Spectral broadening and luminescence quenching of 1.53 μm emission in Er3+-doped zinc tellurite glass

ARTICLE IN PRESS Journal of Luminescence 129 (2009) 270–276

Contents lists available at ScienceDirect

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

Spectral broadening and luminescence quenching of 1.53 mm emission in Er3+-doped zinc tellurite glass N. Jaba a,, H. Ben Mansour a, A. Kanoun a, A. Brenier b, B. Champagnon b a b

´ des Sciences, 5019 Monastir, Tunisia Laboratoire de Physique des Semi-conducteurs, De´partement de Physique, Faculte ˆt. A. Kastler, 10 rue Ampe `re, 69622 Villeurbanne Cedex, France Laboratoire de Physico-chimie des Mate´riaux Luminescents, UMR 5620 CNRS, Universite´ Claude Bernard Lyon 1, Ba

a r t i c l e in f o

a b s t r a c t

Article history: Received 30 January 2008 Received in revised form 9 August 2008 Accepted 13 October 2008 Available online 29 October 2008

The emission spectra and the lifetime of the lasing transition 4I13/2-4I15/2 in Er3+-doped TeO2-ZnO binary glass have been studied. The investigation includes Raman scattering spectroscopy as well as optical absorption, luminescence, and lifetime measurements techniques. The influence of erbium concentration on the line-shape of this electron transition has been analyzed. It was observed that the increasing of Er3+ ion concentration, in the 0.2–4 mol% range, results in a structural changes and a significant spectral broadening of the 1.53 mm emission band. Reabsorption has been evoked to explain the broadening of the 4I13/2-4I15/2 emission line. In the paper, is also reported the effect of the erbium content on the emission intensity of the 4I13/2-4I15/2 transition as well as on the lifetime of the 4I13/2 level. Based on the electrical–dipole interaction theory, the luminescence concentration quenching mechanism by hydroxyl groups is analyzed. The data suggest that o10% of hydroxyl groups are coupled to erbium ions in the zinc tellurite glass network. & 2008 Elsevier B.V. All rights reserved.

Keywords: Tellurite glass Erbium: Er3+ Reabsorption Hydroxyl groups

1. Introduction Due to the increasing demand for information capacity of wavelength division multiplexing (WDM) networks, it is desirable that the erbium-doped fiber amplifier (EDFA) has a broad and flat gain spectrum within the telecommunication window [1,2]. So far, many researchers have paid much attention to silicate [3], germanate [4], phosphate [5], fluoride [6], bismuth [7,8], and tellurite glasses [9], probably better suited for fiber amplifiers. Although the silica-based erbium-doped fiber has a good thermal stability, high chemical durability, and most importantly the mechanical stability, the silica-based EDFA still has some disadvantages especially on the narrow gain spectrum which only permits fewer channels to limit its application. However, telluritebased EDFA was reported to have 80 nm wide and flat gains up to 1.5 mm, which also shows various excellent material properties such as lowest phonon energy (750 cm1) among oxide glasses, high refractive index (2), high dielectric constants, strength and corrosion resistance over fluoride glass and rare-earth (RE) ion solubility [10–13]. Numerous special optical properties i.e., large emission cross-section and broad fluorescence full-width at halfmaximum identify Er3+-doped tellurite glasses as attractive materials for potential applications in high-performance fiber amplifiers, as well as optics and laser technology.

 Corresponding author. Tel.: +216 98 67 60 57; fax: +216 7350 02 78.

E-mail addresses: [email protected], [email protected] (N. Jaba). 0022-2313/$ - see front matter & 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.jlumin.2008.10.006

Recently, as a result of the development of optical network systems [14], there have been intensive studies on compact amplifiers such as short-length fiber amplifier or Er-doped planar waveguide amplifiers [15–19]. For the amplifiers with the short effective length, it is necessary that the doping concentration of Er3+ ions in the devices is sufficiently high. At higher Er3+ concentrations, however, the luminescence shows a quenching by energy transfer processes, due to interactions between Er3+ ions and hydroxyl (OH) groups [20,21], and in addition, another cooperative upconversion process occurs if a high pump power is applied as well [22]. Several works on Er3+-doped tellurite glasses have been reported by authors in the last decade [23–30]. There, mainly the effect of Er3+ concentration on the spectroscopic properties has been studied. However, up to now there is only a few experimental works on the effect of erbium content on structural properties [31–33] and on spectral broadening [34,35] have been reported. At the same time, many research works have been carried out to study the quenching effect of OH groups on the 4 I13/2 level of Er3+ in silicate, borosilicate and phosphate glasses [20,21]. Nevertheless, little is known about the OH quenching on emission properties in different Er concentrations of Er3+-doped tellurite glasses. Thus, it is very necessary to reveal the effects of high erbium doping levels on the local field of erbium sites, the spectral broadening at 1.53 mm and concentration quenching. Such investigations should provide great progress in the engineering of EDFA. This paper reports on the effects of the Er2O3 content on the local structure and on spectroscopic properties of Er3+-doped

ARTICLE IN PRESS N. Jaba et al. / Journal of Luminescence 129 (2009) 270–276

2. Experimental

1.5 1.2 0.9

Int. Raman Band at 750 cm-1

1.8

Raman intensity (a.u.)

tellurite zinc (TZ) glass system as well. The doping concentration ranges from 0.2 to 4 mol% Er2O3. Especially, the effect on the spectral broadening of the 1.53 mm emission is analyzed and discussed. The analysis of OH and Er3+ concentration quenching in erbium-doped TZ glass has been also carried out. Glasses were investigated using Raman spectroscopy, optical absorption, photoluminescence (PL), and time-resolved luminescence measurements at room temperature (RT).

