Organic Solderability Preservative (OSP) is an ultra-thin organic final finish coating for copper pads on printed circuit boards (PCB), which provides protection from oxidation under a wide range of conditions. Several advantages like lower cost, easier processing, excellent coplanarity compared to metallic finishes has made OSP an attractive choice to PCB fabricators and assemblers. Solder joints formed on OSP has negligible tendency of electromigration as they do not contain any metallic impurity coming that may be present in case of other finishes.

OSP finish is being used in over 60% of all PCBs made today and it has proliferated into high reliability markets like automotive which have traditionally used metallized finishes. OSP coatings, however, have a few shortcomings as well. OSP coatings after exposure to one or more thermal excursions can become more challenging to solder especially in wave and selective soldering applications. For example, a thick multi-layered double-sided PCB coated with OSP that has been exposed to one or more SMT solder paste reflow cycles may experience reduced solder wetting or hole fill during subsequent through-hole soldering process. This poses a huge challenge to all the assemblers, especially while working with SAC alloys with higher operating temperature.

Solutions suggested by engineers and researchers to mitigate poor soldering performance on pre-reflowed OSP were limited to PCB and component design improvements or optimizing soldering process parameters. Some groups suggested to use more active fluxes, but these fluxes tend to leave corrosive residues and compromise electrical performance of the assembly. Keeping these challenges in mind, we have tried to understand the chemical changes OSP coatings undergo during thermal excursions and identify elements of soldering flux which will interact with such thermally exposed OSP to convert them into clean copper surface, which in turn will result into better solder wetting performance.

Characterization of OSP Coating

OSP coating is formed by dipping specially treated copper surface into a solution containing active OSP chemicals. During this process OSP molecules attach themselves on the copper surface by forming co-ordination bonds with surface metal ions as well as by electrostatic attraction. When OSP coated copper pads undergoes thermal exposures (baking or reflow), the oxide layer at the metal – organic interface gets thickened depending on the nature of the exposure (Figure 1).

Figure 1. Schematic diagram of oxide thickening at Cu – OSP interfaceFigure 1. Schematic diagram of oxide thickening at Cu – OSP interfaceThe oxide layer formed under different thermal conditions was quantified using Sequential Electrochemical Reduction Analysis (SERA). The results showed that the oxide layer consists of both Cu2O and CuO and its thickness increases with the increase in exposure temperature or time (Figure 2). The study also indicated that there may be a temperature range above which OSP coating softens and changes its morphology. During this transition coating becomes less protective and allows external gas (oxygen) to penetrates through it which in turn increases the oxide layer thickness.

Figure 2. Oxide thickness in OSP coating after different thermal excursionsFigure 2. Oxide thickness in OSP coating after different thermal excursionsWe have studied this morphology change of OSP coating by Atomic Force Microscopy (AFM). Three-dimensional AFM images (Figure 3) showed that as coated OSP surface consists of extensively large number of fine-grained micro-peaks which largely disappear after Pb-free reflow with the appearance of larger grain like features. This morphology change is not noticed when coupons are exposed to 165 °C or less, independent of exposure duration. These observations reinforce the hypothesis of coating softening above 165 °C suggested in SERA study.

Figure 3. Three dimensional AFM images of as coated and thermally treated OSP coatingFigure 3. Three dimensional AFM images of as coated and thermally treated OSP coatingThe obvious question here is “what type of changes in the organic layer result in larger grain morphology after OSP is exposed to higher temperatures?” Images obtained from field emission scanning electron microscopy (FESEM) showed presence of small size random grains present on the surface of as coated OSP which got converted into elongated fibre like structures upon standard Pb-free reflow (in air or nitrogen) (Figure 4). Formation of such fibre may be attributed to supramolecular structure formation of substituted benzimidazole molecules (active OSP chemical) via interactions of pi-electrons of its aromatic rings.

Figure 4. FESEM images of as coated and a thermally treated OSP coatingFigure 4. FESEM images of as coated and a thermally treated OSP coatingFrom the above results it can be summarized that when OSP coated PCBs undergoes thermal excursions, two main changes take place – (1) oxide layer at the organic metal interface thickens and (2) active OSP molecules get rearranged to form supramolecular fibre like structures. Oxide thickening requires both heat and oxygen while fibre formation is affected by the elevated temperature alone.

