Open Access
Add to BibSonomy
Abstract
Nine specimens are prepared on the basis of the L9(33) orthogonal array design to evaluate the optical and electrical properties, morphology, and microstructure of a-IGZO/Ag/a-IGZO (IAI) triple-layer composite films deposited on curved glass substrates with a different radius of curvature efficiently. The experiments are arranged for the change in the three controlling factors, namely the IGZO and Ag thicknesses, and the substrate’s radius of curvature. The radius of curvature has the highest contribution for the RMS surface roughness (SRq) and the mean particle size (PS). The carrier mobility (CM) and carrier concentration (CC) are proportional to each, irrespective of the controlling factors in this study. An increase in the radius of curvature can result in the increase of SRq and PS, and therefore brings in a surface scattering effect that can cause the reduction of CM as well as the enhancement of resistivity (R). In addition, a sufficiently large radius of curvature can elevate the transmittance at 550 nm and Haack’s figure of merit (FOM) effectively, but it can also lower the reflectance for blue, green and red. Via the carrier injections, increasing the Ag thickness can elevate the carrier mobility and concentration significantly. The reflectance for blue, green, and red are also risen by increasing the Ag thickness. As a result of the Burstein-Moss effect, carrier concentration and optical bandgap are elevated by increasing the IGZO thickness. Additionally, the reflectance for blue, green, and red are also increased. Appropriate choices in the IGZO and Ag thicknesses and the radius of curvature can obtain the transmittance >80 %, and elevate the FOM significantly.
© 2019 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Electronic devices such as smart windows, touch screens, and virtual reality glasses much rely on the transparent conducting thin films to achieve a high transparency and conductance [1–3]. Such films are commonly realized by the transparent conducting oxides (TCOs), e.g., fluorine-doped tin oxide (FTO), tin-doped indium oxide (ITO), and aluminum-doped zinc oxide (AZO) [4–6], on account of their superior optical and electrical properties. Amorphous indium gallium zinc oxide (a-IGZO) is a particularly promising material for device applications due to its high transparency in the visible range (380 - 750 nm), wide optical bandgap (> 3.5 eV) and its excellent electrical properties, such as its high electron mobility compared to the conventional amorphous silicon (a-Si) [7–10]. In addition, the a-IGZO possesses the adjustable microstructure, and therefore it can avoid the defects, such as a fracture or bending when it is deposited on a flexible substrate. Besides, it also allows the fabrication process to be implemented at the room temperature which implies a high potential of being developed as the flexible large-area electronic devices with excellent opto-electrical performance and mechanical reliability, according to studies of Zhou et al [11]. and Nomura et al [12].
Curved panel displays are a high potential technology due to their wide field of view (FOV) feature, which can provide the customers with an immersive experience and improve the perception of vision. A previous study by Chien et al. [13] indicates that a-IGZO is a material fully compatible with flexible substrates in a thin film transistor (TFT). It is able to retain its function for a bending radius up to 40 mm.
Several studies have shown that the opto-electrical performance of films is further enhanced and the energy consumption of devices is lowered by the TCO/metal/TCO (TMT) multilayer structures [14–16]. Boscarino et al. [14] reported that the Ag interlayer in a TCO/Ag/TCO structures dramatically reduces the resistivity of the thin films while decreasing the average transmittance in the visible light (380-750 nm) slightly. Ravichandran et al. [15] reported that a TMT system has its optical properties better than those of a metal/TCO double layer structure. Since the top TCO layer plays the role as an anti-reflective coating layer, and therefore can suppress the reflection from the embedded metal layer and thus increase the transmittance. In the study of Chen et al. [17], it shows that the variation of the TCO layers for the TMT thin film can affect the optical properties significantly. Besides, the symmetric TMT structure thin film, in which the thicknesses of two TCO layers are at the same value, possesses the strongest anti-reflection effect, and therefore achieve the high transmittance in the visible light regions.
The optical and electrical properties of the TMT transparent conductive thin films are considerably dependent on the thickness of the embedded metal layer [18–20]. Among all metals, silver (Ag) is the most promising material used as an interlayer due to its highest conductance. With a thin and uniform Ag layer, TMT systems can have high optical and electrical performances in the transparent conductive thin film transistors.
In the present study, 9 kinds of specimen have been prepared by varying the a-IGZO and Ag thicknesses and the radius of curvature of glass substrate via the three-level and three factorial (L9(33)) design. By the design of experiments on the basis of the Taguchi method, an efficient way is implemented to establish the correlations of the three controlling factors with the mean particle size, the morphology, the opto-electrical properties, and the optical bandgap of specimens. The effects of the controlling factors on these parameters mentioned above are evaluated and explained from the experimental evidences. A figure of merit (FOM) function proposed by Haacke [21] is then applied to evaluate the efficiency of the opto-electrical conversion in the transparent conductive thin films.
