- Email: [email protected]
Time- and temperature-resolved luminescence spectroscopy of LiAl4O6F:Mn red phosphors
Journal of Luminescence 216 (2019) 116754
Contents lists available at ScienceDirect
Journal of Luminescence journal homepage: www.elsevier.com/locate/jlumin
Time- and temperature-resolved luminescence spectroscopy of LiAl4O6F:Mn red phosphors
T
Nicholas Khaidukova, Maria Brekhovskikha, Guido Tocib, Barbara Patrizib, Matteo Vanninib, Angela Pirric, Vladimir Makhovd,∗ a
N. S. Kurnakov Institute of General and Inorganic Chemistry, 31 Leninskiy Prospekt, Moscow, 119991, Russia Istituto Nazionale di Ottica, Consiglio Nazionale delle Ricerche, INO-CNR, via Madonna del Piano 10, 50019, Sesto Fiorentino, Florence, Italy c Istituto di Fisica Applicata “Carrara”, Consiglio Nazionale delle Ricerche, IFAC-CNR, via Madonna del Piano 10, 50019, Sesto Fiorentino, Florence, Italy d P. N. Lebedev Physical Institute, 53 Leninskiy Prospekt, Moscow, 119991, Russia b
A R T I C LE I N FO
A B S T R A C T
Keywords: Photoluminescence Time-resolved spectroscopy Red-emitting Mn4+ phosphors Spinel Zero-phonon line Phosphor-converted LED
Ceramic samples of LiAl4O6F phosphors doped with 1.0 mol. % manganese ions have been synthesized by hightemperature solid-state reaction technique and investigated as possible red-emitting light converters. It has been found that the emission spectrum of LiAl4O6F:Mn phosphors consists of a narrow band peaked at 661 nm and a broader band peaked in the region of 675 ÷ 720 nm depending on synthesis conditions. The intensity of the latter band increases and the peak of this band shifts to longer wavelengths with increasing the fraction of MgF2 used as a flux material. Besides that, these two emission bands have strongly different decay times, namely τ1 ~ 240 μs and τ2 ~ several ms, respectively. The decay kinetics and time-resolved spectra of these two emission components have been studied in the temperature range of 10 ÷ 290 K under pulsed laser excitation at 262 nm. The observed luminescence properties have been analyzed within a model which considers the presence of manganese ions having different valence states (Mn4+ and Mn2+) in the specific spinel-type crystal structure of the LiAl4O6F host matrix, which provides many possibilities for different local environment near the doping manganese ions.
1. Introduction As it is now commonly accepted, the phosphor-converted white LEDs (pc-WLEDs) are superior to all other light sources for general lighting because they have low energy consumption, high efficiency, long service life, compact dimensions, are environmentally safe (do not contain mercury), and are stable to external conditions [1]. Most commercially available pc-WLEDs are based on the simplest and most cost-efficient scheme involving a blue emitting semiconductor chip and a yellow emitting phosphor, such as Ce3+ doped garnet Y3Al5O12:Ce3+ (YAG:Ce). However, such light sources emit bluish (cool) white light, which is not suitable for indoor lighting. The problem of obtaining warm white light can be solved by enriching a red component in the spectrum of pc-WLED. In commercial warm pc-WLEDs some Eu2+ doped nitrides as red emitting phosphors are typically used [2–5]. However, the synthesis of nitride phosphors requires high temperature and pressure, which makes them difficult to produce industrially. Also, such Eu2+ doped phosphors show rather broad emission band with a tail into the deep red and near IR spectral range, where human eyes are
∗
not sensitive (the eye sensitivity strongly drops above 650 nm), i.e. a considerable part of light from such phosphors is wasted thus decreasing luminous efficacy of the light source. Therefore, the development of new narrow-band red emitting phosphors is of tremendous importance for the further optimization of solid-state light sources for indoor illumination applications. It is generally accepted that Mn4+ doped phosphors emitting a narrow-band luminescence in the red spectral range are mostly attractive candidates for application in warm pc-WLEDs [6–12]. An important advantage of this type of phosphors is the absence of rare earth elements in their compositions, such as Eu2+, the prices of which are increased permanently, with the only exception of Ce, which is quite abundant worldwide and relatively cheap. By now a rather large number of rare earth free Mn4+ doped red phosphors have been developed, which are based mainly on fluoride hosts, because Mn4+ narrow-band emission in fluorides is located close to the optimal wavelength for red phosphors near 630 nm. However, one major drawback of such luminescence materials is that toxic fluoric acid is needed for their synthesis. Besides that, such fluoride phosphors are typically not well resistant to high
Corresponding author. E-mail address: [email protected] (V. Makhov).
https://doi.org/10.1016/j.jlumin.2019.116754 Received 27 June 2019; Received in revised form 12 September 2019; Accepted 17 September 2019 Available online 18 September 2019 0022-2313/ © 2019 Elsevier B.V. All rights reserved.
Journal of Luminescence 216 (2019) 116754
N. Khaidukov, et al.
temperature and humidity as well as degrade at high excitation densities. One could expect that oxyfluoride compounds as hosts can provide much more stability of phosphors to external conditions but still have acceptable properties of red luminescence from doping Mn4+ ions. Indeed, diverse Mn4+ doped oxyfluoride phosphors have been extensively studied in recent years (the review of luminescence properties of Mn4+ doped oxyfluorides can be found in Ref. [12]), which shows that spectral properties of them are similar to those of Mn4+ doped fluorides. In particular an oxyfluoride LiAl4O6F:Mn4+ phosphor which possesses relatively broad emission band peaked at ~662 nm in contrast to most other Mn4+ doped oxyfluorides showing red line emission has been proposed just recently [13,14]. In our opinion, luminescence properties of this phosphor are still not completely understood. In the present work the specific features of luminescence from red phosphors of such a type have been studied by using technique of time-resolved low-temperature spectroscopy. LiAl4O6F compound has the spinel-like structure but with some specific crystal chemical properties [15,16]. On the other hand, in Ref. [17] it is claimed that the Mn4+ doped spinel phosphors MgAl2O4 and MgGa2O4 prepared with the excess of Mg show higher intensity of Mn4+ luminescence. In this context, a study of the effects of doping with Mg ions on luminescent properties of Mn4+ in LiAl4O6F has been performed too. 2. Experimental A series of LiAl4O6F ceramic phosphors doped with manganese ions was synthesized by high-temperature solid-state reaction technique. LiF - 2Al2O3 precursors mixed with KMnO4 (1.0 mol %) were uniaxially pressed into pellets with a diameter of 10 mm and a thickness of 3–5 mm and annealed at temperatures of 500–700 °C in normal air atmosphere by taking into account that K+ ions are removed from the material as a result of the subsequent multi-stage thermal treatment. Then, the samples were thoroughly ground, re-pelletized and heat treated at various temperatures 1000–1100 °C in argon atmosphere with the addition of different concentrations of MgF2 flux. The molar amount of the flux was varied in the range from 0 to 20%. Eventually, pellets of synthesized ceramic phosphors were polished for later characterization. The structural-phase composition of the obtained ceramic samples was studied on an X-ray powder diffractometer D8 Advance (Bruker) in a monochromatic CuKα radiation. Identification of the synthesized compounds was performed in the EVA (Bruker) software package using the ICDD PDF-2 database. The unit cell parameters were refined by program AXES using the whole powder pattern fitting toolbox. Energy dispersive X-ray (EDX) analysis of the samples was performed by an Energy-Dispersive System, Octane Elect Super Team Basic (EDAX, AMETEK, Mahwah, NJ, USA). Photoluminescence (PL) and PL excitation (PLE) spectra at room temperature were measured using a spectrofluorimeter PerkinElmer LS55. The time-resolved luminescence studies were performed in the temperature range of 10–290 K using the set-up based on a JOBINYVON SPEX 320 spectrometer and an optical cryostat cooled by a closed cycle refrigerator (CTI-CRYOGENICS). The spectrometer was equipped with the 300 l/mm grating which provided spectral resolution of 0.6 nm with the typical entrance slit width 40 μm. Decay kinetics and time-resolved spectra were measured under pulsed laser excitation at 262 nm (4th harmonics of a Nd:YLF laser). In front of the entrance slit of the spectrometer the cut-off filter Schott GG435 was installed in order to prevent a possible penetration of scattered laser light into the spectrometer. Decay curves were recorded using a photomultiplier tube EMI 9816QB and a digital storage oscilloscope Tektronix TDS 680B, with the spectral band-pass (determined by the exit slit width) varied between 2 and 4 nm depending on the signal intensity. The time resolved spectra were measured with a gated intensified multichannel detector EG&G 1420 with an OMA2000 detector controller with the spectral width of one pixel 0.51 nm. The time gate duration and the delay of the gate with respect to the laser excitation pulse were
Fig. 1. XRD patterns of LiAl4O6F:Mn phosphors synthesized without using MgF2 flux (upper panel) and with using 20% MgF2 flux (lower panel); peaks are labeled by Miller indices of corresponding lattice planes.
controlled using a Stanford DG535 delay generator.
3. Results Synthesis experiments have shown that annealing of the precursor mixtures at temperatures above 1000 °C leads to the formation of monophase LiAl4O6F, without traces of any other phases. The XRD patterns of synthesized phosphors are presented in Fig. 1. LiAl4O6F belongs to cubic crystal system and has the spinel-type structure [15] in which Li+ and a part of Al3+ ions are 6-fold coordinated by anions and the rest of Al3+ is 4-fold coordinated although some disordering of Li+ and Al3+ ions entering octahedral and tetrahedral sites is probably possible. The lattice parameters of the cubic unit cell obtained from XRD analysis are 7.9446 and 7.9705 Å for phosphors synthesized without using MgF2 flux and with using 20% MgF2 flux, respectively. The increase of the lattice parameter after adding MgF2 can reflect the fact that Mg2+ ions substitute for the ions of smaller ionic radius, i.e. Al3+, which is rather unexpected. On the other hand, LiAl4O6F chemical composition does not match the spinel-type stoichiometry, in contrast to similar spinel-type material LiAl5O8 (Li0.5Al2.5O4) [14], and according to Ref. [16] the crystal structure of this compound should contain one vacant cation site and one vacant anion site for every chemical formula unit LiAl4O6F (Li0.5Al2O3F0.5). Due to this feature of the crystal structure, Mg2+ ions (as well as additional F‾ ions) can enter the available vacant sites, which will result in increasing the lattice parameter. The compound synthesized with using MgF2 flux can be described by chemical formula LiMgxAl4O6F1+2x up to (Li,Mg,Al)6(O,F)8 having the spinel-like crystal structure. Moreover, as can be seen in Fig. 1 from the comparison of the full width at half maximum 2
Journal of Luminescence 216 (2019) 116754
N. Khaidukov, et al.
Fig. 3. PLE and PL spectra of LiAl4O6F:Mn phosphors synthesized without using MgF2 flux (upper panel) and with using 20% MgF2 flux (lower panel).
