Sol–gel preparation and characterisation of mixed metal tin oxide thin films

Sol –gel preparation and characterisation of mixed metal tin oxide thin filmsChemistry Department, University of Stellenbosch, Matielanad X1, Stellenbosch 7800, South Africa Received 28

Sol –gel preparation and characterisation of mixed metal tin oxide thin films

Chemistry Department, University of Stellenbosch, Matielanad X1, Stellenbosch 7800, South Africa Received 28 June 2005; received in revised form 13 March 2006; accepted 29 January 2007

Available online 3 February 2007

Abstract

Mixed metal oxide stannates were prepared by sol–gel methods and coated onto solid titanium substrates as thin films using spin and dip coating methods Metal oxides such as Sb2O5, ZrO2, CuO, MnOxand PdO were introduced into a SnO2host matrix using sol–gel technology The mixed metal tin oxide materials prepared via the sol–gel route were extensively characterised in terms of surface characterisation and chemical composition Thermogravimetric analysis was performed to confirm that at 600 °C (the calcination temperature) no further structural changes due

to mass loss occur UV spectroscopy of the liquid gels allowed the determination of the band gap energy The surface morphology of the thin film electrodes were characterised by atomic force microscopy and scanning electron microscopy and the effect of the coating method employed i.e spin or dip coating could be clearly seen in the estimated values of surface roughness These techniques were also able to confirm the thickness of the films in the nano range Combined nuclear beam techniques such as Rutherford backscattering spectroscopy and particle induced X-ray emission provided some insight into the chemical composition of the mixed metal tin oxides and confirmed the presence of the dopant element in the SnO2host material

© 2007 Elsevier B.V All rights reserved

Keywords: Sol –gel; Surface characterisation; Metal oxides; thin films; spin and dip coating

1 Introduction

One of the most characteristic chemical properties of

transition metals, is the occurrence of multiple oxidation states

Many transition metal compounds can be used as homogeneous

catalysts, since they readily inter-convert between oxidation

states In this way they can act as intermediates for the exchange

of electrons between the reactants Enzymes act as

homoge-neous catalysts in living systems and their active sites are often

transition metals[1]

Electrocatalysis requires the incorporation of the transition

metal catalyst into a suitable host material, for production of

thin film electrodes The resultant thin film should be

homogeneous in its chemical composition and stable under

anodic conditions, with respect to chemical and mechanical

weathering Therefore, in the production of thin film electrodes

the chemical and electrical nature of the host material is as

important as the catalytic ability of the electrode, induced by the dopant element SnO2films show very little absorbance in the visible IR-range, are very resistant against chemical attack, have high mechanical strength, good electrical conductivity when doped, relatively high hydrogen and oxygen over voltages and can easily be prepared as thin film electrodes Therefore, SnO2

films meet the general requirements for dimensionally stable electrode host material[2]

These properties make it an attractive choice for electrode material in spectroscopic studies SnO2has a band gap of∼ 3.7 eV and when used as a thin film electrode material usually contains oxygen vacancies in the lattice Additionally it can also be prepared to contain donor atoms such as Cl or Sb, which raises its n-type carrier capacity to values in excess of 1020cm− 3 and consequently the potential conductivity to well above

10 ohm− 1 cm− 1 The electrodes retain their semiconducting properties and the exchange currents are low [3] The micro-structure of tin oxide, which can be controlled by deposition parameters, also greatly affects its electrical properties Factors such as film thickness and deposition temperature impact on electrical conductivity [4] Perfect control over deposition parameters in the development of industrial gas sensors, have

www.elsevier.com/locate/tsf

⁎ Corresponding author Tel.: +27 21 808 3535; fax: +27 21 808 3849.

E-mail address: amc@sun.ac.za (A.M Crouch).

1 Present address: Chemistry Department, University of the Western Cape,

Private Bag X17, Bellville 7535, South Africa.

