Fabrication of Smart Card using UV Curable Anisotropic Conductive Adhesive (ACA) Part II: Reliability Performance of the ACA Joints

Transcription

1 Fabrication of Smart Card using UV Curable Anisotropic Conductive Adhesive (ACA) Part II: Reliability Performance of the ACA Joints C. W. Tan, Y M Siu, K. K. Lee, *Y. C. Chan & L. M. Cheng Department of Electronic Engineering 83 Tat Chee Avenue, Kowloon, Hong Kong * EEYCCHAN@cityu.edu.hk Abstract For mobile electronic products, there are different mechanical or environmental attacks like warps, stretching, changing in humidity and temperature, which may affect the performance and reliability of the electronic products. In this work, the reliability of the smart cards fabricated by chip-onflex (COF) bonding using UV curable Anisotropic Conductive Adhesive (ACA) was investigated. Based on the aforementioned attacks, successfully fabricated contactless smart card samples were subjected to autoclave test to investigate the failure mechanisms. The reading distance between the card reader and the sample was measured and compared at each test interval. Besides, shear test was performed to determine the shear strength and the effect of autoclave test to ACA joint. Fracture surfaces were then analyzed using Scanning Electron Microscope (SEM). ACA joints selected from the aged samples were cross-sectioned and examined using the SEM. Bare ACA specimens that were cured under the same curing conditions were also put into autoclave test chamber simultaneously. Chemical analyses were performed using FTIR-ATR analysis to investigate the chemical structure change in these UV curable ACA samples. These results were then collaborated to have a better understanding to the failure mechanism of ACA joints of the smart card that have undergone autoclave test. {keywords} autoclave test, chip-on-flex (COF), smart card, UV curable anisotropic conductive adhesive (ACA) Introduction Conductive Adhesive Bonding using conventional ACA has been used maturely in electronics packaging industry. Analyses of the reliability and failure mechanism have been carried out, by which the product can give better performance. Most of the conventional ACA are only made use of thermal curable epoxy resin. Nowadays, UV curable adhesive are now used successfully in the manufacture of ACA. Due to the rapid cure and little to no emissions of volatile organic compounds without affecting other components in the assembly [1], investigation of the bonding and curing conditions had been done in out previous paper. Therefore, the optimized curing conditions obtained in our previous paper which gave a relatively better performance of the COF bonding were used in this work. In order to determine the curing degree of the UV curable adhesive, Fourier Transform Infrared Spectroscopy (FTIR) [2, 3] measurements was used, this measurement technique is contributed to the epoxy and reference peaks can numerically represent degree of curing of the adhesives. Moreover, it is worth to carried out some relability tests and microscopic observation of the contactless smart card application would be a more superior way to improve the fabrication procedures and curing conditions. Experimental Procedure Materials In this work, the UV curable anisotropic conductive adhesive (ACA) used is made of epoxy resin in which plastic hollow bead with nickel-gold coated particles are dispersed. The density of the conducting particles in the epoxy is ~1200 per mm 2. Each conducting particle is about 6.5µm in diameter. The dimensions of the test chip are 1.11 x 1.06 mm 2 with 4 Au bumps on the active side of the chip. Only the two bumps which placed diagonally at the corners are required to be bonded on the pads of the flexible substrate. The dimensions of the bumps are 104 x 104 x 18 µm 3. The flexible substrate used in this work is 75µm thick PET and the area of the substrate is about 70 x 76 mm 2. A custom designed copper coil is screen-printed on the PET as the antenna of the contactless smart card. The thickness of the copper coil is 17µm. EFOS Novacure Model N2001-A1 was used as the light source with operating range of about nm. Fabrication of Contactless Smart Cards In order to activate the curing reaction of the adheisve, a spot of UV curable ACA weighted about 0.1mg was dispersed on the connection area of the flexible substrate and then exposed to UV light intensity of 400mW/cm 2 for 1.5s. By using a KarlSuss FCM505 Manual Flip Chip Bonder, after the alignment between the bumps and the pads was finished, the test chip was then bonded on the flexible substrate for 10s under the pressure of about 1N at room temperature. The bonded samples were then post-cured in an oven at 140 o C for 10 mins. A total number of 15 contactless smart card samples were successfully fabricated. The read range of all the samples were measured by the card reader before subjecting to autoclave test at 121 o C, 100%RH for 0, 48, 96 and 144 hours to investigate the failure mechanisms. Besides the contactless smart card samples, 3 sets of bare ACA specimens that were cured under the same curing conditions were also put into autoclave test chamber simultaneously. Shear Test At each test interval of the autoclave test, the read range of each set of samples were measured so as to determine the electrical perfornance of the ACA joint. Then some of the samples were sheared by using a Dage Series 4000 Bondtester /04/$ IEEE International Conference on the Business

