DOI: 10.1002/cphc.201500117
Articles
UV Emission of Gd3 + in the Presence of Cu2 + : Towards Luminescence Quenching through Quantum Cutting? Jos A. Jimnez*[a] The first investigation into the ultraviolet (UV) photoluminescence of gadolinium(III) in the presence of copper(II) is reported. A melt-quenched barium phosphate glass was used as a model matrix. The optical spectroscopy assessment shows that with increasing CuO concentration the Cu2 + absorption band grows steadily, whereas the UV emission from Gd3 + ions
is progressively quenched. The data, thus, suggests the existence of a Gd3 + !Cu2 + energy-transfer process ocurring through quantum cutting. A downconversion/cross-relaxation pathway proceeding through a virtual state in Gd3 + is proposed. These findings suggest gadolinium(III) could potentially be used in the optical sensing of copper(II).
1. Introduction Dielectric materials activated with Gd3 + ions are attractive as sources for ultraviolet (UV) emission in phototherapy lamps, which are useful for the treatment of skin diseases.[1–3] Hence, approaches to maximize the efficiency of UV-type B light (280– 320 nm) emission, which is suitable for such applications, are currently under investigation. For instance, co-doping with a sensitizer,[2, 4] variation of host band structure,[4] and control over the atmosphere during the synthesis of the materials[5] have been considered as important parameters. Materials systems containing gadolinium have also been investigated for the development of luminescent materials through quantumcutting processes.[6–9] This phenomenon refers to the production of two or more photons for every photon absorbed, and for a typical case, in which two photons are emitted, quantum cutting is also referred to as two-photon luminescence.[6, 7] In this way, quantum efficiencies of the materials that approach 200 % are realistically envisioned, for example through energytransfer (downconversion) processes,[6] an output highly desirable for applications in fluorescent lamps and displays. Divalent copper has been reported to quench the photoluminescence (PL) of Eu3 + , Tb3 + and Sm3 + rare-earth ions, because considerable spectral overlap exists between their visible (e.g. red) emissions and the absorption due to intraconfiguration d–d transitions of Cu2 + .[10–13] In fact, the underlying energy transfer has been investigated for the development of analytical methods to detect trace amounts of copper(II).[10–13] Nevertheless, the influence of Cu2 + on Gd3 + emission properties remains unexplored to the best of the author’s knowledge. Divalent copper ions dispersed in phosphate glass have been reported to contribute to host absorption deep in the UV region (e.g. around 230 nm), by virtue of O2 !Cu2 + charge-transfer transitions.[14] Electronic transitions leading to emission in triva-
lent gadolinium ions are also excited rather deep in the UV (e.g. 8S7/2 !6IJ transitions around 275 nm).[2, 3] Thus, an energy match between these two transitions could potentially lead to a Gd3 + !Cu2 + electronic excitation transfer, which could reduce emission from the 6P7/2 metastable state in Gd3 + . Furthermore, it may be postulated that if a two-photon quantumcutting process[6, 15] takes place from the 6P7/2 state, which emits at around 310 nm, energies that are relevant to Cu2 + absorption in the visible region (d–d transitions) could be reached, and the resonance condition could be fulfilled. If such a Gd3 + !Cu2 + energy transfer occurs through quantum cutting, this should be reflected by the intensity of the Gd3 + (6P7/2) emission being dependent on the Cu2 + concentration. Hence, following this line of thought, the present work investigates, for the first time, Gd3 + !Cu2 + interactions in a phosphate glass host, as a model matrix for copper and rare-earth codoping.[16–18]
2. Results Figure 1 shows the absorption profiles obtained for Gd and GdCu0.2–0.6 glasses (where GdCu0.2, GdCu0.4 and GdCu0.6 refer to glasses prepared with 2.0 mol % Gd2O3 and 0.2, 0.4, and 0.6 mol %, respectively, of CuO; Table 1). The presence of copper(II) oxide in the GdCu0.2–0.6 glasses results in the development of a broad absorption band at around 850 nm, which is ascribed to 2E!2T2 intraconfigurational (d–d) transitions in Cu2 + .[19, 20] Notably, the broad Cu2 + absorption band grows steadily with increasing CuO concentration. The concentration dependence of the intensity of this band is shown in the plot presented in the inset of Figure 1. A linear relationship is manifested (correlation coefficient r = 0.992), as expected for these relatively low concentrations of CuO if Beer’s Law is satisfied.[13, 20] In contrast, UV absorption edges only show slight variations, as reported for CuO-containing Sm3 + -doped aluminophosphate glasses.[13] To assess the effect of the addition of CuO on UV absorption edges, in relation to likely O2 !Cu2 +
[a] Dr. J. A. Jimnez Department of Chemistry University of North Florida Jacksonville, FL 32224 (USA) E-mail: [email protected]
ChemPhysChem 0000, 00, 0 – 0
These are not the final page numbers! ÞÞ
1
0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
Articles
Figure 2. Plot of (Ea)1/2 versus hn for Gd and GdCu0.2–0.6 glasses. The solid lines are the linear fits to the linear regions of the absorption edges, from which the optical band gap values (Eopt) were determined (intercepts in energy axis). Inset: obtained Eopt values (with error bars) versus CuO concentration in the Gd (0 mol % CuO) and GdCu0.2–0.6 glasses.
