The SEM uses an electron beam to perform a circular scanning motion on the surface of the sample, while monitoring the generation of various signal images in real time, and then modulating the image according to the amount of signal generated.

There are three sources of contrast in SEM images that cause the various signals to be generated:

1. The properties of the sample itself: including the unevenness of the surface, the difference in composition, the difference in orientation, and the difference in surface potential.

2. The properties of the signal itself: mainly including secondary electrons and backscattered electrons.

3. Manual processing of signals.

First, the morphology contrast

Topography contrast refers to the difference in the angle of inclination at different positions of the sample.

A. Relationship between secondary electron yield δ and incident electron beam angle α

When electron beams of the same energy bombard different topographic regions of the sample, the depth of the secondary electrons produced is the same. However, after secondary electrons are generated, they must be able to escape the sample surface in order to be received. Suppose α represents the angle between the incident electron beam and the normal of the specimen surface. For a smooth specimen surface, the yield of secondary electrons is δ ∝ 1/cosα. As the α increases, the incident electrons get closer to the surface layer of the specimen. As shown in Figure 1, when α2 > α1, the path of the incident electron beam (2) is closer to the surface of the specimen than that of the incident electron beam (1), so more secondary electrons escape from the sample surface. In order to obtain a strong secondary electron signal, it is often necessary to tilt the specimen stage, i.e. to change the angle of the incident electron beam (change the α) to induce more secondary electrons to be excited.

Fig.1 Relationship between inclination angle and SE yield

B. Effect of incident electron direction on backscattered electron yield η

Backscattered electron yield η indicates the probability of an electron being produced with an initial electron energy greater than 50 eV and less than the initial energy. When the electron beam is incident perpendicular, the distribution of backscattered electrons is close to the cosine law, and the direction of emission is random (see Figure 2); However, when the electron beam is obliquely incidentified, the angular distribution of the backscattered electrons appears in a forward rod shape (see Figure 2). As the angle of incidence α increases, the backscattered electrons are closer to the surface of the specimen and their yields increase accordingly. When the angle of incidence is close to the grazing angle, the back reflection coefficient η is close to 1.

Although η increases with an increasing inclination angle of α, it does not exactly satisfy the secant relation, as shown in Figure 3. As the α angle increases, the backscattered electron reflectance coefficient of η increases, indicating that the backscattered electron reflectance is also sensitive to the sample surface state.

Figure 2 Angular distribution of BSE

Fig.3 The relationship between the back reflection coefficient η and the tilt angle

C. Topography contrast

For SEM, the direction of the incident electrons is fixed, but due to the unevenness of the specimen surface, the angle of incidence of the electron beam on the specimen surface is different. As shown in Figure 4, the angles of incidence α in the two planes A and B in the specimen are different. According to the reflection law of secondary electrons and backscattered electrons, it can be seen that the larger the incident angle α, the higher the secondary electron yield δ and the backscattered electron reflection coefficient η. As a result, the number of secondary and backscattered electrons received by the SEM detector is also different, resulting in a difference in brightness on the image. For example, the angle of incidence in region A is greater than in region B, so region A receives more secondary electrons and backscattered electrons. As a result, the A region is brighter than the B region in the image, thus revealing the topography contrast of the specimen.

Fig.4. Surface topography contrast

The secondary electron yield and the reflection coefficient of the backscattered electrons can reflect the morphological contrast of the sample, but due to the deep exit depth of the backscattered electrons and the relatively large emission area, the spatial resolution is much lower than that of the secondary electrons, and its three-dimensional perception is not as good as that of the secondary electrons. The backscattered electrons largely reflect the topography of the subsurface. As shown in Figure 5, it can be observed that the secondary electron image is more sensitive to the surface morphology than the backscattered electron image of the lithium battery electrode material. However, many desktop SEMs now use backscattered electrons to observe the topography of the specimen. In addition, in certain cases, the observation topography of backscattered electrons has an advantage over secondary electrons, which will be discussed later.

Fig.5. SE image and BSE image of lithium battery electrode material

However, although the higher the tilt angle, the higher the backscattered electron yield, the divergence angle distribution also changes. Therefore, although the backscattered electron yield is large, it does not mean that all backscattered electrons can be received efficiently. As a result, sometimes the relationship between chiaroscuro and tilt angle in a backscattered electron image is not exactly the same, as shown in Figure 6.

