Abnormal photoluminescence phenomena in (Tb3+, Eu3+) codoped Ga2O3 phosphor

Abnormal photoluminescence phenomena in (Tb3+, Eu3+) codoped Ga2O3 phosphor

Journal of Alloys and Compounds 678 (2016) 448e455 Contents lists available at ScienceDirect Journal of Alloys and Compounds journal homepage: http:...

2MB Sizes 2 Downloads 39 Views

Journal of Alloys and Compounds 678 (2016) 448e455

Contents lists available at ScienceDirect

Journal of Alloys and Compounds journal homepage: http://www.elsevier.com/locate/jalcom

Abnormal photoluminescence phenomena in (Tb3þ, Eu3þ) codoped Ga2O3 phosphor Kenji Sawada, Toshihiro Nakamura*, Sadao Adachi* Division of Electronics and Informatics, Faculty of Science and Technology, Gunma University, Kiryu-shi, Gunma 376-8515, Japan

a r t i c l e i n f o

a b s t r a c t

Article history: Received 27 January 2016 Received in revised form 23 March 2016 Accepted 1 April 2016 Available online 4 April 2016

The structural and optical properties of (Tb3þ, Eu3þ)-codoped b-Ga2O3 crystals were examined using Xray diffraction measurement, photoluminescence (PL) analysis, PL excitation (PLE) spectroscopy, and PL decay measurement. The (Tb3þ, Eu3þ)-codoped samples were synthesized by metal-organic decomposition (MOD). The PL and PLE measurements showed that the (Tb3þ, Eu3þ)-codoped samples exhibit clearly different luminescence properties when excited at wavelengths below or above lex ~ 350 nm, which corresponds to the boundary in wavelength between the 4f 6 / charge transfer state and 4f 6 / 4f 6 transitions to occur in Eu3þ. At lex > 350 nm, efficient energy transfer occurred from Tb3þ to Eu3þ, resulting in the decreased Tb3þ and enhanced Eu3þ emissions. At lex < 350 nm, not only the Tb3þ but also the Eu3þ emissions decreased with increasing Tb3þ/Eu3þ concentration. Furthermore, the PL intensity degradation in the seconds time scale was observed by exciting light at lex < 350 nm. Such unusual PL phenomena at lex < 350 nm seem to be due to an interaction between Tb3þ and Eu3þ with producing nonradiative relaxation channels in the Tb3þ/Eu3þ emission pathway. The temperature dependence of the PL intensity was also measured at T ¼ 20e450 K in increments of 10 K and analyzed using newly developed models. © 2016 Elsevier B.V. All rights reserved.

Keywords: Ga2O3 Phosphor Resonant energy transfer Degradation Lanthanide

1. Introduction The crystal structure of b-form Ga2O3 is monoclinic and comparable to the a form (rhombohedral), which is isostructural with a-Al2O3 [1]. Thus far, only b-Ga2O3 has been easily grown in singlecrystalline form, since it is a high-temperature stable phase. bGa2O3 has a wide band-gap energy of ~5 eV and is chemically and thermally stable, which has practical applications in various functional materials and devices. Rare-earth trivalent ions, such as Eu3þ and Tb3þ, are known to be very efficient activator ions in various hosts for phosphor applications. Excitation energy transfer and cooperative phenomena commonly occur in codoped phosphors and are important concepts to be used for the improvement of phosphor performances [2]. There have been reported many studies on the energy transfer between Tb3þ and Eu3þ in various host materials [3e11]. The majority conclusion obtained from such studies is that an excitation energy transfer occurs from Tb3þ to Eu3þ and between Tb3þ ions.

* Corresponding authors. E-mail addresses: [email protected] (T. Nakamura), [email protected] ac.jp (S. Adachi). http://dx.doi.org/10.1016/j.jallcom.2016.04.004 0925-8388/© 2016 Elsevier B.V. All rights reserved.

An advantage of energy transfer is an increase in the emission intensity of activator (Eu3þ) ions due to sensitizer (Tb3þ) doping. In fact, more than 200 times larger activator emission intensities were observed in various codoped phosphor systems, such as CaCO3:Ce3þ, Tb3þ [12], CaCO3:Ce3þ, Mn2þ [13], and KCl:Ce3þ, Tb3þ [14]. In some (Tb3þ, Eu3þ)-codoped phosphors [3,9], however, decreased rather than increased intensity compared with the singly doped intensity has been observed. Several authors investigated the codoping effects of Tb3þ and Eu3þ in Ga2O3 [15e17]. Sinha and Patra [15] reported that the energy transfer occurs from the b-Ga2O3 nanoparticle host to Tb3þ/ Eu3þ. However, no detailed study was performed on the possibility of energy transfer between Tb3þ and Eu3þ. Cabello et al. [16] observed the characteristic Tb3þ and Eu3þ emissions in (Tb3þ, Eu3þ)-codoped samples with decreased Tb3þ and Eu3þ emission intensities. This fact suggests an energy transfer between Tb3þ and Eu3þ with producing nonradiative relaxation channels in the Tb3þ and Eu3þ emission pathways. Wawrzynczyk et al. [17] studied codoping effects of Tb3þ and Eu3þ on the luminescence properties of cubic g-form Ga2O3 samples. The luminescence from the Tb3þ or Eu3þ singly doped sample showed a broad blue emission band and a series of the sharp emission lines coming from the trivalent ions.

