Surface Chemical and Leakage Current Density Characteristics of
Nanocrystalline Ag–Ba
0.5Sr
0.5TiO
3Thin Films
Kampurath P. Jayadevan, Chi-Yi Liu, and Tseung-Yuen Tseng
wDepartment of Electronics Engineering and Institute of Electronics, National Chiao-Tung University, Hsinchu 300, Taiwan Nanocrystalline xAg–(1x)Ba0.5Sr0.5TiO3(Ag–BST, 0rxr0.1,
where x is the mole fraction of Ag) thin films have been depos-ited on Pt/Ti/SiO2/Si substrates by a sol–gel method. The films have been characterized by X-ray diffraction (XRD), scanning electron microscopy, and X-ray photoelectron spectroscopy (XPS). The core-level XPS of oxygen (O1s) of the Ag–BST films indicate that an optimum amount of Ag (x 5 0.02 or 2 mol%) enhances the binding energy of oxygen, possibly through a mechanism in which the electrophilic oxygen dissociates from the Ag surface and fills the oxygen vacancies. Similarly, the binding energy of Ag (Ag3d) shows a shift toward a higher value with increasing Ag up to 4 mol%, probably because of the chemical shift of Ag in BST along the surface layers, surface relaxation, or changes in the Fermi level of small Ag particles in the solid solution range of Ag in BST films (xr0.04). The leakage current density of 2 mol% Ag-added BST (B106 A/cm2) is less by about an order of magnitude than pure BST at an electric field of 200 kV/cm. A defect model is proposed to explain the observed leakage current density of Ag–BST films satisfactorily.
I. Introduction
T
HEhigh-permittivity (Ba,Sr)TiO3(BST) thin films have beenintensively studied in recent years as a potential dielectric material for the development of capacitors in microwave devices and future ultra high-density dynamic random access memories (DRAM) because of their low dielectric loss, high breakdown field strength, tunable dielectric constant, low leakage current density, and high reliability.1 The present research interests in this class of high-permittivity perovskite oxides are directed toward achieving improved performance of miniaturized BST capacitors by tuning the processing parameters,2 reducing the
leakage current density by suitable dopants,3fabricating
com-posite or multilayer capacitors and superlattices with enhanced properties,4and adopting innovative procedures for the integra-tion of high-permittivity films directly on silicon (Si).5Among the various factors that lead to high leakage current density in BST thin films, oxygen vacancy formation at the film/electrode interface has been identified as a major cause for current con-duction and performance degradation of BST capacitors.6,7 Deposition or annealing of BST films at selected partial pres-sures of oxygen has been adopted as a suitable method to reduce the oxygen vacancies, which are omnipresent in an equilibrium perovskite structure, as demonstrated by Tsai et al.6 for the sputter-deposited Ba0.5Sr0.5TiO3(BST) thin films at various
ox-ygen mixing ratios with argon (O2/Ar1O2). Our present
inves-tigation into the silver (Ag)–BST system has been motivated by
a recent study by Srivastava et al.,8which indicates a significant
reduction in leakage current density of pulse-laser deposited (Ba,Sr)TiO3thin films that are prepared from a BST target with
5 wt% Ag. The proposed explanation for the reduction in leak-age current can be summarized as follows: the unstable silver oxides (AgxO) that are formed during the film deposition and
annealing decompose to Ag, which remains intact with a low drive for oxidation9in the BST under ambient conditions. The irreversible decomposition of AgxO
10
would compensate oxygen vacancies; thereby, a less defective BST would form along with a reduction in leakage current.8However, an experimental study to verify this mechanism has not been reported so far. In order to obtain a possible experimental evidence for the oxygenation of BST films in the presence of Ag, we have carried out an X-ray photoelectron spectroscopic (XPS) investigation of sol–gel-de-rived nanocrystalline Ag–BST thin films. An optimum amount of Ag (2 mol%) is found to enhance the binding energy of sur-face oxygen and the O1s peak intensity. This observation is fur-ther supported by the leakage current density characteristics that show a reduction at an electric field higher than 150 kV/cm for 1 and 2 mol% Ag when compared with pure BST. The grain size and phase have been characterized by X-ray diffraction (XRD), while scanning electron microscopic (SEM) investigation of the surface of the Ag–BST films provides morphology characteris-tics. Sol–gel processing is undoubtedly one of the easiest meth-ods for the deposition of thin films with excellent control over the composition of the reactive components in solution.4It can also be used as an economically affordable pilot-scale technique for the fabrication of novel solid solutions or composites. The nanocrystalline Ag–BST thin films have been prepared by an acetate–precursor sol–gel route.11The method has been chosen because of its relatively simple approach for obtaining a clear and homogeneous solution for deposition.
