Report Submission Date

Download 58.9 Kb.
NameReport Submission Date
A typeReport > manual > Report

NE 320 L Characterization of Materials

University of Waterloo

Nanotechnology Engineering



Rajesh Swaminathan – 20194189

Ryan Iutzi – 20202504

Group Number:


Experiment Name and Number:

#5 SEM Characterization of Nano Materials

Experiment Date:


Report Submission Date:


Report Submitted to T.A.:

Bryan Kuropatwa

Table of Contents

1.Introduction 3

Objective 3

Background and History 3

Comparison to Other Techniques 4

Uses 5

2.Theoretical Principles 6

3.Experimental 8

Equipment 8

Materials and Procedure 10

4.Observations and Results 10

Qualitative Observations 10

Quantitative Observations 11

5. Discussion 11

Error Analysis 12

5.Conclusions 13

7. References 14

Appendices 16

Appendix A – Original Observations 16

Sample #1 - Glass fibre reinforced polyethylene terethphtalate 16

At 5 kV 16

Appendix B – Questions 18

Appendix C – SEM Images and EDX Spectra 20

  1. Introduction


The objective of this experiment is to familiarize with the workings of a scanning electron microscope (SEM) by analyzing some topographical and compositional properties of samples of glass fibre reinforced polyethylene terephthalate and lava rock by detecting secondary and backscattered electron signals, and determining the composition of the lava rock by EDX.

Background and History

Scanning electron microscopy is a technique in which a sample is imaged by exposing it to a high energy-beam of electrons and scanning this beam across the specimen. The electrons interact with the atoms in the sample and are then detected, providing information about the sample’s topography and composition. SEM is often capable of resolving images down to 1-5 nm in resolution. Additionally, due to the mechanism in which the device operates, a SEM provides a very high depth of field, allowing for 3-dimensional-like images to be viewed. [1] The SEM typically has three types of signals: secondary electrons, which provide topographical information, backscattered electrons, which provide compositional information, and X-Rays, which give the identity of the sample. These types of signals are discussed more in depth in the theory section, and Appendix B.
The theory of the scanning electron microscope was first demonstrated by Max Knoll in 1935. The first actual SEM was produced in 1938 by von Ardenne. In 1939, Ruska and von Borries, employees of Siemens, produced the first commercially-available SEM. [2]
Today, SEM has become a cornerstone of materials characterization, and the design of a typical SEM is continually being improved. For example, older SEM’s utilize a thermionic tungsten filament to produce an electron beam. The filament is heated to around 2800 K, above an anode plate, which causes electrons to be pulled from the filament towards the anode, producing the beam. A newer innovation, however, utilizes a field emission source, in which a <111> oriented tungsten crystal is placed at a strong 4 kV potential which causes electron emission from the tip of the crystal, without large amounts of heat. This allows for brightness around 104 times that of the older thermionic tungsten filament. It also allows for a smaller initial spot size and a lower accelerating voltage which creates a much smaller excitation zone. This allows for a better resolution than that of tungsten filament source SEMs. [2]

Comparison to Other Techniques

SEM is similar to transmission electron microscopy (TEM) in the sense that an electron beam is used as the primary source. The main difference, however, is that TEM is a purely optical instrument since it uses a lens to form the image. The electron beam is passed through the sample, and then a lens projects the image onto a plane. SEM, on the other hand, is not an optical instrument in the true sense. It uses electrons to illuminate the sample, and then detects them using various electron detectors. SEM essentially looks at electrons being emitted from the sample, while TEM looks at the electrons that pass through, or get diffracted, depending on the configuration. As a result, TEM needs thin, smooth samples, while SEM can have samples of varying thickness. Also, SEM has a higher depth of focus, allowing it to show depth and giving a three-dimensional-like image, where TEM always gives a 2-dimensional image. This concept will be explained more in depth in the theory section. [2]
Compared to optical microscopy, SEM is much superior in its resolution. The key difference is that SEM uses an electron beam, which is less prone to diffraction effects that lower the resolution, whereas optical microscopy uses light, where diffraction severely limits the resolution. Also, like TEM, optical microscopy requires thin samples and gives a fairly two-dimensional image. In fact, optical microscopy is much more similar to TEM in that they are both optical methods. [2]
Nevertheless, all three methods are similar in the sense that they use lenses to focus the incoming beam. All three techniques involve a condenser lens to originally focus it. Optical microscopy and TEM then have the specimen right below this lens, followed by objective and projective lenses. SEM, has a different type of set-up, which will be shown in the experimental section. SEM and TEM also use electrostatic or electromagnetic lenses, while optical microscopy simply uses a diffractive material. [2]