271

3+

1.7

4 mol. % Er undoped

1.6 1.5 1.4 1.3 0

1

2 3+

Er

3

4

(mol. %)

0.6 0.3

Glasses were prepared from oxide powders of TeO2, ZnO, and Er2O3 as starting materials using the conventional melt-quenching method. Batches of 10 g were melted in platinum crucibles at 900 1C in an electrically heated furnace in the air atmosphere. In order to investigate the effect of OH groups on concentration quenching, no removing OH process by bubbling oxygen was adopted in melting glass. The amount of dopant was varied between 0.2 and 4 mol% Er2O3. The visible (vis) and near-infrared (NIR) absorption spectra were measured using a Perkin-Elmer UV/ vis/NIR Lambda 900 spectrophotometer in the 400–1700 nm range. IR transmission spectra were measured using a PerkinElmer GX II FTIR spectrophotometer. PL spectra were recorded by exciting the samples with a cw near-infrared Ti:saphire laser tuned to 797 nm pumped with an Ar+ ion laser Spectra-Physics 2017. The emitted light was dispersed by a Jobin–Yvon HRD1 monochromator and detected with a Hamamatsu HV 1250 photomultiplier. The signal from the detector was preamplified and passed to a lock-in amplifier whose reference was a variable speed light chopper in the excitation beam. The experimental lifetimes of the levels were obtained by exciting the samples with a laser analytical systems dye laser, tuned to 980 nm, pumped by a pulsed frequency doubled Nd:YAG laser from IBM Industries. The duration of pulses was 8 ns. The emitted light has been focused on a Jobin–Yvon HR S2 spectrophotometer. The detection has been performed using an R 1767 Hamamatsu photomultiplier and a Lecroy 9410 averager oscilloscope. Raman measurements were performed using a double grating spectrometer (XY Dilor) with the 457.9 nm Ar+ ion laser exciting line. All Raman spectra were recorded, in the wavenumber range 300–1000 cm1 under a vertical–vertical (VV) polarization, with a spectral slit width of 0.6 cm1. All measurements were recorded at RT.

3. Results and discussion 3.1. Raman analysis In order to investigate the evolution of the glass structure with the adding of Er2O3, Raman measurements were carried out on undoped and Er3+-doped TZ glass samples. Fig. 1 shows VV Raman spectra of the base line TZ glass and the 4 mol% Er3+-doped glass, both have been measured under 457.9 nm excitation line. In the wave-numbers range 300–1000 cm1, each spectrum exhibits two bands centred around 675 and 750 cm1 and a third smaller amplitude band located at 450 cm1. The band at 675 cm1 is assigned to the stretching vibrations of the TeO4 trigonal bypiramidal (tbp) groups. They are linked through Te–O–Te, with O in a position alternatively axial and equatorial, and form the backbone of pure TeO2-based crystals or glasses [36]. While the band at 750 cm1 arises from TeO3 and TeO3+1 vibration [37,38]. To determine the ratio of the Raman signal of the 750 cm1 Raman band for doped samples to that in the undoped glass, we have normalized Raman spectra relative to the maximum intensity of the band at 675 cm1. As shown, the adding of erbium oxide

0.0 300

450

600

750

900

Raman shift (cm-1) Fig. 1. RT Raman spectra of the TZ glass : (J) undoped glass and (m) 4 mol% Er2O3doped glass. The inset shows the intensity of 750 cm1 Raman band versus erbium ion concentration.

results in a growing of the peak around 750 cm1 (inset of Fig. 1) with a slight blue shift with the increasing Er3+ ion concentration. Controversly, the intensity of the Raman band near 450 cm1 decreases with increased Er2O3 content. A shift towards smaller wave-numbers has been observed for the peak frequency of the band. It should be noticed that the presence of erbium, even in relatively high doping levels, does not lead to the development of structural peaks, indicating a very good dispersion of RE ions, with no evidence of the formation of clusters [36]. The progressive enhancement of the Raman peak intensity at 750 cm1 can be explained as due to the evolution of TeO3+1 and TeO3 structures with the increasing erbium concentration. The results show that the erbium ion acts as a lattice modifier in the binary tellurite glasses, i.e., it converts TeO4 tbp units to primarily TeO3 trigonal pyramids units, and possibly some terminal TeO3+1 polyhedra [39]. The significant change of the 450 cm1 Raman band, associated with the Te–O–Te linkages, is attributed to a structure disruption of the tellurite network and to a decrease in the Te coordination number as the erbium content increases [40]. Initially, TZ glass contains a variety of structural motives (TeO4, TeO3, and TeO3+1) due to the presence of Zn2+ modifier ions, which give rise to a large distribution of structural sites. Adding erbium ions further enhances the variety of the network-forming TeOx species, resulting in the enhancement of structural disorder. As a consequence, the addition of erbium ions in the tellurite glass system induces a diversity of dopant sites and then may gives rise to a broadening of the emission lines. In this context, Jha et al. [35] suggest that with the increasing dopant concentrations in a tellurite glass network, the Stark sublevel splitting increases, thereby indicating that the Er3+ ions occupy new sites at higher concentrations, which leads to a further line-shape modification [35]. The variations in the Er3+ local environment, symmetry, and ligand field strength may induce spectral broadening of the Er3+ emission. 3.2. Emission spectra and cross-section at 1.5 mm Raman analysis demonstrates that, with increasing erbium content in the glass network, some structural changes take place. A spectroscopic study is indeed required to investigate the importance of these changes and their possible role on optical properties. Fig. 2 shows RT luminescence spectra in the wavelength range 1450–1700 nm obtained on the lightly (0.2 mol%) and heavily (4 mol%) doped TZ glass pumped with 797 nm excitation line. The figure also shows the 4I15/2-4I13/2 absorption band of an

