- Email: [email protected]
Quantum efficiency and energy transfer processes in rare-earth doped borate glass for solid-state lighting
Journal of Luminescence 170 (2016) 770–777
Contents lists available at ScienceDirect
Journal of Luminescence journal homepage: www.elsevier.com/locate/jlumin
Quantum efficiency and energy transfer processes in rare-earth doped borate glass for solid-state lighting Franziska Steudel a,n, Sebastian Loos b, Bernd Ahrens a,b, Stefan Schweizer a,b a Fraunhofer Application Center for Inorganic Phosphors, Branch Lab of Fraunhofer Institute for Mechanics of Materials IWM, Lübecker Ring 2, 59494 Soest, Germany b Department of Electrical Engineering, South Westphalia University of Applied Sciences, Lübecker Ring 2, 59494 Soest, Germany
art ic l e i nf o
a b s t r a c t
Article history: Received 7 April 2015 Accepted 21 July 2015 Available online 4 August 2015
Sm3 þ , Eu3 þ and Tb3 þ doped borate glass is investigated for its potential as light converting phosphor for solid-state lighting applications. Concentration-dependent luminescence, quantum efficiency and radiative lifetimes are analysed. The luminescence quantum efficiency exceeds values of more than 80%. Cross-relaxation processes in the single-doped glasses result in luminescence quenching for Sm3 þ with increasing doping level, whereas for Eu3 þ and Tb3 þ an increase is observed. In Sm3 þ /Eu3 þ and Tb3 þ /Eu3 þ double-doped glass the energy transfer processes between the rare-earth ions are investigated in detail. The colour coordinate of the Tb3 þ /Eu3 þ double-doped glass can be shifted from the green to the red spectral range by changing the Tb3 þ -to-Eu3 þ ratio in favour of Eu3 þ . In addition, double doping allows for a change in colour coordinate by using different excitation wavelengths. & 2015 Elsevier B.V. All rights reserved.
Keywords: Photon conversion Rare-earth ion doping Energy transfer LED Colour management
1. Introduction Glass is very versatile and a good host for rare-earth (RE) ions; it provides high optical transparency, good RE ion solubility, and it can be cast in almost any shape or size. Luminescent glasses have attracted much attention in the last decades, in particular for lasers, optical fibres, and optical amplifiers [1]. Borate glass is a suitable optical material with high mechanical, chemical and thermal stability [1,2]. A widespread interest in borate glass is recorded by multiple publications on spectroscopy of RE ions in borate glasses in 2014 [3–8]. For many applications, in particular for high-efficiency devices, the knowledge of the radiative quantum efficiency and the radiative lifetime of the luminescent material is of decisive importance. For solid-state lighting applications, e.g. white lightemitting diodes (LEDs), the colour impression is of great significance and appropriate doping with multiple rare-earth ions becomes important. In particular, the rare-earth ions Eu3 þ and Tb3 þ gained a great technological relevance. Their red and green luminescence is used in cathode ray tubes, fluorescent lamps, and plasma displays and have therefore been intensively studied. In addition, Tb3 þ was found to be a good sensitizer to enhance luminescence efficiency of Eu3 þ via energy transfer [9]. n
Corresponding author. E-mail address: [email protected] (F. Steudel).
http://dx.doi.org/10.1016/j.jlumin.2015.07.032 0022-2313/& 2015 Elsevier B.V. All rights reserved.
In the last years, Eu3 þ and Tb3 þ owned increased interest as a phosphor for white light-emitting diodes. White LEDs are replacing conventional light sources due to their high efficiency, compactness, long operational lifetime, and resultant energy saving. Traditional white LEDs combine a blue LED chip and a yellow YAG:Ce3 þ phosphor [10]. However, this method has a low colourrendering index (CRI) and high colour temperature due to a lack of red emission. For higher CRI conventional phosphors can be double-doped with a red emitter, i.e. Sm3 þ or Eu3 þ or both. Another possibility for white light generation is the combination of a blue LED chip with a Eu3 þ and Tb3 þ doped phosphor. In this work, the optical properties of Sm3 þ , Eu3 þ , and Tb3 þ single-doped barium borate glasses are analysed and crossrelaxation processes between the ions are investigated. Furthermore, the influence of Sm3 þ /Eu3 þ and Tb3 þ /Eu3 þ double-doping and the energy transfer processes between the two different rareearth ions are analysed. Both effects play an essential role in optimization of the luminescent glasses for their implementation in phosphor-converted LEDs (pc-LEDs).
2. Experimental details 2.1. Sample preparation Borate glasses using barium oxide as network modifier were prepared. A ratio of two moles of boron oxide (B2O3) and one mole of barium oxide (BaO) was used. In this ratio the glass network
F. Steudel et al. / Journal of Luminescence 170 (2016) 770–777
771
Table 1 Nominal composition of the double-doped samples under study. Dopant
Composition (mol%)
RE-content (at%)
B2O3
BaO
Sm2O3
Eu2O3
Tb2O3
Sm3 þ /Eu3 þ
65.67 65.33 64.67
32.83 32.67 32.33
1.0 1.0 1.0
0.5 1.0 2.0
– – –
0.50/0.25 / – 0.50/0.50 / – 0.50/1.00 / –
Tb3 þ /Eu3 þ
65.93 65.67 65.33
32.97 32.83 32.67
– – –
0.1 0.5 1.0
1.0 1.0 1.0
–/0.05/0.50 –/0.25/0.50 –/0.50/0.50
consists of the highest possible amount of four-coordinated boron [11]. The glasses were additionally doped with different doping levels of samarium oxide (Sm2O3), europium oxide (Eu2O3), or terbium oxide (Tb2O3). The single-doped glasses are comprised of (66.67 0.67x)B2O3 (33.33 0.33x)BaO xRE2O3 with x ¼0.1, 0.3, 0.5, 1.0, 2.0, 3.0 and 5.0 mol% which equals 0.05–2.5 at%. Table 1 summarizes the nominal chemical composition of the doubledoped glasses. The chemicals were weighed in a platinum gold crucible (Pt/Au 95/5) and melted at 1100 °C for approximately 3 h. The melt was then poured onto a brass block at 500 °C, which is below the glass transition temperature of Tg ¼605 °C for barium borate glass [12]. The glass was kept at this temperature for 3 h to eliminate residual mechanical and thermal stresses before being slowly cooled to room temperature. The glass was then cut into squares of 20 mm 20 mm with a thickness of 3.2 mm and polished to optical quality.
Fig. 1. Energy level diagram of (a) Sm3 þ , (b) Eu3 þ , and (c) Tb3 þ . Possible excitation (arrows from bottom to top) and emission routes (arrows from top to bottom) are indicated [14].
2.2. Experimental set-up Absolute photoluminescence quantum efficiency (QE) measurements were performed with a commercial quantum yields measurement system (Hamamatsu C9920-02G) coupled to a 3.3 inch integrating sphere with a xenon lamp (150 W) as excitation source and a photonic multichannel analyser (PMA 12) as detector. The setup has a measurement accuracy of approx. 3%. The quantum efficiency was determined from emission spectra in the spectral range from 450 to 900 nm. Radiative lifetime measurements were performed under excitation with a temperature-stabilized 405-nm laser-diode (Sanyo DL-3146-151) in case of Sm3 þ single- and Sm3 þ /Eu3 þ double-doped glasses and a 370-nm ultraviolet (UV) LED (Winger WEEUV00-CS) in case of Eu3 þ and Tb3 þ single- and double-doped glasses. A peltier-cooled photomultiplier (EMI 9863/492) coupled to a 300 mm focal length monochromator (Princeton Instruments Acton 2300) was used for detection.
3. Results and discussion 3.1. Single-doped glasses 3.1.1. Photoluminescence Fig. 1 shows the energy level diagrams of (a) Sm3 þ , (b) Eu3 þ , and (c) Tb3 þ [13,14]. Possible excitation (arrows from bottom to top) and emission routes (arrows from top to bottom) are indicated. For all three RE ions, emissions in the blue spectral range (from higher energy levels) are quenched due to the high maximum phonon frequency of 1400 cm 1 in borate glass [15]. Though the spectral positions of rare-earth ion emission are well known, the transition intensity ratios vary due to differences in crystallinity and phonon frequencies of the host material.
Fig. 2. Normalized emission spectra of borate glass doped with (a) Sm3 þ , (b) Eu3 þ and (c) Tb3 þ . The excitation was carried out at 402 nm (Sm3 þ ), 393 nm (Eu3 þ ), and 370 nm (Tb3 þ ). The emission transitions are indicated.
