1. Introduction

Besides the potential application in high brightness display of a crystal generating the three fundamental colors [1, 2, 3], red emission is also useful in excitation of photosensitive drugs in medicine [4] and in pumping solid state laser based on Cr³⁺, like LiCAF and LiSF [5, 6]. Under one-color diode lasers excitation, red upconversion (UC) emission is usually obtained from Er³⁺/Yb³⁺ or Ho³⁺/Yb³⁺ codoped systems [7, 8]. However, these codoped systems are quite prone to green UC emission, so red UC emission is not efficient nor pure. Eu³⁺ is an unsurpassable red emitting ion in fluorescence lamps, cathode-ray tubes, and projection television tubes under ultraviolet or vacuum ultraviolet excitation due to the high luminescent quantum yield, high resolution, and adjustable wavelength. But its scarce UC property excludes it as a red UC emitting ions in the past for a long time. Single Eu³⁺ UC emissions, which originate from three-photon simultaneous absorption of the Eu³⁺ charge transfer state or from second harmonic generation (SHG) of host under pulse laser excitation, are reported [9, 10]. These processes hardly occur under cw diode laser excitation. On the other hand, a red UC emission in Yb³⁺/Eu³⁺ codoped oxyfluororate glass under cw diode laser excitation is reported [11]. In this work, a red UC emission in Er³⁺/Eu³⁺ codoped LiNbO₃ crystal is investigated when excited with 800 nm femtosecond laser.

2. Experiment

Er³⁺/Eu³⁺ codoped LiNbO₃ crystal was grown along the ferroelectric c axis in air by Czochralski technique. 1.0 mol% of Er³⁺ and 0.1 mol% of Eu³⁺ were added to a congruent ([Li]/[Nb]=0.946) melt. The sample was cut from the middle of the boules without being polarized to y-cut slice, and then polished to optical grade. A regeneratively amplified 800 nm Ti:sapphire mode-locked laser (Spectra-Physics, Spitfire) delivering 1 kHz and ~130 fs pulses in duration was loosely focused onto the sample. The emission was collected by a fiber probe, whose other end was connected to the entrance port of a grating-spectrometer (Bruker Optics 250IS/SM), in a direction normal to the incident excitation beam. The spectrometer output was detected by a thermoelectrically cooled intensified CCD (ICCD) detector (Andor, iStar DH720). Power dependencies were measured by using neutral density filters. The displayed emission spectrum was corrected for the wavelength response of the ICCD detector according to its quantum efficiency curve.

3. Results and discussion

The UC emission spectrum of LiNbO₃ crystal codoped with Er³⁺ and Eu³⁺ is shown in Fig. 1. The emission lines are labeled each with the respective electronic transition involved. The emission is mainly consisted of two emission bands around 550 and 626 nm. It is worth noting that the luminescence intensity of 0.1 mol%Eu³⁺ is comparable to that of 1.0 mol%Er³⁺.

 

Fig. 1. Red and green UC emission spectrum of LiNbO₃:Er³⁺,Eu³⁺ crystal under 800 nm femtosecond laser excitation at room temperature.

Download Full Size | PDF

To understand the UC mechanism the pump energy dependence of the UC fluorescent intensities are investigated. For an unsaturated UC process, the number of photons that are necessary to populate the upper emitting state can be obtained by [12]

where I _UC is the integrated fluorescence intensity, P _pump is the pump energy density, and n is the number of laser photons required to produce one UC photon. The pump energy dependence of UC emission intensities are plotted in Fig. 2. n are obtained from the slope of the graph of ln(I _UC) versus ln(P _pump). In the low pump energy side, n close to 2 indicate two-photon process dominating red and green UC emissions. In the high pump energy side, the slope of 1.06 means the occurrence of saturation for Eu³⁺.

To further clarify the UC mechanism, the decay kinetics of Eu³⁺ and Er³⁺ are measured, and the decay curve of Eu³⁺ around 626 nm are shown in Fig. 3. The decay curve shows a rise at initial stage followed by an exponential decay. The rise can be ascribed to ET [13] or to population from higher energy level. Whereas the energy gap between ⁵D₁ and ⁵D₀ is just 2 quantum of cut-off phonon energy (880 cm^-1) in LiNbO₃ crystal, which leads to very effective multiphonon relaxation from ⁵D₁ to ⁵D₀. The effective multiphonon relaxation from ⁵D₁ to ⁵D₀ is demonstrated by the absence of obvious emission band between 500 and 580 nm in the Eu³⁺:LiNbO₃ emission spectrum under 393 nm excitation (Fig. 2 of Ref. [14]) and the direct exponential decay in Eu³⁺:Sr_0.61Ba_0.39Nb₂O₆ under 532 nm excitation into ⁵D₁ level of Eu³⁺ [15]. So the assignation of the rise to population from higher energy level is ruled out. Based on instant decay under excitation in the ⁵D₁ or higher, the UC mechanisms of two-photon simultaneous absorption or SHG can be ruled out. If the UC mechanisms of two-photon simultaneous absorption or SHG is operative, Eu³⁺ will be excited to a higher excited state directly, and then cascade to ⁵D₀ level without delay, which cannot explain the delay of emission signal in experiment. The UC mechanism of SHG can also be excluded by the different temperature dependence between SHG and red UC emission, as shown in Fig. 4. The spectra obtained at 45 and 160 °C are shown in the insert of Fig. 4.

 

Fig. 2. (Color online) Double logarithmic plot of the pump energy dependence of the UC emission intensities of LiNbO₃:Er³⁺,Eu³⁺ under 800 nm laser excitation at room temperature.

