Vailable online at www.sciencedirect.com Energy Procedia 27 (2012 ) 664 669 SiliconPV 2012, 03-05 April 2012, Leuven, Belgium Intermetallic Phase Growth and Reliability of Sn-Ag-Soldered Solar Cell Joints P. Schmitt *, P. Kaiser, C. Savio, M. Tranitz, U. Eitner Fraunhofer ISE, Heidenhofstr. 2, 79110 Freiburg, Germany Abstract Soldering is the most common way to interconnect solar cells into strings to form a PV module. The most prevalent and standard solder material is the SnPbAg-solder. Since lead is a hazardous material and the RoHS-guideline is expected to prohibit its further use in PV, the industry is looking for an alternative. In electronics there are a number of different alloy options. Promising candidates are the SnAg-solders, which are already in use by several PV manufacturers. The growth of intermetallic phases in the joint, both during soldering and lifetime, affects the quality of the joint and its reliability in terms of mechanical behavior. In this work we show the growth of different intermetallic phases, investigate their detrimental effects on the long term stability and compare them to a standard leaded solder. We produce solder joints and let them undergo an accelerated aging at elevated temperatures in an inert nitrogen atmosphere. We use different optical instruments including microscopy, SEM and EDX, as well as peel tests to investigate the diffusion phases and their influences. We show that intermetallic compounds (IMC) form due to diffusion in the solder joints with different intensity for both solders at elevated temperatures. Our study also poses that SnAg3.5 outperforms the leaded solder in terms of adhesion and durability. The results indicate that the failure mode and the adhesion are depending on the amount of diffused tin into the metallization. 2012 Published by by Elsevier Ltd. Ltd. Selection and and peer-review under under responsibility of the of scientific the scientific committee of the committee SiliconPV of 2012 the SiliconPV conference. 2012 Open conference access under CC BY-NC-ND license. Keywords: Solar Cells; Interconnection; Soldering; Intermetallics; Reliability; Lead free; Diffusion 1. Introduction SnAg3.5-solder is an alloy which is often argued to be a good alternative to leaded solders. It has a melting point of approximately 221 C, which is around 40 K higher than leaded solders [1]. Consequently, higher thermo-mechanical stress is induced in the solar cell after cooling down to room * Corresponding author. Tel.: +49-761-4588-5543; fax: +49-761-4588-9000 E-mail address: peter.schmitt@ise.fraunhofer.de 1876-6102 2012 Published by Elsevier Ltd. Selection and peer-review under responsibility of the scientific committee of the SiliconPV 2012 conference. Open access under CC BY-NC-ND license. doi:10.1016/j.egypro.2012.07.126
P. Schmitt et al. / Energy Procedia 27 ( 2012 ) 664 669 665 temperature. Intermetallics arise in the joints due to diffusion processes and have a crucial impact of the quality of solder joints and their reliability [1]. Prevalent intermetallics in solder joints are Cu 3 Sn, Cu 6 Sn 5 and Ag 3 Sn. These diffusion-based intermetallic compounds have a brittle mechanical behavior, which is detrimental to the stability of the joint [2]. The formation of the phases mostly depends on the composition of solder, temperature and time [3]. The purpose of this work is to identify the diffusion phases, quantify their growth before and after isothermal exposure for accelerated aging and evaluate the reliability of SnAg-joints compared to a leaded solder. 2. Experimental 2.1. Production of solder joints We produce joints with the SnAg3.5-solder and a standard leaded SnPb36Ag2-solder as a reference. Standard 156 mm 156 mm multi-crystalline silicon cells with a backside aluminum metallization and two silver busbars are tabbed with a manual contact solder platform at Fraunhofer ISE. The busbars are screen printed and an alcohol-based no-clean flux is used. The higher melting point of the SnAg3.5-solder requires higher process temperatures resulting in a solder temperature of 270 C for the lead-free solder, as opposed to 228 C for the leaded solder. 2.2. Accelerated aging To avoid oxidation processes in the joints during accelerated aging, we expose the samples to a nitrogen atmosphere in the temperature chamber. The choice of the aging temperatures and times is based on the van t-hoff-rule and Arrhenius equation, which states that reactions are faster at elevated temperatures [4, 5]. The joints are exposed to an isothermal aging at a temperature of 130 C for 8.5 h, 42.8 h and 85.6 h. Based on the theoretical model, these times at 130 C correspond approximately to 2 a, 10 a and 20 a at 20 C. 2.3. Joint analysis The joints are analyzed by metallographic methods and peel tests. The methods used in this work are adapted processes conforming to the specific requirements of soldered solar cells we have presented in previous publications [6, 7]. We investigate the cross-sections of the soldered solar cells (including metallization, solder and interconnector) with light microscopy and scanning electron microscopy (SEM) with energy dispersive X-ray spectroscopy (EDX). These techniques reveal the existence of diffusion and intermetallic phase growths with their thicknesses, wetting, voids and cracks. Additional peel tests [8] are performed to measure the adhesion of the joints and to examine the breakage mode. 3. Results Peel tests reveal that the adhesion of all the joints decreases after aging, as shown in Fig. 1. The SnAg3.5-solder exhibits higher adhesion than the SnPb36Ag2-solder and a steady decrease in adhesion from around 3 N to 1 N with respect to a 2 mm wide interconnector. The initial adhesion of SnPb36Ag2 is lower and decreases strongly after the first aging step from 2 N to less than 0.5 N.