OSP Flux Ingredient Interaction

Now with our understanding about the changes in OSP after thermal excursion, we wanted to find what chemicals, often contained in liquid fluxes, can overcome the effects of these changes. In soldering flux industry, few organic solvents are loosely termed as ‘OSP cutters’ because of their ability to dissolve OSP coatings. We studied the relative ‘OSP cutting efficiency’ of few such solvents. As coated and 2x reflowed laminates of fixed dimension with a commercially available OSP finish were dipped in fixed volume of each of these solvents (at 70-80 °C) and 5% HCl. OSP coating dissolved in solvents was quantified by measuring the UV absorption of the resultant solutions. 5% HCl known to dissolve as coated OSP completely and absorption value in of this solution was taken as 100% OSP concentration. Absorption value of other solutions was converted into relative percentage of OSP dissolved (Figure 5). From the absorption plot it is easily understood that most of the solvents taken can easily dissolve as coated OSP, but solubility decreases drastically once the coating undergoes two reflows. Only solvent SOL1 and 5% HCl can dissolve 40-50% of the pre-reflowed coating under the experimental conditions. Solvent SOL1 is a one of the strongest organic solvent and can break down the supramolecular arrangement in pre-reflowed OSP. On the other hand, HCl can react with benzimidazoles (which is a base by nature) and solubilize it.

Figure 5. Relative percentage of as coated and reflowed OSP dissolved in different solventsFigure 5. Relative percentage of as coated and reflowed OSP dissolved in different solvents

Figure 6. Relative percentage of as coated and reflowed OSP dissolved in solvent and activator solutionsFigure 6. Relative percentage of as coated and reflowed OSP dissolved in solvent and activator solutionsSo, if inorganic acids can dissolve thermally treated OSP can organic acids be able to do the job? To confirm this 1% solution of two organic acids, were prepared in SOL3 (having poor performance on reflowed OSP) and dissolution study was carried out as earlier. Results showed that both the acid solutions can dissolve ~ 80% of as coated OSP. But in case of pre-reflowed OSP, while solvent can dissolve only 5-10% of the OSP layer, activator solutions showed better efficiency, dissolving 20 – 30% (Figure 6).

Dissolving the organic layer, however, is only half of the job done. For solder to wet and form a joint, the oxide film beneath the organic layer also needs to be removed. Historically, organic or inorganic acids, halides, halogens etc. combinedly known as activators are used for this purpose. To understand how they react with oxides, differential scanning calorimetry (DSC) was used. Commercially available analytical grade cuprous or cupric oxide was mixed with an activator (90:10 ratio) by dry mixing and was heated in DSC pan at 10 C/min ramp from 30 °C to 300 °C. A baseline DSC study of only activator was carried out to find if activator undergoes any heat change during this thermal ramp. Representative DSC thermograms are shown in Figure 7 (a and b). DSC results reveal endothermic peaks for the only activator runs corresponding to the melting point of the activators. Additional exo- and endothermic peaks are observed for oxide-activator mixtures. These additional peaks we believe are resulting from the reaction between activators and copper oxides.

Figure 7. Representative DSC thermogram of activator copper oxide mixturesFigure 7. Representative DSC thermogram of activator copper oxide mixtures

A closer look at the thermogram also reveals activator ACT1 reacts with cupric oxide just after it melts around 140 °C, while the other, ACT3 reacts with cuprous oxide at similar temperature after melting at 100 °C. ACT3 also showed a smaller reaction exotherm around 220 °C with cupric oxide. This indicates each of the activators react with a particular oxide at a particular temperature range. Thus, a multi component activator package, where different constituents will react with different oxides at different temperature ranges, will be efficient to remove all the oxides from OSP coated copper pads.

Wetting of Solder on OSP

Now with the knowledge on the flux ingredients require to remove thermally treated OSP, we wanted to evaluate their efficiency to wet solder on OSP coated surfaces. Few activators and solvents used in above studies were shortlisted as shown in Table 1. Wetting evaluations were carried out on commercially available OSP coated coupons using Malcomtech SWB-S2 wetting balance tester fitted with SAC305 solder bath. Wetting data were collected on both as coated and 2X Pb-free pre-reflowed coupons with solder bath temperature of 260 °C. Solder score (S) for each measurement was calculated using the equation, S= [3.5 – T0] + [4 – T2/3 – T0] + [Fmax ´ 10], where, Fmax, T0 and T2/3 stand for maximum force, time to zero wetting and time to reach two third of maximum force respectively. T0, T2/3 and Fmax values were obtained from the instrument and solder scores (S) for each run were calculated using that data.

