In this study, by a conventional melt quenching method, we synthesized novel up-conversion phosphors of 60TeO2–30TlO0.5–(9−x)ZnO–xTm2O3–1Yb2O3 (x = 0.1–0.5) glasses, whose system was recently developed in our collaborative group, and their blue up-conversion photoluminescence (UCPL) of Tm3+ ions via three-step energy transfer from near-infrared (NIR) sensitizer of Yb3+ ions was observed. In particular, the substantial rate of the energy transfer <γd5> in the third step from Yb3+ to Tm3+ under excitation at 975 nm, which determined the final blue UCPL intensity, was estimated as a function of the rare-earth concentration. With an aid of analytical methods of PL lifetime and Judd–Ofelt theory, it was revealed that the highest energy transfer rate <γd5> was achieved to be 2.07 × 10−17 cm3/s for x = 0.2, and further increasing Tm2O3 content x in the fixed Yb2O3 resulted in the decrease in the energy transfer rate <γd5>. One of the plausible causes was concentration quenching of Yb3+ ions. The other was back-transfer from Tm3+ to Yb3+ ions. The influence of the condition of glass synthesis and the melting time on <γd5> was also discussed.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Rare-earth up-conversion photoluminescence (UCPL) is a sort of visible light emission excited by near-infrared (NIR) light. This UCPL requires two kinds of optically active ions: the emitter of UCPL (e.g., Er3+, Eu3+, Tm3+, and so on) and energy donor (in cases, Yb3+ ion is widely used.).[1-3] The sensitizer transfers the energy absorbed from excitation light to the activator, resulting in emission of higher energy light than excitation light. Such a mechanism of UCPL is expected for applications to biolabels, display devices (3-D display), sensors, and so on.[1, 3-7]
It is well-known that Tm3+/Yb3+ co-doped phosphors show blue UCPL by irradiation of NIR rays around 980 nm. The energy level diagram of Tm3+/Yb3+ system is shown in Fig. 1, which is helpful for comprehension of UCPL mechanism. Tm3+ ion receives excitation energy from Yb3+ ions, resulting in the excitation to higher energy levels, such as 1G4 and 3H4 levels via several energy transfers. When Yb3+ ions are excited at ~980 nm, Yb3+ ions act as an energy donor for Tm3+ ions. The first and third steps of the energy transfers (see Fig. 1) are in nonresonant type (energy difference is compensated by phonons) and the second is in resonant type. Yb3+ ions have higher oscillator strength (several times higher than those of Tm3+); thus, the most probable second step for Tm3+ to rise up to 3F2,3 levels or 3H4 level (after nonradiative transition from 3F2,3) is resonant energy transfer from Yb3+ to Tm3+ ions ((Yb3+, Tm3+): (2F5/2, 3F4)→(2F7/2, 3F2,3)) as well as excited state absorption (3F4→3F2,3 by 975 nm). On the other hand, the third step for Tm3+ blue UCPL is determined not by excited state absorption but by the energy transfer mechanism ((Yb3+, Tm3+): (2F5/2, 3H4)→(2F7/2, 1G4)), regardless of the excitation process for Tm3+ ions up to 3H4 level. When the excited 4f electrons in Tm3+ ion are relaxed to lower energy levels, blue (1G4→3H6 at 480 nm), red (1G4→3F4 at 650 nm, 3F2,3→3H6 at 700 nm), or NIR (3H4→3H6 at 800 nm) emission can be observed. As shown later, the rate of third energy transfer can be analyzed by the emission ratio I480/I800 as a good measure of the blue up-conversion PL efficiency.
While Er3+/Yb3+ co-doped phosphors, which emit red or green UCPL, have been studied by many researchers,[2, 6, 8] Tm3+/Yb3+ co-doped phosphors capable of emitting blue UCPL have been studied not more than Er3+/Yb3+ co-doped phosphors. The difference between these two kinds of phosphors is the number of steps for energy transfer. While Er3+ emits UCPL via only two-step energy transfer, blue UCPL of Tm3+ ions requires three-step energy transfer; thus, up-conversion efficiency of Tm3+/Yb3+ co-doped phosphor is in general lower than the conversion of Er3+/Yb3+ co-doped phosphor. Therefore, improving blue up-conversion efficiency of Tm3+ ions is an issue to be solved for optical applications.
