Lecture notes 4

Scanning Electron Microscopy (II) General Layout

In simplest terms, an SEM is really nothing more than a television. We use a filament to get electrons, magnets to move them around, and a detector acts like a camera to produce an image.

EE206

1

Lecture notes 4

SEM Instrumentation

Figure: Schematic of a scanning electron microscope.

EE206

2

Lecture notes 4

Detectors Secondary electrons

1. Secondary electrons are accelerated to the front of the detector by a bias

voltage of 100 - 500V. 2. They are then accelerated to the scintillator by a bias of 6 - 12 kV, (10 KV is normal). 3. Scintillator is doped plastic or glass covered with a fluorescent material (for example: Europium). High-energy electrons hit the fluorescent material and light is generated. 4. The light photons travel down the tube to a photocathode, which converts them into electrons 5. The electrons move through the detector, producing more electrons as they strike dynodes. An amplified output electron pulse is then detected. Back-scattered electrons Since BSE have high energies, they cannot be pulled in like secondary electrons since if you placed a potential on a grid to attract them, you would also attract the incident beam! The most common detector used for BSE is called a solid-state diode detector. It sits above the sample below the objective lens. BSE, which strikes the detector, is detected.

EE206

3

Lecture notes 4

Image formation and interpretation The SEM image is formed in a different manner compared with optical microscope. That is the meaning of “scanning” all about. Look at the following schematic, which illustrates the scanning system of the SEM. All signals from electron beam-specimen interaction can be detected by certain detector, and amplified to control the brightness of a cathode ray tube (CRT) scanned in synchronism with the sample beam scan in the SEM. A one-toone correspondence is then established between each point on the display and each point on the sample. Magnification is then defined as:

M = L/l CRT normally has a width of 10 cm. So you can basically obtain magnification if you select the scan range on your sample. Example: If you scan the electron beam within 10 µm, what would be the magnification time that SEM would get?

EE206

4

Lecture notes 4

Figure Schematic illustration of the scanning system of the SEM. Abbreviations: FA, final aperture; SD, solid-state backscattered-electron detector; EDS, energy-dispersive x-ray spectrometer; WDS, wavelength-dispersive x-ray spectrometer; CRTs, cathode ray tubes, and E-T, Everhart-Thornley secondary/backscattered –electron detector, consisting of F, Faraday cage; S, scintillator; LG, light guide; and PM, photomultiplier. Successive beam positions are indicated by the numbered rays of a scanning sequence.

Figure The Principle of image display by area scanning. A correspondence is established between a set of locations on the specimen and one on the CRT. Magnification = L/l.

EE206

5

Lecture notes 4

Resolution The ultimate resolution is the smallest separation of two points, which the microscope can detect as being separate entities, is determined largely by the diameter of the beam of electrons, which is scanning the specimen. The following figure illustrates the concept of resolution in SEM. On the left is a sketch showing the path of the electron beam across the specimen. Each scanning line of the raster is adjacent to the previous line. Only three lines are shown, as they pass over a region of the specimen containing five small features. On the right are shown the equivalent three lines of the cathode-ray tube display. Here the scale is much grosser, the spot is about 0.1 mm in diameter, so we can see it easily on the screen. As the electron beam pass each features A, C, and E on the specimen the detector receives an increased number of secondary electrons and therefore the spot on the CRT becomes brighter at these points (we have shown it darker in the diagram for clarity). On its next pass the electron beam only touches D and E and thus two bright dots appear on the CRT. On the third pass only B is detected. In terms of resolution we can interpret this in the following way. Features A and B were farther apart than the diameter of the electron beam and gave rise to two distinct dots on the display of CRT. Features C and D, however, which were only separated by a distance equal to or less than the diameter of the beam give rise to a slightly elongated dot which cannot be distinguished from the dot representing the feature E. Points C and D are therefore not resolved and we can say that the geometrical resolution limit is the diameter of the electron beam scanning on the sample surface.

Figure The concept of resolution in the scanning electron microscope. The small features A and B are resolved, since their separation is greater than the electron beam diameter, whereas the two features C and D cannot be distinguished in the image from the single larger feature E.

EE206

6

Lecture notes 4

Depth of field As we already pointed out that the large depth of field can be obtained in SEM and is one of their most striking capabilities.

