Structure and Luminescence Properties of Transparent Germanate Glass-Ceramics Co-Doped with Ni2+/Er3+ for Near-Infrared Optical Fiber Application

An investigation of the structural and luminescent properties of the transparent germanate glass-ceramics co-doped with Ni2+/Er3+ for near-infrared optical fiber applications was presented. Modification of germanate glasses with 10–20 ZnO (mol.%) was focused to propose the additional heat treatment process controlled at 650 °C to obtain transparent glass-ceramics. The formation of 11 nm ZnGa2O4 nanocrystals was confirmed by the X-ray diffraction (XRD) method. It followed the glass network changes analyzed in detail (MIR—Mid Infrared spectroscopy) with an increasing heating time of precursor glass. The broadband 1000–1650 nm luminescence (λexc = 808 nm) was obtained as a result of Ni2+: 3T2(3F) → 3A2(3F) octahedral Ni2+ ions and Er3+: 4I13/2 → 4I15/2 radiative transitions and energy transfer from Ni2+ to Er3+ with the efficiency of 19%. Elaborated glass–nanocrystalline material is a very promising candidate for use as a core of broadband luminescence optical fibers.


Introduction
The rapid increase in the capacity of modern information technology (IT) based on the computer network and optical communications demands optical fiber amplifiers with the ultrabroad and efficient gain bandwidth in the telecommunication windows. For this reason, the possible expansion of emission bands in erbium-doped fiber amplifiers (EDFA) in different ways is a hot topic [1]. In particular, the following co-doping system of rareearth ions (RE) has been proposed, including: Er 3+ /Tm 3+ , Pr 3+ /Er 3+ , Er 3+ /Tm 3+ /Pr 3+ , Yb 3+ /Er 3+ /Tm 3+ and others [2][3][4][5]. However, the main general problem in co-doping systems is the narrowing of 4f-4f transitions and unwanted energy transfer mechanism between RE ions which strongly limits the effective emission in the near-infrared spectral range. It is well-known that possible energy transfer channels depend on the phonon energy of the host and the type of electrostatic interionic interactions [6][7][8]. In the case of low-phonon glasses, the upconversion mechanisms (ETU-energy transfer upconversion, ESA-excited state absorption) leads to the depopulation of higher energy levels, and then excitation energy is partially distributed between rare-earth ions [9]. From the other side, in matrices with high-phonon energy, where the upconversion processes are negligible, the cross-relaxation and energy migration mechanisms are more prominent, and finally, the

Materials and Methods
The germanium based GGK glasses were prepared according to the following general molar formula: (1-x-y)75GeO 2 -10Ga 2 O 3 -xZnO-5K 2 O-yNiO-0.1Er 2 O 3 , (x = 10, 15, 20, y = 0, 0.1%) by standard melting and quenching method. The homogenized set (purity of materials 99.99% Sigma-Aldrich, Saint Louis, MO, USA) was melted in a platinum crucible in an electric furnace (CZYLOK Company, Jastrzębie-Zdrój, Poland) at T = 1500 • C for 30 min. The molten glass was poured out onto a stainless-steel plate and then annealed in an air atmosphere at 560 • C for 12 h. Glass-ceramic materials were obtained by additional heat treatment of the as-melted glasses in Czylok tube furnace at 650 • C for 1-5 h. The glasses and glass-ceramics samples were labeled according to the ZnO amount (10ZnO, 15ZnO) and heating time (3 h and 5 h at 650 • C). The photos of obtained samples were presented in Figure 1a,b.
