SURFACE TOPOGRAPHY AND MECHANICAL PROPERTIES OF AU ON PDMS

Seyed M. Allameh^*, Onobu Akogwu⁺ and Wolé Soboyejo⁺

^*Northern Kentucky University, Highland Heights, KY 41099

⁺Princeton Institute for the Science and Technology of Materials, and Department of Mechanical and Aerospace Engineering, Princeton University, Princeton, NJ 08540

Abstract

This paper presents the results of a preliminary study on the surface topography and mechanical properties of Au films deposited on PDMS substrates. The microstructure of the Au films was studied by atomic force microscope. The grain size of the films increased linearly for Au thicknesses of 25-75 nm. Young modulus of the PDMS substrates increases with strain rate. The reduced modulus obtained by nanoindentation varies with the indent size but flattens out at indent sizes more than twice the thickness of the film. The variation in the electrical properties of these films was studied during microtensile testing of strips made of Au on PDMS substrates. Relaxation of stress in the films was shown to lead to the reversal of the of conductivity loss in the Au films due to stretching. The implications of the results are discussed in relationship with the reliability of metallization lines on PDMS substrates used for flexible displays.

I. Introduction

A developing area of macro-electronics is integrated circuits on flexible or deformable substrates [1]. Potential applications for such devices include flat panel displays, e-textiles, and sensitive skin for robotics, medical prostheses and sensor technology [1-5]. Depending on the circuit’s eventual application, the circuit may undergo a one-time deformation or be deformed a number of times. Such a circuit application in sensitive skin will require it to be deformed a number of times through cyclic loading or by tensile stretching. Previously, experiments have been performed on different materials in an effort to produce a more stretchable and electrically conductive interconnect. In certain instances, inherently conductive polymers were explored. In other cases, non-conductive polymers were made electrically conductive by suspending metal particles in the polymer [2]. However, both attempts have proved inadequate as the polymers were found to be brittle and had little ability to withstand sufficient strain, which is desirable for applications. In another approach, electronic components were grown on individual islands on a flexible substrate, and these components were integrated with metal interconnects. The islands help to protect the components from strain-induced deformation [3]. Typically, the conductive interconnects were thin metallic films with low critical strain and was thus not compliant.

In an effort to achieve the desired properties in metal interconnects with the ability to withstand sufficient strains due to tensile or compressive forces; experiments and theoretical studies have been carried on conducting films deposited on flexible substrates [4, 5]. These include: simple stretch tests, deformation into complex shapes [3], mechanical cycling and stress relaxation tests [4-7]. Research has been done to develop our basic understanding of cracking, and how cracking influences the resistance of metallic films on flexible polymeric substrates [8].

Recent experiments have yielded flexible and conductive metal interconnect that can withstand higher percent strain than the theory has predicted [4-6, 9]. In most cases, the metal-films are deposited on poly-dimethyl-siloxane (PDMS). While some prior works suggests that PDMS is elastic; others state its viscoelastic behavior [4]. At slow strain rate, PDMS has a Young’s modulus of 1MPa, and it is not expected to suppress crack formation in the metal-film [1, 4-7]. The crack formation coupled with suggested diffusion of the film between grain boundaries is vital to the resistive behavior of interconnect. The polymer substrate helps to disperse the strain locally allowing specimen elongation beyond that of theoretical free standing film. It has been theoretically shown that a thin film with no initial flaws deposited on plastic Kapton with a Young’s modulus of 1GPa with will allow to film to stretch to 80% whereas the free- standing film fails at 1-2%.

It is observed that stress and strain play critical roles in determining the stretchability. Stress build-up in the sample either in the metal-film or the polymer substrate can be due to a number of reasons: mismatch in strain field due to defects in film: surface scratches, bending of substrate before and during deposition of film and thermal strains that can occur if the film is deposited at temperatures greater than room temperature. In most cases, the metal film accommodates this stress by buckling or by forming micro cracks in the metal film.

