The three up-conversion emission bands centered at 695, 544 and 474 nm have attributed to the transitions Tm³⁺: ³F_{2, 3}→ ³H₆, Ho³⁺: ⁵S₂→ ⁵I₈, as well as Ho³⁺: ⁵F_{2, 3}→ ⁵I₈ and Tm³⁺: ¹G₄→ ³H₆, respectively. The 740 nm emission and its origination have seldom been reported. The three possible passages significantly enhancing 740 nm emission are shown as follows: Firstly, the 740 nm emission originates from energy transition of Ho³⁺. The 740 nm red emission should be assigned to the ⁵S₂→ ⁵I₇ transition of Ho³⁺ ion, although there is no corresponding absorb band near at the 750 nm wavelength in Ho³⁺:ZnWO₄ crystal^[10]. Figure 4 shows the emission spectra of the Tm³⁺:Ho³⁺ co-doped glass samples plus the curve for the pure Tm³⁺ doped glass sample with the same Tm³⁺ concentration. The spectra have been normalized with respect to the 695 nm peak. It can be clearly discovered that the relative intensity of the 740 nm peak decreases linearly with increasing Ho³⁺ concentration. On the other hand, figure 1 shows that the intensity of the 740 nm peak growths with increasing Tm³⁺ concentration while the Ho³⁺ concentration is fixed. These cases indicate that the major reason of enhancing 740 nm emission is not induced by the energy level transition of Ho³⁺. Perhaps stark split of energy level is another reason. The energy difference between 695 nm and 740 nm emission is about 900 cm^-1 which approaches to a phonon energy of matrix. The stark split of energy level ³F_{2, 3} of Tm³⁺ maybe lead 740 nm emission by releasing a phonon energy to matrix. However, the shape of spectrum should not be transformed. The third possible reason is that 695 and 740 nm emissions are originated from different energy levels transition of Tm³⁺, respectively. The energy difference between energy levels ¹D₂ and ³F_{2, 3} of Tm³⁺ is about 13 500 cm^-1 which matches with 740 nm emission. Hence, we conclude that 740 nm emission would be likely to be assigned to the ¹D₂→ ³F_{2, 3} transition of Tm³⁺.

Fig.4

	Figure Option ViewDownloadNew Window
	Fig.4 Fluorescence spectra for the (99.5-y)(70TeO2-20WO3-10ZnO)-0.5Tm2O3-yHo2O3(y=0 and 0.3) temlurite glasses normalied with respect to the 695 nm peak of Tm3+

The main up-conversion processes in pure Tm³⁺ doped tellurite glasses are as follow^{[11, 12, 13]}:

GSA:Tm³⁺: ³H₆+hν ₀→ Tm³⁺: ³H₄ (1)

ESA:Tm³⁺: ³H₅+hν ₀→ Tm³⁺: ¹G₄ (2)

CU: Tm³⁺: ¹G₄+Tm³⁺: ¹G₄→ Tm³⁺: ³F_{2, 3}+Tm³⁺: ¹D₂ (3)

With the input laser radiation of 808 nm, Tm³⁺ populated at the ground state ³H₆ is initially excited to the ³H₄ level by absorbing a photon (GSA). The excited Tm³⁺ in the ³H₄ level relaxes to ³F₅ state through multi-phonon relaxation process and then absorbs a second photon to the ¹G₄ level (ESA). Tm³⁺ ions store energy on the ¹G₄ level and then return back to the ground state (³H₆), resulting in an emission at 474 nm. Energy difference between ¹G₄ and ¹D₂ levels is almost the same as between ¹G₄ and ³F_{2, 3} levels for Tm³⁺. No such energy coincidence is seen for energy levels in Tm³⁺. It is possible that there is a cooperation up-conversion process (CU). A pair of Tm³⁺ ions, both being in the ¹G₄ level, cooperatively transfers their excitation energy to each other. One is deexcited to the ³F_{2, 3} level while the other is promoted to the ¹D₂ level. Tm³⁺ populated at the ¹D₂ level radiate to the ³F_{2, 3} level with yielding 740 nm red light. The 695 nm up-conversion emission process can be described as follow: The increase of population for the ³F_{2, 3} level is not only caused by the process that Tm³⁺ at ¹G₄ level decay rapidly to the ³F₂ meta-stable state and further relax to ³F₃ level through multi-phonon relaxation, but also created by the CU process and 740 nm emission. Tm³⁺ ions are transferred from ³F_{2, 3} level to ³H₆ level, emitting red light at 474 nm. As can be seen from above, Tm³⁺ populated at the ¹D₂ level (740 nm emission ) is mainly originated from the CU process while Tm³⁺ populated at the ³F_{2, 3} level (695 nm emission) can be obtained from the other two passages besides the CU process. Therefore, the intensity of 695 nm emission is stronger than that of 740 nm emission.

