- Email: [email protected]
Corrosion resistance analysis of aluminium-doped zinc oxide layers deposited by pulsed magnetron sputtering
TSF-34468; No of Pages 5 Thin Solid Films xxx (2015) xxx–xxx
Contents lists available at ScienceDirect
Thin Solid Films journal homepage: www.elsevier.com/locate/tsf
Corrosion resistance analysis of aluminium-doped zinc oxide layers deposited by pulsed magnetron sputtering E. G-Berasategui a,⁎, R. Bayón a, C. Zubizarreta a, J. Barriga a, R. Barros b, R. Martins b, E. Fortunato b a b
IK4-Tekniker, Research Centre, c/Iñaki Goenaga, 5, 20600 Eibar, Guipuzcoa, Spain CENIMAT/I3N, Departamento de Ciência dos Materiais, FCT, Universidade Nova de Lisboa (UNL) and CEMOP/UNINOVA, 2829-516 Caparica, Portugal
a r t i c l e
i n f o
Article history: Received 12 November 2014 Received in revised form 8 July 2015 Accepted 8 July 2015 Available online xxxx Keywords: Aluminum-doped zinc oxide Thin films Transparent conductive films Direct-current magnetron sputtering Electrochemical impedance spectroscopy Corrosion properties
a b s t r a c t In this paper an exhaustive analysis is performed on the electrochemical corrosion resistance of Al-doped ZnO (AZO) layers deposited on silicon wafers by a DC pulsed magnetron sputtering deposition technique to test layer durability. Pulse frequency of the sputtering source was varied and a detailed study of the electrochemical corrosion response of samples in the presence of a corrosive chloride media (NaCl 0.06 M) was carried out. Electrochemical impedance spectroscopy measurements were performed after reaching a stable value of the open circuit at 2 h, 192 h and 480 h intervals. Correlation of the corrosion resistance properties with the morphology, and the optical and electrical properties was tested. AZO layers with transmission values higher than 84% and resistivity of 6.54 × 10− 4 Ω cm for a deposition process pressure of 3 × 10 − 1 Pa, a sputtering power of 2 kW, a pulse frequency of 100 kHz, with optimum corrosion resistance properties, were obtained. © 2015 Elsevier B.V. All rights reserved.
1. Introduction Lifetime of optoelectronic devices is related to the durability of the transparent conductor front electrodes as they are the most exposed layers in the systems [1]. These front electrodes are currently tin-doped indium oxide layers but indium has been classified as a critical raw material by the European Union [2]. Thus, alternative conductor materials are being investigated. Al-doped ZnO (AZO) is a promising material due to its high transmittance, low electrical resistivity, resource availability, low toxicity, strong resistance to high radiation and thermal and chemical stability. It is an n-type semiconductor material with high conductivity caused by an increase in carrier concentration and in carrier mobility due to the aluminium impurities. Among the different techniques available to deposit AZO layers, DC pulse magnetron sputtering technology has been used as it is optimal for industrial applications allowing deposition of high quality films at large scale with automated processes with no toxic emission or waste [3]. To make AZO front contacts penetrate the market, it is necessary to further investigate the durability of these layers. To test the durability, an exhaustive analysis of the electrochemical corrosion response of AZO layers grown by DC pulse magnetron sputtering deposition as a function of pulse frequency has been performed. The corrosion ⁎ Corresponding author. E-mail address: [email protected] (E. G-Berasategui).
resistance behaviour is correlated with the optical, electrical and morphological properties.
2. Experimental procedure A DC magnetron sputtering industrial equipment (MIDAS 450 manufactured by Ik4-TEKNIKER) with a ZnO: Al2O3, 98:2 wt.% target was used to deposit the AZO thin films. The power density was 2.9 W/cm 2 , which corresponds to 2000 W. The substrates were heated up to 350 °C. The power source was dual output pulsed-DC (5 to 350 kHz) from AEI Pinnacle Plus + 5/5 kW with a duty cycle of 96%. Two different process pulse frequencies were studied, 50 kHz and 100 kHz. Process pressure was fixed at 3 × 10−1 Pa. Layers of 1000 nm thickness were deposited on glass and silicon wafers. The structural characterization of the films deposited on glass was carried out by X-ray diffraction (XRD, PANalytical, model X'Pert Pro) in grazing incidence geometry in 2θ model with Cu Kα line radiation (λ = 1.5406 Å). Surface morphology was examined by scanning electron microscopy (Zeiss Auriga CrossBeam scanning electron microscope/focused ion beam SEM/FIB system) and surface roughness by atomic force microscopy in tapping mode (Asylum Research 3D Stand Alone). The transmission spectra (350–2500 nm) were measured by a Perkin-Elmer Lambda 950 spectrophotometer on glass substrates (using as a reference Spectralon labsphere certified reflectance standard). The electrical properties were evaluated by Hall effect measurements in a Van der Pauw geometry [4].
http://dx.doi.org/10.1016/j.tsf.2015.07.010 0040-6090/© 2015 Elsevier B.V. All rights reserved.
