Circuits Assembly Online Magazine - Mass Reflow Assembly of 01005 Resistors

News

Tombstone ratios of co-assembled 0201s and 01005s show the latter are more stable across a variety of atmospheres.

Assembly process concerns between 0201 and 01005 are similar; however, the common expectation is the smaller 01005 should be more sensitive to process and design variables. Several articles published on 01005 assembly commonly recommended fine powder solder paste (Type 4 or 5) in conjunction with a thin stencil (0.003").^1,2,3 This work challenges two previous recommendations and attempts to prove or disprove that 01005 assembly process can be performed with Type 3 powder size standard solder paste material using a 0.004"-thick stencil.

Process Design

Several experiment variables have been tested to determine sensitivity to 01005-resistor assembly yield, described in the following categories:

PWB. The double-sided 5.5" x 8.0" x 0.062"FR-4 PCB was designed to accommodate 01005 components on side A and 01005 and 0201 components on side B. Only copper OSP pads were considered. All pads were non-solder mask defined and without mask inside the component outline. The majority of testing was conducted on side A, which comprised 27 combinations of pad design, four levels side-to-side spacing, and two orientations. Table 1 identifies pad designs on side A.

Side B consisted of fixed pad dimensions for both 0201 and 01005 components. Table 2 specifies land details.

Stencil. The assembly tests used three stainless steel laser-cut stencils. All foils accommodated mounting on a 23"x 23" frameless stencil system. Table 3 lists their differences.

The 01005 aperture sizes designed for Stencil 1 were determined by applying pad to aperture size scaling from previous successful experience with 0201 assembly.⁴ Table 4 shows equations used to generate aperture sizes on Stencil 1. Each of the nine pad length and width combinations has both standard size apertures and overprint apertures. The solder paste is intended to print inside the pad boundaries using the “standard” aperture design, while the “oversize” aperture means the solder prints outside the pad borders.

Two additional designed features of Stencil 1 include “grab” testing and print offsets. Grab is defined as the overlap between passive component termination and printed solder paste. The stencil has been designed to produce three grab levels per pad design (Figure 1). Apertures positioned closest together will have the most solder residing under the component and will be considered the highest grab level, indicated by “H.” Note that the pad separation dimension (S) on the board also has a significant influence on the grab factor.

While the majority of the Stencil 1 apertures are designed to be positioned symmetrically about the pad sets, some pads are designated to receive paste that has intentional print misregistration. Two levels of diagonally offset apertures at 0.0005" and 0.001" increments were used to test assembly yield tolerance of print deposit position across pads.

Stencil 2 was designed to investigate 01005 resistor assembly yield tolerance to gradually reduced aperture sizes across all pad designs. All apertures were centered on the pads for these tests. The largest apertures for any pad were set at a 1:1 design. From here, aperture width was reduced in 0.0005" increments if the corresponding pad width exceeded length. If pad length exceeded width then aperture length was reduced in 0.0005" increments instead. The pad length to width ratio was used as the scale factor to determine the remaining aperture dimension so that all decrementing aperture sizes would have the same aspect ratio with respect to the pad design. Matching length and width would both be reduced by 0.0005" increments.

Stencil 3 is the only stencil that prints on PCB side B and contains aperture designs for 0201 and 01005 components. There were three 0201 aperture designs, the smallest being 1:1 with the pad and incrementing 0.001" larger per side on the two other designs. There were eight different 01005 apertures sizes, three of which were square shape from 0.008" to 0.010". The remaining five 01005 apertures were rectangles ranging between 0.007" and 0.011" per side.

Solder paste. Three standard commercially available solder pastes were used in the assembly experiments (Table 5). All were Type 3 powder size and no-clean flux paste formulations. Two SAC 305 materials were tested. The eutectic SnPb paste was the same formulation used to investigate the 0201 assembly process.⁵

Printing. A high-speed automatic stencil printer with programmable fiducial location was used to deposit solder paste onto the PCBs. Printing parameters (Table 6) were fixed for all builds. Settings were based on recommendations described in an earlier report on 01005 printing.⁶

Placement. A standard four-spindle automatic SMT placement machine outfitted with prototype nozzles was used to assemble the 01005 resistors. The nozzle design schematic is shown in Figure 2. A 0.0004" per pixel magnification upward-looking camera was installed to inspect for component presence and location on the nozzles. No modifications were made to the electronic feeder to index the components. Default machine settings were used to operate a normal pick-and-place sequence.