K. Sawada et al. / Journal of Alloys and Compounds 678 (2016) 448e455

After codoping with 1% of Tb3þ and Eu3þ, the sample showed only the broad blue emission band but not any trivalent-ion luminescence. These studies suggest that further investigation needs to make clear energy transfer phenomena between Tb3þ and Eu3þ in Ga2O3 and other hosts. Herein, we report on the structural and photoluminescence (PL) properties of (Tb3þ, Eu3þ)-codoped b-Ga2O3 samples synthesized by the metal organic decomposition (MOD) [18e24]. To examine excitation and relaxation processes of the electrons excited in the uppermost states of the Tb3þ/Eu3þ ions, the PL spectra were measured by excitation at various wavelengths from lex ¼ 266e488 nm for the samples doped with different Tb3þ or Eu3þ concentration. Clearly different luminescence properties, including PL intensity degradation, were observed by excitation at wavelength below or above lex ~350 nm, which corresponded to an excitation via the 4f / 4f5d (or charge transfer state) or 4f / 4f transitions in Eu3þ.

2. Experimental procedure The (Tb3þ, Eu3þ)-codoped Ga2O3 phosphors were synthesized by the MOD [18e24]. A mixture of Ga(RCOO)n, CH3COOC2H5, turpentine, and CxHyOz was supplied from Kojundo Chemical Laboratory Co., Ltd., Japan (Product No.: Ga-03), and used as the starting solution. First, TbCl3$6H2O and Eu2O3 were dissolved in an aqueous acetic acid. Then, the solution was mixed with the MOD solution. The mixed raw solution was at

Ga : Tb : Eu ¼ ð1  x  yÞ : x : y

(1)

in molar ratio. A gelatinous material was obtained by stirring the mixed raw solution. After prebaking in air at 120  C for 10 min, the precursor was calcined on an alumina boat at Tc ¼ 800  C for t ¼ 30 min in air. The reason for the relatively low calcination temperature (Tc ¼ 800  C) was to avoid the influences of energy transfer from the lanthanide ions to Cr3þ, where the Cr3þ ions were unintentionally doped into the Ga2O3 host [18e21,23]. The synthesized phosphors, expressed by the chemical formula (Ga1xyTbxEuy)2O3, were finally grained in an agate mortar. The structural properties of the synthesized phosphors were characterized by X-ray diffraction (XRD) analysis using a RAD-IIC Xray diffractometer (Rigaku) with Cu Ka radiation. PL measurements were performed at T ¼ 300 K using a single monochromator equipped with a charge-coupled device (Princeton Instruments PIXIS 100) and a Nd:YAG laser (MINILITE I, Continuum ElectroOptics, Inc.) at lex ¼ 266 nm with 5 ns pulse duration at 15 Hz, a Nd:YAG laser (STV-01E, Teem Photonics) at lex ¼ 355 nm with 0.3 ns pulse duration at 20 Hz, an Arþ laser continuously operating at lex ¼ 488 nm (Showa Optronics GLG3110), and a 50 W Xe arc lamp (ILC Technology, Inc.) at lex ¼ 300, 395, and 465 nm. PL measurements were also carried out between T ¼ 20 and 450 K in 10-K step at lex ¼ 488 nm using an Arþ laser (Showa Optronics GLG3110). PL excitation (PLE) measurements were performed at 300 K using a monochromator (JASCO CT-25C), a Peltier-device cooled photomultiplier tube (Hamamatsu R375), and a 50 W Xe arc lamp (ILC Technology, Inc.) as the excitation light source. The PL decay time was measured by excitation at lex ¼ 266 and 355 nm (Nd:YAG laser). The signal was detected at 300 K with a Peltier-element-cooled photomultiplier tube (Hamamatsu R375), a multichannel scaler (SR 430, Stanford Research Systems, Inc.), and a preamplifier (SR 445A, Stanford Research Systems, Inc.).

449

3. Results and discussion 3.1. Structural properties Fig. 1 shows the XRD patterns for the Ga2O3 samples doped with x ¼ 0e0.30 (Tb3þ) and y ¼ 0.03 (Eu3þ) followed by calcination at Tc ¼ 800  C. The XRD pattern for b-Ga2O3 taken from the American Society for Testing and Materials (ASTM) card is shown in the lower part of Fig. 1 (#00-041-1103). The diffraction pattern for the undoped sample is in agreement with the ASTM pattern. The bphase Ga2O3 crystal has a monoclinic structure with the space 3  C2=m[1]. This phase is one of the five well-known group of C2h forms of gallium oxide: a-, c-, d-, and ε-Ga2O3 [25], and all of these polymorphs can be converted to b phase at high temperature [26]. After doping with Tb3þ and Eu3þ, the samples produced more diffusive peaks than the undoped sample. The Ga2O3 samples calcined at Tc  600  C were somewhat amorphous; On the other hand, those at Tc  700  C, showed XRD patterns consisting of many diffraction peaks from various crystallographic planes of b-Ga2O3 [18e20]. Using the DebyeScherrer expression and the (111) diffraction peak at 2q ~ 35 in Fig. 1 (Tc ¼ 800  C), the crystalline size of the undoped b-Ga2O3 sample was estimated to be ~15 nm. The doping-induced increase in the diffraction peak widths observed in Fig. 1 may mainly come from the effects of lattice distortion rather than the decreased crystalline size of the Tb3þ/Eu3þ doped samples. Further study needs to confirm this consideration. For the samples doped with x ¼ 0.30, the XRD pattern is clearly different from that of b-Ga2O3. This fact is caused by the formation of cubic terbium garnet, Tb3Ga5O12 (TGG), with the space group of O10  Ia3d [21,23]. The XRD pattern of TGG taken from the ASTM h card (#01-088-0575) is shown in the top of Fig. 1. The sample synthesized with x ¼ 0.20 (y ¼ 0.03) in Fig. 1 is understood to be a mixture of b-Ga2O3 and TGG. In the following, we discuss on the optical properties of the b-Ga2O3:Tb3þ, Eu3þ crystals. Detailed structural and PL properties of TGG:Eu3þ can be found in Refs. [21,23].

