Evaluation of relaxation dynamics from excited states of Ho³⁺ ions in sol–gel nano-glass-ceramic materials

Abstract

In this paper, a series of oxyfluoride SiO₂-LaF₃ nano-glass-ceramics (GCs) activated by Ho³⁺ ions with variable concentrations were synthesized by the sol–gel method. The crystallization of the LaF₃ fluoride phase for all prepared series was verified by XRD measurements and TEM microscopy. For the fabricated GCs, a series of specific luminescence bands within VIS, NIR, and MIR regions were recorded, and the most prominent emissions were assigned to the following 4f¹⁰–4f¹⁰ transitions of Ho³⁺ ions: (⁵S₂,⁵F₄) → ⁵I₈ (green), ⁵F₅ → ⁵I₈ (red), ⁵I₆ → ⁵I₈ (NIR, ∼1.19 μm), and ⁵I₇ → ⁵I₈ (MIR, ∼2.0 μm). Based on the mutual intensities of recorded individual emission bands, their percentage contributions (β) were evaluated. It was found that the calculated values of β coefficients are strongly dependent on the applied annealing temperature (700 or 900 °C), the average size of precipitated LaF₃ crystals, and the concentration of Ho³⁺ ions. Thus, the results indicate that the connotation in the mutual population of the individual excited states (i.e., (⁵S₂,⁵F₄), ⁵F₅, ⁵I₆, and ⁵I₇) – correlated with the relaxation dynamics from the above-mentioned levels – relies on the phonon energy (modified during an increase in annealing temperature (700 °C → 900 °C)) in the nearest proximity around Ho³⁺ ions by activating (or suspending) the non-radiative multiphonon channels (MPR). The performed luminescence studies tentatively showed that increasing the heat-treatment temperature favors Ho³⁺ migration from the silicate network into the low-oscillation energy LaF₃ phase, unfavoring the non-radiative MPR processes between the (⁵S₂,⁵F₄) and the ⁵F₅ states, and also between the excited ⁵I₆/⁵I₇ states and the ⁵I₈ ground level. The prepared Ho³⁺-doped GCs are able to emit green light with high color purity (CP) reaching even 98%, and generate long-lived NIR/MIR emissions at ∼1.19 μm (up to τ(⁵I₆) = 10.19 ms) and ∼2.0 μm (up to τ(⁵I₇) = 8.44 ms), which could predispose them for use in optoelectronic devices.

---

Introduction

Trivalent lanthanide ions – exhibiting a specific electron configuration, i.e., [Xe]4fⁿ (n = 0–14) – are characterized by multiple energy states within a broad range of electromagnetic spectra from ultraviolet (UV), through visible (VIS) up to near- or mid-infrared (NIR/MIR) irradiation.¹ Among them, Ho³⁺ ions with a [Xe]4f¹⁰ configuration emit green ((⁵S₂,⁵F₄) → ⁵I₈) and red (⁵F₅ → ⁵I₈) light, as well as radiation from NIR and MIR ranges, i.e., 1.2 μm (⁵I₆ → ⁵I₈), 1.38 μm (⁵S₂,⁵F₄) → ⁵I₅), 1.5 μm (⁵F₅ → ⁵I₆), 2.0 μm (⁵I₇ → ⁵I₈), and even 3.9 μm.^2–4 The unique properties of Ho³⁺ ions are strictly correlated with their multiple meta-stable levels; those luminescence peculiarities open diverse application possibilities.⁵ Indeed, Ho³⁺-doped materials are mainly dedicated to green lighting applications (e.g., for lasers or color displays), as was demonstrated for plentiful glassy hosts,^2,6–9 and also for several types of ceramic materials (for example, Y₂O₃ or MTiO₃ (M = Ca or Ba)).^10–12 Additionally, the literature mentions the possibility of dominant generation of red luminescence, as was found for glass ceramics containing BaAlBO₃F₂:Ho³⁺ nanocrystals or Ho³⁺-doped fibers.^13,14 Interestingly, as was described for glass-ceramics containing Ba₄Y₃F₁₂:Yb³⁺/Ho³⁺ nanoparticles, co-doping with Ce³⁺ favors the occurrence of appropriate cross-relaxation (CR) processes, responsible for the gradual depopulation of the upper-lying (⁵S₂,⁵F₄) levels with a simultaneous enhancement in the occupation of the lower-lying ⁵F₅ state, resulting in the generation of tunable green-to-red luminescence dependent on the Ce³⁺ concentration.¹⁵ It is expected that visible emission originating from Ho³⁺ ions could be applied in data storage systems and sensors.^6,16,17 Moreover, Ho³⁺ luminescence with a combination of blue-emitting centers (like Tm³⁺) is suitable for producing illumination dedicated to technologies based on white-light emitting diodes (WLEDs), as was presented for Ho³⁺/Yb³⁺/Tm³⁺ triply-doped GeO₂-Nb₂O₅ nanostructured materials.¹⁸

Even with the tremendous application of VIS luminescence from Ho³⁺-doped optical materials, nonetheless, the potential utility area of NIR and MIR emissions has also attracted attention from many researchers. Factually, the characteristic NIR luminescence of Ho³⁺ ions at 1.2 μm – located near the second telecommunication window – could be applied in designing O-band amplifiers, as was presented, e.g., for heavy-metal gallate glasses.¹⁹ Importantly, ∼1.2 μm lasers can be applied in medicine or for oxygen detection, and this fact was first reported by NASA;²⁰ thus, optical fibers operating on the ⁵I₆ → ⁵I₈ electronic transition of Ho³⁺ may be suitable for lasing purposes as was described for Ho³⁺/Yb³⁺ co-doped TeO₂-BaF₂-Y₂O₃ glasses.²¹ Furthermore, the ‘eye-safe’ MIR emission at 2.0 μm, imperceptible to the human eye, is utilized in remote sensing (LIDAR) devices, wind mapping, and laser ranger finders. Additionally, 2.0 μm wavelength is firmly absorbed by water molecules; thus, it could be used in surgery as an optical laser beam ensuring less damaged tissues, less bleeding, and limiting contamination in comparison with metallic ones; this MIR emission could find application not only in the detection of atmospheric pollutions, like greenhouse gases, e.g., methane or carbon dioxide, but also in atmospheric communication systems.^3,22–25 The great fields of 2.0 μm wavelength utilities presented above are a strong motivation for designing and developing appropriate materials; factually, Yb³⁺/Ho³⁺/Ce³⁺ triply-doped sodium-zinc-tellurite glasses,²⁶ Er³⁺/Ho³⁺ co-doped silicate glasses,²⁷ or Yb³⁺/Ho³⁺ co-doped oxyfluoride glass-ceramics²⁸ are considered as promising candidates for improving the Ho³⁺ 2.0 μm fiber laser performance. Thus, considering the tremendous potential of Ho³⁺-doped optical materials, their design and studies are essential for developing the current optoelectronics.

