4+ Red-emitting Mn -doped ﬂuorides are a promising class of materials to improve the color rendering and luminous efﬁcacy of white light-emitting diodes (w-LEDs). For w-LEDs, the luminescence quenching temperature is very important, but surprisingly no systematic research has been conducted to understand the mechanism for thermal 4+ 4+ quenching in Mn -doped ﬂuorides. Furthermore, concentration quenching of the Mn luminescence can be an issue but detailed investigations are lacking. In this work, we study thermal quenching and concentration quenching 4+ 4+ in Mn -doped ﬂuorides by measuring luminescence spectra and decay curves of K TiF :Mn between 4 and 600 K 2 6 4+ 4+ and for Mn concentrations from 0.01% to 15.7%. Temperature-dependent measurements on K TiF :Mn and other 2 6 4+ 4 Mn -doped phosphors show that quenching occurs through thermally activated crossover between the T excited 4 4+ state and A ground state. The quenching temperature can be optimized by designing host lattices in which Mn has a high T state energy. Concentration-dependent studies reveal that concentration quenching effects are limited 4+ 4+ 4+ in K TiF :Mn up to 5% Mn . This is important, as high Mn concentrations are required for sufﬁcient absorption of 2 6 4+ 4+ blue LED light in the parity-forbidden Mn d–d transitions. At even higher Mn concentrations (>10%), the quantum efﬁciency decreases, mostly due to direct energy transfer to quenching sites (defects and impurity ions). Optimization 4+ of the synthesis to reduce quenchers is crucial for developing more efﬁcient highly absorbing Mn phosphors. The 4+ present systematic study provides detailed insights into temperature and concentration quenching of Mn emission 4+ and can be used to realize superior narrow-band red Mn phosphors for w-LEDs. 2+ Introduction their use also has a serious drawback. The Eu emission White light-emitting diodes (w-LEDs) are the next- band is broad and extends into the deep red spectral generation light sources for display and illumination sys- region (λ > 650 nm) where the eye sensitivity is low. This tems because of their small size, high luminous efﬁcacy, causes the luminous efﬁcacy of the w-LED to drop 1–5 and long operation lifetime . Conventional w-LEDs are (reduced lumen/W output). A worldwide search is composed of blue-emitting (In,Ga)N LEDs and green/ therefore aimed at ﬁnding efﬁcient narrow-band red- yellow-emitting and orange/red-emitting phosphors that emitting phosphors that can be excited by blue light. In 5–7 4+ convert part of the blue LED emission . Both phosphors this search, Mn -doped ﬂuoride phosphors, such as 4+ 4+ are necessary to generate warm white light with a high K SiF :Mn and K TiF :Mn , have recently attracted 2 6 2 6 9–13 color rendering index (CRI > 85). The typical red phos- considerable attention . Under blue light excitation, 2+ 4+ phors in w-LEDs are Eu -doped nitrides (e.g., CaAlSiN : Mn -doped ﬂuorides show narrow red line emission 2+ 4,8 13–16 Eu ) . These phosphors exhibit high photo- (λ ~ 630 nm) with high luminescence QEs . Fur- max luminescence (PL) quantum efﬁciencies (QEs > 90%), but thermore, they are prepared through low-cost, simple 11,17 wet-chemical synthesis at room temperature . These 4+ aspects make Mn -doped ﬂuorides very promising red- Correspondence: Tim Senden (email@example.com) emitting phosphors for developing energy-efﬁcient high Condensed Matter and Interfaces, Debye Institute for Nanomaterials Science, color-rendering w-LED systems . Utrecht University, P.O. Box 80000, 3508 TA Utrecht, The Netherlands Soft Condensed Matter, Debye Institute for Nanomaterials Science, Utrecht University, P.O. Box 80000, 3508 TA Utrecht, The Netherlands © The Author(s) 2018 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to theCreativeCommons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. 1234567890():,; 1234567890():,; 1234567890():,; 1234567890():,; Senden et al. Light: Science & Applications (2018) 7:8 Page 2 of 13 understanding of the thermal quenching behavior is Energy migration Quenching site 4+ essential for developing Mn -doped ﬂuoride phosphors 4+ Mn ion with superior quenching temperatures, and thereby improving their potential for application in w-LEDs. Besides thermal quenching, concentration quenching is Excitation 4+ an issue for the application of Mn -doped ﬂuorides in w- 4+ LEDs. As the Mn d–d transitions are parity-forbidden, 4+ high Mn doping concentrations (e.g., 5 mol%) are required for sufﬁcient absorption of the blue LED light . At high dopant concentrations, energy migration among Energy 4+ 26,28 transfer to the Mn ions can result in concentration quenching , 4+ quencher as is illustrated in Fig. 1. If the distance between the Mn ions is small, excitation energy may efﬁciently migrate 4+ from one Mn ion to another until it reaches a quenching site (defect or impurity ion), where the exci- Host cation Radiative emission tation energy is lost non-radiatively (as heat). Studies on 4+ 4 Fig. 1 Concentration quenching for Mn in crystals. At high Mn 4+ + 4+ concentration quenching in Mn -doped ﬂuorides are doping concentrations the Mn ions (orange) are in close 4+ limited. Several works have compared the luminescence proximity in the crystal lattice. If the Mn ions are close together, 4+ 4+ energy transfer between Mn ions (dark blue) causes the excitation properties of ﬂuoride phosphors with varying Mn to migrate through the crystal. Eventually, it may reach a quenching concentrations, but do not measure the actual site such as a vacancy or impurity (dashed circle), where the excitation 4+ Mn concentration in the phosphors by elemental ana- energy is lost as heat. This process competes with radiative emission 29–33 4+ lysis . Determining the Mn concentration is crucial, (red) and reduces the luminescence efﬁciency 4+ as often only a fraction of the Mn ions is incorporated 19,34 during the synthesis . Reports that do perform ele- 4+ 4+ The application of Mn -doped ﬂuoride phosphors in mental analysis study only a small range of Mn doping w-LEDs may, however, be hampered by thermal concentrations and do not provide insight into the role of 4+ 4+ 13,35,36 quenching of the Mn luminescence. Thermal quench- concentration quenching in Mn doped ﬂuorides . ing of the phosphor luminescence is a serious issue, as it An in-depth investigation of concentration quenching in 4+ affects both the efﬁcacy and color stability of the w-LED. Mn -doped ﬂuorides is thus lacking, despite it being very 4+ In high-power w-LEDs, the temperature of the on-chip important for the application of Mn -doped ﬂuorides in phosphor layer easily reaches 450 K. At these elevated w-LEDs. 4+ temperatures, thermal quenching occurs for Mn -doped In this work, we systematically investigate concentration 4+ ﬂuorides. The luminescence quenching temperature T , quenching and thermal quenching in Mn -doped the temperature at which the emission intensity is ﬂuorides. The quenching is studied by measuring lumi- reduced to half of its maximum, is typically between 400 nescence spectra and decay curves in the temperature 15,18,19 4+ and 500 K . Although the temperature dependence range of 4 to 600 K for K TiF :Mn phosphors with 2 6 4+ of the emission intensity has been measured for many Mn concentrations ranging from 0.01 to 15.7 mol% 4+ 4+ Mn -doped ﬂuorides, the understanding of the thermal (actual Mn concentration). The temperature-dependent 4+ quenching behavior is still limited. Most studies do not luminescence measurements of K TiF :Mn and other 2 6 4+ 4+ explain which process quenches the Mn lumines- Mn -doped phosphors demonstrate that thermal 13,20–23 cence . Moreover, the few reports that do propose a quenching occurs because of thermally activated cross- 15 4 4 quenching mechanism disagree. Paulusz states that the over from the T excited state to the A ground state. 