Figure 1. Optical absorption spectra of the Gd and GdCu0.2–0.6 glasses. Inset: plot of peak absorption (850 nm) versus CuO concentration in the Gd (0 mol % CuO) and GdCu0.2–0.6 glasses; the solid line is the linear fit to the data.
Table 1. Matrix composition and additive concentrations (mol %) of Gd2O3 and CuO in the studied glasses. Glass
P2 O 5
BaO
Gd2O3
CuO
Gd GdCu0.2 GdCu0.4 GdCu0.6
50 50 50 50
50 50 50 50
2.0 2.0 2.0 2.0
– 0.2 0.4 0.6
charge-transfer transitions,[14] the optical band gap energies were analyzed. In this context, the power law expression for the absorption coefficient a(n) as a function of photon energy (E = hn) is employed [Eq. (1)]: aðvÞ ¼ k
ðhv Eopt Þn hv
Figure 3. UV emission spectra of the Gd and GdCu0.2–0.6 glasses collected under excitation at 275 nm. The dotted trace is the spectrum of a reference glass containing 0.4 mol % CuO without Gd3 + . Inset: plot of I0/I versus CuO concentration in the Gd (0 mol %) and GdCu0.2–0.6 glasses; I0 and I are the peak emission intensities for the 6P7/2 !8S7/2 transition in Gd3 + in the absence and presence of the copper(II) quencher, respectively. The solid line is the linear fit to the data.
ð1Þ
where the exponent n is related to the nature of the transitions, k is a constant, and Eopt is the optical band gap energy.[21] For glass materials, the value n = 2, connected to indirect transitions, is usually employed in fitting experimental results.[22, 23] Hence, Eopt can be estimated from a plot of (Ea)1/2 versus photon energy (hn) by extrapolating the linear portion of the plot and determining the intercept on the energy axis. Accordingly, such a plot was constructed from the data in Figure 1 and is shown in Figure 2. Linear fits were performed on the linear regions of the absorption edges, and the optical band gap values were determined from the x-axis intercepts of the lines generated by the regression analyses.[23] A plot of the corresponding Eopt values is shown in the inset of Figure 2, together with the associated errors in the determinations. Although there was some variation in the Eopt values for the different glasses, these were within experimental error and no definite trend was observed with respect to the concentration of CuO.
served towards the visible region for the GdCu0.2–0.6 glasses. This is likely due to luminescence from traces of monovalent copper (e.g. with cubic coordination).[14, 24] The intensity of this PL band is similar for the GdCu0.2, GdCu0.4 and GdCu0.6 glasses, thus suggesting they contain comparable residual amounts of luminescent copper (Cu + ions). Also presented in Figure 3, is a spectrum recorded under the same conditions for a glass containing the intermediate CuO concentration of 0.4 mol %, but without any Gd2O3. No significant emission is observed in the visible region of the spectrum. It is similar to that of the Gd glass, which is consistent with the lack of Cu + ions therein. Thus, it seems that the presence of Gd3 + promotes, to some extent, the manifestation of luminescent copper. A similar result was recently reported by Tong et al.[25] for SnO2-doped phosphate glasses, where the presence of Gd2O3 was observed to promote the presence of Sn2 + over Sn4 + , as compared to a glass containing no Gd2O3. The reason
Figure 3 shows the UV emission of Gd3 + in the various glasses, as assessed under excitation at 275 nm (8S7/2 !6IJ transitions).[2, 3] A relatively weak and broad emission band is ob-
&
ChemPhysChem 0000, 00, 0 – 0
www.chemphyschem.org
2
0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!
Articles for gadolinium doping promoting lower oxidation states in transition-metal ions is not well understood at present. However, although the data in Figure 3 suggests that some Cu + is present in the GdCu0.2–0.6 glasses, the comparable intensities of the PL bands indicate that there is an equivalent amount present in each of the glasses. This is in agreement with the fact that Beer’s law is still satisfied with respect to Cu2 + absorption, as seen from the inset of Figure 1. Nonetheless, there is no correlation or trend with respect to trace amounts of emitting Cu + ions in the glasses, which contain a fixed concentration of Gd2O3 and increasing amounts of CuO. In contrast, the UV emission around 315 nm corresponding to 6P7/2 !8S7/2 transitions in Gd3 + decreases with increasing concentration of CuO clearly indicating Cu2 + -induced PL quenching. Typically, the dependence of luminescence quenching through energy transfer on the concentration of the quencher is evaluated according to a linear dependence as [Eq. (2)]: I0 ¼ 1 þ QC ½A I
a medial energy state is not part of the electronic energy levels of Gd3 + ions, which have a 4f7 configuration.[26] However, similar situations have been encountered in the context of quantum cutting, in which a virtual state is located on the donor ions (e.g. Tb3 + ), the presence of which is invoked to account for the operating mechanism.[15] Furthermore, the existence of a medial energy level that is not part of the electronic structure of the rare-earth elements is accounted for by downconversion processes (e.g. Tm3 + ).[27] The origin of such a state is possibly related to defects in the glass matrix, as considered by Shen et al.[27] Accordingly, it is, herein, postulated that the Gd3 + !Cu2 + energy transfer may occur through a downconversion process involving a virtual state in Gd3 + , by which the resonance condition for Cu2 + absorption is fulfilled. The proposed energy-transfer pathway is represented schematically in Figure 4. Steps 1–3 illustrate the common process by which
ð2Þ
where I0 and I are the emission intensities in the absence and presence of the quencher, respectively, [A] is the concentration of the quencher (energy acceptor), and QC is the quenching constant.[11, 13] The inset of Figure 3 shows such a plot constructed with the PL peak intensities of the 6P7/2 !8S7/2 transition versus CuO concentration in the Gd and GdCu0.2–0.6 glasses. A linear relationship is seen; the regression analysis performed yielded a correlation coefficient r = 0.994. The value for the quenching constant extracted from the fit is 6.56 ( 0.52) mol %1. This value is actually larger than that estimated for the resonant transfer between Sm3 + ions and Cu2 + in aluminophosphate glasses (2.59 mol %1).[13] Evidently, the data points towards an effective Gd3 + !Cu2 + energy-transfer process that results in depopulating the 6P7/2 metastable state in Gd3 + .
