Compositionally Dependent Nonlinear Optical Bandgap Behavior of

Dec 19, 2016 - Optical bandgap mapping of Nb–Ti mixed oxides anodically grown on a ... amorphization and an in-depth compositional gradient of the m...
0 downloads 0 Views 2MB Size
Subscriber access provided by UB + Fachbibliothek Chemie | (FU-Bibliothekssystem)

Article

Compositionally dependent non-linear optical bandgap behavior of mixed anodic oxides in niobium-titanium system Petra Bleckenwegner, Cezarina Cela Mardare, Christoph Cobet, Jan Philipp Kollender, Achim Walter Hassel, and Andrei Ionut Mardare ACS Comb. Sci., Just Accepted Manuscript • DOI: 10.1021/acscombsci.6b00162 • Publication Date (Web): 19 Dec 2016 Downloaded from http://pubs.acs.org on December 28, 2016

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

ACS Combinatorial Science is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Combinatorial Science

Compositionally dependent non-linear optical bandgap behavior of mixed anodic oxides in niobium-titanium system Petra Bleckenwegnera, Cezarina Cela Mardareb, Christoph Cobetc, Jan Philipp Kollendera, Achim Walter Hassela,b, Andrei Ionut Mardarea,b,* *

[email protected]

a

Institute for Chemical Technology of Inorganic Materials, Johannes Kepler University Linz,

Altenberger Str. 69, 4040 Linz, Austria b

Christian Doppler Laboratory for Combinatorial Oxide Chemistry, at the Institute for Chemical

Technology of Inorganic Materials, Johannes Kepler University Linz, Altenberger Str. 69, 4040 Linz, Austria c

Center of Surface and Nanoanalytics (ZONA), Johannes Kepler University Linz, Altenberger

Str. 69, 4040 Linz, Austria valve metal alloys, thin films, mixed anodic oxides, spectroscopic ellipsometry, bandgap

ACS Paragon Plus Environment

1

ACS Combinatorial Science

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 36

Abstract Optical bandgap mapping of Nb-Ti mixed oxides anodically grown on a thin film parent metallic combinatorial library was performed via variable angle spectroscopic ellipsometry (VASE). A wide Nb-Ti compositional spread ranging from Nb-90 at.% Ti to Nb-15 at.% Ti deposited by co-sputtering was used for this purpose. The Nb-Ti library was step-wise anodized at potentials up to 10 V SHE and the anodic oxides optical properties were mapped along the Nb-Ti library with 2 at.% resolution. The surface dissimilarities along the Nb-Ti compositional gradient were minimized by tuning the deposition parameters, thus allowing a description of the mixed Nb-Ti oxides based on a single Tauc-Lorentz oscillator for data fitting. Mapping of the Nb-Ti oxides optical bandgap along the entire compositional spread showed a clear deviation from the linear model based on mixing individual Nb and Ti electronegativities proportional to their atomic fractions. This is attributed to the strong amorphisation and an in-depth compositional gradient of the mixed oxides. A systematic optical bandgap decrease toward values as low as 2.0 eV was identified at approximately 50 at.% Nb. Mixing of Nb2O5 and TiO2 with both amorphous and crystalline phases is concluded, while the possibility of complex NbaTibOy oxide formation during anodization is unlikely.

ACS Paragon Plus Environment

2

Page 3 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Combinatorial Science

Introduction Alloys belonging to the Nb-Ti binary system are subject of intense research due to their special properties with direct applications in metallurgy and related areas. This is triggered by the significant role of Nb as stabiliser of the cubic beta-Ti phase.1 Different from the highly anisotropic hexagonal alpha-Ti phase, the beta-stabilised Ti alloys show an isotropic behaviour with higher elastic modulus.2 For intermediate amounts of Nb in this binary system, mixed alpha-beta alloys are normally obtained.3 Addition of Nb to Ti has a positive effect not only on the mechanical properties but also on the passivity of the forming oxides. Both metals are valve metals and it was demonstrated that all different crystallographic orientations of Nb-Ti show good passivation behaviour, as opposed to the case of Ti with its anisotropic passsivity behaviour of various crystallographic orientations.4 Apart from applications directly using metallic Nb-Ti alloys, also oxidised Nb and Ti are used in many different areas. Most recently the use of TiO2 have reached applications in the farand deep-ultraviolet spectral range with tremendous possibilities in various fields such as solar cells, photocatalysis, high-density integrated circuits, high-energy LEDs, energy storage and small optical elements.5 Many studies aim to improve the final properties of TiO2. Recently, Nb and N co-alloying with TiO2 has been shown to decrease the oxide bandgap for enhanced photocatalytic or photovoltaic functionality under visible light while oxidised Fe-Ti-Sr alloys showed enhanced activity for photoelectrochemical water splitting.6 Nb doping resulted in a remarkable increase of the carrier densities yielding higher electrical conductivities of TiO2 with direct applications in transparent conducting oxides (TCOs).7 Besides Nb, Ta substitution in TiO2 represents an important step in further development of TCOs with great potential for applications in photovoltaics, photocatalysis, and water splitting applications.8 Additionally,

ACS Paragon Plus Environment

3

ACS Combinatorial Science

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 36

combinatorial approaches have proven themselves extremely useful for electrical conductivity screening of Nb:TiO2 for identifying the ideal amount of Nb responsible for the highest conductivity.9 Mixtures of Nb and Ti oxides are good candidates as catalysts and Nb-Ti libraries were screened for development of new materials with improved electrochemical activity for the oxygen reduction reaction.10 Accordingly there is a high scientific interest in the Nb-Ti system and in the present study their mixed anodic oxides are investigated via spectroscopic ellipsometry as a function of the parent metal alloy compositions. The non-contact measurements provided by ellipsometry aim at completing the description of compositionally induced bandgap tuning of anodized Nb-Ti alloys usually performed by spectroscopic photoelectrochemistry when an electrolyte is in contact with the investigated surface.

