Upconversion luminescence in Er3+ doped Ga10Ge25S65 glass and glass-ceramic excited in the near-infrared Whualkuer Lozano B., Cid B. de Araújo, Yannick Ledemi, and Younes Messaddeq Citation: Journal of Applied Physics 113, 083520 (2013); doi: 10.1063/1.4793638 View online: http://dx.doi.org/10.1063/1.4793638 View Table of Contents: http://scitation.aip.org/content/aip/journal/jap/113/8?ver=pdfcov Published by the AIP Publishing [This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP: 200.145.3.45 On: Wed, 05 Feb 2014 15:49:01 http://scitation.aip.org/content/aip/journal/jap?ver=pdfcov http://oasc12039.247realmedia.com/RealMedia/ads/click_lx.ads/www.aip.org/pt/adcenter/pdfcover_test/L-37/1479233748/x01/AIP-PT/JAP_Article_DL_0214/aipToCAlerts_Large.png/5532386d4f314a53757a6b4144615953?x http://scitation.aip.org/search?value1=Whualkuer+Lozano+B.&option1=author http://scitation.aip.org/search?value1=Cid+B.+de+Ara�jo&option1=author http://scitation.aip.org/search?value1=Yannick+Ledemi&option1=author http://scitation.aip.org/search?value1=Younes+Messaddeq&option1=author http://scitation.aip.org/content/aip/journal/jap?ver=pdfcov http://dx.doi.org/10.1063/1.4793638 http://scitation.aip.org/content/aip/journal/jap/113/8?ver=pdfcov http://scitation.aip.org/content/aip?ver=pdfcov Upconversion luminescence in Er31 doped Ga10Ge25S65 glass and glass-ceramic excited in the near-infrared Whualkuer Lozano B.,1,a) Cid B. de Ara�ujo,1,b) Yannick Ledemi,2,3 and Younes Messaddeq2,3 1Departamento de F�ısica, Universidade Federal de Pernambuco, Recife-PE 50670-901, Brazil 2Centre d’Optique, Photonique et Laser, Universit�e Laval, Qu�ebec (Qc) G1V 0A6, Canada 3Instituto de Qu�ımica, Universidade Estadual Paulista-UNESP, 14801-970 Araraquara, S~ao Paulo, Brazil (Received 21 November 2012; accepted 13 February 2013; published online 27 February 2013) The infrared-to-visible frequency upconversion was investigated in Er3þ-doped Ga10Ge25S65 glass and in the transparent glass-ceramic obtained by heat-treatment of the glass above its glass-transition temperature. Continuous-wave and pulsed lasers operating at 980 nm and 1480 nm were used as excitation sources. The green (2H11/2! 4I15/2; 4S3/2! 4I15/2) and red (4F9/2! 4I15/2) photoluminescence (PL) signals due to the Er3þ ions were characterized. The PL decay times were influenced by energy transfer among Er3þ ions, by cross-relaxation processes and by energy transfer from the Er3þ ions to the host material. The PL from the Er3þ ions hosted in the crystalline phase was distinguished only when the glass-ceramic was excited by the 1480 nm pulsed laser. The excitation pathways responsible for the green and red PL bands are discussed to explain the differences between the spectra observed under continuous-wave and pulsed excitation. VC 2013 American Institute of Physics. [http://dx.doi.org/10.1063/1.4793638] I. INTRODUCTION The search of transparent materials that can be doped with rare earth (RE) ions has been very intense in the past years motivated by applications in displays, biomedical lasers, optical sensors, optical amplifiers, among other uses.1–6 In particular, special glasses and glass-ceramics are being increasingly studied because they have unique features such as large acceptance for high RE ions doping concentra- tion, large optical homogeneity, wide transparency from the visible to the infrared region, high mechanical strengths and simple manufacture procedures for obtaining good optical quality samples. Among the materials that exhibit such appropriate characteristics for photonics