silicon oxide coatings on plastic substrates

Surface & Coatings Technology 280 (2015) 92–99

Contents lists available at ScienceDirect

Surface & Coatings Technology journal homepage: www.elsevier.com/locate/surfcoat

Barrier property and mechanical flexibility of stress controlled organosilicon/silicon oxide coatings on plastic substrates Shao-Kai Lu a, Shun-Chi Chen a, Tai-Hong Chen b, Li-Wen Lai b, Ren-Mao Liao a, Day-Shan Liu a,⁎ a b

Institute of Electro-Optical and Materials Science, National Formosa University, Huwei, Yunlin 63201, Taiwan, ROC ITRI South, Industrial Technology Research Institute, Liujia Shiang, Tainan 734, Taiwan, ROC

a r t i c l e

i n f o

Article history: Received 12 April 2015 Revised 7 August 2015 Accepted in revised form 31 August 2015 Available online 5 September 2015 Keywords: Organosilicon/silicon oxide Water vapor permeation Mechanical flexibility Residual internal stress Tensile strain

a b s t r a c t In this study, we demonstrated the effect of the organosilicon layer in the pairs of the organosilicon/silicon oxide (SiOx) multi-layered structure on the barrier property to water vapor permeation and on the mechanical flexibility under tensile strain. The residual internal stress in the structure was controlled by introducing the organosilicon layer, thereby enhancing the structural barrier property. The experimental results showed that the thickness of the organosilicon layer required in each paired structure to minimize the structural stress was closely correlated to the surface and material properties of the underlying layer. Accordingly, the internal stress that existed in the 2- and 3-paired structures, as well as their barrier properties to the water vapor permeation, was optimized further by altering the thickness of the organosilicon layer that was deposited onto the surface of the SiOx film. These pairs of multi-layered structures, in which the compressive stress had been minimized, cracked at a lower bending radius under static tensile strain than the SiOx film with the same thickness deposited directly onto the polyethylene terephthalate (PET) substrate. In addition, the degradation in the water vapor transmission rate (WVTR) of the organosilicon/SiOx multi-layered structure deposited onto the PET substrate was less apparent than that of the PET coated only with a single SiOx film. This was ascribed to the structure having fewer cracks than the single SiOx film after being bent by the same tensile strain. Dynamic tensile strain also was conducted on these multi-layered structures to further confirm the structural reliability. © 2015 Elsevier B.V. All rights reserved.

1. Introduction In recent years, significant research has been invested in the development of high-end, high-performance flexible electronics with foldable characteristics, such as solar cells, thin film transistors, light emitting diodes, and displays [1–5]. Generally, there are three choices for flexible substrates, i.e., polymer, thin glass, and metal foil substrates. Among these, polymer substrates are preferred, especially for applications with optoelectronic devices, because of their optical transparency, flexibility, lightness, ease of handling, and cost-effectiveness. However, many obstacles and challenges prevent flexible devices based on polymer substrates from being fully commercialized, because polymer substrates are poor in terms of surface hardness, sensitivity to solvents and other chemicals, and susceptibility to water vapor and oxygen permeation. Thus, transparent oxides or nitride barrier films, such as SiOx, Si3N4, AlOx, TiOx, and IZTO have been used extensively to encapsulate these substrates/devices [6–10]. Although depositing an inorganic film onto a polymer substrate leads to a significant reduction in the water vapor transmission rate (WVTR), this reduction falls short of the orders of magnitude required

⁎ Corresponding author. E-mail address: [email protected] (D.-S. Liu).

http://dx.doi.org/10.1016/j.surfcoat.2015.08.063 0257-8972/© 2015 Elsevier B.V. All rights reserved.

for packaging flexible substrate/device. Therefore, a thick and quality inorganic film must be deposited onto the flexible substrate/device to provide an ultra-high barrier property. However, due to the significant discrepancies in the elasticity and other properties between the inorganic film and the polymer substrate, as well as the low deposition temperature, the resulting WVTR of the polymer substrate coated with a thick inorganic film is limited to the range of about of 0.1–0.01 g/m2/day at 25 °C with 100% relative humidity (RH) originating from the formation of the nanometer- to micrometer-sized defects and cracks in the barrier film [11–13]. Consequently, hybrid structures comprised of inorganic/ inorganic or organic/inorganic multi-layered structures were designed to eliminate these defects in the inorganic layer and function to avoid the penetration of water vapor and oxygen molecules through the polymer substrate [14–17]. In our previous report, we described our development of an organosilicon layer to address adhesion to the polymer substrate [18]. Compressive stress in the SiOx film deposited onto the substrate was demonstrated to be released effectively by introducing an adequate thickness of the organosilicon layer. The barrier property of the 1-paired organosilicon/SiO x multi-layered structure with a minimized internal stress apparently was more effective than a single SiOx film [19]. Accordingly, the PET substrate provided an ultra-low WVTR (approximately 10−5 g/m2/day) after it was coated with a 6-paired organosilicon/SiOx multi-layered structure.