Please cite this article as: E. G-Berasategui, et al., Corrosion resistance analysis of aluminium-doped zinc oxide layers deposited by pulsed magnetron sputtering, Thin Solid Films (2015), http://dx.doi.org/10.1016/j.tsf.2015.07.010
2
E. G-Berasategui et al. / Thin Solid Films xxx (2015) xxx–xxx
3. Results and discussion
Fig. 1. XRD patterns of the AZO films at the two different studied frequencies.
Table 1 Values of the position of the (002) peak, FWHM, grain size and deposition rate. Freq (kHz) 2θ (002) (°) FWHM (rad) Grain size (nm) Deposition rate (nm/m) 50 100
34.37 34.36
0.00384 0.00350
39.40 43.31
13 17
The corrosion behaviour of AZO layers was evaluated using electrochemical impedance spectroscopy (EIS) and polarisation techniques on silicon substrates. The measurements were performed in a three-electrode electrochemical cell using a Ag/AgCl (KCl 3 M) reference electrode (SSC) and a platinum wire counter electrode. Tests were performed at room temperature under aerated conditions on an exposed area of 0.2 cm2 in a standard solution of NaCl 0.06 M. Three immersion times were selected to test the electrochemical evolution of the coatings over time. The first measurement was after 2 h of immersion when the open circuit potential becomes stable. After this, 192 and 480 h of immersion were used to estimate the evolution of the corrosion resistance over longer immersion times. For each immersion time a sinusoidal AC perturbation of ±10 mV of amplitude at a frequency range from 10 kHz to 10 mHz under open circuit potential conditions was applied. After the impedance measurements (480 h immersion), the potentiodynamic polarisation curves were registered for each sample from − 0.4 V to 1.2 V (versus open circuit potential) at a scan rate of 0.5 mV/s. All the potentials are referred to Ag/AgCl electrode (0.207 V vs SHE).
P= 50 kHz
a
Rrms= 23 nm
The X-ray diffraction patterns (Fig. 1) show a strong ZnO (002) peak typical of the hexagonal wurtzite crystalline structure orientated with the c-axis perpendicular to the substrate for both studied frequencies. A weak (004) peak is also observed. The position (2θ) of (002) peak is 34.37° and 34.36° for process frequency of 50 kHz and 100 kHz respectively, smaller than 34.45° corresponding to the ZnO crystal. This shift towards lower diffraction angles shows that the (002) interplanar spacing of these layers is larger than the 002 interplanar space for ZnO crystals because of the increase in the compressive stress of the crystals when doping due to substrate heating [5]. No other phases, such as Al or Al 2 O 3 were detected in these films. So, Al 3 + ions substitute Zn 2 + ions into the ZnO hexagonal wurtzite structure. Al ions may also occupy the interstitial sites of ZnO or Al segregates to the non-crystalline region in the grain boundaries and forms Al–O bonds. The full width at half maximum (FWHM) of the XRD peaks depends on the crystalline quality of each grain and on the distribution of grain orientation [6]. Table 1 shows the analysis of the FWHM for the (002) XRD peak and the derived grain size through the Scherrer's equation [7]. The film grown at 100 kHz presents the highest crystallinity associated with the lowest FWHM value and the highest grain size of 43.31 nm [6]. Morphology of the layers changes with pulse frequency as observed in Fig. 2. For 100 kHz process pulse frequency, the grains are uniformly stacked up and compact, so no grain boundaries are visible (Fig. 2b), while for 50 kHz pulse frequency the grains grow loosely and they are clearly notable (Fig. 2a). These changes in morphology slightly affect the roughness values, with roughness varying from 23 nm for 50 kHz pulse frequency to 17 nm for 100 kHz pulse frequency. There are no significant changes in resistivity for both studied samples (Table 2), although the sample grown at 100 kHz shows slightly lower values of resistivity correlated with the highest carrier mobility. The free carriers in the AZO films are mainly provided by oxygen vacancies, zinc interstitials and Al dopants. This correlates with the highest grain size, so the number of grain boundaries decreases facilitating the mobility of the free carriers. There are no significant changes in transmittance spectra (Fig. 3), with an average transmittance in the visible range (400–800 nm) above 84% for both pulse frequencies. The bandgap of AZO films is higher than that of ZnO films (3.37 eV). The absorption edge shifts towards higher energy with an increase in carrier density due to Burstein–Moss effect [8]. The durability of these layers was tested through corrosion resistance analysis. Three electrochemical impedance measurements were registered after 2, 192 and 480 h of immersion. Electrochemical impedance spectroscopy (EIS) is a non-destructive technique that allows
P= 100 kHz Rrms= 17nm
b
Fig. 2. SEM micrograph of AZO films for the two studied pulse frequencies.