The emission spectra of Sm3 þ , Eu3 þ and Tb3 þ single-doped borate glasses are shown in Fig. 2. For Sm3 þ doping (Fig. 2a), the emission spectra show transitions from the excited state4G5/2 to the ground state levels 6H5/2, 6H7/2, 6H9/2, and 6H11/2 (550– 700 nm); the emission is excited at 402 nm. Upon excitation at 393 nm, the Eu3 þ -doped glass (Fig. 2b) shows the typical Eu3 þ emissions in the red spectral range, which are caused by transitions from the excited state 5D0 to the ground state levels 7F0 (580 nm), 7F1 (592 nm), 7F2 (613 nm), 7F3 (652 nm), and 7F4 (700 nm). The electric-dipole transition 5D0 to 7F2 is hypersensitive to variations in crystal symmetry [16]. The high intensity of this transition in borate glass indicates the amorphous nature of the matrix material with low inversion symmetry for the Eu3 þ ion. In Fig. 2c, the typical Tb3 þ -related emissions at 490 nm, 543 nm,
772
F. Steudel et al. / Journal of Luminescence 170 (2016) 770–777
concentration of 0.05 at% shows a single exponential radiative decay curve due to a low probability of dipole–dipole interaction for the large mean distance between the Sm3 þ ions of 4.5 nm. The radiative decay curves of the Eu3 þ and Tb3 þ single-doped samples are not affected by the rare-earth doping level. For comparison, Fig. 3b gives an overview on the decay behaviour of Eu3 þ and Tb3 þ for a doping level of 0.5 at% each. The poor signal-tonoise ratio in case of the Tb3 þ decay is due to the relatively small absorption coefficient of Tb3 þ [19]. The decay of Eu3 þ and Tb3 þ is single exponential and can be described by
⎛ t ⎞ I (τ ) = I0·exp ⎜ − ⎟. ⎝ τ0 ⎠
Fig. 3. Normalized radiative decay curves of (a) Sm3 þ in barium borate glass with a Sm3 þ doping level from 0.05 to 1 at%, (b) Eu3 þ and Tb3 þ doped barium borate glass with a doping level of 0.5 at% each. The analysed transitions are 4G5/2 to 6H7/2 (Sm3 þ ), 5D0 to 7F2 (Eu3 þ ), and 5D4 to 7F4 (Tb3 þ ). The solid curves represent a best fit on the basis of the Förster model (Sm3 þ ) and a single exponential decay (Tb3 þ , Eu3 þ ). (For interpretation of the references to color in this figure caption, the reader is referred to the web version of this paper.)
583 nm, and 622 nm can be assigned to transitions from the excited state 5D4 to the ground state levels 7FJ (J¼ 6, 5, 4, and 3), respectively. 3.1.2. Lifetime The luminescence decay curves for the transition 4G5/2 to 6H7/2 of Sm3 þ are depicted in Fig. 3a for different doping levels. The lifetime decreases with increasing Sm3 þ doping level and the decay curves show a non-single exponential behaviour due to superposition of radiative and non-radiative relaxation processes. In general, the experimental lifetime, τexp, of an excited level a can be described by
1 = τexp
∑ (A ab + Wab ) b
γ=
π 3/2·cA·R 03
3.1.3. Quantum efficiency Fig. 4 shows the spectrally-resolved photoluminescence quantum efficiency for the 0.5 at% doped samples. The quantum efficiency shows a spectral dependence; the maximum values are collected in Tables 2 and 3. In all cases, the quantum efficiency depends significantly on the rare-earth doping level. For the Sm3 þ doped glass (Fig. 4a), the quantum efficiency decreases significantly with increasing Sm3 þ doping (“luminescence quenching”) from approx. 41% for a concentration of 0.05 at% to 2% for 1 at% doping. The Eu3 þ single-doped borate glass (Fig. 4b) shows a different behaviour: Here, the luminescence quantum efficiency increases from 43% for 0.05 at% to 85% for 1 at% (Table 3). The luminescence quantum efficiency of the Tb3 þ single-doped glass (Fig. 4c) also increases with the doping level from 11% to 75% (Table 3), showing saturation for concentrations higher than 1 at%. 3.1.4. Cross-relaxation The observed changes in luminescence quantum efficiency and luminescence lifetime (Tables 2 and 3) result from cross-relaxation Table 2 Concentration of Sm3 þ ions, c Sm3+ , quantum efficiency, QE, at 402 nm and parameters obtained from an analysis based on the Förster model for different Sm3 þ doping levels. Doping level (at%)
c Sm3+ /1018 cm 3 QE (%)
τexp (ms)
τ0 (ms)
γ
R0 (nm)
0.05 0.15 0.25 0.50 1.00
21.6 62.3 107.1 204.3 426.5
2.41 1.95 1.58 0.97 0.39
2.70 2.53 2.73 2.44 2.47
0.09 0.25 0.62 1.14 2.68
0.66 0.65 0.73 0.72 0.75
40.7 31.1 18.8 9.6 1.9
(2)
where I(t) is the intensity of the radiative decay, I0 the initial intensity, t the time after the excitation pulse, and τ0 the intrinsic lifetime of the donor in the absence of an acceptor. The energy transfer parameter, γ, is defined by 4 3
The obtained intrinsic lifetimes, τ0, are collected in Table 3 and amount to approx. 2.0 ms and 2.5 ms for Eu3 þ and Tb3 þ doping, respectively. These values are comparable with those found in the literature [20].
(1)
with Aab and Wab the radiative and non-radiative transition probabilities from level a to level b. Non-radiative decays include multiphonon relaxation and ion-ion interactions, such as energy transfer and cross-relaxation. For Sm3 þ , the non-single exponential behaviour of the decay curves is caused by non-radiative energy transfer processes between neighbouring Sm3 þ ions arising from dipole–dipole interaction [17]. It can be described by the so-called Förster model [18]
⎛ t t ⎞ I (t ) = I0·exp ⎜ − − γ · ⎟ τ0 ⎠ ⎝ τ0
(4)
(3)
with cA the concentration of acceptors and R0 the critical distance. The results obtained from a best fit on the basis of the abovedescribed Förster model are shown as red solid curves in Fig. 3a; the corresponding fitting parameters are listed in Table 2. The experimental lifetimes, τexp, decrease from 2.6 to 0.4 ms with increasing doping level, whereas the energy transfer parameter, γ, increases from 0.1 to 2.7. The intrinsic lifetime, τ0, is approx. 2.6 ms and the critical distance approx. 0.7 nm. The sample with a Sm3 þ
Table 3 Quantum efficiencies, QE, at 464 nm (Eu3 þ ) and 484 nm (Tb3 þ ) and intrinsic lifetimes, τ0, for different Eu3 þ and Tb3 þ concentrations. Eu3 þ c (at%) 0.05 0.15 0.25 0.50 1.00 1.50 2.50 a
Tb3 þ
QE (%)
τ0 (ms)
QE (%)
τ0 (ms)
43 50 59 60 84a 81a 85a
2.10 2.02 1.95 2.00 2.03 2.02 1.98
11 25 45 63 75 75 70
2.55 2.54 2.57 2.55 2.50 2.48 2.41
Different Eu2O3 raw material.
F. Steudel et al. / Journal of Luminescence 170 (2016) 770–777
Fig. 4. Quantum efficiency spectra of (a) Sm3 þ , (b) Eu3 þ , and (c) Tb3 þ singledoped barium borate glass with a doping level of 0.5 at% each. The most intense transitions are indicated.
773
Fig. 6. Photoluminescence emission spectra of (a) 0.5 at% Sm3 þ and (b) 0.5 at% Eu3 þ single-doped barium borate glass as well as (c) Sm3 þ /Eu3 þ double-doped barium borate glass with a doping level of 0.5 at% each. The emission spectra were recorded upon 480-nm excitation. 5 D4 increases with increasing concentration enabling higher luminescence quantum efficiency values for higher concentrations.
3.2. Double-doped glasses
Fig. 5. Cross-relaxation channels of (a) Sm3 þ , (b) Eu3 þ , and (c) Tb3 þ [22–24].
processes, i.e. non-radiative energy transfer processes between neighbouring Sm3 þ , Eu3 þ , or Tb3 þ ions arising from dipole–dipole interaction. The cross-relaxation rate, kCR, depends significantly on the distance, r, between the interacting ions, namely kCR∼r 6 [21]. Possible cross-relaxation channels are depicted in Fig. 5. Sm3 þ has four cross-relaxation channels (Fig. 5a), all starting from the emission level 4G5/2 [17,22] leading to a depopulation of the emission level and thus a decrease in luminescence quantum efficiency and luminescence lifetime with increasing Sm3 þ concentration. Eu3 þ and Tb3 þ have two cross-relaxation channels each (Fig. 5b and c). Since the cross-relaxation rate depends on the frequency, i.e. kCR∼ν 4 [21], Eu3 þ 's low-energy cross-relaxation channel (5D1, 7F0) → (5D0, 7F3) is favoured over its higher energy cross-relaxation channel (5D2, 7F1) → (7F4, 5D1). This results in a higher population of the emission level 5D0 and thus to increased emission intensities. For Tb3 þ , multipolar transfer interactions are possible due to an equal energy gap of approx. 0.6 eV between the energy levels 5D3 and 5D4 as well as 7F6 and 7F0 enabling a relaxation from 5D3 to 5D4. The population of the emission level
3.2.1. Photoluminescence The photoluminescence emission spectra of the double-doped glasses show the same emission lines as the single-doped glasses but with different intensity ratios: The Sm3 þ /Eu3 þ double-doped glass (Fig. 6c) is not the sum of the corresponding spectra of the singledoped glasses (Fig. 6a and b). For 480-nm excitation, Sm3 þ (a) has a strong absorption band due to a transition from the ground state 6 H5/2 to the excited state 4G7/2. The typical Sm3 þ -related emissions are strong in intensity. For Eu3 þ (b) the situation is different: at 480 nm, the emission is low in intensity (Fig. 6b) due to the small Eu3 þ absorption coefficient at this wavelength. For the double-doped sample (c) the emission spectrum exhibits as expected Sm3 þ emissions, but also intense Eu3 þ -related emissions. Appropriate scaling of the spectra from the single-doped glasses (so that the sum of both fits the spectrum of the double-doped glass) reveals an increase in Eu3 þ emission by a factor of 16, whereas the Sm3 þ intensity is scaled by a factor of 0.62. The same behaviour is observed for Tb3 þ /Eu3 þ double-doping (Fig. 7): upon 484-nm excitation, the intensity of the Eu3 þ emission is increased by a factor of 7.7 while the Tb3 þ emission is only about one third of its intensity in single-doped glass. 3.2.2. Lifetime The decay curves of the Sm3 þ /Eu3 þ double-doped samples (Fig. 8a) after excitation at 405 nm (Sm3 þ -related 6H5/2 to 6P5/2 transition) and detection at 598 nm (Sm3 þ -related 4G5/2 to 6H7/2 transition) show a non-single exponential behaviour similar to the Sm3 þ single-doped samples (Fig. 3a). To minimize Sm3 þ crossrelaxation effects, the Sm3 þ concentration is kept constant and only the Eu3 þ concentration is varied. However, increasing the Eu3 þ doping level does not significantly affect the decay behaviour. The parameters obtained from the Förster model are
774
F. Steudel et al. / Journal of Luminescence 170 (2016) 770–777
are summarized in Table 4. For constant Tb3 þ concentration and increasing Eu3 þ doping level, the energy transfer parameter γ increases from zero to 0.24 while the experimental lifetime decreases from 2.55 ms for Tb3 þ single-doped to 1.65 ms for an additional doping with 0.5 at% Eu3 þ . The critical distance, R0, decreases with higher Eu3 þ concentration.