S.-K. Lu et al. / Surface & Coatings Technology 280 (2015) 92–99

In addition to the urgent need to develop multi-layered structures with an ultra-high barrier property, these structures coated onto the polymer substrates also should be sufficiently robust to avoid cracking during bending. However, the inorganic material in the multi-layered structure usually is stiff and brittle, but the polymer substrate is flexible. The barrier property of the inorganic material is degraded due to the formation of cracks after bending. Accordingly, an investigation into the mechanical stress/strain distribution within the entire barrier structure during bending to a radius of curvature of a few centimeters also is critical to prevent the loss of its functionality. Although some papers have reported and discussed the flexibility of the single barrier layer, relatively few reports have examined the mechanical durability of the organic/ inorganic multi-layered structures [20–23]. As a result, the aim of this study was to investigate the mechanical flexibility of the organosilicon/ SiOx multi-layered structure. To do so, we engineered the residual internal stress in the entire barrier structure by altering each thickness of the organosilicon layer in the multi-layered structures in detail for the first time. The critical radii associated with the initiation of the cracks were determined in these multi-layered structures as they were bent by tensile strain, and these radii were compared to that of the single SiOx films directly deposited onto the PET substrates. The flexible stability of these barrier structures was studied through the evolution of the WVTR for the PET substrates that had been coated with these barrier structures and then bent by different tensile strains. In addition, dynamic bending tests were conducted on these samples to confirm their reliability. 2. Experimental The organosilicon/silicon oxide (SiOx) multi-layered structures were deposited consecutively onto the 200 μm-thick PET and silicon substrates by a plasma-enhanced chemical vapor deposition (PECVD) system, using the tetramethylsilane (TMS) monomer and TMS–oxygen (TMS-O2) gas mixture. The deposition pressure, rf power, and temperature were fixed at 13 Pa, 70 W, and 120 °C, respectively. The gas flow rate of the TMS monomer for synthesizing the organosilicon and SiOx layers was fixed at 60 sccm and the gas flow rate ratio (TMS/O2) of the TMS-O2 gas mixture was controlled at 0.5. The thickness of the SiOx film in the multi-layered structure was controlled at 300 nm to study the barrier property and flexibility affected by the thickness of each organosilicon layer in the pairs of the multi-layered structures. Fig. 1 shows a schematic cross-section of the 3-paired organosilicon/ SiOx multi-layered structure. The thickness of the organosilicon layer in each of the pairs of multi-layered structure ranged from 15 to 90 nm. Film thickness was measured using a surface profile instrument (Dektak 6M). The residual internal stress, σf, in the multi-layered

Fig. 1. Schematic cross-section of the 3-paired organosilicon/SiOx multi-layered structure.

93

structure as a function of the thickness of the organosilicon layer deposited onto the silicon substrate was derived from the Stoney formula expressed as [20]: 2

σf ¼

E s ds 6ð1−νs Þd f



1 1 − R f R0

 ð1Þ

where Es, νs, and ds are Young's modulus, Poisson ratio, and thickness of the substrate, respectively. df is the total thickness of the multi-layered structure. R0 and Rf are the radii of the sample curvature before and after depositing the multi-layered structure measured by the beambending method using a thin film stress measurement instrument (FLX-2320, Tencor). Because the surface of the substrate must be reflective for the incident beam, we used the silicon substrate as a substitute for transparent PET substrate to evaluate the internal stress in the barrier structure. Even though the values of the residual internal stress for these structures deposited onto the silicon substrate undoubtedly were different from the values derived from the structures deposited onto the PET substrate, they exhibited the same evolution of the curvature as confirmed by the surface profile observation. This implied that the discussion of the evolutions of residual internal stress derived from the barrier structures deposited onto the silicon substrate was reasonable. The adhesion behavior between the barrier coating and the PET substrate was evaluated by the American Society for Testing and Materials (ASTM) D3359 Scotch® standard tape-peeling test [21]. The fracture behavior associated with the onset and development of the cracks in these barrier structures under static tensile strain, which was examined using a simple-support bending facility of collapsing radius test, was monitored in-situ using an optical microscope. Also, dynamic bending tests were conducted using the universal testing machine (JIA-802PC) controlled at an external force of 0.5 kg and a bending speed of 100 mm/min at a curvature of 20 mm. The water vapor permeation of the coated samples was measured using a WVTR measurement system (MOCON Inc., PERMATRAN-W 3/61) at a temperature of 40 °C with 95% RH. For the barrier structures with permeation below the MOCON test limitation (i.e., WVTR b 10−3 g/m2/day), a calcium (Ca) degradation test under the same environment was conducted to evaluate the WVTR by observing the percentage of the area in which the color was changed within the dimensions of 15 × 15 mm using an optical microscope. 3. Results and discussion The evolution of the residual internal stress and the WVTR of the 1-paired organosilicon/SiOx multi-layered structure (i.e., 300 nm-thick SiOx barrier film deposited onto various thicknesses of organosilicon layers) is shown in Fig. 2(a) (the values of the single SiOx film deposited onto the substrate are also given for comparison). The internal compressive stress in the SiOx film (approximately 498 MPa) could be balanced effectively by insetting the organosilicon layer (hereafter referred to as the first organosilicon layer). The residual internal stress of the organosilicon/SiOx multi-layered structure was optimized at −90 MPa for the SiOx barrier film deposited onto a 30 nm-thick organosilicon layer. This 1-paired multi-layered structure coated onto the PET substrate exhibited the lowest WVTR value of 0.28 g/m2/day whereas the 300 nm-thick SiOx film directly deposited onto the PET substrate produced a WVTR of 0.38 g/m2/day. However, the internal compressive stress of the 1-paired multi-layered structure was increased when the thickness of the organosilicon layer was greater than 30 nm, which also resulted in the degradation of the barrier performance. The mechanism responsible for the enhancement of the barrier performance of the SiOx film by insetting the organosilicon layer has been ascribed to the increased density of the packing and the improvement of the structural quality of the SiOx film that was deposited onto the organosilicon layer with its abundance of C–H groups [19]. In addition, since single organosilicon layers with thicknesses of 15, 30, and 60 nm that were

