Laser micro joining of thin metal films on flexible substrates for mechanical and electrical connections M. Ehrhardt1, K. Zimmer1,* 1 - Leibniz Institut für Oberflächenmodifizierung e. V., Permoserstr. 15, 04318 Leipzig, Germany

Abstract Joining of similar or dissimilar materials with a thickness in the range of micrometers and sub micrometers is of great interest for a number of applications in, e.g., micro technology, photovoltaic and thin-film technology. A laser micro joining process using 25 ns long KrF excimer laser pulses for joining thin films and foils is demonstrated. Metal films of silver, aluminium, copper, molybdenum and titanium with thicknesses down to 500 nm deposited on polyimide substrates were used for bonding to a 12.5 µm thick silver foil. The laser fluencies used for joining of the foil to the metal films are in the range of 3.5 J/cm². The laser-induced joints were investigated by SEM (scanning electron microscope), optical microscopy, and a tensile strength tester. The shear stress calculated from the tensile force measurements and considering the laser-exposed area to be the bonding area is 0.5 N/mm² for silver/aluminium bonds. The tensile strength is not only determined by the bond between the metal films but also by the adhesion of the thin film to the substrate. Synergetic effects lead to bond formation comprising of thermal, mechanical, and chemical processes. Keywords: micro joining, thin film, welding, polyimide, foil, laser, nanosecond, metal

Introduction Due to the increasing significance of miniaturization and integration of micro devises the relevance of joining thin films for mechanical and electrical connections with sizes from millimetres to sub-micrometers increases dramatically. However, classic techniques such as wire bonding and soldering require specific metal layers, such as gold and a sufficient film thickness. Laser techniques like laser welding [1, 2] and laser soldering [3, 4] have a great potential due to, e.g. the focusing to spot sizes in the micrometer range for micro joining, the precise control of the applied energy, and the huge variety of material combinations which can be joined. Current studies aim also at the development of laser joining processes of thin metal films. Especially in the fast growing field of the so-called “plastic electronics” [5-7] and “flexible electronics” [8-10] there is a great interest for fast and flexible joining methods for packing and assembling. However, new joining options have to be developed (i) due to the much lower thickness of the films used in comparison to classical applications and (ii) due to the usage of thermal sensitive substrates, e.g. polymers or semiconductor materials. One possibility to minimize the influence of the laser-induced heat is the use of laser pulses with shorter pulse duration. Therefore, the needed joining approach is different to the classical welding and laser welding because traditionally the object of welding techniques is the formation of a sufficiently large molten bulk volume of material. This liquid pool propagates through the solid and causes the seam of the original gap between the components to be joined [11]. However, in the case of using laser pulses with a short pulse duration (< µs) the heat-effected zone and thus the liquid volume are drastically reduced. Up to now nanosecond and shorter laser pulses are rarely used for welding and joining. An exception is the so-called “transmission joining”. In this process one joining component has to be transparent for the used laser radiation. Thus, the laser beam penetrates the transparent material and is absorbed by the second absorbing material where heat is generated, which results in the joining. This approach has to be demonstrated for the joining of different material combinations like, e.g., glass/polyimide [12, 13],

Laser-based Micro- and Nanopackaging and Assembly V, edited by Wilhelm Pfleging, Yongfeng Lu, Kunihiko Washio, Jun Amako, Willem Hoving, Proc. of SPIE Vol. 7921, 792109 · © 2011 SPIE · CCC code: 0277-786X/11/$18 · doi: 10.1117/12.874920 Proc. of SPIE Vol. 7921 792109-1

