Available online at www.sciencedirect.com ScienceDirect Energy Procedia 55 (214 ) 331 335 4th International Conference on Silicon Photovoltaics, SiliconPV 214 The mechanical theory behind the peel test Ulrich Eitner and Li Carlos Rendler Fraunhofer ISE, Heidenhofstraße 2, Freiburg 7911, Germany Abstract The peel test is a very simple and fast method to determine the adhesion of interconnector ribbons to solar cell metallizations. It is part of the solar cell standard DIN EN 5461 and is, due to its ease of use, widely accepted to qualify cell metallizations and the soldering process. In the standard a force of 1 N per mm of joint width is specified but other relevant quantities are missing, for example the peeling angle. We show that this lack of specification enables the manipulation of peel testing results. We therefore apply the mechanical theory of Kinloch [1] where measured peel forces are translated into adhesive fracture energies G A. The fracture energy is a geometry-independent parameter that describes the energy to break the interfacial bondings at the peel front. It incorporates the dimensions of the ribbon and its stress-strain-curve. We perform 86 peel experiments at 9, 135 and 18 of ribbons on continuous front side busbars of cells from one stringing batch. While the median forces for 9 (3.7 N), 135 (2.35 N) and 18 (3.39 N) differ by up to 3.4 % we find the median adhesive fracture energies to deviate by only 17.4 %. Using the same adhesive fracture energy (26 J/m 2 ) for a 45 peel test we expect peel forces of 7.45 N which is factor 2.4 (2.2) higher than the 9 (18 ) peel forces. 214 214 Published The Authors. by Elsevier Published Ltd. This by Elsevier is an open Ltd. access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3./). Peer-review under responsibility of the scientific committee of the SiliconPV 214 conference. Peer-review under responsibility of the scientific committee of the SiliconPV 214 conference Keywords: Peel test; metallization; qualification; ribbon; soldering 1. Introduction The first test to qualify the interconnection of crystalline silicon solar cells after soldering is the peel test. The interconnector ribbons are peeled off from the solar cell measuring the force. This easy and fast method is used to accept or reject new cells in a module production line and to optimize the soldering process of a tabber stringer. Although the test is part of the standard DIN EN 5461, various configurations of the test are possible. * Corresponding author. Tel.: +49 761 4588 5825; fax: +49 761 4588 9825. E-mail address: ulrich.eitner@ise.fraunhofer.de 1876-612 214 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3./). Peer-review under responsibility of the scientific committee of the SiliconPV 214 conference doi:1.116/j.egypro.214.8.96
332 Ulrich Eitner and Li Carlos Rendler / Energy Procedia 55 ( 214 ) 331 335 The weakness of the peel test in terms of consistent testing procedures and equipment has been addressed by Klengel and Wendt [2-4]. Their solution is a peel testing machine specifically designed for peel testing solar ribbons where the ribbon is held down at 9 by a low-wear material to avoid cell cracking. However, for existing peel testing equipment no consistent postprocessing procedures are present in the PV community to compare peel test results from 9 and 18 or other angles. Here, we investigate the impact of the peeling angle to the measured forces by applying the theory of adhesive fracture energies from Kinloch[1]. Fig. 1. Configurations of the peel test for different peel angles α. 2. Theory The geometrical configuration of the peel test is shown in Fig 1. Following the work of Kinloch [1] the adhesive fracture energy G A is derived from an energy balance equation where G ext is the external energy, G S is the stored strain energy in the peeling arm, G T is the dissipated energy by plastic tensile deformation of the peeling arm and G B is the dissipated energy by plastic bending deformation at the peel front. These energies are determined by where F is the measured force, α the peeling angle as illustrated in Fig. 1, b the width of the peeling arm, ε the strain and σ the stress in the peeling arm. In case of plastic deformation G T and the σ-ε relation is expressed by a bilinear work-hardening model, (1) (2) (3) (4) The energy dissipated by plastic bending G B is determined by an iterative method that accounts for the different loading states, such as plastic deformation during bending with elastic deformation during unbending and plastic bending and unbending. The calculation procedure is described in detail in the appendix of [1].
Ulrich Eitner and Li Carlos Rendler / Energy Procedia 55 ( 214 ) 331 335 333 Solar ribbons exhibit a certain bending stiffness as well as plastic deformation which is well visible after peeling them from a solar cell. Therefore, the plastic deformation needs to be taken into account and all energy terms need to be determined. The elastic and plastic parameters E (Young s modulus), ε y (yield strain) and a (work hardening parameter) are extracted from the stress-strain curve of the ribbon. Therefore, a tensile test of the ribbon must be performed in order to apply the procedure to determine the adhesive fracture energy G A. 3. Experimental 3.1. Stress-strain curve of ribbon We perform 5 tensile tests on unsoldered copper ribbons with dimensions 16 μm x 1.6 mm until fracture on a Zwick tensile testing machine. The stress-strain curves are then fitted with the bilinear elasto-plastic model, giving E=85 GPa, ε y =.16 % and a=.5. 6 5 18 peel force F [N] 4 3 2 1 2 4 6 8 1 12 14 Peel path s [mm] Fig. 2. Measured peel forces in 36 peel experiments with an angle α = 18 (right). The error in force F is below.6 N (in red) and one exemplary peel curve is shown in black. 3.2. Peel experiments We perform 36 peel tests at a peel angle of 18, 35 tests at an angle of 9 and 15 tests at an angle of 135. We use only the ribbons on continuous front busbars. The 3 cells are identically soldered in an automated tabber stringer using IR soldering. For all tests the cells are attached to a rigid substrate to avoid cell chipping. In all tests very homogeneous cohesive fracture in the cell metallization is observed and the peel fronts remain in perfect shape without sharp kinks during the tests. Figure 2 shows all 36 peel curves of the 18 tests. For each angle we use every datapoint of every peel curve to create the statistical plot for the measured forces in Figure 3.