Fig. 7. Photoluminescence emission spectra of (a) 0.5 at% Tb3 þ and (b) 0.5 at% Eu3 þ single-doped barium borate glass as well as (c) Tb3 þ /Eu3 þ double-doped barium borate glass with a doping level of 0.5 at% each. The emission spectra were recorded upon 480-nm excitation.
3.2.3. Quantum efficiency The absolute photoluminescence quantum efficiency spectra of the double-doped samples with 0.5 at% doping level each are shown in Fig. 9 (solid curves): For Sm3 þ /Eu3 þ doping (Fig. 9a) the quantum efficiency of the double-doped glass is significantly lower than the sum of both single-doped glasses (dashed) due to cross-relaxation processes between neighbouring Sm3 þ ions. The obtained photoluminescence quantum efficiency values are summarized in Table 4 for 480-nm excitation. For constant Sm3 þ concentration and increasing Eu3 þ doping, the efficiency increases slightly from 10% to 12% which is less than found for Eu3 þ singledoped glasses. The quantum efficiency of the Tb3 þ /Eu3 þ double-doped glass (Fig. 9b, solid) is similar to the sum of both single-doped glasses (dashed) and reaches values higher than 80%. Only for wavelengths with high Tb3 þ absorption, i.e. 484 nm and 380 nm, the sum of the quantum efficiency spectra of both single-doped glasses is higher, since at these wavelengths, re-absorption decreases the quantum efficiency. For 484-nm excitation, the quantum Table 4 Parameters obtained from Förster-model analysis for different Sm3 þ /Eu3 þ and Tb3 þ /Eu3 þ doping levels and quantum efficiency values at 480 nm (Sm3 þ /Eu3 þ ) and at 484 nm (Tb3 þ /Eu3 þ ). Donor
RE content (at%)
τexp
τ0 (ms)
γ
R0 (nm)
QE (%)
(ms) cdonor
Eu3 þ
Sm3 þ
0.5 0.5 0.5 0.5
0 0.25 0.5 1.0
0.97 1.03 1.12 1.16
2.44 2.72 2.55 2.35
0 0.56 0.46 0.40
– 0.903 0.669 0.512
10 10 11 12
Tb3 þ
0.5 0.5 0.5 0.5
0 0.05 0.25 0.50
2.55 2.45 2.30 1.65
2.55 2.70 2.56 2.25
0 0.11 0.17 0.24
– 0.70 0.48 0.42
63 63 62 62
Fig. 8. Normalized radiative decay of (a) Sm3 þ /Eu3 þ double-doped barium borate glass after excitation at 405 nm and detection at 598 nm (Sm3 þ -related 4G5/2 to 6 H7/2 transition) and (b) Tb3 þ /Eu3 þ double-doped barium borate glass after excitation at 370 nm and detection at 543 nm (Tb3 þ -related 5D4 to 7F4 transition) and 613 nm (Eu3 þ -related 5D0 to 7F2 transition). The red curves result from Förster model (Sm3 þ /Eu3 þ and Tb3 þ /Eu3 þ at 543 nm) and single exponential decay analyses (Tb3 þ /Eu3 þ at 613 nm).
summarized in Table 4. The radiative lifetime is approx. 1 ms and increases slightly upon increasing the Eu3 þ doping level. The intrinsic lifetime, τ0, is constant at approx. 2.5 ms, the energy transfer parameter, γ, decreases with increasing Eu3 þ concentration due to direct excitation of Eu3 þ at 405 nm. For Tb3 þ /Eu3 þ double-doping (Fig. 8b), the decay curve after excitation at 370 nm (Tb3 þ -related 7F6 to 5G6 transition) and detection at 613 nm (Eu3 þ -related 5D0 to 7F2 transition) is single exponential and does not indicate any energy transfer processes from Eu3 þ to other ions; the lifetime is 2.1 ms, which is nearly the lifetime for Eu3 þ single-doping. However, for detection at 543 nm (Tb3 þ -related 5D4 to 7F5 transition) the radiative decay cannot be described by a simple single exponential decay any longer, so that an analysis based on the Förster model is applied. The parameters
Fig. 9. Experimental quantum efficiency of the double-doped glasses (solid) and sum of the corresponding single-doped glasses (dashed) of (a) Sm3 þ /Eu3 þ and (b) Tb3 þ /Eu3 þ doped glass with a doping level of 0.5 at% each.
F. Steudel et al. / Journal of Luminescence 170 (2016) 770–777
775
efficiency values of the Tb3 þ /Eu3 þ double-doped glass with 0.5 at% Tb3 þ and various Eu3 þ doping levels are collected in Table 4. At this wavelength, the quantum efficiency is not affected by the Eu3 þ concentration. 3.2.4. Energy transfer The photoluminescence spectra of the double-doped glasses show a clear energy transfer from Sm3 þ and Tb3 þ to Eu3 þ . The energy transfer route with the highest energy transfer rate, kET , for Sm3 þ /Eu3 þ double-doping is shown in Fig. 10: from the Sm3 þ -related emission level 4G5/2 the energy is released by relaxation to the ground state 6H5/2 and partially transferred to Eu3 þ , which is excited from its ground state to the excited state 5D0 by resonance between the two transitions [25]. The energy mismatch between the two transitions amount to 700 cm 1 (86 meV), which is the best overlap between Sm3 þ emission and Eu3 þ excitation. The energy transfer route for Tb3 þ /Eu3 þ double-doping with the best overlap between Tb3 þ emission and Eu3 þ excitation is shown in Fig. 11, black arrows. It is found for the 5D4 to 7F4 emission from Tb3 þ and the 7F1 to 5D0 excitation from Eu3 þ . Here, the energy mismatch amounts to only 300 cm 1. The population of the Eu3 þ -related energy level 7F1 is possible at room temperature (RT) due to thermal population according to the Boltzmann distribution. Here, the ratio of the population probabilities of the energy levels 7F1 and 7F0 amounts to N1/N0=0.56, which means that 35% of the electrons occupy 7F1. For low
Fig. 10. Energy transfer route between Sm3 þ and Eu3 þ : the non-radiative relaxation energy from the Sm3 þ -related4G5/2 level is used to excite Eu3 þ to its excited state 5D0.
Fig. 11. Possible energy transfer routes between Tb3 þ and Eu3 þ : The non-radiative relaxation energy from the Tb3 þ -related energy level 5D4 is used to excite Eu3 þ to its excited state 5D0. (For interpretation of the references to color in this figure caption, the reader is referred to the web version of this paper.)
