Zn₂GeO₄:Cr,Mn phosphor for visible to NIR luminescence-based applications – a theoretical and experimental investigation

Abstract

Zn₂GeO₄ (ZGeO) phosphor exhibits a wide bandgap energy, making it highly suitable for luminescence-based applications spanning the entire electromagnetic spectrum, from ultraviolet (UV) to near-infrared (NIR) wavelengths. In this work, nominally undoped and Cr-doped ZGeO (ZGeO:Cr) were prepared by solid-state reaction. X-ray diffraction and Raman spectroscopy confirmed the monophasic willemite crystalline structure of the ZGeO and ZGeO:Cr samples, and X-ray photoelectron spectroscopy corroborated the identification of Zn, Ge, O, and Cr elements. X-ray photoemission indicates an insulator character for ZGeO and ZGeO:Cr. The oxide host revealed a direct bandgap energy of 4.77 eV, assessed by room temperature absorption, in line with the density functional theory (DFT) calculations that predicted 4.8 eV at the Γ point of the first Brillouin zone. Cr⁴⁺ was found to occupy distorted tetracoordinated Ge⁴⁺ sites with C₁ symmetry, in agreement with the measured unfolded ³A₂ → ³T₁, ³T₂ intraionic absorption. Er³⁺ and Mn²⁺ trace impurities occupy distorted Zn²⁺ sites, also with C₁ symmetry. A Mn²⁺–O²⁻ charge transfer state, placed 0.8 eV below the conduction band minimum, was identified by absorption measurements. In addition, as calculated by DFT, Cr³⁺ impurities exhibit lower energy formation when placed in distorted interstitial octahedral Zn–Ge and Zn–Zn rings with C₃ and S₆ symmetry, respectively. The identified site locations were found to be compatible with the measured unfolded ⁴A₂ → ⁴T₂, ⁴T₁ intraionic absorption, with the ions subject to intermediate/low crystalline field strengths. The emission of the samples is dominated by structureless broad bands spanning from the visible to mid-infrared. A broad bluish-white emission results from an overlap of emitting centers related to both intrinsic defects (generating a blue emission) and trace impurities of Mn²⁺ (generating a green emission). Besides, Cr³⁺ and Cr⁴⁺ were found to coexist in the oxide host, and their multiple-site occupation is responsible for the observed broad emission bands in the NIR-I (700–950 nm) and NIR-II (1000–1700 nm) spectral regions, opening the way to the exploitation of the NIR luminescence for light-based devices.

---

1. Introduction

As other binary and ternary wide-bandgap (WBG) semiconductor oxides (E_g > 3.5 eV), Zn₂GeO₄ (ZGeO) appears as a suitable material for a widespread range of technological applications, including ultraviolet (UV) photodetectors and emitters,^1–3 high-power electronics,^4,5 photocatalysis,^6–9 gas sensing,^10,11 bioimaging and biosensing,^12–15 solid-state phosphors,^13,16–22 anticounterfeiting and encryption,^4,9,23–25 among others. In addition, and more recently, WBG oxides have been investigated due to the interest in the development of tunable steady-state near-infrared (NIR) ultra-wideband light sources (NIR pc-LEDs) for biomedical applications and quantitative chemical analysis.^26–36

With a room temperature (RT) bandgap energy reported between 4.4 eV and 4.9 eV,^17,37–42 ZGeO exhibits a willemite-type crystalline structure with a rhombohedral unit cell (space group R3̄, No. 148), with Ge and Zn ions occupying a Wyckoff position 18f,^38,42–48 as is shown in Fig. 1. The crystalline cell is built by ZnO₄ and GeO₄ tetrahedra with cornered-shared oxygen, aligned parallel to the c-axis, and with lattice parameters a = b = 14.231 Å and c = 9.530 Å, α = β = 90° and γ = 120°.^48–50 Zn²⁺ ions possess two inequivalent sites with reported average Zn–O bond lengths of 1.963 Å (Zn1) and 1.967 Å (Zn2). The remaining average bond length is 1.749 Å for Ge–O.^20,47,50 Moreover, the lattice exhibits four-membered rings that encompass alternating corner-sharing ZnO₄ and GeO₄ tetrahedra and two types of six-membered rings: one consisting of corner-sharing Zn(1)O₄ tetrahedra (radius of 196 pm) and the other one consisting of alternating corner-sharing Zn(2)O₄ and GeO₄ tetrahedra (radius of 277 pm).^{20,40,47–53}

Fig. 1 Structural model of zinc germanate obtained with VESTA using the CIF file: 1549041.⁵⁴

Density functional theory (DFT) calculations predict that ZGeO has a direct bandgap, with the extremes of the conduction (CB) and valence (VB) bands occurring in the Γ point of the first Brillouin zone.^38,53 The large energy separation between the CB and VB is prone to incorporating electronic energy levels inside the forbidden bandgap, generated either by native defects or by extrinsic dopants. J. Dolado et al.³⁸ and S. K. Gupta et al.⁴ recently calculated the formation energies of intrinsic defects using DFT, showing that among the vacancy defects, the oxygen vacancy (V_O) is the most favorable, while the germanium vacancy (V_Ge) is the least. In addition, interstitial zinc, Zn_i, in positions 1 (at the center of the Zn ring) and 2 (at the center of the Zn–Ge ring), exhibit the lowest formation energies, and antisite defects (Zn_Ge, Ge_Zn) possess lower formation energy than the V_O.^4,38 Furthermore, the authors^4,38 calculated the ZGeO electronic structure and intrinsic defect energy levels, establishing a correlation between the intrinsic defects and the luminescence observed in ZGeO. A schematic representation of the estimated energies for all the intrinsic defects obtained by the authors^4,38 is shown in Fig. 2.

Fig. 2 Schematic representation of the estimated energies for the ZGeO intrinsic defects based on the DFT calculations from ref. 4 and 38.

Defects of intrinsic and extrinsic nature are typically responsible for the observed luminescence in undoped and intentionally doped ZGeO. Typically, using photons with energy equal to or higher than the oxide bandgap as excitation, the cryogenic and RT photoluminescence (PL) of the nominally undoped material is dominated by broad unstructured luminescence bands in the UV (UVL) and blue spectral regions.^{4,18,22,38,40,42,55–58} A UVL band peaked around 340–360 nm observed upon above bandgap excitation was recently reported on ZGeO microwires synthesized by a catalyst-free thermal evaporation method.^38,58 The band was found to be an overlap of emitting centers, presenting either wide or narrow UVL bands, with the latter (peaked between 370–380 nm) revealing a two-component character and being also excited with below bandgap photon energies.^38,58 Following the authors' DFT calculations,^4,38 such high energy emission bands were attributed to radiative recombination involving electrons in either the V_O or Zn_i donor levels with holes at the VB.^38,58 In ZGeO, besides the UVL bands, a bluish-white band (BL) with a maximum located between 440 and 470 nm (depending on growth/synthesis parameters) was also ascribed to a recombination process involving intrinsic defects.^{4,18,22,40,55–57,59–61} The complex nature of this emission is still under debate in the literature. In particular, Z. Liu et al.⁴⁰ proposed that the BL was related to a donor–acceptor pair (DAP) recombination mechanism also involving the V_O and Zn_i as donors and ionized V_Ge and V_Zn as acceptors. After performing thermal annealing treatments in different atmospheres, Z. Gu et al.⁵⁵ ascribed the recombination in this spectral region to a defect that involves the V_O. J. Dolado et al.⁵⁹ suggested that the V_Zn may be involved in the origin of the BL. In addition, the BL was considered as an overlap of several emitting centers, as proposed recently by M. Tinoco et al.⁵⁷ and earlier by Li et al.⁶⁰ These authors attributed the BL to germanium and oxygen-related defects. Moreover, the analysis of the recombination kinetics of the BL has been described as controversial in the literature. While Z. Liu et al.⁴⁰ found a non-exponential decay behavior for the proposed DAP luminescence, M. Shang et al.¹⁸ reported a single exponential lifetime of 6.4 ns for the same emission, meaning that a complete model for the BL recombination still needs to be addressed.

The luminescence of ZGeO is highly sensitive to various factors, including synthesis/growth method, nature and purity of the precursors, as well as the temperature and atmospheric conditions during the production. In fact, in nominally undoped hosts or intentionally Mn-doped samples, the BL and an asymmetric broad green luminescence (GL) band with a maximum at ∼534 nm/2.32 eV have been observed^{4,20–22,38,41,46,48,51,55,59,61,62} when the material is excited with energies above the bandgap. The GL was also observed when ZGeO is excited either by high-energy electrons, resulting in green cathodoluminescence,^37,63 or via carrier injection, enabling the observation of green electroluminescence.^10,64 Despite the knowledge of the GL in ZGeO since at least the 1970s, its nature continues to be a matter of discussion.^{4,20,21,40,53}

For instance, based on their DFT calculations, S. K. Gupta et al.⁴ recently reported that Zn_i in positions 1 and 2 are involved in the recombination processes of both the GL and BL emissions located at around 515 nm/2.41 eV. As mentioned above, besides the intrinsic defects, trace impurities (e.g., transition metal and/or lanthanide ions), even in nominally undoped samples, can originate optically active defects with electronic energy levels inside the material bandgap, resulting in luminescence bands over a wide range of the electromagnetic spectrum and with different spectral shapes. In the case of ZGeO, there is an enlarged consensus on the attribution of the ∼534 nm/2.32 eV GL band to the ⁴T₁(⁴G) → ⁶A₁(⁶S) transition of Mn²⁺ (3d⁵) impurities in substitutional Zn²⁺ sites.^{18,20–22,37,41,47,56,61,63,65}

Electron paramagnetic resonance (EPR) measurements confirmed that Mn²⁺ occupies the two inequivalent Zn²⁺ sites (with one of the sites slightly preferred over the other one), as expected from the similarity of charge and ionic radius (0.60 Å and 0.66 Å for Zn²⁺ and Mn²⁺ in tetrahedral coordination, respectively).^20,21,26,48 On the other hand, the low energy shift of the Mn²⁺ GL observed with increasing the Mn/Zn content, as reported by Feldman et al.,⁶⁵ may be attributed to variations in crystal field strength arising from the distinct local environments of the two Zn sites.

Additionally, spin exchange interactions between Mn²⁺–Mn²⁺ pairs,^22,66 as was previously observed in Mn-doped Zn₂SiO₄, which is isostructural with ZGeO⁴³ may also contribute to this shift. Furthermore, although some authors³⁷ have suggested that an emission band in the orange/red region (with maximum ca. 650 nm/1.91 eV) could be due to the presence of Mn⁴⁺ in an octahedral position in the ZGeO lattice, J. Q. Hu et al.⁵⁰ recently reported that in ZGeO:Mn samples there is a reduction of Mn⁴⁺ to Mn²⁺ due to charge compensation effects.

