1. Introduction

Photoluminescence (PL) materials have been widely applied in versatile fields. Rare earth (RE) ions have long been preferentially used as activators in the PL materials due to their rich emission spectra and inter-ion energy transfer phenomena [1,2]. Transition metal (TM) ions, on the other hand, exhibit fluorescence mechanism usually involving both parity and spin forbidden transitions and have been frequently reported in different crystalline and glassy solids [2–4]. Recently, it appears that non rare earth ions activated luminescent materials are among the cutting-edge researches [5,6].

As an important TM ion, divalent manganese (Mn²⁺) plays a key role in phosphors for lighting applications. Mn²⁺ can give rise to a broad emission from green to red under the ultraviolet (UV) or near UV excitation, depending on the interaction of crystal field (CF) with the ⁴T₁(G)-⁶A₁(G) transition [2,3]. Thus Mn²⁺ ion has become an attractive component in developing novel PL materials as primary phosphor in combination with soft UV and blue excitation sources or as secondary phosphors to improve color rendering index. However, the emission of Mn²⁺ is inefficient with long decay time due to the electric dipole forbidden d–d transition [7]. In fact, in the process of some applications, e.g., plasma display panels, the emission intensity and lifetime of a phosphor are of equal significance. Great efforts have been made to improve the PL performance of Mn²⁺ doped phosphors via energy transfer (ET) from co-doped ions. For example, Ce³⁺ and Eu²⁺ have been successfully used as sensitizers to increase Mn²⁺ emission intensity greatly, but the decay time was not reduced [7,8]. There is also an approach to enhance the Mn²⁺ emission through relaxing limitations on doubly-forbidden transition of Mn²⁺ by introducing co-activators or so-called quenchers, mostly TM ions, such as Ni²⁺ and Y³⁺ with which both short decay time and high PL intensity were achieved [9,10].

In the present work, the novel Cu⁺, Mn²⁺ co-doped borosilicate glasses are presented and their PL properties are investigated by optical transmission and PL spectra. The univalent copper (Cu⁺) ion exhibits broad excitation in a UV band and wide emission in the visible region overlapping large portions of the Mn²⁺ excitation bands [9]. In the new pair of dopants consisting of two different TM ions, Cu⁺ ions, besides absorbing UV energy and emitting in the blue-green region, are expected to act as both activator and sensitizers. As an activator, Cu⁺ might split the energy levels of Mn²⁺ due to exchange interaction between them, and change the formerly spin forbidden transition of Mn²⁺ to allowed one, leading to increased transition probability and reducing emission lifetime from millisecond to the microsecond level [9,10]. As a sensitizer, a portion of the absorbed energy by Cu⁺ ions is transferred to Mn²⁺ ions, all mechanisms resulting in enhanced Mn²⁺ emission in the green-red spectrum range. Since emissions from Cu⁺ and Mn²⁺ ions cover all visible region, it is thus believed possible to realize white-light emission by proper Cu⁺, Mn²⁺ co-doping, meeting requirements for white light-emitting diodes (W-LEDs) pumped by UV LED chips.

2. Experimental

Mn²⁺ and Cu⁺ singly and co-doped borosilicate glasses were prepared by the melt-quenching process using chemical purity SiO₂, H₃BO₃, Na₂CO₃, BaCO₃, CuO, MnO₂ as starting materials and SnCl₂ as the reducing agent with the molar ratio to CuO as 4:1. Glass batches were prepared according to the nominal compositions in Table 1, including the host glass (G1), Mn²⁺, Cu⁺ singly doped glasses (G2, G3) and Cu⁺, Mn²⁺ co-doped glasses with different host glass composition (G4, G5, G6).

Table 1. Nominal Compositions of As-prepared Glasses (mol %).

View Table | View all tables in this article

The prepared batches were poured into Al₂O₃ crucibles and moved to an electric furnace for melting in air at 1450-1550 °C for 3-4 hrs and annealed at 550 °C for 2-3 hrs. The glasses obtained were transparent in UV and visible region. All samples were cut into rectangular shapes with 2 mm in thickness and well polished to mirror smoothness to achieve good quality for optical measurements.

