This article was downloaded by: [Simon Fraser University] On: 29 December 2012, At: 11:15 Publisher: Taylor

& Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Journal of Adhesion Science and Technology

Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/tast20

Adhesion improvement of epoxy resin/copper lead frame j oints by azole compounds

S.M. Song , C.E. Park , H.K. Yun , C.S. Hwang , S.Y. Oh & J.M. Park
a e f a b c d

Polymer Research Institute, Department of Chemical Engineering, Pohang University of Science and Technology, San 31 Hyoja Dong, Pohang 790-784, Korea
b

Polymer Research Institute, Department of Chemical Engineering, Pohang University of Science and Technology, San 31 Hyoja Dong, Pohang 790-784, Korea
c

Packaging Development Team, Semiconductor R&D Center, Samsung Electronics, P.O. Box 37, Suwon 449-900, Korea
d

Packaging Development Team, Semiconductor R&D Center, Samsung Electronics, P.O. Box 37, Suwon 449-900, Korea
e

Packaging Development Team, Semiconductor R&D Center, Samsung Electronics, P.O. Box 37, Suwon 449-900, Korea
f

Packaging Development Team, Semiconductor R&D Center, Samsung Electronics, P.O. Box 37, Suwon 449-900, Korea Version of record first published: 02 Apr 2012.

To cite this article: S.M. Song , C.E. Park , H.K. Yun , C.S. Hwang , S.Y. Oh & J.M. Park (1998): Adhesion improvement of epoxy resin/copper lead frame j oints by azole compounds, Journal of Adhesion Science and Technology, 12:5, 541-561 To link to this article: http://dx.doi.org/10.1163/156856198X00218

PLEASE SCROLL DOWN FOR ARTICLE

Full terms and conditions of use: http://www.tandfonline.com/page/terms-andconditions This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sublicensing, systematic supply, or distribution in any form to anyone is expressly forbidden. The publisher does not give any warranty express or implied or make any representation that the contents will be complete or accurate or up to date. The accuracy of any instructions, formulae, and drug doses should be independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims, proceedings, demand, or costs or damages whatsoever or howsoever caused arising directly or indirectly in connection with or arising out of the use of this material. Downloaded by [Simon Fraser University] at 11:15 29 December 2012

Adhesion j oints by

improvement azole compounds

of epoxy

resin/copper

lead

frame

Downloaded by [Simon Fraser University] at 11:15 29 December 2012

S. M. SONG 1, C. E. PARK 1 ,*, H. K. YUN 2, C. S. HWANG 2, S. Y. OH 2 and J. M. PARK 2 1 Polymer Research Institute, Department of Chemical Engineering, Pohang University of Science and Technology,San 31 Hyoja Dong, Pohang 790-784, Korea 2 Packaging Development Team, Semiconductor R&D Center, Samsung Electronics, P.O. Box 37, Suwon 449-900, Korea Received in final form 20 December 1997 AbstractThe adhesion strength of epoxy resin/copper joints is often very poor, due to the naturally formed copper oxide having a low mechanical strength. To improve the adhesion strength of epoxy resin/copper lead frame joints, copper lead frames were treated with azole compounds as adhesion promoters. The azole compounds used were benzotriazole (BTA), benzotriazole-5-carboxylic acid (CBTA), 8-azaadenine, imidazole, 2-methyl imidazole, urocanic acid, adenine, benzimidazole, and polybenzimidazole (PBI). The dependence of the adhesion strength of epoxy resin/azole-treated copper joints on the structure of the azole compound, the azole treatment time, and the azole treatment temperature was investigated. The surface coverage of azole-treated copper was examined by contact angle measurements, a surface defect test, optical microscopy, and scanning electron microscopy (SEM), and the locus of failure was studied by X-ray photoelectron spectroscopy (XPS). Triazole compounds such as CBTA and 8-azaadenine showed excellent adhesion strength; imidazolebased azole compounds did not improve the adhesion strength. However, the adhesion strength of CBTA- and 8-azaadenine-treated joints decreased with increasing treatment time, since thick porous Cu-azole complexes had a weaker mechanical strength when formed. The polymeric azole compound PBI showed the highest adhesion strength, 785 N/m, because of complete coverage of the copper surface. The thermal stability of azole compounds and epoxy resin/azole-treated copper joints was also investigated. CBTA and 8-azaadenine did not decompose up to 250C, while PBI was stable up to 500C in an air atmosphere. Keywords: Adhesion strength; epoxy resin; copper; triazole; polybenzimidazole; Cu-azole complex.

