Acoustic Detection of Package Contamination

A pulse of ultrasound launched into a sample by the transducer of an acoustic microscope sends back return echo signals only from material interfaces — not from homogeneous materials. If the sample is a plastic encapsulated microcircuit (PEM), the first internal interface encountered may be the interface between the molding compound and die face.

How the ultrasonic pulse interacts with this interface depends on the acoustic impedance (material density × acoustic velocity) of each of the two materials at the boundary. If the molding compound is firmly bonded to the die face, the difference in acoustic impedance between the two materials will be moderate, and a return echo signal of moderate amplitude will be generated when ultrasound strikes the interface. The single image pixel made from this echo signal will probably appear gray (or some other mid-scale color) in the acoustic image. If the entire die face is well bonded, the color of the die face will be uniform.

A delamination between the molding compound and die face drastically changes the behavior of the pulse. A delamination is generally considered to contain air or another gas, so the first interface is now between the molding compound and a gas. The gas will have a vastly lower density and smaller acoustic velocity than a solid. The acoustic impedance of molding compounds, metals and all other solids is measured in units called megaRayls, but the acoustic impedance of air and other gases is around 410 Rayls — a Rayl being one-millionth of a megaRayl.

The relatively enormous difference in acoustic impedance between the molding compound and air in the delamination means that virtually all of the ultrasound is reflected as an echo. The fact that the acoustic impedance of air is less than that of plastic means that the echo polarity/phase is shifted by 180°, resulting in an inverted (or negative) echo. The amplitude of this echo is very high; on many of the color maps used on acoustic microscopes, this high-amplitude inverted echo will be shown as a red pixel. Because delaminations and other gap-type anomalies have high-amplitude reflections, it has become a more or less accepted convention to display the defects in red.

Over the past several months, numerous PEM samples imaged at our applications laboratories, SonoLab, have generated unusual images in which only part of the delamination is red. The remainder of what appears to be an ordinary delamination is seen as white in the acoustic image. In the color map being used, red indicates a reflection that is both high in amplitude and negative in its polarity -- negative meaning that the pulse traveled from a material of higher acoustic impedance to a material of lower acoustic impedance. In the same color map, white indicates a reflection that is both high in amplitude and positive in polarity, usually meaning that the pulse has traveled from a material of lower acoustic impedance to a material of higher acoustic impedance. The initial question confronting technicians whose job it is to interpret the acoustic image was "How can a delamination between two solid materials possess both positive and negative polarities when, in the case of a possible delamination, the acoustic impedance is unlikely to be greater than that of the molding compound?"

1. SEM of a sectioned sample showing the delamination between the particle-filled molding compound (top) and the die face (bottom). (Source: Riga Analytical Lab Inc.)

The red/white delamination images were obtained during routine imaging, where the return echoes were gated on the delamination. Gating in this case means that a time window was set up to use return echoes from a depth just above the top surface of the die to a depth just below the top surface of the die. Such gating is a standard procedure used to exclude features from other depths from the acoustic image.

An obvious potential contaminant was moisture that could have entered into the delamination. If the gap were filled with water, it might produce the unusual images that had been observed. Sample packages were therefore baked at 125°C for 24 hr, but the red/white appearance of the delaminations persisted. Whatever the contaminant was, it had not been removed by baking.

2. Microdroplets were found on the inner surface of the molding compound after dry separation. (Source: Riga Analytical Lab Inc.)

One delaminated sample was cross-sectioned through the center and lightly etched for SEM imaging (Fig. 1 ). Imaging showed that the delamination between the molding compound and die face was as little as 0.6 μm thick. Gas-filled delaminations and other gaps down to 0.01 μm uniformly reflect virtually all of an ultrasonic pulse, so the thickness of this delamination was not significant. The cross-section did not reveal the presence of a contaminant, because the contaminant was most likely removed during preparation of the sample.

Another delaminated package was disassembled by a dry separation technique to give access to the die surface and inner surface of the molding compound. An examination of the inner surface of the removed molding compound revealed micro-droplets of a colorless fluid (Fig. 2 ). Because there was no evaporation of this fluid after some hours of exposure at room temperature, it was concluded that this fluid was not water alone. The plastic surface was then examined by energy dispersive X-ray (EDX) analysis. In the vacuum chamber, the fluid evaporated and phosphorus was detected on the surface of the plastic. Since phosphorus is hygroscopic, its presence may have prevented the water from initially evaporating. The source of the phosphorus is unknown, as it is rarely used in molding compounds.

3. Surface of the die after dry decapping. Small irregular areas are the residue of a contaminant. Auger electron spectroscopy showed the presence of silicon, carbon and oxygen, suggesting that the contaminant may have been a silicone. (Source: Riga Analytical Lab Inc.)

This analysis does not prove conclusively that the contaminant is silicone, or that its source is the releasing agent, but it does provide an answer that fits very well with the acoustic data. The acoustic pulse used in imaging encounters, in order, the plastic molding compound, which has a lower acoustic impedance of ~4.5 megaRayls; the organic fluid, which, if it is silicone, has an acoustic impedance of ~1.5 megaRayls; and the silicon of the die, which has a much higher acoustic impedance of ~20 megaRayls.

4. Acoustic image made at the depth of the top surface of the die. Red area (1) is an air-filled delamination that stops further penetration of ultrasound and gives a high-amplitude negative echo. The white area (2) that partly surrounds the delamination results from a distorted echo created by a thin layer of fluid contaminant.

The research carried out on these samples demonstrates that delaminations in plastic encapsulated microcircuits may harbor contaminants that produce an otherwise unexpected acoustic signature. The acoustic microscope operator who encounters a "red/white" delamination may be able to at least tentatively identify the contaminant as water if it disappears after baking, and tentatively identify it as an unknown fluid if it does not disappear. The operator can also help determine whether a feature is a contaminant by looking at the echo waveform and noting the distortion resulting from the combination of positive and negative attributes. This ability to more finely discriminate the conditions at a delamination may provide clues for manufacturing engineers on how to more accurately monitor and improve manufacturing processes.