Scanning electron microscopy is useful in a variety of research environments because it is able to provide an actual image of a specimen, as opposed to techniques such as XRD or Raman. Additionally, the image is much more three-dimensional than that of transmission electron microscopy (TEM), as the depth of field is much larger for SEM. As a result, SEM is typically used for rough samples that cannot easily be made into a smooth surface. In the area of materials science, it is very useful for imaging different types of materials, especially for inspections of grain structures, which can be seen well with SEM. In general, an analysis of the structure of a material from the near-nanoscale up to the macroscale can be done with SEM. Additionally, one can determine information on composition, such as where heavier atoms are located in the structure, which gives information about the actual chemistry of the sample from a spatial perspective.
In nanotechnology, SEM is particularly useful for nanomaterials, as the small grains of nanomaterials can often be imaged. Additionally, nanoparticles and nanotubes can be seen and their structures can be analyzed in the same was as is done with materials science research. Furthermore, SEM is useful in the field of nanoelectronics, since nano-scale transistors (which are now around 65 nm minimum feature width [3]), are imaged to determine the robustness of the process. This is frequently done as a characterization method in semiconductor fabrication facilities [2].
Outside of materials and nanotechnology, SEM can essentially be used to characterize anything “small”. For example, biologists frequently use this technique (well before materials scientists and nanotechnology researchers) to study cells and organisms on a micro scale. [2]
  1. Theoretical Principles

The basic principle of SEM is that an electron beam is positioned on the sample, and interactions lead to different types of electrons and photons being emitted from the sample, which are detected. The first type of electrons, secondary electrons, are formed when an electron from the beam passes near an atom in the sample and ionizes it. This means that an electron in the valence band of the sample uses some of the energy of the beam electron to escape from the atom. This energy is small, because of the small work function, that is, the small amount of energy required for an electron in the valence band to escape into vacuum. The energy of secondary electrons is 50 eV, fairly low, and they must be generated near the surface in order to escape. Therefore, they are produced near the surface, within about 10 nm. As a result of this, secondary electrons give information on the topography of the sample. This results in a sharp image which clearly shows the folds and curves of the structure. It also gives a perception of depth, because secondary electrons can leave the sample in various directions and also because the brightness of the image is a function of depth (because the further away the surface is, the fewer secondary electrons can escape from it). This is what allows for a three-dimensional-like image. [2]
The other main type of electron produced is backscattered electrons. They are caused by an incident electron from the beam colliding with the nucleus of an atom in the sample scattering backwards. As there is no energy transfer, the backscattered electrons have equal energy to the electrons in the beam. As a result, they can escape the sample from much deeper. Higher atomic number atoms have much larger nuclei, and as such, are more effective in scattering electrons. As a result, backscattered electrons are produced more from larger atoms. This allows for compositional information to be retrieved from backscattered electrons, as the brighter a spot on the image, the heavier the atoms located in that region must be. This technique is quite useful in detecting heavy atoms. [2]
Finally, X-rays can be produced from the incident beam. This occurs when an electron in a shell of an atom absorbs some energy of the incident electron, and moves to a higher energy level, leaving a hole in the original shell. An electron then drops back down to recombine with this hole, and an X-ray photon is emitted, equal in energy to the difference between the higher and lower energy level. This can occur very deep in the sample, because since X-rays are not charged, it is very easy for them to escape the sample compared to electrons. X-rays are produced and escape deeper in the sample than both secondary and backscattered electrons. The wavelength of the X-Ray must obviously be characteristic of the atom, because the difference in energy between the two energy levels is always characteristic of an atom. Hence, a spectrum of the X-Ray wavelengths emitted from the sample give an indication of the chemical composition, known as EDX. This technique works fairly well with all elements. However, the electron must be at least the energy of the energy level transition, so detecting some atoms with high transitions will require a high accelerating voltage. Typically, 15kV accelerating voltage is required. [2]
Figure 1 shows all three types of signals, and the penetration depth from which these signals arise. As is consistent with what has been explained, secondary electrons come mainly from near the surface, while backscattered go deeper, and characteristics X-Rays come from the deepest point. The overall tear-drop shape is the interaction volume.