As the host material of phosphor, fluoride materials are widely considered because of their low phonon energy,[9, 10] but we chose a TeO2–TlO0.5–ZnO glass system as a host material, which is one of the recently developed in our research group, and we have focused ourselves on the glass composition and rare-earth concentration. TeO2-based glasses are attractive for optical materials because of their low phonon energy among oxide materials (650–750 cm−1), high refractive index, high nonlinear optical properties, high transparency, and so on.[12-14] TeO2–TlO0.5–ZnO glass system is characterized by good chemical stability, homogeneous distribution of lanthanide ions, as well as lower phonon energy. In this study, to investigate the dependence of rare-earth concentration (Tm3+/Yb3+) on Tm3+ blue UCPL, Tm3+/Yb3+ co-doped TeO2–TlO0.5–ZnO glasses in different rare-earth concentrations were synthesized and estimated.
60TeO2–30TlO0.5–(9−x)ZnO–xTm2O3–1Yb2O3 (x = 0.1–0.5) glasses were prepared by a conventional melt quenching method at ambient atmosphere using commercially available chemicals, Tl2CO3, ZnO, Tm2O3, and Yb2O3. Tellurium dioxide TeO2 was only prepared by decomposing commercial telluric acid (H6TeO6, Aldrich) at 550°C for 24 h. The mixture of these powders with stoichiometry ratio was melted in a platinum crucible at 800°C for 30 min (series A) and 8 h (series B) in air. During melt preparation, a part of component in melt is possibly to be evaporated. In the case of this study, ZnO comparatively tended to be evaporated somewhat among TeO2, Tl2O, and ZnO; however, there was no substantial difference between the nominal composition and final composition even though the melting time was long of 8 h. Thus, to confirm the influence of segregation rare-earth elements in glasses, the different melting times were adopted. The obtained glasses were annealed for 10 h at 40°C below Tg (Tg = 207°C) and then cooled down to room temperature. Planar surfaces of these glasses were polished to optical flat. After polishing, the thicknesses of each glass were ~1 mm.
UCPL spectra were measured using the excitation beam from a CW semiconductor laser (JDSU corporation, San Jose, CA) (JDSU, 27-7552) operated at 975 nm. The excitation power density was varied from 1.67 kW/cm2 to 8.68 kW/cm2. Absorption spectra were measured by UV–vis–NIR spectrophotometer (V-570; JASCO, Easton, MD). The wavelength range was 300–2500 nm, which was examined for Tm3+ and Yb3+ ions absorption peaks needed for Judd–Ofelt analysis.[16, 17] Lifetime of 1G4→3H6 transition in Tm3+ ion was measured by dye laser (KEC-160; USHO, Osaka, Japan) using a coumarin 460 dye whose wavelength was tuned to 462 nm.
The f–f transition probabilities of rare-earth ion can analytically be calculated from optical absorption data by use of the so-called Judd–Ofelt theory. This theoretical background was provided for the first time by B. R. Judd and G. S. Ofelt,[16, 17] and many researchers have used this analysis for researches of rare-earth optical materials. Krupke et al. gave the following formulations,[16-20] which were used in this study. From absorption spectra, an oscillator strength was at first determined by the equation:
where ε (ν) is the molar absorption coefficient at wave number. This should be equal to the sum of the electric dipole and magnetic dipole oscillator strengths for the transition band of interest, fed and fmd. Therefore, it can be rewritten as fexp = fed + fmd, with
where ν is the wave number of transmission J→J', n is the refractive index, h is Plank's constant, and Sed and Smd are line strengths for the electric dipole and magnetic dipole transitions, respectively.
The Judd–Ofelt (JO) parameters, Ωλ (λ = 2, 4, 6), are phenomenological ones and material-dependent. With the JO parameters obtained from Eqs. (1)-(5), the oscillator strengths for the transition were recalculated as theoretical values. The accuracy of the least square fitting for the calculation of JO parameters (Eqs. (1)-(5)) can be estimated by the rms deviation.[21, 22]
where N is the number of absorption band involved in the JO calculation. The details on rms deviation and rms error are also compiled in our recent study.
