Wetting of Fresh and Aged Immersion Tin and Silver Surface Finishes by Sn/Ag/Cu Solder Minna Arra Flextronics Tampere, Finland Dongkai Shangguan & DongJi Xie Flextronics San Jose, California, USA Abstract The wetting of alternative PCB surface finishes, including immersion silver (I-Ag) and immersion tin (I-Sn), by Sn/Ag/Cu solder and Sn/Pb solder, was studied in this work, along with electroless nickel/immersion gold (Ni/Au) and organic solderability preservative (OSP) finishes for comparison. Evaluation of wetting was carried out with fresh boards and boards subjected to different pre-conditioning treatments which simulated the effects of aging, storage and multiple reflow. Selected conditions consisted of high temperature aging at 155 C for up to 6 and 12 hours, temperature-moisture exposure at 85 C/85%RH for 6, 12 and 24 hours, and reflow treatments between 2 and 4 reflow. Wetting was studied based on the IPC-TM-65 2.4.45 standard, by a wetting bar test, and by a wave soldering test. The results show that when the boards are fresh, the wetting of the I-Sn finish is excellent and comparable to that of the Ni/Au finish, and the wetting of I-Ag is slightly better than that of the OSP finish. However, after the preconditioning treatments, the wetting of the I-Sn finish degrades the fastest, whereas the wetting of the I-Ag and OSP finishes decrease almost at the same rate after different pre-conditioning treatments, while the wetting of the Ni/Au finish remains excellent through all the pre-conditionings treatments. In all cases, the wetting of surfaces is better by the Sn/Pb solder than by the lead-free solder. Introduction The two main reasons to replace the Sn/Pb HASL (hot air solder levelling) PCB finish with I-Sn (immersion tin), OSP (organic solderability preservative), I-Ag (immersion silver) or Ni/Au (electroless nickel/ immersion gold), are the environmental pressure and the higher coplanarity requirements for the fine pitch surface mount assemblies, which is difficult to achieve with HASL. The advantage of Sn/Pb HASL is its high resistance to ageing under high temperature conditions. Alternative finishes, in turn, offer excellent coplanarity but somewhat inferior surface protection. Immersion finishes, developed for single element coatings, result in relatively thin layers (typically less than 1 µm) because the deposition process halts when the substrate surface is completely covered with the coated material 1. 1/9 Three classes of compounds are commonly used as OSP: benzotirazole, imidazole and benzimidazole compounds. OSP forms a chemical bond with the Cu that prevents the latter s oxidation, and the solderability often depends on the adhesion of the OSP layer on the copper. Shelf life of 6-12 months has been documented for office-type storage environments 1. Temperatures of 12-15 C cause OSP coatings to dissipate 1 ; Therefore, if OSP coated Cu is exposed to a process cycle in the absence of flux and solder, the coating may be lost, thereby limiting the use of OSP for products requiring multiple soldering processes. However, solderability protection improves and may remain adequate for the second reflow cycle if a N 2 soldering atmosphere is applied, because the loss of OSP does not result in as strong oxidation of Cu. Ni/Au has an excellent corrosion resistance and solderability, which is retained even after several process passes 1, 2, 3, and is also wire-bondable 2, 3.
When compared with Cu, the IMC (intermetallic compound) layer growth between Sn-based solders and Ni is relatively slow 1. The nickel salts in the barrier layers are carcinogenic and require special waste treatment in the board manufacturing process, which makes them less environmentally friendly 2. An advantage of the use of I-Ag as opposed to the Au and/or Pd containing coatings is a lower material cost. I-Ag finishes are also wire-bondable 3. Ag surface readily oxidizes into Ag 2 O 1. Customarily, the dark tarnish that may form on Ag coatings and reduce its solderability is largely the result of sulfidication 1, 2. In soldering, the Ag layer dissolves into the molten solder; the Ag coating itself does not melt, which may decrease the speed of wetting. Some recent electromigration test results indicate that the migration of Ag is not a concern 4. So far, I-Sn coatings have found the least favour in the electronics industry because of their limited solderability protection. For the Sn finish, layers of Cu 3 Sn (ε-phase), Cu 6 Sn 5 (η-phase) and Sn form on top of the copper layer. The Cu 6 Sn 5 and Cu 3 Sn layers grow in thickness at the expense of Sn and the growth is temperature