copper/glass [14] and glass/silicon [15] joining. In the case of using femtosecond laser pulses both joining materials can be transparent [16]. However, only few studies have been presented on micro-welding processes of two metal foils which have a thickness in the range of µm and sub-µm [17, 18]. The reasons for that might be the low film thickness and the substrate damage risk as already mentioned. The damage thresholds for pulsed laser irradiation of thin metal films were studied by Matthias et al. with the result that the damage threshold at lower metal film thicknesses decrease [19]. Typical laser-induced damages are cracks, which may occur due to shockwaves during ablation, melting of the surface, or ablation at higher laser fluences. These processes depend on the thin-film material properties, the adhesion to the substrate and the properties of the substrate. Once the laser fluence exceeds a certain threshold explosive boiling and plasma formations occur. The relationship between vaporization of laser-heated metal and plasma formation is studied in Ref. [20]. The laser plasma is capable of absorbing photons and causing laser beam absorption, which results in plasma heating and reduces the metal surface irradiance. The basic processes involved in laser plasma absorption are discussed in Ref. [21, 22]. In this study the joining of thin films by pulsed laser radiation without auxiliary material laser is demonstrated. The laser bonding with ns-UV Laser were studied by SEM investigations and shear force measurements which showed the capabilities of this process

Experimental set-up For the experiments thin metal films (see Table 1) which have thicknesses from 500 nm up to 1.5 µm were deposited by magnetron sputtering on a 25 µm thick polyimide substrate (UPLEX). On the metal film a soft annealed silver (Ag) foil with a thickness of 12 µm was placed.

Material

Melting point [23]

Molybdenum (Mo) Titanium (Ti) Aluminium (Al) Silver (Ag) Cupper (Cu)

2620 °C 1670 °C 659 °C 961 °C 1083 °C

Thickness 600 nm 500 nm 1.5 µm 500 nm 1000 nm

Table 1: Summary of the used materials and the respective thicknesses for the lower metal film which was deposited at the polyimide substrate.

A schema of the geometry used for the joining experiments is given in Figure 1. A vacuum chuck was used to clamping the Ag foil on the top of the metal film (see Figure 2). Because the Ag foil was smaller than the metal film substrates an additional foil which was larger than the Ag foil and the metal films was used to fix the smaller Ag foil on top of the metal films and for sealing the whole vacuum chuck as well.

Figure 1: Schema of the arrangement for the used materials

Proc. of SPIE Vol. 7921 792109-2

Figure 2: Sketch of the vacuum chuck which was used to fasten the material system. An additional 25 µm thick polyimide foil was used to seal the complete chuck.

For the current investigations a KrF Excimer laser (LPX 220i, Lambda Physik) with a pulse length of tp = 25 ns and a wavelength of λ = 248 nm that is embedded in a laser workstation (Exitech, Ltd.) was used. In addition, the workstation comprises a beam shaping and homogenizing optics which provides a flat top beam profile. The flat top profile homogeneous heating across the joining width can be expected in contrast to a Gaussian intensity distribution which might lead to an overheating in the centre of the laser spot and, thus, may result into a damage of the polyimide substrate. The laser fluence and the repetition rate were fixed at a value of F = 3.5 J/cm² and frep. = 50 Hz, respectively. A Schwarzschild objective (15x demagnification with an optical resolution of 1.5 µm) was used for projecting a square aperture onto the sample. The final laser spot has a size of A = 100 µm x 100 µm at the sample surface. An x-y-z stage enables programcontrolled positioning and scanning of the laser beam across the sample surface. A computer-controlled laboratory tensile testing machine was used to measure the maximum shear force which can be loaded onto the join between the Ag foil and the metal films. To get an average bonding strength in the present experiments the bonding area consists of 75 joining points which are symmetrically arranged in three lines as sketched in Figure 3. The tensile direction of the tensile testing machine is indicated by the arrows in Figure 3

The share stress was calculated from the measured forces using the laser-irradiated area (spot size times the spot

Figure 3: Sketch of the joined metal film (lower layer) with the Ag foil (upper layer). The red quadrates indicate the joining points. The arrows show the direction of tension which is induced due to the shear force measurement.

number) to be the contact area of the joins. After the tensile tests of the samples, the failed surface in the joining area on both the Ag foil and the thin metal films covered the polyimide substrate were investigated with a scanning electron microscope (SEM).

