Download Transmission Electron Microscope - Lecture Notes | MSE 421 and more Study notes Materials science in PDF only on Docsity!
IV. The Transmission Electron Microscope
A. The instrument
The first transmission electron microscope was invented in 1933 by Max Knoll and Ernst Ruska at the Technical College in Berlin. The transmission electron microscope is the electronic cousin of the transmission light microscope: a beam of electrons passes through a thin sample followed by a series of lenses, forming a highly magnified image of the sample on a screen. Knoll and Ruska found that they could focus their electron beam with a magnetic lens that was produced by sending the beam through a current-carrying coil. Modern transmission electron microscopes usually consist of a beam column that is about 2.5m tall with a diameter of about 30cm, and they are able to achieve a resolution of about 2Å; however, the addition of aberration correctors has more than doubled this performance.
Source
First condenser lens, C1 (Spot Size)
Second condenser lens, C2 (Brightness) Condenser aperture
Specimen
First demagnified source image
Objective lens Back focal plane
First image plane
Projector Lens(es)
Intermediate or diffraction lens
Objective aperture
Second image plane
Screen
Selected area aperture
Source
First condenser lens, C1 (Spot Size)
Second condenser lens, C2 (Brightness) Condenser aperture
Specimen
First demagnified source image
Objective lens Back focal plane
First image plane
Projector Lens(es)
Intermediate or diffraction lens
Objective aperture
Second image plane
Screen
Selected area aperture
- Electron gun At the top of the column, the electron gun†^ delivers high-energy electrons to the instrument. Thermionic guns (tungsten or LaB 6 ) are the most common types. The appropriate electron energy depends on the nature of the specimen and the kind of information required. Higher electron energies allow thicker samples to be analysed and, due to their smaller wavelengths, increase the resolution possible; however, it is rare now to see TEMs which operate at energies greater than 200 keV. The introduction of field emission guns and improvements in lens design have largely made higher-energy microscopes unnecessary for high-resolution. Additionally, higher energy electrons cause increasing amounts of damage to samples. Biological samples in particular require lower operating voltages.
- Condenser lens system The condenser lens system then acts to control and reduce the diameter of this beam. The first condenser C1 lens (or spot size ) is a strong lens which demagnifies the image of the electron source by about X1/100 to give a small “point” source at the “crossover” which is more coherent than the large (50 μm diameter) filament tip. The second condenser C2 lens ( brightness or intensity ) is a weaker lens which projects the demagnified source image onto the specimen with a magnification of X2, giving an overall demagnification of X1/50. This lens controls the spread of illumination on the screen. The condenser aperture, located just below the condenser lenses (sometimes between them), collimates (makes parallel) the electron beam and modifies its intensity.
Three parameters control the operation of the electron gun: the accelerating voltage , the filament current (and hence its temperature), and the bias voltage on the Wehnelt cap.
The filament current controls the filament tip temperature and hence the number of electrons emitted. The emission is maximised by saturating the filament, i.e. , increasing the filament current until the number of electrons emitted no longer increases. Ramping the filament current to saturation is controlled electronically on the JEOL 2100 and should not need to be adjusted by users.
The gun bias controls the bias resistor setting, which controls the current passing between the high-voltage system and earth. At low bias values, the negative potential of the Wehnelt compared to the filament is ineffective; therefore, the electrons are accelerated towards the anode with relatively little focusing. The beam is consequently spread and appears weak on the screen. As the bias is increased, the focusing action improves so that the effective beam brightness increases; however, above a certain value the Wehnelt is so negative with respect to the filament that the brightness starts to decrease because electrons are prevented from being emitted from the filament or, if they are emitted, are repelled back towards the filament.
The distance between the Wehnelt and the filament is obviously important in determining the point at which the optimum beam brightness is obtained. For this reason, if the filament has recently been changed, a slightly different emission setting may be required.
† (^) Never try taking one of these through airport security!