ACS Paragon Plus Environment

4

Page 5 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Combinatorial Science

Experimental Deposition of the Nb-Ti thin film combinatorial library Borosilicate glass slides (26 × 76 mm2) were used as substrates for deposition of the Nb-Ti combinatorial library. Before utilization they were cleaned consecutively with acetone, isopropanol and deionised water in an ultrasonic bath and subsequently dried with high purity nitrogen. The compositional spread of the Nb-Ti parent metal alloys was deposited using a cosputtering system (Mantis Deposition Ltd) with two opposite DC powered sputtering guns (Ø 50 mm). High purity Nb (99.95%, MaTeck Material-Technologie & Kristalle GmbH) and Ti (99.99%, ITL-Vacuum-Components) targets were used simultaneously. The system (with a base pressure of 10-10 hPa) was operated at room temperature in 5×10-3 hPa Ar and with a deposition distance of 130 mm (no rotation of the substrates). In order to attain a wide range Nb-Ti library, three individual substrate batches with partial overlaps in the compositional gradients were prepared. Desired compositional spreads along each substrate were obtained by varying the power of each sputtering gun. The necessary power values were calculated using a selfdeveloped software based on a mixed matter theory model. For serving as reference and deposition rate calibration (via contact-profilometric thickness measurements) pure Nb and Ti samples were prepared using the same conditions. Deposition rates of 2.268 nm W-1 h-1 for Nb and 1.440 nm W-1 h-1 for Ti were used and all films were approximately 300 nm thick (as measured in the middle of the samples). For the three individual Nb-Ti samples defining the combinatorial library the DC power values for Nb and Ti sputtering guns were 80, 54, 23 W and 40, 82, 103 W, respectively.

ACS Paragon Plus Environment

5

ACS Combinatorial Science

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 36

Anodic oxide formation on Nb-Ti library Anodic oxides were grown along the Nb-Ti library using an electrochemical setup allowing a sequential dipping of the sample into the electrolyte in controllable steps. In this way different thicknesses of the mixed oxides could be obtained for further ellipsometric investigations. The aim was to concurrently grow anodic oxide along the entire Nb-Ti library at once. For this purpose all single metal and/or alloy thin films were dipped into the electrolyte in perpendicular direction to the compositional gradient prior to applying the anodizing potential. At each dip step an anodic oxide stripe approximately 4 mm wide was formed along the entire Nb-Ti compositional spread. The thickest oxide (10 V vs. SHE) was grown at the first dip step and the anodizing potential was decreased in 1 V steps for subsequent dipping until a final anodisation potential of 1 V SHE was used. Due to substrate size limitations, two identical Nb-Ti samples were used for each compositional batch, each one comprising 5 consecutive oxide thicknesses obtained from the corresponding anodising procedure (1-5 and 6-10 V SHE). The anodisations were carried out potentiodynamically with a potential increase rate of 100 mV s-1 and a subsequent potentiostatic polarization for 200 s at the maximum potential. This approach ensures a complete oxide growth for obtaining a high quality, compact and reproducible anodic oxide.11 For all anodisations, a 1287 Solartron electrochemical interface in a three electrode configuration was used with a Hg/Hg2(CH3COO)2 µ-reference electrode and a Pt mesh as counter electrode.12,13 As electrolyte, a pH 6.0 CH3COOH/CH3COONa buffer solution was used. More details of the setup including detailed photos are presented elsewhere.14

ACS Paragon Plus Environment

6

Page 7 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Combinatorial Science

Metallic and oxide thin films characterization Before anodisation of the thin film alloys their microstructure was examined by a field emission Zeiss Gemini 1540 XB SEM with an acceleration voltage of 20 kV and in-lens detection. The compositional spread of each Nb-Ti sample was determined using an EDX analyser integrated in the SEM system (INCA X-sight, Oxford Instruments). The EDX detector was calibrated before each series of measurements with a high purity Co standard (Micro-Analysis Consultants, United Kingdom) and INCA-software was used to process the obtained data. Additionally, the Nb-Ti library was analyzed crystallographically by grazing incidence X-ray diffraction (GIXRD) using a Cu-Kα source (PANalytical X’pert Pro). Optimized incident angles from 1.0 to 1.3° were used to minimise substrate influence. All anodised pure and alloy thin films were analysed using variable angle spectroscopic ellipsometry (VASE). These measurements were performed with a J. A. Woollam M-2000 DI ellipsometer equipped with a quartz tungsten halogen lamp, a deuterium lamp, a rotating compensator and two spectrographs to allow concurrent measurements of 700 wavelengths at photon energies ranging from 0.75 to 6.5 eV. The anodized Nb-Ti compositional spread was investigated with 2 at.% resolution for every oxide thickness (1 to 10 V SHE). Ellipsometric measurements were performed at 5 different angles of incidence (55-75°), for each composition and each oxide thickness. The applied layer model for the ellipsometric data analysis consisted of two layers describing a metal substrate with an oxide layer on top. The multi-layer optical response was simulated with the Berreman 4 × 4 matrix formalism.15 The layer dielectric properties were determined through a fit of the measured Stokes parameter N, C, and S within the Levenberg-Marquardt method by minimizing the mean square error (MSE) function.

16

Therefore, the CompleteEASE software of the J. A. Woollam company was used. The

ACS Paragon Plus Environment

7

ACS Combinatorial Science

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 36

borosilicate glass substrate was neglected in the model based on the assumption that the parent metal thickness of 300 nm is large enough for the metal to be non-transparent. In the model applied for analyzing anodic oxides on pure metals, the metal substrate dielectric function was approximated by a Kramers-Kronig consistent general B-spline layer while a single TaucLorentz (TL) oscillator was used for the oxide film dielectric functions. The dielectric function of the substrate was independently determined on almost pristine Nb and Ti films before anodization. This approach has been successfully used in previous studies for modelling amorphous dielectric and semiconducting materials.