Proc. of SPIE Vol. 7921 792109-3

5HVXOWVDQG'LVFXVVLRQ Typically joining areas of the upper Ag foil are shown in the SEM image in Figure 4. It can be seen that due to the laser irradiation the Ag foil was molten and partially ablated. As a result of that the thickness of the originally

Figure 4: SEM image of a joining area at the upper side of the Ag foil after the joining process. The lines which can be seen in the image are rolled-in grooves caused due to the manufacture process of the Ag foil.

12 µm thick Ag foil is reduced in the laser-irradiated joining area. With the presumption of the complete melting of the laser irradiated foil the remaining thickness of the Ag foil can roughly be estimated under the assumption that the remaining thickness is smaller or equal to the thermal diffusion length. The thermal diffusion length can be calculated with the formula [19]:

Ldiff = 2 * κ * t pulse

(1),

were κ is the thermal diffusivity and tpulse the pulse duration. With formula (1) the calculated thermal diffusion length is Ldiff = 2.9 µm under the given experimental conditions (κ = 173 x 10-6 m2/s and tpulse = 25 ns) is thermal diffusion length. The real thickness is properly thinner because (I) the underlying metal film has also heated up above a critical temperature which enables a joining, (II) the interface between the Ag foil and the metal film and a possibly remaining gap between the Ag foil and the metal film disable the heat flux from the upper Ag foil into the metal film, (III) due to the 3-dimensionality of the joining area a significant amount of heat could flux though the side of the joining area and is no longer available for heating up the actual joining area. The shear force which can be loaded onto the join until the join fails strongly depends on the used material combination. Typical failure shear force values which can be achieved for the different material combinations are summarized in Figure 5.

Figure 5: Failure shear force of a join between an Ag foil and different metal films on polyimide.

The laser pulse numbers per joining point were chosen so that a maximum joining strength between the Ag foil and the metal film was achieved. The used pulse numbers per join were 145, 145, 135, and 105 for Cu, Ag, Ti, and Al metal films, respectively. However, it has keep in mind that several of the applied laser pulses are necessary to ablate the polyimide foil which was used to seal the vacuum chuck. So that the laser pulse number which provide the actual joining is lower than the previous given laser pulse number. The used 75 joining points per join cover a calculated total joining area of A = 0.75 mm². The resulting calculated shear tensions of the join for the different material combinations have values of σtensile = 0.79 N/mm²,

Proc. of SPIE Vol. 7921 792109-4

σtensile = 0.62 N/mm², σtensile = 0.53 N/mm², σtensile = 0.11 N/mm² for a join between Cu/Ag, Ag/Ag, Ag/Al, Ag/Ti, respectively (see Figure 5). The achieved join strengths are much smaller than typical joining strengths on metalmetal joining made with, e.g., laser welding, brazing, or bonding [24]. However, none of these techniques is able to join metal films with thicknesses in the sub-micrometer range on top of a flexible substrate. A connection between the Ag foil and the Mo film was not possible under the given experimental conditions. No joining of the Ag/Mo and only a weak joining between the Ag/Ti material systems can possibly correlate with the high melting points of Mo and Ti (see Table 1). Furthermore, Ag and Mo are insoluble into each other and cannot form an alloy [25] which enables the metallurgical connection between the Ag foil and the Mo film. Because the laser energy is coupled onto the upper Ag foil and must diffuse to the joining area, the combination of the melting points of the materials may of importance. In comparison to the melting point of the Ag the films of Al, Ag and Cu have lower, similar and higher melting points respectively. At the Ag/Al and the Ag/Ag join the material with the highest melting point is the upper Ag. At the Ag/Cu join the two materials have roughly the same melting point. The measured failure shear force of the join between Ag/Al, Ag/Ag, and Ag/Cu does not differ very much; however, the join of the material combination of Ag/Cu which has the highest melting point has the highest joining strength. This can be caused by several effects. One group of effects may correlate with the metal-metal join itself, e.g., the influence of oxide coating or the potential and strength of alloy generation with consideration of the time scale of the used pulse duration.