3. Discussion Reviewing the results obtained, the notable outcome of incorporating Cu2 + ions into a Gd3 + -doped glass matrix is not a change in the UV optical band gap energies, but the consistent growth of the visible absorption band at around 575– 900 nm with increasing concentration of CuO. Furthermore, the quenching of the UV emissions owing to 6P7/2 !8S7/2 transitions in Gd3 + clearly correlated with the CuO concentration. Herein, no trend was demonstrated between such PL quenching and residual Cu + emission or UV absorption analysis aimed at assessing O2 !Cu2 + charge-transfer transitions. In addition, a phonon-assisted energy transfer is unlikely given the large energy difference between the 6P7/2 emitting state and 2E!2T2 transitions in Cu2 + . In contrast, the phenomenon of quantum cutting has been well-documented for gadolinium-containing materials.[6–9] In the present case, it can be noticed that the energy necessary to achieve the resonance condition for Cu2 + absorption is reached by Gd3 + ions if a two-photon quantumcutting process occurs from the 6P7/2 emitting state. Such ChemPhysChem 0000, 00, 0 – 0
www.chemphyschem.org
These are not the final page numbers! ÞÞ
Figure 4. Schematic representation of the proposed downconversion pathway for the energy transfer from Gd3 + to Cu2 + resulting in the observed photoluminescence quenching. The solid-straight and solid-curved arrows indicate radiative and nonradiative transitions, respectively. The dashed arrows depict the quantum cutting involving a virtual state in Gd3 + and consequent cross-relaxation energy transfer to Cu2 + .
the emission around 315 nm is obtained from the 6P7/2 state in Gd3 + . The novel Cu2 + -induced PL quenching is represented by steps 4–6, where steps 4 and 5 involve the quantum-cutting process through the virtual state, which allows the resonance condition to be fulfilled. Consequently, the energy is transferred by cross-relaxation to Cu2 + ions, which decay nonradiatively in step 6. This would tentatively explain the obtained results regarding the role of Cu2 + ions as quenchers of the UV emission of Gd3 + .
3
0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
Articles 4. Conclusions
tonics K. K.) and a photomultiplier tube (R1527P, Hamamatsu Photonics K. K.). The flash lamp was kept operating at a frequency of 125 Hz with the total period of data collection set to 8 ms. The step size used for all spectral acquisitions was 1 nm. All PL measurements were recorded with samples mounted in a solid sample holder at an angle of 408 with particular attention given to keep conditions constant during experiments. All measurements were carried out at room temperature.
An optical spectroscopy study of melt-quenched barium phosphate glasses containing Gd2O3 and varying amounts of CuO was carried out to evaluate the effect of Cu2 + on UV emission from Gd3 + ions. The estimated UV optical band gap energies were similar (within experimental error) for glasses containing no CuO and those containing varying amounts of CuO. The fixed concentration of 2 mol % Gd2O3 used here, seemed to promote the production of residual Cu + , but at equivalent levels in all the different samples. However, the Cu2 + visible absorption peak around 850 nm increased steadily with increasing CuO concentration. Furthermore, an increase in the concentration of CuO resulted in significant quenching of Gd3 + PL from the 6P7/2 UV-emitting state. These results indicate that the PL quenching is caused by a Gd3 + !Cu2 + energy transfer. It was proposed that a cross-relaxation energy transfer may occur through a quantum-cutting process that proceeds via a virtual state in Gd3 + , by which the resonance condition for Cu2 + absorption is met. Thus, the Gd3 + –Cu2 + couple appears promising as a system for further fundamental studies on quantum-cutting and downconversion processes that, for instance, involve a third luminescent center (e.g. another rareearth element). This work also indicates that Cu2 + impurities could limit the efficiency of Gd3 + -activated glasses for phototherapy lamp applications. Moreover, the findings suggest the potential of gadolinium(III) for optical sensing of copper(II), which could be explored for analytical applications.