2 2 4+ luminescence of Mn -doped ﬂuorides is quenched by This insight into the quenching mechanism shows that 4+ 4 4+ thermally activated crossing of the Mn T excited state the Mn quenching temperature can be raised by ﬁnding 4 24 4+ 4 and A ground state. In contrast, Dorenbos ﬁnds a ﬂuoride hosts that have an increased Mn T level 2 2 relation between the quenching temperature and the energy. Concentration studies show that the lumines- − 4+ 4+ energy of the F → Mn charge-transfer (CT) state and cence QE of K TiF :Mn is high, ~80%, for doping 2 6 4+ therefore suggests that quenching involves crossover concentrations up to 5 mol% Mn . Concentration 4 4+ between the CT state and A ground state. This CT state quenching is limited for these relatively high Mn crossover mechanism was also used by Blasse and our dopant concentrations. At even higher doping con- 4+ 4+ group to explain thermal quenching in Mn -doped centrations of >10 mol%, the QE of K TiF :Mn falls 2 6 25–27 oxides . Finally, other reports claim that the quench- below 60%. Luminescence decay curves indicate ing temperature increases if the radius of the cation that the drop in QE can be attributed to an increased 4+ 11,18 substituted by Mn becomes smaller . A better probability for direct energy transfer to quenching sites Senden et al. Light: Science & Applications (2018) 7:8 Page 3 of 13 2+ 3+ (e.g., defects, impurity ions, Mn , and Mn ), the con- Hamamatsu R928 photomultiplier tube (PMT) with a 4+ centration of which increases with the Mn concentra- grating blazed at 500 nm for detection of emission. For PL tion. The present results provide an improved decay measurements, excitation was done with a tunable understanding of thermal quenching and concentration optical parametric oscillator (OPO) Opotek Opolette HE 4+ quenching in Mn -doped solids and can be used to 355II laser (pulse width 10 ns, repetition rate 10 Hz) and 4+ develop superior Mn -doped ﬂuoride phosphors for emission was detected with a Hamamatsu H74220–60 PMT. The PL decay curves between 300 and 600 K were w-LEDs. recorded on a different setup, which had an Ekspla NT Materials and methods 342B OPO laser (pulse width 5 ns, repetition rate 10 Hz) 4+ Synthesis and characterization of K TiF :Mn phosphors as excitation source and a 0.55 m Triax 550 mono- 2 6 4+ The K TiF :Mn (x%) phosphors were synthesized chromator combined with a Hamamatsu H74220–60 2 6 according to the method of Zhu et al. For the synthesis PMT for detection of emission. All PL decay curves were 4+ of K TiF :Mn (0.8%), 0.0488 g of K MnF (prepared obtained by multi-channel scaling (MCS) with a Pico- 2 6 2 6 37,38 4+ following refs. ) was dissolved in 2.5 mL of a 40 wt% Quant TimeHarp 260 computer card. The K TiF :Mn 2 6 HF solution (Fluka, 40 wt% HF in water). Next, the phosphors were cooled down to 4 K with an Oxford obtained yellow-brown solution was mixed with 4.5730 g Instruments liquid helium ﬂow cryostat. For PL mea- of K TiF (Sigma-Aldrich, p.a.) and then stirred for 1 h at surements between 300 and 600 K samples were heated in 2 6 4+ room temperature to form K TiF :Mn crystals. The a Linkam THMS600 temperature controlled stage. The 2 6 4+ K TiF :Mn phosphor was isolated by decanting the HF PL quantum efﬁciencies of the phosphors were deter- 2 6 solution, washing twice with 15 mL of ethanol and then mined with a calibrated home-built setup, which con- drying the phosphor for 7 h at 75 °C. The other K TiF : sisted of a 65 W Xe lamp, excitation monochromator, 2 6 4+ Mn (x%) phosphors were prepared following the integrating sphere (Labsphere) and CCD camera (Avantes same procedure but using other amounts of K MnF AvaSpec-2048). 2 6 4+ and K TiF as to obtain different Mn doping 2 6 concentrations. Results and discussion 4+ Powder X-ray diffraction (see Supplementary Figure S1) Luminescence of K TiF :Mn 2 6 4+ conﬁrms that the K TiF :Mn (x%) phosphors exhibit For our quenching studies, we examine the lumines- 2 6 4+ the hexagonal crystal structure of K TiF up to the highest cence of K TiF :Mn phosphors with a wide range of 2 6 2 6 4+ 4+ doping concentration of 15.7% Mn . Furthermore, no Mn doping concentrations. A photographic image of 4+ impurities of K MnF or other crystal phases are observed the K TiF :Mn (x%) phosphors is displayed in Fig. 2a. 2 6 2 6 4+ in the diffraction patterns. Scanning electron microscopy The Mn doping concentrations x (molar percentages 4+ 4+ (SEM) images show that most K TiF :Mn phosphor with respect to Ti ) were determined by inductively 2 6 particles are irregularly shaped and have sizes ranging coupled plasma optical emission spectroscopy (ICP-OES). 4+ from 1 to 200 µm (see Supplementary Figure S2a). Some The body color of K TiF :Mn becomes more yellow 2 6 4+ particles have a hexagonal shape, in agreement with the with increasing Mn concentration as a result of hexagonal crystal structure of K TiF (see Supplementary enhanced absorption in the blue. All of the investigated 2 6 4+ 4+ Figure S2b). Energy-dispersive X-ray (EDX) spectra K TiF :Mn phosphors exhibit bright red Mn lumi- 2 6 (see Supplementary Figure S2c) conﬁrm that the phos- nescence under UV photoexcitation. phor particles consist of potassium, titanium, ﬂuorine, Figure 2b depicts the Tanabe–Sugano energy level 4+ 3 and manganese ions. The manganese dopant concentra- diagram of Mn (3d electron conﬁguration) in an 4+ 39,40 3 tions in the K TiF :Mn phosphors were determined octahedral crystal ﬁeld . The diagram gives the d 2 6 with inductively coupled plasma optical emission spec- energy levels as a function of the crystal ﬁeld splitting Δ . 4+ troscopy (ICP-OES). The ICP-OES measurements were Due to its high effective positive charge, Mn experi- performed on a Perkin-Elmer Optima 8300DV spectro- ences a strong crystal ﬁeld and therefore the E state is the meter (λ = 257.61 and 259.37 nm). For the ICP-OES lowest energy excited state. Hence, the emission spectrum em 4+ 4+ analyses, the K TiF :Mn phosphors were dissolved in of K TiF :Mn (0.8%) is dominated by narrow red 2 6 2 6 2 4 aqua regia. emission lines due to spin- and parity-forbidden E→ A transitions, as can be seen in Fig. 2c. The other K TiF : 2 6 4+ Optical spectroscopy Mn (x%) phosphors exhibit similar emission spectra. As 2 4 PL measurements were performed on an Edinburgh the potential energy curves of the E and A states are at 2 4 Instruments FLS920 ﬂuorescence spectrometer, except the same equilibrium position, the E→ A emission is for the PL decay measurements between 300 and 600 K characterized by narrow zero-phonon and vibronic (see below). For recording excitation and emission spec- emission lines. The potential energy curves of the E and tra, we used a 450 W Xe lamp as excitation source and a A states are at the same equilibrium position because 2 Senden et al. Light: Science & Applications (2018) 7:8 Page 4 of 13 a c 4+ v Mn concentration (x %) 1.0 2 4 0.8% Mn E A 0.01 0.1 0.8 1.3 3.8 5.4 9.4 15.7 ZPL 0.8 Anti-Stokes Stokes 0.6 0.4 0.2 0.0 2 4 2 A T A 550 600 650 700 2 1 1 Wavelength (nm) 50 T 4 d 2 4 4 2 1.0 A T 2 2 0.8% Mn 4 4 0.8 30 A T 2 1 T 0.6 1 4 2 E 20 2 G 0.4 4 2 A T 2 1 0.2 0.0 F 4 300 400 500 600 0 10 20 30 40 Wavelength (nm) Δ /B 4+ 4+ 4+ Fig. 2 Mn luminescence of K TiF :Mn .a Photographic image of K TiF :Mn (x%) phosphors with x = 0.01, 0.1, 0.8, 1.3, 3.8, 5.4, 9.4, and 15.7. 2 6 2 6 4+ The phosphors have a white to yellow body color under ambient light (top) and show red Mn luminescence under 365 nm UV illumination 3 4 4 4 4 (bottom). b Tanabe−Sugano energy level diagram of the d electron conﬁguration in an octahedral crystal ﬁeld. The A → T , A → T , and 2 1 2 2 2 4 4+ E→ A transitions of Mn are indicated by the purple, blue and red arrows, respectively. Note that the excitation transitions are displaced for 4+ clarity. For a speciﬁc coordination all transitions take place around the same crystal ﬁeld Δ . c Emission spectrum of K TiF :Mn (0.8%) upon O 2 6 4+ 4+ excitation with blue light (λ = 450 nm). d Excitation spectrum of the red Mn luminescence (λ = 630 nm) from K TiF :Mn (0.8%). Spectra are exc em 2 6 recorded at ambient temperature 2 4 the E and A states originate from the same t electron Figure 2d displays the excitation spectrum of the red 2g 41 4+ 4+ conﬁguration . Mn luminescence from K TiF :Mn . The two broad 2 6 2 4 4 4 The E→ A emission spectrum consists of a weak excitation bands correspond to spin-allowed A → T 2 2 1 4 4 zero-phonon line (ZPL) at ~622 nm and more intense and A → T transitions (violet and blue arrows in 2 2 anti-Stokes and Stokes vibronic emissions (labeled ν , ν , Fig. 2b). In addition, some weak peaks are visible around 3 4 4 2 and ν ) on the high and low energy sides of the ZPL, 600 nm. These peaks are assigned to A → E and 6 2 13,15 4+ 4 2 4 2 2 respectively . The ZPL is very weak because Mn is A → T transitions. The A → T , E transitions are 2 1 2 1 located on a site with inversion symmetry in K TiF :Mn spin-forbidden and therefore low in intensity compared to 2 6 + 4 4 4 . Due to the inversion symmetry, there are no odd-parity the spin-allowed A → T , T transitions. 