Analyzing Lead-free Wavesoldering Defects
December 31, 1969 |Estimated reading time: 10 minutes
A European consortium and others have concluded that Pb-free soldering is technologically possible, but implementation issues first must be addressed, including the VOC-free flux technology and whether the process must be modified to accommodate required higher soldering temperatures.
By Gerjan Diepstraten
Figure 1. Example of bridging. The score of this component is 6 because six pins exhibit no bridging while the others do.
The Taguchi design-of-experiment (DOE) method and statistical process control (SPC) are effective ways to evaluate lead-free processing in wave soldering. The goal is to determine the essential control parameters for the optimum setup of a specific application.
The Taguchi method seeks to integrate an innovative quality approach with a traditional experiment design method. A series of interrelated techniques is developed to minimize unwanted variability, reduce production waste and provide greater customer satisfaction. For example, the Taguchi approach for reducing production variations has two steps:
Figure 2. A completely filled through-hole (left) is 2 points. Sides of holes wetted with solder are accorded 1 point (right). Sides of holes unwetted with solder are 0 points.
1. Manufacture the product in the "best" manner to achieve fewer deviations from the target.
2. Produce all products as identically as possible to achieve fewer deviations between the products.
Taguchi experiments employ a specially constructed table or "orthogonal array" to affect the design process so quality is built into a product during its design stage. An orthogonal array is an experimental design that permits a mathematically independent asessment of the factors affecting the experiment.
Experiment PreparationTaguchi experiment preparation begins with a brainstorming session in which a team incorporating different disciplines establishes clear statements listing problems, objectives, desired output characteristics and measurement method for designing the appropriate experiment. Next, all process parameters are identified and relevant factors affecting the outcome are defined:
Figure 3. Bridging relative to process factors -- the higher the number, the higher the quality (200 = no bridging). Contact time and preheat temperature were most influential.
1. Controllable factors: C1 = factors that contribute significantly to the process and can be directly controlled; C2 = factors needed to stop the process if C1 factors are changed.
In this experiment, three C1 factors were selected:B = contact timeC = preheat temperatureD = flux amount
Solder temperature is a C2 factor because of the time required to increase/decrease the temperature.
2. Noise factors are process variables that affect variation but are impossible or cost-ineffective to control. Examples are changes in ambient temperature, humidity, dust, etc., during the production/experiment. For practical reasons, there is no "noise" element factored into the experiment. Rather, the main goal is to estimate the contribution of the individual quality influencing factors. Other experiments must be conducted to quantify their response to process noise.
Finally, a selection of output characteristics to be measured is required.
Two criteria are preferred: the number of pins without bridges and through-hole fill qualification.
Experiment Layout and DesignIn contrast to other methods (which mostly use one-factor-at-a-time studies to identify controllable parameters), this experiment used an L9 orthogonal array. In only nine experimental runs, four factors at three levels are investigated, as shown in Table 1.
Output characteristics. A proper experiment setup obtains the most reliable data. For example, the control parameter range must be as extreme as practical to make problems in this case, bridging and bad hole penetration (Figure 1) apparent. To quantify the effect of bridging, soldered pins without bridges are counted. (Each board had 200 pins, hence the maximum score would be 200.)
For the effects of through-hole penetration, each hole filled with solder is marked as shown in Figure 2. A total of 4,662 points was the maximum for each board.
Materials in the ExperimentPb-free alloy. The most common wavesoldering lead-free alloys are Sn/Cu and Sn/Ag/Cu. Sn/Cu, one of the least expensive Pb-free alloys, has a high melting point (227°C) in addition to poor mechanical properties compared to other lead-free alloys. Sn/Ag/Cu is an improvement over the basic Sn/Ag. Sn/Ag3.8/Cu0.7 solder forms higher reliability joints while its solderability is better than both Sn/Ag and Sn/Cu. The addition of antimony (0.25 to 0.50 percent Sb) provides increased thermal resistance via the intermetallic structures of antimony with silver and copper. However, there is concern regarding Sb toxicity, notwithstanding that toxic Sb oxides only are generated at temperatures above 600°C. Sn/Ag3.8/Cu0.7/Sb0.25 (SACS), with a 217°C melting point, was used in the experiment (Table 2).