The intensities of the individual emission bands of Ho³⁺ ions are sensitive to the oscillation energy in their local proximity in the host matrix.^16,29 Indeed, it is a well-known fact that multiphonon processes become more substantial if the energy gap is comparable with the phonon energy of the local neighborhood of Ln³⁺.²⁴ Thus, the resultant intensities of individual luminescence bands of Ho³⁺ strongly depend on the population of appropriate excited states, affected by the preferable location of dopant ions in the host matrix. Those connotations were described in detail for Ho³⁺/Yb³⁺ co-doped aluminosilicate glasses (G) and glass-ceramics containing NaYF₄ nanocrystals (GC).³⁰ Since the energy gap between the ⁵I₆ and ⁵I₇ levels of Ho³⁺ (ΔE ≈ 2550 cm⁻¹) could be easily matched by two or three phonons of the glass host (hω ≈ 1100 cm⁻¹), the authors established that the ⁵I₆ state could non-radiatively relax to the lower-lying ⁵I₇ level by a multiphonon-assisted process. Conversely, when Ho³⁺ ions are preferentially located inside NaYF₄ nanocrystals with a lower oscillation energy (hω ≈ 324 cm⁻¹), the multiphonon-assisted relaxation rate decreases greatly; thus, the upper-lying ⁵I₆ level is more occupied than the ⁵I₇ state. Resultantly, for glasses, Ho³⁺ ions at the ⁵I₇ level can be efficiently populated to the upper ⁵F₅ state; therefore, the ⁵F₅ → ⁵I₈ red emission dominates in emission spectra. In the case of glass-ceramics, Ho³⁺ ions in the ⁵I₆ state can be mainly pumped into the (⁵S₂,⁵F₄) levels. Also, for Ho³⁺/Yb³⁺ co-doped glass-ceramics containing CaF₂ nanocrystals,³¹ it was noticed that the correlation in mutual intensities of green (⁵S₂,⁵F₄) → ⁵I₈) and red (⁵F₅ → ⁵I₈) emissions is closely associated with the local phonon energy around Ho³⁺ ions; factually, the green luminescence increased gradually with progressive transformation from glasses to glass-ceramics. The authors explained that this trend is correlated with the growing efficiency of Ho³⁺ and Yb³⁺ migration inside the CaF₂ nanocrystal lattice, parallelly with an increase in the heat-treatment temperature from 650 to 725 °C; indeed, due to lower phonon energy inside the CaF₂ phase (hω ≈ 495 cm⁻¹), the non-radiative relaxation from the (⁵S₂,⁵F₄) states is reduced. Similar correlations can also be found based on the analysis of photoluminescence (PL) results shown in the available literature for various types of optical hosts doped with Ho³⁺ ions.^1,7,16,32 Thus, the modification in phonon energies around Ho³⁺ ions may significantly affect their resultant photoluminescence properties. Those peculiarities may be involved in the appropriate tuning of the emission for dedicated applications in optoelectronics.

In this paper, we synthesized and characterized a series of Ho³⁺-doped oxyfluoride nano-glass-ceramics containing the LaF₃ nanophase. The GCs were fabricated by annealing amorphous silicate xerogels at selected temperatures, i.e., 700 and 900 °C. The structural measurements of the fabricated sol–gel samples involved the registration of X-ray diffraction (XRD) patterns and transmission electron microscopy (TEM) images, and the photoluminescence characterization consisted in recording the excitation (PLE) and emission spectra (PL within VIS, NIR and MIR ranges) along with the decay lifetimes from the several excited states of Ho³⁺ ions, i.e., (⁵S₂,⁵F₄), ⁵F₅, ⁵I₆, and ⁵I₇. This work is intended to systematize the correlations between local phonon energy, and the population as well as relaxation dynamics from the individual excited levels of Ho³⁺ ions. Although the optical spectroscopy of Ho³⁺ ions in singly- or multi-doped glass-ceramics (fabricated by high-temperature melt-quenching followed by further annealing) has been described in several previously published papers,^{3,13,31,33–39} in this work, we synthesized and characterized Ho³⁺-doped nano-glass-ceramics fabricated by a sol–gel method. As far as we know, there are only a few published articles concentrated on the synthesis pathway and luminescence characterization of Ho³⁺-doped sol–gel glass-ceramic materials, i.e., SiO₂-PbF₂ triply-doped with Tm³⁺/Yb³⁺/Ho³⁺⁴⁰ and SiO₂-LaF₃ co-doped with Yb³⁺/Ho³⁺;⁴¹ hence – in our opinion – it seems to be justified and reasonable to perform additional studies in this matter.

Experimental

The series of xerogels activated by Ho³⁺ ions were prepared using the sol–gel synthesis described in our previous paper.⁴² All reagents were taken from Sigma Aldrich Chemical Company. The subsequent chemical reactions of tetraethoxysilane (TEOS), e.g., hydrolysis, condensation, and polycondensation, were performed in a solution of ethanol, deionized water, and acetic acid with a molar ratio equal to 1 : 4 : 10 : 0.5 (95 mol%). Simultaneously, the solutions of La(CH₃COO)₃ and Ho(CH₃COO)₃ in trifluoroacetic acid (TFA) and deionized water were added dropwise to the TEOS-based mixtures. The molar ratio of TFA : La³⁺ : Ho³⁺ was set on 5 : (1 − x) : x (5 mol%), where x = 0.003 (GC1_x), 0.006 (GC2_x), 0.012 (GC3_x), 0.03 (GC4_x), 0.06 (GC5_x), and 0.12 (GC6_x). The as-prepared sols were dried at 35 °C for several weeks to form slightly orange-colored (Ho³⁺) rigid xerogels. Finally, the nano-glass-ceramic materials were fabricated during heat-treatment of as-prepared xerogels at 700 and 900 °C, and the samples were appropriately denoted in the text as GC₇₀₀ and GC₉₀₀.

The fabricated glass-ceramics were characterized by powder diffraction measurements using a PANalytical X'Pert Pro MD diffractometer, which utilized the CuK_α1 X-ray line with a Bragg–Brentano standard setup included with a Ge(111) Johansson monochromator in the incident beam. The scanning range was set from 20° to 80°, with a step size of 0.016°. The full pattern fitting procedure was carried out using HighScore Plus (PANalytical) software version 3.1. and the PDF5+ 2024 (ICDD) database. The transmission electron microscopy (TEM) observations were performed using a JEOL high resolution (HR-TEM) JEM 3010 microscope working at an accelerating voltage of 300 kV and equipped with a Gatan 2k × 2k Orius™ 833SC200D CCD camera and an Elite T Energy Dispersive Spectroscopy (EDS) silicon drift detector (SDD) from AMETEK EDAX. The sol–gel samples were suspended in isopropanol, and after ultrasonication for 10 minutes, the resulting materials were deposited on a copper grid coated with an amorphous carbon film standardized for TEM observations. Selected area electron diffraction (SAED) patterns were characterized using ElDyf software (version 2.1). The luminescence measurements were performed using a Photon Technology International (PTI) Quanta-Master 40 (QM40) UV/VIS steady state spectrofluorometer supplied with a tunable pulsed optical parametric oscillator (OPO) pumped with the third harmonic of a Nd:YAG laser. The laser system was coupled with a xenon lamp, a double 200 mm monochromator, a multimode UV/VIS PMT (R928, PTI Model 914) and a Hamamatsu detector (H10330B-75). The excitation and emission spectra were recorded with a resolution of 0.5 nm. The luminescence decay curves were recorded using a PTI ASOC-10 (USB-2500) oscilloscope with an accuracy of ±0.1 μs. All structural and optical measurements were carried out at room temperature.

Results and discussion

Structural characterization of Ho³⁺-activated glass-ceramics

The characterization of prepared Ho³⁺-doped sol–gel materials began from the structural measurements using the XRD technique, and the recorded diffractograms are collected in Fig. 1. For samples from both of the fabricated series, independent of the applied heat-treatment conditions (i.e., 700 or 900 °C), the well-visible diffraction patterns consisted of a set of lines characteristic for the LaF₃ phase with trigonal symmetry (ICDD: 032-0483). The details from the Rietveld refinement, i.e., average crystal size, lattice parameters, unit cell volume, and weighted profile (R(%)) for individual Ho³⁺-doped sol–gel samples are presented in Table S1.†

Fig. 1 The XRD patterns of Ho³⁺-doped sol–gel materials annealed at 700 °C (a) and 900 °C (b).