0040-6090/$ - see front matter © 2007 Elsevier B.V All rights reserved.

doi: 10.1016/j.tsf.2007.01.042

many advantages i.e miniaturisation which implies power

reduction, microelectronic compatibility for use in multi sensors,

mass production of sensors and good reproducibility

We have seen from literature that there are many examples of

mixed metal oxide electrode materials for various applications

Good candidates as starting materials for alloys seem to be early

to late transition metals alloys, since alloying of such metals can

generate electronic structures with high catalytic activity

Amorphous alloys (of which the composition can be gradually

changed) tend to be very homogeneous and have co-ordination

sites, which are unsaturated in many cases and can generate

strong catalytic activity An iron modified cobalt oxide anode

system and a nickel oxide anode have both demonstrated the

successful degradation of phenol to carboxylic acids and CO2

[5,6] A mixed metal oxide system prepared from SnO2as host

material, was also successful in the complete degradation of

phenol SnO2 doped with Sb2O5 showed improved catalytic

ability for phenol oxidation, compared to Ebonex/PbO2

electrodes[7] We would like to investigate the possibility of

using SnO2as host material and doping it with selected transition

metal oxides (PdO, MnOx, CuO, and ZrO2), thereby producing a

new class of electrodes, with improved catalytic activity

2 Experimental details

2.1 Sol–gel preparation

Sol–gel involves solvolysis of organic or inorganic precursors,

followed by condensation and ageing A schematic representation

of the sol–gel process is given inFig 1 The cationic precursors in

sol–gel preparation were either chloride salts (Sb, Pd) or organic

derivatives (Cu, Zr) In the case of manganese doping, citrate

derivatives of manganese and tin were preferred The manganese

citrate derivative was freshly prepared before sol–gel synthesis,

from MnCO3by precipitation and filtration The organic solvent

in all cases was absolute ethanol, except for zirconium doping,

where the solvent was methanol All acids used in the preparation

were Merck Suprapur grade and synthesis reagents were

analytical grade and were used as purchased from the various

suppliers, without any further purification

The doping ratio (Table 1) was determined according to the following formula:

Doping ratio¼ ni=Rn Where niis the number of moles of dopant atom and n is the total number of moles (moles of SnO2+ moles of dopant) The sol–gel synthesis for the Sb, Cu, Mn, Pd and Zr-doped tin oxide materials, was based on synthesis procedures for these metal oxides, derived from literature [7–11] In general the starting materials were placed in a round bottomed flask fitted with a thermometer, and the solutions were well stirred and refluxed for 1 h, before being left to age at room temperature for

at least 24 h to ensure ageing of the gel The gels were found to

be stable at room temperature for up to 6 months when stored in the absence of direct light

2.2 Preparation of thin films For the purpose of producing the thin films, the gels were retained in the liquid form The titanium substrates (Sigma Aldrich, titanium foil, 99.99%; 34,880–5) were cut to size (50 mm × 50 mm) and etched in concentrated HCl (1 h), rinsed with copious amounts of UHQ water and finally rinsed with absolute ethanol The dip coated thin films were prepared using the method of Grimm[7] The substrates were dipped into the gel solution and then slowly withdrawn at a rate of 80 mm min− 1 The spin coated samples were prepared using the rotating disk of a commercial angle grinder as controlled spinning rate surface The titanium substrate onto which thin films were coated, was affixed to the rotating disk surface Two

to three drops of the respective gel material was placed on the substrate and rotation of the sample at full speed (± 1000 rpm) for 10 s gave reproducible thin films The spin coating procedure was repeated 2–3 times to ensure good coverage of the substrate, with drying at 100 °C in air in between each coating application, for about 20 min After the final coating was applied the thin films were annealed in air in a quartz tube furnace at a slow heating rate (1 °C/min) up to 600 °C and then allowed to cool to room temperature, under ambient conditions 2.3 Thermogravimetric analysis

Small sample masses were used in Thermogravimetric analysis (TGA) to ensure uniform heating of sample and to reduce exchange of gas with the surrounding atmosphere A concentrated gel material was prepared for TGA by evaporating

Fig 1 Schematic representation of sol –gel process (M represents the dopant

metal as described in the text).

Table 1 Doping ratios for mixed metal tin oxides

SnCl 2 ·2H 2 O, SbCl 3 [7] Sb 2 O 5 10 mol% SnCl 4 ·5H 2 O; Cu(CH 3 COO) 2 ·H 2 O [8] CuO 2.5 wt% SnCl 2 ·2H 2 O; Zr(OC 3 H 7 ) 4 [9] ZrO 2 1 mol% Citric acid, tin citrate, ethylene glycol,

nitric acid, manganese citrate [10]

