Influence of Al/Nd ratio on light emitting properties of Nd-doped glass prepared by sol-gel process

Transcription

1 nfluence of Al/Nd ratio on light emitting properties of Nd-doped glass prepared by sol-gel process Q. Xiang, Y. Zhou, Y. L. Lam, B. S. Ooi, Y.. han and. H. Kam Photonics Research Group, School of Electrical and Electronic Engineering Nanyang Technological University, Nanyang Avenue, Singapore ABSTRAT For rare earth doped silica-based glasses derived by sot-gel process, Al was used as a modifier in order to improve the dispersion of the rare earth ion in silica lattices, and thus, make the higher rare earth doped silica glasses without clustering possible. n this research, the influence of the ratio of Al to Nd on the fluorescence intensity and lifetime was studied in details to get the material which has a strong fluorescence intensity as well as the long fluorescence lifetime enough for integrated amplifier and laser use. Ten samples in the form of powder with different Al/Nd (in the range of 10 to 80) and different Nd (from 0.25 mole to 3 mole) concentration were prepared by sol-gel process. These powders were gotten by heating the dried gels in a furnace in air environment. The results of the fluorescence intensity and lifetime show that the AlfNd=1O with mole Nd, that is the recipe loosio2 : 1OAlO15 : Nd015 has the strongest fluorescence intensity in the ten samples. But its tile S only 11Ots. For loosio2 : 2OAlO15 : 0.25 Nd01, the obtained tile S 216ts without special OH movement treatment. The research results show that we need to balance the fluorescence intensity and the lifetime to choose the suitable recipe for practice use. Keywords: Nd doped glass, Al/Nd ratio, Active properties 1. NTRODUTON The rare earth doped glasses derived by sol-gel process is a promising candidate for material fabrication ofintegrated optical amplifiers and lasers. By the sol-gel process, it is very easy to tailor the materials composition to realize different functions. The process is also very flexible as the final materials can be made in the form of bulk, film, powder, and even special shapes. Meanwhile, the dopants can disperse in the host homogeneously with relatively high concentration duo to the wet chemistry nature of the process. For optical amplifier and laser use, the active properties, such as fluorescence strength, transmission lifetime, are essential. Since the optical interaction path in integrated optics is much shorter than in fiber, a large rare earth concentration is needed. But the large rare earth concentration will lead to rare earth ion cluster, which will reduce the fluorescence intensity and the lifetime because of the concentration quench and cross relaxation between ions. n rare earth doped silica glass derived by sol-gel process, aluminum or phosphate are added to disperse the rare earth ion in host lattices homogeneously. What is the optimum ratio of aluminum (or phosphate) to rare earth (Nd3 or Er3) ion? Some researchers adopt the ratio (atom ratio) of 10 : 1, and less than 2(at)% rare earth concentratio&'2'3. n general, there exist a optimum rare earth concentration for a certain amount Al or P dopants for the strongest PL intensity. But on that rare earth concentration, the fluorescence lifetime is relatively short. So, what is the suitable rare earth concentration for both strong photoluminescence (PL) intensity and long lifetime enough for lasing operation? n this research, we focused on Nd-doped Si02-Al203 material system and prepared ten samples with different ratio of Al to Nd and different Nd concentration. Their PL spectra and lifetime were measured and analyzed. Because powder is easier than thin film to prepare and to get the active optical properties, we analyzed the properties ofthe powders. But the results are also available for relevant films. The main purpose of this research is to compare the material properties of different compositions for further research. The processing parameters are not optimized. 2. SAMPLE PREPARATON Following the conventional sol-gel processing4, tetraethlorthosilicate(teos) was diluted in ethanol (EtOH) and water, with H added as a catalyst, let it hydrolyze in 60 for one hour. Al(N03)3 9H20 and Nd(N03)3.6H20 were used as precursors orrespondence: eqxiangntu.edu.sg; Tel: ; Fax: Part of the SPE onference on Design, Fabrication, and haracterization of Photonic Devices Singapore November December SPE Vol X/99/$1 0.00