The term |<aJ||Uλ||bJ'>|2 is the square of the matrix elements of the tensorial operator Uλ, which connects states |aJ> to |bJ'>. The values of |<aJ||Uλ||bJ'>|2 are considered to be independent of the host matrix; thus, the list provided previously by Carnall et al. was utilized. The spontaneous emission probabilities (A) of the different electronic transitions are given by
UCPL spectra of 60TeO2–30TlO0.5–8.9ZnO–0.1Tm2O3–1Yb2O3 glass melted for 30 min with different excitation power densities are shown in Fig. 2. Three emission peaks centered at 480, 650, and 800 nm were due to 1G4→3H6, 1G4→3F4, and 3H4→3H6 transition of Tm3+ ions, respectively. Essentially, Tm3+ ion has an energy level 1D2, higher than 1G4 level, resulting in the blue-UV emissions of 450 nm (1D2→3F4) and 360 nm (1D2→3H6).[25, 26] In this study, these emissions could not be observed. Using UCPL spectra, saturated intensity ratio a, which is needed for calculation of energy transfer rate γd5, could be obtained. Figure 3 shows UCPL integrated intensity ratio (I480/I800) of 60TeO2–30TlO0.5–8.9ZnO–0.1Tm2O3–1Yb2O3 glass as a function of the excitation power density, Iexc, whose curve is fitted by the following equation:
where IS is defined by IS= hνexc/τdσd (τd is the experimental lifetime of the donor, and σd is the ground-state absorption cross-section at the excitation wavelength) and means an excitation power density at r = a/2. The obtained saturated intensity ratios are shown in Table 1. The highest saturated intensity ratio is 0.45 in the glass melted for 8 h with 0.2 mol% of Tm2O3. According to Silva et al., the energy transfer rate is proportional to the saturated intensity ratio, a. Thus, the glass with the highest saturated intensity ratio is expected to show the highest energy transfer rate, γd5.
Table 1. Saturated Intensity Ratio, a of 60TeO2–30TlO0.5–(9−x)ZnO-xTm2O3–1Yb2O3 Glasses (x = 0.1–0.5)
Saturated intensity ratio, a
Optical absorption spectra of 60TeO2–30TlO0.5–(9−x)ZnO–xTm2O3–1Yb2O3 (x = 0.1–0.5) glasses are shown in Fig. 4. The solid lines represent for the absorption spectra of the glasses melted for 30 min and the dotted lines for those of glasses melted for 8 h. The absorption peaks centered at 690, 797, 1215, and 1690 nm were attributable to 3H6→3F2,3, 3H6→3H4, 3H6→3H5, and 3H6→3F4 excitation of Tm3+ ions, respectively.[27, 28] The other peak centered at 980 nm was assigned to 2F7/2→2F5/2 of Yb3+ ions. The absorption peak intensities of Tm3+ ions increased with increasing Tm2O3 concentration. In Judd–Ofelt analysis of absorption data of Tm3+ ion, we obtained oscillator strengths of the Tm3+ absorption transition and hence through Eqs. (1)-(5) Judd–Ofelt parameters Ωλ (λ = 2, 4, 6), which are shown in Table 2. The results of JO calculation have a reasonable agreement between experimental and theoretical (recalculated) oscillator strengths with small δRMS. The rms errors were at most 22%, and the minimum was 11.9% for x = 0.1 in series A (30 min.). The Judd–Ofelt parameters were characterized by larger Ω2 values (4–5 pm2) with respect to Ω4 and Ω6, indicating asymmetric ligand structures around Tm3+ ions in tellurite glasses. Emission probabilities A were calculated by the estimated Judd–Ofelt parameters Ωλ via Eqs. (4), (5) and (7), which are shown in Table 3.
