Structural modifications of M₅O₄I₁₁ (M = Nb, Ta) cluster networks from heterogeneous solid-state reactions

Abstract

The cluster compounds M₅O₄I₁₁ (M = Nb, Ta) and Ta₅O₄I₁₁(TaI₅) were obtained from heterogeneous solid-state reactions and structurally characterised by single-crystal X-ray diffraction. Their crystal structures are based on the novel [M₅O₄] cluster core with metal (M) atoms arranged following the motif of a square pyramid. Iodide ligands contribute to different connectivities in the structures, resulting in (van der Waals type) waved layer structures. Two structural modifications exist for Ta₅O₄I₁₁, denoted as o-Ta₅O₄I₁₁ and m-Ta₅O₄I₁₁, and the compound Ta₅O₄I₁₁(TaI₅) encloses [TaI₅] molecules within voids of the structure. The two-dimensional nature of the structures and the presence of metal-to-metal bonding motivated investigations of the electronic properties through optical band-gap measurements, electrical conductivity studies, electronic band structure calculations, and X-ray photoelectron spectroscopy.

---

Introduction

The chemistry of reduced niobium and tantalum halides is characterised by metal cluster compounds that are usually achieved through solid-state synthesis at elevated temperatures.¹ These compounds display a wide range of structural motifs and oxidation states, with a pronounced tendency to form clusters featuring octahedral M₆ (M = Nb, Ta) cores.² Many transition metal (M) halide (X) clusters crystallise in either the [M₆X₁₂]- or [M₆X₈]-structural types.^3–6 The [M₆X₁₂] structure consists of twelve halide ligands bridging the edges of an M₆ octahedron which generally favours combinations of larger metal atoms with smaller halides. In contrast, [M₆X₈] clusters, in which eight halides cap the faces of the octahedron, tend to form when smaller metal atoms are paired with larger halide ions. Due to the spatial separation of adjacent clusters by outer ligands, these crystalline materials typically exhibit semiconducting behaviour with poor electrical conductivity.^7,8 The chemistry of binary niobium halide clusters is further enriched by triangular Nb₃ clusters present in Nb₃X₈ (X = Cl, Br, I), as well as Peierls-distorted NbX₄.^9,10 Metal-rich tantalum halides exist as Ta₆X₁₄, Ta₆X₁₅ and TaX₄.^5,11–15

Beyond these well-established binary systems, the range of cluster compounds and structural diversity expands significantly in heteroanionic halides of niobium and tantalum. Among the most common examples are chalcohalides, which include M₃ clusters, such as Ta₃SBr₇ and ANb₃SBr₇ (A = Rb,Cs),^16–18 as well as M₄ clusters exhibiting planar, butterfly or tetrahedral geometries (e.g. Nb₄OI₁₀; Ta₄S₉Br₈; Ta₄SBr_11; Nb₄Se₄I₄)^19–26 Octahedral M₆ clusters are also observed, as in Nb₆I₉S.²⁷ A notable characteristic of these compounds is the presence of bridging halides along with the interstitial or capping atoms, which are most often chalcogenides, and rarely pnictides (Nb₄PnX₁₁, Pn = N, P; X = Cl, Br, I).²⁵

Occasionally, molecular MX₅ units are found to be incorporated into a cluster network. A notable example is Nb₇S₂I₁₉ which can be represented as (Nb₃SI₇)₂(NbI₅), emphasising the stabilisation of thermodynamically unstable NbI₅ monomers within the inorganic framework of Nb₃SI₇. This structure exemplifies a form of structural synergism, where the enclosed molecule influences the topology of the surrounding cluster framework.²⁸ Similarly, space-filling [ZrCl₅]⁻ units are observed in the compound described as Cs₃Zr₇Cl₂₀Mn and [TaBr₆]⁻ units in (Ta₆Br₁₂)Br₃(TaBr₆)_0.86.^29,30

Less common is the occurrence of pentanuclear M₅ clusters. Reported examples generally fall into two categories: those composed of two different metals, which give rise to a variety of geometries, and metal halide clusters that typically adopt a square-pyramidal cluster core. The latter is commonly observed in compounds with the general formula [M₅X₁₃]ⁿ⁻ (M = Mo, W, Tc; X = Cl, Br, I; n = 0, 1, 2, 3).^31–33

Oxyiodides of niobium and tantalum are reported as MO₂I, MOI₂ and NbOI₃.^34–36 More recently there has been some progress in the development of metal-rich oxyiodides.^19,37 In the course of this progress, we present a new class of metal-rich oxyiodides based on a distinctive M₅ cluster motif, observed for both niobium and tantalum.

These clusters, of the general formula M₅O₄I₁₁ with M = Nb and Ta, are shown to exist in two polymorphic forms for tantalum (o-Ta₅O₄I₁₁, m-Ta₅O₄I₁₁), as well as in the compound Ta₅O₄I₁₁(TaI₅), containing molecular [TaI₅] in its structure. The syntheses and crystal structures of the compounds are reported, and their electrical conductivities and electronic structures are analysed through a combination of experimental measurements and theoretical calculations.

Experimental

Preparations

Manipulations of starting materials, such as charging the silica ampoules (estimated volume 5 cm³) with starting materials were performed in an argon-filled glovebox (MBraun, labmaster 130, O₂ < 1 ppm, H₂O < 1 ppm).

o-Ta₅O₄I₁₁, m-Ta₅O₄I₁₁ and Ta₅O₄I₁₁(TaI₅)

A mixture of TaI₅ (144.8 mg, 0.178 mmol), Ta (6.1 mg, 0.03 mmol, Merck, 99.9%), Li₂O (5.1 mg, 0.169 mmol, ABCR, 96%) and Li₂CN₂ (4.6 mg, 0.085 mmol) was fused into an evacuated silica tube and heated with 0.5 K min⁻¹ to 500 °C in a Simon–Müller furnace. After heating at 500 °C for 48 hours, o-Ta₅O₄I₁₁ and m-Ta₅O₄I₁₁ were obtained as black, plate-like (o-Ta₅O₄I₁₁) and rod-shaped (m-Ta₅O₄I₁₁) crystals, which are sensitive to moisture (estimated yield: ∼30% (o-Ta₅O₄I₁₁); ∼10% (m-Ta₅O₄I₁₁); a powder X-ray diffraction pattern is shown in Fig. S1, top). Reactions at 500 °C with a shorter reaction time of 12 hours resulted in a few black crystals of Ta₅O₄I₁₁(TaI₅). Additional side phases were red TaO₂I and black, fibrous TaOI₂, as well as Ta₂O₃I₄ and LiTa₃O₂I₁₂ which are reported in this study.

