- Email: [email protected]
Investigation of brittle failure in transparent conductive oxide and permeation barrier oxide multilayers on flexible polymers
Thin Solid Films 518 (2010) 3075–3080
Contents lists available at ScienceDirect
Thin Solid Films j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / t s f
Investigation of brittle failure in transparent conductive oxide and permeation barrier oxide multilayers on flexible polymers Gun-Hwan Lee, Jungheum Yun ⁎, Sunghun Lee, Yujeong Jeong, Jae-Hye Jung, Sang-Hyun Cho Materials Processing Division, Korean Institute of Materials Science, 531 Changwondaero, Changwon, Gyeongnam 641-831, Republic of Korea
a r t i c l e
i n f o
Available online 13 October 2009 Keywords: Barrier TCO SiOx IZO Bending Water vapor transmission rate (WVTR) Polymer Flexible substrate
a b s t r a c t An oxide multilayer structure—consisting of an indium zinc oxide (IZO) conductive layer, a silicon oxide (SiOx, x = 1.8) water vapor permeation barrier, and an aluminum oxide (Al2O3) interlayer—coated on polyethylene terephthalate (PET) is proposed as a transparent flexible substrate for display and photovoltaic applications. Vital properties of the multilayer, such as the low water vapor impermeability of the SiOx barrier and the high conductance of the IZO film, degraded considerably because of the crack formation in bend geometries, attributed to the large difference between elastic properties of the oxide films and polymers. In order to suppress the crack formation, a 10-nm-thick Al2O3 interlayer was sputtered on Ar ionbeam treated PET surfaces prior to a SiOx plasma-enhanced chemical vapor deposition (PECVD) process. Changes in the conductance and water vapor impermeability were investigated at different bending radii and bending cycles. It was found that the increases in resistance and water vapor transmission rate (WVTR) were significantly suppressed by the ion-beam PET pretreatment and by the sputtered Al2O3 interlayer. The resistance and WVTR of IZO/SiOx/Al2O3/PET systems could be kept low and invariable even in severely bent states by choosing the SiOx thickness properly. The IZO (135 nm)/SiOx (90 nm)/Al2O3 (10 nm)/PET system maintained a resistance of 3.2 × 10− 4 Ω cm and a WVTR of b 5 × 10− 3 g m2 d− 1 after 1000 bending cycles at a bending radius of 35 mm. © 2009 Elsevier B.V. All rights reserved.
1. Introduction The development of flexible substrates that are based on organic polymers is of critical importance for display and photovoltaic applications. Flexible substrates offer the advantages of mechanical flexibility, design freedom, optical transparency, low weight, and cost effectiveness [1,2]. A number of inorganic functional coatings on polymers have critical properties such as water vapor impermeability and electronic conductance, which are required for display and photovoltaic applications [3]. Water vapor impermeability is achieved by coating a thin oxide or nitride barrier on polymers, and electronic conductance is achieved by using a transparent conductive oxide (TCO), often coated on the barrier. The water vapor transmission rate (WVTR) represents the rate of penetration of water molecules into a barrier. Although recent developments in oxide/polymer multilayers [4,5] and atomic-layerdeposited oxide layers [6,7] can provide extremely low WVTR values of approximately 10− 6 g m− 2 d− 1, the commercial application of these barriers has been limited because of complexities in the manufacturing process and low throughput. However, the fabrication of barriers using plasma-enhanced chemical vapor deposition
⁎ Corresponding author. Tel.: +82 55 280 3515; fax: +82 55 280 3570. E-mail address: [email protected] (J. Yun). 0040-6090/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.tsf.2009.08.057
(PECVD) and physical vapor deposition (PVD) techniques, which are well-established in industry, has been thoroughly investigated to develop simple and cost-effective methods that yield WVTR values in the range of 10− 3–10− 4 g m− 2 d− 1 [8–10]. These WVTR values are sufficient for satisfying requirements for various applications such as organic thin film transistors (TFTs), liquid crystal displays (LCDs), and inorganic photovoltaics [1]. With regard to TCO materials that can be applied to flexible substrates, amorphous indium zinc oxide (IZO) has become the preferred option as a replacement for polycrystalline indium tin oxides (ITOs) because it requires low fabrication temperature and is less fragile [11–13]. The mechanical failure of brittle oxide coatings on flexible polymers is a serious issue. The formation of defects, such as cracks and debonding, in water vapor barriers and TCO films may be inevitable in situations in which bend geometries are required because of the large difference between the elastic properties of oxides and polymers [14,15]. The initiation and subsequent developments of defects mainly depend on the coating material, thickness, and interfacial adhesion [2]. Therefore, it is important to have a clear understanding of the failure behavior of coatings in the case of bend geometries, in order to prevent the destruction of devices in which they are used. The failure behavior of thin films coated on flexible polymers has been reported for individual water vapor permeation oxide barriers [16] and TCOs [14,15,17] in bend geometries. However, the failure behavior of complex multilayer structures in bend
3076
G.-H. Lee et al. / Thin Solid Films 518 (2010) 3075–3080
geometries has not been well understood. The effects of film deformation on the performance of multilayer coatings, which include both a water vapor permeation barrier and a TCO, on polymers are largely unknown. In this study, we achieved advanced water vapor impermeabilities in SiOx barriers by developing a two-step pretreatment process comprising (i) PET treatment with Ar ion beams and (ii) subsequent sputtering of a 10-nm-thick Al2O3 interlayer on the PET, prior to a SiOx PECVD process. This process brought about an improvement in both the morphology of the SiOx barrier and the adhesion of the barrier with a polymer substrate, resulting in significant barrier enhancement. Low WVTR values of b5 × 10− 3 g m− 2 d− 1 were obtained for a SiOx/Al2O3/PET system with a SiOx thickness of approximately 100 nm, and further decrease in the WVTR to the range of 10− 4 g m− 2 d− 1 could be readily expected from an increase in the SiOx thickness beyond 150 nm. Changes in the water vapor impermeability of SiOx/Al2O3/PET systems and the electrical resistance of subsequently grown IZO films were measured as functions of the bending radius and the bending cycle in bend geometries. The increase in the WVTR and resistance during bending tests was significantly suppressed in the case of the IZO/SiOx/Al2O3/PET system. 2. Experimental procedure The barrier system was composed of (i) a 10-nm-thick sputtered Al2O3 interlayer and (ii) a thicker PECVD SiOx (x = 1.8) barrier with a thickness of up to 750 nm. Prior to the growth of the Al2O3 interlayer, a 188-μm-thick PET substrate (Kimoto Co., Ltd.) was pretreated with Ar ion beams. The pretreatment conditions were as follows: working pressure of 1 × 10− 4 Torr, dc power of 220 W, Ar flow rate of 15 sccm, and processing time of 3 min. The Al2O3 interlayer was grown by a dc reactive magnetron sputtering system using a 4-in. Al target (Applied Science Corp.). The sputtering conditions were as follows: working pressure of 3 × 10− 3 Torr, dc power of 300 W, and Ar:O2 flow rate of 45:5 sccm. Subsequently, the SiOx barrier was deposited on the Al2O3 interlayer by an rf-PECVD reactor using a capacitively coupled plasma supply with a frequency of 13.56 MHz. The PECVD conditions were as follows: working pressure of 0.18 Torr, rf power of 200 W, and hexamethyldisiloxane (HMDSO):O2:Ar flow rate of 2.7:10:100 sccm. The growth process and film properties of SiOx are described in detail elsewhere [18]. An IZO film was deposited on the SiOx/Al2O3/PET barrier system by rf (13.56 MHz) magnetron sputtering using an IZO target of 10 wt.% ZnO (Samsung Corning Co., Ltd.). The deposition conditions were as follows: substrate temperature of approximately 90 °C, working pressure of 2.6 × 10− 3 Torr, rf power of 200 W, and Ar flow rate of 32 sccm. A two-point bending technique [1,2,19] was used for determining the degradation of water vapor impermeability and electrical conductance in the bent states. The size of the bending samples was fixed at a length of 10 cm and a width of 3 cm. The oxide coated surface of the bending samples was placed in such a way that it could be subjected to tensile stresses during the bending test, as shown in Fig. 1. The change in the electrical resistance of IZO films was measured as a function of the bending radius, which was approximately half of the distance between the two plates in the bending system [1]. IZO films with two different thicknesses, 10 nm and 135 nm, were deposited on the SiOx/Al2O3 barrier. The IZO film with a thickness of 135 nm exhibited a resistivity of 3.2 × 10− 4 Ω cm. The IZO film with a thickness of 10 nm was deposited on the barriers only for the purpose of measuring resistance changes corresponding to the mechanical failure of the barriers. In the bending tests, the displacement rate of a moving plate was 0.2 mm/s, while it was increased to more than 10 mm/s in the cyclic bending tests. The WVTR of 50-cm2 SiOx/Al2O3/PET samples was measured using a MOCON Permatran-W Model 3/33 instrument at 38 °C and 100% relative humidity. The film thickness was determined on Si wafers by using a surface profiler (P-11, KLA-Tencor). The surface morphology of the
Fig. 1. Schematic diagram of the two-point bending technique used in this study for determining mechanical failures of oxide films grown on flexible polymer substrates.
