Fabrication of aligned convex CNT field emission triode by MPCVD

evier.com/locate/diamond

Diamond & Related Material

Fabrication of aligned convex CNT field emission triode by MPCVD

Y.M. Wong a, W.P. Kang a,*, J.L. Davidson a, B.K. Choi a, W. Hofmeister b, J.H. Huang c

a Department of Electrical Engineering, Vanderbilt Univ., VU Station B 351661, Nashville, TN 37235-1661, USAb Department of Chemical Engineering, Vanderbilt Univ., Nashville, TN 37235, USA

c Department of Materials Science and Engineering, National Tsing Hua Univ., Hsinchu 300, Taiwan, ROC

Available online 26 October 2005

Abstract

In this work, vertically aligned carbon nanotubes (CNTs) were used to form a gated microcathode with a convex surface profile, being

selectively synthesized from Microwave Plasma Chemical Vapor Deposition (MPCVD) with nickel (Ni) as a catalyst. A single-mask

microfabrication process achieved an array of 10 Am�10 Am CNT microtriodes with self-aligned gate. The convex profile is important in

preventing cathode-gate leakage without resorting to more complicated fabrication processes or utilizing a gate over-etching approach. The main

mechanism for the formation of the convex-shaped CNT microcathodes was investigated and is proposed to be the result of plasma etching of

CNTs near the gate opening region due to higher plasma density during the growth process, leading to slower growth rate or shorter CNTs at the

circumferential area. Additionally, previous simulation work has predicted that this type of surface profile is beneficial for more quasi-uniform

electric field distribution on CNT tips. Field emission characteristics of the triode device were investigated, whereby a gate turn-on voltage as low

as 25 V was achieved. The low turn-on of the device is mainly due to the smaller gate aperture made possible by the convex-shaped CNT

microcathodes.

D 2005 Elsevier B.V. All rights reserved.

Keywords: MPCVD; Carbon nanotubes; Field emission; Electrical properties characterization

1. Introduction

Recently, gated CNTs field emission microcathodes or

triode devices have caught attention of researchers [1–11]

due to their low turn-on voltages and potential for high current,

high frequency and high power applications in vacuum

microlelectronics. Unlike diode devices, a field emission triode

is a three-terminal device. Due to the proximity of the gate

electrode to the electron emitting cathode, a gate voltage, often

less than 50 V is sufficient to cause a high electric field on the

emitters for the extraction of electron into vacuum, i.e. the

tunneling phenomena termed as field electron emission. The

field emission triode is an indispensable building block in the

development of high-speed, radiation, and temperature-im-

mune vacuum microelectronics and field emission displays

(FEDs).

Most of the reported CNT triode results thus far do not test

or reveal DC or AC transistor characteristics such as ampli-

0925-9635/$ - see front matter D 2005 Elsevier B.V. All rights reserved.

doi:10.1016/j.diamond.2005.09.022

* Corresponding author. Tel.: +1 615 322 0952; fax: +1 615 343 6614.

E-mail address: [email protected] (W.P. Kang).

fication factor (l), transconductance ( gm), anode resistance

(ra) and voltage gain (Av) of the device. The reported gate

turn-on voltages range from 10 to 60 V for as-grown CNT on

microfabricated triode [1–11]. Previously, we successfully

fabricated a CNT triode with a gate turn-on voltage of ¨40 V

by thermal CVD [1]. Since the CNTs grown were randomly

oriented, an over-etched gate structure [1] was adopted in

order to avoid cathode-gate leakage problems. As a result, the

large cathode-gate spacing (¨12 Am) led to high turn-on

voltage and triode characteristics of amplification factor ¨10

and transconductance ¨47 nS when configured as a triode

amplifier.

