Growth and transmission electron microscopy studies of nanomaterials 3 4

We also directly observe the process of formation of a BN encapsulated metal nanoparticle for the first time by performing in-situ TEM at high vacuum in order to get insight into the gro

Chapter 3 Growth of boron nitride nanostructural

materials

In this chapter, the nucleation and growth of boron nitride nanomaterials in

microwave plasma enhanced CVD system and in-situ TEM is presented in detail

3.1 Introduction

The growth chemistry and atomic structure of BN nanostructures have

attracted much attention because the wide band gap exhibited by BN suggests

interesting applications in novel UV lasers and low-k materials [1-3] Following the

discovery of the carbon nanotube by Iijima [4], similar structures have been

proposed and discovered for BN [5] Compared with the carbon nanotubes, BN

nanotubes have been predicted to display a bandgap of roughly 5.5 eV, independent

of their chirality together with high ultimate strength and oxidation resistance [5, 6]

Thermal stability and oxidation resistance make BN a suitable material to store and

protect some air-sensitive metals inside, such as pure Co and Fe [7]

3.2 Motivation

BN nanotubes have been synthesized by a range of methods from arc

discharge [8, 9], chemical vapor deposition [10, 11] to solid-state ball milling

methods followed by annealing at high temperature [12, 13] Non-catalytic growth

methods have also been demonstrated [14-16] Compared with the growth condition

required for carbon nanotubes, most of these methods require thermal conditions

higher than 1200 ˚C and the yield of BN nanomaterials is often low The growth

temperatures required for BN synthesis are often too high to be compatible with

microelectronic processing on conventional semiconductor substrates Hence, the

properties and technological applications of BN nanotubes, nanowires and

nanospheres have not been fully investigated due to the lack of these materials in

sufficient quantity

Narita and Oku [17-19] have studied the arc-melting of a sequence of borides

such as TiB2, VB2, NbB2, LaB6 for the production of BN nanocapsules and

nanotubes The formation enthalpies of BN from the respective borides were

considered as the key factor in catalyst design However another important factor is

the composition of the boride phase in the design of the catalyst, since this will

influence the eutectic melting point A lower eutectic melting point will promote the

ready formation of a molten phase to initiate the VLS mechanism [20], thus

allowing the phase segregation of BN nanomaterials at lower temperatures

However, the direct plasma nitridation of iron borides at moderately low

temperatures, i.e <1000 ˚C, for the growth of BN nanomaterials has not been

demonstrated The growth of metal encapsulated BN nanoparticles is more difficult

due to the lack of a reactive BN-precursor

The growth of metal encapsulated nanomaterials, where the outer sheath

consists of an inert coating to prevent the oxidation of the inner metal nanoclusters,

offers important technological applications in ferromagnetism and quantum devices

Interesting host-guest chemistry or supra-atom properties may be exhibited by

metallo-fullerene type structures where a metal atom is inserted in the center of the

BN nano-sphere [21] BN nanotubes may also act as a sheath encapsulating metal

nanowire, forming an interesting insulator-metal structure A recent study of BN

nanocage-encapsulated cobalt nanopaticles showed ferromagnetic properties [8]

Cobalt nanoparticles encapsulated in carbon shells have been synthesized by

catalytic chemical vapor deposition with high yield by reducing Mg0.9Co0.1O solid

solution-impregnated MgO catalyst with a H2/CH4 mixture Magnetic measurements

confirmed that carbon-encapsulated Co nanoparticles were protected against

oxidation and remained metallic Fe-based films are the logical choice for next

generation magnetic record heads because pure Fe has a magnetization of 1710

emu/cc Encapsulating the iron, or reacting with nitrogen to produce ultrafine grain

sizes is an effective means of obtaining films with low coercivity and high

permeability It has been suggested that the 2D organization of such magnetic

nanoparticles could be a step forward in the realization of high-density recording

media

In this work, we report the nucleation and growth of various BN

nanomaterials using the heterogeneous reaction between ammonia and iron boride

particles in order to prepare BN nanomaterials at temperatures lower than 1000 ˚C

NH3 or nitrogen is the single source to provide nitrogen to react with boron evolved

from FeB particles We also directly observe the process of formation of a BN

encapsulated metal nanoparticle for the first time by performing in-situ TEM at high

vacuum in order to get insight into the growth mechanism of BN nanomaterials

3.3 Experiment

Reactive FeB nanoparticles [22, 23] were prepared by ball-milling water-free

FeCl3 and NaBH4 powders (1:3.3 mole ratio) for 8 hours, followed by annealing in

an Ar atmosphere at 500 ˚C for 3 hours

In the series of ex-situ TEM studies, samples were prepared by casting FeB

nanoparticles onto Si (100) wafers, where were then introduced into the

MW-PECVD system The sample was heated to 850 ˚C during plasma treatment

The gaseous precursor consisted of a mixture of N2:H2 in a volumetric flow ratio of