films was analyzed by atomic force microscopy (AFM, Digital Instruments Nanoscope IIIa) and field-emission scanning electron microscopy (FE-SEM, Hitachi S-4200). Optical transmittance was measured in the wavelength range of 300–800 nm by using a spectrophotometer (Varian, CARY5000). The surface temperature was measured using thermal labels (Nichiyu Giken Kogyo Co., Ltd.). 3. Results and discussion The excellent water vapor impermeability of the SiOx barriers was achieved by carrying out Ar ion-beam pretreatment and subsequent Al2O3 interlayer sputtering on PET surfaces, as shown in Fig. 2. WVTR values of the SiOx/Al2O3/PET system were at least two orders of magnitude smaller than those of the conventional SiOx/PET system. The SiOx/Al2O3/PET system had a WVTR value in the range of 10− 3 g m− 2 d− 1 at a SiOx thickness of 100 nm, while the SiOx/PET system had a WVTR larger than 1.0 g m− 2 d− 1 at the same SiOx thickness. The WVTR of the SiOx/Al2O3/PET system rapidly decreased as the SiOx thickness increased beyond 100 nm and could be attained a value in the range of 10− 4 g m− 2 d− 1 at SiOx thicknesses of more than 150 nm, as indicated in the inset of Fig. 2. We have previously reported the decrease in the WVTR to the range of 10− 3 g m− 2 d− 1 for a 300-nm-thick SiOx barrier that was grown on an Al2O3 interlayer [18]. In this study, the WVTR was further decreased by Ar ion-beam pretreatment of the PET surface. The PET surface was activated by heavy bombardment of Ar ions in the pretreatment process. This result is in good agreement with the
Fig. 2. Change in WVTR for SiOx/PET and SiOx/Al2O3 (10 nm)/PET systems as a function of SiOx thickness. The inset shows a magnification of the WVTR change in SiOx/Al2O3 (10 nm)/ion-beam treated PET system.
G.-H. Lee et al. / Thin Solid Films 518 (2010) 3075–3080
results of Wuu et al. [20], who reported a decrease in the WVTR as a result of the improvement in adhesion between SiOx and polyethersulfone (PES) by means of ion-beam treatments. A faster decrease in the WVTR with increasing SiOx thickness, as shown in Fig. 2, is associated with the improvement in adhesion between the Al2O3 interlayer and the PET substrate due to the ionbeam pretreatment. Very large mechanical stresses are induced at the interface between the Al2O3 interlayer and the PET substrate because of very different thermal expansion and elastic properties. The continuous stress build-up with increasing SiOx thickness leads to the adhesive failure of the overgrown oxide multilayers on the PET substrate. Delamination of the oxide multilayers from the PET substrate can readily occur in the SiOx/Al2O3/PET system prepared without the ion-beam pretreatment because the adhesion is poor at the interface [1,2]. As a result, the water vapor diffusion into the interface hinders the reduction of WVTR values in thick SiOx barriers. However, very high adhesion levels between the oxide multilayers and the PET, which are achieved by applying the ion-beam pretreatment, suppress the delamination at the interface and promote the fast decrease of the WVTR with an increase in the SiOx thickness. A considerable amount of research has been carried out in recent years on the improvement of adhesion of Al2O3 layers on PET substrates pretreated by ion beams or plasmas [21–24]. It is generally agreed that good adhesion is attributed to the formation of Al–O–C covalent bonds at the interface between the Al2O3 layers and the PET substrates [2,21,23]. In the pretreatment process of PET substrates, the kinetic energy of ions is sufficient to break C = O double bonds and to form the polar functionalities, including − OH, −OOH, and − COOH, at the PET surface [23,25]. The polar functionalities provide more favorable sites for the initial adsorption of Al atoms and for the formation of the strong Al−O−C covalent bonds. The highest adhesion between the Al2O3 layer and the PET substrate can be attained by maximizing the polar functionalities at the PET surface via a pretreatment process using ion beams or plasmas [2]. The increase in the polar functionalities can be exhibited by measuring the polar component of the surface free energy of PET substrates. Here, the polar component was determined from the contact angle of a polar liquid (water) with the PET and the contact angles of various non-polar liquids (glycerol, bromonaphthalene, formamide, and ethylene glycol) with the PET. The surface of untreated PET substrates showed a contact angle of 73° for water. The polar component of the surface free energy was approximately 4.9 mJ/m2 prior to the Ar ion-beam treatment. The ion-beam treatment induced a significant increase in the polar component, 22.8 mJ/m2, and a reduction in the contact angle, 37°. The changes are indicative of the magnitude of the formation of Al–O–C covalent bonds at the PET surface due to the ion-beam treatment. The impermeability of SiOx barriers is strongly dependent on the morphological evolution of the barriers. Fig. 3 shows the morphological discrepancies between SiOx/Al2O3/PET and conventional SiOx/PET systems. Fig. 3(a) shows the surface of an untreated PET substrate having a mean surface roughness of 0.97 nm. There was no apparent morphological modification after the Ar ion-beam treatment and subsequent Al2O3 interlayer sputtering, as illustrated in Fig. 3(b). When a PECVD SiOx barrier was grown directly on the PET surface, a rough granular-type surface was formed, as shown in Fig. 3(c). However, the SiOx barrier on the sputtered Al2O3 interlayer appeared relatively smooth, as shown in Fig. 3(d). Fig. 3(e) shows that large numbers of pinholes and granular boundaries were formed in the 300nm-thick SiO x barrier of the SiO x/PET system. Fig. 3(f) clearly demonstrates the successful elimination of such defects in the barrier of the SiOx/Al2O3/PET system having the same SiOx thickness. Microscopic pinholes, several tens of nanometers in cross-sectional size, were visualized on the surface of the SiOx/PET system with a surface density of 4–5 μm− 2 by carrying out a reactive ion etching process in low-pressure oxygen plasma for 20 min. However, no visible micro-
3077
Fig. 3. Surface morphology evolution in SiOx/PET and SiOx/Al2O3/PET systems. AFM images show the surfaces of (a) an untreated PET substrate, (b) Al2O3 (10 nm)/PET, (c) SiOx (60 nm)/PET, and (d) SiOx (60 nm)/Al2O3 (10 nm)/PET systems. The Rms roughness was measured by scanning 3 × 3 μm2 using the AFM tapping mode. FE-SEM images show SiOx surfaces of (e) SiOx (300 nm)/PET and (f) SiOx (300 nm)/Al2O3 (10 nm)/PET systems.
scopic pinhole was detected on the surface of the SiOx/Al2O3/PET system even after etching by oxygen plasmas. Therefore, the excellent water vapor impermeability of the SiOx/Al2O3/PET system, as demonstrated in Fig. 2, can be understood as to be the result of (i) the improvement in the adhesion between the PET and the sputtered Al2O3 interlayer because of the pretreatment of the PET surfaces with Ar ion beams and (ii) the elimination of microscopic defects in the SiOx barrier grown on the sputtered Al2O3 interlayer. Fig. 4 shows the light transmittance of the SiOx/Al2O3/PET and IZO/ SiOx/Al2O3/PET systems under different SiOx barrier thicknesses: 90 nm and 300 nm. The transmittance of the SiOx/Al2O3/PET system at a wavelength of 550 nm was approximately 87%. The transmittance
Fig. 4. Optical transmittance of SiOx/Al2O3/PET and IZO/SiOx/Al2O3/PET systems.