The average crystallite size (D) for individual sol–gel samples was calculated based on the peak broadening using the Scherrer equation as follows:

in which λ is the wavelength of the X-ray beam, β is the width of the peak, θ is the Bragg angle, and K refers to the Scherrer constant. The calculations were performed with the assumption of a pseudo-Voigt profile. For the glass-ceramics from the GC₇₀₀ series, the resultant crystallite sizes hesitate from 6.48 nm (GC1₇₀₀) to 11.81 nm (GC6₇₀₀), while for samples from the GC₉₀₀ series, the average sizes fluctuate from 13.99 nm (GC1₉₀₀) to 25.06 nm (GC3₉₀₀). It should be noted that the recorded XRD lines for GC2₇₀₀ and GC5₇₀₀ samples are not intense enough to estimate the average nanocrystal sizes.

Transmission electron microscopy (TEM) analysis was conducted for representative GC6₇₀₀ and GC6₉₀₀ samples (Fig. 2). The performed TEM investigations revealed the nanocrystalline nature of the materials; the micrographs recorded in the bright and dark fields apparently show that fabricated sol–gel glass-ceramics contain crystalline grains dispersed inside an amorphous host. The acquired selected area electron diffraction (SEAD) patterns were indexed using phase data sourced from the ICDD database with lattice parameter values obtained from XRD analysis. Based on microscopic images, it could be denoted that fluoride nanocrystals are dispersed more homogeneously within the amorphous silicate as the heat-treatment temperature was elevated from 700 °C to 900 °C. Additional energy-dispersive X-ray spectroscopy (EDS) and fast Fourier transform analysis (FFT) details are available in the ESI (Fig. S1†).

Fig. 2 TEM observations of the GC6₇₀₀ (a) and GC6₉₀₀ (b) samples. The right part of the columns presents corresponding dark field images and recorded SAED patterns. Red circles indicate theoretical Bragg positions.

Another significant aspect according to the structural changes associated with an elevation in the heat-treatment temperature (700 °C → 900 °C) is the removal of residual OH groups from the silicate sol–gel host (characterized by high-oscillation energy, ∼3500 cm⁻¹), as was previously demonstrated for analogous SiO₂-LaF₃ nano-glass-ceramics doped with Pr³⁺ ions.⁴² The elimination of OH groups also remarkably affects the photoluminescence behavior of the obtained materials, as was presented in the subsequent sections of this paper.

Optical properties of fabricated sol–gel samples

Emission spectra of Ho³⁺-doped sol–gel samples in the VIS range. The normalized photoluminescence emission spectra (PL) recorded inside the VIS region with the evaluated percentage contributions of individual luminescence bands are illustrated in Fig. 3. The PL lines were recorded upon excitation at λ_ex = 450 nm, selected from the analysis of excitation spectra (Fig. S2†). The spectra clearly revealed the presence of characteristic luminescence bands of Ho³⁺ ions, assigned to the following electronic transitions: ³F_2,3 → ⁵I₈ (480 nm), (⁵S₂,⁵F₄) → ⁵I₈ (538 nm), ⁵F₅ → ⁵I₈ (640 nm), and (⁵S₂,⁵F₄) → ⁵I₇ (748 nm). It is quite interesting that the relative contribution of green ((⁵S₂,⁵F₄) → ⁵I₈) and red (⁵F₅ → ⁵I₈) emission lines undergo progressive modifications as a result of both elevation in the heat-treatment temperature of as-prepared xerogels (1), as well as a change in the Ho³⁺ concentration of the subsequent samples in the series (2). Finally, the resultant colors of the obtained emissions shift from the yellow-toned light region to green, as was presented in the Commission Internationale de l'Eclairage (CIE) chromaticity diagrams in Fig. S3†. Thus, to evaluate the mutual connotation in the population of the (⁵S₂,⁵F₄) and the ⁵F₅ excited states, as well as to assess the relaxation dynamics from the levels, the percentage involvement of each registered PL band (β) was determined (Table S2†).

Fig. 3 The photoluminescence emission spectra (PL) of Ho³⁺ ions recorded for fabricated sol–gel nano-glass-ceramics upon excitation at λ_ex = 450 nm in the VIS range: presented as a function of increasing concentration of Ho³⁺ ions (a and c), and the percentage involvement of individual luminescence bands (b and d). The data presented in Fig. (a and b) and (c and d) refer to GCs fabricated at 700 and 900 °C, respectively.

The analysis of the percentage contribution values of individual PL bands clearly reveals that the relative share of green-toned luminescence (at λ_em = 538 nm, according to the (⁵S₂,⁵F₄) → ⁵I₈ transition of Ho³⁺) is generally greater for the GCs fabricated by annealing carried out at 900 °C, compared to those heat-treated at 700 °C. Indeed, the calculated β_green factors for subsequent samples from the GC₇₀₀ series modify as follows: 31.9% (GC1₇₀₀), 44.7% (GC2₇₀₀), 61.7% (GC3₇₀₀), 46.6% (GC4₇₀₀), 14.0% (GC5₇₀₀), and 12.5% (GC6₇₀₀); meanwhile for samples from the GC₉₀₀ series, β_green exceed the following values: 91.5% (GC1₉₀₀), 90.5% (GC2₉₀₀), 88.4% (GC3₉₀₀), 78.6% (GC4₉₀₀), 57.1% (GC5₉₀₀), and 40.1% (GC6₉₀₀). The opposite trend was found for red emission (at λ_em = 640 nm, assigned to the ⁵F₅ → ⁵I₈ transition of Ho³⁺); for samples from the GC₇₀₀ series, the β_red factor changed from 62.0% (GC1₇₀₀), through 48.8% (GC2₇₀₀), 31.7% (GC3₇₀₀), 45.5% (GC4₇₀₀), 80.0% (GC5₇₀₀), up to 81.6% (GC6₇₀₀). The β_red factor values are significantly reduced for glass-ceramics from the GC₉₀₀ series, which were estimated as: 2.6% (GC1₉₀₀), 2.9% (GC2₉₀₀), 4.7% (GC3₉₀₀), 16.6% (GC4₉₀₀), 33.1% (GC5₉₀₀) and 52.9% (GC6₉₀₀). Additionally, it is worth noting that regardless of the applied heat-treatment conditions, the percentage contribution of blue emission (λ_em = 480 nm; ³F_2,3 → ⁵I₈ transition) and red/NIR luminescence (λ_em = 748 nm; (⁵S₂,⁵F₄) → ⁵I₇ transition) is negligible (β_blue and β_red/NIR ≤ 5%) and does not change noticeably while the heat-treatment temperature was elevated from 700 to 900 °C.

Based on the above-mentioned mutual correlations in β_green and β_red values, it could be concluded that the population of the (⁵S₂,⁵F₄) excited levels is greater (in comparison with the population of the lower-lying ⁵F₅ state) for samples heat-treated at 900 °C than for glass-ceramics obtained at 700 °C. For an illustration of the differences in the β_green/β_red ratio of analogous samples from both of the prepared series, the PL spectra of representative GC1₇₀₀ and GC1₉₀₀ glass-ceramics were depicted (Fig. S4†). Such a relationship could be explained by considering the energy gap between the (⁵S₂,⁵F₄) and the ⁵F₅ states, which are approximate equivalents to the oscillation energy of the OH group (ΔE ≈ 3500 cm⁻¹); hence, the presence of OH moieties in the immediate vicinity of Ho³⁺ ions straightforwardly promotes the multiphonon relaxation (MPR) processes, resulting in a progressive depopulation of the (⁵S₂,⁵F₄) levels with simultaneous pumping of the ⁵F₅ state. Taking into account that β_red values are generally higher for samples from the GC₇₀₀ series, it could be assumed that the residual OH groups significantly affect the population of the lower-lying ⁵F₅ excited state of this part of Ho³⁺ ions, which remained inside the silicate sol–gel host (even though we expect that part of Ho³⁺ ions migrated into the crystallized LaF₃ nanophase). Furthermore, the dominant contribution of the β_green factor for samples from the GC₉₀₀ series clearly suggests that the MPR to the ⁵F₅ level is suppressed and the electrons occupy primarily the (⁵S₂,⁵F₄) states; the denoted changes in β_green and β_red parameters with an elevation in heat-treatment temperature point to two simultaneous factors. The first is correlated with the efficient removal of residual OH groups from the silicate host; thus, Ho³⁺ ions that remained in the silicate network are surrounded by Qⁿ [SiO₄] tetrahedral units with a lower oscillation energy (hω ≈ 1220 cm⁻¹). The second of the above-mentioned factors is associated with more efficient migration of Ho³⁺ into the LaF₃ crystal lattice with a particularly low phonon energy (hω ≈ 350 cm⁻¹), which significantly reduces the probability of MPR occurring. Those considerations could also be confirmed by the augmented PL intensity for samples heat-treated at 900 °C compared to glass-ceramics obtained at 700 °C, as was shown for the representative GC1₇₀₀ and GC1₉₀₀ samples in Fig. S4.† Based on the conclusions from the analysis of PL spectra, the dominant transitions for samples annealed both at 700 and 900 °C were presented in the energy level diagram for Ho³⁺ ions in Fig. 4.