Figure 6: SEM micrographs of joining points after shear force measurement: a) lower side of the upper Ag foil, b) top view at the lower Ag film. The selected joining points originate from a join between an Ag foil with an Ag film.

To investigate the reasons of the joins failure in detail, SEM investigation of the joining points of the Ag/Ag join after rupture the upper Ag foil from the lower Ag film were done. In Figure 6a) the lower side of the upper Ag foil an imprint of the used laser beam can be seen in the metal foil, due to several processes like fast heating, material softening, thermal expansion, or pressure generation due to metal evaporation, and plasma formation, which were induced by the laser radiation in the metal foil. The laser-induced pressure generation is well known in the literature and is used for several laser processes like laser peening [26] and laser shock forming [27]. At some experimental conditions within the joined area pieces of the Ag film can be found attached onto the Ag foil as seen in Figure 6b).Further, shown in Figure 6b) parts of the deposited Ag film have been peeled off from the polyimide substrate during the shear force measuring. This observation gives an evidence of a main failure mechanism of the join – the insufficient adhesion between the metal foil and the metal film. One main effect which determines the strength of the join is the adhesion strength of the metal film onto the polyimide substrate. These could also be the explanation for the higher measured failure tensions of the Cu/Ag join compared to the Ag/Ag join (see Figure 5). Because the thickness of the Cu film is twice as large as that of the Ag film it can be assumed that the mechanical strength of the Cu film is higher than that of the Ag film. Additionally, the adhesion strength between the Cu film/polyimide substrate differs from the Ag film/polyimide substrate. The black spot in the centre of the SEM image in Figure 6b) is the polyimide substrate which is charged due to the SEM investigations. However, the polyimide substrate seems to be undamaged after the joining process and the peeling off within the laser spot area. This is notably because the decomposition temperature of polyimide (Tdecomp = 550 °C) [28] is lower than the melting temperature of Ag and the other applied metal films.

Proc. of SPIE Vol. 7921 792109-5

That effect could be explained by the decomposition dynamics of polyimide that might allow higher

figure 7: SEM images of the joining area between two 12 µm thick Ag foils after the shear force measurements. In a) one joining point at the lower side of the upper foil can be seen, b) shows a section of the joining area with several joining points.

temperatures within the short laser pulse heating time [29]. To confirm the assumption that the substrate and the adhesive strength between the polyimide and the metal film influence the maximum shear force which can be loaded onto the join a verifying experiment was done. For that two Ag foils with a thickness of 12 µm were joined together with similar experimental conditions as with the joining of the Ag foil and the Ag film described previously. The maximum measured shear tension which can be loaded onto a join between two Ag foils is σtensile = 1.1 N/mm². This value is about two times larger than the maximum shear tension which was measured on a join between the Ag foil and the 500 nm thick Ag film. To verify whether the failure mechanism of the Ag/Ag foil joining is comparable to the failure mechanism of the Ag foil/Ag film joining, SEM images after the shear force measurement were done as well. In figure 7a) the lower side of the top Ag foil and in figure 7b) the upper side of the lower Ag foil are shown. It can be seen in figure 7a) that in the case of joining two Ag foils the joining area of the upper foil tears off. This behaviour can be explained by the previous estimation of the remaining thickness of the joining area of the upper Ag foil. In this estimation it was shown that the remaining thickness of the upper Ag foil is less then 2.9 µm so that the mechanical strength of the upper foil is much smaller than that of the lower foil. The failure SEM image together with the estimation of the remaining thickness of the upper Ag foil shows that a maximum theoretical limitation of the joining strength is given by the mechanical strength of the remaining material thickness of the upper Ag foil. The remaining material thickness is essentially determined by (I) the coefficient of the thermal diffusivity and the pulse duration how it can be derived form formula (1) and (II) and the temperature which is necessary to melt the lower metal film of the stack which has to be joined together. In the SEM images of the lower foil the imprint of the used beam profile can be seen. These imprints in combination with the imprints seen in the upper Ag foil strongly suggest that during the exposure time by the laser the upper and lower Ag foil have a very closed contact. The force which is needed to press the upper Ag foil onto the lower foil is probably provided by the recoil pressure which was induced into the material system by thinning the upper Ag foil. However, this secondary effect helps to overcome the disturbing influence of the air gap between two work pieces which have to be joined together by laser welding. Due to the laser-induced recoil pressure and the flat top beam profile it is ensured that a sufficient large joining area is in a very closed contact so that a sufficient heat is transferred from the upper metal foil and the lower metal film or foil. Future simulations of the laser-induced heating and the heat transfer will more exactly show which temperatures can be expected in the interface region between the lower side of the upper Ag foil and the lower metal film. Such results from the thermal simulations may allow the estimation of suited conditions, e.g. laser exposure conditions, for joining of functional material combination for the desired application