the solvent of a small volume (a few mL) of the original gel The

sample was analysed on a Perkin Elmer TGA (7 series) The

heating rate was 1 °C/min up to a maximum of 600 °C and

nitrogen was used as purging gas

2.4 UV/Vis spectroscopy

Absorption measurements of the gel material in the liquid

phase was done in order to observe the effect of the inclusion of

the dopant metal oxide on the band gap of the host material,

before annealing the thin films A Unicam, Helios Gamma, UV

spectrophotomer and Aurora 1.1 scanning software version was

used for the collection of the data The instrument was used in

single beam mode Absolute ethanol (solvent used in

prepara-tion of mixed metal oxide gels) was used as a reference The gel

samples were diluted 1:1 with absolute ethanol to obtain

absorbance readings on an appropriate scale

2.5 X-ray diffraction

The liquid gels obtained by sol–gel technology, was

evaporated to dryness to produce a powder The gel solution

was subjected to the same annealing cycle (maximum

temperature of 600 °C, ramp rate of 1 °C/min) as that used

for the preparation of the thin films The resultant powder

samples were submitted for X-ray diffraction (XRD) analysis

Measurements were made on a Siemens advanced D8

diffractometer (Bruker AXS) with primary (Cu tube, variable

divergent slit) and secondary (NaI(Th)) scintillation, variable

anti-scatter slit) detectors The incident angle was set as 2Θ for

all measurements

2.6 Scanning electron microscopy

Scanning electron microscopy (SEM) pictures of the thin

film surfaces were obtained at 500× and 5000× magnification

using a Topcon scanning electron microscope, model ABT60

The microscope was operated in secondary electron detection

(SED) mode at an accelerating voltage of 20 kV The working

distance was 10 mm, the samples were tilted 20° and they were

not coated for viewing

2.7 Atomic force microscopy

Atomic force microscopy (AFM) micrographs were taken

at two different positions on each thin film (solid state) that

was prepared using a Topometrix Explorer TMX 2000 The

samples were scanned in non-contact mode (i.e the probe is

oscillated in the attractive regime) A low resonance

frequency cantilever was used and the force spring constant

was 35–65 N/m

2.8 Particle induced X-ray emission/ Rutherford backscattering

Compositional analyses of the mixed metal oxides cast as

thin films (solid state) were obtained by the combination

Particle induced X-ray emission/ Rutherford Backscattering

(PIXE/RBS) technique The sample was bombarded with protons and the energy of the beam was kept at 3 eV A selected section of the surface (5 μm×5 μm) was bombarded with protons and taken to be representative of the total surface composition (5 mm × 5 mm) The backscattered energy was recorded using an X-ray detector Semi-quantitative on-line imaging was made possible by the use of the GEO-PIXE suite

of programmes Compositional analysis (ppm of element) is based on matrix determination and depth profiling (film thickness) provided by RBS PIXE/RBS and XRD analyses were done at the nuclear microprobe facility at Ithemba Labs, Faure, South Africa

3 Characterisation results and discussion 3.1 Thermogravimetric analysis

The TGA profiles obtained for the synthesized gels are in good agreement with published TGA profiles for Zr, Pd and CuO doped SnO2 [12,13] In all cases, an initial 10 to 20% weight loss was assigned to the evaporation of the organic solvent Subsequent major weight loss between 100 to 200 °C is ascribed to evaporation of water, salts and organic compounds The only possibilities for further weight loss after the oxides are formed, is loss of oxygen ZrO2maintains its 4+ oxidation state throughout the heating cycle ZrO2 only decomposes above

2677 °C[14] Sb2O5however decomposes by losing oxygen at

380 °C to form Sb2O3, which is stable up to 655 °C[15] The similarities in the decomposition profiles of antimony doped and zirconium doped tin oxide can be related to the high oxidation state of the final oxide product during the heating cycle CuO and MnO2 doped tin oxide follow a similar decomposition pattern and the similarities can be attributed to the fact that in both of these mixed oxides the dopant metal center has an oxidation number of II CuO is stable up to

1336 °C but MnO2decomposes above 535 °C to form MnO

[16,17] It is possible that the decomposition temperature of MnO2 could be lowered when incorporated in to the SnO2

matrix PdO2decomposes to form PdO at comparatively low temperature i.e 200 °C PdO is then stable up to 750 °C[18] Any further loss of oxygen by this oxide will give rise to elemental Pd, therefore the most stable phase possible for palladium oxides, is reached