Fig. 12. Emission (left) and excitation (right) spectra of Tb3 þ /Eu3 þ double-doped barium borate glass with a doping level of 0.5 at% each, recorded at (a) 10 K and (b) RT. The emission spectra were recorded under 370-nm excitation, the excitation spectra were detected at an emission wavelength of 613 nm. Additional excitation bands from the Eu3 þ -related energy level 7F1 are marked (asterisks).
temperatures, e.g. for 10 K, this ratio amounts to only N1/N0≈10 22, i.e. the energy level 7F1 is not populated. The energy transfer route with the best overlap between populated energy levels is the 5D4 to 7F5 emission from Tb3 þ and the 7F0 to 5D0 excitation from Eu3 þ (Fig. 11, blue arrows). The energy mismatch is now significantly higher than for RT and amounts to approx. 1000 cm 1. This results in a lower energy transfer rate, kET , leading to a lower Eu3 þ emission intensity at 10 K. The amount of this effect is clearly seen in the photoluminescence spectra. Fig. 12 shows the emission (left) and excitation spectra (right) of the Tb3 þ /Eu3 þ double-doped barium borate glass with a doping level of 0.5 at% each, recorded at 10 K and RT. Compared to the 10 K excitation spectrum, the spectrum recorded at RT features additional bands from population of the Eu3 þ -related energy level 7F1 (Fig. 12b, stars). The RT emission spectrum does not show any additional bands, but the intensity ratio changes: at 10 K the most intense emission band is the Tb3 þ -related 5D4 to 7F5 transition whereas at RT the Eu3 þ -related 5 D0 to 7F2 transition dominates the spectrum. This means, that at RT a higher energy transfer is observed than for low temperatures. 3.3. Luminescent glasses as LED phosphor Sm3 þ emits in the orange spectral range and thus represents an interesting rare-earth dopant for pc-LEDs with a poor colour rendering index. However, Sm3 þ exhibits a lower quantum efficiency than Eu3 þ and Tb3 þ and is therefore not a suitable phosphor for high-power LEDs. This section focuses on Eu3 þ and Tb3 þ doped glasses. Fig. 13 shows the colour coordinates of the Eu3 þ and Tb3 þ single-doped borate glasses (black dots) in the Commission Internationale de l'Éclairage (CIE) chromaticity diagram. The colour coordinates have been obtained from the corresponding emission spectra. Both RE ions are close to the edge of the CIE chromaticity diagram, i.e., they provide high colour saturation since the blue emissions are suppressed by the high phonon frequencies of the borate host. Compared to Eu3 þ , the Tb3 þ emission is slightly shifted to the blue because of the 5D4 to 7F6 transition at 490 nm. For the single-doped glasses, the emission spectra do not change with the RE doping level. The emission spectra of the double-doped glasses, however, depend significantly on the RE ion
776
F. Steudel et al. / Journal of Luminescence 170 (2016) 770–777
resulting in a green colour impression, while 380-nm excitation leads to a yellow one. Excitation at 385 nm and 390 nm results in an orange and red colour impression, respectively, due to the strong Eu3 þ absorption bands at these wavelengths. A colour change during use is thus possible. Changing the excitation wavelength, e.g. from 350 nm to 390 nm, the colour coordinate changes from green to red. Using different excitation wavelengths enables an immediate change in colour coordinate with only one phosphor. This method might be used in control lamps to attract attention or in lighting applications where a change in colour coordinate during use is wanted.
4. Conclusion Fig. 13. CIE colour space chromaticity diagram with point of equal energy, E, and the colour coordinates for the Eu3 þ and Tb3 þ single- (crosses) and double-doped borate glasses (full circles) under 380-nm excitation. The dotted lines indicate the complementary blue part. The inset shows four of the five glasses under UV illumination. (For interpretation of the references to color in this figure caption, the reader is referred to the web version of this paper.)
Sm3 þ , Eu3 þ , and Tb3 þ single- and double-doped borate glasses have been investigated. Intrinsic lifetimes amount to between 2 ms and 3 ms. Cross-relaxation and energy transfer processes have been observed in the single- and double-doped glass. For Tb3 þ /Eu3 þ double-doped glass the energy transfer is temperature-dependent due to thermal population of the Eu3 þ -related energy level 7F1. With high quantum efficiency values of more than 80%, the luminescent glass is suitable as phosphor for white LEDs. The rare-earth ions provide high colour saturation since their blue emission bands are suppressed by the high phonon frequency of the borate base glass. With the Tb3 þ emission in the green and the Eu3 þ emission in the red spectral range, colour mixing from green to red is enabled by varying the ratio of the RE doping level accordingly. In addition, different excitation wavelengths can be used to adjust the colour coordinate of the phosphor.
Acknowledgment
Fig. 14. Section of the CIE colour space chromaticity diagram with point of equal energy, E, and the colour coordinate of the 0.5 at% Tb3 þ /0.05 at% Eu3 þ doped borate glass for different excitation wavelengths (full circles) between the colour coordinate of the single-doped glasses (crosses). The FWHM of the excitation is approx. 16 nm, which is typical for LEDs. (For interpretation of the references to color in this figure caption, the reader is referred to the web version of this paper.)
concentration. The colour coordinate of the Tb3 þ /Eu3 þ doubledoped glass can be shifted from the green (Tb3 þ ) to the red (Eu3 þ ) spectral range by increasing the Eu3 þ concentration accordingly (Fig. 13, crosses). The dashed line connecting the colour coordinates of the Eu3 þ and Tb3 þ single-doped glasses indicates all possible colour coordinates of double-doped glasses. Note, that the emission spectrum of the equally double-doped glass (see 0.5 at% Tb3 þ /0.5 at% Eu3 þ ) is shifted to the red spectral range due to the significant energy transfer from Tb3 þ to Eu3 þ (see Section 3.2.4). The dotted lines going through the point of equal energy, E, indicate the complementary blue region. This method of colour mixing by varying the RE ratio is a conventional method, in particular used for white LED phosphors. However, the colour of the phosphor is then fixed and cannot be changed during use. The double-doped borate glass offers the opportunity to vary the colour coordinates upon changing the excitation wavelength accordingly. This is shown in Fig. 14 for the 0.5 at% Tb3 þ /0.05 at% Eu3 þ doped glass. The full width at half maximum (FWHM) of the excitation is approx. 16 nm, which is typical for LEDs. For excitation at 350 nm, the Tb3 þ emission dominates the spectrum
The Fraunhofer Application Center for Inorganic Phosphors in Soest is supported by the “Ministerium für Innovation, Wissenschaft und Forschung des Landes Nordrhein-Westfalen”.
References [1] H. Lin, E.Y.-B. Pun, X. Wang, X. Liu, J. Alloys Compd. 390 (1–2) (2005) 197. [2] N. Soga, K. Hirao, M. Yoshimoto, H. Yamamoto, J. Appl. Phys. 63 (9) (1988) 4451. [3] J. Hu, X. Gong, Y. Chen, J. Huang, Y. Lin, Z. Luo, Y. Huang, Opt. Mater. 38 (0) (2014) 108. [4] T.Y. Lim, H. Wagiran, R. Hussin, S. Hashim, M. Saeed, Physica B: Condens. Matter 451 (0) (2014) 63. [5] M. Mhareb, S. Hashim, S. Ghoshal, Y. Alajerami, M. Saleh, R. Dawaud, N. Razak, S. Azizan, Opt. Mater. 37 (0) (2014) 391. [6] J. Pisarska, A. Kos, W.A. Pisarski, Spectrochim. Acta Part A: Mol. Biomol. Spectrosc. 129 (0) (2014) 649. [7] K. Swapna, S. Mahamuda, A.S. Rao, M. Jayasimhadri, S. Shakya, G.V. Prakash, J. Lumin. 156 (0) (2014) 180. [8] H. Xiong, L. Shen, E. Pun, H. Lin, J. Lumin. 153 (0) (2014) 227. [9] L.G.V. Uitert, R.R. Soden, J. Chem. Phys. 36 (5) (1962) 1289. [10] P. Schlotter, R. Schmidt, J. Schneider, Appl. Phys. A: Mater. Sci. Process. 64 (4) (1997) 417. [11] W. Müller-Warmuth, H. Eckert, Phys. Rep. 88 (2) (1982) 91. [12] S. Kapoor, H.B. George, A. Betzen, M. Affatigato, S. Feller, J. Non-Cryst. Solids 270 (1–3) (2000) 215. [13] G.H. Dieke, H.M. Crosswhite, Appl. Opt. 2 (7) (1963) 675. [14] K. Binnemans, R.V. Deun, C. Görller-Walrand, J.L. Adam, J. Non-Cryst. Solids 238 (1998) 11. [15] M. Shojiya, M. Takahashi, R. Kanno, Y. Kawamoto, K. Kadono, J. Appl. Phys. 82 (12) (1997) 6259. [16] G. Blasse, B.C. Grabmaier, Luminescent Materials, Springer-Verlag, Berlin, Heidelberg, 1994. [17] G.M. Kumar, B.N.S. Bhaktha, D.N. Rao, Opt. Mater. 28 (11) (2006) 1266. [18] T. Förster, Ann. Phys. 437 (1–2) (1948) 55. [19] F. Steudel, S. Loos, B. Ahrens, S. Schweizer, Proc. SPIE 9140 (2014) 91400I.
F. Steudel et al. / Journal of Luminescence 170 (2016) 770–777
[20] D. Umamaheswari, B.C. Jamalaiah, T. Sasikala, T. Chengaiah, I.-G. Kim, L.R. Moorthy, J. Lumin. 132 (5) (2012) 1166. [21] J.R. Lakowicz, Principles of Fluorescence Spectroscopy, Springer Science þ Business Media, LLC, 2006. [22] R. Praveena, V. Venkatramu, P. Babu, C. Jayasankar, Physica B: Condens. Matter 403 (19–20) (2008) 3527. [23] G. Blasse, J. Lumin. 1–2 (0) (1970) 766.
777
[24] G. Lakshminarayana, E.M. Weis, A. Lira, U. Caldi no, D.J. Williams, M.P. Hehlen, J. Lumin. 139 (2013) 132. [25] H. Lin, D.L. Yang, G.S. Liu, T.C. Ma, B. Zhai, Q.D. An, J.Y. Yu, X.J. Wang, X.R. Liu, E.Y.B. Pun, J. Lumin. 113 (1–2) (2005) 121.