Nb₅O₄I₁₁

NbI₄ (160.8 mg, 0.268 mmol), Li₂O (2 mg, 0.067 mmol) and Li₂(CN₂) (7.2 mg, 0.135 mmol) were fused into a silica ampoule and heated to 400 °C with a rate of 0.1 K min⁻¹. The holding time was 24 hours before the reaction was cooled down to room temperature with a rate of 0.1 K min⁻¹. Black, plate-like crystals of Nb₅O₄I₁₁ were obtained at the walls of the ampoule (yield: ∼10%); NbOI₂ and NbI₅ were sublimed to the cooler part of the ampoule, while LiI, Li₃Nb₇O₅I₁₅³⁸ and an amorphous phase were found in the hotter part at the bottom. Crystals behave sensitive in moist air. A powder X-ray diffraction pattern is shown in Fig. S1, bottom.

NbI ₄ and TaI ₅ were synthesised according to literature by reacting appropriate amounts of niobium or tantalum (Merck, 99.9%) with resublimed iodine (Merck, 99.999%) at 400–450 °C in evacuated quartz ampoules.³⁹

Li ₂ (CN ₂ ) was synthesised as described before,⁴⁰ by reacting LiH (99.4%, Alfa Aesar) with melamine (≥99%) in a 6 : 1 molar ratio under argon.

Powder X-ray diffraction

All reaction products were investigated by powder X-ray diffraction (PXRD) using a StadiP diffractometer (Stoe, Darmstadt) with Ge-monochromated Cu-K_α1 radiation, and a Mythen1 Detector.

Single-crystal X-ray diffraction

Data collections were performed on a Rigaku XtaLAB Synergy-S single-crystal X-ray diffractometer equipped with HyPix-6000HE detector and monochromated Mo-K_α radiation (λ = 0.71073 Å) and Cu-K_α radiation (λ = 1.54184 Å) at 150 K. X-ray intensities were corrected for absorption with a numerical method (crystal faces) using CrysAlisPro 1.171.43.121a (Rigaku Oxford Diffraction, 2024). The structures were solved by direct methods (SHELXT) and refined by full-matrix least–squares methods performed with SHELXL–2019/3 as implemented in Olex2 1.5.

EDX and SEM

Energy dispersive X-ray spectroscopy (EDX) and scanning electron microscopy (SEM) were performed on a HITACHI SU8030 scanning electron microscope with a Bruker QUANTAX 6G EDX-detector. Single crystals of o- and m-Ta₅O₄I₁₁ were mounted onto carbon tape under an argon atmosphere during sample preparation. For inert transfer, a vacuum transfer device, similar to the one reported by Yao et al., was used.⁴¹ In our adapted design, it consists of a steel lid with an O-ring seal (Fig. S2a). The lid is placed on top of the sample holder and evacuated within the glovebox airlock. It is then rapidly flooded with argon to create a reduced-pressure environment, ensuring an airtight seal via the O-ring (Fig. S2b and c). Once inserted into the SEM transfer chamber and evacuated, the internal pressure becomes equal to, or slightly higher than, the external pressure. This allows the magnetic lid to be lifted using neodymium magnets from outside the chamber (Fig. S2d).

Electrical conductivity

Conductivity measurements were performed in a Lake Shore Cryotronics CRX-6.5K probe station with a Keithley 2636B source meter unit. Rod-shaped crystals of o- and m-Ta₅O₄I₁₁ were contacted on a silicon substrate with 770 nm oxide layer using silver paste (Fig. S3, right) and transferred into the measuring chamber under protective gas. The conductive silver pads at each end of the crystals were connected to the circuit with gold coated tungsten tips. The chamber was kept under vacuum (<5 × 10⁻⁵ mbar) and the temperature was varied between 140 K and 350 K. Before each measurement, sufficient time was allowed for the sample to reach the chosen temperature. Two-point conductivity measurements were performed by varying the applied source–drain voltage from −2 V to 2 V while detecting the current. For time-resolved photocurrent measurements, using a picosecond pulsed laser driver (Taiko PDL M1, PicoQuant) together with a laser head 779 nm (pulse length < 500 ps) the crystals were illuminated at ∼55 mW cm⁻² laser output power using the continuous wave mode under a constant bias of 1 V. The electrical measurements shown in this work had crystal dimensions (length L; width W; height H) of o-Ta₅O₄I₁₁: L = 205 μm; W = 124 μm; H = 75 μm and m-Ta₅O₄I₁₁: L = 119.4 μm; W = 32.4 μm; H = 27.3 μm.

DRIFT (diffuse reflectance infrared fourier transformation) spectroscopy

Samples were measured at room temperature under inert conditions in diffuse reflectance with a Harrick Praying Mantis attachment using a Bruker Vertex 70 infrared spectrophotometer with a deuterated triglycine sulfate (DTGS) detector and KBr beamsplitter. The background spectra were collected using pure dried KBr in powder form.

TXRF (total internal reflection X-ray fluorescence) spectroscopy

TXRF studies were performed using a S2 Picofox (Bruker AXS Microanalysis, Berlin, Germany) equipped with a Mo X-ray tube, which was operated at 50 kV and 600 μA. The measurement period for each sample was 1000 s (live time). Fitting of the resulting spectra was done using the Spectra software (Bruker Nano GmbH) in the super bias mode (maximum stripping cycles of 2000).

XPS (X-ray photoelectron spectroscopy)

XPS measurements were performed under ultra-high vacuum (UHV) conditions (8 × 10⁻¹⁰ mbar) using an XR50 Al-K_α standard source equipped with a PHOIBOS 100 hemispherical analyser (SPECS GmbH, Berlin, Germany). Since o-Ta₅O₄I₁₁ represents the major phase of the synthesis, the crystals used for XPS measurements were carefully selected under the optical microscope. To avoid exposure to air and prevent oxidation, the synthesised samples were mounted on a double-sided conductive carbon tape under argon atmosphere. For inert transfer into the spectrometer, a custom-built vacuum transfer device (an enlarged version of the design used for SEM/EDX) was employed. This device consists of an aluminium lid with an O-ring that seals the sample holder under reduced pressure within the glovebox airlock (Fig. S4, SI). Upon evacuation of the XPS transfer chamber, the external pressure drops below the internal pressure, causing the lid to detach and fall off, thereby revealing the sample without contact to air. The chamber is then further evacuated to UHV before transferring the sample into the measurement chamber (see Fig. S5, SI for details).

The binding energy scale was calibrated to the signal positions of Au 4f_7/2 (84.0 eV), Ag 3d_5/2 (368.2 eV) and Cu 2p_3/2 (932.6 eV). To take charging effects into account, the acquired Ta 4f spectrum was referenced by setting the I 3d_5/2 peak to a binding energy of 619 eV. Peak fitting of XPS spectra was performed using the software Unifit 2018 (Unifit Scientific Software GmbH, Leipzig, Germany).

