SURFACE PREPARATION FOR IMPROVED ADHESION

Application Note

Published by March Plasma Systems www.marchplasma.com © 2001 March Plasma Systems, Inc

INTRODUCTION photoresist used for photoprocessing substrates, and oil fumes in the atmosphere. Surface preparation through surface Surface contamination is difficult to avoid activation and contamination removal, by during the manufacturing process. Plasma plasma processing is widely used in processing removes the contaminants and industries such as the microelectronic and makes surface clean and active resulting in optoelectronic device assembly, printed improved bonds and decreased circuit board (PCB) manufacturing and occurrence of delamination at the interface as medical device manufacturing industries. surface contamination is a major cause of Surface preparation by plasma removes the poor wire bond pull strength and adhesion [4- contaminants from the surface and activates 6]. Therefore, it has been widely used for the surface for various applications including enhancing improves attachment, removing improving adhesion and promoting fluid flow. oxides, promoting void-free underfill, improving wire-bond strength and eliminating Plasmas are highly reactive mixtures of gas delamination in microelectronic and species consisting of large concentrations of optoelectronic industries [2. 4-9]. ions, electrons, free radicals, and other neutral species. Plasma is a proven EXPERIMENTAL METHODS technology, which provides an efficient, economic, environmentally friendly, and General experimental methods for surface versatile technique for modifying the surface property evaluation are contact angle properties of materials. Plasma treatment measurement, SEM, AFM and XPS. can be used for surface activation and Contact angle measurement is a simple and contamination removal without creating any inexpensive method for evaluating the hazardous by-products and without effectiveness of the surface contamination changing any bulk properties [1, 2]. removal and surface activation processes.

Surface activation is a process where Plasma treatment and surface modification surface functional groups are replaced with of plastics, metals, and ceramic surfaces different atoms or chemical groups from the increases the wettability of those surfaces plasma. Surface activation of materials is as measured by the contact angle. In achieved utilizing plasma source gases such general, the lower the contact angle the as argon, oxygen, hydrogen, or a mixture of higher the surface energy. The increase of these gases [2]. energy and decrease of contact angle, usually correlates directly with improved Surface contamination removal by plasma is adhesion since organic contaminants have an ablation process, where physical been removed during the plasma treatment sputtering and chemical etching are the key and the free radicals and polar function processes involved. The plasma process groups form on the surface allowing for a removes organic contaminants such as better interface between the surface and the residual organic solvents, epoxy residue, typically polar fluid. oxidation, and mold release compounds, on the surface of most industrial materials. The correlation of the level of interfacial These surface contaminants undergo organic contamination as determined by repetitive chain scission under the influence XPS, relative to the contact angle measured of ions, free radicals and electrons of the on the leadframe for Ar and O2 plasma until their molecular weight is plasma treatment is shown in Figure 1. The sufficiently low to volatilize in the vacuum [2, data indicates that as the contact angle 3]. decreases the level of organic contamination decreases proportionally. The result clearly In microelectronic assembly applications the shows that the contact angle measurements surface contaminants include metal oxides are indeed a good indication of the level of and organic substances introduced by organic contamination on copper substrates. overexposure to atmospheric air, bond grease during manual handling, soldering process, Figure 1. 90 components, such as die, , fiber, and 80 O2/400W thermoelectric coolers. A clean die and O2/50W 70 substrate surface is desirable because it 60 Ar/400W 50 Ar/50W promotes better adhesion of the die attach 40 compound to both the die and the substrate. 30 Plasma cleaning prior to component 20 attachment provides better contact, better 10 heat transfer, and minimal voiding. Surface Contact Angle (degrees) 0 0 10203040506070 Plasma Treatment Time (second) Wire Bonding

Figure 1. The presence of oxides and organic contaminants on bond pads inhibits APPLICATIONS successful wire bonding. Assurance of an oxide-and-contaminant-free surface is Component Attach important to obtain good bond yields.

Surface activation and contamination The data shown in Table 1 indicates the removal by plasma process can improve the effect of argon plasma cleaning on wire adhesion between the substrate and bond yield.

