Organic Electronics 14 (2013) 2980–2986

Contents lists available at ScienceDirect

Organic Electronics

journal homepage: www.elsevier .com/locate /orgel

The interplay of electronic structure, molecular orientation andcharge transport in organic semiconductors: Poly(thiophene)and poly(bithiophene)

1566-1199/$ - see front matter � 2013 Elsevier B.V. All rights reserved.http://dx.doi.org/10.1016/j.orgel.2013.08.022

⇑ Corresponding authors. Address: Federal University of Rio de Janeiro,21941-909 Rio de Janeiro, Brazil(M.L.M. Rocco), Federal University ofParaná, 81531-990 Curitiba, Brazil(L.S. Roman). Tel.: +55 21 2562 7786;fax: +55 21 2562 7265.

E-mail addresses: lsroman@fisica.ufpr.br (L.S. Roman), luiza@iq.ufrj.br(M.L.M. Rocco).

Y. Garcia-Basabe a,b, B.G.A.L. Borges a, D.C. Silva c, A.G. Macedo d, L. Micaroni e, L.S. Roman c,f,⇑,M.L.M. Rocco a,⇑a Institute of Chemistry, Federal University of Rio de Janeiro, 21941-909 Rio de Janeiro, Brazilb Department of Physics, Faculty of Engineering, University of Matanzas ‘‘Camilo Cienfuegos’’, Matanzas 40100, Cubac Engineering and Materials Science (PIPE), Federal University of Paraná, 81531-990 Curitiba, Brazild Department of Physics, Technological Federal University of Paraná, 80230-901 Curitiba, Brazile Department of Chemistry, Federal University of Paraná, 81531-990 Curitiba, Brazilf Department of Physics, Federal University of Paraná, 81531-990 Curitiba, Brazil

a r t i c l e i n f o a b s t r a c t

Article history:Received 16 July 2013Received in revised form 23 August 2013Accepted 27 August 2013Available online 8 September 2013

Keywords:Core-hole clock methodMorphologyElectron delocalization dynamicsPoly(thiophene)Poly(bithiophene)Photovoltaics

Ultrathin films of poly(thiophene) (PT) and poly(bithiophene) (PBT) were prepared by elec-trochemical route using ionic liquid (BFEE) as medium and electrolyte. Distinct morpholo-gies and electrical properties were observed in these materials. To evaluate its response inphotovoltaics, these films were used as active layer in bilayer geometry solar cells with theelectron acceptor molecule C60. The best performance was observed for PT films. In order toprobe the differences in molecular dynamics and structural order, ultrafast electrondynamics in the low-femtosecond regime was evaluated by resonant Auger spectroscopyusing the core–hole clock method at the sulfur K absorption edge. Electron delocalizationtimes for the different polymeric films were derived as a function of the excitation energy.Photoabsorption measurements were conducted and molecular orientation derived. Theseresults corroborated with the morphology found for these films and thus the performanceof PT and PBT in the devices, and with the proposed conduction mechanism.

� 2013 Elsevier B.V. All rights reserved.

1. Introduction

The research concerning conjugated polymers for pho-tovoltaic devices has been focused on the power conver-sion efficiency, the issues of stability, operational lifetime,and processing [1–6]. Also, it should be emphasized thatthe combination of high efficiency, long lifetime, and largearea processing has not been achieved for the same mate-rial currently. To meet this challenge of unification, the

delicate interplay between many parameters such as pro-cessing, morphology, carrier transport, photochemistry,and molecular dynamics at the active layer and interfacesneeds to be controlled [4,7–11].

Poly(thiophene) (PT) is one of the most studied con-ducting polymer and shows good stability to oxygen andmoisture in both the dedoped and doped forms but it isnecessary high potentials for the polymerization of the thi-ophene monomer, over the oxidation potential of thepoly(thiophene), that may cause an irreversible oxidationof the polymer chains [12–16]. Lower oxidation potentialhas been found for bithiophene in comparison withthiophene [17,18]. The study performed also suggests thatpoly(bithiophene) (PBT) has the highest averageconjugation length of the polymer chains. PBT is generally

Fig. 1. Height (top) and phase (bottom) AFM images of (a) PT and (b) PBT prepared in BFEE.

