Yb3 + codoped germanate glass

Yb3 + codoped germanate glass

Journal of Non-Crystalline Solids 376 (2013) 26–29 Contents lists available at SciVerse ScienceDirect Journal of Non-Crystalline Solids journal home...

839KB Sizes 2 Downloads 34 Views

Journal of Non-Crystalline Solids 376 (2013) 26–29

Contents lists available at SciVerse ScienceDirect

Journal of Non-Crystalline Solids journal homepage: www.elsevier.com/ locate/ jnoncrysol

Broadband conversion of ultraviolet to visible and near-infrared emission in Gd 3 +/Yb 3 + codoped germanate glass Lili Tao a, Bo Zhou a, b, Gongxun Bai a, Yonggang Wang a, Jianhua Hao a, Yuen H. Tsang a,⁎ a b

Department of Applied Physics and Materials Research Center, The Hong Kong Polytechnic University, Hung Hom, Kowloon, Hong Kong Department of Electrical Engineering, The Hong Kong Polytechnic University, Hung Hom, Kowloon, Hong Kong

a r t i c l e

i n f o

Article history: Received 31 January 2013 Received in revised form 10 April 2013 Available online 11 June 2013 Keywords: Germanate glass; Gd3 +/Yb3 +; Down-conversion; White emission

a b s t r a c t Gd3+/Yb 3+ co-doped germanate glasses were fabricated using the melt-quenching method. Broadband visible emission was observed from the gadolinium doped germanate glasses under ultraviolet (UV) excitation, and white-light was obtained by modifying Yb 3+ co-dopant concentration. Through this co-doped system, the white-light emission can be further converted into the near-infrared range which is matched well with the energy band gap of commercial silicon solar cell. The possible energy transfer processes involved are proposed and discussed according to the excitation, emission and fluorescence decay measurements. The effect of Yb 3+ ion concentration with respect to fixed Gd 3+ ions on the visible and near-infrared (NIR) emission has also been studied. The conversion efficiency from visible to near-infrared is as high as 69.3% in the GZN: 0.2Gd2O3–0.6Yb2O3 sample. The results indicate that this material is potential in various applications, such as white LED, enhancing silicon solar cell efficiency, and so on. © 2013 Elsevier B.V. All rights reserved.

1. Introduction Optical glasses are very useful for the new technological development in the field of optics and photonics due to its mature fabrication techniques, low optical transmission loss, easy-made in different size and shape, as well as high solubility of rare-earth (RE) ions. For example, germanate glass, one type of the attractive optical glasses, possesses low maximum vibrational frequencies in comparison to silicate, phosphate and borate glasses, besides its excellent chemical durability, mechanical properties and broad optical transmission range [1]. Because of its lower phonon energy, the quantum efficiency of luminescence from excited states of RE ions in this matrix is high, making it an efficient medium for laser and photonics devices. In addition, such material has a high damage threshold which is suitable for the use in higher power laser systems [2,3]. The low cost, commercially available silicon solar cells have the efficiency just around 15–25% with an estimated limitation around 29% [4]. The main reason is that most of the solar energy cannot be absorbed by the silicon solar cells and further converted into electrical energy due to the large mismatch between the solar spectrum and spectral response of silicon [5]. Researchers have tried to convert visible and infrared light into the ~ 1 μm range through photon down-conversion and up-conversion, respectively. However, upconversion process depends much on the excitation power density

