Radio frequency sputtered Al:ZnO-Ag transparent conductor: A plasmonic nanostructure with enhanced optical and electrical properties Anna Sytchkova, Maria Luisa Grilli, Antonio Rinaldi, Sylvain Vedraine, Philippe Torchio et al. Citation: J. Appl. Phys. 114, 094509 (2013); doi: 10.1063/1.4820266 View online: http://dx.doi.org/10.1063/1.4820266 View Table of Contents: http://jap.aip.org/resource/1/JAPIAU/v114/i9 Published by the AIP Publishing LLC.
Additional information on J. Appl. Phys. Journal Homepage: http://jap.aip.org/ Journal Information: http://jap.aip.org/about/about_the_journal Top downloads: http://jap.aip.org/features/most_downloaded Information for Authors: http://jap.aip.org/authors
JOURNAL OF APPLIED PHYSICS 114, 094509 (2013)
Radio frequency sputtered Al:ZnO-Ag transparent conductor: A plasmonic nanostructure with enhanced optical and electrical properties Anna Sytchkova,1,a) Maria Luisa Grilli,1 Antonio Rinaldi,1 Sylvain Vedraine,2 Philippe Torchio,2 Angela Piegari,1 and Franc¸ois Flory2 1
ENEA Advanced Material Department, via Anguillarese 301, 00123 Rome, Italy Aix-Marseille University, Institut Mat eriaux Micro electronique Nanosciences de Provence-IM2NP, CNRS-UMR 7334, Domaine Universitaire de Saint-J er^ ome, Service 231, 13 397 Marseille Cedex 20, France 2
(Received 30 May 2013; accepted 19 August 2013; published online 4 September 2013) Optimization of metal-based transparent conductors (MTCs) made of silver and aluminium-doped zinc oxide (AZO) prepared by radio-frequency (r.f.) sputtering has been carried out through tuning of metal film properties. The influence of morphology and related plasmonic features of AZO/Ag/ AZO MTCs on their optical and electrical performance is demonstrated and it is shown that the nominal thickness of the silver layer itself is not the most crucial value determining the MTC performance. The MTC performance has been optimized by a search of deposition conditions ensuring fractal-type metal layer formation up to a certain coalescence state that enables full gaining from silver optical properties, including its plasmonic features. For 150 W- and 200 W-deposited silver, MTCs with maximum transmittance as high as 83.6% have been obtained. These coatings have a figure of merit as good as 0.01 X1 and a remarkably wide spectral transparency region: transmittance higher than 70% down to 1200 nm for 200W-samples. Modelling of the MTC coatings is proposed additionally, based on variable angle spectroscopic ellipsometric measurements, which takes into account the variation of the optical properties of silver when deposited in various C 2013 AIP Publishing LLC. conditions and embedded in a semiconductor stack. V [http://dx.doi.org/10.1063/1.4820266]
I. INTRODUCTION
Metal-based transparent conductors (MTCs) are a hot topic of interest in which it is determined by the ever growing request for transparent electrodes of lower cost and improved performance for renewable energy applications,1 in particular solar energy [e.g., Ref. 2], energy savings like “smart” windows,3 as well as by an emergent research on novel lighting devices and transparent electronics. Above all, it is the search of a valid alternative to indium-tin oxide in thin film devices, a task proven to be challenging up to now. Among various possible metal candidates for application in MTCs (e.g., Au,3,4 Cu,5,6 Al,7 and Pt8), silver5,9,10 is often the most proper choice thanks to its particular optical and electrical characteristics suitable for applications in the visible and near infrared range. Both optical constants and electrical conductivity of any thin film material differ from those of bulk ones due to many factors of film self-organization such as material density, morphology, etc. While for dielectrics this difference is mainly due to film stoichiometry and porosity, for conductors besides these two factors, localized and/or propagating surface plasmons (SPs) contribute to the film dielectric functions.11 Therefore, limits on optical transparency of MTCs posed by metal absorption may be tuned by varying both thickness and optical constants of the metallic layer.
a)
Author to whom correspondence should be addressed. Electronic mail:
[email protected]. Telephone: þ39 06 30484441. Fax: þ39 06 30486364.