First demagnified source image
Back focal plane
First image plane
Second image plane
Source
First condenser lens, C1 (Spot Size)
Second condenser lens, C2 (Brightness) Condenser aperture
Specimen
Objective lens Objective aperture
Intermediate or diffraction lens
Projector Lens(es)
Screen image diffraction pattern
Selected area aperture
First demagnified source image
Back focal plane
First image plane
Second image plane
Source
First condenser lens, C1 (Spot Size)
Second condenser lens, C2 (Brightness) Condenser aperture
Specimen
Objective lens Objective aperture
Intermediate or diffraction lens
Projector Lens(es)
Screen image diffraction pattern
Selected area aperture
- Practical notes The value of the filament saturation will generally decrease with time as the filament material (W, LaB 6 , etc. ) evaporates, making the filament smaller. Very low settings of the filament control are often a sign of imminent failure.
On the JEOL 2100 the optimum emission setting is often about 63%. Never operate the microscope above this pre-set value.
Contamination (“poisoning”) of the filament is also a potential problem. It is very important to maintain a good vacuum in the gun area. Contamination causes high-voltage break-down (arcing) which is manifest as voltage instability or complete shut-down caused by the loss of vacuum as particles of contamination are vaporised. In addition, a poor vacuum might cause oxidation of a W filament.
Only ever turn on the electron beam when there is a good vacuum in the column (≤ 2.5 x 10-5Pa). Similarly, the gun area should be allowed to cool before the column is brought up to atmospheric pressure (this should only be done by BSCMC personnel). All parts of the gun area should be handled using appropriate gloves.
There is no guarantee that when the filament has been changed that it points directly down the optic axis. Clearly, the beam must be on the optic axis to maximise brightness. On the JEOL 2100 the beam is centred using the gun tilt controls.
- Vacuum System Electron microscopes are operated under vacuum for four reasons:
- Because electrons scatter easily, the mean free path of electrons at atmospheric pressure is only about 1cm; however, at 10-6^ Pa they can travel about 6.5m.
- The vacuum acts as an insulator between the anode and cathode (filament) and in the area around the field emitters, thus hindering unwanted electrical discharge in the electron gun.
- The elimination of oxygen around the filament prevents it from being oxidised and eventually “burning out.”
- Reduced interaction between the electron beam and gas molecules decreases contamination on the sample.
The SI unit of pressure is the Pascal (Pa). Unfortunately, there are several other units for pressure also in common use. For convenience, a conversion chart is shown below. A further complication is that, despite the fact that vacuums necessarily require low pressures, we perversely refer to very low pressures as high vacuums!
1 atm 1 bar 760 mm Hg 760 Torr 101325 Pa 14.696 psi
Different levels of vacuum are required for different parts of the microscope. The gun might require 10-9^ Pa, while the specimen can be at 10-6^ Pa and the projection chamber and camera can be at 10-5^ Pa.
There are three main types of vacuum pump: roughing pumps, high-vacuum pumps which need backing, and high-vacuum pumps which do not need backing. Alternatively, vacuums can be categorise vacuums as rough (100 - 0.1 Pa), low (10-1^ - 10-4^ Pa), high (10-4^ - 10-7^ Pa), or ultrahigh (< 10-7^ Pa).
a. Roughing Pump (Rotary Pump) A rotary pump can pump down from atmospheric pressure but can only reach a rather modest vacuum, at best about 1 - 0.1 Pa. They consist of a belt-driven, eccentrically mounted reciprocating mechanism suck air through an inlet valve into a chamber and expels it through an exit valve. Oil is typically used as a lubricant and gas seal, although expensive oil-free models are also available. The high-pressure side can operate at atmospheric pressures, while the low- pressure side is usually limited by the vapour pressure of the pump oil used. These pumps are generally reliable, inexpensive, noisy, and dirty.
pump cannot exhaust directly to atmospheric pressure, the forepump is used to maintain proper discharge pressure conditions.
