Niobium oxides and niobates physical properties: Review and prospects

Niobium oxides and niobates physical properties: Review and prospects

Progress in Materials Science 80 (2016) 1–37 Contents lists available at ScienceDirect Progress in Materials Science journal homepage: www.elsevier...

4MB Sizes 0 Downloads 43 Views

Progress in Materials Science 80 (2016) 1–37

Contents lists available at ScienceDirect

Progress in Materials Science journal homepage: www.elsevier.com/locate/pmatsci

Niobium oxides and niobates physical properties: Review and prospects C. Nico ⇑, T. Monteiro, M.P.F. Graça Department of Physics & I3N, University of Aveiro, Campus Universitário de Santiago, 3810-193 Aveiro, Portugal

a r t i c l e

i n f o

Article history: Available online 16 February 2016 Keywords: Niobium oxide Niobate Electron correlated material Polymorph Transparent conductive oxide Resistive switching Oxygen vacancies

Abstract: For the last 75 years several studies have been reporting on the physical properties of niobium oxides, but there is still many contradictory, inconsistent and insufficient information on these metal oxides. This review will begin by describing the niobium oxygen system and the different stoichiometric and nonstoichiometric phases, specifically Nb, NbO, NbO2, Nb2O5 and Nb2O5d. The crystalline phases and polymorphs of these materials are often inconsistently identified in different works and thus, a clarification of the nomenclature of the several niobium oxides polymorph and their crystalline structure is also presented. Due to their interesting physical properties, many applications of these materials have been suggested such as solid electrolytic capacitors, catalysis, photochromic devices, transparent conductive oxides or memristors, becoming obvious that a good understanding of niobium oxides physical properties and their control is essential and urgent. Additionally, a short review on different types of niobates, namely alkali niobates, columbite niobates and rare earth niobates and the relation of the properties of these materials with niobium oxides will be presented. Ó 2016 Elsevier Ltd. All rights reserved.

Contents 1.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

⇑ Corresponding author. Tel.: +351 936876475. E-mail address: [email protected] (C. Nico). http://dx.doi.org/10.1016/j.pmatsci.2016.02.001 0079-6425/Ó 2016 Elsevier Ltd. All rights reserved.

2

C. Nico et al. / Progress in Materials Science 80 (2016) 1–37

2.

3.

4.

Niobium–oxygen system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 2.1. Nb. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 2.2. NbO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 2.3. NbO2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 2.4. Nb2O5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 2.5. Non-stoichiometric niobium oxides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 Niobates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 3.1. Alkali niobates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 3.2. Columbite niobates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 3.3. Rare-earths orthoniobates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 Summary and prospects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

1. Introduction Niobium oxides can lead to many different and interesting properties, making it a very versatile group of materials. Specifically, niobium oxides have been showing great potentiality in many technological applications such as solid electrolytic capacitors, transparent conductive oxides, photochromic devices, memristors, dye-sensitized solar cells and others. Despite the known potentialities of niobium oxides, and many types of niobates, in several technological applications, the understanding of these oxide systems is still noticeably insufficient. The available literature reveals that niobium oxides are a complex system, with many phases and polymorphs, and many works reporting contradictory or inconsistent information. More specifically, stoichiometry problems are one of the most common complications in niobium oxides and in many different niobates, where NbO6 octahedra are typically the basic structural units. Hence, a careful bibliographic review is shown, which evidences the complexity of these materials, the difficulty in identifying their different phases and polymorphs, as well as in the interpretation of their properties. This work intends to be an important and solid groundwork for future research on niobium oxides and niobates systems, to enable a faster and more sustained scientific research and to evince the wide range of properties and applications of these materials.

2. Niobium–oxygen system In niobium–oxygen system, the niobium element can be found in four different charge states: 0, 2+, 4+ and 5+. Generally, these charge states are related to the phases of metallic Nb and to the NbO, NbO2 and Nb2O5 respectively, as Fig. 1 synthetically illustrates. The special complexity of the niobium oxides system is related to the existence of several stoichiometric and non-stoichiometric phases, some of which have several polymorphs, some of which are

Fig. 1. Schematic illustration of the different oxidation states of niobium.

C. Nico et al. / Progress in Materials Science 80 (2016) 1–37

3

Fig. 2. Niobium–oxygen phase diagram [19].

metastable, and the difficulty in synthesizing a single phase, i.e. without mixture of phases or stoichiometry. Controlling, identifying and determining small variations of the stoichiometry is rather difficult in niobium oxides as the structure of these phases may be extremely similar, and also because the precise quantification of oxygen is, technically, a great challenge. This is probably the major drawback of this system which can present a wide range of interesting physical properties, which are highly dependent on the phase, polymorph and stoichiometry. Although the first studies in niobium oxides are from Brauer in 1940 [1], and from other authors later in the decade of 1960 [2–11], there still has been no consensus in literature concerning the physical properties and the nomenclature of the different niobium oxides phases. The fact that many of the first scientific works on niobium oxides are written in German may partially justify the difficulty of the international scientific community having access to very important information regarding these materials and thus contributing for a better understanding of these materials. David Bach makes an ample collection and review of both early and recent literature of the niobium oxides system in his PhD thesis (2009), defended at the University of Karlsruhe [12]. However, it is stressed that many doubts about the synthesis, structure and physical properties of niobium oxides still remain [12]. Bach et al. [12–17] used electron energy loss spectroscopy (EELS) as a high spatial resolution technique for quantitative analysis of the niobium oxygen ratio and their oxidation state,

Fig. 3. Cubic crystalline structure of metallic Nb [21].

4

C. Nico et al. / Progress in Materials Science 80 (2016) 1–37

which is an extremely useful type of information, hardly obtained by any other technique, especially because the properties of niobium oxides highly depend on their stoichiometry. One of the first works on niobium oxides, by Elliot in 1960 [18], reports the phase diagram of the niobium–oxygen system. Since then, other phase diagrams have been reported, and one of the most recent, from 1990 by Massalski [19], is illustrated in Fig. 2. Massalski shows the existence of four thermodynamically stable phases of niobium–oxygen system (Nb, NbO, NbO2 and Nb2O5) with very narrow single-phase fields and negligible deviations from the exact stoichiometry [12,19]. However, one should consider that this phase diagram is incomplete, as it fails to describe the formation of stable non-stoichiometric phases at room temperature and the formation of different polymorphs, as it will be further mentioned.

2.1. Nb Niobium is a metal which crystallizes in a body-centred cubic (bcc) lattice (O9h spatial group) with a density of 8.57 g/cm3. It is a refractory metal, good thermal conductor, with a melting and boiling points at 2477 °C and 4744 °C, respectively [20]. It has an electrical resistivity of c.a. 15.2 lX cm at 273 K and it is a superconductor below the critical temperature Tc ’ 9.3 K [20] (see Fig. 3). Niobium can form very stable carbides, nitrides, borides and silicides, possessing high interatomic bonding energies [22]. Hence, niobium is used in the stabilization of stainless steels to prevent intergranular corrosion since it helps lowering the content of carbon in the steel. A small niobium amount, from 100 to 5000 ppm, is typically necessary for the prevention and retardation of austenitic grain growth which leads to higher strength and toughness [23]. These high-strength microalloyed steels are commonly used in pipelines for oil and gas transportation, in high rise buildings construction, in tools steels and heat resisting cast steel auto parts [23–25]. Niobium is also commonly added to most nickel-based superalloys for application in aircraft engines, land-based turbines for power generation, chemical processing industries and other applications where a special resistance to abrasion and corrosion is required [24,25]. Pure niobium and niobium-based alloys have a high resistance to corrosion in several environments, which is attributed to a passivating niobium oxide film at the surface [24–26]. These materials also present high melting points and refractory properties, particularly the niobium-based alloys which can surpass the maximum application temperatures of nickel-based superalloys, thus being well suited for space, nuclear and aircraft applications [24,25]. Furthermore, niobium-based alloys are typically characterized by low densities and relatively high ductility at room temperature which is an advantage for cold-working and fabrication of complex structures [23,24]. However, these alloys typically require a protective coating, such as a silicide, to prevent them from easily oxidizing at high temperatures [24]. Other applications of niobium comprise the fabrication of metal insulator metal (MIM) tunnel diodes and hard tissue replacements (given its good biocompatibility) [27,28].

Fig. 4. Cubic structure of NbO [21].

C. Nico et al. / Progress in Materials Science 80 (2016) 1–37

5

Niobium also presents type II superconductivity properties. In fact, from all the chemical elements, metallic niobium is the one with the highest critical temperature (Tc ’ 9.3 K) [20,26]. Other niobiumbased compounds, such as Nb3Ga or Nb0.75(GeAl)0.25, are among the materials with the highest critical temperatures for superconductivity (Tc = 20.0 K and Tc = 20.7 K respectively) [29]. Hence, niobium is frequently used in the fabrication of radio-frequency superconducting cavities (for particle accelerators), production of Josephson tunnel junctions, single photon detectors or superconducting quantum interference devices (SQUIDs) used, for instance, to cover the surface of nearly perfect spheres for the Gravity Probe B mission, which purpose was to measure the space–time curvature near the Earth [30,31]. However, pure metallic niobium is not suitable for superconducting magnets where high upper critical magnetic fields (Bc2) are required. Still, while niobium has a Bc2 ’ 420 mT (at 0 K), an alloy of NbTi can have a Bc2 ’ 11 T (at 4.2 K), and the Nb3Sn compound a Bc2 ’ 25 T (at 4.2 K) [32]. Therefore, it is no surprise that these niobium-based materials are fundamental and used, almost exclusively, in the production of high magnetic fields, required in a series of industrial and scientific applications. One flagrant example is the LHC project, in CERN, that required hundreds of tons of Nb and NbTi alloy. Niobium is also characterized by its high affinity and binding energy to oxygen [26]. Solubility of oxygen in the niobium matrix, where it occupies octahedral interstitial sites, increases with temperature (ranging from 0.8% at 500 °C to 9.0 at.% at 1915 °C) [19]. Oxygen in solid solution leads to an increase of the lattice parameter of metallic niobium, acts as an hardener and decreases the ductility of the metal [33]. Oxygen does also affect the electrical properties increasing the resistivity of niobium c.a. 4.1 lX cm per at.% of oxygen, and decreasing the superconductivity transition temperature Tc  0.93 K per at.% of oxygen [12,26]. Therefore, and also because defects at the surface may produce heat, high-purity niobium is required for superconductivity applications. It is commonly mentioned in literature the existence of a native oxide layer at the surface of niobium exposed to oxygen [26,34]. In a superconductor, the electromagnetic fields have a penetration depth of c.a. 60 nm, and therefore the existence of an oxide layer at the surface is critical. To obtain highly pure Nb surfaces, i.e. without an oxide layer, the metal is typically heated above 2000 K in ultra-high vacuum conditions [26]. The composition and thickness of such native oxide layer is not yet well clarified, but it is accepted that it depends on the processing parameters and techniques of the metal, crystallographic orientation, and ambient conditions such as oxygen partial pressure, temperature and exposure time [26,29,35– 40]. Some authors [26,35,38] suggest that this layer is around 5–6 nm thickness with a composition of nearly stoichiometric Nb2O5, while other authors [12,41] report thicknesses as high as 25 nm and/or with different layers of niobium suboxides [26,29,35,42,43]. More recently (2008), Delheusy et al. [43] suggested a three-layer model of NbO, NbO2 and Nb2O5 between the metallic niobium and the oxide surface.

Fig. 5. Structure of (a) tetragonal and (b) rutile NbO2 [21].

6

C. Nico et al. / Progress in Materials Science 80 (2016) 1–37

2.2. NbO NbO crystallizes in a face-centred cubic structure (similar to NaCl), O1h space group, with 3 Nb atoms in the lattice sites 3(c) (0 ½ ½; ½ 0 ½; ½ ½ 0), and 3 O atoms in the lattice sites 3(d) (½ 0 0; 0 ½ 0; 0 0 ½;) [7], where each Nb atom is coordinated to four O atoms in a square planar array, as illustrated in Fig. 4. Moreover, NbO crystalline structure is a very particular case since it shows 25% ordered vacancies in both sublattices of Nb and O [44]. This is the highest number of point defects among all transition metal monoxides. The density of NbO is c.a. 7.3 g/cm3 and has a melting point at Tm ’ 1940 °C [12,18]. Niobium monoxide presents a typical metallic behaviour, and is widely regarded as a metal [11,44,45], with a resistivity of about 21 lX cm at 25 °C [11,45] that decreases with temperature down to 1.8 lX cm at 4.2 K [45]. Kurmaev et al. [44] performed X-ray fluorescence measurements for different niobium oxides and compared the results of NbO to the band structure calculations of Nb1.0O1.0, finding that there were significant differences. Therefore, Kurmaev et al. tried to mimic the NbO structure by performing calculations of the band structure for Nb0.75O0.75 in order to make account for the 25% of vacancies. Given the good correspondence between these calculations and the experimental results, the authors assigned the electrical properties of NbO to the aforementioned vacancies, characteristic of this structure: ‘‘The electronic DOS of Nb0.75O0.75 is characterized by the formation of additional subbands connected with vacancy states.” [44]. Below the critical temperature, Tc ’ 1.38 K, NbO becomes superconductive [45]. Hulm et al. [45] performed electrical measurements and studied the superconductivity of different transition metal monoxides, and found that, in NbO, there is no evidence of negative temperature-coefficient behaviour of the type that was observed in TiO and VO. Furthermore, the authors verified that a small increase of the oxygen ratio (towards NbO2) would induce a sharp increase of the resistivity. Oppositely, an increase of the niobium ratio (towards metallic Nb) would promote an increase of the critical temperature for superconductivity [45]. The authors attribute this behaviour to a mixed-phase sample with contributions of both NbO and NbO2 (when the O% is increased), or of both Nb and NbO (when the Nb% is increased). Niobium monoxide is not used massively in any major technological application. However, the fact that NbO has improved properties, regarding the oxygen diffusion, in comparison with Nb, makes it a suitable candidate to niobium-based solid electrolytic capacitors [46–50].

2.3. NbO2 The NbO2 has the Nb element in a 4+ charge state, presents a stoichiometry between NbO and Nb2O5, and has a melting point of 1901 °C [20]. NbO2 is typically characterized by a blue colour (associated to the Nb4+ ions [8]) and can be obtained by controlled oxidation of Nb or NbO, or reduction of Nb2O5 [51]. At room temperature NbO2 crystallizes in what can be described as a distorted tetragonal superstructure with a rutile-type sublattice, with a space group C64h, and a density c.a. 5.9 g/cm3 [52,53]. This structure, illustrated in Fig. 5a is characterized by chains of edge-sharing NbO6 octahedrons which are cross-linked by sharing corners [52] where the Nb–O distances in the octahedral are practically the same, c.a. 2.05 Å [52]. Similarly to the VO2 phase at room temperature, in NbO2 the metal ions tend to pair along the fourfold [0 0 1] axis, along the edge-sharing chains, with alternated Nb–Nb distances of 2.80 and 3.20 Å [52,53]. Inherent instabilities of the NbO6 octahedra and the structure of NbO2 have been reported [53,54]. It is also known that between 797 and 808 °C, the NbO2 suffers a reversible second-order phase transition, together with a change of the crystal structure into a regular rutile lattice, illustrated in Fig. 5b [52,53,55,56]. Electrically, this is characterized as being a semiconductor–metal transition, where this high temperature NbO2 phase shows a typical metallic conductivity (c.a. 103 S/cm) [56]. It was also reported that above 1100 °C a small, but measurable, change of the stoichiometry towards NbO2.006 occurs, which can strongly influence the electrical conductivity of the room temperature phase [56,57].

C. Nico et al. / Progress in Materials Science 80 (2016) 1–37

7

Several metal dioxides, such as VO2, TiO2, GeO2 and SnO2, have a rutile-type structure, but their physical properties can be significantly different. In fact, GeO2 with a reported band gap between 5.4 and 5.9 eV [58] is considered an insulator, while SnO2 can be considered as a wide band gap semiconductor (3.6 eV) [59] as also the rutile TiO2 (3.05 eV) [60,61]. TiO2, GeO2 and SnO2 crystallize in a rutile structure at room temperature and, despite their different band gap energies, they all exhibit a fundamental indirect absorption edge with dipole forbidden direct band gap [61,62]. However, NbO2 is more frequently compared to VO2, not only due to their structural similarities, but particularly because they both show a semiconductor–metal transition. VO2 crystallizes in a monoclinic structure at room temperature with band gap between 0.6 and 0.7 eV, and undergoes a sharp transition to the rutile phase (metallic behaviour) near room temperature (340 K) at ultrafast timescales [63]. This makes VO2 an extremely promising material for a broad range of electronic applications (e.g. sensor, switching and memory elements) [63]. However, as aforementioned, NbO2 undergoes such metal–insulator transition at a much higher temperature (c.a. 1081 K), which can be an advantage to isolate the electric properties from temperature effects. In VO2 this type of behaviour is typically associated to a Mott–Peierls transition [63–66], a combination of the effect of the structural change in electronic structure, and strong electronic correlation commonly observed in oxides with 3d transition metals. However, Nb is a 4d transition metal and therefore it would be expectable that electron correlation would be less significant [64–66]. In fact, O’Hara and Demkov have recently shown that the metal–insulator transition in NbO2 is explained by a second-order structural (Peierls) transition via dimerization of chains of Nb atoms, without the need to account for electron correlation effects, unlike VO2 [64]. Wong et al. were able to show that in NbO2 the djj  epg orbital splitting is 0.3 eV larger than the observed in VO2, which may explain its higher metal–insulator transition temperature [66]. Some authors [67,68] have mentioned the existence of a monoclinic crystalline structure of NbO2, citing the work of Terao [69] as the reference for such phase. In addition, the Joint Committee on Powder Diffraction Standards – International Centre for Diffraction Data (JCPDS–ICDD) card No. 19-0859 does in fact report the X-ray powder diffraction (XRD) data of a monoclinic phase of NbO2, with lattice parameters a = 12.03 Å, b = 14.37 Å, c = 10.36 Å and b = 121.17°, also referring to the work of Terao [69]. However, in the abstract of the Terao’s work, it was only mentioned a tetragonal phase of NbO2, among other phases of niobium oxides. It is therefore reasonable to doubt about the true existence of such monoclinic structure, since the work of Terao is the only work reporting, allegedly, a fundamental structural study on an NbO2 monoclinic phase. The physical properties of NbO2, namely the optical and electrical ones, have not been yet extensively explored. Still, the tetragonal phase of NbO2 is usually classified as being an n-type semiconductor with a small band gap (between 0.5 and 1.2 eV) [65] and an electrical resistivity in the order of 104 X cm [44,53,55,56]. Weibin et al. [70] theoretically calculated the electronic structure of a tetragonal phase of NbO2, with an indirect band gap of 0.25 eV, but experimental results on thin films have shown a band gap of 0.5 eV, and a vacuum level (ionization energy) of 5 eV [70]. Recently, in an attempt to put an end to the inconsistent reports on NbO2 electronic structure, and in line with other rutile based oxide structures, theoretical and experimental studies reported an indirect band gap of at least 1.0 eV [65,71]. Also very recently, Wang et al. were able to show that the electrical conduction mechanism in tetragonal NbO2 is described by the Efros–Shklovskii variable range hopping model, rather than Mott’s model which is associated to VO2 [72]. In 2004, Zhao et al. [55] reported the optical and dielectric properties of nanostructured NbO2 (in the form of nanoslices), synthesized by thermal deposition, using a metallic niobium wire deposited in low vacuum as a thin film on a Si substrate. In this work, the authors report for the first time the Raman spectrum of NbO2, showing at least 12 active modes with energies of 139, 170, 247, 333, 343, 405, 436, 463, 580, 631, 699 and 815 cm1 [55]. Furthermore, the photoluminescence (PL) analysis of the NbO2 nanoslices, under 512 nm laser excitation and at room temperature, revealed two emission bands centred at 700 nm (1.77 eV) and 825 nm (1.50 eV). The origin and corresponding mechanisms for the observed PL were not clarified, but it was suggested that, in analogy with other metal dioxides, they may be attribute to intraionic Nb4+ dxy–dyz band transitions [55]. The dielectric characterization performed by Zhao et al., between 1 kHz and 10 MHz, revealed interesting properties of the NbO2 nanostructures. It was observed that the relative dielectric constant of NbO2 remains