3078
G.-H. Lee et al. / Thin Solid Films 518 (2010) 3075–3080
of the SiOx/Al2O3/PET system was not affected by changes in the SiOx thickness. In the case of the IZO/SiOx/Al2O3/PET system having a 135nm-thick IZO film, transmittance decreased at wavelengths larger than 600 nm and the peak transmittance shifted to longer wavelengths. These changes in transmittance can be attributed to the increase in the reflectance due to the IZO film [13]. Fig. 5 shows the change in the resistance of IZO films grown on the SiOx/Al2O3/PET system for different bending radii. When thin and brittle films are subjected to externally induced stresses in bend geometries, cracks are considered to be the main cause of film failure [2,15]. Cracks are formed under induced stress at microscopic defects such as pinholes in the films and surface defects on the underlying polymer substrates [2]. The cracks then propagate from these defects with an increase in stress. Eventually, the cracks propagate to the entire width of the samples along the direction perpendicular to the direction of the loading of stresses. Leterrier et al. [15] found that the initiation of crack propagation could be detected at a resistance change of roughly 10%. Therefore, in this study, the 10% increase in resistance was considered to be the criterion for the conductive failure of the IZO films. The failure behavior of SiOx/Al2O3/PET barrier systems is shown in Fig. 5(a). Very thin IZO films were coated on the barrier systems, and the formation of cracks in the barrier systems was indirectly detected from the resistance change of the IZO films. The thickness of IZO films, 10 nm, is so thin that it could not influence the failure of the barrier systems. By comparing Fig. 5(a) and (b), it is observed that the bending radius could be reduced to approximately 20 mm without causing conductive failure in the case of the IZO (135 nm)/SiOx
Fig. 5. Normalized resistance change in (a) IZO (10 nm) and (b) IZO (135 nm) grown on barrier systems as a function of bending radius. The inset shows the change of stress distribution in a SiOx (300 nm) barrier grown on a PET (188 μm) substrate as a function of bending curvature.
(300 nm)/Al2O3 (10 nm)/PET system, while it could be further reduced to approximately 10 mm in the case of the IZO (10 nm)/ SiOx (300 nm)/Al2O3 (10 nm)/PET system without any conductive failure. Therefore, we concluded that the failure of the IZO (135 nm)/ SiOx/Al2O3/PET systems was dominated by the formation of cracks in the IZO films because the sputtered IZO films were more brittle than the PECVD SiOx barriers in bent states. This result can be explained by structural differences in films grown using PECVD and sputtering processes. Dense SiOx barriers having a small number of defects were grown by the PECVD process because of active surface reactions between energetic gaseous species and adsorbed atoms [26,27]. In contrast, the IZO films grown by the sputtering process without active surface reactions might contain a relatively high number of defects [28]. This led to the poor bending performance of the sputtered IZO films because of the formation of a large number of cracks even under small induced stress. The tensile stress imposed on SiOx barriers with increasing bending radius can be calculated using the following stress formula [29]: 3
σf =
Es ts 6 Rtf ðts + tf Þ
where σf is the tensile stress accumulated in the SiO x barrier, Es (5.53 GPa) is the Young's modulus of the PET substrate, R is the bending radius, and ts and tf are the PET and SiOx thickness, respectively. The inset of Fig. 5(a) shows the stress-bending radius relation of a 300nm-thick SiOx barrier on a mechanically deformed PET substrate. The conductive failure at a bending radius of 10 mm clearly corresponds to the fast increment in stress at the same bending radius. Fig. 5(b) shows that the conductive failure of the IZO films is strongly affected by the existence of the SiOx/Al2O3 barrier films. It is clear that the growth of IZO films on the SiOx/Al2O3/PET system significantly delayed its conductive failure as compared to that in the case of IZO films directly grown on PET. Furthermore, a reduction in the SiOx thickness in IZO/SiOx/Al2O3/PET systems suppressed the increase in the resistance of the IZO films. The conductive failure of 135-nm-thick IZO films was detected at bending radii of 13.4 mm and 16.7 mm at SiOx thicknesses of 90 nm and 300 nm, respectively. An instantaneous increase in the resistance of the IZO film grown on 90nm-thick SiOx was observed when the bending radius was reduced to a value less than 13.4 mm. However, in the case of the IZO film grown on the 300-nm-thick SiOx, there was a significant difference in the bending radius between the initiation of resistance increase and conductive failure. The failure behavior of IZO films grown on the SiOx/Al2O3/PET system is not fully understood. However, two possible reasons are suggested to explain the failure behavior. The first reason can be the improvement in adhesion between the IZO/SiOx/Al2O3 multilayer and PET because of the Al2O3 interlayer. The second reason can be the increase in the internal compressive stress in the PECVD SiOx barrier. A steady build-up of the internal compressive stress has been observed with increasing thicknesses in dense films [2,26]. Increasing internal compressive stress in the SiOx barrier could improve the strength of the IZO/SiOx/Al2O3/PET system by improving its resistance against the tensile cracking failure. However, cracks were eventually formed in the system as a result of the stress relaxation process in thick SiOx barriers [2]. In order to confirm the dependence of the internal compressive stress on the film thickness, the internal compressive stress of SiOx barriers with three different thicknesses, 90 nm, 300 nm, and 600 nm, was measured by a cantilever method using 100-μm-thick 5 × 50 mm2 Si substrates. A general tendency was observed for the internal stress evolution although the stresses were not identical between SiOx/Si and SiOx/PET systems. The stress in SiOx barriers had a tendency to increase from 89 MPa to 214 MPa when the SiOx thickness increased from 90 nm to 300 nm. The stress relaxation process was observed with a further increase in the SiOx thickness to
G.-H. Lee et al. / Thin Solid Films 518 (2010) 3075–3080
3079
Table 1 Change in WVTR under cyclic bendings of SiOx/Al2O3/PET and IZO/SiOx/Al2O3/PET systems. Film sample
SiOx (90 nm)/Al2O3/PET SiOx (300 nm)/Al2O3/PET IZO (135 nm)/SiOx (90 nm)/Al2O3/PET IZO (135 nm)/SiOx (300 nm)/Al2O3/PET
Fig. 6. WVTR change in the SiOx (300 nm)/Al2O3 (10 nm)/PET and IZO (135 nm)/SiOx (300 nm)/Al2O3 (10 nm)/PET systems as a function of bending curvature.
