Investigation on the Assembly Process for m03015 and a Brief Look at m0201 Components
Over the years, components have continued to reduce in size. The latest components—m03015 and m0201—are starting to appear in the markets. The m03015 is the metric designation for the EIA 009005 and m0201 is the metric designation for the EIA 008004.
These types of components will be used in module assembly, which would include items for smart wearables where miniaturization is required for higher functional densities. SiPs—the latest term for these modules—are already seen in wearables such as watches, wristbands, and other devices. These parts will not be mainstream for some time because the 01005 components have only been used for a small portion of products today.
Figure 1 shows how the components have reduced in size over the years.
Figure 1: Component sizes timeline.
Test Vehicle
For this testing, Flex's miniaturized test vehicle was used. This board has many features, including package-on-package (PoP), 0.3-mm pitch chip-scale package (CSP), 01005, 0201, high-density spacing, solder flip chip down to 180-um pitch, and others. The bare board can be seen in Figure 2. The pads for m03015 were placed on the boards in anticipation of these parts about five years ago.
Figure 2: Miniaturized test vehicle.
The land pattern used has the following dimensions: 0.15 x 0.15 mm copper pad with a gap of 0.076 mm. This pad will have a toe of approximately 0.038 mm based on the nominal component design. Figure 3 shows the schematic of the pad design. The pad design is slightly larger than different designs seen in other studies. The spacing between the copper depends on the panel location; the pad spacing is 200, 150, and 100 μm. The board is 130 x 77 mm and 1 mm thick with an organic solderability preservative (OSP) surface finish.
Figure 3: Pad design.
This pad was designed to help printability, and would not be the best for miniaturization; this is being done with a new test board that will be discussed briefly at the end of this article. For a 76-µm stencil, the area ratio (AR) is about 0.49; for a 50-µm stencil, the AR is approximately 0.75.
Process Materials and Parameters
For this testing, it was decided to use a 50-µm fine-grain stainless-steel stencil with 150 x 150 µm apertures. The stencil was also nanocoated to help provide the best release possible. A dedicated support fixture was used for the print process. The equipment being used is typically in a standard SMT manufacturing line. The pick-and-place machine had all the necessary upgrades (e.g., camera, software, nozzles) required to move m03015 components. The equipment was verified before running the actual samples. For reflow, we created a typical profile and ran in a nitrogen environment with 200–600ppm of O2 during processing. Figure 4 shows the profile used.
Figure 4: Reflow profile.
The solder paste material used was one that has been used in volume production—the only difference being that a Type 5 particle size was evaluated. This was a low-residue flux system with halogen-free, ROL0 (RO stands for rosin, L for low-activity, and zero for no detectible halides) materials. A Type 4 material was trialed, but the decision was made to use the Type 5 for this study. Figure 5 shows print comparisons.
Figure 5: Type 4 (L) and Type 5 (R) after printing.
Printing Process Data Analysis
A solder paste inspection system was used to analyze the solder paste distribution from the printing process, and an interesting observation was made during this analysis. Figure 6 shows the distribution of all of the m03015 pads during this testing. The data clearly shows that there are two distributions included in the dataset.
Figure 6: Solder paste data for all m03015 pads.
Looking at the data, there are two distinct sets of distributions included. After further analysis, it was found that the pad designs on the test board were done differently. The 200- and 150-µm pad designs were done with the solder mask throughout the array of pads while the 100-µm pad area had solder mask completely cleared from the pad area (Figure 7).
Figure 7: Solder mask differences in 200-, 150-, and 100-µm pad areas (L to R).
The data was separated into the two types of pad designs, and the distributions were redone. Figure 8 shows the new data.
Figure 8: Cpk analysis for each set of pads (solder paste volume).
The Cpk data looked good with 1.71 for the 200- and 150-µm pads and 1.62 for the 100-µm pad area.
Pick-and-place Operations
We then proceeded into the pick-and-place operations. Initial verification was done on double-sided tape. An issue was found where some of the initial placements damaged the component. Figure 9 shows where the component impacted the board with too much force and shattered it.
Figure 9: Damaged component during placement.
Everything was checked, the placement force was set at a minimum, and this was still seen. It turned out to be a software bug where the height the placement head started its deceleration was not set properly. This only happened on the first placement of the board. This has been resolved since this time. Then, the placements on solder paste were then done with 100% visual inspection to verify placements. No defects were found at any of the spacings—just the initial placement as previously seen. Figure 10 shows a sample of the components placed onto solder paste.
The pick-and-place rates were recorded; the pickup rate was 98.9%, while placement rate was 100%. These results were very encouraging because the rates were in the expected areas, and that the equipment could tell if a part was present on the nozzle or not. Most of the latest equipment does this well.
Figure 10: Component placement on solder paste.
Reflow Results
After placing the components on the board, the assembly was sent through reflow. Overall, the soldering with the Type 5 solder paste was acceptable. However, some issues were observed. With the larger pad with a component with terminations on the bottom side only, the part tended to float on top of the solder. Figure 11 shows the results after reflow.
Figure 11: Components floating on the solder.
These results did not match the results seen in other studies, and it has been found that the pad sizes can cause this because there is an excess of solder that forms a larger bump on the pad (semicircle). The component with bottom-side termination can only solder in that area, so the solder bump raises the component; depending on the location of the component, it will be raised up and on the side of the solder bump. If there are three or more sides covered with solderable surface, the part will flatten out. Further studies have also shown with a pad size closer to the terminal size, the part will not raise as much, and there will not be as rounded of a solder joint that would push the part to become tilted.
Another defect seen was some shorting on the 100-µm spacing due to the design of the traces connecting the pads together. The solder would wick down those traces since there was no solder mask covering these areas for the 100-µm spacing area. The tilting of the component also caused issues with AOI and AXI not being able to detect the parts and determine defects. A separate study covers more on those evaluations.
Further Work
Additional work is in progress, and a new test vehicle has been designed to include m0201 components. Initial printing and placement have been done, and additional trials are scheduled to be done. The pad designs on this SiP test vehicle are designed to the same size as the terminations on the m03015 and m0201 components. The spacings on this board have been reduced as well. The board contains component spacing of 100, 75, and 50 µm, which would be needed for SiP products.
Figure 12 shows some of the initial results of m0201 at 50-µm spacing. The print process still needs to be dialed in further because the printing is inconsistent with many high-volume spikes, which also caused the bridging in Figure 12, and the insufficient and no solder results. More trials are underway with different types of solder pastes to enable better printability. The placement equipment was able to pick and place even smaller components with little issues seen. New software and nozzles were required for the m0201 component on the equipment that was being used.
Figure 12: m0201 after reflow at 50-µm spacing.
Conclusion
This article demonstrates that smaller components can be done in an SMT process with standard equipment with the latest cameras, software, and nozzles with high yield. Printing was shown to be possible with a high Cpk. Pick-and-place rates were at a high level, and a minimum number of defects were seen. The pad size in this study was slightly larger, so there is room for more reduction to enhance miniaturization. This will be seen on the ongoing work with m0201 and m03015 for SIP products.
Acknowledgments
The authors would like to thank everyone involved with this project from inside and outside of our company.
Further Reading
1. N. Heilmann. "03015 Information." ASM Productronica, 2013.
2. D. Geiger, A. Mohammed, M. Kurwa, AEG, & Flextronics International Inc. "Overview Miniaturization on Large Form Factor PCBA." IPC APEX EXPO 2016.
Editor's Note: This article was presented at the IPC APEX EXPO 2018 Technical Conference.