0021-8979/2013/114(9)/094509/9/$30.00
Here, we report on optimization of deposition parameters of a radio-frequency (r.f.) sputtering system used for manufacturing high-quality MTCs made of silver and aluminium-doped zinc oxide (AZO) having an AZO/Ag/ AZO structure. The design of a low emissivity system based on such three-layer coating was proposed 15 years ago.12 This material combination produced by electron beam evaporation was intensively studied, see, for example, Ref. 13, and the island structure of ultra-thin silver was always reasonably referred to as the main difficulty to overcome in order to obtain high quality MTCs. In an attempt to minimize the percolation threshold (10–12 nm for evaporated films), sophisticated methods of evaporation rate control were applied2 such as substrate chilling or the use of a seed layer. Sputtering is known to facilitate creation of smooth thin metallic surfaces thanks to its higher deposition energy. In addition, sputtering is an economic and widely used technique in industry that increases the potential interest in our results in view of their large scale implementation. In fact, the magnetron-sputtered AZO/Ag/AZO three-layer is a typical industrial solution for low-emissivity coatings.1 Researchers, however, continue optimization of this material combination, as there is still vast room for its performance optimization through improvement of carrier transport mechanisms in the layers composing the electrode. Recent studies provide some characteristics of sputtered AZO/Ag/AZO structures: Ref. 14 reports on DC-sputtered AZO-Ag MTCs, while in Ref. 15 such r.f.-sputtered coatings were studied for composition (Rutherford backscattering spectrometry) and optical and electrical properties as a
114, 094509-1
C 2013 AIP Publishing LLC V
094509-2
Sytchkova et al.
function of silver layer thickness, similar to the approach proposed in Ref. 13 for evaporated films. A study of plasmonic excitation in a magnetron-sputtered AZO/Ag/AZO system using electron energy-loss spectroscopy (EELS) revealed the peculiarities of observable bulk and surface plasmons.16 In Ref. 17, ZnO/Ag bi-layers deposited by DC magnetron sputtering on glass were studied for optical and electrical properties as a function of the coalescence of the silver clusters, and plasmonic nature of electrical transport in partially percolated silver was discussed in detail. There, the Stranski–Krastanov growth mode was considered for silver on a smooth surface of the amorphous glass substrate at a chosen fixed value of deposition power. In all the cited works, silver for Ag-AZO MTC was studied as a percolating system, where the silver layer thickness is the only crucial parameter determining the MTC performance. To our best knowledge, in the literature there are no data on the dependence of the performance of the AZOAg MTCs on the sputtering deposition conditions. Silver growth on oxide surfaces was intensively studied18 and the wetting of the surface by the deposited material is known to be the main factor determining whether the process follows the Volmer-Weber or the Stranski–Krastanov growth mode. Weak chemical bonding between silver and oxides is the reason of a poor oxide surface wetting, and hence the difficulty to obtain smooth silver ultrathin layers on AZO. Deposition conditions of silver may play a determining role for the layer electrical and optical quality and we show that equally thin silver layers deposited at different r.f. powers have dramatically different morphologies leading to drastic differences in MTC performance. We searched to minimize the percolation threshold through adjustment of the deposition parameters leading to proper wetting of the sample surface. We show that the nominal thickness of the silver layer itself is not the critical value determining the MTC performance, as it is commonly cited in the literature, but an optimal formation of the metallic film ensuring an adequate percolation state of lowest possible clusters. The opto-electrical performance of r.f. sputtered AZO-Ag MTCs is, therefore, determined by their morphology, and is influenced by plasmonic properties of silver. Our coatings show an improved performance in a wider spectral transparency range compared to the data available in the literature and the proposed approach paves the way for further quality improvement of r.f.-sputtered AZO-Ag MTCs. II. EXPERIMENTAL DETAILS
Silver and AZO films were prepared by r.f. sputtering in a pure argon atmosphere, without breaking the vacuum, starting from 6-in. targets of 99.99% pure silver and 2% wt. Al2O3:ZnO, respectively. Before deposition, a vacuum level better than 1 106 mbar was ensured in the chamber (MRC 8620 sputtering plant, 13.6 MHz, sample-target distance 7 cm ca.). For the AZO target, a magnetron with magnetic field of 48 G at maximum was applied and all AZO depositions were done at 200 W r.f. power. Even though better AZO performances can be achieved on pre-heated substrates, as shown by some of the authors of this paper in a previous
J. Appl. Phys. 114, 094509 (2013)