Please cite this article as: E. G-Berasategui, et al., Corrosion resistance analysis of aluminium-doped zinc oxide layers deposited by pulsed magnetron sputtering, Thin Solid Films (2015), http://dx.doi.org/10.1016/j.tsf.2015.07.010
E. G-Berasategui et al. / Thin Solid Films xxx (2015) xxx–xxx Table 2 Hall mobility (μ), carrier concentration (n) and resistivity values (ρ). Frequency (kHz)
Thickness (nm)
μ (cm2/Vs)
n (cm3)
ρ (Ω cm)
50 100
1050 1090
18.7 20.7
5.05 × 1020 4.6 × 1020
6.62 × 10−4 6.54 × 10−4
Fig. 3. Optical transmittance and energy band gap values of AZO films.
testing the corrosion resistance of a coating immersed into an electrolyte over time. Through modelling of the electrochemical experimental data, values of resistance and capacitance for the coating can be determined. The modelling procedure uses equivalent electrical circuits composed of resistors and capacitors with the same frequency response as the electrochemical reactions to represent the electrochemical behaviour at the coating/electrolyte and the substrate/electrolyte interfaces [9]. Fig. 4 shows the experimental Bode diagrams for the AZO coatings deposited at 50 kHz and 100 kHz pulse frequencies after 2, 192 and 480 h of immersion times. The three Bode diagrams show two time constants represented by two changes in the slope of log f-log|Z| for both samples in all the immersion times. The presence of two time constants suggests the contribution of the silicon substrate to the electrochemical response of the systems. This is due to the presence of defects or discontinuities in the coating structure. For these AZO layers, these defects are mainly associated with the presence of grain boundaries. To determine quantitative values of kinetics electrochemical parameters, the experimental data were analyzed using an equivalent circuit represented in Fig. 5a [10]. This circuit,
3
composed of two pairs of CPE/R elements combined in parallel, is frequently employed to simulate the behaviour of PVD coatings in aqueous media [9]. R1 is related to the resistance across the coating defects and CPE1 represents the capacitance of the coating (interface coating/electrolyte). In the second sub-circuit, R2 represents the corrosion resistance at the coating pores/defects bottom and CPE 2 the capacitance in the interface electrolyte/silicon substrate. Rs represents the resistance of the electrolyte. Fig. 5b presents the corrosion resistance values obtained after simulation of the experimental data for both coatings. In this figure, the value of R (polarisation resistance) is the sum of the resistance across the coating defects (R1) and the resistance at the coating pores/defects bottom (R2). Both samples present very high corrosion resistance values (MΩ). The total corrosion resistance of the coating deposited at 100 kHz is higher than the one measured for the coating deposited at 50 kHz for all immersion times. This is attributed to the morphology of the layers since the sample grown at 100 kHz shows bigger grain sizes reducing the number of grain boundaries and therefore the discontinuities in the layer structure. A reduction in corrosion resistance when increasing exposure time to the corrosive media is observed in all cases. For the 100 kHz coating, total corrosion resistance is reduced by 88% from 2 h to 480 h. The reduction of corrosion resistance in case of 50 kHz coating is 81%. Looking at the coatings capacitive response, it was found that CPE 1 (capacitance of the coating) after 480 h of immersion is lower in case of the AZO deposited at 100 kHz (0.26 × 10 − 6 F) whereas in case of the coating deposited at 50 kHz this value is 17.3 × 10 − 6 F. Lower values of CPE1 imply a better capacitive response. From the CPE and R values, it is obtained that higher deposition frequencies improve corrosion resistance properties of the layers. After 480 h of immersion, a polarisation curve was performed on both AZO coatings (Fig. 6). Polarisation curves allow evaluating the stability of the passive film for wide range of anodic potentials determining the kinetic parameters such as corrosion current or polarisation resistance. The coating deposited at the highest frequency exhibits a lower current density in the anodic branch of the curve implying higher corrosion resistance and electrochemical stability. Table 3 summarizes the resulting kinetic parameters after the Tafel analysis [11]. Corrosion currents were in both cases very low, around 10− 9 A. This implies a very good corrosion response and elevated polarisation resistance (R p): 204 K Ω for coating deposited at 50 kHz and 314 K Ω for coating deposited at 100 kHz. When the potential increases to anodic values, the current in both cases tends to increase. At 1.2 V, the current value in both cases is around 10−5 A, which means that the coatings experience a soft dissolution rate when they are subjected to high anodic potentials. This effect
Fig. 4. Bode analysis as a function of the immersion time.
Please cite this article as: E. G-Berasategui, et al., Corrosion resistance analysis of aluminium-doped zinc oxide layers deposited by pulsed magnetron sputtering, Thin Solid Films (2015), http://dx.doi.org/10.1016/j.tsf.2015.07.010
4
E. G-Berasategui et al. / Thin Solid Films xxx (2015) xxx–xxx
Fig. 5. Equivalent circuit used to simulate the EIS experimental data and corrosion resistance values derived from the analysis as a function of the immersion time.