8

C. Nico et al. / Progress in Materials Science 80 (2016) 1–37

almost constant, c.a. 10, in the measured frequency range, and with low dielectric losses (<1%) for the same frequency range [55]. This type of dielectric characteristics on nanostructures, allied to the simplicity of fabrication, may be quite attractive for the production of MOS devices, as suggested by the authors [55]. In the last couple years, many works have been exploring NbO2 due to its resistive switching properties for the development of memristive devices [73–78]. In fact, Slesazeck et al. [78] have recently proposed a model that describes the behaviour of NbO2 based filamentary switching devices, by a temperature activated Frenkel–Poole conduction mechanism, showing that the such switching effect is independent on the metal insulator transition. NbO2 has also been studied regarding other technological applications, such as impedimetric biosensors using biofunctionalized Nb/NbO2 electrodes [79,80]. Moreover, Sasaki et al. [81] took advantage of the good chemical stability of NbO2 to produce highly efficient electrocatalysts of Pt/NbO2/C nanostructures for oxygen-reduction reactions for application in proton exchange membrane fuel cells. In 2012 Lee et al. [80] reported the synthesis of NbO2 nanowires by chemical vapour transport method which can further enhance the potentiality of this material for such technological applications. 2.4. Nb2O5 Niobium pentoxide (Nb2O5) is the most thermodynamically stable state of the niobium–oxygen system. With an charge state of 5+ in Nb2O5, the electronic structure of the Nb atom is [Kr]4d0, which means that all the 4d electrons are bonded to the O 2p-band, thus justifying the fact that Nb2O5 has a much lower electrical conductivity than the other niobium oxides [50]. Nb2O5 can occur in the amorphous state or in one of many different crystalline polymorphs. Generally, all the Nb2O5 polymorphs have a white colour (in the form of powders) or transparent (in single-crystals). However, most of the physical properties of Nb2O5 depend on its polymorph and on the used synthesis parameters and technique [8,49,50,82–85]. The complexity, confusion and contradictions regarding the niobium oxides system are old and still prevail up to the present. In 1966, Schäfer et al. published a work entitled ‘‘The modifications of Niobium Pentoxide” [8], where it was stated that ‘‘The abundance of observations makes it desirable to subject the available material [niobium pentoxide] to a critical review and classification, and thus simultaneously to create a basis for further investigations”. Already then, there were many different phases of niobium oxides with different nomenclatures and there was, therefore, the need of clarifying, organizing and compiling all the existent information. Some of the Nb2O5 polymorphs were classified with a sequence of Greek letters as it is common in well-known systems. However, since there was not yet a well-established knowledge of all polymorphs, Schäfer et al. decided to use neutral symbols following, and extending, the same type of classification as Brauer’s [1]. Thus, some polymorphs were classified based on the temperature they were obtained: TT, T, M and H (from the German Tief– Tief, Tief, Medium and Hoch, meaning low–low, low, medium and high), while other polymorphs were named after the shape of the particles B, N and R (from the German Blätter, Nadeln and Prismen, meaning leaves/plates, needles and prisms). While Schäfer et al. [8] were able to match some of the same polymorphs reported by different authors, there were other niobium pentoxide modifications, such as the e, I-high and II, that they were not able to reproduce and study, suggesting the possibility of these phases to be metastable and/or non-stoichiometric. From what it is possible to assess from current literature, the opinion shared in this work is that not all Nb2O5 polymorphs are yet well known and studied, what justifies the choice of following the nomenclature suggested and reported by Schäfer et al. [8], also followed by some authors. There are however other authors that misuse this nomenclature by choosing the letter according to the crystalline structure, e.g. T-, H-, O- and M-Nb2O5, respectively standing for Tetragonal, Hexagonal, Orthorhombic and Monoclinic structures of niobium pentoxide [86,87], which can be confusing and misleading. The Nb2O5 phase can exist in an amorphous state, but it may crystallize in several kinds of polymorphs with different physical properties: T (D92h, orthorhombic), B (C62h, monoclinic), H (C12h, mono10 clinic), N (C32h, monoclinic), Z (C12, monoclinic), R (C32h, monoclinic), M (D17 4h, tetragonal), P (D4 , tetragonal) and also TT (pseudohexagonal or monoclinic) [8,12,83,88]. The TT-Nb2O5 polymorph can

C. Nico et al. / Progress in Materials Science 80 (2016) 1–37

9

be interpreted as a less developed phase of the T-polymorph, which is only stabilized by impurities, such as OH or Cl, or vacancies [13,83,89]. Among all the Nb2O5 polymorphs, and niobium oxides in general, the H-phase is thermodynamically the most stable. The H-Nb2O5 is therefore one of the most common, and probably the most studied, niobium pentoxide polymorph. A lot more of metastable Nb2O5 modifications were reported, but not extensively studied [8,90–92]. Generally, the structures of niobium pentoxides are based on NbO6 octahedra which may be more or less distorted depending on the type of linkage between the octahedra. Most of Nb2O5 polymorphs structures are described by the combination of one or both of these types of links between octahedra that can happen by corner-sharing or by edge-sharing. Particularly, the structures characterized by perpendicular edge-sharing octahedra are arranged in such a way that a zigzag chain is produced. This type of chains, linked in parallel by corner-sharing, are very common as an additional constructional element in many forms of niobium pentoxide but also in related compounds [8]. Actually, the fact that there are several possible combinations of octahedral linkages that can produce an O/Nb ratio of 2.5, is pointed as the reason for the multiplicity of Nb2O5 structures [8]. Furthermore, the principle of crystallographic shear formed by the existence of different types of linkage regions (but preserving the cation coordination), but also the formation of point defects, can explain the possibility of variations of stoichiometry in relation to Nb2O5, giving rise to the formation of non-stoichiometric niobium oxide phases [8,93–97]. In a very recent paper, Valencia-Balvín et al. [83] report first-principle studies on the energetics and phase stability of the several Nb2O5 polymorphs under pressure, and also perform a comprehensive and detailed description of their crystalline structures. Additionally, the structural characteristics of the different niobium oxides can vary significantly with the synthesis technique [8,34,82,83,85]. This is clearly evidenced by Reznichenko et al. [82] who make a very complete list of numerous works in the literature with different production methods of several stable and metastable Nb2O5 polymorphs, their structural parameters and different designations or nomenclatures. The most reported techniques used to prepare different Nb2O5 polymorphs involve the oxidation of niobium oxides of lower stoichiometry, promoted by heating in air, or by heat treating other Nb2O5 phases. Other methods, such as heating sulphate or chloride niobic acid, anodization, chemical transport, sol–gel, special hydrothermal conditions, or high temperatures and/or pressure, are also commonly used in the production of different Nb2O5 polymorphs [8,49,50,82,84,85,89]. Generally, the temperature and starting material used in the synthesis method will be the most determinant parameters. One interesting behaviour that is observed in the synthesis of the niobium oxides, is the ‘‘memory of solids”, as it was referred by Schäfer et al. [8]. This behaviour is characterized by how the same Nb2O5 phase will behave differently under the same heat treatment, depending on the original preparation

Fig. 6. Monoclinic structure of the H-Nb2O5 phase [21].

10

C. Nico et al. / Progress in Materials Science 80 (2016) 1–37

Fig. 7. Monoclinic structure of the B-Nb2O5 phase [21].

Fig. 8. Monoclinic structure of the N-Nb2O5 phase [21].

method, i.e. similar crystalline samples undistinguishable by XRD measurements will ‘‘remember” their synthesis method by changing to different phases under the same heat treatment [8]. While this behaviour is not yet well clarified, it was suggested that it might be related to the incorporation of different impurities or specific structural defects, that are dependent on the original synthesis route [8]. The H-Nb2O5 phase is rather easy to be obtained. Starting from any Nb2O5 polymorph, or a lower stoichiometric oxide (such as NbO2, NbO or even the metallic Nb), the H-Nb2O5 is obtained by a heat treatment in air at high temperatures (>1000 °C) [8,12]. Heating any type of niobic acid precipitates (sulphate, chloride, bromide, iodide or fluoride) at high temperatures will also produce this phase [8]. The production of the H phase can also be achieved under hydrothermal conditions and from a melt (if seeded with the same phase) [8]. The monoclinic structure of the H-Nb2O5, which contains 14 formula units per unit cell, is illustrated by Fig. 6. The Nb atoms occupy the Wyckoff positions 1g, 2i (with half occupation), six 2n and seven 2m, while the O atoms occupy the 1f, 1h, sixteen 2n and eighteen 2m positions. This structure is described by layered 3  4 and 3  5 blocks of cornersharing octahedra. These blocks are connected by edge-sharing octahedra, with a partly but systematic tetrahedral site, occupied by one Nb atom per unit cell [83]. The B-Nb2O5 phase can be produced by chemical transport of Nb2O5 or NbOCl3 at temperatures between 750 and 850 °C. However, it was observed that using this method other polymorphs are produced in addition, e.g. N and P-Nb2O5 [8]. It is possible to obtain a single B phase sample from molten Nb2O5 using a selected crystalline seed of the same phase [8]. Alternatively, it was reported that by

C. Nico et al. / Progress in Materials Science 80 (2016) 1–37

11

Fig. 9. Tetragonal structure of the M-Nb2O5 phase [21].

Fig. 10. Tetragonal structure of the P-Nb2O5 phase [21].

heating, in air, between 500 and 800 °C metallic Nb or a lower stoichiometry niobium oxide (NbO or NbO2 for instance), as well as the TT or the T-Nb2O5 polymorph, is a good method for producing the B phase [8]. The monoclinic structure of this polymorph, described by edge-sharing distorted NbO6 octahedra blocks (each block linked in zigzag by corner-sharing) is illustrated by Fig. 7. This structure contains 4 formula units per unit cell, with Nb atoms located at the Wyckoff position 8f and the O atoms at 4e and two 8f positions [83]. The N-Nb2O5 phase can be found, as mentioned, together with the M phase when synthesized from niobic acids. It can also be prepared by chemical transport of Nb2O5 at 840 °C, but only in the presence of small amounts of fluoride, producing additional phases like the H and B polymorphs [8]. The N polymorph can also be produced under hydrothermal conditions where the OH groups may play the same role as the F [8]. It is also possible to prepare this polymorph by thermal decomposition of NbO2F, at c.a. 1000 °C in high vacuum, though it is a very sensible method given that the formation of Nb3O7F or H-Nb2O5 can easily occur if the time and temperature are not optimal [8]. The monoclinic

12

C. Nico et al. / Progress in Materials Science 80 (2016) 1–37

Fig. 11. Monoclinic structure of the R-Nb2O5 phase [21].

Fig. 12. Monoclinic structure of the Z-Nb2O5 phase [21].

Fig. 13. Orthorhombic structure of the T-Nb2O5 phase [21].

structure of the N-Nb2O5, illustrated by Fig. 8, is described by 4  4 blocks of corner-sharing octahedra, where the blocks are interlinked by edge-sharing. The Nb and O atoms occupy, respectively, eight and twenty 4i Wyckoff positions [83].

C. Nico et al. / Progress in Materials Science 80 (2016) 1–37

13

The M-Nb2O5 phase is usually formed by heating sulphate or chloride niobic acid between 900 and 950 °C for some hours (or even at higher temperatures but for shorter times). This phase can also be detected on oxidized Nb and Nb alloys when heat treated at 1000 °C in air [8]. Nevertheless, these methods also produce additional polymorphs, especially the N phase, hampering the preparation of a single phase sample of M-Nb2O5 [8]. Very recently, Dhawan et al. [98] were able to prepare spherical quantum dots of M-Nb2O5, by physical vapour deposition, with controllable sizes between 1 and 20 nm. For quantum dots with diameters below 5 nm, quantum confinement effects were observed [98]. The tetragonal structure of the M phase is illustrated in Fig. 9. The unit cell contains 16 formula units, where the Nb atoms occupy two 8i, two 8h and three 16l Wyckoff positions, and the O atoms at two 8j, two 8h and three 16l. The structure can be described by 4  4 blocks of corner-sharing octahedra, with adjacent blocks linked by octahedra edges [83]. The P-Nb2O5 phase can be obtained by chemical transport, either in the chloride, bromide or iodide systems, at a temperature c.a. 750 °C [8]. It was found that the presence of small amounts of water was fundamental to promote the growth of the P phase [8]. Alternatively, the P-Nb2O5 can be formed by very slow thermal decomposition of NbO2F [8]. The idealized structure of the P polymorph, illustrated in Fig. 10 (courtesy of Professor Osorio-Guillén from the Physics Institute of the University of Antioquia, Colombia), contains 4 formula units per unit cell where the Nb atoms occupy the Wyckoff position 8c while the O atoms occupy a 4i and two 8c positions [83]. The structure is composed by distorted octahedra organized in blocks of two edges-sharing, where the blocks are linked by corner-sharing [83]. The R-Nb2O5 phase was first reported by Gruehn in 1966 [90]. It is not easy to find works which have identified this phase, though this polymorph is commonly referred in literature [83,99,100]. Originally, Gruehn obtained the R phase by chemical transport of Nb2O5 at temperatures between 600 and 800 °C. Using the same method with chloride niobic acid also produced this phase. Additionally, Gruehn said that the R-Nb2O5 was also found in the product of NbOCl3 hydrolysis heated at 275 °C [90]. These methods, however, did not produce a single phase material, but constantly a mixture of R-Nb2O5 with other polymorphs (e.g. P or TT) [90]. The R phase has one of most simple structures among niobium pentoxides, illustrated in Fig. 11. This monoclinic structure has two formula units per unit cell with the Nb atoms occupying the Wyckoff positions 4i and the O atoms the 2a and 4i [83]. This structure is described by distorted octahedra linked by edge-sharing, forming zigzag chains along the b direction which are interlinked by corner-sharing [90]. The Z-Nb2O5 phase was first reported by Zibrov et al. in 1998 [101]. This modification was identified, together with the B polymorph, after heat treating the H-Nb2O5 phase, between 800 and 1100 °C, in a high-pressure chamber at 8.0 GPa during 1–10 min [101]. The structure was determined and refined by the Rietveld method from the XRD measurements, and is illustrated is Fig. 12. It was found that this monoclinic structure is one of the few niobium pentoxides without sixfold coordinated niobium, presenting mono-capped trigonal prisms with sevenfold coordination Nb atoms occupying the Wyckoff position 4c, while the O atoms occupy a 2b two 4c positions [83,101]. The T-Nb2O5 phase is one of the most commonly studied niobium pentoxide phases and one of the first to be reported (originally by Brauer in 1941 [1]). This phase can be obtained by heating sulphate of chloride niobic acid between 600 and 800 °C. Heating the TT-Nb2O5 polymorph, a lower niobium oxide (NbO2 or NbO), or metallic Nb, to the same range of temperatures (600–800 °C) does also produce the T phase [8]. Additionally, this polymorph can be obtained under hydrothermal conditions (starting from amorphous niobic acid), by quenching Nb2O5 into supercooled melts, or even by chemical transport of Nb2O5 [8]. The orthorhombic structure of the T phase, illustrated in Fig. 13, is not simple and contains 8.4 unit formulas per unit cell. Nb atoms with six and sevenfold coordination produce distorted octahedra and pentagonal bipyramids, respectively. These polyhedra are linked both corner and edge-sharing along the [0 0 1] direction but are only connected by corner-sharing along the c axis [83]. There are 16 Nb atoms, distributed in parallel with the (0 0 1) plane and placed at four 8i Wyckoff positions with half-occupancy, while 0.8 Nb atoms are distributed in three 4g positions with 0.08, 0.08 and 0.04 occupancies [83]. The occupancy of the Nb atoms is represented in Fig. 13 by the percentage of blue that each atom is coloured. The O atoms occupy one 2b, four 4g and six 4h Wyckoff positions [83].