3:1 at a pressure of 20 torr The microwave power was maintained at 350 W The

product was scraped off the Si substrate, ultrasonically cleaned in ethanol, and then

cast onto a Cu grid for observation with a Phillips CM300 TEM

For direct in-situ TEM observation, detailed information on sample

preparation was given in chapter 2.1.4 Ammonia gas was admitted into the TEM

column through a leak valve The in-situ TEM has a base pressure of ~3-4×10-6 torr

during reaction and the sample was heated to 1000 ˚C

3.4 Growth of BN nanomaterials from FeB catalysts

BN(002)

10 min

30 min60min

Figure 3.1 Grazing angle XRD spectra of FeB before and after nitridation

The original FeB powders are fine black particles TEM observation of the

FeB particles before reaction shows that the nanoparticles are approximately 20–30

nm in diameter Following microwave plasma treatment, a whitish-grey product

covered the sample Grazing angle XRD was performed to determine the crystal

structure of FeB before and after reaction at different nitridation times under the

same experimental conditions The diffraction pattern, shown in figure 1, reveals

that the crystalline phase of the initial FeB particles is alpha-Fe, and the Fe-B is also

present as an amorphous phase Hu et al [22] also found FeB to be amorphous using

Mössbauer spectroscopy Following nitrogen plasma treatment at 850 ˚C in a

microwave CVD system, we can detect the peaks assignable to the products of iron

and nitrogen and phase recrystallization from FeB, such as Fe23B6, Fe8B, Fe2N and

FeN, as well as a small peak due to BN (002), suggesting the formation of

crystalline h-BN The intensity of the BN (002) peak is similar for all samples after

nitridation, resulting from the formation of BN mainly in the first 10 minutes

CL measurements (figure 3.2) were carried out to characterize the quality of

the BN products No peak can be detected for FeB particles before N2 plasma

treatment After reaction, in each case, emission is detected centered at ~370 nm,

corresponding to an energy of 3.36 eV which is considered to be the impurity level

of the BN layers (5.5 eV) Impurities such as oxygen (leakage) or hydrogen can be

incorporated into the BN during N2 and H2 treatment [24, 25] The CL signal

increases substantially with increasing nitridation time, and the FWHM is observed

to decrease with nitridation time, from ~50 nm after 10 minutes, to ~20 nm after 60

minutes nitridation This is most likely due to the reduction of defects in the BN

products with increasing reaction time The asymmetry of the peak also decreases

with nitridation time, as longer wavelength emission is suppressed This is most

likely due to the reduction in mid-gap defects due to surface and related effects CL

measurement proves the high quality of BN can be achieved after N2 treatment of

FeB particles

Generally, several types of BN nanostructures were formed from the grey

deposits following the nitrogen-plasma treatment of the FeB nanoparticles A

comprehensive analysis of the morphologies of the BN nanomaterial and the

adjacent FeB particles from which it ensues were studied carefully under TEM

TEM results indicate a relationship between the size of the FeB particle and the

shape and form of its BN derivative In actual experimental conditions, where we

subjected the FeB-cast silicon substrate to a N2-plasma treatment, the distribution of

products on the silicon substrate is quite heterogeneous and the nanotube diameter

correlation is complicated by nanoparticle-substrate interaction, mobility of particles

at elevated temperatures and fragmentation and coagulation process Underpinning

this relationship is the relative ease by which the particle undergoes liquid-flow

Table 3.1 Relationship between the FeB particle size and the morphology of the nanomaterials ensuing from it

Size of associated

FeB particle (nm) Morphology Length Quantity Reference

Iron nanowires 0.5–1 µm Few Fig 3.7

Table 3.1 summarizes the apparent relationship between the size of the FeB

particle and the BN nanomaterial that ensued from it, as was observed in this work

It is more difficult for the larger particles to achieve fluid-like motion compared to

the smaller particles So if the smaller FeB particles aggregate to form a FeB particle

larger than 200 nm, the mobility of this particle is restricted and only an outer BN

microcapsule grows on it Nanotubules are generally observed to grow from FeB

particles with sizes in the range of 50–200 nm, because these particles melt and flow

readily, and trace out nanotubules with bamboo-like segments in its motion What is

interesting are FeB particles with sizes ranging from 50 to 100 nm, smaller BN

nanotubes and nanocapsules are readily created from these, and in addition, these

