ARTICLE pubs.acs.org/Langmuir
Ultrasmooth Silver Thin Film Electrodes with High Polar Liquid Wettability for OLED Microcavity Application Cristina Cioarec,†,‡ Patrizia Melpignano,*,†,‡,§ Nicolas Gherardi,†,‡ Richard Clergereaux,†,‡ and Christina Villeneuve|| †
)
Universite de Toulouse, UPS, INPT, LAPLACE (Laboratoire Plasma et Conversion d'Energie), 118 route de Narbonne, F-31062 Toulouse cedex 9, France ‡ CNRS, LAPLACE, F-31062 Toulouse, France § OR-EL.doo, Volariceva Ulica 6, 5222 Kobarid, Slovenija CNRS-LAAS, 7, Avenue du Colonel Roche, 31077 Toulouse Cedex 4, France ABSTRACT:
For a lab-on-chip application, we fabricate a blue bottom emitting strong microcavity organic light emitting diode (OLED), using very smooth and optically thin (25 nm) silver film as anode on a glass substrate. To improve the hole injection in the OLED device, PEDOT-PSS (poly(3,4-ethylenedioxythiophene)-poly(styrene sulfonic acid)) has been used, so the silver anode must present not only a very smooth surface but also a strong adherence on the glass and a high wettability to allow a good PEDOT-PSS spin coating deposition. To obtain these physical properties, different 5 nm thick nucleation layers (germanium, chromium, and hydrogenated amorphous carbon) have been used to grow the silver thin films by e-beam deposition. The Ge/Ag bilayer presents all the desired properties: this bilayer, investigated by ellipsometry, optical profilometry, contact angle measurements, and XPS analysis, highlights an ultrasmooth surface correlated with the film growth mode and a high wettability related to its surface chemical composition.
’ INTRODUCTION Since their beginning, computers and communication systems have been dominated by electronic technology, but recently photonic technology is making serious inroads throughout the optical telecommunication systems where devices such as lasers, light-emitting diodes, photodetecting diodes, optical switches, optical amplifiers, optical modulators, and optical fibers are widely used. Most of these innovative devices and circuits require the use of ultrasmooth thin films of noble metals such as silver (Ag), platinum (Pt), or gold (Au),1-4 both as semitransparent optical elements and as electrical contacts. One specific photonic application of ultrasmooth Ag thin film is the fabrication of semitransparent electrodes for organic light emitting diodes (OLEDs) in microcavity configuration. This particular application requires atomic scale thin film roughness to avoid short circuits between the electrodes that sandwich the thin organic multilayer stacks. The aim of the present study is to optimize the anode of a blue bottom emitting strong microcavity OLED to be used as excitation source for a fluorophore emission in a biochip application. The biochip has been specifically developed for point of care (POC) diagnostics, in which an entire test is self-contained in a hand-held device, delivering onthe-spot medical advice.5 r 2011 American Chemical Society
The optimized architecture of our microcavity OLED requires a 25 nm thin silver film anode, e-beam deposited, with a 40 nm thick PEDOT-PSS layer deposited on it to improve the holes injection into the OLED. PEDOT-PSS is a polymer widely used in the fabrication of OLEDs: PEDOT is a polythiophene derivative, poly(3,4-ethylenedioxythiophene), and PSS is a water-soluble polyelectrolyte, poly(styrene sulfonic acid). PEDOT and PSS chains are linked by ionic interaction and form an ionic polymer complex dispersed in aqueous solution with good film-forming properties, high conductivity, high visible-light transmission, and excellent stability.6 The 25 nm silver film thickness is calculated by optical simulations to obtain an OLED emission spectrum optimally matched to the absorption spectrum of the tag fluorophore used in our biochip application. However, preliminary tests showed that deposition of the Ag thin film directly onto a borosilicate glass substrate exhibits high surface roughness and poor film/ substrate adhesion, incompatible with the fabrication of an efficient OLED device. It is important to note that, in our Received: November 30, 2010 Revised: January 31, 2011 Published: March 10, 2011 3611
dx.doi.org/10.1021/la104760a | Langmuir 2011, 27, 3611–3617
Langmuir application, the ultrasmooth roughness must be combined also to a good adhesion of Ag thin film on borosilicate glass and to a suitable wettability to allow PEDOT-PSS layer spin-coating deposition without degrading the Ag anode or its roughness. It has been reported7,8 that the use of nucleation layers can improve the Ag thin film quality, obtaining a semitransparent, homogeneous, spike free, conductive, and adherent layer. To verify the applicability of this approach to our specific problem, a careful study of Ag thin film deposition on cleaned borosilicate glass coated by nucleation layers, such as chromium (Cr), germanium (Ge), and hydrogenated amorphous carbon (a-C: H), has been carried out. The choice of these nucleation layers has been driven by several factors: the chromium’s well-known properties of noble metal adhesion promoter on glass, the recent results reported in literature7 for Ge, and the possibility to obtain a very cheap nucleation layer by using carbon. In this paper, we demonstrate that a bilayer based on a 5 nm thick germanium nucleation layer allows to obtain the desired characteristics of the 25 nm thick silver film: the extremely low roughness but also the high substrate adhesion and the high wettability with polar liquids such as PEDOT-PSS.