Fig.6. SE and BSE images with pyramidal morphology (anomalous contrast)

D. Edge effects

Fig.7 Edge effect and tip effect in topography contrast

However, there are obvious edge effects and tip effects in the topography contrast of secondary electrons or backscattered electrons, as shown in Figure 7. At some protruding tips, small particles, steeper inclined planes, and at the intersection of multiple planes, there are more paths for secondary electrons to escape to the surface than normal, so the yield of secondary electrons tends to be much higher than in normal planes at these locations. This phenomenon appears in the image as a very bright area of tip effect or edge effect, forming a white dot or white outline, as shown in Figure 8.

Fig.8 Edge effect and tip effect

2. Atomic number Z contrast (component contrast)

A. Relationship between secondary electron yield δ and atomic number Z

Fig.9 The SE yield of different substances is different

In addition to the morphology, the secondary electron yield is also related to the energy of the incident electron beam and the atomic number Z, as shown in Figure 9. The relationship between secondary electron yield and atomic number is complex. Because substances with different atomic numbers have different numbers of extranuclear electrons and different ionization energies, there is a certain relationship between the secondary electron yield and the atomic number. In addition, the yield of backscattered electrons for different atomic number pairs will also be different, and the backscattered electrons will produce secondary electrons.

In general, the yield of secondary electrons increases with the increase of atomic number. When the atomic number Z is less than 20, the secondary electron yield increases with the increase of atomic number Z. However, when the atomic number Z is greater than 20, the secondary electron yield basically does not change with the change of atomic number. Only the secondary electron yield of the element with the lower atomic number is related to the composition of the sample. Therefore, in general, secondary electrons are used to observe surface topography and not to observe component distribution; However, in the case of low atomic number or large differences, the secondary electrons are also able to show the contrast of the atomic number.

B. Relationship between the backscattered electron coefficient η and atomic number Z

The backscattered electron coefficient η has the following relation to the atomic number Z: η increases with the increase of Z, as shown in Figure 10.

η＝-0.0254+0.016Z-0.000186Z²+8.3×10^-7Z³

Fig.10. Relationship between backscatter coefficient and atomic number

Fig.11 Relationship between SE and BSE yields and atomic number

C. Atomic number contrast

Figure 11 illustrates the relationship between secondary and backscattered electron yields and atomic numbers. As can be seen from the figure, both secondary electrons and backscattered electrons increase in their yields as the atomic number increases. Therefore, when analyzed, the region with higher atomic number in the sample will emit more secondary electrons and backscattered electrons than the region with lower atomic number, so the region with higher atomic number will show brighter features in the image, which is the basic principle of atomic number contrast.

In addition, Figure 11 reveals the problem that the atomic number Z contrast reflected by the secondary electrons is much weaker than that of the backscattered electrons.

When the atomic number Z is small or the Z difference is very large, the secondary electrons can still exhibit good atomic number Z contrast. As shown in Figure 12, the left figure shows the secondary electron image of the carbon-silver hybrid material, the atomic number of carbon is small, and the atomic number of silver is large, and there is still a large difference in the secondary electron yield between the two, so carbon and silver can be easily distinguished from the secondary electron image. The backscattered electron image on the right has a more pronounced compositional contrast, but its surface detail is much less than that of the secondary electron image.

Fig.12 SE and BSE images and carbon and silver electronic yields of carbon-silver mixed materials

Although both secondary and backscattered electrons are capable of exhibiting atomic number contrast, the sensitivity of backscattered electrons to atomic number is always much higher than that of secondary electrons, regardless of atomic number Z. As a result, backscattered electrons are more widely used for compositional analysis. Based on the difference in brightness of the backscattered electron images, combined with the knowledge of the specimen, we can quickly and qualitatively determine the phase type.

For example, as shown in Figure 13, the specimen is a cross-section of a copper-clad aluminum wire material with an outer layer of copper and an inner layer of aluminum. Since the difference in atomic number between aluminum and copper is not large, the difference in their secondary electron yield becomes small, so the secondary electron image can no longer distinguish the different components well. However, there is still a large difference in the yield of backscattered electrons, so the contrast of the backscattered electron image is very pronounced, and the distribution of copper and aluminum can be easily distinguished.