and time dependent. The formation of Sn/Cu IMC layer finally consumes all the tin during aging and/or the assembly process. Solderability has been claimed to depend both on the quantity of free tin remaining on the surface and on the degree of the oxidation of the Sn finish 5, 6. The mechanism of oxide growth has been found to be accelerated in the presence of moisture since it is catalysed by the presence of water or hydroxyl ions 7. The objective of this study was to investigate the suitability of the alternative surface finishes to the lead-free soldering processes by studying the solderability of both fresh and artificially aged surfaces. Materials and Processes Solder Alloys, PCBs and Components Among the lead-free alloys, Sn/Ag/Cu alloy was selected since it is believed to be the leading choice of the electronics industry for lead-free solder and is recommended by numerous international industry and research consortia. A eutectic 63Sn/37Pb (wt-%), served as the reference alloy. Some properties of the solder pastes are shown in Table 1. Table 1 - Properties of the Solder Pastes Sn/3,9Ag/,6Cu Sn/37Pb Metal content 88 9 (wt-%) Flux ROL REL1 classification per J-STD-4 Particle size 2-38 µm 25-45 µm For wave soldering, the Sn/3,5Ag/,7Cu (wt -%) alloy and a water-based flux (Table 2) were used. Table 2 - Properties of the Wave Soldering Flux Solvent Waterbased J-STD-4 Solids Acid value classification content ORL 3,25 27 A special test vehicle was designed for the study. The overall size of the vehicle was 21,5 cm x 15,1 cm x 1,6 mm. Each board contained four sites for the IPC wetting test, eight sites for the wetting bar test, and six DIL14 component sites for the wave soldering test. Half of the DIL14 components mounted were coated with Ni/Pd and half with Sn/Pb. Besides the wetting tests, the test vehicle contained sites for the assembly of different components, for example CSPs and QFPs down to,5 mm pitch; These sites were not populated, which allowed to monitor the wetting on several different types of pads. Four PCB surface finishes were studied: organic solderability preservative (OSP) of type Entek 16, electroless nickel/ immersion gold (Ni/Au), immersion silver (I-Ag), and immersion tin (I-Sn). The laminate base material of the PCBs was conventional FR-4. The thickness of copper and coating layers are given in Table 3. Table 3 - Thickness of Copper and Coating Layers in the PCBs (Data from the supplier) Measured nominal thicknesses (µm) I-Ag I-Sn OSP Ni/Au Cu: 36- Cu: 36- Cu: Cu: 3-48 5 34-44 44 Ag: Sn: Ni:4,5-,23-,9-1,2 5,6,36 Au:,8 Screen Printing and Reflow Processes Stencils were manufactured with laser-cutting process followed by electropolishing. The screen printer settings were verified by measuring the volume of the paste on specified locations of the board so that a process capability level of 1,33 was achieved. For reflow, a 9-zone forced convection oven was employed. The profiles for the lead-free and Sn/Pbreflow are presented in Figure 1. The peak 2/9
temperature was 224 C in the Sn/Pb profile and 248 C in the Sn/Ag/Cu profile. 3 25 Temperature ( ) 2 15 1 5 1 2 3 4 Time (sec) Sn/Pb Sn/Ag/Cu Figure 1 - Reflow profiles Figure 2 - IPC wetting test pattern Wave Soldering Process For wave soldering in air atmosphere, a solder bath temperature of 27 C in combination with a conveyor speed of 1,6 m/min was used in the wetting tests, giving a contact time of less than 1, sec at the chip wave and ~1, sec at the main wave. Methodology WettingTests Wetting tests consisted of three parts. Two of the tests were carried out with solder paste and one test with a wave soldering process. Wetting Bar Test In the wetting bar test, solder paste was printed on test pads with varying coverage and the board was then sent to reflow. Stencil thickness was kept at 2 µm. The smallest printed paste coverage was 4 % and the largest 12 % of the pad area. This means that the paste was overprinted when the coverage exceeded 1 %. The printed area changed with 5 % intervals as seen from Figure 3. The minimum coverage that allowed the pad to be fully wetted after reflow was recorded. Full wetting was judged, when the solder touched all the edges of the pad; the corners could still be uncovered. Wetting Test IPC-TM-65, 2.4.45 8 In this test, solder paste was printed onto a coated area of PCB through a 2 µm thick stencil having a round opening of 6,5 mm diameter. After printing, the board was sent to reflow. The diameter of the solder after reflow was measured using a stereomicrosope. An example of the test pattern is shown in Figure 2. Figure 3 - Part of the wetting bar test pattern Besides the numeric results gathered from the two wetting tests, defects such as de-wetting and abnormal spreading were inspected on the wetting test areas and other areas such as QFP and CSP pads. 3/9