As aforementioned, WBG oxides doped with chromium ions are of particular interest for the development of NIR pc-LEDs.^{26–36,67,68} For example, Cr⁴⁺ (3d²), which tends to occupy tetrahedral sites, is known to exhibit broad absorption bands in the spectral region where the main commercially available pump lasers operate, as is, for instance, the case of the chromium ion in Zn₂SiO₄, Mg₂SiO₄, and Ca₂SiO₄ crystals.^{26,27,69–77} In these oxides, when Cr⁴⁺ replaces Si⁴⁺ in distorted tetrahedral sites, strong intraionic polarized absorption bands take place between 500 nm and 800 nm and from 800 nm to 1200 nm due to the spin-allowed ³A₂ → ³T₁ (³F), ³T₂ (³F) transitions,^{26,27,69–77} respectively. At RT, the broad emission of Cr⁴⁺ due to the ³T₂ → ³A₂ transition typically occurs in the NIR-II spectral region (1000–1700 nm), with a maximum ranging from 1000 nm to 1500 nm, depending on the crystal field strength.^{26,27,69–77} As mentioned, due to their broad pump absorption bands, Cr-ions have been extensively used as laser active media in several hosts for tunable solid-state lasers,⁷⁵ and more recently, Cr-doped oxides are at the forefront of device development for applications in the NIR-I (700 to 950 nm) and NIR-II spectral regions.^{26,34,36,78,79}

Considering the above, it is of great interest to explore the role of Cr ions in the ZGeO host, which, to our knowledge, has not yet been properly addressed. To the best of our knowledge, no red/NIR luminescence of Cr-activated ZGeO has been reported to date. As for the case of Mn²⁺ in Zn²⁺ sites, Cr⁴⁺ and Ge⁴⁺ exhibit the same valence and similar ionic radii in fourfold coordination (0.41 Å and 0.39 Å, respectively²⁷) and, therefore, in Cr-doped ZGeO, it is expected that Cr⁴⁺ in Ge⁴⁺ sites could be optically activated, as will be further discussed in this work. On the other hand, Cr³⁺ (3d³) has an ionic radius of 0.46 Å in four-fold coordination,⁶⁸ and 0.61 Å in a six-fold coordination (close to the Zn²⁺ in four-fold coordination), which, according to Y. Cong et al.,⁵⁶ is expected to substitute Ge⁴⁺ in ZGeO. According to the Tanabe and Sugano diagrams⁸⁰ for this electronic configuration, a ⁴A₂(⁴F) ground state is expected for the ion in octahedral coordination and, depending on the crystal field strength, the first excited state is either the ²E(²G) (strong fields) or ⁴T₂(⁴F) (weaker fields). Therefore, the luminescence from the first excited level to the ground state is mainly characterized by narrow emission lines, the well-established R-lines, due to the ²E → ⁴A₂ transition, or by a broad emission band due to the ⁴T₂ → ⁴A₂ transition.^{30,31,74,76,77,81,82} In addition, ⁴T₂ → ⁴T₁ NIR-I luminescence due to Cr³⁺ in a tetrahedral coordination environment has also been reported in several oxide hosts, in which the ion experiences weaker crystal field strengths than those in octahedral sites.^68,82–84 However, as mentioned by Shang et al.,⁶⁸ Jahn–Teller effect on the ground state and/or inherent lattice distortion enhances the nonradiative transition probability between the aforementioned electronic states, which may be the origin of the literature ambiguities concerning Cr³⁺ luminescence on T_d site symmetry, with widely accepted Cr³⁺ activators tending to occupy O_h (or near) lattice sites.^68,85

In this work, nominally undoped (with Mn and Cr as trace impurities) and intentionally Cr-doped ZGeO were synthesized by conventional high-temperature solid-state reaction, and an in-depth optical characterization of these oxide materials was performed. Apart from the BL and GL bands, unreported emission bands in the NIR-I and NIR-II spectral regions were observed. Based on the absorption, temperature, and excitation energy dependence on luminescence, the emission bands were assigned to Cr³⁺ in distinct distorted octahedral sites and to Cr⁴⁺ in tetrahedral Ge⁴⁺ sites, respectively, which agrees well with the performed structural and elemental analysis. The experimental evidence is supported by first-principles calculations on the formation energies of Cr in ZGeO, their site occupation, and valence states. Moreover, from photoluminescence excitation (PLE) measurements, the preferential population mechanisms of the emitting centers were established, highlighting the role of intrinsic defects, Mn²⁺–O²⁻ charge transfer state (CTS), and Cr³⁺/Cr⁴⁺ excitation bands. Therefore, the present theoretical and experimental investigation elucidates the mechanisms of the broad emission bands covering the visible, NIR-I, and NIR-II spectral regions in ZGeO:Mn,Cr materials, opening the way to future developments of this emerging oxide host for luminescence-based applications.

2. Methods

2.1. Material synthesis

The nominally undoped Zn₂GeO₄ and Cr-doped Zn₂GeO₄ (ZGeO and ZGeO:Cr) were synthesized via a conventional, scalable solid-state reaction method. Zinc oxide (ZnO, 99.99% trace metal basis, with Mn and Cr as trace contaminants, confirmed by inductively coupled plasma optical emission spectroscopy (ICP–OES)) and germanium(IV) oxide (GeO₂, 99.999% trace metal basis) were acquired from Aldrich, whereas chromium(III) oxide (Cr₂O₃, 99% metal basis) was purchased from Alfa Aesar, to be used as reagent powders. Ethanol absolute anhydrous used in the synthesis was acquired from Carlo Erba. The synthesis method was based on our previous work,³⁰ adapted for this material. For the undoped Zn₂GeO₄, the following precursors were mixed: ZnO (3.00 g, 36.86 mmol) and GeO₂ (1.93 g, 18.43 mmol). In the case of Zn₂GeO₄ doped with chromium, a stoichiometric ratio of Cr/Ge = 0.01 (1.0 mol%) was used, mixing the following precursors: ZnO (2.99 g, 36.74 mmol), GeO₂ (1.92 g, 18.37 mmol), and Cr₂O₃ (0.014 g, 0.092 mmol). The precursor powders were first mixed in an agate ball mill, with a small amount of ethanol covering the powders, for 4 h. To evaporate all the ethanol, this mixture was then dried in the oven for 7 h at 70 °C. Afterward, the dried powder was homogenized manually and pressed into 13 mm diameter pellets with a 1.5 t uniaxial press. The pellets were then fired at 900 °C for 2 h and then ground and reshaped into pellets with the same press. These pellets suffered a sintering treatment in air at 1350 °C for 36 h in a ThermoLab oven (heating rate of 5 °C min⁻¹).

2.2. Structural, morphological, elemental, and optical characterization

The structural properties of the synthesized materials were determined by X-ray diffraction (XRD) and Raman spectroscopy. The XRD patterns were acquired with a Philips X’Pert MPD employed in a θ–2θ mode with Cu K_α1 radiation (λ = 1.54056 Å). The Raman spectra were measured at RT by exciting the samples with the 442 nm line of a cw He–Cd laser (Kimmon IK Series). The scattered light was analyzed using a Horiba Jobin Yvon HR800 spectrometer equipped with a 600 grooves mm⁻¹ grating. The measurements were performed with a 50× objective (numerical aperture, NA, of 0.5) in the backscattering mode.

For the morphological and elemental composition characterization, the scanning electron microscopy (SEM) images were obtained using a field-emission Schottky-gun (FEG-SEM) SEM Hitachi SU70, operated at 15 kV, and equipped with an AXS Bruker energy dispersive X-ray fluorescence (EDX) detector. Additionally, for the identification of trace impurities, ICP-OES analysis was performed using a Horiba JobinYvon Activa M instrument.

X-ray photoelectron spectroscopy (XPS) measurements were carried out using a Kratos Axis Supra instrument with monochromated Al K_α radiation (1486.6 eV). Detailed scans were obtained at an X-ray power of 150 W and a pass energy of 10 eV. Electron flood gun charge neutralization was used, and all peaks were referenced to C 1s at 284.8 eV. Data analysis was conducted with CasaXPS software.

Diffuse reflectance (DR) of the undoped and intentionally doped pellets was measured at RT in the wavelength range 190–1600 nm using a UV/Visible GBC Cintra-303 equipment in the absorbance mode, with a bandwidth of 2 nm, a speed velocity of 100 nm min⁻¹, and a data pitch of approximately 0.520.

Fourier-transform infrared (FTIR) absorption measurements were conducted in an FTIR Bruker Tensor 27 using a diamond ATR Golden Gate, in absorbance mode, with a resolution of 2 cm⁻¹ and 256 scans in the range 4000–400 cm⁻¹.

The RT steady-state PL and PLE data were obtained in a Fluorolog®-3 from Horiba Scientific. A continuous Xe lamp of 450 W was used as the excitation source, coupled with an excitation monochromator, Gemini 180, with a diffraction grating of 1200 grooves mm⁻¹ and a focal distance of 180 mm. The scanning emission monochromator is an iHR550 with a diffraction grating of 1200 grooves mm⁻¹ and a focal distance of 550 mm. The detector used is an R928P PMT, which is sensitive to wavelengths in a range of 200 to 850 nm. For the RT time-resolved PL (TRPL) measurements, the same Fluorolog®-3 system was used, in this case with the aid of a pulsed Xe lamp (operating at up to 25 Hz, pulse lamp width of 3 μs) coupled to the same monochromator. The measurement conditions were set to variable windows (duration of signal acquisition) of 1, 2, and 5 ms, and several time delays after flash were employed (from 0.08 to 30 ms). For the lifetime measurements, the samples were excited with appropriate excitation wavelengths to maximize the PL signal, and the time evolution of the PL intensity was assessed.

For the RT and low temperature PL data acquisition in the UV-visible range, a dispersive system SPEX 1074 Czerny–Turner monochromator with a diffraction grating of 1200 grooves mm⁻¹ and a focal distance of 1 m was used. This equipment was coupled to a RCA C31034 photomultiplier detector. The excitation source used was the 325 nm line of a cw He–Cd laser. The samples were placed in a cold finger of a closed-cycle He cryostat under vacuum conditions, which allowed the temperature to reach 15 K. Temperature-dependent PL spectra were performed between 15 K and the RT. The spectra acquired in both systems were corrected for the optics of the system and the detector response. For quantitative analysis, the recorded data were corrected for the Jacobian transformation.

Vis-NIR/MIR PL measurements were also carried out in the range of 22 000–5000 cm⁻¹ (455–2000 nm) using a Bruker Vertex 80v Fourier Transform Infrared (FTIR) spectrometer equipped with InGaAs and Si detectors. To remove the alignment laser line at 632.8 nm, a long-pass optical filter with a cut-off wavelength of 700 nm was used. The samples were inserted and mounted on an OptistatCF (Oxford Instruments) cryostat, and the spectra were measured from 70 K to RT. As excitation sources, the 457.9 nm (DPSS (CVI Melles Griot)), 532 nm (MGL-F-532 (CNI)) and 808 nm (Roithener LaserTechnik, RLDB808-500-5) laser lines were used.

2.3. Theoretical calculation methods

The DFT calculations were performed using the VASP code⁸⁶ and considering spin polarization. The exchange–correlation functional used was HSE06⁸⁷, a hybrid functional, with the fraction of exact exchange set to 0.35 to provide a better agreement between the calculated and measured bandgaps, 4.80 and 4.77 eV, respectively. Slightly different setups have been used in other works. For example, Dolado et al.³⁸ used the HSE06 functional as well, but with a mixing fraction of 0.25, corresponding to a bandgap of 4.62 eV, while Gupta et al.⁴ used the PBE0 functional instead, yielding a bandgap of 4.39 eV.

The simulations were run on a Zn₂GeO₄ unit cell, with periodic boundary conditions in all cartesian directions. The explicitly treated electrons were 3d¹⁰ 4s² for Zn, 3d¹⁰ 4s² 4p² for Ge, 2s² 2p⁴ for O, 3p⁶ 3d⁵ 4s¹ for Cr, 5p⁶ 5d¹ 6s² for Er, and 3p⁶ 3d⁶ 4s¹ for Mn, with the remaining electrons included in the frozen core and tackled using the projector augmented-wave method.⁸⁸ The Brillouin zone was sampled using only the Gamma k-point, due to the extreme computational cost of hybrid DFT calculations when using the plane-wave formalism.⁸⁹ The plane-wave basis had a cutoff value of 415 eV, and the force (energy) convergence criterion for structural (electronic) relaxation was 0.01 eV Å⁻¹ (10⁻⁵ eV).

3. Results and discussion

3.1. Structure, morphology, and elemental analysis

Fig. 3 shows the X-ray diffraction patterns, Raman and FTIR absorption spectra, SEM images, and EDX spectra of undoped Zn₂GeO₄ (ZGeO) and Zn₂GeO₄ doped with Cr (1.0 mol%) (ZGeO:Cr). The produced samples exhibit prominent and narrow diffraction maxima (Fig. 3a), corresponding to a single phase willemite ZGeO structure with random crystallographic orientations of crystallites (JCPD #00-011-0687), representing a polycrystalline nature. No extra diffraction maxima were observed in the XRD patterns for the doped ZGeO:Cr sample, indicating that no additional crystalline phases were formed. The lattice parameters of the rhombohedral cell were measured as a = 14.23 Å and c = 9.52 Å, which are in line with the reported values.^48–50 The calculated lattice parameters of ZGeO are a = b = 14.31 Å and c = 9.57 Å, which differ by less than 1% from the measured ones.