and has little practical application under natural sunlight radiation, while the down-conversion may occur at low power excitation. Conversion of visible radiation into NIR has been extensively studied, but little research has been focused on the conversion of high energy UV to NIR. As been well known, trivalent Yb 3 + plays a significant role in producing NIR photons in ~ 0.9–1.1 μm spectral region [6,7] and serving as luminescence sensitizers in both up- and downconversion configurations [8–10] due to its unique energy level structure with no excited state absorption, broad output tunable wavelength range and low concentration quenching by interionic energy transfer. Gd 3 + co-doped with Yb 3 + could absorb photons in the UV region (250 nm to 425 nm) and emit infrared photons at ~ 1 μm where silicon solar cells exhibit their maximum spectral response [4]. Gd 3 +/Yb 3 + co-doped systems are thus potentially applicable in enhancing the efficiency of the silicon solar cells. In this study, in order to make use of the broad UV high-energy region of solar spectrum, we propose Gd 3 +/Yb 3 + co-doped germanate glass as a potential spectral modified material for enhancing silicon photovoltaic efficiency. To our best knowledge, it is the first time that the visible and NIR emission under UV excitation in Gd 3 +/Yb 3 + codoped glass are reported. The luminescent mechanism was investigated according to the excitation and emission spectra, lifetime measurements, and its potential application in silicon photovoltaic improvement was discussed. 2. Experimental

⁎ Corresponding author. E-mail address: [email protected] (Y.H. Tsang). 0022-3093/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jnoncrysol.2013.04.056

Glass samples with composition of 80GeO2–10ZnO–10Na2O– 0.2Gd2O3–xYb2O3 (x = 0, 0.2, 0.4, 0.6 mol.%) were prepared by

L. Tao et al. / Journal of Non-Crystalline Solids 376 (2013) 26–29

melt-quenching methods [11]. GeO2 (analytical reagent), ZnO (analytical reagent), Na2CO3 (analytical reagent), Gd2O3 (99.99%) and Yb2O3 (99.99%) were used as raw materials. Batches of ~ 10 g were mixed sufficiently and melted at 1200–1250 °C for 3 h in a platinum crucible. The melts were then thermally quenched by casting the melt into a copper mold preheated to 400 °C and then annealed at this temperature for 2 h before cooling down to room temperature at a rate of 0.3 °C/min. All the as-prepared glass samples were optically polished for measurements. The excitation spectrum, emission spectrum and the fluorescence decay curves in both the visible and NIR regions were recorded at room temperature by an FLS920p fluorescence spectrophotometer (Edinburgh Instrument Ltd., UK). 3. Results and discussion Fig. 1(a) shows the visible emission spectra of the glasses with an increasing Yb2O3 concentration from 0 to 0.6 mol.% at an excitation of 300 nm. Broad visible emission from 400 to 800 nm, covering blue, green and red range, is obtained from the glasses under UV (300 nm) excitation. The digital photo of sample D under excitation shown in bottom inset of Fig. 1(a) indicates that the luminescence is close to white color. It is well known that there are no transitions in Gd 3+ or Yb3+ which are capable of producing this broad luminescence, therefore, it is believed due to the defects in the host glass as reported in Refs. [12] and [13]. In order to verify it, the excitation and emission spectra of the germanate host glass without doping any RE ions were performed and given in the upper inset of Fig. 1(a). It can be found that, under excitation at 300 nm, the germanate host glass also gives a broad emission, confirming the presence of the defects and radiative transitions from them in the host glass (the sharp peak observed at around 600 nm is due to the second harmonics of the excited source). For comparison, the excitation spectrum of GZN: 0.2Gd2O3 is also plotted in the upper inset of Fig. 1. The red-shift of the peak wavelength as well as the much higher intensity (around 20 times) of the excitation spectrum of the Gd 3+-doped germanate host glass confirms the involvement of the Gd3+ in the luminescent properties. It is also noticed that there is a change in profile of the broad visible emission band, such as the blue-shift of the peak wavelength from ~600 to ~550 nm and disappearance of the shoulder at around 390 nm in the sample doped with Gd 3+. This might be due to the change of the defect energy levels resulted from the incorporation of Gd 3+ [12,13]. The color coordinates