In order to overcome the gate leakage problems, which

limits the performance of the CNT triode, three approaches

have been suggested in the literature, namely (i) over-etched

gate electrode or reduced gate overhang [1,2], (ii) sidewall

protector [3–6], and (iii) post-growth processing which

includes utilizing chemical mechanical polishing (CMP)

technique [7], or plasma trimming of the grown CNTs [8]. In

this respect, the over-etched gate and sidewall protector

techniques did alleviate the gate leakage problems but at the

expense of higher gate turn-on voltages. Apart from the above,

s 15 (2006) 334 – 340

www.els

Y.M. Wong et al. / Diamond & Related Materials 15 (2006) 334–340 335

potential solution to the cathode-gate leakage problems is a

gated triode structure with a second gate or focus electrode

[9]. This dual-gate configuration is capable of reducing gate-

leakage currents as well as providing focused electron beams

but more complicated and costly fabrication processes are

involved. In addition, by operating the field emission triode in

a ‘‘collector-assisted’’ mode [12] or biasing the anode such

that the cathode is on the verge of emission [13] also

improves electron emission uniformity and gate current

leakage problem.

In this work, an alternative approach with gated convex

CNT microcathodes was performed to minimize cathode-gate

leakage problems. The CNT microtriode was fabricated with

self-aligned gate utilizing a single-mask microfabrication

process. Vertically aligned CNTs were synthesized selectively

inside the triode structure by MPCVD method with Ni as a

catalyst and without shorting the gate electrode. H2 plasma

pretreatment of the catalysts prior to the CNT synthesis

process was successfully employed to obtain a convex-shaped

surface profile, which is important in preventing cathode-gate

leakage without resorting to more complicated fabrication

processes or utilizing a gate over-etching approach. Field

emission characteristics of the aligned CNTs field emission

triode were investigated.

LPCVD & diffusion of poly-Si gate

SiO2

Thermal oxidation of Si substrate

SiO2

Patterning &RIE of poly-Si

SiO2

SiO2

Poly-Si

Photoresist n+-Si

SiO2

Fig. 1. Schematic diagrams of the single-mask fabrication process for the aligne

2. Experimental method

In this study, vertically aligned CNT microcathodes with

self-aligned gate were synthesized selectively inside the triode

mold utilizing MPCVD method. The single-mask microfabri-

cation process achieved an array of 10 Am�10 Am CNT

microtriodes with 2856 array cells and 20 Am array spacing. As

illustrated in the schematic fabrication process in Fig. 1, the

process began with a low-pressure CVD deposition of

polysilicon as the gate electrode on thermally oxidized highly

doped n-type silicon (100) substrate. A spin-on-diffusion

source was utilized to dope the polysilicon gate at high-

temperature (1050 -C) for good conductivity. The thickness of

the thermal oxide and the polysilicon gate layers was ¨1.5 Amand 0.8 Am, respectively.

After conventional photolithography patterning, the poly-

silicon gate was subsequently dry-etched by a gas mixture of

sulfur hexafluoride (SF6) and oxygen (O2) at 150 mTorr in a

reactive-ion-etch system (MRC). Next, the thermal oxide was

isotropically wet-etched with buffered-oxide-etch (BOE) to

obtain an undercut structure as illustrated in Fig. 1. A thin

film of titanium (Ti) ¨20 nm, a diffusion barrier layer and Ni

¨5–10 nm, acting as nanocluster catalytic centers for CNT

nucleation growth catalyst, were sputter deposited in sequence

Wet-etching of SiO2

(BOE)

SiO2

SiO2

Catalyst deposition& PR liftoff

SiO2

SiO2

SiO2

SiO2

MPCVD growth ofaligned CNTs

Convex CNTs

d CNT field emission triode amplifier with a convex-shaped surface profile.

Fig. 2. SEM micrographs of the aligned CNT microtriode grown with MPCVD

and H2 plasma pretreatment at 400 W microwave power and 400 -C (a) tilted

cross-sectional view of a single 10 Am�10 Am array cell, (b) high

magnification of the aligned CNTs, and (c) part of the 2856 array (3�3) with

20 Am array spacing.

Y.M. Wong et al. / Diamond & Related Materials 15 (2006) 334–340336

using a DC magnetron sputtering system (MRC) on the triode

sample without breaking the vacuum. The surplus catalyst on

the polysilicon gate was removed utilizing photoresist lift-off

technique by acetone.