ARTICLE IN PRESS N. Jaba et al. / Journal of Luminescence 129 (2009) 270–276

1.0

Absorption 0.2 mol. % Er2O3 4 mol. % Er2O3

0.8

115 110 105 100 95 90 85 80 75 70

10

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 Er2O3 concentration (mol. %)

0.6

Absorption McCumber Meas. PL

8 Cross-section (x 10-21 cm2)

Normalized emission (a.u)

1.2

Effective bandwidth (nm)

272

0.4 0.2

6

4

2

0.0 1450

1500

1550 1600 Wavelength (nm)

1650

1700

1450

Fig. 2. Normalized absorption and emission spectra versus the Er3+ concentration. The inset depicts the Er3+ concentration-dependent effective bandwidth for 1.53 mm emission band.

Er3+-doped TZ glass. The intensities of both absorption and emission lines have been normalized relative to the intensity at the peak wavelength lp ¼ 1.53 mm. The main features observed are: (i) the peak position of absorption and emission lines remain unchanged, (ii) large overlap of the absorption and emission spectra around the peak wavelength lp, and (iii) the higher the dopant concentration, the broader the 1.53 mm emission. It is worth noticing that absorption analysis shows that no significant change is observed for the line-shape and/or for the bandwidth of absorption bands as erbium content is increased. Because the 1.53 mm emission band of Er3+ ions in glasses is asymmetric, choosing the effective bandwidth as a semi-quantitative indication is more meaningful rather than the full-width at half-maximum [41,42]. The definition of the effective bandwidth R according to Weber [11] is Dleff ¼ I(l)dl/I(lp), where I(l) is the emission intensity at the wavelength l, and I(lp) is the intensity at the peak wavelength lp. The inset of Fig. 2 illustrates the dependence of effective bandwidth as a function of the Er3+ concentration. As can be seen, Dleff increases linearly from 77 to 108 nm with the erbium concentration. To obtain the intrinsic 1.53 mm emission line free from reabsorption, the emission spectra of Er3+ in TZ glass are calculated according to the reciprocity method of McCumber–Miniscalco [43,44], which relates the absorption and emission crosssections by

se ðnÞ ¼ sa ðnÞ exp ½ð  hn=kTÞ

0

(1)

where se is the stimulated emission cross-section, sa is the absorption cross-section, n is the photon frequency, h is the Plank’s constant, k is the Boltzmann constant, and e is the net free energy required to excite one Er3+ ion from the 4I15/2 to 4I13/2 state at temperature T. Using absorption data, the estimated value of absorption cross-section at peak value is 8.16  1021 cm2. The energy e can be estimated using the method proposed by Miniscalco et al. [44]. Using this latter procedure [44], we obtain the value of e ¼ 6547 cm1. The peak emission cross-section, as calculated using Eq. (1), is 8.81 1021 cm2 at 1531 nm. Fig. 3 depicts the absorption cross-section spectrum, the calculated emission cross-section by McCumber theory, and the measured fluorescence line-shape of the 0.2 mol%-doped glass sample. It is worth noticing that the two cross-sections are reported with the indicated scale (1021 cm2), but the luminescence is reported in an arbitrary scale for comparison of shapes. As can be seen in Fig. 3, the shape of the calculated stimulate emission cross-section is similar to that of the luminescence of the

1500

1550 Wavelength (nm)

1600

1650

Fig. 3. Absorption cross-section, calculated emission cross-section and measured PL for the 1.53 mm emission band. The two cross-sections are reported with the indicated scale (1021 cm2), the luminescence (Meas. PL) is reported in an arbitrary scale for comparison of shapes.

sample with the lowest erbium content, the differences being due to reabsorption, as it will discussed later. Because of the overlap of the absorption and the emission spectra of Er3+ ions at 1.53 mm, the broadening trend of the emission bandwidth should mainly be attributed to reabsorption (self-absorption) [34]. This phenomenon always occurs in a typical 3-level system when the absorption and emission fluorescence spectra overlap [45,46], such as 5I7-5I8 in Ho3+ at 2.1 mm, 3F4-3H6 in Tm3+ at 2.0 mm and 4I13/2-4I15/2 in Er3+ at 1.53 mm. With the increasing erbium content, the emitted fluorescence may be affected significantly by self-absorption [36]. Then, the measured fluorescence should be largely deformed relative to the intrinsic emission calculated by McCumber theory, particularly at the peak wavelength. In fact, due to the important overlap of the absorption and emission spectra around the 1.53 mm peak, reabsorption around this wavelength energy is much more efficient than that at other wavelengths [34]. Consequently, the measured emission intensity around the peak wavelength is lower than their intrinsic values and then one can observe bandwidth broadening with the increasing erbium content. It is worth to mention, based on Raman analysis, that erbium ions influence the Te–O local structure as oxide modifiers in binary or ternary tellurite-based glasses [47–50]. However, optical properties of active RE ions are not sharply affected by the structural changes induced by the increasing erbium content. Absorption analysis confirms this assertion. In fact, optical absorption does not see any effect as erbium content increases even for highly doping levels. Then, one can conclude that the structural change are important, but, contrarily to the suggestion of jha [35], do not substantially affect the local field of erbium sites and that the observed increase of line broadening in luminescence is practically only due to reabsorption. 3.3. Concentration quenching We have investigated the effects of the Er3+ ion concentration on the PL intensity of the 4I13/2-4I15/2 electron transition. It was found that this intensity shows a quenching behavior beyond a critical Er3+ ion concentration at 1 mol% (Fig. 4). The inset of Fig. 4 shows the plot of intensity/concentration as a function of Er3+ ion concentration. As can be seen, the quenching concentration