2. Experimental
2.1 Methodology
The L9(33) orthogonal array based on the Taguchi method were used to set up the design of experiments, and it provided us a way to conduct the experiments more efficiently and effectively than the traditional full-factors experiment design [22]. The L9(33) orthogonal array can be used for the experiment of both 3 factors and 4 factors with 3 levels [23,24]. The analysis of variance (ANOVA) is applied to evaluate the contribution rates of three controlling factors for the optical, electrical, and microstructural properties. The residual error of each analysis is also calculated to present the effect of noise factor (< 15%). The optimal settings for the optical, electrical, and microstructural properties were determined from the analysis in the Taguchi method. This method was implemented to determine the individual interaction of factors in a 95 % confidence level [25].
2.2 Specimen preparation
In the present study, the IAI composite thin films, as shown in Fig. 1(a), were deposited in a co-sputtering system (ACS-4000-C3, ULVAC) for the IGZO layer by a radio frequency power source of 70W, and for the Ag interlayer with a direct current power source of 50W. Prior to the set-up of experiment, the deposition rates of Ag and IGZO on the substrate with different radius of curvature were measured and calculated. It was found that the variations of substrate’s radius of curvature had no significant effect on the thin film deposition rate. Thus, the thicknesses of a-IGZO and Ag films were controlled simply by the deposition time, and this method of controlling the TMT thin film thickness was also adapted in the previous study [26]. Both a-IGZO and Ag were deposited at room temperature (25℃) with an Argon flow of 30 sccm and at a rotational speed of 15 rpm to ensure the uniformity of the thin film. The bare quartz is used as the substrate material due to its high transparency (>90 %) and insulating features. Besides, the quartz possesses the high strength and durability, and thus allows the manufacturing of the curved substrate with the smooth surface. Figures 1(b)-(d) show the schematic plots of the curved glass substrates (25 mm × 25 mm) with 35, 45 and 55 mm, respectively as the radius of curvature of the top surface. The a-IGZO is a transparent amorphous oxide, which is compatible with the flexible and curved substrate due to its high flexibility and uniformity features [27]. Thus, the crack-free thin film layer with the high mechanical durability deposited on the curved substrate can be achieved via the use of a-IGZO. Three controlling factors, namely substrate’s radius of curvature, Ag thickness, and IGZO thickness were set to be the three-level and three-factorial designs as shown in Table 1. Nine specimens were prepared on the basis of the L9 (33) Taguchi orthogonal table as shown in Table 2. For the IAI structure specimens, the top and bottom a-IGZO layers were prepared with the same thickness. An IGZO target (In:Ga:Zn:O = 1:1:1:1, 99.999% pure) was used for the oxide layers and pure silver (99.999% pure) was deposited as the metal interlayer. Before the film depositions, glass substrates were cleaned in ethanol for 6 minutes, in acetone for 6 minutes, and then cleaned in the deionized water for 6 minutes by an ultrasonic cleaner. Besides, the IGZO and Ag targets were pre-sputtering for 5 minutes to remove contamination substances on the surface.
Fig. 1. (a) Triple-layer thin films deposited on a curved glass substrate with its top surface having the radius of curvature of (b) 35 mm, (c) 45 mm, (d) 55 mm.
Table 1. Design of the controlling factors.
Table 2. L9 (33) orthogonal table.
2.3 Characterization
An UV/Visible/NIR spectrophotometer (HITACHI U4100, Japan) was applied to measure transmittance (T), absorptance (A), and reflectance (Re) spectra in the light wavelengths between 300 and 1000 nm. The Multipurpose X-ray Diffractometer (XRD) (Bruker AXS, model D8 DISCOVER with GADDS, Germany) was used for the microstructure analysis by using its grazing incidence diffraction (GID) identification function in the Cu Kα radiation. The parallel beam x-ray diffraction analysis of GID-XRD can adjust the distance between the x-ray source, sample, and the detectors. Therefore, the flexibility of this system enables the measurement to be implemented for the samples with different shape and size. It overcomes the restrictions of conventional XRD measurement such as non-flat specimen and rough specimen. The parallel beam x-ray system eliminates the errors associated with sample displacement, transparency, and axial divergence [28]. In parallel beam x-ray system, the Gobel mirrors convert the divergent x-ray into a parallel beam, and the geometry of Gobel mirror is prepared to be a parabolic shape [29]. Prior to the measurement, the Z height and rocking curve adjustments are performed to ensure the XRD signal is detected. Surface morphology and the root mean square (RMS) roughness (SRq) of specimen were determined using an Atomic Force Microscope (AFM) (Bruker Dimension ICON, Germany). The microstructure and crystallization behaviors in the Ag interlayer and a-IGZO layer were detected and identified using a High-Resolution Transmission Electron Microscopy (HRTEM; JEM-2100F, JEOL, Japan). Samples for cross-sectional TEM images were implemented using a Dual-Beam Focused Ion Beam (DB-FIB) (FEI Nova-200 NanoLab Compatible, America). A Pt/Ink/Pt composite layer was deposited as the protection layer to achieve a damage-free specimen. A Scanning Electron Microscope (SEM; JEOL JSM-7001, Japan) was then used to investigate particle size and pattern formed on specimen surface. Electrical properties, such as carrier mobility and concentration, and resistivity were measured using a Hall Effect Analyzer.