94

S.-K. Lu et al. / Surface & Coatings Technology 280 (2015) 92–99 Table 1 Barrier performance of the single SiOx film and 1-, 2-, and 3-paired organosilicon/SiOx multi-layered barrier structures deposited on the PET substrate, as well as their residual internal stresses when deposited onto the silicon substrate (the transmission rate of the bare PET is also shown for comparison). Structure

Film thickness, dc (nm)

Transmission rate, ΠT (g/m2/day)

Effective permeability, Pc (μm-g/m2/day)

Compressive stress (MP)

Bare PET Single SiOx 1-paired 2-paireda 2-pairedb 3-paireda 3-pairedb

– 300 330 660 690 990 1050

3.6 0.38 0.28 0.12 0.08 0.03 b0.01

– 0.127 0.100 0.082 0.056 0.029 b0.01

– 498 90 265 217 346 205

Measurements taken at 40 °C with 95% RH. a The thicknesses of the 2nd and 3rd organosilicon layers both are 30 nm. b The thicknesses of the 2nd and 3rd organosilicon layers both are 60 nm.

layer, pairs of the multi-layered structures using the same thickness of the organosilicon and SiOx layers were consecutively deposited onto the PET substrate. Table 1 summarizes the WVTR of the pairs of the organosilicon/ SiOx multi-layered structure with the same thickness of the organosilicon (30 nm) and SiOx (300 nm) films coated onto the PET substrates. Also, as shown in Table 1, with the effect of the PET substrate eliminated, the effective permeability, Pc, for the entire barrier structure was calculated according to the ideal laminate theory (ILT) [22]: 1 dT ds dc ¼ ¼ þ ΠT PT Ps Pc

Fig. 2. The evolution of the residual internal stress and the WVTR for the (a) 1-paired organosilicon/SiOx multi-layered structures as a function of the thickness of the first organosilicon layer (the values of the 300 nm-thick SiOx film deposited directly onto the substrate also is given for comparison), (b) 2-paired organosilicon/SiOx multi-layered structure as a function of the thickness of the second organosilicon layer, and (c) 3-paired organosilicon/SiOx multi-layered structure as a function of the thickness of the third organosilicon layer deposited on the substrate (the inset figure shown the image of the Ca test result for the 3-paired multi-layered structure with the first, second, and third organosilicon layer thicknesses of 30 nm, 60 nm, and 60 nm, respectively).

deposited onto the silicon substrate induced internal tensile stress of 134, 888, and 122 MPa, respectively, the internal compressive stress in the SiOx barrier layer could be balanced, thereby minimizing the formation of nanometer-sized cracks in the SiOx film. The arrows in Fig. 1 also show the residual compressive and tensile stress in the SiOx and organosilicon layers of the barrier structure. Since the residual internal stress of the SiOx film that was deposited onto the PET substrate was optimized by introducing an adequate thickness of the organosilicon

ð2Þ

where Π is the transmission rate, P is permeability, and d is thickness; the subscripts s, c and T denote the substrate, barrier coating, and combination structure (s + c), respectively. The WVTR of the PET substrate coated with the 2-paired organosilicon/SiOx multi-layered structure, which corresponded to a 600 nm-thick SiOx layer, further decreased to 0.12 g/m2/day, whereas the WVTR of the PET substrate coated with a single 600 nm-thick SiOx film was about 0.51 g/m2/day. This revealed that the thick SiOx film which was stacked in a multi-layered structure served as a quality barrier to water vapor permeation. In addition, the WVTR was reduced significantly (to approximately 0.03 g/m2/day) for the 3-paired organosilicon/SiOx structure that was deposited consecutively onto the PET substrate, and this structure also corresponded to a very low effective permeability of 0.029 μm-g/m2/day. In general, the mechanism responsible for the improvement on the barrier performance was ascribed to the decrease in the diffusion pathways for water vapor permeation [23–25]. Since the defects and pinholes in the SiOx barrier film were decoupled and/or covered by the neighboring organosilicon layer, the above-mentioned multi-layered barrier structure therefore could obstruct the water vapor permeation more effectively than the single SiO x barrier film. Fig. 3(a)–(c) shows the surface morphologies of the single SiOx film and the 1-paired and 2-paired organosilicon/SiOx multi-layered structures with the same thickness of the organosilicon layer (30 nm) (high magnification photographs also are provided in the inset figures). Fig. 3(a) shows that many white protrusions were distributed over the surface of the SiO x film. These spherical protrusions, with diameters of about 100 nm, as observed in the inset figure, were linked to the coverage of the voids that originated from the large compressive stress excited in the film [26–27]. Compared to the SiOx film that was deposited directly onto the substrate, the density of the protrusions (Fig. 3(b)) apparently was reduced due to the effective balancing of the compressive stress that existed in this structure. In addition, since the inset organosilicon layer was beneficial for decoupling the defect, these protrusions were essentially absent on the surface of the 2-paired organosilicon/SiOx multi-layered structures, as shown in Fig. 3(c).