2 1 2 crystal ﬁeld components to admix opposite parity states 4 2 2 4 4+ into the A and E states and, as a result, the E→ A Temperature dependence of the Mn luminescence 2 2 2 4 4+ transition is electric dipole forbidden. The E→ A To study the thermal quenching of the Mn emission, 4+ transition can become partly allowed, however, by cou- we measure the PL intensity and Mn emission lifetime 4+ pling with asymmetric vibrations that induce odd-parity of K TiF :Mn (0.01%) as a function of temperature 2 6 4+ crystal ﬁeld components. The most intense lines in Fig. 2c between 4 and 600 K. We use a very low Mn doping 2 4 4+ are assigned to E→ A transitions coupling with the concentration of 0.01%, as for higher Mn concentra- asymmetric ν , ν , and ν vibrational modes (phonons) of tions reabsorption of emission and energy transfer 3 4 6 4+ the MnF group. Thermal population of phonons at between Mn ions can occur. These processes will 4+ room temperature allows coupling with ν , ν , and ν inﬂuence (the temperature dependence of) the Mn 3 4 6 2 6 phonon modes in the E excited state (giving rise to the luminescence spectra and decay curves . As a result, with 4+ anti-Stokes lines), while transitions to these phonon a high concentration of Mn ions, the observations may modes in the A ground state can occur at all tempera- not reﬂect the intrinsic thermal quenching properties of 4+ tures (Stokes lines). Mn . E /B Normalized intensity Normalized intensity Senden et al. Light: Science & Applications (2018) 7:8 Page 5 of 13 ac 6 K 2 4 4.0 E A = 631 nm 2 4 K em 126 K 102 K 3.0 175 K 175 K 227 K 2.0 294 K 292 K –1 1.0 373 K 423 K 0.0 448 K 4.0 303 K 473 K 373 K −2 3.0 448 K 2.0 473 K 498 K 1.0 573 K −3 0.0 580 600 620 640 660 680 0 20406080 Wavelength (nm) Delay time (ms) bd 1.2 1.0 0.8 T = 457 K 0.6 T = 462 K 0.4 ½ 4 0.2 0.0 0 0 100 200 300 400 500 0 100 200 300 400 500 Temperature (K) Temperature (K) 4+ 4+ 4+ Fig. 3 Temperature dependence of the Mn luminescence from K TiF :Mn (0.01%). a Emission spectra (λ = 450 nm) of K TiF :Mn 2 6 exc 2 6 4+ (0.01%) at various temperatures between 0 and 600 K. b Integrated PL intensity of K TiF :Mn (0.01%) as a function of temperature. The integrated 2 6 PL intensity I is scaled to the integrated PL intensity at room temperature I . The red and green lines represent ﬁts to Eqs. 6 and 7, respectively. c PL PL RT 4+ 4+ decay curves of the Mn emission from K TiF :Mn (0.01%) at various temperatures between 0 and 600 K (λ = 450 nm and λ = 631 nm). 2 6 exc em 4+ 4+ d Temperature dependence of the Mn emission lifetime for K TiF :Mn (0.01%). The red and green lines represent ﬁts to Eqs. 4 and 8, respectively. 2 6 The cyan line gives the ﬁt for Eq. 4 (red line) divided by two 4+ Figure 3a shows emission spectra of K TiF :Mn decay times. Figure 3c shows a selection of PL decay 2 6 4+ (0.01%) at various temperatures between 4 and 600 K. At curves of K TiF :Mn (0.01%) measured between 4 and 2 6 4+ 2 4 4+ 4 K the Mn E→ A emission spectrum consists of 600 K. The decay of the Mn emission is single expo- zero-phonon and Stokes vibronic lines. Upon raising the nential and becomes faster with increasing temperature. temperature, phonon modes are thermally populated and The PL decay time is on the order of milliseconds, which 2 4 anti-Stokes emission lines appear (solid arrow in Fig. 3a). is expected as the transition between the E and A states 4+ With the appearance of anti-Stokes lines, the relative is both parity- and spin-forbidden. In Fig. 3d, the Mn intensity of the Stokes emission decreases between 4 and emission lifetime (determined from single exponential 300 K. Above 400 K the intensities of both the anti-Stokes ﬁtting) is plotted as a function of temperature. The life- and Stokes emission lines begin to decrease (dashed arrow time shows a steady decrease, starting above 50 K. The in Fig. 3a), which indicates the onset of non-radiative decrease levels off between 300 and 400 K but then shows transitions from the E excited state. The luminescence is a rapid decrease above 400 K. quenched at 600 K. From the measurements, we obtain The temperature dependences observed in Fig. 3b and d the temperature dependence of the integrated PL inten- are quite exceptional. For most luminescent materials, the sity (I ) relative to the integrated PL intensity at room PL intensity and lifetime are relatively constant with PL temperature (I ) (Fig. 3b). The PL intensity of K TiF : temperature and both begin to decrease once thermal RT 2 6 4+ 6,42,43 4+ Mn (0.01%) gradually increases between 4 and 350 K quenching sets in . The PL intensity of K TiF :Mn , 2 6 but then rapidly drops due to the onset of non-radiative however, rises by 40% between 4 and 350 K while the transitions (luminescence quenching). lifetime decreases before thermal quenching takes place. An alternative method to determine the luminescence To understand this peculiar temperature dependence, we quenching temperature is by measuring luminescence ﬁrst discuss how the radiative decay rate of the E state Integrated I (T )/I PL intensity (10 counts) PL RT Normalized intensity Lifetime (ms) Senden et al. Light: Science & Applications (2018) 7:8 Page 6 of 13 2 4 changes with temperature. The E→ A emission of with vibrations (for more details on the vibronic structure 4+ 4 4 15,16,44 K TiF :Mn mainly consists of anti-Stokes and Stokes of the A → T excitation band, see refs. ). As a 2 6 2 2 vibronic emissions (Fig. 2c). Their transition probabilities result, the PL intensity I will scale with temperature PL 20,41,45 increase with phonon population. The population of as : phonon modes is given by the phonon occupation num- hv ber n, which increases with temperature according to : I ðÞ T ¼ IðÞ 0 coth ð6Þ PL 2k T n ¼ ð1Þ expðÞ hv=k T 1 with I(0) being the PL intensity at T = 0 K. The results in where k is the Boltzmann constant and hν is the energy Fig. 3b show that the increase in PL intensity between 4 2 4 of the phonon coupling to the E→ A transition. The and 350 K follows the temperature dependence given by transition probabilities P of the anti-Stokes and Stokes Eq. 6. This conﬁrms that the higher PL intensity at 350 K vibronics scale with n by: is due to a stronger absorption of excitation light. An increase in PL intensity between 4 and 350 K due to Anti Stokes : P ðÞ T ¼ P ðÞ 0 ½ n ð2Þ R R 4+ enhanced absorption is observed for all investigated Mn Stokes : P ðÞ T ¼ P ðÞ 0½ n þ 1 ð3Þ R R doping concentrations (see Supplementary Information). Although the temperature dependence of the PL intensity where P (0) is the transition probability at T = 0 K. As the follows Eq. 6, there is deviation between the ﬁt of Eq. 6 radiative lifetime τ is proportional to 1/[P (anti- R R and the measured data (see red line in Fig. 3b). The model Stokes) + P (Stokes)], it follows from Eqs. 1–3 that: of Eq. 6 is simple and does not take into account the shift τ ðÞ 0 4 4 τ ðÞ T ¼ ð4Þ and broadening of the A → T absorption band with 2 2 cothðhv=2k TÞ temperature. Both these effects also inﬂuence the tem- perature dependence of the PL intensity, and this can Here, τ (0) is the radiative lifetime at T = 0 K. In Fig. 3d, explain the deviation between the model and the experi- Eq. 4 (red line) has been plotted for τ (0) = 12.3 ms and −1 mental data. Including the effect of a shift and broadening hν = 216 cm (phonon energy of the intense ν mode 4 4 