ARTICLE IN PRESS N. Jaba et al. / Journal of Luminescence 129 (2009) 270–276

400 300 200 100

Intensity (a.u)

Intensity/concentration

appears more evident. The quenching in luminescence intensity may be assigned to a non-radiative energy transfer. To analyze the dynamic of this energy transfer, we have studied the luminescence decay of the 4I13/2-4I15/2 transition as a function of the doping concentration. Fig. 5 depicts the PL decay profiles of Er3+, measured for three different erbium concentrations: 0.5, 2, and 4 mol% Er2O3. Three main observations are revealed: (i) the decays show an initial rise time which is due to the populating of 4I13/2 level from the higher 4I11/2 state directly excited, (ii) the decay curves show an exponential behavior with ion-concentrationdependent time constant, and (iii) high ion concentration levels

0 0

1 Er

0

1

3+

2

3

4

concentration (mol.%)

2

3

4

Er3+ concentration (mol. %) Fig. 4. Dependence of 1.53 mm emission line intensity on the Er2O3 content. The inset illustrates the plot of Intensity/concentration as a function of erbium concentration.

0.5 mol.% 2.0 mol.% 4.0 mol.%

Intensity (a.u.)

1

0.1

0.01

1E-3

0

4

2

6

8

10

Time (ms) Fig. 5. RT PL decays of the level 4I13/2 under 980 nm laser excitation line for three different Er3+ ion concentration.

273

speed up the luminescence decay significantly. Total experimental decay time (tTot), as deduced from the decay curves, are summarized in Table 1 versus the Er3+ ion concentration. Two types of doping-related processes can influence the experimental lifetime detected: concentration quenching and self-absorption. The first process causes a decrease of the excited state lifetime and luminescence time decay, whereas self-absorption causes a delay in the detection of the excited luminescence because of successive reabsorption and re-emission processes inside the sample, depending on the thickness L and absorption cross-section sa [51]. When self-absorption occurs, according to Auzel’s approach [52], we have tTot ¼ tPL[1+saNErL], where tPL is the value not influenced by self-absorption and NEr is the erbium ion concentration. Corrected tPL values have been calculated and gathered in Table 1. The reliability of the correction for selfabsorption to the lifetime has been checked through timeresolved luminescence measurements on the 1 mol%-doped sample that has been powdered to minimize self-absorption effects. The experimental lifetime value observed in the powdered sample, indeed shorter than in bulk (value in square brackets in Table 1), is in good agreement with the value of bulk sample after correction. It is worth to notice that the measured lifetime decreases with the increasing erbium content without relevant change in the kinetics (Fig. 5), which remains a single-exponential process. In fact, in heavily Er3+-doped glasses, the lifetime shortening is usually accompanied by pronounced deviations from the singleexponential kinetics. Such a phenomenon is not observed in the TZ glass system investigated. Therefore, the lifetime shortening evidences the activation of ion–ion energy transfer that makes the erbium emission faster and faster. Nevertheless, the singleexponential kinetics gives clear evidence that the mechanism of energy transfer is homogeneous inside the network, without preferential positions regarding the excitation transfer towards sites of non-radiative decay [51,53]. This requires that the energy transfer process has to be fast, in order to allow a spatial equilibrium of the excitation within the RE ions system in a time shorter than the decay time. These results are little surprising for glasses where inhomogeneous broadening of the line-shapes is large, decreasing the number of excited RE impurity ions in perfect resonance, and thus the speed of the energy migration. Moreover, from Table 1, it is seen that the extrapolated value of tPL at zero concentration is 4.09 ms. The estimation of the radiative lifetime value is indeed required to understand deexcitaion mechanisms. To evaluate the radiative decay rate of the 4 I13/2-4I15/2 transition, the intensity parameters Ot ¼ 2,4,6 have been calculated using J–O formulae [54,55]. The relevant values are O2 ¼ 5.93  1020 cm2, O4 ¼ 1.50  1020 cm2, and O6 ¼ 1.07  1020 cm2 [56]. From these parameters, the spontaneous emission probability, WR, can be evaluated. In the framework of the J–O theory [54,55], the spontaneous emission probability is expressed 3 as W R ¼ ð64p4 =3hð2J þ 1Þl¯ ÞwSED þ W MD , where l¯ is the average transition’s wavelength, (2J+1) is the ground state degeneracy, and w is the field correction factor w ¼ (n(n2+2)2/9). n is the refractive index. SED is the line strength of the electric-dipole transition.