Fig.13 SE and BSE images and aluminum and copper electronic yields of copper-clad aluminum wire cross-sections

Fig.14 SE image and BSE image of the coated surface

Figure 14, for example, illustrates the case of a coating material. Due to the influence of large surface morphology contrast on the surface surface and the existence of edge effects, the secondary electron yield can no longer show atomic number contrast at all. However, the backscattered electrons are less affected by the topography at this time and can still show a distinct compositional contrast.

3. The relationship between morphology contrast, atomic number contrast, secondary electrons, and backscattered electrons

At present, in many places, secondary electron images are called topography maps, and backscattered electron images are called composition maps, which are not very rigorous. From the previous introduction, we already know that the secondary electron yield is mainly more sensitive to morphology, and the backscattered electron yield is mainly more sensitive to the composition. However, the secondary electron image can also reflect a certain component contrast, and the backscattered electron image also contains a certain morphological contrast. Therefore, whether it is a secondary electron image or a backscattered electron image, it is always a mixture of at least these two contrasts. The effects of other contrasts on the yields of these two electrons will be discussed later.

The respective characteristics of secondary electrons and backscattered electrons in the SEM are summarized in the following table:

Table 15 Comparison of the characteristics of SE and BSE

In many experimental exchanges, it is found that many people tend to understand the secondary electrons and ignore the backscattered electrons. Some SEMs are not even equipped with a backscattered electron detector and therefore cannot capture backscattered electron images. Only understanding the morphology contrast and ignoring the atomic number contrast, and treating the SEM as an instrument with a better resolution than the optical microscope to observe the morphology, are all not enough to understand the SEM and not give full play to its role.

Secondary electrons and backscattered electrons are the two most commonly used signals in SEM, and topography contrast and atomic number contrast are also the most common contrasts in the properties of the sample itself. As can be seen from Table 15, the two electronic signals and the two contrasts complement each other and are indispensable. Only when the relationship between the two electrons and the two contrasts is fully understood, can the SEM be fully utilized and more information can be obtained from the EM images.

For example, as shown in Figure 16, the secondary electron image (left) and backscattered electron image (right) of the fracture of a metal material, combined with the secondary electron and backscattered electron images, can be combined to comprehensively consider the morphology contrast and composition contrast, and the failure of the fracture can be analyzed faster and more accurately.

Fig.16. SE image and BSE image of fracture

Fourth, magnetic contrast

In some specimens, such as magnetic domains in ferromagnetic materials, magnetic fields on video tapes, or thin film wires in integrated circuits, epitaxial magnetic fields can form on the surface of the specimen. The secondary electrons with a certain law will be deflected by the influence of this magnetic field to form a certain contrast, which is the first type of magnetic contrast, which is manifested as a fringe contrast.

The free path of the backscattered electrons in the sample is longer, and the external magnetic field may affect the trajectory of the electrons, and the electrons that bend to the surface of the trajectory are easy to escape, on the contrary, the electrons that bend to the surface are not easy to escape, and the reflection coefficient of the backscattered electrons formed is η, and the resulting contrast is the second type of magnetic contrast.

5. Potential contrast

If there are differences in potential distribution on the surface of the sample, such as the P-N junction of a semiconductor, an integrated circuit with bias, etc., these local potential differences will affect the trajectory and intensity of the secondary electrons. In the positive potential region, the secondary electrons seem to be pulled and it is not easy to escape, so in these regions, the secondary electron yield is less, and the image appears darker. On the contrary, in the negative potential region, the secondary electrons are easily pushed out, the yield is higher, and the image appears brighter, which is the potential contrast. We can use potential contrast to study the process structure of materials and devices. Figure 17 shows an image of an integrated circuit board with both unbiased and biased boards.

Fig.17 Image of the integrated circuit board before and after the bias is applied

However, in the usual observation, there are not many specimens with significant potential contrast. However, with the popularity of deceleration techniques, some of the potential contrast in the deceleration mode may be amplified, allowing the sample to observe a significant potential contrast as well.