Wetting Test in Wave Soldering DIL14 type of components were assembled on the boards through the wave soldering process. Regarding the wetting in general, solder joints fulfilled the IPC-A-61 specification 9. Therefore, a more sensitive categorization system was created. The through-hole fill and topside pad wetting of each lead of the six DIL14 components, shown in Figure 4, on every assembled board, were inspected. The results were divided into three different categories. In the first category (Figure 5), the solder had filled the hole and the topside pad was fully wetted by the solder. In the second category (Figure 6), the hole was filled but the topside pad wetting was not complete. In the third and worst case (Figure 7), the solder had not completely filled the hole and the topside pad was not wetted. Figure 4 - DIL14 -Component used for wetting test in wave soldering Figure 5 Category 1: Solder has filled the hole completely and the topside pad is fully wetted Figure 6 - Category 2: Solder has filled the hole completely but the topside pad is not fully wetted Figure 7 Category 3: Solder has not completely filled the hole and topside pad is not wetted Pre-conditioning of the Boards The wetting was studied on fresh boards (with time from the manufacturing to testing less than 2 weeks for I-Sn boards and ~5 weeks for the other boards) and on boards that had undergone different preconditioning treatments to simulate the aging of boards during storage and handling as well as multiple reflow. Three types of pre-treatment conditions were used for the wetting tests with solder paste: a high temperature storage at 155 C, a high temperature-moisture storage at 85 C/85%RH, and multiple reflow treatment for fresh and preconditioned boards. Lead-free reflow profile was used for lead-free solder and Sn/Pb profile for Sn/Pb solder. The selected pre-treatment combinations included: (1) Fresh, no pre-conditioning (2) 155 C, 6 h (3) 155 C, 12 h (4) 85 C/85%RH, 6 h (5) 85 C/85 %RH, 12 h (6) 85 C/85%RH, 24 h (7) Two reflow (8) Four reflow (9) 85 C/85%RH, 6 h, followed by two reflow (1) 155 C, 6 h, followed by two reflow (11) 85 C/85%RH, 12 h, followed by two reflow (12) 155 C, 6 h, followed by one reflow cycle The pre-conditioning #4 (85 C/85%RH, 6 h) is similar to IPC/EIA J-STD-3A (draft) 1 which specifies steam aging to simulate the aging of Sn and Sn/Pb coatings. After the pre-conditioning treatment, solder paste was printed on the boards and the boards were sent to reflow. The sample size for both the IPC wetting test 4/9
and the wetting bar test was minimum 12 for all combinations of pre-treatments and board finishes. The wave soldering part was carried out at a later stage, and therefore the boards had been stored for ~5 months in ambient conditions before testing. Fresh boards and boards undergone one and two lead-free reflow were included in this test. Results and Discussion Wetting Test Results with Sn/Ag/Cu Solder Paste I-Ag The wetting of the I-Ag surface was better than that of OSP but clearly worse than that of I-Sn and Ni/Au in the fresh condition (Figure 8). This is somewhat contradictory to the results from an earlier study, where the wetting of I-Ag surface by a Sn/Ag/Cu solder paste was found to be poorer than that of OSP surface 4. Four reflow before soldering still produced acceptable wetting, which indicates that relatively fresh I-Ag finished boards would be fairly resistant to solderability degradation through multiple lead-free reflow. Wetting Diameter (mm) 7,6 7,4 7,2 7 6,8 6,6 6,4 6,2 6 5,8 Fresh 155, 6 I-Ag Finish 155, 12 85 /85 %RH, 12 Four reflow 155, 6, followed by two reflow 12 1 8 6 4 2 Printed Paste Coverage for Full Wetting (% Wetting bar Wetting diameter Figure 8 - Wetting test results for I-Ag surface The wetting degraded slightly more during the storage at 155 C than during the 85 C/85%RH storage, which correlates well with some earlier test results 11. The pre-conditioning at 155 C for 6 has been estimated to correspond to aging at room temperature for ~15 months 11. As can be seen from Figure 8 and Figure 9, two reflow after the storage at 155 C for 6 resulted in poor wetting. An overprint of paste was needed to wet the whole pad in the wetting bar test and the wetting diameter was smaller. Therefore it can be expected that after 15 months of storage at the room temperature, the I- Ag surfaces are no longer suitable for the assembly processes, where multiple lead-free reflow may be required. Figure 9 - Wetting of QFP1 pads after 6 at 155 C, followed by two reflow (upper), as compared with a fresh board (lower), on I-Ag surface finish I-Sn The wetting of I-Sn surface was excellent and comparable to Ni/Au, in the fresh condition. However, as known from previous studies 5, 11, this finish is very 5/9