Nanomaterials 2021, 11, x FOR PEER REVIEW 3 of 16 in an electric furnace (CZYLOK Company, Jastrzębie-Zdrój, Poland) at T = 1500 °C for 30 minutes. The molten glass was poured out onto a stainless-steel plate and then annealed in an air atmosphere at 560 °C for 12 h. Glass-ceramic materials were obtained by additional heat treatment of the as-melted glasses in Czylok tube furnace at 650 °C for 1-5 h. The glasses and glass-ceramics samples were labeled according to the ZnO amount (10ZnO, 15ZnO) and heating time (3 h and 5 h at 650 °C). The photos of obtained samples were presented in Figure 1a, b. FTIR spectra were recorded with a Bruker Company Vertex 70v spectrometer (Rheinstetten, Germany). The MIR spectra were normalized to the highest peak. FTIR spectra have been decomposed using Fityk software (0.9.8 software, open-source (GPL2+)) based on the analysis of second derivatives with different degrees of smoothing. The coefficient of determination was 0.99. The standard deviation of the position of each of the component bands was ±4 cm −1 .
The excitation wavelength of 532 nm was used, and power was about 10 mW. Acquisition time was set to 30 seconds. The formed crystallites were examined by the X-ray powder diffraction (XRD) method in the range of 10° to 90° using an X'Pert Pro diffractometer (PANalytical, Eindhoven, The Netherlands). The Cu X-ray tube with Kα radiation was used. DSC measurement was performed with 10 K/min using the SETARAM Labsys thermal analyzer (Setaram Instrumentation, Caluire, France). The glass-ceramics morphology was observed by an FEI Company Nova Nano SEM 200 scanning electron microscope (Hillsboro, OR, USA) and the analyses were performed in the secondary electron FTIR spectra were recorded with a Bruker Company Vertex 70v spectrometer (Rheinstetten, Germany). The MIR spectra were normalized to the highest peak. FTIR spectra have been decomposed using Fityk software (0.9.8 software, open-source (GPL2+)) based on the analysis of second derivatives with different degrees of smoothing. The coefficient of determination was 0.99. The standard deviation of the position of each of the component bands was ±4 cm −1 .
The excitation wavelength of 532 nm was used, and power was about 10 mW. Acquisition time was set to 30 seconds. The formed crystallites were examined by the X-ray powder diffraction (XRD) method in the range of 10 • to 90 • using an X'Pert Pro diffractometer (PANalytical, Eindhoven, The Netherlands). The Cu X-ray tube with K α radiation was used. DSC measurement was performed with 10 K/min using the SETARAM Labsys thermal analyzer (Setaram Instrumentation, Caluire, France). The glass-ceramics morphology was observed by an FEI Company Nova Nano SEM 200 scanning electron microscope (Hillsboro, OR, USA) and the analyses were performed in the secondary electron mode (SE). The absorbance and emission spectra of glasses in the range of 1 µm were measured using the Stellarnet Green-Wave spectrometer (Setaram Instrumentation, Caluire, France) and a high-power LIMO laser diode (λ p = 808 nm, P opt (max) = 0.1-30 W). Spectral measurements in the range of 1000-1700 nm were carried out using Acton 2300 i monochromator (Acton Research Corporation, Acton, MA, USA) equipped with an InGaAs detector.

Structural Properties
In Figure 2, MIR spectra of 15 ZnO glass, and glass-ceramics samples doped with Ni 2+ and Er 3+ in the 400-1300 cm −1 range were presented. Due to the low-resolution technique, the presence of the nanocrystals has not been detected in spectra of glass-ceramics samples (lack of the bands in MIR spectra assigned to the bonds in crystals lattice).
Spectral measurements in the range of 1000-1700 nm were carried out using Acton 2300 i monochromator (Acton Research Corporation, Acton, MA, USA) equipped with an In-GaAs detector.

Structural Properties
In Figure 2, MIR spectra of 15 ZnO glass, and glass-ceramics samples doped with Ni 2+ and Er 3+ in the 400-1300 cm −1 range were presented. Due to the low-resolution technique, the presence of the nanocrystals has not been detected in spectra of glass-ceramics samples (lack of the bands in MIR spectra assigned to the bonds in crystals lattice).