Figure 1: SEM Primary Signals
The size of the teardrop depends on the atomic weight of atoms in the sample, and on the accelerating voltage (proportional top the electron beam intensity). The relationship is demonstrated in figure 2. As can be seen, as accelerating voltage is increased, or atomic weight is decreased, the interaction volume becomes much larger. This leads to a lower resolution, as there will be overlap in signals from adjacent regions on the specimen. [2]

Figure 2: Effect of atomic weight and accelerating voltage on interaction volume

  1. Experimental


In this investigation, the scanning electron microscope was an S-3500N. SEM manufactured by Hitachi High Technologies Canada. The basic setup of the SEM is shown in Figure 3.

Figure 3: SEM Setup
The electron source is what produces the electron beam (traveling down) with all electrons being monochromatic. The SEM in this investigation uses a tungsten filament. The condenser lens forms a beam (there may be multiple lenses). The objective lens focuses the beam onto the specimen. The beam is scanned across the specimen surface using the scan coils. The scanning process is controlled by the scan generator in the magnification control. The electrons from the beam interact with the specimen and the detector then counts the electrons emitted when the beam is focused at a given location. The intensity is outputted, amplified and shown on the display. [2]
Another important instrument that usually accompanies an SEM is a sputter coater, which coats a conductive material onto the substrate. This helps deflect secondary electrons, giving a sharp image for topography. Also, this prevents the sample from charging. Charging occurs when electrons become trapped in an insulating sample then slowly get released causing very high brightness in the image. Adding a conducting film allows the electrons to discharge, effectively grounding the specimen.
The setup of a sputter coater is shown in Figure 4. The gold is located on the target. A large voltage is established, which ionizes the argon that is pumped into the chamber. The positively charged ions are driven to the top electrode (at negative potential) and strike the anode, liberating gold atoms which are then deposited onto the specimen. [3]

Figure 4: Sputter Coater

The main disadvantage to sputtering a gold film is that gold absorbs X-Rays, which interferes with the EDX spectrum of the specimen. Hence, if EDX analysis is required, the carbon coatings must be used instead. This can bee done by carbon filament heating or carbon arc coating. [4] Carbon coating was not done in this investigation, and any EDX analysis was performed on a conducting sample.

Materials and Procedure

The following materials were used in this investigation

  • Glass fibre reinforced polyethylene terethphtalate (examined by secondary electrons and at various accelerating voltages)

  • Lava rock sample (examined by secondary electrons, backscattered electrons, and EDX)

  • Gold coated non-conductive sample (not examined with the SEM)

  • Lava rock nanotubes (examined by secondary electrons)

The operational procedure for the use of the scanning electron microscope is outlined in the lab manual Characterization of Materials [5], under experiment #5. Only the specific steps for acquiring an image or EDX spectrum were used. The overall procedure is described in SEM Lab Task [6]. This procedure was followed as is written. For the fourth section, carbon nanotubes grown from lava rocks were examined. However, these nanotubes are not involved in the analysis.

  1. Observations and Results

Qualitative Observations

The high resolution image obtained of the glass fiber reinforced PET composite at 15 kV in SE mode was about the same quality as that obtained at 5 kV. However, the high resolution picture was slightly “bent”. This effect was not visible while using the BSE and EDX detectors. At 25 kV, we had much more contrast which led to several bright and dark spots. As we jumped back from 15 kV to 5 kV, we noticed the image to drift slowly from one side of the screen to the other and this made the sample more difficult to image.
Next, a BSE detector was used to obtain images of the sample. This required us to turn off the camera in the sample chamber to avoid interference. Appendix C contains images of the lava rock taken using both the SE detector as well as the BSE detector. The SE image contains more information about the surface topography of the sample whereas the BSE detector yielded more details pertaining to the morphology of the sample.

Quantitative Observations

We obtained 8 spectra from different parts of the lava rock sample by conducting an EDX analysis. The percentage of abundance for each sample was obtained by comparing the peak heights. On average, the sample contained 56.1% tellurium, 32.9% copper, 9.5% barium, and 1.5% selenium.
Note that all original observations and data are included in Appendix A.