’ EXPERIMENTAL SECTION The samples were prepared using high quality borosilicate glass substrates, 1 mm thick, which were ultrasonically cleaned with organic solvents (acetone and isopropyl alcohol), rinsed with deionized water, and dried with purified nitrogen in a class 10 clean room. After the cleaning procedure, the glass samples were transferred in different deposition chambers to evaporate the nucleation layers: • A 5 nm thick chromium (99.999%) layer was deposited by thermal evaporation in a VEECO evaporator at a pressure of 2 10-6 mbar and a deposition rate of 0.4 Å 3 s-1. The film thickness was determined using a quartz microbalance film thickness monitor (FTM) system included in the VEECO reactor. • A 5 nm thick hydrogenated amorphous carbon layer has been deposited on the glass substrates in a microwave multipolar plasma excited at distributed electron cyclotron resonance (MMP-DECR) using methane (CH4) as the precursor.9 The substrate holder was kept at room temperature. The plasma was produced using a working pressure of about 10-3 mbar. The microwave (MW) plasma was fixed at 400 W, and the deposition time lasted 2 min (the film thickness deposition was calibrated at 2.5 nm/min for these plasma evaporation conditions). Before being used for the silver deposition, the a-C:H samples were ultrasonically cleaned with organic solvents and rinsed with deionized water. • A 5 nm thick germanium (99.999%) layer was evaporated by e-beam with a rate of 0.1 Å 3 s-1 at a pressure of 5 x 10-6 mbar in a BOC Edwards 500 reactor. Pure silver thin films (99.999%) with thicknesses ranging from 1 to 50 nm were evaporated by electron beam technique in a BOC Edwards 500 reactor with a rate of 0.5 Å 3 s-1. The silver films were deposited without breaking the vacuum on the germanium samples only. The silver and germanium film thicknesses were determined using a quartz microbalance FTM system included in the BOC Edwards 500 reactor. The chamber’s pressure during the evaporations was 1.5 10-6 mbar. The growth modes of the silver thin films deposited on different nucleation layers have been analyzed ex situ by using a spectroscopic ellipsometer GES-5 from SOPRA, with the incidence angle being fixed to the glass Brewster angle (70°). The ellipsometric angles Ψ and Δ were measured at the wavelength of 632 nm. The surface roughness was measured by using a noncontact whitelight interferometer precision optical profiler WYKO NT 3000 in phase shifting interferometry (PSI) mode. This technique allows to measure
ARTICLE
rather large surfaces (around 1 mm2 using a 5 objective) with a vertical resolution as high as 0.3 nm. For each sample, measurement of the following parameters has been performed over the probed surface: • Mean roughness (Ra), defined as the arithmetic mean distance between successive peaks and valleys over a defined distance 1 Ra ¼ L L0
ZL ZL0 zðx, yÞdxdy
ð1Þ
0 0
where z(x,y) is the local height at point x,y and L and L0 are the total measured distances. • Root mean square roughness (Rq) defined as vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi u ZL ZL0 u u 1 t Rq ¼ ðz - mÞ2 dxdy ð2Þ L L0 0 0
where m is the mean of height z over the measured area LL0 . • Maximum peak-to-valley height of the profile (Rt), that corresponds to the sum of the maximum peak height and the maximum valley heights on the measured area LL0 . In order to calculate the wettability and the related surface free energy of our bilayers, contact angles were recorded for three different pure liquids: diiodomethane (Sigma Aldrich), used as apolar liquid; formamide (UCB, Brussels, Belgium) and deionized water, used as polar liquids, by using a contact angle meter (DIGIDROP, GBX Co., France) equipped with a CCD camera (25 frames/s). Surface free energy is further obtained using the Owens-Wendt model.10 The final control on the physicochemical structure of Ag thin films has been carried out by using X-ray photoelectron spectroscopy (XPS). A NOVA-KRATOS XPS (BIOPHY RESEARCH S.A.-France) instrument, using a monochromatic Al KR source in normal detection geometry, has been used to analyze an area of 300 700 μm2. The analyses have been performed both on pristine surfaces and on subsurfaces. The Ag subsurface is reached by sputtering with Arþ primary ions at 0.5 keV, with a rate of 1 nm 3 min-1. The sputtered surface was 3 3 mm2, and the sputtered thickness was 3 nm.