Fig. 3 (a) X-ray diffractograms of the ZGeO and ZGeO:Cr samples, as well as the standard pattern of ZGeO (JCPD #00-001-0687). (b) RT Raman and (c) FTIR absorbance spectra for both samples. SEM images of the surface of (d) ZGeO and (e) ZGeO:Cr pellets with the same magnification. (f) EDX spectra for both samples (*Si, Al, and Ca arise from the accessories used in the equipment).

Zn₂GeO₄ has S₆ symmetry and the following lattice sites: all Ge⁴⁺ sites are equivalent and tetrahedral (GeO₄); there are two inequivalent Zn²⁺ sites, both tetrahedral (ZnO₄); and there are four inequivalent O²⁻ sites, all trigonal planar, with O bonded to two Zn²⁺ and one Ge⁴⁺.

The crystallinity of the samples was also assessed by Raman spectroscopy, as shown in Fig. 3b. The spectra of ZGeO and ZGeO:Cr are similar, with the vibrational modes located at the same frequency with almost the same relative intensity and full-width half maximum (FWHM). The strongest resonance occurs at 802 cm⁻¹ (A⁽²⁾_g) and has been assigned to the stretching vibration O-Ge–O bonds in GeO₄ tetrahedra.^{42,58,62,90–93} Besides the observation of a vibrational mode at 861 cm⁻¹, two other maxima were measured at 746 cm⁻¹ (A⁽¹⁾_g) and 779° cm⁻¹ (E⁽⁴⁾_g), which are associated with Ge–O–Zn symmetric and asymmetric vibrations, respectively.^{42,58,62,90–93} The spectral shape of this peak is compatible with the presence of another peak at 751 cm⁻¹ (E⁽³⁾_g) which, although unresolved, cannot be definitively ruled out. The low-frequency modes (<600 cm⁻¹) corresponding to O–Zn–O and Zn–O–Ge bending and stretching vibrations^91,93 were also identified in the studied samples.

The vibrational host absorption is also shown in Fig. 3c, which presents the FTIR absorption spectra for both undoped and doped samples, in which eight modes were identified. In line with what has been reported in the literature,^22,93,94 the absorption bands at 450–650 cm⁻¹ and 700–900 cm⁻¹ were assigned to the vibrational modes of ZnO₄ and GeO₄ tetrahedra, respectively.

Fig. 3d and e show SEM images of the ZGeO and ZGeO:Cr samples, respectively, with the same magnification. The microscopy images display a similar cohesive and dense morphology for both samples. Additionally, Fig. 3f shows the EDX spectra with the corresponding indexation of elements. EDX analysis indicates the presence and homogeneous distribution of the constituent chemical elements, Zn, Ge, and O. The Cr element was also identified in the doped sample, as were impurities of other elements such as Si, Al, and Ca, which arise from the accessories used for SEM-EDX observation.

A more detailed chemical analysis of the phosphors was performed by XPS. The survey XPS spectra (Fig. 4a) exhibit the presence of Ge, Zn, O, and C in both samples; additionally, in the doped sample, Cr was also identified. The spectra were corrected to adventitious carbon (C 1s) at 284.8 eV. For the assessment of the XPS detailed spectra of the elements, Fig. 4b–d, a Shirley background was applied for fitting purposes.

Fig. 4 (a) Survey XPS spectra of undoped and doped ZGeO; (b) O 1s, (c) Zn 2p, (d) Ge 3d, and (e) Cr 3p fitted spectra. Open circles – acquired data; yellow solid line – Shirley background; red solid lines – fitting data (envelope); dashed lines – deconvoluted peaks; blue solid line – manual polynomial background.

Fig. 4b shows the O 1s spectrum. The emission was deconvoluted into two peaks at 530.6 and 531.4 eV for ZGeO, as well as 530.9 and 531.9 eV for ZGeO:Cr, which were assigned to lattice oxygen and surface-adsorbed oxygen.^90,95,96 The observed increase in the binding energies from the non-doped sample to the doped one can be related to the electron charge compensation⁹⁷ due to the introduction of chromium in the matrix, for instance, when Cr⁴⁺ replaces a Ge⁴⁺. The fine XPS spectra (Fig. 4c) show one of the two Zn 2p peaks characteristic of ZGeO and ZGeO:Cr at 1021.1 and 1021.4 eV, which are in agreement with the ones reported previously.^90,95 Moreover, the second Zn 2p peak was identified in both samples (data not shown), with an energy separation of approximately 23.1 eV between the two Zn 2p peaks, consistent with previously reported values for this host material.^90,95 The fine XPS spectra displayed in Fig. 4d reveal the presence of the Ge 3d peak located at 31.6 eV and 32.0 eV for ZGeO and ZGeO:Cr, respectively, which are close to the values reported elsewhere for this host material.^90,95,98

In the undoped sample, no evidence of Cr was found. However, as expected, in the ZGeO:Cr sample, the Cr 3p was identified at circa 43.5 eV, confirming the presence of this dopant (Fig. 4e).⁹⁹ The region around 590 to 575 eV, corresponding to Cr 2p (2p_1/2 and 2p_3/2 orbitals of chromium), has been used to identify the presence of trivalent and tetravalent states of chromium in similar phosphors (without Zn), such as Mg₁₄Ge₅O₂₄:Cr³⁺, Cr⁴⁺ (ref. 100) and BaTiO₃:Cr³⁺, Cr⁴⁺.⁹⁶ However, Zn LMM Auger also occurs at this binding energy,¹⁰¹ so an accurate confirmation of the valence state of chromium is hindered for ZGeO and ZGeO:Cr. Nevertheless, as will be discussed later, Cr³⁺ and Cr⁴⁺, as well as Mn²⁺ (not detected by XPS), were identified by optical measurements in undoped and doped samples.

To quantify the elements present in each sample, the Cr 3p emission was used, but as it is shallow, a custom polynomial background correction was applied in this case, intended to mimic the background of the undoped sample. Additionally, the Zn 3p emission was used due to its kinetic energy being closer to that of the other cations than in the case of Zn 2p. The results can be found in Table 1, where a ratio of Zn : Ge of 2 : 1 for the ZGeO sample was determined, whereas the doped sample showed a zinc deficiency (Zn : Ge = 1.82 : 1) and a chromium fraction of 0.01 related to Ge, which is in agreement with the nominal percentage of Cr used for sample preparation.

Table 1 XPS quantification for undoped and doped samples

Name	Position (eV)	ZGeO	ZGeO : Cr	x/Ge	Area	At%	x/Ge
		Area	At%
a The C percentage was removed to determine the ratio between an element (x) and Ge.
O 1s	531.0	16 318.58	40.83	4.81	16 990.55	47.42	4.35
C 1s	285.0	4767.34	33.47		2781.51	21.78
Ge 3p	124.0	4854.36	8.48	1	5589.31	10.89	1
Zn 3p	89.0	8282.86	17.22	2.03	8525.18	19.77	1.82
Cr 3p	43.5	—	—	—	17.61	0.14	0.01

---

3.2. Optical characterization

3.2.1. RT absorption. Fig. 5a shows the RT absorption spectra obtained by applying the Kubelka–Munk function^102,103 to the measured DR spectra of ZGeO and ZGeO:Cr samples. For comparison purposes, the spectra were normalized to the absorption maxima. The oxide band edge absorption has a maximum at 260 nm/4.77 eV, corresponding to the RT host bandgap. The measured value agrees well with the theoretically calculated bandgap using the HSE06 functional, which indicates a 4.80 eV direct bandgap material at the Γ point of the first Brillouin zone, in line with the previous reported DFT calculation.^38,53 In addition, the spectra of both samples are dominated by broad absorption bands in a wide spectral range, from UV to NIR. From shorter to longer wavelengths, shoulders/maxima are observed at ∼313 nm/3.96 eV, 415 nm/3.0 eV, 455 nm/2.73 eV, 584 nm/2.12 eV, 636 nm/1.95 eV, 694 nm/1.79 eV, 767 nm/1.62 eV, 898 nm/1.38 eV, and 1129 nm/1.1 eV. The similar spectral shapes of the absorption bands observed in both samples indicate that the absorbing defects are already present in the non-intentionally doped ZGeO, likely due to the same intrinsic defects and trace impurities (e.g., manganese and chromium). Particularly in the case of the sample intentionally doped with chromium ions, the relative intensity of the absorption bands in the UV, visible, and NIR regions is seen to increase, indicating that chromium ions should be involved in the absorption processes.

Fig. 5 (a) RT normalized absorption spectra of ZGeO and ZGeO:Cr samples. (b)–(d) adapted Tanabe and Sugano diagrams for the Cr⁴⁺ 3d² electronic configuration in T_d site symmetry, Cr³⁺ 3d³ electronic configuration in O_h site symmetry (for simplicity, the gerade (g) sub-indices were omitted), and Mn²⁺ 3d⁵ electronic configuration in O_h site symmetry, respectively. (b) to (d) Adapted from ref. 104 and 105.

For the ZGeO lattice, and due to the charge and size of the Cr⁴⁺ that closely match that of Ge⁴⁺, it is expected that Cr⁴⁺ will enter the lattice by substituting the fourfold coordinated germanium cations. In this valence state, and by the Tanabe and Sugano diagram for the 3d² electron configuration in tetrahedral symmetry (Fig. 5b), the ground ³F free ion term splits, in increasing energies, into ³A₂, ³T₂, and ³T₁ electronic states, while the ¹D term is split into the ¹E, ¹T₂ levels, and the ³P excited state into the ³T₁ level.¹⁰⁴ Accordingly, spin and electric dipole-allowed transitions are expected to be observed from the ³A₂ electronic ground state to the ³T₁ (³F) and ³T₁ (³P) excited levels, and if the relaxation of the selection rules occurs due to symmetry lowering, the ³A₂ →³T₂ (³F) absorption can also be observed to some extent at lower energies. This is, for instance, the case corresponding to the ion in a distorted tetrahedral environment (e.g., a trigonal distortion (C_3v) followed by a C_s component), which is known to split the triplet states into three orbital components. Such symmetry lowering results in pronounced polarized absorption transitions as were already reported for the case of the Mg₂SiO₄ and Ca₂GeO₄ forsterite and cunyite olivine crystals, in which the ³A₂ →³T₂ (³F) absorption transitions are seen to occur between 800 nm/1.55 eV and 1100 nm/1.13 eV and the ³A₂ →³T₁ (³F) between 500 nm/2.48 eV and 800 nm/1.55 eV.^{71–73,106,107}

Theoretical calculations performed on the Cr⁴⁺ site location on ZGeO support a similar interpretation. Indeed, although we mentioned that all Ge sites are equivalent and tetrahedral (GeO₄), these are not regular tetrahedra (with T_d symmetry). The Ge–O bond lengths and angles are not all equal, reducing the symmetry of Ge sites to the lowest one, C₁. Therefore, any substitutional atom at a Ge site does not see a local T_d symmetry, but C₁ instead, even before any lattice distortions occur because of the substitution by a distinct element. According to our DFT calculations, replacing a Ge⁴⁺ with a Cr⁴⁺ causes nearly no lattice strain, indicating very high structural stability of this defect, leading to a fundamental electronic state placed at 1.25 eV above the VB maximum (Fig. S1 of SI). Two of the bond lengths with the neighboring O atoms do not change (up to 0.01 Å), the third one increases by 0.01 Å, and the fourth decreases by 0.01 Å (Table 2), so the average Cr–O distance is the same as the average Ge–O distance.