14

C. Nico et al. / Progress in Materials Science 80 (2016) 1–37

Table 1 Niobium oxides crystalline phases. Phase

Crystal system

Lattice parameters

Space group

References

NbO NbO2

Cubic Tetragonal

a = 4.210 Å a = 13.695 Å c = 5.981 Å a = 4.55 Å/a = 4.841 Å c = 2.86 Å/c = 2.992 Å a = 12.03 Å b = 14.37 Å c = 10.36 Å b = 121.17° a = 15.686 Å b = 3.831 Å c = 20.71 Å b = 121.17° a = 3.832 Å b = 2.740 Å c = 28.890 Å a = 15.749 Å b = 3.824 Å c = 17.852 Å b = 102.03° a = 6.175 Å/a = 6.144 Å b = 29.175 Å/b = 29.194 Å c = 3.930 Å/c = 3.940 Å a = 12.73 Å b = 4.88 Å c = 5.56 Å b = 105.1° a = 21.153 Å/a = 21.163 Å b = 3.8233 Å/b = 3.824 Å c = 19.356 Å/c = 19.355 Å b = 119.80° a = 28.51 Å b = 3.830 Å c = 17.48 Å b = 120.8° a = 5.219 Å b = 4.699 Å c = 5.928 Å b = 108.56° a = 12.79 Å b = 3.826 Å c = 3.983 Å b = 90.75° a = 20.44 Å c = 3.832 Å a = 3.876 Å c = 25.43 Å a = 3.607 Å/a = 3.600 Å c = 3.925 Å/c = 3.919 Å a = 7.23 Å b = 15.7 Å c = 7.18 Å b = 119.08°

O1h C64h

[7,13] [52,55]

D14 4h

[54,179]

C2h

[53,67,68]

C32h

[173,175,178]

D17 2h

[173,175,178]

C12h

[173]

D92h

[99,100,102,180]

C62h

[99,100,181,182]

C12h

[99,102,180–183]

3 C2h

[99,100,181,182]

C12

[101]

C32h

[99,100]

D17 4h

[99,100]

D410

[99,100]



[86,99,100,102,180]

Rutile Monoclinic

Nb12O29

Monoclinic

Orthorhombic

Nb22O54

Monoclinic

T-Nb2O5

Orthorhombic

B-Nb2O5

Monoclinic

H-Nb2O5

Monoclinic

N-Nb2O5

Monoclinic

Z-Nb2O5

Monoclinic

R-Nb2O5

Monoclinic

M-Nb2O5

Tetragonal

P-Nb2O5

Tetragonal

TT-Nb2O5

Pseudohexagonal Monoclinic

Finally, the TT-Nb2O5 phase is also frequently reported in literature, and is commonly referred as a less crystalline form of the T phase, stabilized by vacancies or impurities, such as OH or Cl [83,85,89]. It is reported that the TT-Nb2O5 can be obtained by heating sulphate or chloride niobic acid to temperatures c.a. 500 °C, or by promoting the oxidation of lower niobium oxides by heating between 320 and 350 °C in air. In both methods, amorphous Nb2O5 is observed as an intermediate.

C. Nico et al. / Progress in Materials Science 80 (2016) 1–37

15

This phase is also observed in oxidized Nb alloys at higher temperatures (800 °C), and as one of the products of the reaction between NbO2 and Cl2 carried out between 270 and 320 °C. It is reported that the TT-Nb2O5 can crystallize in either monoclinic or pseudo-hexagonal structure [83,85,99,100,102], but the unit cell was never fully refined. It is affirmed, nevertheless, that the structure of the T phase is characterized by the presence of distorted octahedra, pentagonal and hexagonal bipyramids, i.e. NbO6, NbO7 and NbO8 polyhedra [83,102], which are reported to be the same structural units as the amorphous phase of Nb2O5 [13,103]. A list of these Nb2O5 polymorphs (and other niobium oxides), with the respective crystal lattice parameters and space groups, is presented in Table 1. Regarding the physical properties of niobium pentoxides, one should expect that, depending on the Nb2O5 polymorph, some of the properties may be different. In fact, not only the physical properties depend on the polymorph, but also on the synthesis method [8,13,50,82,104]. Because of this, it is common to find a wide range of values for some properties reported for Nb2O5. Nonetheless, independently on the polymorph, Nb2O5 is generally considered a wide band gap semiconductor (or insulator, depending on the classification criteria) [44,50]. Filho et al. performed photoelectrochemical studies and measured the band gap of thin films of amorphous niobium oxides, which was consistently around 3.4 eV, in agreement with results of previous works [105]. The authors also measured a high quantum yield on certain niobium oxides films and discussed the charge transfer model between this material and an electrolyte. Schultze and Lohrengel [106] report niobium pentoxide passive films with band gap energies between 3.4 and 5.3 eV. Brayner and Bozon-Verduraz [89] report a nearly constant band gap of 3.4 eV for niobium pentoxide, in the amorphous, TT, T and H phases. However, the authors also report that, using a soft route method they were able to produce nanoparticles, with sizes between 40 nm and 4.5 nm, which band gap varied between 3.4 and 4.2 eV, respectively [89]. This variation was assigned to quantum confinement effects. It was also suggested that the local coordination of the Nb atoms can be a determinant factor in the band gap of niobium pentoxides and other niobates [89]. Additionally, it was reported that Nb2O5 particles heat treated at low temperatures (around 400 °C) presented a brownish-white colour, turning into white after being heated, in air, to temperatures higher than 600 °C (the characteristic colour of niobium pentoxide), which was assigned to oxygen vacancies originated during the synthesis process and respective vacancy filling with increasing temperature [89]. Soares et al. [50] reported about the effect of the processing method on the optical and electrical properties of niobium pentoxides. A band gap of c.a. 3.0 eV for Nb2O5 single crystals (H phase) at room temperature, increasing 100 meV when cooled down to 14 K, was observed. Additionally, photoconductivity measurements revealed identical band gap values for polycrystalline samples of H-Nb2O5, while for polycrystalline T-Nb2O5 two photoconductivity peaks at 3.4 and 4.7 eV were respectively assigned to the fundamental absorption edge and to a transition from the conduction band to a higher energy band structure of this phase [50]. In 2011, Kovendhan et al. [107] reported a band gap of 3.75 eV in MNb2O5 crystalline thin films. Very recently, Dhawan et al. [98] observed, through optical absorption measurements, a variation of the band gap from 4.15 to 3.36 eV in M-Nb2O5 spherical quantum dots between 1 and 20 nm in diameter, respectively. While such variation was assigned to quantum confinement effects, it was only detected in quantum dots with less than 5 nm in diameter. For bigger sizes the band gap was nearly constant, therefore indicating that the exciton Bohr radius in the MNb2O5 phase should be c.a. 5 nm. Additionally, a small effective mass of the exciton l ffi 0.532 was reported [98]. Weibin et al. [70] calculated the electronic structure of B-Nb2O5 with density functional theory, which results have indicated an indirect band gap of 2.55 eV. Additionally, the authors performed X-Ray and Ultraviolet Photoelectron Spectroscopy (XPS and UPS) measurements on Nb2O5 films deposited by pulsed laser deposition on a Si(1 0 0) substrate, not specifying however the crystalline structure (or amorphous nature) of the samples, and the results pointed to a band gap of 3.7 eV and a vacuum level (ionization energy) of approximately 4.2 eV [70]. It is occasionally stated that Nb2O5 presents n-type conductivity [83,104], but in fact this type of conductivity was reported, and it is today strongly accepted, as being associated to small stoichiometry deviations [108,109]. Nevertheless, reported values for conductivity of Nb2O5 are wide-ranging and inconsistent. For amorphous Nb2O5, Fischer et al. [110] reported electrical conductivities of 1011 S/cm, while Cavigliasso et al. [111] reported 1013–9  1013 S/cm, and Macek and Orel [112]

16

C. Nico et al. / Progress in Materials Science 80 (2016) 1–37

reported conductivities in the order of 1012–1011 S/cm, all at room temperature. Regarding crystalline Nb2O5 samples, Soares et al. [50] report conductivity values, at room temperature, that range between 3.7  1010 S/cm (for the T phase) and 7.6  107 S/cm (for the H phase). Graça et al. [84], prepared polycrystalline Nb2O5 samples by the sol–gel method and observed conductivities ranging from 106 S/cm (for the H phase) to 1013 S/cm (for the T phase). Schäfer et al. [57], by its turn, reported a conductivity of 3  106 S/cm for the H-Nb2O5 phase. While all these results are in fact inconsistent, it appears that the H phase is often reported to be more conductive than other stoichiometric niobium pentoxides. Since the electrical conductivity of these materials are essentially related to oxygen deficiencies, a more detailed discussion on the conductivity mechanisms is presented in the following section regarding the non-stoichiometric niobium oxides. The reported values for the dielectric constant of the Nb2O5 are not consistent as well. For amorphous Nb2O5 films formed by anodization, depending on the forming electrolyte Cavigliasso et al. [111] report dielectric constants between 49 and 120, while Schultze and Lohrengel [106] report values between 41 and 46. For the T-Nb2O5, there are reported dielectric constants between 40 and 200 at 100 kHz (at room temperature) depending on the synthesis method and orientation [50,84]. For the H phase, values between 38 and 170 at 100 kHz (at room temperature) have been reported [50,84]. In most works, is not clear at which frequency and temperature these dielectric constant values are referring to, but it is typically stated [13,113–115] that the static dielectric constant of niobium pentoxide is 41 when comparing with the static dielectric constant at room temperature of other materials for the purpose of dielectric applications. Very recently, El-Shazly et al. [116] calculated of the electronic structure and optical properties of the B-Nb2O5 polymorph, reporting a static dielectric constant of 5 and a refractive index from 2.5 to 3 within the visible range of the electromagnetic spectrum. In fact, besides the reported potential [13,46,110,114] for the production of solid electrolytic capacitors, the niobium pentoxides possess a wide range of interesting properties which makes this system suitable for many different applications. The high dielectric constant makes Nb2O5 an interesting material for complementary metal–oxide–semiconductor (CMOS) devices, or MIM tunnel diodes, or MIM capacitors [117]. Nb2O5 is also commonly reported for its photo and electrochromic properties [112,118–125] being possible to change the colour of thin films by applying a voltage. It has also been used as photoelectrode for dye-sensitized solar-cells (DSSCs) as an alternative to, or together with, TiO2 [87,118,126,127], offering higher open circuit voltages [128] and the possibility to achieve a higher light absorption coefficient by inducing oxygen deficiencies [129]. Another active topic of research regarding the niobium pentoxides is their application as a catalyst [130]. Given its high catalytic activity, selectivity at low temperatures and stability, specially of amorphous hydrated Nb2O5nH2O, it has been used to catalyse different types of reactions such as esterification, hydrolysis, dehydration, condensation or alkylation [103,130]. The use of niobium pentoxide as a catalyst for hydrogen storage has also been reported [131]. The application of Nb2O5 in lithium batteries [86,118], humidity sensors [132], electrochemical biosensors [133], and as a biomaterial (given its chemical stability and low cytotoxicity) [134–136] have been reported. The incorporation of Nb2O5 in different glass systems [137–141] has been explored essentially for their interesting non-linear optical properties. The production of rare-earth doped Nb2O5 thin films for optical waveguides and amplifiers have also been recently reported [142]. Niobium pentoxide also finds applications as hard coatings for optical glasses and lenses by taking advantage of its low optical absorption coefficient, high refractive index, chemical and thermal stability and mechanical resistance [143]. Optical transparency of Nb2O5 is typically reported to be around 90% in the visible range, strongly decreasing for wavelengths smaller than 400 nm [107,129,144,145]. However, the transparency is highly dependent on oxygen deficiencies, allowing to achieve black Nb2O5 nanostructures for enhanced solar light absorption purposes [129]. Amorphous Nb2O5 has a reported refraction index between 2.1 and 2.6 for visible wavelengths, depending on the synthesis method, microstructure and internal stress [144,146,147], higher than HfO2 and Ta2O5 (which are also used in optical components). In fact, mechanical stress is one of the major issues to solve in order to achieve high refraction index and low extinction coefficients [146,147]. Vinnichenko et al. have reported the synthesis of Nb2O5 films with low mechanical stress (90 MPa) by reactive pulsed magnetron sputtering in a high plasma flow configuration, which allowed them to achieve refraction indexes as high as 2.54 at 400 nm and an extinction coefficient as low as 6  104 at 400 nm [146].

C. Nico et al. / Progress in Materials Science 80 (2016) 1–37

17

Fig. 14. Monoclinic structure of the Nb22O54 phase [21].

A very recent and promising application of Nb2O5, based on the resistive switching behaviour reported in niobium oxides in 1965 by Hiatt and Hickmott [4], is the application as resistive random access memories (ReRAM). This type of application and working principle fits the concept of a passive electronic component called memristor, which existence was first suggested in 1971 by Chua [148] and proved to exist only in 2008 by IBM Labs using TiO2 [149]. The memristor is considered to be a fourth basic circuit element in electronics, along with the concept of resistor, capacitor and inductor, and is defined by the dependence of its resistance on the history (i.e. integral over time) of the current that passed through it. This is currently an active and fast-growing area of research [150–158], with many related works based on niobium oxides [77,159–167]. Recently, a comprehensive review paper was published by Rani et al. [85] regarding the fundamental properties, some recent applications and particularly synthesis methods of different niobium pentoxides micro and nanostructures, such as thin films, nanorods, nanobelts or nanospheres. Such work corroborates what was aforementioned, particularly the inconsistencies on the identification of the Nb2O5 polymorphs, their wide range of applications and also stoichiometry issues. 2.5. Non-stoichiometric niobium oxides It is known that crystallographic shear and point defects are the most common types of defects that produce/accommodate stoichiometry changes in certain metal oxides (such as TiO2) [93,94,168] but which are especially easy to occur in niobium oxides [51,95–97,169]. However, point defects cannot explain high variations of stoichiometry since their concentration in a crystal lattice is rather limited (usually not exceeding 104) [95]. Crystallographic shear on the other side, leads to the formation of planes where the metal oxide octahedra changed their linkage (e.g. from corner to edge-sharing) thus accommodating large oxygen deficiencies keeping the metal ion coordination [93,95]. In addition, the formation of such ordered defects can lead to the formation of different phases of the metal oxide. Citing Mrowec [94], ‘‘[. . .] the mutual interaction between point defects results in formation of complexes and defect clusters called extended defects. These extended defects can become further ordered, which leads to superstructure ordering and to formation of intermediate phases. In some cases point defects can become eliminated in the process of crystallographic shear which is connected with formation of a whole series of intermediate phases.” It is possible to find several non-stoichiometric niobium oxides reported in literature. Essentially, these can be divided in two groups: one with stoichiometry between Nb and NbO, and other with stoichiometry between NbO2 and Nb2O5, as illustrated before by Fig. 1. There are three reported nonstoichiometric niobium oxides between Nb and NbO, which were found to be metastable and were

18

C. Nico et al. / Progress in Materials Science 80 (2016) 1–37

Fig. 15. (a) Structure of the orthorhombic Nb12O29 phase; (b) crystallographic shear planes depicted with the light blue NbO6 octahedra [21].

classified as NbOx, NbOy and NbOz [170]. NbOx was reported to be formed at 270–500 °C with a stoichiometry equivalent to Nb6O, NbOy was formed at 330–500 °C with stoichiometry equivalent to Nb4O, and NbOz was formed at 400–700 °C with an unknown exact stoichiometry [170]. To date, there are very few published works on this group of metastable non-stoichiometric niobium oxides. Most of the reported non-stoichiometric niobium oxides are about those having a stoichiometry between NbO2 and Nb2O5, and more precisely between NbO2.4 and NbO2.5. Schäfer et al. [8] referred the existence of such phases, with close but different stoichiometries, classified as ox I to ox VI, and suggested that all, or some of such phases might belong to the same mixed-crystal phase. Subsequent works, Schäfer et al. [57] and Kimura [171], reported the observation of measurable single phase ranges of non-stoichiometric niobium oxides, such as Nb12O29, Nb22O52, Nb47O116, Nb25O62 and Nb53O132. Later studies, between 900 and 1300 °C, reported different results which were summarized and classified as ‘‘incomplete and inconsistent” by Marucco [108]. Hence, Marucco [108] performed a critical review and a detailed study in order to clarify the thermodynamic existence of such phases, ‘‘[. . .] by the method of equilibration between oxides and buffer gaseous mixtures, which is the most reliable”. However, between 1000 and 1100 °C, Marucco only found two stable phases within the same composition range (2.4 < x < 2.5)—Nb12O29 (x = 2.42) and Nb25O62 (x = 2.47)—suggesting that the previously reported non-stoichiometric phases were metastable or were associated with the sensitivity of the different methods that were used, or by the dif-

C. Nico et al. / Progress in Materials Science 80 (2016) 1–37

19

Fig. 16. (a) Structure of the monoclinic Nb12O29 phase; (b) crystallographic shear planes depicted with the light blue NbO6 octahedra [21].

ficulty in keeping the phases unchanged after quenching. In a later study regarding the electrical resistance of niobium oxides, Marucco [172] explained that phases with very small stoichiometry deviations from the stable phases of Nb2O5, Nb12O29 and Nb25O62, were in fact possible to be found and that this deviation was not related to coherent intergrowth in the crystal, but instead to point defects, oxygen vacancies in stable phases, and niobium interstitials in metastable phases (e.g. Nb22O54 was found to be stable only above 1250 °C). Overall, Marucco [172] concluded that the only stable phases of niobium oxides with a stoichiometry between NbO2.4 and NbO2.5 are Nb2O5, Nb12O29 and Nb25O62 where such variations in stoichiometry are possible to occur, and can be interpreted as single or doubly charged oxygen vacancies in their structure, resulting in significant variations of the electrical resistance. Since then, apart from the Nb12O29, there have been almost no reports on non-stoichiometric niobium oxide phases that could clarify these doubts about the stability of phases and their physical properties. Still, there is a recent work reporting the synthesis and structural characterization of the crystalline phase of Nb22O54 [173].