DFT (Density functional theory)

DFT calculations were performed in the software package Abinit (v. 10).⁴² The Perdew–Burke–Ernzerhof exchange–correlation functional was used with the dispersion correction of Grimme.^43,44 Calculations were performed using a plane wave basis set and the PAW formalism,⁴⁵ with an energy cut-off of 100 Ha inside the PAW spheres and 24 Ha outside. A 4 × 2 × 2 (o-Ta₅O₄I₁₁) or 3 × 2 × 2 (m-Ta₅O₄I₁₁) Monkhorst–Pack grid⁴⁶ of k-points was used to sample reciprocal space. These quantities were chosen following convergence studies. Methfessel–Paxton smearing was used to determine band occupation.⁴⁷ PAW data files were used as received from the Abinit library. The structures were relaxed to an internal pressure of 2 MPa prior to calculations of the electronic band structure.

Results and discussion

Synthesis and crystal structure

The two most stable compounds reported in the Ta–O–I system appear to be TaO₂I and TaOI₂, both of which were synthesised via transport reactions involving Ta, I₂ and Ta₂O₅ in temperature gradients of 450–550 °C.^34,35,48 In our study, we aimed to access more metal-rich compounds in this system by further reducing tantalum below the Ta(IV) oxidation state in TaOI₂. For this purpose we have explored Li₂CN₂ as an unconventional reduction agent, which has previously been shown its reducing nature, as demonstrated in metathesis reactions of NiCl₂ with Li₂CN₂, yielding elemental Ni, along with LiCl, C₃N₄ and N₂.⁴⁹

The employment of Li₂CN₂ into the Ta–O–I system triggers reduction reactions likely above 300 °C, marked by the emergence of a violet gas phase. As the temperature exceeded 400 °C, this gas phase darkened significantly. Upon cooling, crystalline iodine was found deposited on the cooler part of the ampoule, indicating that molecular iodine is present in the gas phase, likely along with reactive precursors that contribute to the observed cluster formation.

Additionally, slightly increased pressure was observed in the ampoules, indicating the formation of N₂. Amorphous C₃N₄ was identified as a side product via infrared spectroscopy (see Fig. S6, SI). These findings further confirm that Li₂CN₂ acts as a reducing agent in our system, consistent with its previously reported behaviour in the reduction of NiCl₂ to elemental Ni.

Thermal analysis of reactions aimed to investigate the impact of Li₂CN₂ on the Ta–O–I system proved to be complex, as the reaction pathways are intricate. In addition to the observed cluster compounds (discussed in detail in this study), various crystalline and amorphous by-products were detected.

Each experiment resulted in the formation of multiple products, indicating that several reactions occurred simultaneously. The spatial distributions of these products within the reaction container suggests the occurrence of partial transport reactions along a narrow temperature gradient.

Among the consistently identified side products were LiI, TaOI₂ and TaO₂I, as well as two previously unreported compounds: LiTa₃O₂I₁₂ and Ta₂O₃I₄. While these compounds are not the focus of this study, their structures have been characterised by us and are briefly described as follows. Both compounds feature Ta⁵⁺ ions in distorted octahedral coordination environments. The structure of LiTa₃O₂I₁₂ forms kinked strands of [Ta₃O₂I₁₀I_4/2]⁻ units which are separated by disordered lithium ions (Fig. 1 left). In Ta₂O₃I₄, each Ta⁵⁺ ion is coordinated by three oxygen and three iodide ions, forming octahedra that connect via corner-sharing oxygen atoms and edge-sharing iodides, with one terminal iodide per octahedra. This connectivity results in zig-zag layers (Fig. 1 right).

Fig. 1 Sections of the crystal structures of LiTa₃O₂I₁₂ (left) and Ta₂O₃I₄ (right). Tantalum atoms are depicted in blue, iodine pink, oxygen red and lithium grey.

Efforts to minimise the amount of these side products by adjusting reactant ratios frequently resulted in lower yields of the desired cluster compounds. Likewise, prolonging the reaction duration to steer the system toward thermodynamic equilibrium did not improve reaction selectivity or reproducibility. This suggests that the cluster formation may proceed under kinetic control, potentially influenced by the side phases. Notably, the cluster compounds were consistently found in close proximity to TaO₂I, implying that this phase could serve as a precursor in the cluster-forming process.

Comparable unconventional reduction methods with Li₂CN₂ were recently reported in the Nb–O–I system, resulting in several novel niobium oxyiodides, such as the oxygen centred Nb₄OI_12−x clusters with x = 0, 1, 2.^19,37

In extension of these findings, analogous reactions were also performed in the Ta–O–I system, yielding distinct cluster networks, referred to as o-M₅O₄I₁₁ (M = Nb, Ta), m-Ta₅O₄I₁₁, and Ta₅O₄I₁₁(TaI₅), where o and m denote the orthorhombic and monoclinic polymorphs, respectively.

The crystal structure of o-Ta₅O₄I₁₁ was solved and refined from single-crystal X-ray diffraction data in the orthorhombic space group Pmc2₁ (No. 26). The characteristic motif of o-Ta₅O₄I₁₁ is depicted in Fig. 2: The Ta₅ cluster core adopts the motif of a distorted square pyramid. At the base of the pyramid, four μ₃-oxygen atoms bridge the edges between the metal atoms. Each of the four outer metal atoms exhibits distorted octahedral coordination by oxygen and iodide atoms with additional metal-to-metal bonding with the central metal atom on top of the pyramid (Fig. 2) with intermetallic distances given in Table 1.

Fig. 2 Building block of isotypic M₅O₄I₁₁ structures (M = Nb, Ta), with the [M₅O₄] cluster core displayed for M = Ta, corresponding to o-Ta₅O₄I₁₁ (iodide atoms are shown in pink).

Table 1 Comparison of selected interatomic distances and bond angles in o-Ta₅O₄I₁₁ and Nb₅O₄I₁₁. For atom labelling, see Fig. 2

Distance/pm; ∠/°	o-Ta5O4I11	Nb5O4I11
Intra cluster
M(1)–M(2)	283.2(1)	285.2(1)
M(1)–M(3)	288.7(1)	287.8(4)
M(1)–M(4)	290.7(1)	290.4(4)
∠M(2)–M(1)–M(2)	149.3(1)	148.9(1)
∠M(3)–M(1)–M(4)	136.9(1)	131.3(1)
Inter cluster
M(2)–M(2)	435.7(1)	430.1(3)
M(3)–M(4)	388.4(1)	377.5(4)

---

The connectivity within and between clusters can be described using the notation [(Ta₅O₄)Iⁱ₂I^a₄I^a–a_10/2] where two inner (i) iodide ligands cap the edges at the apex of the distorted square-pyramidal cluster core, and four outer (a) iodides are terminally coordinated, one positioned above and three below the base of the pyramid. Additionally, ten iodide ligands are shared (10/2) between adjacent clusters, referred to as bridging outer (a–a) iodides. This notation follows the convention originally developed for describing octahedral cluster compounds.^1,7,50

Each cluster is connected to four neighbouring clusters at its corners through bridging iodide ligands, forming extended layers expanding into the ac-plane. Along the a-axis, the Ta(2) atoms are linked by two bridging iodides positioned perpendicular to an (m_x) mirror plane (Fig. 3 top). Ta(3) and Ta(4) are connected by three iodide bridges along a (2₁) screw axis that runs parallel to the c-axis, creating a tilted arrangement of clusters, as illustrated in Fig. 3, bottom. In the extended crystal structure, this leads to the formation of undulating layers which are separated by a van der Waals gap (Fig. 4).

