Scanning Electron Microscopy

Overview

Scanning Electron Microscopy (SEM) utilizes a high energy beam of focused electrons to generate images or micrographs of the topographic detail and/or material variation in small specimens. SEMs are capable of few-nanometer resolution with a well aligned beam and appropriate sample preparation and operating conditions. Images are formed by scanning an electron beam point-by-point across a specimen surface, which generates various signals that are collected, converted to an intensity grayscale, and correlated to a location for display on the monitor.

Equipment

• JEOL 7500F HRSEM | 7500F Reference Guide Image with secondary electrons, backscattered electrons, and transmission electrons at high vacuum and high resolution, even at very low voltages. Elemental analysis with x-rays available with EDS by EDAX.
• TFS Quanta 600 FEG ESEM | Quanta Reference Guide Image non-conductive or vacuum-sensitive specimens in high vacuum or in a low vacuum water vapor environment. EDAX EDS elemental analysis and EBSD crystallography available. This microscope is equipped with a large chamber for in-situ imaging and high- and low-temperature stages.

Techniques

• EDS - Energy Dispersive X-Ray Spectroscopy: A technique to chemically characterize specimens with overview spectra or spectrum maps that can yield an approximate composition of specimens on any of our SEMs.
• EBSD - Electron Backscatter Diffraction: A technique to characterize the crystallographic structure of specimens, preferably specimens that have been polished, to show details such as grain size and orientation. Available on the Quanta.
• Low Voltage SEM - One method to image fine or thin surface details, including on specimens of low to no conductivity. The 7500F is equipped with a gentle beam mode that enables ultra low landing voltages through a stage biasing system.
• Low Vacuum SEM - A charge compensation method for imaging nonconductive specimens without applying a conductive coating. The Quanta is designed to operate under low vacuum water vapor environment up to 1.5 torr, and an environmental mode to higher pressures and higher humidity. Humidity at standard low vacuum settings is significantly lower than room humidity and will not adversely affect specimens.

Sample Preparation

Contact staff for specific sample preparation advice. We are not a specimen preparation facility, however we can advise on what makes a good specimen and how to mount specimens to stubs for optimal imaging.

General Considerations

• Specimens (and anything going inside the chamber) should always be handled with gloves.
• Know the sample size limits for the SEM: The Quanta has a large chamber that can accommodate large specimens up to 1kg. The 7500F has a more restrictive chamber: specimens should be no taller than the specimen holder being used, however it can accommodate 4 inch wafers.
• Specimens should be well secured with no loose particles. Contact staff for assistance preparing powder specimens to help prevent contamination inside the chamber.
• Specimens should be clean and vacuum compatible. Outgassing and contamination can both deteriorate the quality of the vacuum and thus image quality.

Conductivity

• With the exception of the Low Vacuum Mode available on the Quanta, specimens should be conductive for the best image quality. This means creating a conductive pathway from the specimen surface to the stub and stage so that excess electrons can be drawn away from the specimen surface.
• Know the approximate conductivity of your sample: even semiconductors have enough conductivity for good images, but insulators are more difficult and may require additional preparation, such as sputter coating.
• Know the approximate conductivity of your substrates: conductive substrates contribute to the conductive pathway. For instance, Si wafer pieces are easier to image than glass or sapphire.
• All provided materials for sample preparation are conductive, including:
  • Double sided carbon tape: Useful for adhering most specimens to stubs or holders, this double sided adhesive is also conductive. Use a small amount for most specimens.
  • Single sided copper tape: Useful for adding a conductive pathway to e.g. nonconductive films on conductive substrates. Place copper tape so it is in contact with the specimen of interest and the stub or holder. Ensure that the copper tape is pressed firmly onto the specimen.
  • Pin stubs: Stubs are small pieces of aluminum for mounting common specimens. Pin stubs have a pin on the end for secure loading onto specimen holders or stages. Pin stubs come in a variety of forms, including common flat stubs, large area stubs, cross section, intermediate angles, and more.
  • Specimen holders: Hold either pin stubs or specimens directly. Some holders also have pins for easy loading. Others have large areas and unique geometry or mechanics to hold large or odd shaped specimens.

Cross Section

• Clean, flat edges image best. Cleaved wafers typically have good edges. Ask for cleaving help if you need it.
• Position cross section specimens as close to the flat surface of a cross section stub as possible for best results.
• A thin strip of copper tape across the surface of a cross sectioned specimen may help improve conductivity.