(100 kHz), AZO layer looks degraded with respect to Fig. 2. In the left picture (50 kHz), the AZO layer has peeled off after polarisation tests. 4. Conclusions
Fig. 6. Polarisation curves of AZO coatings developed at different pulse frequencies.
is more pronounced in case of the coating deposited at 50 kHz that shows higher corrosion current when the potential applied reaches 1.2 V. Coating degradation after being polarized to elevated anodic potentials was correlated to SEM analysis (Fig. 7). In the right picture
AZO layers with optimum optical and electrical properties were deposited by DC pulse magnetron sputtering. The electrochemical corrosion performance of these films were thoroughly analyzed in the presence of a standard corrosive medium of NaCl 0.06 M allowing to determine the influence of the process parameters on the durability of the coatings when exposed to aggressive environments. From periodic EIS measurements it was found that both AZOs exhibit a high corrosion resistance of the order of MΩ. Looking into the influence of the process frequency, it was found that coatings developed at higher frequencies present higher corrosion resistances and better capacitive response. This improved electrochemical behaviour is attributed to their microstructure. When the process is developed at 100 kHz, the coatings present higher crystallinity, higher grain size and lower roughness. These surface and structural properties enhance the corrosion resistance of the surface. It was also observed that when increasing exposure time to the aggressive media, the corrosion resistance and the capacitive response of the coatings decrease. When the samples are submitted to higher potentials (anodic ones), a surface degradation takes place dealing with the coating detachment in case of the samples developed at 50 kHz. Acknowledgements
Table 3 Tafel analysis of the polarisation curves after 480 h of immersion. Frequency (kHz)
Ecorr (V)
Icorr (10−9 A)
Rp (KΩ)
50 100
−0.069 −0.036
120 131
204 314
The authors acknowledge financial support from the European Commission obtained in the IN-LIGHT project (FP7; Grant agreement nr: 314233) and from the Spanish Ministry of Science and Innovation obtained in the project: CSD2008-00023 FUNCOAT.
Fig. 7. SEM image of AZO's surface after polarisation tests.
Please cite this article as: E. G-Berasategui, et al., Corrosion resistance analysis of aluminium-doped zinc oxide layers deposited by pulsed magnetron sputtering, Thin Solid Films (2015), http://dx.doi.org/10.1016/j.tsf.2015.07.010
E. G-Berasategui et al. / Thin Solid Films xxx (2015) xxx–xxx
References [1] T. Sasabayashi, N. Ito, E. Nishimura, M. Kon, P.K. Song, K. Utsumi, A. Kaijo, Y. Shigesato, Comparative study on structure and internal stress in tin-doped indium oxide and indium–zinc oxide films deposited by r.f. magnetron sputtering, Thin Solid Films 445 (2003) 219–223. [2] http://ec.europa.eu/enterprise/policies/raw-materials/critical/index_en.htm. [3] F. Ruske, V. Sittinger, W. Werner, B. Szyszka, K.-U. vanOsten, K. Dietrich, R. Rix, Hydrogen doping of DC sputtered ZnO:Al films from novel target material, Surf. Coat. Technol. 200 (2005) 236–240. [4] L.J. Van der Pauw, A method of measuring the resistivity and Hall coefficient on lamellae of arbitrary shape, Philips Tech. Rev. 20 (1958) 220–224. [5] S. Lin, J. Huang, D. Lii, The effects of r.f. power and substrate temperature on the properties of ZnO films, Surf. Coat. Technol. 176 (2004) 173–181.
5
[6] N. Matsunami, O. Fukuoka, Masato Tazawa, Masao Sataka, Electronic structure modification of ZnO and Al-doped ZnO films by ions, Surf. Coat. Technol. 196 (2005) 50–55. [7] B.E. Warren, X-Ray Diffraction, Dover Publishing, New York, 1990. [8] E. Burstein, Anomalous optical absorption limit in InSb, Phys. Rev. 93 (1954) 632–633. [9] C. Liu, Q. Bi, A. Leyland, A. Matthews, An electrochemical impedance spectroscopy study of the corrosion behaviour of PVD coated steels in 0.5 NaCl aqueous solution: Part I. Establishment of the equivalent circuits for EIS data modelling, Corros. Sci. 45 (2003) 1243–1256. [10] A.J. Bard, L.R. Faulkner, Electrochemical Methods: Fundamentals and Applications, second ed. JohnWiley and Sons Inc., New York, 2001. [11] N.W. Khun, E. Liu, Effect of substrate temperature on corrosion performance of nitrogen doped amorphous carbon thin films in NaCl solution, Thin Solid Films 517 (2009) 4762–4766.