Acknowledgements The author thanks undergraduate student Martin Garnero from the Chemistry Department at UNF for assistance with the experiments. Keywords: copper · gadolinium · glasses · luminescence · rare earths [1] D. D. Ramteke, R. S. Gedam, J. Rare Earths 2014, 32, 389 – 393. [2] P. P. Mokoena, I. M. Nagpure, V. Kumar, R. E. Kroon, E. J. Olivier, J. H. Neethling, H. C. Swart, O. M. Ntwaeaborwa, J. Phys. Chem. Solids 2014, 75, 998 – 1003. [3] V. Singh, G. Sivaramaiah, J. L. Rao, S. H. Kim, J. Electron. Mater. 2014, 43, 3486 – 3492. [4] Y. Shimizu, Y. Takano, K. Ueda, Thin Solid Films 2014, 559, 23 – 26. [5] C. Tang, S. Liu, L. Liu, D. P. Chen, J. Lumin. 2015, 160, 317 – 320. [6] R. T. Wegh, H. Donker, K. D. Oskam, A. Meijerink, Science 1999, 283, 663 – 666. [7] N. Kodama, S. Oishi, J. Appl. Phys. 2005, 98, 103515. [8] D. Wang, N. Kodama, J. Solid State Chem. 2009, 182, 2219 – 2224. [9] L. Han, Y. Wang, J. Zhang, Y. Tao, Mater. Chem. Phys. 2014, 143, 476 – 479. [10] M. A. Kessler, Anal. Chim. Acta 1998, 364, 125 – 129. [11] C. Cano-Raya, M. D. Fernndez Ramos, L. F. Capitn Vallvey, O. S. Wolfbeis, M. Schferling, Appl. Spectrosc. 2005, 59, 1209 – 1216. [12] M. Turel, A. Duerkop, A. Yegorova, Y. Scripinets, A. Lobnik, N. Samec, Anal. Chim. Acta 2009, 644, 53 – 60. [13] J. A. Jimnez, J. Lumin. 2015, 161, 352 – 357. [14] K. Tanaka, T. Yano, S. Shibata, M. Yamane, S. Inoue, J. Non-Cryst. Solids 1994, 178, 9 – 14. [15] P. Vergeer, T. J. H. Vlugt, M. H. F. Kox, M. I. den Hertog, J. P. J. M. van der Eerden, A. Meijerink, Phys. Rev. B 2005, 71, 014119. [16] J. A. Jimnez, J. B. Hockenbury, J. Mater. Sci. 2013, 48, 6921 – 6928. [17] J. A. Jimnez, J. Phys. Chem. Solids 2014, 75, 1334 – 1339. [18] J. A. Jimnez, M. Sendova, J. Appl. Phys. 2014, 116, 033518. [19] T. Murata, K. Morinaga, Proc. SPIE 2000, 4102, 316 – 323. [20] J. A. Jimnez, J. Mater. Sci. 2014, 49, 4387 – 4393. [21] J. Tauc, A. Menth, J. Non-Cryst. Solids 1972, 8 – 10, 569 – 585. [22] G. Venkateswara Rao, H. D. Shashikala, J. Adv. Ceram. 2014, 3, 109 – 116. [23] F. S. De Vicente, F. A. Santos, B. S. Simes, S. T. Dias, M. Siu Li, Opt. Mater. 2014, 38, 119 – 125. [24] R. Debnath, S. K. Das, Chem. Phys. Lett. 1989, 155, 52 – 58. [25] Y. Tong, Z. Yan, H. Zeng, G. Chen, J. Lumin. 2014, 145, 438 – 442. [26] G. H. Dieke, H. M. Crosswhite, Appl. Opt. 1963, 2, 675 – 686. [27] C. Shen, S. Baccaro, Z. Xing, Q. Yan, S. Wang, G. Chen, Chem. Phys. Lett. 2010, 492, 123 – 126.
Experimental Section Barium-phosphate glasses with a 50P2O5 :50BaO (mol %) composition were prepared from high purity compounds (P2O5 and BaCO3) by the melt-quenching technique.[16] Batch materials were thoroughly mixed and melted in porcelain crucibles at 1150 8C for 15 min under normal atmospheric conditions and immediately quenched. Gadolinium- and/or copper-doping was achieved by adding Gd2O3 and/or CuO quantities in mol %, in relation to network former P2O5. The gadolinium concentration was kept constant at 2.0 mol % in all Gd-doped samples. A glass containing only gadolinium was prepared as a reference, and is referred to as Gd glass. A set of glasses were prepared having a fixed Gd2O3 concentration (2.0 mol %) with different amounts of CuO (0.2, 0.4 and 0.6 mol %) and are referred to as GdCu0.2, GdCu0.4 and GdCu0.6, respectively. The nominal compositions of the glasses are summarized in Table 1. All glasses were cut and polished in order to produce glass slabs with final thicknesses of about 1.0 mm. Optical absorption measurements were performed using a PerkinElmer 35 UV/Vis double-beam spectrophotometer. All absorption spectra were recorded with air as reference. Emission spectra were obtained with a Photon Technology International QuantaMaster 30 spectrofluorometer equipped with a Xe flash lamp having a pulse width of about 2 ms (L4633, Hamamatsu Pho-
&
ChemPhysChem 0000, 00, 0 – 0
www.chemphyschem.org
Received: March 4, 2015 Published online on && &&, 2015
4
0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!
ARTICLES A quantum cut above: Photoluminescence quenching of Gd3 + ions in the presence of Cu2 + impurities in glass is demonstrated. The phenomenon of quantum cutting is proposed to be involved in the energy transfer between Gd3 + and Cu2 + ions. The findings suggest the potential of gadolinium(III) for optical sensing of copper(II), which could be explored for analytical applications.
ChemPhysChem 0000, 00, 0 – 0
www.chemphyschem.org
These are not the final page numbers! ÞÞ
J. A. Jimnez* && – && UV Emission of Gd3 + in the Presence of Cu2 + : Towards Luminescence Quenching through Quantum Cutting?