CIE of the PL spectra are calculated to be (0.339, 0.380), (0.326, 0.377), (0.312, 0.361) and (0.310, 0.352) for the four glasses from A to D, respectively, and they are labeled in Fig. 1(b). It shows that the color of the emission was tuned closer to the standard white light emission (0.33, 0.33) by increasing the Yb3+ concentration, although the luminescence intensity has a decrease because of the energy transfer to neighboring Yb3+ via the down-conversion process. The NIR emission spectra of the glasses with different Yb2O3 concentration are shown in Fig. 2. The NIR emission of Yb 3+ is composed of two peaks: a sharp peak located at 980 nm and a broad peak centered at around 1020 nm. They are assigned to the transitions from the ground state 2F7/2 to two the upper 2F5/2 multiplets, respectively [14]. The NIR emission intensity shows an increase with Yb 3+ concentration increasing from 0 to 0.4 mol.%. However, a decrease in emission intensity is observed as Yb 3+ further rises up to 0.6 mol.%. Luminescence decay curves of Yb 3+ are displayed in the inset of Fig. 2. Lifetime values fitted by a single exponential function are equal to 1.0, 1.9 and 1.4 ms for the glasses with 0.2, 0.4 and 0.6 mol.% Yb 3+ dopings, respectively. This lifetime values show a similar tendency change to the intensity that increase first and then decrease as the Yb 3+ concentration increase, might be due to the concentration quenching effect resulted from cross relaxations between Yb 3+ ions, self-absorption and energy migration to quenching centers such as defects [15,16]. The optimal Yb 3+ concentration for the maximum NIR emission is around 0.4 mol.% Yb2O3 in the present work. The spectral response curve (black solid line in Fig. 3) of silicon solar cell shows that it has the best response to the photons with energy close to the bandgap of silicon solar cell while nearly has no response to the UV light [5]. This response characteristic causes an optic-electric conversion efficiency limit which was estimated to be 29% by Shockley and Queisser [4]. So, if the radiation of the sunlight in low response region was converted into the range with higher response, the optic-electric conversion efficiency of silicon can be improved and even break the theoretical efficiency limit. The excitation spectrum of GZN: 0.2Yb2O3–0.2Gd2O3 monitored at 1020 nm is also given in Fig. 3. A broad and strong UV absorption is observed which should be due to the abundant energy levels of Gd 3 + in the UV range. The transmittance of the germanate glass was also measured (not shown), and the result shows that it has a high transparency in the UV–vis-NIR range; even at 300 nm it has a transparency as high as 85% (0.5 mm thick) enabling UV light

Fig. 1. (a) Visible emission spectra of the glasses with different Yb2O3 concentration under 300 nm excitation. Upper inset shows excitation spectrum monitored at 550 nm (black line) and emission spectrum under 300 nm excitation (red line) of the germanate host glass, the purple line is excitation spectrum of GZN: 0.2Gd2O3 monitored at 550 nm. Bottom inset shows the digital photo of glass D under 300 nm excitation. (b) CIE chromaticity coordinates of the four glasses in the CIE 1931 chromaticity diagram.


L. Tao et al. / Journal of Non-Crystalline Solids 376 (2013) 26–29

Fig. 2. NIR emission spectra of the glasses with different Yb2O3 concentration under 300 nm excitation. The inset shows the corresponding luminescence decay curves of the Yb3+:2F5/2 → 2F7/2 transition.

absorbed by Gd 3 + as well as further converted into visible and NIR. The strong visible and NIR emission indicated that UV was efficiently converted into the wavelength regions which has better response to silicon solar cell. Here the conversion efficiency cannot be calculated according to the lifetime change at a certain wavelength like other reports [17–19], because the emission is so broad and it shows different lifetimes, as shown in Fig. 4 in which the decay curves of the broad visible emissions at different wavelengths (450, 500, 550, 600 and 650 nm) are plotted for samples with various Yb 3 + concentration. The luminescence decay curves show non-exponential characteristics, and a mean decay lifetime τm is calculated by τm = ∫ tI(t)dt/∫ I(t)dt [20]. The calculated lifetime values show a decline with the increase of Yb 3 + concentration, confirming the ET process from the defects of the host glass to Yb 3 + as shown in Fig. 4. Therefore, the energy transfer efficiency (ηET) can be calculated by the equation ηET = 1 − ∫ Ixdλ/∫ I0dλ, where Ix is the visible emission intensity of the samples with different Yb 3 + concentration, I0 is the visible emission intensity of the sample without Yb 3 +, and the ηET is calculated to be 11.0%, 25.1% and 69.3% for the glasses with 0.2%, 0.4% and 0.6 mol.% of Yb 3 +, respectively. Fig. 5 shows the schematic energy-level diagram of Gd 3+, Ge-related oxygen defects, and Yb 3+, possibly involved transitions

and energy transfer processes. Because of the abundant 4f energy levels in the UV range, Gd 3+ can be efficiently excited by a broad UV, and then this absorbed energy may be transferred to the defects, leading to a broad visible emission when it decays from the conduction band to valence band. When Yb 3+ is introduced and sited around the defect centers, energy transferred to Yb 3+ occurs with a rapid decline of the white light emission especially at the longer wavelength side. The tuning of the broad visible emission to standard white light also indicates its application for white LED. More importantly, intense NIR luminescence ranging from 950 to 1100 nm from Yb 3+ was observed under UV excitation, due to the radiative transition 2 F5/2 → 2F7/2. This result means that the UV energy has been converted into the NIR by employing the Gd 3+/Yb 3+ codoped germanate glass. Since such UV-to-NIR down conversion was realized via the germanate host through its defect-associated radiation, it is of significance to investigate the possibly involved down-conversion process. In general, only the white light photons with higher energy than 550 nm may have second-order energy transfer (ET) process in which a high-energy photon from the defect center excites two Yb 3+ ions simultaneously, and the photons with lower energy mainly via the first-order ET process, as schematically shown in Fig. 5 [21]. Therefore, it is believed that the ET from defects to Yb 3+ is a combination of first-order and second-order processes, and thus the quantum efficiency (QE) must be higher than ηET. These energy transfer processes were schematically illustrated in Fig. 5. It has been shown that the probability of the first-order ET is much higher than the second-order process [22], which agrees well with the larger lifetime decrease at the longer wavelength with increasing the Yb 3+ concentration, as shown in Fig. 4; this also realizes the tuning of the yellow-white to standard white emission despite of a decline in intensity. Due to such an efficient UV-to-NIR energy down conversion, deposition of a film of this high transparent Gd3+/Yb 3+ codoped germanate glass on the surface of silicon solar cell, as shown in Fig. 5 inset, by pulsed laser deposition, magnetic sputtering or some other methods can thus effectively convert UV of the sunlight into visible and NIR range with much better response and finally result in an improvement of efficiency, like the calculation result reported by Trupke et al. [23] that the theoretical efficiency of silicon solar cell can be improved from 29% to 36.6% with an ideal down-converter film on the front side of it.

4. Conclusions Intense white light emission was generated in Gd 3 + doped germanate glass under UV excitation. Through Yb 3 + co-doping, the white light was tuned and further converted into NIR luminescence, which matches with the bandgap of silicon solar cell. The optimal Yb 3 + ion concentration for the maximum NIR emission is found to be 0.4 mol.% when Gd 3 + ion concentration is fixed at 0.2 mol.%. The possible energy transfer processes involved have been proposed and discussed according to the excitation, emission and fluorescence decay measurements, and the conversion efficiency from visible light to near-infrared luminescence is as high as 69.3% for the GZN: 0.2Gd2O3–0.6Yb2O3 sample. This result is promising for further enhancement of UV-violet to visible and NIR conversion in some solar energy collection and photonic devices.


Fig. 3. Excitation spectrum monitored at 1020 nm, visible and NIR emission spectrum excited at 300 nm of GZN: 0.2Gd2O3–0.2Yb2O3, and spectral response of silicon.

This work is financially supported by the Research Grants Council of Hong Kong, China (Project Number: GRF 526511/PolyU B-Q26E) and The Hong Kong Polytechnic University (grants G-YJ20, G-YH91, A-PK72, and 1-ZV8N).

