All electron beam instruments are built around an electron column, which produces a stable electron beam, controls beam current, beam size and beam shape, and rasters the beam. Electron optics are a very close analog to light optics, and most of the principles of an electron beam column can be understood by thinking of the electrons as rays of light and the electron optical components as simply their optical counterparts.
There have been great technical advances in microprobes and scanning electron microscopes during the last 30 years (mostly in automation and electronics), but all instruments are built around an electron column, which produces a stable electron beam, controls beam current, beam size and beam shape, and rasters the beam for SEM work. Scanning electron microscopes and electron microprobes have very similar electron columns. Electron optics are a very close analog to light optics, and most of the principles of an electron beam column can be understood by thinking of the electrons as rays of light and the electron optical components as simply their optical counterparts.
Casmeca Camebax MBX Microprobe overview.
Scanning Electron Microscope overview.
A pumping system must be employed to remove air from the electron column. In most microprobes and scanning electron microscopes, vacuum is achieved through a combination of mechanical and diffusion pumps. Various valves allow sequential pumping of the electron column and sample chamber. The two main units used to measure pressure (vacuum) are the torr and Pascal. Standard atmospheric pressure (STD) is 760 torr or 1.01 x 105 Pascal. (1 torr = 133.32 Pascal; 1 Pascal = 0.0075 torr). The electron column and wavelength dispersive spectrometers must be operated under high vacuum for five reasons:
- To produce a mean-free path for electrons greater than the length of the electron column. This corresponds to a vacuum of better than about 10-4 torr (<0.1 Pa);
- To avoid arcing between the cathode (filament) and the anode plate. There is very high voltage between these two components and stray air or gas molecules can cause electrical arcing between them. The dielectric strength of air depends strongly on pressure. To maintain a voltage of 20 keV between the Wehnelt and anode plate at a pressure of 10-4 torr, requires a gap of about 2 mm; higher voltages require better vacuums.
Effect of system pressure on the dielectric strength of air. Dielectric strength is given in volts/mil (1 mil = 0.0254 mm). Image source: http://www-rhvd.fnal.gov/meetings/monday_meetings/presentations/sergio1.html.
- To avoid collisions between electrons of the beam and stray molecules. These collisions can result in spreading or diffusing of the electron beam or, more seriously, can result in volatilization event if the molecule is organic in nature (for example, vacuum oil). Volatilizations can severely contaminate the microscope column, especially apertures, and degrade the image;
- To avoid damaging the filament. The volume around the electron gun must be kept free of gas molecules especially oxygen, which will greatly shorten the filament life.
- To prevent absorption of X-rays produced from the sample by air molecules. At high vacuum, even soft X-rays (such as B-Kα) are transmitted without loss
The gas in a vacuum system can be in a viscous state, in a molecular state or in a state that is intermediate between these two. Evacuation of the sample chamber from atmospheric pressure to operating vacuums proceeds from the viscous flow regime into the molecular flow regime. The type of flow regime reflects the mean free path of the the air molecules. The mean free path (MFP) for air at 25 °C can be approximated by:
where, P = pressure (torr) and lMFP = the mean free path length (cm).
Thus, MFP of gas molecules is very small at atmospheric pressure (6 x 10-6 cm) and the flow of the gas is limited by its viscosity. Under these conditions, gas molecules collide more readily with one another than with the chamber walls, and will move as a viscous mass in the general direction of low pressure. However, when the pressure drops to the point where the MFP exceeds the size of the chamber, the gas molecules do not collide with one another and no longer exhibit viscous flow. At this pressure, gas movement is independent of any pressure gradient and molecules may "backstream" from the pump into the sample chamber, moving against the pressure gradient. Backstreaming of pump oil vapor can be an important issue with diffusion pumps. Usually, water-cooled plates or baffles are positioned above the diffusion pump to promote condensation of pump oil before it enters the electron column.
Flow regimes. Note that the type of flow depends strongly on the diameter of the volume. Redrafted from image at: http://www.rdmag.com/images/0508/vactech_lrg.jpg
Electron microprobes and scanning electron microscopes have similar pumping systems. These usually include mechanical (rotary) pumps and oil diffusion pumps. When an ultra clean vacuum environment is required, a turbomolecular pump often replaces the oil diffusion pump. A mechanical pump is used to evacuate the sample chamber from atmospheric pressure to a moderate vacuum. Once the chamber pressure low enough (~10-3 torr), a gate valve is opened, and high vacuum achieved using an oil diffusion pump. Cameca’s SX-50 and SX-100 designs use an additional ion pump to help maintain vacuum around the electron gun.
Cameca microprobes have high- and low-vacuum sections. The wavelength-dispersive spectrometers are isolated from the electron column by thin windows made of polypropylene or Mylar. The spectrometers are evacuated by a separate mechanical pump and kept at a pressure of about 10-3 torr = 1 Pa, whereas the column is kept at a pressure of about 10-6 torr. Polypropylene windows are used on the light-element spectrometer(s) because they absorb fewer X-rays; however, they are more fragile than Mylar windows. This arrangement using windows allows a smaller volume (the electron column) to be kept at high vacuum. Additionally, they help keep the column clean by limiting out-gasses from gear and fitting oils and detector gas leaking from the spectrometers.