Please cite this article as: E. G-Berasategui, et al., Corrosion resistance analysis of aluminium-doped zinc oxide layers deposited by pulsed magnetron sputtering, Thin Solid Films (2015), http://dx.doi.org/10.1016/j.tsf.2015.07.010
Contents lists available at ScienceDirect
Thin Solid Films journal homepage: www.elsevier.com/locate/tsf
Corrosion resistance analysis of aluminium-doped zinc oxide layers deposited by pulsed magnetron sputtering E. G-Berasategui a,⁎, R. Bayón a, C. Zubizarreta a, J. Barriga a, R. Barros b, R. Martins b, E. Fortunato b a b
IK4-Tekniker, Research Centre, c/Iñaki Goenaga, 5, 20600 Eibar, Guipuzcoa, Spain CENIMAT/I3N, Departamento de Ciência dos Materiais, FCT, Universidade Nova de Lisboa (UNL) and CEMOP/UNINOVA, 2829-516 Caparica, Portugal
a r t i c l e
i n f o
Article history: Received 12 November 2014 Received in revised form 8 July 2015 Accepted 8 July 2015 Available online xxxx Keywords: Aluminum-doped zinc oxide Thin films Transparent conductive films Direct-current magnetron sputtering Electrochemical impedance spectroscopy Corrosion properties
a b s t r a c t In this paper an exhaustive analysis is performed on the electrochemical corrosion resistance of Al-doped ZnO (AZO) layers deposited on silicon wafers by a DC pulsed magnetron sputtering deposition technique to test layer durability. Pulse frequency of the sputtering source was varied and a detailed study of the electrochemical corrosion response of samples in the presence of a corrosive chloride media (NaCl 0.06 M) was carried out. Electrochemical impedance spectroscopy measurements were performed after reaching a stable value of the open circuit at 2 h, 192 h and 480 h intervals. Correlation of the corrosion resistance properties with the morphology, and the optical and electrical properties was tested. AZO layers with transmission values higher than 84% and resistivity of 6.54 × 10− 4 Ω cm for a deposition process pressure of 3 × 10 − 1 Pa, a sputtering power of 2 kW, a pulse frequency of 100 kHz, with optimum corrosion resistance properties, were obtained. © 2015 Elsevier B.V. All rights reserved.
1. Introduction Lifetime of optoelectronic devices is related to the durability of the transparent conductor front electrodes as they are the most exposed layers in the systems [1]. These front electrodes are currently tin-doped indium oxide layers but indium has been classified as a critical raw material by the European Union [2]. Thus, alternative conductor materials are being investigated. Al-doped ZnO (AZO) is a promising material due to its high transmittance, low electrical resistivity, resource availability, low toxicity, strong resistance to high radiation and thermal and chemical stability. It is an n-type semiconductor material with high conductivity caused by an increase in carrier concentration and in carrier mobility due to the aluminium impurities. Among the different techniques available to deposit AZO layers, DC pulse magnetron sputtering technology has been used as it is optimal for industrial applications allowing deposition of high quality films at large scale with automated processes with no toxic emission or waste [3]. To make AZO front contacts penetrate the market, it is necessary to further investigate the durability of these layers. To test the durability, an exhaustive analysis of the electrochemical corrosion response of AZO layers grown by DC pulse magnetron sputtering deposition as a function of pulse frequency has been performed. The corrosion ⁎ Corresponding author. E-mail address: [email protected] (E. G-Berasategui).
resistance behaviour is correlated with the optical, electrical and morphological properties.
2. Experimental procedure A DC magnetron sputtering industrial equipment (MIDAS 450 manufactured by Ik4-TEKNIKER) with a ZnO: Al2O3, 98:2 wt.% target was used to deposit the AZO thin films. The power density was 2.9 W/cm 2 , which corresponds to 2000 W. The substrates were heated up to 350 °C. The power source was dual output pulsed-DC (5 to 350 kHz) from AEI Pinnacle Plus + 5/5 kW with a duty cycle of 96%. Two different process pulse frequencies were studied, 50 kHz and 100 kHz. Process pressure was fixed at 3 × 10−1 Pa. Layers of 1000 nm thickness were deposited on glass and silicon wafers. The structural characterization of the films deposited on glass was carried out by X-ray diffraction (XRD, PANalytical, model X'Pert Pro) in grazing incidence geometry in 2θ model with Cu Kα line radiation (λ = 1.5406 Å). Surface morphology was examined by scanning electron microscopy (Zeiss Auriga CrossBeam scanning electron microscope/focused ion beam SEM/FIB system) and surface roughness by atomic force microscopy in tapping mode (Asylum Research 3D Stand Alone). The transmission spectra (350–2500 nm) were measured by a Perkin-Elmer Lambda 950 spectrophotometer on glass substrates (using as a reference Spectralon labsphere certified reflectance standard). The electrical properties were evaluated by Hall effect measurements in a Van der Pauw geometry [4].
http://dx.doi.org/10.1016/j.tsf.2015.07.010 0040-6090/© 2015 Elsevier B.V. All rights reserved.