20

C. Nico et al. / Progress in Materials Science 80 (2016) 1–37

Digesting the available information to date, one can say that only three crystalline phases, stable at room temperature, of non-stoichiometric niobium oxides were reported: the monoclinic Nb22O54, and the monoclinic and orthorhombic Nb12O29 phases. The several other non-stoichiometric phases are reported to be metastable, i.e. to exist only at high temperatures. The synthesis of these non-stoichiometric phases may be achieved by a controlled oxidation of lower niobium oxides (e.g. NbO2, NbO or even metallic Nb), or by heat treating Nb2O5 samples in a reducing atmosphere at high temperatures (from 900 to 1300 °C) [8,57,108,172]. It was also recently reported the synthesis of crystalline Nb12O29 (both monoclinic and orthorhombic phases), in the form of thin films, by dc magnetron off-axis sputtering at room temperature, using metallic Nb as a target, after what they were annealed in vacuum at 1000 °C to promote the crystallization [174]. Moreover, McQueen et al. [173] successfully grown single crystals of orthorhombic Nb12O29 and Nb22O54 by mixing the right amounts of NbO2 and Nb2O5 powders into the desired stoichiometry, and by subsequently melting the mixtures in a vacuum furnace back-filled with argon at 1425 °C. The single crystals obtained by McQueen et al. [173], allowed them to refine and report, for the first time, the crystal structure of the Nb22O54. This phase crystallizes in a monoclinic structure (space group C12h) with two formula units per unit cell, as illustrated in Fig. 14. The structure is described by 3  3 and 4  3 blocks of NbO6 octahedra linked by corner-sharing, and by Nb atoms in a tetrahedral coordination. The Nb atoms with octahedral coordination occupy six 2m, four 2n and one 1e Wyckoff positions, while the Nb atoms with tetrahedral coordination occupy a 2i, a 2n and a 2m positions with occupancies of 0.403, 0.068 and 0.029 respectively. The O atoms occupy thirteen 2m, thirteen 2n, one 1c and one 1d Wyckoff positions. It should be noticed that in the 3  3 and 4  3 blocks, the niobium atoms are displaced from the middle of the octahedra towards the centre of the block, forming an antiferroelectric ordering of the electric dipoles [173]. The Nb12O29 phase is the most commonly reported non-stoichiometric niobium oxide phase. This phase can crystallize into an orthorhombic structure (space group D17 2h), or in a monoclinic structure (space group C32h) [173,175]. The structures of Nb12O29, both orthorhombic and monoclinic polymorphs, are based on 4  3 blocks of corner-sharing NbO6 octahedra, where the blocks are linked by edge-sharing. The orthorhombic structure of Nb12O29, illustrated in Fig. 15, has six formula units per unit cell, where the Nb atoms occupy six 8f Wyckoff positions and the O atoms occupy thirteen 8f and three 4c positions [173,175]. The structure of the monoclinic polymorph of Nb12O29, illustrated in Fig. 16, has three formula units per unit cell, with the Nb atoms occupying six 4i Wyckoff positions and the O atoms occupying one 2d and fourteen 4i positions [175]. Similarly to Nb22O54, both Nb12O29 have the Nb atoms displaced from the centre of the NbO6 octahedra, towards the centre of each block. These displacements are typically responsible for a spontaneous polarization and electric polarization in ferroelectrics, but in this case they are arranged in such way that each block has no net electrical polarization, and therefore these materials are considered to be antiferroelectric [173]. These structures are a clear example of how stoichiometry deficiencies can give rise to crystallographic shear planes, as illustrated in Figs. 15 and 16 by the light blue octahedral connected by edge-sharing. Even though the stoichiometry of these materials is near of the Nb2O5, which is a semiconductor, the electrical properties are quite different. Schäfer et al. [57], Janninck and Whitmore [176] and later Marucco [172], investigated on the electrical properties of non-stoichiometric niobium oxides, verifying that even a small change of the O/Nb ratio produced a significant variation of the conductivity, showing a typical n-type conductivity. Schäfer et al. [57] reported a variation from 3  106 to 4  101, and to 3  103 S/cm, by changing the O/Nb ratios from 2.5 to 2.495, and to 2.489, respectively (at room temperature). One should notice that these conductivity variations for different stoichiometries do not refer to metastable phases or a mixture of phases. Instead, the Nb2O5 sub stoichiometries are achieved by having different concentrations of oxygen vacancies, which can give rise to phase changes at very high concentrations, i.e. Nb22O54 and Nb12O29. Janninck and Whitmore [176] studied the electrical conductivity, between 77 and 1273 K, of the HNb2O5 polymorph with more than 20 different stoichiometry deviations, from Nb2O4.9992 to Nb2O4.8568, and recorded values between 103 and 103 S/cm (depending on temperature and composition) in an inert atmosphere. The authors confirmed a dependence of the electrical conductivity with the oxygen partial pressure to the power of 1/6, which is consistent with a defect structure model based essentially on double ionized oxygen vacancies. Additionally, the thermal dependence of the

21

C. Nico et al. / Progress in Materials Science 80 (2016) 1–37

electrical conductivity has shown two distinct regions, one with a positive temperature coefficient at low temperatures, and another with a negative temperature coefficient at higher temperatures. An explanation for this behaviour was proposed by the authors [176] as being due to the thermal ionization of the electrons from the donor centres (positive temperature coefficient), until a saturation regime is reached, i.e. the number of charge carriers remain constant. As the temperature further increases, the decrease of the conductivity is explained by the decrease of the mobility of the charge carriers which was found to be proportional to T0.5, in agreement with an optical mode scattering mechanism. By assuming that the oxygen vacancies are all double ionized at 1000 °C, the magnitude of the carrier mobility was calculated to be 0.218 cm2/(V s), and found to be independent of the studied stoichiometries [176]. Marucco [172] performed a more detailed study on the conductivity of niobium oxides, at high temperatures, in a broad range of stoichiometries between Nb12O29 and Nb2O5, as a function of oxygen pressure and temperature, also observing an increase of conductivity with the decrease of O/Nb ratio. Marucco verified that for an oxygen partial pressure above 107 atm, the electrical conductivity behaviour suggested that the defect structure on these materials was predominantly characterized by single ionized oxygen vacancies. However, for lower pressures (down to 1017 atm) all the stable nonstoichiometric phases have shown a dependency consistent with double ionized oxygen vacancies, in agreement with what Janninck and Whitmore [176] observed. Additionally, Marucco observed pronounced discontinuities of the electrical conductivity in metastable phases (compared to stable phases), which was related to the presence of niobium interstitials. It was suggested that the formation of crystallographic shear planes was enabled by the presence of such Nb interstitials [172]. Later studies [175,177] specifically about Nb12O29, report a typical metallic behaviour regarding the electrical conductivity, where Cava et al. [177] recorded values in the order of 103 S/cm, but without showing superconductivity at least down to 0.25 K. The electrical conduction of amorphous Nb2O5 thin films, at low pressures (1010 atm), was investigated by Jouve [115] who was able to identify different conduction processes depending on the temperature and applied field. While the author never suggests a stoichiometry deficiency in the samples, the results are consistent with the presence of oxygen vacancies in the amorphous Nb2O5 films. Jouve [115] emphasized the role of localized trap levels, below the conduction band, in the electrical properties of these materials. At low fields and at room temperature, typical ohmic conductivity was observed and associated to variable-range hopping. Above 360 K and at low fields, the results were consistent with small-polaron model of nearest-neighbour hopping, which is activated by an optical multiphonon process. As the voltage is increased (at room temperature), space-charged limited current density obeying the Mott–Gurney law was observed. At fields higher than 5  106 V m1, as the space-charge effects becomes negligible, the observed current density was associated to the

Table 2 Band gap energy values reported for niobium oxides. Phase

Eg (eV)

NbO2 (tetragonal)

0.5 0.7 0.88 1.0 1.16 3.4 3.4–5.3 3.4 3.4 3.4 3.4 3.4 3.0 2.55 3.75 3.36–4.15

Amorphous Nb2O5

TT-Nb2O5 T-Nb2O5 H-Nb2O5 B-Nb2O5 M-Nb2O5

Ref. Ceramic pellet Thin films Thin films Thin films Ceramic pellet Thin films Thin films Sol–gel powders Sol–gel powders Sol–gel powders Single-crystal Sol–gel powders Single-crystal Calculated Thin films Quantum dots

[184] [70] [185] [65,71] [186] [105] [106] [89] [89] [89] [50] [89] [50] [70] [107] [98]

22

C. Nico et al. / Progress in Materials Science 80 (2016) 1–37

Table 3 Electrical conductivity and dielectric constants of niobium oxides. Phase Nb NbO NbO2 (tetragonal) NbO2 (rutile) Amorphous Nb2O5

T-Nb2O5 H-Nb2O5

B-Nb2O5 NbOx (2.5 > x > 2.489)

rdc (S cm1) @ 300 K 4

Ref.

6.5  10 4.8  104 104 103 @ 1000 K 1011 1013 1012–1011 3.7  1010 1013–1011 7.6  107 3  106 108–106

[20] [11] [56] [56] [110] [111] [112] [50] [84] [50] [57] [84]

106–103

[57]

e0 @ 300 K

Ref.

10

1 kHz–10 MHz

[55]

41 41–46

Static (dc) Static (dc)

[113–115] [106]

39 60–77 38–160 16–130

100 kHz 100 kHz 100 kHz 100 kHz

[50] [84] [50] [84]

5

Static (dc)

[116]

lowering of the barriers around a trapping centre localized 0.34 eV below the conduction band, according to Jouve’s calculations [115]. The author clearly shows that such behaviour cannot be explained by a Schottky effect, but indeed by a Poole–Frenkel mechanism. Jouve also reported an irreversible switching behaviour for fields higher than 5  107 V m1 coherent with a filamentary conduction process [115], similarly to what happens with NbO2 [78]. In 1991, Cava et al. [177] were the first to report on the magnetic properties of non-stoichiometric niobium oxides. While the Nb2O5 is diamagnetic, Cava et al. verified that the orthorhombic Nb12O29 phase shows an approximated Curie–Weiss behaviour, i.e. paramagnetic, and an antiferromagnetic ordering at 12 K. The fact that orthorhombic Nb12O29 shows, simultaneously, a Curie–Weiss behaviour, antiferromagnetism and metallic conductivity (even below the ordering temperature), which are characteristics of localized and delocalized electrons, led the author to suggest that the structure contained some Nb4+ ions with a localized 4d1 configuration, which give rise to the magnetism, while the other Nb5+ ions at other sites contribute with delocalized electrons, responsible for the conductivity. In 2001, Waldron et al. [178] further extend this explanation to monoclinic Nb12O29, but only in 2004 the same authors [175] experimentally verify for this phase an identical metallic conductivity, allied to a Curie–Weiss behaviour and a antiferromagnetic transition at 12 K, the same ordering temperature as the orthorhombic polymorph. However, the advanced structural study reported by McQueen et al. [173], in 2007, shows that based on bond valence results, no Nb4+ sites are detected in the orthorhombic Nb12O29 phase (at least at 200 K), which suggests that the magnetic behaviour observed in these materials is not related with the valence electrons of individual Nb atoms, but instead, as the authors suggest, ‘‘[. . .] each magnetic electron may be delocalized across a 4  3 block [. . .]”. Up to date, this question about the magnetic nature of the Nb12O29 phases remains unclear. Moreover, it should be noticed that Cava et al. [177] verified that other non-stoichiometric niobium oxide phases displayed paramagnetic behaviour, but not any kind of ordering, at least down to 2 K. In 2011 Ohsawa et al. [174] reported the synthesis of a Nb12O29 thin films (with a thickness of 120 nm) which have shown a high transmittance in the visible range (50% in the red and 70% in the blue regions), with a refractive index n = 2.2 at 400 nm. The films were deposited on a glass substrate with a transmittance between 90% and 95%. The high transparency, allied to the high electrical conductivity of about 300 S/cm (even after annealing at 1000 °C in vacuum) without the need of doping, makes Nb12O29 a new class of transparent conductive oxides (TCO). Classified by the authors as an ‘‘intrinsically doped d-electron-based TCO” [174], and though there is no evidence if it has a direct or indirect band gap, the Nb12O29 may be a promising material for use in high-performance and lowcost optoelectronic devices. With the description of the niobium oxides, the complexity of this system is evident. It also becomes obvious that an in-depth knowledge about the structures and stoichiometry of these oxides is fundamental for solid scientific research or towards any type of technological application. Table 1

C. Nico et al. / Progress in Materials Science 80 (2016) 1–37

23

Fig. 17. Typical perovskite structure of an alkali niobate [21].

enumerates all the stable niobium oxides phases reported to date, while Tables 2 and 3 summarize reported values of band gap energy, electrical conductivity and dielectric constant for these materials. 3. Niobates 3.1. Alkali niobates Lithium Niobate (LiNbO3), Sodium Niobate (NaNbO3) and Potassium Niobate (KNbO3) are known as the alkali niobates. These oxides, structural classified as perovskites, have been reported by many of their interesting properties such as piezoelectricity, pyroelectricity, electro-optic and nonlinear optical behaviour [187,188]. More particularly, LiNbO3, possibly the most studied of the alkali niobates, is a well-known transparent semiconductor (with a band gap of 3.7 eV [189]) with an extremely technological importance. It presents large pyroelectric, piezoelectric, acousto-optic, nonlinear and electro-optic coefficients, and it is mainly used, in solid-state based optical devices, as optical modulators, for Q-switching, for second harmonic generation (SHG), optical filters, waveguides or acoustic-wave transducers, either in bulk or nanostructured forms [187,190]. Due to its wide band gap, LiNbO3 has also been reported as an interesting host for achieving optically active lanthanides via suitable doping processes [191,192]. Similarly, KNbO3 is reported for having large nonlinear optical coefficients which makes it suitable for SHG, frequency mixing, or optical parametric oscillators [190]. Moreover, KNbO3 and also NaNbO3 are pointed as promising alternatives to the currently used piezoelectrics ceramics, mainly based on lead zirconate titanate (PZT). The need of finding a substitute for PZT is essentially related with environmental and health concerns, since it presents a weight percentage of lead 600 times greater than the maximum limit imposed by governmental regulations [188]. Thus, the alkali niobate (K,Na)NbO3, typically abbreviated as KNN, is probably the top candidate to replace PZT, with typical reported piezoelectric constant d33 around 80 pC/N for an optimal K/Na ratio 1 [188]. The most common approach for property optimization of KNN involves forming solid solutions with other materials such as LiNbO3, LiTaO3, LiSbO3, BaTiO3, CaTiO3, (Bi0.5Na0.5)TiO3, or a combination of these, resulting in d33 constants in the order of 200 and 300 pC/N [188]. The crystalline structure of KNN ceramics is not simple. While the pervoskite subcell of this material has a monoclinic symmetry (illustrated in Fig. 17), the KNN crystalline structure is orthorhombic with a space group C14 2v . The structure of K0.5Na0.5NbO3

24

C. Nico et al. / Progress in Materials Science 80 (2016) 1–37

Fig. 18. Typical orthorhombic structure of most columbite niobates [21].

is not yet refined and does not have a standard JCPDS–ICDD file, and therefore compromises a correct crystal indexing literature. Furthermore, the KNN ceramics have different polymorphs depending on the temperature and on typically used dopants such as Li, Ta, Sb or others [188]. Despite the fact that these alkali niobates have been presenting extremely good piezoelectric properties, their application is still limited by difficulties in their synthesis. In fact, one of the current biggest challenges in the study of alkali niobates is the control and reproducibility of their properties, which are typically associated to the synthesis methods and sintering processes. It is known that small variations of the sintering temperature may easily result in deteriorated performance [193]. Additionally, there are two types of mechanisms that were proposed to justify the piezoelectric property enhancement – polymorphism phase transition (PPT) and morphotropic phase boundary (MPB). Summarizing, it becomes obvious that the investigation and control of the phase and crystalline structures of these materials is extremely important [188]. One of the most common precursors for the synthesis of niobates is some sort of niobium oxide, particularly Nb2O5. However, as explained above, due to the ‘‘memory of solids” the original synthesis method of the niobium oxide can be determinant in the properties of the final product, thus affecting the properties of the niobate in different ways. Furthermore, it is known [8] that there is a direct structural relationship between the form of Nb2O5 and the related ternary oxides. Moreover, Kuznetsova et al. [194] ‘‘[. . .] have demonstrated that the polymorphous phase state of Nb2O5 significantly influences the characteristics of the alkali metal niobate solid solution powders and determines the extreme variation of the properties of synthesized ceramics, which must be taken into account in development of the Nbcontaining ferroelectric piezoceramics”. Hrešcˇaka et al. [195] have recently published a work on this issue, alerting for the need of a good identification of the Nb2O5 polymorph when reporting the synthesis of alkali niobates. This is why it is believed that the experience and knowledge acquired, while exploring the complex system of niobium oxides, may constitute an important basis in the research of different niobates. Additionally, there have been also recent works regarding multiple hetero-nanostructures based on alkali niobates. This new class of materials are very interesting for the possibility of effectively coupling different properties. Due to their properties and multifunctional character, heteronanostructured systems can be more advantageous than single component systems. Particularly, Yan et al. [196] reported the synthesis of multiple NaNbO3 nanoplates inside hollow Nb2O5 nanotubes. The authors classified these structures as a novel class of multiple ferroelectric/semiconductor heterostructures as they exhibit distinct ferroelectric switching due to the presence of the strain states at the NaNbO3/Nb2O5 interface.

