Chemistry 250 Lab: Scanning Electron Microscopy (SEM)

Objectives: (1) to become familiar with the basic operating principles of the Hitachi S3500N
SEM, (2) to become familiar with the different signals of interest in the SEM and their utility (3)
and to learn techniques for improving imaging resolution and the compromises you make in
doing so.

Background

Scanning Electron Microscopy

Scanning Electron Microscopy (SEM) is one of the most versatile methods for analysis of solid
materials. Since its inception in the middle of the 20th Century, the technique has evolved to
allow for imaging of nanoscale features and integration with electron probe microanalysis
(EPMA) techniques. In SEM, an electron beam is moved in a raster pattern across the surface of
a sample. The beam interacts with the sample surface, producing a number of different signals,
which can be analyzed to provide useful information about topography, composition,
crystallography, etc. In many respects, the optical system in an SEM is analogous to that in an
optical microscope, however, instead of glass lenses used to focus light, electromagnetic lenses
are used to focus and deflect the electron beam. The primary motivation for using SEM instead
of light microscopy is related to the fact that electrons have a much shorter wavelength than light
(higher resolution) and SEM uses a longer focal length (greater depth of focus).

The main classification for different types of SEM instruments is related to the source of electron
illumination. The most common electron guns employ a tungsten filament (e.g. S3500N SEM in
EPIC), which is resistively heated to nearly 3000K until electrons have sufficient energy to
overcome the work-function (Ew) energy barrier (thermionic emission). By choosing materials
with a lower Ew than tungsten, such as LaB6 (e.g. S570 SEM in EPIC), the brightness of the
electron source can be increased by an order of magnitude. Field emission electron guns
represent another class of electron sources and are generally comprised of sharpened tip of single
crystal tungsten. Field emission guns have two orders of magnitude greater brightness than is
possible with LaB6 thermionic emission, exhibit a longer lifetime, have a smaller virtual source
and a lower energy spread. There are three main types of field emission electron guns: cold field
emission (e.g S4500 SEM in EPIC), thermal field emission and schottky emission (e.g. LEO
SEM in EPIC). While the enhanced resolution of field emission gun equipped SEM instruments
is useful, even the thermionic gun equipped instrument used for this lab has a resolution better
than 5nm, however there are several tradeoffs involved in achieving this performance.

Laboratory procedures

Part 1: General Operation of the SEM

Before we can do any imaging or analysis with the SEM, we first need to prepare a sample
appropriate for the instrument. Fortunately, sample preparation for the SEM is generally quite
straightforward and doesn’t usually require any rigorous and time-consuming processing. Some
general sample requirements are listed below:
SEM laboratory manual p. 2 of 11

1. Sample size – the sample should be small. Of course, the maximum sample size
depends on the particular instrument and the 3500 can actually accommodate a fairly
large sample - roughly 40x80x40mm (LxWxH). However, it is generally more
practical to choose a sample size such that it can fit on the standard 12.5mm diameter
aluminum SEM stub. This will allow you to load several samples at once and avoid
unnecessary sample exchanges.
2. Sample conductivity – the sample should be electrically conductive. Sample
preparation of metals or other conductive materials in the SEM is trivial. However,
electrically insulating samples or poor conductors generally require a conductive
coating on the surface, which is grounded during imaging. This is typically very thin
(2-10nm) of a noble metal with small grain size (Au, Au/Pd, Pt) deposited by a
physical vapor deposition technique, e.g. sputter coating. Without this coating, the
sample will charge under electron bombardment creating image distortion and
increasing sample damage.
3. High Vacuum Compatibility – low vapor pressure materials or materials that can be
readily volatilized by the electron beam should not be loaded in a high vacuum SEM.
The S3500N is a variable pressure SEM, allowing for imaging of some wet samples,
but generally speaking your sample should be dry and relatively stable under electron
bombardment.
4. Surface Information – the SEM obtains information only from near the surface of
your sample. Therefore, you should provide a surface which is as clean as possible
and presents the features of interest. For example, if you want to look at the internal
structure of a polymer, it may be necessary to cleave it under liquid nitrogen to
preserve the morphology. When possible, you should avoid cleaning samples with
organic solvents as hydrocarbons are readily decomposed by the electron beam
leaving black ‘contamination’ marks on your sample.