FeB particles readily melt and recrystallize into crystalline Fe nanowires ensheathed

by BN A detailed discussion of the various structural polytypes follows below

Figure 3.3 TEM images of BN nanotubes recovered from the FeB nanoparticles after nitrogen-plasma treatment at 850 ˚C

Some BN nanotubes can be observed in the nitridation products, as shown in

figure 3.3a These are BN nanotubes with diameters of 20-30 nm and length to 200

nm Higher magnification images were recorded in order to observe the internal

structure of these nanotubes Unfortunately, the low contrast between the BN

nanotubes and the carbon support film leads to poor quality images However, we

can still see tat the walls of the BN nanotube are poor by crystalline with many

defects which make the BN nanotubes unstable under the strong electron beam

These nanotubes may have evolved from a partial crystalline BN film during the

nitridation process For the growth of BN nanotubes, a high temperature and a slow

cooling rate would be needed In our plasma system, 850 ˚C may not be a high

enough temperature to form high quality BN nanotubes

Figure 3.4 TEM images of BN nanocapsules

For FeB particles with sizes between 50 and 100 nm, nanotubules or

nanocapsules are readily formed from them We can observe the Fe or FeB-FeN

nano-sized particles encapsulated in the BN nanocages Isolated FeB particles that

escaped sintering afforded an interesting observation point for the growth of BN

nanocages Figure 3.4 shows TEM images of BN nanocages consisting of

multi-layered hexagonal BN with some FeB particles leaving the BN nanocages

When some of the particles were leaving the BN capsules, the movement of the

particles stopped due to the rapid reduction of the temperature This indicated that

the BN nanocages formed at the surface of FeB particles

Figure 3.5 TEM image of (a) (b)BN bamboo-like nanotubules, HRTEM images of (c) center and (d) tip of the nanotubules

Moreover, we can observe for example a one-micron long nanotubule traced

out by one single FeB particle, as shown in figure 3.5 HRTEM lattice images of the

nanotubule close to the FeB particles and at the center are shown in figure 3.5c and

d It is noteworthy to point out two facts: First, the nanotubule formed is a

shape-transform of the FeB particle as judged by the BN walls surrounding the

exterior of the FeB particle Second, the nanotubule walls are curvy and

twisted-looking, and are formed by a string of interconnected BN nanotubles

circumscribed by the FeB particle in its liquid flow

Figure 3.6 HR-SEM images of (a) BN microcapsules on the top of FeB particles which resembles cracked egg-shells; (b) single crystalline FeB or Fe particles beside the BN microcapsules; (c) TEM image BN microcapsules (d) HRTEM image of walls of these capsules

(a)

(d) (b)

(c)

When the size of the FeB particles are large than 200 nm due to aggregation,

the mobility of the FeB particles will decrease compared with small particles The

outer BN micro-capsules will grow around these big FeN particles Following the

nitrogen plasma treatment and cooling, many micron-sized egg shell-like BN

microcapsules were formed, as shown by the SEM images in figure 3.6a What is

noteworthy is that all the BN microcapsules have a crack in the shell Next to these

cracked BN ‘egg-shells’ can be found single crystal, well-faceted Fe, as shown in

figure 3.6b This suggests that molten metal drops created from the aggregation of

FeB particles are contained in the micron-sized solid BN capsule during the plasma

treatment and high-temperature annealing Compressive stress created by the

inward-growing BN shells caused the molten Fe to break through, escape and

recrystallize into single crystals next to it, leaving behind the cracked BN egg-shells

This suggests that the compressive force is strong enough to crack thick BN

microcapsules The internal microstructure of some of the thinner, transparent BN

microcapsules is shown by the TEM images in figure 3.6c and 3.6d These thinner

capsules become disengaged from the FeB particles after ultrasonication The walls

of the capsules are well ordered with the interlayer spacing of 3.4 Å, which is in

good agreement with the spacing of 3.34 Å of bulk hexagonal BN Such shell-like

BN sub-micron and micron scale structures have also been observed in laser ablation

studies by Komatsu and coworkers [26, 27]

Figure 3.7 TEM images showing: (a) metallic nanoparticles exhibiting capillary flow and being entrapped in the mass of BN sheets; (b) image of a long iron nanowire encapsulated in

