Yao N. Focused Ion Beam Systems: Basics and Applications

Подождите немного. Документ загружается.

Finally, gallium can be easily distinguished from other elements, so if

implantation occurs, the gallium ions will not interfere with the analysis of

the sample. Although other ions can be used, gallium has become the ion of

choice for the focused ion beam system.

Ultimately, the purpose of precision engineering and attention to detail in

FIB design is to produce a focused ion beam that impacts the surface at a

desired point. It is important to consider the implications of this impact. A

beam of light incident upon a surface causes local temperature increase, small

electromagnetic ﬂuctuations, and photoelectron emission. A beam of charged

particles does all of this and more. The incident particles raise the tem-

perature of the area they impact, although generally not to a signiﬁcant

degree; they change the local charge densities of the region, resulting in a

temporary charge imbalance; the addition of their kinetic energy to the

energy of the sample causes secondary electrons to be emitted, which can be

captured and used to image the surface of the sample; they produce a degree

of characteristic X-ray emission, which can be used for spectroscopy pur-

poses; and, of course, they cause damage to the surface structure at different

levels depending on the physical nature of charged particles.

The emission of secondary electrons from the surface is what gives both

beam types their imaging abilities. Adding the energy of the particles in the

beam to that of the electrons in the sample allows some of those electrons to

escape from within the material, depending on the penetration depth of the

beam and the conductivity and work function of the sample. These electrons

come from a roughly spherical region around and beneath the beam spot,

with increased numbers of electrons exiting from the sides of topographical

features that would not have escaped from a ﬂat surface otherwise. In

addition to the secondary electrons, there is a degree of radiation produced

by the rapid deceleration of the charged particles, generally in the X-ray

spectrum. This radiation can later be used for X-ray spectroscopy. Between

this emitted radiation and the radiation from the accelerating potentials in

the sources, the chamber requires good shielding in order to preserve the

safety of the operator. All the modern electron and ion instruments have

handled this issue satisfactorily.

The ion beam is capable of efﬁciently and precisely depositing material.

By creating a cloud of, for example as shown in Figure 1.3, platinum atoms

above the sample, platinum can be deposited by letting the ion beam strike

these atoms, imparting some kinetic energy to them and causing them to

impact the surface, on which they remain, adsorbed onto the sample by

Van der Waals forces. This technique can be used to deposit both con-

ductive and resistive materials, since the type of atom in the suspended

Focused ion beam systems8

material does not matter, making the ion beam quite useful for nano-

fabrication purposes. This technique can also be used to improve electron-

beam imaging, by depositing a thin layer of conductor over the surface of a

feature to be imaged. Similar to sputter- or deposition-coating of a sample,

this method preserves more detail at a small scale by using a precisely

controlled, very thin layer of conductor to reduce charging effects and

deﬁne the surface [8].

The electron beam can be used, to some degree, to perform deposition.

However, due to the low mass of the electrons, the deposition occurs slowly,

and it is more feasible if the beam is very intense and focused precisely, to

increase the probability of electrons intersecting atoms of the deposition

material. In general, electron-beam deposition is not done with the beams

found in smaller scanning electron microscopes, instead requiring a larger

and more powerful apparatus. Until the advent of the focused ion beam, this

type of precision deposition was extremely difﬁcult to achieve.

While the electron beam barely affects the surface, the heavy particles of

the ion beam penetrate deeper within the lattice, kicking out atoms as they go

and embedding them in the sample. Therefore, the lithographic abilities of

Cloud of platinum atoms

with organic atoms attached

Raw platinum beam

Ion beam

Reflected ion

Organic atoms

Sample

Purified platinum

deposition

Figure 1.3 A Schematic diagram of an ion beam induced platinum

deposition process.

Introduction to the focu sed ion beam system 9

the ion beam are extremely useful. It is capable at relatively low beam cur-

rents of removing atoms from a surface in a very precise and controlled

manner; it is able to make very small cuts or take large cross sections, all

without changing the chemical or structural composition of the sample.