Table 1 Total decay time (tTot) of photoluminescence, PL decay time (tPL) corrected for self-absorption, and emission quantum efficiency (Z) of 4I13/2 excited state. Er2O3 (mol%) NEr (1020 cm3) tTot (ms) tPL (ms) Z (%)

0 4.08 extrapolated 97

0.2 0.94 3.33 2.23 46

0.5 2.28 2.69 1.59 37

The value in square brackets is obtained on powdered sample doped with 1 mol% Er2O3.

1.0 4.56 1.81 [1.03] 1.12 26

2.0 9.13 1.25 0.66 16

3.0 13.69 0.85 0.45 11

4.0 17.65 0.53 0.35 9

ARTICLE IN PRESS 274

N. Jaba et al. / Journal of Luminescence 129 (2009) 270–276

According to Weber [57], SED is given by 4

4

SED ½ I13=2 ; I15=2  ¼ 0:019 O2 þ 0:118 O4 þ 1:462 O6

(2)

WMD is the magnetic-dipole contribution independent on ligand fields [58,59]. It can be estimated using the relationship, WMDn3 [60,61], and the value reported for LaF3 [57]. Using the J–O intensity parameters, we have calculated the radiative decay rate with n ¼ 1.95 at 1.53 mm. As has been found WR is equal to 239 s1. The radiative lifetime of an emitting level is defined as: tR ¼ 1/WR. The emission quantum efficiency of a related electronic transition can be calculated from the relation: Z ¼ (tPL/tR) ¼ WRtPL. The relevant radiative lifetime value tR is equal to 4.18 ms. This value is very close to the extrapolated value of tPL at zero concentration (4.09 ms). Consequently, the quantum efficiency at zero erbium concentration is around 96%. This means that the deexcitation of the 4I13/2 level is mainly governed by radiative decay process [62]. However, the experimental results show a rapid decrease in the quantum efficiency with increased Er3+ concentration (Table 1). The highest efficiency of about 55% is achieved for the sample with the lowest erbium content. It reduces to 26% for 1 mol% of Er3+-doped sample. The observed decrease in the quantum efficiency of the 4I13/2 level with the increasing Er3+ concentration may be related to an increase in the energy transfer rate involving both Er3+ ions and quenching centres, probably OH groups. 3.4. Modelling of Er3+ luminescence quenching It has been suggested by several authors that OH groups are serious quenchers of the photoluminescence of Er3+ ions in phosphate and alkali silicate glasses [20]. OH groups are also considered as important quenching centres in tellurite glasses as well [63,64]. IR absorption spectroscopy can provide information, in particular, about OH groups in the glassy network. Fig. 6 shows the typical IR absorption spectrum of undoped TZ glass in the spectral range 1800–3800 cm1. The broad and strong absorption band located between 2500 and 3500 cm1 is associated with the stretching vibration of OH. The same melting process being adopted, there is no noticeable influence upon lineshape and/or for the bandwidth of absorption bands with the increasing erbium content. Additionally, the absorption coefficient of the OH vibration band at 3000 cm1, aOH, for a sample of thickness L is defined as aOH ¼ (Ln(I0/I)/L), where I0 and I are, respectively, the incident and transmitted intensities. The free OH

content can be estimated, in terms of the OH absorption coefficient at 3000 cm1, as [62] NOH ¼

N



aOH

(3)

where N is the Avogadro constant and e is the molar absorptivity of the free OH groups in the glasses equal to 49.1 103 cm2/mol. [62]. The estimated absorption coefficient of the OH vibration band at 3000 cm1 for this modified tellurite glass does not show a significant change as the erbium ion concentration varies. It, however, presents an average trend with respect to the value of 4.45 cm1. Thus, the free OH content in the TZ glass studied is found to be in the order of 5.48  1019 cm3. In Ref. [65], the authors concluded that Er3+ PL lifetime reduction has a combined rather than a separate dependence on both the Er3+ and OH concentrations. A possible two-phonon quenching mechanism by OH groups can occur [65]. In fact, the energy gap of Er3+: 4I13/2-4I15/2 transition (6500 cm1) corresponds to the energy of the second harmonic of the OH stretching vibration. Non-radiative relaxation of the 4I13/2 level can take place by exciting two OH vibrational quanta. Based on this assumption, a simplified energy transfer and quenching model, is described here to explain the observed Er3+ luminescence quenching. According to Fo¨rster–Dexter theory [66,67] for the dipole– dipole interaction, the resonant energy transfer rate between two nearby Er3+ ions can be expressed using the following formula:  6 Z 4 3h c4 Q a f d ðEÞf a ðEÞ R0 1 dE ¼ (4) W dd ¼ 5 4 6 4 R tR 64p n R tR E where c is the speed of light, E is the energy of the 4I13/2-4I15/2 transition, the factors fd(E) and fa(E) are the normalized line-shape R functions, with f(E)dE ¼ 1 for donor emission and acceptor absorption band, respectively. R is a distance between donor and acceptor determined by the Er3+ density, NEr, using the formula R ¼ (4pNEr/3)1/3, R0 is called the critical transfer distance in the sense that the energy transfer rate for an isolated donor–acceptor pair separated by R0 occurs with the same rate as the spontaneous R deactivation in the donor itself (WR) and Qa ¼ sa(E)dE is the integrated absorption cross-section of the energy-acceptor level. By rearranging Eq. (4), R0 is given by 4

R60 ¼

3h c4 Q 64p5 n4 a

Z

f d ðEÞf a ðEÞ E4

dE

(5)

The interaction micro-parameter for the migration rate of 4I13/ transition, CErEr, can be calculated by the following equation if migration occurs through dipole–dipole interaction [68]:  6 C ErEr R0 W dd ¼ ¼ W (6) R R R6

4 2- I15/2

Absorption coefficient (cm-1)

5

4

3

2

1

0 2000

2500

3000

3500

Wavenumber (cm-1) Fig. 6. Infrared absorption spectra for the undoped TZ glass.