As the nanocrystals have not been observed by the MIR spectrum the diffraction patterns of the samples were measured (Figure 5a). The amorphous state of the 15ZnO glass sample has been confirmed by a broad diffraction hump in the range of 2θ from 15 • to 40 • . In the diffraction patterns of the glass heated at 650 • C in three (15ZnO_6500C_3 h) and five (15ZnO_6500C_5 h) hours, the peaks were observed. They were assigned to the zinc gallium oxide (ZnGa 2 O 4 ) cubic phase (JCPDF: 01-086-0413) [35]. The average size of the nanocrystals was ca. 11 nm (seen also at SEM Figure 5b) which is much smaller than the visible light wavelength. Thus, obtained glass-ceramics are transparent (Figure 1a,b).
The size of the ZnGa 2 O 4 nanocrystals presented in Figure 5b was also checked according to the calculation by Scherrer's Equation (1) [36,37]. The obtained results for 15ZnO_650 • C_3 h and 15ZnO_650 • C_5 h samples have been presented in Tables 3 and 4. The crystalline size of crystals has been evaluated for diffraction peaks visible in Figure 5a. According to the calculated data, the crystalline size of nanocrystals is in the 3-18 nm range for both samples. These data are in line with the SEM results and give an overview of the crystal size distribution in the glass-ceramic samples. Moreover, the longer heating time of Ni/Er co-doped glass sample from 3 h to 5 h did not cause an increase in crystal size. The difference in the size of the crystallites for the analyzed samples is within the measurement bar (±2 nm). This suggests that with an increase in heating time of the glass sample, the number of nanocrystallines grew [38].
where: the nanocrystals was ca. 11 nm (seen also at SEM Figure 5b) which is much smaller than the visible light wavelength. Thus, obtained glass-ceramics are transparent (Figure 1a,b). The size of the ZnGa2O4 nanocrystals presented in Figure 5 b was also checked according to the calculation by Scherrer's Equation (1) [36,37]. The obtained results for 15ZnO_650 °C_3h and 15ZnO_650 °C_5h samples have been presented in Tables 3 and 4. The crystalline size of crystals has been evaluated for diffraction peaks visible in Figure  5a. According to the calculated data, the crystalline size of nanocrystals is in the 3-18 nm range for both samples. These data are in line with the SEM results and give an overview of the crystal size distribution in the glass-ceramic samples. Moreover, the longer heating time of Ni/Er co-doped glass sample from 3h to 5h did not cause an increase in crystal size. The difference in the size of the crystallites for the analyzed samples is within the measurement bar (±2nm). This suggests that with an increase in heating time of the glass sample, the number of nanocrystallines grew [38].   Thermal treatment conditions have been investigated using the DSC analysis of the as-melted precursor GGK glass at a heating rate of 10 • C/min and shown in Figure 6. One exothermic, crystallization peak located around 703 • C was observed. The glass transition temperature T g and the onset of the crystallization temperature were about 560 • C and 675 • C, respectively. The thermal stability parameter ∆T defined as T x-T g was calculated to be 115 • C. It proofs the good thermal stability of glass, which is essential in the case of controllable crystallization of glasses and fabrication of the glass-ceramics material with embedded nanocrystals. Moreover, optical fiber technology ∆T above 100 • C describes glass as a suitable material for the optical fiber drawing [39].