C. Nico et al. / Progress in Materials Science 80 (2016) 1–37

25

3.2. Columbite niobates The columbite niobates (which have their name from the same crystalline structure of the columbite mineral), have the general formula M2+Nb2O6 where M2+ is a divalent cation of Mg or Ca or of a transition metal such as Cu, Cd, Zn, Ni, Mn, Co or Fe, and have been reported especially as microwave ceramics [197]. These materials crystallize in a orthorhombic structure (space group D14 2h) characterized by zigzag chains linked by corner sharing, each chain formed by MO6 and NbO6 edge-sharing octahedra, as illustrated by Fig. 18 [197]. The exceptions to this type of structure, in columbite niobates, are those in which the divalent ion has a ionic radius higher than 1.0 Å, such as Ba2+ and Sr2+, promoting the crystallization in other types of orthorhombic structure [197]. Microwave ceramics are characterized as being dielectric resonators with a frequency in the microwave region (500 MHz to 10 GHz), where a high dielectric constant and low dielectric losses are determinant characteristics. These materials find an important application in wireless communications which is, and has been, a highly demanding technology. Therefore, it is natural that there is a constant search for less expensive and high performance dielectric ceramics, especially because the materials currently used are tantalum-based complex perovskites, such as BaZn0.33Ta0.67O3 (BZT) or BaMg0.33Ta0.67O3 (BMT) [197]. As an alternative, there are niobium-based perovskites, such as BaZn0.33Nb0.67O3 (BZN) or Ba(Zn/ Co)0.33Nb0.67O3 (BZCN), which are less expensive, but require high sintering temperatures (c.a. 1600 °C) and times, and while the relative permittivity is superior, the overall dielectric properties are not yet competitive enough. The reported interest in columbite niobates (especially ZnNb2O6) as microwave ceramics is related to their dielectric properties, similar to BZN or BZCN, and significantly better than columbite tantalates (M2+Ta2O6). In contrast with the tantalate and niobate perovskites, these materials do not require sintering at such high temperatures, typically between only 1100 and 1200 °C, and also have just a superficial degree of ordering [197]. Additionally, because of the simpler chemistry of binary compounds, these materials require a less complex processing than the perovskites. Nevertheless, it can be concluded from the literature that many columbite niobates are not synthesized simply by one step. Difficulty in forming a single-phase samples or requiring two or more sintering steps are commonly reported. Furthermore, the occurrence of point defects in the lattice, dislocations, grain boundaries and porosity are pointed as frequent problems that compromise and justify the variation of the dielectric properties of these materials. It should be noticed that, also in these materials, non-stoichiometric phases are commonly produced as well. It was for instance reported that a slightly reduced manganese niobate (MnNb2O6x) presents a much higher electrical conductivity, assigned to the production of delocalized electrons upon a small degree of Nb reduction [197]. Generally, it was observed that a small stoichiometry variation for columbite niobates produce huge variations in their dielectric properties, thus compromising their application [197]. It seems possible the cause of these type of observations may be closely related to the analogous behaviour of niobium oxides, and also related with the phase of niobium pentoxide used as precursor for the niobates, but further investigation is nevertheless necessary. Besides the application as dielectric ceramics, columbite niobates have also been reported by their magnetic properties, which in general present antiferromagnetic ordering below temperatures 4–7 K with special interest on CoNb2O6 that shows some unusual magnetic ordering due to geometrical frustration [197]. The reported NiNb2O6 photocatalytic properties, for water hydrolysis, are quite interesting but it was shown that this material was easily corroded, thus compromising its application. On the other hand, Zn and Mn niobates were also reported to have interesting photocatalytic properties for isobutene oxidation, presenting a good chemical stability and selectivity. The optical properties of both pure and rare-earth-doped columbite niobates have been reported in literature [197]. The pure columbite niobates can be considered wide band gap semiconductors with reported values ranging from 2.2 to 3.9 eV, depending on the M2+ ion [198–201], and have shown (at least for Ca, Zn, Cd and Mg niobates) intrinsic blue luminescence with broad bands centred c.a. at 430– 470 nm [197]. Hence, these materials were classified as self-activated phosphors. It was reported that the luminescence efficiency decreased with decreasing size of the divalent ion, i.e. Ca niobate was found to have the most efficient emission. Particularly for Ca niobate, it was found that the

26

C. Nico et al. / Progress in Materials Science 80 (2016) 1–37

464.5 nm emission was preferentially excited with 200–260 nm wavelength radiation [197]. Partial substitution of Ca by other elements, such as Zn and Cd, were reported to enhance the blue luminescence intensity, which was assigned to the participation of the 4d electrons in the charge transfer within the NbO6 octahedra [197]. Ca niobate is transparent in a wide spectral range (300 nm to 5.5 lm), it has relatively low phonon cut-off energy (900 cm1) and a high refractive index is commonly reported (between 2.07 and 2.2) although it is not specified at which wavelengths [197,202,203]. Additionally, single crystals of Ca niobate are rather easy to grow by laser-heated pedestal growth technique, which allied to these properties makes it a preferable host for rare-earth doping, aiming their application as optical devices [197]. Thus, there are works reporting laser action in Ca niobate doped with Nd, Ho, Pr and Tm, or even co-doped with Eu and Ti which was proposed as a low cost lamp phosphor [197]. Europium was found to be optically active in doped Cd niobate (band gap between 3.35 and 3.5 eV) [204,205], showing higher intensity, when excited and detected under the same experimental conditions, compared with the Ca niobate host and with strong evidences of multiple Eu3+ emitting sites [197]. Single crystals of CaNb2O6 doped with neodymium have shown stimulated emission at 1064 nm, due to intraionic Nd3+ transitions, with a lifetime of only 145 ls; the samples doped with erbium showed a pink colouration and an intense emission at 1550 nm at room temperature (due to transition between the first excited multiplet and the ground one of the Er3+), and similarly, those doped with praseodymium showed a sharp and intense band centred at 610 nm due to the intraionic emission of the Pr3+ [197]. Additionally, CaNb2O6 was reported to have a strong blue mechanoluminescence, observed during grinding [197]. On the other hand, nanoparticles of Ni niobate doped with Dy, prepared by a sol–gel combustion method, did not show any characteristic luminescence of the lanthanide ion in its trivalent charge state, but it was found that the intrinsic blue emission of the host was enhanced, which was assigned to the energy transfer from the Dy3+ ion to the NbO6 octahedra. This behaviour was observed also for Zn niobate doped with Dy, but it was also found that for >5 mol% Dy there was an opposite behaviour assigned to concentration quenching effect [197]. There has also been some works reporting on another type of niobate perovskite based on rareearths with very interesting optical properties, particularly the CaRENb2O10 (where RE is a trivalent rare-earth ion) [206]. These materials are classified as triple-layered perovskites, as they are built by layers of corner-sharing octahedra. In a very recent work (2013), Qin et al. [206] report the synthesis and characterization of these ‘‘novel luminescent materials” with an estimated band gap of 4.42 eV. With 266 nm (4.7 eV) laser excitation, and so above the material band gap, the authors observed a broad emission band which was assigned to the NbO6 octahedra. Depending on the lanthanide, the intrinsic NbO6 broad band varied its spectral position between 400 and 500 nm, from blue to greenish blue, which is preferentially excited at 270 nm. The characteristic intra-4f emissions of the trivalent rare-earth ions were also observed under 266 nm excitation, namely for the Sm3+, Er3+, Dy3+ and Eu3+ ions, which shows that there was a charge transfer mechanism from host to the lanthanide. It is noted

Fig. 19. Structure of the monoclinic phase of RENbO4 [21].

C. Nico et al. / Progress in Materials Science 80 (2016) 1–37

27

Fig. 20. Structure of the tetragonal phase of RENbO4 [21].

that these materials show a high critical point for luminescence concentration quenching effects. This, and the high luminescence intensity of the lanthanides characteristic emissions at room temperature, lead the authors to suggest these materials to be used as phosphors in near-ultraviolet or blue light emitting diodes (LEDs), especially the CaEuNb3O10 [206]. 3.3. Rare-earths orthoniobates There are works reporting rare-earth orthoniobates (RENbO4), as well as rare-earth orthotantalates (RETaO4), from decade of 1960 [207–211], but until recently these systems were not a research topic which has received much attention from the scientific community. The RENbO4 are known to crystallize in a fergusonite-type structure, i.e. in a monoclinic structure (space group C62h), undergoing a pure and reversible ferroelastic phase transformation to a scheelitetype structure, with a tetragonal system lattice (space group C64h), only stable at high temperatures (>700 °C) [211–213]. In the monoclinic phase the Nb and RE atoms occupy the Wyckoff positions 4e while the O atoms occupy the 8f position, which structure is illustrated by Fig. 19 where it is possible to identify the NbO6 octahedra linked by edge-sharing forming chains along the c axis. The tetragonal phase is not built on NbO6 octahedra, but instead on unlinked NbO4 tetrahedra, as illustrated in Fig. 20, with the RE, Nb and O atoms occupying, individually and respectively, the 4b, 4a and 16f Wyckoff positions. The lattice parameters of the RENbO4 compounds will obviously depend on the lanthanide ion. One of the most detailed structural studies on rare-earths orthoniobates, which covers all the lanthanide ions (from La to Lu, except Pm), is a work reported by Siqueira et al. [212]. These authors report the synthesis of rare-earth niobates by solid-state reaction (by mixing niobium pentoxide and rare-earths oxides), and their characterization by XRD and Raman spectroscopy. It was found that only for heat treatment temperatures above 1150 °C, the formation of secondary phases could be avoided. The Raman analysis on the RENbO4 samples revealed very complex spectra, with 18 active modes, which was found to be in complete agreement with group theory calculations [212]. Furthermore, it was found that some of the vibrational modes follow a decrease of energy with the increase of the lanthanide ionic radii (which is a natural consequence of the unit cell expansion), while other modes followed the inverse trend (justified by a mass effect) [212]. Some of the Raman bands in the range of 316–343 cm1 showed a tendency to split for lower lanthanide ionic radii, which was assigned to a higher monoclinic distortion. From all the RENbO4 samples, the Raman spectra of the Ce and La niobates were classified as an exception to the general trend because they presented a significant broadening and down-shifting of their modes energy. This was linked to the proximity of the transition temperature from the monoclinic to the tetragonal phase, since it decreases from the Lu to

28

C. Nico et al. / Progress in Materials Science 80 (2016) 1–37

the La ion. The authors additionally refer the identification of 15 active modes by Fourier-transform infrared spectroscopy (FTIR) [212]. Rare-earth tantalates are known to have very similar properties and structures [212,214,215]. Still, besides the fergusonite-type structure, recent studies report on the identification of two additional different monoclinic polymorphs, of rare-earth orthotantalates stable at room temperature, depending on the lanthanide ion and on the synthesis/processing conditions [214,216]. Such polymorphs are mainly characterized by different arrangement and linkage of distorted TaO6 octahedra, presenting different physical properties. The same was not observed for orthoniobates [212], but given the obvious similarities between these two systems, it would not be surprising that such polymorphs also exist for RENbO4, and for that reason further research is necessary to clarify this point. While the synthesis of RENbO4 crystals has been previously reported by using Czochralski or flux growth methods [211], only in 2004 Octaviano et al. [217] report the synthesis of single crystals by a floating zone technique. The authors have grown single crystals of La, Ce, Pr, Sm, Eu, Dy Ho, and Er orthoniobates, and a few more doped with Er3+, and mentioned the particular difficulty in controlling the crystal diameter during the growth. It was verified by X-ray Laue diffraction analysis that the crystals tend to grow along the [1 1 0] direction. Octaviano et al. [217] mainly focused on the Er-doped LaNbO4 material which showed very interesting optical properties, with high absorption coefficient and a strong luminescence of Er3+ when pumped with a green 514 nm laser line at 300 K. Recently, Graça et al. [218] report on the synthesis of EuNbO4 crystals, grown by laser floating zone, and their optical and dielectric properties. The luminescence analysis of the EuNbO4 crystals revealed intense red luminescence at room temperature with ultraviolet excitation, and the existence of multiple Eu-related optical centres whose emission is observed at low temperatures some of which are quenched at c.a. 50 K. Other authors, Xiao and Yan [219], report on the synthesis, by a modified insitu chemical co-precipitation method, and the luminescence of different doped rare-earth orthoniobates. Very recently, from the end of 2013, a patent registered by Sandia National Laboratories [220] claims on the production of rare-earth niobates and tantalates, by an hydrothermal technique, aiming their application as nanophosphors to lighting and display technologies, which testifies the interesting optical properties of these materials. Additionally, rare-earth orthoniobates have been reported to have a characteristic ferroelasticity attributed to the formation of domain walls during phase transitions [213,217]. The microwave dielectric properties and effects of such ferroelasticity on these properties were investigated by Kim et al. [213]. The authors estimated the mean static dielectric constants of several orthoniobates using the Clausius–Mossotti relation. The experimental dielectric constants of the RENbO4 samples, however, were significantly lower: 20, almost half than the estimated values. This was related to a monoclinic distortion which has its origin in the phase transformation from the tetragonal to the monoclinic structure, resulting in ferroelastic domain structure that resembles twinning, with two orientation states identical in structure but different in orientation [213]. Despite the low dielectric constants measured by Kim et al. [213], the RENbO4 samples presented excellent quality factors, ranging between 33,000 and 56,600 GHz, depending on the lanthanide, and also high temperature coefficients of resonant frequency (sf) Particularly, LaNbO4 presented a positive sf of 9 ppm/K. The transition temperature of several RENbO4, listed by the authors, was found to be tunable by mixing two different lanthanides, thus allowing to control the strain induced by the lattice distortion, and therefore the dielectric properties. The high-temperature stability of the resonant frequency and quality factor, make the rare-earth orthoniobates interesting ceramics for microwave applications [213]. In other frequency range, from 100 Hz to 2 MHz, Graça et al. [218] verified that, for EuNbO4 crystalline fibres, the dielectric constant depends on the crystal growth speed, being highest (e0 ffi 40) for the slowest pulling rate. Furthermore, it was found that from 80 to 400 K the dielectric constant of the EuNbO4 samples remained constant in a remarkable way, which may be interesting for dielectric applications which require stable operation in a wide temperature range. Probably, the major interest in the rare-earth orthoniobates is related to the reported proton conductivity, thus having the potential application in sensors and fuel cells electrolytes [215]. The first work reporting on the protonic conduction on these materials is from 2006 by Haugsrud and Norby [215], and though the authors have studied both Ca and Sr-doped rare-earth orthoniobates (and also orthotantalates), the main focus was on proton conductivity and solubility of protonic defects in wet

C. Nico et al. / Progress in Materials Science 80 (2016) 1–37

29

H2 atmospheres of Ca-doped RENbO4 samples. The authors clearly identified a pure protonic conductivity for these samples, increasing with the temperature between 400 and 1200 °C. Generally, between 400 and 600 °C it was possible to notice a break in the slope of the conductivity, which was associated with the phase transformation from monoclinic to tetragonal. The Ca-doped LaNbO4 material presented the highest conductivity, reaching 103 S/cm. The analogous orthotantalates were found to behave similarly to the niobates but with a protonic conductivity almost half an order of magnitude lower. Additionally, since their phase transition is much higher, the tantalates remained with the monoclinic structure up to 1200 °C. The authors refer that ‘‘the tetragonal polymorph of the niobates, in general, shows more exothermic (more negative) hydration enthalpies and is thus more easily hydrate than the monoclinic polymorph of the tantalates”. Still, since the phase transformation may lead to thermal expansion, Haugsrud and Norby [215] suggest a partial substitution of Nb by Ta in order to minimize this effect. The fact that these materials are the oxides (without Ba or Sr as the main components) with the highest reported protonic conductivities, make them particularly interesting for fuel-cell electrolytes and humidity sensors, as aforementioned [215]. It is therefore no surprise that, since this work, there have been many other reported studies on the protonic conductivity of these materials, especially on LaNbO4 [221–229]. It is also known that some RENbO4 present antiferromagnetic ordering at temperatures below 4 K, while most of them follow a Curie–Weiss law [210]. Additionally, in 1998, Baughman et al. [230] report a large negative Poisson’s ratio (in at least one direction) for LaNbO4, thus being classified as an auxetic material. This type of property makes these materials interesting for high pressure applications as, for instance, optical line systems installed at deep sea level.

4. Summary and prospects The available literature on niobium oxides reveals that this is a complex metal oxide system, with many phases and polymorphs, and many works reporting contradictory or inconsistent information. Therefore, it should be emphasized that the review of the physical properties and state of art of niobium oxides, is an important and solid groundwork for future research on these materials. Moreover, like most electron correlated materials, it was shown that this metal oxide system can lead to many different and exotic properties, making it a very versatile group of materials (but difficult to work with). More specifically, stoichiometry problems are one of the most common complications in niobium oxides and in many different niobates, where NbO6 octahedra are typically the basic structural units. Niobium oxides have been pointed as an alternative to tantalum for the production of solid electrolyte capacitors, but the diffusion of oxygen easily promoted by the application of a voltage has been their main limitation [49]. Conversely, these oxides have been showing great potentiality in many other technological applications, most of them based on one important matter that can lead to different properties: oxygen vacancies. This is the case of the reported photochromic properties of niobium oxides [231] and the synthesis of Nb12O29 thin films to be used as a transparent conductive oxide (TCO) [174]. Also, it should be emphasized one of the most interesting and hot topics to which niobium oxides are candidates: the memristors. In the case of niobium oxides, what may be a problem for the application in the solid electrolytic capacitors, i.e. the oxygen diffusion and variation of conductivity, can be the key factor to achieve a fully operational memristor. There are some, but still few, works that have been reported regarding resistive switching in niobium oxides (and other metal oxides) [4,77,155–167,232–235], and it seems clear that following this research topic will lead the near future reports on these materials. The synthesis of niobium oxides thin films and nanostructures, and their electrical and optical characterization is encouraged for future work. While the most common experimental techniques used to characterize memristive properties are based on cyclic voltammetry [235–238], i.e. recording hysteretic current-voltage profiles, Messerschmitt et al. [235] have done an interesting work showing that chronoamperometry and also bias-dependent resistive measurements can be powerful methods to understand the defect kinetics that give rise to the resistive switching behaviour.