Even though SEM's are expensive and technically advanced analytical tools, don't let them be
intimidating. For the SEM on which you will be working, there are only a few ways in which
you can damage the scope. If you remember the following things you will be fine:

1. Always wear gloves when handling your sample and the SEM sample holders. This
helps keep the vacuum system clean and avoids sample contamination.
2. Shut off the high voltage before venting the chamber. If you don’t shut off the high
voltage and bring the chamber up to atmosphere, the filament will rapidly oxidize and
burn out.
3. To change the accelerating voltage, follow this procedure: 1) Shut of the high
voltage, 2) Change the voltage level, 3) Turn the voltage back on, 4) Re-saturate the
filament. Never change the voltage ‘on the fly’ or you can destroy the filament.
4. Set the sample height correctly. Use the sample height gauge to make sure your
sample is not too tall. Always make sure the stage is in the exchange position and the
BSE detector is withdrawn when you are doing a sample exchange. This will prevent
you from accidentally hitting the objective lens or BSE detector with your sample.
SEM laboratory manual p. 3 of 11

For optimum imaging in the SEM, it is necessary to go through a series of steps to align the
microscope. The general idea of the alignment procedure is to make sure all the lenses and
apertures in the instrument are concentric about the same axis. The alignment of the microscope
is affected by many factors, including the accelerating voltage, condenser lens strength, objective
aperture size, working distance, etc. Every time you change a major parameter, it is necessary to
repeat the alignment procedure in order to obtain the best image. Unfortunately, as you will
come to see, optimum conditions vary with the sample and the type of information you wish to
obtain. Therefore, it is in your best interest to practice the alignments and become familiar with
this process.

For step-by-step operation instructions of the S3500N SEM, please refer to the Appendix.

For this portion of the laboratory, you should capture two images of the Au nanoparticles
sample - one before alignment and the other after alignment using the following conditions:

Sample: Au nanoparticles
Accelerating Voltage: 25kV
Beam Current: 30
Detector: SE
Objective Aperture: 2
Working Distance: 10mm
Magnification: >15,000X

Part 2: Signals in the SEM

There are a number of useful signals produce by the interaction of the electron beam with the
sample. Among these signals are secondary electrons (SE), backscattered electrons (BSE),
characteristic x-rays, auger electrons and light (cathodoluminescence). These signals can be
utilized to provide information about the sample topography, composition, electronic structure,
crystal structure, magnetic properties, etc. In this laboratory, we will be primarily concerned
with information provided by the SE and BSE signals. When the electron beam enters the
sample, it is scattered in three-dimensions to produce a so-called interaction volume (IV). The
dimensions of the IV depend on the primary beam energy and sample properties, such as atomic
number and density. The various signals are generated throughout the IV, but their detection
depends on their ability to escape the surface of the sample. This is shown schematically next to
a Monte Carlo simulation of primary electron trajectories for 20keV electrons striking a copper
sample.
SEM laboratory manual p. 4 of 11

SE
BSE

X-rays

Relative escape depths

SEM is renowned for its capability to achieve high resolution topographical images with
remarkable depth of focus and the ability to convey this information without any real understand
of how the image is formed. These images are typically generated by detection of the SE signal.
The SE signal is generated via inelastic collisions between the high energy electrons in the
primary beam and valence electrons in the sample. Once a valence electron has been ejected
from its orbit, it may escape the sample surface provided it has sufficient energy. The energy of
SE’s is generally quite low (<50eV), which results in a short mean free path, hence a shallow
escape depth (~10nm). This makes the information obtained with the SE signal inherently
surface sensitive with high spatial resolution. However, the situation gets a bit more complicated
as the SE signal we detect does not only come from the ‘footprint’ of the electron probe, which
we define as the SE1 signal. There are other SE signals present as well, including SE2’s, which
are generated by backscattered electrons BSE, and other SE signals generated by interaction
between primary electrons or BSE with various parts of the SEM (apertures, chamber walls,
etc.).
The BSE signal is produced by primary beam electrons that are elastically scattering by the
nuclei of atoms in the sample. As such, backscattered electrons have an energy that is some
fraction of the primary beam energy. The number of backscattered electrons generated defined
by a backscatter coefficient (η), which is a function of the atomic number (Z):
η = -0.0254 + 0.016Z – 1.86x10-4Z2 + 8.3x10-7Z3.
This atomic number dependence allows for the differentiation of different materials in a sample
based on their average atomic number. The BSE signal, since it is generated by much higher
energy electrons, is an inherently lower resolution signal that the SE1 signal, coming from up to
~40% of the maximum penetration depth. The SE2 signal, created by BSE’s exiting the surface
carries the same low resolution information and can contribute compositional contrast to the SE
image. It is a common misconception, however, that the BSE signal does not contain
topographic information. There is both an angular dependence on the generation of BSE’s and a
fixed collection angle, which provide for topographic information.