Schematic diagram of the MBX vacuum system. (1) Electron gun isolation valve; (2) Window separating spectrometers from column; (3) Sample change airlock; (4) Ballast tank; (5) Mechanical pumps; (D) Oil diffusion pump.
The pump-down sequences for a microprobe and scanning electron microscope are very similar. Initially, a mechanical pump use used to evacuate the column and sample chamber. Pressure is monitored using a thermocouple gauge. Thermocouple gauges work down to a pressure of about 10-3 torr. Once the vacuum is sufficiently good, the "gate" valve that isolates an oil diffusion pump from the column is opened. This allows the diffusion pump to evacuate the column further. There is a water-cooled baffle located above the pump to prevent oil vapor from entering the electron column. A cold-cathode (ion) vacuum gauge, which operates at pressures down to 10-6 torr, is used to monitor the pressure at higher vacuums. A cold-cathode gauge gets dirty with time and reads a better vacuum than actually exists.
Schematic diagram of the MBX vacuum system. During initial pumping (green and blue) the column is evacuated using mechanical pump 1. Once a vacuum of about 1 torr is achieved the gate valve opens to access the oil diffusion pump (green and red), isolating mechanical pump 1. The diffusion pump is kept at operating vacuum by mechanical pump 2. The ballast tank serves as a vacuum reservoir and allows the diffusion pump to function for short intervals with pumps 1 and 2 turned off. This minimizes mechanical vibration and permits high magnification operation.
In general, high vacuum pumps can be divided into capture and momentum transfer pumps. Capture pumps, such as ion-getter and cryo, operate by sequestering air molecules onto a surface where they are held it either temporarily (cryo) or permanently (ion-getter). Momentum transfer pumps, which include oil diffusion, and turbomolecular pumps, move gas by compressing it using some form of mechanical impact, and exhausting it at a higher pressure into another volume at lower pressure. These pumps require initial pumping (“roughing”) to reach operating vacuum and continued pumping to maintain a low exhaust pressure (“backing”).
Water vapor is a major problem in vacuum systems because it desorbs slowly from the internal surfaces of the sample chamber. It is the major residual gas after a chamber is pumped below 10-3 torr. Because water is released so slowly from surfaces, microprobes are constantly pumped. It may take many hours to reach operating pressures after an electron column has been vented to atmosphere because of the attendant admission of water vapor into the instrument.
There are two mechanical (or "sealed-oil") pumps on the Cameca microprobe and JEOL scanning electron microscope. One is dedicated to pumping on ("backing") the diffusion pump; the other serves to evacuate the column, sample chamber and spectrometers during initial pump-down. Mechanical pumps work at a rate of ~12 m2/h and can achieve a final vacuum of ~10-2 torr.
Schematic diagram of a mechanical rotary pump. After Potts 1987.
Mechanical pumps operate by rotation of an off-center cylinder. Longitudinal vanes move out of the cylinder and make contact with the sides of the pump housing. The vanes are either spring-loaded (E) or extended by the centripetal force of the spinning cylinder. Pumping proceeds as follows:
- The pump takes in air from the volume to be pumped through the inlet. Air molecules are sucked into the space between a pair of vanes (C and D) by the depressurization caused by rotation of the cylinder
- The isolated packet of air is compressed by the off-center rotation of the cylinder (E) until it has sufficiently high pressure high to force open the exhaust valve
- The air molecules exit through the oil to the outside
Schematic diagram showing the operation of a mechanical pump. Image source: http://whome.phys.au.dk/~philip/q1_05/surflec/node7.html.
The outlet valve (F) and moving parts of the pump are bathed in oil that serves to both lubricate the moving parts of the pump but also to trap air molecules. The process operates until there are too few air molecules to push open the exhaust valve. Gas molecules in the pressure range of a mechanical pump pressure range move via viscous flow.
The diffusion pump is used to produce and maintain low pressures in the electron column and the sample chamber (and in the Oxford Spectrometer on the SEM). An oil diffusion pump requires a initial vacuum of about 50 millitorr (5 to 10 Pa) to operate. Therefore, it is isolated from the column and sample chamber by a "gate" valve until the mechanical pump has produced a sufficiently good vacuum in the column as described above. Diffusion pumps are so named because rather than viscously "pull" air molecules out of a volume, they wait for the molecules to diffuse into the active part of the pump where they are then trapped and removed.
A diffusion pump operates by boiling oil at the base of the pump. Oil molecules escape to produce a supersonic vapor (400-400 m/s) that is directed into a funnel-shaped set of baffles (known as a “chimney stack”) and jetted towards the sides. Air molecules that have diffused into the vicinity are compressed against the jets and entrained. The upper part of the pump wall is cooled by circulating cold water through coils wrapped around it. When the oil vapor strikes the upper pump wall, it condenses into liquid and sinks to the bottom of the pump taking the adsorbed air molecules with it. At the bottom, the heater reboils the oil, releasing the air molecules, producing a build up of air molecules in the lower region. A port that is attached to a mechanical pump removes these air molecules. The diffusion pump produces a working pressure in the sample chamber of 10-5 to 10-6 torr.
Schematic diagram of an oil diffusion pump. Red arrows indicate the movement of boiling oil vapor (after Potts 1987).
