Electrochemical Fabrication and Physicochemical Characterization of

Article pubs.acs.org/JPCC

Electrochemical Fabrication and Physicochemical Characterization of Metal/High‑k Insulating Oxide/Polymer/Electrolyte Junctions F. Di Franco, M. Santamaria,* and F. Di Quarto Electrochemical Materials Science Laboratory, Dipartimento di Ingegneria Civile, Ambientale, Aerospaziale e dei Materiali (DICAM), Università di Palermo, Viale delle Scienze, 90128 Palermo, Italy

R. Macaluso, M. Mosca, and C. Calì Thin Films Laboratory, Dipartimento di Energia, Ingegneria dell’Informazione e modelli Matematici (DEIM), Università di Palermo, Viale delle Scienze, 90128 Palermo, Italy ABSTRACT: Photoelectrochemical polymerization of poly(3,4-ethylenedioxythiophene), PEDOT, was successfully realized on anodic film grown to 50 V on magnetron sputtered Ti−6 atom % Si alloys. Scanning electron microscopy allowed us to evidence formation of compact and uniform polymer layers on the oxide surface. Photoelectrochemical and impedance measurements showed that photopolymerization allows one to grow PEDOT in its conducting state, while a strong cathodic polarization is necessary to bring the polymer in its p-type semiconducting state. Information on the optical and electrical properties of metal/ oxide/polymer/electrolyte junctions proves that PEDOT has promising performance as an electrolyte in metal/insulator/metal structures, while its use in metal/oxide/semiconductorbased devices is negatively influenced by the occurrence of lithium intercalation phenomena during the dedoping process.

Since 1999,8 poly(3,4-ethylenedioxythiophene) has been successfully proposed as a suitable conducting polymer for a tantalum solid capacitor due to its excellent environmental stability. PEDOT is usually synthesized chemically in aqueous medium in the presence of an oxidizer agent (i.e., ferric sulfate), even if this method can leave strongly acidic oxidant on the polymer chains, which can cause degradation of the dielectric. Thus, more recently, a photoelectrochemical method has been suggested to grow conducting polymers on high k oxides.9−13 Oxides with selected εox/dox can be easily prepared by anodizing, a room-temperature wet electrochemical process, which gives the chance to tune the electronic properties of the films by accurate selection of the alloy composition and anodizing conditions. In ref 14 it is proved how alloying of Ti with a small amount of Si allows one to grow amorphous anodic oxides with a high dielectric constant up to a formation voltage of 100 V. The presence of silicon is reported to hinder the occurrence of crystallization phenomena responsible for anodic TiO2 breakdown. Since a low electronic current is expected to circulate after oxide formation, due the dielectric properties of these oxides, a photoelectrochemical process can be designed to allow for growth of the semiconducting polymer necessary to complete the oxide/conducting polymer junction.

1. INTRODUCTION After the discovery of conducting polymers and the possibility to modify their electrical properties (from insulating to metallic) by doping and a careful choice of the processing conditions, a large amount of research efforts have been devoted to exploit their possible application in many technological fields, including large area organic electronics, polymer photovoltaic cell, sensors, and electrolytic capacitors.1−3 The latter consist of a valve metal (e.g., aluminum or tantalum), covered by an anodic film, the dielectric, in direct contact with an electrolyte, and the true counter electrode, which in turn is connected to a current collector, usually made by the same valve metal. The performance of these devices is strictly related to the dielectric properties of the oxide (i.e., the dielectric constant, ε, and the leakage current) and the employed electrolyte. Anodizing valve metals alloys of selected composition are currently being investigated as a viable route to prepare oxides of tailored solid state properties, such as band gap, flat band potential, and dielectric constant.4−7 However, the capacitance and impedance characteristics are also influenced by the composition of the electrolyte and by its state (liquid or solid). An electrolytic capacitor with electron-conducting polymer electrolyte showed low impedance and high capacitance in the high-frequency range, even if they have the disadvantage of a difficult control of the thickness and a complicated manufacturing process.1−3 © 2014 American Chemical Society

Received: October 8, 2014 Revised: November 24, 2014 Published: November 26, 2014 29973

dx.doi.org/10.1021/jp510169r | J. Phys. Chem. C 2014, 118, 29973−29980

The Journal of Physical Chemistry C

Article

In this work we report on the photoelectrochemical polymerization of PEDOT on an anodic film grown to 50 V on magnetron-sputtered Ti−6 atom % Si. The morphology and composition of the polymer were studied by scanning electron microscopy (SEM) and Raman spectroscopy. The prepared junctions were characterized by photocurrent spectroscopy and impedance measurements in order to test their performance as MIM structures, while their use as MOS to be integrated in metal/oxide/semiconductor field effect transistor (MOSFET), proposed in ref 13, was critically reviewed on the basis of new electrochemical experimental findings on the occurrence of lithium intercalation phenomena, which change the electrical properties of the oxide.