Contents lists available at ScienceDirect
Journal of Luminescence journal homepage: www.elsevier.com/locate/jlumin
Quantum efficiency and energy transfer processes in rare-earth doped borate glass for solid-state lighting Franziska Steudel a,n, Sebastian Loos b, Bernd Ahrens a,b, Stefan Schweizer a,b a Fraunhofer Application Center for Inorganic Phosphors, Branch Lab of Fraunhofer Institute for Mechanics of Materials IWM, Lübecker Ring 2, 59494 Soest, Germany b Department of Electrical Engineering, South Westphalia University of Applied Sciences, Lübecker Ring 2, 59494 Soest, Germany
art ic l e i nf o
a b s t r a c t
Article history: Received 7 April 2015 Accepted 21 July 2015 Available online 4 August 2015
Sm3 þ , Eu3 þ and Tb3 þ doped borate glass is investigated for its potential as light converting phosphor for solid-state lighting applications. Concentration-dependent luminescence, quantum efficiency and radiative lifetimes are analysed. The luminescence quantum efficiency exceeds values of more than 80%. Cross-relaxation processes in the single-doped glasses result in luminescence quenching for Sm3 þ with increasing doping level, whereas for Eu3 þ and Tb3 þ an increase is observed. In Sm3 þ /Eu3 þ and Tb3 þ /Eu3 þ double-doped glass the energy transfer processes between the rare-earth ions are investigated in detail. The colour coordinate of the Tb3 þ /Eu3 þ double-doped glass can be shifted from the green to the red spectral range by changing the Tb3 þ -to-Eu3 þ ratio in favour of Eu3 þ . In addition, double doping allows for a change in colour coordinate by using different excitation wavelengths. & 2015 Elsevier B.V. All rights reserved.
Keywords: Photon conversion Rare-earth ion doping Energy transfer LED Colour management
1. Introduction Glass is very versatile and a good host for rare-earth (RE) ions; it provides high optical transparency, good RE ion solubility, and it can be cast in almost any shape or size. Luminescent glasses have attracted much attention in the last decades, in particular for lasers, optical fibres, and optical amplifiers [1]. Borate glass is a suitable optical material with high mechanical, chemical and thermal stability [1,2]. A widespread interest in borate glass is recorded by multiple publications on spectroscopy of RE ions in borate glasses in 2014 [3–8]. For many applications, in particular for high-efficiency devices, the knowledge of the radiative quantum efficiency and the radiative lifetime of the luminescent material is of decisive importance. For solid-state lighting applications, e.g. white lightemitting diodes (LEDs), the colour impression is of great significance and appropriate doping with multiple rare-earth ions becomes important. In particular, the rare-earth ions Eu3 þ and Tb3 þ gained a great technological relevance. Their red and green luminescence is used in cathode ray tubes, fluorescent lamps, and plasma displays and have therefore been intensively studied. In addition, Tb3 þ was found to be a good sensitizer to enhance luminescence efficiency of Eu3 þ via energy transfer [9]. n
Corresponding author. E-mail address: [email protected] (F. Steudel).
http://dx.doi.org/10.1016/j.jlumin.2015.07.032 0022-2313/& 2015 Elsevier B.V. All rights reserved.
In the last years, Eu3 þ and Tb3 þ owned increased interest as a phosphor for white light-emitting diodes. White LEDs are replacing conventional light sources due to their high efficiency, compactness, long operational lifetime, and resultant energy saving. Traditional white LEDs combine a blue LED chip and a yellow YAG:Ce3 þ phosphor [10]. However, this method has a low colourrendering index (CRI) and high colour temperature due to a lack of red emission. For higher CRI conventional phosphors can be double-doped with a red emitter, i.e. Sm3 þ or Eu3 þ or both. Another possibility for white light generation is the combination of a blue LED chip with a Eu3 þ and Tb3 þ doped phosphor. In this work, the optical properties of Sm3 þ , Eu3 þ , and Tb3 þ single-doped barium borate glasses are analysed and crossrelaxation processes between the ions are investigated. Furthermore, the influence of Sm3 þ /Eu3 þ and Tb3 þ /Eu3 þ double-doping and the energy transfer processes between the two different rareearth ions are analysed. Both effects play an essential role in optimization of the luminescent glasses for their implementation in phosphor-converted LEDs (pc-LEDs).
2. Experimental details 2.1. Sample preparation Borate glasses using barium oxide as network modifier were prepared. A ratio of two moles of boron oxide (B2O3) and one mole of barium oxide (BaO) was used. In this ratio the glass network
F. Steudel et al. / Journal of Luminescence 170 (2016) 770–777
771
Table 1 Nominal composition of the double-doped samples under study. Dopant
Composition (mol%)
RE-content (at%)
B2O3
BaO
Sm2O3
Eu2O3
Tb2O3
Sm3 þ /Eu3 þ
65.67 65.33 64.67
32.83 32.67 32.33
1.0 1.0 1.0
0.5 1.0 2.0
– – –
0.50/0.25 / – 0.50/0.50 / – 0.50/1.00 / –
Tb3 þ /Eu3 þ
65.93 65.67 65.33
32.97 32.83 32.67
– – –
0.1 0.5 1.0
1.0 1.0 1.0
–/0.05/0.50 –/0.25/0.50 –/0.50/0.50
consists of the highest possible amount of four-coordinated boron [11]. The glasses were additionally doped with different doping levels of samarium oxide (Sm2O3), europium oxide (Eu2O3), or terbium oxide (Tb2O3). The single-doped glasses are comprised of (66.67 0.67x)B2O3 (33.33 0.33x)BaO xRE2O3 with x ¼0.1, 0.3, 0.5, 1.0, 2.0, 3.0 and 5.0 mol% which equals 0.05–2.5 at%. Table 1 summarizes the nominal chemical composition of the doubledoped glasses. The chemicals were weighed in a platinum gold crucible (Pt/Au 95/5) and melted at 1100 °C for approximately 3 h. The melt was then poured onto a brass block at 500 °C, which is below the glass transition temperature of Tg ¼605 °C for barium borate glass [12]. The glass was kept at this temperature for 3 h to eliminate residual mechanical and thermal stresses before being slowly cooled to room temperature. The glass was then cut into squares of 20 mm 20 mm with a thickness of 3.2 mm and polished to optical quality.
Fig. 1. Energy level diagram of (a) Sm3 þ , (b) Eu3 þ , and (c) Tb3 þ . Possible excitation (arrows from bottom to top) and emission routes (arrows from top to bottom) are indicated [14].
2.2. Experimental set-up Absolute photoluminescence quantum efficiency (QE) measurements were performed with a commercial quantum yields measurement system (Hamamatsu C9920-02G) coupled to a 3.3 inch integrating sphere with a xenon lamp (150 W) as excitation source and a photonic multichannel analyser (PMA 12) as detector. The setup has a measurement accuracy of approx. 3%. The quantum efficiency was determined from emission spectra in the spectral range from 450 to 900 nm. Radiative lifetime measurements were performed under excitation with a temperature-stabilized 405-nm laser-diode (Sanyo DL-3146-151) in case of Sm3 þ single- and Sm3 þ /Eu3 þ double-doped glasses and a 370-nm ultraviolet (UV) LED (Winger WEEUV00-CS) in case of Eu3 þ and Tb3 þ single- and double-doped glasses. A peltier-cooled photomultiplier (EMI 9863/492) coupled to a 300 mm focal length monochromator (Princeton Instruments Acton 2300) was used for detection.
3. Results and discussion 3.1. Single-doped glasses 3.1.1. Photoluminescence Fig. 1 shows the energy level diagrams of (a) Sm3 þ , (b) Eu3 þ , and (c) Tb3 þ [13,14]. Possible excitation (arrows from bottom to top) and emission routes (arrows from top to bottom) are indicated. For all three RE ions, emissions in the blue spectral range (from higher energy levels) are quenched due to the high maximum phonon frequency of 1400 cm 1 in borate glass [15]. Though the spectral positions of rare-earth ion emission are well known, the transition intensity ratios vary due to differences in crystallinity and phonon frequencies of the host material.
Fig. 2. Normalized emission spectra of borate glass doped with (a) Sm3 þ , (b) Eu3 þ and (c) Tb3 þ . The excitation was carried out at 402 nm (Sm3 þ ), 393 nm (Eu3 þ ), and 370 nm (Tb3 þ ). The emission transitions are indicated.