Optical Properties
The absorbance spectra of 0.1NiO/Er2O3 co-doped glasses with 10, 15, and 20 mol% ZnO is shown in Figure 7a. The spectra consist of five bands that are assigned to the

Optical Properties
The absorbance spectra of 0.1NiO/Er 2 O 3 co-doped glasses with 10, 15, and 20 mol% ZnO is shown in Figure 7a. The spectra consist of five bands that are assigned to the transition from the ground state of Er 3+ to the 4 G 11/2 , 4 F 7/2 , 4 H 11/2 , 4 F 9/2 , 4 I 9/2 higher levels, and one band at 430 nm which is derived from the transition of 3 A 2 ( 3 F) → 3 T 1 ( 3 P) in five folded Ni 2+ ions. It should be noted that the presence of ZnO does not affect the background absorption. However, the shape of the absorption spectrum has changed after the heat treatment process (Figure 7b). It was also proofed by changing the color of the glass and glass-ceramics (Figure 1a,b). This effect is related to the variation of the Ni 2+ ions in sites: from tetrahedral in glasses to octahedral in glass-ceramics [40,41]. Therefore, the Ni 2+ ions should be embedded in the ZnGa 2 O 4 nanocrystals after the heat treatment. It was reported in the literature that blue, brown, and yellow-green glass colors are observed in the case of 4 Ni 2+ , 5 Ni 2+ , and 6 Ni 2+ coordination, respectively [42].

Optical Properties
The absorbance spectra of 0.1NiO/Er2O3 co-doped glasses with 10, 15, and 20 mol% ZnO is shown in Figure 7a. The spectra consist of five bands that are assigned to the transition from the ground state of Er 3+ to the 4 G11/2, 4 F7/2, 4 H11/2, 4 F9/2, 4 I9/2 higher levels, and one band at 430 nm which is derived from the transition of 3 A2( 3 F) → 3 T1 ( 3 P) in five folded Ni 2+ ions. It should be noted that the presence of ZnO does not affect the background absorption. However, the shape of the absorption spectrum has changed after the heat treatment process (Figure 7b). It was also proofed by changing the color of the glass and glass-ceramics (Figure 1a,b). This effect is related to the variation of the Ni 2+ ions in sites: from tetrahedral in glasses to octahedral in glass-ceramics [40,41]. Therefore, the Ni 2+ ions should be embedded in the ZnGa2O4 nanocrystals after the heat treatment. It was reported in the literature that blue, brown, and yellow-green glass colors are observed in the case of 4 Ni 2+ , 5 Ni 2+ , and 6 Ni 2+ coordination, respectively [42]. Besides, after heat treatment at 650 °C for 5 h, UV absorption is red shifted and background absorption is slightly higher in comparison to as-melted glass. A similar effect was reported in the literature [15].
In Figure 8, luminescence spectra of the GGK glass and glass-ceramics with 10, 15, and 20 ZnO (mol.%) doped with 0.1NiO (λexc = 808 nm) are presented. The broadband  Besides, after heat treatment at 650 • C for 5 h, UV absorption is red shifted and background absorption is slightly higher in comparison to as-melted glass. A similar effect was reported in the literature [15].
In Figure 8, luminescence spectra of the GGK glass and glass-ceramics with 10, 15, and 20 ZnO (mol.%) doped with 0.1NiO (λ exc = 808 nm) are presented. The broadband 1050-1650 nm emission centered at 1280 nm can be assigned to the 3 T 2 ( 3 F) → 3 A 2 ( 3 F) transition of octahedral Ni 2+ ions in the ZnGa 2 O 4 nanocrystals [43]. It is seen that the emission spectrum is composed of two broad sub-peaks located at 1280 nm and 1460 nm. It is known that ZnGa 2 O 4 is a two-site spinel compound [33]. It was also found that luminescence intensity increases with the ZnO content. It might be the result of the increase of Ni 2+ in octahedral sites with the ZnGa 2 O 4 nanocrystals [12]. These results correspond well to the further discussed emission decay curves.