sensitive to high temperature exposure. The exposure of fresh boards to two lead-free reflow before soldering resulted in poor wetting bar test results (Figure 1). Therefore, two or more reflow after any pre-conditioning treatment would not provide acceptable results either. The storage at 155 C significantly degraded the solderability of the I-Sn finish (Figure 11), but it survived relatively well the 85 C/85%RH storage. It has been reported that 6 at 155 C or 8 at 85 C/85%RH both correspond to approximately 1,5 years of shelf life and hence provide similar wetting test results 2. However, these assumptions did not correlate with the data obtained in this work. One explanation can be the different abilities of the fluxes used in the tests to clean the Sn-oxide layers, which have been reported to grow rapidly under the presence of moisture 7 during the 85 C/85%RH storage. The degradation during 6 at 155 o C can in turn be linked to the excessive consumption of the Sn layer due to the IMC growth 5. For some unknown reason, the correlation between the two wetting tests was quite weak for I-Sn. Wetting Diameter (mm) 7,6 7,4 7,2 7 6,8 6,6 6,4 6,2 6 5,8 Fresh 155, 6 I-Sn Finish 85 /85 %RH, 12 85 /85 Two reflow %RH, 24 12 1 8 6 4 2 Printed Paste Coverage for Full Wetting (% Wetting bar Wetting diameter Figure 1 - Wetting test results for I-Sn surface Figure 11 - Wetting of QFP28 pads after 6 at 155 C (upper), compared with a fresh board (lower) on I-Sn surface finish Ni/Au Ni/Au was by far the best surface finish for wetting, which remained excellent after all the selected preconditionings treatments (Figure 12). Wetting Diameter (mm) 7,6 7,4 7,2 7 6,8 6,6 6,4 6,2 6 5,8 Fresh 155, 12 Ni/Au Finish 85 /85 %RH, 24 Four reflow 155, 6, followed by two reflow 12 1 8 6 4 2 Printed Paste Coverage for Full Wetting (%) Wetting bar Wetting diameter Figure 12 - Wetting test results for Ni/Au surface OSP The wetting of the OSP surface was the worst among all the tested surfaces. Solder did not spread even on the fresh board surface. When subjected to high temperature or high temperature-moisture aging for six hours followed by two reflows, the wetting results became very poor. For example on the QFP1 pads, the solder pulled back considerably towards the centre of the pad leaving large corner areas exposed (Figure 14). Similarly to the I-Ag surface, the differences in the wetting after different pre - conditioning treatments were relatively small with the OSP surface (Figure 13). 6/9
Wetting Diameter (mm) 7,6 7,4 7,2 7 6,8 6,6 6,4 6,2 6 5,8 Fresh 155, 6 OSP Finish 155, 12 85 /85 %RH, 12 Four reflow 155, 6, followed by two reflow 12 1 8 6 4 2 Printed Paste Coverage for Full Wetting (%) Wetting bar Wetting diameter Sn/Ag/Cu solder. It is widely known that the Sn/Ag/Cu solder generally does not wet as well as the Sn/Pb solder. However, it is to be noted that the difference reported in this study can be due not only to the alloy but also to the different flux in the Sn/Ag/Cu and the Sn/Pb pastes, as well as to the reflow peak temperature above the melting temperature of the alloy (41 C superheat for Sn/Pb vs. 27 C superheat for Sn/Ag/Cu). Interestingly enough, I-Sn boards exhibited good solderability after storage at 85 C/85%RH for 6 hours followed by two Sn/Pb reflows (Figure 16), but the solderability was poor after the storage at 85 C/85%RH for 6 hours followed by two Sn/Ag/Cu reflows, due to the sensitivity of the I-Sn finish to exposure at elevated temperatures. Figure 13 - Wetting test results for OSP surface Wetting bar test 12 1 Paste coverage for full wetting (%) 8 6 4 2 Sn/Pb Sn/Ag/Cu Fresh, I-Ag 85 /85 %RH, 6, followed by two reflow, I-Ag 155, 6, followed by two reflow, I-Ag Fresh, I-Sn 85 /85 %RH, 6, followed by two reflow, I-Sn 155, 6, followed by two reflow, I-Sn Figure 14 - Wetting of QFP1 pads after 6 at 155 C, followed by two reflow (upper), as compared with a fresh board (lower), on OSP surface finish Figure 15 - Wetting bar test results for Sn/Pb solder paste as compared with Sn/Ag/Cu solder paste Wetting Test Results with Sn/Pb Solder Paste Some comparative testing was carried out with the eutectic Sn/Pb solder. The pre -conditioning reflow as well as the soldering reflow were done with the typical Sn/Pb profile. In general, the wetting was considerably better with the Sn/Pb solder than with the Sn/Ag/Cu solder, as can be seen in Figure 15. For I-Ag, the wetting bar results with the Sn/Pb solder were consistently 2-3 % better than those with the 7/9