The emission spectra of Sm3 þ , Eu3 þ and Tb3 þ single-doped borate glasses are shown in Fig. 2. For Sm3 þ doping (Fig. 2a), the emission spectra show transitions from the excited state4G5/2 to the ground state levels 6H5/2, 6H7/2, 6H9/2, and 6H11/2 (550– 700 nm); the emission is excited at 402 nm. Upon excitation at 393 nm, the Eu3 þ -doped glass (Fig. 2b) shows the typical Eu3 þ emissions in the red spectral range, which are caused by transitions from the excited state 5D0 to the ground state levels 7F0 (580 nm), 7F1 (592 nm), 7F2 (613 nm), 7F3 (652 nm), and 7F4 (700 nm). The electric-dipole transition 5D0 to 7F2 is hypersensitive to variations in crystal symmetry [16]. The high intensity of this transition in borate glass indicates the amorphous nature of the matrix material with low inversion symmetry for the Eu3 þ ion. In Fig. 2c, the typical Tb3 þ -related emissions at 490 nm, 543 nm,
772
F. Steudel et al. / Journal of Luminescence 170 (2016) 770–777
concentration of 0.05 at% shows a single exponential radiative decay curve due to a low probability of dipole–dipole interaction for the large mean distance between the Sm3 þ ions of 4.5 nm. The radiative decay curves of the Eu3 þ and Tb3 þ single-doped samples are not affected by the rare-earth doping level. For comparison, Fig. 3b gives an overview on the decay behaviour of Eu3 þ and Tb3 þ for a doping level of 0.5 at% each. The poor signal-tonoise ratio in case of the Tb3 þ decay is due to the relatively small absorption coefficient of Tb3 þ [19]. The decay of Eu3 þ and Tb3 þ is single exponential and can be described by
⎛ t ⎞ I (τ ) = I0·exp ⎜ − ⎟. ⎝ τ0 ⎠
Fig. 3. Normalized radiative decay curves of (a) Sm3 þ in barium borate glass with a Sm3 þ doping level from 0.05 to 1 at%, (b) Eu3 þ and Tb3 þ doped barium borate glass with a doping level of 0.5 at% each. The analysed transitions are 4G5/2 to 6H7/2 (Sm3 þ ), 5D0 to 7F2 (Eu3 þ ), and 5D4 to 7F4 (Tb3 þ ). The solid curves represent a best fit on the basis of the Förster model (Sm3 þ ) and a single exponential decay (Tb3 þ , Eu3 þ ). (For interpretation of the references to color in this figure caption, the reader is referred to the web version of this paper.)
583 nm, and 622 nm can be assigned to transitions from the excited state 5D4 to the ground state levels 7FJ (J¼ 6, 5, 4, and 3), respectively. 3.1.2. Lifetime The luminescence decay curves for the transition 4G5/2 to 6H7/2 of Sm3 þ are depicted in Fig. 3a for different doping levels. The lifetime decreases with increasing Sm3 þ doping level and the decay curves show a non-single exponential behaviour due to superposition of radiative and non-radiative relaxation processes. In general, the experimental lifetime, τexp, of an excited level a can be described by
1 = τexp
∑ (A ab + Wab ) b
γ=
π 3/2·cA·R 03
3.1.3. Quantum efficiency Fig. 4 shows the spectrally-resolved photoluminescence quantum efficiency for the 0.5 at% doped samples. The quantum efficiency shows a spectral dependence; the maximum values are collected in Tables 2 and 3. In all cases, the quantum efficiency depends significantly on the rare-earth doping level. For the Sm3 þ doped glass (Fig. 4a), the quantum efficiency decreases significantly with increasing Sm3 þ doping (“luminescence quenching”) from approx. 41% for a concentration of 0.05 at% to 2% for 1 at% doping. The Eu3 þ single-doped borate glass (Fig. 4b) shows a different behaviour: Here, the luminescence quantum efficiency increases from 43% for 0.05 at% to 85% for 1 at% (Table 3). The luminescence quantum efficiency of the Tb3 þ single-doped glass (Fig. 4c) also increases with the doping level from 11% to 75% (Table 3), showing saturation for concentrations higher than 1 at%. 3.1.4. Cross-relaxation The observed changes in luminescence quantum efficiency and luminescence lifetime (Tables 2 and 3) result from cross-relaxation Table 2 Concentration of Sm3 þ ions, c Sm3+ , quantum efficiency, QE, at 402 nm and parameters obtained from an analysis based on the Förster model for different Sm3 þ doping levels. Doping level (at%)
c Sm3+ /1018 cm 3 QE (%)
τexp (ms)
τ0 (ms)
γ
R0 (nm)
0.05 0.15 0.25 0.50 1.00
21.6 62.3 107.1 204.3 426.5
2.41 1.95 1.58 0.97 0.39
2.70 2.53 2.73 2.44 2.47
0.09 0.25 0.62 1.14 2.68
0.66 0.65 0.73 0.72 0.75
40.7 31.1 18.8 9.6 1.9
(2)
where I(t) is the intensity of the radiative decay, I0 the initial intensity, t the time after the excitation pulse, and τ0 the intrinsic lifetime of the donor in the absence of an acceptor. The energy transfer parameter, γ, is defined by 4 3
The obtained intrinsic lifetimes, τ0, are collected in Table 3 and amount to approx. 2.0 ms and 2.5 ms for Eu3 þ and Tb3 þ doping, respectively. These values are comparable with those found in the literature [20].
(1)
with Aab and Wab the radiative and non-radiative transition probabilities from level a to level b. Non-radiative decays include multiphonon relaxation and ion-ion interactions, such as energy transfer and cross-relaxation. For Sm3 þ , the non-single exponential behaviour of the decay curves is caused by non-radiative energy transfer processes between neighbouring Sm3 þ ions arising from dipole–dipole interaction [17]. It can be described by the so-called Förster model [18]
⎛ t t ⎞ I (t ) = I0·exp ⎜ − − γ · ⎟ τ0 ⎠ ⎝ τ0
(4)
(3)
with cA the concentration of acceptors and R0 the critical distance. The results obtained from a best fit on the basis of the abovedescribed Förster model are shown as red solid curves in Fig. 3a; the corresponding fitting parameters are listed in Table 2. The experimental lifetimes, τexp, decrease from 2.6 to 0.4 ms with increasing doping level, whereas the energy transfer parameter, γ, increases from 0.1 to 2.7. The intrinsic lifetime, τ0, is approx. 2.6 ms and the critical distance approx. 0.7 nm. The sample with a Sm3 þ
Table 3 Quantum efficiencies, QE, at 464 nm (Eu3 þ ) and 484 nm (Tb3 þ ) and intrinsic lifetimes, τ0, for different Eu3 þ and Tb3 þ concentrations. Eu3 þ c (at%) 0.05 0.15 0.25 0.50 1.00 1.50 2.50 a
Tb3 þ
QE (%)
τ0 (ms)
QE (%)
τ0 (ms)
43 50 59 60 84a 81a 85a
2.10 2.02 1.95 2.00 2.03 2.02 1.98
11 25 45 63 75 75 70
2.55 2.54 2.57 2.55 2.50 2.48 2.41
Different Eu2O3 raw material.
F. Steudel et al. / Journal of Luminescence 170 (2016) 770–777
Fig. 4. Quantum efficiency spectra of (a) Sm3 þ , (b) Eu3 þ , and (c) Tb3 þ singledoped barium borate glass with a doping level of 0.5 at% each. The most intense transitions are indicated.
773
Fig. 6. Photoluminescence emission spectra of (a) 0.5 at% Sm3 þ and (b) 0.5 at% Eu3 þ single-doped barium borate glass as well as (c) Sm3 þ /Eu3 þ double-doped barium borate glass with a doping level of 0.5 at% each. The emission spectra were recorded upon 480-nm excitation. 5 D4 increases with increasing concentration enabling higher luminescence quantum efficiency values for higher concentrations.
3.2. Double-doped glasses
Fig. 5. Cross-relaxation channels of (a) Sm3 þ , (b) Eu3 þ , and (c) Tb3 þ [22–24].
processes, i.e. non-radiative energy transfer processes between neighbouring Sm3 þ , Eu3 þ , or Tb3 þ ions arising from dipole–dipole interaction. The cross-relaxation rate, kCR, depends significantly on the distance, r, between the interacting ions, namely kCR∼r 6 [21]. Possible cross-relaxation channels are depicted in Fig. 5. Sm3 þ has four cross-relaxation channels (Fig. 5a), all starting from the emission level 4G5/2 [17,22] leading to a depopulation of the emission level and thus a decrease in luminescence quantum efficiency and luminescence lifetime with increasing Sm3 þ concentration. Eu3 þ and Tb3 þ have two cross-relaxation channels each (Fig. 5b and c). Since the cross-relaxation rate depends on the frequency, i.e. kCR∼ν 4 [21], Eu3 þ 's low-energy cross-relaxation channel (5D1, 7F0) → (5D0, 7F3) is favoured over its higher energy cross-relaxation channel (5D2, 7F1) → (7F4, 5D1). This results in a higher population of the emission level 5D0 and thus to increased emission intensities. For Tb3 þ , multipolar transfer interactions are possible due to an equal energy gap of approx. 0.6 eV between the energy levels 5D3 and 5D4 as well as 7F6 and 7F0 enabling a relaxation from 5D3 to 5D4. The population of the emission level
3.2.1. Photoluminescence The photoluminescence emission spectra of the double-doped glasses show the same emission lines as the single-doped glasses but with different intensity ratios: The Sm3 þ /Eu3 þ double-doped glass (Fig. 6c) is not the sum of the corresponding spectra of the singledoped glasses (Fig. 6a and b). For 480-nm excitation, Sm3 þ (a) has a strong absorption band due to a transition from the ground state 6 H5/2 to the excited state 4G7/2. The typical Sm3 þ -related emissions are strong in intensity. For Eu3 þ (b) the situation is different: at 480 nm, the emission is low in intensity (Fig. 6b) due to the small Eu3 þ absorption coefficient at this wavelength. For the double-doped sample (c) the emission spectrum exhibits as expected Sm3 þ emissions, but also intense Eu3 þ -related emissions. Appropriate scaling of the spectra from the single-doped glasses (so that the sum of both fits the spectrum of the double-doped glass) reveals an increase in Eu3 þ emission by a factor of 16, whereas the Sm3 þ intensity is scaled by a factor of 0.62. The same behaviour is observed for Tb3 þ /Eu3 þ double-doping (Fig. 7): upon 484-nm excitation, the intensity of the Eu3 þ emission is increased by a factor of 7.7 while the Tb3 þ emission is only about one third of its intensity in single-doped glass. 3.2.2. Lifetime The decay curves of the Sm3 þ /Eu3 þ double-doped samples (Fig. 8a) after excitation at 405 nm (Sm3 þ -related 6H5/2 to 6P5/2 transition) and detection at 598 nm (Sm3 þ -related 4G5/2 to 6H7/2 transition) show a non-single exponential behaviour similar to the Sm3 þ single-doped samples (Fig. 3a). To minimize Sm3 þ crossrelaxation effects, the Sm3 þ concentration is kept constant and only the Eu3 þ concentration is varied. However, increasing the Eu3 þ doping level does not significantly affect the decay behaviour. The parameters obtained from the Förster model are
774
F. Steudel et al. / Journal of Luminescence 170 (2016) 770–777
are summarized in Table 4. For constant Tb3 þ concentration and increasing Eu3 þ doping level, the energy transfer parameter γ increases from zero to 0.24 while the experimental lifetime decreases from 2.55 ms for Tb3 þ single-doped to 1.65 ms for an additional doping with 0.5 at% Eu3 þ . The critical distance, R0, decreases with higher Eu3 þ concentration.