Fig. 2. EDX spectra of LiAl4O6F:Mn phosphors synthesized without using MgF2 flux (upper panel) and with using 20% MgF2 flux (lower panel).
intensities for the XRD patterns, the degree of crystallinity for the sample synthesized with using MgF2 flux is higher than that for the sample prepared without using MgF2 flux. Experimental EDX spectra obtained from the surface of the samples are shown in Fig. 2. However, the content of Li cannot be determined from our EDX measurements, being too light to be detected. The elemental ratio Al:O well enough corresponds to chemical compositions of the compounds, whereas there is a considerable deficiency of F, which can be lost during heat treatment. The content of Mn is also smaller than that introduced in the starting mixture. As can be seen in Fig. 2 in the EDX spectrum of the sample synthesized using 20 mol.% MgF2 flux a considerable amount of Mg (~2.6 at. %) is detected thus confirming that Mg ions enter the crystal lattice of the phosphor but without changing the spinel structural type as confirmed by the XRD data. On the other hand, it should be emphasized that K+ is not detected at all. PL and PLE spectra of LiAl4O6F:Mn phosphors measured at room temperature are shown in Fig. 3. The PL spectrum measured under excitation in the visible spectral range is dominated by the narrow band peaked at ~661 nm for the sample prepared without using flux. With increasing concentration of MgF2 flux this peak shifts slightly to longer wavelengths reaching ~663 nm for the sample synthesized with 20% flux, the PL spectrum becomes broader, especially towards longer wavelengths, and the intensity of the broad part of spectrum increases. For the sample obtained without using flux the shape of the band depends weakly on the excitation wavelength whereas for the sample prepared using MgF2 flux the PL spectrum strongly changes under excitation in the deep UV region (262 nm), namely it becomes dominated by a broad band peaked at ~702 nm. The PLE spectra of the narrow-band emission (661–663 nm) contain three main bands peaked at ~458, ~369 and ~286 ÷ 300 nm, although the spectrum can be well enough decomposed (in the linear
Fig. 4. Decomposition of PL (excitation at 300 nm) and PLE (emission at 663 nm) spectra of the LiAl4O6F:Mn phosphor, synthesized using 20% MgF2 flux, into the sum of Gaussian-like sub-bands. The spectra were first converted into linear energy scale E = hc/λ. The PL spectrum was additionally converted according to I(E) = I(λ)*λ2/hc. After fitting with multi-Gaussian functions the spectra were converted back to the wavelength scale.
energy scale) into at least 5 Gaussian-like sub-bands (see Fig. 4 as an example). In PLE spectra of a broad-band emission (740 nm) for samples prepared using MgF2 flux the peaks at ~458 and ~369 nm become less pronounced, the additional long-wavelength band peaked at ~560 nm appears and the UV band shifts to shorter wavelengths. In addition to the emission in the red spectral range some weak broadband emission is also observed in the green region with a peak maximum at ~500 nm (not shown). The intensity of this green emission increases with amount of used MgF2 flux. One can suggest that the red emission band observed from LiAl4O6F:Mn phosphors consists of at least two sub-bands: the narrowband and the broad-band one, although the decomposition of the emission band into at least four sub-bands gives good enough fitting result (Fig. 4). The measurements of spectrally-resolved decay kinetics have shown that these two emission sub-bands have strongly different decay kinetics: the narrow band (~661 nm) has the characteristic decay time in the range of hundreds of microseconds whereas the broad band 3
Journal of Luminescence 216 (2019) 116754
N. Khaidukov, et al.
Fig. 5. Luminescence decay curves measured from the LiAl4O6F:Mn phosphor (synthesized without using flux) monitoring 661 and 700 nm emissions at 290 K.
Fig. 7. Decay curves of fast (λem = 661 nm, τ ~ 290 μs) and slow (λem = 720 nm, τ ~3.5 ms) components of luminescence recorded from the LiAl4O6F:Mn phosphor, synthesized using 20% MgF2 flux, at temperature 9.6 K. The positions and widths of time gates are shown which were chosen for timeresolved emission spectra measurements.
(675–720 nm) has typical decay time of several milliseconds (Fig. 5). Decay kinetics curves recorded at any wavelength within the width of emission spectrum cannot be well fitted by single-exponential decay but two-exponential fitting gives good enough simulation of these decay curves. This is because of spectral overlapping of different emission components which prevents selection of the pure “fast” or the pure “slow” component. However, decay kinetics of the 661 nm emission is almost single-exponential with a decay time τ ~240 μs with very small contribution of the slower component. The decay curve recorded at 700 nm has the dominant decay component with τ ~5.8 ms but has a contribution of the “fast” decay component just with τ ~240 μs. Decay kinetics of 661 nm emission measured for samples synthesized with the use of flux is essentially the same one as that measured for the sample synthesized without flux whereas decay curves recorded at 720 nm for samples synthesized with flux show slightly shorter decay times than that obtained for 700 nm emission of the sample synthesized without flux. For the sample with maximal (20%) amount of flux the decay time of the “slow” emission is τ ~ 3.0 ms at 290 K. The time-integrated and time-resolved spectra measured at temperature 290 K from the LiAl4O6F:Mn phosphor (synthesized without using flux) are shown in Fig. 6. The optimal time gate durations for
recording time-resolved spectra corresponding to the fast and the slow emission component have been chosen as follows: 100 μs without any delay with respect to excitation pulse and 3 ms with 3 ms delay with respect to the laser pulse, respectively. The positions and widths of time gates are shown in Fig. 7 where decay curves of the “fast” and the “slow” emission component measured at 9.6 K from the LiAl4O6F:Mn phosphor synthesized with 20% MgF2 flux are also presented. As is indicated in Fig. 7, some measurements of time-resolved spectra for the slow emission component have been performed also with 6 ms delay with respect to the laser pulse. Under cooling the luminescence intensity increases but time-resolved spectral shapes show only small changes, although in the PL spectra the features of the fast and the slow component become better resolved. In particular, some fine structure appears in the spectrum of the slow emission component, namely a narrow-line becomes well pronounced in the spectrum (Fig. 8). The shape of the slow-component spectrum practically does not change when the delay is increased to 6 ms (Fig. 9), i.e. this narrow line and a broad sub-band have the same decay rate and accordingly belong to radiative transitions from the same electronic state. This observation allows one to ascribe this narrow line to zero-phonon line (ZPL) of the respective electronic transition. The energy of this ZPL is slightly higher than the energy of the intense line of the fast component (the difference is ~1.5 nm) in the spectrum. On the other hand, this ZPL is not seen in the spectra of LiAl4O6F:Mn phosphors synthesized using high concentration of MgF2 flux. Decay kinetics of the narrow-line emission at 661 nm practically does not change with temperature but small increase of lifetime is observed under cooling from τ ~240 μs at 290 K to τ ~290 μs at 10 K. Similar weak dependence on temperature is observed for decay kinetics of the “slow” broad-band emission. 4. Discussion By taking into account all the experimental results described above, as a first approximation, one can suggest that two kinds of manganese emission centers exist in LiAl4O6F:Mn phosphors, which are responsible for “fast” narrow-band and “slow” broad-band red luminescence. By analogy with the results presented in Refs. [13,14] the narrow-band emission peaked at 661 nm can be attributed to luminescence of Mn4+ ions. Because of very similar ionic radii of Mn4+ (0.53 Å) and Al3+
Fig. 6. Normalized time-integrated and time-resolved emission spectra measured from the LiAl4O6F:Mn phosphor (synthesized without using flux) at 290 K. 4
Journal of Luminescence 216 (2019) 116754
N. Khaidukov, et al.
Fig. 8. Time-resolved emission spectra of the “fast” and “slow” components of luminescence from the LiAl4O6F:Mn phosphor measured in the temperature range of 10–290 K.
transitions and one more broad band in the UV region typically ascribed to the O2− - Mn4+ charge-transfer transition. Using obtained spectral data on Mn4+ luminescence and the standard formulas [20], CF strength parameter Dq as well as Racah parameters B and C can be calculated for the Mn4+ ion in the studied phosphor. For this calculation the peak energies of emission band as well as of 4T2 and 4T1 bands in the excitation spectra were used. The obtained values are: Dq = 2185 cm−1, B = 475 cm−1, C = 3791 cm−1, Dq/B = 4.6, which are similar to those obtained in Ref. [13] and correspond to rather large value of the CF strength. On the other hand, this narrow-band red luminescence observed in LiAl4O6F:Mn phosphors has some specific properties which are different compared to typical properties of most other Mn4+ doped phosphors. First, its decay time is of the order of hundreds of microseconds, even at low temperatures, which looks too short compared to typical decay times of several milliseconds observed in most Mn4+ doped phosphors for this parity- and spin-forbidden transition 2E → 4A2 of Mn4+. Also, the shape of the emission spectrum, namely the ratios of intensities between the central intense band and weaker sub-bands, only slightly change with temperature, which is not typical for Mn4+ emission spectrum where at low temperatures the spectrum is dominated by Stokes vibronic side-bands having longer wavelengths with respect to very weak ZPL and at higher temperatures the shorter-wavelength antiStokes vibronic side-bands appear in the spectrum. These properties of narrow-band luminescence of LiAl4O6F:Mn phosphors can be explained, in principle, if one suggests that in this phosphor the respective radiative transition of Mn4+ becomes partially dipole-allowed, which results in the increase of radiative transition probability, i.e. in the decrease of decay time, as well as to the increase of intensity of ZPL corresponding to pure electronic transition. Accordingly, the central intense line in the emission spectrum of LiAl4O6F:Mn can be ascribed to ZPL corresponding to pure electronic 2E → 4A2 transition of Mn4+. The weaker narrow shorter-wavelength band (at ~640 nm) cannot be ascribed to vibronic side-band of this ZPL because anti-Stokes vibronics with such high intensity (see Fig. 8) cannot be observed at 10 K. Thus, the weaker narrow shorter-wavelength side-band should be ascribed to ZPL related to luminescence of Mn4+ ions entering octahedral sites with different local environment, which can be really expected for such a specific crystal structure of the LiAl4O6F host. Although, according to the Tanabe-Sugano diagram for d3 electron configuration (Mn4+), the energy
Fig. 9. Time-resolved emission spectra of the “fast” and “slow” components of luminescence from the LiAl4O6F:Mn phosphor measured at 10 K. The spectrum of the “slow” component was measured with two delays of the time gate (3 and 6 ms) with respect to the laser pulse.
