Wang Zh.M. One-Dimensional Nanostructures

Подождите немного. Документ загружается.

x Contributors

Suklyun Hong

Department of Physics and Institute

of Fundamental Physics,

Sejong University, Seoul 143–747,

Korea

Byoung-Kye Kim

Department of Physics,

Chonbuk National University,

Jeonju, Korea

Hyo-Suk Kim

Department of Physics,

Chonbuk National University,

Jeonju, Korea

Ju-Jin Kim

Department of Physics,

Chonbuk National University,

Jeonju, Korea

Hideo Kohno

Graduate School of Science, Osaka

University 1-1 Machikaneyama,

Toyonaka, Osaka 560-0043, Japan

Thierry Laroche

Institut Charles Delaunay, Universit

ede

technologie de Troyes, CNRS FRE

2848 Laboratoire de Nanotechnologie

et d’Instrumentation Optique 12,

rue Marie Curie BP-2060,

10010 Troyes Cedex, France

Jeong-O Lee

Fusion-Biotechnology Research Center,

Advanced Materials Division,

Korea Research Institute of Chemical

Technology, Daejeon 305–600,

Korea

W. Lei

Key Laboratory of Semiconductor

Materials Science, Institute of

Semiconductors, Chinese Academy of

Sciences, PO Box 912, Beijing 100083,

P.R. China

M. Meyyappan

NASA Ames Research Center,

Moffett Field, CA 94035, USA

Guoquan Min

Shanghai Nanotechnology Promotion

Center, Shanghai 200037, China

Otto L. Muskens

FOM Institute for Atomic and

Molecular Physics, AMOLF,

c/o Philips Research Laboratories,

High-Tech Campus 4, 5656 AE

Eindhoven, The Netherlands

Frank Neubrech

Kirchhoff Institute for Physics,

University of Heidelberg,

Im Neuenheimer Feld 227,

69120 Heidelberg, Germany

Garrick Ng

NASA Ames Research Center,

Moffett Field, CA 94035, USA

Xiaoming Niu

Shanghai Nanotechnology Promotion

Center, Shanghai 200037, China

Hyun D. Park

US Naval Research Lab,

4555 Overlook Ave., SW,

Washington, DC 20375, USA

Noejung Park

Department of Applied Physics,

Dankook University,

Yongin 448–701, Korea

S.M. Prokes

US Naval Research Lab,

4555 Overlook Ave., SW, Washington,

DC 20375, USA

Contributors xi

Annemarie Pucci

Kirchhoff Institute for Physics,

University of Heidelberg,

Im Neuenheimer Feld 227,

69120 Heidelberg, Germany

Jaime G

omez Rivas

FOM Institute for Atomic

and Molecular Physics, AMOLF,

c/o Philips Research Laboratories,

High-Tech Campus 4, 5656 AE

Eindhoven,

The Netherlands

Hye-Mi So

Fusion-Biotechnology Research

Center, Advanced Materials Division,

Korea Research Institute

of Chemical Technology,

Daejeon 305–600, Korea

Xuhui Sun

NASA Ames Research Center,

Moffett Field, CA 94035, USA

Alexandre Vial

Institut Charles Delaunay,

Universit

e de technologie de Troyes,

CNRS FRE 2848 Laboratoire de

Nanotechnologie et d’Instrumentation

Optique 12, rue Marie Curie BP-2060,

10010 Troyes Cedex, France

Z.G. Wang

Key Laboratory of Semiconductor

Materials Science,

Institute of Semiconductors,

Chinese Academy of Sciences,

PO Box 912, Beijing 100083, P.R.

China

Zhiguo Wang

Department of Applied Physics,

University of Eletronic Science

and Technology of China,

Chengdu 610054, P.R. China

William J. Weber

Paciﬁc Northwest National

Laboratory, PO Box 999, Richland,

WA 99352, USA

Hongjun Xiang

Hefei National Laboratory for Physical

Sciences at the Microscale, University

of Science and Technology

of China, Hefei, Anhui 230026, P.R.

