Process Considerations for Optimizing a Reflow Profile

Estimated reading time: 9 minutes

Implementing a lead-free reflow process with alloys that have higher melting points is challenging. This article addresses the question of whether the materials used in the reflow process, especially the solder paste, can meet these high-speed reflow soldering requirements.

By Denis Barbini, Gerjan Diepstraten and Ursula Marquez

One main issue when implementing a lead-free reflow process is the optimization and repeatability of the process within lead-free paste and component specifications. The result is a narrow reflow soldering process window, which typically requires current reflow soldering equipment to decrease conveyor speed and meet material specifications while producing a reliable solder joint. This scenario has two consequences that compromise a manufacturer’s ability to produce an acceptable end product. The first is reduced throughput. This compromise may be unacceptable, but a necessity. The second compromise, which depends on board complexity, component selection and oven type, is the exposure of materials to higher temperatures for longer times with minimal tolerance levels. Small deviations in the reflow process, while insignificant in SnPb processing, may lead to increased defect levels in lead-free processing.

Due to precise process controls, newly-designed reflow ovens characterized by improved heat-transfer coefficients can sculpt a profile for SnAgCu solder pastes with the smallest product temperature range and wider tolerance levels. This prevents the overheating of sensitive components and enables the end user to implement “high-speed reflow soldering”, which is defined as running a lead-free profile at the speed of the analogous SnPb reflow process - without making any concessions toward throughput (Figure 1).

Figure 1. Reflow time temperature profile SnPb vs. SnAgCu with a reduced throughput.

For this study, ten lead-free solder pastes from major paste suppliers were qualitatively and quantitatively ranked by a series of analytical tests, as well as their behavior when exposed to two lead-free profiles. Analysis of the results are used to illustrate the impact that specific process parameters, such as peak temperature and time above liquidus (TAL), have on wetting, intermetallic formation, residue amount and cosmetics on a lead-free test vehicle. Sensitivity of the formation of a lead-free joint indicates that small changes in the reflow process can impact solder-joint characteristics significantly. This compounds the challenge to maintain a robust and repeatable reflow process.

Process Considerations

The ultimate goal of the reflow soldering process is to form reliable solder joints while maintaining component integrity. The solder must melt, coalesce and wet to appropriate terminations. None of the materials (neither components nor board material) should be damaged. Consequently, control over the reflow process is critical in achieving these goals.

There are a number of reflow parameters that are critical during solder heating and cooling. Many solder-paste manufacturers define these parameters for their respective pastes:

• Temperature gradient in preheat zone;
• Soak temperatures and time;
• Temperature gradient from soak to maximum temperature;
• TAL temperature of the solder;
• Maximum peak temperatures;
• Cooling gradient in cooling zones;
• Total heating time.

Developing a “high-speed” lead-free profile requires process settings that are at, or near, extremes of current lead-free solder-paste specifications. To ascertain the behavior and characteristics of a paste, two lead-free profiles were developed. Details of the two profiles are listed in Table 1. A selection of ten lead-free solder pastes from major paste suppliers was studied to verify if “high-speed” reflow soldering is feasible.

null

Solder Paste Fundamentals and Verifications

A number of different test procedures are defined in DIN and JEDEC standards. However, none are modified for lead-free solder pastes. The tests applied in this study were developed for SnPb assembly, but adjusted to accommodate the higher temperatures required by lead-free pastes.

Of the selected solder pastes, seven were SnAgCu-based, but with different silver and copper content. The other pastes were SnAg, SnCuNi and SnPbAg.

Solder ball diameters - A sample of each solder paste was inspected under a microscope to measure ball diameters. All solder pastes meet DIN32513 and J-STD-005 Type 3 specifications, but several showed a wide variance in solder-ball size, resulting in Cp values of approximately 1, where at least Cp of 1.33 is required.

Wetting test - Copper coupons were scoured with sandpaper and cleaned with alcohol. Four deposits of solder paste were printed on the coupons. One sample from each paste type was soldered according to a “high-speed” reflow profile, one hour after printing. Another sample was soldered after 72 hours to verify solder-paste robustness at room temperature.

Qualification rankings of these tests need to be reviewed because lead-free solder pastes have different spreading characteristics. Class 1 is specified that after soldering, the melted deposit should be larger than the printed diameter. A lead-free paste will not spread differently than SnPb after heating (when printed on copper).

This test showed that lead-free solder pastes are stable over a longer period of time. Of all pastes, the SnPbAg showed the biggest difference between those samples soldered after printing, and those soldered after 72 hours (Figure 2). Only one of the lead-free solder pastes met the Class 1 specification after being soldered with a “high-speed” reflow profile.

Figure 2. Different appearances of the Sn3.0Ag0.5Cu and the Sn0.7Cu0.1Ni molten solder.

Intermetallic layers - Cross sections were made from the samples to study intermetallic thickness. Thickness was measured at 20 spots on each sample, and the average was calculated. Due to higher tin content and solder temperatures, intermetallic thickness for lead-free soldered joints is thicker than SnPbAg joints. The intermetallic thickness for SnPbAg was 1.4 µm. The lead-free solder averaged 1.7 to 2.9 µm, representing an increase in thickness from 20 to 100%.

Standard deviation of intermetallic thickness indicates layer shape. A small number signifies a homogeneous layer thickness.

The wetting angle of the solder was higher for the lead-free solder due to the minor solderability of the lead-free alloys. The SnPbAg had a solder angle of 4.42º, whereas the SnAg-based solder paste showed an angle of 7.32º.