600 nm; the stress decreased to 158 MPa at 600 nm. The externally induced tensile stress continuously increased with increasing barrier thickness; cracks were formed when the compressive stress was overcome by the tensile stress. This relaxation process can explain the conductive failure of 300-nm-thick SiOx at a bending radius that is considerably larger than that of 90-nm-thick SiOx. The water vapor impermeability of the systems was severely limited by cracks that were formed at fairly large bending radii considerably before the increase in resistance, as shown in Fig. 6. The WVTR of the SiOx (300 nm)/Al2O3 (10 nm)/PET system increased from b10− 3 g m− 2 d− 1 to 2.5 × 10− 2 g m− 2 d− 1 as the bending radius was reduced from 30 mm to 15 mm. However, no change in resistance was detected at these bending radii in the case of a 10-nm-thick IZO film grown on the SiOx
Bending radius (mm)
WVTR (g m− 2 d− 1) Initial
After 1000 cycles
25 25 35 35
2 × 10− 2 b 5 × 10− 3 b 5 × 10− 3 b 5 × 10− 3
2 × 10− 2 1 × 10− 2 b 5 × 10− 3 7 × 10− 3
(300 nm)/Al2O3 (10 nm)/PET system. The impermeability of the IZO (135 nm)/SiOx (300 nm)/Al2O3 (10 nm)/PET system severely degraded with a reduction in the bending radius, and its WVTR value increased to 1.9 × 10− 2 g m− 2 d− 1 at a 20-mm bending radius. For both the SiOx (300 nm)/Al2O3 (10 nm)/PET and the IZO (135 nm)/SiOx (300 nm)/ Al2O3 (10 nm)/PET systems, the WVTR at bending radii leading to conductive failure was more than two orders of magnitude larger than the WVTR at a bending radius of 30 mm leading to no increase in resistance. The changes in resistance and WVTR under recurring deformation conditions were determined by performing cyclic bending tests. Fig. 7(a) and (b) illustrate that Al2O3 interlayer sputtering is a critical factor in suppressing the formation of cracks. The IZO/SiOx/PET systems without the interlayer failed in the early stages of the bending cycles; however, the IZO/SiOx/Al2O3/PET systems showed no such failure in these stages. In the case of the IZO/SiOx/Al2O3/PET system, Table 1 shows that conductive failure and impermeability degradation were suppressed during the bending cycles by the reduction in the SiOx barrier thickness from 300 nm to 90 nm. These results imply that we can achieve excellent conductance and water vapor impermeability in bend geometries by controlling the SiOx thickness. An IZO (135 nm)/SiOx (90 nm)/Al2O3 (10 nm)/PET system showed excellent mechanical and electrical properties, a TCO resistivity of 3.2 × 10− 4 Ω cm and a WVTR of b5 × 10− 3 g m− 2 d− 1, after 1000 bending cycles at a 35-mm bending radius. Further reduction in the WVTR to the range of 10− 4 g m− 2 d− 1 could be achieved by slightly increasing the SiOx thickness to 150 nm, without any noticeable change in the bending characteristic of the IZO/ SiOx/Al2O3/PET system. 4. Conclusion
Fig. 7. Dependence of normalized resistance change on bending cycles for (a) IZO (10 nm) and (b) IZO (135 nm) grown on barrier systems: SiOx (90 and 300 nm)/Al2O3 (10 nm)/PET and SiOx (300 nm)/PET.
The water vapor impermeability and electrical conductance of IZO/SiOx/Al2O3/PET systems were investigated by carrying out the two-point bending and cyclic bending tests. Low WVTR values of b5 × 10− 3 g m− 2 d− 1 were obtained in a SiOx/Al2O3/PET system as the SiOx thickness increased beyond 100 nm. A two-step pretreatment process involving (i) Ar ion-beam treatment of PET surfaces and (ii) subsequent sputtering of a 10-nm-thick Al2O3 interlayer was carried out on PET surfaces prior to SiOx PECVD processes. IZO films sputtered on the SiOx/Al2O3/PET system exhibited excellent mechanical stabilities in bend geometries, whereas the films grown directly on PET substrates underwent fatal crack formation. In the case of the SiOx/Al2O3/PET system, the conductive failure in the IZO films was significantly delayed by the improvement in cohesion between oxide coatings and PET substrates. The degradation of resistance and WVTR in severe bend geometries was suppressed by applying the IZO/SiOx/Al2O3/PET system with a properly chosen SiOx thickness. Large increments in the SiOx thickness negatively affect the WVTR and electrical conductance in bend geometries. An IZO (135 nm)/SiOx (90 nm)/Al2O3 (10 nm)/PET system exhibited a WVTR of b5 × 10− 3 g m− 2 d− 1 and a TCO resistivity of 3.2 × 10− 4 Ω cm after 1000 bending cycles at a 35-mm bending radius. However, there were degradations in WVTR and electrical conductance during the bending cycles at the same bending radius when the SiOx barrier thickness increased from 90 nm to 300 nm. The WVTR and
3080
G.-H. Lee et al. / Thin Solid Films 518 (2010) 3075–3080
resistance increased to 7 × 10− 3 g m− 2 d− 1 and 4.0 × 10− 4 Ω cm, respectively. Acknowledgment This work was financially supported by the fundamental R&D Program for Core Technology of Materials funded by the Ministry of Knowledge Economy, Republic of Korea. References [1] J. Lewis, Mater. Today 9 (2006) 38. [2] Y. Leterrier, Prog. Mater. Sci. 48 (2003) 1. [3] 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. 99. [4] M.S. Weaver, L.A. Michalski, K. Rajan, M.A. Rothman, J.A. Silvernail, J.J. Brown, P.E. Burrows, G.L. Graff, M.E. Gross, P.M. Martin, M. Hall, E. Mast, C. Bonham, W. Bennett, M. Zumhoff, Appl. Phys. Lett. 81 (2002) 2929. [5] J.S. Lewis, M.S. Weaver, IEEE J. Quantum Electron. 10 (2004) 45. [6] M.D. Groner, S.M. George, R.S. McLean, P.F. Carcia, Appl. Phys. Lett. 88 (2006) 051907. [7] E. Langereis, M. Creatore, S.B.S. Heil, M.C.M. van de Sanden, W.M.M. Kessels, Appl. Phys. Lett. 89 (2006) 081915. [8] Y.G. Tropsha, N.G. Harvey, J. Phys. Chem., B 101 (1997) 2259. [9] A.G. Erlat, R.J. Spontak, R.P. Clarke, T.C. Robinson, P.D. Haaland, Y. Tropsha, N.G. Harvey, E.A. Vogler, J. Phys. Chem., B 103 (1999) 6047. [10] N. Inagaki, V. Cech, K. Narushima, Y. Takechi, J. Appl. Polym. Sci. 104 (2007) 915.
[11] Y.S. Jung, J.Y. Seo, D.W. Lee, D.Y. Jeon, Thin Solid Films 445 (2003) 63. [12] B. Yaglioglu, Y.-J. Huang, H.-Y. Yeom, D.C. Paine, Thin Solid Films 496 (2006) 89. [13] N. Ito, Y. Sato, P.K. Song, A. Kaijio, K. Inoue, Y. Shigesato, Thin Solid Films 496 (2006) 99. [14] S.K. Park, J.I. Han, D.G. Moon, W.K. Kim, Jpn. J. Appl. Phys. 42 (2003) 623. [15] Y. Leterrier, L. Médico, F. Demarco, J.-A.E. Månson, U. Betz, M.F. Escola, M. Kharrazi Olsson, F. Atamny, Thin Solid Films 460 (2004) 156. [16] Y. Leterrier, L. Boogh, J. Andersons, J.A.E. Mansons, J. Polym. Sci., B 323 (1999) 63. [17] D.R. Cairns, R.P. Witte, D.K. Sparacin, S.M. Sachsman, Appl. Phys. Lett. 76 (2000) 1425. [18] J. Yun, S. Lee, Y. Jeong, H.-R. Lee, J.-D. Kwon, G.-H. Lee, Jpn. J. Appl. Phys. 48 (2009) 055503. [19] H. Gleskova, S. Wagner, Z. Suo, Appl. Phys. Lett. 75 (1999) 3011. [20] D.S. Wuu, W.C. Lo, C.C. Chiang, H.B. Lin, L.S. Chang, R.H. Horng, C.L. Huang, Y.J. Gao, Surf. Coat. Technol. 197 (2005) 253. [21] J.F. Silvain, A. Veyrat, J.J. Ehrhardt, Thin Solid Films 221 (1992) 114. [22] F. Milde, K. Goedicke, M. Fahland, Thin Solid Films 279 (1996) 169. [23] C.H. Bichler, H.-C. Langowski, U. Moosheimer, B. Seifert, J. Adhes. Sci. Technol. 11 (1997) 233. [24] B. Gupta, J. Hilborn, C.H. Hollenstein, C.J.G. Plummer, R. Houriet, N. Xanthopoulos, J. Appl. Phys. Sci. 78 (2000) 1083. [25] I. Tepermeister, H.H. Sawin, J. Vac. Sci. Technol., A, Vac. Surf. Films 10 (1992) 3149. [26] D.S. Wuu, T.N. Chen, C.C. Wu, C.C. Chiang, Y.P. Chen, R.H. Horng, F.S. Juang, Chem. Vap. Depos. 12 (2006) 220. [27] A.S. da Silva Sobrinho, N. Schühler, J.E. Klemberg-Sapieha, M.R. Wertheimer, M. Andrews, S.C. Gujrathi, J. Vac. Sci. Technol., A, Vac. Surf. Films 16 (1998) 2021. [28] A.S. da Silva Sobrinho, M. Latréche, G. Czeremuszkin, J.E. Klemberg-Sapieha, M.R. Wertheimer, J. Vac. Sci. Technol., A, Vac. Surf. Films 16 (1998) 3190. [29] 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.