’ RESULTS AND DISCUSSION In general, silver film growth on glass starts from initially isolated metallic islands to a connecting network and finally to a complete coverage of the whole area.11,12 The presence of a nucleation layer really affects the Ag growth mode. Indeed, the growth mode can be extrapolated from the simultaneous variation of the two ellipsometric angles Ψ and Δ, or their reduced parameter sin2Ψ sin Δ, as a function of the silver layer thickness. The evolution of the reduced parameter sin2Ψ sin Δ with the silver thickness is plotted in Figure 1a, where the continuous curves represent three simulations of the different growth modes and the crosses are the experimental data. The calculated curves reported in Figure 1a were obtained by simulation using the growth ellipsometric models schematized in Figure 1b:13,14 • M1: Franck-van der Merwe layer-by-layer growth mechanism, where the film thickness increases linearly with time and the film roughness is independent of its thickness. • M2 and M3: Volmer-Weber growth mechanism, where from a nucleation dot the film grows in a 3D growth mode up to the coalescence reached for the critical thickness, tc. In our simulations, tc was fixed at 5 nm. For thicknesses higher than tc, two different evolutions are possible, in particular: • M2: Conformal growth, where the thickness increases linearly with time while the roughness keeps constant. Note that M2 never took place as shown in Figure 2, so 3612
dx.doi.org/10.1021/la104760a |Langmuir 2011, 27, 3611–3617
Langmuir the curve of M2 is reported only in the first panel of Figure 1a while in the other panels only M1 and M3 have been traced. • M3: Planarization, where the thickness increases linearly with time while the roughness decreases. Simulations were performed using Winelli 2 software from SOPRA. Based on the optical indexes for the wavelength of 632 nm of pure Ag (n = 1.433, k = 0.7611) and of pure Cr (n = 3.12, k = 3.3), Ge (n = 5.44, k = 0.75), and a-C:H (n = 1.482, k = 1.8 10-3), M1 was obtained only by increasing the Ag
Figure 1. (a) Growth model curves and measured ellipsometric data (crosses) as a function of Ag thickness on various nucleation layers. (b) Description of different models of Ag thin film growth: M1, Franck-van der Merwe layer-by-layer growth mechanism; M2 and M3, VolmerWeber growth mechanism.
ARTICLE
thickness, and M2 and M3 were simulated for an Ag coalescence at 5 nm. These last two models were obtained by increasing the Ag volume fraction in a 5 nm thick layer compound of silver with voids. Above coalescence, the Ag layer deposition was simulated by a layer-by-layer Ag deposition on the substrate with a top layer described by an effective medium of silver with voids film. Opposite to the M2 model, where the Ag volume fraction was kept constant, in the M3 model the Ag volume fraction was increased up to 100%. From our measurements, it appears that silver growth process on the Cr nucleation layer, as the one on a glass substrate, is well simulated with Volmer-Weber model considering a coalescence thickness of tc = 5 nm. The smoothest surface is reached for about 10 nm thick layer (intersection of M1 and M3). In contrast, for the Ag growth process on the a-C:H and Ge nucleation layers, the model M1 seems to be more appropriate to simulate the experimental data for Ag thicknesses larger than 10 nm. It means that, even if in the very first growth step, a VolmerWeber model with very small and dense nucleation dots can be considered,8 the coalescence is really fast (tc < 1 nm), leading to a smoother surface for both a-C:H and Ge nucleation layers. In parallel, the roughness of the Ag thin film deposited on the different nucleation layers has been measured with the optical profiler as reported in Figure 2. For all samples, the roughness decreases with increasing silver thickness. The highest Ra values are measured below 5 nm, confirming that the coalescence for Volmer-Weber grown films takes place below this thickness. When increasing the silver thickness, the roughness decreases to a constant value which highlights that the smoothest silver layer is reached above 10 nm. Moreover, the surface roughness depends on the nucleation layer as well, decreasing in the order glass > Cr > a-C:H > Ge. The surface planarization is then really affected by the nucleation layer. These experimental observations are in good agreement with the ellipsometry measurements: the roughness of the Ag thin film deposited on glass and Cr (tc > 5 nm) is about 2 or 3 times higher than that measured on a-C:H and Ge (tc < 1 nm). One possible explanation for the silver 3D growth mode on Cr and glass substrates is the presence of oxide on the nucleation layer. In fact, it is well-known that Ag (and in general the noble
