J-STD-004 was updated about five years ago, but the comingling of designations from the old (A) and new (B) revisions in industry literature has created much confusion among users. What changed, what didn’t, how does it affect flux selection criteria, and what does a user need to know? Here’s a quick overview:
Flux designation has three components. The first two letters, RO, RE or OR, represent the basic chemical composition: rosin-based, resin-based or organic acid-based. Nothing has changed there. But the next component to flux designation, L, M or H, which describes the flux’s activity level as low, medium or high, and the final component, 0 or 1, which indicates halide content, are affected by the new revision.
A flux’s activity is determined by a number of tests. Most of these tests remain the same, but two important ones have changed. The most notable change is to the SIR testing, with updates to the test’s environment, electrical bias and sampling frequency. The other major change is the elimination of the qualitative halide test; this can significantly affect the halide content portion of the designation for a large number of fluxes. Table 1 summarizes the changes.
Table 1. Significant Differences Between J-STD-004A and J-STD-004B
What the PCB assembly engineer needs to know:
Don’t dismiss L1 fluxes just because they have some halides. Considering Revision B’s closure of the spot test loophole, the old L0 flux could contain more halides than the new L1 flux. Keep in mind that the newer flux, which may be labeled L1, passed a far more stringent SIR test than the older one labeled L0.
Know the relationship between halides and halogens. Halides are ionic compounds typically used as activators and often associated with corrosion. IPC standards address test methods and thresholds for halide content. Non-ionic halogenated compounds, or halogens for short, are sometimes used to improve the stability of fluxes and are the subject of environmental watch lists. European Environmental Standards (EN-14582) address test methods and thresholds for halogen content. A halide-free flux may not necessarily be halogen-free, but a halogen-free flux will also be halide-free.
Understand the subtle differences between rosin and resin. The terms are often used interchangeably, but rosin is a naturally occurring substance, and resin is either a modified rosin or completely synthetic material. Rosins are plant products and are subject to more natural variation than resins; resins are commonly used in newer flux formulations due to their more consistent performance. When more than one rosin/resin are combined, the IPC classification is based on the larger constituent. Therefore, if the distinction between RO and RE is important to the user, they should inquire with the flux’s manufacturer to better understand the exact details of the content and classification criteria.
Ask about the classification standard. The revised standard has been in force for roughly five years. If a flux formulation is over five years old, it was likely classified under the previous system; if it is fewer than five years old, it was likely classified under the current system. Some fluxes are classified under both. If the technical data sheet does not specify whether the product was classified to J-STD-004A or J-STD-004B, just ask.
The expiration of the RoHS exemptions is driving most of the industry to Pb-free materials and processes. One of the key tools that significantly improves the joint formation capability of lead-free solders is … wait for it … halides. Yes, halides. Taking the more stringent SIR/electromigration requirements of J-STD-004B into account, the rejection of a flux product simply because the IPC classification is “1” and not “0” can handcuff an operation and adversely impact yields, productivity and solder joint quality.
All PCB assemblers share a similar goal: to make many good solder joints with flux characteristics that meet the mission profile of the assembly. Blindly adhering to a specification without regard to the implications is a classic case of the tail wagging the dog; summarily limiting assemblers’ chemistry options based solely on nomenclature compromises soldering performance and, ultimately, the end-products’ long-term reliability.
Nickel-modified tin-copper solder, known commercially as Sn100C, is a leading lead-free alloy for PTH soldering, rework and hot air-leveled PCB final finishes. Because it contains no silver, it is much more economical than SAC alloys containing as little as 1% of the precious metal, and it produces smooth, shiny, easy-to-inspect solder joints. Why has it not gained major acceptance as an SMT alloy? In large part, fear. Fear of full compatibility with SAC reflow processes.
SnCuNi melts at 227°C. SAC 305 begins to melt at 217°C, reaching its fully liquid state at approximately 220.6°C. Recommended reflow temperatures are typically at least 13°C higher than melting temperatures, hence the SAC 305 peak temperature window of 233° to 255°C. Applying the 13°C guideline to the SnCuNi alloy results in a minimum peak temperature recommendation of 240°C. It’s that 7°C difference in minimum reflow temperatures that generates the fear – of cold joints, incomplete wetting, inconsistent IMC formation or other reliability issues – if the peak temperature of a solder joint falls into that questionable space between 233° and 240°C.
Laboratory tests have demonstrated good wetting at lower temperatures, but they’re just that: lab tests. Until recently, there was no real-world, production-based data on the soldering behavior of the SnCuNi alloy near the low margin of the reflow window. But a newly published study1 explored exactly that situation. It dropped the Sn100C alloy into a reflow process considered cool for SAC 305; the temperature peaked at 234°C with only 60 sec. above 217°. The cool process shown in Figure 1 was dictated by thermally sensitive components on the PCB, which was a mixed-technology industrial controller of low-medium complexity.