Diffusion pumps use a high velocity stream of molecules to kinetically ‘ trap ’ random gas molecules from the vacuum system that blunder into the stream. They basically consist of a stainless steel chamber containing vertically stacked cone-shaped jet assemblies, each of which will support pressure ratios of approximately 10:1 or greater. Typically there are three jet assemblies of diminishing sizes, with the largest at the bottom. The pressure on the low-pressure side is typically around 10-4^ Pa or so, while the maximum pressure on the high pressure side is typically on the order of around 7 Pa. At the base of the chamber is a pool of a specialized type of oil having a low vapour pressure. The oil is heated to boiling by an electric heater beneath the floor of the chamber. The vaporized oil moves upward and is expelled through the jets in the various assemblies. The high-energy oil droplets travel downward in the space between the jet assemblies and the chamber wall at speeds up to 335 ms-1. The droplets may actually exceed the speed of sound, but thankfully there is no sonic boom because the molecules in the partial vacuum are too far apart to transmit the sound energy. The capture efficiency of the vapour jet depends on its density, velocity, and molecular weight. The high-velocity jet collides with gas molecules that happen to enter it due to their thermal motion. This typically imparts a downward motion to the molecules and transports them towards the pump outlet, creating higher vacuum at the top of the pump chamber which is connected to the microscope column. At the base of the chamber, the condensed molecules of atmospheric gases are removed by the forepump, while the condensed oil begins another cycle. Water circulated through coils on the outside of the chamber cool it to prevent thermal runaway and permit operation over long periods of time.
The first designs of diffusion pumps go back to 1915, when the pump was invented by Irving Langmuir. The original diffusion pump fluid used was mercury, which could withstand elevated temperatures but had the disadvantage of being toxic - and volatile. Indeed, the oil in a modern diffusion pump would contaminate the vacuum in the TEM if oil vapor were to escape into the column. To minimize this problem, several kinds of synthetic non-hydrocarbon oils with low vapour pressures, based on silicones, esters, perfluorals, and polyphenyl ethers, can be used. The polyphenyl ether Santovac®^ 5 and the perfluorinated polyethers Fomblin®^ and Krytox®^ have been worldwide standards for some time. These oils are expensive ($1-$2 per ml) and may eventually crack or char (oxidise) if subjected to both high temperature and pressure, forming a tar-like substance that can be extremely difficult to remove.
Exactly what happens when diffusion pump oil cracks depends on the type of oil in the chamber and the temperature at which it breaks down. Breakdowns with hydrocarbon-based oils are hardest to clean up. The residue is much like tar and is tenacious. Silicone-based oils leave a residue that is somewhat easier to remove. Oils containing perfluorals decompose to form fluorine compounds that can be extremely toxic and very damaging to the aluminum jet assemblies. The least messy and least damaging are the polyphenyl ethers, in part because their breakdown temperature is much higher (350°C vs 300°C), and in part because polyphenyl ethers tend to decompose into small non-toxic molecules such as water and carbon dioxide. Thankfully, breakdowns involving polyphenyl ethers are rare.
The diffusion pump is usually positioned at the bottom of the column, near the back. It is used to pump the viewing chamber and the photographic film casement. There is usually a large empty volume between the back end of the diffusion pump and the roughing manifold, so that when the microscope is being used, the roughing pump can be switched off (it creates a lot of vibration). The empty volume is very slowly filled with the gas being pumped out of the chamber by the diffusion pump. From time to time, the rotary pump will come on to empty the backing line. This is the loud mechanical noise you hear every few hours. If the rotary pump or cooling system fails, oil can backstream into the microscope. In all diffusion pumps, a small amount of backstreaming occurs. A liquid nitrogen cryotrap is sometimes used to remove oil particles before they can reach the microscope column (the JEOL 2100 does not have this feature); but for most applications of microscopy, the use of polyphenyl ethers makes a cryotrap unnecessary, since the high purity of polyphenyl ethers minimizes backstreaming.
Although they have been replaced in some applications by more advanced designs such as cryopumps or ion pumps, vacuum diffusion pumps are still plentiful because they have several advantages: they are reliable, they are simple in design, they run without noise or vibration, and they are relatively inexpensive to operate and maintain. In fact, diffusion pumping is still the most economical means of creating high vacuum environments.
c. Turbomolecular Pump/Turbo Pump Turbo pumps are basically like jet engines. Multiple stages of rotating blades (rotors) and fixed blades (stators), which act as axial compressors (turbines), create a powerful downdraft and force air out of the microscope. The blades are designed like airfoils to enhance the flow of gas, and the turbines spin very high speed (typically 20,000 - 50,000 rpm) so are more likely to fail than diffusion pumps. The typical description of a turbo pump failure is of a brief “scream of anguished metal”, followed by the entire vacuum system being filled with aluminium confetti. Given the high speeds involved, a foreign particle, faulty bearings, etc. will cause a rather spectacular failure.