17,18

The valve metal characteristic of the

Nb-Ti alloys and their fast passivation behaviour in air makes the pure metal surface inaccessible for an ellipsometric analysis. For this reason a Bruggeman effective medium approximation (EMA) was used to determine the substrate dielectric function, comprising a mixing of the Bspline fit data of the pure metals according to the Nb-Ti composition (as measured by EDX). For the mixed Nb-Ti oxide layer a single TL-oscillator was employed again. In this approach the imaginary part of the dielectric function depends on the photon energy E as:

ε 2 (E) =

AE0 B ( E − E g ) 2 2

E (( E 2 − E0 ) 2 + ( BE ) 2 ) , E ≥ E

(1)

g,

ε 2 (E) = 0 , E < E . g The amplitude of the function is represented by the factor A, the peak transition energy by E0, the peak broadening by B and the band gap energy by Eg. The real part ε1 is calculated by the Kramers-Kronig relation:

ε1 ( E ) = ε ∞ +

2

π

P∫



Eg

E 'ε 2 ( E ' ) dE ' E '2 − E 2

(2)

ACS Paragon Plus Environment

8

Page 9 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Combinatorial Science

Here, ε∞ accounts for the contribution from high frequencies and P denotes the Cauchy principal value of the integral.19 The high quality assessment of the model fits as compared with the experimental data is performed using the mean square error function.

ACS Paragon Plus Environment

9

ACS Combinatorial Science

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 36

Results and discussion For defining the spatial distribution of various Nb-Ti alloys along the compositional spread, the thin film library was mapped by EDX immediately after the deposition. In this way, each thin film alloy along the entire Nb-Ti library under study can be immediately accessed by a simple length measurement from one reference edge (e.g. left edge) of the library containing all three samples. The obtained results are presented in Fig. 1 together with a schematic representation of the three individual Nb-Ti samples. Individual sputtering guns are colour coded for defining the colour gradient along each sample symbolizing the compositional gradient. Vertical dashed lines are added for separating the EDX data belonging to each individual sample. The EDX analysis shows almost linear composition variations along each sample which are due to the chosen sputtering geometry. The link between the three Nb-Ti samples defining the overall library was carefully done in order to exclude any edge effects. For this purpose, the compositional spread of each sample was tuned as partly overlapping its neighbour. This is observable in Fig. 1 in the vicinity of the vertical dashed lines. These regions will be further used for checking the continuity of various parameters or properties. A total compositional spread ranging from Nb90 at.% Ti to Nb-15 at.% Ti was obtained. This wide compositional range defines a compositional resolution of 0.3 at.% mm-1 which is very convenient for allowing further localized investigations with very good compositional resolution. At first the surface microstructure was analyzed at various Nb/Ti ratios using SEM. This was performed before the anodic oxide growth in order to observe compositionally induced changes which may further impact the oxide formation. In a previous study, a similar Nb-Ti library was analyzed and significant surface microstructure changes were identified.20 In the present work, the proposed ellipsometric study for mapping the optical bandgaps along the Nb-Ti

ACS Paragon Plus Environment

10

Page 11 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Combinatorial Science

compositional spread would ideally require parent metal surfaces as smooth as possible for avoiding unnecessary complications during the optical layer modelling. The complexity of the model would generally increased when describing rough surfaces mainly due to possible depolarization effects. The anodic oxide growth is performed at potentials up to 10 V SHE and due to oxide formation factors of Nb and Ti below 3 nm V-1, an overall surface smoothening effect triggered by the presence of the oxide is not expected here.21 Moreover, for minimizing the roughness variation from one Nb-Ti alloy to the next along the analyzed library, the individual deposition rates of Nb and Ti were significantly reduced here (as compared to the previous study) while both metallic sources were now placed at 180 ° for a head-on atomic collision The results of the SEM imaging along the Nb-Ti thin film combinatorial library before anodisation are presented in Fig. 2. A table of SEM images is presented containing three rows corresponding to each of the Nb-Ti sample defining the entire library. The amount of Nb in at.% is given in each SEM image and the surfaces of pure Ti and Nb films (deposited in identical conditions) are additionally presented as references and are indexed as 0 and 100 at.% Nb, respectively. The compositionally overlapping regions at the boundaries between the Nb-Ti samples (30 and 60 at.% Nb) are also imaged in Fig. 2 for properties continuity confirmation. The surface of pure Ti thin film was drastically changed here (as compared to the previously mentioned study) by the decrease of the deposition rate. Unlike previously reported, the pure Ti grains are small (not exceeding 75 nm) forming a compact surface, while decreasing the deposition rate did not significantly affect the pure Nb microstructure with grains in the range of 30 nm.20 Starting from the features observed on pure Ti, small amounts of Nb triggered slight modifications on the surface. Slightly larger and seemingly flatter surface grains can be observed at 10 at.% Nb. Increasing the Nb content along the thin film library resulted in a sudden change

ACS Paragon Plus Environment

11

ACS Combinatorial Science

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 36

of the microstructure at 20 at.% Nb where much smaller grains can be observed on a more compact surface. From this point on, increasing further the Nb content did not produce any significant change along the library. The surface microstructure resembles the pure Nb, the only observable difference being slightly larger grains for the Nb-Ti alloys. The imaged regions of composition overlapping between the samples at 30 and 60 at.% Nb show identical surface features indicating that changing the power ratio between the Nb and Ti guns from one sample to the next did produce the expected effect of surface microstructure conservation in the present situation. The crystallographic properties of the Nb-Ti thin film combinatorial library were characterized by GIXRD and the results of the compositional mapping are presented in Fig. 3. All diffractograms are presented in log scale for better visualizing the presence of crystalline peaks. Similar to the previous case of SEM mapping, the diffractograms corresponding to pure Nb and pure Ti are also shown (in blue and red, respectively) as references. The reference Ti film has a hexagonal symmetry with several peaks being identifiable in the diffractogram while the pure Nb film shows a cubic symmetry with three distinct peaks being indexed. Even though some of the cubic and hexagonal peaks have almost the same 2θ positions (e.g. (211) with (103) or (110) with (002), respectively) plotting their compositionally induced evolution indicates changes from one symmetry to the other. This is the case of the present Nb-Ti library. A low amount of Nb in the alloy preserves the hexagonal symmetry of pure Ti as observable for Nb-90 at.% Ti. The hexagonal (002) peak appears much increased in intensity while (101) remains only as a slight shoulder. Due to the peak overlapping position with the cubic (110), it may be suggested that the alloy with 10 at.% Nb actually shows a mixture of cubic and hexagonal phases. This behaviour is not singular, such transitory alloys being encountered before in binary mixtures of valve

