The continuing reduction in the feature sizes of electronics creates many difficulties in the analysis of problems. Determining the composition of small areas of contamination or the distribution of an element requires high spatial and elemental resolution. A scanning electron microscope (SEM) with energy dispersive X-ray (EDX) analysis is the standard method of examination but it has limitations.

In EDX analysis, the minimum electron beam energy required to excite the X-rays providing the elemental information is between 5 to 20 KV. When this energetic beam strikes the surface of the sample at the point of interest, it is scattered and generates X-rays from a small volume with an approximately spherical shape at least a micron across. This finite size and the closeness of certain X-ray lines creates difficulties in interpretation.

ERA uses the “INCA” data analysis system from Oxford Instruments which provides powerful capabilities that help in these situations. It has the standard facilities of automatic identification of the elements from the peak positions and heights, and the ability to generate maps of the distributions of selected elements within the field of view. In addition there are two methods of visualising the distribution which aid interpretation.

Feature-specific composition analysis
- example: application to pcb interfacial defects

The composition at any point or defined rectangle can be determined in the standard manner, but in addition the area over which the identification is made can be set at any shape with a fixed intensity in the backscattered image. The shape can have features down to the resolution of the backscattered image which is far finer than that of the EDX spot size. This makes it possible to assess the spread into areas where it is difficult to be certain of its position from the EDX data alone. An example is shown in Figure 1.

Figure 1 Part of a cross section of a badly damaged board, where the numbered segments contain copper

The purple lines mark the position of copper that has spread through a printed circuit board by means of Conductive Anodic Filaments. These filaments form within contamination at the interface between the glass fibres and the matrix of printed circuit boards under certain circumstances. Copper is driven along the interface and can cause short circuits in the board. The traces can be very narrow and might not appear continuous on a standard EDX map, but by linking and analysing areas with a common intensity, the distribution is shown on a far finer scale.

False colour imaging
- example: application to plating defects

Another useful method of display is called “Cameo”. A false colour image is generated from the EDX maps by assigning the visible spectrum to a suitable band of X-ray energies. An example of a cross section of a gold layer over nickel on copper is shown in Figure 2. The spectral range has been chosen so that gold appears as yellow, the nickel blue and the copper green. The way the nickel (blue) has penetrated into the gold is clearly shown. The faint green areas within the yellow show where there is a mixture of gold and nickel. The wire bonds to semiconductor devices were not reliable as they were partially made onto a nickel surface rather than the intended gold.

Figure 2 False colour “Cameo” view applied to gold plating on nickel on copper. [Yellow=gold, green=copper, blue=nickel]

While these methods do not increase the actual spatial resolution, which is limited by the physics of the interaction of the electron beam with the sample and the emergence of X-rays, they do enable an improved appreciation of the distribution and variation of composition. This information is often key to establishing the nature of a fault and in turn to reaching the right conclusions regarding the root cause.

Further information
ERA has been using SEM and EDX for decades in the analysis of problems in electronic systems. Considerable experience has been gained in a wide variety of failure situations. To find out more call us on +44 (0)1372 367444 or email.