The business of lead-free assembly

With enforcement of the EU WEEE directive less than 18 months away, assemblers need to be producing faultless lead-free assemblies, at commercial yield rates, before the start date of July 2006.

As well as investing in time and training to understand and practice with the revised processes, research is also showing that assemblers may need to re-optimise some aspects of their equipment, such as stencils. In any case, the cost of not being ready is too high to contemplate.

Convert All Processes
The world has been talking about lead-free, with some sectors working feverishly behind the scenes, for some years. Replacement alloys are close to completion, and dual lead-rich/lead-free compatible components have already entered the market. But it appears that many assemblers have not devoted enough effort to studying the practical changes that are about to take effect. Some apparently expect their OEM customers to provide precise instructions to accompany new lead-free contracts. Although some OEMs are keen to assert control in this way, none can provide 100% of the information assemblers need.

Some processes, such as screen printing in particular, rely substantially on knowledge to arrive at an optimal setup. This information needs to be reacquired, and the only way is through practical contact with the materials, trial, error, and rectification of problems and mistakes. In short, assemblers need to invest time and effort right now, to ensure their key production staff get their hands dirty.

Screen Printing with Lead-free Solder Pastes
Table 1 shows the composition of lead-free solder pastes from the major brands. All samples feature Type 3 particle size (25 - 45µm), suspended in a no clean flux medium. The alloy types and rheology packages are the main variable parameters. All samples shown in table 1 are commercially available.

Paste label	Alloy type	Metal content (% weight)
W	Sn 95.5Ag3.8Cu0.7	89.3
X	Sn96.5Ag3.0Cu0.5	89.0
Y	Sn96.5Ag3.0Cu0.5	88.0
Z	Sn63Pb37	90.25

 Table 1. Variation of lead-free solder paste characteristics.

To begin to build a picture of the effects of Lead-free pastes on the mass imaging process, a Design of Experiments based on print speed, paste pressure and separation speed, at HIGH, LOW and MEDIUM levels, enables a rapid assessment of process capability for each paste type. Including a widely-used lead-rich paste in the DoE provides a benchmark against which the lead-free processes can be compared.

Experiment number	Print Speed (A)	Paste Pressure (B)	Separation Speed (C)
1	H	H	H
2	L	H	L
3	M	M	M
4	L	L	L
5	L	L	H
6	H	H	L
7	L	H	H
8	H	L	H
9	H	L	L

Where L = low; M = medium; H= high
Table 2. Design of Experiments to Investigate Print Speed, Paste Pressure and Separation Speed.

With all other process parameters, such as print gap, system pressure, temperature and humidity held constant, the results can be presented in a three-dimensional diagram, where each corner of the cube represents one of the nine experiments. Analysis of paste deposits using a solder paste inspection system allows the average solder paste volume and the standard deviation across all deposits for each solder paste on test to be calculated. Colour coding the experimental results enables a concise comparison of all parameters and their effect on process capability: red indicates low paste volume and poor repeatability that would be unacceptable in a production process; yellow indicates acceptable results. Green areas highlight the point at which paste volume is closest to the nominal value, with the lowest standard deviation: that is, a close to ideal process. Graphical results from the DoE for each solder paste are shown in figures 1 to 4.

These diagrams clearly show that each factor has a significant effect on process performance. For each solder paste, it is possible to find a "sweet spot". This demonstrates that all lead-free pastes are compatible with enclosed head mass imaging technology. Moreover, high excursion speed and low print pressure tend to deliver generally better results, suggesting a greater imperative on adopting enclosed head printing using a system such as ProFlow® DirEKt Imaging. Separation speed also appears to be a critical factor in the search for a robust process.

Lead-free Process Analysis
The printing process can be simplified down to two sub processes: an aperture filling process, and an aperture release process. The first depends predominantly on the relationship between print speed, paste pressure, and the solder paste properties. The second depends on the stencil aperture characteristics as well as the physical properties of the solder paste. It has also been found to vary with separation speed. This is a rediscovery, as processes based on lead-rich pastes have been almost completely insensitive to this variable.

Aperture filling
Aperture filling is dominated by the relationship between paste rheology and the transfer head excursion speed, as well as the transfer head technology (such as enclosed head printing). Paste characteristics can be expected to converge as lead-free technology matures. But in the immediate term, simply switching to the lead-free paste offered by an incumbent paste partner may not deliver the best possible results. To be sure of achieving the highest yields as quickly as possible, we recommend re-evaluating lead-free pastes from all suitable suppliers, the effect of enclosed head printing on aperture filling has already been well investigated with lead-free processes, and the advantages will not be diminished in the lead-free era. In fact, the wider process window afforded by using ProFlow or a similar enclosed print head will be of even greater benefit considering the changes in rheology from lead-rich to lead-free.