Unlike traditional etching methods, it does not require masking or resist

stages. The ion beam can be used to etch and mill almost any material, with

little or no sample preparation. Operated at higher currents, it can achieve

very high resolution etching at rapid speeds, with high reproducibility. In

addition, the beam can be used to implant ions within the surface of the

sample in order to tune the electronic properties of the material.

Unlike from an electron beam, collisions that result from the use of a

gallium ion beam induce many secondary processes such as recoil and

sputtering of constituent atoms, defect formation, electron excitation and

emission, and photon emission. Thermal and radiation-induced diffusion

that result from these collisions contribute to various phenomena of inter-

diffusion of constituent elements, phase transformation, amorphization,

crystallization, track formation, permanent damage, and so on. Also, pro-

cesses such as ion implantation and sputtering will change the surface mor-

phology of the sample, possibly creating craters, facets, grooves, ridges,

pyramids, blistering, exfoliation, or a spongy surface.

Because of the interrelatedness of these processes, no single phenomenon

can be understood without the discussion of several others. Therefore, it is

imperative that one possesses a quantitative understanding of the experi-

mental observations as well as creativity in design so that new and more

sophisticated combinations of these versatile processes can be applied in the

ﬁeld of nanotechnology. With it, we can aim at more advanced material

modiﬁcation, deposition techniques, implantation, erosion, nano-fabrication,

surface analysis, and many other applications.

1.3 The ion source and electron source

In order to properly understand the focused ion beam, it is necessary to

consider the source of the beam itself, as illustrated in Figure 1.4. In almost

all focused ion beam systems, a reservoir of heavy-metal atoms (typically

gallium for the aforementioned reasons) is heated to near evaporation, after

which it ﬂows and wets a sharp, heat-resistant tungsten needle with a tip

radius of 2–5 mm. Once heated, the Ga can remain liquid for weeks without

further heating due to its super-cooling properties. The Ga atoms then ﬂow

to the very end of the needle, drawn there by an annular electrode concentric

with the tip of the needle and positioned close to it, called the extractor.

Focused ion beam systems10

A potential difference between needle tip and aperture creates an electric ﬁeld

on the order of 10

V/m and causes energetic ions in the region immediately

above the tip to accelerate toward the extractor. These ions exist in a region

where the balance between electrostatic and surface tension forces has drawn

the liquid-metal into a sharp ‘‘Taylor cone’’ whose apex is only about 5 nm in

diameter. The cone tip is small enough such that the extraction voltage of the

aperture can pull Ga from the needle tip and efﬁciently ionize it by ﬁeld

evaporation of the Ga at the end of the Taylor cone, after which it is accelerated

by a potential down the ion column. The current density of ions that may be

extracted from such an LMIS system is on the order of 10

A/cm

,as

the evaporated ions are continuously replaced by a ﬂow of liquid Ga to the

cone. The source is generally operated at low emission currents of 1–3 mAto

reduce the energy spread of the beam and to yield a stable beam. At higher

emission currents, the probability of the formation of dimers, trimers, and

droplets increases, which both increases energy spread and decreases source

lifetime.

The tip of the tungsten needle is situated just above the extractor, which is

held at a voltage on the order of 6 kV relative to the source. The resulting

intense electric ﬁeld at the source tip draws the liquid Ga into a Taylor cone

and creates a tiny cusp at its end (called the ‘‘incipient jet’’). After ﬁeld

evaporation causes ion emission to occur, the ions begin to accelerate down

the column. The current emitted from the tip is known as the extraction

current. It is regulated by both the suppressor and the extractor, as shown

in Figure 1.4, which roughly correspond to ‘‘ﬁne’’ and ‘‘coarse’’ controls of

Suppress

Extract

Ion

Tungsten tip

Primary ion

Beam aperture

Electrical

feedthroughs

Insulating

mount

Coil heater

and Ga

reservoir

Figure 1.4 A cross-sectional diagram of a liquid-me tal ion source (LMIS)

found in FIB systems.