3. Results and discussion
AFM and SEM were used to characterize the surface morphology. The SEM images with a 200000× magnification for these 9 specimens have been obtained. For example, the SEM images 1, 4, 7, as shown in Figs. 2(a)-(c), show the surface particles to be typical granular structure with different size distribution on the top surface. The data of particle size were calculated from the SEM images using an image software package (ImageJ, National Institute of Health, Bethesda, MD). The pixels were converted into the particle sizes according to the scale factor of images, and therefore provided to calculate the mean particle size (PS) of specimen.
Two statistical parameters, skewness (Sk) and kurtosis (Ku) were introduced here to describe the distribution of particle diameters [30]. Ku is a measure of the peakedness or sharpness of the geometrical parameter distributions (surface height, particle size, etc.). A central distribution profile has a Ku value >3 (leptokurtic) whereas the Ku of a comparatively uniform distribution is <3 (platykurtic). Sk denotes the measure of the parameter asymmetry in distribution. Sk may be positive if the distribution has a longer tail at the upper side of the mean plane, and vice versa. A comparatively large positive Sk may indicate the presence of comparatively few spikes in the distribution of parameters.
A histogram was applied to present the mean particle size (PS) distribution from the SEM images. For example, the particle size distributions of the specimens 1, 4, and 7, as shown in Figs. 2(a)-(c), were used to calculate the Sk and Ku values of the particle size distributions.
In order to describe the correlation of parameters, the regression equation is obtained from the polynomial fitting to the experimental data, and the order of equation is determined when the fitting curve possesses the highest Adjusted R-squared. A higher Adjusted R-squared indicates a higher reliability of the fitting curve. In this study, the order of regression equation in the figures are all determined following this rule. Figure 3(a) shows the results that (Ku)PS and (Sk)PS are highly proportional to each other. For the specimens with (Sk)PS >0 and (Ku)PS <3, the particle size distribution is presented to have relatively few particles with large sizes. For the rest specimens with (Sk)PS >0 and (Ku)PS >3, the particle size distribution is presented to have relatively many particles with large sizes. Besides, specimens 1, 3, 4, and 5 show their (Ku)PS value lower than 3, while the other specimens have their value higher than 3. The distributions of PS for the specimens with (Ku)PS <3 are relatively uniform compared to those with (Ku)PS >3. The (Sk)PS values for all specimens except for specimen 1 are found to be >1. Furthermore, the smallest PS and the lowest (Ku)PS and (Sk)PS are formed in specimen 1. This implies that a specimen with larger values of (Ku)PS and (Sk)PS is prepared with a larger particle size. Figure 3(b) shows that the (Sk)PS and (Ku)PS values are varied nonlinearly proportional to each other. Figure 3(c) indicates a positive dependence between SRq and PS. When a larger PS is formed on the specimen’s surface, it can bring in a larger SRq as a result. Besides, it can be noticed that specimens 7, 8, and 9, which have the 55 mm substrate’s radius of curvature, have the comparatively larger SRq and MPS of all specimens.
Fig. 3. (a) Results of the kurtosis and skewness of particle size distribution as a function of mean particle size; (b) plot of skewness vs kurtosis for particle size distribution. (c) plot of particle size vs RMS surface roughness.
In order to clarify the influences of controlling factors on the morphology of specimens, the ANOVA was applied to evaluate their contribution rates on PS and SRq. The results of ANOVA indicate that substrate’s radius of curvature produces the highest contribution rate for PS (71.7%) and SRq (68.9%). Relatively higher contribution rates indicate that the substrate’s radius of curvature becomes the dominant factor for the two morphological parameters; besides, it also provides the explanation why specimens 7, 8, and 9 have the comparatively larger SRq and PS values of all specimens.
The experimental data in Figs. 3 are provided to evaluate the effects of substrate’s radius of curvature on SRq and PS based on the average of the calculations for fixing the radius of curvature but changing the thicknesses of Ag and IGZO through the Taguchi method. For example, the points in Fig. 4 corresponding to 35-mm substrate’s radius of curvature are obtained to be the average values of the SRq and PS data for specimens 1, 2, and 3. This methodology was also applied to tackle the results for the other controlling factors. In this study, the values of residual error (noise) for the structural, optical, and electrical properties are obtained to be always lower than 15%, as shown in Tables 3 and 4, respectively. Therefore, the independent factor effects on the optical, electrical, and microstructural properties have their contribution rates much higher than those of the residual errors. The inaccuracy of measurements and the interactions between or among the controlling factors are the main sources of the residual error, and both of them are thus neglected in this study. Thus, in the following sections, the influence of the factors on the properties will be discussed on the basis of the independent factors, and their interaction effects will be neglected.
Fig. 4. Results of mean RMS surface roughness (SRq) and mean particle size (PS) expressed as a function of substrate’s radius of curvature.