400 450 500 550 600 650 7000.0

0.2

0.4

0.6

0.8

1.0

Nor

mal

ised

Abs

orba

nce

(arb

. uni

ts)

Wavelength (nm)

Fig. 2. Absorbance spectra of PT (solid line) and PBT (dotted line) films.Inset: the structure of thiophene and bithiophene monomers.

Y. Garcia-Basabe et al. / Organic Electronics 14 (2013) 2980–2986 2981

synthesized by electrochemical oxidation of bithiophenemonomer in organic media and tends to present higheramount of regular a, a0- linking of the monomeric unitswhile PT films present a disordered cross-linked polymerstructure. However, the reduced solubility of bithiophenemonomer, when compared with the thiophene monomer,in different mediums may reflect negatively in themorphology of the thin film [18,19].

Studies of interfacial electronic structure and ultrafastcharge transfer (CT) are relevant topics for understandingthe transport properties of conjugated polymer andimprove their performance in photovoltaic device applica-tions [20,21]. Pulse laser pump–probe spectroscopy is themost used techniques for CT dynamic studies [22–25]. Inthis case the dynamic of optically excited state as a func-tion of delay time after excitation are investigated. Thistechnique is restricted by the extension of the laser pulse,allowing studying only processes occurring in the timescale above hundred femtoseconds (10�15 s). Besides la-ser-based pump–probe methods the so-called core–holeclock (CHC) approach using resonant Auger spectroscopyemerge as an alternative, with some advantages [26,27].The short core hole lifetimes used as internal clock in thismethodology allow access to dynamic processes in verylow time scale (in the range of femtoseconds down to hun-dred attoseconds (10�18 s)) and also to probe the atomicspecificity of core levels.

In this work we investigate the impact of usingthiophene or bithiophene monomer in electrochemicalsynthesis of thiophene based polymer using ionic liquidBoron Trifluoride Diethyl Etherate (BFEE). We studied themorphology, electronic structure, optical and electrical

behavior in solar cells. Molecular orientation and chargetransfer dynamics using the core–hole clock methodologywere also investigated in order to add information aboutthe polymeric films regarding their use in photovoltaicdevices.

2. Experimental

A layer of PT or PBT was electrochemically deposited,using a potentiostat AUTOLAB 3530 with GPES system,onto ITO/glass substrates by using three electrodes cell(WE:ITO; CE:Pt plate 1 cm2 and RE:Ag, Ø 2 mm) apply-ing+1.5 V and+0.6 V, to obtain PT and PBT respectively, ina potentiostatic mode from a Boron Trifluoride Diethyl

2982 Y. Garcia-Basabe et al. / Organic Electronics 14 (2013) 2980–2986

Etherate (BFEE) ionic liquid with PT or PBT monomers in aconcentration of 0.02 mol/L. This solution and electropoly-merisation were prepared inside a glove box under a drynitrogen atmosphere. Bilayer solar cells were prepared asfollows: a 30 nm PEDOT:PSS layer was spincoated ontoITO and after annealing at 100 �C under reduced pressure,resulting a 30 nm layer, the as-prepared ITO|PEDOT:PSS|PTor PBT film received layers of 30 nm of C60 and 100 nm ofAluminium (Al) thermally evaporated through a shadowmask at vacuum pressure of 6 � 10�6 mbar. Thicknesswas determined in a Dektak 3 profilometer. Topographyimages were acquired by using an atomic force microscope(AFM, Shimadzu SPM 9500J3) in dynamic mode. Photovol-taic characterization was performed with a Keithleypicoammeter with power supply, model 6487 and amonochromator/spectrometer (1/4 m Oriel). The solar sim-ulation was made using air mass (AM1.5) filter with apower illumination of 100 mW cm�2 from a 150 W OrielXenon lamp. Hole only devices were fabricated in theITO|PT or PBT|Al geometry.