Reflow. An eight-heating zone forced hot air convection furnace with board edge supporting adjustable conveyer rails was used to reflow assembled boards. The system is compatible with achieving elevated temperatures required for Pb-free assembly and plumbed for N2 capability. For all assemblies reflowed in N2, the atmosphere consistently records O2 levels below 50 ppm in the final heat zone. Convection was set to the maximum 3500 rpm fan speed, and top/bottom zone temperature settings were balanced. Temperature profiles were programmed to match the paste vendors’ recipes listed in the material technical data sheets (Figure 3).

Assembly Tests

The 01005 components used for assembly were resistors. The termination metal on these passives is Sn100. Table 7 shows averaged dimensions from a sample of measured components. Note this component more accurately resembles a metric 0402 rather than a true “01005” size.

A total of 12 boards were assembled. Builds are grouped according to the stencil used and listed with the specific experimental condition details in Tables 8 to 10.

Stencil 1 Builds. Builds 1 and 2: The strategy for these builds was to determine if the 01005 resistor assembly process would show sensitivity to pad design, stencil induced grab, and standard vs. oversize print deposit. Nitrogen was present in reflow to promote wetting and enhance tombstoning occurrence. Not all pad designs were assembled on the 0.006" pad-to-pad spacing as it was considered impractical to assemble components onto sites printed with aperture area ratio below 0.5.7 Component orientation was selected at 90° only so that one termination would follow the other into reflow, which was found to increase tombstoning in earlier 0201 assembly testing.^4,5 The difference between builds 1 and 2 is the solder alloy and reflow profile. Under inspection, reflowed boards from both builds showed no bridging, very low occurrence of solder balls, and no open solder joints. This introduces evidence that a wider-than-expected pad and stencil design process window may exist to achieve high-yield 01005 resistor assembly using SAC 305 and eutectic SnPb. Several pad designs printed with the “oversize” apertures and assembled showed instances of floating components (Figure 4). Figure 5 shows pad designs with excess solder delivery by oversized apertures.

Inspection. Builds 3 and 4: More challenging component placement locations were programmed in this run to encourage assembly yield loss. Nitrogen was again active in reflow to promote wetting. One board was printed with SnPb37 and the other with SAC 305. “Oversize” apertures were assembled on 0.010" side-to-side pad spacing, while “standard” aperture designs were assembled everywhere else. The finest side-to-side pad spacing of 0.004" was populated at 90° for all three stencil grab levels wherever the print area ratio exceeded 0.5 (Figure 6). Two tombstoned components were located on pad CFI at the lowest stencil designed grab level for Build 3, corresponding to the eutectic SnPb assembly. The Pb-free assembled Build 4 board showed no defects. There were no bridging defects observed on both builds at this component placement density.

Figure 7 shows a comparison between leaded and non-leaded solder paste assembly on pad design BEG using the largest paste grab “standard” aperture design at 0.004" pad spacing. Note the significant difference in wetting between the two alloys. Even with a high concentration of nitrogen present in reflow, the Pb-free solder paste tends to form bulging solder joints at the terminations due to reduced wetting.

Pad sites were also added to the assembly list that printed with intentional diagonal print misregistration of 0.0005" and 0.001" at 0.006" pad side-to-side spacing. Pad designs were not assembled for print area ratios below 0.5. All three stencil grab levels were assembled on print area ratio-qualified pads. The eutectic SnPb Build 3 showed 10 tombstone defects, while the Pb-free Build 4 showed none. One tombstone defect was observed for pad width “B” and the remaining nine occurred in pad width “C”. These defect locations are shown in the board map (Figure 8).

