1. Introduction

Rare earth ions doped luminescent materials have been deeply and exhaustively investigating for several decades in that these ions generally have plenty of well-shielded 4f states that can emit fluorescence covering different wavelength from ultraviolet to infrared. Therefore, these materials has generated much interest in applications such as plasma display panels, bulk lasers, fiber lasers/amplifiers, solar concentrators, displays, fluorescent lamps and transparent phosphors [1–3]. Recent ten years, white-LEDs have attracted lots of attention in solid state lighting for the replacement of conventional incandescent and fluorescent lamps due to their excellent advantages in energy saving and environment protection [2,4]. White-LEDs can basically classified into two types, one is multi-chip type by packaging red, green, blue LEDs together and the other one is single-chip type with one blue or ultraviolet light source coated with some phosphors. The multi-chip type must combine two or three different LEDs and the drive circuit is very complex. The single-chip type has only one type of LED and the circuit design is easier. In comparison with the commercial white LED fabricated with a blue chip coated with yellow phosphor YAG: Ce³⁺, the white LED based on near ultraviolet (NUV) chip coated with tri-phosphor fluorescent powder has higher color stability and better color index because all of the colors are determined by the phosphors [1,5]. Luminescent glass has many advantages, and widely used as fluorescent material. There is plenty of investigation on rare earth ion-doped luminescent glass excited by LED; most of them are limited in non-silicate glass system such as borate and phosphate glass. Compared to the non-silicate glass systems, the silica glass has some advantages such as higher mechanical, thermal and chemical stability [5–7]. The silica glass is rarely used as fluorescent materials other than optical fiber, because there is a natural tendency for rare-earth ions with higher concentration to form clusters and quench the luminescence in the silica glass [6–9]. Recently we developed a new method to overcome these drawbacks mentioned above. The method is primarily dependent on sintering of porous glass impregnated with rare-earth ions to prepare high-silica luminous glass. The component of the porous glass is very close to that of silica glass; therefore the sintered high-silica glass has excellent mechanical, thermal and chemical properties similar to silica glass [6,10,11]. Optically active elements such as metal ions and nanocrystals can be incorporated into the nano pores and channels in the porous glass. The sintered porous glasses can tolerate a higher concentration of dopants than traditional melt glasses without apparent concentration quenching [10,11]. By this route, we fabricated a series of transparent high-silica luminous glass with different emission. In this manuscript, we report on a new kind of Ce³⁺-Tb³⁺-Ca²⁺ co-doped green-emitting high silica luminous glass. The luminescence spectra show that there are energy transfers between Ce³⁺ and Tb³⁺, and also between ⁵D₃ and ⁵D₄ energy level of Tb³⁺. The excitation wavelength can be tuned from short-wavelength ultraviolet (238nm) towards long-wavelength ultraviolet (310nm) by energy transfer processes. Furthermore, the energy transfer process can be adjusted by addition of Ca²⁺ into the porous glass, and the transfer rate can be enhanced about four times than that of Ce³⁺-Tb³⁺ co-doped porous glass.

2. Experiment

Preparation of the rare earth ions doped porous glass followed the process described in earlier work of Chen and Yang et al [6,10,11]. We prepared the samples by sintering of the porous glass impregnated with Tb³⁺, Ce³⁺, and Ca²⁺ ions. In this system Ce³⁺ acts as sensitizer transferring energy to activator Tb³⁺ and Ca²⁺ acts as a modifier to this process. The ions were introduced into the glasses by immersing the porous glass into 0.2−1.7 mol% solutions containing corresponding nitrate compounds. After drying in air for 1 hour, the glasses were sintered at 1120°C in a reduced atmosphere for 2 hours. Finally, a series of compact, transparent and colorless glasses were fabricated. Then the large surfaces of glasses were optically polished for subsequent measurements with thickness of 1mm. Excitation and emission spectra in the ultra-violet and visible wavelength ranges were recorded on a JASCO FP-6500 fluorescence spectrophotometer equipped with a tunable excitation source range from 200nm to 750nm. Photoluminescence decay measurements of Ce³⁺ and Tb³⁺ were performed by using an FLS920 fluorescence spectrophotometer. Decay curves for Tb³⁺was recorded at 379nm under 238nm excitation and that of Ce³⁺ was recorded at 395nm with the excitation at 307nm. In order to reduce the test error, we put every sample at the same position and keep them on same angle related to excitation light. All of measurements were carried out at room temperature.