Introduction to the focu sed ion beam system 11

the emitted ion distribution. The suppressor uses an applied electric ﬁeld of

up to þ2 kV to work alongside the extractor in maintaining a constant

beam current [3]. This is particularly important to control the etching rate

during milling operations. Adjusting the suppressor voltage will change the

extraction current, which means that the extraction current may be regu-

lated without changing the voltage of the extractor. This is generally the

preferred method of adjustment, as changing the extractor voltage can result

in spatial displacement of the Taylor cone and apparent beam drift on the

sample surface. This instability corresponds to the fact that LMIS have a

highly nonlinear current–voltage relationship [9]. Adjusting the suppressor is

also very useful to offset the gradual downward drift in extraction current

that occurs in LMIS as the surrounding electric ﬁeld causes electrons to

collide with contaminants in the vacuum and ‘‘ﬁx’’ them to the source.

Thus, the ability of the suppressor to maintain constant extraction voltage

without altering the source tip is an essential requirement for FIB system

stability.

Field evaporation, the process responsible for ion production in the LMIS,

is a physically complex phenomenon, and complete treatment is beyond the

scope of this chapter. Fundamentally, however, ﬁeld evaporation takes place

when the potential barrier preventing evaporation has been lowered by the

presence of a ﬁeld and can only be crossed by the ionization of the evapor-

ating atom on the surface of the ﬁeld emitter. It can be described analytically

by ﬁrst calculating the ﬁeld needed to produce ions in free space, that is, in

the absence of any other ﬁeld. The energy Q

needed to produce an ion in free

space in this case is given by Q

¼ H

þ I

– n, where H

denotes the heat of

atomic desorption, I

refers to the ionization energy to produce an n-fold ion,

and  is the work function of the ﬁeld emitter. The term n corresponds to

the released energy when n electrons return to the metal. Taking into account

the presence of a ﬁeld E, which lowers the potential barrier for ﬁeld ionization

ﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ

, the required energy for evaporation of the liquid Ga is

¼ Q



ﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ

: ð1:1Þ

Alternatively, the atoms in the Taylor cone will ionize when

I   ¼ eEx

; ð1:2Þ

where I is the ionization potential near a metal with work function  in the

presence of an electric ﬁeld E. It has been found that the critical distance x

from the tungsten needle tip required for ion production is approximately

0.2 nm at a ﬁeld strength of 10

V/m.

Focused ion beam systems12

For comparison, we can consider one type of source of electrons in an

electron beam column. In the most common conﬁguration, a tungsten ﬁlament

is heated by a large current, causing it to emit a spectrum of radiation

accompanied by a number of loose electrons that have gained sufﬁcient energy

to overcome the work function of the metal and escape. These electrons are

then accelerated away from the tip by a set of electrostatic ﬁelds generated by

large coils, and reduced to a relatively clean beam by an aperture below the

source. The ﬁlament itself is formed into a sharp point, since this shape causes

charge to cluster at the tip, giving a greater output current from the source.

The basics of the electron gun revolve around raising an electron’s energy

above the Fermi level of the cathode material. In a standard thermal emission

gun, the energy of the electrons in the emission region is raised by the

addition of heat in the cathode and a potential from the anode, resulting in a

current density j

that follows Richardson’s Law:

¼ ATc



; ð1:3Þ

where A is a constant related to the cathode material, T

is the temperature of

the emissive tip, 

is the work function (related to the Fermi level) of the

cathode, and k is Boltzmann’s constant. This current density produces a

roughly Gaussian beam proﬁle:

jrðÞ¼j

ð

; ð1:4Þ

where r

is the radius of the emissive region on the cathode tip, usually

ranging from 10 to 50 micrometers.

The tungsten ﬁlament is also usually zirconated (ZrO) to increase the

thermionic emission of electrons by means of the Schottky effect, in which an

accelerating ﬁeld for electrons exists at the surface of the ﬁlament due to an

external applied electrostatic ﬁeld. Zirconated tungsten exhibits a decreased

work function (2.5 eV versus 4.5 eV) and provides adequate electron emission

at only about 1600 K, compared with 2500–3000 K.

Two other classes of enhanced electron guns are worth mentioning. The

potential of rare-earth hexaboride materials, especially lanthanum hexaboride

(LaB

), for thermionic electron emission was ﬁrst reported in 1951 by J. M.