30

C. Nico et al. / Progress in Materials Science 80 (2016) 1–37

The fact that Nb12O29 can be a TCO, can enable its application as an interface layer in organic light emitting diodes (OLEDs) to facilitate the charge injection and extraction to and from the device [239,240]. Also in this context and inspired in reported works [126] regarding dye-sensitized solar cells using mesoporous Nb2O5 tubular structures as an alternative to TiO2, it would be interesting to try using Nb12O29 instead (transparent and conductive) by performing a thermal treatment in a reducing atmosphere to the Nb2O5 mesoporous structure. Therefore, future studies on niobium oxides focused in understanding and controlling the phase transformations, oxidations and oxygen vacancies are crucial to take advantage of the broad range of properties. Not only fundamental characterization is necessary, but also the synthesis and fabrication of micro and nanostructures are worth to be further explored. Additionally, the research and development of different types of niobates, particularly alkali niobates, is often limited by the difficulties on the synthesis and phase stability. Many of such problems are related to niobium oxygen system and therefore it is believed that a good understanding of the fundamental properties of the niobium oxides, presented and discussed in this work, are an important contribution to better understand such niobates. Acknowledgements This work is funded by FEDER funds through the COMPETE 2020 Programme and National Funds through FCT – Portuguese Foundation for Science and Technology under the project UID/ CTM/50025/2013 and RECI/FIS-NAN/0183/2012 (FCOMP-01-0124-FEDER-027494). C. Nico thanks to FCT for his PhD grant SFRH/BD/68295/2010. References [1] Brauer G. Die oxyde des niobs. Z Für Anorg Allg Chem 1941;248:1–31. [2] Schwartz N, Gresh M, Karlik S. Niobium solid electrolytic capacitors. J Electrochem Soc 1961;108:750–8. http://dx.doi.org/ 10.1149/1.2428210. [3] Ling HW, Kolski TL. Niobium solid electrolyte capacitors. J Electrochem Soc 1962;109:69–70. http://dx.doi.org/10.1149/ 1.2425333. [4] Hiatt WR, Hickmott TW. Bistable switching in niobium oxide diodes. Appl Phys Lett 1965;6:106. http://dx.doi.org/ 10.1063/1.1754187. [5] Cox B, Johnston T. The oxidation and corrosion of niobium (columbium). Trans AIME 1963;227. [6] Gatehouse BM, Wadsley AD. The crystal structure of the high temperature form of niobium pentoxide. Acta Crystallogr 1964;17:1545–54. http://dx.doi.org/10.1107/S0365110X6400384X. [7] Bowman AL, Wallace TC, Yarnell JL, Wenzel RG. The crystal structure of niobium monoxide. Acta Crystallogr 1966;21:843. http://dx.doi.org/10.1107/S0365110X66004043. [8] Schäfer H, Gruehn R, Schulte F. The modifications of niobium pentoxide. Angew Chem Int Ed Engl 1966;5:40–52. http:// dx.doi.org/10.1002/anie.196600401. [9] Fromm E. Thermodynamische beschreibung der festen lösung von kohlenstoff, stickstoff und sauerstoff in niob und tantal. J Common Met 1968;14:113–25. [10] Allpress JG, Sanders JV, Wadsley AD. Electron microscopy of high-temperature Nb2O5 and related phases. Phys Status Solidi B 1968;25:541–50. http://dx.doi.org/10.1002/pssb.19680250206. [11] Pollard Jr ER. Electronic properties of niobium monoxide. Massachusetts Institute of Technology; 1968. [12] Bach D. EELS investigations of stoichiometric niobium oxides and niobium-based capacitors. Universität Karlsruhe, Fakultät für Physik; 2009. [13] Bach D, Störmer H, Schneider R, Gerthsen D, Verbeeck J. EELS investigations of different niobium oxide phases. Microsc Microanal 2006;12:416–23. http://dx.doi.org/10.1017/S1431927606060521. [14] Bach D, Störmer H, Schneider R, Gerthsen D. Quantitative EELS analysis of niobium oxides. Microsc Microanal 2007;13. http://dx.doi.org/10.1017/S143192760708195. [15] Bach D, Störmer H, Schneider R, Gerthsen D, Sigle W. EELS investigations of reference niobium oxides and anodically grown niobium oxide layers. Microsc Microanal 2007;13:1274–5. http://dx.doi.org/10.1017/S143192760707359X. [16] Bach D, Schneider R, Gerthsen D, Verbeeck J, Sigle W. EELS of niobium and stoichiometric niobium-oxide phases—part I: plasmon and near-edges fine structure. Microsc Microanal 2009;15:505. http://dx.doi.org/10.1017/S143192760999105X. [17] Bach D, Schneider R, Gerthsen D. EELS of niobium and stoichiometric niobium-oxide phases—part II: quantification. Microsc Microanal 2009;15:524. http://dx.doi.org/10.1017/S1431927609991061. [18] Elliott RP. Columbium–oxygen system. Trans Am Soc Met 1960;52:990–1014. [19] Massalski TB, International ASM. Binary alloy phase diagrams. 2nd ed. Materials Park (OH): ASM Intl; 1990. [20] Lide DR. CRC handbook of chemistry and physics. 85th ed. CRC Press; 2004. [21] Momma K, Izumi F. VESTA 3 for three-dimensional visualization of crystal, volumetric and morphology data. J Appl Crystallogr 2011;44:1272–6. [22] Williams WS. Transition metal carbides, nitrides, and borides for electronic applications. JOM 1997;49:38–42. http://dx. doi.org/10.1007/BF02914655.

C. Nico et al. / Progress in Materials Science 80 (2016) 1–37

31

[23] DeArdo AJ. Niobium in modern steels. Int Mater Rev 2003;48:371–402. http://dx.doi.org/10.1179/095066003225008833. [24] Heisterkamp F, Carneiro T. Niobium: future possibilities – technology and the market place. In: Niobium sci technol proc int symp, niobium, Orlando, Florida, USA; 2001. [25] Kommel L, Kimmari E, Saarna M, Viljus M. Processing and properties of bulk ultrafine-grained pure niobium. J Mater Sci 2013;48:4723–9. http://dx.doi.org/10.1007/s10853-013-7210-3. [26] Halbritter J. On the oxidation and on the superconductivity of niobium. Appl Phys A 1987;43:1–28. http://dx.doi.org/ 10.1007/BF00615201. [27] Periasamy P, Guthrey HL, Abdulagatov AI, Ndione PF, Berry JJ, Ginley DS, et al. Metal–insulator–metal diodes: role of the insulator layer on the rectification performance. Adv Mater 2013;25:1301–8. http://dx.doi.org/10.1002/adma.201203075. [28] Matsuno H, Yokoyama A, Watari F, Uo M, Kawasaki T. Biocompatibility and osteogenesis of refractory metal implants, titanium, hafnium, niobium, tantalum and rhenium. Biomaterials 2001;22:1253–62. http://dx.doi.org/10.1016/S01429612(00)00275-1. [29] Lindau I, Spicer WE. Oxidation of Nb as studied by the UV-photoemission technique. J Appl Phys 1974;45:3720–5. http:// dx.doi.org/10.1063/1.1663849. [30] Bouchiat V, Faucher M, Thirion C, Wernsdorfer W, Fournier T, Pannetier B. Josephson junctions and superconducting quantum interference devices made by local oxidation of niobium ultrathin films. Appl Phys Lett 2001;79:123–5. http:// dx.doi.org/10.1063/1.1382626. [31] Everitt CWF, DeBra DB, Parkinson BW, Turneaure JP, Conklin JW, Heifetz MI, et al. Gravity probe B: final results of a space experiment to test general relativity. Phys Rev Lett 2011;106:221101. http://dx.doi.org/10.1103/ PhysRevLett.106.221101. [32] Seeber B, editor. Handbook of applied superconductivity, vol. 2. Bristol: CRC Press; 1998. [33] Gupta CK, Suri AK. Extractive metallurgy of niobium. CRC Press; 1993. [34] Shamrai VF, Blagoveshchenski YV, Gordeev AS, Mitin AV, Drobinova IA. Structural states and electrical conductivity of oxidized niobium nanopowders. Russ Metall Met 2007;2007:322–6. http://dx.doi.org/10.1134/S0036029507040106. [35] Grundner M, Halbritter J. XPS and AES studies on oxide growth and oxide coatings on niobium. J Appl Phys 1980;51:397–405. [36] Bai XD, Zhu DH, Liu BX. Anodic and air oxidation of niobium studied by ion beam analysis with implanted Xe marker. Nucl Instrum Methods Phys Res Sect B Beam Interact Mater Atoms 1996;114:64–9. [37] d’Acapito F, Mobilio S, Cikmacs P, Merlo V, Davoli I. Temperature modification of the Nb oxidation at the Nb/Al interface studied by reflEXAFS. Surf Sci 2000;468:77–84. [38] Wu AT. Investigation of oxide layer structure on niobium surface using a secondary ion mass spectrometry. Phys C Supercond 2006;441:79–82. [39] Batchelor AD, Leonard DN, Russell PE, Stevie FA, Griffis DP, Myneni GR. TEM and SIMS analysis of (1 0 0), (1 1 0), and (1 1 1) single crystal niobium. Single cryst – large grain niobium technol int niobium workshop, vol. 927. AIP Publishing; 2007. p. 72–83. [40] Yoon KE, Seidman DN, Bauer P, Boffo C, Antoine C. Atomic-scale chemical-analyses of niobium for superconducting radiofrequency cavities. IEEE Trans Appl Supercond 2007;17:1314–7. http://dx.doi.org/10.1109/TASC.2007.898059. [41] Kovacs K, Kiss G, Stenzel M, Zillgen H. Anodic oxidation of niobium sheets and porous bodies. J Electrochem Soc 2003;150:B361–6. http://dx.doi.org/10.1149/1.1585052. [42] Franchy R, Bartke TU, Gassmann P. The interaction of oxygen with Nb(1 1 0) at 300, 80 and 20 K. Surf Sci 1996;366:60–70. http://dx.doi.org/10.1016/0039-6028(96)00781-9. [43] Delheusy M, Stierle A, Kasper N, Kurta RP, Vlad A, Dosch H, et al. X-ray investigation of subsurface interstitial oxygen at Nb/oxide interfaces. Appl Phys Lett 2008;92:101911. http://dx.doi.org/10.1063/1.2889474. [44] Kurmaev EZ, Moewes A, Bureev OG, Nekrasov IA, Cherkashenko VM, Korotin MA, et al. Electronic structure of niobium oxides. J Alloys Compd 2002;347:213–8. http://dx.doi.org/10.1016/S0925-8388(02)00765-X. [45] Hulm JK, Jones CK, Hein RA, Gibson JW. Superconductivity in the TiO and NbO systems. J Low Temp Phys 1972;7:291–307. http://dx.doi.org/10.1007/BF00660068. [46] Qiu Y, Smyth D, Kimmel J. The stabilization of niobium-based solid electrolyte capacitors. Act Passive Electron Compon 2002;25:201–9. [47] Karnik T. Ceramic powder for use in forming an anode of an electrolytic capacitor. GB 2454049 (A); 2009. [48] Karnik T. Doped ceramic powder for use in forming capacitor anodes. Patent US 7760487 B2; 2010. [49] Nico C, Soares MRN, Rodrigues J, Matos M, Monteiro R, Graça MPF, et al. Sintered NbO powders for electronic device applications. J Phys Chem C 2011;115:4879–86. http://dx.doi.org/10.1021/jp110672u. [50] Soares MRN, Leite S, Nico C, Peres M, Fernandes AJS, Graça MPF, et al. Effect of processing method on physical properties of Nb2O5. J Eur Ceram Soc 2011;31:501–6. http://dx.doi.org/10.1016/j.jeurceramsoc.2010.10.024. [51] Forghany SKE, Anderson JS. Reduction and polymorphic transformation of B-Nb2O5. J Chem Soc, Dalton Trans 1981:255–61. http://dx.doi.org/10.1039/DT9810000255. [52] Cheetham AK, Rao CNR. A neutron diffraction study of niobium dioxide. Acta Crystallogr B 1976;32:1579–80. http://dx. doi.org/10.1107/S0567740876005876. [53] Rimai DS, Sladek RJ. Pressure dependences of the elastic constants of semiconducting NbO2 at 296 K. Phys Rev B 1978;18:2807. http://dx.doi.org/10.1103/PhysRevB.18.2807. [54] Gervais F, Kress W. Lattice dynamics of oxides with rutile structure and instabilities at the metal–semiconductor phase transitions of NbO2 and VO2. Phys Rev B 1985;31:4809. http://dx.doi.org/10.1103/PhysRevB.31.4809. [55] Zhao Y, Zhang Z, Lin Y. Optical and dielectric properties of a nanostructured NbO2 thin film prepared by thermal oxidation. J Phys Appl Phys 2004;37:3392–5. http://dx.doi.org/10.1088/0022-3727/37/24/006. [56] Janninck RF, Whitmore DH. Electrical conductivity and thermoelectric power of niobium dioxide. J Phys Chem Solids 1966;27:1183–7. http://dx.doi.org/10.1016/0022-3697(66)90094-1. [57] Schäfer H, Bergner D, Gruehn R. Beiträge zur Chemie der Elemente Niob und Tantal. LXXI. Die thermodynamische Stabilität der sieben zwischen 2,00 und 2,50 O/Nb existierenden Phasen. Z Für Anorg Allg Chem 1969;365:31–50. http:// dx.doi.org/10.1002/zaac.1969365010.

32

C. Nico et al. / Progress in Materials Science 80 (2016) 1–37

[58] Afanas’ev VV, Stesmans A, Delabie A, Bellenger F, Houssa M, Meuris M. Electronic structure of GeO2-passivated interfaces of (1 0 0)Ge with Al2O3 and HfO2. Appl Phys Lett 2008;92:022109. http://dx.doi.org/10.1063/1.283166. [59] Batzill M, Diebold U. The surface and materials science of tin oxide. Prog Surf Sci 2005;79:47–154. http://dx.doi.org/ 10.1016/j.progsurf.2005.09.002. [60] Pascual J, Camassel J, Mathieu H. Resolved quadrupolar transition in TiO2. Phys Rev Lett 1977;39:1490–3. http://dx.doi. org/10.1103/PhysRevLett.39.1490. [61] Pascual J, Camassel J, Mathieu H. Fine structure in the intrinsic absorption edge of TiO2. Phys Rev B 1978;18:5606–14. http://dx.doi.org/10.1103/PhysRevB.18.5606. [62] Robertson J. Electronic structure of SnO2, GeO2, PbO2, TeO2 and MgF2. J Phys C: Solid State Phys 1979;12:4767. http://dx. doi.org/10.1088/0022-3719/12/22/018. [63] Yang Z, Ko C, Ramanathan S. Oxide electronics utilizing ultrafast metal-insulator transitions. Annu Rev Mater Res 2011;41:337–67. http://dx.doi.org/10.1146/annurev-matsci-062910-100347. [64] O’Hara A, Demkov AA. Nature of the metal-insulator transition in NbO2. Phys Rev B 2015;91:094305. [65] O’Hara A, Nunley TN, Posadas AB, Zollner S, Demkov AA. Electronic and optical properties of NbO2. J Appl Phys 2014;116:213705. [66] Wong FJ, Hong N, Ramanathan S. Orbital splitting and optical conductivity of the insulating state of NbO2. Phys Rev B 2014;90:115135. [67] Vezzoli GC. Electrical properties of NbO2 and Nb2O5 at elevated temperature in air and flowing argon. Phys Rev B 1982;26:3954. http://dx.doi.org/10.1103/PhysRevB.26.3954. [68] Passier F, Wouters Y, Galerie A, Caillet M. Thermal oxidation of metallic niobium by water vapor. Oxid Met 2001;55:153–63. [69] Terao N. Structures des oxydes de niobium. Jpn J Appl Phys 1963;2:156–74. http://dx.doi.org/10.1143/JJAP.2.156. [70] Weibin Z, Weidong W, Xueming W, Xinlu C, Dawei Y, Changle S, et al. The investigation of NbO2 and Nb2O5 electronic structure by XPS, UPS and first principles methods. Surf Interface Anal 2013;45:1206–10. http://dx.doi.org/10.1002/ sia.5253. [71] Posadas AB, O’Hara A, Rangan S, Bartynski RA, Demkov AA. Band gap of epitaxial in-plane-dimerized single-phase NbO2 films. Appl Phys Lett 2014;104:092901. http://dx.doi.org/10.1063/1.4867085. [72] Wang Y, Comes RB, Kittiwatanakul S, Wolf SA, Lu J. Epitaxial niobium dioxide thin films by reactive-biased target ion beam deposition. J Vac Sci Technol A 2015;33:021516. http://dx.doi.org/10.1116/1.4906143. [73] Kim S, Park J, Woo J, Cho C, Lee W, Shin J, et al. Threshold-switching characteristics of a nanothin-NbO2-layer-based Pt/ NbO2/Pt stack for use in cross-point-type resistive memories. Microelectron Eng 2013;107:33–6. http://dx.doi.org/ 10.1016/j.mee.2013.02.084. [74] Liu X, Nandi SK, Venkatachalam DK, Belay K, Song S, Elliman RG. Reduced threshold current in NbO2 selector by engineering device structure. IEEE Electron Device Lett 2014;35:1055–7. http://dx.doi.org/10.1109/LED.2014.2344105. [75] Joshi T, Senty TR, Borisov P, Bristow AD, Lederman D. Preparation, characterization, and electrical properties of epitaxial NbO2 thin film lateral devices. J Phys Appl Phys 2015;48:335308. http://dx.doi.org/10.1088/0022-3727/48/33/335308. [76] Zhang J, Norris K, Samuels K, Ge N, Zhang M, Park J, et al. Electron energy-loss spectroscopy (EELS) study of NbOx film for resistive memory applications. Microsc Microanal 2015;21:285–6. http://dx.doi.org/10.1017/S1431927615002226. [77] Nandi SK, Liu X, Venkatachalam DK, Elliman RG. Threshold current reduction for the metal–insulator transition in NbO2x-selector devices: the effect of ReRAM integration. J Phys Appl Phys 2015;48:195105. http://dx.doi.org/10.1088/ 0022-3727/48/19/195105. [78] Slesazeck S, Mähne H, Wylezich H, Wachowiak A, Radhakrishnan J, Ascoli A, et al. Physical model of threshold switching in NbO2 based memristors. RSC Adv 2015;5:102318–22. http://dx.doi.org/10.1039/C5RA19300A. [79] Helali S, Abdelghani A, Hafaiedh I, Martelet C, Prodromidis MI, Albanis T, et al. Functionalization of niobium electrodes for the construction of impedimetric biosensors. Mater Sci Eng, C 2008;28:826–30. http://dx.doi.org/10.1016/j. msec.2007.10.078. [80] Lee S, Yoon H, Yoon I, Kim B. Single crystalline NbO2 nanowire synthesis by chemical vapor transport method. Bull Korean Chem Soc 2012;33:839. [81] Sasaki K, Maier J. Re-analysis of defect equilibria and transport parameters in Y2O3-stabilized ZrO2 using EPR and optical relaxation. Solid State Ionics 2000;134:303–21. http://dx.doi.org/10.1016/S0167-2738(00)00766-9. [82] Reznichenko LA, Akhnazarova VV, Shilkina LA, Razumovskaya ON, Dudkina SI. Invar effect in n-Nb2O5, aht-Nb2O5, and LNb2O5. Crystallogr Rep 2009;54:483–91. http://dx.doi.org/10.1134/S1063774509030183. [83] Valencia-Balvín C, Pérez-Walton S, Dalpian GM, Osorio-Guillén JM. First-principles equation of state and phase stability of niobium pentoxide. Comput Mater Sci 2014;81:133–40. http://dx.doi.org/10.1016/j.commatsci.2013.07.032. [84] Graça MPF, Meireles A, Nico C, Valente MA. Nb2O5 nanosize powders prepared by sol–gel – structure, morphology and dielectric properties. J Alloys Compd 2013;553:177–82. http://dx.doi.org/10.1016/j.jallcom.2012.11.128. [85] Rani RA, Zoolfakar AS, O’Mullane AP, Austin MW, Kalantar-Zadeh K. Thin films and nanostructures of niobium pentoxide: fundamental properties, synthesis methods and applications. J Mater Chem A 2014;2:15683. http://dx.doi.org/10.1039/ C4TA02561J. [86] Viet AL, Reddy MV, Jose R, Chowdari BVR, Ramakrishna S. Nanostructured Nb2O5 polymorphs by electrospinning for rechargeable lithium batteries. J Phys Chem C 2010;114:664–71. http://dx.doi.org/10.1021/jp9088589. [87] Le Viet A, Jose R, Reddy MV, Chowdari BVR, Ramakrishna S. Nb2O5 photoelectrodes for dye-sensitized solar cells: choice of the polymorph. J Phys Chem C 2010;114:21795–800. http://dx.doi.org/10.1021/jp106515k. [88] Kumagai N, Koishikawa Y, Komaba S, Koshiba N. Thermodynamics and kinetics of lithium intercalation into Nb2O5 electrodes for a 2 V rechargeable lithium battery. J Electrochem Soc 1999;146:3203–10. http://dx.doi.org/10.1149/ 1.1392455. [89] Brayner R, Bozon-Verduraz F. Niobium pentoxide prepared by soft chemical routes: morphology, structure, defects and quantum size effect. Phys Chem Chem Phys 2003;5:1457–66. [90] Gruehn R. Eine weitere neue Modifikation des niobpentoxids. J Common Met 1966;11:119–26. http://dx.doi.org/10.1016/ 0022-5088(66)90076-2.