ACS Paragon Plus Environment

12

Page 13 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Combinatorial Science

metals.22 Above 20 at.% Nb the presence of a cubic phase is evidenced by the disappearance of hexagonal (102) accompanied by the appearance (very weak at first) of cubic (200). However, the cubic (200) peak was weakly identified also in the pure Nb film, likely due to the modified deposition conditions as previously discussed. The XRD results are matching the SEM mapping shown in Fig. 2. The compositional threshold of approximately 20 at.% Nb, where the microstructure completely changes towards small grains and compact surface, represents the transition from hexagonal to cubic phase. These results confirm previous findings while the decrease of the deposition rates allows observation of the transition between the two symmetries in the present study.20 All Nb-Ti alloys above 20 at.% Nb have therefore a cubic symmetry sharing the same microstructure and their anodic oxides may be directly comparable to each other. A systematic properties mapping based on the same ellipsometric model is therefore expected to provide at least a correct qualitative evaluation of the trend/behaviour of the mixed anodic oxides in the Nb-Ti system. The anodic oxide growth was performed simultaneously along the entire Nb-Ti thin film compositional spread by applying anodic potentials (potentiodynamically and potentiostatically up to 10 V SHE) during stepwise dipping the samples in electrolyte. A typical series of cyclic voltammograms (CV) as recorded during anodisation of Nb-Ti thin film alloys is exemplified in Fig. 4. The presence of both, the current overshoot and current plateau are typical to valve metal anodisation process. According to the well established and experimentally confirmed high field regime (describing the anodic oxide formation on valve metals) the initial current peak is due to a delay of the oxide formation when charge is accumulated at both metal/oxide and oxide/electrolyte interfaces.23 However, on the external circuit (monitored by the potentiostat) charge is still transferred, therefore the current increases while no actual oxide growth occurs.

ACS Paragon Plus Environment

13

ACS Combinatorial Science

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 36

Only when both positive (due to cations hopping) and negative (due to anions hopping) space charge layers are fully overlapped (defining the maximum current overshoot during the carriers injection into the already formed oxide) new oxide may grow. After this point the current settles down in a plateau describing the continuous oxide growth concomitant at both interfaces.23 The change of the CVs while modifying the final anodisation potential is represented in Fig. 4 by an arrow indicating the direction of the potential increase. Attempting to observe an anodisation trend in Fig. 4, other than the presence of the current overshoot and/or the oxidation plateau, is not very useful due to the vertical absolute scale used for the current. A current density value is normally used in such plots and then various parameters (e.g. anodic oxide formation factor) become directly accessible. Unfortunately, the nature of the present experimental approach does not allow this. Many different Nb-Ti alloys are simultaneously exposed to the electrolyte in an open system which may lead to uncontrolled effects (e.g. electrocapillarity) during the anodisation process. Therefore, the real exposed area of the working electrode at each dip step cannot be precisely known. The purpose of the present study is not focused on interpretation of electrochemical data (which was already reported 20) but on growing high quality anodic oxides under precise potential control (precise thickness definition) for further ellipsometric investigations. However, an increase of the anodisation plateau length with the potential is clearly observable in the CVs from Fig. 4 as directly related to thicker oxides growth. The final potentiostatic anodisation step is applied at the end of each CV in order to avoid ionic species entrapment into the newly formed oxides. The inlet of Fig. 4 exemplifies such potentiostatic final treatment (e.g. 1 V SHE applied potential) for 200 s. The removal of the mobile charges and subsequent potentiostatic oxide growth is directly described by the time exponential decay of the current (represented by the red fitting curve).

ACS Paragon Plus Environment

14

Page 15 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Combinatorial Science

The entire anodized Nb-Ti thin film library was investigated (with a compositional resolution of 2 at.%) using VASE for compositional mapping of the oxide optical properties. A simultaneous evaluation of the optical data sequentially obtained for various oxide thicknesses up to 10 V SHE was performed. As a by-product of the data fitting procedure, the oxide thickness values for various Nb-Ti parent metal compositions were obtained. Additionally, the imaginary part of the dielectric function (ε2) is approximated with a single TL oscillator dispersion during the VASE data fit. This describes electronic excitations in the Nb-Ti mixed oxides triggering absorption processes. Selected results from the total compositional mapping of the TL absorption are summarized in Fig. 5 as obtained after evaluation of the entire optical dataset for all investigated thicknesses. Even though more data are available, the values of ε2 are plotted as a function of the energy for the same parent metal compositions as previously discussed in Fig. 2 and Fig. 3. Additionally, the values of the absorption edge (as provided by the VASE analysis in terms of Eg) are presented in the figure as individual points. All analyzed anodic oxides are transparent toward IR below 2.0 eV (above 620 nm) and fully absorb UV radiation above 3.5 eV (below 354 nm). In the energy range 2.0-3.5 eV all Nb-Ti oxides showed varying optical properties depending on composition. The anodic oxide grown on pure Ti thin film has an absorption band edge located at 3.2 eV and small additions of Nb into the parent metal alloys trigger a decrease of this value by almost 1 eV. The previously discussed cubic phase stabilization of the parent metal alloys above 20 at.% Nb increases the absorption edge value above 3.0 eV. This may be linked to the parent metals transition from hexagonal to cubic through a mixed hexagonal-cubic phase (identified by XRD) which may directly influence the anodic oxide properties. Further increasing the Nb content of the anodized alloys resulted in a gradual decrease of the absorption onsets down to energies around 2 eV as observed for the compositional region between 40 and

