Reliability Assessment of Immersion Silver Finished Circuit Board Assemblies Using Clay Tests

Reliability Assessment of Immersion Silver Finished Circuit Board Assemblies Using Clay Tests Yilin Zhou Research Lab of Electric Contacts, Automation School Beijing University of Posts and Telecommunications Beijing, 100876, P. R. China, E-mail: ylzhou@bupt.edu.cn Michael Pecht CALCE Electronic Products and Systems Center University of Maryland, College Park, 20742, USA E-mail: pecht@calce.umd.edu Abstract This paper discusses issues regarding the reliability of immersion silver (ImAg) surface finish on printed circuit boards (PCBs) in high-sulfur environments with a focus on creep corrosion. The test approach used clay with sulfur to drive corrosion. It was found that silver sulfide had formed on the ImAg surfaces, but dendrite-shaped creep corrosion products were formed on the edges of the ImAg finished copper traces by galvanic corrosion. The creep distance and the content of CuS and Ag 2 S in the corrosion products were dependent on the relative location of the pad edges and the solder mask, which caused different extents of coverage of ImAg finish on the copper traces. A Weibull distribution with two parameters was used to analyze the length of the creep corrosion products, the values of which were used along with corrosion probability to assess the corrosion resistance of ImAg finished PCBs. Key words: Immersion silver, sulfur, creep corrosion, Weibull distribution I. INTRODUCTION Surface finishes are commonly used to protect the copper metallization of a PCB from oxidation to enable good solderability of the PCB [1]. With lead-free legislation (e.g., the EU s Restrictions on Hazardous Substances (RoHS)) impacting PCB manufacturing on a global scale, the ImAg finish is rapidly gaining popularity as the lead-free surface finish of choice. Studies of high temperature aging show that silver coatings can remain solderable for up to 12 months prior to assembly [2], [3]. The manufacturing costs of plating ImAg are half the price of electroless nickel/immersion gold (ENIG) [4]. Silver is compatible with most assembly processes and is becoming more commonly available in the electronics industry. A high rate of failure was recently found in computers with ImAg finished boards in an environment containing high levels of elemental sulfur. Many industries release elemental sulfur into the air. For example, the clays used in the prototype modeling of vehicles, paper mills where sulfur is used in the bleaching process, and power plants where geothermal sources are used to turn steam turbines. Corrosion was found mostly inside the VIA holes because copper remained exposed inside the VIA holes due to incomplete silver finish covering. The corrosion products crept over the VIA holes and caused shorts in the circuits [5]. Schueller [6] selected clays containing 30 50% elemental sulfur and heated the clays with water to simulate an actual environment. Creep corrosion was found on different finishes depending on the weight and heating times of clays. However, the corrosion mechanism and the evaluation methods for ImAg finish were not discussed deeply. This study explains the failure mechanisms of ImAg finished circuit board assemblies in high sulfur environments. By the use of corrosion probability of ImAg finished pads and the statistics of the length of corrosion products, the reliability of ImAg finished PCBs in high sulfur environments will be assessed. II. CORROSION MECHANISMS A silver finish is corroded by direct chemical corrosion in a sulfur environment: 2Ag+S Ag 2 S (1) The electrochemical corrosion that occurs on ImAg finished PCBs is galvanic corrosion. The standard electrode potential of Ag is 0.800 V, and that of Cu is 0.340 V. Ag contacts with Cu to form a conductive circuit. Water from the atmosphere is present in monolayer quantities, which is enough to promote the reactions. Ag finish becomes the cathode, and Cu becomes the anode. The reactions are as shown below. CuS is finally formed [5]. Anode reaction: Cu Cu 2+ +2e - (2) Cathode reaction: S+2e - S 2- (3) The total reaction becomes: Cu+S CuS (4) Creep corrosion is driven by the concentration gradients of the corroded product, so the chemical species move from areas with a higher concentration to those with a lower concentration [7], [8]. Creep corrosion products on ImAg finishes appear to begin with the growth of dendrites that take place equally in all directions, unlike the electric potential-driven dendrite growth of silver migration [6]. If it were in a high sulfur and high humidity environment, the exposed copper would form Cu 2+ and disperse in the condensed water film on the PCB surface. Then the copper ions would react with the sulfur to form copper sulfide. This corrosion product has semiconductor properties [9]. As the corroded products creep, the insulation resistance between adjacent pads on the PCB decreases until functional shorting occurs. 978-1-4244-4905-7/09/$25.00 2009 IEEE 1212

III. TEST METHODS To study creep corrosion on ImAg finishes and assess the quality of PCBs, it is necessary to devise a test method to drive creep corrosion and develop an evaluation method to assess the creep corrosion resistance of ImAg finished PCBs. A. Experimental samples Three ImAg finished FR-4 PCBs were cut into 5 3 cm testing samples. All of the samples were ultrasonically cleaned in deionized water for 30 min and were dried before the corrosion experiments. The thickness of the ImAg finish was measured with an X-Ray Fluorescent Coating Thickness Gauge. Five pads were randomly measured on each PCB. The average thicknesses of the ImAg finish on PCBs No. 1, No. 2, and No. 3 were 0.36 μm, 0.22μm and 0.21μm, respectively. Their standard deviations were all approximately 0.04μm. B. Corrosion test method Referring to Schueller s experiment, Industrial Hard Styling Clay (J-525) made by the Chavant Company was used in this experiment. This clay is used to design prototypes of vehicles [6]. The content of sulfur in this clay was measured by an X-ray energy dispersive spectroscope (XEDS) and was about 17% (weight percentage). Its softening point is 57 C~66 C. To make the clay soft and workable, 700 g of clay was put in a plastic container with 10 ml of water, the lid was closed, and the clay was heated for 5 min in a microwave oven. Then, the PCBs to be tested, which had been stored in a refrigerator for 10 min to enhance condensation, were put into the container and the lid was closed immediately. All of the samples were held by clamps, and none directly contacted the clay. The relative humidity in the container was close to 100% and the temperature was room temperature. The corrosion experiment lasted for 3 days uninterrupted. After 3 days, the relative humidity in the container was 75%. IV. EXPERIMENTAL RESULTS Figure 1. PCB No.3 after corrosion TABLE I. THE ATOMIC PERCENTAGE OF ELEMENT COMPOSITIONS ON IMAG FINISHED PADS BEFORE AND AFTER CORROSION PCB No. 2 Before After corrosion corrosion C 30.07 34.44 O 7.08 11.99 S 0.78 5.96 Cu 45.94 32.49 Ag 16.14 15.11 Many intrinsic pores and defects were found on the PCB pads. However, no creep corrosion product was found on the pad surfaces after corrosion, which indicated that the thin Ag finish still covered the defects and the inside of the pores. This explains why galvanic corrosion did not occur, as shown in Figure 2. B. Creep corrosion Creep corrosion products like dendrites occurred on the edges of the ImAg finished PCB pads after 3 days of corrosion. The dendrites grew from the pad edges and crept outside. According to the statistical results, the longer creep corrosion products formed at the boundary between the ImAg pad edges and the solder mask. Some were hundreds of micrometers long if the pad edges were covered by the solder mask. A backscattered electron image is shown in Figure 3a. If the pad edges were not covered by the solder mask, the dendrites that grew from the ImAg pad edges were shorter, only several micrometers, as shown in Figure 3b. Pores A. Corrosion of ImAg finish by sulfur The color of all of the ImAg pads changed