Figure 2. Left: Graph of mean roughness vs thickness for silver films grown on different nucleation layers. Right: Optical profilometer 3D surface profile of Ag (25 nm) on (a) glass, (b) a-C:H, (c) Cr, and (d) Ge. 3613
dx.doi.org/10.1021/la104760a |Langmuir 2011, 27, 3611–3617
Langmuir metals) grows on oxidized surfaces in 3D island mode.15-18 The Cr surface is easily oxidized when extracted from the evaporator and exposed to atmosphere. This results in the formation of a dense Cr(III) oxide (Cr2O3) layer.19 So the successive Ag growth follows the three-dimensional Volmer-Weber growth mode. In contrast, Ag growth does not take place following this mode on Ge that can be oxidized as well as Cr. It grows as on a-C:H that is more resistant to oxygen ambient.20 Then, the growth mode is not only due to the oxidation state of the surface. The layer-by-layer growth of Ag on a-C:H can be attributed to another mechanism: the formation of a thin alloy between carbon and Ag at the interface. Indeed, it is well-known that the solid solubility of graphite in Ag can be obtained in the temperature range of 785-957 °C, with 0.036 atom % carbon at the silver melting point temperature (Ag evaporation temperature).21 This hypothesis could explain the silver coalescence at less than 1 nm on this nucleation layer. In the same way, the early silver coalescence on the germanium layer can be attributed to the formation of an Ag-Ge solid solution. It is known that, depending on Ag concentration,22,23 the formation of a Ge/Ag alloy is possible. In our experimental conditions, considering that the eutectic temperature of a Ag/Ge alloy is 924 °C, an Ag/Ge alloy could be formed in the very first stage of Ag evaporation on Ge. In our bilayers, we have observed differences not only related to the growth mode, and the associated surface roughness, but also in their wettability with polar liquids. This parameter is of importance in our application that requires the spin coating deposition of an aqueous solution, in our case a solution of PEDOT-PSS. To measure the wettability of our Ag based bilayers, sessile drop static contact angle measurements were made with two polar liquids and one apolar liquid, as described before. Contact angles of PEDOT-PSS were also measured in order to analyze its wettability on all the substrates, but these values were neglected in the calculation of the surface free energy, γS. The surface wettability of a film is related to its surface free energy which is determined by the surface chemical bonds, structure, topography, and presence of adsorbates. The overall wettability is determined by the interplay of these factors, with some of them promoting hydrophobicity and others enhancing hydrophilicity. The contact angles measured on 25 nm thick silver thin films deposited on the different nucleation layers are reported in Table 1. The snapshots of PEDOT-PSS contact angles on these samples are shown in Figure 3. Moreover, the wettability on Ag films of different thicknesses has been measured on the nucleation layers that present a planarized surface which allows the minimization of roughness effect on the surface wettability. The contact angles are really dependent on the nucleation layer, and surprisingly Ge and a-C: H nucleation layers do not lead to the same behavior. For Ag deposited on Ge, the contact angle is lower for deionized water, formamide, and PEDOT-PSS, while it is higher for diiodomethane, as in the case of a pure Ge layer. Furthermore, a dependence of the water contact angles with the Ag thickness is also evident, different from the case of the a-C:H nucleation layer. The contact angle of a liquid drop (θY) on a solid surface is defined by the mechanical equilibrium of the drop under the action of three interfacial tensions: solid-vapor (γSV), solidliquid (γSL), and liquid-vapor (γLV). This equilibrium relation
ARTICLE
Table 1. Measured Contact Angles on Different Thickness Ag Based Bilayers θDI water
θformamide
substrate
(°)
(°)
θdiiodomethane θPEDOT-PSS (°)
(°)
glass/Ag 25 nm Cr 5 nm/Ag 25 nm
41.4 51.1
32.3 38.1
5.3 7.7
52.1 48.9 12.5
Ge5 nm
4.4
7.1
20.1
Ge 5 nm/Ag 10 nm
4.4
4.3
25.1
8.7
12.1
6.8
25.9
18.2
21.7
6.5
14.2
24.2
54.5 57.7
45.8 22 2
26.9 5.5
58.6 64.9
56.3
26.1
4.8
68.6
56.5
35.8
6.4
77.8
Ge 5 nm/ Ag 25 nm Ge 5 nm/ Ag 50 nm a-C:H 5 nm a-C:H 5 nm/ Ag 10 nm a-C:H 5 nm/ Ag 25 nm a-C:H 5 nm/ Ag 50 nm
Figure 3. PEDOT-PSS contact angle snapshots showing the drop shape in contact with the following bilayers: (a) Cr 5 nm/Ag 25 nm, (b) glass/Ag 25 nm, (c) a-C:H 5 nm/Ag 25 nm, and (d) Ge 5 nm/ Ag 25 nm.