(0.535 Å) in octahedral coordination [18] and taking into account the commonly accepted statement that Mn4+ ions can be stabilized in the lattice only in octahedral coordination it is natural to suggest that Mn4+ ions occupy predominantly Al3+ octahedral sites of the LiAl4O6F lattice. The charge compensation can be reached by the occupation of neighbor anion vacancy by F‾ ion or by substitution of neighbor F‾ by O2−. According to the energy level scheme of Mn4+ given by the wellknown Tanabe-Sugano diagram for d3 electron configuration in an octahedral crystal field (CF) [19] the Mn4+ doped phosphors show broad absorption bands in the blue and near UV spectral range which are due to relatively strong spin-allowed transitions of Mn4+: 4A2 → 4T2 and 4 A2 → 4T1, respectively, while a narrow-band emission observed from Mn4+ doped phosphors originates from the Mn4+ spin-forbidden transition 2E → 4A2. Indeed, the PLE spectrum of LiAl4O6F:Mn monitoring 661 nm emission shows two characteristic broad bands at ~458 and ~369 nm well corresponding to Mn4+ 4A2 → 4T2 and 4A2 → 4T1 5
Journal of Luminescence 216 (2019) 116754
N. Khaidukov, et al.
the emitting 2E state does not depend on the CF strength, i.e. “goes parallel to the x-axis” in the Tanabe-Sugano diagram for d3 electron configuration. The qualitative analysis of the Tanabe-Sugano diagrams for d7 (Mn4+ in tetrahedron), d4 and d6 (Mn3+ in octahedron and tetrahedron, respectively), and d5 (Mn2+) electron configurations in octahedral CF, shows that practically all these options can be excluded from consideration. The only variant, which can be discussed, is Mn3+ in octahedral CF, by taking into account also that the substitution of Al3+ by Mn3+ does not require any charge compensation. However, in this case the narrow-band emission is typically observed in the IR range [27] and excitation spectrum will be completely different from that obtained in the experiment. Anyway, the possible influence of Mn3+ ions on luminescence properties of LiAl4O6F:Mn phosphors cannot be completely excluded but there is no experimental data for real analysis of this possibility. Some CT transitions, e.g. between manganese ions of different valence states, also cannot be responsible for this narrow-band luminescence because such transitions have typically much faster decay rate, and give a broad-band emission because they are assisted by large lattice relaxation due to the strong redistribution of the charge density. The nature of the “slow” broad-band red emission component in the PL spectrum of the LiAl4O6F:Mn phosphor can be different. The simplest explanation for the existence of such type of luminescence observed from the samples synthesized without MgF2 flux is the presence of some impurities. In particular, the spectral shape of this luminescence is very similar to that of Fe3+ ions (iso-electronic analog of Mn2+ ions) doped into disordered type of the LiAl4O6F matrix [14]. Also, the spectrum and decay time of this luminescence are similar to those observed from Fe3+ ions doped into so-called ordered phase of spinel-type compound LiAl5O8 [28]. Moreover, at low temperature the Fe3+ emission spectrum measured from LiAl5O8:Fe3+ shows well-resolved ZPL as is the case for the low-temperature spectrum of the “slow” broad-band emission component of the LiAl4O6F:Mn phosphor synthesized without MgF2 flux. This is a strong argument in favor of the impurity (Fe3+) nature of “slow” broad-band luminescence observed from the LiAl4O6F:Mn phosphor synthesized without MgF2 flux. On the other hand, the PL spectra of the “slow” broad-band red emission observed from the phosphors synthesized with the use of MgF2 flux are shifted to longer wavelengths and ZPL is practically not seen in the low-temperature spectra. One can suggest that this luminescence is due to the presence of some other optical center in such phosphors. The results of XRD and EDX analysis show that during synthesis of LiAl4O6F:Mn4+ using MgF2 flux, the Mg2+ and additional F‾ ions enter the crystal lattice of synthesized compound, most probably, by the occupation of the vacant sites in the spinel crystal lattice of LiAl4O6F. In fact this creates some more complex multi-component host compound of the type LiMgxAl4O6F1+2x as well as provide the sites for entering manganese ions of different valence states, in particular Mn2+, which can substitute for Mg2+ without charge compensation, and ionic radii of Mn2+ and Mg2+ do not differ so strongly as those of Mn2+ and Al3+. Accordingly, the “slow” (several ms) broad-band deep red emission observed from the LiAl4O6F:Mn phosphors synthesized with the use of MgF2 flux can be ascribed to luminescence of Mn2+ ions entering the octahedral sites. Since all absorption transitions of Mn2+ are spin-forbidden, in PLE spectra the respective bands are hidden behind the stronger spin-allowed and CT transitions of Mn4+ although in the longwavelength region of PLE spectra some additional band peaked at ~560 nm can be recognized as possibly due to the Mn2+ 6A1 → 4T1 transition. Besides that, under excitation in the deep UV region the PL spectrum consists of almost only broad deep-red emission band, i.e. the respective excitation band can be ascribed to the O2− - Mn2+ CT band. Weak green emission, whose intensity also increases with the amount of used MgF2 flux, can be ascribed to luminescence of Mn2+ ions entering the tetrahedral sites. The proposed model of observed luminescence properties fits well the decomposition structure of PL spectrum shown in Fig. 4. Two narrow bands peaked at 663 nm (higher intensity) and 646 nm (lower
of the emitting 2E state does not depend on the CF strength, still the Mn4+ emission spectrum can be remarkably shifted by changes of local environment around the Mn4+ ion because of the nephelauxetic effect (see, e.g. Ref. [21]). One could also mention that the spectral width of this ZPL is rather large (it is not determined by spectral resolution of the set-up) compared to typical ZPL widths. However, in the crystal structure of the LiAl4O6F host matrix with large concentrations of vacancies a considerable structural disorder around Mn4+ ions can be expected because a variety of combinations of ions/vacancies neighbor to Mn4+ ions is possible resulting in fact in the presence of the multiplicity of Mn4+ optical centers having different local environments, which gives inhomogeneous broadening of ZPL. The unusually intense ZPL in the emission spectra of Mn4+ ions has been already observed in some other oxyfluoride hosts, e.g. in Na2WO2F4:Mn4+ [22], but all other features such as Stokes and antiStokes side-bands with typical temperature dependence have been also detected in those spectra. The reason of this effect is the deviation of the Mn4+ site environment from centrosymmetric nature because of large distortions of Mn4+ octahedral coordination observed in oxyfluorides. If the Mn4+ site loses the inversion symmetry, the parity selection rule for the Mn4+ 2E → 4A2 transition will be relaxed by the admixture of the high-lying odd-parity states, which force this transition to become partially electric-dipole allowed thus increasing the decay rate of pure electronic transition and the intensity of ZPL. In oxyfluoride hosts the non-centrosymmetric distortion of octahedral coordination for Mn4+ ions arises because of the difference in charge of O and F ligand anions forming octahedrons around the Mn4+ sites. In the LiAl4O6F host the additional strong distortion of octahedral environment appears because of specific spinel-like crystal structure of this compound which suggests the presence of one vacant cation site and one vacant anion site for every chemical formula unit LiAl4O6F. This feature of the crystal structure can be the reason of ultra-intense ZPL in emission spectrum of Mn4+ ions doped into the LiAl4O6F crystal host. To the best of our knowledge and according to data of review papers [9,12] there is only one example of Mn4+ doped phosphor, namely αAl2O3:Mn4+, which shows an emission spectrum dominated by ZPLs of Mn4+ luminescence (so-called R1 and R2 lines observed in this phosphor at 677 and 673 nm, respectively) without a remarkable contribution of phonon side-bands even at 300 K [23–26]. The reason of such a specific behavior of this phosphor is not explained. However, one could mention that α-Al2O3 has a hexagonal crystal lattice in which the Al3+ cations occupy the non-centrosymmetric trigonally distorted octahedral sites (point group C3). Moreover, Al3+ ions occupy only 2/3 of such sites, i.e. in the α-Al2O3 lattice 1/3 of Al-sites are empty. These properties can be considered to combine α-Al2O3 and LiAl4O6F into one matrix type for Mn4+ doped phosphors. In fact, one could expect that there exist some other Mn4+ phosphors, which demonstrate Mn4+ emission spectra with dominating ZPL, and accordingly the classification of Mn4+ phosphors proposed in review papers [9,12] can be probably corrected. It should be mentioned also that high intensity of ZPL shifts the barycenter of Mn4+ emission spectrum towards shorter wavelengths compared to typical Mn4+ emission spectra, in which the Stokes vibronic side bands dominate, i.e. the increase of ZPL intensity in the emission spectrum of Mn4+ entering strongly distorted octahedral sites can be considered as one of the ways for tuning red emission of Mn4+ doped phosphors to shorter wavelengths. Generally speaking, one could consider also other possible origins of this luminescence: Mn4+ in tetrahedral sites, Mn3+ in tetrahedral and octahedral sites, Mn2+ in tetrahedral and octahedral sites (as first approximation, without taking into account the difference of ionic radii of manganese and substituted ions). However, the narrow band-shape of this “fast” emission suggests that radiative transition occurs between electronic states, the energy difference between which does not depend on the CF strength, i.e. the respective energy levels “go parallel to each other” in the Tanabe-Sugano diagram. The particular case of such transitions is realized for Mn4+ in octahedral CF where the energy of 6
Journal of Luminescence 216 (2019) 116754
N. Khaidukov, et al.
intensity) correspond to Mn4+ luminescence from two Mn4+ emission centers with different local environment, and two broader bands peaked at 669 and 713 nm to luminescence from impurity (Fe3+) and Mn2+ emission centers, respectively. The decomposition of the PLE spectrum into sub-bands cannot be well described in a similar way because of strongly different absorption coefficients for spin-allowed transitions in Mn4+ and spin-forbidden transitions in Fe3+ and Mn2+. The latter weak spin-forbidden transitions can be hardly seen in the excitation spectrum, only somewhere in between spectral bands of strong spin-allowed transitions of Mn4+, and accordingly cannot be well identified. Only the presence of the O2− - Mn2+ CT band in the deep UV region can be recognized. There is one more conclusion which can be formulated on the basis of the performed work. Although the use of some fluxes for the improvement of solid-state synthesis conditions is a common practice, this procedure should be applied with care because the ions of the flux compounds can enter the host crystal lattice and provide sites for doping ions of different valence states, which can degrade luminescence properties of synthesized phosphors. In any case, in contrast to results of work [17], no any increase of Mn4+ luminescence intensity has been observed from the LiAl4O6F:Mn phosphor with increasing amount of MgF2 flux.