X-ray absorption (XAS) and Auger decay measurementson PT and PBT films were performed at the soft X-ray spec-troscopy (SXS) beamline at the Brazilian Synchrotron LightSource (LNLS). The beamline is equipped with a double-crystal type monochromator using the Si(111) plane withan energy resolution of 0.38 eV at the sulphur K-edge. Theexperimental set-up includes a sample manipulator and aconcentric hemispherical electron energy analyzer housedin an UHV chamber with a base pressure of 10�8 mbar.Photoabsorption spectra were recorded in the total elec-tron yield mode (electron current at the sample) simulta-neously with a photon flux monitor (Au grid). The finaldata was normalized by this flux spectrum to correct forfluctuations in beam intensity. The energy calibrationwas performed by taking the well-known value for the LIII

transition (2p3/2 ? 4d) of metallic molybdenum. Polariza-tion dependence was evaluated by measuring XAS spectraat different incoming X-rays incident angles. Auger decayspectra were measured by the hemispherical electronenergy analyzer employing a pass energy of 20 eV. No

Fig. 3. Angular dependence of sulphur K-edge XAS spectra of PT (left) and PBT (riincident angle is defined as the angle between the incoming photon and the sam

beam damage effects were observed in the photoabsorp-tion and photoemission spectra. The deconvolution of theresonant Auger spectra was performed using linear combi-nation of Gaussian and Lorentzian line-profile functionsand an adjustable background.

3. Results and discussion

The morphology of PT or PBT films are quite different ascan be seen in Fig. 1, where AFM images of a top surface ofPT and PBT films prepared onto ITO|PEDOT:PSS glasssubstrates is presented. PT films are quite homogeneousand compact with low roughness. PBT films are lessuniform and it is formed by big particles, which bring highroughness to the layer. These differences in morphologyare reflected later on in the devices performance.

The films present absorption in the visible range, theabsorption spectra for both polymers are shown in Fig. 2.As observed, they present similar energy gaps of 2 eV, ascalculated from the absorption onsets [28]. The vibronicfeatures on the spectra are quite pronounced for PBT films,indicating that the film is formed by aggregates of thepolymer chains, reflected by its morphology.

The packing of the polymer chains can dictate the speedof the charges can be moved in the material. In order toinvestigate these packing and derive electron delocaliza-tion dynamics in both films, obtained from PT and PBTmonomers onto ITO substrates, we performed X-rayabsorption (XAS) and Auger decay measurements. Fig. 3shows S1s XAS spectra of PT and PBT polymeric films mea-sured as a function of X-ray incident angles. The spectrumshows a sharp peak followed by broad bands, assigned toelectronic transitions from the sulphur 1s electron. Thefirst intense peak corresponds to the overlapping of the S1s p* and S 1s r*(S–C) transitions and the other minorfeatures to higher energy excitations. They are probablydue to Rydberg features and r*(C–C) shape resonances,appearing above sulfur 1s ionization potential. Similar fea-tures were observed in XAS spectra of thiophene-based

ght) polymers. Inset: main transitions region (S1s ? p* and S1s ? r*). Theple surface.

Y. Garcia-Basabe et al. / Organic Electronics 14 (2013) 2980–2986 2983

polymers [29–31]. We will concentrate our analysis onthe two main transitions (S 1s p* and S 1s r*(S–C)). Anincrement in intensity of the S1s–r* transition as opposedto the S1s–p* behavior with the increase of the incidentangle (measured with respect to the substrate surface)can be clearly observed in both PT and PBT XAS spectraof Fig. 3. This opposite behavior suggests that the thio-phene-units in these polymers are preferably orientedwith the molecular plane parallel to the substrate surface.