is known as Young’s equation: γLV cos θY ¼ γSV - γSL
ð3Þ
The surface free energy of our samples was determined using the Owens and Wendt model.10 This method considers γS as a sum of two components such that: p
γS ¼ γdS þ γS
ð4Þ
γSd
gives the long-range dispersion (Lifshitz-van der where Waals) component and γSp denotes the short-range polar (hydrogen bonding) component of surface free energy. To calculate the surface free energy, the measured contact angles were introduced in the Owens-Wendt eq 5, where γdL is the energy of the dispersive liquid and γpL is the energy of the polar liquid, known from the literature and reported in Table 2: qffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffi p p ð5Þ γLV ð1 þ cos θY Þ ¼ 2 γdS γdL þ 2 γS γL 3614
dx.doi.org/10.1021/la104760a |Langmuir 2011, 27, 3611–3617
Langmuir
ARTICLE
Table 2. Values of the Lifshitz-van der Waals (γdL) and Polar (γpL) Components of Liquid Surface Tension (γL) for Different Test Liquids γL (mJ 3 m-2)
γdL (mJ 3 m-2)
γpL (mJ 3 m-2)
PEDOT-PSS (polar)6
71.2
40.6
30.6
deionized water (polar)24
72.8
21.8
51.0
fomamide (polar)24
58.0
39.0
19.0
diiodomethane (apolar)24
50.8
50.8
0.0
liquids
Table 3. Total Free Surface Energy (γS), and Its Dispersive (γdS) and Polar (γpS) Components Calculated Using the Owens-Wendt Model substrate
γS (mJ 3 m-2)
γdS (mJ 3 m-2)
γpS (mJ 3 m-2)
glass/Ag 25 nm
58.6
34.8
23.9
Cr 5 nm/Ag 25 nm
54.0
36.4
17.6
Ge5 nm
71.3
29.9
41.4
Ge 5 nm/Ag 10 nm
71.3
28.9
42.4
Ge 5 nm/Ag 25 nm
70.2
29.9
41.1
Ge 5 nm/Ag 50 nm
68.1
32.6
35.5
a-C:H 5 nm
49.6
32.0
17.7
a-C:H 5 nm/Ag 10 nm a-C:H 5 nm/Ag 25 nm
55.1 57.6
42.6 37.3
12.6 20.3
a-C:H 5 nm/Ag 50 nm
53.1
39.4
13.7
Replacing the contact angles and the liquid surface tension components in eq 5, the surface free energy (γS) with its dispersive (γdS) and polar (γpS) components is determined. These data are reported in Table 3, which, together with Table 1, show that the silver deposited on the germanium nucleation layer exhibits the highest values of the surface free energy and polar component as well as the highest wettability with PEDOT-PSS and water, while the free energy and the polar component for the other bilayers are similar. From Table 3, it is also possible to note that in the case of Ge based bilayers these parameters exhibit a dependence on the Ag film thickness which is much less evident in the case of a-C:H based silver films. These measurements suggest that the different behavior of the Ge/Ag bilayer could be ascribed to a surface effect, so further investigations on the 25 nm silver films deposited on the germanium nucleation layers and on the glass substrates have been performed by using X-ray photoelectron spectroscopy (XPS). The XPS survey spectra of these samples have been recorded for both the pristine and sputtered surface (Figures 4 and 5). For both samples, the spectra of the pristine surfaces highlight the presence of carbon, oxygen, and silver. However, the survey spectra realized on the Ge/Ag sample showed also a small concentration of germanium at the surface (Figure 5) that suggests the diffusion of Ge in the Ag layer. Moreover, the oxygen concentrations (atom %) measured on the analyzed samples, and reported in Table 4, highlight the presence of oxygen at the surface of silver due to its exposure to the atmosphere.25 In the first sample, this element is located only on the surface: Ag deposited on glass does not exhibit the presence of oxygen 3 nm below the surface. In contrast, the presence of oxygen is still evident on the sputtered Ge/Ag surface, suggesting the presence of an oxide, as will be explained in the following.