2 In this test, the shear blade was equipped with a shear hook and then fixed to the crosshead. Each sample was placed on the fixture and then clamped on the bondtester. The shear blade was placed just 20µm right above the copper coil, which applied a tensile load with 100.0µm/s loading rate moving horizontally towards the test chip. Result & Discussion FTIR-ATR Results One of the key functions of the epoxy matrix in ACA is used to adhere the chip and the substrate. The curing degree of epoxy can affect the mechanical property as well as the electrical property of the ACA [4-6]. In current study, FTIR- ATR spectroscopy is used to check the curing degree of epoxy in ACA with the following equation (1). α = 1 [(A epoxy,t / A ref,t ) / (A epoxy,0 / A ref,0 )] (1) Where A epoxy,0 and A ref,0 are the initial area of epoxy and reference peak respectively, and their corresponding values at time t are A epoxy,t and A ref,t. For the UV curable anisotropic conductive adhesive used in this work, the most suitable epoxy and reference peaks were suggested to be 789 and 1455 cm -1 [2]. The reference and specimen spectra represented the average of 16 scans recorded at 4cm -1 resolution. From Figure 1, it was found that the curing degree still increased after 48 hours autoclave test. Also, it was still relatively slightly increment of the curing degree for the bare UV curable ACA specimens from 48 to 144 hours. adhesive used for the ACA joint. Thus the curing of the adhesive and the number of conductive particles may affect the read range of the contactless smart card sample. Each sample was placed horizontally right above the card reader, the longest distance that the sample could be recognized by the card reader was measured as the read range of the sample as shown in Figure 2. Before subjecting to the autoclave test, the read range of all the samples was ~8.2mm with standard deviation of ~0.3mm. Read Range (mm) Sample No. Read Range Figure 2. Read range of all the good contactless smart card samples before autoclave test Read Range (mm) 9 8 Curing Degree (%) Read Range 5 80 Curing Degree Figure 3. Read range of contactless smart card samples measured at each test interval of autoclave test Figure 1. Curing degree of bare UV curable ACA measured at each test interval of autoclave test Read Range Measurement As mentioned in our previous paper, the ACA joint of the contactless smart card application can be represented as a circuit diagram of the two resistances R J and one capacitance C J connected in serious. R J is contributed to the number of conductive particles in between the bump and the pad while the C J is contributed to the dielectric properties of the After autoclave test, the flexible substrates became more brittle and they were warped more seriously for longer test intervals. The read range of the samples after autoclave test were also measured as shown in Figure 3. The result was found that the read range was still increased after 48 hours autoclave test which was similar to the phenomenon of the result of curing degree. Moreover, it was relatively slightly increase in the read range from 48 to 96 hours. However, after 144hours, the flexible substrate was cracked and warped seriouly that the samples cannot be recognized by the card reader. Nevertheless, by comparing the contactless smart card using conventional wire bonded technology, the samples /04/$ IEEE International Conference on the Business

3 after 48 and 96 hours autoclave test in this work can reach nearly 92% and 98% performance of the general wire bonded contactless smart card. Shear Results After read range meansurement, the 3 sets of contactless smart card samples were then undergone shear test. The result is shown in Figure 4. It was found that the shear strength of the samples increased from 0 to 48 hours and then drops down until 144 hours. Although the curing degree and read range kept on increasing, the shear strength started to drop at 48 hours. Therefore, in order to investigate the occurrence of the shear result, SEM observation was carried out. Obviously, more adhesive was left on the bumps in Figure 6 rather than in Figure 7. Therefore, the electrical connection of samples after 144 hours might not be so good that those samples cannot be recognized by the card reader Shear Load (g) Figure 5. Active side of sheared test chip with after 48 hrs autoclave test 200 Shear Result 0 Figure 4. Shear strength of the contactless smart card samples after autoclave test SEM Observation From the read range measurement, all the samples were failed after 144 hours autoclave test. One of them was put into cross-section observation by using Scanning Electron Microscope (SEM) with Energy Dispersive X-ray Spectrometer (EDX) (Philips Model XL40). On the other hand, after shear test, the sheared test chips were also observed by using the SEM to find out the fracture mode and the distribution of the adhesive and the conductive particles on the active sides of the chips. Figures 5, 6 and 7 show the active sides of the sheared test chip of the samples were undergone autoclave test for 48, 96 and 144 hours respectively. When comparing the figures, the overall remnants on the chip surface is in declining trend. There is almost no ACA on the chip surface; it s a adhesive failure. This implies that the autoclave test has caused corrosion attack at the ACA-chip layer. At zero hour, the failure mode of shear test is more to cohesive failure which is highly dependent on the curing degree of ACA. High temperature environment in autoclave test provide hest energy to the ACA and causing further curing that has increase the curing degree and thus the shear strength of the ACA joints. However, the corrosion attack at the ACAchip interface has degraded the joint strength and dominant in the shear test failure mode. Figure 6. Active side of sheared test chip with after 96 hrs autoclave test Figure 7. Active side of sheared test chip with after 144 hrs autoclave test /04/$ IEEE International Conference on the Business