666 P. Schmitt et al. / Energy Procedia 27 ( 2012 ) 664 669 4 front side back side 3 adhesion [N] 2 1 0 initial 8.5h 42.8h85.6h -- initial 8.5h 42.8h85.6h SnAg3.5 SnPb36Ag2 Fig. 1. Adhesion measured by peel tests after heat storage at 130 C for 8.5 h, 42.8 h and 85.6 h for SnPb36Ag2 and SnAg3.5-solder materials in comparison to initial samples. Shown is the absolute adhesion force for a 2 mm wide copper ribbon and the standard deviation In the cross-sections, different diffusion zones can be found at the interface of copper and solder and between solder and cell metallization. The copper of the interconnector interacts with the tin of the solder to form Cu 3 Sn and Cu 6 Sn 5. Additionally, the silver interacts with the tin to form Ag 3 Sn [9]. Fig. 2 shows a light microscopy image of a cross-section visualizing the different diffusion zones. interconnector solder CuSn-phase 1000x metallization wafer AgSn-phase 10μm Fig. 2. Microscopic image of SnAg3.5-solder after 85.6 h of accelerated aging, magnification 1000x. Cu 3Sn- and Cu 6Sn 5-phases are formed at the transition from copper to the SnAg3.5-solder. A spacious diffusion zone of tin in the metallization is formed The diffusion zones can be seen on cross-section with an optical microscope and SEM. EDX determines the allocation of elements. Fig. 3 shows an EDX mapping of tin in the solder, in the metallization and in a zone between solder and interconnector. interconnector CuSn-phase 2500x solder AgSn-phase metallization wafer 10μm EDX element map: tin Fig. 3. SEM image from SnPb36Ag2-solder after 85.6 h accelerated aging with a magnification of 2500x. Cu 3Sn- and Cu 6Sn 5-
P. Schmitt et al. / Energy Procedia 27 ( 2012 ) 664 669 667 phases are formed at the transition from copper to the solder. A large amount of tin is found in the metallization. EDX element determination in the right image shows the allocation of the tin in the joint as red dots in the order of 0.2 % mass fraction Small diffusion zones can be found after soldering in the initial samples which increase with time of aging. The thickness of the intermetallics in the initial state is higher in SnAg3.5 than in SnPb36Ag2. However, after aging it is higher in SnPb36Ag2. Fig. 4 shows the intermetallic layer thickness of the CuSn-phase over aging time of the interface interconnector/solder. layer thickness [μm] 4 3 2 1 initial 8.5h 42.8h 85.6h 0 front side back side -- front side back side SnAg3.5 SnPb36Ag2 CuSn-phase Fig. 4. Growth of the intermetallic layer thickness at the interface copper/solder after heat storage at 130 C for 8.5 h, 42.8 h and 85.6 h for SnAg3.5- and SnPb36Ag2-solder materials in comparison to initial samples. Shown is the layer thickness in μm and the standard deviation In Fig. 5 the growth of the AgSn-phase between solder and cell metallization is shown. A large diffusion zone is measured after the heat storage for 85.6 h in the leaded sample. In comparison to Fig. 3 the migration zone of tin reaches down to the wafer. layer thickness [μm] 16 14 12 10 8 6 4 initial 8.5h 42.8h 85.6h 2 0 front side back side -- front side back side SnAg3.5 SnPb36Ag2 AgSn-phase Fig. 5. Growth of the diffusion layer thickness at the solder/metallization by heat storage at 130 C for 8.5 h, 42.8 h and 85.6 h for SnAg3.5- and SnPb36Ag2-solder materials in comparison to initial samples. Shown is the layer thickness in μm and the standard deviation