Material

The PDMS substrates used to fabricate the Au/PDMS structures were fabricated as follows: Initially the pre-polymer Slygard 184 (Dow Corning, Cleveland OH) was mixed with a curing agent in a 10:1 ratio at room temperature. The mixture was then cast into plastic Petri dishes and left to cure for 12 h at a temperature of 60 ˚C. The time required to cure varies with the curing temperature. The substrate surface was not treated before metal evaporation. The metal film was deposited on the substrate by electron beam evaporation. The shadow mask used to pattern the sample was a flexible polyimide foil (DuPont, Kapton) that provided an easy and dry technique for patterning metal films. Prior to evaporation the Kapton mask was cleaned with acetone and isopropanol (IPA). The PDMS and Kapton were mounted on a rotational stage and held flat by clips. A 5-nm adhesion layer of chromium was evaporated on the PDMS substrate. The metal film of gold was then evaporated after the chromium adhesive layer on the PDMS. The metals were deposited on rotating substrates with a deposition rate of 2Å/s at a vacuum of 8 x 10^-6Torr. The temperature of the substrate holder was nearly 26 ˚C throughout the deposition process. The cast PDMS substrates were cut into thin strips with a width of 3-5 mm and lengths of ~ 25 mm for microstructural characterization and mechanical testing. The schematic of the test structures is shown in (Figure 1)

Experimental Procedure

For the evaluation of mechanical properties of the Au films deposited on the PDMS substrates, a microtensile tester, developed at Princeton University was used [microtensile]. The system included an Entran 2-lb load cell, connected to a uni-slide displacement system at one end and to an air bearing at the other end. Grips appropriate for the PDMS substrates were attached to the air bearing and to the stationary platform. The PDMS strips with and without gold films were mounted onto these grips. The samples were of 4-5 mm wide and 25 mm long. The samples were placed onto the grips with the gold film facing the camera and the light source. The latter were mounted in front of the tester to record images during the deformation of the sample. Image acquisition was performed using Labview program along with IMAQ Vision and Vision Builder[1] from National Instruments. Image analysis was performed on the images to measure the local strains in the Au film using a block matching technique built in the Vision Builder. The load data collected by the Entran load cell were recorded along with the images were acquired by a data/image acquisition system from National Instruments.

In order to study the variation in the electrical behavior of the Au films on the PDMS substrate, the resistance of the Au films was measured using a NI-4060 DMM¹ digital multimeter card from National Instruments in conjunction with Labview program. One side of the PDMS strip with gold film on top was directly in contact with the grip, to which a probe was attached. The other end of the PDMS strip was mounted on the grip using insulation. On top of the gold film, a conductive strip was placed between the insulator spacer and the gold film. This conductive strip was connected to the second probe of the DMM card. Resistance data were collected during the tensile test by the Labview/DMM system with a frequency of approximately 1 Hz.

The load data were analyzed with respect to the corresponding images. Stress-strain curves were constructed using the analyzed load and image data. These were then combined with the resistance data obtained from the sample. Relaxation plots were made by the load-resistance-time data combination.

AFM images were obtained from the samples (with and without the Au film) using a DI-3000 atomic force microscope in Tapping Mode. This was performed on the samples before and after the tensile test. An in-situ tensile test was performed under the AFM to examine the change in the morphology due to testing. This was performed using a manual loading system developed for this purpose.

Nano indentation experiments were performed using a Hysitron tribo-indenter[2] mounted on a DI-3000[3] nanoscope from Digital Instruments. A Berkovich tip was used for nano-hardness measurements. All tests were performed at room temperature.

IV. Results and Discussions

(a) Microstructure

The results of microstructural characterization of the films are presented in (Figure 2)(Figure 3)(Figure 4). The grain size of the films increases with the thickness as shown in (Figure 5). The values of grain sizes are also reported in Table 1. As seen from this table grain size of Au films scale with the thickness of the films starting with 17 nm for the 25 nm-thick films and continues with 40 nm for the 50 nm-thick film and 77 nm for the 75 nm-thick gold film. The variation of the grain size with the thickness of the film is illustrated in (Figure 5). The grain size of the 100 nm-thick film is seen to be nearly double that of the thickness of the film. The 25 nm-thick gold films on the PDMS substrate are depicted in (Figure 6). The film is seen to have formed islands which are about 200 nm in size. These islands are separated by boundaries that are of ~ 10-20 nm thickness. The edges of the islands are seen to be raised while the centers are depressed. The fact that the islands are nearly equiaxed indicates that the stresses developed in the film are comparable in different directions in the plane of the film.

The more interesting microstructure of the 50-nm thick film is presented in (Figure 2). It does not show the microcrack grid of the 25 nm-thick films. As seen from the phase data image of (Figure 2), there are randomly oriented wrinkles in the gold film that gives a brain-morphology appearance to it. The surface morphology of the film shows two types of periodic structures, the first of which is associated with wider microcracks. The periodicity of these microcracks is seen to be ~ 7 mm. The second periodic feature is finer and consist of valleys the film with a spacing of ~ 4 mm. Both the microcracks and the valleys appear in one direction, nearly parallel with the width direction. There are finer microcracks that run in the direction of the length of the strips. These microcracks are much shorter with an average length that scales with the valley spacing (~ 4 mm). The spacing between these secondary microcracks is ~ 3 mm.