3. Results and discussion

Figure 1 shows the emission spectra of Tb³⁺ doped, Ce³⁺ doped,Tb³⁺–Ce³⁺ co-doped, and Tb³⁺–Ce³⁺–Ca²⁺ co-doped sintered porous glasses. For the Tb³⁺ doped porous glass (a), there is almost no emission under 310nm excitation and for the Ce³⁺ doped porous glass (b) there is only one broad emission bands around 375 nm. However, for the Tb³⁺–Ce³⁺ co-doped glass (c) and the Tb³⁺–Ce³⁺–Ca²⁺ co-doped glass (d), there are a broad emission band around 375 nm, and several narrow peaks centered at 488, 543, 588, and 622 nm. The broad emission band can be ascribed to the 5d→4f transition of Ce³⁺ and the narrow emission peaks are due to ⁵D₄→⁷F_J transitions of Tb³⁺ [4–8].The narrow emission peaks from Tb³⁺ indicates presence of energy transfer from Ce³⁺ to Tb³⁺. It was shown that by co-doping of a small amount of Ce³⁺ with Tb³⁺ into glass, the emission of Tb³⁺ can be significantly increased due to energy transfer from Ce³⁺ to Tb³⁺. Furthermore, it is amazing that compared with the Tb³⁺–Ce³⁺ co-doped glass, the Tb³⁺ ions in the Tb³⁺–Ce³⁺–Ca²⁺ co-doped glass have stronger emission whose intensity is enhanced at least three times than that in Tb³⁺–Ce³⁺ co-doped glass. It is obvious that Ca²⁺ plays an important role in increasing the emission of Tb³⁺, and the energy transfer processes can be adjusted by addition of Ca²⁺. The left inset shows the photographs of luminous glasses under 312nm excitation. It can be seen the glasses exhibit gradually increased green emission from sample c to sample d, which corresponds to the Tb³⁺–Ce³⁺ co-doped, Tb³⁺–Ce³⁺–Ca²⁺ co-doped glasses.

 

Fig. 1 Emission spectra of Tb³⁺ doped (a),Ce³⁺ doped (b), Tb³⁺–Ce³⁺ co-doped (c), and Tb³⁺– Ce³⁺–Ca²⁺ co-doped (d) sintered porous glasses. The excitation wavelength is 310 nm. Tb³⁺ doped glass has a Tb³⁺ concentration of 0.235 mol%. Tb³⁺–Ce³⁺ co-doped glass has a Tb³⁺ concentration of 0.235 mol% and a Ce³⁺ concentration of 0.08 mol%. Tb³⁺– Ce³⁺–Ca²⁺ co-doped glass has a Tb³⁺ concentration of 0.235 mol%, a Ce³⁺ concentration of 0.08 mol% and a Ca²⁺ concentration of 1.2 mol%. The left inset gives the photographs of luminous glasses exhibiting blue and green emission under 310nm irradiation. The a, b, c, and d samples correspond to Tb³⁺ doped glass, Ce³⁺ doped glass, Tb³⁺-Ce³⁺ co-doped glass, Tb³⁺-Ce³⁺-Ca²⁺ co-doped glass, respectively.

Download Full Size | PDF

To investigate the actual role of Ca²⁺ in the glass, we prepared a series of glass samples with the same Tb³⁺ and Ce³⁺ concentration and different Ca²⁺ concentrations. Figure 2 exhibits the ⁵D₄ emission spectra and the corresponding excitation spectra of Tb³⁺. The excitation and monitoring wavelengths is 310nm and 544nm respectively.

 

Fig. 2 (A) describes the ⁵D₄ emission spectra of Tb³⁺-Ce³⁺-Ca²⁺-co-doped glasses under 310nm excitation. All of the glasses have Tb³⁺ concentration of 0.235 mol% and Ce³⁺ concentration of 0.08 mol% respectively. The concentration of Ca²⁺ is changed from 0.4 mol%, 0.6 mol%,0.8 mol%, 1.0 mol% to 1.2 mol%. Figure 2(B) shows the excitation spectra of the corresponding glasses monitored at 544 nm.

Download Full Size | PDF

Figure 2(A) shows the ⁵D₄ emission spectra of Tb³⁺-Ce³⁺-Ca²⁺ co-doped glasses under 310 nm excitation, the concentration of Tb³⁺ and Ce³⁺ are constant but the concentration of Ca²⁺ is varied from 0.4 mol% to 1.2 mol%. It can be seen that the emissions from ⁵D₄-⁷F_j transitions of Tb³⁺ increased with increased Ca²⁺ concentration. Corresponding to the ⁵D₄ emission spectra, Fig. 2(B) shows increased excitation band with increased Ca²⁺ concentration. The excitation band centered at 320 nm is due to the 4f-5d absorption of Ce³⁺, and the weaker band centered at 238 nm is contribute to the absorption of Tb³⁺.