Figure 9 shows a pad design BEI with simulated 0.001" diagonal offset print and accurate 01005 placement. The west positioned paste deposit barely contacts the left termination, while the east positioned paste deposit still overlaps the right termination. The grab imbalance caused by the print misregistration at highest pad separation “I” and lowest designed stencil grab level creates a condition that favors paste climbing the right termination during reflow and pulling the component to the tombstoned state, as shown. Pb-free paste has shown more resistance to tombstoning, likely explained by a weaker wetting force exerted on the terminations, permitting greater tolerance for paste misregistration and grab imbalance.

A sample of 01005 resistors was also assembled at 0° rotation to determine orientation sensitivity. No influence on assembly yield was observed.

Build 5: This build used the same component placement locations and Pb-free solder paste as Build 4. The only difference with this assembly is that the profile atmosphere was changed from nitrogen to air. This combination of solder paste and profile produced a significantly defective result. Nearly all deposits showed incomplete solder coalescence. Larger deposits, particularly on the largest width pads, show better results (Figure 10). Assembly yield conclusions for this build were unfeasible because of challenges with reflow.

Stencil 2 builds. Builds 6, 7, 9 and 10: This stencil was designed with largest aperture designs to be sized 1:1 with pads, and then decreasing in size to 0.46 area ratio, to determine the influence of solder volume on assembly yield. Figure 11 illustrates the aperture design and assembly plan.

All four boards were printed with Type 3 SAC 305 Pb-free solder paste, two using vendor A and two from vendor B. For each paste vendor, one board was reflowed in nitrogen and the other in air. A new jar of the same paste printed in Build 5 was opened and used to determine if fresh material would improve air reflow performance. Indeed, there was some improvement in Pb-free reflow noted with new Paste 1, as reflow coalescence did not appear as sensitive to changes in printed solder volume. However, solder joints formed from Paste 1 in air reflow did exhibit some granularity on most deposits. SAC 305 Paste 2 displayed better reflow performance in air compared to Paste 1, with incomplete coalescence occurring only on pads where paste transfer was starved because of aperture clogging. Figure 12 shows the difference in appearance between solder joints formed with the two solder pastes across nitrogen and air reflow conditions.

No occurrence of shorting or solder balling defects was witnessed. At locations where both component pads receive printed solder, no open joints resulted. As the gradually reduced aperture sizes starve solder away from the pads, there is also a tendency for unbalanced paste volume distribution to occur across terminations. In some extreme cases, apertures may be totally clogged at low area ratios so that only one component termination is placed into solder. This condition was observed to produce no instances of tombstones, but did result in a few occurrences of low-angle drawbridging. Figure 13 shows an example of such an observation. Further investigation into this behavior is the focus of testing with Stencil 3.

Stencil 3 builds. Builds 11, 12 and 13: With the unexpected absence of open solder joints from all previous builds, these assemblies’ purpose was to study the tombstone trends between two Pb-free solder pastes on the effect of printing solder on only one of the two termination attachment pads for both 0201 and 01005 resistors using PCB Side B. The design of the 0201 pads reflected a combination found to produce the highest assembly yield from earlier testing.^4,5 Components exhibiting an angle of inclination and not lying flat on any of the non-printed pads were counted as a tombstone. Results are plotted in Figure 14.

The frequency of tombstones occurring on 0201 resistors is much greater than 01005s when the reflow atmosphere is nitrogen. This observation is consistent across Paste 1 and Paste 2 materials using their respective unique thermal profiles. Figure 15 shows a typical view of 0201s and 01005s from an assembled board reflowed in nitrogen. Tombstoning occurrences across both passive types are nearly the same level when the reflow atmosphere is switched air. Although both component types show greater resistance to tombstoning when reflowed in air, the effect of inerting the reflow atmosphere is significantly more influential on causing 0201s to tombstone. There was no correlation found to show the influence of aperture size on tombstone trends for either 0201 or 01005 components.

Samples of each component type were also assembled on pairs of printed pads as the experiment control. No tombstones were found under this assembly condition for either component across the three builds.