Fig. 1. XRD patterns for the MOD-synthesized Ga2O3 samples doped with x ¼ 0e0.30 (Tb3þ) and y ¼ 0.03 (Eu3þ), together with that for the undoped (x ¼ y ¼ 0) Ga2O3 sample. The XRD patterns for b-form Ga2O3 and cubic TGG taken from the American Society for Testing and Materials (ASTM) cards are also shown in the bottom (#00-0411103) and top parts (#01-088-0575), respectively.

450

K. Sawada et al. / Journal of Alloys and Compounds 678 (2016) 448e455

3.2. PL spectra: excitation wavelength dependence Fig. 2 shows the PL spectra for the b-Ga2O3 samples doped with x ¼ 0.03 (Tb3þ) and y ¼ 0e0.10 (Eu3þ). The PL spectra were measured by excitation at three different wavelengths: (a) lex ¼ 266 nm, (b) 355 nm, and (c) 488 nm. The singly doped sample exhibits a series of the sharp PL peaks at ~490, ~545, ~590, and ~625 nm coming from the Tb3þ 4f 8 (5D0) / 4f 8 (7FJ) transitions with J ¼ 6, 5, 4, and 3, respectively [20]. By doping with Eu3þ, a new series of the PL peaks appears at ~595, ~615, and ~700 nm due to the Eu3þ 4f 6 (5D0) / 4f 6 (7FJ) transitions with J ¼ 1, 2, and 4, respectively ([19], see also Fig. 7 below). The strongest Tb3þ emission peak at ~545 nm in Fig. 2 gradually decreases with increasing Eu3þ concentration (y), regardless of the excitation wavelength (lex ¼ 266, 355, or 488 nm). Fig. 3 shows, as an example, the Tb3þ emission intensity IPL at lem ~545 nm vs y plots obtained from the PL spectra in Fig. 2(a) (lex ¼ 266 nm). As shown in Fig. 3, IPL decreases with increasing y in the manner IPL f ya with a ~1.0. The a values for lex ¼ 355 nm [Fig. 2(b)] and 488 nm [Fig. 2(c)] are also determined to be almost the same as that for lex ¼ 266 nm (a ~ 1.0). Further, the Eu3þ emission intensity in Fig. 2 sublinearly increases with increasing y and then shows saturation or decrease with the further increase of y. If any efficient energy transfer occurs from Eu3þ to Tb3þ, one can expect a strong Tb3þ emission and its intensity would increase with increasing y. However, no increase in the Tb3þ luminescence intensity has been observed in our (Tb3þ, Eu3þ)-codoped samples (x ¼ 0.03, y s 0). This may be due to nonradiative losses of the excited electrons via the exchange interaction, radiation reabsorption, and others. The PL spectra for the b-Ga2O3 samples doped with y ¼ 0.03 (Eu3þ) and x ¼ 0e0.10 (Tb3þ) are shown in Fig. 4. The PL spectra were measured by excitation at three different wavelengths: (a) lex ¼ 266 nm, (b) 355 nm, and (c) 488 nm. The sharp PL peaks at ~595, ~615, and ~700 nm in the Eu3þ singly doped (x ¼ 0) sample are attributed to the Eu3þ 4f 6 (5D0) / 4f 6 (7FJ) transitions with J ¼ 1, 2, and 4, respectively [19]. The Eu3þ emission intensity IPL vs x data taken from Fig. 4 are shown in Fig. 5. The Eu3þ emission intensity observed by excitation at lex ¼ 488 nm [Fig. 5(a)] increases with increasing Tb3þ concentration (x). The dashed line in Fig. 5(a) shows the result calculated using IPL(x)/IPL(0) ¼ 250x þ 1.0. A slight increase in IPL with increasing x is also observed by excitation at lex ¼ 355 nm

Fig. 2. Room-temperature PL spectra for the b-Ga2O3 samples doped with x ¼ 0.03 (Tb3þ) and y ¼ 0e0.10 (Eu3þ). The PL spectra were measured by excitation at three different wavelengths: (a) lex ¼ 266 nm, (b) 355 nm, and (c) 488 nm. The sample synthesized with y ¼ 0 corresponds to that singly doped with Tb3þ (x ¼ 0.03), and thus their PL peaks come only from the Tb3þ intra-4f-shell transitions.

Fig. 3. Tb3þ emission intensity IPL at lem ~545 nm vs y plots for the (Ga0.97yTb0.03Euy)2O3 samples. The plotted data were obtained from the PL spectra at lex ¼ 266 nm in Fig. 2(a). The solid line represents the relation of IPL and y, given by IPL f y1.0.

Fig. 4. Room-temperature PL spectra for the b-Ga2O3 samples doped with x ¼ 0e0.10 (Tb3þ) and y ¼ 0.03 (Eu3þ). The PL spectra were measured by excitation at three different wavelengths: (a) lex ¼ 266 nm, (b) 355 nm, and (c) 488 nm. The sample synthesized with x ¼ 0 corresponds to that singly doped with Eu3þ (y ¼ 0.03), and thus their PL peaks come only from the Eu3þ intra-4f-shell transitions.

Fig. 5. Eu3þ emission intensity IPL at lem ~703 nm vs x plots for the (Ga0.97xTbxEu0.03)2O3 samples obtained from the PL spectra at (a) lex ¼ 488 nm and (b) lex ¼ 355 and 266 nm in Fig. 4. The dashed line in (a) shows the result calculated using IPL(x)/ IPL(0) ¼ 250x þ 1.0.

[Fig. 5(b)]. On the contrary, the IPL vs x data in Fig. 5(b) (lex ¼ 266 nm) show gradual decrease with increasing x.