Table 2 Calculated bond lengths

Ge^4+–O (Å)	Cr^4+–O (Å)
1.76	1.75
1.76	1.76
1.76	1.76
1.76	1.77

---

Therefore, from the spectral shape of the unpolarized RT absorption spectra shown in Fig. 5a, and by analogy with the above-mentioned behavior of the chromium ions in the tetravalent charge state in other hosts, we attribute the longer wavelength absorption bands to Cr⁴⁺ in distorted tetrahedral sites on the ZGeO lattice. The bands at 636 nm/1.95 eV, 694 nm/1.79 eV, and 767 nm/1.62 eV correspond to the split ³A₂ → ³T₁ (³F) transition and the weaker bands at 898 nm/1.38 eV and 1129 nm/1.1 eV are due to the ³A₂ → ³T₂ (³F) transition of Cr⁴⁺ in C₁ lower site symmetry.

For the case of the intentionally Cr-doped sample (ZGeO:Cr), besides the absorption of the Cr⁴⁺, an increase in the relative intensity of the absorption bands located at shorter wavelengths (between 360 nm and 640 nm) is noticed, which are likely to be related to the introduced dopant. A careful inspection allows the identification of the absorption bands already present in the ZGeO sample, in which chromium was found as a trace impurity. Considering this fact, the absorption bands in this spectral region were attributed to the trivalent chromium in the ZGeO host. In this charge state, Cr³⁺ is usually found in octahedral coordination in several hosts, which is energetically favorable.^{74,100,108,109} To evaluate the preferred ion site location in the ZGeO host, we performed theoretical calculations, also considering the possibility of the ion in four-fold coordination. As in the case of Ge, the ZnO₄ tetrahedra are not regular. There are two inequivalent Zn sites. In one of them, the calculated Zn–O bond lengths range from 1.95 to 1.97 Å, whereas in the second one, these lengths are between 1.95 and 1.99 Å. As before, a substitutional Cr³⁺ at a Ge or Zn site does not see a local T_d symmetry, but C₁ instead. Moreover, none of the substitutional sites of the Zn₂GeO₄ lattice is octahedral. However, it is possible for Cr³⁺ to bind to six O atoms in a distorted octahedron configuration: in one of the two interstitial positions, at the center of a Zn–Ge ring, or of a Zn ring. The tetrahedral sites are almost regularly tetrahedral. As mentioned, the bond lengths are the same up to 0.01 Å. The perfect tetrahedral angle is 109.5°, and the O–Ge–O angles in Zn₂GeO₄ have values very close to this (Fig. 6a). For the octahedral sites, two possibilities must be considered: (i) the center of a Zn–Ge ring and (ii) the center of a Zn ring. Fig. 6b shows a Cr³⁺ at the interstitial site at the center of a Zn–Ge ring (C₃ symmetry), viewed along the c-direction. All bond lengths are equal (2.16 Å) up to differences of 0.01 Å, but the O–Cr–O bond angles are not 90° and 180°, as in a regular octahedron. In a regular octahedron, all the angles would be 90°, as shown in Fig. 6b. Instead, they are all lower than that, so this corresponds to a squeezed octahedron, with a side view represented in Fig. 6c. A similar conclusion can be drawn for the case of an interstitial Cr³⁺ at the center of the Zn ring (S₆ symmetry), as shown in Fig. 6d, where all bond lengths are similar (2.14 Å), but the bond angles are 75.9°.

Fig. 6 (a) GeO₄ tetrahedral site in ZGeO with bond lengths mentioned in Table 2 and O–Ge–O close to the 109.5° expected for a perfect tetrahedron. (b) Interstitial Cr³⁺ site at the center of a Zn–Ge ring (C₃ symmetry), viewed along the c-direction. (c) Side view of the octahedron shown in (b). (d) Interstitial Cr³⁺ site at the center of a Zn ring (S₆ symmetry), viewed along the c-direction.

To find the thermodynamically preferred site of Cr³⁺, we calculated its formation energy on each site, E_form, using the following expression:^89,110

E_form = E_defect − E_pristine + E_removed − E_added + q(E_VBM + E_Fermi) (1)

where E_defect and E_pristine are the energies of the defective and pristine cell, respectively, E_removed and E_added are the energies of the atoms that were removed and added to generate the defect, q is the charge state of the defect, E_VBM corresponds to the energy of the valence band maximum and E_Fermi to the Fermi energy that can be anywhere inside the bandgap, between 0 and 4.80 eV in this case. The energies of single atoms were calculated assuming the atoms are available in their standard phases (bulk solids for Zn, Ge, or Cr and an O₂ molecule for O). The formation energy diagram of each Cr³⁺ site considered is shown in Fig. 7. The calculations evidence that (i) when the Fermi level is between the VB maximum and the center of the energy bandgap, Cr³⁺ prefers to occupy interstitial octahedral positions, at the center of a Zn–Ge ring (C₃ site symmetry) or a Zn ring (S₆), with similar formation energy; (ii) when the Fermi level is between the center of the energy band gap and the CB minimum, the formation energies of all the defects considered becomes very similar, so all of them should be equally likely to exist in the material; and (iii) when the Fermi level is very close to the CB minimum, Cr³⁺_Ge (C₁ site symmetry) becomes the most stable position for Cr³⁺.

Fig. 7 Formation energy diagram for Cr³⁺ sites in Zn₂GeO₄.

Hence, as the investigated samples evidenced a substantial isolating behavior as assessed by XPS measurements (Fig. S2), the Fermi level energy should be placed near the middle of the bandgap energy, meaning that Cr³⁺ ions are likely to preferentially occupy the two distorted octahedral interstitial sites at Zn–Ge and Zn rings, adding occupied states near 0.6 eV above the VB maximum and unoccupied states near the CB minimum (Fig. S1). A similar trend for interstitial Cr³⁺ was recently reported for the case of isostructural Zn₂SiO₄.¹⁰⁸ The free ion terms of the Cr³⁺ (3d³) split in the cubic field are illustrated by the Tanabe and Sugano diagram of Fig. 5c. The ⁴F ground state unfolds into the ⁴A₂, ⁴T₂, and ⁴T₁ electronic states, the excited ⁴P term to the ⁴T₁, and the ²G state to the ²E, ²T₁, ²T₂ and ²A₁. Under this site symmetry, Cr³⁺ is known to give rise to strong absorption bands in the visible spectral range in several hosts,^{74,100,108,109} mainly due to the spin-allowed and parity-forbidden ⁴A₂ → ⁴T₂, ⁴T₁ (⁴F) transitions. The spin and parity forbidden ⁴A₂ → ²E, ²T₁ (²G) may appear as very narrow absorption lines overlapped with the ⁴A₂ → ⁴T₂ (⁴F) absorption band.¹⁰⁵ From Fig. 5a, an increase in the relative intensity of the visible and UV absorption peaked at 584 nm/2.12 eV, 455 nm/2.73 eV, and 415 nm/3.0 eV is noticeable, which corresponds to the typical spectral region of the ⁴A₂ → ⁴T₂, ⁴T₁ (⁴F) transitions of Cr³⁺ ions in intermediate crystal field values. Therefore, and by comparison with the theoretical calculations, we assign the high energy absorption bands to the trivalent chromium in octahedral distorted sites with C₃ and S₆ symmetry in the ZGeO lattice. It should be emphasized that the low energy Cr³⁺ absorption transition partially overlaps with the absorption of Cr⁴⁺ on distorted tetrahedral sites (C₁). As a first conclusion, we can confirm that the measured RT absorption is consistent with the simultaneous presence of the trivalent and tetravalent chromium charge states in the ZGeO lattice.

From the precedent analysis, the average crystal field strength (Dq) felt by chromium ions in the tetra and trivalent charge states, when located in a distorted tetrahedral or octahedral environment, can be estimated.^104,111 Considering the mean energy value of the ³A₂ → ³T₂, ³T₁ (³F) transitions of the Cr⁴⁺ and ⁴A₂ → ⁴T₂, ⁴T₁ (⁴F) transitions of Cr³⁺, and the corresponding energy differences,^104,111Dq/B values of ∼2.34 and ∼2.44 were estimated for Cr⁴⁺ and Cr³⁺, with B Racah parameters of ∼475 cm⁻¹ and ∼700 cm⁻¹, respectively. These average values suggest that the intermediate crystal field strength of Cr³⁺ ions is close to the crossing point between the ⁴T₂ (⁴F) and ²E (²G) levels, impacting the expected observed luminescence, as will be further discussed.

Regarding the absorption band at 313 nm/3.96 eV, its corresponding signature is less evident than the previous ones, and several hypotheses should be considered. From the estimated Dq/B values for Cr⁴⁺, and considering the Tanabe and Sugano diagram (Fig. 5b), a transition from the ground ³A₂ (³F) to the excited ³T₁ (³P) state can be excluded, as a high Dq/B value (∼3) would be required. In addition, and using a similar argument, a lower Dq/B value (∼2) would be necessary to observe the ⁴A₂ (⁴F) → ⁴T₁ (⁴P) transition of Cr³⁺. Therefore, it is unlikely that the absorption band at 313 nm/3.96 eV could be assigned to internal transitions of the Cr⁴⁺ or Cr³⁺. As stated in the introductory section, a wide variety of defects are expected to be present in both samples, as is the case of intrinsic defects and other trace impurities. As mentioned before, DFT calculations^4,38 place one of the V_O energy levels close to the energies of the observed absorption band. Therefore, we cannot exclude that the 313 nm/3.96 eV absorption band could be due to such a defect. On the other hand, the presence of Mn²⁺ (3d⁵) as a common trace impurity in ZGeO could also lead to absorption bands at the mentioned higher energies, as discussed in the next section.

3.2.2. Photoluminescence UV/vis/NIR-I spectral regions.
3.2.2.1. Blue and green emission. Fig. 8a to d show the RT PL/PLE spectra of the ZGeO and ZGeO:Cr samples, as well as the PL spectra obtained at 70 K in the NIR-I spectral region. By using either above or below-bandgap excitation, the RT luminescence is dominated by wide and structureless emission bands. A broad BL with a maximum at 480 nm/2.58 eV dominates the spectrum when the ZGeO sample is preferentially excited via band-to-band absorption. The BL spectral shape is similar to the ones reported by other authors,^{4,18,22,40,55,57,59–61,112} which is ascribed to the recombination of overlapped emitting centers related to the presence of intrinsic and extrinsic defects. Fig. 8e compares the steady-state RT PL spectrum of the ZGeO sample with those measured under transient conditions for a fixed sample window and an increased sample delay (SD). As seen, by using excitation photons with energy equal to or higher than the bandgap, the BL is still observed even for longer delay times (e.g., 30 ms), meaning that the recombination process in the studied sample is of the order of tens of ms rather than in the ns scale as previously reported by M. Shang et al.¹⁸ for a DAP recombination. In addition, no significant shifts on the band maximum were observed with increasing delay times as expected for a DAP transition, as was pointed out by Z. Liu et al.⁴⁰ However, the comparison of the spectral shape of the steady-state PL and TRPL spectra indicates that the BL is an overlap of emitting centers as proposed recently by M. Tinoco et al.⁵⁷ and M. P. Dias et al.,¹¹² involving, in the samples studied here, the BL and the Mn²⁺ GL. Indeed, and as shown in Fig. 8e, besides the BL, the asymmetric GL peaked at 534 nm/2.32 eV is detected, which is ascribed to the ⁴T₁→ ⁶A₁ transition of Mn²⁺ trace impurities, which is a transition that is forbidden by selection rules (we notice that the ⁶A₁→ ⁴T₁ transition was not observed in the absorption spectra (Fig. 5a) due to its low oscillator strength). DFT calculations indicate that when Mn²⁺ occupies the two distorted tetrahedral Zn sites with C₁ symmetry (Fig. S3), occupied localized states close to the valence band maximum are added, as shown in Fig. S1. PLE spectra monitored at the BL and GL (Fig. 8a) allow us to conclude that both emissions are preferentially excited by photons with energies higher than or equal to the ZGeO bandgap.