are the chalcoge- nide glasses and chalcogenide glass-ceramics that exhibit high refractive index (�2.2), low cutoff phonon energy (�400 cm�1), and high stability against moisture and devitri- fication.2–6 In previous papers, we reported on the photoluminescence (PL) properties of the chalcogenide glass Ga10Ge25S65 (GGS) doped with Nd3þ, Pr3þ, and Er3þ ions. For the Nd3þ-doped GGS glass (GGS-GLASS), we determined transition probabil- ities, radiative lifetimes, and branching ratios related to the Nd3þ ions. Green and red emissions were observed for excita- tion at 1064 nm due to two-photon absorption by isolated ions and energy transfer among Nd3þ pairs.7 In Pr3þ-doped GGS glasses, we observed orange-to-blue frequency upconversion (UC) and investigated the influence of silver nanoparticles on the UC efficiency.8 The UC enhancement observed was attrib- uted to the large local field acting on the Pr3þ ions due to their proximity with silver nanoparticles. Er3þ-doped GGS glass was studied under laser excitation at 532 nm, 800 nm, and 980 nm.9 Er3þ parameters such as transition probabilities, radi- ative lifetimes, and branching ratios were determined. Also, the mechanisms leading to Stokes and anti-Stokes PL were dis- cussed. Other authors investigated the UC luminescence in Er3þ-doped GGS-CsCl glass-ceramic excited at 800 nm.10 Enhancements of the green and red PL were observed in com- parison with the base-glass. In the present article, we report on the infrared-to-visible UC luminescence of Er31-doped GGS glass and Er31-doped GGS glass-ceramics (GGS-GC) excited at 980 nm and 1480 nm. The experiments were performed with continuous- wave (CW) and pulsed lasers. The intensity and the time behavior of the visible PL due to the Er3þ ions were studied. The excitation mechanisms contributing for the UC process and for the relaxation rates of the electronic states involved in the PL process are discussed. II. EXPERIMENTAL DETAILS The base-glass (GGS-GLASS) with composition Ga10Ge25S65:(Er2S3)0.25 (mol. %) was prepared by the melt- quenching method using a mixture of highly pure raw mate- rials (Ga2S3, Ge, S: 99.99% and Er2S3: 99%) in a silica ampoule sealed under vacuum (10�4 mbar). The ampoule of 9 mm inner diameter was placed in a rocking furnace, slowly heated up to 900 �C and maintained at this temperature for 12 h. The silica tube was quenched in water at room tempera- ture, annealed near the glass transition temperature, Tg, for 3 h to minimize inner constraints, and finally slowly cooled down to room temperature. The glass rods of 8 g weight were cut into slices of 2 mm thickness and polished to obtain perpendicular polished faces for the optical measurements. The value of Tg¼ 430 �C, and the onset crystallization a)Permanent address: Facultad de Ciencias F�ısicas, Universidad Nacional Mayor de San Marcos-UNMSM, Lima, Peru. b)Author to whom correspondence should be addressed. Electronic mail: cid@df.ufpe.br. 0021-8979/2013/113(8)/083520/6/$30.00 VC 2013 American Institute of Physics113, 083520-1 JOURNAL OF APPLIED PHYSICS 113, 083520 (2013) [This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP: 200.145.3.45 On: Wed, 05 Feb 2014 15:49:01 http://dx.doi.org/10.1063/1.4793638 http://dx.doi.org/10.1063/1.4793638 http://dx.doi.org/10.1063/1.4793638 http://dx.doi.org/10.1063/1.4793638 http://dx.doi.org/10.1063/1.4793638 http://dx.doi.org/10.1063/1.4793638 http://dx.doi.org/10.1063/1.4793638 mailto:cid@df.ufpe.br http://crossmark.crossref.org/dialog/?doi=10.1063/1.4793638&domain=pdf&date_stamp=2013-02-27 temperature, Tx¼ 540 �C were determined by differential scanning calorimetry, with a heating rate of 10 �C/min. A description of the fabrication procedure is presented in Refs. 11 and 12. The glass-ceramics (GGS-GC) were prepared by heat- treatment of polished slices of the base-glass at 450 �C for 15 h. The volume crystallization was checked by X ray dif- fraction and transmission electron microscopy that demon- strate the presence of Ga2S3 nanocrystals as reported in Refs. 10–12. The samples were optically homogeneous to the naked eye. The linear optical absorption spectra were recorded using a commercial spectrophotometer. For the PL experiments with excitation at 980 nm (10 204 cm�1) and 1480 nm (6757 cm�1), CW diode lasers (chopped at 8 Hz) with maximum output power of 300 mW and one optical parametric oscillator (5 ns, 5 Hz) pumped by a Nd: YAG laser were used. In all cases, the beams were focused onto the samples using a lens with focal length of 10 cm. The PL signals collected along a direction perpendic- ular to the laser beam were dispersed by a 0.25 m grating spectrometer attached to a photomultiplier tube. The signals were recorded using a digital oscilloscope connected to a personal computer. All measurements were made at room temperature. III. RESULTS AND DISCUSSION A. Linear absorption spectra and CW laser excitation Figure 1 shows the absorbance spectra of the GGS- GLASS and the GGS-GC samples. The bands centered at � 492 nm, � 526 nm, � 547 nm, and � 661 nm are due to transitions from the ground state (4I15/2) to the excited states 4F7/2, 2H11/2, 4S3/2, and 4F9/2, respectively. The bands related to the Er3þ ion states with higher energies than the 4F7/2 state cannot be seen because they lie above the absorption edge of the samples. The bands’ positions in Fig. 1 are in agreement with the ones observed in other Er3þ doped glasses1–3,10,13,14 and the spectrum in the infrared region was already pre- sented in Ref. 9. All Er3þ transitions are inhomogeneously broadened due to site-to-site variations of the crystalline field. Notice that the GGS-GC absorption edge is red-shifted with respect to the GGS-GLASS; it is an indication that the crystallization of the inner crystalline phase has been com- pleted after 15 h of heat-treatment.11 The optical gap energy, Eg, for both samples was determined from Fig. 1 by plotting the square of the optical absorption coefficient, a2 0, versus the photon energy and determining the crossing point between a straight line fitted to the a2 0 curve and the horizontal axis. The results obtained were Eg¼ 2.42 eV (19 526 cm�1) for the GGS-GLASS and Eg¼ 2.40 eV (19 353 cm�1) for the GGS-GC. Typical UC spectra under CW laser excitation at 980 nm and at 1480 nm emitted by the GGS-GLASS and GGS-GC are shown in Figs. 2(a) and 2(b), respectively. The PL bands centered at �660 nm, �549 nm, and �530 nm are due to the Er3þ transitions 4F9/2 ! 4I15/2, 4S3/2 ! 4I15/2, and 2H11/2 ! 4I15/2, respectively. The amplitude of the green (red) signal is �1000% (�100%) times larger for excitation at 980 nm than at 1480 nm. A very weak emission centered at �493 nm [4F7/2 ! 4I15/2] was observed only for excitation at 980 nm which is resonant with transitions 4I15/2! 4I11/2 and 4I11/2! 4F7/2. However, the signal-to-noise ratio was poor and cannot be clearly observed in Fig. 2. The intensity ratio (R¼ I549 nm/I530 nm) between the two green PL bands excited at 980 nm was twice higher than the FIG. 1. Absorbance spectra of the GGS-GLASS (green solid line) and GGS- GC (red dashed line). Sample thickness: 2 mm. FIG. 2. Room temperature UC emission spectra obtained for excitation at 980 nm and 1480 nm with continuous-wave lasers: (a) GGS-GLASS; (b) GGS-GC. 083520-2 Lozano et al. J. Appl. Phys. 113, 083520 (2013) [This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP: 200.145.3.45 On: Wed, 05 Feb 2014 15:49:01 one obtained with excitation at 1480 nm. The reason for this discrepancy is attributed to the dominant UC mechanism in each case, as discussed below. The green PL is visible to the naked eye for 1480 nm CW laser powers higher than 10 mW; however, it is weaker when the pulsed laser at 1480 nm was used. The dependence of the UC intensity, IUC, versus the laser intensity, I, in the absence of saturation, is described by IUC / IN , where N is the number of photons that participate in the UC process. Hence, the value of N corresponding to each UC transition was obtained from the slope of the straight line representing IUC versus I in the double- logarithmic plot of Fig. 3. It is shown in Figs. 3(a) and 3(b) that the laser intensity dependence of the green and the red PL bands when excited at 1480 nm presents cubic power law corresponding to the absorption of three laser photons. Fig. 3(c), for excitation at 980 nm, shows quadratic depend- ence of IUC versus I indicating that two laser photons contrib- ute for the UC process in the GGS-GC. The results for the GGS-GLASS sample were reported in Ref. 9 and also pres- ent quadratic dependence. Figure 4 shows the relevant Er3þ energy levels together with indication of possible UC pathways for both excitation wavelengths and the observed PL lines. A simplified scheme of the conduction and valence bands of the host matrix is also shown in Fig. 4. The main processes that may lead to transitions involving the excited Er3þ ion states in experiments with low-power infrared lasers are excited state absorption, cross-relaxation (CR), and energy transfer (ET) among the ions. Another pos- sibility would be the excitation of the host matrix by multi- photon absorption followed by ET from the matrix to the Er3þ ions; however, this process is less probable than the one mentioned above. Nevertheless it is important when high-power lasers are used as shown in Sec. III B. For excitation at 980 nm, the laser wavelength is in reso- nance with the transitions 4I15/2 ! 4I11/2 and 4I11/2 ! 4F7/2. Then, after promotion of Er3þ ions to level 4F7/2, nonradiative (NR) decay to levels 2H11/2 and 4S3/2 takes place followed by the green emissions due to transitions 4S3/2! 4I15/2 and 2H11/2 ! 4I15/2. The red emission (transition 4F9/2 ! 4I15/2) is due to the population that reached level 4F9/2 by phonon relaxation from the 4S3/2 level and CR involving ions in the 4I9/2 and 4S3/2 states.15,16 Therefore, in the experiments at 980 nm, the contri- bution for the PL signals is mainly due to isolated ions as well FIG. 3. Photoluminescence intensity versus laser intensity. Excitation wave- length: 1480 nm (a) and (b); 980 nm (c). FIG. 4. Simplified energy levels of Er3þ ions. The bold upward arrows rep- resent the laser induced transitions. The downward arrows represent lumi- nescence and the dotted arrows represent energy transfer between ions. Also presented is a scheme of conduction and valence bands of the host matrix. 083520-3 Lozano et al. J. Appl. Phys. 113, 083520 (2013) [This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP: 200.145.3.45 On: Wed, 05 Feb 2014 15:49:01 as a fraction of the total number of ions that interact weekly with their neighbors. For excitation at 1480 nm, the absorption transition 4I15/2 ! 4I13/2 takes place but the promotion from the 4I13/2 level to the high energy states is negligible because the laser wavelength is not resonant with transition 4I13/2! 4I9/2. However, it is prob- able that two Er3þ ions in the state 4I13/2 transfer their energy to a third ion, already in the state 4I13/2, that is promoted to the state 2H11/2; the corresponding excitation pathway is indi- cated in Fig. 4 by the dashed-dotted lines. This is the most prob- able pathway for samples having Er3þ concentration of 0.1 mol. % and larger, as demonstrated for other glasses.13,14,17,18 PL at 530 nm (2H11/2 ! 4I15/2) is observed as well as at 549 nm (4S3/2 ! 4I15/2) that occurs after NR decay from 2H11/2 to the 4S3/2 state. Since the contribution of isolated ions is negligible the whole PL spectrum is weaker than the spectrum observed for excitation at 980 nm. In order to understand the discrepancy between the in- tensity ratio R¼ I549 nm/I530 nm for excitation at 980 nm and 1480 nm, we recall that R depends on the ratio between the amplitudes of transitions 2H11/2 ! 4I15/2 and 4S3/2 ! 4I15/2 as well as on the number of Er3þ ions in the states 2H11/2 and 4S3/2. Since only the interacting ions contribute for the UC signal when the sample is excited at 1480 nm, the oscillator strengths of the transitions should be affected in different ways with respect to the case of the isolated ions that partici- pates in the UC for excitation at 980 nm. B. Pulsed laser excitation at 980 nm and 1480 nm The UC process excited with a pulsed laser at 980 nm in the GGS-GLASS was previously reported in Ref. 9 and the results obtained in the present experiment were similar. The signal corresponding to transition 2H11/2 ! 4I15/2 presents a rise time of �188 ns attributed to the NR decay from the level 4F7/2; the decay time of the signal from states 2H11/2 and 4S3/2 was �24 ls and �31 ls, respectively. These values are smaller than the radiative lifetime, calculated using the Judd-Ofelt theory9 and indicate relevant contributions of NR processes such as ET among the Er3þ ions, CR and multi- phonon relaxation. We recall that, in general, the measured lifetime of RE levels for Stokes and anti-Stokes excitation are not equal due to the nonequivalent ions inside the inho- mogeneous bandwidth and/or the Stark sublevels that partici- pate in the processes. For example for the GGS-GLASS, the 4S3/2 level lifetime for the excitation at 532 nm was 25 ls.9 In order to identify the UC pathway for pulsed excitation at 1480 nm, the PL spectra were studied with respect to the laser intensity dependence and time behavior. The dependence of the PL intensity versus the laser intensity was cubic indicat- ing that three laser photons are absorbed for each emitted UC photon. The ET process involving three excited Er3þ ions, described in Sec. III A, is not probable because the excitation occurs during each laser pulse while in the experiment with CW lasers there is dominance of the ET between the ions due to the large lifetime of levels 4I13/2 and 4I9/2. Two possible ways to reach the 2H11/2 level may be consid- ered. One possibility is the direct excitation of the Er3þ ions: the resonant one-photon absorption 4I15/2! 4I13/2, followed by the two-photon absorption transition 4I13/2 ! 2H11/2 without real intermediate resonant level. Other possible way is the three-photon absorption (3PA) by the host matrix followed by ET to the Er3þ ions. With basis on the temporal behavior of the UC signals described below, we could identify that the 3PA is the more efficient process. Fig. 5 shows the PL spectra of the GGS-GLASS and the GGS-GC for pulsed excitation at 1480 nm. Notice a splitting of the PL bands in the GGS-GC indicating that the Er3þ ions participating in the UC process are located in the crystalline phase. The weak PL signal observed for the GGS-GC is understood considering: (1) the small fraction of the total number of Er3þ ions inside the nanocrystals; (2) the time behavior of the PL signals shown in Fig. 6 that indicates larger NR relaxation of the Er3þ ions in the GGS-GC. As discussed below the smaller band gap of the GGS-GC favors a larger energy back transfer (EBT) rate from the Er3þ ions to the host. These considerations support the statement that the Er3þ ions contributing for UC in the pulsed experiment are not the same ions probed in the CW experiment and explain why the splitting of the PL bands are not observed in Fig. 2. The time evolution of the UC signals, recorded using a detection system having response of �10 ns, is illustrated in Fig. 6. Figs. 6(a) and 6(b) show the PL signals associated to transitions 4S3/2 ! 4I15/2 and 4F9/2 ! 4I15/2 for the GGS- GLASS, and Figs. 6(c) and 6(d) refer to the GGS-GC. Notice that the signals grow after the laser pulse and reach a maximum value followed by a decay of several microsec- onds. The signals corresponding to the GGS-GLASS show two decay time components: fast decay (�20 ls) and slow decay (>100 ls). However, the slow decay is not observed in the GGS-GC PL signal. Figs. 6(a) and 6(b) show fits of the expression [exp(�t/ sr)–Aexp(�t/sd1)–Bexp(�t/sd2)], where sr and sd1, sd2 are the rise and decay times, respectively. The fitting of the 4S3/2 ! 4I15/2 signal provided sr¼ 800 ns, sd1¼ 15ls, and sd2¼ 65 ls. For the 4F9/2 ! 4I15/2 transition, we determined sr¼ 800 ns, sd1¼ 20 ls, and sd2¼ 950 ls. The data shown in Figs. 6(c) and 6(d) were fit using [Cexp(�t/sr) – Dexp(�t/sd)] with FIG. 5. UC emission spectra under pulsed laser excitation at 1480 nm: GGS- GLASS (green solid line) and GGS-GC (red dashed line). 083520-4 Lozano et al. J. Appl. Phys. 113, 083520 (2013) [This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP: 200.145.3.45 On: Wed, 05 Feb 2014 15:49:01 sr¼ 600 ns and sd¼ 3.8 ls for the transition 4S3/2! 4I15/2 and sr¼ 2 ls and sd¼ 20 ls for the transition 4F9/2! 4I15/2. The temporal behavior indicates that the direct excita- tion of the Er3þ ions can be ruled out because this process would correspond to �5 ns signal rise time. Therefore we concluded that the 3PA by the host matrix followed by ET to the Er3þ ions is the dominant pathway and it is relevant because of the large excitation intensity obtained with the pulsed laser (1.34 GW/cm2). To model the PL decay from level 4S3/2 we recall that sd is determined by radiative (WR) and nonradiative (WNR) probability rates in such way that sd �1¼WRþWNR. The val- ues of WR are obtained from Ref. 9. Generally, WNR includes contributions from multi-phonon relaxation and CR with probability rates WMP and WCR, respectively. In a multi- phonon relaxation process, the energy stored in the RE ion is released to the host by emission of p phonons of energy �hx. On the other hand, the efficient CR processes (4S3/2; 4I15/2) ! (4I13/2; 4I9/2) and (4S3/2; 4I15/2) ! (4I9/2; 4I13/2) were reported for Er3þ doped glasses and crystals by various authors.15–18 In the present case, due to the proximity between the 4S3/2 level energy and the host band edge an extra term related to the energy back transfer probability, WEBT, from the excited Er3þ ions to the host has to be included in such way that sd �1¼WRþWMPþWCRþWEBT. The multi-phonon decay rate22 can be written as WMPðTÞ ¼ B½1þ nðTÞ� pe�a DE, where a and B are phenome- nological parameters, nðTÞ ¼ ½e�hx=kT � 1��1 represents the density number of thermally generated phonons per mode at the absolute temperature T, and k is Boltzmann0s constant. If the energy gap to the next lower energy level is DE, the num- ber of phonons emitted will be p ¼ DE=�hx. Notice that although phonons of any energy may be involved in the relaxation process, phonons of larger energy dominate and we considered for the calculations of WMP(T) the maximum phonon energy of our samples (345 cm�1) and the values B� 106 s�1 and a¼ 2.9� 10�3 cm found in Ref. 23. Table I presents the values for sd �1, WMP, and (WCRþWEBT) for the GGS-GLASS and the GGS-GC. From the results, we notice that the value of WCRþWEBT is larger for the GGS-GC sample. The increase with respect to the GGS-GLASS is attributed to the smaller band gap of the glass-ceramic sample and because the Er3þ ions participating in the UC process are located inside the nanocrystals. We recall that WEBT¼ 3� 105 s�1 was determined in Refs. 19–21 by analyzing the light induced conductivity of samples con- taining 1.8 at. % of Er3þ ions. The values obtained here have an approximate order of magnitude. Notice also that the value of WNR corresponding to state 4F9/2 does not change due to the heat-treatment of the GGS-GLASS because the energy of this state is smaller than Eg. FIG. 6. Dynamics of the UC signal for excitation at 1480 nm. (a) and (b): GGS-GLASS; (c) and (d): GGS-GC. TABLE I. Decay time (sd) and (WMP—multi-phonon; WCR—cross-relaxation; WEBT—energy back transfer) relaxation rates corresponding to transitions 4S3/2 ! 4I15/2 and 4F9/2! 4I15/2. The values of sd �1 for the GGS-GLASS refer only to the fast decay part. Sample sd �1 (103 s�1) WMP (s�1) WCRþWEBT (103 s�1) GGS-GLASS 4S3/2! 4I15/2 66.7 6 3.3 114 6 6 60.3 6 3.0 4F9/2! 4I15/2 50.0 6 2.5 349 6 18 43.2 6 2.2 GGS-GC 4S3/2! 4I15/2 263.2 6 13.2 114 6 6 256.8 6 12.8 4F9/2! 4I15/2 50.0 6 2.5 349 6 18 43.2 6 2.2 083520-5 Lozano et al. J. Appl. Phys. 113, 083520 (2013) [This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP: 200.145.3.45 On: Wed, 05 Feb 2014 15:49:01 IV. SUMMARY In summary, we investigated the luminescence behavior of Er3þ-doped Ga10Ge25S65 glass and glass-ceramic samples with respect to the laser excitation intensity and the time evolution of the frequency upconverted emissions. CW and pulsed lasers operating at 980 nm and 1480 nm were used. The experiments allowed identification of the mechanisms contributing to the infrared-to-visible wavelength conversion. Frequency upcon- version processes involving excited-state-absorption by the Er3þ ions, three-photon absorption by the matrix host followed by ET to the Er3þ ions, and energy transfer among Er3þ ions, were characterized. Besides the relaxation mechanisms of energy transfer among the erbium ions and multi-phonon decay, our results indicate an important contribution of energy back transfer from the excited ions to the host that also contributes to reduce of the decay time of the upconverted luminescence. ACKNOWLEDGMENTS We acknowledge financial support from the Conselho Nacional de Desenvolvimento Cient�ıfico e Tecnol�ogico (CNPq) and the Fundaç~ao de Amparo �a Ciência e Tecnologia do Estado de Pernambuco (FACEPE). The work was performed in the framework of the Photonics National Institute project (INCT de Fotônica) and PRONEX. 1F. Auzel, Chem. Rev. 104, 139 (2004). 2M. Yamane and Y. Asahara, Glasses for Photonics (Cambridge University Press, Cambridge, 2000). 3R. A. H. El-Mallawany, Tellurite Glasses Handbook—Physical Properties and Data (CRC, Boca Raton, FL, 2001). 4B. J. Eggleton, B. Lutter-Davies, and K. Richardson, Nature Photon. 5, 141 (2011) and references therein. 5N. Mehta, J. Sci. Ind. Res. 65, 777 (2006). 6S. Kasap, K. Koughia, G. Soundararajan, and M. G. Brik, IEEE J. Sel. Top. Quantum Electron. 14, 1353 (2008). 7V. K. Rai, C. B. de Ara�ujo, Y. Ledemi, B. Bureau, M. Poulain, and Y. Messaddeq, J. Appl. Phys. 106, 103512 (2009). 8V. K. Rai, C. B. de Ara�ujo, Y. Ledemi, B. Bureau, M. Poulain, X. H. Zhang, and Y. Messaddeq, J. Appl. Phys. 103, 103526 (2008). 9M. L. Frej, E. Valdez, C. B. de Ara�ujo, Y. Ledemi, and Y. Messaddeq, J. Appl. Phys. 108, 093514 (2010). 10R. Balda, S. Garcia-Revilla, J. Fern�andez, V. Seznec, V. Nazabal, X. H. Zhang, J. L. Adam, M. Allix, and G. Matzen, Opt. Mater. 31, 760 (2009). 11C. G. Lin, L. Calvez, M. Roz�e, H. Tao, X. H. Zhang, and X. J. Zhao, Appl. Phys. A 97, 713 (2009). 12Y. Ledemi, B. Bureau, L. Calvez, M. Le Floch, M. Roz�e, C. Lin, X. H. Zhang, M. Allix, G. Matzen, and Y. Messaddeq, J. Phys. Chem. B 113, 14574 (2009). 13C. B. de Ara�ujo, L. S. Menezes, G. S. Maciel, L. H. Acioli, A. S. L. Gomes, Y. Messaddeq, A. Florez, and M. A. Aegerter, Appl. Phys. Lett. 68, 602 (1996). 14D. M. da Silva, L. R. P. Kassab, S. R. Luthi, C. B. de Ara�ujo, A. S. L. Gomes, and M. J. V. Bell, Appl. Phys. Lett. 90, 081913 (2007). 15V. Q. Shi, M. Bass, and M. Birnbaum, J. Opt. Soc. Am. B 7, 1456 (1990). 16A. Lupei, V. Lupei, S. Georgescu, I. Ursu, V. I. Zhekov, T. M. Murina, and A. M. Prokhorov, Phys. Rev. B 41, 10923 (1990). 17R. El-Mallawany, A. Patra, C. S. Friend, R. Kapoor, and P. N. Prasad, Opt. Mater. 26, 267 (2004). 18H. T. Amorim, M. T. de Araujo, E. A. Gouveia, A. S. Gouveia-Neto, J. A. Medeiros Neto, and A. S. B. Sombra, J. Lumin. 78, 271 (1998). 19T. Yu. Ivanova, A. A. Man0shina, A. V. Kurochkin, Yu. S. Tver’yanovich, and V. B. Smirnov, J. Non-Cryst. Solids 298, 7 (2002). 20A. Tverjanovich, Ya. G. Grigoriev, S. V. Degtyarev, A. V. Kurochkin, A. A. Man0shina, T. Yu. Ivanova, A. Povolotskiy, and Yu. S. Tveryanovich, J. Non-Cryst. Solids 326&327, 311 (2003). 21T. Yu. Ivanova, A. A. Man0shina, A. V. Povolotskiy, Yu. S. Tver‘yanovich, S.-K. Liaw, and Y.-S. Hsieh, J. Phys. D: Appl. Phys. 41, 175110 (2008). 22R. S. Quimby and B. G. Aitken, J. Non-Cryst. Solids 320, 100 (2003). 23J. Heo and Y. B. Shin, J. Non-Cryst. Solids 196, 162 (1996). 083520-6 Lozano et al. J. Appl. Phys. 113, 083520 (2013) [This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP: 200.145.3.45 On: Wed, 05 Feb 2014 15:49:01 http://dx.doi.org/10.1021/cr020357g http://dx.doi.org/10.1038/nphoton.2011.309 http://dx.doi.org/10.1109/JSTQE.2008.921415 http://dx.doi.org/10.1109/JSTQE.2008.921415 http://dx.doi.org/10.1063/1.3259439 http://dx.doi.org/10.1063/1.2927402 http://dx.doi.org/10.1063/1.3493239 http://dx.doi.org/10.1016/j.optmat.2008.06.006 http://dx.doi.org/10.1007/s00339-009-5304-1 http://dx.doi.org/10.1007/s00339-009-5304-1 http://dx.doi.org/10.1021/jp905649v http://dx.doi.org/10.1063/1.116481 http://dx.doi.org/10.1063/1.2679798 http://dx.doi.org/10.1364/JOSAB.7.001456 http://dx.doi.org/10.1103/PhysRevB.41.10923 http://dx.doi.org/10.1016/j.optmat.2004.01.002 http://dx.doi.org/10.1016/S0022-2313(98)00002-7 http://dx.doi.org/10.1016/S0022-3093(01)01044-4 http://dx.doi.org/10.1016/S0022-3093(03)00420-4 http://dx.doi.org/10.1088/0022-3727/41/17/175110 http://dx.doi.org/10.1016/S0022-3093(03)00030-9 http://dx.doi.org/10.1016/0022-3093(95)00579-X