ACS Paragon Plus Environment

15

ACS Combinatorial Science

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 36

60 at.% Nb. This decrease is systematic and is followed by a similar increase for even higher Nb amounts until a value above 3 eV is finally reached for anodic oxides on pure Nb. Since electric field crystallization of both Nb and Ti anodic oxides is possible24, the observed behaviour of the absorption onset may be linked to it. Even if generally the anodic oxides are expected to be quasi-amorphous their behaviour must be affected by the properties of their parent metals which directly influence the anodisation process.25,26 Along the entire anodized Nb-Ti library the absorption maximum of the TL resonance slightly shifts to higher photon energies with increasing Nb concentrations while the peak broadening increases. As previously mentioned, the values of the anodic oxide thickness are calculated during the VASE analysis and can be directly used for obtaining the oxide formation factors. These factors (usually k, expressed in nm V-1) are material constants depending on the oxide and/or parent metal composition being particularly relevant for characterization of mixed anodic oxides. In Fig. 6 the oxide formation factors as obtained from the optical data are mapped for the entire NbTi compositional spread under study. In order to check for possible discontinuities and/or edge effects present at the borders of each Nb-Ti sample forming the entire library, the measurement values obtained from each sample are colour coded and data from overlapping regions are clearly visible. No indication of strong discontinuities or edge effects is notable. The compositional mapping of the oxide formation factors is in agreement with previously reported values directly obtained from electrochemical investigations.20 Due to missing experimental data for low Nb contents up to 20 at.%, a gradual decrease of the oxide formation factors starting from the value of pure Ti oxide (2.5 nm V-1) was only hinted in the electrochemical study. This behaviour is demonstrated now as being correct using a completely different investigation technique. The increase of the Nb amount along the library resulted in a further small decrease of the oxide

ACS Paragon Plus Environment

16

Page 17 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Combinatorial Science

formation factors toward 1.5 nm V-1 measured at approximately 40 at.% Nb. Above this composition the k factor starts to continuously increase with the Nb amount up to the value of 2.6 nm V-1 corresponding to pure Nb (similar to electrochemical data). However, small discrepancies are observable (as expected) in the mapping of k factors when using both techniques due to their different nature. The fact that the oxide formation factor measured during growth of a mixed Nb-Ti oxide is smaller as compared to the factor corresponding to each of the pure oxides suggests a hindrance of the ion transport mechanism. At Nb/Ti ratios close to 50% (in the parent metal alloys) the electric field induced ion hopping is disturbed likely due to the competition of both Nb5+ and Ti4+ for O2- ions. This idea would explain why increasing the concentration of one species (Nb or Ti) results in approaching the known oxide formation factor of that pure element anodized. The main result obtained during VASE analysis is the optical bandgap i.e. the absorption onset of the investigated oxide. During the current study, the bandgap compositional mapping of the anodized Nb-Ti library was obtained directly from the fitting procedure of the specialized VASE software. However, the oxide film thickness has been proven to influence the quality of the fit results. When datasets corresponding to very thin anodic oxides (up to 4 V SHE) were used, the bandgap values could not be calculated for certain compositions with a reasonable MSE and parameter uniqueness which otherwise resulted in a fit error of less than 1%. This may be due to deficiencies in the model chosen for describing these thin oxides. It is likely that oxides thinner than 10 nm are not properly modelled by a smooth homogeneous layer, as opposite to thicker oxides. For this reason the optical bandgap compositional mapping of the Nb-Ti mixed anodic oxides was performed as a function of oxide thickness and the obtained results are summarized in Fig. 7. This finding reconfirms previously reported hints that mixed anodic oxides on valve

ACS Paragon Plus Environment

17

ACS Combinatorial Science

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 36

metals may change their properties with thickness which inadvertently may lead to bandgap modifications.

14

In Fig. 7 a second vertical axis was added on the right side for providing the

absorbed radiation wavelength equivalent of the bandgap energy. Two different curves are plotted as corresponding to VASE fitting performed on oxides grown between 5-7 V and 8-10 V SHE (as directly indicated in the figure). Each bandgap mapping dataset was additionally fitted using Gaussian multi-peak functions (full lines) in order to provide a mean of following the bandgap changing trend as a function of composition. Similar trends are observable for both selected anodic oxide thickness ranges. Similar to the previous discussion, the edge effects and measurement continuity across the three Nb-Ti samples can be evaluated by observing the overlapping data. Values obtained from each Nb-Ti sample are shape-coded in Fig. 7. A good continuity along the entire Nb-Ti library can be concluded. Since the anodic oxides on pure Nb and Ti thin films were also investigated, their values give a profound basis of the borderline cases and thus can be used as references. These measured values are in the expected literature ranges (3.0-3.2 eV for TiO2 and 3.2-3.4 eV for Nb2O5).23,6,27 While increasing Nb concentrations the optical bandgap values go first through a minimum approximately at 10 at.% Nb where a mixure of hexagonal and cubic phases were identified in the parent metal alloys (see Fig. 3). The measured optical bandgaps are raising back to their original values around the compositional threshold corresponding to the complete cubic stabilization of the parent metal alloys (Nb80 at.% Ti). Increasing even more the Nb concentration leads to a second decay of the bandgap values. A minimum is attained approximately at 50 at.% Nb (or Ti) followed by increased values toward pure Nb oxide at the end of the compositional spread. Values as low as 2.2 eV were obtained for the thickest analyzed anodic oxides (8-10 V SHE) grown on Nb-50 at.% Ti. This represents a bandgap decrease by almost 1 eV as compared to any of the value corresponding to

ACS Paragon Plus Environment

18

Page 19 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Combinatorial Science

the pure oxides. The precise reasons for this decrease cannot be clearly concluded from the present data. However, it may be argued here that a clear compositionally induced transformation occurs. This is evident from the gradual change of the bandgap along the Nb-Ti library. Even assuming a limit case scenario when the used model for the ellipsometry data fitting is completely inappropriate, still the obtained results would be a good indication of the optical bandgap variation trend along the library. This conclusion is fully supported by the microstructure and crystallography tuning purposefully performed in the present study in order to expose the same surface features to the same fitting model independent on composition. Only in this way the bandgaps of neighbouring oxides can be directly comparable. The measured optical bandgap decrease in the anodized Nb-Ti system at 50 at.% Nb may be triggered by changes in the mixed Nb-Ti oxides crystallinity. Both Nb and Ti oxides can crystallize in the electric field applied during their formation. Previous studies already showed that small amounts of Ti alloyed with Nb can lead to anodic oxide field crystallization at low anodisation potentials (below 20 V).24 Whether the mixed anodic oxide remains amorphous or not (as a function of composition) may dictate the optical bandgap behaviour. In a completely disordered oxide the optical bandgap is directly represented by the mobility gap which means that smaller values will be optically measured as compared to a crystalline oxide. The difference between the actual bandgap and the optically measured one can be as large as 0.5 eV for valve metals.28 Such strong disturbance may be responsible for the sudden drop of the Nb-Ti anodic oxide bandgaps at 10 at.% Nb which one would expect during the transition between two different symmetries of the parent metal alloys. Additionally, it was recently shown that metallic Nb plays a crucial role in decreasing the TiO2 bandgap down to 2.0 eV when co-alloyed with N.6 The possibility of trapping metallic Nb inside the anodic oxide together with electrolyte species