The experiment consisted of two runs, one with SAC 305 solder paste and one with SnCuNi solder paste. Thirty boards of each were built. Both pastes used the same flux, and the two products ran down the assembly line sequentially with no changes to any portion of the SMT process. Following assembly, the boards were subjected to numerous tests and analyses.
Inspection and electrical test. AOI took place on the assembly line, and visual inspection was performed by the line’s IPC-certified inspectors. No defects were found by either inspection process. The SnCuNi solder joints formed under the cooler profile did not exhibit the typical luster that is characteristic of the alloy, but were not as dull as SAC 305. Figure 2 shows photos of the typical solder joints produced in this process.
Five PCBs from each alloy group were then fully assembled with PTH, wave-soldered, and installed into chasses for functional test. They all passed, and the assemblies were earmarked for thermal cycling.
X-ray analysis showed more voiding in the SnCuNi solder joints than the SAC 305 (Figure 3). Although the voiding levels are acceptable, they could likely be mitigated by adding a soak to the reflow profile.
Microstructural analysis. Optical microscopy showed good wetting to both leaded and leadless terminations, and good IMC formation. In the as-reflowed state, both alloys formed continuous IMCs approximately 3µm thick. They were then subjected to two thermal aging cycles: an initial one at 125°C for 96 hr. and a subsequent cycle at 150°C for 240 hr. The initial cycle had no significant effects on the IMC or the solder joint shear strength. Both alloys demonstrated similar IMC growth during the second aging (Figure 4).
Joint strength. Next, 0805 components were shear tested at 15° angles before and after thermal aging. Shear strengths averaged 4 to 6kg, typical for this component size and comparable to previous tests run on assemblies with peak temperatures of 245°C. Thermal aging had no significant impact on shear strength values (Figure 5).
Thermal cycling. Original experimental plans included thermal cycling five full assemblies from each alloy set. However, given the good appearance, microstructure and solder joint strength, it is highly unlikely any solder joints would fail during cycling, and no new knowledge would be gained from the effort. A similar study that compared the two alloys on a more complex PCB completed 3000 cycles (0° to 100°C) with no remarkable results. Therefore, thermal cycling was eventually eliminated from the test plan.
The results of this test were eye-opening. A reflow process with a peak temperature of only 234°C and 60 sec. above 217°C is considered the low end of the window for SAC 305. Despite SnCuNi’s higher melting point, it formed good joints, passing every standard test to which it was subjected. It demonstrated full compatibility with the cool range of the SAC 305 window, establishing itself as a viable drop-in replacement for almost any SAC-based SMT process.
The silver-free alloy is currently used in numerous consumer applications, but concerns over low temperature reflow processing have slowed broader implementation
and delayed its resulting cost reductions, which can be substantial. Replacing SAC 305 with SnCuNi can save as much as 20% on solder paste costs due to the elimination of silver.
This study was instrumental in generating production-based data for the previously uncharted area of the reflow window. It should help to dispel fears of poor reliability if SnCuNi is soldered on the low end of the established SAC 305 process. Many thanks to Andy Monson and Walter Machado of Hayward Industries for making this experiment possible and sharing the results with the industry.
1. The data presented here are part of a larger study presented and published at SMTA International; it can be downloaded at http://www.aimsolder.com/technical-articles or at http://www.CircuitsAssembly.com.
Needed: Solder joint forensic scientist. Common sense preferred.
Wanted: X-ray engineer. A test engineer with an interest in x-ray technology will suffice. So will a skilled and teachable technician. Hell, an intelligent person with a pulse will do in this economy. We’re open-minded. Just show us. No shrinking violets here. Honesty still matters to us (like being honest about the state of the economy and its effects on available talent). You should be honest, too, if you’d like us to hire you. Bring the aptitude; we’ll give you the qualifications.
We will train you.
Is the reflow profile the problem? X-ray can help.
Looking through some recent x-ray images of what I would call “good bad” boards (at least, that is what they are for me, as they showcase “good” examples of how certain “bad” types of failure look under x-ray inspection), I came across a number of different issues that are different from “traditional” BGA/QFN problems mentioned in this space before. To wit, I noted some images showed where solder paste had not reflowed under the devices, and there was the presence of foreign object(s), such as discrete components, trapped under the package.
FIGURES 1 and 2 show how unreflowed solder paste typically looks under QFN joints in an x-ray image. In the magnified view (Figure 2), individual grains of the solder paste are seen clearly, instead of appearing as a typical single smooth continuous joint. The cause of this is probably not an insufficient reflow profile. Rather, it is more likely the board has not been reflowed at all. As it may be desired, or necessary, to x-ray inspect (representative) boards after placement but before reflow as part of a quality control process, it is worth noting this characteristic shape of the solder under the components is different from what would be expected post-reflow.