668 P. Schmitt et al. / Energy Procedia 27 ( 2012 ) 664 669 Conclusion We show that intermetallic compounds form due to diffusion in the solder joints with different intensity for both solders at elevated temperatures. The intermetallics appear mainly at the interface of copper/solder and solder/metallization and decreases the performance of the joints in terms of mechanical stability [1]. The main failure mode of the joint which we find in our experiments is an adhesive paste breakage between metallization and silicon. This failure possibility is increased after accelerated aging where an increased diffusion of tin into the metallization is observed and decreased adhesion is measured. This behavior is exacerbated in the leaded solder compared to the SnAg3.5-solder. Our results are in good agreement with the explanation of adhesive failure given in [10]. There, the diffusion of tin into the metallization is made responsible for the paste breakage. The tin leads to a swelling in the porous thick film metallization [10]. The interface of the swollen metallization and the wafer is under stress and tends to break. A degradation of the joints leads to a lower fill factor and inactive cell areas [11]. Our study shows that SnAg3.5 outperforms the leaded solder in terms of adhesion. The results indicate that the failure mode is depending on the amount of diffused tin into the metallization. SnAg3.5-solders also have a slower diffusion rate, are non-toxic and are, therefore, regarded as ideal substitutes for the standard leaded solders. Further tests will determine whether tin is able to diffuse into the wafer as reported in [12]. Reliability tests should be extended to temperature cycle tests with minimum temperatures under 13 C to evaluate the impact of the alpha-modification of tin and the effect of tin pest under the influence of Ostwald ripening. Acknowledgements This work was supported by the German Federal Ministry for the Environment, Nature Conservation and Nuclear Safety (Contract Number 0325082). The Authors would like to thank Jutta Zielonka for conducting the SEM-images. The data is based partially on the Bachelor Thesis of Pablo Kaiser with the title Untersuchung der metallischen Diffusion an bleifrei gelöteten Solarzellen, Fraunhofer ISE and Albstadt-Sigmaringen University, 2011. References [1] A. Grusd and A. Miric, "Lead-free Alloys", Soldering & Surface Mount Technology, vol. 10, pp. 19-25, 1998. [2] R. J. K. Wassink, Soldering in Electronics, 2nd ed. Ayr: Electrochemical Publications Ltd., 1998. [3] J. W. Evans, A Guide to Lead-free Solders, 1 ed. London: Springer Science+Business Media Ltd., 2007. [4] A. F. Holleman, E. Wiberg, and N. Wiberg, Lehrbuch der anorganischen Chemie, 101 ed. Berlin, Germany: Walter de Gruyter, 1995. [5] M. Hongtao, et al., "Isothermal aging effects on the dynamic performance of lead-free solder joints", presented at the Electronic Components and Technology Conference. ECTC 2009. 59th, San Jose, CA, USA, 2009. [6] P. Schmitt, et al., "Metallographic Preparation of Solar Cell Samples for Quality Assurance and Material Evaluation", Energy Procedia, vol. 8, pp. 402-408, 2011. [7] D. Eberlein, P. Schmitt, and P. Voss, "Metallographic Sample Preparation of Soldered Solar Cells", Practical Metallography, vol. 2011, pp. 239-260, 2011. [8] J. Wendt, et al., "Improved quality test method for solder ribbon interconnects on silicon solar cells", in Thermal and Thermomechanical Phenomena in Electronic Systems (ITherm), 2010 12th IEEE Intersociety Conference on, 2010, pp. 1-4.
P. Schmitt et al. / Energy Procedia 27 ( 2012 ) 664 669 669 [9] G. Cuddalorepatta, et al., "Durability of Pb-free solder between copper interconnect and silicon in photovoltaic cells", Progress in Photovoltaics: Research and Applications, vol. 18, pp. 168-182, 2010. [10] J. T. Borenstein and R. C. Gonsioraski, "Method for Forming Solar Cell Contacts and Interconnecting Solar Cells", United States of America Patent, 1993. [11] J. Moyer, et al., "The role of silver contact paste on reliable connectivity systems", presented at the 25th European Photovoltaic Solar Energy Conference and Exhibition, Valencia, Spain, 2010. [12] T. H. Yeh, S. M. Hu, and R. H. Kastl, "Diffusion of Tin into Silicon", Journal of Applied Physics, vol. 39, pp. 4266-4271, 1968.