To investigate ultra-broadband emission in the range of 1050-1650 nm GGK glass was co-doped with 0.1NiO/0.1Er 2 O 3 . The effect of the heat treatment of GGK glasses with 10, 15, 20 ZnO on the luminescence spectra under 808 nm laser excitation is presented in Figures 9 and 10. Moreover, the influence of the heat treatment time on the I 1550nm /I 1300nm intensity ratio has been presented in Figure 11. Emission spectra in all investigated glassceramic samples are the result of a superposition of the 1280 nm and 1550 nm luminescence bands. Ni 2+ : 3 T 2 ( 3 F) level is populated directly by the ground-state absorption of the 808 nm laser pump whereas erbium ions also directly absorb the same excitation wavelength. Next, within the Er 3+ : 4 I 9/2 → 4 I 11 / 2 → 4 I 13/2 the nonradiative relaxation occurs. At the same time, part of the energy is transferred from Ni 2+ : 3 T 2 ( 3 F) to Er 3+ : 4 I 13/2 level. It was confirmed by analysis of the luminescence decay curves (Figures 12-14). Finally, the 1280 nm luminescence band can be assigned to the 3 T 2 ( 3 F) → 3 A 2 ( 3 F) transition of octahedral Ni 2+ ions, and 1550 nm emission is a result of the Er 3+ : 4 I 13/2 → 4 I 15/2 radiative transition. Normalized luminescence spectra enabled analysis of the influence of the heat treatment time on the I 1550nm /I 1300nm intensity ratio. It is seen that the luminescence intensity of Ni 2+ ions increases with the longer heat treatment time. The curve of I 1550nm /I 1300nm vs. heat treatment time shows the smallest changes (0-2 h) in GGK glass with 20 ZnO (Figure 11). It might be the result of the faster nucleation of the ZnGa 2 O 4 nanocrystals with incorporated Ni 2+ ions. However, after 5 h, the I 1550nm /I 1300nm ratio has stabilized and was estimated to be 14.5 (10ZnO), 15.6 (15ZnO), 17.8 (20ZnO). Moreover, a decrease in FWHM (full width at half maximum) of 1550 nm luminescence band from 77 nm (as-melted glasses) to 44 nm (HT for 5 h glass-ceramics) was observed in all investigated 0.1NiO/0.1Er 2 O 3 -co-doped glasses. It suggests modifications in the site symmetry around erbium through the incorporation part of the Er 3+ ions into the crystalline phase [44]. If only Ni 2+ ions (not Er 3+ ) are incorporated into ZnGa 2 O 4 nanocrystals the distance between Ni 2+ and Er 3+ is elongated and energy transfer Ni 2+ and Er 3+ may be suppressed. This effect was observed in silicate glass ceramics [15]. Our research indicates that Ni 2+ to Er 3+ energy transfer occurs in the crystalline phase of germanate glass-ceramics. 1050-1650 nm emission centered at 1280 nm can be assigned to the 3 T2( 3 F) → 3 A2( 3 F) transition of octahedral Ni 2+ ions in the ZnGa2O4 nanocrystals [43]. It is seen that the emission spectrum is composed of two broad sub-peaks located at 1280 nm and 1460 nm. It is known that ZnGa2O4 is a two-site spinel compound [33]. It was also found that luminescence intensity increases with the ZnO content. It might be the result of the increase of Ni 2+ in octahedral sites with the ZnGa2O4 nanocrystals [12]. These results correspond well to the further discussed emission decay curves. To investigate ultra-broadband emission in the range of 1050-1650 nm GGK glass was co-doped with 0.1NiO/0.1Er2O3. The effect of the heat treatment of GGK glasses with 10, 15, 20 ZnO on the luminescence spectra under 808 nm laser excitation is presented in Figures 9 and 10. Moreover, the influence of the heat treatment time on the I1550nm/I1300nm intensity ratio has been presented in Figure 11. Emission spectra in all investigated glassceramic samples are the result of a superposition of the 1280 nm and 1550 nm luminescence bands. Ni 2+ : 3 T2( 3 F) level is populated directly by the ground-state absorption of the 808 nm laser pump whereas erbium ions also directly absorb the same excitation wavelength. Next, within the Er 3+ : 4 I9/2 → 4 I11/2 → 4 I13/2 the nonradiative relaxation