ACS Paragon Plus Environment

19

ACS Combinatorial Science

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 36

should not be excluded in the present case when both Nb and Ti ions compete for O during the high field anodic oxide formation process. Even though metallic inclusions may form, the subsequent potentiostatic anodisation step used in this work was precisely aiming at removal of ionic leftovers from the oxide thus rendering such possibility as unlikely. In a previous study, it was indicated that potentiostatic polarizations following potentiostatic anodizations decreased the charge carrier concentration in the oxide. This was evidenced through a decrease of dielectric relaxation currents and resonance tunnelling currents during subsequent anodic pulses.29 The presently observed non-linear behaviour of the optical bandgap of mixed Nb-Ti anodic oxides represents a deviation from the model previously implemented for mixing of d-group metal anodic oxides. A semiempirical correlation between the optical bandgaps and electronegativity of the alloying elements was previously derived by using a single electronegativity value for the metal ions ( χ M ) calculated as a linear superposition of the electronegativities of constituent elements (A and B) weighted by their atomic fractions (xa and xb):30

χ M = xa χ A + xb χ B

(3)

The optical bandgap for amorphous d-metal oxides is further calculated as: 31,32 E gopt − Eam = 1.35(χ M − χ O ) − 1.49 2

where

E gopt

(4)

is the optical bandgap of the oxide, Eam describes the lattice disorder affecting both,

density of states near the valence and conduction band edges and χM , χ,O represent the metal and oxygen electronegativities.33 A value of 0.15 eV is typically taken for Eam as suggested in studies of various oxides on Nb and Ti alloys.27,28 In order to better observe the non-linearity of the

ACS Paragon Plus Environment

20

Page 21 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Combinatorial Science

bandgap behaviour in the present study, theoretically calculated values of

E gopt

(using Eq. 3 and

4) are plotted in Fig. 8 together with the experimental values obtained here from ellipsometric investigations. The compositional dependence of the theoretical bandgap value (

E gtheor

) is linear

due to Eq. 3. However, the experimental values show clearly a non-linear behaviour (

E gexp

refer

to the data fitted using the middle range anodic oxide thicknesses of 5-7 V). Additionally, in Fig. 8 the compositional deviation of mixed Nb-Ti anodic oxides from their parent alloy composition is presented as previously reported from surface XPS measurements.20 If the low Nb concentration bandgap deviation of the mixed Nb-Ti oxide from the theoretical values may be justified by the previously discussed hexagonal to cubic transition of the Nb-Ti alloys, the experimentally found bandgap lowering by 1 eV around 50 at.% Nb is intriguing. Coincidently, this gradual decrease of the optical bandgap occurs in a compositional region where the anodic oxides are strongly surface enriched by TiO2 (up to 60% as compared to the metal alloy composition) as indicated by previous studies.20 This strong deviation of the mixed oxide composition from the parent metals can indicate a strong disorder occurring during the anodisation process which would confirm the observed bandgap decrease. Moreover, a recent similar study on anodisation of Nb-Ta thin film alloys show very similar findings and point out the possibility of an in-depth compositional change of the anodic oxides.14 Following this idea, the driving force for strong oxide disordering will be enhanced if stronger in-depth compositional gradients are present, as it can be assumed for the Nb-Ti system due to the discussed surface enrichment of TiO2 and its complementary depletion of Nb2O5. A similar idea justifies very recent findings regarding inhomogeneous compositions identified at nm scale in anodized valve metal alloys.34 In conclusion, the reason for the present deviation (see Fig. 8) of experimental findings from the theoretical bandgap values may be the assumption of linear

ACS Paragon Plus Environment

21

ACS Combinatorial Science

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 36

electronegativities mixing provided by Eq. 3. Detailed studies applying this approach demonstrate its validity for many binary oxides described as AaBbOy,27,28,30 which is not the case in the present study. This conclusion is also supported by a direct observation at atomic scale of various Nb-Ti oxides showing amorphous and partly crystalline Nb2O5 and TiO2 together forming the mixed anodic oxide.24 For such cases Eq. 3 needs to be empirically re-formulated.

ACS Paragon Plus Environment

22

Page 23 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Combinatorial Science

Conclusions The present study focuses on the optical bandgap measurement of Nb-Ti mixed oxides anodically grown on a thin film parent metal combinatorial library. A wide Nb-Ti compositional spread ranging from Nb-90 at.% Ti to Nb-15 at.% Ti was deposited on glass substrates using a co-sputtering system. The obtained Nb-Ti library was anodized at potentials up to 10 V SHE. Variable angle spectroscopic ellipsometry (VASE) was used for optical properties mapping of the anodic oxides with 2 at.% resolution. The microstructure of the parent metal library was tuned for minimizing the surface dissimilarities along the compositional change for a better interpretation of ellipsometric results. This allowed a description of the mixed Nb-Ti oxides based on a single Tauc-Lorentz oscillator for data fitting. A compositional mapping of the anodic oxide formation factors was performed using the oxide thickness ellipsometrically measured. The results are in good agreement with electrochemical data. The optical bandgap values of NbTi oxides along the entire compositional spread show a clear deviation from the linear model based on mixing individual Nb and Ti electronegativities proportional to their atomic fractions developed using spectroscopic photoelectrochemistry. This behaviour is explained via an indepth compositional gradient of the oxides related to strong amorphisation within the mixed oxide. At approximately 50 at.% Nb this resulted in a systematic optical bandgap (i.e. the absorption onset) decrease toward a value of 2.0 eV. The mixing of Nb and Ti oxides is concluded as being formed from Nb2O5 and TiO2 with both amorphous and crystalline phases, while the possibility of complex NbaTibOy oxide formation is unlikely. The compositionally dependent non-linear character of the ellipsometrically measured optical bandgap supports this conclusion. The complete understanding of all reasons for which ellipsometry (as a "dry" noncontact technique) and photoelectrochemistry (as a "wet" contact technique) are providing

ACS Paragon Plus Environment

23

ACS Combinatorial Science

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 36

discrepancies regarding the measured bandgap values likely requires additional studies of other valve metal alloys combinations.