The decomposition profiles obtained by TGA are not influ-enced by the decomposition of the host material, which is stable

as SnO2up to temperatures of 1630 °C[19] TGA has confirmed that all mixed metal tin oxides form stable phases before the maximum temperature used in the annealing of prepared mixed metal oxide thin films, i.e 600 °C (Fig 2) The effect of temperature cycle and heating atmosphere on doping content was not evaluated by any other thermal treatment programmes 3.2 UV/Vis spectroscopy

The band gap energies for all mixed metal tin oxide gel materials (liquid form) were determined to be between 2.37 to 3.37 eV (Table 2)

The band gap energy (eV) for each mixed metal oxide was

determined by extrapolating the wavelength of the onset of

absorption in the UV region, as illustrated (Fig 3) and solving

the equation E = hc /λ[16,17]

PdO–SnO2was the only mixed metal tin oxide which gave a

band gap energy lower than that of undoped tin oxide Such a

lowering of the band gap energy means that the conductivity of

the doped material is improved as a result of the addition of the

dopant The MnO2–SnO2sample gel was obtained in the form

of a dispersion rather than a clear solution and this complicated

the absorbance measurements, even after dilution It was

therefore not possible to determine the band gap energy for

this particular mixed metal oxide composition

3.3 X-ray methods

XRD was able to confirm the characteristic peaks for SnO2

when compared to the Joint Committee on Powder Diffraction

Standards (JCPDS) file for cassiterite syn SnO2for the mixed

metal oxide powders (Fig 4)[18] Peaks for the identification of

the dopant atom were however not always observed since the

very low dopant ratio of 1 mol% is beyond the capability of the

instrument used In a similar study of SnO2/ZrO2prepared as

95:5 wt % composition, XRD could only confirm the crystalline

cassiterite structure Thus at 5 wt % ZrO2doping the refinement

of the lattice parameters were in good agreement with that of bulk SnO2 [19] Energy dispersive X-ray fluorescence (EDXRF) was able to positively identify the copper in CuO/ SnO2 XRD could not confirm the presence of the dopant atoms 3.4 Scanning electron microscopy

Scanning electron micrographs reflected well coated sur-faces for all mixed metal tin oxides when prepared by spin coating as well as dip coating The thin films were homogeneous in composition and morphology after being annealed at 600 °C[20]

The Zr-doped spin coated surface showed dramatic frag-mentation at 500× magnification Increase in the doping level of

Zr in the tin oxide host material gave rise to clusters or island formation as a result of exclusion Pd-doped spin coated surface produced a cauliflower like surface as a result of agglomeration

of metal oxide particles (Fig 5) Noble metals like palladium, readily oxidize in air forming a layer of touching oxide particles These oxide particles can be reduced back to metallic palladium

by further heating (up to 500 °C) in a reducing atmosphere, giving rise to the agglomeration of metal oxide and metallic particles observed[21,22]

Fig 2 TGA of metal oxide doped tin oxide gels (a) ZrO 2 –SnO 2 (b) Sb 2 O 5 –

SnO 2 (c) CuO –SnO 2 (d) MnO x –SnO 2 (e) PdO –SnO 2

Table 2

Band gap energy values of the mixed metal tin oxide gels prepared

Fig 3 An example of how band gap energy was measured for ZrO 2 –SnO 2

Fig 4 XRD pattern for Zr –SnO 2 powder showing only the characteristic peaks for SnO

The agglomeration observed for zirconium and palladium

doped tin oxide, were confirmed by repeat coatings The spin

coated thin films showed greatly enhanced surface roughness

compared to dip coated films of the same composition

3.5 Atomic force microscopy

Atomic force micrographs of all mixed metal tin oxide thin

films confirmed the increased surface roughness of dip coated

thin films as opposed to spin coated surfaces Dip coated

samples were observed to have greater surface roughness as

compared to the samples prepared by spin coating, for example

the maximum peak-to-valley value of measured heights for

Mn-doped tin oxide films, ∼ 247 nm for dip coated films and

∼40 nm for spin coated films (Fig 6) The surface roughness

values are based on average depth profiles of the pores for

which measurements were made from the low point (dark spots)

to the high point (light spots) on the film The regularly

repeating triangular surface morphology observed for ZrO2

doped tin oxide thin films are indicative of the stable

m-tetragonal phases that are formed in the temperature range 400–

650 °C (Fig 7) Sb2O5, CuO, PdO and MnO2doped tin oxide

thin films also exhibited this typical triangular morphology

[23,24]