Fig. 3 Cluster connectivities in the structure of M₅O₄I₁₁ along a (top) and c (bottom), corresponding to o-Ta₅O₄I₁₁. M atoms are depicted in blue, oxygen in red and iodine in pink.

Fig. 4 Waved layer structure of M₅O₄I₁₁, corresponding to o-Ta₅O₄I₁₁. M atoms are depicted in blue, oxygen in red and iodine in pink.

A comparison of selected interatomic distances and bond angles in o-Ta₅O₄I₁₁ and the isostructural Nb₅O₄I₁₁ is given in Table 1. As expected, given the similarity of ionic radii of niobium and tantalum,⁵¹ the shortest M–M bond lengths for both compounds fall within the same rang. However, the ∠M(3)–M(1)–M(4) bond angle, which traces the undulation of the layers along the c-axis, differs by around 6°. Furthermore, the inter cluster M(3)–M′(4) distance in this direction is 11 pm shorter in Nb₅O₄I₁₁, resulting in a slightly more pronounced corrugation of the layers.

The crystal structure of the monoclinic m-Ta₅O₄I₁₁ contains a [Ta₅O₄] cluster core similar to that found in o-Ta₅O₄I₁₁. However, the core in m-Ta₅O₄I₁₁ is less distorted due to a different spatial arrangement of the iodine atoms, leading to variations in how the cluster interconnect (Fig. 5). Each [(Ta₅O₄)Iⁱ₂I^a₅I^a–a_8/2] cluster includes two inner (i) μ₂-bridging iodides located at the base of the pyramid, five terminal outer (a) iodides each bonded to one of the five tantalum atoms towards the pyramid's apex, and eight outer (a–a) iodides that are shared between two clusters (8/2) (see Table S1 in SI for comparison of interatomic distances).

Fig. 5 Building block of the m-Ta₅O₄I₁₁ structure. Tantalum atoms are depicted in blue, oxygen atoms in red and iodine atoms in pink.

Each cluster is interconnected with four neighbouring clusters at all four corners through two iodide bridges, forming an extended layered network. In this structure, the surrounding cluster pyramids are oriented oppositely to the central one: when the central cluster points upwards, the adjacent clusters point downwards, and vice versa (Fig. 6); this alternating orientation results in a sinusoidal layering pattern. These undulating layers are further separated by a van der Waals gap (Fig. 7).

Fig. 6 Section of the m-Ta₅O₄I₁₁ structure depicting the inter-cluster connectivity within one layer. Tantalum atoms are depicted in blue, oxygen in red and iodine in pink.

Fig. 7 Sinusoidal connectivity of clusters in the layered structure of m-Ta₅O₄I₁₁. Note that there is no connectivity between adjacent layers. Tantalum atoms are depicted in blue, oxygen in red and iodine in pink.

The cluster network of Ta₅O₄I₁₁(TaI₅) is closely related to that of m-Ta₅O₄I₁₁, but with the additional incorporation of molecular [TaI₅] units. This leads to a significant transformation of the overall structural architecture. Fig. 8 shows a section of the Ta₅O₄I₁₁(TaI₅) structure (top) along with the arrangement of [Ta₅O₄I₁₁] layers (bottom left) and [TaI₅] units (bottom right). A closer comparison of the layer formation and organisation in Ta₅O₄I₁₁(TaI₅) (Fig. 8) and m-Ta₅O₄I₁₁ (Fig. 7) reveals clear structural differences, even though the cluster connectivity pattern [(Ta₅O₄)Iⁱ₂I^a₅I^a–a_8/2] remains the same in both structures.

Fig. 8 Projected crystal structure of Ta₅O₄I₁₁(TaI₅) (top) and segmentation of the structure into [Ta₅O₄I₁₁] (left) and [TaI₅], appearing as two superimposed TaI₅ units (right).

Notable differences between both structures are observed in the amplitude of the sinusoidal waves of the layers and by the directional shift of these waves relative to one another. In m-Ta₅O₄I₁₁, the clusters alternate in an up-and-down fashion, whereas in Ta₅O₄I₁₁(TaI₅), for each directional shift of the wave, two of the four adjacent clusters point in the same direction as the central cluster and two are inverted and point the other direction (compare Fig. 7 and Fig. 8). This more pronounced corrugation creates voids that accommodate [TaI₅] molecules, which contribute to the stabilisation of the structure of Ta₅O₄I₁₁(TaI₅).

This finding aligns with the observations made for the compound Nb₇S₂I₁₉ ≙ (Nb₃SI₇)₂(NbI₅), which Miller and Lin described as exhibiting a synergistic relationship between the extended framework and the enclosed [NbI₅] molecule.²⁸ In this structure, [NbI₅] molecules are packed within hexagonal channels formed between the [Nb₃SI₇]. In contrast, when [NbI₅] molecules are absent, Nb₃SI₇ adopts a van der Waals layered structure that lacks these hexagonal voids.¹⁷

As with the [NbI₅] units in Nb₇S₂I₁₉, the [TaI₅] molecules in Ta₅O₄I₁₁(TaI₅) are incorporated as trigonal bipyramidal monomers (see Fig. 9) similar to the geometries reported for TaCl₅ and TaBr₅ in the vapour phase.^52,53 However, in the solid state, TaI₅ typically crystallises as (TaI₅)₂ dimers composed of edge-sharing bioctahedra.⁵⁴

Fig. 9 Molecular [TaI₅] in the structure of Ta₅O₄I₁₁(TaI₅) with 80% displacement ellipsoids. Blue: Ta, pink: I.

The Ta–I bond lengths of molecular [TaI₅] in the structure of Ta₅O₄I₁₁(TaI₅) range from 260(1) pm to 271(1) pm, with an average of 266(1) pm. This closely matches the average Nb–I distance of 267(1) pm of the NbI₅ monomers in Nb₇S₂I₁₉, as well as the average Ta–I distance of 267(1) pm in Ta₂I₁₀, not considering edge-sharing iodides.⁵⁴

Although all M₅O₄I₁₁-based compounds share a similar penta-nuclear cluster core, they display distinct structural frameworks. These differences arise from variations in the inner (i) and outer (a) iodine ligand environments, the presence of incorporated [TaI₅] molecules, as well as the cluster connectivity (a–a).

