Effect of Ni-Cr Seed Layer Thickness on the Adhesion Characteristics of

Met. Mater. Int., Vol. 16, No. 5 (2010), pp. 779~784doi: 10.1007/s12540-010-1013-8 Published 26 October 2010Effect of Ni-Cr Seed Layer Thickness on the Adhesion Characteristics of Flexible Copper Clad Laminates Fabricated using a Roll-to-Roll ProcessBo-In Noh, Jeong-Won Yoon,and Seung-Boo Jung*School of Advanced Materials Science and Engineering, Sungkyunkwan University,Suwon-si, Gyeonggi 440-746, Korea(received date: 16 October 2009/accepted date: 26 April 2010)The adhesion strength of a Cu/Ni-Cr/polyimide flexible copper clad laminate (FCCL), which was manufacturedvia a roll-to-roll process, was evaluated according to the thickness of the Ni-Cr seed layer using a 90° peeltest. The changes in the morphology, chemical bonding, and adhesion properties were characterized usinga scanning electron microscopy (SEM), an atomic force microscopy (AFM), and an X-ray photoelectronspectroscopy (XPS). The thickness of the Ni-Cr (Ni:Cr = 80:20) seed layer in which the maximum peelstrength of the FCCL was observed was 200Å. The higher FCCL peel strength was attributed to the lowerproportion of C-N bonds and higher proportion of C-O and carbonyl (C = O) bonds in the polyimidesurface compared with the FCCL with a lower adhesion strength. The FCCL with a higher peel strengthhad a fractured polyimide surface with a higher surface roughness. The adhesion strength between themetal and polyimide was pirmarily attributed to the chemical interaction between the metal layer and thefunctional groups of the polyimide.Keywords: electronic materials, plating, interfaces, atomic force microscopy (AFM), flexible copper clad lami-nate (FCCL)1. INTRODUCTIONA printed circuit board (PCB) is a micro-thin conducting circuit, either rigid or flexible, that is fabricated on a dielectric board for locating electronic devices. A flexible PCB (FPCB) has many advantages compared with a rigid board due to its flexibility, especially in applications demanding more integrated electronic devices, and its ability to be mass produced using roll-type manufacturing systems. A flexible copper clad laminate (FCCL) is generally employed as a raw material for FPCBs. FCCLs fabricated by sputtering have advantages in terms of fine pitch etching and dimensional accuracy compared with FCCLs fabricated by casting or laminating. FCCL is a system which brings together an elec-tric conductor such as copper (Cu) and an insulator such as polyimide. Polyimide has desirable properties of high tem-perature stability, good mechanical strength, good dimen-sional stability, good chemical stability, and a low dielectric constant. Accordingly, Cu coated polyimide films are com-monly used as FPCBs in the electronics industry [1,2]. Adhesion between the polyimide and Cu is an important issue in the field of electronic packaging devices, such as multi-chip modules and FPCBs. In general, the adhesion between a polymer and a metal is known to depend on the strength of the chemical bonding, physical interactions, and mechanical interlocking at the interface. The poor adhesion properties of Cu to polyimide need to be overcome to increase the flexibility of the fabricated devices because the stress applied during the bending is increasingly focused on the interface between the metal lines and the polymer substrate with a decreasing line width. In order to improve the interfacial adhesion strength, an additional metal layer is deposited at the Cu/polyimide interface. The adhesion between the polyimide surface and Cu can be improved by pre-treating the former with a plasma, ion beam or chemi-cal solutions, which induces mechanical, chemical or phys-ical changes. In addition, an adhesion promoting layer such as chromium (Cr), nickel (Ni), titanium (Ti), Ni-Cr or indium