Fig. 8 RT absorption and PLE spectra monitored at the BL, GL, and NIR-I band maxima for (a) ZGeO and (b) ZGeO:Cr samples. (c) RT and 70 K PL spectra obtained with above and below bandgap excitation for the ZGeO and (d) ZGeO:Cr samples. (e) Comparison between the steady-state and time-resolved PL for the ZGeO sample, as described in the text. (f) Rise and decay times for the BL and GL observed in the undoped and Cr-doped samples acquired with different photon excitation (symbols: experimental data; solid lines: fits to eqn (2)). All spectra are normalized to the intensity maximum.

In addition, the Mn²⁺ GL is also populated by lower energy photons, namely via the excitation band centered at ∼313 nm/3.96 eV, which coincides with the unassigned defect absorption band mentioned above when discussing Fig. 5a. Although this absorption band is close in energy to one of the theoretically DFT-predicted oxygen vacancy donor energy levels,³⁸ a more likely explanation is that the absorption band arises from the Mn²⁺–O²⁻ CTS, as reported by others,^22,41,46,50 with the CTS energy level placed ca. 0.8 eV below the conduction band. Therefore, and as mentioned, for observing the Mn²⁺ GL, two main population paths should be considered: (i) pumping the samples with above bandgap energy photons, and (ii) exciting the samples with subgap energy photons coincident with the CTS absorption.

The interplay between the two overlapped emitting centers (BL and Mn²⁺ GL) can be further visualized in Fig. 9 and in the multimedia file in the SI. Here, upon continuous band-to-band excitation, the RT Mn²⁺ GL of ZGeO can be seen with the naked eye for a few seconds, after which the BL recombination prevails. A similar observation was also recently reported by Yue et al., who assumed both the BL and GL as self-activated DAP recombination involving intrinsic defects.⁶¹ The identified behavior indicates that carrier trapping/de-trapping mechanisms and distinct feeding times of the defect emitting levels from which the GL and BL originate should be present in the studied samples. Such findings are corroborated by the measured transient behavior of the luminescence at RT (Fig. 8f) upon band-to-band (260 nm) or sub-bandgap (310 nm) photon excitation. As seen, before the decay, a rise time of the BL and GL intensities occurs, meaning that distinct population pathways are responsible for feeding the BL and Mn²⁺ GL ⁴T₁ emitting levels, respectively. The temporal dependence of the emission intensity I(t) shown in Fig. 8f was fitted considering that I(t) is proportional to the number of electrons in the emitting level,¹¹³ which, in our case, follows the approach described by V. Lojpur et al.,¹¹⁴

where A and B are intensity normalization parameters and C is a constant related to fast and slow populations of the defects emitting levels. τ_r and τ_d stand for the rise and decay times, respectively. The best fits of the experimental data using above and below bandgap excitation are indicated in Table 3 for both samples.

Fig. 9 Frames of the multimedia file showing the color shifting emission upon continuous band-to-band photon excitation (254 nm) for the undoped (left) and doped (right) samples at different times.

Table 3 Rise (τ_r) and decay times (τ_d1, τ_d2) estimated for the BL, ⁴T₁ → ⁶A₁ Mn²⁺ GL, and RL/NIR-I (τ_d3, τ_d4) by using band-to-band (τ_exc = 260 nm) and below bandgap (τ_exc = 310 nm) photon excitation

Emission band		BL	GL	RL
Sample		τ exc = 260 nm	τ exc = 310 nm	τ exc = 310 nm
ZGeO	τ r (ms)	0.044	0.237	—
ZGeO:Cr		0.014	0.335	—
ZGeO	τ d1 (ms)	14.6	—	—
ZGeO:Cr		13.8	—	—
ZGeO	τ d2 (ms)	2.72	3.61	—
ZGeO:Cr		1.67	3.36	—
ZGeO	τ d3 (ms)	—	—	1.25
ZGeO:Cr		—	—	1.73
ZGeO	τ d4 (ms)	—	—	3.98
ZGeO:Cr		—	—	4.45

---

When exciting the samples at the Mn²⁺–O²⁻ CTS absorption band (310 nm/4.00 eV), a single exponential decay of ca. 3.6 ms is measured for the spin and parity forbidden ⁴T₁ → ⁶A₁ transition of Mn²⁺ GL for the ZGeO sample. This intrinsic radiative decay is slightly shortened when the band-to-band excitation is performed and monitored at 480 nm, suggesting that additional energy transfer processes may occur, as expected from the identification of spectral overlap between the GL emission and Cr ions absorption (Fig. 5a). In addition, and as expected, the BL presents a slower radiative decay of 14.6 ms, in line with the observed transient PL (Fig. 8e). Moreover, it is important to mention that the rise times measured for the BL and GL under the different excitation conditions differ by almost one order of magnitude. These findings allow us to hypothesize that, upon band-to-band excitation, nonradiative carrier relaxation takes place from the host to higher excited defect levels (e.g., Mn²⁺–O²⁻ CTS and/or other defects), slowly populating the Mn^{2+ 4}T₁ emitting level, leading to the observed GL. In addition, competitive nonradiative processes through the crossover of the defect levels and/or energy transfer between other intrinsic/extrinsic defects should arise, explaining the de-trapping process that leads to the change of GL to BL under continuous band-to-band excitation, as shown in Fig. 9 and in the multimedia file in SI. From our data and following the DFT calculations performed by Gupta et al.⁴ and Dolado et al.,³⁸ it is hypothesized that the mentioned electronic states crossover may involve other intrinsic defects, e.g. Zn_i, which are the native defects with lower formation energy, and the estimated location of their energy levels might justify the observation of the BL. Therefore, our data are consistent with a description of the broad bluish-white emission band as an overlap of the intraionic GL luminescence of Mn²⁺ in Zn²⁺ sites with the BL likely related to Zn_i intrinsic defects.

For the case of the doped ZGeO:Cr sample, as shown in Fig. 8b and d, both BL and Mn²⁺ GL are observed, although with distinct relative intensities compared with the nominally undoped ZGeO. In addition, and despite the overlapped emissions, for this sample, only the BL is seen with the naked eye upon continuous band-to-band excitation, as visualized in Fig. 9 (multimedia file in SI). Also, as shown in Fig. 8f, the luminescence intensity decay of the Mn²⁺ GL for the ZGeO:Cr sample is shortened when a comparison is made with the non-intentionally doped sample. As previously stated, energy transfer processes are expected to occur in line with the increase in the relative intensity of the Cr-related absorption, likely explaining the decrease in the measured lifetime. The same tendency is observed for the BL, showing that the populations of the GL and BL are strongly conditioned by trapping/de-trapping and energy transfer phenomena arising from the presence of the intrinsic and extrinsic defects.

Temperature-dependent PL measurements between 15 K and RT were also performed on ZGeO and ZGeO:Cr samples. With 325 nm photon excitation, which corresponds to an excitation on the Mn²⁺–O²⁻ CTS defect absorption (Fig. 8a and b), the PL intensity of the intraionic ⁴T₁ → ⁶A₁ transition of Mn²⁺ trace impurities decreases with increasing temperature for both samples, with different thermal quenching factors, as shown in Fig. S4. In particular, 30% and 19% of its original intensity at 15 K were found at RT, for ZGeO and ZGeO:Cr samples, respectively. Under steady-state conditions, the quenching of the PL intensity due to any thermally activated process depends on the rate equations that describe the population and depopulation of the emitting excited state. Considering negligible nonradiative processes at low temperatures, the radiative quantum efficiency can be defined as η = k_rad/(k_rad − k_nrad) where k_rad,nrad are the radiative (k_rad = 1/τ_rad) and nonradiative (k_nrad = 1/τ_nrad) transition probabilities. Assuming a classical Mott law, the nonradiative transition probability can be written as k_nrad = A exp(−E_a/k_BT) and, therefore, the temperature dependence of the Mn²⁺ spectrally integrated GL intensity can be well accounted for,

I(T) = I₀[1 + C exp(−E_a/k_BT)]⁻¹ (3)

where I₀ is the intensity at 0 K, C is a temperature-independent frequency factor that accounts for the degeneracy of the states, k_B is the Boltzmann constant, T the absolute temperature, and E_a the activation energy that accounts for the nonradiative deexcitation process. Using eqn (3), the best fit for the measured experimental values of ZGeO:Cr sample was achieved considering an activation energy of 13.6 ± 2.3 meV, as indicated in Fig. S4. This value is similar to the one reported by M. P. Dias et al.¹⁰⁰ for the same emission detected on ZGeO nanorods. Contrary to the ZGeO:Cr, the integrated intensity of the GL Mn²⁺ emission in the undoped ZGeO evidences a steeper decrease at low temperatures, described by the same model and with an activation energy of ca. 3 meV (Fig. S4). The largest reduction of the GL PL intensity between 15 K and 100 K reveals additional nonradiative process contributions, even at low temperatures. The observed behavior suggests that energy transfer between the Mn²⁺ sensitizer and activator species can occur. By comparing Fig. 5 and 8, both samples exhibit Mn²⁺ GL that partially overlaps with the Cr⁴⁺ and Cr³⁺ absorption bands, indicating that energy transfer between these ions is favored. In addition, as will be shown in the next subsection, for the undoped ZGeO sample, Er³⁺ was identified as a trace impurity and, as for Mn²⁺, Er³⁺ occupies tetrahedral distorted Zn²⁺ sites (Fig. S3) with C₁ symmetry. Due to its ladder scheme levels, the 4f¹¹ lanthanide ion can emit/absorb in the visible and infrared spectral regions. As shown in Fig. 10a, the transition between the first excited and ground state multiplets, ⁴I_13/2 → ⁴I_15/2, was found to occur at ca. 1.54 μm/0.8 eV.¹¹⁵ To mention a few, besides the ⁴I_15/2 ground state multiplet, the ion excited multiplets in increasing energy are ⁴I_13/2, ⁴I_11/2, ⁴I_9/2, ⁴F_9/2, ⁴S_3/2, ²H_11/2, ⁴F_7/2, ⁴F_5/2, ⁴F_3/2, with the latter known to absorb in the blue and green spectral regions.¹¹⁵ In particular, the energy location of ⁴S_3/2 and ²H_11/2 match the spectral region of the Mn²⁺ GL (Fig. S5), indicating that further energy transfer/cross-relaxation effects may arise between the Mn²⁺ activator and the Er³⁺ sensitizer, which explains either the distinct thermal quenching process of the Mn²⁺ GL, as well as the color shifting observed upon band-to-band excitation in the ZGeO sample.

Fig. 10 RT and 70 K PL in deep-red/IR region obtained with 457.9, 532, and 808 nm excitation wavelengths for (a) ZGeO and (b) ZGeO:Cr. Figures (c)–(e) represent the same data in three different wavelength ranges to better visualize the different Cr centers. The spectra were vertically shifted for clarity. All spectra were normalized to the intensity maximum.

3.2.2.2. Deep-red, NIR-I and NIR-II luminescence. Besides the already discussed BL and Mn²⁺ GL, the ZGeO and ZGeO:Cr samples also exhibit broad deep-red (RL)/NIR-I and NIR-II emission bands, as shown in Fig. 8c and d and 10a and b (see also Fig. S6), due to the multiple optical centers of Cr³⁺ and Cr⁴⁺ in C₃/S₆ and C₁ sites symmetry, as we will discuss in this subsection.