Transmission spectra were recorded by a double-beam Lambda 950 UV-VIS Spectrometer (Perkin-Elmer, USA). The PL and PLE spectra were collected by a high-resolution spectrofluorometer (Fluorolog-3, Horiba Jobin Yvon Inc., Edison, NJ) using a 450 W Xe-lamp as the excitation source. The pump light beams were incident and the emitted light beams were collected both at a 45 degree angle to the plate normal with the same slit width for all samples, while reported data were corrected for instrumental response. The PL decay lifetime was measured by FLSP920 (Edinburgh Instruments, Livingston, UK) using nF900 microsecond pulsed Xe-lamp as source with the pulsed width of 2-3 microseconds. An angle of 90 degrees was applied between the light source and the detector. The reflected light signal was detected by the PMT visible light detector after excitation of the sample.

3. Result and discussion

3.1 Transmission and PL spectra

Transmission spectra of G1-G4 are jointly presented in Fig. 1(a). Compared with G1, G2 exhibits two typical Mn²⁺ absorption bands at 353 nm (⁶A₁(S) →⁴E(D)/ ⁴T₂(D)) and 419 nm (⁶A₁(S) →⁴E(G)/⁴A₁(G) and ⁴T₂(G)) [8], G3 suffers a significant red shift of the absorption edge relative to G1, corresponding to the transition of Cu⁺ from ground d state to excited s states [11,12], while G4 is characterized by absorptions of both Cu⁺ and Mn²⁺. For all doped samples, there appear no obvious absorption bands of Cu nanoparticles or Cu²⁺ from 510 nm to 800 nm [11,12], and/or characteristic Mn³⁺ absorption band at around 490 nm [3], indicating that SnCl₂ is an effective reducing agent to make copper and manganese ions exist both in the lower oxide states (Cu⁺, Mn²⁺) in the present glass samples.

 

Fig. 1 Transmission spectra of samples G1-G4 (a) and PL and PLE spectra of glasses G2 (b), G3 (c) and G4 (d). Inset in (a): the digital photos of the samples G1-G4.

Download Full Size | PDF

Figures 1(b)-1(d) present PL and PLE spectra of G2-G4. When excited at 419 nm, G2 exhibits a characteristic green-red emission of Mn²⁺ peaking around 598 nm (⁴T₁(G)→⁶A₁(S)), and so does at 353 nm but much weaker (b). Consistently, the PLE spectrum for the 598 nm emission shows an intense band around 419 nm and a comparatively weak band around 353 nm. For the PL spectrum of G3 excited at 300 nm, a blue-green emission of Cu⁺ ions (s→d) appears peaking at around 482 nm (c), and by reversely monitoring the Cu⁺ emission, a broad excitation band covering 260 nm to 360 nm is observed with a maximum around 300 nm. In the case of G4, it is impressively noted that pumping Cu⁺ at 300 nm produces a super-broad emission band which combines exactly the Cu⁺ blue-green (482 nm) and Mn²⁺ green-red (598 nm) emission bands (d), composing the quasi white-light emission with the chromaticity coordinate of emission as (0.29, 0.34). In particular, the Mn²⁺ emission at 598 nm is enhanced by about 5 times compared with G2 excited at 419 nm, while the Cu⁺ emission at 482 nm is reduced around one time relative to G3. In order to obtain more information on Mn²⁺ emission via Cu⁺ excitation, PLE spectra of G4 for Cu⁺ and Mn²⁺ emissions are compared. The PLE spectrum for Cu⁺ emission at 482 nm in G4 resembles that in G3, while for the Mn²⁺ emission at 598 nm, the UV excitation band is much enhanced relative to the originally stronger visible excitation band at 419 nm, and broadened as well as blue shifted to around 320 nm, close to the Cu⁺ excitation band. Nevertheless, the Mn²⁺ green-red emission under excitation at 419 nm is still nearly one-time enhanced in G4 compared with that in G2.

Hence, the data consistently presents evidence of an enhanced Mn²⁺ emission in association with the presence of Cu⁺ ions in the host glass. There are two possible explanations for the above observations, one is via an ET process from Cu⁺ by absorption energy transfer to Mn²⁺ ions in the UV region, and another is due to an exchange interaction between Cu⁺ and Mn²⁺ ions leading to change of doubly-forbidden transition of Mn²⁺ ions.