Before sputtering, the high-resolution spectra of the Ag3d peak components, Ag3d5/2 and Ag3d3/2, showed a shoulder on the side of the weak binding energies corresponding to the presence of silver oxide phase. A deconvolution of the Ag3d peak is shown in Figure 4b. The binding energy corresponding to the silver metallic form of the Ag3d5/2 peak has been found at 368.2 eV, while the binding energy peak corresponding to the Ag oxide phase has been found at 367.7 eV, in fairly good agreement with literature results for Ag2O component.26 In the same sample, also the O1s peak of the pristine surface shows two components, as shown in Figure 4c: the first one, found at 530.6 eV, corresponds to a Ag-O bound, while the second, found at 531.75 eV, is attributed to CdO bound. The concentrations of Ag and C bound oxygen are 5.3 and 2.0% respectively, in good agreement with the presence of a native silver oxide at the thin film surface. The same oxidation degree of silver can be observed also on the Ag/Ge pristine surface. Moreover, for the pristine Ge/Ag sample, the surface oxidation was significantly greater (22 atom %) and the survey spectra show the evidence of germanium segregation (both in metal and oxide form) at the silver layer surface. The evidence of Ge spectral peaks also in the sputtered surface can be attributed to a gradient diffusion of germanium and its oxides in the silver film deposited over the nucleation layer (Figure 5a). The amount of germanium elemental and oxidized forms (GeOx and GeO2) on the surface is around 7 atom % before sputtering the surface. The high-resolution spectrum of germanium (Figure 5b) has been recorded after the sample’s surface sputtering. The Ge3d peak shows three components: the first is located at 28.9 eV and corresponds to the germanium metallic form (1.2 atom %). The other two components correspond to different germanium oxides and exhibit maximum intensity at, respectively, 30.6 and 32.5 eV with a characteristically large full width at halfmaximum. The component situated at 32.5 eV is attributed to GeO2 (1.7 atom %), and the one at 30.6 eV to GeOx (x < 2) (0.7 atom %) (Figure 5b). As we previously discussed, silver has a high affinity with germanium and in our deposition conditions can form an alloy in the first deposition stages.27 The fact that germanium and its oxides could migrate at the surface of an alloy has already been discussed in literature.28 These oxides can be easily produced in contact with a small amount of oxygen or moisture. As reported in the literature, the most probable product due to germanium interaction with oxygen is crystalline or vitreous following the reaction: Ge (solid) þ O2 (gas) f GeO2 (solid).29 In the first stage of Ge oxidation, a suboxide GeOx (x < 2) structure is formed, breaking the Ge-Ge bonds which were attacked by oxidizing agents such as water vapor and oxygen. Subsequently, GeO2 is formed by penetration and diffusion of oxygen into the oxidized Ge layer.30 This can explain the germanium oxide presence at the silver surface and its elevated oxygen concentration. The amount of oxygen in the outer layers of the samples and the depth of oxygen incorporation can influence the magnitude of the polar share.31 By these considerations, we can infer that the high oxygen electronegativity is responsible for the high polarity at the interfacial region (solid-liquid interface) in the Ge/Ag bilayer, with it being known that the germanium covered by its native oxide is very hydrophilic.32 On the contrary, we observe that there is no big difference between the surface free energy and polar component of the 25 nm silver film grown on simple glass substrate or on the Cr and a-C:H nucleation layers (Table 3). This can be explained by 3615
dx.doi.org/10.1021/la104760a |Langmuir 2011, 27, 3611–3617
Langmuir
ARTICLE
Figure 4. (a) XPS survey spectra of Ag (25 nm) on glass. The black curve depicts the pristine Ag surface, and the blue curve the Ag surface after a 3 min Arþ sputtering. (b) High-resolution spectrum of the Ag3d components of the pristine surface after deconvolution and background subtraction. It is clearly visible the Ag oxide component in the weak energy side. (c) High-resolution spectrum of the O1s component of the pristine surface after deconvolution and background subtraction. Also in this case the presence of a Ag oxide component is clearly observable at the position 530.6 eV, while the CdO component (531.75 eV) can be ascribed to a sample surface contamination. Both the Ag3d and O1s high-resolution spectra confirm the presence of a Ag oxide at the Ag sample surface.