Fig. 7. Photoluminescence emission spectra of (a) 0.5 at% Tb3 þ and (b) 0.5 at% Eu3 þ single-doped barium borate glass as well as (c) Tb3 þ /Eu3 þ double-doped barium borate glass with a doping level of 0.5 at% each. The emission spectra were recorded upon 480-nm excitation.
3.2.3. Quantum efficiency The absolute photoluminescence quantum efficiency spectra of the double-doped samples with 0.5 at% doping level each are shown in Fig. 9 (solid curves): For Sm3 þ /Eu3 þ doping (Fig. 9a) the quantum efficiency of the double-doped glass is significantly lower than the sum of both single-doped glasses (dashed) due to cross-relaxation processes between neighbouring Sm3 þ ions. The obtained photoluminescence quantum efficiency values are summarized in Table 4 for 480-nm excitation. For constant Sm3 þ concentration and increasing Eu3 þ doping, the efficiency increases slightly from 10% to 12% which is less than found for Eu3 þ singledoped glasses. The quantum efficiency of the Tb3 þ /Eu3 þ double-doped glass (Fig. 9b, solid) is similar to the sum of both single-doped glasses (dashed) and reaches values higher than 80%. Only for wavelengths with high Tb3 þ absorption, i.e. 484 nm and 380 nm, the sum of the quantum efficiency spectra of both single-doped glasses is higher, since at these wavelengths, re-absorption decreases the quantum efficiency. For 484-nm excitation, the quantum Table 4 Parameters obtained from Förster-model analysis for different Sm3 þ /Eu3 þ and Tb3 þ /Eu3 þ doping levels and quantum efficiency values at 480 nm (Sm3 þ /Eu3 þ ) and at 484 nm (Tb3 þ /Eu3 þ ). Donor
RE content (at%)
τexp
τ0 (ms)
γ
R0 (nm)
QE (%)
(ms) cdonor
Eu3 þ
Sm3 þ
0.5 0.5 0.5 0.5
0 0.25 0.5 1.0
0.97 1.03 1.12 1.16
2.44 2.72 2.55 2.35
0 0.56 0.46 0.40
– 0.903 0.669 0.512
10 10 11 12
Tb3 þ
0.5 0.5 0.5 0.5
0 0.05 0.25 0.50
2.55 2.45 2.30 1.65
2.55 2.70 2.56 2.25
0 0.11 0.17 0.24
– 0.70 0.48 0.42
63 63 62 62
Fig. 8. Normalized radiative decay of (a) Sm3 þ /Eu3 þ double-doped barium borate glass after excitation at 405 nm and detection at 598 nm (Sm3 þ -related 4G5/2 to 6 H7/2 transition) and (b) Tb3 þ /Eu3 þ double-doped barium borate glass after excitation at 370 nm and detection at 543 nm (Tb3 þ -related 5D4 to 7F4 transition) and 613 nm (Eu3 þ -related 5D0 to 7F2 transition). The red curves result from Förster model (Sm3 þ /Eu3 þ and Tb3 þ /Eu3 þ at 543 nm) and single exponential decay analyses (Tb3 þ /Eu3 þ at 613 nm).
summarized in Table 4. The radiative lifetime is approx. 1 ms and increases slightly upon increasing the Eu3 þ doping level. The intrinsic lifetime, τ0, is constant at approx. 2.5 ms, the energy transfer parameter, γ, decreases with increasing Eu3 þ concentration due to direct excitation of Eu3 þ at 405 nm. For Tb3 þ /Eu3 þ double-doping (Fig. 8b), the decay curve after excitation at 370 nm (Tb3 þ -related 7F6 to 5G6 transition) and detection at 613 nm (Eu3 þ -related 5D0 to 7F2 transition) is single exponential and does not indicate any energy transfer processes from Eu3 þ to other ions; the lifetime is 2.1 ms, which is nearly the lifetime for Eu3 þ single-doping. However, for detection at 543 nm (Tb3 þ -related 5D4 to 7F5 transition) the radiative decay cannot be described by a simple single exponential decay any longer, so that an analysis based on the Förster model is applied. The parameters
Fig. 9. Experimental quantum efficiency of the double-doped glasses (solid) and sum of the corresponding single-doped glasses (dashed) of (a) Sm3 þ /Eu3 þ and (b) Tb3 þ /Eu3 þ doped glass with a doping level of 0.5 at% each.
F. Steudel et al. / Journal of Luminescence 170 (2016) 770–777
775
efficiency values of the Tb3 þ /Eu3 þ double-doped glass with 0.5 at% Tb3 þ and various Eu3 þ doping levels are collected in Table 4. At this wavelength, the quantum efficiency is not affected by the Eu3 þ concentration. 3.2.4. Energy transfer The photoluminescence spectra of the double-doped glasses show a clear energy transfer from Sm3 þ and Tb3 þ to Eu3 þ . The energy transfer route with the highest energy transfer rate, kET , for Sm3 þ /Eu3 þ double-doping is shown in Fig. 10: from the Sm3 þ -related emission level 4G5/2 the energy is released by relaxation to the ground state 6H5/2 and partially transferred to Eu3 þ , which is excited from its ground state to the excited state 5D0 by resonance between the two transitions [25]. The energy mismatch between the two transitions amount to 700 cm 1 (86 meV), which is the best overlap between Sm3 þ emission and Eu3 þ excitation. The energy transfer route for Tb3 þ /Eu3 þ double-doping with the best overlap between Tb3 þ emission and Eu3 þ excitation is shown in Fig. 11, black arrows. It is found for the 5D4 to 7F4 emission from Tb3 þ and the 7F1 to 5D0 excitation from Eu3 þ . Here, the energy mismatch amounts to only 300 cm 1. The population of the Eu3 þ -related energy level 7F1 is possible at room temperature (RT) due to thermal population according to the Boltzmann distribution. Here, the ratio of the population probabilities of the energy levels 7F1 and 7F0 amounts to N1/N0=0.56, which means that 35% of the electrons occupy 7F1. For low
Fig. 10. Energy transfer route between Sm3 þ and Eu3 þ : the non-radiative relaxation energy from the Sm3 þ -related4G5/2 level is used to excite Eu3 þ to its excited state 5D0.
Fig. 11. Possible energy transfer routes between Tb3 þ and Eu3 þ : The non-radiative relaxation energy from the Tb3 þ -related energy level 5D4 is used to excite Eu3 þ to its excited state 5D0. (For interpretation of the references to color in this figure caption, the reader is referred to the web version of this paper.)