occurs. At the same time, part of the energy is transferred from Ni 2+ : 3 T2( 3 F) to Er 3+ : 4 I13/2 level. It was confirmed by analysis of the luminescence decay curves (Figures 12-14). Finally, the 1280 nm luminescence band can be assigned to the 3 T2( 3 F) → 3 A2( 3 F) transition of octahedral Ni 2+ ions, and 1550 nm emission is a result of the Er 3+ : 4 I13/2 → 4 I15/2 radiative transition. Normalized luminescence spectra enabled analysis of the influence of the heat treatment time on the I1550nm/I1300nm intensity ratio. It is seen that the luminescence intensity of Ni 2+ ions increases with the longer heat treatment time. The curve of I1550nm/I1300nm vs. heat treatment time shows the smallest changes (0-2 h) in GGK glass with 20 ZnO (Figure 11). It might be the result of the faster nucleation of the ZnGa2O4 nanocrystals with incorporated Ni 2+ ions. However, after 5 h, the I1550nm/I1300nm ratio has stabilized and was estimated to be 14.5 (10ZnO), 15.6 (15ZnO), 17.8 (20ZnO). Moreover, a decrease in FWHM (full width at half maximum) of 1550 nm luminescence band from 77 nm (as-melted glasses) to 44 nm (HT for 5 h glass-ceramics) was observed in all investigated 0.1NiO/0.1Er2O3 -co-doped glasses. It suggests modifications in the site symmetry around erbium through the incorporation part of the Er 3+ ions into the crystalline phase [44]. If only Ni 2+ ions (not Er 3+ ) are incorporated into ZnGa2O4 nanocrystals the distance between Ni 2+ and Er 3+ is elongated and energy transfer Ni 2+ and Er 3+ may be suppressed. This effect was observed in silicate         Analysis of luminescence decay curves in glass-ceramics singly doped with Ni 2+ and co-doped with Ni 2+ /Er 3+ presented in Figure 12 enables to calculate the efficiency of Ni 2+ →Er 3+ energy transfer according to the equation: where: is the lifetime of Ni 2+ : 3 T2( 3 F) in the presence of Er 3+ , is the lifetime of Ni 2+ : 3 T2( 3 F) in singly Ni 2+ -doped glass-ceramics. The lifetime of Ni 2+ : 3 T2( 3 F) is characterized by double-exponential behavior. This effect suggests multiple side effects of Ni 2+ and nonradiative multipolar interactions among Ni 2+ [42]. The luminescence decay curves of the glass-ceramics singly doped with Ni 2+ and co-doped with Ni 2+ /Er 3+ were fitted by the sum of two exponential decay components from: where τ1 and τ2 were short-and long-decay components, respectively. Parameters A1 and A2 were fitting constants. According to Equation (3), the average lifetime <τ> was given by: According to Equation (4), the average lifetime of Ni 2+ : 3 T2( 3 F) in the fabricated GGK with 15ZnO glass-ceramics (HT@650 °C for 5 h) was calculated to be 433 μs (0.1NiO), and 349 μs (0.1NiO/0.1Er2O3). Thus, the efficiency of Ni 2+ → Er 3+ energy transfer (ET) was determined to be 19%. In the literature, in CaZrO3 nanocrystals co-doped with 0.2Ni/10Er (mol.%), the efficiency of the nickel to erbium energy transfer was estimated to be 86% [42]. In contrast, in silicate glass-ceramics with ZnGa2O4 nanocrystals (HT@ 850-950 °C) energy transfer between Ni 2+ and Er 3+ was suppressed [15]. It indicates that in the analyzed  nanocrystals lifetime of Ni 2+:3 T2( 3 F) was 300 μs. In general, the Ni 2+ lifetime has a positive correlation with the crystal field strength [12].  Analysis of the Er 3+ : 4 I13/2 luminescence decay (Figure 13 b) showed an increase in lifetime with increasing HT time (from 2.18 ms-as melted to 5.98 ms-HT 5h). It suggests that erbium ions are incorporated in the low-phonon energy ZnGaO4 nanocrystal phase. This effect is also consistent with the narrowing of the 1.55 μm luminescence band. Moreover, the Er 3+ : 4 I13/2 lifetime increases with increasing the ZnO content (Figure 14), which could be the cause for more erbium ions in the nanocrystal phase. The Er 3+ : 4 I13/2 lifetime was well fitted by a single-exponential curve which suggests one dominant structural position.