Theory (under construction)

Beam Generation

Electron microscopy begins at the top of the column where a source is energized with specific settings, some customizable by the user, to extract a beam of electrons which are then focused through a series of electromagnetic lenses before being emitted from the end of the pole piece to interact with the specimen, where it then generates other signals to be collected by detectors.

Accelerating Voltage

The accelerating voltage, sometimes called the landing voltage, high voltage, or high tension (although these terms may have slightly different meanings,) defines the energy supplied to the electron beam at the source. This voltage is the voltage difference that extracts and accelerates electrons away from the source and is typically measured in kilovolts (kV,) but sometimes displayed as keV (kiloelectronvolts, aka a kilovolt applied to an electron.) The accelerating voltage determines what is visible in images, as higher voltages have larger penetration depths, lower voltages reveal more surface detail (although too low of a voltage can reveal irrelevant surface contamination.) Techniques such as EDS tend to require higher voltages to generate the appropriate signals. Too high of a voltage may introduce charging in less conductive specimens, or may hide important thin surface details due to a larger penetration depth and interaction volume within the specimen. Theoretically, higher voltages yield sharper probes, and therefore higher resolution.

Emission Current

The emission current determines how many electrons are extracted from the source at a time. This is measured in microamps (μA.) Typically, the best emission current for imaging is the highest available current to yield the greatest signal from the specimen. However, some cases (especially for nonconductive specimens) may warrant a lower emission current.

Image Formation

Understanding how images are formed helps us understand how to visualize the important aspects of a specimen, interpret images, and know what to do when artifacts or other problems interfere with good image formation. In short:

• the beam rasters, or scans, across a specimen surface where
• at each pixel it interacts with a certain volume of the specimen and then
• generates various signals that can show topographic detail, material contrast, or specific elements.

Beam-Specimen Interactions

When high energy electrons hit a surface, most of the electrons penetrate some distance into the specimen and scatter. Various scattering events yield different types of signal, all within some volume known as the interaction volume. The size of this interaction volume is determined by the beam energy (accelerating voltage) and the specimen material. Higher incident beam energies yield larger interaction volumes. Generally, non-conductive materials have larger interaction volumes than conductive materials for a given energy, so images of conductive specimens may appear sharper than images under the same conditions of non-conductive materials.

In SEM, three key signals are generated:

• Secondary electrons are produced from the smallest interaction volume closest to the specimen surface, usually a few to tens of nanometers. These electrons are the result of multiple inelastic scattering events and so have less energy to escape the specimen surface from deeper within the specimen - hence the small interaction volume. Secondary electrons show topographic detail, where the amount of surface exposed at a given point largely determines how many electrons can reach the detector. This means that edges tend to appear bright, while holes or pits appear dark.
• Backscattered electrons come from a single elastic scattering event and therefore have higher energy than secondary electrons. Backscattered electrons "bounce" off atoms in a specimen, and tend to scatter more from larger or heavier elements. These conditions combined mean that backscattered electrons come from a larger interaction volume than secondary electrons - typically on the scale of a hundred nanometers - and that they carry material contrast, so brighter regions show the presence of heavier elements.
• Characteristic x-rays are produced when an incident beam electron displaces an atomic electron and a higher energy atomic electron falls to the lower energy vacancy. The energy difference between these levels is emitted as a photon with a characteristic wavelength or energy. These energies are known for all atoms and dedicated software is used to measure these signals for semi-quantitative chemical characterization. Since characteristic x-rays are a release of photons rather than electrons, they can be generated and detected from approximately a micron of interaction volume. More information on this form of characterization is found on the EDS page.

Scanning

The electron beam scans, or rasters, across a specimen surface point by point, and the signal generated from each point is displayed as a greyscale intensity on a coordinating pixel. This greyscale intensity is based on the number of electrons (intensity) detected. Scattered electrons are able to escape specimen surfaces more efficiently at edges or peaks, which results in higher intensity in what is known as the edge effect. If a specimen is not conductive or does not have a conductive pathway to ground, electrons may accumulate at certain locations - especially edges - if the beam scans too slowly. This effect is known as charging and distorts images.