K. Sawada et al. / Journal of Alloys and Compounds 678 (2016) 448e455

Fig. 6 shows the PL and PLE spectra for (a) the b-Ga2O3:Tb3þ (x ¼ 0.03), (b) b-Ga2O3:Eu3þ (y ¼ 0.03), and (c) b-Ga2O3:Tb3þ, Eu3þ samples (x ¼ y ¼ 0.03) measured by excitation at lex ¼ 488 nm (PL) and by monitoring at lem ¼ 543 nm (Tb3þ) [Fig. 6(a)] and lem ¼ 703 nm (Eu3þ) [Fig. 6(b) and (c)] (PLE). The dashed line in Fig. 6(a) also shows the PLE spectrum for TGG [23]. In Tb3þ-activated phosphors [12,14,20,23], many PLE peaks due to the 4f 8 (7F6) / 4f 8 (5DI) transitions are usually observed in the 325e500 nm spectral region [Fig. 6(a)]. Several PLE peaks can also be observed below ~300e325 nm and assigned to the spinforbidden 4f 8 (7F6) / 4f 75d (9E) and spin-allowed 4f 8 (7F6) / 4f 75d (7E) transitions in Tb3þ [12,14,20,23]. In Eu3þ-activated phosphors [19,23,24,27,28], several sharp PLE peaks due to the 4f 6 (7F0) / 4f 6 (5DI, 5L6, 5G2) transitions are observed in the wavelength region of ~350e500 nm [19,23,24,27,28] [see Fig. 6(b)]. A large absorption peak at ~300 nm in Fig. 6(b) is due to the 4f 6 (7F0) / charge transfer state (CTS) transitions. The CTS (2poxide) levels in Eu3þ, are usually lower in energy than its 4f 55d states (9E and 7E) [29]. Thus, only the CTS band, but not any 4f 55d-related peaks, can be observed in the optical spectra of Eu3þ-activated phosphors. It should be noted that its ordering is in direct contrast to that in Tb3þ (i.e., the 4f 75d states in Tb3þ are lower in energy than its CTS levels) [29]. It is interesting to note that the PL and PLE spectra for the Eu3þ singly doped sample (Fig. 6) are almost the same as those for the (Tb3þ, Eu3þ)-codoped sample. The only difference is that the Eu3þ emission intensity in the latter sample is about 10 times larger than that in the former sample [see also Fig. 5(a)]. This is to be due to the effects of an efficient energy transfer from Tb3þ to Eu3þ in the latter sample by excitation at lex ¼ 488 nm. The energy-level scheme for the Eu3þ ion, together with the diffuse reflectance absorption (a) and PL spectra for the bGa2O3:Tb3þ, Eu3þ sample (x ¼ y ¼ 0.03), is shown in Fig. 7. The a spectrum was measured by means of diffuse reflectance spectroscopy using a spectrometer (JASCO V-570), whereas the PL spectra were obtained by excitation at lex ¼ 266 and 488 nm. The position of Eg at ~4.8 eV for b-Ga2O3 is shown by the vertical arrow. Various 4f N states for Tb3þ (N ¼ 8) and Eu3þ (N ¼ 6) are also shown by the horizontal lines. The 4f 8 / 4f 75d and 4f 8 / CTS transitions in

451

Fig. 7. (a) Electronic energy-level scheme for Tb3þ and Eu3þ in b-Ga2O3:Tb3þ, Eu3þ phosphor. (b) Room-temperature diffuse reflectance absorption (a) and PL spectra for the b-Ga2O3:Tb3þ, Eu3þ sample (x ¼ y ¼ 0.03). The a spectrum was measured by means of diffuse reflectance spectroscopy, whereas the PL spectra were obtained by excitation at lex ¼ 488 and 266 nm. An efficient resonant energy transfer (RET) occurs from Tb3þ to Eu3þ by excitation at lex ¼ 488 nm, whereas by excitation at lex ¼ 266 nm the luminescence quenching occurs via an interaction between neighbor Tb3þ and Eu3þ ions followed by the nonradiative relaxation (NRR) events. The bold vertical arrow in the top part of (a) indicates the boundary between the [4f 6 (5D0) / CTS (4f 5d)] and [4f 6 5 ( D0) / 4f 6 (7FJ)] transitions to occur at l ~ 350 nm in Eu3þ.

Fig. 7(a) occur at wavelengths 320 nm (Tb3þ) and 350 nm (Eu3þ). The electrons excited at the high photon energy (lex ¼ 266 nm) will experience more nonradiative relaxation events than those excited at the low photon energy (lex ¼ 488 nm). Such nonradiative relaxation events would occur more dramatically if samples were more highly doped, known as the concentration quenching. As a result, only the decreased Eu3þ emission intensity was observed with increasing x (Tb3þ) at lex ¼ 266 nm [Figs. 4(a) and 5(b)]. After excitation at lex ¼ 488 nm, only the limited excitation states can involve in the Eu3þ luminescence process. These states are: 5D4 (Tb3þ; ~488 nm), 5D1 (Eu3þ; ~529 nm, see Ref. [28]) and 5D0 (Eu3þ; ~580 nm) [Fig. 7(b)]. Note that the 5D4 state in Tb3þ occurring at ~2.54 eV (~488 nm) accidentally coincides with an oscillation wavelength of Arþ laser at 488 nm. Thus, an excitation at lex ¼ 488 nm enables to induce energy transfer and cross relaxation between Tb3þ and Eu3þ. The possible energy-level scheme responsible for these mechanisms are: 5

D4 ðTbÞ þ

7

F0 ðEuÞ /

7

F6;5;4 ðTbÞ þ

5

D1;0 ðEuÞ± DEJp ; (2)

where DEJp is energy mismatched and can be satisfied by the absorption or emission of the phonons or by the involvement of energy levels of other ions in a many-body process. In Fig. 6(a), the emission of Tb3þ at ~543 nm is dominated. The emission of Tb3þ decreases, accompanied by an increase in the Eu3þ emission intensity, with increasing y (Eu3þ) [see Fig. 2(c); also cf. Fig. 6(a)e(c)]. This is caused by the enhanced energy transfer from Tb3þ to Eu3þ.