In particular, o-Ta₅O₄I₁₁ features clusters interconnected by two and three μ₂-bridging iodides at the corners (Fig. 3), resulting in shorter inter cluster Ta–Ta distances and potentially stronger electronic interactions, since m-Ta₅O₄I₁₁ clusters are only connected by two iodide bridges at all four corners (Fig. 6). This difference is expected to influence the electronic properties of the material, particularly electrical conductivity, which is discussed later in this study.

The [Ta₅O₄] cluster motive demonstrates both structural stability and geometric flexibility, as it persists across different environments and tolerates considerable distortions. However, despite these characteristics, the o- and m-modifications of Ta₅O₄I₁₁ are not interconvertible by thermal treatment. Heating beyond 650 °C results in decomposition and the formation of TaOI₂, rather than phase transition between the two forms.

For o-M₅O₄I₁₁, the structures with niobium and tantalum are isotypic. However, no analogues of m-Ta₅O₄I₁₁ or Ta₅O₄I₁₁(TaI₅) could be observed for niobium. Instead, niobium forms distinct cluster types such as Nb₄OI_12−x, reflecting differences in cluster stability and connectivity between the two elements.^19,37

Ta₅O₄I₁₁ carries six cluster electrons, but their distribution across the four short Ta–Ta bonds is not straightforward. From a formal oxidation state perspective, the four basal tantalum atoms can be assigned to Ta⁴⁺, while the apical atom is considered more reduced, with an oxidation state of Ta³⁺. To further probe this, XPS measurements are employed in the next section.

Analysis of Ta–Ta distances provide additional insights into the presence of metal–metal bonding in o-Ta₅O₄I₁₁. The average Ta–Ta bond length is approximately 286 pm, which is slightly shorter than in other tantalum halide clusters such as Ta₆I₁₄ (∼292 pm), Ta₃SeI₇ (∼295 pm) or Ta₄SBr₁₁ (∼305 pm). This comparison supports the presence of bonding interactions between the tantalum atoms. Additional insight into electron distribution and bonding can be gained through ELF (Electron Localisation Function) analysis.

Scanning electron microscopy (SEM), energy dispersive X-ray spectroscopy (EDX) and total reflection X-ray fluorescence (TXRF)

When in contact with moist air, samples of Ta₅O₄I₁₁ show pronounced surface degradation and partial delamination (Fig. S7 in SI). To prevent this, a custom-designed vacuum transfer device was developed, enabling inert handling and transfer of samples. Details of the vessel design and operation are provided in the Experimental section and illustrated in Fig. S2 of the SI.

The scanning electron micrographs in Fig. 10 were obtained from samples transferred under inert conditions. Fig. 10 shows crystal agglomerates and surfaces morphologies of o-Ta₅O₄I₁₁ (Fig. 10a–c) and m-Ta₅O₄I₁₁ (Fig. 10d–f). The two compounds clearly exhibit different morphologies: o-Ta₅O₄I₁₁ frequently shows well defined individual layers (Fig. 10c), suggesting weaker interlayer interactions. In contrast, m-Ta₅O₄I₁₁ displays more compact, block-like surface features (Fig. 10f), indicating stronger interactions between adjacent layers.

Fig. 10 Scanning electron micrographs of o-Ta₅O₄I₁₁ crystals (a–c) and m-Ta₅O₄I₁₁ crystals (d–f), showing a layered morphology.

To verify the composition, determined by X-ray diffraction, EDX measurements of multiple crystals were performed, resulting in an average Ta : O : I ratio of 5 : 4.3(5) : 10.9(3). TXRF measurements verified the Ta : I ratio to be 5 : 11.3(2).

X-ray photoelectron spectroscopy (XPS)

Further insights on the oxidation state of tantalum in o-Ta₅O₄I₁₁ were achieved using X-ray photoelectron spectroscopy (XPS).

The Ta 4f spectrum shown in Fig. 11 can be described by two doublets (Ta 4f_7/2 and Ta 4f_5/2), which are assigned to the oxidation states of Ta⁴⁺ and Ta³⁺.^55,56 The fit parameters are summarised in Table S2 (SI). The intensity ratio of Ta⁴⁺ and Ta³⁺ components of 4.14 : 1 is in good agreement with the proposition of a Ta³⁺ at the central position on top of the distorted pyramid and four Ta⁴⁺ at the corners of the pyramid base.

Fig. 11 XPS spectrum of Ta 4f depicting Ta(IV) and Ta(III) oxidation states in a ratio of 4.14 : 1.

The survey spectrum in Fig. S8 (SI) reveals the presence of iodine and oxygen (notably, the additional carbon and silicon signals can be attributed to the carbon tape substrate) giving further evidence for the proposed structure.

Optical band gap determination from DRIFTS (Diffuse reflectance infrared fourier transform spectroscopy)

The absorption coefficient F was obtained using Kubelka–Munk analysis following:
where
is the reflectance of an infinitely thick specimen, α is the absorption and S is the scattering coefficient. For particle sizes greater than the light wavelengths measured, the scattering coefficient is understood to be approximately independent of frequency (F(R_∞) ∼ α) and therefor F(R_∞) could be understood as a “pseudo-absorbance” coefficient.^57–63

The band gap determination was performed on the DRIFTS data according to Zanata et al.⁶⁴ The energy is plotted against the absorption coefficient α and fitted with a sigmoid-Boltzmann function:

where α_min (α_max) stands for the minimum (maximum) absorption coefficient; E^Boltz₀ is the energy coordinate at which the absorption coefficient is halfway between α_min and α_max; and δE is associated with the slope of the sigmoid, indicating the energy range over which most optical transitions occur.⁶⁴

The band gap can then be calculated by following equation with n^Boltz_dir= 0.3 and n^Boltz_indir= 4.3:

E^Boltz_g= E^Boltz₀− n^Boltz_dir/indir·δE

The optical band gaps of o-Ta₅O₄I₁₁ were determined to be 0.241 eV (direct) and 0.226 eV (indirect) indicating semiconducting behaviour (see Fig. S9, SI). The band gap of m-Ta₅O₄I₁₁ could not be measured within the detectable range of the DRIFTS setup. To further assess the electronic properties, conductivity measurement were subsequently performed.

Electrical properties

We reasoned that the structural differences in o-Ta₅O₄I₁₁vs. m-Ta₅O₄I₁₁ as well as their distinct ligand environments should manifest in different electrical properties. Hence, two-point current–voltage scans of both modifications (o and m) were carried out at varying temperatures. Exemplarily results, obtained at 300 K are displayed in Fig. S3 (SI).^19,65 The measurements revealed that o-Ta₅O₄I₁₁ is roughly one order of magnitude more conductive than m-Ta₅O₄I₁₁ with typical values of 5 × 10⁻⁵ S m⁻¹ to 2 × 10⁻⁴ S m⁻¹ for o-Ta₅O₄I₁₁vs. 4 × 10⁻⁶ S m⁻¹ to 2 × 10⁻⁵ S m⁻¹ for m-Ta₅O₄I₁₁.