C. Nico et al. / Progress in Materials Science 80 (2016) 1–37

33

[91] Hibst H, Gruehn R. Weitere neue Nb2O5-Modifikationen, metastabile Oxidationsprodukte von NbOx-Phasen (2,4 < x < 2,5). Z Für Anorg Allg Chem 1978;442:49–78. [92] Meyer B, Gruehn R. Investigation of metastable niobium oxides by transmission electron microscopy. Comments Inorg Chem 1982;1:361–77. http://dx.doi.org/10.1080/02603598208078104. [93] Stoneham AM, Durham PJ. The ordering of crystallographic shear planes: theory of regular arrays. J Phys Chem Solids 1973;34:2127–35. http://dx.doi.org/10.1016/S0022-3697(73)80060-5. [94] Mrowec S. On the defect structure in nonstoichiometric metal oxides. Ceramurg Int 1978;4:47–58. http://dx.doi.org/ 10.1016/0390-5519(78)90119-9. [95] Anderson JS, Browne JM, Cheetham AK, Dreele RV, Hutchison JL, Lincoln FJ, et al. Point defects and extended defects in niobium oxides. Nature 1973;243:81–3. http://dx.doi.org/10.1038/243081a0. [96] Iijima S, Kimura S, Goto M. High-resolution microscopy of nonstoichiometric Nb22O54 crystals: point defects and structural defects. Acta Crystallogr Sect A 1974;30:251–7. http://dx.doi.org/10.1107/S0567739474000519. [97] Iijima S, Kimura S, Goto M. Direct observation of point defects in Nb12O29 by high-resolution electron microscopy. Acta Crystallogr Sect A 1973;29:632–6. http://dx.doi.org/10.1107/S0567739473001610. [98] Dhawan S, Dhawan T, Vedeshwar AG. Growth of Nb2O5 quantum dots by physical vapor deposition. Mater Lett 2014;126:32–5. http://dx.doi.org/10.1016/j.matlet.2014.03.107. [99] Tamura S. High-pressure phase research on Nb2O5. J Mater Sci 1972;7:298–302. http://dx.doi.org/10.1007/BF00555631. [100] Madelung O, Rössler U, Schulz M, editors. Non-tetrahedrally bonded binary compounds II, vol. 41D. Berlin/ Heidelberg: Springer-Verlag; 2000. [101] Zibrov IP, Filonenko VP, Werner P-E, Marinder B-O, Sundberg M. A new high-pressure modification of Nb2O5. J Solid State Chem 1998;141:205–11. http://dx.doi.org/10.1006/jssc.1998.7954. [102] Jehng JM, Wachs IE. Structural chemistry and Raman spectra of niobium oxides. Chem Mater 1991;3:100–7. http://dx.doi. org/10.1021/cm00013a025. [103] Nowak I, Ziolek M. Niobium compounds: preparation, characterization, and application in heterogeneous catalysis. Chem Rev 1999;99:3603–24. http://dx.doi.org/10.1021/cr9800208. [104] de Sá AI, Rangel CM, Skeldon P. Semiconductive properties of anodic niobium oxides. Port Electrochim Acta 2006;24:305–11. [105] de Filho DAB, Franco DW, Filho PPA, Alves OL. Niobia films: surface morphology, surface analysis, photoelectrochemical properties and crystallization process. J Mater Sci 1998;33:2607–16. http://dx.doi.org/10.1023/A:1004361420740. [106] Schultze JW, Lohrengel MM. Stability, reactivity and breakdown of passive films. Problems of recent and future research. Electrochim Acta 2000;45:2499–513. http://dx.doi.org/10.1016/S0013-4686(00)00347-9. [107] Kovendhan M, Joseph DP, Manimuthu P, Ganesan S, Sambasivam S, Maruthamuthu P, et al. Spray deposited Nb2O5 thin film electrodes for fabrication of dye sensitized solar cells. Trans Indian Inst Met 2011;64:185–8. http://dx.doi.org/ 10.1007/s12666-011-0036-2. [108] Marucco JF. Thermodynamic study of the system NbO2–Nb2O5 at high temperature. J Solid State Chem 1974;10:211–8. http://dx.doi.org/10.1016/0022-4596(74)90028-0. [109] Naito K, Kamegashira N, Sasaki N. Phase equilibria in the system between NbO2 and Nb2O5 at high temperatures. J Solid State Chem 1980;35:305–11. http://dx.doi.org/10.1016/0022-4596(80)90526-5. [110] Fischer V, Stormer H, Gerthsen D, Stenzel M, Zillgen H, Ivers-Tiffee E. Niobium as new material for electrolyte capacitors with nanoscale dielectric oxide layers. In: Proc 7th int conf prop appl dielectr mater, vol. 3, ICPADM 2003, Nagoya, Japan; 2003. p. 1134–7. [111] Cavigliasso GE, Esplandiu MJ, Macagno VA. Influence of the forming electrolyte on the electrical properties of tantalum and niobium oxide films: an EIS comparative study. J Appl Electrochem 1998;28:1213–9. [112] Macek M, Orel B. Electrochromism of sol–gel derived niobium oxide films. Sol Energy Mater Sol Cells 1998;54:121–30. http://dx.doi.org/10.1016/S0927-0248(98)00062-2. [113] Störmer H, Weber A, Fischer V, Ivers-Tiffée E, Gerthsen D. Anodically formed oxide films on niobium: microstructural and electrical properties. J Eur Ceram Soc 2009;29:1743–53. http://dx.doi.org/10.1016/j.jeurceramsoc.2008.10.019. [114] Zednicek T, Vrana B. Tantalum and niobium technology roadmap. AVX Corp; n.d. [115] Jouve G. Electrical conduction mechanisms in electrochemically formed amorphous films of Nb2O5. Philos Mag Part B 1991;64:207–18. http://dx.doi.org/10.1080/13642819108207614. [116] El-Shazly TS, Hassan WMI, Abdel Rahim ST, Allam NK. Unravelling the interplay of dopant concentration and band structure engineering of monoclinic niobium pentoxide: a model photoanode for water splitting. Int J Hydrogen Energy 2015;40:13867–75. http://dx.doi.org/10.1016/j.ijhydene.2015.08.05. [117] Kukli K, Ritala M, Leskelä M. Development of dielectric properties of niobium oxide, tantalum oxide, and aluminum oxide based nanolayered materials. J Electrochem Soc 2001;148:F35–41. http://dx.doi.org/10.1149/1.1343106. [118] Aegerter MA. Sol–gel niobium pentoxide: a promising material for electrochromic coatings, batteries, nanocrystalline solar cells and catalysis. Sol Energy Mater Sol Cells 2001;68:401–22. [119] Maruyama T, Arai S. Electrochromic properties of niobium oxide thin films prepared by radio-frequency magnetron sputtering method. Appl Phys Lett 1993;63:869–70. [120] Özer N, Rubin MD, Lampert CM. Optical and electrochemical characteristics of niobium oxide films prepared by sol–gel process and magnetron sputtering: a comparison. Sol Energy Mater Sol Cells 1996;40:285–96. http://dx.doi.org/10.1016/ 0927-0248(95)00147-6. [121] Pawlicka A, Atik M, Aegerter MA. Synthesis of multicolor Nb2O5 coatings for electrochromic devices. Thin Solid Films 1997;301:236–41. http://dx.doi.org/10.1016/S0040-6090(96)09583-1. [122] Costa ED, Avellaneda CO, Pawlicka A. Alternative Nb2O5–TiO2 thin films for electrochromic devices. J Mater Sci 2001;36:1407–10. http://dx.doi.org/10.1023/A:1017576125092. [123] Huang Y, Zhang Y, Hu X. Structural, morphological and electrochromic properties of Nb2O5 films deposited by reactive sputtering. Sol Energy Mater Sol Cells 2003;77:155–62. http://dx.doi.org/10.1016/S0927-0248(02)00318-5. [124] Pehlivan E, Tepehan FZ, Tepehan GG. Comparison of optical, structural and electrochromic properties of undoped and WO3-doped Nb2O5 thin films. Solid State Ionics 2003;165:105–10. http://dx.doi.org/10.1016/j.ssi.2003.08.021.

34

C. Nico et al. / Progress in Materials Science 80 (2016) 1–37

[125] Romero R, Dalchiele EA, Martín F, Leinen D, Ramos-Barrado JR. Electrochromic behaviour of Nb2O5 thin films with different morphologies obtained by spray pyrolysis. Sol Energy Mater Sol Cells 2009;93:222–9. http://dx.doi.org/10.1016/ j.solmat.2008.10.012. [126] Abdul Rani R, Zoolfakar AS, Subbiah J, Ou JZ, Kalantar-zadeh K. Highly ordered anodized Nb2O5 nanochannels for dyesensitized solar cells. Electrochem Commun; n.d. http://dx.doi.org/10.1016/j.elecom.2013.12.011. [127] Barea E, Xu X, González-Pedro V, Ripollés-Sanchis T, Fabregat-Santiago F, Bisquert J. Origin of efficiency enhancement in Nb2O5 coated titanium dioxide nanorod based dye sensitized solar cells. Energy Environ Sci 2011;4:3414–9. http://dx.doi. org/10.1039/C1EE01193F. [128] Jose R, Thavasi V, Ramakrishna S. Metal oxides for dye-sensitized solar cells. J Am Ceram Soc 2009;92:289–301. http://dx. doi.org/10.1111/j.1551-2916.2008.02870.x. [129] Cui H, Zhu G, Xie Y, Zhao W, Yang C, Lin T, et al. Black nanostructured Nb2O5 with improved solar absorption and enhanced photoelectrochemical water splitting. J Mater Chem A 2015;3:11830–7. http://dx.doi.org/10.1039/ C5TA01544H. [130] Tanabe K. Catalytic application of niobium compounds. Catal Today 2003;78:65–77. [131] Fan M-Q, Sun L-X, Zhang Y, Xu F, Zhang J, Chu H. The catalytic effect of additive Nb2O5 on the reversible hydrogen storage performances of LiBH4–MgH2 composite. Int J Hydrogen Energy 2008;33:74–80. [132] Kurioka N, Watanabe D, Haneda M, Shimanouchi T, Mizushima T, Kakuta N, et al. Preparation of niobium oxide films as a humidity sensor. Catal Today 1993;16:495–501. http://dx.doi.org/10.1016/0920-5861(93)80090-N. [133] Choi J, Lim JH, Rho S, Jahng D, Lee J, Kim KJ. Nanoporous niobium oxide for label-free detection of DNA hybridization events. Talanta 2008;74:1056–9. http://dx.doi.org/10.1016/j.talanta.2007.07.007. [134] Eisenbarth E, Velten D, Müller M, Thull R, Breme J. Nanostructured niobium oxide coatings influence osteoblast adhesion. J Biomed Mater Res A 2006;79:166–75. [135] Velten D, Eisenbarth E, Schanne N, Breme J. Biocompatible Nb2O5 thin films prepared by means of the sol–gel process. J Mater Sci – Mater Med 2004;15:457–61. [136] Ramírez G, Rodil SE, Arzate H, Muhl S, Olaya JJ. Niobium based coatings for dental implants. Appl Surf Sci 2011;257:2555–9. [137] Cardinal T, Fargin E, Le Flem G, Leboiteux S. Correlations between structural properties of Nb2O5–NaPO3–Na2B4O7 glasses and non-linear optical activities. J Non-Cryst Solids 1997;222:228–34. [138] Praveena R, Venkatramu V, Babu P, Jayasankar CK. Fluorescence spectroscopy of Sm3+ ions in P2O5–PbO–Nb2O5 glasses. Phys B Condens Matter 2008;403:3527–34. [139] Sanghi S, Rani S, Agarwal A, Bhatnagar V. Influence of Nb2O5 on the structure, optical and electrical properties of alkaline borate glasses. Mater Chem Phys 2010;120:381–6. http://dx.doi.org/10.1016/j.matchemphys.2009.11.016. [140] Graça MPF, Valente MA, Ferreira Da Silva MG. The electric behavior of a lithium-niobate-phosphate glass and glassceramics. J Mater Sci 2006;41:1137–44. http://dx.doi.org/10.1007/s10853-005-3652-6. [141] Lin H, Jiang S, Wu J, Song F, Peyghambarian N, Pun EYB. Er3+ doped Na2O–Nb2O5–TeO2 glasses for optical waveguide laser and amplifier. J Phys Appl Phys 2003;36:812. [142] Aquino FT, Ferrari JL, Ribeiro SJL, Ferrier A, Goldner P, Goncalves RR. Broadband NIR emission in novel sol–gel Er3+-doped SiO2–Nb2O5 glass ceramic planar waveguides for photonic applications. Opt Mater 2013;35:387–96. [143] Eckert J, Reichert K, Schnitter C, Seyeda H. Processing of columbite-tantalite ores and concentrates for niobium and niobium compounds in electronic applications. In: Proc int symp niobium 2001, Niobium 2001 Ltd, Bridgeville, USA; n.d. p. 67–87. [144] Michael Scherer HH. Innovative production of thin film laser components 2005; 5963. http://dx.doi.org/10.1117/12. 625226. [145] Dhar A, Alford TL. Optimization of Nb2O5/Ag/Nb2O5 multilayers as transparent composite electrode on flexible substrate with high figure of merit. J Appl Phys 2012;112:103113. http://dx.doi.org/10.1063/1.4767662. [146] Vinnichenko M, Rogozin A, Grambole D, Munnik F, Kolitsch A, Möller W, et al. Highly dense amorphous Nb2O5 films with closed nanosized pores. Appl Phys Lett 2009;95:081904. http://dx.doi.org/10.1063/1.3212731. [147] Hunsche B, Vergöhl M, Neuhäuser H, Klose F, Szyszka B, Matthée T. Effect of deposition parameters on optical and mechanical properties of MF- and DC-sputtered Nb2O5 films. Thin Solid Films 2001;392:184–90. http://dx.doi.org/ 10.1016/S0040-6090(01)01025-2. [148] Chua LO. Memristor – the missing circuit element. IEEE Trans Circuit Theory 1971;18:507–19. http://dx.doi.org/10.1109/ TCT.1971.1083337. [149] Strukov DB, Snider GS, Stewart DR, Williams RS. The missing memristor found. Nature 2008;453:80–3. http://dx.doi.org/ 10.1038/nature06932. [150] Yang JJ, Strukov DB, Stewart DR. Memristive devices for computing. Nat Nanotechnol 2013;8:13–24. http://dx.doi.org/ 10.1038/nnano.2012.240. [151] Jo SH, Chang T, Ebong I, Bhadviya BB, Mazumder P, Lu W. Nanoscale memristor device as synapse in neuromorphic systems. Nano Lett 2010;10:1297–301. http://dx.doi.org/10.1021/nl904092h. [152] Liu X, Sadaf SM, Son M, Park J, Shin J, Lee W, et al. Co-occurrence of threshold switching and memory switching in cells for crosspoint memory applications. Electron Device Lett IEEE 2012;33:236–8. [153] Linn E, Rosezin R, Kügeler C, Waser R. Complementary resistive switches for passive nanocrossbar memories. Nat Mater 2010;9:403–6. [154] Rosezin R, Linn E, Nielen L, Kugeler C, Bruchhaus R, Waser R. Integrated complementary resistive switches for passive high-density nanocrossbar arrays. Electron Device Lett IEEE 2011;32:191–3. [155] Sawa A. Resistive switching in transition metal oxides. Mater Today 2008;11:28–36. http://dx.doi.org/10.1016/S13697021(08)70119-6. [156] Yang JJ, Zhang M-X, Strachan JP, Miao F, Pickett MD, Kelley RD, et al. High switching endurance in TaOx memristive devices. Appl Phys Lett 2010;97:232102. [157] Lee M-J, Lee CB, Lee D, Lee SR, Chang M, Hur JH, et al. A fast, high-endurance and scalable non-volatile memory device made from asymmetric Ta2O5x/TaO2x bilayer structures. Nat Mater 2011;10:625–30.