The above results imply that Ca²⁺ promote the energy transfer from Ce³⁺ to Tb³⁺. For the porous glass, there are some non-bridging oxygen (NNO) atoms at the internal surface of pores and channels and the rare-earth ions have high coordination numbers so the doped ions accumulated to share the limited non-bridging oxygen atoms [8,11]. Therefore in the sintered glass, parts of Tb³⁺ and Ce³⁺ tend to be clustered and have no contribution to emission. At low RE concentrations, the dopants can be distributed uniformly, bound to NNO atoms in RE–O–Si bonds. As the RE concentration is increased, clusters are formed through RE–O–RE bonding [8]. Clustering facilitates energy migration and fluorescence quenching. It is suggested from the above results that the Ca²⁺ can prevent the cluster formation because addition of Ca²⁺ can provide more non-bridging oxygen. Addition of Ca²⁺ facilitates the break-up of silica tetrahedron network; this will generate more non-bridging oxygen. At the same time, the lower coordination number of Ca²⁺ increases the probability of forming Ca–O–RE bond, so Ca²⁺ can substitute parts of clustered REs and make them free. Therefore there will be more rare-earth ions contributing to emission, and this will also decrease the distance between free Ce³⁺ and Tb³⁺ which can promote the energy transfer efficiency.

Furthermore, we also prepared Tb³⁺-Ca²⁺-codoped glasses with different Ca²⁺ concentration. Figure 3 shows the ⁵D₃ and ⁵D₄ emission spectra and excitation spectra of Tb³⁺. The excitation and monitoring wavelength is 238nm and 379 nm for ⁵D₃ emission,and 238 and 544 nm for ⁵D₄ Emission. Figure 3(A) shows the ⁵D₃ emission spectra, there are several emission peaks at 379, 415, 438, and 458 nm corresponding to ⁵D₃→⁷F_J (J = 6, 5, 4, and 3) transitions, respectively. It is shown that the emission intensity of ⁵D₃ gradually decreases with increased Ca²⁺ concentration. However, for the ⁵D₄ emission [Fig. 3(B)] peaked at 488, 544, 588, and 622 nm, the intensity can be greatly increased with increased Ca²⁺. Corresponding to the ⁵D₃ and ⁵D₄ emission spectra, Fig. 3(C) and 3(D) show the excitation spectra of Tb³⁺ obtained by monitoring 379nm (⁵D₃→⁷F₆) and 543nm (⁵D₄→⁷F₅)emissions, respectively [4,10–12]. The excitation band of 379nm decreases when increasing of Ca²⁺ concentration. However, the excitation band of 544nm increases with the increasing of Ca²⁺ concentration. The two excitation spectra just like the corresponding emission spectra have absolutely contrary tendency, this implies more and more energy of ⁵D₃ were transferred to ⁵D₄ with the increase of Ca²⁺. All above results implies the occurrence of cross-relaxation processes among Tb³⁺ ions, and Ca²⁺ play an important role in this process [13,14]. The cross-relaxation process can be represented as (⁵D₃, ⁷F₆) →(⁵D₄,⁷F₀), and was facilitated for the closely matched energy level difference between ⁵D₄ and ⁵D₃ levels and the ⁷F₆ and ⁷F₀ levels [8]. The energy transfer rate of this process strongly depends on the inter-ion distance, which is determined byTb³⁺ concentration.

 

Fig. 3 (A) and (B) are ⁵D₃ and ⁵D₄ emission spectra of Tb³⁺–Ca²⁺ co-doped glasses under 238-nm excitation. (C) and (D) are the excitation spectra of ⁵D₃ and ⁵D₄ emission of Tb³⁺ obtained by monitoring 379nm and 544nm emissions, respectively. All of the glasses have Tb³⁺concentration of 0.235 mol%. The concentration of Ca²⁺ is changed from 0.4 mol%, 0.6mol%, 0.8mol% to 1.0 mol%.