Foundation (RSF) project No. 18-13-00407. V. M. would like to thank “Short Term Mobility” Program of CNR for the support of research visit in Istituto di Fisica Applicata “Carrara” (Italy). We want to thank Dr. Cristina Salvatici of the service CEME CNR of Sesto Fiorentino for the time spent in achieving EDX spectra. References [1] Y.-C. Lin, M. Karlsson, M. Bettinelli, Inorganic phosphor materials for lighting, Top. Curr. Chem. (Z) 374 (2016) 21. [2] K. Uheda, N. Hirosaki, Y. Yamamoto, A. Naito, T. Nakajima, H. Yamamoto, Luminescence properties of a red phosphor, CaAlSiN3: Eu2 +, for white light-emitting diodes, Electrochem. Solid State Lett. 9 (2006) H22–H25. [3] R.J. Xie, N. Hirosaki, Silicon-based oxynitride and nitride phosphors for white LEDs - a review, Sci. Technol. Adv. Mater. 8 (2007) 588–600. [4] P. Pust, A.S. Wochnik, E. Baumann, P.J. Schmidt, D. Wiechert, C. Scheu, W. Schnick, Ca[LiAl3N4]:Eu2+ - a narrow-band red-emitting nitridolithoaluminate, Chem. Mater. 26 (2014) 3544–3549. [5] R.-J. Xie, Y.Q. Li, N. Hirosaki, H. Yamamoto, Nitride Phosphors and Solid-State Lighting, CRC Press, Boca Raton, FL, 2011. [6] M.G. Brik, A.M. Srivastava, On the optical properties of Mn4+ ions in solids, J. Lumin. 133 (2013) 69–72. [7] Z. Zhou, N. Zhou, M. Xia, M. Yokoyama, H. T. (Bert) Hintzen, Research progress and application prospect of transition metal Mn4+-activated luminescent materials, J. Mater. Chem. C 4 (2016) 9143–9161. [8] S. Adachi, Photoluminescence spectra and modeling analyses of Mn4+-activated fluoride phosphors: a review, J. Lumin. 197 (2018) 119–130. [9] S. Adachi, Photoluminescence properties of Mn4+-activated oxide phosphors for use in white-LED applications: a review, J. Lumin. 202 (2018) 263–281. [10] Q. Zhou, L. Dolgov, A.M. Srivastava, L. Zhou, Z. Wang, J. Shi, M.D. Dramićanin, M.G. Brik, M. Wu, Mn2+ and Mn4+ red phosphors: synthesis, luminescence and applications in WLEDs. A review, J. Mater. Chem. C 6 (2018) 2652–2671. [11] T. Senden, R.J.A. van Dijk-Moes, A. Meijerink, Quenching of the red Mn4+ luminescence in Mn4+-doped fluoride LED phosphors, Light Sci. Appl. 7 (2018) 8. [12] S. Adachi, Review - Mn4+-activated red and deep red-emitting phosphors, ECS J. Solid State Sci. Technol. 9 (2020) 016001. [13] Q. Wang, J. Liao, L. Kong, B. Qiu, J. Li, H. Huang, H.-r. Wen, Luminescence properties of a non-rare-earth doped oxyfluoride LiAl4O6F: Mn4+ red phosphor for solid-state lighting, J. Alloy. Comp. 772 (2019) 499–506. [14] R. Kobayashi, H. Tamura, Y. Kamada, M. Kakihana, Y. Matsushima, A New Host Compound of Aluminum Lithium Fluoride Oxide for Deep Red Phosphors Based on Mn4+, Fe3+ and Cr3+, ECS Transactions vol. 88, (2019), pp. 225–236. [15] S.F. Belov, T.I. Abaeva, I.A. Rozdin, M.B. Varfolomeev, Interaction of Al2O3 with LiF, Inorg. Mater. 20 (1984) 1675–1676. [16] H.-a. Takahashi, H. Takahashi, K. Watanabe, H. Kominami, K. Hara, Y. Matsushima, Fe3+ red phosphors based on lithium aluminates and an aluminum lithium oxyfluoride prepared from LiF as the Li source, J. Lumin. 182 (2017) 53–58. [17] Y. Wakui, Y.J. Shan, K. Tezuka, H. Imoto, M. Ando, Crystal-site engineering approach for preparation of Mg B2O4:Mn2+, Mn4+ (B = Al, Ga) phosphors: Control of green/red luminescence properties, Mater. Res. Bull. 90 (2017) 51–58. [18] R.D. Shannon, Revised effective ionic radii and systematic studies of interatomic distances in halides and chalcogenides, Acta Crystallogr. A 32 (1976) 751–767. [19] Y. Tanabe, S. Sugano, On the absorption spectra of complex ions II, J. Phys. Soc. Jpn. 9 (1954) 776–779. [20] B. Henderson, G.F. Imbush, Optical Spectroscopy of Inorganic Solids, Clarendon Press, Oxford, 1989. [21] T. Jansen, T. Jüstel, M. Kirm, S. Vielhauer, N.M. Khaidukov, V.N. Makhov, Composition-dependent spectral shift of Mn4+ luminescence in silicate garnet hosts CaY2M2Al2SiO12 (M = Al, Ga, Sc), J. Lumin. 198 (2018) 314–319. [22] T. Hu, H. Lin, Y. Cheng, Q. Huang, J. Xu, Y. Gao, J. Wang, Y. Wang, A highlydistorted octahedron with a C2v group symmetry inducing an ultra-intense zero phonon line in Mn4+-activated oxyfluoride Na2WO2F4, J. Mater. Chem. C 5 (2017) 10524–10532. [23] S. Geschwind, P. Kisliuk, M.P. Klein, J.P. Remeika, D.L. Wood, Sharp-line fluorescence, electron paramagnetic resonance, and thermoluminescence of Mn4+ in αAl2O3, Phys. Rev. 126 (1962) 1684–1686. [24] H. Riesen, T. Monks-Corrigan, N.B. Manson, Temperature dependence of the R1 linewidth in Al2O3:Mn4+: a spectral hole-burning and FLN study, Chem. Phys. Lett. 515 (2011) 241–244. [25] Y. Xu, L. Wang, B. Qu, D. Li, J. Lu, R. Zhou, The role of co-dopants on the luminescent properties of α-Al2O3:Mn4+ and BaMgAl10O17:Mn4+, J. Am. Ceram. Soc. 102 (2019) 2737–2744. [26] S.V. Zvonareva, E.I. Frolova, K. Yu. Chesnokov, N.O. Smirnov, V.A. Pankov, V.Y. Churkin, Luminescent properties of alumina ceramics doped with manganese and magnesium, Opt. Mater. 91 (2019) 349–354. [27] S. Kück, S. Hartung, S. Hurling, K. Petermann, G. Huber, Optical transitions in Mn3+-doped garnets, Phys. Rev. B 57 (1998) 2203–2216. [28] G.T. Pott, B.D. McNicol, Zero-phonon transition and fine structure in the phosphorescence of Fe3+ ions in ordered and disordered LiAl5O8, J. Chem. Phys. 56 (1972) 5246–5254.