Random noise is also collected by the detector, resulting in grainy images. Random noise is by nature random, so increasing the amount of signal generated (and therefore collected) improves this signal to noise ratio (SNR,) thus yielding smoother images and improving the ability to resolve fine features. Increasing the SNR can be accomplished by:

• Increasing the dwell time, also known as the scan speed: A longer dwell time means more electrons interact with the specimen to produce more signal, while the noise contribution remains constant. Longer dwell times are typical for capturing images and fine alignments, while short dwell times are great for coarse focus or changing the magnification and stage position.
• Increasing the probe current, also known as the spot size: A larger probe current means there is more current at the specimen surface, and therefore more electrons. This method increases the rate of electron interactions, whereas changing the dwell time keeps the rate constant. Increasing the probe current may result in blurry images at high magnifications due to an increase in the size of the beam spot at the specimen surface.
• Using frame integration or averaging: Averaging or integrating the signal from multiple frames is similar to increasing the dwell time, but reduces the likelihood of charging since there is time for charge to dissipate between scans. Using more frames is especially effective when the individual frame time is fast and many frames are integrated. In general, to achieve the SNR of a comparable single frame at a slow speed, divide the target dwell time by the dwell time of an individual (stable) frame. Some SEMs have a drift compensation mode available, which compares the position of bright and dark pixels from one frame to the next. Drift compensation is not effective for a single frame, nor for multiple slow frames when the drift occurs within a frame. Use of drift compensation may reduce the appearance of drifting features when an individual frame contains no obvious drift, but the specimen appears to drift from one frame to the next.

Signal Collection

Various detectors attract, collect, and process generated signals to convert them into an image. Depending on the type of signal desired, different detectors may be used:

• Secondary electrons are relatively low energy, so a bias is applied to the detector to attract these electrons. It is common for secondary electron detectors to also collect backscattered electrons, however, some detectors have tunable voltage biases to reduce this contribution, if desired.
• Backscattered electrons are relatively high energy and may be collected by reducing the bias of a standard secondary electron detector. Some systems have dedicated backscattered electron detectors which can be mounted directly under the pole piece to take advantage of the difference in takeoff angle between secondary and backscattered electrons, and to reduce the appearance of shadows. Some of these detectors have multiple segments that can be tuned to show shadows or behave similar to a secondary electron detector.
• Characteristic x-rays are very high energy and are collected by a dedicated x-ray detector (an EDS detector) that is often offset to the side and calibrated for a specific working distance. These detectors may need to be installed for each use.

With some detectors, images appear to have shadows. This is most common for detectors that are offset from the beam. Shadows appear in images when signal generated from one part of a specimen is blocked by another, taller part of the specimen. If shadows are obscuring an important portion of a specimen, you may be able to use a different detector if the system has multiple, or rotate the stage. Be aware of the difference between stage and image (or scan) rotation. Stage rotation will change the orientation of the specimen in the chamber, and therefore the position of shadows, while image rotation changes the way the beam scans across the specimen, which does not affect shadows.

Each detector may have a different optimal working distance. Ask staff for assistance or check documentation if you are unsure of the optimal working distance for a given microscope and detector. In some cases, shortening the working distance (carefully) may yield greater signal collection. This is especially true if the detector is mounted above the specimen. However, offset detectors yield diminishing returns when the working distance is too short since the signal is less able to reach the detector.

Charging

Charging appears in different forms depending on the specimen and the type of charging. In general, charging refers to a buildup or accumulation of electrons on a specimen surface. This buildup tends to occur in nonconductive specimens or specimens without a good conductive pathway connecting the surface to ground (via the stub and stage.) Distortions from charging tend to be more apparent at slower scan speeds since the beam dwells longer on a given point. The accumulation of charged particles influences and slows the incident electron beam, which leads to a higher secondary electron emission. This emission may clear the charge buildup, but it may return on the next scan. Charging may appear as: (make this a table with images)

• Drift: The sample appears to move continuously in one direction when the stage is stationary. Vibrations are not typically due to charging but rather due to an unsecured specimen.
• High contrast: For example, a very bright feature surrounded by an annular dark ring that does not seem to respond to changing the image contrast.
• Pulsing or fluctuating brightness: The same features may appear dark in one scan and bright in the next, or may actively appear to fluctuate if using a fast scanning speed.
• Streaks: Bright (or dark) streaks may appear across an image from a point of charging along the row of pixels. This is distinct from streaks seen when a beam is astigmatic since the direction of the streak always follows the scan direction along a row of pixels.
• Bloom: growing and changing "features" on the outer edge of cross section specimens.

Charging is influenced by the electron dose (beam energy and current) and sample conductivity. Sample conductivity refers not only to the conductivity of the material itself, but also the conductive pathway connecting the specimen to ground and the morphology of the specimen (e.g. clumps of powders or bundles of fibers tend to have poor conductivity if they are too thick.)