3.6 Rutherford backscattering

The RBS data was used to confirm that the substrate that was

used (solid titanium) and the coated metal oxide thin films gave

well separated peaks for simulation and fitting purposes

Therefore RBS could be employed as a suitable analytical tool for compositional analysis Simulation was done using the Rutherford Universal Manipulation Programme (RUMP)

In most cases the elements were found to be more abundant

in the spin coated (larger peak area) thin films than in the dip coated thin films

Compositional analysis of the thin films as presented in the data is biased by the composition of the specific surface area that was bombarded with α-particles (RBS) and protons (PIXE) Usually an area 5 μm×5 μm near the center of the samples was selected for characterization and taken as representative of the entire sample However, it is reasonable

to expect the concentration of dopant element to be higher in spin coated samples as compared to dip coated samples, since the spread of the gel material under centrifugal force (spin coating) is more homogeneous and thinner as opposed to the spread or drag created by the gravitational force in the dip coating method

An estimate of the film thickness was obtained from RBS data and was found to be in the order of 20 to 40 nm for spin coated films and less for dip coated films The only exception was the Zr spin coated film, which had a film thickness of

140 nm This is in agreement with AFM estimates of film thickness, in the nanometer range

3.7 Particle induced X-ray emission Three sets of prepared thin films were analysed In all cases the dopant element was positively identified The spin coated samples generally gave thicker films than the dip coated samples did The detection of the dopant element was consequently also more successful for spin coated samples than for dip coated samples For the third set of samples, detection of the dopant element in the dip coated samples was generally so low that no quantitative evaluation of the dopant element was performed

Tungsten (W) was identified as a contaminant in all samples,

as determined by performing a blank run on an uncoated titanium sample Further tests with blank samples and standards eliminated the detection chamber as the source of contamination

Fig 5 SEM of spin coated metal Ti/SnO 2 –PdO thin film.

Fig 6 Comparison of surface roughness for dip and spin coated thin films.

Fig 7 AFM topographic view of Ti/SnO 2 –ZrO 2

The W is thought to be present as a contaminant in the titanium.

Fe and Ag contaminants were introduced as part of the

composition of the starting materials and were therefore

unavoidable The zinc contamination of the palladium dip

coated sample was included in the table since it emerged as a

possibility during the characterisation of the major lines, in the

PIXE spectrum The zinc line was only observed once in the

characterisation of the three sets of palladium samples and could

be due to contamination of the sample or sample chamber

However, none of the contaminants showed any interference

with the catalytic behaviour of the thin film electrodes

Evaluating the electrochemical response of a one electron

shuttle such as K3Fe(CN)6at these electrodes showed only the

oxidation and reduction peaks associated with the Fe3+/Fe2+

couple and no other interfering peaks (Fig 8)

By using intrinsic doping during sol–gel synthesis, a fusion of

the dopant atom with the SnO2host material is obtained after

annealing the thin films The surface composition of mixed metal

oxide thin films (expressed as a percentage) based on ppm

quantification, will therefore not necessarily reflect the calculated

composition (Table 3) Consistency in the estimations for Pd, Zr,

Sb and Cu is observed, at their respective levels of doping The

dopant ratios for the different mixed metal tin oxides vary in

accordance with the prescribed dopant ratio in the reference from

which the synthesis conditions were taken (cf.Table 1)

For detailed structural analysis, tools such as X-ray

crystallography could assist in identifying the exact location

of the dopant molecules within the SnO2host structure This would involve the growth of good quality crystals This is a time consuming process and a specialized skill in its own right Crystallographic results were not pursued as part of this work

4 Conclusion From TGA analysis it emerges that PdO–SnO2forms the most stable phase (± 90% weight loss and 180 °C) whilst all the other mixed metal oxides stabilize at much higher temperatures (around 400 °C) ZrO2–SnO2and Sb2O5–SnO2follow a similar temperature degradation profile with a maximum % weight loss

of 60% CuO–SnO2and MnOx–SnO2show 80% weight loss The mixed metal tin oxides form stable phases in the order PdONCuONSb2O5NZrO2NMnOx

From UV/Vis measurements it was possible to determine the band gap energies for all mixed metal oxide liquid gels to be in the range 2.37 to 3.37 eV The addition of these metal oxide dopants do not decrease the band gap energies of undoped SnO2 ZrO2–SnO2 gels show a lower band gap energy i.e improved electron conduction, with prolonged ageing since the sample becomes more crystalline PdO–SnO2was the only mixed metal oxide that had a lower band gap than undoped SnO2 XRD analysis of the transition metal doped samples could only confirm the cassiterite structure of SnO2; it was not useful

in identifying any of the dopant atoms PIXE/RBS combination was used for determination of chemical composition An accurate assessment of whether the calculated doping ratio was obtained, needs to be based on raster or grid-like compositional analysis of the entire surface and taking into account the effect

of depth profiling This is possible with the PIXE/RBS combination technique but was not done due to limitations in availability of the nuclear probe facility