of the A → T band on the absorption strength is emission). Equation 4 accurately describes the measured 2 2 4+ complex and will not aid a more accurate determination temperature dependence of the Mn emission lifetime of T . up to 375 K, conﬁrming that the decay of the E state is ½ mainly radiative up to this temperature. The radiative 4+ 4+ Above 400 K the PL intensity of K TiF :Mn (0.01%) 2 6 lifetime of the Mn emission shortens with temperature begins to decrease due to the onset of non-radiative due to thermal population of odd-parity vibrational transitions (Fig. 3a, b). The non-radiative decay prob- modes at higher temperatures. ability rapidly increases with temperature above 400 K and as a result the luminescence is quenched, with no Next, we investigate the increase in PL intensity emission intensity remaining at 600 K. The quenching between 4 and 350 K. The PL intensity I equals the PL 4+ temperature T is determined to be 462 K. The Mn product of the PL QE and number of absorbed photons ½ emission lifetime also rapidly decreases once thermal (as I scales with the number of absorbed photons, the PL 4+ quenching sets in (Fig. 3d). Above 400 K the Mn excitation wavelength can have a large inﬂuence on the emission lifetime is shorter than the radiative lifetime τ temperature dependence observed for I ; see Supple- R PL 4+ predicted by Eq. 4 (red line). The lifetime shortens mentary Information). The PL QE η of K TiF :Mn can 2 6 because of an additional thermally activated non-radiative be expressed as: contribution to the decay of the E state. From the tem- η ¼ ð5Þ perature dependence of the lifetime, T can be deter- γ þ γ R NR mined by locating the temperature at which the lifetime where γ and γ are the radiative and non-radiative has decreased to half of its radiative lifetime value. To R NR decay rates of the emitting E state, respectively. The estimate T , we divide the value from the ﬁt of Eq. 4 for τ ½ R results in Fig. 3d show that the decay of the E state is by a factor of 2 (Fig. 3d, cyan line). The cyan line crosses mainly radiative up to 375 K, so we can assume that γ is the data points at 457 K. This value for T is very close to NR ½ negligible between 0 and 350 K. The value for η is the T of 462 K obtained from the PL intensity therefore approximated as a constant close to unity measurements. 4 4 between 0 and 350 K. On the other hand, the A → T Thermal quenching can be described as a thermally 2 2 absorption will change with temperature. Like the activated process with an activation energy ΔE. The 2 4 4 4 E→ A transition, the A → T transition is electric activation energy is obtained by ﬁtting a modiﬁed 2 2 2 dipole (parity) forbidden and gains intensity by coupling Arrhenius equation to the temperature dependence of the Senden et al. Light: Science & Applications (2018) 7:8 Page 7 of 13 43,46 −1 PL intensity I between 350 and 600 K : quenching process is ~8000 cm . The rate constants A PL and 1/τ should be approximately equal to the vibra- NR IðÞ 0 I ðÞ T ¼ ð7Þ PL tional frequencies of the MnF group. The ν vibrational 1 þ A ´ expðÞ ΔE=k T B 12 −1 mode has a frequency of 6.5 × 10 s , close to the rate constants found by ﬁtting the data to Eqs. 7 and 8. The In Eq. 7, I(0) is the maximum PL intensity, k is the variation in activation energy values and prefactors can be Boltzmann constant and A is a rate constant for the explained by the fact that thermal quenching is not a thermal quenching process. The best ﬁt to Eq. 7 (green −1 simple thermally activated process. Struck and Fonger line in Fig. 3b) gives an activation energy ΔE of 9143 cm 12 have shown that the temperature dependence of a non- and a rate constant A of 2.5 × 10 . We can also determine 4+ radiative process is accurately described by considering ΔE by ﬁtting the temperature dependence of the Mn 47 ground and excited state vibrational wave function over- emission lifetime τ(T) to the following expression : 46,48 lap . According to the Struck–Fonger model, the non- τ ðÞ T radiative process occurs through tunneling (crossover) τðÞ T ¼ ð8Þ τ ðÞ T from a vibrational level of the excited state to a high 1 þ expðÞ ΔE=k T NR vibrational level of the ground state. The tunneling rate, i.e., the non-radiative decay rate, depends on the wave Here, 1/τ is the non-radiative decay rate and τ (T)is NR R function overlap of the vibrational levels involved. The the radiative lifetime as described by Eq. 4 with τ (0) = −1 4+ tunneling rate will be faster for a larger overlap between 12.3 ms and hν = 216 cm .We ﬁt Eq. 8 to the Mn the wave functions and when the vibrational levels are in emission lifetimes (green line in Fig. 3d) and ﬁnd an −1 resonance. For the present discussion, analysis of the data activation energy ΔE of 7100 cm and a prefactor 1/τ NR 12 −1 using complex models such as the Struck–Fonger model of 1.5 × 10 s . On the basis of the two similar values for is not relevant, but it is important to realize that the ΔE, we conclude that the activation energy of the thermal ac 40 CT T 500 ΔE R R 20,750 21,250 21,750 22,250 4 4 –1 A T energy (cm ) 2 2 b d 40 Fluorides CT Oxides ΔE 0 100 R R 17,000 21,000 25,000 4 4 –1 A T energy (cm ) 2 2 4+ Fig. 4 Thermal quenching in Mn -doped ﬂuorides. a, b Conﬁguration coordinate diagrams showing luminescence quenching due to a − 4+ 4+ 4 thermally activated crossover via the F → Mn charge-transfer (CT) state and b thermally activated crossover via the Mn T excited state. c 4+ 4 4 Quenching temperature T of Mn -doped ﬂuoride phosphors as a function of the A → T transition energy. The red dashed line is a linear ﬁtto ½ 2 2 4+ 4+ 4 4 the data points. d Quenching temperature T of Mn -doped ﬂuorides (blue dots) and Mn -doped oxides (red dots) as a function of the A → T ½ 2 2 transition energy 4 –1 4 –1 Energy (10 cm ) Energy (10 cm ) Quenching temperature T (K) Quenching temperature T (K) ½ Senden et al. Light: Science & Applications (2018) 7:8 Page 8 of 13 4 4 Struck–Fonger model gives a more correct description of crossing of the T and A states. The offset of the CT 2 2 the actual quenching process. state is typically larger than the offset of the T state. Note that the diagrams in Fig. 4a and b are schematic 4+ Thermal quenching in Mn -doped ﬂuorides conﬁguration coordinate diagrams to illustrate the dif- 4+ To obtain insight into the thermal quenching of Mn ferent quenching mechanisms. luminescence, we will discuss four possible quenching In Fig. 4a, the CT state has a larger offset ΔR than the 4 4 processes: (1) multi-phonon relaxation, (2) thermally T state, which causes the CT parabola to cross the A 2 2 activated photoionization, (3) thermally activated cross- parabola at lower energies than the T parabola. − 4+ over via the F → Mn charge-transfer (CT) state, and Thermal activation over the energy barrier ΔE will allow 4+ 4 2 (4) thermally activated crossover via the Mn T excited crossover from the E state into the CT state followed by state. non-radiative relaxation to the ground state via the In the conﬁgurational coordinate diagram, the parabolas crossing of the CT and A parabolas. Alternatively, 4+ 2 4 4+ of the Mn E and A states do not cross and lumi- thermal quenching of the Mn luminescence may 2 4 nescence quenching by crossover from the E to the A be due to the mechanism depicted in Fig. 4b. Here, the CT states is not possible (Fig. 4a). The A ground state may state has a smaller offset ΔR compared to that shown in however be reached by multi-phonon relaxation. In Mn Fig. 4a, and its potential curve is therefore at higher + −1 4 -doped ﬂuorides more than 30 phonons of ~500 cm energies. In addition, the T state has a slightly larger 2 4 4 are needed to bridge the energy gap between the E and offset. As a result, the crossing of the T and A para- 2 2 4 49 A states . For such high numbers of phonons (p > 30), bolas is now at a lower energy and non-radiative relaxa- 4 4 it is unrealistic that non-radiative multi-phonon relaxa- tion will proceed via the crossing of the T and A 2 2 tion is responsible for thermal quenching (see Supple- parabolas. mentary Information for a more detailed discussion). The activation energies ΔE in the conﬁguration coor- −1 Alternatively, the thermal quenching can be due to ther- dinate diagrams are ~8000 cm , similar to the ΔE values mally activated photoionization of an electron from the obtained from the temperature-dependent measurements. 