Fig. 4 The energy level diagrams of Ho³⁺ ions representing the dominant PL bands at 700 and 900 °C.

Another interesting correlation is associated with the modification in the mutual participation of β_green and β_red parameters as a function of progressively growing Ho³⁺ concentration in subsequent samples from the both prepared series. In the case of the GC₇₀₀ series, the change in the La³⁺ : Ho³⁺ molar ratio from 0.997 : 0.003 (GC1₇₀₀) to 0.988 : 0.012 (GC3₇₀₀) resulted in the increasing involvement of the (⁵S₂,⁵F₄) → ⁵I₈ green emission; meanwhile, a further decrease in the La³⁺ : Ho³⁺ molar ratio to 0.88 : 0.12 (GC6₇₀₀) favors the contribution of ⁵F₅ → ⁵I₈ red luminescence. It is anticipated that the observed tendency in values of β_green and β_red factors for GC1₇₀₀–GC3₇₀₀ glass-ceramics is probably correlated with the slightly enhanced ability of Ho³⁺ ions to be incorporated into precipitated LaF₃ nanocrystals (as the concentration of Ho³⁺ progressively grows). However, unexpectedly, for GC3₇₀₀–GC6₇₀₀ the trend in β_green and β_red factor values assumed the opposite direction, and finally, the involvement of the (⁵S₂,⁵F₄) → ⁵I₈ green luminescence noticeably decreased. Herein, it should be noted that the concentration quenching process (CQ_Ho) started to occur from the GC3₇₀₀ sample; thus, the observed mutual modifications in the correlation between β_green and β_red parameters seem to be strictly associated with relaxation dynamics from the individual (⁵S₂,⁵F₄) and the ⁵F₅ excited levels. Indeed, the obtained results indicate that the relaxation dynamics (realized by, i.e., CR mechanisms, resonance energy transfer (RET), etc.) from the (⁵S₂,⁵F₄) levels are much faster than that from the ⁵F₅ state. Such differences are responsible for faster quenching of the (⁵S₂,⁵F₄) → ⁵I₈ green luminescence line compared to the ⁵F₅ → ⁵I₈ red emission, causing the involvement of the latter band on the overall PL spectra to grow as the Ho³⁺ concentration increases in GC3₇₀₀–GC6₇₀₀ samples. Similarly, for the glass-ceramics fabricated at 900 °C, the CQ_Ho process begins from the GC1₉₀₀ sample with the lowest Ho³⁺ concentration in the series; thus, a progressive diminishment in the contribution of the (⁵S₂,⁵F₄) → ⁵I₈ green emission was noted, conversely to the ⁵F₅ → ⁵I₈ red luminescence, the percentage involvement of which undergoes augmentation. The observed modifications in the β_green/β_red ratio are closely related to the relaxation dynamics from the (⁵S₂,⁵F₄) and the ⁵F₅ excited states. Our assumptions about the relaxation dynamics could also be confirmed by analysis of the τ(⁵S₂,⁵F₄) and the τ(⁵F₅) decay times (see section Decay analysis of Ho³⁺-doped GCs). The tunability in the color of emitted light, correlated color temperatures (CCTs), and color purities (CPs) were discussed and are depicted in the ESI (Table S3†).

Emission spectra of Ho³⁺-doped glass-ceramics in NIR and MIR regions. The PL spectra recorded in the NIR range with an estimated contribution of each luminescence band for both series of Ho³⁺-doped glass-ceramics are shown in Fig. 5. As a result, several emission bands assigned to the following transitions: ⁵F₅ → ⁵I₇ (∼0.97 μm), (⁵S₂,⁵F₄) → ⁵I₆ (∼1.02 μm), ⁵I₆ → ⁵I₈ (∼1.19 μm), (⁵S₂,⁵F₄) → ⁵I₅ (∼1.35 μm), and ⁵F₅ → ⁵I₆ (∼1.45 μm), were registered. Additionally, in the farther-located MIR region (also shown in Fig. 5), another luminescence band with a maximum located near ∼2.0 μm (corresponding to the ⁵I₇ → ⁵I₈ electronic transition) was recorded. It was observed that the intensities of PL bands inside the NIR region for nano-glass-ceramics fabricated at 700 °C overall increase for GC1₇₀₀–GC3₇₀₀ samples, but for GC4₇₀₀–GC6₇₀₀ glass-ceramics, the recorded intensity diminishes. For samples from the series fabricated at 900 °C, the NIR luminescence intensities decrease from the GC2₉₀₀ sample, clearly indicating that the CQ_Ho process occurred parallelly with an increase in the concentration of Ho³⁺ ions. In the case of emission recorded at ∼2 μm (MIR region), the CQ_Ho process starts to reveal from GC4₇₀₀ nano-glass-ceramics (for samples from GC₇₀₀ series), and from GC2₉₀₀ (for samples from GC₉₀₀ series, adequately).

Fig. 5 The PL spectra of Ho³⁺ ions recorded for prepared GCs upon excitation at λ_ex = 450 nm in the NIR range with the percentage contribution of individual emission bands (a and b), and the PL spectra registered in the MIR range for samples from the prepared GC₇₀₀ and GC₉₀₀ series (c).

Interestingly, it is well-visible that the PL profile in the NIR range is clearly determined by applied heat-treatment conditions; factually, the luminescence from the ⁵F₅ state is significantly suppressed for samples from the GC₉₀₀ series, as is also shown in Fig. S5† for the representative GC4_x samples. Moreover, the PL intensity for glass-ceramics obtained at 900 °C is significantly greater compared to those registered for analogous samples from the GC₇₀₀ series. Therefore, to establish the mutual connotations in the preferable location of Ho³⁺ ions (amorphous sol–gel framework or fluoride crystal environment), and in the population of the (⁵S₂,⁵F₄) and the ⁵F₅ excited levels, the percentage involvement of individual NIR emission bands in the range of 950 nm to 1500 nm (β) was calculated, similar to that done for the VIS spectral range (Table S4†).