They also do not contain oil, but the moving parts can introduce some vibration (although the best turbo pumps are very quiet and almost vibration-free). In addition, the tight tolerances required in the machining of the blades means that these pumps are generally expensive.
Pa is more common. It is common to add ion pumps directly to the stage or gun chamber of a TEM to concentrate their pumping action on these important parts.
e. Cryogenic/Adsorption Pumps Cryogenic pumps are also oil-free and rely on liquid nitrogen to cool molecular sieves with large surface areas. The cold surface removed air molecules from ambient pressure down to ~10-4^ Pa.
The anticontamination device (ACD) is a cold trap that immediately surrounds the specimen in most TEMs and acts as a minicryopump, trapping volatiles as they are produced from interaction of the beam with the specimen. The ACD provides an alternative site (to your specimen) for condensation of residual contamination in your vacuum. This is an important way to keep the internal components of the TEM clean. Once the beam is off and the trap warms up the trapped gasses are released and removed via the normal pumping system. It is important that the sample be removed before this happens.
Ion pumps and crypumps are trapping pumps. They keep the air molecules within them and release them when turned off or warmed up, respectively.
f. Vacuum Tube Gauge (Pirani Gauge) A Pirani gauge, named for its German inventor, Marcello Pirani, in 1906, uses a platinum wire in a sealed vacuum tube and a second wire in specimen chamber. A voltage of 6-12 V is applied to heat the wires. A heated metal filament suspended in a gas will lose heat to the gas as its molecules collide with the wire, removing heat and accelerating in the process. If the gas pressure is reduced the number of molecules present will fall proportionately and the wire will rise in temperature due to the reduced cooling effect. So, the hotter the wire, the better the vacuum. The vacuum is measured indirectly by the current which flows through the wire. The higher the temperature of the wire, the greater its electrical resistance and the less current will flow. The difference in current flow between the known vacuum in the closed tube and the unknown vacuum in the instrument gives an indication of the vacuum in the chamber. In many systems the wire is maintained at a constant temperature and the current required to achieve this is therefore a measure of the vacuum being studied.
Pirani gauges are useful for pressures between about 70 Pa to 0.01 Pa. The thermal conductivity of the gas may effect the readout from the meter, and therefore the apparatus may need calibrating before accurate readings are obtainable
g. Ion Discharge Gauge (Penning Gauge) Penning gauges are the most sensitive gauges for very low pressures (10-1^ to ~10-8^ Pa). They sense pressure indirectly by bombarding a gas with electrons, thermionically emitted from a filament, and then measuring the current produced by the positive ions created. The ions are attracted to the cathode, known as the collector , which is biased by several kV with respect to the anode or grid. The current in the collector is proportional to the rate of ionization, which is a function of the pressure in the system. Fewer ions will be produced by lower density gases. Hence, measuring the collector current gives the gas pressure.
The calibration of a Penning gauge is difficult and dependent on the nature of the gases being measured, which is not always known.
Objective aperture removes scattered off-axis beams
Sample
Objective Aperture
Objective aperture removes diffracted beams
Electron Beam
Objective aperture removes scattered off-axis beams
SampleSample
Objective Aperture
Objective Aperture
Objective aperture removes diffracted beams
Electron BeamElectron Beam
a. Perfect crystals If a beam of electrons is incident upon a crystal of thickness t , the diffracted intensity for a given reflection can be calculated as:
2
ξ π
πsinπ
s
ts I g
g
where s is the deviation parameter (the deviation of the g vector from the Ewald sphere, or the distance in reciprocal space from the exact Bragg condition) and ξ g is a material constant (for a given g ) called the extinction distance. Although the maximum diffracted intensity will always be for s = 0, some diffraction will also occur for s ≠ 0. The thinner a crystal is, the further it can deviate from the Bragg condition and still diffract. This rule (which is also true for other forms of diffraction, eg. XRD) can be thought of with reference to the shape and size of the rel rods (II.C.7.) in reciprocal space and their intersections with the Ewald sphere, which we have already discussed. As the crystal thickness increases, the rel rods become shorter and so make fewer intersections with the Ewald sphere – there are fewer reflections in such diffraction patterns. Absorption and increased inelastic scattering also means such patterns show more diffuse scattering and eventually, when the thickness is too great, no intensity at all gets through.