*To whom correspondence should be addressed. E-mail: cep@postech.ac.kr

542 1. INTRODUCTION In packaging encapsulation, proper adhesion between the epoxy molding comBut the ad(EMC) and the lead frame is important for device reliability. pound hesion strength between the EMC and copper in a lead frame is not strong because the mechanical strength of a natural copper oxide layer is very weak and the copper oxide degrades the epoxy resin at high temperature [1-5]. Especially in the chip packaging process, since a copper lead frame is frequently exposed to high temperature through the process of die attachment and gold wire bonding, weak copper oxide is formed on the copper lead frame. Moreover, it has been reported that the mechanical strength of copper oxide formed at high temperature is weaker than that formed at room temperature [6]. Therefore, adhesion promoters are required to improve the adhesion strength between the EMC and the copper lead frame. Azole compounds such as benzotriazole can be used as adhesion promoters for improving the adhesion strength of epoxy resin/copper joints since the azole moiety can react with copper to form a Cu-azole complex [7-10], which acts as a corrosion inhibitor for copper [8, 9]. Park and Bell improved the adhesion strength between epoxy resin and copper by treating copper with triazole compounds [7]. However, azole compounds such as benzotriazole do not have a functional group to react with epoxy resin. In this study, azole compounds having carboxylic acid or amine were employed to react with the epoxy resin. It has also been reported cannot prevent oxidation of the copper surface at high that azole compounds [ 11 ] . Therefore, the polymeric azole compound polybenzimidazole temperature was used because it has good thermal stability and better surface coverage of the copper surface.

Downloaded by [Simon Fraser University] at 11:15 29 December 2012

2. EXPERIMENTAL 2.1. Treatment of copper lead frame coupons with azole compounds

The triazole compounds used were benzotriazole (BTA), benzotriazole-5-carboxylic the imidazole-based azole compounds used were acid (CBTA), and 8-azaadenine; In adimidazole, 2-methyl imidazole, urocanic acid, adenine, and benzimidazole. These dition, polybenzimidazole (PBI) was used as a polymeric azole compound. azole compounds were purchased from Aldrich Chemical Co. Figures 1 and 2 show azole the chemical structures of the triazole compounds and the imidazole-based compounds used. the copper coupons (C 19400, Before treatment with the azole compounds, and then rinsed with Co., Korea) were degreased with dichloromethane Poongsan acetone. The copper coupons were immersed in 5 wt% sulfuric acid solution for 5 min to remove the weak, naturally forming copper oxide and rinsed with distilled water; they were then immersed in the azole solution. The azole solutions, except of 3 x 10-3 M with distilled water; PBI solution, were made to a concentration

543

Downloaded by [Simon Fraser University] at 11:15 29 December 2012

Figure 1. Chemical structures of the triazole compounds.

Figure 2. Chemical structures of the imidazole-based azole compounds and PBI. 0.5 wt% PBI solution was made with dimethyl formamide (DMF; Aldrich Chemical Co.). The immersion time and temperature of the above solutions were varied. After treatment of the copper coupons, they were removed, gently blown with air, and then dried at 150C for 30 min.

544 For thermal stability measurement of the azole compounds, copper coupons were heated at 200C in a convection oven for treated with azole compounds different times before molding of epoxy resin. 2.2. Preparation of peel test samples of epoxy/copper joints

Downloaded by [Simon Fraser University] at 11:15 29 December 2012

210O-Cresol novolac epoxy resin (LER N-673, epoxy equivalent weight = 230 g/eq; Lucky Epoxy Co., Korea), 100 parts, and nadic methyl anhydride (NMA; Aldrich Chemical Co.), 80 parts, as a curing agent were mixed in a I 00 C oil bath, and benzyl dimethylamine (BDMA, Aldrich Chemical Co.), 0.3 part, was added as a catalyst to the epoxy resin mixture. The epoxy resin mixture was poured onto the copper coupons treated with azole compounds and then cured at 90C for 2 h and 150C for 4 h to make the epoxy resin/copper joints. 2.3. Peel strength measurements for epoxy resinlcopper joints