Fig. 4. Kinetic energies of the S KL2,3L2,3 Auger electrons as a function of photonAuger are as following: p* in blue, r* (S–C) in green and Rydberg in violet linessubtracted. (For interpretation of the references to colour in this figure legend,

Aygül et al. [32] report a vertical orientation of the molec-ular plane with respect to the substrate for poly(3-hexyl-thiophene), prepared by doctor blade casting. However, itis known that molecular orientation is highly dependenton synthesis conditions [32–34] and also on the presenceof side groups [9]. As a result of strong polarizationdependence a parallel to the surface adsorption geometrywas derived for thin layers of thiophene adsorbed onNi(100) [35].

energy for PT (left) and PBT (right). Results of curve fitting for spectator. Normal Auger peaks are in red lines. For all spectra the background wasthe reader is referred to the web version of this article.)

Fig. 4 (continued)

Fig. 5. Charge transfer times as a function of photon energy. Theexponential fittings are also shown.

2984 Y. Garcia-Basabe et al. / Organic Electronics 14 (2013) 2980–2986

On the other hand, the intensity of the S1s–p* transitionfor PBT polymer presents a very weak angular dependencecompared to PT (see inset of Fig. 3) suggesting higher uni-formity for PT films. For a quantitative analysis the figureof merit R was calculated following the formula given byJ. Stöhr [36,37]:

R ¼ ½Ið90�Þ � Ið0�Þ�=½Ið90�Þ þ Ið0�Þ� ð1Þ

where I represents the relative integral intensity of the1s–p* transition obtained by a deconvolution procedureof the XAS spectra. The intensity at 0� is obtained by anextrapolation procedure of the fitted spectra [36,37]. Thevalues of R parameters were �0.67 and �0.40 for PT andPBT polymers, respectively. Considering that for an X-raybeam totally linearly polarized (P = 1) the figure of meritR = �1 corresponds to a pure face orientation, we can con-clude that the molecular plane of PT is more likely parallelto the substrate surface. Indeed, PT molecular plane pre-sents a tilt angle with respect to the normal to the sub-strate of 69.7 ± 4� while for PBT it is equal to 60.5 ± 3�.

Fig. 4 shows sulphur KL2,3L2,3 Auger decay spectra for PTand PBT polymers measured at different photon energies.The results of the deconvolution procedure are also shown.Similar to previous report by Arantes et al. [38], threefeatures are observed in the resonant Auger (RAS) spectrafor PT: two narrow and well-separated peaks at the maxi-mum resonance energy of 2472.6 eV related to the S 1s p*

and S 1s r* (S–C) transitions, and appearing at 2113.7 and2115.4 eV kinetic energies, respectively, and anotherfeature located around 2112 eV kinetic energy. This lastfeature is characteristic of a normal Auger peak asdiscussed previously [38] and the other two peaks corre-spond to spectator Auger electrons. Their intensities com-pete and change with photon energy: the intensity of the

0

4

8

12

16

20EQ

E (%

)

Wavelength (nm)400 450 500 550 600 650 700 0.0 0.1 0.2 0.3 0.4

-4

-3

-2

-1

0

J (m

A/cm

2 )

Voltage (V)

Fig. 6. EQE spectra acquired from bilayer photovoltaic devices having PT (j) or PBT (s) as active layer (left) and J�V characteristics acquired upon AM1.5illumination having PT (j) or PBT (s) as active layer (right).

Y. Garcia-Basabe et al. / Organic Electronics 14 (2013) 2980–2986 2985

normal Auger signal increases, while the intensity of theother two spectator Auger signals decrease for both poly-mers. The presence of these three peaks at the resonanceenergy position is an evidence of electron delocalizationduring the Auger decay occurring in femtosecond regime.However, different behavior between normal and specta-tor Auger intensities is observed for both polymers, whichwill be discussed below. Another feature clearly emergesfor higher photon energies and with �2117 eV kinetic en-ergy. Its position also shifts to higher kinetic energies withphoton energy and therefore this is considered anotherspectator decay channel and assigned to a Rydberg state.The large width of the normal Auger signal (2112.2 eV)compared to the other Auger decay features is related withthe delocalized character of the electron involved in thecharge transfer process [39]. Constraints were applied forthe fitting procedure shown in Fig. 4, namely: the positionof the normal Auger peak was fixed while the other peakscould shift, and the FWHM (Full Width at Half Maximum)values of the normal Auger contribution was kept higherthan the spectator contributions.