In most cases, decreasing the probe current and/or accelerating voltage reduces charging. Some samples will charge regardless of these conditions and must be imaged at fast speeds with frame averaging or integration for noise reduction. In rare cases, a higher accelerating voltage will reduce charging. One such case is for very thin films.

There are some specimen preparation techniques that can help reduce charging, including:

• Sputter coating: the application of a few nanometer layer of a metal to the surface of a specimen
• Applying copper tape: this is useful for conductive films on nonconductive substrates. The copper tape acts as an adhesive wire to give the electrons an escape route that bypasses the nonconductive substrate.
• Ensuring powders and fibers are in a single layer, or as thin a layer as possible to reduce the likelihood of large (to an SEM) clumps or bundles that inhibit the flow of electrons.
• Using conductive substrates whenever possible: glass slides are notoriously difficult to image in an SEM.

If specimen preparation cannot be suitably modified, the specimen may be imaged in a low vacuum SEM. Low vacuum imaging bleeds a small amount of gas or water vapor into the imaging chamber. In the presence of the high energy incident electron beam, this gas or vapor will ionize and be drawn to charged areas of a specimen where the ions remove electron buildup.

Some SEMs have multiple available detectors, some of which are more suited for imaging samples prone to charging. In rare cases where secondary electron imaging is impossible due to charging, imaging with backscattered electrons or altering the bias on the detector may reduce the appearance of charging.

Good Images

Good images in microscopy are self-evident. A viewer should be able to understand and clearly see the intended features of an image, and the caption may provide additional context. Achieving good images may mean adjusting settings and ensuring proper alignment as much as achieving reasonable contrast and providing visual context.

• A well aligned beam is essential. The alignment notes on this page and operating guides provide the tools and theory behind alignment, but practice yields the difference between good alignment and great alignment. An image that is unfocused and astigmatic lacks key details and may distort features in a misleading way.
• Selecting the appropriate accelerating voltage and currents can likewise reveal or conceal features. Low voltages show surface detail, but too low of a voltage may show undesired surface contamination. High currents may improve the signal to noise ratio, but will blur features at high magnifications. It is best to try a few different settings to understand which combination yields the best view of desired features, while still considering scientific integrity.
• Image contrast enables the viewer to see differences in topography and material. Both low and high contrast obscure features. A good greyscale image avoids pure black or pure white pixels as much as possible while using as much of the available greyscale spectrum. Images may appear noisy with good contrast, which may be compensated for by scanning for a longer time to collect more signal. When in doubt, most systems allow the user to see the greyscale histogram and ensure that the contrast spread is reasonable.
  • As a slightly more advanced method, capturing images in 16-bit format (available on most SEMs in the NCF) with somewhat low contrast is acceptable since the 16-bit format allows for more postprocessing.
• Capturing images at appropriate magnifications facilitates comprehension. Images should include visual context such as nearby features or devices. Serial magnifications should be in reasonable increments to avoid confusion. In general, a 2-5X increase in magnification maintains enough visual context for understanding.
  • When capturing serial magnifications, it is often beneficial to start at the highest magnification then zoom out in steps so the key field of view remains centered. However, for specimens prone to charging or beam damage, it may be better to start at the lowest magnification then zoom in to avoid the presence of raster boxes within lower magnification images.

Alignment

There are three key steps to aligning an SEM: Focus, aperture alignment, and stigmator alignment. After performing initial alignment at the beginning of a session at a sufficiently high magnification, only small changes may need to be performed when moving between specimens at different heights. The beam will likely need to be realigned if beam settings are changed, including the accelerating voltage and probe current.

Focus

This is the first step in aligning an electron beam. Focusing means moving the smallest cross section of the electron probe to the specimen surface. The plane normal to the beam containing this cross section is called the focal plane. The position of the focal plane is defined as the Working Distance when measured from the pole piece (where the beam is emitted from the column) down. When a specimen is in focus, the working distance describes where the specimen surface is. In some systems the working distance can be linked to the stage height. In general, the position of the specimen according to any stage height indicators or displays is considered an estimate, and the real position of the specimen is defined by the working distance when the specimen is in focus. When an image is out of focus, it appears blurry because the probe cross section is larger than the features being imaged, or because the probe is larger than the size of the image pixels. If an image is still blurry when in focus, or if it is difficult to focus because no one working distance appears sharp, then the beam is likely astigmatic and the stigmators need to be tuned to shape the beam. On highly topographic specimens, a portion of the image may appear blurry while other portions are in focus. This is due to the limited depth of field. Some SEMs have a control that enables the user to tune the depth of field for textured samples or respond to high degrees of tilt.