The following notations were also adopted for the energy transfer processes investigated in pure Tm³⁺ doped tellurite glass:

CR1: Tm³⁺: ³H₄+Tm³⁺: ³H₆→ Tm³⁺: ³F₄+Tm³⁺: ³F₄ (4)

SR: Tm³⁺: ³H₄→ Tm³⁺: ³H₅+hν (5)

Because the energy difference between ³F₄ and ³H₄ levels better matches that between ³F₄ and ³H₆ levels for Tm³⁺, Tm³⁺ at ³H₄ level decay rapidly to the ³F₄ meta-stable state via cross relaxation (CR1) process while the Tm³⁺ at ³H₆ level are raised to ³F₄ level. As a result, the population accumulated at ³F₄ level of Tm³⁺ gets more densely, and a much stronger luminescence of the ³F₄ level of Tm³⁺ at 1.8 μ m can be expected (see Fig.5). Tm³⁺ at ³H₄ level can be transferred to³H₅ level via spontaneous radiation (SR) process, and further relax to³F₄ level through multi-phonon relaxation process. These processes would result in that a strong emission at 1.47 μ m is occurred and Tm³⁺ at ³H₄ level is heavily populated.

Fig.5

	Figure Option ViewDownloadNew Window
	Fig.5 Emission spectra for pure Tm3+ doped bulk glass samples upon excitation at 808 nm LD

There is no emission band at around 544 nm in the pure Tm³⁺ doped tellurite glass, However, the novel emission at 544 nm is clearly observed in Tm³⁺/Ho³⁺ co-doped tellurite glasses although Ho³⁺ ion has no absorption band at around 808 nm^{[6, 14]}. These indicate that the 544 nm emission of Ho³⁺ can only caused by energy transfer (ET) of Tm³⁺ to Ho³⁺ ion. The possible ET processes are as follows^{[12, 15]}:

RET1:Tm³⁺: ¹G₄+ Ho³⁺: ⁵I₈→ Ho³⁺: ⁵F_{2, 3}+Tm³⁺: ³H₆ (6)

RET3: Tm³⁺: ³F₄+ Ho³⁺: ⁵I₈→ Tm³⁺: ³H₆+Ho³⁺: ⁵I₇ (7)

CR2: Tm³⁺: ³F₄+Ho³⁺: ⁵I₇→ Tm³⁺: ³H₆+Ho³⁺: ⁵I₅ (8)