Proc. of SPIE Vol. 7921 792109-6

6XPPDU\ In this study, laser microjoining of two silver metal foils and of silver foils with thin films of various metals on polyimide foil are shown. This is of technical interest for the fabrication of electrical contacts or mechanical bonding in microelectronic and microsystem technology. Especially the joining of 12 µm thick Ag foil and metal films from 1.5 µm down to 500 nm was investigated. The obtained results demonstrate that a laser microjoining of a thin foil with a submicron thin film at least for the material combinations Ag/Al, Ag/Ag, Ag/Cu, and Ag/Ti is possible. The joining between the Ag foil and the Mo film, however, was not achieved which could be the result of the very different melting temperatures of the metals. The joining strength of the connections between the Ag foil and the various metal films were tested mechanically using a tensile testing machine. The observed shear stress values that were calculated simply by using the laserirradiated areas were in a range between 0.11 N/mm² for Ag/Ti and 0.79 N/mm² for Ag/Cu joining. However, it is suggested that the contact area is less the laser spot size although a homogenized beam was used. The joining strength is strongly influenced by the adhesion of the thin film on the polyimide substrate as the SEM investigations suggest. The polyimide substrate, however, is obviously undamaged due to the joining process although the decomposition temperature of polyimide is well below the achieved joining temperatures. During the laser-induced joining process the upper, laser-irradiated Ag foil was thinned in the joined areas by laser ablation before a joining of both the Ag foil and the metal film was achieved. The comparison of the thermal diffusion length at the laser pulse length with the remaining thickness of the Ag foil that is about 2.9 µm suggests the importance of melting and material softening for the bond forming. The maximum bonding strength at the joining of two Ag foils is limited by the remaining material thickness of the upper Ag foil after the joining process as it can be suggested from the shear force measurements and the failure analysis by SEM.

Reference [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16]

C. T. Dawes, [Laser Welding: A Practical Guide] Crc Press, (1992). W. W. Duley, [Laser Welding] John Wiley &Sons, (1998). F. Bachmann, “Industrial applications of high power diode lasers in materials processing,” Applied Surface Science, 208, 125-136 (2003). M. Brandner, G. Seibold, C. Chang et al., “Soldering with solid state and diode lasers: Energy coupling, temperature rise, process window,” Journal of Laser Applications, 12(5), 194-199 (2000). A. Scandurra, G. F. Indelli, N. G. Sparta et al., “Low-temperature sintered conductive silver patterns obtained by inkjet printing for plastic electronics,” Surface and Interface Analysis, 42(6-7), 1163-1167. W. A. MacDonald, “Engineered films for display technologies,” Journal of Materials Chemistry, 14(1), 4-10 (2004). H. Klauk, [Organic Electronics] Wiley-VCH Verlag, 446 (2006). W. S. Wong, [Flexible Electronics: Materials and Applications], (2009). P. Gecys, G. Raciukaitis, M. Ehrhardt et al., “ps-laser scribing of CIGS films at different wavelengths,” Applied Physics a-Materials Science & Processing, 101(2), 373-378. F. Kessler, D. Herrmann, and M. Powalla, “Approaches to flexible CIGS thin-film solar cells,” Thin Solid Films, 480, 491-498 (2005). V. A. Ventrella, J. R. Berretta, and W. de Rossi, “Pulsed Nd:YAG laser seam welding of AISI 316L stainless steel thin foils,” Journal of Materials Processing Technology, 210(14), 1838-1843. A. Mian, G. Newaz, D. G. Georgiev et al., “Performance of laser bonded glass/polyimide microjoints in cerebrospinal fluid,” Journal of Materials Science-Materials in Medicine, 18(3), 417-427 (2007). N. Amanat, N. L. James, and D. R. McKenzie, “Welding methods for joining thermoplastic polymers for the hermetic enclosure of medical devices,” Medical Engineering & Physics, 32(7), 690-699. Y. Ozeki, T. Inoue, T. Tamaki et al., “Direct welding between copper and glass substrates with femtosecond laser pulses,” Applied Physics Express, 1(8), 3 (2008). C. Luo, and L. W. Lin, “The application of nanosecond-pulsed laser welding technology in MEMS packaging with a shadow mask,” Sensors and Actuators a-Physical, 97-8, 398-404 (2002). T. Tamaki, W. Watanabe, J. Nishii et al., “Welding of transparent materials using femtosecond laser pulses,” Japanese Journal of Applied Physics Part 2-Letters & Express Letters, 44(20-23), L687-L689 (2005).