5. Conclusions Ceramic samples of LiAl4O6F phosphors doped with 1.0 mol % manganese ions have been synthesized by high-temperature solid-state reaction technique with the addition of different amounts of MgF2 flux (from 0 to 20 mol %). The phosphors demonstrate bright red luminescence with the spectrum composed of a narrow band peaked at 661 nm and a broader and longer-wavelength band peaked in the range 675–720 nm (at 290 K) depending on amount of the used flux. Besides that, these two kinds of emissions have strongly different decay times: τ1 ~ 240 μs, τ2 ~ several milliseconds, respectively (at 290 K). Under cooling down to 10 K the intensity of luminescence from these phosphors increases but decay kinetics and time-resolved spectral shapes show only small changes, although the decay rate slightly decreases and the spectra become better resolved with some additional fine structure. The specific spinel-like crystal structure of LiAl4O6F suggests the presence of vacant cation and anion sites, which provides many possibilities for different local environment near the doping manganese ions entering Al3+ octahedral sites, in particular with the strong deviations from centrosymmetry. Accordingly, the “fast” (hundreds microseconds) and “narrow” emission band observed from the LiAl4O6F:Mn phosphors has been interpreted as ZPL in the emission spectrum of Mn4+ entering strongly distorted octahedral sites for which the pure electronic transition becomes partially dipole-allowed. Because of structural disorder around Mn4+ octahedral sites caused by specific spinel-like crystal structure of the host matrix this band undergoes inhomogeneous broadening. The “slow” (several milliseconds) broad-band emission spectrally overlapped with the “fast” one in the LiAl4O6F:Mn phosphor synthesized without MgF2 flux can be due to the presence of some impurities, e.g. Fe3+. On the other hand, the more-intense “slow” broad-band deep red emission observed from the LiAl4O6F:Mn phosphors synthesized with the use of MgF2 flux has been attributed to luminescence of Mn2+ ions entering the octahedral sites. Conflicts of interest None. Acknowledgements This work was done under support of the Russian Science
7
Contents lists available at ScienceDirect
Journal of Luminescence journal homepage: www.elsevier.com/locate/jlumin
Time- and temperature-resolved luminescence spectroscopy of LiAl4O6F:Mn red phosphors
T
Nicholas Khaidukova, Maria Brekhovskikha, Guido Tocib, Barbara Patrizib, Matteo Vanninib, Angela Pirric, Vladimir Makhovd,∗ a
N. S. Kurnakov Institute of General and Inorganic Chemistry, 31 Leninskiy Prospekt, Moscow, 119991, Russia Istituto Nazionale di Ottica, Consiglio Nazionale delle Ricerche, INO-CNR, via Madonna del Piano 10, 50019, Sesto Fiorentino, Florence, Italy c Istituto di Fisica Applicata “Carrara”, Consiglio Nazionale delle Ricerche, IFAC-CNR, via Madonna del Piano 10, 50019, Sesto Fiorentino, Florence, Italy d P. N. Lebedev Physical Institute, 53 Leninskiy Prospekt, Moscow, 119991, Russia b
A R T I C LE I N FO
A B S T R A C T
Keywords: Photoluminescence Time-resolved spectroscopy Red-emitting Mn4+ phosphors Spinel Zero-phonon line Phosphor-converted LED
Ceramic samples of LiAl4O6F phosphors doped with 1.0 mol. % manganese ions have been synthesized by hightemperature solid-state reaction technique and investigated as possible red-emitting light converters. It has been found that the emission spectrum of LiAl4O6F:Mn phosphors consists of a narrow band peaked at 661 nm and a broader band peaked in the region of 675 ÷ 720 nm depending on synthesis conditions. The intensity of the latter band increases and the peak of this band shifts to longer wavelengths with increasing the fraction of MgF2 used as a flux material. Besides that, these two emission bands have strongly different decay times, namely τ1 ~ 240 μs and τ2 ~ several ms, respectively. The decay kinetics and time-resolved spectra of these two emission components have been studied in the temperature range of 10 ÷ 290 K under pulsed laser excitation at 262 nm. The observed luminescence properties have been analyzed within a model which considers the presence of manganese ions having different valence states (Mn4+ and Mn2+) in the specific spinel-type crystal structure of the LiAl4O6F host matrix, which provides many possibilities for different local environment near the doping manganese ions.
1. Introduction As it is now commonly accepted, the phosphor-converted white LEDs (pc-WLEDs) are superior to all other light sources for general lighting because they have low energy consumption, high efficiency, long service life, compact dimensions, are environmentally safe (do not contain mercury), and are stable to external conditions [1]. Most commercially available pc-WLEDs are based on the simplest and most cost-efficient scheme involving a blue emitting semiconductor chip and a yellow emitting phosphor, such as Ce3+ doped garnet Y3Al5O12:Ce3+ (YAG:Ce). However, such light sources emit bluish (cool) white light, which is not suitable for indoor lighting. The problem of obtaining warm white light can be solved by enriching a red component in the spectrum of pc-WLED. In commercial warm pc-WLEDs some Eu2+ doped nitrides as red emitting phosphors are typically used [2–5]. However, the synthesis of nitride phosphors requires high temperature and pressure, which makes them difficult to produce industrially. Also, such Eu2+ doped phosphors show rather broad emission band with a tail into the deep red and near IR spectral range, where human eyes are
∗
not sensitive (the eye sensitivity strongly drops above 650 nm), i.e. a considerable part of light from such phosphors is wasted thus decreasing luminous efficacy of the light source. Therefore, the development of new narrow-band red emitting phosphors is of tremendous importance for the further optimization of solid-state light sources for indoor illumination applications. It is generally accepted that Mn4+ doped phosphors emitting a narrow-band luminescence in the red spectral range are mostly attractive candidates for application in warm pc-WLEDs [6–12]. An important advantage of this type of phosphors is the absence of rare earth elements in their compositions, such as Eu2+, the prices of which are increased permanently, with the only exception of Ce, which is quite abundant worldwide and relatively cheap. By now a rather large number of rare earth free Mn4+ doped red phosphors have been developed, which are based mainly on fluoride hosts, because Mn4+ narrow-band emission in fluorides is located close to the optimal wavelength for red phosphors near 630 nm. However, one major drawback of such luminescence materials is that toxic fluoric acid is needed for their synthesis. Besides that, such fluoride phosphors are typically not well resistant to high
Corresponding author. E-mail address: [email protected] (V. Makhov).
https://doi.org/10.1016/j.jlumin.2019.116754 Received 27 June 2019; Received in revised form 12 September 2019; Accepted 17 September 2019 Available online 18 September 2019 0022-2313/ © 2019 Elsevier B.V. All rights reserved.
Journal of Luminescence 216 (2019) 116754
N. Khaidukov, et al.
temperature and humidity as well as degrade at high excitation densities. One could expect that oxyfluoride compounds as hosts can provide much more stability of phosphors to external conditions but still have acceptable properties of red luminescence from doping Mn4+ ions. Indeed, diverse Mn4+ doped oxyfluoride phosphors have been extensively studied in recent years (the review of luminescence properties of Mn4+ doped oxyfluorides can be found in Ref. [12]), which shows that spectral properties of them are similar to those of Mn4+ doped fluorides. In particular an oxyfluoride LiAl4O6F:Mn4+ phosphor which possesses relatively broad emission band peaked at ~662 nm in contrast to most other Mn4+ doped oxyfluorides showing red line emission has been proposed just recently [13,14]. In our opinion, luminescence properties of this phosphor are still not completely understood. In the present work the specific features of luminescence from red phosphors of such a type have been studied by using technique of time-resolved low-temperature spectroscopy. LiAl4O6F compound has the spinel-like structure but with some specific crystal chemical properties [15,16]. On the other hand, in Ref. [17] it is claimed that the Mn4+ doped spinel phosphors MgAl2O4 and MgGa2O4 prepared with the excess of Mg show higher intensity of Mn4+ luminescence. In this context, a study of the effects of doping with Mg ions on luminescent properties of Mn4+ in LiAl4O6F has been performed too. 2. Experimental A series of LiAl4O6F ceramic phosphors doped with manganese ions was synthesized by high-temperature solid-state reaction technique. LiF - 2Al2O3 precursors mixed with KMnO4 (1.0 mol %) were uniaxially pressed into pellets with a diameter of 10 mm and a thickness of 3–5 mm and annealed at temperatures of 500–700 °C in normal air atmosphere by taking into account that K+ ions are removed from the material as a result of the subsequent multi-stage thermal treatment. Then, the samples were thoroughly ground, re-pelletized and heat treated at various temperatures 1000–1100 °C in argon atmosphere with the addition of different concentrations of MgF2 flux. The molar amount of the flux was varied in the range from 0 to 20%. Eventually, pellets of synthesized ceramic phosphors were polished for later characterization. The structural-phase composition of the obtained ceramic samples was studied on an X-ray powder diffractometer D8 Advance (Bruker) in a monochromatic CuKα radiation. Identification of the synthesized compounds was performed in the EVA (Bruker) software package using the ICDD PDF-2 database. The unit cell parameters were refined by program AXES using the whole powder pattern fitting toolbox. Energy dispersive X-ray (EDX) analysis of the samples was performed by an Energy-Dispersive System, Octane Elect Super Team Basic (EDAX, AMETEK, Mahwah, NJ, USA). Photoluminescence (PL) and PL excitation (PLE) spectra at room temperature were measured using a spectrofluorimeter PerkinElmer LS55. The time-resolved luminescence studies were performed in the temperature range of 10–290 K using the set-up based on a JOBINYVON SPEX 320 spectrometer and an optical cryostat cooled by a closed cycle refrigerator (CTI-CRYOGENICS). The spectrometer was equipped with the 300 l/mm grating which provided spectral resolution of 0.6 nm with the typical entrance slit width 40 μm. Decay kinetics and time-resolved spectra were measured under pulsed laser excitation at 262 nm (4th harmonics of a Nd:YLF laser). In front of the entrance slit of the spectrometer the cut-off filter Schott GG435 was installed in order to prevent a possible penetration of scattered laser light into the spectrometer. Decay curves were recorded using a photomultiplier tube EMI 9816QB and a digital storage oscilloscope Tektronix TDS 680B, with the spectral band-pass (determined by the exit slit width) varied between 2 and 4 nm depending on the signal intensity. The time resolved spectra were measured with a gated intensified multichannel detector EG&G 1420 with an OMA2000 detector controller with the spectral width of one pixel 0.51 nm. The time gate duration and the delay of the gate with respect to the laser excitation pulse were
Fig. 1. XRD patterns of LiAl4O6F:Mn phosphors synthesized without using MgF2 flux (upper panel) and with using 20% MgF2 flux (lower panel); peaks are labeled by Miller indices of corresponding lattice planes.
controlled using a Stanford DG535 delay generator.
3. Results Synthesis experiments have shown that annealing of the precursor mixtures at temperatures above 1000 °C leads to the formation of monophase LiAl4O6F, without traces of any other phases. The XRD patterns of synthesized phosphors are presented in Fig. 1. LiAl4O6F belongs to cubic crystal system and has the spinel-type structure [15] in which Li+ and a part of Al3+ ions are 6-fold coordinated by anions and the rest of Al3+ is 4-fold coordinated although some disordering of Li+ and Al3+ ions entering octahedral and tetrahedral sites is probably possible. The lattice parameters of the cubic unit cell obtained from XRD analysis are 7.9446 and 7.9705 Å for phosphors synthesized without using MgF2 flux and with using 20% MgF2 flux, respectively. The increase of the lattice parameter after adding MgF2 can reflect the fact that Mg2+ ions substitute for the ions of smaller ionic radius, i.e. Al3+, which is rather unexpected. On the other hand, LiAl4O6F chemical composition does not match the spinel-type stoichiometry, in contrast to similar spinel-type material LiAl5O8 (Li0.5Al2.5O4) [14], and according to Ref. [16] the crystal structure of this compound should contain one vacant cation site and one vacant anion site for every chemical formula unit LiAl4O6F (Li0.5Al2O3F0.5). Due to this feature of the crystal structure, Mg2+ ions (as well as additional F‾ ions) can enter the available vacant sites, which will result in increasing the lattice parameter. The compound synthesized with using MgF2 flux can be described by chemical formula LiMgxAl4O6F1+2x up to (Li,Mg,Al)6(O,F)8 having the spinel-like crystal structure. Moreover, as can be seen in Fig. 1 from the comparison of the full width at half maximum 2
Journal of Luminescence 216 (2019) 116754
N. Khaidukov, et al.
Fig. 3. PLE and PL spectra of LiAl4O6F:Mn phosphors synthesized without using MgF2 flux (upper panel) and with using 20% MgF2 flux (lower panel).