The sample for this portion of the laboratory is a silicon wafer with features patterned with
electron beam lithography (eBL). This sample was created with two separate eBL steps in
order to create features with two different metal films – pure Titanium and a 60/40 Au/Pd
SEM laboratory manual p. 5 of 11

alloy. You should capture images of patterns with each type of metal using both the SE
and BSE detectors.

SEM Conditions for Part 2:

Sample: eBL patterned silicon
Accelerating Voltage: 25kV
Beam Current: 50
Detector: SE and BSE
Objective Aperture: 2
Working Distance: 15mm
Magnification: ~10,000X

Part 3: Resolution vs. Signal to Noise

Tanstaafl. There ain’t no such thing as a free lunch. This axiom is quite applicable to the
problems of achieving high resolution images in the SEM. For this portion of the laboratory, we
will concentrate on methods for improving imaging resolution in the SEM and the trade-offs
involved. There are two major aspects to achieving high resolution images: the ability to form
small probe of electrons and the ability to detect high resolution signals produced by the
interaction with the sample. Let’s take the issue of probe formation first, which starts with the
electron gun. One of the reasons field emission guns can achieve higher resolution is due to the
smaller source size. Starting with a small source (3-5nm for cold field emission), means that less
demagnification to achieve a ~1nm probe. The tungsten source used in the S3500N SEM has a
size of ~50µm and therefore requires 4 orders of magnitude of demagnification to reach the 5nm
resolution specified by the manufacturer. This demagnification is achieved through the use of
electromagnetic lenses, called condenser lenses, which introduce a number of issues we need to
consider.

By increasing the strength of the condenser lens(es), we can achieve a smaller probe size as
shown schematically in the diagram below.
SEM laboratory manual p. 6 of 11

However, you should note that by increasing the strength of the lens, we are effectively
discarding a portion of the beam. This contributes to a signal loss and a resulting decrease in the
signal to noise ratio. In addition, one of the problems with electromagnetic lenses is that they are
fraught with aberrations. There are 4 major lens aberrations to consider: chromatic aberration,
spherical aberration, aperture diffraction and astigmatism. We can generally ignore astigmatism,
which is caused by asymmetry in the objective lens, because it can be completely corrected for in
the alignment procedure. Chromatic aberration is caused by the fact that electrons of slightly
different energies will be focused to slightly different locations in the image plane. This causes
the formation of a disk of least confusion (dc) as defined by:

dc = Ccα(∆E/E0)

where Cc is the chromatic aberration coefficient, α is the beam convergence angle, E0 is the
primary beam energy and ∆E is the energy spread. This is of major concern with thermionic
sources (high ∆E), particularly when operated at low E0. Spherical aberration is due to electrons
further away from the optic axis being more strongly effected by the lens and resulting a similar
disk of least confusion (ds) as defined by:

ds = 1/2Csα3

where Cs is the spherical aberration coefficient. Finally, the objective aperture can introduce
diffraction effects increasing the spot size (dd) as defined by:

dd ≈ 0.76/(α E01/2).

The aberration coefficients are defined by the lenses and short of buying a new SEM, there isn’t
much we can do about them. However, we do have control over the other major factors, namely
the convergence angle (α) and the primary beam energy (E0). The convergence angle is
determined by the size of the objective aperture and the working distance. For this lab, we are
going to keep the working distance fixed and look at aperture effects alone. From the spherical
aberration equation, a strong dependence between α and the probe size can be seen. However,
using a small aperture to decrease spherical aberration effects has two important consequences:
and increase in aperture diffraction effects and a decrease in the current reaching the sample
(decreases signal to noise).

