The Effectiveness of 75% IPA/25% DI Extraction Solution on No-clean Flux Residues
The continuous challenge in the electronics industry is to keep up with the demand for smaller, faster, and more reliable electronics. This places great strains on the manufacturers of those electronics. The strain on manufacturing electronics comes from every direction. Environmental regulations have removed lead from the majority of electronics, forcing materials changes. Packages have become smaller with more voltage requirements. Complex devices are expected to work reliably for years under harsh environments that would be unimaginable to a circuit designer one decade ago. These examples and many more combine to create a multitude of challenges.
This rapid development of manufacturing and reliability challenges is juxtaposed with the slow pace of new test method development. Resistance of Solvent Extract (ROSE) testing is the de facto ionic contamination testing method primarily because it was developed in 1970, when PCBs were becoming more common in military and other high-reliability applications. In the ROSE method, a PCB to be tested is submerged in a solution of 75% isopropanol or IPA (but more properly propan-2-ol) and 25% DI water. The change in the resistance of the solution is monitored. As ionic materials dissolve into solution, the resistance of the solution decreases, indicating that there is ionic contamination.
This dip in resistance is then converted into the mass of sodium chloride per unit area that would be expected to produce an equivalent dip in the resistance of the solution. Traditionally, this is expressed as micrograms of sodium chloride equivalent per square centimeter (μg NaCl eq. /cm2). Shortly after ROSE’s development, the United States Department of Defense decided that a value of 1.56 μg NaCl eq. /cm2 was to be used as a criterion for product acceptance. If a particular assembly was below this value, it would be acceptable for field use.
Relevant to the current work is the decision to use the extraction solution of 75%/25% IPA/DI water. This decision was made because the vast majority of solder fluxes on the market were rosin based, and therefore, highly soluble in IPA. This is because the major constituent of rosin is abietic acid and is highly soluble in IPA [1]. The true systematic name for abietic acid is abieta-7,13-dien-18-oic acid (CAS Number 514-10-3) and is shown in Figure 1 [2]. It is commonly cited as virtually insoluble in water, but it does exhibit some solubility in water [3]. It is also known to be highly soluble in various organic solvents, although exact numbers are hard to come by [2, 4, & 5].
Once dissolved into solution, abietic acid can ionize into an electrically conductive species that can be measured by resistance. This is what characterizes abietic acid as what our industry calls weak organic acids (WOAs). The degree to which the ionization occurs is highly dependent on pH, but will cause a change in resistance none the less.
Figure 1: Abietic acid structure [1].
As time progressed, newer fluxes entered the market that were not based on rosin, namely water soluble (WS), or sometimes called “water wash flux,” and no-clean (NC), or “low-residue flux.” There is a multitude of reasons why these fluxes came onto the market and why they necessitated different WOAs, but those are beyond the scope of this article. The flux manufacturers turned toward a different type of organic acid, linear dicarboxylic acids. Dicarboxylic acids are organic molecules that have two carboxylic acid groups, and linear dicarboxylic acids are simply the dicarboxylic acid derivatives of straight chain hydrocarbons.
The exact identities of the dicarboxylic acids used are considered proprietary, but there are a series of dicarboxylic acids that have been reported in the literature, such as adipic acid, succinic acid, glutaric acid, malic acid [6,7], and more recently, pimelic acid [8]. Table 1 shows the structure of these acids [9]. The important thing to note from the structures in Table 1 is that they are substantially different from that of abietic acid, and as a result, they all have some degree of solubility in water.
Table 1: Common dicarboxylic acids.
One additional consideration is that the WOAs can react. One known chemical reaction that they undergo is acid-base reactions. In this reaction, WOAs react with a base, such as an amine activator, metals, or other bases that are present. These result in the formation of a salt. Typically, organic salts are more soluble in water than organic solvents; this property is a common way to purify WOAs [10]. It is suspected by the authors that these WOA salts are capable of causing current leakage paths or even ECM. When dissolved, these salts result in the formation of an electrolyte solution.
During this time, a new cleanliness assessment method was introduced to the industry in the form of ion chromatography (IC). IC is an analytical technique from the liquid chromatography family and can be used to identify and quantify ionic species in solution. A small portion of a liquid sample (typical volume ~25 μL) is injected into a moving fluid called an eluent. The eluent carries the sample to an analytical column. This column consists of very fine particles coated with an ionic compound. This coating attracts the ions of interest, slowing them down. The degree of this slowing depends on a number of factors, but the ultimate goal is to slow each ion of interest down enough that they travel as discrete bands in the plumbing of the instrument. This causes the ions to be removed by the eluent from the column at different times.