Therefore, CErEr can be expressed as: CErEr ¼ WRR06. We have applied this model to the TZ glass under study. The following input parameters used are:n ¼ 1.95, tR ¼ 4.18 ms, Qa ¼ 4.92  R 1045 m2 J and fd(E)fa(E)/E4 dE ¼ 26.80  1094 J5. We have obtained a value of CErEr in the order of 53  1040 cm6 s1 and a critical distance R0 ¼ 16.82 A˚. It worth noticing that the excitation transfer constant CErEr , in this work, is in the same range as the value found by Dai et al. [63] (46  1040 cm6 s1) but much larger than that of sodalime silicate glass and phosphate glass (23 and 7  1040 cm6 s1) [21]. According to Dai [63], the difference can be explained to some extent by the fact that, due to inhomogeneous broadening, the Er3+ PL spectrum in tellurite glass is broader and then gives rise to a larger mean overlap integral than that in silicate and phosphate glasses. Furthermore, due to the large index value in tellurite glass (2), radiative rates will be

ARTICLE IN PRESS N. Jaba et al. / Journal of Luminescence 129 (2009) 270–276

content estimated to NOH ¼ 5.48  1019 cm3. This result indicates that o10% of the free OH groups inside TZ glass are coupled to Er3+ ions, which causes the PL quenching. It should be noted that, at Er-doping concentrations below 4  1020 Er ions/cm3, the data no longer follow the linear dependence predicted by Eq. (8). Apparently, the energy transfer between Er ions is less severe with a low Er concentration, which leads to less quenching. Comparing the w value in the tellurite glass studied to that in phosphate glasses i.e., w60% [20], one can say that tellurite glasses are promising candidates for optical applications due to their lower quenching rates by OH groups compared to other oxide glasses. At the same time, the large spectral bandwidth of the stimulated emission cross-section makes this TZ glass very attractive candidate for broadband amplifiers in WDM systems.

3000 2500 WOH= 1/τPL - WR (s-1)

275

2000 1500 1000 500 0 2

0

4

6 3+

Er

8

10

14

12 20

16

18

20 4. Conclusion

-3

concentration (10 cm )

Fig. 7. Energy transfer rate (1/tET ¼ WOH) as a function of Er3+ concentration. The solid line is the fit of experimental data as obtained from Eq. (8).

much larger and therefore tellurite glass systems should have large values of excitation transfer constants. It is worth to mention that, the Fo¨rster–Dexter theory [66,67] deals with the microscopic case of two ions interacting with one another. Concerning the macroscopic case with many ions, the energy transfer process in a glassy matrix can be treated as an energy diffusion process [69]. The concentration-dependent PL lifetime, free from reabsorption, can be written as 1

tPL

¼

1

tR

þ

1

(7)

tET

1/tET accounts for the quenching rate due to energy transfer by electrical dipole–dipole interactions, and is dependent on both the Er3+ and OH concentrations. Further explanation for this term will be given later. Because of the low phonon energy (750 cm1), the multiphonon decay rate is too weak, so that it is neglected in the total decay rate. The following assumptions are proposed to build the present Er3+ concentration quenching model [20]: (i) the OH quenching centres are only coupled to a fraction of erbium ions, (ii) the amount of Er3+ ions coupled to OH groups is dependent on the OH concentration in the glass network, and (iii) non-radiative quenching occurs after the excited energy is transferred to Er3+ ions coupled to OH groups, possibly via two-phonon mechanism [20]. Thus, with the increasing erbium content beyond a critical concentration, an energy transfer between excited Er3+ ions and ground state Er3+ ions increasingly takes place. The excited energy will be lost whenever it is transferred to an Er3+ ion coupled to an OH quenching centre. According to the later assumptions, Eq. (7) can be rewritten as 1

tET

¼ W OH ¼

1

tPL

 W R ¼ 8pC ErEr NEr N ErOH

(8)

Here, WOH is the energy transfer rate between Er3+ ions and OH groups. Such a rate is proportional to the activators (Er ions) and quenchers (Er ions coupled to OH groups) concentrations. In Eq. (8), NErOH is the erbium ion concentration coupled to OH groups. By fitting the measured PL lifetime with use of Eq. (8), we have obtained from the slope of the plot erbium ion concentration coupled to OH groups NErOH ¼ 4.96  1018 cm3 (Fig. 7). As has been noticed by Houde-Walter et al. [70], the density of quenched Er3+ ions, is directly proportional to the OH group concentration through the empirical factor w, i.e., NErOH ¼ wNOH with 0owo1. Thus, we obtain w ¼ 9% for this TZ glass corresponding to free OH