Fig. 6. Room-temperature PL and PLE spectra for (a) the b-Ga2O3:Tb3þ (x ¼ 0.03), (b) bGa2O3:Tb3þ (y ¼ 0.03), and (c) b-Ga2O3:Tb3þ, Eu3þ samples (x ¼ y ¼ 0.03) measured by excitation at lex ¼ 488 nm [PL; (a)e(c)] and by monitoring at (a) lem ¼ 543 nm (Tb3þ emission) [(a)] and 703 nm (Eu3þ emission) [(b) and (c)] (PLE). The dashed line in (a) also shows the PLE spectrum for TGG [23].

3.3. PL decay characteristics To investigate whether an interaction between Tb3þ and Eu3þ actually occurs in the (Tb3þ, Eu3þ)-codoped sample, we measured the PL decay curves for the b-Ga2O3 samples doped with x ¼ 0.03

452

K. Sawada et al. / Journal of Alloys and Compounds 678 (2016) 448e455

The open and solid circles in Fig. 8(b) show the results calculated using Eqs. (5) and (6), respectively. The effective decay time teff can be calculated using

Z



tIðtÞdt

teff ¼ Z0 ∞

(7) IðtÞdt

0

where I(t) is the PL intensity at time t. The teff values calculated from Fig. 8(a) using Eq. (7) are plotted by the open triangles in ð1Þ Fig. 8(b). We can see that teff has nearly the same value as tav defined by Eq. (6). The experimental data in Fig. 8 indicate that the faster the decay time the larger the Eu3þ concentration y. This means that the energy transfer actually occurred from Tb3þ to Eu3þ in the (Tb3þ, Eu3þ)-codoped b-Ga2O3 samples. The energy transfer efficiency h(y) from Tb3þ to Eu3þ can be calculated by Fig. 8. (a) Room-temperature PL decay curves obtained by excitation at lex ¼ 355 nm and by monitoring at lem ¼ 543 nm (Tb3þ) in the (Ga0.97yTb0.03Euy)2O3 samples with y ¼ 0, 0.001, 0.003, 0005, 0.01, 0.02, 0.03, 0.05, and 0.10. The solid lines show the results calculated using Eq. (3). (b) Average and effective times vs y plots for the (Ga0.97yTb0.03Euy)2O3 samples obtained from (a). The average times are calculated from Eq. (5) (open circles) and Eq. (6) (solid circles), whereas the effective times are obtained from Eq. (7) (open triangles). The energy transfer efficiency h(y) from Tb3þ to Eu3þ, calculated using Eq. (8), vs y plots are also shown in (b).

(Tb3þ) and y as a parameter (y ¼ 0e0.10). Fig. 8(a) shows the results of this experiment. The PL decay curves were measured at lem ¼ 543 nm (Tb3þ) by excitation at lex ¼ 355 nm (Nd:YAG laser). An evidence of the Tb3þEu3þ ion interaction can be understood from Fig. 8(a). Each solid line in Fig. 8(a) is calculated using the triple-exponential function [30]:

IðtÞ ¼

  t ai exp 

3 X

ti

i¼1

(3)

with 3 X

ai ¼ 1

(4)

i¼1

For examples, we obtained the decay parameters for the samples with x ¼ 0.03 and y ¼ 0 as a1 ¼ 0 (t1 ¼ 0 ms), a2 ¼ 0.70, t2 ¼ 0.70 ms, a3 ¼ 0.30, and t3 ¼ 1.55 ms (double-exponential function); x ¼ 0.03 and y ¼ 0.005 as a1 ¼ 0.35, t1 ¼ 0.09 ms, a2 ¼ 0.55, t2 ¼ 0.50 ms, a3 ¼ 0.10, and t3 ¼ 1.30 ms; x ¼ 0.03 and y ¼ 0.10 as a1 ¼ 0.50, t1 ¼ 0.05 ms, a2 ¼ 0.42, t2 ¼ 0.25 ms, a3 ¼ 0.08, and t3 ¼ 0.90 ms. These results suggest that the smaller the y-value sample, the lower the fast decay-component product a1  t1. In fact, in the limit y / 0 the sample has no fast decay component (a1 ¼ 0). We can thus understand that the higher doping of Eu3þ (y) produces the more striking non-radiative decay pathway appearing in the fast decay-time regime of the charge recombination (relaxation) process. Average decay times tav are calculated using [30].

P3

ð0Þ tav ¼ Pi¼1 3

ai ti

i¼1

ai

¼

3 X

ai ti

(5)

i¼1

or

P3

ð1Þ tav ¼ Pi¼1 3

i¼1

ai t2i ai ti

(6)

hðyÞ ¼ 1 

teff ðyÞ teff ð0Þ

(8)

where teff(0) and teff(y) represent the effective decay times of the Tb3þ donor in the absence and presence of the Eu3þ acceptor, respectively. The results of this calculation are shown in Fig. 8(b). The energy transfer between Tb3þ and Eu3þ occurs due to the different mechanisms, such as the exchange and multipolar interactions. Based on the Dexter's energy transfer formula [31], the nature of energy transfer can be understood from

teff ð0Þ s=3 fy teff ðyÞ

(9)