From Figure 3, it can be observed that the energy of the ¹G₄ and ³H₅ levels on the Tm³⁺ ion approach the energy of the ⁵F_{2, 3} and ⁵I₆ levels on the Ho³⁺ ion, respectively. Hence, in Tm³⁺/Ho³⁺ co-doped tellurite glasses, most likely two resonant energy transfer (RET) mechanisms work. In the first mechanism (RET1) the excited Tm³⁺ ions are transferred from ¹G₄ level to ³H₆ level with releasing energy. Ho³⁺ ions absorb the energy and then are excited from ⁵I₈ level to ⁵F_{2, 3} level. Similarly, another pathway for resonant energy transfer (RET3) from Tm³⁺ to Ho³⁺ ions may be as follow: the excited Tm³⁺ ions are transferred from ³H₅ level to ³H₆ level and then the Ho³⁺ ions at ⁵I₈ level are excited to ⁵I₆ level. The energy difference between ³F₄ and ³H₆ levels of Tm³⁺ well matches that between ⁵I₅ and ⁵I₇ levels of Ho³⁺, Tm³⁺ at ³H₄ level decay rapidly to the ³H₆ level via cross relaxation (CR2) process while the Ho³⁺ at ⁵I₇ level are raised to ⁵I₅ level^[13]. Ho³⁺ at ⁵I₅ level can be transferred non-radiatively to ⁵I₆ level by the multiphonon relaxation process.

The intensities of 695, 740 and 474 nm emission in the Tm³⁺/Ho³⁺ co-doped tellurite glass are larger than those in the pure Tm³⁺ doped tellurite glass, respectively, when the glasses samples are all doped at the concentration of 0.5 mol% of Tm₂O₃ (see Fig.1). These indicate that there is an intense backward energy transfer from Ho³⁺ to Tm³⁺ which enhances the fluorescence emission of Tm³⁺. The backward energy transfer from Ho³⁺ to Tm³⁺ is mainly in two ways:

RET2: Ho³⁺: ⁵F_{2, 3}+Tm³⁺: ³H₆→ Tm³⁺: ¹G₄+Ho³⁺: ⁵I₈ (9)

CR3: Ho³⁺: ⁵I₆+Tm³⁺: ³H₄→ Tm³⁺: ¹G₄+Ho³⁺: ⁵I₈ (10)

Firstly, the backward resonant energy transfer (RET2) from Ho³⁺ to Tm³⁺ is possibly taken place when the forward RET1 occurs. Ho³⁺ at the ⁵F_{2, 3} level could be transferred to the ⁵I₈ level and then the Tm³⁺ at the ³H₆ level are excited to the ¹G₄ level. Another backward energy transfer channel is the cross relaxation (CR3) process because the energy difference between ⁵I₆ and ⁵I₈ levels in the Ho³⁺ matches very well that between ¹G₄ and ³H₄ levels in the Tm³⁺. Since Ho³⁺ ion has no absorption band at around 808 nm, the population of ⁵F_{2, 3} level at the Ho³⁺ should be originated from the forward energy transfer from Tm³⁺ to Ho³⁺. Therefore, the speed of forward energy transfer from Tm³⁺ to Ho³⁺must be larger than that of backward energy transfer from Ho³⁺ to Tm³⁺. The intensities of the up-conversion fluorescence peaks of Tm³⁺ should significantly decrease with increasing Ho³⁺ concentration. On the contrary, the intensities of up-conversion fluorescence peaks become stronger with increasing Ho³⁺ concentration (see Fig.1). These indicate that there are other backward energy transfer pathways enhancing emission of the ¹G₄ level at Tm³⁺ besides RET2 process. According the aforementioned, the population of the ¹G₄ level at Tm³⁺ is also reinforced by the cross relaxation process CR2 and CR3. Because of the cross relaxation CR2 and CR3, as well as backward resonant energy transfer RET2, there is a net increase of electron population for the level ¹G₄ of Tm³⁺ in the Tm³⁺/Ho³⁺ co-doped tellurite glass compared to the Tm³⁺ single-doped tellurite glass while the two glasses are conformed with the concentration of Tm³⁺. The net increase of population for the level ¹G₄ of Tm³⁺ is possible to reinforce the multi-phonon relaxation process of this level, emission of 740 nm red light, as well as CU process, and further enhance the population for the level ³F_{2, 3} of Tm³⁺. Therefore, the intensities of red and blue fluorescence in the Tm³⁺/Ho³⁺ co-doped tellurite glass are clearly greater than those in the Tm³⁺ single-doped tellurite glass (see Fig.1).