On some microprobes the oil diffusion pump is replaced by a turbomolecular pump. Turbomolecular pumps use no oil and operate like jet engines with multiple, angled blades rotating at very high speed. Most turbomolecular pumps employ a series of rotor/stator pairs mounted in series. As gas molecules enter through the inlet, the rotor, consisting of a number of angled blades, strikes the molecules, imparting energy to them. Gas molecules are propelled into gas transfer holes in a plate below the blades called a “stator”. Gas captured by the first rotor is sent into the lower rotor/stator pairs and is successively compressed until it reaches a pressure where it can be removed by a mechanical backing pump. The blades as thin as possible and slightly bent for max compression. Turbo pumps can reach 10-7 to 10-10 torr. Unfortunately, because of the high rotation speeds (10-20,000 rpm), turbomolecular pumps have shorter life spans than oil diffusion or sealed-oil mechanical pumps.
Cross-section of a turbomolecular pump. Image source: http://whome.phys.au.dk/~philip/q1_05/surflec/node7.html
Sputter ion pumps are a type of capture pump, wherein air molecules are removed from the chamber by plating (gettering) them onto a surface. The gettering can be accomplished by making the pump wall very cold (cryopumps) or by ionizing gas within a magnetically confined cold cathode discharge (ion pumps). The ultimate pressure achieved by an ion pump is generally in the region of 2 x 10-11 Torr. Ion pumps require an initial starting pressure is 5 x 10-3 Torr or lower. Operating an ion pump at high pressure for extended periods shortens the pump life.
Ion pumps consist of short stainless steel cylinders (anodes) sandwiched between two metal (Ti, or Ti and Ta) plates (cathodes), all situated within a strong magnetic field aligned parallel to the cylinder axes. A high voltage is applied between the anodes and cathodes, and the electrons produced from the cathodes move in long helical trajectories through the anode tubes. The long electron paths increase the probability of collision with and ionization of gas molecules. In the main pumping mechanism, the ionized molecules are accelerated toward one of the cathodes where they are buried by Ti atoms. Ion impacts also sputter titanium from the cathode and the resulting Ti atoms acts as a getter for reactive gases (O, N, CO and H) and are deposited as stable oxides, carbides, nitrides and hydrides elsewhere in the pump.
Ion pump. (left) Diode ion pump, which has two cathodes. Image source: http://www.cae2k.com/howto.html. (right) Mechanism of operation of an ion pump. Image source: ??
No single gauge can measure pressure from atmosphere to high vacuum, so different gauges are used to monitor the level of vacuum present within the electron column and spectrometers at different pressures. Microprobes and scanning electron microscopes are usually fitted with thermocouple (1 to 10-3 torr range) and cold cathode gauges (10-3 to 10-10 torr).
A thermocouple (or T/C) gauge is used to monitor column pressure during initial pump-down (roughing) and the foreline pressure of the high vacuum pumps. It operates by measuring the thermal conductivity of a gas. The T/C tube contains a filament heated with a constant current and a thermocouple in contact with the filament. As the pressure decreases, the filament becomes hotter because the number of gas molecules hitting the wire and conducting heat away from the wire decreases. As temperature rises, the thermocouple voltage increases and is measured by a sensitive meter that has previously been calibrated against a manometer to determine the pressure. Each type of T/C tube has its own calibration curve.
Components of a thermocouple gauge. Image source: http://www.rdmag.com/images/0410/FEVac_a.jpg
A cold cathode (Penning) gauge is used to monitor column vacuum during operation. Cold cathode ionization vacuum gauges contain just two unheated electrodes, a cathode and an anode. A "cold" discharge is initiated by an external event (cosmic ray, radioactive decay) and maintained between the cathode and anode by a voltage of ~2 keV. A strong magnetic field makes the electron paths long enough so that they collide with enough gas molecules to maintain the discharge. The magnetic field is oriented so that the field lines cross the electric field lines, confining electrons to a helical path. The resulting current depends on pressure, which is indicated on the meter. The upper measuring range is limited by the glow discharge region (shown above). Unfortunately, the calibration of an ion gauge changes over time as the result of gradual contamination of the components.
Cross section of a Penning gauge.
Electron guns provide electrons for an electron beam by allowing them to escape from a cathode material. However, an electron must be supplied sufficient energy to kick it into a high energy state within the material and additional energy for it to escape the surface. Electrons with a component of velocity at right angles to the surface and kinetic energy at least equal to the work done in passing through the surface will be emitted. This total energy required to for a material to give up electrons is related to its work function, Ew. The work function of a material is given by:
where, E is the total amount of energy needed to remove an electron to infinity from the lowest free energy state, Ef is the highest free energy state of an electron in the material, and Ew is the work function or work required to achieve the difference.
There are three main types of electron sources used in SEMs and microprobes:
Thermionic sources, in which electrons are produced by heating a conductive material to the point where the outer orbital electrons gain sufficient energy to overcome the work function barrier and escape. There are two main types of thermionic sources: tungsten metal filaments and LaB6 crystals. These two types of sources require vacuums of ~10-5 and ~10-7 torr, respectively.
Thermionic Emission. Thermionic Emission occurs when enough heat is supplied to the emitter so that electrons can overcome the work-function energy barrier EW of the material and escape from the material. Electrons have a range of energies with the highest energy state called the Fermi Level, EF. The energy required to place an electron in the vacuum from the lowest energy state in the metal is E. When the emitter material is heated to a high temperature, a small fraction of the electrons at the Fermi level acquire enough energy to overcome EW and escape into the vacuum.