The 75 nm-thick film showed characteristics similar to that of 50 nm-thick film. Two types of periodic features are observed in (Figure 3). These corresponded to the microcracks with a spacing of 10 mm and to the valleys with a spacing of mm.

The 100 nm-thick gold film deposited on the PDMS substrates showed morphology similar to that of the 50 nm- and the 75 nm-thick films. The spacing of the larger microcracks was ~ 20 mm which were longer compared the thinner films. (e.g. over 50 mm). The valley structure of the film had a spacing of ~ 4 mm.

All Au films deposited on PDMS substrates were found to be cracked. Microcracks observed in these films were characterized by AFM images obtained from height and phase data. The roughness of the films is reported in Table 1. The RMS roughness depended on the scanned area over which the RMS is calculated. The reported roughness data are reported along with the scanned area size for each film. The roughness values presented were computed for uncracked regions of the films. The RMS roughness is seen to increase with the thickness of the films as seen from Table 1. The smallest RMS was found to be 7 nm obtained for the 25 nm-thick film over an area of 150 nm.. This value increased to 10 nm over larger areas of 500-1000 nm. The RMS roughness was higher for the 50 nm-thick film and its variation with scanned area is seen to be larger than that of the 25 nm-thick film. In general, the roughness of the film increased at higher thicknesses. However, the roughness of the 75 nm- and the 100 nm-thick films was comparable. This trend is continued for the 75 and 100 nm thick films. The roughness of the gold films on PDMS can be compared with those of plain PDMS measured in this study (e.g. 7 nm) and by other researchers (e.g. 6 nm) [10].

(b) Mechanical Properties

The results of mechanical testing on the PDMS films are presented in (Figure 7)(Figure 8). The stress-strain behavior of the PDMS substrates without Au film are is illustrated in (Figure 7). The behavior is seen to be nearly linear but the slope of the stress-strain curves are time dependent.

Effect of Strain Rate: The effect of strain rate on the mechanical behavior of PDMS substrates is shown in (Figure 7)(Figure 8). Three PDMS substrates without Au coating were examined in tensile tests performed at various strain rates. The magnitude of the modulus increases from ~2.1 MPa for a strain rate of 10^-4 to a ~ 3.0 MPa for a strain rate of 10^-3. At an intermediate strain rate of 8 x 10^-4 the modulus was found to be ~2.9 MPa. The log-log plot of (Figure 8) shows a linear behavior with a slope of ~ 0.4 when Young’s modulus is expressed in MPa.

Nano indentation of Au films: Nano indentation tests were performed on plain PDMS to characterize nano-scale mechanical properties including nano-hardness and nano-scale reduced modulus. The results are presented in (Figure 9). The hardness and reduced modulus are seen to greatly depend on indent sizes up to around 200 nm beyond which the curves flatten and give a nearly constant value of ~ 5 MPa for the reduced modulus.

AFM In-situ test: In order to study the morphological changes due to testing, AFM in-situ testing were performed. To conduct this test a manual loading device was used to stretch the PDMS samples to desired strain levels. The surface of the samples was imaged before and right after deformation. The surface could not be imaged during the deformation due to large changes in the height position of the sample upon stretching. The in-situ test allowed characterization of the morphology immediately after testing. The results are shown in (Figure 10) (Figure 11). These figures show the height data-based images of the surfaces of the 100 nm gold film stretched to strains levels of 2.7 and 11%. As seen from these figures, the width of the microcracks has significantly increased. In fact, the bottom of the crack is sealed in the section profile of the crack at the 2.7% strain level. However, the bottom of the crack is seen to have flattened out in the 11% strain level image. The height of the Au film is seen to scale with the depth of the crack ( ~ 100 nm).

(c) Electrical Properties

Relaxation experiments were performed to study the variation of stress and electrical resistance in the Au/PDMS films at constant strains. This was achieved by straining the films for 10 s followed by a 180 s hold time during which the variation of load and resistance was monitored. The relaxation tests were performed on three PDMS samples with Au coatings of 50, 75 and 100 nm. The results are presented in (Figure 12) (Figure 13) (Figure 14). The stress relaxation is clear for these films with the stress dropping significantly during the hold time. This effect is particularly more prominent for the 75 and 100 nm-thick films. The extent of this relaxation is up to ~ 10% for the 100 nm-thick films. The variation of resistance with stress is also seen from (Figure 12) (Figure 13) (Figure 14). The reduction in resistance varies with the degree of strain and the thickness of the films. The highest observed change in resistance is seen in (Figure 13) with a reduction of nearly 50%.