tin oxide (ITO) can be inserted between the modified poly-imide and Cu via physical vapor deposition, in order to minimize the adhesion loss, which is normally caused by the hostile conditions of temperature and humidity [3]. In particular, Ni and Cr have been used as a typical interface metal layer [3-13]. Cr is used as an interface metal layer between the Cu and polyimide films due to the formation of strong Cr-C and Cr-O bonds [5,7]. The use of Cr was also intended to prevent Cu diffusion into the polyimide*Corresponding author: sbjung@skku.ac.kr ©KIM and Springer780Bo-In Noh et al. and to maintain the insulating properties of the polyimide [8]. Also, Ni-Cr thin films find applications in integrated circuits where low noise, good power dissipation, and anear zero temperature coefficient of resistance (TCR) are important requirements. Some papers have been publishedon the deposition of Ni-Cr films by thermal evaporationand radio frequency (RF) sputtering [14,15].Many studies have been performed on the adhesion prop-erties of various Cu/polyimide films with different interlay-ers and pre-treatments [11,16-18]. Nevertheless, the current knowledge of the adhesion properties of the Cu/Ni-Cr/ polyimide system remains insufficient. Therefore, in this study, the adhesion properties of Cu/Ni-Cr/polyimide filmswere investigated according to the thickness of the Ni-Crseed layer using the 90° peel test. The test results are dis-cussed in connection with the fracture analysis results thatwere also observed in this study.2. EXPERIMENTAL PROCEDUREThe FCCL structure investigated in this study was Cu/Ni-Cr/ polyimide. The substrate was 25µm thick polyimide film (Kaneka, Japan) that had not undergone any plasma or chemical surface treatment. The ion beam treatment andtwo sputtering steps were performed in the same vacuum chamber using the roll-to-roll process. The base pressureand roll speed of the sputtering system were 10−4 to 10−6 Paand 0.5 m/min, respectively, for both the ion beam treat-ment and sputtering processes. Firstly, the ion beam treat-ment was performed on the polyimide under argon (Ar) gas conditions. The constant current and flow rate were 3.5 A and 15 standard cubic centimeters per minute (sccm), respectively. Secondly, a Ni-Cr seed layer was sputtered on the polyimide under Ar gas conditions. The Ni:Cr ratio of the Ni-Cr seed layer was 80:20. Three Ni-Cr thicknesses were investigated: 100Å, 200Å, and 300Å. The power and gas flow rate for the three samples are listed in Table 1. Lastly, a 0.2µm thick Cu layer was sputtered on the poly-imide under Ar gas conditions. The power and flow rate were 2.0 kW and 50 sccm, respectively. After the sputter-ing of the Cu, the samples were subsequently electroplated with pure Cu to a thickness of 8µm. The current density and roll speed for the Cu electroplating were 1.3 A/dm2 (ASD) and 320 mm/min, respectively. Figure 1 shows the manufacturing process of the FCCL used in this study. The dimensions of the polyimide and Cu electroplating layers for the peel test were 100 mm × 10 mm and 100 mm × 3 mm, respectively. Figure 2(a) shows a schematic of the FCCL manufactured for the peel test. The samples under-went 90° peel testing conducted at 4 mm/min with 10 spec-imens at each condition. Figure 2(b) shows the schematic of the 90° peel test. After the peel test, the fracture surface and composition on the Cu and polyimide surfaces were identified using a scanning electron microscopy (SEM, S-3000H, Hitachi, Japan) and an energy dispersive X-ray spectroscopy (EDS). The morphologies of the fracture sur-faces were also studied using an atomic force microscopy (AFM, Thermo-Microscopes, CP Research, USA). In each case, an area of 10µm × 10µm was scanned using a tap-ping mode. The AFM observation was undertaken at atmo-sphere