2. EXPERIMENTAL SECTION Ti−6 atom % Si alloy was prepared by cosputtering of 99.5% titanium and 99.999% silicon using a dc magnetron sputtering method. The alloy was deposited as a layer about 200 nm thick on glass substrates. Deposited layers were anodized to 50 V at 5 mA cm−2 in phosphoric acid at room temperature.14−17 According to transmission electron micrographs of ultramicrotomed sections relating to anodic films grown on sputter-deposited Ti−6 atom %vSi in 1 mol dm−3 H3PO4 at 5 mA cm−2,15−17 an anodizing ratio of ∼2 nm V−1 can be estimated, which allows one to estimate for the films a thickness of ∼100 nm. Polymerization was performed in 0.1 M LiClO4 propylene carbonate (PC, 99.8 Sigma-Aldrich) with 0.04 M 3,4ethylenedioxythiophene (EDOT) under irradiation (λ = 320 nm).11,12 The PEDOT morphology was investigated using a Philips XL30 ESEM scanning electron microscope. Some of the PEDOT films were detached from the substrate with copper-conducting adhesive in order to allow the view of the electrode side polymer. The metal/oxide/conducting polymer/electrolyte junction was investigated by PCS. The photoelectrochemical setup is described elsewhere.18−20 Briefly, a 450 W UV−vis xenon lamp coupled with a monochromator irradiates the specimen through a quartz window in the anodizing cell. A two-phase lock-in amplifier with a mechanical chopper (chopping frequency =13 Hz) enables separation of the photocurrent from total current. During the measurements a UV filter (λcutoff = 400 nm) was inserted to avoid a doubling effect on the measured photocurrent in the high-wavelength region. Electrochemical impedance spectroscopy (EIS) data were obtained using a Parstat 2263 (PAR), controlled by a computer via Electrochemistry PowerSuite software. A three-electrode arrangement was used consisting of the anodized specimen (with and without polymeric film on the surface), a reference electrode (silver/silver chloride), and a Pt net having a very high specific area, immersed in different solution (0.25 M Na2SO4 and 0.1 M LiClO4). Impedance spectra were generated by applying a sinusoidal signal of amplitude 10 mV over the frequency range 0.1 Hz−100 kHz. Resultant spectra were analyzed with Zview software.19

Figure 1. (a) Photocurrent vs potential curve relating to an anodic film grown on Ti−6 atom % Si to 50 V, recorded at 10 mV s−1. (b) Current vs time curve recorded at 8 V (Ag/AgCl). (c) Current vs time curve recorded at −1.2 V (Ag/AgCl). D = dark and L = light. Sol: 0.1 M LiClO4. Propylene carbonate, λ = 320 nm.

In Figure 1a we report the photocurrent vs potential curve (photocharacteristic) recorded in 0.1 M LiClO4 PC electrolyte by irradiating the anodic oxide grown on Ti−6 atom % Si at λ = 320 nm, i.e., for a photon energy of ∼3.87 eV, thus higher than the band gap of the oxide, Eg ≈ 3.45 eV according to previous experimental findings.14 An inversion of the photocurrent sign is revealed by moving the polarizing voltage toward the cathodic direction, as suggested by the dependence of the phase angle on the electrode potential and as confirmed by comparing the current circulating in the dark and under irradiation (see Figure 1b and 1c). The time necessary to reach the steady state Iph value depends on hν, the electric field, and the impinging photon flux. Inversion of the photocurrent sign is typical of an insulating material, for which both anodic and cathodic photocurrent can be measured depending on the direction of the electric field, i.e., on the applied potential with respect to the flat band potential, UFB.20 Thus, the inversion photocurrent potential (ca. −0.2 vs Ag/AgCl) can be assumed as a proxy of UFB for the investigated oxide. This value is comparable to that estimated in aqueous solution for the same oxide14 if we assume the anodic film at the isoelectric point in propylene carbonate (pHpzc = 5.8 for TiO2,21) and if we consider the Nerstian dependence of the flat band potential on pH valid in aqueous solution

3. RESULTS AND DISCUSSION 3.1. Energetics of the Metal/Oxide/Electrolyte Interface. Photoelectrochemical measurements were performed to get information on the energetics of the anodic film/electrolyte interface and, thus, to design the photoelectrochemical deposition of conducting polymer on anodic oxide.