Solder-slump resistance - A pattern of solder paste was printed on copper with distances between the deposits increasing from 0.2 to 1 mm. The slump test (as described in the standards) must be reviewed for the higher-temperature lead-free solders. In this test, samples were heated for 2 min. at 180ºC, and paste slump was inspected. Results were not consistent. Some solder paste performed poorly, while others met expectations (Figure 3). Solder-slump resistance is critical because it affects defects such as solder beading and bridging.

Figure 3. Slump resistance of solder paste: Samples 7 and 10.

null

Experiments Focused on Flux Chemistry

More advanced research on these solder pastes was performed in a laboratory using Thermal Gravimetric Analysis (TGA) equipment. This instrument measures the weight change as a function of time and temperature. The first experiment involved heating different solder pastes according to 3- and 5-min. heating profiles. During profiling, the mass of the solder paste was weighed continuously. This allowed the following calculation:

• Amount of flux evaporating during the reflow process;
• Amount of residue left on the assembly;
• Where the flux is active in the process, and where it is not;
• Where, and in which oven zones, does solder-paste chemistry evaporate?

The “high-speed” 3-min. profile will be compared to the slower 5-min. profile.

Calculation of Flux Residues on the Assembly After Reflow

The mass of a lead-free solder-paste sample after reflow was 92.6% of the initial amount. For this solder paste, the metal content is specified as 88.5% of paste weight. Thus, we can calculate the volume of flux left after reflow:

(92.6-88.5)/(100-88.5) = 35.4% of total flux.

This showed that for the 3-min. profiles, all solder pastes left between 35 to 60% of residues on the assembly after reflow soldering, whereas between 30 to 55% remained for the 5-min. profiles.

For three of the solder pastes tested, only a small difference between the two profiles was noted, indicating that these solder pastes are more compatible to reflow at a higher conveyor speed than the others (Table 2).

null

Where Flux is Active - and Where It is Not

The derivative of weight loss measured with the TGA indicates where the flux is evaporating. If flux is evaporating, it indicates that the solder paste remains active in that area. Critical areas in the profile are at the end of soak and peak zones. During those times, temperatures are high, solder melts to form a proper solder joint and the potential for oxidation is high. We recognized that many lead-free solder pastes lost most of their activity before reaching peak zone. Results of solder-paste evaluations showed that most suppliers improved the chemistry to be more compatible with the higher temperatures of lead-free reflow. Two solder pastes showed too much activity in the first part of soak zones, resulting in poor performance in the higher-temperature areas of the solder process. Upon inspection, the solder showed a bad solder coalescence due to this poor activity in peak zones.

One can divide the process into four sections: preheat, soak, peak and cooling. Typically, for all solder pastes, fluxes start to evaporate at about 120ºC. Therefore, only 1% of the flux is evaporating in the preheat section. In the soak area, the amount that evaporates depends on the chemistry, but typically is 20 to 30% of the flux. Extreme high values that have been measured in this experiment are 50%, indicating the poor performance of these fluxes in the other parts of the process.

Flux chemistry loses typically 45 to 60% of its weight in peak. All pastes showed that 10 to 20% of the flux still evaporates in the cooling area. The risk is that these residues evaporating in the cooling zone condense on colder spots in the oven, unless the extraction system can remove them from the process area before condensing (Figure 4).

Figure 4. The evaporation chemistry in the different peak zones of a reflow oven.* Sample 7 shows poor activity in the peak, where Sample 1 performs correctly in a three-minute profile.

null

Results of the Experiment and Machine Design

The information obtained from the experiment is good input when designing new generations of reflow ovens. Understanding the flux chemistry of solder paste gives designers knowledge of where to exhaust air or nitrogen in a reflow oven, and how to filter residues. Lead-free solder-paste residues are more difficult to deal with than SnPb. Condensation of flux in the cooling zone will contaminate internal parts. A sufficient exhaust will remove vapors from the area where the assembly is soldered. A separate filter system with easy access for cleaning is necessary.

Conclusion

Solder paste suppliers have made significant improvements with respect to the performance of flux chemistry used in lead-free solder-paste applications. For those solder pastes with poor activity in peak zones, optimizations are required to meet high-temperature specifications.

All SnAgCu solder pastes show rough surface structure and micro-cracking as a result of shrinkage characteristics of this alloy. Nitrogen, flux or fast cooling have little to no impact on the formation of these micro-cracks. Other alloys, such as SnCuNi and SnPbAg, show a smoother solder surface.

Flux residues left after soldering vary in composition, color and amount. Because some solder pastes showed poor performance, slump resistance should be noted; possibly increasing the prevalence of solder beading or bridging defects.

Experiments have proven that a “high-speed”, 3-min. heating profile can be created without making concessions to solder-paste performance. From Table 2, we learn that some solder pastes need a longer profile. For those pastes, the difference between shorter and longer heating profiles is significant.

*My Reflow 930 oven.

For a complete list of tables, contact the authors.

REFERENCES

1. All samples are soldered according to the three-min. reflow profile in air.
2. All TGA experiments were done in nitrogen.
3. There were no TGA experiments for this solder paste.

Denis Barbini, Ph.D., advanced technologies manager; Gerjan Diepstraten, senior research engineer; and Ursula Marquez, process and research engineer, Vitronics Soltec, may be contacted at (603) 772-7778 e-mail: info@us.vitronics-soltec.com.