The eluent exiting the column passes through a conductivity cell where the conductivity is continuously measured. When these bands of ions pass through the detector, a large peak forms from the ion. The size of the peak is proportional to the number of ions present, and the time from injection it takes to appear is characteristic (but not conclusive) of different ions [11]. Despite these changes, no changes to the extraction solvent were made. It was assumed that the WOAs of Table 1 would be equally soluble in 75%/25% IPA/DI water as abietic acid is. An informal survey within the industry revealed that no one knew the answer to the question, “What is the solubility of common WOAs in IPA/DI water mixtures?” This is a critical question as for detecting ions via conductivity, which is used in both ROSE and IC; they must be dissolved in solution.
It is relatively easy to find the solubility of most WOAs in water. It is somewhat more difficult, but still possible, to find the solubility of WOAs in common solvents, such as IPA. What is virtually impossible to find in literature is the solubility of WOAs in a mixed solvent system. Many complex factors and interactions govern solubility. These factors are not additive nor linear. A literature review found only a few cases where the solubility of abietic acid or any dicarboxylic acids of interest were measured [12 & 13]. Generally, this was under conditions that were not similar to those in industry. It was with this stumbling block the solubility was to be investigated.
Methods
1. Gravimetric
To assess the solubility of abietic acid, glutaric acid, malic acid, succinic acid, and sodium succinate, saturated solutions of each were prepared in the solvent systems indicated in Table 2. To do this, each acid was obtained from Sigma-Aldrich and used without any further purification. Next, the acids were placed in 250-mL jars with a magnetic stir bar and a quantity of the chosen solvent. The jars were sealed and vigorously stirred at room temperature (20–25°C) for at least one week.
A substantial excess of acid was used to ensure that solubility was reached as indicated by a large amount of undissolved acid. Then, each solution was transferred into four pre-weighed vials using a syringe and a 5-μm membrane filter to remove any suspended solids. The weight of all of the vials was recorded. The vials were then placed in a dry bath at 110°C until they reached a constant weight, indicating dryness. For all weights, an analytical balance with an accuracy of ±0.0001 g was used.
Table 2: Solvent systems.
2. Ion Chromatography
A parallel study was undertaken to approach the issue for the other direction, namely looking at changes in ion concentration as a function of extraction solvent as measured by IC. A total of 25 boards were printed with an industry standard NC solder paste. The boards were reflowed without populating the QFN components in a standard reflow profile. After reflow, boards were randomly assigned an extraction solution from Table 2. A total of five boards were used for each solvent tested. Each board was extracted individually per IPC-TM-650 2.3.28 (80°C for one hour). The extraction volume was held constant for all boards. Twelve of the samples were analyzed immediately by Laboratory A who produced the boards. Thirteen of the samples were sent to Laboratory B for analysis.
Results
Gravimetric
The results of the gravimetric method are presented in Figure 2. The plots show all four replicates plotted for each IPA concentration as well as a LOESS fit with the standard error shown.
Figure 2: Gravimetric results.
Ion Chromatography
Due to instrument errors, the full data set is not available at the time of writing. Preliminary data is available and discussed along with future work. IC results for the total WOA extracted from each board is presented only to protect proprietary information (Figure 3).
Inferences
The results of these findings would seem to have practical impacts on measuring the reliability of PCBAs. The gravimetric results indicate that more “modern” WOAs are soluble in pure water about as much as they are in IPA. However, if the WOA forms a salt with a base, the trend is the opposite. These salts would still be capable of electrochemical migration and to serve as a leakage path for current. The IC data shows a different picture, however. The maximum extraction appears to be around 50%/50% IPA/DI water. This indicates that it is possible that ROSE and IC under report contamination as there could be more ions present than are soluble under a different concentration of IPA. It is suspected that a non-extraction based technique, such as SIR, may have an advantage over ROSE or IC.
Figure 3: Preliminary IC results.
Conclusions
Gravimetric
At the present time, there is not enough data in the literature to confirm the gravimetric findings, so additional work will be needed. The WOA with the most available solubility data from reputable sources is succinic acid. The solubility in pure DI water ranges from 77 mg/mL [14] to 83 mg/mL [15]. Our reported value was 84.8 ± 0.069 mg/mL, which is in good agreement. However, the solubility in IPA and/or IPA+ water is not readily available. The solubility of succinic acid in similar alcohols is 57 mg/mL for ethanol and 158 mg/mL for methanol [14]. Our value for pure IPA was 60.0 ± 0.024 mg/mL, which seems plausible. The data for malic acid and glutaric acid is suspect as our values are substantially less than those found in the literature.
While there is a lack of solid data in the literature for the solubility of the WOAs in solvent mixtures, the trends do match what might be expected. Abietic acid needs a critical amount of organic solvent to begin to be solubilized. While not discussed in this article, sodium succinate was expected to follow an opposite trend than the other materials, as it is a salt, because salts tend to have less solubility in organic solvents than their acid counterparts. The flux reactions should generate different salts of WOAs, which would be less detectable, but equally as harmful as their parent acid compounds.
Ion Chromatography
The IC data is interesting but does not have enough of a sample size to draw any firm conclusions. What is clear from the data is that the solubility of all of the WOAs is not linear with IPA concentration for this particular solder flux. It appears that, for the flux studied, there is an optimum IPA concentration, which may be located around 50% IPA 50% DI water. This indicates that the commonly used 75% IPA 25% DI water system may not be optimal.
Further Work
The findings of this study are interesting and warrant further investigation. To better refine the gravimetric data for the WOAs understudy a complimentary solubility determination method will be used. Top contenders for this method are titrations, ultraviolet-visible spectroscopy (UV-VIS), laser solubility determination, or a synthetic method [16 & 17]. These methods have the potential for being much more sensitive than the gravimetric method. The UV-VIS method is currently being explored. In addition to additional experimental work for the solubility of WOA in mixed solvent systems, a theoretical approach using Hansen Solubility Parameters is being undertaken [18]. In addition to improving the data as a function of IPA concentration, the temperature dependence of the solubility is of critical importance.
The IC work is being continued as well. Currently, additional IC samples are being run at Laboratory B. This work will be updated and expanded to additional fluxes. The specific flux used is suspected to be a substantial factor in the amount of extractable WOA ions.
References
1. Lee, N.-C. “Reflow Soldering Processes and Troubleshooting: SMT, BGA, CSP, and Flip-Chip Technologies,” Lee, N.-C. B. T.-R. S. P., (Ed.) Newnes, 2002.
2. Sigma-Aldrich. “Abietic acid,” https://www.sigmaaldrich.com/catalog/product/sial/00010?lang=en®ion=US.
3. Nyren, V.; Back, E. “The Ionization Constant, Solubility Product, and Solubility of Abietic and Dehydroabietic Acid,” Förlagsföreningen Acta Chemica Scandinavica, 12, pp. 1516–1520, 1958.
4. Lide, D. R. CRC Handbook of Chemistry and Physics, CRC Press, 2005.
5. Armarego, W. L. F.; Chai, C. L. L. Purification of Laboratory Chemicals: Sixth Edition; Armarego, W. L. F., Chai, C. L. L. B. T.-P. of L. C. (Eds.), Butterworth-Heinemann, 2009.
6. IPC-TM-650 2.3.28 Ionic Analysis of Circuit Boards, Ion Chromatography Method. IPC 2004.
7. Zhou, Y.; Turbini, L. J.; Ramjattan, D.; Christian, B.; Pritzker, M. “Characterizing Corrosion Effects of Weak Organic Acids Using a Modified Bono Test,” Journal of Electronic Materials, 42 (12), pp. 3609–3619, 2013.
8. Verdingovas, V.; Jellesen, M. S.; Ambat, R. “Solder Flux Residues and Humidity-Related Failures in Electronics: Relative Effects of Weak Organic Acids Used in No-Clean Flux Systems,” Journal of Electronic Materials, 44 (4), pp. 1116–1127, 2015.
9. ResearchGate. “Evaluating the Efficacy of a Group of Nontraditional Plasticizers on the Glass Transition Temperature of Ethyl Cellulose Polymer,” https://www.researchgate.net/Example-of-dicarboxylic-acids_tbl1_47413566.
10. Lehman, J. W. Operational Organic Chemistry: First Edition, Prentice-Hall, 2002.
11. Harris, D. C. Quantitative Chemical Analysis: Seventh Edition, W. H. Freeman and Company, 2007.
12. Nong, W.; Chen, X.; Wang, L.; Liang, J.; Wang, H.; Huang, Y.; Tong, Z. “Measurement and Correlation of Solubility of Abietic Acid in Ethanol + Water Mixtures,” Journal of Chemical Thermodynamics, 68, pp. 199–204, 2014.
13. Nong, W.; Chen, X.; Wang, L.; Liang, J.; Wang, H.; Long, L.; Huang, Y.; Tong, Z. “Measurement and Correlation of Solid-Liquid Equilibrium for Abietic Acid + Alcohol Systems at Atmospheric Pressure,” Fluid Phase Equilibria, 367, pp. 74–78, 2014.
14. Royal Society of Chemistry. “The Merck Index Online,” 2013.
15. National Center for Biotechnology Information. “Succinic Acid CID=1110,” https://pubchem.ncbi.nlm.nih.gov/compound/1110.
16. Jouyban, A.; A. Fakhree, M. A. “Experimental and Computational Methods Pertaining to Drug Solubility, Toxicity, and Drug Testing,” InTech, 2012.
17. Yi, Y.; Hatziavramidis, D.; Myerson, A. S.; Waldo, M.; Beylin, V. G.; Mustakis, J. “Development of a Small-Scale Automated Solubility Measurement Apparatus,” Industrial & Chemical Engineering Research, 44 (15), pp. 5427–5433, 2005.
18. Hansen, C. M. Hansen Solubility Parameters: A User’s Handbook: Second Edition, CRC Press, 2007.
Editor's Note: This technical paper was originally published in the proceedings of SMTA International, 2018.