Discussion

Under the constraint of a 0.004" thick stencil, the solder paste transfer efficiency prediction for 01005-scaled apertures is low due to low area ratio stencil designs used. Despite this, several pad designs showed promise for achieving consistent solder deposit formation using “standard” aperture designs to offer promising assembly yield results. The replication of print success shown here will be influenced by proper seal formation between apertures and pads, particularly using low area ratio stencil designs where solder adhesion is already biased more strongly on aperture walls. The effect of board topography (i.e., solder mask thickness and registration), which may vary significantly across applications, is expected to have considerable impact on 01005 process stencil gasketing.

Using Type 3 solder paste, the print deposit shapes produced by <0.6 area ratio apertures will not be well-formed bricks. Rather, they more closely resemble rounded mounds of solder that contour wider at their base. Placement machines with delicate nozzles that drop components onto printed solder deposits may be challenged to cope with such print deposit shapes because free-falling components tend to land in odd positions on non-flat solder deposits. For conventional direct forced placement into printed solder deposits, tapered print deposit shapes are compatible.

Other notable observations regarding 01005 resistor placement include nozzle design and ESD. Component insertion into the paste not only flattens the solder, it squeezes material out laterally and also upward along the termination sides. As the solder deposit thickness is nearly the height of the 01005 resistor itself, the solder paste can easily access the nozzle during placement. The nozzles used in this study (Figure 2) experienced periodic episodes of solder paste accumulation on the ends of the hard deck tip. Solder debris contaminating the nozzles contributed significantly to pick yield loss and required regular nozzle cleaning to minimize rejected components. Pad designs with widest separation (S) and low stencil grab were found to have the highest influence on solder contamination on the nozzle. Nozzle tip design should consider suitable sizing to inhibit this.

Design of 01005 component placement should also consider implementation ample ESD controls. Selection of static-resistant materials for nozzles and feeders that rapidly dissipate any charge produced due to removal of the component cover tape is strongly suggested.

Air reflow is considered the most attractive atmosphere choice from a cost perspective. This process produces lower molten solder wetting strength, which minimizes tombstones and bridging, but tends to be more difficult generating the molten condition for 01005 size solder deposits, particularly for Pb-free solder. This study identified paste age as one factor that affected air reflow performance (i.e., Build 5 vs. Build 7). Evidence was found that the useful life of Pb-free solder paste could be extended when implementing a nitrogen reflow atmosphere. A difference in air reflow performance between two Pb-free materials, Paste 1and Paste 2, was also observed. It is unclear whether this difference is a result of material ingredient variations alone, or if there is any thermal profile sensitivity. It has been suggested that improved air reflow results can be achieved using faster heating.3 This theory is supported by observations in this research showing better air reflow coalescence of Paste 2, which uses more aggressive heating than Paste 1.

Recommendations

The following designs are suggested for 01005 resistor assembly compatibility with stencil thickness of 0.004" and Type 3 solder paste. The ranking strategy for pad designs is shown in Table 11.

Four categories are used to evaluate the designs. Only “standard” aperture sizes are considered, with designs with print area ratios below 0.5 scored with a demerit, indicated by red font. The second category, “narrow,” compares the pad width dimension to the 01005 resistor width. Because all pad width “A” designs were narrower than the component width, these are all designated “poor,” as this scenario generally is not advised. The third category, “symmetry,” assigns demerits to aperture designs most rectangular. The logic here is that nonsymmetrical apertures will have unequal print transfer results comparing 0° and 90° orientations. The final category of tombstones represents locations where this occurs on Build 3. This is the only PCB side A assembly where such defects were observed. The final column tabulates all the demerits (red labels). The summation of demerits is used to determine the qualified pad and stencil designs.

The top three pad and stencil designs are BF, CE and BE. Ranking in this order follows area ratio from high to low, so BF should produce better print results than the other two designs. Of the three pad separations, level “G” did not show tombstone defects and is considered the recommended design. At pad separation “G,” the influence of aperture grab level was found to be insignificant as a source of defects. The highest aperture design grab level “H,” where apertures are centered on the pads, is suggested, as this should help relax print and placement accuracy requirements.