3.2. PL decay curves

To demonstrate above speculations, samples G2-G4 are characterized by PL decays as shown in Fig. 2.All the decay curves show strong non-single exponential character; for singly doped samples as G2 and G3, this is a clear evidence of either the presence of different emitting centers or of energy migration due to interaction among them.

 

Fig. 2 PL decay curves of G2-G4: pumped at 300 nm for G3 and G4 (a); pumped at 419 nm for G4 and G2 (b). Note the log-log scale in (b) is used to evidence the fast initial decay of G4.

Download Full Size | PDF

Investigation of the real nature of these processes is beyond the scope of this work. Nevertheless, it is useful to introduce a phenomenological quantification of the decay times of the examined samples. Here we use the equivalent decay time τ_eq defined as [13]

Table 2. Equivalent decay times of the Mn²⁺ (598 nm) and Cu⁺ (482 nm) emissions for samples G2-G4.

View Table | View all tables in this article

The ET efficiency (η_T) can be evaluated using the empirical equation: η_T = 1-τ_s/τ_s0, where τ_s and τ_s0 are the decay lifetimes of the sensitizer Cu⁺ with and without the activator Mn²⁺ [8]. The calculated η_T is around 16%, which seems not high enough to account for the drastically enhanced Mn²⁺ emission.

Also, the emission at 598 nm from the co-doped sample G4 excited at 300 nm is short (τ_eq = 0.040 ms); however, pumping at 300 nm excites both Cu⁺ and Mn²⁺ ions as Cu⁺ has a small but sizeable emission at 598 nm while Mn²⁺ overlaps much excitation source at 300 nm. To disentangle the effect of Cu⁺ emission on Mn²⁺ emission lifetime, the decay of the Mn²⁺ emission in G4 was measured under 419 nm excitation [Fig. 2(b)]. It is observed that in this case the equivalent decay time τ_eq of the Mn²⁺ emission in the co-doped G4 sample (τ_eq = 0.17 ms) is much shortened with respect to that of the singly-doped G2 sample (τ_eq = 0.93 ms) as shown in Table 2. This situation can be explained by considering that the Mn²⁺ luminescent emission is produced by the spin-forbidden transition ⁴T₁-⁶A₁, thus the associated probability is low, giving long (millisecond) lifetime. With Cu⁺ co-doping it is most likely that an exchange interaction between Mn²⁺ and Cu⁺ occurs, changing the nature of the forbidden Mn²⁺ transition to allowed one [9,10]. Consequently, high Mn²⁺ emission intensity together with fast decay time down to tens of microsecond level were concurrently achieved. Similar phenomena have been observed in Ni co-doped ZnS:Mn²⁺ and Y³⁺/Li⁺ co-activated Zn₂SiO₄:Mn²⁺ phosphors [9,10], but scarcely reported for glasses. With this evidence it is convincing that the fivefold time enhanced Mn²⁺ emission of G4 under excitation at 300 nm is due to the joint effects of the ET from Cu⁺ and exchange interaction with Cu⁺.

As the energy level structure of Cu⁺ is related to distances of Cu⁺ ions due to the change of surrounding CF [12], Cu⁺ (S-S) and Cu + (L-W) models are used herein to stand for the long wavelength emission of Cu⁺ with short distance in strong CF and the short wavelength emission with long distance in weak CF, respectively (Fig. 3). ET1 from Cu⁺ to Mn²⁺ ions can easily occur via the excited state of Cu⁺ which is energetically close to the ⁴T₁(P) level of Mn²⁺ ions, resulting in population of the Mn²⁺ excited state ⁴T₁(P) from the ground state. Mn²⁺ ions then undergo multi-phonon relaxation to the emitting level ⁴T₁(P) followed by radiative relaxation to the ground state, emitting at 598 nm. ET2 is also possibly proceeded via the following way: part of the emitted photons from Cu⁺ pumps Mn²⁺ from the ground state to the ⁴T₂(G) and ⁴A₁(G) excited states, and the Mn²⁺ ions in the populated state relax nonradiatively to the emitting state ⁴T₁(P), resulting in the characteristic emission of Mn²⁺ at 598 nm.

 

Fig. 3 Simplified energy-level diagrams of Cu⁺ and Mn²⁺ ions and possible ET processes.

Download Full Size | PDF

3.3. CIE chromaticity coordinate

As characterized by CIE chromaticity coordinate of emission (0.29, 0.34), G4 emits at the edge between green and white lights. To tune the PL of the co-doped sample to the white-light region, the CF on Cu⁺ ions was adjusted by modifying the host glass composition. PLE spectra of G5 and G6 for Cu⁺ and Mn²⁺ emissions show the similarity to G4 (not presented herein) while the PL spectra of G4-G6 are given in Fig. 4(a).G5 containing the higher B₂O₃ content than G4 for substituting Na₂O and BaO suffers an obvious blue shift of the Cu⁺ emission from 482 nm of G4 to 470 nm with the intensity ratio of Mn²⁺/Cu⁺ emissions increasing from 0.65 to 0.70. On the contrary, substitution of BaO for SiO₂ (G6) results in a red shift of Cu⁺ emission to 525 nm while the Mn²⁺/Cu⁺ intensity ratio decreases to 0.60. Correspondingly, the CIE chromaticity coordinate of emission [Fig. 4(b) ] moves from G4 green-white edge (0.29, 0.34) to G6 green region (30, 40) and G5 white-light region (0.30, 0.31) (as shown by the inset image of luminescence from G5 in Fig. 4(a), respectively. According to Cu⁺ (S-S) and (L-W) models it is supposed that G5 offers the weaker CF than G4 on Cu⁺ ions while Cu⁺ ions in G6 are in the stronger CF. In other words, the Cu⁺ emission can be tuned effectively by modifying the host glass composition.

 

Fig. 4 PL spectra of (a, the inset shows an image of white emission from G5 excited at 300 nm) G4-G6 and (c) G5, (b) and (d) are their CIE chromaticity diagrams, respectively.

Download Full Size | PDF

Previous work shows that effect of CF on the Cu⁺ emission can also be exerted via change of excitation wavelength (EW) [12]. Thus with the help of an efficient ET from Cu⁺ to Mn²⁺ in G5, the tunable PL in the white-light region might be possible by altering the EW successively. Figures 4 (c) and d present PL spectra of G5 excited by the light from 280 nm to 350 nm with 10 nm steps and the CIE chromaticity diagram indexing an evolution of PL colors. With the EW shifting to the longer wavelength, the Cu⁺ emission suffers red shift with the intensity reaching the maximum at 290 nm and then decreasing while the intensity ratio of Mn²⁺/Cu⁺ emission becomes much increased. In particular, Cu⁺ and Mn²⁺ emissions become of equal intensity at 320 nm, while moving to 350 nm, the Cu⁺ emission disappears. The CIE chromaticity diagram shows that emissions are almost all within the white-light region and tuned from cold white moving along the green-white edge to the warm white and finally into the yellow region. Thus the tunable white-light emissions are realized in G5 with the lower alkaline earth and alkali oxides by altering the EW, indicating that Cu⁺, Mn²⁺ co-doped borosilicate glasses may be used as converting phosphors for UV LED chips to generate W-LEDs.

4. Conclusion

Enhanced and shortened Mn²⁺ green-red emission is obtained in the Mn²⁺ doped borosilicate glasses by co-doping Cu⁺, composing a super-broad band emission via combination of the Cu⁺ blue-green emission. PLE and decay lifetime data consistently present evidences of an enhanced Mn²⁺ emission in presence of Cu⁺ ions via energy transfer from Cu⁺ to Mn²⁺ ions. Moreover, the Mn²⁺ emission is much shortened from 0.93 ms to 0.17 ms, indicating the change of the spin forbidden transition of Mn²⁺ to allowed one due to an exchange interaction with the co-doped Cu⁺. White-light emission is realized in the co-doped borosilicate glass with comparatively lower alkali and alkaline earth oxides, and can be continuously tuned from cold to warm white by altering the excitation wavelength. Our work indicates potential applications of Cu⁺, Mn²⁺ co-doped borosilicate glasses as converting phosphors for white LEDs pumped by UV LED chips.