Using the two-point conductivities obtained at different temperatures, we arrive at the Arrhenius plot in Fig. 12, indicating temperature-activated, Arrhenius type transport for both phases, which is typical for semiconductors.^66,67 From the slopes of both plots, we calculate activation energies of E_A = 0.10 eV for o-Ta₅O₄I₁₁ and E_A = 0.40 eV for m-Ta₅O₄I₁₁, respectively. Due to the poor signal strength from m-Ta₅O₄I₁₁ at temperatures below 240 K, additional conductivities were measured above 300 K up to 350 K to improve the fit confidence. o-Ta₅O₄I₁₁ was measured between 140 K–300 K.

Fig. 12 Arrhenius-plot of the temperature-dependent conductivity between 140 K–300 K for o-Ta₅O₄I₁₁ (red) and 240 K–350 K for m-Ta₅O₄I₁₁ (blue). The inset shows the linear plot of conductivity versus temperature.

To a first order approximation, the electrical band gap of an intrinsic semiconductor can be gauged as two times the activation energy, hence 0.2 eV for o-Ta₅O₄I₁₁ and 0.8 eV for m-Ta₅O₄I₁₁, respectively.

Both crystal species exhibit low photocurrents toward optical excitation at 779 nm illuminated with 40 mW output power. o-Ta₅O₄I₁₁ showed higher photocurrents around 1 nA while m-Ta₅O₄I₁₁ exhibited around 15 pA at 300 K (Fig. S10, SI). At lower temperatures, the photocurrent decreases gradually until it is below the noise level at ∼0.1 pA.

Band structure

The spin-unpolarised electronic band structures of o-Ta₅O₄I₁₁ and m-Ta₅O₄I₁₁ were calculated using density functional theory (DFT). At the DFT level, o-Ta₅O₄I₁₁ is a band metal (Fig. 13), unlike the small-gap semiconductor observed in our optical and electrical measurements. Since the experimental gap of o-Ta₅O₄I₁₁ is small, the DFT calculations were performed with spin orbit coupling, as this could plausibly open a small gap, however this was not observed. The opening of the gap could therefore be due to electronic correlations and magnetic interactions between Ta atoms, as was determined in the related Mott insulating material Ta₄SBr₁₁.²² However, due to the large size of the unit cell of o-Ta₅O₄I₁₁, computational exploration of potential Mott behaviour was not possible in this case.

Fig. 13 Calculated electronic band structure of o-Ta₅O₄I₁₁, with bands coloured by their Ta character. Spin–orbit coupling was included in the calculation. Special points in and paths through reciprocal space were chosen following the literature.⁶⁸

In contrast, DFT calculations showed m-Ta₅O₄I₁₁ to be a semiconductor with a gap of 0.4 eV (Fig. 14), in accordance with experiments. The existence of a band gap in m-Ta₅O₄I₁₁ is therefore most likely not due to magnetic interactions or spin–orbit coupling. Significant differences in the electronic structure and chemical bonding between o-Ta₅O₄I₁₁ and m-Ta₅O₄I₁₁ can be seen. In o-Ta₅O₄I₁₁, the states near the Fermi energy have predominantly Ta d character, with several distinct band manifolds being visible. The upper manifolds are nearly flat, indicating localised electrons. In m-Ta₅O₄I₁₁, a much greater degree of hybridisation can be seen near the Fermi level, with the valence electrons having some O and I character in addition to Ta character (see Fig. S11–S14 in the SI for details). The formation of a band gap in m-Ta₅O₄I₁₁ could then be ascribed to the formation of covalent bonds between Ta and other cluster atoms, whereas Ta atoms in o-Ta₅O₄I₁₁ are more ionic. In o-Ta₅O₄I₁₁, the Ta atoms would therefore be expected to behave closer to idealised Ta³⁺ and Ta⁴⁺, which contain unpaired d electrons, leading to partially filled bands at the DFT level.

Fig. 14 Calculated electronic band structure of m-Ta₅O₄I₁₁, with bands coloured by their Ta character. Spin–orbit coupling was not included in the calculation. Special points in and paths through reciprocal space were chosen following the literature.⁶⁸

Electron localisation function (ELF)

We have also calculated from DFT the electron localisation function (ELF) of o-Ta₅O₄I₁₁ and m-Ta₅O₄I₁₁, further elucidating the electronic structure and chemical bonding in these materials.^69,70 Like the band structures, the ELFs of the two polymorphs are strikingly different. The ELFs around the Ta and O atoms in o-Ta₅O₄I₁₁ are largely spherical, indicating ionic bonding (Fig. 15). The monoclinic modification shows the ELFs of the atoms to be more distorted, as well as the presence of some interstitial electron density, which are indicators of covalent bonding. As Fig. 16 shows, this extra electron density lies in the planes between pairs of Ta atoms, corresponding to the formation of Ta–Ta bonds. This feature is absent in o-Ta₅O₄I₁₁, where there are instead nodal planes between the Ta atoms, again indicative of ionic interactions (compare Fig. S15 in SI).

Fig. 15 Calculated ELF of o-Ta₅O₄I₁₁ (top) and m-Ta₅O₄I₁₁ (bottom), showing the presence of increased covalent bonding between Ta and O in the monoclinic phase, as demonstrated by the deviation of the shapes of the ELFs from idealised ionic spheres.

Fig. 16 Calculated ELF of o-Ta₅O₄I₁₁ (top) and m-Ta₅O₄I₁₁ (bottom), showing the presence of interstitial localised electrons in the monoclinic phase, corresponding to Ta–Ta bonds. No similar feature can be seen in the orthorhombic phase.

Charge-disproportionated Ta³⁺ and Ta⁴⁺ ions, as seen in the XPS measurements of o-Ta₅O₄I₁₁ (Fig. 11), do not appear in our DFT calculations, where the ELFs and electron counts around the Ta atoms are nearly identical at each site. Therefore, this disproportionation could be driven by whichever factors which lead to the opening of the band gap and are not accounted for in our DFT calculations.⁷¹

Conclusions

Heterogeneous solid-state reactions involving an unidentified gas phase have led to the formation of the compounds Nb₅O₄I₁₁, o-Ta₅O₄I₁₁, m-Ta₅O₄I₁₁, and Ta₅O₄I₁₁(TaI₅), all featuring an [M₅O₄] cluster core. In these structures, each cluster is connected to four neighbouring clusters at the corners via two or three bridging iodide ligands, creating distinct extended layered networks. These connectivity patterns are summarised in Fig. 17. A common structural characteristic among these compounds is a sinusoidal layering pattern within the crystal lattice, with layers separated by van der Waals gaps. The presence of bridging outer ligands spatially separates adjacent clusters, resulting in semiconducting behaviour with low electrical conductivity primarily confined to the layers, particularly evident in both forms of Ta₅O₄I₁₁.