II. Experimental Procedure
The scheme for the preparation of the precursor solutions for xAg–(1x)Ba0.5Sr0.5TiO3(Ag–BST), where x is the mole
frac-tion of Ag in the BST thin films, is summarized as follows: the Ag contents in the Ag–BST thin films were selected as x 5 0.01, 0.02, 0.03, 0.04, and 0.1 (1, 2, 3, 4, and 10 mol%), with the cor-responding variations in the BST precursor as 0.99, 0.98, 0.97, 0.96, and 0.90. For convenience, the composition of the films would be presented henceforth as Ag(X)–BST, where X refers to the mol% of Ag in BST films. The stoichiometric amounts of the reactants, Ba(OAc)2, Sr(OAc)2, Ti(i-OPr)4, and Ag(OAc) (OAc
acetate and i-OPr isopropoxide), were dissolved in 25 mL acetic acid at about 901C. The solution was stabilized by the addition ofB8.5 mL ethylene glycol so as to maintain the ratio of acetic acid to ethylene glycol as 3:1, which would lead to a uniform microstructure as reported by Tahan et al.11The clear and ho-mogeneous solutions (0.15 M) obtained after 60 min of stirring were kept at room temperature and aged for about 12–24 h before spin coating. An undoped BST precursor solution was prepared to deposit the reference BST film. The solutions were filtered using a syringe attached with a 0.2 mm filter onto
Journal
DOI: 10.1111/j.1551-2916.2005.00441.x r2005 The American Ceramic Society
2456 I. W. Chen—contributing editor
This work was supported by the National Science Council (NSC) of Republic of China under the Project No. NSC 92-2215-E009-016.
w
Author to whom correspondence should be addressed. e-mail: [email protected] Manuscript No. 20260. Received August 20, 2004; approved March 5, 2005.
Pt (100 nm)/Ti (5 nm)/SiO2(500 nm)/Si substrates and then spin
coated in two steps (1000 rpm for 15 s and 4000 rpm for 60 s). The coated layers were dried at 2001C for 10 min and then pyrolyzed at 5001C for 30 min. The coating and pyrolysis steps were repeated 11 times to obtainB210 nm thickness for each composition. The composition of Ag–BST films has not been modulated by the graded layer deposition technique in this study. During the deposition of each layer, a small portion of the substrate was masked so as to preserve the bottom Pt layer as an electrode for measurements. The films were finally an-nealed at 7001C for 2 h. The anan-nealed films were characterized for phase formation and grain size using XRD (Rigaku D-Max/ IIB X-ray Diffractometer, Rigaku, Japan, CuKa l 5 0.15405 nm, 0.021 2y step). The microstructure and cross-section of the films were examined using a SEM (Hitachi S4700, Hitachi, Japan). XPS were recorded under ultra high-vacuum conditions using a VG Microlab 310F spectrometer (VG Scientific UK) with AlKa radiation of 1486.6 eV energy. For electrical measurements of the films, metal–insulator–metal (MIM) structures were fabri-cated by depositing 100 nm thick Pt top electrodes with a di-ameter of 250 mm by electron beam evaporation through a shadow mask. The current–voltage (I–V) characteristics were recorded by using an Agilent 4155C semiconductor parameter analyzer (Agilent Technologies, Japan) under bias voltage in the range from5 to 5 V.
III. Results and Discussion (1) Phase, Crystallinity, and Grain Size
The spin coated and thermally processed Ag–BST thin films on Pt/Ti/SiO2/Si substrates were subjected to phase
characteriza-tion and crystallite size determinacharacteriza-tion using XRD. Figure 1 de-picts the XRD patterns of the prepared films with Ag content up to 4 mol%. All the peaks are indexed to the BST phase, with no appearance of a second phase in the sample. The sharp (110) and (200) peaks of BST phase indicate that the films have good crystallinity, with the average grain size (tXRD) varying from
B19 nm for Ag(2)–BST (the number in brackets refers to mol% of Ag in BST according to the scheme described under the ex-perimental section) toB25 nm for Ag(4)–BST as determined by using the Scherrer equation
tXRD¼ 0:9l=b cos yB (1)
where l is the wavelength of the CuKa radiation, b is the full-width at half-maximum, (FWHM) and yBis the diffraction
an-gle. The amounts of Ag added in the present study do not seem to suppress the average BST crystallite size. No significant var-iations in peak positions are detected as the Ag content is in-creased up to 4 mol%.