5. Discussion

In general, the quality of the experimental data and SEM images was relatively good and was reasonable with respect to the theory revolving SEM imaging. Any unreliability in the images (such as “bending”) were caused due to surface charging effects which could be eliminated using a beam with a low energy spread [7].
Secondary electrons had much better resolution than back-scattered electrons. We attributed this observation to the fact that the back-scattered electron beam begins to broaden because of strong elastic scattering effects. The beam also undergoes inelastic interactions that cause an energy loss of the electrons.
We attributed the “bend” in our SEM images to localized electrostatic charging of the surface sample. Electrons are charged particles that rush from the anode to the cathode. If the cathode (in this case, the aluminum sample holder and the sample itself) isn’t effective in absorbing the electrons, the electrons tend to stay on the surface of the sample. Thus, surface charging effects are prominent when working with non-conductive samples but can be reduced by coating the sample with a good conductor like gold. The gold helps discharge the free electrons consequently removing them from the sample surface. This charging effect was not visible while using the BSE and EDX detector since these detectors are not influenced by surface characteristics [8].
When we shifted back from 15 kV to 5 kV, a similar charging effect was observed, once again due to the build-up of static charge that caused the image to drift from one side of the screen to the other.
We deduced that the major constituents of the lava rock sample we analyzed were tellurium, copper, barium and selenium in that order of abundance. We obtained 8 spectra from different parts of the sample and found that the composition was fairly uniform throughout since the standard deviation of the sample compositions were relatively low (~5-10%).
Sometimes the sample is coated with a conductive material (e.g. gold), especially non-conductive samples. One of the major reasons for coating a non-conductive specimen with a conductive material like gold is to increase the number of secondary electrons that will be emitted from the sample. This gives us a much sharper image and a higher resolution. The other reason is to provide a free path for electrons to discharge thus preventing surface charging effects and their resulting bright spots.

Error Analysis

One big source of error is the excess charge build-up on the surface of the sample. This led to bright spots that obscured certain key features of the sample during imaging. The charge may have also interfered heavily with the X-ray absorption and emission, thereby altering the signal. All of these effects contribute towards error in the EDX measurements. This charging effect could have been reduced by sputter coating the surface with a thin uniform layer of gold [8].
The results of the EDX analysis, as computed by the computer software are in Appendix C. The software also calculates the error associated with each weight %. The error associated with the most abundant element tellurium is only 3.2%, whereas the error associated with the second-most abundant element copper is 6.8%, more than double. We could identify a few sources for these errors. Free gases like oxygen and nitrogen, and contaminants in the air could alter the composition of the lava rock. Improper handling of the sample, like for example touching it with bare hands, could introduce moisture. Manually selecting regions of the sample for scanning due to their fascinating morphologies could have been improved upon by gradually rastering the sample to cover its entire sample area [9]. This would mean much more data from throughout the sample, thereby removing any bias in the measurements and improving accuracy.
In summary, three simple ways of improving accuracy are:

  • Manipulate the sample only with provided tweezers and never with bare hands.

  • Store the sample in a clean, dry and possibly inert environment (e.g. argon or nitrogen gas)

  • Perform the EDX measurements throughout the sample to obtain a more uniform and representative sample space. This can be achieved by rotating the sample by a fixed angle after every measurement.

  • Coat the sample with a thin uniform layer of gold to reduce surface charging. This may interfere with the EDX results to the introduction of a new element, so the software should be asked to ignore this element before performing any analysis assuming of course that all gold originates from the coating and there is none in the sample to begin with.
  1. Conclusions

The major objective of this lab was to familiarize ourselves with the SEM instrument and to learn how to prepare and analyze a given samples using the SEM equipment. As a consequence, we were required to learn how the instrument worked and the difference between secondary electrons, back-scattering electrons and x-rays. Each of these provides a different kind of information based on the interaction of the x-ray beam with the sample as a function of the beam’s energy. This is critical to knowing what kind of information can be gleaned from each type of detector. We learned why the SEM was capable of producing images of much more resolution as compared to a simple optical microscope and this is mainly due to the fact that the SEM uses electrons which have a much smaller de Broglie wavelength than regular photons.
A key observation is that increasing the accelerating voltage of the electron beam increases the resolution of the images obtained, but does so at a compromise. A higher energy electron beam increases the amount of surface charge accumulation which interferes with the measurements and produces bright white spots that distorts the SEM scans.
The lava rock sample that was analyzed was concluded to contain elements such as tellurium, copper, barium and selenium in that order of abundance. The error in the percent weight of the composition was fairly acceptable for the two most abundant elements tellurium and copper – 3.2% and 6.8% respectively.
A few improvements for future labs were also stated to improve the accuracy of our results. Effective storage in an inert environment and careful handling using specified tweezers is critical to preserve the integrity of the sample. Coating the surface with a thin uniform layer of gold helps minimize surface charging effects, although the analyst must remember to tell the computer software about this lest it interfere with any quantitative analyses.