S.-K. Lu et al. / Surface & Coatings Technology 280 (2015) 92–99

95

Fig. 4. Tape-peeling test results for the (a) single SiOx and the (b) 1-paired, (c) 2-paired, and (d) 3-paired organosilicon/SiOx multi-layered structures with the thicknesses of the organosilicon and SiOx layers controlling at 30 and 300 nm, respectively, deposited onto the PET substrates.

internal stress in this structure, σtotal , presumably was optimized according to the following equation [28]: σ total ¼

Fig. 3. Surface morphologies of the (a) single SiOx and the (b) 1-paired and (c) 2-paired organosilicon/SiOx multi-layered structures with the thickness of each organosilicon layer of 30 nm, respectively (high magnification photographs also are given in the inset figures).

Fig. 4(b)–(d) shows the adhesion behavior of the 1-, 2-, and 3-paired organosilicon/SiOx multi-layered structures with the thickness of the organosilicon and SiOx layers controlled at 30 and 300 nm, respectively, deposited onto the PET substrates using tape-peeling tests (the adhesion behavior of the single 300 nm-thick SiOx film directly deposited onto the PET substrate also is presented in Fig. 4(a) for comparison). Based on ASTM D3359, the poor adhesion of the 300 nm-thick SiOx film deposited onto the PET substrate (Fig. 4(a)) apparently was improved to rank 5B (no peel-off, i.e., none of the squares was detached) by insetting a 30 nm-thick organosilicon layer, as shown in Fig. 4(b). The significant enhancement of the organosilicon/SiOx structure's adherence to the PET substrate was ascribed to the reduction of the compressive stress existing in the coating system [18]. However, extensive peel-off areas are apparent in Fig. 4(c) and (d), indicating that the internal stress that existed in the 2- and 3-paired organosilicon/SiOx multi-layered structures must be addressed and reduced by conducting additional research. Fig. 2(b) shows the residual internal stress and the WVTR of the 2-paired organosilicon/ SiOx multi-layered structures deposited on the substrate as a function of the thickness of the second organosilicon layer. The figure shows that the 2-paired multi-layered structure with the same thickness of the organosilicon and SiOx layers, i.e., 30 and 300 nm, respectively, did not exhibit the lowest residual internal stress, although the total residual

σ 1 d1 þ σ 2 d2 þ σ 3 d3 þ :::σ n dn d1 þ d2 þ d3 þ :::dn

ð3Þ

where σ is the internal stress that existed in the film when it is individually deposited onto the substrate and d is the thickness of each layer in the multi-layered structure. The measured internal compressive stress (~265 MPa) that existed in the 2-paired multi-layered structure with the same thickness of the organosilicon and SiOx layers apparently were higher than that of the 1-paired multi-layered structure (~ 90 MPa), indicating that the residual internal stress was induced extensively when the second organosilicon and SiOx layers were deposited. The lowest internal compressive stress that existed in the 2-paired multi-layered structure shown in Fig. 2(b) was obtained when the thickness of the second organosilicon layer was 60 nm (~217 MPa), indicating that the thickness of the second organosilicon layer required to minimize the structural stress was different from that of the first organosilicon layer. The reason for the different thickness of the organosilicon layer to optimize the residual internal stress of the multi-layered structure was attributed to these two organosilicon layers that were respectively deposited onto the underlying silicon substrate and the SiOx film would induce different internal stresses due to the discrepancies in the surface and material properties. When the thickness of the second organosilicon layer was 90 nm, the internal compressive stress in the 2-paired multi-layered structure apparently increased to 482 MPa. As quoted from previous reports and the above-mentioned study, the barrier structure with the optimal internal stress is beneficial for lowering the formation of nanometer-sized defects and cracks in the SiOx film [13,19]. Accordingly, the WVTR of these 2-paired organosilicon/SiOx multi-layered structures deposited onto the PET substrate shown in Fig. 2(b) was correlated strongly with the evolution of their residual internal stress. The lower the compressive stress that existed in the 2-paired structure resulted in fewer nanometer-sized defects and cracks, so the performance of the barrier was better in reducing the permeation of water vapor permeation. The lowest WVTR obtained from the PET substrate coated with the 2-paired organosilicon/SiOx multi-layered structure was about