Table 2. Refractive Indices, Experimental and theoretical oscillator strengths, Judd–Ofelt Parameters (Ω2,4,6 and δRMS (with rms error Defined by Eqs. (7)–(9) in Ref.(23) of 60TeO2–30TlO0.5–(9−x)ZnO–xTm2O3–1Yb2O3 Glasses (x = 0.1–0.5)
fexp × 10−6
δRMS × 10−6
30 min (series A)
8 h (series B)
Table 3. Transition Probabilities Ai0 and the Ratio of Products for Transition Probabilities Ai0 and Transition Energy hν i0 of 60TeO2–30TlO0.5–(9−x)ZnO–xTm2O3–1Yb2O3 Glasses (x = 0.1–0.5). The Subscripts of i0 (i = 3 and 5) Stands for f–f Transitions of Tm3+ ions (30: 3→0 (3H4→3H6), 50: 5→0 (1G4→3H6), See also Fig. 1)
Lifetime decay curves of 1G4→3H6 transition in Tm3+ ion for all synthesized glasses were examined. Figure 5 (a) shows lifetime decay curves of the synthesized glasses for 30 min, and Fig. 5 (b) shows those for 8 h. By using the obtained lifetime decay curves, lifetime of 1G4→3H6 transition in Tm3+ ion, τ5, was calculated by the following equation.
Calculated lifetimes τ5 are shown in Table 4. Regardless of the melting time, the lifetimes monotonically decreased with an increase in Tm2O3.
Table 4. Lifetime τ5 for 1G4 Level of Tm3+ ions for 60TeO2–30TlO0.5–(9−x)ZnO–xTm2O3–1Yb2O3 Glasses (x = 0.1–0.5)
Life time τ5/μs
In this study, we carried out several experiments and obtained such parameters as saturated intensity ratio a, transition probabilities A50, A30, and lifetime τ5. Using these parameters, we attempted to calculate energy transfer rate, γd5, by the following equation:
where Nd is donor concentration of Yb3+ ion. Energy transfer rate γd5 is defined as being a probability that excitation energy on Yb3+ ion is transferred to Tm3+ ion in the excited state ((Yb3+,Tm3+):(2F5/2, 3H4)→(2F7/2, 1G4)). Although it is plausible that Nd is equal to the Yb concentration doped in glasses, NYb (=2.29 × 1020 cm−3), due to the saturation condition for the blue UCPL obtained, possible relaxation processes for the excited Yb3+ ions, for example, Yb3+ concentration quenching and multiphonon relaxation could reduce the population Nd of the excited state 2F5/2 of Yb3+. Moreover, the population of Yb3+ ions in the excited state would be influenced by back-transfer from Tm3+ in 1G4 level to Yb3+ in ground state. Thus, here, we define substantial energy transfer rate <γd5> as follows,
where ηYb is the population yield of Yb3+ ions in the excited state (2F5/2). Although Nd is not determined by the experiments employed in this study, we can instead estimate the substantial energy transfer rate <γd5>, which is directly related with blue UCPL efficiency. The calculated substantial energy transfer rates <γd5> are shown in Table 5 and Fig. 6. According to Table 5, the highest energy transfer rate was found to be 2.07 × 10−17 cm3/s in the glass doped with 0.2 mol% Tm2O3 and melted for 8 h, which was higher than the reported value by Silva et al. that was 1.21 × 10−17 cm3/s for Tm3+/Yb3+ co-doped Al2O3–CaO–SiO2–MgO glass. In the optical absorption spectra, the absorption edges of the synthesized glasses were located around 420 nm and had a small broad tail to 500 nm regardless of the rare-earth concentration, as shown in Fig. 4. As the central location of the emission peak of blue UCPL of Tm3+ ion is 476 nm, it is supposed that the blue UCPL would be more or less decreased by the absorption of host glass. If the optical window of host glass were extended to lower wavelength, the intensity of blue UCPL could be increased, resulting in further increase of saturated intensity ratio a and energy transfer rate <γd5>.