Download Full Size | PDF

The fluorescence decay curves of ⁵D₃→⁷F₅ emission are examined for the samples with different Ca²⁺ concentration. Figure 4 represents the decay curves of ⁵D₃→⁷F₅ emission for these samples. The cross relaxation between ⁵D₃ and ⁵D₄ leads to the decrease in ⁵D₃ decay with increased Ca²⁺ concentration and produces a nonexponential component in the initial regime of the decay curves. As can be observed in Fig. 4, the cross relaxation is inconspicuous at lower Ca²⁺ concentration, and the decay are nearly in the form of single exponential and the variation of decay with Ca²⁺ concentration can be hardly observed. The decay curves are getting more nonexponential as Ca²⁺ concentration increase. In order to give a reasonable interpretation to the nonexponential behavior, a fundamental approach is adopted based on the multipolar interaction scheme.

 

Fig. 4 Decay curves of the Tb³⁺ (⁵D₃→⁷F₅, 379nm) transition in the sintered porous silica glasses with different Ca²⁺ concentration under 238nm excitation. The concentration of Ca²⁺ is 0 mol% (a), 0.4 mol% (b), 0.6 mol% (c) and 0.8 mol% (d) respectively. The Tb³⁺ concentration is 0.235 mol%. The lines in colors are fitted lines of decay curves corresponding to sample a, b, c, and d, respectively.

Download Full Size | PDF

Assuming that the interaction is multipolar type and no migration takes place, the ⁵D₃ decay curves could be analyzed using the well-known direct quenching mechanism. There is no energy migration effect or less energy migration effect in the ⁵D₃ level because it is obviously suppressed by faster cross relaxation. Usually, the interaction scheme of ⁵D₃ emission involves both the dipole-dipole and the dipole-quadrupole interactions, which can be described by Inokuti-Hirayama formula [8,13,14]. The rare earth ions doped in porous glass tend to accumulate at the internal surface of the pores and channels of porous glass, therefore the concentration of rare-earth ions is different everywhere. During subsequent sintering process, partsk of rare-earth ions tend to cluster in the glass matrix. Therefore parts of Tb³⁺ ions doped in porous glass clustered and these ions have no or less contribution to the emission. The non-clustered Tb³⁺ ions can efficiently emit fluorescence from the ⁵D₃ and ⁵D₄ levels, and can be classified into two types due to the distance between Tb³⁺ ions, one isolated and with no interaction with other ions, another with cross-relaxation by multipolar interaction. Therefore, the fluorescence from the former should decay exponentially and the latter should decay according to Inokuti-Hirayama formula. So the resultant fitting equation could be formulated as [11,15]:

The fitting equation consists of the emission without cross-relaxation and the emission taking account of the cross-relaxation. Where, A is the contributive parameter; τis the intrinsic lifetime of a single ion (0.85msec for our samples measured for a sample with very low Tb³⁺ concentration of 0.05mol%); C is the number of acceptors per unit volume; C₀⁻¹ is the volume of donor’s sphere of influence ( = 4πR₀ ³/3, R₀ is the critical separation between donor and acceptor, at which the nonradiative rate equals that of the internal single ion relaxation); the ratio of C/C₀ means the number of acceptors in donor’s sphere of influence. The fitted curves are depicted in Fig. 4 and the contributive parameter (A) values and the resultant C/C₀ values are given in Table 1. The overall fitting result shows that the analysis using the multipolar interaction scheme produces a relatively good fit between the theoretical calculation and the measured data. From the Table 1, it is shown that the contributive parameter A gradually decrease with the increasing of Ca²⁺ concentration which means that the increased Ca²⁺ concentration facilitated the cross-relaxation. The ratio of C/C₀ has a increasing tendency with increased Ca²⁺ concentration. C is the number of acceptors per unit volume, that is, Tb³⁺ concentration itself in this case. According to our previous assumption, it refers in particular to the concentration of non-clustered Tb³⁺ with cross-relaxation. The addition of Ca²⁺ facilitates the cross-relaxation process and results in the increase of ⁵D₄ emission. Furthermore, we calculated the mean lifetime (τ_m) and the energy transfer efficiency (η_ET) as follows [4]:

The calculated lifetime (τ_m-xCa) and energy transfer efficiency (η_ET) is also exhibited in Table 1.

Table 1. Parameters for the fitted equation, decay life time and energy transfer efficiency of samples with different Ca²⁺ concentration

View Table | View all tables in this article

The mechanism of the energy transfer between Ce³⁺ and Tb³⁺ can also be analyzed using the approach of Dexter’s theory or Inokuti-Hirayama formula. Usually the interaction between Ce³⁺ and Tb³⁺ is an electric dipole-dipole nature for lower acceptor concentrations.