Ufb = Ufb pHpzc − 0.059(pH − pH pzc)

(1)

The supralinear behavior of the photocharacteristic (see Figure 1) is very similar to that shown for anodic film grown on Ti−6 atom 29974

dx.doi.org/10.1021/jp510169r | J. Phys. Chem. C 2014, 118, 29973−29980

The Journal of Physical Chemistry C

Article

% Si at a final voltage of 40 V,14 according to the amorphous nature of these anodic films. By polarizing the electrode at a potential more negative than the inversion photocurrent potential it was possible to record cathodic photocurrent spectra, i.e., photocurrent vs wavelength curves, as shown in Figure 2 at UE = −2 V vs Ag/AgCl. From

where Iph is the photocurrent yield, assumed proportional to the light absorption coefficient, hν is the photon energy, Eg is the optical band gap, and an n value equal to 0.5 has been assumed for indirect optical transitions and n value equal to 2 for direct optical transitions.22 By assuming nondirect optical transition (see inset), an optical band gap of ∼3.5 eV can be estimated for this oxide by extrapolating to zero the (Iphhν)0.5 vs hν plot, which is very close to the value estimated under anodic polarization, the small difference being attributed to the low electric field. This value is higher than that reported for crystalline anatase (Eg = 3.2 eV) and rutile (3.05 eV) and is very close to the mobility gap expected for amorphous anodic TiO2.23 Knowing the mobility gap and the flat band potential of the oxide it was possible to sketch the energetics of the metal/oxide/ electrolyte interface as depicted in Figure 3, where the valence band mobility edge of amorphous Si-containing anodic TiO2 was assumed coincident with that of pure TiO2.24 In both cases the valence band is constituted by the occupied O2− orbitals; thus, the presence of silicon in the outer part of the film is not expected to have a significant influence on the energy location of the valence band edge. As shown in the scheme, such energetic level was ∼2.2 eV below the reference electrode, i.e., ∼0.4 eV below the oxidation potential of propylene carbonate (Eox = 1.8 V vs Ag/AgCl, according to ref 25). 3.2. Photoelectropolymerization. In Figure 4a we report the current vs time curve recorded under irradiation (λ = 320 nm) by polarizing the anodic oxide grown on Ti−6 atom % Si at UE = 8 V (Ag/AgCl) in 0.1 M LiClO4 PC solution containing 0.04 M EDOT. By irradiating the anodic oxide surface with a photon energy higher than the band gap of the

Figure 2. Raw photocurrent spectrum relating to an anodic film grown to 50 V on Ti−6 atom % Si recorded by polarizing the electrode in 0.1 M LiClO4 Propylene carbonate at −2 V (Ag/AgCl). (Inset) (Iphhν)0.5 vs hν plot.

the photocurrent spectra for photon energy in the vicinity of the band gap it is possible to estimate the optical band gap value of the oxide according to the following equation (Iphhν)n ∝ (hν − Eg )

(2)

Figure 3. Sketch of the energetic levels at the Ti−6 atom % Si/anodic film/organic solution interface. 29975