Fig. 2. EDX spectra of LiAl4O6F:Mn phosphors synthesized without using MgF2 flux (upper panel) and with using 20% MgF2 flux (lower panel).
intensities for the XRD patterns, the degree of crystallinity for the sample synthesized with using MgF2 flux is higher than that for the sample prepared without using MgF2 flux. Experimental EDX spectra obtained from the surface of the samples are shown in Fig. 2. However, the content of Li cannot be determined from our EDX measurements, being too light to be detected. The elemental ratio Al:O well enough corresponds to chemical compositions of the compounds, whereas there is a considerable deficiency of F, which can be lost during heat treatment. The content of Mn is also smaller than that introduced in the starting mixture. As can be seen in Fig. 2 in the EDX spectrum of the sample synthesized using 20 mol.% MgF2 flux a considerable amount of Mg (~2.6 at. %) is detected thus confirming that Mg ions enter the crystal lattice of the phosphor but without changing the spinel structural type as confirmed by the XRD data. On the other hand, it should be emphasized that K+ is not detected at all. PL and PLE spectra of LiAl4O6F:Mn phosphors measured at room temperature are shown in Fig. 3. The PL spectrum measured under excitation in the visible spectral range is dominated by the narrow band peaked at ~661 nm for the sample prepared without using flux. With increasing concentration of MgF2 flux this peak shifts slightly to longer wavelengths reaching ~663 nm for the sample synthesized with 20% flux, the PL spectrum becomes broader, especially towards longer wavelengths, and the intensity of the broad part of spectrum increases. For the sample obtained without using flux the shape of the band depends weakly on the excitation wavelength whereas for the sample prepared using MgF2 flux the PL spectrum strongly changes under excitation in the deep UV region (262 nm), namely it becomes dominated by a broad band peaked at ~702 nm. The PLE spectra of the narrow-band emission (661–663 nm) contain three main bands peaked at ~458, ~369 and ~286 ÷ 300 nm, although the spectrum can be well enough decomposed (in the linear
Fig. 4. Decomposition of PL (excitation at 300 nm) and PLE (emission at 663 nm) spectra of the LiAl4O6F:Mn phosphor, synthesized using 20% MgF2 flux, into the sum of Gaussian-like sub-bands. The spectra were first converted into linear energy scale E = hc/λ. The PL spectrum was additionally converted according to I(E) = I(λ)*λ2/hc. After fitting with multi-Gaussian functions the spectra were converted back to the wavelength scale.
energy scale) into at least 5 Gaussian-like sub-bands (see Fig. 4 as an example). In PLE spectra of a broad-band emission (740 nm) for samples prepared using MgF2 flux the peaks at ~458 and ~369 nm become less pronounced, the additional long-wavelength band peaked at ~560 nm appears and the UV band shifts to shorter wavelengths. In addition to the emission in the red spectral range some weak broadband emission is also observed in the green region with a peak maximum at ~500 nm (not shown). The intensity of this green emission increases with amount of used MgF2 flux. One can suggest that the red emission band observed from LiAl4O6F:Mn phosphors consists of at least two sub-bands: the narrowband and the broad-band one, although the decomposition of the emission band into at least four sub-bands gives good enough fitting result (Fig. 4). The measurements of spectrally-resolved decay kinetics have shown that these two emission sub-bands have strongly different decay kinetics: the narrow band (~661 nm) has the characteristic decay time in the range of hundreds of microseconds whereas the broad band 3
Journal of Luminescence 216 (2019) 116754
N. Khaidukov, et al.
Fig. 5. Luminescence decay curves measured from the LiAl4O6F:Mn phosphor (synthesized without using flux) monitoring 661 and 700 nm emissions at 290 K.
Fig. 7. Decay curves of fast (λem = 661 nm, τ ~ 290 μs) and slow (λem = 720 nm, τ ~3.5 ms) components of luminescence recorded from the LiAl4O6F:Mn phosphor, synthesized using 20% MgF2 flux, at temperature 9.6 K. The positions and widths of time gates are shown which were chosen for timeresolved emission spectra measurements.
(675–720 nm) has typical decay time of several milliseconds (Fig. 5). Decay kinetics curves recorded at any wavelength within the width of emission spectrum cannot be well fitted by single-exponential decay but two-exponential fitting gives good enough simulation of these decay curves. This is because of spectral overlapping of different emission components which prevents selection of the pure “fast” or the pure “slow” component. However, decay kinetics of the 661 nm emission is almost single-exponential with a decay time τ ~240 μs with very small contribution of the slower component. The decay curve recorded at 700 nm has the dominant decay component with τ ~5.8 ms but has a contribution of the “fast” decay component just with τ ~240 μs. Decay kinetics of 661 nm emission measured for samples synthesized with the use of flux is essentially the same one as that measured for the sample synthesized without flux whereas decay curves recorded at 720 nm for samples synthesized with flux show slightly shorter decay times than that obtained for 700 nm emission of the sample synthesized without flux. For the sample with maximal (20%) amount of flux the decay time of the “slow” emission is τ ~ 3.0 ms at 290 K. The time-integrated and time-resolved spectra measured at temperature 290 K from the LiAl4O6F:Mn phosphor (synthesized without using flux) are shown in Fig. 6. The optimal time gate durations for
recording time-resolved spectra corresponding to the fast and the slow emission component have been chosen as follows: 100 μs without any delay with respect to excitation pulse and 3 ms with 3 ms delay with respect to the laser pulse, respectively. The positions and widths of time gates are shown in Fig. 7 where decay curves of the “fast” and the “slow” emission component measured at 9.6 K from the LiAl4O6F:Mn phosphor synthesized with 20% MgF2 flux are also presented. As is indicated in Fig. 7, some measurements of time-resolved spectra for the slow emission component have been performed also with 6 ms delay with respect to the laser pulse. Under cooling the luminescence intensity increases but time-resolved spectral shapes show only small changes, although in the PL spectra the features of the fast and the slow component become better resolved. In particular, some fine structure appears in the spectrum of the slow emission component, namely a narrow-line becomes well pronounced in the spectrum (Fig. 8). The shape of the slow-component spectrum practically does not change when the delay is increased to 6 ms (Fig. 9), i.e. this narrow line and a broad sub-band have the same decay rate and accordingly belong to radiative transitions from the same electronic state. This observation allows one to ascribe this narrow line to zero-phonon line (ZPL) of the respective electronic transition. The energy of this ZPL is slightly higher than the energy of the intense line of the fast component (the difference is ~1.5 nm) in the spectrum. On the other hand, this ZPL is not seen in the spectra of LiAl4O6F:Mn phosphors synthesized using high concentration of MgF2 flux. Decay kinetics of the narrow-line emission at 661 nm practically does not change with temperature but small increase of lifetime is observed under cooling from τ ~240 μs at 290 K to τ ~290 μs at 10 K. Similar weak dependence on temperature is observed for decay kinetics of the “slow” broad-band emission. 4. Discussion By taking into account all the experimental results described above, as a first approximation, one can suggest that two kinds of manganese emission centers exist in LiAl4O6F:Mn phosphors, which are responsible for “fast” narrow-band and “slow” broad-band red luminescence. By analogy with the results presented in Refs. [13,14] the narrow-band emission peaked at 661 nm can be attributed to luminescence of Mn4+ ions. Because of very similar ionic radii of Mn4+ (0.53 Å) and Al3+
Fig. 6. Normalized time-integrated and time-resolved emission spectra measured from the LiAl4O6F:Mn phosphor (synthesized without using flux) at 290 K. 4
Journal of Luminescence 216 (2019) 116754
N. Khaidukov, et al.
Fig. 8. Time-resolved emission spectra of the “fast” and “slow” components of luminescence from the LiAl4O6F:Mn phosphor measured in the temperature range of 10–290 K.
transitions and one more broad band in the UV region typically ascribed to the O2− - Mn4+ charge-transfer transition. Using obtained spectral data on Mn4+ luminescence and the standard formulas [20], CF strength parameter Dq as well as Racah parameters B and C can be calculated for the Mn4+ ion in the studied phosphor. For this calculation the peak energies of emission band as well as of 4T2 and 4T1 bands in the excitation spectra were used. The obtained values are: Dq = 2185 cm−1, B = 475 cm−1, C = 3791 cm−1, Dq/B = 4.6, which are similar to those obtained in Ref. [13] and correspond to rather large value of the CF strength. On the other hand, this narrow-band red luminescence observed in LiAl4O6F:Mn phosphors has some specific properties which are different compared to typical properties of most other Mn4+ doped phosphors. First, its decay time is of the order of hundreds of microseconds, even at low temperatures, which looks too short compared to typical decay times of several milliseconds observed in most Mn4+ doped phosphors for this parity- and spin-forbidden transition 2E → 4A2 of Mn4+. Also, the shape of the emission spectrum, namely the ratios of intensities between the central intense band and weaker sub-bands, only slightly change with temperature, which is not typical for Mn4+ emission spectrum where at low temperatures the spectrum is dominated by Stokes vibronic side-bands having longer wavelengths with respect to very weak ZPL and at higher temperatures the shorter-wavelength antiStokes vibronic side-bands appear in the spectrum. These properties of narrow-band luminescence of LiAl4O6F:Mn phosphors can be explained, in principle, if one suggests that in this phosphor the respective radiative transition of Mn4+ becomes partially dipole-allowed, which results in the increase of radiative transition probability, i.e. in the decrease of decay time, as well as to the increase of intensity of ZPL corresponding to pure electronic transition. Accordingly, the central intense line in the emission spectrum of LiAl4O6F:Mn can be ascribed to ZPL corresponding to pure electronic 2E → 4A2 transition of Mn4+. The weaker narrow shorter-wavelength band (at ~640 nm) cannot be ascribed to vibronic side-band of this ZPL because anti-Stokes vibronics with such high intensity (see Fig. 8) cannot be observed at 10 K. Thus, the weaker narrow shorter-wavelength side-band should be ascribed to ZPL related to luminescence of Mn4+ ions entering octahedral sites with different local environment, which can be really expected for such a specific crystal structure of the LiAl4O6F host. Although, according to the Tanabe-Sugano diagram for d3 electron configuration (Mn4+), the energy
Fig. 9. Time-resolved emission spectra of the “fast” and “slow” components of luminescence from the LiAl4O6F:Mn phosphor measured at 10 K. The spectrum of the “slow” component was measured with two delays of the time gate (3 and 6 ms) with respect to the laser pulse.
