Yao N. Focused Ion Beam Systems: Basics and Applications

Подождите немного. Документ загружается.

hard samples are also applicable to biological samples. In-situ sample

manipulation allows for site-speciﬁc morphological and structural analysis

through quasi-real time viewing of the sample during processing. The ability

to quickly mill cross sections in a variety of directions provides a strong

understanding of the relationship between cell structure and morphology.

Imaging and sectioning of biological specimens from single yeast cells to

small arthropods have provided high-resolution images of biological struc-

tures from the subcellular level to the microstructural level of tissue and

organs [3,6]. Preliminary results indicate the sample preparation must be of

primary concern during the imaging and manipulation of biological samples

[3,5,9]. Care must be taken to ensure the ﬁdelity of the biological structures

on all size scales and to minimize artefacts due to milling or sample charging.

Novel techniques for the study of ultrastructure and morphology in bio-

logical samples include FIB tomography with subsequent three-dimensional

reconstruction as well as the proposal of nano-biomachining [5,8,11]. The

incorporation of a cooled sample chamber holds the promise for minimizing

sample damage while helping to retain integrity of biological structures

without artefacts introduced by chemical ﬁxation. This chapter covers the

topics important to successful FIB milling and imaging of biological samples.

13.2 Sample preparation

Existing procedures for sample preparation have been developed and opti-

mized for effective TEM and SEM imaging. These techniques exist to

strengthen biological structures from bulk tissue down to subcellular

nanostructures. Only with extensive processing will these samples be able to

retain native conﬁgurations with minimal loss of structural components. The

general procedure for sample preparation consists of the following steps:

chemical ﬁxation, dehydration, embedment, and staining or coating. Other

alternatives to chemical ﬁxation are cryotechniques in which structural pre-

servation is accomplished through the ‘‘freezing’’ of cellular components.

Recent developments have shown that cryo-dual stage SEM/FIB systems are

adequate for imaging unprepared biological samples (Figure 13.1)[31,32].

However, chemical methods are more widely employed as samples prepared

through cryotechniques require the availability of a cold stage, and ‘‘frozen’’

cells are thermodynamically unstable at low temperatures. FIB imaging and

sectioning of biological samples is a relatively unexplored ﬁeld and use of

varying degrees of sample preservation have been employed in existing

literature [1–10,22]. Although there exist some instances in which no pre-

paration was required, chemical ﬁxation and cryotechniques remain popular.

Focused ion beam systems338

For an in-depth discussion of sample preparation the reader is referred to

Introduction to Biological Electron Microscopy: Theory and Techniques by

Clinton Dawes [12].

13.2.1 Chemical methods for sample preparation

Chemical ﬁxation

By beginning with chemical ﬁxation, cellular structure is preserved in a near

in vivo state for the subsequent processing. The success of these procedures

has been supported by comparison with samples preserved by freezing

techniques. Chemicals used for ﬁxation function by forming inter- and intra-

molecular cross linkages in proteins to form a gel, preserving cellular support

structures and the membrane system of various organelles with little struc-

tural change. The quality of a ﬁxation is indicated by the appearance of lipid

membranes making up the endoplasmic reticulum and golgi systems. The

choice of ﬁxing agent is determined by what cellular constituents are to be

preserved with the highest integrity. Common chemicals for ﬁxation include

osmium tetroxide, aldehydes, and permanganates. Osmium tetroxide reacts

with nearly all cellular components, making it the most popular choice for

chemical ﬁxation. The considerations important to successful staining with

osmium tetroxide are discussed next.

Figure 13.1 Cryo-FIB milled yeast sample originally published in an

application note from FEI company [31].

Applications for biological materials 339

Osmium tetroxide (OsO

) is a nonpolar compound, allowing unhindered

diffusion through cellular membranes. OsO

also has the ability to penetrate

hydrophilic lipids as well as aqueous areas of the cell. Osmium has ﬁve stable

oxidation states, allowing for reactions with most cellular constituents [12].

However, OsO

reacts most strongly with carbon–carbon double bonds. As a

result, osmium tetroxide reacts most strongly with unsaturated fats and

phospholipids. OsO

also partially ﬁxes proteins by reaction with phenols

and SH groups. This also allows some ﬁxation of nucleic acids by reacting

with histone proteins which are involved in DNA packaging. However, OsO

is a strong oxidizing agent and may result in the complete oxidation of

double bonds to form diols if concentration and ﬁxation duration are not

closely monitored [13].