96

S.-K. Lu et al. / Surface & Coatings Technology 280 (2015) 92–99

0.08 g/m2/day, and it also exhibited the lowest residual internal stress, as shown in Fig. 2(b). The residual internal stress and the WVTR of the 3-paired organosilicon/SiOx multi-layered structures deposited onto the substrate as a function of the thickness of the third organosilicon layer are shown in Fig. 2(c) (the thicknesses of the first and second organosilicon layers and the thickness of each of the SiOx films are controlled at 30, 60, and 300 nm). Similar to the evolution of the residual internal stress of the 2-paired organosilicon/SiOx multi-layered structures, the compressive stress that existed in the 3-paired structure also was optimized as the thickness of the third organosilicon layer reached 60 nm (~ 205 MPa), which corresponded to an ultra-low WVTR value that was below the detection limit of the MOCON Permatran-W3/61 (0.01 g/m2/day). The inset figure shows the image of the Ca test for this 3-paired multi-layered structure coated onto the PET substrate after aging in the 40 °C/95% RH environment for 14 days. The corresponding WVTR was about 4.4 × 10− 4 g/m2/day based on the evaluation of the percentage of the hydrolyzation areas in the inset figure. The barrier performance and residual internal stress of these multi-layered structures with the second and third organosilicon layers of 60 nm also are summarized in Table 1. The effective permeability of the 2- and 3-paired multi-layered structures could be improved significantly by altering the thicknesses of the second and third organosilicon layers, which also had the lowest residual internal stress. Furthermore, such 2- and 3-paired organosilicon/SiOx multi-layered structures with low residual internal stress also showed an excellent adhesion grade of 5B as ranked by the standard tape-peeling test. Fig. 5(a) shows the crack density as a function of the bending radius under tensile strain for the single SiOx films with thicknesses of 100, 300, 600, and 900 nm. The initiation of the cracks in the thinner SiOx

film occurred at a lower bending radius than that of the thicker SiOx film, revealing the better mechanical flexibility of the thinner SiOx film. After the initiation of the cracks, the crack density in these SiOx films increased significantly as the tensile strain increased, and the crack density gradually became saturated due to the delamination of the films, which would prohibit the applied tensile strain from being transmitted from the PET substrate to the SiOx film. In addition, as the thicker SiOx film cracked at a higher bending radius (i.e., lower applied tensile strain), the corresponding saturated crack density, which had a strong correlation with the delamination of the film, was lower than that in the thinner SiOx film. The critical radius of curvature, R0, that represented the initiation of the cracks under bending was obtained by fitting the evolution of the crack density, using the following exponential form [29]: y ¼ y0 f1− exp½−bðR−R0 Þg

ð4Þ

In this equation, y is the crack density, y0 is the saturated crack density, R is the bending radius of curvature, and b is an adjustable parameter. The corresponding critical radius and saturation crack density for these SiOx films are summarized in Table 2. The critical radius apparently was increased from 8.7 mm for the 100 nm-thick SiOx to 17.7 mm for the 900 nm-thick SiOx film, implying that a thinner SiOx film was more flexible and able to withstand higher applied tensile strain without cracking. Fig. 5(b) shows the critical tensile strain for the initiation of crack as a function of the single SiOx film thickness. The critical tensile strain, εC, was calculated from the critical radius of curvature, R0, as listed in Table 2, adopted the following equation [30]: εC ¼

ds þ d f 2R0


ð5Þ

where ds and df are the thickness of the PET substrate and SiOx film, respectively. Consistent with other reports, the critical tensile strain initially was inversely proportional to the thickness of the film, but it gradually became saturated as the thickness of the film increased [31–33]. The relationship between the critical tensile strain and film thickness, df, was obtained by fitting these data points according to the form [34]: −1

εC  f ðmaterial propertiesÞ  d f

=n

ð6Þ

The fitting exponent n was determined to be 2.85. Fig. 6(a)–(d) shows the images of the cracks in the single 100, 300, 600, and 900 nm-thick SiOx films that were deposited directly onto the PET substrate while they were bent by a tensile strain to the radius of 5 mm observed in situ using an optical microscope (the tensile strain was applied in the horizontal direction in these figures). All of the cracks that appeared in the SiOx film were straight, and they were perpendicular to the direction of the tensile strain load. Under the same tensile strain, the density of these channel cracks in the 100 nm-thick SiOx

Table 2 Critical radius and saturation crack density for the single SiOx films and the multi-layered structures. Structure Single SiOx

Fig. 5. (a) Crack density as a function of the bending radius under tensile strain for different thicknesses of the single SiOx films and (b) critical tensile strain as a function of the thickness of the SiOx film.