(0.535 Å) in octahedral coordination [18] and taking into account the commonly accepted statement that Mn4+ ions can be stabilized in the lattice only in octahedral coordination it is natural to suggest that Mn4+ ions occupy predominantly Al3+ octahedral sites of the LiAl4O6F lattice. The charge compensation can be reached by the occupation of neighbor anion vacancy by F‾ ion or by substitution of neighbor F‾ by O2−. According to the energy level scheme of Mn4+ given by the wellknown Tanabe-Sugano diagram for d3 electron configuration in an octahedral crystal field (CF) [19] the Mn4+ doped phosphors show broad absorption bands in the blue and near UV spectral range which are due to relatively strong spin-allowed transitions of Mn4+: 4A2 → 4T2 and 4 A2 → 4T1, respectively, while a narrow-band emission observed from Mn4+ doped phosphors originates from the Mn4+ spin-forbidden transition 2E → 4A2. Indeed, the PLE spectrum of LiAl4O6F:Mn monitoring 661 nm emission shows two characteristic broad bands at ~458 and ~369 nm well corresponding to Mn4+ 4A2 → 4T2 and 4A2 → 4T1 5
Journal of Luminescence 216 (2019) 116754
N. Khaidukov, et al.
the emitting 2E state does not depend on the CF strength, i.e. “goes parallel to the x-axis” in the Tanabe-Sugano diagram for d3 electron configuration. The qualitative analysis of the Tanabe-Sugano diagrams for d7 (Mn4+ in tetrahedron), d4 and d6 (Mn3+ in octahedron and tetrahedron, respectively), and d5 (Mn2+) electron configurations in octahedral CF, shows that practically all these options can be excluded from consideration. The only variant, which can be discussed, is Mn3+ in octahedral CF, by taking into account also that the substitution of Al3+ by Mn3+ does not require any charge compensation. However, in this case the narrow-band emission is typically observed in the IR range [27] and excitation spectrum will be completely different from that obtained in the experiment. Anyway, the possible influence of Mn3+ ions on luminescence properties of LiAl4O6F:Mn phosphors cannot be completely excluded but there is no experimental data for real analysis of this possibility. Some CT transitions, e.g. between manganese ions of different valence states, also cannot be responsible for this narrow-band luminescence because such transitions have typically much faster decay rate, and give a broad-band emission because they are assisted by large lattice relaxation due to the strong redistribution of the charge density. The nature of the “slow” broad-band red emission component in the PL spectrum of the LiAl4O6F:Mn phosphor can be different. The simplest explanation for the existence of such type of luminescence observed from the samples synthesized without MgF2 flux is the presence of some impurities. In particular, the spectral shape of this luminescence is very similar to that of Fe3+ ions (iso-electronic analog of Mn2+ ions) doped into disordered type of the LiAl4O6F matrix [14]. Also, the spectrum and decay time of this luminescence are similar to those observed from Fe3+ ions doped into so-called ordered phase of spinel-type compound LiAl5O8 [28]. Moreover, at low temperature the Fe3+ emission spectrum measured from LiAl5O8:Fe3+ shows well-resolved ZPL as is the case for the low-temperature spectrum of the “slow” broad-band emission component of the LiAl4O6F:Mn phosphor synthesized without MgF2 flux. This is a strong argument in favor of the impurity (Fe3+) nature of “slow” broad-band luminescence observed from the LiAl4O6F:Mn phosphor synthesized without MgF2 flux. On the other hand, the PL spectra of the “slow” broad-band red emission observed from the phosphors synthesized with the use of MgF2 flux are shifted to longer wavelengths and ZPL is practically not seen in the low-temperature spectra. One can suggest that this luminescence is due to the presence of some other optical center in such phosphors. The results of XRD and EDX analysis show that during synthesis of LiAl4O6F:Mn4+ using MgF2 flux, the Mg2+ and additional F‾ ions enter the crystal lattice of synthesized compound, most probably, by the occupation of the vacant sites in the spinel crystal lattice of LiAl4O6F. In fact this creates some more complex multi-component host compound of the type LiMgxAl4O6F1+2x as well as provide the sites for entering manganese ions of different valence states, in particular Mn2+, which can substitute for Mg2+ without charge compensation, and ionic radii of Mn2+ and Mg2+ do not differ so strongly as those of Mn2+ and Al3+. Accordingly, the “slow” (several ms) broad-band deep red emission observed from the LiAl4O6F:Mn phosphors synthesized with the use of MgF2 flux can be ascribed to luminescence of Mn2+ ions entering the octahedral sites. Since all absorption transitions of Mn2+ are spin-forbidden, in PLE spectra the respective bands are hidden behind the stronger spin-allowed and CT transitions of Mn4+ although in the longwavelength region of PLE spectra some additional band peaked at ~560 nm can be recognized as possibly due to the Mn2+ 6A1 → 4T1 transition. Besides that, under excitation in the deep UV region the PL spectrum consists of almost only broad deep-red emission band, i.e. the respective excitation band can be ascribed to the O2− - Mn2+ CT band. Weak green emission, whose intensity also increases with the amount of used MgF2 flux, can be ascribed to luminescence of Mn2+ ions entering the tetrahedral sites. The proposed model of observed luminescence properties fits well the decomposition structure of PL spectrum shown in Fig. 4. Two narrow bands peaked at 663 nm (higher intensity) and 646 nm (lower
of the emitting 2E state does not depend on the CF strength, still the Mn4+ emission spectrum can be remarkably shifted by changes of local environment around the Mn4+ ion because of the nephelauxetic effect (see, e.g. Ref. [21]). One could also mention that the spectral width of this ZPL is rather large (it is not determined by spectral resolution of the set-up) compared to typical ZPL widths. However, in the crystal structure of the LiAl4O6F host matrix with large concentrations of vacancies a considerable structural disorder around Mn4+ ions can be expected because a variety of combinations of ions/vacancies neighbor to Mn4+ ions is possible resulting in fact in the presence of the multiplicity of Mn4+ optical centers having different local environments, which gives inhomogeneous broadening of ZPL. The unusually intense ZPL in the emission spectra of Mn4+ ions has been already observed in some other oxyfluoride hosts, e.g. in Na2WO2F4:Mn4+ [22], but all other features such as Stokes and antiStokes side-bands with typical temperature dependence have been also detected in those spectra. The reason of this effect is the deviation of the Mn4+ site environment from centrosymmetric nature because of large distortions of Mn4+ octahedral coordination observed in oxyfluorides. If the Mn4+ site loses the inversion symmetry, the parity selection rule for the Mn4+ 2E → 4A2 transition will be relaxed by the admixture of the high-lying odd-parity states, which force this transition to become partially electric-dipole allowed thus increasing the decay rate of pure electronic transition and the intensity of ZPL. In oxyfluoride hosts the non-centrosymmetric distortion of octahedral coordination for Mn4+ ions arises because of the difference in charge of O and F ligand anions forming octahedrons around the Mn4+ sites. In the LiAl4O6F host the additional strong distortion of octahedral environment appears because of specific spinel-like crystal structure of this compound which suggests the presence of one vacant cation site and one vacant anion site for every chemical formula unit LiAl4O6F. This feature of the crystal structure can be the reason of ultra-intense ZPL in emission spectrum of Mn4+ ions doped into the LiAl4O6F crystal host. To the best of our knowledge and according to data of review papers [9,12] there is only one example of Mn4+ doped phosphor, namely αAl2O3:Mn4+, which shows an emission spectrum dominated by ZPLs of Mn4+ luminescence (so-called R1 and R2 lines observed in this phosphor at 677 and 673 nm, respectively) without a remarkable contribution of phonon side-bands even at 300 K [23–26]. The reason of such a specific behavior of this phosphor is not explained. However, one could mention that α-Al2O3 has a hexagonal crystal lattice in which the Al3+ cations occupy the non-centrosymmetric trigonally distorted octahedral sites (point group C3). Moreover, Al3+ ions occupy only 2/3 of such sites, i.e. in the α-Al2O3 lattice 1/3 of Al-sites are empty. These properties can be considered to combine α-Al2O3 and LiAl4O6F into one matrix type for Mn4+ doped phosphors. In fact, one could expect that there exist some other Mn4+ phosphors, which demonstrate Mn4+ emission spectra with dominating ZPL, and accordingly the classification of Mn4+ phosphors proposed in review papers [9,12] can be probably corrected. It should be mentioned also that high intensity of ZPL shifts the barycenter of Mn4+ emission spectrum towards shorter wavelengths compared to typical Mn4+ emission spectra, in which the Stokes vibronic side bands dominate, i.e. the increase of ZPL intensity in the emission spectrum of Mn4+ entering strongly distorted octahedral sites can be considered as one of the ways for tuning red emission of Mn4+ doped phosphors to shorter wavelengths. Generally speaking, one could consider also other possible origins of this luminescence: Mn4+ in tetrahedral sites, Mn3+ in tetrahedral and octahedral sites, Mn2+ in tetrahedral and octahedral sites (as first approximation, without taking into account the difference of ionic radii of manganese and substituted ions). However, the narrow band-shape of this “fast” emission suggests that radiative transition occurs between electronic states, the energy difference between which does not depend on the CF strength, i.e. the respective energy levels “go parallel to each other” in the Tanabe-Sugano diagram. The particular case of such transitions is realized for Mn4+ in octahedral CF where the energy of 6
Journal of Luminescence 216 (2019) 116754
N. Khaidukov, et al.
intensity) correspond to Mn4+ luminescence from two Mn4+ emission centers with different local environment, and two broader bands peaked at 669 and 713 nm to luminescence from impurity (Fe3+) and Mn2+ emission centers, respectively. The decomposition of the PLE spectrum into sub-bands cannot be well described in a similar way because of strongly different absorption coefficients for spin-allowed transitions in Mn4+ and spin-forbidden transitions in Fe3+ and Mn2+. The latter weak spin-forbidden transitions can be hardly seen in the excitation spectrum, only somewhere in between spectral bands of strong spin-allowed transitions of Mn4+, and accordingly cannot be well identified. Only the presence of the O2− - Mn2+ CT band in the deep UV region can be recognized. There is one more conclusion which can be formulated on the basis of the performed work. Although the use of some fluxes for the improvement of solid-state synthesis conditions is a common practice, this procedure should be applied with care because the ions of the flux compounds can enter the host crystal lattice and provide sites for doping ions of different valence states, which can degrade luminescence properties of synthesized phosphors. In any case, in contrast to results of work [17], no any increase of Mn4+ luminescence intensity has been observed from the LiAl4O6F:Mn phosphor with increasing amount of MgF2 flux.