to dark red after three days of corrosion by clay. The corrosion on the ImAg finished pads changed more strongly on the bottom than on the top of the PCBs, since the bottom of the PCB was closer to the clay sulfur source. PCB No. 3 is shown in Figure 1. The element compositions on the ImAg finished pads were detected by XEDS and are listed in TABLE I. The content of the sulfur increased from 0.78% to 5.96% after corrosion, which proved that the ImAg finish was sulfurized and that the dark red corrosion product was silver sulfide. Figure 2. Defects No creep corrosion was found on the pores and defects on PCB No. 2 after corrosion The compositions of the dendrites as detected by XEDS are listed in TABLE II. The main corrosion products were CuS and a little Ag 2 S. C came from organic materials evaporated from the clays. O, Si, and Ba were PCB materials. It was found that the ratio of Cu:Ag was 2.5:1 in the longer dendrites, but the ratio was 11:1 in the shorter dendrites. These results will be discussed later. 1213

Figure 3. Creep corrosion products (a) Long creep corrosion products Pad Solder mask Creep corrosion products Solder mask (b) Short creep corrosion products Backscattered electron images of the creep corrosion products No. of PCB 1 2 3 No. of pads with creep corrosion 7 18 13 No. of pads within 1.5 cm of the 80 130 130 bottom Occurrence Probability 8.75% 13.85% 10.00% B. Lengths of creep corrosion products The lengths of the creep corrosion products on PCB pads were measured from the root to the end of the dendrite. The accumulation probability of the occurrence of various lengths of creep corrosion products on the three PCBs is shown in Figure 4, which was fitted by a two parameter Weibull distribution. The confidence levels of the three plots were all higher than 95%. The lengths of 79, 256, 155 dendrites on three PCB samples were measured respectively. The longest dendrite was 217.83μm, and the shortest dendrite was 2.19μm. The mean lengths of the creep corrosion products on PCBs No. 1, 2, and 3 were 23.5μm, 47.0μm, and 58.3μm, respectively, as listed in TABLE IV. The mean length of creep corrosion products on PCB No.1 was only 50% of that on PCB No. 2 and 40% of that on PCB No. 3. According to the accumulation probability of the occurrence of various lengths of creep corroded products on the three PCBs, it can be deduced that PCB No.1 had the highest corrosion resistance. PCB No.2 was a little better than PCB No.3, which was not consistent with the order of creep corrosion probability. TABLE II. ELEMENT COMPOSITIONS ON CREEP CORROSION PRODUCTS (ATOMIC PERCENTAGE) Elements Longer dendrites Shorter dendrites C 54.27 64.66 O 11.23 13.44 Si 1.31 1.02 S 10.88 2.23 Cu 13.66 16.36 Ag 5.45 1.48 Ba 3.20 0.81 V. ASSESSMENT OF CREEP CORROSION RESISTANCE OF IMAG FINISHED PCBS Creep corrosion products can cause shorts in circuits, so the occurrence probability of creep corrosion and the lengths of dendrites were used to qualitatively assess the corrosion resistance of ImAg finished PCBs. A. Creep corrosion probability Creep corrosion did not occur on all of the ImAg finished PCB pads. The occurrence probability of creep corrosion on the PCBs was defined as the number of pads with creep corrosion divided by the total number of pads within 1.5 cm of the bottom of the PCB samples, as listed in TABLE III. PCB No. 1 had an 8.75% probability of forming creep corrosion, which was lower than PCB No.2, 13.85%, and PCB No. 3, 10.00%. Therefore, PCB No. 1 had the best corrosion resistance. Figure 4. Accumulation probability and Weibull fitting of the length of the creep corroded products on the three PCBs TABLE IV. LENGTH OF CREEP CORROSION PRODUCTS ON PCB PADS No. of PCB 1 2 3 No. of Dendrites 79 256 155 Max. dendrite length (μm) 74.10 170.65 217.83 Min. dendrite length (μm) 4.18 2.19 7.91 Mean length of creep corrosion products (μm) 23.5 47.0 58.3 TABLE III. PROBABILITY OF OCCURRENCE OF CREEP CORROSION ON PCBS 1214