with s ¼ 3, 6, 8, and 10 correspond to the exchange, dipoledipole, dipolequadrupole, and quadrupolequadrupole interactions, respectively. The values of teff(y) taken from Fig. 8(b) indicated that the exchange interaction (s ¼ 3) is responsible for the Tb3þ / Eu3þ energy transfer in the b-Ga2O3:Tb3þ, Eu3þ samples. To investigate the Tb3þ / Eu3þ energy transfer process in more detail, we examined the decay curves for the b-Ga2O3 samples doped with x ¼ 0.03 (Tb3þ) and y (Eu3þ) as a parameter (y ¼ 0e0.10) at excitation with lex ¼ 266 and 355 nm (Nd:YAG laser) and by monitoring at lem ¼ 703 nm (Eu3þ). These results are shown in Fig. 9. As in Fig. 8(a), increasing y deviated the decay curves from the single-exponential function. It should be noted that by excitation at lex ¼ 355 nm [Fig. 9(b)] the decay curves have an enhancement component at the early stage and then exhibit usual decay process. The enhancement process indicates that the photo-excited electrons in the Tb3þ 5D4 state can transfer to the Eu3þ 5D1 (and 5D0) state. This enhancement process occurs faster as more Eu3þ ions are doped into the Tb3þ doped samples. Such initial enhancement process disappears at y  0.02, as shown in the inset of Fig. 9(b). No enhancement process can be found in the decay curves when excited at lex ¼ 266 nm [Fig. 9(a)]. The decreased Eu3þ emission intensity at lex < 350 nm [Figs. 4(a) and 5(b)] should be due to an interaction between the Tb3þ and Eu3þ ions, known as the killer ion effect [2,32]. 3.4. PL intensity degradation at lex < 350 nm Recently, we observed that the red-emitting TGG:Eu3þ phosphor exhibits remarkable degradation in the PL intensity under UV light exposure in the seconds time scale [33]. The phenomenon was reversible and, therefore, the emission efficiency slowly returned to its initial value after keeping in the dark or in room light. To check

K. Sawada et al. / Journal of Alloys and Compounds 678 (2016) 448e455

I0 þ c0 ¼ 1:0

Fig. 9. Room-temperature PL decay curves obtained by excitation at (a) lex ¼ 266 nm and (b) 355 nm and by monitoring at lem ¼ 703 nm (Eu3þ) in the (Ga0.97yTb0.03Euy)2O3 samples with y ¼ 0.001, 0.003, 0005, 0.01, 0.02, 0.03, 0.05, and 0.10.

whether such phenomenon occurs in the (Tb3þ, Eu3þ)-codoped bGa2O3 phosphor, we measured PL intensity I(t) as a function of excitation light exposure time t at various wavelengths. Fig. 10 shows the I(t) at lem ¼ 703 nm vs t data at lex from 266 to 488 nm. The Xe arc lamp was used as the continuous excitation light source. As evidenced from Fig. 10, no clear intensity change is observed when excited at lex  395 nm. However, we observed dramatic decrease in the PL intensity with increasing t at lem  355 nm. As in TGG:Eu3þ [33], the degraded PL intensity slowly returned to its initial value after blocking excitation light. The solid lines in Fig. 10 represent the fitted results using

 IðtÞ ¼ I0 exp



t



t

þ c0

453

(11)

Examples of the fit-determined parameters for lex ¼ 355 nm are: I0 ¼ 0.14 and t ¼ 50 s with c0 ¼ 0.86. Since no decreased PL intensity was observed under excitation at lex  395 nm, an agreement with the experimental data can be achieved by introducing t ¼ ∞ into Eq. (10). The same PL intensity degradation phenomena were also observed in CaTiO3:Eu3þ [33]. The degradation constant t vs lex plots for the b-Ga2O3:Tb3þ, 3þ Eu phosphor are shown in Fig. 11(a). The PLE spectrum, the same as that in Fig. 6(c), is also shown in Fig. 11(b). No any degradation in the PL intensity occurs by excitation at lex > 350 nm. When excited at lex  350 nm, an exponential decay in the PL intensity with respect to t can be observed. The (4f 6 4 4f 6)/(4f 6 4 CTS) [or (4f 8 4 4f 8)/(4f 8 4 4f 75d)] transition boundary takes place in Eu3þ (or Tb3þ) at wavelength of ~350 nm. Thus, there is a correlation between no efficient energy transfer among Tb3þ and Eu3þ by excitation at lex  350 nm (Sec. 3.2) and excitation light-induced PL intensity degradation occurring at the same excitation light wavelengths (Sec. 3.4). Note that the excitation wavelength of lex ~350 nm corresponds to the boundary between the [4f 6 (5D0) / CTS (4f 5d)] and [4f 6 (5D0) / 4f 6 (7FJ)] transitions to occur in Eu3þ. Because the degradation phenomena occurred in the seconds time scale, no clear decrease in the Eu3þ emission intensity was observed when excited by the short-pulse Nd:YAG lasers at lex ¼ 266 nm (5 ns pulse duration at 15 Hz) and at lex ¼ 355 nm (0.3 ns pulse duration at 20 Hz) [34]. Moreover, no degradation was observed when excited at longer wavelengths than ~350 nm (see Fig. 11) even if the light source was “continuous” as the Arþ laser we used. Therefore, it is considered that the PL data in Figs. 2e9 do not include any vagueness or large error arising from the excitation light-induced degradation phenomena (Figs. 10 and 11). 3.5. Lattice temperature dependence of PL intensity Fig. 12 shows the PL spectra for the (Tb3þ, Eu3þ)-codoped bGa2O3 phosphor with x ¼ y ¼ 0.03 measured at lex ¼ 488 nm at T ¼ 20e440 K with 60-K increments. The corresponding Eu3þ emission intensity IPL vs T1 plots are shown in Fig. 13(a), together with those obtained by excitation at lex ¼ 266 nm in Fig. 13(b). Note that the peculiar temperature dependence of IPL observed in Fig. 13(a) cannot be successfully explained by any conventional

(10)

with

Fig. 10. Plots of Eu3þ emission intensity at lem ¼ 703 nm vs continuous irradiation time of the excitation light for the (Ga0.94Tb0.03Eu0.03)2O3 sample (x ¼ y ¼ 0.03) at 300 K. The excitation wavelengths from a Xe arc lamp were varied in the range of lex ¼ 266e488 nm. The solid lines show the results calculated using Eq. (10).