Table 3. Residual errors for particle size, RMS surface roughness, carrier mobility, carrier concentration, resistivity, and optical bandgap.
Table 4. Residual errors for transmittance and reflectance in the blue, green and red light regions.
Figure 4 shows that increasing substrate’s radius of curvature from 35 to 55 mm can bring in a SRq increase from 1.12 to 1.53 nm, and a PS increase from 41.26 to 65.70 nm. The effect of SRq on the electrical and optical properties will be discussed in the later sections.
GID-XRD and HR-TEM characterizations were used to investigate the microstructural and interfacial behaviors of thin films and evaluate their influence on surface morphology. Figure 5 shows the XRD patterns for the 9 specimens, which indicate that all specimens are shown to be mostly an amorphous microstructure in IGZO. A minor peak is observed to be present at 2θ = 38.5° corresponding to the Ag (1,1,1) orientation, and it is comparatively more noticeable for specimens 3, 6, and 9, in which the Ag layer was prepared with the largest designed thickness (=12nm). Previous study [31] shows that the variations of the metal interlayer follow the island-shaped growth mechanism. In the beginning, it starts from an isolated island-shaped structure and then gradually merges together to form a successive layer when the metal interlayer thickness increases to be sufficiently large. Figures 6(a-(1))-(c-(1)) and Figures 6(a-(2))-(c-(2)) are the cross-sectional TEM images for a relatively larger region and their magnification for a smaller region of the specimens with the designed Ag thickness to be 8, 10, and 12 nm, respectively. The images show that the continuity of the Ag film with its two adjacent IGZO layers is strongly dependent on its thickness. In present study, a significant transformation of Ag microstructure occurs when the Ag thickness is around 10 nm, and the discontinuity of Ag interlayer exists until its thickness is up to be 10 nm.
Fig. 6. Cross-sectional TEM images for specimens (a) 4, (b) 5, and (c) 6 and their SAED analyses for Ag and IGZO layer, respectively.
The selected area electron diffraction (SAED) patterns, as shown in Figs. 6(a-(3))-(c-(3)) indicate the polycrystalline structure of the Ag interlayer. In addition, diffuse diffraction rings of IGZO are also shown in Figs. 6(a-(4))-(c-(4)), and they represent to be nearly the amorphous phase. The SAED patterns images are used to calculate the d-spacing between two adjacent lattice planes via the image processing technique, and determine their associated Miller induces. It is found that the Ag layer growing along the (1,1,1), (2,0,0), (2,2,0), (3,1,1), and (2,2,2) orientations have the d-spacing values of 2.359, 2.044, 1.445, 1.231, 1.1796 Å, respectively.
A Hall effect analyzer was used to measure the carrier mobility (CM), carrier concentration (CC), resistivity (R) for all 9 specimens operating at the room temperature.
Figure 7 shows a high positive correlation between carrier concentration and carrier mobility exists, and the carrier mobility increases from 4.31 to 10.90 cm2/Vs as carrier concentration increases from 3.57×1021 cm−3 to 9.73×1021 cm−3. Owing to the random distributions of Ga+ and Zn2+ ions in the crystal structure, the carrier transport in both amorphous and crystalline IGZOs is found to be governed by percolation conduction over the distribution of potential barriers around the conduction band edge, and such barrier is overcame when the carrier concentration exceeds 3×1018 cm−3 [7,12]. The carrier mobility is elevated as the carrier concentration increases because the carrier concentrations in this study are larger than this critical value.
Fig. 7. The correlation among Hall carrier mobility, resistivity, and carrier concentration of the IAI thin films.
A decrease in resistivity of the IAI thin films with the increasing IGZO and Ag thicknesses can be explained by the formula defined for electrical resistivity (R):
where R is the resistivity; CC is the carrier concentration; e denotes the electron charge, and CM is the carrier mobility. According to Eq.(1), the resistivity of film is inversely proportional to the product solution of carrier mobility and concentration. It implies that the changes in carrier mobility and carrier concentration would lead to the inverse trend in resistivity, as the results are shown in Fig. 7.The scattering mechanisms has been reported to be the dominant mechanism for the electron transport in semiconductor materials [1,6], and their contribution on carrier mobility is given by the Matthiessen’s rule [6]:
The absorption coefficient (α) is calculated from the measured spectral transmittance via the formula [32] :
where t is the total thickness of the composite film, and T denotes the mean transmittance. The value of optical bandgap for the thin films are determined using Tauc plot by extrapolating the linear portion of α2 to intersect the hν axis. The relation of absorption coefficient (α) and optical bandgap energy (Eg) can be described as follow [33].Fig. 8. The absorption coefficient of IAI thin films on curved quartz slides as a function of bandgap energy.