The peak position of the high energy emission band strongly depends on temperature, excitation energy, and Cr³⁺ amount (Fig. 10c). As The RL/NIR-I is preferentially excited by photons with an energy that matches the defect absorption band, i.e., the Mn²⁺–O²⁻ CTS, as well as by intraionic Cr³⁺ absorption (Fig. 8a and b), the luminescence could likely be assigned to the Cr³⁺ ions in line with what is observed in other hosts, even knowing that the emission is strongly dependent on the crystal field strength.^{30,31,74,76,77,81,105,111,116–119} Typically, in octahedral coordination, the Tanabe and Sugano energy levels diagram for the 3d³ electron configuration (Fig. 5c) allows the identification of the first excited electronic state of Cr³⁺ as ²E or ⁴T₂, for high or low crystal field strengths, respectively. In addition, Fig. 5c shows that the energy location of the ²E state is almost insensitive to an increase in the strength of the crystal field, while the energy of the ⁴T₂ and ⁴T₁ (⁴F) excited states increases with increasing Dq/B. As stated in the Introduction, in several hosts and when in octahedral sites, the intraionic Cr³⁺ luminescence to the ⁴A₂ ground state is characterized by: (i) well-defined zero phonon lines (ZPL), the so-called R-lines, and by rich vibrational sidebands for the case of the ²E→⁴A₂ transition, which preferentially occurs when the ²E state is placed at lower energies than the ⁴T₂, i.e., when the ion is placed in sites with high crystalline field strengths;^{30,74,81,105,118,119} or (ii) by broadband recombination corresponding to the ⁴T₂ →⁴A₂ transition, usually observed when the ⁴T₂ is at lower energy than the ²E, i.e., when the ion is under a low crystalline field strength environment.^{31,74,76,77,105,111,116,117} Furthermore, we note that from our DFT calculations, for an insulating ZGeO, Cr³⁺ is expected to occupy two interstitial sites. This suggests that at least two Cr³⁺ centers, located in distorted octahedral sites (C₃ and S₆ symmetries), may be optically activated. From the absorption measurements, we have identified that the ions are, on average, under an intermediate crystal field, meaning that the luminescence could arise from the components of the ²E and ⁴T₂ excited states, upon symmetry lowering (a doublet and a singlet for the case of the triplet state in the case of a trigonal/rhombohedral distortion; or a fully degeneracy splitting in singlet states if a further orthorhombic distortion is considered). These assumptions allow us to explain the observed RL/NIR-I luminescence measured with two distinct spectroscopic systems (see experimental details) and with different excitation wavelengths, as shown in Fig. 8c, d and 10c (and Fig. S6). For the case of the ZGeO sample, and as displayed in Fig. 10c, the 70 K PL spectra evidence structureless broad emission bands with maxima at 812 nm/1.53 eV, 800 nm/1.55 eV, and a short wavelength shoulder at ca. 740 nm/1.67 eV when the excitation is performed with 325 nm, 457.9 nm, and 532 nm photons. The absence of lines/structure at this temperature may indicate that: (i) if the transitions are due to the ²E→⁴A₂, the ZPLs are already quenched at 70 K; (ii) the emission could preferentially arise from the split ⁴T₂→⁴A₂, which usually gives rise to a broad unstructured band.

As the bands’ peak position depends on the excitation wavelength, one cannot discard the hypothesis that overlapped distinct Cr³⁺ emitting centers could occur. Therefore, for the ZGeO sample, in the 70 K spectra, we designated by Cr³⁺(1) the center with maximum at 812 nm/1.53 eV preferentially excited with 457.9 nm, and by Cr³⁺(2) the one with maximum at 800 nm/1.55 eV, that is 23 meV apart from the former and is preferentially observed with 532 nm photons. In addition, a third center can be distinguished for the case of the ZGeO:Cr sample (Cr³⁺(3)). The latter exhibits a similar peak position to Cr³⁺(1) and is detected with photon excitation at 325 nm, 457.9 nm, and 532 nm.

This site-selective wavelength excitation 70 K PL allows us to conclude that at least two optically active Cr³⁺ centers coexist in the ZGeO sample, in line with the theoretical DFT predictions. Moreover, the RL/NIR-I emission bands from the Cr³⁺ (1 to 3) centers are sensitive to the temperature rising from 70 K to RT. As seen in Fig. 10c (see also Fig. S6), a high-energy shift of the emission maxima with increasing temperature was identified, indicating that a thermal population of higher excited states occurs. Particularly, an energy separation of ca. 73 meV and 50 meV was estimated from the band maxima for the case of Cr³⁺(1) and Cr³⁺(2) in ZGeO sample, respectively, whereas a value of ca. 23 meV was found for the Cr³⁺(3) center in the ZGeO:Cr sample. These results also evidence that Cr³⁺(2) and Cr³⁺(3) are not the same center, despite their similar spectral shape at 70 K.

At longer wavelengths (Fig. 10d), and still in the NIR-I region, two main bands are observed centered at 880 nm/1.41 eV and 940 nm/1.32 eV. These emission bands are detected upon excitation on the Cr³⁺ absorption (e.g., 458 nm and 532 nm) and are absent for longer wavelength photon excitation (808 nm), in which the Cr⁴⁺ absorption takes place (Fig. 5a). Therefore, they are tentatively attributed to the intraionic trivalent transitions of Cr³⁺ placed in a site subjected to a lower crystal field strength than the previous ones. Even though the spectral shape of this Cr³⁺(4) center is similar to that of Cr³⁺(3), some differences occur with increasing temperatures. While the latter shows a thermal population when the temperature rises, the Cr³⁺(4) luminescence arises simultaneously from two split excited states (maxima are separated by 90 meV), and no changes in the peak position were observed with increasing temperature from 70 K to RT. In NIR-I region and on the low energy side of the Cr³⁺(4) center, we also observe the ⁴I_11/2 → ⁴I_15/2 transition from Er³⁺, for the ZGeO sample under 457.9 nm excitation.

The presence of multiple Cr³⁺ optical centers hinder a full deconvolution analysis when temperature-dependent PL measurements are made. Fig. S6a, c and e show the overall PL emission of Cr³⁺ centers in both samples, between 15 K and RT under 325 nm photon excitation and between 70 K and RT under 532 nm photon excitation. As a general tendency, the overall PL intensity of the RL/NIR-I decreases ca. 72% (Fig. S6b) and 95% (Figures S6d and f) of its original intensity for the ZGeO and ZGeO:Cr samples, respectively.

In the NIR-II spectral region (Fig. 10e), and besides the ⁴I_13/2 → ⁴I_15/2 Er³⁺ luminescence for the ZGeO sample, the undoped and Cr-doped ZGeO exhibit additional broad bands with the peak position also dependent on the excitation energy and temperature. In particular, at 70 K, by using long wavelength (808 nm) photon excitation, matching the unpolarized absorption bands of Cr⁴⁺, a dominant emission with a maximum at 1493 nm/0.83 eV is observed. Hence, it is reasonable to assume that the 1493 nm transition corresponds to the split intraionic ³T₂ → ³A₂ transition of the Cr⁴⁺ in C₁ site symmetry (Fig. 6a) with weak crystal field strength, as also reported for the isostructural Zn₂SiO₄ willemite host.⁶⁶ In addition, and for the case of the undoped sample, the 1493 nm emission is also detected when 457.9 nm photon excitation (via Cr³⁺ absorption) is performed. However, if the sample is subjected to 532 nm photon excitation, a shift of the band maximum towards higher energies is noticed, with the emission peak at 1444 nm/0.86 eV. As for the case of Cr³⁺, such findings suggest that Cr⁴⁺ ions could be subjected to slightly different crystalline strengths with sites preferentially excited with different photon energies. As so, we assign the 1493 nm and 1444 nm emission bands to Cr⁴⁺(1) and Cr⁴⁺(2) centers, respectively, and these are the dominant ones for the ZGeO sample. Furthermore, while the Cr⁴⁺(2) emission band keeps its emission maximum independent of temperature, in the case of the Cr⁴⁺(1), raising the temperature between 70 K and RT leads to a 30 meV shift of the peak position to high energies, suggesting a thermal population of higher excited states. The temperature-dependent PL spectra of the multiple Cr⁴⁺centers of ZGeO sample, as well as the ⁴I_13/2 → ⁴I_15/2 of Er³⁺ obtained with 532 nm photon excitation, are shown in Figures S7a and c. As for Cr³⁺ the overlap of the emitting centers (Cr⁴⁺ and Er³⁺) in the measured spectral region hinders a full spectral deconvolution. Therefore, from the overall integrated intensity of all Cr⁴⁺ centers, a thermal population occurs between 70 K and 130 K with further nonradiative deexcitation up to RT. As a general tendency, the overall PL intensity decreases ca. 42% of its original intensity for the ZGeO sample (Fig. S7d). In addition, the Er^{3+ 4}I_13/2→⁴I_15/2 integrated PL was found to increase by 140% (Fig. S7b), indicating that the emission is thermally populated. For the case of the ZGeO:Cr sample, besides the Cr⁴⁺(1) center, exciting the sample with 457.9 nm and 532 nm results, at 70 K, an emission maximum at ca. 1320 nm/0.94 eV with a shoulder at low energy. Such asymmetry suggests that the emission could be due to an overlap between the 1320 nm band (designated as the Cr⁴⁺(3) center) and the Cr⁴⁺(1) center. Furthermore, raising the temperature between 70 K and RT shows that with 532 nm excitation, the Cr⁴⁺(1) is strongly quenched and the emission of the Cr⁴⁺(3) center prevails (see Fig. S7e). The overall integrated luminescence of Cr⁴⁺centers for the ZGeO:Cr sample center was seen to decrease 85% in the 70 K to RT interval (Fig. S7f). For the 457.9 nm photon excitation, the broadening of the Cr⁴⁺(3) center increases, not only due to thermal processes, but also likely due to the contribution of Cr⁴⁺(2). Therefore, for the doped sample, three Cr⁴⁺ optically active centers are present. Nevertheless, it should be stressed that, as measured by RT absorption (Fig. 5a) and estimated by DFT (Fig. 6a), the Cr⁴⁺ ions are in C₁ site symmetry when replacing the germanium ions in the ZGeO host. Hence, as happens for Ca₂GeO₄ and Mg₂GeO₄,⁷² the triplet states are expected to unfold under the low site symmetry, giving rise to additional transitions (Fig. S5). As such, we cannot discard that some of the assigned centers could be related to the luminescence of excited states from the same optically active defect. However, and as a main conclusion, our samples clearly evidence that besides the Er³⁺ contamination in the ZGeO sample, the detected emission in NIR-I and NIR-II spectral regions is due to Cr³⁺ and Cr⁴⁺ that coexist in low site symmetries and weaker crystalline strength on the ZGeO lattice, in good agreement with the theoretical DFT predictions and RT absorption (Fig. 5a).

4. Conclusions

Nominally undoped and chromium-doped zinc germanate pellets were successfully synthesized via a high-temperature solid-state method exhibiting a monophasic polycrystalline willemite structure, as confirmed by XRD and Raman measurements. The morphological analysis performed by SEM revealed that the microstructure was similar for both samples and presented a cohesive and dense morphology. The constituent elements of the host material, Zn, Ge, and O, were identified by EDX and XPS. These two advanced techniques allowed us to confirm chromium as a dopant, whereas chromium and manganese were identified as trace impurities from the reagent precursors, by ICP. RT absorption spectra evidence that the samples studied here have a direct bandgap energy of 4.77 eV, in line with the 4.8 eV estimated by DFT at the Γ point. Furthermore, Cr absorption bands were identified in both samples. Absorption bands were assigned to intraionic transitions of Cr³⁺ ions for shorter wavelengths, while intraionic absorption bands of Cr⁴⁺ were identified for the longer wavelength region. The latter are in line with the ions in low site symmetry, as also estimated by DFT calculations, when the ions are placed in Ge⁴⁺ sites with C₁ symmetry. The former are consistent with the ions in distorted octahedral sites, in agreement with the DFT predictions when the ions occupy interstitial positions in the Zn–Ge and Zn–Zn rings with C₃ and S₆ site symmetries, respectively. In addition, for our insulator samples, these correspond to the Cr³⁺ defects with lowest formation energies. In the near-band edge region, an additional absorption band located at 313 nm was observed and assigned to a Mn²⁺–O²⁻ charge transfer state. Mn²⁺ occupying the two Zn²⁺ sites was found to be a trace impurity in the studied samples. When the ZGeO and ZGeO:Cr sample are excited with above or below bandgap excitation, two wide and structureless bands were observed in the visible region. For band-to-band excitation, our data are consistent with a description of the bluish-white band centered at 480 nm as an overlap of the intraionic ⁴T₁ → ⁶A₁ green luminescence of Mn²⁺ with a blue luminescence likely related to Zn_i intrinsic defects, the intrinsic defects with lower energy formation. Moreover, both samples exhibit deep-red/NIR-I/NIR-II emission bands that are related to the unfolded energy levels of Cr ions in trivalent and tetravalent valence states in low site symmetries, respectively. For shorter wavelengths (deep-red/NIR-I), four different Cr³⁺ centers were identified, while in the NIR-II region, the emission is dominated by Cr⁴⁺ emission bands corresponding to three different Cr⁴⁺ centers. From the temperature dependence analysis of the integrated luminescence, recombination models and thermal quenching mechanisms were proposed for the different optically active centers. Moreover, trapping/detrapping mechanisms and energy transfer were found to be effective in the defective samples studied. In particular, for the ZGeO sample, which also exhibits Er³⁺ contamination, a color shift from green to blue was found to occur at room temperature upon band-to-band excitation. The mechanism was assigned to an energy transfer process between the Mn²⁺ activator and the Er³⁺ sensitizer.