Fig. 11. (a) Degradation constant t vs excitation light wavelength (lex) plots for the (Ga0.94Tb0.03Eu0.03)2O3 sample obtained from the PL data in Fig. 10. The degradation constants are determined by fitting with Eq. (10). (b) PLE spectrum measured by monitoring at lem ¼ 703 nm in the (Ga0.94Tb0.03Eu0.03)2O3 sample. The bold vertical arrow in the top part of (a) indicates the boundary between the [4f 6 (7F0) / CTS (4f 5 d)] and [4f 6 (7F0) / 4f 6 (5DI, 5L6, etc.)] transitions to occur at l ~ 350 nm in Eu3þ.

454

K. Sawada et al. / Journal of Alloys and Compounds 678 (2016) 448e455

Boltzmann constant. The first and second terms in the bracket of Eq. (12) correspond to the Stokes and anti-Stokes components, respectively. Eq. (12) can then be written as

(Ga0.94Tb0.03Eu0.03)2O3

PL iintensity (arb. units)

T (K) 20

IPL ðTÞ ¼

80

260 320

440

600

650

700

 1þ

2 expðhn=kB TÞ  1



(15)

200

550

I0    ai exp  Eqi kB T

i

140

380

500



P

750

Wavelength (nm) Fig. 12. Temperature dependence of the PL spectra for the (Ga0.94Tb0.03Eu0.03)2O3 sample (x ¼ y ¼ 0.03) measured by excitation at lex ¼ 488 nm and at T ¼ 20e440 K with 60-K increments.

The solid line in Fig. 13(a) shows the best-fit result using Eq. (15). The fit-determined parameters are: I0 ¼ 0.49, a1 ¼ 1.0  104, Eq1 ¼ 0.35 eV, a2 ¼ 6.0, Eq2 ¼ 45 meV, and hn ¼ 18 meV. The hn value of 18 meV seems to be smaller than the G-point optical phonon energies in b-Ga2O3, but falls in the experimental Raman frequency range [35,36]. The increase in IPL with increasing T at >50 K can be explained by the Stokes and anti-Stokes activation processes in Eq. (15). The decrease in IPL with increasing T at >300 K can also be explained by the thermal quenching energy of Eq1 ¼ 0.35 eV. We next consider simple model based on a charge transfer process. This process may take place from a trap or a reservoir state (Et) to Eu3þ by

IPL ðTÞ ¼



P

I0    ai exp  Eqi kB T





 Et 1 þ At exp  kB T

i

(16) The charge transfer on the second term in the square bracket of Eq. (16) can be activated by thermal energy. The dashed line in Fig. 13(a) shows the result calculated using Eq. (16) with I0 ¼ 0.51, a1 ¼ 1.0  104, Eq1 ¼ 0.35 eV, a2 ¼ 60, Eq2 ¼ 45 meV, At ¼ 80, and Et ¼ 37 meV. The increased IPL at T from 50 to 300 K can be explained by Eq. (16), but its fitting quality is poorer than that using Eq. (15). At present, however, we cannot say which model is actually occurring in the present phosphor, Eq. (15) or Eq. (16). Interestingly, the IPL data obtained by excitation at lex ¼ 266 nm in Fig. 13(b) show no peculiar temperature dependence of IPL. Thus, the IPL vs T1 data can be well explained by the conventional thermal quenching model that is obtained by neglecting the second term in the bracket of Eq. (15), namely,

IPL ðTÞ ¼ 3þ

Fig. 13. Eu emission intensity vs reciprocal temperature (1/T) plots obtained by excitation at (a) lex ¼ 488 nm (Arþ laser) and (b) 266 nm (Nd:YAG laser) for the (Ga0.94Tb0.03Eu0.03)2O3 sample (x ¼ y ¼ 0.03) at T ¼ 20e450 K (see, e.g., Fig. 12). The solid and dashed lines in (a) show the results calculated using Eqs. (15) and (16), respectively, whereas the solid line in (b) is obtained from Eq. (17).

theoretical model. Because the Eu3þ 4f 6 / 4f 6 transition is parity forbidden, we consider that the luminescence intensity can be further gained by the activation of vibronic quanta with an energy hn. The corresponding luminescence intensity is given by

IPL ðTÞ ¼



P

  I0    nþ  p þ np ai exp  Eqi kB T

(12)

with

np ¼

1 expðhn=kB TÞ  1

P

I0    ai exp  Eqi kB T

(17)

i

The solid line in Fig. 13(b) represents the result calculated using Eq. (17) with I0 ¼ 1.0, a1 ¼ 1.2  105, Eq1 ¼ 0.35 eV, a2 ¼ 9.0, and Eq2 ¼ 45 meV. The thermal quenching energies are found to be the same as those in Fig. 13(a). The different PL behaviors observed below and above lex ~350 nm suggest that the radiative recombination/non-radiative relaxation processes of the excited electrons in the b-Ga2O3:Tb3þ, Eu3þ system may be strongly connected with the fundamental vibronic quanta. Further study needs to completely understanding these unique PL properties. 4. Conclusions

i

1 1 n± p ¼ ± þ np 2 2



(13)

(14)

where Eqi is the thermal quenching (activation) energy and kB is the

We synthesized the (Tb3þ, Eu3þ)-codoped Ga2O3 samples by the MOD and subsequent calcination at Tc ¼ 800  C. The dopant concentrations were varied in the ranges of x ¼ 0e0.10 (Tb3þ) and y ¼ 0e0.10 (Eu3þ) in (Ga1xyTbxEux)2O3. The samples, examined by the XRD measurement, were of the b form (monoclinic). The luminescence properties were investigated by PL analysis, PLE spectroscopy, and PL decay measurement, by varying an excitation wavelength from lex ¼ 266e488 nm. It was found that the bGa2O3:Tb3þ, Eu3þ samples exhibit clearly different PL behaviors