In the present work, we have investigated the spectral features and concentration quenching of the 4I13/2-4I15/2 electron transition of Er3+ in a TZ glass. To this end, activated TZ glasses were prepared and characterized using Raman scattering, optical absorption, luminescence, and lifetime measurements techniques. The study is particularly focussed on two aspects: (i) the effect of the dopant ion concentration on spectral broadening of the 1.53 mm emission line, and (ii) the concentration quenching of the 4 I13/2-4I15/2 electron transition. Raman data show structural changes when increasing the concentration of optically active ions. Optical absorption does not see any effect with the increasing erbium content. The 1.53 mm emission band, however, shows a large broadening with increased Er3+ concentration. The relevant values of the effective bandwidth of 1.53 mm emission range from 77 to 108 nm. It is concluded from this study that structural changes induced by erbium incorporation do not substantially affect the local field of erbium sites and then do not induce spectral broadening. The broadening trend, as observed for the 1.53 mm emission line, is practically only due to reabsorption. On the other hand, concentration quenching by OH groups in Er3+-doped TZ glasses was investigated as a function of Er3+ ion concentration. The observed decrease in the PL intensity beyond a critical Er3+ ion concentration around 1 mol% and in the lifetimes of the lasing transition 4I13/2-4I15/2 with the increasing doping levels has been related to the increase in the energy transfer rate involving both Er3+ ions and the quenching centres as OH groups. A concentration quenching model is proposed. It is estimated that less than 10% of OH groups in this TZ glass are coupled to erbium ions.

References [1] M. Yamada, H. Ono, T. Kanamori, Y. Ohishi, Electron. Lett. 33 (1997) 710. [2] Y. Ohishi, A. Mori, M. Yamada, H. Ono, Y. Nishida, K. Oikawa, Opt. Lett. 23 (1998) 274. [3] J. Yang, S. Dai, N. Dai, S. Xu, L. Wen, L. Hu, Z. Jiang, J. Opt. Soc. Am. B 20 (2003) 810. [4] Z.M. Yang, S.Q. Xu, L.L. Hu, Z.H. Jiang, J. Opt. Soc. Am. B 21 (5) (2004) 951. [5] S. Jiang, T. Luo, B.C. Hwang, et al., J. Non-Cryst. Solids 263–264 (2004) 364. [6] K. Soga, H. Inoue, A. Mkishima, J. Non-Cryst. Solids 274 (2000) 69. [7] S. Tanabe, N. Sugimoto, S. Ito, T. Hanada, J. Lumin. 87–89 (2000) 670. [8] X. Zhang, T. Xu, Q. Nie, et al., Spectrochim. Acta Part A 67 (2007) 246. [9] H. Lin, E.Y.B. Pun, S.Q. Man, X.R. Liu, J. Opt. Soc. Am. B 18 (2001) 602. [10] J.S. Wang, E.M. Vogel, E. Snitzer, Opt. Mater. 3 (1994) 187. [11] M.J. Weber, J.D. Myers, D.H. Blackburn, J. Appl. Phys. 52 (1981) 2944. [12] Z. Pan, S.H. Morgan, K. Dyer, A. Ueda, H. Liu, J. Appl. Phys. 79 (1996) 8906. [13] L.L. Neindre, S.B. Jiang, et al., J. Non-Cryst. Solids 255 (1999) 97. [14] H. Hayashi, N. Sugimoto, S. Tanabe, S. Ohara, J. Appl. Phys. 99 (2006) 093105. [15] E. Snoeks, G.N. van den Hoven, A. Polman, J. Appl. Phys. 73 (1993) 8179. [16] E. Snoeks, G.N. van den Hoven, B. Hendriksen, M.B.J. Diemeer, F. Priolo, J. Opt. Soc. Am. B 12 (1995) 1468.