Download Full Size | PDF

 

Fig. 3. (Color online) Time evolution of red UC emission after 800 nm femtosecond laser excitation with pump energy of 50μJ/pulse at room temperature. The open dots are the experimental points and full line corresponds to the best fitting to Eq. (2). The inset shows the Er³⁺ decay curve at 550 nm.

Download Full Size | PDF

So the delay in the emission signal is ascribed to ET. ET from Er³⁺ to Eu³⁺ must occur because the energy level structure of Eu³⁺ excludes the possibility of ET UC emission within Eu³⁺ ions. In view of energy levels matching condition, two channels of ET from Er³⁺ to Eu³⁺ can lead to Eu³⁺ red UC emission under 800 nm excitation: cooperative sensitization (Er³⁺:⁴I_11/2 + Er³⁺:⁴I_11/2 → Eu³⁺:⁵D₂), as occurred in Yb³⁺/Eu³⁺ system [11], and ET [Er³⁺:(²H_11/2,⁴S_3/2) → Eu³⁺:⁵D₁]. The cooperative sensitization from ⁴I_11/2 level under ⁴I_9/2 excitation is possible because the population from ⁴I_9/2 to ⁴I_13/2 will stay at ⁴I_11/2 level for 210 μs. If the cooperative sensitization is operative, the decay curve of Eu³⁺ can be expressed as I(t) = k[exp(-2t/τ _Er(⁴I_11/2)) − exp(−t/τ)] [16]. But using this expression to fit experimental points, the obtained τ _Er(⁴I_11/2) (50 μs) is much shorter than the ⁴I_11/2 excited state lifetime of Er³⁺ (210 μs [17]). This antinomy result excludes the possibility of cooperative sensitization, and the only remained possibility is ET.

 

Fig. 4. (Color online) Temperature dependence of the integrated red UC emission and the SHG intensities and the intensity ratios of red emission to SHG under 800 nm laser excitation. The insets show the emission spectra obtained at 45 and 160 °C.

Download Full Size | PDF

The Inokuti-Hirayama approach [18] is generally used in ET process, the time-dependent acceptor luminescence has been given by [19]

where k is a coefficient related to the efficiency of the emission process, γ ₆ is the ET parameter, and τ _A and τ _D are the luminescent lifetimes of acceptor and donor ions, respectively. The obtained lifetimes from the fitted line in Fig. 3 are τ _D = 25 ± 1 μs and τ _A = 420 ± 20 μs. τ _D is close to the thermally coupled (²H_11/2,⁴S_3/2) excited states lifetime (24.5 μs) (see the inset of Fig. 3). This coincides with ET of [Er³⁺:(²H_11/2,⁴S_3/2) → Eu³⁺:⁵D₁]. τ _A is ascribed to the ⁵D₀ excited state lifetime of Eu³⁺.

Based on the above discussions, our UC emission mechanism is shown in Fig. 5. The Er³⁺ UC emission mechanisms are extensively investigated, and excited-state absorption (ESA) is adopted here due to the absence of delay in the Er³⁺ UC emission decay curve at 550 nm (see the inset of Fig. 3). The photon required for the ESA from ⁴I_13/2 is from the posterior pulse after n (n=1, 2, 3…) ms [20]. The ET from Er³⁺:(²H_11/2,⁴S_3/2) to Eu³⁺:⁵D₁ excites Eu³⁺. Populations of Eu³⁺ at ⁵D₁ level nonradiatively relax to ⁵D₀, from where red UC emission occurs.

According to Ref. [21], we calculate the ET efficiency under Er³⁺ excitation based on emission spectrum from the expression

 

Fig. 5. Energy levels scheme of Er³⁺ and Eu³⁺ and proposed UC emission processes. The solid, jagged, and dashed arrows denote radiative transitions, multiphonon relaxations, and ET, respectively. Ground state absorption is abbreviated to GSA.

Download Full Size | PDF

According to available literature [22, 23], the radiative quantum efficiency of the ⁵D₀ state of Eu³⁺ (η _Eu) is 0.43, the effective fluorescence quantum efficiency of the (²H_11/2,⁴S_3/2) states of Er³⁺ (η _Er) is 0.25, the effective fluorescence branching ratio of (²H_11/2,⁴S_3/2) → ⁴I_15/2 (β _Er) is 72%. α is the coupling constant, which includes the mean emission wavelength and the absolute spectral response of the experimental setup. α1 and α2 are assumed to be the same since the red emission wavelength is in the proximity of green. The estimated transfer efficiency is about 30%. This value can be enhanced via increasing the concentration of acceptor ion (Eu³⁺) since the ET efficiency increases with the concentration of acceptor ion [21].

4. Conclusion

In summary, red UC emission is observed in a LiNbO₃ crystal codoped with Er³⁺ and Eu³⁺ under 800 nm femtosecond laser excitation. This red UC emission is a two-photon process and can be ascribed to ET of Er³⁺:(²H_11/2,⁴S_3/2) → Eu³⁺:⁵D₁. The ET efficiency of about 30% is obtained in the LiNbO₃ crystal codoped with 1.0 mol % of Er³⁺ and 0.1 mol % of Eu³⁺. These results imply that a red UC emission from Eu³⁺ can occur in Eu³⁺/Er³⁺ codoped or Eu³⁺/Er³⁺/Yb³⁺ triply doped systems after Er³⁺ ions are excited to ⁴S_3/2 level by cw diode laser.