7. References

[1] Shackelford, James; “Introduction to Materials Science for Engineers”, Pearson Prentice Hall, Upper Saddle River, NJ, 2004
[2] T. Hesjedal, Scanning Electron Microscopy, NE 226 Course Notes, University of Waterloo, Waterloo (2007)

[3] Plummer, James; “Silicon VLSI Technology”, Pearson Prentice Hall, Upper Saddle River, NJ, 2000

[4] Lee, William; “Ceramic Microstructures”, Chapman & Hall, London, UK, 1994
[5] Q. Xie, F. McCourt, Nanotechnology Engineering NE 320L Lab Manual,

University of Waterloo, Waterloo, pp 35-39 (2008).

[7] Wischnitzer, Saul; “Introduction to Electron Microscopy”, Pergamon Press Inc,

Elmsford, NY, 1970

[8] Ian Watt; “The Principles of Electron Microscopy”, Cambridge University Press,

Cambridge, 1985

[9] Newbury, Dale. “Advanced Scanning Electron Microscopy and X-Ray

Microanalysis”, New York: Plenum Press, 1986.


Appendix A – Original Observations

Sample #1 - Glass fibre reinforced polyethylene terethphtalate

At 5 kV

100 μm scale

  • stigmated, had to lower contrast

30 μm scale

  • captured HR image (Image 1), image was not good

  • Found another feature and captured (Image 2), better quality

At 15 kV

50 μm scale

  • Fibers visible, ~60-70 μm

20 μm scale

  • Took Image (Image 3)

  • Some “wooziness” on screen due to charging

  • Moved to other area, bright spots present due to static charge

At 5 kV Again

100 μm scale

  • Charging effects still visible, very bright image

  • Took Image (Image 4)

Sample #2 Lava Rock

At 25 kV

50 μm scale

  • Crystal appears to be ~160 μm

  • Acquired Image on SE (Image 5)

  • Acquired Image on BSE (Image 6)


Acquired first spectrum (Spectrum1)

  • Presence of S, Cu, Te, Ba, Se

Acquired second spectrum (Spectrum 2)

  • Presence of Ti and Si as well

Appendix B – Questions

Question 1

Why is an electron beam with a low energy spread desirable in an SEM?

A high energy spread essentially means more noise in the experimental data. A low energy spread is desired because all points on the sample need to be exposed to the same energy of electrons. Otherwise, higher energy areas will have more scattering, leading to brighter spots in these areas, and darker spots in areas where the energy is lower.

Another issue is that electrons of different energies will be affected differently by the electrostatic and electromagnetic lenses used in the SEM to focus the beam. This leads to chromatic aberration, which distorts the image. [2]

Question 2

Describe each of the important signals produced by the interactions between the electron beam and a sample, and explain which information can be derived from each signal

The answer to this question has already been addressed at the beginning of the theory section. For simplicity, the answer will be summarized in a chart. For a more in-depth explanation, please refer to the theory section.


Caused By


Penetration Depth

Information Derived

Secondary Electrons

Ionization of atoms, freeing a valence electron




Backscattered Electrons

Elastic scattering



Composition (brightness proportional to Z)

X-Rays (EDX)

X-Ray emission from atom after electron falls from higher energy level



Composition (energies are characteristic of atom)

Appendix C – SEM Images and EDX Spectra

Image 1

Image 2

Image 3

Image 4

Image 5

Image 6

Spectrum 1

Spectrum 2

Share in:


Report Submission Date iconDue Date for full compliance of the Report – (15 days from the date of submission)

Report Submission Date iconReport Submission Date

Report Submission Date iconDate of report (Date of earliest event reported) September 10, 2010

Report Submission Date iconDate of report (Date of earliest event reported): January 7, 2011

Report Submission Date iconSubmission of Annual Quality Assurance Report (aqar) of

Report Submission Date iconSubmission of Annual Quality Assurance Report (aqar) of

Report Submission Date iconReport Date

Report Submission Date iconReport Date

Report Submission Date iconReport Date

Report Submission Date iconReport no.   Date  

Report Submission Date iconReport date

Report Submission Date iconReport Date

Report Submission Date iconReport Date: Tuesday February 10, 2009

Report Submission Date iconReport Run Date: 2011-03-10 12: 23: 07. 696 Est

Report Submission Date iconDate Title of Report in Initial Capital Letters

Report Submission Date iconReport Run Date: 2011-02-12 10: 55: 27. 751 Est

Report Submission Date iconReport (Date of earliest event reported): November 21, 2007

Report Submission Date iconReport (Date of Earliest Event Reported): February 21, 2017

Report Submission Date iconReport (Date of earliest event reported) February 28, 2007

Report Submission Date iconManual Defibrillators Begin Date 02/15/2005 End Date 02/14/2006


When copying material provide a link © 2017