(2) Surface Morphology
The microstructures of the prepared nanocrystalline Ag–BST films have been examined by SEM. Surface morphology char-acteristics of a selected set of films, undoped BST, Ag(2)–BST, and Ag(4)–BST, and cross-sectional SEM picture for Ag(2)– BST are shown in Fig. 2. Upon 2 mol% addition of Ag, signif-icant modification of the microstructure is observed with the realization of a highly dense and uniform microstructure of an Ag(2)–BST film (Fig. 2(b)) in comparison with undoped BST (Fig. 2(a)). The average grain or agglomerate diameter (dSEM) of
Ag(2)–BST film isB25 nm, which is comparable with the grain size obtained from the XRD analysis. The improvement of microstructure of Ag(2)–BST as compared with undoped BST could have been caused by the increased presence of Ag along the grain boundaries8close to the surface layers that lead to the formation of dense layers for the Ag(2)–BST film annealed at 7001C. The possibility of a significant reduction in the actual melting point of Ag (B9601C)9
in the nanometer regime or an increase in the mobility of Ag in the shallow potential wells along the grain boundaries in the layers close to the surface might cause a quasi-melting12of the small Ag particles, as the
direction of diffusion of Ag is toward the surface. The analysis of binding energy of Ag at the surface will be discussed in the next section and based on the leakage current density charac-teristics of Ag–BST films, a possible microstructure for the Ag– BST film will be proposed. The large island-like agglomerate formation in Ag(4)–BST (Fig. 2(c)) with significantly high crys-tallinity could have been caused by the further enhancement of grain boundary diffusion, as the mobility of Ag increases along the grain boundaries close to the surface. The dense and fine-grained microstructure of Ag(2)–BST film is also illustrated as a cross-sectional SEM picture in Fig. 2(d). The average thickness of the film is 210710 nm, as obtained from the cross-sectional SEM analysis (Fig. 2(d)).
(3) Surface Chemical States of Ag and Oxygen (O) XPS corresponding to oxygen (O1s) in Ag–BST nanocrystalline thin films up to 2 mol% Ag and silver (Ag3d) up to 10 mol% Ag are presented in Fig. 3(a) and (b), respectively. The binding en-ergy of O1s corresponding to undoped BST is 528.2 eV, which is less than the reported values of oxygen that range between 529 and 531 eV in perovskite (Ba,Sr)TiO3,13–16possibly because of
oxygen deficiency in our samples. As the Ag content is increased, there is a systematic shift of O1s peaks to higher binding energy (529.6 eV) with increasing intensity up to 2 mol% Ag (Ag(2)– BST). An example that illustrates the evidence of oxygenation
20 30 40 50 60 Ag(4)-BST Ag(3)-BST Ag(2)-BST Ag(1)-BST BST (211) (200) Pt (110) (100)
Intensity (Arb. Units)
2Θ (Degrees)
Fig. 1. X-ray diffraction (XRD) patterns of undoped BST, Ag(1)–BST,
Ag(2)–BST, Ag(3)—BST, and Ag(4)–BST films crystallized at 7001C on
Pt/Ti/SiO2/Si substrates. The number in brackets refers to mol% of Ag.
Fig. 2. Scanning electron micrographs (SEM) of the surface
morph-ologies of (a) undoped BST (b) Ag(2)–BST, (c) Ag(4)–BST films, and (d)
cross-section forB210 nm thick Ag(2)–BST film to illustrate the dense
microstructure. The scale bar for (a)–(c) is 500 nm and 1 mm for (d).
from the XPS analysis is the shift of the O1s profile toward higher binding energy with an increase in the intensity that is observed for the Si substrate because of the incorporation of oxygen as the substrate is heated.17In our study, the
enhance-ment of O1s signal of BST could have been caused by a catalytic oxidation in the presence of Ag. The reactivity of oxygen at the surface of Ag has been a topic of immense interest for the ex-ploitation of Ag in catalysis, for the oxidation of organic mol-ecules and removal of toxic gases from the environment.18–23 The potential for catalytic oxygenation depends on the nature of the oxygen that is adsorbed on the Ag surface. The selective oxygenation of organic molecules using Ag as a catalyst implies that the oxygen adsorbed on the surface of Ag would be elect-rophilic, covalent, and weakly polarized.18When Ag is formed as a solid solution with BST, the electrophilic oxygen adsorbed on the surface of Ag would find a potentially stable site for oc-cupancy, probably driven by the electrostatic potential in asso-ciation with titanium ion (Ti41)24in the BST. The observation that the O1s peaks in BST, Ag(1)–BST, and Ag(2)–BST do not show any splitting or distinct higher energy peak25and the
com-parable FWHM of O1s peaks (1.7–1.8 eV) indicate the oxygen species as O2ion in the BST or Ag–BST films. Therefore, the shift in the O1s peak and the increase in the intensity could have been caused by the filling of the oxygen vacancies, which is similar to the XPS study of the oxygenation of an oxygen-de-ficient MgO surface as reported by Peterka et al.26The increase in the oxygen content with increasing Ag in BST can be estimated from the XPS data and the value is anB2 at.% in-crease in oxygen for Ag(2)–BST in comparison with pure BST. As the Ag content is increased to 4 mol%, the maximum binding energy value remains on the average in the shifted range. However, some irregularity of peak position in the binding en-ergy curve has been observed beyond 2 mol% Ag (not shown here), which could probably be because of the onset of many competitive surface reactions such as Ag–Ag and Ag–BST in-teractions or changes in the electrostatic potential that depend on the coordination surrounding the titanium ion (Ti41)24in the BST. A detailed surface analysis of the Ag–BST films would help to delineate the various competitive interactions at the Ag–BST surface.