In addition, potential contrast plays an important role in the field of failure analysis of semiconductor chips. Some semiconductor chips may have open circuits, short circuits, or other anomalies at specific locations, often not on the surface of the sample. Therefore, in this case, the topography contrast and atomic number contrast of the surface may not play much of a role. However, by scanning the surface of the sample with an electron beam over a period of time, in the case of a defect inside the chip, the potential of the surface position corresponding to the defect may be different from the surrounding one, resulting in a distinct bright or dark spot on the image, as shown in Figure 18. With this method, the location of chip defects can be quickly found, and then the defect location can be further processed and further analyzed.

Fig. 18 Potential contrast caused by semiconductor chip defects

6. Electronic Channel Contrast (ECC)

If the specimen is crystalline, the yield of secondary electrons and backscattered electrons is also related to the relative orientation of the initial electron beam to the crystal plane. The difference in crystal orientation will lead to a difference in the probability of atomic scattering of the initial electron sample, which in turn affects the yield of secondary electrons and backscattered electrons, which is called electron channel contrast, also known as ECC (Electron Channeling Contrast), as shown in Figure 19.

Fig.19 Schematic diagram of the contrast of the electron channel

However, the electron channel contrast is much weaker than the topography contrast and requires a good crystal structure on the specimen surface. Therefore, to observe a distinct electron channel contrast, it is often necessary to remove the effect of topography contrast. Generally speaking, only very flat specimens of polycrystalline materials such as metals and crystals with no excess residual stresses can clearly observe the channel contrast generated by grains with different orientations. To get an ideal ECC image, there are a few things that need to be met:

In particular, the first two items basically need to be prepared according to the standards of EBSD:

(1) The sample must be flat enough to reduce the effect of topography contrast;

(2) The residual stress on the surface of the sample should be reduced as much as possible, and the crystal structure of the surface should be relatively intact;

(3) The channel contrast of the backscattered electrons is usually stronger than that of the secondary electrons, so it is better to observe the ECC using the backscattered electrons;

(4) At larger beam currents, the electron channel contrast is more obvious.

As shown in Figure 20, the electron channel contrast of the nickel-metal material observed with secondary electrons and backscattered electrons under different beam conditions.

Fig.20. ECC of metal Ni under different beam conditions

SEM-ECC technology is suitable for observing the grain size of polycrystalline materials, especially the subcrystalline structure. Lattice distortion occurs near defects in the crystal, and the contrast in these places is different from the surroundings, so that their contrast image can be displayed on the display. However, due to the limited resolution of the SEM, the SEM ECC is not yet able to observe individual dislocations.

7. Density contrast

In addition to the common types of contrast, there is also a less common type of contrast that has been mistaken for channel contrast or atomic number contrast in the past. Some substances, while having exactly the same composition and similar morphology, may be isomerized. Due to the different structures between different isomers, there may be large density differences, which in turn will lead to different scattering probabilities of the initial electron beam in the sample, resulting in differences in the yield of secondary electrons or backscattered electrons. Denser species have higher electron yields and appear brighter grayscales, while less dense species exhibit duller grayscales. This contrast is actually a density contrast.

Fig. 21 Density contrast exhibited by diamond and graphite

As a result, it is sometimes possible for substances with isomers (e.g. carbon materials) to produce this contrast due to different densities. However, to explain this contrast, it is first necessary to ensure that the composition is essentially the same and that there is essentially no effect of topography contrast. As shown in Figure 21, the diamond with a specific gravity of about 3.5 is on the left and graphite with a specific gravity of about 2.3 on the right, and it can be seen that the denser diamond has a higher electron yield.

ZEPTOOLS ZEM18 Benchtop Scanning Electron Microscope

ZEPTOOLS is a scientific instrument company with fully independent intellectual property rights. Since the 1990s, our R&D team has been committed to providing excellent instruments for nanoscience research. At present, the company has a number of product lines including PicoFemto series of in-situ TEM measurement system, in-situ SEM measurement system, ZEM series of benchtop scanning electron microscope, JS series of step profiler, nano-displacement stage, two-dimensional material transfer stage, probe stage and cryogenic system, grating Ruler etc., which have gained a high degree of attention at home and abroad, and filled the gaps of the country's high-end precision instruments in the field of a number of blank.