dx.doi.org/10.1021/jp510169r | J. Phys. Chem. C 2014, 118, 29973−29980

The Journal of Physical Chemistry C

Article

At longer photoelectropolymerization time a deviation from the linear slope is observed that can be explained by considering that photo-oxidation occurs on a surface progressively larger with respect to the initial one owing to a spot light smaller that the whole exposed surface. A possible variation in the efficiency of the photo-oxidation of EDOT cannot be excluded, although we have no experimental evidence supporting this hypothesis. According to the energetics of the oxide/electrolyte interface of Figure 3 it is evident that oxidations of both EDOT and PC are thermodynamically possible. The occurrence of photoelectrochemical polymerization was confirmed by direct visual inspection for longer photodeposition times and SEM analysis of the samples (see below) at short times. The polymer shows good adhesion to the substrate, remaining stuck on the oxide surface even after in situ electrochemical and photoelectrochemical experiments (see below) and after drying once out on the electrolyte. A morphological study of PEDOT photoelectrochemically grown on anodic oxides on Ti−6 atom % Si electrodes was carried out, and in Figure 5 we report the SEM micrographs relating to PEDOT surface solution side (Figure 5a and 5b), oxide side (Figure 5c and 5d), as well as cross-section (Figure 5e). The morphologies in both cases are really different with respect to those evidenced for electrochemically grown polymer on gold11 but similar with respect to those showed for photoelectrochemically grown polymer on anodic films.11,12 In fact, solution-side micrographs (Figure 5a and 5b) reveal that the surface of the polymer is slightly corrugated, with the presence of compact and flat regions delimited by interconnected ridges of polymeric material. This morphology has been previously described for polypyrrole film electrodeposited or chemically prepared on different substrates.26 Shrinkage and drying of the polymer can be discarded as possible source of wrinkle formation and originate at the oxide/polymer interface during the growth process due to polymer−oxide interaction governing the adhesion process26 and/or substrate morphology.27−29 Such ridges are richer in dopant ions, which are reported to be incorporated into the volume of the wrinkles during their thickening.29 The polymer is very uniform and compact all over the surface. By averaging the measured thickness on several areas of the same sample and several samples it was possible to estimate an efficiency of ∼70% for the growth process, which is very close to that estimated for electrochemically grown polymer.11 SEM micrographs of Figure 5 show a very compact PEDOT film grown on anodic oxide. 3.3. MIM Junction. In Figure 6a we report the EIS spectra in the Bode representation relating to the 50 V anodic film grown on Ti−6 atom % Si before and after photopolymerization, recorded by polarizing the electrode in 0.1 M Na2SO4 at 3 V vs Ag/AgCl. As clearly visible in the figure, both the impedance modulus and the phase angle dependence on the ac signal frequency are not strongly influenced by the presence of PEDOT. This can be better appreciated by comparing the parameters derived for polymer-free and polymer-covered anodic film (see Table 1), derived by a best-fitting procedure according to the equivalent circuit of Figure 6a, where R is the series resistance, Rox is the anodic film resistance, and Qox is a constant phase element (CPE) introduced to model the oxide capacitance. The high-frequency |Z| asymptotes, i.e., the resistance of the electrolyte, are almost coincident, thus suggesting that the series contribution due the presence of polymer in its conductive state is on the order of 20%.

Figure 4. (a) Current vs time curve recorded by polarizing the electrode at UE = 8 V (Ag/AgCl) in 0.1 M LiClO4−0.04 M Edot propylene carbonate under irradiation at constant wavelength (λ = 320 nm). D = Dark (interrupting irradiation). (b) Fitting of the ln(Iph/ Iph(0)) vs circulated charge density (q) from Figure 4a.

anodic film (λ = 320 nm), growth of polymer can occur according to the following photon-assisted electropolymerization reaction11,12 + − TiO2 + hυ → hVB + eCB

(3a)

+ n Edot(sol) + nhVB → n Edot+·(sol)

(3b)

n Edot+·(sol) +zClO−4 → (n − z)H + + [(Edot n)z + (ClO−4 )z ] (3c)

with step 3bbeing thermodynamically possible according to Figure 3. Thus, an anodic photocurrent circulates sustained by oxidation of EDOT and its consequent polymerization. As shown in Figure 4, soon after irradiation, Iph increases, reaches a maximum, and at t ≈ 10 min starts to decrease almost exponentially as a function of the circulated charge, i.e., polymer layer thickness. This behavior can be explained by the reduced number of photons reaching the oxide surface with increasing polymer thickness.10−12 If we assume that the growing polymer behaves like an absorbing metallic layer with a constant light absorption coefficient, α, it is possible to fit the experimental curve (see Figure 4b) according to the following equation Iph(qph) = Iph(0)exp( −4.9 × 10−4αPEDOTqph)

(4)

where Iph is the photocurrent, α the light absorption coefficient of the growing polymer, and qph the photocirculated charge per unit area. From the initial slope of the Iph vs qph curve a light absorption coefficient of ∼4.5 × 104 cm−1 can be estimated by eq 4, which is in agreement with the values estimated with the same procedure during the early stage of photoelectropolymerization of PEDOT on Nb2O5 and of polypyrrole on Ta2O5.8,9 29976

dx.doi.org/10.1021/jp510169r | J. Phys. Chem. C 2014, 118, 29973−29980

The Journal of Physical Chemistry C

Article

Figure 5. SEM micrographs relating to PEDOT grown photoelectrochemically (λ = 320 nm) at UE = 8 V (Ag/AgCl) on anodic film after circulation of ∼60 mC cm−2: (a and b) solution-side surface morphologies, (c and d) oxide-side morphology, and (e) cross section.