After curing the epoxy resin/copper joints, the copper coupons were cut to dimensions of 5 x 70 mm for a 90 peel test. The adhesion strength of the epoxy resin/copper joints was measured by a 90 peel test at a peel rate of 5 mm/min using an Instron (Model 4206). The number of specimens tested was more than 15 for each experimental condition. 2.4. Wettability measurements The wettability of copper coupons treated with azole compounds was monitored by measuring the equilibrium contact angles of water on the azole-treated copper coupons, using a contact angle goniometer (CA-A, Kyowa Surface Science Co., Tokyo, Japan). The contact angles were measured 20 times for each specimen. 2.5. Sujhce defect test and morphology of the azole-treated copper coupons

The azole-treated copper coupons were immersed in 0.001 us aqueous silver nitrate solution for 10 min, rinsed with water, and then dried at 110 C for 10 min [5]. The surface morphology of the azole-treated copper coupons was also investigated using a scanning electron microscope (Hitachi S-570). The copper area that does not react with the azole compound in the azole-treated copper surface reacts with silver nitrate, forming a silver spot, which was observed using an optical microscope (Carl Zeiss, Germany) in the reflection mode. 2.6. X-ray photoelectron spectroscopy (XPS) analysis

of the azole-treated The atomic concentrations copper coupons and the peeled surfaces of epoxy resin/azole-treated copper joints were determined with a PerkinElmer Phi-5400 X-ray photoelectron using a Mg Ka X-ray source spectrometer (1253.6 eV, 350 W). The take-off angle between the analyzer and the sample surface was varied from 30 to 90 in order to examine the variation in the atomic composition as a function of the depth from the surface [12].

545 2.7. Thermogravimetric analysis (TGA)

TGA of the azole compounds was performed from 50C to 700C using PerkinElmer TGA-7 equipment at a scanning rate of 10C/min in an air environment.

3. RESULTS AND DISCUSSION 3.1. Adhesion strength of epoxy resinltriazole-treated copper joints

Downloaded by [Simon Fraser University] at 11:15 29 December 2012

Figure 3 shows the adhesion strength between epoxy resin and the CBTA-treated copper coupon as a function of the temperature of the CBTA solution. The adhesion strength increased as the treatment temperature increased. With the increase of the CBTA solution temperature, the extent of the reaction between CBTA and copper increased and the coverage of copper with CBTA was more complete. Based on these results, the treatment temperature of the azole compound was fixed at 80C. The adhesion strength between epoxy resin and the azole-treated copper coupons is shown in Fig. 4 as a function of the azole treatment time. The adhesion strength between epoxy resin and an untreated copper coupon was 54 N/m. Therefore, triazole compounds, CBTA and 8-azaadenine, were very effective in especially improving the adhesion strength between the epoxy resin and a copper coupon, up to 301 and 353 N/m, respectively. The adhesion strength of epoxy resin to BTAtreated copper did not vary much with the treatment time. But the adhesion strengths of CBTA- and 8-azaadenine-treated copper coupons decreased with increasing azole treatment time. These results were contrary to our expectation, i.e. when the reaction time between copper and an azole compound increases, the copper surface is completely covered with a Cu-azole complex and a higher adhesion strength would be expected.

Figure 3. Peel strength of epoxy resin/CBTA-treated copper joints as a function of the treatment temperature of the azole solution.

546

Downloaded by [Simon Fraser University] at 11:15 29 December 2012

Figure 4. Peel strength of epoxy resin/copper joints as a function of the azole treatment time. (0) BTA; (8) CBTA; (A) 8-azaadenine. To investigate the surface coverage of azole-treated copper surfaces, a surface defect test was performed. As shown in Fig. 5, the results of the surface defect test confirmed that the coverage of copper treated with an azole compound was more complete with the increase of the azole treatment time. A surface defect was not seen in the coupons treated with BTA, CBTA, and 8-azaadenine for 10 min and this result was in line with our expectation. the contact angles of water on azole-treated copper surfaces were Furthermore, measured to investigate the reaction of copper and azole, because the coverage of the copper surface by an azole treatment decreases the polarity of the copper surface. The contact angle data for different azole treatment times are shown in Fig. 6. The contact angle on the BTA-treated copper did not vary with the BTA treatment time, but those on CBTA- and 8-azaadenine-treated copper surfaces decreased sharply the contact angles measured after with the azole treatment time. Furthermore, a treatment time of 10 min were reduced to less than 10 for both CBTA- and This phenomenon was not expected since 8-azaadenine-treated copper coupons. the contact angle of water to the surface covered with the organic material cannot be less than 10. Therefore, it can be surmised that a porous Cu-azole complex was formed on the copper surface with the increase of the treatment time and the contacting liquid, water, was absorbed into the pores to reach less than 10 of contact angle. To confirm this hypothesis, the surface morphology of azole-treated copper was observed using SEM. As shown in Fig. 7, no change in morphology of BTA-treated copper was observed for copper treated for either 15 s or 10 min. The copper treated with CBTA and 8-azaadenine, however, showed a large difference between the surface morphologies at treatment times of 15 s and 10 min. As the azole treatment time was increased, particles were observed on the surface of copper and the thickness of the Cu-azole complex layer increased.