VI. ANALYSIS AND DISCUSSION A. The influencing factors on the creep corrosion products Chemical corrosion occurred on the ImAg finish surface in a high sulfur environment. Ag 2 S was formed, which caused the color of the pads to change to dark red, but this corroded product did not creep. The corrosion products of dendrites formed on the pad edges and depended on the relative location between pad and solder mask. Since the mechanism of the creep corrosion was galvanic corrosion, only if the silver finish could not fully cover the copper trace would the exposed copper be corroded in a high sulfur and high humidity environment. The copper exposed on the pad edges lost electrons to form copper cations in the water film condensed on the PCB surface, and the cations dispersed outside to form copper sulfide with sulfur evaporated from the clays. The PCB was covered by solder mask first and then was coated by immersion silver in the manufacturing process. Since the solder mask was not covered on the pad edges, the silver finish was coated on the pad edges, and a few areas of copper were exposed to form short dendrites. But if the solder mask was covered on the pad edges, the copper along the boundary between the pad and the solder mask was fully exposed to form long dendrites. And as the amount of Cu 2+ diffused and crept outward, the Ag on the surface migrated and floated on the crept Cu 2 S, which was also corroded to form Ag 2 S by direct chemical corrosion. This is why the content of Ag in the long dendrites was much higher than that in short dendrites. B. Assessment methods for creep corrosion resistance Since the fact that the ImAg finish did not fully cover the copper trace is the root reason behind the creep corrosion, increasing the thickness of ImAg finish should reduce the probability and areas of exposed copper in order to decrease the amount of creep corrosion. The experimental results proved that the probable occurrence and the size of creep corrosion on PCB No.1, where the thickness of ImAg was double of that on PCBs No. 2 and No. 3, was lower than the others. On the other hand, although the thicknesses of ImAg finish on PCB No. 2 and No. 3 were similar, the fact that the solder mask covered the pad edges on PCB No. 3 more than it did on PCB No. 2 caused the mean length of the dendrites on PCB No. 3 to be longer than the mean length of the dendrites on PCB No.2. Shorting in the circuit is the failure mode, so the length of creep corrosion products should be taken as the main assessment factor, which can be combined with creep corrosion probability. Therefore, PCB No. 3 had the worst creep corrosion resistance. VII. CONCLUSIONS (1) Heating the clay with water can be used as simulation method to drive sulfur corrosion on ImAg finished PCB. (2) The ImAg finish surface was sulfurized. The dendrite-shaped creep corrosion products were formed by galvanic corrosion on ImAg finished pad edges under a high sulfur and high humidity environment. CuS was used as the main component of creep corrosion product, accompanied by a little Ag 2 S. (3) The occurrence probability and the length of the creep corrosion products were dependent on the coverage of ImAg finish on the copper trace. It was found that the occurrence probability and the length of the creep corrosion products were reduced if the thickness of the ImAg finish was increased and the solder mask was separated from the pad edges. (4) A two-parameter Weibull distribution can be used to calculate the mean length of creep corrosion products, which can be taken as the main assessment factor of creep corrosion resistance for ImAg finished PCBs. REFERENCES [1] Ganesan S. and Pecht M., Lead-Free Electronics, John Wiley and Sons, Inc., New York, 2006. [2] Lopez E., Vianco P., Buttry R., Lucero S., and Rejent J., Effect of Storage Environments on the Solderability of Immersion Silver Board Finishes with Pb-Based and Pb-Free Solders, SMTA Journal, Vol. 18, Iss. 4, pp. 19-28, Oct. 2005. [3] Vianco P., Lopez E., Buttry R., Kilgo A., and Lucero S., Effects of Accelerated Storage Environments on the Solderability of Immersion Silver-Coated Printed Circuit Boards, Pan Pacific Symposium, Hawaii, Jan. 