Please cite this article as: E. G-Berasategui, et al., Corrosion resistance analysis of aluminium-doped zinc oxide layers deposited by pulsed magnetron sputtering, Thin Solid Films (2015), http://dx.doi.org/10.1016/j.tsf.2015.07.010
2
E. G-Berasategui et al. / Thin Solid Films xxx (2015) xxx–xxx
3. Results and discussion
Fig. 1. XRD patterns of the AZO films at the two different studied frequencies.
Table 1 Values of the position of the (002) peak, FWHM, grain size and deposition rate. Freq (kHz) 2θ (002) (°) FWHM (rad) Grain size (nm) Deposition rate (nm/m) 50 100
34.37 34.36
0.00384 0.00350
39.40 43.31
13 17
The corrosion behaviour of AZO layers was evaluated using electrochemical impedance spectroscopy (EIS) and polarisation techniques on silicon substrates. The measurements were performed in a three-electrode electrochemical cell using a Ag/AgCl (KCl 3 M) reference electrode (SSC) and a platinum wire counter electrode. Tests were performed at room temperature under aerated conditions on an exposed area of 0.2 cm2 in a standard solution of NaCl 0.06 M. Three immersion times were selected to test the electrochemical evolution of the coatings over time. The first measurement was after 2 h of immersion when the open circuit potential becomes stable. After this, 192 and 480 h of immersion were used to estimate the evolution of the corrosion resistance over longer immersion times. For each immersion time a sinusoidal AC perturbation of ±10 mV of amplitude at a frequency range from 10 kHz to 10 mHz under open circuit potential conditions was applied. After the impedance measurements (480 h immersion), the potentiodynamic polarisation curves were registered for each sample from − 0.4 V to 1.2 V (versus open circuit potential) at a scan rate of 0.5 mV/s. All the potentials are referred to Ag/AgCl electrode (0.207 V vs SHE).
P= 50 kHz
a
Rrms= 23 nm
The X-ray diffraction patterns (Fig. 1) show a strong ZnO (002) peak typical of the hexagonal wurtzite crystalline structure orientated with the c-axis perpendicular to the substrate for both studied frequencies. A weak (004) peak is also observed. The position (2θ) of (002) peak is 34.37° and 34.36° for process frequency of 50 kHz and 100 kHz respectively, smaller than 34.45° corresponding to the ZnO crystal. This shift towards lower diffraction angles shows that the (002) interplanar spacing of these layers is larger than the 002 interplanar space for ZnO crystals because of the increase in the compressive stress of the crystals when doping due to substrate heating [5]. No other phases, such as Al or Al 2 O 3 were detected in these films. So, Al 3 + ions substitute Zn 2 + ions into the ZnO hexagonal wurtzite structure. Al ions may also occupy the interstitial sites of ZnO or Al segregates to the non-crystalline region in the grain boundaries and forms Al–O bonds. The full width at half maximum (FWHM) of the XRD peaks depends on the crystalline quality of each grain and on the distribution of grain orientation [6]. Table 1 shows the analysis of the FWHM for the (002) XRD peak and the derived grain size through the Scherrer's equation [7]. The film grown at 100 kHz presents the highest crystallinity associated with the lowest FWHM value and the highest grain size of 43.31 nm [6]. Morphology of the layers changes with pulse frequency as observed in Fig. 2. For 100 kHz process pulse frequency, the grains are uniformly stacked up and compact, so no grain boundaries are visible (Fig. 2b), while for 50 kHz pulse frequency the grains grow loosely and they are clearly notable (Fig. 2a). These changes in morphology slightly affect the roughness values, with roughness varying from 23 nm for 50 kHz pulse frequency to 17 nm for 100 kHz pulse frequency. There are no significant changes in resistivity for both studied samples (Table 2), although the sample grown at 100 kHz shows slightly lower values of resistivity correlated with the highest carrier mobility. The free carriers in the AZO films are mainly provided by oxygen vacancies, zinc interstitials and Al dopants. This correlates with the highest grain size, so the number of grain boundaries decreases facilitating the mobility of the free carriers. There are no significant changes in transmittance spectra (Fig. 3), with an average transmittance in the visible range (400–800 nm) above 84% for both pulse frequencies. The bandgap of AZO films is higher than that of ZnO films (3.37 eV). The absorption edge shifts towards higher energy with an increase in carrier density due to Burstein–Moss effect [8]. The durability of these layers was tested through corrosion resistance analysis. Three electrochemical impedance measurements were registered after 2, 192 and 480 h of immersion. Electrochemical impedance spectroscopy (EIS) is a non-destructive technique that allows
P= 100 kHz Rrms= 17nm
b
Fig. 2. SEM micrograph of AZO films for the two studied pulse frequencies.