The charge transfer time (sCT) was calculated for bothpolymers from the ratio between normal Auger signal(with constant kinetic energy of 2112.2 eV) and spectatorsignals and using S 1s core–hole lifetime sCH = 1.27 fs cor-responding to 0.52 eV lifetime width [40]. The result of thisprocedure is presented in Fig. 5 for photon energies abovethe sulphur K-edge. As reported previously [38] the chargetransfer time presents an exponential behavior as a func-tion of the photon energy. However, shorter time valueswere found for PT especially at photon energies close to S1s p* and S 1s r*(S–C) transitions. This behavior may beassociated to the molecular orientation of PT, which seemsto present a more organized structure, and thereforefavouring a more efficient ultrafast electron delocalization

Table 1Device characteristics.

Active layer Thickness (nm) Rrms (nm) Voc (V)

PT 25 ± 02 3.00 0.35PBT 42 ± 11 9.00 0.46

process. The importance of inter-chain interactions forcharge transport in conjugated polymers was previouslydiscussed for PT [38].

Another important point present in Fig. 5 is related tothe different slope of the curves. Similar analysis per-formed by Ikeura-Sekiguchi et al. [41,42] in a study ofcharge transfer in DNA relates the slope of the curves withthe degree of delocalization of the unoccupied conductionband. Slower slope is observed for PT and therefore it pre-sents a more delocalized conduction band as compared toPBT.

In order to evaluate these differences in PT and PBTfilms regarding their charge transport, we prepare devicesin sandwich structure using two metallic electrodes and PTor PBT active layer in between. The electrodes were chosento be ITO and Aluminum to select hole only transport ofcharges (ITO|(PT or PBT)|Al). From the current–voltagecharacteristics acquired in dark from these devices waspossible to obtain the mobility (l) value of charge carriersusing the Mott–Gurney law, J = 9 e e0lV2/8L3, where V isthe applied voltage, J is the current density, e0 is the per-mittivity of vacuum, e is the permittivity of the polymer,and L is the thickness of the active layer. The mobility cal-culated from the fitting this equation to the experimentalJ�V data was three fold higher for PT than PBT. It was inthe order of 10�6 cm2/Vs for PT films while that for PBTin the order of 10�9 cm2/Vs. Similar devices constructedusing PT and PBT films were used as photovoltaic cells inbi-layer structure with C60. In Fig. 6 it is presented the ac-tion spectra (EQE–External Quantum Efficiency–mono-chromatic incident photon to converted electronefficiency) for PT and PBT based solar cells as well as thecurrent–voltage characteristics upon white light illumina-tion (100 mW/cm2 at AM1.5). The action spectra followthe absorption spectra of the polymer films as can be seen

Jsc (mA/cm2) FF (%) ge (%) EQE (510 nm) (%)

3.60 28.00 0.39 14.00.17 19.00 0.01 03.3

2986 Y. Garcia-Basabe et al. / Organic Electronics 14 (2013) 2980–2986

in Fig. 2. The EQE efficiency at the wavelength of 510 nm is14% for PT based devices and 3.3% for PBT based devices.

From the J�V characteristics obtained under illumina-tion one can obtain important parameters for photovoltaicdevice characterization, following: short-circuit currentdensity (Jsc), open-circuit voltage (Voc) and the fill factor.Our devices presented a fill factor (FF) of 28% and 19%and power conversion efficiency (ge) of 0.39% and 0.01%with an average Voc of 350 mV and 460 mV, for PT andPBT cells, respectively. See Table 1 with the summary.The monomer used in the synthesis affects the efficiencyof OPVs significantly, due to the packing of the polymerchains resulted from the oxidation potential and the poly-merization steps followed by the monomers. In the presentpaper, where the films are synthesized in ionic liquid, PTfilms indeed shows better optoelectronic properties thanPBT.