Figure 9 shows the results of mean resistivity, optical bandgap and SRq evaluated for the substrates with the three radii of curvature. A bandgap widening, namely a blue-shifting of the absorption edge, of 0.04 eV occurs when the substrate’s radius of curvature decreases from 55 to 35 mm. The decrease in substrate’s radius of curvature can be interpreted equivalent to an increase in the deposition angle during deposition because the shape of substrate becomes more curved. In the previous study [34], the bandgap widening effect was observed via the increase in the deposition angle in the deposition of TCO films. It indicates that a high deposition angle can induce the rise of strain between substrate and thin film, thus leading to an increase in optical bandgap. An increase in the resistivity from 2.01× 10−4 to 2.17× 10−4 Ω·cm is observed when the optical bandgap decreases from 3.56 to 3.52 eV, and the surface roughness increases from 1.12 to 1.53 nm. It can be noticed that an increase in SRq along with a decrease in optical bandgap result in a noticeable decline of resistivity. Surface and grain boundary scattering effects are produced in association with the surface morphology of specimen. However, an increase in particle size, and thus SRq caused by increasing substrate’s radius of curvature can bring in a reduction of the grain boundary scattering of thin films. Hence, it is believed that the effect of surface scattering is stronger than that of the grain boundary scattering in this study.
Fig. 9. Mean resistivity as a function of mean SRq and mean optical bandgap with the variation of substrate’s radius of curvature.
Generally, an increase in surface roughness can enhance the resistance of carrier transport in the thin film surface due to the surface scattering, and hence lead to the rise of resistivity as a result. For the extrinsic materials in this study, the number of electrons lying in the conduction band are unequal to the one of holes in the valence band, and the Fermi level is usually not located at the middle of the bandgap. Therefore, the correlation between optical bandgap and resistivity may be not always proportional to each other [35].
The results of ANOVA show that the variations of Ag thickness and IGZO thicknesses have the contribution rates of 80.26% and 15.2% respectively on the resistivity. Several studies [18,20,36] show that Schottky carrier injections occur in the TCO/Metal/TCO (TMT) thin films when the metal and TCOs layers are brought into contacts, due to a decrease in effective Schottky barrier height and an increase in the probability of tunneling at the junctions induced by the inserted metal interlayer. The work function is the energy difference between the vacuum and Fermi levels. When the metal and TCOs material are connected together, the electrons will flow from the low work function metal to the high work function TCOs material. Reported works [36,37] indicate that the work function of Ag (Φ = 4.2 eV) is lower than that of IGZO (Φ = 4.7 eV), so electrons are easy to inject from Ag to IGZO like that there is no barrier for electron flow. In other words, an Ohmic contact, i.e. the contact with a pretty small resistance, was formed at the junction of Ag and IGZO. The significant increase in carrier concentration by increasing the Ag thickness is thus explained reasonably. Figure 10(a) shows that the optical bandgap decreases while the mean carrier concentration is risen by increasing the Ag thickness. The injected electrons occupy the conduction band and cause the bandgap narrowing of thin films [18]. Figure 10(b) shows the dependences of the mean carrier mobility, carrier concentration and resistivity on the Ag thickness. The mean carrier concentration and mobility are varied to be proportional to each other; however, the resistivity is presented to have the trend opposite to that of carrier concentration and mobility. Figure 10(b) shows that the Ag thickness increase from 8 nm to 12 nm can lead to the rise of carrier mobility from 4.93 to 9.69 cm2/Vs. The interface regions of IGZO/Ag and Ag/IGZO are applied to account for the interface scattering of the whole thin film volume. When the thickness of Ag increases steadily, the ratio of the interface regions to the whole composite film is gradually reduced, which ultimately leads to a decline of interface scattering and an increase of carrier mobility.
Fig. 10. (a) Mean optical bandgap and mean carrier concentration; and (b) mean resistivity, carrier mobility, and carrier concentration expressed as a function of Ag thickness.
Figure 11(a) shows that both the mean optical bandgap and mean carrier concentration are elevated by increasing the IGZO thickness, demonstrating a result exactly opposite to the tendency shown in Fig. 10(a). When the carrier concentration exceeds the Mott critical density (∼1019−1020 cm−3) [38], the bandgap widening effect, namely the red-shifting of absorption edge caused by increasing the IGZO thickness, can be ascribed to the Burstein-Moss effect (BM effect) [39]. For heavily doping TCOs, such as the a-IGZO, the BM effect is presented because the lower energy levels in the conduction band are filled by the electrons of dopants. Therefore, it requires more energy to promote electrons from the valence band to the conduction band, and hence brings in the increase of optical bandgap as a result of band filling [40]. As Fig. 11(b) shows, mean carrier mobility and concentration are elevated and resistivity is lowered by increasing the IGZO thickness.
Fig. 11. (a) Mean optical bandgap and mean carrier concentration as a function of IGZO thickness; (b) mean resistivity, carrier mobility, and carrier concentration as a function of IGZO thickness.
Optical properties, such as transmittance (T) and reflectance (Re) for these 9 specimens in the wavelength range of 300 to 1000 nm have been observed and shown in Figs. 12(a) and (b), respectively. The transmittances arise at the wavelengths of green, yellow, orange, and red for most of these 9 specimens are relatively higher compared to those for the violet, blue and near infrared regions. The reflectance corresponding to the regions of these four colors with relatively higher transmittance are conversely lower.