Please cite this article as: E. G-Berasategui, et al., Corrosion resistance analysis of aluminium-doped zinc oxide layers deposited by pulsed magnetron sputtering, Thin Solid Films (2015), http://dx.doi.org/10.1016/j.tsf.2015.07.010
E. G-Berasategui et al. / Thin Solid Films xxx (2015) xxx–xxx Table 2 Hall mobility (μ), carrier concentration (n) and resistivity values (ρ). Frequency (kHz)
Thickness (nm)
μ (cm2/Vs)
n (cm3)
ρ (Ω cm)
50 100
1050 1090
18.7 20.7
5.05 × 1020 4.6 × 1020
6.62 × 10−4 6.54 × 10−4
Fig. 3. Optical transmittance and energy band gap values of AZO films.
testing the corrosion resistance of a coating immersed into an electrolyte over time. Through modelling of the electrochemical experimental data, values of resistance and capacitance for the coating can be determined. The modelling procedure uses equivalent electrical circuits composed of resistors and capacitors with the same frequency response as the electrochemical reactions to represent the electrochemical behaviour at the coating/electrolyte and the substrate/electrolyte interfaces [9]. Fig. 4 shows the experimental Bode diagrams for the AZO coatings deposited at 50 kHz and 100 kHz pulse frequencies after 2, 192 and 480 h of immersion times. The three Bode diagrams show two time constants represented by two changes in the slope of log f-log|Z| for both samples in all the immersion times. The presence of two time constants suggests the contribution of the silicon substrate to the electrochemical response of the systems. This is due to the presence of defects or discontinuities in the coating structure. For these AZO layers, these defects are mainly associated with the presence of grain boundaries. To determine quantitative values of kinetics electrochemical parameters, the experimental data were analyzed using an equivalent circuit represented in Fig. 5a [10]. This circuit,
3
composed of two pairs of CPE/R elements combined in parallel, is frequently employed to simulate the behaviour of PVD coatings in aqueous media [9]. R1 is related to the resistance across the coating defects and CPE1 represents the capacitance of the coating (interface coating/electrolyte). In the second sub-circuit, R2 represents the corrosion resistance at the coating pores/defects bottom and CPE 2 the capacitance in the interface electrolyte/silicon substrate. Rs represents the resistance of the electrolyte. Fig. 5b presents the corrosion resistance values obtained after simulation of the experimental data for both coatings. In this figure, the value of R (polarisation resistance) is the sum of the resistance across the coating defects (R1) and the resistance at the coating pores/defects bottom (R2). Both samples present very high corrosion resistance values (MΩ). The total corrosion resistance of the coating deposited at 100 kHz is higher than the one measured for the coating deposited at 50 kHz for all immersion times. This is attributed to the morphology of the layers since the sample grown at 100 kHz shows bigger grain sizes reducing the number of grain boundaries and therefore the discontinuities in the layer structure. A reduction in corrosion resistance when increasing exposure time to the corrosive media is observed in all cases. For the 100 kHz coating, total corrosion resistance is reduced by 88% from 2 h to 480 h. The reduction of corrosion resistance in case of 50 kHz coating is 81%. Looking at the coatings capacitive response, it was found that CPE 1 (capacitance of the coating) after 480 h of immersion is lower in case of the AZO deposited at 100 kHz (0.26 × 10 − 6 F) whereas in case of the coating deposited at 50 kHz this value is 17.3 × 10 − 6 F. Lower values of CPE1 imply a better capacitive response. From the CPE and R values, it is obtained that higher deposition frequencies improve corrosion resistance properties of the layers. After 480 h of immersion, a polarisation curve was performed on both AZO coatings (Fig. 6). Polarisation curves allow evaluating the stability of the passive film for wide range of anodic potentials determining the kinetic parameters such as corrosion current or polarisation resistance. The coating deposited at the highest frequency exhibits a lower current density in the anodic branch of the curve implying higher corrosion resistance and electrochemical stability. Table 3 summarizes the resulting kinetic parameters after the Tafel analysis [11]. Corrosion currents were in both cases very low, around 10− 9 A. This implies a very good corrosion response and elevated polarisation resistance (R p): 204 K Ω for coating deposited at 50 kHz and 314 K Ω for coating deposited at 100 kHz. When the potential increases to anodic values, the current in both cases tends to increase. At 1.2 V, the current value in both cases is around 10−5 A, which means that the coatings experience a soft dissolution rate when they are subjected to high anodic potentials. This effect
Fig. 4. Bode analysis as a function of the immersion time.
Please cite this article as: E. G-Berasategui, et al., Corrosion resistance analysis of aluminium-doped zinc oxide layers deposited by pulsed magnetron sputtering, Thin Solid Films (2015), http://dx.doi.org/10.1016/j.tsf.2015.07.010
4
E. G-Berasategui et al. / Thin Solid Films xxx (2015) xxx–xxx
Fig. 5. Equivalent circuit used to simulate the EIS experimental data and corrosion resistance values derived from the analysis as a function of the immersion time.