Board finish. The organic solderability preservative (OSP)*, a high-performance copper coating that protects and maintains through-hole solderability, was selected for the test board. (Earlier studies demonstrated that Sn/Ag/Cu is compatible with OSP finishes.) OSP was a replacement for hot-air solder leveling (HASL) and other metallic printed circuit board (PCB) finishes. With higher preheat settings, the thin organic coating (0.2 to 0.5 µm thickness) loses activity. While the OSP is compatible with water-soluble fluxes, acids and solvents contained within the flux quickly dissolve the OSP coating, which becomes part of the flux and volatilizes when the molten solder contacts the board.
Flux. For this test, 396-RX synthetic volatile organic compound (VOC)-free flux with less than 2 percent solids was selected (Table 3). A halide-free, low-residue flux was selected because of its good solderability on copper surfaces plus its effectiveness in preventing bridging. Flux preheat, measured in topside temperatures, may range from 100° to 112°C, depending on the assembly configuration.
Flux application. Of the flux application technologies available, a nozzle-spray fluxer was found to best apply an appropriate layer to the board. With VOC-free fluxes, achieving the finest (smallest) droplets possible is key to achieving good through-hole penetration and successful water film volatilization. Accordingly, a water-based flux should have carefully formulated surface properties to yield a smooth interface with metallic and nonmetallic surfaces.
The nozzle fluxer permits precise control of applied flux from approximately 300 to 750 mg/dm2 (wet flux). The maximum is 750 mg/dm2 because excess flux begins to drip off the board beyond this point.
Test board design and material. Test board dimensions were 160 x 100 x 1.6 mm.
Material was FR-4 with copper double-side plated through-holes. The connectors feature 10 pins, double row with a 0.2 µm Au/Ni finish.
Test ResultsEighteen boards (nine boards with one repetition) were run to obtain the data needed for orthogonal-array analysis to achieve the following objectives:
Figure 4. Bridging pie chart showing the impact of each control parameter.
- Estimate the contribution of individual quality-influencing factors
- Gain the optimum condition for the Pb-free process
- Approximate the response of the control parameters under optimum conditions.
Analysis of variance (ANOVA), a statistical treatment, assessed the results of the orthogonal-array experiment and determined how much each factor contributed. Table 4 shows data related to bridging obtained from the experiment.
Figure 3 illustrates the incidence of bridging relative to process factors, i.e., the higher the number, the higher the quality (200 = no bridging). It shows that with respect to bridging, contact time and preheat temperature affected output data the most, i.e., changing one of the settings will have the most dramatic effect on the number of bridgings.
Optimal settings for bridging based on experiment data were determined to be A2, B1, C1 and D2. Although the difference between A2 and A3 is very small, A2 was selected because the 260°C solder temperature was preferred because of lower energy consumption requirements. Plus, at that level, less thermal shock to components and board material was experienced. Figure 4 shows the impact in percent of each control parameter with respect to bridging; Table 5 lists the experimental results related to through-hole wetting.
In Figure 5, once again, the higher the count, the greater the result (4,662 = a 100 percent "good soldered" board).
Preheat temperature (130°C) affects the process the most while the other factors' impact was about equal. Also, the experiment repetition error was small for through-hole penetration. (Figure 6 shows the impact in percent of each control parameter related to through-hole penetration.) Optimal settings established for the best through-hole solder penetration based on experiment data were A3, B1, C2 and D2. Figure 7 displays cross-section microphotographs of the SACS' solder joints of testboards used in the experiment.
Experiment ResultsSolder temperature impact was minor relative to its effect on bridging. For through-hole penetration, a higher solder temperature is better. However, this option may be limited because of potential component, flux activator and board material damage.
Figure 5. Through-hole penetration relative to process factors. Preheat temperature affects the process the most while the effects of other factors are about equal.
Shorter contact times yielded better results for this experiment, possibly because the flux activator system was compromised at the higher preheat and solder-pot temperature settings. This is typical for this flux type under these process conditions. Other tests reveal that longer contact times could be benificial if the flux activation system is strong enough to withstand higher temperatures. Otherwise, contact times of 2.5 to 4 seconds are recommended.
Figure 6. The input in percent of each control parameter related to through-hole penetration.