Contents lists available at ScienceDirect
Thin Solid Films j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / t s f
Investigation of brittle failure in transparent conductive oxide and permeation barrier oxide multilayers on flexible polymers Gun-Hwan Lee, Jungheum Yun ⁎, Sunghun Lee, Yujeong Jeong, Jae-Hye Jung, Sang-Hyun Cho Materials Processing Division, Korean Institute of Materials Science, 531 Changwondaero, Changwon, Gyeongnam 641-831, Republic of Korea
a r t i c l e
i n f o
Available online 13 October 2009 Keywords: Barrier TCO SiOx IZO Bending Water vapor transmission rate (WVTR) Polymer Flexible substrate
a b s t r a c t An oxide multilayer structure—consisting of an indium zinc oxide (IZO) conductive layer, a silicon oxide (SiOx, x = 1.8) water vapor permeation barrier, and an aluminum oxide (Al2O3) interlayer—coated on polyethylene terephthalate (PET) is proposed as a transparent flexible substrate for display and photovoltaic applications. Vital properties of the multilayer, such as the low water vapor impermeability of the SiOx barrier and the high conductance of the IZO film, degraded considerably because of the crack formation in bend geometries, attributed to the large difference between elastic properties of the oxide films and polymers. In order to suppress the crack formation, a 10-nm-thick Al2O3 interlayer was sputtered on Ar ionbeam treated PET surfaces prior to a SiOx plasma-enhanced chemical vapor deposition (PECVD) process. Changes in the conductance and water vapor impermeability were investigated at different bending radii and bending cycles. It was found that the increases in resistance and water vapor transmission rate (WVTR) were significantly suppressed by the ion-beam PET pretreatment and by the sputtered Al2O3 interlayer. The resistance and WVTR of IZO/SiOx/Al2O3/PET systems could be kept low and invariable even in severely bent states by choosing the SiOx thickness properly. The IZO (135 nm)/SiOx (90 nm)/Al2O3 (10 nm)/PET system maintained a resistance of 3.2 × 10− 4 Ω cm and a WVTR of b 5 × 10− 3 g m2 d− 1 after 1000 bending cycles at a bending radius of 35 mm. © 2009 Elsevier B.V. All rights reserved.
1. Introduction The development of flexible substrates that are based on organic polymers is of critical importance for display and photovoltaic applications. Flexible substrates offer the advantages of mechanical flexibility, design freedom, optical transparency, low weight, and cost effectiveness [1,2]. A number of inorganic functional coatings on polymers have critical properties such as water vapor impermeability and electronic conductance, which are required for display and photovoltaic applications [3]. Water vapor impermeability is achieved by coating a thin oxide or nitride barrier on polymers, and electronic conductance is achieved by using a transparent conductive oxide (TCO), often coated on the barrier. The water vapor transmission rate (WVTR) represents the rate of penetration of water molecules into a barrier. Although recent developments in oxide/polymer multilayers [4,5] and atomic-layerdeposited oxide layers [6,7] can provide extremely low WVTR values of approximately 10− 6 g m− 2 d− 1, the commercial application of these barriers has been limited because of complexities in the manufacturing process and low throughput. However, the fabrication of barriers using plasma-enhanced chemical vapor deposition
⁎ Corresponding author. Tel.: +82 55 280 3515; fax: +82 55 280 3570. E-mail address: [email protected] (J. Yun). 0040-6090/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.tsf.2009.08.057
(PECVD) and physical vapor deposition (PVD) techniques, which are well-established in industry, has been thoroughly investigated to develop simple and cost-effective methods that yield WVTR values in the range of 10− 3–10− 4 g m− 2 d− 1 [8–10]. These WVTR values are sufficient for satisfying requirements for various applications such as organic thin film transistors (TFTs), liquid crystal displays (LCDs), and inorganic photovoltaics [1]. With regard to TCO materials that can be applied to flexible substrates, amorphous indium zinc oxide (IZO) has become the preferred option as a replacement for polycrystalline indium tin oxides (ITOs) because it requires low fabrication temperature and is less fragile [11–13]. The mechanical failure of brittle oxide coatings on flexible polymers is a serious issue. The formation of defects, such as cracks and debonding, in water vapor barriers and TCO films may be inevitable in situations in which bend geometries are required because of the large difference between the elastic properties of oxides and polymers [14,15]. The initiation and subsequent developments of defects mainly depend on the coating material, thickness, and interfacial adhesion [2]. Therefore, it is important to have a clear understanding of the failure behavior of coatings in the case of bend geometries, in order to prevent the destruction of devices in which they are used. The failure behavior of thin films coated on flexible polymers has been reported for individual water vapor permeation oxide barriers [16] and TCOs [14,15,17] in bend geometries. However, the failure behavior of complex multilayer structures in bend
3076
G.-H. Lee et al. / Thin Solid Films 518 (2010) 3075–3080
geometries has not been well understood. The effects of film deformation on the performance of multilayer coatings, which include both a water vapor permeation barrier and a TCO, on polymers are largely unknown. In this study, we achieved advanced water vapor impermeabilities in SiOx barriers by developing a two-step pretreatment process comprising (i) PET treatment with Ar ion beams and (ii) subsequent sputtering of a 10-nm-thick Al2O3 interlayer on the PET, prior to a SiOx PECVD process. This process brought about an improvement in both the morphology of the SiOx barrier and the adhesion of the barrier with a polymer substrate, resulting in significant barrier enhancement. Low WVTR values of b5 × 10− 3 g m− 2 d− 1 were obtained for a SiOx/Al2O3/PET system with a SiOx thickness of approximately 100 nm, and further decrease in the WVTR to the range of 10− 4 g m− 2 d− 1 could be readily expected from an increase in the SiOx thickness beyond 150 nm. Changes in the water vapor impermeability of SiOx/Al2O3/PET systems and the electrical resistance of subsequently grown IZO films were measured as functions of the bending radius and the bending cycle in bend geometries. The increase in the WVTR and resistance during bending tests was significantly suppressed in the case of the IZO/SiOx/Al2O3/PET system. 2. Experimental procedure The barrier system was composed of (i) a 10-nm-thick sputtered Al2O3 interlayer and (ii) a thicker PECVD SiOx (x = 1.8) barrier with a thickness of up to 750 nm. Prior to the growth of the Al2O3 interlayer, a 188-μm-thick PET substrate (Kimoto Co., Ltd.) was pretreated with Ar ion beams. The pretreatment conditions were as follows: working pressure of 1 × 10− 4 Torr, dc power of 220 W, Ar flow rate of 15 sccm, and processing time of 3 min. The Al2O3 interlayer was grown by a dc reactive magnetron sputtering system using a 4-in. Al target (Applied Science Corp.). The sputtering conditions were as follows: working pressure of 3 × 10− 3 Torr, dc power of 300 W, and Ar:O2 flow rate of 45:5 sccm. Subsequently, the SiOx barrier was deposited on the Al2O3 interlayer by an rf-PECVD reactor using a capacitively coupled plasma supply with a frequency of 13.56 MHz. The PECVD conditions were as follows: working pressure of 0.18 Torr, rf power of 200 W, and hexamethyldisiloxane (HMDSO):O2:Ar flow rate of 2.7:10:100 sccm. The growth process and film properties of SiOx are described in detail elsewhere [18]. An IZO film was deposited on the SiOx/Al2O3/PET barrier system by rf (13.56 MHz) magnetron sputtering using an IZO target of 10 wt.% ZnO (Samsung Corning Co., Ltd.). The deposition conditions were as follows: substrate temperature of approximately 90 °C, working pressure of 2.6 × 10− 3 Torr, rf power of 200 W, and Ar flow rate of 32 sccm. A two-point bending technique [1,2,19] was used for determining the degradation of water vapor impermeability and electrical conductance in the bent states. The size of the bending samples was fixed at a length of 10 cm and a width of 3 cm. The oxide coated surface of the bending samples was placed in such a way that it could be subjected to tensile stresses during the bending test, as shown in Fig. 1. The change in the electrical resistance of IZO films was measured as a function of the bending radius, which was approximately half of the distance between the two plates in the bending system [1]. IZO films with two different thicknesses, 10 nm and 135 nm, were deposited on the SiOx/Al2O3 barrier. The IZO film with a thickness of 135 nm exhibited a resistivity of 3.2 × 10− 4 Ω cm. The IZO film with a thickness of 10 nm was deposited on the barriers only for the purpose of measuring resistance changes corresponding to the mechanical failure of the barriers. In the bending tests, the displacement rate of a moving plate was 0.2 mm/s, while it was increased to more than 10 mm/s in the cyclic bending tests. The WVTR of 50-cm2 SiOx/Al2O3/PET samples was measured using a MOCON Permatran-W Model 3/33 instrument at 38 °C and 100% relative humidity. The film thickness was determined on Si wafers by using a surface profiler (P-11, KLA-Tencor). The surface morphology of the
Fig. 1. Schematic diagram of the two-point bending technique used in this study for determining mechanical failures of oxide films grown on flexible polymer substrates.