Figure 5. (a) XPS survey spectra of Ag (25 nm) on Ge (5 nm). The black curve depicts the pristine Ag surface, and the blue curve the Ag surface after a 3 min Arþ sputtering. The Ge3d peak is visible in both spectra. (b) High-resolution spectrum of the Ge3d components of the sputtered surface. The Ge metallic component is shown in the narrow peak at 28.9 eV. The large peak on the high energy side is attributed to the overlapping of the GeOx and GeO2 oxides with peaks at, respectively, 30.6 and 32.5 eV.
3616
dx.doi.org/10.1021/la104760a |Langmuir 2011, 27, 3611–3617
Langmuir
ARTICLE
Table 4. Oxygen Elemental Concentrations (atom %) at the Ag Surface of Pristine and Sputtered Ag Based Bilayers Samples oxygen (%) substrate
nonsputtered surface
sputtered surface
7.3
0.0
22.0
3.6
glass/Ag 25 nm Ge 5 nm/Ag 25 nm
the fact that the oxide presence at the silver surface, Ag2O, has a slight hydrophobic nature and interacts weakly with coming water molecules because of the small quantity of surface OH groups.33 So, the reason for the low values of the surface free energy and the polar component of these samples can be ascribed to the hydrophobic behavior of the surface silver oxide and to its weak hydrogen bonding.
’ CONCLUSION The Ag thin anode for our OLED microcavity needs to be smooth and to show a high wettability for an aqueous solution of PEDOT-PSS. The first property is highly related to Ag growth mechanisms: layer-by-layer deposition has to be favored to form the smoothest surface. This point excludes the use of Cr as the nucleation layer, as the presence of oxide at its surface leads to Ag growth by 3D islands and consequently to a high roughness. In contrast, a-C:H and Ge lead to a Volmer-Weber growth mode with very small and dense nucleation sites that allows to obtain a very smooth and planar surface very similar to a layer-by-layer growth mode. Finally, we demonstrate that the use of a thin germanium (5 nm) nucleation layer gives also a hydrophilic character to the surface of silver thin film due to Ge and Ge oxides diffusion through the Ag layer. Thus, this property can be useful to obtain uniform deposition of polar liquids such as PEDOTPSS. So, a Ge 5 nm/Ag 25 nm bilayer has been successfully used as a well-suitable anode for the fabrication of a long lifetime, large area (12.5 mm2), and short-circuit-free blue bottom emitting strong microcavity OLED.8 ’ AUTHOR INFORMATION Corresponding Author
*Address: Universite Paul Sabatier-LAPLACE, 118, Rue de Narbonne, 31062 Toulouse Cedex 4, France. Telephone: þ39 348 3626473. Fax: þ33(0)6 63540984. E-mail: patrizia.
[email protected] and offi
[email protected].
’ ACKNOWLEDGMENT The authors thank Mr. Benoit Schlegel of Service Instrumentation-LAPLACE, Univ. Paul Sabatier, Toulouse (France) for the chromium deposition and the “Lumiere et Matiere” group of the LAPLACE for the support in the research activities. The work at OR-EL.doo was partially supported by Slovenski Podjetniski Sklad under Contract SIA10/00020.