Figure 16 - Wetting bar test pattern with Sn/Pb solder paste after pre-conditioning for 6 at 85 C/85%RH, followed by two reflow (upper), as compared with a fresh board (lower), on I-Sn surface finish Wetting Tests Results from Wave Soldering I-Ag and I-Sn surfaces were further studied with the wave soldering process. 3 DIL14 components were soldered for each combination of surface finish and pre-conditioning. The inspection results showed that there were no leads in the wetting category 3, which means that in all cases, the acceptability requirements of IPC-A-61C workmanship standard were met. However, leads belonging to category 2 were observed as shown in Figure 17, with the rest of the leads belonging to category 1. % of leads 35 3 25 2 15 1 5 fresh Pre-conditioning 1 reflow cycle 2 reflow I-Ag I-Sn Figure 17 - Percentage of DIL14 leads belonging to wetting Category 2 after different board preconditionings in the wave soldering test The results for I-Sn confirm that the solderability of this surface suffers considerably from reflow treatments. The wetting of I-Ag surface, on the other hand, was not sensitive to the reflow treatments. 8/9 Conclusions In general, the wetting bar test, the solder spreading test, and the wave soldering test, have yielded consistent results. I-Sn and Ni/Au surfaces provided the best wetting results on fresh boards, followed by I-Ag and OSP. Fresh I-Ag boards can withstand multiple lead-free reflow, but after aging for 6 at 155 C, multiple reflows are no longer to be recommended for I-Ag. Neither fresh nor aged I-Sn finished boards can withstand multiple lead-free reflow without significant degradation in wetting. Up to two reflow before wave soldering did not significantly change the through-hole wetting performance of the I-Ag boards, but had a considerable impact on the wetting of the I-Sn boards. Therefore, I-Sn boards should not be used in products requiring lead-free reflow prior to wave soldering. It is important to notice that even though the I-Sn boards did not withstand multiple lead-free reflow, they seem to withstand multiple Sn/Pb reflow quite well, due to the lower temperatures in the Sn/Pb reflow. A detailed surface analysis study is being carried out on intermetallic compound growth and oxidation in order to better understand the solderability degradation mechanism of the surfaces. Acknowledgements The authors would like to thank Hoang Phan at Flextronics San Jose, USA, and David Khoo at Flextronics Melaka, Malaysia, for their assistance in the experimental work. References 1 Vianco, P. T., An Overview of Surface Finishes and Their Role in Printed Circuit Board Solderability and Solder Joint Performance, Circuit World, Vol. 25, No. 1 (1999), pp. 6-24. 2 Bratin, P. et al., Tin finishes, PC Fab, February 2, pp. 1-9. 3 Willis, B., The Changing Face of PCB Solderable Finishes, Electronics Engineering, July 1997, pp. 61-63. 4 Chada, S. et al., Investigation of Immersion Silver PCB Finishes for Portable Product Applications, Proceedings of SMTA International, Chicago, IL, October 21, pp. 64-611. 5 Lamprecht, S. et al., Ageing characteristics of immersion tin surface finishes, Proceedings of SMTA International, Chicago, IL, USA, September, 22.
6 Ormerod, D. H., Immersion Sn as a high performance solderable finish for fine pitch PW Bs, Circuit World, Vol. 26, No. 3 (2), pp. 11-16. 7 Ray, U. et al., Influence of Temperature and Humidity on the Wettability of Immersion Tin Coated Printed Wiring Boards IEEE Transactions on Components, Packaging, and Manufacturing Technology- Part A, Vol. 18, No. 1 (1995), pp. 153-162. 8 Institute for Interconnection and Packaging Electronics Circuits (IPC), IPC-TM-65 Test Methods Manual, Number 2.4.45 Solder Paste Wetting Test, 1995, available at http://www.ipc.org/html/fsresources.htm 9 Institute for Interconnection and Packaging Electronics Circuits (IPC), IPC-A-61C Acceptability of Electronic Assemblies, 2. 1 Electronics Industries Alliance (EIA) and Institute for Interconnection and Packaging Electronics Circuits (IPC), IPC/EIA J-STD-3A, 2. 11 D. Xie et al., Solderability and Process Integration Studies of Immersion Silver and Tin Surface Finishes, Proceedings of SMTA International, Chicago, IL, September 22. "This paper was originally published in the Proceedings of APEX Exhibition and Conference, Anaheim, CA, USA, March 31-April 3, 23". 9/9