As the O1s peaks do not show any splitting and exhibit only a single maximum with increasing Ag in BST, we conclude that there is no silver oxide formation in the film. The absence of silver oxides is further confirmed by the Ag3d spectra, which will be discussed later in this section. The broad hump with a low relative intensity at a higher binding energy of O1s may be be-cause of the oxygen that binds to carbonaceous impurities13,14as this is observed in all samples including undoped BST. The en-hanced oxygen binding energy in Ag–BST films could be related to the decomposition of unstable silver oxides9or short-lived transient compounds (½Ag::O, oxygen chemisorbed on the sur-face of Ag)27 well below the annealing temperature (7001C) of the BST films. In our study, the ½Ag::O species (weakly adsorbed electrophilic oxygen species) are more likely to be
present, as the deposition is not carried out in the presence of energetic oxygen species in a plasma (e.g., oxidation of an Ag foil to Ag2O in an oxygen microwave plasma28or pulsed laser
ablation of an Ag–BST target in an oxygen atmosphere8). The
½Ag::Othat is formed from the Ag precursor in the film would decompose according to the reaction
½Ag::OðsÞ ! AgAgðsÞ þ 1=2O2ðgÞ (2) The filling of the oxygen vacancies can be written according to the Kro¨ger–Vink notation as
1=2 O2ðgÞ þ VO þ 2e 0¼ Ox
O (3)
As the sticking coefficient of oxygen on Ag surface is low,27 and the favorable energetic path is filling of oxygen vacancies, the dissociated oxygen from ½Ag::O (weakly electrophilic ox-ygen on Ag) may be involved in a surface chemical binding in BST films, which is manifested as a chemical shift in the O1s XPS spectrum as described earlier. The estimation of an increase in oxygen by about 2 at.% for Ag(2)–BST film when compared with pure BST indicates the possibility of reactions (2) and (3). The evidence for Ag at the surface is derived from Ag3d5/2
(BEB369.5–370.5 eV) and Ag3d3/2(BEB375.5–376.5 eV) levels
of Ag3d XPS spectra (Fig. 3(b)) of Ag–BST films that have binding energy values closer to those reported for pure Ag (3d5/2
B368 eV and 3d3/2B374 eV),29but higher than those of Ag2O
(3d5/2B367.6 eV and 3d3/2B373.6 eV), 29
AgO (3d5/2B367.4 eV),
and Ag2CO3 (3d5/2B367.6 eV). 27
The higher binding energy value of Ag in Ag–BST than in its pure form may be attributed to the difference in the chemical environment of Ag, the relax-ation effect, or the Fermi level shift. The total shift in the bind-ing energy (DEB) can be expressed as
DEB¼ DECþ DERþ DEF (4)
where DECis the shift in the binding energy because of the
dif-ference in the chemical configuration (chemical shift), DERis the
shift because of the contribution from the relaxation, and DEFis
the Fermi level contribution to the total binding energy shift.21
The DECin Eq. (4) represents the change in the electrostatic
potential between the initial state and the final state because of the chemical shift.30Although the solubility of Ag in stoichio-metric BST-related bulk perovskite oxides is stated to be negli-gible,31,32the shifts of the Ag3d peaks in the XPS spectra to higher binding energy values as the Ag content is increased to 4 mol% Ag (Fig. 3(b)) indicate that along the surface layers of the Ag–BST film, a small amount of Ag could be soluble as the Ag radius would be comparable with (Ba,Sr)-site radius.33 There-fore, the oxygen vacancy gets compensated through a Schottky equilibrium at the surface layers. The DECwould be more
pro-nounced when Ag is present in low concentration in the solid solution xAg–(1x)Ba0.5Sr0.5TiO3(xr0.04) because of the
de-fective coordination environment close to the surface, difference in size between Ag and the (Ba,Sr)21ions, and the reduced cou-lomb interaction of substituted Ag with the oxide ions. The XPS analysis that confirms the solid solubility of Ag in perovskite BST has not been reported in the literature, but that of the Ag-doped perovskite LaCoO3indicates that the neutral Ag exists
at the surface of LaCoO3 monoliths on catalyst support.34
The study does not show any positive shift in the binding energy. The binding energy of Ag3d level shifts to that of Ag1ion, when
the surface of the LaCoO3is etched to a depth of a few
nano-meters. The presence of Ag1ion in the bulk has been taken as the probable evidence for the A-site substitution in LaCoO3.