Likewise, increasing the primary electron energy decreases chromatic aberration effects,
allowing the formation of a much smaller probe, but has some serious consequences. The higher
beam energy creates a much larger IV, resulting in significant reduction in resolution for the BSE
(hence SE2) and X-ray signals. The SE1 signal does not experience the same loss of spatial
resolution, but since most of the SE’s are generated deeper in the sample and cannot escape,
there is a decrease in signal to noise ratio. However, this is offset to some extent by the higher
brightness achieved in the electron gun with higher beam energy. In addition, with a typical
Everhart-Thornley SE detector, it is not possible to distinguish between the SE1 and SE2 signals
so there is a loss of surface sensitivity among other effects.
SEM laboratory manual p. 7 of 11

The sample for this portion of the laboratory is ZnO nanorods on silicon. We will explore
the effects of objective aperture size and condenser lens strength on resolution and signal to
noise ratio. You will capture at least 4 images for this section comparing the at least two
different aperture sizes and two different condenser lens strengths (beam currents).

SEM Conditions for Part 3:

Sample: ZnO nanorods on silicon
Accelerating Voltage: 25kV
Beam Current: 30 and 60
Detector: SE
Objective Aperture: 1 and 4
Working Distance: 15mm
Magnification: >20,000X

Questions for laboratory write-up

1. You may have noticed that to get a good BSE image, you need to have weak condenser
lens settings and a large objective aperture. Why do you think this is the case?

2. For the Au/Pd patterns on silicon, why do the smaller features appear darker than the
larger ones if they have the same composition?

3. Note that even though the image on the screen may get noisier with decreasing aperture size
or increasing condenser lens strength, the photo taken may not show this. Please explain
why this occurs.

4. Since SEM is essentially a surface analysis technique, why doesn’t sputter coating
obscure useful surface information? Can you think of any cases where the coating might
become a problem?

5. Low voltage microscopy is a useful method for minimizing charging effects. What are
some of the other benefits and tradeoffs? Why is cold field emission well suited for this
technique?

6. If you didn’t have a dedicated BSE detector in your SEM, can you think of a way to use
the E-T SE detector for this purpose?

7. The image below is a pattern produced by electron beam lithography in PMMA on a silicon
substrate. The PMMA was first spun onto the wafer (~150nm thick) and then the inner and
outer lines of the letters in the ‘NUANCE’ pattern were written by taking control of the
SEM scan coils. After exposure to the electron beam, the PMMA becomes more soluble in
the developer and can be removed. After developing, the sample was then sputter coated
with a uniform thin film of Pt/Pd prior to imaging. For reference, the letters are ~2µm tall
and the lines are ~30nm wide.
SEM laboratory manual p. 8 of 11

Why do the N, A and E have different contrast than the rest of the letters and the rest of the
sample? Why are they dark? (This pattern was written several times on the sample and different
letters were dark in each, so there is no difference in processing from letter to letter besides random
variation.) Also, try to explain why the center of the ‘A’ is darker that the rest and why the ‘E’ is
darker on the horizontal parts.

References

Joseph I. Golstein et al, Scanning Electron Microscopy and X-Ray Microanalysis, 2nd ed,
(New York: Plenum Press, 1992).

P E J Flewitt and R K Wild, Physical Methods for Materials Characterization, (London:

Institute of Physics Publishing, 1994).
SEM laboratory manual p. 9 of 11

Appendix
Hitachi S-3500N VP SEM
Operation Instructions
For additional assistance, please contact the facility manager.

Please contact under emergency:

SEM manager: Mr. Ben Myers, 1-3439 (O), 312-593-8298 (cell)
b-myers3@northwestern.edu
EPIC manager: Dr. Jian-Guo Zheng, 1-7807 (O), 847-675-7387(h),
j-zheng3@northwestern.edu
EPIC director: Prof. Vinayak P Dravid, 7-1363 (O), 847-486-1705 (h),
v-dravid@northwestern.edu

S-3500 reservations are made using the EPIC login system. Please follow all EPIC facility rules
for using this system.

Note: It is imperative that gloves be worn during all sample exchange procedures. If you cannot
find any gloves, please ask!

You are asked to make a copy of your data on your own disk IMMEDIATELY after your session
is finished. You may save your data on a 100MB Zip disk or transfer your data by FTP. The data
may be deleted at any time without notice. EPIC is not responsible for any data loss.

System Startup
1. Login to the reservations PC and on the paper log at the SEM.
2. Prepare the sample as necessary and mount on a Hitachi sample holder. Check the height
of the sample using the sample height gauge. The total height of the sample holder and
sample should not exceed the top of the height gauge.
3. Verify that the Hitachi PC_SEM software is running.