17, 2006. [4] United States Environmental Protection Agency, Alternative Technologies for Surface Finishing: Cleaner Technologies for Printed Wiring Board Manufacturers, Technical Report, EPA-744-R-01-001, pp. 1-42, June 2001. [5] Mazurkiewicz P., Accelerated Corrosion of Printed Circuit Boards Due to High Levels of Reduced Sulfur Gasses in Industrial Environments, Proceedings of the 32nd International Symposium for Testing and Failure Analysis, pp. 469-473, Austin, TX, Nov. 12-16, 2006. [6] Schueller R., Creep Corrosion on Lead-free Printed Circuit Boards in High Sulfur Environments, SMTA International Conference, Orlando, FL, pp. 643-654, Oct. 11, 2007. [7] Zhao P., Pecht M., Field Failure Due to Creep Corrosion on Components with Palladium Pre-plated Leadframes, Microeletronics Reliability, Vol. 43, No. 5, pp. 775-783, 2003. [8] Zhao P., Pecht M., Mixed Flowing Gas Studies of Creep Corrosion on Plastic Encapsulated Microcircuit Packages with Noble Metal Pre-plated Leadframes, IEEE Transactions on Device and Materials Reliability, Vol. 5, No. 2, pp. 268-276, June 2005. [9] Xu C., Flemming D., Demerkin K., Corrosion Resistance of PWB Surface Finishes, Apex, Los Angeles, CA, Feb. 20-22, 2007. Yilin Zhou, Automation School, Beijing University of Posts and Telecommunications, internal box 71, Beijing, 100876, P. R. China, Email: ylzhou@bupt.edu.cn, Tel/Fax: 86-10-62283173 She received the B.S. degree in mechanical and electrical engineering and the Ph.D. degree in circuits and systems from Beijing University of Posts & Telecommunications, China, in 1994 and 1999, respectively. She is now an associate professor in the Automation School, Beijing University of Posts & Telecommunications. She worked in the Center for Advanced Life Cycle Engineering (CALCE), University of Maryland, College Park as a visiting scholar from Aug. 2007 to Aug. 2008. She is interested in the study of contact failure caused by contamination and fretting, quality analysis of gold plating for connectors, and the mechanism and application of lubricants 1215

for electric contacts. She has published more than 30 papers. Michael Pecht, Visiting Professor in Electrical Engineering City University of Hong Kong, and Director CALCE Electronics Products and Systems Center, Room 1103, Building 089, University of Maryland, College Park, 20742, USA, Tel.: +301-405-5323; Fax: +301-405-9269, E-mail address: pecht@calce.umd.edu Prof Michael Pecht is currently visiting Professor in Electrical Engineering at City University of Hong Kong. He has an MS in Electrical Engineering and an MS and PhD in Engineering Mechanics from the University of Wisconsin at Madison. He is a Professional Engineer, an IEEE Fellow, an ASME Fellow and an IMAPS Fellow. He was awarded the highest reliability honor, the IEEE Reliability Society s Lifetime Achievement Award in 2008. He served as chief editor of the IEEE Transactions on Reliability for eight years and on the advisory board of IEEE Spectrum. He is chief editor for Microelectronics Reliability and an associate editor for the IEEE Transactions on Components and Packaging Technology. He is the founder of CALCE (Center for Advanced Life Cycle Engineering) at the University of Maryland, College Park, where he is also the George Dieter Chair Professor in Mechanical Engineering and a Professor in Applied Mathematics. He has written more than twenty books on electronic products development, use and supply chain management and over 400 technical articles. He has been leading a research team in the area of prognostics for the past ten years. He has consulted for over 100 major international electronics companies, providing expertise in strategic planning, design, test, prognostics, IP and risk assessment of electronic products and systems. He has previously received the European Micro and Nano-Reliability Award for outstanding contributions to reliability research, 3M Research Award for electronics packaging, and the IMAPS William D. Ashman Memorial Achievement Award for his contributions in electronics reliability analysis. 1216