SEM and AFM confirmed the formation of good quality thin films with very few cracks or pin holes The spin coated films were smoother than dip coated films with surface roughness estimates usually one order of magnitude lower than for dip coated thin films

Consistency in the compositional analysis of Pd, Zr, Sb and

Cu thin films (% composition based on ppm concentrations) was observed with RBS/PIXE combination, at their respective level of doping The inclusion of the transition metal oxide active centres in the host material was confirmed

Fig 8 Cyclic voltammetry of the Ti/SnO 2 –ZrO 2 spin coated thin film showing

the oxidation of K 3 Fe(CN) 6 in acidic medium at 50 mV/s.

Table 3

Typical compositional analysis of mixed metal tin oxide thin films (% composition based on ppm analysis) as determined by PIXE/RBS

The authors wish to acknowledge the financial support of the

National Research Foundation (NRF) South Africa and Ithemba

Labs (Faure, South Africa) for assistance with PIXE/RBS

analysis of the thin film samples

References

[1] M.S Silberberg, Chemistry: the Molecular Nature of Matter and Change,

2nd edition McGraw-Hill Higher Education, 2000.

[2] C Iwakura, M Inai, T Uemura, H Tamura, Electrochim Acta 26/4 (1981)

579.

[3] W Badawy, K Doblhofer, I Eiselt, H Gerischer, S Krause, J.

Melsheimer, Electrochim Acta 29/12 (1984) 1617.

[4] L Bruno, R Lalauze, C Pijolat, Adv Inorg Films Coatings 5 (1995) 497.

[5] St.G Christoskova, M Stoyanova, Water Res 35/8 (2001) 2073.

[6] St.G Christoskova, M Stoyanova, M Georgieva, Appl Catal., A Gen.

208 (2001) 243.

[7] J Grimm, D Bessarabov, W Maier, S Storck, R Sanderson, Desalination

115 (1998) 295.

[8] G Fang, Z Liu, Z Zhang, Phys Status Solidi A 156 (1996) 15.

[9] G Fang, Z Liu, Z Zhang, Phys Status Solidi A 156 (1996) 81.

[10] O Gouvea, J Varela, E Longo, A Smith, J Bonnet, Eur J Solid State Inorg Chem 30 (1993) 915.

[11] X Cao, L Cao, W Yao, X Ye, Surf Interface Anal 24 (1996) 662 [12] S Tiwari, S Nema, Y Agarwal, Thermochim Acta 317 (1998) 175 [13] S Qadri, H Kim, H Khan, E Skelton, Thin Solid Films 377378 (2000) 750.

[14] K MacKay, R MacKay, W Henderson, Introduction to Modern Inorganic Chemistry, 6th edition CRC Press, 2002.

[15] V Henrich, P Cox, The Surface Science of Metal Oxides, Cambridge University Press, 1996.

[16] G Martínez-Castañón, M Sánchez-Loredo, J Martínez-Mendoza, F Ruiz, Adv Technol Mater Mater Process 7/2 (2005) 171.

[17] F Zhou, K Kang, T Maxisch, G Ceder, D Morgan, Solid State Commun.

132 (2004) 181.

[18] A Garje, R Aiyer, Int J Appl Ceram Technol 3/6 (2006) 477 [19] S Qadri, H Kim, H Khan, E Skelton, Thin Solid Films 377 –378 (2000) 50.

[20] D Gang, D Zhang, L Zhao, J Xu, D Zhou, Key Eng Mater (2005) 280 [21] J Mizsei, L Pirttiaho, M Karppinen, V Lanto, Sens Actuators, B, Chem.

65 (2000) 195.

[22] H.M Villullas, F Mattos-Costa, P Nascente, L.O.S Bulhoes, Electrochim Acta 49/22 –23 (2005) 3909.

[23] S Chang, R Doong, Chem Mater 17 (2005) 4837.

[24] M Makgae, C Theron, W Przybylowicz, A Crouch, Mater Chem Phys.

92 (2005) 559.