Based on the conclusions previously drawn, concerning the population of the (⁵S₂,⁵F₄) and ⁵F₅ levels (from PL spectra recorded in the VIS spectral range), the percentage contributions of the following luminescence bands were also estimated: ⁵F₅ → ⁵I₇ (β_{NIR/0.97 μm}), (⁵S₂,⁵F₄) → ⁵I₆ (β_{NIR/1.02 μm}), (⁵S₂,⁵F₄) → ⁵I₅ (β_{NIR/1.35 μm}), and ⁵F₅ → ⁵I₆ (β_{NIR/1.45 μm}). For samples fabricated at 700 °C, the β values changed in the GC1₇₀₀–GC6₇₀₀ sequence as follows: 16.0%, 9.5%, 5.0%, 7.7%, 23.7%, and 20.3% (β_{NIR/0.97 μm}); 6.5%, 8.5%, 11.8%, 6.8%, and 1.5% (β_NIR/1.02μm); 24.3%, 25.4%, 29.5%, 18.5%, 11.4%, and 8% (β_{NIR/1.35 μm}); 43.7%, 42.5%, 23.9%, 26.5%, 51.7%, and 63.1% (β_{NIR/1.45 μm}). For samples heat-treated at 900 °C, β_{NIR/0.97 μm} values did not exceed 0.5%; β_{NIR/1.02 μm} values initially changed from 6.3% (GC1₉₀₀), through 3.8% (GC2₉₀₀), up to 7.4% (GC3₉₀₀), but then – for GC4₉₀₀–GC6₉₀₀ samples – drastically decreased to ≤0.5%. The β_{NIR/1.35 μm} values modify as follows: 14.6%, 9.0%, 19.3%, 11.0%, 1.7%, and 1.1%. The β_{NIR/1.45 μm} values do not exceed 2.1%. Herein, it could also be observed that the percentage involvement of PL bands associated with the electronic transitions from the (⁵S₂,⁵F₄) levels, β_{NIR/1.02 μm} and β_{NIR/1.35 μm} gradually increases for samples heat-treated at 700 °C (GC1₇₀₀-GC3₇₀₀) with a nominal La³⁺ : Ho³⁺ molar ratio modified from 0.997 : 0.003 to 0.988 : 0.012; meanwhile, the appropriate β parameters evaluated for transitions from the ⁵F₅ state (i.e., β_{NIR/0.97 μm} and β_{NIR/1.45 μm}) tend to decline. Thus, we could assume that such a correlation is probably associated with the slightly growing efficiency of Ho³⁺ migration into the LaF₃ crystal lattice, simultaneously as the concentration of optically active ions increases (which may decrease the involvement of MPR from the upper-lying (⁵S₂,⁵F₄) levels to the lower-lying ⁵F₅ one). Upon exceeding the La³⁺:Ho³⁺ molar ratio beyond 0.988 : 0.012 (GC3₇₀₀), the trend in the percentage contribution starts to reverse, i.e., β_{NIR/1.02 μm} and β_NIR/1.35μm coefficients diminished, while the values of β_NIR/0.97μm and β_{NIR/1.45 μm} parallelly increased. The indicated tendency originates in the CQ_Ho process, which progressively occurs for GC4₇₀₀–GC6₇₀₀ glass-ceramics. Again, the non-radiative relaxation from the (⁵S₂,⁵F₄) levels proceeds more dynamically than from the ⁵F₅ state, which explains the mutual correlations in the intensities of the following PL bands: ⁵F₅ → ⁵I₇, (⁵S₂,⁵F₄) → ⁵I₆, (⁵S₂,⁵F₄) → ⁵I₅, and ⁵F₅ → ⁵I₆. Additionally, it is worth noting that the (⁵S₂,⁵F₄) → ⁵I₆ (∼1.02 μm) and ⁵F₅ → ⁵I₆ (∼1.45 μm) electronic transitions favor the occupation of the ⁵I₆ state; hence, the ⁵I₆ → ⁵I₈ (∼1.19 μm) transition undergoes in the analyzed NIR spectral range. Considering the subsequent values of the β_NIR/1.19μm coefficient for glass-ceramics fabricated at 700 °C: 9.5% (GC1₇₀₀), 14.1% (GC2₇₀₀), 29.8% (GC3₇₀₀), 40.5% (GC4₇₀₀), 11.7% (GC5₇₀₀), and 12.8% (GC6₇₀₀), it was assumed that for samples with the lowest Ho³⁺ concentration, the relatively low probability for the ⁵I₆ → ⁵I₈ transition to occur (≤14.1%) should be related to the high impact of OH groups on the luminescence of Ho³⁺ ions located inside the amorphous silicate sol–gel host. Indeed, since Ho³⁺ ions are surrounded by functional groups or bonds with a high oscillation energy, like OH moieties, the non-radiative deactivation of the ⁵I₆ level could easily occur (the energy gap between the ⁵I₆ and the lower-lying ⁵I₇ state equals about ∼3550 cm⁻¹), resulting in a relative low intensity of luminescence observed at ∼1.19 μm. However, as the concentration of Ho³⁺ ions increases, the probability of MPR from the ⁵I₆ state gradually decreases and the share of radiative ⁵I₆ → ⁵I₈ emission enlarges. For the last samples from the series with the highest Ho³⁺ concentration, i.e., GC5₇₀₀ and GC6₇₀₀ (for which the CQ_Ho occurs), another decrease in β_{NIR/1.19 μm} parameter values was denoted; it indicates a diminishing probability of radiative relaxation from the ⁵I₆ level. Moreover, it is also well-observed that the luminescence from the ⁵I₆ state quenched faster in comparison with the emission from the ⁵F₅ level; thus, we concluded that the non-radiative relaxation from the former excited level is more dynamic than from the latter one. The dominant transitions inside the IR region for the GC₇₀₀ series (and also for the GC₉₀₀ one) are presented in the energy level scheme in Fig. 4.

In the case of glass-ceramics from the GC₉₀₀ series, it is well-observed that the luminescence from the ⁵F₅ excited state has a little contribution to the overall recorded PL spectra, i.e., β_{NIR/0.97 μm} ≤ 0.5% and β_{NIR/1.45 μm} ≤ 2.1%. Factually, the population of the ⁵F₅ level for samples heat-treated at 900 °C is decreased (compared to samples from the GC₇₀₀ series) due to the more efficient location of Ho³⁺ ions inside the fluoride crystal environment, where depopulation from the (⁵S₂,⁵F₄) states to the lower-lying ⁵F₅ is restricted; additionally, the depopulation of the ⁵F₅ excited level also results from elimination of residual OH groups from the silicate sol–gel host. On the other hand, it is worth noting that the percentage contribution of emissions from the (⁵S₂,⁵F₄) levels is also reduced in comparison with analogous glass-ceramics fabricated at 700 °C, and the resultant values of β_{NIR/1.02 μm} and β_NIR/1.35μm parameters do not exceed 7.4% as well as 19.3%, respectively. The dominant involvement in PL spectra was attributed to the emission corresponding to the ⁵I₆ → ⁵I₈ transition and the values of the β_NIR/1.19μm parameter for individual samples were changed from 79.1% (GC1₉₀₀) to 97.3% (GC6₉₀₀). Such a domination of the ⁵I₆ → ⁵I₈ transition could also be explained by the lowering of the local phonon energies around Ho³⁺ ions, realized by more efficient incorporation of Ho³⁺ inside the LaF₃ nanophase and removal of OH groups from the silicate network. Additionally, it was tentatively assumed that such a dominant contribution of the ⁵I₆ → ⁵I₈ luminescence could probably be also associated with the CR process leading to the population of the ⁵I₆ state from the upper-lying (⁵S₂,⁵F₄) by the involvement of the {(⁵S₂,⁵F₄) + ⁵I₈ → ⁵I₆ + ⁵I₆} mechanism.⁴³

Similarly, as for the NIR spectral region, the PL spectra in the MIR range were also compared for representative samples, as is shown in Fig. S6.† It is well-visible that the luminescence near ∼2 μm is more intense for nano-glass-ceramics fabricated at 900 °C, than for samples fabricated by controlled heat-treatment performed at 700 °C. Thus, the differences in the mutual band intensity are determined by two factors: efficient removal of residual OH groups from the immediate vicinity of Ho³⁺ ions in the host, and their entry into the LaF₃ nanophase, which favors the radiative emission of the ⁵I₇ → ⁵I₈ transition. In addition, the population of the ⁵I₇ excited state could also be associated with the CR process from the (⁵S₂,⁵F₄) levels, realized by the following {(⁵S₂,⁵F₄) + ⁵I₈} → {⁵I₄ + ⁵I₇} simultaneous electronic transitions.⁴³