Table 5. Substantial Energy Transfer Rate, <γd5>, of 60TeO2–30TlO0.5–(9−x)ZnO–xTm2O3–1Yb2O3 Glasses (x = 0.1–0.5)
ETR, <γd5> (×10−17 cm3/s)
In general, energy transfer rate γd5 depends on the mean distance of two ions (energy donor and acceptor) and on the spectral overlap of the emission spectrum of the donor ion (Yb3+) and the absorption spectrum of the acceptor ion (Tm3+). In the condition of constant glass composition and Yb3+ concentration, γd5 could only depend on the Tm3+ concentration and exhibit a monotonic increase with Tm3+ concentration. Nevertheless, our experimental results show a maximum in the Tm3+ concentration range studied. It is because the substantial energy transfer rate <γd5> includes the influence of back-transfer from excited Tm3+ in 1G4 to ground state Yb3+ as well as Yb3+ concentration quenching and other Yb3+ decay processes. With an increase in Tm3+ concentration, the total concentration of rare-earth ions (Tm3+ and Yb3+ ions) was increased. As a result, Yb3+–Tm3+ and Yb3+-Yb3+ mean distances were decreased. In the increasing Tm2O3 concentration region ([Tm2O3] > 0.2 mol%), Yb3+ concentration quenching possibly occurred and ηYb could then be decreased. As for Tm3+ ions, the lifetime was not varied largely for Tm3+ concentration of x = 0.2 to 0.3 mol%, and thus, Tm3+ concentration quenching was not very significant. The further decreased Yb3+-Tm3+ mean distance in the higher Tm3+ concentration (x ~0.4, 0.5 mol%) could decay Tm3+ population due to Tm3+→Yb3+ back-transfer, as evidenced by the shortened Tm3+ lifetime (See Table 4).
In comparison with series A and B with the different melting time of 30 min and 8 h, respectively, the substantial energy transfer rates <γd5> of the glasses with 0.1 and 0.2 mol% of Tm2O3 in series B (the melting time 8 h) are relatively higher than those of the glasses with the same doping concentration but 30-min melting time. These results are caused by the long melting time of 8 h. That is, the long melting time was expected to made the glass more homogeneous and enough to disperse rare-earth ions in glasses and eventually avoided the concentration quenching of Yb3+. Therefore, the excitation energy on Yb3+ ions could be transferred efficiently and allowed Tm3+ ions to the higher energy levels to give blue UCPL. In this study, total rare-earth concentration (Yb3+ and Tm3+) was varied from 1.1 to 1.5 mol%, where Tm concentration was x = 0.1 to 0.5 mol%. Due to relatively high Yb3+ concentration in comparison with Tm3+ ions for x = 0.1 and 0.2 mol%, Tm3+–Tm3+ distance is supposed to be far enough even if the rare-earths are segregated. This is a reason why the PL lifetime of Tm3+ in 1G4 level is almost independent of the segregation of rare-earths (Tm3+ and Yb3+) (See Table 4). On the other hand, Yb3+ ions are closer each other after the segregation. In this case, Yb3+–Yb3+ energy migration could easily occur, and if it was combined with quenching centers, the excited state of Yb3+ ions could be depolulated, meaning the reducing Yb3+ population yield of ηYb and resultantly the reducing substantial energy transfer rate <γd5>. In Tm3+ concentrations of 0.1 to 0.2 mol%, <γd5> was increased for 8-h melting time glasses, corresponding to the increasing energy transfer rate γd5 due to shorter distance between Yb3+ and Tm3+ ions.
From Table 2, the systematic difference in the refractive indices between the short (series A) and long (series B) melting times was notified. Thus, density data were evaluated for all the sample studied and were resultantly found to be constant, ~6.4 g/cm3, regardless of Tm2O3 concentration and the melting times. On the one hand, it was because Tm2O3 content was low, 0.1–0.5 mol%. On the other hand, the melting times had no significant influence on the density data, which would be due to compositional changes, in particular, comparably volatile ZnO, for longer melting time and more developed tellurite networks of four oxygen coordinated Te (TeO4, trigonal bipyramids (tbp)), which has higher polarizability than three oxygen coordinated Te (TeO3, trigonal pyramid (tp)). This increased refractive index would also result in increased transfer rates.