The presented results are a step forward in the clarification of the nature of the active optical centers in zinc germanate, which evidences a strong dependence on trace contaminants, intrinsic defects, and deliberately introduced chromium dopants, contributing to exploring the host potential as an emerging broad band emitter either in visible and NIR spectral regions.

Author contributions

Maria S. Batista: conceptualization, formal analysis, investigation, methodology, visualization, writing – original draft, writing – review & editing. Miguel P. Dias: investigation, writing – review & editing. Maria B. Candeias: investigation, writing – review & editing. Gabriel Marques: investigation, writing – review & editing. Daniel D. Gouveia: formal analysis, methodology, writing – review & editing. Ana V. Girão: investigation, formal analysis, investigation, writing – review & editing. Florinda M. Costa: funding acquisition, formal analysis, resources, writing – review & editing. Joaquim Leitão: investigation, resources, writing – review & editing. Luís Rino: formal analysis, writing – review & editing. Jonas Deuermeier: investigation, writing – review & editing. Elvira Fortunato: funding acquisition, resources, writing – review & editing. Rodrigo Martins: funding acquisition, resources, writing – review & editing. Ana Pimentel: investigation, supervision, writing – review & editing. Joana Rodrigues: funding acquisition, investigation, methodology, writing – review & editing. Teresa Monteiro: conceptualization, funding acquisition, formal analysis, methodology, resources, supervision, writing – original draft, writing – review & editing. Sónia O. Pereira: conceptualization, formal analysis, investigation, methodology, supervision, writing – original draft, writing – review & editing.

Conflicts of interest

There are no conflicts to declare.

Data availability

The data supporting this article have been included in the main article and as part of the SI. Supplementary information: Additional theoretical data, additional XPS data, temperature-dependent PL spectra, and schematic diagrams of the defect energy levels. See DOI: https://doi.org/10.1039/d5tc02709h

Acknowledgements

This study has the support of the Institute for Nanostructures, Nanomodelling and Nanofabrication (i3N), through projects UIDB/50025/2020-2023 (DOI: 10.54499/UIDB/50025/2020) & UIDP/50025/2020-2023 (DOI: 10.54499/UIDP/50025/2020) & LA/P/0037/2020 (DOI: 10.54499/LA/P/0037/2020), as well as IonProGO (2022.05329.PTDC), DOI: 10.54499/2022.05329, 2023.00054.RESTART, LIGHEART (2022.08597.PTDC), CO2RED (PTDC/EQU-EPQ/2195/2021), and POEMS (101217446) projects, all funded by the Foundation for Science and Technology (FCT). J. Rodrigues acknowledges FCT support via grant CEECINSTLA/00005/2022. J. D. Gouveia acknowledges the FCT grant 2023.06511.CEECIND, in the scope of the Individual Call to Scientific Employment Stimulus – 6th Edition, M. S. Batista thanks i3N and FCT for the PhD grant (UI/BD/152567/2022). G. Marques acknowledges the financial support from Project “Agenda ILLIANCE” [C644919832-00000035 | Project no. 46], financed by PRR – Plano de Recuperação e Resiliência under the Next Generation EU from the European Union. A. V. Girão also acknowledges the project CICECO-Aveiro Institute of Materials, UIDB/50011/2020, UIDP/50011/2020 & LA/P/0006/2020, financed by national funds through the FCT/MEC (PIDDAC). The authors thank MSc. Celeste Azevedo and Doctor Rosário Soares (CICECO – Aveiro Institute of Materials), for the DR and XRD experiments, respectively.