The MPCVD system used to grow aligned CNTs in this

study is an ASTeX 1.5 kW microwave plasma system equipped

with an RF induction heater for substrate heating. The same

system has been routinely used to synthesize polycrystalline

and nanocrystalline diamonds by adopting different growth

parameters [14–16]. The triode sample was then transferred to

the CVD chamber and heated to 400 -C in flowing H2 with a

working pressure of 20 Torr. A critical H2 plasma pretreatment

of the catalysts at microwave power of 400 W was performed

for a specified duration of time. This pre-growth catalyst

pretreatment is important in obtaining the required convex-

shaped cathode structure. After the pretreatment, the substrate

temperature was raised to 650 -C, the synthesis temperature.

In the meantime, methane (CH4) gas and microwave power of

1 kW was applied instantly to initiate the CNTs growth for

¨60–90 s. The specified growth time was critical to obtain the

optimum height of the vertically aligned CNT cathodes without

cathode-gate shorting. The flow ratio of the gas mixtures, CH4/

H2 was maintained at 1:8 with a total flowrate of 135 sccm

during the CNTs growth.

The CNT samples were examined with a Hitachi 4200

scanning electron microscope (SEM) for the morphologies and

cross-sectional surface profile of the aligned CNTs grown, and

the catalyst size distribution. Raman scattering spectra was

taken with a JOBIN YVON micro-Raman LabRam system

with a 632.8 nm He–Ne laser as the light source. The CNT

samples were imaged with a 100� objective lens under the

microscope and a laser power of 1.1 mW at the sample.

In device characterization, the aligned CNT triode was

tested in a common emitter amplifier configuration for DC field

emission characteristics, as described in detail elsewhere [1].

The measurements were performed at room temperature in a

vacuum chamber evacuated to a base pressure of 10�6 Torr. A

standard test procedure began by applying a fixed anode

voltage, Va of 400 V to the anode or electrons collector made of

highly doped Si. A computerized data acquisition system using

Labview program was then employed to measure the anode/

collector current, Ia as a function of the applied gate voltage, Vg

at a constant Va.

3. Results and discussion

The SEM micrographs in Fig. 2 show the morphologies

and cross-sectional surface profile of the aligned CNTs grown

on the gated device with H2 plasma pretreatment of the

catalysts. A tilted cross-sectional view of a single 10 Am�10

Am array cell of the CNT microtriode with self-aligned gate is

depicted in Fig. 2(a). Vertically aligned CNTs with diameters

ranging from 15 to 30 nm were observed in Fig. 2(b). Fig.

2(c) shows part of the 2856 array of CNT microtriodes with

20 Am array spacing. A pre-growth H2 plasma pretreatment of

the sputtered catalysts was utilized successfully to achieve a

gated CNT cathode with a convex-shaped surface profile, i.e.

shorter nanotubes on the edges, as exhibited in Fig. 3(a). This

convex profile is important in preventing cathode-gate


processes such as sidewall protector [3–6] or gate over-

etching method [1,2] that result in higher turn-on voltage. It is

also predicted that the convex-shaped surface profile of the

aligned CNT microcathode may become Spindt-type [17], i.e.

cone-shaped if the gate aperture is reduced to ¨4 Am or less

in diameter. The cathode-gate spacing was ¨2.0 Am, as

estimated from the gap between the edge of the gate and the

edge of the CNT cathode, which has the highest electric field.

The integrity of the gate-cathode insulation of the fabricated

CNT field emission triode was checked by a Fluke 187

multimeter before device characterization. The resistance was

Fig. 3. Cross-sectional SEM micrographs of the aligned CNT microtriode grown with MPCVD (a) with H2 plasma pretreatment at 400 W microwave power and 400

-C, and (b) without H2 plasma pretreatment. Insets are high magnification.


found to be totally Fopen_, i.e. greater than 500 MV measura-

ble by the multimeter.

On the other hand, Fig. 3(b) shows a gated CNT

microcathode grown without H2 plasma pretreatment. Al-

though aligned CNTs were also observed, this cathode with a

concave-shaped surface profile, i.e. taller nanotubes on the

circumferential region, is prone to cathode-gate shorting that

limits the performance of the field emission triode device. A

similar growth phenomenon is also observed in [18]. They

attributed the higher growth rate of the CNTs to smaller grain

size of the catalysts on the circumferential area. The variation

in the initial thickness distribution which eventually determines

the grain size of the catalysts, between the middle and the

circumferential portions is most likely due to poor contact of

the shadow mask [18].