Lafferty [10]. The work function of LaB

is lower than that of W (2.7 eV

compared to 4.5 eV), which, combined with its low vapor pressure at high

temperature, makes it a superior thermionic electron source for the electron

microscopes (both scanning and transmission). LaB

crystals can provide

considerably higher current densities, and LaB

offers improved coherence

and a smaller energy spread [ 11]. Extensive research in the 1970s and 1980s

Introduction to the focu sed ion beam system 13

followed Lafferty’s work in an attempt to optimize the performance of this

class of materials. The drawback to this source is mainly on its relative larger

energy spread and low beam intensity for an electron probe of nanometer

size. Cold ﬁeld emission guns (CFEGs) are another recent development in

SEM source technology. Unlike thermionic guns, they operate at relatively

low temperatures of about 300 K and offer superior resolution and perfor-

mance. FEGs are simpler in their operation – they use a pair of anodes below

the tungsten tip to generate intense electric ﬁelds that extract electrons by

enabling them to tunnel from the extremely sharp tip (Figure 1.5). As a result

of the small tip size and low operating temperature, the electron beam is

highly spatially coherent and experiences almost no energy spread, which

limits the deleterious effects of chromatic aberration; its current density is

also remarkably high [12]. The weakness of FEGs is that they can only be

under ultra high vacuum (< 10

10

), which requires more expensive

machinery.

Both ion and electron sources are subject to a pair of limitations. First,

the total current i n the beam leaving the extraction aperture is only a

fraction of that produced by the sources; the majority of the particl es

generated are blocked by the sides of the aperture, as their velocity vectors

are not pointed along the direction of the beam. This means that the

‘‘brightness’’ of the resulting beam depends on both aperture size and

source current. A larger aperture means a brighter beam with better ima-

ging and milling characteristics; however, the beam may not be a s uniform

or as precisely focused as one traveling through a smaller aperture. A

higher sour ce current means more charg ed p article s produced, bu t on ly a

First anode

Second anode

Field emission tip

Tungsten needle

Primary electron beam

Figure 1.5 A cross-sectional diagram of a cold ﬁeld emission electron source

found in high-resolution SEM systems.

Focused ion beam systems14

fraction of those will go into the beam itself, so a subst antial increase i n

current is nee ded for a marked increase in beam brightness. The inc reased

current can cause source instability and lower the li fetim e of the source

element.

A second limitation is that of uniformity. While the electrons emitted

from the zirconated tungsten are relatively evenly distributed in energy, the

ions emitted by a liquid-metal ion source (LMIS) as described above tend

to follow a Gaussian energy distribution, sometimes asymmetrical, which

leads to chromatic aberration. The energy required to evaporate ions from

a liquid is dependent on local temperature, ﬁeld strength, and surface

tensions in the area around each atom, which can vary from point to point

within the emissive region [13]. The ion emission current is strongly

dependent on the tip radius and on the tip surface condition. The sharper

the tip is, the higher the ﬁeld; and the higher the ﬁeld, the stronger the ion

emission. However, the inter-particle repulsion effect at high emission and

liquid ﬂow rates with capillary characteristics need to be balanced out in

order to maintain a stable, consistent ion beam emission. Electron emission

is determined by the work function of the metal used and the ﬁeld strength

at the source, with electrons escaping as soon as they have reached a well-

deﬁned critical energy.

Both sources experience an additional degree of energy spread due to

mutual repulsion between the charges just beyond the source. The like

charges of the ions and the electrons repel each other, imparting a small but

signiﬁcant random velocity that changes the overall energy proﬁle, causing

further chromatic aberration and forming the fundamental limit for the

focusing ability of the system. Variations in emission characteristics can be

controlled to a degree, but before the limit of that control is reached, it is

overshadowed by the aberrations arising from these mutually repulsing

charged particles.