Aperture release
After carrying out the DoE, Plotting paste release efficiency, as a percentage of the nominal paste volume, against the aspect ratio of the apertures yields paste release curves as shown in figures 5 and 6 for rectangular QFP apertures and round CSP apertures respectively.

Note that an aspect ratio of 0.6 is generally accepted within industry to be the lowest useable value, below which paste release efficiency becomes unacceptable for production purposes. By including the Lead-rich solder paste, material Z, this experiment shows that established SnPb pastes support greater process robustness than Lead-free at all practical stencil geometries. Only when the aspect ratio is reduced below 0.6 does paste release for Lead-rich paste fall below that of the Lead-free formulations. Given that the Lead-free pastes are relatively new, and the technology is at the beginning of its lifecycle, this is perhaps to be expected.

Implications for Stencils
The recent analysis has shown that lead-free pastes display lower aperture release efficiency compared to lead-rich paste, if used directly with an existing stencil technology. Process development for lead-free screen printing must therefore aim to recover this lost efficiency. This may require investment in replacement stencils for products converting to lead-free assembly.

All aspects of the stencil should be scrutinised to achieve this, including aperture dimensions and aspect ratio, but also in terms of material properties and manufacturing process. Each of these aspects influences the characteristics of the aperture wall, with a corresponding effect on paste release. 

In particular, increasing the nickel content appears to aid paste release. Laser cut stencils made using high nickel-content stainless steel have performed better than regular stainless steel stencils in other laboratory tests. However pure nickel stencils, created by laser cutting, display even higher efficiency. Electro-formed stencils in pure nickel, on the other hand, have shown only marginally higher release efficiency than laser cut pure nickel samples, with identical apertures. This indicates that the coefficient of friction of the aperture surface is more important than the surface finish in governing paste release efficiency for lead-free pastes.

This apparent dependence on high nickel-content steels and pure nickel will tend to increase the price of stencils optimised for forthcoming lead-free pastes. On the other hand, turnaround time is generally shorter for laser cut stencils than for electro-forming, provided the stencil manufacturer has good availability of blanks in the chosen material, whether pure nickel or high nickel-content stainless steel..

In the latest testing it has been found that a standard stainless steel laser cut stencil will perform to a modest level of capability. But the dimensional accuracy, linear integrity and wall finish has to be fabricated to the highest standard. Data has shown that if any of the 3 requirements are violated then the effect on Volume Cp/Cpk is dramatically reduced.

To increase the release robustness and performance of lead free materials the latest study has shown that the inclusion of nickel into the base material has improved the release capability. The findings have shown that the "Classic" Eform stencil outperformed all stencil technologies this follows the trend of previous work. But the inclusion of a nickel foil with apertures laser cut showed a performance which virtually mirrored that of the classic E-Form, this solution would give a manufacturing facility the advantage of the lower surface tension of nickel but with out the premium cost and overhead of delivery.

Conclusion
The industry expects lead-free assembly to be more expensive, mostly due to the higher tin and precious metals content of suitable solder alloys. But higher value stencils may also be necessary in order to maintain print process repeatability - and therefore end of line yield - comparable to lead-rich processes.

However, assemblers must look harder to evaluate the true costs of conversion, since the incremental costs of a higher-specification stencil material, even if moving to pure nickel is preferred, are less in the short to medium term than the costs for personnel to reacquire essential process knowledge. Nevertheless, this is an important expenditure, in terms of time and any applicable off-site training fees, as the price of poor expertise will be manifest in high levels of scrap and poor productivity resulting from low end of line yield. In the extreme, assemblers who cannot complete lead-free process development will be unable to build product within the EU or for sale in the EU.

The results from the Design of Experiments described in this article are not presented as a definitive set of process recommendations, but are intended to stimulate further practical research into optimal mass imaging process settings and selection of solder pastes. At this stage, less than eighteen months before lead-free legislation comes into force, it is extremely important for manufacturers to maximise their hands-on experience of lead-free materials and processes - ideally offline in a suitable environment. One way is by finding out about vendors' lead-free workshops such as those regularly hosted at DEK technical centres in the UK and internationally.

Figure 1. Process window for solder paste W.

Figure 2. Process window for solder paste X.

Figure 3. Process window for solder paste Y.

Figure 4. Process window for solder paste Z (lead-rich paste).

Figure 5. Paste release curve for rectangular apertures.

Figure 6. Paste release curve for round apertures.

Clive Ashmore is responsible for the Global Applied Process Engineering group for DEK.  Prior to joining DEK in 1998, Ashmore held senior process and manufacturing engineering posts at Ericsson and Philips.  He specializes in all aspects of manufacturing engineering, with special emphasis over the last five years on mass imaging technologies.  Ashmore has also done extensive research in other areas of surface mount assembly, quality engineering and equipment analysis.  One of Ashmore's most recent projects involves the study of mass imaging to enable wafer bumping and encapsulation.