Decay analysis of Ho³⁺-doped GCs. To further verify the photoluminescence behavior of fabricated Ho³⁺-doped GC materials, the luminescence decay curves from the individual excited states, i.e., (⁵S₂,⁵F₄), ⁵F₅, ⁵I₆, and ⁵I₇ were evaluated. The registered curves are presented in Fig. 6 (for samples from the GC₇₀₀ series) and Fig. 7 (for glass-ceramics from the GC₉₀₀ series). The decay profiles follow the second-order exponential nature; therefore, the average lifetimes were calculated using the equation given below:in which A₁ and A₂ are the residual weighting factors, and τ₁ and τ₂ are the fast and slow decay components, respectively. It could be observed that for GC1₇₀₀–GC3₇₀₀ glass-ceramic samples, the evaluated lifetimes – independent of the considered excited level of Ho³⁺ ions – undergo gradual prolongation, while after exceeding the critical concentration of Ho³⁺, a successive shortening in decay times was denoted (GC4₇₀₀–GC6₇₀₀). The identified progressive reduction in luminescence lifetimes is strictly associated with CQ_Ho. It is well-observable that among the evaluated lifetimes for the samples from the GC₇₀₀ series, i.e., τ(⁵S₂,⁵F₄), τ(⁵F₅), τ(⁵I₆), and τ(⁵I₇), the longest values were denoted for the ⁵I₇ excited level (with τ(⁵I₇) = 1.5 ms for the GC3₇₀₀ sample). Based on the available literature data (Table S5†), the estimated τ(⁵I₇) lifetimes are very close to those determined for sol–gel silica glasses, as was presented in the previous work.⁴⁴ It should be noted that for ceramics, e.g., CaF₂ : Ho³⁺ (in which optically active Ho³⁺ ions are efficiently incorporated into the fluoride crystal lattice with a low-oscillation energy), the resultant lifetimes are much elongated, even to τ(⁵I₇) ≈ 30 ms.⁴⁵ Thus, it could be justified that according to the fabricated glass-ceramics from the GC₇₀₀ series, Ho³⁺ ions are mostly located inside the amorphous silicate sol–gel host; meanwhile, the minor part of accessible Ho³⁺ is incorporated inside the LaF₃ nanocrystal fraction. These assumptions are consistent with the analysis of PL spectra, inside both the VIS and the NIR ranges (Fig. 3 and 5), which indicates that for samples from the GC₇₀₀ series, the population of the lower-lying ⁵F₅ excited level (typical of high-oscillation environments) is more efficient in comparison with analogous samples from the GC₉₀₀ series; similarly, the population of the upper-lying (⁵S₂,⁵F₄) states is less efficient for the GC₇₀₀ nano-glass-ceramics in comparison with samples fabricated at 900 °C. Hence, in the case of glass-ceramics annealed at 700 °C, the major part of Ho³⁺ ions are probably located inside the high-phonon energy environment with residual OH groups, favoring the relaxation realized by the (⁵S₂,⁵F₄) → ⁵F₅ non-radiative MPR channel.

Fig. 6 Luminescence decay curves for the individual excited states, i.e., (⁵S₂ + ⁵F₄), ⁵F₅, ⁵I₆, and ⁵I₇ of Ho³⁺ in the prepared glass-ceramics from the series heat-treated at 700 °C.

Fig. 7 Luminescence decay curves for the individual excited states, i.e., (⁵S₂ + ⁵F₄), ⁵F₅, ⁵I₆, and ⁵I₇ of Ho³⁺ in the fabricated sol–gel glass-ceramics from the series annealed at 900 °C.

To evaluate the impact of CQ_Ho on luminescence in the studied glass-ceramic samples, the efficiencies of the process were calculated based on the below equation:

where τGCx refers to the lifetime of the appropriate excited state for the selected sample, while τGC_ref corresponds to the longest decay time of the same excited level evaluated in the series (the reference sample for the series heat-treated at 700 °C is GC3₇₀₀; in the case of materials fabricated by annealing at 900 °C, the reference samples are GC1₉₀₀ (according to the lifetimes from the (⁵S₂,⁵F₄) and ⁵F₅ excited levels) and GC2₉₀₀ (lifetimes from the ⁵I₆ and the ⁵I₇ states)). The obtained results for each of the excited states of Ho³⁺ in glass-ceramics heat-treated at 700 °C are graphically demonstrated in Fig. S7.† By analyzing the decay lifetimes for the (⁵S₂,⁵F₄) and the ⁵F₅ levels, it could be stated that η_CQ(⁵S₂,⁵F₄) for the subsequent samples increases in the following order: 45.4% (GC4₇₀₀), 68.8% (GC5₇₀₀), and 90.2% (GC6₇₀₀), while η_CQ(⁵F₅) increases as follows: 22.8% (GC4₇₀₀), 32.8% (GC5₇₀₀) and 57.0% (GC6₇₀₀). The performed calculations clearly suggest that η_CQ reached greater values for the (⁵S₂,⁵F₄) levels compared to the lower-lying ⁵F₅ state, and therefore, the dynamics of the non-radiative relaxation from the former progresses faster with a gradual increase of the Ho³⁺ concentration in the subsequent samples from the series. The observed differences in η_CQ values for these two excited levels explain why the β_red coefficients increase (from 31.7% (GC3₇₀₀) to 81.6% (GC6₇₀₀)), but β_green coefficients parallelly diminish (from 61.7% (GC3₇₀₀) to 12.5% (GC6₇₀₀)) with the growing Ho³⁺ concentration. In addition, the calculations of η_CQ(⁵I₆) were also performed, and the resultant values were estimated as: 44.9% (GC4₇₀₀), 53.5% (GC5₇₀₀), and 78.7% (GC6₇₀₀). Based on the previously performed analysis of the individual β_NIR coefficients, it was stated that the luminescence quenching from the ⁵I₆ level (but also from the (⁵S₂,⁵F₄) states) is faster compared to the quenching from the ⁵F₅ excited state; the same conclusion could be drawn from the analysis of calculated η_CQ(⁵I₆) and η_CQ(⁵F₅) efficiencies. Finally, the evaluated values of η_CQ(⁵I₇) changed as follows: 38.6% (GC4₇₀₀), 62.1% (GC5₇₀₀), and 87.0% (GC6₇₀₀). Thus, it could be stated that emission from the ⁵I₇ state also quenched faster than from the ⁵F₅ level; generally, the concentration quenching efficiencies decrease in the following order: (⁵S₂,⁵F₄) > ⁵I₇ > ⁵I₆ > ⁵F₅.

The further analysis of the obtained results involved the luminescence decay times for glass-ceramics from the GC₉₀₀ series. Hence, it could be observed that the gradual shortening in the decay times is achieved from GC1₉₀₀ (for τ(⁵S₂,⁵F₄) and τ(⁵F₅)) or from the GC2₉₀₀ sample for (τ(⁵I₆) and τ(⁵I₇)). The progressive shortening in luminescence lifetimes is obviously correlated with the CQ_Ho process; for GC₉₀₀ samples, the CQ_Ho process begins from samples with lower Ho³⁺ concentrations compared with glass ceramics fabricated at 700 °C. The longest lifetimes (in the millisecond range) were evaluated for the ⁵I₆ level (up to 10.19 ms for the GC2₉₀₀ sample) and the ⁵I₇ one (up to 8.44 ms for the same sample) in comparison with the decay times estimated for the (⁵S₂,⁵F₄) and ⁵F₅ excited states (in the microsecond range). Moreover, regardless of the considered excited level, the determined lifetimes are prolonged for nano-glass-ceramic samples from the GC₉₀₀ series, compared to those evaluated for samples annealed at 700 °C (it is particularly noticeable for τ(⁵I₆) and τ(⁵I₇)). Those differences in luminescence lifetimes for glass ceramics from the GC₇₀₀ and GC₉₀₀ series are strictly correlated with changes in phonon energies in the local environment around Ho³⁺ ions. These assumptions are in good correlation with our earlier expectations about the greater population of the (⁵S₂,⁵F₄) and ⁵I₆ levels for the GC₉₀₀ samples than for the GC₇₀₀ glass-ceramics, and conversely – lower occupancy of the ⁵F₅ excited state for the GC₉₀₀ series than for the GC₇₀₀ samples. It should also be noted that estimated luminescence lifetimes, i.e., τ(⁵S₂,⁵F₄), τ(⁵F₅), and τ(⁵I₆), for glass-ceramics annealed at 900 °C, are elongated in comparison with other glass-ceramics or ceramic materials described in the literature, as was demonstrated in Table S5.† Only the τ(⁵I₇) lifetimes are comparable to those reported for ceramics, e.g., Y₂O₃:Ho³⁺ and Y₃Al₅O₁₂:Ho³⁺,²⁴ (Lu,Sc)₂O₃:Ho³⁺,⁴⁶ or Y₃NbO₇:Ho³⁺.⁴⁷