Field emission source, in which a large electrical field, 105 to 108 V/cm, is placed between cathode and anode. This field decreases the Ew of the cathode, a phenomenon called the "Schottky effect". At sufficiently high field strengths, Ew vanishes. The Schottky effect occurs even at room temperature and depends only very slightly on temperature, indicating that it is not a temperature activated process. Instead, it is a purely quantum mechanical effect called "tunneling". Field emission sources require vacuums of ~10-9 torr.
Field emission gun. Schematic diagram showing the components of a field emission source: emitter tip (T), first anode (FA) and second anode (SA). Voltages applied between the tip and first anode are ~ -3000 V; V0 is the accelerating voltage. Figure 3.23 from Heinrich, 1981, p. 38.
Thermal-field (TF) source, in which the tungsten point in a field emission source is heated, incorporating both thermionic and field emissions; this is also referred to as a "'Schottky cathode". A TF source requires a vacuum of ~10-8 torr.
Microprobes exclusively use thermionic electron guns. The electron source in an SEMs may use either type of thermionic cathode or utilize field emission. Field emission machines are markedly more expensive and are employed for ultrahigh resolution work. Most geological SEM applications, however, require relatively low magnifications and use thermionic guns.
The electron guns used in NAU's Cameca Camebax MBX microprobe and JEOL-6480LV scanning electron microscope employ tungsten filaments and have a triode (three part) configuration consisting of a cathode, Wehnelt cap and anode.
Electron gun. (left) Schematic diagram of the electron gun of the MBX microprobe. (right) Wehnelt assembly of the JEOL-6480LV SEM.
The tungsten filament is a thin (about 0.1 mm) wire bent into an "inverted V" to localize emission at the tip. This yields a coherent source of electrons emitted from a fairly small area; however, because the filament is bent in a single plane the geometry of the emission region is not perfectly circular. Tungsten is used because it withstands high temperatures without melting or evaporating. Unfortunately, the filament has a very high operating temperature (2700 K). Higher temperatures can deliver greater beam current, but the tradeoff is an exponentially decreasing lifetime due to thermal evaporation of the cathode material.
Configuration of self-biased electron gun. The distance between the Wehnelt and the filament can be adjusted in most SEMs, allowing the shape of the electrostatic field to be changed and optimization of the electron gun (after Goldstein et al. 1981).
Electrons leave a heated filament with an average energy of:
where k = Boltzmann's constant (8.617398 x 10-5 eV/K), and T = filament temperature (K). At 2700 K, emitted electrons have energies of about 0.23 eV.
Another type of thermionic source uses LaB6 for the cathode material. LaB6 has Ew of ~2.5 eV, yielding higher currents at lower cathode temperatures than tungsten. Typically, LaB6 cathodes exhibit 10 times the brightness and more than 10 times the service life of tungsten cathodes. Additionally, the smaller, circular, emission region improves the final resolution of the SEM. However, LaB6 is reactive at the high temperatures needed for electron emission.
Three designs of LaB6 cathodes. The Broers design uses a tungsten coil to heat the tip of the LaB6 rod. A heat sink at the base helps cool the rod and decrease its overall reactivity. The Vogel design heats the LaB6 rod by passing a current perpendicular to the length of the rod and uses graphite spacers between the rod and contacts to limit chemical reactivity between them. The Ferris design uses a short LaB6 rod is supported by a ribbon or strip through which an electrical current is passed for heating. The ribbon is made of a material that does not react with LaB6, such as graphite or tantalum. Image source an design explanations: http://www.sem.com/analytic/sem.htm.
Electrons are emitted in all directions from the entire heated filament, so a way is needed to localize emission at one spot. This is accomplished by surrounding the filament with a negatively biased Wehnelt cap. The Wehnelt is biased -200 to -300 V with respect to the filament, producing a repulsive electrostatic field that condenses the cloud of primary electrons produced from the filament. Emission from the filament is localized at the tip above an aperture in the Wehnelt.
Filament and Wehnelt for the JEOL-6480LV SEM.
It is important that the filament is properly centered in relation to the opening of the Wehnelt cap and be the proper distance from the opening. Otherwise, an off center beam that is either weak/condensed or bright/diffuse will be produced. The Wehnelt cap acts as a convergent electrostatic lens and serves to focus the cloud of electrons. The electrons converge at a point (10-100 μm in diameter) located between the base of the Wehnelt cap and the anode plate. This point is called the "cross-over" and is the location of the effective electron source. The distance between the tip of the filament and the Wehnelt aperture is critical in determining the geometry of the lens. Movement of the filament tip is the major source of beam instability and even a displacement of 1o will produce a significant change.
The potential difference between the filament and Wehnelt is maintained using a bias resistor, which allows the gun to be self-regulating. Recall from high-school physics that V = I R, where V = voltage, I = current, and R = resistance. As the filament emits electrons, an emission current (I) flows from filament to Wehnelt. Any increase the emission current causes a larger voltage drop (V) across the bias resistor and a larger negative voltage is applied to the Wehnelt, reducing the current. As the emission increases, so does the voltage difference between Wehnelt and filament, causing the emission to plateau. Proper bias voltage also optimizes the electron beam brightness (current density per solid unit angle) providing the most focused electron beam.