(Note: The software will occasionally crash and the PC will need to be rebooted. It is
sometimes sufficient to simply restart the PC, but it is often necessary to re-sync the PC
and SEM. To do this, simply shutdown windows, turn off the display power on the left
side of the column control panel, wait ~10 seconds and turn the display power back on.
The PC should automatically boot up – hit ‘cancel’ at the network prompt - and the
PC_SEM software should load automatically. If the software still doesn’t load, try it
again or contact the facility manager.)

4. Check to make sure the SEM stage is in the sample exchange position: X=30, Y=20,
Z=EX, Tilt = 0.
5. Make sure the backscattered electron (BSE) detector is withdrawn and the accelerating
voltage (HV) is off.
6. Depress the AIR/EVAC button on the column control panel to vent the sample chamber.
Do not try to force the door open by pulling on it – wait for the chamber to vent and the
door to release on its own.
SEM laboratory manual p. 10 of 11

7. Place the sample holder in the chamber. Using the sample chamber height gauge,
double-check the height of the sample to insure the holder is fully inserted.
8. Select Vacuum Mode under the Setup menu and set the vacuum level. For standard,
high vacuum operation, choose SEM and for variable pressure operation, chose the VP-
SEM mode.

(Note: For VP-SEM mode the standard lower SE detector cannot be used and the system
will automatically select the BSE detector. To insert the BSE detector, lift gently on the
bottom rod to release the detector and slowly insert the detector by pushing it toward the
column.)

8. While holding the door firmly shut, press the AIR/EVAC button to evacuate the sample
chamber.
9. Watch the vacuum indicator lights on the screen as they change from RED to GREEN to
BLUE. The BLUE light indicates that the chamber has reached the desired vacuum
level.
10. Once the chamber reaches vacuum, select HV Control under the Setup menu.
11. Select the desired accelerating voltage and press the HV ON button.
12. As soon as the high voltage is on, select Low, Med or High filament saturation and hit
AFS to engage the automatic filament saturation.

(Note: Never select high filament saturation if the sample chamber has just been at
atmospheric pressure since the filament will likely become over-saturated as the vacuum
improves. Choose Med or Low for normal operation to extend the filament life. When
changing the accelerating voltage, never just switch it on the fly. First, select HV OFF,
then change the accelerating voltage, select HV ON and then re-saturate the filament to
the desired level.)

General Operation and Alignment

1. Locate your sample using the X, Y and rotate stage controls.
2. Select Column Setup under the Setup menu.
3. Set the working distance (WD) to the desired level and then adjust the Z-axis control to
raise/lower your sample until it comes into focus.

(Note: For higher resolution, use a shorter working distance. For better depth of focus,
use a longer working distance. For EDS analysis, the system is optimized at a 15mm
working distance – this is of particular importance for any quantitative analysis.)

4. Set the desired beam current level – this controls the strength of the condenser lenses in
the column. Higher values of beam current will result in greater signal strength, but poor
resolution.
5. Insert the desired aperture. Align the aperture for the brightest image using the X and Y
knobs on the aperture.
6. Zoom into a relatively high magnification (>10,000X) and find a feature on your sample
to use for alignment.
7. Select Alignment under the Operate menu.
SEM laboratory manual p. 11 of 11

8. Select Aperture Alignment - the focus wobble will turn on automatically. Adjust the X
and Y knobs on the aperture to minimize translation of the image.
9. Select Gun Shift and hit the AGA button.
10. Select Gun Tilt and hit the AGA button.
11. Select Stig. X and use the X and Y stigmator/alignment knobs on the control surface to
minimize translation of the image. Select Stig. Y and repeat.
12. Close the alignment menu and use the X/Y stigmator/alignment knobs on the control
surface to correct for astigmatism.

Image Capture
1. The preferred method for digital image capture is through the PC_SEM software.
2. Select H.R. Capture under the Scan menu.
3. Select the desired scan speed and resolution then hit the H.R. Capture button to capture
the image.
4. To save the image, select Save Image As under the File menu.

X-ray Microanalysis

1. Energy Dispersive X-ray analysis (EDS) is available through the attached PGT system.
2. To capture an x-ray spectrum, select X-ray Collection under the X-ray menu. Set the
desired parameters and hit start.
3. Image capture and x-ray mapping are available through this system as well by select
Image Collection under the Image menu.
4. For more details on the PGT system, see the operation manual or contact the facility
manager.