China

Jinlong Yang

Hefei National Laboratory for Physical

Sciences at the Microscale,

University of Science and Technology

of China, Hefei, Anhui 230026,

P.R. China

Sung Soo Yi

Philips Lumileds Lighting, 370 West

Trimble Road, San Jose, CA 95131,

USA

Bin Yu

NASA Ames Research Center,

Moffett Field, CA 94035, USA

Yafei Zhang

Research Institute of Micro/Nano

Science and technology,

Shanghai Jiaotong University,

Shanghai 200030, China

Weimin Zhou

Shanghai Nanotechnology Promotion

Center, Shanghai 200037, China

Xiaotao Zu

Department of Applied Physics,

University of Eletronic Science

and Technology of China,

Chengdu 610054,

P.R. China

Chapter 1

Study of Nanowire Growth Mechanisms:

VLS and Si Assisted

Hyun D. Park and S.M. Prokes

Abstract In this chapter, we have examined several of our recent results on InAs

nanowires that have implications to the vapor–liquid–solid (VLS) growth mecha-

nism as well as the newly proposed Si-assisted growth mechanism. In summary,

the study on the effect of oxygen during the nanowire growth showed the inhibiting

effect of oxygen on the VLS growth mechanism. The results on the observation of

size-dependent liquidus depression, more importantly, do not seem applicable on the

results of Ti-catalyzed grown Si nanowires, but bring into question the validity of the

vapor–solid–solid (VSS) growth mechanism in the Au-catalyzed grown GaAs and

InAs nanowires. In the newly proposed nanowire growth mechanism, namely the

Si-assisted mechanism using SiO

, a growth model is proposed based on the phase

separation of SiO at higher temperature, which forms a stable SiO

and reactive,

nanometer-sized Si clusters. It is suggested that these clusters consequently serve

as the nucleating/catalyst sites for the growth of InAs nanowires with the growth

mechanism different from VLS, VSS, and OA.

1.1 Introduction

Two primary nanowire growth mechanisms have been reported in the literature:

vapor–liquid–solid (VLS) and oxide assisted (OA). On the one hand, the OA growth

mechanism does not use any metal catalyst, but somehow utilizes the reaction of the

metastable oxide. Previously, nanowires such as Ge [1], C [2], GaAs [3], and Si [4]

have been grown using GeO, CO, Ga

, and SiO, respectively. Based on these re-

sults, it has been suggested that the metastable oxides serve as a nucleating center for

subsequent growth of the nanowires, but much work is still needed to fully under-

stand this growth mechanism. On the other hand, the VLS growth mechanism uti-

lizes a metal catalyst (such as Au, Ni, and Fe) to enhance the nanowire growth. The

key feature of this growth mechanism is the liquid state of the metal alloy tip during

the nanowire growth, which occurs by the formation of a eutectic alloy. Once the

2 H.D. Park, S.M. Prokes

alloy tip forms a eutectic, any excess material that is adsorbed from the vapor pre-

cipitates out in the form of a nanowire. Over the years, numerous nanowires such as

Si [5,6], Ge [7], and InAs [8–10], to name a few, have been successfully grown using

this growth mechanism. Recently, however, a different growth mechanism, namely

the vapor–solid–solid (VSS) (where the metal alloy tip is supposedly in solid state

during the nanowire growth, unlike the VLS) was proposed to explain the growth of

Au-catalyzed GaAs [11] and InAs [12] nanowires in the CBE and MOVPE growth

systems, respectively. The growth of InAs nanowires in the MOVPE system, in par-

ticular, was shown to occur at temperatures up to 480

◦

C when using an Au catalyst

(this is near the Au–In eutectic temperature of 450

◦

C), but no growth was noted at

higher temperatures. However, when a 13-

ASiO

layer was deposited along with

the Au catalyst, the growth of InAs nanowires at temperature of 580

◦

Cwasre-

ported. Based on this preliminary work, it was suggested that the growth at higher

temperature was possible only with the presence of the oxide, where its effect was

to prevent the Au–In tip from melting, thereby extending the VSS growth process.

Our own recent results (where we observed a size-dependent liquidus depression of

Au–In alloy tip) [13], however, reported InAs nanowire growth up to 660

◦

C (with

no SiO

), and the Au–In tip retained its liquid feature down to 400

◦

C. Since our

results conﬁrmed that the VLS is valid in the InAs system, there is some question

as to the validity of the VSS growth mechanism in the Au-catalyzed grown GaAs

and InAs nanowires as well as the real role of the oxide employed in the work of

Dick et al.