films was analyzed by atomic force microscopy (AFM, Digital Instruments Nanoscope IIIa) and field-emission scanning electron microscopy (FE-SEM, Hitachi S-4200). Optical transmittance was measured in the wavelength range of 300–800 nm by using a spectrophotometer (Varian, CARY5000). The surface temperature was measured using thermal labels (Nichiyu Giken Kogyo Co., Ltd.). 3. Results and discussion The excellent water vapor impermeability of the SiOx barriers was achieved by carrying out Ar ion-beam pretreatment and subsequent Al2O3 interlayer sputtering on PET surfaces, as shown in Fig. 2. WVTR values of the SiOx/Al2O3/PET system were at least two orders of magnitude smaller than those of the conventional SiOx/PET system. The SiOx/Al2O3/PET system had a WVTR value in the range of 10− 3 g m− 2 d− 1 at a SiOx thickness of 100 nm, while the SiOx/PET system had a WVTR larger than 1.0 g m− 2 d− 1 at the same SiOx thickness. The WVTR of the SiOx/Al2O3/PET system rapidly decreased as the SiOx thickness increased beyond 100 nm and could be attained a value in the range of 10− 4 g m− 2 d− 1 at SiOx thicknesses of more than 150 nm, as indicated in the inset of Fig. 2. We have previously reported the decrease in the WVTR to the range of 10− 3 g m− 2 d− 1 for a 300-nm-thick SiOx barrier that was grown on an Al2O3 interlayer [18]. In this study, the WVTR was further decreased by Ar ion-beam pretreatment of the PET surface. The PET surface was activated by heavy bombardment of Ar ions in the pretreatment process. This result is in good agreement with the
Fig. 2. Change in WVTR for SiOx/PET and SiOx/Al2O3 (10 nm)/PET systems as a function of SiOx thickness. The inset shows a magnification of the WVTR change in SiOx/Al2O3 (10 nm)/ion-beam treated PET system.
G.-H. Lee et al. / Thin Solid Films 518 (2010) 3075–3080
results of Wuu et al. [20], who reported a decrease in the WVTR as a result of the improvement in adhesion between SiOx and polyethersulfone (PES) by means of ion-beam treatments. A faster decrease in the WVTR with increasing SiOx thickness, as shown in Fig. 2, is associated with the improvement in adhesion between the Al2O3 interlayer and the PET substrate due to the ionbeam pretreatment. Very large mechanical stresses are induced at the interface between the Al2O3 interlayer and the PET substrate because of very different thermal expansion and elastic properties. The continuous stress build-up with increasing SiOx thickness leads to the adhesive failure of the overgrown oxide multilayers on the PET substrate. Delamination of the oxide multilayers from the PET substrate can readily occur in the SiOx/Al2O3/PET system prepared without the ion-beam pretreatment because the adhesion is poor at the interface [1,2]. As a result, the water vapor diffusion into the interface hinders the reduction of WVTR values in thick SiOx barriers. However, very high adhesion levels between the oxide multilayers and the PET, which are achieved by applying the ion-beam pretreatment, suppress the delamination at the interface and promote the fast decrease of the WVTR with an increase in the SiOx thickness. A considerable amount of research has been carried out in recent years on the improvement of adhesion of Al2O3 layers on PET substrates pretreated by ion beams or plasmas [21–24]. It is generally agreed that good adhesion is attributed to the formation of Al–O–C covalent bonds at the interface between the Al2O3 layers and the PET substrates [2,21,23]. In the pretreatment process of PET substrates, the kinetic energy of ions is sufficient to break C = O double bonds and to form the polar functionalities, including − OH, −OOH, and − COOH, at the PET surface [23,25]. The polar functionalities provide more favorable sites for the initial adsorption of Al atoms and for the formation of the strong Al−O−C covalent bonds. The highest adhesion between the Al2O3 layer and the PET substrate can be attained by maximizing the polar functionalities at the PET surface via a pretreatment process using ion beams or plasmas [2]. The increase in the polar functionalities can be exhibited by measuring the polar component of the surface free energy of PET substrates. Here, the polar component was determined from the contact angle of a polar liquid (water) with the PET and the contact angles of various non-polar liquids (glycerol, bromonaphthalene, formamide, and ethylene glycol) with the PET. The surface of untreated PET substrates showed a contact angle of 73° for water. The polar component of the surface free energy was approximately 4.9 mJ/m2 prior to the Ar ion-beam treatment. The ion-beam treatment induced a significant increase in the polar component, 22.8 mJ/m2, and a reduction in the contact angle, 37°. The changes are indicative of the magnitude of the formation of Al–O–C covalent bonds at the PET surface due to the ion-beam treatment. The impermeability of SiOx barriers is strongly dependent on the morphological evolution of the barriers. Fig. 3 shows the morphological discrepancies between SiOx/Al2O3/PET and conventional SiOx/PET systems. Fig. 3(a) shows the surface of an untreated PET substrate having a mean surface roughness of 0.97 nm. There was no apparent morphological modification after the Ar ion-beam treatment and subsequent Al2O3 interlayer sputtering, as illustrated in Fig. 3(b). When a PECVD SiOx barrier was grown directly on the PET surface, a rough granular-type surface was formed, as shown in Fig. 3(c). However, the SiOx barrier on the sputtered Al2O3 interlayer appeared relatively smooth, as shown in Fig. 3(d). Fig. 3(e) shows that large numbers of pinholes and granular boundaries were formed in the 300nm-thick SiO x barrier of the SiO x/PET system. Fig. 3(f) clearly demonstrates the successful elimination of such defects in the barrier of the SiOx/Al2O3/PET system having the same SiOx thickness. Microscopic pinholes, several tens of nanometers in cross-sectional size, were visualized on the surface of the SiOx/PET system with a surface density of 4–5 μm− 2 by carrying out a reactive ion etching process in low-pressure oxygen plasma for 20 min. However, no visible micro-
3077
Fig. 3. Surface morphology evolution in SiOx/PET and SiOx/Al2O3/PET systems. AFM images show the surfaces of (a) an untreated PET substrate, (b) Al2O3 (10 nm)/PET, (c) SiOx (60 nm)/PET, and (d) SiOx (60 nm)/Al2O3 (10 nm)/PET systems. The Rms roughness was measured by scanning 3 × 3 μm2 using the AFM tapping mode. FE-SEM images show SiOx surfaces of (e) SiOx (300 nm)/PET and (f) SiOx (300 nm)/Al2O3 (10 nm)/PET systems.
scopic pinhole was detected on the surface of the SiOx/Al2O3/PET system even after etching by oxygen plasmas. Therefore, the excellent water vapor impermeability of the SiOx/Al2O3/PET system, as demonstrated in Fig. 2, can be understood as to be the result of (i) the improvement in the adhesion between the PET and the sputtered Al2O3 interlayer because of the pretreatment of the PET surfaces with Ar ion beams and (ii) the elimination of microscopic defects in the SiOx barrier grown on the sputtered Al2O3 interlayer. Fig. 4 shows the light transmittance of the SiOx/Al2O3/PET and IZO/ SiOx/Al2O3/PET systems under different SiOx barrier thicknesses: 90 nm and 300 nm. The transmittance of the SiOx/Al2O3/PET system at a wavelength of 550 nm was approximately 87%. The transmittance
Fig. 4. Optical transmittance of SiOx/Al2O3/PET and IZO/SiOx/Al2O3/PET systems.