5
0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
Articles
UV Emission of Gd3 + in the Presence of Cu2 + : Towards Luminescence Quenching through Quantum Cutting? Jos A. Jimnez*[a] The first investigation into the ultraviolet (UV) photoluminescence of gadolinium(III) in the presence of copper(II) is reported. A melt-quenched barium phosphate glass was used as a model matrix. The optical spectroscopy assessment shows that with increasing CuO concentration the Cu2 + absorption band grows steadily, whereas the UV emission from Gd3 + ions
is progressively quenched. The data, thus, suggests the existence of a Gd3 + !Cu2 + energy-transfer process ocurring through quantum cutting. A downconversion/cross-relaxation pathway proceeding through a virtual state in Gd3 + is proposed. These findings suggest gadolinium(III) could potentially be used in the optical sensing of copper(II).
1. Introduction Dielectric materials activated with Gd3 + ions are attractive as sources for ultraviolet (UV) emission in phototherapy lamps, which are useful for the treatment of skin diseases.[1–3] Hence, approaches to maximize the efficiency of UV-type B light (280– 320 nm) emission, which is suitable for such applications, are currently under investigation. For instance, co-doping with a sensitizer,[2, 4] variation of host band structure,[4] and control over the atmosphere during the synthesis of the materials[5] have been considered as important parameters. Materials systems containing gadolinium have also been investigated for the development of luminescent materials through quantumcutting processes.[6–9] This phenomenon refers to the production of two or more photons for every photon absorbed, and for a typical case, in which two photons are emitted, quantum cutting is also referred to as two-photon luminescence.[6, 7] In this way, quantum efficiencies of the materials that approach 200 % are realistically envisioned, for example through energytransfer (downconversion) processes,[6] an output highly desirable for applications in fluorescent lamps and displays. Divalent copper has been reported to quench the photoluminescence (PL) of Eu3 + , Tb3 + and Sm3 + rare-earth ions, because considerable spectral overlap exists between their visible (e.g. red) emissions and the absorption due to intraconfiguration d–d transitions of Cu2 + .[10–13] In fact, the underlying energy transfer has been investigated for the development of analytical methods to detect trace amounts of copper(II).[10–13] Nevertheless, the influence of Cu2 + on Gd3 + emission properties remains unexplored to the best of the author’s knowledge. Divalent copper ions dispersed in phosphate glass have been reported to contribute to host absorption deep in the UV region (e.g. around 230 nm), by virtue of O2 !Cu2 + charge-transfer transitions.[14] Electronic transitions leading to emission in triva-
lent gadolinium ions are also excited rather deep in the UV (e.g. 8S7/2 !6IJ transitions around 275 nm).[2, 3] Thus, an energy match between these two transitions could potentially lead to a Gd3 + !Cu2 + electronic excitation transfer, which could reduce emission from the 6P7/2 metastable state in Gd3 + . Furthermore, it may be postulated that if a two-photon quantumcutting process[6, 15] takes place from the 6P7/2 state, which emits at around 310 nm, energies that are relevant to Cu2 + absorption in the visible region (d–d transitions) could be reached, and the resonance condition could be fulfilled. If such a Gd3 + !Cu2 + energy transfer occurs through quantum cutting, this should be reflected by the intensity of the Gd3 + (6P7/2) emission being dependent on the Cu2 + concentration. Hence, following this line of thought, the present work investigates, for the first time, Gd3 + !Cu2 + interactions in a phosphate glass host, as a model matrix for copper and rare-earth codoping.[16–18]
2. Results Figure 1 shows the absorption profiles obtained for Gd and GdCu0.2–0.6 glasses (where GdCu0.2, GdCu0.4 and GdCu0.6 refer to glasses prepared with 2.0 mol % Gd2O3 and 0.2, 0.4, and 0.6 mol %, respectively, of CuO; Table 1). The presence of copper(II) oxide in the GdCu0.2–0.6 glasses results in the development of a broad absorption band at around 850 nm, which is ascribed to 2E!2T2 intraconfigurational (d–d) transitions in Cu2 + .[19, 20] Notably, the broad Cu2 + absorption band grows steadily with increasing CuO concentration. The concentration dependence of the intensity of this band is shown in the plot presented in the inset of Figure 1. A linear relationship is manifested (correlation coefficient r = 0.992), as expected for these relatively low concentrations of CuO if Beer’s Law is satisfied.[13, 20] In contrast, UV absorption edges only show slight variations, as reported for CuO-containing Sm3 + -doped aluminophosphate glasses.[13] To assess the effect of the addition of CuO on UV absorption edges, in relation to likely O2 !Cu2 +
[a] Dr. J. A. Jimnez Department of Chemistry University of North Florida Jacksonville, FL 32224 (USA) E-mail: [email protected]
ChemPhysChem 0000, 00, 0 – 0
These are not the final page numbers! ÞÞ
1
0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
Articles
Figure 2. Plot of (Ea)1/2 versus hn for Gd and GdCu0.2–0.6 glasses. The solid lines are the linear fits to the linear regions of the absorption edges, from which the optical band gap values (Eopt) were determined (intercepts in energy axis). Inset: obtained Eopt values (with error bars) versus CuO concentration in the Gd (0 mol % CuO) and GdCu0.2–0.6 glasses.