# of # of Wire Size Pull # of Bond Failure Devices (mils) Test Failures Rate Lab#1 Plasma 25 1380 1.5 5 g 6 0.43% Cleaned Plasma 100 1.0 3 g 11 11% Cleaned Control 25 1378 1.5 5 g 10 0.73% Control 94 1.0 3 g 23 24.5% Lab#2 Plasma 50 1375 3.5 g 8 0.58% Cleaned Control 50 1375 3.5 g 26 1.89% Lab#3 Plasma 10 840 1 0.12% Cleaned Control 10 840 29 3.45%

Table 1 The samples were plasma cleaned with the ability to form good adhesion with argon for 10 minutes using the following package components and to remain bonded plasma condition: 100 watts and 0.2 Torr, is of paramount importance as delamination and were then subjected to pull tests. The along the interfaces is a major reliability plasma cleaned samples showed average issue for plastic encapsulated microcircuits pull strength of 6.65 grams with a standard (PEMs). Plasma treatment improves this deviation of 1.57. The control showed adhesion and bond strength. average pull strength of 5.3 grams with a standard deviation of 1.89. The data The data shown in Figure 3 demonstrates indicates that the bonding strength has been about a factor of two increase in the bond improved after plasma cleaning. strength. The material used in this case was a PPS plastic molded into a multi-pin connector. Cadmium and nickel wires were bonded into position by epoxy cement The unique challenge in flip chip packaging (Abelbond #789-3), cured, and the bonds is the underfill process, particularly designs tested. Plasma treatment was run in the that use large dies, tight gaps, and high- March PX-500 system. density ball placement. Plasma has proven to increase surface energy promoting Gas: Argon adhesion, minimizing voiding and increasing Power: 200 watts wicking speeds. The contact angle under the Pressure: 180 mTorr die and on the covered substrate surface Time: 15 minutes decreases with the increasing plasma treatment time as shown Figure 2. Also Bond Strength displayed in Figure 2 is the effect of die size; the larger the die the more difficult it is for the plasma to penetrate between the die and 40 the substrate. 30

20 120 ITRAK, Gap: 0.58 mm, 120 W, 170 mtorr 360 sec 10 100 180 sec 80 120 sec 0 0 sec 60 No plasma Plas ma treatment treatment 40 20 Figure 3 Contact Angle (degrees) 0 The adhesion between leadframe and encapsulant in plastic encapsulated on die 4.5*5.0on die 6.5*6.5on die 8.5*7.5 on PBGA on4.5*5.0 PBGA on6.5*6.5 PBGA 8.5*7.5 microcircuits is characterized by leadframe Figure 2. Surface contact angle under the pull-out test (See Figure 4) [4]. The beneath of die after plasma treatment with maximum de-bond load decreases with the different plasma exposure time. increase of plasma exposure time. The debond load is a measure of the strength of Encapsulation and Mold the bond between the encapsulant and the leadframe. The larger the debond load the The purpose of the plastic encapsulant for better the adhesion. Figure 4b displays the applications is to provide relationship between contact angle and adequate mechanical strength, adhesion to debond load. In general, with decreasing various package components, good contact angle the debond load increases. corrosion and chemical resistance, matched Thus the contact angle measurement coefficient of thermal expansion to the method is a good indicator of bond strength materials it interfaces with, high thermal in encapsulation processes. conductivity and high moisture resistance in the temperature range used. In particular, Marking 90 80 Plasma surface preparation is also used in 70 Plasma treated 60 marking. The activated surface can improve 50 the adhesion of aqueous ink. The plasma 40 prepared surface improves adhesion of 30 aqueous based inks. 20 As received 10 Max. Debond Load (N) 0 Hermetic Sealing 50 60 70 80 90 Contact angle (degrees) Plasma technology can be used to prepare Figure 4. Maximum de-bond load as a the surface prior to hermetic sealing of laser function of after plasma exposure time (a) device. Plasma cleaned surface and surface contact angle (b). The plasma improves the adhesion at the interface condition: H2 (50%) and Ar (50%), 5 min. allowing for a more reliable weld. 234-300 mTorr, and 400 W. [4]

CONSIDERATIONS 100 Contact Angle (degrees) Plasma Conditions 80

60 H2+Ar As received Plasma conditions are very important for the Maximum Debond Load (N) H2+Ar 40 plasma surface activation and contamination As received removal. The important factors of plasma 20 process include gases, input power,