4. Conclusions

Morphologies and electrical properties were conductedin ultrathin films of poly(thiophene) (PT) and poly(bithi-ophene) (PBT) prepared by electrochemical route using io-nic liquid. PT films present low roughness as compared toPBT. In order to evaluate molecular orientation of thesefilms XAS spectra of PT and PBT was obtained and a moreoriented configuration emerges for the former. ResonantAuger spectroscopy following S K-edge photoexcitationwas applied in order to derive electron delocalizationtimes. Shorter delocalization time was derived for PT.These spectroscopic findings corroborate the resultsobtained when using these films as active layer in bilayergeometry solar cells with C60 as the electron acceptormolecule. Finally, we emphasize the importance of thecore–hole clock methodology to probe femtosecond elec-tron delocalization dynamics in polymeric ultrathin filmsas well its contribution to evaluate material efficienciesin photovoltaic devices.

Acknowledgments

Research partially supported by LNLS–National Syn-chrotron Light Laboratory, Brazil. M.L.M.R. and L.S.R. wouldlike to thank CNPq for financial support. The authors wouldalso like to acknowledge CAPES, CNPq, and the technicalassistance of the soft X-ray group from LNLS.

References

[1] O. Inganas, F. Zhang, M.R. Andersson, Acc. Chem. Res. 42 (2009)1731–1739.

[2] L. Akcelrud, Prog. Polym. Sci. 28 (2003) 875–962.[3] Y.-J. Cheng, S.-H. Yang, C.-S. Hsu, Chem. Rev. 109 (2009) 5868–5923.[4] J. Song, C. Zhang, C. Li, W. Li, R. Qin, B. Li, Z. Liu, Z.J. Bo, Polym. Sci.

APolym. Chem. 48 (2010) 2571–2578.

[5] W. Cai, X. Gong, Y. Cao, Sol. Energy Mater. Sol. Cells 94 (2010) 114–127.

[6] X. Gong, Polym. 53 (2012) 5437–5448.[7] V.I. Arkhipov, P. Heremans, E.V. Emelianova, G.J. Adriaenssens, H.

Bassler, Appl. Phys. Lett. 82 (2003) 3245–3247.[8] R.J. Kline, M.D. McGehee, J. Macromol, Sci. Polym. Rev. 46 (2006) 27–

45.[9] M. Jaiswal, R. Menon, Polym. Int. 55 (2006) 1371–1384.

[10] D. Wynands, M. Levichkova, M. Riede, P. Baeuerle, R. Rentenberger,P. Denner, K.J. Leo, Appl. Phys. 107 (2010). 014517-014517-6.

[11] G. Hukic-Markosian, T. Basel, S. Singh, A.V. Vardeny, Li S, D. Laird,Appl. Phys. Lett. 100 (2012). 213903-213903-5.

[12] R. Valaski, C.D. Canestraro, L. Micaroni, R.M.Q. Mello, L.S. Roman, Sol.Energy Mater. Sol. Cells 91 (2007) 684–688.

[13] R. Valaski, L.S. Roman, L. Micaroni, I.A. Hümmelgen, Eur. Phys. J. E 12(2003) 507–511.

[14] J.M. Pringle, M. Forsyth, D.R. MacFarlane, K. Wagner, S.B. Hall, D.L.Officer, Polym. 46 (2005) 2047–2058.

[15] P. Vanlaeke, G. Vanhoyland, T. Aernouts, D. Cheyns, C. Deibel, J.Manca, P. Heremans, J. Poortmans, Thin Solid Films 511–512 (2006)358–361.

[16] A.G. Macedo, D.C. Silva, N.A.D. Yamamoto, L. Micaroni, R.M. Q Mello,L.S. Roman, Synth. Met. 170 (2013) 63–68.