4+ 2 Mn E state to the ﬂuoride host conduction band. This indicates that both mechanisms in Fig. 4a, b can 4+ Thermally activated photoionization typically quenches explain the thermal quenching of Mn luminescence. To the emission from a luminescent center if the emitting determine which of these two mechanisms is responsible 26,50 state is close in energy to the host conduction band . for the luminescence quenching, we compare the 4+ quenching temperature T In density functional theory (DFT) calculations, large of K TiF :Mn to the T of ½ 2 6 ½ 4+ band gaps of around 8 eV have been found for ﬂuoride other Mn -doped materials. A relation between the 51,52 hosts like K SiF and K TiF . It is therefore expected quenching temperature and the energy of either the CT or 2 6 2 6 4+ 2 4 that the Mn E state is well below the host conduction T state in a variety of hosts will give insight. If band levels. Based on this, we conclude that thermal quenching occurs by crossover from the CT state 4+ 4 4+ quenching in Mn -doped ﬂuorides is not caused by to the A state, T will be higher for Mn -doped solids 2 ½ 4+ thermally activated photoionization. However, more evi- with higher CT transition energies. In K TiF :Mn and 2 6 4+ − 4+ dence is necessary to exclude this quenching mechanism. other Mn -doped ﬂuorides the F → Mn CT transi- 4+ −113,15 4+ Photoconductivity measurements on Mn phosphors at tion is at ~40,000 cm .Mn -doped oxides have 2− 4+ elevated temperatures need to be performed to provide lower O → Mn CT transition energies of −1 convincing evidence for a possible role of photoionization 30,000–35,000 cm and are therefore expected to have 4+ in the thermal quenching of Mn emission. lower T values than ﬂuorides if quenching occurs by the 4+ 26,27,53,54 4+ Thermal quenching in Mn -doped ﬂuorides has been mechanism in Fig. 4a . Some Mn -doped oxides, suggested to occur by thermally activated crossover via however, have much higher quenching temperatures than 4+ 4 − 4+ 4+ 4+ the Mn T state or the F → Mn charge-transfer Mn -doped ﬂuorides. For example, Mg GeO :Mn , 2 4 6 15,24,26 4+ 4+ (CT) state . Both these states are displaced relative Mg Ge O F :Mn , and Mg As O :Mn have a T 28 7.5 38 10 6 2 11 ½ 4 55–57 4+ 4 to the potential curve of the A ground state (Fig. 4a, b). of ~700 K , while K TiF :Mn and other Mn 2 2 6 4 4 + Hence, the T and CT state parabolas cross the A -doped ﬂuorides have a T of 400–500 K (see also 2 2 ½ ground state parabola. The difference between the Tables 1 and 2). No correlation is found between the Mn potential curve equilibrium positions is given by the offset luminescence quenching temperature and the energy of 4 2 ΔR = R ′ − R . By using the energies of the A → E, the CT transition (see Supplementary Information for an 0 0 2 4 4 4 4+ A → T and A → CT transitions in K TiF :Mn overview and a plot of quenching temperatures and CT 2 2 2 2 6 (Fig. 2d and ref. ) and assuming speciﬁc offsets ΔR for energies). From this we conclude that thermal quenching 4 4+ the T and CT states, we can construct the diagrams in in Mn -doped ﬂuorides is not caused by thermally − 4+ Fig. 4a and b, where non-radiative relaxation occurs either activated crossover from the F → Mn CT state to the 4 4 via (a) the crossing of the CT and A states or (b) the A ground state. 2 2 Senden et al. Light: Science & Applications (2018) 7:8 Page 9 of 13 4+ 4 4 Table 1 Quenching temperature T (K) and A → T Alternatively, thermal quenching of the Mn lumi- ½ 2 2 −1 4+ energy (cm ) for Mn -doped ﬂuoride materials nescence can be caused by thermally activated crossover 4+ 4 via the Mn T excited state (Fig. 4b). To investigate the 4 4 −1 Host lattice A → T energy (cm ) T (K) References 2 2 ½ validity of this mechanism, we compare the T and 4 4 4+ A → T transition energies for K TiF :Mn and a 2 2 2 6 K TiF 21,459 462 This work 2 6 4+ variety of other Mn -doped ﬂuorides. From the litera- K SiF 22,099 518 This work 2 6 4+ ture and measurements on Mn luminescence we have K SiF 22,120 490 15 2 6 collected quenching temperatures and luminescence K GeF 21,280 470 15 spectra, preferably for systems with low doping con- 2 6 4+ centrations. Figures 2d and 3b show that K TiF :Mn 2 6 K TiF 21,190 450 15 2 6 4 4 −1 has a A → T energy of 21,459 cm (maximum of the 2 2 K TiF 21,368 478 13 2 6 4+ excitation band) and a T of 462 K. For K SiF :Mn ,we ½ 2 6 4 4 −1 Na SiF 21,739 488 21 2 6 measured a A → T energy of 22,099 cm and a T of 2 2 ½ 4+ Rb SiF 21,739 480 18 518 K (Supplementary Figure S6, K SiF :Mn BR301-C 2 6 2 6 commercial phosphor from Mitsubishi Chemical, Japan). Rb TiF 21,186 450 18 2 6 In Fig. 4c we plot the quenching temperature T against Rb GeF 21,739 513 60 2 6 4 4 4+ 4+ the A → T energy for K TiF :Mn ,K SiF :Mn and 2 2 2 6 2 6 4+ Cs GeF 21,277 420 22 2 6 many other Mn -doped ﬂuoride phosphors reported in Cs SiF 21,368 430 22 the literature (displayed data also listed in Table 1). The 2 6 data show that the T increases with the energy of the T ½ 2 Cs HfF 20,964 403 44 2 6 state. The clear trend shows that the thermal quenching BaSiF 21,322 430 23 4+ in Mn -doped ﬂuorides is due to thermally activated 4 4 BaSnF 21,008 400 45 crossover from the T excited state to the A ground 2 2 BaTiF 21,142 425 61 state. Further conﬁrmation for this quenching mechanism 4+ is provided by Mn spectra measured at elevated tem- peratures (see Supplementary Information). Supplemen- 4+ tary Figure S7 shows emission spectra of K SiF :Mn at 2 6 4 4 4 4 Table 2 Quenching temperature T (K) and A → T ½ 2 2 T = 573 and 673 K. At 573 K a broad T → A emission 2 2 −1 4+ energy (cm ) for Mn -doped oxide materials band is observed, which is almost completely quenched at 4 4 4 4 −1 673 K. The initial rise of the T → A emission at ele- 2 2 Host lattice A → T energy (cm ) T (K) References 2 2 ½ vated temperatures conﬁrms thermal population of the Mg GeO 23,697 730 55 T level, which eventually leads to thermal quenching of 4 6 4+ all Mn emission via this state. Mg Ge O F 23,923 700 26,55,56 28 7.5 38 10 To investigate whether thermally activated crossing via K Ge O 21,739 373 62 2 4 9 the T state is also responsible for temperature 4+ K Ge O (site 1) 19,231 160 63 2 4 9 quenching in Mn -doped oxides, we extend the K Ge4O (site 2) 21,700 379 63 data set of Fig. 4c with quenching temperatures reported 2 9 4+ for Mn -doped oxides. Figure 4d shows the quenching Rb Ge O (site 1) 19,231 162 63 2 4 9 4 4 temperature T as a function of the A → T energy for ½ 2 2 Rb Ge O (site 2) 20,850 346 63 2 4 9 4+ the Mn -doped ﬂuorides and oxides listed in Tables 1 Y Mg Ge O 23,753 850 64 2 3 3 12 and 2. The results show that T increases with the 4 4 La GaGe O 21,413 420 65 energy of the A → T transition. This indicates that the 3 5 16 2 2 4+ Mn emission in ﬂuorides and oxides are both La ZnTiO 19,608 230 66 2 6 quenched due to thermally activated crossover from the La MgTiO 20,000 250 66 2 6 4 T excited state, and not the CT state as previously 24–27 CaZrO 18,500 300 25,26 suggested in some reports . The present results and 4+ Mg As O 23,810 680 57 analysis provide strong evidence that in many Mn 6 2 11 phosphors the thermal quenching mechanism involves Y Al O 20,619 300 67 3 5 12 thermally activated crossover via the T excited state. A Y Al O 20,833 300 68 3 5 12 contribution from other mechanisms cannot be ruled out Sr Al O 22,222 423 69 4 14 25 and further research, for example, photoconductivity SrLaAlO 19,231 300 53 measurements and high pressure studies, can give addi- tional information on the role of alternative quenching LiGa O 20,000 350 70 5 8 mechanisms. Senden et al. Light: Science & Applications (2018) 7:8 Page 10 of 13 a c 10 6.2 1.0 Quantum efficiency 6.0 0.8 –1 5.8 0 2.5 5 4+ 0.6 Mn (x %) Delay time (ms) 0.01% 0.1% 5.6 0.4 0.8% –2 1.3% 3.8% 5.4 0.2 5.4% Emission lifetime 9.4% 15.7% –3 10 