Figure 1. Optical absorption spectra of the Gd and GdCu0.2–0.6 glasses. Inset: plot of peak absorption (850 nm) versus CuO concentration in the Gd (0 mol % CuO) and GdCu0.2–0.6 glasses; the solid line is the linear fit to the data.
Table 1. Matrix composition and additive concentrations (mol %) of Gd2O3 and CuO in the studied glasses. Glass
P2 O 5
BaO
Gd2O3
CuO
Gd GdCu0.2 GdCu0.4 GdCu0.6
50 50 50 50
50 50 50 50
2.0 2.0 2.0 2.0
– 0.2 0.4 0.6
charge-transfer transitions,[14] the optical band gap energies were analyzed. In this context, the power law expression for the absorption coefficient a(n) as a function of photon energy (E = hn) is employed [Eq. (1)]: aðvÞ ¼ k
ðhv Eopt Þn hv
Figure 3. UV emission spectra of the Gd and GdCu0.2–0.6 glasses collected under excitation at 275 nm. The dotted trace is the spectrum of a reference glass containing 0.4 mol % CuO without Gd3 + . Inset: plot of I0/I versus CuO concentration in the Gd (0 mol %) and GdCu0.2–0.6 glasses; I0 and I are the peak emission intensities for the 6P7/2 !8S7/2 transition in Gd3 + in the absence and presence of the copper(II) quencher, respectively. The solid line is the linear fit to the data.
ð1Þ
where the exponent n is related to the nature of the transitions, k is a constant, and Eopt is the optical band gap energy.[21] For glass materials, the value n = 2, connected to indirect transitions, is usually employed in fitting experimental results.[22, 23] Hence, Eopt can be estimated from a plot of (Ea)1/2 versus photon energy (hn) by extrapolating the linear portion of the plot and determining the intercept on the energy axis. Accordingly, such a plot was constructed from the data in Figure 1 and is shown in Figure 2. Linear fits were performed on the linear regions of the absorption edges, and the optical band gap values were determined from the x-axis intercepts of the lines generated by the regression analyses.[23] A plot of the corresponding Eopt values is shown in the inset of Figure 2, together with the associated errors in the determinations. Although there was some variation in the Eopt values for the different glasses, these were within experimental error and no definite trend was observed with respect to the concentration of CuO.
served towards the visible region for the GdCu0.2–0.6 glasses. This is likely due to luminescence from traces of monovalent copper (e.g. with cubic coordination).[14, 24] The intensity of this PL band is similar for the GdCu0.2, GdCu0.4 and GdCu0.6 glasses, thus suggesting they contain comparable residual amounts of luminescent copper (Cu + ions). Also presented in Figure 3, is a spectrum recorded under the same conditions for a glass containing the intermediate CuO concentration of 0.4 mol %, but without any Gd2O3. No significant emission is observed in the visible region of the spectrum. It is similar to that of the Gd glass, which is consistent with the lack of Cu + ions therein. Thus, it seems that the presence of Gd3 + promotes, to some extent, the manifestation of luminescent copper. A similar result was recently reported by Tong et al.[25] for SnO2-doped phosphate glasses, where the presence of Gd2O3 was observed to promote the presence of Sn2 + over Sn4 + , as compared to a glass containing no Gd2O3. The reason
Figure 3 shows the UV emission of Gd3 + in the various glasses, as assessed under excitation at 275 nm (8S7/2 !6IJ transitions).[2, 3] A relatively weak and broad emission band is ob-
&
ChemPhysChem 0000, 00, 0 – 0
www.chemphyschem.org
2
0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!
Articles for gadolinium doping promoting lower oxidation states in transition-metal ions is not well understood at present. However, although the data in Figure 3 suggests that some Cu + is present in the GdCu0.2–0.6 glasses, the comparable intensities of the PL bands indicate that there is an equivalent amount present in each of the glasses. This is in agreement with the fact that Beer’s law is still satisfied with respect to Cu2 + absorption, as seen from the inset of Figure 1. Nonetheless, there is no correlation or trend with respect to trace amounts of emitting Cu + ions in the glasses, which contain a fixed concentration of Gd2O3 and increasing amounts of CuO. In contrast, the UV emission around 315 nm corresponding to 6P7/2 !8S7/2 transitions in Gd3 + decreases with increasing concentration of CuO clearly indicating Cu2 + -induced PL quenching. Typically, the dependence of luminescence quenching through energy transfer on the concentration of the quencher is evaluated according to a linear dependence as [Eq. (2)]: I0 ¼ 1 þ QC ½A I
a medial energy state is not part of the electronic energy levels of Gd3 + ions, which have a 4f7 configuration.[26] However, similar situations have been encountered in the context of quantum cutting, in which a virtual state is located on the donor ions (e.g. Tb3 + ), the presence of which is invoked to account for the operating mechanism.[15] Furthermore, the existence of a medial energy level that is not part of the electronic structure of the rare-earth elements is accounted for by downconversion processes (e.g. Tm3 + ).[27] The origin of such a state is possibly related to defects in the glass matrix, as considered by Shen et al.[27] Accordingly, it is, herein, postulated that the Gd3 + !Cu2 + energy transfer may occur through a downconversion process involving a virtual state in Gd3 + , by which the resonance condition for Cu2 + absorption is fulfilled. The proposed energy-transfer pathway is represented schematically in Figure 4. Steps 1–3 illustrate the common process by which
ð2Þ
where I0 and I are the emission intensities in the absence and presence of the quencher, respectively, [A] is the concentration of the quencher (energy acceptor), and QC is the quenching constant.[11, 13] The inset of Figure 3 shows such a plot constructed with the PL peak intensities of the 6P7/2 !8S7/2 transition versus CuO concentration in the Gd and GdCu0.2–0.6 glasses. A linear relationship is seen; the regression analysis performed yielded a correlation coefficient r = 0.994. The value for the quenching constant extracted from the fit is 6.56 ( 0.52) mol %1. This value is actually larger than that estimated for the resonant transfer between Sm3 + ions and Cu2 + in aluminophosphate glasses (2.59 mol %1).[13] Evidently, the data points towards an effective Gd3 + !Cu2 + energy-transfer process that results in depopulating the 6P7/2 metastable state in Gd3 + .