Conclusions
In summary, the effects of the ZnO content in transparent glass-ceramics containing ZnGa2O4 nanocrystals doped with Ni 2+ and co-doped with Ni 2+ /Er 3+ on structural and NIR luminescence properties were investigated. The controlled crystallization process enabled creation of ca. 11 nm size nanocrystals with incorporated Ni 2+ ions. According to the decomposed MIR spectra, it was found that the germanate-gallate networks of glass and glass-ceramics with 15    nanocrystals lifetime of Ni 2+:3 T2( 3 F) was 300 μs. In general, the Ni 2+ lifetime has a positive correlation with the crystal field strength [12].  Analysis of the Er 3+ : 4 I13/2 luminescence decay (Figure 13 b) showed an increase in lifetime with increasing HT time (from 2.18 ms-as melted to 5.98 ms-HT 5h). It suggests that erbium ions are incorporated in the low-phonon energy ZnGaO4 nanocrystal phase. This effect is also consistent with the narrowing of the 1.55 μm luminescence band. Moreover, the Er 3+ : 4 I13/2 lifetime increases with increasing the ZnO content (Figure 14), which could be the cause for more erbium ions in the nanocrystal phase. The Er 3+ : 4 I13/2 lifetime was well fitted by a single-exponential curve which suggests one dominant structural position.

Conclusions
In summary, the effects of the ZnO content in transparent glass-ceramics containing ZnGa2O4 nanocrystals doped with Ni 2+ and co-doped with Ni 2+ /Er 3+ on structural and NIR luminescence properties were investigated. The controlled crystallization process enabled creation of ca. 11 nm size nanocrystals with incorporated Ni 2+ ions. According to the decomposed MIR spectra, it was found that the germanate-gallate networks of glass and glass-ceramics with 15  Analysis of luminescence decay curves in glass-ceramics singly doped with Ni 2+ and co-doped with Ni 2+ /Er 3+ presented in Figure 12 enables to calculate the efficiency of Ni 2+ →Er 3+ energy transfer according to the equation: where: τ Ni is the lifetime of Ni 2+ : 3 T 2 ( 3 F) in the presence of Er 3+ , τ Ni is the lifetime of Ni 2+ : 3 T 2 ( 3 F) in singly Ni 2+ -doped glass-ceramics. The lifetime of Ni 2+ : 3 T 2 ( 3 F) is characterized by double-exponential behavior. This effect suggests multiple side effects of Ni 2+ and non-radiative multipolar interactions among Ni 2+ [42]. The luminescence decay curves of the glass-ceramics singly doped with Ni 2+ and co-doped with Ni 2+ /Er 3+ were fitted by the sum of two exponential decay components from: where τ 1 and τ 2 were short-and long-decay components, respectively. Parameters A 1 and A 2 were fitting constants. According to Equation (3), the average lifetime <τ> was given by: According to Equation (4), the average lifetime of Ni 2+ : 3 T 2 ( 3 F) in the fabricated GGK with 15ZnO glass-ceramics (HT@650 • C for 5 h) was calculated to be 433 µs (0.1NiO), and 349 µs (0.1NiO/0.1Er 2 O 3 ). Thus, the efficiency of Ni 2+ → Er 3+ energy transfer (ET) was determined to be 19%. In the literature, in CaZrO 3 nanocrystals co-doped with 0.2Ni/10Er (mol.