work,19 no substrate heating was applied to avoid silver layer damage. The working gas pressure was 5 103 mbar for the AZO films, and 1.2 103 mbar for the silver films. Film thickness was monitored by manual timing, which introduces an uncertainty in reproducibility of the MTC structure (total thicknesses reported in Table I). However, such deposition control imperfection is compensated by a precise post-deposition characterization of the coatings. Electrical sheet resistance of the samples was measured with a home-made setup working in Van der Pauw configuration. Direct transmittance measurements were performed with a double-beam spectrophotometer (Perkin Elmer Lambda 950) in the spectral range of 300–2500 nm. No specific observations in the acquired scattering spectra (total reflectance and transmittance, as well as more sensitive angular-resolved reflected scattering (ARS) measurements) suggest a hypothesis of purely absorptive mechanisms of losses in the coatings. Ellipsometric investigations were performed with a rotating analyzer ellipsometer (W-VASE J.A. Woolam) in the range of 300–2500 nm at incidence angles of 55 , 65 , and 75 . Besides the ellipsometric angles w and D, the depolarization factor was acquired allowing a more precise modelling of effects due to specimen surface morphology. The samples were observed with a field-emission-gun SEM (FEG-SEM, LEO 35) at low-moderate accelerating voltage of 3–10 kV to reduce electron damage of the coating stack. SEM imaging allowed indirect assessment of relative differences in the electrical surface conductivity from sample to sample. Accordingly, the lower value of accelerating voltage (3 kV) was used for the less conductive surfaces to avoid charging. AFM measurements were performed with a FLEX-AFM system (NANOSURF, Switzerland) with 1 1 lm2 and 10 10 lm2 sampling area. The SEM micrographs were also analysed to assess the fractality of the deposited Ag clusters by checking whether the area R of the clusters evaluated over an observation window of length K (with the SEM scan size L K) would scale by a power law scaling R(K) KD with fractal exponent D, also known as "box counting dimension." After converting the SEM images into black and white format, the computation was performed by an automated algorithm provided by Image-J# (NIH, USA), which rendered D as the asymptotic slope of the regression line fitting log R vs. log K over a K range spanning two orders of magnitude P log ðKÞ : D ¼ lim K!L logK
(1)
III. RESULTS AND DISCUSSION A. Optical and electrical properties of coatings versus their nanostructure
The strategy of this study was to vary not only silver film thickness but also its morphology by varying the applied r.f. power in order to guarantee an efficient wetting and hence fractal growth of silver on AZO, aiming to diminish the percolation threshold. The goal morphology should
094509-3
Sytchkova et al.
J. Appl. Phys. 114, 094509 (2013)
TABLE I. Electro-optical characteristics and thicknesses of the AZO layer, silver layers, and multilayer stacks. Thicknesses were obtained from ellipsometric modelling. Here, Rs is the coating sheet resistance, q is the coating resistivity, T% is the coating transmittance either as a mean value in the range of 400–1000 nm or the value at peak in the vicinity of 450 nm, T is the transmittance value without the substrate glass contribution at peak in the vicinity of 450 nm, and FTC is the electrical revaluing merit of the coatings.
Sample AZO Bi-layer _60 W Bi-layer _150 W Sample 1 _60 W Sample 2 _60 W Sample 1 _150 W Sample 2 _150 W Sample 4 _150 W Sample 1 _200 W Sample 4 _200 W
Rs 4 points, Total thickness, X/w nm 28 000 >2M 14.5 20 130 8 14 20 16 15
61.3 52.4 41.6 53.5 52.3 48.7 49.7 56.5 58.3 61
Ag thickness, nm
q, lX cm
0 2.5 105 2.5 þ 27.9 at 70.4% porosity nd 19.9 þ 2.8 at 86.4% porosity 60 11.2 107 13 680 13.2 39 12.2 70 10 113 13.5 93 12.8 92
provide itself maximum trade-off between transparency and conductivity. However, additional yields may be expected from the enhanced plasmonic properties of semi-continuous films. In the case of almost continuous films, remaining nano-sized holes may induce the effect of enhanced transmittance,20 while a random metal grating may gain from SPs, which re-emit efficiently the incident light in a wide spectral range.17 1. Ag/AZO/glass bi-layers
We started with investigation of a silver film surface before the deposition of the top AZO layer of the final desired
T, at max near T% average T% at max 450 nm w/o glass FTC ¼ T10/Rs, 400–1100 nm near 450 nm contribution m X1 80.8
44.3 45.2 50.1 64.4 67.1 69.5 73.1
83.6 63 58.7 73.9 74.5 70.2 75 74.4 80.4 78.6
0.871 0.656 0.611 0.770 0.776 0.731 0.781 0.775 0.836 0.819
0.009 … 0.5 3.663 0.609 5.446 6.031 3.908 10.422 9.052
AZO/Ag/AZO structure. The thickest curves in Figures 1(a) and 1(b) represent direct transmittance spectra of the two Ag/ AZO/glass bi-layers deposited at 60 W and 150 W r.f. power under similar other conditions. Sample Bi-layer_150 W shows a pronounced optical loss band centred at 530 nm approximately and a weaker one at 340 nm (marked with an arrow), while the sample Bi-layer_60 W has a more marked 340 nm band. On the other hand, the 530 nm band of the last is immersed into the metal absorption wide band with a maximum at 750 nm, approximately. These absorption bands are imprints of different morphologies of the samples (SEM images in Figures 1(c) and 1(d)), determining differences in optical losses due to localized SPs.