The extinction distance is given by:
g
c F
V
λ
π cosθ ξ (^) g = B
where Vc is the volume of the unit cell, θB is the Bragg angle, λ is the electron wavelength, and Fg is the structure factor. The amount of contrast in a specimen, the apparent size of a defect, the
appearance of stacking fault fringes, thickness fringes, and bend contours are all determined by ξg.
In general, sharp images are only obtained when ξ g is small (a few tens of nm). According to the equation above, in order to minimise ξ g , θB must be made small and Fg large. These two conditions are only satisfied for low-index reflections.
A plot of Ig versus s is called a rocking curve , since variation in s can be achieved by rocking a flat specimen through the Bragg condition. Such a rocking curve is shown below. The yellow curve represents the equation as given above, which is only valid for very thin crystals where the intensity of diffracted beams is negligible (kinematical conditions). The orange curve allows for dynamical effects, and the red curve allows for both dynamical effects and absorption.
Deviation Parameter (nm-1)
Intensity of Diffracted Beam
t /ξ
g = 1.
ξ g /ξ g’ = 0.
ξ g /ξ o’ = 0.
Deviation Parameter (nm-1)
Intensity of Diffracted Beam
t /ξ
g = 1.
ξ g /ξ g’ = 0.
ξ g /ξ o’ = 0.
A buckled specimen provides a range of s values without the need for rocking and so produces fringes just like a rocking curve. The bend or extinction contours which result often reflect the symmetry of the crystal and, in fact, have been used in the past to orient crystals.
Bend contours can easily be distinguished from actual crystalline defects by tilting the specimen. Bend contours will seem to sweep across the specimen as it is tilted and different
Dark-field bend contours in Pb 3 Nb 2 O 8.
can be used to describe the amplitude of both the diffracted (φ g ) and undiffracted (φo) beams as a function of z , the distance through the crystal:
g ( ) g
g g i isz
i dz
d
isz i i dz
d
φ ξ
π φ exp 2 π ξ
φ π
φ exp 2 π ξ
π φ ξ
φ π
o
o o
o o
o
The first term arises from the scattering from the transmitted beam and the second from the diffracted beam. The amplitude of each wave changes with z due to contributions from the other. The possibility of absorption (high-angle inelastic scattering) can be accounted for by replacing 1/ξ with the complex parameter 1/ξ + i /ξ'.
In order to calculate the intensity, the equations must be integrated over the entire thickness to give φo and φ g at the exit surface of the specimen. The bright-field intensity is then given by
φo φo and the dark-field intensity by φ φ* g g , where^
* (^) indicates the complex conjugate. The
intensity of the diffracted beam is then:
2
ξ π '
πsinπ '
s
ts I g
g
which is exactly the same as the kinematical solution except for the use of the effective deviation parameter, s', where:
2 ' 2 1
g
s s
b. Defects A defect which disturbs the crystal planes will locally modify the deviation parameter. In this case, the Howie-Whelan equations can be re-written as:
g ( ) g
g g i isz i dz
d
isz
i i dz
d
φ ξ
π φ exp 2 π ( ξ
φ π
φ exp 2 π ( ξ
π φ ξ
φ π
o
o o
o o
o
g R
g R
where R is the displacement of atoms from their lattice positions due to the defect ( eg. , a Burgers vector), and g • R modifies the product sz. When g • R = 0 (or an integer), the defect has no
effect on the diffracting planes and so it is invisible. This condition is called the invisibility criterion , and occurs when g is perpendicular to R. Stacking faults, grain boundaries, and phase
boundaries can all be studied in this way. The larger g • R is, the more obviously visible will be
the defect.
Stacking faults will result in defect fringes which are identical in contrast in both bright-field and dark-field images above the fault but complementary below it.
If one grain oriented in a two-beam condition overlaps another which is not, then the first grain can show thickness fringes just as if it were a tapered single crystal (which it is). Such fringes will be parallel to the intersection of the grain boundary with the surface and can easily be distinguished from stacking fault fringes by dark-field imaging, in which case only the strongly diffracting grain will appear bright.