Schematic relationship between bias voltage, emission current, and beam brightness. Figure after Goldstein and Yakowitz, 1975, p. 25).
The electron flux from a tungsten filament is minimal until a temperature of approximately 2500 K. Above 2500 K, the relationship predicts that the electron flux will increase essentially exponentially with increasing temperature, until the filament melts at about 3100 K. However, in practice, the electron emission reaches a plateau termed saturation due to the self-biasing effects of the Wehnelt cap. Emission currents, those that flow between the filament and anode, generally are on the order of 200 pA (1 pA = 10-12 A), in SEMs. Proper saturation is achieved at the edge of the plateau; higher emission currents serve only to reduce filament life.
Sample current as a function of filament voltage in a self-biased gun. Emission and sample currents as a function of filament voltage in a self-biased gun (think of voltage as analogous to filament temperature). The operating voltage was 20 keV. Notice the false peak at about 3.4 volts caused by region of filament that reaches emission temperature before tip. After Figure 3.19 in Heinrich, 1981, p. 32.
The flux of electrons from the filament-Wehnelt configuration can be expressed by the ‘Richardson-Dushman’ equation (1923), which describes the current density emitted by a heated filament:
where A is a material constant (60 amp cm-2 K-2 for tungsten), and Ew is the thermionic work function (about 4.5 eV for tungsten).
The electrons emitted from the filament are drawn away from the cathode-Wehnelt assembly by the anode plate, which is a circular plate with a hole in its center. The anode is biased from +1 to +50 keV with respect to the filament (actually, the electron gun is held at a negative voltage relative to the anode, which is grounded). The voltage potential between the cathode and the anode plate accelerates the electrons down the column and is known as the accelerating voltage, which is usually given in terms of KeV and abbreviated E0. Together the Wehnelt cylinder and anode plate serve to condense and roughly focus the beam of primary electrons.
Observation of the absorbed current (that remaining in the sample during electron beam bombardment) allows alignment of the electron gun. The electron gun is aligned by shifting the position of the filament assembly relative to the anode and the column beneath it, to maximize the absorbed current. In the Camebax microprobe this is accomplished by moving the electron gun using two alignment knobs. Alignment may have to be periodically checked.
The JEOL-6480LV SEM uses an electromagnetic alignment coil located beneath the anode to align the gun. The coil is software controlled.
Configuration of the JEOL scanning electron microscope.
The electron beam is divergent after passing through the anode plate and must be refocused. The simplest type of electron lenses are electrostatic, which deflect beam electrons using electrically charged plates. While a charged particle is in an electric field, a force acts upon it. The faster the particle the smaller the accumulated impulse. Thus substantial lenses are requires to deflect a high-voltage electron beam. Additionally, electrostatic lenses require a very clean high vacuum environment to prevent arcing across the plates.
Cross-section Through a Cylinder Electrostatic Lens. A cylinder lens consists of several cylinders aligned along the electron optical axis; there are small gaps between the cylinders. When each cylinder has a different voltage, the gaps between the cylinders work as a lens. The magnification can be changed by using different voltage combinations. Image source: www.uga.edu/caur/EMOptics.ppt
At present, electrostatic lenses most commonly are used to deflect and focus ion beams in mass spectrometers. Microprobes and SEMs use magnetic lenses. Although electron lenses in principle behave the same as optical lenses, there are differences and the quality of electron lenses is not nearly as good as optical lenses in terms of aberrations.
The first magnetic electron lenses were developed by M. Knoll and E. Ruska in Germany in 1932. Their action is similar in principle to optical lenses, but electron lenses can be made only to converge, not diverge. Magnetic lens consist of two circularly symmetric iron pole-pieces with copper windings with a hole in center through which beam passes. The two pole pieces are separated by "air-gap" where focusing actually takes place. The magnetic flux diverges along the electron beam axis (figure below). Consequently, an off-axis electron is acted on by a magnetic force proportional to the cross product of the vectors v and B:
where, v = electron velocity and B = the magnetic field. This force causes the electron to move perpendicular to the axis of the lens. The resulting change in direction, in turn, yields a direction of force, deflecting the electron toward the the axis. The combination of forces cause the electron to spiral around the beam axis.
Electron path in Magnetic Lens. (left) Magnetic field and initial velocity of electron. (middle) Magnetic forces cause the electron to spiral around the lens axis (A) in increasingly small twists, focusing it. Note that the beam diverge past the focus point. (left) The result is similar to optical lens.
Electron lenses are not as good as optical lenses in terms of defects of focus, called aberrations. Aberrations are of two types. Spherical aberrations, in which the outer zones of a lens focus more strongly than inner zones, are most important in magnetic lenses. The result is that electrons along beam axis are deflected less than electrons passing through beam periphery, yielding more than one focal point. Chromatic aberrations, in which electrons of slightly different energies are focused differently, are relatively minor because the electron gun produces electrons with essentially uniform velocities. Spherical aberrations are minimized by placing a spray aperture in front of the magnetic lens, confining electrons to the center. This results in greatly reduced, but still acceptable, beam currents.
The condenser lens controls the amount of current that passes down the rest of the column. This is accomplished by focusing the electron beam to variable degrees onto a lower aperture. The sharper the focus, the less of the beam intercepted by the aperture and the higher the current. The Cameca MBX microprobe has a second condenser lens that is used to provide a better focus for SEM work, but during normal operation, this lower condenser lens is "decoupled" (not used).