With respect to this work, the Ni catalytic thin film was

sputter-deposited into a triode structure, Fig. 4(a), with

photoresist as the lift-off mask. As a result, the height of the

sandwiched layers of thermal oxide and polysilicon resulted in

incomplete contact between the photoresist and the Si

substrate, leading to thinner catalyst layer or smaller grain size

of the catalyst on the circumferential area of the microcathode.

Two possible scenarios could have happened that result in the

convex-shaped CNT cathode. The first one is that with H2

plasma pretreatment, Ni catalysts on the circumferential area

were subjected to a stronger plasma treatment due to higher

microwave plasma density near the edges of the polysilicon

gate opening, resulting in local heating or higher temperature

on the affected area, and leading to larger grain size of the

catalysts. It has been known that when heated, high temper-

ature results in formation of nano-sized islands of the Ni

catalysts. In addition, higher temperature results in larger grain

size of the catalytic thin film due to agglomeration of the

smaller grains [19]. Therefore, it is believed that the shorter

CNTs were the cause of slower growth rate from the larger

catalytic grains on the circumferential region of the micro-

cathode, leading to a convex-shaped surface profile of the CNT

microcathode. Also, it has been shown that CNTs can be etched

by H2 plasma [8]. Therefore, the second conjecture is that the

growing CNTs were subjected to strong plasma etching at the

same time due to higher plasma density near the edges of the

polysilicon gate opening, leading to shorter CNTs on the

circumferential region. The strong plasma right underneath the

gate opening has the heaviest etching capability and thus

results in the shortest CNTs while longer ones were found

towards the middle region where the plasma is weaker.

Fig. 4. SEM micrographs showing the catalytic grain size distribution of (a) a single 10 Am�10 Am microtriode after heating up to 650 -C without CNTs synthesis,

high magnifications of the catalyst distribution (b) on the middle region (without H2 plasma pretreatment), (c) on the circumferential region (without H2 plasma

pretreatment), (d) on the middle region (with H2 plasma pretreatment), and (e) on the circumferential region (with H2 plasma pretreatment).


Further investigation of the catalyst thickness distribution

and grain size prior to CNTs synthesis was performed to verify

the mechanism involved in the synthesis of convex-shaped

CNT emitters. In this case, two triode molds were subjected to

the same CNTs synthesis processes without introduction of

CH4, without synthesis of CNTs. Fig. 4(a) shows the SEM

micrograph of a single array cell of the CNT triode without

CNT growth, while Fig. 4(b) to (e) display high magnification

of the catalyst size distribution on the middle and the

circumferential regions of the two samples. Fig. 4(b) shows

that the middle region of the Ni/Ti thin film (without H2 plasma

pretreatment) had not transformed completely into isolated

nanoislands but rather semi-continuous particles, which are not

favorable for catalytic growth of CNTs. This corresponds well

to the lack of CNTs found on the middle region of the

microcathodes, as exhibited in Fig. 3(b). Furthermore, scattered

nanoparticles were observed on the circumferential region and

their size and density increases towards the middle region, as

displayed in Fig. 4(c). The average size of the nanoparticles is

¨20–30 nm, which again corresponds well with the diameter

of the CNTs grown, as illustrated in Fig. 3(b). In addition, the

few randomly aligned CNTs spotted on the edges of the CNT

cathode is most likely due to the low density of the catalytic

nanoparticles. On the other hand, well-isolated nanoparticles

were observed on the entire cathode area of the Ni/Ti thin film

pretreated with H2 plasma pretreatment. The average catalyst

size ranges from 15 to 30 nm and 5 to 10 nm for the middle and

the circumferential portions, respectively. Surprisingly, the

catalysts on the circumferential area are not larger as expected

from the first postulation. As a result, it is believed that plasma

etching of the CNTs during the MPCVD growth may be the

prominent factor that leads to the convex-shaped CNT cathode.