FIG. 1. Measured direct transmittance spectra in comparison with FDTD simulations for bi-layers Ag/AZO/B270 deposited at different r.f. power values and their morphology: (a) and (c) insulating sample deposited at 60 W, (b) and (d) conducting sample deposited at 150 W. “Measurement” denotes the measured transmittance; the simulated transmittance curves were calculated using optical constants as follows: Ellips—our ellipsometric simulation, Woolam—WVASE32 database, CRC—Ref. 23, JC—Ref. 22, Palik— Ref. 24.
094509-4
Sytchkova et al.
Despite rather similar filling factors, f ¼ 0.86 for Bilayer_60 W and f ¼ 0.82 for Bi-layer_150 W, these bi-layers possess very different morphologic organization: if Bilayer_60 W is a dense agglomerate of separated domains, the surface of Bi-layer_150 W is an almost continuous although sparse net, so that only Bi-layer_150 W is a percolating system. This leads to their opposite electrical properties: the sample Bi-layer_60 W is insulating, while Bi-layer_150 W is conducting (Table I). The fractal exponents D of these samples deduced from their SEM images according to Eq. (1) were D 1.48 for “Bi-layer_150 W” and D ¼ 2 for “Bilayer_60 W,” which corresponds to the Euclidean dimension and is typical of "fat" fractals. An important factor to take into consideration when studying multilayers, and specially sputtered metal-dielectric stacks, is a probable difference in the layer thickness and optical constants, when a material is deposited singularly or within a stack, due to material re-sputtering and creation of imperfect interfaces. The characterization of such multilayers is a sophisticated task, being the properties of a “sandwiched” ultrathin metal are highly variable due to both deposition conditions and embedding materials influencing the plasmonic properties of the metal layer. Here, we propose an approach to model the optical properties of MTC coatings based on variable angle spectroscopic ellipsometry (VASE) measurements that take into consideration the variation of the optical properties of silver when embedded in a semiconductor stack. For the data elaboration, the WVASE32 software was used implementing the general oscillator models for both materials. Discontinuity of the silver layers within MTC was modelled by their modified optical functions. Additionally, the porosity of the bi-layers was described by the Effective Medium Approximation (EMA) model [e.g., Ref. 21], where
J. Appl. Phys. 114, 094509 (2013)
the outer part of the silver layer (so-called overlayer) is a mixture of the metal material with air. For the two bi-layers, the overall deposited silver amount given by composition of a continuous silver layer and a porous EMA-modelled overlayer, as retrieved from ellipsometric data, is rather similar: for Bi-layer_60 W it is 2.5 nm plus an overlayer of 27.9 nm at 70.4% porosity, while for Bi-layer_150 W it is 19.9 nm plus an overlayer of 2.8 nm at 86.4% porosity. Despite similar nominal thickness of these silver layers, the thicknesses of the continuous parts and the EMA overlayers are very different for the two samples. This puts in evidence the difference in film morphology due to different r.f. power values during the deposition: the Bilayer_150 W is far denser and is composed of lower and larger clusters compared to Bi-layer_60 W. Simulations of direct transmittance data for the samples Bi-layer_60 W and Bi-layer_150 W were performed with finite-difference time-domain (FDTD) software, starting from the SEM images of the samples and using silver refractive indices we determined from ellipsometric measurements and those available in the literature22–24 and in the WVASE32 software data base. For all simulations, we used AZO optical constants determined from our ellipsometric characterization. The numerical results are plotted in Figures 1(a) and 1(b) in comparison with the experimental curves. Similar modelling was performed for calculation of optical properties in thin film coatings with integrated silver nanoparticles25 and in metal-oxide structures.26,27 For each of the two samples, the simulated spectral curves in Fig. 1 may be considered to follow coarsely the shape of the experimental curve. The observed discrepancy is likely attributed to inadequate surface sampling for the simulation program. In fact, due to limitations of the calculation facility, the maximum possible
FIG. 2. (a) Survey on direct transmittance curves for three-layer samples, for a bare glass substrate (B270) and for a single AZO layer. (b) Morphology of an AZO (60 nm)/glass sample. (c) Morphology of an AZO/Ag/AZO/glass sample (sample 1_60 W): (d) backscattering image of the last.
094509-5
Sytchkova et al.
J. Appl. Phys. 114, 094509 (2013)
FIG. 3. Ellipsometric angles and depolarization factor for (a) sample 1_60 W, (b) sample 1_150 W, and (c) sample 1_200 W. Dashed curves are experimental data, solid curves are model fit.
094509-6
Sytchkova et al.
J. Appl. Phys. 114, 094509 (2013)
FIG. 3. (Continued.)
surface taken for image acquisition was about 100 100 nm2, which cannot be considered representative for the surface morphology. The position of main transmission peak is better predicted when using the ellipsometrically determined silver n and k. In general, in this case the qualitative similarity of experimental and FDTD-simulated transmittance spectra of AZO-Ag bi-layers may be interpreted as a confirmation of the correctness of ellipsometric data modelling, despite the corresponding geometrical model of the coating does not consider silver discontinuity, but operates in terms of ideal plane-parallel layers. The resulting silver optical constants should be understood as averaged values corresponding to larger surface integration. In fact, use of “bulk” silver n and k for FDTD simulations brings better overall value of the coating transmittance. However, this last approach loses in prediction of the coating spectral behaviour, when compared to the experimentally derived silver n and k. This is especially true for the Bi-layer_150 W at longer wavelengths. Summarizing, the electrical and optical properties of the two bi-layers are very different: while Bi-layer_60 W is insulating and absorbing in the NIR, Bi-layer_150 W is conducting despite its lower filling factor and has relatively good NIR transparency. Consequently, the filling factor value and nominal film thickness are insufficient for prediction of optical and electrical properties of the film, while the fractal exponent is more representative for this purpose. Coalescence of the metal islands is the first factor to take into consideration when
estimating the MTC conductivity, but not the only factor. Light re-emission by the SPs of the random grating of the coalescing silver may play an important role in the case of ultrathin metal layers.17 It should also be taken into consideration that there is a correlation between enhanced fluctuations of the local density of optical states (LDOS) in disordered fractal metallic films and existence of localized SP modes.28 The LDOS is an essential quantity for characterization of optical properties of complex systems as it drives the spontaneous light emission by dipoles.29 It determines, therefore, macroscopic transport properties of the films, in particular the photon transport regime.30 Moreover, as shown experimentally,28 the localized SP modes dominate around the percolation threshold, although not exactly at percolation, and can substantially modify LDOS. As for the electrical conductivity of disordered metallic fractal films, the lateral extension of the fractal clusters is evidently the most essential factor driving the path of free electrons. Propagating SP modes are most related to conducting properties of semi-continuous films, and the electrical conductivity simply increases with increase of film thickness, achieving its maximum for “bulky” films. On the other hand, the optical properties of disordered metallic fractal films are also connected to SP modes, but both localized and extended. The peculiarity consists in enhanced fluctuation of the LDOS in the vicinity of percolation threshold, which renders the optical properties of such films particularly interesting at this
094509-7
Sytchkova et al.
formation state. The optimization of the MTC structures may be proposed as a search of the deposition conditions ensuring just initial metal percolation, whereas the height of the clusters should be maintained as low as possible. A lower and denser cluster net may be obtained by increasing energy of the deposition process ensuring improved surface wetting. Conversely, some additional percolation may be induced in the metal due to energetic bombardment by the molecules of the next (top AZO) layer during its deposition. An optimal resulting formation state of the metallic film should be found, so that the SP modes of partially percolated clusters ensure enhanced transmission of the MTC stack. 2. Three-layers AZO/Ag/AZO/glass
The thicknesses and main electrical and optical parameters of the coatings are listed in Table I. The transmittance curves of the deposited three-layer structures AZO/Ag/AZO/ glass are reported in Figure 2(a) in comparison with the transmittance curves of both a bare glass substrate and an AZO layer with a thickness of about 60 nm, which is approximately the sum of the two AZO layers sandwiching the silver layer in the three-layered structures. The SP absorption positioned at approximately 385 nm for the bi-layers is less evident for the three-layers. It is mainly seen at 55 ellipsometric observation, being redshifted to 389 nm, in agreement with the literature data.31 The resonance peak appearing at 530 nm for bi-layers vanishes once the silver layer is covered by another AZO layer, both in transmittance curves (Figure 2(a)) and in the spectra of the ellipsometric w-angle (Figure 3). The deposition of the third antireflection layer has probably favoured an additional percolation between the silver clusters, so that only the intrinsic absorption at plasma frequency of silver remains visible in the spectral curves at higher incidence angles. In fact, surface morphology of the three-layer sample 1_60 W is shown in Figure 2(c) and represents a superposition of morphologies of a single AZO layer on glass sample of thickness 20 nm approximately (Figure 2(b)), and a conducting Ag/AZO/glass sample Bi-layer_150 W (Figure 1(d)), despite this last sample was deposited at higher r.f. power. Notice that the back-scattering image of sample 1_60 W, Figure 2(d), proves the chemical homogeneity of the sample surface composed only from AZO material without silver migration to the sample surface. In fact, this agrees with the results of Ref. 15, where AZO is reported as a good barrier layer preventing silver diffusion that would be detrimental for Si-based solar cells (Rutherford back scattering measurements on r.f. sputtered AZO/Ag/AZO structure). Figure 3 shows examples of ellipsometric data fit for three representative samples deposited at three chosen sputtering power values. For the ellipsometric spectra fitting, the dielectric function of the AZO layers was maintained the same for all the samples in the fitting process, while their thickness was a free parameter. Other parameters for the fit were the silver layer thickness together with its refractive index and extinction coefficient. Refractive indices and extinction coefficients of the silver layers within the three-layer samples are reported in
J. Appl. Phys. 114, 094509 (2013)
Figures 4(a) and 4(b), respectively. The thicknesses of the layers within the three-layer coatings are listed in Table I. Notice how different the transmittance spectra (Figure 2(a)) are for the samples deposited at three different r.f. power values, sample 2_60 W, sample 1_150 W, and sample 1_200 W, despite their almost equally thick silver layers (about 13 nm). In general, the samples deposited at 200 W have the highest transmittance and widest transparency window that, combined with good conductivity of these samples, determines the highest quality of 200 W-samples. The best among all of the samples deposited at 150 W (differing in silver layer thickness), sample 4_150 W, is comparable with 200 Wsamples in the transparency range. The refractive index and extinction coefficient of silver in these highest-quality samples are different from the optical constants of silver in the other 150 W-deposited samples, and both differ significantly from the optical constants of the 60 W-deposited samples. The reason for this should be looked for in different coalescence states of these films. Higher r.f.-power employed during silver deposition induces the Stranski-Krastanov growth of the metal islands already for 150 W-grown silver. The almost continuous morphology of sample 1_200 W is shown in Fig. 5(a). The peculiarity of this sample consists,
FIG. 4. Refractive indices (a) and extinction coefficients (b) of silver layers within the three-layer structures AZO/Ag/AZO/glass.
094509-8
Sytchkova et al.
FIG. 5. Morphology (a) and topography (b) of AZO/Ag/AZO/glass sample with silver deposited at 200 W (sample 1_200 W).
moreover, in the presence of nano-sized holes (AFMacquired topography is shown in Fig. 5(b)), which might additionally improve the electrode performance thanks to enhanced transmittance through such plasmonic structures.20 To confirm this hypothesis, additional investigations should be performed, including near-field measurements. In Table I, the coating resistivity is reported together with the figure of merit FTC ¼ T10/Rs used to evaluate the performance of the bi- and three-layers. Notice that to enable comparison with the data reported in the literature, we applied here the most common definition of the figure of merit, i.e., T is taken as the maximum value of the coating transmittance in the vicinity of k ¼ 450 nm, and Rs is the coating sheet resistance. However, the width of the transparency range differs noticeably from sample to sample, and most of them appear transparent over a significantly wider wavelength range compared to the data reported in the literature. Therefore, the transmittance value averaged in the range of 400–1100 nm is also reported to illustrate the enhanced optical quality of the MTCs deposited at high r.f. power, especially useful for wide-range applications such as solar cells. The best sample (with 200 W-deposited silver layer) has the mean transmittance value of 73%. IV. CONCLUSIONS
A systematic study was performed for optimization of r.f. sputtered AZO/Ag/AZO MTCs. The MTC performance optimization is proven to be possible through tuning of the
J. Appl. Phys. 114, 094509 (2013)
deposition conditions aimed at diminishing the percolation threshold of the silver layer. Resulting low-height coalescent metal clusters not only have evidently lowest possible absorbance but also might gain from plasmonic features such as re-emission of the light coupled by the SPs at the random silver-island grating or enhanced transmittance through nano-apertures. Proper modelling of VASE data has allowed calculation of the optical and geometrical parameters of the coating components. The optical functions of ultrathin silver layers sandwiched between thin AZO films were modelled by multiple general oscillators. The oscillators’ parameters change with variation of the deposition conditions, and may be interpreted by considering silver film plasmonic features determined by the film morphology. Lower and better-wetting silver clusters on the AZO film may be obtained when applying higher r.f.-power during silver deposition; this induces the Stranski-Krastanov growth of the metal islands. The consequent fractal-type development of ultrathin silver layers ensures both better electrical conductivity at thinner metal film clusters as well as selforganization of a random metal grating with a variety of space constants matching the variety of wave-vectors of the incident light, hence enlarging the transparency range of the MTC. On the other hand, almost continuous films may additionally benefit from transmission through nano-sized holes. Near-field investigations are planned to confirm these hypotheses. Indirect confirmation may be seen in the spectral features of the dielectric functions modelled of silver layers. For 150 W- and 200 W-deposited silver layers, MTCs with at-peak transmittance as high as 83.6% have been obtained. These coatings may have FTC as high as 0.01 X 1 and have a remarkably wide spectral extension of the coating transparency region (transmittance higher than 70% down to 1200 nm). The observed higher transparency in a wider spectral range for the MTC with silver deposited at 200 W when compared to conventionally 60 W-deposited films and even to best MTC with 150 W-deposited silver paves the way to a general process optimization. ACKNOWLEDGMENTS
The authors thank V. Brissonneau for the help with ARS measurements. 1
K. Ellmer, “Past achievements and future challenges in the development of optically transparent electrodes,” Nat. Photonics 6, 809 (2012). 2 N. P. Sergeant, A. Hadipour, B. Niesen, D. Cheyns, P. Heremans, P. Peumans, and B. P. Rand, “Design of transparent anodes for resonant cavity enhanced light harvesting in organic solar cells,” Adv. Mater. 24, 728 (2012). 3 P. C. Lansaker, J. Backholm, G. A. Niklasson, and C. G. Granqvist, “Influence of Ag thickness on structural, optical, and electrical properties of ZnS/Ag/ZnS multilayers prepared by ion beam assisted deposition,” Thin Solid Films 518, 1225 (2009). 4 Y. S. Kim, J. H. Park, D. H. Choi, H. S. Jang, J. H. Lee, H. J. Park, J. I. Choi, D. H. Ju, J. Y. Lee, and D. Kim, “ITO/Au/ITO multilayer thin films for transparent conducting electrode applications,” Appl. Surf. Sci. 254, 1524 (2007). 5 C. Guillen and J. Herrero, “ITO/metal/ITO multilayer structures based on Ag and Cu metal films for high-performance transparent electrodes,” Sol. Energy Mater. Sol. Cells 92, 938 (2008).
094509-9 6
Sytchkova et al.
D. R. Sahu and J.-L. Huang, “Deposition of Ag-based Al-doped ZnO multilayer coatings for the transparent conductive electrodes by electron beam evaporation,” Thin Solid Films 516, 208 (2007). 7 C. Lee, R. P. Dwivedi, C. Hong, W. I. Lee, and H. W. Kim, “IZO/Al/GZO multilayer films to replace ITO films,” J. Mater. Sci.: Mater. Electron. 19, 981 (2008). 8 C.-H. Cheng, W.-Y. Wu, and J.-M. Ting, “Nanoscaled multilayer thin films based on GZO,” J. Nano Res. 2, 61 (2008). 9 M. Fahland, T. Vogt, W. Schoenbergen, and N. Schiller, “Optical properties of metal based transparent electrodes on polymer films,” Thin Solid Films 516, 5777 (2008). 10 K. Koike, F. Yamazaki, T. Okamura, and S. Fukuda, “Improvement of corrosion resistance of transparent conductive multilayer coating consisting of silver layers and transparent metal oxide layers,” J. Vac. Sci. Technol. A 25, 527 (2007). 11 V. P. Drachev, U. K. Chettiar, A. V. Kildishev, H.-K. Yuan, W. Cai, and V. M. Shalaev, “The Ag dielectric function in plasmonic Metamaterials,” Opt. Express 16, 1186 (2008). 12 P. Grosse, R. Hertling, and T. Mueggenburg, “Design pf low-emissivity system based on a three-layer coating,” J. Non-Cryst. Solids 218, 38 (1997). 13 D. R. Sahu, S.-Y. Lin, and J.-L. Huang, “Investigation of conductive and transparent Al-doped ZnO/Ag/Al-doped ZnO multilayer coatings by electron beam evaporation,” Thin Solid Films 516, 4728 (2008). 14 H.-K. Park, J.-W. Kang, S.-I. Na, D.-Y. Kim, and H.-K. Kim, “Characteristics of indium-free GZO/Ag/GZO and AZO/Ag/AZO multilayer electrode grown by dual target DC sputtering at room temperature for low-cost organic photovoltaics,” Sol. Energy Mater. Sol. Cells 93, 1994 (2009). 15 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, 4432 (2012). 16 M. Philipp, M. Knupfer, B. B€ uchner, and H. Gerardin, “Plasmonic excitations in ZnO/Ag/ZnO multilayer systems: Insight into interface and bulk electronic properties,” J. Appl. Phys. 109, 063710 (2011). 17 D. Zhang, H. Yabe, E. Akita, P. Wang, R. Murakami, and X. Song, “Effect of silver evolution on conductivity and transmittance of ZnO/Ag thin films,” J. Appl. Phys. 109, 104318 (2011).
J. Appl. Phys. 114, 094509 (2013) 18
C. T. Campbell, “Ultrathin metal films and particles on oxide surfaces: Structural, electronic and chemisorptive properties,” Surf. Sci. Rep. 27, 1 (1997). 19 M. L. Grilli, A. Sytchkova, S. Boycheva, and A. Piegari, “Transparent and conductive Al-doped ZnO films for solar cells applications,” Phys. Status Solidi A 210, 748 (2013). 20 F. van Beijnum, C. Retif, C. B. Smiet, and M. P. van Exter, “Transmission processes in random patterns of subwavelength holes,” Opt. Lett. 36, 3666 (2011). 21 L. Ward, The Optical Constants of Bulk Materials and Films (JW Arrowsmith, Bristol, 1988). 22 P. B. Johnson and R. W. Christy, “Optical constants of noble metals,” Phys. Rev. B 6, 4370 (1972). 23 W. M. Haynes, CRC Handbook of Chemistry and Physics (CRC Press, Oxford, 2013). 24 E. D. Palik, Handbook of Optical Constants of Solids II (Academic Press, Boston, 1991). 25 S. Vedraine, P. Torchio, D. Duche, F. Flory, J. J. Simon, J. Le Rouzo, and L. Escoubas, “Intrinsic absorption of plasmonic structures for organic solar cells,” Sol. Energy Mater. Sol. Cells 95, S57 (2011). 26 M. D. Thoreson, J. Fang, A. V. Kildishev, L. J. Prokopeva, P. Nyga, U. K. Chettiar, V. M. Shalaev, and V. P. Drachev, “Fabrication and realistic modeling of three-dimensional metal-dielectric composites,” J. Nanophotonics 5, 051513 (2011). 27 D.-T. Nguyen, S. Vedraine, L. Cattin, P. Torchio, M. Morsli, F. Flory, and J. C. Berne`de, “Effect of the thickness of the MoO3 layers on optical properties of MoO3/Ag/MoO3 multilayer structures,” J. Appl. Phys. 112, 063505 (2012). 28 V. Krachmalnikov, E. Castanie, Y. De Wilde, and R. Carminati, “Fluctuations of the local density of states probe localized surface plasmons on disordered metal films,” Phys. Rev. Lett. 105, 183901 (2010). 29 J. M. Wylie and J. E. Sipe, “Quantum electrodynamics near an interface,” Phys. Rev. A 30, 1185 (1984). 30 R. Pierrat and R. Carminati, “Spontaneous decay rate of a dipole emitter in a strongly scattering disordered environment,” Phys. Rev. A 81, 063802 (2010). 31 G. Xu, M. Tazawa, P. Jin, and S. Nakao, “Surface plasmon resonance of sputtered Ag films: Substrate and mass thickness dependence,” Appl. Phys. A 80, 1535 (2005).