Downloaded by [Simon Fraser University] at 11:15 29 December 2012

547

Downloaded by [Simon Fraser University] at 11:15 29 December 2012

548

549

Downloaded by [Simon Fraser University] at 11:15 29 December 2012

Figure 6. Contact angle of water on azole-treated copper for different azole treatment times. () BTA; (0) CBTA; (A) 8-azaadenine. Table 1. XPS analysis of copper treated for I S s and10 min with 3 x 1 O-4M CBTA solution at 80C

To clarify the formation of a thick Cu-azole complex layer, Table I shows that performed with varying take-off angle. of nitrogen is nearly constant for take-off angles larger than concentration of Cu increases with increasing take-off angle for

XPS analysis was the concentration 45 and that the the CBTA-treated

Downloaded by [Simon Fraser University] at 11:15 29 December 2012

550

Downloaded by [Simon Fraser University] at 11:15 29 December 2012

551

552

Downloaded by [Simon Fraser University] at 11:15 29 December 2012

Figure 8. Optical micrograph of CBTA-treated copper after treatment for 5 h. copper for the treatment time of 15 s. This means that the Cu-azole complex layer is very thin and the copper concentration increases with increasing depth from the copper surface. For the treatment time of 10 min, the concentrations of both nitrogen and copper increase with increasing take-off angle. This means that the Cu-azole of complex layer is thicker than that after 15 s treatment because the concentrations and copper are higher in the Cu-azole nitrogen complex layer. From the combined results of the peel test, the contact angle measurements, and Cu-azaadenine SEM, and XPS, it appears that the thick Cu-CBTA complex It layers on the copper coupon are porous and weak in mechanical strength. was also confirmed that the thick Cu-azole was weak since the complex layer Cu-CBTA complex layer that formed after 5 h was broken even with gentle air The optical blowing during the preparation of CBTA-treated copper coupons. micrograph shows the delaminated and broken Cu-CBTA layer as white and dark areas, respectively, in Fig. 8. To investigate the locus of failure in the epoxy resin/copper joint, the atomic of peeled surfaces of epoxy resin and copper were determined by compositions XPS after treatment times of 15 s and 10 min (Table 2). The concentrations of C, 0, N, and Cu were monitored with varying take-off angle. The nitrogen concentrations on the peeled surfaces of both epoxy resin and copper treated with CBTA solution for 10 min are much higher than those on the peeled surfaces of epoxy resin and copper treated with CBTA solution for 15 s, and the peeled surface of copper treated for 15 s shows a much higher copper concentration than that of copper treated

553 Table 2. Atomic concentrations of the peeled surfaces of epoxy resin/copper joints with copper treated with 3 x 10-4 M CBTA solution

Downloaded by [Simon Fraser University] at 11:15 29 December 2012

for 10 min. Furthermore, the atomic composition of the peeled surface of copper treated with CBTA for 10 min is similar to that of unjointed copper treated with CBTA for 10 min as shown in Table 1. Therefore, the locus of failure of epoxy treated with CBTA for 10 min is in the Cu-azole complex layer. resin/copper From the experimental results, it is concluded that a thin Cu-azole complex layer can prevent the copper surface from forming weak copper oxide and thus improve the adhesion strength between epoxy resin and copper, but a thick Cu-azole complex layer can act as a weak layer with a concomitant reduction in adhesion strength. 3.2. Adhesion strength of epoxy resinlimidazole-based azole-treated copper joints