Fig. 17 Comparison of the inter cluster connectivity of o-Ta₅O₄I₁₁, m-Ta₅O₄I₁₁ and Ta₅O₄I₁₁(TaI₅).

In all compounds, the [Ta₅O₄] cluster core possesses six cluster electrons involved in metal–metal bonding, that could likely be localised along the four shorter metal–metal contacts between the pyramidal base and apex. According to a localised bonding model, the metal atoms in the basal plane may be assigned to Ta⁴⁺, while the apical metal atom corresponds to Ta³⁺ in o-Ta₅O₄I₁₁, ensuring charge neutrality. This interpretation is supported by XPS, and it is partially consistent with charge distribution predicted by DFT calculations.

Conflicts of interest

There are no conflicts to declare.

Data availability

Data are available within the article.

The data that support the findings of this study are available on request from the corresponding author, H.-J. Meyer. Computational data are availiable from DOI: 10.5281/zenodo.17424860.

Supplementary information (SI): See DOI: https://doi.org/10.1039/d5dt02097b.

CCDC 2351105 (Ta₅O₄I₁₁(TaI₅)), 2380271 (o-Ta₅O₄I₁₁), 2389419 (Ta₂O₃I₄), 2390289 (LiTa₃O₂I₁₂), 2400835 (Nb₅O₄I₁₁) and 2424946 (m-Ta₅O₄I₁₁) and contain the supplementary crystallographic data for this paper.^72a–f

Acknowledgements

Funding by the Deutsche Forschungsgemeinschaft through grant ME 914/32-1 and SCHE1905/9-1 is gratefully acknowledged. Computational resources were provided by the state of Baden-Württemberg through bwHPC and the DFG through grant no INST 40/467-1 FUGG (JUSTUS cluster). C. P. R. acknowledges support from the project FerrMion of the Ministry of Education, Youth and Sports, Czech Republic, co-funded by the European Union (CZ.02.01.01/00/22\_008/0004591). The authors thank Ms. Elke Nadler (University Tübingen) for recording SEM images and EDX data.