Similarly, as for the samples from the GC₇₀₀ series, the impact of the CQ_Ho process on luminescence behavior was evaluated. The results are graphically illustrated in Fig. S8;† the η_CQ(⁵S₂,⁵F₄) values were evaluated as 6.4% (GC2₉₀₀), 11.8% (GC3₉₀₀), 48.0% (GC4₉₀₀), 62.5% (GC5₉₀₀), and 74.9% (GC6₉₀₀). Subsequently, the η_CQ(⁵F₅) reached the following values: 8.6% (GC2₉₀₀), 6.8% (GC3₉₀₀), 25.7% (GC4₉₀₀), 37.4% (GC5₉₀₀), and 45.3% (GC6₉₀₀). It could be observed that similar to the samples from the GC₇₀₀ series, the η_CQ(⁵S₂,⁵F₄) values increase faster with the growing Ho³⁺ concentration than the η_CQ(⁵F₅) values. It means that the non-radiative relaxation dynamics from the (⁵S₂,⁵F₄) levels is higher than from the ⁵F₅ excited state, and it explains the progressive diminishment in the β_green values (from 91.5% (GC1₉₀₀) to 40.1% (GC6₉₀₀)) in comparison with the progressively growing β_red parameter (from 2.6% (GC1₉₀₀) to 52.9% (GC6₉₀₀)). Hence, the increasing contribution of the ⁵F₅ → ⁵I₈ PL line with the growing concentration of Ho³⁺ (in comparison with the decreasing contribution of the (⁵S₂,⁵F₄) → ⁵I₈ emission band) is also observed for samples from the GC₉₀₀ series with more efficiently incorporated Ho³⁺ ions inside the fluoride crystal phase, than in the case of the GC₇₀₀ glass-ceramics. In addition, by analyzing the correlations in the mutual percentage contribution of the (⁵S₂,⁵F₄) → ⁵I₆/⁵I₅ and ⁵I₆ → ⁵I₈ NIR luminescence bands, it could be assessed that the non-radiative relaxation from the (⁵S₂,⁵F₄) excited states is more dynamic than from the lower-lying ⁵I₆ level, which was also justified by the calculations of the η_CQ(⁵I₆) values, which changed as follows: 10.6% (GC3₉₀₀), 19.4% (GC4₉₀₀), 23.0% (GC5₉₀₀), and 45.5% (GC6₉₀₀). Finally, the η_CQ efficiencies were also calculated for the ⁵I₇ level, and the resultant values change in the following order: 9.0% (GC3₉₀₀), 18.6% (GC4₉₀₀), 39.3% (GC5₉₀₀) and 45.4% (GC6₉₀₀). The values are lower than the calculated η_CQ(⁵S₂,⁵F₄) efficiencies for analogous glass-ceramic samples with the same concentration of Ho³⁺ ions. In this case, the concentration quenching efficiencies decrease in general in the following order: (⁵S₂,⁵F₄) > ⁵F₅ ≈ ⁵I₆ ≈ ⁵I₇.

Conclusions

In this work, we have synthesized and characterized a series of sol–gel nano-glass-ceramics (GCs) containing a LaF₃ nanocrystal phase and activated by Ho³⁺ ions. The prepared sol–gel GCs exhibit luminescence bands within the VIS, NIR, and MIR spectral ranges, which were assigned to the intra-configurational electronic transitions within the 4f¹⁰ manifold of Ho³⁺, and the most prominent emission lines were assigned to the following ones: (⁵S₂,⁵F₄) → ⁵I₈ (green), ⁵F₅ → ⁵I₈ (red), ⁵I₆ → ⁵I₈ (NIR, ∼1.19 μm), and ⁵I₇ → ⁵I₈ (MIR, ∼2.0 μm). Based on the mutual intensities of the recorded PL bands, their percentage contributions (β) were assessed; for the VIS range, it was established that – in general – the relative contribution of the ⁵F₅ → ⁵I₈ red emission band is greater for samples fabricated at 700 °C; however, the green emission ((⁵S₂,⁵F₄) → ⁵I₈) dominates for samples annealed at 900 °C. It was concluded that the relative population of the (⁵S₂,⁵F₄) and lower-lying ⁵F₅ states is strongly correlated with the phonon energy in the nearest proximity around Ho³⁺ ions in the host and (un)involvement of the (⁵S₂,⁵F₄) → ⁵F₅ MPR channel. Indeed, based on the correlations in β_red and β_green coefficients, it was stated that the residual OH groups with high-oscillation energy (from the silicate sol–gel host) affect the population of the ⁵F₅ level for GCs heat-treated at 700 °C. Simultaneously, their removal by temperature elevation to 900 °C and the more efficient entry of Ho³⁺ ions into the LaF₃ nanocrystal lattice favor the suppression of the relaxation from the (⁵S₂,⁵F₄) states. Additionally, the emission recorded in the NIR (∼1.19 μm) and MIR (∼2.0 μm) regions is enhanced for nano-glass-ceramics fabricated at 900 °C. Notably, it was proven that the relaxation dynamics decrease in the following order: (⁵S₂,⁵F₄) > ⁵I₇ > ⁵I₆ > ⁵F₅ for samples from the series fabricated at 700 °C, and as follows: (⁵S₂,⁵F₄) > ⁵F₅ ≈ ⁵I₆ ≈ ⁵I₇ for nano-glass-ceramics heat-treated at 900 °C. In summary, the prepared Ho³⁺-doped GCs are able to emit green light with high color purity (CP) reaching even 98%, and generate long-lived NIR/MIR emissions at ∼1.19 μm (up to τ(⁵I₆) = 10.19 ms) and ∼2.0 μm (up to τ(⁵I₇) = 8.44 ms), which could predispose them for use in optoelectronic devices.

Author contributions

N.P.: writing – original draft, supervision, methodology, formal analysis, investigation, conceptualization, and data curation. J.Ś.: investigation, review and editing. B.H. and M.Z.: investigation, formal analysis, data curation, review and editing. M.S. and W.A.P.: resources, review and editing.

Data availability

The data that support the findings of this study have been included as part of the ESI† and are also available from the corresponding author upon reasonable request.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

The research activities are co-financed by the funds granted under the Research Excellence Initiative of the University of Silesia in Katowice.