Multi-layered structure

100 nm 300 nm 600 nm 900 nm 1-paired 2-paired 3-paired

Critical radius (mm)

Saturated crack density (mm−1)

8.7 14.3 16.6 17.7 13.2 15.2 17.3

203 56 42 28 44 28 23

S.-K. Lu et al. / Surface & Coatings Technology 280 (2015) 92–99

97

be expressed as a function of the bending radius, R, as addressed in the following equation [36]: 3

σ induce ¼

Fig. 6. Images of the cracks in the single SiOx film with thickness of (a) 100 nm, (b) 300 nm, (c) 600 nm, and (d) 900 nm directly deposited on the PET substrate bent by a tensile strain with the radius of 5 mm.

film (Fig. 6(a)) was approximately 188 mm−1, and their density in the 900 nm-thick SiOx film was only about 38 mm−1, as evaluated from Fig. 6(d). In general, the film delamination originating from the interfacial shear stress between the film and the substrate was more significant in the thick SiOx film [35]. The features of the cracks and the width of the cracking line width appeared in the thick SiOx film was more apparent than that in the thin SiOx film under a high applied tensile strain. Fig. 7 shows the crack density as a function of the bending radius under tensile strain for the 1-, 2-, and 3-paired organosilicon/SiOx multi-layered structures (the crack density of the 300 nm-thick SiOx film bent by tensile strain at various radii also are provided for comparison). It was evident that the critical radius (~13.2 mm) for the initiation of the crack in the 300 nm-thick SiOx film deposited onto the 30 nm organosilicon layer, which had a low residual internal compressive stress, was lower than the same thickness of the SiOx film directly deposited onto the PET substrate (~14.3 mm). When the tensile strain was imposed on the samples, an additive compressive stress was accumulated in the barrier structures. The induced stress, σinduce, can

Fig. 7. Crack density as a function of the bending radius under tensile strain for the 1-, 2-, and 3-paired organosilicon/SiOx multi-layered structures, respectively (the crack density of the 300 nm-thick SiOx film bent by tensile strain at various radii also is provided for comparison).

Es ds
6Rd f ds þ d f

ð7Þ

where Es and ds are Young's modulus and thickness of the PET substrate, respectively. df is the total thickness of the multi-layered structure. It is well known that cracks are formed as a result of the relaxation of stress when the barrier structure experienced an excessive compressive stress after being bent by the tensile strain. Accordingly, since the SiOx barrier film deposited onto a 30 nm-thick organosilicon layer had a lower internal compressive stress than the SiOx film deposited directly onto the substrate, this multi-layered structure could withstand a high tensile strain without the formation of cracks. Furthermore, the corresponding saturated crack density (~ 44 mm−1) was reduced by inserting the organosilicon layer, implying that the organosilicon layer could also buffer the formation of cracks in the SiOx film under a high tensile strain. The derived critical radius and saturation crack density for these barrier structures deposited onto the PET substrate are listed in Table 2. The evolution of the mechanical flexibility for these structures was similar to that of the SiO x film deposited directly onto the PET substrate. The thicker the SiO x film in the paired structure was, the earlier cracks appeared and the less saturated the crack density was. In addition, because the inset organosilicon layer in the multi-layered structure helped balance the structural stress and buffer the propagation of the tensile strain load from the substrate to the barrier film, both the critical radius determined from the initiation of the cracks and the saturated crack density were improved as compared to those of the same SiO x film deposited directly onto the PET substrate. Fig. 8(a)–(d) shows the images of the cracks in the 300 nm-thick SiOx film and 1-, 2-, and 3-paired organosilicon/SiOx multi-layered structures, respectively, deposited onto the PET substrate bent by a tensile strain to the radius of 5 mm observed in situ using an optical microscope. The features of the cracks in the 1-paired organosilicon/SiO x multi-layered structure were almost identical to those in the single SiOx film (Fig. 8(a)), except for the reduction in the crack density from 78 mm − 1 to 66 mm − 1 due to the contribution of the organosilicon layer to balancing the structural stress. The crack density in these multi-layered structures decreased

Fig. 8. Images of the cracks in the (a) single SiOx film (300 nm) and the (b) 1-paired, (c) 2-paired, and (d) 3-paired organosilicon/SiOx multi-layered structures deposited onto the PET substrate bent by a tensile strain with the radius of 5 mm.