(100 kHz), AZO layer looks degraded with respect to Fig. 2. In the left picture (50 kHz), the AZO layer has peeled off after polarisation tests. 4. Conclusions
Fig. 6. Polarisation curves of AZO coatings developed at different pulse frequencies.
is more pronounced in case of the coating deposited at 50 kHz that shows higher corrosion current when the potential applied reaches 1.2 V. Coating degradation after being polarized to elevated anodic potentials was correlated to SEM analysis (Fig. 7). In the right picture
AZO layers with optimum optical and electrical properties were deposited by DC pulse magnetron sputtering. The electrochemical corrosion performance of these films were thoroughly analyzed in the presence of a standard corrosive medium of NaCl 0.06 M allowing to determine the influence of the process parameters on the durability of the coatings when exposed to aggressive environments. From periodic EIS measurements it was found that both AZOs exhibit a high corrosion resistance of the order of MΩ. Looking into the influence of the process frequency, it was found that coatings developed at higher frequencies present higher corrosion resistances and better capacitive response. This improved electrochemical behaviour is attributed to their microstructure. When the process is developed at 100 kHz, the coatings present higher crystallinity, higher grain size and lower roughness. These surface and structural properties enhance the corrosion resistance of the surface. It was also observed that when increasing exposure time to the aggressive media, the corrosion resistance and the capacitive response of the coatings decrease. When the samples are submitted to higher potentials (anodic ones), a surface degradation takes place dealing with the coating detachment in case of the samples developed at 50 kHz. Acknowledgements
Table 3 Tafel analysis of the polarisation curves after 480 h of immersion. Frequency (kHz)
Ecorr (V)
Icorr (10−9 A)
Rp (KΩ)
50 100
−0.069 −0.036
120 131
204 314
The authors acknowledge financial support from the European Commission obtained in the IN-LIGHT project (FP7; Grant agreement nr: 314233) and from the Spanish Ministry of Science and Innovation obtained in the project: CSD2008-00023 FUNCOAT.
Fig. 7. SEM image of AZO's surface after polarisation tests.
Please cite this article as: E. G-Berasategui, et al., Corrosion resistance analysis of aluminium-doped zinc oxide layers deposited by pulsed magnetron sputtering, Thin Solid Films (2015), http://dx.doi.org/10.1016/j.tsf.2015.07.010
E. G-Berasategui et al. / Thin Solid Films xxx (2015) xxx–xxx
References [1] T. Sasabayashi, N. Ito, E. Nishimura, M. Kon, P.K. Song, K. Utsumi, A. Kaijo, Y. Shigesato, Comparative study on structure and internal stress in tin-doped indium oxide and indium–zinc oxide films deposited by r.f. magnetron sputtering, Thin Solid Films 445 (2003) 219–223. [2] http://ec.europa.eu/enterprise/policies/raw-materials/critical/index_en.htm. [3] F. Ruske, V. Sittinger, W. Werner, B. Szyszka, K.-U. vanOsten, K. Dietrich, R. Rix, Hydrogen doping of DC sputtered ZnO:Al films from novel target material, Surf. Coat. Technol. 200 (2005) 236–240. [4] L.J. Van der Pauw, A method of measuring the resistivity and Hall coefficient on lamellae of arbitrary shape, Philips Tech. Rev. 20 (1958) 220–224. [5] S. Lin, J. Huang, D. Lii, The effects of r.f. power and substrate temperature on the properties of ZnO films, Surf. Coat. Technol. 176 (2004) 173–181.
5
[6] N. Matsunami, O. Fukuoka, Masato Tazawa, Masao Sataka, Electronic structure modification of ZnO and Al-doped ZnO films by ions, Surf. Coat. Technol. 196 (2005) 50–55. [7] B.E. Warren, X-Ray Diffraction, Dover Publishing, New York, 1990. [8] E. Burstein, Anomalous optical absorption limit in InSb, Phys. Rev. 93 (1954) 632–633. [9] C. Liu, Q. Bi, A. Leyland, A. Matthews, An electrochemical impedance spectroscopy study of the corrosion behaviour of PVD coated steels in 0.5 NaCl aqueous solution: Part I. Establishment of the equivalent circuits for EIS data modelling, Corros. Sci. 45 (2003) 1243–1256. [10] A.J. Bard, L.R. Faulkner, Electrochemical Methods: Fundamentals and Applications, second ed. JohnWiley and Sons Inc., New York, 2001. [11] N.W. Khun, E. Liu, Effect of substrate temperature on corrosion performance of nitrogen doped amorphous carbon thin films in NaCl solution, Thin Solid Films 517 (2009) 4762–4766.
Please cite this article as: E. G-Berasategui, et al., Corrosion resistance analysis of aluminium-doped zinc oxide layers deposited by pulsed magnetron sputtering, Thin Solid Films (2015), http://dx.doi.org/10.1016/j.tsf.2015.07.010