1.4 Ion optics and electron optics

After a beam of charged particles is produced, whether of ions or electrons, it

must be focused to the desired spot size by a series of lenses, usually with one

or more apertures along the beam path as well to help control aberrations. A

lens for an ion or electron beam can be thought of in the abstract as being

almost identical to a lens for light, with a similar function and similar

parameters, such as focal length and refractive index. Standard light optics

concerns thus enter the picture, including chromatic aberration, spherical

Introduction to the focu sed ion beam system 15

aberration, and astigmatism. An additional concern speciﬁc to electron and

ion beam applications is that of apparent source size, which is related to the

inter-particle repulsion mentioned above.

Electron and ion lenses function very much like light optics, but their

construction is quite different. Instead of using a material with a certain

geometry and index of refraction to bend the path of the light, electron and

ion optics use magnetic and electrostatic ﬁelds to change the paths of the

particles.

Electron beams, consisting of fast-moving, low-mass particles, are gen-

erally focused using only magnetic ﬁelds. The advantage of a magnetic ﬁeld is

that it is relatively easy to produce a uniform ﬁeld over a region, and that

magnetic ﬁelds do no net work on objects within them, so that the kinetic

energy of the electrons is not changed. This makes it easier to determine the

results of the electron’s impact on the surface and to process and understand

the resulting data. The magnetic lenses in an electron beam column are

generally washer-shaped coils with small central holes, so that the ﬁeld is

intensiﬁed by being compressed into a smaller space. The deﬂection of the

beam is proportional to the distance from the axis of the lens, as in light

optics, so that the ﬁeld behaves just as a glass lens would.

Charged particles, however, require stronger ﬁelds to focus higher-energy

particles because charged particles have kinetic energy vectors that must be

diverted by an applied force, while light can be focused by changing the

nature of the medium through which it propagates. The higher the kinetic

energy, the higher the force needed and the higher the lens ﬁelds must be.

Higher kinetic energies correspond to shorter wavelengths, which provide

higher resolution. In order to achieve this higher resolution, however, ever-

stronger ﬁelds must be produced in the focusing column.

Ion beams, in contrast to electron beams, use electrostatic lenses almost

exclusively. The reason for this stems from the fact that the force on a

charged particle due to an electric ﬁeld E, F

¼qE is independent of the

particle’s velocity, whereas the force exerted by a magnetic ﬁeld B, F ¼qv · B,

depends directly on the velocity. A particle accelerated by a potential drop

1V will gain a speed

ﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ

2q1V=m

, where q1V is the energy and m is the

particle’s mass. Given the higher mass of ions, their velocity is 0.0028 that of

an electron accelerated by the same potential, while their momentum is 370

times higher. Magnetic optics would need to be impractically large to provide

enough focusing power for an ion beam. Electrostatic lenses, however, can be

made extremely small and are capable of producing much faster response for

Focused ion beam systems16

beam deﬂections, an important capability also used for beam blanking (a

toggling of the beam’s incidence on the sample by deﬂecting it from the

optical axis upstream from an aperture).

1.4.1 The lenses

Immediately after emission in an SEM, the electrons are accelerated by the

voltage V between cathode and anode to a kinetic energy of E ¼eV, where e is

the electron charge. After the beam is reduced to a relatively coaxial column

by an aperture, it enters the top of the electron optics column and passes into

the magnetic ﬁeld of the ﬁrst electron lens, an axial ﬁeld with rotational

symmetry about the axis of the column. This ﬁeld B

has a bell-shaped dis-

tribution along the axis where it swells out from the center of the lens coil (see

Figure 1.6). B

, the radial component of the magnetic ﬁeld, obeys the fol-

lowing equation:

r



: ð1:5Þ

Field

parallel to axis

perpendicular to axis

Figure 1.6 A cross section side view of the cylindrical magnetic lens within

an SEM column. Magnetic ﬁeld lines in a magnetic lens are shown in thin

lines, while the thick light lines represent two possible electron paths. The

smaller inset to the left gives an approximation of the magnetic ﬁeld strength

along the optical axis and perpendicular to the optical axis. The narrowing

helical rotation of electrons within the lens cannot be faithfully represented

in two dimensions but can be visualized.

Introduction to the focu sed ion beam system 17