3. Discussion Reviewing the results obtained, the notable outcome of incorporating Cu2 + ions into a Gd3 + -doped glass matrix is not a change in the UV optical band gap energies, but the consistent growth of the visible absorption band at around 575– 900 nm with increasing concentration of CuO. Furthermore, the quenching of the UV emissions owing to 6P7/2 !8S7/2 transitions in Gd3 + clearly correlated with the CuO concentration. Herein, no trend was demonstrated between such PL quenching and residual Cu + emission or UV absorption analysis aimed at assessing O2 !Cu2 + charge-transfer transitions. In addition, a phonon-assisted energy transfer is unlikely given the large energy difference between the 6P7/2 emitting state and 2E!2T2 transitions in Cu2 + . In contrast, the phenomenon of quantum cutting has been well-documented for gadolinium-containing materials.[6–9] In the present case, it can be noticed that the energy necessary to achieve the resonance condition for Cu2 + absorption is reached by Gd3 + ions if a two-photon quantumcutting process occurs from the 6P7/2 emitting state. Such ChemPhysChem 0000, 00, 0 – 0
www.chemphyschem.org
These are not the final page numbers! ÞÞ
Figure 4. Schematic representation of the proposed downconversion pathway for the energy transfer from Gd3 + to Cu2 + resulting in the observed photoluminescence quenching. The solid-straight and solid-curved arrows indicate radiative and nonradiative transitions, respectively. The dashed arrows depict the quantum cutting involving a virtual state in Gd3 + and consequent cross-relaxation energy transfer to Cu2 + .
the emission around 315 nm is obtained from the 6P7/2 state in Gd3 + . The novel Cu2 + -induced PL quenching is represented by steps 4–6, where steps 4 and 5 involve the quantum-cutting process through the virtual state, which allows the resonance condition to be fulfilled. Consequently, the energy is transferred by cross-relaxation to Cu2 + ions, which decay nonradiatively in step 6. This would tentatively explain the obtained results regarding the role of Cu2 + ions as quenchers of the UV emission of Gd3 + .
3
0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
Articles 4. Conclusions
tonics K. K.) and a photomultiplier tube (R1527P, Hamamatsu Photonics K. K.). The flash lamp was kept operating at a frequency of 125 Hz with the total period of data collection set to 8 ms. The step size used for all spectral acquisitions was 1 nm. All PL measurements were recorded with samples mounted in a solid sample holder at an angle of 408 with particular attention given to keep conditions constant during experiments. All measurements were carried out at room temperature.
An optical spectroscopy study of melt-quenched barium phosphate glasses containing Gd2O3 and varying amounts of CuO was carried out to evaluate the effect of Cu2 + on UV emission from Gd3 + ions. The estimated UV optical band gap energies were similar (within experimental error) for glasses containing no CuO and those containing varying amounts of CuO. The fixed concentration of 2 mol % Gd2O3 used here, seemed to promote the production of residual Cu + , but at equivalent levels in all the different samples. However, the Cu2 + visible absorption peak around 850 nm increased steadily with increasing CuO concentration. Furthermore, an increase in the concentration of CuO resulted in significant quenching of Gd3 + PL from the 6P7/2 UV-emitting state. These results indicate that the PL quenching is caused by a Gd3 + !Cu2 + energy transfer. It was proposed that a cross-relaxation energy transfer may occur through a quantum-cutting process that proceeds via a virtual state in Gd3 + , by which the resonance condition for Cu2 + absorption is met. Thus, the Gd3 + –Cu2 + couple appears promising as a system for further fundamental studies on quantum-cutting and downconversion processes that, for instance, involve a third luminescent center (e.g. another rareearth element). This work also indicates that Cu2 + impurities could limit the efficiency of Gd3 + -activated glasses for phototherapy lamp applications. Moreover, the findings suggest the potential of gadolinium(III) for optical sensing of copper(II), which could be explored for analytical applications.