Condenser Lens. Schematic of a condenser lens, showing upper spray aperture and lower limiting aperture.
Controlling Beam Current with the Condenser Lens. Change in probe current resulting from varying the condenser lens current. The points a, b and c correspond in both diagrams. The condenser lens should be operated to keep the focus point above the lower limiting aperture (currents along BC). Redrafted after Potts 1987.
Operation of the JEOL-6480LV condenser lens. The zoom condenser lens closely maintains beam focus at the sample, without needing operator intervention. The diagram does not depict the limiting aperture inside the condenser lens assembly.
The ultimate electron beam spot size depends on the beam current, which is controlled by the condenser lens, and the type of filament in use. A field emission gun provides the largest current with smallest beam diameter. However, these are expensive and tungsten filaments are most commonly used.
Beam Current and Diameter. The beam diameter is a function of beam current. Tungsten filaments yield the largest final diameters, reflecting the larger initial size of the beam produced from the filament. LaB6 and field emission guns yield smaller ultimate diameters. In essence, for all filaments beams with higher currents, having more electrons, spread due to repulsive electrostatic forces.
Microprobe electron beam currents are usually from about 10-9 A (10-9 A = 1 nanoamp = 1 nA) to 10-7 A (100 nA). SEM electron beam currents of are usually on the order of 100 pA (10-12 A = 1 picoamp = 1 pA). Microprobe beam currents are larger (resulting in a larger focus point) because there is no reason for beam diameter to be smaller than diameter of excited region (~2 µm). In contrast, SEM beam size is not limited by excitation area because SE are produced from just the upper ~100 pm of the sample. Scanning electron microscopes typically use beam diameters much less than 1 µm and use smaller final apertures to limit beam diameter. We tend to operate the JEOL-6480LV SEM as if it were a microprobe, using high sample currents.
The beam below the condenser lens is again divergent and must be refocused before striking the sample. Before this occurs, it is useful to be able to monitor and regulate the beam current. Emission from a tungsten filament can vary by up to 1% in 10 minutes; although this seems a small change, it will result in unacceptable analytical results. Thus, beam regulation is of critical importance in the electron microprobe. Such variations can be ignored in a scanning electron microscope, where the primary function is imaging; it takes no longer than about 2 minutes to acquire an image.
The beam regulator is located between the condenser and objective lenses in the electron microprobe column. The beam current may vary due to shifts in filament emission or changes in the current through the condenser lens. The regulator senses changes in the current and adjusts the condenser lens power accordingly.
Beam regulation. Schematic of the microprobe electron column showing the electron beam in red with its width greatly expanded. The effects of the condenser lens (1) and objective or probe-forming lens (PFL) are shown schematically by divergent black arrows; the sample is located at (6). The beam regulation system, located below the cross-over from the condenser has two apertures: a limiting aperture (2) and lower aperture with a smaller diameter hole (3). The amount of current impinging on the lower aperture (3) is amplified (5) and used to control the condenser lens power supply (4). Increased current through the condenser causes the beam to widen and more of it is blocked by the limiting aperture. The lower aperture senses the small beam current and decreases the condenser lens current, allowing more beam to pass through it. If the beam current decreases and narrows, less is blocked by the limiting aperture, more strikes the lower aperture. The current through the condenser lens is increased, decreasing the beam current.
The beam can be measured by interposing a sensor along its path. In the microprobe, the beam is measured before every analysis. X-ray counts are normalized to the measured value as an additional control on drift in the beam current.This procedure requires that there be a linear correlation between beam current and X-rays produced from the sample; however, in practice this does not seem to be the case for large changes in current. for example, the beam normalized count rates on a standard vary significantly between a 10 nA and 25 nA beam current. Obviously, one would not expect such large changes to occur as a result of drift!
Beam monitoring is software controlled in the microprobe. The user must push a button to activate the SEM beam monitor. A Faraday cup is used to block and measure the electron beam current. This device consists of a cup of electron absorbing material, such as carbon, capped by a 25 to 100 µm aperture. The Faraday cup is swung into the beam current, completely blocking it. The beam electrons enter the cup through the hole in the aperture and are trapped. Those backscattered hit the bottom of the aperture; similarly secondary and Auger electrons are unable to escape. The cup is grounded through a picoammeter, which measures the resulting current.
Faraday Cup. Schematic diagram of a Faraday cup with the electron beam shown in blue.
The Cameca MBX microprobe does not use a true Faraday Cup, although later models (SX-50, SX-100) do. Instead an aperture-capped cylinder of electron absorbing material, through which the beam always passes, is part of the electron column. The beam is measured by swinging a "trap door" to block the bottom of the cylinder, blocking the beam.
JEOL-6480LV SEM beam monitor. The user must press a button to move the Faraday cup in and out of the beam.
The objective, or "probe-forming", lens is located at the base of electron column just above sample. The beam is again divergent after passing through the apertures below the condenser lens and must be refocused. The objective lens focuses the electron beam onto the sample and controls final size and position. Hosted within it are scanning coils that allow the beam to be rastered across the sample surface, the astigmation correction coils, the beam shift coils, and the visible light microscope optics.