pressure and room temperature. The root-mean-square (RMS) roughness was calculated from the rough-ness profile determined by the AFM. An X-ray photoelectron spectroscopy (XPS, ESCA 2000 LAB MK-II spectrometer, VG Microtech, England) analysis with a MgKαX-ray source was performed on the peeled polyimide surface to elucidate the chemical bonding state at the interface. The base pressure in the sample chamber was controlled in the range of 10−8 to 10−9 Torr.3. RESULTS AND DISCUSSIONThe 90° peel test has been widely used to measure the adhesion strength of a metal film on a polymer substrate. Figure 3 shows the results of the peel strength test of thethree FCCLs according to the thickness of the Ni-Cr seed Fig. 1.Manufacturing process of the FCCL used in this study.Fig. 2. Schematics of (a) the top view of the FCCL and (b) the 90°peel test.Effect of Ni-Cr Seed Layer Thickness on the Adhesion Characteristics of Flexible Copper Clad Laminates (781)layer. The peel strength of the FCCLs increased until the thickness of the Ni-Cr seed layer reached 200Å and then slightly decreased for the Ni-Cr seed layer with a thickness of 300Å. However, the magnitude of this decrease in the peel strength was not significant: the peel strength was sat-urated at the thickness of the Ni-Cr seed layer at 200Å.This means that the FCCL with the 200Å thick Ni-Cr seed layer had a higher peel strength than that of the FCCLs with the 100Å thick and 300Å thick Ni-Cr seed layers.These results demonstrate that the optimum thickness of the Ni-Cr seed layer in order to maximize the peel strength was 200Å for the FCCL with a Ni:Cr ratio of 80:20 and 8µm thick electroplated Cu layer. Previous studies have shown that the peel strength of the FCCL increased with an increasing Cr content and Cu electroplating layer thick-ness, and that the Cr content had a greater effect on the peel strength than the Cu electroplating layer thickness [17,18].Generally, good adhesion between the metal layer and polymer was attributed not only to mechanical interlock-ing, but also to interfacial chemical bonds. In particular, it is known that the adhesion strength is significantly affected by the chemical interaction between the metal layer and the functional groups of the polymer [5,7]. It has been reported that some metals can interact with oxygen containing groups on polymer surfaces, thereby creating a metal-oxygen-car-bon bonding [5]. From these results, it was inferred that the improvement in adhesion was due to the resulting changes in the characteristics of the polyimide surface. It is well known that the ion bombardment of a polymer induces the scission of the chains, due to their collision with the inci-dent ions, causing them to undergo further reactions such as cross linking, carbonization, and chemical reactions.These reactions lead to changes in the polymer surface,resulting in increased chemical bonding, physical interac-tion, and mechanical interlocking between the polyimideand metal layer. Therefore, in this study, the peel strength increased with an increasing Ni-Cr thickness from 100Å to 200Å, due to the increasing ion bombardment caused by the higher power used in the Ni-Cr sputtering process, as shown in Table 1.Figure 4 shows the SEM images and EDS analysis results of the FCCL with the 200Å thick Ni-Cr seed layer, before and after the peel test. The examination of Figs. 4(a) and (b)revealed that the adhesion interface of the manufactured FCCL was very uniform and that the fracturing observed after the peel test occurred between the polyimide and metal layer (Ni-Cr/Cu). Figures 4(c) and (d) show that after the peel test, oxygen, copper, nickel, and chromium elements were detected on the Cu side of the FCCL. Only carbon and oxygen were detected on the polyimide side; thus, no metal elements were detected on the polyimide side. These results indicate that a failure occurred at the metal/polyimide inter-face. The failure site and fracture mode on the polyimide of the FCCL with the 100Å thick or 300Å thick Ni-Cr seed lay-ers were similar to the results shown in Fig. 4.Figure 5 shows the AFM images of the fracture surfaces (polyimide sides) for the three samples after the peel test.The RMS roughnesses of the fracture surfaces of the FCCLswith the 100Å thick, 200Å thick and 300Å thick Ni-Cr seedFig. 