Fig. 12. (a) The transmittance, and (b) reflectance of all 9 specimens in a wavelength region of 300 to 1000 nm.
The data in Figs. 12(a) and (b) are provided to determine the variations of reflectance (%) and transmittance (%) with the three controlling factors for different colors. Figures 13(a) and (b) show that the mean values of reflectance (Re) and transmittance (T) for the representative wavelengths of blue, green, and red are evaluated as a function of the three controlling factors. The values of reflectance are elevated by increasing both the Ag and IGZO thicknesses, but they are lowered by increasing the substrate’s radius of curvature in the wavelengths of 300-1000 nm. (Re)red ≦ (Re)green ≦ (Re)blue is always valid for the three controlling factors operating at different level. Furthermore, it can be noticed that the change of reflectance is influenced by the IGZO thickness mostly, and the Ag thickness secondly. (T)Red is elevated by increasing these three controlling factors. (T)blue is always lower than (T)Red and (T)Green, irrespective of the controlling factors and their levels.
Fig. 13. The mean reflectance and transmittance as a function of (a) substrate’s radius of curvature; (b) Ag thickness; (c) IGZO thickness.
In order to determine the optimal parameters, which is able to express the best opto-electrical conversion efficiency with the lowest power consumption and the highest efficiency in the transparent conductive electrodes, a Haacke’s figure of merit (FOM) is applied to evaluate [21]:
where T is the transmittance for a typical wavelength of 550 nm, and Rsh denotes the sheet resistance of thin films. The values of FOM for the 9 specimens are shown in Table 5. It shows that specimen 9 has the highest FOM while specimen 1 has the lowest FOM value of these 9 specimens. In our study, specimens 4, 5, 6, 7, 8, and 9, show a relatively higher FOM compared to the work by Chee et al. [41], where the IGZO/Ag/IGZO specimens were prepared with the top and bottom IGZO thicknesses were 65 and 25 nm, respectively, and the Ag thickness were 9, 14, and 36 nm, respectively. This result can be ascribed to two reasons: (1) in this study, specimens prepared with 45 and 55 mm as substrate’s radius of curvature show the average transmittance at the wavelength of 550 nm higher than 80 %, which is also higher than the reported results [36]; (2) it has been reported that when the Ag thickness is made too large (> 14 nm), it will cause a significant decrease in transmittance at the wavelength of 550 nm, and thus bring in a reduction of FOM even though the resistivity is also reduced [36]. According to the reasons mentioned above, it is believed that the ranges of the controlling factors in our study are arranged appropriately. The FOM is efficiently enhanced by adopting the 45 and 55 mm as substrate’s radius of curvature, although specimen’s resistivity is slightly elevated. The Ag thickness of 12 nm along with the IGZO thickness of 30 nm can have the highest T at 550 nm as well as the highest FOM.Table 5. Optical and electrical properties, FOM and optical bandgap of the 9 specimens.
4. Conclusions
- (1) An increase in the substrate’s radius of curvature results in increases in surface roughness (SRq) and the mean particle size of thin films. A high SRq brings in a surface scattering effect, which can reduce the carrier mobility and enhance the resistivity. Increasing substrate’s radius of curvature to be sufficiently large is helpful to elevate the transmittance at the wavelength of 550 nm, and improve the figure of merit (FOM) efficiently. The reflectance for blue, green, and red are lowered by increasing the radius of curvature.
- (2) The Ag interlayer is found to have discontinuity when its thickness is prepared below 10 nm. As the result of carrier injections, the carrier mobility and carrier concentration are dramatically elevated by increasing the Ag thickness, while the optical bandgap and resistivity are lowered by increasing the Ag thickness.
- (3) As a result of the Burstein-Moss effect, the carrier concentration and the optical bandgap are risen, and the resistivity is conversely lowered by increasing the IGZO thickness. The reflectance for blue, green, and red are elevated by increasing the IGZO thickness.
- (4) Appropriate choices in the IGZO and Ag thicknesses are beneficial to the elevations of transmittance and the FOM. An IGZO thickness of 30 nm along with a Ag thickness of 12 nm can result in the highest transmittance at the wavelength of 550 nm.
Funding
Ministry of Science and Technology, Taiwan (MOST) (106-2221-E-006-099-MY2).
Acknowledgements
This work was financially supported by the Ministry of Science and Technology, Taiwan, R.O.C., under grant 106-2221-E-006-099-MY2.