The Ag–BST system that we have investigated is nanocrystalline with the possibility of defective anionic sites at the surface that would promote the dissolution of Ag35along the surface layers of BST. Further detailed investigation of nanocrystalline thin films of Ag–BST would be required to understand the exact electronic state of Ag in BST.
360 365 370 375 380 385
(b)
Intensity (Arb. Units)
Binding Energy (eV)
520 525 530 535 540 545
(a)
Intensity (Arb. Units)
Binding Energy (eV)
Fig. 3. X-ray photoelectron spectra (XPS) of (a) oxygen (O 1s), and (b)
From the XPS results of our study on Ag–BST films and those on Ag–LaCoO3in which neutral Ag has been identified at
the surface,34it can be believed that the Ag diffuses toward the
surface. The dense and uniform layer that we observe in the ex-amination of surface morphology of Ag(2)–BST and large island-like agglomerates for Ag(4)–BST film in Figs. 2(b) and (c), respectively, could have been because of the presence of Ag, which has two roles when it diffuses to the surface layers: (1) modification of surface morphology of BST as derived from the SEM analysis and illustrated in Figs. 2 (b) and (c), and (2) dis-solution of Ag in BST grains close to the surface as obtained from the XPS analysis (Fig. 3(b)).
The second term in Eq. (4), DER, is the relaxation energy that
is required to excite a core-level electron to the valence state to neutralize the localized holes that are associated with the excited metallic atoms so that the translational symmetry of the infinite solid is preserved.36The contribution from DERwill be
negligi-ble if there is pronounced screening effect from the valence elec-trons that helps to maintain the local charge neutrality. For very small particle sizes of Ag that are likely to be present in the Ag– BST films for low mole fractions of Ag (xr0.04), the positively charged final state (excited state) would not be well screened and would therefore contribute to the shift of the binding energy in the form of relaxation energy (DER). For the BST film with 10
mol% Ag (x 5 0.1), the shift toward binding energy that is lower than the samples with x 5 0.01, 0.02, 0.03, and 0.04 indicates an increased Ag–Ag interaction and hence an improvement in screening.
In addition to DECand DER, there could also be a
contribu-tion from the DEFin Eq. (4), as the original Fermi level of Ag
could shift because of small particle sizes of Ag that is present in the Ag–BST solid solution. For low mole fractions or an opti-mum amount of Ag in BST (xr0.04), the shift in the Fermi level of Ag would contribute to the DEB, while undergoing an
equi-libration process with BST through space charge effect.37For the BST film with x 5 0.1, the increased Ag–Ag interaction would reduce the contribution of DEFto the DEB. The three
energy terms, DEC, DER, and DEF, would add to give a resultant
shift in the Ag3d spectra (Fig. 3(b)). A detailed surface analysis would be required to quantify these effects and identify the exact mechanism in this system. From our XPS analysis, we suggest that the oxygenation and Schottky equilibration of oxygen va-cancies are probably operative in Ag–BST films up to a small amount of Ag.
The absence of oxide phases of Ag is further confirmed by the Ag3d XPS spectrum of a 10 mol% Ag-added BST (Ag(10)– BST) (Fig. 3(b)). The spectrum indicates well-defined 3d5/2and
3d3/2 peaks with the respective binding energy maxima of
B368.8 and B374.7 eV that are close to those of pure Ag.27,29
The shift toward binding energy value of pure Ag may be be-cause of the increase in the Ag–Ag interaction that leads to the formation of Ag islands or clusters on the surface that form as a separate phase from BST. An additional evidence for the diffu-sion of Ag toward the surface could be derived from our ca-pacitance–voltage (C–V) measurements of BST, Ag(1)–BST, and Ag(2)–BST films, which show an increasing asymmetry in the C–V curves with increasing Ag because of the surface mod-ification of the Ag–BST films by Ag as reported in a recent letter.38
(4) Leakage Current Density Characteristics
The leakage current densities of MIM capacitors of configura-tion, Pt/Ag(0y4)–BST/Pt have been investigated and presented in Fig. 4. The asymmetry of the current density curves with higher leakage currents under reverse bias indicates the differ-ence in the interface characteristics of the top and bottom elec-trodes with the film. The major contributing factor to the leakage current density of undoped BST films is believed to be because of the electrons associated with the oxygen vacancies. In the present study, we have observed a reproducible reduction in leakage current density by about an order of magnitude for the
Ag(2)–BST-based MIM capacitor (B106 A/cm2) at a high electric field of 200 kV/cm when measured under either biases in comparison with undoped BST (B105A/cm2). The dominant mechanism in Ag(1)–BST and Ag(2)–BST films might be steps (2) and (3) as indicated before. According to expression (3), it can be concluded that electrons are consumed during the oxy-genation of the Ag(1 and 2)–BST films, which makes the film relatively p-type39compared with pure BST up to an optimum Ag content. The lowest leakage current density at 200 kV/cm is obtained for an Ag(2)–BST film, where a number of mecha-nisms may be working in concert so that an effective reduction in leakage current is obtained.