azole compounds have been known as corrosion inhibitors for Imidazole-based copperlike BTA [8, 13], and are cheaper and more soluble in water than triazole compounds such as CBTA and 8-azaadenine. The adhesion strengths between epoxy resin and copper treated with imidazolebased azole compounds were much lower than those between epoxy resin and copper treated with triazole, and were nearly the same as that between epoxy resin and untreated copper (Table 3). azole of copper treated with imidazole-based Table 4 shows the wettability of water on copper treated with imidazole-based The contact angle compounds. azoles increased with increasing treatment time. This phenomenon is well known: as a nonpolar organic material covers the surface of a polar material such as metal, the contact angle of water on the metal increases. A silver spot test was performed to investigate the extent of coverage of the copper As shown in Fig. 9, the surface azole compounds. surface with imidazole-based azole compounds was very poor, of copper treated with imidazole-based coverage even with 10 min of treatment time.

Downloaded by [Simon Fraser University] at 11:15 29 December 2012

554

Downloaded by [Simon Fraser University] at 11:15 29 December 2012

555

556 Table 3. Adhesion strengths of epoxy resin/copper joints treated with 3 x 10-3 M imidazolebased azole compounds for different azole treatment times at 80C

Downloaded by [Simon Fraser University] at 11:15 29 December 2012

Table 4. Contact angle of water on copper treated with 3 x compounds for different azole treatment times at 80C

M imidazole-based azole

It appears that the formation of a Cu-azole azole complex with imidazole-based compounds is difficult, compared with triazole compounds. Therefore, treatment by imidazole-based azole compounds cannot prevent oxidation of the copper surface, i.e. the formation of a weak copper oxide layer. 3.3. Polymeric azole adhesion

promoter for

epoxy resin/copper Joints

such as CBTA and 8-azaadenine Although triazole compounds improved the adhesion strength between epoxy resin and copper by suppressing the oxidation of copper, the coverage of the copper surface was not complete with 15 s of treatment. However, further treatment with triazole compounds reduced the adhesion strength sharply, due to the formation of a thick porous Cu-triazole complex layer having

557

Downloaded by [Simon Fraser University] at 11:15 29 December 2012

Figure 10. Optical micrographs of the surface defect test of PBI-treated copper for PBI treatment times of (a) 15 s and (b) 10 min.

Figure 11. Scanning electron micrographs of PBI-treated copper surfaces for PBI treatment times of (a) 15 s and (b) 10 min. weak mechanical strength, even though the coverage of the copper surface with triazole compounds was complete with 10 min of treatment. Therefore, the polymeric azole compound PBI was used to improve the coverage of the copper surface and avoid the formation of a porous Cu-azole complex

558

Downloaded by [Simon Fraser University] at 11:15 29 December 2012

Figure 12. Contact angle of water on the PBI-treated copper surface as a function of the PBI treatment time.

Figure 13. Peel strength of epoxy resin/PBI-treated copper joints as a function of the PBI treatment time.

structure. Figures 10 and 11show optical micrographs of PBI-treated copper after a surface defect test and scanning electron micrographs of PBI-treated copper for treatment times of 15 s and 10 min, respectively. Unlike the low-molecular-weight azole compounds, the copper surface was completely covered, even with 15 s of treatment. The contact angle of water on the PBI-treated copper surface did not change, even with 10 min of treatment (Fig. 12). This indicates that the copper surface can be completely covered with 15 s of PBI treatment and that a porous Cu-PBI complex is not formed. Therefore, as shown in Fig. 13, a higher adhesion strength, 785 N/m, was obtained with PBI treatment and an adhesion strength higher than 500 N/m was maintained even with 10 min of PBI treatment.

559 3.4. Thermal stability of triazoles and PBI

The thermal stability of azole compounds as adhesion promoters is important since lead frames experience high temperature the azole-treated during the packaging (above 200C) and solder reflow process, for example during die attachment (above 225 C). Figure 14 shows TGA thermograms of BTA, CBTA, 8-azaadenine, and PBI. Since BTA melts at 90C and then vaporizes above the melting temperature under an air stream, the weight of BTA decreases sharply above 100 C. The thermal stability of CBTA and 8-azaadenine was maintained up to 250C, and that of PBI up to 500C.

Downloaded by [Simon Fraser University] at 11:15 29 December 2012

Figure 14. TGA thermograms of BTA (-), CBTA (- -), at a heating rate of 10 Cin an air atmosphere.