Acknowledgements The author thanks undergraduate student Martin Garnero from the Chemistry Department at UNF for assistance with the experiments. Keywords: copper · gadolinium · glasses · luminescence · rare earths [1] D. D. Ramteke, R. S. Gedam, J. Rare Earths 2014, 32, 389 – 393. [2] P. P. Mokoena, I. M. Nagpure, V. Kumar, R. E. Kroon, E. J. Olivier, J. H. Neethling, H. C. Swart, O. M. Ntwaeaborwa, J. Phys. Chem. Solids 2014, 75, 998 – 1003. [3] V. Singh, G. Sivaramaiah, J. L. Rao, S. H. Kim, J. Electron. Mater. 2014, 43, 3486 – 3492. [4] Y. Shimizu, Y. Takano, K. Ueda, Thin Solid Films 2014, 559, 23 – 26. [5] C. Tang, S. Liu, L. Liu, D. P. Chen, J. Lumin. 2015, 160, 317 – 320. [6] R. T. Wegh, H. Donker, K. D. Oskam, A. Meijerink, Science 1999, 283, 663 – 666. [7] N. Kodama, S. Oishi, J. Appl. Phys. 2005, 98, 103515. [8] D. Wang, N. Kodama, J. Solid State Chem. 2009, 182, 2219 – 2224. [9] L. Han, Y. Wang, J. Zhang, Y. Tao, Mater. Chem. Phys. 2014, 143, 476 – 479. [10] M. A. Kessler, Anal. Chim. Acta 1998, 364, 125 – 129. [11] C. Cano-Raya, M. D. Fernndez Ramos, L. F. Capitn Vallvey, O. S. Wolfbeis, M. Schferling, Appl. Spectrosc. 2005, 59, 1209 – 1216. [12] M. Turel, A. Duerkop, A. Yegorova, Y. Scripinets, A. Lobnik, N. Samec, Anal. Chim. Acta 2009, 644, 53 – 60. [13] J. A. Jimnez, J. Lumin. 2015, 161, 352 – 357. [14] K. Tanaka, T. Yano, S. Shibata, M. Yamane, S. Inoue, J. Non-Cryst. Solids 1994, 178, 9 – 14. [15] P. Vergeer, T. J. H. Vlugt, M. H. F. Kox, M. I. den Hertog, J. P. J. M. van der Eerden, A. Meijerink, Phys. Rev. B 2005, 71, 014119. [16] J. A. Jimnez, J. B. Hockenbury, J. Mater. Sci. 2013, 48, 6921 – 6928. [17] J. A. Jimnez, J. Phys. Chem. Solids 2014, 75, 1334 – 1339. [18] J. A. Jimnez, M. Sendova, J. Appl. Phys. 2014, 116, 033518. [19] T. Murata, K. Morinaga, Proc. SPIE 2000, 4102, 316 – 323. [20] J. A. Jimnez, J. Mater. Sci. 2014, 49, 4387 – 4393. [21] J. Tauc, A. Menth, J. Non-Cryst. Solids 1972, 8 – 10, 569 – 585. [22] G. Venkateswara Rao, H. D. Shashikala, J. Adv. Ceram. 2014, 3, 109 – 116. [23] F. S. De Vicente, F. A. Santos, B. S. Simes, S. T. Dias, M. Siu Li, Opt. Mater. 2014, 38, 119 – 125. [24] R. Debnath, S. K. Das, Chem. Phys. Lett. 1989, 155, 52 – 58. [25] Y. Tong, Z. Yan, H. Zeng, G. Chen, J. Lumin. 2014, 145, 438 – 442. [26] G. H. Dieke, H. M. Crosswhite, Appl. Opt. 1963, 2, 675 – 686. [27] C. Shen, S. Baccaro, Z. Xing, Q. Yan, S. Wang, G. Chen, Chem. Phys. Lett. 2010, 492, 123 – 126.
Experimental Section Barium-phosphate glasses with a 50P2O5 :50BaO (mol %) composition were prepared from high purity compounds (P2O5 and BaCO3) by the melt-quenching technique.[16] Batch materials were thoroughly mixed and melted in porcelain crucibles at 1150 8C for 15 min under normal atmospheric conditions and immediately quenched. Gadolinium- and/or copper-doping was achieved by adding Gd2O3 and/or CuO quantities in mol %, in relation to network former P2O5. The gadolinium concentration was kept constant at 2.0 mol % in all Gd-doped samples. A glass containing only gadolinium was prepared as a reference, and is referred to as Gd glass. A set of glasses were prepared having a fixed Gd2O3 concentration (2.0 mol %) with different amounts of CuO (0.2, 0.4 and 0.6 mol %) and are referred to as GdCu0.2, GdCu0.4 and GdCu0.6, respectively. The nominal compositions of the glasses are summarized in Table 1. All glasses were cut and polished in order to produce glass slabs with final thicknesses of about 1.0 mm. Optical absorption measurements were performed using a PerkinElmer 35 UV/Vis double-beam spectrophotometer. All absorption spectra were recorded with air as reference. Emission spectra were obtained with a Photon Technology International QuantaMaster 30 spectrofluorometer equipped with a Xe flash lamp having a pulse width of about 2 ms (L4633, Hamamatsu Pho-
&
ChemPhysChem 0000, 00, 0 – 0
www.chemphyschem.org
Received: March 4, 2015 Published online on && &&, 2015
4
0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!
ARTICLES A quantum cut above: Photoluminescence quenching of Gd3 + ions in the presence of Cu2 + impurities in glass is demonstrated. The phenomenon of quantum cutting is proposed to be involved in the energy transfer between Gd3 + and Cu2 + ions. The findings suggest the potential of gadolinium(III) for optical sensing of copper(II), which could be explored for analytical applications.
ChemPhysChem 0000, 00, 0 – 0
www.chemphyschem.org
These are not the final page numbers! ÞÞ
J. A. Jimnez* && – && UV Emission of Gd3 + in the Presence of Cu2 + : Towards Luminescence Quenching through Quantum Cutting?
5
0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
