Publication date: 26 August 2010
Package On Package PoP assembly is new to many engineers. PoP introduces different processes, materials and challenges, which will undoubtedly highlight some process defects. The trick with any new production process is to gain as much hands on practical experience as possible. It’s great fun seeing engineers presented with a PCB, PoP components, tub of dip paste and rubber gloves!! Much better than looking at a computer screen all day and a better learning experience? Talking of education in cooperation with Electronics Production World we are running online seminars on PoP Technology along with many other webinars, details are available on the next PoP seminar by clicking here
When you look at a PoP x-ray for the first time don’t be surprised, its confusing but this example is perfectly normal and represents a satisfactory assembly. The high resolution view shows two parallel rows of balls on two layers of a PoP assembly. The larger balls are on the topside component and the smaller balls on the bottom side, close to the printed board. As with any high resolution x-ray system it is possible to see the pad to ball interface visible in the joint area indicating satisfactory reflow and connection.
Microsection image taken of an area array solder joint termination. The image shows the bottom of the ball to copper pad interface, a small void cavity is present in the solder connection and just below the pad is a crack in the PCB resin. This is a common fault seen on boards which have been exposed to drop testing, flexture or vibration. The crack is through the resin in the interface next to the copper circuitry. It has become a more common defect on high Tg laminates and with the increased rigidity of lead-free terminations
Solder paste dipping is a common process used in PoP assembly. The second level package is dipped on to a flat plate of specially formulated dip solder paste during the placement cycle. In this case the application plate depth control has been ineffective. When dipping the depth of paste on the plate must be controlled accurately otherwise excess or insufficient paste application will occur. In the example there is a paste short between four terminations and one ball has a limited deposit
The x-ray image shows a four stack PoP assembly with columns of solder balls visible on the same pitch. In addition the image shows chip component terminations with some evidence of voiding on the surface of the printed board. There is also some evidence of solder balling on the left of the image, however it’s difficult to say on which layer of the PoP assembly the random balls are located. Dip paste has a lower metal content, often designed for nitrogen reflow, and has a smaller solder particle size than stencil printing grades hence more prone to solder balling and slump during reflow.
One of the reasons solder paste is used on PoP assembly as opposed to dip flux is coplanarity of the ball terminations. Possible warping of the components during reflow can cause the balls to be separated from the surface of the matting pads resulting in an open connection as shown. Dip flux is less forgiving in filling the separation gap whereas dip paste can bridge the gap between the two surfaces when reflowed. The JEDEC specification for PoP packages can provide an indication of the ball to ball coplanarity of the package but not the potential variation across the packages during reflow. Variations will be dependent on the package construction, soldering parameters and the base PCB flatness.
Just like dip solder paste on a rotary or linear application plate, the surface and the depth control of the material must be controlled accurately to avoid excess or missed application of material to the terminations. The image shows an uneven flux layer on a rotary plate, as the doctor blade has passed over the plate there are groves in the surface. This can be an indication of material drying on the doctor blade, the flux surface is staring to dry, limited blade sweep is being used or damage to the blade surface has occured. Regular checks on the flux plate are important to eliminate the possibility of skipped flux application.
The image shows the bottom of a PoP device prior to placement onto the first package, the liquid flux should only be on the surface of the balls. In this case the part has been drowned with flux covering all of the balls and the surface of the package. Excess application of flux can result in component float during reflow, depending on the solvent in the flux the part could even jump off the board during reflow. Excess flux should be avoided, it shows poor control, may cause problems during cleaning and will impact successful underfill if conducted.
In process inspection of a PoP assembly is shown after the second level component has been placed. The flux is visible at the ball to pad interface prior to reflow. With the blue colour the level of the original flux dip height is discernable as being approximately 50% of the ball. This is considered to be satisfactory and should result in a reliable connection after reflow.
Sideways view of a PoP assembly showing one row of terminations after underfilling. The black underfill can be seen between the top and bottom of adjacent packages. The underfill is present but has failed to fill the cavity and cover all of the interconnections. Currently there are no international standards for underfilling and it’s probably not known if this incomplete underfill would have any impact on the reliability for either temperature cycling or drop testing. It is however, good practice to have a standard for reference and an in process control. References for PoP assembly inspection are available at SolderingStandards.com
Underfilling is common practice on portable products where large area array packages are used in combination with thin substrates. This practice is to provide, in most cases mechanical strength to the interconnecting surfaces rather than thermal expansion relief. It has been good practice over the years to apply underfill to the edge of parts and allow capillary action to suck the material under the parts to fill all the void areas under the packages and between the balls. Due to the decreased space on portable products it is often the case that materials are now applied in larger quantity to adjacent area array devices and allowed to fill two parts at the same time. The result is other chip parts being coated with underfill, in reality is this an issue of a poorly controlled process or a well engineered application?
As previously stated dip solder paste has a lower metal content, more often designed for nitrogen reflow and also has a smaller solder particle size, type 5-6 powder as opposed to a stencil printing grade of type 3-4, hence more prone to solder balling during reflow. It’s important for engineers to go back to school and learn how to conduct solder balling and solder slump measurements of paste from different vendors.
Bob Willis is a process engineer working in the electronics industry, providing training, consultancy and process/product failure analysis. Bob will be running workshops at SMTA International and APEX 2011 covering PoP, Conformal Coating Application and Counterfeit Component Inspection. Bob offers on site workshops on conventional and lead-free manufacture, he is also happy to assist you set-up, fault find and optimise production lines for users and also provides conferences and workshops worldwide www.bobwillis.co.uk