Foundation (RSF) project No. 18-13-00407. V. M. would like to thank “Short Term Mobility” Program of CNR for the support of research visit in Istituto di Fisica Applicata “Carrara” (Italy). We want to thank Dr. Cristina Salvatici of the service CEME CNR of Sesto Fiorentino for the time spent in achieving EDX spectra. References [1] Y.-C. Lin, M. Karlsson, M. Bettinelli, Inorganic phosphor materials for lighting, Top. Curr. Chem. (Z) 374 (2016) 21. [2] K. Uheda, N. Hirosaki, Y. Yamamoto, A. Naito, T. Nakajima, H. Yamamoto, Luminescence properties of a red phosphor, CaAlSiN3: Eu2 +, for white light-emitting diodes, Electrochem. Solid State Lett. 9 (2006) H22–H25. [3] R.J. Xie, N. Hirosaki, Silicon-based oxynitride and nitride phosphors for white LEDs - a review, Sci. Technol. Adv. Mater. 8 (2007) 588–600. [4] P. Pust, A.S. Wochnik, E. Baumann, P.J. Schmidt, D. Wiechert, C. Scheu, W. Schnick, Ca[LiAl3N4]:Eu2+ - a narrow-band red-emitting nitridolithoaluminate, Chem. Mater. 26 (2014) 3544–3549. [5] R.-J. Xie, Y.Q. Li, N. Hirosaki, H. Yamamoto, Nitride Phosphors and Solid-State Lighting, CRC Press, Boca Raton, FL, 2011. [6] M.G. Brik, A.M. Srivastava, On the optical properties of Mn4+ ions in solids, J. Lumin. 133 (2013) 69–72. [7] Z. Zhou, N. Zhou, M. Xia, M. Yokoyama, H. T. (Bert) Hintzen, Research progress and application prospect of transition metal Mn4+-activated luminescent materials, J. Mater. Chem. C 4 (2016) 9143–9161. [8] S. Adachi, Photoluminescence spectra and modeling analyses of Mn4+-activated fluoride phosphors: a review, J. Lumin. 197 (2018) 119–130. [9] S. Adachi, Photoluminescence properties of Mn4+-activated oxide phosphors for use in white-LED applications: a review, J. Lumin. 202 (2018) 263–281. [10] Q. Zhou, L. Dolgov, A.M. Srivastava, L. Zhou, Z. Wang, J. Shi, M.D. Dramićanin, M.G. Brik, M. Wu, Mn2+ and Mn4+ red phosphors: synthesis, luminescence and applications in WLEDs. A review, J. Mater. Chem. C 6 (2018) 2652–2671. [11] T. Senden, R.J.A. van Dijk-Moes, A. Meijerink, Quenching of the red Mn4+ luminescence in Mn4+-doped fluoride LED phosphors, Light Sci. Appl. 7 (2018) 8. [12] S. Adachi, Review - Mn4+-activated red and deep red-emitting phosphors, ECS J. Solid State Sci. Technol. 9 (2020) 016001. [13] Q. Wang, J. Liao, L. Kong, B. Qiu, J. Li, H. Huang, H.-r. Wen, Luminescence properties of a non-rare-earth doped oxyfluoride LiAl4O6F: Mn4+ red phosphor for solid-state lighting, J. Alloy. Comp. 772 (2019) 499–506. [14] R. Kobayashi, H. Tamura, Y. Kamada, M. Kakihana, Y. Matsushima, A New Host Compound of Aluminum Lithium Fluoride Oxide for Deep Red Phosphors Based on Mn4+, Fe3+ and Cr3+, ECS Transactions vol. 88, (2019), pp. 225–236. [15] S.F. Belov, T.I. Abaeva, I.A. Rozdin, M.B. Varfolomeev, Interaction of Al2O3 with LiF, Inorg. Mater. 20 (1984) 1675–1676. [16] H.-a. Takahashi, H. Takahashi, K. Watanabe, H. Kominami, K. Hara, Y. Matsushima, Fe3+ red phosphors based on lithium aluminates and an aluminum lithium oxyfluoride prepared from LiF as the Li source, J. Lumin. 182 (2017) 53–58. [17] Y. Wakui, Y.J. Shan, K. Tezuka, H. Imoto, M. Ando, Crystal-site engineering approach for preparation of Mg B2O4:Mn2+, Mn4+ (B = Al, Ga) phosphors: Control of green/red luminescence properties, Mater. Res. Bull. 90 (2017) 51–58. [18] R.D. Shannon, Revised effective ionic radii and systematic studies of interatomic distances in halides and chalcogenides, Acta Crystallogr. A 32 (1976) 751–767. [19] Y. Tanabe, S. Sugano, On the absorption spectra of complex ions II, J. Phys. Soc. Jpn. 9 (1954) 776–779. [20] B. Henderson, G.F. Imbush, Optical Spectroscopy of Inorganic Solids, Clarendon Press, Oxford, 1989. [21] T. Jansen, T. Jüstel, M. Kirm, S. Vielhauer, N.M. Khaidukov, V.N. Makhov, Composition-dependent spectral shift of Mn4+ luminescence in silicate garnet hosts CaY2M2Al2SiO12 (M = Al, Ga, Sc), J. Lumin. 198 (2018) 314–319. [22] T. Hu, H. Lin, Y. Cheng, Q. Huang, J. Xu, Y. Gao, J. Wang, Y. Wang, A highlydistorted octahedron with a C2v group symmetry inducing an ultra-intense zero phonon line in Mn4+-activated oxyfluoride Na2WO2F4, J. Mater. Chem. C 5 (2017) 10524–10532. [23] S. Geschwind, P. Kisliuk, M.P. Klein, J.P. Remeika, D.L. Wood, Sharp-line fluorescence, electron paramagnetic resonance, and thermoluminescence of Mn4+ in αAl2O3, Phys. Rev. 126 (1962) 1684–1686. [24] H. Riesen, T. Monks-Corrigan, N.B. Manson, Temperature dependence of the R1 linewidth in Al2O3:Mn4+: a spectral hole-burning and FLN study, Chem. Phys. Lett. 515 (2011) 241–244. [25] Y. Xu, L. Wang, B. Qu, D. Li, J. Lu, R. Zhou, The role of co-dopants on the luminescent properties of α-Al2O3:Mn4+ and BaMgAl10O17:Mn4+, J. Am. Ceram. Soc. 102 (2019) 2737–2744. [26] S.V. Zvonareva, E.I. Frolova, K. Yu. Chesnokov, N.O. Smirnov, V.A. Pankov, V.Y. Churkin, Luminescent properties of alumina ceramics doped with manganese and magnesium, Opt. Mater. 91 (2019) 349–354. [27] S. Kück, S. Hartung, S. Hurling, K. Petermann, G. Huber, Optical transitions in Mn3+-doped garnets, Phys. Rev. B 57 (1998) 2203–2216. [28] G.T. Pott, B.D. McNicol, Zero-phonon transition and fine structure in the phosphorescence of Fe3+ ions in ordered and disordered LiAl5O8, J. Chem. Phys. 56 (1972) 5246–5254.
5. Conclusions Ceramic samples of LiAl4O6F phosphors doped with 1.0 mol % manganese ions have been synthesized by high-temperature solid-state reaction technique with the addition of different amounts of MgF2 flux (from 0 to 20 mol %). The phosphors demonstrate bright red luminescence with the spectrum composed of a narrow band peaked at 661 nm and a broader and longer-wavelength band peaked in the range 675–720 nm (at 290 K) depending on amount of the used flux. Besides that, these two kinds of emissions have strongly different decay times: τ1 ~ 240 μs, τ2 ~ several milliseconds, respectively (at 290 K). Under cooling down to 10 K the intensity of luminescence from these phosphors increases but decay kinetics and time-resolved spectral shapes show only small changes, although the decay rate slightly decreases and the spectra become better resolved with some additional fine structure. The specific spinel-like crystal structure of LiAl4O6F suggests the presence of vacant cation and anion sites, which provides many possibilities for different local environment near the doping manganese ions entering Al3+ octahedral sites, in particular with the strong deviations from centrosymmetry. Accordingly, the “fast” (hundreds microseconds) and “narrow” emission band observed from the LiAl4O6F:Mn phosphors has been interpreted as ZPL in the emission spectrum of Mn4+ entering strongly distorted octahedral sites for which the pure electronic transition becomes partially dipole-allowed. Because of structural disorder around Mn4+ octahedral sites caused by specific spinel-like crystal structure of the host matrix this band undergoes inhomogeneous broadening. The “slow” (several milliseconds) broad-band emission spectrally overlapped with the “fast” one in the LiAl4O6F:Mn phosphor synthesized without MgF2 flux can be due to the presence of some impurities, e.g. Fe3+. On the other hand, the more-intense “slow” broad-band deep red emission observed from the LiAl4O6F:Mn phosphors synthesized with the use of MgF2 flux has been attributed to luminescence of Mn2+ ions entering the octahedral sites. Conflicts of interest None. Acknowledgements This work was done under support of the Russian Science
7