%), the efficiency of the nickel to erbium energy transfer was estimated to be 86% [42]. In contrast, in silicate glass-ceramics with ZnGa 2 O 4 nanocrystals (HT@ 850-950 • C) energy transfer between Ni 2+ and Er 3+ was suppressed [15]. It indicates that in the analyzed germanate glass-ceramics Ni 2+ to Er 3+ energy transfer occurs in the crystalline phase. Figure 13 presents the effect of the HT time GGK with 15 ZnO glass on the lifetime of Ni 2+ : 3 T 2 ( 3 F) and Er 3+ : 4 I 13/2 metastable levels. Due to the incorporation of the nickel ions into ZnGa 2 O 4 nanocrystals, the lifetime of the Ni 2+ level increases with increasing the HT time and was calculated to be 62 µs (as-melted glass) and 433 µs (glass ceramics, HT-5 h). In silicate glass-ceramics with ZnGa 2 O 4 and CaZrO 3 nanocrystals, this value was 730 µs [12] and 600 µs [45], respectively. On the other hand, in silicate glass-ceramics with LiGa 5 O 8 nanocrystals lifetime of Ni 2+:3 T 2 ( 3 F) was 300 µs. In general, the Ni 2+ lifetime has a positive correlation with the crystal field strength [12].
Analysis of the Er 3+ : 4 I 13/2 luminescence decay ( Figure 13b) showed an increase in lifetime with increasing HT time (from 2.18 ms-as melted to 5.98 ms-HT 5 h). It suggests that erbium ions are incorporated in the low-phonon energy ZnGaO 4 nanocrystal phase. This effect is also consistent with the narrowing of the 1.55 µm luminescence band. Moreover, the Er 3+ : 4 I 13/2 lifetime increases with increasing the ZnO content ( Figure  14), which could be the cause for more erbium ions in the nanocrystal phase. The Er 3+ : 4 I 13/2 lifetime was well fitted by a single-exponential curve which suggests one dominant structural position.

Conclusions
In summary, the effects of the ZnO content in transparent glass-ceramics containing ZnGa 2 O 4 nanocrystals doped with Ni 2+ and co-doped with Ni 2+ /Er 3+ on structural and NIR luminescence properties were investigated. The controlled crystallization process enabled creation of ca. 11 nm size nanocrystals with incorporated Ni 2+ ions. According to the decomposed MIR spectra, it was found that the germanate-gallate networks of glass and glass-ceramics with 15  The structural changes of heat-treated precursor glasses showed local environment changes connected with the incorporation of parts of the Ni 2+ and Er 3+ ions into the crystalline phase of ZnGa 2 O 4 . This was in agreement with optical analysis. In particular, the broadband emission in the range of 1000-1650 nm (λ exc -808 nm) as a result of a superposition of the Ni 2+ : 3 T 2 ( 3 F) → 3 A 2 ( 3 F) octahedral Ni 2+ ions and Er 3+ : 4 I 13/2 → 4 I 15/2 radiative transitions and energy transfer from Ni 2+ to Er 3+ with an efficiency of ca. 19%. Moreover, increasing the Er 3+ : 4 I 13/2 lifetime and narrowing the 1.55 µm emission band suggest that Er 3+ ions are partially incorporated into ZnGa 2 O 4 nanocrystals. It was also observed that ZnO content has a positive impact on the luminescent properties of GGK Ni 2+ /Er 3+ -co-doped glass-ceramics. The highest Ni 2+ luminescence intensity and the longest lifetime of Er 3+ : 4 I 13/2 level have been obtained in GGK glass-ceramics with 20ZnO (mol.%). It is known that a longer lifetime of the metastable level is beneficial for obtaining high optical gain. It should be also emphasized that analyzed glass is thermally stable. Thus, investigated glass-ceramics are promising candidates for applications in broadband fiber emission sources of radiation or optical fiber amplifiers. Further research should be focused on the flattening of the emission in the whole 1000-1650 nm spectral range.