Proc. of SPIE Vol. 7921 792109-7

[17]

[18]

[19]

[20] [21] [22] [23] [24] [25] [26] [27] [28] [29]

D. P'ng, and P. Molian, “Q-switch Nd : YAG laser welding of AISI 304 stainless steel foils,” Materials Science and Engineering a-Structural Materials Properties Microstructure and Processing, 486(1-2), 680-685 (2008). T. Hirano, M. Demura, K. Kishida et al., “Laser spot welding of cold-rolled boron-free Ni3Al foils,” Metallurgical and Materials Transactions a-Physical Metallurgy and Materials Science, 38A(5), 10411047 (2007). E. Matthias, M. Reichling, J. Siegel et al., “THE INFLUENCE OF THERMAL-DIFFUSION ON LASER-ABLATION OF METAL-FILMS,” Applied Physics a-Materials Science & Processing, 58(2), 129-136 (1994). S. Petzoldt, J. Reif, and E. Matthias, “Laser plasma threshold of metals,” Applied Surface Science, 968, 199-204 (1996). P. Mora, “THEORETICAL-MODEL OF ABSORPTION OF LASER-LIGHT BY A PLASMA,” Physics of Fluids, 25(6), 1051-1056 (1982). P. Mora, and H. Yahi, “THERMAL HEAT-FLUX REDUCTION IN LASER-PRODUCED PLASMAS,” Physical Review A, 26(4), 2259-2261 (1982). U. Fischer, M. Heinzler, F. Näher et al., [Tabellenbuch Metall] Verlag Europa Lehrmittel, Düsseldorf, 432 (2008). J. D. Majumdar, and I. Manna, “Laser processing of materials,” Sadhana-Academy Proceedings in Engineering Sciences, 28, 495-562 (2003). H. Okamoto, [Phase Diagrams for Binary Alloys] ASM International, (2000). C. S. Montross, T. Wei, L. Ye et al., “Laser shock processing and its effects on microstructure and properties of metal alloys: a review,” International Journal of Fatigue, 24(10), 1021-1036 (2002). H. X. Liu, Z. B. Shen, X. Wang et al., “Micromould based laser shock embossing of thin metal sheets for MEMS applications,” Applied Surface Science, 256(14), 4687-4691. S. Heltzel, and C. O. A. Semprimoschnig, “A detailed study on the thermal endurance of Kapton HN (R) and Upilex S (R),” High Performance Polymers, 16(2), 235-248 (2004). N. Arnold, and N. Bityurin, “Model for laser-induced thermal degradation and ablation of polymers,” Applied Physics a-Materials Science & Processing, 68(6), 615-625 (1999).

Proc. of SPIE Vol. 7921 792109-8