References

1. H. Schäfer and H.-G. v. Schnering, Angew. Chem., 1964, 76, 833–849.
2. F. A. Cotton, Q. Rev., Chem. Soc., 1966, 20, 389–401.
3. P. Lemoine, J.-F. Halet and S. Cordier, in Ligated Transition Metal Clusters in Solid-state Chemistry : The legacy of Marcel Sergent, ed. J.-F. Halet, Springer International Publishing, Cham, 2019, pp. 143–190,  DOI:10.1007/430_2019_39.
4. L. Pauling, The nature of the chemical bond and the structure of molecules and crystals, Cornell University Press, 1949.
5. D. Bauer and H.-G. Schnering, Z. Anorg. Allg. Chem., 1968, 361, 259–276.
6. R. E. McCarley, J. Lewis and M. L. H. Green, Philos. Trans. R. Soc. London, Ser. A, 1982, 308, 141–157.
7. A. Simon, H.-G. Schnering, H. Wöhrle and H. Schäfer, Z. Anorg. Allg. Chem., 1965, 339, 155–170.
8. D. G. Blight and D. L. Kepert, Phys. Rev. Lett., 1971, 27, 504–504.
9. S. L. Benjamin, Y.-P. Chang, A. L. Hector, M. Jura, W. Levason, G. Reid and G. Stenning, Dalton Trans., 2016, 45, 8192–8200.
10. D. R. Taylor, J. C. Calabrese and E. M. Larsen, Inorg. Chem., 1977, 16, 721–722.
11. P. J. Kuhn and R. E. McCarley, Inorg. Chem., 1965, 4, 1482–1486.
12. H.-G. v. Schnering, D. Vu, S.-L. Jin and K. Peters, Z. Kristallogr. - New Cryst. Struct., 1999, 214, 15–16.
13. G. Meyer, R. Wiglusz, I. Pantenburg and A.-V. Mudring, Z. Anorg. Allg. Chem., 2008, 634, 825–828.
14. F. Grahlow, M. Ströbele and H.-J. Meyer, ICSD Commun., 2022 DOI:10.25505/fiz.icsd.cc2cy0dr.
15. B. Bajan and H.-J. Meyer, Z. Kristallogr. – Cryst. Mater., 1996, 211, 818–818.
16. M. Smith and G. J. Miller, J. Solid State Chem., 1998, 140, 226–232.
17. G. V. Khvorykh, A. V. Shevelkov, V. A. Dolgikh and B. A. Popovkin, J. Solid State Chem., 1995, 120, 311–315.
18. F. Grahlow, F. Strauß, M. Scheele, M. Ströbele, A. Carta, S. F. Weber, S. Kroeker, C. P. Romao and H.-J. Meyer, Phys. Chem. Chem. Phys., 2024, 26, 11789–11797.
19. J. Beitlberger, M. Ströbele, F. Strauß, M. Scheele, C. P. Romao and H.-J. Meyer, Eur. J. Inorg. Chem., 2024, 27, e202400329.
20. H. B. Yaich, J. C. Jegaden, M. Potel, M. Sergent, A. K. Rastogi and R. Tournier, J. Less-Common Met., 1984, 102, 9–22.
21. M. N. Sokolov, A. L. Gushchin, P. A. Abramov, A. V. Virovets, E. V. Peresypkina, S. G. Kozlova, B. A. Kolesov, C. Vicent and V. P. Fedin, Inorg. Chem., 2005, 44, 8756–8761.
22. F. Grahlow, F. Strauß, P. Schmidt, J. Valenta, M. Ströbele, M. Scheele, C. P. Romao and H.-J. Meyer, Inorg. Chem., 2024, 63, 19717–19727.
23. F. A. Cotton and M. Shang, J. Am. Chem. Soc., 1990, 112, 1584–1590.
24. J. L. Seela, J. C. Huffman and G. Christou, J. Chem. Soc., Chem. Commun., 1987, 1258–1260.
25. M. Ströbele, O. Oeckler, M. Thelen, R. F. Fink, A. Krishnamurthy, S. Kroeker and H.-J. Meyer, Inorg. Chem., 2022, 61, 17599–17608.
26. A. Broll, A. Simon, H.-G. von Schnering and H. Schäfer, Z. Anorg. Allg. Chem., 1969, 367, 1–18.
27. G. J. Miller, J. Alloys Compd., 1995, 217, 5–12.
28. G. J. Miller and J. Lin, Angew. Chem., Int. Ed. Engl., 1994, 33, 334–336.
29. J. Zhang and J. D. Corbett, Inorg. Chem., 1995, 34, 1652–1656.
30. K. Habermehl, A.-V. Mudring and G. Meyer, Eur. J. Inorg. Chem., 2010, 2010, 4075–4078.
31. T. C. Zietlow and H. B. Gray, Inorg. Chem., 1986, 25, 631–634.
32. J. D. Franolic, J. R. Long and R. Holm, J. Am. Chem. Soc., 1995, 117, 8139–8153.
33. M. Ströbele and H.-J. Meyer, Dalton Trans., 2019, 48, 1547–1561.
34. S. Hartwig and H. Hillebrecht, Z. Naturforsch., B:J. Chem. Sci., 2007, 62, 1543–1548.
35. M. Ruck, Acta Crystallogr., Sect. C:Cryst. Struct. Commun., 1995, 51, 1960–1962.
36. S. Hartwig and H. Hillebrecht, Z. Anorg. Allg. Chem., 2008, 634, 115–120.
37. J. Beitlberger, M. Martin, M. Scheele, P. Schmidt, M. Stroebele and H.-J. Meyer, Dalton Trans., 2025, 54, 5486–5493.
38. J. Beitlberger, M. Ströbele, P. Schmidt, C. P. Romao and H.-J. Meyer, Dalton Trans., 2025, 54, 14376–14383.
39. G. Brauer, Handbuch der präparativen anorganischen Chemie, Enke, 1975.
40. M. Ströbele, E. Bayat and H.-J. Meyer, Inorg. Chem., 2024, 63, 16565–16572.
41. Y. Yao, J. Liu, Z. Wang, S. Yao and F. Du, Micron, 2024, 187, 103720.
42. M. J. Verstraete, J. Abreu, G. E. Allemand, B. Amadon, G. Antonius, M. Azizi, L. Baguet, C. Barat, L. Bastogne and R. Bejaud, arXiv, 2025, preprint, arXiv:2507.08578,  DOI:10.48550/arXiv.2507.08578.
43. J. P. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett., 1996, 77, 3865.
44. S. Grimme, J. Antony, S. Ehrlich and H. Krieg, J. Phys. Chem., 2010, 132(15), 154104.
45. M. Torrent, N. A. W. Holzwarth, F. Jollet, D. Harris, N. Lepley and X. Xu, Comput. Phys. Commun., 2010, 181(11), 1862–1867.
46. H. J. Monkhorst and J. D. Pack, Phys. Rev. B, 1976, 13, 5188.
47. M. Methfessel and A. Paxton, Phys. Rev. B:Condens. Matter Mater. Phys., 1989, 40, 3616.
48. G. Brauer and A. Simon, Handbook of Inorganic Synthesis, 1985, vol. 5, pp. 1563–1564.
49. K. Gibson, M. Ströbele, B. Blaschkowski, J. Glaser, M. Weisser, R. Srinivasan, H.-J. Kolb and H.-J. Meyer, Z. Anorg. Allg. Chem., 2003, 629, 1863–1870.
50. B. Baján and H.-J. Meyer, Z. Naturforsch., B:J. Chem. Sci., 1995, 50, 1373–1376.
51. R. Shannon, Acta Crystallogr., Sect. A, 1976, 32, 751–767.
52. H. A. Skinner and L. E. Sutton, Trans. Faraday Soc., 1940, 35, 668–680.
53. G. L. Carlson, Spectrochim. Acta, 1963, 19, 1291–1307.
54. K. Habermehl, I. Pantenburg, P. Held and G. Meyer, Z. Anorg. Allg. Chem., 2008, 634, 829–831.
55. I. Perez, V. Sosa, F. G. Perera, J. T. E. Galindo, J. L. Enríquez-Carrejo and P. G. Mani-González, Vacuum, 2019, 165, 274–282.
56. Y. Abbas, Y.-R. Jeon, A. S. Sokolov, S. Kim, B. Ku and C. Choi, Sci. Rep., 2018, 8, 1228.
57. W. W. Wendlandt and H. G. Hecht, Reflectance spectroscopy, Interscience, New York, 1966.
58. P. Kubelka, J. Opt. Soc. Am., 1948, 38, 448–457.
59. P. Kubelka and F. Munk, Z. Tech. Phys., 1931, 12, 259–274.
60. G. Kortüm and H. Koffer, Ber. Bunsen-Ges., 1963, 67, 67–75.
61. G. Kortum and J. E. Lohr, Reflectance Spectroscopy, Springer My Copy, UK, 1969.
62. A. B. Murphy, Sol. Energy Mater. Sol. Cells, 2007, 91, 1326–1337.
63. R. López and R. Gómez, J. Sol-Gel Sci. Technol., 2012, 61, 1–7.
64. A. R. Zanatta, Sci. Rep., 2019, 9, 11225.
65. F. Fetzer, A. Maier, M. Hodas, O. Geladari, K. Braun, A. J. Meixner, F. Schreiber, A. Schnepf and M. Scheele, Nat. Commun., 2020, 11, 6188.
66. A. M. Al-Fa'ouri, O. A. Lafi, H. H. Abu-Safe and M. Abu-Kharma, Arabian J. Chem., 2023, 16, 104535.
67. M. M. A. Imran and O. A. Lafi, Phys. B, 2013, 410, 201–205.
68. Y. Hinuma, G. Pizzi, Y. Kumagai, F. Oba and I. Tanaka, Comput. Mater. Sci., 2017, 128, 140–184.
69. A. D. Becke and K. E. Edgecombe, J. Phys. Chem., 1990, 92, 5397–5403.
70. A. Savin, R. Nesper, S. Wengert and T. F. Fässler, Angew. Chem., Int. Ed. Engl., 1997, 36, 1808–1832.
71. A. Walsh, A. A. Sokol, J. Buckeridge, D. O. Scanlon and C. R. A. Catlow, J. Phys. Chem. Lett., 2017, 8, 2074–2075.
72. (a) CCDC 2351105: Experimental Crystal Structure Determination, 2025,  DOI:10.5517/ccdc.csd.cc2jxj33; (b) CCDC 2380271: Experimental Crystal Structure Determination, 2025,  DOI:10.5517/ccdc.csd.cc2kwvy8; (c) CCDC 2389419: Experimental Crystal Structure Determination, 2025,  DOI:10.5517/ccdc.csd.cc2l6d18; (d) CCDC 2390289: Experimental Crystal Structure Determination, 2025,  DOI:10.5517/ccdc.csd.cc2l7938; (e) CCDC 2400835: Experimental Crystal Structure Determination, 2025,  DOI:10.5517/ccdc.csd.cc2ll89s; (f) CCDC 2424946: Experimental Crystal Structure Determination, 2025,  DOI:10.5517/ccdc.csd.cc2mdc2h.

---