BN Higher magnification images in (c) and (d) reveal that the interface between the BN and

Fe consists of disordered h-BN

Figure 3.8 Bright field image and elemental mapping of Fe, N and B of BN ensheathed–Fe nanowire The results show that the core is filled with elemental iron, whilst the sheath is made of BN

In addition to BN nanotubes and nanospheres, metallic nanoparticles with

high shape anisotropy can be found among the products after the nitrogen plasma

treatment The relative ease of liquidification and capillary flow of smaller FeB

particles, followed by recrystallization, result in the growth of iron nanowires

encapsulated by disordered hexagonal BN sheets Figure 3.7a shows TEM images of

metallic nanoparticles entrapped in the mass of BN sheets after capillary flow It can

be seen that some wires grow from the catalyst by a "root-growth" mechanism, with

the source of the wire in the FeB catalyst The diameter of the wire is about 5 nm

The catalyst was observed to move with capillary fluidity in the tube and crystalline

planes could be seen inside the tubule Figure 3.7b shows an image of a long iron

nanowire encapsulated in BN Higher magnification images in figure 3.7 c, d reveal

that the interface between the BN and Fe consists of disordered h-BN, and

crystalline h-BN layers nucleates on top of the disordered h-BN The composition

was verified by elemental mapping with EELS in figure 3.8, which confirmed that

the inner core is elemental iron, whilst the outer coating is BN The wrap around of

the entire iron wire by BN layers attests to the good wetting properties between iron

and BN Babonneau and coworkers reported the synthesis of vertically elongated

nanoparticles of Fe2N encapsulated in disordered hexagonal BN by the ion beam

sputtering of BN–Fe nanocomposite film [28] They attributed the high shape

anisotropy to the activated surface diffusion of the incoming deposited species

assisted by ion beams In the work described here, many BN laminar sheets, as well

as small isolated FeB particles were left behind in an extended matrix after the

encapsulated FeB particles escaped and coagulated into bigger particles These

laminar sheets may curl up to form BN nanotubes with the metal particle

encapsulated within Plasma-activated nitrogen and hydrogen species in our

experiment may also impart energy for the activated diffusion of Fe on the

dimensional BN laminar sheets, resulting in the anisotropic crystallization of

metallic nanowires

Based on these experimental observations, we can deduce that the growth

occurs at the inner face between the BN walls and the FeB particle

Nitrogen-bearing species may be supplied to the FeB particles either by diffusion

through the shell wall, or by diffusion along the particle:shell interface Diffusion of

N–BN layers has a high activation barrier with a characteristic time that is much

larger compared to diffusion through the liquid particle Low diameter liquefied

nanoparticles are characterized by significantly shorter times of N diffusion through

the particle body which is important for BN nanotube formation by the segregation

process The dissolution of nitrogen in the molten FeB results in the formation of

BN, which is insoluble and segregates as an outer BN shell Following the

segregation of a single BN shell, growth of further BN shells proceeded between the

inner face of the shell and the FeB particle which is wetting it by the continuous

epitaxial precipitation of BN from the molten FeB, i.e the growth occurs between a

wetted liquid-solid interface This process is similar to the VLS growth method The

inward directed growth of the BN shells will result in a continuously shrinking inner

volume and increasing tensional forces at the Fe–BN interface At any aperture in

the BN shell, capillary action will draw the molten FeB out from within and through

it as the compressive forces build up to a critical level

Figure 3.9 Phase diagram of Fe-B binary alloy

The FeB phase diagram [29] indicates that the FeB is stable to 1650 ˚C No

liquid Fe-B phase can be detected in the phase diagram But as the B content

decreases below 50% at a temperature of 1000 ˚C, Fe2B (solid) and liquid FeB are

the equilibrium phases (stable to 1389 ˚C), whilst below 33% B, solid Fe and liquid

Fe2B are expected, with a eutectic at 1174 ˚C In addition, the melting point of the

small particles decreases dramatically compared to that of the bulk states The

surface melting will occur when the particles size reaches nanometer size [30, 31]

The well-known Gibbs–Thompson effect predicts that for particles with sizes below

100 nm, melting occurs at a significantly lower temperature compared to values in

the standard phase diagram, thus fluid-like motion can be attained more readily at