Apertures

Of the initial beam generated from the source, only a small portion of it is selected for final imaging. This portion of the beam is selected with an aperture, which may need to be positioned in the beam path so it is centered. If this aperture is not centered, then the beam is not emerging from the pole piece straight (which results in other alignment distortions, especially when imaging highly topographic specimens or using long working distances,) and the beam may not be as bright as possible. To ensure the brightest, most central portion on-axis portion of the beam is selected, a wobbler is used to shift the position of the focal plane up and down automatically, enabling the user to observe whether the image shifts - which indicates that the aperture is not centered. This function may be called the wobbler, lens align, or aperture align, and typically has X and Y axis controls to move the aperture inside the SEM. At low magnifications (typically below 10kX) the image may appear to rotate when the wobbler is engaged. This is normal. Zoom in to a higher magnification than is desired for imaging for this step.

Stigmators

An ideal electron beam is circular in cross section regardless of focus. If the cross section is not circular when out of focus, the beam is considered astigmatic and will be blurry when in focus. This is because the electrons are not reaching a proper point at the focal plane so the probe is larger than it should be. The stigmators knobs on an SEM enable the user to control and reshape the beam with a set of magnetic fields. These are labeled X and Y, but do not correspond to the X and Y directions, but instead control (typically) two sets of perpendicular axes, offset by about 45 degrees to one another.

Astigmatic beams cause the image to stretch when out of focus, and this stretching effect is directionally perpendicular to itself on either side of focus. This means that astigmatism can be identified by moving the focal plane to either side of focus and observing whether the image appears distorted in opposite directions. If so, then the astigmatism can be corrected by first going to focus (directly between these two stretching directions) then adjusting one stigmator at a time until the image appears sharpest. This can also be corrected when out of focus by observing how the stretching changes. Ask staff for an explanation.

Troubleshooting

Achieving a sharp image in SEM takes practice, and is not always due to the focus. It is important to ask the right questions when troubleshooting to find the best solution to improve image quality:

Are my beam settings appropriate for this sample?

Every sample interacts with the electron beam differently. Consider:

• What part of the sample do I want to see?
• Is the sample charging?

You may need to change the beam voltage or probe current/spot size. There will always be a balance between beam energy, specimen damage, charging, and image quality. Try changing one setting at a time in either direction (higher/lower, larger/smaller) and observe how your sample responds.

Have I aligned the beam?

There are three key steps to aligning an SEM:

• Focus: The focal plane should be on the sample surface.
• Aperture/Lens alignment: The beam should straight from the pole piece, and so that the brightest, most central part of the beam is selected by the aperture.
• Stigmators: The cross section of the beam both in focus and out of focus should be circular, yielding the smallest probe diameter at focus.

If you are having trouble focusing, the number one reason is that the stigmators are not aligned. Try aligning them at a higher magnification and vary the focus (in and out) until the image is sharpest at focus and does not stretch or smear directionally on either side of focus.

Is it how I'm collecting signal?

SEMs collect signal with a single-point detector, not an array like the detector in a camera. Because of this, we must consider:

• Beam-Specimen Interactions, or how the beam generates signal.
• How signal reaches the detector, or the position of the detector relative to the specimen.
• How images are formed, including scanning rates and image size or pixel resolution.

In some cases, slowing the scan speed or integrating more frames of a faster speed can improve the signal-to-noise ratio (SNR) in response to different types of charging. In other, the sample may need to be moved to a shorter working distance to improve contrast. Be careful if moving the stage closer to the pole piece: use the chamberscope, be aware of stage limits, and ask for help.

Do I need to rethink my sample preparation?

Sample preparation can help or impede signal generation. Generally speaking, well prepared samples produce better images. Consider:

• How the sample itself is prepared.
• How the sample is secured to the stub.

With some exceptions for low-vacuum SEM, ensure there is a conductive pathway between the specimen surface and the stub. This may mean coating, adding copper tape or a conductive clip, pressing specimens better to carbon tape, using a conductive substrate, or ensuring powders or fibers are adhered in as thin a layer as possible. Remember: what may seem small to us, may still seem large to the SEM. Failure to secure or ground specimens well may result in charging or drift.

Remember that SEM is a skill and every sample is different - take time to practice and observe how images change in response to beam settings and alignments.