 |
 |
 导航:老古开发网首页→文章索引→索引第119页 →文章分类→英文资料→第1页→[用于高产量生产的自动微声像技术]
|
| | -文章搜索 - 最新文章 - | |
第7048篇:用于高产量生产的自动微声像技术 |
| 发布时间:2003年5月21日 点击次数:1041 |
| 来源: 作者: |
Internal delaminations and voids are most easily imaged and analyzed by acoustic micro imaging. Traditionally, this imaging has been done off-line as part of a reliability study or as part of failure analysis, and not as part of production. Now, however, the same imaging and analyzing concepts have been developed in automated in-line acoustic systems that rapidly screen trays of packages in order to identify those packages that will not meet the assembler’s reliability requirements. Automated systems from Sonoscan (www.sonoscan.com) are already operating 24 hours a day in numerous facilities to ensure that components moving into production meet future as well as current reliability demands. How Acoustic Micro Imaging Works Although the hardware is different, both laboratory and in-line acoustic microscope systems follow the same principles. Several thousand times a second, two operations take place: the transducer pulses ultrasound into the target and receives the return echoes from the interior of the target. The transducer scans over the surface of the target and at each x,y position data is collected from the depth of interest. When a pulse of ultrasound enters the top surface of a component package, it usually travels first through a homogeneous layer of material - the silicon of a flip chip die, for example, or the molding compound of a plastic IC package. When traveling through a homogeneous material, the ultrasound sends back no return echoes. Penetrating the surface of the target, perhaps 50 nanoseconds later the ultrasonic pulse reaches an interface between two materials. Each of these two materials has a particular density (g/cc) and a particular acoustic velocity (m/s) - the speed at which sound travels through the material. For each material, the density multiplied by the acoustic velocity yields a material characteristic known as acoustic impedance. What matters to the ultrasound pulse at the interface is the difference in acoustic impedance between the two materials. If the difference in acoustic impedances is relatively low, a small portion of the ultrasonic pulse is reflected back to the transducer as an echo signal. The remaining portion of the pulse crosses the interface and travels on to the next interface, where it is again reflected. Well bonded materials produce predictable echoes at an interface. If an IC package has good bonding between materials at all depths, an acoustic image will display those bonds in somewhat different gray scales or in somewhat different colors. Delaminations, voids, cracks and other gap-type defects have a very different type of echo because one of the materials is the air or other gas filling the gap. A gas has a very low density and low acoustic velocity; its acoustic impedance is so different from the acoustic impedance of any solid that nearly 100% of the ultrasonic pulse is reflected by these types of defects. In the acoustic image, delaminations and similar defects stand out in high contrast. Both laboratory and production-line acoustic micro imaging systems take advantage of this high contrast to identify and image internal defects. Production-line systems scan bigger targets (trays rather than single packages), are much faster, and are built to withstand continuous 24-hour operation.  Operation of the Automated System The first thing an observer is likely to notice is a stack of JEDEC-style trays at one end of the automated system. The JEDEC-style trays are designed to hold a specific component type, usually a high-value BGA, CSP or Flip Chip whose reliability must be assured. There are actually two types of trays, those for holding singulated components, and those for holding strips of components whose lead frames have not been cut. One tray at a time is automatically removed from the stack, fed into the system, and imaged. When it exits, the tray is automatically restacked. [Figure 1] On most systems, human attention is needed only to place and remove stacks of trays. A recent modification makes even these steps unnecessary by linking the tray carriers at each end of the system to other machines in the production line. Inside the system itself, the tray is rapidly scanned by the ultrasonic transducer. The scanning process itself has been engineered to achieve the highest scanning and throughput speed and the level of resolution required by the user. For this reason, the transducer does not pause to scan individual components. Instead, it scans the entire tray. Diagnostic software separates out the individual component images and looks for internal defects. If there are gaps between rows of components, scanning can be programmed to skip these gaps, since they contain no useful information.  As the transducer scans the tray, it ignores all of the echo signals except those within a time window defined by the engineer in charge. Echoes from different depths within the components arrive at the transducer at slightly different times. [Figure 2] For most component types, the engineer knows where defects are most likely to occur. In flip chips, for example, the critical depth is the interface between the die face and the underfill. This is the depth where the solder bumps are bonded to the pads on the die face, where solder bumps are most likely to have cracks, and where delaminations between the die face and the underfill occur. Depending on the thickness of the silicon, return echoes from this depth might arrive at the transducer in 80 to 120 nanoseconds, so an electronic gate would be set with these time boundaries. Other package types – BGAs, CSPs, etc. – have different depths of interest and would require different gates. Electronic gates can be narrow or wide, and multiple gates can be set to accommodate more than one depth of interest. The spatial resolution in the acoustic images is important for automated analysis of the components and, when needed, for analysis by an engineer. Resolution is determined by the transducer frequency chosen and by the number of pixels created in scanning. To meet different requirements, resolution can be adjusted over a wide range. As device and interconnect sizes have become smaller, the demand for higher resolution has increased. If an engineer is interested, for example, only in detecting reasonably large voids in a die attach layer, then moderate resolution across the entire tray may be adequate. In a flip chip, the critical defects can be much smaller. A solder bump that is only 200 microns wide can be destroyed by an adjacent void that is itself only 100 microns wide, or even less. High resolution makes it possible to detect even these tiny defects reliably.  Figure 3 is one portion of one high-resolution image, and shows the value of high resolution in such applications. The solder bumps in this flip chip are all intact - there is a white center surrounded by a slightly darker outer ring. But there are also numerous small voids (arrows) in the cured underfill. These voids are highly reflective to ultrasound, and therefore appear very bright. Note that some of these voids are right next to solder bumps. A coommon failure mechanism in flip chips is for the solder to gradually flow into a nearby void until the solder bump itself cracks and the interconnect is broken.  Figure 4 shows a portion of the automated acoustic image of a tray of tiny BGAs. The electronic gate was set to include both the die attach layer and the bond between the molding compound and the substrate. Red areas in figure 3 are delaminations. There are no red areas in the die attach, indicating that die attach processes are under control. But there are numerous delaminations between the molding compound and the substrate. Something – surface contamination, perhaps, or a mismatch of materials – is causing a severe bonding problem Summary Although both laboratory-based and production line-based acoustic micro imaging systems use the same basic principles, production systems are specifically designed to achieve the highest possible throughput at the r esolution level required by the user. The result is that engineers expend little time or labor in removing from the production stream those components that threaten product reliability. Figure captions for Automated Acoustic Imaging Of Components: Figure 1 Automatically taking one tray of components at a time from the stack, an automated acoustic inspection system finds internal defects according to the user’s requirements. Removing defective components increases product reliability. Figure 2 When scanning an entire tray of components, automated acoustic imaging systems are gated on specific depths within the components where defects are likely to occur. Figure 3 High speed is combined with high resolution for components such as this flip chip, where very tiny voids threaten reliability. Figure 4 These chip-scale packages (CSPs) have no defects in the die attach area, but they have numerous delaminations (red) in the bonding of the molding compound to the substrate. Contact information: Sonoscan, Inc. 2149 E. Pratt Blvd. Elk Grove Village IL USA 60007 Phone: 847 437-6400 Fax: 847 437-1550 E-mail: info@sonoscan.com Website: www.sonoscan.com SI_封装_2003.05.22 |
|
|
|
 欢迎进入老古论坛进行讨论
[英文资料] 相关文章: |
|
|
|
下一个:
上一个: