Down- and up-conversion emissions in Er3+–Yb3+ codoped TeO2–ZnO–ZnF2 glasses

Down- and up-conversion emissions in Er3+–Yb3+ codoped TeO2–ZnO–ZnF2 glasses

Journal of Luminescence 158 (2015) 142–148 Contents lists available at ScienceDirect Journal of Luminescence journal homepage: www.elsevier.com/loca...

900KB Sizes 4 Downloads 40 Views

Journal of Luminescence 158 (2015) 142–148

Contents lists available at ScienceDirect

Journal of Luminescence journal homepage: www.elsevier.com/locate/jlumin

Down- and up-conversion emissions in Er3 þ –Yb3 þ codoped TeO2–ZnO–ZnF2 glasses A. Miguel a, M.A. Arriandiaga b, R. Morea c, J. Fernandez a,d, J. Gonzalo c, R. Balda a,d,n a

Departamento de Física Aplicada I, Escuela Superior de Ingeniería, Universidad del País Vasco UPV/EHU, Alda. Urquijo s/n, 48013 Bilbao, Spain Departamento de Física Aplicada II, Facultad de Ciencia y Tecnología, Universidad del País Vasco UPV/EHU, Apartado 644, Bilbao, Spain c Instituto de Optica, Consejo Superior de Investigaciones Científicas CSIC, Serrano 121, 28006 Madrid, Spain d Materials Physics Center CSIC-UPV/EHU and Donostia International Physics Center, 20018 San Sebastian, Spain b

art ic l e i nf o

a b s t r a c t

Article history: Received 2 June 2014 Received in revised form 12 September 2014 Accepted 26 September 2014 Available online 7 October 2014

In this work, we report the near infrared and upconversion emissions of Er3 þ –Yb3 þ codoped fluorotellurite TeO2–ZnO–ZnF2 glasses for different YbF3 concentrations ranging between 0.5 and 2 wt%. The study includes absorption and emission spectra and lifetime measurements for the infrared and visible fluorescence. The energy transfer between Yb3 þ and Er3 þ ions is confirmed by the temporal behavior of the near-infrared luminescence of Yb3 þ ions as well as by the enhancement of the 1532 nm emission of Er3 þ ions in the codoped samples. The Yb3 þ -Er3 þ energy transfer efficiency is calculated from the Yb3 þ lifetimes in single and codoped samples. Back transfer from Er3 þ to Yb3 þ ions is present under near infrared and visible excitation of Er3 þ ions at 798 and 488 nm respectively. An enhancement of the visible upconversion fluorescence is also observed in the codoped samples due to energy transfer from Yb3 þ to Er3 þ ions. The standardized value for the efficiency of the green upconversion emission is 1.06  10  4 for the codoped sample with 2 wt% of YbF3 which is comparable to that reported in lead–zinc–tellurite glasses. The possible upconversion processes and mechanisms leading to the population of several excited levels are discussed. & 2014 Elsevier B.V. All rights reserved.

Keywords: Laser spectroscopy Fluorotellurite glasses Er3 þ /Yb3 þ codoping Energy transfer Upconversion

1. Introduction The interest in Erbium-doped tellurite glasses has increased in the last years due to their important optical properties which make them suitable for applications in photonics such as optical amplifiers and frequency upconverters [1–9]. In particular, the broad bandwidth of the 1.5 mm emission in tellurite glasses, which is about twice the one of silica-based Er-doped fiber amplifiers (EDFA), makes these glasses attractive candidates for broadband amplifiers. In fact, excellent performance of a tellurite based EDFA, with a gain of 25.3 dB and a noise figure of less than 6 dB from 1561 to 1611 nm has been reported by Mori et al. [3]. Moreover, tellurite glasses combine good mechanical stability, chemical durability, high linear and nonlinear refractive indices, and low phonon energies (750 cm  1) with a wide transmission window (typically 0.4–6 μm), and high rare-earth solubility. The low phonon energy reduces the multiphonon relaxation rates and increases the quantum efficiency of excited states of rare-earth ions. Furthermore, the high values of the refractive index lead to large stimulated emission cross-sections at 1.5 μm as known to occur in rare-earth ions [1]. Mixed fluorotellurite glasses which combine the

n

Corresponding author. E-mail address: [email protected] (R. Balda).

http://dx.doi.org/10.1016/j.jlumin.2014.09.052 0022-2313/& 2014 Elsevier B.V. All rights reserved.

low phonon energies of fluorides with the high chemical durability and thermal stability of tellurites can reduce the OH content which has a great influence in the quenching processes of the radiative emission of excited levels of Er3þ ions [10,11]. The energy levels of the Er3þ ions for optical amplification at 1.5 μm act as a three level gain system which requires a high pump rate. A well known method for increasing the optical pumping efficiency of Er3þ -doped glasses is the sensitization with Yb3 þ ions since the spectral region of the 2F5/2-2F7/2 emission of the Yb3 þ ions overlaps that of 4I15/2-4I11/2 absorption of Er3þ ions, which favors an effective Yb3 þ to Er3þ energy transfer [12]. As a consequence, Yb3þ ions are often used as sensitizer to enhance the absorption crosssection and pumping efficiency of Er3þ ions at 980 nm in compact laser devices and optical amplifiers [13–15]. Moreover, Yb3þ codoping has demonstrated to be a successful way to enhance the upconversion efficiency in ytterbium sensitized Er3 þ systems [16–20]. In this work, we report the near infrared and upconversion emissions of Er3 þ –Yb3 þ codoped fluorotellurite TeO2–ZnO–ZnF2 glasses with a fixed ErF3 concentration (0.5 wt%) and different YbF3 concentrations (0.5, 1, and 2 wt%). The study includes absorption and emission spectra and lifetime measurements for the infrared fluorescence. The energy transfer between Yb3 þ and Er3 þ ions is confirmed by the temporal behavior of the nearinfrared luminescence of Yb3 þ ions as well as by the enhancement

A. Miguel et al. / Journal of Luminescence 158 (2015) 142–148

2. Experimental techniques Fluorotellurite host glasses, having a nominal composition: 74.6TeO2–8.8ZnO–16.6ZnF2 mol%, were prepared by meltquenching. Prior to the melting, a 10 g batch of high purity (99.99%) oxides and fluorides (Aldrich chemicals: TeO2, ZnO, ZnF2, ErF3, and/or YbF3) were mixed in an agate ball mill for 10 min at 320 rpm. Then, this powder mixture was introduced in a covered platinum crucible and melted in an electrical vertical furnace at temperatures in the 750–800 1C range during 30 min. Glass melts were homogenized by using an electrical platinum stirrer and finally they were poured onto a preheated brass mold. The obtained glass blocks (about 1 cm3 size) were immediately introduced into the annealing furnace, kept for 10 min at 300 1C and then cooled down to room temperature. The codoped samples were doped with 0.5 wt% of ErF3 (7.3  1019 Er3 þ ions/cm3) and 0.5, 1, and 2 wt% of YbF3 which correspond to 7.2  1019, 1.5  1020, and 2.9  1020 Yb3 þ ions/cm3 respectively. Single doped samples with 0.5 wt% of ErF3 and 0.5, 1, and 2 wt% of YbF3 were also prepared. The optical measurements were carried out on polished planoparallel glass slabs of about 1 mm thickness. Conventional absorption spectra were performed with a Cary 5 spectrophotometer. The steady-state emission measurements were made with a Ti–sapphire ring laser (0.4 cm  1 linewidth) and an Argon laser as exciting light. The fluorescence was analyzed with a 0.25 monochromator, and the signal was detected by an extended IR Hamamatsu H10330A-75 photomultiplier and finally amplified by a standard lock-in technique. Visible emission was detected by a Hamamatsu R636 photomultiplier. Lifetime measurements were obtained by exciting the samples with a dye laser pumped by a pulsed nitrogen laser and a Ti–sapphire laser, and detecting the emission with Hamamatsu R636 and H10330A-75 photomultipliers. Data were processed by a Tektronix oscilloscope. All measurements were performed at room temperature.

3. Results and discussion

0.5Er-2Yb

6 4 -1

Absorption coefficient (cm )

of the 1532 nm emission in the codoped samples. Back transfer from Er3 þ to Yb3 þ ions is present under near infrared (NIR) and visible (VIS) excitation of Er3 þ ions at 798 and 488 nm respectively. Bright upconverted green emission is observed at room temperature from all glasses together with a red 4F9/2-4I15/2 emission, which increases with Yb3 þ concentration. An enhancement of the visible upconversion fluorescence in Er3 þ /Yb3 þ codoped samples is confirmed due to efficient energy transfer from Yb3 þ to Er3 þ ions. The possible upconversion processes and mechanisms leading to the population of several excited levels are discussed.

143

2 0 6

2Yb

4 2 0 4

0.5Er

3 2 1 0 350

550

750

950 1150 1350 1550 1750

Wavelength (nm) Fig. 1. Room temperature absorption spectra of the single doped samples with 0.5 wt% of ErF3 and 2 wt% of YbF3 and the codoped sample with 0.5 wt% of ErF3 and 2 wt% of YbF3.

2.93  10  20 cm2 in the codoped samples. The spectra for the other co-doped samples are similar, except for the intensity of the 2 F7/2-2F5/2 absorption band, which increases as Yb3 þ concentration increases. The values of the integrated absorption coefficient for the 2 F7/2-2F5/2 transition increase linearly with the Yb3 þ content. In a previous work, the Judd–Ofelt theory was used to calculate the spontaneous emission probabilities and the radiative lifetimes for the excited levels of Er3þ ion in the same host [21]. The intensity values obtained for the Judd–Ofelt parameters were Ω2 ¼4.71  10  20 cm2, Ω4 ¼1.57  10  20 cm2, and Ω6 ¼1.13  10  20 cm2, with a root-meansquared deviation equal to 3.24  10  7. With these parameters a radiative lifetime of 3.23 ms was obtained for the 4I13/2 level. However, the Judd–Ofelt theory cannot be used to calculate the radiative lifetime of Yb3 þ ion due to its characteristic energy level structure. In this case, the radiative lifetime can be calculated from the expression [22], Z 1 g f 8πcn2 ¼ αðλÞdλ ð1Þ τR g i λ4p N were gi and gf are the degeneracies of the initial (2F5/2) and final (2F7/2) states, λp is the mean wavelength of the 2F5/2-2F7/2 electronic transition, n is the refractive index, N is the Yb3 þ concentration, c is the light velocity, and α is the absorption coefficient of the 2F7/2-2F5/2 transition. The calculated lifetime is 380 μs.

3.1. Absorption spectra

3.2. NIR down-conversion luminescence under NIR excitation

The room temperature absorption spectra were measured for all samples in the 350–1750 nm spectral range. As an example, the absorption spectra corresponding to the single-doped samples with 0.5 wt% of ErF3 and 2 wt% of YbF3, and the codoped sample with 0.5 wt% of ErF3 and 2 wt% of YbF3, are shown in Fig. 1. The observed bands in the absorption spectrum of the Er3 þ singledoped sample are assigned to the transitions from the 4I15/2 ground state to each one of the excited states of Er3 þ ions. In the case of the ytterbium doped sample, only one transition is observed from the ground state 2F7/2 to the unique excited state 2 F5/2. This transition overlaps the 4I15/2-4I11/2 absorption of Er3 þ ions. The absorption cross-section at 980 nm increases from 3.18  10  21 cm2 in the case of the Er3 þ single doped sample to

The NIR emission spectra of the Er3 þ /Yb3 þ codoped samples have been obtained in the 900–1700 nm spectral range by exciting at different excitation wavelengths in resonance with the 4I9/2 (Er3 þ ) and 4I11/2(Er3 þ )–2F5/2(Yb3 þ ) levels. As an example, Fig. 2 shows the fluorescence spectra for the Er3 þ single doped sample and the codoped samples measured under excitation at 798 nm in resonance with the 4I9/2 (Er3 þ ) level. After excitation of this level, the next lower levels are populated by multiphonon relaxation. The spectrum of the single doped sample shows the emissions corresponding to the 4I11/2-4I15/2 and 4I13/2-4I15/2 transitions whereas in the case of the codoped samples the spectra show, in addition to the Er3 þ emissions, the Yb3 þ emission corresponding to the 2F5/2-2F7/2 (Yb3 þ ) transition. The presence of the Yb3 þ

144

A. Miguel et al. / Journal of Luminescence 158 (2015) 142–148

Fig. 2. Room temperature emission spectra of Er3 þ and Yb3 þ in the codoped samples together with the emission spectrum of Er3 þ ions in the single doped glass obtained by exciting at 798 nm.

Fig. 3. Room temperature emission spectra of Er3 þ ions in the single doped and codoped samples obtained by exciting at 980 nm.

emission around 1000 nm indicates that an energy transfer from Er3 þ to Yb3 þ ions takes place. The ratio between the (4I11/24 I15/2 þ 2F5/2-2F7/2) and 4I13/2-4I15/2 emission bands increases from 0.1 for the single doped sample to 1.01 for the codoped sample with 2 wt% of YbF3 which indicates that the Er3 þ -Yb3 þ energy transfer increases with Yb3 þ concentration. The increase of Yb3 þ concentration decreases the Yb3 þ –Er3 þ distances which results in an increase of the energy transfer rate from Er3 þ to Yb3 þ favoring the Yb3 þ emission around 1000 nm. Fig. 3 shows the emission spectra obtained for all samples by exciting at 980 nm. Under 980 nm excitation, 4I11/2 (Er3 þ ) and 2F5/2 (Yb3 þ ) levels are populated. As it may be seen, the peak position ( 1532 nm) and linewidth ( 65 nm) of the 4I13/2-4I15/2 (Er3 þ ) emission are similar for all samples but the fluorescence intensity increases significantly as Yb3 þ concentration increases. As we have seen in the absorption spectra, for a fixed concentration of Er3 þ ions the introduction of Yb3 þ ions increases the optical absorption at 980 nm, which increases the population of level 4I11/2 due to the energy transfer from Yb3 þ to Er3 þ and consequently the emission from the 4I13/2 level. The energy transfer from Yb3 þ to Er3 þ is also confirmed by the emission spectra obtained under 890 nm excitation where only Yb3 þ ions absorb. The spectra show in addition to the Yb3 þ emission, the 4I13/2-4I15/2 (Er3 þ ) band which indicates that an Yb3 þ -Er3 þ energy transfer takes place (see Fig. 4). The experimental decays of the 2F5/2-2F7/2 (Yb3 þ ) emission have been measured in the Yb3 þ single-doped and in the Er3 þ /Yb3 þ codoped samples to have a further evidence of energy transfer from Yb3 þ to Er3 þ at room temperature. Lifetime measurements were performed under excitation at 980 and the fluorescence was monitored at 1070 nm to avoid the luminescence due to the 4I11/2-4I15/2 (Er3 þ ) emission. The curves show an exponential behavior even at

Fig. 4. Room temperature emission spectra of Er3 þ and Yb3 þ in the codoped samples obtained by exciting at 890 nm.

Fig. 5. Semilogarithmic plot of the fluorescence decays of the 2F5/2 level of Yb3 þ in the single doped sample with 2 wt%YbF3 and in the codoped sample with 0.5 wt% ErF3–2 wt%YbF3.

high concentration of Yb3 þ . The lifetime values for the samples with 0.5 and 1 wt% of YbF3 are 546 and 550 μs respectively, and increases up to 702 μs for the sample with 2 wt% of YbF3 due to radiation trapping effect [23]. In spite of following the same rising trend, the decay of the 2F5/2 (Yb3 þ ) level in the codoped samples is faster than in the single-doped samples, which confirms the non-radiative energy transfer from Yb3 þ to Er3 þ . As an example, the fluorescence decays for the single-doped sample with 2 wt% of YbF3 and the codoped sample with 0.5 wt% ErF3–2 wt%YbF3 are shown in Fig. 5. The energy transfer (ET) efficiency η from Yb3 þ to Er3 þ can be obtained by using the expression: ηYb–Er ¼ 1 

τYb–Er τYb

ð2Þ

where τYb–Er and τYb are the measured lifetimes of the 2F5/2 (Yb3 þ ) level, with and without Er3 þ ions, respectively. The energy transfer efficiency increases with the Yb3 þ concentration from 5% for the 0.5 wt%ErF3–0.5 wt%YbF3 sample to 22% for the codoped sample with the highest YbF3 concentration. Higher concentration of rare earth ions leads to smaller distances among donor and acceptor ions, which favors the energy transfer. However, the ET efficiency is lower than in tellurite glasses [24,25]. The low value of the energy transfer efficiency can be related with the presence of back transfer from Er3 þ to Yb3 þ ions and the lower multiphonon relaxation rate from 4I11/2 (Er3 þ ) level. It has been reported that the ET mainly depends on the ratio between the back-transfer rate and the multiphonon relaxation rate of the 4I11/2 (Er3 þ ) level [12]. The existence of back transfer has been demonstrated by the presence of the Yb3 þ luminescence under excitation in the Er3 þ absorption band at 798 nm. The absorption spectra show that the 2F7/2-2F5/2 (Yb3 þ ) and 4I15/2-4I11/2 (Er3 þ ) absorption bands overlap, being the

A. Miguel et al. / Journal of Luminescence 158 (2015) 142–148

respective transitions of both ions nearly resonant. As a consequence, the excitation can be transferred in both directions. On the other hand, in this flurotellurite glass the presence of ZnF2 weakens the electron–phonon coupling strength of the phonon mode locally coupled to Er3 þ ions, which reduces the multiphonon relaxation rate from 4I11/2(Er3 þ ) level and makes the level lifetime longer if compared to pure tellurite glasses [10,11]. The back transfer rate depends on the lifetime of 4I11/2(Er3 þ ) level, increasing as lifetime increases. Therefore, the Yb3 þ -Er3 þ energy transfer efficiency in this fluorotellurite glass is lower than in tellurite glasses. The temporal evolution of the luminescence from 4I13/2 level has been obtained for all samples under excitation at 980 nm. The experimental decays can be fitted to a single exponential function and lifetimes present similar values ranging from 4.7 ms for the Er3 þ single-doped sample to 4.8, 4.5, and 4.4 ms for the codoped samples with 0.5, 1, and 2 wt% YbF3 respectively. As can be seen, in this glass the lifetime of the 4I13/2 level is sligthly reduced by the presence of Yb3 þ ions.

3.3. NIR down-conversion luminescence under VIS excitation The NIR emission spectra were also measured for all samples under visible excitation in the 4F7/2 (Er3 þ ) level at 488 nm. In this case, one visible photon can produce two emitted photons from lower lying levels of Yb3 þ and Er3 þ . As can be seen in Fig. 6, in the codoped samples, the spectra show the emission around 1000 nm corresponding to the overlapping between the 4I11/2-4I15/2 (Er3 þ ) and 2F5/2-2F7/2 (Yb3 þ ) transitions together with the 4I13/2-4I15/2 (Er3 þ ) emission of Er3 þ ions centered at around 1532 nm. The 2 F5/2-2F7/2 (Yb3 þ ) emission increases with Yb3 þ concentration whereas the 4I13/2-4I15/2 (Er3 þ ) emission decreases in the codoped samples if compared with the Er3 þ single doped sample which indicates that an Er3 þ -Yb3 þ energy transfer process takes place. Once the 4S3/2 (Er3 þ ) is populated nonradiatively from the excited 2 H11/2 (Er3 þ ) level, there are two possible downconversion processes [26]. The first one is a cross-relaxation from Er3þ (4S3/2-4I11/2) to Yb3þ (2F7/2-2F5/2) with an energy mismatch of  1900 cm  1. In this process the absorption of around 3 phonons is needed to reach the 2F5/2 (Yb3þ ) level, since the maximum phonon energy in these glasses is around 750 cm  1 [10]. The second one is a phonon-assisted cross-relaxation (4S3/2-4I13/2) (Er3 þ ); (2F7/2-2F5/2) (Yb3þ ) with an energy mismatch of þ 1000 cm  1. This process involves the emission of around 2 phonons and is expected to be more probable. After these cross-relaxation processes the populated 4I11/2 and 4I13/2 levels relax radiatively to the ground state. Part of the energy in 4I11/2 level is transferred to the 2F5/2 level of Yb3 þ by the resonant energy transfer process: Er3þ (4I11/2-4I15/2); Yb3þ (2F7/2-2F5/2) which generates the emission around 1000 nm.

Fig. 6. Room temperature emission spectra of Er3 þ and Yb3 þ in the codoped samples together with the emission spectrum of Er3 þ ions in the single doped glass obtained by exciting at 488 nm.

145

Fig. 7. Semilogarithmic plot of the fluorescence decays of the 4S3/2 level of Er3 þ in the single doped sample with 0.5 wt% ErF3 and in the codoped samples with 0.5 and 2 wt% YbF3.

To further investigate the Er3þ -Yb3þ energy transfer under VIS excitation, the temporal evolution of the 4S3/2-4I15/2 (Er3þ ) emission has been obtained for all samples as a function of Yb3 þ concentration under excitation at 488 nm. As an example, Fig. 7 shows the decays of the 4S3/2 level of Er3þ in the single doped sample with 0.5 wt% ErF3 and in the codoped samples with 0.5 and 2 wt% YbF3. The fluorescence decay curves corresponding to the 4S3/2-4I15/2 (Er3þ ) emission show a non-exponential behavior for the co-doped samples. The nonexponential character of the decays together with the reduction of the lifetimes can be explained by the additional probability for energy transfer to Yb3þ ions. The average lifetime of the 4S3/2 level, calculated R from hτi ¼ ðI ðt Þ U dt=I max Þ decreases from 71 μs to 44.6 μs for the codoped sample with the highest Yb3 þ concentration, which confirms an energy transfer mechanism involving the 4S3/2 (Er3 þ ) level. The Er3 þ -Yb3 þ transfer efficiency has been evaluated from the lifetimes of 4S3/2 (Er3 þ ) level in single-doped and co-doped samples by using expression (2). The ET efficiency increases from 15% for the codoped sample with 0.5 wt% YbF3 up to 37% for the codoped sample with the highest YbF3 concentration (2 wt%). 3.4. Direct and upconverted visible emissions The visible emission spectra in the 500–700 nm range were obtained under 488 nm excitation in the 4F7/2 level. After excitation of this level, the next lower levels are populated by multiphonon relaxation. Fig. 8(a) shows the emission spectra for all samples. As can be seen, the main emission corresponds to the green one associated to the (2H11/2,4S3/2)-4I15/2 thermalized transitions. The weak red emission from level 4F9/2 is due to its population from level 4S3/2 through multiphonon relaxation. The energy gap between both levels is 3100 cm  1 and the maximum phonon energy is 750 cm  1. The phonon order involved in this process is about 4, which indicates that multiphonon relaxation occurs with a moderate rate. It can be also observed that, the green emission from (2H11/2, 4S3/2) levels is reduced for a fixed Er3þ concentration as the Yb3þ concentration is increased, whereas the intensity of the weak red emission from 4F9/2 level remains unchanged. Although a reduction of the red emission with Yb3þ concentration should be expected, the observed behavior has been previously reported and attributed to the cross relaxation mechanism 4F7/2-4I11/2 (Er3þ ); 2F7/2-2F5/2 (Yb3þ ) which populates the 4I11/2 (Er3þ ) level and favors the energy transfer mechanism between Er3þ ions (4F7/2-4I11/2); (4I11/2-4F9/2) populating the 4F9/2 level [27]. The room temperature upconverted emission spectra obtained under 980 excitation, in resonance with the 4I11/2 (Er3 þ ) and 4F5/2 (Yb3 þ ) levels, are shown in Fig. 8(b). The shape of the emission spectra is similar to the one obtained under direct excitation (488 nm). However, the green emission intensity is greatly

146

A. Miguel et al. / Journal of Luminescence 158 (2015) 142–148

enhanced when Yb3 þ concentration increases, becoming E40 times more intense than in the single-doped sample. This enhancement, which is higher than that found recently in zinc–tellurite glasses [28], confirms the existence of an effective Yb3 þ –Er3 þ energy transfer process. Moreover, as can be observed in this figure, there is an increase in the red emission intensity relative to the green one with increased Yb3 þ concentration. If we compare the upconversion emission spectrum obtained in the single doped sample and those obtained for the codoped samples, the ratio between the intensities of red and green emissions changes from 0.038 in the single doped sample to 0.68 for the codoped

Fig. 8. Room temperature emission spectra obtained (a) under excitation at 488 nm in the single doped and codoped samples and (b) under 980 nm excitation.

sample with 2 wt% YbF3 which suggests the existence of additional mechanisms that populate level 4F9/2 (Er3 þ ). To investigate the excitation mechanisms for populating the 4 S3/2 and 4F9/2 levels after 980 nm excitation, we have obtained the evolution of the upconverted emission intensities at 544 and 652 nm for different pumping powers. The upconversion emission intensity (Iem) depends on the incident pump power (Ppump) according to the relation Iem p (Ppump)n, where n is the number of photons involved in the pumping mechanism. Fig. 9 shows a logarithmic plot of the integrated emission intensity of the upconverted green and red emissions as a function of the pump laser intensity. The dependence of the intensity on the pump power is quadratic which indicates that a two photon (TP) upconversion process populates levels 4S3/2 and 4F9/2. This in turn may be associated to excited state absorption (ESA) and/or to energy transfer upconversion (ETU) [29]. Fig. 10 shows the possible upconversion mechanism for the red and green emissions under 980 nm excitation. The Er3 þ ions can be excited to the 4I11/2 state by ground state absorption and by resonant energy transfer from the Yb3 þ ions through the

Fig. 9. Logarithmic plot of the integrated emission intensities of the upconversion from 4S3/2, and 4F9/2 levels as a function of the pump laser intensity obtained under excitation at 980 nm. Symbols correspond to the experimental data and solid lines are linear fits. The slope values of the linear fits are also included.

Fig. 10. Energy level diagram of Er3 þ and Yb3 þ ions and possible upconversion mechanisms.

A. Miguel et al. / Journal of Luminescence 158 (2015) 142–148

(2F5/2-2F7/2) (Yb3 þ ); (4I15/2-4I11/2) (Er3 þ ) process (ET1). This process seems to be important, since the Yb3 þ ions have a larger absorption cross-section than the Er3 þ ions at 980 nm and the upconversion fluorescence is more efficient in the codoped samples. The Er3 þ ions in the 4I11/2 level can be excited to the 4F7/2 state by excited state absorption of a second infrared photon 4I11/2 (Er3 þ )þphoton-4F7/2 (Er3 þ ), by energy transfer from an Yb3 þ ion (2F5/2-2F7/2) (Yb3 þ );(4I11/2-4F7/2) (Er3 þ ) (ET2), and/or by energy transfer from an adjacent Er3 þ ion (4I11/2 -4I15/2) (Er3 þ ); (4I11/2-4F7/2) (Er3 þ ). This last process depends on the Er3 þ concentration. The populated 4F7/2 level may relax nonradiatively to the lower levels 2H11/2 and 4S3/2, which produces the intense green emission. The population of the 4I11/2 level increases with Yb3 þ doping as a result of the energy transfer from Yb3 þ (2F5/2) to Er3 þ (4I11/2). As a consequence, the probabilities of an ESA and/or ETU processes from the populated 4I11/2 level are higher in the codoped glasses than those in the Er3 þ single doped system. Concerning the red emission from level 4F9/2, as we mentioned before, the population of this level increases as the Yb3 þ concentration increases. Multiphonon relaxation from the 4S3/2 level together with ESA and/or ETU from the 4I13/2 level can populate the 4F9/2 level responsible for the red emission. A possible explanation to this behavior is an excited state absorption from the 4I13/2 level populated nonradiatively from the 4I11/2 and/or an energy transfer process described by (4F5/2-4F7/2) (Yb3 þ );(4I13/2-4F9/2) (Er3 þ ) (ET3). This energy transfer process can explain the enhancement of the ratio between the red and green upconverted emissions as Yb3 þ concentration increases. The upconversion efficiency for the green emission obtained under 980 nm excitation has been estimated by comparing the upconversion luminescence intensity with the one obtained under one photon excitation by using the expression [30,31]:   P ðVISÞ I em ðupconversionÞ η ¼ ηq abs ð3Þ P abs ðIRÞ I em ðdirect Þ where ηq is the luminescence quantum yield of the emitting level after direct excitation, Pabs(VIS) and Pabs(IR) are the absorbed light power for VIS and IR excitation respectively, and Iem(upconversion) and Iem(direct) are the luminescence intensities of the green emission collected under the same experimental conditions. The values of the absorbed light power were calculated from the measured laser power incident on the sample, the absorption coefficient at the excitation wavelength, and the absorption path length on the sample. The luminescence quantum yield for direct excitation has been calculated from the expression ηq ¼ ðτexp =τR Þ, where τexp is the measured lifetime (44.6 μs) and τR is the radiative lifetime obtained from Judd–Ofelt calculation (290 μs). To compare the upconversion efficiency of this glass with other Er–Yb codoped glasses, we have calculated the standardized efficiency by using an incident light power density of 1 W/cm2. The obtained value for the green emission in the codoped sample with 2 wt% YbF3 concentration is 1.06  10  4. This value is much higher than those reported for lead-germanate (3.1  10  6), silicate (2  10  7), phosphate (6  10  8), and lead–tellurite–germanate (0.79  10  6) glasses, lower than those reported for fluoride (1  10  3) and alkali bismuth gallate glasses (1.3  10  3), and comparable to that reported in lead–zinc–tellurite glasses (8.8  0  4) [18,19,30,31].

4. Conclusions In this work we demonstrate the existence of Yb3 þ 2Er3 þ energy transfer in 74.6TeO2–8.8ZnO–16.6ZnF2 glasses doped with 0.5 wt% of ErF3 and codoped with 0.5, 1, and 2 wt% YbF3. The energy transfer between Yb3 þ and Er3 þ ions is evidenced both by the temporal behavior of the near-infrared luminescence of Yb3 þ

147

ions and by the enhancement of the 1532 nm emission in the codoped samples. The Yb3 þ -Er3 þ energy transfer efficiency obtained from the Yb3 þ (2F5/2) lifetime with and without Er3 þ ions, reaches 22% for the codoped sample with 0.5% ErF3 and 2% YbF3. The relatively low value for the energy transfer efficiency is attributed to the lower multiphonon relaxation rates from the 4 I11/2 level in these flurotellurite glasses if compared with other glasses which favor an efficient back transfer from Er3 þ to Yb3 þ ions. Back transfer from Er3 þ to Yb3 þ ions is also observed under near infrared excitation and visible excitation of Er3 þ ions at 798 and 488 nm respectively. In the last case, the energy transfer efficiency from Er3 þ to Yb3 þ ions obtained from the 4S3/2 lifetime increases up to 37% for the codoped sample with the highest YbF3 concentration. Intense green emission due to the (2H11/2,4S3/2)4 I15/2 transitions together with a red emission corresponding to the 4 F9/2-4I15/2 transition is observed in all samples and attributed to a two photon process. The increase in the population of the 4I11/2 level of Er3 þ ions as a consequence of the Yb3 þ codoping, increases the upconversion emission intensity in the codoped samples becoming E40 times more intense than in the single-doped sample. The standardized upconversion efficiency of the green emission for the codoped sample with the highest Yb3 þ concentration (2 wt% YbF3) is 1.06  10  4. This value is comparable to that reported in lead–zinc– tellurite glasses. Finally, the increase of the red emission as YbF3 concentration increases could be attributed to an excited state absorption from the 4I13/2 level populated nonradiatively to the 4 I11/2 and/or an energy transfer process described by (4F5/2-4F7/2) (Yb3 þ );(4I13/2 -4F9/2) (Er3 þ ). The results presented in this work might provide useful information for further development of photonic devices based on Er3 þ –Yb3 þ codoped flurotellurite glasses.

Acknowledgments This work has been supported by the Spanish Government MINECO (Projects MAT2009-14282-C02-02, MAT2009-14282-C02-01, FIS2011-27968, and MAT2013-48246-C2-2-P) and the Basque Country Government (Project IT-659-13). A. Miguel and R. Morea acknowledge FPI grants from the Spanish Government. References [1] J.S. Wang, E.M. Vogel, E. Snitzer, Opt. Mater. 3 (1994) 187. [2] A. Mori, Y. Ohishi, S. Sudo, Electron. Lett. 33 (1997) 863. [3] A. Mori, T. Sakamoto, K. Kobayashi, K. Shikano, K. Oikawa, K. Hoshino, T. Kanamori, Y. Ohishi, M. Shimizu, IEEE J. Lightwave Technol. LT-20 (2002) 822. [4] X. Feng, S. Tanabe, T. Hanada, J. Am. Ceram. Soc. 84 (2001) 165. [5] M. Yamada, A. Mori, K. Kobayashi, H. Ono, T. Kanamori, K. Oikawa, Y. Nishida, Y. Ohishi, IEEE Photon. Technol. Lett. 10 (1998) 1244. [6] G.C. Righini, M. Brenci, G.N. Conti, S. Pelli, M. Ferrari, M. Bettinelli, A. Speghini, B. Chen, Proc. SPIE 5061 (2002) 34. [7] N. Jaba, A. Kanoun, H. Mejri, A. Selmi, S. Alaya, H. Maaref, J. Phys: Condens. Matter 12 (2002) 4523. [8] F. Vetrone, J.C. Boyer, J.A. Capobianco, A. Speghini, M. Bettinelli, Appl. Phys. Lett. 80 (2002) 1752. [9] K. Kumar, S.B. Rai, D.K. Rai, J. Non-Cryst. Solids 353 (2007) 1383. [10] V. Nazabal, S. Todoroki, A. Nukui, T. Matsumoto, S. Suehara, T. Hondo, T. Araki, S. Inoue, C. Rivero, T. Cardinal, J. Non-Cryst. Solids 325 (2003) 85. [11] A. Miguel, M. Al-Saleh, J. Azkargorta, R. Morea, J. Gonzalo, M.A. Arriandiaga, J. Fernández, R. Balda, Opt. Mater. 35 (2013) 2039. [12] V.P. Gapontsev, S.M. Matitsin, A.A. Isineev, V.B. Kravchencko, Opt. Laser Technol. 14 (1982) 189. [13] P. Laporta, S. de Silvestri, V. Magni, O. Svelto, Opt. Lett. 16 (1991) 1952. [14] J.A. Hutchinson, T.H. Allik, Appl. Phys. Lett. 60 (1992) 1424. [15] K. Hsu, C.H. Miller, J.T. Kringlebotn, D.N. Payne, Opt. Lett. 20 (1995) 377. [16] V.K. Tikhomirov, V.D. Rodríguez, J. Méndez-Ramos, J. del-Castillo, D. Kirilenko, G. Van Tendeloo, V.V. Moshchalkov, Sol. Energy Mater. Sol. Cells 100 (2012) 209. [17] J. Yang, L. Zhang, L. Wen, S. Dai, L. Hu, J. Zhonghong, J. App. Phys. 95 (2004) 3020. [18] H. Lin, E.Y.B. Pun, S.Q. Man, X.R. Liu, J. Opt. Soc. Am. B 18 (2001) 602. [19] Z. Jin, Q. Nie, T. Xu, S. Dai, X. Shen, X. Zhang, Mater. Chem. Phys. 104 (2007) 62.

148

A. Miguel et al. / Journal of Luminescence 158 (2015) 142–148

[20] M. Pokhrel, G.A. Kumar, S. Balaji, R. Debnath, D.K. Sardar, J. Lumin. 132 (2012) 1910. [21] A. Miguel, R. Morea, J. Gonzalo, M.A. Arriandiaga, J. Fernandez, R. Balda, J. Lumin. 140 (2013) 38. [22] M.J. Weber, Phys. Rev. B 4 (1971) 3153. [23] F. Auzel, G. Baldacchini, L. Laversenne, G. Boulon, Opt. Mater. 24 (2003) 103. [24] H. Desirena, E. De la Rosa, A. Shulzgen, S. Shabet, N. Peyghambarian, J. Phys. D: Appl. Phys. 41 (2008) 095102 (7pp). [25] H. Desirena, E. De la Rosa, V.H. Romero, J.F. Castillo, L.A. Díaz-Torres, J.R. Oliva, J. Lumin. 132 (2012) 391.

[26] B. Fan, C. Point, J.L. Adam, X. Zhang, X. Fan, H. Ma, J. Appl. Phys. 110 (2011) 113107 (8pp). [27] F. Vetrone, J.C. Boyer, J.A. Capobianco, A. Speghini, M. Bettinelli, J. Appl. Phys. 96 (1) (2004) 661. [28] D.K. Mohanty, V.K. Rai, Spectrochim. Acta Part A: Mol. Biomol. Spectrosc. 121 (2014) 9. [29] R. Balda, A.J. Garcia-Adeva, J. Fernandez, J.M. Fdez-Navarro, J. Opt. Soc. Am. B 21 (2004) 744. [30] Z. Pun, S.H. Morgan, K. Dyer, A. Ueda, H. Liu, J. Appl. Phys. 79 (1996) 8906. [31] R.S. Quimby, M.G. Drexhage, M.J. Suscavage, Electron. Lett. 23 (1987) 32.