~1000 ˚C

3.5 in-situ TEM study of growth of BN nanocapsules from FeB particles

Based on our experimental observations we propose that growth of the BN

nanocages proceeds by the dissolution of nitrogen in the molten FeB results in the

formation of BN, which is insoluble and segregates as an outer BN shell at the

surface of the FeB particle, with the Fe playing the role of catalyst To fully

understand the growth mechanism of the BN nanomaterials, the FeB catalysts were

nitrided in the in-situ TEM This allowed the direct observation of the nitridation

process with ammonia, sintering of the FeB particles, as well as the growth of BN

layers on the periphery of the aggregated particles could be observed

The sample was heated in the presence of a partial pressure of 3-4×10-6 torr

of NH3 background Around ~1000 ˚C in high vacuum, a rapid and significant

degree of sintering and agglomeration of the FeB nanoparticles was observed, which

made electrons difficult to go through the melted thick part of the nanoparticles

Fluid-like motion of the particle profiles was apparent at and above ~900 ˚C, which

confirmed the liquid like behavior of FeB concluded from the phase diagram and the

small size effect on the melting point To avoid sintering, the sample was then heated

to a maximum temperature of 900 ˚C Despite their fluid-like motion, the particles

remained crystalline, as determined by the diffraction contrast associated with the

particles In the heating process, due to thermal drift and the influence of a high

concentration of NH3 on the electrons, we were not able to follow closely the growth

of all dimensional structures other than the nanospheres described above

Figure 3.10 Bright-field image of the Fe–B nanoparticles prior to heating

TEM observation of the particles revealed excellent size uniformity, and

figure 3.10 presents a bright-field image of a typical area of the sample prior to

reaction with ammonia The FeB particles are approximately 30 nm in diameter

EELS was performed at this stage and revealed the presence of the B K-edge and C

K-edge with low concentrations, as shown in figure 3.11 (lower trace); no evidence

of the presence of nitrogen was observed prior to reaction EELS spectra recorded

post-nitridation is shown in figure 3.11 (upper trace) The spectrum now shows two

characteristic edges at 188 eV and 401 eV corresponding to the K-shell ionization

edge of B and N, with the absence of carbon The ionization edges π*, σ* and the

energy loss fine structure correspond to the sp2 hybridization of BN, as expected for

the formation of hexagonal BN

The composition can be derived from the EELS spectrum using the

following equation [32]:

),(),(),(

),(),(),(

2 2

2

1 1

1

2 1

βσββ

II

II

C

C

c c

where Ic1(β, ∆) is the integrated core-loss intensity integrated over an collecting

semi-angel β (10 mRad) and an energy window of width ∆ (50 eV), I1(β, ∆) is the

integrated intensity of the low-loss region of the spectrum, σ(β, ∆)is the

cross-section for core-loss scattering up to the β and ∆, C is the number of atoms per

unit area of corresponding element Since we only need relative concentrations, the

low loss integral can be ignored From the above calculation, the B: N atomic ratio

of the BN nanocapsules is around 0.93:1, which is consistent with the stoichiometry

of BN The estimated error of the calculated result is around 10 % This suggests

that stoichiometric BN with B/N atomic ratios of approximately 1 can be achieved

after nitridation

Detailed insights into the heterogeneous reaction come from elemental

mapping by energy-filtered imaging in figure 3.12 The red particle indicates iron,

which is surrounded by thin BN cages We can conclude that the iron particle is

encapsulated by a BN nano-cocoon

Figure 3.12 Elemental map of FeB particles after annealing in NH3 at 900 ˚C, Blue- Nitrogen, Green-Boron, red- iron

Figure 3.13 A bright-field image shows the series of BN nanocages produced by the nanoparticle

After a few minutes exposure to NH3 at 900 ˚C, melting of the catalyst and

its resultant “activation” could be detected At this point, crystalline layers were seen

to nucleate around the molten catalyst The particle located at ‘E’ (figure 3.13) was

initially located within the BN nanocage at ‘A’; the series of BN nanocages formed

around the particle can be clearly seen in TEM, in positions ‘A’-‘E’ The following

sequence of events was observed in real time, as shown in figure 3.13: (i) growth of

BN nanocages around iron boride particle; (ii) After reaching a certain thickness, the

aperture of the cage is closing, capillary force at the interface expels the molten iron

boride particle into region B, leaving an incipient void, and a new BN cage grows in

B; the FeB particles were seen to be ejected from the core of the shells that had

formed around them, being subsequently located outside the shell but still in contact

with it (iii) the particle was expelled into C after the BN nanocage reaches a certain

size, and (iv) process repeats itself, and then into ‘E’ finally No further movement

was detected in ‘E’ thereafter even though the reaction conditions were maintained

Figure 3.14 High-resolution phase contrast image of the BN layer structure The observed lattice spacing corresponds to the (002) planes of hexagonal BN