During this procedure, temperature, pH, and osmolarity of the ﬁxing

solution are of utmost importance. Buffer solutions are required to maintain

a constant pH. As OsO

inﬁltrates the cell, the pH drops drastically,

resulting in the denaturation of proteins, and the degradation of protein

structures such as microtubules, microﬁlaments, and intermediate ﬁlaments.

For animal tissue, the buffer should maintain a pH of 7, while for plant

matter a slightly acidic environment of pH 5 to 6 is required. The most

common buffers are cacodylate and phosphate due to low reactivity with

ﬁxing agents. The ﬁxation solution should also be nearly isotonic with the

cellular environment to prevent shrinking and swelling. Also, a hypotonic

solution will result in ion leaching, which could destabilize the membrane

structures of organelles [12].

As the rate of ﬁxation is directly related to diffusion, temperature begins to

play a large role in ﬁxation duration and quality. Room temperature allows

for shorter durations, which is less disruptive to the cell; however, leaching of

cellular components is highly likely and autolysis of cell membranes will also

occur. Temperatures as low as 4



C prevent these deleterious effects, but require

a longer ﬁxation time. If the ﬁxation duration with OsO

is too long, however,

cellular components may be oxidized too far, resulting in the degradation of

cellular structures [12]. Due to the difﬁculties associated with OsO

ﬁxation,

aldehydes are often employed for primary ﬁxation of proteins, followed by

secondary ﬁxation with OsO

to provide electron density and stabilize lipid

membranes. By the end of such ﬁxation procedures, electron micrographs will

show a granular cytoplasm due to stabilization of cytoplasmic lipo-proteins

and dark staining of unsaturated fatty acids. Cell walls will appear darkened

along the edges due to low penetration into the dense cell wall. Loss of up

to 70% of carbohydrates, 21–50% of lipids, and 12–50% of proteins is com-

mon, but these numbers can be reduced by lower temperatures [12].

Focused ion beam systems340

Post-ﬁxation procedures

For preparation for vacuum, water must be removed to avoid cellular

changes upon water evaporation from biological samples. When embedding

the sample, dehydration is also required as most embedding media are not

miscible with water. Polymerization of plastics will also be hindered by the

presence of water. During dehydration, water is slowly replaced by another

solvent, usually ethanol or acetone, by slowly increasing the concentration

of dehydrant in the inﬁltrating solution. Acetone is the most popular choice

due to miscibility with many plastics, and a lack of reaction with osmium

tetroxide. It also will not cause removal of phospholipids. Ethanol is

used more sparingly due to cell shrinkage. Special dehydrating agents such as

2,2-dimethoxypropane or 2,2-diethoxypropane can be used to remove water

by reaction to produce methanol and acetone or ethanol and acetone [12].

Nonpolar dehydrants can give improved results due to increased retention of

lipids and proteins. However, cell compression and membrane expansion will

occur during freeze-drying or freeze-substitution.

Although embedding the sample in a plastic or resin is most essential when

sectioning in preparation for TEM imaging, sample embedding is also

important for the imaging of animal cells without cell walls. Embedment

requires the use of a low viscosity plastic to inﬁltrate the sample. Such plastics

can not have volume change upon polymerization, and have low electron

scattering effects. Most importantly, the plastic must be stable under ion

beam bombardment. To provide the most uniform sectioning, the polymer

embedment must be similar to the sample in both density and chemical

composition as both parameters affect etching rate. Common embedments

include vinyl plastics, methacrylates, epoxy resins, and polyester resins [14].

Drobne et al. have used paraplast for the embedding of a crustacean digestive

system [8]. Parafﬁn is another popular choice, although maintaining hot

parafﬁn during preparation is difﬁcult. If the sample is to be imaged in a

dual-beam system, stability under electron beam bombardment must also be

considered. Epoxies are most stable, while methacrylates are least stable [14].

Staining or coating is required for biological samples in which a large content

of carbon, oxygen, and nitrogen prevent high contrast due to low electron

density. As a result, coating with heavy metals or post-ﬁxation heavy metal

staining is required. Staining refers to either chemical or physical incorporation

of heavy atoms into the sample. Common stains include metals such as tung-

sten, ruthenium, osmium, and lead. These metal ions form coordinate com-

plexes with active groups within the sample, adhering to certain structures to

increase contrast. For surface morphology, thin ﬁlms of gold or palladium are

Applications for biological materials 341

deposited on sample surfaces by sputter coating or high vacuum evaporation.