Recently, simulations by SIMION 3D 7.0 software (Scien-

tific Instrument Services, Inc.) on the electric field distribution

-24.5

-24

-23.5

-23

-22.5

-22

-21.5

-21

-20.5

0 0.01 0.02 0.03 0.04 0.05 0.06

1/Vg

Ln(

I a/V

g2)

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

0 10 20 30 40

Vg (V)

Ia,

Ig (

µA

)

Ig

Ia

-24.5

-24

-23.5

-23

-22.5

-22

-21.5

-21

-20.5

0 0.02 0.04 0.061/Vg

Ln(

Ia/V

g2 )

Fig. 6. Plot of the anode currents, Ia and the gate currents, Ig vs. the gate

voltage, Vg of the aligned CNTs field emission triode amplifier at an applied

anode voltage, Va=400 V. Inset: the F–N plot of the corresponding data of Iavs. Vg.


in a nanotriode using CNTs as emitters are reported [20,21]. The

nanotriode model has a gate aperture of 75 nm and vertically

aligned CNTs with 2 nm diameter. Two types of emitters namely

aligned CNTs with the same height and variable height were

simulated. The latter is similar to the aligned CNT micro-

cathodes with a convex-shaped surface profile studied in this

work. As expected, it was found that the nanotriode with uniform

CNTs height exhibit substantial electric field screening and non-

uniformity of the field on the circumferential regions of the CNT

emitters. As a result, not all the CNTs are participating in the

field emission process and this led to over-active emitters with

high possibility of burnt-out. On the other hand, CNTs with

variable height display more quasi-uniform electric fields on the

CNT tips. There optimized CNT emitters are capable of

producing higher emission current density at the same electric

field level than the one with uniform height [20,21]. Ideally,

individual vertically aligned CNTs should be spaced apart by

twice their height to minimize field screening effects and to

further optimize the emitted current density [22]. However, such

precise CNT placement can only be achieved through advanced

electron-beam lithography technique [9].

The structural compositions of the MPCVD synthesized

aligned CNTs were analyzed by Raman spectroscopy. The

Raman spectrum of the CNTs grown with Ni catalysts is shown

in Fig. 5. In general, there are two distinctive peaks observed in

the Raman spectra of the CNT samples, D (¨1322 cm�1) and

G (¨1596 cm�1) modes, corresponding to the disorder-induced

D-band and the stretching mode in the graphite plane,

respectively [23,24]. The third peak located at 521 cm�1 is

due to the background Si substrate. The broad D-band peak is

attributed to the disorder-induced features caused by finite size

effects in sp2 carbon [23,24]. In addition, the D-line signal

intensity, I(D) of the CNT samples is stronger than the G-line

graphitic signal intensity, I(G). A I(G)/I(D) ratio less than 1

(¨0.59) indicates that the synthesized CNTs have a significant

amount of defects, which is confirmed by the SEM images in

Figs. 2 and 3. These defects however may be favorable for

promoting the field enhancement factor in the filed emission

process [23].

0

50

100

150

200

250

300

350

400

450

500

100 400 700 1000 1300 1600 1900

Raman shift (cm-1)

Inte

nsit

y (a

.u.)

1322

1596

Si

Fig. 5. Raman spectrum of the aligned CNTs grown with MPCVD utilizing Ni

as catalysts.

Fig. 6 shows a plot of the anode emission current, Ia and the

gate current, Ig vs. the applied gate voltage, Vg at a constant

anode voltage of 400 V, whereby a gate turn-on voltage as low

as 25 V was observed. This gate turn-on voltage is lower than

our previously reported triode synthesized by thermal CVD [1],

and compared favorably to other reported results [2–11]. The

gate current was found to be less than 10% of the anode

current, as shown in Fig. 6. Utilizing the same plasma

pretreatment technique, we have recently reported [25] Ig/Iaof less than 3% for a differential amplifier using a rectangular

array. The low gate currents observed in our cases demonstrate

that the convex-shaped CNT cathode profile is effective in

reducing cathode-gate leakage and gate interception of anode

emission current. Thus far, most of the reported CNT triodes

[1–5,7–11] either have high gate leakage currents or did not

report the gate current data at all, except Ref. [6] demonstrated

Ig/Ia of ¨2.5% from a gated array with sidewall SiO2 spacer.