K. Sawada et al. / Journal of Alloys and Compounds 678 (2016) 448e455

when exciting at lex below or above ~350 nm, where this wavelength of lex ~350 nm corresponds to the boundary between the (4f 6 / CTS) and (4f 6 / 4f 6) transitions to occur in Eu3þ. By excitation at lex > 350 nm, we observed efficient energy transfer from Tb3þ to Eu3þ, resulting in the decreased Tb3þ and enhanced Eu3þ emissions. When excited at lex < 350 nm, not only the Tb3þ but also the Eu3þ emission intensities decreased with increasing Tb3þ/Eu3þ ion concentration. The PL intensity degradation in the seconds time scale was also observed by continuously exciting at lex < 350 nm. Such unique PL properties are thought to be due to an interaction between Tb3þ and Eu3þ with producing nonradiative relaxation channels in the Tb3þ/Eu3þ emission pathway when excited at lex < 350 nm. An interaction between Tb3þ and Eu3þ was confirmed by the PL decay studies. The temperature dependence of the Eu3þ emission intensity was also measured at T ¼ 20e450 K in increments of 10 K and analyzed on the basis of newly developed models. Acknowledgments This work was supported by a Grant-in-Aid for Exploratory Research (25630120) from the Ministry of Education, Culture, Sports, Science and Technology, Japan. The authors would like to thank T. Miyazaki and H. Oike for their experimental supports and useful discussion. References [1] S. Geller, J. Chem. Phys. 33 (1960) 676. [2] E. Nakazawa, in: W.M. Yen, S. Shionoya, H. Yamamoto (Eds.), Phosphor Handbook, CRC, Boca Raton, 2007, p. 99. [3] F. Li, H. Liu, S. Wei, W. Sun, L. Yu, J. Rare Earths 31 (2013) 1063. [4] L. Jin, X. Du, X. Lei, L. Ren, Y. Feng, W. Chen, Appl. Phys. A 114 (2014) 631. [5] A. Gupta, N. Brahme, D.P. Bisen, J. Lumin. 155 (2014) 112.

455

[6] T. Jiang, X. Yu, X. Xu, H. Yu, D. Zhou, J. Qiu, Opt. Mater. 36 (2014) 611. [7] Y. Jiang, Y. Liu, G. Liu, X. Dong, J. Wang, W. Yu, Q. Dong, Opt. Mater. 36 (2014) 1865. , P. Marie, C. Frilay, X. Portier, Thin Solid Films [8] C. Davesnne, A. Ziani, C. Labbe 553 (2014) 33. _ [9] J. Pisarska, A. Kos, M. Sołtys, L. Zur, W.A. Pisarski, J. Non-Cryst. Solids 388 (2014) 1. [10] K. Pavani, J.S. Kumar, L.R. Moorthy, J. Alloys Compd. 586 (2014) 722. ~es, G.S. Maciel, J. Alloys Compd. 618 (2015) [11] N. Rakov, S.A. Vieira, R.B. Guimara 127. [12] J. Nara, S. Adachi, ECS J. Solid State Sci. Technol. 2 (2013) R135. [13] J. Nara, S. Adachi, J. Appl. Phys. 113 (2013) 033519. [14] Y. Tosaka, S. Adachi, J. Lumin. 156 (2014) 157. [15] G. Sinha, A. Patra, Chem. Phys. Lett. 473 (2009) 151. [16] G. Cabello, A. Araneda, L. Lillo, C. Caro, C. Venegas, M. Tejos, B. Chornik, Solid State Sci. 27 (2014) 24. [17] D. Wawrzynczyk, M. Nyk, M. Samoc, Chem. Phys. 456 (2015) 73. [18] Y. Tokida, S. Adachi, J. Appl. Phys. 112 (2012) 063522. [19] Y. Tokida, S. Adachi, Jpn. J. Appl. Phys. 52 (2013) 101102. [20] Y. Tokida, S. Adachi, ECS J. Solid State Sci. Technol. 3 (2014) R100. [21] K. Sawada, S. Adachi, ECS J. Solid State Sci. Technol. 3 (2014) R238. [22] T. Tsuchiya, T. Matsuura, K. Shinoda, T. Nakajima, J. Akimoto, Y. Idemoto, Jpn. J. Appl. Phys. 53 (2014) 05FB14. [23] K. Sawada, S. Adachi, J. Lumin. 165 (2015) 138. [24] Y. Onishi, S. Adachi, ECS J. Solid State Sci. Technol. 4 (2015) R153. [25] R. Roy, V.G. Hill, E.F. Osborn, J. Am. Chem. Soc. 74 (1952) 719. € llbauer, E. Born, H. Meixner, J. Am. Ceram. Soc. 80 (1997) [26] M. Fleischer, L. Ho 2121. [27] T. Arai, S. Adachi, J. Lumin. 153 (2014) 46. [28] T. Orihashi, S. Adachi, J. Alloys Compd. 646 (2015) 1116. [29] C.K. Jørgensen, R. Pappalardo, E. Rittershaus, Z. Naturforsch 20a (1964) 54. [30] J.R. Lakowicz, Principles of Fluorescence Spectroscopy, Springer, Baltimore, 2006. [31] D.L. Dexter, J. Chem. Phys. 21 (1953) 836. [32] S. Shionoya, in: W.M. Yen, S. Shionoya, H. Yamamoto (Eds.), Phosphor Handbook, CRC, Boca Raton, 2007, p. 247. [33] K. Sawada, S. Adachi, J. Appl. Phys. 118 (2015) 103106. [34] K. Sawada, T. Nakamura, S. Adachi, ECS J. Solid State Sci. Technol. 5 (2016) R67. [35] D. Machon, P.F. McMillan, B. Xu, J. Dong, Phys. Rev. B 73 (2006) 094125. [36] T. Onuma, S. Fujioka, T. Yamaguchi, Y. Itoh, M. Higashiwaki, K. Sakaki, T. Masui, T. Honda, J. Cryst. Growth 401 (2014) 330.