Pad designs BFG, CEG, BEG and “standard” size high grab stencil aperture designs are capable of comfortably tolerating ±20 µm print and ±60 µm placement misregistrations while still maintaining component contact into printed solder deposits. The worst-case scenario is simulated in Figure 16. Larger pad separation or lower aperture grab designs will further challenge the ability of this offset scenario to recover in reflow.

Given concerns over achieving sustained, consistent paste transfer results for low area ratio apertures, some additional changes can increase printing success. Upgrading to electroformed stencils should increase print process stability, as the lower surface roughness aperture walls encourage the highest paste release efficiency performance.8 Another enhancement is replacing the squeegees with commonly available low friction metal blade options.8 Such materials improve solder paste roll circulation during the print and also reduce the tendency for paste pulling off the stencil when the squeegee is retracted after print completion. Enclosed printhead technology can also be used in place of squeegees to provide a boost to transfer efficiency performance.⁹

The performance of “oversize” aperture designs was broadly successful in achieving repeatable high transfer efficiency solder deposits. However, the prevailing flaw with this stencil design strategy was the delivery of excess solder to pads, causing bulging solder joints and the occasional floating component. Referring to Figure 5, the only pad design recommended for use with the oversize aperture is pad design BD. Again, the lowest pad separation “G” and highest grab aperture design are advised.

Conclusion

A successful 01005 resistor assembly process has been demonstrated using a 0.004"-thick stencil with Type 3 SnPb and Pb-free alloy solder pastes across a variety of copper OSP pad sizes. Side-to-side pad spacing as narrow as 0.004" showed promising assembly yield results. The pad design BFG with 0.009" wide x 0.010" long apertures centered on pad is recommended. The 01005 resistors assembled with Pb-free solder were found to be significantly less sensitive to tombstoning when reflowed in nitrogen compared with 0201 resistors assembled in the same manner. Similar defect ratios are found between the two component types if reflow is conducted in air. A Pb-free air reflow 01005 assembly process is considered the most desirable industry practice, and proof of achieving this has been realized. However, this work also reveals observations of incomplete solder coalescence, suggesting that air reflow compatibility challenges may exist with small lead-free solder deposits.

Acknowledgments
Thanks to Ashok Viswanathan of Binghamton University for assistance with the component assembly process and many constructive discussions that helped shape this work. Also, compliments to the staff of the Unovis-Solutions Advanced Process Laboratory for supporting the assembly and inspection equipment requirements to complete this project.

Ed.: This article was first presented at IPC Apex in February 2007, and is published here with permission.

References

F. Mattsson, D. Geiger, D. Shangguan, and T. Castello, “Design and Assembly of 01005 Passives Using Pb-Free Solder,” Circuits Assembly, May 2005.
P. Grundy, and M. Magnell, “Process Optimization of an 01005,” SMT, May 2006.
T. Borkes, and L. Groves, “Process Characterization of the 01005 (English) Component Package,” SMTA International, September 2006.
J. Adriance and J. Schake, “Mass Reflow Assembly of 0201 Components,” IPC Apex, March 2000.
G. Westby, J. Adriance, W. Prinz von Hessen, J. Schake and D. Barbini, 0201 Issues and Process Window, September 2000 webcast, http://webevents.broadcast.com/cahners/universal0900/home.asp.
A. Viswanathan, J. Schake and K. Srihari, “Process Characterization for the Assembly of 01005 Components,” SMTA International, September 2006.
Area ratios calculated per IPC-7525, Stencil Design Guidelines.
W.E. Coleman and M.R. Burgess, “Choosing a Stencil,” SMT, July 2006.
M. Mukadam, K. Srihari and P. Borgesen, “Assembly of 0.4mm Pitch Wafer Level Chip Scale Packages (WLCSPs),” Area Array Consortium 2003 Report, Universal Instruments Corp., 2003.