References

1. Z. Liu, H. Huang, B. Liang, X. Wang, Z. Wang, D. Chen and G. Shen, Opt. Express, 2012, 20, 2982.
2. M. Muzafa Jumidali, M. R. Hashim and K. Al-Heuseen, Mater. Sci. Semicond. Process., 2013, 16, 1360–1364.
3. C. Yan, N. Singh and P. S. Lee, Appl. Phys. Lett., 2010, 96(5), 053108.
4. S. K. Gupta, K. Sudarshan, B. Modak and R. Gupta, ACS Appl. Electron. Mater., 2023, 5, 1286–1294.
5. D. K. Bharti, S. Badatya, P. Tanwar, J. Tawale, A. K. Srivastava and M. K. Gupta, Mater. Sci. Eng., B, 2020, 259, 114602.
6. C. Karthikeyan, P. Arunachalam, K. Ramachandran, A. M. Al-Mayouf and S. Karuppuchamy, J. Alloys Compd., 2020, 828, 154281.
7. L. Sun, Y. Qi, C.-J. Jia, Z. Jin and W. Fan, Nanoscale, 2014, 6, 2649.
8. J. Sato, H. Kobayashi, K. Ikarashi, N. Saito, H. Nishiyama and Y. Inoue, J. Phys. Chem. B, 2004, 108, 4369–4375.
9. Y. Yuan, H. Dai, H. Chi, W. Gao, Q. Liu, C. Ding, Y. Shen, Z. Tang, C. Zhuang, Y. Yang, Y. Zhang, Z. Zou and Y. Zhou, ACS Mater. Lett., 2022, 4, 2631–2637.
10. R. Li, R. Zhang, Z. Lou, T. Huang, K. Jiang, D. Chen and G. Shen, Nanoscale, 2019, 11, 12116–12123.
11. Y.-C. Liang and T.-Y. Lin, Nanoscale Res. Lett., 2014, 9, 344.
12. R. M. Calderón-Olvera, E. Arroyo, A. M. Jankelow, R. Bashir, E. Valera, M. Ocaña and A. I. Becerro, ACS Appl. Mater. Interfaces, 2023, 15, 20613–20624.
13. F. Chi, B. Jiang, Z. Zhao, Y. Chen, X. Wei, C. Duan, M. Yin and W. Xu, Sens. Actuators, B, 2019, 296, 126640.
14. Z. Yin, L. Zhu, Z. Lv, M. Li and D. Tang, Talanta, 2021, 233, 122563.
15. J. Wang, Q. Ma, W. Zheng, H. Liu, C. Yin, F. Wang, X. Chen, Q. Yuan and W. Tan, ACS Nano, 2017, 11, 8185–8191.
16. D. T. Palumbo and J. J. Brown, J. Electrochem. Soc., 1970, 117, 1184.
17. J. S. Lewis and P. H. Holloway, J. Electrochem. Soc., 2000, 147, 3148.
18. M. Shang, G. Li, D. Yang, X. Kang, C. Peng, Z. Cheng and J. Lin, Dalton Trans., 2011, 40, 9379.
19. F. Chi, X. Wei, B. Jiang, Y. Chen, C. Duan and M. Yin, Dalton Trans., 2018, 47, 1303–1311.
20. M. Shang, G. Li, D. Yang, X. Kang, C. Peng and J. Lin, Dalton Trans., 2012, 41, 8861.
21. Q. Zhang and J. Wang, Appl. Phys. A: Mater. Sci. Process., 2012, 108, 943–948.
22. S. Takeshita, J. Honda, T. Isobe, T. Sawayama and S. Niikura, J. Solid State Chem., 2012, 189, 112–116.
23. K. Huang, X. Dou, Y. Zhang, X. Gao, J. Lin, J. Qu, Y. Li, P. Huang and G. Han, Adv. Funct. Mater., 2021, 31(22), 2009920.
24. Z. Yang, L. D. Spencer, D. R. Patocka, A. Putta, E. E. Schnetzer, J. Kellar and C. Jiang, Opt. Mater., 2023, 142, 114036.
25. D. Gao, K. Ma, P. Wang, X. Zhang, Q. Pang, H. Xin, Z. Zhang and H. Jiao, Dalton Trans., 2022, 51, 553–561.
26. F. Zhao, Z. Song and Q. Liu, Laser Photonics Rev., 2022, 16(11), 2200380.
27. S. He, Y. Liu, T. Gao, R. Liu, G. Chen, M. Duan and M. Cao, ACS Omega, 2023, 8, 15698–15707.
28. H. Suo, Y. Wang, X. Zhao, X. Zhang, L. Li, K. Guan, W. Ding, P. Li, Z. Wang and F. Wang, Laser Photonics Rev., 2022, 16(7), 2200012.
29. H. Zhou, H. Cai, J. Zhao, Z. Song and Q. Liu, Inorg. Chem. Front., 2022, 9, 1912–1919.
30. M. S. Batista, J. Rodrigues, M. S. Relvas, J. Zanoni, A. V. Girão, A. Pimentel, F. M. Costa, S. O. Pereira and T. Monteiro, Appl. Sci., 2022, 12, 2104.
31. S. Liu, J. Du, Z. Song, C. Ma and Q. Liu, Light: Sci. Appl., 2023, 12, 181.
32. Y. Wang, G. Liu and Z. Xia, Laser Photonics Rev., 2024, 18(1), 2300717.
33. X. Guo, B. Liu, R. Kuang, W. Gan, L. Huang and J. Wang, J. Mater. Chem. C, 2023, 11, 7611–7618.
34. R. T. Parayil, S. K. Gupta and M. Mohapatra, Coord. Chem. Rev., 2025, 522, 216200.
35. D. M. Esteves, M. S. Batista, J. Rodrigues, A. V. Girão, L. C. Alves, A. L. Rodrigues, M. I. Dias, F. M. Costa, K. Lorenz, S. O. Pereira, T. Monteiro and M. Peres, J. Mater. Chem. C, 2024, 12, 19359–19370.
36. H. Suo, Y. Wang, X. Zhang, W. Zheng, Y. Guo, L. Li, P. Li, Y. Yang, Z. Wang and F. Wang, Matter, 2023, 6, 2935–2949.
37. L. C. Williams, D. Norton, J. Budai and P. H. Holloway, J. Electrochem. Soc., 2004, 151, H188.
38. J. Dolado, R. Martínez-Casado, P. Hidalgo, R. Gutierrez, A. Dianat, G. Cuniberti, F. Domínguez-Adame, E. Díaz and B. Méndez, Acta Mater., 2020, 196, 626–634.
39. T. K. Tran, W. Park, J. W. Tomm, B. K. Wagner, S. M. Jacobsen, C. J. Summers, P. N. Yocom and S. K. McClelland, J. Appl. Phys., 1995, 78, 5691–5695.
40. Z. Liu, X. Jing and L. Wang, J. Electrochem. Soc., 2007, 154, H500.
41. G. Anoop, K. M. Krishna and M. K. Jayaraj, J. Electrochem. Soc., 2008, 155, J7.
42. S. Luo, L. Trefflich, S. Selle, R. Hildebrandt, E. Krüger, S. Lange, J. Yu, C. Sturm, M. Lorenz, H. von Wenckstern, C. Hagendorf, T. Höche and M. Grundmann, Appl. Phys. Lett., 2023, 122(3), 031601.
43. R. Stevens, B. F. Woodfield, J. Boerio-Goates and M. K. Crawford, J. Chem. Thermodyn., 2004, 36, 349–357.
44. Q. Bai, Z. Wang, P. Li, S. Xu, T. Li, J. Cheng and Z. Yang, Mater. Des., 2016, 108, 597–607.
45. J. Breternitz, D. Fritsch, A. Franz and S. Schorr, Z. Anorg. Allg. Chem., 2021, 647, 2195–2200.
46. C. He, Z. Xiao, L. Zeng, T. Jin, J. Liu, G. Li, H. He, L. Zhou, J. Yang, C. Li and J. Tang, Ceram. Int., 2023, 49, 35254–35263.
47. X. Hou, C. Du, Q. Pang, C. Jia, X. Wang and D. Gao, J. Lumin., 2024, 265, 120217.
48. S. Zhang, Y. Hu, R. Chen, X. Wang and Z. Wang, Opt. Mater., 2014, 36, 1830–1835.
49. J. Huang, K. Ding, Y. Hou, X. Wang and X. Fu, ChemSusChem, 2008, 1, 1011–1019.
50. J. Q. Hu, E. H. Song, S. Ye, B. Zhou and Q. Y. Zhang, J. Mater. Chem. C, 2017, 5, 3343–3351.
51. Y. Takahashi, M. Ando, K. Iwasaki, H. Masai and T. Fujiwara, Appl. Phys. Lett., 2010, 97(7), 071906.
52. H. He, Y. Zhang, Q. Pan, G. Wu, G. Dong and J. Qiu, J. Mater. Chem. C, 2015, 3, 5419–5429.
53. Z.-Y. Xie, H.-L. Lu, Y. Zhang, Q.-Q. Sun, P. Zhou, S.-J. Ding and D. W. Zhang, J. Alloys Compd., 2015, 619, 368–371.
54. M. Kanzaki, J. Mineral. Petrol. Sci., 2018, 113, 41–46.
55. Z. Gu, F. Liu, X. Li and Z. W. Pan, Phys. Chem. Chem. Phys., 2013, 15, 7488.
56. Y. Cong, Y. He, B. Dong, Y. Xiao and L. Wang, Opt. Mater., 2015, 42, 506–510.
57. M. Tinoco, J. M. Lendínez, J. M. González-Calbet, B. Méndez, P. Hidalgo and J. Ramírez-Castellanos, preprints,  DOI:10.20944/preprints202309.0623.v1.
58. J. Dolado, P. Hidalgo and B. Méndez, Phys. Status Solidi RRL, 2022, 16, 1–6.
59. J. Dolado, J. García-Fernández, P. Hidalgo, J. González-Calbet, J. Ramírez-Castellanos and B. Méndez, J. Alloys Compd., 2022, 898, 162993.
60. L. Li, Y. Su, Y. Chen, M. Gao, Q. Chen and Y. Feng, Adv. Sci. Lett., 2010, 3, 1–5.
61. Y. Yue, T. Wang, Y. Yan, L. Guo, X. Zhu, W. Bu, G. Wang, N. Zhu, L. Zhao and X. Yu, ACS Appl. Mater. Interfaces, 2024, 16, 32402–32410.
62. J. Ren, X. Xu, H. Zeng, G. Chen, D. Kong, C. Gu, C. Chen, Z. Liu and L. Kong, J. Am. Ceram. Soc., 2014, 97, 3197–3201.
63. K. H. Yoon and J. H. Kim, Thin Solid Films, 2010, 519, 1583–1586.
64. B. Wagstaff and A. Kitai, J. Lumin., 2015, 167, 310–315.
65. D. W. Feldman, J. Chem. Phys., 1973, 58, 363–366.
66. C. R. Ronda and T. Amrein, J. Lumin., 1996, 69, 245–248.
67. T. Zheng, H. Yang, H. Lu, Y. Liu, Y. Li, C. Peng, L. Zhang and X. Li, J. Lumin., 2024, 271, 120598.
68. X. Zou, H. Zhang, W. Li, M. Zheng, M. S. Molokeev, Z. Xia, Y. Zheng, Q. Li, Y. Liu, X. Zhang and B. Lei, Adv. Opt. Mater., 2022, 10(19), 2200882.
69. Z. Ye, Z. Wang, H. Yang, X. Huo, Y. Wang, Q. Wu, D. Wang, J. Zhao, H. Suo and P. Li, Dalton Trans., 2022, 51, 13499–13506.
70. R. G. Pappalardo, W. J. Miniscalco, T. E. Peters and K. Lee, J. Lumin., 1993, 55, 87–93.
71. M. F. Hazenkamp, U. Oetliker, H. U. Güdel, U. Kesper and D. Reinen, Chem. Phys. Lett., 1995, 233, 466–470.
72. M. F. Hazenkamp, H. U. Güdel, M. Atanasov, U. Kesper and D. Reinen, Phys. Rev. B: Condens. Matter Mater. Phys., 1996, 53, 2367–2377.
73. K. V. Yumashev, N. V. Kuleshov, P. V. Prokoshin, A. M. Malyarevich and V. P. Mikhailov, Appl. Phys. Lett., 1997, 70, 2523–2525.
74. S. Kück, Appl. Phys. B, 2001, 72, 515–562.
75. A. Sennaroglu, Prog. Quantum Electron., 2002, 26, 287–352.
76. Y. Nanai, R. Ishida, Y. Urabe, S. Nishimura and S. Fuchi, Jpn. J. Appl. Phys., 2019, 58, SFFD02.
77. H. Cai, H. Chen, H. Zhou, J. Zhao, Z. Song and Q. L. Liu, Mater. Today Chem., 2021, 22, 100555.
78. W. Huang, V. Rajendran, M. Chan, M. Hsiao, H. Chang and R. Liu, Adv. Opt. Mater., 2023, 11(11), 2202061.
79. J. Zhong, D. Chen, Y. Peng, Y. Lu, X. Chen, X. Li and Z. Ji, J. Alloys Compd., 2018, 763, 34–48.
80. Y. Tanabe and S. Sugano, J. Phys. Soc. Japan, 1954, 9, 766–779.
81. M. S. Relvas, M. R. N. Soares, S. O. Pereira, A. V. Girão, F. M. Costa and T. Monteiro, J. Phys. Chem. Solids, 2019, 129, 413–423.
82. L. Shang, M. Liu and C.-K. Duan, J. Phys. Chem. Lett., 2022, 13, 10635–10641.
83. V. Rajendran, M.-H. Fang, G. N. De Guzman, T. Lesniewski, S. Mahlik, M. Grinberg, G. Leniec, S. M. Kaczmarek, Y.-S. Lin, K.-M. Lu, C.-M. Lin, H. Chang, S.-F. Hu and R.-S. Liu, ACS Energy Lett., 2018, 3, 2679–2684.
84. S. Ding, P. Feng, J. Cao, X. Ma and Y. Wang, ACS Appl. Mater. Interfaces, 2022, 14, 44622–44631.
85. Z. Li, J. Xie, M. Faizan, S. Wang, X. Chen and L. Zhang, Adv. Funct. Mater., 2025, 35(29), 2422716.
86. G. Kresse and J. Furthmüller, Phys. Rev. B: Condens. Matter Mater. Phys., 1996, 54, 11169–11186.
87. A. V. Krukau, O. A. Vydrov, A. F. Izmaylov and G. E. Scuseria, J. Chem. Phys., 2006, 125(22), 224106.
88. P. E. Blöchl, Phys. Rev. B: Condens. Matter Mater. Phys., 1994, 50, 17953–17979.
89. J. D. Gouveia and J. Coutinho, Electron. Struct., 2019, 1, 015008.
90. S. Wu and Q. Ma, J. Alloys Compd., 2013, 567, 40–46.
91. Y. Zhao, S. Yang, J. Zhu, G. Ji and F. Peng, Solid State Ionics, 2015, 274, 12–16.
92. J. Dolado, P. Hidalgo and B. Méndez, Phys. Status Solidi, 2018, 215, 1–5.
93. P. Taret, Spectrochim. Acta, 1962, 18, 467–483.
94. L. Z. Pei, Y. Yang, L. J. Yang, C. G. Fan, C. Z. Yuan and Q.-F. Zhang, Solid State Commun., 2011, 151, 1036–1041.
95. L. Yu, R. Zou, Z. Zhang, G. Song, Z. Chen, J. Yang and J. Hu, Chem. Commun., 2011, 47, 10719.
96. I. C. Amaechi, G. Kolhatkar, A. H. Youssef, D. Rawach, S. Sun and A. Ruediger, RSC Adv., 2019, 9, 20806–20817.
97. B. C. Babu, G.-G. Wang, B. Yan, Q. Yang and A. P. Baker, Ceram. Int., 2018, 44, 938–946.
98. J. Dolado, B. Rodríguez, R. Martínez-Casado, I. Píš, E. Magnano, P. Hidalgo and B. Méndez, Acta Mater., 2023, 245, 118606.
99. P. S. Bagus, E. S. Ilton and J. R. Rustad, Phys. Rev. B: Condens. Matter Mater. Phys., 2004, 69, 205112.
100. X. Wang, Z. Wang, M. Zheng, J. Cui, Y. Yao, L. Cao, M. Zhang, Z. Yang, H. Suo and P. Li, Dalton Trans., 2021, 50, 311–322.
101. J. F. Moulder, W. F. Stickle, P. E. Sobol and K. D. Bomben, Handbook of X-ray Photoelectron Spectroscopy, PerkinElmer Corporation Physical Electronics Division, Minnesota, USA, 1995.
102. P. Kubelk and F. Munk, Z. Technol. Phys., 1931, 12, 593–601.
103. P. Makuła, M. Pacia and W. Macyk, J. Phys. Chem. Lett., 2018, 9, 6814–6817.
104. T. Matsuzawa, Y. Aoki, N. Takeuchi and Y. Murayama, J. Electrochem. Soc., 1996, 143, 2670–2673.
105. B. Henderson and G. F. Imbusch, Optical Spectroscopy of Inorganic Solids, Oxford University Press, New York, 1989.
106. W. Jia, H. Liu, S. Jaffe, W. M. Yen and B. Denker, Phys. Rev. B: Condens. Matter Mater. Phys., 1991, 43, 5234–5242.
107. V. Petričević, A. B. Bykov, J. M. Evans and R. R. Alfano, Opt. Lett., 1996, 21, 1750.
108. B. C. Babu, G.-G. Wang, B. Yan, Q. Yang and A. P. Baker, Ceram. Int., 2018, 44, 938–946.
109. V. Rajendran, H. Chang and R.-S. Liu, Opt. Mater. X, 2019, 1, 100011.
110. C. Freysoldt, B. Grabowski, T. Hickel, J. Neugebauer, G. Kresse, A. Janotti and C. G. Van de Walle, Rev. Mod. Phys., 2014, 86, 253–305.
111. B. Struve and G. Huber, Appl. Phys. B, 1985, 36, 195–201.
112. M. P. Dias, M. S. Batista, A. Pimentel, E. Fortunato, R. Martins, F. M. Costa, S. O. Pereira, J. Rodrigues and T. Monteiro, Mater. Chem. Phys., 2025, 334, 130463.
113. R. M. Ranson, E. Evangelou and C. B. Thomas, Appl. Phys. Lett., 1998, 72, 2663–2664.
114. V. Lojpur, Ž. Antić and M. D. Dramićanin, Phys. Chem. Chem. Phys., 2014, 16, 25636–25641.
115. ed. G. Liu and B. Jacquier, Spectroscopic Properties of Rare Earths in Optical Materials, Springer, 2006.
116. W. H. Fonger and C. W. Struck, Phys. Rev. B, 1975, 11, 3251–3260.
117. Y. Zhuang, S. Tanabe and J. Qiu, J. Am. Ceram. Soc., 2014, 97, 3519–3523.
118. M. Peres, D. M. Esteves, B. M. S. Teixeira, J. Zanoni, L. C. Alves, E. Alves, L. F. Santos, X. Biquard, Z. Jia, W. Mu, J. Rodrigues, N. A. Sobolev, M. R. Correia, T. Monteiro, N. Ben Sedrine and K. Lorenz, Appl. Phys. Lett., 2022, 120(26), 261904.
119. J. Rodrigues, A. F. Carvalho, J. Zanoni, A. J. S. Fernandes, F. M. Costa and T. Monteiro, Appl. Sci., 2024, 14, 528.

---