5.2 0.0 0 4 8 12 16 020 40 4+ Mn concentration (%) Delay time (ms) b d e 0 0 0 10 10 10 4+ 4+ 4+ 0.8% Mn 15.7% Mn 15.7% Mn T = 298 K T = 298 K T = 4 K –1 –1 –1 10 10 10 –2 –2 –2 10 10 10 = 5.6 ms = 5.4 ms = 10.6 ms fit fit fit 020 40 020 40 04 20 0 60 80 Delay time (ms) Delay time (ms) Delay time (ms) 4+ 4+ Fig. 5 Luminescence decay and quantum efﬁciency of K TiF :Mn as a function of the Mn doping concentration. a Room-temperature PL 2 6 4+ 4+ decay curves of the Mn emission from K TiF :Mn (x%) for 0.01% (pink), 0.1% (blue), 0.8% (green), 1.3% (orange), 3.8% (purple), 5.4% (cyan), 9.4% 2 6 4+ 4+ (yellow), and 15.7% (red) Mn (λ = 450 nm and λ = 631 nm). b PL decay curve of K TiF :Mn (0.8%) at T = 298 K. The decay time exc em 2 6 4+ corresponding to the mono-exponential ﬁt (red line) is 5.6 ms. The bottom panel shows the ﬁt residuals. c Mn emission lifetime (blue squares) and 4+ 4+ 4+ PL quantum efﬁciency (red dots) of K TiF :Mn with different Mn doping concentrations. d, e PL decay curves of K TiF :Mn (15.7%) at d T = 298 2 6 2 6 K and e T = 4 K. The decay times corresponding to the mono-exponential ﬁts (red lines) are 5.4 and 10.6 ms, respectively. The bottom panels show the ﬁt residuals 4+ As quenching occurs by thermally activated crossover that there is a variation in ΔR for Mn -doped ﬂuorides. via the T excited state, the quenching temperature T of The variation in ΔR is small, however, compared to the 2 ½ 4+ 4 the Mn luminescence is controlled by the energy of the differences in the T energy, and no correlation is 4+ 4 Mn T state (the dependence of T on the energy of observed between the spectral width and quenching 2 ½ 4 4 the T state is shown in Fig. 4c,d). In addition, the T of temperatures. This indicates that the T level energy has 2 ½ 2 4+ the Mn luminescence depends on the offset ΔR the largest inﬂuence on the quenching temperature of 4 4 4+ between the T and A states, as ΔR also determines Mn -doped ﬂuorides. 2 2 4 4 where the T and A states cross in the conﬁguration Finally, in view of applications, it is interesting to see 2 2 coordinate diagram (Fig. 4a,b). The horizontal displace- how we can control the T level energy (and thereby T ) 2 ½ ment of the T parabola will inﬂuence the quenching through the choice of the host lattice. The energy of the 4+ 4 temperature. A variation in ΔR can explain the spread Mn T state depends on the crystal ﬁeld splitting Δ 2 O observed in the data of Fig. 4c and d. To investigate the (Fig. 2b), where Δ is typically larger for shorter Mn–F 4+ 44,58 4+ variation in the offset ΔR for Mn -doped ﬂuorides, we distances . For Mn -doped ﬂuorides the lumines- 4 4 compare the bandwidth of the A → T excitation band cence quenching temperature can therefore be raised by 2 2 4+ 4+ 4+ 4+ − in K TiF :Mn ,K SiF :Mn and Cs HfF :Mn selecting host lattices with short M –F distances 2 6 2 6 2 6 (see Supplementary Figure S9). The width of the (see Supplementary Figure S10a). This is consistent with 4 4 4+ A → T excitation band is controlled by the displace- ﬁndings that T increases if the radius of the M host 2 2 ½ ment of the T state and therefore gives a good indication cation decreases, as expected based on crystal ﬁeld the- 4 4 11,18 4+ of ΔR. Comparison of the A → T bandwidths shows ory . If, however, T is plotted against the M -ligand 2 2 ½ Normalized intensity Normalized intensity Residuals Normalized intensity Residuals Lifetime (ms) Normalized intensity Residuals Quantum efficiency Senden et al. Light: Science & Applications (2018) 7:8 Page 11 of 13 4+ 4+ 4+ distance for both Mn -doped ﬂuorides and Mn -doped transfer for Mn ions close to a quencher. In case of oxides (see Supplementary Figure S10b), no correlation energy migration, a faster decay is also expected for longer 4+ between T and the M -ligand distance is found. This times after the excitation pulse. As this is not observed, 4 4+ shows that the crystal ﬁeld splitting and T energy give a the contribution of energy migration via many Mn ions better indication of the quenching temperature for Mn to quenching sites seems to be small. -doped phosphors. To further investigate the role of energy migration in 4+ the concentration quenching of the Mn emission, we 4+ Concentration quenching measure a PL decay curve of K TiF :Mn (15.7%) at T = 2 6 In addition to insight into thermal quenching, con- 4 K, which is displayed in Fig. 5e. At T = 4 K energy 4+ 4+ centration quenching in Mn -doped ﬂuorides is impor- migration among the Mn ions (blue arrows in Fig. 1) tant for application in w-LEDs. The weak parity-forbidden will be hampered, as there is almost no spectral overlap 4 4 4+ 2 4 4 2 A → T absorption requires that commercial phos- between the Mn E→ A emission and A → E 2 2 2 2 4+ phors have high Mn concentrations. If there is effective excitation lines (see Supplementary Figure S11). Hence, at concentration quenching, the PL decay time and QE will 4 K non-radiative decay due to energy migration to 4+ 4+ decrease when the Mn doping concentration is quenching sites will be suppressed. The Mn decay 26,28 raised . We therefore investigate concentration dynamics in Fig. 5e, however, show that the non-radiative 4+ quenching in K TiF :Mn by measuring the PL decay decay is not suppressed at 4 K. The deviation from single 2 6 4+ 4+ times and QEs of K TiF :Mn phosphors with Mn exponential behavior is similar to that at 300 K. There is 2 6 4+ concentrations ranging from 0.01 to 15.7% Mn . an initial faster decay (single-step energy transfer to Figure 5a presents room-temperature PL decay curves quenching sites) followed by an exponential decay with a 4+ 4+ 4+ of the Mn emission from K TiF :Mn with increasing decay time very close to that measured for Mn at low 2 6 4+ Mn doping concentration x. It can be seen that the PL doping concentrations. This suggests that the decrease in 4+ 4+ decay becomes slightly faster as the Mn concentration QE at higher Mn concentrations is not due to energy increases. We analyze the decay dynamics by single migration. The absence of strong concentration quench- exponential ﬁtting of the PL decay curves. The ﬁt for ing by energy migration is conﬁrmed by the thermal 4+ 4+ K TiF :Mn (0.8%) is shown in Fig. 5b. The ﬁt residuals quenching behavior measured for the different Mn 2 6 (bottom panel) are random and the PL decay thus concentrations. In Supplementary Figure S4, it can be resembles a single exponential. This indicates that the seen that the luminescence quenching temperature is decay of the approximately the same for doping concentrations of E state is mainly radiative. Consequently, the 4+ 4+ K TiF :Mn (0.8%) phosphor has a very high QE of 90%. 0.01% and 15.7% Mn , which shows that effects due to 2 6 Figure 5c gives an overview of the ﬁtted decay times (blue thermally activated energy migration (i.e., concentration 4+ squares) and QEs (red dots) of K TiF :Mn with differ- quenching) are weak. Hence, we conclude that the non- 2 6 4+ 4+ ent Mn concentrations. The emission lifetime barely radiative decay at high Mn concentrations is not caused 4+ shortens if the Mn concentration is increased (5.7 ms by energy migration. Inefﬁcient energy migration can be 4+ 4+ for 0.01% Mn to 5.4 ms for 15.7% Mn ). This suggests understood based on the strongly forbidden character of 2 4 4+ 4+ that energy migration to quenching sites is inefﬁcient in the E→ A transition. This allows only Mn –Mn 4+ K TiF :Mn . To verify this, we look at the QE values energy transfer via short range exchange interaction 2 6 4+ obtained for the K TiF :Mn (x%) phosphors. The QE (see Supplementary Information for details). 