The thickness of each nanocage wall was found to be ~3-5nm In figure 3.14

we present a high resolution image of a representative wall, showing the layered

structure of hexagonal BN as viewed along the (002) planes The observed lattice

spacing corresponds to the expected (002) spacing of 0.34 nm The (002) plane,

being the lowest free-energy layers for hexagonal BN, manifests as many two

dimensional (002) sheets in many areas The cages shown here are still filled with a

metallic particle with a second cage overlapping this cage in the lower left portion of

the image The presence of occasional defects within the cage walls was observed,

and a typical example is found in this image, labeled ‘D’ At position ‘F’, above the

through-thickness defect, a section of the outer shell wall has apparently ‘folded’

over during the initial stages of growth The feature has some similarities with the

observations of Bengu and Marks [33] where characteristic shapes were calculated

for capped single-walled BN nanostructures that avoid B-B or N-N bonds which

would be unstable Subsequent growth in this part of the wall has been disturbed,

resulting in the defect observed at ‘D’ which has propagated through the entire

thickness of the wall The defects may permit the permeation of N atoms or

molecules into FeB particles

Figure 3.15 Video frames captured from the particle expulsion sequence showing the dewetting of the nanoparticle from the BN cage wall The frames were recorded at (a) t ~ 0

ms, (b) t ~ 360 ms, (c) t ~ 400 ms and (d) t ~ 1280 ms

The details of the ejection process may be seen in the video frames presented

in figure 3.15, which were recorded immediately before (figure 3.14(a)) and during

(figure 3.15(b)-(d)) the ejection process Frames (b), (c) and (d) were recorded 360,

400 and 1280 ms after frame (a) From the image in figure (b), a dynamic contact

angle of 150º can be measured between the BN:Fe-B interface and the dewetting

Fe-B particle surface

The equilibrium value of the contact angle between the liquid and solid

defines the wetting behaviour of the liquid, and obeys the classical equation of

Young:

LV

SL SV

σ

σσ

cos

where σSV, σSL and σLV are the solid: vapour, solid: liquid and liquid: vapour

interfacial energies, respectively Although in our experiments the Fe-B particle

remains crystalline, the observed behavior is liquid-like for the purposes of this

equation A contact angle in excess of 90º implies a non-wetting BN: Fe-B interface

The energy of the interface between FeB and BN is therefore higher than the

combined energies of the “liquid-like” FeB: vacuum and the solid BN: vacuum

interfaces This is consistent with our observations, where no detectable Fe-B

remains on the inner walls of the shells following particle expulsion

In our synthetic method, the BN growth relied on the material supply of B

from the FeB and not from external supply The diffusion of nitrogen into the

peripheral regions allows BN to crystallize from the molten iron boride The growth

of BN will be quenched once the supply of boron in the iron boride is depleted and

no further volume diffusion of boron reacts with the ammonia at the interface

occurred The expulsion force of the FeB particle arises from the wetting

incompatibility of BN with iron and the increasingly smaller volume in the cage

following the inward growth of the BN hexagonal layers The reaction in this study

is a vapor-liquid-solid heterogeneous reaction involving the difffusion of gaseous

ammonia into the molten FeB and subsequent recrystallization into h-BN

Figure 3.16 Schematic diagram of (a) initial FeB particles before dosing NH3, (b) the growth

of a BN shell around an Fe–B nanoparticle, followed by (c) strain-induced ejection, and (d) the subsequent formation of a second shell connected to the first

We propose that initial nucleation of BN occurs at the interface between the

particle and the support, with the subsequent growth of the BN shell to coat the

particle free surface A region of uncoated Fe-B surface appears to remain, as may

FeN

BN

FeB FeN/FeB

be seen at the leading edge ‘A’ of the particle (arrowed) in figure 3.13 Growth then

continues at the interface between the particle and the shell Nitrogen-bearing

species may be supplied to this interface either by diffusion through the shell wall,

or by diffusion along the particle: shell interface Dissolution of nitrogen species in

the Fe-B crystal, followed by incorporation into the growing shell is also a

possibility; these processes are illustrated schematically in figure 3.16(a)

