structure at low temperature [10].
Moreover, the hydrogen in the plasma is also critical for forming
nanosheets. In addition, during the growth of CVD nanotubes,
structural defects and non-hexagonal member rings (topological
lattice defects) are always observed. These defects are highly
reactive and sensitive to the plasma etching, and therefore may
provide nucleation sites for nanosheets [20,21]. It seems that
there are two competitive effects to influence the nanosheets
growth. One is the deposition of the carbon-based ions, and
another is the etching effect caused by the hydrogen ions. With
low hydrogen content in the discharge gases, the deposition rate
of carbon-based ions is higher than the etching rate, resulting in
that the graphite sheets form.
Fig. 4. SEM images of Co-catalyzed wing-like CNTs tilted at 451 grown for (a) 10,
b) 15, and (c) 30 min, respectively, from which the insets (a-1) and (c-1) exhibit
(
the corresponding SEM images of (a) and (c), tilted at 701.
Fig. 7(a) plots the FEE current density as a function of electric
field for pure CNTs (A) and the tree-like CNTs with branches
having a diameter of 4 (B), 6 (C), and 20 nm (D), respectively, as
well as the wing-like CNTs grown for 10 (E), 15 (F), and 30 min (G),
A competitive growth mechanism between CNTs and graphite
sheets is proposed as follows: at the initial stage of deposition,
CNTs grow very fast due to the high activity of catalyst Co. With
increasing deposition time, the activity of catalyst Co is lowered
and the abundant carbon radicals rapidly accumulate, thus
respectively. The turn-on electric fields, which are corresponding
2
to an electron emission density of 10
m
A/cm , can be determined

ARTICLE IN PRESS
G.M. Yang et al. / Journal of Solid State Chemistry 182 (2009) 3393–3398
3397
Fig. 6. TEM images of wing-like CNTs grown for (a) 15 and (b) 30 min, respectively, and (c) and (d) are the corresponding TEM images of nanosheets denoted by arrows in
b).
(
in Fig. 7(a). The results and the characteristic parameters
associated with the FEE property of CNTs are listed in Table 1.
From this table, the turn-on fields for samples B, C, and D are all
much lower than that for sample A, indicating that the tree-like
CNTs have much better FEE properties compared to pure CNTs.
This is attributed to the tip effect in the tree-like CNTs. It is
reported that the amplification factor of the tip is approximately
deposition time increases up to 30 min, too many nanosheets
almost cover the tips of CNTs, impeding the FEE. Consequently,
sample F has better FEE properties than sample G. Furthermore,
too many nanosheets produced by relatively long deposition time
(e.g., 30 min) shield all surfaces of CNTs, causing an increase in the
diameters of CNTs [3] and a decrease in the tip emission of the
CNTs, as well as an increase in the electrostatic screen effect [23].
This result in a depressed emission performance compared to
pure CNTs.
125 times as large as that of the nanotube body [22]. As shown in
Fig. 2(d), the branches appear on the as-prepared CNTs for our
obtained samples, acting as the tips. These branches provide more
electrons emission sites to enhance the FEE properties of tree-like
CNTs than pure CNTs. In general, the emission current density is
inversely proportional to the square of the emitter radius [3]. In
this work, the FEE property of sample B should be better than that
of C due to that sample B has a smaller branch diameter. However,
the turn-on field of the sample C is lower than that of either
sample B or D. Also, sample C has the highest emission current
density among seven samples. This can be explained by the well
oriented branches in sample C. From Table 1, it is worth noting
that sample B has a better FEE property than sample D, which can
be attributed to that sample B has much smaller branch (4 nm)
than D (20 nm).
As shown in Fig. 7(b), the Fowler–Nordheim (F–N) plots of
samples A–G observe the Fowler–Nordheim equation as follows:
2
2
3=2=
ꢂ6
ꢂ2
lnðJ=E Þ ¼ lnðA
b
=
j
Þ ꢂ B
j
b
E, where A=1.54 ꢁ10 A eV V
3
ꢂ3/2
ꢂ1
3/2
and B=6.83 ꢁ10 eV
V
mm
,
b
= ꢂB
j
/SFN is the field en-
hancement factor with SFN being the slope of the FN plots, and
the work function of emitters. With CNT=5.0 eV, the values are
estimated and listed in Table 1, from which the value of the
j is
j
b
b
pristine CNTs is much smaller than that of either tree- or wing-
like CNTs except for sample G. The reason is that most of the tips
on CNTs for sample G are covered by too many nanosheets, as
noted before.
From Fig. 7(a) and Table 1, the wing-like CNTs grown for 15 min
4. Conclusions
(sample F) have the highest emission current density among
samples E, F, and G. Compared with sample E, the graphite sheets
in sample F are well mixed with CNTs, resulting in the formation
of more tips on the sidewall or tips of CNTs. This leads to that
sample F has better FEE properties than sample E. However, as
A simple technique has been developed to synthesize the tree-
and wing-like hybrid CNTs using PECVD. The FEE properties for
these hybrid materials have been significantly enhanced com-
pared with pristine CNTs. The diameters of the branches and

ARTICLE IN PRESS
G.M. Yang et al. / Journal of Solid State Chemistry 182 (2009) 3393–3398
Table 1
3
398
Characteristic parameters of samples A–G associated with their FEE properties.