Thin ﬁlm coatings have the advantage of being an in-situ process that can be

performed within the FIB as new sections are generated.

13.2.2 Cryotechniques for sample preparation

Crytotechniques can be used with or without prior chemical ﬁxation and can

be performed on living or ﬁxed samples. These techniques are often com-

bined to prepare samples for vacuum by solvent removal. For bulk samples

such as arthropods with strong exoskeletons or even yeast cells with thick cell

walls, sample embedment is not required; however, removal of liquid within

the sample is important [3,9,10]. During ordinary drying, surface tension

causes specimen distortion as the liquid volume within the sample decreases

[14]. Cryotechniques minimize these effects through increased evaporation or

elimination of surface tension effects.

The favored techniques for this task are freeze-drying and critical point

drying. Freeze-drying employs vacuum induced sublimation of the solvent to

accomplish solvent removal. The largest drawback of freeze-drying is sample

shrinkage during freezing. As a result, critical point drying is the more

popular choice for solvent removal. During critical point drying, the sample

is heated until the solvent reaches its critical point, allowing the solvent to be

bled off without surface tension effects or other interactions that could

destabilize cellular structures. The critical points of liquid solvents occur at

temperatures and pressures too high for practical application, therefore,

carbon dioxide or Freon are the preferred solvents. Nevertheless, surface

tension damage is not completely avoidable, and a minimal amount of

shrinkage is still observed.

Yonehara et al.[1] have imaged the small intestine of Wister mice which

had been prepared using critical point drying. These samples were ﬁrst ion

beam etched to reveal the small intestine where microvilli were imaged by

SEM to obtain high-resolution images after coating with 10 nm of heavy

metal [1] (Figures 13.2 and 13.3).

13.2.3 FIB imaging without sample preparation

Ballerini et al.[6] have directly imaged and sectioned yeast cells without prior

sample preparation (Figure 13.4). Yeast cells on ﬁlter paper were directly

transferred into the sample chamber. The semipermeable membrane allows

for the evaporation of water from within the cell, while the thick cellular wall

provided the structure to prevent sample shrinkage and collapse. Despite a

Focused ion beam systems342

Figure 13.2 Apical view of microvilli of mouse small intestine taken with

SEM at 3 kV after ion beam etching to remove surface mucosa [1].

Figure 13.3 Side view of microvilli of mouse small intestine taken with SEM

at 3 kV after ion beam etching to remove surface mucosa [1].

Applications for biological materials 343

lack of chemical ﬁxation, ultracellular support structures were clearly visible.

They have suggested that imaging of cells such as lymphocytes or chon-

drocytes will also be possible without much sample preparation (Figure 13.5).

Yeast cells may be a special case, however, as they are known to survive

extreme conditions including complete dehydration.

13.2.4 Sample mounting

Due to sample embedding, the capsule must be trimmed to reveal the

specimen before imaging. Special considerations must be taken to avoid

build up of charge on the sample during ion beam or electron beam

bombardment depending on the system. Copper tape connecting the sample

to the sample holder is an option that is simple and easy to apply. Con-

ductive paints such as silver or carbon have also proven successful in

maintaining conduction between nonconducting polymer samples and the

sample holder [15]. Complete coatings of iridium, platinum, or gold have

been successful for SEM imaging; however, due to ion beam bombardment

in the FIB, these methods are less successful as they will be removed. In-

situ coatings are possible using ion assisted deposition for small cross

Figure 13.4 FIB microsection of yeast cells taken by Ballerini et al. with no

sample preparation [6].

Focused ion beam systems344

sections, which can increase the effectiveness of dual-beam systems in which

SEM imaging is possible.

If serial ultra-thin sectioning is to be performed in a dual-beam system, the

angle of the sample face with regard to the ion and electron beams is

extremely important. The ion beam must have a grazing angle of incidence to

decrease damage to the sample such as amorphous surface layers and ion

incorporation into the sample. The conﬁguration of the ion and electron

beams must be considered when serial cross sections are examined. Dual-

beam machines currently on the market have an FIB at between 48 and 52



from the normal electron beam. Due to the limited tilt of the stage, to obtain

a grazing incidence angle of the ion beam, the sample face must be tilted as

illustrated in Figure 13.6.