Table 1. Details of solvent activator combinations for wetting balance studyTable 1. Details of solvent activator combinations for wetting balance study

A solder score plot (Figure 8) shows that all solvent – activator combinations except 7 have excellent wetting properties on as coated OSP. All these combinations have solder score above 5.93, a requirement for the sample to be called as AA-class per IPC J-STD 003C. However, when tested on 2X pre-reflowed OSP coupons, solder scores for only a few combinations remained above 5.93, indicating that only these combinations were able to remove the pre-reflowed OSP coating effectively. Solder score data obtained from such simple wetting experiment helps flux formulators to shortlist ingredients to develop new products with best in class wetting on pre-reflowed OSP.

Figure 8. Solder score plot for solvent – activator combinationsFigure 8. Solder score plot for solvent – activator combinations

Another interesting observation from this experiment is that whenever a secondary solvent is used – solder scores of those combinations have declined. Combinations 7 and 8 contains the same activator as combination 1 with addition of two different secondary solvents and combination 11 have same activator as 10 with a secondary solvent. We have seen in the earlier section that these solvents help to remove thermally treated OSP. Authors assume this anomaly of results between two set of experiments is due to the partial drying of the coupons in absence of preheat system in wetting balance. Whenever a partially dry coupon is dipped in the solder bath, volatiles present in coupons dilutes the activator system and also causes spattering, which disturbs and delays the wetting process. Wetting balance is an excellent tool to screen the efficiency of activators, but it may not be the most reliable method to compare flux performance. Especially for fluxes containing higher boiling solvents which preserve the activity at higher temperatures. These fluxes require more complex evaluation processes like wave or selective soldering to compare the activity and other properties.

To verify this, we selected few commercially available liquid fluxes containing some of these activator – solvent combinations. Details of the solvents and activators present in the fluxes are described in Table 2. Wave soldering performance of these fluxes were benchmarked on a 2.4 mm thick 6-layer PCB with a commercially available OSP finish. PCBs were subjected to 2X reflow prior to wave process. Wave soldering were carried out on ERSA Power Wave machine using SAC305 bath at 265 °C. Flux loading for all the fluxes were controlled in such way that activator amounts on the board remain constant. Wetting performance of the fluxes were evaluated by comparing the x-ray images of the PCI connector barrel filling (Figure 9).

Table 2. Details of activators and solvents present in commercial fluxes taken for wave soldering performanceTable 2. Details of activators and solvents present in commercial fluxes taken for wave soldering performance

From the barrel fill images, it is very clear that Flux A containing solvent – activator combination 3 has the worst performance among the fluxes. Flux B and C with combination 5 and Flux D with combination 6 showed excellent barrel filling. Wave soldering performance of these fluxes corroborates the results obtained in wetting balance experiments. But with addition to these, Flux E having solvent SOL3 and activator combination 5 and Flux F with combination 11, also demonstrate soldering performance equivalent to Flux B, C and D. For Fluxes E and F, secondary solvent is helping the flux to remove the thermally treated OSP, thus assisting in solder wetting process. In many Pb-free capable fluxes secondary solvents help dissolving OSP and also preserve the activity for soldering higher temperatures.

Figure 9. X-Ray images of PCI connectors showing the barrel filling for different fluxesFigure 9. X-Ray images of PCI connectors showing the barrel filling for different fluxes

Conclusion

The study showed when OSP-coated copper pads undergo thermal excursions, two major changes take place. Copper oxide film at the organic metal interface thickens due to formation of Cu2O and CuO. In addition, the organic layer becomes compact as substituted benzimidazole molecules rearrange themselves to form elongated fiber-like structures via intermolecular attraction. This organic layer of pre-reflowed OSP can be dissolved by some solvents or solution of organic acids in these solvents. Flux activators can react with the oxides and remove them at elevated temperatures. Different activators reacted selectively with one of the oxides at different temperature ranges, indicating the necessity of a multi-activator package for soldering fluxes.

Wetting balance method was found to be useful in screening activators to effectively remove pre-reflowed OSP, but the results sometimes led to false failure especially where a secondary solvent is used in addition to IPA to make the activator solutions. Thus, for fluxes with secondary solvent, it is advisable to compare the flux activity using wave or selective soldering processes. Results obtained from both wetting balance and wave soldering experiments, however, confirm the effectiveness of fluxes with multi activator packages in achieving superior wetting on thermally treated OSP.

For more information, visit MacDermid Alpha Electronics Solutions.