The depth of field criterion depends on where the beam reaches a condition of overlapping adjacent pixels, in the above figure, the vertical distance D/2 required to broaden the beam of minimum size to a radius r is given by tan α =

r D/2

for small angles,

tan α = α

Thus, D / 2 ≈ r / α ⇒ D ≈ 2r / α Consider that r equals the pixel size of the image. On a high-resolutions CRT(spot size =0.1mm=100µm) most observers find that defocusing becomes objectionable when two pixels are fully overlapped. The pixel size on the specimen is then given by r=0.1/M mm, therefore depth of field is D≈

0 .2 mm αM

Example: for M=1000, α=5×10-3 rad, D ≈ 40µm At this point, we have known the principle of a topographic image formation from an SEM.

EE206

7

Lecture notes 4

SEM specimen Preparation for nanoscale characterizations One of the great strength of SEM is the fact that many specimens can be examined with virtually no specimen preparation. For the examination of images of topographic contrast, the only specimen preparation necessary is to ensure that the specimen is thoroughly degreased so as to avoid hydrocarbon contamination and, in the case of insulators, to provide a conductive coating. Sample preparation for semiconductor nanostructures, nano-electronic devices, etc: 1) cut to fit specimen holder, 2) degrease (acetone ultrasonic rinse, then methanol wash), 3) mount onto a specimen stub either mechanically with a clamp or with conductive paint, conductive double-sticky tape etc. 4) Dry in an oven (75C). Finally maybe, coat a conductive layer immediately before the introduction of the sample to SEM chamber. Conductive Coating: is necessary to eliminate or reduce the electric charge that builds up rapidly in a nonconductive specimen when it is scanned by a beam of high-energy electrons. See figure for charging phenomena. In addition to charging phenomena, which result in image distortion (incoming electrons are repelled and deviated from their normal path), the primary beam also causes thermal and radiation damage, which can lead to a significant loss of material from the specimen.

Figure Charging that could be mistaken as a specimen feature. (a) The polymer spheres seem to show a complicated interior structure; E0 = 2.4 keV. (b) At E0=1.53 keV, the structure has changed into a pupil, making each sphere look like an eye. (c) At E0 =0.87 keV and a reduction in magnification of a factor of two to reduce the dose per unit area, the charging artifact has been eliminated.

EE206

8

Lecture notes 4

Thin film coating methods (2 typical methods) 1) Thermal evaporation method

Figure Diagram of a high-vacuum thermal evaporation unit.

EE206

9

Lecture notes 4

2) Sputtering a) Diode or Direct-current sputtering b) Plasma-Magnetron Sputtering c) Ion-Beam Sputtering d) Penning sputtering Example: Plasma-Magnetron Sputtering

Figure

Diagram of a plasma-magnetron sputter coater.

Figure a plasma-magnetron sputter coater.

EE206

10

Lecture notes 4

Choice of coating materials:

Figure Cleaned diatom frustules coated with 10nm of various metals and examined at 20 keV; (a) gold, (b) aluminum, (c) copper, (d) silver, (e) chromium, (f) goldpalladium, (g) titanium, (h) tin, (i) calcium. The degree of granularity from each material would be used for researcher to choose the ultimate coating material.

EE206

11

Lecture notes 4

Choices of thickness of the coated material:

Figure Cleaned diatom frustules coated by evaporation, and increasing thickness of gold examined using secondary electrons at 30keV: (a) 2.5 nm, (b) 5.0 nm, (c) 10 nm, (d) 20nm, (e) 50nm, (f) 100nm. Note that the very thin layers give optimal information, whereas the thicker layers obscure the surface of the specimen.

EE206

12

Lecture notes 4

Summary for SEM specimen conductive coating 1. For high-resolution SEM (spatial resolution 1-3 nm), where one is attempting to push the instrumental capabilities to their limit, a mixture of Pt-Ir and C would appear to be suitable for most morphological problems. For high-resolution SEM, where one can be sure to be primarily collecting only the secondary electrons, then platinum, chromium, or niobium can also be used. In all instances, a film thickness of 1-2 nm appears to give the best result. 2. For medium-resolution SEM (spatial resolutions 3-5 nm), either goldpalladium or platinum-iridium gives the best result. The final film thickness is normally in the range of 3-8 nm. 3. For routine SEM on coated specimens (spatial resolutions greater than 810 nm), Au-Pd or Pt-Ir gives good results. The final film thickness can be in the range 8-12 nm.

Example: Pt-Ir coating seems very good for SEM purpose; do you also use this for EDS? If so, why? If not, what would be the material you use for EDS?

EE206

13