2 ofa2o3 and Nd203 respectively. They are dissolved in EtOH, and then dropped into the pre-hydrolyzed TEOS solution. The end solutions stirred at room temperature in a covered bottle for one week, and then removed the cover. Several days later, they became dried gels. The high temperature treatment was carried out on all dried gels in a furnace by heating to 1000 in 12.5 hours and then holding at 1000 for 5 hours. The ten samples with different ratio of Alto Nd and Nd concentration are listed in Table. Table Sample List Sample No. ompositions Al/Nd ratio Nd (mole) 1 loosio2: 10 Al015: 0.25 Nd SiO2: 10A1015:0.5 Nd SiO2: 10A1015: 1 Nd SiO2: 20 A1015: 0.25 Nd SiO2:20A1015:0.5 Nd SiO2: 20 Al015: 1 Nd SiO2: 20 Al015: 2 Nd loosio2:40alo15: 1 Nd SiO2:40A1015:2 Nd loosio2:40a1015:3 Nd PHOTOLUMNESENE PROPERTES The experimental apparatus used for obtaining the photoluminescence spectra of Nd-doped Si02-A1203 glass powder is shown in Figure. 1. An Ar laser with 488nm emission was used as the pump source. The pump signal was mechanically chopped at 80Hz. The fluorescence light emitted is collected by two lenses, analyzed with an OREL (model 77200) monochromator and detected with an OREL (model 71905) detector associated with a high pass optical filter. The reflected pump light was cut by a long pass filter located in front of the monochrometor. Allthe spectra are recorded at room temperature. n order to get comparable results, every sample was held in a same quartz cell, all the other parametersin the measurement system kept constant during the measurement. The results are shown in Figure 2. sample Fig.1 Setup for measurement of PL Sample 1 to 3(called group one) has the same 10 mole Al but the Nd concentration was changed from 0.25 to 1 mole. As you can see in the Figure 2(a), the sample with loosio2: loalo15: 1NdO15 composition has strongest PL inthis group. There exist no any difference ofthe peak wavelength, they are all at lo6onm. The FWHM (full width ofhalfmaximum) are 56nm, 52nm and 44nm for 0.25at%Nd, 0.Sat%Nd sample and lat%nd respectively. Sample 4 to 7(called group two) has the same 20 mole Al but Nd concentration was changed from 0.25 to 2 mole. The results in Figure 2(b) show that the sample with 434

3 loosio2: 20A1015 : 2NdO15 composition has the strongest PL in this group. The peak wavelengths are all lo6onm as well. The FWHM are 58nm, 58nm, 56nm, and 54nm for O.25at%Nd, O.5at%Nd, lat%nd and 2at%Nd sample respectively. Similarly, in group three (sample 8 to 10) with 40 mole Al concentration, the recipe loosio2 : 40A101 :lndo1 has the strongest PL shown in Figure 2(c). The FWHM are 55nm, 57nm and 6Onm for lat%nd, 2at%Nd, and 3at%Nd sample respectively. No peak wavelength shift is observed for the ten samples. We think this results from the wavelength resolution limitation (2nm) of our measurement system. But the FWHM increases with the increment of Al concentration due to the wide distribution of Nd bonding sites with the presence of Al in host lattices. The maximum in each group is compared in figure 3. As you can see, the peak of the recipe loosio2:loalo15: 1NdO15 is highest and also strongest in ten samples. But the intensity difference among the three samples in Figure 3 is not great. That implies these three recipes are available only considering the PL intensity. 8.E-03 o > 0) U O.E+OO (a) Al/Nd = 10:x(x =0.25-4) (b) Al/Nd = 20 : x(x = ) 7.E-03 1.E-02 5.E-03 8.E-03 o ( E-03 1.E-03,) 0) O.E+OO (c)al/nd = 40 : x (x = 1 3) E Fig. 2 PL spectra of Nd : Si02-Al203 ofnd:si02-al203 glasses 4. LFETME Fig. 3 omparing ofthe PL spectra The fluorescence lifetimes for Nd doped silica glass were measured by exciting the samples with a pulsed semiconductor laser with the wavelength of 808nm. The pulse duration is about 100 ns, and the repeating rate is set at 1000 times per second. A Hamamatsu R5108 photomultiplier tube (PMT) was used as the detector, whose anode pulse rising time is 1.2 ns and the electron transiting time is 18 ns. A Tektronix TDS 360 digital oscilloscope was used in this system to record the decay. Figure 4 shows this setup. 435