98

S.-K. Lu et al. / Surface & Coatings Technology 280 (2015) 92–99

crack density (~ 45 mm− 1) than that observed from the single SiOx film (~56 mm−1), as shown in Fig. 7, was alleviated because this structure had less diffusion pathways for water vapor permeation. It is also worth noting that only a limited increase occurred in the WVTR of the PET substrate coated with the two-paired organosilicon/SiOx multilayered structure after bending at a radius of 15 mm (by a factor of about 1.03) because this radius was only slightly lower than the initiation of the cracks (~ 15. 2 mm) as derived in Fig. 7. By contrast, the WVTR of the PET substrate coated with the 3-paired structure, which showed the initiation of the cracks at a bending radius of 17.3 mm, obviously was increased after the same tensile strain was applied (approximately by a factor of 1.36). When the bending radius reached 10 mm, the barrier properties of all these multi-layered structures were degraded as a result of the growth of the cracks, especially for the PET substrate coated with the 3-paired organosilicon/SiOx structure. However, although the sample coated with the 3-paired multi-layered structure had the worst mechanical flexibility, the degree of the resulting WVTR still was about two orders of magnitude lower than that of the PET substrate coated with the 2-paired multi-layered structure after being bent by a tensile radius of 10 mm. The changes in the WVTR as well as the standard deviations acquired from five samples for these multilayered structures coated onto the PET substrates after being bent by a tensile radius of 20 mm for 1, 10, 100, and 1000 times are illustrated in Fig. 9(b). Even though the samples had a lower WVTR than the initial value, possibly because of the slight enhancement of the structural densification after cycle bending, all of the PET substrates coated with the multi-layered structures exhibited superior reliability with an almost constant WVTR value (measurement repeatability within 3%) when the bending radius was larger than the critical radius as derived from Fig. 6.

4. Conclusions

Fig. 9. Variations of the WVTR for the single SiOx film and the organosilicon/SiOx multilayered structures coated onto the PET substrate after (a) static tensile bending at various radii and (b) dynamic tensile bending at the radius of 20 mm for 1, 10, 100, and 1000 times.

as the thickness of the SiOx film increased, and the cracks' features and the width of the cracking line also became apparent. Fig. 9(a) shows the variations of the WVTR for the single SiOx film and the organosilicon/SiOx multi-layered structures coated onto the PET substrate bent by various radii under tensile strain. The variations of the WVTR (ΔWVTR) shown in this figure were obtained using the expression: ΔWVTR ¼

W−W 0  100% W0

ð8Þ

where W0 and W are the WVTR of the sample before and after the tensile bending. The WVTR of the PET substrate coated with a 300 nm-thick SiOx film remained almost unchanged after bending by the tensile strain with a radius of 15 mm, but it was about 1.8 times higher than the initial value as the bending radius reached 10 mm. For the 1-paired multilayered structure with the same thickness as the SiOx film coated onto the PET substrate, the WVTR also was unchanged at a bending radius of 15 mm, but it increased slightly (by a factor of 1.15) when the bending radius reached 10 mm as compared to the non-bending sample. The degradation in the WVTR of the single and 1-paired barrier structures after the tensile strain bending was attributed to the formation of the cracks, which provided the diffusion pathways for the water vapor permeation. Although the WVTR of these samples were both degraded after bending at a radius of 10 mm as a consequence of the appearance of the cracks, the degradation in the WVTR of the PET substrate coated with the 1-paired multi-layered structure, which exhibited a lower

The compressive stress that existed in the SiOx film prepared by a PECVD system using the TMS-O2 gas mixture was balanced effectively by insetting an organosilicon layer using only the TMS monomer. The internal compressive stress of the 300 nm-thick SiOx film deposited onto the silicon substrate was decreased markedly, from 498 MPa to 90 MPa, after insetting a 30 nm-thick organosilicon layer. This organosilicon/SiO x multi-layered structure coated onto the PET substrate also had a lower WVTR (0.28 g/m2/day) than the SiOx film that was deposited directly onto the PET substrate (0.38 g/m2/day). However, due to the differences in the surface and material properties between the substrate and the SiOx film, the optimal internal stress that existed in the 2- and 3-paired multi-layered structures occurred as the thickness of both the second and third organosilicon layers reached 60 nm. The corresponding WVTR of the PET substrates coated with the 2- and 3-paired multi-layered structures were decreased further to 8.0 × 10−2 and 4.4 × 10−4 g/m2/day, respectively. In addition, since the organosilicon layer that was introduced into the barrier structure was beneficial for buffering the structural stress, these multi-layered structures could withstand a higher tensile strain without cracking than the same thickness of SiOx film deposited directly onto the PET substrate. Accordingly, the critical radii associated with the initiation of cracks under tensile strain, which were derived from the crack density on the 2- and 3-paired multi-layered structures, were decreased to 15.2 and 17.3 mm, respectively. In addition, the degradation in the barrier property of the 300 nm-thick SiOx film that was deposited directly onto the PET substrate was more obvious than it was deposited onto an organosilicon layer after being bent by the same tensile strain. The result was attributed to the fewer cracks for the water vapor permeation as obtained from the organosilicon/SiOx multi-layered structure. Furthermore, the barrier properties of all these multi-layered structures had excellent reliability after being bent at a radius of 20 mm for 1000 times.