[17] B. Krische, M. Zagorska, Synth. Met. 33 (1989) 257–267.[18] J. Roncali, Chem. Rev. 92 (1992) 711–738.[19] H.S.O. Chan, S.C. Ng, Prog. Polym. Sci. 23 (1998) 1167–1231.[20] L. Wang, W. Chen, A.T. Shen Wee, Surf. Sci. Rep. 63 (2008) 465–486.[21] R. Friedlein, S. Braun, M.P. de Jong, W. Osikowicz, M. Fahlman, W.R.

Salaneck, J. Electron. Spectrosc. Relat. Phenom. 183 (2011) 101–106.[22] T. Hannapel, B. Burfeindt, W. Storck, F.J. Willig, Phys. Chem. 101

(1997) 6799–6802.[23] X-Y. Zhu, Surf. Sci. Rep. 56 (2004) 1–83.[24] X.-Y. Zhu, J. Phys. Chem. B 108 (2004) 8778–8793.[25] D. Polli, L. Luer, G. Gerullo, Rev. Sci. Instrum. 78 (2007) 103108–

103109.[26] P.A. Bruhwiler, O. Karis, N. Martensso, Rev. Mod. Phys. 74 (2002)

703–740.[27] D. Menzel, Chem. Soc. Rev. 37 (2008) 2212–2223.[28] L. Micaroni, F.C. Nart, I.A. Hümmelgen, J Solid State Electr. 7 (2002)

55–59.[29] M.L.M. Rocco, D.E. Weibel, L.S. Roman, L. Micaroni, Surf. Sci. 560

(2004) 45–52.[30] G. Araújo, C. Arantes, L.S. Roman, A.J.G. Zarbin, M.L.M. Rocco, Surf.

Sci. 603 (2009) 647–652.[31] J.R. Santa Rita, C. Arantes, G. Araújo, L.S. Roman, L. Micaroni, M.L.M.

Rocco, J. Elect, Spectrosc. Relat. Phenom. 184 (2011) 265–269.[32] U. Aygül, D. Batchelor, U. Dettinger, S. Yilmaz, S. Allard, U. Scherf, H.

Peisert, T. Chassé, J. Phys. Chem. C 116 (2012) 4870–4874.[33] J.F. Chang, B.Q. Sun, D.W. Breiby, M.M. Nielsen, T.I. Solling, M. Giles, I.

McCulloch, H. Sirringhau, Chem. Mater. 16 (2004) 4772–4776.[34] M.C. Gurau, D.M. deLongchamp, B.M. Vogel, E.K. Lin, D.A. Fischer, L.J.

Richter, Langmuir 23 (2007) 834–842.[35] Y. Takata, Y. Kitajima, H. Aga, S. Yagi, T. Asahi, T. Yokoyama, K.

Tanaka, T. Ohta, Jpn. J. Appl. Phys. 32–2 (1993) 350–352.[36] D.A. Outka, J. Stöhr, J.P. Rabe, J.D. Swale, J. Chem. Phys. 88 (1988)

4076–4087.[37] J. Stöhr, NEXAFS spectroscopy, Springer-Verlag, New York, 1992.[38] C. Arantes, B. Beck, B.G.A.L. Borges, G. Araújo, L.S. Roman, M.L.M.

Rocco, J. Phys. Chem. C 117 (2013) 8208–8213.[39] S. Lizzit, G. Zampieri, K.L. Kostov, G. Tyuliev, R. Larciprete, L. Petaccia,

B. Naydenov, D. Menzel, New. J. Phys. 11 (2009) 053001–053017.[40] J.L. Campbell, T. Papp, At. Data. Nucl. Data. Tables 77 (2001) 1–56.[41] H. Ikeura-Sekiguchi, T. Sekiguchi, Phys. Rev. Lett. 99 (2007) 228102–

228104.[42] H. Ikeura-Sekiguchi, T. Sekiguchi, Surf. Interface Anal. 42 (2010)

1085–1088.