4 12 1o OA 025Nd LD 808nm > Sample Filter _1' Digital Osiloscope 4 1OA: Nd 2 0 ióo T(ts) Fig.4 setup for measurement of the fluorescence decay ofnd:sio2-alo15 Fig.5(a) Al/Nd = 10:x (x = 0.25 i) T(ts) (b) Al/Nd = 20:x (x = ) T(jtS) (c) Al/Nd = 40:x(x = i 3) Fig.5 Fluorescence decay Figure 5(a), (b) and (c) shows the fluorescence decay of group one, two and three respectively. n each group, the higher the Nd concentration is, the faster the fluorescence decays. The decay curve of the samples from No. 1 to No.5 is pure exponential, the samples from No.6 to No. 10 is non-exponential. When the Nd concentration is increased, the average distance between ions is smaller and some degree of non-exponentiality starts to occur, even without clustering, because of cross relaxation. ross-relaxation is a special case of energy transfer where the original system (E3-E) losses the energy by decaying to the lower state E2 (which may also be the ground state E1 in Figure 6) and another system (E3'-E2') acquires the energy by going to a higher state E2'. The cross-relaxation between a pair of RE ions is graphically presented in Figure 6. t may take place between the same lanthanide (being a major mechanism for quenching at higher concentration in a given material) or between two differing elements, which happen to have two pairs of energy levels separated by the same amount. The two energy gaps may be equal or can be matched by one or two phonons. The possible cross-relaxation channelsfor Nd3 are 4F / tile 5 the fluorescence lifetime (the time is the time for fluorescence signal reaches a value of 1(O)/c where 1(0) is the value of the fluorescence signal at t=o). The ri/e value for the ten samples is listed in Table. At lower Nd concentration (O.25at%), the increment ofal can obviously increase the r11, for example, loal : O.25Nd, t11= 172ts, and 2OAl : O.25Nd, tl/e= 2l6pS. But at higher Nd concentration (more than O.5at%), the increment of Al can not only increase the 'ri/e but also decrease the i/e As examples, loal : O.5Nd, t ts, 20A1:O.5Nd, ri/e 14Ots; 1OA1: ind, t= hops, 20A1:2Nd, -r11= 68.ts. So, the function of the Al is more effective at low Nd concentration than at higher concentration (more than O.5at%). 436

5 Table Lifetime of Different Al/Nd and Nd samples Samples No Al/Nd ratio Nd (mole) tile Energy level on A 3' on A' 3> 12> 1'> *12'> Fig.6 Scheme for cross-relaxation between two ions ofthe same or different nature 5. DSUSSON AND ONLUSON For loat%a sample group, the PL intensity increases as the increment ofnd up to lat%, the samples with Nd concentration more than lat% were not prepared, we don't find the maximum Nd concentration in this group. For 2Oat%Al sample group, the PL intensity also increases with the increment ofnd concentration up to 2at%. But the PL intensity of2oal : 2Nd sample is a little lower than the intensity of loal : ind sample. Moreover the lifetime of2oal : 2Nd sample is much shorter than that of loal : ind sample. So, in practical, it is impossible increasing Nd concentration any more in group two samples. As above said, the intensity difference among the three samples in Figure 3 is not great, but their lifetime are quite different. For practical use, at the same level PL intensity and lifetime, we should choose the lower Nd and Al concentration recipe considering the possible crystalline and quench. Thus, as a conclusion, we think the recipe loosio2 : : lndo1 is the suitable one. REFERENES 1. Y. Zhou, Y. L. Lam, S. S. Wang, H. L. Liu,. H. Kam, and Y.. han, Fluorescence " enhancement of Er3-doped solgel glass by aluminium codoping" Appi. Phys. Lett. 71(5), August 1997, pp S.R. Natarajan, T.Srinivas, M. J. Joseph, and A. Selvarajan, "Fabrication of rare earth doped sol-gel based composite planar optical waveguides on glass", 1998 nternational onference on Applications of Photonic Technology, Ottawa, ANADA, July 27-30, X. Origaac, D. Barbier, X. M. Du, and R. M. Almeida, "Fabrication and characterization of sol-gel planer waveguides doped with rare earth ions", App!. Phys. Lett. 69(7), 12 August 1996, an M. Thomas, Stephen A. Payne and Gary D. Wilke, " Optical properties and laser demonstration ofnd-doped sol-gel Silica glasses",.j. Non-ry Solids, 151(1992)