Reading time ( words)
Many companies are shifting away from cleaning with pure de-ionized (DI) water to chemistry-assisted processes. Numerous reasons can be cited supporting the recent trend toward cleaning with chemistry. Results from this experiment, presented by Harald Wack, Ph.D., Zestron, support the hypothesis that cleaning temperature and cleaning chemistry concentration can be manipulated to meet cost and cleanliness requirements for the majority of solder pastes.
Upon examination of the electronics manufacturing industry in North America, a clear trend becomes apparent. Many companies are shifting away from cleaning with pure de-ionized (DI) water to chemistry-assisted processes. Numerous reasons can be cited supporting the recent trend toward cleaning with chemistry. For one, an increased use of lead-free solder requires higher soldering temperatures. This results in more burnt-in fluxes, which are much harder to remove as they begin to produce water-insoluble contamination. DI-water alone has a limited ability to solubilize non-ionic residues on the boards’ surface.
Secondly, cleaning of leaded and lead-free water-soluble fluxes (especially under low-standoff components) has become much more difficult. In other words, water, with a high surface tension of over 70 dynes/cm, cannot effectively penetrate between the PCB and low-standoff components. As standoff heights decrease and component densities increase, companies will have to improve their existing cleaning processes.
Table 1. Process parameters used during the study.A recent research study compared various commonly used water-soluble, lead-free solder pastes. Cleanliness assessments on bare FR-4 areas as well as surfaces under chip capacitors were conducted. All test assemblies were reflowed in a state-of-the-art oven to simulate production conditions as closely as possible. Table 1 shows variable as well as fixed process settings used during the test series.
During a design of experiment (DOE) study, a fully factorial analysis evaluated the variables of wash temperature, cleaning agent technology, cleaning agent concentration, and brands of solder pastes used. The baseline for evaluating relative cleanliness was DI-water and unpopulated assemblies with the same pastes. State-of-the-art in-line equipment was used to perform all experiments. A visual inspection (40×) was performed and the data collected.
Results Figure 1. Excerpt of obtained results for a commonly used OA solder paste.The results of the cleanliness tests for all water-soluble pastes showed significant differences when tested on bare reflowed assemblies. The limitations of DI-water are already evident when compared to a 3% and 5% (concentrations) chemically supported cleaning process. Temperature and concentration mostly factored into the cleaning results for DI-water, but there was no significant difference noticeable for both chemistries (cleaning agents 1 and 2) between 3% and 5% active concentration. This is an important conclusion as it suggests that 3% can be a feasible cleaning chemistry concentration and points to potentially lower operating temperatures when using a cleaning agent rather than straight DI water.
Apart from a few exceptions, the chemistry-assisted processes significantly outperformed straight DI-water usage in this study. An increase in temperature generally supports slightly better results, but numerous cases indicated full flux removal at temperatures as low as 120°F. The relative concentration levels show that chemistry levels of 3% might be sufficient for cleaning with various solder pastes currently used in the industry. A minority of pastes was not fully cleanable, which indicates that the tested chemistries are chemically capable of removing residues from the majority of commonly used fluxes.
The interaction of variables, with respect to the cleaning results, was investigated. This led to the set of final conclusions expressed in Figures 2–4. These figures illustrate cleaning performance in relation to different technical set-ups of the DOE experiment. The tested cleaning agent outperforms the pure DI-water cleaning process when cleaning water-soluble flux residues at lower wash temperatures. Secondly, there is a noticeable improvement when cleaning with an increased wash temperature.
While evaluating the concentration levels of the two cleaning agents, it was furthermore found that the cleaning results were not significantly different at 3% versus 5% concentration levels. Out of all tested pastes, five were more responsive to an increase in wash temperature in terms of cleanability, regardless of the cleaning agent concentration levels. Here, the removability of the remaining pastes correlates with the exposure time between the flux residues and the tested cleaning agents in the cleaning equipment. In other words, longer wash exposure times would assist in removing the flux residues to achieve 100% cleanliness. As expected, cleaning under low-standoff components was more challenging (at least 25% or more) than cleaning around the components.
The use of cleaning agents at concentrations as low as 3% can quantifiably provide up to 111% better cleaning results under low-standoff components when compared to a pure DI-water cleaning process.
ConclusionIn the long run, the use of cleaning chemistries seems to offer a number of previously unknown benefits. Despite the additional process cost of a cleaning agent, the “value added” benefits are sizable and should exceed that expenditure.
Recent supplemental studies have also demonstrated better bonding and coating results after the introduction of chemistry assisted cleaning to remove OA fluxes. To offset the added cost, users can operate cleaning systems at lower temperatures. The wider process window allows for cleaning not only OA but also RMA and no-clean fluxes. That will become a requirement in the North American market as contract manufacturers are moving to lower-volume/higher-mix assembly of significantly higher-reliability products. The introduction of a chemistry-assisted cleaning process will increase your cleaning process window and permit de-fluxing of all production boards in a single cleaning process.
Various customers are currently investigating our hypothesis in real-time. The results should be available shortly and will be presented at upcoming industry conferences.
Harald Wack, Ph.D., an SMT Editorial Advisory Board member, is president of Zestron. He holds M.S. and Ph.D. degrees in organic chemistry from Johns Hopkins University and received the Whittacker Chambers and Sonneborn Fellowships. Wack holds numerous patents and is an active member in industry organizations such as the IPC and SMTA. He is a member of the IPC Cleaning & Alternatives Handbook task group and chairs the European IPC/SMTA Cleaning Symposium 2009. He may be contacted at (703) 393-9880 or via e-mail at firstname.lastname@example.org. Read his other recent articles: What is Innovation in Chemistry? Reduce the Cost of Cleaning ProcessesIPA-Water (75/25): Ineffective for Cleanliness Test with Modern Contaminations