The Cameca MBX microprobe has a complex "minicoil" objective lens. It is an asymmetrical magnet with less of iron casing located below its air gap to minimize the magnetic field at the sample. It is designed to accommodate the light optical system and scanning coils inside it and has apertures that allow X-rays from the sample to pass to the spectrometers at a 40° take-off angle. The apertures have electron traps, which use negative voltages to prevent backscattered electrons from entering X-ray spectrometers.
MBX microprobe objective lens. The light optics (4) and scanning coils (1) are located inside the minicoil probe-forming lens (2), which has a pole piece (7) made of one solid piece of metal. X-rays (3) are collimated by small apertures (6) in the lens. Electron traps (5) prevent backscattered electrons from entering the X-ray spectrometers.
The objective lens in the JEOL-6480LV SEM is much simpler than that used in the MBX microprobe, being a solid conical lens with the scanning coils housed inside it. There is no need for the objective lens to accommodate a light microscope system, allowing it to have a more pointed configuration. Consequently, there is no need for apertures for X-ray spectrometers to achieve comparable take-off angles.
JEOL-6480LV SEM objective lens. Only the scanning coils are located inside the scanning electron microscope's probe-forming lens.
Scanning (or "deflector") coils raster the beam across the sample for textural imaging. Beam deflection is accomplished by coils located within the objective lens. The coils consist of four magnets oriented radially that produce fields perpendicular to optical axis of the electron beam. The beam is rastered by varying the current through these magnets. Deflecting the beam off-axis introduces aberrations that limit the maximum deflection that can be used and still keep the sample in focus. Additionally, deviations from linearity in X and Y increase as amount of deflection increases.
MBX Scanning Coils. The location of the scanning coils, beam centering coils and the stigmators is shown in this detailed schematic of the interior of the MBX microprobe's objective lens.
Magnetic lenses do not have perfect symmetry and the electron beam may be elliptical, something termed astigmatism (stigme means "mark" or "spot" in Latin). An astigmatic beam will collect signals from a elliptical region that will be rendered as a round spot on viewing screen or final image, producing a reduction in resolution. Astigmatism results from machining errors, asymmetry in magnetic windings, inhomogeneous magnetic fields in the iron, the buildup of contamination within electron column, and misalignment of the electron column. Only misalignment of the column is easily addressed because cleaning the column is a difficult and prolonged exercise.
Effects of astigmatism. If the effect of a magnetic lens is slightly tilted (or, equivalently, if the electron beam is tilted off axis), the focusing pattern becomes astigmatic. The beam first focuses along a horizontal segment (tangential focus), has a more or less circular shape at the focal plane, and focuses again along a vertical segment (sagittal focus) behind the focal plane. Image source: http://www.optonlaser.com/pages_communes/glossaire/GlossaireLaser.htm.
Both microprobes and scanning electron microscopes have "stigmators" housed with the scanning coils inside objective lens that compensate for imperfections in construction and those not corrected by alignment of column. These consist of eight weak electromagnetic lenses with alternating coils connected in series but wound in opposite directions to achieve different polarities. Stigmators apply a correcting magnetic field to produce symmetrical electron beam at the sample. The operator can change both the strength and orientation (angle) of the magnetic field produced by stigmators to control the final beam shape.
Stigmators. Schematic top view of the arrangement and polarities of the stigmator coils.
The alignment of the microprobe column can be tested by changing the objective lens focus and examining the over- and under-focus relationships of the electron beam on a fluorescent sample. If the limiting aperture is not centered along the axis of the electron column, the image of the aperture hole will rotate as the setting of the objective lens is changed due to the helicoidal movement of electrons in the beam. The aperture should be centered using the limiting aperture knobs to eliminate this movement.
Alignment of the limiting aperture. A. Perfect alignment: the beam changes diameter symmetrically when changing objective lens focus. B. Poor alignment: the beam sweeps out an arc as the focus is changed; r1 is the distance from the focused beam to the electron column axis. C. Fair alignment: this is the most common case, where alignment is slightly off. Note that rather than sweeping, the beam seems to expand and contract from one side as focus is changed; r2 is the distance from the focused beam to the electron column axis. After Figure 3.25 from Heinrich, 1981, p. 42.
Geological microprobe samples are usually petrographic thin sections that have been ground flat and polished. In contrast, SEM samples are far more variable ranging from thin sections to actual large pieces of rock. The sample chambers of the two instruments reflect this difference. The Camebax microprobe has a sample chamber just large enough to permit movement of two entire thin sections under the electron beam (one usually reserved for standards materials). In contrast, the JEOL-6480LV scanning electron microscope has a sample chamber that permits large and complex movements of the sample stage; for example, the SEM can accommodate samples up to 200 mm in diameter and ~80 mm high.
JEOL-6480LV SEM sample chamber. The silver stage motor in the front is ~11 cm long.
The JEOL SEM chas the added capability to also work with a relatively poor vacuum in the specimen chamber. Air (or another gas) is admitted into the sample chamber to neutralize the build-up of the negative electrostatic charge on the sample. In "low-vacuum (LV) mode," the sample chamber pressure, typically 20-30 Pa, is maintained at a specifiable value by a separate extra pump with a large foreline trap. The electron column is separated from the chamber by a vacuum orifice (aperture) that permits the column and objective lens apertures to be maintained at a higher pressure. Unfortunately, secondary electrons are absorbed by the air in the sample chamber. Instead, the backscattered detectors are used to image the sample. LV mode allows the study of samples with poor or no electrical conductivity. Normally, these materials would be either coated with an electrically conductive coating, which takes time and hides the true surface, or studied at low voltages, which do not give backscatter or X-ray information (E0 < Ec for most elements of interest). Additionally, samples containing volatile substances (water) can often be studied directly.