2 6 4+ remains above 80% for Mn doping concentrations of 5% We instead assign the non-radiative decay to direct 4+ or less, which shows that concentration quenching is transfer of excitation energy from Mn ions to quench- 4+ indeed limited up to a concentration of 5% Mn ions. ers (green arrow in Fig. 1). This process can occur at all 4+ This result is important for applications in w-LEDs, as temperatures and becomes more efﬁcient at higher Mn 4+ 4+ these high Mn doping concentrations (e.g., 5 mol%) are dopant concentrations. With an increasing Mn dopant required for sufﬁcient absorption of the blue LED light in concentration, the stress on the K TiF lattice grows and 2 6 the parity-forbidden d–d transitions . as a result more crystal defects (i.e., quenchers) may be 4+ 2+ For higher Mn concentrations (x > 10%), non- formed. In addition, Mn in different valence states (Mn 2 3+ 4+ radiative decay from the E excited state becomes stron- and Mn ) may be incorporated at higher Mn con- 4+ 4+ ger, however, and as a result the QE of K TiF :Mn falls centrations. Even if a very small fraction of Mn ions has 2 6 below 60% (Fig. 5c). The non-radiative decay is also visible a different valence state than 4+, effective quenching can 4+ in the PL decay curve of K TiF :Mn (15.7%), shown in occur via metal-to-metal charge-transfer states or direct 2 6 Fig. 5d. The decay is multi-exponential, which proves that energy transfer. Consequently, the probability for energy 4+ 2 with 15.7% Mn the E state decays both radiatively and transfer to quenchers increases, resulting in faster initial 4+ 4+ non-radiatively. The faster initial decay indicates that PL decay and lower QEs for K TiF :Mn at high Mn 2 6 there is enhanced quenching by single-step energy dopant concentrations. Optimized synthesis procedures Senden et al. Light: Science & Applications (2018) 7:8 Page 12 of 13 to reduce quenchers (defects and impurity ions) are thus Conﬂict of interest 4+ The authors declare that they have no conﬂict of interest. crucial for obtaining highly luminescent Mn -doped ﬂuoride phosphors (see also recent work of Garcia- Supplementary information is available for this paper at https://doi.org/ Santamaria et al. on concentration quenching in K SiF : 2 6 10.1038/s41377-018-0013-1. 4+ Mn ). Received: 12 October 2017 Revised: 21 February 2018 Accepted: 7 March 2018 Accepted article preview online: 13 March 2018 Conclusions 4+ Narrow-band red-emitting Mn phosphors form an important new class of materials for LED lighting and displays. For these applications, it is important to under- References stand and control the luminescence efﬁciency. We have 1. The Economist. Charge of the LED brigade: a global switch to LEDs will 4+ change the lighting business. 20 Aug (2011). therefore investigated quenching of the Mn lumines- 4+ 2. Krames,M.R.etal. Status andfutureof high-power light-emitting diodes for cence in Mn -doped ﬂuorides by measuring the PL solid-state lighting. J. Disp. Technol. 3,160–175 (2007). 4+ intensity and luminescence lifetimes of K TiF :Mn 2 6 3. Harbers, G., Bierhuizen, S. J. & Krames, M. R. Performance of high power light 4+ emitting diodes in display illumination applications. J. Disp. Technol. 3,98–109 between 4 and 600 K and for Mn concentrations from (2007). 0.01 to 15.7%. Temperature-dependent measurements of 4. Setlur,A.A.Phosphors forLED-based solid-state lighting. Electrochem. Soc. 4+ 4+ the Mn emission intensity and lifetime for K TiF :Mn 2 6 Interface 18,32–36 (2009). 4+ 5. Smet,P.F., Parmentier, A. B.& Poelman, D. Selectingconversion phosphors for and other Mn -doped phosphors show that thermal white light-emitting diodes. J. Electrochem. Soc. 158,R37–R54 (2011). quenching is caused by thermally activated crossover via 6. Bachmann, V., Ronda, C. & Meijerink, A. Temperature quenching of yellow Ce 4+ 4 the Mn T excited state. As a result, the quenching 2 luminescence in YAG:Ce. Chem. Mater. 21,2077–2084 (2009). 4+ 7. Ye,S., Xiao,F., Pan, Y. X.,Ma, Y. Y. &Zhang,Q.Y.Phosphorsinphosphor- temperature is higher in Mn -doped materials with 4 converted white light-emitting diodes: recent advances in materials, techni- higher T state energies. These ﬁndings can be used to ques and properties. Mater. Sci. Eng. R. Rep. 71,1–34 (2010). 4+ engineer Mn -doped ﬂuoride phosphors with higher 8. Xie, R. J. & Hirosaki, N. Silicon-based oxynitride and nitride phosphors for white LEDs—areview. Sci. Technol. Adv. Mater. 8,588–600 (2007). quenching temperatures for application in high-power w- 9. Setlur,A.A.etal. Energy-efﬁcient, high color-rendering LED lamps using LEDs. oxyﬂuoride and ﬂuoride phosphors. Chem. Mater. 22, 4076–4082 (2010). Furthermore, quantum efﬁciency and luminescence 10. Lin, C. C., Meijerink, A. & Liu, R. S. Critical red components for next-generation 4+ white LEDs. J. Phys. Chem. Lett. 7,495–503 (2016). decay measurements for a wide range of Mn doping 4+ 11. Nguyen, H. D. &Liu,R.S.Narrow-band red-emitting Mn -doped hexaﬂuoride concentrations show that no concentration quenching phosphors: synthesis, optoelectronic properties, and applications in white 4+ 4+ occurs up to 5% Mn in K TiF :Mn . This is important 2 6 light-emitting diodes. J. Mater. Chem. C 4, 10759–10775 (2016). 4+ 12. Radkov, E. V.,Grigorov, L. S.,Setlur, A. A.,& Srivastava, A.M.Red lineemitting for the application of Mn -doped materials in w-LEDs, 4+ phosphor materials for use in LED applications: US 7497973B2. 2009-03-03. as high Mn doping concentrations (e.g., 5 mol%) are 13. Zhu, H. et al. Highly efﬁcient non-rare-earth red emitting phosphor for warm required for sufﬁcient absorption of the blue LED light in white light-emitting diodes. Nat. Commun. 5, 4312 (2014). 4+ 14. McKittrick, J. & Shea-Rohwer, L. E. Review: down conversion materials for solid- the parity-forbidden Mn d–d transitions. At very high 4+ state lighting. J. Am.Ceram.Soc. 97,1327–1352 (2014). Mn doping concentrations (>10 mol%) the quantum 15. Paulusz,A.G.Efﬁcient Mn(IV) emission in ﬂuorine coordination. J. Electrochem. 4+ efﬁciency of K TiF :Mn decreases due to enhanced 2 6 Soc. 120,942–947 (1973). 4+ 4+ 16. Takahashi, T. & Adachi, S. Mn -activated red photoluminescence in K SiF direct energy transfer from Mn to quenching sites. 2 6 4+ 4+ phosphor. J. Electrochem. Soc. 155,E183–E188 (2008). Concentration quenching by Mn –Mn energy 17. Wei, L. L. et al. A low-temperature co-precipitation approach to synthesize 4+ migration is limited. To optimize the efﬁciency in highly ﬂuoride phosphors K MF :Mn (M = Ge, Si) for white LED applications. J. 2 6 4+ Mater. Chem. C 3,1655–1660 (2015). doped Mn phosphors, a synthesis procedure aimed at 4+ 4+ 2+ 18. Sakurai,S., Nakamura, T.&Adachi,S.Rb SiF :Mn and Rb TiF :Mn red- 2 6 2 6 reducing quenching sites (defects, impurity ions, Mn , emitting phosphors. ECS J. Solid State Sci. Technol. 5,R206–R210 (2016). 3+ and Mn ) will be crucial. 19. Sijbom, H. F., Joos, J. J., Martin, l. I. D. J., Van den Eeckhout, K. & Poelman, D. et al 4+ Luminescent behavior of the K SiF :Mn red phosphor at high ﬂuxes and at 2 6 the microscopic level. ECS J. Solid State Sci. Technol. 5, R3040–R3048 (2016). 4+ 20. Kasa, R. & Adachi, S. Red and deep red emissions from cubic K SiF :Mn and 2 6 Acknowledgements hexagonal K MnF synthesized in HF∕KMnO ∕KHF ∕Si solutions. J. Electrochem. 2 6 4 2 We thank Stephan Zevenhuizen and Hans Meeldijk for performing the SEM Soc. 159,J89–J95 (2012). and EDX measurements. Mart Peeters is acknowledged for measuring the PL 4+ 21. Nguyen, H. D., Lin,C.C., Fang, M.H. & Liu, R. S. Synthesis of Na SiF :Mn red 2 6 quantum efﬁciencies. Suzanne Verkleij is acknowledged for taking the 4+ phosphors for white LED applications by co-precipitation. J. Mater. Chem. C 2, photographic images of the K TiF :Mn phosphors. This work is ﬁnancially 2 6 10268–10272 (2014). supported by Technologiestichting STW, which is part of the Nederlandse 22. Arai, Y. & Adachi, S. Optical transitions and internal vibronic frequencies of Organisatie voor Wetenschappelijk Onderzoek (NWO). 2− MnF ions in Cs SiF and Cs GeF red phosphors. J. Electrochem. Soc. 158, 6 2 6 2 6 J179–J183 (2011). 23. Sekiguchi,D., Nara,J.&Adachi,S. Photoluminescence andRaman scattering 4+ spectroscopies of BaSiF :Mn red phosphor. J. Appl. Phys. 113, 183516 (2013). Author details 6 24. Dorenbos, P. Charge transfer bands in optical materials and related defect Condensed Matter and Interfaces, Debye Institute for Nanomaterials Science, level location. Opt. Mater. 