The observed leakage current for a BST-based capacitor comprises the contributions from the electronic conduction and polarization loss.6As observed and discussed in our
previ-ous communication on dielectric characteristics of Ag(1)–BST and Ag(2)–BST,38 there is a significant reduction in dielectric constant of Ag(2)–BST as compared with undoped BST, with the lowest dielectric loss tangent in the entire range of electric field measured. The observed reduction in loss factor has been ascribed to the dense microstructure, enhanced polarization pin-ning, and a series configuration of low-dielectric interface layers or metal–ceramic configuration of Ag and BST.38From this re-sult, we conclude that the polarization loss is minimum for an Ag(2)–BST film. Moreover, as explained with the help of Eqs. (2) and (3), filling of oxygen vacancies may be predominant up to a certain amount of Ag. Upon increasing the Ag content to 3 mol%, the leakage current density increases rapidly beyond 100 kV/cm because of an increase in the concentration of elec-trons that results in increased electronic conductivity. The grad-ual increase in the electronic conduction with an increase in Ag is obvious as the current density curve rises with a higher slope at the beginning of the measurement at a low electric field as obtained in the case of Ag(4)–BST. Therefore, we conclude that a low leakage current is maintained only up to a certain amount of Ag in BST for which a series metal–ceramic configuration exists and oxygenation is dominant. In our samples, this series configuration is disrupted as the Ag content is increased to 3 mol% and beyond.
IV. Conclusions
In summary, an optimum amount of Ag (1 and 2 mol%) en-hances the binding energy of surface oxygen and reduces the leakage current density of chemical solution-deposited Ag–BST nanocrystalline thin-film capacitors through a combination of a microstructure that maintains a series metal–ceramic
configura-–300 –200 –100 0 100 200 300 102 101 100 10–1 10–2 10–3 10–4 10–5 10–6 10–7 10–8
Leakage Current Density (A/cm
2) Electric Field (kV/cm) BST Ag(1)-BST Ag(2)-BST Ag(3)-BST Ag(4)-BST
Fig. 4. Leakage current density characteristics of
metal–insulator–met-al (MIM) capacitors of configuration, Pt/Ag(0y4)–BST/Pt under for-ward and reverse biases.
tion for which the polarization loss is less and the filling of ox-ygen vacancy is dominant. In addition, there is also a slight en-hancement of binding energy of Ag up to 4 mol% because of its possible chemical shift or dissolution along the surface layers of BST, surface relaxation effects, and Fermi-level shifts of Ag in the Ag–BST films. The leakage current increases and becomes higher than pure BST for Ag content greater than or equal to 3 mol% because of increased electronic conduction.
References
1
M. Nayak, S. Ezhilvalavan, and T. Y. Tseng, ‘‘High-Permittivity (Ba,Sr)TiO3
Thin Films;’’ pp. 99–167 in Handbook of Thin Film Materials, Vol. 3, Edited by H. S. Nalwa. Academic Press, New York, 2002.
2
Y. Gao, C. L. Perkins, S. He, P. Alluri, T. Tran, S. Thevuthasan, and M. A. Henderson, ‘‘Mechanistic Study of Metal-Organic Chemical Vapor Deposition of (Ba,Sr)TiO3Thin Films,’’ J. Appl. Phys., 87 [10] 7430–7 (2000).
3
S. Y. Lee and T. Y. Tseng, ‘‘Electrical and Dielectric Behavior of MgO Doped Ba0.7Sr0.3TiO3Films on Al2O3Substrate,’’ Appl. Phys. Lett., 80 [10] 1797–9 (2002).
4
K. P. Jayadevan and T. Y. Tseng, ‘‘Composite and Multilayer Ferroelectric Thin Films: Processing, Properties and Applications,’’ J. Mater. Sci.: Mater. Elec-tron., 13 [8] 439–59 (2002).
5
C. Y. Liu, H. T. Lue, and T. Y. Tseng, ‘‘Effects of Nitridation of Silicon and Repeated Spike Heating on the Electrical Properties of SrTiO3Gate Dielectrics,’’
Appl. Phys. Lett., 81 [23] 4416–8 (2002).