0 operating pressure, plasma exposure time, -5 0 5 10 15 20 25 30 location of the sample in the chamber, and Exposure time (hours) electrode configuration. All of the Figure 4b parameters should be determined carefully for the different applications. Principally, a Figure 5 shows the surface contact angle on lower operating pressure needs to be copper leadframe with the plasma treatment applied in argon plasma process as it is a time. The surface contact angle decreases physical plasma process. However, a with increasing plasma treatment time. The higher operating pressure is necessary in surface contact angle also depends on the oxygen or other reactive gas plasma as plasma operating conditions, such as gas chemical reaction is dominant on the selection, power input, pressure, and time. surface. The figure displays that reducing the power impacts the plasma treatment effectiveness. Life Time

The question often asked is "how long does 90 80 O2/400W a surface remain active?" since the 70 O2/50W activated surface is sensitive to the 60 Ar/400W environment. Generally, the activated Ar/50W 50 surface will gradually lose its wettability 40 30 because of air contamination, self- 20 contamination and storage contamination. 10 One example, using the same PPS plastic Surface Contact Angle (degrees) 0 0 10203040506070 and plasma treatment conditions is shown in Plasma Treatment Time (second) Figure 6. The data illustrates the change in Figure 5 contact angle as a function of time. Due to the surface recontamination illustrated in Figure 6, the adhesion strength will decline with increasing exposure time after plasma treatment. 90 80 REFERENCE Contact angle of PPS W/O plasma treatment 70 60 50 1. Handbook of Plasma Processing 40 Technology, Fundamentals, Etching, 30 Deposition, and Surface Interactions, 20 Edited by S. M. Rossnagel, J. J. Cuomo 10 and W. D. Westwood, Noyes

Contact Angle (degrees) 0 0 20 40 60 80 100 120 140 160 Publication, Westwood, NJ, USA. Times (hours) 2. E. Finson, S. L. Kaplan, and L. Wood, Figure 6 Plasma Treatment of Webs and Films, Society of Vacuum Coaters, 38th Annual Storage Technical Conference Proceedings (1995) Another issue that arises is how to store the 3. H. Yasuda, Plasma Polymerization, treated samples. In this study, the same Academic Press, Orlaando, FL, USA, PPS plastic samples were plasma treated 1985. and placed in a Teflon FEP bag, a 4. Y. Sung, J. -K. Kim, C Y. Yue, and J. -H polyethylene bag, or wrapped in a plasma Hsieh. Bonding strengths at plastic treated aluminum foil (see Figure 7 below). encapsulant--plated copper The surface activation of all samples leadframe interface, Microelectronics degrades with time, except for those stored Reliability 40 (2000) 1207-1214. in FEP. 5. L. Wood, C. Fairfield, and K. Wang, Plasma Cleaning of Chip Scale

90 Polyethylene Packages for Improvement of Wire 80 Aluminum foil Bond Strength, TAP technology, Second FEP 70 edition, 75-78. 60 50 6. F. Djennas, E. Prack, Y. Matsuda, 40 Investigation of Plasma Effects on 30 Plastic Packages delamination and 20 Cracking, IEEE Trans CHMT 16 (1993) 10 0 919-924 No plasma Immediately 24 hour after 48 hour after treatment after plasma plasma plasma 7. L. J. Matienzo and F. D. Egutto, treatment treatment treatment Adhesion Issues in Electronic Figure 7 Packaging, Solid State Technology, July (1995), 99-106. APPLICATIONS LABORATORY 8. R. N. Booth and P. E. Ongley, Plasma Treatment in Hybrid and Conventional The technical staff at March is pleased to Electronic Assemblies, Hybrid Circuits, 7 offer its experience in plasma technology for (1995). your applications. We would also be happy 9. H. K. Kim, Plasma Cleaning of to publish any data you would like to share Spacecraft Hybrid Microcircuits and Its with others in the field. Direct your calls and Effect on Electronic Components, 1989 faxes to March Plasma Systems, Attention: The International Society for Hybrid Applications Laboratory. Microelectronics, 144-148.