V. Implications

The advent of flexible displays has prompted studies of reliability of metal contacts and metallization lines on flexible substrates. This study sheds light on the electrical and mechanical behavior of Au films deposited on PDMS substrates. Clearly, thinner (~ 25nm thick) Au films do not provide reliable conductive paths for electrons on flexible PDMS substrates with roughnesses examined in this study. Further, conductivity of these gold films is shown to last at strains as high as 10%. Interestingly, the rise in resistively reverses itself when the load is removed. As the stress mitigates, so does the resistance, by what appears to be a self-healing process that may take place in the microstructure of the film. The implications of the results are robust metallization lines on flexible substrates that maintain their electronic properties at acceptable levels under high levels of strains. These are necessary to make the flexible displays that are reliable under service conditions.

VI. Summary and Conclusions

The surface topography and mechanical properties of Au films with the thicknesses of 25-100 nm on PDMS substrates were studied using atomic force microscopy, nanoindentation technique and microtensile testing of the samples. Based on the results of the experiments the following conclusions are made:

The grain size and roughness of the topical Au films deposited on the PDMS substrates increased with the thickness of the film. For thinner films (25-75 nm range) the grains size scaled with the thickness of the film.
The 25 nm thick Au film showed nano-scale cracks with island morphology. The spacing of the cracks was about 172 nm. Thicker films with thicknesses of 50-100 nm showed periodic features including larger microcracks that were spaced 7-20 mm. A second periodic feature parallel to the microcracks where observed to valleys with a spacing of 4 mm.
The Young’s modulus of the Au films increased linearly with the logarithm of the strain rate. The Young’s modulus varied from ~2 MPa to ~ 3 MPa for strain rates of 1 x 10^-4 to 1 x 10^-3. The hardness and reduced modulus obtained by nanoindentation varied with indent size but flattened out at ~ 200 nm indent size. For the 100 nm thick film, the reduced modulus remained nearly unchanged in this region at about 5 MPa.
The results of resistance measurement during microtensile experiments showed an increase in the resistance with increasing strain. This was attributed to the formation of localized necking and microcracks that partially healed during relaxation. The stress in the PDMS dropped with time when the stretching of the sample stopped at intervals of 3 minutes. This phenomenon was observed for the 50-100 nm thick films. The thinner 25 nm films did not show any measurable conductivity.

Table 1: Roughness of Au films on PDMS

Film Thickness (nm)	Grain Size (nm)	Feature Length Scale 1 (mm)	Feature Length Scale-II (mm)	Roughness/ length of area measured
25	17	0.172		7 nm/150 nm
50	40	4	7	15 nm/0.5 mm
75	77	4	10	38 nm /10 mm
100	200	4	20	44nm/50 mm
100 deformed	170	4	20	56 nm/10 mm

REFERENCES

1. Lacour, S.P., S. Wagner, and Z. Suo, Stretchable conductors: films on silicone elastomer. Mat.Res.Soc.Symp.Proc, 2004. 795.

2. Gilleo, K., Polymer thick film. 1996: Van Nostrand Reinhold.

3. Hsu, P.I., et al., Thin-film transistor circuits on large-area spherical surfaces. Applied Physics Letters, 2002. 81(9): p. 1723-5.

4. Lacour, S.P., et al., Stretchable gold conductors on elastomeric substrates. Applied Physics Letters, 2003. 82(15): p. 2404-6.

5. Lacour, S.P., et al. Deformable interconnects for conformal integrated circuits. in Electronics on Unconventional Substrates - Electrotextiles and Giant-Area Flexible Circuits. Symposium. 2003. Warrendale, PA, USA

Boston, MA, USA: Mater. Res. Soc.

6. Lacour, S.P., et al., Design and performance of thin metal film interconnects for skin-like electronic circuits. IEEE Electron Device Letters, 2004. 25(4): p. 179-81.

7. Lacour, S.P., et al. 5-kV switches on flexible plastic substrate for electro-active polymer actuators. in 61st Device Research Conference. 2003. Piscataway, NJ, USA

Salt Lake City, UT, USA: IEEE.

8. Caims, D.R., et al., Strain-dependent electrical resistance of tin-doped indium oxide on polymer substrates. Applied Physics Letters, 2000. 76(11): p. 1425-7.

9. Jones, J., et al., Stretchable wavy metal interconnects. Journal of Vacuum Science & Technology A (Vacuum, Surfaces, and Films), 2004. 22(4): p. 1723-5.

10. Chambers, C., Super-Elastic Gold Conductors on Elastomeric Substrates. 2003, Princeton University: Princeton. p. 1-50.

List of Figures