ꢂ1
ꢂ2
)
Sample D or T
Alignment
R
V (V
mm
)
J (
mA cm
b
3
4
4
4
4
4
A
B
C
D
E
/
/
0.65
3
300
7.5 ꢁ 10
3.1 ꢁ 10
4.2 ꢁ 10
1.6 ꢁ 10
1.4 ꢁ 10
3.9 ꢁ 10
4 nm
6 nm
20 nm
10 min
15 min
30 min
Not good
0.72 1.7
0.81 1.3
0.96 2.1
0.59 2.3
0.82 1.6
0.97 2.9
470
633
410
376
477
271
Good
Not good
/
/
/
F
3
G
7 ꢁ 10
D denotes the branch diameter, T deposition time, R crystallinity, V turn on field,
ꢂ1
J current density at 8 V
mm
, and b field enhancement factor.
Acknowledgments
The authors would like to thank the supports from National
Natural Science Foundation of China (Grant nos. 50832001 and
50525204), and special Ph.D. program from Chinese Education
Ministry (Grant no. 200801830025).
References
[
[
1] S. Iijima, Nature 354 (1991) 56.
2] H.J. Kim, I.T. Han, P.Y. Jun, J.M. Kim, J.B. Park, B.K. Kim, N.S. Lee, Chem. Phys.
Lett. 396 (2004) 6.
[
[
3] H.Y. Jung, S.M. Jung, L. Kim, J.S. Suh, Carbon 46 (2008) 969.
4] L.P. Biro, Z.E. Horvath, G.I. Mark, Z. Osvath, A.A. Koos, A.M. Benito, W. Maser,
P. Lambin, Diamond Relat. Mater. 13 (2004) 241.
[
[
[
5] D. Zhou, S. Seraphin, Chem. Phys. Lett. 238 (1995) 286.
6] H. Yu, Z. Li, G. Luo, F. Wei, Diamond Relat. Mater. 15 (2006) 1447.
7] Y. Abdi, S. Mohajerzadeh, J. Koohshorkhi, M.D. Robertson, C.M. Andrei, Carbon
4
6 (2008) 1611.
8] X. Sun, R. Li, B. Stansfield, J.P. Dodelet, S. Desilets, Chem. Phys. Lett. 394 (2004)
66.
9] G.M. Yang, X. Wang, Q. Xu, S.M. Wang, H.W. Tian, W.T. Zheng, J. Solid State
Chem. 182 (2009) 966.
[
[
2
[
[
[
[
[
[
10] S. Trasobares, C.P. Ewels, J. Birrell, O. Stephan, B.Q. Wei, J.A. Carlisle, D. Miller,
P. Keblinski, P.M. Ajayan, Adv. Mater. 16 (2004) 610.
11] A. Malesevic, S. Vizireanu, R. Kemps, A. Vanhulsel, C.V. Haesendonck,
G. Dinescu, Carbon 45 (2007) 2932.
12] W.Z. Li, S.S. Xie, L.X. Qian, B.H. Chang, B.S. Zou, W.Y. Zhou, Science 274 (1996)
1
701.
13] M. Meyyappan, L. Delzeit, A. Cassell, D. Hash, Plasma Sources Sci. Technol. 12
2003) 205.
14] J.W. Liu, X. Wang, W.T. Zheng, J.X. Li, Q.F. Guan, Y.D. Su, J.L. Qi, Q. Jiang, Carbon
5 (2007) 668.
15] E.T. Thostenson, W.Z. Li, D.Z. Wang, Z.F. Ren, T.W. Chou, J. Appl. Phys. 91
2002) 6034.
Fig. 7. (a) Field emission current density as a function of electric field for (A) pure
CNTs, tree-like CNTs with branches having different diameters: (B) 4, (C) 6, and (D)
(
2
3
0 nm, respectively, as well as wing-like CNTs grown for (E) 10, (F) 15, and (G)
0 min, respectively, and (b) Fowler–Nordheim (F–N) plots corresponding to
4
samples A–G.
(
[
[
16] Q.T. Li, Z.C. Ni, J.L. Gong, D.Z. Zhu, Z.Y. Zhu, Carbon 46 (2008) 434.
17] W.Z. Li, H. Zhang, C.Y. Wang, Y. Zhang, L.W. Xu, K. Zhu, S.S. Xie, Appl. Phys. Lett.
70 (1997) 2684.
branch alignment in CNTs are the main factors to influence the
FEE properties of tree-like CNTs. In contrast, the graphitic-sheet
density in CNTs may play a key role in determining the FEE
properties of wing-like CNTs. The findings in this work provide a
design principle for developing CNT-based nanomaterials with
superior field electron emission properties and new insight into
their potential applications as field emission devices.
[18] A. Hirsch, Angew. Chem. Int. Ed. 41 (2002) 1853.
[
[
19] A.C. Ferrari, J. Robertson, Phys. Rev. B 61 (2000) 14095.
20] A.V. Krasheninnikov, K. Nordlund, M. SirviO, E. Salonen, J. Keinonen, Phys. Rev.
B 63 (2001) 245405.
[21] J.A.V. Pomoell, A.V. Krasheninnikov, K. Nordlund, J. Keinonen, J. Appl. Phys. 96
2004) 2864.
(
[
[
22] G. Zhou, W. Duan, B. Gu, Appl. Phys. Lett. 79 (2001) 836.
23] L. Nilsson, O. Groening, C. Emmenegger, O. Kuettel, E. Schaller, L. Schlapbach,
H. Kind, J.-M. Bonard, K. Kern, Appl. Phys. Lett. 76 (2000) 2071.