This behavior is confirmed by the differential capacitance, i.e., the capacitance vs potential curves recorded by superimposing to the continuous signal a small (10 mV) ac signal (f = 10 kHz), reported in Figure 6b. The curves are almost overlapped for a wide potential range (1−5 V). The metallic behavior of the polymer is confirmed by the photocurrent spectrum recorded in 0.1 M LiClO4 in PC at UE = 1 V vs Ag/AgCl, as shown in Figure 7. The presence of a PEDOT layer in its metallic state adherent to the surface of the anodic film does not modify appreciably the shape of the anodic photocurrent spectra with respect to those recorded in the absence of polymer14 nor the value of the measured optical band gap. In fact, its value (3.45 eV) is coincident with that reported in ref 14, thus suggesting that the measured photocurrent is

sustained by the photoholes generated in the valence band of the titanium oxide. 3.4. MOS Junction. In order to obtain a semiconducting polymer to fabricate a metal−oxide−semiconductor junction employed in a OFET (ref 13), a reduction step, i.e., PEDOT dedoping, was performed according to eq 4 [(EDOTn)z + (ClO−4 )z ] + x e− → [(EDOTn)(z − x) + (ClO−4 )(z − x)] + x(ClO4 )−

(5)

The redox equilibrium potential for PEDOT dedoping/ reduction on Au is −0.5 V (Ag/AgCl) in PC,11 but a more negative potential is expected to be necessary due to potential drop across the insulating layer.11,12 However, under strong 29977

dx.doi.org/10.1021/jp510169r | J. Phys. Chem. C 2014, 118, 29973−29980

The Journal of Physical Chemistry C

Article

Figure 8. Cyclic voltammetry relating to anodic film grown to 50 V on Ti−6 atom % Si (without and with PEDOT on the surface, after circulation of ∼60 mC cm−2) recorded at 10 mV·s−1 in 0.1 M LiClO4 propylene carbonate.

at −1 V vs Ag/AgCl for CV recorded in the presence of lithium ions, which can be attributed to their intercalation inside the anodic film.30 This behavior is confirmed by impedance measurements recorded in 0.1 M LiClO4 in PC solution. In Figure 9a and 9b we report the EIS spectra recorded at several electrode potentials relating to 50 V anodic films on Ti−6 atom % Si. It is evident that by changing the polarizing voltage both the impedance modulus and the phase angle significantly change in the low-frequency range. More specifically, the lower |Z| is the more negative the polarizing voltage, while the phase angle suggests a resistive behavior of the film. These findings can be explained by the occurrence of lithium ions intercalation phenomena (see Figure 8), which induce an increase in the film conductivity.30 As shown in Figure 8, when the oxide is covered by PEDOT, a higher current circulates in the same potential range due to the occurrence of polymer dedoping;11 thus, if we want to bring PEDOT in its p-type semiconductor state we cannot avoid oxide lithium insertion with dramatic consequences on the dielectric properties of the oxide. This is confirmed by the impedance spectra reported in Figure 9c and 9d recorded after photoelectropolymerization, which are strongly dependent on the polarizing voltage. At 1 V Ag/AgCl, the impedance almost coincides with that of polymer-free oxide, but under more negative potential the overall impedance decreases by increasing the polarizing potential toward the cathodic direction down to −2 V. For more cathodic potential the presence of the dedoped polymer (i.e., p-type semiconductor) is responsible for a higher |Z| with respect to the corresponding values recorded in the case of polymer-free oxide (compare the low-frequency region of Figure 9a and 9c). The best-fitting procedure of EIS spectra reported in Figure 9 (not reported here) suggests that the impedance of the junction can be modeled by introducing several time constants accounting for a nonuniform lithium insertion across the oxide (occurring for both PEDOT-free and PEDOT-covered oxide) and also for the presence of p-type PEDOT for polymer-covered samples. The dramatic change in the dielectric properties of the anodic film on Ti−6 atom % Si can account for the poor performance of the prepared junction as MOSFET, reported in ref 13. This statement is confirmed by the photoelectrochemical investigation performed on the metal/oxide/polymer junction after dedoping for ∼30 min at −3 V vs Ag/AgCl. A cathodic

Figure 6. (a) EIS spectra in the Bode representation relating to anodic film grown to 50 V on Ti−6 atom % Si (without and with PEDOT) recorded by polarizing the film at 3 V vs Ag/AgCl in 0.25 M Na2HPO4. (b) Measured series capacitance vs potential curves recorded in 0.25 M Na2HPO4, frequency =1 kHz.