References
1. H. Liu, V. Avrutin, N. Izyumskaya, Ü. Özgür, and H. Morkoç, “Transparent conducting oxides for electrode applications in light emitting and absorbing devices,” Superlattices Microstruct. 48(5), 458–484 (2010). [CrossRef]
2. A. Yamamoto, S. Nagasawa, H. Yamamoto, and T. Higuchi, “Electrostatic tactile display with thin film slider and its application to tactile telepresentation systems,” IEEE Trans. Vis. Comput. Graphics 12(2), 168–177 (2006). [CrossRef]
3. D. Zhang, K. Ryu, X. Liu, E. Polikarpov, J. Ly, M. E. Tompson, and C. Zhou, “Transparent, conductive, and flexible carbon nanotube films and their application in organic light-emitting diodes,” Nano Lett. 6(9), 1880–1886 (2006). [CrossRef]
4. C. G. Granqvist and A. Hultåker, “Transparent and conducting ITO films: new developments and applications,” Thin Solid Films 411(1), 1–5 (2002). [CrossRef]
5. F. Mei, T. Yuan, and R. Li, “Effects of second-phase particles and elemental distributions of ITO targets on the properties of deposited ITO films,” Ceram. Int. 43(12), 8866–8872 (2017). [CrossRef]
6. S. Calnan and A. Tiwari, “High mobility transparent conducting oxides for thin film solar cells,” Thin Solid Films 518(7), 1839–1849 (2010). [CrossRef]
7. T. Kamiya and H. Hosono, “Material characteristics and applications of transparent amorphous oxide semiconductors,” NPG Asia Mater. 2(1), 15–22 (2010). [CrossRef]
8. K. Nomura, A. Takagi, T. Kamiya, H. Ohta, M. Hirano, and H. Hosono, “Amorphous oxide semiconductors for high-performance flexible thin-film transistors,” Jpn. J. Appl. Phys. 45(5B), 4303–4308 (2006). [CrossRef]
9. T. Kamiya, K. Nomura, and H. Hosono, “Present status of amorphous In–Ga–Zn–O thin-film transistors,” Sci. Technol. Adv. Mater. 11(4), 044305 (2010). [CrossRef]
10. T.-C. Li, C.-F. Han, K.-C. Hsieh, and J.-F. Lin, “Effects of thin titanium and graphene depositions and annealing temperature on electrical, optical, and mechanical properties of IGZO/Ti/graphene/PI specimen,” Ceram. Int. 44(6), 6573–6583 (2018). [CrossRef]
11. X. Zhou, J. Xu, L. Yang, X. Tang, Q. Wei, and Z. Yu, “Amorphous In 2 Ga 2 ZnO 7 films with adjustable structural, electrical and optical properties deposited by magnetron sputtering,” Opt. Mater. Express 5(7), 1628–1634 (2015). [CrossRef]
12. K. Nomura, H. Ohta, A. Takagi, T. Kamiya, M. Hirano, and H. Hosono, “Room-temperature fabrication of transparent flexible thin-film transistors using amorphous oxide semiconductors,” Nature 432(7016), 488–492 (2004). [CrossRef]
13. C.-W. Chien, C.-H. Wu, Y.-T. Tsai, Y.-C. Kung, C.-Y. Lin, P.-C. Hsu, H.-H. Hsieh, C.-C. Wu, Y.-H. Yeh, and C.-M. Leu, “High-performance flexible a-IGZO TFTs adopting stacked electrodes and transparent polyimide-based nanocomposite substrates,” IEEE Trans. Electron Devices 58(5), 1440–1446 (2011). [CrossRef]
14. S. Boscarino, I. Crupi, S. Mirabella, F. Simone, and A. Terrasi, “TCO/Ag/TCO transparent electrodes for solar cells application,” Appl. Phys. A 116(3), 1287–1291 (2014). [CrossRef]
15. K. Ravichandran, K. Subha, A. Manivasaham, M. Sridharan, T. Arun, and C. Ravidhas, “Fabrication of a novel low-cost triple layer system (TaZO/Ag/TaZO) with an enhanced quality factor for transparent electrode applications,” RSC Adv. 6(68), 63314–63324 (2016). [CrossRef]
16. S. Y. Lee, Y. S. Park, and T.-Y. Seong, “Optimized ITO/Ag/ITO multilayers as a current spreading layer to enhance the light output of ultraviolet light-emitting diodes,” J. Alloys Compd. 776, 960–964 (2019). [CrossRef]
17. K.-N. Chen, C.-F. Yang, C.-C. Wu, and Y.-H. Chen, “Development of the α-IGZO/Ag/α-IGZO Triple-Layer Structure Films for the Application of Transparent Electrode,” Materials 10(3), 226 (2017). [CrossRef]
18. H. Han, N. Theodore, and T. Alford, “Improved conductivity and mechanism of carrier transport in zinc oxide with embedded silver layer,” J. Appl. Phys. 103(1), 013708 (2008). [CrossRef]
19. I. Crupi, S. Boscarino, V. Strano, S. Mirabella, F. Simone, and A. Terrasi, “Optimization of ZnO: Al/Ag/ZnO: Al structures for ultra-thin high-performance transparent conductive electrodes,” Thin Solid Films 520(13), 4432–4435 (2012). [CrossRef]
20. J. Leng, Z. Yu, W. Xue, T. Zhang, Y. Jiang, J. Zhang, and D. Zhang, “Influence of Ag thickness on structural, optical, and electrical properties of ZnS/Ag/ZnS multilayers prepared by ion beam assisted deposition,” J. Appl. Phys. 108(7), 073109 (2010). [CrossRef]