69,8–22 (2017). Utrecht University, P.O. Box 80000, 3508 TA Utrecht, The Netherlands. Soft 25. Blasse, G. & de Korte, P. H. M. The luminescence of tetravalent manganese in Condensed Matter, Debye Institute for Nanomaterials Science, Utrecht CaZrO :Mn. J. Inorg. Nucl. Chem. 43,1505–1506 (1981). University, P.O. Box 80000, 3508 TA Utrecht, The Netherlands 3 Senden et al. Light: Science & Applications (2018) 7:8 Page 13 of 13 26. Blasse, G. & Grabmaier, B. C. Luminescent Materials (Springer-Verlag, Berlin, 49. de Jong, M., Seijo, L., Meijerink, A. & Rabouw, F. T. Resolving the ambiguity in 1994). the relation between Stokes shift and Huang–Rhys parameter. Phys. Chem. 4+ 2 27. Senden, T., Broers, F. T. H. & Meijerink, A. Comparative study of the Mn E → Chem. Phys. 17, 16959–16969 (2015). 4 4+ 3+ 3+ 3 2+ 2+ A luminescence in isostructural RE Sn O :Mn pyrochlores (RE = Y ,Lu 50. Dorenbos, P. Anomalous luminescence of Eu and Yb in inorganic 2 2 2 7 + 3+ or Gd ). Opt. Mater. 60,431–437 (2016). compounds. J. Phys. Condens Matter 15, 2645–2665 (2003). 4+ 28. Dexter, D. L. & Schulman, J. H. Theory of concentration quenching in inorganic 51. Jin, Y. et al. Narrow red emission band ﬂuoride phosphor KNaSiF :Mn for phosphors. J. Chem. Phys. 22,1063–1070 (1954). warm white light-emitting diodes. ACS Appl. Mater. Interfaces 8, 11194–11203 4+ 29. Zhou, Q. et al. Mn -activated BaSiF red phosphor: hydrothermal synthesis (2016). and dependence of its luminescent properties on reaction conditions. Mater. 52. Brik, M. G. & Srivastava, A. M. Ab initio studies of the structural, electronic, and Chem. Phys. 170,32–37 (2016). optical properties of K SiF single crystals at ambient and elevated hydrostatic 2 6 30. Jiang, X. et al. Hydrothermal synthesis and photoluminescence properties of pressure. J. Electrochem. Soc. 159,J212–J216 (2012). 4+ 4+ red phosphor BaSiF :Mn for LED applications. J. Mater. Chem. C 2, 53. Zhydachevskii, Y. et al. Spectroscopic properties of Mn ions in SrLaAlO . Opt. 6 4 2301–2306 (2014). Mater. 35, 1664–1668 (2013). 31. Lv, L.,Jiang,X., Huang, S.,Chen, X. & Pan, Y. The formationmechanism, 54. Geschwind, S., Kisliuk, P., Klein, M. P., Remeika, J. P. & Wood, D. L. Sharp-line improved photoluminescence and LED applications of red phosphor K SiF : ﬂuorescence, electron paramagnetic resonance, and thermoluminescence of 2 6 4+ 4+ Mn . J. Mater. Chem. C 2,3879–3884 (2014). Mn in α-Al O . Phys. Rev. 126,1684–1686 (1962). 2 3 32. Zhong, J. et al. Synthesis and optical performance of a new red-emitting 55. Thorington, L. Temperature dependence of the emission of an improved 4+ ZnTiF ·6H O:Mn phosphor for warm white-light-emitting diodes. J. Alloy. manganese-activated magnesium germanate phosphor. J. Opt. Soc. Am. 40, 6 2 Compd. 662,232–239 (2016). 579–583 (1950). 33. Xi, L. & Pan, Y. Tailored photoluminescence properties of a red phosphor 56. Kemeny, G. & Haake, C. H. Activator center in magnesium ﬂuorogermanate 4+ BaSnF :Mn synthesized from Sn metal at room temperature and its for- phosphors. J. Chem. Phys. 33,783–789 (1960). mation mechanism. Mater. Res. Bull. 86,57–62 (2017). 57. Travniçek,M., Kröger,F.A., Botden,T.P.J.&Zalm,P.The luminescence of basic 34. Kim, M., Park, W. B., Bang, B., Kim, C. H. & Sohn, K. S. Radiative and non-radiative magnesium arsenate activated by manganese. Physica 18,33–42 (1952). 4+ decay rate of K SiF :Mn phosphors. J. Mater. Chem. C 3,5484–5489 (2015). 58. Srivastava, A. M. & Brik, M. G. The dependence of 10 Dq crystal ﬁeld parameter 2 6 4+ 3 7 3+ 35. Liao, J., Nie, L., Zhong, L., Gu, Q. & Wang, Q. Co-precipitation synthesis and for Mn (3d conﬁguration) and the magnitude of F level splitting for Eu 4+ 6 luminescence properties of K TiF :Mn red phosphors for warm white light- (4f conﬁguration) on pyrochlore compositions. Opt. Mater. 35,196–200 2 6 emitting diodes. Luminescence 31,802–807 (2016). (2012). 36. Yeo,B.E., Cho, Y. S. &Huh,Y.D.Synthesis andphotoluminescence properties 59. Garcia-Santamaria,F., Murphy,J.E., Setlur,A.A.&Sista, S. P. Concentration 4+ 4+ of a red-emitting phosphor, K SiF :Mn , for use in three-band white LED quenching in K SiF :Mn phosphors. ECS J. Solid State Sci. Technol. 7, 2 6 2 6 applications. Opt. Mater. 51,50–55 (2016). R3030–R3033 (2018). 4+ 37. Bode, H., Jenssen, H. & Bandte, F. Über eine neue darstellung des kalium- 60. Wu, W. L. et al. High color rendering index of Rb GeF :Mn for light-emitting 2 6 hexaﬂuoromanganats(IV). Angew. Chem. 65, 304 (1953). diodes. Chem. Mater. 29,935–939 (2017). 4+ 38. Roesky, H. W. Efﬁcient Preparations of Fluorine Compounds (John Wiley & Sons, 61. Mo, G. et al. Deep red BaTiF :Mn phosphor: synthesis, optical properties and Inc, Hoboken, 2012). application for warm WLED devices. J. Mater. Sci. Mater. Electron 28,8155–8159 39. Tanabe, Y. & Sugano, S. On the absorption spectra of complex ions. I. J. Phys. (2017). 4+ Soc. Jpn. 9, 753–766 (1954). 62. Li, P., Wondraczek, L., Peng, M. & Zhang, Q. Tuning Mn red photo- 4+ 40. Tanabe, Y. & Sugano, S. On the absorption spectra of complex ions II. J. Phys. luminescence in (K,Rb) Ge O :Mn solid solutions by partial alkali substitu- 2 4 9 Soc. Jpn. 9, 766–779 (1954). tion. J. Am.Ceram.Soc. 99,3376–3381 (2016). 4+ 41. Henderson, B. & Imbusch, G. F. Optical Spectroscopy of Inorganic Solids (Oxford 63. Baur, F. & Jüstel, T. Dependence of the optical properties of Mn activated University Press, Oxford, 1989). A Ge O (A = K, Rb) on temperature and chemical environment. J. Lumin. 2 4 9 42. Ogiegło, J. M. et al. Luminescence and luminescence quenching in 177,354–360 (2016). 3+ Gd (Ga,Al) O scintillators doped with Ce . J. Phys. Chem. A 117,2479–2484 64. Jansen, T. et al. Narrow-band deep red photoluminescence of Y Mg Ge O : 3 5 12 2 3 3 12 4+ + (2013). Mn ,Li inverse garnet for high power phosphor converted LEDs. ECS J. Solid 43. Bachmann, V.,Jüstel, T.,Meijerink,A., Ronda, C. &Schmidt,P.J. Luminescence State Sci. Technol. 7, R3086–R3092 (2018). 4+ properties of SrSi O N doped with divalent rare earth ions. J. Lumin. 121, 65. Zhang, S. et al. Novel La GaGe O :Mn based deep red phosphor: a 2 2 2 3 5 16 441–449 (2006). potential color converter for warm white light. RSC Adv. 5,90499–90507 44. Senden, T., van Harten, E. J. & Meijerink, A. Synthesis and narrow red lumi- (2015). 4+ nescence of Cs HfF :Mn , a new phosphor for warm white LEDs. J. Lumin. 66. Takeda, Y., Kato, H., Kobayashi, M., Kobayashi, H. & Kakihana, M. Photo- 2 6 4+ 194,131–138 (2018). luminescence properties of Mn -activated perovskite-type titanates, La M- 4+ 45. Hoshino, R., Nakamura, T. & Adachi, S. Synthesis and photoluminescence TiO :Mn (M = Mg and Zn). Chem. Lett. 44, 1541–1543 (2015). 4+ 4+ properties of BaSnF :Mn red phosphor. ECS J. Solid State Sci. Technol. 5, 67. Chen, D., Zhou, Y. & Zhong, J. A review on Mn activators in solids for warm R37–R43 (2016). white light-emitting diodes. RSC Adv. 6,86285–86296 (2016). 46. Struck, C. W. & Fonger, W. H. Uniﬁed model of the temperature quenching of 68. Riseberg, L. A. &Weber,M.J.Spectrum and anomaloustemperature 2 4 4+ narrow-line and broad-band emissions. J. Lumin. 10,1–30 (1975). dependence of the E → A emission of Y Al O :Mn . Solid State Commun. 2 3 5 12 47. Srivastava, A. M. et al. Unusual luminescence of octahedrally coordinated 9,791–794 (1971). 2+ 2+ divalent europium ion in Cs M P O (M = Ca, Sr). J. Lumin. 129,919–925 69. Peng, M.,Yin,X., Tanner, P.A., Brik,M.G. & Li, P.Site occupancy preference, 2 2 7 4+ (2009). enhancement mechanism, and thermal resistance of Mn red luminescence 3+ 4+ 48. Fonger, W. H. & Struck, C. W. Temperature dependences of Cr radiative and in Sr Al O :Mn for warm WLEDs. Chem. Mater. 27,2938–2945 (2015). 4 14 25 nonradiative transitions in ruby and emerald. Phys.Rev.B 11, 3251–3260 70. Da Fonseca,R.J.M.&Abritta,T. Radiative andnonradiativeprocesses in 4+ (1975). LiGa O :Mn . Phys. B Condens Matter 190,327–332 (1993). 5 8
Light: Science & Applications – Springer Journals
Published: May 23, 2018
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.
All for just $49/month
Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly
Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.
All the latest content is available, no embargo periods.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud