Divergence in Test Results Using IPC Standard SIR and Ionic Contamination Measurements
Controlled humidity and temperature controlled surface insulation resistance (SIR) measurements of flux covered test vehicles, subject to a direct current (DC) bias voltage, are recognized by a number of global standards organizations as the preferred method to determine if no-clean solder paste and wave soldering flux residues are suitable for reliable electronic assemblies. The Association Connecting Electronics Industries (IPC), Japanese Industry Standard (JIS), Deutsches Institut fur Normung (DIN) and International Electrical Commission (IEC) all have industry reviewed standards using similar variations of this measurement.
Ionic contamination testing is recognized by the IPC as a standard for evaluating the cleanliness of assemblies that have been subjected to a cleaning process. IPC J-STD001F calls for a cleanliness level of
IPC-TM-650 Method 2.3.25 contains standard test methods for extracting contaminants from circuit boards using heated isopropanol (IPA) / water mixtures. Test method 2.3.25 is commonly referred to as the ROSE (resistivity of solvent extract) test. Previous work1,2 has shown poor correlation between the presence of extractable, corrosive weak organic acids and results from IPC-TM-650 2.3.25 test results, partially due to the lack of solubility of materials found in no-clean fluxes, and the higher SIR values imparted by rosins and resins in modern no-clean soldering materials.
This study will compare the results from testing two solder pastes using the IPC-J-STD-004B IPC TM-650 2.6.3.7 surface insulation resistance test and IPC TM-650 2.3.25 in an attempt to investigate the correlation of ROSE methods as predictors of electronic assembly electrical reliability.
Introduction
Ionic contamination testing has been traced back to work done at the United States Naval Avionics Center in Indianapolis in the early 1970s by Hobson and DeNoon3. This work eventually led to the development of the 1.56 µg/cm2 (10 µg/inch²) NaCl equivalent standard for ionics extracted using an IPA/water mixture. High volume circuit assembly at the time used only wave soldering processes, employing foam fluxers to apply RMA flux followed by a post soldering cleaning process with fluorocarbon solvents.
This ionic contamination limit became part of now defunct Mil Spec P-288094 and Mil STD-2000A, but has been carried through versions A through F of ANSI/J-STD-001. This manual procedure has become more automated with the invention of descriptively branded test equipment such as the Contaminometer, Ionograph and Omega Meter. Although these measuring devices improve the efficiency and accuracy of measuring ionic contamination soluble in alcohol/water mixtures, they also increase the amount of ionic material measured5, 6.
IPC-9202 describes a procedure for qualifying a process for electrical reliability by measuring SIR using IPC TM-650 2.6.3.7 and IPC TM-650 2.3.28 by using the IPC-B-52 test coupon. This coupon is shown in Figure 1. The standard calls for a minimum SIR value of 100MΩ, but only calls for a measure and report of the ionic contamination. The reported ionic contamination then becomes a benchmark for “future trouble shooting or process improvement efforts.”
The experiments carried out in this work were designed to use methodology derived from IPC-9202 to determine if it is possible to have a solder paste that passes SIR standard of >100 MΩ, but fails the ionic contamination level of ANSI/J-STD-001f, and determine if a second solder paste fail the SIR test and pass the ROSE test standard.
Figure 1: IPC B-52 test coupon comprising a SIR test coupon (SIR) and a section for ionic contamination measurements (SIR). Component ID: 1 – TH connector 4 x 24 pins; 2 – Capacitor, 10 pF, 0402 package; 3 – BGA, 256 IO, 1 mm pitch, isolated; 4 – SM connector IEEE 1386, 2 x 16 pins; 5 – Capacitor, 10 pF, 0805 package; 6 – QFP160, 0.65 mm pitch, isolated; 7 – QFP80, 0.5 mm pitch, isolated; 8 – Capacitor, 10 pF, 0603 package; 9 – SOIC16, 1.27 mm pitch, isolated; 10 – Capacitor, 10 pF, 1206 package.
Methodology
Two different SAC305 solder pastes were printed and reflowed on IPC B-52 test coupons (Figure 1). The assembled coupons were broken into two separate test vehicles after the solder pastes were printed, populated and reflowed. The section of the board on the center right was used to measure ionic contamination. The left portion of the test vehicle was used to measure SIR. The four smaller panels on the far right were discarded. A schematic diagram of the Ionograph that was used is depicted in Figure 2. The Ionograph is considered a “dynamic” ROSE measurement in which the extraction solution is continuously passed through ion exchange columns that remove the ionic material in the solution. A conductivity bridge detects ions in solution, and a flow meter measures the volume of solution passing by the conductivity bridge, allowing ionic contamination to be integrated with extraction solution volume.
A second measurement using three IPC-B-24 coupons (usually used for single material SIR measurements) for each of the two pastes was made.
Figure 2: Schematic diagram of the “Ionograph.”
Experimental Procedure
1. IPC-B-52 Coupon Preparation
Ten IPC-B-52 coupons were processed for each of the two selected solder pastes. These coupons were used as delivered from the board fabricator; no further cleaning was done. The pastes were printed using a 0.125 mm (5 mil) stencil. Positions 2, 3, 5, 7, 8 and 10, as shown in Figure 1, were populated with dummy components. The coupons were then reflowed using an OmniFlo 7 oven with a 1.1 °C/s straight ramp to a peak temperature of 243°C with a time above liquidus of 53 s using a nitrogen atmosphere (600–800ppm O2), as shown in Figure 3.
Figure 3: Applied reflow profile comprising a straight ramp to 243°C under nitrogen (1.1 °C/s, 53 s TAL).
2. SIR Measurements
The SIR portion of twenty coupons and two unprocessed control coupons were mounted in a temperature-humidity chamber. Teflon-insulated leads were hand soldered to the coupons. The chamber was programmed to run at 40°C +-1°C and 90–93% RH, and a GEN3 AutoSIR was programmed to apply 12V bias and to measure SIR at bias voltage every 20 minutes for seven days.
SIR data were recorded for pattern No. 5, 6, 7 and 8 (Figure 1). The boards were mounted in the temperature/humidity chamber and connected to a “Gen3 AutoSIR” instrument for measuring the SIR. The chamber was programmed to record SIR readings at 12 V every 20 minutes. No edge card connectors were used. Teflon-insulated wires, were soldered with ROL0 solder cored wire.
3. Ionic Contamination Measurements of IPC-B52-Coupons
Ten IPC B 52 ionics break away with three populated patterns were measured for each of the two pastes. An Ionograph 500M SMD II was used to measure the ionic contamination of each sub coupon. A 75% isopropanol 25% water extraction solution heated to 45°C was used. A dwell time of 15 minutes in the ionograph was used. This time was a balance between complete removal of ionic contamination, versus CO2 absorption increasing the apparent conductivity of the solution from a source other than the test vehicle. A PCB surface area of 65 cm2 was used in the calculations.
4. IPC-B-24 Coupon Preparation
A. Pre-cleaning
Modified IPC-B-24 test coupons, with bare copper FR4 were immersed in a 75% isopropanol/25% water solution in an Ionograph 500M SMD II. The solution was heated and circulated. The boards remained in the solution until a solution resistivity of >300 ohm-cm was achieved. The boards were then baked at 50°C for one hour.
B. Coupon Assembly
The solder paste was printed on three test coupons per paste using a 150 µm (6 mil) stencil. The coupons were then exposed the temperature profile similar to that shown in Figure 3 below using a OmniFlo 7 oven in N2 (600-800ppm O2).
Results and Discussion
Log SIR results for solder paste A and solder paste B measured on the leads around the small QFP on the IPC-B-52 coupons are depicted in Figures 4 and 5 respectively. Figure 6 shows the mean Log SIR values for all six patterns measured for both solder pastes. As can be seen from the 7-day test, paste A has a consistent SIR reading of three orders of magnitude greater than paste B. Paste B is a marginal fail if 100MΩ is used as the pass/fail standard.
Figure 4: IPC-B-52 SIR test results of paste A at small QFP leads.
Figure 5: IPC-B-52 SIR test results of paste B at small QFP leads.
Figure 6: Mean B52-SIR all patterns paste A and paste B.
Figure 7 compares the results for ionic contamination measured on the IC part of the B-52 test vehicles with solder paste A and B. Contrary to the SIR readings, the test coupons processed with solder paste A exhibit an ionic contamination level that were three times larger than solder paste B. The discrepancy is a result of the easy extraction of benign ionics in the alcohol/water mixture from solder paste A, and insoluble contributors to reduced surface insulation in solder paste B. A detailed analysis of the chemical nature of the ionic residues was not performed. The results are a borderline pass for solder paste A with high SIR values and a pass with a wide margin for the samples that used solder paste that failed to consistently maintain a SIR above 100 MΩ during the constant climate test.
Figure 7: Box plots of ionic contamination measurements of the IC part of B-52 test vehicles processed with solder pastes A and B.
Log SIR per J-STD-004C is shown for pastes A and B in Figure 8. In this case, SIR values are 4.5 orders of magnitude higher for paste A than paste B.
Figure 8: Paste A and Paste B SIR per J-STD-004B/IPC-TM-650 Method 2.6.3.7.
Conclusions
These divergent test results emphasize why a ROSE test should be used as a relative test to qualify a process, and to use the baseline SIR/Ionic contamination measured using the dual purpose test as a benchmark for “future trouble shooting or process improvement efforts.” Test coupons with ionic equivalent measurements just below the 1.56 µg/cm2 NaCl equivalent limit used in ANSI/J-STD-001f were shown to have three orders of magnitude greater surface insulation resistance, but still have 3X measured ionic contamination. Thus, the ionic contamination of a PCBA does not predict any reliability of the electronic control unit under high temperature and high humidity conditions. It is highly recommended that the 5-22A task group review and amend ANSI/IPC-J-STD-001f to account for the known divergence in SIR and Ionic contamination results on the electrical reliability if SMT assemblies.
*Co-Authors: M. Holtzer, T. Cucu, M. Liberatore, M. Schmidt, Alpha Assembly Solutions; and S. Moser, L. Henneken, P. Eckold, U. Welzel, R. Fritsch, D. Schlenker, Robert Bosch GmbH.
References
1. Seelig, K. “VOC-Free Flux Study-Not All Weak Organic Acids are the Same,” APEX 2012 Proceedings (2012).
2. Chan, A.S.L., Shankoff, T.A. "A Correlation between Surface Insulation Resistance and Solvent Extract Conductivity Cleanliness Tests," Circuit World, Vol. 14 Iss: 4, pp.23–26 (1988).
3. Mittal, K.L. “Treatise on Clean Surface Technology,” Vol. 1, p.81, (1987).
4. Hymes, L. “Cleaning Printed Wiring Assemblies in Today’s Environment” (1991).
5. Tegehall, P-E., “Cleanliness and Reliability,” IVF Research Publication 96846 (1996).
6. Crawford, T., “An In-Depth Look at Ionic Cleanliness Testing,” IPC-TR-583 (1993).
This article appeared in the July 2016 issue of SMT Magazine.