When both crystals are strongly diffracting, moiré fringes may appear. These are common in images of thin crystalline materials deposited on each other, where two crystals are diffracting with slightly different values of g or are rotated slightly with respect to each other.
d 1 d 2 Dp
d 1 d 1
Dr
Parallel Moiré Patte rn
Rotation Moiré Patte rn
dd 11 dd 22 Dp
dd 11 dd 11
Dr
Parallel Moiré Patte rn
Rotation Moiré Patte rn
In the case of parallel lattices, the net effect is a set of fringes running perpendicular to g with a spacing:
1 2
1 2 d d
dd Dp −
and in the case of lattices only rotated by an angle α, the spacing is:
1 2 sin
d Dr =
g = 002
g = 113
g = 111
g = 220
B = [110]
g = 002
g = 113g = 113g = 113
g = 111g = 111
g = 220g = 220
B = [110]
For a bright-field image, the objective aperture is placed around the transmitted beam, excluding the diffracted one. The vector from the transmitted beam (000) to the diffracted one ( hkl ) is called g.
For a dark-field image, the incident electron beam is tilted so that the diffracted beam is parallel
to the optic axis by using deflectors. In fact, it is ( h kl ) which is brought to the centre, and the associated vector is therefore g.
Io
object
objective lens
objective aperture
It
Io
Id=Io-It
2 θ (^2) θ
g (^) T T g
Bright field Dark field
Io
object
objective lens
objective aperture
It
Io
Id=Io-It
2 θ (^2) θ
g (^) T T gg
Bright field Dark field
[101]
v g = 440
[110]
(111) layering visible (111) layering visible
[101]
[110]
(111) layering visible(111) layering visible (111) layering visible(111) layering visible
200 nm 200 nm
[101]
[110]
(111) layering visible(111) layering visible (111) layering visible(111) layering visible
[101]
v g = 440
[110]
(111) layering visible(111) layering visible (111) layering visible(111) layering visible
200 nm 200 nm
v g
[101]
v g v g == 440440
[110]
(111) layering visible(111) layering visible (111) layering visible(111) layering visible
[101]
[110]
(111) layering visible(111) layering visible (111) layering visible(111) layering visible
200 nm200 nm 200 nm200 nm
[101]
[110]
(111) layering visible(111) layering visible (111) layering visible(111) layering visible
[101]
v g
v g = 440
[110]
(111) layering visible(111) layering visible (111) layering visible(111) layering visible
200 nm200 nm 200 nm200 nm
v g
v g
Bright Field
[101]
[110]
200 nm 200 nm
(111) layering visible (111) layering visible
[101]
[110]
200 nm 200 nm
(111) layering visible(111) layering visible (111) layering visible(111) layering visible
[101]
[110]
200 nm200 nm 200 nm200 nm
(111) layering visible(111) layering visible (111) layering visible(111) layering visible
[101]
[110]
200 nm200 nm 200 nm200 nm
(111) layering visible(111) layering visible (111) layering visible(111) layering visible v g
v g
[101]
[110]
200 nm200 nm 200 nm200 nm
(111) layering visible(111) layering visible (111) layering visible(111) layering visible
[101]
[110]
200 nm200 nm 200 nm200 nm
(111) layering visible(111) layering visible (111) layering visible(111) layering visible
[101]
[110]
200 nm200 nm 200 nm200 nm
(111) layering visible(111) layering visible (111) layering visible(111) layering visible
[101]
[110]
200 nm200 nm 200 nm200 nm
(111) layering visible(111) layering visible (111) layering visible(111) layering visible v g
v g
v g v g
Dark Field
- Phase contrast and high-resolution imaging Unlike absorption and diffraction contrast mechanisms, which rely on the amplitude of scattered waves, phase contrast results whenever electrons of a different phase pass through the objective aperture. If spots along a systematic row are allowed through, a lattice image is formed. Such images can be used to show the extent of crystallisation of a grain-boundary film or the habit plane of planar defects. If more diffracted beams are allowed to contribute, then a structure image can be formed. Interpreting such images is not trivial and requires knowledge of specimen thickness, defocus, and TEM resolution (itself dependent on Cs of the objective lens