Table 1. Fitting Parameters Relating to EIS Spectra of Figure 6a without polymer with polymer

R (Ω·cm2)

Rox (Ω·cm2)

Qox (S·sncm−2)

n

29 36

1.0 × 10 7.7 × 106

7.8 × 10−7 6.7 × 10−7

0.87 0.80

7

Figure 7. Raw photocurrent spectrum relating to anodic film grown to 50 V on Ti−6 atom % Si with PEDOT grown photoelectrochemically (after circulation of ∼60 mC cm−2) recorded by polarizing the metal/ oxide/PEDOT interface in 0.1 M LiClO4 propylene carbonate at 1 V (Ag/AgCl). (Inset) (Iphhν)0.5 vs hν plot.

cathodic polarization both lithium ions intercalation in TiO230 and solvent reduction25 can occur. Therefore, a cyclic voltammetry (CV) based investigation was necessary to study such processes. In Figure 8 we report the CV recorded by sweeping the electrode potential at 10 mV s−1 in 0.1 M LiClO4 in PC. It is evident that a significant cathodic current starts 29978

dx.doi.org/10.1021/jp510169r | J. Phys. Chem. C 2014, 118, 29973−29980

The Journal of Physical Chemistry C

Article

Figure 10. (a) Raw photocurrent spectrum relating to PEDOT photoelectrochemically grown after circulation of ∼60 mC cm−2 on anodic oxide electrode, recorded by polarizing the metal/oxide/ PEDOT interface in 0.1 M LiClO4 propylene carbonate at −3 V (Ag/ AgCl). (b) (Iphhν)2 vs hν plot, and (c) (Iphhν)0.5 vs hν plot.

Figure 11. Raman spectra relating to PEDOT films photoelectrochemically grown after circulation of ∼60 mC cm−2 on anodic oxide before and after dedoping electrochemical processes. Figure 9. EIS spectra in the Bode representation relating to anodic film grown on Ti−6 atom % Si to 50 V (without and with PEDOT on the surface, after circulation of ∼60 mC cm−2) recorded by polarizing the electrode at several voltage vs Ag/AgCl in 0.1 M LiClO4 propylene carbonate: (a and b) without PEDOT and (c and d) with PEDOT.

oligomers extrapolated to infinite chain length.32 The lower band gap measured in the hypothesis of indirect optical transitions could be attributed to formation of a band of defects (polaron and/or bipolaron) near the valence band of PEDOT during the dedoping process.33 The occurrence of the dedoping process is supported by Raman analysis. In Figure 11 we report the Raman spectra relating to PEDOT soon after photoelectrodeposition and after dedoping process. According to ref 34 a shift toward lower wavenumber of the symmetric CαCβ bond is expected as a consequence of the dedoping process. As shown in the inset of Figure 11 such band shifts from 1436 cm−1 for as-prepared PEDOT to 1424.7 cm−1 for dedoped polymer are as expected for a not fully completed dedoping process.34

photocurrent spectrum is recorded for photon energies lower than the band gap of the anodic oxide (see Figure 10a). From such spectra and according to eq 2 we can estimate a direct and an indirect band gap of 1.63 (Figure 10b) and 1.43 eV (Figure 10c), respectively, both values being very close to those reported in previous work for PEDOT.31 Moreover, the direct optical band gap of PEDOT agrees quite well with the theoretically estimated HOMO−LUMO gap (1.68−1.83 eV) of 29979