In addition, in the work of Kolb et al. [14], SiO evaporation and VLS were com-

bined to grow Si nanowires. In this case, the suggested growth model included the

VLS mechanism to grow Si nanowires, and the oxide was thought to deposit only

around the Si core but not on the Au catalyst. In this chapter, we review some of our

recent work, which includes the effect of oxygen in the VLS growth mechanism,

the controversy regarding the VSS–VLS growth mechanisms, and the newly pro-

posed Si-assisted growth mechanism that was used to explain the nanowire growth

using SiO

. The outline of the chapter is as follows. In Sect. 1.2, the VLS growth

mechanism is discussed, which includes the basic description (Sect. 1.2.1), the ef-

fect of oxygen on VLS (Sect. 1.2.2), and the controversy regarding VSS and VLS

(Sect. 1.2.3). In Sect. 1.3, the Si-assisted growth mechanism is discussed. Conclud-

ing remarks are given in Sect. 1.4.

1.2 VLS Growth Mechanism

1.2.1 Basic Description

In its simplest term, the VLS growth mechanism is a nanowire growth process,

which utilizes a nanometer-sized metal alloy that is in liquid state during the

nanowire growth. It is perhaps the most well-known nanowire growth mechanism,

1 Study of Nanowire Growth Mechanisms: VLS and Si Assisted 3

gold nanoparticles

Sivapor

Si (111)

Fig. 1.1 VLS growth process of Si nanowires using Au nanoparticles on a Si(111) substrate

ﬁrst put forward by Wagner and Ellis [15] and subsequently reviewed in detail by

Wagner and Givargizov [16, 17]. The primary feature of this growth mechanism is

the liquid metal alloy droplet which serves as the preferred site for nucleation due to

its large accommodation coefﬁcient compared with the surrounding solid surface.

It should be noted that the dimension of the nanometer-sized droplet also serves to

determine the diameter of the nanowire.

The schematic of the basic steps involved in the VLS growth mechanism is

shown in Fig. 1.1 in the epitaxial growth of Si nanowires using Au nanoparticles on

a Si(111) substrate. The growth proceeds with the initial deposition of Au nanopar-

ticles (Au thin ﬁlm can also be deposited instead) on an oxide-free Si(111) substrate

using poly-

L-lysine. The Au-deposited Si(111) substrate is then placed inside the

reaction zone and the temperature is raised above the Au–Si eutectic, where the Au

nanoparticles form nanometer-sized Au–Si liquid alloy droplets. At the nanowire

growth temperature, the Si vapor (from such sources as silane, SiCl

, or laser ab-

lation of Si target) impinging on these droplets allows supersaturation to occur.

Further incorporation of Si vapor into the droplet then causes the nanowire to grow

by precipitation above the substrate in the 111 growth direction, with the liquid

alloy droplet remaining at the top. When the nanowire growth is ﬁnished and cooled

to room temperature, the alloy droplet still remains at the tip of the nanowire, but

in solid form. The remnant of this metallic alloy nanoparticle found at the tip of

the nanowire is typically used as evidence of the VLS growth mechanism. Over the

years, numerous types of nanowire growths have been successfully attributed to this

growth mechanism [5,18–20].

1.2.2 Effect of Oxygen on VLS

Recently, we have examined the effect of oxygen on the VLS growth mechanism

in the growth of InAs nanowires inside a torch-sealed quartz tube [21]. The de-

tails of the experimental procedure for this work can be found elsewhere, but the

4 H.D. Park, S.M. Prokes

Fig. 1.2 Growth on InAs(111) with 5

◦

miscut as seen from the top with (a) 50 mTorr (no oxygen

backﬁll). Parts (b)and(c) are with 200 and 800 mTorr of oxygen backﬁll, respectively (reprinted

experiment essentially consisted of a quartz tube (initially ﬁlled with bare and Au-

deposited InAs substrates) evacuated to 50 mTorr and then backﬁlled with 200 and

800 mTorr of oxygen and torch sealed. The quartz tube was then annealed inside

the open-end furnace for the nanowire growth time of 30 min. The InAs nanowires

were grown on InAs(111) substrate with 5

◦

miscut with Au catalyst. The results

are shown in the scanning electron microscope (SEM) images in Fig. 1.2a–c, where

there was a corresponding decrease in the nanowire lengths with the increase in the

oxygen level, clearly indicating the deleterious effect of oxygen on the VLS growth

mechanism.

To further ascertain the effect of oxygen on the VLS nanowire growth, nanowires

were ﬁrst grown for only 5min using an Au catalyst on InAs(111)B substrate. These

nanowires were then oxidized in air at room temperature for 3 days, placed back into

the quartz tube, and then grown for an additional 25 min. The quartz tubes were all

sealed at 50 mTorr. The results are shown in Fig. 1.3a and b for 5 min and addi-

tional 25 min (after room temperature air oxidation), respectively. The nanowires

grown on InAs(111)B substrate under identical, nonoxidized growth condition typ-

ically yielded a length greater than 2µm (not shown), but as can be seen in the

SEM images, we only observed a small change in the length of the nanowires from

5 min (250–300 nm) to 25 additional minutes (300–600 nm). This result indicates a

noticeable decrease in the growth rate after the oxidation step.