6
M. S. Tsai, S. C. Sun, and T. Y. Tseng, ‘‘Effect of Oxygen to Argon Ratio on Properties of (Ba,Sr)TiO3Thin Films Prepared by Radio-Frequency Magnetron
Sputtering,’’ J. Appl. Phys., 82 [7] 3482–7 (1997).
7
S. Safar, R. E. Jones, B. Jiang, B. White, P. Chu, D. Taylor, and S. Gillespie, ‘‘Oxygen Vacancy Mobility Determined from Current Measurements in Thin Ba0.5Sr0.5TiO3Films,’’ Appl. Phys. Lett., 73 [2] 173–7 (1998).
8
A. Srivastava, D. Kumar, R. K. Singh, H. Venkataraman, and W. R. Eisestadt, ‘‘Improvement in Electrical and Dielectric Behavior of (Ba, Sr)TiO3Thin Films by
Ag Doping,’’ Phys. Rev. B, 61 [11] 7305–7 (2000).
9
J. Assal, B. Hallstedt, and L. J. Gauckler, ‘‘Thermodynamic Assessment of the Silver–Oxygen System,’’ J. Am. Ceram. Soc., 80 [12] 3054–6 (1997).
10
Y. C. Her, Y. C. Lan, W. C. Tsu, and S. Y. Tsai, ‘‘Effect of Constituent Phases of Reactively Sputtered AgOxFilm on Recording and Readout Mechanisms of
Super-Resolution Near-Field Structure Disk,’’ J. Appl. Phys., 96 [3] 1283–8 (2004).
11
D. M. Tahan, A. Safari, and L. C. Klein, ‘‘Preparation and Characterization of BaxSr1xTiO3Thin Films by a Sol–Gel Technique,’’ J. Am. Ceram. Soc., 79 [6]
1593–8 (1996).
12
P. M. Ajayan and L. D. Marks, ‘‘Experimental Evidence of Quasi-Melting in Small Particles,’’ Phys. Rev. Lett., 63 [3] 279–82 (1989).
13
C. Miot, E. Husson, C. Proust, R. Erre, and J. P. Coutures, ‘‘X-Ray Photo-electron Spectroscopy Characterization of Barium Titanate Ceramics Prepared by Citric Route. Residual Carbon Study,’’ J. Mater. Res., 12 [9] 2388–92 (1997).
14
Y. Fujisaki, Y. Shimamoto, and Y. Matsui, ‘‘Analysis of Decomposed Layer Appearing on the Surface of Barium Strontium Titanate,’’ Jpn. J. Appl. Phys. Part 2, 38 [1A/B] L52–4 (1999).
15
V. Craciun and R. K. Singh, ‘‘Characteristics of the Surface Layer of Barium Strontium Titanate Thin Films Deposited by Laser Ablation,’’ Appl. Phys. Lett., 76 [14] 1932–4 (2000).
16
F. Amy, A. Wan, A. Kahn, F. J. Walker, and R. A. McKee, ‘‘Surface and Interface Chemical Composition of Thin Epitaxial SrTiO3and BaTiO3Films:
Photoemission Investigation,’’ J. Appl. Phys., 96 [3] 1601–6 (2004).
17
S. J. Wang and C. K. Ong, ‘‘Epitaxial Y-Stabilized ZrO2Films on Silicon:
Dynamic Growth Process and Interface Structure,’’ Appl. Phys. Lett., 80 [14] 2541–3 (2002).
18
A. I. Boronin, S. V. Koscheev, K. T. Murzakhmetov, V. I. Avdeev, and G. M. Zhidomirov, ‘‘Associative Oxygen Species on the Oxidized Silver Surface Formed Under O2Microwave Excitation,’’ Appl. Surf. Sci., 165 [1] 9–14 (1999).
19
X. Bao, U. Wild, M. Muhler, B. Pettinger, R. Schlo¨gl, and G. Ertl, ‘‘Coad-sorption of Nitric Oxide and Oxygen on the Ag(110) Surface,’’ Surf. Sci., 425 [2–3] 224–32 (1999).
20
V. I. Avdeev and G. M. Zhidomirov, ‘‘Ethylene and Oxygen Species Adsorbed on a Defect Oxidized Surface Ag(111): Theoretical Analysis by DFT Method,’’ Surf. Sci., 492 [1–2] 137–51 (2001).
21
D. Y. Zemlyanov, E. Savinova, A. Scheybal, K. Doblhofer, and R. Schlo¨gl, ‘‘XPS Observation of OH Groups Incorporated in an Ag(111) Electrode,’’ Surf. Sci., 418 [2] 441–56 (1998).
22
K. S. Song, S. K. Kang, and S. D. Kim, ‘‘Preparation and Characterization of Ag/MnOxPerovskite Catalysts for CO Oxidation,’’ Catal. Lett., 49 [1–2] 65–8
(1997).