Figure 5 compares the Ce³⁺ fluorescence decay for several samples doped with different Ca²⁺ concentration. With increased Ca²⁺ the decay curves are non-exponential and the tendency is getting more prominent as Ca²⁺ concentration increase. The deviation from the single exponential behavior is due to the radiationless process involving energy transfer. Assuming a dipole-dipole radiationless energy transfer and without energy migration occurs over the donor subsystem, the decay can be fitted by such a formula [11,15]:

 

Fig. 5 Decay curve of Ce³⁺ emission at 395nm in the sintered porous silica glasses with different Ca²⁺ concentration under 307nm excitation. The concentration of Ca²⁺ is 0.2 mol% (a), 0.8 mol% (b) and 1.4 mol% (c), respectively. The concentration of Ce³⁺ is 0.08 mol% and Tb is 0.235 mol%. The red lines are fitted lines of decay curves corresponding to sample a, b, and c, respectively.

Download Full Size | PDF

Where I(0) represents the initial intensity of the emission, τ₀ is the intrinsic lifetime measured with very low Ce³⁺ concentration (56ns for our samples) and other parameters have the same meaning with Eq. (1) mentioned above. From Fig. 5, it is seen that there is a good fit between measured data and the fitted curves. Therefore, the energy transfer between Ce³⁺ and Tb³⁺ is based on dipole-dipole interaction. We also calculated the mean lifetime (τ) and the energy transfer efficiency (η_ET) according to formula (2) and formula (3).

The parameter C/C₀, calculated mean lifetime (τ) and energy transfer efficiency (η_ET) are show in Table 2.

Table 2. Parameters for the fitted equation, decay life time and energy transfer efficiency of Ce³⁺-Tb³⁺-Ca²⁺ co-doped samples with different Ca²⁺ concentration. CTCx mol% represent Ce³⁺-Tb³⁺-Ca²⁺ co-doped glasses with Ca²⁺ concentration of x mol%

View Table | View all tables in this article

The fitting of Tb³⁺ decay and Ce³⁺ decay by formula (1) and formula (4) is successful respectively. The Ce³⁺ decay cannot be fitted by formula (1), which means there is no isolated Ce³⁺, and almost all of the non-clustered Ce³⁺ ions have interaction with Tb³⁺. The above result is also approved by the fact that the critical distance of cross-relaxation between Tb³⁺ ions is usually shorten than that of Ce³⁺-Tb³⁺ dipolar-dipolar interaction [16–18].

It is obvious that the rare earth ions of Ce³⁺ and Tb³⁺ doped in porous glass results in Ce³⁺,Tb³⁺ and Ce³⁺-Tb³⁺ clusters as well as non-clustered Ce³⁺ and Tb³⁺ ions. In porous glasses, ion clustering occurs because the rare earth ions have large coordination number and must share the limited non-bridging oxygen, therefore the local concentration of non-clustered ions is low and the distance between non-clustered Ce³⁺ and Tb³⁺ is large, so the energy transfer efficiency between non-clustered ions is low and the emission intensity of Tb³⁺ is weak. Addition of Ca²⁺ into Ce³⁺-Tb³⁺ co-doped glass can increase the emission intensity of Tb³⁺ greatly, because the Ca²⁺ ions lead to more non-network oxygen and replace clustered Ce³⁺ and Tb³⁺. On the one hand, the more non-network oxygen results in more dispersed rare earth ions, on the other hand, the replaced Ce³⁺ and Tb³⁺ will increase the local concentration non-clustered Ce³⁺ and Tb³⁺. Therefore the local concentration of non-clustered Ce³⁺ and Tb³⁺ is greatly increased. The distance among non-clustered rare earth ions can be greatly shortened. Therefore, the addition of Ca²⁺ facilitates the energy transfer between Ce³⁺ and Tb³⁺ and the cross-relaxation between the ⁵D₃ and ⁵D₄ levels of Tb³⁺. Therefore, the ⁵D₄ emission of Tb³⁺ can be enhanced greatly.

4. Conclusions

In conclusion, the sintered porous glasses impregnated with Tb³⁺ ions were successfully prepared as intense green emitters under long wavelength ultraviolet excitation. Luminescence spectra and fluorescence decay of Tb³⁺-Ca³⁺ co-doped glasses and Ce³⁺-Tb³⁺ co-doped glasses were measured and studied. It is suggested that the addition of optically inert glass network modifier oxides (such as Ca²⁺) play an important role in facilitating the energy transfer between non-clustered sensitizers and activators. Due to their strong emissions under long wavelength ultraviolet excitation, the colorless transparent green-emitting glasses have a potential for application in lasers, solar concentrators, LED displays and fluorescent lamps.