In terms of preheater temperature, a 110°C setting is the "champ" for this process. At a higher setting (130°C), the process window narrows considerably (the OSP coating and flux may lose activity). Preheating water-based VOC-free flux requires special consideration. Once the flux is applied, it is essential to improve the chemical bond between the flux and board surface, which can be achieved by heating the flux. Thus, at the end of the first process zone (600 mm), the topside board temperature should be about 70° to 80°C. For this test, medium-wave Calrod infrared (IR) generating elements were selected. The elements provide the appropriate IR energy volume and wavelength to initiate activity without forcing the water to boil out of the material at the start. Forced convection heating used in the second and third zones eliminates excess water before entering the solder waves.
A continuous, uniform flux spray pattern across the entire board is essential. The finest possible droplets applied with the lowest possible air pressure will deliver the best results. Higher settings may cause a droplet bouncing effect, rather than improved board surface wetting. The D2 setting was the "paper" leader in this experiment.
The overall "best" settings for this experiment are listed as follows: A3, B1, C2 and D2:
- A Solder temperature = 275°C; to avoid thermal damage, between 265° and 270°C.
- B Contact time = 1.8 seconds.
- C Preheat temperature (topside) = 110°C.
- D Wet-flux volume = 474 mg/dm2.
ConclusionFor successful Pb-free wavesoldering implementation in production volumes, the entire process must be reviewed, i.e., it is not simply a matter of dropping new chemicals and materials into the process. This review can be facilitated, and process development expedited, by conducting a Taguchi analysis and appropriately designed experiment. It will enable the process engineer to reach a practical understanding of what will be required in his own specific application with only a small number of test runs. Although new Pb-free wavesoldering processes will have a smaller process window because of higher temperatures and other materials, SPC can be a valuable tool to help the engineer remain within correct specifications and achieve minimum variation in the newly developed process.
Figure 7. Microphotographs of experiment joints. Micrograph (a) is the bridge shown in Figure 1; cross-section (b) shows a pin illustrating that no fillet lifting occurred in the experiment; cross-section (c) is a close-up of the intermetallic Ni/Sn layer between the pin and the lead-free solder.
The test board used in this experiment was run on a standard wavesoldering machine** equipped with a spray-nozzle fluxer, a three-stage preheater and nitrogen, which can obtain repeatable results in this lead-free process at a conveyor speed of 2 m per minute without problems. This illustrates that special or nonstandard equipment (or options) typically will not be required to convert to a Pb-free process. The spray-fluxer used was capable of delivering adequate flux to the through-hole barrels, which is encouraging because foam fluxers in general will not work properly with most water-based VOC-free flux (special, newly formulated fluxes are required).
*Entek Plus Cu-106A.** Vitronics Soltec Delta 6622.
WORKS CONSULTED
- P. Langeveld, D. Schwarzbach and E. de Kluizenaar, "Lead-free Wave Soldering Feasibility Study," Philips Electronic Packaging & Joining, CTR594-98-0051, 1998.
- Ibid. (CTR594-99-0012, 1999).
- B.P. Richards, C.L Levoguer, C.P. Hunt, K. Nimmo, S. Peters and P. Cusack, "An Analysis of the Current Status of Lead-free Soldering," NPL and ITRI, 1998.
- Marconi Materials Technology, "Improved Design Life and Environmentally Aware Manufacturing of Lead-free Soldered Assemblies, IDEALS," BRPR-CT96-0140, 1999.
- D. Suraski, A Study of Antimony in Solder, AIM.
- K. Hollevoet, "Organic Surface Protection of Copper Results Are Superior to HASL," Back to Soldering Basics, Interflux Belgium, EPP 1999.
- K. Wengenroth, "OSPs: Guidelines for Successful Soldering," Enthone-OMI, Inc., West Haven, Conn.
- http://kernow.curtin.edu.au/Taguch1.
- E. Westerlaken and H. Trip, "Process Considerations for the VOC-free Concept," Cobar Europe BV, 1999.
- M. Warwick, "Consortium Research: Implementing Lead-free Soldering," Multicore Solders Ltd., Hemel Hempstead, U.K., SMTA 1999.
GERJAN DIEPSTRATEN may be contacted at Vitronics Soltec, P.O. Box 143, 4900 AC, Oosterhout, Netherlands; 31-162-483236; Fax: 31-162-483253; E-mail: gdiepstraten@nl.vitronics-soltec.com.