334 Ulrich Eitner and Li Carlos Rendler / Energy Procedia 55 ( 214 ) 331 335 Measured peel force F [N] 6 5 4 3 2 1 6 5 4 3 2 1 9 135 18 9 135 18 Fig. 3. Statistical plot (median, upper and lower quartiles and whiskers) of all 87231 measured force values grouped by peel angle (left). Applying the method of Kinloch gives the adhesive fracture energy values grouped by peel angles from the original experiments (right). 3.3. Calculation of the adhesive fracture energies The datapoints of each peel experiment are transformed into values of the adhesive fracture energy by using the model of Kinloch in Matlab taking into account the elasto-plastic parameters E, ε y and a as well as the ribbon dimensions b and h. Each force value of each peel curve is transformed into a corresponding value for the adhesive fracture energy to avoid the distortion of the original values by averaging. The resulting adhesive fracture energies are shown in Figure 3 (right). 4. Results and discussion The median force values differ significantly for 135 (2.35 N) compared to 9 (3.7 N) or 18 (3.39 N). The difference adds up to 3.4 %. The 5 % of the force data between lower and upper quartile (box) at 135 have no overlap with the corresponding boxes for 9 and 18. When transformed into adhesive fracture energies, the boxes for all three peeling angles have a wide overlap and the median values are on the same level. Here, the median adhesive fracture energies differ by 17.4 %. However, we find the scattering in the force data to increase relatively when transformed to the adhesive fracture energies as the range of the whiskers increases. It indicates that the algorithm operates reasonably if the variation in the force values is small. With these experiments we show that the transformation of force values into adhesive fracture energies is possible and allows to compare the peel test results under different angles. Calculating the G A for peel angles below 9 as shown in Fig.4 indicates that much higher peel forces can be measured for the identical adhesive fracture energies. With an adhesive fracture energy of 26 J/m 2 the forces can be increased by a factor of 2.4 from 9 to 45. In practice, by switching to lower peel angles a manufacturer might promote an increased adhesion strength in terms of higher peel forces although the quality of the joint has not improved.
Ulrich Eitner and Li Carlos Rendler / Energy Procedia 55 ( 214 ) 331 335 335 8 7 6 5 4 3 2 1.5 N 1. N 1.5 N 2. N 2.5 N 3. N 3.5 N 4. N 4.5 N 5. N 2 18 16 14 12 1 8 6 4 2 135 9 18 45 45 9 135 18 1 2 3 4 5 6 7 8 Peel force F [N] Fig. 4. Calculation of the adhesive fracture energies over different peel angles for distinct force levels (left). The relation of peel forces to adhesive fracture energies for different peel angles (right). The calculations are only valid for the elasto-plastic material model and for the dimensions of the ribbon used for this investigation. 5. Conclusion We apply the adhesive fracture energy method by Kinloch to solar cells and ribbons and can thereby describe the adhesive forces in a solder joint independent of the peeling angle. The measured forces (median) differ by 1.3 N between 18 and 135 and are homogenized to energy levels of 26 J/m 2 (9 : 257 J/m 2, 135 : 243 J/m 2, 18 : 294 J/m 2 ). According to the theory, a given joint strength in terms of the adhesive fracture energy gives the lowest peel forces at an angle of 135 while for angles below 45 the peel forces will show 2 to 6 times higher values than for 135. Using this information a manufacturer can thus increase or decrease the probability of rejecting novel cell types in his production if his peel test guidelines do not specify the peel angle. This subject is of major importance for qualifying novel technologies that come along with lower adhesion such as plated contacts or conductive gluing. Acknowledgements The authors like to thank T. Barp for performing the peel experiments as well as T. Fischer and A. Riethmüller from Teamtechnik for providing the soldered solar cells and ribbons. References [1] A.J. Kinloch, C.C. Lau and J.G. Williams, "The peeling of flexible laminates", International Journal of Fracture 66, 45-7 (1994). [2] R. Klengel, R. Härtel, S. Schindler, D. Schade, B. Sykes, "Evaluation of the Mechanical Strength of Solar Cell Sol-der Joint Interconnects and Their Microstructural Properties by Developing a New Test and Inspection Equipment", Proceedings of the 26th European Photovoltaic Solar Energy Conference, 157-1511 (211). [3] R. Klengel, M. Petzold, D. Schade, B. Sykes, "Improved testing of soldered busbar interconnects on silicon solar cells", Proceedings of 18th European Microelectronics and Packaging Conference (EMPC), 211. [4] J. Wendt, M. Träger, R. Klengel, M. Petzold, D. Schade, R. Sykes, "Improved quality test method for solder ribbon interconnects on silicon solar cells", 12th IEEE Intersociety Conference on Thermal and Thermomechanical Phenomena in Electronic Systems (ITherm), (21).