dx.doi.org/10.1021/jp510169r | J. Phys. Chem. C 2014, 118, 29973−29980

The Journal of Physical Chemistry C

Article

(13) Mosca, M.; Macaluso, R.; Randazzo, G.; Di Bella, M.; Caruso, F.; Calì, C.; Di Franco, F.; Santamaria, M.; Di Quarto, F. Anodized TiSi Alloy as Gate Oxide of Electrochemically-Fabricated Organic FieldEffect Transistors. Electrochem. Solid-State Lett. 2014, 3, P7−P9. (14) Di Quarto, F.; Di Franco, F.; Monarca, C.; Santamaria, M.; Habazaki, H. Photoelectrochemical Characterization of Amorphous Anodic Films on Ti-6 atom % Si. Electrochim. Acta 2013, 110, 517− 525. (15) Habazaki, H.; Uozumi, M.; Konno, H.; Shimizu, K.; Skeldon, P.; Thompson, G. E. Crystallization of Anodic Titania on Titanium and Its Alloys. Corros. Sci. 2003, 45, 2063−2073. (16) Habazaki, H.; Shimizu, K.; Nagata, S.; Skeldon, P.; Thompson, G. E.; Wood, G. C. Ionic Transport in Amorphous Anodic Titania Stabilised by Incorporation of Silicon Species. Corros. Sci. 2002, 44, 1047−1055. (17) Tauseef Tanvir, M.; Fushimi, K.; Shimizu, K.; Nagata, S.; Skeldon, P.; Thompson, G. E.; Habazaki, H. Influence of Silicon on the Growth of Barrier-Type Anodic Films on Titanium. Electrochim. Acta 2007, 52, 6834−6840. (18) Di Quarto, F.; La Mantia, F.; Santamaria, M. Modern Aspects of Electrochemistry; Springer: New York, 2009. (19) La Mantia, F.; Santamaria, M.; Di Quarto, F.; Habazaki, H. Physicochemical Characterization of Thermally Aged Anodic Films on Magnetron-Sputtered Niobium. J. Electrochem. Soc. 2010, 157, C258− C267. (20) Santamaria, M.; Di Quarto, F.; Habazaki, H. Influences of Structure and Composition on the Photoelectrochemical Behaviour of Anodic Films on Zr and Zr-20 at.%Ti. Electrochim. Acta 2008, 53, 2272−2280. (21) McCafferty, E. A. Surface Charge Model of Corrosion Pit Initiation and of Protection by Surface Alloying. J. Electrochem. Soc. 1999, 146, 2863−2869. (22) Mott, N. F.; Davis, E. A. Electronic Processes in Non-crystalline Materials; Clarendon Press: Oxford, U.K., 1979. (23) Piazza, S.; Calà, L.; Sunseri, C.; Di Quarto, F. Influence of the Crystallization Process on the Photoelectrochemical Behaviour of Anodic TiO2 Films. Ber. Bunsen-Ges. Phys. Chem. 1997, 101, 932−942. (24) Gerisher, H. Topics in Applied Physics, Solar Energy Conversion; Springer-Verlag Heildeberg: New York, 1979. (25) Bard, A. J.; Faulkner, L. R. Electrochemical Methods: fundamentals and applications; John Wiley & Sons: Hoboken, 2000. (26) Yuan, W.-L.; O’Rear, E. A.; Cho, G.; Funkhouser, G. P.; Glatzhofer, D. T. Thin Polypyrrole Films Formed on Mica and Alumina with and without Surfactant Present: Characterization by Scanning Probe and Optical Microscopy. Thin Solid Films 2001, 385, 96−108. (27) Shapiro, J. S.; Smith, W. T.; MacRae, C. Wrinkle Morphology of Polypyrrole Films Grown in Thin-Layer Cells. Synth. Met. 1995, 69, 505−506. (28) Shapiro, J. S.; Smith, W. T.; MacRae, C. Morphology of Polypyrrole Films Grown in Thin-Layer Cells and on Indium-Tin Oxide Conductive Glass. Polymer 1995, 36, 1133−1140. (29) Miles, M. J.; Smith, W. T.; Shapiro, J. S. Morphological Investigation by Atomic Force Microscopy and Light Microscopy of Electropolymerised Polypyrrole Films. Polymer 2000, 41, 3349−3356. (30) Churikovz, A. V.; Zobenkova, V. A.; Pridatko, K. I. Lithium Intercalation into Titanium Dioxide Films from a Propylene Carbonate Solution. Russ. J. Electrochem. 2004, 40, 63−68. (31) Kertesz, M.; Choi, C. H.; Yang, S. Conjugated Polymers and Aromaticity. Chem. Rev. 2005, 105, 3448−3481. (32) Zade, S. S.; Bendikov, M. From Oligomers to Polymer: Convergence in the HOMO-LUMO Gaps of Conjugated Oligomers. Org. Lett. 2006, 8, 5243−5246. (33) Vardeny, Z. V.; Wei, X. Handbook of Conducting Polymers; Marcell Decker: New York, 1998. (34) Chiu, W. W.; Travas -Sejdic, J.; Cooney, R. P.; Bowmaker, G. A. Studies of Dopant Effects in Poly(3,4-ethylenedioxythiophene) Using Raman Spectroscopy. J. Raman Spectrosc. 2006, 37, 1354−1361.