The linearity of the Fowler–Nordheim (F–N) plot of the

corresponding emission data of Ia vs. Vg (inset of Fig. 6)

confirms that electron tunneling behavior is observed [26,27].

The local electric field on an emitting surface is given by F =bEwhere E is the composite macroscopic applied electric field from

the gate and the anode, and b is the field enhancement factor. In

this case, E can be expressed as [1,28–30]:

E ¼Vg þ Va=l� �

dð1Þ

where Vg, Va, l and d are the applied gate voltage, anode

voltage, amplification factor and cathode-gate spacing, respec-

tively. Based on Eq. (1), the linearity of the F–N plot in Fig. 6

further suggest that electron emission of the CNT triode is solely

extracted by Vg and the anode acting primarily as an electron

collector. Detailed DC large-signal and AC small-signal


characterization of the triode amplifier will be presented

elsewhere [31].

4. Conclusions

Vertically aligned CNT field emission triodes were success-

fully fabricated with a single-mask process and MPCVD with

Ni as catalyst. H2 plasma pretreatment of the catalysts prior to

the CNT synthesis process was employed to obtain a convex-

shaped surface profile, critical in preventing cathode-gate


processes or utilizing a gate over-etching approach. In addition,

literature shows that CNTs with variable height display more

quasi-uniform electric fields on the CNT tips and favorable for

high current density. A plausible explanation for the formation

of aligned CNTs microcathode with a convex-shaped surface

profile is the concurrent synthesis and plasma etching of CNTs

during the growth process due to higher plasma density near

the gate opening region, leading to slower growth rate or

shorter CNTs on the circumferential area. The vertically

aligned CNT triode shows a lower gate turn-on voltage than

the CNT triode grown with thermal CVD method [1], in part

due to the smaller gate aperture made possible by the convex-

shaped CNT microcathodes. This CNT triode amplifier could

be a candidate for high power, high gain and high frequency

amplifier applications that requires radiation-hardness and

temperature-immune capability.

References

[1] Y.M. Wong, W.P. Kang, J.L. Davidson, W. Hofmeister, S. Wei, J.H.

Huang, J. Vac. Sci. Technol., B 23 (2005) 868.

[2] K.J. Chen, W.K. Hong, C.P. Lin, K.H. Chen, L.C. Chen, H.C. Cheng,

IEEE Electron Device Lett. 22 (2001) 516.

[3] S.D. Ahn, Y.H. Song, S.Y. Choi, J.B. Park, J.I. Sohn, S.H. Lee, J.H. Lee,

K.I. Cho, Proceedings of the 14th International Vacuum Microelectronics

Conference, 12–16 Aug., Davis, CA, 2001, p. 55.

[4] Y. Huh, J.Y. Lee, J.H. Lee, T.J. Lee, S.C. Lyu, C.J. Lee, Chem. Phys. Lett.

375 (2003) 388.

[5] Y.Y. Jang, C.H. Choi, B.K. Ju, J.H. Ahn, Y.H. Lee, Thin Solid Films 436

(2003) 298.

[6] D.S.Y. Hsu, Appl. Phys. Lett. 80 (2002) 2988.

[7] W.J. Yu, Y.S. Cho, G.S. Choi, D.J. Kim, Y.H. Song, J.H. Lee, Diamond

Relat. Mater. 13 (2004) 1017.

[8] S. Kang, C. Bae, W. Son, M.H. Kim, J. Yi, A. Chang, J.J. Kim, C.R. Lee,

Technical Digest of the 16th International Vacuum Microelectronics

Conference, 7–11 July, Osaka, Japan, 2003, p. 51.

[9] M.A. Guillorn, X. Yang, A.V. Melechko, D.K. Hensley, M.D. Hale, V.I.

Merkulov, M.L. Simpson, R. Baylor, L. Gardner, D.H. Lowndes, J. Vac.