In some cases, high resolution TEM images recorded after cooling to room

temperature suggest the presence of metallic species within the cage wall, trapped

between BN layers The presence of the metallic species within the wall would

indicate that the metal became incorporated during the growth process, becoming

trapped by the overgrowth of further BN layers This lends further support to our

proposal that BN growth is taking place at the particle: shell interface (growth by

addition to the inner wall of the shell) rather than at the interface between the shell

and the vacuum

As the thickness of the shell increases, and inward growth continues, strain

energy will build up in both the BN shell wall and the Fe-B particle At a critical

shell thickness, the strain energy of the system will exceed that required for

expulsion of the particle from the cage by dewetting, and the ejection process of

figure 3.13 will take place This is illustrated schematically in figure 3.16(b)

Provided sufficient B remains within the particle, the formation of a second BN shell

around the freshly exposed particle surface, figure 3.16(c), may then occur We note

that the shells remain in fixed positions relative to the substrate throughout the

growth sequence, with only migration of the Fe-B particle being observed

Calculations based on simple mass balance show that the maximum amount

of BN that could be produced by a single particle of FeB, intitially 80 nm in

diameter, is approximately consistent with the amount of BN observed to form in

our experiments (figure 3.12) Assuming complete conversion of an 80 nm diameter

FeB particle to Fe, a total of four complete shells, each ~3 nm in diameter could be

generated by the complete reaction of the elemental boron with nitrogen species

This would be approximately sufficient to generate the 5 shells observed in figure

3.12, allowing for the sharing of walls between shells Multiple nanocages were

formed from a single Fe-B particle by a mechanism of strain-induced particle

expulsion The observed liquid-like behavior at temperatures well below the melting

temperature provides direct evidence of the possibility of growth models for related

nanostructural materials which involve liquid-like flow of metal nanoparticles

If the sample is heated to 900 ˚C, significant sintering and agglomeration of

the FeB nanoparticles are observed The low partial pressure of NH3 during the

reaction, yet rapid rate of BN growth, implies a high reaction efficiency between

impinging ammonia molecules and boron The occasional defects observed in high

resolution lattice images (such as ‘D’ in figure 3.13) wall could act as rapid diffusion

paths for the reactive gas species to the Fe-B interface, in addition to diffusion along

the particle: shell interface

The influence of the illuminating electron beam on the observed shell growth

was checked by examining areas of the sample that were not subjected to irradiation

during growth In addition, periodic observation of fresh areas of the sample during

the growth process was also made, with no significant differences being observed

Chen and coworkers [12, 13] reported observations of bamboo-type BN

nanostructures with diameters up to 100 nm, synthesized by annealing of ball milled

BN powders at 1300 ˚C under nitrogen TEM images showed a continuous

endohedral capillary structure with Fe particles trapped within the free ends of the

tubes A growth mechanism involving continuous shape transforms of the Fe

particles as they became coated with BN (by solid state diffusion) was suggested

Our observations lend substantial support to this proposed mechanism, since we

have demonstrated that liquid-like capillary flow of high melting point material such

as Fe-B is possible at temperatures as low as 900 ˚C This is 400 degrees lower than

the temperature at which Chadderton and Chen’s experiments were performed Our

experiments thus provide direct evidence of the possibility of liquid-like flow in a

broader class of experiments involving fine metal particles at elevated temperatures

3.6 Conclusion

We have demonstrated that the nitrogen plasma treatment of FeB

nanoparticles at the eutectic melting point of the alloy can give rise to a range of BN

nanomaterials, ranging from BN bamboo-like nanotubes, to nanotubules and

microspheres, as well as iron nanowires encapsulated in BN The mechanism of the

phase segregation of BN from a dynamically flowing liquid FeB particle is

interesting, because in its path, nanotubules, and nanospheres can be created by a

shape transform of its outline, similar to mold-casting Our results suggest that

capillary flow may provide a controlled way of creating dimensional nanostructures,

i.e by creating nanochannels on a substrate to facilitate the capillary flow of FeB to

make BN nanotubes, or BN-ensheathed Fe nanowires

Reference

[1] Mishima, O.; Tanaka, J.; Yamaoka, S.; Fukunaga, O Science 1987, 238, 181

[2] Sugino, T.; Tagawa, S Appl Phys Lett 1999, 74, 889

[3] Sugiyama, T.; Tai, T.; Okamoto, A.; Toshitake, M.; Sugino, T Diamond Relat

Mater 2003, 12, 1113

[4] Iijima, S Nature 1991, 354, 56

[5] Rubio, A.; Corkill, J.L.; Cohen, M.L Phys Rev B 1994, 49, 5081

[6] Blase, X.; Rubio, A.; Louie, S.G.; Cohen, M.L Europhys Lett 1994, 28, 335

[7] Flahaut, E.; Agnoli, F.; Sloan, J.; Connor, C.O.; Green, M.L.H., Chem Mater

2002, 14, 2553

[8] Chopra, N.G.; Luyken, R.J.; Cherrey, K.; Crespi, V.H.; Cohen, M.L.; Louie,

S.G.; Zettl, A Science 1995, 269, 966

[9] Kuno, M.; Oku, T.; Suganuma, K Diamond Relat Mater 2001, 10, 1231

[10] Lourie, O.R.; Jones, C.R.; Bartlett, B.M.; Gibbons, P.C.; Ruoff, R.S.; Buhro,