ACKNOWLEDGMENTS The financial support by the Austrian Federal Ministry of Economy, Family and Youth and the National Foundation for Research, Technology and Development for the Christian Doppler Laboratory for Combinatorial Oxide Chemistry (COMBOX) is gratefully acknowledged.

ACS Paragon Plus Environment

24

Page 25 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Combinatorial Science

REFERENCES

(1) Raabe, D.; Sander, B.; Friak, M.; Ma, D.; Neugebauer, J. Theory-guided bottom-up design of beta-titanium alloys as biomaterials based on first principles calculations: Theory and experiments Acta Materialia 2007, 55, 4475-4487. (2) Kim, H. Y.; Ikehara, Y.; Kim, J. I.; Hosoda, H.; Miyazaki, S. Martensitic transformation, shape memory effect and superelasticity of Ti–Nb binary alloys Acta Materialia 2006, 54, 2419– 2429. (3) Sander, B.; Raabe, D. Texture inhomogeneity in a Ti-Nb-based beta-titanium alloy after warm rolling and recrystallization Mater. Sci. Eng. A 2008, 479, 236-247. (4) Woldemedhin, M. T.; Raabe, D.; Hassel, A. W. Scanning droplet cell microscopy on single grains of Ti-Nb alloys, submitted. (5) (a) Tanabe, I.; Ozaki, Y. Far- and deep-ultraviolet spectroscopic investigations for titanium dioxide: electronic absorption, Rayleigh scattering, and Raman spectroscopy J. Mater. Chem. C 2016, 4, 7706-7717. (b) Kong, L.; Wang, C.; Zheng, H.; Zhang, X.; Liu, Y. Defect-Induced Yellow Color in Nb-Doped TiO2 and Its Impact on Visible-Light Photocatalysis J. Phys. Chem. C, 2015, 119, 16623–16632. (c) Fehse, M.; Cavaliere, S.; Lippens, P. E.; Savych, I.; Iadecola, A.; Monconduit, L.; Jones, D. J.; Rozière, J.; Fischer, F.; Tessier, C.; Stievano, L. Nb-Doped TiO2 Nanofibers for Lithium Ion Batteries J. Phys. Chem. C, 2013, 117, 13827–13835. (6) (a) Breault, T. M.; Bartlett, B. M. Lowering the band gap of anatase-structured TiO2 by coalloying with Nb and N: Electronic structure and photocatalitic degradation of methylene blue dye J. Phys. Chem. C, 2012, 116, 5986-5994. (b) Kusama, H.; Wang, N.; Miseki, Y.; Sayama, K. Combinatorial Search for Iron/Titanium-Based Ternary Oxides with a Visible-Light Response J. Comb. Chem., 2010, 12, 356–362. (7) Han, X.; Song, K.; Lu, L.; Deng, Q.; Xia X.; Shao, G. Limitation and extrapolation correction of the GGA + U formalism: a case study of Nb-doped anatase TiO2 J. Mater. Chem. C, 2013, 1, 3736-3746. (8) (a) Nursam, N. M.; Wang, X.; Caruso, R. A. High-Throughput Synthesis and Screening of Titania-Based Photocatalysts ACS Comb. Sci., 2015, 17, 548–569. (b) Gopinadhan, K.; Kumar, B.; Palina, N.; Motapathula, M.; Pallecchi, I.; Sarkar, T. P.; Zhihua, Y.; Chen, J. Q.; Annadi, A.; Rana, A.; Srivastava, A.; Marre, D.; Chen, J.; Ariando, Dhar, S.; Rusydi, A.;. Venkatesan, T Effect of Nb and Ta substitution on donor electron transport and ultrafast carrier dynamics in anatase TiO2 thin films, J. Mater. Chem. C, 2015, 3, 6329-6333. (9) Kafizas, A.; Noor, N.; Carmalt, C. J.;. Parkin, I. P TiO2-based transparent conducting oxides; the search for optimum electrical conductivity using a combinatorial approach, J. Mater. Chem. C, 2013, 1, 6335-6346. (10) Schäfer, D.; Mardare, C. C.; Savan, A.; Sanchez, M. D.; Mei, B.; Xia, W.; Muhler, M.; Ludwig, A.; Schuhmann, W. High-Throughput Characterization of Pt Supported on Thin Film Oxide Material Libraries Applied in the Oxygen Reduction Reaction, Anal. Chem., 2011, 83, 1916–1923. (11) Hassel, A. W.; Diesing, D. Breakdown of ultrathin anodic valve metal oxide films in metalinsulator-metal-contacts compared with metal-insulator-electrolyte contacts, Thin Solid Films, 2002, 414, 296-303. (12) Hassel, A. W.; Fushimi, K.; Seo, M. An agar-based silver| silver chloride reference electrode for use in micro-electrochemistry, Electrochem. Comm., 1999, 1, 180-183.