Sci. Technol., B 22 (2004) 35.

[10] L. Gangloff, E. Minoux, K.B.K. Teo, P. Vincent, V. Semet, V.T. Binh, M.H.

Yang, R.G. Lacerda, I.Y.Y. Bu, G. Pirio, J.P. Schnell, D. Pribat, D.G.

Hasko, G.A.J. Amaratunga, W.I. Milne, P. Legagneux, Nano Lett. 4 (2004)

1575.

[11] D.S.Y. Hsu, J.L. Shaw, J. Vac. Sci. Technol., B 23 (2005) 694.

[12] H.H. Busta, J.E. Pogemiller, B.J. Zimmerman, IEEE Trans. Electron

Devices 40 (1993) 1537.

[13] D.M. Garner, J. Vac. Sci. Technol., B 22 (2004) 1250.

[14] W.P. Kang, J.L. Davidson, A. Wisitsora-at, Y.M. Wong, R. Takalkar, K.

Subramanian, D.V. Kerns, W.H. Hofmeister, Diamond Relat. Mater. 14

(2005) 685.

[15] A. Wisitsora-at, W.P. Kang, J.L. Davidson, M. Howell, W. Hofmeister,

D.V. Kerns, J. Vac. Sci. Technol., B 21 (2003) 1671.

[16] W.P. Kang, J.L. Davidson, Y.M. Wong, K. Holmes, Diamond Relat.

Mater. 13 (2004) 975.

[17] C.A. Spindt, C.E. Holland, R.D. Stowell, Appl. Surf. Sci. 16 (1983) 268.

[18] Y.C. Choi, Y.M. Shina, D.J. Baea, S.C. Lima, Y.H. Lee, B.S. Lee,

Diamond Relat. Mater. 10 (2001) 1457.

[19] Y. Huh, J.Y. Lee, J. Cheon, Y.K. Hong, J.Y. Koo, T.J. Leed, C.J. Lee, J.

Mater. Chem. 13 (2003) 2297.

[20] D. Nicolaescu, L.D. Filip, S. Kanemaru, J. Itoh, J. Vac. Sci. Technol., B 21

(2003) 366.

[21] D. Nicolaescu, L.D. Filip, S. Kanemaru, J. Itoh, Jpn. J. Appl. Phys. 43

(2004) 485.

[22] 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.

[23] Y.M. Wong, S. Wei, W.P. Kang, J.L. Davidson, W. Hofmeister, J.H.

Huang, Y. Cui, Diamond Relat. Mater. 13 (2004) 2105.

[24] A.M. Rao, A. Jorio, M.A. Pimenta, M.S.S. Dantas, R. Saito, G.

Dresselhaus, M.S. Dresselhaus, Phys. Rev. Lett. 84 (2000) 1820.

[25] Y.M. Wong, W.P. Kang, J.L. Davidson, B.K. Choi, J.H. Huang, Oral

presented at the 18th International Vacuum Nanoelectronics Conference,

Oxford, U.K., July 10–14, 2005.

[26] R.H. Fowler, L.W. Nordheim, Proc. R. Soc. London 119 (1928) 173.

[27] L.W. Nordheim, Proc. R. Soc. London, A 121 (1928) 626.

[28] A. Wisitsora-at, W.P. Kang, J.L. Davidson, C. Li, D.V. Kerns, M. Howell,

J. Vac. Sci. Technol., B 21 (2003) 1665.

[29] W.J. Orvis, C.F. McConaghy, D.R. Ciarlo, J.H. Yee, E.W. Hee, IEEE

Trans. Electron Devices 36 (1989) 2651.

[30] R.E. Neidert, P.M. Phillips, S.T. Smith, C.A. Spindt, IEEE Trans. Electron

Devices 38 (1991) 661.

[31] Y.M. Wong, W.P. Kang, J.L. Davidson, B.K. Choi, W. Hofmeister, J.H.

Huang, Abstract of the 10th International Conference on New

Diamond Science and Technology (ICNDST), Tsukuba, Japan, May

11–14, 2005.

Fabrication of aligned convex CNT field emission triode by MPCVD

Documents