From the growth experiments performed at various vacuum pressures, these re-

sults suggest that the presence of oxygen inhibits nanowire growth in the VLS

growth mechanism. Although some nanowires were still seen growing even at

800 mTorr, the inhibiting effect of oxygen was quite evident and it was more pro-

nounced with worsening vacuum conditions. The room temperature oxidation of the

Au–In tips in the air offers additional proof of the deleterious effect of oxygen on

the nanowire growth. For the Au–In alloy, the interdiffusion of Au and In occurs

quickly even at room temperature, and various intermetallic compounds are formed

depending on the wt% of In and Au. After 1 day, intermetallic compounds such

as AuIn

, AuIn, and Au

have been identiﬁed [22]. As In readily oxidizes, the

In in the Au–In alloy will be subjected to oxidation. Although the literature on the

1 Study of Nanowire Growth Mechanisms: VLS and Si Assisted 5

Fig. 1.3 SEM images showing growth at (a) 5 min and (b) 25 additional minutes after 3 days room

temperature oxidation

oxidation of Au–In alloy is scarce, the work of Pasquevich et al. [23] has shown

that In oxidation occurs primarily on the surface, not internally, regardless of the In

content in the Au–In alloy. Consequently, the surface oxide, which forms either at

room or at higher temperature (at 580

◦

C during the growth at poor vacuum pres-

sures), will then serve as a diffusion barrier for impinging In and As atoms, which

are in the vapor state, thus inhibiting the nanowire growth. This is consistent with

our reported growth results.

6 H.D. Park, S.M. Prokes

1.2.3 VLS and VSS

A recent and more debated development regarding the VLS growth mechanism

is the newly proposed VSS growth mechanism that was used to explain the

Au-catalyzed growth of GaAs nanowires in CBE by Persson et al. [11] (and

subsequently by Dick et al. [12] in the Au-catalyzed growth of InAs nanowires

in MOVPE). The basis of the VSS growth mechanism is the solid state of the

metal alloy tip used during the nanowire growth, and not liquid as in the case of

VLS. Generally, the VLS growth mechanism had been quite successful over the

years in explaining the nanowire growth process that used metal catalysts in the

growth. Recently, however, there was one result, namely the growth of Ti-catalyzed

Si nanowires by Kamins et al. [24], that did not seem consistent with the VLS

growth mechanism. In their work, Kamins et al. found the growth of Ti-catalyzed Si

nanowire at approximately 600

◦

C, much below the Si–Ti eutectic temperature (the

lowest published Ti–Si eutectic temperature is about 1,300

◦

C). Because the growth

temperature was considerably lower than that required for the metal alloy tip to

be in liquid state, and with the size-dependent melting point depression interpreta-

tion insufﬁcient, the authors then concluded that the nanowire should have grown

with the metal alloy tip in solid state. Subsequently, Persson et al. investigated the

growth of GaAs nanowires by CBE using an Au catalyst, where Au–Ga alloy is

formed. Then, through in situ transmission electron microscope (TEM) analysis and

X-ray energy dispersive spectroscopy (EDS) of the GaAs nanowires, Persson et al.

observed not only the crystallinity of the Au–Ga seed alloy particle at the growth

temperature, but also a low Ga concentration in the Au–Ga alloy below the level

required for the eutectic melt.

Based on these observations, VSS growth mechanism was then proposed. Soon

thereafter, Dick et al. [12] used this growth mechanism to explain the growth of InAs

nanowires by MOVPE as well (with Au–In alloy particle), suggesting the failure

of the VLS growth mechanism. In their work, the authors also concluded that the

growth of Au-catalyzed InAs nanowires was possible only when the metal alloy tip

of Au–In was in solid state during the nanowire growth, which limited the growth of

InAs nanowires below the eutectic temperature. It should be mentioned that for both

Au-catalyzed GaAs and InAs nanowires, the element As is not present (or present

only in trace amount) in these metal alloy tips and thus only a binary phase diagram

is necessary.