23
V. R. Choudhary, B. S. Uphade, and S. G. Pataskar, ‘‘Low Temperature Complete Combustion of Methane Over Ag-Doped LaFeO3and LaFe0.5Co0.5O3
Perovskite Oxide Catalysts,’’ Fuel, 78 [8] 919–21 (1999).
24
W. X. Xu, S. Zhu, J. Zhou, X. C. Fu, and X. N. Zhao, ‘‘Atomic-Dimensional Film of TiO2xand its Quantum Size Effect: UV–VIS, XPS, SEM and TEM
study,’’ J. Chem. Soc., Faraday Trans., 93 [23] 4187–95 (1997).
25
M. Bielmann, P. Schwaller, P. Ruffieux, O. Gro¨ning, L. Schlapbach, and P. Gro¨ning, ‘‘AgO Investigated by Photoelectron Spectroscopy: Evidence for Mixed Valence,’’ Phys. Rev. B, 65 [23] 5431–5 (2002).
26
D. Peterka, C. Tegenkamp, K. M. Schro¨der, W. Ernst, and H. Pfnu¨r, ‘‘Oxygen Surplus and Oxygen Vacancies on the Surface of Epitaxial MgO Layers Grown on Ag(100),’’ Surf. Sci., 431 [1–3] 146–55 (1999).
27
G. I. N. Waterhouse, G. A. Bowmaker, and J. B. Metson, ‘‘Interaction of a Polycrystalline Silver Powder with Ozone,’’ Surf. Inerface Anal., 33 [5] 401–9 (2002).
28
K. P. Jayadevan, N. V. Kumar, R. M. Mallya, and K. T. Jacob, ‘‘Use of Metastable, Dissociated and Charged Gas Species in Synthesis: A Low Pressure Analogue of the High Pressure Technique,’’ J. Mater. Sci., 35 [10] 2429–34 (2000).
29
L. H. Tjeng, M. B. Meinders, J. van Elp, J. Ghijsen, G. A. Sawatzky, and R. L. Johnson, ‘‘Electronic Structure of Ag2O,’’ Phys. Rev. B, 41 [5] 3190–9 (1990).
30
J. H. Pei, C. G. Olson, and R. Manzke, ‘‘Surface Core-Level Binding Energy Shifts of InAs,’’ Chinese J. Phys., 26 [2] 90–5 (1988).
31
S. J. Shih and W. H. Tuan, ‘‘Solubility of Silver and Palladium in BaTiO3,’’ J.
Am. Ceram. Soc., 87 [3] 401–7 (2004).
32
T. A. Byrne and D. P. Cann, ‘‘Effect of Silver on Barium Titanate as a Func-tion of Stoichiometry,’’ J. Am. Ceram. Soc., 87 [5] 875–80 (2004).
33
R. D. Shannon, ‘‘Revised Effective Ionic Radii and Systematic studies of Interatomic Distances in Halides and Chalcogenides,’’ Acta Crystallogr., 32A, 751–67 (1972).
34
B. Kucharczyk and W. Tylus, ‘‘Effect of Pd or Ag Additive on the Activity and Stability of Monolithic LaCoO3Perovskites for Catalytic Combustion of
Methane,’’ Catal. Today, 90 [1–2] 121–6 (2004).
35
C. Y. Chen and W. H. Tuan, ‘‘Effect of Silver on the Sintering and Grain-Growth Behavior of Barium Titanate,’’ J. Am. Ceram. Soc., 83 [12] 2988–92 (2000).
36
A. R. Williams and N. D. Lang, ‘‘Core-Level Binding Energy Shifts in Met-als,’’ Phys. Rev. Lett., 40 [14] 954–7 (1978).
37
N. Halder, A. D. Sharma, S. K. Khan, A. Sen, and H. S. Maiti, ‘‘Effect of Silver Addition on the Dielectric Properties of Barium Titanate Based Low Tem-perature Processed Capacitors,’’ Mater. Res. Bull., 34 [4] 544–50 (1999).
38
K. P. Jayadevan, C. Y. Liu, and T. Y. Tseng, ‘‘Dielectric Characteristics of Nanocrystalline Ag–Ba0.5Sr0.5TiO3Composite Thin Films,’’ Appl. Phys. Lett., 87
[2] 1211–3 (2004).
39
R. Waser and D. M. Smyth, ‘‘Defect Chemistry, Conduction, and Breakdown Mechanism of Perovskite-Structure Titanates;’’ pp. 47–92 in Ferroelectric Thin Films: Synthesis and Properties, Edited by C. P. de Araujo, J. F. Scott, and G. W. Taylor. Gordon and Breach Publishers, UK, 1996. &