ACS Paragon Plus Environment

25

ACS Combinatorial Science

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 26 of 36

(13) Lill, K. A.; Hassel, A. W. A combined µ-mercury-reference-gold-counter-electrode system for microelectrochemical applications J. Solid State Electrochem., 2006, 10, 941-946. (14) Limberger, W.; Mardare, C. C.; Cobet, C.; Zuo, J.; Hassel, A. W.; Mardare, A. I. Spectroscopic ellipsometry for compositionally induced bandgap tuning of combinatorial niobium–tantalum anodic oxides RSC Adv., 2016, 6, 79934-79942. (15) Yeh, P. Optical Waves in Layered Media; John Wiley & Sons Inc.: New York, 1988; pp. 239 (16) Tompkins, H.; Irene, E. A. Handbook of Ellipsometry; Springer: Heidelberg, 2005; pp. 64 (17) von Blanckenhagen, B.; Tonova, D.; Ullmann, J. Application of the Tauc-Lorentz formulation to the interband absorption of optical coating materials Appl. Opt., 2002, 41, 31373141. (18) Jellison, G. E.; Modine, F. A. Parameterization of the optical functions of amorphous materials in the interband region, Appl. Phys. Lett., 1996, 69, 371-373. (19) Masse, J.-P.; Szymanowski, H.; Zabeida, O.; Amassian, A.; Klemberg-Sapieha, J. E.; Martinu, L. Stability and effect of annealing on the optical properties of plasma-deposited Ta2O5 and Nb2O5 films, Thin Solid Films, 2006, 515, 1674–1682. (20) Mardare, A. I.; Savan, A.; Ludwig, A.; Wieck, A. D.; Hassel, A.W. High-throughput synthesis and characterization of anodic oxides on Nb-Ti alloys, Electrochim. Acta, 2009, 54, 5973-5980. (21) Schultze, J. W.; Lohrengel, M. M. Stability, reactivity and breakdown of passive films. Problems of recent and future research, Electrochim. Acta, 2000, 45, 2499–2513. (22) Mardare, A. I.; Ludwig, A.; Savan, A.; Hassel, A. W. Electrochemistry on binary valve metal combinatorial libraries:niobium-tantalum thin films, Electrochim. Acta, 2014, 140, 366– 375. (23) Lohrengel, M. M. Thin anodic oxide layers on aluminium and other valve metals: high field regime, Mater. Sci. Eng. R, 1993, 11, 243-294. (24) Semboshi, S.; Bando, K.; Ohtsu, N.; Shim, Y.; Konno, T. J. Structural and dielectric properties of anodic oxide film on Nb–Ti alloy, Thin Solid Films, 2008, 516, 8613–8619. (25) Habazaki, H.; Ogasawara, T.; Konno, H.; Shimizu, K.; Nagata, S.; Skeldon, P.; Thompson, G. E. Field Crystallization of Anodic Niobia, Corr. Sci., 2007, 49, 580–593. (26) Habazaki, H.; Shimizu, K.; Skeldon, P.; Thompson, G. E.; Wood, G. C. Inter–relationships between ionic transport and composition in amorphous anodic oxides, Proc. Royal Soc. A, 1997, 453, 1593-1609. (27) Piazza, S.; Santamaria, M.; Sunseri, C.; Di Quarto, F. Recent advances in photocurrent spectroscopy of passive films, Electrochim. Acta, 2003, 48, 1105-1114. (28) 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. (29) Hassel, A. W.; Diesing, D. Modification of trap distributions in anodic aluminum tunnel barriers, J. Electrochem. Soc., 2007, 154, C558-C561. (30) Di Quarto, F.; Sunseri, C.; Piazza, S.; Romano, M. C. Semiempirical Correlation between Optical Band Gap Values of Oxides and the Difference of Electronegativity of the Elements. Its Importance for a Quantitative Use of Photocurrent Spectroscopy in Corrosion Studies, J. Phys. Chem. B, 1997, 101, 2519-2525. (31) Di Quarto, F.; Piazza, S.; Santamaria, M.; Sunseri C. Handbook of Thin Film Materials; Academic Press: S. Diego, 2002; pp 373-380.

ACS Paragon Plus Environment

26

Page 27 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Combinatorial Science

(32) Di Quarto, F.; Santamaria, M.; Sunseri, C. Photoelectrochemical techniques in corrosion studies. in Analytical Methods in Corrosion Science and Technology; Marcus P., Mansfeld F., Eds.; Taylor and Francis: Boca-Raton, 2006; pp 697-702. (33) Mott, N. F.; Davis, E. A. Electronic Processes in Non-crystalline Materials, 2nd Ed., Clarendon Press: Oxford, 1979 (34) Zaffora, A.; Santamaria, M.; Di Franco, F.; Habazaki, H.; Di Quarto, F. Photoelectrochemical evidence of inhomogeneous composition at nm length scale of anodic films on valve metals alloys, Electrochim. Acta, 2016, 201, 333–339.

ACS Paragon Plus Environment

27

ACS Combinatorial Science

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Fig. 1: EDX compositional mapping of the Nb-Ti thin film combinatorial library. 254x190mm (300 x 300 DPI)

ACS Paragon Plus Environment

Page 28 of 36

Page 29 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Combinatorial Science

Fig. 2: SEM surface microstructure mapping of Nb-Ti thin film compositional spread (the at.% amount of Nb is specified in each image). 244x190mm (300 x 300 DPI)

ACS Paragon Plus Environment

ACS Combinatorial Science

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Fig. 3: Selected grazing incidence X-ray diffractograms measured along the Nb-Ti compositional spread. 254x166mm (300 x 300 DPI)

ACS Paragon Plus Environment

Page 30 of 36

Page 31 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Combinatorial Science

Fig. 4: Typical series of cyclic voltammograms (with step-wise increase of the maximum potential) as measured on the Nb-Ti compositional spread. Inlet: Typical potentiostatic anodisation performed at the end of each voltammogram. 246x190mm (300 x 300 DPI)

ACS Paragon Plus Environment

ACS Combinatorial Science

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Fig. 5: The imaginary part of the dielectric function presented for selected compositions as measured along the anodized Nb-Ti library. 254x173mm (300 x 300 DPI)

ACS Paragon Plus Environment

Page 32 of 36

Page 33 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Combinatorial Science

Fig. 6: Compositional mappings of the oxide formation factors as measured by VASE along the anodized NbTi library. 252x178mm (300 x 300 DPI)

ACS Paragon Plus Environment

ACS Combinatorial Science

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Fig. 7: Anodized Nb-Ti optical bandgap mappings measured by VASE for different thickness dependent datasets. 234x190mm (300 x 300 DPI)

ACS Paragon Plus Environment

Page 34 of 36

Page 35 of 36

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Combinatorial Science

Fig. 8: Non-linear behaviour of the measured optical bandgaps as compared to theoretical values and mapping of mixed oxide compositional deviations from parent metal composition (data from 18). 254x190mm (300 x 300 DPI)

ACS Paragon Plus Environment

ACS Combinatorial Science

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Graphical abstract 254x80mm (300 x 300 DPI)

ACS Paragon Plus Environment

Page 36 of 36