4 Figure 8. Cross-section of the ACA joint after 96 hrs autoclave test read range measurement. The results showed that when the curing degree of the UV curable ACA can reach more than 95%, the electrical performance of the fabricated samples in this work can behave ~98% of the general performance of wire bonded contactless smart cards. However, the samples cannot be recognized by the card reader when for further heat post-curing treatment. According to the shear test result, the shear strength of the samples after 96 hours autoclave test started to degrade due to the highly cured of the adhesive that led to the weakly adhesion at the ACA-chip interface. The corrosion attack at the ACA-chip interface has degraded the joint strength and dominant in the shear test failure mode. By observing the cross-section of ACA joints at 96 and 144 hours, the cause of the ACA joint deformation might due to the warp of the flexible substrate. Consequently, the test chip was lifted up and then the spacing between the bump and the pad was increased that caused open joint. This led to the failure of the read range mesurement after 144 hours autoclave test. In conclusion, the optimized curing conditions obtained from the previous paper could not provide a very good electrical performance due to the read range and curing degree measurement. For that reason, the UV curable ACA should be cured up to around 95% to obtain a longer read range. However, the adhesive cannot be fully cured as it would probably cause the degradation of the shear strength performance and even cause open joint of the contactless smart card and thus the reliability of the contactless smart card. Figure 9. Cross-section of the ACA joint after 144 hrs autoclave test In order to have a in-depth analysis of the failure of the electrical performance and degradation of shear strength of the ACA joint after 144 hours autoclave test, cross-sections of the samples after 96 and 144 hours were inspected and compared as shown in Figures 8 and 9. There was still electrical contact of the ACA joint in Figure 8 while that of in Figure 9 did not. Therefore, not only the fracture of the flexible substrate influenced the failure of the contactless smart card recognization, but also the lift up of the test chip caused the swelling of the ACA joint. The deformation of the ACA joint might be contributed to the warp of the flexible substrate after autoclave test. On the other hand, the separation between the adhesive and the test chips in the figures were become larger. The spalling and the delamination between the chip and the ACA would cause the declination of shear test data after 96 & 144 hours autoclave test. Conclusion From the FTIR-ATR measurement, it was determined that the curing degree still had an increment after the autoclave test. The increment phenomenon also slimilar to that of the Acknowledgments The authors would like to acknowledge the financial support from the City University of Hong Kong (ITF Project Conductive Adhesive Technology Programme for Fine Pitch Electronic Packaging, Project No.: ITS/182/00). The authors would like to thank Ms S C Tan for her consultation on chemical background of the UV curable ACA and also would like to thank MaCaPs International Limited for the technical advice on the read range measurement of contactless smart card. References 1. William S. Pataki, Optimization of Free-radical Initiation Reactions in the Electrical Industry, Electrical Insulation Conference and Electrical Manufacturing & Coil Winding Conference Proceedings, 22 25, (1997), pp J. D. Cho, H. K. Kim, Y. S. Kim and J. W. Hong, Dual curing of cationic UV-curable clear and pigmented coating systems photosensitized by thioxanthone and anthracene, Polymer Testing, Vol. 22, (2003), pp H. Nagata, M. Shiroishi, Y. Miyama, N. Mitsugi and N. Miyamoto, Evaluation of New UV-Curable Adhesive Material for Stable Bonding between Optical Fibers and Waveguide Devices: Problems in Device Packaging, Optical Fiber Technology 1, (1995), pp /04/$ IEEE International Conference on the Business

5 4. Liu, J., An Overview of Advances of Conductive Adhesive Joining Technology in Electronics Applications Materials Technology, Vol. 10, pp Watanabe, I. and Takemura, K., pp [A book reference ] 6. Lai, Z. and Liu, J., Anisotropically Conductive Adhesive Flip-Chip Bonding on Rigid and Flexible Printed Circuit Substrates, IEEE Transactions on Components, Packaging and Manufacturing Technology Part B, Vol. 19, No. 3, pp /04/$ IEEE International Conference on the Business