References

1. Y. Ye, Z. Tang, Z. Ji, H. Xiao, Y. Liu, Y. Qin, L. Liang, J. Qi and T. Lu, Opt. Mater., 2021, 121, 111643.
2. V. Reddy Prasad, S. Damodaraian and Y. C. Ratnakaram, Opt. Mater., 2018, 78, 63–71.
3. W.-J. Zhang, Q.-J. Chen, Q. Qian and Q.-Y. Zhang, J. Am. Ceram. Soc., 2012, 95, 663–669.
4. L. Gomes, V. Fortin, M. Bernier, R. Vallée, S. Poulain, M. Poulain and S. D. Jackson, Opt. Mater., 2016, 60, 618–626.
5. M. Rajesh, D. Siva Raju and T. Kanagasekaran, Ceram. Int., 2024, 50, 49200–49209.
6. M. Ravi Prakash, G. Neelima, V. K. Kummara, N. Ravi, C. S. Dwaraka Viswanath, T. Subba Rao and S. Mahaboob Jilani, Opt. Mater., 2019, 94, 436–443.
7. M. Venkateswarlu, S. Mahamuda, K. Swapna, M. V. V. K. S. Prasad, A. Srinivasa Rao, S. Shakya, A. Mohan Babu and G. Vijaya Prakash, J. Lumin., 2015, 163, 54–71.
8. P. Rekha Rani, M. Venkateswarlu, K. Swapna, S. Mahamuda, M. V. V. K. Srinivas Prasad and A. S. Rao, Solid State Sci., 2020, 102, 106175.
9. K. Mariselvam and J. Liu, J. Solid State Chem., 2021, 293, 121793.
10. Y. Liu, X. Qin, L. Gan, G. Zhou, S. Hu, Z. Wang, J. Jiang, T. Zhang and H. Chen, Materials, 2024, 17, 402.
11. V. Singh, A. A. Bhat, A. R. Kadam, S. Saravanakumar, P. K. Tripathi, S. J. Dhoble and J. B. Joo, J. Electron. Mater., 2024, 53, 6384–6394.
12. V. Singh, A. A. Bhat, C. M. Mehare and S. J. Dhoble, J. Electron. Mater., 2024, 53, 4857–4868.
13. K. Sathya Moorthy, E. Sailatha, P. Muralimanohar, K. S. Nagaraja and C. Karnan, Inorg. Chem. Commun., 2023, 153, 110770.
14. S. Ji, Y. Song, Z. Wang, C. Shen, J. Lin, B. Xiao, Q. Feng, Q. Du, H. Xu and Z. Cai, Nanophotonics, 2022, 11, 1603–1609.
15. Y. Li, W. Wang, X. Huang, H. Chen, Y. Pan and X. Wei, J. Lumin., 2021, 238, 118280.
16. S. A. Jupri, S. K. Ghoshal, M. F. Omar and N. N. Yusof, J. Alloys Compd., 2018, 753, 46–456.
17. A. S. Alqarni, R. Hussin, S. N. Alamri and S. K. Ghoshal, Res. Phys., 2020, 17, 103102.
18. V. dos Santos de Souza, F. J. Caixeta, K. de Oliveira Lima and R. R. Gonçalves, J. Lumin., 2022, 248, 118978.
19. B. J. Chen, L. F. Shen, E. Y. B. Pun and H. Lin, Opt. Commun., 2011, 284, 5705–5709.
20. X. Zhu, J. Zong, A. Miller, K. Wiersma, R. A. Norwood, N. S. Prasad, A. Chavez-Pirson and N. Peyghambarian, Proc. SPIE, Fiber Lasers X: Technology, System, and Applications, 2013, 8601, 86010Y.
21. S. Wang, C. Li, C. Yao, S. Jia, Z. Jia, G. Qin and W. Qin, Opt. Mater., 2017, 64, 421–426.
22. S. Mahamuda, K. Swapna, P. Packiyaraj, A. Srinivasa Rao and G. Vijaya Prakash, Opt. Mater., 2013, 36, 362–371.
23. M. Liu, J. Zhang, J. Zhang, Z. Zhang, G. Farrell, G. Brambilla, S. Wang and P. Wang, J. Lumin., 2021, 238, 118265.
24. P. Loiko, L. Basyrova, R. Maksimov, V. Shitor, M. Baranov, F. Starecki, X. Mateos and P. Camy, J. Lumin., 2021, 240, 118460.
25. Y. Liu, Q. Li, L. Meng, Q. Pang, Y. Wang, H. Yao and Q. Chen, Opt. Mater., 2024, 154, 115694.
26. L. Tao, Y. H. Tsang, B. Zhou, B. Richards and A. Jha, J. Non-Cryst. Solids, 2012, 358, 1644–1648.
27. R. Cao, Y. Lu, Y. Tian, F. Huang, Y. Guo, S. Xu and J. Zhang, Sci. Rep., 2016, 6, 37873.
28. J. Pan, R. Xu, Y. Tian, K. Li, L. Hu and J. Zhang, Opt. Mater., 2010, 32, 1451–1455.
29. H. Ebendorff-Heidepriem, I. Szabó and Z. E. Rasztovits, Opt. Mater., 2000, 14, 127–136.
30. Q. Liu, Y. Tian, C. Wang, F. Huang, X. Jing, J. Zhang, X. Zhang and S. Xu, Phys. Chem. Chem. Phys., 2017, 19, 29833–29839.
31. P. Babu, U. R. Rodríguez-Mendoza, V. Lavín and R. Praveena, Opt. Mater., 2024, 153, 115609.
32. M. A. Marzouk, I. M. Elkashef and H. A. Elbatal, Appl. Phys. A, 2019, 125, 97.
33. Y. Gao, Y. Hu, D. Zhou and J. Qiu, J. Nanosci. Nanotechnol., 2016, 16, 3744–3748.
34. P. Liu, C. Han, X. Zhang, X. Li, Y. Wan, H. Zhang and C. Su, Ceram. Int., 2024, 50, 31164–31172.
35. F. Lahoz, S. E. Hernández, N. E. Capuj and D. Navarro-Urrios, Appl. Phys. Lett., 2007, 90, 201117.
36. K. Driesen, V. K. Tikhomirov, C. Görller-Walrand, V. D. Rodriguez and A. B. Seddon, Appl. Phys. Lett., 2006, 88, 073111.
37. G. Zhang, H. Wang, S. Xu, F. Jia, H. Zhang and C. Su, J. Non-Cryst. Solids, 2023, 619, 122567.
38. A. Santana-Alonso, A. C. Yanes, J. Méndez-Ramos and J. Del-Castillo, Sci. Adv. Mater., 2013, 5, 592–597.
39. S. Jiang, H. Guo, X. Wei, C. Duan and M. Yin, J. Lumin., 2014, 152, 195–198.
40. J. del-Castillo, A. C. Yanes, J. Méndez-Ramos, V. K. Tikhomirov, V. V. Moshchalkov and V. D. Rodríguez, J. Sol-Gel Sci. Technol., 2010, 53, 509–514.
41. J. J. Velázquez, A. C. Yanes, J. del Castillo, J. Méndez-Ramos and V. D. Rodríguez, Phys. Status Solidi A, 2007, 204, 1762–1768.
42. N. Pawlik, T. Goryczka, M. Zubko, J. Śmiarowska and W. A. Pisarski, Nanoscale, 2024, 16, 4249–4265.
43. N. Bednarska-Adam, M. Kuwik, E. Pietrasik, W. A. Pisarski, T. Goryczka, B. Macalik and J. Pisarska, Materials, 2022, 15, 5263.
44. X. Wang, W. Xu, S. Wang, C. Yu, D. Chen and L. Hu, J. Alloys Compd., 2016, 657, 478–482.
45. Z. Wan, W. Li, B. Mei, Z. Liu and Y. Yang, J. Lumin., 2020, 223, 117188.
46. W. Jing, P. Loiko, J. M. Serres, Y. Wang, E. Kifle, E. Vilejshikova, M. Augiló, F. Díaz, U. Griebner, H. Huang, V. Petrov and X. Mateos, J. Lumin., 2018, 203, 145–151.
47. L. Cornet, S. Guene-Girard, J.-M. Heintz, R. Boulesteix, A. Maître and V. Jubera, Opt. Mater., 2023, 144, 114319.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d5dt00342c

---