21. G. Haacke, “New figure of merit for transparent conductors,” J. Appl. Phys. 47(9), 4086–4089 (1976). [CrossRef]
22. J. R. Philip, “Taguchi techniques for quality engineering,” McGraw-Hill, New York (1996).
23. B. Raju, U. C. Sekhar, and D. Drakshayani, “Optimizing multiple quality characteristics of stereolithography process via Taguchi method-based grey analysis for SL5530 epoxy resin material to enhance part quality,” Procedia Mater. Sci. 5, 2532–2541 (2014). [CrossRef]
24. M. Nalbant, H. Gökkaya, and G. Sur, “Application of Taguchi method in the optimization of cutting parameters for surface roughness in turning,” Mater. Des. 28(4), 1379–1385 (2007). [CrossRef]
25. P. J. Ross and P. J. Ross, Taguchi Techniques for Quality Engineering: Loss Function, Orthogonal Experiments, Parameter and Tolerance Design (McGraw-Hill, 1988).
26. X.-Y. Liu, Y.-A. Li, S. Liu, H.-L. Wu, and H.-N. Cui, “ZnO/Cu/ZnO multilayer films: Structure optimization and investigation on photoelectric properties,” Thin Solid Films 520(16), 5372–5377 (2012). [CrossRef]
27. Z. Wang, H. Xu, X. Li, X. Zhang, Y. Liu, and Y. Liu, “Flexible resistive switching memory device based on amorphous InGaZnO film with excellent mechanical endurance,” IEEE Electron Device Lett. 32(10), 1442–1444 (2011). [CrossRef]
28. S. R. Byrn, G. Zografi, and S. Chen, Solid State Properties of Pharmaceutical Materials (Wiley Online Library, 2017).
29. H. Wang, X.-Z. Yuan, and H. Li, PEM Fuel Cell Diagnostic Tools (CRC press, 2011).
30. V. Şenay, S. Özen, S. Pat, and Ş. Korkmaz, “Optical, structural, morphological and compositional characterization of a Co-doped GaAs semiconducting thin film produced by thermionic vacuum arc,” J. Alloys Compd. 663, 829–833 (2016). [CrossRef]
31. N. Ren, J. Zhu, and S. Ban, “Highly transparent conductive ITO/Ag/ITO trilayer films deposited by RF sputtering at room temperature,” AIP Adv. 7(5), 055009 (2017). [CrossRef]
32. D. Mardare, M. Tasca, M. Delibas, and G. Rusu, “On the structural properties and optical transmittance of TiO2 rf sputtered thin films,” Appl. Surf. Sci. 156(1-4), 200–206 (2000). [CrossRef]
33. Z.-N. Ng, K.-Y. Chan, and T. Tohsophon, “Effects of annealing temperature on ZnO and AZO films prepared by sol–gel technique,” Appl. Surf. Sci. 258(24), 9604–9609 (2012). [CrossRef]
34. K. Kim, J. H. Park, H. Kim, J. K. Kim, E. Fred Schubert, and J. Cho, “Energy bandgap variation in oblique angle-deposited indium tin oxide,” Appl. Phys. Lett. 108(4), 041910 (2016). [CrossRef]
35. J. Morrison, Modern Physics: For Scientists and Engineers (Academic Press, 2015).
36. A. Indluru and T. Alford, “Effect of Ag thickness on electrical transport and optical properties of indium tin oxide–Ag–indium tin oxide multilayers,” J. Appl. Phys. 105(12), 123528 (2009). [CrossRef]
37. H. Aziz, Z. D. Popovic, N.-X. Hu, A.-M. Hor, and G. Xu, “Degradation mechanism of small molecule-based organic light-emitting devices,” Science 283(5409), 1900–1902 (1999). [CrossRef]
38. J. Lu, S. Fujita, T. Kawaharamura, H. Nishinaka, Y. Kamada, T. Ohshima, Z. Ye, Y. Zeng, Y. Zhang, and L. Zhu, “Carrier concentration dependence of band gap shift in n-type ZnO: Al films,” J. Appl. Phys. 101(8), 083705 (2007). [CrossRef]
39. X. Li, S. Chen, T. Chen, and Y. Liu, “Thickness dependence of optical properties of amorphous indium gallium zinc oxide thin films: effects of free-electrons and quantum confinement,” ECS Solid State Lett. 4(3), P29–P32 (2015). [CrossRef]
40. D. B. Potter, M. J. Powell, I. P. Parkin, and C. J. Carmalt, “Aluminium/gallium, indium/gallium, and aluminium/indium co-doped ZnO thin films deposited via aerosol assisted CVD,” J. Mater. Chem. C 6(3), 588–597 (2018). [CrossRef]
41. K. W. Chee, F. Meng, D. C. Lai, and F. Huang, “Measurement-based optimization and analysis of α-IGZO/Ag/α-IGZO transparent conducting electrodes fabricated using DC magnetron sputter deposition,” Ceram. Int. 44(17), 20939–20946 (2018). [CrossRef]