C. Nico et al. / Progress in Materials Science 80 (2016) 1–37

35

[158] Chen Y-S, Lee H-Y, Chen P-S, Chen W-S, Tsai K-H, Gu P-Y, et al. Novel defects-trapping TaOX/HfOX RRAM with reliable self-compliance, high nonlinearity, and ultra-low current; 2014. [159] Liu X, Sadaf SM, Park S, Kim S, Cha E, Lee D, et al. Complementary resistive switching in niobium oxide-based resistive memory devices. Electron Device Lett IEEE 2013;34:235–7. [160] Grishin AM, Velichko AA, Jalalian A. Nb2O5 nanofiber memristor. Appl Phys Lett 2013;103:053111. http://dx.doi.org/ 10.1063/1.4817302. [161] Sadaf SM, Liu X, Son M, Park S, Choudhury SH, Cha E, et al. Highly uniform and reliable resistance switching properties in bilayer WOx/NbOx RRAM devices. Phys Status Solidi A 2012;209:1179–83. [162] Kundozerova TV, Grishin AM, Stefanovich GB, Velichko AA. Anodic Nb2O5 nonvolatile RRAM. IEEE Trans Electron Devices 2012;59:1144–8. http://dx.doi.org/10.1109/TED.2011.2182515. [163] Liu X, Sadaf SM, Son M, Shin J, Park J, Lee J, et al. Diode-less bilayer oxide (WOx–NbOx) device for cross-point resistive memory applications. Nanotechnology 2011;22:475702. [164] Jo Y, Sim H, Hwang H, Ahn K, Jung M-H. Reverse resistance switching in polycrystalline Nb2O5 films. Thin Solid Films 2010;518:5676–8. http://dx.doi.org/10.1016/j.tsf.2009.10.026. [165] Sim Hyunjun, Choi Dooho, Lee Dongsoo, Seo Sunae, Lee Myong-Jae, Yoo In-Kyeong, et al. Resistance-switching characteristics of polycrystalline Nb2O5 for nonvolatile memory application. IEEE Electron Device Lett 2005;26:292–4. http://dx.doi.org/10.1109/LED.2005.846592. [166] Mitin A, Shamray V, Gordeev A. Effects of resistance switching in niobium oxides: metastable nanochannels. Bull Russ Acad Sci Phys 2009;73:125–8. http://dx.doi.org/10.3103/S106287380901033X. [167] Baek H, Lee C, Choi J, Cho J. Nonvolatile memory devices prepared from sol–gel derived niobium pentoxide films. Langmuir 2013;29:380–6. http://dx.doi.org/10.1021/la303857b. [168] Roberts MW, Thomas JM, Anderson JS, Tilley RJD. Crystallographic shear and non-stoichiometry. In: Surf defect prop solids. vol. 3; 2007. [169] Iijima S. Direct observation of lattice defects in H-Nb2O5 by high resolution electron microscopy. Acta Crystallogr Sect A 1973;29:18–24. http://dx.doi.org/10.1107/S0567739473000045. [170] Niebuhr J. Die niederen Oxide des Niobs. J Common Met 1966;11:191–203. [171] Kimura S. Phase equilibria in the system NbO2–Nb2O5: phase relations at 1300 and 1400 °C and related thermodynamic treatment. J Solid State Chem 1973;6:438–49. [172] Marucco JF. Electrical resistance and defect structure of stable and metastable phases of the system Nb12O29–Nb2O5 between 800 and 1100 °C. J Chem Phys 1979;70:649–54. http://dx.doi.org/10.1063/1.437545. [173] McQueen T, Xu Q, Andersen EN, Zandbergen HW, Cava RJ. Structures of the reduced niobium oxides Nb12O29 and Nb22O54. J Solid State Chem 2007;180:2864–70. [174] Ohsawa T, Okubo J, Suzuki T, Kumigashira H, Oshima M, Hitosugi T. An n-type transparent conducting oxide: Nb12O29. J Phys Chem C 2011;115:16625–9. [175] Waldron JEL, Green MA, Neumann DA. Structure and electronic properties of monoclinic Nb12O29. J Phys Chem Solids 2004;65:79–86. http://dx.doi.org/10.1016/j.jpcs.2003.09.001. [176] Janninck RF, Whitmore DH. Electrical conduction in nonstoichiometric a-Nb2O5. J Chem Phys 1962;37:2750–4. http://dx. doi.org/10.1063/1.1733100. [177] Cava RJ, Batlogg B, Krajewski JJ, Gammel P, Poulsen HF, Peck WF, et al. Antiferromagnetism and metallic conductivity in Nb12O29. Nature 1991;350:598–600. http://dx.doi.org/10.1038/350598a0. [178] Waldron JEL, Green MA, Neumann DA. Charge and spin ordering in monoclinic Nb12O29. J Am Chem Soc 2001;123:5833–4. http://dx.doi.org/10.1021/ja015506b. [179] Xu J, Jarlborg T, Freeman A. Self-consistent band structure of the rutile dioxides NbO2, RuO2, and IrO2. Phys Rev B: Condens Matter 1989;40:7939–47. [180] Ikeya T, Senna M. Change in the structure of niobium pentoxide due to mechanical and thermal treatments. J Non-Cryst Solids 1988;105:243–50. http://dx.doi.org/10.1016/0022-3093(88)90313-4. [181] McConnell AA, Aderson JS, Rao CNR. Raman spectra of niobium oxides. Spectrochim Acta Part Mol Spectrosc 1976;32:1067–76. http://dx.doi.org/10.1016/0584-8539(76)80291-7. [182] Huang BX, Wang K, Church JS, Li Y-S. Characterization of oxides on niobium by Raman and infrared spectroscopy. Electrochim Acta 1999;44:2571–7. http://dx.doi.org/10.1016/S0013-4686(98)00385-5. [183] Balachandran U, Eror NG. Raman spectrum of the high temperature form of Nb2O5. J Mater Sci Lett 1982;1:374–6. http:// dx.doi.org/10.1007/BF00724842. [184] Adler D. Mechanisms for metal–nonmental transitions in transition-metal oxides and sulfides. Rev Mod Phys 1968;40:714–36. http://dx.doi.org/10.1103/RevModPhys.40.714. [185] Lee JC, Durand WW. Electrically stimulated optical switching of NbO2 thin films. J Appl Phys 1984;56:3350–2. http://dx. doi.org/10.1063/1.333863. [186] Sakai Y, Tsuda N, Sakata T. Electrical properties of semiconducting NbO2. J Phys Soc Jpn 1985;54:1514–8. http://dx.doi. org/10.1143/JPSJ.54.1514. [187] Manzo M, Laurell F, Pasiskevicius V, Gallo K. Lithium niobate: the silicon of photonics! In: Bartolo BD, Collins J, editors. Nano-opt enhancing light-matter interact mol scale. Netherlands: Springer; 2013. p. 421–2. [188] Wang K, Li J-F. (K, Na)NbO3-based lead-free piezoceramics: phase transition, sintering and property enhancement. J Adv Ceram 2012;1:24–37. http://dx.doi.org/10.1007/s40145-012-0003-3. [189] Thierfelder C, Sanna S, Schindlmayr A, Schmidt WG. Do we know the band gap of lithium niobate? Phys Status Solidi C 2010;7:362–5. http://dx.doi.org/10.1002/pssc.200982473. [190] Rachel, Dutto F, Radenovic A. Niobates nanowires: synthesis, characterization and applications. In: Hashim A, editor. Nanowires – implement. Appl. InTech; 2011. [191] Graça MPF, Valente MA, Peres M, Cruz A, Soares MJ, Neves AJ, et al. Structural and optical properties of Er3+ ion in sol–gel grown LiNbO3. J Phys: Condens Matter 2007;19:016213. http://dx.doi.org/10.1088/0953-8984/19/1/016213. [192] Nico C, Graça MPF, Monteiro T, Valente MA. Lithium niobosilicate glass doped with samarium: dielectric and optical analysis. Phys Status Solidi A 2011;208:2288–92. http://dx.doi.org/10.1002/pssa.201000764.

36

C. Nico et al. / Progress in Materials Science 80 (2016) 1–37

[193] Fang J, Wang X, Tian Z, Zhong C, Li L, Zuo R. Two-step sintering: an approach to broaden the sintering temperature range of alkaline niobate-based lead-free piezoceramics. J Am Ceram Soc 2010;93:3552–5. http://dx.doi.org/10.1111/j.15512916.2010.04085.x. [194] Kuznetsova E, Reznichenko L, Razumovskaya O, Shilkina L. The polymorphism of niobium pentoxide and the properties of alkali metal niobates for ferroelectric piezoceramics. Tech Phys Lett 2001;27:189–92. http://dx.doi.org/10.1134/ 1.1359821. [195] Hrešcˇak J, Bencan A, Rojac T, Malicˇ B. The influence of different niobium pentoxide precursors on the solid-state synthesis of potassium sodium niobate. J Eur Ceram Soc 2013;33:3065–75. http://dx.doi.org/10.1016/j.jeurceramsoc.2013.07.006. [196] Yan C, Nikolova L, Dadvand A, Harnagea C, Sarkissian A, Perepichka DF, et al. Multiple NaNbO3/Nb2O5 heterostructure nanotubes: a new class of ferroelectric/semiconductor nanomaterials. Adv Mater 2010;22:1741–5. http://dx.doi.org/ 10.1002/adma.200903589. [197] Pullar RC. The synthesis, properties, and applications of columbite niobates (M2+Nb2O6): a critical review. J Am Ceram Soc 2009;92:563–77. http://dx.doi.org/10.1111/j.1551-2916.2008.02919.x. [198] Cho I-S, Bae ST, Kim DH, Hong KS. Effects of crystal and electronic structures of ANb2O6 (A = Ca, Sr, Ba) metaniobate compounds on their photocatalytic H2 evolution from pure water. Int J Hydrogen Energy 2010;35:12954–60. http://dx. doi.org/10.1016/j.ijhydene.2010.04.057. [199] Hsiao Y-J, Liu C-W, Dai B-T, Chang Y-H. Sol–gel synthesis and the luminescent properties of CaNb2O6 phosphor powders. J Alloys Compd 2009;475:698–701. http://dx.doi.org/10.1016/j.jallcom.2008.07.142. [200] Cho I-S, Bae ST, Yim DK, Kim DW, Hong KS. Preparation, characterization, and photocatalytic properties of CaNb2O6 nanoparticles. J Am Ceram Soc 2009;92:506–10. http://dx.doi.org/10.1111/j.1551-2916.2008.02894.x. [201] Arroyo y de Dompablo ME, Lee Y-L, Morgan D. First principles investigation of oxygen vacancies in columbite MNb2O6 (M = Mn, Fe Co, Ni, Cu). Chem Mater 2010;22:906–13. http://dx.doi.org/10.1021/cm901723. [202] Thornton JR, Fountain WD, Flint GW, Crow TG. Properties of neodymium laser materials. Appl Opt 1969;8:1087–102. http://dx.doi.org/10.1364/AO.8.001087. [203] Ballman AA, Porto SPS, Yariv A. Calcium niobate Ca(NbO3)2—a new laser host crystal. J Appl Phys 1963;34:3155–6. http:// dx.doi.org/10.1063/1.1729154. [204] Hsiao Y-J, Chang Y-S, Chen G-J, Chang Y-H. Synthesis and the luminescent properties of CdNb2O6 oxides by sol–gel process. J Alloys Compd 2009;471:259–62. http://dx.doi.org/10.1016/j.jallcom.2008.03.081. [205] Hsiao Y-J, Fang T-H, Ji L-W, Chi S-S. Surface and photoluminescence characteristics of CdNb2O6 nanocrystals. Open Surf Sci J 2009;1:30–3. http://dx.doi.org/10.2174/1876531900901010030. [206] Qin L, Wei D, Huang Y, Kim SI, Yu YM, Seo HJ. Triple-layered perovskite niobates CaRNb3O10 (R = La, Sm, Eu, Gd, Dy, Er, Yb, or Y): new self-activated oxides. Inorg Chem 2013;52:10407–13. http://dx.doi.org/10.1021/ic401854r. [207] Rooksby HP, White EAD, Langston SA. Perovskite-type rare-earth niobates and tantalates. J Am Ceram Soc 1965;48:447–9. [208] Rooksby HP, White EAD. Rare-earth niobates and tantalates of defect fluorite-and weberite-type structures. J Am Ceram Soc 1964;47:94–6. [209] Rooksby HP, White EAD. The structures of 1:1 compounds of rare earth oxides with niobia and tantala. Acta Crystallogr 1963;16:888–90. [210] Cashion JD, Cooke AH, Leask MJM, Thorp TL, Wells MR. Crystal growth and magnetic susceptibility of some rare-earth compounds. J Mater Sci 1968;3:402–7. http://dx.doi.org/10.1007/BF00550984. [211] Garton G, Wanklyn BM. Crystal growth and magnetic susceptibility of some rare-earth compounds. J Mater Sci 1968;3:395–401. http://dx.doi.org/10.1007/BF00550983. [212] Siqueira KPF, Moreira RL, Dias A. Synthesis and crystal structure of lanthanide orthoniobates studied by vibrational spectroscopy. Chem Mater 2010;22:2668–74. [213] Kim D-W, Kwon D-K, Yoon SH, Hong KS. Microwave dielectric properties of rare-earth ortho-niobates with ferroelasticity. J Am Ceram Soc 2006;89:3861–4. http://dx.doi.org/10.1111/j.1551-2916.2006.01302.x. [214] Siqueira KPF, Carvalho GB, Dias A. Influence of the processing conditions and chemical environment on the crystal structures and phonon modes of lanthanide orthotantalates. Dalton Trans 2011;40:9454–60. http://dx.doi.org/10.1039/ C1DT10783F. [215] Haugsrud R, Norby T. Proton conduction in rare-earth ortho-niobates and ortho-tantalates. Nat Mater 2006;5:193–6. http://dx.doi.org/10.1038/nmat1591. [216] Nyman M, Rodriguez MA, Rohwer LES, Martin JE, Waller M, Osterloh FE. Unique LaTaO4 polymorph for multiple energy applications. Chem Mater 2009;21:4731–7. http://dx.doi.org/10.1021/cm9020645. [217] Octaviano ES, Reyes Ardila D, Andrade LHC, Siu Li M, Andreeta JP. Growth and evaluation of lanthanoids orthoniobates single crystals processed by a miniature pedestal growth technique. Cryst Res Technol 2004;39:859–63. [218] Graça MPF, Peixoto M, Ferreira N, Rodrigues J, Nico C, Costa F, et al. Optical and dielectric behaviour of EuNbO4 crystals. J Mater Chem C 2013;1:2913–9. [219] Xiao X, Yan B. Synthesis and luminescent properties of novel RENbO4:Ln3+ (RE = Y, Gd, Lu; Ln = Eu, Tb) micro-crystalline phosphors. J Non-Cryst Solids 2005;351:3634–9. http://dx.doi.org/10.1016/j.jnoncrysol.2005.09.018. [220] Nyman MD, Rohwer LES. Rare-earth tantalates and niobates suitable for use as nanophosphors. US8585928 B1; 2013. [221] Kato K, Toyoura K, Nakamura A, Matsunaga K. First-principles analysis on proton diffusivity in La3NbO7. Solid State Ionics 2013;262:472–5. [222] Ivanova M, Ricote S, Meulenberg WA, Haugsrud R, Ziegner M. Effects of A-and B-site (co-) acceptor doping on the structure and proton conductivity of LaNbO4. Solid State Ionics 2012;213:45–52. [223] Kepaptsoglou D-M, Hadidi K, Løvvik O-M, Magraso A, Norby T, Gunnaes AE, et al. Interfacial charge transfer and chemical bonding in a Ni–LaNbO4 cermet for proton-conducting solid-oxide fuel cell anodes. Chem Mater 2012;24:4152–9. [224] Amsif M, Marrero-López D, Ruiz-Morales JC, Savvin S, Núñez P. Low temperature sintering of LaNbO4 proton conductors from freeze-dried precursors. J Eur Ceram Soc 2012;32:1235–44. [225] Mather GC, Fisher CAJ, Islam MS. Defects, dopants, and protons in LaNbO4. Chem Mater 2010;22:5912–7. http://dx.doi. org/10.1021/cm1018822.

C. Nico et al. / Progress in Materials Science 80 (2016) 1–37

37

[226] Tolchard JR, Lein HL, Grande T. Chemical compatibility of proton conducting LaNbO4 electrolyte with potential oxide cathodes. J Eur Ceram Soc 2009;29:2823–30. http://dx.doi.org/10.1016/j.jeurceramsoc.2009.03.034. [227] Reijers R, Haije W. Literature review on high temperature proton conducting materials. Energy Res Cent Neth; 2008. [228] Mokkelbost T, Kaus I, Haugsrud R, Norby T, Grande T, Einarsrud M-A. High-temperature proton-conducting lanthanum ortho-niobate-based materials. Part II: sintering properties and solubility of alkaline earth oxides. J Am Ceram Soc 2008;91:879–86. http://dx.doi.org/10.1111/j.1551-2916.2007.02232.x. [229] Mokkelbost T, Andersen Ø, Strøm RA, Wiik K, Grande T, Einarsrud M-A. High-temperature proton-conducting LaNbO4based materials: powder synthesis by spray pyrolysis. J Am Ceram Soc 2007;90:3395–400. http://dx.doi.org/10.1111/ j.1551-2916.2007.01904.x. [230] Baughman RH, Stafström S, Cui C, Dantas SO. Materials with negative compressibilities in one or more dimensions. Science 1998;279:1522–4. [231] Pan L, Wang Y, Wang X, Qu H, Zhao J, Li Y, et al. Hydrogen photochromism in Nb2O5 powders. Phys Chem Chem Phys 2014. http://dx.doi.org/10.1039/C4CP02834A. [232] Joshua Yang J, Miao F, Pickett MD, Ohlberg DAA, Stewart DR, Lau CN, et al. The mechanism of electroforming of metal oxide memristive switches. Nanotechnology 2009;20:215201. http://dx.doi.org/10.1088/0957-4484/20/21/21520. [233] Park S-J, Lee J-P, Jang JS, Rhu H, Yu H, You BY, et al. In situ control of oxygen vacancies in TiO2 by atomic layer deposition for resistive switching devices. Nanotechnology 2013;24:295202. http://dx.doi.org/10.1088/0957-4484/24/29/295202. [234] Waser R, Aono M. Nanoionics-based resistive switching memories. Nat Mater 2007;6:833–40. [235] Messerschmitt F, Kubicek M, Schweiger S, Rupp JLM. Memristor kinetics and diffusion characteristics for mixed anionicelectronic SrTiO3-d bits: the memristor-based cottrell analysis connecting material to device performance. Adv Funct Mater 2014;24:7448–60. http://dx.doi.org/10.1002/adfm.201402286. [236] Valov I. Redox-based resistive switching memories (ReRAMs): electrochemical systems at the atomic scale. ChemElectroChem 2014;1:26–36. http://dx.doi.org/10.1002/celc.201300165. [237] Lee S, Na H, Kim J, Moon J, Sohn H. Anion-migration-induced bipolar resistance switching in electrochemically deposited TiOx films. J Electrochem Soc 2011;158:H88–92. http://dx.doi.org/10.1149/1.3516464. [238] Waser R, Dittmann R, Staikov G, Szot K. Redox-based resistive switching memories – nanoionic mechanisms, prospects, and challenges. Adv Mater 2009;21:2632–63. [239] Hamwi S, Meyer J, Kröger M, Winkler T, Witte M, Riedl T, et al. The role of transition metal oxides in charge-generation layers for stacked organic light-emitting diodes. Adv Funct Mater 2010;20:1762–6. http://dx.doi.org/10.1002/ adfm.201000301. [240] Liang J, Zu F-S, Ding L, Xu M-F, Shi X-B, Wang Z-K, et al. Aqueous solution-processed MoO3 thick films as hole injection and short-circuit barrier layer in large-area organic light-emitting devices. Appl Phys Exp 2014;7:111601. http://dx.doi. org/10.7567/APEX.7.111601.