L. Tao et al. / Journal of Non-Crystalline Solids 376 (2013) 26–29


Fig. 4. Decay curves of the broad visible emissions at different wavelength under 300 nm excitation. Calculated lifetime τm is also shown in the figures; subscripts 1, 2, 3 and 4 correspond to the glasses with different Yb2O3 concentration from 0 to 0.6 mol.%.

Fig. 5. Schematic energy-level diagram of Gd3+, Yb3+, and defect center. Possible energy transfer processes are also indicated.

References [1] A. Iniguez P, J.D. Tapia-Takaki, R.P. Duarte-Zamorano, J.A. Duarte-Moller, E. Alvarez R, Opt. Mater. 30 (2008) 1796–1799. [2] J. Wu, S. Jiang, T. Luo, J. Geng, N. Peyghambarian, N.P. Barnes, IEEE Photon. Technol. Lett. 18 (2006) 334–336. [3] J. Wu, Z. Yao, J. Zong, S. Jiang, Opt. Lett. 32 (2007) 638–640. [4] W. Shockley, H.J. Queisser, J. Appl. Phys. 32 (1961) 510–519. [5] J. Zhou, Y. Teng, S. Ye, X. Liu, J. Qiu, Opt. Mater. 33 (2010) 153–158. [6] P. Yan, S. Yin, J. He, C. Fu, Y. Wang, M. Gong, IEEE Photon. Technol. Lett. 23 (2011) 697–699. [7] Y.H. Tsang, D.J. Binks, Appl. Phys. B 96 (2009) 11–17. [8] Y.H. Tsang, T.A. King, T. Thomas, C. Udell, M.C. Pierce, Opt. Commun. 215 (2003) 381–387. [9] G. Sobon, P. Kaczmarek, A. Antonczak, J. Sotor, K.M. Abramski, Opt. Express 19 (2011) 19104–19113. [10] B. Zhou, E.Y.B. Pun, H. Lin, D. Yang, L. Huang, J. Appl. Phys. 106 (2009) 103105.

[11] B. Richards, S. Shen, A. Jha, Y. Tsang, D. Binks, Opt. Express 15 (2007) 6546–6551. [12] Y. Zhuang, Y. Teng, J. Zhou, S. Ye, X. Liu, G. Lin, J. Ruan, J. Qiu, J. Opt, J. Opt. Soc. Am. B 26 (2009) 2185–2188. [13] G. Lin, B. Zhu, S. Zhou, H. Yang, J. Qiu, Opt. Express 15 (2007) 16980–16985. [14] D. Chen, Y. Wang, Y. Yu, E. Ma, Mater. Chem. Phys. 101 (2007) 464–469. [15] M. Ito, G. Boulon, A. Bensalah, Y. Guyot, C. Goutaudier, H. Sato, J. Opt, J. Opt. Soc. Am. B 24 (2007) 3023–3033. [16] J. Liao, Y. Lin, Y. Chen, Z. Luo, E. Ma, X. Gong, Q. Tan, Y. Huang, J. Opt, J. Opt. Soc. Am. B 23 (2006) 2572–2580. [17] H. Lin, D. Chen, Y. Yu, A. Yang, Y. Wang, Opt. Lett. 36 (2011) 876–878. [18] Q. Luo, X. Qiao, X. Fan, X. Zhang, Opt. Lett. 36 (2011) 2767–2769. [19] B. Zhou, H. Lin, B. Chen, E.Y.B. Pun, Opt. Express 19 (2011) 6514–6523. [20] B. Zhou, L. Tao, Y.H. Tsang, W. Jin, E.Y.B. Pun, Opt. Express 20 (2012) 3803–3813. [21] J. Ueda, S. Tanabe, J. Appl. Phys. 106 (2009) 043101. [22] X. Huang, S. Han, W. Huang, X. Liu, Chem. Soc. Rev. 42 (2013) 173–201. [23] T. Trupke, M. Green, P. Wurfel, J. Appl. Phys. 92 (2002) 1668–1674.