13.3 Ion–sample interactions

Understanding the reactions that occur when energetic ions interact with

sample surfaces is essential to preparing high-quality images and uniform cross

sections. As with traditional FIB applications within the ﬁeld of semi-

conductors and metals, interactions ranging from ion implantation to sample

sputtering and secondary electron emission involve the transfer of energy from

Figure 13.5 Chondrocyte imaged after FIB sectioning by Milani et al.[7].

Applications for biological materials 345

the incident ion to the solid surface. Knowledge of these interactions, speciﬁ-

cally molecular sputtering theory, enhances the ability to utilize the focused ion

beam as an effective milling device. Unfortunately the FIB’s imaging cap-

abilities are compromised by artefacts resulting from ion bombardment. Due

to the nature of organic materials, artefacts speciﬁc to imaging of biological

samples become a matter of importance. Careful study of reactions resulting

from an incident beam of ions provides a foundation for the analysis and

minimization of artefacts generated during FIB milling and sectioning.

With both hard (semiconductors, ceramics, metals, etc.) and soft (organic)

materials transfer of momentum from incident ions to the sample surface is

the result of elastic collisions and inelastic interactions. The former leads to

physical sputtering of surface molecules if the kinetic energy of the

impinging ion beam is greater than the bonding energy of those molecules.

Complex inelastic interactions with soft materials give similar results to

those noted in hard materials such as the generation of secondary electrons

and phonons. All ion interactions with the sample result in a decrease in

kinetic energy of the impinging ions. If the ions are not backscattered, they

will remain implanted within the material. Artefacts such as defects, sample

heating, and amorphization typical of hard materials may have even more

complex effects in organic materials. Denaturation of proteins, instability of

lipid proteins and preferential sputtering may result. Detailed analysis of

ion–solid interactions in traditional hard materials are provided in the work

by Nastasi et al.[16].

13.3.1 Molecular sputtering theory

Physical ejection of atoms from the surface of a target depends on the kinetic

energy of the incident electron beam and the energy with which atoms are

Electron beam

Ion beam

Embedment

Figure 13.6 Proposed ample embedment angle to achieve grazing surface

angle to mini mize damage during cross-sectioning with eucentric stage.

Focused ion beam systems346

bound to the surface (sublimation energy). The energy transfer depends on

the size, mass, and charge of the ions and atoms. The power potential law

proposed by Lindhard et al.[17] suggests that the power potential between

ions and atoms varies as

VðrÞ¼

n1

; ð13:1Þ

where Z

and Z

are the atomic numbers of the incident ion and target atom,

respectively, r is the distance between the two interacting atoms, and e is the

charge of an electron. n depends on the type of collision: n =1 for Rutherford

type, n ¼2 for weakly screened collisions, and n ¼5 for collisions of hard

spheres. a is the effective screen radius of the atoms as given by

a ¼ 0:8853 · a

· ðZ

2=3

þ Z

2=3

1=2

; ð13:2Þ

where a

is the Bohr radius for the hydrogen atom.

Kanaya et al. (1988) enhance this theory to develop atomic sputtering

theory by accounting for the weak-screening effects and hard sphere colli-

sions [2]. Molecular sputtering theory as proposed by Kanaya et al. (1992)

builds upon atomic sputtering theory to account for bonding energies of

molecules [18]. For samples of a pure solid, the sputtering yield is dependent

on the nuclear stopping cross section, the number of atoms per unit volume,

and the number of primary knock-on atoms per incident particle, as well as

the number of atoms that recoil as a result of collision cascade effects.

The proposed semi-empirical relationship for the sputtering yield (ejected

atoms per incident ion) as a function of the energy of the incident beam (E)

derived by Kanaya et al.[2,18] takes the form

¼ Y



1=21=n

: ð13:3Þ

The maximum theoretical yield Y

is given by

0:45 · Z

ð1 þ M

ðZ

2=3

þ Z

2=3

1=2

; ð13:4Þ

and E

is given by

0:4Z

ðM

þ M

; ð13:5Þ

where the coefﬁcients 0.45 and 0.4 have been determined experimentally, M

and M

are the atomic masses of the incident ion and target atom, respectively,

Applications for biological materials 347