4. CONCLUSIONS Photoelectropolymerization of EDOT on wide band gap anodic oxide grown to 50 V on Ti−6 atom % Si was realized. According to SEM characterization a compact polymer layer can be grown under potentiostatic polarization and under irradiation with photon energies higher than the band gap of the oxide. The experimental results arising from the photoelectrochemical investigation and impedance measurements suggest that the PEDOT is in its metallic state soon after polymerization, behaves as a good solid state electrolyte, and thus is a promising candidate to be integrated in MIM-based devices. Due to the large potential drop inside the oxide, a strong cathodic polarization is necessary to induce polymer dedoping; thus, a parallel lithium ion insertion reaction occurs with dramatic consequences on the dielectric properties of the oxide, which can account for the poor performances of the prepared junction in MOSFET-based devices.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

REFERENCES

(1) Toita, S.; Inoue, K. One-Pot Synthesis of Polypyrrole Film on an Aluminum Oxide Layer by Electropolymerization in the Presence of Ammonium Borodisalicylate in Acetonitrile. J. Power Sources 2007, 164, 905−910. (2) Jang, K. S.; Moon, B.; Oh, E. J.; Lee, H. Characteristics of Tantalum Electrolytic Capacitors Using Soluble Polypyrrole Electrolyte. J. Power Sources 2003, 124, 338−342. (3) Tsai, M.-L.; Chen, P.-J.; Do, J.-S. Preparation and Characterization of Ppy/Al2O3/Al Used as a Solid-State Capacitor. J. Power Sources 2004, 133, 302−311. (4) Di Franco, F.; Zampardi, G.; Santamaria, M.; Di Quarto, F.; Habazaki, H. Characterization of the Solid State Properties of Anodic Oxides on Magnetron Sputtered Ta, Nb and Ta-Nb Alloys. J. Electrochem. Soc. 2012, 159, C33−C39. (5) Santamaria, M.; Di Franco, F.; Di Quarto, F.; Skeldon, P.; Thompson, G. E. Tailoring of the Solid State Properties of Al-Nb Mixed Oxides: a Photoelectrochemical Study. J. Phys. Chem. C 2013, 117, 4201−4210. (6) Di Franco, F.; Santamaria, M.; Di Quarto, F.; La Mantia, F.; De Sà, A. I.; Rangel, C. M. Dielectric Properties of Al-Nb Amorphous mixed oxides. ECS J. Solid State Sci. Technol. 2013, 2, N205−N210. (7) Santamaria, M.; Di Quarto, F.; Skeldon, P.; Thompson, G. E. Effect of Composition on the Photoelectrochemical Behavior of Anodic Oxides on Binary Aluminum Alloys. J. Electrochem. Soc. 2006, 153, B518−B526. (8) Kudoh, Y.; Akami, K.; Matsuya, Y. Solid Electrolytic Capacitor with Highly Stable Conducting Polymer as a Counter Electrode. Synth. Met. 1999, 102, 973−974. (9) Ueno, K.; Dominey, L.; Uchi, H.; Alwitt, R. S. Impedance Properties of Photoelectrochemically and Chemically Prepared Ta/ PEDOT Capacitors. J. Surf. Finish. Soc. Jpn. 2009, 60, 179−185. (10) Di Quarto, F.; Figà, V.; Bocchetta, P.; Santamaria, M. Photoelectrochemical Synthesis of Polypyrrole on Anodic Ta2O5 Films. Electrochem. Solid-State Lett. 2007, 10, H305−H308. (11) Di Franco, F.; Bocchetta, P.; Santamaria, M.; Di Quarto, F. Light Induced Electropolymerization of Poly(3,4-ethylenedioxythiophene) on Niobium Oxide. Electrochim. Acta 2010, 56, 737−744. (12) Di Franco, F.; Bocchetta, P.; Calì, C.; Mosca, M.; Santamaria, M.; Di Quarto, F. Electrochemical Fabrication of Metal/Oxide/ Conducting Polymer Junction. J. Electrochem. Soc. 2011, 158, H50− H54. 29980

dx.doi.org/10.1021/jp510169r | J. Phys. Chem. C 2014, 118, 29973−29980