W.E Chem Mater 2000, 12, 1808

[11] Huo, K.F.; Hu, Z.; Chen, F.; Fu, J J.; Chen, Y.; Liu, B.H.; Ding, J.; Dong Z L.;

White, T Appl Phys Lett 2002, 80, 3611

[12] Chadderton L T.; Chen, Y J Cryst Growth 2002, 240, 164

[13] Chen, Y.; Fitzgerald, J.; Williams J S.; Bulcock, S Chem Phys Lett 1999,

299, 260

[14] Loiseau, A.; Willaime, F.; Demoncy, N.; Hug, G.; Pascard, H Phys Rev Lett

1996, 76, 4737

[15] Golberg, D.; Bando, Y.; Eremets, M.; Takemura, K.; Kurashima K.; Yusa, H

Appl Phys Lett 1996, 69, 2045

[16] Oleg A Louchev, Appl Phys Lett 71 (1997) 3552

[17] Narita, I.; Oku, T Diamond Relat Mater 2002, 11, 949

[18] Narita, I.; Oku, T Solid State Commun 2002, 122, 465

[19] Narita, I.; Oku, T Diamond Relat Mater 2003, 12, 1146

[20] Wagner, R S.; Ellis, W C Appl Phys Lett 1964, 4, 39

[21] Oku, T.; Suganuma, K Diam Related Mater 2001, 10, 1205

[22] Hu, Z.; Fan, Y.; Wu Y.; Yan, Q.; Chen, Y J Mater Sci 1996, 31, 611

[23] Huo, K.F.; Hu, Z.; Fu, J.J.;Xu, H.; Wang, X.Z.; Lu, Y.N J Phys Chem B 2003,

[27] Komatsu, S.; Shimizu, Y.; Moriyoshi, Y.; Okada K.; Mitomo, M J Appl Phys

2002, 91, 618

[28] Babonneau, D.; Jaouen, M.; Denanot, M.F.; Guérin, P.; and Petroff, F Appl

Phys Lett 2003, 82, 3056

[29] Massalski, T.B.; Murray, J.L.; Bennett, L.H.; Baker, U.; (Eds.), Binary Alloy

Phase Diagrams, American Society for Metals, Metals Park, OH, 1986, 481

[30] Buffat, P.; Borel, J.P Phys Rev A 1976, 13, 2287

[31] Wang, Z.L.; Petroski, J.M.; Green, T.C.; El-Sayed, M A J Phys Chem B

1998, 102, 6145

[32] Egerton, R.F Electron Energy-Loss Spectroscopy in the Electron Microscope,

Plenum Press, NY, 1996

[33] Bengu, E.; Marks, L D Phys Rev Lett 2001, 86, 2385

Chapter 4 Mechanistic study of zinc sulfide nanowires

In last chapter, we successfully synthesized BN nanomaterials using VLS

method In this chapter, the growth process of one-dimensional ZnS nanowires using

VLS method will be investigated by AFM The mechanism of morphology transfer

from ZnS into ZnO nanowires will also be discussed in this chapter

4.1 Introduction

A record of observation of luminescence from zinc sulfide (ZnS) has a long

history; it dates back to 1886 when a French chemist Sidot found phosphorescence

from a ZnS crystal he grew In recent years, due to its direct and wide-bandgap of

3.72 eV, ZnS has attracted great attention for its technological utility such as

manufacture of pigments, phosphorluminescent screens and optoelectronics

Currently, ZnS nanowires can be prepared by the simple evaporation of ZnS powder

on Si substrate with Au as catalyst [1, 2]

ZnO is another important II-VI semiconductor compound with a direct and

wide band-gap of 3.37 eV and an exciting binding energy of 60 meV at room

temperature [3] Promising ultraviolet lasing action and a low power threshold for

optical pumping make it as a suitable material for blue optoelectronic applications

From previous research, high quality ZnO film can be achieved from the oxidation

of ZnS thin films at over 700 ˚C for the substitution of S by O [4-7] In the past five