S.-K. Lu et al. / Surface & Coatings Technology 280 (2015) 92–99

Acknowledgements This work was supported by the National Science Council and Industrial Technology Research Institute (ITRI South) under no. FY104-AE2 and the Ministry of Science and Technology under 1032221-E-150-067. References [1] N. Bano, S. Zaman, A. Zainekabdin, S. Hussain, I. Hussain, O. Nur, M. Willander, J. Appl. Phys. 108 (2010) 043103. [2] P. Cosseddu, S. Lai, M. Barbaro, A. Bonfiglio, Appl. Phys. Lett. 100 (2012) 093305. [3] F.M. Keely, L. Meunier, C. Cochrane, V. Koncar, Displays 34 (2013) 1. [4] S.C. Li, L.H. Qiu, C.Z. Shi, X.J. Chen, F. Yan, Adv. Mater. 26 (2014) 1266. [5] C.B. Park, K.M. Kim, J.E. Lee, H. Na, S.S. Yoo, M.S. Yang, Org. Electron. 15 (2014) 3538. [6] M.A. El Khakani, M. Chaker, A. Jean, S. Boily, H. Pépin, J.C. Kieffer, S.C. Gujrathi, J. Appl. Phys. 74 (1993) 2834. [7] G.H. Lee, J. Yun, S. Lee, Y. Jeong, J.H. Jung, S.H. Cho, Thin Solid Films 518 (2010) 3075. [8] Z. Liu, X.X. Ma, Z.H. Sun, F.J. Xu, Rare Metal Mater. Eng. 41 (2012) 268. [9] M.R. Saleem, R. Ali, S. Honkanen, J. Turunen, Thin Solid Films 542 (2013) 257. [10] S.K. Lu, J.T. Huang, T.H. Lee, J.J. Wang, D.S. Liu, Smart Sci. 2 (2014) 7. [11] G. Rossi, M. Nulman, J. Appl. Phys. 74 (1993) 5471. [12] B.M. Henry, F. Dinelli, C.R.M. Grovenor, O.V. Kolosov, G.A.D. Briggs, A.P. Roberts, R.S. Kumar, R.P. Howson, Thin Solid Films 355–356 (1999) 500. [13] M. Hanika, H.C. Langowski, U. Moosheimer, W. Peukert, Chem. Eng. Technol. 26 (2003) 605. [14] L. Murrayt, P. McCarry, J. Plast. Film. Sheet. 19 (2003) 55. [15] M. Gupta, Y.J. Lin, T. Deans, E. Baer, A. Hiltner, D.A. Schiraldi, Macromolecules 43 (2010) 4230.

99

[16] A. Morlier, S. Cros, J.P. Garandet, N. Alberola, Sol. Energy Mater. Sol. Cells 115 (2013) 93. [17] N. Kim, S. Graham, K.J. Hwang, Mater. Res. Bull. 58 (2014) 24. [18] D.S. Liu, C.Y. Wu, J. Phys. D. Appl. Phys. 43 (2010) 175301. [19] C.Y. Wu, R.M. Liao, L.W. Lai, M.S. Jeng, D.S. Liu, Surf. Coat. Technol. 206 (2012) 4685. [20] V. Bhatt, S. Chandra, S. Kumar, Indian J. Pure Appl. Phys. 45 (2007) 377. [21] ASTM, D3359, standard test methods for measuring adhesion by tape test, ASTM International, West Conshohocken, PA, 2008. [22] D.G. Howells, B.M. Henry, J. Madocks, H.E. Assemder, Thin Solid Films 516 (2008) 3081. [23] J. Levis, Mater. Today 9 (2006) 38. [24] D.A. Spee, M.R. Schipper, C.H.M. van der Werf, J.K. Rath, R.E.I. Schropp, Thin Solid Films 532 (2013) 84. [25] S. Park, W.M. Yun, L.H. Kim, S. Park, S.H. Kim, C.E. Park, Org. Electron. 14 (2013) 3385. [26] S.E. Babayan, J.Y. Jeong, A. Schütze, V.J. Tu, M. Moravej, G.S. Selwyn, R.F. Hicks, Plasma Sources Sci. Technol. 10 (2001) 573. [27] C.H. Xu, X.M. Qin, W.H. Guo, Y. Li, Mater. Sci. Technol. 27 (2011) 1271. [28] Y.C. Lin, Q.K. Le, L.W. Lai, R.M. Liao, M.S. Jeng, D.S. Liu, Int. J. Eng. Technol. Innov. 2 (2012) 184. [29] Y. Leterrier, Prog. Mater. Sci. 48 (2003) 1. [30] Z. Suo, E.Y. Ma, H. Gleskova, S. Wagner, Appl. Phys. Lett. 74 (1999) 1177. [31] Y. Leterrier, L. Médico, F. Demarco, J.A.E. Månson, U. Betz, M.F. Escolà, M.K. Olsson, F. Atamny, Thin Solid Films 460 (2004) 156. [32] N.E. Jansson, Y. Leterrier, J.A.E. Månson, Eng. Fract. Mech. 73 (2006) 2614. [33] S.H. Jen, J.A. Bertrand, S.M. George, J. Appl. Phys. 109 (2011) 084305. [34] M. Yanaka, T. Miyamoto, Y. Tsukahara, N. Takeda, Compos. Interface 6 (1999) 409. [35] C. Peng, Z. Jia, D. Bianculli, T. Li, J. Lou, J. Appl. Phys. 109 (2011) 103530. [36] P.C.P. Bouten, P.J. Slikkerveer, Y. Leterrier, in: G.P. Crawford (Ed.), Flexible Flat Panel Displays, John Wiley & Sons Ltd., Chichester, England 2005, p. 125.