An environmental SEM utilizes relatively low vacuum pressures (up to 50 torr ~ 6700 Pa) not only to neutralize charges, but to provide signal amplification. The electron column is gradually brought up to pressure by a series of pumps and apertures.
ESEM schematic. Image modified to emphasize variable pumping. Image source: aic.stanford.edu/jaic/articles/jaic33-02-008.html.
A positively charged detector electrode is placed at the base of the objective lens at the top of the sample chamber. Secondary electron emitted from the sample are attracted to the positive electrode, undergoing acceleration. As they travel through the gaseous environment, collisions occur between electrons and a gas particles, producing more electrons and ionization of the gas molecules and effectively amplifying the original secondary electron signal (electron cascade). This form of signal amplification is identical to that which occurs in a gas-flow X-ray detector. Positively charged gas ions are attracted to the negatively biased specimen and offset charging effects. The vapor pressures in environmental SEMs can high enough that liquids may be observed!
Secondary electron detection in an environmental SEM (ESEM). Image source: http://www.calce.umd.edu/general/Facilities/ESEM.pdf.
Dissolution and precipitation of halite observed in an ESEM. A series of electron micrographs showing the dissolution (a–c) and reprecipitation (d–f) of NaCl as water is condensed on the surface of the salt crystals and then evaporated. White bar at the lower center part of each micrograph = 20 μm. Image source: http://aic.stanford.edu/jaic/articles/jaic33-02-008.html.
The sample chamber for a microprobe does not need to be large since all samples are flat. The only requirement is the the sample chamber have sufficient room for the motors and gearing that allow movement of the sample under the electron beam. Most microprobes can accommodate standard polished thin sections or 1" polished round sections and most permit examination of the sample with either reflected or transmitted light (see below). The number of sections and the maximum amount of movement depend upon the individual instrument.
The X, Y, and Z positioning of samples is accomplished by moving the stage with stepper motors. The MBX microprobe has a transmitted light stage installed that can be moved 20 mm in the X-direction (left/right) and 50 mm in the Y-direction (in/out). Movement in the Z-direction is limited to avoid striking the optical system. Each stepper-motor step represents 1 µm of stage travel. Consequently, stage motion is reproducible to about 1 µm. The worm gears on the stage drives and spectrometers (see below) are subject to mechanical "slop", even with very high tolerance machining. As a result, peak positions appear to shift when approached from different directions (up or down the drive screw) and stage positions are not reproducible. The MBX microprobe analytical programs use a "backlash" to compensate for this slop in which all peaks and stage positions bypassed by a certain number of steps and always approached from the same direction. The newest Cameca microprobes (SX-50, SX-100) use optically encoded light bars to control stage (and spectrometer) movement. The direction of movement does not matter, since the positions are located using absolution positions on the encoders. The stage and spectrometers are moved using DC motors rather than AC stepper motors, and movement is very rapid and precise.
MBX Sample stage. (left) The drive motors are located on the left of the stage assemble. Note the forked sample holder that can accommodate two samples simultaneously. The technician is holding the transmitted light assembly prior to installation. (right) Sample hold with a sample installed in right-hand fork. The transmitted light assembly inside the column is positioned as indicated.
In microprobes, sample changes are accomplished using an airlock to avoid having to vent the entire column. The airlock cover plate is removed and a sample change unit attached. This changer is evacuated to a pressure of about 2 Torr, whereupon the airlock door can be opened and the sample inserted. In the MBX microprobe the sample slides into a narrow forked holder that allows a transmitted light source to be positioned beneath them.
The MBX microprobe is equipped with a light microscope that allows simultaneous examination of electron beam and specimen during operation. Beam position, size and shape are observed using a fluorescent sample such as willemite, or benitoite. This allows the operator to locate the beam position prior to analysis of minerals that do not fluoresce. The microscope visible light optics are housed within the objective lens. The optical system consists of a series of mirrors with holes in them through which electron beam passes. The resolution of the microscope is 0.55 µm (the smallest tick marks on optical reticule are 1 µm apart); the field of view is 500 µm wide. The light optics have a small depth of focus (~0.9 µm), insuring good focus for analysis.
Light optics system. The sample is observed (and illuminated) by a series of mirrors, the latter three of which are coaxial with the electron beam. The illumination/reflection path of the bottom two mirrors is shown schematically by a dashed line.
Samples can be observed in both reflected and transmitted light. The transmitted light assembly consists of a right angle mirror positioned beneath the optical microscope and a horizontally oriented light source located outside of the evacuated electron column. There is only one magnification (400x) available for observation on the MBX probe; this magnification is higher than that used during standard petrographic examination, making it sometimes difficult to locate points of interest for analysis. Other microprobes have additional magnifications available. Regardless, all samples should be thoroughly characterized petrographically and grains of interest be carefully located before microprobe analysis. One popular method is to use thin lines of India ink leading to or circling grains of interest. Additionally, photographs taken in transmitted or (better) reflected light serve a useful maps of a sample.