As stated in Introduction, Tm3+ ion can emit blue photoluminescence via a three-step energy transfer from Yb3+ ions. Thus, it is speculated that the appropriate rare-earth ions ratio of Tm3+ and Yb3+ for the highest energy transfer rate could be Tm3+:Yb3+ = 1:3. However, the obtained result was Tm3+:Yb3+ = 1:5. This can be explained as follows. Because rare-earth ions were well-dispersed, there occurred energy migration from excited Yb3+ ion to ground-state Yb3+ ion. Importantly to be mentioned, the migration itself does not reduce a population of the excited state of Yb3+ ions. The excitation energy was traveled via several Yb3+ ions and finally to Yb3+ ions located close to Tm3+ ions on excited/ground state, which was enabled to promote the capture of the excitation energy by Tm3+ ions. The highest energy transfer rate was obtained for Tm2O3 concentration of 0.2 mol%, and then the energy transfer rates were decreased with an increase in Tm2O3 over 0.2 mol%.
The reduced substantial energy transfer rate <γd5> = ηYb·γd5 can be explained by the term of segregation of rare-earth elements doped and concentration quenching due to the depopulation of the excited state of Yb3+ ions ηYb. Tm3+→Yb3+ back-transfer is also one of the other possible causes of decreasing <γd5> in higher concentration region, where γd5 is not constantly increased with decreasing distance between Yb3+ and Tm3+ ions. The higher total concentration of rare-earth ions (totally 1.2–1.5 mol%) decreased a mean distance not only between Tm3+ and Yb3+ ions but also between Yb3+ ions, the latter of which would induce energy migration among Yb3+ ions, as mentioned above. Such higher amounts of such rare-earth impurities and rare-earth segregation (for 30-min melting time) could produce quenching centers related with Yb3+ concentration quenching. In the argument, Yb3+ concentration quenching, which is allowed to be called energy-migration-enhanced quenching, can be defined as being a combinational process of energy migration among Yb3+ ions and subsequent energy capture by quenching sites. As for whether energy migration can be desired or not, the presence of quenching sites (although quenching sites themselves are still not identified) should be considered because energy migration itself is not harmful. The requirement of Yb3+ concentration higher than Tm3+ ions (here, the optimized ratio was [Yb3+]/[Tm3+] = 5) shows that the energy migration among Yb3+ ions is more or less effective for Tm3+ ion to capture excitation energy traveling on Yb3+ ions if quenching sites are not very significant.
In this study, we synthesized Tm3+/Yb3+ co-doped 60TeO2–30TlO0.5–(9–x)ZnO–xTm2O3–1Yb2O3 (x = 0.1–0.5) glasses with varying rare-earth concentration by a conventional melt quenching method and estimated their blue UCPL properties. As a result, it was found that the rare-earth concentration ratio for the highest substantial energy transfer rate <γd5> (=ηYb·γd5), 2.07 × 10−17 cm3/s, for the third step from Yb3+ to Tm3+ ions was Tm3+:Yb3+ = 1:5 when the melting time was 8 h and x = 0.2 mol%. The excess Yb3+ ions were responsible for the energy migration from excited Yb3+ to ground-state Yb3+ ions, and the long melting time of 8 h resulted in the elimination of segregation effect leading to Yb3+ concentration quenching. By contrast, the short melting time of 30 min caused rare-earth ions to be much segregated, resulting in a decrease of the substantial energy transfer rate. In higher concentration region (x > 0.2 mol%), a mean distance between Yb3+ and Yb3+ (Tm3+) was decreased, which would induce further energy migration among Yb3+ ions and energy transfer from Yb3+ to Tm3+ ions. However, closer distance between Yb3+ and Tm3+ ions could reduce Tm3+ population in 1G4 level (the lifetime was practically reduced) as a result of Tm3+→Yb3+ back-transfer and eventually γd5 or <γd5> was decreased. Moreover, the higher amounts of such rare-earth impurities (totally 1.3–1.5 mol%) could produce quenching centers, which would be combined with energy migration process among Yb3+ ions (energy-migration-enhanced quenching, or Yb3+ concentration quenching) so as to decrease the Yb3+ population yield in the excited state ηYb. This was also one of the reasons of the decreasing <γd5>.
This work was supported by the JSPS International Training Program (ITP), “Young Scientist-Training Program for World Ceramics Networks” and by grant from Institute of Ceramics Research and Education (ICRE) in Nagoya Institute of Technology.