Optimum Conditions for an Observation of Replicas of Metallographic 
Specimens in the Scanning Electron Microscope 
Jan Hejna, Institute of Material Science and Applied Mechanics, Wroclaw University of 
Technology, Wroctaw, Poland 
Introduction. 
Replica method is one of basic techniques of a nondestructive testing of power plant components (1, 
2) and it is also useful for an inspection of any large element that cannot be directly placed in the 
i} chamber of the scanning electron microscope (SEM). A place of interest on the component is 
Seas grinded, polished, etched and a replica of the surface in a thermoplastic foil is made. The 
impression on this one-stage replica is a negative copy of the relief on the specimen but because of 
2, RS; Holmes, simplicity of preparation usually such replicas are used for observations. Before an observation in 
On Bchvee 2 the SEM the replica needs only to be coated with a conductive metal layer by sputtering or 
4 evaporation in a vacuum coater. Better representation of the surface topography can be obtained 
= rap. when a two-stage (positive) replica is made. In this case the plastic replica taken from the specimen 
MOL S (1994) © is covered with a relatively thick (= 100 nm) metal layer and next with an epoxy resin. The epoxy 
after hardening is separated from the plastic replica - the metal layer is glued to the epoxy and is 
k Strength. els. detached with it from the foil. To facilitate removal of the positive replica the plastic foil can be 
085 of ses, | covered with a carbon layer prior to deposition of the metal layer. The surface relief of the epoxy 
covered with the metal layer is a positive copy of the specimen topography but the preparation of 
Che the positive replica is more complex and there is more chance for introducing artifacts that can lead 
us ds DVL to false interpretation of the specimen structure. An appearance of the surface relief in an image 
at stoffen 24. obtained in the SEM depends on detectors used for observations and on conditions of their 
a operation. Using different detectors we can obtain different images with different appearance of the 
EEE topography, so it is important to find optimum conditions for an observation of the surface relief. 
he Stallforth i Detection of topographic contrast in the SEM. 
em Ne York, For obtaining topographic contrast in the SEM secondary electrons (SE) and backscattered electrons 
41-6), 5. 1250 (BSE) are used. In order to form an image electrons have to be collected by appropriate detectors. 
‚weitungsverhal- SE have low energy and a majority of them can be easily attracted by the detector biased with a 
fung 1999 „Fort- positive potential of several hundreds volts — in result SE images are not very sensitive to the 
Sichem 87, 12205 detector geometry. BSE have high energy and are collected by the detector if they are emitted in its 
direction — in result BSE images depend very much on the geometry of the detector. At high take- 
Stents. in Stalk off angles of BSE (in respect to the specimen surface) material contrast dominates in the BSE 
rs, New York, signal, at low take-off angles topographic contrast dominates. Every SEM is equipped with a 
detector of SE, part of them is equipped with BSE detectors that are used mainly for detection of 
material contrast but some of them can be also used for detection of topographic contrast. Figure 1 
shows some detectors that are used for detection of topographic contrast. SE are detected with the 
Everhart-Thomley detector (3) (Fig. 1a), it consists of a scintillator coupled by a light-guide with a 
photomultiplier. Such detector not only collects SE emitted from the specimen surface (SEI 
generated by the primary beam and SE2 generated by BSE escaping from the specimen) but also SE 
generated by BSE in a specimen chamber (SE3), mainly at a lens pole-piece. A simple way of 
indirect detection of BSE is collection of SE3 by the E-T detector. In this case SE from the 
specimen should be suppressed in the signal — it can be done by placing a grid biased with a 
negative potential over the specimen (4) (Fig. 1b) or by biasing the specimen with a positive 
potential (5) (Fig. 1c). The drawback of a conversion is a fact that we do not know exactly an origin 
of SE3 and we cannot control the process of detection. Some microscopes are equipped with 
semiconductor detectors in an annular-split form (6) (Fig. 1d), when the difference of signals of 
347