Fig. 12. Emission (left) and excitation (right) spectra of Tb3 þ /Eu3 þ double-doped barium borate glass with a doping level of 0.5 at% each, recorded at (a) 10 K and (b) RT. The emission spectra were recorded under 370-nm excitation, the excitation spectra were detected at an emission wavelength of 613 nm. Additional excitation bands from the Eu3 þ -related energy level 7F1 are marked (asterisks).
temperatures, e.g. for 10 K, this ratio amounts to only N1/N0≈10 22, i.e. the energy level 7F1 is not populated. The energy transfer route with the best overlap between populated energy levels is the 5D4 to 7F5 emission from Tb3 þ and the 7F0 to 5D0 excitation from Eu3 þ (Fig. 11, blue arrows). The energy mismatch is now significantly higher than for RT and amounts to approx. 1000 cm 1. This results in a lower energy transfer rate, kET , leading to a lower Eu3 þ emission intensity at 10 K. The amount of this effect is clearly seen in the photoluminescence spectra. Fig. 12 shows the emission (left) and excitation spectra (right) of the Tb3 þ /Eu3 þ double-doped barium borate glass with a doping level of 0.5 at% each, recorded at 10 K and RT. Compared to the 10 K excitation spectrum, the spectrum recorded at RT features additional bands from population of the Eu3 þ -related energy level 7F1 (Fig. 12b, stars). The RT emission spectrum does not show any additional bands, but the intensity ratio changes: at 10 K the most intense emission band is the Tb3 þ -related 5D4 to 7F5 transition whereas at RT the Eu3 þ -related 5 D0 to 7F2 transition dominates the spectrum. This means, that at RT a higher energy transfer is observed than for low temperatures. 3.3. Luminescent glasses as LED phosphor Sm3 þ emits in the orange spectral range and thus represents an interesting rare-earth dopant for pc-LEDs with a poor colour rendering index. However, Sm3 þ exhibits a lower quantum efficiency than Eu3 þ and Tb3 þ and is therefore not a suitable phosphor for high-power LEDs. This section focuses on Eu3 þ and Tb3 þ doped glasses. Fig. 13 shows the colour coordinates of the Eu3 þ and Tb3 þ single-doped borate glasses (black dots) in the Commission Internationale de l'Éclairage (CIE) chromaticity diagram. The colour coordinates have been obtained from the corresponding emission spectra. Both RE ions are close to the edge of the CIE chromaticity diagram, i.e., they provide high colour saturation since the blue emissions are suppressed by the high phonon frequencies of the borate host. Compared to Eu3 þ , the Tb3 þ emission is slightly shifted to the blue because of the 5D4 to 7F6 transition at 490 nm. For the single-doped glasses, the emission spectra do not change with the RE doping level. The emission spectra of the double-doped glasses, however, depend significantly on the RE ion
776
F. Steudel et al. / Journal of Luminescence 170 (2016) 770–777
resulting in a green colour impression, while 380-nm excitation leads to a yellow one. Excitation at 385 nm and 390 nm results in an orange and red colour impression, respectively, due to the strong Eu3 þ absorption bands at these wavelengths. A colour change during use is thus possible. Changing the excitation wavelength, e.g. from 350 nm to 390 nm, the colour coordinate changes from green to red. Using different excitation wavelengths enables an immediate change in colour coordinate with only one phosphor. This method might be used in control lamps to attract attention or in lighting applications where a change in colour coordinate during use is wanted.
4. Conclusion Fig. 13. CIE colour space chromaticity diagram with point of equal energy, E, and the colour coordinates for the Eu3 þ and Tb3 þ single- (crosses) and double-doped borate glasses (full circles) under 380-nm excitation. The dotted lines indicate the complementary blue part. The inset shows four of the five glasses under UV illumination. (For interpretation of the references to color in this figure caption, the reader is referred to the web version of this paper.)
Sm3 þ , Eu3 þ , and Tb3 þ single- and double-doped borate glasses have been investigated. Intrinsic lifetimes amount to between 2 ms and 3 ms. Cross-relaxation and energy transfer processes have been observed in the single- and double-doped glass. For Tb3 þ /Eu3 þ double-doped glass the energy transfer is temperature-dependent due to thermal population of the Eu3 þ -related energy level 7F1. With high quantum efficiency values of more than 80%, the luminescent glass is suitable as phosphor for white LEDs. The rare-earth ions provide high colour saturation since their blue emission bands are suppressed by the high phonon frequency of the borate base glass. With the Tb3 þ emission in the green and the Eu3 þ emission in the red spectral range, colour mixing from green to red is enabled by varying the ratio of the RE doping level accordingly. In addition, different excitation wavelengths can be used to adjust the colour coordinate of the phosphor.
Acknowledgment
Fig. 14. Section of the CIE colour space chromaticity diagram with point of equal energy, E, and the colour coordinate of the 0.5 at% Tb3 þ /0.05 at% Eu3 þ doped borate glass for different excitation wavelengths (full circles) between the colour coordinate of the single-doped glasses (crosses). The FWHM of the excitation is approx. 16 nm, which is typical for LEDs. (For interpretation of the references to color in this figure caption, the reader is referred to the web version of this paper.)
concentration. The colour coordinate of the Tb3 þ /Eu3 þ doubledoped glass can be shifted from the green (Tb3 þ ) to the red (Eu3 þ ) spectral range by increasing the Eu3 þ concentration accordingly (Fig. 13, crosses). The dashed line connecting the colour coordinates of the Eu3 þ and Tb3 þ single-doped glasses indicates all possible colour coordinates of double-doped glasses. Note, that the emission spectrum of the equally double-doped glass (see 0.5 at% Tb3 þ /0.5 at% Eu3 þ ) is shifted to the red spectral range due to the significant energy transfer from Tb3 þ to Eu3 þ (see Section 3.2.4). The dotted lines going through the point of equal energy, E, indicate the complementary blue region. This method of colour mixing by varying the RE ratio is a conventional method, in particular used for white LED phosphors. However, the colour of the phosphor is then fixed and cannot be changed during use. The double-doped borate glass offers the opportunity to vary the colour coordinates upon changing the excitation wavelength accordingly. This is shown in Fig. 14 for the 0.5 at% Tb3 þ /0.05 at% Eu3 þ doped glass. The full width at half maximum (FWHM) of the excitation is approx. 16 nm, which is typical for LEDs. For excitation at 350 nm, the Tb3 þ emission dominates the spectrum
The Fraunhofer Application Center for Inorganic Phosphors in Soest is supported by the “Ministerium für Innovation, Wissenschaft und Forschung des Landes Nordrhein-Westfalen”.
References [1] H. Lin, E.Y.-B. Pun, X. Wang, X. Liu, J. Alloys Compd. 390 (1–2) (2005) 197. [2] N. Soga, K. Hirao, M. Yoshimoto, H. Yamamoto, J. Appl. Phys. 63 (9) (1988) 4451. [3] J. Hu, X. Gong, Y. Chen, J. Huang, Y. Lin, Z. Luo, Y. Huang, Opt. Mater. 38 (0) (2014) 108. [4] T.Y. Lim, H. Wagiran, R. Hussin, S. Hashim, M. Saeed, Physica B: Condens. Matter 451 (0) (2014) 63. [5] M. Mhareb, S. Hashim, S. Ghoshal, Y. Alajerami, M. Saleh, R. Dawaud, N. Razak, S. Azizan, Opt. Mater. 37 (0) (2014) 391. [6] J. Pisarska, A. Kos, W.A. Pisarski, Spectrochim. Acta Part A: Mol. Biomol. Spectrosc. 129 (0) (2014) 649. [7] K. Swapna, S. Mahamuda, A.S. Rao, M. Jayasimhadri, S. Shakya, G.V. Prakash, J. Lumin. 156 (0) (2014) 180. [8] H. Xiong, L. Shen, E. Pun, H. Lin, J. Lumin. 153 (0) (2014) 227. [9] L.G.V. Uitert, R.R. Soden, J. Chem. Phys. 36 (5) (1962) 1289. [10] P. Schlotter, R. Schmidt, J. Schneider, Appl. Phys. A: Mater. Sci. Process. 64 (4) (1997) 417. [11] W. Müller-Warmuth, H. Eckert, Phys. Rep. 88 (2) (1982) 91. [12] S. Kapoor, H.B. George, A. Betzen, M. Affatigato, S. Feller, J. Non-Cryst. Solids 270 (1–3) (2000) 215. [13] G.H. Dieke, H.M. Crosswhite, Appl. Opt. 2 (7) (1963) 675. [14] K. Binnemans, R.V. Deun, C. Görller-Walrand, J.L. Adam, J. Non-Cryst. Solids 238 (1998) 11. [15] M. Shojiya, M. Takahashi, R. Kanno, Y. Kawamoto, K. Kadono, J. Appl. Phys. 82 (12) (1997) 6259. [16] G. Blasse, B.C. Grabmaier, Luminescent Materials, Springer-Verlag, Berlin, Heidelberg, 1994. [17] G.M. Kumar, B.N.S. Bhaktha, D.N. Rao, Opt. Mater. 28 (11) (2006) 1266. [18] T. Förster, Ann. Phys. 437 (1–2) (1948) 55. [19] F. Steudel, S. Loos, B. Ahrens, S. Schweizer, Proc. SPIE 9140 (2014) 91400I.
F. Steudel et al. / Journal of Luminescence 170 (2016) 770–777
[20] D. Umamaheswari, B.C. Jamalaiah, T. Sasikala, T. Chengaiah, I.-G. Kim, L.R. Moorthy, J. Lumin. 132 (5) (2012) 1166. [21] J.R. Lakowicz, Principles of Fluorescence Spectroscopy, Springer Science þ Business Media, LLC, 2006. [22] R. Praveena, V. Venkatramu, P. Babu, C. Jayasankar, Physica B: Condens. Matter 403 (19–20) (2008) 3527. [23] G. Blasse, J. Lumin. 1–2 (0) (1970) 766.
777
[24] G. Lakshminarayana, E.M. Weis, A. Lira, U. Caldi no, D.J. Williams, M.P. Hehlen, J. Lumin. 139 (2013) 132. [25] H. Lin, D.L. Yang, G.S. Liu, T.C. Ma, B. Zhai, Q.D. An, J.Y. Yu, X.J. Wang, X.R. Liu, E.Y.B. Pun, J. Lumin. 113 (1–2) (2005) 121.