(3) Morandi, V.; Marabelli, F.; Amendola, V.; Meneghetti, M.; Comoretto, D. Adv. Funct. Mater. 2007, 17, 2779–2786. (4) Jonas, K. L.; Von Oeynhausen, V.; Bansmann, J.; Meiwes-Broer, K. H. Appl. Phys. A: Mater. Sci. Process. 2006, 82, 131–137. (5) Pais, A.; Banerjee, A.; Klotzkin, D.; Papautsky, I. Lab Chip 2008, 8, 794–800. (6) Vacca, P.; Petrosino, M.; Miscioscia, R.; Nenna, G.; Minarini, C.; Della Sala, D.; Rubino, A. Thin Solid Films 2008, 516, 4232–4237. (7) Logeeswaran, V. J.; Kobayashi, N. P; Islam, M. S.; Wu, W.; Chaturvedi, P.; Fang, N. X.; Wang, S. Y.; Williams, R. S. Nano Lett. 2009, 9 (1), 178–182. (8) Melpignano, P.; Cioarec, C.; Clergereaux, R.; Gherardi, N.; Villeneuve, C.; Datas, L. Org. Electron. 2010, 11, 1111–1119. (9) Kihel, M.; Clergereaux, R.; Escaich, D.; Calafat, M.; Raynaud, P.; Sahli, S.; Segui, Y. Diamond Relat. Mater. 2008, 17, 1710–1715. (10) Owens, D. K.; Wendt, R. C. J. Appl. Polym. Sci. 1969, 13 (8), 1741–1747. (11) Gardiner, T. M.; Stiddard, M. H. B. Thin Solid Films 1981, 77, 335–340. (12) Jung, Y. S. Appl. Surf. Sci. 2004, 221, 281–287. (13) Antoine, A. M.; Drevillon, B. J. Appl. Phys. 1988, 63, 360–367. (14) Antoine, A. M.; Drevillon, B.; Roca i Cabarrocas, P. J. Appl. Phys. 1987, 61 (7), 2501–2508. (15) Anders, A.; Byon, E.; Kim, D. H.; Fukuda, K.; Lim, S. H. N. Solid State Commun. 2006, 140, 225–229. (16) Hu, M.; Noda, S.; Komiyama, H. Surf. Sci. 2002, 513, 530–538. (17) Ohring, M. Materials Science of Thin Films. Deposition and Structure, 2nd ed.; Academic Press, San Diego, 2002. (18) Campbell, C. T. Surf. Sci. Rep. 1997, 27, 1–111. (19) Jacobs, J. A.; Testa, S. M. Overview of Chromium(VI) in the Environment: Background and History. In Chromium(VI) Handbook; Jacobs, J. A., Guertin, J., Avakian, C., Eds.; CRC Press: Boca Raton, FL, 2005; pp 1-23. (20) Clergereaux, R.; Escaich, D.; Martin, S.; Gaillard, F.; Raynaud, P. Thin Solid Films 2005, 482 (1-2), 216–220. (21) Karakaia, I.; Thompson, W. T. Bull. Alloy Phase Diagrams 1988, 9 (3), 226–227. (22) Renji, Z.; Li, L.; Ziquin, W. J. Mater. Sci. 1993, 28, 1705–1724. (23) Oughaddou, H.; Sawaya, S.; Goniakowski, J.; Aufray, B.; Le Lay, G.; Gay, J. M.; Treglia, G.; Biberian, J. P.; Barrett, N.; Guillot, C.; Mayne, A.; Dujardin, G. Phys. Rev. B 2000, 62 (24), 16653–16656. (24) Ja nczuk, B; Biazopiotrowicz, T.; Zdziennicka, A. J. Colloid Interface Sci. 1999, 211, 96–103. (25) Bartell, F. E.; Smith, J. T. J. Phys. Chem. 1953, 57 (2), 165–172. (26) Abe, Y.; Hasegawa, T.; Kawamura, M.; Sasaki, K. Vacuum 2004, 76, 1–6. (27) Kovalskiy, A.; Miller, A. C.; Jain, H.; Mitkova, M. J. Am. Ceram. Soc. 2008, 91 (3), 760–765. (28) Ferri, D.; B€urgi, T.; Baiker, A. J. Phys. Chem. B 2001, 105, 3187–3195. (29) Peksheva, N. P.; Strukov, V. M. Russ. Chem. Rev. 1979, 48 (11), 1092–1108. (30) Park, K.; Lee, Y.; Lee, J.; Lim, S. Appl. Surf. Sci. 2008, 254, 4828–4832. (31) Lugscheider, E.; Bobzin, K.; B€arwulf, St.; Hornig, Th. Surf. Coat. Technol. 2000, 133-134, 540–547. (32) Janssen, D.; De Palma, R.; Verlaak, S.; Heremans, P.; Dehaen, W. Thin Solid Films 2006, 515, 1433–1438. (33) Kuroda, Y.; Watanabe, T.; Yoshikawa, Y. Langmuir 1997, 13, 3823–3826.
’ REFERENCES (1) Rill, M.; Plet, C.; Thiel, M.; Wegener, M. Quantum Electronics and Laser Science Conference (QELS), San Jose, California May 4, 2008. (2) Juodkazis, S.; Yamaguchi, A.; Ishii, H.; Matsuo, S.; Takagi, H.; Misawa, H. Jpn. J. Appl. Phys. 2001, 40, 4246–4251. 3617
dx.doi.org/10.1021/la104760a |Langmuir 2011, 27, 3611–3617