Since then, several contradicting works by Harmand et al. [25] and by us [13,21]

have been published, reexamining the validity of the VSS growth mechanism. In the

work of Harmand et al., the elemental composition of the Au–Ga seed alloy particle

was examined at different growth durations in the case of GaAs nanowires grown in

MBE. From their analysis, the authors have identiﬁed three different metallic com-

pounds at room temperature: the hexagonal β



structure, the orthorhombic

AuGa structure, and an almost pure Au face-centered cubic structure, and observed

that the ﬁnal composition of the metallic particle (determined at room temperature)

depended on the growth history of the wire. Thus, it was suggested that the Au–

Ga seed alloy particle was indeed in liquid state during the nanowire growth and

1 Study of Nanowire Growth Mechanisms: VLS and Si Assisted 7

that GaAs nanowires grew via the VLS mechanism in contrast to the VSS growth

mechanism of CBE-grown nanowires.

In our recent work, we studied the size-dependent melting point depression of

a metallic binary system undergoing the VLS growth mechanism. In our work, a

point elemental composition EDS was performed on various Au–In tip sizes (20–

100 nm) in the InAs nanowires that were grown in the temperature range 400–660

◦

(not shown). The highest growth temperature reported was approximately 100

◦

higher than the temperature (which is above the eutectic temperature) at which InAs

nanowires were grown in the MOVPE system of Dick et al. The results of EDS

analysis are shown in Table 1.1, where tips having diameters approximately 60 nm

and below showed decreased In content compared with diameters larger than 60nm.

The results show a 30% decrease in the In content in all nanowires with diameters

60 nm or less, as compared with the 100-nm nanowires. Interestingly, in the Au–

In phase diagram (not shown), the measured In content for the tip size 100 nm is

within the liquid region above the liquidus line, while the tip sizes 60 nm or less are

in the solid plus liquid region. These observations suggest the occurrence of size-

dependent melting point depression, but to support this claim, the InAs nanowires

were also grown below the Au–In eutectic at 400

◦

C using 20 and 60-nm-sized

Au nanoparticles. The results, as shown in Fig. 1.4, show the liquid-like migration

of Au–In alloy particle (as well as subsequent nanowire growth) despite the tem-

perature being below the level required for the eutectic melt, thereby conﬁrming

the occurrence of size-dependent melting point depression as well as the nanowire

growth via VLS growth mechanism.

One possible origin of the melting temperature depression is the Gibbs–Thomson

effect owing to the ﬁnite size of a crystal [26]. To illustrate the order of magnitude

associated with this effect [13], consider a single-component spherical crystal of ra-

dius r. The melting temperature T

of the ﬁnite size crystal is given by the expression

−T

= 2

γΩ

/r∆S,

Table 1.1 Summary of EDS elemental composition analyses on the gold alloy tip of various

nanowires

Size (nm) per

element

580

◦

C (liq. 26.6%) 610

◦

C (liq. 25.5%) 660

◦

C (liq. 23%)

Au In Au In Au In

25 79.8 19.5

35 80.2 19.5

40 80.4 19.4

60 79.8 19.6 78.1 20.9 80.4 19.3

100 71.3 28.2 71.0 28.3 77.2 22.5

The values are in atm%. The standard deviation was ∼2atm%. Trace amount of As (<0.5atm%)

Elsevier)

8 H.D. Park, S.M. Prokes

Fig. 1.4 InAs nanowires grown at 400

◦

Cfor(a)60nmand(b) 20 nm sized Au nanoparticles

where T

is the bulk melting temperature,

is the free surface energy,

Ω

is the

molar volume, and ∆S is the molar entropy change for melting. Assuming ∆S to be

approximately independent of temperature, then ∆S = L/T

, where L is the molar

latent heat of melting, so that (T

−T

)/T

≈ 2

γΩ

/rL. Using the values for Au

(

= 1.4Jm

−2

Ω

= 10.2cm

mol

−1

, L = 12.55kJmol

−1

) and a radius of 40 nm

gives a melting temperature reduction of 6%. This would also apply to a single-

component sphere sitting on a “noninteracting” substrate. For the case of Au on the

InAs substrate, however, it is a two-component alloy on an “interacting” substrate.

Nevertheless, since the expressions for Gibbs–Thomson effects in multicomponent

systems have the same dependence on the size of the crystal and similar depen-

dences on the surface thermodynamic parameters [26], it is expected that a ﬁnite

size Au–In alloy would display a similar order of magnitude reduction for the liq-

uidus temperature. Also, since the free surface area of the crystal is larger than the