8-azaadenine (......)

and PBI Hz -)

Figure 15. Peel strength of epoxy resin/azole-treated copper joints where azole-treated copper was exposed to 200C in an air atmosphere before molding of epoxy resin: (0) BTA; (0) CBT.4; (A) 8-azaadenine; (0) PBI.

560 When the azole-treated copper was exposed to 200C in an air atmosphere before the molding of epoxy resin, the adhesion strength between epoxy resin and azole-treated copper decreased with increasing exposure time (Fig. 15). The adhesion strength of the BTA-treated joint decreased to that of an untreated joint after 10 min of exposure. It appears that BTA is not effective as an adhesion the strength of the CBTA- and 8-azaadenine-treated promoter. Although joints decreased, the same adhesion strength as the initial adhesion strength of BTAtreated joints was maintained after 30 min of exposure at 200C. The adhesion strength of the PBI-treated joint was maintained at more than 500 N/m after 30 min of exposure at 200C. Since CBTA, 8-azaadenine, and PBI are thermally stable at 200 C, it can be speculated that thermally oxidized copper is formed and reduces the adhesion strength. From the results of peel strength and thermal stability, CBTA, 8-azaadenine, and PBI can be used as adhesion promoters and to improve the reliability of semiconductor devices by improving the adhesion strength between the epoxy molding compound and the copper lead frame even after the packaging process.

Downloaded by [Simon Fraser University] at 11:15 29 December 2012

4. CONCLUSIONS strength between epoxy resin and a copper lead frame was improved by employing triazole compounds as adhesion promoters. When BTA, CBTA, and 8-azaadenine were used as adhesion promoters, the maximum adhesion strength was 160, 301, and 353 N/m, respectively. The adhesion strength of epoxy resin/ I 8-azaadenine-treated was seven times that of epoxy resin/untreated copper joints copper joints. However, the adhesion strengths of CBTA and 8-azaadenine-treated joints decreased with increasing treatment time because a thick, porous Cu-azole The imidazole-based complex having a weak mechanical strength was formed. azole compounds did not improve the adhesion because the formation of a Cu-azole complex was very slow. When the polymeric azole compound PBI was employed, an adhesion strength of 785 N/m was obtained since coverage of the copper surface with PBI was complete. Although the adhesion strength decreased upon storing the azole-treated copper coupons at 200C for 30 min in a convection oven, a good adhesion strength above 180 N/m was still maintained for CBTA and 8-azaadenine, and 500 N/m for PBI-treated joints. It appears that the decrease in adhesion strength comes from the oxidation of copper at 200C. In conclusion, the application of triazole compounds to a copper lead frame can improve the adhesion strength of EMC/copper lead frame joints and the of semiconductor devices by preventing the formation of a weak oxide. reliability A polymeric azole such as PBI is a better adhesion promoter than low-molecularweight triazole compounds because it provides complete coverage of the copper surface and superior thermal stability. The adhesion

561 Acknowledgement.s This work was supported by Samsung Electronics Center for Advanced Functional Polymers. and partially supported by the

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. A. J. Kinloch, Adhesion and Adhesives. Chapman & Hall, London (1987). S. G. Hong and T. C. Wang, J. Appl. Polym. Sci. 52, 2339 (1994). S. G. Hong, Angew. Makromol. Chem. 215, 161 (1994). B. J. Love and P. F. Packman, J. Adhesion 40, 139 (1993). H. K. Yun, K. Cho, J. H. Ahn and C. E. Park, J. Mater. Sci. 27, 5811 (1992). S. Kim, IEEE Trans. CHMT 14, 809 (1991). J. M. Park and J. P. Bell, in: Adhesion Aspects of Polymeric Coatings, K. L. Mittal (Ed.), pp. 205-204. Plenum Press, New York (1983). F. Mansfeld and T. Smith, Corrosion 27, 289 (1971). F. Mansfeld and T. Smith, Corrosion 29, 105 (1973). D. Chadwick and T. Hashemi, Surface Sci. 89, 649 (1979). F. P. Eng and H. Ishida, in: Corrosion Protection by Coatings and Surface Modification, M. D. Kending and K. Sugimoto (Eds). American Chemical Society, Washington, DC (1986). D. T. Clark, in: Handbook of X-rayand Ultraviolet Photoelectron Spectroscopy, D. Briggs (Ed.), pp. 211-247. Heyden, London (1977). J. Benard, Acta Metall. 8, 272 (1960).

Downloaded by [Simon Fraser University] at 11:15 29 December 2012