3078
G.-H. Lee et al. / Thin Solid Films 518 (2010) 3075–3080
of the SiOx/Al2O3/PET system was not affected by changes in the SiOx thickness. In the case of the IZO/SiOx/Al2O3/PET system having a 135nm-thick IZO film, transmittance decreased at wavelengths larger than 600 nm and the peak transmittance shifted to longer wavelengths. These changes in transmittance can be attributed to the increase in the reflectance due to the IZO film [13]. Fig. 5 shows the change in the resistance of IZO films grown on the SiOx/Al2O3/PET system for different bending radii. When thin and brittle films are subjected to externally induced stresses in bend geometries, cracks are considered to be the main cause of film failure [2,15]. Cracks are formed under induced stress at microscopic defects such as pinholes in the films and surface defects on the underlying polymer substrates [2]. The cracks then propagate from these defects with an increase in stress. Eventually, the cracks propagate to the entire width of the samples along the direction perpendicular to the direction of the loading of stresses. Leterrier et al. [15] found that the initiation of crack propagation could be detected at a resistance change of roughly 10%. Therefore, in this study, the 10% increase in resistance was considered to be the criterion for the conductive failure of the IZO films. The failure behavior of SiOx/Al2O3/PET barrier systems is shown in Fig. 5(a). Very thin IZO films were coated on the barrier systems, and the formation of cracks in the barrier systems was indirectly detected from the resistance change of the IZO films. The thickness of IZO films, 10 nm, is so thin that it could not influence the failure of the barrier systems. By comparing Fig. 5(a) and (b), it is observed that the bending radius could be reduced to approximately 20 mm without causing conductive failure in the case of the IZO (135 nm)/SiOx
Fig. 5. Normalized resistance change in (a) IZO (10 nm) and (b) IZO (135 nm) grown on barrier systems as a function of bending radius. The inset shows the change of stress distribution in a SiOx (300 nm) barrier grown on a PET (188 μm) substrate as a function of bending curvature.
(300 nm)/Al2O3 (10 nm)/PET system, while it could be further reduced to approximately 10 mm in the case of the IZO (10 nm)/ SiOx (300 nm)/Al2O3 (10 nm)/PET system without any conductive failure. Therefore, we concluded that the failure of the IZO (135 nm)/ SiOx/Al2O3/PET systems was dominated by the formation of cracks in the IZO films because the sputtered IZO films were more brittle than the PECVD SiOx barriers in bent states. This result can be explained by structural differences in films grown using PECVD and sputtering processes. Dense SiOx barriers having a small number of defects were grown by the PECVD process because of active surface reactions between energetic gaseous species and adsorbed atoms [26,27]. In contrast, the IZO films grown by the sputtering process without active surface reactions might contain a relatively high number of defects [28]. This led to the poor bending performance of the sputtered IZO films because of the formation of a large number of cracks even under small induced stress. The tensile stress imposed on SiOx barriers with increasing bending radius can be calculated using the following stress formula [29]: 3
σf =
Es ts 6 Rtf ðts + tf Þ
where σf is the tensile stress accumulated in the SiO x barrier, Es (5.53 GPa) is the Young's modulus of the PET substrate, R is the bending radius, and ts and tf are the PET and SiOx thickness, respectively. The inset of Fig. 5(a) shows the stress-bending radius relation of a 300nm-thick SiOx barrier on a mechanically deformed PET substrate. The conductive failure at a bending radius of 10 mm clearly corresponds to the fast increment in stress at the same bending radius. Fig. 5(b) shows that the conductive failure of the IZO films is strongly affected by the existence of the SiOx/Al2O3 barrier films. It is clear that the growth of IZO films on the SiOx/Al2O3/PET system significantly delayed its conductive failure as compared to that in the case of IZO films directly grown on PET. Furthermore, a reduction in the SiOx thickness in IZO/SiOx/Al2O3/PET systems suppressed the increase in the resistance of the IZO films. The conductive failure of 135-nm-thick IZO films was detected at bending radii of 13.4 mm and 16.7 mm at SiOx thicknesses of 90 nm and 300 nm, respectively. An instantaneous increase in the resistance of the IZO film grown on 90nm-thick SiOx was observed when the bending radius was reduced to a value less than 13.4 mm. However, in the case of the IZO film grown on the 300-nm-thick SiOx, there was a significant difference in the bending radius between the initiation of resistance increase and conductive failure. The failure behavior of IZO films grown on the SiOx/Al2O3/PET system is not fully understood. However, two possible reasons are suggested to explain the failure behavior. The first reason can be the improvement in adhesion between the IZO/SiOx/Al2O3 multilayer and PET because of the Al2O3 interlayer. The second reason can be the increase in the internal compressive stress in the PECVD SiOx barrier. A steady build-up of the internal compressive stress has been observed with increasing thicknesses in dense films [2,26]. Increasing internal compressive stress in the SiOx barrier could improve the strength of the IZO/SiOx/Al2O3/PET system by improving its resistance against the tensile cracking failure. However, cracks were eventually formed in the system as a result of the stress relaxation process in thick SiOx barriers [2]. In order to confirm the dependence of the internal compressive stress on the film thickness, the internal compressive stress of SiOx barriers with three different thicknesses, 90 nm, 300 nm, and 600 nm, was measured by a cantilever method using 100-μm-thick 5 × 50 mm2 Si substrates. A general tendency was observed for the internal stress evolution although the stresses were not identical between SiOx/Si and SiOx/PET systems. The stress in SiOx barriers had a tendency to increase from 89 MPa to 214 MPa when the SiOx thickness increased from 90 nm to 300 nm. The stress relaxation process was observed with a further increase in the SiOx thickness to
G.-H. Lee et al. / Thin Solid Films 518 (2010) 3075–3080
3079
Table 1 Change in WVTR under cyclic bendings of SiOx/Al2O3/PET and IZO/SiOx/Al2O3/PET systems. Film sample
SiOx (90 nm)/Al2O3/PET SiOx (300 nm)/Al2O3/PET IZO (135 nm)/SiOx (90 nm)/Al2O3/PET IZO (135 nm)/SiOx (300 nm)/Al2O3/PET
Fig. 6. WVTR change in the SiOx (300 nm)/Al2O3 (10 nm)/PET and IZO (135 nm)/SiOx (300 nm)/Al2O3 (10 nm)/PET systems as a function of bending curvature.
600 nm; the stress decreased to 158 MPa at 600 nm. The externally induced tensile stress continuously increased with increasing barrier thickness; cracks were formed when the compressive stress was overcome by the tensile stress. This relaxation process can explain the conductive failure of 300-nm-thick SiOx at a bending radius that is considerably larger than that of 90-nm-thick SiOx. The water vapor impermeability of the systems was severely limited by cracks that were formed at fairly large bending radii considerably before the increase in resistance, as shown in Fig. 6. The WVTR of the SiOx (300 nm)/Al2O3 (10 nm)/PET system increased from b10− 3 g m− 2 d− 1 to 2.5 × 10− 2 g m− 2 d− 1 as the bending radius was reduced from 30 mm to 15 mm. However, no change in resistance was detected at these bending radii in the case of a 10-nm-thick IZO film grown on the SiOx
Bending radius (mm)
WVTR (g m− 2 d− 1) Initial
After 1000 cycles
25 25 35 35
2 × 10− 2 b 5 × 10− 3 b 5 × 10− 3 b 5 × 10− 3
2 × 10− 2 1 × 10− 2 b 5 × 10− 3 7 × 10− 3
(300 nm)/Al2O3 (10 nm)/PET system. The impermeability of the IZO (135 nm)/SiOx (300 nm)/Al2O3 (10 nm)/PET system severely degraded with a reduction in the bending radius, and its WVTR value increased to 1.9 × 10− 2 g m− 2 d− 1 at a 20-mm bending radius. For both the SiOx (300 nm)/Al2O3 (10 nm)/PET and the IZO (135 nm)/SiOx (300 nm)/ Al2O3 (10 nm)/PET systems, the WVTR at bending radii leading to conductive failure was more than two orders of magnitude larger than the WVTR at a bending radius of 30 mm leading to no increase in resistance. The changes in resistance and WVTR under recurring deformation conditions were determined by performing cyclic bending tests. Fig. 7(a) and (b) illustrate that Al2O3 interlayer sputtering is a critical factor in suppressing the formation of cracks. The IZO/SiOx/PET systems without the interlayer failed in the early stages of the bending cycles; however, the IZO/SiOx/Al2O3/PET systems showed no such failure in these stages. In the case of the IZO/SiOx/Al2O3/PET system, Table 1 shows that conductive failure and impermeability degradation were suppressed during the bending cycles by the reduction in the SiOx barrier thickness from 300 nm to 90 nm. These results imply that we can achieve excellent conductance and water vapor impermeability in bend geometries by controlling the SiOx thickness. An IZO (135 nm)/SiOx (90 nm)/Al2O3 (10 nm)/PET system showed excellent mechanical and electrical properties, a TCO resistivity of 3.2 × 10− 4 Ω cm and a WVTR of b5 × 10− 3 g m− 2 d− 1, after 1000 bending cycles at a 35-mm bending radius. Further reduction in the WVTR to the range of 10− 4 g m− 2 d− 1 could be achieved by slightly increasing the SiOx thickness to 150 nm, without any noticeable change in the bending characteristic of the IZO/ SiOx/Al2O3/PET system. 4. Conclusion
Fig. 7. Dependence of normalized resistance change on bending cycles for (a) IZO (10 nm) and (b) IZO (135 nm) grown on barrier systems: SiOx (90 and 300 nm)/Al2O3 (10 nm)/PET and SiOx (300 nm)/PET.