ARTICLE IN PRESS 276

N. Jaba et al. / Journal of Luminescence 129 (2009) 270–276

[17] T. Ohtsuki, N. Peyghambarian, S. Honkanen, S.I. Najafi, J. Opt. Soc. Am. B 14 (1997) 1838. [18] T. Ohtsuki, S. Honkanen, S.I. Najafi, N. Peyghambarian, J. Appl. Phys. 78 (1995) 3617. [19] Y.C. Yan, A.J. Faber, H. de Waal, P.G. Kik, A. Polman, Appl. Phys. Lett. 71 (1997) 2922. [20] Y. Yan, A.J. Faber, H. de Waal, J. Non-Cryst. Solids 181 (1995) 283. [21] E. Snoeks, P.G. Kit, A. Polman, Opt. Mater. 5 (1995) 91. [22] M. Arnaduov, Phys. Chem. Glasses 127 (1986) 48. [23] W.J. Chung, A. Jha, S. Shen, P. Joshi, Philos. Mag. 84 (2004) 1197. [24] G.A. Kumar, E. De la Rosa, H. Desirena, Opt. Commun. 260 (2006) 601. [25] R. El-Mallawany, A. Patra, C.S. Friend, R. Kapoor, P.N. Prasad, Opt. Mater. 12 (2004) 34. [26] H. Lin, S. Jiang, J. Wu, F. Song, N. Peyghambarian, E.Y.B. Pun, J. Phys. D: Appl. Phys. 36 (2003) 812. [27] H. Lin, G. Meredith, S. Jiang, X. Peng, T. Luo, N. Peyghambarian, E.Y.B. Pun, J. Appl. Phys. 93 (2003) 186. [28] S. Dai, J. Wu, J. Zhang, G. Wang, Z. Jiang, Spectrochim. Acta A 62 (2005) 431. [29] R. Rolli, M. Mongtagna, S. Chaussedent, A. Monteil, V.K. Tikhomirov, M. Ferrari, Opt. Mater. 21 (2003) 743. [30] S. Zhao, X. Wang, S. Xu, L. Hu, Chalcogen. Lett. 2 (2005) 97. [31] N.M. Souza Neto, A.Y. Ramos, L.C. Barbosa, J. Non-Cryst. Solids 304 (1995) 195. [32] S. Khatir, J. Bolka, B. Capoen, S. Turrell, M. Bouazaoui, J. Mol. Struct. 563–564 (2001) 283. [33] H. Desirena, E. De la Rosa, L.A. Diaz-Torres, G.A. Kumar, Opt. Mater. 28 (2006) 560. [34] X. Feng, S. Tanabe, T. Hanada, J. Am. Ceram. Soc. 84 (2001) 165. [35] A. Jha, S. Shen, M. Naftaly, Phys. Rev. B 62 (2000) 6215. [36] L.M. Fortes, L.F. Santos, et al., Opt. Mater. 29 (2007) 503. [37] P. Charton, P. Armand, J. Non-Cryst. Solids 333 (2004) 307. [38] L. Fortes, L.F. Santos, M.C. Gonc-alves, R.M. Almeida, J. Non-Cryst.Solids 352 (2006) 690. [39] C. Duverger, J.M. Nedelec, M. Benatsou, M. Bouazaoui, B. Capoen, M. Ferrari, S. Turrell, J. Mol. Struct. 480–481 (1999) 169. [40] N. Jaba, A. Mermet, E. Duval, B. Champagnon, J. Non-Cryst. Solids 351 (2005) 833. [41] O.K. Deutschbein, C.C. Pautrat, I.M. Svirchevsky, Rev. Phys. Appl. 2 (1967) 29.

[42] M.J. Weber, J.E. Lynch, D.H. Blackburn, D.J. Cronin, IEEE J. Quantum Electron. QE-19 (1983) 1600. [43] D.E. McCumber, Phys. Rev. 136 (1964) A954. [44] W.J. Miniscalco, R.S. Quimby, Opt. Lett. 16 (1991) 258. [45] P.R. Ehrmann, J.H. Campbell, J. Am. Ceram. Soc. 85 (2002) 1061. [46] S. Dai, C. Yu, G. Zhou, J. Zhang, G. Wang, L. Hu, J. Lumin. 117 (2006) 39. [47] T. Sekiya, N. Mochida, A. Soejima, J. Non-Cryst. Solids 191 (1995) 115. [48] C. Duverger, M. Bouazaoui, S. Turrell, J. Non-Cryst. Solids 220 (1997) 169. [49] S. Sakida, S. Hayakawa, T. Yoko, J. Non-Cryst. Solids 243 (1999) 13. [50] J.C. Sabadel, P. Armand, et al., J. Non-Cryst. Solids 244 (1999) 143. [51] S. Brovelli, A. Galli, R. Lorenzi, F. Meinardi, G. Spinolo, S. Tavazzi, V. Sigaev, S. Sukhov, P. Pernice, A. Aronne, E. Fanelli, A. Paleari, J. Non-Cryst. Solids 353 (2007) 2150. [52] F. Auzel, F. Bonfigli, S. Gagliari, G. Baldacchini, J. Lumin. 94 (2001) 293. [53] M.J. Weber, Phys. Rev. B 4 (1971) 2932. [54] B.R. Judd, Phys. Rev. B 127 (1962) 750. [55] G.S. Ofelt, J. Chem. Phys. 37 (1962) 511. [56] N. Jaba, A. Kanoun, H. Mejri, A. Selmi, S. Alaya, H. Maaref, J. Phys. Condens. Matter 12 (2000) 4523. [57] M.J. Weber, Phys. Rev. 157 (1967) 262. [58] J. Yang, S. Dai, N. Dai, L. Wei, L. Hu, Z. Jiang, J. Lumin. 106 (2004) 9. [59] S. Shen, A. Jha, Opt. Mater. 25 (2004) 321. [60] R. Reisfeld, Struct. Bonding 22 (1975) 123. [61] Y. Ding, S. Jiang, et al., Opt. Mater. 15 (2000) 123. [62] X. Feng, S. Tanabe, T. Hanada, J. Non-Cryst. Solids 281 (2001) 48. [63] S. Dai, C. Yu, G. Zhou, J. Zhang, G. Wang, L. Hu, J. Lumin. 117 (2006) 39. [64] S. Dai, J. Zhang, C. Yu, G. Zhou, G. Wang, L. Hu, Mater. Lett. 59 (2005) 2333. [65] A.J. Bruce, W.A. Reed, A.E. Neeves, L.R. Copeland, W.H. Grodkiiewicz, A. Lidgardet, Mater. Res. Soc. Symp. Proc. 244 (1992) 157. [66] T. Fo¨rster, Ann. Phys. (NY) 2 (1948) 55. [67] D.L. Dexter, J. Chem. Phys. 21 (1953) 836. [68] B. Moine, A. Brenier, C. Pedreni, IEEE J. Q-E 25 (1989) 88. [69] F. Auzel, in: B. Di Bartolo (Ed.), Radiationless Processes, Plenum, New York, 1979, p. 213. [70] S.N. Houde-Walter, P.M. Peters, J.F. Stebbins, Q. Zeng, J. Non-Cryst. Solids 286 (2001) 118.