3. Peel strength of the FCCLs according to the thickness of theNi-Cr seed layer.Table 1. Power and flow rate used in the Ni-Cr sputtering process Thickness of seed layer (Å)Power (kW)Flow rate of gas (sccm)1000.5502000.83001.1Fig. 4. SEM images and EDS analysis results of the FCCL with the 200Å-thick Ni-Cr seed layer: (a) cross-section of the FCCL before peel testing, (b) cross-section of the FCCL after peel testing, (c) Cu side of the FCCL after peel testing and (d) polyimide side of the FCCL after peel testing.782Bo-In Noh et al.layers were 18.96 nm, 24.63 nm, and 20.35 nm, respectively.From these results, it was identified that the fracture sur-face of the FCCL with the 200Å thick Ni-Cr seed layer had the highest RMS roughness. Some previous studies indi-cated that surface roughness is an important factor in enhanc-ing adhesion properties [5,6]. In fact, the surface roughnessFig. 5. AFM images of the fracture surfaces (polyimide sides) for the three samples after the peel test: (a) 100Å, (b) 200Å and (c) 300Å.The scale of the vertical range in the figure is different.Fig. 6. XPS results for C 1s spectra of the fracture surfaces (polyimide side) for the three samples: (a) 100Å, (b) 200Å and (c) 300Å.Effect of Ni-Cr Seed Layer Thickness on the Adhesion Characteristics of Flexible Copper Clad Laminates (783)of the fractured polyimide surface increased with the increas-ing adhesion strength of the FCCL. This rougher fracture surface for the FCCL with a higher adhesion strength was attributed to the higher bond ratio between the Ni-Cr seed layer and the functional groups on the polyimide surface, as described in the next two paragraphs.An XPS analysis was conducted to investigate the chem-ical bonding state at the peeled interface of Ni-Cr/polyim-ide. Figure 6 shows the XPS results of the C1s spectra for the fractured polyimide surfaces after the peel test. The FCCL with the 100Å thick Ni-Cr seed layer had a C-N ratio of 48.77 %, a C-O ratio of 37.27 %, and a carbonyl (C = O) ratio of 13.96 %. However, the FCCL with the 200Åthick Ni-Cr seed layer had a C-N ratio of 39.15 %, a C-O ratio of 41.17 %, and a carbonyl (C = O) ratio of 19.68 %. Also, the FCCL with the 300Å thick Ni-Cr seed layer had a C-N ratio of 41.43 %, a C-O ratio of 39.74 %, and a car-bonyl (C = O) ratio of 18.83%. The ratio of C-N bonds decreased with an increasing Ni-Cr thickness from 100Åto 200Å, whereas the ratios of the C-O and carbonyl (C = O) bonds increased with an increasing thickness of the Ni-Cr seed layer. These results indicate that the FCCL with the 200Å thick Ni-Cr seed layer has a lower C-N ratio, but higher C-O and carbonyl (C = O) ratios, than that of the 100Å thick Ni-Cr seed layer. Thus, a connec-tion can be made between thickness of the Ni-Cr seed layer in the Cu/Ni-Cr/polyimide structure or the power used in the Ni-Cr sputtering process affecting the chemi-cal groups on the polyimide surface.Previous studies reported that suitable functional groups, such as the hydroxyl (–OH), carbonyl (C = O), and carboxyl (COOH) species, promote the adhesion strength at the metal/polymer interface [5,11]. It has also been reported that the adhesion of Cr to polyimide occurs through the for-mation of Cr-O complexes or Cr oxides at the interface as a result of the reaction of Cr with the