The water vapor impermeability and electrical conductance of IZO/SiOx/Al2O3/PET systems were investigated by carrying out the two-point bending and cyclic bending tests. Low WVTR values of b5 × 10− 3 g m− 2 d− 1 were obtained in a SiOx/Al2O3/PET system as the SiOx thickness increased beyond 100 nm. A two-step pretreatment process involving (i) Ar ion-beam treatment of PET surfaces and (ii) subsequent sputtering of a 10-nm-thick Al2O3 interlayer was carried out on PET surfaces prior to SiOx PECVD processes. IZO films sputtered on the SiOx/Al2O3/PET system exhibited excellent mechanical stabilities in bend geometries, whereas the films grown directly on PET substrates underwent fatal crack formation. In the case of the SiOx/Al2O3/PET system, the conductive failure in the IZO films was significantly delayed by the improvement in cohesion between oxide coatings and PET substrates. The degradation of resistance and WVTR in severe bend geometries was suppressed by applying the IZO/SiOx/Al2O3/PET system with a properly chosen SiOx thickness. Large increments in the SiOx thickness negatively affect the WVTR and electrical conductance in bend geometries. An IZO (135 nm)/SiOx (90 nm)/Al2O3 (10 nm)/PET system exhibited a WVTR of b5 × 10− 3 g m− 2 d− 1 and a TCO resistivity of 3.2 × 10− 4 Ω cm after 1000 bending cycles at a 35-mm bending radius. However, there were degradations in WVTR and electrical conductance during the bending cycles at the same bending radius when the SiOx barrier thickness increased from 90 nm to 300 nm. The WVTR and
3080
G.-H. Lee et al. / Thin Solid Films 518 (2010) 3075–3080
resistance increased to 7 × 10− 3 g m− 2 d− 1 and 4.0 × 10− 4 Ω cm, respectively. Acknowledgment This work was financially supported by the fundamental R&D Program for Core Technology of Materials funded by the Ministry of Knowledge Economy, Republic of Korea. References [1] J. Lewis, Mater. Today 9 (2006) 38. [2] Y. Leterrier, Prog. Mater. Sci. 48 (2003) 1. [3] 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. 99. [4] M.S. Weaver, L.A. Michalski, K. Rajan, M.A. Rothman, J.A. Silvernail, J.J. Brown, P.E. Burrows, G.L. Graff, M.E. Gross, P.M. Martin, M. Hall, E. Mast, C. Bonham, W. Bennett, M. Zumhoff, Appl. Phys. Lett. 81 (2002) 2929. [5] J.S. Lewis, M.S. Weaver, IEEE J. Quantum Electron. 10 (2004) 45. [6] M.D. Groner, S.M. George, R.S. McLean, P.F. Carcia, Appl. Phys. Lett. 88 (2006) 051907. [7] E. Langereis, M. Creatore, S.B.S. Heil, M.C.M. van de Sanden, W.M.M. Kessels, Appl. Phys. Lett. 89 (2006) 081915. [8] Y.G. Tropsha, N.G. Harvey, J. Phys. Chem., B 101 (1997) 2259. [9] A.G. Erlat, R.J. Spontak, R.P. Clarke, T.C. Robinson, P.D. Haaland, Y. Tropsha, N.G. Harvey, E.A. Vogler, J. Phys. Chem., B 103 (1999) 6047. [10] N. Inagaki, V. Cech, K. Narushima, Y. Takechi, J. Appl. Polym. Sci. 104 (2007) 915.
[11] Y.S. Jung, J.Y. Seo, D.W. Lee, D.Y. Jeon, Thin Solid Films 445 (2003) 63. [12] B. Yaglioglu, Y.-J. Huang, H.-Y. Yeom, D.C. Paine, Thin Solid Films 496 (2006) 89. [13] N. Ito, Y. Sato, P.K. Song, A. Kaijio, K. Inoue, Y. Shigesato, Thin Solid Films 496 (2006) 99. [14] S.K. Park, J.I. Han, D.G. Moon, W.K. Kim, Jpn. J. Appl. Phys. 42 (2003) 623. [15] Y. Leterrier, L. Médico, F. Demarco, J.-A.E. Månson, U. Betz, M.F. Escola, M. Kharrazi Olsson, F. Atamny, Thin Solid Films 460 (2004) 156. [16] Y. Leterrier, L. Boogh, J. Andersons, J.A.E. Mansons, J. Polym. Sci., B 323 (1999) 63. [17] D.R. Cairns, R.P. Witte, D.K. Sparacin, S.M. Sachsman, Appl. Phys. Lett. 76 (2000) 1425. [18] J. Yun, S. Lee, Y. Jeong, H.-R. Lee, J.-D. Kwon, G.-H. Lee, Jpn. J. Appl. Phys. 48 (2009) 055503. [19] H. Gleskova, S. Wagner, Z. Suo, Appl. Phys. Lett. 75 (1999) 3011. [20] D.S. Wuu, W.C. Lo, C.C. Chiang, H.B. Lin, L.S. Chang, R.H. Horng, C.L. Huang, Y.J. Gao, Surf. Coat. Technol. 197 (2005) 253. [21] J.F. Silvain, A. Veyrat, J.J. Ehrhardt, Thin Solid Films 221 (1992) 114. [22] F. Milde, K. Goedicke, M. Fahland, Thin Solid Films 279 (1996) 169. [23] C.H. Bichler, H.-C. Langowski, U. Moosheimer, B. Seifert, J. Adhes. Sci. Technol. 11 (1997) 233. [24] B. Gupta, J. Hilborn, C.H. Hollenstein, C.J.G. Plummer, R. Houriet, N. Xanthopoulos, J. Appl. Phys. Sci. 78 (2000) 1083. [25] I. Tepermeister, H.H. Sawin, J. Vac. Sci. Technol., A, Vac. Surf. Films 10 (1992) 3149. [26] D.S. Wuu, T.N. Chen, C.C. Wu, C.C. Chiang, Y.P. Chen, R.H. Horng, F.S. Juang, Chem. Vap. Depos. 12 (2006) 220. [27] A.S. da Silva Sobrinho, N. Schühler, J.E. Klemberg-Sapieha, M.R. Wertheimer, M. Andrews, S.C. Gujrathi, J. Vac. Sci. Technol., A, Vac. Surf. Films 16 (1998) 2021. [28] A.S. da Silva Sobrinho, M. Latréche, G. Czeremuszkin, J.E. Klemberg-Sapieha, M.R. Wertheimer, J. Vac. Sci. Technol., A, Vac. Surf. Films 16 (1998) 3190. [29] 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.