pendent oxygen atoms in the polyimide [13] or through the reaction of Cr with the oxygen functionality, in particular carbonyl groups, in the pyromellitic dianhydride (PMDA) units in the polyimide [5,12]. Therefore, the carbon atoms in the polyimide film, which became a constituent of the C-N bonding, may have been rearranged during the Ni-Cr sputtering. This increase in the ratios of C-O or C = O bonds for the FCCL with a thicker Ni-Cr layer was attributed to the reaction of Cr with the oxygen atoms during the Ni-Cr sputtering. These results suggest that the increase in the C-O and carbonyl (C = O) ratios increased the adhesion strength between the Ni-Cr layer and the polyimide. In other words, the increased for-mation of functional groups identified as C-O and carbonyl (C = O) bonds, as well as the reduction in the C-N compo-nent ratio, may have contributed to the enhanced chemical interactions between the Ni-Cr and polyimide surfaces, thereby increasing the peel strength.4. CONCLUSIONSIn this study, the mechanical reliability, as measured by adhesion strength, of FCCLs with Ni-Cr seed layers of dif-ferent thickness were investigated using the 90° peel test. The FCCL with the Ni-Cr (Ni:Cr = 80:20) seed layer with a thickness of 200Å had a higher adhesion strength than that with a thickness of 100Å. However, increasing the thickness of the Ni-Cr seed layer to 300Å did not signifi-cantly affect the adhesion strength between the Cu and polyimide. The morphological observation of the fracture surface revealed that the surface roughness of the fractured polyimide surface increased with an increasing adhesion strength of the FCCL. The FCCL with the 200Å thick Ni-Cr seed layer had a lower C-N bond ratio, but higher C-O and carbonyl (C = O) bond ratios. These results suggest that the increase in the C-O and carbonyl (C = O) ratios increased the adhesion strengths between the Ni-Cr seed layer and polyimide. Therefore, the adhesion strength between metal and polyimide was mostly attributed to the chemical effects arising from the enhanced interaction between the metal layer and the functional groups of the polyimide. ACKNOWLEDGMENTSThis study was supported by a grant from the Fundamen-tal R&D Program for the Core Technology of Materials funded by the Ministry of Knowledge Economy, Republic of Korea.REFERENCES1.S. Kamiya, H. Furuta, and M. Omiya, Surf. Coat. Technol.202, 1084 (2007).2.F. Barlow, A. Lostetter, and A. Elshabini, Microelectron.Reliab.42, 1091 (2002).3.E. Liston, L. Martinu, and M. Wertheimer, J. Adhes. Sci.Technol.7, 1091 (1993).4.J. Y. Song and J. Y u, Acta mater.50, 3985 (2002).5.S. H. Kim, S. W. Na, N. E. Lee, Y. W. Nam, and Y. H. Kim,Surf. Coat. Technol.200, 2072 (2005).6.Y. B. Park, I. S. Park, and J. Y u, Mater. Sci. Eng. A266, 261(1999).7.T. Miyamura and J. Koike, Mater. Sci. Eng. A.445-446,620 (2007).8.L. P. Buchwalter and K. Holloway, J. Adhes. Sci. Technol.12, 95 (1998).9.K. J. Min, M. H. Jeong, K. H. Lee, Y. S. Jeong, and Y. B.Park, J. Kor. Inst. Met. & Mater. 47, 675 (2009).10.J. L. Jordan, P. N. Sanda, J. F. Morar, C. A. Kovac, F. J.Himpsel, and R. A. Pollak, J. Vac. Sci. Technol. A4, 1046 (1986).11.J. S. Eom and S. H. Kim, Thin Solid Films516, 4530 (2008).784Bo-In Noh et al.12.R. Haight, R. C. White, B. D. Silverman, and P. S. Ho, J.V ac. Sci. Technol. A6, 2188 (1988).13.N. J. Chou, D. W. Dong, J. Kim, and A. C. Liu, J. Electro-chem. Soc.131, 2335 (1984).14.M. I. Birjega, C. A. Constantin, I. T. Florescu, and C. Sarbu,Thin Solid Films92, 315 (1982).15.S. Schiller, U. Heisig, K. Goedicke, and H. Bilz, Thin SolidFilms119, 211 (1984).16.I. S. Park and J. Y u, Acta mater.46, 2947 (1998).17.B. I. Noh and S. B. Jung, J. Electron. Mater.38, 46 (2009).18.B. I. Noh, J. W. Yoon, B. Y. Lee, and S. B. Jung, J. Mater.Sci. Mater. Electron.20, 885 (2009).。